<<

The Mechanism of Action of Alkylating Agents G. P. WARWICK

(Chester Beatty Research In3tiiute, Institute of Cancer Research: Royal Cancer Hospital, @ Fuiham Road; S.W. 3, England)

SUMMARY The term “alkylating agent― has been defined and a detailed discussion made of the mechanisms (Sni and Sn@) by which they interact with nucleophilic centers. The various nucleophiles likely to be encountered in vivo are discussed in terms of their relative reactivities toward alkylating agents. With regard to Sn@ reactors it was possible to group them into four sections on the basis of their chemical reactivity and the “spread―intheir affinity toward different nucleophiles. The groups comprised (a) the epoxides, ethyleneimines, and $-lactones; (b) the primary alkyl methanesulfonates; (c) primary alkyl halides; and (d) a-halo acids and ketones. An attempt has been made to correlate the reactivity of the Sn@ and Sni reactors with their known pharmacological properties and the likelihood of their reaction with the various cellular constituents. In particular, the effects of the alkylating agents on nuclear material and cell division, the blood components, male fertility, and the im mune response have been discussed in detail, and the chemistry involved in the inter action of the alkylating agents with compounds of biological importance such as the nucleic acids, proteins, and peptides has been critically examined. The importance and limitations of the alkylating agents as chemothenapeutic drugs and as tools for examining the basic mechanisms involved in carcinogenesis and muta tion have been considered in the light of recent findings.

No attempt will be made to review the whole of in the elucidation of fundamental mechanisms of the literature relevant to the mode of action of action. alkylating agents, since this has been done many The term “alkylating agent―in its widest sense times in the past, and more recently by Ross (7@) denotes those compounds capable of replacing a and by Wheeler (87). Because of the growing com hydrogen atom in another molecule by an alkyl plexity of the field, it might be instructive to ex radical, and this of course involves electrophilic amine the chemical reactivities and alkylating po attack by the alkylating agent so that the defini tentials of some of the different classes of alkylat tion must be extended to include those reactions ing agents in relation to what is known about their involving addition of the radical to a molecule con pharmacological properties. The different types of taming an atom in a lower valency state—for ex compound which can function as alkylating agents ample, formation of a sulfonium compound from a or electrophilic reagents include alkyl halides, sulfide. The alkylating agents exhibit a diversity of alkyl methanesulfonates, alkyl sulfates, alkyl phos pharmacological properties including the capacity phates, halogenomethyl ethers, @2-chloroethyl sul to interfere with mitosis, to cause mutations, and fides, @-ch1oroethylamines, epoxides, @-lactones, to initiate and promote malignant tumors; some ethyleneimines and ethyleneimides, diazoalkanes, of them have a stimulatory action on the nervous activated ethylenic compounds, halogenomethyl system, and others are powerful vesicants or ketones and esters, methylolamines, etc., and, al lachrymators. When attempting to find a parallel though only a few of these classes are useful as ism between the pharmacological effects elicited palliatives in clinical therapy, all are capable of re by the alkylating agents and their chemical reac acting with at least some nucleophilic centers with tions at particular cellular sites, it is important to the result that even inactive compounds can help consider, among other variables, the particular 1315

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 1316 Cancer Research Vol. p23, September 1963

mechanism of alkylation involved, since this will Ogston (68) in the case of . Mustard largely determine the chemical groups which will gas (see later) yields a carbonium ion which is sta be attacked inside the cell and their position bilized in the form of a threo-membered ring, and within the cell. Almost every review which has this accounts for its greaten capacity to discrimi been written on the alkylating agents has included nate between anions. This stability is even greater at least some description of the mechanisms in in the case of the immonium ion formed from au volved in unimolecular and bimolecular reactions, phatic @-chloroethylamines due to the greater and the justification for indulging in this again in basicity of nitrogen than sulfur (Chart @). detail is inherent in the problem with which we are Each nucleophile (canbonium ion, immonium concerned. The alkylating agents, unlike many ion, etc.) has a competition factor which was meas other drugs used in medicine, produce their effects ured by Ogston in the case of mustard gas by by binding covalently with cell constituents, and observing the extent of reaction in aqueous solu tion with various anions (Table 1). In the case of compounds such as thiols, amines, and acids, ac R@ ___ • 9 @ —x R+X count must be taken of the amount in the reactive ® 9 form at a particular pH in considering likelihood @ R+Y RY of reaction with an alkylating agent. If one no

CHART 1.—Mechanism of first-order nucleophilic substitu members that alkylating agents pick out centers tion Sn 1 reaction. of high electron density, thiols and acids will be most reactive when ionized, in contrast to amines which become unreactive when protonated. If the R1 competition factor for a particular group is F, c is its concentration, and is the amount in the reac :(@2@ZC1 tive form, then in a particular system the relative V @cH2 amount of any group reacting is proportional to

CHART 2.—Ethyleneimmonium ion formed from an au phatic nitrogen mustard. TABLE 1 COMPETITION FACTORS FOR their sites and extents of binding, although de MUSTARD G@s (63) pendent to some extent on factors such'as solubil ity, will be determined mainly by their electro (Competi. SubstanceF . philic (or alkylating) potential; and this involves factor)pHThiosulfate tion consideration of ease of electron displacement, resonance energies, bond strengths, steric factors, Cysteine ethyl ester 1 .SX1O' 7 etc. Methylamine 3.9X10' 13 Phosphate 75 8 There are two generally accepted basic mecha Pyridine 54 7 nisms of alkylation, first-order nucleophilic sub Chloride 21 7 Acetate 10 7 stitution (Sni) and second-order nucleophilic Formate 3 7 substitution (Sn@), although care should be taken Nitrate 0.2 7 in making too rigid a classification. In the former, Glucose2.7X104 08 7 the rate-controlling stage in the reaction is ioniza tion to form a solvated carbonium ion, followed by a rapid combination with a solvent molecule or Fl c. However, Alexander (1) has drawn attention new nucleophilic reagent such as an anion or sul to the fact that apparent competition factors of fide (Chart 1). Since the driving force of the reac groups in macromolecules may be increased by tion is a displacement of electrons away from R, adsorption of the agent onto the macromolecule, then reactions of this type will be facilitated by the thereby increasing effective concentration. Ross presence of electron-repelling substituents (such as (74) assessed the reactivity of a number of chloro methyl) in R. The carbonium ion, being an ex ethylarylamines by measuring their rates of hy tremely unstable and hence reactive species, will drolysis in aqueous acetone and found that electron be to some extent indiscriminate in its combination releasing substituents in the benzyne ring such as with nucleophilic centers. The more unstable car p-methyl greatly increased the rate of hydrolysis, bonium ions, such as those derived from i-propyl whereas it was retarded by p-nitro-, p-phenylazo-, methanesulfonate, will tend to react rapidly with or other electron-attracting substituents. solvating water molecules, whereas in some cases The other essential features of Sni reactions more discrimination is possible as demonstrated by are (a) that the rate of any particular reaction will

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARwICK—Mechanism of Action of Alkylating Agents 1317 be essentially independent of both the concentra sition state is high compared with the initial state. tion and the nature of the other reagent. Thus, an Bond strength is obviously important, since, ionized thiol, a powerful nucleophile because of whereas electron affinity increases in the order the electron distribution on the sulfur atom, will I < Br < Cl < F, replacement by an alkylating react at the same rate as, e.g., an amine or hy agent increases in the reverse order, I being most droxide ion, less powerful nucleophiles, and (b) easily expelled. that no reaction will take place until primary Compounds such as primary alkyl halides, ionization has occurred. Both these factors are in methanesulfonates, and phosphates are not par contrast to what is observed in Sn@ reactions and ticularly powerful reactors, their rate of reaction are probably very significant when comparing the with water and other nucleophiles being usually pharmacological properties of the two classes of relatively low. Thus, whereas mustard gas has a reactant. half-life of 3 mm. in water at 87°C., ethyl meth Mustard gas (dichlorodiethyl sulfide) and the anesulfonate has a half-life of 36.3 hours and aromatic nitrogen mustards are the Sni reactors that of methyl bromide is 116 hours (75). The no with which we will be concerned in this discussion. action rate increases considerably in the presence Mustard gas is an extremely reactive compound of centers of high electron density, especially (half-life in water about 3 mm. at 37°C.) because of the presence and position of the electron-releas TABLE 2 ing sulfur atom relative to the atoms RELATiVE REACT1VITIES OF NUCLEOPRILES which facilitates formation of a carbonium ion TOWARD METHYL BROMIDE (84) through the unstable ring system shown. The carbonium ion (which may exist in the form of a Nucleophilek/kwNucleophilek/kwH,O threo-membered ring) can discriminate between nucleophiles to some extent, but it is far too reac CH,C00 500 Thiosulfate1.3X10' 4 .0X10 tive and, hence, toxic to be of any use in therapy Pyridine 4000 (Chart 3). }IPOi 6300 0H1 16000HS The aromatic nitrogen mustards were developed because they retained the capacity to react by an Snl mechanism; but the electron-withdrawing capacity of the aromatic ring greatly reduced the anions. Thus, the ability of Y to replace X in R-X rate of carbonium ion formation so that corn by the Sn@mechanism can be assessed by compar pounds could be synthesized which could reach ing reaction rate with V as compared with water. distant sites in the organism before reacting with This has been done by Swain for methyl bromide tissue constituents. (84). The major nucleophiic centers present inside cells are ionized thiol groups, ionized carboxylic ci@—@ci acids, ionized phosphates, and unionized amines, /1 acH,cH,—@-cI@ —k cicii@cii@scii,cij, and, ignoring accessibility and concentration fac tors, reaction would proceed most readily in the CHART 3.—Mechanism of ionization of mustard gas order thiol > amine > phosphate > carboxyl. Compounds such as a-halogenocarboxylic acids and ketones are highly reactive Sn@ reagents toward certain nucleophules by virtue of the elec Y@RL@—*Y_R + Xe tron-attracting capacity of the carbonyl group CHART 4.—Mechanism of Sn2 reaction (Chart 5) . However, for reasons to be discussed later, its reactivity toward weak nucleophiles such In second-order nucleophiic substitution a car as the carboxyl ion or the phosphate ion is low. In bonium ion is not formed, but a transition com this context a significant factor in considering plex is formed involving both reactants, and hence Sn@ reactors is the variation in relative reactivity the rate of reaction is dependent upon both con of a particular reagent toward different nucleo centrations (Chart 4). In this case the reagent philic centers. For example, helps in the breaking of the R-X bond, and this rate constant (thiosulfate) will occur most readily if the energies of the Y-R rate constant (acetate) and R-X bonds in the transition state are high, if the electron affinity of X is high compared with is @05forethyl methanesulfonate, 4380 for methyl that of Y, and if the resonance energy of the tran bromide, and 16,600 for the iodoacetate ion (75,

