Integration into the host chromosome of the double stranded DNA product of reverse transcription is an essential step in the life cycle of HIV-1.1-4 This process is facilitated by integrase (IN), a 32-kDa virally encoded enzyme. After reverse transcription, HIV-1 IN, along with other viral and cellular proteins, binds to specific sequences in the long-terminal repeats of the viral cDNA to form the pre-integration complex. IN cleaves 2 conserved nucleotides from the 3′-ends of both strands of the viral cDNA and, after nuclear importation, ligates the 5′-ends of the viral cDNA to the host chromosomal DNA by a process called strand transfer. IN strand-transfer inhibitors (INSTIs) block infection by preventing integration of viral double stranded DNA into the host cell DNA.5
Raltegravir is the first IN inhibitor approved for the treatment of HIV-1 infection.6 As with other antiretroviral drugs, resistance to RAL can emerge in the setting of incomplete viral suppression. Data from clinical trials show that RAL resistance involves IN mutations Y143C, Q148H or R or K or N155H, together with associated secondary mutations that result in higher levels of resistance.7-10 These mutations are located within the catalytic core domain of IN and have been shown to reduce viral replication capacity.11 The N155H mutant generally emerges first and is eventually replaced by Q148H mutants, usually in combination with G140S.11-13 To date, only limited information is available on the effect of RAL resistance mutations on relative viral fitness.
To explore virologic consequences of INSTI resistance, we constructed a series of recombinant viruses carrying various INSTI resistance mutations and compared their effect on viral fitness as determined by growth competitions assays in the absence and presence of RAL.
MATERIALS AND METHODS
Reagents and Cells
Raltegravir tablets (Issentress, 400 mg; Merck, Whitehouse Station, NJ) were purchased from the Massachusetts General Hospital pharmacy. MT-2 cells were grown in R-10 medium [RPMI 1640 (Cellgro, Herndon, Virginia) supplemented with 10% fetal calf serum, 2 mM L-glutamine, penicillin (100 U/mL), and streptomycin (100 μg/mL)]. The cell lines 293T, U87-CD4-X4, and TZM-HeLa CD4-long-terminal repeat/β-gal, obtained from the AIDS Research and Reference Reagent Program, were propagated in Dulbecco modified Eagle medium supplemented with 10% fetal calf serum, hygromycin B (0.2 mg/mL), and geneticin (0.2 mg/mL).
Construction of HIV-1 Vectors With an IN Deletion
A 3868 basepair fragment encompassing the entire coding sequence of the HIV-1 pol gene [corresponding to nucleotides 1970 to 5838 of the HIV-1NL4-3 sequence (http://hiv-web.lanl.gov)] was amplified by PCR (polymerase chain reaction) from proviral clone pNL4-3 and cloned into pGEM-T-Easy vector to generate plasmid pGEM-Pol. Subsequently, 864 nucleotides of the IN coding sequence (corresponding to nucleotides 4230 to 5094 in HIV-1NL4-3) were deleted, and a unique PmeI restriction enzyme site introduced at the deletion junction by PCR-based cloning. The pol gene carrying this deletion was then cloned into pNL4-3 to yield pHIVΔIN.PmeI. A segment of the Salmonella enterica serovar Typhimurium histidinol dehydrogenase gene14 or the green fluorescent protein gene (GFP) were then introduced into the XhoI site in nef to serve as sequence tags, yielding plasmids pHIVΔin.PmeInef-hisD or pHIVΔin.PmeInef-GFP, respectively.
The wild-type IN-coding region of HIV-1NL4-3 pol, along with approximately 1026 basepairs of flanking sequence (corresponding to nucleotides 1970 to 2253 and 5095 to 5838 of the HIV-1NL4-3 sequence [http://hiv-web.lanl.gov]) was cloned into the pGEM-T-easy-vector (Promega, Madison, WI). Mutations at IN codons 74, 92, 138, 140, 148, 155, and/or 163 were introduced using the QuickChange site-directed mutagenesis kit (Stratagene). The presence of mutant sequences was confirmed by automated sequencing of the final plasmid clones on an ABI 377 automated sequencer.
