Xem mẫu

  1. Section IX. Chemotherapy of Neoplastic Diseases Introduction Among the subspecialties of internal medicine, medical oncology may have had the greatest impact in changing the practice of medicine in the past four decades, as curative treatments have been identified for a number of previously fatal malignancies such as testicular cancer, lymphomas, and leukemia. New drugs have entered clinical use for disease presentations previously either untreatable or amenable to only local means of therapy, such as surgery and irradiation. At present, adjuvant chemotherapy routinely follows local treatment of breast cancer, colon cancer, and rectal cancer, and chemotherapy is employed as part of a multimodality approach to the initial treatment of many other tumors, including locally advanced stages of head and neck, lung, cervical, and esophageal cancer, soft tissue sarcomas, and pediatric solid tumors. The basic approaches to cancer treatment are constantly changing. Clinical protocols are now exploring genetic therapies, manipulations of the immune system, stimulation of normal hematopoietic elements, induction of differentiation in tumor tissues, and inhibition of angiogenesis. Research in each of these new areas has led to experimental or, in some cases, routine applications for both malignant and nonmalignant disease. The same drugs used for cytotoxic antitumor therapy have become important components of immunosuppressive regimens for rheumatoid arthritis (methotrexate and cyclophosphamide), organ transplantation (methotrexate and azathioprine), sickle cell anemia (hydroxyurea), antiinfective chemotherapy (trimetrexate and leucovorin), and psoriasis (methotrexate). Thus, a broad spectrum of medical, surgical, and pediatric specialists employ these drugs for both neoplastic and nonneoplastic disease. At the same time, few categories of medication in common use have a narrower therapeutic index and a greater potential for causing harmful side effects than do the antineoplastic drugs. A thorough understanding of their pharmacology, drug interactions, and clinical pharmacokinetics is essential for safe and effective use in human beings. Traditionally, cancer drugs were discovered through large-scale screening of synthetic chemicals and natural products against animal tumor systems, primarily murine leukemias. The agents discovered in the first two decades of cancer chemotherapy (1950 to 1970) largely interacted with DNA or its precursors, inhibiting the synthesis of new genetic material or causing irreparable damage to DNA itself. An overview of such agents is given in Figure IX–1. In recent years, the discovery of new agents has extended from the more conventional natural products such as paclitaxel and semisynthetic agents such as etoposide, both of which target the proliferative process, to entirely new fields of investigation that represent the harvest of new knowledge about cancer biology. The first successful applications of this knowledge include diverse drugs. One agent, interleukin-2, regulates the proliferation of tumor-killing T lymphocytes and so-called natural killer cells; this agent has proven able to induce remissions in a fraction of patients with malignant melanoma and renal cell carcinoma, diseases unresponsive to conventional drugs. Another agent, all-trans-retinoic acid, elicits differentiation and can be used to promote remission in acute promyelocytic leukemia, even after failure of standard chemotherapy. The related compound 13- cis-retinoic acid prevents occurrence of second primary tumors in patients with head and neck cancer. Initial success in characterizing unique tumor antigens and oncogenes has introduced new possible therapeutic opportunities based on an understanding of tumor biology. Thus the bcr-abl translocation in chronic myelocytic leukemia codes for a tyrosine kinase essential to cell proliferation and survival. Inhibition of the kinase by imatinib (STI-571), a new molecularly targeted drug, has produced a high response rate in chronic-phase patients resistant to standard therapy. In a similar, though immunological, approach tumor-associated antigens, such as the her- 2/neu receptor in breast cancer cells, have become the target for monoclonal antibody therapy that
  2. has shown activity in patients. These examples emphasize that the care of cancer patients is likely to undergo revolutionary changes as entirely new treatment approaches are identified, based on new knowledge of cancer biology (Kaelin, 1999). The diversity of agents useful in treatment of neoplastic disease is summarized in Table IX–1. The classification used in Chapter 52: Antineoplastic Agents, which follows, is a convenient framework for describing various types of agents. Figure IX–1. Summary of the Mechanisms and Sites of Action of Chemotherapeutic Agents Useful in Neoplastic Disease. PALA =N- phosphonoacetyl-L-aspartate; TMP = thymidine monophosphate. It is unlikely that new therapies will totally replace existing drugs, as these drugs have become increasingly effective and their toxicities have become more manageable and predictable in recent years. Improvements in their use are the result of a number of factors, including the following: 1. Drugs now are routinely used earlier in the course of the patient's management, often in conjunction with radiation or surgery, to treat malignancy when it is most curable and when the patient is best able to tolerate treatment. Thus, adjuvant therapy and neoadjuvant chemotherapy are used in conjunction with irradiation and surgery in the treatment of head and neck, esophageal, lung, and breast cancer patients. 2. The availability of granulocyte colony-stimulating factor (G-CSF; see Chapter 54: Hematopoietic Agents: Growth Factors, Minerals, and Vitamins) has shortened the period of leukopenia after high-dose chemotherapy, increasing the safety of bone marrow–ablative
  3. regimens and decreasing the incidence of life-threatening infection. A similar megakaryocyte growth and development factor has been cloned but has not yet achieved a useful place as an adjunct to chemotherapy. 3. A greater insight into the mechanisms of tumor cell resistance to chemotherapy has led to the more rational construction of drug regimens and the earlier use of intensive therapies. Drug-resistant cells may be selected from the larger tumor population by exposure to low-dose, single-agent chemotherapy. The resistance that arises may be specific for the selecting agent, such as the deletion of a necessary activating enzyme (deoxycytidine kinase for cytosine arabinoside), or more general, such as the overexpression of a general drug-efflux pump such as the P-glycoprotein, a product of the MDR gene. This membrane protein is one of several ATP-dependent transporters that confer resistance to a broad range of natural products used in cancer treatment. More recently, it has become appreciated that mutations underlying malignant transformation, such as the loss of the p53 suppressor oncogene, may lead to drug resistance. (A suppressor gene is essential for normal control of cell proliferation; its loss or mutation allows cells to undergo malignant transformation.) Mutation of p53, or its loss, or the overexpression of the bcl-2 gene that is translocated in nodular non-Hodgkin's lymphomas, inactivates a key pathway of programmed cell death (apoptosis) and leads to survival of highly mutated tumor cells that have the capacity to survive DNA damage. Drug discovery efforts are now directed toward restoring apoptosis in tumor cells, as this process, or its absence, seems to have profound influence on tumor cell sensitivity to drugs. Each of these topics concerning drug resistance is covered in greater detail in Chapter 52: Antineoplastic Agents. In designing specific regimens for clinical use, a number