Antibiotics in Laboratory Medicine, 6 Ed.

Chapter 13. Evaluation of Antimicrobials in Experimental Animal Infections

David R. Andes, Alexander J. Lepak, Niels Frimodt-Møller

In vitro studies provide important information on the potency and spectrum of activity of new antimicrobials, but animal models form the link between in vitro testing and anticipated clinical results. The results of animal studies suggest appropriate indications and clinical trials, uncover potential toxicity problems, and provide insight into the pharmacokinetics (PKs) of new agents in relation to those of known agents. Furthermore in recent years, animal studies have become crucial in evaluating the importance of bacterial resistance mechanisms and which antibiotics to use against them. Thus, it is essential that new and old agents shown to be of interest following in vitro evaluation exhibit sufficient activity in vivo to justify their continued clinical development. Clinical evaluation guidelines for antiinfective drugs place experimental evaluation of new compounds (or novel combinations or therapeutic modalities) in animals as a prerequisite for clinical trials (1). Specifically, indications of the PKs of new molecules, including their metabolism and pharmacodynamics (PDs) (e.g., the effects of the interaction of drug, host, and infecting microbe, including postantibiotic effects [PAEs] and the efficacy of the drug in animal models mimicking human disease), are required and indeed may assist in the planning of clinical trials of new antibiotics (1).

Due to restrictions of space we have in the present text deliberately focused on experimental animal models, which test some aspect of antimicrobial drugs (antibiotics, antifungals, etc.), that is, PK, PD, or effect, which means that we excluded parts only describing virulence factors of microorganisms, immunologic treatment, or host factors. The reader is referred to the extensive specialized literature concerning such subjects. Further, we have not included nonvertebrate models, because these models in our view cannot evaluate a true humanlike in vivo approach for the effect of antimicrobials incorporating a sensible PK/PD analysis. Following a previous edition of the book by Zak and Sande (3), a subsequent “how-to” book was commissioned (4) that covers the technical aspects of almost every major animal model used in infection research.

When considering in vitro tests, the challenge is to consider whether they are really sufficient indicators of the efficacy and safety of a compound. When considering in vivo tests, the challenges include not only assessing whether such tests are reliable indicators of efficacy and safety but also considering the morality of exploitation of animals in research (5). A “practical-minded” discussion of ethics of animal use is available (6). In any discussion of the ethics of animal experimentation, it is critical to expand the discussion to consider the rights of the afflicted, whose suffering may be alleviated based on information gained through animal experimentation (7,8); there is a certain “cost” involved in not using animals, just as there is a cost in their use. Although open and direct discussion between proponents of each view is apparently scarce, it is the key to resolution of this conflict. However, because the goal of the animal rights or animal liberation activists is the complete discontinuation of animal experimentation (9–11), the opposing positions appear fundamentally irreconcilable. What remains clear is that individual researchers remain responsible for their own conduct. In light of this, readers are urged to carefully consider both points of view in order to resolve, as far as possible, this issue for themselves, as well as to provide for themselves a basis for intelligent discussion with fellow biomedical researchers, the general public, and those opposed to vivisection.

Acceptance of the position that it is ethically justified to perform animal experiments does not solve all of the ethical problems associated with such research (8). In answering these questions, it is clear that the bulk of the responsibility again lies with the individual researchers, but the questions should not be decided in isolation. Ethics committees need to be consulted prior to embarking on any experimentation. In brief, the other questions include the following:

■  Is this experiment necessary and can it answer the proposed questions? (12)

■  Is the welfare of the animals being used in this experiment given due consideration?

■  Are the data gained by such experiments being utilized to best advantage?

Perhaps the most difficult issue is the degree of pain and suffering that animals experience during the course of infection. An excellent review of the recognition of pain and distress in animals and the physiologic basis and consequences of pain has been published (13).

Most governments have passed legislation concerning the use of animals in experiments and require that certain basic requirements be met before such experiments can be carried out. The European Union (EU) has in 2010 renewed and expanded on the recommendations for the use of animals for research (14), which appear to be stricter than the Animals Welfare Act imposed in the United States (15). There has, in Europe, been a general trend toward considerably stricter rules for animal research in the wake of very active animal rights organizations and thereby higher public interest in this issue. The EU directive is based on the three Rs—replacement, reduction, and refinement—and encompasses ruling within the following list of headings:

■  Relevance and justification of the following:

a.  Use of animals including their origin, estimated numbers, species, and life stages

b.  Procedures

■  Application of methods to replace, reduce, and refine the use of animals in procedures

■  The planned use of anesthesia, analgesia, and other pain relieving methods

■  Reduction, avoidance, and alleviation of any form of animal suffering, from birth to death where appropriate

■  Use of humane end points

■  Experimental or observational strategy and statistical design to minimize animal numbers, pain, suffering, distress, and environmental impact where appropriate

■  Reuse of animals and the accumulative effect thereof on the animals

■  The proposed severity classification of procedures

■  Avoidance of unjustified duplication of procedures where appropriate

■  Housing, husbandry, and care conditions for the animals

■  Methods of killing

■  Competence of persons involved in the project

Examples include the restriction in using death as an end point, where the text reads as follows: “The methods selected should avoid, as far as possible, death as an end point due to the severe suffering experienced during the period before death. Where possible, it should be substituted by more humane end points using clinical signs that determine the impending death, thereby allowing the animal to be killed without any further suffering.” The directive includes rules for the methods used for killing of each category of animals used. Death as end point is not allowed for LD- or ED50 determination purposes.

Important issues of reporting of animal studies have recently been raised (16,17). Detailed and relevant reporting is essential for peer review and to inform future research, both to be able to repeat experiments but also to avoid redundant use of animals. Thus, there is also a responsibility of the peer reviewers of studies to impose such rules and demand that missing details are provided. Ideally, scientific publications should present sufficient information to allow a knowledgeable reader to understand what was done, why, and how, and to assess the biologic relevance of the study and the reliability and validity of the findings. A review of animal welfare issues in studies on murine tuberculosis from 1997 to 2009 (17) found that although the quality of the studies had improved in the way of focus on handling of the animals, avoiding mice being unduly exposed to serious, debilitating disease and dying from infection, they found that 80% of papers did not report the method for euthanasia, and information on the sex of the animals used was not available in 34% of the studies. Further, spontaneous death was the chosen end point for 66% of so-called lethal studies. Kilkenny et al. (16) and others (18) have suggested guidelines for publication of animal research, which is a good step forward in this field, although the list is not comprehensive enough (17). The next step forward is for scientific journals to impose such guidelines as a minimal requirement for authors to have their results published.

Standards for animal care differ in their details, and various governments have issued guidelines that are periodically updated. Controlling the factors that contribute to animal health will lead to more uniform experimental results (19). The impact of animal health on the outcomes of experiments using infection models has been reviewed (20,21). Key factors to monitor and control are the environmental conditions the animals are housed under, the adequacy and consistency of the animals’ nutritution, and restriction of access to animal rooms in order to protect the animals and minimize risk to the investigator, particularly when animals are experimentally infected.

Selecting an Animal Model of Infection

The objectives of a particular study form the basis for model selection, but as complete an understanding of the model as possible is needed to ascertain that the model is appropriate to meet these objectives. Furthermore, at least in testing antimicrobial agents, the choice of model may be dependent on the nature of the compound, the quantities available (in the case of medicinal chemistry programs, the compound supply is limiting at the early stages), and the extent of information available (e.g., the spectrum of action, the PKs). For example, medicinal chemistry programs aimed at discovering a new chemical entity effective against a novel target would normally use a mouse screening model (e.g., thigh infection or peritoneally initiated sepsis with parenteral and oral compound administration to answer the question: Which of the many newly synthesized compounds that are active in vitro are orally active in vivo?) early in the program to select compounds for later profiling using the more complex discriminative models needed to provide a basis for clinical trial design (e.g., to determine the PK/PD relationships and demonstrate efficacy in a model closely mimicking a clinical indication so as to discover the best treatment regimen using the selected compound to treat, say, endocarditis alone and in combination with other antibiotics). Therefore, the “ideal” model varies depending on the questions asked. However, the preferred models will be similar to humans in terms of tolerability, drug absorption, distribution, metabolism, and excretion and if possible will demonstrate a pathophysiology of infection similar to that observed in humans. Lastly, using the model should be technically feasible.

Categories of Animal Models of Infection

Animal models of infection have been classified according to the complexity of the model (22). Basic screening models, ex vivo models (where implants such as fibrin clots are infected and subsequently removed for further analysis), monoparametric/polyparametric models (which are similar to screening models but where a single parameter or, preferably, many parameters are examined during the experiment), and discriminative models are models designed to simulate human infection as closely as possible. A discriminative model is ideal if there is a simple technique of infection; the causative organism, the route of entry, and the spread in the body are identical or at least similar to the human equivalents; the tissue involvement and the severity, course, and duration of the disease should be predictable, reproducible, and amenable to analysis; and the response of the model to chemotherapy should be measurable and reproducible.

Reproducibility of Animal Models

The response (e.g., the rate and extent of bacterial growth or the onset of clinical signs) observed in the control group has to be in the same range every time the experiment is performed. Furthermore, the response to reference compounds should be reliable and dose-dependent. It is difficult and potentially misleading to uncategorically rely on response ranges found in the literature for the interpretation of data collected in a new experiment. For this reason, investigators should establish in-house reference ranges, and these should be periodically reconfirmed. There are many other sources of variation that can affect the results of individual animal experiments and thereby complicate data interpretation. Mainly, these are strain and gender differences, age-related changes, animal well-being, and the mode and nature of infection initiation and compound administration.

Limitations of Animal Models

Differences in adsorption, distribution, metabolism, and excretion can be profoundly different, and care has to be taken to study drug doses that are “reasonable” so that the effective dose found in the model is similar to one envisioned for patients; despite the general trend of longer drug half-lives in larger animals, the requirement of large doses in mice is normally a sign that the compound is too weakly active. Ideally, drug exposure should utilize regimens that produce humanlike PK profiles. However, what such a regimen would be is obviously not known for experimental drugs at the preclinical stage. The resulting limitations can be profound and must be taken seriously when performing an animal experiment and interpreting the data obtained. Animal models should be used cautiously, their limitations should be recognized, and only those questions that the models can answer should be asked.

Model Validation

Once the most appropriate model is selected, the next step is to validate model, that is, to establish its suitability mostly in terms of responsiveness to chemotherapeutic intervention (therapy or prophylaxis). The results of the validation tests must delineate the similarities and differences between the disease as it occurs in the model and in humans. As will be discussed later, redefining or modifying the animal model to better reflect the clinical situation, particularly in terms of response to treatment, apparently seldom occurs but can produce dramatic improvements in the predictive value of the model (see Eichaker et al. [23]). Standard drugs are usually marketed drugs with proven efficacy in patients. Due consideration should be given to devising an appropriate therapeutic regimen so that the outcome in the model is similar to outcome clinically (e.g., creating a severity of infection and devising a treatment regimen that results in a “cure” rate that is similar to the clinical cure rate [see reference 23]).

Validation should be seen as a continuous process. This implies that a positive control (reference compound) should be included as an internal standard every time the model is used. A negative control (vehicle treatment) is almost always needed to account for time-related changes or other “hard-to-control” variables during the experiment.

The nature and extent of variability depends on the homogeneity of the experimental animals used and to a lesser extent on the variability of the analytical methods used. Therefore, experimental animals, especially small rodents, are often derived from inbred populations in order to reduce this variability. Moreover, individual variability due to the physical status of the animals or the procedure itself can occur. Adaptation or acclimatization times may vary depending on the animals used and the time of day of the experiment (many indicators of a disease process are subject to individual circadian rhythms). Differences may be in part controlled by stratifying the treatment groups. In the process of stratification, animals are placed into groups (blocks) based on criteria defined before the experiment, and then the groups are randomized with respect to treatments. However, stratification, though it is likely to reduce variability between groups, may increase variability within groups.

Finally, the use of the most appropriate statistical test to analyze the data is critical to successfully conclude an in vivo study. The sizes of the groups of experimental subjects and the exact statistical tests should be chosen based on the expected or desired minimal responses to therapy and the inherent variability of the groups. The downside of these fully justified efforts to minimize variability is that such experiments give no hints of the variability to be expected under clinical conditions, where many factors (e.g., pharmacogenomics, the time of the disease presentation, and the presence of underlying diseases) render extrapolation of preclinical data to the clinical situation more difficult.

Statistics and Experimental Design

The following is extensively based on previous publications (24–27).

Experimental Design

Any experiment comprises experimental units or subjects, questions posed, experimental design, facilities for performing the experiment, and the logistics of material supply and labor, all of which need to be considered carefully. For several reasons, both ethical and cost-related, in vivo pharmacologists have an obligation to consistently use the best possible experimental design. A good experimental design should make efficient use of resources and extract the maximum amount of information from the available material, and the number of animals per group should be neither so small that treatment effects remain undetected and incorrect conclusions are reached nor so large that animals and other resources are wasted. However, it is important to state that if an animal experiment is embarked on, care should be taken to include so many animals in the experiment that a statistically valid answer to the research question can be delivered.

Experimental design involves the following:

■  The question(s) to be answered must be clearly formulated (but the researchers must be prepared to analyze unexpected observations). The researchers must also recognize the constraints that the model possesses in terms of “relevance” (i.e., they must limit the questions to those that the model can really address).

■  Sources of bias must be minimized, normally by including randomization steps.

■  The natural subdivisions of experimental subjects (or the treatments) must be accounted for. This is normally accomplished by using specialized experimental designs.

■  Whenever possible, the numbers of experimental subjects should be the same in each group to facilitate statistical analysis. Plans for dealing with “dropouts” that reduce the number in a group or cause the loss of data (i.e., censored data) should be in place.

■  The researchers must take into account the chronobiologic aspects (e.g., orally administered treatments at night will likely be administered to mice with full stomachs).

■  They must weigh feasibility versus perfection and the logistics of carrying out the experiment.

■  They must plan statistical analyses prior to initiation of the experiment and include questions indicating how large a difference between groups is expected (or desired) and the likely variability of data, both of which are used in estimating the experimental group sizes.

■  The researchers must consider the final display of the data. Complex experiments are not easily amenable to graphic display (normally more than four to five lines on a single graph are difficult to follow, especially during oral presentations), and therefore, it may be necessary to break a single experiment into several parts for effective display of the data.

In a typical experiment, only one set of treatments are administered in order to isolate a single variable. However, whenever possible, the design should allow several independent variables (or factors) to be considered simultaneously. Further, the design should be such that each combination of variables is represented in the form of treatments, preferably with replicates in a single experiment. Experimental designs of this type can be used to look at the effect of individual variables but also at possible interactions between the variables. Such “multifactorial” designs clearly require built-in statistical analysis plans (e.g., two-way or three-way analysis of variance [ANOVA]). Essential to good experimental design is the inclusion of appropriate controls. Vehicle controls and, if needed, separate controls for each treatment formulation (e.g., saline groups as well as an ethanol group if the test compounds are formulated in saline or ethanol) or each route of administration (e.g., intravenous administration and oral administration) should be included. Controls for the ageing of the experimental animals during long-term experiments are often overlooked.

Prevention of bias and appropriate randomization are also important factors to consider. Bias occurs when interfering factors have dissimilar effects on different groups, with the end result that the data are unreliable. The main ways of reducing this are through randomization and elimination of investigator bias through use of blinded treatment and observations. Animals or stratified groups can be assigned to treatments randomly (using random number tables or a computer program). Although somewhat laborious, blinding investigators to treatments by the coding of compounds is recommended, particularly when clinical observation or (histo)pathology is used for the evaluation of effects.

Statistical Analyses

The main purpose of statistical evaluation of animal experiments is to ensure that any findings are not due to chance variation within or between the experimental groups. However, statistical evaluation should also be used to illuminate the data, helping to uncover effects that might otherwise be overlooked. It is critical to understand that statistical analyses do not prove biologic “cause and effect” hypotheses nor do they necessarily prove the biologic relevance of an effect of the compound in question. Although statistical significance may be shown, the effect may be so small as to be of little interest biologically (or clinically). A second purpose of statistical analyses is to show that the experiment was carried out so as to be free from the kind of experimental errors that may compromise the data (e.g., placement of all animals having the highest body weight in one group). Statistical analyses should demonstrate that before beginning of the treatment, the groups were balanced and that during the course of the experiment, no detectable unwanted bias occurred (e.g., high mortality in a treatment group in which the survivors were cured of the disease).

Owing to both the ethical aspects of animal experimentation and the costs involved in performing an in vivo experiment, it behooves researchers to obtain results characterizing the substance(s) in vitro. Tests to consider include the following.

Determination of In Vitro Antimicrobial Properties

The minimum inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) of a substance used against a particular microorganism indicate the inherent susceptibility of the microorganism to that substance. When combinations are proposed, potential interactions (interference or antagonism, indifference, additivity, or synergy) should be evaluated by the use of checkerboard titrations or time kill studies. Determination of the MIC and MBC should be considered the minimum prerequisite before proceeding to in vivo evaluations. Also, the goal should be to test the in vitro activity of antibiotics under conditions likely to exist in vivo because environmental factors (e.g., pH, pO₂, and pCO₂) can dramatically affect antibiotic activity.

Although requiring specialized equipment that is not always available, the following supplemental experiments may be considered. The activity of antimicrobials against intracellular microorganisms should be determined in specialized models of intracellular growth (e.g., Mycobacterium avium growing in J774A cells). Because biofilms have been proposed to be an essential part of the in vivo situation (28–30), and the activity of antibiotics against adherent bacteria can differ from their activity against planktonic bacteria (31), determination of antimicrobial activity against adherent bacteria using the “Robbins device” (32) or other procedures (33–38) may yield insight into the effectiveness of a particular agent. Furthermore, microorganisms may grow at slower rates in vivo than in vitro (29,34), and this may affect their susceptibility to antimicrobial agents (30,34,39,40); this can be appropriately determined by evaluation of antimicrobial activity in chemostats, where the growth rate of bacteria can be controlled. Lastly, in vitro PK models (41,42) may assist in planning dosing schedules, for such models can be used to predict critical parameters of antimicrobial drug concentration and effectiveness (e.g., time above MIC or peak area). However, a problem with in vitro kinetic models is that it is often difficult to create so low elimination half-lives in the in vitro models as found in small animals such as mice. Therefore, extrapolations from dose-kinetic studies may be tested, or multiple dosing in the mice can better simulate the in vitro kinetics.

Preparation of Suitable Formulations for Administration of Antimicrobial Agents to Animals

The formulation of compounds may have a profound effect on the activity of antimicrobials. Practical reviews of formulations are available (43). Although generally not a problem with well-characterized antiinfective agents, the solubility of novel agents may dictate the routes of administration and the in vivo pharmacologic activity of an agent. Generally, the substance should be prepared so that it is in the most soluble form possible. However, specialized delivery systems such as liposomes, depot formulations, or formulations for topical administration may not require highly soluble substances. The choice of formulation ultimately affects the bioavailability and PK attributes of any particular substance, and this should be carefully considered when making comparisons of different chemical classes. In particular, formulations can dramatically affect the oral absorption of compounds and also tissue distribution. Note, however, that formulation effects can differ between species (44), and subtle differences can occur between animal strains (45). Some formulations can also reduce the toxicity of compounds (e.g., cyclodextrin [46]).

When considering oral administration of substances, it is important to recognize that the degree of uptake of antibiotics from the gastrointestinal tract varies between species. For example, in vitro tests using intestinal brush-border vesicles to study the kinetics and inhibition of cephalosporin uptake have indicated that the characteristics of transport of β-lactams by brush-border membranes are similar for rat and human tissues but that rabbit tissues possess distinct properties (47). The plasma PKs of amoxicillin are nonlinear in both humans and rats, but rats have lower oral bioavailability (48,49). These studies showed that mathematical modeling could result in false predictions of human PKs of amoxicillin in humans from rat plasma profiles (48,49) and also showed that precipitation of a portion of the total orally administered dose in rats may occur in proximal gastrointestinal areas, complicating prediction of human oral bioavailability from data obtained from rats. Furthermore, the lower amoxicillin bioavailability in rats is in part owing to degradation within the intestine (50).

In general, however, use of formulations acceptable for use in humans is recommended whenever possible. When used parenterally, some formulations may provide more of a depot of active substance rather than resulting in immediate high blood levels of the antimicrobial. When comparing a few antimicrobials, the best strategy is to use the best formulation for each compound in order to ensure maximal bioavailability. However, this approach may not be useful for large-scale screening. In this case, a standard formulation optimized for the class of substances to be compared should be used; however, one must accept that some substances may fail due to poor formulation and will be considered poorly active in vivo. Adequate formulation of novel antimicrobials is often neglected during in vivo evaluations, and following are suggestions that should be considered prior to performing animal experiments. Aqueous solvents are preferred, and poorly soluble substances can be formed into fine suspensions by sonication. At least in initial evaluations, a fine precipitate, if kept in even suspension, can be well tolerated, particularly with oral application. Note, however, that particulate material, especially when administered parenterally, has altered PKs and bioavailability compared with soluble compounds, and this may complicate the interpretation of the findings. Suspensions intended for intravenous application need to consist of nanoparticles (<100 nm in diameter), and because this requires specialized technology, suspensions made by sonication should generally not be administered intravenously. Poorly soluble compounds can be prepared using a variety of mixed solvent systems. In these cases, the substance is first dissolved in an organic solvent and then carefully diluted in an aqueous solvent system. For example, a widely suitable method for preparing compounds is to dissolve the substance in N-methylpyrrolidine (NMP) and add PEG300 to a final 90% v/v. This procedure can use ethanol (10% v/v) or benzyl alcohol (6% v/v). Mixed solvents may additionally contain tetraglycol (polyethylene glycol monotetrahydrofurfuryl ether), polyethylene glycol 400, or propylene glycol (50% w/v, maximum final concentration). Substances can be dissolved in these solutions (sometimes adding Tween 80 initially) and then diluted in aqueous solutions (e.g., dissolution in PEG300 and then dilution with 0.9% saline to a final 30% PEG30). In all cases, the aqueous solvent should not contain high concentrations of salts, and often, physiologic saline is the best cosolvent. Specialized formulations in cyclodextrins, chemically modified celluloses, Gelucire, liposomes, and so on, have been described previously (51). Due to such constraints in solubility is of paramount importance, that PK studies are performed with different doses in order to evaluate the dosages needed to obtain a possible effect in vivo, that is, serum antimicrobial concentrations need to surpass the MIC for a certain period for the drug to show antimicrobial activity in vivo. In this context, the degree of serum protein binding should be determined prior to PK studies, because only free, unbound drug is active in vivo. Intelligent pilot studies of the PK in the animal to be tested should be performed also for the purpose of avoiding senseless use of animals for useless effect studies.

Stability of Dissolved or Formulated Compounds

Although generally not a problem if compounds are prepared immediately before use, the stability of the compound in solution may become a problem when continuous dosing or prolonged fractional dosing is proposed as an administration technique. Although variations in the biologic activity (MIC/MBC) of the substance when it is stored for varying lengths of time may indicate severe stability problems, chemical assay of the substance (normally by high-performance liquid chromatography [HPLC] or combined with mass spectrum analysis) is perhaps best, given the crude activity estimates achievable by in vitro activity assessments. Determination of biologic activity may not be appropriate for some formulations (e.g., Gelucire), because this would require rescue of the compound from fine suspensions, which may be incomplete.

Administration of infectious agents, antimicrobial compounds, or other substances to animals is the central technique in experimental chemotherapy. Excellent introductions to these procedures are available (52,53). Substances are routinely administered subcutaneously (s.c.), intraperitoneally (i.p.), intramuscularly (i.m.), intravenously (i.v.), orally (p.o.), or topically (e.g., on the skin or in a wound). More specialized methods include intranasal, intratracheal, and intragastric administration and injection directly into the cerebrospinal fluid (CSF) or into the vitreous humor of the eye.

The administration of the antimicrobial compound should take into consideration that the purpose of testing the compound in vivo is to study the process of the compound reaching the infectious site via the bloodstream and subsequent diffusion or transport out of the blood vessels. Therefore, intraperitoneal injection of the drug for treatment of a peritonitis infection is more a direct treatment than a systemic treatment, whereas i.p. administration of a drug for treatment of a thigh infection can simulate i.v. or s.c. administration, because the drug is taken up in the peritoneum to the blood and then distributed to the infectious site.

This subject has been thoroughly covered previously (13,54). Anesthetics and analgesics can be of variable duration of action, and care should be taken to provide a suitable period of anesthesia (neither too long nor too short). Postoperative pain relief should be administered, with compensation made in planning the experiment to allow a degree of “washout” of the substance prior to initiation of the experiment, because analgesics may alter normal host responses to infection or provide a source of unwanted drug interaction. However, in certain cases (e.g., models of infection associated with surgery), the use of analgesics may serve to make the models better reflect the clinical situation.

Certain anesthetics may not be applicable to all types of surgical interventions involved in experimental infections, and this should be considered and experimentally tested prior to widespread application. For example, agents that strongly depress respiratory rates (e.g., pentobarbital) are not suitable for pulmonary infection models involving tracheal exposure, and another anesthetic should be used. Furthermore, due consideration should be given to the stress placed on an animal due to anesthesia (long-term depression of normal body function and disorientation of the animal during recovery) versus the stress placed on the animal by not using anesthetics. For example, most routes of compound administration do not require anesthetics, and their use may place additional stress on the animal.

In vitro activity of a drug, as measured by MIC or MBC, provides a means of comparing potency for antibacterial drugs; however, PK measurements are necessary to ensure an agent will be active at a given site of infection in a mammalian host. The integration of PK, MIC, and outcome gave birth to the science of PD. The main goal of these studies is to examine relationships between antimicrobial concentration at the site of infection and drug activity over time (56–60). PK, as it relates to PD, is primarily measured in terms of elimination half-life (t_1/2), area under the drug concentration curve (AUC), and maximal concentrations achieved (C_max). Practically, this is performed by administering a group of mice the same concentration of drug by the same route and sampling serum or plasma at regular time intervals to determine drug concentration. From this information, the PK parameters of interest can be determined. A further step can then be performed by modeling the relationship of drug concentration and efficacy to determine which PK/PD index best correlates with outcome (for a general review of PK/PD principles, see references 56–59,61–67).

General Considerations

Analytical Method

The analytical method must be both reliable and accurate, thus balancing sensitivity of detection, precision of measurement, and reproducibility of results. Currently, HPLC and mass spectroscopy measurement methods have largely supplanted traditional methods such as bioassay. One caveat that deserves recognition with these newer methods is that they measure the presence and amount of a chemical. They do not necessarily indicate biologic activity, which is included in the bioassay. Appropriate controls should be performed to ensure sample matrices (blood, plasma, serum, or tissue) do not interfere with the analysis and that an appropriate extraction method is available and validated.

Experimental Design

The experimental design must balance the potential use of large numbers of animals, the logistics of completing the experiment, and the PK information that is deemed necessary. In general, pilot studies with a small number of animals are useful to identify optimal sampling approaches. It is important to note the frequency and number of samples can greatly influence the accuracy of results. For example, if a drug has a very short half-life such that PK measurements are planned at 10-minute intervals over an hour, this would prove very difficult in terms of accuracy if one or two lab technicians are responsible for drawing serum on a large number of animals at each time point.

In addition to determining the appropriate time points for sampling and number of animals, one must also determine the dose range and route of administration. The dose range employed for PK studies often utilizes at least three different doses that vary from each other by two- to fourfold. The route is usually determined by the intended route of administration should the drug make it to human trials and is usually oral or subcutaneous in most animal models. The administration of oral drug does bring up the potential influence of food on PK and therefore it must be decided whether animals need to be fasted or not. Again, a small pilot study to determine food effect is often helpful to determine whether or not this consideration is necessary for a full PK study.

Organ Tissue Sampling

Although bloodstream (whole blood, serum, or plasma) measurements of drug concentration are most common, there are situations in which tissue or target-organ drug concentration measurement is necessary. This can include determination antimicrobial activity or drug-related toxicity in specific tissues. Traditionally, antimicrobial activity has been evaluated based on plasma or serum drug concentration levels unless the site of infection is considered sequestered (e.g., brain, CSF, urine, eye, placenta) or for pathogens that are primarily intracellular. In the absence of these aforementioned scenarios, plasma drug concentrations have correlated well with outcome at many infection sites (68–70). However, there is still debate on the merits of tissue-specific drug concentration sampling (68,70). For example, many preclinical PK investigations now include measurement of drug concentration in the epithelial lining fluid (ELF) compartment for investigational compounds in development for pulmonary infections. However, in most studies, the plasma PK concentration-effect relationships (PK/PD) accurately predict the outcome in animal model pulmonary infections. Thus, it is not clear whether ELF PK/PD relationships offer a clear advantage over serum PK/PD. An additional limitation to tissue sampling is in the processing tissue samples. The most common method of processing tissue samples for drug concentration measurement is tissue homogenization (71,72). However, tissues have two distinct fluid components consisting of the interstitial and intracellular compartments. When homogenized, these two compartments are irrevocably mixed. Since the intracellular compartment is usually of larger volume, drugs that concentrate more in the interstitial compartment will appear to be much lower in total concentration than drugs that accumulate in the intracellular compartment (e.g., β-lactams vs. fluoroquinolones). More recently, a technique to determine tissue-specific drug concentrations via microdialysis has been developed (73,74).

Pharmacokinetics Parameters

Several informative reviews on PK parameters have been previously published (75,76). In general, there are five parameters of interest including elimination half-life (t^1/2), apparent volume of distribution (V), total plasma clearance (C_L), absolute bioavailability (BAV) (relative bioavailability can also be useful), and free fraction of drug (nonprotein bound). With the exception of the last, these PK parameters can all be calculated using the plasma or serum drug concentration measurements. Most calculations are now performed with the aid of powerful computer programs that can provide the parameters of interest from the input of raw PK data. As mentioned earlier, separate parameter estimates should ideally be performed for tissue sites when relevant (i.e., sequestered sites of infection).

Free Fraction of Drug

Protein constituents in blood and tissues (chiefly albumin) can, and often do, have a high capacity to bind antimicrobial agents (77,78). Generally, drugs are pharmacologically active, metabolized, or excreted only in their nonprotein-bound state (i.e., free fraction). Therefore, it is often critical to know the level of protein binding in the animal model to determine relative total and free concentrations of drug. It is important to note that the relative amount of protein in circulation and the degree of protein binding can change under certain disease states (79,80). Protein binding can also markedly slow clearance of drugs that undergo glomerular filtration (81) and protein binding can change depending on the host animal model. Therefore, the extent of protein binding should be determined for an antimicrobial agent in each animal model used.

Factors Affecting the Pharmacokinetics of Antibiotics in Animals

Effect of Animal Species on Antibiotic Pharmacokinetics

The host animal species can have profound effects on the PK of a drug. It is often noted that smaller mammals (i.e., rodents) possess much more rapid routes of metabolism and elimination, and therefore, half-lives in these models can be considerably shorter than in larger mammals such as humans (81). Even among rodents, though, PK parameters can differ. For example, moxifloxacin PK profiles differ between mice and rats (T_max 0.25 vs. 0.08 hour, C_max0.137 vs. 0.312 mg/L, AUC 0.184 vs. 0.305 mg × h/L, respectively) following administration of 9.2 mg/kg p.o. (82). The route of administration can also affect drug PK in different animal species, as demonstrated by rifampicin where the t^1/2 in rats was 4.72 hours following i.v. administration, but increased to 9.31 hours following oral administration of the same drug dose (83). This same increase was not evident in mice. Finally, even the strain of animal can affect the PK. For example, BALB/c mice and DBA/2 mice display markedly different serum drug concentrations of itraconazole over time in multiple administration experiments (84). In sum, the aforementioned examples highlight the need to measure PK in each animal model used for preclinical antimicrobial evaluation.

Effect of Infection on Antibiotic Pharmacokinetics

The infection process can have a dramatic effect on the PK of a drug. Perhaps, the most well-known clinical scenario that has long been recognized to demonstrate this effect is bacterial meningitis, where bacterial and host inflammatory-induced damage to the blood–brain barrier produces profound changes in the penetration of antibiotics into this otherwise privileged site (85–87). For example, one of the most commonly relied upon drugs to treat gram-positive bacterial meningitis is vancomycin, which penetrates poorly through an intact blood–brain barrier due to the presence of tight junctions (86,88). However, significant damage occurs to the tight junctions during bacterial meningitis leading to increased permeability. In a study using rabbits, there was a near fourfold increase in CSF vancomycin levels in animals with meningitis versus healthy controls (86). Sepsis can also alter drug PK via a variety of mechanisms including increased volume of distribution and organ dysfunction leading to altered metabolism and elimination (89–91). The translatability of preclinical animal model PK to patients therefore usually includes both uninfected and infected animal PK to determine if the disease state significantly alters drug PK.

Effect of Fever on Antibiotic Pharmacokinetics

The physiologic effects of fever could potentially alter PK of a drug although this is not well studied (92). The reasons for this are likely related to confounders in the febrile model. For example, most febrile models use sepsis or endotoxemia to stimulate the febrile response (93–96). However, PK changes in this disease state may be due to vascular and/or organ dysfunction associated with the sepsis/endotoxemia and not necessarily attributable to fever itself. One PK parameter that is likely affected at higher body temperatures is protein binding, which has been shown to be reduced at higher temperatures (97). Thus, fever could potentially increase the free fraction of drug, which could enhance distribution and microbiologic activity, although could also hasten metabolism and excretion. Outside of protein binding, the effects of fever on drug PK are largely unknown in animal models due to difficulty in the ability to dissociate fever from other confounders induced in the febrile state.

Effect of Animal Age on Antibiotic Pharmacokinetics

As would be expected, age can have a profound effect on drug PK. Recognition of these differences is important; however, the clinical applicability of using age-related PK in an animal model and correlating it to age-related PK in a human is limited. The main reason for this is the need to prove age-related changes in the animal model mimic those noted in humans. For example, plasma PK of five β-lactam antibiotics are markedly different in neonatal versus adult mice (98). Without a study in neonatal humans, it is unknown whether these differences are applicable from the animal model. When differences do occur in the animal model, it can provide the stimulus to study the PK in the age groups the antibiotic is being developed for in humans. However, when age-related differences do not occur in the animal model, it does not necessarily indicate that there are no significant clinical differences in drug PK in different aged humans. With this caveat aside, there are examples of age-related changes in antimicrobial PK in animal models (99,100).

Effects of Various Factors on Antibiotic Pharmacokinetics

Many other factors may affect antibiotic PK. A number of notable examples include animal gender, administration of concomitant drugs, timing of administration in regard to circadian rhythm, presence of organ dysfunction, and genetic background. Each of these may or may not affect a specific antibiotic’s PK properties, and unfortunately, it is not always predictable which antibiotic may possess one or more of these less common influences on PK. Oftentimes, these less common factors are investigated in animal models only after significant differences in PK are noted in different populations of humans.

Pharmacodynamics

One advantage of experimental animal infections is the ability to monitor drug effect over time in an in vivo system. A variety of questions can be addressed in these preclinical models to help direct clinical dosing regimens as well as address clinical problems such as toxicity or drug resistance. The principal analysis tool in most of these studies is PD. There are too numerous to cite examples of how informative and predictive PD animal studies can be. The reader is directed to the many insightful reviews cited here (53,56–59,61,63–67,81) for further information.

PD examines the relationships between an antimicrobial agent and the target pathogen over time. Drug effect can be concentration-independent or concentration-dependent and time-independent or time-dependent. Consideration of these concentration- and time-related activities lead to three common PK/PD indices used to describe optimal drug concentration-effect relationships. They include the 24-hour area under the serum concentration time curve (AUC) over the minimum inhibitory concentration (MIC) ratio (AUC/MIC), the peak serum drug concentration level over MIC ratio (C_max/MIC), or the time that serum drug levels remain above the MIC (T>MIC) over a defined period (usually 24 hours; often, the total T>MIC of several doses during 24 hours is calculated as the percentage covered of the dosing intervals, %T>MIC). Determining which of the three PD indices is predictive of efficacy provides a framework for dosing regimen design. For example, concentration-dependent antimicrobials demonstrate enhanced effect as the drug concentration increases over the MIC. The dosing design that optimizes this activity is administration of large doses infrequently. The concentration-dependent indices, C_max/MIC and AUC/MIC, are the PD indices associated with optimal treatment effect for this dosing design. Conversely, drugs that exhibit optimal efficacy at concentrations near the MIC but lack increased effect as concentrations exceed the MIC are considered time-dependent killing. These agents therefore exert optimal effect when smaller doses are given frequently to keep the concentration relatively stable just above the MIC. The predictive index in this case is T>MIC.

Dose Fractionation

Two common experimental approaches are used to find the predictive PD index of an antimicrobial agent and include dose escalation and dose fractionation. The former examines two important aspects of concentration-effect. The first is the impact of escalating drug concentrations on the extent and rate of organism killing or growth inhibition over time. The second outcome considered is the antimicrobial effect after the drug concentration has fallen below the MIC. The phenomenon of growth suppression following antimicrobial exposure is called the postantibiotic effect. For some compounds, organism growth suppression persists for prolonged periods of time after drug exposure, allowing for lengthening of the dosing interval. The PAE is usually concentration-dependent (i.e., the duration and effect usually increase with higher concentrations or doses). Thus, the activity of drugs exhibiting prolonged postantifungal effects (PAFEs) is best described by the C_max/MIC or AUC/MIC indices.

Dose fractionation is performed by administering the same total dose level while changing the interval of administration. For example, a total 24-hour antimicrobial dose of 100 mg/kg can be fractionated as follows: 100mg/kg q24h, 50 mg/kg q12h, 25 mg/kg q6h, and 12.5 mg/kg q3h. In this situation, each group of animals is receiving the same total daily dose, thus the AUC for each regimen is similar; however, the peak level and T>MIC will vary dramatically. Using this scheme, one can investigate which dosing interval results in optimal efficacy. If regimens using higher, infrequent dosing result in lower burdens, then the PD index predictive of efficacy is related to peak concentrations (C_max). If regimens using frequent, small doses result in lower burdens, then the PD index predictive of efficacy is the %T>MIC. If efficacy is similar in each of the dosing fractionations, then outcome depends on total drug exposure (AUC/MIC). Table 13.1 lists predictive PK/PD indices for commonly used antimicrobial drug classes.

The crucial step to making animal model PD studies clinically relevant, and thus translatable, is to study dosing regimens that mimic human drug concentrations over time at the site of infection. This can be more problematic than it first appears. The first issue is accounting for differences in PK of a drug from animals to humans, which at times can vary dramatically. For example, small mammals, such as rodents, often have much more rapid metabolism and elimination of antimicrobial agents. This can usually be accounted for by adjusting the dosing regimen to more closely approximate drug concentration profiles in humans, for example, by more frequent administration of drug or increased concentration on a milligram per kilogram basis. Other strategies include attempting to slow the metabolism or rate of elimination. The second potential difficulty is that human PK, concentration profiles over time, and potential dosing strategy in humans may not be known at the time of preclinical animal model studies. Thus, it can be difficult to predict what dosing strategy or drug exposure is best to use in the animal model to mimic human drug exposures.

Strategies to Prolong Drug Concentrations in Animal Models

The two most common strategies to attempt to mimic human PK in an animal model where there is rapid metabolism or clearance of the drug is to either directly alter the clearance or provide a means of very rapid drug replenishment by frequent or continuous dosing systems. Impairment in renal function can result in slower elimination of antimicrobials if this mechanism is the major clearance organ (e.g., cephalexin [101]). In mice, this has been accomplished by a single subcutaneous injection of uranyl nitrate (10 mg/kg) 3 days prior to animal infection (102). Uranyl nitrate produces acute tubular necrosis and subsequent stable but decreased renal glomerulofiltration for a duration of 7 days (103–105). As shown by Craig and colleagues (104), the administration of uranyl nitrate to mice in the study of amikacin increased the half-life of the drug, the peak concentration for each dose, and the AUC for each dose when compared to non–renally impaired mice. The resultant PK parameters and concentration-time curves more appropriately simulated human PK. It also led to a 10-fold greater potency (as measured by total daily dose required to reduce 1 log CFU/g tissue) in renally impaired mice. Thus, human-simulated PK in this model was more effective than dosing more frequently in renally sufficient mice. Antimicrobial agents actively secreted by renal tubular cells can be competitively blocked by other compounds that use the same excretion process. An example of this is probenicid, a weak organic acid which blocks the secretion of penicillin and other cephalosporins that are excreted by renal tubular cells (105).

A variety of renal impairment mechanisms have also been reported for rats (101). This includes proximal tubular necrosis induced by cisplatin (one dose at 5 mg/kg IP), papillary necrosis induced by 2-bromoethylamine hydrobromide (one dose at 75 mg/kg IV), glomerulonephritis induced by sodium aurothiomalate (six weekly injections of 0.05 mg/kg IV) or anti–rabbit antibodies to rat glomerular basement membrane (single IV injection).

Continuous dosing of antimicrobials has been used to counteract the effect of rapid antimicrobial clearance that can be marked in small rodents. There are a number of systems that have achieved continuous antimicrobial levels and include tissue cage infusion (106), a variety of pump techniques (107–111), and more recently, sophisticated computer programmable pumps (112). These systems work best from an efficacy standpoint for time-dependent drugs in which the time above MIC is the driving PD index. Roosendaal and colleagues (109) demonstrated an example of this approach and correlation with PD indices. They examined the efficacy of intermittent versus continuous administration of ceftazidime, gentamicin, and ciprofloxacin in a rat endobronchial Klebsiella pneumoniae infection model. Despite subinhibitory serum concentrations, continuous dosing of ceftazidime was more effective than intermittent dosing, in which serum levels exceeded the inhibitory concentration with each dose. The converse was true for gentamicin and less so ciprofloxacin.

Osmotic pumps are advantageous as they can provide a constant release of drug over periods of several days to weeks. They can also overcome interspecies differences in antibiotic elimination. However, limitations involve many prerequisite factors, including the compound must be prepared in a fully soluble and highly concentrated solution, dosing times need to be limited, surgical implant of the device, and necessary stability of drug for days to weeks at 37°C. Additional requirements include the compatibility of the solvent system with the pump, confirmation of consistency in the release of active drug over time, and that no precipitation of the compound occurs on the outside of the pump after coming into contact with biologic fluids. These latter requirements can be determined by immersing a filled pump in a suitable isotonic buffer containing 5% to 50% fetal bovine serum and incubating it at 37°C followed by periodic sampling.

Continuous Dosing to Mimic Human Pharmacokinetic Profiles

This subject has been thoroughly reviewed by Mizen (113). Continuous- or variable-rate infusion of antibiotics into animals has been used to obtain plasma antibiotic clearance similar to that found in humans administered bolus or drip infusions. Based on careful determination of temocillin PKs in both humans and rabbits, a continuous-rate infusion system was developed to deliver temocillin to rabbits with meningitis due to K. pneumoniae in such a manner that human plasma elimination rates following a 2-g dose were obtained (114). The femoral artery was cannulated to allow continuous infusion of antibiotic. Rabbits received first a bolus dose (82 mg/kg) to mimic the temocillin distribution phase and then an infusion of continuously diluted temocillin to mimic the β-elimination phase observed in humans. Phosphate-buffered saline was administered at a constant rate (equivalent to the human temocillin elimination rate corrected for the rate of temocillin elimination by rabbits) into a fixed-volume, stirred reservoir containing temocillin such that the concentration of temocillin infused into the animal was constantly declining. Rabbits receiving infusions were treated with a total of 758 mg/kg temocillin over 12 hours at 2.0 mL/hour. Compared to bolus dosing (82 mg/kg), infusion dosing resulted in a dramatically prolonged temocillin half-life and larger AUC values without altering the percent penetration into the CSF. Note that, after a 2-g dose to humans, the plasma half-life was 5.0 ± 0.2 hours and the AUC was 784.5 ± 47.1 µg × h/mL. However, considering the serum binding of temocillin (60% to 85% in human serum, depending on the temocillin concentration; 35% in rabbit serum, concentration independent), this mode of temocillin administration to rabbits would result in free temocillin concentrations similar to that observed after a 4-g dose to humans. Humanlike PKs resulted in a dramatically improved therapeutic outcome, in that K. pneumoniae was rapidly removed from the CSF (to less than log 2 CFU/mL within 6 hours) during infusion but remained essentially unaltered during bolus dosing. The authors did not evaluate the effect of the same total temocillin dose in bolus infusion (114).

This infusion method was adapted to rats in order to mimic the human plasma PKs of cefazolin, piperacillin, and the β-lactamase inhibitor BRL 42715 (115). Rats were infected intraperitoneally with either Escherichia coli or Serratia marcescens, and treatment began 1 hour after infection; both microorganisms demonstrated susceptibility to cefazolin or piperacillin only in the presence of BRL 42715 (concentrations more than ~0.1 µg/mL). Simulation of human plasma PKs was obtained. The half-lives for BRL 42715, cefazolin, and piperacillin in humans were 0.6, 1.6, and 1.1 hours, respectively, considerably different from those in rats (0.1, 0.51, and 0.33 hours, respectively). Despite the plasma concentration of the β-lactamase inhibitor falling below the level predicted to be effective within 3 hours, coadministration of BRL 42715 with piperacillin (E. coli infection) or cefazolin (S. marcescens infection) dramatically improved efficacy in terms of survival and bacterial counts in blood and peritoneal fluid (115); these results indicate that β-lactamase inhibitors need not have plasma PKs identical to those of their partner antibiotics in order to have synergistic effects. Comparison of the effectiveness of human-simulated PKs and the efficacy of these combinations in bolus administration (rat PK) was not reported (115).

