FOR a number of years basic biochemical and biological research has been directed toward gaining new knowledge regarding the mechanism of action of purine antagonists, particularly 6-mercaptopurine (MP) because of its useful temporary anti-leukaemic activity in man1. One of the goals of such efforts has been to gain information which could serve as a rationale for design and synthesis of purine antagonists that might circumvent the most prevalent biochemical mechanisms associated with cellular resistance to MP. A common biochemical alteration (observed in bacteria, experimental cancer cells including leukaemic cells, and human cancer cells in culture) associated with the development of cellular resistance to MP is the loss of activity of a specific nucleotide pyrophosphorylase (IMP-GMP pyrophosphorylase) responsible for catalysing the conversion of MP to its cytotoxic form, MP-ribonucleotide2. It has not been possible to circumvent this loss of pyrophosphorylase activity by exposure of MP-resistant cells to MP-ribonucleotide, because this nucleotide either fails to enter the cell, or is degraded in the process of entry. MP-ribonucleoside is also ineffective against MP-resistant cells; its inactivity may be due either to its rapid degradation to MP (which has been observed in Ehrlich ascites cells3), to the absence of the necessary nucleoside phospho-kinase activity (presumably ?inosine kinase?, an enzyme which has not been reported4), or to both these factors. Efforts to overcome resistance to MP by employment of bis (hioinosine)-5',5'''-phosphate have shown some promise5, but MP-resistant cells in culture show some degree of cross-resistance to this compound (about 20- to 40-fold as compared with > 1,000-fold resistance to MP). Such cross-resistance is too great to offer encouragement for in vivo activity against MP-resistant neoplasms, because of the relatively low therapeutic index of all known purine antagonists.Table l. RELATIVE EFFECTIVENESS OF G-METHYLTHIOPTJKINE RIBONTJCLEO-SIDE AND RELATED COMPOUNDS AGAINST H. Ep. No. 2 CELLS IN CITLTURE AND A 6-MERCAPTOPURINE-RESISTANT VARIANTAverage ED50 (jug/ml.) * Compound H.Ep.-No.ZjS H.Ep.Ko.ZIMP 6-Mercaptopurine 0-21 400 6-Mercaptopurine ribonucleoside 0-36 > 100 6-Mercaptopurine ribonucleotide 0-61 > 100 6-Methylthiopurine 73 111 6-Methylthiopurine ribonucleoside f 0-10 0-07 6-Methylthiopurine ribonucleotide j 0-17 0-04 6-Ethylthiopurine ribonucleoside 0-29 0-29 6-Benzylthiopurme ribonucleoside 3-7 2-8 e-CtycZopentylthiopurine ribonucleoside > 100 > 100 6-Thioguanine 0-04 52 6-Thioguanosine 0-38 > 100 6-Thioguanylic acid 0-28 > 100 6-Methylthioguanine 44 32 6-Methylthioguanosine > 100 > 100Compound 6-Mercaptopurine 6-Mercaptopurine ribonucleoside 6-Mercaptopurine ribonucleotide 6-Methylthiopurine 6-Methylthiopurine ribonucleoside f 6-Methylthiopurine ribonucleotide j 6-Ethylthiopurine ribonucleoside 6-Benzylthiopurme ribonucleoside e-CtycZopentylthiopurine ribonucleoside 6-Tnioguanine 6-Thioguanosine 6-Thioguanylic acid 6-Methylthioguanine 6-Methylthioguanosine* ED50 is that concentration of drug inhibiting cell growth by 50 per cent. Cells were grown on glass and growth was measured by determination of protein content (method of Oyama and Eagle (ref. 30)) after 4 days growth in the presence of drug. ?Average of 9 independent and internally consistent experiments.?Barium salt monohydrate. The present article is concerned with the biological activity (including the in vivo anti-leukaemic activity), some intermediary metabolic effects, and the metabolism of 6-methylthiopurine ribonucleoside (MeMP-ribonucleoside), a simple derivative of MP with high activity in culture against human cancer cells resistant to MP.Inhibition studies with H. Ep. No. 2 cells in culture. The isolation of a line of human epidermoid cancer cells (H. Ep. No. 2) resistant to MP has already been described6; this line is completely devoid of IMP-GMPpyrophosphory-lase, which converts MP to MP-ribonucleotide7. A second line of H. Ep. No. 2 cells resistant to MP was also used in studies of metabolic effects described later; this line shows the same loss of IMP-GMP pyrophosphorylase activity as the first MP-resistant line, but since it represents an independently isolated cell line and since it has been carried in the presence of 6-thioguanine rather