|Home | About | Journals | Submit | Contact Us | Français|
The human immunodeficiency virus type 1 (HIV-1) integrase mutations N155H and Q148R(H)(K) that reduce susceptibility to the integrase inhibitor raltegravir have been identified in patients failing treatment regimens containing raltegravir. Whether these resistance mutations occur individually or in combination within a single virus genome has not been defined, nor do we fully understand the impact of these primary mutations and other secondary mutations on raltegravir susceptibility and viral replication capacity. To address these important questions, we investigated the raltegravir susceptibility and replication capacity of viruses containing mutations at positions 155 and 148 separately or in combination with secondary mutations selected in subjects failing treatment regimens containing raltegravir. Clonal analysis demonstrated that N155H and Q148R(H)(K) occur independently, not in combination. Viruses containing a Q148R(H)(K) mutation generally displayed larger reductions in raltegravir susceptibility than viruses with an N155H mutation. Analysis of site-directed mutants indicated that E92Q in combination with N155H resulted in a higher level of resistance to raltegravir than N155H alone. Viruses containing a Q148R(H) mutation together with a G140S mutation were more resistant to raltegravir than viruses containing a Q148R(H) mutation alone; however, viruses containing G140S and Q148K were more susceptible to raltegravir than viruses containing a Q148K mutation alone. Both N155H and Q148R(H)(K) mutations reduced the replication capacity, while the addition of secondary mutations either improved or reduced the replication capacity depending on the primary mutation. This study demonstrates distinct genetic pathways to resistance in subjects failing raltegravir regimens and defines the effects of primary and secondary resistance mutations on raltegravir susceptibility and replication capacity.
Antiretroviral drug combinations have proven efficacious in the treatment of human immunodeficiency virus type 1 (HIV-1). To date, there are five different drug classes approved for use in clinics. The majority of drugs target the viral protease or reverse transcriptase (RT). More recently, drugs that target virus entry and integration steps of the HIV-1 life cycle have received regulatory approval, including the entry inhibitors enfurvitide (15) and maraviroc (16), and the integrase (IN) inhibitor raltegravir (RAL) (5). Many HIV-1-infected individuals develop resistance to antiretroviral therapy; thus, the increasing number of drug classes with distinct targets and mechanisms of action provides much-needed new options for the treatment of HIV-1 infection.
Provirus formation is a critical step in the HIV-1 life cycle and is mediated by the virally encoded IN protein, one of the three essential enzymes encoded by the HIV-1 pol gene. IN is a 32-kDa protein consisting of 288 amino acids and comprised of the following three functional domains: the N-terminal zinc binding domain, the central catalytic core domain, and the C-terminal DNA binding domain (8). IN catalyzes two main steps of integration, 3′ processing and strand transfer. 3′ processing involves the removal of a dinucleotide from both 3′ ends of the double-stranded viral DNA after reverse transcription of the viral genomic RNA template (2, 3). Subsequently, strand transfer occurs in the nucleus and involves the joining of the 3′ ends of viral DNA to the 5′ ends of the host genome (1, 9, 10). Finally, the expression of viral genes following integration results in the generation of new virions. Current IN drug candidates, including RAL, target the strand transfer process (18, 24, 27).
Consistent with the other drug classes approved for the treatment of HIV-1, the development of resistance to RAL is associated with treatment failure, i.e., partial or complete loss of viral suppression. Therefore, it is important to understand the impact of RAL-selected resistance mutations on both the susceptibility and IN-mediated replication capacity (RC) of patient viruses, which provide additional information regarding treatment decisions and future drug discovery in this new class of antiretrovirals. Resistance to IN inhibitors involves the accumulation of mutations located primarily within the IN catalytic core domain (12, 13, 17-19, 26). Early IN inhibitor candidates that targeted the strand transfer step, like the diketo acids and naphthyridine carboxamides, were structurally distinct and gave rise to nonoverlapping patterns of resistance mutations in vitro (17). Diketo acid derivatives selected mutations mainly at positions 66, 153, and 154 of IN, while the naphthyridine carboxamide (L-870,810) selected mutations at positions 72, 121, 125, and 151, with the exception of mutations at position 155 which reduced susceptibility to both compounds. RAL, the first approved IN strand transfer inhibitor, is a hydroxypyrimidinone carboxamide (MK-0518) and structurally distinct from the diketo acid derivatives and naphthyridine carboxamides (11). In addition, elvitegravir (GS-9137), a quinolone-based strand transfer inhibitor, is currently being evaluated in phase III clinical trials (7).
