Two classes of drug that inhibit HIV-1 replication by targeting the viral reverse transcriptase (RT) have been developed : the nucleoside reverse transcriptase inhibitors (NRTI) and the non-nucleoside reverse transcriptase inhibitors (NNRTI). The seven NRTI currently approved for treatment of HIV infection in the United States (abacavir, didanosine, lamivudine, stavudine, tenofovir, zalcitabine, and zidovudine) are competitive inhibitors of RT . The three NNRTI approved for treatment of HIV (delavirdine, efavirenz, and nevirapine) are non-competitive inhibitors of RT that occupy a hydrophobic pocket in heterodimeric RT . In the three-dimensional structure of HIV-1 RT, the NNRTI-binding pocket is adjacent to the active site  and it is believed that NNRTI inhibit HIV-1 RT by distorting the normal structural conformation of the enzyme [4,5]. Resistance to the NRTI is associated with substitutions at amino acid positions 40–75, 115–118, 151, 184 and 210–219 in RT [6,7]. NNRTI resistance is associated with substitutions at amino acid positions 97–108, 179–192, and 224–236 of RT . Because NRTI and NNRTI inhibit RT through very different mechanisms, the mutations that cause resistance to one class do not generally affect susceptibility to drugs of the other class. However, there are a few clear examples of mutations in RT that have opposite effects on NRTI and NNRTI susceptibility. For example, the Y181C and L100I mutations that are selected by certain NNRTI can increase susceptibility to zidovudine [9,10] and other NRTI . The close proximity of the NNRTI-binding pocket to the RT active site makes it relatively easy to appreciate how changes in one site can affect the susceptibility of both classes of inhibitor.
During routine commercial testing of clinical HIV-1 isolates for antiretroviral drug susceptibility (PhenoSense HIV; ViroLogic, South San Francisco, California, USA), it was observed that many samples exhibited significant hypersusceptibility to drugs of the NNRTI class compared with that of a ‘wild-type’ reference virus. In many cases, these viruses also displayed reduced susceptibility or resistance to members of the NRTI class of drugs. The purpose of this study was to define NNRTI hypersusceptibility, to assess the prevalence of NNRTI hypersusceptibility among HIV-1 clinical isolates, to explore the relationship between NRTI resistance and NNRTI hypersusceptibility, and to identify the genotypic and phenotypic correlates of NNRTI hypersusceptibility in HIV-1.
Clinical virus isolates
The HIV-1 isolates analyzed in this study were derived from over 17 000 patient plasma samples submitted to ViroLogic Inc. (South San Francisco, California, USA) during 1999–2001 for routine drug resistance testing. Samples were tested either for routine patient management or as part of a clinical trial.
Drug susceptibility was measured using recombinant viruses and a rapid single replication cycle assay, PhenoSense HIV (ViroLogic) . Drug susceptibility results of the the assay are reported as IC50 values (inhibitory concentration of 50%) and as fold-change in susceptibility (patient virus IC50/drug-sensitive reference virus IC50) compared with a wild-type reference virus derived from NL4-3. Viruses with drug susceptibility > 2.5-fold were defined as demonstrating reduced susceptibility to the drug. Viruses with susceptibility < 0.4-fold (reciprocal of 2.5-fold) were defined as being ‘hypersusceptible’ to the drug. Viruses with drug susceptibility between 0.4-fold and 2.5-fold were defined as being ‘sensitive’ to the drug.
The reproducibility of NNRTI drug susceptibility measurements was demonstrated by repeatedly testing (n = 6) a panel of 12 patient-derived virus isolates that exhibited NNRTI drug susceptibilities (fold-change in IC50) ranging from 0.04-fold to 10-fold. Fold-change measurements for each NNRTI were highly reproducible in all viruses tested. Replicate drug susceptibility determinations varied by less than twofold. The mean coefficient of variation of the replicate fold-change results for each NNRTI ranged from 14 to 18% .
