Handbook of Cancer Chemotherapy (Lippincott Williams & Wilkins Handbook Series), 8th Ed.

1. Biologic and Pharmacologic Basis of Cancer Chemotherapy

Roland T. Skeel


The purpose of treating cancer with chemotherapeutic agents is to prevent cancer cells from multiplying, invading, metastasizing, and ultimately killing the host (patient). Most traditional chemotherapeutic agents currently in use appear to exert their effect primarily on cell proliferation. Because cell multiplication is a characteristic of many normal cells as well as cancer cells, most nontargeted cancer chemotherapeutic agents also have toxic effects on normal cells, particularly those with a rapid rate of turnover, such as bone marrow and mucous membrane cells. The goal in selecting an effective drug, therefore, is to find an agent that has a marked growth-inhibitory or controlling effect on the cancer cell and a minimal toxic effect on the host. In the most effective chemotherapeutic regimens, the drugs are capable not only of inhibiting but also of completely eradicating all neoplastic cells while sufficiently preserving normal marrow and other target organs to permit the patient to return to normal, or at least satisfactory, function and quality of life.

Ideally, the cell biologist, pharmacologist, and medicinal chemist would like to look at the cancer cell, discover how it differs from the normal host cell, and then design a chemotherapeutic agent to capitalize on that difference. Until the last decade, less rational means were used for most of the chemotherapeutic agents that are now in use. The effectiveness of agents was discovered by treating either animal or human neoplasms, after which the pharmacologist attempted to discover why the agent worked as well as it did. With few exceptions, the reasons that traditional chemotherapeutic agents are more effective against cancer cells than against normal cells have been poorly understood. With the rapid expansion of information about cell biology and the factors within the neoplastic cell that control cell growth, the strictly empiric method of discovering effective new agents has changed. For example, antibodies against the protein product of the overexpressed HER2/neu oncogene have been demonstrated to be effective in controlling metastatic breast cancer and reducing recurrences after primary therapy in patients whose tumors overexpress this gene. Discovery of the constitutively activated Bcr-Abl tyrosine kinase created as a consequence of the chromosomal translocation in chronic myelogenous leukemia has led to a burgeoning era of orally administered small molecular inhibitors of antibodies targeting critical molecular changes in cancer cells and their environment. These sentinel events have presaged the development of a host of new therapeutic agents that are directed at known specific targets within and around the cancer cell. These targets have been selected because they are altered in the cancer cell and are critical for cancer cell growth, invasion, and metastasis. This increased understanding of cancer cell biology has already provided more specific and selective ways of controlling cancer cell growth in several human cancers and will continue to dominate systemic therapy drug development in the decades to come.

Inhibition of cell multiplication and tumor growth can take place at several levels within the cell and its environment:

A. Classic chemotherapy agents

Most agents currently in use, with the exception of immunotherapeutic agents, other biologic response modifiers, and molecular targeted therapies appear to have their primary effect on either macromolecular synthesis or function. This effect means that they interfere with the synthesis of DNA, RNA, or proteins or with the appropriate functioning of the preformed molecule. When interference in macromolecular synthesis or function in the neoplastic cell population is sufficiently great, a proportion of the cells die. Some cells die because of the direct effect of the chemotherapeutic agent. In other instances, the chemotherapy may trigger differentiation, senescence, or apoptosis, the cell’s own mechanism of programmed death.

Cell death may or may not take place at the time of exposure to the drug. Often, a cell must undergo several divisions before the lethal event that took place earlier finally results in the death of the cell. Because only a proportion of the cells die as a result of a given treatment, repeated doses of chemotherapy must be used to continue to reduce the cell number (Fig 1.1). In an ideal system, each time the dose is repeated, the same proportion of cells—not the same absolute number—is killed. In the example shown in Figure 1.1, 99.9% (3 logs) of the cancer cells are killed with each treatment, and there is a 10-fold (1-log) growth between treatments, for a net reduction of 2 logs with each treatment. Starting at 1010 cells (about 10 g or 10 cm3 leukemia cells), it would take five treatments to reach fewer than 100, or 1, cell. Such a model makes certain assumptions that rarely are strictly true in clinical practice:

bull All cells in a tumor population are equally sensitive to a drug.

bull Drug accessibility and cell sensitivity are independent of the location of the cells within the host and of local host factors such as blood supply and surrounding fibrosis.

bull Cell sensitivity does not change during the course of therapy.

