Occurrence and significance of diastereomers of methotrexate alpha-peptides.

The L,L diastereomer of methotrexate-alpha-alanine (L,L-MTX-Ala) was synthesized by reaction of alpha-L-glutamyl-L-alanine di-tert-butyl ester with 4-amino-4-deoxy-10-methylpteroic acid, followed by removal of the blocking groups. It was identified by HPLC (C18 reversed-phase silica gel; acetic acid/CH3OH) as the slower of two closely spaced components in DL,L-MTX-Ala prepared previously by a different route [Kuefner et al. (1989) Biochemistry 28, 2288-2297]. The L,L diastereomer was hydrolyzed by pancreatic carboxypeptidase A (to yield MTX and Ala) twice as rapidly as the DL,L mixture. Analysis of the latter by HPLC established that the slower component was hydrolyzed to MTX and that the unreactive, faster component was D,L-MTX-Ala. DL,L-MTX-Arg was resolved by HPLC (NH4OAc/CH3CN) into two closely spaced components, and the diastereomers were partially separated by chromatography on DEAE-Trisacryl (H2O----2% NH4HCO3). Serum carboxypeptidase N hydrolyzed only the slower HPLC component (to yield MTX and Arg), thereby identifying it as the L,L diastereomer. When tested for cytotoxicity against L1210 cells, L,L-MTX-Arg (ID50 = 1.6 X 10(-8) M) was more effective than the D,L diastereomer (ID50 = 2.2 X 10(-7) M). Treatment of MTX with dicyclohexylcarbodiimide and N-hydroxysuccinimide (NHS), followed by hydrolysis of the NHS ester, led to racemization in the L-glutamate moiety of MTX as shown by the fact that the product was hydrolyzed by carboxypeptidase G2 (at the pteroate-Glu bond) only to the extent of ca. 50% compared to the untreated control. These observations have a broad significance, since coupling agents are employed extensively in the derivatization of MTX for attachment to affinity supports and monoclonal antibodies.

Carboxypeptidase-mediated release of methotrexate from methotrexate alpha-peptides.

Methotrexate (MTX) alpha-peptides containing representative neutral (alanine), acidic (aspartic acid), and basic (arginine) amino acids were synthesized by a regiospecific route. Purity and authenticity of MTX-Ala, MTX-Asp, and MTX-Arg were established by TLC, HPLC, elemental analysis, and NMR and absorbance spectra. These peptides were hydrolyzed by carboxypeptidases to yield MTX and the amino acids. Reactions were monitored by using a ninhydrin assay for the amino acids and HPLC and spectrophotometric assays for MTX. Pancreatic carboxypeptidase A (CP-A) hydrolyzed MTX-Ala and, at a much slower rate, MTX-Asp and MTX-Arg. MTX-Ala was also a substrate for pancreatic carboxypeptidase B (CP-B); marginal activity was observed with this enzyme and MTX-Arg. Human serum hydrolyzed only MTX-Arg; biphasic inhibition of this activity by 2-(mercaptomethyl)-3-(guanidinoethyl)thiopropionate was consistent with the known presence of two types of endogenous carboxypeptidase (CP-N). Cytotoxicity of the MTX peptides toward L1210 cells in culture was enhanced considerably in the presence of the appropriate carboxypeptidases. MTX-Ala was much less toxic than MTX (ID50 values of 2.0 X 10(-6) M and 2.4 x 10(-8) M, respectively), but in the presence of CP-A the ID50 of the peptide improved to 8.5 X 10(-8) M. Similar results were obtained with MTX-Asp/CP-A and MTX-Ala/CP-B combinations. MTX-Arg showed good cytotoxicity (ID50 of 5.0 X 10(-8) M), due to CP-N activity in the fetal bovine serum of the culture medium; inclusion of CP-B lowered the ID50 to that of MTX. Possible clinical uses of MTX peptides are discussed.

Chemotherapeutic potential of methotrexate peptides.

MTX peptides in which the amino acid was linked to the alpha-carboxyl group have been prepared and examined for cytotoxicity before and after treatment with proteolytic enzymes. The alanine, aspartic acid and arginine derivatives (MTX-ala, MTX-asp and MTX-arg) were synthesized by a regio-specific route, following the general procedures of Rosowsky and Montgomery. Each compound was obtained in good yield, and purity was established by TLC, HPLC, absorbance spectra and elemental analyses. The MTX peptides were not hydrolyzed by a variety of proteolytic enzymes (e.g., trypsin, plasmin, urokinase, aminopeptidase). Pancreatic carboxypeptidase A, however, hydrolyzed MTX-ala readily, MTX-asp slowly and MTX-arg not at all. The MTX-ala and, to a lesser extent, MTX-arg were substrates for pancreatic carboxypeptidase B. MTX-arg was also hydrolyzed by the endogenous carboxypeptidase N in human serum. The cytotoxicity of these MTX peptides toward L1210 cells was measured in a microculture assay system using a tetrazolium dye. MTX-ala was weakly cytotoxic (ID50 = 2.0 x 10(-6)M) compared to MTX (ID50 = 2.4 x 10(-8)M). When MTX-ala was tested in the presence of carboxypeptidase A, the ID50 value improved to 8.5 x 10(-8)M. MTX-arg gave an ID50 of 5.0 x 10(-8)M, which was not unexpected in view of its susceptibility to hydrolysis by the carboxypeptidase activity present in the fetal calf serum of the culture medium. Inclusion of carboxypeptidase B lowered the ID50 value to 2.5 x 10(-8)M. Possible clinical uses of MTX peptides are discussed.