In the RAL phase III trials (BENCHMRK I and II), viruses from subjects who failed regimens containing RAL had primary IN mutations at positions 148 and 155 and to a lesser extent 143 (5). Secondary mutations were also observed, including L74I or L74M [L74I(M)], E92Q, T97A, E138K, G140S or G140A [G140S(A)], V151I, G163R, I203M, and S230R, as described previously (D. Cooper and B. Nguyen, presented at the 14th Conference on Retroviruses and Opportunistic Infections, Los Angeles, CA, 25 to 28 February 2007). In this study, we investigated the relationships between phenotypic alterations in RAL susceptibility and RAL-associated resistance mutations of patient samples from the BENCHMRK I and II studies by examining the following: (i) the genetic linkage of RAL primary resistance mutations at positions 148 and 155, (ii) the genotypic relationship of RAL primary and secondary mutations, and (iii) the impact of primary and secondary RAL mutations on RAL susceptibility and IN-mediated RC.
Baseline and first RAL treatment failure samples from a subset of subjects (n = 69) participating in the BENCHMRK I and II studies were analyzed. BENCHMRK study participants were comprised of subjects infected with triple-class drug-resistant HIV-1 and failing their current antiretroviral therapy. The investigation described here was conducted only for subjects containing viruses with RAL-associated resistance mutations detected at early failure time points; the subjects failing treatment without RAL resistance mutations were not included in this study.
RAL susceptibility and IN-mediated RC were assessed using the PhenoSense integrase assay, a new phenotypic assay built on the PhenoSense technology platform initially created to determine susceptibility to protease and RT inhibitors (23). The PhenoSense integrase assay has been validated for patient management applications in compliance with the College of American Pathologists and Clinical Laboratories Improvement Amendments specifications. Briefly, 1.6 kb of the HIV-1 pol sequence containing the C-terminal domain of RT, RNase H, and IN was amplified from patient plasma by RT-PCR and transferred into a resistant test vector (RTV) containing a luciferase reporter gene. Cotransfections of HEK293 cells with IN-specific RTVs and an amphotropic murine leukemia virus envelope expression vector were performed to produce pseudovirus stocks that contain patient-derived IN sequences. Virus stocks were used to infect fresh HEK293 cells in the absence or presence of serial dilutions of IN inhibitor. Susceptibility to an IN inhibitor was calculated by plotting the percent inhibition of virus replication (luciferase activity) versus the log10 drug concentration to derive the drug concentration required to inhibit virus replication by 50% (IC50). The fold change (FC) in drug susceptibility was calculated by comparing the IC50 of the sample virus to the IC50 of a wild-type reference strain (NL4-3). Validation studies have determined the reproducibility of FC values to be within twofold. IN-mediated RC was determined and expressed as a percentage of viral infectivity (luciferase activity) in the absence of drug relative to the wild-type reference strain. The nucleotide sequences of IN coding regions were determined using conventional dye-deoxy chain terminator chemistry (ABI, Foster City, CA). Differences in nucleotide and deduced amino acid sequences were recorded relative to the NL4-3 reference sequence.
