The essence of antimicrobial chemotherapy is selective toxicity—to kill or inhibit the microbe without harming the patient. So far as bacteria are concerned, a prime target for such an attack is the cell wall, since practically all bacteria (with the exception of mycoplasmas) have a cell wall, whereas mammalian cells lack this feature.
Not all bacterial cell walls are the same. Indeed, of the many species that have been investigated, no two have been found to be identical. However, in general they conform to two basic patterns which may readily be distinguished by that most familiar of all microbiological techniques, the Gram stain.
The cell walls of Gram-positive and Gram-negative bacteria differ in many fundamental respects but both groups possess a cross-linked chain of peptido-glycan (also called mucopeptide or murein), which gives the wall its strength. Cell-wall-active antibiotics interfere with the biosynthesis of this structure. Peptidoglycan consists of a backbone of alternating units of N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAMA). Each NAMA molecule carries a short peptide made up of five amino acids which cross-link (via an interpeptide bridge composed of further amino acids in most Gram-positive organisms) to provide the characteristic rigidity.
In Gram-positive organisms the peptidoglycan is thick (about 30 nm), tightly cross-linked, and interspersed with polysugarphosphates (teichoic acids), some of which have a lipophilic tail buried in the cell membrane (lipoteichoic acids). Gram-negative bacteria, in contrast, have a relatively thin (2–3 nm), loosely cross-linked peptidoglycan layer.
External to the Gram-negative peptidoglycan is a membrane-like structure, composed chiefly of lipopolysaccharide and lipoprotein, which may prevent large hydrophilic molecules from reaching an otherwise susceptible cellular target. Small hydrophilic molecules enter Gram-negative bacilli through aqueous channels, porins, within the outer membrane. Differential activity among some groups of antibiotics, notably the penicillins and cephalosporins, is influenced by their
ability to negotiate these porin channels and this, in turn, reflects the size and ionic charge of substituents carried by the individual agents.
The bacterial peptidoglycan in both Gram-positive and Gram-negative organisms is assembled from units of NAG, initially linked to uridine diphosphate (UDP). UDP–NAMA units are manufactured from UDP–NAG by the addition of a lactic acid moiety derived from phosphoenolpyruvate. The NAMA then receives, one by one, three amino acids—usually L-alanine, D-glutamic acid, and either L-lysine (in Gram-positive organisms) or meso-diaminopimelic acid (in Gram-negative organisms). Meanwhile, two D-alanine residues, produced from L-alanine by an enzyme called alanine racemase, are joined together by another enzyme, D-alanine synthetase. The linked unit, D-ala-D-ala, is added to the tripeptide side-chain of NAMA and the NAMA–pentapeptide thus formed is passed to a lipid carrier in the cell membrane. Here a UDP–NAG unit transfers its NAG to the NAMA-pentapeptide and any amino acids needed for interpeptide bridges are added to the L-lysine of the pentapeptide side-chain. The lipid carrier transports the whole building block across the cell membrane and the unit is added to the end of the growing peptidoglycan chain of the existing cell wall, where the final cross-linking reaction takes place. The process is illustrated in outline in Fig. 1.1.
Fig. 1.1 Simplified scheme of bacterial cell wall synthesis showing site of action of cell-wall-active antibiotics.
The most important groups of antibiotics that interfere with cell wall synthesis are the glycopeptides and the β-lactam agents, but bacitracin, cycloserine, and fosfomycin also act at this level. In addition, isoniazid and some other anti-mycobacterial agents appear to act on the specialized mycobacterial cell wall (p. 57).
Bacitracin is one of a group of antibiotics, also including gramicidin and tyro-cidine (p. 57), that exhibit a cyclic structure made up of about ten amino acids. Bacitracin itself (actually a mixture of three closely related compounds, baci-tracin A, B, and C, of which bacitracin A is the major component) was first obtained from a strain of Bacillus subtilis grown from the infected wound of a 7 year-old girl, Margaret Tracy, in whose honour the antibiotic was named.
The spectrum of activity of bacitracin and related cyclic peptides is virtually restricted to Gram-positive organisms. They are too toxic for systemic use but are found in topical preparations. Bacitracin also finds a place in microbiology laboratories in the presumptive identification of Streptococcus pyogenes, which is exquisitely susceptible to its action. It was formerly used as a growth promoter in animal feedstuffs.
