the mechanism of action of alkylating agents...substancef (competi.. factor)phthiosulfate tion...

G. P. WARWICK

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

SUMMARY

The term â€œalkylating agentâ€• has been defined and a detailed discussion made of themechanisms (Sni and Sn@) by which they interact with nucleophilic centers. Thevarious nucleophiles likely to be encountered in vivo are discussed in terms of theirrelative reactivities toward alkylating agents.

With regard to Sn@ reactors it was possible to group them into four sections on thebasis of their chemical reactivity and the â€œspreadâ€•in their affinity toward differentnucleophiles. The groups comprised (a) the epoxides, ethyleneimines, and $-lactones;(b) the primary alkyl methanesulfonates; (c) primary alkyl halides; and (d) a-haloacids and ketones.

An attempt has been made to correlate the reactivity of the Sn@ and Sni reactorswith their known pharmacological properties and the likelihood of their reaction withthe various cellular constituents. In particular, the effects of the alkylating agents onnuclear material and cell division, the blood components, male fertility, and the immune response have been discussed in detail, and the chemistry involved in the interaction of the alkylating agents with compounds of biological importance such as thenucleic acids, proteins, and peptides has been critically examined.

The importance and limitations of the alkylating agents as chemothenapeutic drugsand as tools for examining the basic mechanisms involved in carcinogenesis and mutation have been considered in the light of recent findings.

No attempt will be made to review the whole ofthe literature relevant to the mode of action ofalkylating agents, since this has been done manytimes in the past, and more recently by Ross (7@)and by Wheeler (87). Because of the growing complexity of the field, it might be instructive to examine the chemical reactivities and alkylating potentials of some of the different classes of alkylating agents in relation to what is known about theirpharmacological properties. The different types ofcompound which can function as alkylating agentsor electrophilic reagents include alkyl halides,alkyl methanesulfonates, alkyl sulfates, alkyl phosphates, halogenomethyl ethers, @2-chloroethyl sulfides, @-ch1oroethylamines, epoxides, @-lactones,ethyleneimines and ethyleneimides, diazoalkanes,activated ethylenic compounds, halogenomethylketones and esters, methylolamines, etc., and, although only a few of these classes are useful aspalliatives in clinical therapy, all are capable of reacting with at least some nucleophilic centers withthe result that even inactive compounds can help

in the elucidation of fundamental mechanisms ofaction.

The term â€œalkylating agentâ€• in its widest sensedenotes those compounds capable of replacing ahydrogen atom in another molecule by an alkylradical, and this of course involves electrophilicattack by the alkylating agent so that the definition must be extended to include those reactionsinvolving addition of the radical to a molecule contaming an atom in a lower valency stateâ€”for example, formation of a sulfonium compound from asulfide. The alkylating agents exhibit a diversity ofpharmacological properties including the capacityto interfere with mitosis, to cause mutations, andto initiate and promote malignant tumors; someof them have a stimulatory action on the nervoussystem, and others are powerful vesicants orlachrymators. When attempting to find a parallelism between the pharmacological effects elicitedby the alkylating agents and their chemical reactions at particular cellular sites, it is important toconsider, among other variables, the particular

1315

The Mechanism of Action of Alkylating Agents

Research. on February 3, 2020. © 1963 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

SubstanceF(Competi..

tionfactor)pHThiosulfate

Cysteine ethyl esterMethylaminePhosphatePyridineChlorideAcetateFormateNitrateGlucose2.7X104

1 .SX1O'3.9X10'

7554211030.208

7138777777

1316 Cancer Research Vol. p23, September 1963

mechanism of alkylation involved, since this willlargely determine the chemical groups which willbe attacked inside the cell and their positionwithin the cell. Almost every review which hasbeen written on the alkylating agents has includedat least some description of the mechanisms involved in unimolecular and bimolecular reactions,and the justification for indulging in this again indetail is inherent in the problem with which we areconcerned. The alkylating agents, unlike manyother drugs used in medicine, produce their effectsby binding covalently with cell constituents, and

R@ ___ â€¢ 9â€”x@ R+XÂ® 9

R+Y@ RY

CHART 1.â€”Mechanism of first-order nucleophilic substitu

tion Sn 1 reaction.

:(@2@ZC1

Ogston (68) in the case of mustard gas. Mustardgas (see later) yields a carbonium ion which is stabilized in the form of a threo-membered ring, andthis accounts for its greaten capacity to discriminate between anions. This stability is even greaterin the case of the immonium ion formed from auphatic @-chloroethylamines due to the greaterbasicity of nitrogen than sulfur (Chart @).

Each nucleophile (canbonium ion, immoniumion, etc.) has a competition factor which was measured by Ogston in the case of mustard gas byobserving the extent of reaction in aqueous solution with various anions (Table 1). In the case ofcompounds such as thiols, amines, and acids, account must be taken of the amount in the reactiveform at a particular pH in considering likelihoodof reaction with an alkylating agent. If one nomembers that alkylating agents pick out centersof high electron density, thiols and acids will bemost reactive when ionized, in contrast to amineswhich become unreactive when protonated. If thecompetition factor for a particular group is F, c isits concentration, and is the amount in the reac

tive form, then in a particular system the relativeamount of any group reacting is proportional to

TABLE 1

COMPETITION FACTORS FORMUSTARD G@s (63)

Fl c. However, Alexander (1) has drawn attentionto the fact that apparent competition factors of

groups in macromolecules may be increased byadsorption of the agent onto the macromolecule,thereby increasing effective concentration. Ross(74) assessed the reactivity of a number of chloroethylarylamines by measuring their rates of hydrolysis in aqueous acetone and found that electronreleasing substituents in the benzyne ring such asp-methyl greatly increased the rate of hydrolysis,whereas it was retarded by p-nitro-, p-phenylazo-,or other electron-attracting substituents.

The other essential features of Sni reactionsare (a) that the rate of any particular reaction will

R1

V @cH2CHART 2.â€”Ethyleneimmonium ion formed from an au

phatic nitrogen mustard.

their sites and extents of binding, although dependent to some extent on factors such'as solubil

ity, will be determined mainly by their electrophilic (or alkylating) potential; and this involvesconsideration of ease of electron displacement,resonance energies, bond strengths, steric factors,etc.

There are two generally accepted basic mechanisms of alkylation, first-order nucleophilic substitution (Sni) and second-order nucleophilicsubstitution (Sn@), although care should be takenin making too rigid a classification. In the former,the rate-controlling stage in the reaction is ionization to form a solvated carbonium ion, followedby a rapid combination with a solvent molecule ornew nucleophilic reagent such as an anion or sulfide (Chart 1). Since the driving force of the reaction is a displacement of electrons away from R,then reactions of this type will be facilitated by thepresence of electron-repelling substituents (such asmethyl) in R. The carbonium ion, being an extremely unstable and hence reactive species, willbe to some extent indiscriminate in its combinationwith nucleophilic centers. The more unstable carbonium ions, such as those derived from i-propylmethanesulfonate, will tend to react rapidly withsolvating water molecules, whereas in some casesmore discrimination is possible as demonstrated by



Nucleophilek/kwNucleophilek/kwH,O

CH,C00Pyridine}IPOi0H1

50040006300

16000HS

Thiosulfate1.3X10' 4 .0X10

WARwICKâ€”Mechanism of Action of Alkylating Agents 1317

be essentially independent of both the concentration and the nature of the other reagent. Thus, anionized thiol, a powerful nucleophile because of

the electron distribution on the sulfur atom, willreact at the same rate as, e.g., an amine or hydroxide ion, less powerful nucleophiles, and (b)that no reaction will take place until primaryionization has occurred. Both these factors are incontrast to what is observed in Sn@ reactions and

are probably very significant when comparing thepharmacological properties of the two classes ofreactant.

Mustard gas (dichlorodiethyl sulfide) and thearomatic nitrogen mustards are the Sni reactorswith which we will be concerned in this discussion.Mustard gas is an extremely reactive compound(half-life in water about 3 mm. at 37Â°C.) becauseof the presence and position of the electron-releasing sulfur atom relative to the chlorine atomswhich facilitates formation of a carbonium ionthrough the unstable ring system shown. Thecarbonium ion (which may exist in the form of athreo-membered ring) can discriminate betweennucleophiles to some extent, but it is far too reactive and, hence, toxic to be of any use in therapy(Chart 3).

The aromatic nitrogen mustards were developedbecause they retained the capacity to react by an

Snl mechanism; but the electron-withdrawingcapacity of the aromatic ring greatly reduced therate of carbonium ion formation so that cornpounds could be synthesized which could reachdistant sites in the organism before reacting withtissue constituents.

ci@â€”@ci/1

acH,cH,â€”@-cI@ â€”k cicii@cii@scii,cij,

CHART 3.â€”Mechanism of ionization of mustard gas

Y@RL@â€”*Y_R + Xe

CHART 4.â€”Mechanism of Sn2 reaction

In second-order nucleophiic substitution a carbonium ion is not formed, but a transition complex is formed involving both reactants, and hencethe rate of reaction is dependent upon both concentrations (Chart 4). In this case the reagenthelps in the breaking of the R-X bond, and thiswill occur most readily if the energies of the Y-Rand R-X bonds in the transition state are high, ifthe electron affinity of X is high compared withthat of Y, and if the resonance energy of the tran

sition state is high compared with the initial state.Bond strength is obviously important, since,whereas electron affinity increases in the orderI < Br < Cl < F, replacement by an alkylatingagent increases in the reverse order, I being mosteasily expelled.

Compounds such as primary alkyl halides,methanesulfonates, and phosphates are not particularly powerful reactors, their rate of reactionwith water and other nucleophiles being usuallyrelatively low. Thus, whereas mustard gas has ahalf-life of 3 mm. in water at 87Â°C., ethyl methanesulfonate has a half-life of 36.3 hours andthat of methyl bromide is 116 hours (75). The noaction rate increases considerably in the presenceof centers of high electron density, especially

TABLE 2

RELATiVE REACT1VITIES OF NUCLEOPRILESTOWARD METHYL BROMIDE (84)

anions. Thus, the ability of Y to replace X in R-Xby the Sn@mechanism can be assessed by comparing reaction rate with V as compared with water.This has been done by Swain for methyl bromide(84).

The major nucleophiic centers present insidecells are ionized thiol groups, ionized carboxylicacids, ionized phosphates, and unionized amines,and, ignoring accessibility and concentration factors, reaction would proceed most readily in theorder thiol > amine > phosphate > carboxyl.

Compounds such as a-halogenocarboxylic acidsand ketones are highly reactive Sn@ reagentstoward certain nucleophules by virtue of the elec

tron-attracting capacity of the carbonyl group(Chart 5) . However, for reasons to be discussedlater, its reactivity toward weak nucleophiles suchas the carboxyl ion or the phosphate ion is low. Inthis context a significant factor in consideringSn@ reactors is the variation in relative reactivityof a particular reagent toward different nucleophilic centers. For example,

rate constant (thiosulfate)rate constant (acetate)

is @05for ethyl methanesulfonate, 4380 for methylbromide, and 16,600 for the iodoacetate ion (75,



1318 Cancer Research Vol. @Z3,September 1963

p. @).Thus, it is obvious that the â€œspreadâ€•in noactivity toward different nucleophilic centersvaries considerably from reagent to reagent and isof fundamental importance in determining thereactions which a particular drug will undergoin vivo and hence its pharmacological properties.

