CS-92

A growing concern in the pursuit of new therapies for HIV-1 infection is the potential for the virus to develop drug resistance. With the advent of modern antiretroviral therapy and the common use of combined modalities, it is difficult to identify in the clinic the mutations associated with a specific drug. In general, drug selection of mutants using a relevant cell system, such as primary human lymphocytes, is a good prognosticator of what will happen in humans. In this study, HIV-infected human peripheral blood mononuclear cells were exposed, at a concentration of 1- to 10-fold the median effective antiviral concentration, to the nucleosides (-)-beta-2',3'-dideoxy-3'-thia-5-fluorocytidine [(-)-FTC] (-)-beta-2',3'-dideoxy-3'-thiacytidine (3TC), 3'-azido-2',3'-dideoxyuridine (CS-87, AZDU), 3'-azido-2',3'-dideoxy-5-methylcytidine (CS-92, AZMC), 2',3'-didehydro-3'-deoxythymidine (d4T), beta-L-2',3'-didehydro-2',3'-dideoxy-5-fluorocytidine (beta-L-D4FC), beta-L-2',3'-dideoxyadenine SATE[beta-L-ddAMP-bis(tbutylSATE)], beta-L-5-fluoro-2',3'-dideoxycytidine (L-FddC), and the protease inhibitors nelfinavir and amprenavir (VX-478). Virus from the culture supernatant was amplified by PCR and analysed by both HIV-1 reverse transcriptase and protease line probe assay. All the L-nucleoside analogues tested selected for the V184 mutation, including the L-pyrimidine nucleosides 3TC (-)-FTC, beta- L-FddC, beta-L-D4FC and the beta-L-purine nucleoside. beta-L-D4FC also selected for K/R65 in addition to V184, indicating that these two mutations are linked and compatible in vitro. No pattern of mutations leading to resistance or reduced susceptibility was discerned with d4T. Rapid genotyping analysis revealed the different kinetics and mutations obtained by in vitro selection in HIV-infected cells exposed to nucleoside analogues and protease inhibitors.

3'-Azido-2',3'-dideoxy-5-methylcytidine (CS-92, AzddMeC) is an antiviral nucleoside analogue structurally related to 3'-azido-3'-deoxythymidine (AZT). CS-92 is a potent and selective inhibitor of HIV-1 reverse transcriptase and HIV-1 replication in human lymphocytes and macrophages. The EC50 for CS-92 in HIV-1-infected human PBM cells was 0.09 microM. In HIV-1-infected human macrophages, the EC50 was 0.006 microM. This compound was also effective against human immunodeficiency virus type 2 in lymphocytes. The replication of Friend murine virus was only weakly inhibited, and no effect was observed against herpes simplex virus type 1 and type 2 and coxsackievirus B4. CS-92 was not toxic to PBM or Vero cells when tested up to 200 microM and was, furthermore, at least 40 times less toxic to granulocyte-macrophage and erythroid precursor cells in vitro than was AZT. The interaction of the 5'-triphosphate of CS-92 with HIV-1 reverse transcriptase indicated competitive inhibition (the inhibition constant, Kis, was 0.0093 microM) with a 30-fold greater affinity for CS-92-TP than for ddCTP. CS-92-TP inhibited HIV-1 reverse transcriptase by 50% at a concentration 6,000-fold lower than that which was required for a similar inhibition of DNA polymerase alpha. Pharmacokinetic studies showed that CS-92 was not deaminated to AZT in rats, but this compound was found to have a half-life of 2.7 hours. In rhesus monkeys, however, a compound with a retention time and ultraviolet spectra characteristics similar to AZT was detected. The mean half-life in rhesus monkeys for CS-92 was 1.52 and 1.74 h after intravenous and oral administration, respectively, and the oral bioavailability was about 21 percent. Additional preclinical studies with CS-92 will determine the ultimate utility of this antiviral agent for the treatment of HIV-1 infections.