A similar approach was used to mimic the PK of 3- and 0.1-g doses of ticarcillin/clavulanic acid and a 2-g dose of ceftazidime administered to humans (116) in rabbits with meningitis due to K. pneumoniae. The infusion system was modified to include two pumps, one infusing a constant dose of agent for a short time (to produce a peak serum concentration similar to that seen in humans) and the other constantly infusing a continuously diluted solution of drug (to mimic the serum elimination PKs manifest in humans). This system was successful in overriding the more rapid elimination of ticarcillin, clavulanic acid, and ceftazidime by rabbits. Single doses of ticarcillin/clavulanic acid were able to reduce (by 99% at 4 hours) but not eliminate the drug combination–susceptible K. pneumoniae present in the CSF due to regrowth of the organisms after clavulanic acid levels fell below the MBC. Multiple dosing of ticarcillin/clavulanic acid (three doses every 4 hours) according to simulated human PKs resulted in higher AUCs in both plasma and CSF (without altering the CSF penetration) and correspondingly greater antibacterial efficacy (99.99% reduction of colony-forming units [CFU] per milliliter in the CSF). Two ceftazidime doses (every 8 hours) reduced the counts of the drug-susceptible microbe below the limit of detectability and sterilized the CSF in two of three rabbits at 12 hours. The efficacy of bolus doses of these drugs (rabbit PKs) was not reported (116).

A computer-controlled, variable-speed pump was used to mimic human serum concentrations of amoxicillin in a study to determine the effectiveness of amoxicillin prophylaxis in preventing streptococcal endocarditis (117). Sterile aortic vegetations were produced by placement of a polyethylene catheter through the right carotid artery across the aortic valve. Amoxicillin was administered by infusion through a Silastic catheter placed into the jugular vein and brought through the skin of the intercapsular region. Intravenous infection with Streptococcus intermedius or Streptococcus sanguis (1, 10, or 100 × the 90% inhibitory dose) occurred 1 hour after administration of 40 mg/kg amoxicillin (rat PKs) or amoxicillin dosage to mimic human PKs following a 3-g oral dose. At the time of bacterial challenge, the serum antibiotic levels were similar (bolus dose, 18 ± 0.3 µg/mL; infusion, 16 ± 5 µg/mL), but the durations of detectable amoxicillin levels were different (bolus, 4.5 hours; infusion, 9 hours). Simulation of human serum PKs was decidedly more effective than bolus dosing in the protection of rats from developing endocarditis (117). A similar procedure was used to deliver ceftriaxone to obtain humanlike PKs in rats with Streptococcus sanguis or Streptococcus mitis (118) or methicillin-resistant Staphylococcus epidermidis (119) endocarditis.

Use of animal models in the evaluation of antimicrobial compounds is considered when a clinical study involving humans is not possible because (a) the toxicity of the compound is unknown, (b) its antibacterial ability in vivo is unknown, (c) the type of infection under consideration is rarely encountered or impossible to encounter in humans, or (d) the effect parameters needed (e.g., bacterial counts in tissues or fluids) cannot easily be obtained in patients. Otherwise, a clinical study will be the optimal method for studying any drug for clinical use, since it automatically answers the question (which is always asked after an experimental animal study), “Can the results be extrapolated to the clinic?” Animals are always used when new compounds have shown relevant antimicrobial activity in vitro and their in vivo effects are questioned. Furthermore, experimental animal infections are considered when other important issues need to be solved prior to clinical studies, such as the advantages or disadvantages of the compound in activity, its PK profile (i.e., its dosing advantage), the best mode of administration (i.e., oral or parenteral), and potential toxicity problems (diarrhea, nephrotoxicity, etc.). Given the ethical considerations and government legislation involved, the testing of novel compounds, or novel combinations of known compounds, requires comparative testing in animal experiments in order to indicate efficacy in vivo. As in the evaluation of antimicrobials from known classes (e.g., new cephalosporins), the questions include not just whether the agent will be active in vivo but how it will compare in spectrum and potency to other antibiotics. β-Lactamase inactivation seen in vitro might not occur in vivo. In many instances, very low MIC values in vitro are not reflected by the in vivo results. Thus, for new compounds or derivatives of known antibiotics, the in vivo test is also a tool for selecting the potentially best candidate from a number of active agents. The following discussion mostly concerns the evaluation of antibacterial or antifungal substances in animal models, although the evaluation of antiviral compounds is not covered. Furthermore, for purpose of simplicity and because they can be obtained elsewhere (3,4), details on the establishment of models have been reduced to a minimum.

General considerations for working safely with infectious agents have been summarized in Richmond and Quimby (121), and readers are encouraged to read this review prior to embarking on establishing animal models of infection in their laboratory.

Factors Influencing Antimicrobial Activity in the In Vivo Tests

A number of the factors can influence the activity of antimicrobial agents in vivo (e.g., see 122):

■  Inoculum size and vehicle. If the inoculum is too small, an infection will not be established; if it is too large, overwhelming endotoxemia can occur.

■  Virulence or pathogenicity of the infecting strain. Highly virulent strains may produce rapidly fatal disease, necessitating early treatment initiation.

■  Generation time in vivo. Slow-growing bacteria in vivo are phenotypically very different from fast-growing bacteria in vitro. Similarly, biofilm growth in vivo is different from planktonic growth in vitro.

■  Timing of treatment. A delay in treatment initiation often results in greater difficulty curing the infection. This is often related to the inoculum, see the following text.

■  Method of antimicrobial administration. The lack of oral uptake of a drug highly active in vitro may render it inactive in vivo.

■  The PK/PD of the antimicrobial. Generally, more rapidly eliminated drugs need to be administered more frequently.

■  The development of resistance in vivo. Unique resistance patterns in vivo may render a drug highly active in vitro inactive in vivo.

■  In vivo growth of an intracellular compartment that the drug cannot penetrate.

■  Inactivation of the compound in vivo. The host metabolism may render a drug that is highly active in vitro inactive in vivo.

Inoculum size has a major influence on the in vivo activity of antimicrobials. An increase of 1 log unit or even less in the challenge bacterial dose can render an antibacterial ineffective. In addition, virulence or pathogenicity is closely connected to the inoculum size, since high virulence to a particular animal species often leads to lower inocula being used in order not to induce an overwhelming infection. On the other hand, higher doses are still needed to achieve an effect against a highly virulent strain, compared with strains with lower virulence. For example, the heavily capsulated Streptococcus pneumoniaeserotype 3 (penicillin MIC, 0.01 mg/L) has an LD₅₀ of 10² CFU for intraperitoneal infection in CF1 mice, in comparison with 10⁷ CFU for S. pneumoniae serotype 6B (penicillin MIC, 0.01 mg/L). Still, the ED₅₀ for single-dose benzylpenicillin against serotype 3 is 180 mg/kg, in comparison with 0.8 to 2 mg/kg against serotype 6B strains (123,124).

For Streptococcus pyogenes, increasing the in vitro starting inoculum (from log 3 to 7 CFU/mL) had no effect on the MIC of cefoxitin (0.5 µg/mL) and mezlocillin (0.05 µg/mL). In contrast, increasing the inoculum of K. pneumoniae from log 5 to log 8 CFU/mL had no effect on the MIC of cefoxitin (4 µg/mL) but dramatically altered the MIC of mezlocillin (4 µg/mL at log 5 or 6 CFU/mL, 32 µg/mL at log 7 CFU/mL, and >128 µg/mL at log 9 CFU/mL).

Preparation of the microorganism for inoculation in animal experiments can dramatically affect the results obtained, and this fact is often overlooked. Fundamentally, the microbe should be at maximal viability, and care should be taken to obtain suitable cultures for inoculum preparation. Whether the microbe should be taken from the in vitro culture in lag phase, log phase, or stationary phase has to our knowledge never been validated. The use of overnight broth cultures may be problematic for bacteria such as S. pneumoniae or Haemophilus influenzae, which undergo autolysis shortly after reaching the stationary phase (125). Agar plate cultures have the advantage that they can be directly studied to determine if any contamination has occurred and whether loss of potential capsule has occurred. Many investigators prefer to bring the microbe into the exponential growth phase in a broth culture prior to its use or inoculation. None has proven, however, that the bacteria will stay in this exponential phase after the procedures used (e.g., washing) to achieve the exact inoculum needed.

An increase in the virulence of certain microorganisms can be achieved in various ways. In vitro growth in specialized media can alter virulence. For example, growth of Neisseria meningitidis under conditions of low pH and low growth medium iron content increases the virulence of this organism 1,200-fold, relative to bacteria grown in neutral-pH, iron-replete medium (126). Growth of uropathogenic Escherichia coli in human urine increased siderophore production and renal pathogenicity in ascending pyelonephritis in mice (127). Furthermore, virulence-associated gene expression by Enterococcus faecalis is modulated during the growth phase and affected by the growth medium (128). Using the guinea pig subcutaneous chamber model as a test system, iron-limited gonococci were found to be extremely virulent, whereas cystine-limited (iron-replete) gonococci did not survive in the chambers despite retention of pili. Loss of piliation also occurred during the shift from iron-limited to glucose-limited growth, but the bacteria remained virulent. No change in susceptibility to normal human serum killing occurred, and the lipooligosaccharide composition remained similar despite varied culture conditions. Some membrane proteins traditionally associated with iron limitation were produced by cystine- or glucose-limited bacteria (129). Note, however, that iron restriction apparently does not affect all bacteria. The rate and extent of in vitro growth of Salmonella typhimurium are unaffected by the addition of deferoxamine, and treatment of mice with deferoxime prior to infectious challenge exacerbates Salmonella typhimurium infection (130). Virulence can also be enhanced by serial passage in animals; for example, intraperitoneal inoculation with subsequent subculture from peritoneal wash, blood, or organs such as the liver or spleen and use of this growth either directly or after subculture will enhance the virulence of Streptococcus pneumoniae, Streptococcus pyogenes, or H. influenzae. In spite of great care taken to standardize the inoculum, one of the major problems encountered in animal experiments is the variation in the virulence and growth of organisms. This problem highlights the need for control groups for every new infection experiment considered rather than relying on historical controls.

Generation time (rate of bacterial cell division) is also a factor of major importance that differs between in vivo and in vitro test conditions. In vivo, the generation time seems to increase progressively during the course of infection and, depending on the site of infection, may last up to 20 hours. Little is known about nutrient limitations on bacterial growth in infected tissues, with the exception of iron, which has been found to limit bacterial growth in serum. It has been demonstrated that prolonged lag phase as well as prolonged generation time may adversely affect the clinical activity of antimicrobials (especially β-lactams) that are most effective against bacterial cells that are rapidly dividing (31,34,39,118). The nature of bacterial growth in vivo in tissues is not well studied. Good evidence is available indicating that bacteria grow as biofilms in many urinary tract infections (131,132), cases of otitis media (133), catheter-related infections (134–136), and pulmonary infections (131,137–139). For a review, see Costerton et al. (28,29). Further, bacteria growing as biofilm have a dramatically different physiology and antibiotic sensitivity (30). A set of genes are specifically activated during the establishment of a biofilm, both in vitro and in vivo (140). However, not all bacterial growth in vivo occurs as biofilms (141). Multicolor fluorescence microscopy has been used to delineate the nature of Salmonella enterica in the livers of infected mice (142). The growth of Salmonella occurred by the formation of new foci of infection from initial ones as well as by the expansion of each focus. Each focus consisted of phagocytes containing low numbers of bacteria and of independently segregating bacterial populations. The net increase in bacteria paralleled the increase in the number of infected phagocytes in the tissues (142).

Mouse Protection Test

The animal models most frequently used in the evaluation of antibacterials may be categorized as basic screening, ex vivo, monoparametric, or discriminative (143,144). For the preliminary evaluation of new agents, the basic screening system is usually employed. The ex vivo and monoparametric models are used to measure specific variables (e.g., dosage schedule, serum binding, or penetration into extravascular spaces). The discriminative systems are employed to differentiate the new agents from related or unrelated active agents. Screening models involve simple one-step infections, simple techniques and schedules of treatment, short-duration experiments, reproducible courses of infection, simple evaluation (all-or-nothing models), economy of test drugs, and low costs. These requirements are best met by the mouse protection test, which is the most widely used in vivo screening model in antibacterial research. The mouse protection test is suitable for determining the efficacy and toxicity of new antibacterials, and it can indicate whether a drug is likely to be active orally or parenterally. The features and use of the mouse protection test have been previously reviewed (145–147).

Various mouse strains are the primary hosts used for the following reasons: (a) good correlation between the clinical response to an antimicrobial agent and the agent’s activity in mice; (b) the ease of obtaining large numbers; (c) the economy of the compound to be tested; (d) the relatively small cost per unit test; and (e) normal use of an outbred strain of mice, which provides a heterogenous population and allows for immunologic and other host factor variations (however, in special situations [e.g., Mycobacterium infection models], inbred, genetically defined strains may be required in order to provide a suitably susceptible host). It should be clearly recognized, though, that the mouse protection model represents an unnatural infection in which the host is usually subjected to an overwhelming challenge (148).

Correlation of In Vitro and In Vivo Results

This subject—the correlation of in vitro and in vivo activity—has a long history (149) and is critical for a medicinal chemistry program, given the costs of in vivo screening.

Many substances that are active in vivo are also active in vitro; however, the converse is not always true. Many antibacterials that are active in vitro either are inactive when tested against systemic infections in vivo or are only active in the more sensitive topical infections. Zak and Sande (150) reported a correlation of in vitro and in vivo activity in only 14.8% of 2,000 compounds randomly screened for antimicrobial activity. Of the 2,000 compounds, 45.3% were inactive in vitro and in vivo. Of those inactive in vitro, 0.3% showed in vivo activity. Of those active in vitro, 36.6% were inactive in vivo. Of the total, 14.8% displayed activity both in vitro and in vivo. Given that currently only compounds with in vitro activity are tested in vivo, recalculation based only on those having in vitro activity would change the figures to 73% for in vitro activity only and 27% for activity in both tests. The latter percentage is typical of those noted by investigators. Furthermore, a major problem in correlating in vitro and in vivo results occurs when the agent being tested is very active in vitro but inactive or moderately active in vivo. Because in vitro and in vivo tests differ in their general characteristics and specific variables, discrepancies are likely to occur. However, they may be understood and interpreted if the limitations of the tests are taken into account. Causes of missing activity in animal experiments in spite of good activity in vitro can include PK factors, such as minimal distribution in the host due to poor uptake and rapid metabolism or other inactivation (e.g., high protein binding) of the drug, resulting in insufficient dosing. Differences in the pharmacology of compounds between humans and animals commonly used for experimentation can lead to effective drugs being wasted because their potential clinical effect is never tested. Beneficial effects of drugs other than their antimicrobial activity (e.g., immunostimulative behavior) have a risk of being overlooked if they do not reach in vivo testing (e.g., the potential immunity-stimulating activity of the macrolides has only recently been detected using experimental animal testing).

Commonly used in vitro tests do appear to fail to predict outcome in certain types of infections, especially device-related infections (151). Characteristically, bacteria involved in device-related infections are adherent, slow-growing bacteria that are phenotypically distinct from the rapidly multiplying bacteria that grow during in vitro susceptibility testing (see reference 30). Consequently, specialized in vitro techniques are needed to obtain a better correlation between in vitro and in vivo (experimental or clinical) activity. Using a model of S. aureus device-related infections (subcutaneous chamber implant) in guinea pigs, Zimmerli et al. (151) found that, as a single agent, only rifampicin was active, in contrast to vancomycin, teicoplanin, ciprofloxacin, and fleroxacin, despite the fact that the S. aureus strain was sensitive to all compounds in vitro using standard tests. Determination of peak and trough drug levels in tissue cage fluid showed that, at the doses given, the drug levels of rifampicin, vancomycin, and teicoplanin exceeded the MIC constantly throughout the 4-day experiment. Further experimentation demonstrated a dramatic loss of drug activity against stationary phase bacteria (the minimal loss occurring with rifampicin) and that testing the killing effect of antibiotics and their combinations against bacteria adherent to glass beads at drug levels achieved in vivo did provide an accurate prediction of treatment effect in vivo.

Anaissie et al. (152) and Rex et al. (153) have studied the correlation between in vitro susceptibility and in vivo activity in a Candida sepsis model in mice. Lack of in vitro susceptibility in a microbroth dilution test correlated well with fungal kidney colonization 4 days postinfection and with prolongation of survival (152). Follow-up studies indicated that MICs determined at 24 hours (as opposed to 48 hours) correlated better with in vivo outcomes (153).

Acute Toxicity Assays to Determine Tolerated Doses

Important technical issues that must be considered include the amount of drug administered during a primary screening program and the most suitable route of administration. If possible, some measure of toxicity should be obtained in vitro. Although more often done only if anomalous results are obtained, before use in the treatment of infected animals, the maximum tolerated dose of a substance should be determined by administering single injections of the substance to groups of mice (N = 3 to 6) by oral, subcutaneous, and intraperitoneal routes. The animals are then observed for survival for periods from 24 hours to 7 days. Care should be taken to observe the mice continuously after drug administration and note their clinical condition in order to kill them in a humane way before they die from the toxic activity. Such killed mice are still counted as dead from toxicity. These acute toxicity studies establish for each route the 100% toxic dose (LD₁₀₀), the 50% lethal dose (LD₅₀), and the maximum tolerated dose (LD₀)—the dose at which all animals survive. Various methods of determining the LD₅₀ (or infectious dose) have been previously reviewed (154). Approximately one-fifth of the maximum tolerated dose of a substance can be well tolerated when treatments are given once daily for 5 days or longer. For 1 to 3 days of treatment, one-half of the maximum dose can usually be given. When using these crude guidelines, it would be reasonable to assume that animals dying after multiple treatments succumb to the effects of the particular infection rather than to drug toxicity. The use of LD₅₀models is currently under debate regarding the mortality in the high-dose groups (14). Toxicity models where toxic effects are analyzed by histology or other means should be preferred to simple LD₀ experiments, because this may decrease unnecessary harm to the animals (14).

In addition to initial information on toxicity, some information on oral bioavailability may also be obtained. For example, if a substance is tolerated at 1,000 mg/kg when given orally but is toxic when given at a dose of 50 mg/kg intraperitoneally or intravenously, the lack of toxicity by the oral route probably reflects poor oral absorption.

Choice of Organism

In developing an experimental mouse model for in vivo testing, it is desirable to use human pathogens whenever possible. It is also desirable to infect with strains of microorganisms that are sufficiently virulent so that conditioning procedures to lower the host’s resistance are unnecessary. Natural infections typically result from inoculation with Streptococcus pneumoniae, Streptococcus pyogenes, certain strains of K. pneumoniae, Salmonella typhi, Salmonella typhimurium, Mycobacterium tuberculosis, and Cryptococcus neoformans. When reproducible infections cannot be achieved by inoculation of the organisms alone, it is necessary to reduce the resistance of the animal. A common procedure is to suspend the organism in 3% to 10% hog gastric mucin and to administer 0.5 mL amounts by the intraperitoneal route. In the case of infections with Candida albicans and Histoplasma capsulatum, the animals are infected intravenously with virulent strains; for less virulent strains, the animals are conditioned by injection of 0.1 mL of a 1% suspension of cortisone in saline twice daily by the intramuscular route before the introduction of the organisms. Alternatively, immunosuppression can be achieved by rendering the mice leukopenic by administering cyclophosphamide.

Preparation of Inoculum for Infection: Virulence Titration

In order to obtain reproducible infections, it is necessary to determine the degree of virulence of each strain to be studied. To carry out virulence tests, suitable broth cultures (where high viability of the culture is maintained) are serially diluted in broth to obtain 10-fold decreases in the number of organisms. If mucin is to be used, one part of each broth dilution is combined with nine parts of mucin. Groups of four to six mice are injected intraperitoneally with 0.5 mL of each dilution. Samples are taken from peritoneum, blood, or organs such as liver and spleen and bacteria quantified in order to construct in vivo growth curves. Concomitantly and before sampling, mice (or control groups of mice) are evaluated for symptoms and signs of systemic infection (for rating schemes, see reference 14). When animals are considered to progress to clinical stages, which will lead to death of the animals, the animals are killed and quantification of bacteria in vivo ascertained and compared to the in vivo growth curves. In this manner, a surrogate LD₅₀ (or LD₁₀₀) can be determined, which will enable an intelligent choice of the inoculation strategy for that pathogen/mouse combination. The original calculation of an LD₅₀ from mortality of mice has increasingly been replaced by bacterial quantification studies (154).

The virulence of many organisms is so low for an unnatural animal host that some type of stressing agent or adjuvant is usually required to achieve a reproducible infection. As previously stated, mucin is usually required to provide reproducible bacterial infections in mice. Without mucin, infections would require the large numbers of organisms provided by cultures that are undiluted or only marginally diluted. Such an inoculum may be overwhelming, either because of toxic effects (e.g., endotoxic shock or the introduction of toxic components of spent culture broth) or because the antibacterial would not be able to inhibit the large numbers of organisms (see reference 155). We have found in the past that mucin, at a concentration of 5%, gives consistent results with few deaths directly attributable to the effect of mucin, although with certain batches of mucin, an 8% to 10% concentration may be required to give consistent results. The quality of the mucin available is variable, and thus the mucin used should be evaluated in separate experiments. One problem often encountered is contamination of the commercial hog gastric mucin powder. The mucin can be autoclaved without loosing its macrophage-inhibiting abilities. Furthermore, in all experiments, or at least periodically, a group of uninfected animals should receive mucin alone to ensure that deaths are not due to the stressing effects of this adjuvant.

Previous studies (156) clearly show the enhancing effect of mucin on the proliferation of bacteria in the murine host. Mice were infected intraperitoneally with 0.5 mL of an overnight broth culture of E. colieither as a saline suspension or in 3% gastric mucin. Groups of five mice were killed, samples of blood were collected from the axillary region at 10 minutes and at hourly intervals after infection, peritoneal lavage was performed, and CFU/mL determinations were made of both fluids. A count of 10⁵ CFU/mL was obtained from intraperitoneal washings immediately after infection with bacteria suspended in saline. The count dropped to approximately 10³ CFU/mL within 10 minutes and remained in this range for the next 7 hours. The viable count in the blood rose to approximately 50 CFU/mL within 10 minutes and then showed little increase over the next 7 hours (approximately 100 CFU/mL). In contrast, when 3% mucin was used as an adjuvant, the initial count in the peritoneal washings of 10⁵ CFU/mL increased stepwise with time to a count in excess of 10⁹ CFU/mL by the end of 8 hours. The viable count in the blood closely paralleled that seen in the peritoneal washings (increasing from 5 × 10³ CFU/mL after 1 hour to 7 × 10⁸ CFU/mL at 8 hours).

For certain slowly growing, fastidious organisms (e.g., S. pyogenes and S. pneumoniae), mucin is not required in order to obtain a reproducibly virulent infection. The virulence of these organisms can be maintained by passage of the cultures in mice. One or two mice are infected intraperitoneally with 1 mL of an overnight broth culture. After 6 to 8 hours, when the animals show signs of illness, they are anesthetized, the hearts are removed aseptically, and several drops of heart blood are added to a tube of appropriate broth medium. For S. pyogenes and S. pneumoniae, trypticase soy broth containing 10% serum is suitable. The serum of any animal species is suitable. At the same time, the blood is also streaked on blood agar plates. After overnight incubation, serial 10-fold dilutions of the broth cultures are prepared for use as the infecting inoculum. The blood agar plates are used for confirmation of the purity and identity of the infecting inoculum. The quellung reaction can be used to type the pneumococci, and standard procedures are used to confirm the identity of group A streptococci. In vivo passage can dramatically affect the virulence of many other pathogens (e.g., some strains of H. influenzae), and this method should be considered for all strains proposed to be used for many experiments. Following in vivo passage, stocks of the organisms should be made from exponential-phase cultures of low passage and then stored frozen. If available, liquid nitrogen is preferred; otherwise, a −80°C freezer provides sufficient stability, while storage at −40°C can lead to loss of virulence in S. pneumoniaestrains often used for animal experiments.

If there is doubt whether animals died from drug toxicity rather than infection, samples of blood from the hearts of dead animals as well as from some of the survivors should be inoculated onto agar plates. The cultures from the dead animals should show the infecting organism, whereas the cultures from surviving animals should be sterile.

The three methods most frequently used for calculating the 50% dose parameter are the method of Reed and Muench (157), the probit method (158), and method of the sigmoidal dose-response (variable slope), also known as the Hill equation (159).

Treatment Routes and Times

By altering the treatment route or schedule, differences in activity can be demonstrated. In addition, the relative efficacy of oral and subcutaneous administration of a substance can be compared (as discussed earlier).

Studies in which mice infected with S. pneumoniae serotypes 1 and 2 were treated once, orally or subcutaneously, with doses of ampicillin or amoxicillin demonstrate the influence of the treatment route (160). Treatment by the subcutaneous route was more effective than treatment by the oral route. When administered orally, amoxicillin was more active against the type 1 S. pneumoniae than was ampicillin. Otherwise, the two agents were equivalent in activity (160).

Tests with ampicillin and amdinocillin (mecillinam) in which the antibacterials were administered to mice subcutaneously immediately (0 hour) or at 1, 2, or 4 hours after infection indicate the potential problems with delay in treatment initiation (160). With treatment 4 hours after inoculation, the regimens were considerably less effective than when administered immediately or at 1 hour (160). Similarly, multiple-dose regimens were more effective than single-dose regimens, which are typical for all β-lactam antibiotics. This can be explained by the importance of T>MIC (i.e., the duration the antibiotic concentration remains above the MIC). Interestingly, the greater effectiveness of multiple dosing was also apparent for ampicillin against enterobacteria resistant toward the drug (160).

For drugs such as the aminoglycosides or the fluoroquinolones, single-dose regimens would result in lower PD₅₀s than those of multiple-dose regimens due to the importance of the AUC/MIC ratio for these types of compounds (e.g., reference 149). In this manner, PK/PD relationships can be demonstrated by relatively simple dosing experiments using the mouse protection test.

Differences in the PK patterns of various agents may also be determined by prophylactic-type experiments (i.e., treatment before infection). The activities of ceftriaxone and cefotaxime were similar when treatment was administered immediately (0 hour) after treatment (161). However, when treatment was administered at 24 or 8 hours before infection, the activity of ceftriaxone was clearly superior against the gram-negative bacteria. No such differences were seen against S. aureus in this model.

Duration of treatment can affect the outcome (for reviews, see references 51 and 162). For example, extending ciprofloxacin treatment (p.o., 20 mg/kg twice a day.) of systemic S. typhimurium–infected mice from 17 to 28 days improved the outcome (163).

Synergy or Antagonism In Vivo in Screening Models

In order to study interaction between two antibiotics in vivo, graded doses of the combined agents and the single agents are administered to groups of four to six mice after infection. The PD₅₀ can be calculated for the combination and compared to the 50% doses for the single agents. A fractional index (FIC) can be calculated for the combination doses by dividing the PD₅₀ value for each of the components in the combination by the PD₅₀ value obtained for each component alone and adding the two quotients. Synergy can then be defined using the FIC (e.g., a value of ≤0.5–0.6), similar to the method of studying synergy in vitro (164). Delay of treatment causes loss of synergistic activity.

Setting the PD₅₀ value of the combination at one-fourth that of the most active single agent is perhaps an easier way of evaluating the results of the experiments. Combining different dosing regimens over a 24-hour dosing period after inoculation will further allow estimation of the importance of different PK parameters for synergy in vivo. This approach has been studied in the neutropenic mouse thigh model (160,165,166) but could also be used in the mouse peritonitis model.

Antagonism between antibiotics in vivo has been studied in the mouse peritonitis model (167). The combination of erythromycin and penicillin against S. pneumoniae in vivo resulted in the same mortality as erythromycin alone and significantly higher mortality than penicillin alone. The inhibiting effect of erythromycin on the activity of penicillin could also be demonstrated by in vivo time-kill curves of bacterial counts in peritoneal wash. This apparent antagonism between erythromycin and penicillin has been contested by others, at least in vitro (168). The use of animal models and PK/PD relationships in determining in vivo antibiotic synergy has been reviewed (160,164,166,169–171).

Den Hollander et al. (172) used in vitro–derived FIC determinations to attain PD parameters of the combination of tobramycin and ceftazidime. They first determined the MICcombi, which is the MIC of tobramycin in the presence of ceftazidime. Using humanlike PK profiles, they then divided the tobramycin and ceftazidime concentrations at each time point along the dosing interval to construct FIC over time curves, which were used to derive PD parameters. T_>FICI (time above a specified FIC value during the dosing interval) appears to be the key PD parameter for this combination. Although difficult, to date, little application of this methodology has occurred with in vivo data.

Mouse Peritonitis Model for the Study of Antibiotic Activity against Intracellular Bacteria

The mouse peritonitis model has for many years been used to harvest leukocytes for in vitro purposes (173). The extravasation into the peritoneal fluid of leukocytes can be stimulated by microbiologic or chemical means. Although in vitro cell culture has long been in use for the study of antibiotic activity against intracellular bacteria (174), few have extended this to the in vivo situation. Sandberg et al. (175–179) studied the mouse peritonitis model with i.p. inoculation with Staphylococcus aureus and used peritoneal wash to enumerate extra- and intracellular staphylococci after a series of centrifugation steps and washing, addition of lysozyme to remove extracellular bacteria, and lysing of the leukocytes in plain, sterile water. Enumeration of colony counts in the supernatant, without cells, and the lysed pellet with cells resulted in reasonable estimates of extra- and intracellular staphylococcal cells, which were validated by electron microscopy. Treatment with various antibiotics showed slower killing effect intracellularly, which correlated with slower intracellular growth and the the theoretical penetrability of the antibiotics (175). PK/PD analyses of dicloxacillin and linezolid showed surprisingly good effect of the former, which is standard treatment of staphylococcal infections in many European countries, whereas the latter drug showed almost no intracellular activity (176,177). Similar studies have been performed with Salmonella sp revealing excellent both extra- and intracellular activity of ceftriaxone but less so of carbapenems (180).

Thigh Lesion (Selbie) Model

The rodent thigh lesion model was originally described by Selbie and Simon in 1952 (181) and continues to be a fundamental experimental method in animal model antimicrobial efficacy studies. This model is commonly employed in the development of new antimicrobial agents given its relative simplicity compared to other sites of infection. It can also allow for more limited numbers of animals as each thigh can represent one biologic replicate in the infection model (i.e., An investigator only needs to use two mice, four thighs, to achieve statistically evaluable data). Note, however, that using both thighs from same mice for infection is not allowed by all animal ethics committees in Europe. In general, the model involves intramuscular injection of an inoculum into the dorsal thighs of the animal. Mice are then treated with an antimicrobial agent for a defined period, euthanized at study end point, and CFU enumerated from each thigh. In order to make the data most meaningful, 0-hour control mice are necessary to determine the viable burden at the start of therapy, which allows one to determine whether infectious burden increased, decreased, or remained stable over time. Untreated controls are also necessary to prove fitness in the animal model. Most studies use a neutropenic mouse model. There are a number of reasons for this but two are likely the most important. First, an unconfounded evaluation of drug–pathogen effect can be performed if the immune system is removed or significantly inhibited from affecting the outcome. Secondly, many animals are inherently resistant to some types of infection and/or species, or the organism has limited fitness, unless the immune system is compromised. While CFU determination of pathogen abundance is most commonly performed, novel techniques such as fluorescent protein markers, serum biomarkers of infection, image scoring, quantitative polymerase chain reaction (qPCR), and antigen/antibody testing have been developed for certain pathogens as a means to monitor infectious burden in animals.

Animal model research using fungal pathogens has increased dramatically in the past decade. Similar to bacteria, fungal pathogens have the ability to cause localized disease (e.g., esophageal candidiasis or dermatophytosis), disseminated disease (e.g., invasive candidiasis), or a combination of the two (e.g., invasive aspergillosis or cryptococcosis). Therefore, models have been developed to mimic either site-specific inoculation routes or disseminated routes (e.g., intravenous inoculation). A comprehensive review and practical descriptions of many of these models is presented in the following citations (4,150,182,183). Important differences from bacterial models do require consideration and include the inoculum size and immune suppression. Many fungal pathogens require a high inoculum to produce disease, most commonly on the order of 7 to 9 log₁₀. Given a very high inoculum, at times, this can lead to relatively small growth in end organ burden prior to study end point or animal death (i.e., 1 log₁₀ or less) (184). Additionally, although many bacterial studies use a neutropenic host animal, it is almost universally employed for fungal studies. This is required as most fungal infections cannot be established in the animal host without significant immune suppression, which is likely intuitive as a major risk factor to these pathogens in humans is significant immune suppression. An often-employed additional step in immune suppression in pulmonary mold infection models is the use of high doses of corticosteroids (185–190). A few investigations have employed rodents deficient in specific immune components (191,192). For example, severe combined immunodeficient (SCID) mice devoid of B- or T-cell immunity have been used for a mucosal candidiasis model to mimic infection in patients with HIV (192). Finally, a murine model of diabetic ketoacidosis, a major risk factor for disseminated and cerebral zygomycosis, has been successfully described and used to examine antifungal therapy in this setting (193,194).

Animal model investigation for fungal pathogens is most robust for Candida species and includes oropharyngeal and esophageal candidiasis (195,196), vaginitis (49), and invasive candidiasis (too numerous to cite examples, however, representative studies are referenced here [84,182,190,197–237]). The most common model used in the study of antifungal agents in murine models is disseminated infection. This is achieved through direct intravenous injection of the organism inoculum into the tail vein of a mouse. Provided a large enough inoculum is introduced (6 to 7 log), C. albicans disseminated infection in the neutropenic mouse model will rapidly progress to death in 24 to 72 hours (209). If one wishes to study organism growth or decline over longer treatment periods, this can be accomplished using a lower starting inoculum (18). For other Candida spp (such as Candida glabrata), severe infection is more difficult to establish and often requires longer experimental durations to find discernable differences in treatment groups (199,200,211,238). Nonetheless, the disseminated candidiasis model continues to be one of the most commonly used fungal models for drug development and dosing regimen refinement.

A relatively clinical complication of invasive candidiasis is dissemination to the eye with subsequent endophthalmitis. Animal models examining drug efficacy in animal models of endophthalmitis have been developed (239–247). These models have provided important guidance on therapeutic options for which drug and immune system penetration is limited. Models mimicking human fungal keratitis have also been described (248–259).

Filamentous fungal pathogen models (most commonly Aspergillus) have also undergone significant experimental refinement. Many of these pathogens are acquired via inhalation and therefore primary pulmonary infection models with dissemination most closely mimic human disease (189). However, disseminated models via intravenous injection of the organism inoculum have also been used (260). As stated previously, these models often employ significant immune suppression in the form of a combination of chemotherapy-induced neutropenia and corticosteroid treatment. Treatment durations of 7 days or longer are often required as in general, despite the immunosuppression, filamentous fungi require longer incubation periods to grow to significant levels and/or disseminate via a pulmonary infection route. A previous limitation to robust filamentous fungal pathogen investigation has been difficulty in quantitation, as filamentous fungi do not grow in discrete colonies as do bacteria and yeast. Additionally, concern has been raised that homogenization can fracture a filament into multiple pieces leading to overestimation of organism burden. Current molecular surrogates of organism burden have largely alleviated this limitation. The most common methods of organism burden include galactomannan measurement (185) or real-time qPCR (261,186,262). For example, a recent publication evaluated whether qPCR was a good surrogate marker for disease progression, treatment outcome, and animal mortality in a 7-day study of invasive pulmonary aspergillosis (IPA) in a murine model using numerous Aspergillus fumigatus isolates (186). The authors demonstrated a very strong relationship between qPCR result and treatment efficacy. In fact, every 1-log increased growth of organism based on qPCR resulted in a 17% increase in mortality. Additionally, the increase in survival was most profound at the dose exposure that was associated with net stasis (static dose) of organism burden. Thus, stasis or net cidal drug activity based on qPCR is a very strong predictor of clinical survival in this model. Other measures of organism viability and abundance have also been used such as XTT, DiBAC staining, chitin measurement, histopathology grading, lung weights, and pulmonary infarct scoring (263,264). These types of models have also been recently expanded to examine less common filamentous fungal pathogens including Zygomycetes(194,265–268).

Models mimicking fungal meningitis for pathogens that commonly cause primary central nervous system (CNS) infection or have a high likelihood of dissemination to the CNS have been developed for a number of pathogens (e.g., Cryptococcus, Aspergillus, and dimorphic pathogens Blastomyces, Histoplasma, and Coccidioides). Infection is induced by either intravenous or intracisternal injection of a defined inoculum and organism burden is quantified in the CSF and brain parenchyma to determine outcome. One of the more common pathogens examined in animal models given its predilection to cause CNS infection is Cryptococcus sp (269–278). Animal hosts in these experiments have included rabbits, mice, and guinea pigs. Animal models of CNS aspergillosis, many developed by Clemons and Stevens, have been described and used to determine therapeutic efficacy of various antifungal agents (279–286). Additionally, CNS models of invasive candidiasis have been developed to better understand treatment strategies for this rare pediatric complication (287–289).

Dimorphic fungal pathogens are acquired via the pulmonary route but can also disseminate to involve the CNS. This is not uncommon for Coccidioides sp and a number of investigations have examined antifungal therapy in animal models of CNS coccioidomycosis (290–296). For reference, a thorough review of dimorphic fungal animal models is provided by Sorensen et al. (297). Finally, animal models of CNS phaeohyphomycosis have been described for this rare but severe infectious entity (298–300).

The efficacy of agents directed at dermatophytes has been evaluated in animal models with cutaneous infection (301–305). Most commonly, the site of infection (skin, foot pad, or nail) is mechanically abraded prior to topical inoculation to predispose the tissue to infection. The infection often takes several days or weeks to establish and therefore initiation of systemic or local topical therapy is delayed. After therapy, which may also require a prolonged period of time, tissue samples are cultured and examined by histopathology to determine drug efficacy.

Rare fungal infections that can occur in patients with severe or prolonged immunosuppression have also been studied to a limited extent. Some examples include and blastoschizomycosis (306–308), fusariosis (309–312), scedosporiosis (313–315), and trichosporonosis (316,317).

Animal Models of Urinary Tract Infections

The models for experimental acute urinary tract infections (UTIs) that are commonly used to evaluate antibacterials produce either hematogenous or ascending infections, depending on whether the inoculum is administered intravenously or intravesically, with or without the addition of a foreign body (318). Mice and rats are the most common species used in experimental UTIs, and they have been used to determine the pathogenesis of this infection as well as test experimental chemotherapy. Note, however, that naturally occurring vesicoureteral reflux (backflow of the urine from the bladder to the kidneys) normally occurs in rodents but only infrequently in humans (319). Although some bacteria have a trophism for the urinary tract, even when inoculated intraperitoneally (e.g., Borrelia burgdorferi [320]), and some models of UTIs utilize bloodstream inoculation with bacteria to generate pyelonephritis (321), normally, manipulation of the urinary tract of rodents is a prerequisite for establishing infection. Some of the bacterial virulence factors necessary to establish UTI in humans (e.g., type 1 and type P fimbriae of E. coli) are also required to establish ascending UTI in mice and rats with the same binding mechanisms to the epithelium of the urinary tract. With focus on these virulence factors in strains used for inoculation, little manipulation of the urinary tract is actually needed for creating ascending infection in these rodents. In addition to testing antimicrobials, these models have been used to evaluate adjunct antiinflammatory agents (e.g., pentoxifylline [322]). Previous reviews (144,321,323,324) describe additional models for establishing UTI, in particular, pyelonephritis.

Ascending Obstructive Pyelonephritis

The original model for pyelonephritis (325) was further developed (323). Rats are operated on and bacteria are inoculated directly into the bladder, whereafter one of the ureters is obstructed by ligation, which is removed 18 to 24 hours later. This model has been used to demonstrate that, following acute infection, inflammation leading to chronic pyelonephritis is the major contributor to renal damage. Early antibiotic therapy suppresses renal damage (326) by rapid eradication of bacterial infection, but antiinflammatory treatment with dexamethasone failed to suppress renal damage (327). This model has also been used to compare the efficacy of various antibiotics (328).

Chronic Cystitis and Subacute Pyelonephritis

A model of persistent bladder infection has been described that requires placement of a foreign body into the bladder (329). A small cylinder of polyurethane foam (4 × 2 mm) was introduced into the bladder via a needle pushed bluntly into the bladder via the urethra. Two weeks later, the surgically exposed bladder was inoculated directly with E. coli. Chronic bacteriuria ensued for as long as 8 weeks after infection, leading to focal and diffuse inflammation of the bladder wall. Furthermore, bilateral pyelonephritis developed in the majority of the animals. The model is amenable to antimicrobial therapy (329). Several drugs have been tested for 7 days duration, and the effect was measured as reduction of CFU in the bladder wall and kidney homogenates (329).

Acute and Subclinical Pyelonephritis by Intrarenal Infection

This procedure is commonly used to establish kidney infections for the evaluation of antimicrobial agents, usually in rats. The kidney is surgically exposed, and the inoculum (50 to 100 CFU/µL) is injected directly into one (330,331) or both poles (332) of the kidney. Usually, only one of the kidneys is used, but infection of the contralateral kidney can ensue (332). The passive infection of the contralateral kidney has been used to study subclinical pyelonephritis in comparison with the acute infection in the directly inoculated kidney (332).

The model has been used to study the effect as well as toxicity of gentamicin (330,331) and the importance of duration of therapy in pyelonephritis (330). Several different antibiotics have been compared in this model (329,330). The most effective drugs have been gentamicin, ceftriaxone, and various fluoroquinolones, while ampicillin and co-trimoxazole have shown lower efficacy (329,330).

Ascending Pyelonephritis following Direct Bladder Inoculation

Direct inoculation of surgically exposed bladders has long been used to induce experimental UTI. Rodents generally do not need to be water deprived. The urethra is clamped and the bladder exposed by surgical intervention. Bacteria (50 to 200 CFU/µL) are slowly injected into the bladder, and the urethra remains clamped for 2 to 4 hours in order to avoid discharge of the inoculum. Ascending infection with development of bilateral pyelonephritis will follow. Renal scarring for up to 6 weeks later has been used as a parameter to study interventional therapy (333). Early quinolone treatment eliminated the incidence of renal scarring, while delayed treatment resulted in renal damage in approximately 50% of the animals (333). Reduction as compared to untreated controls of CFU of E. coli after homogenizing the kidneys was used to study the effect of nemonoxacin, a novel nonfluorinated quinolone, as compared to common fluoroquionlones as part of screening for in vivo efficacy of this compound in several animal models (334).

Ascending Urinary Tract Infection by Bladder Inoculation via Urethral Catheter

During the last 10 to 15 years, this model in mice has been the most widely used to study virulence factors and host resistance in UTI. With this model, detailed knowledge of the binding mechanisms between bacteria harboring various virulence traits and epithelial cells in the urinary tract has been discovered, and it has further been utilized to reveal the various facets of the host mechanisms of resistance to infection, both in the bladder and in the kidney. A detailed review of this literature is beyond the scope of this chapter, and the reader is referred to recent reviews (335–338).

The model is probably by far the easiest to use for the study of UTI when the technique of bladder cauterization has been learned, since no other surgical manipulation is needed. With the correct bacterial strain harboring the virulence factors needed (i.e., type 1 or type P fimbriae of E. coli or other enterobacteria), UTI with moderate to high bacterial counts in the urine, bladder wall, and kidneys (40% to 70% of the infected mice) will ensue (339–341). The presence of type 1 fimbriae in E. coli can easily be tested for by agglutination with bakers yeast cells or with sheep erythrocytes (342). CFUs appear to decrease after 2 to 3 weeks, which is why antibiotic treatment studies should preferably be performed 1 to 8 days after inoculation. Higher colony counts can be obtained by pretreating the mice with 5% glucose in the drinking water starting 3 days prior to infection (341,342).

Antibiotic concentrations can be measured simultaneously in serum, urine, and renal tissue, and these can be related to the effect of the antibiotics according to the MICs of the infecting strains (340,342). PD relationships for dosing of antibiotics in UTI can be studied with this model (343), which has also been used to study antibiotic effect against UTI caused by Enterococcus faecalis and Pseudomonas aeruginosa (344). The model has been used to show effect of different antibiotics and has revealed that no antibiotics are able to eradicate bacteria in the bladder wall because the bacteria are situated intracellularly in bacterial colonies, perhaps in a kind of biologic biofilm, which renders the bacteria resistant to antibiotics present both in blood and urine (340). Bacteria in the urine and in the kidneys are more easily removed, as long as urine and serum antibiotic concentrations lie above the MIC, while resistance correlates with missing antibiotic effect even for lowly ciprofloxacin-resistant Escherichia colibearing the qnr-genes (345,346).

Models of Urinary Tract Infections Associated with Indwelling Catheters

Models of short-term and long-term indwelling catheter infections have been described (347,348). Both long (25-mm) and short (4-mm) segments of tubing have been used in mice; the short not secured and therefore expelled, with 3 to 7 days serving as a short-term model. When the longer segment has been secured to the bladder, it has been left for up to 12 months, thus mimicking a long-term indwelling urinary catheter (348). Spontaneous bacteriuria (>10²CFU/mL urine) was not reported in the mice with unsecured tubing but occurred intermittently in 44% of the animals with secured bladder catheters and was predominantly due to Proteus mirabilis. Apparently, no colonization of the kidneys occurred unless the infection was persistent and of high density (10⁵ CFU/mL). Postsurgery ampicillin treatment for 7 days and housing on wire platforms reduced the incidence of bacteriuria to 7% over 12 months.