DNA sequencing was performed by a thermocycling method using fluorescent dye labeled dideoxynucleotide chain terminator chemistry and reaction products were resolved using a parallel 96 capillary gel electrophoresis system (ABI, Foster City, California, USA). Base calling and amino acid sequence derivations were performed using customized Sequencher software (GeneCodes, Ann Arbor, Michigan, USA). RT amino acid sequences were compared with the HIV-1 reference strain (NL4-3; GenBank Accession number M19921).
Viruses generated by site-directed mutangenesis
The megaprimer method of site-directed mutagenesis  was used to construct a series of isogenic viruses containing specific amino acid substitutions in RT. Mutations were introduced into a 1.5 kb ApaI/PinAI fragment that encodes protease and RT and cloned directly into the resistance test vector . The entire 1.5 kb region was sequenced to confirm all mutations and that no additional mutations were introduced into the vector.
Regression analysis and non-parametric statistics were performed using Stat-View (version 5.0: SAS, Cary, North Carolina, USA) and GraphPad Prism (version 3.0: GraphPad Software, San Diego, California, USA).
Prevalence of NNRTI hypersusceptibility
Phenotypic drug susceptibility testing was performed on 17 697 patient plasma samples between late 1999 and early 2001. Reduced drug susceptibility was common in this dataset and 68% of samples had reductions in susceptibility to one or more NRTI, 53% to an NNRTI and 52% to a protease inhibitor. Prevalence of reduced susceptibility (fold-change > 2.5-fold) ranged from 45% (efavirenz) to 51% (nevirapine) among the NNRTI drugs. NNRTI hypersusceptibility (fold-change < 0.4-fold) was detected in 17% of samples and was slightly higher for efavirenz (10.8%) and delavirdine (10.7%) than for nevirapine (8.0%) (Fig. 1a). The prevalence of viruses hypersusceptible to all three NNRTI was 6.2%. In contrast, hypersusceptibility to NRTI drugs was uncommon, being identified in 3.8% of isolates for zidovudine, but less than 0.03% of isolates for abacavir, didanosine, stavudine, zalcitabine, and lamivudine.
Impact of NRTI treatment history and resistance on NNRTI hypersusceptibility
Drug susceptibility results were available for 778 HIV-1 isolates obtained from NNRTI treatment-naive patients. These isolates could be divided into two groups with respect to prior NRTI exposure: NRTI treatment-naive (n = 331) and NRTI treatment-experienced (n = 447). The prevalence of viruses showing hypersusceptibility to one or more of the NNRTI was significantly higher in the NRTI-experienced group compared with the NRTI-naive group (P < 0.001; Fig. 1b,c). The prevalence of hypersusceptibility to delavirdine, efavirenz, and nevirapine was 5, 9, and 11%, respectively, among the NRTI-naive patients compared with 29, 26, and 21%, respectively, among NRTI-experienced patients (P < 0.001 for all drugs).
To assess further the association between NRTI experience and NNRTI drug susceptibility, the distributions of NRTI and NNRTI susceptibilities were evaluated in each group. In the group of NRTI/NNRTI-naive patients, drug susceptibilities for both drug classes clustered around 1-fold, indicating that the virus isolates were uniformly sensitive to the NRTI and NNRTI drugs (Fig. 1b). However, among the HIV-1 isolates from the NRTI-experienced patients, a significant inverse relationship was observed between NRTI susceptibility and NNRTI susceptibility (Fig. 1c). Increasing drug resistance to NRTI was associated with increasing hypersusceptibility to NNRTI. The associations between NRTI resistance and NNRTI hypersusceptibility in clinical isolates from NRTI-experienced patients are demonstrated by a comparison of zidovudine and efavirenz susceptibilities (rho = 0.485; P < 0.0001). Similar relationships were observed among the other NRTI/NNRTI pairs in the subset of NRTI-experienced/NNRTI-naive patients. In the larger unselected database of all viruses that included both NNRTI-experienced and NNRTI-naive patients, no relationship between NRTI resistance and NNRTI hypersusceptibility was observed (Fig. 1a).
