POTENTIAL FOLIC ACIDABTAGONISTS. I11
May 1968
583
Potential Folic Acid Antagonists. 111. The Antitumor and Folate Reductase Inhibitory Properties of Some Irreversibly Acting 6-N-Substituted 2,4,6-Triamino-5-(4-carbethoxyphenylazo)pyrim~dines J. HAMPSHIRE, P. HEBBORN, A. 11. TRIGGLE, AND D. J. TRIGGLE Depaitnaent of Biochemical Pharmacology, School of Pharmacy, and the Center for Theoretical Biology, State L'niziersity o j ,Vew York at Buffalo, Buffalo, &YewYork lSWl4 Received Sovember 24, 1967 The synthesis of some irreversibly acting inhibitors of folic acid reductase is described. These compounds, the 6-Nw-(N-eth~1-N-2-chloroethyl)alkyl-2,4,6-triamino-5-(4-carbethoxyphenylazo)pyrimidines(XVI-XX), were tested for antitumor activity against the Ehrlich ascites system. Their ability to reduce the in vivo levels of folic acid reductase in mouse tissues was also determined. Only a partial correlation between the in vitro effectiveness of these compoiinds as folate reductase inhibitors and their ability to reduce folate reductase levels in vivo in mouse tissues and to inhibit the growth of the Ehrlich ascites tumor was discovered. Some reasons for this discrepancy are discussed. The possible mechanism of action of these compounds a t the enzyme level is discussed in terms of Baker's extensive work on pyrimidine binding to folic and dihydrofolic acid reductase.
In continuation of our study of 5-arylazopyrimidines as inhibitors of folic acid reductase1v2me have investigated the effect of incorporating alkylating functions at the 6 position of 2,4,6-triamino-5-(4-carbethoxypheny1azo)pyrimidine. The alkylating function selected was the 2-chloroethylamino group attached to the 6 substituent of the parent pyrimidine by a polymethylene chain. In a preliminary report 011 these compounds3 we noted that the extent of inactivation of folic acid reductase by the B-S-w-(N-ethylS-2 - chloroethyl) alkyl- 2,4,6-triamino-S- (4carbethoxypheny1azo)pyrimidines showed some dependence on the chain length of the 6 substituent and that the relative irieff ectiveness of S-2-chloroethyl-N-ethyl-n-butylamine (which represents the alkylating side chain of XVIII) was not increased in the presence of 2,4,6-triamino-S-(4-carbethoxyphenylazo)pyrimidine. These results suggested that the irreversible inactivation of the enzyme is not caused by a random alkylation of a nucleophilic Eite but is due to the initial formation of a reversible enzyme-inhibitor complex and subsequent alkylation, a process which Baker4 has termed activesite-directed irreversible inhibition. In this paper we report the synthesis and a more detailed analysis of the biological activities of the B-K-w-(N-ethyl-N-2-chloroethyl) alkyl-2,4,6-triamino-5(4-carbethoxypheny1azo)pyrimidines.
Experimental Section Synthetic Procedures.5-The physical constants of compounds used in this study are presented in Tables I and 11. N-Ethyl-N-2-hydroxyethyl-1,w-alkanediamines (VI-X, Table I).-N-Ethyl-2-hydroxyethylamine (0.3 mole) and N-w-bromoalkylphthalimide (0.23 mole) dissolved in C B H (200 ~ ml) were refluxed for 12 hr. The benzene solution of N-w-(N-ethyl-N-2hydroxyethy1)aminoalkylphthalimide was decanted from the oily layer of N-ethyl-2-hydroxyethylamine hydrobromide and (1) J. Hampshire, P. Hebborn, A. &I. Triggle, a n d D. J. Triggle, J . M e d . Chem.,B, 745 (1965). (2) 1M. Chadwick, J. Hampshire, P. Hebborn, A. AI. Triggle. a n d D. J. Triggle, ibid., 9, 874 (1966). (3) J. Hampshire, P. Hebborn, A. hI. Triggle, a n d D. J. Triggle, J. Pharm. S c i . . 66, 453 (1966). (4) B. R. Baker, i b i d . , 63,347 (1964). ( 5 ) hlelting points are corrected and \!-ere determined on a ThomasKofler hot stage. Analyses were performed by Dr. .A. E. Bernhardt, Malheim, West Germany. Where analyses are indicated only b y symbols of t h e elements or functions, analytical results obtained for those elements or functions were within 0.4% of t h e theoretical values.
