McNicholas, Paul M PhD*; Mann, Paul A MPH*; BSc, Lisa Wojcik*; PhD, Ping Qiu†; Lee, Erin RN‡; McCarthy, Michael MD‡; Shen, Junwu PhD§; Black, Todd A PhD*; Strizki, Julie M PhD*
To enter CD4+ cells HIV-1 binds the CD4 receptor and 1 of 2 chemokine coreceptors, CCR5 and CXCR4. Depending on the coreceptor used, the virus is designated CCR5-tropic (R5) or CXCR4-tropic (X4).1 A dual-tropic (DT) virus can utilize both coreceptors. Because commercial phenotypic assays cannot distinguish DT populations from mixtures of R5/X4 viruses, viral isolates that give a positive signal in cells with either coreceptor are referred to as dual-mixed (DM) tropic. The CCR5 antagonist's vicriviroc (VCV) and maraviroc (MCV) bind with high affinity to the receptor and induce conformational changes that specifically block entry of R5 viruses.2-6 HIV-1 can evade the inhibitory effects of CCR5 antagonists through alterations in the envelope protein, gp160, that either allow it to utilize an alternate coreceptor (primarily CXCR4), or the drug-bound form of CCR5. Due to the allosteric mechanism of MCV and VCV, phenotypic resistance is primarily manifested as a decrease in the maximum level of viral suppression or percent of inhibition (termed MPI) and not as an increase in the IC50 value.5-7
The incidence of subjects with virus resistant to VCV was monitored in phase 2 studies in both treatment-naive (P03802) and treatment-experienced subjects (ACTG 5211 and P03672/VICTOR-E1).8-10 In the ACTG 5211 trial, only 1 subject developed resistance to VCV during treatment, and the resistant isolates had multiple changes in the V3-loop. However; these changes were associated with a loss in fitness as VCV susceptible virus re-emerged after cessation of VCV treatment. In P03802 and P03672, virus from 4 and 5 subjects displayed phenotypic VCV resistance, respectively; this despite the fact that the P03802 study was terminated early due to elevated failure rates in the low-dose groups.11 Analysis of virus from the 4 subjects from P03802 failed to identify a consistent pattern of mutations in gp160 associated with VCV resistance.8
In this study, we characterize additional VCV-resistant isolates from 5 highly treatment-experienced subjects enrolled in P03672 and extend our initial results by investigating of the contribution of mutations within the third variable (V3) loop to the overall resistance phenotype.
The P03672 study design was previously reported.9 Briefly, subjects failing 3-drug ART determined to have R5 virus by the original Trofile Assay at baseline and a plasma HIV RNA level of ≥1000 copies per milliliter were randomized 1:1:1 to receive once-daily VCV 30 mg (n = 39), 20 mg (n = 40), or placebo (n = 37) in conjunction with a boosted protease containing optimized background regimen (OBT). Two subjects in the placebo group discontinued before taking a single dose.
Population Pharmacokinetic Analysis
Blood samples for VCV plasma concentrations were obtained at weeks 4, 12, and 24. Population pharmacokinetic analysis was performed using the NONMEM V software with additional PK data from healthy subjects in study P04825 (unpublished data).
Drug Susceptibility and Tropism
Susceptibility testing was performed at screening, baseline (BL), and at time of treatment failure or study discontinuation. The PhenoSense GT assay (Monogram Biosciences, San Francisco, CA), was used to determine susceptibility to drugs in the OBT. A sensitivity score of 1 (full susceptibility) or 0 (partial susceptibility or resistant) was assigned for each drug, the overall susceptibility score (OSS) represents the sum for all drugs in the OBT.
HIV coreceptor tropism and VCV susceptibility were assayed using the Monogram Biosciences original Trofile and PhenoSense Entry assays, respectively, at BL, weeks 8, 20, and 48, or at discontinuation for subjects with HIV RNA >1000 copies per milliliter. Drug concentrations required to inhibit virus replication by 50% (IC50), and MPI values were calculated from inhibition curves (Fig. 1) as described previously.8 VCV susceptibility was also expressed as relative-MPI (R-MPI) and fold change-IC50, (FC-IC50) which are the ratios of the subject and control virus values (MPI or IC50, respectively) tested on the same day.
