Human Immunodeficiency Virus Type 1 Reverse Transcriptase Mutation Selection during In Vitro Exposure to Tenofovir Alone or Combined with Abacavir or Lamivudine

Abstract

Mutations selected or deselected during passage of human immunodeficiency virus strain HXB2 or resistant variants with tenofovir (TFV), abacavir (ABC), and lamivudine (3TC) differed depending on the drug combination and virus genotype. In the wild-type virus, TFV-ABC and TFV-3TC selected K65R (with reduced susceptibility to all three inhibitors) and then Y115F. TFV-containing regimens might increase K65R selection, which confers multiple nucleoside reverse transcriptase inhibitor resistance.

Resistant human immunodeficiency virus type 1 strains may be selected in patients undergoing antiretroviral therapy (10) or by in vitro selection with increasing inhibitor concentrations (24, 25). Abacavir (ABC) (6) and lamivudine (3TC) (12) are nucleoside reverse transcriptase inhibitors (NRTI), while tenofovir (TFV) is a nucleotide analogue (3, 22) but is grouped with the NRTI on the basis of its mode of action (7). Substitutions in reverse transcriptase (RT) with reduced susceptibility to ABC include K65R, L74V, Y115F, and M184V, but alone they give only a modest (threefold) reduction in ABC susceptibility (23). M184V, generally selected first, does not appear to influence the clinical response to ABC (11, 14). The most significant mutations for cross-resistance to ABC are multiple thymidine analogue mutations (TAMs) when present with M184V (14, 23). For 3TC, selection of M184V or M184I alone gives high-level resistance (12, 24, 26). Maintenance of the M184V mutation may confer a clinical benefit by delaying TAMs (2) and reducing replicative capacity (13, 18). For TFV, K65R was the only resistance-conferring mutation selected in vitro and in vivo (25; M. D. Miller, N. A. Margot, D. J. McColl, T. Wrin, D. F. Coakley, and A. K. Cheng, XII Int. HIV Drug Resist. Workshop, abstr. 135, 2003), with a three- to fourfold susceptibility reduction, and confers cross-resistance to ABC, 2′,3′-dideoxyinosine, 2′,3′-dideoxycytidine, dioxolane (−)-β-d-dioxolane-guanosine, and 3TC (4, 8-10, 20, 23, 25). Clinical trials with the prodrug tenofovir disoproxil fumarate (TDF) demonstrated significantly reduced antiviral response in the presence of three TAMs including M41L or L210W (19). In contrast, M184V increases virus susceptibility to TFV (17, 18).

Resistance selection in the presence of TFV was studied in vitro both alone and in drug combinations against virus with and without NRTI resistance-conferring mutations. Separate passage series were performed with MT-4 cells and wild-type virus (HXB2) and resistant variants (HXB2_{41L,184V,215Y} and HXB2_65R,74V,184V) as described previously (23, 24). The first passage used drug concentrations that were two and four times the 50% inhibitory concentration (IC₅₀). Virus supernatants were taken for further passage, and drug concentrations were maintained and increased exponentially (two- and fourfold) for subsequent passages. Infected-cell pellets were used to sequence the RT region including codons 1 to 400 by ABI 3700 technology. The phenotypes of selected viruses were determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay with MT-4 cells (5).

Details of passage series, mutations observed, and virus susceptibility at different passages are shown in Table 1. The IC₅₀s of all three drugs against HXB2 were similar, ranging from 1.9 to 2.6 μM. The IC₅₀s for mutant viruses HXB2_{M41L,M184V,T215Y} and HXB2_{K65R,L74V,M184V} showed 8- and 16-fold reduced susceptibility to ABC, respectively, and 1.8- and 1.2-fold reduced susceptibility to TFV, respectively.

TABLE 1.