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 1318 Cancer Research Vol. @Z3,September 1963

p. @).Thus, it is obvious that the “spread―inno halides leads to a further drop in Sn@ rate, but activity toward different nucleophilic centers now the Snl mechanism becomes of importance, varies considerably from reagent to reagent and is and the two mechanisms operate simultaneously of fundamental importance in determining the and at a comparable rate (37) . The reason for reactions which a particular drug will undergo this is that the extra methyl group facilitates pni in vivo and hence its pharmacological properties. mary ionization of the halide ion (Sni) but retards The factors which determine whether an agent the Sn@process for electronic reasons and because will react by an Snl or an Sn@ mechanism are it hinders the approach of the nucleophile and sub obviously complex, and it is dangerous to be too sequent formation of the transition complex (pni rigid in this type of classification, since even sol many stenic effect) . The presence of three methyl vent changes can sometimes lead to a modification groups as in tert-butyl halides causes a complete in mechanism. There occur also those compounds changeover to the Sni mechanism, and the rate of such as the epoxides and ethyleneimines which, by hydrolysis becomes faster than the Sn@ hydrolysis virtue of their strained three-membered ring struc of methyl halides under the same conditions (61). ture, fall into a different category from most other A change in mechanism can also be observed Sn@ reactors, as will be considered later. In the sometimes if an attacking nucleophile is replaced case of alkyl halides and alkyl methanesulfonates, by one of less nucleophilic potential under the the rate and mechanism of reaction depend on same experimental conditions. At first the effect will be to decrease the velocity of the Sn@ process, e ‘@2 but with even weaker nucleophilic reagents a point x + IcH,—@c-oH—@x--cH,cooH+ I will be reached in which the reagent is no longer able to assist in the breaking of the R-X bond CHART 5.—Mechanism of reaction of iodoacetic acid which must, therefore, occur unaided by the reagent, and the mechanism will become Sni cHa CH@ probably coupled with a decreased reaction rate IS (39). For example, in the substitution of the tn @ CH,S0,O•CCH3. coso,cH3 @H35020(@H2@,0S0PI3methylsulfonium ion by various anions it was

CHART 6.—Dimethylmyleran and Myleran found that the specific rates were OH(743), PhO—(13.4), CO@(7.4), Br(7.8), Cl(7.3), whether the reactant is primary, secondary, or ten ously indicating an Sn@ mechanism for the first tiary. In some cases, such as in the reaction of two anions and an Snl mechanism at a lesser rate alkyl bromides with pynidine, the rate of reac for the others. A situation similar to this probably tion to form an ethylene derivative is in the order occurs in the case of a-haloketones and acids, mdi tent. > sec. > prim., whereas in the simultaneous eating their tremendous potential to react with reaction to form quaternary salts the order is no powerful nucleophiles such as ionized thiols by an versed. The differences in the reaction rates of Sn@ process and their relative unreactivity toward primary, secondary, and tertiary alkyl compounds weaken nucleophiles such as ionized acid groups are generally attributed to the inductive effect. (hence the misnomer “specificthiol reactors―). The replacement of a hydrogen atom by a methyl The earlier data concerning the alkyl halides is group is acid-weakening—i.e., it induces a nega probably of fundamental importance in consider tive charge on the atom on which substitution has ing the biological properties of the various agents. occurred. It therefore increases the ease of ioniza For example, the greater reactivity of methyl tion of a halide ion from the molecule. This ease halides (or methanesulfonates), as compared with of separation can be more than compensated for the corresponding ethyl derivatives, probably ex by the smaller attraction of the reacting base. Con plains the greater toxicity of the former in vivo. sequently, the primary halides require a stronger One surprising finding is the fact that dimethyl basic reagent than the tert-halides, but this order myleran (@,5-dimethanesulfonyloxyhexane) is so is reversed when the reagent acts as an ionizing similar to Myleran (1,4-dimethanesulfonyloxybu rather than a basic reagent (61). tane) in its pharmacological properties (see later) In general methyl halides react almost exclu (Chart 6). Unlike Myleran, it will tend toward an sively by an Sn@ process, but introduction of an Sni mechanism in its reactions, forming a very Un other methyl (electron-releasing) substituent (giv stable carbonium ion extremely reactive toward ing ethyl) leads to a reduction in the rate of reac its own solvating water in aqueous media and tion by the Sn@ mechanism, the factor usually be hence unreactive toward other nucleophiles. How ing about 10, as in the reaction with amines. Intro ever, the possibility of its reacting by an Sn@ duction of another methyl group as in iso-propyl mechanism may become more important in non

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARwICK—Mechanism of Action of Alkylating Agents 1319 aqueous media, and the implications of this will be anticipate epoxides and ethyleneimines to react considered later. readily with a range of nucleophiles in a somewhat Striking differences in pharmacological effects similar way to the mustards. They have in fact have been found on the one hand between alkylat been shown to react readily with thiols, amines, ing agents which react essentially by an Sni and acids (73), and this ease of reaction with di mechanism, and many of those which react by the verse nucleophiles is reflected in their biological Sn@ mechanism, between monofunctional and bi properties, since they produce in many respects functional agents of both types, and between mdi end-products (inhibition of mitosis, mutation, vidual members of the Sn@ type. The differences chromosome breaks, cancer, effects on the hemo observed between the various Sn@ reactors are poietic system) very similar to the mustards (35). probably explicable on the basis of the foregoing Group (b), the primary methanesulfonates, differ discussion, and from the point of view of chemical from the epoxides in their potential of reactivity, reactivity and “spread―inaffinity toward differ because no strained ring is present and the Sn@ ent nucleophiles the Sn@ reactors can be broadly mechanism will mainly operate in which bond classified into four groups: making will be more important than bond-break a) the epoxides, ethyleneimines, and fl-lactones; ing, so that both the concentration of an attack b) the primary alkyl methanesulfonates (e.g., ing reagent and its nucleophilicity will be of prime Myleran and ethyl methanesulfonate); importance, although it should be mentioned that c) primary alkyl halides (e.g., dibromobutane); and, ® 0® d) a-haloacids and ketones (e.g., iodoacetic R—CH—CH2+X + H—@RCH--CH,X acid and co-bromoacetophenone). Group (a) .—Compnises compounds which bear @1 6H a resemblance in structure to the ethyleneim monium ions which are derived from aliphatic CHART 7.—Mechanism of epoxide reactions nitrogen mustards and which react by an Sn@ mechanism with nucleophiles (see earlier) (Chart the true Sn@ tendency will be greater in the 7). methyl esters than the ethyl. Because of their rela Epoxide reactions can be acid-catalyzed, but tive unreactivity compared with epoxides, there this will probably be of little significance in vivo. is found a greater “spread―in their affinity toward The presence of a strained three-membered ring nucleophiles, and their reactivity toward thiol in these compounds makes them somewhat unlike groups will be far greater than toward amines on the other Sn@Zreactors in the finer details of their acids. In this connection Brookes and Lawley (@1) reaction kinetics. In a normal Sn@ process, in have found that both ethyl methanesulfonate and which an entering nucleophile causes the expulsion Myleran have much less reactivity toward guanine of a weaker nucleophile such as in the reaction of in DNA than the mustards and epoxides, whereas an ionized thiol with methyl iodide, the process of Roberts and Warwick (65, 66) have demonstrated bond-making is considered a more important dniv that almost all the urinary reactivity excreted ing force of the reaction than that of bond-break after injection of these compounds could be ac ing. In the transition state of such a reaction the counted for by reaction with the powerfully process of bond-making (new -S-C-bond) will al nucleophilic thiol group in vivo. In the case of ready have progressed to a considerable extent. If ethyl methanesulfonate much of the C'4 in the as mentioned earlier the process of bond-breaking drug was exhaled as C―02which could have been (-C-halogen) is the more important driving force derived either by hydrolysis of esters formed by of the reaction, then the reaction kinetics tend reaction with carboxyl groups on phosphate toward unimolecular and the rate of ionization be groups. In the case of Myleran, much less of the comes the rate-determining step. There are obvi drug was metabolized in this way. The capabilities ously gradations between the two extreme mecha of an alkylating agent to react with various nucleo nisms as already mentioned, and the epoxides of philes is reflected faithfully in the biological and biological interest fall into a class in which, while pharmacological effects which it elicits, leading in second-order kinetics are followed, the process of the case of these two agents to effects much less bond-breaking has progressed to a considerable potent than were observed with the mustards, extent in the transition complex because of the epoxides, etc. Thus Myleran is much less toxic strained three-membered ring structure. Since than mustard gas, possesses none of its vesicant bond-breaking is a more important driving force properties, has only moderate activity against than bond-making in such reactions, one would transplanted tumors, causes far fewer chromosome

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 13@20 Cancer Research Vol. @3,September 1963