The lack of curability of most initially sensitive tumors is probably a reflection of the degree to which these assumptions do not hold true.


FIGURE 1.1 The effect of chemotherapy on cancer cell numbers. In an ideal system, chemotherapy kills a constant proportion of the remaining cancer cells with each dose. Between doses, cell regrowth occurs. When therapy is successful, cell killing is greater than cell growth.

B. Biologic response modifiers and molecular targeted therapy

Within individual cells and cell populations are intricate interrelated mechanisms that promote or suppress cell proliferation, facilitate invasion or metastasis when the cell is malignant, lead to cell differentiation, promote (relative) cell immortality, or set the cell on the path to inevitable death (apoptosis). These activities are controlled in large part by normal genes and, in the case of cancer, by mutated cancer promoter genes, tumor suppressor genes, and their products. Included in these products are a host of cell growth factors that control the machinery of the cell. Some of these factors that affect normal cell growth have been biosynthesized and are now used to enhance the production of normal cells (e.g., epoetin alfa and filgrastim) and to treat cancer (e.g., interferon).

The recent expansion of our understanding of the biologic control of normal cells and tumor growth at the molecular level has only begun to offer improved therapy for cancer, though it has helped to explain differences in response among populations of patients. New discoveries in cancer cell biology have provided insights into apoptosis, cell cycling control, angiogenesis, metastasis, cell signal transduction, cell surface receptors, differentiation, and growth factor modulation. New drugs in clinical trials have been designed to block growth factor receptors, prevent oncogene activity, block the cell cycle, restore apoptosis, inhibit angiogenesis, restore lost function of tumor suppressor genes, and selectively kill tumors containing abnormal genes. Further understanding of each of these holds a great potential for providing powerful and more selective means to control neoplastic cell growth and may lead to more effective cancer treatments in the next decade. The fundamental principles related to this group of antineoplastic agents is discussed in Chapter 2.


Cancer cells, unlike other body cells, are characterized by a growth process whereby their sensitivity to normal controlling factors has been partially or completely lost. As a result of this uncontrolled growth, it was once thought that cancer cells grew or multiplied faster than normal cells and that this growth rate was responsible for the sensitivity of cancer cells to chemotherapy. Now it is known that most cancer cells grow less rapidly than the more active normal cells such as bone marrow. Thus, although the growth rate of many cancers is faster than that of normal surrounding tissues, growth rate alone cannot explain the greater sensitivity of cancer cells to chemotherapy.

A. Tumor growth

The growth of a tumor depends on several interrelated factors.

1. Cell cycle time, or the average time for a cell that has just completed mitosis to grow, redivide, and again pass through mitosis, determines the maximum growth rate of a tumor but probably does not determine drug sensitivity. The relative proportion of cell cycle time taken up by the DNA synthesis phase may relate to the drug sensitivity of some types (synthesis phase–specific) of chemotherapeutic agents.

2.Growth fraction, or the fraction of cells undergoing cell division, contains the portion of cells that are sensitive to drugs whose major effect is exerted on cells that are dividing actively. If the growth fraction approaches 1 and the cell death rate is low, the tumor-doubling time approximates the cell cycle time.

3.Total number of cells in the population (determined at some arbitrary time at which the growth measurement is started) is clinically important because it is an index of how advanced the cancer is; it frequently correlates with normal organ dysfunction. As the total number of cells increases, so does the numberof resistant cells, which in turn leads to decreased curability. Large tumors may also have greater compromise of blood supply and oxygenation, which can impair drug delivery to the tumor cells as well as impair sensitivity to both chemotherapy and radiotherapy.