Placing a catheter precolonized with P. aeruginosa in the rat bladder with subsequent 3-day treatment showed higher reduction of catheter biofilm counts with a combination of fosfomycin and a novel fluoroquinolone, prulifloxacine, than with the latter drug alone (349). A bladder cathether model in mice for the study of Candida biofilm on catheters with or without silver coating was introduced by Wang and Fries (350). A rabbit indwelling bladder catheter model was used to show that silver-coated catheters significantly reduced bacteriruia as compared to noncoated catheters (351).

Models of Urinary Tract Infection Resulting from Bloodstream Inoculation

Hematogenous infection has been long used to establish UTI (321,352). Normally, no manipulations of the animals are required, but strain selection is essential in order to have selective colonization of the kidneys. Examples of this model have used E. faecalis (353–355) or S. aureus or K. pneumoniae (356). Trovafloxacin and rifampin alone or in combination were compared in a rat model against E. faecalispyelonephritis (354). Although antagonism is usually considered the result of combining these two types of antibiotics, no such effect was evident from the reduction in renal colony counts found, which was similar to those achieved with the two drugs given alone. A new cephalosporin with gram-positive activity was compared with ampicillin and vancomycin against E. faecalis in the hematogenous model in mice (353). With the higher doses used, the cephalosporin was as effective as the two other antibiotics.

Animal Models of Foreign Body Infections

Advances in the development of prostheses and a variety of vascular grafts and permanently residing catheters have been limited by problems of bacterial infections, which are exceedingly difficult to cure and often necessitate removal of the device. Reviews of animal models of foreign body infections have appeared (357–360), prompting review of strategies for dealing with such infections clinically (361,362). Foreign bodies and bone and joint infections are discussed in “Animal Models of Osteomyelitis” section.

One of the most common foreign body infection models utilizes the subcutaneous implantation of a perforated plastic cylinder (363). Although commonly used in guinea pigs and rats, this model can also be used in mice. This model is suitable for evaluation of antibiotic treatment, although the infections are difficult to treat. For example, one study demonstrated that only triple therapy with 50 mg/kg vancomycin, 50 mg/kg fleroxacin, and 25 mg/kg rifampin (i.p. every 12 hours for 21 days) provided an adequate response (364). The same model has recently been used to study the effect of vancomycin, gentamicin, and daptomycin against E. faecalis (365); daptomycin against methicillin-resistant Staphylococcus aureus (MRSA) (366); and levofloxacin and rifampicin, alone or in combination, against methicillin-susceptible Staphylococcus aureus (MSSA) (367). In an effort to study infections related to artificial vascular devices, Artini et al. (368) studied the antimicrobial properties of two kinds of vascular prosthesis to prevent early-onset infections and the efficacy of the concomitant action of a systemic antibiotic treatment. In adult male Wistar, they subcutaneously implanted in four groups a silver-coated prosthesis fragment, and a rifampicin-soaked prosthesis fragment in the remaining four groups. They inoculated the site of implant with S. aureus and administered systemic levofloxacin for 7 days in four groups representing the two kinds of prosthesis; after 21 days, the rats were sacrificed, prosthesis fragments were sonicated, and the corresponding supernatants were plated for bacterial counts. The rifampicin-soaked prostheses explanted from rats treated with levofloxacin were sterile, regardless of the bacterial inoculum. In other groups, some prostheses were colonized (368). Jean-Baptiste et al. (369) studied the efficacy, safety, and healing properties of polyester vascular prostheses coated with a hydroxypropyl-β-cyclodextrin (HPβCD)–based polymer (PVP-CD) and loaded with one or two antibiotics in vitro and in an experimental model in dogs. The study end points included hemolysis, platelet aggregation, antibacterial efficacy, polymer biodegradation, acute toxicity, and chronic tolerance and PVP-CD was proved safe and demonstrated excellent biocompatibility, healing, and degradation properties. Effective antimicrobial activity was achieved with PVP-CD in conditions consistent with a sustained-release mechanism (369).

Infected Sutures

Intramuscular implantation of infected sutures into the thighs of mice has been described (370). Lengths of cotton 2–0 suture are sterilized in broth, which is subsequently inoculated (in this example, with K. pneumoniae). The contaminated suture is then attached to a sterile needle and drawn through the thigh muscle of a mouse, the exposed ends are trimmed flush with the skin, and the ends are buried under the skin. Culture of homogenized suture material facilitates determination of the infecting CFU. The infection spreads from the suture to the surrounding muscle, and subsequently, a sepsis develops. This model is amenable to antibiotic intervention (371); continuous infusion of 180 mg/kg/day cefazolin by use of intraperitoneally implanted osmotic pumps was superior to an equal amount of antibiotic given by intramuscular bolus dosing (every 8 hours).

Animal Models of Skin, Burn, and Surgical Wound Infections

Several models of infection associated with surgical intervention (in the absence of foreign bodies) have been described and utilized for experimental evaluation of antimicrobial chemotherapy. These models rely on direct inoculation of the wound site and generally achieve simulated prophylaxis of postsurgical infection by administering antibiotics at the time of infection, or just before, and ascertain efficacy of treatment by determining infection remaining at the inoculation site as well as dissemination to internal organs. These models can be used to evaluate oral, parenteral, or topical antibiotic administration. Several reviews of postsurgical infectious complications and burn wound infection are available (361,372). A model of clean wound infections has been described (373).

Surgical wounds in mice infected with either MSSA or MRSA have been used to study the effect of teicoplanin, antimicrobial peptides, and hyaluronic acid (374,375). Ozcan et al. (376) induced sternal wound infection with MRSA and compared topical, systemic, or combination of topical and systemic vancomycin with untreated controls and found best effect of the combination as measured by reduction of bacteria in the wound. Mihu and coworkers (377) examined the capacity of a nitric oxide–releasing nanoparticle (NO-np) to treat wounds in mice infected with Acinetobacter baumannii. They demonstrated that NO-np treatment reduced suppurative inflammation, decreased microbial burden, reduced the degradation of collagen, and altered the local cytokine milieu.

Models of Infected Burn Wounds

Infections following burns are difficult to treat and have been modeled in several systems. Typically, partial-thickness or full-thickness burn wounds are made on the shaved skin of anesthetized animals by using a metal stamp or by partially immersing the animals in heated water; a minimum of about 30% of the skin area of the animal needs to be damaged. Normally, the burn wound is then directly inoculated, but translocation of normal intestinal bacteria often follows; interestingly, animal models have demonstrated that infection of a skin burn apparently promotes translocation of intestinal bacteria, compared with that occurring when wounds are kept sterile (378,379). Experimental burn-induced wound infections for the study of antimicrobial treatment have been performed in rats or mice with MRSA (380,381), MSSA (382), K. pneumoniae (383), and C. albicans (384). Treatment modalities investigated have ranged from topical agents, systemic antibiotics, phage therapy, and photodynamic therapy.

Research in this area has also evaluated promotion of host resistance to infection by use of cytokines, whose expression in some cases also occurs as part of the normal healing process. In experimental full-thickness murine wounds, the expression of inducible nitric oxide synthase (iNOS) by infiltrating inflammatory cells is not a part of normal repair processes but is a response to bacterial colonization due to S. aureus (385).

Skin Infection Models

Skin infection models in guinea pigs and mice have recently been developed and used for treatment effect studies with topical and systemic antibiotic treatment (304,386,387). In both animal species, the skin is prepared under anesthesia by first removal of hair, then scraping the skin with, for example, fine sand paper until the superficial corneal skin layers are removed. Bacteria (Staphylococcus aureus or Streptococcus pyogenes) or dermatophytes are then inoculated directly on the denuded skin, which creates a skin infection spreading only to the fascia but shows up as a red, inflamed, and edematous lesion (386,387). Usually, the infection remains localized to the skin without signs of systemic infection and the animals appear healthy apart from the local infection. After treatment, a skin biopsy is taken from the infected skin and homogenated, from which bacteria or fungi can be quantified. Both topical and systemic treatments have shown significant reduction in bacterial or fungal counts as compared to untreated controls (387).

Animal Models of Pneumonia

Rodent Models of Acute Pneumonia

A useful model for production of pneumococcal pneumonia in rats has been described by Ansfield et al. (388). Lung bacterial counts progressively increased, reaching 10⁷ CFU per lung within 48 hours. This increase was associated with localized atelectasis and consolidation. Bacterial multiplication could be inhibited with 50 mg/kg tetracycline given once intraperitoneally prior to infection or at 4 or 12 hours after infection. Viable pneumococci were rapidly killed by lung defenses if bacterial multiplication was inhibited within 12 hours of the onset of infection. No change occurred in the bacterial population if tetracycline treatment was delayed until 24 hours after infection.

A similar rat model of pneumococcal pneumonia was used to determine the role of the host defense system in antimicrobial therapy by impairing the phagocytic system by complement depletion using cobra venom factor (389). There was a consistent decrease in body weight with time (approximately 15% to 20% loss by 108 hours postinfection). The temperature was initially elevated (approximately 1°C to 2°C, up from normal levels of 37°C to 37.4°C), but by 108 hours, it was depressed (by as much as 4°C). The weight of the left lung increased (from about 0.6 ± 0.15 g to 3 to 4 g) with the involvement of the lung tissue in the infectious process. Pulmonary lesions were very extensive in the left lung by 108 hours, and the number of pneumococci increased from 6 × 10⁷ CFU to approximately 10⁹ CFU per lung. By the end of 108 hours, the infectious process had spread, so that both the blood (approximately 10³ to 10⁶ CFU/mL) and the pleural fluid were positive for pneumococci. Treatment of infected rats with 2 mg/kg penicillin G every 12 hours starting from 36 hours postinfection was very effective in preventing weight loss (only a transient 6% to 8% loss by 36 hours, 0% to 5% by 132 hours), normalizing the temperature (approximately 38°C to 39°C at 36 hours and within normal range from 84 hours onward), and maintaining the weight of the left lung near normal (a transient rise up to 2 g at 36 hours, within the normal range from 84 hours onward). By 84 hours, there was a significant but highly variable fall in the number of viable pneumococci, and by 132 hours, all lung, blood, and pleural fluid cultures were sterile. After treatment with cobra venom factor, the whole complement hemolytic activity was decreased to less than 2% of normal values. When cobra venom factor–treated rats were also administered penicillin, the results very closely paralleled the course of infection seen in normal untreated infected rats, indicating the importance of an intact innate immune system for the outcome of antimicrobial therapy (389).

Rat models have also been used to demonstrate that chronic alcohol ingestion increases susceptibility to infection (390), as does liver cirrhosis (391) and neutropenia (392). Intratracheal injection of soft agar (0.7%) encased penicillin-resistant S. pneumoniae to create an acute pneumonia model in immunocompetent rats (393). Lung CFU and mortalities were dependent on the size of the infectious challenge (inocula of 2.9 × 10⁸ or 4.2 × 10⁹ were uniformly fatal, but only the higher inocula produced stable lung CFU of approximately 10⁹ CFU/g over 3 days). Despite an in vitro MIC of 2 µg/mL for penicillin, 100,000 IU/kg penicillin G every 2 hours for eight administrations reduced lung CFU (the change was approximately log 2/g vs. controls) and promoted survival (13% mortality vs. 33% for controls), whereas 250,000 IU/kg reduced the mortality to 7% and the change in CFU/g was approximately log 3. Similar activity was seen with cefpirome (200 mg/kg) and cefotaxime (100 mg/kg). With the latter, the lung CFU/g was similar to that resulting from the penicillin G treatment, but the mortality was apparently higher (19%). Two administrations of 50 mg/kg vancomycin every 8 hours produced a change in CFU/g of log 4 and 7% mortality.

Murine models of acute pneumonia are increasingly incorporated into drug development and PK/PD studies. An early description published in 1981 by Beskid and colleagues (394) examined a novel cephalosporin (ceftriaxone) efficacy in a murine model of pneumococcal pneumonia. Mice were infected by intranasal instillation of the inoculum and treated with either ceftriaxone or another comparator antibiotic (cefotaxime, ampicillin, piperacillin, cefamandole, and carbenicillin). After 48 hours, the mice were sacrificed and lungs aseptically harvested and a touch prep culture of a freshly sectioned area of lung was performed on blood agar plates. Overall, ceftriaxone was superior to the comparators in terms of PD₅₀ (mg/kg): ceftriaxone, 0.88; ampicillin, 11; cefotaxime, 16; piperacillin, 79; cefamandole, 79; and carbenicillin, 84. This is a prime example of the use of preclinical animal model studies as ceftriaxone is now a cornerstone antimicrobial agent in community-acquired pneumonia.

Many contemporary models have also been described and a thorough review of murine models to mimic human pneumonia is provided (395). The models have been adapted to include common community pulmonary pathogens including S. pneumoniae, H. influenzae, Chlamydia pneumoniae, and Mycoplasma pneumoniae. Important considerations in the model include host immune dysfunction, organism pathogenicity in mice, route of infection, inoculum size, experimental duration, and end point (e.g., mortality, organism burden, etc.). Mice are usually made neutropenic by methods previously described in this chapter. This allows for the most accurate evaluation of drug efficacy by removing immune function as a confounder on treatment effect. Additionally, many bacterial organisms will not produce disease in the model without some level of immune suppression. For a few organisms, even with immune suppression, it is difficult to obtain reproducible results in biologic replicates. One strategy to overcome this has been to use pulmonary administration of a chemical irritant (1% formalin) just prior to organism inoculation (396). A decreased ability to cause infection (i.e., decreased fitness) has also been noted in drug-resistant isolates, where presumably the loss of fitness is due to genetic changes in the isolate (397). Therefore, it is important at the outset to ensure all organisms used have similar degrees of fitness and pathogenesis in the animal model if one is attempting to compare efficacy against resistant strains.

The routes of infection used for production of murine pneumonia vary and include aerosolization of the inoculum, intranasal instillation with subsequent aspiration, injection into the trachea via percutaneous puncture with a fine needle, or direct instillation into the lungs by tracheal intubation. One advantage of aerosolized inoculation is the ability to infect large numbers of animals in large chambers at the same time with similar inoculum burden. For example, nebulization of 10⁸ CFU/mL of K. pneumoniae via a Collison nebulizer for 45 minutes produced a similar degree of pneumonia in up to 100 mice at the same time (398). Experiment duration can vary depending on pathogenicity of the infecting organism but usually does not need to be prolonged more than 24 to 48 hours. Finally, determination of organism burden is most commonly performed by quantitative culture techniques (CFU determination).

The earlier mentioned techniques are now increasingly used for hospital-/health care–acquired pneumonia including multiple drug–resistant (MDR) organisms, difficult-to-treat gram-negative organisms, and MRSA. For example, A. baumannii is an increasingly recognized respiratory pathogen in patients who are mechanically ventilated and develop hospital-acquired pneumonia. Many of these isolates have limited therapeutic options and therefore animal models have been helpful to delineate treatment strategies (399–408). Dudhani and colleagues (399) examined the predictive PD index for colistin in treatment of experimentally infected mice. The free drug AUC/MIC correlated best with efficacy (R² = 0.80). The PD drug target in the pneumonia model was a free drug AUC/MIC of 1.57 to 6.52 for net stasis. This was essentially identical to the PD target in a murine thigh model with the same organism. Unfortunately, drug resistance emergence was detected in subpopulations. Many other common health care–associated pulmonary pathogens have also been studied in rodent models such as MRSA (409–417), P. aeruginosa (418–426), and K. pneumoniae (427–434). Additionally, anthrax models have been actively pursued in recent years given the continued threat of bioterrorism in the United States and other parts of the world (435–442). These models are particularly relevant as clinical studies are not possible.

Chronic Pneumonia Models

Many chronic pneumonia models are based on the premise of preexisting lung conditions, such as chronic obstructive pulmonary disease (COPD) and cystic fibrosis (CF), which produce intermittent obstruction and subsequent risk for chronic infection. These preexisting conditions are difficult to mimic in rodents and bacteria are inherently rapidly cleared from the airways in these animals (443–445). Therefore, strategies to mimic periodic obstruction and prevent bacterial clearance have been developed. To date, the most common method is the use of agarose or alginate beads (444,446). This method successfully resulted in a persistent P. aeruginosa infection in rats for up to 35 days (444).

Most commonly, P. aeruginosa is the pathogen of choice for studies of chronic pneumonia as it is not only one of the most common isolates found in colonization and infection in COPD and CF patients but is also particularly difficult to treat and has high propensity of acquiring drug resistance. One of the earliest studies examined antimicrobial therapy in a guinea pig model of chronic P. aeruginosa pneumonia (447). Infection was established using agar bead–encased bacteria and the compounds tested included ticarcillin (120 mg/kg), ciprofloxacin (10 mg/kg), and tobramycin (1.7 mg/kg). Three days after infection, the drugs were administered as monotherapy for 5 days. Ciprofloxacin was judged to be most effective, followed by tobramycin and ticarcillin, which was ineffective, based on CFU counts. Notably, no single drug treatment was able to eradicate the infecting organism completely. Rodent models more recently have been used to examine antimicrobial therapy in chronic P. aeruginosa pneumonia (448–452). For example, Macia and colleagues (449) examined ciprofloxacin and tobramycin monotherapy and combination therapy in a murine model of chronic pneumonia using a reference strain and its hypermutable derivative. After exposure to ciprofloxacin, the hypermutable isolate of P. aeruginosa demonstrated a profound and rapid increase in drug-resistant subpopulations despite having the same in vitro susceptibility as the reference strain, which did not show any resistant subpopulations after drug exposure. This effect was not observed with tobramycin monotherapy. Finally, the combination of the two drugs appeared synergistic against the hypermutable isolate. Inhaled therapeutics are an additional area of investigation garnering more interest for chronic pneumonia (448,453–457). The advantage of this method is directly targeting antimicrobial therapy at the site of infection as well as limiting systemic toxicity that can be problematic for certain antimicrobial agents.

Animal Models of Otitis Media

Otitis media, an infection of the middle ear, is largely a childhood infection that apparently very few people avoid contracting. Caused chiefly by H. influenzae and S. pneumoniae, the infection, despite being painful, is often self-resolving, and although antibiotic therapy hastens its resolution, recurrence is common. Spread of the infection from the ear to produce sepsis and/or meningitis can occur, as can damage to the ear, suggesting that antibiotics still have a role in its treatment. Antibiotic treatment often results in the transformation of acute otitis media (AOM) into sterile otitis media with effusion (OME). However, culture-negative OME fluid may contain viable bacteria (458–460), likely due to the growth of bacteria as biofilm (133,458–460), which conveys different physiologic character to the bacteria, including antibiotic sensitivity.

Excellent technical descriptions of the models are available for the chinchilla (461), guinea pig (462), gerbil (463), and rat (464,465). Briefly, anesthetized animals have their ear canals thoroughly cleaned and are infected by direct administration of bacteria into the ear canal, normally by injection through the thin bone structures of the cephalad bulla. A thorough review of the histopathology and pathophysiology of experimental models of AOM is available (465).

A gerbil model of bilateral AOM induced by either penicillin-resistant or penicillin-sensitive S. pneumoniae, combined with the measurement of ear fluid CFU and drug levels, was used to establish PK/PD parameters for linezolid (466). Following intrabullar injection of S. pneumoniae, peak infection occurred at day 2 for the penicillin-resistant strain and at day 3 for the penicillin-sensitive strain. Linezolid, amoxicillin, or vehicle was administered twice per day over 4.5 days. Amoxicillin was effective only against the sensitive strain, whereas linezolid doses of 10 mg/kg or greater produced cure rates above 72% versus both strains. A similar study (467) demonstrated that, for penicillin-sensitive S. pneumoniae, doses of amoxicillin (>2.5 mg/kg) resulting in ear fluid concentrations = 1.4 µg/mL or serum concentrations greater than the MIC for = 14% of the dosing interval were effective in resolving clinical signs and reducing bacterial CFU.

Using a model of mixed infection with S. pneumoniae and H. influenzae (468), amoxicillin/clavulanate and cefuroxime were evaluated as treatment by stratifying the gerbils according to the presence of effusions. Mixed infections had lower effusion rates than AOM due just to H. influenzae, and treatment of OME was more difficult. Additionally, the mixed infection model was treatable, but more than 80% of the animals developed culture-negative OME. Furthermore, AOM models have been used to demonstrate the potential of antibiotic treatment to promote a protective immune response (469) and immunization (483). Initiation of penicillin treatment early in S. pneumoniaeinfection in chinchillas produced greater inflammation than late treatment (470). The importance of delay of treatment was studied in the gerbil model for amoxicillin against S. pneumoniae with varying susceptibility to penicillin (471). Independent of penicillin susceptibility delay of amoxicillin therapy resulted in lower effect, which was considered due to changes in metabolic activity of the pathogens. The same group showed the same lack of effect with delay of treatment with erythromycin (472).

The addition of dexamethasone to antibiotic treatment reduced the structural damage associated with this infection as compared with antibiotics alone (473). However, treatment of experimental AOM in gerbils with antibiotics plus acetaminophen delayed eradication of H. influenzae as compared with antibiotics alone, possibly due to a reduction of phagocyte recruitment to the site caused by the antiinflammatory agent (474). Addition of ibuprofen therapy to amoxicillin or erythromycin against experimental pneumococcal infection also in the gerbil model did not interfere with antibiotic therapy (475). In comparing mixed S. pneumoniae and H. influenzae with monoinfection by H. influenzae, it was demonstrated that the exact characteristics of AOM or OME in models depends, at least in part, on the time from the appearance of clinical symptoms until the diagnosis/intervention, on the bacteria involved, and on previous antibiotic treatment. Furthermore, poorer eradication rates occurred with lower levels of inflammation, and PK/PD relationships in middle ear fluid provide better predictive value than serum PK/PD parameters (468). These models have also used β-lactamase–positive H. influenzae (476), and in one study, this pathogen did not protect S. pneumoniae against the activity of amoxicillin (477).

Animal Models of Meningitis

Animal models of bacterial meningitis have been considered extremely useful in delineating the pathophysiology of meningitis and elucidating optimal antibiotic and adjunct therapies.

Mouse Models of Meningitis

An increasing number of studies using mouse models of meningitis have been performed recently, in particular, with the use of gene-modulated knockout mice. Mice have been infected with various pathogens (e.g., S. pneumoniae, N. meningitidis, group B streptococci [GBS], E. coli, H. influenzae, and C. neoformans) and using three different routes of inoculation (systemic [478], intracisternal [479–481], and intracerebral [482]). These models primarily involve evaluating survival and to some degree brain histopathologic alterations and have been useful in the study of antibiotic therapy efficacy, adjunctive therapy, and the pathophysiology of meningitis.

Rat Models of Meningitis

A model for the induction of H. influenzae type b meningitis in infant rats, which appears to be both simple and reproducible, has been described (483). Five-day-old rats were inoculated intranasally with H. influenzae type b, and bacteremic rats and rats with meningitis were identified by sampling of the CSF. The rats were sacrificed, the skin and soft tissue over the cisterna magna were removed by dissection for exposure of the dura, and the cisterna magna was entered by puncturing the dura with a sterile dissecting needle. This model system provides a simple method for determining the effectiveness of antibacterials in an acute meningitis infection similar to that seen in human infants (483). The infant rat pneumococcal meningitis model was used to investigate the importance of bacteriolytical properties of daptomycin (nonbacteriolytic) and ceftriaxone (bacteriolytic) on inflammation and brain damage. Daptomycin treatment resulted in more rapid bacterial killing, lower CSF inflammation, and less brain damage than ceftriaxone treatment as measured by lower CSF concentrations of interleukin (IL)-1β, IL-10, IL-18, monocyte chemoattractant protein-1 (MCP-1), and macrophage inflammatory protein (MIP)-1α (484).

An excellent infant rat model for studying survival, brain damage, and learning deficiency has been described (485). This model, which uses intracisternal inoculation of GBS and S. pneumoniae, seems to be able to induce histopathologic alterations that mimic the findings in human meningitis and has provided significant knowledge about the pathophysiology of meningitis. Less brain damage seems to develop in the adult rat model of pneumococcal meningitis, but it has been very useful in the study of cerebrovascular alterations, CSF and brain tissue cytochemistry, and hearing loss (486).

Meningitis in adult rats can be induced via cisterna magna tap with a 23-gauge needle (487). The animals received either 10 µL of sterile saline as a placebo or an equivalent volume of S. pneumoniaesuspension and treatment for 7 days with daptomycin or ceftriaxone was compared (487). Apart from spinal taps for efficacy, the animals underwent separately to four behavioral tasks: habituation to an open field, step-down inhibitory avoidance task, continuous multiple trials step-down inhibitory avoidance task and object recognition. Although both antibiotics were effective in clearing the infection, the investigators found evidence suggesting the potential alternative of the treatment with daptomycin in preventing learning and memory impairments caused by pneumococcal meningitis. In the same model, the same group showed that early ceftriaxone (8 hours vs. 16 hours) administration was an effective strategy to prevent long-term cognitive impairment (488).

The adult rat model has also been used to study hearing loss and cochlear damage as related to pneumococcal meningitis. Hearing loss and cochlear damage were assessed by distortion product otoacoustic emission (DPOAE), auditory brainstem response (ABR), and histopathology in rats treated with ceftriaxone 28 hours after infection. Furthermore, rats were treated with granulocyte colony-stimulating factor (G-CSF) initiated prior to infection, 28 hours after infection or with ceftriaxone only. Rats were followed for 7 days, and assessment of hearing was performed before infection and 24 hours and day 8 after infection. Pretreatment with G-CSF increased hearing loss 24 hours after infection and on day 8 compared to untreated rats and this was associated with significantly decreased spiral ganglion cell counts, increased damage to the organ of Corti, increased areas of inflammatory infiltrates and increased white blood cell (WBC) counts in CSF on day 8 after infection. Initiation of G-CSF 28 hours after infection did not significantly affect hearing loss or cochlear pathology compared to controls (489).

Because of the limited access for repetitive CSF sampling, the rat meningitis model and the mouse meningitis model have only been used sporadically for the study of antibacterial PKs (490). Excellent technically orientated reviews of the infant rat (491) and adult rat (492) meningitis models are available.

Guinea Pig Model of Meningitis

Force and coworkers (493) used both the guinea pig and the rabbit model of meningitis to evaluate the efficacy of meropenem against cephalosporin-susceptible and cephalosporin-resistant pneumococcal infection. Results with meropenem in the experimental rabbit model of penumococcal meningitis have been controversial perhaps due to the possible role of renal dehydropeptidase I in meropenem efficacy, why the investigators wanted to determine the efficacy of meropenem in two meningitis models, and the possible influence of the animal model over results. Meropenem was bactericidal at 6 hours in the guinea pig model against both strains with a reduction of greater than 4 log CFU/mL. In the rabbit model, it was bactericidal at 6 hours against the susceptible strain, but against the resistant, 3/8 therapeutical failures were recorded at 6 hours, being bactericidal at 24 hours. The authors concluded that meropenem showed bactericidal activity in both experimental models and that the guinea pig should be considered the best choice among laboratory animal species when assessing meropenem efficacy (493).

Rabbit Model of Meningitis

The optimal model for studying the PKs and PDs of antibiotics in CNS infections seems to be the rabbit meningitis model, which provides a controlled system for testing antibacterial penetration and efficacy in inflamed and normal CSF (494). Rabbits have been challenged with various pathogens (e.g., Streptococcus pneumoniae, N. meningitidis, Staphylococcus aureus including in later years MRSA, H. influenzae, Listeria monocytogenes, and enterobacteria). The rabbit model allows simultaneous and repetitive sampling of CSF and blood, and it is therefore very useful in kinetic studies of CSF bacterial killing and CSF cytochemistry (495), though less useful in the study of brain damage and survival. For more than 20 years, the rabbit model has yielded considerable information about the PKs and efficacy of antibacterials as well as the pathophysiology of meningitis. The efficacy of antifungals has also been studied using a rabbit model of C. neoformans (496); an excellent technical review of this model has appeared previously (497). Surgical intervention is required to attach prosthesis to the rabbit’s skull, facilitating immobilization of the deeply anesthetized animal. Blood samples (normally 1 to 3 mL) and simultaneous CSF samples (normally 0.1 to 0.2 mL) can be collected at frequent intervals (e.g., 0, 2, 4, 6, and 8 hours after initiation of therapy). The rate of removal of CSF should not exceed the rate of its synthesis (approximately 0.4 mL/hour [498]). In this manner, the antibacterial levels in both the blood and CSF and the bactericidal or bacteriostatic titers were determined. This model provides a controlled system for testing antibacterial penetration and efficacy in inflamed and normal meninges (499).

This model yields considerable information about the PKs and efficacy of antibacterials. In one study (500) using the basic model system described, the PK profile and bacteriologic efficacy of a single dose or continuous infusion of six antibacterials (penicillin, cefoperazone, ceftriaxone, cefuroxime, moxalactam, and chloramphenicol) were determined in two infections (S. pneumoniae and H. influenzae). The PK results following continuous infusion allow a comparative determination of the penetration of the antibacterial into the CSF of infected animals. It is evident that the rabbit meningitis model system can provide a great deal of information about the activity of different agents in a very difficult infection. Antibiotic–pathogen combinations tested in the rabbit model in later years have included daptomycin, vancomycin, and linezolid against MRSA (501); daptomycin, ceftriaxone, and vancomycin against S. pneumoniae (503); moxifloxacin, ampicillin, and gentamicin against L. monocytogenes (504); and doripenem against E.coli and K. pneumoniae (505).

The extrapolation of the results in rabbits to efficacy in humans has generally been good, with one or two exceptions (506).

Animal Models of Brain Abscess

The mortality associated with brain abscesses ranges from 0% to 24%, with neurologic sequellae in 30% to 55% of survivors and the incidence of brain abscess appears to be increasing, likely due to an increase in the population of immunosuppressed patients. A rat model of brain abscess/cerebritis was developed by Nathan and Scheld (507) and used to determine the relative efficacy of trovafloxacin as compared to ceftriaxone in animals infected with S. aureus. Very slowly, 10⁵ CFU of S. aureus in 1 µL was injected with a Hamilton syringe, through a 2-mm burr hole created with a spherical carbide drill just posterior to the coronal suture and 4-mm lateral to the midline. Eighteen hours later, treatment was initiated and continued for 4 days three times a day with ceftriaxone, trovafloxacin, or saline for controls. The brains were removed and the entire injected hemisphere was homogenized and quantitative cultures performed. Both ceftriaxone and trovafloxacin reduced bacterial counts with a factor 1,000, with no significant difference between the two drugs. The authors concluded that trovafloxacin or other quinolones may provide a viable alternative to intravenous antibiotics in patients with brain abscess/cerebritis (507).

Animal Models of Infectious Endocarditis

Experimental endocarditis in rabbits and rats has been well studied and been shown to be reliable for the evaluation of the pathogenesis of the disease and the effectiveness of antibacterials (508–512); it is considered highly predictive of the clinical situation. Technical aspects of the rabbit model and some examples of the type of data that can be obtained have been well described (513). Essentially, a polyethylene catheter is inserted into the right carotid artery and advanced toward the heart; after it crosses the aortic valve, it is secured in place by suturing at the site of insertion. The presence of a catheter in the heart results in the development of sterile vegetations consisting of small, rough, whitish nodules 1 to 2 mm in size, usually at points of contact between the catheter and the endocardium. The sterile vegetations were infected by a single injection of bacteria into an ear vein; S. epidermidis, S. aureus, C. albicans, Proteus mirabilis, and Pseudomonas aeruginosa, among many other strains, have been used in this model. This basic approach has also been applied to rats (511,514).

Combinations of antibiotics in most cases are required to effectively treat endocarditis clinically, and this has been largely predicted by animal models (515). In the study by Batard et al. (516), in vitro checkerboard assays and time-kill curves showed an indifferent response by various Staphylococcus aureus strains (with different antibiotic resistance mechanisms) to the combination of quinupristin-dalfopristin and gentamicin. Using a rabbit endocarditis model and simulated human PK, the authors found no benefit from the combination in vivo, a result predicted by the in vitro testing. An in vitro infection model, which uses simulated endocardial vegetations, has been shown to produce results similar to these of the rabbit model when the PK parameters are known (517). Recent studies have used animal models for evaluating prophylaxis, including the use of azithromycin or ampicillin (518) and trovafloxacin or ampicillin (519) for Streptococcus oralis infection and azithromycin or vancomycin for MRSA (520).

Inflammation of the heart valves occurs in human and experimental endocarditis. The rabbit model was used to evaluate possible benefits of adjunctive dexamethasone regarding the course of experimental aortic valve endocarditis and the degree of valve tissue damage. Using a methicillin-resistant strain of S. aureus, researchers found that combining low-dose dexamethasone with an effective dose of vancomycin had no effect on survival, the blood culture sterilization rate, or the valve bacterial CFU. Dexamethasone adjunct treatment did reduce the inflammation and structural damage to the valves; this study also was able to demonstrate an inverse correlation between neutrophil number in vegetations and degree of tissue damage (520).

A number of studies have evaluated treatment of MRSA and glycopeptide nonsusceptible MRSA aortic endocarditis in rabbits or rats with a range of antibiotics including vancomycin and generics, teicoplanin, daptomycin, telavancin, linezolid, rifampicin, gentamicin, ceftobiprole, tigecycline, garenoxacin, levofloxacin, quinopristin/dalfopristin, and the experimental antimicrobial peptide, plectasin (521–535). Generally, combination treatment was more active than single drugs and daptomycin was more active than vancomycin, which was more bactericidal than linezolid. Two studies evaluated treatment of daptomycin-resistant MRSA isolates in the rabbit aortic endocarditis model and found that either telavancin (531) or a combination of oxacillin with daptomycin (537) was effective.

Several studies have evaluated the optimal treatment for E. faecalis endocarditis in rabbit or rat aortic endocarditis models (537–542). An interesting result was the effect of ceftriaxone in combination with ampicillin for vancomycin-susceptible E. faecalis; however, daptomycin showed superior effect against all types of strains, both vancomycin-susceptible E. faecalis and vancomycin-resistant Enterococcus faecium (540). In a study by Boutoille and coworkers (543), the in vivo impact of the MexAB-OprM efflux system in P. aeruginosa on antipseudomonal β-lactam efficacy (ticarcillin, piperacillin/tazobactam, and ceftazidime) was investigated in the aortic endocarditis model in rabbits comparing two isogenic strains with and without the resistance mechanism. Against the resistant strain, only the high-dose regimens of ceftazidime were effective, with the most significant effect being achieved by continuous infusion. In contrast, all the tested regimens were effective against the susceptible wild-type. In the same model, with a susceptible P. aeruginosa, Navas and coworkers (544) showed that constant infusion with cefepime or imipenem at plasma concentrations three to four times the MIC was sufficient for effect and that addition of tobramycin did not add to the killing effect of the β-lactams.

Colistin’s effect against A. baumannii aortic endocarditis in rabbits was studied by Rodríguez-Hernández et al. (545), who found that although colistin cleared the bloodstream, it could not sterilize the aortic vegetations. Thus, endocarditis models are used extensively to support clinical decisions how to treat endocarditis caused by the rapidly appearing and troublesome antibiotic-resistant pathogens.

Animal Models of Eye Infections

Experimental eye infections have received much attention for the evaluation of antiinfective therapy. Animal models of keratitis, endophthalmitis, and eye injury and conjunctivitis are available (546). Technical descriptions of the rabbit model of conjunctivitis (547) and the mouse model of bacterial keratitis (548) have been provided.

Experimental Keratitis

Keratitis can be established by inoculation of the surface of an eye of an anesthetized animal damaged by scratching the surface with a syringe needle (e.g., a 26-gauge needle) or by direct injection into the cornea. The rabbit is used most often due to the size of its eye also for clinical, macroscopic evaluation, but rats and mice have also been studied. For therapeutic studies, antibacterial therapies by a parenteral or topical (or combined) route are started at different times, but usually within 24 hours after infection. Topical treatment usually is frequently applied to the surface of the eye or administered by less frequent (often only once) intravitreal injection. Normally, the concentration of antibiotic in the aqueous humor correlates more closely with therapeutic efficacy than does the concentration in the cornea. Although many antibacterials can extensively reduce the number of bacteria in the cornea, typically by more than 99% in the first 24 hours of therapy, sterilization of the cornea is difficult and may require several additional days of continuous therapy (548).

P. aeruginosa mutants with a lipopolysaccharide (LPS) core and O antigen defects exhibit reduced viability after internalization by corneal epithelial cells, and a complete core LPS is required for full epithelial invasion (549). Despite effective antibacterial therapy, disease resolution can be delayed with respect to the time of bacterial eradication (550). Ofloxacin (mammalian cell penetrable) and tobramycin (less cell permeable) have been tested against invasive and noninvasive P. aeruginosa in a mouse model of keratitis. Topical ofloxacin and tobramycin, with or without prednisolone acetate, were administered hourly as eye drops for 12 hours postinfection. Tobramycin was less effective than ofloxacin against the invasive strain, but in the other groups, antibiotic treatment was effective against both strains. However, despite effective antibacterial treatment, disease progression continued in all groups, and differences in responses to treatment were not manifest until day 7 (550).

The rabbit keratitis model has been used to study antimicrobial treatment, in most cases by topical application of antimicrobial solutions, for infections caused by Staphylococcus aureus, both methicillin susceptible and resistant (551–560), Staphylococcus epidermidis (555), P. aeruginosa (559–564), Serratia marcescens (562,563), Mycobacterium chelonae (565,566), C. albicans (248), Fusarium sp (567,568), and Acanthamoeba sp (569). The following drug–pathogen combinations have been studied: MSSA and methicillin-susceptible Staphylococcus epidermidis (MSSE): cefazolin, vancomycin, tobramycin, chlorohexidine, benzalkonium, and a number of fluoroquinolones (ciprofloxacin, ofloxacin, levofloxacin, moxifloxacin, gatifloxacin, gemifloxacin, and besifloxacin) (551–560); P. aeruginosa: gentamicin, tobramycin, ciprofloxacin, ofloxacin, moxifloxacin, levofloxacin, gatifloxacin, and chlorhexidine (559–564); M. chelonae: amikacin, clarithromycin, ciprofloxacin, levofloxacin, and gatifloxacin (565,566); C. albicans: amphotericin B, natamycin, and caspofungin (248); Fusarium sp: amphotericin B, caspofungi, itraconazole, and voriconazole (567,568); and for Acanthamoeba sp: chlorhexidine and neosporin (569). In general terms, most antimicrobials administered topically show effect in reducing pathogen CFUs significantly as compared to controls if the pathogens are susceptible to the drugs in vitro and doses are high enough. Furthermore, early treatment is better than late treatment (553). Even chlorhexidine showed effect against MRSA and P. aeruginosa (560).

Experimental Intraocular Infections

Several methods to obtain reproducible intraocular infections in laboratory animals have been described and many are modifications of the method discussed here (570). In this method, S. aureus, E. coli, or P. aeruginosa is inoculated into the center of the rabbit cornea, the anterior chamber, or the vitreous of the eye, and samples of the vitreous humor, irises, and anterior chamber as well as the retina are used to determine the progress of the infection. The authors (570) found that, when 3 × 10⁶ CFU/0.2 mL of broth were inoculated into the corneas, anterior chamber, and vitreous of rabbit eyes, a virulent panophthalmitis was produced within 24 to 48 hours, and destruction of the eye took place within 72 hours regardless of the site of inoculation. When 5 × 10³ CFU/0.2 mL were inoculated into the corneas, anterior chambers, and vitreous, a panophthalmitis resulted in 72 hours. The infections were most severe following intravitreal inoculations and less intense when the anterior chamber was the site of inoculation. When 7 × 10² CFU/0.02 mL were used as the inoculum, the infections were eliminated in the corneas and anterior chambers within 24 hours but not in the vitreous. The anterior chambers were most resistant, the cornea slightly less, and the vitreous the least resistant to virulent infection.

Infectious endophthalmitis is characterized by an inflammatory reaction in a sensitive, normally immune-privileged or protected tissue. Depletion of circulating neutrophils by i.v. administration of specific antibody at 6 or 12 hours after intravitreal injection of S. aureus into rats resulted in diminished neutrophil influx, lower and delayed clinical and histopathologic evidence of disease, but also a reduction in bacterial clearance from the eye (571). The inflammatory response appears to lag behind bacterial growth of either S. epidermidis or P. aeruginosa, for a maximum in the number of microorganisms was reached earlier than the influx of leukocytes (572). In S. epidermidisendophthalmitis, the number of microorganisms reached a maximum at day 2 after intraviteral inoculation and then declined spontaneously; clinical scores were the worst on day 5 but poor scores persisted in the absence of detectable bacteria. In P. aeruginosa endophthalmitis, the number of microorganisms reached a maximum 36 hours after inoculation, and bacteria were detectable for 15 days.

Ravindranath et al. (573) measured the immune response during endophthalmitis. Rats received an intravitreal injection of viable S. epidermidis that resolved by day 14. The inflammatory cell content of the vitreous switched from neutrophilic to monocytic-macrophagic/lymphocytic by day 3 postinfection. B cells (CD45+/CD3−) were also detected, and IgM and IgG antibodies but not IgA antibodies to glycerol teichoic acid were found in the vitreous of injected eyes; IgM antibodies declined by day 7 postinfection. Anti-GTA IgM was observed in vitreous and serum, anti-GTA IgM antibodies were significantly elevated, but a weak IgG response and no IgA response were observed in serum S. epidermidis–infected rats (573).

The use of adjunct antiinflammatory agents in endophthalmitis is controversial. Intravitreal vancomycin once plus 7 days of i.m. methylprednisolone was not as effective as vancomycin alone in reducing ocular inflammation and improving retinal function in experimental S. aureus endophthalmitis (574). Using a variety of antibiotics in an animal model of S. aureus endophthalmitis, researchers found that the combination of vitrectomy and injection of intraocular vancomycin was the most effective regimen and that no improvement resulted from administering adjunct i.v. corticosteroids (575). The timing of dexamethasone treatment appears to influence outcome (576). Bilateral eye S. aureus infection in rabbits was treated once with vancomycin in one eye and vancomycin plus dexamethasone in the other at 24, 36, 48, or 72 hours after intraviteral infection. Early combination treatment (at 24 or 36 hours) produced reduced ocular inflammation as compared with antibiotic alone, but only when treated at 36 hours postinfection did the combination group preserve retinal function better than vancomycin alone; no treatment was able to eradicate infection. However, dexamethasone has improved antibacterial and antifungal effect and reduced inflammation in several subsequent studies of endophthalmitis (577–582).

Many studies have evaluated the PK of antibiotics in eye tissue. PK modeling of antibiotic eye-blood barrier penetration data has been used to compare different formulations (583) and explain in part why eye penetration in rodents may overestimate antibiotic penetration due to the dependence of larger species on convective fluid flow (584). A single intravenous administration of 5 or 20 mg/kg moxifloxacin demonstrated good penetration into the vitreous, but apparently, the penetration was dose-independent and increased when inflammation was present (585). Microdialysis has been used to both measure drugs (e.g., ceftazidime [586] and vancomycin [587]) and dispense drugs (588). Inflammation or eye trauma appears to increase the penetration and residency of many topically applied antibiotics (e.g., levofloxacin [579], ofloxacin [589], ciprofloxacin [590], and vancomycin [575]), but apparently not ceftazidime, whose half-life is decreased by inflammation and eye surgery (575). A 1% vancomycin hydrochloride ophthalmic ointment was administered to the corners of the eyes of rabbits with Bacillus subtilis infection. Vancomycin reached effective concentrations in the aqueous humor and extraocular tissues but was not detected in the aqueous humor, iris-ciliary body, vitreous, or serum in uninfected animals; nonetheless, the presence of inflammation permitted concentrations to reach potentially therapeutic levels in these tissues (591).

Systemic treatment with moxifloxacin demonstrated effectiveness against MRSA and MSSA (585), as did treatment with trovafloxacin against S. epidermidis (592); this is somewhat surprising, as most antibiotics show poor ocular penetration. Despite inflamed eyes demonstrating improved penetration of i.v. gentamicin or amikacin, aminoglycoside levels in the eye failed to reach therapeutic concentrations sufficient for either Pseudomonas spp or S. epidermidis (593). However, systemic administration of sparfloxacin, pefloxacin, or imipenem (though not vancomycin or amikacin) was effective as a prophylaxis against intravitreal S. aureus challenge (594). Combined topical and oral ofloxacin (590) or ciprofloxacin (589) increased the ocular levels of the drug in a model of posttraumatic endophthalmitis due to S. aureus infection (590). Topical treatment with liposomal formulation increased the half-life of fluconazole (595) but was inferior to free fluconazole in the treatment of C. albicans endophthalmitis (596; see also 597). Intravitreal treatment with vancomycin plus amikacin was effective against vancomycin-sensitive E. faecalis, and intravitreal ampicillin plus gentamicin was effective against vancomycin-resistant E. faecalis endophthalmitis (598). Gatifloxacin or ofloxacin ophthalmic ointments prevented E. faecalis endophthalmitis when administered 1-hour postinfection, but this effect decreased with application of antibiotics with longer intervals after infection (599). Intravitreal treatment has been shown to reduce pathogen counts significantly for vancomycin and moxifloxacin against Bacillus cereus (600), linezolid, vancomycin, imipenem, ceftazidime, amikacin, moxifloxacin, and other fluorquinolones against MSSA (601–603); levofloxacin, piperacillin/tazobactam, and ceftazidime against P. aeruginosa(604,605); tigecycline against A. baumannii (606); and amphotericin B and caspofungin but not itraconazole or voriconazole against C. albicans (243).