Bacitracin acts by preventing regeneration of the lipid carrier in the cell membrane, which is left in an unusable phosphorylated form after transporting cell wall subunits (Fig. 1.1). Gramicidin and tyrocidine have a different mechanism of action, interfering with the integrity of the cell membrane.
Cycloserine bears a stuctural resemblance to the D-isomer of alanine and inhibits alanine racemase. It also blocks the synthetase enzyme that links two D-ala molecules together before they are inserted into the cell wall (Fig. 1.1). The drug has broad-spectrum, but rather feeble, antibacterial activity. Its chief attraction lies in its activity against Mycobacterium tuberculosis, but even against this important pathogen it is used only as a reserve drug, mainly because of toxicity problems.
Fosfomycin (formerly known as phosphonomycin) is a naturally occurring antibiotic originally obtained from a species of Streptomycesisolated in Spain. Structurally, it is the simplest of all antibiotics (Fig. 1.2). The molecule inhibits the pyruvyl transferase enzyme which brings about the condensation of phos-phoenolpyruvate and NAG in the formation of NAMA (Fig. 1.1).
Fig. 1.2 Structure of fosfomycin.
Gram-positive cocci are rather less susceptible than Gram-negative rods. The precise level of activity is a matter of dispute, since the in-vitro activity can be manipulated by altering the test medium: the presence of glucose-6-phosphate potentiates the activity, whereas glucose and phosphate individually have an adverse effect. The effect of glucose-6-phosphate appears to be due to the fact that fosfomycin can enter bacteria by an active transport process that is inducible by glucose-6-phosphate, but not by fosfomycin itself.
Fosfomycin is formulated as the sodium salt for parenteral use, but this is unsuitable for oral administration. The trometamol (tromethamine) salt, which is highly soluble, well absorbed, and excreted in high concentration in urine, is preferable to the calcium salt for oral therapy.
Fosfomycin is well tolerated, and the ready emergence of bacterial resistance that is observed in vitro does not appear to have been a major problem in treatment. The trometamol salt has been successfully used in high dosage for the single-dose treatment of cystitis.
The glycopeptides vancomycin and teicoplanin are complex heterocyclic molecules consisting of a heptapeptide backbone to which are attached various substituted sugars. These compounds bind to acyl-D-alanyl-D-alanine in peptido-glycan, thereby preventing the addition of new building blocks to the growing cell wall (Fig. 1.1). Glycopeptides are too bulky to penetrate the external membrane of Gram-negative bacteria, so the spectrum of activity is generally restricted to Gram-positive organisms. Acquired resistance used to be uncommon, but resistant strains of enterococci are now widely prevalent and staphylococci exhibiting reduced susceptibility are causing concern. Avoparcin, a glycopeptide formerly used in animal husbandry in some countries (now banned in the European Union), has been implicated in generating resistance in enterococci, but human use of glycopeptides is equally important. Some Gram-positive genera, including Lactobacillus spp.,Pediococcus spp., and Leuconostoc spp. are inherently resistant to glycopeptides, but these organisms are seldom implicated in disease.
This compound is widely used for the treatment of infections caused by staphylococci that are resistant to methicillin and other β-lactam antibiotics, and for serious infections with Gram-positive organisms in patients who are hypersensitive to penicillin. It is also effective in the treatment of antibiotic-associated diarrhoea caused by toxigenic strains of Clostridium difficile (p. 251), but such use is discouraged because of the fear of undermining the value of this compound by promoting the emergence of resistance in Gram-positive cocci.
Early preparations of vancomycin contained impurities that gave the drug a reputation for toxicity. The purified formulations now available are much safer, but renal and ototoxicity still occur, particularly with high dosage. The drug is given by slow intravenous infusion to avoid ‘red man syndrome’ (p. 203).
This is a naturally occurring mixture of several closely related compounds with a spectrum of activity similar to that of vancomycin, although some coagulase-negative staphylococci are less susceptible to teicoplanin. Some strains of enterococci that are resistant to vancomycin (those with the VanB phenotype) retain susceptibility to teicoplanin. Unlike vancomycin, teicoplanin can be administered by intramuscular injection; it also has a much longer plasma half-life than vancomycin and appears to have a reduced propensity to cause adverse reactions.