4. Intrinsic cell death rate of tumors is difficult to measure in patients but probably makes a major and positive contribution by slowing the growth rate of many solid tumors.

B. Cell cycle

The cell cycle of cancer cells is qualitatively the same as that of normal cells (Fig. 1.2). Each cell begins its growth during a postmitotic period, a phase called G1, during which enzymes necessary for DNA production, other proteins, and RNA are produced. G1 is followed by a period of DNA synthesis, in which essentially all DNA synthesis for a given cycle takes place. When DNA synthesis is complete, the cell enters a premitotic period (G2), during which further protein and RNA synthesis occurs. This gap is followed immediately by mitosis, bat the end of which actual physical division takes place, two daughter cells are formed, and each cell again enters G1. G1phase is in equilibrium with a resting state called G0. Cells in G0 are relatively inactive with respect to macromolecular synthesis and are consequently insensitive to many traditional chemotherapeutic agents, particularly those that affect macromolecular synthesis.


FIGURE 1.2 Cell cycle time for human tissues has awide range (16 to 260 hours), with marked differences among normal and tumor tissues. Normal marrow and gastrointestinal lining cells have cell cycle times of 24 to 48 hours. Representative durations and the kinetic or synthetic activity are indicated for each phase.

C. Phase and cell cycle specificity

Most classic chemotherapeutic agents can be grouped according to whether they depend on cells being in cycle (i.e., not in G0) or, if they depend on the cell being in cycle, whether their activity is greater when the cell is in a specific phase of the cycle. Most agents cannot be assigned to one category exclusively. Nonetheless, these classifications can be helpful for understanding drug activity.

1. Phase-specific drugs. Agents that are most active against cells in a specific phase of the cell cycle are called cell cycle phase–specific drugs. A partial list of these drugs is shown in Table 1.1.

a. Implications of phase-specific drugs. Phase specificity has important implications for cancer chemotherapy.

(1) Limitation to single-exposure cell kill. With a phase-specific agent, there is a limit to the number of cells that can be killed with a single instantaneous (or very short) drug exposure because only those cells in the sensitive phase are killed. A higher dose kills no more cells.


(2) Increasing cell kill by prolonged exposure. To kill more cells requires either prolonged exposure to, or repeated doses of, the drug to allow more cells to enter the sensitive phase of the cycle. Theoretically, all cells could be killed if the blood level or, more importantly, the intracellular concentration of the drug remained sufficiently high while all cells in the target population passed through one complete cell cycle. This theory assumes that the drug does not prevent the passage of cells from one (insensitive) phase to another (sensitive) phase.

(3) Recruitment. A higher number of cells could be killed by a phase-specific drug if the proportion of cells in the sensitive phase could be increased (recruited).

b. Cytarabine. One of the best examples of a phase-specific agent is cytarabine (ara-C), which is an inhibitor of DNA synthesis and thus is active only in the synthesis phase (at standard doses). When used in doses of 100–200 mg/m2 daily (i.e., not “high-dose ara-C”), ara-C is rapidly deaminated in vivo to an inactive compound, ara-U, and rapid injections result in very short effective levels of ara-C. As a result, single doses of ara-C are non-toxic to the normal hematopoietic system and are generally ineffective for treating leukemia. If the drug is given as a daily rapid injection, some patients with leukemia respond well but not nearly as well as when ara-C is given every 12 hours. The apparent reason for the greater effectiveness of the 12-hour schedule is that the synthesis phase (DNA synthesis) of human acute myelogenous leukemia cells lasts about 18 to 20 hours. If the drug is given every 24 hours, some cells that have not entered the synthesis phase when the drug is first administered will not be sensitive to its effect. Therefore, these cells can pass all the way through the synthesis phase before the next dose is administered and will completely escape any cytotoxic effect. However, when the drug is given every 12 hours, no cell that is “in cycle” will be able to escape exposure to ara-C because none will be able to get through one complete synthesis phase without the drug being present.