Molecular IN clones were isolated from a subset of patient viral populations containing mixtures at positions known to confer resistance to RAL (codons 155 and 148). Specifically, for each sample, IN RTV plasmid DNA preparations representing viral IN sequence populations were used to retransform bacteria, and 48 or more individual clones were obtained from each virus population. IN coding regions of each clone were sequenced, and a subset of clones from each population was tested for RAL susceptibility and IN-mediated RC using the PhenoSense integrase assay.
Single or double mutations were introduced into the IN coding region of HIV-1 IIIB using site-directed mutagenesis. The following panel of site-directed mutants (SDMs) was transferred into IN RTVs and tested for susceptibility to RAL and IN-mediated RC: N155H, Q148R, Q148H, Q148K, N155H and E92Q, Q148R and G140S, Q148R and G140A, Q148R and E138K, Q148H and G140S, Q148H and G140A, Q148H and E138K, Q148K and G140S, Q148K and G140A, and Q148K and E138K.
A paired t test was performed to compare the IN-mediated RC of viruses at baseline and treatment failure using GraphPad Prism version 4.03.
The subset of virologic failure subjects (n = 69) from the BENCHMRK I and II studies evaluated here was limited to the subjects containing viruses with RAL-associated resistance mutations detected at early failure time points. The majority of subjects failed with a primary mutation at either position 155 or 148 (n = 54) or a combination of the two (n = 10). A small number of subjects failed with another primary resistance mutation at position 143 only (n = 4). Table Table11 presents a description of the genotypes of the 69 viruses studied, detailing primary mutations at position 155, 148, or 143, and the secondary mutations E92Q and G140S(A). The secondary mutation E92Q was observed in viruses containing N155H alone or N155H in combination with Q148R or Q148H or Q148K [Q148R(H)(K)] but never in viruses containing Q148R(H)(K) alone. Conversely, the secondary mutation G140S(A) was observed in viruses containing Q148R(H)(K) alone or Q148R(H)(K) in combination with N155H but never in viruses containing N155H alone. Furthermore, viruses that contained primary mutations at both positions 155 and 148 were observed only as mixed virus populations (n = 10), and notably in all 10 cases, both E92Q and G140S(A) mutations were either present as mixtures or absent. Thus, from this genotypic sorting, it appeared that mutations at positions 155 and 148 may be mutually exclusive and that additional mutations at positions 92 and 140 may also segregate, with E92Q associated with mutations at position 155 and G140S(A) associated with mutations at position 148 (Table (Table1).1). The association of mutations at position 143 with E92Q or G140S(A) was not investigated in this study as the number of samples containing 143 mutations was small. The one virus that lacked mutations at positions 155, 148, and 143 contained a mixture of E and Q at position 92 (Table (Table1).1). In some subjects, additional resistance-associated mutations, such as L74I(M), T97A, E138K(A), V151I, G163R, I203M, and S230R, were observed at the failure time point. Some of these mutations (L74I, T97A, E138K, V151I, G163R, and I203M) were also identified as baseline polymorphisms, which is consistent with findings in a previous study (25).
To determine whether viral populations containing 155 and 148 mutations were comprised of distinct subpopulations with either N155H or Q148R(H)(K), we characterized 20 to 30 molecular clones each from four virus populations that contained mixed amino acid sequences at positions 148 and 155 (Table (Table2).2). Two subjects had N155N/H and Q148Q/R, one subject had N155N/H and Q148Q/H, and one subject contained N155N/H and Q148Q/K. Mutations at position 92 and/or 140 were also observed in three out of four subjects. For each subject, the amino acid in the viral population at each of these four positions (155, 148, 92, and 140) was either wild type or a mixture of wild type and mutant. Clones isolated from each viral population bearing different substitutions at these four positions are shown in Table Table2.2. In all cases, the clones containing N155H had a wild-type amino acid (Q) at position 148. Conversely, all clones containing a mutation at position 148 (R, H, or K) contained a wild-type amino acid (N) at position 155. Clones containing the E92Q mutation were identified in two subjects (subjects 7 and 10) in the presence or absence of N155H. However, E92Q was not found in clones containing 148 mutations. In the subjects studied here, all clones bearing mutations at position 140 had mutations at position 148 but not 155. In summary, the mutations at position 155 occurred independently from mutations at position 148. The secondary mutation E92Q was preferentially selected in viruses bearing the N155H mutation, while the G140S(A) mutation was preferentially selected in viruses bearing 148 mutations.
