Antiviral Activity of Tenofovir Alafenamide Against HIV-1 Subtypes and Emergence of K65R

Combined antiretroviral therapy (cART) leads to improved survival and increased life expectancy in the HIV-1-infected population. Tenofovir (TFV), a widely used nucleotide HIV reverse transcriptase inhibitor (NRTI), and its prodrug, tenofovir disoproxil fumarate (TDF), can be an effective part of cART. Tenofovir alafenamide (TAF) is a novel prodrug of tenofovir and has an improved safety profile and antiviral potency than TDF. ¹

Tenofovir has been shown to be effective against all HIV-1 subtypes. ² Although subtype B is found primarily in North America and Europe, subtype C is the most prevalent subtype and found predominantly in sub-Saharan Africa. It was previously reported that the reverse transcriptase (RT) mutation K65R, the primary resistance mutation for tenofovir, ³ could emerge more rapidly in vitro in HIV-1 subtype C isolates treated with tenofovir than other subtypes, ^4,5 suggesting that tenofovir could be less effective against those isolates.

Clinical data from studies with TDF-containing regimens failed to find a correlation between K65R and subtype C. ⁶ More recent clinical data have linked subtype C with an increased rate of K65R emergence, ⁷ whereas additional reports have found no correlation between subtype C and virologic failure on TDF-containing regimens. ⁸ This inconsistent data could be attributed to a variety of factors, such as the different types of NRTI backbones being used with tenofovir, the limited number of patients in some of the analyses, and the geography and clinical characteristics of the population studied. Here we present in vitro results and data from clinical studies to further evaluate K65R emergence in subtype C virus.

Dose escalation selection studies were performed with three subtype B and four subtype C viruses to investigate potential differences in K65R emergence in vitro. Consensus sequence variations at RT codons 64–65 between subtype B (AAG AAA/G) and subtype C (AAA AAG) were suggested to contribute to the earlier emergence of K65R. ⁹ To further investigate this, we also included two subtype B viruses with subtype C-like sequence at codons 64–65 (AAA AAG) (labeled subtype B*) for a total of nine viruses (Supplementary Tables S1 and S2; Supplementary Data are available online at www.liebertpub.com/aid).

Recombinant mutant HIV-1 viruses were created by direct polymerase chain reaction of patient isolates and profiled in MT-2 cells as described previously. ¹⁰ TAF antiviral activity was similar for all viruses regardless of the HIV-1 subtype, ranging from 0.7- to 1.1-fold of the reference value (Supplementary Table S1), similar to previous observations. The dose escalation resistance selection studies were done in MT-2 cells over a period of 3–6 months as described previously. ¹¹ Because of their stochastic nature, selections were conducted at least in duplicate for each of the nine viruses, resulting in 18 experiments with TAF and 22 experiments with TFV.

K65R emerged in 16 of 18 experiments with TAF, and in 20 of 22 experiments with TFV (Supplementary Table S1). All four cases where K65R did not emerge were conducted with HIV-1 subtype B viruses. Replicate experiments showed some variability in the time to gain the K65R mutation (subtype B: 53–137 days for TAF and 28–124 days for TFV; subtype C: 18–46 days for TAF and 28–97 days for TFV: subtype B*: 32–53 days for TAF and 28–115 days for TFV) (Supplementary Fig. S1a–c). On average, subtype B viruses acquired K65R after 96 days with TAF and 78 days with TFV (Table 1). For subtype C viruses, K65R emerged after 35 days with TAF and after 48 days with TFV. Subtype B* viruses gained K65R in a similar timeframe as subtype C viruses (42 days with TAF and 65 days with TFV). Overall, K65R emerged earlier in subtype C than in subtype B viruses, although the time to emergence of K65R varied widely within each subtype (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Average time to K65R breakthrough by subtype. Each point on the graph represents one dose escalation selection experiment and which day K65R emergence was initially seen. For the samples where K65R did not emerge, the last day of the selection was used. All TAF and TFV selection experiments are grouped together based on subtype (●: Subtype B AAG,AAA; ■: Subtype B* AAA,AAG; ▲: Subtype C AAA,AAG). TAF, tenofovir alafenamide; TFV, tenofovir.