For selected clones, full length gp160 was cloned into replication competent pNL4-3 backbone and tropism determined by measurement of p24 antigen levels (p24 ELISA) on day 3 in infected U87-CD4-CCR5 or U87-CD4-CXCR4 cells as previously described.12
Envelope Clonal Analysis
Full length gp160 clones were generated by polymerase chain reaction amplification from BL and end of therapy (EOT; defined as the end of 48 weeks or study discontinuation, but not including the open-label extension period) plasma samples. The clones were sequenced and tested for susceptibility to VCV as described above. Sequence alignments and phylogenetic trees were constructed using the VectorNTI AlignX program employing the Clustal W and neighbor-joining analytical methods. Tropism predictions were made using the web position-specific scoring matrices (webPSSM) and Geno2Pheno algorithms; default settings were used for both algorithms.13,14 Changes in the V3 loop sequences were made by site-directed mutagenesis and susceptibility to VCV and MCV assessed in triplicate using the PhenoSense Entry assay. Amino acids in gp160 were numbered according to HXB2 reference strain coordinates using the Los Alamos National Laboratory HIV sequence locator tool: (http://www.hiv.lanl.gov/content/sequence/LOCATE/locate.html). The sequences are deposited in GenBank (HQ377375-HQ377506).
Identification of Subjects With Virus With Reduced Susceptibility to VCV
Of the 114 subjects that received at least 1 dose of VCV or placebo, VCV susceptibility data for BL plasma samples were available for 105 subjects. The mean MPI at BL was 93.7% (SD 4.6%) and the range was 77%-100%. The corresponding values for the control virus (JrCSF) tested in parallel were 87.4% (3.1%) and 79%-94%. To control for the relatively high level of day to day variability in the MPI parameter, R-MPI values were used. In this study, phenotypic resistance to VCV was defined as an R-MPI value that was 2 standard deviations lower than the mean R-MPI value (1.07, SD: 0.065) for all BL samples; this value was 0.94.
Using this cutoff value, all but 1 subject had VCV susceptible virus at BL. The exception, subject 011 from the 30 mg arm, had an R-MPI of 0.89 and an OSS of 3 at BL. However, by week 12, this subject was fully suppressed, and viral RNA remained undetectable until week 48 thus precluding additional measurements.
During the treatment period, virus with reduced susceptibility to VCV was recovered from 1 and 4 subjects from the 30 and 20 mg VCV dose groups, respectively (Table 1, Figs 1, 2). Resistance to VCV was not detected in subjects receiving placebo. As previously observed for other VCV-resistant viruses, the IC50 values for the viral populations with reduced MPI values were either similar to or lower than the BL IC50 values. Resistant virus was detected as early as week 8 (subject 498) and as late as week 53 (subject 247) of therapy.
All 5 subjects with VCV-resistant virus were virologic failures. Three subjects had subtype B virus, and the remaining 2 were infected with subtypes F and F1. Four of the subjects had OSS scores of zero and so were essentially on VCV monotherapy. The remaining subject (subject 498) had a single active drug (tenofovir), and the viral swarm remained susceptible to this drug until EOT. After completion of the study at 48 weeks, subjects had the option of continuing to receive 30 mg VCV along with OBT in an open-label extension study. From the 5 subjects listed above, subjects 485 and 500 continued. At the start of the extension, both subjects had OSS scores of zero, and virus from both subjects became progressively more resistant to VCV over time (Table 1).
Changes in Viral Tropism in Subjects With Virus That Developed Reduced Susceptibility to VCV
Although the viral swarms from all 5 subjects replicated robustly in CCR5 cells, virus from 3 subjects, 247, 485, and 500, also infected CXCR4 cells, although the signal for subject 485 was relatively weak (Table 1). In subjects 485 and 500, the VCV-resistant and DM viruses were detected at the same time points, whereas in subject 247, DM virus was detected >40 weeks earlier than VCV-resistant virus. In the open-label extension period, the viral swarm from subject 485 changed from DM (week 72) to R5 (week 96), whereas the swarm from subject 500 remained DM.