We measured phenotypic susceptibility to RAL in the 28 subjects containing N155H and the 14 subjects containing Q148R or Q148H (Fig. (Fig.1).1). The range of reductions in RAL susceptibility was broad in subjects with N155H (median FC of 33; range of 5 to >150) versus those with Q148R(H), which in most cases showed larger reductions. Twelve of 14 viruses with mutations at position 148 had FC values exceeding the maximal FC measurable in the assay (>150-fold). Notably, 13/14 viruses with mutations at position 148 had a secondary mutation [E138K and/or G140S(A)], and 22/28 viruses with N155H contained at least one secondary mutation (L74M, E92Q, T97A, V151I, G163R, and I203M). The one subject with a virus containing a mutation at position 148 (Q148R) without any other resistance-associated mutations displayed a smaller reduction in RAL susceptibility (FC = 21).
To examine the association between RAL resistance and IN-mediated RC, we compared the IN-mediated RC of baseline and paired treatment failure virus in 42 study subjects with viruses containing a mutation at position 155 or 148 of IN (Fig. (Fig.2).2). Relative to the paired baseline viruses, reductions in IN-mediated RC were observed at the time of the first treatment failure in 37 of 42 subjects (88%; P < 0.0001; paired t test) (Fig. (Fig.2A).2A). We also compared changes in IN-mediated RC between baseline and failure viruses for different mutational pathways and observed decreased IN-mediated RC in 26/28 viruses with a mutation at position 155 (P < 0.0001) and 11/14 viruses with a mutation at position 148 (P = 0.0023) (Fig. (Fig.2B).2B). Interestingly, when we examined IN-mediated RC at position 148 more closely, we observed that viruses with Q148R displayed a greater reduction in IN-mediated RC (P = 0.01) than viruses containing Q148H (nonsignificant; P = 0.07) (Fig. (Fig.2C2C).
We next verified the effect of N155H and Q148R(H)(K) alone and in combination with other mutations on RAL susceptibility and IN-mediated RC by using SDMs (Table (Table3).3). The N155H mutation alone resulted in a 16-fold reduction in RAL susceptibility and was further reduced to >150-fold when in combination with E92Q. Relative to the reference virus, the N155H mutation alone reduced IN-mediated RC to approximately 70%, and the combination of N155H and E92Q resulted in a further reduction in IN-mediated RC to 50% (Table (Table3).3). Conversely, the effect of mutations at position 148 alone and in combination with G140S on RAL susceptibility varied depending on which mutation was selected at position 148. Table Table33 summarizes the RAL FC and IN-mediated RC for all three different mutations at position 148. The addition of Q148R, H, or K resulted in reductions in RAL susceptibility (34-fold for Q148R, 18-fold for Q148H, and 43-fold for Q148K). Similar results of reduction in RAL susceptibility of mutations at positions 148 and 155 have been observed in a previous study (14). The addition of G140S to Q148R or Q148H further decreased susceptibility to RAL (>150-fold). In contrast, the addition of G140S to Q148K increased susceptibility to RAL relative to Q148K alone (sixfold versus 43-fold). For all three mutations at position 148, the IN-mediated RC was reduced relative to the wild-type reference (Q148R, 59%; Q148H, 43%; and Q148K, 32%). The addition of G140S to Q148R had no effect on IN-mediated RC relative to Q148R alone. However, increases in IN-mediated RC were observed with the addition of G140S to Q148H (99%) or Q148K (67%).