Genotypic associations with NNRTI hypersusceptibility
A dataset containing 2050 clinical HIV-1 isolates with paired phenotypic and genotypic data was used to analyze the genotypic correlates of NNRTI hypersusceptibility. A case–control analysis was performed by selecting a random sample of 100 ‘case’ viruses that were hypersusceptible to all three NNRTI (fold change < 0.4) and 100 ‘control’ viruses that were susceptible to all NNRTI (fold-change 0.7–1.4). The distributions of genotypic mutations for both the NNRTI hypersusceptible and NNRTI wild-type viruses were compared. First, the relative frequencies of common NRTI resistance-associated mutations (M41, D67, T69, K70, L74, V75, M184, L210, T215, and K219) were compared for the two groups (Fig. 2). The NNRTI hypersusceptible viruses had significantly more of these common resistance mutations (mean, 4.3 substitutions per sequence; range, 0–10) than the wild-type viruses (mean, 0.7 substitutions; range, 0–6) (P < 0.0001). Although NNRTI hypersusceptibility was clearly associated with increased numbers of NRTI mutations, nearly 15% of hypersusceptible viruses possessed none or one NRTI mutations, and viruses with as many as six NRTI mutations frequently did not exhibit NNRTI hypersusceptibility.
A second analysis was performed to identify specific amino acid substitutions in RT (amino acids 1–305) that were significantly more frequent in NNRTI-hypersusceptible viruses than NNRTI-susceptible viruses. This analysis included substitutions at 60 positions in RT that were present in at least 5% of the total study group. Mutations at amino acid positions 41, 44, 67, 69, 74, 75, 118, 184, 210, 215, and 219 correlated strongly with NNRTI hypersusceptibility (Table 1). These mutations have been associated with NRTI resistance [6,7]. Other amino acid positions that showed highly significant association (P < 0.01) with NNRTI hypersusceptibility include 39, 43, 203, 208, 211, 228, and 284, which are not typically considered to be NRTI resistance-associated mutations. Mutations at two amino acid positions, 178 and 281, were negatively associated with NNRTI hypersusceptibility.
NNRTI susceptibility in genetically engineered viruses
Testing of genetically engineered viruses containing specific mutations confirmed that NNRTI hypersusceptibility can be caused by combinations of NRTI-resistance mutations (Table 2). The M184V mutation, which causes high-level lamivudine resistance and hypersusceptibility to zidovudine, caused slight increases in susceptibility to all three NNRTI. The M41L and T215Y mutations caused reduced susceptibility to zidovudine and increased susceptibility to delavirdine and efavirenz, but not to nevirapine. The M41L and T215Y mutations in combination with the M184V resulted in a virus with hypersusceptibility to all three NNRTI. NNRTI hypersusceptibility was also demonstrated in viruses possessing the T69SSA insertion plus M41L, A62V and T215Y. Conversely, another multi-NRTI-resistance mutation, Q151M, was not associated with NNRTI hypersusceptibility but did produce a small reduction in susceptibility to the three NNRTI. The K65R substitution reduces susceptibility to lamivudine and increases susceptibility to zidovudine but did not result in NNRTI hypersusceptibility.
Clinical case study of contemporaneous loss of NNRTI hypersusceptibility and NRTI resistance during antiretroviral treatment interruption
Longitudinal plasma samples were collected from an NNRTI-naive patient undergoing a structured treatment interruption after failure of a potent combination therapy regimen containing stavudine, lamivudine, ritonavir, and indinavir (in addition, the patient had extensive prior zidovudine experience) . The virus population from each sample was tested for susceptibility to the NRTI and NNRTI drugs for a period of 15 weeks following discontinuation of therapy (Fig. 3). The baseline, on-treatment, virus demonstrated resistance to zidovudine and lamivudine and hypersusceptibility to nevirapine, efavirenz, and delavirdine. During the treatment interruption, resistance to the NRTI faded to wild-type susceptibility concomitant with a shift in NNRTI susceptibility from hypersusceptible to wild type.
Drug resistance tests are routinely utilized to determine the sensitivity or resistance of HIV-1 to available antiretroviral therapies [2,16]. Numerous prospective and retrospective studies have demonstrated that results of drug resistance tests are associated with significantly better treatment outcomes when the therapy regimen selected contains more drugs to which the patient's virus is sensitive . The current study presents a description of hypersusceptibility of HIV-1 to the NNRTI and sets the stage for asking whether or not this increased susceptibility measured in vitro is predictive of clinical response.