evaporated in vacuo to an uricrystallizable oil a small portion of which was converted to the methiodide (I-V, Table I ) for identification purposes. The oily N-w-(N-ethyl-K-2-hydroxyethyl)aminoalkylphthalimide was dissolved in concentrated HCl (300 ml, d 1.42) and refluxed for 12 hr. H,O (100 ml) was added to assist the precipitation of phthalic acid and the cooled mixture was filtered. The filtrate was evaporated under reduced pressure to give the oily N-ethyl-r\T-2-hydroxyethyl-l,w-alkanediamine dihydrochlorides. These were apparently very hygroscopic and could not be maintained in a crystalline state but were converted to the free bases by addition of the Calculated amount of ethanolic sodium ethoxide. Filtration of XaC1 and fractional distillation of the filtrate afforded VI-X (Table I ) which were further characterized as dipicrates (VIa-Xa, Table I). 6-N-w-(N-Ethyl-N-2-hydroxyethyl)alkyl-2,4,6-triamino-5-( 4carbethoxypheny1azo)pyrimidines (XI-XV, Table 11) were prepared from 2,4-diamino-6-chloro-3-(4-carbethoxyphenylazo)pyrimidine (3.2 g, 0.01 mole) and N-ethyl-K-2-hydroxyethyl-la-alkanediamine (0.025 mole) in EtOH (73 ml) heated to 100" for 13 hr. On cooling XI-XV crystallized as orange plates in 80-937, yields and were recrystallized to purity. 6-N-w-(N-Ethyl-N-2-chloroethyl)aIkyl-2,4,6-triamino-5-( 4carbethoxypheny1azo)pyrimidine hydrochlorides (XVI-XX, Table 11) were prepared from finely powdered 6-N-w-( N-ethyl-X4-carbethoxypheny1azo)2 - hydroxyethyl)alkyl-2,4,6-triamino-5-( pyrimidine (0.005 mole) suspended in anhydrous ether a t 0" to which SOCl, (0.9 g, 0.0075 mole) was added with vigorous stirring; the mixture was maintained a t 33' for 1.5 hr. Excess SOCl, was destroyed by the addition of EtOH and the mixture was filtered, washed (Et,O), and recrystallized from 90% EtOH to give XVI-XX as the hydrated hydrochlorides in near quantitative yields. Prolonged drying a t 100" (0.1 mm) did not remove the water of hydration. Biological Test Methods.-Toxicity determinations were carried out using random-bred male Swiss mice (20-25 g). The compound dissolved in 10% gum acacia was administered by intraperitonal injection to groups of three to six mice per dose level. Deaths within a 21-day period were recorded. LDbo values were calculated using the method of Litchfield and Wilcoxon.6 Antitnmor activities were determined with the Ehrlich ascites system. Ehrlich ascites cells were taken from mice bearing a 10-11-day-old tumor. The cells were washed with cold sterile saline (0.85%) and mice were inoculated with 106 cells. Compounds dissolved in saline or 10y0 gum acacia were injected intraperitoneally on the day after inoculation once only, or once daily for 5 days. The mice were killed on the 10th day after inoculation and packed-cell volumes were determined. IDso values were calculated according to the method of Litchfield and Wilcoxon.6 Folic Acid Reductase Assay.-Folic acid reductase activity was determined as previously d e s ~ r i b e d with ~ ~ * the ~ ~ following modifications: in the assay of intestinal enzyme MgC1, was replaced (6) J. T. Litchfield and Y9 (1949).
(7)
F. Wilcoxon, J . Pharmacol. Ezptl. Therap., 96,
W.C. Werkheiser. J . Biol. Chem., 236, 888
(1961).