The factors which determine whether an agentwill react by an Snl or an Sn@ mechanism areobviously complex, and it is dangerous to be toorigid in this type of classification, since even solvent changes can sometimes lead to a modificationin mechanism. There occur also those compoundssuch as the epoxides and ethyleneimines which, byvirtue of their strained three-membered ring struc

ture, fall into a different category from most otherSn@ reactors, as will be considered later. In thecase of alkyl halides and alkyl methanesulfonates,the rate and mechanism of reaction depend on

e â€˜@2x + IcH,â€”@c-oHâ€”@x--cH,cooH+ I

CHART 5.â€”Mechanism of reaction of iodoacetic acid

cHa CH@I SCH,S0,Oâ€¢CCH3.@ coso,cH3

halides leads to a further drop in Sn@ rate, butnow the Snl mechanism becomes of importance,and the two mechanisms operate simultaneouslyand at a comparable rate (37) . The reason forthis is that the extra methyl group facilitates pnimary ionization of the halide ion (Sni) but retardsthe Sn@process for electronic reasons and becauseit hinders the approach of the nucleophile and subsequent formation of the transition complex (pnimany stenic effect) . The presence of three methylgroups as in tert-butyl halides causes a completechangeover to the Sni mechanism, and the rate ofhydrolysis becomes faster than the Sn@ hydrolysisof methyl halides under the same conditions (61).

A change in mechanism can also be observedsometimes if an attacking nucleophile is replacedby one of less nucleophilic potential under thesame experimental conditions. At first the effectwill be to decrease the velocity of the Sn@ process,but with even weaker nucleophilic reagents a pointwill be reached in which the reagent is no longerable to assist in the breaking of the R-X bondwhich must, therefore, occur unaided by thereagent, and the mechanism will become Sniprobably coupled with a decreased reaction rate(39). For example, in the substitution of the tnmethylsulfonium ion by various anions it wasfound that the specific rates were OH(743),PhOâ€”(13.4), CO@(7.4), Br(7.8), Cl(7.3),ously indicating an Sn@ mechanism for the firsttwo anions and an Snl mechanism at a lesser ratefor the others. A situation similar to this probablyoccurs in the case of a-haloketones and acids, mdieating their tremendous potential to react withpowerful nucleophiles such as ionized thiols by an

Sn@ process and their relative unreactivity towardweaken nucleophiles such as ionized acid groups(hence the misnomer â€œspecificthiol reactorsâ€•).The earlier data concerning the alkyl halides isprobably of fundamental importance in considering the biological properties of the various agents.For example, the greater reactivity of methylhalides (or methanesulfonates), as compared withthe corresponding ethyl derivatives, probably explains the greater toxicity of the former in vivo.One surprising finding is the fact that dimethylmyleran (@,5-dimethanesulfonyloxyhexane) is sosimilar to Myleran (1,4-dimethanesulfonyloxybutane) in its pharmacological properties (see later)(Chart 6). Unlike Myleran, it will tend toward anSni mechanism in its reactions, forming a very Unstable carbonium ion extremely reactive towardits own solvating water in aqueous media andhence unreactive toward other nucleophiles. However, the possibility of its reacting by an Sn@mechanism may become more important in non

@H35020(@H2@,0S0PI3

CHART 6.â€”Dimethylmyleran and Myleran

whether the reactant is primary, secondary, or tentiary. In some cases, such as in the reaction ofalkyl bromides with pynidine, the rate of reaction to form an ethylene derivative is in the ordertent. > sec. > prim., whereas in the simultaneousreaction to form quaternary salts the order is noversed. The differences in the reaction rates ofprimary, secondary, and tertiary alkyl compounds

are generally attributed to the inductive effect.The replacement of a hydrogen atom by a methylgroup is acid-weakeningâ€”i.e., it induces a negative charge on the atom on which substitution hasoccurred. It therefore increases the ease of ionization of a halide ion from the molecule. This easeof separation can be more than compensated forby the smaller attraction of the reacting base. Consequently, the primary halides require a strongerbasic reagent than the tert-halides, but this orderis reversed when the reagent acts as an ionizingrather than a basic reagent (61).

In general methyl halides react almost exclusively by an Sn@ process, but introduction of another methyl (electron-releasing) substituent (giving ethyl) leads to a reduction in the rate of reaction by the Sn@ mechanism, the factor usually being about 10, as in the reaction with amines. Introduction of another methyl group as in iso-propyl



WARwICKâ€”Mechanism of Action of Alkylating Agents 1319

aqueous media, and the implications of this will beconsidered later.

Striking differences in pharmacological effectshave been found on the one hand between alkylating agents which react essentially by an Snimechanism, and many of those which react by theSn@ mechanism, between monofunctional and bifunctional agents of both types, and between mdividual members of the Sn@ type. The differencesobserved between the various Sn@ reactors areprobably explicable on the basis of the foregoingdiscussion, and from the point of view of chemicalreactivity and â€œspreadâ€•in affinity toward different nucleophiles the Sn@ reactors can be broadlyclassified into four groups:

a) the epoxides, ethyleneimines, and fl-lactones;b) the primary alkyl methanesulfonates (e.g.,

Myleran and ethyl methanesulfonate);c) primary alkyl halides (e.g., dibromobutane);

and,d) a-haloacids and ketones (e.g., iodoacetic

acid and co-bromoacetophenone).Group (a) .â€”Compnises compounds which bear

a resemblance in structure to the ethyleneimmonium ions which are derived from aliphaticnitrogen mustards and which react by an Sn@mechanism with nucleophiles (see earlier) (Chart7).

Epoxide reactions can be acid-catalyzed, butthis will probably be of little significance in vivo.The presence of a strained three-membered ringin these compounds makes them somewhat unlikethe other Sn@Zreactors in the finer details of theirreaction kinetics. In a normal Sn@ process, inwhich an entering nucleophile causes the expulsionof a weaker nucleophile such as in the reaction ofan ionized thiol with methyl iodide, the process ofbond-making is considered a more important dniving force of the reaction than that of bond-breaking. In the transition state of such a reaction theprocess of bond-making (new -S-C-bond) will already have progressed to a considerable extent. Ifas mentioned earlier the process of bond-breaking(-C-halogen) is the more important driving forceof the reaction, then the reaction kinetics tendtoward unimolecular and the rate of ionization becomes the rate-determining step. There are obviously gradations between the two extreme mechanisms as already mentioned, and the epoxides ofbiological interest fall into a class in which, whilesecond-order kinetics are followed, the process ofbond-breaking has progressed to a considerableextent in the transition complex because of thestrained three-membered ring structure. Sincebond-breaking is a more important driving forcethan bond-making in such reactions, one would

anticipate epoxides and ethyleneimines to reactreadily with a range of nucleophiles in a somewhatsimilar way to the mustards. They have in factbeen shown to react readily with thiols, amines,and acids (73), and this ease of reaction with diverse nucleophiles is reflected in their biologicalproperties, since they produce in many respects

end-products (inhibition of mitosis, mutation,chromosome breaks, cancer, effects on the hemopoietic system) very similar to the mustards (35).

Group (b), the primary methanesulfonates, differfrom the epoxides in their potential of reactivity,because no strained ring is present and the Sn@mechanism will mainly operate in which bondmaking will be more important than bond-breaking, so that both the concentration of an attacking reagent and its nucleophilicity will be of primeimportance, although it should be mentioned that

Â® 0Â®Râ€”CHâ€”CH2+ X + Hâ€”@RCH--CH,X

@1 6HCHART 7.â€”Mechanism of epoxide reactions

the true Sn@ tendency will be greater in themethyl esters than the ethyl. Because of their relative unreactivity compared with epoxides, thereis found a greater â€œspreadâ€•in their affinity towardnucleophiles, and their reactivity toward thiolgroups will be far greater than toward amines onacids. In this connection Brookes and Lawley (@1)have found that both ethyl methanesulfonate andMyleran have much less reactivity toward guaninein DNA than the mustards and epoxides, whereasRoberts and Warwick (65, 66) have demonstratedthat almost all the urinary reactivity excretedafter injection of these compounds could be accounted for by reaction with the powerfullynucleophilic thiol group in vivo. In the case ofethyl methanesulfonate much of the C'4 in thedrug was exhaled as Câ€•02which could have beenderived either by hydrolysis of esters formed byreaction with carboxyl groups on phosphategroups. In the case of Myleran, much less of thedrug was metabolized in this way. The capabilitiesof an alkylating agent to react with various nucleophiles is reflected faithfully in the biological andpharmacological effects which it elicits, leading inthe case of these two agents to effects much lesspotent than were observed with the mustards,epoxides, etc. Thus Myleran is much less toxicthan mustard gas, possesses none of its vesicantproperties, has only moderate activity againsttransplanted tumors, causes far fewer chromosome



Cancer Research Vol. @3,September 196313@20

â€œbreaksâ€•(33), and, although it decreases the number of circulating granulocytes, it has far less effeet on the peripheral lymphocyte count than the

mustards and epoxides (30) ; and its effects on theimmune response and on male fertility in rats

again differs fundamentally from the mustardsand ethyleneimines (see later).

Group (c), comprising the primary alkyl halides,represents, with the exception of the methyl corn

pounds, agents even more unneactive than the pnimary alkyl methanesulfonates. Differences in thenucleophilicity of the Br and CH3SO2O anionscoupled with differences in the C-Br and C-OSO2CH3bond strengths and a weak capacityto undergo Sn@ reactions makes the halides (e.g.,

ethyl bromide and dibromobutane) even poorercandidates than the methanesulfonates as far asdiversity of reactions in vivo are concerned. Thus adecreased reactivity toward all nucleophiles is ohserved, so that even fewer biological effects wouldbe anticipated; and this is so, since 80 mg. of dibromobutane can be administered safely to a300-gm. rat compared with 4 mg. of Myleran, andthis compound produces none of the pharmacological effects observed with Mylenan. The increased reactivity of methyl compounds has been

discussed earlier, and this is reflected in the hightoxicity of methyl bromide compared with ethyl

bromide.Roberts and Warwick showed that dibromo

butane reacted with cysteine in t@itroin a mannersimilar to that of Myleran leading to a sulfoniumion (66). Since dibromobutane lacks the pharmacological properties of Myleran, this evidence couldbe used to discount the importance of such a reaction in vivo and, indeed, the importance of thiolreaction in general. However, such reasoning isprobably not valid, because it does not take account of the different thiol groups in the cell, theirdifferent reactivity, or the difference in alkylatingpotential between Myleran and dibromobutane.The major thiol reaction in vivo will be with glutathione, which will probably have little permanenteffect on cellular function ; reaction with thiol proteins or enzymes might have greater significance,and there is no evidence yet that dibromobutanewill react with these in vivo. The site of thiol reaction inside the cell is probably very important,since it is now almost certain that every alkylatingagent will combine extensively with thiol groupsin vivo, and yet compounds such as the mustards,rnethanesulfonates, halides, and a-haloketonesand acids manifest very different effects on cells.There are two conclusions to be drawn from this:(a) that thiol reaction in no case contributes topharmacological end-effects, or (b) that thiol re

action is a contributor and that the site at whichit occurs and its extent differ with those of thevarious agents leading to different biological effects. There is no proof yet which alternative iscorrect, and yet certain vital experiments, whichremain to be done, may make the situation clearer.The past literature is full of examples implicatingthe importance of thiol reactions in all sorts of biological phenomena (10), and yet a really criticalassessment of some of these claims can leave littledoubt that they were based on naÃ¯ve or oversimplified assumptions leading to a very confusedsituation. The potential importance of thiol neaction remains (58), and yet more research needs tobe done to discard outmoded ideas and to bringthe whole field into line with modern concepts.Speculative â€œpaperchemistryâ€• leads us nowhere,except possibly in circles, and there is no substitute for doing the vital experiments which canfinally clear up these points. More will be said on-SR reaction later.