Animal Models of Osteomyelitis

Although of relatively low incidence, bone and joint infections are difficult to cure, largely due to limited penetration of antibiotics, coupled with the fact that slow-growing or adherent bacteria are likely to be more resistant to antibiotics. Furthermore, designing and executing clinical trials is difficult due to the likelihood of low recruitment, the heterogeneity of the disease, and the many hard-to-control factors influencing treatment outcome. Consequently, advances in clinical management have heavily relied on the contribution of animal models (607).

The experimental conditions for the rabbit model initially described by Norden (608,609) have recently been reviewed (610). In one study using this procedure, 89% or more of the animals developed osteomyelitis (608,609). The infecting organism was recovered from 91% of rabbits sacrificed 60 to 180 days after infection. If blood samples were taken 6 hours after infection, more than 80% were positive for the infecting organism, but by 24 hours, less than 20% were positive. Injection of a bacterial suspension or sodium morrhuate alone did not cause osteomyelitis, as evidenced by radiologic examination or culture of the bone. Antibacterial therapy was initiated 1 to 14 days after infection. Because the radiologic changes of chronic osteomyelitis were present at day 14, treatment at this time was considered to represent therapy of chronic osteomyelitis.

The rabbit model has also been used to study treatment of MRSA osteomyelitis with ceftaroline as compared to linezolid and vancomycin and tigecycline or vancomycin in combination with or without rifampicin (611).

The rat is another commonly used animal for osteomyelitis (612). This model is widely used for experimental chemotherapy, and the practical aspects of this model have been previously described (613). Occasionally overlooked, it is critical to culture at least some remaining crushed bone, initially in broth with subculture on agar plates, in order estimate bone sterility. Typically, single agents are weakly active, and combination chemotherapy generally produces superior results (614). More thorough determinations of PK/PD relationships for osteomyelitis are warranted. The rat tibia osteomyelitis model has been used to study the effect of fosfomycin (615) and tigecycline as compared to teicoplanin (616) against MRSA-induced infection, all drugs showing significant effect as compared to untreated controls.

These models have also been used to evaluate various drug delivery systems (e.g., a sulbactam-cefoperazone polyhydroxybutyrate-co-hydroxyvalerate depot formulation [617], several depot formulations [618], tobramycin pellets [619], tobramycin fibrin sealant [620], and antibiotic-impregnated hydroxyapatite [621]), hyperbaric oxygen combined with antibiotics (622), and adjuvant treatment with granulocyte-macrophage colony-stimulating factor (GM-CSF) (623). Further, recent studies have used vancomycin-loaded borate glass or calcium sulfate (624), vancomycin-coated titanium plates (625), vancomycin or daptomycin-loaded beads (626), polyelectrolyte multilayers with gentamicin (627), biodegradable dilactide polymer releasing ciprofloxacin (628), a synthetic semihydrate form of calcium sulphate impregnated with moxifloxacin (629), gentamicin-vancomycin–impregnated polymethylmethacrylate (PMMA) coating nail (630), and calcium-deficient apatites with linezolid (631), in all cases with preventive effect on staphylococcal infections.

An interesting study performed by Nijhof et al. (632) demonstrated that combined tobramycin bone cement and systemic cefazolin was superior to monotherapies, suggesting that combining local and systemic treatments might be a useful approach to treatment of osteomyelitis. Furthermore, this study demonstrated the utility of measuring bacterial DNA, as persistence of DNA may occur despite effective treatment.

Osteomyelitis models with precontaminated (methicillin-susceptible or resistant S. aureus or S. epidermidis) foreign bodies inserted into the medulla of either tibia or femur of rabbits or rats have been used to study treatment effect with various systemically administered antibiotics or even electric current (633–636). Various treatment durations from 7 to 28 days were used, why comparisons among studies are difficult.

Animal Models of Mycobacterium Infections

Models of Disseminated Mycobacterium avium Infections

Models of disseminated M. avium infection have usually used beige mice (bearing an Nramp1 mutation), although alternatives include C57BL/6 mice (637); hamsters (638); and immunosuppressed, cyclosporine-treated rats (approximately 0.03 mg cyclosporine/kg [639]). Reviews of the technical aspects of the beige mouse model of M. avium infection have been published (640,641).

Several details of the model have been described (e.g., the influence of the route of infection) (641,642). Infections are initiated by intravenous injection of large inocula (10⁷ to 10⁸ CFU). Although the disease is usually nonfatal, high numbers of M. avium are found in the liver, spleen, and lung, and determination of the efficacy of treatment is accomplished by ascertaining the organ bacteria loads. Furthermore, mild chronic CNS infection develops in the mice during sustained systemic M. avium infection, similar to what has been reported in most human cases. In one study, M. avium was detected initially in the parenchyma of the choroid plexus but also in the ventricles and meninges. However, the mice did not develop clinical signs nor did they die due to CNS involvement (643). Iron restriction inhibits the in vitro and intramacrophagic growth of M. avium, and mice fed an iron-poor diet experienced reduced M. avium proliferation; administration of iron chelators had small effects, as they impacted little on the iron status of mice (644).

In some studies, the beige mouse model demonstrated poor outcome against M. avium in the testing of marketed antibiotics (645). Note that in vitro tests do not always accurately predict in vivo susceptibility (646). Although previous reports indicated a benefit to infected mice, administration of recombinant G-CSF failed to improve the course of M. avium infection in C57Bl/6 or beige mice and did not enhance the activity of the combination of clarithromycin plus ethambutol plus rifabutin (647). A rather extensive combination of antimicrobials was tested by Fattorini et al. (648). The activity of 18 anti–M. aviumregimens was evaluated. Mice were treated with clarithromycin, ethambutol, amikacin, rifabutin, ciprofloxacin, or clofazimine alone or in combination. Monotherapies were less effective than combinations, and resistant M. avium emerged. Some two-drug combinations were active, but none more than clarithromycin alone. The triple combination of clarithromycin, amikacin, and ethambutol was the most effective (648). Moxifloxacin is active against M. avium in combination with other agents (649).

Animal Models of Mycobacterium tuberculosis Infections

Animal models have been used to model tuberculosis since Robert Koch started his pioneering work on the infecting pathogen, M. tuberculosis. The use of M. tuberculosis–infected animals for testing antituberculosis drugs date back to the start of the 19th century, but have gained broad and increased importance in later years due to better understanding of the PK/PD of available drugs as well as for testing a range of novel drugs or other treatment strategies in the era of multi- of panresistant M. tuberculosis (650). In 2012, a group of experts, funded by the Bill and Melinda Gates Foundation, published a comprehensive review on the analysis of methods used for the evaluation of compounds against M. tuberculosis (651). In addition to in vitro methods, the review presents a detailed discussion of the available animal models based on the literature as well as personal visits and interviews with most of the pharmaceutical and academic investigators working with these models. This review is highly recommended for both the interested infectious disease specialist as well as the researcher who intends to embark on an animal model for the study of treatment of experimental tuberculosis (651). The availability of this review also precludes the need for a detailed review of animal tuberculosis models in this book. In short, although a range of animals have been used for experimental tuberculosis including nonhuman primates, guinea pigs, hamsters, rabbits, rats, and mice, the guinea pig as the historic model for M. tuberculosis infections has largely been supplanted by mice (in- and outbred) models which now remain the most commonly used (651). Among the many explanations for its popularity are small size, ease of handling, low prize, small volumes only needed for expensive drugs, ease of inoculation, similarity to humans of pulmonary infection induced and proven predictability of treatment of human infections. On the negative side are factors such as variability in response to infection and tolerance to drugs of different mouse strains, PK behavior of drugs different from humans, and many others. Further adding to problems with animal tuberculosis models are major differences in virulence of and host response to the M. tuberculosis strains generally used for animal studies, size (1 to 4 aerosolized bacilli up to 10⁶ CFU installed intratracheally) and route of inoculum (inhalation, intratracheal installation, nasal application, intravenous injection), type of infectious processes occurring in lungs and other organs (extra- or intracellular bacteria, silent or latent infection, biofilm), relapse, re- or superinfection, etc. The review called for standardization and comparison among models which has already resulted in published studies on these issues (652). Most of the discussion on the earlier mentioned issues is extremely relevant and pertinent for all animal models discussed in this chapter, why the review is recommended reading for all scientists working with experimental animal models.

Animal Models of Sexually Transmitted Diseases

Animal models of human sexually transmitted infections (STIs) can be problematic owing to high-level specificity many of the STI pathogens display for a human host. Despite this, a number of models have been successfully developed. In some cases, investigators have used a microbial species that is distinct from that which causes human disease but is related and specific for the urogenital tract of the animal model used.

Disseminated Gonococcal Infection in Mice

Mice were traditionally considered resistant to disseminated gonococcal infection (653) despite some initial descriptions by Corbeil and colleagues (654). Many early studies used an infection route with subcutaneous chambers to study pathogenicity and therapeutic effects of antimicrobial therapy (655–658). However, more recently, a murine model has been developed and used by a number of investigators (659,660). Mice are made susceptible to colonization and infection with Neisseria gonorrhoeae via the combination of pretreatment with antibiotics (e.g., vancomycin and TMP-sulfa) and estradiol (659,661). A recent interesting finding was a functioning MtrCDE multidrug efflux system enhanced experimental genital tract infection in female mice (662). This discovery has led to additional studies on delineating the genetic mechanisms of regulating expression for the efflux system and its effect on in vivo fitness (663,664). The model has also been used recently to examine the effects of fluoroquinolone resistance mutation development and associated compensatory mutations to restore wild-type fitness (665).

Syphilis

Animal models examining antimicrobial efficacy in syphilis have been developed for localized disease (i.e., dermal), genitourinary tract disease, CNS/disseminated disease, and congenital disease. Most models use rabbits as the animal host; however, a model in the hamster has also been well described (666).

Localized dermal infection models were the first to be developed in rabbits. Localized infection is induced by intradermal injection of live spirochetes. Antimicrobial therapy is usually withheld until signs of an active syphilitic lesion are present and confirmed by dark-field microscopic analysis of a skin scraping. Once active infection is confirmed, antimicrobial therapy is administered to the rabbits. Representative examples of the use of this protocol to test antimicrobial efficacy for localized disease includes study of penicillin G (667), aztreonam (668), cefetamet (669), cefmetazole (670), an investigational penem (667), ceftriaxone (671), and azithromycin (672). A similar hamster model of intradermal infection has been employed for study of clarithromycin efficacy (673).

Genitourinary tract disease models have primarily been limited to orchitis infection models. In this model, the rabbits receive an inoculum of syphilis spirochetes directly into the testes. It has been used on a limited basis to determine drug efficacy, with encouraging results from a study of ceftriaxone, ceftizoxime, and penicillin G (674,675).

Disseminated disease can be induced by intravenous or intraperitoneal inoculation, although there has been limited use of this model for examining antimicrobial therapy (667). More commonly disseminated disease that includes CNS infection is used. CNS infection with syphilis can be achieved by direct intracisternal injection of Treponema pallidum. Marra and colleagues (676) used this procedure to develop a rabbit model of CNS syphilis that very closely mimicked human disease including a 6% rate of uveitis in the animal model (676). A year later, the same group of authors demonstrated enhanced efficacy for penicillin G versus ceftriaxone in the rabbit model (677). The model has also been used to examine whether certain strains of T. pallidum exhibit increased neuroinvasion (678).

Finally, congenital syphilis has been described in a rabbit and hamster model (679–681). However, antimicrobial therapy in congenital syphilis models has not been well explored.

Chlamydia Genital Tract Infections

Chlamydia trachomatis is a major cause of STI worldwide and remains the most common STI in the United States. Despite its long history and medical importance, attempts at developing an animal model have proved difficult (682). Reproducible establishment of infection of the upper genital tract in female mice with human isolates of C. trachomatis requires hormonal manipulation, inbred animals, and surgical intervention to place the organisms directly into the site (i.e., to produce salpingitis) (683). However, in 1994, Beale and Upshon (683) developed a novel upper genital tract infection model in mice. They used C. trachomatis MoPN (primarily a mouse respiratory pathogen) and were able to demonstrate upper genital tract disease in progesterone-treated mice administered the inoculum by intravaginal injection. Treatment studies with minocycline, doxycycline, amoxicillin-clavulanate, and azithromycin each demonstrated efficacy when initiated 1 or 7 days postinfection. Both doxycycline and azithromycin were highly effective in restoring animal fertility. A study of azithromycin efficacy in female mice demonstrated the antimicrobial agent could reverse chlamydial-induced damage and restore fertility if administered within 2 or 7 days of infection (684). Conversely, if administered 12 or more days after infection, even at very high doses, antimicrobial therapy failed to prevent infertility. A male murine model of genital tract disease caused by C. trachomatis MoPN has also been described (685) that also closely mimics disease in human males. However, evaluation of antimicrobial therapy in this model has not been performed. Another strategy to circumvent problems with establishing genital tract disease with human isolates of C. trachomatis is to utilize a microbial species that does cause intrinsic genital tract infection in the animal host. This has been accomplished using the isolate Chlamydia muridarum to infect the urogenital tract of mice (686–691) and Chlamydophila caviae in guinea pigs (692–694).

Animal Models of Peritonitis

Peritonitis can be established in animals either by direct intraperitoneal infection of fecal material (often encased in gelatin capsules) or by puncture of the cecum to provide a focus of infection. These models have been reviewed previously (695).

Intraperitoneal Inoculation

A rat model for simulating intraabdominal sepsis, either with known organisms or with mixed fecal flora cultures, has been developed by Weinstein et al. (696). A uniform inoculum was prepared from the pooled cecal and large bowel contents of 15 rats that had been maintained on a diet of lean ground meat and water for 2 weeks. The value of this mixed infection model for evaluating agents for effectiveness in preventing mortality and the formation of intraabdominal abscesses was confirmed as follows. A clindamycin–gentamicin combination reduced mortality rates and the formation of the abscesses, whereas either drug alone reduced mortality rates (gentamicin) or abscess formation (clindamycin) but not both (696).

Cecal Ligation and Puncture

The cecal ligation and puncture (CLP) model has been described for both mice (697), rats (698), and rabbits (699). The description given by Hyde et al. (700) is typical of this approach to induction of lethal peritonitis in animals as a model of postsurgical sepsis. By measurement of bacterial counts in peritoneum, blood, and various organs, as well as antigens and immunologic factors in peritoneum and blood (LPS, TNF-α, ILs, etc.), the model has been used to study pathogenicity and effect of antimicrobial agents and biologic response modifiers (701). Combination immunotherapy with soluble tumor necrosis factor receptors plus IL-1 receptor antagonist decreases mortality in the CLP model (702). Similarly, a range of different antimicrobial peptides (tachyplesin III, pexiganan, cathelicidine [LL-37], S-thanatin, tritrpticin, indolicidin) have been shown to both have antibacterial effect in vivo as well as reducing unwanted immunologic responses, and these effects have shown reduced mortality when combining with various antibiotics (703–708).

However, the cautionary suggestions by Eichaker and associates (23) need to be heeded when using this model, as it is apparent that more severe sepsis, which is usual for the model, benefits best from chemotherapeutic intervention but may not represent a typical clinical presentation.

The CLP model has been used to evaluate some parameters concerning the differential sensitivity of internal organs to damage during sepsis (709). Increases in microcirculatory permeability were greater in the lung than in the liver 12 hours after CLP, and increases in water mass fraction were greater and occurred earlier in the lung than in the liver. The CLP model has been used to evaluate cardiovascular responses during sepsis characterized by an early hyperdynamic phase followed by a late hypodynamic phase (710). Mice made septic by CLP demonstrated hypotension and a hyperdynamic state that could be monitored using manometric catheters and echocardiography and could be modulated with fluid resuscitation and antibiotics (711).

The administration of the antioxidant phenyl N-tert-butyl nitrone (PBN) (150 mg/kg 30 minutes after CLP), followed by the antibiotic imipenem (10 mg/kg 1 hour after CLP), significantly increased survival compared with other single treatment groups. However, the increase in survival found in the PBN plus imipenem-treated group was abrogated by anti–IL-10 antibody, indicating that endogenous IL-10 is an effective protective factor (712).

Animal Models of Infected Abscesses

The technical aspects of intraabdominal abscess models have been presented previously (713). As an example of the formation of abscesses by anaerobes, a model of subcutaneous abscesses in mice caused by Bacteroides fragilis has been described (714). Among the advantages of this model are that host factors involved in abscess formation can be studied and that the PK properties of the antimicrobial agents, especially penetration of the abscess, can be assessed. Both the inoculum size and the time of treatment significantly altered the effectiveness of both clindamycin and cefoxitin. In a larger experiment, the rank order of antibiotic efficacy could clearly be determined and related to the peak levels of antimicrobials in serum and abscesses (714).

A model of intraperitoneal abscess formation by S. aureus has been described (715).

Animal Models of Gastrointestinal Infections

As with sexually transmitted diseases, animal models of gastrointestinal tract infections are hampered because of the apparent specificity of the causative pathogens for humans, particularly in the case of Helicobacter infections.

Clostridium difficile Enterocolitis

Diarrhea following antibiotic chemotherapy can be caused by Clostridium difficile and is referred to as pseudomembranous colitis or antibiotic-associated colitis (716). Proliferation of C. difficile following suppression of normal gastrointestinal flora results in an enteric intoxication owing to the release of exotoxins TcdA and TcdB from C. difficile. Current standard treatments include vancomycin (125 to 200 mg/kg orally four times per day) or metronidazole (250 to 500 mg/kg orally four times per day) (717). C. difficile infection has been studied in a number of animal models, including hamsters, guinea pigs, rabbits, germ-free piglets, germ-free mice, and conventional mice (718–723).

Inoculation of experimental animals can be done either with spore preparations via the oral route or direct instillation into the colon, or by instillation of toxin preparations into the lumen of the colon. Typically, antibiotics (e.g., mixtures of aminoglycosides, colistin, metronidazole, and vancomycin) are administered prior to inoculation in order to change the intestinal flora, which renders the animals more susceptible to C. difficile infection.

Golden Syrian hamsters are widely used to model this disease because they have the causative organism as part of their normal flora, are known to succumb to antibiotic-induced colitis (724,725), and are sensitive to the activity of C. difficile toxins (726). Typically, clindamycin (0.8 to 100 mg/kg intraperitoneally or subcutaneously once) is administered, and death attributed to C. difficile (by virtue of culture and detection of toxins) ensues rapidly (up to 100% mortality within 1 week), though it is dependent on the clindamycin dose. Modifications of this model include the orogastric infection of clindamycin-treated hamsters with axenic C. difficile (4 × 10⁵ spores/clindamycin-treated hamster [727]). Modulation of the diet of hamsters can alter the extent of C. difficile disease. As compared with hamsters fed a normal fat and cholesterol diet, hamsters fed an atherogenic (defined high-fat) diet have increased susceptibility to C. difficile (728).

Peptidic antibiotics (e.g., vancomycin) appear to be efficacious in this model, but most must be given continuously because regrowth of C. difficile occurs following cessation of treatment (729). Following p.o. clindamycin treatment and inoculation with a toxigenic C. difficile strain, hamsters developed C. difficile–associated ileocecitis and 3 days later were treated with intragastric nitazoxanide (30 to 150 mg/kg), vancomycin (50 mg/kg), or metronidazole (150 mg/kg). All three compounds inhibited the appearance of C. difficile gastroenteritis symptoms, but upon treatment cessation, the hamsters relapsed, indicating failure to eradicate C. difficile. In a prophylactic mode, only nitazoxanide produced hamsters free of clinical symptoms, histopathology, or residual bacteria (730). C. difficle is affected in vitro by sub-MIC levels of antibiotics that establish conditions that precipitate disease (e.g., amoxicillin, clindamycin, cefoxitin, and ceftriaxone) and those antibiotics used for treatment of established infection (vancomycin and metronidazole). The sub-MIC effects are, however complex, strain-dependent and affect both bacterial growth (increasing lag time and overall growth rate) and the timing of initiation of toxin production (faster initiation of production). These effects were observed with clindamycin, metronidazole, and amoxicillin, rarely with vancomycin, and never with cefoxitin (731). However, the in vivo significance of the sub-MIC effects remains to be determined.

The hamster model has been criticized for being too sensitive to C. difficile developing clinical signs of disease already after 2 to 3 days and thereafter showing high mortality with death in a few days (732). Mice models have gained increasing interest due to more closely resembling the disease seen in humans (733). Although true relapses do not occur in hamsters due to the acute infection, a relapse model has been developed in conventional mice (733). Using the mouse models, several investigators have shown that vancomycin treatment only delays disease occurrence, while, for example, antibodies against the two exotoxins do protect animals against relapse (733,735). One of the reasons for the missing posttreatment effect of vancomycin is its lack of activity against C. difficile spores (736). Other glycol- or lipopeptides such as ramoplanin or oritavancin do not appear to induce germination of C. difficile(736,737). The mouse model has also shown the importance of antibiotic-induced reduction of the anaerobic gram-negative flora for promoting colonization and infection with C. difficile (738,739).

Models of Helicobacter Gastric Infection

The association of Helicobacter pylori with gastric ulcers and gastritis has dramatically altered approaches to gastroduodenal disease to now include antimicrobial chemotherapy as a therapeutic modality (740). Normal laboratory animals apparently are difficult to colonize with H. pylori but are more susceptible to the related species Helicobacter felis and Helicobacter mustelae. Development of a suitable animal model has delayed evaluation of novel antibacterial strategies because in vitro antimicrobial susceptibility is apparently not consistently predictive of in vivo efficacy (740). Some models have used predisposition of the animals with acetic acid to induce a gastric ulcer, followed by orogastric inoculation; H. pylori is capable of efficiently colonizing these heavily damaged areas. Marchetti et al. (741) presented evidence suggesting that passage of clinical H. pylori isolates in the gastrointestinal tracts of mice selects for those bacteria with increased colonizing ability while maintaining several of the features of the disease in humans.

Various models have demonstrated that combination chemotherapy is more effective than monotherapy in eradicating Helicobacter organisms from gastric tissues. Except for treatment with metronidazole, monotherapy and dual therapies involving amoxicillin, bismuth subcitrate, and clarithromycin (with or without the proton pump inhibitor omeprazole) did not cure mice bearing H. pylori Sydney strain gastric infection (742). The triple therapies of OMC (omeprazole, metronidazole, clarithromycin) and BMT (bismuth subcitrate, metronidazole, tetracycline) were more successful in eradicating infection. However, these treatments also produced a different pattern of stomach colonization, suggesting the antrum-body transitional zone is a “sanctuary site” harboring H. pylori in cases of treatment failure. The authors concluded that there was good correlation between the Sydney strain mouse model and antibiotic therapy outcome in humans (except for metronidazole monotherapy and OAC triple therapy).

Combination treatments involving proton pump inhibitors or antiulcer agents have been evaluated in animal models. The apparent synergic anti–H. pylori effects of the proton pump inhibitor lansoprazole are apparently due to enhanced penetration of orally administered amoxicillin in gastric mucus and tissue by lansoprazole-produced increased intragastric pH. Supplemental treatment of rats with clarithromycin did not affect this drug interaction (743). Using a C57BL/6 mouse model, the cytoprotective antiulcer agent plaunotol was shown to enhance the activity of clarithromycin or amoxicillin in the treatment of H. pylori infection (744). In the same model hyperimmune bovine colostrum with N-acetylcysteine and zinc in combination with amoxicillin in high-dose eradicated H. pylori in all mice treated for 10 days (745).

Animal Models for Determining the Effect of Age on Susceptibility to Infection

Host susceptibility and response to infection change dramatically based on age. It has been well established that very young (i.e., neonates) and elderly humans are more susceptible to certain infectious diseases, and this can often be attributed to changes in immune function at the extremes of age. Neonatal animal models have been established to study common infectious diseases noted in this group including GBS infection (746,747), staphylococcal infection (748–751), and invasive candidiasis (287,751). A limited number of these studies also included an evaluation of antimicrobial therapy. Conversely, advanced age animal models have focused primarily on specific immune function such as studying innate immune responses (e.g., cytokine response and neutrophil function) as well as adaptive immune responses (e.g., T- and B-cell specific responses). For example, Boyd and colleagues (692) studied the toll-like receptor (TLR)-2 responses in mature and aged mice infected with pneumococcal pneumonia. In a previous study, they demonstrated senescent mice were more susceptible to pneumococcal pneumonia through potential priming effect of chronic inflammation and TLR dysfunction (60). In a more recent study, they were able to show that aged macrophages exhibit impaired TLR-2–dependent recognition of pneumococcus and this was associated with a delayed proinflammatory cytokine response in vivo along with enhanced susceptibility to pneumococcal pneumonia. Other common infections noted in elderly humans have also been modeled in aged animals including bacterial peritonitis, intraabdominal abscess and sepsis via cecal ligation, invasive candidiasis, and C. difficile (752–756). Unfortunately, while susceptibility to infection has been examined in aged animal models, study of antimicrobial efficacy in these models has been limited.

In comparison with other disease areas (e.g., cancer), antiinfectives are generally exceedingly well tolerated within a wide therapeutic window (the difference between the dosage required for maximal desired pharmacologic effect and the dosage at which unwanted effects become apparent). This is likely in part due to the drug substance being directed to targets unique to microorganisms or distinct from mammalian counterparts. Preclinical in vitro and in vivo toxicity/tolerability testing of novel antibiotics is required prior to initiating clinical trials (757), and although this will reveal toxic effects, they generally occur at doses far in excess of those required for efficacy studies in animal models. There are five main types of toxicity associated with antimicrobials (758): direct effects, where the molecule acts against the host tissues (e.g., anemia caused by chloramphenicol is inhibition of mitochondrial protein synthesis); hypersensitivity, which can include allergic responses (e.g., anaphylactic reaction) or other nonallergic reactions such as diarrhea; disorders resulting from antibiotic-produced changes in bacterial flora (e.g., vaginal yeast infections); drug interactions, where coadministered drugs interact deleteriously together (e.g., ketoconazole interference with cytochrome P450–mediated drug metabolism); and microbial lysis, with the associated massive release of proinflammatory bacterial products that can cause widespread tissue damage, as occurs in bacterial meningitis.

An excellent overview of the use and tolerability of some older (but still clinically used) antibiotics in laboratory animals has been prepared by Morris (759). Experience has indicated that guinea pigs and, to a lesser extent, rabbits tolerate antibiotic treatment much less well than mice or rats. There are species and strain differences in tolerability to antibiotics (760). Strain is critical to the tolerance of certain compounds; for example, tobramycin is more toxic to Fischer rats than to Sprague-Dawley rats (759). Mice show strain differences in susceptibility to the toxic effects of chloramphenicol (761). Qualitative and quantitative strain differences were observed in the hematologic response to chloramphenicol succinate (500 to 2,500 mg/kg administered orally for 7 days). When administered to several inbred mice strains (C3H/He, CBA/Ca, BALB/c, and C57BL/6) and one outbred strain (CD-1), the inbred strains were more susceptible to the toxic effects and produced more variable results. Only the inbred strain developed leukopenia, despite the fact that all strains displayed reticulocytopenia and anemia. The toxicity of chloramphenicol is apparently due to minor metabolites, which may vary due to subtle differences in the metabolism of chloramphenicol, and the levels of antioxidants (e.g., glutathione and vitamin E) may account in part for the differences in tolerability (762).

In a broad sense, animals tolerate drugs better than humans, and in most efficacy studies, tolerability issues are not revealed; often relevant effects are simply not observed by the experimenter or the study plan does not facilitate their being observed. This may partly be due to inherent physiologic properties of the animal that result in faster elimination of the compound (or deleterious metabolites) from the body, generally short treatment courses because of the fulminant nature of the infection, and the masking of tolerability problems by the symptoms of the infection. Additionally, the psychobehavioral nature of the animal may mask the toxicity of the antibiotic. Anatomic and physiologic differences—for example, differences in the gastrointestinal tract resulting in different degrees of exposure, rates of absorption and total fraction of drug absorbed, and drug transit time—may affect tolerability as well as efficacy (763). Antibiotic treatment of rats will change the gastrointestinal bacterial flora and/or composition of the feces but may have little effect on transit time (764). Due to their coprophilic nature, antibiotic treatment of pregnant rats and the resultant changes in intestinal microflora lead to abnormal gastrointestinal flora in suckling rats along with the establishment of potential pathogens in the gastrointestinal tract (765); the same study also found that antibiotic treatment had less impact on skin and vaginal microflora than on gastrointestinal tract microflora (765).

The nephrotoxicity of antiinfectives in animals is often the result of proximal tubule damage, which is related to its relative size and the drug concentration-time profile, the latter being the sum of the effects of drug movement across the lumen and the contraluminal secretory transport processes (766). These processes can be profoundly different in different species. Enhanced renal excretion by rats with streptozotocin-induced diabetes reduces cephaloridine renal toxicity due to higher renal excretion rates (producing lower kidney cephaloridine levels) than normal rats (767); serum PK profiles were similar.

Macrolides possess potential arrhythmogenic properties. The PKS/PDS of the prolongation of the Q-T interval by clarithromycin, roxithromycin, erythromycin, and azithromycin has been determined from electrocardiograms (ECGs) in rats (768). Compounds were administered by infusion in order to override the different PK properties of these compounds and produce controlled serum levels. The data for clarithromycin and erythromycin fit an E_maxmodel, whereas the data for roxithromycin and azithromycin fit better a linear model. The order of potency of the compounds was erythromycin > clarithromycin > azithromycin > roxithromycin. Futher, the Q-T–prolonging activity of erythromycin and clarithromycin occurred at serum concentrations required for antibacterial effect, whereas such activity occurred only at supraantibacterial levels for azithromycin and roxithromycin.

Norfloxacin was also tested for CNS effects in rats. The PK/PD profile was monitored using an electroencephalogram (EEG) in order to arrive at a PK/PD relationship for norfloxacin (769). Rats received 5 mg/kg norfloxacin over 30 minutes, and the EEG demonstrated a pattern suggestive of eliptogenic potential; in addition, PK/PD modeling indicated that longer diffusion times have a greater eliptogenic potential (769).

Many factors contribute to the tolerability of animals to antimicrobial agents (759). Age, which affects metabolic enzymes and renal function, is another factor. Given that most laboratory animals are nocturnal, time-related functions, such as metabolic rates, food and water intake, and sleep time/activity, may also alter the tolerability (and effectiveness) of antibiotics; for example, exercise can alter the PK profile of mice. The single most important mechanism of antibiotic toxicity in small animals is disruption of the normal bacterial flora. In guinea pigs and hamsters, antibiotic toxicity leading to death is very often due to overgrowth of C. difficile and consequent toxin release (759), although L. monocytogenes, Clostridium perfringens, and Clostridium spiroforme have all been implicated as mediators of toxicity. Induction of overgrowth appears to be most prominent with some macrolides, notably clindamycin and lincomycin. Allergic reactions sensu stricto do not readily occur in lab animals, and the typical “allergic response” of guinea pigs to penicillin is likely due to enteric overgrowth. Newborn and germ-free guinea pigs are not susceptible to enterocolitis, suggesting the lack of a true allergic response (759). This may also be the case with many of the antibiotics given to laboratory animals.

Commonly, tolerability to compounds is reported without data presentation. Occasionally, some objective measure of tolerability is included. For example, Van Etten et al. (770) reported the maximally tolerated dose of amphotericin B in different formulations and determined renal and liver toxicity based on blood chemistry measurements. Additionally, in rats with subcutaneous staphylococcal abscesses, daptomycin was superior to vancomycin in treating both MSSA and MRSA. No detectable renal damage was elicited by daptomycin. The combination of daptomycin and tobramycin produced less renal injury (as determined by function tests and histology) than tobramycin alone (771). Rarely has observation of the extent of well-being of the experimental animal (i.e., the sum of illness due to infection, toxicity due to the drug substance, and relief from symptoms of infection due to antimicrobial effect) been included in the study design. Radiotelemetry was used to monitor a panel of physiologic indices to devise a “sickness behavior” index, which was then applied by Bauhofer et al. (772,773) to monitor the response to antibiotics with or without G-CSF in parallel with more classic indicators of the CLP model such as fever and mortality. Their study indicated a mild improvement from combined treatment over antibiotics alone, prompting their suggestion that the sickness behavior index could serve as the equivalent of a human quality-of-life index.

PD parameters such as the rate of bactericidal activity with increasing drug concentrations, the PAE, sub-MIC effects, postantibiotic leukocyte enhancement, and the first-exposure effect more accurately describe the time course of antimicrobial activity than the MIC and MBC (62). The PAE is the persistent inhibition of bacterial growth after a brief exposure to an antibiotic. Most easily demonstrated in the in vitro systems, this effect is observable in vivo (104,774–782), although it is complicated by compounding sub-MIC effects (e.g., the postantibiotic sub-MIC effect [PASME]; 783–785) and by the enhancement of the phagocytic activity of leukocytes that occurs during the PAE phase and is partially due to sub-MIC effects on bacteria; this effect is termed the postantibiotic leukocyte effect (PALE) (e.g., aminoglycosides and quinolones demonstrate a longer in vivo PAE in the presence of neutrophils [104,774,775]).

The PAE may be nonexistent or of varying duration depending on the compound class and target bacterium (62,774). Typically, β-lactams and cephalosporins produce a substantial PAE against streptococci, but a small PAE or none at all against gram-negative bacteria, although there are exceptions (e.g., imipenem against P. aeruginosa) (786). Nucleic acid synthesis inhibitors and protein synthesis inhibitors (notably the aminoglycosides) produce an in vivo PAE with many bacteria (e.g., streptococci and gram-negatives). An in vivo PAE occurs with staphylococci treated with many different antibiotics. Furthermore, the use of humanlike PK prolongs the duration of the PAE in vivo. Indeed, the expected duration of the PAE (in combination with sub-MIC effects or the PALE) can be incorporated into estimates of clinical treatment regimens (46,104,787). However, Fantin et al. (776) found neither the duration of the PAE in vitro nor the MIC nor bactericidal activity in vivo correlated with the duration of the in vivo PAE.

The following are critical for defining the presence and duration of a PAE: the specific microorganism–antimicrobial combination, the antimicrobial combination and the experimental conditions, including the antimicrobial concentration and the length of the antimicrobial exposure (788,789). The PAE is probably also affected by the density of bacteria, their growth rate and metabolic activity, and the extent of inflammation at the site of infection (230). Proposed mechanisms by which the PAE occurs include both nonlethal damage induced by the antimicrobial agent and a limited persistence of the antimicrobial agent at the bacterial binding site (788).

Determination of the PAE in vivo is complicated by sub-MIC effects on the bacteria, which can suppress growth and have other physiologic effects that contribute to a composite PAE in vivo (785). The PASME is the growth suppression that may occur when a low concentration of antibiotic (typically ≤0.3 × MIC) is in the presence of bacteria previously exposed to a suprainhibitory concentration. As low levels of antibiotic are likely still at the infection site at the time of the next treatment, the PASME probably reflects the in vivo situation more closely than the PAE (62,774).

Taken together, dissection of the various components of the protracted effects of drug near the end of the treatment interval, where drug levels fall to the sub-MIC level of activity and then fall still further to the “true” postantibiotic level, is complicated. Required groups include immunocompromised (leukopenic) and immunocompetent animals (to control for the PALE), PK measurements of bloodstream and infected tissue (to determine the duration of the sub-MIC levels and identify the initiation of the true PAE phase) in both groups of animals (to control for the influence of neutrophil influx on PK properties), and perhaps a group treated with an antibiotic-inactivating enzyme to further control for sub-MIC effects. Furthermore, groups treated with antibiotic producing a humanlike PK profile would be desirable. With so many groups being followed over time, the mouse thigh infection model is most frequently used for determination of the PAE.

The capability of producing a long PAE facilitates therapy using those antibiotics (e.g., aminoglycosides) to be administered infrequently, and continuous or frequent administration is required for those compounds that lack an in vivo PAE (e.g., β-lactams) (789). Extending the dosing interval of an antimicrobial agent that has a PAE has several potential advantages, among them reduced cost, less toxicity, and better compliance among outpatients receiving antimicrobial therapy (62,66,774).

It is fairly well established that severity of clinical infection, usually associated with a delay in presentation or diagnosis, is related to treatment failure, particularly in the case of neutropenic patients (791,792). Delay in beginning antimicrobial therapy, and hence, progression of infection, has been demonstrated to have profound effects on the outcome of antibiotic treatment of experimental infections; however, this issue has not yet been thoroughly studied.

Using the thigh infection model in both leukopenic and normal mice, Gerber et al. (793) found that the age of infection had a profound influence on the outcome of antibacterial therapy of P. aeruginosainfection using gentamicin, ticarcillin, and ceftazidime.

Several factors were proposed to account for these observations:

■  Alteration of the physiology of the bacteria (e.g., alteration to nongrowing or slowly growing phenotypes typically more resistant to antibiotics or production of an extracellular matrix inhibitory to antibiotics). Note, however, that in the study by Gerber et al. (793), P. aeruginosa grew at a constant rate during the entire infection, and therefore, slow growth cannot be the only contributing factor.

■  The presence of an in vivo correlate of the inoculum effect observed in vitro. In this case, bacteria within clusters or microcolonies may be protected from the effects of antibiotics or antibiotics may poorly penetrate these foci.

■  Pathophysiology at the site of the lesion, resulting in reduction of the activity of antibiotics. For example, low pH and low oxygen at the infected site are nonoptimal for aminoglycosides, and there is reduced penetration of antibiotics in large cardiac vegetations in advanced endocarditis (794).

The effect of delay in treatment of experimental pneumonia due to S. pneumoniae with temafloxacin was studied in a mouse model (795). Whereas initiation of doxycycline (1.5 mg/kg intraperitoneally once) at 4 hours postinfection with anthrax spores in mice was effective (90% survival), initiation of treatment at 24 and 48 hours had no substantial effect on mortality rates, although the onset of death was delayed to 4 days in the 24-hour treatment group and to 2 to 3 days in the 48-hour treatment group and the control group (796). As death from anthrax infection occurs due to toxin production, it is not surprising that a delay in treatment initiation past a threshold point would have a very small impact on outcome.

Noninvasive imaging encompasses a rapidly developing palette of techniques that are continuing to set new standards for clinical diagnosis and disease monitoring. However, such techniques are only slowly becoming established for use in experimental animals and in particular models of infection (797–799); a general review of imaging techniques specifically for application in small animals has been published (800). An obvious advantage of the use of noninvasive imaging in animal models is the possibility of a direct comparison of response in animals with that of human patients. As compared with human and human systems, animals require machines with increased resolution before particular techniques can be used really successfully.

Imaging of the Host’s Response to Infection

Many imaging techniques are available to noninvasively monitor disease states and their response to therapy. Difficulty in accessing the required equipment and designing of animal-specific imaging systems has perhaps hampered the use of these approaches. Access to specialized reagents may also be limiting (e.g., ¹⁸F for use in PET imaging has a short half-life). Labeled cytokines have found use in a variety of imaging techniques for monitoring different leukocyte subsets in vivo during infection (801).

Magnetic Resonance Imaging

Using magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS), histopathology and segmentation maps were obtained by the mathematical processing of three-dimensional T2-weighted MRI data via a neural network. The MRI patterns varied according to the nature and extent of infection with A. fumigatus. The MRS results show a statistically significant increase in inorganic phosphate and a significant decrease in phosphocreatine levels in the inflamed region (802). MRI along with contrast agent gadolinium-diethylene-triamine-pentaacetic acid was used to monitor the transient modulation of the blood–brain barrier following intravenous injection of bacterial glycopeptides (803).

The S. aureus thigh infection model, including treatment with vancomycin and imipenem/cilastatin, showed that MRI images closely paralleled histologic changes occurring as the infection progressed or resolved with antibiotic therapy (804). Therefore, this noninvasive procedure could be used repeatedly on an individual animal to monitor at least some aspects of antimicrobial chemotherapy. Experimental S. aureus osteomyelitis in New Zealand white rabbits was examined by MRI, computed tomography (CT), and plain film radiography (PF). MRI detected periostitis despite the absence of periosteal ossification and was more sensitive than CT or PF (805).

Positron Emisson Tomography

Positron emisson tomography (PET) can be used for PK studies as long as the compound can be synthesized to contain the short-lived isotopes used in PET (806,807). In an elegant study of the dynamics of tubercle lesions in the rabbit lungs and effect of antituberculosis treatment PET was used to illustrate the lesions as well as quantify the development during antibiotic treatment (808).

Computed Tomography

Contrast material–enhanced CT and MRI were used on rabbits with osteomyelitic S. aureus lesions, and the detection rates were similar (809). Three-phase technetium-99m methylene diphosphonate gallium-67 MRI images were obtained from New Zealand rabbits with S. aureus osteomyelitis (810). There was no significant difference between radionuclide studies and MRI images in the detection of osteomyelitis, but MRI was significantly more sensitive in the detection of soft tissue infection, including cellulitis and abscesses. Arthritic knee joints of rabbits have been monitored using CT. The use of the perfluorocarbon macrophage-labeling contrast agent perfluoroctylbromide facilitated discerning the response of rabbits aseptic or septic to tetracycline therapy (811). Microcomputed tomography has been used to study bone formation within and adjacent to the defect around implanted polyacetyl plates and Kirchner wires in rats infected with S. aureus (812).

Echocardiography

Serial transthoracic echocardiography, during and after treatment of experimental S. aureus endocarditis in rabbits, was used as a physiologic indicator of the relative benefits of different antibiotic regimens (813). It has also been used with murine CLP models (711).

As for emerging technologies, fluorescence-mediated molecular tomography (FMT) can three-dimensionally image gene expression by resolving fluorescence activation in deep tissues (814).

Detection of Bacterial Activity by Use of Bacterial Labeling with lux Genes

General reviews of the use of cellular labeling by luciferase expression have been published (798,815), and a review of the use of bioluminescence techniques to study gene expression is available (816).

Use of bioluminescence to detect bacteria in vivo provides a noninvasive way of monitoring the progress of infection and the response to antiinfective therapies (for reviews, see 797–799,817). First described by Contag et al. (818), construction of virulent bacteria bearing lux operons constitutes a sensitive marker of bacterial viability, as the bioluminescence reaction is adenosine triphosphate (ATP)–dependent. The article by Contag et al. (798) describes the construction and use of this technique in detail.

A thorough description of this approach has been presented by Rocchetta et al. (819), who used lux-transformed E. coli and a neutropenic mouse thigh infection model. The incorporation of an intensified charge-coupled device (ICCD) camera system facilitated the determination of light emission and therefore the sequential evaluation of individual mice. Dose-dependent bacterial CFU and light emission values were obtained in vitro and in vivo for controls and the bacteria exposed to ceftazidime, tetracycline, or ciprofloxacin. The detection methods were found to be highly correlated. Similar results were obtained with murine lung infections due to S. pneumoniae (820).

Given the extreme differences in the nature of bacterial growth as biofilms or planktonic cultures, bioluminescence may help in understanding the chemotherapy of device-related infections. lux-Transposed S. aureus and P. aeruginosa were used to establish Teflon catheter biofilm infection model mice, with either precolonized or postimplant-infected catheters being used (134). The effectiveness of various antibiotics, the determination of the in vivo PAE, and the monitoring of the emergence of antibiotic resistance in this model have also been discussed (135). Seven days after subcutaneous implantation of catheters precolonized with S. aureus, treatment with rifampin, tobramycin, or ciprofloxacin was initiated. Tobramycin and ciprofloxacin were poorly active, but rifampin led to an initial decline of bacterial CFU, after which resistance developed as expected.

Bioluminiscense imaging has been evaluated for in vivo illustration as well as quantification of antibiotic effect in S. aureus–infected skin wounds in mice (821).

Detection of Bacterial Activity by Use of Bacterial Labeling with Green Fluorescent Protein

Green fluorescent protein, a 31-kDa gene product of the jellyfish Aequorea victoria, has been extensively used to label mammalian and microbial cells. It is capable of “nonspecifically” labeling bacteria (822,823) and of being inserted to act as a reporter gene system (e.g., in organisms [824], Salmonella and Pseudomonas spp and an Alcaligenes sp [825], a Yersinia sp [826], Legionella pneumophila [824], Streptococcus pneumoniae [827], Yersina pseudotuberculosis [828], C. neoformans [829], and a Helicobacter sp [830]). Care must be taken in the construction of such cell lines in order to optimize the system for in vivo activity (831). This technology is advancing, and color variants are being developed (simulation of gene expression [832], profiling of a genetically modified gfp [831], and profiling of a genetically modified gfp [828]). Single-copy gene insertion has been noted for P. aeruginosa (824), salmonellae (268), and Streptococcus pneumoniae (1). Improved data processing (833) and the ability to see single cells have contributed to enhancing this technology. The GFP expression levels of 7,000 to 200,000 copies per cell have been viewed as nondeleterious to the virulence of Salmonella typhimurium (834).

Genomic approaches utilizing animal models of infection fall into two distinct categories: (a) bacterial genomics, whose goal is to identify bacterial genes involved in the initiation, survival, and propagation of bacteria during pathogenic infection and to uncover the genetic basis of sensitivity (or resistance) to antibiotics; and (b) pharmacogenomics, which in the specialized case of infection, concerns itself with understanding the genetics of differences in antibiotic metabolism and disposition and their toxic effects without considering the effects of the drugs on the host physiology as the drugs are targeted to the microorganism. A good overview of the study of microorganism and host genomics using DNA microarrays has been provided by Bryant et al. (835). They outline the use of DNA microarrays in studying the infectious process from the aspect of the pathogen (diagnostics, epidemiology, pathogenicity, and antimicrobial resistance), the host (innate immunity, adoptive or learned immunity, physiological differences related to susceptibility, and drug metabolism and toxicity), and the host–microbe interaction (“normal” and abnormal responses to infectious insult, pathogenic processes such as microbe-induced host cell apoptosis, responses to antibiotic treatment such as inflammatory responses to antibiotic-lysed bacteria, and vaccination).