Due to the different effects observed on both RAL susceptibility and IN-mediated RC for combinations of Q148R(H)(K) with G140S, we also tested combinations of these three mutations at position 148 with two other previously reported RAL-selected secondary mutations, G140A and E138K (D. Cooper and B. Nguyen, presented at the 14th Conference on Retroviruses and Opportunistic Infections, Los Angeles, CA, 25 to 28 February 2007). In Table Table3,3, the double mutant E138K and Q148R further reduced susceptibility to RAL relative to Q148R alone (79-fold versus 34-fold), and the double mutant G140A and Q148R produced even greater reductions in susceptibility (>150-fold). Both of these secondary mutations (E138K and G140A) had no effect on IN-mediated RC, relative to the single mutant Q148R, similar to G140S. However, the double mutant G140A and Q148H conferred high-level resistance to RAL (>150-fold) and a marked reduction in IN-mediated RC relative to Q148H alone (10% versus 43%), while the addition of E138K had no effect on RAL susceptibility or IN-mediated RC relative to Q148H alone. These results contrasted with the increased IN-mediated RC observed with the G140S and Q148H mutant, relative to Q148H alone. Both secondary mutations E138K and G140A, when combined with Q148K, displayed high-level resistance (>150-fold), but only E138K and Q148K increased IN-mediated RC relative to Q148K alone (71% versus 32%), similar to results observed with G140S and Q148K. Taken together, the effect of 148 mutations on RAL susceptibility and IN-mediated RC may differ dramatically in patient viruses depending on the specific mutation at position 148 (R, H, or K), as well as the secondary mutation selected.
In this study, we have demonstrated that the IN mutations N155H and Q148R(H)(K) are independently selected in HIV-1 from subjects failing RAL-containing treatment regimens. Extensive clonal analysis of four virus populations with mixed amino acid substitutions detected at positions 155 (N155N/H) and 148 (Q148Q/R, Q148Q/H, Q148Q/K) conclusively demonstrated that mutations at positions 155 and 148 do not coexist in the same viral genome. In addition, the secondary mutation E92Q preferentially resides on virus clones bearing the N155H mutation while G140S(A) is selected by viruses bearing mutations at position 148. In this study population (a subset of virologic failures from BENCHMRK I and II), the identification of E92Q in the absence of N155H was rare (1/69 RAL failures), consistent with other reports (4, 20). We also observed primary IN mutations at position 143, albeit at a low frequency (4/69 or 6%) compared to the selection of mutations at position 155 only (31/69 or 45%) and 148 only (R, H, K; 20/69 or 29%) at the time of the first virologic failure.
In the subjects studied here, the RAL susceptibility of viruses bearing N155H varied from fivefold to >150-fold (average, 33-fold), whereas viruses bearing mutations at position 148 generally exhibited large reductions in RAL susceptibility, with the majority of viruses exceeding the upper limit of quantification of our phenotypic susceptibility assay. All RAL treatment failure viruses with mutations at position 148 had an additional mutations, either G140S(A) or E138K, except for one virus which contained Q148R alone. Similarly, the majority of viruses containing N155H (22/28) also had at least one secondary mutation, L74M, E92Q, T97A, V151I, or G163R. The effect of the secondary mutations in combination with mutations at position 155 or 148 on RAL susceptibility and IN-mediated RC was characterized further using SDMs. A single mutation at position 148 or 155 had a similar effect on RAL susceptibility (approximately 20- to 40-fold reduction in susceptibility) (Table (Table3);3); similar results were reported by Goodman et al. (14). The addition of the E92Q to N155H markedly reduced RAL susceptibility relative to the N155H mutation alone. Similarly, in most cases, the addition of other secondary mutations G140S(A) or E138K to 148 mutations (R, H, and K) conferred large reductions in RAL susceptibility, with the exception of G140S and Q148K, E138K and Q148R, and E138K and Q148H. These results are highly consistent with our observations in patient samples, in which the viruses with N155H and E92Q or Q148R(H) and G140S mutations exhibited large reductions in RAL susceptibility. The fact that patient viruses containing N155H without E92Q exhibited a broad range of RAL susceptibility indicates that other amino acid substitutions in the IN proteins of these viruses also influence RAL susceptibility.