Generation of Recombinant Marker Viruses
Infectious recombinant marker viruses carrying the desired IN mutations were generated by cotransfecting linearized pHIVΔin.PmeInef-hisD or pHIVΔin.PmeInef-GFP along with the wild-type or mutant IN gene of interest into 293T cells. The IN-coding sequence was amplified from proviral DNA of infected cells at the end of virus culture and analyzed by automated DNA sequencing to verify presence of the correct alleles at codons 74, 92, 138, 140, 148, 155, and/or 163. Homologous recombination between flanking sequences 5′- and 3′- to IN and the IN-deleted HIV plasmid resulted in generation of intact, fully infectious viruses.
Raltegravir Susceptibility Testing
The susceptibility of HIV-1 recombinants to RAL was determined by a single-cycle HIV-1 infectivity assay using TZM-bl cells. Raltegravir stock solutions were prepared by dissolving crushed tablets containing 400 mg of drug in 90 mL of water to yield a nominal concentration of 0.01 mM. The solution was passed through a 0.22-micron filter to remove insoluble excipients and to sterlize the solution. Two-fold serial dilutions of raltegravir (range, 5000-5 nM) in a total of 100 μL Dulbecco modified Eagle medium were added to the wells of a 96-well microtiter plate. An amount of virus sufficient to produce ß-galactosidase activity equivalent to 1 × 105 counts per second (CPS) was added to each well except for uninfected control wells, after which the cell suspension was added into each well (1 × 104 TZM-bl cells in 50 μL/well); the final volume in each well was 200 μL. After incubation at 37°C for 48 hours, ß-galactosidase activity was quantified using Galacto-Light Plus (Applied Biosystems, Foster City, CA) and expressed as chemiluminescence units (CPS). All infections were performed in triplicate and experiments repeated at least twice. Percent inhibition was expressed as CPS generated at various inhibitor concentrations relative to the no-drug control. Drug susceptibility was calculated by plotting the percent inhibition of virus replication (ß-galactosidase activity) versus the log10 drug concentration to derive the IC50. The drug susceptibility curves were fitted by nonlinear regression using GraphPad Prism 5 (GraphPad Software, Ann Arbor, MI).
Growth Competition Assays
The fitness of various raltegravir-resistant mutants was compared with wild-type virus and to each other in pair-wise growth competition assays as described.14,15 Briefly, recombinant marker viruses of interest carrying the hisD or GFP sequence tags were mixed together at ratios of 50:50; 80:20; or 20:80, respectively, and inoculated onto 1.5 × 106 MT-2 cells suspended in 300 μL of R-10 medium to yield an m.o.i. of 0.001. After incubation at 37°C for 2 hours, cells were washed twice with phosphate-buffered saline, resuspended in 10 mL of R-10 medium at a concentration of 0.15 × 106 cells/mL in 25-cm2 tissue culture flasks and re-incubated. Cultures were passaged by inoculating 200 μL of supernatant onto 10 × 106 fresh MT-2 cells every 3 or 4 days. The proportion of the 2 competing variants was estimated by quantifying GFP and hisD sequences present in culture supernatants on days 1, 4, 7, 11, and 14 using real-time reverse trancriptase-polymerase chain reaction with an ABI PRISM 7000 Sequence Detection System (Applied Biosystems; Supplemental Information for primer and probe sequences). Viral RNA was extracted from culture supernatants using the Qiagen kit and treated with RNase-free DNase (Qiagen, Valencia, CA). Parameters for the real-time PCR were as described,15 except that the initial reverse transcription reaction was performed at 50°C for 30 minutes. Quantitative real-time reverse trancriptase-polymerase chain reaction was performed in triplicate for each sample. Control experiments in which two wild-type recombinants tagged with the hisD and GFP markers, respectively, were competed against each other showed no change in relative proportions of the 2 viruses over time, indicating that the sequence tags had no significant effect on relative fitness of the recombinants (data not shown). For each pair of viruses tested, reciprocal growth competition assays were conducted in which the IN genes of interest were linked to hisD or GFP and vice versa; in each case, similar results were obtained. Therefore, data shown represent the means ± SD of reciprocal experiments.