Pharmacogenomics of Pathogens

Clearly, genomic investigation of the microorganisms’ genotype/phenotype during infection will uncover novel genes, perhaps expressed only in vivo, that represent potential novel targets for antimicrobial therapy or the means of antimicrobial resistance (836–839). However, despite the sequencing of entire genomes of several bacteria, exploitation of this new knowledge especially in in vivo models has been slow. Despite this, there is great potential for genomics to “customize” clinical treatment of infection, providing information on both the pathogen and the host to aid in drug selection (838). De Backer and Van Dijck (840) have provided a review of the role pharmacogenomics might play in the development of novel antifungal chemotherapeutics.

Pharmacogenomics of the Infected Host

Pharmacogenomics focuses on the host, evaluating genetic-based differences in host drug transport and metabolism, both of which form the basis of PK and tolerability features of any antibiotic (841). For example, polymorphisms occur in P-glycoprotein, an ATP-binding cassette transporter that pumps drugs out of cells and thus affects drug uptake from the gastrointestinal tract as well as tissue distribution. Polymorphisms also occur in drug-metabolizing enzymes that are components of both the “phase I” metabolizing group (e.g., the cytochrome P450 family, which metabolize drugs through oxidation, reduction, or hydrolysis steps) and members of the “phase II” family (e.g., UDP glucuronosyltransferases, glutathione transferases, methyltransferase, and acetyl transferases) and that by conjugating the drug help to facilitate its elimination from the body. In addition to genetic predisposition to infection, inbred mouse strains also carry alterations in drug metabolism, such as the differences in tolerability of inbred strains to chloramphenicol toxicity (see reference 841). Genetic mapping of inbred strains to identify loci and eventually genes that are involved in drug metabolism and display polymorphisms will likely lead to the identification of human homologues. In addition, use of genetically modified mice will aid in identifying drug-metabolizing enzymes (842), and such information can be applied to medicinal chemistry programs to design antibiotics with better PK and metabolic/tolerability profiles. In this regard, the apparent failure or poor activity of antibiotics in some animal models may be due to differences in drug metabolism between different mouse strains or different species.

Perhaps the ultimate question that arises following the preclinical evaluation of an antiinfectives in an animal model of infectious disease is, what is the true utility of the data obtained from the animal model, namely, how predictive is the model for the clinical situation? The answer to this lies in understanding both the model and the clinical situation. In understanding the model, it is as important to know the limitations of the model as much as its advantages and to ask only those questions that the particular model can answer. Few studies have systematically reviewed the correlation of the effect of antibiotics in animal models and clinical outcome. One such a review (612) suggests that the rat osteomyelitis model does have a good measure of predictability. Although discriminative models of infection are designed to mimic the clinical situation more closely than screening models, all models can provide useful information. This is probably due to the mode of action of antibiotics, for they primarily target pathogens of another hierarchical kingdom (prokaryotes). Given this likelihood of high specificity of drug-target interaction, the infecting organism bearing the target remains “constant” irrespective of host, be it an experimental one (mouse, rat, etc.) or one that acquires the infection naturally (humans among other animals) (63). Therefore, the issue of predictability becomes an issue of PK/PD relationship: can the drug attain a sufficient PK profile (and be well tolerated) in the patient and satisfy the PD criteria established during the experimental evaluation (e.g., plasma concentration above the MIC for 80% of the administration interval). Allometry or physiologically based PK/PD modeling can be used to predict dosages for the target patient. But this serves only as a guide to aid decisions; it is on the basis of animal efficacy model data that clinical trials are most often initiated. Such a modeling exercise requires input from many different models to provide a database sufficient for making an educated prediction concerning clinical success.

However, in spite of careful experimentation, clinical failure of a compound effective in preclinical testing will sometimes occur. In such cases, the researchers will ask, “What went wrong?”

If only to provide a basis for discussion, consideration should be given to an excellent publication by Eichaker and associates (23). In this study, the authors review clinical data (from 22 trials) and preclinical data (from 95 “trials”) concerning the testing of antiinflammatory agents in bacterial sepsis. They argue that the vast amount of preclinical data, do not reflect clinical experience. Using meta-regression techniques to pool the data, they test the hypothesis that a strong relationship exists between risk of death from disease and the effectiveness of antiinflammatory agents in treating disease. If preclinical studies show an exceedingly high risk of death and the clinical trials a more moderate one, then the preclinical data would be misleading; for the findings to be truly comparable, the disease outcomes should be similar in both the preclinical and the clinical trials. The authors show that the mortality rates of control animals (88% [79% to 96%]; median [25th to 75th quartiles]) were highly different from those of the clinical trials (39% [32% to 43%], p = 0.0001). Regression analysis of the animal data indicated that, of the factors evaluated, 70% of the variability of the effect of antiinflammatory agents could be attributed to the risk of death. Normally, the clinical trials did not categorize the patients regarding risk of death, but in two trials that did, the greatest benefit of the antiinflammatory agents was to those at the highest risk of death, similar to the preclinical studies. Plotting the control odds ratio of dying and the odds ratio of the treated group dying for both animal and human studies shows a clustering of the human trial data at a lower risk than in the majority of the animal trials. The authors then modified their animal model to have a range of severity of disease when testing antiinflammatory agents in sepsis and produced data yielding an odds ratio plot very similar to that obtained from the clinical trials. In a reiterative process, the animal model was modified to produce data that better fit the spectrum of disease observed clinically, thus obtaining an animal model of greater predictive quality.

In some cases, definitive clinical trials are difficult, if not impossible, to perform. In such cases, reliance on animal model data is extensive, if not exclusive. For example, endocarditis is a severe, costly infection, and therefore identification of situations invoking prophylaxis against this infection is clearly warranted. Prophylaxis requires the ability to identify clinical procedures that might result in bacteremia, identify the patients at risk, and determine the optimal prophylactic antibiotic regimen, one that maximizes effectiveness and minimizes the risk of side effects. Owing to the difficulty of performing appropriate clinical studies, Moreillon (844) summarized the information obtained from animal studies of endocarditis and used it as a basis for a proposal regarding guidelines for clinical prophylaxis of endocarditis.

In the comprehensive review of animal models used for the study of antibiotic treatment of tuberculosis, Franzblau et al. (651) discuss the predictability of the models for human disease in detail. Although mouse models have been fairly useful in predicting multidrug treatment effect for humans, the lack of standardization of the models and effect parameters may be seen as an obstacle to such evaluation. In the quest for the future collaborative efforts in standardizing animal tuberculosis models researchers should take a cautionary note from the advice on animal welfare issues connected with murine tuberculosis models (17).

In summary, there is a vast amount of information on animal models of infection, but it is likely that only when the models are used wisely and modified to better reflect the clinical situation can predictions based on them be relied upon.

The authors would like to acknowledge the contributions of Terry O’Reilly, PhD, made to the 5th edition of this book.

  1.  Beam TR, Gilbert DN, Kunin CM. General guidelines for the clinical evaluation of anti-infective drug products. Clin Infect Dis 1992;15(Suppl 1):S5–S32.

  2.  O’Reilly T, Andes DA, Østergaard C, et al. Evaluation of antimicrobials in experimental animal infections. In: Lorian V, ed. Antibiotics in laboratory medicine. 5th ed. New York: Lippincott Williams & Wilkins, 2005:654–718.

  3.  Zak O, Sande MA, eds. Experimental models in antimicrobial chemotherapy. London: Academic Press, 1986.

  4.  Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999.

  5.  Zak O, O’Reilly T. Animal infection models and ethics: the perfect infection model. J Antimicrob Chemother 1993;31(Suppl D):193–205.

  6.  Morton DB. Ethical aspects of the use of animal models of infection. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:29–48.

  7.  Cohen C. The case for use of animals in biomedical research. N Engl J Med 1986;315:865–870.

  8.  Obrink KJ. Animal models and ethics. In: Keusch G, Wadstrom T, eds. Experimental bacterial and parasitic infections. New York: Elsevier Biomedical, 1982:3–10.

  9.  Regan T. The case for animal rights. In: Singer P, ed. Defense of animals. New York: Harper & Row, 1985:13–26.

 10.  Regan T. The case for animal rights. London: Routledge, 1988.

 11.  Singer P. Animal liberation. New York: Avon Books, 1990.

 12.  Bateson P. Do animals feel pain? New Sci 1992;134(1818):30–33.

 13.  Flecknall P, Waterman-Pearson A. Pain management in animals. London: WB Saunders, 2000.

 14.  European Union. Directive 2010/63/EU of the European parliament and of the council of 22 September 2010 on the protection of animlas used for scientific purposes. Official Journal of the European Union. September 22, 2010:33–79.

 15.  Cowan T. The Animal Welfare Act: background and selected legislation. http://nationalaglawcenter.org/wp-content/uploads/assets/crs/RS22493.pdf. Congressional Research Service RS22493. Published June 2013.

 16.  Kilkenny C, Browne WJ, Cuthill IC, et al. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLos Biol 2010;8:e1000412.

 17.  Franco NH, Correia-Neves M, Olsson AS. Animal welfare in studies of murine tuberculosis: assessing progress over a 12-year period and the need for further improvement. PLos One 2012;7:e47723.

 18.  Hooijmanns CR, Leenaars M, Ritskes-Hoitinga M. A gold standard publicatin checklist to improve the quality of animal studies, to fully integrate the three Rs, and to make systematic reviews more feasible. Altern Lab Anim 2010;38:167–182.

 19.  van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993.

 20.  Hansen AK. The impact of general laboratory animal health on experimental models in antimicrobial chemotherapy. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:49–59.

 21.  Torres-Molina F, Peris-Ribera JE, Garcia-Carbonell MC, et al. Nonlinearities in amoxicillin pharmacokinetics. II. Absorption studies in the rat. Biopharm Drug Dispos 1992;13:39–53.

 22.  Harter DH, Petersdorf RG. A consideration of the pathogenesis of bacterial meningitis: review of experimental and clinical studies. Yale J Biol Med 1960;32:280–309.

 23.  Eichacker PQ, Parent C, Kalil A, et al. Risk and the efficacy of antiinflammatory agents retrospective and confirmatory studies of sepsis. Am J Respir Crit Care Med 2002;166:1197–1205.

 24.  Beynen AC, Festing MFW. Phases in an animal experiment. In: van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993:197–208.

 25.  Beynen AC, Festing MFW, van Montfort MAJ. Design of animal experiments. In: van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993:209–240.

 26.  Hanfelt JJ. Statistical approaches to experimental design and data analysis of in vivo studies. Breast Cancer Res Treat 1997;46:279–302.

 27.  Hermans PGC, Fosse RT, van der Gulden WJI, et al. Organisation and management of animal experiments. In: van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993:241–254.

 28.  Costerton JW, Cheng K-J, Geesey GG, et al. Bacterial biofilms in nature and disease. Annu Rev Microbiol 1987;41:435–464.

 29.  Costerton JW, Veeh R, Shirtliff M, et al. The application of biofilm science to the study and control of chronic bacterial infections. J Clin Invest 2003;112:1466–1477.

 30.  Foley I, Brown MRW. Activity of antibiotics against adherent/slow-growing bacteria reflecting the situation in vivo. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:117–123.

 31.  Schwank S, Rajacic Z, Zimmerli W, et al. Impact of bacterial biofilm formation on in vitro and in vivo activities of antibiotics. Antimicrob Agents Chemother 1998;42:895–898.

 32.  Domingue G, Ellis B, Dasgupta M, et al. Testing antimicrobial susceptibilities of adherent bacteria by a method that incorporates guidelines of the National Committee for Clinical Laboratory Standards. J Clin Microbiol 1994;32:2564–2568.

 33.  Anwar H, Strap JL, Chen K, et al. Dynamic interactions of biofilms of mucoid Pseudomonas aeruginosa with tobramycin and piperacillin. Antimicrob Agents Chemother 1992;36:1208–1214.

 34.  Brown MRW, Collier PJ, Gilbert P. Influence of growth rate on susceptibility to antimicrobial agents: modification of the cell envelope and batch and continuous culture studies. Antimicrob Agents Chemother 1990;34:1623–1628.

 35.  Dalhoff A, Matuat S, Ullmann U. Effect of quinolones against slowly growing bacteria. Chemotherapy 1995;41:92–99.

 36.  Duguid IG, Evans E, Brown MRW, et al. Growth-rate–independent killing by ciprofloxacin of biofilm-derived Staphylococcus epidermidis: evidence for cell-cycle dependency. J Antimicrob Chemother 1992;30:791–802.

 37.  Eng RHK, Hsieh A, Smith SM. Antibiotic killing of bacteria: comparison of bacteria on surfaces and in liquid, growing and nongrowing. Chemotherapy 1995;41:113–120.

 38.  Vergeres P, Blaser J. Amikacin, ceftazidime, and flucloxacillin against suspended and adherent Pseudomonas aeruginosa and Staphylococcus epidermidis in an in vitro model of infection. J Infect Dis 1992;165:281–289.

 39.  Cozens RM, Tuomanen E, Tosch W, et al. Evaluation of the bactericidal activity of β-lactam antibiotics on slowly growing bacteria cultured in the chemostat. Antimicrob Agents Chemother 1986;29:797–802.

 40.  Davey PG, Renneberg J, Speller DCE, eds. Bacterial infection models in antimicrobial chemotherapy. J Antimicrobial Chemother 1993(Suppl D);31:1–205.

 41.  Blaser J, Zinner SH. In vitro models for the study of antibiotic activities. Prog Drug Res 1987;31:349–381.

 42.  Dudley MN, Blaser J, Gilbert D, et al. Combination therapy with ciprofloxacin plus azlocillin against Pseudomonas aeruginosa: effect of simultaneous versus staggered administration in an in vitro model of infection. J Infect Dis 1991;164:499–506.

 43.  Cozens RM. Formulation of compounds and determination of pharmacokinetic parameters. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:83–92.

 44.  Martinez MN, Pedersoli WM, Ravis WR, et al. Feasibility of interspecies extrapolation in determining the bioequivalence of animal products intended for intramuscular administration. J Vet Pharmacol Ther 2001;24:125–135.

 45.  Perkins RJ, Liu W, Drusano G, et al. Pharmacokinetics of ofloxacin in serum and vitreous humor of albino and pigmented rabbits. Antimicrob Agents Chemother 1995;39:1493–1498.

 46.  Bhardwaj R, Dorr RT, Blanchard J. Approaches to reducing toxicity of parenteral anticancer drug formulations using cyclodextrins. PDA J Pharm Sci Technol 2000;54:233–239.

 47.  Sugawara M, Toda T, Iseki K, et al. Transport characteristics of cephalosporin antibiotics across intestinal brush-border membrane in man, rat and rabbit. J Pharm Pharmacol 1992;44:968–972.

 48.  Torres-Molina F, Peris-Ribera JE, Garcia-Carbonell MC, et al. Nonlinearities in amoxicillin pharmacokinetics. I. Disposition studies in the rat. Biopharm Drug Dispos 1992;13:23–38.

 49.  Torres-Molina F, Peris-Ribera JE, Garcia-Carbonell MC, et al. Nonlinearities in amoxicillin pharmacokinetics. II. Absorption studies in the rat. Biopharm Drug Dispos 1992;13:39–53.

 50.  Chesa-Jimenez J, Peris JE, Torres-Molina F, et al. Low bioavailablity of amoxicillin in rats as a consequence of presystematic degradation in the intestine. Antimicrob Agents Chemother 1994;38:842–847.

 51.  O’Reilly T, Cleeland R, Squires E. Evaluation of antimicrobials in experimental animal infections. In: Lorian V, ed. Antibiotics in laboratory medicine. 4th ed. Baltimore: Williams & Wilkins, 1996:599–759.

 52.  Baumans V, ten Berg RGM, Beterns APMG, et al. Experimental procedures. In: van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993:299–318.

 53.  Van Dongen JJ, Remie R, Rensema JW, et al. Manual of microsurgery on the laboratory rat. Part 1. General information and experimental techniques. Amsterdam: Elsevier, 1990.

 54.  Beterns APMG, Booij LHDJ, Flecknell PA, et al. Anaesthesia, analgesia and euthanasia. In: van Zutphen LFM, Baumans V, Beyen AC, eds. Principles of laboratory animal science. Amsterdam: Elsevier, 1993:267–298.

 56.  Ambrose PG, Bhavnani SM, Rubino CM, et al. Pharmacokinetics-pharmacodynamics of antimicrobial therapy: it’s not just for mice anymore. Clin Infect Dis 2007;44:79–86.

 57.  Andes D. Clinical utility of antifungal pharmacokinetics and pharmacodynamics. Curr Opin Infect Dis 2004;17:533–540.

 58.  Craig WA. Pharmacokinetic/pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin Infect Dis 1998;26:1–10;quiz 11–12.

 59.  Drusano GL. Pharmacokinetics and pharmacodynamics of antimicrobials. Clin Infect Dis 2007;45(Suppl 1):S89–S95.

 60.  Hinojosa E, Boyd AR, Orihuela CJ. Age-associated inflammation and toll-like receptor dysfunction prime the lungs for pneumococcal pneumonia. J Infect Dis 2009;200:546–554.

 61.  Andes D. Pharmacokinetics and pharmacodynamics of antifungals. Infect Dis Clin North Am 2006;20:679–697.

 62.  Craig W. Pharmacodynamics of antimicrobial agents as a basis for determining dosage regimens. Europ J Clin Microbiol Infect Dis 1993;12(Suppl 1):S6–S8.

 63.  Drusano GL. Antimicrobial pharmacodynamics: critical interactions of “bug and drug.” Nat Rev Microbiol 2004;2:289–300.

 64.  Frimodt-Møller N. How predictive is PK/PD for antibacterial agents? Int J Antimicrob Agents 2002;19:333–339.

 65.  Hope W, Drusano GL. Antifungal pharmacokinetics and pharmacodynamics: bridging from the bench to bedside. Clin Microbiol Infect 2009;15:602–612.

 66.  Nicolau DP. Predicting antibacterial response from pharmacodynamic and pharmacokinetic profiles. Infection 2001;29(Suppl 2):11–15.

 67.  Schentag JJ, Gilliland KK, Paladino JA. What have we learned from pharmacokinetic and pharmacodynamic theories? Clin Infect Dis 2001;32(Suppl 1):S39–S46.

 68.  Barza M. 1993. Pharmacokinetics of antibiotics in shallow and deep compartments. J Antimicrob Chemother 1993;31(Suppl D):17–27.

 69.  Bergan T. Pharmacokinetic properties of the cephalosporins. Drugs 1987;34(Suppl 2):89–104.

 70.  Carbon C. Significance of tissue levels for prediction of antibiotic efficacy and determination of dosage. Eur J Clin Microbiol Infect Dis 1990;9:510–516.

 71.  Mouton JW, Theuretzbacher U, Craig WA, et al. Tissue concentrations: do we ever learn? J Antimicrob Chemother 2008;61:235–237.

 72.  Redington J, Ebert SC, Craig WA. Role of antimicrobial pharmacokinetics and pharmacodynamics in surgical prophylaxis. Rev Infect Dis 1991; 13(Suppl 10):S790–S799.

 73.  Brunner M, Derendorf H, Muller M. Microdialysis for in vivo pharmacokinetic/pharmacodynamic characterization of anti-infective drugs. Curr Opin Pharmacol 2005;5:495–499.

 74.  Liu P, Muller M, Derendorf H. Rational dosing of antibiotics: the use of plasma concentrations versus tissue concentrations. Int J Antimicrob Agents 2002;19:285–290.

 75.  Clark B, Smith D. An introduction to pharmacokinetics. 2nd ed. Oxford: Blackwell Scientific Publications, 1986.

 76.  Rowland M, Tozer T. Clinical pharmacokinetics: concepts and applications. 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 1995.

 77.  Craig WA, Ebert SC. Protein binding and its significance in antibacterial therapy. Infect Dis Clin North Am 1989;3:407–414.

 78.  Zeitlinger MA, Derendorf H, Mouton JW, et al. Protein binding: do we ever learn? Antimicrob Agents Chemother 2011;55:3067–3074.

 79.  Zini R, Riant P, Barre J, et al. Disease-induced variations in plasma protein levels. Implications for drug dosage regimens (Part I). Clin Pharmacokinet 1990;19:147–159.

 80.  Zini R, Riant P, Barre J, et al. Disease-induced variations in plasma protein levels. Implications for drug dosage regimens (Part II). Clin Pharmacokinet 1990;19:218–229.

 81.  Andes D, Craig WA. Animal model pharmacokinetics and pharmacodynamics: a critical review. Int J Antimicrob Agents 2002;19:261–268.

 82.  Siefert HM, Domdey-Bette A, Henninger K, et al. Pharmacokinetics of the 8-methoxyquinolone, moxifloxacin: a comparison in humans and other mammalian species. J Antimicrob Chemother 1999;43(Suppl B):69–76.

 83.  Bruzzese T, Rimaroli C, Bonabello A, et al. Pharmacokinetics and tissue distribution of rifametane, a new 3-azinomethyl-rifamycin derivative, in several animal species. Arzneimittelforschung 2000;50:60–71.

 84.  MacCallum DM, Odds FC. Influence of grapefruit juice on itraconazole plasma levels in mice and guinea pigs. J Antimicrob Chemother 2002;50:219–224.

 85.  Andes DR, Craig WA. Pharmacokinetics and pharmacodynamics of antibiotics in meningitis. Infect Dis Clin North Am 1999;13:595–618.

 86.  Krontz DP, Strausbaugh LJ. Effect of meningitis and probenecid on the penetration of vancomycin into cerebrospinal fluid in rabbits. Antimicrob Agents Chemother 1980;18:882–886.

 87.  Scheld WM. Drug delivery to the central nervous system: general principles and relevance to therapy for infections of the central nervous system. Rev Infect Dis 1989;11(Suppl 7):S1669–S1690.

 88.  Levy RM, Gutin PH, Baskin DS, et al. Vancomycin penetration of a brain abscess: case report and review of the literature. Neurosurgery 1986;18:632–636.

 89.  Ganzinger U, Haslberger A. Pharmacokinetics of cephalosporins in normal and septicemic rabbits. Antimicrob Agents Chemother 1985;28:473–477.

 90.  Mimoz O, Jacolot A, Padoin C, et al. Influence of experimental rat model of multiple organ dysfunction on cefepime and amikacin pharmacokinetics. Antimicrob Agents Chemother 1996;40:819–821.

 91.  Ngeleka M, Auclair P, Tardif D, et al. Intrarenal distribution of vancomycin in endotoxemic rats. Antimicrob Agents Chemother 1989;33:1575–1579.

 92.  Mackowiak PA. Influence of fever on pharmacokinetics. Rev Infect Dis 1989;11:804–807.

 93.  Ahmad M, Raza H, Murtaza G, et al. Pharmacokinetic variations of ofloxacin in normal and febrile rabbits. Pak Vet J 2008;28:181–185.

 94.  Goudah A, Mounier SM, Shim JH, et al. Influence of endotoxin induced fever on the pharmacokinetics of intramuscularly administered cefepime in rabbits. J Vet Sci 2006;7:151–155.

 95.  Marier JF, Beaudry F, Ducharme MP, et al. A pharmacokinetic study of amoxycillin in febrile beagle dogs following repeated administrations of endotoxin. J Vet Pharmacol Ther 2001;24:379–383.

 96.  Pennington JE, Dale DC, Reynolds HY, et al. Gentamicin sulfate pharmacokinetics: lower levels of gentamicin in blood during fever. J Infect Dis 1975;132:270–275.

 97.  Sarwari AR, Mackowiak PA. The pharmacologic consequences of fever. Infect Dis Clin North Am 1996;10:21–32.

 98.  Morita E, Mizuno N, Nishikata M, et al. Comparison of the pharmacokinetics of five beta-lactam antibiotics between neonatal and adult rats. Dev Pharmacol Ther 1990;14:223–230.

 99.  Tanira MOM, Ali BH, Bashir AK. Effect of endotoxin on gentamicin pharmacokinetics in old and young adult rats. Life Sci 1997;60:413–424.

100.  Thadepalli H, Reddy U, Chuah SK, et al. Evaluation of trovafloxacin in the treatment of Klebsiella pneumoniae lung infection in tumour-bearing mice. J Antimicrob Chemother 2005;45:69–75.

101.  Maiza A, Daley-Yates PT. Variability in the renal clearance of cephalexin in experimental renal failure. J Pharmacokinet Biopharm 1993;21:19–30.

102.  Giacomini KM, Roberts SM, Levy G. Evaluation of methods for producing renal dysfunction in rats. J Pharm Sci 1981;70:117–121.

103.  Andes D, Craig WA. In vivo activities of amoxicillin and amoxicillin-clavulanate against Streptococcus pneumoniae: application to breakpoint determinations. Antimicrob Agents Chemother 1998;42:2375–2379.

104.  Craig WA, Redington J, Ebert SC. Pharmacodynamics of amikacin in vitro and in mouse thigh and lung infections. J Antimicrob Chemother 1991;27(Suppl C):29–40.

105.  Nierenberg DW. Drug inhibition of penicillin tubular secretion: concordance between in vitro and clinical findings. J Pharmacol Exp Ther 1987;240:712–716.

106.  Thonus A, de Lange-Macdaniel AV, Otte CJ, et al. Tissue cage infusion: a technique for the achievement of prolonged steady state in experimental animals. J Pharmacol Methods 1979;2:63–69.

107.  Astry Cl, Nelson S, Karam GH, et al. Interactions of clindamycin with antibacterial defenses of the lung. Am Rev Respir Dis 1987;135:1015–1019.

108.  Naziri W, Cheadle WG, Trachtenberg LS, et al. Second place winner of the Conrad Jobst Award in the gold medal paper competition. Increased antibiotic effectiveness in a model of surgical infection through continuous infusion. Am Surg 1995;61:11–15.

109.  Roosendaal R, Bakker-Woudenberg IAJM, van den Berghe-van Raffe M, et al. Impact of the dosage schedule on the efficacy of ceftazidime, gentamicin and ciprofloxacin in Klebsiella pneumoniae pneumonia and septicemia in leukopenic rats. Eur J Clin Microbiol Infect Dis1989;8:878–887.

110.  Thauvin C, Eliopoulos GM, Willey S, et al. Continuous-infusion ampicillin therapy of enterococcal endocarditis in rats. Antimicrob Agents Chemother 1987;31:139–143.

111.  Tran Ba Huy P, Meulemans A, Wassef M, et al. Gentamicin persistence in rat endolymph and perilymph after a two-day constant infusion. Antimicrob Agents Chemother 1983;23:344–346.

112.  Robaux MA, Dube L, Caillon J, et al. In vivo efficacy of continuous infusion versus intermittent dosing of ceftazidime alone or in combination with amikacin relative to human kinetic profiles in a Pseudomonas aeruginosa rabbit endocarditis model. J Antimicrob Chemother2001;47:617–622.

113.  Mizen L. Methods for obtaining human-like pharmacokinetic patterns in experimental animals. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:93–103.

114.  Woodnutt G, Catherall EJ, Kernutt I, et al. Temocillin efficacy in experimental Klebsiella pneumoniae meningitis after infusion into rabbit plasma to simulate antibiotic concentrations in human serum. Antimicrob Agents Chemother 1988;32:1705–1709.

115.  Woodnutt G, Berry V, Mizen L. Simulation of human serum pharmacokinetics of cefazolin, piperacillin, and BRL 42715 in rats and efficacy against experimental intraperitoneal infections. Antimicrob Agents Chemother 1992;36:1427–1431.

116.  Mizen L, Woodnutt G, Kernutt I, et al. Simulation of human serum pharmacokinetics of ticarcillin-clavulanic acid and ceftazidime in rabbits, and efficacy against experimental Klebsiella pneumoniae meningitis. Antimicrob Agents Chemother 1989;33:693–699.

117.  Fluckiger U, Moreillon P, Blaser J, et al. Simulation of amoxicillin pharmacokinetics in humans for the prevention of streptococcal endocarditis in rats. Antimicrob Agents Chemother 1994;38:2846–2849.

118.  Entenza JM, Blatter M, Glauser MP, et al. Parenteral sparfloxacin compared with ceftriaxone in treatment of experimental endocarditis due to penicillin-susceptible and -resistant streptococci. Antimicrob Agents Chemother 1994;38:2638–2688.

119.  Entenza JM, Fluckiger U, Glauser MP, et al. Antibiotic treatment of experimental endocarditis due to methicillin-resistant Staphylococcus epidermidis. J Infect Dis 1994;170:100–109.

121.  Richmond JY, Quimby F. Considerations for working safely with infectious disease agents in research animals. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:69–73.

122.  Frimodt Møller N, Thomsen VF. The pneumococcus and the mouse protection test: inoculum, dosage and timing. Acta Path Microbiol Immunol Scand 1986;94:33–37.

123.  Frimodt-Møller N, Thomsen VF. Interaction between beta-lactam antibiotics and gentamicin against Streptococcus pneumoniae in vitro and in vivo. APMIS 1987;95:269–275.

124.  Knudsen JD, Odenholt I, Erlendsdottir H, et al. Selection of resistant Streptococcus pneumoniae during penicillin treatment in vitro and in three animal models. Antimicrob Agents Chemother 2003;47:2499–2506.

125.  O’Reilly T, Niven DF. Tryptone-yeast extract broth as a culture medium for Haemophilus pleuropneumoniae and Haemophilus parasuis to be used as challenge inocula. Can J Vet Res 1986;50:441–443.

126.  Brenner D, DeVoe IW, Holbein BE. Increased virulence of Neisseria meningitidis after in vitro iron-limited growth at low pH. Infect Immun 1981;33:59–66.

127.  Sharma S, Harjai K, Mittal R. Enhanced siderophore production and mouse kidney pathogenicity in Escherichia coli grown in urine. J Med Microbiol 1991;35:325–329

128.  Shepard BD, Gilmore MS. Differential expression of virulence-related genes in Enterococcus faecalis in response to biological cues in serum and urine. Infect Immun 2002;70:4344–4352.

129.  Keevil CW, Davies DB, Spillane BJ, et al. Influence of iron-limited and replete continuous culture on the physiology and virulence of Neisseria gonorrhoeae. J Gen Microbiol 1989;135:851–863.

130.  Collins HL, Kaufmann SHE, Schaible UE. Iron chelation via deferoxamine exacerbates experimental salmonellosis via inhibition of the nicotinamide adenine dinucleotide phosphate oxidase-dependent respiratory burst. J Immunol 2002;168:3456–3463.

131.  Bagge N, Ciofu O, Skovgaard LT, et al. Rapid development in vitro and in vivo of resistance to ceftazidime in biofilm-growing Pseudomonas aeruginosa due to chromosomal beta-lactamase. APMIS 2000;108:589–600.

132.  Li X, Zhao H, Lockatell CV, et al. Visualization of Proteus mirabilis within the matrix of urease-induced bladder stones during experimental urinary tract infection. Infect Immun 2002;70:389–394.

133.  Ehrlich GD, Veeh R, Wang X, et al. Mucosal biofilm formation on middle-ear mucosa in the chinchilla model of otitis media. JAMA 2002;287:1710–1715.

134.  Kadurugamuwa JL, Sin L, Albert E, et al. Direct continuous method for monitoring biofilm infection in a mouse model. Infect Immun 2003;71:882–890.

135.  Kadurugamuwa JL, Sin LV, Yu J, et al. Rapid direct method for monitoring antibiotics in a mouse model of bacterial biofilm infection. Antimicrob Agents Chemother 2003;47:3130–3137.

136.  Saint S, Chenoweth CE. Biofilms and catheter-associated urinary tract infections. Infect Dis Clin North Am 2003;17:411–432.

137.  Chmiel JF, Davis PB. State of the art: why do the lungs of patients with cystic fibrosis become infected and why can’t they clear the infection? Respir Res 2003;4:8.

138.  Ciofu O, Bagge N, Hoiby N. Antibodies against beta-lactamase can improve ceftazidime treatment of lung infection with beta-lactam-resistant Pseudomonas aeruginosa in a rat model of chronic lung Infection. APMIS 2002;110:881–891.

139.  Mongodin E, Bajolet O, Cutrona J, et al. Fibronectin-binding proteins of Staphylococcus aureus are involved in adherence to human airway epithelium. Infect Immun 2002;70:620–630.

140.  Vandecasteele SJ, Peetermans WE, Merckx R, et al. Expression of biofilm-associated genes in Staphylococcus epidermidis during in vitro and in vivo foreign body infections. J Infect Dis 2003;188:730–737.

141.  Francois P, Tu Quoc PH, Bisognano C, et al. Lack of biofilm contribution to bacterial colonisation in an experimental model of foreign body infection by Staphylococcus aureus and Staphylococcus epidermidis. FEMS Immunol Med Microbiol 2003;35:135–140.

142.  Sheppard M, Webb C, Heath F, et al. Dynamics of bacterial growth and distribution within the liver during Salmonella infection. Cell Microbiol 2003;5:593–600.

143.  Peterson LR. Animal models: the in vivo evaluation of ciprofloxacin. J Antimicrob Chemother 1986;18(Suppl D):55–64.

144.  Zak O. Scope and limitations of experimental chemotherapy. Expertmentia 1980;36:479–483.

145.  Frimodt-Møller N. Correlation of in vitro activity and pharmacokinetic parameters with effect in vivo for antibiotics: observations from experimental pneumococcus infection. Dan Med Bull 1988;35:422–437.

146.  Frimodt-Møller N. The mouse peritonitis model: present and future use. J Antimicrob Chemother 1993;31(Suppl D):55–60.

147.  Frimodt Møller N, Knudsen JD, Espersen F. The mouse peritonitis/sepsis model. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:127–136.

148.  Cross AS, Opal SM, Sadoff JC, et al. Choice of bacteria in animal models of sepsis. Infect Immun 1993;61:2741–2747.

149.  Fantin B, Legget J, Ebert S, et al. Correlation between in vitro and in vivo activity of antimicrobial agents against Gram-negative bacilli in a murine infection model. Antimicrob Agents Chemother 1991;35:1413–1422.

150.  Zak O, Sande MA. Correlation of in vitro antimicrobial activity of antibiotics with results of treatment in experimental animal models and human infection In: Sabath LD, ed. Action of antibiotics in patients. Bern, Switzerland: Hans Huber Publishers, 1982:55–67.

151.  Zimmerli W, Frei R, Widmer AF, et al. Microbiologic tests to predict treatment outcome in experimental device-related infections due to Staphyloccus aureus. J Antimicrob Chemother 1994;33:959–967.

152.  Anaissie EJ, Karyotakis NC, Hachem R, et al. Correlation between in vitro and in vivo activity of antifungal agents against Candida species. J Infect Dis 1994;170:384–389.

153.  Rex JH, Nelson PW, Paetznick VL, et al. Optimizing the correlation between results of testing in vitro and therapeutic outcome in vivo for fluconazole by testing critical isolates in a murine model of invasive candidiasis. Antimicrob Agents Chemother 1998;42:129–134.

154.  Welkos S, O’Brien A. Determination of median lethal and infectious doses in animal model systems. Methods Enzymol 1994;235:29–39.

155.  Johnson JA, Lau BH, Nutter RL, et al. Effect of L1210 leukemia on the susceptibility of mice to Candida albicans infections. Infect Immun 1978;19:146–151.

156.  Comber KR, Osborne CD, Sutherland R. Comparative effects of amoxicillin and ampicillin in the treatment of experimental mouse infections. Antimicrob Agents Chemother 1975;7:179–185.

157.  Reed LJ, Muench H. A simple method of estimating fifty percent endpoints. Am J Hyg 1938;27:4934–4997.

158.  Finney DJ. Probit analysis. Cambridge: Cambridge University Press, 1971.

159.  Hoogeterp JJ, Mattie H, Krul AM, et al. The efficacy of rifampicin against Staphylococcus aureus in vitro and in an experimental infection in normal and granulocytopenic mice. Scand J Infect Dis 1988;20:649–656.

160.  Isenberg HD, Sampson-Scherer J, Cleeland R, et al. Correlation of the results of antibiotic synergy and susceptibility testing in vitro with results in experimental mouse infections. Crit Rev Microbiol 1982;10:1–76.

161.  Beskid G, Christenson JG, Cleeland R, et al. In vivo activity of ceftriaxone (Ro 13–9904), a new broad-spectrum semisynthetic cephalosporin. Antimicrob Agents Chemother 1981;20:159–167.

162.  Norrby SR, O’Reilly T, Zak O. Efficacy of antimicrobial agent treatment in relation to treatment regimen: experimental models and clinical evaluation. J Antimicrob Chemother 1993;31(Suppl D):41–54.

163.  Brunner H, Zeiler HJ. Oral ciprofloxacin treatment for Salmonella typhimurium infection of normal and immunocompromised mice. Antimicrob Agents Chemother 1988;32:57–62.

164.  Grunberg E, Cleeland R. In vivo activity of the 6-amidino-penicillanic acid derivative, mecillinam, chemically linked or combined in varying ratios with 6-aminopenicillanic acid derivatives. J Antimicrob Chemother 1977;3(Suppl B):59–68.

165.  Mouton JW, Punt N. Use of the t>MIC to choose between different dosing regimens of β-lactam antibiotics. J Antimicrobial Chemother 2001;47:500–501.

166.  Mouton JW, van Ogtrop ML, Andes D, et al. Use of pharmacodynamic indices to predict efficacy of combination therapy in vivo. Antimicrob Agents Chemother 1999;43:2473–2478.

167.  Johansen HK, Jensen TG, Dessau RB, et al. Antagonism between penicillin and erythromycin against Streptococcus pneumoniae in vitro and in vivo. J Antimicrob Chemother 2000;46:973–980.

168.  Deshpande LM, Jones RN. Antagonism between penicillin and erythromycin against Streptococcus pneumoniae: does it exist? Diagn Microbiol Infect Dis 2003;46:223–225.

169.  Fantin B, Carbon C. In vivo antibiotic synergism: contribution of animal models. Antimicrob Agents Chemother 1992;36:907–912.

170.  Grunberg E. The effect of trimethoprim on the activity of sulfonamides and antibiotics in experimental infections. J Infect Dis 1973;128(Suppl):S478–S485.

171.  Schentag JJ, Strenkoski-Nix LC, Nix DE, et al. Pharmacodynamic interactions of antibiotics alone and in combination. Clin Infect Dis 1998;27:40–46.

172.  den Hollander JG, Mouton JW, Verbrugh HA. Use of pharmacodynamic parameters to predict efficacy of combination therapy by using fractional inhibitory concentration kinetics. Antimicrob Agents Chemother 1998;42:744–748.

173.  Nishi Y, Hasegawa MM, Ohkawa Y, et al. Mouse peritoneal lymphocytes, a new target for analyzing induction of sister chromatid exchanges on in vivo exposure to a genotoxic agent. Cancer Res 1986;46:3341–3347.

174.  Barcia-Macay M, Seral C, Mingeot-Leclerq MP, et al. Pharmacodynamic evaluation of the intracellular activities of antibiotics against Staphylococcus aureus in a model of THP-1 macrophages. Antimicrob Agents Chemother 2006;50:841–851.

175.  Sandberg A, Hessler JHR, Skov RL, et al. Intracellular activity of antibiotics against Staphylococcus aureus in a mouse peritonitis model. Antimicrob Agents Chemother 2009;53:1874–1883.

176.  Sandberg A, Jensen KS, Baudoux P, et al. Intra- and extracellular activities of dicloxacillin against Staphylococcus aureus in vivo and in vitro. Antimicrob Agents Chemother 2010;54:2391–2400.

177.  Sandberg A, Jensen KS, Baudoux P, et al. Intra- and extracellular activity of linezolid against Staphylococcus aureus in vivo and in vitro. J Antimicrob Chemother 2010;65:962–973.

178.  Brinch KS, Sandberg A, Baudoux P, et al. Plectasin shows intracellular activity against Staphylococcus aureus in human THP-1 monocytes and in a mouse peritonitis model. Antimicrob Agents Chemother 2009;53:4801–4808.

179.  Sandberg A, Lemaire S, Van Bambeke F, et al. Intra- and extracellular activities of dicloxacillin and linezolid against a clinical Staphylococcus aureus strain with a small-colony-variant phenotype in an in vitro model of THP-1 macrophages and an in vivo mouse peritonitis model. Antimicrob Agents Chemother 2011;55:1443–1452.

180.  Tang H, Chen C-C, Zhang C-C, et al. Use of carbapenems against clinical, nontyphoid Salmonella isolates: results from in vitro and in vivo animal studies. Antimicrob Agents Chemother 2012;56:2916–2922.

181.  Selbie FR, Simon RD. Virulence to mice of Staphylococcus pyogenes: its measurement and its relation to certain in vitro properties. Br J Exp Pathol 1952;33:315–326.

182.  Capilla J, Clemons KV, Stevens DA. Animal models: an important tool in mycology. Med Mycol 2007;45:657–684.

183.  Guarro J. Lessons from animal studies for the treatment of invasive human infections due to uncommon fungi. J Antimicrob Chemother 2011;66:1447–1466.

184.  Dannaoui E, Mouton JW, Meis JFGM, et al. Efficacy of antifungal therapy in a nonneutropenic murine model of zygomycosis. Antimicrob Agents Chemother 2002;46:1953–1959.

185.  Howard SJ, Lestner JM, Sharp A, et al. Pharmacokinetics and pharmacodynamics of posaconazole for invasive pulmonary aspergillosis: clinical implications for antifungal therapy. J Infect Dis 2011;203:1324–1332.

186.  Lepak AJ, Marchillo K, Vanhecker J, et al. Posaconazole pharmacodynamic target determination against wild-type and Cyp51 mutant isolates of Aspergillus fumigatus in an in vivo model of invasive pulmonary aspergillosis. Antimicrob Agents Chemother 2013;57:579–585.

187.  Lewis RE, Prince RA, Chi J, et al. Itraconazole preexposure attenuates the efficacy of subsequent amphotericin B therapy in a murine model of acute invasive pulmonary aspergillosis. Antimicrob Agents Chemother 2002;46:3208–3214.

188.  Patterson TF, George D, Ingersoll R, et al. Efficacy of SCH 39304 in treatment of experimental invasive aspergillosis. Antimicrob Agents Chemother 1991;35:1985–1988.

189.  Sheppard DC, Rieg G, Chiang LY, et al. Novel inhalational murine model of invasive pulmonary aspergillosis. Antimicrob Agents Chemother 2004;48:1908–1911.

190.  Wiederhold NP, Kontoyiannis DP, Chi J, et al. Pharmacodynamics of caspofungin in a murine model of invasive pulmonary aspergillosis: evidence of concentration-dependent activity. J Infect Dis 2004;190:1464–1471.

191.  Clemons KV, Stevens DA. Efficacy of ravuconazole in treatment of mucosal candidosis in SCID mice. Antimicrob Agents Chemother 2001;45:3433–3436.

192.  Ju JY, Polhamus C, Marr KA, et al. Efficacies of fluconazole, caspofungin, and amphotericin B in Candida glabrata-infected p47phox-/- knockout mice. Antimicrob Agents Chemother 2002;46:1240–1245.

193.  Ibrahim AS, Avanessian V, Spellberg B, et al. Liposomal amphotericin B, and not amphotericin B deoxycholate, improves survival of diabetic mice infected with Rhizopus oryzae. Antimicrob Agents Chemother 2003;47:3343–3344.

194.  Luo GT, Gebremariam T, Lee H, et al. Efficacy of liposomal amphotericin B and posaconazole in intratracheal models of murine mucormycosis. Antimicrob Agents Chemother 2013l;57:3340–3347.

195.  Foldvari M, Jaafari MR, Radhi J, et al. Efficacy of the antiadhesin octyl O-(2-acetamido-2-deoxy-beta-D-galactopyranosyl)-(1-4)-2-O-propyl-beta-D-galactopy ranoside (Fimbrigal-P) in a rat oral candidiasis model. Antimicrob Agents Chemother 2005;49:2887–2894.

196.  Takakura N, Sato Y, Ishibashi H, et al. A novel murine model of oral candidiasis with local symptoms characteristic of oral thrush. Microbiol Immunol 2003;47:321–326.

197.  Andes D. Use of an animal model of disseminated candidiasis in the evaluation of antifungal therapy. Methods Mol Med 2005;118:111–128.

198.  Andes D, Ambrose PG, Hammel JP, et al. Use of pharmacokinetic-pharmacodynamic analyses to optimize therapy with the systemic antifungal micafungin for invasive candidiasis or candidemia. Antimicrob Agents Chemother 2011;55:2113–2121.

199.  Andes D, Diekema DJ, Pfaller MA, et al. In vivo comparison of the pharmacodynamic targets for echinocandin drugs against Candida species. Antimicrob Agents Chemother 2010;54:2497–2506.

200.  Andes D, Diekema DJ, Pfaller MA, et al. In vivo pharmacodynamic characterization of anidulafungin in a neutropenic murine candidiasis model. Antimicrob Agents Chemother 2008;52:539–550.

201.  Andes D, Forrest A, Lepak A, et al. Impact of antimicrobial dosing regimen on evolution of drug resistance in vivo: fluconazole and Candida albicans. Antimicrob Agents Chemother 2006;50:2374–2383.

