Wild-type reverse transcriptase has evolved for the survival of human immunodeficiency virus type 1 (HIV-1) by natural selection. In contrast, therapy relying on inhibitors of reverse transcriptase by nucleosides like zidovudine (AZT) or dideoxyinosine (ddI), and by non-nucleosides like pyridinones or nevirapine, may exert different selection pressures on this enzyme. Therefore the acquisition of resistance to reverse transcriptase inhibitors by selection of mutations in the pol gene may require compromises in enzyme function that affect viral replication. As single mutations are unlikely to confer broad resistance when combinations of reverse transcriptase inhibitors are used, multiple mutations may occur that result in further compromises. Certain drug combinations may prevent the co-existence of adequate reverse transcription function and multi-drug resistance (MDR). Unlike bacterial or eukaryotic drug resistance, retroviral drug resistance is conferred only by mutations in its own genome and is limited by genome size. Combining drugs directed against the same essential viral protein may thus prevent HIV-1 MDR, whereas the conventional approach of targeting different HIV-1 proteins for combination therapy may not, because genomes with resistance mutations in different HIV-1 genes might recombine to develop MDR. Here we show that several mutations in the HIV-1 reverse transcriptase gene that confer resistance to inhibitors of this enzyme can attenuate viral replication. We tested whether combinations of mutations giving rise to single-agent resistance might further compromise or even abolish viral replication, and if multidrug-resistant viruses could be constructed. Certain combinations of mutations conferring resistance to AZT, ddI and pyridinone are incompatible with viral replication. These results indicate that evolutionary limitations exist to restrict development of MDR. Furthermore, a therapeutic strategy exploiting these limitations by using selected multidrug regimens directed against the same target may prevent development of MDR. This approach, which we call convergent combination therapy, eliminated HIV-1 replication and virus breakthrough in vitro, and may be applicable to other viral targets. Moreover, elimination of reverse transcription by convergent combination therapy may also limit MDR.

No abstract

No abstract

ACKNOWLEDGEMENTS. We thank G. Frey for M. kandleri cells, L. Liu for antibody aga inst human topoisomerase I, H. Nash for). Int protein , Y. C. Tse-Dinh for E. coli topoisomerase I. J. Wang tor S. cerevisiae topoisomerase Ill and antibody against E. coif topoisomerase I, and C. Scheim er and K Eichinger for technical assistance. The work was supported in part by funds from the Leibniz Preis (K.S.) and the NSF (JAL.), and by Alexa nder von Humboldt and Alfred P. Sloan fellowships (to A.LS.).