Reduction of CCR5 with Low-Dose Rapamycin Enhances the Antiviral Activity of Vicriviroc Against Both Sensitive and Drug-Resistant HIV-1

Home , Enfuvirtide, Integrase inhibitor, Vicriviroc

Reduction of CCR5 with low-dose rapamycin enhances the antiviral activity of vicriviroc against both sensitive and drug-resistant HIV-1

Alonso Heredia1,2, Olga Latinovic1, Robert C. Gallo2, Gregory Melikyan, Marv Reitz, Nhut Le, and Robert R. Redﬁeld2

Institute of Human Virology, University of Maryland School of Medicine, Baltimore, MD 21201

Contributed by Robert C. Gallo, October 27, 2008 (sent for review August 25, 2008) Vicriviroc (VCV) is a chemokine (C-C motif) receptor 5 (CCR5) We previously have demonstrated that low doses of the antagonist with potent anti-HIV activity that currently is being transplant drug rapamycin (RAPA) reduce CCR5 density, in- evaluated in phase III clinical trials. In the present study, donor hibit R5 HIV-1, and enhance T-20 antiviral activity (14, 15). In CCR5 density (CCR5 receptors/CD4 lymphocytes) inversely corre- the present study we show that the combination of RAPA and /VCV has considerable antiviral synergy. Moreover, the RAPA ؍ lated with VCV antiviral activity (Spearman’s correlation test; r Low doses of the transplant drug rapamycin VCV combination is active against VCV-resistant HIV-1. These .(0.0034 ؍ P ,0.746 (RAPA) reduced CCR5 density and enhanced VCV antiviral activity. findings suggest the use of RAPA in patients as a strategy for In drug interaction studies, the RAPA/VCV combination had con- enhancing the antiviral activity of VCV and perhaps of other siderable antiviral synergy (combination indexes of 0.1–0.04) in CCR5 antagonists as well. both multicycle and single-cycle infection of lymphocytes. The synergy between RAPA and VCV translated into dose reduction Results indexes of 8- to 41-fold reductions for RAPA and 19- to 658-fold Impact of CCR5 Density on the Antiviral Activity of VCV. We first reductions for VCV. RAPA enhanced VCV antiviral activity against evaluated the effect of CCR5 levels on the antiviral activity of both B and non-B clade isolates, potently suppressing clade G VCV using cell lines JC10 (ϳ 2 ϫ 103 CCR5/cell) and JC57 (ϳ viruses with reported reduced sensitivities to VCV and to the 9 ϫ 103 CCR5/cell), which express CD4 levels similar to those in licensed CCR5 antagonist maraviroc. Importantly, RAPA reduction normal CD4 T cells and whose CCR5 densities correspond to the of CCR5 density in lymphocytes sensitized VCV-resistant strains to approximate lower and upper levels of CCR5 expression in ϳ VCV, inhibiting virus production by 90%. We further demon- donors (9–12). We infected each clone with the R5 HIV-1 strains strated the role of CCR5 density on VCV activity against resistant JRFL and CC1/85 (multiplicity of infection, MOI ϭ 0.01) in the virus in donor lymphocytes and in cell lines expressing varying presence of varying concentrations of VCV (0.03–30 nM) for 3 h. CCR5 densities. Together, these results suggest that low doses of For comparison, cells also were infected in the presence of the RAPA may increase the durability of VCV-containing regimens in reverse transcriptase inhibitor efavirenz (0.003–3 nM). On day 4 patients by enhancing VCV viral suppression, by allowing the use after infection, HIV-1 p24 levels in the culture supernatants of lower doses of VCV with reduced potential for toxicity, and by were determined by ELISA (National Cancer Institute). VCV controlling emerging VCV-resistant variants. inhibited HIV-1 JRFL replication in the JC-10 and JC-57 clones with EC values of 0.02 and 0.25 nM, respectively (Fig. 1A). In chemokine receptor 5 antagonists ͉ coreceptor density ͉ HIV resistance ͉ 50 contrast, EC values for efavirenz were similar (less than vicriviroc resistance ͉ maraviroc 50 twofold change) in both cell lines (data not shown). Similar results were