Before enrollment, all subjects were screened using the original Trofile assay to ensure that they had R5 virus only. However, this assay can only reliably detect X4 virus when it constitutes at least 10% of the viral swarm.15 Prior data suggested that in subjects receiving a CCR5 antagonist, the appearance of X4 virus resulted from the unmasking of a minor population of X4 virus that was not detected at screening.16 To determine if low levels of X4 virus were present before therapy, BL samples from all subjects were rescreened using an enhanced sensitivity Trofile assay (Trofile ES) that reportedly detects X4 virus when it constitutes as little as 0.3% of the viral swarm.17 Among the 5 subjects with VCV-resistant virus, only subject 500 was categorized as having DM virus at BL by the Trofile ES assay.
VCV Plasma Concentrations Were Not Associated With Loss of VCV Susceptibility
A population PK model, generated by combining PK data from P03672 and Phase I studies, was used to derive steady-state Cmax, Cmin, and AUC plasma concentrations of VCV for each subject (data not shown). Subjects 247, 485, and 500 had lower than average Cmin concentrations of VCV (96-114 ng/mL) whereas the levels in subject 784 were similar and those in subject 498 above the average values calculated for all the subjects in the respective dose groups.
Clonal Analysis of Viral gp160 Envelopes From Subjects With Reduced Susceptibility to VCV
Full-length gp160 clones (12/time point) were generated from BL and EOT plasma samples for each of the 5 subjects (clones were also generated from the 48-week sample for subject 247). Using the BL and EOT gp160 nucleotide sequences, phylogenetic trees were constructed, and the relative heterogeneity of the virus populations assessed by calculating the mean branch length of individual clones from each time point. In general the EOT clones clustered together and were distinct from the BL clones (data not shown). For subjects 247, 500, and 784, the mean branch lengths of the EOT clones were shorter than those of the corresponding BL clones suggesting that VCV treatment selected for more homogeneous viral populations. Genotypic tropism predictions were made using webPSSM and Geno2Pheno algorithms (Table 2). EOT clones predicted to be X4 clustered separately from the R5 clones from the same subject (data not shown).
Consistent with the susceptibility data for the viral swarms, all but 4 of the BL clones were susceptible (ie, R-MPI > 0.94) to VCV (Table 2, Fig. 2). The exceptions were from subject 498; interestingly, the V3 loop sequences of the 4 VCV-resistant BL clones were identical to those of the 8 susceptible clones. The EOT clones exhibited a mixture of susceptibilities to VCV ranging from fully susceptible to highly resistant in some subjects. With the exception of the clones from subject 247, the VCV-resistant clones showed similar infectivity (reflected in the RLU values) to the BL clones in CCR5-expressing cells.
Comparing the sequences of the BL and EOT clones, the majority of the resistant clones from the 5 subjects had acquired at least 1 and often several, amino acid substitutions in the V3 loop. However, overall there were no conserved patterns of substitutions between subjects either within the V3 loop (Table 2) or flanking regions (data not shown). A subject by subject description follows.
All 12 BL clones were susceptible to VCV, as were 8 and 3 clones from weeks 48 and 53, respectively. Seventeen clones (11 and 6 from weeks 48 and 53, respectively) had a 3 amino acid insertion in the V3 loop, and both algorithms predicted that the 17 clones are X4-tropic. “The gp160 gene from one representative week 53 clone (GenBank accession number…) with the 3 amino acid insertion was introduced into replication-competent pNL4-3 backbone, the resultant virus infected cells expressing CXCR4 but not CCR5 suggesting that the clone was not DT (data not shown)”.
Consistent with the genotypic predictions, 6 of the 17 clones (3 each from weeks 48 and 53), replicated poorly in CCR5 cells (RLU values were similar to uninfected controls), and as a result, VCV susceptibility could not be determined. The remaining 11 clones were susceptible to VCV and infected CCR5 cells more efficiently (RLU values of 2000-85,000). Of note, several of these clones had identical V3 loop sequences to the 6 clones that failed to replicate in the CCR5 cells (footnote 6 in Table 2), suggesting that regions outside V3 may influence the ability of these clones to utilize CCR5 and possibly CXCR4.
