Background:  A major obstacle to the efficient pre-clinical evaluation of new non-nucleoside reverse transcriptase inhibitors (NNRTI) is the lack of a cell culture assay that will predict the dominant-resistance mutations in clinical isolates during therapy. Historically, passage experiments of wild type HIV-1 in the presence of escalating drug have been used for this purpose. These assays notably did not predict the frequent occurrence of K103N and Y181C during therapy with delavirdine (DLV) and K103N during therapy with efavirenz (EFV). Since NNRTI-resistant variants likely pre-exist in patients before therapy is initiated, we postulated that infecting cells with a mixture of wild type and NNRTI-resistant variants followed by growth in the presence of drug would be a better model system.

Methods:  We produced wild type and mutant NL4-3 stocks (K103N, Y181C, P236L, V106A, G190A, G190S) separately by transfection into 293 cells. PM1 cells (7 x 10⁶) were infected with a total of 300 ng of a virus mixture (88% wild type and 2% of each mutant, based on p24 content of each virus stock). Replicate cultures were infected and propagated in the presence of different concentrations of EFV (0, 5, 10, 25, 50, 100, 250, 500, 750, and 1000 nM) or DLV (0, 100, 300, 1000, and 3000 nM). To determine the proportion of each variant, HIV protease and RT were amplified from genomic cellular DNA at day 6. Polymerase chain reaction (PCR) products were then sequenced from protease codons 49 to 99 and RT codons 1 to 320 to determine the dominant variant. The relative proportion of each variant was estimated by measuring relative peak heights on the electropherogram, and viral growth was quantified using p24 antigen concentration in the culture supernatant.

Results:  Only cultures with demonstrable virus growth were analyzed (all DLV concentrations and ≤100 nM EFV). The dominant variant present in the no drug controls was wild type (100%). The dominant mutant present at all time points and in all EFV concentrations <100 nM was K103N (100%). A minority G190S variant (20%) was present in addition to K103N (80%) at 100 nM EFV. The dominant variants present at all DLV concentrations were K103N (55 to 70%) and Y181C (15 to 40%). No other mutations were observed.

Conclusions:  We have developed a growth competition assay with pre-existing minority NNRTI-resistant variants that better mimics the selection of NNRTI-resistant mutants during clinical therapy with EFV and DLV. This assay may prove useful in predicting which variants will dominate during therapy with newer NNRTI.