obtained with HIV-1 CC1/85 (data not shown). ighly active antiretroviral therapy (HAART) has improved These data suggest that variations in CCR5 density within the Htreatment of HIV-1 infected individuals considerably (1). range observed in most individuals (ϳ 2 ϫ 103-1 ϫ 104 receptors/ However, the success of current therapies is limited by the CD4 T cell) greatly affect the antiviral activity of VCV. emergence of drug-resistant strains, the need for sustained We next evaluated the effect of CCR5 density on VCV activity adherence to complex regimens, and the potential for drug in infectivity assays using lymphocytes from different donors as toxicity (2, 3). The two new antiretroviral classes of integrase and target cells. Coreceptor CCR5 density was quantified for each entry inhibitors may help overcome some of the current limita- donor cell by quantitative flow cytometry analysis. Cells were tions of HAART (4, 5). Entry inhibitors are especially attractive infected with the R5 strain HIV-1 JRFL (MOI ϭ 0.001) for 3 h because they target HIV at the earliest step of the viral cycle and in the presence of varying concentrations of VCV. Virus pro- are effective against strains resistant to inhibitors of protease and duction was measured on day 5 after infection by p24 levels in reverse transcriptase. Entry inhibitors interfere with HIV bind- the supernatants. Results are summarized in Fig. 1B. Among ing to CD4 receptor (attachment inhibitors) or chemokine (C-C different donors, CCR5 density on CD4 T cells varied between 1.9 motif) receptor 5 (CCR5)/chemokine (C-X-C- motif) receptor 4 and 7 ϫ 103 CCR5 receptors/cell. In contrast, measured CD4 (CXCR4) coreceptors (coreceptor antagonists) or by preventing density values were similar among the donors (range of 22.4–27.0 ϫ fusion between cellular and viral membranes (fusion inhibitors) (5). Currently, the fusion inhibitor T-20 (enfuvirtide) and the

CCR5 antagonist maraviroc (Selzentry) are the only licensed Author contributions: A.H., O.L., R.C.G., G.M., and R.R.R. designed research; A.H., O.L., and entry inhibitors (6, 7). The CCR5 antagonist vicriviroc (VCV) N.L. performed research; A.H., O.L., G.M., M.R., and R.R.R. analyzed data; and A.H. and presently is in phase III clinical trials (8). Coreceptor CCR5 R.R.R. wrote the paper. antagonists inhibit CCR5-tropic HIV-1 (referred to as ‘‘R5 The authors declare no conﬂict of interest. HIV-1’’) strains, which are responsible for most transmissions 1These authors contributed equally to this work. and generally are present throughout the course of infection. In 2To whom correspondence may be addressed. E-mail: [email protected], normal individuals, CCR5 density (receptors/cell) on CD4 lym- [email protected], or rredﬁ[email protected]. 3 3 phocytes ranges between ϳ 2 ϫ 10 and ϳ 10 ϫ 10 (9–12), and This article contains supporting information online at www.pnas.org/cgi/content/full/ this variation influences disease progression and response to 0810843106/DCSupplemental. therapy in HIV-1 infection (11, 13). © 2008 by The National Academy of Sciences of the USA

20476–20481 ͉ PNAS ͉ December 23, 2008 ͉ vol. 105 ͉ no. 51 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810843106 Downloaded by guest on September 28, 2021 Fig. 1. CCR5 density impacts the antiviral activity of VCV. (A) Effect of CCR5 density on VCV antiviral activity in cell lines. The HeLa clones HI-JC10 and HI-JC57 were infected with the R5 strain HIV-1 JRFL in the presence of the indicated concentrations of VCV. Virus production was measured by p24 ELISA on day 4 after infection, and data were used to determine EC50 values. Data are means Ϯ SD of two experiments. (B) CCR5 density on primary CD4ϩ cells correlates with antiviral activity of VCV. Activated PBMCs from different donors were subjected to flow cytometry analysis and infection with the R5 strain HIV-1 JRFL in the presence