“breaks―(33),and, although it decreases the num action is a contributor and that the site at which ber of circulating granulocytes, it has far less ef it occurs and its extent differ with those of the feet on the peripheral lymphocyte count than the various agents leading to different biological ef mustards and epoxides (30) ; and its effects on the fects. There is no proof yet which alternative is immune response and on male fertility in rats correct, and yet certain vital experiments, which again differs fundamentally from the mustards remain to be done, may make the situation clearer. and ethyleneimines (see later). The past literature is full of examples implicating Group (c), comprising the primary alkyl halides, the importance of thiol reactions in all sorts of bio represents, with the exception of the methyl corn logical phenomena (10), and yet a really critical pounds, agents even more unneactive than the pni assessment of some of these claims can leave little mary alkyl methanesulfonates. Differences in the doubt that they were based on naï or over nucleophilicity of the Br and CH3SO2O anions simplified assumptions leading to a very confused coupled with differences in the C-Br and C- situation. The potential importance of thiol neac OSO2CH3bond strengths and a weak capacity tion remains (58), and yet more research needs to to undergo Sn@ reactions makes the halides (e.g., be done to discard outmoded ideas and to bring ethyl bromide and dibromobutane) even poorer the whole field into line with modern concepts. candidates than the methanesulfonates as far as Speculative “paperchemistry― leads us nowhere, diversity of reactions in vivo are concerned. Thus a except possibly in circles, and there is no sub decreased reactivity toward all nucleophiles is oh stitute for doing the vital experiments which can served, so that even fewer biological effects would finally clear up these points. More will be said on be anticipated; and this is so, since 80 mg. of di -SR reaction later. bromobutane can be administered safely to a Group (d) comprises those agents, the a-halo 300-gm. rat compared with 4 mg. of Myleran, and acids and ketones, which in their reaction mecha this compound produces none of the pharmaco nism lie in the other extreme from undoubted Sni logical effects observed with Mylenan. The in reactors. With these potent Sn@ reactors the creased reactivity of methyl compounds has been process of bond-making far outweighs in im discussed earlier, and this is reflected in the high portance the process of bond-breaking, and so the toxicity of methyl bromide compared with ethyl nature of the attacking nucleophile is vitally im bromide. portant, leading to extreme reactivity toward Roberts and Warwick showed that dibromo powerful nucleophiles such as ionized thiol groups butane reacted with cysteine in t@itroin a manner and a very weak level of reactivity toward less similar to that of Myleran leading to a sulfonium nucleophilic centers. Thus as mentioned earlier ion (66). Since dibromobutane lacks the pharmaco k(thiosulfate) logical properties of Myleran, this evidence could k(acetate) be used to discount the importance of such a reac tion in vivo and, indeed, the importance of thiol is 16,600, as compared with 9205for ethyl methane reaction in general. However, such reasoning is sulfonate. probably not valid, because it does not take ac Probably many resonance forms contribute to count of the different thiol groups in the cell, their the transition state (920), but the peculiar reactiv different reactivity, or the difference in alkylating ity of compounds of this type is a direct result of potential between Myleran and dibromobutane. the presence of the powerful electron-attracting The major thiol reaction in vivo will be with gluta carbonyl group. Because of this, compounds of this thione, which will probably have little permanent type were once regarded as specific thiol reactors, effect on cellular function ; reaction with thiol pro and all their effects from enzyme inhibition to gen teins or enzymes might have greater significance, eral toxicity in vivo were ascribed on the basis of and there is no evidence yet that dibromobutane -511 reaction, despite warning by Michaelis and will react with these in vivo. The site of thiol reac Schubert (59) that amine reaction could also oc tion inside the cell is probably very important, cur. In this connection Anson and Stanley (5) later since it is now almost certain that every alkylating showed that, although tobacco mosaic virus was agent will combine extensively with thiol groups not inactivated by p-chionomercunibenzoate (a in vivo, and yet compounds such as the mustards, powerful 511 reactor), it was inactivated by iodo rnethanesulfonates, halides, and a-haloketones acetarnide and iodoacetic acid, implicating reac and acids manifest very different effects on cells. tion at another site. More recently, Stark, Stein, There are two conclusions to be drawn from this: and Moore (83) have shown that the inactivation (a) that thiol reaction in no case contributes to of nibonuclease is due to reaction with the ring pharmacological end-effects, or (b) that thiol re nitrogen of histidine. This reaction with tertiary

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARWICK—Mechanism of Action of Alkylating Agents 13f21

amines is of interest, since Roberts and Warwick THE EFFECT OF ALKYLATING AGENTS ON (67) have shown that bromoacetic acid reacts Nucj@n MATERIAL ANDCELL DIVISION readily with the punine portion of deoxyguanylic When it was discovered that in many respects acid and that a low level of reaction with the the bifunctional chloroethylamines and chioro guanine of DNA also occurred. It remains to be ethylsulfides produced cytological and pharmaco seen how significant such reactions are in tiivo, logical effects (e.g., cell death, chromosome ab however, because of the high competition by other normalities, degenerative changes in the bone groups such as thiol in the cytoplasm. In this con marrow and testis, suppression of the immune no nection, the same authors have shown (67) that, sponse) similar to those produced by ionizing radia after incubating Krebs ascites cells with C'4- tion, the term “radiomimetic―was applied to them bromoacetic acid, there is a steady uptake of label (@8). Although certain of the alkylating agents and the ratio produce end-results which appear superfluously specific activity (protein) like those produced by radiation, detailed investi gation has often revealed significant differences specific activity (DNA) between them, and the term in many respects leads was approximately 2@; and mthas yet to be estab to confusion in the minds of chemists and biolo lished whether any DNA reaction other than that gists alike. Loveless and Revell (56) set forth their associated with residual DNA protein has oc precise biological criteria upon which chemical curred. It is significant that in the case of mustard substances were to be classified as radiomimetic, gas and Myleran the ratio was nearer 1 (@1). and this involved restriction “tothose substances From the foregoing discussion it is probably true which can be shown unequivocally to react upon to say that, with some exceptions, the different the resting cell in such a way as to produce an types of alkylating agent are potentially capable alteration of the genetic material which is revealed of reacting with all nucleophilic centers and that by the appearance of chromosome breakage and the pharmacological properties of a particular rearrangement in subsequent divisions.― Certain agent will depend on (a) ease of penetration, (b) to compounds which caused pyknosis, and which some extent mechanism of alkylation, and (c) Dustin had referred to as radiomimetic, were not “spread―incapacity to react with different nucleo covered by this definition. Others have preferred to retain the original broad definition, but since philes. The mustards, for example, probably pro this time so many alkylating agents have been duce such diverse effects because they can easily discovered to mimic superfluously the effects of penetrate cells, their chance of reaching most radiation in, say, one respect only, that the term parts of the cell is high because of their unreactive ceases to have great significance. Ethyl methane nature until they are ionized (a relatively slow sulfonate, for example, shows none of the properties process), and the carbonium ions once formed are of radiation except that it is a powerful mutagen reactive toward most nucleophiles, leading to (53). Mylenan, possibly atypical in some respects, some reaction at most parts of the cell. The van nevertheless has a profound destructive effect on ous Sn@ reactors fall into different categories, and granulocyte production (3@), causes certain chro compounds such as the epoxides and ethylenei mosomal abnormalities (45), depresses the im mines produce effects similar to the mustards be mune response (1@) (in a manner very similar to cause of their relatively low reactivity which en radiation) and causes infertility in rats (48) (also ables them to reach most parts of the cell un in a manner reminiscent of radiation) ; but, unlike altered, coupled with a satisfactory gradient in the certain mustards, it has only a moderate inhibit “spread―oftheir affinity toward different nucleo ing effect on the transplanted Walker rat car philes. cinoma. Thus these compounds would seem to be “radiomimetic,― partially “radiomimetic,― or Some of the biological and pharmacological “nonradiomimetic,― depending on the particular properties of the alkylating agents will now be dis biological phenomenon being considered. Koller cussed in the light of their chemical reactivities. (46) has compared the effectsof several alkylating The time will surely come when the barriers which agents with those of radiation and concludes divide chemistry from biology must crumble and “thatin view of the dissimilarities, it would be an they will merge into one conceptual framework. error to infer a similarity of the mode of action of If the static concepts of chemistry could be made alkylating agents and x-nays, basing the inference more dynamic, and the dynamic aspects of biology on the similarity of some end products.― more static, then the meeting point would seem Of the various types of alkylating agent which much nearer. have been tested for tumor growth-inhibiting

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 13@2 Cancer Research Vol. 23, September 1963 properties, those which have emerged as the most function by direct combination with cellular com active include the Sni reactors such as the aro ponents such as nuclear and cytoplasmic material. matic nitrogen mustards and the Sn@ reactors of The most “radiomimetic―chemicals will be those the HN@ type, epoxides, ethyleneirnines. Other capable of penetrating deep into the cell and of no Sn@ reactors such as the aliphatic methanesulfo acting with a diversity of cellular sites, whereas nates (Myleran) show a low order of activity less reactive chemicals or those which alkylate by against the transplanted Walker tumor, whereas a different basic mechanism may possess far fewer those Sn@ reactors such as the a-halo acids and “radiomimetic―properties. ketones have been found to be inactive while Tumor inhibition with the alkylating agents no possessing a high order of general toxicity. The sults from their cytotoxic action, which probably other factor of importance in determining efficacy is the result of a complex series of changes varying in this respect is functionality, and it is now well in fundamental mechanism with different types of known that bifunctionality leads to a profound alkylating agent, and even from cell type to cell increase in capacity to interfere with mitosis, pro type. Lethal cytotoxic action is usually of two vided the reactive centers are spaced at an opti types called interphase death and mitotic death. mum distance apart. Thus in the chemical sense Interphase death occurs without cells' going into greatest activity is associated with undoubted division or attempting to go into division. Alkylat Snl reactors or those Sn@ reactors which, as dis ing agents vary in their capacity to kill cells at this cussed earlier, show some tendency toward the stage, and individual cell types vary in their sus Sni mechanism in the bond-breaking and bond ceptibility to this type of death. Thus fully dif making process and which have a reasonable no ferentiated lymphocytes tend to be extremely activity toward most nucleophiles. Biesele et al. sensitive to the action of mustards (but not to (18), in discussing the chemistry of alkylating Myleran), whereas most other differentiated cells agents, postulated that the most active compounds do not. The mechanism of interphase death is in causing chromosomal aberrations were those probably complex and results from damage to which possessed or could be transformed into un many cellular sites. stable threo-membered rings. In a comment on this In mitotic death the primary damage is thought postulate, Loveless and Ross (57) drew attention to occur in the “nestingstage,―and chromosome to the “radiomimetic―properties of 1,4-dimeth aberrations and other effects become cytologically anesulfoxyloxybutane (Myleran) in inhibiting apparent at the time of division. Many changes the growth of Sarcoma 180 and in its effect on the occur, such as clumping and bridging of chromo blood and bone marrow, and the capacity of somes, and some of the effects are cumulative dur dimethylsulfate to produce chromosome breakage ing successive cell divisions leading eventually to in plant material. Since both these compounds are nonviable cells. The action is not restricted to incapable of forming three-membered ring sys malignant cells, but extends to proliferating cells tems, the postulate of Biesele et at. was obviously of all types (35). an oversimplification, and Loveless and Ross (57) Various differences in detail of action have argued that capacity to form a carbonium ion was been reported to exist between the mustards (e.g., a more general feature of “radiomimetic―chemi HN@) and the methanesulfonates (e.g., dimethyl cals. Actually, it is doubtful whether epoxides, myleran) . Thus in comparing their actions on ethyleneimines, and methanesulfonates do, in fact, lymphoma cells in culture, Alexander (3) found form true carbonium ions under conditions likely that HN@ immediately arrested cell division, to be met with in the organism, and it is difficult whereas dimethylmyleran behaved like x-rays, therefore to accept this generalization as it is and at least one division occurred before cell pro worded. As indicated earlier the term “nadiomi liferation stopped. Koller (45) has compared in metic― is often applied far too indiscriminately to detail the cytological effects produced by various sometimes ill-defined biological phenomena, and alkylating agents and again found differences in it would seem more profitable and less misleading their modes of action and also differences in no to define carefully the biological effects manifested sponse to varying doses of alkylating agent. He by each alkylating agent rather than to refer to found that in a series of dimethanesulfonyloxy some property as “nadiomimetic―because it may esters, the greatest diversity of effects on cells of vaguely correspond to a particular property of the Walker tumor was shown by Myleran. He also radiation. Any apparent similarity between the found that it could produce suppression of mitosis, biological properties of radiation and the alkylat and “radiomimetic― and cytotoxic effects, al ing agents undoubtedly results from the capacity though these probably differed in detail from those of each entity to modify cellular structure and observed with the mustards. Alexander reported