202.  Andes D, Lepak A, Nett J, et al. In vivo fluconazole pharmacodynamics and resistance development in a previously susceptible Candida albicans population examined by microbiologic and transcriptional profiling. Antimicrob Agents Chemother 2006;50:2384–2394.

203.  Andes D, Marchillo K, Conklin R, et al. Pharmacodynamics of a new triazole, posaconazole, in a murine model of disseminated candidiasis. Antimicrob Agents Chemother 2004;48:137–142.

204.  Andes DK, Marchillo K, Lowther J, et al. In vivo pharmacodynamics of HMR 3270, a glucan synthase inhibitor, in a murine candidiasis model. Antimicrob Agents Chemother 2003;47:1187–1192.

205.  Andes D, Marchillo K, Stamstad T, et al. In vivo pharmacodynamics of a new triazole, ravuconazole, in a murine candidiasis model. Antimicrob Agents Chemother 2003;47:1193–1199.

206.  Andes D, Marchillo K, Stamstad T, et al. In vivo pharmacokinetics and pharmacodynamics of a new triazole, voriconazole, in a murine candidiasis model. Antimicrob Agents Chemother 2003;47:3165–3169.

207.  Andes D, Safdar N, Marchillo K, et al. Pharmacokinetic-pharmacodynamic comparison of amphotericin B (AMB) and two lipid-associated AMB preparations, liposomal AMB and AMB lipid complex, in murine candidiasis models. Antimicrob Agents Chemother2006;50:674–684.

208.  Andes D, Stamsted T, Conklin R. Pharmacodynamics of amphotericin B in a neutropenic-mouse disseminated-candidiasis model. Antimicrob Agents Chemother 2001;45:922–926.

209.  Andes D, van Ogtrop M. Characterization and quantitation of the pharmacodynamics of fluconazole in a neutropenic murine disseminated candidiasis infection model. Antimicrob Agents Chemother 1999;43:2116–2120.

210.  Andes D, van Ogtrop M, et al. In vivo characterization of the pharmacodynamics of flucytosine in a neutropenic murine disseminated candidiasis model. Antimicrob Agents Chemother 2000;44:938–942.

211.  Andes DR, Diekema DJ, Pfaller MA, et al. In vivo pharmacodynamic target investigation for micafungin against Candida albicans and C. glabrata in a neutropenic murine candidiasis model. Antimicrob Agents Chemother 2008;52:3497–3503.

212.  Arendrup MC, Perlin DS, Jensen RH, et al. Differential in vivo activities of anidulafungin, caspofungin, and micafungin against Candida glabrata isolates with and without FKS resistance mutations. Antimicrob Agents Chemother 2012;56:2435–2442.

213.  Clancy CJ, Cheng S, Nguyen MH. Animal models of candidiasis. Methods Mol Biol 2009;499:65–76.

214.  Gumbo T, Drusano GL, Liu W, et al. Once-weekly micafungin therapy is as effective as daily therapy for disseminated candidiasis in mice with persistent neutropenia. Antimicrob Agents Chemother 2007;51:968–974.

215.  Gumbo T, Drusano GL, Liu W, et al. Anidulafungin pharmacokinetics and microbial response in neutropenic mice with disseminated candidiasis. Antimicrob Agents Chemother 2006;50:3695–3700.

216.  Hope WW, Drusano GL, Moore CB, et al. Effect of neutropenia and treatment delay on the response to antifungal agents in experimental disseminated candidiasis. Antimicrob Agents Chemother 2007;51:285–295

217.  Hope WW, Mickiene D, Petraitis V, et al. The pharmacokinetics and pharmacodynamics of micafungin in experimental hematogenous Candida meningoencephalitis: implications for echinocandin therapy in neonates. J Infect Dis 2008;197:163–171.

218.  Hope WW, Warn PA, Sharp A, et al. Derivation of an in vivo drug exposure breakpoint for flucytosine against Candida albicans and impact of the MIC, growth rate, and resistance genotype on the antifungal effect. Antimicrob Agents Chemother 2006;50:3680–6388.

219.  Howard SJ, Livermore J, Sharp A, et al. Pharmacodynamics of echinocandins against Candida glabrata: requirement for dosage escalation to achieve maximal antifungal activity in neutropenic hosts. Antimicrob Agents Chemother 2011;55:4880–4887.

220.  Lepak A, Castanheira M, Diekema D, et al. Optimizing Echinocandin dosing and susceptibility breakpoint determination via in vivo pharmacodynamic evaluation against Candida glabrata with and without fks mutations. Antimicrob Agents Chemother 2012;56:5875–5882.

221.  Lepak A, Nett J, Lincoln L, et al. Time course of microbiologic outcome and gene expression in Candida albicans during and following in vitro and in vivo exposure to fluconazole. Antimicrob Agents Chemother 2006;50:1311–1319.

222.  Louie A, Banarjee P, Drusano GL, et al. Interaction between fluconazole and amphotericin B in mice with systemic infection due to fluconazole-susceptible or -resistant strains of Candida albicans. Antimicrob Agents Chemother 1999;43:2841–2847.

223.  Louie A, Deziel M, Liu W, et al. Pharmacodynamics of caspofungin in a murine model of systemic candidiasis: importance of persistence of caspofungin in tissues to understanding drug activity. Antimicrob Agents Chemother 2005;49:5058–5068.

224.  Louie A, Drusano GL, Banerjee P, et al. Pharmacodynamics of fluconazole in a murine model of systemic candidiasis. Antimicrob Agents Chemother 1998;42:1105–1109.

225.  MacCallum DM, Coster A, Ischer A, et al. Genetic dissection of azole resistance mechanisms in Candida albicans and their validation in a mouse model of disseminated infection. Antimicrob Agents Chemother 2010;54:1476–1483.

226.  Marine M, Pastor FJ, Guarro J, et al. Efficacy of posaconazole in a murine disseminated infection by Candida tropicalis. Antimicrob Agents Chemother 2010;54:530–532.

228.  Sanati H, Ramos CF, Bayer AS, et al. Combination therapy with amphotericin B and fluconazole against invasive candidiasis in neutropenic-mouse and infective-endocarditis rabbit models. Antimicrob Agents Chemother 1997;41:1345–1348.

229.  Spellberg B, Fu Y, Edwards JE Jr, et al. Combination therapy with amphotericin B lipid complex and caspofungin acetate of disseminated zygomycosis in diabetic ketoacidotic mice. Antimicrob Agents Chemother 2005;49:830–832.

230.  Spellberg B, Ibrahim AS, Edwards JE Jr, et al. Mice with disseminated candidiasis die of progressive sepsis. J Infect Dis 2005;192:336–343.

231.  Sugar AM. Interactions of amphotericin B and SCH 39304 in the treatment of experimental murine candidiasis: lack of antagonism of a polyene-azole combination. Antimicrob Agents Chemother 1991;35:1669–1671.

232.  Sugar AM, Goldani LZ, Picard M. Treatment of murine invasive candidiasis with amphotericin B and cilofungin: evidence for enhanced activity with combination therapy. Antimicrob Agents Chemother 1991;35:2128–2130.

233.  Sugar AM, Hitchcock CA, Troke PF, et al. Combination therapy of murine invasive candidiasis with fluconazole and amphotericin B. Antimicrob Agents Chemother 1995;39:598–601.

234.  Sugar AM, Liu XP. Interactions of itraconazole with amphotericin B in the treatment of murine invasive candidiasis. J Infect Dis 1998;177:1660–1663.

235.  Wiederhold NP, Najvar LK, Bocanegra R, et al. Comparison of anidulafungin’s and fluconazole’s in vivo activity in neutropenic and non-neutropenic models of invasive candidiasis. Clin Microbiol Infect 2012;18:E20–E23.

236.  Wiederhold NP, Najvar LK, Bocanegra R, et al. In vivo efficacy of anidulafungin and caspofungin against Candida glabrata and association with in vitro potency in the presence of sera. Antimicrob Agents Chemother 2007;51:1616–1620.

237.  Wiederhold NP, Najvar LK, Bocanegra RA, et al. Caspofungin dose escalation for invasive candidiasis due to resistant Candida albicans. Antimicrob Agents Chemother 2011;55:3254–3260.

238.  Brieland J, Essig D, Jackson C, et al. Comparison of pathogenesis and host immune responses to Candida glabrata and Candida albicans in systemically infected immunocompetent mice. Infect Immun 2001;69:5046–5055.

239.  Demant E, Easterbrook M. An experimental model of candida endophthalmitis. Can J Ophthalmol 1977;12:304–307.

240.  Deren YT, Ozdek S, Kalkanci A, et al. Comparison of antifungal efficacies of moxifloxacin, liposomal amphotericin B, and combination treatment in experimental Candida albicans endophthalmitis in rabbits. Can J Microbiol 2010;56:1–7.

241.  Edwards JE Jr, Montgomerie JZ, Foos RY, et al. Experimental hematogenous endophthalmitis caused by Candida albicans. J Infect Dis 1975;131:649–657.

242.  Gupta SK, Dhingra N, Velpandian T, et al. Efficacy of fluconazole and liposome entrapped fluconazole for C. albicans induced experimental mycotic endophthalmitis in rabbit eyes. Acta Ophthalmol Scand 2000;75:448–450.

243.  Kusbeci T, Avci B, Cetinkaya Z, et al. The effects of caspofungin and voriconazole in experimental Candida endophthalmitis. Curr Eye Res 2007;57–64.

244.  Livermore JL, Felton TW, Abbott J, et al. Pharmacokinetics and pharmacodynamics of anidulafungin for experimental Candida endophthalmitis: insights into the utility of echinocandins for treatment of a potentially sight-threatening infection. Antimicrob Agents Chemother2013;57:281–282.

245.  Louie A, Liu W, Miller DA, et al. Efficacies of high-dose fluconazole plus amphotericin B and high-dose fluconazole plus 5-fluorocytosine versus amphotericin B, fluconazole, and 5-fluorocytosine monotherapies in treatment of experimental endocarditis, endophthalmitis, and pyelonephritis due to Candida albicans. Antimicrob Agents Chemother 1999;43:2831–2840.

246.  Omuta J, Uchida K, Yamaguchi H, et al. Histopathological study on experimental endophthalmitis induced by bloodstream infection with Candida albicans. Jpn J Infect Dis 2007;60:33–39.

247.  Savani DV, Perfect JR, Cobo LM, et al. Penetration of new azole compounds into the eye and efficacy in experimental Candida endophthalmitis. Antimicrob Agents Chemother 1987;31:6–10

248.  Goldblum D, Frueh BE, Sarra GM, et al. Topical caspofungin for treatment of keratitis caused by Candida albicans in a rabbit model. Antimicrob Agents Chemother 2005;49:1359–1363.

249.  Hu J, Wang Y, Xie L. Potential role of macrophages in experimental keratomycosis. Invest Ophthalmol Vis Sci 2009;50:2087–2094.

250.  Ishibashi Y, Kaufman HE. Topical ketoconazole for experimental Candida keratitis in rabbits. Am J Ophthalmol 1986;102:522–526.

251.  Mitchell BM, Wu TG, Jackson BE, et al. Candida albicans strain-dependent virulence and Rim13p-mediated filamentation in experimental keratomycosis. Invest Ophthalmol Vis Sci 2007;48:774–780.

252.  O’Day DM, Head WS, Csank C, et al. Differences in virulence between two Candida albicans strains in experimental keratitis. Invest Ophthalmol Vis Sci 2000;41:1116–1121.

253.  Ohno S, Fuerst DJ, Okumoto M, et al. The effect of K-582, a new antifungal agent, on experimental Candida keratitis. Invest Ophthalmol Vis Sci 1983;24:1626–1629.

254.  Pleyer U, Legmann A, Mondino BJ, et al. Use of collagen shields containing amphotericin B in the treatment of experimental Candida albicans-induced keratomycosis in rabbits. Am J Ophthalmol 1992;113:303–308.

255.  Ray WA, O’Day DM, Head WS, et al. Variability in isolate recovery rates from multiple and single breeds of outbred pigmented rabbits in an experimental model of Candida keratitis. Curr Eye Res 1984;3:949–953.

256.  Wu TG, Wilhelmus KR, Mitchell BM. Experimental keratomycosis in a mouse model. Invest Ophthalmol Vis Sci 2003;44:210–216.

257.  Yuan X, Hua X, Wilhelmus KR. The corneal expression of antimicrobial peptides during experimental fungal keratitis. Curr Eye Res 2010;35:872–879.

258.  Yuan X, Mitchell BM, Wilhelmus KR. Gene profiling and signaling pathways of Candida albicans keratitis. Mol Vis 2008;14:1792–1798.

259.  Zhong W, Yin H, Xie L. Expression and potential role of major inflammatory cytokines in experimental keratomycosis. Mol Vis 2009;15:1303–1311.

260.  Mavridou E, Bruggeman RJ, Melchers WJ, et al. Efficacy of posaconazole against three clinical Aspergillus fumigatus isolates with mutations in the cyp51A gene. Antimicrob Agents Chemother 2010;54:860–865.

261.  Bowman JC, Abruzzo GK, Anderson JW, et al. Quantitative PCR assay to measure Aspergillus fumigatus burden in a murine model of disseminated aspergillosis: demonstration of efficacy of caspofungin acetate. Antimicrob Agents Chemother 2001;3474–3481.

262.  Vallor AC, Kirkpatrick WR, Navjar LK, et al. Assessment of Aspergillus fumigatus burden in pulmonary tissue of guinea pigs by quantitative PCR, galactomannan enzyme immunoassay, and quantitative culture. Antimicrob Agents Chemother 2008;52:2593–2598.

263.  Hope WW, V Petraitis, Petraitiene R, et al. The initial 96 hours of invasive pulmonary aspergillosis: histopathology, comparative kinetics of galactomannan and (1->3) beta-d-glucan and consequences of delayed antifungal therapy. Antimicrob Agents Chemother2010;54:4879–4886.

264.  Lewis RE, Ben-Ami R, Best L, et al. Tacrolimus enhances the potency of posaconazole against Rhizopus oryzae in vitro and in an experimental model of mucormycosis. J Infect Dis 2013;207:834–841.

265.  Ibrahim AS, Bowman SJC, Avanessian V, et al. Caspofungin inhibits Rhizopus oryzae 1,3-beta-D-glucan synthase, lowers burden in brain measured by quantitative PCR, and improves survival at a low but not a high dose during murine disseminated zygomycosis. Antimicrob Agents Chemother2005;49:721–727.

266.  Lamaris GA, Ben-Ami R, Lewis RE, et al. Increased virulence of Zygomycetes organisms following exposure to voriconazole: a study involving fly and murine models of zygomycosis. J Infect Dis 2009;199:1399–1406.

267.  Lewis RE, Albert ND, Liao G, et al. Comparative pharmacodynamics of amphotericin B lipid complex and liposomal amphotericin B in a murine model of pulmonary mucormycosis. Antimicrob Agents Chemother 2010;54:1298–1304.

268.  Rodriguez MM, Serena C, Marine M, et al. Posaconazole combined with amphotericin B, an effective therapy for a murine disseminated infection caused by Rhizopus oryzae. Antimicrob Agents Chemother 2008;52:3786–3788.

269.  Calvo E, Pastor FJ, Rodriguez MM, et al. Antifungal therapy in a murine model of disseminated infection by Cryptococcus gattii. Antimicrob Agents Chemother 2010;54:4074–4077.

270.  Clemons KV, Stevens DA. Comparison of fungizone, Amphotec, AmBisome, and Abelcet for treatment of systemic murine cryptococcosis. Antimicrob Agents Chemother 1998;42:899–902.

271.  Larsen RA, Bauer M, Thomas AM, et al. Correspondence of in vitro and in vivo fluconazole dose-response curves for Cryptococcus neoformans. Antimicrob Agents Chemother 2005;49:3297–3301.

272.  Lengerova M, Kocmanova I, Racil K, et al. Detection and measurement of fungal burden in a guinea pig model of invasive pulmonary aspergillosis by novel quantitative nested real-time PCR compared with galactomannan and (1,3)-beta-D-glucan detection. J Clin Microbiol2012;50:602–608.

273.  Schwarz P, Dromer F, Lortholary O, et al. Efficacy of amphotericin B in combination with flucytosine against flucytosine-susceptible or flucytosine-resistant isolates of Cryptococcus neoformans during disseminated murine cryptococcosis. Antimicrob Agents Chemother2006;50:113–120.

274.  Serena C, Pastor FJ, Marine M, et al. Efficacy of voriconazole in a murine model of cryptococcal central nervous system infection. J Antimicrob Chemother 2007;60:162–165.

275.  Thompson GR 3rd, Wiederhold NP, Najvar R, et al. A murine model of Cryptococcus gattii meningoencephalitis. J Antimicrob Chemother 2012;67:1432–1438.

276.  Widmer F, Wright LC, Obando D, et al. Hexadecylphosphocholine (miltefosine) has broad-spectrum fungicidal activity and is efficacious in a mouse model of cryptococcosis. Antimicrob Agents Chemother 2006;50:414–421.

277.  Wiederhold NP, Navjar LK, Bocanegra R, et al. Limited activity of miltefosine in murine models of cryptococcal meningoencephalitis and disseminated cryptococcosis. Antimicrob Agents Chemother 2013;57:745–750.

278.  Zaragoza O, Mihu C, Casadevall A, et al. Effect of amphotericin B on capsule and cell size in Cryptococcus neoformans during murine infection. Antimicrob Agents Chemother 2005;49:4358–4361.

279.  Chiller TM, Sobel RA, Luque JC, et al. Efficacy of amphotericin B or itraconazole in a murine model of central nervous system Aspergillus infection. Antimicrob Agents Chemother 2003;47:813–815.

280.  Clemons KV, Espiritu M, Parmar R, et al. Comparative efficacies of conventional amphotericin b, liposomal amphotericin B (AmBisome), caspofungin, micafungin, and voriconazole alone and in combination against experimental murine central nervous system aspergillosis. Antimicrob Agents Chemother 2005;49:4867–4875.

281.  Clemons KV, Parmar R, Martinez M, et al. Efficacy of Abelcet alone, or in combination therapy, against experimental central nervous system aspergillosis. J Antimicrob Chemother 2006;58:466–469.

282.  Clemons KV, Schwartz JA, Stevens DA. Experimental central nervous system aspergillosis therapy: efficacy, drug levels and localization, immunohistopathology, and toxicity. Antimicrob Agents Chemother 2012;56:4439–4449.

283.  Clemons KV, Stevens DA. The contribution of animal models of aspergillosis to understanding pathogenesis, therapy and virulence. Med Mycol 2005;43(Suppl 1):S101–S110.

284.  Imai J, Singh G, Fernandez B, et al. Efficacy of Abelcet and caspofungin, alone or in combination, against CNS aspergillosis in a murine model. J Antimicrob Chemother 2005;56:166–171.

285.  Imai JK, Singh G, Clemons KV, et al. Efficacy of posaconazole in a murine model of central nervous system aspergillosis. Antimicrob Agents Chemother 2004;48:4063–4066.

286.  Singh G, Imai J, Clemons KV, et al. Efficacy of caspofungin against central nervous system Aspergillus fumigatus infection in mice determined by TaqMan PCR and CFU methods. Antimicrob Agents Chemother 2005;49:1369–1376.

287.  Flattery AM, Hickey E, Gill CJ, et al. Efficacy of caspofungin in a juvenile mouse model of central nervous system candidiasis. Antimicrob Agents Chemother 2011;55:3491–3497.

288.  Kang CI, Rouse MS, Mandrekar JN, et al. Anidulafungin treatment of candidal central nervous system infection in a murine model. Antimicrob Agents Chemother 2009;53:3576–3578.

289.  Warn PA, Livermore J, Howard S, et al. Anidulafungin for neonatal hematogenous Candida meningoencephalitis: identification of candidate regimens for humans using a translational pharmacological approach. Antimicrob Agents Chemother 2012;56:708–714.

290.  Capilla J, Clemons KV, Sobel RA, et al. Efficacy of amphotericin B lipid complex in a rabbit model of coccidioidal meningitis. J Antimicrob Chemother 2007;60:673–676.

291.  Clemons KV, Capilla J, Sobel RA, et al. Comparative efficacies of lipid-complexed amphotericin B and liposomal amphotericin B against coccidioidal meningitis in rabbits. Antimicrob Agents Chemother 2009;53:1858–1862.

292.  Gonzalez GM, Gonzalez G, Najvar LK, et al Therapeutic efficacy of caspofungin alone and in combination with amphotericin B deoxycholate for coccidioidomycosis in a mouse model. J Antimicrob Chemother 2007;60:1341–1346.

293.  Kamberi P, Sobel RA, Clemons A, et al. Comparison of itraconazole and fluconazole treatments in a murine model of coccidioidal meningitis. Antimicrob Agents Chemother 2007;51:998–1003.

294.  Sorensen KN, Sobel RA, Clemons KV, et al. Comparative efficacies of terbinafine and fluconazole in treatment of experimental coccidioidal meningitis in a rabbit model. Antimicrobial Agents Chemother 2000;44:3087–3091.

295.  Sorensen KN, Sobel RA, Clemons KV, et al. Comparison of fluconazole and itraconazole in a rabbit model of coccidioidal meningitis. Antimicrob Agents Chemother 2000;44:1512–1517.

296.  Williams PL, Sobel RA, Sorensen KN, et al. A model of coccidioidal meningoencephalitis and cerebrospinal vasculitis in the rabbit. J Infect Dis 1998;178:1217–1221.

297.  Sorensen KN, Clemons KV, Stevens DA. Murine models of blastomycosis, coccidioidomycosis, and histoplasmosis. Mycopathologia 1999;146:53–65.

298.  Al-Abdely HM, Najvar LK, Bocanegra R, et al. Antifungal therapy of experimental cerebral phaeohyphomycosis due to Cladophialophora bantiana. Antimicrob Agents Chemother 2005;49:1701–1707.

299.  Calvo E, Pastor FJ, Guarro J. Antifungal therapies in murine disseminated phaeohyphomycoses caused by Exophiala species. J Antimicrob Chemother 2010;65:1455–1459.

300.  Calvo E, Pastor FJ, Rodriguez MM, et al. Murine model of a disseminated infection by the novel fungus Fonsecaea monophora and successful treatment with posaconazole. Antimicrob Agents Chemother 2010;54:919–923.

301.  Itoyama T, Uchida K, Yamaguchi H. Therapeutic effects of omoconazole nitrate on guinea-pigs experimentally infected with Trichophyton mentagrophytes. J Antimicrob Chemother 1997;39:825–827.

302.  Itoyama T, Uchida K, Yamaguchi H, et al. Therapeutic effects of omoconazole nitrate on experimental tinea pedis, an intractable dermatophytosis, in guinea-pigs. J Antimicrob Chemother 1997;40:441–444.

303.  Koga H, Nanjoh Y, Kaneda H, et al. Short-term therapy with luliconazole, a novel topical antifungal imidazole, in guinea pig models of tinea corporis and tinea pedis. Antimicrob Agents Chemother 2012;56:3138–3143.

304.  Saunte DM, Simmel F, Frimodt-Møller N, et al. In vivo efficacy and pharmacokinetics of voriconazole in an animal model of dermatophytosis. Antimicrob Agents Chemother 2007;51:3317–3321.

305.  Shimamura T, Kobota N, Nagasaka S, et al. Establishment of a novel model of onychomycosis in rabbits for evaluation of antifungal agents. Antimicrob Agents Chemother 2011;55:3150–3155.

306.  Serena C, Marine M, Marimon R, et al. Effect of antifungal treatment in a murine model of blastoschizomycosis. Int J Antimicrob Agents 2007;29:79–83.

307.  Serena C, Rodriguez MM, Marine M, et al. Combined therapies in a murine model of blastoschizomycosis. Antimicrob Agents Chemother 2007;51:2608–2610.

308.  Serena C, Rodriguez MM, Marine M, et al. Micafungin combined with fluconazole, an effective therapy for murine blastoschizomycosis. J Antimicrob Chemother 2008;61:877–879.

309.  Lionakis MS, Chamilos G, Lewis RE, et al. Pentamidine is active in a neutropenic murine model of acute invasive pulmonary fusariosis. Antimicrobial agents and chemotherapy 50:294-297.

310.  Ruiz-Cendoya M, Marine M, Guarro J. Combined therapy in treatment of murine infection by Fusarium solani. J Antimicrob Chemother 2008;62:543–546.

311.  Ruiz-Cendoya M, Marine M, Rodriguez MM, et al. Interactions between triazoles and amphotericin B in treatment of disseminated murine infection by Fusarium oxysporum. Antimicrob Agents Chemother 2009;53:1705–1708.

312.  Wiederhold NP, Najvar LK, Bocanegra R, et al. Efficacy of posaconazole as treatment and prophylaxis against Fusarium solani. Antimicrob Agents Chemother 2010;54:1055–1059.

313.  Bocanegra R, Najvar LK, Hernandez S, et al. Caspofungin and liposomal amphotericin B therapy of experimental murine scedosporiosis. Antimicrob Agents Chemother 2005;49:5139–5141.

314.  Rodriguez MM, Calvo E, Serena C, et al. Effects of double and triple combinations of antifungal drugs in a murine model of disseminated infection by Scedosporium prolificans. Antimicrob Agents Chemother 2009;53:2153–2155.

315.  Rodriguez MM, Pastor FJ, Salas V, et al. Experimental murine scedosporiosis: histopathology and azole treatment. Antimicrob Agents Chemother 2010;54:3480–3984.

316.  Serena C, Gilgado M, Marine M, et al. Efficacy of voriconazole in a guinea pig model of invasive trichosporonosis. Antimicrob Agents Chemother 2006;50:2240–2243.

317.  Serena C, Pastor FJ, Gilgado F, et al. Efficacy of micafungin in combination with other drugs in a murine model of disseminated trichosporonosis. Antimicrob Agents Chemother 2005;49:497–502.

318.  Ryan DM. The usefulness of experimental models of urinary tract infections in the assessment of chemotherapeutic compounds. In: Williams JD, Gedes AM, eds. Chemotherapy. Vol. 2. New York: Plenum, 1976:205–215.

319.  Roberts JA. Vesicoureteral reflux and pyelonephritis in the monkey: a review. J Urol 1992;148:1721–1725.

320.  Goodman JL, Jurkovich P, Kodner C, et al. Persistent cardiac and urinary tract infections with Borrelia burgdorferi in experimentally infected Syrian hamsters. J Clin Microbiol 1991;29:894–896.

321.  Kaijser B, Larsson P. Experimental acute pyelonephritis caused by enterobacteria in animals: a review. J Urol 1982;127:786–790.

322.  Yagmurlu A, Boleken ME, Ertoy D, et al. Preventive effect of pentoxifylline on renal scarring in rat model of pyelonephritis. Urology 2003;61:1037–1041.

323.  Glauser MP, Ransley P, Bille J. Urinary tract infections, pyelonephritic scars, and chemotherapy. In: Zak O, Sande MA, eds. Experimental models in antimicrobial chemotherapy. Vol. 1. London: Academic Press, 1986:319–346.

324.  Rank RG. Animal models for urogenital infections. Methods Enzymol 1994;235:83–93.

325.  Brooks SJD, Lyons JM, Braude AL. Immunization against retrograde pyelonephritis. I. Production of an experimental model of severe ascending Escherichia coli pyelonephritis without bacteremia in rats. Am J Pathol 1974;74:345–358.

326.  Glauser MP, Lyons JM, Braude AI. Synergism of ampicillin and gentamicin against obtructive pyelonephritis due to Escherichia coli in rats. J Infect Dis 1979;139:133–140.

327.  Meylan PR, Glauser MP. Failure of dexamethasone to prevent polymorphonuclear leukocyte infiltration during experimental acute exudative pyelonephritis and to reduce subsequent chronic scarring. J Infect Dis 1988;157:480–485.

328.  Glauser MP, Bonard M. Treatment of experimental ascending Escherichia coli pyelonephritis with ceftriaxone alone and in combination with gentamicin. Chemotherapy 1982;28:410–416.

329.  Lecamwasam JP, Miller TE. Antimicrobial agents in the management of urinary tract infection: an experimental evaluation. J Lab Clin Med 1989;114:510–519.

330.  Bergeron MG, Marois Y. Benefit from high intraenal levels of gentamicin in the treatment of E. coli pyelonephritis. Kidney Int 1986;30:481–487.

331.  LeBrun M, Grenier L, Gourde P, et al. Effectiveness and toxicity of gentamicin in an experimental model of pyelonephritis: effect of time of administration. Antimicrob Agents Chemother 1999;43:1020–1026.

332.  Miller TE, Findon G, Rainer SP, et al. The pathobiology of subclinical pyelonephritis: an experimental evaluation. Kidney Int 1992;41:1356–1365.

333.  Haraoka M, Matsumoto T, Takashi K, et al. Suppression of renal scarring by prednisolone combined with ciprofloxacin in ascending pyelonephritis in rats. J Urol 1994;151:1078–1080.

334.  Li C-R, Li Y, Li G-Q et al. In vivo antibacterial activity of nemonoxacin, a novel non-fluorinated quinolone. J Antimicrob Chemother 2010;65:2411–2415.

335.  Anderson GG, Palermo JJ, Schilling JD, et al. Intracellular bacterial biofilm-like pods in urinary tract infections. Science 2003;301:105–107.

336.  Mulvey MA, Schilling JD, Hultgren SJ. Establishment of a persistent Escherichia coli reservoir during the acute phase of a bladder infection. Infect Immun 2001;69:4572–4579.

337.  Mulvey MA, Schilling JD, Martinez JJ, et al. Bad bugs and beleaguered bladders: interplay between uropathogenic Escherichia coli and innate host defenses. Proc Natl Acad Sci USA 2000;97:8829–8835.

338.  Schilling JD, Martin SM, Hung CS, et al. Toll-like receptor 4 on stromal and hematopoietic cells mediates innate resistance to uropathogenic Escherichia coli. Proc Natl Acad Sci USA 2003;100:4203–4208.

339.  Hagberg L, Engberg I, Freter R, et al. Ascending unobstructed urinary tract infection in mice caused by pyelonephritogenic Escherichia coli of human origin. Infect Immun 1983;40:273–283.

340.  Hvidberg H, Struve C, Krogfelt KA, et al. Development of a long-term ascending urinary tract infection mouse model for antibiotic treatment studies. Antimicrob Agents Chemother 2000;44:156–163.

341.  Johnson JR, Brown JJ. Defining inoculation conditions for the mouse model of ascending urinary tract infection that avoid immediate vesicoureteral reflux yet produce renal and bladder infection. J Infect Dis 1996;173:1306–1311.

342.  Kerrn MB, Frimodt-Møller N, Espersen F. Effects of sulfamethizole and amdinocillin against Escherichia coli strains (with various susceptibilities) in an ascending urinary tract infection mouse model. Antimicrob Agents Chemother 2003;47:1002–1009.

343.  Frimodt-Møller N. Correlation between pharmacokinetic/pharmacodynamic parameters and efficacy for antibiotics in the treatment of urinary tract infection. Int J Antimicrob Agents 2002;19:546–553.

344.  Tsuchimori N, Yamasaki T, Okonogi K. Therapeutic effects of cefozopran against experimental mixed urinary tract infection with Enterococcus faecalis and Pseudomonas aeruginosa in mice. J Antimicrob Chemother 1997;39:423–425.

345.  Allou N, Cambau E, Massias L, et al. Impact of low-level resistance to fluoroquinolones due to qnrA1 and qnrS1 genes or a gyrA mutation on ciprofloxacin bactericidal activity in a murine model of Escherichia coli urinary tract infection. Antimicrob Agents Chemother2009;53:4292–4297.

346.  Jakobsen L, Cattoir V, Jensen KS, et al. Impact of low-level fluoroquinolone resistance genes qnrA1, qnrB19 and qnrS1 on ciprofloxacin treatment of isogenic Escherichia coli strains in a murine urinary tract infection model. J Antimicrob Chemother 2012;67:2438–2444.

347.  Johnson DE, Lockatell CV. Mouse model of ascending UTI involving short- and long-term indwelling catheters. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:441–445.

348.  Johnson DE, Lockatell CV, Hall-Craggs M, et al. Mouse models of short- and long-term foreign body in the urinary bladder: analogies to the bladder segment of urinary catheters. Lab Anim Sci 1991;41:451–455.

349.  Mikuniya T, Kato Y, Ida T, et al. Treatment of Pseudomonas aeruginosa biofilms with a combination of fluoroquinolones and fosfomycin in a rat urinary tract infection model. J Infect Chemother 2007;13:285–290.

350.  Wang X, Fries BC. A murine model for catheter-associated candiduria. J Med Microbiol 2011;60:1523–1529.

351.  Evliyaoğlu Y, Kobaner M, Celebi H, et al. The efficacy of a novel antibacterial hydroxyapatite nanoparticle-coated indwelling urinary catheter in preventing biofilm formation and catheter-associated urinary tract infection in rabbits. Urol Res 2011;39:443–449.

352.  Yuste J, Jado I, Fenoll A, et al. β-lactam modification of the bacteraemic profile and its relationship with mortality in a pneumococcal mouse sepsis model. J Antimicrob Chemother 2002;49:331–335.

353.  Griffith DC, Harford L, Williams R, et al. In vivo antibacterial activity of RWJ-54428, a new cephalosporin with activity against Gram-positive bacteria. Antimicrob Agents Chemother 2003;47:43–47.

354.  Montgomerie JZ, Schick DG. Treatment of enterococcal pyelonephritis with trovafloxacin and rifampin: in vitro in vivo contrast. Antimicrob Agents Chemother 1998;42:188–189.

355.  Sapico FL, Ginunas VJ, Montgomerie JZ, et al. Cefpirome, alone and in combination with gentamicin, for enterococcal pyelonephritis in the rodent model. Diagn Microbiol Infect Dis 1991;14:297–300.

356.  Fu KP, Foleno BD, Lafredo SC, et al. In vitro and in vivo antibacterial activities of FK037, a novel parenteral broad-spectrum cephalosporin. Antimicrob Agents Chemother 1993;37:301–307.

357.  Attardo Genco C, Arko RJ. Animal chamber models for study of host-parasite interactions. Methods Enzymol 1994;235:120–140.

358.  Espersen F, Frimodt-Møller N, Corneliussen L, et al. Experimental foreign body infection in mice. J Antimicrob Chemother 1993;31(Suppl D):103–111.

359.  Zimmerli W. Experimental models in the investigation of device-related infections. J Antimicrob Chemother 1993;31(Suppl D):97–102.

360.  Zimmerli W. Tissue cage infection model. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:409–417.

361.  de Lalla F. Antimicrobial chemotherapy in the control of surgical infectious complications. J Chemother 1999;11:440–445

362.  Pascual A. Pathogenesis of catheter-related infections: lessons for new designs. Clin Microbiol Infect 2002;8:256–264.

363.  Zimmerli W, Waldvogel FA, Vaudaux P, et al. Pathogenesis of foreign body infection: description and characteristics of an animal model. J Infect Dis 1982;146:487–497.

364.  Chuard C, Vaudaux P, Waldvogel FA, et al. Susceptibility of Staphylococcus aureus growing on fibronectin-coated surfaces to bactericidal antibiotics. Antimicrob Agents Chemother 1993;37:625–632.

365.  Furustrand Tafin U, Majic I, Zalila Belkhodja, et al. Gentamicin improves the activities of daptomycin and vancomycin against Enterococcus faecalis in vitro and in an experimental foreign-body infection model. Antimicrob Agents Chemother 2011;55(10):4281–4827.

366.  John A-K, Schmaler M, Khanna N, et al. Reversible daptomycin tolerance of adherent staphylococci in an implant infection model. Antimicrob Agents Chemother 2011;55(7):3510–3516.

367.  Murillo O, Pachon ME, Euba G, et al. Antagonistic effect of rifampin on the efficacy of levofloxacin at high doses in staphylococcal experimental foreign-body infection. Antimicrob Agents Chemother 2008;52(10):3681–3686.

368.  Artini M, Scoarughi GL, Papa R, et al. Comparison of anti-bacterial prophylactic properties of two different vascular grafts: action of anti-bacterial graft coating and systemic antibiotic treatment. Int J Immunopathol Pharmacol 2010;23:383–386.

369.  Jean-Baptiste E, Blanchemain N, Martel B, et al. Safety, healing, and efficacy of vascular prostheses coated with hydroxypropyl-β-cyclodextrin polymer: experimental in vitro and animal studies. Eur J Vasc Endovasc Surg 2012;43:188–197.

370.  Polk HC, Lamont PM, Galland RB. Containment as a mechanism of nonspecific enhancement of defenses against bacterial infection. Infect Immun 1990;58:1807–1811.

371.  Naziri W, Cheadle WG, Trachtenberg LS, et al. Increased antibiotic effectivness in a model of surgical infection through continuous infusion. Am Surg 1995;61:11–15.

372.  Dai T, Kharkwal GB, Tanaka M, et al. Animal models of external traumatic wound infections. Virulence 2011;2,296–315.

373.  Kaiser AB, Kernodle DS. Low-inoculum model of clean wound infection. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:205–211.

374.  Simonetti O, Cirioni O, Goteri G, et al. Temporin A is effective in MRSA-infected wounds through bactericidal activity and acceleration of wound repair in a murine model. Peptides 2008;29:520–528.

375.  Zaleski KJ, Kolodka T, Cywes-Bentley C, et al. Hyaluronic acid binding peptides prevent experimental staphylococcal wound infection. Antimicrob Agents Chemother 2006;50:3856–3860.

376.  Ozcan AV, Demir M, Onem G, et al. Topical versus systemic vancomycin for deep sternal wound infection caused by methicillin-resistant Staphylococcus aureus in a rodent experimental model. Tex Heart Inst J 2006;33:107–110.

377.  Mihu MR, Sandkovsky U, Han G, et al. The use of nitric oxide releasing nanoparticles as a treatment against Acinetobacter baumannii in wound infections. Virulence 2010;1:62–67.

378.  Dijkstra HM, Manson WL, Klasen HJ, et al. Effect of polymixin B on intestinal bacterial translocation in Pseudomonas aeruginosa wound-colonized burned mice. Eur Surg Res 1992;24:69–76.

379.  Manson WL, Coenen JMFH, Klasen HJ, et al. Intestinal bacterial transolcation in experimentally burned mice with wounds colonized by Pseudomonas aeruginosa. J Trauma 1992;33:654–658.

380.  Simonetti O, Cirioni O, Orlando F, et al. Effectiveness of antimicrobial photodynamic therapy with a single treatment of RLP068/Cl in an experimental model of Staphylococcus aureus wound infection. Br J Dermatol 2011;164:987–995.

381.  Ulkür E, Oncul O, Karagoz H et al. Comparison of silver-coated dressing (Acticoat), chlorhexidine acetate 0.5% (Bactigrass), and fusidic acid 2% (Fucidin) for topical antibacterial effect in methicillin-resistant Staphylococci-contaminated, full-skin thickness rat burn wounds. Burns 2005;31:874–877.

382.  Simonetti O, Cirioni O, Lucarini G et al. Tigecycline accelerates staphylococcal-infected burn wound healing through matrix metalloproteinase-9 modulation. J Antimicrob Chemother 2012;67:191–201.

383.  Kumari S, Harjai K, Chhibber S. Evidence to support the therapeutic potential of bacteriophage Kpn5 in burn wound infection caused by Klebsiella pneumoniae in BALB/c mice. J Microbiol Biotechnol 2010;20:935–941.

384.  Acar A, Uygur F, Diktaş H et al. Comparison of silver-coated dressing (Acticoat^®), chlorhexidine acetate 0.5% (Bactigrass^®) and nystatin for topical antifungal effect in Candida albicans-contaminated, full-skin-thickness rat burn wounds. Burns 2011;37:882–885.

385.  Mahoney E, Reichner J, Bostom LR, et al. Bacterial colonization and the expression of inducible nitric oxide synthase in murine wounds. Am J Pathol 2002;161:2143–2152.

386.  Kugelberg E, Norström T, Petersen TK, et al. Establishment of a superficial skin infection model in mice by using Staphylococcus aureus and Streptococcus pyogenes. Antimicrob Agents Chemother 2005;49:3435–3441.

387.  Vingsbo Lundberg C, Frimodt-Møller N. Efficacy of topical and systemic antibiotic treatment of methicillin-resistant Staphylococcus aureus in a murine superficial skin wound infection model. Int J Antimicrob Agents 2013;42:272–275.

388.  Ansfield MJ, Woods DE, Johanson WG. Lung bacterial clearance in murine pneumococcal pneumonia. Infect Immun 1977;17:195–204.

389.  Bakker-Woudenberg IAJM, Jong-Hoenderop YT, Michel MF. Efficacy of antimicrobial therapy in experimental rat pneumonia: effects of impaired phagocytosis. Infect lmmun 1979;25:366–375.

390.  Davis CC, Mellencamp MA, Preheim LC. A model of pneumococcal pneumonia in chronically intoxicated rats. J Infect Dis 1991;163:799–805.

391.  Mellencamp MA, Preheim LC. Pneumococcal pneumonia in a rat model of cirrhosis: effects of cirrhosis on pulmonary defense mechanisms against Streptococcus pneumoniae. J Infect Dis 1991;163:102–108.

392.  Wakebe H, Imada T, Yoneda H, et al. Evaluation of OPC-17116 against important pathogens that cause respiratory tract infections. Antimicrob Agents Chemother 1994;38:2340–2345.

393.  Gavalda J, Capdevila JA, Almirante B, et al. Treatment of experimental pneumonia due to penicillin-resistant Streptococcus pneumoniae in immunocompetent rats. Antimicrob Agents Chemother 1997;41:795–801.

394.  Beskid G, Christenson JG, Cleeland R, et al. In vivo activity of ceftriaxone (Ro 13-9904), a new broad-spectrum semisynthetic cephalosporin. Antimicrob Agents Chemother 1981;20:159–167.

395.  Mizgerd JP, Skerrett SJ. Animal models of human pneumonia. American journal of physiology. Lung Cell Mol Physiol 2008;294:L387–L398.

396.  Miyazaki S, Nunoya T, Matsumoto T, et al. New murine model of bronchopneumonia due to cell-bound Haemophilus influenzae. J Infect Dis 1997;175:205–209.

397.  Fukuoka T, Kawada H, Kitayama A, et al. Efficacy of CS-834 against experimental pneumonia caused by penicillin-susceptible and -resistant Streptococcus pneumoniae in mice. Antimicrob Agents Chemother 1998;42:23–27.

398.  Leggett J. Murine models of pneumonia using aerosol infection. San Diego, CA: Academic Press, 1999.

399.  Dudhani RV, Turnidge JD, Nation RL, et al. fAUC/MIC is the most predictive pharmacokinetic/pharmacodynamic index of colistin against Acinetobacter baumannii in murine thigh and lung infection models. J Antimicrob Chemother 2010;65:1984–1990.

400.  Harris G, Kuo Lee R, Lam CK, et al. A mouse model of Acinetobacter baumannii associated pneumonia using a clinically isolated hypervirulent strain. Antimicrob Agents Chemother 2013;57:3601–3613.

401.  Koomanachai P, Kim A, Nicolau DP. Pharmacodynamic evaluation of tigecycline against Acinetobacter baumannii in a murine pneumonia model. J Antimicrob Chemother 2009;63:982–987.

402.  Mutlu Yilmaz E, Sunbul M, Aksoy A, et al. Efficacy of tigecycline/colistin combination in a pneumonia model caused by extensively drug-resistant Acinetobacter baumannii. Int J Antimicrob Agents 2012;40:332–336.

403.  Pachon-Ibanez ME, Docobo-Perez F, Jimenez-Mejias ME, et al. Efficacy of rifampin, in monotherapy and in combinations, in an experimental murine pneumonia model caused by panresistant Acinetobacter baumannii strains. Eur J Clin Microbiol Infect Dis 2011;30:895–901.

404.  Pachon-Ibanez ME, Dacobo-Perez F, Lopez-Rojas R, et al. Efficacy of rifampin and its combinations with imipenem, sulbactam, and colistin in experimental models of infection caused by imipenem-resistant Acinetobacter baumannii. Antimicrob Agents Chemother2010;54:1165–1172.

405.  Pichardo C, Pachon-Ibanez ME, Docobo-Perez F, et al. Efficacy of tigecycline vs. imipenem in the treatment of experimental Acinetobacter baumannii murine pneumonia. Eur J Clin Microbiol Infect Dis 2010;29:527–531.

406.  Song JY, Cheong HJ, Lee J, et al. Efficacy of monotherapy and combined antibiotic therapy for carbapenem-resistant Acinetobacter baumannii pneumonia in an immunosuppressed mouse model. Int J Antimicrob Agents 2009;33:33–39.

407.  Tang HJ, Chuang YC, Ko WC, et al. Comparative evaluation of intratracheal colistimethate sodium, imipenem, and meropenem in BALB/c mice with carbapenem-resistant Acinetobacter baumannii pneumonia. Int J Infect Dis 2012;16:e34–e40.

408.  Yuan Z, Ledesma KR, Singh R, et al. Quantitative assessment of combination antimicrobial therapy against multidrug-resistant bacteria in a murine pneumonia model. J Infect Dis 2010;201:889–897.

409.  Bhalodi AA, Crandon JL, Biek D, et al. Efficacy of ceftaroline fosamil in a staphylococcal murine pneumonia model. Antimicrob Agents Chemother 2012;56:6160–6165.

410.  Crandon JL, Kuti JL, Nicolau DP. Comparative efficacies of human simulated exposures of telavancin and vancomycin against methicillin-resistant Staphylococcus aureus with a range of vancomycin MICs in a murine pneumonia model. Antimicrob Agents Chemother2010;54:5115–5119.