Group (d) comprises those agents, the a-haloacids and ketones, which in their reaction mechanism lie in the other extreme from undoubted Snireactors. With these potent Sn@ reactors theprocess of bond-making far outweighs in importance the process of bond-breaking, and so the

nature of the attacking nucleophile is vitally important, leading to extreme reactivity towardpowerful nucleophiles such as ionized thiol groups

and a very weak level of reactivity toward lessnucleophilic centers. Thus as mentioned earlier

k(thiosulfate)k(acetate)

is 16,600, as compared with 9205for ethyl methanesulfonate.

Probably many resonance forms contribute tothe transition state (920), but the peculiar reactivity of compounds of this type is a direct result ofthe presence of the powerful electron-attractingcarbonyl group. Because of this, compounds of thistype were once regarded as specific thiol reactors,

and all their effects from enzyme inhibition to general toxicity in vivo were ascribed on the basis of-511 reaction, despite warning by Michaelis andSchubert (59) that amine reaction could also occur. In this connection Anson and Stanley (5) latershowed that, although tobacco mosaic virus wasnot inactivated by p-chionomercunibenzoate (apowerful 511 reactor), it was inactivated by iodo

acetarnide and iodoacetic acid, implicating reaction at another site. More recently, Stark, Stein,and Moore (83) have shown that the inactivationof nibonuclease is due to reaction with the ringnitrogen of histidine. This reaction with tertiary



WARWICKâ€”Mechanism of Action of Alkylating Agents 13f21

amines is of interest, since Roberts and Warwick(67) have shown that bromoacetic acid reactsreadily with the punine portion of deoxyguanylicacid and that a low level of reaction with theguanine of DNA also occurred. It remains to beseen how significant such reactions are in tiivo,however, because of the high competition by othergroups such as thiol in the cytoplasm. In this connection, the same authors have shown (67) that,after incubating Krebs ascites cells with C'4-bromoacetic acid, there is a steady uptake of labeland the ratio

specific activity (protein)specific activity (DNA)

was approximately 2@; and mthas yet to be established whether any DNA reaction other than thatassociated with residual DNA protein has occurred. It is significant that in the case of mustardgas and Myleran the ratio was nearer 1 (@1).

From the foregoing discussion it is probably trueto say that, with some exceptions, the differenttypes of alkylating agent are potentially capableof reacting with all nucleophilic centers and thatthe pharmacological properties of a particularagent will depend on (a) ease of penetration, (b) tosome extent mechanism of alkylation, and (c)â€œspreadâ€•in capacity to react with different nucleophiles. The mustards, for example, probably produce such diverse effects because they can easilypenetrate cells, their chance of reaching mostparts of the cell is high because of their unreactive

nature until they are ionized (a relatively slowprocess), and the carbonium ions once formed arereactive toward most nucleophiles, leading tosome reaction at most parts of the cell. The vanous Sn@ reactors fall into different categories, andcompounds such as the epoxides and ethyleneimines produce effects similar to the mustards because of their relatively low reactivity which enables them to reach most parts of the cell unaltered, coupled with a satisfactory gradient in theâ€œspreadâ€•of their affinity toward different nucleophiles.

Some of the biological and pharmacologicalproperties of the alkylating agents will now be discussed in the light of their chemical reactivities.The time will surely come when the barriers whichdivide chemistry from biology must crumble andthey will merge into one conceptual framework.If the static concepts of chemistry could be mademore dynamic, and the dynamic aspects of biologymore static, then the meeting point would seemmuch nearer.

THE EFFECT OF ALKYLATING AGENTS ON

Nucj@n MATERIAL ANDCELL DIVISION

When it was discovered that in many respectsthe bifunctional chloroethylamines and chioroethylsulfides produced cytological and pharmacological effects (e.g., cell death, chromosome abnormalities, degenerative changes in the bonemarrow and testis, suppression of the immune nosponse) similar to those produced by ionizing radiation, the term â€œradiomimeticâ€•was applied to them(@8). Although certain of the alkylating agentsproduce end-results which appear superfluouslylike those produced by radiation, detailed investigation has often revealed significant differencesbetween them, and the term in many respects leadsto confusion in the minds of chemists and biologists alike. Loveless and Revell (56) set forth theirprecise biological criteria upon which chemicalsubstances were to be classified as radiomimetic,and this involved restriction â€œtothose substanceswhich can be shown unequivocally to react uponthe resting cell in such a way as to produce analteration of the genetic material which is revealedby the appearance of chromosome breakage andrearrangement in subsequent divisions.â€• Certaincompounds which caused pyknosis, and whichDustin had referred to as radiomimetic, were notcovered by this definition. Others have preferredto retain the original broad definition, but sincethis time so many alkylating agents have beendiscovered to mimic superfluously the effects ofradiation in, say, one respect only, that the termceases to have great significance. Ethyl methanesulfonate, for example, shows none of the propertiesof radiation except that it is a powerful mutagen(53). Mylenan, possibly atypical in some respects,nevertheless has a profound destructive effect ongranulocyte production (3@), causes certain chromosomal abnormalities (45), depresses the immune response (1@) (in a manner very similar toradiation) and causes infertility in rats (48) (alsoin a manner reminiscent of radiation) ; but, unlikecertain mustards, it has only a moderate inhibiting effect on the transplanted Walker rat carcinoma. Thus these compounds would seem to beâ€œradiomimetic,â€• partially â€œradiomimetic,â€• orâ€œnonradiomimetic,â€• depending on the particularbiological phenomenon being considered. Koller(46) has compared the effectsof several alkylatingagents with those of radiation and concludesâ€œthatin view of the dissimilarities, it would be anerror to infer a similarity of the mode of action ofalkylating agents and x-nays, basing the inferenceon the similarity of some end products.â€•

Of the various types of alkylating agent whichhave been tested for tumor growth-inhibiting



Cancer Research Vol. 23, September 196313@2

properties, those which have emerged as the mostactive include the Sni reactors such as the aromatic nitrogen mustards and the Sn@ reactors ofthe HN@ type, epoxides, ethyleneirnines. OtherSn@ reactors such as the aliphatic methanesulfonates (Myleran) show a low order of activityagainst the transplanted Walker tumor, whereasthose Sn@ reactors such as the a-halo acids andketones have been found to be inactive whilepossessing a high order of general toxicity. Theother factor of importance in determining efficacyin this respect is functionality, and it is now wellknown that bifunctionality leads to a profoundincrease in capacity to interfere with mitosis, provided the reactive centers are spaced at an optimum distance apart. Thus in the chemical sensegreatest activity is associated with undoubted

Snl reactors or those Sn@ reactors which, as discussed earlier, show some tendency toward theSni mechanism in the bond-breaking and bondmaking process and which have a reasonable noactivity toward most nucleophiles. Biesele et al.(18), in discussing the chemistry of alkylatingagents, postulated that the most active compoundsin causing chromosomal aberrations were thosewhich possessed or could be transformed into unstable threo-membered rings. In a comment on thispostulate, Loveless and Ross (57) drew attention

to the â€œradiomimeticâ€•properties of 1,4-dimethanesulfoxyloxybutane (Myleran) in inhibitingthe growth of Sarcoma 180 and in its effect on theblood and bone marrow, and the capacity ofdimethylsulfate to produce chromosome breakagein plant material. Since both these compounds areincapable of forming three-membered ring systems, the postulate of Biesele et at. was obviouslyan oversimplification, and Loveless and Ross (57)argued that capacity to form a carbonium ion was

a more general feature of â€œradiomimeticâ€•chemicals. Actually, it is doubtful whether epoxides,ethyleneimines, and methanesulfonates do, in fact,form true carbonium ions under conditions likelyto be met with in the organism, and it is difficulttherefore to accept this generalization as it isworded. As indicated earlier the term â€œnadiomimeticâ€• is often applied far too indiscriminately tosometimes ill-defined biological phenomena, andit would seem more profitable and less misleadingto define carefully the biological effects manifestedby each alkylating agent rather than to refer tosome property as â€œnadiomimeticâ€•because it mayvaguely correspond to a particular property of

radiation. Any apparent similarity between thebiological properties of radiation and the alkylating agents undoubtedly results from the capacityof each entity to modify cellular structure and

function by direct combination with cellular components such as nuclear and cytoplasmic material.The most â€œradiomimeticâ€•chemicals will be thosecapable of penetrating deep into the cell and of noacting with a diversity of cellular sites, whereasless reactive chemicals or those which alkylate bya different basic mechanism may possess far fewerâ€œradiomimeticâ€•properties.

Tumor inhibition with the alkylating agents nosults from their cytotoxic action, which probablyis the result of a complex series of changes varyingin fundamental mechanism with different types ofalkylating agent, and even from cell type to celltype. Lethal cytotoxic action is usually of twotypes called interphase death and mitotic death.Interphase death occurs without cells' going intodivision or attempting to go into division. Alkylating agents vary in their capacity to kill cells at thisstage, and individual cell types vary in their susceptibility to this type of death. Thus fully differentiated lymphocytes tend to be extremelysensitive to the action of mustards (but not toMyleran), whereas most other differentiated cellsdo not. The mechanism of interphase death isprobably complex and results from damage tomany cellular sites.

