Background:  As few as 1 or 2 amino acid changes in envelope can change the co-receptor use of HIV-1 from CCR5 to CXCR4. Given the high mutation rate of HIV-1, co-receptor switch variants should be readily generated, but CXCR4-using (X4 or R5X4) viruses take years to emerge in a subset of infected individuals. We hypothesize that intrinsic obstacles determined by the virus as well as extrinsic obstacles determined by the host limit the emergence of X4 variants.

Methods:  Co-receptor switch variants were selected in vitro by diluting target cells expressing CCR5 into target cells expressing CXCR4. After each round of target cell switching, viruses were tested for their ability to replicate on target cells expressing only the new co-receptor. Envelope sequences for switch variants were determined, and the relative sensitivity of the co-receptor switch variants to the CCR5 inhibitor PSC-RANTES or the CXCR4 inhibitor AMD3100 was compared to the starting isolate. Relative replication rates on CCR5- or CXCR4-expressing target cells were determined.

Results:  Multiple independent co-receptor switch variants were selected from 6 R5 parental strains after 12 to 120 days of culture; 1 to 7 amino acid sequence changes accompanied co-receptor switching. Arg and Lys substitutions in V3 and less frequently in V1-V2 were the most common mutations, and G-to-A base transitions accounted for most substitutions. Independent switch variants from the same R5 isolate shared common mutations. Most switch variants were R5X4 in phenotype, with a wide range of target cell preference. Some of the early switch variants displayed a transitional phenotype characterized by growth on U87-CD4-X4 cells but not MT-2 cells and high sensitivity to AMD3100 inhibition. R5X4 switch variants were 2- to 50-fold more sensitive to inhibition by PSC-RANTES than the parental isolate. The same R5X4 variants were more sensitive to AMD3100 inhibition on target cells expressing only CXCR4 than a panel of primary X4 isolates, and transitional intermediates were more sensitive than MT-2 adapted final variants.

Conclusions:  These results suggest that the transition from CCR5 to CXCR4 use involves unique and rare mutation pathways for each isolate examined. The time until isolation of switch variants was independent of the number of mutations. Susceptibility of both CCR5 and CXCR4 inhibitors increased during the transition phase, suggesting less efficient co-receptor use as an intrinsic obstacle to switching.

Keywords: coreceptor use; mutation; coreceptor switch