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. \\ARWICK—Mechafl?$m of Action of Alkylating Agents 13@23

(3) that, after all treatments, there was an increase poiesis was prolonged, cytological abnormalities in volume of cultured lymphoma cells; with HN@ were not seen (cf. 45) at the doses used, and mitosis the cell volume doubled, and then there was de proceeded normally. They explain these different generation, whereas with x-rays and dimethyl effects on the basis of the capacity of the mustards myleran a wide spectrum of cell sizes was obtained, to attack ruthlessly nuclear material in the resting including “giants,―although nuclear granulation stage and to lyse fully differentiated lymphocytes, similar to that seen with HN@ was not observable whereas they consider that Myleran prolongs the after @4hours. intermitotic interval in all proliferating cells inre Alexander' has also found that iodoacetic acid spective of the stage in the mitotic cycle at which produces different effects from either the mustards they are exposed. Differences in normal inter or the methanesulfonates on lymphoma cells. It mitotic levels which vary with different cell types seems to have little effect on cell division, but could explain the effects. Lajtha (47) has mdi when a critical dose level is reached the cells cated that erythropoietic cells normally divide break up. This would be consistent with the no about twice as often as granulopoietic cells, and activity of this compound mentioned earlier, re delay in each series would explain why depletion sulting in its major reactions occurring in the cy of erythropoiesis occurs more rapidly than that of toplasm (67). The need for bifunctionality and a granulopoietic cells. The authors do not consider certain critical level of nuclear reaction (whether that Myleran has any selectivity for particular this be with nucleic acid, protein, or cofactor) seem cell types and “thatthe apparent resistance of essential for “radiomimetic―cell death. “Cyto lymphoid tissue to Myleran may depend as much plasmic death― may be a factor of importance in on the ratio of mitotic frequency to life-plan as to considering mustard action on fully differentiated the lack of damage to mature lymphocytes.― lymphocytes and indeed in the inhibition of tumor In their effects on the circulating gnanulocytes growth in general. Koller (45) considers that many and lymphocytes, Elson has found (30) that the factors must be considered when assessing the epoxides and ethyleneimines produce effects simi causes of cell death in a dynamic population and lan to those produced by the mustards—not sun that “theability of the alkylating agents to pro pnising in view of the similarity in reaction kinetics duce nadiomimetic injuries may not be the most discussed earlier—whereas he found the effects important property that determines the effective produced by the carcinogens 4-aminostilbene and ness of these drugs in the therapy of cancer.― the polycyclic hydrocarbons to be more like those of Myleran, a result difficult to explain at present EFFECTS ON THE BLOOD but possibly of great significance. Ethylmethane Elson (30) showed that the effects of whole sulfonate (half-Myleran) was not active. body x-radiation on the number of circulating As in the case of tumor growth-inhibition, bi platelets, lymphocytes, and neutrophils could be functional alkylating agents of all series were much essentially reproduced by the administration of more efficient than the corresponding monofunc suitable combinations of (a bifunc tional derivatives, and in the Mylenan series opti tional mustard) and Myleran, and he concluded mum activity was found when n = 4 or 5 (31). therefore that each of these chemicals was only Potent Sn@ reactors such as a-halo acids and partially “radiomimetic.―Elson, Galton, and Till ketones apparently have not been tested as ex (3@) consider that Myleran appears to act on the tensively as the foregoing compounds, whereas resting stage and prolongs the intermitotic inter alkyl halides such as dibromubutane (group [c] val, whereas mustards appear to act mainly on Sn@2reactors) possess no detectable activity (note cells undergoing mitosis and possibly during the lower chemical reactivity). latter part of the DNA synthetic period. They also exert a particularly destructive action on mature EFFECTS ON MALE FERTILITY lymphocytes which Mylenan does not. They found Very interesting results which may contribute that the effects of Chlorambucil appeared rapidly, to an elucidation of the mode of action of x-rays were transient, and were accompanied by con and the alkylating agents have been found by com spicuous mitotic abnormalities. Recovery pro paring the effects of these agents on the process of ceeded rapidly except in the lymphoid tissue, and spermatogenesis in rats and mice. The process fol there was a transient overcompensation in the lows the path spermatogonia —@spenmatocytes —+ granulocyte series. Myleran, on the other hand, spenmatids —+spermatozoa in epididymis, and it produced gradual effects, depression of hemo has been found that different agents can affect dif

I p@ Alexander, personal communication. ferent stages in the process. Once again it tran

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 18@4 Cancer Research Vol. @23,September 1963

spires that Myleran most closely follows the effects was found that irradiation and Myleran sup of radiation. pressed antibody production maximally if given Intrapenitoneal injection into mice of aiiphatic about @—4daysbefore the antigen, whereas nitno mustards produced destructive changes in the gen mustard and TEM did so when given about @ testis, and recovery was observed 3—4weeks later days after the antigen, and antibody production (48), whereas in rats potency returned after @—4could be suppressed for some weeks in each case, months (34) . It was of interest, however, that two following a single injection given at the appropni aromatic nitrogen mustards, Melphelan and Chlo ate time interval. Once antibody production is rambucil, did not interfere with the fertility of under way, however, these agents have no further male rats (4@, 44) at doses which were adequate inhibiting effect. Berenbaum tentatively explains to cause transient leukopenia in the peripheral these relative effects by assuming that immuno white count (@9) and inhibit tumor growth (13). logically competent cells are produced from pnimi Jackson (43) concludes, therefore, that capacity to tive stem cells which give rise to more differenti inhibit tumor growth or to interfere with the bone ated cells which can produce antibody. He en marrow does not imply capacity to damage visages early cell types to be sensitive to the action germinal epithelium. This, of course, may have of x-nays and Mylenan, but not to TEM or the been due to poor penetration rather than to a dif mustards, whereas the later, more differentiated ferent mode of action. forms are sensitive to the latter agents, but not to 2,@,6-Triethyleneimino-1,3,5-triazine (TEM), the former. Other alkylating agents have unfortu which was placed with the epoxides as far as nately not yet been tested, but the foregoing no chemical reactivity is concerned, produced effects sults will be discussed later in the light of known indicative of selective interference with certain reactions of these agents in vivo. stages of spermatogenesis (19), mainly, probably, with spermatocytes (11, 4@), since the infertility CHEMICAL INTERACTIONS WITH was found at 4 weeks and recovery was rapid after CELL CONSTITUENTS a single dose of 0.@mg/kg. Five daily doses of 0.@ General conaiderations.—In discussing the basic mg/kg, however, caused sterility for several weeks, chemistry of the alkylating agents it was found and early and late stages of spermatogenesis were logical to separate these compounds into various affected, and probably mature sperm in the epi types or groups depending on whether they were didymis were also attacked (as with the mustards). essentially Sni reactors on Sn@ reactors, and fun The relative insusceptibility of spermatogenia is ther subdivision was possible when a more detailed in contrast in that they are the germinal tissue assessment was made. In considering some of the most sensitive to radiation (6@, 79) and to My pharmacological properties of the alkylating leran. agents, it was found that, where comparative After one dose of Myleran (11 mg/kg) the testis assessments have been made, in a broad sense histology remained normal until the 8th week, these faithfully reflected the type of alkylating when sterility rapidly developed (@6, 4@), indicat agent involved. Thus, one can distinguish between ing an effect on an early stage of spermatogenesis. the pharmacological properties manifested by the The other germinal cells present at the time of mustards, the epoxides and ethyleneimines on the treatment appeared to develop normally for one hand, and the alkyl methanesulfonates on the about 45 days, resulting in a systematic depletion other, in considering their pharmacological effects. of spermatogonia, spermatocytes, spermatids, and Although fewer biological studies have been car spermatozoa, in that sequence (4@). Increased ned out on the alkyl halides and the a-haloko doses still did not affect other spermatogenic cells tones and acids, it is evident that again these differ which continued to proliferate. from one another and from the above compounds. Dimethylmyleran produced similar effects at a Since these complex processes involving cells lower dose level (44). This apparent specificity of and alteration of cellular structure and function action of Myleran is probably in some way related almost certainly involve direct chemical combina to its chemical reactivity. The fact that the effects tion with cellular constituents, and since the level which it produces are similar to those produced by of any particular combination will depend on radiation and unlike those produced by the mus many variables, the most important of which is tards and TEM is probably highly significant. probably chemical reactivity, then an examination of cellular components after treatment, or of EFFECTS ON THE IMMUNE RESPONSE (1@) altered biochemical integrity after treatment, Different effects were found on the immune no should help to pin-point the lesions involved. sponse, depending on the agent tested. Thus, it This type of reasoning, valid on misleading, has