In mitotic death the primary damage is thoughtto occur in the â€œnestingstage,â€• and chromosomeaberrations and other effects become cytologicallyapparent at the time of division. Many changesoccur, such as clumping and bridging of chromosomes, and some of the effects are cumulative during successive cell divisions leading eventually tononviable cells. The action is not restricted tomalignant cells, but extends to proliferating cells

of all types (35).Various differences in detail of action have

been reported to exist between the mustards (e.g.,HN@) and the methanesulfonates (e.g., dimethylmyleran) . Thus in comparing their actions onlymphoma cells in culture, Alexander (3) foundthat HN@ immediately arrested cell division,whereas dimethylmyleran behaved like x-rays,and at least one division occurred before cell proliferation stopped. Koller (45) has compared indetail the cytological effects produced by variousalkylating agents and again found differences intheir modes of action and also differences in nosponse to varying doses of alkylating agent. Hefound that in a series of dimethanesulfonyloxyesters, the greatest diversity of effects on cells ofthe Walker tumor was shown by Myleran. He alsofound that it could produce suppression of mitosis,and â€œradiomimeticâ€• and cytotoxic effects, although these probably differed in detail from thoseobserved with the mustards. Alexander reported



\\ARWICKâ€”Mechafl?$m of Action of Alkylating Agents 13@23

(3) that, after all treatments, there was an increasein volume of cultured lymphoma cells; with HN@the cell volume doubled, and then there was degeneration, whereas with x-rays and dimethylmyleran a wide spectrum of cell sizes was obtained,including â€œgiants,â€•although nuclear granulationsimilar to that seen with HN@ was not observableafter @4hours.

Alexander' has also found that iodoacetic acidproduces different effects from either the mustardsor the methanesulfonates on lymphoma cells. Itseems to have little effect on cell division, butwhen a critical dose level is reached the cellsbreak up. This would be consistent with the noactivity of this compound mentioned earlier, resulting in its major reactions occurring in the cytoplasm (67). The need for bifunctionality and acertain critical level of nuclear reaction (whetherthis be with nucleic acid, protein, or cofactor) seemessential for â€œradiomimeticâ€•cell death. â€œCytoplasmic deathâ€• may be a factor of importance inconsidering mustard action on fully differentiatedlymphocytes and indeed in the inhibition of tumorgrowth in general. Koller (45) considers that many

factors must be considered when assessing thecauses of cell death in a dynamic population andthat â€œtheability of the alkylating agents to produce nadiomimetic injuries may not be the mostimportant property that determines the effectiveness of these drugs in the therapy of cancer.â€•

EFFECTS ON THE BLOOD

Elson (30) showed that the effects of wholebody x-radiation on the number of circulatingplatelets, lymphocytes, and neutrophils could beessentially reproduced by the administration ofsuitable combinations of Chlorambucil (a bifunctional mustard) and Myleran, and he concludedtherefore that each of these chemicals was onlypartially â€œradiomimetic.â€•Elson, Galton, and Till(3@) consider that Myleran appears to act on the

resting stage and prolongs the intermitotic interval, whereas mustards appear to act mainly oncells undergoing mitosis and possibly during thelatter part of the DNA synthetic period. They alsoexert a particularly destructive action on maturelymphocytes which Mylenan does not. They foundthat the effects of Chlorambucil appeared rapidly,were transient, and were accompanied by conspicuous mitotic abnormalities. Recovery proceeded rapidly except in the lymphoid tissue, andthere was a transient overcompensation in thegranulocyte series. Myleran, on the other hand,produced gradual effects, depression of hemo

poiesis was prolonged, cytological abnormalitieswere not seen (cf. 45) at the doses used, and mitosisproceeded normally. They explain these differenteffects on the basis of the capacity of the mustardsto attack ruthlessly nuclear material in the restingstage and to lyse fully differentiated lymphocytes,whereas they consider that Myleran prolongs theintermitotic interval in all proliferating cells inrespective of the stage in the mitotic cycle at whichthey are exposed. Differences in normal intermitotic levels which vary with different cell typescould explain the effects. Lajtha (47) has mdicated that erythropoietic cells normally divideabout twice as often as granulopoietic cells, anddelay in each series would explain why depletionof erythropoiesis occurs more rapidly than that ofgranulopoietic cells. The authors do not considerthat Myleran has any selectivity for particularcell types and â€œthatthe apparent resistance oflymphoid tissue to Myleran may depend as muchon the ratio of mitotic frequency to life-plan as tothe lack of damage to mature lymphocytes.â€•

In their effects on the circulating gnanulocytesand lymphocytes, Elson has found (30) that theepoxides and ethyleneimines produce effects similan to those produced by the mustardsâ€”not sunpnising in view of the similarity in reaction kineticsdiscussed earlierâ€”whereas he found the effectsproduced by the carcinogens 4-aminostilbene andthe polycyclic hydrocarbons to be more like thoseof Myleran, a result difficult to explain at presentbut possibly of great significance. Ethylmethanesulfonate (half-Myleran) was not active.

As in the case of tumor growth-inhibition, bifunctional alkylating agents of all series were muchmore efficient than the corresponding monofunctional derivatives, and in the Mylenan series optimum activity was found when n = 4 or 5 (31).Potent Sn@ reactors such as a-halo acids andketones apparently have not been tested as extensively as the foregoing compounds, whereasalkyl halides such as dibromubutane (group [c]Sn@2reactors) possess no detectable activity (notelower chemical reactivity).

EFFECTS ON MALE FERTILITY

Very interesting results which may contributeto an elucidation of the mode of action of x-raysand the alkylating agents have been found by comparing the effects of these agents on the process ofspermatogenesis in rats and mice. The process follows the path spermatogonia â€”@spenmatocytes â€”+spenmatids â€”+spermatozoa in epididymis, and ithas been found that different agents can affect different stages in the process. Once again it tranI p@ Alexander, personal communication.



Cancer Research Vol. @23,September 196318@4

spires that Myleran most closely follows the effectsof radiation.

Intrapenitoneal injection into mice of aiiphaticmustards produced destructive changes in thetestis, and recovery was observed 3â€”4weeks later(48), whereas in rats potency returned after @â€”4months (34) . It was of interest, however, that twoaromatic nitrogen mustards, Melphelan and Chlorambucil, did not interfere with the fertility ofmale rats (4@, 44) at doses which were adequateto cause transient leukopenia in the peripheralwhite count (@9) and inhibit tumor growth (13).Jackson (43) concludes, therefore, that capacity toinhibit tumor growth or to interfere with the bonemarrow does not imply capacity to damage

germinal epithelium. This, of course, may havebeen due to poor penetration rather than to a different mode of action.

2,@,6-Triethyleneimino-1,3,5-triazine (TEM),which was placed with the epoxides as far aschemical reactivity is concerned, produced effectsindicative of selective interference with certain

stages of spermatogenesis (19), mainly, probably,with spermatocytes (11, 4@), since the infertilitywas found at 4 weeks and recovery was rapid aftera single dose of 0.@mg/kg. Five daily doses of 0.@mg/kg, however, caused sterility for several weeks,and early and late stages of spermatogenesis wereaffected, and probably mature sperm in the epididymis were also attacked (as with the mustards).The relative insusceptibility of spermatogenia isin contrast in that they are the germinal tissuemost sensitive to radiation (6@, 79) and to Myleran.

After one dose of Myleran (11 mg/kg) the testishistology remained normal until the 8th week,when sterility rapidly developed (@6, 4@), indicating an effect on an early stage of spermatogenesis.The other germinal cells present at the time oftreatment appeared to develop normally forabout 45 days, resulting in a systematic depletionof spermatogonia, spermatocytes, spermatids, andspermatozoa, in that sequence (4@). Increaseddoses still did not affect other spermatogenic cellswhich continued to proliferate.

Dimethylmyleran produced similar effects at alower dose level (44). This apparent specificity ofaction of Myleran is probably in some way relatedto its chemical reactivity. The fact that the effectswhich it produces are similar to those produced byradiation and unlike those produced by the mustards and TEM is probably highly significant.

EFFECTS ON THE IMMUNE RESPONSE (1@)

Different effects were found on the immune nosponse, depending on the agent tested. Thus, it

was found that irradiation and Myleran suppressed antibody production maximally if givenabout @â€”4days before the antigen, whereas nitnogen mustard and TEM did so when given about

@ days after the antigen, and antibody productioncould be suppressed for some weeks in each case,following a single injection given at the appropniate time interval. Once antibody production isunder way, however, these agents have no furtherinhibiting effect. Berenbaum tentatively explainsthese relative effects by assuming that immunologically competent cells are produced from pnimitive stem cells which give rise to more differentiated cells which can produce antibody. He envisages early cell types to be sensitive to the actionof x-nays and Mylenan, but not to TEM or themustards, whereas the later, more differentiatedforms are sensitive to the latter agents, but not tothe former. Other alkylating agents have unfortunately not yet been tested, but the foregoing nosults will be discussed later in the light of knownreactions of these agents in vivo.

CHEMICAL INTERACTIONS WITHCELL CONSTITUENTS

General conaiderations.â€”In discussing the basicchemistry of the alkylating agents it was foundlogical to separate these compounds into varioustypes or groups depending on whether they wereessentially Sni reactors on Sn@ reactors, and funther subdivision was possible when a more detailedassessment was made. In considering some of thepharmacological properties of the alkylatingagents, it was found that, where comparativeassessments have been made, in a broad sensethese faithfully reflected the type of alkylatingagent involved. Thus, one can distinguish betweenthe pharmacological properties manifested by themustards, the epoxides and ethyleneimines on theone hand, and the alkyl methanesulfonates on theother, in considering their pharmacological effects.Although fewer biological studies have been carned out on the alkyl halides and the a-halokotones and acids, it is evident that again these differfrom one another and from the above compounds.

Since these complex processes involving cellsand alteration of cellular structure and functionalmost certainly involve direct chemical combina

tion with cellular constituents, and since the levelof any particular combination will depend onmany variables, the most important of which is

probably chemical reactivity, then an examinationof cellular components after treatment, or ofaltered biochemical integrity after treatment,should help to pin-point the lesions involved.This type of reasoning, valid on misleading, has



WARwICKâ€”Mechanism of Action of Alkylating Agents 13@2@5

guided most workers studying the mechanism ofaction of alkylating agents for many years. Thegeneral opinion seems to be that, in the case ofcompounds such as the carcinogenic hydrocarbons,

the carcinogenic azo dyes, etc., we must keep anopen mind whether these agents cause their effectsby actually binding to cellular constituents onwhether they are doing something much moresubtleâ€”such as finding their way between macromolecular strands and exerting their effects by essentially physical means ; however, in the case ofthe alkylating agents, cellular binding is of primeimportance.

To digress a little, there is good reason for believing that the carcinogenic hydrocarbons andazo dyes do, in fact, need optimum structural requirements and particular electron distributions

(7) which may enable them to alter cellular function by some kind of physical means; but there isgrowing evidence that compounds of this type canin fact bind firmly to cellular constituents, both toproteins and to nucleic acids. To quote but twoexamples, Heidelberger (36) has shown dimethylbenzanthracene to combine both with protein andDNA in vivo, whereas the Millers (60) and othershave shown dye binding to liver proteins followingadministration of carcinogenic azo dyes, and, morerecently, Roberts and Warwick (68) have shownthe combination of a metabolite of dimethylaminoazobenzene with nucleic acids in vitro.These findings tend if anything to cloud the complicated situation even further; they may be misleading, but if direct covalent combination withcellular constituents is proved to be a potent forcein producing all the pharmacological properties ofthe alkylating agents, then these findings mightbe of the utmost importance in elucidating at leastsome of the phases of the complex carcinogenicprocess with these and other agents.