\'ol. I 1
\l,LIl> h l . 5
1, s c, 11, I, s ( > ,11,
CI, I T , 1, s c, 11, I, s
\nalyse,
No.
c, IT, N C, TI, N c, €1,N c, TI, N r, IT, N c, 11, c1, s (-,11, c1, s
.xI 4111 SI11
SIV XV x\1 XVII SVIII NIX
C', IT, C1, ?;
c, 11, c1, N c:, IT, c1, N
ss
by 2 G.lf lliiClp, mid glucose 6-phosphitte (0,s m J I ) and gliicow ti-phosphate dehydrogenase (6.4 iiiiit s / m l ) were added to provide an N;ADPII regenerating system. Sicotiiiamide (,XI Mmoles) wah present to iiihibit NA41)Pase activity. 111 tlie ashay of folic acid a fluin Ehrlich ascites tiiiiiilr JIgC12 w a s replaced by 30 f KCl a i d the NAI)PII regeueratirig h o v e was present. Preparation of Folic Acid Reductase.--Female inice (ICRjtIii, 40 g) were iiiociilated intraperitoiieally with Ehrlich hscites cell;: ( 106)). Six days after inoculation the animals were killed and tumor, liver, and iiitestiiie were removed. Tissues from amiiials were pooled. The livers were cleaiied aiid homogenized it] ice-cold 0 . 2 3 JI siicrose coiitaining 0.001 JI EDTA and the homogenate w&s ceiitrifuged at 104,000y for 40 min. The auperiiataiit fractioii was collected aiid stored :it -20" uiitil used. Iiitestind eiizynie wad prepared similarly from tlie i i i testitial iii~icosn. 111 the preparatioii of the tumor eiizyme the i IiIiior c e h were washed with ice-cold 0.13 J I sahie, suapeiided i i i distilled water for 30 sec t o lyse red blood cells, mixed with 2 vu1 of 0.6 JI saline to restore isotoiiicity, and ceiitrifuged a t .iOOOy for 10 niiri. The cell pack pias washed twice with 0.15 JI saliiie :trid the saspensio~i vas hc~moge~iized f o r 2 min a t 45,000 rpni i t i :i \.irt is homogeirizrr. The Iioniogeiiate was centrifiiged at 1 0 X , O 0 0 ~for. 40 mill aiid the siiperriat:tiit was removed and stored :it -20". l'roieiti was determilieti by the method of Lowry.* Determination of Folic Acid Reductase Levels.---The levels of folic :wid reductase iii the 11101.ise liver, iiit estiiie, a ~ i dtuiii(rr 8)
L.. I , L I ~ I I .~i. i d / i ~ X f ir~ t a ~ , m d . 5, , P i 1 1 ~ 6 2 ) .
hoiiioge~iarespreparcct as described itbovc were deterniiiied by I itrntioti with riiethotresnte.~ For the liver arid intestine the
;is.ay coiidit ioiih were thohe de5cribed by \Verkheiser.' For 1,he IF t i i n i o i ~eiizyriie lIgC12 was replac y of folic ;icid reductase levels iii ihese rnini~tratioii of XV &iy after iiiocuht iori with tumor cells, e t i s w were reinovcd I
Determination of Comparative Alkylating Activities.-Thct relative chemical reactivities of compounds were determined iisiiig a model niicleophile, ~-riitrobeiizylp~.ridine, by the met hod of lkirdos niid roworkersln :is modified by Baker." The conipoiiiids were dissolved iii 2-methoxyethanol at a concentration of 0.2 mJI. ,Iliquots of solutioti (0.25 ml) were pipetted into a 11 of a w/v soliition of 4-nitroaeries of test iiit)eh i u i d 0. I~eiuylpyridineiii 2-nietho h a d , 0.23 ml of 2-methoxyethaiiol, [im pht>halatebuffer (pFI 4.2) were :iiid 0.2.5 nil of 0.03 .\I po :tdded t o each tube. Blaiik tubes were identical but did 1101 contain the 2-haloge~ioethylariiiiies. The tubes were m:tiiitairicd ttt 37", removed at timed intervals, and placed on ice. Theii, 0.1 nil of 0.2 ?I KO11 (XOC, aqueous 2-methoxyethanol) was added atrd tlic c~iii~eiits o f the t,itbe was vigorously distiirbed ~.