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARwICK—Mechanism of Action of Alkylating Agents 13@2@5 guided most workers studying the mechanism of cellular constituents, the nature of this binding in action of alkylating agents for many years. The some cases, and that in vitro studies can often help general opinion seems to be that, in the case of to elucidate tantalizing in vivo problems. The compounds such as the carcinogenic hydrocarbons, progress made so far, much as it seems when one the carcinogenic azo dyes, etc., we must keep an assesses it against the almost complete void which open mind whether these agents cause their effects existed 10 years ago, is still extremely limited; and, by actually binding to cellular constituents on although we are in a position to make some intelli whether they are doing something much more gent guesses as to the mechanism by which the subtle—such as finding their way between macro alkylating agents exert some of their effects (e.g., molecular strands and exerting their effects by es their cytotoxic action), it would be unwise to make sentially physical means ; however, in the case of any dogmatic assessments at this stage, no matter the alkylating agents, cellular binding is of prime how justifiable they may seem. However, the sun importance. face has been scratched, and doubtless the task of To digress a little, there is good reason for be a reviewer at the end of another 10 years will be a lieving that the carcinogenic hydrocarbons and comparatively reassuring one. azo dyes do, in fact, need optimum structural re Metabolism of ethyl methanesulfonale and My quirements and particular electron distributions leran.—The first clear-cut demonstration that a (7) which may enable them to alter cellular func pharmacologically active alkylating agent did in tion by some kind of physical means; but there is fact alkylate nucleophilic centers in vivo was oh growing evidence that compounds of this type can tamed when Roberts and Warwick (65) demon in fact bind firmly to cellular constituents, both to strated the presence of N-acetyl-S-ethylcysteine proteins and to nucleic acids. To quote but two and related compounds in the urine of rats which examples, Heidelberger (36) has shown dimethyl had received C'4-ethyl methanesulfonate by intra benzanthracene to combine both with protein and penitoneal injection. The bulk of the radioactivity DNA in vivo, whereas the Millers (60) and others which appeared during the first few hours was have shown dye binding to liver proteins following probably derived from reaction with glutathione administration of carcinogenic azo dyes, and, more known to be present in high concentration in recently, Roberts and Warwick (68) have shown mammalian cells and with which ethyl methane the combination of a metabolite of dimethyl sulfonate reacted readily in vitro. That which ap aminoazobenzene with nucleic acids in vitro. peared later probably came from ethylated pro These findings tend if anything to cloud the com tein. Because of the related structure of 1,4- plicated situation even further; they may be mis dimethanesulfoxyloxybutane (Mylenan) and its leading, but if direct covalent combination with enchanced pharmacological properties, it was con cellular constituents is proved to be a potent force sidered of interest to examine the urinary products in producing all the pharmacological properties of excreted after its injection. Because of the bifunc the alkylating agents, then these findings might tional nature of the drug it was considered that be of the utmost importance in elucidating at least some type of cross-linked product might be formed some of the phases of the complex carcinogenic in vivo. However, the major urinary metabolite process with these and other agents. was found to be 3-hydroxytetrahydrothiophene In the case of the alkylating agents where one 1,1-dioxide (66), which was probably derived from compound can be cytotoxic, carcinogenic, and a sulfonium compound formed by the interaction mutagenic, the correlation between chemical reac of both alkylating arms with one thiol group. This tions at the cellular level with any of these mani was followed by the removal of tetrahydrothio festations becomes exceedingly difficult. However, phene and its further metabolism to the hydroxy attempts have been made to study the types of sulfone. The reaction is of interest, because it in chemical reactions which some of these reagents volves not only modification of an existing thiol can undergo inside cells, and in the main the no group but its later removal from the amino acid sults obtained have indicated a close parallelism moiety in the form of tetnahydrothiophene. In the between the types of reactions observed in vitro case of protein reaction, this could lead to the pro using purified chemicals and those observed in duction of a new sequence of amino acids by the vivo, and it has also been found that levels of bind replacement of an existing molecule of bound ing to various constituents were as one might cysteine by a molecule of bound lanthionine or have anticipated on the basis of the reactivity of bound aminoacrylic acid (69). Bound lanthionine the drug concerned. These studies are probably would result if the sulfonium compound was at important if only to show that definite covalent tacked by another molecule of ionized glutathione, binding can occur between alkylating drugs and whereas bound aminoacrylic acid would result by

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 13@?26 Cancer Research Vol. p23, September 1963

hydrolytic decomposition (Chant 8). The dehydro one glutamic acid unit by valine in a polypeptide peptide or dehydroprotein might undergo addition chain of 300 units. However, these ideas are specu of a molecule of glutathione or other thiol forming lative, and more data regarding the extent of pro bound lanthionine. tein alteration by Myleran in vivo and the func It is extremely difficult to assess the importance tions which these proteins play in cellular integrity of these reactions when considering the mechanism are needed. The related compound, dimethyl of action of Myleran. It has been stated (@1) that myleran, which in many respects is similar to My the low level of reaction of Mylenan with nucleic leran in its pharmacological properties, should be a acids in vitro and in vivo at therapeutically effec useful tool in elucidating some of these problems, tive doses makes these centers seem unlikely vital since if reactivity toward anions is low and for points of attack in this case. It is possibly relevant steric reasons ring closure and dethiolation are not that, in the series of dimethanesulfonyloxy esters, possible in this case, then the relevance of these those having four or five methylene groups are the changes to the mode of action of Myleran would be most active in depressing granulocyte count and questionable. Work on this aspect of the problem is are also those which could undergo the ring closure in progress, with highly tritiated dimethylmyleran. In connection with the mode of action of Myleran, NHR Timmis (85), on the basis of its similarity in CH2—CH2o I action to thioguanine, has suggested that it might 012 CH2@― .—@ I ::s @ G@ CH2—Q1@ combine with nucleic acids in vivo or with nucleo tides to form a type of which could act by lengthening intermitotic intervals rather ,NHR than directly killing cells. He suggested that + Gs—cIl,—cH Myleran might first interact with homocysteine to e @COR' form a compound capable of alkylating purine bases in a similar manner to adenosyl-methionine. i@ NHR r2_cI@4H_c@ CH,—cH2 ,NHR Metabolism of aromatic nitrogen mustards and —* I S+@H@zC mustard gas.—Itwas considered of interest in view cii@—cH@ @0R' cH2—cH2 \coR S of the ring closure reaction of Myleran to deter mine whether similar reactions occurred in vitro CHART 8.—Reactions of the sulfonium ion formed from Myleran and a cysteinyl moiety with glutathione and with and in vivo in the case of the bifunctional mus hydroxide ion. tards. Reaction with thiol groups in vivo seemed probable, and in the case of the bifunctional com pounds ring closure also seemed probable in view @ (_@ N'2 G ---@ I\-N@ CH2IG of the optimum spacing of the reactive centers, \=J @cH2—cii@ \@=/ ‘Qi2@CH2-SG since six-membered rings are normally particularly stable. @.sG Under in vitro conditions with the use of bis CHART9.—Reaction of an aromatic nitrogen mustard with chloroethylaniline, bis-chloroethyl-p-anisidine, or glutathione to form a cross-linked product. mustard gas, and either cysteine or glutathione, evidence was obtained that intermediate sul and dethiolation depicted earlier. Also in this con fonium compounds were in fact formed (70), but nection one should recall the many similarities their instability was such that isolation proved im which exist between the pharmacological effects possible. In the presence of excess cysteine or produced by Mylenan and ionizing radiations. glutathione the only products isolated were the The change in amino acid sequence in a protein cross-linked cysteinyl or cross-linked bound cys or peptide chain constitutes a type of chemical teinyl compounds, in contrast to the situation mutation induced without the intermediate inter found with Myleran. Thus, in this case dethiola vention of DNA or RNA. If certain proteins, tion did not occur. Assuming the formation of an either in the nucleus or the cytoplasm, act as sup intermediate sulfonium ion, the reaction sequence pressons or cell-regulators, then one could envisage followed in the case of glutathione and aniline such a change having possibly far reaching effects, mustard would be as shown in Chart 9. The sec and here may be recalled the drastic effects of the ondary reaction of the sulfonium ion with gluta change of one amino acid unit in a protein in rela thione was obviously so fast that hydrolytic cleav tion to the production of sickle cell anemia (38). age to 4-phenylthiazan was not possible. The inter The difference between the normal hemoglobin esting feature of these reactions is, of course, the and that in the mutant cell is the replacement of point of attack on the sulfonium compound by the

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WAnwI@—Mechanism of Action of Alkylating Agents 13@27 glutathione molecule, leading to ring opening, to porter of any theory of mode of action whose form cross-linked products instead of dethiolation foundations nest on particular reactions at par and ring release as in the case of Myleran. ticular cellular sites, which excludes other possi On the basis of these observations, cross-linking bilities, but he wishes to draw attention to certain of protein strands in vivo following treatment with oversimplifications in reasoning which have led to reactive bifunctional mustards seems likely (71), premature conclusions regarding -511 reaction, and because of the chemical properties of the where in fact a great deal of work remains to be mustards discussed earlier, restriction as to the done. It may transpire that much of the reasoning cellular sites involved would seem less likely used in the past to explain the mode of action of than with powerful Sn@ reactors such as iodoace alkylating agents has its foundation in ovensimpli tate. Thus, the mustards because of their unre fication, and although theories to explain certain active nature prior to ionization would be more pharmacological events are often useful in leading capable of deeper penetration into cells and, to further experimentation, they can be a drag on hence, reaction with a greater range of thiols than rapid progress if their exponents and others regard the Sn@reactors of the iodoacetate type. The argu them as an end in themselves. ments used in the past to dismiss the importance Reaction of aikylating agents with nucleic acids.— of thiol reaction in determining any of the im Since the importance of the nucleic acids in cellu portant pharmacological properties of the alkylat lar metabolism was established, many workers ing agents are for these reasons not valid (74, 8@). have been attracted to the idea that all the im The range of so-called specific -SH reactors tested pontant biological properties of the alkylating for their capacity to inhibit tumor growth, etc., has agents can be explained on the basis of reaction been limited; the compounds have often been with nucleic acids, particularly DNA. Since the monofunctional, when it is known that bifunc alkylating agents cause gross mitotic abnormali tionality is a requisite for high antitumor activity; ties and can effect gene mutations, this hypothesis and most of the bifunctional agents tested have has always had many followers. Nobody would not had their reactive centers spaced an optimum doubt that the alkylating agents lead to an altena distance apart (5@). In addition, account was tion in the structure and function of nuclear DNA obviously not taken of the probability that, be and consequently in cellular. integrity, but there is cause of the inherent chemical reactivity of such evidence that some of these alterations might be compounds, they would probably never reach cel mediated by indirect effects at many sites, such as lular sites accessible to agents such as the mus those indicated earlier when thiol reaction was tards, epoxides, ethyleneimines, and, to a lesser being considered. extent, the aliphatic methanesulfonates. Now that The effects of the alkylating agents on DNA, it is known that both the mustards and the au RNA, and protein synthesis have been recently phatic methanesulfonates are capable of extensive well reviewed by Wheeler (87). It transpires that thiol reaction with -SH groups in vivo, coupled the mustards are capable of inhibiting DNA syn with the knowledge that the mustards can cross thesis to a greater extent than RNA synthesis, link, while Myleran can bring about a unique type whereas the methanesulfonates such as Myleran of dethiolation and at cellular sites probably in have little effect on either—results which may help accessible to -SH reactors of the iodoacetate type, to clarify the apparent differences in phanmaco it will be necessary to reconsider the possible nelo logical properties of the two classes of compound. vance of particular types of -SH reaction to the Considering the popularity of the concept of di mode of action of the alkylating agents. Greater rect action on DNA it seems incredible that the knowledge of the site and extent of reaction is no actual loci of the reactions should be overlooked quired and, in particular, whether thiol reaction in for so long. Despite the work of Young and Camp the nucleus plays any role in inhibition of cell bell (90), who first drew attention to the reactivity division or in those processes leading to ma of the guanine moiety, and of Wheeler, Morrow, lignancy. If in the future alkylation of proteins or and Skipper (88, 89), who pointed out that peptides and, in particular, alkylation of those quaternization of a ring nitrogen atom was likely bearing reactive thiol groups is shown to have no in the reaction of certain punines with alkylating part to play in determining any of the important agents, attention continued to be directed toward pharmacological propertiesof the alkylatingagents, reaction with ionized phosphate groups as the most the experiments leading to such conclusions will probable points of reaction in the nucleotides (@, be mainly those which now remain to be done, 74). Although estenification of phosphate groups since those which have been done can lead us to seems probable in some cases it is a difficult neac no definite conclusion. The writer is not a sup tion to demonstrate because of the lability of the

