Background:  The CAPRISA 004 trial (peri-coital prophylaxis using 1% tenofovir [TDF] vaginal gel) showed 39% efficacy against HIV infection. This study assesses whether low-frequency drug resistant mutants contributed disproportionately to breakthrough infections in women who acquired HIV despite TDF prophylaxis.

Methods:  Using deep 454 pyrosequencing, we examined HIV sequences from breakthrough infections in the treatment arm of the CAPRISA 004 trial (16 women to date) and from incident infections in the placebo arm (11 women to date). Two overlapping regions of HIV-1 RT were amplified from plasma and sequenced on a Roche 454 FLX sequencer to a target depth of 10,000X per participant. Translations of error-corrected DNA sequences were scanned for tenofovir-related resistance mutations, including K65R, K70, D67, Q151, and T215, and mutant frequencies in the TDF arm were compared with frequencies in the placebo arm. To assess error background and quantification, we sequenced serial dilutions of cloned K65R with a wild-type plasmid, and compared site entropies at resistance- and non-resistance sites.

Results:  The most common resistance mutation, K65R, was found at similar frequencies (above-background, but non-clinically significant) in both treatment groups. Mutations at the other examined resistance sites were found at sub-1% levels, and were not elevated in the treatment group. D67 mutations were slightly elevated in the control group; the other mutations assessed (V75, K70R, T215, and Q151) had nearly identical frequencies in both treatment groups. Low levels of TDF-related resistance mutations were found in women from both placebo and treatment groups; resistance-mutant frequencies were not significantly increased in women assigned to TDF gel.

Conclusions:  Deep sequencing detected TDF-related resistance mutants at very low levels in all women studied, showing substantially higher sensitivity than population sequencing or allele-specific PCR for potential drug-resistance detection. The lack of enrichment of resistance mutations indicates that in this low to moderate drug-resistance epidemiological setting, topically applied prophylactic TDF does not appreciably increase the incidence of TDF drug resistance in breakthrough infections.