Penicillins, cephalosporins, and certain other antibiotics belong to a family of compounds, collectively known as β-lactam antibiotics, which share the stuctural feature of a β-lactam ring. In the penicillins the β-lactam ring is fused to a five-membered thiazolidine ring, whereas the cephalosporins display a fused β-lactam/ dihydrothiazine ring structure (Fig. 1.3). The β-lactam ring is the Achilles’ heel of this group of antibiotics because many bacteria possess enzymes (β-lactamases; see p. 146) that are capable of breaking open the ring and rendering the molecule antibacterially inactive.
Fig. 1.3 Structures of benzylpenicillin and cephalosporin C, forerunners of the penicillin and cephalosporin groups, respectively. The fused-ring systems and the side-chains which offer the possibility of modifications introduced in semi-synthetic derivatives are indicated.
The original preparations of penicillin were found on analysis to be mixtures of four closely related compounds which were called penicillin F, G, K, and X. Benzylpenicillin (penicillin G), often simply called ‘penicillin’, was chosen for further development because it exhibited the most attractive properties and because a manufacturing process was developed in which Penicillium chryso-genum was persuaded to produce benzylpenicillin almost exclusively.
Early attempts to modify this structure relied on presenting the Penicillium mould used to produce penicillin with different side-chain precursors during the
manufacturing process. Later a method was discovered of removing the acyl side-chain of benzylpenicillin to liberate the penicillin nucleus, 6-aminopenicil-lanic acid (6-APA). Various chemical groupings could then be added to 6-APA according to the ingenuity of the chemist; a large number of compounds, collectively called semi-synthetic penicillins, have been prepared in this way.
Benzylpenicillin revolutionized the treatment of many potentially lethal bacterial infections, such as scarlet fever, puerperal sepsis, bacterial endocarditis, pneumococcal pneumonia, staphylococcal sepsis, meningococcal meningitis, gonorrhoea, syphilis (and other spirochaetal diseases), anthrax, and many anaerobic infections. The overwhelming importance of benzylpenicillin as a major breakthrough in therapy may be gauged from the fact that it remains today the treatment of choice for all these diseases.
However, resistance has eroded the value of benzylpenicillin. Nearly all staphylococci and many strains of gonococci are now resistant. Moreover, pneumococci exhibiting reduced susceptibility to benzylpenicillin are increasingly prevalent. Such strains are of two types: those for which the minimum inhibitory concentration (MIC) of benzylpenicillin is increased from the usual value of about 0.02 mg/l to 0.1–1 mg/l, and those for which the MIC exceeds 1 mg/l. The former are sufficiently sensitive to enable the antibiotic to be successfully used in high dosage, except in pneumococcal meningitis. However, penicillin is not clinically reliable in infections with strains exhibiting the higher level of resistance.
Despite its attractive properties benzylpenicillin is not the perfect antimicrobial agent: it exhibits a restricted antibacterial spectrum; it causes hypersensitivity reactions in a small proportion of persons to whom it is given; it is broken down by gastric acidity when administered orally; it is eliminated from the body at a spectacular rate by the kidneys; and it is hydrolysed by β-lactamases produced by many bacteria, including staphylococci. Subsequent developments have been aimed at overcoming these inherent disabilities while retaining the attractive properties of benzylpenicillin: high intrinsic activity and lack of toxicity.
The first major success in improving the pharmacological properties of penicillin was achieved with phenoxymethylpenicillin (penicillin V). This compound has properties very similar to those of benzylpenicillin, but it is acid stable and thus achieves better and more reliable serum levels when given orally, at the expense of some intrinsic antibacterial activity. Azidocillin, phenethicillin, and propicillin exhibit similar properties, but are not widely used.