of varying concentrations of VCV. On day 5 Fig. 2. Effect of RAPA on the antiviral activities of VCV and efavirenz in a after infection, virus production was determined by measuring p24 levels. VCV multicycle replication assay. RAPA-pretreated PBMCs were infected with EC50 values were plotted against CCR5 density values on the day of infection. HIV-1 JRFL in the presence of the indicated concentrations of VCV and efa- Line fit was done by linear regression. r, Spearman correlation coefficient. virenz. (A) Virus production was determined by measuring p24 levels on day 5 after infection. (B) Cell viability was determined by the MTT assay on day 5 after infection. Data (means Ϯ SD) are from one representative experiment of 3 10 CD4 receptors/cell) (data not shown). These density values for three independent experiments. CCR5 and CD4 are within the ranges reported by others (9–12). Interestingly, VCV EC50 values varied ϳ 20-fold among the different donors and correlated positively with CCR5 densities (Spear- indexes (CI) at RAPA/VCV ratios of 3/10 and 3/100. We chose man’s correlation test; r ϭ 0.746, P ϭ 0.0034). A similar correlation these drug ratios based on previous determinations of EC50 between CCR5 density and VCV EC50 was obtained using HIV-1 values for each drug. Data were analyzed using the CalcuSyn CC1/85 (data not shown). In a control experiment assessing the software, and results are summarized in Table S1. When used donor cells’ ability to support virus replication, we infected cells alone, RAPA and VCV inhibited virus replication with dose at with the X4 strain HIV-1 IIIb and found p24 values varying by less 50% inhibition (Dm, equivalent to EC50) values of 0.25 and 2.78 than threefold among the different donors (data not shown). nM, respectively. However, when combined at a RAPA/VCV Together, these results using donor cells (Fig. 1B) suggest that ratio of 3/10, the Dm value for RAPA was reduced to 0.03 nM, variations of CCR5 density among individuals considerably affect and the Dm value for VCV was reduced to 0.10 nM. These the antiviral activity of VCV. reductions in Dm values for each drug suggested a synergistic interaction. The corresponding CI values for the combination The Antiviral Activities of RAPA and VCV Are Synergistic. The results ranged from 0.174 to 0.048. Because CI values Ͻ 1 indicate described in the preceding paragraph (Fig. 1) suggested that synergy, and values are proportional to the amount of synergy, reduction of CCR5 density on cells could increase VCV potency. the CI values obtained for the RAPA/VCV combination rep- We previously have shown that the immunomodulatory drug resent a high degree of synergy between the drugs (Table S1). RAPA reduces CCR5 density on CD4 lymphocytes (14). We This synergy translated into dose reduction index values of an 8- thus assessed the antiviral activity of VCV against HIV-1 JRFL to 41-fold reduction for RAPA and a 19- to 658-fold reduction in peripheral blood mononuclear cells (PBMCs) cultured in the for VCV. A similar degree of synergistic interaction was ob- MEDICAL SCIENCES presence of RAPA. As a control, RAPA was used in combina- tained using a RAPA/VCV ratio of 3/100 (data not shown). In tion with efavirenz, a drug that acts intracellularly. Results are contrast, the CI values for the RAPA/efavirenz combination shown in Fig. 2. In the absence of VCV or efavirenz, RAPA (3/10 ratio) ranged from 0.9–1.2, suggesting only an additive inhibited viral replication by up to ϳ 80%, in agreement with our interaction between the drugs. previous studies (14). In combination with 1 nM RAPA, the The antiviral activity of RAPA/VCV was assessed further in amount of VCV required to inhibit virus by 50% was reduced by a single-cycle replication assay based on the infection of lym- approximately fivefold (from 0.5 to 0.15 nM), whereas the amount phocytes with an HIV luciferase