In the case of the alkylating agents where onecompound can be cytotoxic, carcinogenic, andmutagenic, the correlation between chemical reactions at the cellular level with any of these manifestations becomes exceedingly difficult. However,attempts have been made to study the types ofchemical reactions which some of these reagentscan undergo inside cells, and in the main the nosults obtained have indicated a close parallelismbetween the types of reactions observed in vitrousing purified chemicals and those observed invivo, and it has also been found that levels of binding to various constituents were as one mighthave anticipated on the basis of the reactivity ofthe drug concerned. These studies are probablyimportant if only to show that definite covalentbinding can occur between alkylating drugs and

cellular constituents, the nature of this binding insome cases, and that in vitro studies can often helpto elucidate tantalizing in vivo problems. Theprogress made so far, much as it seems when one

assesses it against the almost complete void whichexisted 10 years ago, is still extremely limited; and,although we are in a position to make some intelligent guesses as to the mechanism by which thealkylating agents exert some of their effects (e.g.,their cytotoxic action), it would be unwise to makeany dogmatic assessments at this stage, no matterhow justifiable they may seem. However, the sunface has been scratched, and doubtless the task ofa reviewer at the end of another 10 years will be acomparatively reassuring one.

Metabolism of ethyl methanesulfonale and Myleran.â€”The first clear-cut demonstration that apharmacologically active alkylating agent did infact alkylate nucleophilic centers in vivo was ohtamed when Roberts and Warwick (65) demonstrated the presence of N-acetyl-S-ethylcysteineand related compounds in the urine of rats whichhad received C'4-ethyl methanesulfonate by intrapenitoneal injection. The bulk of the radioactivitywhich appeared during the first few hours wasprobably derived from reaction with glutathioneknown to be present in high concentration inmammalian cells and with which ethyl methanesulfonate reacted readily in vitro. That which appeared later probably came from ethylated protein. Because of the related structure of 1,4-dimethanesulfoxyloxybutane (Mylenan) and itsenchanced pharmacological properties, it was considered of interest to examine the urinary productsexcreted after its injection. Because of the bifunctional nature of the drug it was considered thatsome type of cross-linked product might be formedin vivo. However, the major urinary metabolitewas found to be 3-hydroxytetrahydrothiophene1,1-dioxide (66), which was probably derived froma sulfonium compound formed by the interactionof both alkylating arms with one thiol group. Thiswas followed by the removal of tetrahydrothiophene and its further metabolism to the hydroxysulfone. The reaction is of interest, because it involves not only modification of an existing thiolgroup but its later removal from the amino acidmoiety in the form of tetnahydrothiophene. In thecase of protein reaction, this could lead to the production of a new sequence of amino acids by thereplacement of an existing molecule of boundcysteine by a molecule of bound lanthionine orbound aminoacrylic acid (69). Bound lanthioninewould result if the sulfonium compound was attacked by another molecule of ionized glutathione,whereas bound aminoacrylic acid would result by



Cancer Research Vol. p23, September 196313@?26

hydrolytic decomposition (Chant 8). The dehydropeptide or dehydroprotein might undergo additionof a molecule of glutathione or other thiol formingbound lanthionine.

It is extremely difficult to assess the importanceof these reactions when considering the mechanismof action of Myleran. It has been stated (@1) thatthe low level of reaction of Mylenan with nucleicacids in vitro and in vivo at therapeutically effective doses makes these centers seem unlikely vitalpoints of attack in this case. It is possibly relevantthat, in the series of dimethanesulfonyloxy esters,those having four or five methylene groups are themost active in depressing granulocyte count andare also those which could undergo the ring closure

NHRCH2â€”CH2o I012 CH2@â€• .â€”@ I ::s

G@@ CH2â€”Q1@

e

i@ NHRr2_cI@4H_c@

cii@â€”cH@ @0R'

CHART 8.â€”Reactions of the sulfonium ion formed from

Myleran and a cysteinyl moiety with glutathione and withhydroxide ion.

(_@ N'2@ G ---@ I\-N@ CH2IG\=J @cH2â€”cii@ \@=/ â€˜Qi2@CH2-SG

@.sGCHART9.â€”Reaction of an aromatic nitrogen mustard with

glutathione to form a cross-linked product.

and dethiolation depicted earlier. Also in this connection one should recall the many similaritieswhich exist between the pharmacological effectsproduced by Mylenan and ionizing radiations.

The change in amino acid sequence in a proteinor peptide chain constitutes a type of chemicalmutation induced without the intermediate intervention of DNA or RNA. If certain proteins,either in the nucleus or the cytoplasm, act as suppressons or cell-regulators, then one could envisagesuch a change having possibly far reaching effects,and here may be recalled the drastic effects of thechange of one amino acid unit in a protein in relation to the production of sickle cell anemia (38).The difference between the normal hemoglobinand that in the mutant cell is the replacement of

one glutamic acid unit by valine in a polypeptidechain of 300 units. However, these ideas are speculative, and more data regarding the extent of protein alteration by Myleran in vivo and the functions which these proteins play in cellular integrityare needed. The related compound, dimethylmyleran, which in many respects is similar to Myleran in its pharmacological properties, should be auseful tool in elucidating some of these problems,since if reactivity toward anions is low and forsteric reasons ring closure and dethiolation are notpossible in this case, then the relevance of thesechanges to the mode of action of Myleran would bequestionable. Work on this aspect of the problem isin progress, with highly tritiated dimethylmyleran.In connection with the mode of action of Myleran,Timmis (85), on the basis of its similarity inaction to thioguanine, has suggested that it mightcombine with nucleic acids in vivo or with nucleotides to form a type of antimetabolite which couldact by lengthening intermitotic intervals ratherthan directly killing cells. He suggested thatMyleran might first interact with homocysteine toform a compound capable of alkylating purinebases in a similar manner to adenosyl-methionine.

Metabolism of aromatic nitrogen mustards andmustard gas.â€”Itwas considered of interest in viewof the ring closure reaction of Myleran to determine whether similar reactions occurred in vitroand in vivo in the case of the bifunctional mustards. Reaction with thiol groups in vivo seemedprobable, and in the case of the bifunctional compounds ring closure also seemed probable in viewof the optimum spacing of the reactive centers,since six-membered rings are normally particularlystable.

Under in vitro conditions with the use of bischloroethylaniline, bis-chloroethyl-p-anisidine, ormustard gas, and either cysteine or glutathione,evidence was obtained that intermediate sulfonium compounds were in fact formed (70), buttheir instability was such that isolation proved impossible. In the presence of excess cysteine orglutathione the only products isolated were thecross-linked cysteinyl or cross-linked bound cysteinyl compounds, in contrast to the situationfound with Myleran. Thus, in this case dethiolation did not occur. Assuming the formation of anintermediate sulfonium ion, the reaction sequencefollowed in the case of glutathione and anilinemustard would be as shown in Chart 9. The secondary reaction of the sulfonium ion with gluta

thione was obviously so fast that hydrolytic cleavage to 4-phenylthiazan was not possible. The interesting feature of these reactions is, of course, thepoint of attack on the sulfonium compound by the

,NHR+ Gsâ€”cIl,â€”cH

@COR'

CH,â€”cH2 ,NHRâ€”* I S+@H@zC

cH2â€”cH2 \ ScoR



WAnwI@â€”Mechanism of Action of Alkylating Agents 13@27

glutathione molecule, leading to ring opening, toform cross-linked products instead of dethiolationand ring release as in the case of Myleran.

On the basis of these observations, cross-linkingof protein strands in vivo following treatment withreactive bifunctional mustards seems likely (71),and because of the chemical properties of themustards discussed earlier, restriction as to thecellular sites involved would seem less likelythan with powerful Sn@ reactors such as iodoacetate. Thus, the mustards because of their unreactive nature prior to ionization would be morecapable of deeper penetration into cells and,hence, reaction with a greater range of thiols thanthe Sn@reactors of the iodoacetate type. The arguments used in the past to dismiss the importanceof thiol reaction in determining any of the important pharmacological properties of the alkylating agents are for these reasons not valid (74, 8@).The range of so-called specific -SH reactors testedfor their capacity to inhibit tumor growth, etc., hasbeen limited; the compounds have often beenmonofunctional, when it is known that bifunctionality is a requisite for high antitumor activity;and most of the bifunctional agents tested havenot had their reactive centers spaced an optimumdistance apart (5@). In addition, account wasobviously not taken of the probability that, because of the inherent chemical reactivity of suchcompounds, they would probably never reach cellular sites accessible to agents such as the mustards, epoxides, ethyleneimines, and, to a lesserextent, the aliphatic methanesulfonates. Now thatit is known that both the mustards and the auphatic methanesulfonates are capable of extensivethiol reaction with -SH groups in vivo, coupledwith the knowledge that the mustards can crosslink, while Myleran can bring about a unique typeof dethiolation and at cellular sites probably inaccessible to -SH reactors of the iodoacetate type,it will be necessary to reconsider the possible nelovance of particular types of -SH reaction to themode of action of the alkylating agents. Greaterknowledge of the site and extent of reaction is noquired and, in particular, whether thiol reaction inthe nucleus plays any role in inhibition of celldivision or in those processes leading to malignancy. If in the future alkylation of proteins orpeptides and, in particular, alkylation of thosebearing reactive thiol groups is shown to have nopart to play in determining any of the importantpharmacological propertiesof the alkylatingagents,the experiments leading to such conclusions willbe mainly those which now remain to be done,since those which have been done can lead us tono definite conclusion. The writer is not a sup

porter of any theory of mode of action whosefoundations nest on particular reactions at particular cellular sites, which excludes other possibilities, but he wishes to draw attention to certainoversimplifications in reasoning which have led topremature conclusions regarding -511 reaction,where in fact a great deal of work remains to bedone. It may transpire that much of the reasoningused in the past to explain the mode of action ofalkylating agents has its foundation in ovensimplification, and although theories to explain certainpharmacological events are often useful in leadingto further experimentation, they can be a drag onrapid progress if their exponents and others regardthem as an end in themselves.

Reaction of aikylating agents with nucleic acids.â€”Since the importance of the nucleic acids in cellular metabolism was established, many workershave been attracted to the idea that all the impontant biological properties of the alkylatingagents can be explained on the basis of reactionwith nucleic acids, particularly DNA. Since thealkylating agents cause gross mitotic abnormalities and can effect gene mutations, this hypothesishas always had many followers. Nobody woulddoubt that the alkylating agents lead to an altenation in the structure and function of nuclear DNAand consequently in cellular. integrity, but there isevidence that some of these alterations might bemediated by indirect effects at many sites, such asthose indicated earlier when thiol reaction wasbeing considered.