Prolongation of plasma levels
Most β-lactam antibiotics are rapidly excreted, with a plasma half-life of 1–3 h. Benzylpenicillin is even more rapidly eliminated and several strategies are used in order to maintain effective levels in the body. The blockbuster approach is simply to give enormous doses of this non-toxic drug. Alternatively, oral probenecid
can be administered with the penicillin. Probenecid competes for sites of active tubular secretion in the kidney, slowing down the elimination of penicillin. Another solution is to use insoluble derivatives of penicillin. These are injected intramuscularly and act as depots from which penicillin is slowly liberated. Originally, mixtures of penicillin with oily or waxy excipients were used, but insoluble salts, such as procaine penicillin, were later developed. In this way an inhibitory concentration of penicillin can be maintained in the bloodstream for up to 24 h; extremely insoluble salts, such as benzathine or benethamine penicillin, release penicillin even more slowly, but the concentrations achieved are, of course, correspondingly lower.
Extension of spectrum
Broadening the spectrum of benzylpenicillin to encompass Gram-negative bacilli was first achieved by adding an amino group to the side-chain to form ampicillin. Ampicillin is slightly less active than benzylpenicillin against Gram-positive cocci and is equally susceptible to staphylococcal β-lactamase. However, it displays much improved activity against some enterobacteria, including Escher-ichia coli, Salmonella enterica, and Shigella spp. as well as against Haemophilus influenzae. Oral absorption is relatively poor, but can be improved by esterifying the molecule to form so-called pro-drugs, such as pivampicillin. Such compounds are split by non-specific tissue esterases in the intestinal mucosa to release ampi-cillin during absorption. Improved absorption has also been achieved by a minor modification to the molecule to produce amoxycillin.
A change of spectrum was brought about by altering the form of the linkage at the 6-position of the penicillanic acid nucleus to amidino (N–CH=N) instead of acyl (CO–NH). The only penicillin of this type to become available, mecil-linam (known as amdinocillin in the US), is active against ampicillin-sensitive enterobacteria and some of the more resistant Gram-negative rods. However, mecillinam displays no useful activity against Gram-positive cocci. It is poorly absorbed when given orally, but a pro-drug form, pivmecillinam, can be given by mouth.
Temocillin, a penicillin in which the β-lactam ring carries a stabilizing methoxy group (as in cephamycins; see p. 23), is not hydrolysed by most β-lactamases elaborated by Gram-negative bacilli, but has no useful activity against Gram-positive or anaerobic organisms.
None of the agents so far mentioned has any activity against Pseudomonas aerug-inosa, an important opportunist pathogen, especially in burns, cystic fibrosis, and immunocompromised patients. However, carbenicillin, a simple carboxyl derivative of benzylpenicillin, and ticarcillin, the thienyl variant of carbenicillin, exhibit useful antipseudomonal activity. Somewhat more active are a group of ureido derivatives of ampicillin including azlocillin and piperacillin. These antipseudomonal
penicillins must be administered by injection, but two esterified pro-drugs of carbenicillin, carfecillin and carindacillin, are available in some countries.
By the end of the 1950s 80 per cent of staphylococci isolated in hospitals were resistant to benzylpenicillin because of their ability to produce penicillinase (β-lactamase). The appearance of these resistant organisms, which often gave rise to serious cross-infection problems, stimulated research into derivatives that were insusceptible to β-lactamase hydrolysis. Success was achieved with methicillin, nafcillin and the isoxazolylpenicillins: oxacillin, cloxacillin, dicloxacillin, and flucloxacillin. The isoxazolylpenicillins, particularly flucloxacillin, are well absorbed when given orally and are most widely used. They are highly bound to serum protein in the body (see p. 188), but this does not seem to affect their therapeutic efficacy.
Resistance to penicillinase-stable penicillins is caused not by inactivating enzymes, but by alterations to the penicillin target (p. 152). Staphylococci of this type were originally characterized by resistance to methicillin and, although this compound is no longer used in treatment, they are still known as methi-cillin-resistant staphylococci. Resistance extends to all β-lactam agents and often accompanies resistance to gentamicin and other antibiotics (multi-resistant staphylococci). Some strains fully display the resistance phenotype only at a reduced growth temperature or in the presence of high salt concentrations. Particularly troublesome are methicillin-resistant Staphylococcus aureus strains (MRSA), which cause persistent problems in some units; some strains have a propensity to spread to give rise to mini-epidemics (EMRSA).