reporter virus pseudotyped of efavirenz required for the same effect changed only slightly (from with the R5 JRFL envelope (Env). Cells were infected in the 0.4 to 0.3 nM) (Fig. 2A). Cell viability, assessed by MTT assays on presence of RAPA and VCV, alone and in combination, as day 5 after infection, was not affected by RAPA (Fig. 2B). Similar described previously in the p24 multicycle replication assay. differences between the RAPA/VCV and RAPA/efavirenz com- Pseudovirus replication was determined on day 3 after infection binations were demonstrated using HIV-1 CC1/85 (data not by determining luciferase levels in cell lysates. Data then were shown). These results suggested that RAPA had a greater impact analyzed for synergy using the median-effect principle, and CI on VCV antiviral activity than on efavirenz antiviral activity. values were determined. When RAPA/VCV were tested at To quantify and compare the effect of RAPA on VCV and ratios of 3/10 and 3/100, the CI values ranged from 0.194–0.012 efavirenz activities, we next performed drug interaction studies and from 0.189–0.04, respectively. In contrast, CI values for using the Median Effect Principle (16). To that end, we infected RAPA/efavirenz were ϳ 1, suggesting additivity. These results PBMCs with HIV-1 JRFL (MOI ϭ 0.001) in the absence or are similar to those from the multicycle p24 assay and further presence of both drugs (alone and in combination) over a range demonstrate synergy between RAPA and VCV. of concentrations using a checkerboard design. Drugs were The results of the multicycle and single-cycle infectivity assays tested at threefold dilutions (0.01–3 nM range for RAPA and show that RAPA synergizes with the entry inhibitor VCV. To 0.045–100 nM range for VCV). The antiviral data (p24 produc- evaluate the impact of the RAPA/VCV combination in the early tion) were analyzed for synergy by determining combination steps of viral entry, we tested the drug combination in a cell–cell

Heredia et al. PNAS ͉ December 23, 2008 ͉ vol. 105 ͉ no. 51 ͉ 20477 Downloaded by guest on September 28, 2021 Fig. 4. The RAPA/VCV combination is active against VCV-resistant HIV-1. Fig. 3. Inhibition of R5 HIV-1 cell–cell fusion by the RAPA/VCV combination. Seven-day RAPA-treated PBMCs were infected with replication-competent Fusion was evaluated by determining the extent of fusion between RAPA- D1/85.16 viral clone in the absence and presence of varying VCV concentra- treated lymphocytes (targets) and 293T cells expressing HIV JRFL Env (effec- tions. Viral production was determined by p24 on day 5 after infection. Data tors) by staining with ﬂuorescent dyes and ﬂow cytometry analysis. Data are (% of p24 production compared with VCV-untreated) are mean Ϯ SD of results Ϯ mean SD of two independent experiments, with duplicate replicates in each obtained in PBMCs from four different donors. experiment.

85.16 molecular clone in the presence of varying concentrations fusion assay (17). RAPA-treated lymphocytes and HIV-1 JRFL of VCV and measured p24 production on day 5 postinfection. At Env-transfected cells were stained with fluorescent dyes and the time of infection, cells treated in the absence and presence incubated in the presence of VCV, as described in Methods. of RAPA (0.1, 0.3, and 1 nM) yielded CCR5 density values of Inhibition of cell–cell fusion was evaluated by flow cytometry. 