than MP, it is given a separate designation (H. Ep. No. 2/MP/TG) to distinguish it from the first resistant line (H. Ep. No. 2/MP) and the parent line (H. Ep. No. 2/S). These cell culture lines were propagated in stationary culture in Eagle's basal medium supplemented with 10 per cent calf serum. The drug response was quantitatively assessed by a standardized procedure8. The results, presented in Table 1, show the inability of MP, MP-ribonucleoside and MP-ribonucleotide to inhibit the resistant line, and also show the cross-resistance of this line to 6-thioguanine,6-thioguanosine, and 6-thioguanylic acid. In contrast, MeMP-ribonucleoside was equally effective against H. Ep. No. 2/S and H. Ep. No. 2/MP cells, as were MeMP-ribonucleotide and the S-ethyl and S-benzyl (but not the S-cyclopeutyl) derivatives of MP-ribonucleoside. The free base, 6-methylthiopurine (MeMP), was relatively non-toxic to both the sensitive and resistant lines, the ED50 concentrations being 700- to 1,000-fold higher than those of MeMP-ribonucleoside. It is also of interest that a similar relationship did not hold in the thioguanine series: both 6-methylthioguanine and 6-methylthio-guanosine were relatively without inhibitory activity against both MP-sensitive and MP-resistant cell lines. Evaluation of the response in vivo of L1210 leukaemia and an MP-resistant variant thereof to MeMP-ribonucleoside. Using standardized procedures9, MeMP-ribonucleoside and MP have been compared with regard to capacity to increase life-span of BDF1 mice bearing L1210 or L1210/MP leukaemias. The design and results of these experiments are shown in Table 2. MeMP-ribonucleoside was about as effective against the MP-resistant leukaemia as was MP against the parent sensitive leukaemia, and MeMP-ribonucleoside produced about the same increase in life-span in mice with sensitive and with MP-resistant leukaemic cells. These results have been confirmed repeatedly.Pseudo-feedback inhibition of purine biosynthesis. Many purine analogues, including MP, inhibit an early step of purine biosynthesis10-13 presumably the first step of the pathway14, as a result of the capacity of their nucleotides to mimic the action of natural purine nucleotides as feedback inhibitors. Earlier investigations with MP and other purine analogues in resistant cells have shown that pseudo-feedback inhibition of purine biosynthesis occurs only in cells that have the capacity to form the nucleotides 13,15. Since the activity of MeMP-ribonucleoside against MP-resistant cells suggests that MeMP-ribonucleoside is being converted to a nucleotide in cells where such conversion of MP or MP-ribonucleoside does not occur, it was of interest to determine the effectiveness of this compound as a pseudo-feedback type of inhibitor.Table 2. A COMPARISON or THE in vivo RESPONSE OF G-MERCAPTOPTTRINE-RESISTANT 1210 LEUKAEMIA TO 6-METHYLTmopuRiNE RIBONUCLEOSIDE AND 6-MERCAPTOPURINE *No. of L12W/MP cells inoculated Treatment Dosage (qd 1-15 days) Median host life-span (days) % Increase in host life-span! mg/kg mg/m2 Relative to the LDW Untreated controls Treated10" 6-Mercaptopurine 40 120 1-4 10 9-0 -10 27 81 0-96 9-5 -518 54 0-64 9-0 -10 12 36 0-43 10-0 08 24 0-29 10-0 0 5 15 0-18 10-0 010 6-Methylthiopurine 36 108 ca. 2-4 10 12-5 25 ribonucleoside 24 72 1-6 17-0 7016 48 1-1 15-0 50 11 33 0-73 14-0 407 21 0-47 12-0 20 5 15 0-33 13-0 303 9 0-20 12-0 20 10* 6-Mercaptopurine 40 120 1-4 12 11-5 -427 81 0-96 12-5 -4 18 54 0-64 13-0 812 36 0-43 12-0 0 8 24 0-29 12-0 05 15 0-18 12-0 0 10* 6-Methylthiopurine 36 108 ca. 2-4 12 8-5 -29ribonucleoside 24 72 1-6 15-5 29 16 48 1-1 19-0 5811 33 0-73 18-0 50 7 21 0-47 18-0 505 15 0-33 18-0 50 3 Q 0-20 15-0 25* The above are results from concurrent experiments in which randomly distributed leukaemic mice were used. Leukaemic cells were inoculated intra -peritoneally and all therapy was by the intraperitoneal route. The MP-resistant leukaemia used was a line L210/MP (III)) isolated by Hutchison (ref. 31). The maximum percentage increase in host life-span of animals bearing LI210/MP leukaemia provided by MeMP-ribonucleoside is approximately the same as that observed in animals bearing the parent L1210 leukaemia on treatment with the optimal chronic dosage of MP (ref. 32). In experiments carried out so far, the activity of MeMP-ribonucleoside