...
(!I1 .J. It. I3ertin75 lOO(63160) 75(54104) >l50
13 (8-22) 7 , 6 (5-12) 6 . 8 (4-13) 13.0(5-27) 70(30-190) < 100
>75 72 (33-155)
Therapeutic index'
1.8
5.7 4.4 4.0 75
150
...
55 (47-54) 13 (7-23)
50 (37-78) 1.1 0 . 9 (0.713.3 1.3) a In parentheses, confidence limits, P = 0.05. Tumor (106 cells) inoculated on day 0 and compound admiiiistered intraperitoneally on days 1-5. LD:o/ID90.
bethoxypheny1azo)pyrimidines (XI-XV) and their chlorinated counterparts the %S-U-(N-ethyl-S- 2-chloroethyl) alkyl - 2,4,6 - triamino - 5- (4-carbet hoxyphenylazo)pyrimidines (XVI-XX) . The nonchlorinated com-
SVI-XX
pounds XI-XV had insigriificarit antitumor activities (therapeutic indices me, :tiit1 ilia1 muro.al enzymc levels. c:iti
Discussion
c x p r e s d a s percentage of coiitrol values, in mouse ti sue^ following udmiriistratiori of' the coinpourids xt (lore 1evc.l~ that hraclceted the EDyo's determiiietl i i i acute c.xperimeiit s (Table J-). These data iiidicate tlial
t"rom t lie re-ult< prt~eiitecli n tlii- paper it iy appareii(, t1i:it :ittempt. to correlate the effecti~eiieb~ of com1 7 ) z itro inhibil or- of folic acid reducta5e n it h their iihility to inhibit the growth of a tumor L iuo :$re buhject to a iiumher of difficulties. The (lata in T:ible 111 iridicate :t broad mal;imuni 111 the therapeutic i~idicebof the alkylating compoutitlh XVI -XX, with X171 being the most effective :tgciit iuitl X \ X l iriitl XIX hcing .omewhat 1ebs effective. Tlicrc i b otd) :I partiiil correlation between these reiiilts :iii(l those from thc effectiveiiesr of t h c conipouiitlh :tb 1 ) ~L ~ ~ Ji~iliibitor~h O of folic acid reductaie. These latter result. (Table IT) ihow that XIX is tlie mo-t eff'ective in iwersihl>r iiiliibitirig the enzyme from the mouse liver. intestine, and ascites tissues. However, I\ hen these coiiipouiidi ~ w r ecompared at COKlCellt i*atiori. producing e b w i t i d l > the bame conceritralioii of reversiblj bound eriz) me-inhibitor coinplex, their XIX arid XTXI were approximately equally eff ectivc iri produciiig .iOC/;, irrever.ible inactivation of the liver erizj me, but XVIII was rat lier more efficient iri producing the same degree of in:tctivatiori of the irit estirinl iiiucosd arid ascit e i e i i q me-. When the abilit ies of XVI-XX to reduce the enzyitirl lc~vcl~ iii iIieie ti\sue. 2 1 1 ~compared (Table 1'11) it 1' apprimit t h a t X\?III :itid S I X were the most effectivt.