Relative fold change in susceptibility to HXB2 of molecular clones and variants derived from passage in the presence of RT inhibitors

Starting virus RT inhibitor(s) (passage no.; drug concn[s] [μM]a)	Passage variant(s)	Fold resistance
		ABC	TFV	3TC	ZDV
Wild-type virus (HXB2)	HXB (wild type)	1.0	1.0	1.0	1.0
ABC (4; 36)	M184V	3.1, 4.3	0.6, 0.3
ABC (2; 36)	M184V	3.8, 5.8	0.8, 0.6	>34b	0.3b
ABC (4; 73)	Y115F, M184V	9.2, 11.5	1.2, 0.8		0.6, 0.3
ABC (8; 73)	K65k/r, L741/v, Y115F, M184V	>20, >20	1.6	>12, >30	0.4, 1.3
3TC (3; 60)	M1841	3.2b	0.6b	>80, >30	<0.2b
ABC-TFV (5; 18, 20)	K65R	4.3, 7.6	6.2, 5.0
ABC-TFV (7; 18, 20)	K65R, Y115F	8.5, 8.1	6.6, 5.0	>12, >30	4.3, 3.0
3TC-TFV (3; 15, 21)	K65R	4.2, 4.2	3.3b	7.7, 5.5	0.5b
3TC-TFV (7; 30, 41)	K65R, Y115Fc	7.5, 8.0	4.7, 6.2	>12, >30	1.2, 1.4
3TC-TFV (7; 30, 41)	K65R, Y115Fc	9.0, >9.0	5.4, 10.2	>12, >30	2.1, 2.5
TFV (5; 41)	A62a/t, K65R	3.7, 2.6	4.2, 4.1
ZDV-3TC-resistant virus	M41L, M184V, T215Y	8.5, 7.2	1.8, 2.0		8.1, 1.9
ABC (5; 77)	M41L, Y115F, M184V, T215Y	12.4, 20.2	2.1, 1.2		3.8, 1.8
ABC (7; 77)	M41L, L74V, Y115F, M184V, T215Y	>20, >20	1.0, 2.7	>12, >30	13.0, 13.7
ABC-TFV (4; 39, 9)	M41L, M184V, T215Y, N348I	6.5, 13.6	1.4, 2.3		6.2, 3.8
ABC-TFV (6; 39, 9)	M41L, Y115F, M184V, T215Y	>11, 11	2.4, 3.1	>12, >30	62.7, 20.4
TFV (5; 74)	M41L, T215Y	2.8, 2.7	4.2, 1.8		102b
ABC-TFV-resistant virus	V21I, K65k/r, L74V, M184V	16.7, 9.4, 12.4	1.2, 1.6, 1.7
ABC (6; 76)	V21I, K65R, L74V, Y115F, M184V	>19, >40	2.9, 2.8		1.0, 0.3
ABC-TFV (3; 76, 12)	V21I, K65R, L74V, Y115F, M184V	>40, >50	2.0, 2.6
ABC-TFV (5; 76, 25)	V21v/l, K65R, S68s/l, L74V, Y115F, M184V	>19, >60	2.4, 4.3
TFV (3; 49)	K65R, S68R, L74V, Y115F, M184V	>19, 42	2.5, 4.1
TFV (5; 49)	K65R, S68R, L74V, Y115F	13.7, 12.0	7.2, 7.8		1.0, 2.3

^aDrug concentrations are listed in same order as inhibitors.

^bAssay not repeated.

^cViruses are from different passage series.

Passage of the wild-type virus with individual inhibitors confirmed previous findings (23-25). In all combinations with 3TC-TFV or ABC-TFV, M184I or M184V was not observed despite strong selective pressure, particularly from 3TC, as shown by the early selection of M184I (Table 1). This finding is in keeping with previous reports (24). Instead, the virus acquired first K65R, with three- to sixfold reductions in susceptibility to all three drugs, and then Y115F. The K65R-Y115F genotype showed small increases in resistance to ABC and TFV and was above the assay cutoff for 3TC. This profile might be favored over M184V because K65R reduced susceptibility to both inhibitors and avoided M184V-mediated TFV hypersusceptibility. This might in part explain the relatively high frequency of selection of K65R (24% of virologic failures) during clinical trial GS-903, when antiretroviral therapy-naive subjects were treated with 3TC-TDF-EFV (Miller et al., XII Int. HIV Drug Resist. Workshop). However, in this trial, M184V was also selected, which highlights some limitations of in vitro selection experiments. The presence of K65R reduces both replication capacity and RT processivity in vitro (27) and confers increased susceptibility to ZDV (J. W. Mellors, H. Bazmi, C. K. Chu, and R. F. Schinazi, Fifth Int. Workshop HIV Drug Resist., abstr. 7, 1996). Thus, while TFV promotes the selection of K65R, ZDV might contribute to its deselection.