Estimation of Viral Fitness
Quantitative estimates of relative viral fitness were calculated as described.15 For each pair of recombinant viruses tested, the final ratio of the 2 viruses was determined by quantitative real-time PCR at day 11 as described above. The relative fitness (w) of each virus was obtained from the average of the results of 3 independent growth competition assays (inoculated at ratios of 80:20, 50:50, and 20:80). The fitness difference (WD) was estimated by the ratio of the relative fitness values (WD = WM/WL), where WM is the more fit and WL the less fit virus in the growth competition assay. Growth competition assays that compared 2 wild-type recombinants carrying the his D and GFP markers, respectively, gave a mean fitness difference (±SD) of 1.0 ± 0.02-fold. Therefore, we considered fitness differences greater than 1.1-fold to be significant.
Raltegravir Susceptibility of IN Mutants
Table 1 shows the raltegravir susceptibility of recombinant viruses carrying various mutations in HIV-1NL4-3 IN. Compared with WT, the N155H mutation conferred 19-fold resistance, and the Q148H(R) (K) mutations conferred 7-fold to 22-fold resistance to raltegravir. The addition of secondary mutations L74M or E92Q to N155H resulted in 28-fold and 55-fold resistance, respectively, but addition of G163R did not result in any substantial change in raltegravir resistance. The addition of secondary mutations E138K or G140S to Q148H resulted in 36-fold and 245-fold raltegravir resistance, respectively.
Fitness of N155H Mutants
The relative replicative fitness of recombinant viruses carrying the N155H mutation was compared with WT in growth competitions assays. In these experiments, the N155H recombinants were substantially less fit than WT in the absence of raltegravir and more fit than WT in presence of 5.0 μM RAL (Fig. 1). In the absence of drug, the relative fitness difference for WT versus N155H was 3.4-fold, whereas in the presence of drug the N155H mutant had a relative fitness that was 2.0-fold greater than wild type (Table 2).
To determine how the secondary RAL resistance mutations L74M, E92Q, and G163R affect the fitness of N155H mutants, we compared the relative replicative fitness of the N155H mutants to that of the L74M/N155H, E92Q/N155H, or N155H/G163R double mutants. Introduction of E92Q or G163R into an N155H backbone resulted in a virus with greater fitness than N155H mutant both in the absence and presence of RAL. The relative fitness difference for E92Q/N155H versus N155H was 2.8-fold in absence of drug and 6.7-fold in the presence of RAL (Fig. 1 and Table 2). The relative fitness difference for N155H/G163R versus N155H was 3.4-fold in absence of drug and 2.1-fold in the presence of RAL. By contrast, introduction of L74M into a N155H backbone further reduced viral fitness in the absence of drug but resulted in increased fitness in the presence of drug. The N155H mutant had a 5.7-fold relative fitness advantage over the L74M/N155H mutant in the absence of raltegravir, whereas in the presence of RAL, the L74M/N155H mutant had a 2.1-fold relative fitness advantage (Table 2). When competed against each other in the presence of raltegravir, the N155H/G163R mutant showed 2.1-fold greater fitness than the L74M/N155H mutant and the E92Q/N155H mutant in turn had a 3.7-fold relative fitness advantage over the N155H/G163R mutant (Table 2).