The effects of the alkylating agents on DNA,RNA, and protein synthesis have been recentlywell reviewed by Wheeler (87). It transpires thatthe mustards are capable of inhibiting DNA synthesis to a greater extent than RNA synthesis,whereas the methanesulfonates such as Myleranhave little effect on eitherâ€”results which may helpto clarify the apparent differences in phanmacological properties of the two classes of compound.

Considering the popularity of the concept of direct action on DNA it seems incredible that theactual loci of the reactions should be overlookedfor so long. Despite the work of Young and Campbell (90), who first drew attention to the reactivityof the guanine moiety, and of Wheeler, Morrow,and Skipper (88, 89), who pointed out thatquaternization of a ring nitrogen atom was likelyin the reaction of certain punines with alkylatingagents, attention continued to be directed towardreaction with ionized phosphate groups as the mostprobable points of reaction in the nucleotides (@,74). Although estenification of phosphate groupsseems probable in some cases it is a difficult neaction to demonstrate because of the lability of the



Cancer Research Vol. 23, September 196313@8

products formed and because differences in theliberation of acid, interpreted as being due toesterification, also occur following quaternizationof the ring nitrogen atoms. It remained for Lawleyand Wallick (50), possibly stimulated by the remarks of Skipper (80), to demonstrate that dimethylsulfate reacted with guanylic acid to form7-methylguanine, and it was further shown byLawley (49) that the product formed from deoxyguanylic acid and dimethylsulfate was much morelabile and that liberation of 7-methylguanine occurred appreciably at pH 7.@ (37Â° C.). Thus, a

mechanism was established by which alkylationof guanine moieties of DNA could effect changes

0H(cH3),S04

@IcHHZN'@N'@@I@

R

at pH 7, and in this case breakdown probably occurs with the formation of an aldehydopynimidine derivative by ring opening (Chart 11).

In the case of reaction with DNA, the apuninicacid formed after removal of the alkylated guaninewill be relatively unstable and could lead to mainchain breakdown.

These reactions may be of fundamental importance in explaining some of the phanmacological effects of the alkylating in mechanistic chemical terms. Thus it was postulated (55) that thegreater efficiency of certain bifunctional alkylating

agents in killing cells and interfering with mitosismight be due to their capacity to cross-link partsof DNA, or DNA to protein, or protein to protein.We have seen how proteins might be cross-linked(71), some evidence is available for linking ofnucleic acid to protein (77), and Alexander, Cousens, and Stacey (4) obtained evidence of crosslinking of DNA in nucleoprotein with nitrogenmustards. This they interpreted as being solelydue to reaction with phosphate group, but in thelight of the findings of Brookes and Lawley (@1)that bis-guanine products are formed after reaction of various alkylating agents with nucleicacids, they have had to reconsider their interpretation, and it is the belief of Lett, Parkins, and Alexander (51) that in some cases the alkyl groups ofphosphate esters can transalkylate to the 7-position of guanine accounting for the apparentlyhigher reactivity of guanine when in DNA thanwhen it appears in nucleosides (75, p. 79). Themain evidence which they present for this comesfrom spectral data, the U.V. changes which accompany base alkylation being delayed in somecases as if two reactions were occurring; and theyalso interpret changes in â€œcoilingâ€•as being due toestenification. Whatever the true interpretation,the fact remains that parts of DNA can be crosslinked in vivo, and DNA will be more or less inactivated and degraded depending on the extent ofreaction. Some of the cytotoxic effects of, say, themustards will be explicable on this basis, butwhether the whole complex process of lysis of fullydifferentiated lymphocytes, for example, can beexplained entirely by such mechanisms remains tobetested.

Brookes and Lawley (@1, @)have studied theextent of combination of various alkylating agentswith DNA and have found a simple rule toapplyâ€”the more reactive the compound and thegreaten its tendency to react with all nucleophiles,the greater the extent of reaction with the guaninemoiety. Thus the mustards reacted extensively,whereas the methanesulfonates, halides, and sul

CHART 11.â€”Breakdown of an alkylated guanylic acid to

form an aldehydropyrimidine derivative.

due to quaternization, elimination of the alkylated guanine moieties at physiological pH, andpossibly fission of the polymer chain.

The reaction can be depicted as in Chart 10,when R = nibophosphate or deoxynibophosphate,and the product will have another form contnibuting to the resonance hybrid in which the positivecharge resides on the alkylated nitrogen atom.Where R = deoxynibophosphate, breakdown canprobably occur in two ways, leading in the onecase to the 7-methylguanine and a carbonium ionR@which will react with a molecule of water leadingto the liberation of acid and an apuninic acid, orbreakdown could occur at the other nitrogen atomleading to a regeneration of the original deoxyguanylic acid and liberation of CH@. Alkylated (e.g.,methylated) guanylic acids are much more stable

oH?@'

H@N@%

R- raos.@@ Ã @p@sta

CHART 10.â€”R.eaction of deoxyguanylic acid or guanylicacid with dimethyl sulfate.

0HcH3

H,NR



WARWICKâ€”Mechanism of Action of Al/cylating Agents 13@9

fates reacted much less extensively and more slowly. These findings again reflect the differences inpharmacological properties displayed by the vanous compounds. Thus the mustards, highly reactive toward most nucleophiles, able to penetratedeep into cells, able to cross-link nucleic acids andprobably proteins, are able to create havoc amongcells, causing effects from inhibition of cell divisionto complete cellular disruption and lysis in contrast to the less reactive reagents such as Myleranwhich would be expected to react very little withDNA at therapeutic doses and unable to cross-linkproteins through their thiol groups (see earlier).The greater efficiency of mustards in causing celldeath probably results from their heightenedcapacity to react with all nucleophiles within thecell, not merely to their greater reactivity towardDNA.

The low level of reaction of Myleran withnucleic acids in vivo leads one to seek for otherexplanations for its pharmacological effects. Thereis no scope in this discussion for considering otherimplications of DNA reaction such as mutagenesisand the difficulties which have been found in interpreting some of the biological phenomena encounteredâ€”for example, the â€œuniquenessof ethylationâ€• (@3, 51, 53) ; but let it suffice to say that,although difficulties have been encountered ininterpreting mutation in the light of phosphate

esterification (f), a working hypotheth has beenput forward by Lawley and Brookes (%4) whichseems to fit most of the facts and which can prob..

ably be tested.

MUTATION AND CANCER

Although bifunctionality in alkylating agents isa normal requisite for high killing action againstcells, this is not the case with regard to mutagenesis and carcinogenesis, since in these respectsmonofunctional agents are often as active or moreactive than their bifunctional counterparts, andthis is in line with the concept that cross-linking islethal. Some evidence exists that the physiologicalstate of the cell influences the occurrence ofâ€˜spontaneous' or induced chromosomal abnormalities (15â€”17) in some cases; Loveless (54) hasstrongly supported the hypothesis that mutationin the case of the alkylating agents involves directattack on nucleic acid. The fact that monofunctional alkylating agents can be both mutagenicand carcinogenic coild be used as evidence tosupport the mutational theory of cancer. On theother hand, one should not ignore appealing alternatives which involve much more extensivealteration of cellular integrity and alteration in

delicately balanced cellular homeostasis. There ismuch evidence to support an alteration in the integrity of mitochondria and the endoplasmicreticulum during administration of carcinogens (8,9), a consequence of the alteration of lipo-proteinmembranes which constitute the structural support of the organized multi-enzyme complexes.Thus 3'-DAB (dimethylaminoazobenzene) has aprofound effect on microsomal swelling whenfed for 4 weeks (6), in contrast to the inactive

@-methyl DAB, and the former compound alsoproduces a significant decrease in the mitochondrial population in the cell (64, 78). During theearly stages of carcinogenesis with any agent thecell population is being continuously modified, andmany cells die as a result of excessive interferencewith their cytoplasmic and nuclear elements.Those extensively modified cells which survive willbe those which have been able to adapt themselvesto the unfavorable conditions; but with many oftheir complex feedback controls and fine structuredamaged they will be unable to respond to thenormal stimuli of intercellular dependence and willcease to obey the stringent rules of organismalorganization (81). The great complexity of theintracellular effects produced by the alkylatingcarcinogens, ranging from slight alterations in cellstructure to cell death, makes the problem of sorting out the chain of events leading to malignancyone of the utmost difficulty. The relatively longtime lapse between treatment with the carcinogenand the emergence of malignant tumors is probably related to the multitude of changes occurringin the cell population during the â€˜adaptive'period. The oncogenic viruses are probably morespecific in their effect and faster acting because oftheir information content and may eventuallyprove to be better tools for elucidating the chainof events leading to malignancy than the chemicalcarcinogens which in many ways have not livedup to earlier expectations.

CONCLUSION

No attempt has been made to discuss the actualbiochemical lesions such as effects on cellular synthetic mechanisms which result from the action ofthe alkylating agents. Effects on glycolysis, enzymes, and so on have been well reviewed byWheeler (87). Rather, an attempt has been madeto correlate the basic alkylating capacities of thevarious compounds with gross biological endproducts, and some attempt has been made toclassify the various agents into groups or classesas far as possible.

Some conclusions can be reached regarding



Cancer Research Vol. @23,September 19631330

these comparisons, the main one being that themost reactive compounds of the Snl or pseudoSni type with a satisfactory â€œspreadâ€•in their reactivity toward all nucleophiles are those whichmanifest the most diverse pharmacological effects.It must also be relatively obvious from the foregoing discussion that none of the studies carriedout on the mechanism of action of alkylatingagents has given any information as to how betterdrugs with a greater selectivity of action mightbe designed, although suggestions have been madeas to why certain drugs are more efficacious thanothers. This statement is depressing, and yet theobvious must be facedâ€”and this is that thealkylating agents so far designed are fan too crudein their killing properties to be of any permanentuse in cancer therapy. Of course one must alwaysleave room for the unexpected breakthrough whichmay take the form of a drug with a special â€œcarrierâ€• group (14) or a drug possessing â€œlatentâ€•activity, and yet attempts which have been madeso far to make this kind of breakthrough haveproved somewhat disappointing. Thus, for example, Ross and Warwick (76, 86) made a seriesof mustard derivatives of azobenzene which wouldremain chemically unreactive until reduction ofthe azo linkage occurred in vivo, and it was hopedthat some selectivity of action might be achieved.However, the compounds are relatively toxic, andone (40) has undergone clinical trial with disappointing results. Yet more of this type of researchis needed.

Compounds possessing all the various types ofchemical reactivity have been tried, and noobvious solution to the problem of what to try nexthas been forthcoming. Thus, compounds such asMyleran, Chiorambucil, etc., which are now usedas palliatives in the treatment of chronic leukemiaand Hodgkin's disease, have certainly played apart in relieving distressing symptoms and haveenabled certain patients to remain active whenotherwise they may have been bedridden; and inthis sense the efforts which have gone intoplanning them and synthesizing them has probably been justifiable. However, whether manymore of these compounds should be made is adebatable point.