If all cells were in active cycle, that is, if none were resting in a prolonged G1 or G0 phase, it would be theoretically possible to kill any cells in a population by a continuous or scheduled exposure equivalent to one complete cell cycle. Experiments with patients who have acute leukemia have shown that if tritiated thymidine is used to label cells as they enter DNA synthesis, it may be 7 to 10 days before the maximum number of leukemia cells have passed through the synthesis phase. This means that, barring permutations caused by ara-C or other drugs, for ara-C to have a maximum effect on the leukemia, the repeated exposure must be continued for a 7- to 10-day period. Clinically, continuous infusion or administration of ara-C every 12 hours for 5 days or longer appears to be most effective for treating patients with newly diagnosed acute myelogenous leukemia. However, even with such prolonged exposure, it appears that a few of the cells do not pass through the synthesis phase.

2. Cell cycle-specific drugs. Agents that are effective while cells are actively in cycle but that are not dependent on the cell being in a particular phase are called cell cycle-specific (or phasenonspecificdrugs.This group includes most of the alkylating agents, the antitumor antibiotics, and some miscellaneous agents, examples of which are shown in Table 1.2. Some agents in this group are not totally phase–nonspecific; they may have greater activity in one phase than in another, but not to the degree of the phase-specific agents. Many agents also appear to have some activity in cells that are not in cycle, although not as much as when the cells are rapidly dividing.

3. Cell cycle-nonspecific drugs. A third group of drugs appears to be effective whether cancer cells are in cycle or are resting. In this respect, these agents are similar to photon irradiation; that is, both types of therapy are effective irrespective of whether or not the cancer cell is in cycle. cell cycle–nonspecific drugs and include mechlorethamine (nitrogen mustard) and the nitrosoureas (see Table 1.2).

D Changes in tumor cell kinetics and therapy implications

As cancer cells grow from a few cells to a lethal tumor burden, certain changes occur in the growth rate of the population and affect the strategies of chemotherapy. These changes have been determined by observing the characteristics of experimental tumors in animals and neoplastic cells growing in tissue culture. Such model systems readily permit accurate cell number determinations to be made and growth rates to be determined. (Because tumor cells cannot be injected or implanted into humans and permitted to grow, studies of growth rates of intact tumors in humans must be limited largely to observing the growth rate of macroscopic tumors.)


1. Stages of tumor growth. Immediately after inoculation of a tissue culture or an experimental animal with tumor cells, there is a lag phase, during which there is little tumor growth; presumably, the cells in this phase are becoming accustomed to the new environment and are preparing to enter into cycle. The lag phase is followed by a period of rapid growth called the log phase, during which there are repeated doublings of the cell number. In populations in which the growth fraction approaches 100% and the cell death rate is low, the population doubles within a period approximating the cell cycle time. As the cell number or tumor size becomes macroscopic, the doubling time of the tumor cell population becomes prolonged and levels off (plateau phase). Most clinically measurable human cancers are probably in the plateau phase, which may account, in part, for the slow doubling time observed in many human cancers (30 to 300 days). Because the rate of change in the slope of the growth curve during the premeasurable period is unknown for most human cancers, extrapolation from two points when the mass is measurable to estimate the onset of the growth of the malignancy is subject to considerable error. The prolongation in tumor-doubling time in the plateau phase may be due to a smaller growth fraction, a change in the cell cycle time, an increased intrinsic death rate (predominantly apoptosis, which is a programmed and highly orchestrated cell death that occurs both naturally and under the influence of many types of chemotherapy), or a combination of these factors. Factors responsible for these changes include decreased nutrients or growth promotion factors, increased inhibitory metabolites or inhibitory growth factors, and inhibition of growth by other cell–cell interactions. In the intact host, new blood vessel formation is a critical determinant of these factors.