N a y 1968
I’OTEKTIAL FOLICACIDASTAGONISTS. I11
in producing a general reduction in the enzyme levels of all three tissues. These findings accord with the i n vitro studies (Table VI) on the folic acid reductases from these three tissues which showed that X V I I I and XIX were the most effective in producing irreversible inactivation. h number of factors probably contribute to the lack of a complete correlation between the antitumor and folic reductase inhibitory effects of XT’I-XX. I t seems improbable that diff erences in chemical reactivity alone are sufficient to provide an explanation. In the model system with 4-nitrobenzylpyridine (Table IT), X U - X I X showed essentially similar reactivities, but XX was distinctly less reactive. This difference is not extended, however, to the folic reductase systems where the t8iniesfor 50% inactivation of the enzyme by XT’III-XX were comparable. A possible explanation of these findings is that the 6 substituent of XX is sufficiently flexible to offer hindrance (by coiling, partial micelle formation, or internal bonding) to its reaction with 4-nitrobenzylpyridine; however, on the enzyme surface the intermolecular forces exerted on the side chain may cause its extension and realization of its full alkylating potential. I n connection with the role of relative chemical reactivities it is relevant to draw attention to the recent work of Schaeffer and his co-workers15 who have reemphasized that the rate of inactivation of an enzyme by an irreversible inhibitor within the initially formed reversible enzyme-inhibitor complex is dependent both on K i arid kz. They point out that the study of irreversible inhibitors at only one concentration can lead to erroneous conclusions about their relative effectiveness in producing irreversible inactivation, and inversions in this order can occur if different concentrations are employed. I n our study of the in. vivo effectiveness of XPI-XX in reducing folic reductase levels, it is not known whether equivalent concentrations of the inhibitors reach all tissues or whether the cells of various tissues are equally permeable to an inhibitor molecule. The data of Table VI1 show that while XT’III and XIX are the most effective in producing a general reduction in reductase content of the three tissues, X V I is able to produce a pronounced reduction of tumor enzyme without affecting liver or intestinal enzymes. The apparently selective action of X V I may be attributable to a more favorable binding to the tumor enzyme or, and this is equally probable, to the operation of the factors discussed above. An additional factor, which may be of equal importance to those already discussed, concerns the relative folic acid levels in the tissues; high folic acid levels mill, through competition with the inhibitor, tend to reduce its effectiveness. Work is in progress to determine the relative importance of these factors. Finally, it should be noted that it is most improbable that XT’I-XX confine their actions to the inactivation of folic reductase. At lethal dose levels these compounds have marked central effects, probably because of the nature of the alkylating side chain, producing convulsions and death within 30-60 min. This indicates that their toxic effects at these dose levels are not due to inactivation of folic reductase since methotrexate produces deaths within 3-5 days, (15) H. J. Scliaefler, 68b (1967).
AI. .I. ScliuartL, a n d E. Odin,
J . M e d . Chem., 10,
587
as a consequence of interference with folate metabolism. It is of interest to compare our data for the inhibition of folic reductase by the 5-arylazopyrimidines reported in this paper with the extensive data of Baker and his co-workers on the binding of pyrimidine inhibitors to folic and dihydrofolic reductases. Such comparisons must, however, by tempered with the recognition that they are being made between enzymes from different sources and that some of the differences observed may be due to species variation of the enzyme. Baker has synthesized 17,18 irreversible inhibitors of the 2-aniino-ipyriniidinol class (XXI) which are believed to bind in (CH
C
H
O&~C~2~nC6H4NHCOCH2Br
IN “2
XXI
the conformations depicted in which the hydrophobic interaction between the enzyme and the 5-phenylbutyl substituent determines the binding conformation of the inhibitor molecule. However, the corresponding 2,4diaminopyrimidine (XXII), while a good reversible tc H~ I4c6n5
I
“2
XXII
inhibitor, did not produce irreversible inactivation suggesting that XXII was bound in a conformation in which the alkylating group did not bridge to a nucleophilic site. The reorientation of the alkylating chain of XXII is probably due to repulsion between the 4amino substituent and the enzymelg in conformation XXII leading to its adoption of conformation X X I I a .
Y “2
XXIIa
Our investigat,ioris int’o the structure-activity relationship of the 5-arylazopyrimidines, while incomplete, suggest that t’he ;-aryl group is binding to a hydrophobic site20 and since t,here is probably only one such (16) B. R . Baker, “Design of Active-SiteDirected Irreversible Enzyme Inhibitors,” John \Tiley and Sons, Inc., New York, N. Y., 196i, p p 192-266. This book includes a convenient summary of the work of Baker and his colleagues. (17) B. R. Baker and J. H. Gordaan, J . Pharm. S c i . . 66, 1417 (1966). (18) B. R. Baker a n d H. S. Shapiro, ibid., 66, 1422 (1966). (19) B. R. Baker a n d H. S. Shapiro, ibid., 66, 314 (1966). (20) J. F. Moran, A . 21. Triggle, D. J. Tripgle. a n d A , Wayne, unpulilished data.