411.  Docobo-Perez F, Lopez-Rojas R, Dominguez-Herrera J, et al. Efficacy of linezolid versus a pharmacodynamically optimized vancomycin therapy in an experimental pneumonia model caused by methicillin-resistant Staphylococcus aureus. J Antimicrob Chemother2012;67:1961–1967.

412.  Karau MJ, Tilahun AY, Schmidt SM, et al. Linezolid is superior to vancomycin in experimental pneumonia caused by Superantigen-Producing Staphylococcus aureus in HLA class II transgenic mice. Antimicrob Agents Chemother 2012;56:5401–5405.

413.  Koomanachai P, Crandon JL, Banevicius L, et al. Pharmacodynamic profile of tigecycline against methicillin-resistant Staphylococcus aureus in an experimental pneumonia model. Antimicrob Agents Chemother 2009;53:5060–5063.

414.  Laohavaleeson S, Tessier PR, Nicolau DP. Pharmacodynamic characterization of ceftobiprole in experimental pneumonia caused by phenotypically diverse Staphylococcus aureus strains. Antimicrob Agents Chemother 2008;52:2389–2394.

415.  Lepak AJ, Marchillo K, Pichereau S, et al. Comparative pharmacodynamics of the new oxazolidinone tedizolid phosphate and linezolid in a neutropenic murine Staphylococcus aureus pneumonia model. Antimicrob Agents Chemother 2012;56:5916–5922.

416.  Reyes N, Skinner R, Kaniga K, et al. Efficacy of telavancin (TD-6424), a rapidly bactericidal lipoglycopeptide with multiple mechanisms of action, in a murine model of pneumonia induced by methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother2005;49:4344–4346.

417.  Tessier PR, Keel RA, Hagihara M, et al. Comparative in vivo efficacies of epithelial lining fluid exposures of tedizolid, linezolid, and vancomycin for methicillin-resistant Staphylococcus aureus in a mouse pneumonia model. Antimicrob Agents Chemother 2012;56:2342–2346.

418.  Aoki N, Tateda K, Kikuchi Y, et al. Efficacy of colistin combination therapy in a mouse model of pneumonia caused by multidrug-resistant Pseudomonas aeruginosa. J Antimicrob Chemother 2009;63:534–542.

419.  Bretonniere C, Jacqueline C, Caillon J, et al. Efficacy of doripenem in the treatment of Pseudomonas aeruginosa experimental pneumonia versus imipenem and meropenem. J Antimicrob Chemother 2010;65:2423–2427.

420.  Crandon JL, Nicolau DP. Human stimulated studies of aztreonam and aztreonam-avibactam to evaluate activity against challenging gram-negative organisms, including metallo-beta-lactamase producers. Antimicrob Agents Chemother 2013;57:3299–3306.

421.  Crandon JL, Schuck VJ, Banevicius MA, et al. Comparative in vitro and in vivo efficacies of human simulated doses of ceftazidime and ceftazidime-avibactam against Pseudomonas aeruginosa. Antimicrob Agents Chemother 2012;56:6137–6146.

422.  Hagihara M, Crandon JL, Urban CM, et al. KPC presence in Pseudomonas aeruginosa has minimal impact on the in vivo efficacy of carbapenem therapy. Antimicrob Agents Chemother 2013;57:1086–1088.

423.  Jacqueline C, Roquilly A, Desessard C, et al. Efficacy of ceftolozane in a murine model of Pseudomonas aeruginosa acute pneumonia: in vivo antimicrobial activity and impact on host inflammatory response. J Antimicrob Chemother 2013;68:177–183.

424.  Louie A, Fregeau C, Liu W, et al. Pharmacodynamics of levofloxacin in a murine pneumonia model of Pseudomonas aeruginosa infection: determination of epithelial lining fluid targets. Antimicrob Agents Chemother 2009;53:3325–3330.

425.  Morinaga Y, Yanagihara K, Nakamura S, et al. In vivo efficacy and pharmacokinetics of tomopenem (CS-023), a novel carbapenem, against Pseudomonas aeruginosa in a murine chronic respiratory tract infection model. J Antimicrob Chemother 2008;62:1326–1331.

426.  Sabet M, Miller CE, Nolan TG, et al. Efficacy of aerosol MP-376, a levofloxacin inhalation solution, in models of mouse lung infection due to Pseudomonas aeruginosa. Antimicrob Agents Chemother 2009;53:3923–3928.

427.  Bakker-Woudenberg IA, ten Kate MT, Goessens WH, et al. Effect of treatment duration on pharmacokinetic/pharmacodynamic indices correlating with therapeutic efficacy of ceftazidime in experimental Klebsiella pneumoniae lung infection. Antimicrob Agents Chemother2006;50:2919–2925..

428.  Docobo-Perez F, Nordmann P, Dominguez-Herrera J, et al. Efficacies of colistin and tigecycline in mice with experimental pneumonia due to NDM-1-producing strains of Klebsiella pneumoniae and Escherichia coli. Int J Antimicrob Agents 2012;39:251–254.

429.  Goessens WH, Mouton JW, ten Kate MT, et al. The therapeutic effect of tigecycline, unlike that of Ceftazidime, is not influenced by whether the Klebsiella pneumoniae strain produces extended-spectrum beta-lactamases in experimental pneumonia in rats. Antimicrob Agents Chemother 2013;57:643–646.

430.  Hilliard JJ, Melton JL, Hall L, et al. Comparative effects of carbapenems on bacterial load and host immune response in a Klebsiella pneumoniae murine pneumonia model. Antimicrob Agents Chemother 2011;55:836–544.

431.  Hirsch EB, Guo B, Chang KT, et al. Assessment of antimicrobial combinations for Klebsiella pneumoniae carbapenemase-producing K. pneumoniae. J Infect Dis 2013;207:786–793.

432.  Padilla E, Alonso D, Domenech-Sanchez A, et al. Effect of porins and plasmid-mediated AmpC beta-lactamases on the efficacy of beta-lactams in rat pneumonia caused by Klebsiella pneumoniae. Antimicrob Agents Chemother 2006;50:2285–2260.

433.  Pichardo C, del Carmen Conejo M, Bernabeu-Wittel M, et al. Activity of cefepime and carbapenems in experimental pneumonia caused by porin-deficient Klebsiella pneumoniae producing FOX-5 beta-lactamase. Clin Microbiol Infect 2005;11:31–38.

434.  Pichardo C, Rodriguez-Martinez JM, Pachon-Ibanez ME, et al. Efficacy of cefepime and imipenem in experimental murine pneumonia caused by porin-deficient Klebsiella pneumoniae producing CMY-2 beta-Lactamase. Antimicrob Agents Chemother 2005;49:3311–3316.

435.  Ambrose PG, Forest A, Craig WA, et al. Pharmacokinetics-pharmacodynamics of gatifloxacin in a lethal murine Bacillus anthracis inhalation infection model. Antimicrob Agents Chemother 2007;51:4351–4355.

436.  Deziel MR, Heine H, Louie A, et al. Effective antimicrobial regimens for use in humans for therapy of Bacillus anthracis infections and postexposure prophylaxis. Antimicrob Agents Chemother 2005;49:5099–5106.

437.  Gill SC, Rubino CM, Bassett J, et al. Pharmacokinetic-pharmacodynamic assessment of faropenem in a lethal murine Bacillus anthracis inhalation postexposure prophylaxis model. Antimicrob Agents Chemother 2010;54:1678–1683.

438.  Heine HS, Bassett J, Miller L, et al. Efficacy of oritavancin in a murine model of Bacillus anthracis spore inhalation anthrax. Antimicrob Agents Chemother 2008;52:3350–3357.

439.  Heine HS, Bassett J, Miller L, et al. Determination of antibiotic efficacy against Bacillus anthracis in a mouse aerosol challenge model. Antimicrob Agents Chemother 2007;51:1373–1379.

440.  Heine HS, Bassett J, Miller L, et al. Efficacy of Daptomycin against Bacillus anthracis in a murine model of anthrax spore inhalation. Antimicrob Agents Chemother 2010;54:4471–4473.

441.  Heine HS, Purcell BK, Bassett J, et al. Activity of dalbavancin against Bacillus anthracis in vitro and in a mouse inhalation anthrax model. Antimicrob Agents Chemother 2010;54:991–996.

442.  Kao LM, Bush K, Barnewall R, et al. Pharmacokinetic considerations and efficacy of levofloxacin in an inhalational anthrax (postexposure) rhesus monkey model. Antimicrob Agents Chemother 2006;50:3535–3542.

443.  Hansen EJ, Toews GB. Animal models for the study of noninvasive Haemophilus influenzae disease: pulmonary clearance systems. J Infect Dis 1992;165(Suppl 1):S185–S187.

444.  O’Reilly T. Relevance of animal models for chronic bacterial airway infections in humans. Am J Respir Crit Care Med 1995;151:2101–2107; discussion 2107–2108.

445.  Toews GB, Gross GN, Pierce AK. The relationship of inoculum size to lung bacterial clearance and phagocytic cell response in mice. Am Rev Respir Dis 1979;120:559–566.

446.  Johansen HK, Høiby N. Rat model of chronic Pseudomonas aeruginosa. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:517–532.

447.  Schiff JB, Small GJ, Pennington JE. Comparative activities of ciprofloxacin, ticarcillin, and tobramycin against experimental Pseudomonas aeruginosa pneumonia. Antimicrob Agents Chemother 1984;26:1–4.

448.  Hraiech S, Bregeon F, Brunel JM, et al. Antibacterial efficacy of inhaled squalamine in a rat model of chronic Pseudomonas aeruginosa pneumonia. J Antimicrob Chemother 2012;67:2452–2458.

449.  Macia MD, Borrell N, Segura M, et al. Efficacy and potential for resistance selection of antipseudomonal treatments in a mouse model of lung infection by hypermutable Pseudomonas aeruginosa. Antimicrob Agents Chemother 2006;50:975–983.

450.  Moser C, Van Gennip M, Bjarnsholt T, et al. Novel experimental Pseudomonas aeruginosa lung infection model mimicking long-term host-pathogen interactions in cystic fibrosis. APMIS 2009;117:95–107.

451.  Song Z, Wu H, Mygind P, et al. Effects of intratracheal administration of novispirin G10 on a rat model of mucoid Pseudomonas aeruginosa lung infection. Antimicrob Agents Chemother 2005;49:3868–3874.

452.  van Gennip M, Moser C, Christensen LD, et al. Augmented effect of early antibiotic treatment in mice with experimental lung infections due to sequentially adapted mucoid strains of Pseudomonas aeruginosa. J Antimicrob Chemother 2009;64:1241–1250.

453.  Beaulac C, Sachatelli S, Lagace J. Aerosolization of low phase transition temperature liposomal tobramycin as a dry powder in an animal model of chronic pulmonary infection caused by Pseudomonas aeruginosa. J Drug Target 1999;7:33–41.

454.  Cannon CL, Hogue LA, Vajravelu RK, et al. In vitro and murine efficacy and toxicity studies of nebulized SCC1, a methylated caffeine-silver(I) complex, for treatment of pulmonary infections. Antimicrob Agents Chemother 2009;53:3285–3293.

455.  Ferrari F, Lu Q, Girardi C, et al. Nebulized ceftazidime in experimental pneumonia caused by partially resistant Pseudomonas aeruginosa. Intensive Care Med 2009;35:1792–1800.

456.  Goldstein I, Wallet F, Nicolas-Robin A, et al. Lung deposition and efficiency of nebulized amikacin during Escherichia coli pneumonia in ventilated piglets. Am J Respir Crit Care Med 2002;166:1375–1381.

457.  Lu Q, Girardi C, Zhang M, et al. Nebulized and intravenous colistin in experimental pneumonia caused by Pseudomonas aeruginosa. Intensive Care Med 2010;36:1147–1155.

458.  Post JC. Direct evidence of bacterial biofilms in otitis media. Laryngoscope 2001;111:2083–2094.

459.  Post JC, Aul JJ, White GJ, et al. PCR-based detection of bacterial DNA after antimicrobial treatment is indicative of persistent, viable bacteria in the chinchilla model of otitis media. Am J Otolaryngol 1996;17:106–111.

460.  Post JC, Preston RA, Aul JJ, et al. Molecular analysis of bacterial pathogens in otitis media with effusion. JAMA 1995;273:1598–1604.

461.  Hajek DM, Yuan Z, Quartey MK, et al. Otitis media: the chinchilla model. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:389–401.

462.  Estrem SA. Bacterial otitis externa in the guinea-pig model. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:385–387.

463.  Barry B, Muffat-Joly M. Gerbil model of acute otitis media. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:375–384.

464.  Lim DJ, Hermansson A, Hellström SO, et al. Recent advances in otitis media. 3. Animal models; anatomy and pathology; pathogenesis; cell biology and genetics. Ann Otol Rhinol Laryngol Suppl 2005;194:31–41.

465.  Caye-Thomasen P, Hermansson A, Tos M, et al. Effect of penicillin on experimental acute otitis media: a histopathological study of goblet cell density, bone modelling dynamics, polyp and adhesion formation. Acta Otolaryngol [Suppl] (Stockh) 2000;543:56–57.

466.  Humphrey WR, Shattuck MH, Zielinski RJ, et al. Pharmacokinetics and efficacy of linezolid in a gerbil model of Streptococcus pneumoniae-induced acute otitis media. Antimicrob Agents Chemother 2003;47:1355–1363.

467.  Parra A, Ponte C, Cenjor C, et al. Optimal dose of amoxicillin in treatment of otitis media caused by a penicillin-resistant pneumococcus strain in the gerbil model. Antimicrob Agents Chemother 2002;46:859–862.

468.  Soriano F, Parra A, Cenjor C, et al. Role of Streptococcus pneumoniae and Haemophilus influenzae in the development of acute otitis media and otitis media with effusion in a gerbil model. J Infect Dis 2000;181:646–652.

469.  Westman E, Melhus A. The treatment of Haemophilus influenzae acute otitis media with amoxicillin protects against reinfection but not against structural changes. J Antimicrob Chemother 2002;49:141–147.

470.  Sato K, Quartey MK, Liebeler CL, et al. Timing of penicillin treatment influences the course of Streptococcus pneumoniae–induced middle ear inflammation. Antimicrob Agents Chemother 1995;39:1896–1898.

471.  Parra A, Ponte C, Cenjor C, et al. Effect of antibiotic treatment delay on therapeutic outcome of experimental acute otitis media caused by Streptococcus pneumoniae strains with different susceptibilities to amoxicillin. Antimicrob Agents Chemother 2004;48:860–866.

472.  Martínez-Marín C, Huelves L, del Prado G, et al. Effect of erythromycin treatment delay on therapeutic outcome of experimental acute otitis media caused by Streptococcus pneumoniae. J Antimicrob Chemother 2005;56:783–786.

473.  Park SN, Yeo SW. Effects of antibiotics and steroid on middle ear mucosa in rats with experimental acute otitis media. Acta Otolaryngol 2001;121:808–812.

474.  Parra A, Ponte C, Cenjor C, et al. Is it possible to achieve bacterial eradication in otitis media with effusion by empirical antibiotic high doses and concomitant administration of acetaminophen? A microbiological and pharmacological study in the gerbil model. Int J Antimicrob Agents 2003;22:508–515.

475.  del Prado G, Martínez-Marín C, Huelves L, et al. Impact of ibuprofen therapy in the outcome of experimental pneumococcal acute otitis media treated with amoxicillin or erythromycin. Pediatr Res 2006;60:555–559.

476.  Ponte C, Cenjor C, Parra A, et al. Antimicrobial treatment of an experimental otitis media caused by a beta-lactamase positive isolate of Haemophilus influenzae. J Antimicrob Chemother 1999;44:85–90.

477.  Westmann E, Lundin S, Hermansson A, et al. β-Lactamase-producing nontypeable Haemophilus influenzae fails to protect Stretococcus pneumoniae from amoxicillin during experimental acute otitis media. Antimicrob Agents Chemother 2004;48:3536–3542.

478.  Zwijnenburg PJ, van der PT, Florquin S, et al. Experimental pneumococcal meningitis in mice: a model of intranasal infection. J Infect Dis 2001;183:1143–1146.

479.  Echchannaoui H, Frei K, Schnell C, et al. Toll-like receptor 2-deficient mice are highly susceptible to Streptococcus pneumoniae meningitis because of reduced bacterial clearing and enhanced inflammation. J Infect Dis 2002;186:798–806.

480.  Gazzoni AF, Capilla J, Mayayo E, et al. Efficacy of intrathecal administration of liposomal amphotericin B combined with voriconazole in a murine model of cryptococcal meningitis. Int J Antimicrob Agents 2012;39:223–227.

481.  Mook-Kanamori BB, Rouse MS, Kang C, et al. Daptomycin in experimental murine pneumococcal meningitis. BMC Infect Dis 2009;9:50.

482.  Nau R, Wellmer A, Soto A, et al. Rifampin reduces early mortality in experimental Streptococcus pneumoniae meningitis. J Infec Dis 1999;179:1557–1560.

483.  Moxon ER, Ostrow PT. Haemophilus influenzae meningitis in infant rats: role of bacteremia in pathogenesis of age-dependent inflammatory responses in cerebrospinal fluid. J Infect Dis 1977;135:303–307.

484.  Grandgirard D, Oberson K, Bühlmann A, et al. Attenuation of cerebrospinal fluid inflammation by the nonbacteriolytic antibiotic daptomycin versus that by ceftriaxone in experimental pneumococcal meningitis. Antimicrob Agents Chemother 2010;54:1323–1326.

485.  Leib SL, Heimgartner C, Bifrare YD, et al. Dexamethasone aggravates hippocampal apoptosis and learning deficiency in pneumococcal meningitis in infant rats. Pediatr Res 2003;54:353–357.

486.  Klein M, Koedel U, Pfister HW, et al. Meningitis-associated hearing loss: protection by adjunctive antioxidant therapy. Ann Neurol 2003;54:451–458.

487.  Barichello T, Gonçalves JCN, Generoso JJS, et al. Attenuation of cognitive impairment by the nonbacteriolytic antibiotic daptomycin in Wistar rats submitted to pneumococcal meningitis. BMC Neurosci 2013;14:42.

488.  Barichello T, Silva GZ, Batista AL, et al. Early antibiotic administration prevents cognitive impairment induced by meningitis in rats. Neurosci Lett 2009;465:71–73.

489.  Brandt CT, Cayé-Thomasen P, Lund SP, et al. Hearing loss and cochlear damage in experimental pneumococcal meningitis, with special reference to the role of neutrophil granulocytes. Neurobiol Dis 2006;23:300–311.

490.  Dupuis A, Limosin A, Paquereau J, et al. Pharmacokinetic-pharmacodynamic modeling of electroencephalogram effect of imipenem in rats with acute renal failure. Antimicrob Agents Chemother 2001;45:3607–3609.

491.  Vogel U, Frosch M. Infant rat model of acute meningitis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:619–626.

492.  Townsend GC, Scheld WM. Adult rat model of bacterial meningitis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:631–638.

493.  Force E, Taberner F, Cabellos C, et al. Experimental study of meropenem in the therapy of cephalosporin-susceptible and -resistant pneumococcal meningitis. Eur J Clin Microbiol Infect Dis 2008;27:685–690.

494.  Cottagnoud P, Pfister M, Cottagnoud M, et al. Activities of ertapenem, a new long-acting carbapenem, against penicillin-sensitive or -resistant pneumococci in experimental meningitis. Antimicrob Agents Chemother 2003;47:1943–1947.

495.  Østergaard C, Benfield T, Gesser B, et al. Pretreatment with granulocyte colony–stimulating factor attenuates the inflammatory response but not the bacterial load in cerebrospinal fluid during experimental pneumococcal meningitis in rabbits. Infect Immun 1999;67:3430–3436.

496.  Kartalija M, Kaye K, Tureen JH, et al. Treatment of experimental cryptococcal meningitis with fluconazole: impact of dose and addition of flucytosine on mycologic and pathophysiologic outcome. J Infect Dis 1996;173:1216–1221.

497.  Tureen J, Tuomanen E. Rabbit model of bacterial meningitis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:631–638.

498.  Spector R, Lorenzo AV. Inhibition of penicillin transport from cerebral spinal fluid after intracisternal inoculation of bacteria. J Clin Invest 1974;54:316–325.

499.  Scheld WM, Brown RS, Sande MA. Comparison of netilmicin with gentamicin in the therapy of experimental Escherichia coli meningitis. Antimicrob Agents Chemother 1978;13:899–904.

500.  McCracken GH, Nelson JD, Grimm L. Pharmacokinetics and bacteriological efficacy of cefoperazone, cefuroxime, ceftriaxone, and moxalactam in experimental Streptococcus pneumoniae and Haemophilus influenzae meningitis. Antimicrob Agents Chemother 1982;21:262–267.

501.  Bardal-Ozcem S, Turhan T, Sipahi OR, et al. Daptomycin versus vancomycin in treatment of methicillin-resistant Staphylococcus aureus meningitis in an experimental rabbit model. Antimicrob Agents Chemother 2013;57:1556–1558.

502.  Bardak-Ozcem S, Turhan T, Sipahi OR, et al. Daptomycin versus vancomycin in treatment of methicillin-resistant Staphylococcus aureus meningitis in an experimental rabbit model. Antimicrob Agents Chemother 2013;57:1556–1558.

503.  Egermann U, Stanga Z, Ramin A, et al. Combination of daptomycin plus ceftriaxone is more active than vancomycin plus ceftriaxone in experimental meningitis after addition of dexamethasone. Antimicrob Agents Chemother 2009;53:3030–3033.

504.  Sipahi OR, Turhan T, Pullukcu H, et al. Moxifloxacin versus ampicillin + gentamicin in the therapy of experimental Listeria monocytogenes meningitis. J Antimicrob Chemother 2008;61;670–673.

505.  Stucki A, Cottagnoud M, Acosta F, et al. Efficacy of doripenem against Escherichia coli and Klebsiella pneumoniae in experimental meningitis. J Antimicrob Chemother 2012;67:661–665.

506.  Ernst JD, Sande MA. Selected examples of failure of in vitro testing to predict in vivo response to antibiotics. In: Sabath LD, ed. Action of antibiotics in patients. Bern, Switzerland: Hans Huber Publishers, 1982:68–73.

507.  Nathan BR, Scheld WM. The efficacy of trovafloxacin versus ceftriaxone in the treatment of experimental brain abscess/cerebritis in the rat. Life Sci 2003;73:1773–1782.

508.  Bayer AS, Greenberg DP, Yih J. Correlates of therapeutic efficacy in experimental methicillin-resistant Staphylococcus aureus endocarditis. Chemotherapy 1988;34:46–55.

509.  Braddour LM. Virulence factors among Gram-positive bacteria in experimental endocarditis. Infect Immun 1994;62:2143–2148.

510.  Carbon C. Experimental endocarditis: a review of its relevance to human endocarditis. J Antimicrob Chemother 1993;31(Suppl D):71–86.

511.  Eliopoulos GM, Thauvin-Eliopoulos C, Moellering RC. Contribution of animal models in the search for effective therapy for endocarditis due to enterococci with high-level resistance to gentamicin. Clin Infect Dis 1992;15:58–62.

512.  Gutschik E. The Enterococcus endocarditis model in experimental animals and its relevance to human infections. J Antimicrob Chemother 1993;31(Suppl D):87–96.

513.  Lefort A, Fantin B. Rabbit model of bacterial endocarditis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:611–618.

514.  Santoro J, Levison ME. Rat model of experimental endocarditis. Infect Immun 1978;19:915–918.

515.  Le T, Bayer AS. Combination antibiotic therapy for infective endocarditis. Clin Infect Dis 2003;36:615–621.

516.  Batard E, Jacqueline C, Boutoille D, et al. Combination of quinupristin-dalfopristin and gentamicin against methicillin-resistant Staphylococcus aureus: experimental rabbit endocarditis study. Antimicrob Agents Chemother 2002;46:2174–2178.

517.  Hershberger E, Coyle EA, Kaatz GW, et al. Comparison of a rabbit model of bacterial endocarditis and an in vitro infection model with simulated endocardial vegetations. Antimicrob Agents Chemother 2000;44:1921–1924.

518.  Tsitsika A, Pefanis A, Perdikaris GS, et al. Single-oral-dose azithromycin prophylaxis against experimental streptococcal or staphylococcal aortic valve endocarditis. Antimicrob Agents Chemother 2000;44:1754–1756.

519.  Katsarolis I, Pefanis A, Iliopoulos D, et al. Successful trovafloxacin prophylaxis against experimental streptococcal aortic valve endocarditis. Antimicrob Agents Chemother 2000;44:2564–2566.

520.  Siaperas P, Pefanis A, Iliopoulos D, et al. Evidence of less severe aortic valve destruction after treatment of experimental staphylococcal endocarditis with vancomycin and dexamethasone. Antimicrob Agents Chemother 2001;45:3531–3537.

521.  Galani L, Pefanis A, Sakka V, et al. Successful treatment with moxifloxacin of experimental aortic valve endocarditis due to methicillin-resistant Staphylococcus aureus (MRSA). Int J Antimicrob Agents 2009;33:65–69.

522.  Jacqueline C, Amador G, Batard E et al. Comparison of ceftaroline fosamil, daptomycin and tigecycline in an experimental rabbit endocarditis model caused by methicillin-susceptible, methicillin-resistant and glycopeptide-intermediate Staphylococcus aureus. J Antimicrob Chemother 2011;66:863–866.

523.  Jacqueline C, Caillon J, Grossi O, et al. In vitro and in vivo assessment of linezolid combined with ertapenem: a highly synergistic combination against methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 2006;50:2547–2549.

524.  Miro JM, Garcia-de-la-Maria C, Armero Y, et al. Addition of gentamicin or rifampin does not enhance the effectiveness of daptomycin in treatment of experimental endocarditis due to methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 2009;53:4172–4177.

525.  Miró JM, García-de-la-Mària C, Armero Y, et al. Efficacy of telavancin in the treatment of experimental endocarditis due to glycopeptide-intermediate Staphylococcus aureus. Antimicrob Agents Chemother 2007;51:2373–2377.

526.  Sacar, M, Sacar S, Cevahir N, et al. Comparison of antimicrobial agents as therapy for experimental endocarditis caused by methicillin-resistant Staphylococcus aureus. Tex Heart Inst J 2010;37:400–404.

527.  Seidl K, Chen L, Bayer A, et al. Relationship of agr expression and function with virulence and vancomycin treatment outcomes in experimental endocarditis due to methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 2011;55:5631–5639.

528.  Tattevin P, Basuino L, Bauer D, et al. Evaluation of ceftobiprole in a rabbit model of aortic valve endocarditis due to methicillin-resistant and vancomycin-intermediate Staphylococcus aureus. Antimicrob Agents Chemother 2010;54:610–613.

529.  Tattevin P, Saleh-Mghir A, Davido B, et al. Comparison of six generic vancomycin products for treatment of methicillin-resistant Staphylococcus aureus experimental endocarditis in rabbits. Antimicrob Agents Chemother 2013;57:1157–1162.

530.  Tsaganos T, Skiadas I, Koutoukas P, et al. Efficacy and pharmacodynamics of linezolid, alone and in combination with rifampicin, in an experimental model of methicillin-resistant Staphylococcus aureus endocarditis. J Antimicrob Chemother 2008;62:381–383.

531.  Xiong YQ, Hady WA, Bayer AS, et al. Telavancin in therapy of experimental aortic valve endocarditis in rabbits due to daptomycin-nonsusceptible methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 2012;56:5528–5533.

532.  Entenza JM, Veloso TR, Vouillamoz J, et al. In vivo synergism of ceftobiprole and vancomycin against experimental endocarditis due to vancomycin-intermediate Staphylococcus aureus. Antimicrob Agents Chemother 2011;55:3977–3984.

533.  Entenza JM, Vouillamoz J, Glauser MP, et al. Efficacy of garenoxacin in treatment of experimental endocarditis due to Staphylococcus aureus or viridans group streptococci. Antimicrob Agents Chemother 2004;48:86–92.

534.  Marco F, de la Mària CG, Armero Y, et al. Daptomycin is effective in treatment of experimental endocarditis due to methicillin-resistant and glycopeptide-intermediate Staphylococcus aureus. Antimicrob Agents Chemother 2008;52:2538–2543.

535.  Fox PM, Lampen RJ, Stumpf KS, et al. Successful therapy of experimental endocarditis caused by vancomycin-resistant Staphylococcus aureus with a combination of vancomycin and beta-lactam antibiotics. Antimicrob Agents Chemother 2006;50:2951–2956.

537.  Yang S-J, Xiong YQ, Boyle-Vavra S, et al. Daptomycin-oxacillin combinations in treatment of experimental endocarditis caused by daptomycin-nonsusceptible strains of methicillin-resistant Staphylococcus aureus with evolving oxacillin susceptibility (the “seesaw effect”). Antimicrob Agents Chemother 2010;54:3161–3169.

538.  Dubé L, Caillon J, Jacqueline C, et al. The optimal aminoglycoside and its dosage for the treatment of severe Enterococcus faecalis infection. An experimental study in the rabbit endocarditis model. Eur J Clin Microbiol Infect Dis 2012;11:2545–2547.

539.  Dubé L, Caillon J, Gras-Le Guen C, et al. Simulation of human gentamicin pharmacokinetics in an experimental Enterococcus faecalis endocarditis model. Antimicrob Agents Chemother 2003;47:3663–3666.

540.  Gavaldá J, Onrubia PL, Gómez MTM, et al. Efficacy of ampicillin combined with ceftriaxone and gentamicin in the treatment of experimental endocarditis due to Enterococcus faecalis with no high-level resistance to aminoglycosides. J Antimicrob Chemother 2003;52:514–517.

541.  Pavleas J, Skiada A, Daikos GL, et al. Efficacy of teicoplanin, administered in two different regimens, in the treatment of experimental endocarditis due to Enterococcus faecalis. J Chemother 2008;20:208–212.

542.  Vouillamoz J, Moreillon P, Giddey M, et al. Efficacy of daptomycin in the treatment of experimental endocarditis due to susceptible and multidrug-resistant enterococci. J Antimicrob Chemother 2006;58:1208–1214.

543.  Boutoille D, Jacqueline C, Le Mabecque V, et al. In vivo impact of the MexAB-OprM efflux system on beta-lactam efficacy in an experimental model of Pseudomonas aeruginosa infection. Int J Antimicrob Agents 2009;33:417–420.

544.  Navas D, Caillon J, Gras-Le Guen C, et al. Comparison of in vivo intrinsic activity of cefepime and imipenem in a Pseudomonas aeruginosa rabbit endocarditis model: effect of combination with tobramycin simulating human serum pharmacokinetics. J Antimicrob Chemother 2004;54:767–771.

545.  Rodríguez-Hernández M-J, Jiménez-Mejias ME, Pichardo C, et al. Colistin efficacy in an experimental model of Acinetobacter baumannii endocarditis. Clin Microb Infect 2004;10:581–584.

546.  Rank RG, Whittum-Hudson JA. Animal models for ocular infections. Methods Enzymol 1994;235:69–83.

547.  Motschmann M, Behrens Baumann W. Rabbit model of bacterial conjunctivitis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:353–359.

548.  Kernacki KA, Hobden JA, Hazlett LD. Murine model of bacterial keratitis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:361–366.

549.  Evans D, Kuo T, Kwong M, et al. Pseudomonas aeruginosa strains with lipopolysaccharide defects exhibit reduced intracellular viability after invasion of corneal epithelial cells. Exp Eye Res 2002;75:635–643.

550.  Lee EJ, Truong TN, Mendoza MN, et al. A comparison of invasive and cytotoxic Pseudomonas aeruginosa strain-induced corneal disease responses to therapeutics. Curr Eye Res 2003;27:289–299.

551.  Romanowski EG, Mah FS, Yates KA, et al. The successful treatment of gatifloxacin-resistant Staphylococcus aureus keratitis with Zymar (gatifloxacin 0.3%) in a NZW rabbit model. Am J Ophthalmol 2005;139:867–877.

552.  Oguz H, Ozbilge H, Oguz E, et al. Effectiveness of topical taurolidine versus ciprofloxacin, ofloxacin, and fortified cefazolin in a rabbit Staphylococcus aureus keratitis model. Curr Eye Res 2005;30:155–161.

553.  Wu X, Jiang H, Xu Y, et al. Efficacy of gemifloxacin for the treatment of experimental Staphylococcus aureus keratitis. J Ocul Pharmacol Ther 2012;28:420–427.

554.  Wu XG, Xin M, Chen H, et al. Novel mucoadhesive polysaccharide isolated from Bletilla striata improves the intraocular penetration and efficacy of levofloxacin in the topical treatment of experimental bacterial keratitis. J Pharm Pharmacol 2010;62:1152–1157.

555.  Romanowski EG, Mah FS, Kowalski RP, et al. Benzalkonium chloride enhances the antibacterial efficacy of gatifloxacin in an experimental rabbit model of intrastromal keratitis. J Ocul Pharmacl Ther 2008;24:380–384.

556.  Sanders ME, Norcross EW, Moore QC 3rd, et al. Efficacy of besifloxacin in a rabbit model of methicillin-resistant Staphylococcus aureus keratitis. Cornea 2009;28:1055–1060.

557.  Eguchi H, Shiota H, Oguro S, et al. The inhibitory effect of vancomycin ointment on the manifestation of MRSA keratitis in rabbits. J Infect Chemother 2009;15:279–283.

558.  Balzli CL, McCormick CC, Caballero AR, et al. The effectiveness of an improved combination therapy for experimental Staphylococcus aureus keratitis. Adv Ther 2010;27:933–940.

559.  Aliprandis E, Ciralsky J, Lai H, et al. Comparative efficacy of topical moxifloxacin versus ciprofloxacin and vancomycin in the treatment of P. aeruginosa and ciprofloxacin-resistant MRSA keratitis in rabbits. Cornea 2005;24:201–205.

560.  Bu P, Riske PS, Zaya NE, et al. A comparison of topical chlorhexidine, ciprofloxacin, and fortified tobramycin/cefazolin in rabbit models of Staphylococcus and Pseudomonas keratitis. J Ocul Pharmacol Ther 2007;23:213–220.

561.  McCormick C, Caballero A, Tang A, et al. Effectiveness of a new tobramycin (0.3%) and dexamethasone (0.05%) formulation in the treatment of experimental Pseudomonas keratitis. Curr Med Res Opin 2008;24:1569–1575.

562.  Mah FS, Romanowski EG, Kowalski RP, et al. Zymar (Gatifloxacin 0.3%) shows excellent Gram-negative activity against Serratia marcescens and Pseudomonas aeruginosa in a New Zealand White rabbit keratitis model. Cornea 2007;26:585–588.

563.  Thibodeaux BA, Dajcs JJ, Caballero AR, et al. Quantitative comparison of fluoroquinolone therapies of experimental gram-negative bacterial keratitis. Curr Eye Res 2004;28:337–342.

564.  Frucht-Pery J, Raiskup F, Mechoulam H, et al. Iontophoretic treatment of experimental pseudomonas keratitis in rabbit eyes using gentamicin-loaded hydrogels. Cornea 2006;25:1182–1186.

565.  Hyon JY, Joo MJ, Hose S, et al. Comparative efficacy of topical gatifloxacin with ciprofloxacin, amikacin, and clarithromycin in the treatment of experimental Mycobacterium chelonae keratitis. Arch Ophthalmol 2004;122:1166–1169.

566.  Sarayba MA, Shamie N, Reiser BJ, et al. Fluoroquinolone therapy in Mycobacterium chelonae keratitis after lamellar keratectomy. J Cataract Refract Surg 2005;31:1396–1402.

567.  Ozturk F, Yavas GF, Kusbeci T, et al. Efficacy of topical caspofungin in experimental fusarium keratitis. Cornea 2007;26:726–728.

568.  Yavas GF, Ozturk F, Kusbeci T, et al. Antifungal efficacy of voriconazole, itraconazole and amphotericin b in experimental fusarium solani keratitis. Graefes Arch Clin Exp Ophthalmol 2008;246:275–279.

569.  Polat ZA, Vural A. Effect of combined chlorhexidine gluconate and neosporin on experimental keratitis with two pathogenic strains of Acanthamoeba. Parasitol Res 2012;110:1945–1950.

570.  Maylath FR, Leopold IH. Study of experimental intraocular infection I. The recoverability of organisms inoculated into ocular tissues and fluids. II. The influence of antibiotics and cortisone, alone and combined, on intraocular growth of these organisms. Am J Ophthalmol1955;40:86–101.

571.  Giese MJ, Rayner SA, Fardin B, et al. Mitigation of neutrophil infiltration in a rat model of early Staphylococcus aureus endophthalmitis. Invest Ophthalmol Vis Sci 2003;44:3077–3082.

572.  Kim IT, Park SK, Lim JH. Inflammatory response in experimental Staphylococcus and Pseudomonas endophthalmitis. Ophthalmologica 1999;213:305–310.

573.  Ravindranath RM, Hasan SA, Mondino BJ. Immunopathologic features of Staphylococcus epidermidis–induced endophthalmitis in the rat. Curr Eye Res 1997;16:1036–1043.

574.  Yoshizumi MO, Kashani A, Palmer J, et al. High dose intramuscular methylprednisolone in experimental Staphylococcus aureus endophthalmitis. J Ocul Pharmacol Ther 1999;15:91–96.

575.  Callegan MC, Hill JM, Insler MS, et al. Methicillin-resistant Staphylococcus aureus keratitis in the rabbit: therapy with ciprofloxacin, vancomycin and cefazolin. Curr Eye Res 1992;11:1111–1119.

576.  Yoshizumi MO, Lee GC, Equi RA, et al. Timing of dexamethasone treatment in experimental Staphylococcus aureus endophthalmitis. Retina 1998;18:130–135.

577.  Smith MA, Sorenson JA, D’Aversa G, et al. Treatment of experimental methicillin-resistant Staphylococcus epidermidis endophthalmitis with intravitreal vancomycin and intravitreal dexamethasone. J Infect Dis 1997;175:462–466.

578.  Park SS, Samiy N, Ruoff K, et al. Effect of intravitreal dexamethasone in treatment of pneumococcal endophthalmitis in rabbits. Arch Ophthalmol 1995;113:1324–1329.

579.  Yildirim O, Oz O, Aslan G, et al. The efficacy of intravitreal levofloxacin and intravitreal dexamethasone in experimental Staphylococcus epidermidis endophthalmitis. Ophthalmic Res 2002;34:349–356.

580.  De Kaspar HM, Ta CN, Engelbert M, et al. Effects of intravitreal corticosteroid in the treatment of Staphylococcus aureus-induced experimental endophthalmitis. Retina 2008;28:326–332.

581.  Ermis SS, Cetinkaya Z, Kiyici H, et al. Effects of intravitreal moxifloxacin and dexamethasone in experimental Staphylococcus aureus endophthalmitis. Curr Eye Res 2007;32:337–344.

582.  Liu F, Kwok AK, Cheung BM. The efficacy of intravitreal vancomycin and dexamethasone in the treatment of experimental bacillus cereus endophthalmitis. Curr Eye Res 2008;33:761–768.

583.  Fukuda M, Sasaki K. General purpose antimicrobial ophthalmic solutions evaluated using new pharmacokinetic parameter of maximum drug concentration in aqueous. Jpn J Ophthalmol 2002;46:384–390.

584.  Xu J, Heys JJ, Barocas VH, et al. Permeability and diffusion in vitreous humor: implications for drug delivery. Pharm Res 2000;16:664–669.

585.  Bronner S, Jehl F, Peter JD, et al. Moxifloxacin efficacy and vitreous penetration in a rabbit model of Staphylococcus aureus endophthalmitis and effect on gene expression of leucotoxins and virulence regulator factors. Antimicrob Agents Chemother 2003;47:1621–1629.

586.  Waga J, Nilsson-Ehle I, Ljungberg B, et al. Microdialysis for pharmacokinetic studies of ceftazidime in rabbit vitreous. J Ocul Pharmacol Ther 1999;15:455–463.

587.  Fernandez de Gatta MM, Fruns I, Calvo MV, et al. Influence of pharmacokinetic model on vancomycin peak concentration targets. Ther Drug Monit 1996;18:145–148.

588.  Waga J, Ehinger B. Intravitreal concentrations of some drugs administered with microdialysis. Acta Ophthalmol Scand 1997;75:36–40.

589.  Ozturk F, Kortunay S, Kurt E, et al. Effects of trauma and infection on ciprofloxacin levels in the vitreous cavity. Retina 1999;19:127–130.

590.  Ozturk F, Kurt E, Inan UU, et al. Penetration of topical and oral ofloxacin into the aqueous and vitreous humor of inflamed rabbit eyes. Int J Pharm 2000;204:91–95.

591.  Fukuda M, Hanazome I, Sasaki K. The intraocular dynamics of vancomycin hydrochloride ophthalmic ointment (TN-011) in rabbits. J Infect Chemother 2003;9:93–96.

592.  Ng EW, Samiy N, Ruoff KL, et al. Treatment of experimental Staphylococcus epidermidis endophthalmitis with oral trovafloxacin. Am J Ophthalmol 1998;126:278–287.

593.  el-Massry A, Meredith TA, Aguilar HE, et al. Aminoglycoside levels in the rabbit vitreous cavity after intravenous administration. J Ophthalmol 1996;122:684–689.

594.  Marrakchi-Benjaafar S, Cochereau I, Pocidalo JJ, et al. Systemic prophylaxis of experimental staphylococcal endophthalmitis: comparative efficacy of sparfloxacin, pefloxacin, imipenem, vancomycin, and amikacin. J Infect Dis 1995;172:1312–1316.

595.  Gupta SK, Dhingra N, Velpandian T, et al. Efficacy of fluconazole and liposome entrapped fluconazole for C. albicans induced experimental mycotic endophthalmitis in rabbit eyes. Acta Ophthalmol Scand 2000;78:448–450.

596.  Gupta SK, Velpandian T, Dhingra N, et al. Intravitreal pharmacokinetics of plain and liposome-entrapped fluconazole in rabbit eyes. J Ocul Pharmacol Ther 2000;16:511–518.

597.  Park SS, D’Amico DJ, Paton B, et al. Treatment of exogenous Candida endophthlamitis in rabbits with oral fluconazole. Antimicrob Agents Chemother 1995;39:958–963.

598.  Choi S, Hahn TW, Osterhout G, et al. Comparative intravitreal antibiotic therapy for experimental Enterococcus faecalis endophthalmitis. Arch Ophthalmol 1996;114:61–65.

599.  Wada T, Kozai S, Tajika T, et al. Prophylactic efficacy of ophthalmic quinolones in expeimental endophthalmitis in rabbits. J Ocul Pharmacol Ther 2008;24:278–289.

600.  Sakalar YB, Ozekinci S, Celen MK.Treatment of experimental Bacillus cereus endophthalmitis using intravitreal moxifloxacin with or without dexamethasone. J Ocul Pharmacol Ther 2011;27:593–598.

601.  Saleh M, Lefèvre S, Acar N, et al. Efficacy of intravitreal administrations of linezolid in an experimental model of S. aureus-related endophthalmitis. Invest Ophthalmol Vis Sci 2012;53:4832–4841.

602.  Engelbert M, Miño de Kaspar H, Thiel M, et al. Intravitreal vancomycin and amikacin versus intravenous imipenem in the treatment of experimental Staphylococcus aureus endophthalmitis. Graefes Arch Clin Exp Ophthalmol 2004;242:313–320.

603.  Engelbert M, Miño de Kaspar H, Mette M, et al. Intravenous treatment of experimental Staphylococcus aureus endophthalmitis: imipenem versus the combination of ceftazidime and amikacin. Graefes Arch Clin Exp Ophthalmol 2003;241:1029–1036.

604.  Ferrer C, Rodríguez A, Abad JL, et al. Bactericidal effect of intravitreal levofloxacin in an experimental model of endophthalmitis. Br J Ophthalmol 2008;92:672–682.

605.  Ozkiriş A, Evereklioglu C, Akgün H, et al. A comparison of intravitreal piperacillin/tazobactam with ceftazidime in experimental Pseudomonas aeruginosa endophthalmitis. Exp Eye Res 2005;80:361–367.

606.  Horozoglu F, Metan G, Sever O, et al. Intravitreal tigecycline treatment in experimental Acinetobacter baumannii endophthalmitis. J Chemother 2012;24:101–106.

607.  Cremieux AC, Carbon C. Experimental models of bone and prosthetic joint infections. Clin Infect Dis 1997;25:1295–1302.

608.  Norden CW. Experimental osteomyelitis. I. A description of the model. J Infect Dis 1970;122:410–418.

609.  Norden CW. Experimental osteomyelitis. II. Therapeutic trials and measurement of antibiotic levels in bone. J Infect Dis 1971;124:565–571.

610.  Lazzarini L, Overgaard KA, Conti E, et al. Experimental osteomyelitis: what have we learned from animal studies about the systemic treatment of osteomyelitis? J Chemother 2006;18:451–460.

611.  Yin LY, Lazzarini L, Li F, et al. Comparative evaluation of tigecycline and vancomycin, with and without rifampicin, in the treatment of methicillin-resistant Staphylococcus aureus experimental osteomyelitis in a rabbit model. J Antimicrob Chemother 2005;55:995–1002.

612.  O’Reilly T, Mader JT. Rat model of osteomyelitis of the tiba. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:561–576.