5318, 3900, 3535, and 2791 coreceptors/CD4 T cell, respectively. We evaluated each RAPA concentration in combination with As expected from previous studies (20, 21), VCV alone had no threefold dilutions of VCV. Results are shown in Fig. 3. The impact on virus replication (Fig. 4), probably reflecting viral use VCV EC values at 0, 0.1, 0.3, and 1 nM RAPA were 4.1, 1.9, 50 of VCV-bound CCR5 for infection. In contrast, RAPA alone 0.6, and 0.003 nM, respectively. Comparable reductions in VCV reduced viral replication by 30% to 60%, consistent with viral EC were observed using cells expressing HIV-1 CC1/85 Env 50 dependence on CCR5. Interestingly, when RAPA and VCV (data not shown). These reductions of VCV EC values in the 50 were combined, VCV further increased the RAPA antiviral presence of RAPA are consistent with the results obtained in the effect to ϳ 70% to 90%. Cell viability, as determined by MTT multicycle (p24) and single-cycle (luciferase) infectivity assays assays, was not affected by the drugs (data not shown). These and demonstrate that the RAPA/VCV combination acts early in results suggest that VCV-resistant HIV becomes partially sen- viral entry. sitive to VCV when CCR5 density decreases. To investigate a possible relationship between CCR5 density RAPA Enhances the Antiviral Activity of VCV Against R5 HIV-1 Isolates and the antiviral activity of VCV against VCV-resistant HIV-1, from Different Clades. The experiments described in the previous we evaluated the antiviral effect of 300 nM VCV in donors paragraphs were conducted using the HIV-1 clade B strains JRFL and CC1/85, which have been used frequently in studies expressing different CCR5 densities. As a control, the antiviral evaluating entry inhibitors (18, 19). Because of the remarkable effect also was determined on the sensitive HIV-1 CC1/85 strain. variability in sequence diversity among HIV-1 clades, we also As expected, 300 nM VCV completely inhibited HIV-1 CC1/85 evaluated the antiviral activity of RAPA/VCV in viruses from in all donors (Fig. 5). In contrast, HIV-1 D1/85.16 resistance to non-B clades. For these experiments, we tested the RAPA/VCV VCV was affected by donor CCR5 density, with partial viral combination in PBMCs using replication-competent viral inhibition observed at decreased CCR5 densities. In a control strains. Serial dilutions of VCV were tested in the presence of experiment assessing HIV-1 IIIb (X4 HIV-1) replication, similar RAPA concentrations of 0, 0.1, 0.3, and 1 nM. Virus production p24 levels were obtained in all donors (data not shown). These (p24 levels) and cell viability (MTT assay) were determined on results demonstrate that CCR5 density affects the resistance of day 5 after infection. Results are summarized in Table S2.Inthe HIV-1 D1/85.16 to VCV and further suggest that the RAPA absence of RAPA, VCV EC50 values ranged from 0.09–10 nM. The three clade G isolates tested (RU570, JV1083, and G3) presented the highest EC50 values, in agreement with previous reports demonstrating reduced sensitivity of clade G viruses to VCV (18, 19). In the presence of RAPA concentrations of 0.1, 0.3, and 1 nM, VCV EC50 values were reduced by up to 7.5-, 10-, and 60-fold, respectively. Cell viability was not affected (not shown). These results demonstrate that RAPA enhances the antiviral activity of VCV against isolates from different clades.

The RAPA/VCV Combination Is Active Against the VCV-Resistant HIV-1 D1/85.16 Strain. The HIV-1 strain D1/85.16 is resistant to VCV (20). It is strictly dependent on CCR5 as coreceptor and uses VCV-bound CCR5, albeit less efficiently than free CCR5 (21). Based on D1/85.16 dependence on CCR5 for infection, we reasoned that RAPA could reduce D1/85.16 infection by de- creasing CCR5 density. In addition, reduction of CCR5 density Fig. 5. Density of CCR5 on donor CD4 lymphocytes impacts viral clone by RAPA would increase the proportion of CCR5 receptors D1/85.16 level of resistance to VCV. Donor PBMCs were stimulated by IL-2 occupied by VCV and possibly further reduce the overall effi- treatment for 7 days. On day 7, cells were subjected to CCR5 density evaluation ciency of D1/85.16 infection. To test this possibility, we infected (ﬂow cytometry analysis) and infection in the absence (control) or presence of 7-day RAPA-treated PBMCs with a replication-competent D1/ 300 nM VCV. Results are p24 values measured on day 5 after infection.