and MP against the parent line of L1210 leukaemia appear to be about the same. ?Groups of 20 control and 10 treated animals were used. There were no 30-day survivors in either control or treated groups.Azaserine specifically blocks an early step in purine biosynthesis, and the feedback or pseudo-feedback inhibition can be assayed conveniently by determining the effectiveness of a given agent in decreasing the amount of 14C-labelled formylglycinamide ribonucleotide (FGAR) that accumulates in cells provided 14C-formate and treated with azaserine. The procedures used, which have been described in detail elsewhere13,15,16, are modifications of those used by LePage, Henderson et al.11,12. The details of the experiments are contained in Table 3, which presents the results obtained with the two separate MP-resistant H. Ep. No. 2 lines and the parent sensitive line in culture and with L1210/MP cells in vivo. Each of these resistant cell lines previously has been shown to be cross-resistant to MP-ribonucleoside and to be devoid of IMP-GMP pyrophosphorylase activity, the enzyme which converts MP and 6-thioguanine to nucleotides7,13. TABLE 3. EFFECTS OF 6-METHYLTHIOPUBINE RIBONUCLEOSIDE AND RELATED COMPOUNDS ON THE AZASERINE-INDUCED ACCUMULATION OF FORMYL-GLYCINAMIDE RlBONUCLEOTIDE (FGAR) "Gin FGARCell line Inhibitor and concentration (%ofazaser-ine control) Exp. 1* /ug/ml. (stationary culture)//. Ep. No. ZIS 6-Methylthiopurine ribonucleoside 5-0 1 0-5 3 0-1 15 6-Mercaptopurine ribonucleoside 5-0 1 11. Ep. No. 21 MP 6-Methylthiopurine ribonucleoside 5-0 1 0-5 13 0-1 57 6-Mercaptopurine ribonucleoside 50 92Exp. 2* (suspension culture) //. Ep. No. ZIS 6-Methylthiopurine ribonucleoside 9-82 3 0-98 3 II. Ep. No. 2/ 6-Methylthiopurine ribonucleoside 98-2 4 MPITG 9-82 5 0-98 4 6-Mercaptopurine 50 95 6-Methylthiopurine 48-8 78 6-Mercaptopurine ribonucleoside 83-5 60 6-Thioguanine 50-0 913 xp. 3 (in vivo) mg/kg JA2WIMP 6-Methylthiopurine ribonucleoside 982 9 49-1 7 24-5 10 6-Mercaptopurine ribonucleoside 83-5 94* Cells were treated with azaserine (10 mg/ml.), and sodium formate-14C (25 mc. per flask or bottle) was added 1 h later. Purine analogues were added 30 min after the azaserine. Controls (?azaserine control?) received only azaserine and formate-14C. Cells were collected 2 h after addition of the radioactive formate. ? Azaserine (0-5 mg/kg) was administered intraperitoneally to groups of 3 mice each, bearing 7-day-old implants of leukaemia L1210/MP (ref. 31), followed 30 min thereafter by the purine analogue. Sodium formate-14C (50 mc./mouse) was given 1 h after the azaserine, and the cells were collected 2 h after administration of the tracer. The soluble fraction of the cells was prepared and assayed by the same methods used for H. Ep. No. 2 cells.MeMP-ribonucleoside at concentrations of 0-5 mg/ml, or less blocked the formation of FGAR almost completely in both the sensitive and resistant cell culture lines. MP-ribonucleoside was effective in blocking FGAR synthesis in H. Ep. No. 2/S cells, but not in H. Ep. No. 2/MP or H. Ep. No. 2/MP/TG cells. MP was also without effect on FGAR synthesis in the H. Ep. No. 2/MP/TG line; its high effectiveness in the parent line and its ineffectiveness in the H. Ep. No. 2/MP line have been reported earlier15. MeMP-ribonucleoside, at therapeutic doses, was also highly effective in blocking synthesis of FGAR in LI2W/MP leukaemic cells in vivo, whereas MP-ribonucleoside was without effect (Table 3). The effectiveness of MeMP-ribonucleoside as a pseudo-feedback inhibitor has also been noted by Henderson12 in studies with Ehrlich ascites cells in vitro. Metabolism of MeMP-ribonucleoside-35 S. MP-ribo-imcleoside-35S was prepared by exchange between sulphur-35 dissolved in pyridine and unlabelled MP-ribonucleoside as described by Moravek and Nejedly17. Methylation of MP-ribonucleoside-35S with methyl iodide gave MeMP-ribonucleoside-35S, obtained in about 80 per cent overall yield from sulphur-35. Two batches of labelled compound with specific activities of 2.6 mc./mg and 16.5 were used in metabolism studies.H.E. No.2/S and H. Ep. No. 2/MP/TG cells in suspension culture (4-6 x 107 cells in 200 ml. medium) were exposed to MeMP-ribonucleoside-35S (5 (mg/ml.) for 4 h. At this time, the cells were collected by centrifugation, washed, and extracted