Of particular interest with TFV-3TC pressure on wild-type virus was the selection of Y115F after K65R. Previously, this mutation has been observed uncommonly in subjects treated with ABC (20). Thus, Y115F might confer a selective advantage on the virus when it is exposed to TFV. However, Y115F had not been observed during TDF clinical trials, except in a recently reported trial with TFV-ABC-3TC (J. E. Gallant, A. E. Rodriguez, W. Weinberg, B. Young, D. Berger, M. L. Lim, Q. Liao, L. Ross, L. J. Johnson, and M. S. Shaefer, 43rd Intersci. Conf. Antimicrob. Agents Chemother., abstr. H-1722a, 2003).

The HXB2_{M41L,M184V,T215Y} genotype did not change during passage in the absence of an inhibitor. Passage with TFV-ABC or ABC alone selected the Y115F mutation, which resulted in a further decrease in susceptibility to ABC but not in susceptibility to TFV. Passage with TFV alone resulted in rapid deselection of M184V and reduced susceptibility (mean, 1.9- to 2.7-fold). Mean resistance to ABC decreased to 2.7-fold, compared to 7.8-fold for the original variant.

The HXB2_{K65R,L74V,M184V} genotype lost K65R in the absence of inhibitors. ABC selection pressure in vitro and in the clinic has not been found to produce the K65R-L74V genotype other than in a mixture (1, 23), and so these mutations together might produce an unstable genotype. Passage with TFV, ABC, or TFV-ABC retained K65R. Again, TFV-ABC or ABC selected Y115F, and TFV selected S68R and Y115F but deselected M184V. The combination of K65R, S68R, L74V, and Y115F resulted in 7.5- and 12.9-fold mean increases in resistance to TFV and ABC, respectively.

Whenever present, TFV alone always deselected M184V. Studies with simian immunodeficiency virus also demonstrated deselection of M184V with TFV-3TC in vitro and in vivo (21). The present studies with human immunodeficiency virus type 1 show that, in the context of multiple RT mutations, the combination of ABC with TFV was sufficient to maintain the M184V mutation in vitro. In TDF clinical trial GS-902 (17), most of the subjects enrolled with virus containing M184V, many maintained 3TC therapy, and M184V was retained. Overall, there might be a potential benefit of TDF treatment for subjects possessing M184V mutant virus. Furthermore, the clinical efficacy of TDF in another trial showed improved response in subjects with M184V (19). This is similar to ZDV, where M184V with TAMs partially reverses ZDV resistance with a clinical benefit (15).

Interestingly, with TAMs present, TFV alone did not select K65R, suggesting that TAMs alone may confer a sufficient selective advantage. This finding is supported by data obtained with treatment-experienced subjects (17), in whom TFV only rarely selected for K65R (2%). While the presence of K65R conferred only a three- to fourfold increase in resistance to TFV (17), this might be clinically significant because the clinical cutoff for a reduced response to this compound is 1.4-fold (16).

These in vitro data suggest that combination of TDF with ABC and/or 3TC might lead to selection of K65R in the clinic. In recently reported clinical trials with antiretroviral therapy-naive subjects treated with TDF-ABC-3TC, K65R with M184V was selected rapidly and led to virological failure (C. Farthing, H. Khanlou, and V. Yeh, 2nd IAS Conf. HIV Pathogenesis Treatment, abstr. 43, 2003; Gallant et al., 43rd Int. Conf. Antimicrob. Agents Chemother.). The K65R mutation confers reduced susceptibility to most NRTI and may severely restrict future options for therapy (U. Parikh, D. Koontz, J. Hammond, L. Bacheler, R. Schinazi, P. Meyer, W. Scott, and J. Mellors, XII Int. HIV Drug Resist. Workshop, abstr. 136, 2003).

These in vitro studies with dual-drug combinations and with different starting genotypes have offered additional insight into the potential genotypic adaptability of the virus in response to different drug combinations.