20478 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810843106 Heredia et al. Downloaded by guest on September 28, 2021 effect on VCV activity against resistant virus results from We demonstrate that CCR5 density (CCR5 receptors/CD4 decreases in CCR5 density. lymphocyte) in donors inversely correlates with VCV potency. We next investigated the effect of CCR5 density on VCV Specifically, VCV EC50 values were ϳ 20 times higher in donors activity against VCV-resistant HIV-1 by infecting cell lines that with ϳ 7 ϫ 103 CCR5 molecules/cell than in donors with ϳ 2 ϫ express same CD4 levels but varying CCR5 densities with 103 CCR5 molecules/cell. Previous studies with cell lines have replication-defective pseudoviruses bearing Env from CC1/85 shown that CCR5 levels influence the antiviral activity of the (VCV-sensitive) and D1/85.16 (VCV-resistant) clones. We used CCR5 antagonist TAK-779, with higher CCR5 levels resulting in JC57 (ϳ 9 ϫ 103 CCR5/cell) and JC10 (ϳ 2 ϫ 103 CCR5/cell) reduced antagonist activity (24–26). Ketas et al. (27) have shown cells, which express CCR5 densities corresponding to the ap- recently that donor variation in CCR5 levels (determined by proximate upper and lower limits of CCR5 expression on mean fluorescence intensity) correlates with the antiviral activity primary cells. For comparison, we also used JC53 (ϳ 160 ϫ 103 of VCV. We extended these previous studies by showing that CCR5/cell), which overexpresses CCR5. We confirmed these donor CCR5 density (number of receptors/CD4 lymphocyte) CCR5 density values by quantitative flow cytometry analysis. As correlates with sensitivity to VCV. expected, VCV completely inhibited CC1/85 pseudovirus infec- We further show that RAPA, a drug that downregulates tion (data not shown). In contrast, VCV inhibition of D1/85.16 CCR5 density on lymphocytes (14), greatly enhances VCV pseudovirus infection was incompletely inhibited, with maximal activity against R5 HIV-1. In drug-interaction studies, RAPA/ inhibition plateaus of 18- and 265-fold for JC53 and JC57, VCV combinations demonstrated considerable synergistic ac- respectively (Fig. S1A, Left). Interestingly, VCV inhibited D1/ tivity that translated into VCV dose reductions of up to 650-fold. 85.16 pseudovirus infection of JC-10 cells by ϳ 800-fold, to levels RAPA enhanced VCV antiviral activity against primary strains that were undistinguishable from background (determined by of both B and non-B clades. Of note, the combination of RAPA fully inhibitory concentrations of T-20 and efavirenz). These and VCV was especially potent against clade G isolates, which results clearly confirm that D1/85.16 resistance to VCV depends have reduced sensitivities to VCV (19) and maraviroc (18). At upon CCR5 density and suggest that VCV can inhibit VCV- a concentration of 1 nM, RAPA reduced VCV EC50 values by resistant HIV infection at low CCR5 densities. We confirmed 30- to 60-fold against the tested clade G isolates. These exper- the results obtained with D1/85.16 pseudovirus using the Env iments were conducted in PBMCs pretreated with RAPA for 7 from HIV-1 CC101.19, another VCV-resistant virus (20, 22, 23). days because under these conditions RAPA-treated cells express We prepared pseudoviruses bearing the Env of CC101.19 and reduced CCR5 density (15). It is possible that RAPA will reduce infected the same JC clones as before plus JC37 (ϳ 15 ϫ 103 CCR5 at earlier time points, but this effect is difficult to CCR5/cell) and JC24 (ϳ 90 ϫ 103 CCR5/cell). Both JC37 and determine because Ficoll separation of PBMCs alters coreceptor expression levels, which are not restored until after several days JC24 express same CD4 densities as the other JC clones. Results in culture with IL-2 (28, 29). Although we have observed CCR5 are similar to those with D1/85.16 (Fig. S1A, Right). Normal- downregulation by day 7, our results indicate that CCR5 down- ization of inhibition plateaus to infectivity values in the absence regulation in the presence of RAPA is sustained by day 14 of of drug for each cell line yielded 40%, 80%, 92%, 95%, and 100% culture (not shown). Moreover, in vivo administration of RAPA inhibition for JC53, JC24, JC37, JC57, and