Five of the 6 CCR5-tropic clones from week 53 (lacking the 3 amino acid insertion) were highly resistant and demonstrated enhanced replication in the presence of VCV as evidenced by the negative R-MPI values. All 5 clones were predicted to be R5 and introduction of the gp160 gene from one representative clone (GenBank accession number HQ377387) into pNL4-3-generated virus that was exclusively R5 (data not shown). The five clones had identical V3 loop sequences with 3 amino acid changes (G314A, R/G315E and D/E325G) compared with BL and only minor sequence differences outside of V3 (data not shown). Interestingly, a previous study reported similar changes at positions 314 and 315 that were associated with enhanced replication in a Clade D-resistant isolate from an earlier VCV trial.18 Clones from this isolate had changes at either R314W or Q315E, and neither mutation alone was sufficient to confer phenotypic resistance.
All BL clones were susceptible to VCV, and all week 48 clones were resistant. Although the viral swarm at week 48 was categorized as DM (the signal in CXCR4 cells was just above background, Table 1), the RLU values for the 12 clones from week 48 suggested that they were able to efficiently infect CCR5 cells. The BL clones showed a mixture of lysine and arginine and aspartic acid and glutamic acid at positions 10 and 25, respectively. However, in the resistant clones, only arginine and aspartic acid were selected at these positions and an additional change of H308D/P was present in all clones.
All 12 BL clones had identical V3 loop sequences. Eight of the BL clones were susceptible to VCV and, as noted above, the remainder had R-MPI values ranging from 0.89 to 0.9. Outside of the V3 loop, there were no consistent patterns of substitutions that differentiated susceptible from resistant BL clones (data not shown). The week 20 clones were all resistant to VCV, and all exhibited similar levels of infectivity in CCR5 cells.
There was considerable heterogeneity in the V3 loop among the 12 VCV susceptible BL clones. The V3 loops from the clones from week 48 showed less variability and all 12 were resistant to VCV. At week 48, 2 clones (with identical V3 loops) were predicted to be R5, whereas a second group of five clones, with V3 loops that differed from the two R5 clones by a single amino acid, were predicted to be R5/X4. The gp160 gene from a representative R5 and R5/X4 clone (GenBank accession numbers HQ377471, HQ377472) was introduced into pNL4-3; virus from both clones were DT because they infected cells expressing either CCR5 or CXCR4 (data not shown).
The 12 BL clones were susceptible to VCV. Ten of the clones from week 48 were resistant to VCV; 2 clones, which had identical V3 loop sequences to the 5 of the resistant clones, remained fully susceptible to VCV.
Contributions of V3 Loop Amino Changes to the Resistance Phenotype
The contribution of the V3 loop to the resistance phenotype was further explored by introduction of the V3 substitutions that were unique to the resistant EOT clones into the corresponding susceptible BL clones. In the reciprocal exchange, residues in the EOT clones were reverted back to match the sequence of the corresponding BL clones. The susceptibility of the resultant recombinant envelopes to both VCV and MCV was then determined (Table 3). All clones generated robust signals in CCR5 cells; entry in CXCR4 cells was not measured. The V3 loops from resistant clones #2 and #11 from subject 485 each differed from a group of fully susceptible BL clones (typified by clone #24) by a single residue (H308D/P). Introduction of either the proline or aspartic acid into the BL clone did not confer reduced susceptibility to VCV or MVC. Similarly, changing the V3 loop of the resistant EOT clones to match the BL clone had no impact for clone #11 and only partially restored susceptibility to both drugs for clone #2. For clones from subject 498, there was no impact on susceptibility to either drug when the corresponding single amino acid substitutions were made in the resistant (clones #31 and #33) and susceptible BL clone (#23). However, the V3 loop of the EOT clone #20 from subject 784 differed from the BL clone (#14) by 2 amino acids. Changing the V3 loop in the EOT clone to match the BL clone restored susceptibility to both VCV and MCV, whereas the reciprocal exchange had no measurable effect.
This is the third study to utilize a clonal analysis to examine the molecular basis of the evolution of resistance to the CCR5 antagonist VCV in HIV-1-infected subjects.7,8,10 Despite the differences in prior treatment history for the subjects in the 3 studies, both the frequencies at which resistant virus was detected and the time to development of resistance were similar. In all 3 studies, VCV resistance was relatively uncommon and was not the principal cause of virologic failure. In the treatment-naive subjects (study P03802), a marked reduction in susceptibility to VCV was detected at 24 weeks; in ACTG 5211, partially resistant virus was detected at weeks 16 and 19. In this study (P03672), resistant virus was detected at weeks 20-53 for 4 of the 5 subjects. In the fifth subject, a resistant viral swarm was detected at week 8; surprisingly, this subject was the only 1 of the 5 with a fully active drug in their OBT at BL.