Interestingly, we observed an increase in susceptibility to RAL when comparing the SDMs G140S and Q148K to Q148K alone. Although Q148K is not selected frequently by RAL, this observation demonstrates that not all mutations at position 148, or combinations of resistance mutations, will have the same effect on RAL resistance. This fact may become more important if other IN inhibitors are developed that more readily select Q148K and thus retain low-level activity to RAL. In addition, our observations demonstrate that patient viruses containing N155H display a broad range of RAL susceptibility. This finding may also become increasingly important as more compounds in this class are approved for clinical use.
In this BENCHMRK I and II study population, IN-mediated RC measurements at the first failure time point were significantly lower relative to the corresponding baseline virus. IN-mediated RC reductions were strongly associated with mutations at positions 148 or 155. Using SDMs, we reproduced similar reductions in IN-mediated RC when RAL-associated resistance mutations were introduced into wild type IN at either position 148 or 155. The addition of E92Q to N155H markedly decreased RAL susceptibility and IN-mediated RC, strongly suggesting that this secondary mutation is selected in vivo because it further reduces RAL susceptibility rather than compensating for losses in viral RC conferred by the primary N155H mutation. In the majority of cases, the addition of the secondary mutation G140S(A) or E138K to primary mutations at position 148 further reduced RAL susceptibility in the absence of considerable changes in IN-mediated RC. Notable exceptions were G140S and Q148K, which became more susceptible to RAL than Q148K alone, and G140S and Q148H, which increased IN-mediated RC compared to Q148H alone. Overall, the RAL susceptibility and IN-mediated RC of the SDM panel that we generated were concordant with our results in the BENCHMRK patient viruses that we analyzed. Based on our in vitro and in vivo observations, we conclude that both the selection of amino acid substitutions at primary resistance positions and the secondary mutations are driven primarily by the need to maximize virus replication at high drug concentrations (RAL resistance) and secondarily by preserving or restoring virus replication in the absence of drug (IN-mediated RC). Our observations are highly consistent with those reported by Nakahara et al. in relation to changes in the RC of these SDMs, with the exception of E138K and Q148R, in which we found that IN-mediated RC was similar to Q148R alone, while they reported that E138K and Q148R had a significant increase in infectivity (22).
Recently, several reports have described the evolution of IN resistance mutations under continued RAL pressure in the absence of complete viral suppression (4, 6, 21; M. D. Miller and D. J. Hazuda, presented at the 48th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, 25 to 28 October 2008). In phase II and III RAL clinical trials, some early treatment failure virus populations were found to contain mixed amino acids at positions 155 and 148 of IN but over time switched to virus populations containing a mutation at position 148 alone. Furthermore, in some cases, early treatment failure virus populations containing N155H alone were observed to switch to virus populations containing Q148R(H) (21; M. D. Miller and D. J. Hazuda, presented at the 48th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, 25 to 28 October 2008). Additional case studies have also documented evolution from N155N/H to Q148H (4) or N155H to Y143R in subjects with incomplete viral suppression (6). Our observations reported here establish that the N155H and Q148R(H)(K) pathways to RAL resistance are mutually exclusive, which may help to explain the shift in the resistance mutation pathway reported by others. Further studies are needed to determine how the development of RAL resistance through these distinct resistance pathways may impact the efficacy of future IN inhibitors.
This study was supported by National Institutes of Health (NIH), U.S. National Institute of Allergy and Infectious Diseases (NIAID) Small Business Innovation Research (SBIR) grant R44 AI057074.
We thank the Monogram Biosciences Clinical Reference Laboratory for performing the PhenoSense integrase and GeneSeq integrase assays and Cynthia Sedik for administrative assistance.
Published ahead of print on 16 September 2009.