2. Growth rate and effectiveness of chemotherapy. Chemotherapeutic agents are most effective during the period of logarithmic growth. As might be expected, this result is particularly true for the antimetabolites, which are largely synthetic-phase specific. As a result, when human tumors become macroscopic, the effectiveness of many chemotherapeutic agents is reduced because only part of the cell population is dividing actively. Theoretically, if the cell population could be reduced sufficiently by other means such as surgery or radiotherapy, chemotherapy would be more effective because a higher fraction of the remaining cells would be in logarithmic growth. The validity of this theoretical premise is supported by the varying degrees of success of surgery plus chemotherapy or radiotherapy plus chemotherapy in the treatment of breast cancer, colon cancer, Wilms tumor, ovarian cancer, small cell anaplastic cell carcinoma of the lung, non–small-cell carcinoma of the lung, head and neck cancers, and osteosarcomas.


Combinations of drugs are frequently more effective in producing responses and prolonging life than are the same drugs used sequentially. Combinations are likely to be more effective than single agents for several reasons.

A. Reasons for effectiveness of combinations

1. Prevention of resistant clones. If 1 in 105 cells is resistant to drug A and 1 in 105 cells is resistant to drug B, it is likely that treating a macroscopic tumor (which generally would have more than 109 cells) with either agent alone would result in several clones of cells that are resistant to that drug. If, after treatment with drug A, a resistant clone has grown to macroscopic size (if the same mutant frequency persists for drug B), resistance to that agent will also emerge. If both drugs are used at the outset of therapy or in close sequence, however, the likelihood of a cell being resistant to both drugs (excluding, for a moment, the situation of pleiotropic drug resistance) is only 1 in 1010. Thus, the combination confers considerable advantage against the emergence of resistant clones. Compounding the problem of preexisting resistant clones is the resistance that develops through spontaneous mutation in the absence of drug exposure. The use of multiple drugs with independent mechanisms of action or alternating non–cross-resistant combinations (as well as the use of surgery or radiotherapy to eliminate macroscopic tumor) theoretically minimizes the chances for outgrowth of resistant clones and increases the likelihood of remission or cure.

2. Cytotoxicity to resting and dividing cells. The combination of a drug that is cell cycle–specific (phase–nonspecific) or cell cycle–non-specific with a drug that is cell cycle phase–specific can kill cells that are dividing slowly as well as those that are dividing actively. The use of cell cycle–nonspecific drugs can also help recruit cells into a more actively dividing state, which results in their being more sensitive to the cell cycle phase–specific agents.

3. Biochemical enhancement of effect

a. Combinations of individually effective drugs that affect different biochemical pathways or steps in a single pathway can enhance each other. This may apply to some newer agents whereby blocking more than one molecular target in the interacting signal transduction pathways may magnify the interference of cell proliferation compared with that seen with either agent alone.

b. Combinations of an active agent with an inactive agent can potentially result in beneficial effects by several mechanisms, but have limited clinical utility.

(1) An intracellular increase in the drug or its active metabolites, by either increasing influx or decreasing efflux (e.g., calcium channel inhibitors with multiple agents affected by multidrug resistance [MDR] due to P-glycoprotein overexpression).

(2) Reduced metabolic inactivation of the drug (e.g., inhibition of cytidine deaminase inactivation of ara-C with tetrahy-drouridine).

(3) Cooperative inhibition of a single enzyme or reaction (e.g., leucovorin enhancement of fluorouracil inhibition of thymidylate synthetase).

(4) Enhancement of drug action by inhibition of competing metabolites (e.g., N-phosphonacetyl-L-aspartic acid inhibition of de novo pyrimidine synthesis with resultant increased incorporation of 5-fluorouridine triphosphate into RNA).

4. Sanctuary access. Combinations can be used to provide access to sanctuary sites for reasons such as drug solubility or affinity of specific tissues for a particular drug type.

5. Rescue. Combinations can be used in which one agent rescues the host from the toxic effects of another drug (e.g., leucovorin administration after high-dose methotrexate).

B. Principles of agent selection

When selecting appropriate agents for use in a combination, the following principles should be observed.