JC10, respectively. for 12 weeks resulted in sustained downregulation of CCR5 These data demonstrate that VCV inhibition of VCV-resistant levels in macaques (30). viruses correlates with CCR5 density. In addition, as in the It is thought that HIV requires the engagement of several results obtained with HIV-1 D1/85.16 pseudoviruses, VCV (about four to six) CCR5 molecules to form a fusion pore (31). abrogated infection of JC10 by HIV-1 CC101.19 pseudoviruses. We postulate that RAPA reduction of CCR5 density could Together, these results in cell lines confirm the association prevent R5 HIV virions from engaging a minimum number of between VCV activity against resistant HIV and CCR5 density CCR5s required for effective fusion. In the RAPA/VCV com- observed in primary cells. However, the results in cell lines bination, reduction of CCR5 density and partial blocking of demonstrate that VCV can overcome viral resistance at low CCR5 by VCV would effectively prevent CD4 bound viral CCR5 density (ϳ 2 ϫ 103 CCR5/cell, JC10 cells) but only

envelopes from engaging sufficient CCR5 molecules. As a result, MEDICAL SCIENCES ϳ partially ( 90%) overcomes resistance in primary cells express- CD4-triggered envelopes unable to progress in the entry cascade ing CCR5 densities similar to those in JC10 (either in donors would be exposed for prolonged periods of time and subjected with low CCR5 expression or following RAPA treatment (Fig. to inactivation (24). We currently are investigating these mech- 4 and Fig. 5). To determine the basis for this discrepancy, we anisms in an effort to characterize the basis for antiviral synergy evaluated the antiviral activity of VCV in JC10 cells using the by the RAPA/CCR5 antagonist combination. replication-competent HIV molecular clones used in the PBMC We also have shown that RAPA reduction of CCR5 density assay. Virus production was measured by p24 on day 5 after on lymphocytes decreased VCV-resistant HIV-1 D1/85.16 rep- infection. VCV completely inhibited sensitive virus (HIV-1 lication and, in combination with VCV, inhibited viral replica- CC1/85) as determined by p24 levels (Ͻ 6 pg/ml, the assay tion by ϳ 90%. These results suggested that VCV can regain sensitivity limit) (Fig. S2). However, in contrast to the results antiviral activity against VCV-resistant HIV when CCR5 density from pseudovirus infections, VCV inhibited resistant viruses is reduced by RAPA treatment. Consistent with this notion, we D1/85.16 and CC101.19 molecular clones by ϳ 90% and 80%, found that the extent of HIV-1 D1/85.16 resistance to VCV in respectively. These results indicate that VCV inhibition of primary cells was affected by donor CCR5 density, with VCV resistant virus in cells with low CCR5 density (ϳ 2 ϫ 103 CCR5 partially inhibiting VCV-resistant virus in those donors with cells) depends upon the type of assay used. lower densities. Previous studies using viruses resistant to VCV (21) and maraviroc (32) have shown that titration of drug against Discussion the resistant virus in phytohemagglutinin-activated PBMCs gives Coreceptor CCR5 antagonists are a new class of antiretroviral flat inhibition curves. Our results in IL-2-stimulated PBMCs agents with a unique mechanism of action that offer additional (without use of mitogens) reveal that the shape of the inhibition treatment options to patients, especially those infected with curves is affected by the density of CCR5, suggesting that flat drug-resistant HIV-1 (5). In the present study we evaluated the inhibition curves observed in previous studies reflect a higher effect of CCR5 density and its modulation by RAPA on the cellular activation state that leads to CCR5 overexpression. We antiviral activity of the CCR5 antagonist VCV using primary confirmed the impact of CCR5 density on resistant virus resis- cells. tance to VCV in single-cycle infectivity assays using pseudovi-