of factors must be taken into account. Drugs are generally more effective in combination and may be synergistic through biochemical interactions. These interactions are useful in designing new regimens. It is more effective to use drugs that do not share common mechanisms of resistance and that do not overlap in their major toxicities. Drugs should be used as close as possible to their maximum individual doses and should be given as frequently as possible to discourage tumor regrowth and to maximize dose intensity (the dose given per unit time, a key parameter in the success of chemotherapy). Since the tumor cell population in patients with visible disease exceeds 1 g, or 109 cells, and since each cycle of therapy kills less than 99% of the cells, it is necessary to repeat treatments in multiple cycles to kill all the tumor cells. The Cell Cycle An understanding of cell-cycle kinetics is essential for the proper use of the current generation of antineoplastic agents. Many of the most potent cytotoxic agents act by damaging DNA. Their toxicity is greater during the S, or DNA synthetic, phase of the cell cycle, while others, such as the vinca alkaloids and taxanes, block the formation of the mitotic spindle in M phase. These agents have activity only against cells that are in the process of division. Accordingly, human neoplasms that are currently most susceptible to chemotherapeutic measures are those with a high percentage of cells undergoing division. Similarly, normal tissues that proliferate rapidly (bone marrow, hair follicles, and intestinal epithelium) are subject to damage by most antineoplastic drugs, and such toxicity often limits the usefulness of drugs. On the other hand, slowly growing tumors with a small growth fraction (for example, carcinomas of the colon or lung) often are unresponsive to cytotoxic drugs. Although differences in the duration of the cell cycle occur between cells of various types, all cells display a similar pattern during the division process. This cell cycle may be characterized as follows (see Figure IX–2): (1) There is a presynthetic phase (G1); (2) the synthesis of DNA occurs (S); (3) an interval follows the termination of DNA synthesis, the postsynthetic phase (G 2); and (4) mitosis (M) ensues—the G2 cell, containing a double complement of DNA, divides into two
  4. daughter G1 cells. Each of these cells may immediately reenter the cell cycle or pass into a nonproliferative stage, referred to as G0. The G0 cells of certain specialized tissues may differentiate into functional cells that no longer are capable of division. On the other hand, many cells, especially those in slow-growing tumors, may remain in the G0 state for prolonged periods, only to reenter the division cycle at a later time. Damaged cells that reach the G1/S boundary undergo apoptosis, or programmed cell death, if the p53 gene is intact and if it exerts its normal checkpoint function. If the p53 gene is mutated and the checkpoint function fails, damaged cells will not be diverted to the apoptotic pathway. These cells will proceed through S phase and some will emerge as a drug- resistant population. Thus, an understanding of cell-cycle kinetics and the controls of normal and malignant cell growth is crucial to the design of current therapy regimens and the search for new drugs. Figure IX–2. The Cell Cycle and the Relationship of Antitumor Drug Action to the Cycle. G1 is the period between mitosis and the beginning of DNA synthesis. Resting cells (cells that are not preparing for cell division) are said to be in a subphase of G1, G0. S is the period of DNA synthesis; G2 the premitotic interval; and M the period of mitosis. Examples of cell-cycle–dependent anticancer drugs are listed in blue below the phase in which they act. Drugs that are cytotoxic for cells at any point in the cycle are called cycle-phase-nonspecific drugs. (Modified from Pratt et al., 1994 with permission.) Achieving Therapeutic Balance and Efficacy
  5. While not the subject of this chapter, it must be emphasized that the treatment of most cancer patients requires a skillful interdigitation of multiple modalities of treatment, including surgery, irradiation, and drugs. Each of these forms of treatment carries its own risks and benefits. It is obvious that not all drugs and not all regimens are safe or appropriate for all patients. Numerous factors must be considered, such as renal and hepatic function, bone marrow reserve, and the status of general performance and accessory medical problems. Beyond those considerations, however, are less quantifiable factors such as the likely natural history of the tumor being treated, the patient's willingness to undergo harsh treatments, the patient's physical and emotional tolerance for side effects, and the likely long-term gains and risks involved. The emphasis in Chapter 52: Antineoplastic Agents is placed upon the drugs, but it is essential to point out the importance of the role played by the patient. It is generally agreed that patients in good nutritional state and without severe metabolic disturbances, infections, or other complications have better tolerance for chemotherapy and have a better chance for significant improvement than do severely debilitated individuals. Ideally, the patient should have adequate renal, hepatic, and bone marrow function, the latter uncompromised by tumor invasion, previous chemotherapy, or irradiation (particularly of the spine or pelvis). Nevertheless, even patients with advanced disease have improved dramatically with chemotherapy. Although methods that would enable accurate prediction of the responsiveness of a particular tumor to a given agent are still investigational, in the future, molecular studies of tumor specimens may allow prediction of response and the rational selection of patients for specific drugs. Despite efforts to anticipate the development of complications, anticancer agents have variable pharmacokinetics and toxicity in individual patients. The causes of this variability are not always clear and often may be related to interindividual differences in drug metabolism, drug interactions, or bone marrow reserves. In dealing with toxicity, the physician must provide vigorous supportive care, including, where indicated, platelet transfusions, antibiotics, and hematopoietic growth factors (see Chapter 54: Hematopoietic Agents: Growth Factors, Minerals, and Vitamins). Other delayed toxicities affecting the heart, lungs, or kidneys may not be reversible and may lead to permanent organ damage or death. Fortunately, such toxicities will be uncommon if the physician adheres to standard protocols and respects the guidelines for drug usage detailed in the following discussion. Chapter 52. Antineoplastic Agents Alkylating Agents History Although synthesized in 1854, the vesicant properties of sulfur mustard were not described until 1887. During World War I, medical attention was first focused on the vesicant action of sulfur mustard on the skin, eyes, and respiratory tract. It was appreciated later, however, that serious systemic toxicity also follows exposure. In 1919, Krumbhaar and Krumbhaar made the pertinent observation that the poisoning caused by sulfur mustard is characterized by leukopenia and, in cases that came to autopsy, by aplasia of the bone marrow, dissolution of lymphoid tissue, and ulceration of the gastrointestinal tract. In the interval between World Wars I and II, extensive studies of the biological and chemical actions of the nitrogen mustards were conducted. The marked cytotoxic action on lymphoid tissue prompted Gilman, Goodman, and T.F. Dougherty to study the effect of nitrogen mustards on transplanted lymphosarcoma in mice, and in 1942 clinical studies were initiated. This launched the era of modern cancer chemotherapy (Gilman, 1963).
