272 Journal of Medicinal Chemistry, 1975, Vol 18, No 3:
Kawi, Mead. Urobniak Zakrzeu5ki
Synthesis and Antifolate Activity of New Diaminopyrimidines and Diaminopurines in Enzymatic and Cellular Systems Ivan Kavai, Lawrence H. Mead, Jadwiga Drobniak, and Sigmund F. Zakrzewski* Department of Experimental Therapeutics and J. T . Grace, J r , Cancer Drug Center, Rosu'ell Park Memorial Institute, Neu York State Department of Health, Buffalo, New York 14203. Received September 9, 1974 The following new 2,4-diamino-6-methylpyrimidines, 5-cyclohexylmethyl, 5-cyclohexylethyl, and 5-(2-naphthyl),as well as 2,6-diaminopurines, 8-adamantyl and 8-adamantylmethyl, were synthesized as potential antifolates. These, as well as three known compounds, 2,4-diamino-5-cyclohexyl-6-methylpyrimidine, 2,4-diamino-5-(l-naphthyl)-6-
methylpyrimidine, and 2,6-diaminopurine, were compared with respect to the inhibition of growth of mammalian cells in culture (TA 3) and with respect to the inhibition of partially purified dihydrofolate reductase. All of the pyrimidines except for the 5-(l-naphthyl) derivative were competitive inhibitors of dihydrofolate reductase, with K , values ranging from 0.07 to 0.4 fiM. They were 2-5 times better as inhibitors of the isolated dihydrofolate reductase was a noncompetitive inhibitor of the enthan of the cell growth. 2,4-Diamino-5-(l-naphthyl)-6-methylpyrimidine zyme with a K ; value of 56 fiM. This compound was more potent in inhibiting cell growth than the isolated enzyme, indicating that its biological activity was not related to the inhibition of dihydrofolate reductase. All of the purine derivatives were poor growth inhibitors and although some of them inhibited isolated dihydrofolate reductase, their mode of action in cellular system did not seem to concern folate metabolism, as judged by the inability of hypoxanthine, thymidine, and glycine to provide protection. The implication of these findings as to the structural requirements for inhibition of dihydrofolate reductase is discussed. The pitfalls of the determination of IDSOvalues instead of a complete kinetic analysis in structure-activity studies are emphasized. Earlier work from this laboratory indicated that t h e affinity of 2,4-diaminopyrimidines to t h e enzyme dihydrofolate reductase increased with t h e size of the hydrocarbon substituent a t C-5.' Furthermore, hydrocarbon substituents with a bulky and rigid conformation, such as cyclohexane or adamantane, imparted a much stronger affinity to the enzyme than straight-chain hydrocarbons of similar size. T h e exploration of the biological activity of 2,4-diaminopyrimidines having at C-5 either a combination of cyclohexane with a short straight hydrocarbon or a bulky substituent other than adamantane or cyclohexane was undertaken to learn more about t h e structural requirements for inhibition of dihydrofolate reductase. Also, since in the case of diaminopyrimidines1>'and diaminopteridines' the presence of a lipophilic group adjacent t o the heterocyclic ring increased considerably their biological activity, it was of interest to see if a similar effect would be observed in t h e case of lipophilic substituents at C-8 of 2,6-diaminopurines. T h e parent compound 2,6-diaminopurine has been reported to be a weak competitive inhibitor of dihydrofolate reductase from chicken liver4 and was shown to protect this enzyme from mouse sarcoma cells against proteolytic inactivation." A preliminary report on this work has been published.fi Syntheses. Two synthetic approaches were used for t h e preparation of diaminopyrimidines. 2,4-Diamino-5-cyclohexylmethyl-6-methylpyrimidine (2) and 2,4-diamino-5cyclohexylethyl-6-methylpyrimidine( 4 ) were synthesized by the condensation of guanidine with the appropriately substituted ethylacetoacetate to form t h e 2-amino-4-hydroxy-5-substituted 6-methylpyrimidines. These compounds were converted to t h e corresponding 2.4-diaminopyrimidines by chlorination followed by amination. 2,4Diamino-5-(2-naphthyI)-6-methylpyrimidine( 5 ) was prepared essentially as described for t h e preparation of t h e corresponding 5 - (1-naphthyl) derivative 11.7 2,5-Diamino8-adamantylpurine (8) and 2,6-diamino-8-adamantylmethylpurine (10) were synthesized as described for other 8-substituted purines8 by reacting 2,4,6-triamino-5-nitrosopyrimidine with either 1-adamantanecarboxylic acid chloride or I -adamantaneacetic acid chloride to form corresponding 4,6-adamantoylamino or adamantylacetylamino pyrimidines, respectively (Scheme I). Reduction of the nitroso group led t o closure of the purine ring and desired compounds were obtained after hydrolysis of 6-acyl groups.
