To the editor—In their recent News and Views article1, Michael and Moore concluded that "given recent demonstrations of both viral resistance and persistent provirus in the face of the anti-retroviral drug regimes now in use, the careful clinical assessment of HIV-1 entry inhibitors is urgently needed." Although we agree wholeheartedly with this conclusion, we have data that challenge one of their assumptions regarding the emergence of alternative HIV-1 variants.

Michael and Moore state "the frequency with which X4 [CXCR4-tropic] viruses emerge in individuals who are heterozygous for the CCR5 Δ32 allele is no greater than in individuals with a wild-type allele2, despite the reduced CCR5 expression associated with the CCR5 Δ32 heterozygous genotype." The study they cite examined newly infected patients2. This is important, as there are other data to suggest that the relatively slow disease progression seen in CCR5 Δ32 heterozygotes is limited to the first few years after infection3. Among the moderately advanced HIV-infected patients who entered the NIAID AIDS clinical trials group protocol 241, baseline syncytium-inducing (SI) viral phenotype was more common among the CCR5/Δccr5 heterozygotes (7of 7, 100%) than among the CCR5/CCR5 homozygotes (29 of 88, 33%; P < 0.001, Fisher's exact test). The SI phenotype indicates tropism for cells expressing CXCR4 or both CCR5 and CXCR4, rather than for cells expressing only CCR5. Other factors that might explain this observation did not differ at study entry by CCR5 genotype, including median CD4 cell counts, plasma HIV-1 RNA levels, infectious HIV-1 titer in peripheral blood mononuclear cells, and history of prior anti-retroviral treatment4. These data are consistent with that possibility that reduced functional availability of CCR5 will increase selection for X4-tropic, or dual-tropic, viruses after several years of infection.

Given these data, we would like to further emphasize the need for a careful assessment of potential CCR5-active inhibitors as components of combination regimens. We also feel that an inhibitor active against CXCR4 would be a theoretically attractive agent to combine with a CCR5-active inhibitor, to attempt to limit broadening of co-receptor usage as a viral escape route.