613.  Rissing JP, Buxton TB, Weinstein RS, et al. Model of experimental osteomyelitis in rats. Infect Immun 1985;47:581–586.

614.  Dworkin R, Modin G, Kunz S, et al. Comparative efficacies of ciprofloxacin, pefloxacin, and vancomycin in combination with rifampin in a rat model of methicillin-resistant Staphylococcus aureus chronic osteomyelitis. Antimicrob Agents Chemother 1990;34:1014–1016.

615.  Poeppl W, Tobudic S, Lingscheid T, et al. Efficacy of fosfomycin in experimental osteomyelitis due to methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 2011;55:931–933.

616.  Kandemir O, Oztuna V, Colak M, et al. Comparison of the efficacy of tigecycline and teicoplanin in an experimental methicillin-resistant Staphylococcus aureus osteomyelitis model. J Chemother 2008;20:53–57.

617.  Yagmurlu MF, Korkusuz F, Gursel I, et al. Sulbactam-cefoperazone polyhydroxybutyrate-co-hydroxyvalerate (PHBV) local antibiotic delivery system: in vivo effectiveness and biocompatibility in the treatment of implant-related experimental osteomyelitis. J Biomed Mater Res 1999;46:494–503.

618.  Korkusuz F, Korkusuz P, Eksioglu F, et al. In vivo response to biodegradable controlled antibiotic release systems. J Biomed Mater Res 2001;55:217–228.

619.  Nelson CL, McLaren SG, Skinner RA, et al. The treatment of experimental osteomyelitis by surgical debridement and the implantation of calcium sulfate tobramycin pellets. J Orthop Res 2002;20:643–647.

620.  Mader JT, Stevens CM, Stevens JH, et al. Treatment of experimental osteomyelitis with a fibrin sealant antibiotic implant. Clin Orthop 2002;403:58–72.

621.  Shirtliff ME, Calhoun JH, Mader JT. Experimental osteomyelitis treatment with antibiotic-impregnated hydroxyapatite. Clin Orthop 2002;401:239–247.

622.  Mimoz O, Leotard S, Jacolot A, et al. Efficacies of imipenem, meropenem, cefepime, and ceftazidime in rats with experimental pneumonia due to a carbapenem-hydrolyzing β-lactamase–producing strain of Enterobacter cloacae. Antimicrob Agents Chemother 2000;44:885–890.

623.  Subasi M, Kapukaya A, Kesemenli C, et al. Effect of granulocyte-macrophage colony–stimulating factor on treatment of acute osteomyelitis: an experimental investigation in rats. Arch Orthop Trauma Surg 2001;121:170–173.

624.  Xie Z, Liu X, Jia W, et al. Treatment of osteomyelitis and repair of bone defect by degradable bioactive borate glass releasing vancomycin. J Controlled Release 2009;139:118–126.

625.  Stewart S, Barr S, Engiles J, et al. Vancomycin-modified implant surface inhibits biofilm formation and supports bone-healing in an infected osteotomy model in sheep: a proof-of-concept study. J Bone Joint Surg Am 2012;94:1406–1415.

626.  Rouse MS, Piper KE, Jacobson M, et al. Daptomycin treatment of Staphylococcus aureus experimental chronic osteomyelitis. J Antimicrob Chemother 2006;57:301–305.

627.  Moskowitz JS, Blaisse MR, Samuel RE, et al. The effectiveness of the controlled release of gentamicin from polyelectrolyte multilayers in the treatment of Staphylococcus aureus infection in a rabbit bone model. Biomaterials 2010;31:6019–6030.

628.  Kanellakopoulou K, Thivaios GC, Kolia M, et al. First identification of Pseudomonas aeruginosa isolates producing a KPC-type carbapenem-hydrolyzing β-lactamase. Antimicrob Agents Chemother 2007;51:1553–1555.

629.  Kanellakopoulou K, Galanopoulos I, Soranoglou V, et al. Treatment of experimental osteomyelitis caused by methicillin-resistant Staphylococcus aureus with a synthetic carrier of calcium sulphate (Stimulan) releasing moxifloxacin. Int J Antimicrob Agents 2009;33:354–359.

630.  Giavaresi G, Borsari V, Fini M, et al. Preliminary investigations on a new gentamicin and vancomycin-coated PMMA nail for the treatment of bone and intramedullary infections: an experimental study in the rabbit. J Orthop Res 2008;26:785–792.

631.  Gaudin A, Jacqueline C, Gautier H, et al. A delivery system of linezolid to enhance the MRSA osteomyelitis prognosis: in vivo experimental assessment. Eur J Clin Microbiol Infect Dis 2013;32:195–198.

632.  Nijhof MW, Fleer A, Hardus K, et al. Tobramycin-containing bone cement and systemic cefazolin in a one-stage revision: treatment of infection in a rabbit model. J Biomed Mater Res 2001;58:747–753.

633.  Del Pozo JL, Rouse MS, Euba G, et al. The electricidal effect is active in an experimental model of Staphylococcus epidermidis chronic foreign body osteomyelitis. Antimicrob Agents Chemother 2009;53:4064–4068.

634.  Ozturan KE, Yucel I, Kocoglu E, et al. Efficacy of moxifloxacin compared to teicoplanin in the treatment of implant-related chronic osteomyelitis in rats. J Orthop Res 2010;28:1368–1372.

635.  Vergidis P, Rouse MS, Euba G, et al. (2011). Treatment with linezolid or vancomycin in combination with rifampin is effective in an animal model of methicillin-resistant Staphylococcus aureus foreign body osteomyelitis. Antimicrob Agents Chemother 2011;55:1182–1186.

636.  Schroeder K, Simank H, Lorenz H, et al. Implant stability in the treatment of MRSA bone implant infections with linezolid versus vancomycin in a rabbit model. J Orthop Res 2012;30:190–195.

637.  Cohen Y, Perronne C, Lazard T, et al. Use of normal C57BL/6 mice with established Mycobacterium avium infections as an alternative model for evaluation of antibiotic activity. Antimicrob Agents Chemother 1995;39:735–738.

638.  Yangco BG, Lackman-Smith C, Espinoza CG, et al. The hamster model of chronic Mycobacterium avium complex infection. J Infect Dis 1989;159:556–561.

639.  Brown ST, Edwards FF, Bernhard EM, et al. Azithromycin, rifabutin, and rifapentine for treatment and prophylaxis of Mycobacterium avium complex in rats treated with cyclosporine. Antimicrob Agents Chemother 1993;37:398–402.

640.  Cynamon MH, DeStefano MS. Beige mouse model of disseminated Mycobacterium avium complex infection. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:321–330.

641.  Gangadharam PRJ, Ashtekar DR, Flasher DL, et al. Therapy of Mycobacterium avium complex infections in beige mice with streptomycin encapsulated in sterically stabilized liposomes. Antimicrob Agents Chemother 1995;39:725–730.

642.  Gangadharam PR. Beige mouse model for Mycobacterium avium complex disease. Antimicrob Agents Chemother 1995;30:1647–1654.

643.  Wu HS, Kolonoski P, Chang YY, et al. Invasion of the brain and chronic central nervous system infection after systemic Mycobacterium avium complex infection in mice. Infect Immun 2000;68:2979–2984.

644.  Gomes MS, Dom G, Pedrosa J, et al. Effects of iron deprivation on Mycobacterium avium growth. Tuberc Lung Dis 1999;79:321–328.

645.  Siso JP, Yao Y, Klemper CA, et al. Treatment of Mycobacterium avium complex infection: does the beige mouse model predict therapeutic outcome in humans? J Infect Dis 1996;173:750–753.

646.  Bermudez LE, Inderlied CB, Kolonoski P, et al. Telithromycin is active against Mycobacterium avium in mice despite lacking significant activity in standard in vitro and macrophage assays and is associated with low frequency of resistance during treatment. Antimicrob Agents Chemother2001;45:2210–2214.

647.  Goncalves AS, Appelberg R. Effects of recombinant granulocyte-colony stimulating factor administration during Mycobacterium avium infection in mice. Clin Exp Immunol 2001;124:239–247.

648.  Fattorini L, Xiao Y, Mattei M, et al. Activities of eighteen antimicrobial regimens against Mycobacterium avium infection in beige mice. Microb Drug Resist 1999;5:227–233.

649.  Bermudez LE, Inderlied CB, Kolonoski P, et al. Activity of moxifloxacin by itself and in combination with ethambutol, rifabutin, and azithromycin in vitro and in vivo against Mycobacterium avium. Antimicrob Agents Chemother 2001;45:217–222.

650.  Nuermberger E. Using animal models to develop new treatments for tuberculosis. Semin Respir Crit Care Med 2008;29:542–551.

651.  Franzblau SG, DeGroote MA, Cho SH, et al. Comprehensive analysis of methods used for the evaluation of compounds against Mycobacterium tuberculosis. Tuberculosis 2012;92:453–488.

652.  De Groote MA, Gilliland JC, Wells CL, et al. Comparative studies evaluating mouse models used for efficacy testing of experimental drugs against Mycobacterium tuberculosis. Antimicrob Agents Chemother 2011;55:1237–1247.

653.  Johnson AP, Tuffrey M, Taylor-Robinson D. Resistance of mice to genital infection with Neisseria gonorrhoeae. J Med Microbiol 1989;30:33–36.

654.  Corbeil LB, Wunderlich AC, Corbeil RR, et al. Disseminated gonococcal infection in mice. Infect Immun 1979;26:984–990.

655.  Arka RJ, Balows A. Animal models of experimental gonococcal infection. In: Zak O, Sande MA, eds. Experimental models in antimicrobial chemotherapy. Vol 1. London: Academic Press, 1986:355–369.

656.  Demarco de Hormaeche R, Macpherson A, Bowe F, et al. Alterations of the LPS determine virulence of Neisseria gonorrhoeae in guinea-pig subcutaneous chambers. Microb Pathog 1991;11:159–170.

657.  Elmros T, Holm SE, Kjellberg E, et al. Ampicillin treatment of Neisseria gonorrhoeae in vivo. An experimental study in rabbits. Acta Pathol Microbiol Scand B 1981;89:143–148.

658.  Keevil CW, Davies DB, Spillane BJ, et al. Influence of iron-limited and replete continuous culture on the physiology and virulence of Neisseria gonorrhoeae. J Gen Microbiol 1989;135:851–863.

659.  Jerse AE. Experimental gonococcal genital tract infection and opacity protein expression in estradiol-treated mice. Infect Immun 1999;67:5699–5708.

660.  Kita E, Katsui N, Emoto M, et al. Virulence of transparent and opaque colony types of Neisseria gonorrhoeae for the genital tract of mice. J Med Microbiol 1991;34:355–362.

661.  Song W, Condron S, Mocca BT, et al. Local and humoral immune responses against primary and repeat Neisseria gonorrhoeae genital tract infections of 17beta-estradiol-treated mice. Vaccine 2008;26:5741–5751.

662.  Jerse AE, Sharma ND, Simms AN, et al. A gonococcal efflux pump system enhances bacterial survival in a female mouse model of genital tract infection. Infect Immun 2003;71:5576–5582.

663.  Warner DM, Folster JP, Shafer WM, et al. Regulation of the MtrC-MtrD-MtrE efflux-pump system modulates the in vivo fitness of Neisseria gonorrhoeae. J Infect Dis 2007;196:1804–1812.

664.  Warner DM, Shafer WM, Jerse AE. Clinically relevant mutations that cause derepression of the Neisseria gonorrhoeae MtrC-MtrD-MtrE Efflux pump system confer different levels of antimicrobial resistance and in vivo fitness. Mol Microbiol 2008;70:462–478.

665.  Kunz AN, Begum AA, Wu H, et al. Impact of fluoroquinolone resistance mutations on gonococcal fitness and in vivo selection for compensatory mutations. J Infect Dis 2012;1821–1829.

666.  Alder JD. The hamster model of syphilis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:285–289.

667.  Baughn RE, Adams C, Musher DM. Evaluation of Sch 29482 in experimental syphilis and comparison with penicillin G benzathine in disseminated disease and localized infection. Antimicrob Agents Chemother 1984;26:401–404.

668.  Lukehart SA, Baker-Zander SA, Holmes KK. Efficacy of aztreonam in treatment of experimental syphilis in rabbits. Antimicrob Agents Chemother 1984;25:390–391.

669.  Fitzgerald TJ. Effects of cefetamet (Ro 15-8074) on Treponema pallidum and experimental syphilis. Antimicrob Agents Chemother 1992;36:598–602.

670.  Baker-Zander SA, Lukehart SA. Efficacy of cefmetazole in the treatment of active syphilis in the rabbit model. Antimicrob Agents Chemother 1989;33:1465–1469.

671.  Johnson RC, Bey RF, Wolgamot SJ. Comparison of the activities of ceftriaxone and penicillin G against experimentally induced syphilis in rabbits. Antimicrob Agents Chemother 1982;21:984–989.

672.  Lukehart SA, Fohn MJ, Baker-Zander SA. Efficacy of azithromycin for therapy of active syphilis in the rabbit model. J Antimicrob Chemother 1990;25(Suppl A):91–99.

673.  Alder J, Jarvis K, Mitten M, et al. Clarithromycin therapy of experimental Treponema pallidum infections in hamsters. Antimicrob Agents Chemother 1993;37:864–867.

674.  Korting HC, Haag R, Walter D, et al. Efficacy of ceftizoxime in the treatment of incubating syphilis in rabbits. Chemotherapy 1993;39:331–335.

675.  Korting HC, Walther D, Riethmuller U, et al. Ceftriaxone given repeatedly cures manifest syphilis in the rabbit. Chemotherapy 1987;33:376–380.

676.  Marra C, Baker-Zander SA, Hook EW 3rd, et al. An experimental model of early central nervous system syphilis. J Infect Dis 1991;163:825–829.

677.  Marra CM, Slatter V, Tartaglione TA, et al. Evaluation of aqueous penicillin G and ceftriaxone for experimental neurosyphilis. J Infect Dis 1992;165:396–397.

678.  Tantalo LC, Lukehart SA, Marra CM. Treponema pallidum strain-specific differences in neuroinvasion and clinical phenotype in a rabbit model. J Infect Dis 2005;191:75–80.

679.  Kajdacsy-Balla A, Howeedy A, Bagasra O. Experimental model of congenital syphilis. Infect Immun 1993;61:3559–3561.

680.  Kajdacsy-Balla A, Howeedy A, Bagasra O. Syphilis in the Syrian hamster. A model of human venereal and congenital syphilis. Am J Pathol 1987;126:599–601.

681.  Wicher V, Wicher K. Guinea-pig model of acquired and congenital syphilis. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:291–301.

682.  Moller BR, Mardh PA. Animal models for the study of chlamydial infections of the urogenital tract. Scand J Infect Dis Suppl 1982;32:103–108.

683.  Beale AS, Upshon PA. Characteristics of murine model of genital infection with Chlamydia trachomatis and effects of therapy with tetracyclines, amoxicillin-clavulanic acid, or azithromycin. Antimicrob Agents Chemother 1994;38:1937–1943.

684.  Tuffrey MC, Woods C, Inman C, et al. The effect of a single oral dose of azithromycin on chlamydial infertility and oviduct ultrastructure in mice. J Antimicrob Chemother 1994;34:989–999.

685.  Pal S, Sarcon AK, de la Maza LM. A new murine model for testing vaccines against genital Chlamydia trachomatis infections in males. Vaccine 2010;28:7606–7612.

686.  Carey A, Cunningham K, Andrew D, et al. A comparison of the effects of a chlamydial vaccine administered during or after a C. muridarum urogenital infection of female mice. Vaccine 2011;29:6505–6513.

687.  Imtiaz MT, Schripsema JH, Sigar IM, et al. Inhibition of matrix metalloproteinases protects mice from ascending infection and chronic disease manifestations resulting from urogenital Chlamydia muridarum infection. Infect Immun 2006;74:5513–5521.

688.  Imtiaz MT, Schripsema JH, Sigar IM, et al. Outcome of urogenital infection with Chlamydia muridarum in CD14 gene knockout mice. BMC Infect Dis 2006;6:144.

689.  Johnson RM, Kerr MS, Slaven JE. Plac8-dependent and inducible NO synthase-dependent mechanisms clear Chlamydia muridarum infections from the genital tract. J Immunol 2012;188:1896–1904.

690.  Johnson RM, Yu H, Kerr MS, et al. PmpG303-311, a protective vaccine epitope that elicits persistent cellular immune responses in Chlamydia muridarum-immune mice. Infect Immun 2012;80:2204–2211.

691.  Reeves DM, Nagarajan U, O’Connell C, et al. Lack of an effect of antibiotic treatment on prolonged detection of chlamydial DNA in murine genital tract infection. Antimicrob Agents Chemother 2007;51:2646–2648.

692.  Boyd AR, Shivshankar P, Jiang S, et al. Age-related defects in TLR2 signaling diminish the cytokine response by alveolar macrophages during murine pneumococcal pneumonia. Exp Gerontol 2012;47:507–518.

693.  Frazer LC, Darville T, Chandra-Kuntal K, et al. Plasmid-cured Chlamydia caviae activates TLR2-dependent signaling and retains virulence in the guinea pig model of genital tract infection. PLoS One 2012;7:e30747.

694.  Wang Y, Nagarajan U, Hennings L, et al. Local host response to chlamydial urethral infection in male guinea pigs. Infect Immun 2010;78:1670–1681.

695.  Sayek I. Animal models for intra-abdominal infection. Hepatogastroenterology 1997;44:923–926.

696.  Weinstein WM, Onderdonk AB, Bartlen JC, et al. Antimicrobial therapy of experimental intraabdominal sepsis. J Infect Dis 1975;132:282–286.

697.  Su H, Morrison R, Messer R, et al. The effect of doxycycline treatment on the development of protective immunity in a murine model of chlamydial genital infection. J Infect Dis 1999;180:1252–1258.

698.  Dupont H, Montravers P. Rat polymicrobial peritonitis infection model. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:189–194.

699.  Atmatzidis S, Koutelidakis I, Chatzimavroudis G, et al. Clarithromycin modulates immune responses in experimental peritonitis. Int J Antimicrob Agents 2011;37:347–351.

700.  Hyde SR, Stitih RD, McCallum RE. Mortality and bacteriology of sepsis following cecal ligation and puncture in aged mice. Infect Immun 1990;58:619–624.

701.  Vianna RC, Gomes RN, Bozza FA, et al. Antibiotic treatment in a murine model of sepsis: impact on cytokines and endotoxin release. Shock 2004;21:115–120.

702.  Remick DG, Call DR, Ebong SJ, et al. Combination immunotherapy with soluble tumor necrosis factor receptors plus interleukin 1 receptor antagonist decreases sepsis mortality. Crit Care Med 2001;29:473–481.

703.  Cirioni O, Ghiselli R, Kamysz W, et al. Tachyplesin III and granulocyte-colony stimulating factor enhance the efficacy of tazobactam/piperacillin in a neutropenic mouse model of polymicrobial peritonitis. Peptides 2008;29:31–38.

704.  Cirioni O, Giacometti A, Ghiselli R, et al. LL-37 protects rats against lethal sepsis caused by gram-negative bacteria. Antimicrob Agents Chemother 2006;50:1672–1679.

705.  Cirioni O, Wu G, Li L, et al. S-thanatin enhances the efficacy of tigecycline in an experimental rat model of polymicrobial peritonitis. Peptides 2010;31:1231–1236.

706.  Ghiselli R, Cirioni O, Giacometti A, et al. The cathelicidin-derived tritrpticin enhances the efficacy of ertapenem in experimental rat models of septic shock. Shock 2006;26:195–200.

707.  Ghiselli R, Giacometti A, Cirioni O, et al. Efficacy of the bovine antimicrobial peptide indolicidin combined with piperacillin/tazobactam in experimental rat models of polymicrobial peritonitis. Crit Care Med 2008;36:240–245.

708.  Giacometti A, Cirioni O, Ghiselli R, et al. Effects of pexiganan alone and combined with betalactams in experimental endotoxic shock. Peptides 2005;26:207–216.

709.  Wu RQ, Xu YX, Song XH, et al. Relationship between cytokine mRNA expression and organ damage following cecal ligation and puncture. World J Gastroenterol 2002;8:131–134.

710.  Yang S, Chung CS, Ayala A, et al. Differential alterations in cardiovascular responses during the progression of polymicrobial sepsis in the mouse. Shock 2002;17:55–60.

711.  Hollenberg SM, Dumasius A, Easington C, et al. Characterization of a hyperdynamic murine model of resuscitated sepsis using echocardiography. Am J Respir Crit Care Med 2001;164:891–895.

712.  Kotake Y, Moore DR, Vasquez-Walden A, et al. Antioxidant amplifies antibiotic protection in the cecal ligation and puncture model of microbial sepsis through interleukin-10 production. Shock 2003;19:252–256.

713.  Brook I. Intra-abdominal abscess. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:163–172.

714.  Joiner K, Lower B, Dzink J, et al. Comparative efficacy of 10 antimicrobial agents in experimental infections with Bacteroides fragilis. J Infect Dis 1982;145:561–568.

715.  Kapral FA, Godwin JR, Dye ES. Formation of intraperitoneal abscesses by Staphylococcus aureus. Infect Immun 1980;30:204–211.

716.  Stoddart B, Wilcox MH. Clostridium difficile. Curr Opin Infect Dis 2002;15:513–518

717.  Tran MC, Claros MC, Goldstein EJ. Therapy of Clostridium difficile infection: perspectives on a changing paradigm. Expert Opin Pharmacother 2013;14:2375–2386.

718.  Abrams GD, Allo M, Rifkin GD, et al. Mucosal damage mediated by clostridial toxin in experimental clindamycin-associated colitis. Gut 1980;21:493–499.

719.  Czuprynski CJ, Johnson WJ, Balish E, et al. Pseudomembranous colitis in Clostridium difficile-monoassociated rats. Infect Immun 1983;39:1368–1376.

720.  Fekety R, Silva J, Toshniwal R, et al. Antibiotic-associated colitis: effects of antibiotics on Clostridium difficile and the disease in hamsters. Rev Infect Dis 1979;1:386–397.

721.  Knoop FC. Clindamycin-associated enterocolitis in guinea pigs: evidence for a bacterial toxin. Infect Immun 1979;23:31–33.

722.  Pawlowski SW, Calabrese G, Kolling GL, et al. Murine model of Clostridium difficile infection with aged gnotobiotic C57BL/6 mice and a BI/NAP1 strain. J Infect Dis 2010;202:1708–1712.

723.  Steele J, Feng H, Parry N, et al. Piglet models of acute or chronic Clostridium difficile illness. J Infect Dis 2010;201:428–434.

724.  Bartlett JG, Chang TW, Onderdonk AB. Comparison of five regimens for treatment of experimental clindamycin-associated colitis. J Infect Dis 1977;136:81–86.

725.  Bartlett JG, Onderdonk AB, Cisneros RL. Clindamycin-associated colitis due to a toxin-producing species of Clostridium in hamsters. J Infect Dis 1977;136:701–705.

726.  Lyerly DM, Saum KE, MacDonald DK, et al. Effects of Clostridium difficile toxins given intragastrically to animals. Infect Immun 1985;47:349–352.

727.  Wilson KH, Sheagren JN, Freter R. Population dynamics of infested Clostridium difficile in the gastrointestinal tract of the Syrian hamster. J Infect Dis 1985;151:355–361.

728.  Blankenship-Paris TL, Walton BJ, Hayes YO, et al. Clostridium difficile infection in hamsters fed an atherogenic diet. Vet Pathol 1995;32:269–273.

729.  Dong M-Y, Chang T-W, Gorbach SL. Treatment of Clostridium difficile colitis in hamsters with a lipopeptide antibiotic, LY146032. Antimicrob Agents Chemother 1987;31:1135–1136.

730.  McVay CS, Rolfe RD. In vitro and in vivo activities of nitazoxanide against Clostridium difficile. Antimicrob Agents Chemother 2000;44:2254–2258.

731.  Drummond LJ, Smith DG, Poxton IR. Effects of sub-MIC concentrations of antibiotics on growth of and toxin production by Clostridium difficile. J Med Microbiol 2003;52:1033–1038.

732.  Sun X, Wang H, Zhang Y, et al. Mouse relapse model of Clostridium difficile infection. Infect Immun 2011;79:2856–2864.

733.  Chen X, Katchar K, Goldsmith JD, et al. A mouse model of Clostridium difficile-associated disease. Gastroenterology 2008;135:1984–1992.

735.  Li Y, Figler RA, Kolling G, et al. Adenosine A2A receptor activation reduces recurrence and mortality from Clostridium difficile infection in mice following vancomycin treatment. BMC Infect Dis 2012;12:342.

736.  Freeman J, Baines SD, Jabes D, et al. Comparison of the efficacy of ramoplanin and vancomycin in both in vitro and in vivo models of clindamycin-induced Clostridium difficile infection. J Antimicrob Chemother 2005;56:717–725.

737.  Freeman J, Marquis M, Crowther GS, et al. Oritavancin does not induce Clostridium difficile germination and toxin production in hamsters or a human gut model. J Antimicrob Chemother 2012;67:2919–2926.

738.  Jump RL, Li Y, Pultz MJ, et al. Tigecycline exhibits inhibitory activity against Clostridium difficile in the colon of mice and does not promote growth or toxin production. Antimicrob Agents Chemother 2011;55:546–549.

739.  Adams DA, Riggs MM, Donskey CJ. Effect of fluoroquinolone treatment on growth of and toxin production by epidemic and nonepidemic Clostridium difficile strains in the cecal contents of mice. Antimicrob Agents Chemother 2007;51:2674–2678.

740.  Lee A, Fox J, Hazell S. Pathogenicity of Helicobacter pylori: a perspective. Infect Immun 1993;61:1601–1610.

741.  Marchetti M, Arico B, Burroni D. Development of a mouse model of Helicobacter pylori infection that mimics human disease. Science 1995;267:1655–1658.

742.  Van Zanten SJ, Kolesnikow T, Leung V, et al. Gastric transitional zones, areas where Helicobacter treatment fails: results of a treatment trial using the Sydney strain mouse model. Antimicrob Agents Chemother 2003;47:2249–2255.

743.  Endo H, Yoshida H, Ohmi N, et al. Effects of lansoprazole, clarithromycin and pH gradient on uptake of [¹⁴C]amoxicillin into rat gastric tissue. J Antimicrob Chemother 2001;47:405–410.

744.  Koga T, Inoue H, Ishii C, et al. Effect of plaunotol in combination with clarithromycin or amoxicillin on Helicobacter pylori in vitro and in vivo. J Antimicrob Chemother 2002;50:133–136.

745.  Tran CD, Kritas S, Campbell MA, et al. Novel combination therapy for the eradication of Helicobacter pylori infection in a mouse model. Scand J Gastroenterol 2010;45:1424–1430.

746.  Madoff LC, Michel JL, Gong EW, et al. Protection of neonatal mice from group B streptococcal infection by maternal immunization with beta C protein. Infect Immun 1992;60:4989–4994.

747.  Rodewald AK, Onderdonk AB, Warren HB, et al. Neonatal mouse model of group B streptococcal infection. J Infect Dis 1992;166:635–639.

748.  Kronforst KD, Mancuso CJ, Pettengill M, et al. A neonatal model of intravenous Staphylococcus epidermidis infection in mice <24 h old enables characterization of early innate immune responses. PLoS One 2012;7:e43897.

749.  Oluola O, Kong L, Fein M, et al. Lysostaphin in treatment of neonatal Staphylococcus aureus infection. Antimicrob Agents Chemother 2007;51:2198–2200.

750.  Placencia FX, Kong L, Weisman LE. Treatment of methicillin-resistant Staphylococcus aureus in neonatal mice: lysostaphin versus vancomycin. Pediatr Res 2009;65:420–424.

751.  Venkatesh MP, Pham D, Fein M, et al. Neonatal coinfection model of coagulase-negative Staphylococcus (Staphylococcus epidermidis) and Candida albicans: fluconazole prophylaxis enhances survival and growth. Antimicrob Agents Chemother 2007;51:1240–1245.

752.  Ashman RB, Papadimitriou JM, Fulurija A. Acute susceptibility of aged mice to infection with Candida albicans. J Med Microbiol 1999;48:1095–1102.

754.  Murciano C, Villamon E, Yanez A, et al. Impaired immune response to Candida albicans in aged mice. J Med Microbiol 2006;55:1649–1656.

755.  Pawlowski SW, Calabrese G, Kolling GL, et al. Murine model of Clostridium difficile infection with aged gnotobiotic C57BL/6 mice and a BI/NAP1 strain. J Infect Dis 2010;202:1708–1712.

756.  Turnbull IR, Wlzorek JJ, Osborne D, et al. Effects of age on mortality and antibiotic efficacy in cecal ligation and puncture. Shock 2003;19:310–313.

757.  Bass R, Lehnert T. Basic requirements for the toxicity testing of antimicrobial agents. Eur J Clin Microbiol Infect Dis 1990;9:488–491.

758.  Mandell LA, Ball P, Tillotson G. Antimicrobial safety and tolerability: differences and dilemmas. Clin Infect Dis 2001;32:S72–S75.

759.  Morris TH. Antibiotic therapeutics in laboratory animals. Lab Anim 1995;29:16–36.

760.  Kacew S, Festing MFW. Role of rat strain in the differential sensitivity to pharmaceutical agents and naturally occurring substances. J Toxicol Environ Health 1996;47:1–30.

761.  Festing MF, Diamanti P, Turton JA. Strain differences in haematological response to chloramphenicol succinate in mice: implications for toxicological research. Food Chem Toxicol 2001;39:375–383.

762.  Holt D, Harvey D, Hurley R. Chloramphenicol toxicity. Adverse Drug React Toxicol Rev 1993;12:83–95.

763.  Kararli TT. Comparison of the gastrointestinal anatomy, physiology, and biochemistry of humans and commonly used laboratory animals. Biopharm Drug Dispos 1995;16:351–380.

764.  Cherbut C, Ferre JP, Corpet DE, et al. Alterations of intestinal microflora by antibiotics: effects on fecal excretion, transit time, and colonic motility in rats. Dig Dis Sci 1991;36:1729–1734.

765.  Brunel A, Gouet P. Influence of the destabilization of the maternal digestive microflora on that of the newborn. Biol Neonate 1993;63:236–245.

766.  Tune BM. Renal tubular transport and nephrotoxicity of beta lactam antibiotics: structure-activity relationships. Miner Electrolyte Metab 1994;20:221–231.

767.  Valentovic MA, Ball JG, Rogers BA. Comparison of renal accumulation and urinary excretion in normoglycemic and diabetic animals. Toxicology 1996;108:93–99.

768.  Ohtani H, Taninaka C, Hanada E, et al. Comparative pharmacodynamic analysis of Q-T interval prolongation induced by the macrolides clarithromycin, roxithromycin, and azithromycin in rats. Antimicrob Agents Chemother 2000;44:2630–2637.

769.  Chenel M, Barbot A, Dupuis A, et al. Pharmacokinetic-pharmacodynamic modeling of the electroencephalogram effect of norfloxacin in rats. Antimicrob Agents Chemother 2003;47:1952–1957.

770.  Van Etten EWM, ten Kate MT, Stearne LET, et al. Amphotericin B liposomes with prolonged circulation in blood: in vitro antifungal activity, toxicity, nd efficacy in systemic candidiasis in leukopenic mice. Antimicrob Agents Chemother 1995;39:1954–1958.

771.  Wood CA, Finkbeiner HC, Kohlhepp SJ, et al. Influence of daptomycin on staphylococcal abscesses and experimental tobramycin nephrotoxicity. Antimicrob Agents Chemother 1989;33:1280–1285.

772.  Bauhofer A, Witte K, Celik I, et al. Sickness behaviour, an animal equivalent to human quality of life, is improved in septic rats by G-CSF and antibiotic prophylaxis. Langenbecks Arch Surg 2001;386:132–140.

773.  Bauhofer A, Witte K, Lemmer B, et al. Quality of life in animals as a new outcome for surgical research: G-CSF as a quality of life improving factor. Eur Surg Res 2002;34:22–29.

774.  Craig WA. Post-antibiotic effects in experimental infection models: relationship to in-vitro phenomena and to treatment of infections in man. J Antimicrob Chemother 1993;31(Suppl D):149–158.

775.  Craig WA, Legget J, Totsuka K, et al. Key pharmacokinetic parameters of antibiotic efficacy in experimental animal infections. J Drug Dev 1988;1(Suppl 3):7–15.

776.  Fantin B, Ebert S, Leggett J, et al. Factors affecting duration of in-vivo postantibiotic effect for aminoglycosides against Gram-negative bacilli. J Antimicrob Chemother 1991;27:829–836.

777.  Gudmundsson S, Vogelman B, Craig WA. The in-vivo postantibiotic effect of imipenem and other new antimicrobials. J Antimicrob Chemother 1986;18(Suppl E):67–73.

778.  MacKenzie FM, Gould IM. The post-antibiotic effect. J Antimicrob Chemother 1993;32:519–537.

779.  Minguez F, Izquierdo J, Caminero MM, et al. In vivo postantibiotic effect of isepamicin and other aminoglycosides in a thigh infection model in neutropenic mice. Chemotherapy 1992;38:179–184.

780.  Vogelman B, Gudmundsson S, Leggett J, et al. Correlation of antimicrobial pharmacokinetic indices with therapeutic efficacy in an animal model. J Infect Dis 1988;158:831–847.

781.  Vogelman B, Gudmundsson S, Tumidge J, et al. In vivo postantibiotic effect in a thigh infection in neutropenic mice. J Infect Dis 1988;157:287–298.

782.  Zhanel GG, Craig WA. Pharmacokinetic contributions to postantibiotic effects: focus on aminoglycosides. Clin Pharmacokinet 1994;27:377–392.

783.  Cars O, Odenholt-Tornqvist I. The post-antibiotic sub-MIC effect in vitro and in vivo. J Antimicrob Chemother 1993;31(Suppl D):159–166.

784.  Odenholt-Tornqvist I. Pharmacodynamics of beta-lactam antibiotics: studies on the paradoxical and postantibiotic effects in vitro and in an animal model. Scand J Infect Dis Suppl 1989;58:1–55.

785.  Oshida T, Onta T, Nakanishi N, et al. Activity of sub-minimal inhibitory concentrations of aspoxicillin in prolonging the postantibiotic effect against Staphylococcus aureus. J Antimicrob Chemother 1990;26:29–38.

786.  Majcherczyk PA, Kunz S, Hattenberger M, et al. Isolation and in-vitro and in-vivo characterisation of a mutant of Pseudomonas aeruginosa PA01 that exhibited a reduced postantibiotic effect in response to imipenem. J Antimicrob Chemother 1994;34:485–505.

787.  Andes D, Marchillo K, Lowther J, et al. In vivo pharmacodynamics of HMR 3270, a glucan synthase inhibitor, in a murine candidiasis model. Antimicrob Agents Chemother 2003;47:1187–1192.

788.  Spivey JM. The postantibiotic effect. Clin Pharm 1992;11:865–875.

789.  Zhanel GG, Hoban DJ, Harding GK. The postantibiotic effect: a review of in vitro and in vivo data. DICP 1991;25:153–163.

791.  Hughes WT, Armstrong D, Bodey GP, et al. Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. J Infect Dis 1990;161:381–396.

792.  Love LJ, Schimpff SC, Schiffer CA, et al. Improved prognosis for granulocytopenic patients with Gram-negative bacteremia. Am J Med 1980;68:643–647.

793.  Gerber AU, Greter U, Segessenmann C, et al. The impact of pre-treatment interval on antimicrobial efficacy in a biological model. J Antimicrob Chemother 1993;31(Suppl D):29–39.

794.  Cremieux A-C, Saleh-Mghir A, Vallois J-M, et al. Influence of the pre-treatment duration of infection on the efficacies of various antibiotic regimens in experimental streptococcal endocarditis. J Antimicrob Chemother 1993;32:843–852.

795.  Azoulay-Dupuis E, Bedos J-P, Vallee E, et al. Anti-pneumococcal activity of ciprofloxacin, ofloxacin, and temafloxacin in an experimental mouse pneumonia model at various stages of disease. J Infect Dis 1991;163:319–324.

796.  Kalns J, Morris J, Eggers J, et al. Delayed treatment with doxycycline has limited effect on anthrax infection in BKL57/B6. Biochem Biophys Res Commun 2002;297:506–509.

797.  Contag PR. Whole-animal cellular and molecular imaging to accelerate drug development. Drug Discov Today 2002;7:555–562.

798.  Contag PR, Olomu AB, Contag CH. Non-invasive monitoring of infection and gene expression in living animal models. In: Zak O, Sande MA, eds. Handbook of animal models of infection. San Diego, CA: Academic Press, 1999:61–68.

799.  Contag PR, Olomu IN, Stevenson DK, et al. Bioluminescent indicators in living mammals. Nat Med 1998;4:245–247.

800.  Balaban RS, Hampshire VA. Challenges in small animal noninvasive imaging. ILAR J 2001;42:248–62.

801.  Signore A, Procaccini E, Annovazzi A, et al. The developing role of cytokines for imaging inflammation and infection. Cytokine 2000;12:1445–1454.

802.  Ruiz-Cabello J, Regadera J, Santisteban C, et al. Monitoring acute inflammatory processes in mouse muscle by MR imaging and spectroscopy: a comparison with pathological results. NMR Biomed 2002;15:204–214.

803.  Spellerberg B, Prasad S, Cabellos C, et al. Penetration of the blood-brain barrier: enhancement of drug delivery and imaging by bacterial glycopeptides. J Exp Med 1995;182:1037–1043.

804.  Marzola P, Nicolato E, Di Modugno E, et al. Comparison between MRI, microbiology and histology in evaluation of antibiotics in a murine model of thigh infection. MAGMA 1999;9:21–28.

805.  Spaeth HJ, Chandnani VP, Beltran J, et al. Magnetic resonance imaging detection of early experimental periostitis: comparison of magnetic resonance imaging, computed tomography, and plain radiography with histopathologic correlation. Invest Radiol 1991;26:304–308.

806.  Fischman AJ, Alpert NM, Babich JW, et al. The role of positron emission tomography in pharmacokinetic analysis. Drug Metab Rev 1997;29:923–956.

807.  Fischman AJ, Babich JW, Alpert NM, et al. Pharmacokinetics of ¹⁸F-labeled trovafloxacin in normal and Escherichia coli–infected rats and rabbits studied with positron emission tomography. Clin Microbiol Infect 1997;3:63–72.

808.  Via LE, Schimel D, Weiner DM, et al. Infection dynamics and response to chemotherapy in a rabbit model of tuberculosis using ¹⁸F-2-Fluoro-deoxy-D-glucose positron emission timographt and computed tomography. Antimicrob Agents Chemother 2012;56:4391–4402.

809.  Chandnani VP, Beltran J, Morris CS, et al. Acute experimental osteomyelitis and abscesses: detection with MR imaging versus CT. Radiology 1990;174:233–236.

810.  Beltran J, McGhee RB, Shaffer PB, et al. Experimental infections of the musculoskeletal system: evaluation with MR imaging and Tc-99m MDP and Ga-67 scintigraphy. Radiology 1988;167:167–172.

811.  Sartoris DJ, Guerra J Jr, Mattrey RF, et al. Perfluoroctylbromide as a contrast agent for computed tomographic imaging of septic and aseptic arthritis. Invest Radiol 1986;21:49–55.

812.  Chen X, Schmidt AH, Mahjouri S, et al. Union of a chronically infected internally stabilized segmental defect in the rat femur after debridement and application of rhBMP-2 and systemic antibiotics. J Orthop Trauma. 2007;21:693–700.

813.  Xiong YQ, Kupferwasser LI, Zack PM, et al. Comparative efficacies of liposomal amikacin (MiKasome) plus oxacillin versus conventional amikacin plus oxacillin in experimental endocarditis induced by Staphylococcus aureus: microbiological and echocardiographic analyses. Antimicrob Agents Chemother 1999;43:1737–1742.

814.  Ntziachristos V, Bremer C, Weissleder R. Fluorescence imaging with near-infrared light: new technological advances that enable in vivo molecular imaging. Eur Radiol 2003;13:195–208.

815.  Greer LF 3rd, Szalay AA. Imaging of light emission from the expression of luciferases in living cells and organisms: a review. Luminescence 2002;17:43–74.

816.  Contag CH, Bachmann MH. Advances in in vivo bioluminescence imaging of gene expression. Annu Rev Biomed Eng 2002;4:235–260.

817.  Burns SM, Jon D, Francis KP, et al. Revealing the spatiotemporal patterns of bacterial infectious diseases using bioluminescent pathogens and whole body imaging. Animal testing in infectology. Contrib Microbiol 2001;9:71–88.

818.  Contag CH, Contag PR, Mullins JI, et al. Photonic detection of bacterial pathogens in living hosts. Mol Microbiol 1995;18:593–603.

819.  Rocchetta HL, Boylan CJ, Foley JW, et al. Validation of a noninvasive, real-time imaging technology using bioluminescent Escherichia coli in the neutropenic mouse thigh model of infection. Antimicrob Agents Chemother 2001;45:129–137.

820.  Francis KP, Yu J, Bellinger-Kawahara C, et al. Visualizing pneumococcal infections in the lungs of live mice using bioluminescent Streptococcus pneumoniae transformed with a novel Gram-positive. Infect Immun 2001;69:3350–3358.

821.  Guo Y, Ramos RI, Cho JS, et al. In vivo bioluminiscense imaging to evaluate systemic and topical antibiotics against commuity-acquired methicillin-resistant Staphylococcus aureus-infected skin wounds in mice. Antimicrob Agents Chemother 2013;57:855–863.

822.  Valdevia RH, Falkow S. Probing bacterial gene expression within host cells. Trends Microbiol 1997;5:360–363.

823.  Valdevia RH, Falkow S. Fluorescence-based isolation of bacterial genes expressed within host cells. Science 1997;26:2007–2011.

824.  Suarez A, Guttler A, Stratz M, et al. Green fluorescent protein-based reporter systems for genetic analysis of bacteria including monocopy applications. Gene 1997;196:69–74.

825.  Bongaerts RJ, Hautefort I, Sidebotham JM, et al. Green fluorescent protein as a marker for conditional gene expression in bacterial cells. Methods Enzymol 2002;358:43–66.

826.  Jacobi CA, Roggenkamp A, Rakin A, et al. In vitro and in vivo expression studies of yopE from Yersinia enterocolitica using the gfp reporter gene. Mol Microbiol 1998;30:865–882.

827.  Acebo P, Nieto C, Corrales MA, et al. Quantitative detection of Streptococcus pneumoniae cells harbouring single or multiple copies of the gene encoding the green fluorescent protein. Microbiology 2000;146:1267–1273.

828.  Valdivia RH, Hromockyj AE, Monack D, et al. Applications for green fluorescent protein (GFP) in the study of host-pathogen interactions. Gene 1996;173(1 Spec No):47–52.

829.  del Poeta M, Toffaletti DL, Rude TH, et al. Cryptococcus neoformans differential gene expression detected in vitro and in vivo with green fluorescent protein. Infect Immun 1999;67:1812–1820.

830.  Josenhans C, Friedrich S, Suerbaum S. Green fluorescent protein as a novel marker and reporter system in Helicobacter sp. FEMS Microbiol Lett 1998;161:263–273.

831.  Scholz O, Thiel A, Hillen W, et al. Quantitative analysis of gene expression with an improved green fluorescent protein, p6. Eur J Biochem 2000;267:1565–1570.

832.  Maksimow M, Hakkila K, Karp M, et al. Simultaneous detection of bacteria expressing GFP and DsRed genes with a flow cytometer. Cytometry 2002;47:243–247.

833.  Leveau JH, Lindow SE. Predictive and interpretive simulation of green fluorescent protein expression in reporter bacteria. J Bacteriol 2001;183:6752–6762.

834.  Wendland M, Bumann D. Optimization of GFP levels for analyzing Salmonella gene expression during an infection. FEBS Lett 2002;521:105–108.

835.  Bryant PA, Venter D, Robins-Browne R, et al. Chips with everything: DNA microarrays in infectious diseases. Lancet Infect Dis 2004;4:100–111.

836.  Chalker AF, Lunsford RD. Rational identification of new antibacterial drug targets that are essential for viability using a genomics-based approach. Pharmacol Ther 2002;95:1–20.

837.  Fritz B, Raczniak GA. Bacterial genomics: potential for antimicrobial drug discovery. BioDrugs 2002;16:331–337.

838.  Hayney MS. Pharmacogenomics and infectious diseases: impact on drug response and applications to disease management. Am J Health Syst Pharm 2002;59:1626–1631.

839.  Knowles DJ, King F. The impact of bacterial genomics on antibacterial discovery. Adv Exp Med Biol 1998;456:183–195.

840.  De Backer MD, Van Dijck P. Progress in functional genomics approaches to antifungal drug target discovery. Trends Microbiol 2003;11:470–478.

841.  Watters JW, McLeod HL. Using genome-wide mapping in the mouse to identify genes that influence drug response. Trends Pharmacol Sci 2003;24:55–58.

842.  Kimura S, Gonzalez FJ. Applications of genetically manipulated mice in pharmacogenetics and pharmacogenomics. Pharmacology 2000;61:147–153.

843.  Drusano GL. Antimicrobial pharmacodynamics: critical interactions of “bug and drug.” Nat Rev Microbiol 2004;2:289–300.

844.  Moreillon P. Endocarditis prophylaxis revisited: experimental evidence of efficacy and new Swiss recommendations. Swiss Working Group for Endocarditis Prophylaxis. Schweiz Med Wochenschr 2000;130:1013–1026.