Scheme I THOCR
n
>" I 1
HJ
/
8.10
7
tu
9,lO.R = -CH,
6-8,R=
M
Inhibition of Dihydrofolate Reductase. All of the five newly synthesized compounds were tested as inhibitors of partially purified dihydrofolate reductase from an a m ethopterin-resistant subline (AT/3000) of sarcoma 180 cells. T h e purification of the enzyme, the assay procedure, a n d t h e use and advantages of a transformed LineweaverBurk equation at substrate concentrations much above t h e K , value were reported and discussed previously. l o T h e effects of the newly synthesized compounds on t h e dihydrofolate reductase are presented in Figure 1. 2,4-Diamino-5(l-naphthyl)-6-methylpyrimidine(11) is a compound described previously7 but it has not been tested for inhibition of dihydrofolate reductase and therefore was included for comparison in Figure 1. Growth Inhibition of Mammalian Cells in Culture. T h e culture medium, conditions, and assay procedures used for mouse mammary adenocarcinoma cells ( T A 3) were as described elsewhere.11 In order t o determine whether the site of action of the inhibitors was related to folate metabolism, the T A 3 cells were grown under conditions in which folate metabolism was blocked, i . e . , in a medium supplemented with 1 p M amethopterin, 30 gM thymidine, 100 g M hypoxanthine,
Journal ojhfedicinal Chemistry, 1975, Vol. 18, No. 3 273
New Diaminopyrimidines and Diaminopurines
Table I ID,,/TA 3, pn'f Compd
2 4 5 11 12
8
Ring structure NH.
R
FA-meda
HTG-medb
Met hy 1cy c lohexyl Ethylcyclohexyl 2 -Naphthyl 1-Naphthyl Cyclohexyl
0.74 1.30 0.28 2.5 0.39
10 41 43
Adamantyl Methyladamantyl H
NH.
10 13
Ki (DFR),' p.2.I 0.41e 0.27e 0.07e 56 0.13e
420
48 1O d 35
4.8e 13.3e 1200e
42 50
Wells grown in the normal medium supplemented with folic acid. *Cells grown in the medium supplemented with 1 /uV amethopterin, 100 K M hypoxanthine, 30 @A4thymidine, and 100 +M glycine. CDihydrofolate reductase. dOnly 20% of growth inhibition was obtained at this concentration. ecompetitive inhibitors. INoncornpetitive inhibitors.
and 30 p M glycine (HTG-medium).I2 If compounds are inhibitory under such conditions, their mode of action must be unrelated to folate metabolism. Table I summarizes t h e results of the biological testing. For comparison with the newly synthesized compounds three known compounds were included: 2,4-diamino-5-cyclohexyl-6-methylpyrimidine (12),13 2,4-diamino-5-(l-naphthyl)-6-methylpyrimidine (11),7 and 2,6-diaminopurine (13). T h e purines included in this study were poor inhibitors of cell growth and although 8 and 10 inhibited dihydrofola t e reductase t o a certain extent, their mode of action in cellular system did not seem to concern folate metabolism. T h u s in the case of 8 and 13 t h e hypoxanthine-thymidineglycine mixture did not prevent t h e growth inhibition. In the case of 13 this was not surprising since 2,6-diaminopurine is known t o be an analog of adenine. As such it is converted to nucleotides through phosphoribosyltransferase and thus it acts as a n inhibitor in t h a t form.14 Whether 8 and 10 can be converted t o nucleotides is not known. In t h e pyrimidine series all compounds except 11 were moderately strong growth inhibitors. Their growth inhibitory potency was decreased by one to two orders of magnitude under conditions where folate metabolism is inoperative, indicating t h a t their primary mode of action involves interference with folate metabolism. Indeed, all of these pyrimidines were even more potent (2-5 times) in inhibiting the enzyme than t h e cell growth. In each case the inhibition was competitive with dihydrofolate. Since in the case of amethopterin about 100-1000 times higher concentration is needed t o inhibit t h e cell