As might be expected, X4-using virus was detected more frequently in the treatment-experienced subjects from study P03672 than in naive subjects from P03802. This was true for both the subset of subjects with resistant virus (3 of 5 versus 1 of 4, respectively) and in all subjects receiving VCV (16 of 79 and 5 of 68, respectively). In ACTG 5211, 13 of the 29 virologic failures showed evidence of emergence of X4 virus at the time of failure. These differences likely reflect the subject's prior treatment history and/or the stage of disease progression because the appearance of X4 virus has been associated with lower CD4 counts and more advanced disease.19
The impact of VCV plasma levels and OBT on development of resistance was monitored in both P03802 and P03672. Although P03802 was halted prematurely due to higher failure rates in the low-dose VCV arms, the VCV plasma levels in 3 of the 4 subjects that developed resistance exceeded the mean plasma concentrations for subjects in the same dose group. In this study, 2 of the 5 subjects had mean plasma levels that exceeded the mean value for their respective dose group. However, the Cmin VCV plasma concentrations for all 5 subjects were several fold higher than the BL IC50 values for each subjects' virus and exceeded the mean levels measured in all subjects in the highest dose group from the prior study (data not shown). Therefore, there was not a clear association between VCV concentration and emergence of resistance.
With regard to OBT, 4 of the 5 subjects did not have a fully active drug in their OBT at BL, whereas the fifth subject remained susceptible to tenofovir through out the study. This later subject also had the highest VCV exposure of the 5 subjects. Overall, for the subset of subjects with a BL OSS of ≤1, the frequency of virologic failure was lower among subjects receiving VCV (22.5% and 18% for 20 and 30 mg dose groups, respectively) than among those receiving placebo (41%).
Although signature resistance mutations have been identified for many antiviral drugs, the molecular correlates associated with resistance to CCR5 antagonists have not been determined. Prior work established the importance of the V3 loop of gp120 in making critical interactions with the CCR5 coreceptor, and several studies have implicated mutations in this region as being key determinants of resistance.6-8,10 However, other regions of gp160 also seem to play an important role, these include flanking domains that act in concert with V3 loop substitutions and more distal regions, such as gp41, that seem to be independent of the V3 loop.20 The clonal analysis described herein, like the analysis of resistant viruses from P03802, failed to identify any conserved patterns of mutations that provide a signature for resistance to VCV. With the exception of 3 clones from subject 498, all clones exhibiting resistance to VCV had acquired substitutions in the V3 loop (ie, when compared with the susceptible BL clones), and in general, clones with identical V3 loops exhibited similar susceptibility profiles to VCV and MVC. However, reversing the substitutions in resistant clones to match the sequence of the corresponding BL clone restored susceptibility in only 1 of the clones tested. The reciprocal exchange, introduction of V3 loop substitution(s) from resistant clones into the matched BL clone, had no impact. Clearly, substitutions outside of the V3 loop are required to recapitulate the resistance phenotype, however, the limited number of resistant isolates analyzed to date has not revealed any specific pattern of compensatory mutations outside V3 that are linked to VCV resistance.
A subset of viral clones, recovered from an on-treatment sample from subject 500, were able use both the inhibitor-bound form of the CCR5 receptor and the alternate receptor CXCR4. It remains to be determined if the amino acid substitutions responsible for the change in susceptibility to VCV are the same as or distinct from those substitutions confer the ability to utilize the CXCR4 coreceptor.
In summary, like previous Phase 2 studies of VCV, de novo resistance to the CCR5 antagonist occurred infrequently and was not the primary cause of virologic failure among subjects in the VCV arms. Clonal analysis identified multiple mutations in the viral envelope associated with VCV resistance, however, no clear pattern of resistance mutations was observed among subjects. Currently, only phenotypic assays are available for assessing resistance to CCR5 antagonists and additional work is required to better determine appropriate R-MPI cutoffs for viral susceptibility and resistance in the clinical setting.