Fitness of the Q148H(R)(K) Mutants
The relative replicative fitness of Q148H(R)(K) mutants was compared with WT in growth competitions assays. Recombinants carrying each of these mutations were substantially less fit than WT in the absence of raltegravir and more fit than WT in presence of 5.0 μM drug (Fig. 1). The relative fitness difference for WT versus Q148H was 13.7-fold in absence of raltegravir, whereas the mutant had a 1.8-fold relative fitness advantage over WT in the presence of drug (Table 2). Similarly the Q148K and R mutants were less fit than WT in the absence of RAL and more fit than WT in the presence of drug (Table 2). To determine the effect of the secondary raltegravir resistance mutations E138K and G140S on the fitness of the Q148H mutant, we compared the relative replicative fitness of the Q148H mutant to that of double mutants E138K/Q148H and G140S/Q148H. Introduction of either E138K or G140S into a Q148H backbone increased viral fitness both in the absence and presence of raltegravir (Fig. 1 and Table 2). In the presence of raltegravir, the E138K/Q148H had a 3.0-fold relative fitness advantage over the Q148H mutant. Similarly, the G140S/Q148H mutant had a relative fitness advantage over Q148H. However, the E138K/Q148H mutant was less fit than G140S/Q148H in the absence and presence of drug. The relative fitness difference for G140S/Q148H versus E138K/Q148H was 3.7-fold in absence of drug and 3.1-fold in the presence of RAL (Table 2).
Relative Fitness of Q148H Versus N155H Mutants
To help understand why the N155H mutant usually emerges before the Q148H mutant, we compared the relative replicative fitness of recombinant viruses carrying the Q148H and N155H mutations in growth competitions assays. In these experiments, the Q148H recombinant was substantially less fit than N155H both in the absence and presence of 5.0 μM RAL (Fig. 2). The N155H mutant had a relative fitness advantage over the Q148H mutant of 3.4-fold in absence of raltegravir and a 2.4-fold advantage in the presence of drug (Table 2). The Q148R and Q148K mutants were fitter than the N155H mutant in absence of drug, but less fit in the presence of 5.0 μM RAL (data not shown).
We then compared the fitness of various Q148H and N155H double mutants to each other in pair-wise growth competition assays. In these experiments, the Q148H/G140S recombinant was substantially fitter than E92Q/N155H both in the absence and presence of raltegravir (Fig. 2). The Q148H/G140S mutant had a relative fitness advantage of 2.9-fold over the E92Q/N155H mutant in absence of drug, and 3.9-fold in the presence of 5.0 μM raltegravir (Table 2). Similarly, the E138K/Q148H mutant was fitter than the L74M/N155H mutant both in the absence and presence of 5.0 μM RAL (Table 2). By contrast, the E92Q/N155H mutant was fitter than the E138K/Q148H mutant both in the absence and presence of 5.0 μM RAL (Table 2).
In this study, we determined the relative fitness of INSTI-resistant mutants of HIV-1 in comparison to wild-type virus and to each other, in the presence and absence of raltegravir. The results of growth competition assays showed that in the absence of drug, the Q148H and N155H mutations, which confer raltegravir resistance, are associated with a substantial reduction in viral fitness as compared to wild type. As expected, the mutant viruses were fitter than wild type in the presence of drug, with the N155H mutant being more fit than the Q148H mutant. However, the G140S/Q148H double mutant was fitter than the E92Q/N155H double mutant. These results are consistent with the patterns of raltegravir resistance observed in samples from patients with virologic failure on a raltegravir-containing regimen and suggest that N155H mutants emerge first because they are fitter than the N148H mutants. With the addition of the G140S mutation, the G140S/N148H mutants enjoy a substantial fitness advantage over other raltegravir-resistant mutants and, therefore, become the dominant species in the virus population.9,11,16 Our findings compliment results of another study in which the relative infectivity of various raltegravir-resistant mutants was compared with that of wild-type virus at different drug concentrations.17 That study did not perform direct growth competition assays, however, so relative fitness differences could not be calculated.
