Several investigational INIs that retain antiviral activity against the most commonly observed INI resistance mutations have been identified, suggesting that the development of a second generation of INIs with higher genetic barriers to resistance may be feasible (5, 14, 16). This is desirable, particularly if INIs are to have an impact in low-income countries where drug resistance testing is not widely available and the selection of active inhibitors to prevent the rapid development of INI resistance may not always be possible.

A number of investigators have shown that, in vivo, raltegravir resistance is caused by a limited repertoire of primary integrase mutations, specifically, G140S together with Q148H/R/K, N155H with or without E92Q, and Y143C/R (1, 3, 6). To assist researchers in developing new INIs, we created a vector for cloning patient-derived HIV-1 integrase coding regions into a full-length HIV-1 clone lacking integrase. We then used this vector to create a panel of replication-competent prototypical raltegravir-resistant clones.

We used the high-fidelity, thermostable enzyme Pfu to amplify a 1,661-bp region from viral cDNA derived from each sample. The second-round PCR primers contained PacI and EcoRI sites to enable restriction digestion of IN-containing amplicons. PacI/EcoRI double digests of PCR products were followed by cycling T4 ligation to insert IN-containing amplicons into linearized pNL4.3ΔIN. The ligation products were transformed into competent Escherichia coli cells (Stbl3; Invitrogen, Carlsbad, CA).

The highest levels of raltegravir resistance (>200-fold decrease in susceptibility) were observed in clones containing both G140S and Q148H and in the clone containing Y143C. Lower levels of resistance (48- to 116-fold decreases in susceptibility) were found in clones containing N155H. These results are consistent with those reported by several other groups (3, 8, 10). The clone with Y143C and T97A had an extraordinarily high level of raltegravir resistance, confirming the highly synergistic interaction previously demonstrated only in site-directed mutants (4). Indeed, the level of raltegravir resistance associated with this pair of mutations is striking considering that Y143C alone reduces raltegravir susceptibility by only 3.5-fold (4) and T97A is a polymorphism with a 1 to 2% prevalence in raltegravir-naïve patients (12) that does not reduce raltegravir susceptibility (15).

Our data set contains the first reported clinical sequences with the mutations H51Y and Y143H. H51Y has been selected in vitro by elvitegravir and shown, using the PhenoSense assay, to reduce elvitegravir susceptibility 5-fold (7). A comparison of clones containing N155H/E92Q with and without H51Y suggests that H51Y has little impact on raltegravir resistance but may lower virus fitness (Table ​(Table1).1). Susceptibility data for Y143H have not been published previously.

E157Q, a polymorphism occurring in about 1 to 2% of INI-naïve patients (12), has been associated with virological failure during raltegravir therapy in one patient (11) and with a 3.3-fold reduction in elvitegravir susceptibility (7). However, the one sample in our panel with this mutation was fully susceptible to raltegravir.

Other than H51Y and S230R, each of the minor mutations listed in Table ​Table11 is polymorphic, occurring at a frequency of >1.0% in the absence of therapy (12). With the exception of D41H, K160T, Q216H, I220M, and D229E, each of the mutations listed under “Other” in Table ​Table11 is also polymorphic (12). Their potential impact, if any, on raltegravir susceptibility is not known.