1. Berger EA, Murphy PM, and Farber JM. Chemokine receptors as HIV-1 coreceptors: roles in viral entry, tropism, and disease. Ann Rev Immunol
2. Dorr P, Westby M, Dobbs S, et al. Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 Activity. Antimicrob Agents Chemother
3. Kondru R, Zhang J, Ji C, et al. Molecular interactions of CCR5 with major classes of small-molecule anti-HIV CCR5 antagonists. Mol Pharmacol
4. Strizki JM, Tremblay C, Xu S, et al. Discovery and characterization of vicriviroc (SCH 417690), a CCR5 antagonist with potent activity against human immunodeficiency virus type 1. Antimicrob Agents Chemother
5. Pugach P, Marozsan AJ, Ketas TJ, et al. HIV-1 clones resistant to a small molecule CCR5 inhibitor use the inhibitor-bound form of CCR5 for entry. Virology
6. Westby M, Smith-Burchnell C, Mori J, et al. Reduced maximal inhibition in phenotypic susceptibility assays indicates that viral strains resistant to the CCR5 antagonist maraviroc utilize inhibitor-bound receptor for entry. J Virol
7. Ogert RA, Wojcik L, Buontempo C, et al. Mapping resistance to the CCR5 co-receptor antagonist vicriviroc using heterologous chimeric HIV-1 envelope genes reveals key determinants in the C2-V5 domain of gp120. Virology
8. McNicholas P, Wei Y, Whitcomb J, et al. Characterization of emergent HIV resistance in treatment naive subjects enrolled in a vicriviroc phase 2 trial. J Infect Dis
9. Suleiman J, Zingman BS, Diaz RS, et al. Vicriviroc in combination therapy with an optimized regimen for treatment experienced subjects: 48 week results of the VICTOR-E1 phase 2 trial. J Infect Dis
10. Tsibris AMN, Sagar M, Gulick RM, et al. In vivo emergence of vicriviroc resistance in a human immunodeficiency virus type 1 subtype C-infected subject. J Virol
11. Landovitz RJ, Angel JB, Hoffmann C, et al. Phase II study of vicriviroc versus efavirenz (both with zidovudine/lamivudine) in treatment-naive subjects with HIV-1 infection. J Infect Dis
12. Strizki JM, Xu S, Wagner NE, et al. SCH-C (SCH 351125), an orally bioavailable, small molecule antagonist of the chemokine receptor CCR5, is a potent inhibitor of HIV-infection in vitro and in vivo. Proc Natl Acad Sci U S A
13. Sing T, Low AJ, Beerenwinkle N, et al. Predicting HIV coreceptor usage on the basis of genetic and clinical covariates. Antivir Ther
14. Jensen MA, Li F-S, van't Wout AB, et al. Improved coreceptor usage prediction and genotypic monitoring of R5-to-X4 transition by motif analysis of human immunodeficiency virus type 1 env V3 loop sequences. J Virol
15. Whitcomb JM, Huang W, Fransen S, et al. Development and characterization of a novel single-cycle recombinant-virus assay to determine human immunodeficiency virus type 1 coreceptor tropism. Antimicrob Agents Chemother
16. Westby M, Lewis M, Whitcomb J, et al. Emergence of CXCR4-using human immunodeficiency virus type 1 (HIV-1) variants in a minority of HIV-1-infected patients following treatment with the CCR5 antagonist maraviroc is from a pretreatment CXCR4-using virus reservoir. J Virol
17. Reeves JD, Coakley E, Petropoulos CJ, et al. An enhanced-sensitivity Trofile™ HIV coreceptor tropism assay for selecting patients for therapy with entry inhibitors targeting CCR5: a review of analytical and clinical studies. J Viral Entry
18. Ogert RA, Hou Y, Ba L, et al. Clinical resistance to vicriviroc through adaptive V3 loop mutations in HIV-1 subtype D gp120 that alter interactions with the N-terminus and ECL2 of CCR5. Virology
19. Goetz MB, Leduc R, Kostman JR, et al. Relationship between HIV coreceptor tropism and disease progression in persons with untreated chronic HIV infection. J Acquir Immune Defic Syndr
20. Cleo G, Anastassopoulou CG, Ketas TJ, et al. Resistance to CCR5 inhibitors caused by sequence changes in the fusion peptide of HIV-1 gp41. Proc Natl Acad Sci USA
© 2011 Lippincott Williams & Wilkins, Inc.