  6. In their early phases, all these investigations were conducted under secrecy restrictions imposed by the use of classified chemical-warfare agents. At the termination of World War II, however, the nitrogen mustards were declassified; a general review was presented by Gilman and Philips (1946). A more recent review is provided by Ludlum and Tong (1985). Thousands of variants of the basic chemical structure of the nitrogen mustards have been prepared, but only a few of these agents have proven more useful than the original compound in specific clinical circumstances (see below). At present five major types of alkylating agents are used in the chemotherapy of neoplastic diseases: (1) the nitrogen mustards, (2) the ethylenimines, (3) the alkyl sulfonates, (4) the nitrosoureas, and (5) the triazenes. Chemistry The chemotherapeutic alkylating agents have in common the property of becoming strong electrophiles through the formation of carbonium ion intermediates or of transition complexes with the target molecules. These reactions result in the formation of covalent linkages by alkylation of various nucleophilic moieties such as phosphate, amino, sulfhydryl, hydroxyl, carboxyl, and imidazole groups. The chemotherapeutic and cytotoxic effects are directly related to the alkylation of DNA. The 7 nitrogen atom of guanine is particularly susceptible to the formation of a covalent bond with bifunctional alkylating agents and may well represent the key target that determines their biological effects. It must be appreciated, however, that other atoms in the purine and pyrimidine bases of DNA—particularly, the 1 and 3 nitrogens of adenine, the 3 nitrogen of cytosine, and the 6 oxygen of guanine—also may be alkylated, as will be the phosphate atoms of the DNA chains and amino and sulfhydryl groups of proteins. To illustrate the actions of alkylating agents, possible consequences of the reaction of mechlorethamine (nitrogen mustard) with guanine residues in DNA chains are shown in Figure 52– 1. First, one 2-chloroethyl side chain undergoes a first-order (SN1) intramolecular cyclization, with release of Cl– and formation of a highly reactive ethyleniminium intermediate (Figure 52–1A). By this reaction, the tertiary amine is converted to an unstable quaternary ammonium compound, which can react avidly, through formation of a carbonium ion or transition complex intermediate, with a variety of sites that possess high electron density. This reaction proceeds as a second-order (S N2) nucleophilic substitution. Alkylation of the 7 nitrogen of guanine residues in DNA (Figure 52–1B), a highly favored reaction, may exert several effects of considerable biological importance. Normally, guanine residues in DNA exist predominantly as the keto tautomer and readily make Watson–Crick base pairs by hydrogen bonding with cytosine residues. However, when the 7 nitrogen of guanine is alkylated (to become a quaternary ammonium nitrogen), the guanine residue is more acidic and the enol tautomer is favored. The modified guanine can mispair with thymine residues during DNA synthesis, leading to the substitution of an adenine–thymine base pair for a guanine–cytosine base pair. Second, alkylation of the 7 nitrogen labilizes the imidazole ring, making possible the opening of the imidazole ring or depurination by excision of guanine residues. Either of these seriously damages the DNA molecule and must be repaired. Third, with bifunctional alkylating agents, such as nitrogen mustard, the second 2-chloroethyl side chain can undergo a similar cyclization reaction and alkylate a second guanine residue or another nucleophilic moiety, resulting in the cross-linking of two nucleic acid chains or the linking of a nucleic acid to a protein, alterations that would cause a major disruption in nucleic acid function. Any of these effects could adequately explain both the mutagenic and the cytotoxic effects of alkylating agents. However, cytotoxicity of bifunctional alkylators correlates very closely with interstrand cross-linkage of DNA (Garcia et al., 1988).
  7. Figure 52–1. Mechanism of Action of Alkylating Agents. The ultimate cause of cell death related to DNA damage is not known. Specific cellular responses include cell-cycle arrest, DNA repair, and apoptosis, a specific form of nuclear fragmentation termed programmed cell death (Fisher, 1994). The p53 gene product senses DNA damage and initiates apoptosis in response to DNA alkylation. Mutations of p53 lead to alkylating-agent resistance (Kastan, 1999). All nitrogen mustards are chemically unstable but vary greatly in their degree of instability. Therefore, the specific chemical properties of each member of this class of drugs must be considered individually in therapeutic applications. For example, mechlorethamine is very unstable, and it reacts almost completely in the body within a few minutes of its administration. By contrast, agents such as chlorambucil are sufficiently stable to permit oral administration. Cyclophosphamide requires biochemical activation by the cytochrome P450 system of the liver before its cytotoxicity becomes evident. The ethylenimine derivatives such as chlorambucil and melphalan react by an SN2 reaction; since the opening of the ethylenimine intermediate is acid-catalyzed, they are more reactive at acidic pH. Structure–Activity Relationship The alkylating agents used in chemotherapy encompass a diverse group of chemicals that have in common the capacity to contribute, under physiological conditions, alkyl groups to biologically
  8. vital macromolecules such as DNA. In most instances, physical and chemical parameters, such as lipophilicity, capacity to cross biological membranes, acid dissociation constants, stability in aqueous solution, and sites of macromolecular attack, determine drug activity in vivo. With several of the most valuable agents (e.g., cyclophosphamide and the nitrosoureas), the active alkylating moieties are generated in vivo after complex metabolic reactions. The nitrogen mustards may be regarded as nitrogen analogs of sulfur mustard. The biological activity of both types of compounds is based upon the presence of the bis-(2-chloroethyl) grouping. While mechlorethamine has been widely used in the past, various structural modifications have resulted in compounds with greater selectivity and stability and therefore less toxicity. Bis-(2- chloroethyl) groups have been linked to amino acids (phenylalanine), substituted phenyl groups (aminophenyl butyric acid, as in chlorambucil), pyrimidine bases (uracil), and other chemical entities in an effort to make a more stable and orally available form. Although none of these modifications has produced an agent highly selective for malignant cells, some have unique pharmacological properties and are more useful clinically than is mechlorethamine. Their structures are shown in Figure 52–2. Figure 52–2. Nitrogen Mustards Employed in Therapy. The addition of substituted