In our study, a two-compartment model described vicriviroc pharmacokinetics in treatment-experienced HIV infected subjects also taking ritonavir-containing protease inhibitor regimens, when combined with data from healthy HIV-seronegative subjects using population pharmacokinetic modeling. At week 2, higher vicriviroc concentrations were associated with a greater decrease in viral load. However, after subjects received optimized ART regimens, no PK-PD correlation could be defined at weeks 16 and 24. The virologic response to vicriviroc reported here is comparable to that observed by other investigators. Schurman et al. (12
) observed mean HIV RNA reductions of 1.08, 1.56 and 1.62 log copies/ml at doses of 10 mg, 25 mg and 50 mg BID (without ritonavir), respectively, in treatment-naïve individuals on vicriviroc monotherapy at day 14. Their observed values of mean Cmax
at steady-state were 63 ng/ml, 142 ng/ml and 276 ng/ml, respectively, compared to our results of 33 ng/ml, 69 ng/ml, 95 ng/ml at vicriviroc doses of 5 mg, 10 mg and 15 mg QD (with ritonavir), respectively. Landovitz et al. (15
) reported mean HIV RNA reductions of 0.93, 1.18 and 1.34 log10
using vicriviroc doses of 25 mg, 50 mg and 75 mg once daily respectively in naïve subjects on vicriviroc monotherapy (without ritonavir) at day 14.
Data generated from pharmacokinetic-pharmacodynamic (PK-PD) studies can reveal important aspects of a drug’s behavior, including absorption and elimination kinetics, and identify a range of doses where drug concentrations produce therapeutic and toxic effects. Further, these data can be used to optimize dosing in phase III studies where the clinical efficacy of the drug will be most rigorously evaluated. For example, exposure-effect relationships have been explored for abacavir, using population pharmacokinetic data (16
) and for elvitegravir in a randomized, double-blind placebo-controlled trial (17
). Using step-wise additive logistic modeling, McFadyen et al
) identified an average maraviroc concentration of 75 ng/ml as having an 80% probability of producing HIV RNA <50 copies/ml in the MERIT study of treatment-naïve HIV-infected patients. Rosario et al
) developed a comprehensive PK-PD disease model which sought to combine drug effects on viral elimination dynamics from cellular compartments, pharmacokinetics and virologic response to the CCR5 inhibitor maraviroc. This model was designed to incorporate data from various in vivo
and in vitro
sources including maraviroc pharmacokinetics in healthy volunteers and maraviroc viral inhibitory constants in vitro
. This PK-PD disease model was used to estimate a mean decay rate of actively infected cells expressed as the rate of exponential HIV RNA decline, determined to be −0.58 ± 0.12 day−1
at a dose of 300 mg maraviroc BID.
An important finding of our analysis was the observation that viral load reduction was associated with higher vicriviroc plasma concentrations after two weeks of vicriviroc therapy, while remaining on an otherwise failing regimen. Landovitz et al. (15
) found that the reduction in viral load by day 14 correlated with the vicriviroc Cmin
, but that subjects who subsequently failed therapy had a lower mean Cmin
(43.2 vs. 66.2 ng/ml) and AUC (1896.9 vs. 2788.3 ng.h/ml). Consistent with these findings, we observed maximum virologic suppression in subjects achieving a steady-state plasma concentration of 56 ng/ml. The vicriviroc 5 mg arm had a substantial number of subjects who were switched from this dose to higher doses due to suboptimal virologic responses and a trend towards emergence of D/M tropism during the study (14
The most likely explanation for the lack of PK-PD correlation at weeks 16 and 24 is the effect of the optimized background regimen started at week 2. The mean change in viral load from baseline to week two in the pooled vicriviroc groups was −1.13 log10
copies/ml, while the mean change from baseline to weeks 16 and 24 was −1.98 and −2.05 log10
copies/ml, respectively. The additional decrease in HIV RNA comparing weeks 2 with weeks 16 and 24 did not correlate with the assigned dose of vicriviroc. For example, subjects in the 10 mg vicriviroc arm had a greater additional decline in log10
HIV RNA copies/ml than subjects in the highest (15 mg) dose arm at week 16 (10 mg, −1.01 log vs. 15 mg, − 0.47 log), and similarly at week 24 (not shown). Moreover, the majority of subjects had sustained plasma concentrations of vicriviroc above 56 ng/ml at weeks 16 and 24, and this concentration was shown by recursive partitioning analysis to be associated with increased likelihood of viral suppression. This concentration is very close to the in vitro
of vicriviroc against wild-type virus, and this finding is consistent with the observation in studies of other antiretroviral agents that optimal pharmacologic response occurs at trough plasma drug concentrations at or above the IC90
). Valdez et al. (20
) developed a weighted susceptibility score to rank the efficacy of optimized background regimens in treatment experienced patients from the MOTIVATE study who were randomized to maraviroc with OBT or OBT alone. The number of active drugs in the OBT regimen combined with maraviroc was the strongest predictor of virologic suppression at week 48. Collectively, these observations support the importance of the optimized background regimen in the virologic responses seen at weeks 16 and 24. For achieving durable virologic suppression, maintaining plasma concentrations comfortably above of the IC90
may be warranted, particularly in treatment experienced patients. Analysis of larger numbers of treatment-experienced subjects has suggested favorable long-term outcomes in subjects achieving trough concentrations of greater than 100 ng/ml (21
The lack of correlation between pharmacokinetic parameters and virologic response has been observed in a number of HIVtreatment scenarios. For example, Mould et al. (22
) observed that virologic response to enfuvirtide was independent of enfuvirtide plasma concentrations, likely because dosing 90 mg BID produced drug concentrations in the plateau portion of the dose response curve. The lack of correlation between plasma vicriviroc concentration and virologic response at weeks 16 and 24 might suggest that all vicriviroc doses produced concentrations in the plateau range of the dose response curve in the presence of an OBT regimen; however, the inferiority of the 5 mg dose makes this explanation unlikely. There is also the possibility that the maximum effect of vicriviroc was not yet achieved by the two weeks when the optimal background regimen was initiated, but the inferiority of the 5 mg dosing arm makes this explanation unlikely.