with hot 80 per cent ethanol as described elsewhere16. Three to four per cent of sulphur-35 added was present in the ethanol extract. The alcohol extract was analysed by two-dimensional paper chromatography -autoradiography (phenol-water followed by n-butanol-propionic acid-water) and by one-dimensional paper chromatography (butanol-propionic acid-water), as described previously16. Eighty to eighty-five per cent of the total sulphur-35 on the chromatograms was located in a spot falling in the nucleotide area of the chromatograms (RF 0.30 in butanol-propionic acid); 5?11 per cent of the total sulphur-35 was in a spot of Rp 0.71, and the remainder in several small spots falling below the principal spot. Similar uptake and distribution were observed in both cell lines, and 24 h exposure to the labelled drug did not change either the extent of uptake or the distribution of radioactivity among the spots on the chromatogram. By co-chromatography in three solvents the spot with RF value 0.71 was identified as unchanged MeMP-ribonucleoside. The principal spot (Rp 0.30) was identified as the 5'-mono-phosphate of MeMP-ribonucleoside by (a) co-chromatography with an authentic synthetic sample of 6-methylthioinosine-5'phosphate (prepared by the methy-lation of thioinosinic acid (J. A. Montgomery and H. J. Thomas, unpublished observations) ), in three solvents (butanol-propionic acid-water; dicyclohexylamine, buta-none, H2O (2:10:5, v/v); t-butyl alcohol, butanone, H2O, diethylamine (4:4:2:4, v/v) ), in all of which there was complete coincidence of ultra-violet absorption and radioactivity; and (b) by treatment with calf intestinal phosphatase (Pentex, Inc., Kanakee, 111.) for 1.5 h, 37 C, pH. 9?10, which converted the principal spot completely to a compound indistinguishable from MeMP-ribonucleoside by co-chromatography. The identities of the spots with RF values less than that of the nucleoside mono-phosphate have not been determined. MeMP has been observed to undergo some demethylation in vivo and in vitro18-20; MP-ribonucleotide was therefore a probable metabolite of MeMP-ribonucleotide. However, in the butanol-propionic acid solvent, MP-ribonucleotide (Rp 0-09-0-13) and MeMP-ribonucleotide can be distinguished easily; very little radioactivity at the RF of MP-ribonucleotide was noted on chromatography-autoradiography of extracts of cells exposed to MeMP-ribonucleoside-35S; and when the material falling at this RF value was eluted and treated with alkaline phosphatase, no radioactivity was detected at the RF of MP-ribonucleoside. There was, therefore, little or no conversion of MeMP-ribonucleoside to MP-ribonucleotide. The possible incorporation of MeMP-ribonucleoside into polynucleotides is a question yet unanswered. When H. Ep. No. 2/S cells were grown in the presence of MeMP-ribonucleoside-35S, very small amounts of radioactivity were isolated with the nucleic acids. Determination of whether this radioactivity represents only contamination or a true incorporation is made difficult by the very small amount of sulphur-35 isolated with the polynucleotides and by the instability of MeMP-ribonucleotide?an instability which we observed in the present work and which has recently been noted also by Pfleiderer et al.21. Additional evidence against incorporation into polynucleotides is the fact that on chromatograms made from the 80 per cent alcohol extracts of cells grown in the presence of MeMP-ribonucleoside-35S there was little radioactivity in the area to which the di- and tri-phosphates of MeMP-ribonucleoside would be expected to migrate.It is also worth noting that no MeMP was detected on the chromatograms made from the 80 per cent alcohol extracts?results indicating that MeMP-ribonucleoside is not degraded by nucleoside phosphorylases in intact cells. This finding is in accord with the recent report of Paterson and Sutherland3 that MeMP-ribonucleoside is not converted to MeMP-by cell-free extracts of Ehrlich ascites cells. In separate experiments, cell-free extracts of H. Ep. No. 2/S cells were incubated for varying lengths of time with MeMP-ribonucleoside-35S (in tris buffer, 0-05 M, pH. 7-4, containing ATP at a final concentration of 3 1 mmoles/ml.), after which the incubation mixtures were analysed by the usual chromatography-autoradiography procedures. The autoradiographic