Data generated in the current study show that NNRTI hypersusceptibility is relatively common among viruses from patients who have extensive antiretroviral treatment experience, and much less common in wild-type viruses from treatment-naive patients. In addition, we present data demonstrating a significant inverse correlation between NRTI and NNRTI susceptiblities, a significant association between the presence of NRTI mutations and NNRTI hypersusceptibility, and detection of NNRTI hypersusceptibility in viruses engineered to contain NRTI-resistance mutations. Together, these data strongly suggest that NNRTI hypersusceptibility usually emerges as a consequence of selection by NRTI drug pressure rather than natural variation in HIV-1. Although the presence of certain NRTI-resistance mutations was strongly associated with NNRTI hypersusceptibility, genotypic assessments of resistance mutations do not readily predict NNRTI hypersusceptibility. Some viruses with NNRTI hypersusceptibility had zero or one NRTI resistance-associated mutation while others with numerous NRTI-resistance mutations did not demonstrate NNRTI hypersusceptibility. Interestingly, a number of genetic polymorphisms in HIV-1 RT were also significantly correlated with the presence (or absence) of NNRTI hypersusceptibility. These observations are consistent with the hypothesis that NNRTI hypersusceptibility is a consequence of alterations in the overall enzyme structure owing to the complex interplay among drug-resistance mutations and/or polymorphic changes in RT. Additional studies will be required to define more precisely a genotypic algorithm that better predicts NNRTI hypersusceptibility.
The biochemical explanation for the association between NNRTI hypersusceptibility and NRTI resistance is currently a matter of conjecture. The three-dimensional structure of HIV-1 RT, places the NNRTI-binding pocket in close proximity to the active site , suggesting a structural explanation for NNRTI hypersusceptibility. Mutations in HIV-1 that cause reduced susceptibility to one drug and increased susceptibility to another drug have been described previously. Examples of such RT mutations include M184V , L74V , and Y181C  or L100I , which are associated with resistance to lamivudine, zalcitabine/didanosine, and NNRTI drugs, respectively, but enhanced sensitivity to zidovudine. Mutations that confer resistance to the NRTI, either by directly altering the active site of RT or by complex changes that affect enzymatic function (e.g., phosphorolysis) , may exert structural changes on the contiguous NNRTI-binding pocket, in some cases leading to enhanced affinity for the NNRTI drugs.
Following the initial description of NNRTI hypersusceptibility by our group , four independent clinical trials have reported a significant association between hypersusceptibility to a NNRTI (as measured by PhenoSense HIV) and improved virologic outcomes (Haubrich et al. accompanying manuscript, [22–24]). The accompanying paper by Haubrich, et al. in this issue of AIDS presents an analysis of the association between baseline NNRTI hypersusceptibility and treatment outcomes and discusses these clinical studies in greater detail.
Perhaps more importantly, accumulating data suggest potential clinical benefit of NNRTI hypersusceptibility during treatment of NRTI-experienced patients with NNRTI-based regimens. These findings may provide additional opportunities to use and sequence the currently available antiretroviral drugs in alternative treatment strategies to achieve improved patient outcomes. Additional clinical studies will be required to define fully the implications of NNRTI hypersusceptibility for optimum patient management.