OPEN IN VIEWER

Table 1.

Time to K65R Emergence in Tenofovir Alafenamide and Tenofovir Resistance Selections

	Time to K65R emergence (days)
	Subtype B		Subtype B *		Subtype C
	TAF	TFV	TAF	TFV	TAF	TFV
Experiment 1	111.0 ± 4.0	60.0 ± 16.0	37.0 ± 2.0	52.0 ± 24.0	28.5 ± 4.5	62.5 ± 15.1
Experiment 2	81.0 ± 28.0	86.7 ± 13.7	46.0 ± 7.0	73.7 ± 22.2	40.8 ± 3.4	33.8 ± 1.8
Average	96.0 ± 14.2	77.8 ± 8.9	41.5 ± 3.9	65.0 ± 15.3	34.6 ± 3.5	48.1 ± 8.9

The numbers shown are the mean time (in days) and standard deviation for the emergence of K65R for each subtype and drug for the two sets of dose escalation selection experiments. Subtype B^*  = Subtype B virus with codons 64–65 similar to subtype C.

TAF, tenofovir alafenamide; TFV, tenofovir.

Although the mentioned dose escalation selection assays were initiated with subtherapeutic concentration of drugs to allow for viral growth and resistance selection, viral breakthrough assays were performed using clinically relevant drug concentrations to mimic clinical drug exposure. In brief, MT-2 cells were infected with the nine subtype B and C viruses as described previously, ¹² with drug exposure started at the time of infection. Four experimental drug conditions were set up (twofold dilutions) to determine the concentration where virus could survive and grow, with the highest concentration (Setup #4; Supplementary Table S2) mimicking physiological conditions. The TAF and TFV concentrations utilized captured the fourfold difference in drug concentration of the active moiety, tenofovir diphosphate (TFV-DP), observed after in vivo dosing of TAF or TDF. ¹

At the physiological drug concentration (Setup #4; 400 nM of TAF and 25 μM of TFV), no viral breakthrough was observed for the duration of the experiment for either drug with all viruses (Supplementary Table S2). At lower concentrations (Setups #3 and #2; 200 or 100 nM of TAF and 12.5 or 6.3 μM of TFV), viral breakthrough was observed after 6 days for all viruses with TFV but not with TAF (since TFV-DP is fourfold higher in the TAF-treated cells). In the experimental setup with the lowest drug concentration (Setup #1; 50 nM of TAF and 3.1 μM of TFV), viral breakthrough was seen for all viruses with both TAF and TFV (Supplementary Table S2). Overall, no differences in time to viral breakthrough were observed between subtypes B and C viruses even at suboptimal levels of tenofovir.

The available clinical data for subtype C and emergence of K65R were also evaluated. A previous analysis for TDF found no evidence of subtype C and K65R emergence, with 10 subtype C patients included in the analysis. ⁶ An updated analysis from 5556 patients in 9 clinical studies with TDF- or TAF-containing regimens examined the emergence of resistance mutations and subtype. Only 1 subtype C patient, out of 77 included in the analysis, was classified as having virologic failure but had no development of resistance. Of note is the small number of subtype C patients (77 out of 5556), which limits the significance of this analysis. Overall, a low incidence (0.3%) of K65R emergence was observed in nonsubtype C patients.

As seen in previous in vitro reports, we observed that subtype C HIV-1 isolates could select K65R earlier than subtype B viruses in dose escalation resistance selections, which correlated with codon usage at positions 64–65 in RT. We also report a large variability in time to emergence of K65R between samples and experiments in subtype B, including subtype B*, confirming the stochastic nature of resistance emergence. Finally, when used at physiological concentrations, both TAF and TFV were able to prevent the breakthrough of the K65R mutation in both subtypes. However, at subtherapeutic concentrations, TAF maintained antiviral activity against subtypes B and C viruses, suggesting that the higher intracellular drug concentrations achieved with TAF may result in reduced resistance development among individuals with lower drug adherence regardless of subtype.