growth than the enzyme, i t appears that the transport of these pyrimidines through the cell membrane is less hindered than t h a t of amethopterin. T h e insertion of one or two methyl groups between cyclohexyl group and t h e pyrimidine ring decreased the biological activity relatively little. This is in contrast to the insertion of the NHCO group between adamantyl and t h e pyrimidine ring which lowered t h e biological activity of the analogs by three orders of magnitude.ln T h e latter observation was originally interpreted t h a t only lipophilic substituents adjacent t o the pyrimidine ring fit into t h e hydrophobic area of t h e enzyme. T h e length of the -CH2 CH2linkage is not much different from t h a t of -NHCO- linkage; however, t h e latter may impose specific conformations not necessarily enoountered with the -CH2 CH2- bridge. T h u s it appears that the hydrophobic area of this dihydrofolate reductase is not limited to a strictly confined region,
-
-
15
8
I5t 5
Discussion
1
I
I
2
4
6
I
2
11s Figure 1. Inhibition of dihydrofolate reductase of sarcoma 180 AT/3000 cells by 2,4-diaminopyrimidines and 2,fi-diaminopurines. Numbers refer to compounds designated in the text. For competitive inhibition Vlu = 1 K,/Ki X [I]/[S],where V is maximal velocity; u is initial velocity; K , and Ki are Michaelis and inhibition constants, respectively; and [I] and [SI are molar concentration of
+
inhibitor and substrate, respectively. For noncompetitive inhibition V/u = 1 + [S]/Ki X [I]/[S]. For calculation of Ki's the published K , value was used, i.e., 6 P M . 0, ~ substrate (dihydrofolate) M. concentration 1 X lOW4 M; 0 , substrate concentration 2 X directly adjacent to the pyrimidine binding site. T h e location of hydrophobic area with respect to the active site of dihydrofolate reductase has been discussed a t length by Baker.2 In agreement with t h e results of a n earlier study,' replacement of t h e cyclohexyl group by 2-naphthyl increased the inhibitory activity of diaminopyrimidine only slightly. T h a t study showed that a linear relationship existed between the molar volume of the 5-substituent of 2,4-diaminopyrimidines and t h e hydrophobicity, as determined from the water-heptane partition coefficient. Further, the affinity of t h e pyrimidines to dihydrofolate reductase increased with the hydrophobicity of the 5-substituent. Consequently, t h e larger t h e molar volume of the substituent t h e greater should be the affinity to the enzyme. T h e estimated
271 Joitrnal o f M e d i c i n a l Chemistry, 2975, Vol. 18. Y o . 3
molar volumes of cyclohexane and naphthalene are very similar, namely 108 and 112 ml, respectively, and the free dissociation energies of enzyme-inhibitor complex as calculated from the Ki values were also found to be similar, +9.8 kcal for 12 and +10 kcal for 5 . It is of interest that the 1-naphthyl derivative 11 differs from the 2-naphthyl isomer 5 by being a weak, noncompetitive inhibitor of dihydrofolate reductase. This suggests that the compound does not fit in the active site of the enzyme. Comparison of the molecular models of the two isomers reveals striking differences. In both cases the naphthyl residue can neither rotate freely nor assume a planar conformation with the pyrimidine ring. However, whereas the bulk of the naphthyl residue in t h e case of 2-naphthyl isomer deviates from the longitudinal axis of the molecule only by 30' or less, depending on conformation, this deviation reaches 90' in t h e case of 1-naphthyl isomer. It a p pears t h a t such deviation does not allow a proper fit of t h e inhibitor into the substrate binding site of the enzyme. T h e fact that the 1-naphthyl isomer inhibits cell growth a t substantially