A previous study compared the IN-mediated replication capacity of plasma viruses from 2 raltegravir-treated patients using the PhenoSense IN assay.11 In that study, the appearance of viruses with the Q148R or N155H mutations initially was associated with reduced replication capacity, but replication capacity returned toward wild-type levels as Q148H and G140S mutations emerged. Clones with G140S/Q148H showed higher IN-mediated replication capacity as compared with clones with Q148R or N155H alone; the G140S/Q148H clones also exhibited higher levels of raltegravir resistance.
Another study performed clonal analysis of raltegravir-resistant viruses obtained from patients after failure of a raltegravir-containing regimen in the phase 3 clinical trials of that drug.9 That study showed that the N155H and Q148H(R)(K) mutations seem to be mutually exclusive. The secondary mutation E92Q occurred only in combination with N155H, whereas the G140S mutation occurred only with Q148H(R)(K). Site-directed mutants carrying N155H showed lower replication capacity than wild type but higher replication capacity than Q148H(R)(K). The E92Q/N155H double mutant had a lower replication capacity than the N155H mutant, whereas the G140S/Q148H mutant showed a replication capacity that was close to that of the wild-type control. These results are generally similar to the results of our study using growth competition experiments in the absence of drug, except that the E92Q/N155H mutant showed greater fitness than the N155H single mutant. One possible explanation for these conflicting results is the difference in viral backbones used in the 2 sets of experiments (HIVIIIB versus HIV-1NL4-3, respectively). A third study found the G140A mutation in place of G140S in viruses that carried G148R substitution.12 It would be interesting to compare the fitness of the G140A and S mutants in combination with the different codon 148 mutants.
Reductions in relative fitness and replication capacity associated with INSTI resistance mutations are most likely due to impaired enzyme function.18-21 The N155H mutation results in reduced strand-transfer activity, whereas the E92Q mutation is associated with modest reductions in both 3′-processing and strand transfer activities of the enzyme; both activities of the enzyme are substantially reduced in the G140S/Q148H mutant.8,20,22 Whether these alterations in IN function and the associated reductions in viral fitness and replication capacity are manifested clinically by lower level viremia at the time of raltegravir failure remains uncertain, as no consistent pattern has emerged in the examples reported to date.11-13 Partial treatment interruption studies in which raltegravir is discontinued in patients with raltegravir-resistant virus while maintaining the background antiretroviral regimen are needed to resolve this question.
One limitation of our study is that the mutants we analyzed were constructed by site-directed mutagenesis of IN in an HIV-1NL4-3 backbone. Different results might have been obtained using other viral backbones, including those of non-B subtypes or using raltegravir-resistant clinical isolates. The close parallels between the results of the growth competition studies performed in this study and previously reported replication capacity data support the generalizability of our results. As with other in vitro studies, our study also suffers from the limitations inherent in comparing relative fitness in a system in which target cells are not limiting and in the absence of an antiviral immune response. The fitness effect of raltegravir resistance mutations could be measured in vivo by determining the rate at which these mutations disappear after discontinuation of the drug.23,24
In summary, mutations that confer resistance to the INSTI raltegravir are associated with reduced viral fitness when compared with wild-type virus in the absence of drug. The greater fitness of the N155H mutant as compared with Q148H in the presence of drug likely contributes to the earlier emergence of N155H mutants in the setting of raltegravir failure. The eventual predominance of the G140S/Q148H double mutant is explained in part by its greater fitness in the presence of drug as compared with other raltegravir-resistant mutants. Additional in vivo studies are needed to establish the clinical significance of these findings.
We thank Daniel Eggers for technical support and Jaclyn Coté and Janet Steele for administrative support.