As with maraviroc, the response to vicriviroc is optimal in subjects having only R5 virus detected before initiating treatment (23
). While all subjects in our study had R5-only virus at entry, the first generation tropism assay used in this study is slightly less sensitive in detecting dual/mixed and X4 viruses than the currently available enhanced tropism assay (24
). For example, in a re-analysis of the MERIT study, 15% of the subjects with virus classified as R5 (only) prior to receiving maraviroc were re-classified as having dual/mixed tropism virus by the enhanced tropism assay (25
). When these subjects were removed from the analysis, maraviroc demonstrated an effect superior to that observed in the previous analysis of efficacy. Because the predominant quasispecies were R5-only viruses, patients may have an initial virologic response, but subsequently experience virologic failure as the dual/mixed or X4 quasi-species emerge. In our study, a subsequent analysis revealed that 15 out of the 73 subjects designated initially as having R5 virus and randomized to vicriviroc were found to have dual-mixed virus by the enhanced tropism assay (26
). These subjects likely experienced a virologic response initially which gradually diminished and may have contributed to the lack of PK-PD correlation at weeks 16 and 24.
Mutations in the env gene that alter HIV-1 gp120 can confer resistance to maraviroc and vicriviroc (27
). Ogert et al. (28
) have identified key determinants for vicriviroc resistance in the C2-V5 domain of gp-120. Acquisition of these mutations during the vicriviroc “monotherapy” phase might have altered the association between pharmacokinetics and pharmacodynmics at later points in the study. Initially, vicriviroc resistant virus was only identified in one subject in this study (a subject with clade C HIV, reference 29
), but three other cases of resistance were subsequently detected. Since the target of vicriviroc is a host receptor, the ability of chemokine receptor antagonists to modulate the expression of these receptors should be considered as a factor that may affect drug exposure–response relationships, although long-term down-regulation of receptor expression has not been reported.
Structurally diverse small-molecule CCR5 antagonists all appear to bind to the same site within the transmembrane domain of CCR5 (30
). In clinical studies with vicriviroc as short-course montherapy (14 days), suppression of viral load persisted 2–3 days beyond the end of treatment (12
), an effect observed with other CCR5 antagonists (31
). In some studies, viral load continued to decrease after discontinuation of the drug, suggesting prolonged CCR5 receptor occupancy. While the unique effects of these drugs at the receptor could confound the interpretation of PK-PD relationships at 16 and 24 weeks, this is unlikely the case here since such a relationship was observed at week 2.
Consideration of potential pharmacokinetic interactions with other antiretrovirals is important in assessing vicriviroc pharmacokinetics and pharmacodynamics. Metabolic pathways for vicriviroc include N-oxidation, O-demthylation, N,N-dealkylation, N-dealkylation and oxidation to a carboxylic acid . Cytochrome P450 3A4 catalyzed production of all these metabolites whereas CYP3A5 and CYP2C9 produced restricted sets of these metabolites (32
). When dosed at 10 mg QD, vicriviroc exposure is increased to a similar extent by ritonavir at any dose equal to or greater than 100 mg (33
). Co-administration of virciviroc (15 mg QD) with ritonavir-boosted protease inhibitors produced no significant changes in any vicriviroc pharmacokinetic parameters compared to vicriviroc dosed with 100 mg of ritonavir QD or BID (34
). Efavirenz can induce vicriviroc metabolism resulting in significant reductions in vicriviroc AUC and Cmax
when dosed at 10 mg; however, the addition of ritonavir 100 mg attenuates this interaction (33
). The findings from our study can be informative in the selection of doses and the design of phase III studies with vicriviroc in treatment-experienced subjects also taking ritonavir. Matthias et al
) demonstrated that doses of 100 mg of ritonavir can produce maximal “boosting” of plasma concentrations of different CYP3A substrates, including the HIV integrase inhibitor elvitegravir. A dose of 10 mg vicriviroc with 100 mg of ritonavir produced an average Cmin
(91 ng/ml) which is higher than the steady-state plasma concentration we found to be associated with 90% maximal viral suppression (56 ng/ml). The virologic response we observed at this dose of vicriviroc is comparable to the response observed in studies of treatment-naïve subjects using unboosted vicriviroc at higher doses (12
). On the other hand, a dose of 15 mg vicriviroc with ritonavir 100 mg QD produced a higher Cmin
and AUC than the 10 mg dose, but no improvement in virologic response at week 2. Long-term virologic suppression was confirmed in treatment-experienced patients receiving 20 mg (QD) or 30 mg (QD) vicriviroc with ritonavir-boosted PI-containing regimen in a phase 2b study (36
). The PK-PD findings from our analysis, combined with the data identifying the dose producing the most durable suppression in treatment-experienced patients, led to the selection of the 30 mg dose for phase III studies.
We have shown a relationship between the pharmacokinetics of vicriviroc in treatment-experienced subjects and the magnitude of short-term virologic benefit. Our results highlight the importance of optimizing the other antiretrovirals in a combination regimen to achieve durable suppression with vicriviroc. Utilization of enhanced tropism assays as well as genotypic and phenotypic resistance testing to select an optimized background regimen are important adjuncts to maximizing benefit from this class of drugs.