lower concentration than it does dihydrofolate reductase suggests t h a t this enzyme is not the target of action of this compound. T h e observation t h a t one of two compounds, chemically as similar as 1-naphthyl and 2-naphthyl derivatives of diaminopyrimidines, is a noncompetitive whereas the other a competitive inhibitor of a n enzyme points t o the general necessity for determination of actual K , values and of t h e type of inhibition, when structure-activity relationship is studied. T h e determination of IDio values alone for enzyme inhibition in such studies is highly inadequate and can easily lead to erroneous conclusions. T h e other pitfalls in t h e use of' ID,,, values in enzyme inhibitor studies have been discussed. l T 1 . l t i E x p e r i m e n t a l Section All melting points were taken on a Fisher-Johns apparatus and are uncorrected. Elemental analyses were performed by G. I. Robertson. Florham Park, N.J. Those analyses in which the results are within 0.4°;1 of the calculated values are denoted by the symbols for these elements. TICwas carried out on Rrinkman silica gel (F-254) plate\ o n aluminum. Ir confirmed t h e assigned structure of all conipc)unds discussed. 1 :v spectra were recorded on Cary 14 spect rophotimeter and were performed in absolute ethanol unless stated otherwise. No attempt was made to optimize yields in the reactions described helow. 2-Amino-4-hydroxy-~-cyclohexylmethyl-6-methylpyrimid i n e (1). A solution of sodium ethoxide (0.i9 g of Na, 0.0034 ga t o m ) i n 50 ml of absolute ethanol was cooled in an ice bath and guanidine hydrochloride (1.65 g, 17.2 minol) was added. Ethyl 2cyclohexylmethylacetoacetate" (3.89 g, 17.2 mmol) was added and the mixture was refluxed for 7 2 hr. T h e mixture was flash evaporated to a paste and neutralized with 5 M HC1 to give a glassy tan precipitate, which melted on the filter as filtration was attempted. l ' h e gum was treated with ether and dissolved in hot ethanol. I . p i i n cooling a white solid precipitated. This was collected by filiration and recrystallized from absolute ethanol to give t h e product (1..5 g, : W h , : m p 272-273O. Anal. ( C ~ ~ H I ~ C, N H, ~ ON.) 2,-1-Diamino-5-cyclohexylmethyl-6-methylpyrimidine(2). 1 ( 1.O g. 4.R mmol) was refluxed for 2 hr in a solution of 10 ml of N X ' I i and 1 .O :: of PCI:. T h e mixture was poured on crushed ice. 'I'he resulting white d i d was filtered off and dried in uacuo over P.0; lor 4 days. T h e resulting gum was dissolved in 125 ml of abs i r l u t e ethanol which was saturated with ammonia a t Oo and heated i n a htrmb at 160' for 24 hr. After evaporation of t h e solvent t h e residue was stirred for 1 hr with 0.5 M N a O H . T h e solid was filtered off. washed with ether, and recrystallized twice from hot absolute ethanol: yield 1.0 :: (100%): m p 209-210O. Anal. (C12H20Nd) C, H.
ti. 2-Amino-4-hydroxy-5-cyclohexylethyl-~-methylpyrimid i n e (3). Guanidine hydrochloride (10.1 g, 105 mmol) in a solution of' ;;odium ethoxide-absolute ethanol (4.86 g, 0.211 g-atom of Na) (23.9 g, 992 m M ) were m d ethyl 2-cyclohexylethylacet~acetate~~ refluxed for 48 hr. T h e reaction mixture was cooled and evapo-
