Genotypic resistance testing is part of the routine management of patients with human immunodeficiency virus type 1 (HIV-1) infection. To optimize genotypic resistance test interpretation, it is essential to track virus mutations that cause or contribute to HIV-1 drug resistance. Although the reverse transcriptase (RT) mutations responsible for most nucleoside RT inhibitor (NRTI)- and nonnucleoside RT inhibitor (NNRTI)-resistant viruses are known, there have been recent reports that several less widely recognized RT mutations may also decrease RT inhibitor susceptibility, usually in combination with one or more of the known drug resistance mutations (reviewed in reference 10
However, two rare, novel mutations, Q145M/L, have been reported to cause high-level resistance to multiple NRTIs and NNRTIs, even in the absence of other known RT inhibitor resistance mutations. When placed in an HXB2 clone, pHXB2delta2-261RT, these mutations were reported to cause more than 10- to 100-fold resistance to the NRTIs zidovudine, lamivudine, stavudine, didanosine, tenofovir, and abacavir and to the NNRTIs nevirapine and efavirenz in both cell culture and enzymatic assays (5
). Despite the potential importance of this report, no subsequent studies have confirmed nor contradicted these findings in the above-cited HXB2 backbone, in another HIV-1 clone, or in clinical isolates.
We undertook several analyses and experiments to determine whether Q145M/L should be considered drug resistance mutations and be included in genotypic resistance test reports. Specifically, we determined whether mutations at RT position 145 were selected by RT inhibitors, contributed to decreased RT inhibitor susceptibility, or interfered with a virological response to RT inhibitors.
Table shows that six mutations at position 145 occur in about 0.1% to 0.2% of HIV-1-infected patients. Columns 2 through 5 of Table show that Q145M and other mutations at this position are not associated with NRTI or NNRTI therapy in the HIV Drug Resistance Database. Columns 6 through 8 show that in a large database of HIV-1 RT sequences from a commercial reference laboratory, Q145 mutations were as likely to occur in viruses without RT mutations as they were to occur in viruses with RT inhibitor resistance mutations. This lack of association with RT inhibitor therapy and RT inhibitor resistance mutations demonstrates that Q145 mutations are not selected by RT inhibitor therapy.
Prevalence of Q145 mutations in HIV-1-infected persons by RT inhibitor history (HIV Drug Resistance Database) and cooccurrence with other RT inhibitor resistance mutations (Quest Diagnostics laboratory database)
To assess the phenotypic impact of Q145M, we performed in vitro susceptibility testing on three infectious molecular clones containing Q145M and one containing Q145V (PhenoSense; Monogram, South San Francisco, CA) (7
). One of the three infectious molecular clones with Q145M was a site-directed mutant created on a pNL4-3 backbone using a QuikChange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA) to change the RT codon 145 of pNL4-3 from CAG to ATG. The remaining three infectious molecular clones were created by ligating patient-derived RT amplicons into a vector lacking RT codons 24 to 311, as previously described (3
). Each of the four recombinant infectious molecular clones was transfected into C8166 cells and expanded in SupT1 cells to create multiple aliquots of cell-free virus stocks that were tested for RT inhibitor susceptibility (PhenoSense assay; Monogram, South San Francisco, CA) (7
). Table shows that each of the three infectious molecular clones with Q145M and the clone with Q145V were fully susceptible to each of the FDA-licensed NRTIs and the first three licensed NNRTIs.
In vitro inhibitor susceptibilities of infectious molecular clones with RT codon 145 mutationsa
Among the patients undergoing HIV-1 genotypic resistance testing at the Stanford University Hospital Virology Laboratory for whom antiretroviral treatment histories and clinical follow-ups were available, mutations at position 145 did not interfere with the response to standard first-line treatment regimens. Among two of two patients with Q145M, two of two with Q145L, eight of nine with Q145V, one of one with Q145H, and three of three with Q145E, treatment with a standard first-line treatment regimen led to sustained virological suppression (<75 plasma HIV-1 RNA copies/ml; Siemens bDNA assay).
Examination of amino acids 143 through 157 in the three-dimensional structure of HIV-1 RT shows that amino acids 142 through 147 are part of beta-sheet 8; positions 155 to 157 are part of alpha-helix E; and positions 148 to 154 form a connecting loop (2
). However, in contrast to Q151, which is within 8 to 11 Å from the template, primer, and incoming deoxynucleoside triphosphate, Q145 is more than 20 Å from each of these structural entities (2
RT amino acids 143 through 157 are conserved in group M HIV-1 viruses, with the rare mutations at position 145 being the only mutations occurring in untreated individuals and the multinucleoside resistance mutation Q151M occurring in about 2% of NRTI-treated persons. In HIV-1 group O and HIV-1CPZ isolates, Q145C/H have rarely been reported. In HIV-2, Q145I is the consensus variant, and Q145V/M/T are other common variants at this position. In other primate lentiviruses, positions 145 to 148 are highly variable, whereas positions 149 to 157 are nearly completely conserved (Table ).
Amino acid variability of RT positions 143 to 157 among primate lentivirusesa
Although the previously published site-directed mutagenesis experiment and in vitro susceptibility results were performed in a pHXB2 backbone and ours were performed in a pNL4-3 backbone, this is not likely to explain the differences in results that we obtained because there is no example for such marked differences in in vitro drug susceptibility results obtained using pHXB2 or pNL4-3 vectors. Indeed, the Antivirogram assay (Virco Lab, Mechelen, Belgium) uses an HXB2-derived vector (1
), whereas the PhenoSense assay uses a pNL4-3-derived vector (7
). Although differences in reproducibility between the Antivirogram and PhenoSense assays have been reported, the results of these two assays are generally concordant (8
In conclusion, multiple lines of evidence suggest that the RT mutation Q145M and other mutations at this position do not confer RT inhibitor resistance and should not be reported as RT inhibitor resistance mutations on current genotypic resistance test reports. Novel drug resistance mutations should ideally be confirmed using a standardized phenotypic assay to validate their biological and potential clinical significance.