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 13@8 Cancer Research Vol. 23, September 1963

products formed and because differences in the at pH 7, and in this case breakdown probably oc liberation of acid, interpreted as being due to curs with the formation of an aldehydopynimi esterification, also occur following quaternization dine derivative by ring opening (Chart 11). of the ring nitrogen atoms. It remained for Lawley In the case of reaction with DNA, the apuninic and Wallick (50), possibly stimulated by the re acid formed after removal of the alkylated guanine marks of Skipper (80), to demonstrate that di will be relatively unstable and could lead to main methylsulfate reacted with guanylic acid to form chain breakdown. 7-methylguanine, and it was further shown by These reactions may be of fundamental im Lawley (49) that the product formed from deoxy portance in explaining some of the phanmacologi guanylic acid and dimethylsulfate was much more cal effects of the alkylating in mechanistic chemi labile and that liberation of 7-methylguanine oc cal terms. Thus it was postulated (55) that the curred appreciably at pH 7.@ (37° C.). Thus, a greater efficiency of certain bifunctional alkylating mechanism was established by which alkylation agents in killing cells and interfering with mitosis of guanine moieties of DNA could effect changes might be due to their capacity to cross-link parts of DNA, or DNA to protein, or protein to protein. 0H(cH3),S04 We have seen how proteins might be cross-linked oH?@' (71), some evidence is available for linking of @IcH nucleic acid to protein (77), and Alexander, Cou HZN'@N'@@I@ H@N@% sens, and Stacey (4) obtained evidence of cross R linking of DNA in nucleoprotein with nitrogen @@ R- raos. à @p@sta mustards. This they interpreted as being solely due to reaction with phosphate group, but in the CHART 10.—R.eaction of deoxyguanylic acid or guanylic light of the findings of Brookes and Lawley (@1) acid with dimethyl sulfate. that bis-guanine products are formed after reac tion of various alkylating agents with nucleic acids, they have had to reconsider their interpreta 0HcH3 tion, and it is the belief of Lett, Parkins, and Alex ander (51) that in some cases the alkyl groups of phosphate esters can transalkylate to the 7-posi H,N tion of guanine accounting for the apparently R higher reactivity of guanine when in DNA than when it appears in nucleosides (75, p. 79). The main evidence which they present for this comes from spectral data, the U.V. changes which ac company base alkylation being delayed in some cases as if two reactions were occurring; and they

CHART 11.—Breakdown of an alkylated guanylic acid to also interpret changes in “coiling―asbeing due to form an aldehydropyrimidine derivative. estenification. Whatever the true interpretation, the fact remains that parts of DNA can be cross due to quaternization, elimination of the alkyl linked in vivo, and DNA will be more or less in ated guanine moieties at physiological pH, and activated and degraded depending on the extent of possibly fission of the polymer chain. reaction. Some of the cytotoxic effects of, say, the The reaction can be depicted as in Chart 10, mustards will be explicable on this basis, but when R = nibophosphate or deoxynibophosphate, whether the whole complex process of lysis of fully and the product will have another form contnibut differentiated lymphocytes, for example, can be ing to the resonance hybrid in which the positive explained entirely by such mechanisms remains to charge resides on the alkylated nitrogen atom. be tested. Where R = deoxynibophosphate, breakdown can Brookes and Lawley (@1, @)have studied the probably occur in two ways, leading in the one case to the 7-methylguanine and a carbonium ion extent of combination of various alkylating agents R@which will react with a molecule of water leading with DNA and have found a simple rule to to the liberation of acid and an apuninic acid, or apply—the more reactive the compound and the breakdown could occur at the other nitrogen atom greaten its tendency to react with all nucleophiles, leading to a regeneration of the original deoxygua the greater the extent of reaction with the guanine nylic acid and liberation of CH@. Alkylated (e.g., moiety. Thus the mustards reacted extensively, methylated) guanylic acids are much more stable whereas the methanesulfonates, halides, and sul

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARWICK—Mechanism of Action of Al/cylating Agents 13@9

fates reacted much less extensively and more slow delicately balanced cellular homeostasis. There is ly. These findings again reflect the differences in much evidence to support an alteration in the in pharmacological properties displayed by the van tegrity of mitochondria and the endoplasmic ous compounds. Thus the mustards, highly reac reticulum during administration of carcinogens (8, tive toward most nucleophiles, able to penetrate 9), a consequence of the alteration of lipo-protein deep into cells, able to cross-link nucleic acids and membranes which constitute the structural sup probably proteins, are able to create havoc among port of the organized multi-enzyme complexes. cells, causing effects from inhibition of cell division Thus 3'-DAB (dimethylaminoazobenzene) has a to complete cellular disruption and lysis in con profound effect on microsomal swelling when trast to the less reactive reagents such as Myleran fed for 4 weeks (6), in contrast to the inactive which would be expected to react very little with @-methyl DAB, and the former compound also DNA at therapeutic doses and unable to cross-link produces a significant decrease in the mitochon proteins through their thiol groups (see earlier). drial population in the cell (64, 78). During the The greater efficiency of mustards in causing cell early stages of carcinogenesis with any agent the death probably results from their heightened cell population is being continuously modified, and capacity to react with all nucleophiles within the many cells die as a result of excessive interference cell, not merely to their greater reactivity toward with their cytoplasmic and nuclear elements. DNA. Those extensively modified cells which survive will The low level of reaction of Myleran with be those which have been able to adapt themselves nucleic acids in vivo leads one to seek for other to the unfavorable conditions; but with many of explanations for its pharmacological effects. There their complex feedback controls and fine structure is no scope in this discussion for considering other damaged they will be unable to respond to the implications of DNA reaction such as mutagenesis normal stimuli of intercellular dependence and will and the difficulties which have been found in inter cease to obey the stringent rules of organismal preting some of the biological phenomena en organization (81). The great complexity of the countered—for example, the “uniquenessof ethyl intracellular effects produced by the alkylating ation― (@3, 51, 53) ; but let it suffice to say that, carcinogens, ranging from slight alterations in cell although difficulties have been encountered in structure to cell death, makes the problem of sort interpreting mutation in the light of phosphate ing out the chain of events leading to esterification (f), a working hypotheth has been one of the utmost difficulty. The relatively long put forward by Lawley and Brookes (%4) which time lapse between treatment with the carcinogen seems to fit most of the facts and which can prob.. and the emergence of malignant tumors is prob ably be tested. ably related to the multitude of changes occurring in the cell population during the ‘adaptive' MUTATION AND CANCER period. The oncogenic viruses are probably more Although bifunctionality in alkylating agents is specific in their effect and faster acting because of a normal requisite for high killing action against their information content and may eventually cells, this is not the case with regard to muta prove to be better tools for elucidating the chain genesis and carcinogenesis, since in these respects of events leading to malignancy than the chemical monofunctional agents are often as active or more carcinogens which in many ways have not lived active than their bifunctional counterparts, and up to earlier expectations. this is in line with the concept that cross-linking is lethal. Some evidence exists that the physiological CONCLUSION state of the cell influences the occurrence of No attempt has been made to discuss the actual ‘spontaneous' or induced chromosomal abnormal biochemical lesions such as effects on cellular syn ities (15—17) in some cases; Loveless (54) has thetic mechanisms which result from the action of strongly supported the hypothesis that mutation the alkylating agents. Effects on glycolysis, en in the case of the alkylating agents involves direct zymes, and so on have been well reviewed by attack on nucleic acid. The fact that monofunc Wheeler (87). Rather, an attempt has been made tional alkylating agents can be both mutagenic to correlate the basic alkylating capacities of the and carcinogenic coild be used as evidence to various compounds with gross biological end support the mutational theory of cancer. On the products, and some attempt has been made to other hand, one should not ignore appealing al classify the various agents into groups or classes ternatives which involve much more extensive as far as possible. alteration of cellular integrity and alteration in Some conclusions can be reached regarding

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 1330 Cancer Research Vol. @23,September 1963