phenyl groups has produced a series of relatively stable derivatives that retain the ability to form reactive charged intermediates; the electron-withdrawing capacity of the aromatic ring greatly reduces the rate of cyclization and carbonium ion formation, and these compounds therefore can reach distant sites in the body before reacting with components of blood and other tissues. Chlorambucil and melphalan are the most successful examples of such aromatic mustards. These compounds can be administered orally if desired. A classical example of the role of host metabolism in the activation of an alkylating agent is seen with cyclophosphamide—now the most widely used agent of this class. The design of this molecule was based on two considerations. First, if a cyclic phosphamide group replaced the N-methyl of mechlorethamine, the compound might be relatively inert, presumably because the bis-(2- chloroethyl) group of the molecule could not ionize until the cyclic phosphamide was cleaved at the phosphorus–nitrogen linkage. Second, it was hoped that neoplastic tissues might possess high phosphatase or phosphamidase activity capable of accomplishing this cleavage, thus resulting in the selective production of an activated nitrogen mustard in the malignant cells. In accord with these predictions, the parent cyclophosphamide displays only weak cytotoxic, mutagenic, or alkylating activity in vitro and is relatively stable in aqueous solution. However, when administered to experimental animals or patients bearing susceptible tumors, it causes marked chemotherapeutic effects, as well as mutagenicity and carcinogenicity. The postulated role for phosphatases or
  9. phosphamidases in the mechanism of action of cyclophosphamide has proven incorrect. Rather, the drug undergoes metabolic activation (hydroxylation) by the cytochrome P450 mixed-function oxidase system of the liver (Figure 52–3), with subsequent transport of the activated intermediate to sites of action, as discussed below. The selectivity of cyclophosphamide against certain malignant tissues may result in part from the capacity of normal tissues, such as liver, to protect themselves against cytotoxicity by further degrading the activated intermediates via aldehyde dehydrogenase and other pathways. Figure 52–3. Metabolism of Cyclophosphamide. Ifosfamide is an oxazaphosphorine, similar to cyclophosphamide. Cyclophosphamide has two chloroethyl groups on the exocyclic nitrogen atom, whereas one of the two chloroethyl groups of ifosfamide is on the cyclic phosphamide nitrogen of the oxazaphosphorine ring. Like cyclophosphamide, ifosfamide is activated in the liver by hydroxylation. However, the activation of ifosfamide proceeds more slowly, with greater production of dechlorinated metabolites and chloroacetaldehyde. These differences in metabolism likely account for the higher doses of ifosfamide required for equitoxic effects and the possible differences in antitumor spectrum of the two agents. Although initially considered an antimetabolite, the triazene derivative 5-(3,3-dimethyl-1-triazeno)- imidazole-4-carboxamide, usually referred to as dacarbazine or DTIC, functions through alkylation. Its structural formula is shown below:
  10. Dacarbazine requires initial activation by the cytochrome P450 system of the liver through an N- demethylation reaction. In the target cell, spontaneous cleavage of the metabolite yields an alkylating moiety, diazomethane. A related triazene, temozolomide undergoes spontaneous activation, and has significant activity against gliomas and melanoma in human beings (Agarwala and Kirkwood, 2000). It has the same profile of toxicity as DTIC, and is active against malignant gliomas and melanoma. Its structure is shown below: The nitrosoureas, which include compounds such as 1,3-bis-(2-chloroethyl)-1-nitrosourea (carmustine, BCNU), 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (lomustine, CCNU), and its methyl derivative (semustine, methyl-CCNU), as well as the antibiotic streptozocin (streptozotocin), exert their cytotoxicity through the spontaneous breakdown to alkylating and carbamoylating moieties. The structural formula of carmustine is as follows: The antineoplastic nitrosoureas have in common the capacity to undergo spontaneous, nonenzymatic degradation with the formation of the 2-chloroethyl carbonium ion (from CNU compounds). This strong electrophile can alkylate a variety of substances; guanine, cytidine, and adenine adducts have been identified (Ludlum, 1990). Displacement of the halogen atom can then lead to interstrand or intrastrand cross-linking of the DNA. The formation of the cross-links after the initial alkylation reaction is relatively slow and can be interrupted by the DNA repair enzyme guanine O6-alkyl transferase (Dolan et al., 1990). The same enzyme, when overexpressed in gliomas, produces resistance to nitrosoureas and various methylating agents, including DTIC, temozolomide, and procarbazine. As with the nitrogen mustards, it is generally agreed that interstrand cross-linking is associated with the cytotoxicity of nitrosoureas (Hemminki and Ludlum, 1984). In addition to the generation of carbonium ions, the spontaneous degradation of BCNU,
  11. CCNU, and methyl-CCNU liberates organic isocyanates that attach carbamoyl groups to lysine residues of proteins, a reaction that apparently can inactivate certain DNA repair enzymes. The reactions of the nitrosoureas with macromolecules are shown in Figure 52–4. Figure 52–4. Degradation of Carmustine (BCNU) with Generation of Alkylating and Carbamoylating Intermediates. Since the formation of the ethyleniminium ion constitutes the initial reaction of the nitrogen mustards, it is not surprising that stable ethylenimine derivatives have antitumor activity. Several compounds of this type, including triethylenemelamine (TEM) and triethylene thiophosphoramide (thiotepa), have been used clinically. In standard doses, thiotepa produces little toxicity other than myelosuppression and is thus increasingly used for high-dose chemotherapy regimens. Altretamine (hexamethylmelamine; HMM) is mentioned here because of its chemical similarity to TEM. The methylmelamines are N-demethylated by hepatic microsomes, with the release of formaldehyde, and there is a relationship between the degree of the demethylation and their activity against murine tumors. Altretamine requires microsomal activation to display cytotoxicity (Friedman, 2001). Several interesting compounds have emerged from a large group of esters of alkanesulfonic acids. One of these, busulfan, is of value in the treatment of chronic granulocytic leukemia and in high- dose chemotherapy; its structural formula is as follows:
  12. Busulfan is a member of a series of symmetrical bis-substituted methanesulfonic acid esters in which the length of a bridge of methylene varies from 2 to 10. The compounds of intermediate length (n= 4 or 5) possess the highest activities and therapeutic indices. Cross-linked guanine residues have been identified in DNA incubated in vitro with busulfan (Tong and Ludlum, 1980). Pharmacological Actions The pharmacological actions of the various groups of alkylating agents are considered together in the following discussion. Although there are many similarities, some notable differences also are evident. Cytotoxic Actions The most important pharmacological actions of the alkylating agents are those that disturb DNA synthesis and cell division. The capacity of these drugs to interfere with DNA integrity and function in rapidly proliferating tissues provides the basis for their therapeutic applications and for many of their toxic properties. Whereas certain alkylating agents may have damaging effects on tissues with normally low mitotic indices—for example, liver, kidney, and mature lymphocytes—they are most cytotoxic to rapidly proliferating tissues in which a large proportion of the cells are in division. These compounds may readily alkylate nondividing cells, but cytotoxicity is markedly enhanced if DNA is damaged in cells programmed to divide. Thus, DNA alkylation itself may not be a lethal event if DNA repair enzymes can correct the lesions in DNA prior to the next cellular division. In contrast to many other antineoplastic agents, the effects of the alkylating drugs, although dependent on proliferation, are not cell-cycle–specific, and the drugs may act on cells at any stage of the cycle. However, the toxicity is usually expressed when the cell enters the S phase and progression through the cycle is blocked. While not strictly cell-cycle–specific, quantitative differences may be detected when nitrogen mustards are applied to synchronized cells at different phases of the cycle. Cells appear more sensitive in late G1 or S than in G2, mitosis, or early G1. Polynucleotides are more susceptible to alkylation in the unpaired state than in the helical form; during replication of DNA, portions of the molecule are unpaired. The actual mechanism(s) of cell death related to DNA alkylation are not well understood. There is evidence that, in normal cells of the bone marrow and intestinal epithelium, DNA damage activates a checkpoint dependent on the presence of a normal p53 gene. Cells thus blocked in the G1/S interface either repair DNA alkylation or undergo apoptosis. Malignant cells with mutant or absent p53 fail to suspend cell-cycle progression and do not undergo apoptosis (Fisher, 1994). The great preponderance of evidence indicates that the primary target of pharmacological doses of alkylating agents is DNA, as illustrated in Figure 52–1. A crucial distinction that must be emphasized is between the bifunctional agents, in which cytotoxic effects predominate, and the monofunctional methylating agents (procarbazine, temozolomide), which, although cytotoxic, have greater capacity for mutagenesis and carcinogenesis. This suggests that the cross-linking of DNA strands represents a much greater threat to cellular survival than do other effects, such as single-
  13. base alkylation and the resulting depurination and chain scission. On the other hand, the latter reactions may cause permanent modifications in DNA structure and sequence that are compatible with continued life of the cell and are transmissible to subsequent generations; such modifications may result in mutagenesis or carcinogenesis. The remarkable DNA repair systems found in most cells likely play an important but as yet poorly defined role in the relative resistance of nonproliferating tissues, the selectivity of action against particular cell types, and acquired resistance to alkylating agents. Although alkylation of a single strand of DNA often may be repaired with relative ease, interstrand cross-linkages, such as those produced by the bifunctional alkylating agents, require more complex mechanisms for repair. Many of the cross-links formed in DNA by these agents at low doses also may be corrected; higher doses cause extensive cross-linkage, and DNA breakdown occurs. Specific repair enzymes for removing alkyl groups from the O-6 of guanine (guanine O6-alkyl transferase) and the N-3 of adenine and N-7 of guanine (3-methyladenine-DNA glycosylase) have been identified (Matijasevic et al., 1993). The presence of sufficient levels of guanine O6-alkyl transferase protects cells from cytotoxic effects of nitrosoureas and methylating agents (Pegg, 1990) and confers drug resistance. Detailed information is lacking on mechanisms of cellular uptake of alkylating agents. Mechlorethamine appears to enter murine tumor cells by means of an active transport system, the natural substrate of which is choline. Melphalan, an analog of phenylalanine, is taken up by at least two active transport systems that normally react with leucine and other neutral amino acids. The highly lipophilic drugs, including nitrosoureas, carmustine, and lomustine, diffuse into cells passively. Mechanisms of Resistance to Alkylating Agents Acquired resistance to alkylating agents is a common event, and the acquisition of resistance to one alkylating agent often but not always imparts cross-resistance to others; thus, there are at least theoretical reasons to combine alkylating agents in high-dose therapy. While definitive information on the biochemical mechanisms of clinical resistance is lacking, specific biochemical changes have been implicated in the development of such resistance by tumor cells. Among these changes are (1) decreased permeation of actively transported drugs (mechlorethamine and melphalan); (2) increased production of nucleophilic substances, principally thiols such as glutathione, that can conjugate with and detoxify electrophilic intermediates; (3) increased activity of the DNA repair enzymes, such as the guanine O6-alkyl transferase, that repair nitrosourea-produced alkylation; and (4) increased rates of metabolism of the activated forms of cyclophosphamide to its inactive keto and carboxy metabolites by aldehyde dehydrogenase (see Figure 52–3; Tew et al., 2001). To reverse cellular changes that lead to resistance, strategies have been devised and appear to be effective in selected experimental tumors. These include the use of compounds that deplete glutathione, such as L-buthionine-sulfoximine; sulfhydryl compounds, such as WR-2721, that selectively detoxify alkylating species in normal cells and thereby prevent toxicity; compounds such as O6-benzylguanine that inactivate the guanine O6-alkyl transferase DNA repair enzyme; and compounds such as ethacrynic acid that inhibit the enzymes (glutathione transferases) that conjugate thiols with alkylating agents. While each of these modalities has experimental evidence to support its use, the clinical efficacy has not yet been proven for these strategies. Of these, O6- benzylguanine has advanced to phase II trials used in conjunction with carmustine (BCNU) or procarbazine against malignant gliomas (Schilsky et al., 2000).