1. Choose individually active drugs. Do not use a combination in which one agent is inactive when used alone unless there is a clear, specific biochemical or pharmacologic reason to do so, for example, high-dose methotrexate followed by leucovorin rescue or leucovorin followed by fluorouracil. This principle is not applicable to the combined use of chemotherapeutic agents with biologic response modifiers or molecular targeted agents because the cooperativity of chemotherapy and these drugs may not depend on the independent cytotoxic effect of these nonclassic agents.

2. When possible, choose drugs in which the dose-limiting toxicities differ qualitatively or in time of occurrence. Often, however, two or more agents that have marrow toxicity must be used, and the selection of a safe dose of each is critical. As a starting point, two cytotoxic drugs in combination can usually be given at two-thirds of the dose used when the drugs are given alone. Whenever a new drug combination is tried, a careful evaluation of both expected and unanticipated toxicities must be carried out. Unexpected results such as the increased cardiotoxicity of the combination of trastuzumab with doxorubicin may occur, and this latter case has precluded the use of these agents together.

3. Select agents for a combination for which there is a biochemical or pharmacologic rationale. Preferably, this rationale has been tested in an animal tumor system and in the appropriate model system, and the combination has been found to be better than either agent alone.

4. Be cautious when attempting to improve on a successful two-drug combination by adding a third, fourth, or fifth drug simultaneously. Although this approach may be beneficial, two undesirable results may be seen, as follows:

bull An intolerable level of toxicity that leads to excessive morbidity and mortality.

bull Unchanged or reduced antitumor effect because of the necessity to reduce the dose of the most effective drugs to a level below which antitumor responses are not seen, despite the theoretical advantages of the combination. Therefore, the addition of each new agent to a combination must be considered carefully, the principles of combination therapy closely followed, and controlled clinical trials carried out to compare the efficacy and toxicity of any new regimen with a more established (standard) treatment program.

C. Clinical effectiveness of combinations

Combinations of drugs have been clearly demonstrated to be better than single agents for treating many, but not all, human cancers. The survival benefit of combinations of drugs compared with that of the same drugs used sequentially has been marked in diseases such as acute lymphocytic and acute nonlymphocytic leukemia, Hodgkin lymphoma, non-Hodgkin lymphomas with more aggressive behavior (intermediate- and high-grade), breast carcinoma, anaplastic small-cell carcinoma of the lung, colorectal carcinoma, ovarian carcinoma, and testicular carcinoma. The benefit is less notable in cancers such as non–small-cell carcinoma of the lung, non-Hodgkin lymphomas with favorable prognoses, head and neck carcinomas, carcinoma of the pancreas, and melanoma, although reports exist for each of these tumors in which combinations are better in one respect or another than single agents.


Resistance to antineoplastic chemotherapy is a combined characteristic of a specific drug, a specific tumor, and a specific host whereby the drug is ineffective in controlling the tumor without excessive toxicity. Resistance of a tumor to a drug is the reciprocal of selectivity of that drug for that tumor. The problem for the medical oncologist is not simply to find an agent that is cytotoxic but to find one that selectively kills neoplastic cells while preserving the essential host cells and their function. Were it not for the problem of resistance of human cancer to antineoplastic agents or, conversely, the lack of selectivity of those agents, cancer chemotherapy would be similar to antibacterial chemotherapy in which complete eradication of infection is regularly observed. Such a utopian state of cancer chemotherapy has not yet been achieved for most human cancers. The problem of resistance, including ways to overcome or even exploit it, remains an area of major interest for the oncologist, pharmacologist, and cell biologist. This reductionist description glosses over the fact that each of these factors is a consequence of the complex genetic characteristics and changes of the cancer cell as it evolves.

Resistance to antineoplastic chemotherapeutic agents may be either natural or acquired. Natural resistance refers to the initial unresponsiveness of a tumor to a given drug, and acquired resistance refers to the unresponsiveness that emerges after initially successful treatment. There are three basic categories of resistance to chemotherapy: kinetic, biochemical, and pharmacologic.