patterns thus obtained were similar to those obtained in intact cells; MeMP-ribonucleotide was the principal spot, and the maximum amount of nucleotide was produced during the first 20 min of incubation.The various types of data presented here show MeMP-ribonucleoside to be an interesting compound on several counts. From the practical point of view, it is a simple derivative of MP that overcomes a mechanism of resistance to MP that is found in many species. From the biochemical point of view, it is of interest that S-methylation of MP-ribonucleoside apparently completely changes its substrate specificity for some of the enzymes of purine metabolism. Thus, in intact cells MeMP-ribonucleoside would appear not to be a substrate for nucleoside phosphorylases, and, in vitro, MeMP-ribonucleoside remains intact under conditions that produce extensive cleavage of MP-ribonucleoside to the free base3. The extensive conversion of MeMP-ribonucleoside to the nucleotide in cells which fail to convert MP-ribonucleoside to the nucleotide shows MeMP-ribonucleoside to be a substrate for a nucleoside kinase that does not act on MP-ribonucleoside. Kinases for inosine and guanosine have not been reported4, whereas adenosine kinase is known to occur in mammalian cells22; other ribonucleoside kinases known are uridine kinase, riboflavin kinase, and ribosylnicotinamide kinase22. These facts, and the fact that MeMP-ribonucleoside acts as a purine analogue in inhibiting purine synthesis by a pseudo-feedback action, make it likely that it is adenosine kinase that converts MeMP-ribonucleoside to the nucleotide. With regard to the possibility that MeMP-ribonucleoside is an analogue of adenosine, it is worth noting that, whereas inosine (I) and MP-ribonucleoside (II) both bear a proton at N-l (since they exist largely in the lactam and thiolactam forms23), adenosine (III)24,25 and MeMP-ribonucleoside (IV) do not. The absence of a proton at N-l of the pyrimidine ring might, therefore, be more critical for activity as a substrate for the kinase than the nature of the group at the 6-position of the ring. In this connexion, the inhibitory activity of other S-substituted derivatives of MP-ribonucleoside should be noted (Table 1). Another interesting point relating to substrate specificity is the fact that MeMP was not a substrate for any of the purine nucleotide pyrophosphorylases when assays were carried out with cell-free extracts of H. Ep. No. 2/S cells by-procedures--that. have been described elsewhere7,26. This finding would suggest that the activity of MeMP in vivo27,28 is the result of demethylation to MP and its consequent conversion to the nucleotide by sensitive cells. In so far as its metabolism in cultured cells is concerned, MeMP-ribonucleoside (and possibly its congeners) appears, then, to be a unique inhibitor in that it is not a substrate for enzymes degrading other purine nucleosides but is an excellent substrate for a kinase that converts it to the nucleotide, apparently the active inhibitory form. Since MeMP-ribonucleoside is converted to MeMP-ribonucleotide and not to 6-MP-ribonucleotide, it is an interesting question whether these two nucleotides inhibit cell division by means of the same mechanism. All facets of the action of MP are not yet understood; however, with regard to the mechanisms of action of MP and MeMP-ribonucleoside and to the possibility that MeMP-ribonucleoside acts as an adenosine analogue, it is noteworthy that both MP and MeMP-ribonucleoside (after intracellular conversion to the nucleotide) are potent pseudo-feedback inhibitors of purine biosynthesis in intact mammalian cells, and that this is an action that they have in common with several cytotoxic analogues of adenine and adenosine, that is, 2-fluoroadenine, 2-fiuoroadenosine, 7-deazaadeno-sine (tubercidin), and others29.The in vivo activity of MeMP-ribonucleoside against leukaemia L1210/MP suggests that this agent is worth considering for clinical trial against MP-resistant acute lymphatic and chronic myelocytic leukaemias. This work was supported by the Cancer Chemotherapy National Service Center, National Cancer Institute, National Institutes of Health, contracts PH-43?6451, SA-43-ph-3784 and SA.-43-Ph-2433, and by grants from the Charles F. Kettering Foundation and the Alfred P. Sloan Foundation.