rated to give a solid residue which was crystallized from aqueous ethanol; 10.9 g (46%, needles) was obtained. m p 2?6-2:?9° .Intri (C1:IHrjN:jO) C, H, N. 2,4-Diamino-~-cyclohexylethyl-6-meth~~lpyrimidine ( 4 1 . This compound was prepared by chlorination and amination ( i t 3 (2.0 g, 8.5 mmol) essentially as described for 2 except f o r the worku p of the final product which was as follow. After evaporation of ethanol-ammonia the solid was taken up i n sinall amount of 0.5 :I/ ethanolic sodium hydroxide, stirred for 1 hr, and filtered. 'T'lic~ solid was dissolved in acetone and filtered. E\,aporation of' acetone gave 1.6 g of crude product which was recrystallized t h r w times from absolute ethanol: yield 0.52 g ( 2 6 % ~of white needle>: nip 157-159'; , , ,A 232, 288 i r m a r 6.79). Anal. i ( ' ;,HiiNli Cy, H. N. 2,4-Diamino-5-(2-naphthyl)-6-methylpyrimidine (3). 'l'hii compound was prepared as described f o r the 1-naphthyl dvrivativfa 11; except that 3-(2-naphthyI)acetoiitrilewas the .tartin:: iiiaterial. T h e yield was ,57"6: mp 223-2241": t l r in ahsoiutc, ctha!ii,!. R , 0.35. Anal. (CIsH14N4i C. H , N. 2-Amino-4,6-bis[ ( 1-adamantoyl)amino]-5-nitrosopyrimid i n e (6). 2,4,6-Triamino-5-nitrosopyrimidine (6.16 g, 40 m m o l ~ was added to a solution of I-adamantanecarboxylic acid chloride (17.0 g, 85 mmol) in 600 ml of dry D M F at 0'. After t h e starting material dissolved the solution iurned blue and a h l u ~d i d precipitated. This was filtered and washed with anhydrous ether, yield 10.4 g (54%),and recrystallized from hot mixture of absolute etha no1 and chloroform. Ana/. (C2,,H:l4N6. 2H.01 1:. H. N , (:I. 2-Amino-6-( 1-adamantoy1)amino-8-( 1-adamanty1)purine (7). 6 (1 mmol, 178 mg) was placed in a mixture of' 1.1 mi of ethanol, 2.3 ml of acetic acid, and 2.3 ml of Ivater. T o t h i h \va- ,idded portionwise 0.69 g of zinc dust while the mixture i v a i heated ar 60'. Thereafter the mixture was refluxed for 90 m i n After tittering the remaining zinc the volume of the solution was reduced t o 5 ml, A white solid crystallized: yield 500 mg 199%); nip 216--217°: tlc iethoxyethanol -ethyl acerate-4% formic a t id. l.l:).'?f 11'. i\,O): I I L ,X,I q'120. . . l t ~ ~ l(CL(;H,~JIX~;O . .AcOHi C', H. N. 2,6-Diamino-8-( 1-adamanty1)purine ( 8 ) . f ( 2 0 0 III;. 11. mmol) was refluxed in 40 ml of ethanol while HC1 gas vas ~ I I ( J \ v H to pass through the reaction mixture for :ihr. T h e precipitate which formed was filtered and washed with acetone and eiher. 7'hc cream-colored solid (150 mg) wa5 dissolved in tioiling ivatrr, filtered, and precipitated with concentrated ammonium hydroxidr.. yield of dry material, GO mg (54%). A sample for an tained by recrystallization from aqueous ethanol: nil) (ethoxyethanoLethy1 acetate-4% formic acid. I :I 2:") A,,, 253, 285, 321 ( c , , ~ :3:1.6. ~ 2%8, 23.6 X 10 '1. .Inn/.
H, N. 2-Amino-4,6-bis(adamantaneacetylamino)-5-nit rosopyrimi d i n e (9). To 15.5 g (73 mmol) of 1-adamanty!acetic acid ~ ~ h l i i r i d ~ in 70 ml of dry D M F at 0' was added quickly 5.6 g i36.t; i i i m o i l i i t 2 , 4,6-triamino-5-nitrosopyrimidine. T h e mixture ~ v n , irrrc! I C J ~ i hr a t 0" and for 2 hr at room temperature A hluc solill l . \ i i i L , i : formed was collected. washed with anhydrous ether. : i l l ( { iir1v.tl: vield 10.3 g (86%);m p 263-264'. h i i / i('?,Hc,N,,O,. 0 5H('Ii ( H. N. 2,6-Diamino-8-(l-adamantylmethyl)purine ( I O ) . 9 1 1 tntnot. 525 mg) was placed in a mixture of 14 ml of ethanol. L..inil ( J ~ ' ~ I < w tic acid. and 2.3 ml of water, and 690 mp ot' zinc dust \ \ t i \ atitieti portionwise. T h e mixture was refluxed for 3 hr. Ai'1er re!ii(;\ i l l i ~ l the remaining zinc the solution was flash evapciratrLtl t i l :i ~ I a i i v mass. This was dissolved in 60 ml of absolute ethanol anti rrfliiieci for 3 hr while HCI gas was slowly bubbled throuph. 'The < , r v i t a l h n c solid which formed was filtered: yield 260 An analytical sample was obtained by re ethanol. i l m i . (Clc,H2zNl; * O.SHzOI H, N 64.03.
Acknowledgment. This work was supported i n part I>>.
LJ. S. Public Health Service Grants from t h e Nation;il C a n cer Institute, CA-14078 and CA-13038. References (1) E'. K. Ho, S. F. Zakrzewski. and I,. H. Mead, hi,w/>(su?i,