The spectrum of activity of the most important penicillins in clinical use is shown in Table 1.1.
Table 1.1 Summary of the antibacterial properties of selected penicillins
Cephalosporins generally exhibit a somewhat broader spectrum than penicillins, though, idiosyncratically, they lack activity against enterococci. They are mostly stable to staphylococcal β-lactamase and lack cross-allergenicity with penicillins (see p. 204).
The original cephalosporin, cephalosporin C, was never marketed, but has given rise to a large family of compounds that continues to expand. The extra carbon atom in the fused ring (Fig. 1.3), offers the possibility of modifications at the C-3 position. Alterations at either end of the molecule may profoundly affect the antibacterial activity but, as a generalization, substituents at the C-3 position have more effect on pharmacokinetic properties. Cephalosporins such as cephalothin and cefotaxime, which have an acetoxymethyl group at the C-3 position, are slowly deacetylated by liver enzymes to a hydroxymethyl form; this is usually less active and may display altered pharmacokinetic behaviour, but there is little evidence that the clinical effectiveness is impaired.
Several cephalosporins, including cephamandole, cefotetan, cefmenoxime, cefoperazone, and the oxa-cephem latamoxef possess a methyltetrazole-thiomethyl side-chain at the C-3 position. This substituent has been implicated in haematological side-effects in some patients (see p. 210).
The earliest cephalosporins, cephalothin and cephaloridine, are not absorbed when given orally. Moreover, it soon became clear that the Gram-negative organisms within their spectrum were capable of elaborating a wide variety of enzymes that exhibited potent cephalosporinase activity (see p. 146). Consequently, as with penicillins, developments within the cephalosporin family were aimed at devising compounds with more attractive properties: oral absorption or other improved pharmacological properties; stability to inactivating enzymes; better intrinsic activity; or a combination of these features.
Cephalosporins are commonly described as first, second, third, or even fourth generation compounds. These loose terms, which are best avoided, refer to:
In fact, the cephalosporins display such diverse properties that they defy any rigid categorization, but it is helpful to distinguish between those (the majority) that have to be administered parenterally and those that can be given orally. Among injectable compounds, it is useful to consider separately those with improved β-lactamase stability and those notable for their antipseudomonal activity (Table 1.2).
Table 1.2 Categorization of cephalosporins in clinical use
Parenteral compounds hydrolysed by enterobacterial β-lactamases
Cephalosporins in this group are of limited clinical value and have been largely superseded by other derivatives. Only two are presently available in the UK: cephazolin, which has the unusual property of being excreted in fairly high concentration in bile, and cephamandole, which exhibits a modestly expanded spectrum. Others, including cefacetrile, cefapirin, ceforanide and cefonicid, offer no discernible advantage over earlier congeners such as cephaloridine and cephalothin.
Parenteral compounds with improved β-lactamase stability
An important advance was achieved with the development of cephalosporins that exhibit almost complete stability to the common enterobacterial β-lactamases.
The first of these were cefuroxime and cefoxitin, the latter being one of a group of cephalosporins, collectively called cephamycins, which have a stabilizing methoxy grouping on the β-lactam ring. Other cephamycins available in some countries include cefotetan, cefmetazole, and cefminox. The cephamycins are unusual in displaying useful activity against anaerobes of the Bacteroides fragilis group.
These compounds have been somewhat overshadowed by the appearance of cephalosporins that combine almost complete stability to most β-lactamases with exceptional intrinsic activity. Cefotaxime was the forerunner of this group of compounds, but several others are available: ceftizoxime and cefmenoxime are similar to cefotaxime; ceftriaxone displays a sufficiently long plasma half-life to warrant once-daily administration; cefodizime is said to possess immuno-modulating properties.
Latamoxef (moxalactam), which is strictly an oxa-cephem (see below), also displays activity analogous to that of cefotaxime and its relatives, but differs in possessing useful activity against B. fragilis and related anaerobes. However, latamoxef has lost favour owing to toxicity problems and it is no longer widely used.