  14. Toxicities of Alkylating Agents The alkylating agents differ in their patterns of antitumor activity and in the sites and severity of their side effects. Most cause dose-limiting toxicity to bone marrow elements and, to a lesser extent, intestinal mucosa. Most alkylating agents, including nitrogen mustard, melphalan, chlorambucil, cyclophosphamide, and ifosfamide, produce an acute myelosuppression, with a nadir of the peripheral blood granulocyte count at 6 to 10 days and recovery in 14 to 21 days. Cyclophosphamide has lesser effects on peripheral blood platelet counts than do the other agents. Busulfan suppresses all blood elements, particularly stem cells, and may produce a prolonged and cumulative myelosuppression lasting months. For this reason, it is used as a preparative regimen in allogenic bone marrow transplantation. BCNU and other chloroethylnitrosoureas cause delayed and prolonged suppression of both platelets and granulocytes, reaching a nadir 4 to 6 weeks after drug administration and reversing slowly thereafter. Both cellular and humoral immunity are suppressed by alkylating agents, which have been used to treat various autoimmune diseases. Immunosuppression is reversible at doses used in most anticancer protocols. In addition to effects on the hematopoietic system, alkylating agents are highly toxic to dividing mucosal cells, leading to oral mucosal ulceration and intestinal denudation. The mucosal effects are particularly significant in high-dose chemotherapy protocols associated with bone marrow reconstitution, as they predispose to bacterial sepsis arising from the gastrointestinal tract. In these protocols, melphalan and thiotepa have the advantage of causing less mucosal damage than the other agents. In high-dose protocols, a number of toxicities not seen at conventional doses become dose-limiting. They are listed in Table 52–1. While mucosal and bone marrow toxicities occur predictably with conventional doses of these drugs, other organ toxicities, although less common, can be irreversible and at times lethal. All alkylating agents have caused pulmonary fibrosis, and in high-dose regimens, endothelial damage that may precipitate venoocclusive disease of the liver; the nitrosoureas, after multiple cycles of therapy, may lead to renal failure; ifosfamide in high-dose regimens frequently causes a central neurotoxicity, with seizures, coma, and at times death; and all such agents are leukemogenic, particularly procarbazine (a methylating agent) and the nitrosoureas. Cyclophosphamide and ifosfamide release a nephrotoxic and urotoxic metabolite, acrolein, which causes a severe hemorrhagic cystitis, a side effect that in high-dose regimens can be prevented by coadministration of the sulfhydryl-releasing agent mesna (2-mercaptoethanesulfonate). Mesna, when administered with the offending agent at 60% of the drug dosage, conjugates toxic metabolites in urine. The more unstable alkylating agents (particularly nitrogen mustard and the nitrosoureas) have strong vesicant properties, damage veins with repeated use, and, if extravasated, produce ulceration. Topical application of nitrogen mustard is an effective treatment for cutaneous neoplasms such as mycosis fungoides. Most alkylating agents cause alopecia. Central nervous system (CNS) toxicity is manifest in the form of nausea and vomiting, particularly after intravenous administration of nitrogen mustard or BCNU. Ifosfamide is the most neurotoxic of this class of agents, producing altered mental status, coma, generalized seizures, and paralysis. These side effects have been linked to the release of chloroacetaldehyde from the phosphate-linked chloroethyl side chain of ifosfamide. High-dose busulfan may cause seizures; in addition, it accelerates the clearance of phenytoin, an antiseizure medication (see Chapter 21: Drugs Effective
  15. in the Therapy of the Epilepsies). As a class of drugs, the alkylating agents are highly leukemogenic. Acute nonlymphocytic leukemia, often associated with partial or total deletions of chromosome 5 or 7, peaks in incidence about four years after therapy and may affect up to 5% of patients treated on regimens containing alkylating drugs (Levine and Bloomfield, 1992). Melphalan, the nitrosoureas, and the methylating agent procarbazine have the greatest propensity to cause leukemia, while cyclophosphamide is less potent in this regard. Finally, all alkylating agents have toxic effects on the male and female reproductive systems, causing an often permanent amenorrhea, particularly in perimenopausal women, and an irreversible azoospermia in men. Nitrogen Mustards The chemistry and the pharmacological actions of the alkylating agents as a group, and of the nitrogen mustards, have been presented above. Only the unique pharmacological characteristics of the individual agents are considered below. Mechlorethamine Mechlorethamine, the first nitrogen mustard to be introduced into clinical medicine, is the most reactive of the drugs in this class. Absorption and Fate Severe local reactions of exposed tissues necessitate intravenous injection of mechlorethamine for most clinical uses. In either water or body fluids, at rates affected markedly by pH, mechlorethamine rapidly undergoes chemical transformation and combines with either water or nucleophilic molecules of cells, so that the parent drug has an extremely short mean residence time in the body. Therapeutic Uses Mechlorethamine HCl (MUSTARGEN) is used primarily in the combination chemotherapy regimen MOPP [mechlorethamine, ONCOVIN (vincristine), procarbazine, and prednisone] in patients with Hodgkin's disease (DeVita et al., 1972). It is given by intravenous bolus administration in doses of 6 mg/m2 on days 1 and 8 of the 28-day cycles of each course of treatment. It has been largely replaced in other regimens by cyclophosphamide, melphalan, and other, more stable, alkylating agents. Clinical Toxicity The major acute toxic manifestations of mechlorethamine are nausea, vomiting, and lacrimation as well as myelosuppression. Leukopenia and thrombocytopenia limit the amount of drug that can be given in a single course. Like other alkylating agents, nitrogen mustard blocks reproductive function and may produce menstrual irregularities or premature menopause in women and oligospermia in men. Since fetal abnormalities can be induced, this drug as well as other alkylating agents should not be used in the
  16. first trimester of pregnancy and should be used with caution in later stages of pregnancy. Breast- feeding should be terminated before therapy with mechlorethamine is initiated. Local reactions to extravasation of mechlorethamine into the subcutaneous tissue result in a severe, brawny, tender induration that may persist for a long time. If the local reaction is unusually severe, a slough may result. If it is obvious that extravasation has occurred, the involved area should be promptly infiltrated with a sterile isotonic solution of sodium thiosulfate (1/6 M); an ice compress then should be applied intermittently for 6 to 12 hours. Thiosulfate provides an ion that reacts avidly with the nitrogen mustard and thereby protects tissue constituents. Cyclophosphamide Pharmacological and Cytotoxic Actions Although the general cytotoxic action of this drug is similar to that of other alkylating agents, there are notable differences. Thrombocytopenia is less severe, while alopecia is marked. There are no severe acute or delayed central nervous system (CNS) manifestations either in conventional doses or in high-dose regimens. Nausea and vomiting, however, may occur. The drug is not a vesicant, and there is no local irritation. Absorption, Fate, and Excretion Cyclophosphamide is well absorbed orally. As mentioned above, the drug is activated by the hepatic cytochrome P450 system (see Figure 52–3). Cyclophosphamide is first converted to 4- hydroxycyclophosphamide, which is in a steady state with the acyclic tautomer aldophosphamide. In vitro studies with human liver microsomes and cloned P450 isoenzymes have shown that cyclophosphamide is activated by the CYP2B group of P450 isoenzymes, while a closely related oxazaphosphorine, ifosfamide, is hydroxylated by the CYP3A system (Chang et al., 1993). This difference may account for the somewhat different patterns of antitumor activity, the slower activation of ifosfamide in vivo, and the interpatient variability in toxicity of these two closely related molecules. 4-Hydroxycyclophosphamide may be oxidized further by aldehyde oxidase either in liver or in tumor tissue and perhaps by other enzymes, yielding the metabolites carboxyphosphamide and 4-ketocyclophosphamide, neither of which possesses significant biological activity. It appears that hepatic damage is minimized by these secondary reactions, whereas significant amounts of the active metabolites, such as 4-hydroxycyclophosphamide and its tautomer, aldophosphamide, are transported to the target sites by the circulatory system. In tumor cells, the aldophosphamide cleaves spontaneously, generating stoichiometric amounts of phosphoramide mustard and acrolein. The former is believed to be responsible for antitumor effects. The latter compound may be responsible for the hemorrhagic cystitis seen during therapy with cyclophosphamide. Cystitis can be reduced in intensity or prevented by the parenteral administration of mesna (MESNEX), a sulfhydryl compound that reacts readily with acrolein in the acid environment of the urinary tract (Tew et al., 2001). Pretreatment with P450 inducers such as phenobarbital enhances the rate of drug activation but does not alter toxicity or therapeutic activity in human beings. Urinary and fecal recovery of unchanged cyclophosphamide is minimal after intravenous administration. Maximal concentrations in plasma are achieved 1 hour after oral administration, and the half-life in plasma is about 7 hours.