.itc near the active center the 2,4,6-triamino-5-arylazol)yrirnidines may be binding in conformatioil X X I I I . In this conforniation the alkylating A SUI)-
7"""
COOEt
I
CICH2CH2(Et>N(CH2),NH
,& N'
H N
NH(C H2)n N( Et )CH,C
"2
H2C I
ssI I1R
"2
SSlII
htitueiit would be able to bridge to the same general :irw on the enzyme as the corresponding substituent as i i i XXI. In the case of XXII ionic repulsion between t hc J-l-imino substituent and the enzyme was considered to force it5 adoption of conformation X X I I a : presumably the same repulsion factor would be operative for the L',~,G-trianiitiopyriiiiidiiies but since there :ire
:mino groups a t the 4 :~ndt j 1)osition. i-epulsioii will probably be :q~proximittely equivalent regardless of whether conforniation XXIII or XXIIla is xdoptcd. Such speculation affords an explanation of the irreversible iriactivatiiig properties of ti-S-w-(S-ethyl-2-chloroethyl)alky1-2,4,G-trianiino - 3 - (4 -'carbet hoxj phenyliizo) 1)yrimidinesand also of their relatively low potency. Acknowledgment.-This n-ork was supported by it grant (CA 06645) from the S:ttional Cancer In+titutc, Satioiial Iiistitutes of Health, I-. S. Public Health Service.
Peptides of Pyrimidine Amino Acids'
The a-amino-p-( 2-mel.capto-6-oso)-4-pyrinlidylpropi~)tlicacids were fouiid to be pheiiylalaiiine antagonials. A number of dipeptides of these amino acids have been synthedized and tested as inhibitors of growth and prot ein synthesis iii Ehrlich ascites carcinoma in mice. Glyryl-, or~-phenylalanyl-,and ~,-phenylalanyl-ol-:rnliiioa-(~--mercapto-6-oxo-3-methyl)-4-pyrimidylpropio1lic acid showed an enhanced :ict.ivit,y over n-:imiiio-p-( 2-mercapto-6-oxo3-methyl)-4-pyrimidylpropionic acid.
liesenrch efforts within recent years have resulted in the synthesis of many antimetabolites of pyrimidines, l)urities, amino acids, vita,niins, and other met'abolites :IS I)oteiitinl anticancer agents. One of tlie major t w k s of cancer chemotheraiiy is to search for antitumor diugs n i t h diverse and select spectra of action. In ordcr t o get better selectivity arid transport' into the cell iritcrior large numbers of antitumor drugs h:tve heeii y . t i t hcsized in which cytotoxic groups have been ititroduced into natur:il carrier^.^ ;Iniong the natural c:irriers. the amino acids and peptides have been shon-11 t o ~)I:LJ. important roles. The best example of such ii usc of :in :imino acid or a peptide is found in the work of I k ~ g e l:ind St0clq~8-c Im-ionov and S ~ p h i n a ,and ~~ Sol)hiii:t, et uL3e They have synthesized DL-, L-, and ~~-~~-di(L'-chloroet~liyl)iinii~iophenylala~~i~ie and their pep( l i 'l'tii.* inreiriration wab supported 1j.v l'ulilic Health Service Itesearcli ( : r a n t s ('.\-0636-1-0-1 and CA-0636-1-05 irorn tlie Kational Cancer Inscitilte. ( 2 ) Tu I\ liom iniluiries sliorild be sent. 4)la) F, Bergel and J. .I, Stock, J . Cheni. Soc., 2409 (1954); (b) F. I , V . (', E. Burnop, and J . .\, Stock, ibid., 1223 (19.55); ( e ) F. Bergel, ,I. \ . S t u c k , and K. \Vade, "Bii,logical .lgproaclies t o Cancer Cliernother" .\riuIeinic P r e ~ s Inc., New Y o r k , N. Y., 1961, 11 125; (11) I,.F. 1,ariiin o v and %. 1'. Sopiiiina. IJohI. . 4 1 j ~ d ..I'uuk SSSIZ, 114, 1070 (1857); ( e ) %. 1'. S