1.Parniak MA, Sluis-Cremer N. Inhibitors of HIV-1 reverse transcriptase. Adv Pharmacol
2.Carpenter CC, Cooper DA, Fischl MA et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society–USA Panel. Jama
3.Kohlstaedt LA, Wang J, Friedman JM, Rice PA, Steitz TA. Crystal structure at 3.5 Å resolution of HIV-1 reverse transcriptase complexed with an inhibitor. Science
4.Spence RA, Kati WM, Anderson KS, Johnson KA. Mechanism of inhibition of HIV-1 reverse transcriptase by nonnucleoside inhibitors. Science
5.Rittinger K, Divita G, Goody RS. Human immunodeficiency virus reverse transcriptase substrate-induced conformational changes and the mechanism of inhibition by nonnucleoside inhibitors. Proc Natl Acad Sci USA
6.D'Aquila RT, Shapiro JM, Brun-Vezinet F et al. Update on drug resistance mutations in HIV-1. Topics HIV Med
7.Loveday C. Nucleoside reverse transcriptase inhibitor resistance. J Acquir Immune Defic Syndr
8.Tantillo C, Ding J, Jacobo-Molina A et al. Locations of anti-AIDS drug binding sites and resistance mutations in the three-dimensional structure of HIV-1 reverse transcriptase. Implications for mechanisms of drug inhibition and resistance. J Mol Biol
9.Byrnes VW, Emini EA, Schleif WA et al. Susceptibilities of human immundeficiency virus type 1 enzyme and viral variants expressing multiple resistance-engendering amino acid substitutions to reverse transcriptase inhibitors. Antimicrob Agents Chemother
10.Larder BA. 3′-Azido-3′-deoxythymidine resistance suppressed by a mutation conferring human immunodeficiency virus type 1 resistance to nonnucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother
11.van Laethem K, Witvrouw M, Pannecouque C et al. Mutations in the non-nucleoside binding-pocket interfere with the multi- nucleoside resistance phenotype. AIDS
12.Petropoulos CJ, Parkin NT, Limoli KL et al. A novel phenotypic drug susceptibility assay for human immunodeficiency virus type 1. Antimicrob Agents Chemother
13.Huang W, Paxinos E, Petropoulos C, Whitcomb J. Measuring moderate differences in baseline NNRTI susceptibility. Antiviral Ther
14.Sarkar G, Sommer SS. The ‘megaprimer’ method of site-directed mutagenesis. Biotechniques
15.Deeks SG, Wrin T, Liegler T et al. Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viremia. N Engl J Med
16.Demeter L, Haubrich R. International perspectives on antiretroviral resistance. Phenotypic and genotypic resistance assays: methodology, reliability, and interpretations. J Acquir Immune Defic Syndr
17.Hellmann NS. The role of phenotypic drug susceptibility testing in HIV treatment decisions. Curr Opin Anti-Infect Invest Drugs
18.Tisdale M, Kemp SD, Parry NR, Larder BA. Rapid in vitro selection of human immunodeficiency virus type 1 resistant to 3′-thiacytidine inhibitors due to a mutation in the YMDD region of reverse transcriptase. Proc Natl Acad USA
19.St Clair M, Martin J, Tudor-Williams G et al. Resistance to ddI and sensitivity to AZT induced by a mutation in HIV-1 reverse transcriptase. Science
20.Boyer PL, Sarafianos SG, Arnold E, Hughes SH. Selective excision of AZTMP by drug-resistant human immunodeficiency virus reverse transcriptase. J Virol
21.Whitcomb J, Deeks S, Huang W et al. Reduced susceptibility to NRTI is associated with NNRTI hypersensitivity in virus from HIV-1 infected patients. Seventh Conference on Retroviruses and Opportunistic Infections.
San Francisco, January, 2000 [abstract 234].
22.Shulman N, Zolopa AR, Passaro D et al. Phenotypic hypersusceptibility to non-nucleoside reverse transcriptase inhibitors in treatment-experienced HIV-infected patients: impact on virological response to efavirenz-based therapy. AIDS
23.Katzenstein D, Bosch RJ, Wang N for the ACTG 364 Study Team. Baseline phenotypic susceptibility and virologic failure over 144 weeks among nucleoside RT inhibitor experienced subjects in ACTG 364. Ninth Conference on Retroviruses and Opportunistic Infections.
Seattle, February 2002 [abstract 591T].
24.Mellors J, Bennett FVK, Hellmann NS for the ACTG 398 Study Team. Efavirenz hypersusceptibility improves virologic response to multidrug salvage regimens in ACTG 398. Ninth Conference on Retroviruses and Opportunistic Infections
2002. Seattle, February 2002 [abstract 45].