these comparisons, the main one being that the facile explanations for their mode of action. We most reactive compounds of the Snl or pseudo now have some idea of the chemical sites with Sni type with a satisfactory “spread―intheir re which a few of them interact intracellularly, but activity toward all nucleophiles are those which from a very mechanical organic chemical view manifest the most diverse pharmacological effects. point. Much more work now needs to be directed It must also be relatively obvious from the fore to the study of their effects in relation to cell popu going discussion that none of the studies carried lation dynamics and their effects in relation to out on the mechanism of action of alkylating intercellular dependence and to dedifferentiation, agents has given any information as to how better since it is in these dynamic aspects that the drugs with a greater selectivity of action might explanation of the carcinogenic process lies. be designed, although suggestions have been made Wheeler (87) indicates at the conclusion to his as to why certain drugs are more efficacious than review that there is a need “topinpoint the critical others. This statement is depressing, and yet the site of alkylation and the real cause of anticancer obvious must be faced—and this is that the activity.― Assuming that there is a “criticalsite of alkylating agents so far designed are fan too crude alkylation,― which is debatable, one wonders in their killing properties to be of any permanent whether knowing it would help in designing better use in cancer therapy. Of course one must always chemotherapeutic agents. We now have drugs leave room for the unexpected breakthrough which which will attack almost every cellular site, and may take the form of a drug with a special “car it is difficult to envisage a new type of alkylating rier― group (14) or a drug possessing “latent― agent which has not yet been tried. activity, and yet attempts which have been made More research on the effects of combinations of so far to make this kind of breakthrough have drugs on tumors might be worth while—there proved somewhat disappointing. Thus, for ex would seem to be endless possibilities to exploit ample, Ross and Warwick (76, 86) made a series and the recent work of Dean and Alexander (f27) of mustard derivatives of azobenzene which would demonstrating the potentiating effect of iodoacet remain chemically unreactive until reduction of amide on the destructive effect of radiation on the azo linkage occurred in vivo, and it was hoped cultured lymphoma cells and bacteria is encourag that some selectivity of action might be achieved. ing. In this connection, Roberts and Warwick' ob However, the compounds are relatively toxic, and served the effect of pretreatment of rats with a one (40) has undergone clinical trial with disap large dose of ethyl methanesulfonate prior to injec pointing results. Yet more of this type of research tion with Myleran on the peripheral blood count. is needed. It was hoped that ethyl methanesulfonate, while Compounds possessing all the various types of inactive itself against granulocytes, might potenti chemical reactivity have been tried, and no ate the effects of Mylenan, since it is itself a thiol obvious solution to the problem of what to try next reactor. Preliminary results have proved disap has been forthcoming. Thus, compounds such as pointing, but in some respects rats are not ideal Myleran, Chiorambucil, etc., which are now used test subjects from the point of view of following as palliatives in the treatment of chronic leukemia minor variations in peripheral blood count due to and Hodgkin's disease, have certainly played a large background fluctuations, and the experiment part in relieving distressing symptoms and have will probably be repeated in some other way. enabled certain patients to remain active when Bearing in mind the chemical, as well as the bio otherwise they may have been bedridden; and in logical, properties of the various agents, some this sense the efforts which have gone into interesting combinations which have not yet been planning them and synthesizing them has prob tried might come to light. ably been justifiable. However, whether many At this Institute further attempts are being more of these compounds should be made is a made to exploit differences between normal and tumor tissue in several ways which have been do debatable point. scnibed at length by Ross (7@2),and the attempts The alkylating agents have stimulated much re being made by Connors and Elson (@25)to lower search on fundamental problems concerning cell general toxicity while maintaining antitumor ac metabolism, mutagenesis, cancinogenesis, and so tivity by prior treatment with certain compounds on, and it is probably in these fundamental ways such as thiols, etc., is worthy of comment. Some of that they have proved most useful. However, these studies assume different pH levels between their very number, the diversity of pharmacologi tumor tissue and normal tissue, and results are cal effects which they manifest, and the number of cellular sites with which they interact preclude 2 Unpublished data.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. WARWICK—Mechanism of Action of Alkylating Agents 1331 awaited with interest. Studies on the mechanisms some Alteration and Tumour Inhibition by Nitrogen of resistance are urgently required, but this sub Mustards: The Hypothesis of Cross-Linking Alkylation. Nature, 166:1112—13,1950. ject is to be covered by Dr. Wheeler. 19. BocK, M., and JACKSON, H. The Action of TEM on the Some fresh approaches are needed to lift cancer Fertility of Male Rats. Brit. J. Pharmacol., 12: 1—7,1957. from the somewhat unsatisfactory 20. BRANCH, G. E. K., and C@vzx, M. The Theory of Organic situation in which it finds itself, but most of all is Chemistry, p. 439. New York: Prentice-Hall, Inc., 1947. needed a really critical assessment from all points 21. BROOKES,P., and LAWLET, P. D. The Reaction of Mono and Di-Functional Alkylating Agents with Nucleic Acids. of view of the situation as it stands at the moment. Biochem. J., 80:496—503, 1961. 22. . The Alkylationof GuanosineandGuanylicAcid. REFERENCES J. Chem. Soc., pp. 3923—28,1961. 1. ALEXANDER,P. The Reaction of Carcinogens with Macro 23. . Ann. Rep. Brit. Empire Cancer Campaign, pp. molecules. Adv. Cancer Res., 2: 1—72,1954. 8—13,1960. 2. . Comparisonofthe ChangesProducedbyIonizing 24. . Ionization of DNA Bases or Base Analogues as a Radiations and by the Alkylating Agents: Evidence for a Possible Explanation of Mutagenesis, with Special Refer Similar Mechanism at the Molecular Level. Ann. N.Y. ence to 5-Bromodeoxyuridine. J. Mol. BioL, 4:216-19, Acad. Sci., 68:1225—1337, 1958. 1962. 3. . Comparison of the Action of Ionizing Radiations 25. CONNORS,T. A., and ETaON, L. A. Reduction of the and Biological Alkylating Agents (Nitrogen Mustards) at Toxicity of Alkylating Agents in Rats by Thiol Pretreat the Cellular and Sub-Cellular Level. Sci. Rep. Inst. Super. ment. Biochem. Pharmacol., 11:722—32, 1962. Sanita, 1:537—50,1961. 26. CRAIG, A. W.; Fox, B. W.; and JACKSON, H. Sensitivity of 4. ALEXANDER,P. ; COUSENS,S. F. ; and STACEY,K. A. Drug the Spermatogenic Process in the Rat to Radiomimetic Resistance in Microorganisms. Ciba Foundation Sym Drugs and X-Rays. Nature, 181:353—54, 1958. posium, p. 294. London: Churchill, 1957. 27. DEAN, C. J., and ALEXANDER, P. Sensitization of Radio 5. ANSON, M. L., and STANLEY, W. M. Some Effects of Resistant Bacteria to X-Rays by Iodoacetamide. Nature, Iodine and Other Reagents on the Structure and Activity 196:1324—26, 1963. of Tobacco Mosaic Virus. J. Gen. Physiol., 24:679-90, 28. DUSTIN, P. Some New Aspects of Mitotic Poisoning. 1941. Nature,159:794-97,1947. 6. ARCOS, J. C., and ARCOS,M. Fine Structural Alterations in 29. ELSON, L. A. Comparison of the Physiological Response to Cell Particles during Chemical Carcinogenesis. 1. The In Radiation and Radioniimetic Chemicals. Patterns of fluence of Feeding of Aminoazo Dyes on the Swelling and Blood Response. In: G. C. DE HEVESY (ed.), Adv. in Solubilization of Rat-Liver Microsomes. Biochim. Bio Radiobiology, pp. 372—75. Edinburgh: Oliver & Boyd, phys. Acta, 28:9—20, 1958. 1957. 7. . Molecular Geometry and Mechanism of Action of 30. . Hematological Effects of the Alkylating Agents. Chemical Carcinogens. Fortsch. Arzneimittelforschung, Ann. N.Y. Acad. Sci., 68:826—33, 1958. 4:407—581,1962. 31. . The Effects of Myleran and Homologous Com 8. ARCOS, J. C. ; GRIFFITH, G. W. ; and CUNNINGHAM, R. W. pounds on the Blood. Biochem. Pharmacol., 1:39—47, 1958. Studies on the Swelling of Rat Liver Mitochondria in 32. ELSON, L. A.; GALTON,D. A. G.; and TIu@, M. The Action Relation to Tumour Incidence during Feeding of Aminoazo of Chlorambucil and Busuiphan (Myleran) on the Haemo Dyes. Experientia, 15:316—17, 1959. poietic Organs of the Rat. Brit. J. Haematol., 4:355—74, 9. . Fine Structural Alterations in Cell Particles during 1958. Chemical Carcinogenesis. J. Biophys. Biochem. Cytol., 33. FARMY,0. G., and FAHMY,M. J. Cytogenetic Analysis of 7:49-60, 1960. the Action of Carcinogens and Tumour Inhibitors in 10. BARRON, E. S. G. Thiol Groups of Biological Importance. Drosophila Melanogaster. IX. The Cell-Stage Response of Adv. Enzymol., 11:201—66, 1951. the Male Germ Line to the Mesyloxy Esters. J. Genetics, 11. BATEMAN, A. J., and JACKSON,H. Mutagenic Action of 46:361—72,1961. TEM. Ann. Rep. Brit. Emp. Cancer Campaign, p. 473, 34. GOLDECK,H., and HAGENAR, H. Der Stickstoff host Em 1958. fuss auf die Fertilitat und Spermiogenese der Labora 12. BERENBAUM, M. C. The Effect of Cytotoxic Agents on the toriumsratte. Z. gus. exp. Med., 117:67—80,1951. Production of Antibody to T.A.B. Vaccine in the Mouse. 35. HADDOW, A. The Chemical and Genetic Mechanisms of Biochem. Pharmacol., 11:29—44, 1962. Carcinogenesis. II. Biologic Ailcylating Agents. In: F. 13. BERGEL,F. and STOCK,J. A. Cytotoxic a-Amino Acids and HOMBURGER (ed.), The Physiopathology of Cancer, pp. Peptides. Ann. Rep. Brit. Emp. Cancer Campaign, p. 6, 602—85,2d ed. New York: Hoeber, 1959. 1953. 36. HEIDELBERGER,C. ; and DAVENPORT,G. R. Local Func 14. BERGEL, F.; STocic, J. A.; and WADE, R. Peptides and tional Components of Carcinogenesis. Acta Internat. Unio Macromolecules as Carriers of Cytotoxic Groups. In: contra Cancrum, 17:55—62, 1961. Biological Approaches to Cancer Chemotherapy, pp. 125— 37. HUGHES, E. D.; INGOLD, C. K.; and SHAPIRO, U. G. 34, London: Academic Press, 1961. Mechanism of Substitution at a Saturated Carbon Atom. 15. BIESELE, J. J. Induced Imitations of Mitotic Anomalies. Part VI. Hydrolysis of Iso-Propyl Bromide. J. Chem. Soc., Ann. N.Y. Acad. Sci., 71: 1054—67,1958. pp. 225—36,1936. 16. . Studies on Mitosis in Purine Treated Tissue Cul 38. Hu@r, J. A., and INGRAM, V. M. The Chemical Effects of tures. In: S. C. PALAY (ed.), Frontiers in Cytology, pp. Gene Mutations in Some Abnormal Human Haemoglobins. 84—112. New Haven: Yale Univ. Press, 1958. Symposium on Protein Structure, pp. 148—51. London: 17. . Mitotic Poisons and the Cancer Problem. New Methuen & Co., Ltd., 1958. York: ELsevier, 1958. 39. INGOLD, C. K. Structure and Mechanism of Organic 18. BIESELE, J. J.; PHILIPS, F. S.; TrnEasdu, J. B.; BmtcHE Chemistry, p. 335. Ithaca: Cornell Univ. Press; and Lon NAL, J. H. ; BUCKLEY, S. M.; and STOCK, C. C. Chromo don: Messrs. G. Bell & Sons, Ltd., 1953.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. 133@ Cancer Research Vol. @23,September 1963