  17. Therapeutic Uses Cyclophosphamide (CYTOXAN, NEOSAR) is administered orally or intravenously. Recommended doses vary widely, and published protocols for the dosage of cyclophosphamide and other chemotherapeutic agents and for the method and sequence of administration should be consulted. As a single agent, a daily dose of 100 mg/m2 orally for 14 days has been recommended for patients with more susceptible neoplasms, such as lymphomas and chronic leukemias. A higher dosage of 500 mg/m2 intravenously every 3 to 4 weeks in combination with other drugs often is employed in the treatment of breast cancer and lymphomas. The leukocyte count generally serves as a guide to dosage adjustments in prolonged therapy. An absolute neutrophil count between 500 and 1000 cells per cubic millimeter is recommended as the desired target. In regimens associated with bone marrow or peripheral stem cell rescue, cyclophosphamide may be given in doses of 5 to 7 g/m 2 over a 3-day period. Gastrointestinal ulceration, cystitis (counteracted by mesna and diuresis), and, less commonly, pulmonary, renal, hepatic, and cardiac toxicities may occur after high-dose therapy. The clinical spectrum of activity for cyclophosphamide is very broad. It is an essential component of many effective drug combinations for non-Hodgkin's lymphomas. Complete remissions and presumed cures have been reported when cyclophosphamide was given as a single agent for Burkitt's lymphoma. It is frequently used in combination with methotrexate (or doxorubicin) and fluorouracil as adjuvant therapy after surgery for carcinoma of the breast. Notable advantages of this drug are the availability of the oral route of administration and the possibility of giving fractionated doses over prolonged periods. For these reasons it possesses a versatility of action that allows an intermediate range of use, between that of the highly reactive intravenous mechlorethamine and that of oral chlorambucil. Beneficial results have been obtained in multiple myeloma; chronic lymphocytic leukemia; carcinomas of the lung, breast, cervix, and ovary; and neuroblastoma, retinoblastoma, and other neoplasms of childhood. Because of its potent immunosuppressive properties, cyclophosphamide has received considerable attention for the control of organ rejection after transplantation and in nonneoplastic disorders associated with altered immune reactivity, including Wegener's granulomatosis, rheumatoid arthritis, and the nephrotic syndrome in children. Caution is advised when the drug is considered for use in these conditions, not only because of its acute toxic effects but also because of its potential for inducing sterility, teratogenic effects, and leukemia. Clinical Toxicity Nausea and vomiting, myelosuppression with platelet sparing, and alopecia are common to virtually all regimens using cyclophosphamide. Mucosal ulcerations and, less frequently, interstitial pulmonary fibrosis also may result from cyclophosphamide treatment. Extravasation of the drug into subcutaneous tissues does not produce local reactions, and thrombophlebitis does not complicate intravenous administration. The occurrence of sterile hemorrhagic cystitis has been reported in 5% to 10% of patients. As noted above, this has been attributed to chemical irritation produced by acrolein. Its incidence is significantly reduced by coadministration of mesna (Brock and Pohl, 1986). For routine clinical use, ample fluid intake is recommended. Administration of the drug should be interrupted at the first indication of dysuria or hematuria. The syndrome of inappropriate secretion of antidiuretic hormone (ADH) has been observed in patients receiving cyclophosphamide, usually at doses higher than 50 mg/kg (DeFronzo et al., 1973). It is important to be aware of the possibility of water intoxication, since these patients usually are vigorously
  18. hydrated. Ifosfamide Ifosfamide, an analog of cyclophosphamide, also is activated by ring hydroxylation in the liver. Severe urinary tract toxicity limited the use of ifosfamide when it was first introduced in the early 1970s. However, adequate hydration and coadministration of mesna now permit effective use of ifosfamide. Therapeutic Uses Ifosfamide currently is approved for use in combination with other drugs for germ cell testicular cancer and is widely used to treat pediatric and adult sarcomas. Clinical trials also have shown ifosfamide to be active against carcinomas of the cervix and lung and against lymphomas. It is a common component of high-dose chemotherapy regimens with bone marrow or stem cell rescue; in these regimens, in total doses of 12 to 14 g/m2, it may cause severe neurological toxicity, including coma and death. This toxicity is thought to result from a metabolite, chloracetaldehyde (Colvin, 1982). In addition to hemorrhagic cystitis, ifosfamide causes nausea, vomiting, anorexia, leukopenia, nephrotoxicity, and CNS disturbances (especially somnolence or confusion) (see Brade et al., 1987). Ifosfamide (IFEX) is infused intravenously over at least 30 minutes at a dose of 1.2 g/m 2 per day for 5 days. Intravenous mesna is given as bolus injections in a dosage equal to 20% of the ifosfamide dosage concomitantly and again 4 and 8 hours later, for a total mesna dose of 60% of the ifosfamide dose. Alternatively, mesna may be given in a single dose equal to the ifosfamide dose concomitantly. Patients also should receive at least 2 liters of oral or intravenous fluid daily. Treatment cycles are usually repeated every 3 to 4 weeks. Pharmacokinetics Ifosfamide has a half-life in plasma of approximately 15 hours after doses of 3.8 to 5.0 g/m 2 and a somewhat shorter half-life at lower doses. Toxicity Ifosfamide has virtually the same toxicity profile as does cyclophosphamide, with perhaps greater platelet suppression, neurotoxicity, and, in the absence of mesna, urothelial damage. Melphalan Pharmacological and Cytotoxic Actions The general pharmacological and cytotoxic actions of melphalan, the phenylalanine derivative of nitrogen mustard, are similar to those of other nitrogen mustards. The drug is not a vesicant. Absorption, Fate, and Excretion When given orally, melphalan is absorbed in an incomplete and variable manner, and 20% to 50% of the drug is recovered in the stool. The drug has a half-life in plasma of approximately 45 to 90 minutes, and 10% to 15% of an administered dose is excreted unchanged in the urine (Alberts et al.,