A. Cell kinetics and resistance

Resistance based on cell population kinetics relates to cycle and phase specificity, growth fractions and the implications of these factors for responsiveness to specific agents, and schedules of drug administration. A particular problem with many human tumors is that they are in a plateau growth phase with a small growth fraction. This factor renders many of the cells insensitive to the antimetabolites and relatively unresponsive to many of the other chemotherapeutic agents. Strategies to overcome resistance due to cell kinetics include the following:

1. Reducing tumor bulk with surgery or radiotherapy

2. Using combinations to include drugs that affect resting populations (with many G0 cells)

3. Scheduling of drugs to prevent phase escape or to synchronize cell populations and increase cell kill.

B. Biochemical causes of resistance

Resistance can occur for biochemical reasons including the inability of a tumor to convert a drug to its active form, the ability of a tumor to inactivate a drug, or the location of a tumor at a site where substrates are present that bypass an otherwise lethal blockade. How cells become resistant is only partially understood. There can be decreased drug uptake, increased efflux, changes in the levels or structure of the intracellular target, reduced intracellular activation or increased inactivation of the drug, or increased rate of repair of damaged DNA. In one pre–B-cell leukemia cell line, bcl-2 overexpression or decreased expression of the homolog bax renders cells resistant to several chemotherapeutic agents.

Because bcl-2 blocks apoptosis, it has been proposed that its over-expression blocks chemotherapy-induced apoptosis. The interrelationship between mutations of p53, overexpression of HER2, and similar changes in a host of other oncogenes and tumor suppressor genes and resistance to the cytotoxic effects of radiotherapy and chemotherapeutic, hormonal, and biologic agents, when better understood, may further our understanding of resistance and provide new therapeutic strategies.

MDR, also called pleiotropic drug resistance, is a phenomenon whereby treatment with one agent confers resistance not only to that drug and others of its class but also to several other unrelated agents. MDR is commonly mediated by an enhanced energy-dependent drug efflux mechanism that results in lower intracellular drug concentrations. With this type of MDR, overexpression of a membrane transport protein called P-glycoprotein (“P” meaning pleiotropic or permeability) is observed commonly. Other MDR proteins are the MDR protein found in human lung cancer lines and the lung resistance protein. These proteins appear to have differing expression in different sets of neoplasms. Drugs that are effective in reversing resistance to P-glycoprotein do not reverse these latter MDR proteins. Combination chemotherapy can overcome biochemical resistance by increasing the amount of active drug intracellularly as a result of biochemical interactions or effects on drug transport across the cell membrane. Calcium channel blockers, antiarrhythmics, cyclosporine A analogs (e.g., PSC-833, a nonimmunosuppressive derivative of cyclosporine D), and other agents have been found to modulate the P-glycoprotein MDR effect in vitro, but limited beneficial effects have been observed clinically.

The use of a second agent to rescue normal cells may also permit the use of high doses of the first agent, which can overcome the resistance caused by a low rate of conversion to the active metabolite or a high rate of inactivation. Another way to overcome resistance is to follow marrow-lethal doses of chemotherapy by posttherapy infusion of stem cells obtained from the peripheral blood or bone marrow. This technique is effective for the treatment of some patients with lymphoma, leukemia, multiple myeloma, and a few other cancers. A more widely applicable technique is to combine higher or more frequent doses of chemotherapy with granulocyte-colony–stimulating factor or granulocyte–macrophage-colony–stimulating factor. These and other marrow-protective and marrow-stimulating agents are being used increasingly and may enhance the effectiveness of chemotherapy in the treatment of several types of cancer.