40. ISRAEL8, L. G. ; RITZMANN, S. Laboratory and Clinical 62. OAKBERG, E. F. Sensitivity and Time of Degeneration of Studies on a New Azo-Mustard, 4-Di-2―-chloroethyl Spermatogenic Cells Irradiated in Various Stages of amino-2-methyl-2'-carboxyazobenzene(CB1414). Acta Un. Maturation in the Mouse. Rad. Res., 2:369—91, 1955. mt.Cancer.16:665—71,1960. 63. OGSTON, A. G. The Replacement Reactions of $-@‘- 41. JACKSON, H. Side Effects of Some Tumour Inhibiting Dichiorodiethyl Suiphide and of Some Analogues in Aque Drugs. J. Soc. Chem. md., p. 949, 1957. ous Solution: The Isolation of $-chloro-$'-hydroxy 42. . The Effects of Radiomimetic Chemicals on the Diethyl Sulphide. Trans. Faraday Soc., 44:45—50, 1948. Fertility of Male Rats. J. Fac. Radiol., 9:217—20, 1958. 64. PRICE, J. M.; Mn@Erc, E. C.; MILLER, J. A.; and WEBER, 43. . AntifertilitySubstances.Pharm.Rev., 11:135—72, G. M. Studies on the Intracellular Composition of Livers 1959. from Rats Fed Various Aminoazo Dyes. II. 3'-Methyl, 44. JACKSON,H.; Ciwo, A.W.;and Fox,B. W.Comparative 2'-Methyl and 2-Methyl-4-dimethylaminoazobenzene; 3- Effects of Certain Alkylating Agents on the Proliferation of Methyl-4-monomethylaminoazobenzene and 4'-Fluoro-4- Normal and Tumour Tissues. 1. Spermatogenesis and dimethylaminoazobenzene. Cancer Res., 10: 18—27,1950. Tumour Growth. Acta Unio. Internat. contra Cancrum, 65. ROBERTS, J. J., and WARWICK, G. P. Studies of the Mode 16:673—75,1960. of Action of Tumour Growth Inhibiting Alkylating Agents. 45. Kou@zm, P. C. Comparative Effects of Alkylating Agents I. The Fate of Ethyl Methanesulphonate (“HalfMy on Cellular Morphology. Ann. N.Y. Acad. Sci., 68:783— leran―) in the Rat. Biochem. Pharmacol., 1:60—75, 1958. 801,1958. 66. . Studies of the Mode of Action of Tumor Growth 46. . Comparison of the Biological Effects of X-Rays Inhibiting Alkylating Agents. III. The formation of Tetra and Radiomimetic Chemical Agents. Proc. Third Aus hydrothiophene-1,1-dioxide from 1,4-Dimethanesulpho tralasian Conference on Radiobiology. Radiobiol, pp. 218— nyloxybutane (Myleran), S-@9-L-Alanyltetrahydrothiophe 86, 1960. niummesylate, Tetrahydrothiophene, and Tetrahydrothio 47. LAJTHA, L. G. Bone Marrow Cell Metabolism. Physiol. phene-1,1-dioxide, in the Rat, Rabbit and Mouse. Rev., 37:50-05, 1957. Biochem. Pharmacol., 6:217—27, 1961. 48. LANDING, B. H.; Goum@, A.; and Nos, H. A. Testicular 67. . The Reaction of Some Compounds of Biological Lesions in Mice Following Parenteral Administration of Interest with Nucleic Acids. Biochem. J., 84: 119 r@,1962. Nitrogen Mustards. Cancer, 2: 1075—82,1949. 68. . Azo-Dye Liver Carcinogenesis: Reaction of Hy 49. LAwLET, P. D. The Hydrolysis of Methylated Deoxy droxymethylaminoazobenzene with Nucleic acids in Vitro. guanylic Acid at pH 7 To Yield 7-Methylguanine. Proc. Nature, 197:87—88, 1963. Chem. Soc., pp. 290—91,1957. 69. . The Mode of Action of Alkylating Agents. II. 50. LAWLEY, P. D., and WAu@icu, C. A. The Action of Studies of the Metabolism of Myleran. The Reaction of Alkylating Agents on Deoxyribonucleic Acid and Guanylic Myleran with Some Naturally Occurring Thiols in Vitro. Acid. Chem. & lad., p. 633, 1957. Biochem. Pharmacol., 6:205—16, 1961. 51. Ls'rr, J. T.; PARKINS,G. M.; and ALEXANDER,P. Physico 70. . Studies of the Mode of Action of Tumour Growth chemical Changes Produced in DNA after Alkylation. Inhibiting Alkylating Agents. IV. In Vitro Reactions of Arch. Biochem. Biophys., 97:80—93, 1962. 2-Chloroethylarylamines. Biochem. Pharmacol. (in press). 52. LoVELESS,A.QualitativeAspectsof the Chemistryand 71. . Studies of the Mode of Action of Tumour Growth Biology of Radiomimetic (Mutagenic) Substances. Nature, Inhibiting Alkylating Agents. VI. The Metabolism of Bis 167:338—42,1951. @9-Chloroethyl Sulphide (Mustard Gas) and Related Corn 53. . The Influence of Radiomimetic Substances on De pounds. Biochem. Pharmacol. (in press). oxyribonucleic Acid Synthesis Studied in E8cherichia coli 72. Ross, W. C. J. Biological Alkylating Agents. London: Phage Systems. III. Mutation of T2 Bacteriophage as a Butterworths, 1962. Consequence of Alkylation in Vitro: The Uniqueness of 73. . The Chemistry of Cytotoxic Alkylating Agents. Ethylation. Proc. Roy. Soc., s.B., 150:497—508, 1959. Adv. Cancer Res., 1:397—449, 1953. 54. . The Involvement of Deoxyribonucleic Acids in 74. . in Vitro Reactions of Biological Alkylating Agents. Chemical Mutagenesis. Biochem. Pharmacol., 4:29—33, Ann. N.Y. Acad. Sci., 68:669—81, 1958. 1960. 75. . Biological Alkylating Agents, p. 25. London: 55. LOVELESS,A.;GOLDACRE,R.J.; andRoss, W. C. J. Mode Butterworths, 1962. of Production of Chromosome Abnormalities by the Nitro 76. Ross, W. C. J., and WARWICK, G. P. AryI-2-halogeno gen Mustards: the Possible Role of Cross-Linking. Nature, alkylamines. Part XVIII. The Rates of Reduction of Sub 163:667—69,1949. stituted 4-Di-(2-chloroethyl)arninoazobenzenes by Stan 56. LovELEss, A., and REVEL, S. New Evidence on the Mode nous Chloride, , and the Xanthine Oxidase of Action of Mitotic Poisons. Nature, 184:938—44, 1949. Xanthine System. J. Chern. Soc., pp. 1724—32,1956. 57. LovELEss, A., and Ross, W. C. J. Note Following Biesele, 77. RUTMAN, R. J. ; STEELE, W. J. ; and PRICE, C. C. Observa J. J. St at. Chromosome Alteration and Tumour Inhibition tions on the in Vise Alkylation of Ehrlich Tumour Cell by Nitrogen Mustards: the Hypothesis of Cross-Linking DNA by Nitrogen Mustards. Biochem. Biophys. Res. Alkylation. Nature, 168: 1113—14,1950. Comm., 4:278—82, 1961. 58. M@zi.&, D. Biochemistry of the Dividing CelL Ann. Rev. 78. SCHNEIDER, W. C. Intracellular Distribution of Succiic Biochem., 30:669—88, 1961. Dehydrogenase, Cytochrome Oxidase, Adenosinetriphos 59. Mxcn.&zus, L., and SCHUBERT, M. P. The Reaction of phatase and Phosphorus Compounds in Normal Rat Liver lodoacetic Acid on Mercaptans and Amines. J. BioL and in Rat Hepatornas. Cancer Res., 6:685, 1946. Chem., 106:331—41, 1934. 79. SHAVER, S. L. X-Irradiation Injury and Repair in the 60. Mni@ica,E.C.,and Mu@su,J. A.BiochemicalInvestiga Germinal Epitheium of Male Rats. I. Injury and Repair tions on Hepatic Carcinogenesis. J. Nail. Cancer Inst., 15: in Adult Rats. Am. J. Anat., 92:391-432, 1953. 1571—90,1955. 80. SKIPPER, H. E. Discussion of the Biological Effects of 61. Nou@m@,C. R., and DnisMous, R. The Reaction of Alkyl Ailcylating Agents. Ann. N.Y. Acad. Sci., 68:808, 1958. Bromides with Pyridine. J. Am. Chem. Soc., 54:1025-84, 81. SMITHERS, D. W. Cancer: An Attack on Cytologism. 1932. Lancet, pp. 493—99,1962.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. W@wic@x—Mechanism of Action of Alkylating Agents 1333

82. STACEY, K. A.; COBB, M.; Cousm@s, S. F.; and ALzw@ 87. WHEmms, G. P. Studies Related to the Mechanisms of DEE, P. The Reactions of the “Radiomimetic― Alkylating Action of Cytotoxic Alkylating Agent@ Cancer Res., 22: Agents with Macromolecules in Vitro. Ann. N.Y. Acad. 651—88,1962. Sci.,68:682—701,1958. 88. WHEELER, G. P.; Moimow, J. S.; and S@xppm@,H. E. 83. ST@urn,G. R.; STm1@@,W.H.; and MooRz, S. Relations be Studies with Mustards. L Ion Exchange Chromatography tween the Conformation of Ribonuclease and Its Reactiv of Reaction Mixtures of Purines and Pyrimidines with ity toward Iodoacetate. J. Biol. Chem., 236:436—42, 1961. Nitrogen Mustard. Arch. Biochem. Biophys., 57:124—32, 84. SWAIN,C.C. Abstr.13thNail. Org.Chem.Symposium, 1955. Am. Chem. Soc., p. 65, 1953. 89. . Studies with Mustards. II. Paper Chromatog 85. Tmssxs, G. M. The Action of and Bio raphy and Radioautography of Reaction Mixtures of logical A&ylating Agents on the Synthesis of Deoxyribo Various Compounds with S's-Sulphur Mustard. ibid., nucleic Acid, and a Possible Relation between the Mech pp. 133—39. anisms of Action. Biochem. Pharmacol., 4:49—56, 1960. 90. YOUNG,E.G., and [email protected]@,ILB. The Reactionof 86. W@tuwicx, G. P. Relative Rates of Reduction of a Series of fl-fl'-Dichlorodiethyl Sulphide with Proteins. Canad. J. Azo Compounds. J. Soc. DyersColourists, 75:291—99, 1959. Res., 25B:37—48, 1947.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research. The Mechanism of Action of Alkylating Agents

G. P. Warwick

Cancer Res 1963;23:1315-1333.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/23/8_Part_1/1315

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/23/8_Part_1/1315. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1963 American Association for Cancer Research.