  19. 1979b). Therapeutic Uses The usual oral melphalan (ALKERAN) dose for multiple myeloma is 6 mg daily for a period of 2 to 3 weeks, during which time the blood count should be carefully observed. A rest period of up to 4 weeks should then intervene. When the leukocyte and platelet counts are rising, maintenance therapy, ordinarily 2 to 4 mg daily, is begun. It usually is necessary to maintain a significant degree of bone marrow depression (total leukocyte count in the range of 2500 to 3500 cells per cubic millimeter) in order to achieve optimal results. The usual intravenous dose is 16 mg/m 2 infused over 15 to 20 minutes. Doses are repeated at 2-week intervals for four doses and then at 4-week intervals based on response and tolerance. Dosage adjustments should be considered based on blood cell counts and in patients with renal impairment. Although the general spectrum of action of melphalan seems to resemble that of other nitrogen mustards, the advantages of administration by the oral route have made the drug useful in the treatment of multiple myeloma. Clinical Toxicity The clinical toxicity of melphalan is mostly hematological and is similar to that of other alkylating agents. Nausea and vomiting are infrequent. Alopecia does not occur at standard doses, and changes in renal or hepatic function have not been observed. Chlorambucil Pharmacological and Cytotoxic Actions The cytotoxic effects of chlorambucil on the bone marrow, lymphoid organs, and epithelial tissues are similar to those observed with the nitrogen mustards. Although CNS side effects can occur, these have been observed only with large doses. Nausea and vomiting may result from single oral doses of 20 mg or more. Absorption, Fate, and Excretion Oral absorption of chlorambucil is adequate and reliable. The drug has a half-life in plasma of approximately 1.5 hours, and it is almost completely metabolized (Alberts et al., 1979a). Therapeutic Uses The standard initial daily dosage of chlorambucil (LEUKERAN) is 0.1 to 0.2 mg/kg, continued for at least 3 to 6 weeks. The total daily dose, usually 4 to 10 mg, is given at one time. With a fall in the peripheral total leukocyte count or clinical improvement, the dosage is reduced; maintenance therapy (usually 2 mg daily) is feasible and may be required, depending on the nature of the disease. Other dosage schedules also are used. At the recommended dosages, chlorambucil is the slowest-acting nitrogen mustard in clinical use. It is a standard agent for patients with chronic lymphocytic leukemia and primary (Waldenström's) macroglobulinemia.
  20. Clinical Toxicity In chronic lymphocytic leukemia, chlorambucil may be given orally for months or years, achieving its effects gradually and often without toxicity to a precariously compromised bone marrow. Clinical improvement comparable to that with melphalan or cyclophosphamide has been observed in some patients with plasma cell myeloma. Beneficial results also have been reported in disorders with altered immune reactivity, such as vasculitis associated with rheumatoid arthritis and autoimmune hemolytic anemia with cold agglutinins. Although it is possible to induce marked hypoplasia of the bone marrow with excessive doses of chlorambucil administered over long periods, its myelosuppressive action is usually moderate, gradual, and rapidly reversible. Gastrointestinal discomfort, azoospermia, amenorrhea, pulmonary fibrosis, seizures, dermatitis, and hepatotoxicity may be rarely encountered. A marked increase in the incidence of leukemia and other tumors has been noted in a large controlled study of its use for the treatment of polycythemia vera by the National Polycythemia Vera Study Group, as well as in patients with breast cancer receiving long-term adjuvant chemotherapy (Lerner, 1978). Ethylenimines and Methylmelamines Triethylenemelamine (TEM), Thiotepa (Triethylene Thiophosphoramide), and Altretamine (Hexamethylmelamine; HMM) Pharmacological and Cytotoxic Effects Although nitrogen mustards have largely replaced ethylenimines in general clinical practice, this class of agents continues to have specific use. Thiotepa (THIOPLEX) is active as an intravesicular agent in bladder cancer and is used as a component of experimental high-dose chemotherapy regimens (Kletzel et al., 1992), and altretamine (HEXALEN), formerly known as hexamethylmelamine, is used in patients with advanced ovarian cancer after failure of first-line therapies. Both thiotepa and its primary metabolite, triethylenephosphoramide (TEPA), to which it is rapidly converted by hepatic mixed-function oxygenases (Ng and Waxman, 1991), are capable of forming DNA cross-links. The aziridine rings open after protonation of the ring-nitrogen, leading to a reactive molecule. Absorption, Fate, and Excretion TEPA becomes the predominant form of the drug present in plasma within 5 minutes of thiotepa administration. The parent compound has a plasma half-life of 1.2 to 2 hours, as compared to a half- life of 3 to 24 hours for TEPA. Thiotepa pharmacokinetics are essentially the same in children as in adults at conventional doses (up to 80 mg/m2), and drug and metabolite half-lives are unchanged in children receiving high-dose therapy of 300 mg/m2 per day for 3 days (Kletzel et al., 1992). Less than 10% of the administered drug appears in urine as the parent drug or the primary metabolite. The remainder is metabolized, interacts with biological molecules, or undergoes spontaneous chemical degradation. Clinical Toxicities The toxicities of thiotepa are essentially the same as those of the other alkylating agents, namely
nguon tai.lieu . vn