The alkylating agents have stimulated much research on fundamental problems concerning cellmetabolism, mutagenesis, cancinogenesis, and soon, and it is probably in these fundamental waysthat they have proved most useful. However,their very number, the diversity of pharmacological effects which they manifest, and the number ofcellular sites with which they interact preclude

facile explanations for their mode of action. Wenow have some idea of the chemical sites withwhich a few of them interact intracellularly, butfrom a very mechanical organic chemical viewpoint. Much more work now needs to be directedto the study of their effects in relation to cell population dynamics and their effects in relation tointercellular dependence and to dedifferentiation,since it is in these dynamic aspects that theexplanation of the carcinogenic process lies.

Wheeler (87) indicates at the conclusion to hisreview that there is a need â€œtopinpoint the criticalsite of alkylation and the real cause of anticanceractivity.â€• Assuming that there is a â€œcriticalsite ofalkylation,â€• which is debatable, one wonderswhether knowing it would help in designing betterchemotherapeutic agents. We now have drugswhich will attack almost every cellular site, andit is difficult to envisage a new type of alkylatingagent which has not yet been tried.

More research on the effects of combinations ofdrugs on tumors might be worth whileâ€”therewould seem to be endless possibilities to exploitand the recent work of Dean and Alexander (f27)demonstrating the potentiating effect of iodoacetamide on the destructive effect of radiation oncultured lymphoma cells and bacteria is encouraging. In this connection, Roberts and Warwick' observed the effect of pretreatment of rats with alarge dose of ethyl methanesulfonate prior to injection with Myleran on the peripheral blood count.It was hoped that ethyl methanesulfonate, whileinactive itself against granulocytes, might potentiate the effects of Mylenan, since it is itself a thiolreactor. Preliminary results have proved disappointing, but in some respects rats are not idealtest subjects from the point of view of followingminor variations in peripheral blood count due tolarge background fluctuations, and the experimentwill probably be repeated in some other way.Bearing in mind the chemical, as well as the biological, properties of the various agents, someinteresting combinations which have not yet beentried might come to light.

At this Institute further attempts are beingmade to exploit differences between normal andtumor tissue in several ways which have been doscnibed at length by Ross (7@2),and the attemptsbeing made by Connors and Elson (@25)to lowergeneral toxicity while maintaining antitumor activity by prior treatment with certain compoundssuch as thiols, etc., is worthy of comment. Some ofthese studies assume different pH levels betweentumor tissue and normal tissue, and results are

2 Unpublished data.



WARWICKâ€”Mechanism of Action of Alkylating Agents 1331

awaited with interest. Studies on the mechanismsof resistance are urgently required, but this subject is to be covered by Dr. Wheeler.

Some fresh approaches are needed to lift cancerchemotherapy from the somewhat unsatisfactorysituation in which it finds itself, but most of all isneeded a really critical assessment from all pointsof view of the situation as it stands at the moment.

REFERENCES

1. ALEXANDER,P. The Reaction of Carcinogens with Macromolecules. Adv. Cancer Res., 2: 1â€”72,1954.

2. . Comparisonof the ChangesProducedby IonizingRadiations and by the Alkylating Agents: Evidence for aSimilar Mechanism at the Molecular Level. Ann. N.Y.Acad. Sci., 68:1225â€”1337, 1958.

3. . Comparison of the Action of Ionizing Radiationsand Biological Alkylating Agents (Nitrogen Mustards) atthe Cellular and Sub-Cellular Level. Sci. Rep. Inst. Super.Sanita, 1:537â€”50,1961.

4. ALEXANDER,P. ; COUSENS,S. F. ; and STACEY,K. A. DrugResistance in Microorganisms. Ciba Foundation Symposium, p. 294. London: Churchill, 1957.

5. ANSON, M. L., and STANLEY, W. M. Some Effects ofIodine and Other Reagents on the Structure and Activityof Tobacco Mosaic Virus. J. Gen. Physiol., 24:679-90,1941.

6. ARCOS, J. C., and ARCOS,M. Fine Structural Alterations inCell Particles during Chemical Carcinogenesis. 1. The Influence of Feeding of Aminoazo Dyes on the Swelling andSolubilization of Rat-Liver Microsomes. Biochim. Biophys. Acta, 28:9â€”20, 1958.

7. . Molecular Geometry and Mechanism of Action ofChemical Carcinogens. Fortsch. Arzneimittelforschung,4:407â€”581,1962.

8. ARCOS, J. C. ; GRIFFITH, G. W. ; and CUNNINGHAM, R. W.Studies on the Swelling of Rat Liver Mitochondria inRelation to Tumour Incidence during Feeding of AminoazoDyes. Experientia, 15:316â€”17, 1959.

9. . Fine Structural Alterations in Cell Particles duringChemical Carcinogenesis. J. Biophys. Biochem. Cytol.,7:49-60, 1960.

10. BARRON, E. S. G. Thiol Groups of Biological Importance.Adv. Enzymol., 11:201â€”66, 1951.

11. BATEMAN, A. J., and JACKSON,H. Mutagenic Action ofTEM. Ann. Rep. Brit. Emp. Cancer Campaign, p. 473,1958.

12. BERENBAUM, M. C. The Effect of Cytotoxic Agents on theProduction of Antibody to T.A.B. Vaccine in the Mouse.Biochem. Pharmacol., 11:29â€”44, 1962.

13. BERGEL,F. and STOCK,J. A. Cytotoxic a-Amino Acids andPeptides. Ann. Rep. Brit. Emp. Cancer Campaign, p. 6,1953.

14. BERGEL, F.; STocic, J. A.; and WADE, R. Peptides andMacromolecules as Carriers of Cytotoxic Groups. In:Biological Approaches to Cancer Chemotherapy, pp. 125â€”34, London: Academic Press, 1961.

15. BIESELE, J. J. Induced Imitations of Mitotic Anomalies.Ann. N.Y. Acad. Sci., 71: 1054â€”67,1958.

16. . Studies on Mitosis in Purine Treated Tissue Cultures. In: S. C. PALAY (ed.), Frontiers in Cytology, pp.84â€”112. New Haven: Yale Univ. Press, 1958.

17. . Mitotic Poisons and the Cancer Problem. NewYork: ELsevier, 1958.

18. BIESELE, J. J.; PHILIPS, F. S.; TrnEasdu, J. B.; BmtcHENAL, J. H. ; BUCKLEY, S. M.; and STOCK, C. C. Chromo

some Alteration and Tumour Inhibition by NitrogenMustards: The Hypothesis of Cross-Linking Alkylation.Nature, 166:1112â€”13,1950.

19. BocK, M., and JACKSON, H. The Action of TEM on theFertility of Male Rats. Brit. J. Pharmacol., 12: 1â€”7,1957.

20. BRANCH, G. E. K., and C@vzx, M. The Theory of OrganicChemistry, p. 439. New York: Prentice-Hall, Inc., 1947.

21. BROOKES,P., and LAWLET, P. D. The Reaction of Monoand Di-Functional Alkylating Agents with Nucleic Acids.Biochem. J., 80:496â€”503, 1961.

22. . The Alkylationof Guanosineand GuanylicAcid.J. Chem. Soc., pp. 3923â€”28,1961.

23. . Ann. Rep. Brit. Empire Cancer Campaign, pp.8â€”13,1960.

24. . Ionization of DNA Bases or Base Analogues as aPossible Explanation of Mutagenesis, with Special Reference to 5-Bromodeoxyuridine. J. Mol. BioL, 4:216-19,1962.

25. CONNORS,T. A., and ETaON, L. A. Reduction of theToxicity of Alkylating Agents in Rats by Thiol Pretreatment. Biochem. Pharmacol., 11:722â€”32, 1962.

26. CRAIG, A. W.; Fox, B. W.; and JACKSON, H. Sensitivity ofthe Spermatogenic Process in the Rat to RadiomimeticDrugs and X-Rays. Nature, 181:353â€”54, 1958.

27. DEAN, C. J., and ALEXANDER, P. Sensitization of RadioResistant Bacteria to X-Rays by Iodoacetamide. Nature,196:1324â€”26, 1963.

28. DUSTIN, P. Some New Aspects of Mitotic Poisoning.Nature,159:794-97,1947.

29. ELSON, L. A. Comparison of the Physiological Response toRadiation and Radioniimetic Chemicals. Patterns ofBlood Response. In: G. C. DE HEVESY (ed.), Adv. inRadiobiology, pp. 372â€”75. Edinburgh: Oliver & Boyd,1957.

30. . Hematological Effects of the Alkylating Agents.Ann. N.Y. Acad. Sci., 68:826â€”33, 1958.

31. . The Effects of Myleran and Homologous Compounds on the Blood. Biochem. Pharmacol., 1:39â€”47, 1958.

32. ELSON, L. A.; GALTON,D. A. G.; and TIu@, M. The Actionof Chlorambucil and Busuiphan (Myleran) on the Haemopoietic Organs of the Rat. Brit. J. Haematol., 4:355â€”74,1958.

33. FARMY,0. G., and FAHMY,M. J. Cytogenetic Analysis ofthe Action of Carcinogens and Tumour Inhibitors inDrosophila Melanogaster. IX. The Cell-Stage Response ofthe Male Germ Line to the Mesyloxy Esters. J. Genetics,46:361â€”72,1961.

34. GOLDECK,H., and HAGENAR, H. Der Stickstoff host Emfuss auf die Fertilitat und Spermiogenese der Laboratoriumsratte. Z. gus. exp. Med., 117:67â€”80,1951.

35. HADDOW, A. The Chemical and Genetic Mechanisms ofCarcinogenesis. II. Biologic Ailcylating Agents. In: F.HOMBURGER (ed.), The Physiopathology of Cancer, pp.602â€”85,2d ed. New York: Hoeber, 1959.

36. HEIDELBERGER,C. ; and DAVENPORT,G. R. Local Functional Components of Carcinogenesis. Acta Internat. Uniocontra Cancrum, 17:55â€”62, 1961.

37. HUGHES, E. D.; INGOLD, C. K.; and SHAPIRO, U. G.Mechanism of Substitution at a Saturated Carbon Atom.Part VI. Hydrolysis of Iso-Propyl Bromide. J. Chem. Soc.,pp. 225â€”36,1936.

38. Hu@r, J. A., and INGRAM, V. M. The Chemical Effects ofGene Mutations in Some Abnormal Human Haemoglobins.Symposium on Protein Structure, pp. 148â€”51. London:Methuen & Co., Ltd., 1958.

39. INGOLD, C. K. Structure and Mechanism of OrganicChemistry, p. 335. Ithaca: Cornell Univ. Press; and London: Messrs. G. Bell & Sons, Ltd., 1953.



133@ Cancer Research Vol. @23,September 1963

40. ISRAEL8, L. G. ; RITZMANN, S. Laboratory and ClinicalStudies on a New Azo-Mustard, 4-Di-2â€•-chloroethylamino-2-methyl-2'-carboxyazobenzene(CB1414). Acta Un.mt.Cancer.16:665â€”71,1960.