1. Asante-Appiah E, Skalka AM. Molecular mechanisms in retrovirus DNA integration. Antiviral Res
2. Hindmarsh P, Leis J. Retroviral DNA integration. Microbiol Mol Biol Rev
3. Engelman A, Mizuuchi K, Craigie R. HIV-1 DNA integration: mechanism of viral DNA cleavage and DNA strand transfer. Cell
4. Esposito D, Craigie R. HIV integrase structure and function. Adv Virus Res
5. Hazuda DJ, Felock P, Witmer M, et al. Inhibitors of strand transfer that prevent integration and inhibit HIV-1 replication in cells. Science
6. Anker M, Corales RB. Raltegravir (MK-0518): a novel integrase inhibitor for the treatment of HIV infection. Expert Opin Investig Drugs
7. Roquebert B, Blum L, Collin G, et al. Selection of the Q148R integrase inhibitor resistance mutation in a failing raltegravir containing regimen. AIDS
8. Malet I, Delelis O, Valantin MA, et al. Mutations associated with failure of raltegravir treatment affect integrase sensitivity to the inhibitor in vitro. Antimicrob Agents Chemother
9. Fransen S, Gupta S, Danovich R, et al. Loss of raltegravir susceptibility by human immunodeficiency virus type 1 is conferred via multiple nonoverlapping genetic pathways. J Virol
10. Delelis O, Thierry S, Subra F, et al. Impact of Y143 HIV-1 integrase mutations on resistance to raltegravir in vitro and in vivo. Antimicrob Agents Chemother
11. Fransen S, Karmochkine M, Huang W, et al. Longitudinal analysis of raltegravir susceptibility and integrase replication capacity of human immunodeficiency virus type 1 during virologic failure. Antimicrob Agents Chemother
12. Ferns RB, Kirk S, Bennett J, et al. The dynamics of appearance and disappearance of HIV-1 integrase mutations during and after withdrawal of raltegravir therapy. AIDS
13. Canducci F, Sampaolo M, Marinozzi MC, et al. Dynamic patterns of human immunodeficiency virus type 1 integrase gene evolution in patients failing raltegravir-based salvage therapies. AIDS
14. Lu J, Kuritzkes DR. A novel recombinant marker virus assay for comparing the relative fitness of hiv-1 reverse transcriptase variants. J Acquir Immune Defic Syndr
15. Hu Z, Giguel F, Hatano H, et al. Fitness comparison of thymidine analog resistance pathways in human immunodeficiency virus type 1. J Virol
16. Buzón MJ, Marfil S, Puertas MC, et al. Raltegravir susceptibility and fitness progression of HIV type-1 integrase in patients on long-term antiretroviral therapy. Antivir Ther
17. Quercia R, Dam E, Perez-Bercoff D, et al. Selective-advantage profile of human immunodeficiency virus type 1 integrase mutants explains in vivo evolution of raltegravir resistance genotypes. J Virol
18. Dicker IB, Samanta HK, Li Z, et al. Changes to the HIV long terminal repeat and to HIV integrase differentially impact HIV integrase assembly, activity, and the binding of strand transfer inhibitors. J Biol Chem
19. Gerton JL, Ohgi S, Olsen M, et al. Effects of mutations in residues near the active site of human immunodeficiency virus type 1 integrase on specific enzyme-substrate interactions. J Virol
20. Johnson AA, Santos W, Pais GC, et al. Integration requires a specific interaction of the donor DNA terminal 5'-cytosine with glutamine 148 of the HIV-1 integrase flexible loop. J Biol Chem
21. Lee MC, Deng J, Briggs JM, et al. Large-scale conformational dynamics of the HIV-1 integrase core domain and its catalytic loop mutants. Biophys J
22. King PJ, Lee DJ, Reinke RA, et al. Human immunodeficiency virus type-1 integrase containing a glycine to serine mutation at position 140 is attenuated for catalysis and resistant to integrase inhibitors. Virology
23. Marconi V, Bonhoeffer S, Paredes R, et al. Viral dynamics and in vivo fitness of HIV-1 in the presence and absence of enfuvirtide. J Acquir Immune Defic Syndr
24. Paredes R, Sagar M, Marconi VC, et al. In vivo fitness cost of the M184V mutation in multidrug-resistant human immunodeficiency virus type 1 in the absence of lamivudine. J Virol