Compounds distinguished by antipseudomonal activity
Ps. aeruginosa is not susceptible to most cephalosporins and, as with penicillins, considerable efforts have been made to find derivatives that include this important opportunist pathogen in their spectrum. Some cephalosporins with anti-pseudomonal activity, such as cefoperazone, cefpimizole, and cefpiramide, are not distinguished by any unusual activity against other organisms and cefsulodin is extraordinary in being virtually inactive against bacteria other than Ps. aerugi-nosa. Among other derivatives, ceftazidime, cefpirome, and cefepime exhibit broad-spectrum activity comparable to that of cefotaxime and its congeners. These compounds have established a useful role in the management of Ps. aerugi-nosa infections in seriously ill patients. However, the antistaphylococcal activity is suspect and cefpirome may have some advantage in this respect. Cefepime retains activity against some opportunist Gram-negative bacilli that develop resistance to cefotaxime and its relatives.
Early development of the cephalosporins yielded cephalexin, a compound that exhibits modest activity, particularly in terms of its bactericidal action against Gram-negative bacilli, but which is virtually completely absorbed when given orally. Many other oral derivatives are structurally minor variations on the cephalexin theme. Such compounds include cephradine (the properties of which are indistinguishable from those of cephalexin), cefaclor (which is more active against the important respiratory pathogen H. influenzae), cefadroxil (which exhibits a modestly extended plasma half-life) and cefprozil (which exhibits
improved intrinisic activity). Loracarbef is a carbacephem (carbon replacing sulphur in the fused-ring structure), but is otherwise structurally identical to cefa-clor. Not surprisingly, its properties closely resemble those of cefaclor.
Cefixime and ceftibuten are structurally unrelated to cephalexin. They display much improved activity against most Gram-negative bacilli, but at the expense of antistaphylococcal (and, in the case of ceftibuten, antipneumococcal) activity, which is very poor. Another compound of this type, cefdinir, appears to lack these defects.
The principle of esterification to produce pro-drugs with improved oral absorption has also been applied to cephalosporins. Two such compounds, cefuroxime axetil and cefpodoxime proxetil, are available in the UK; cefteram pivoxil, cefetamet pivoxil, cefotiam hexetil and cefditoren pivoxil are marketed elsewhere. These esters are fairly well absorbed by the oral route and deliver the parent drug into the bloodstream. Cefpodoxime, cefteram and cefetamet are more active than the others against most organisms within the spectrum, although cefetamet has poor activity against staphylococci.
A summary of the antimicrobial spectrum of the most important cephalosporins is presented in Table 1.3.
Table 1.3 Summary of the spectrum of antibacterial activity of cephalosporins available in the UK (2000)
Other β-lactam agents
In addition to penicillins and cephalosporins, various other compounds display a β-lactam ring in their structure (Fig. 1.4). The cephamycins, the oxa-cephems latamoxef and flomoxef, and the carbacephem loracarbef—all of which share the general properties of cephalosporins—are examples of such structural variants. Fundamentally different are clavulanic acid, a naturally occurring substance obtained fromStreptomyces clavuligerus, and two penicillanic acid sulphones, sulbactam and tazobactam. These compounds have little useful antibacterial activity, but act as β-lactamase inhibitors. They are used in combination with β-lactamase-labile agents with a view to restoring their activity (see p. 120). Structurally novel compounds that exhibit antibacterial activity in their own right include the carbapenems (imipenem, meropenem, panipenem, and related investigational compounds); aztreonam, one of a group of compounds, collectively known as monobactams, which have a β-lactam ring but no associated fused-ring system; and the trinems (tribactams), which have a tricyclic fused-ring structure.
Fig. 1.4 Basic molecular structures of β-lactam antibiotics currently available (examples in parentheses).
The carbapenems are stable to most bacterial β-lactamases, and exhibit the broadest spectrum of all β-lactam antibiotics, with high activity against nearly all Gram-positive and Gram-negative bacteria other than intracellular bacteria such as chlamydiae. However, imipenem (but not meropenem) is readily hydrolysed by a dehydropeptidase located in the mammalian kidney and is administered together with a dehydropeptidase inhibitor, cilastatin. Aztreonam is also β-lactamase stable, but, in contrast to carbapenems, the activity is restricted to aerobic Gram-negative bacteria. The trinems are presently under development.
One such compound, sanfetrinem, appears to be very broad spectrum and stable to many β-lactamases. It can be given orally as the hexetil ester.