41. JACKSON, H. Side Effects of Some Tumour InhibitingDrugs. J. Soc. Chem. md., p. 949, 1957.

42. . The Effects of Radiomimetic Chemicals on theFertility of Male Rats. J. Fac. Radiol., 9:217â€”20, 1958.

43. . AntifertilitySubstances.Pharm.Rev., 11:135â€”72,1959.

44. JACKSON,H.; Ciwo, A.W.;and Fox,B. W.ComparativeEffects of Certain Alkylating Agents on the Proliferation ofNormal and Tumour Tissues. 1. Spermatogenesis andTumour Growth. Acta Unio. Internat. contra Cancrum,16:673â€”75,1960.

45. Kou@zm, P. C. Comparative Effects of Alkylating Agentson Cellular Morphology. Ann. N.Y. Acad. Sci., 68:783â€”801,1958.

46. . Comparison of the Biological Effects of X-Raysand Radiomimetic Chemical Agents. Proc. Third Australasian Conference on Radiobiology. Radiobiol, pp. 218â€”86, 1960.

47. LAJTHA, L. G. Bone Marrow Cell Metabolism. Physiol.Rev., 37:50-05, 1957.

48. LANDING, B. H.; Goum@, A.; and Nos, H. A. TesticularLesions in Mice Following Parenteral Administration ofNitrogen Mustards. Cancer, 2: 1075â€”82,1949.

49. LAwLET, P. D. The Hydrolysis of Methylated Deoxyguanylic Acid at pH 7 To Yield 7-Methylguanine. Proc.Chem. Soc., pp. 290â€”91,1957.

50. LAWLEY, P. D., and WAu@icu, C. A. The Action ofAlkylating Agents on Deoxyribonucleic Acid and GuanylicAcid. Chem. & lad., p. 633, 1957.

51. Ls'rr, J. T.; PARKINS,G. M.; and ALEXANDER,P. Physicochemical Changes Produced in DNA after Alkylation.Arch. Biochem. Biophys., 97:80â€”93, 1962.

52. LoVELESS,A. QualitativeAspectsof the ChemistryandBiology of Radiomimetic (Mutagenic) Substances. Nature,167:338â€”42,1951.

53. . The Influence of Radiomimetic Substances on Deoxyribonucleic Acid Synthesis Studied in E8cherichia coliPhage Systems. III. Mutation of T2 Bacteriophage as aConsequence of Alkylation in Vitro: The Uniqueness ofEthylation. Proc. Roy. Soc., s.B., 150:497â€”508, 1959.

54. . The Involvement of Deoxyribonucleic Acids inChemical Mutagenesis. Biochem. Pharmacol., 4:29â€”33,1960.

55. LOVELESS,A.; GOLDACRE,R. J.; andRoss, W. C. J. Modeof Production of Chromosome Abnormalities by the Nitrogen Mustards: the Possible Role of Cross-Linking. Nature,163:667â€”69,1949.

56. LovELEss, A., and REVEL, S. New Evidence on the Modeof Action of Mitotic Poisons. Nature, 184:938â€”44, 1949.

57. LovELEss, A., and Ross, W. C. J. Note Following Biesele,J. J. St at. Chromosome Alteration and Tumour Inhibitionby Nitrogen Mustards: the Hypothesis of Cross-LinkingAlkylation. Nature, 168: 1113â€”14,1950.

58. M@zi.&, D. Biochemistry of the Dividing CelL Ann. Rev.Biochem., 30:669â€”88, 1961.

59. Mxcn.&zus, L., and SCHUBERT, M. P. The Reaction oflodoacetic Acid on Mercaptans and Amines. J. BioLChem., 106:331â€”41, 1934.

60. Mni@ica,E. C.,and Mu@su,J. A.BiochemicalInvestigations on Hepatic Carcinogenesis. J. Nail. Cancer Inst., 15:1571â€”90,1955.

61. Nou@m@,C. R., and DnisMous, R. The Reaction of AlkylBromides with Pyridine. J. Am. Chem. Soc., 54:1025-84,1932.

62. OAKBERG, E. F. Sensitivity and Time of Degeneration ofSpermatogenic Cells Irradiated in Various Stages ofMaturation in the Mouse. Rad. Res., 2:369â€”91, 1955.

63. OGSTON, A. G. The Replacement Reactions of $-@â€˜-Dichiorodiethyl Suiphide and of Some Analogues in Aqueous Solution: The Isolation of $-chloro-$'-hydroxyDiethyl Sulphide. Trans. Faraday Soc., 44:45â€”50, 1948.

64. PRICE, J. M.; Mn@Erc, E. C.; MILLER, J. A.; and WEBER,G. M. Studies on the Intracellular Composition of Liversfrom Rats Fed Various Aminoazo Dyes. II. 3'-Methyl,2'-Methyl and 2-Methyl-4-dimethylaminoazobenzene; 3-Methyl-4-monomethylaminoazobenzene and 4'-Fluoro-4-dimethylaminoazobenzene. Cancer Res., 10: 18â€”27,1950.

65. ROBERTS, J. J., and WARWICK, G. P. Studies of the Modeof Action of Tumour Growth Inhibiting Alkylating Agents.I. The Fate of Ethyl Methanesulphonate (â€œHalfMyleranâ€•) in the Rat. Biochem. Pharmacol., 1:60â€”75, 1958.

66. . Studies of the Mode of Action of Tumor GrowthInhibiting Alkylating Agents. III. The formation of Tetrahydrothiophene-1,1-dioxide from 1,4-Dimethanesulphonyloxybutane (Myleran), S-@9-L-Alanyltetrahydrothiopheniummesylate, Tetrahydrothiophene, and Tetrahydrothiophene-1,1-dioxide, in the Rat, Rabbit and Mouse.Biochem. Pharmacol., 6:217â€”27, 1961.

67. . The Reaction of Some Compounds of BiologicalInterest with Nucleic Acids. Biochem. J., 84: 119 r@,1962.

68. . Azo-Dye Liver Carcinogenesis: Reaction of Hydroxymethylaminoazobenzene with Nucleic acids in Vitro.Nature, 197:87â€”88, 1963.

69. . The Mode of Action of Alkylating Agents. II.Studies of the Metabolism of Myleran. The Reaction ofMyleran with Some Naturally Occurring Thiols in Vitro.Biochem. Pharmacol., 6:205â€”16, 1961.

70. . Studies of the Mode of Action of Tumour GrowthInhibiting Alkylating Agents. IV. In Vitro Reactions of2-Chloroethylarylamines. Biochem. Pharmacol. (in press).

71. . Studies of the Mode of Action of Tumour GrowthInhibiting Alkylating Agents. VI. The Metabolism of Bis@9-Chloroethyl Sulphide (Mustard Gas) and Related Cornpounds. Biochem. Pharmacol. (in press).

72. Ross, W. C. J. Biological Alkylating Agents. London:Butterworths, 1962.

73. . The Chemistry of Cytotoxic Alkylating Agents.Adv. Cancer Res., 1:397â€”449, 1953.

74. . in Vitro Reactions of Biological Alkylating Agents.Ann. N.Y. Acad. Sci., 68:669â€”81, 1958.

75. . Biological Alkylating Agents, p. 25. London:Butterworths, 1962.

76. Ross, W. C. J., and WARWICK, G. P. AryI-2-halogenoalkylamines. Part XVIII. The Rates of Reduction of Substituted 4-Di-(2-chloroethyl)arninoazobenzenes by Stannous Chloride, Hydrazine, and the Xanthine OxidaseXanthine System. J. Chern. Soc., pp. 1724â€”32,1956.

77. RUTMAN, R. J. ; STEELE, W. J. ; and PRICE, C. C. Observations on the in Vise Alkylation of Ehrlich Tumour CellDNA by Nitrogen Mustards. Biochem. Biophys. Res.Comm., 4:278â€”82, 1961.

78. SCHNEIDER, W. C. Intracellular Distribution of SucciicDehydrogenase, Cytochrome Oxidase, Adenosinetriphosphatase and Phosphorus Compounds in Normal Rat Liverand in Rat Hepatornas. Cancer Res., 6:685, 1946.

79. SHAVER, S. L. X-Irradiation Injury and Repair in theGerminal Epitheium of Male Rats. I. Injury and Repairin Adult Rats. Am. J. Anat., 92:391-432, 1953.

80. SKIPPER, H. E. Discussion of the Biological Effects ofAilcylating Agents. Ann. N.Y. Acad. Sci., 68:808, 1958.

81. SMITHERS, D. W. Cancer: An Attack on Cytologism.Lancet, pp. 493â€”99,1962.



W@wic@xâ€”Mechanism of Action of Alkylating Agents 1333

82. STACEY, K. A.; COBB, M.; Cousm@s, S. F.; and ALzw@DEE, P. The Reactions of the â€œRadiomimeticâ€• AlkylatingAgents with Macromolecules in Vitro. Ann. N.Y. Acad.Sci.,68:682â€”701,1958.

83. ST@urn,G. R.; STm1@@,W. H.; and MooRz, S. Relations between the Conformation of Ribonuclease and Its Reactivity toward Iodoacetate. J. Biol. Chem., 236:436â€”42, 1961.

84. SWAIN,C. C. Abstr.13thNail. Org.Chem.Symposium,Am. Chem. Soc., p. 65, 1953.

85. Tmssxs, G. M. The Action of Antimetabolites and Biological A&ylating Agents on the Synthesis of Deoxyribonucleic Acid, and a Possible Relation between the Mechanisms of Action. Biochem. Pharmacol., 4:49â€”56, 1960.

86. W@tuwicx, G. P. Relative Rates of Reduction of a Series ofAzo Compounds. J. Soc. DyersColourists, 75:291â€”99, 1959.

87. WHEmms, G. P. Studies Related to the Mechanisms ofAction of Cytotoxic Alkylating Agent@ Cancer Res., 22:651â€”88,1962.

88. WHEELER, G. P.; Moimow, J. S.; and S@xppm@,H. E.Studies with Mustards. L Ion Exchange Chromatographyof Reaction Mixtures of Purines and Pyrimidines withNitrogen Mustard. Arch. Biochem. Biophys., 57:124â€”32,1955.

89. . Studies with Mustards. II. Paper Chromatography and Radioautography of Reaction Mixtures ofVarious Compounds with S's-Sulphur Mustard. ibid.,pp. 133â€”39.

90. YOUNG,E. G., and [email protected]@,IL B. The Reactionoffl-fl'-Dichlorodiethyl Sulphide with Proteins. Canad. J.Res., 25B:37â€”48, 1947.



1963;23:1315-1333. Cancer Res G. P. Warwick The Mechanism of Action of Alkylating Agents

Updated version

http://cancerres.aacrjournals.org/content/23/8_Part_1/1315

Access the most recent version of this article at:

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

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/23/8_Part_1/1315To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



the mechanism of action of alkylating agents...substancef (competi.. factor)phthiosulfate tion...

Documents