C. Pharmacologic causes of resistance

Apparent resistance to cancer chemotherapy can result from poor tumor blood supply, poor or erratic absorption, increased excretion or catabolism, and drug interactions, all leading to inadequate blood levels of the drug. Strictly speaking, this result is not true resistance; but to the degree that the insufficient blood levels are not appreciated by the clinician, resistance appears to be present. The variation from patient to patient at the highest tolerated dose has led to dose modification schemes that permit dose escalation when the toxicities of the chemotherapy regimen are minimal or nonexistent, as well as dose reduction when toxicities are great. This regulation is particularly important for some chemotherapeutic agents for which the dose–response curve is steep or for patients who have genetically altered drug metabolism, such as can occur with irinotecan. Selection of the appropriate dose on the basis of predicted pharmacologic behavior is essential for some agents not only to avoid serious toxicity but also to optimize effectiveness. This has been applied successfully to dose selection of carboplatin by predicting the time × concentration product (area under the curve) based on the individual patient’s creatinine clearance.

True pharmacologic resistance is caused by the poor transport of agents into certain body tissues and tumor cells. For example, the central nervous system (CNS) is a site that many drugs do not reach well. Several drug characteristics favor transport into the CNS, including high lipid solubility and low molecular weight. For tumors that originate in the CNS or metastasize there, the drugs of choice should be those that achieve effective antitumor concentration in the brain tissue and that are also effective against the tumor cell type being treated.

D. Nonselectivity and resistance

Nonselectivity is not a mechanism for resistance but rather an acknowledgment that for most cancers and most drugs, the reasons for resistance and selectivity are only partially understood. Given a limited understanding of the biochemical differences between normal and malignant cells prior to the last 10 years, it is gratifying that chemotherapy has been as successful as frequently as it has. With the burgeoning of knowledge about the cancer cell, there is reason to hope that in 20 years, we will view current chemotherapy regimens as a fledgling—if not crude—beginning and will have found many more tumor molecular target–directed agents that have a high potential for curing the human cancers that now resist effective treatment.

Selected Readings

Adjei AA, Hidalgo M. Intracellular signal transduction pathway proteins as targets for cancer therapy. J Clin Oncol. 2005;23:5386–5403.

Baguley BC, Holdaway KM, Fray LM. Design of DNA intercalators to overcome topoisomerase II-mediated multi-drug resistance. J Natl Cancer Inst. 1990;82:398–402.

Baserga R. The cell cycle. N Engl J Med. 1981;304:453–459.

Clarkson B, Fried J, Strife A, et al. Studies of cellular proliferation in human leukemia. 3. Behavior of leukemic cells in three adults with acute leukemia given continuous infusions of 3H-thymidine for 8 or 10 days. Cancer.1970;25:1237–1260.

Endicott JA, Ling U. The biochemistry of P-glycoprotein-mediated multidrug resistance. Annu Rev Biochem. 1989;58:137–171.

Friedland ML. Combination chemotherapy. In: Perry MC, ed. The chemotherapy source book. Baltimore: Williams & Wilkins, 1996:63–78.

Giaccone G. HERl/EGFR-targeted agents: predicting the future for patients with unpredictable outcomes to therapy. Ann Oncol. 2005;16:538–548.

Goldie JH, Coldman AJ. A mathematical model for relating drug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat Rep. 1979;63:1727–1733.

Goldie JH. Drug resistance. In: Perry MC, ed. The chemotherapy source book. Baltimore: Williams & Wilkins, 1992:54–66.

Kinzler KW, Vogelstein B. Cancer therapy meets p53. N Engl J Med. 1994;331:49–50.

Quintas-Cardama A, Cortes JE. Chronic myeloid leukemia: diagnosis and treatment Mayo clin Proc. 2006;81:973–988.

Schabel FM Jr. The use of tumor growth kinetics in planning “curative” chemotherapy of advanced solid tumors. Cancer Res. 1969;29:2384–2398.

Sikic BI. Modulation of multidrug resistance: at the threshold. J Clin Oncol. 1993;11: 1629–1635.

Slingerland JM, Tannock IF. Cell proliferation and cell death. In: Tannock IF, Hill RP, eds. The basic science of oncology. New York: McGraw-Hill, 1998:134–165.

Yarbro JW. The scientific basis of cancer chemotherapy. In: Perry MC, ed. The chemotherapy source book Baltimore: Williams & Wilkins, 1996:3–18.