Mode of action of β-lactam agents
All β-lactam antibiotics interfere with the final transpeptidation reaction that gives the cell wall its strength (Fig. 1.1). However, several forms of transpep-tidase, some of which also have transglycosylase activity, are needed to maintain the molecular architecture of the cell and these are differentially inhibited by various β-lactam agents. These target enzymes belong to a group of proteins to which penicillin and other β-lactam antibiotics bind (penicillin-binding proteins; PBPs). Esch. coli, the best-studied species, has seven of these proteins, numbered la, lb, 2, 3, 4, 5, and 6 in order of decreasing molecular weight. PBPs 4–6 are thought to be unconnected with the antibacterial effect of β-lactam agents, since mutants lacking these proteins do not seem to be disabled in any way. Binding to the remainder has been correlated with the various morphological effects of β-lactam antibiotics on Gram-negative bacilli. Thus, cephalexin and its close congeners, as well as aztreonam and temocillin, bind almost exclusively to PBP
3 and inhibit the division process, causing the bacteria to grow as long filaments. The amidinopenicillin, mecillinam, binds preferentially to PBP 2 and causes a generalized effect on the cell wall so that the bacteria gradually assume a spherical shape. Most other β-lactam antibiotics bind to PBPs 1–3 and induce the formation of osmotically fragile, wall-deficient forms (spheroplasts) which typically emerge through cell wall lesions situated at incipient division points. The morphological events are illustrated in Fig. 1.5.
Fig. 1.5 Morphological effects of penicillins and cephalosporins on Gram-negative bacilli (scanning electron micrographs): (A) NormalEsch. coli cells; (B) Esch. coli exposed to cephalexin, 32 mg/l, for 1 h; (C) Esch. coli exposed to mecillinam, 10 mg/l, for 2 h; (D) Esch. coli exposed to ampicillin, 64 mg/l, for 1 h, showing lysed debris, central wall lesions, and a spheroplast; higher concentrations of most β-lactam antibiotics cause this effect. (A–C from Greenwood D and O'Grady F, Journal of Infectious Diseases 128 (1973), 791–4; D from Greenwood D and O'Grady F, Journal of Medical Microbiology 2 (1969), 435–41.)
In Gram-negative bacilli, rupture of spheroplasts can be quantitatively prevented by raising the osmolality of the growth medium, so cell death appears to be an
osmotic phenomenon. Agents such as cephalexin and mecillinam that do not cause spheroplast formation are much more slowly bactericidal than others. The lethal event in Gram-positive organisms, which have much thicker cell walls, appears to be autolysis triggered by the release of lipoteichoic acid following exposure to β-lactam antibiotics.
Optimal dosage effect
A further complication in Gram-positive organisms is that increasing the concentration of β-lactam antibiotics often results in a reduced bactericidal effect. The mechanism of this effect (known as the Eagle phenomenon after its discoverer) is obscure, but may be related to the to the multiple sites of penicillin action, in that rapid bacteristasis achieved by blocking one cellular function may prevent the lethal events that normally follow inhibition of another by lower drug levels.
Persisters and penicillin tolerance
In both Gram-positive and Gram-negative bacteria, a proportion of the population, called persisters, survive exposure to concentrations of β-lactam antibiotics lethal to the rest of the culture. They remain dormant so long as the antibiotic is present and resume growth when it is removed. In addition, some strains of staphylococci and streptococci display tolerance to β-lactam antibiotics in that they succumb much more slowly than usual to the lethal action of β-lactam agents. The therapeutic significance, if any, of persisters is unknown, but penicillin tolerance has been implicated in therapeutic failures in bacterial endocarditis where bactericidal activity is crucial to the success of treatment (p. 264).
Much has also been made of laboratory observations that the antimicrobial activity of β-lactam agents may persist for an hour or more after the drug is removed. This effect is not confined to β-lactam agents and is more consistently demonstrated with Gram-positive than with Gram-negative organisms. Theoretically, knowledge of post-antibiotic effects might influence the design of dosage regimens, but in practice they are too erratic to be used in this way, even if the in-vitro observations could be convincingly shown to have clinical relevance, which is presently not the case.