Heredia et al. PNAS ͉ December 23, 2008 ͉ vol. 105 ͉ no. 51 ͉ 20479 Downloaded by guest on September 28, 2021 ruses bearing Env from two different VCV-resistant viruses VCV will be best used in combination with other antiretrovirals (D1/85.16 and CC101.19 strains) and cell lines expressing varying active against both R5 and X4 HIV-1. CCR5 densities. In these assays, we demonstrated that the RAPA currently is used as an immunosuppressant in kidney magnitude of the maximal inhibition plateau depends on CCR5 transplantation at concentrations ranging from 1–10 nM (41). density, with greater inhibition in cell lines with lower CCR5 Several studies in HIV-infected individuals receiving kidney densities. Interestingly, VCV potently inhibited VCV-resistant transplants and treated with both antiretrovirals and RAPA HIV in JC10 cells that express CCR5 levels (ϳ 2 ϫ 103 suggest that RAPA use in HIV patients is safe (42–44). Our data CCR5/cell) similar to those in primary cells treated with RAPA. (14, 15) and those of others (40) demonstrate that RAPA However, the extent of VCV inhibition of resistant virus at low concentrations in the 0.1–1 nM range inhibit HIV and synergize CCR5 density was greater in pseudovirus infection of cell lines with VCV (this report) in the absence of cell toxicity in vitro, (Ͼ 99%) than in infection of PBMCs with replication-competent suggesting that RAPA doses lower than the ones used in molecular clones (ϳ 90%). Subsequent experiments evaluating transplantation will be effective against HIV in patients. Syn- the antiviral activity of VCV against replication-competent ergistic enhancement of VCV antiviral activity by RAPA may resistant viruses in JC10 cells suggested that the greater inhibi- allow the use of a lower VCV effective dose in patients, thus tion observed in cell lines is not likely to be caused by differences minimizing potential drug toxicities such as hepatotoxicity (34, in CCR5 levels between cell types but rather by the modality of 45, 46). VCV dose reductions may prove particularly beneficial the assay (single-cycle versus multicycle infection). Similar dif- in the treatment of HIV patients co-infected with hepatitis C ferences in drug antiviral activity in the two infectivity assays virus or hepatitis B virus (up to 50% of HIV patients in regions have been reported recently by Heyndrickx, et al. (33). We where the viruses coexist), because these patients are more likely believe that infectivity assays using replication-competent HIV to experience hepatotoxicity (47). In addition, the RAPA/VCV and primary PBMCs are more relevant to physiologic conditions. combination may provide a valuable treatment option for HIV Collectively, these findings may have relevant clinical impli- patients diagnosed with end-stage renal disease and undergoing cations for the use of VCV, and perhaps maraviroc, in the kidney transplantation, because these patients need treatment treatment of HIV-1 infection. Based on guidelines currently with both immunosuppressants and antiretrovirals (44, 48). available for maraviroc (34), it is likely that VCV will be Moreover, our results demonstrating that the RAPA/VCV recommended for use in combination with other antiretrovirals combination is active against VCV-resistant HIV suggest that in treatment-experienced patients infected with R5 HIV-1. Our therapeutic modulation of CCR5 density may help control HIV findings suggest the potential clinical utility of RAPA in VCV- resistance to VCV in patients. Together, these results suggest that containing drug regimens as an approach to reduce CCR5 low doses of RAPA may increase the durability of VCV-containing density and enhance VCV activity against both VCV-sensitive regimens in patients by enhancing VCV viral suppression, by and VCV-resistant R5 HIV-1. Because CCR5 density and VCV allowing the use of lower doses of VCV with reduced potential for EC50 are positively correlated, and RAPA reduces CCR5 density toxicity, and by controlling emerging VCV-resistant variants. to ϳ 2 ϫ 103 CCR5/CD4 T cell (15), it is conceivable that RAPA enhancement of VCV activity will be more apparent in subjects Methods with higher CCR5 levels. The fact that immune activation in HIV Quantification of CD4, CCR5, and CXCR4. Quantification of CCR5, CXCR4 and patients upregulates CCR5 expression (35) suggests that RAPA CD4 was done using Quantibrite-PE beads, as described in SI Methods. will enhance VCV activity significantly in patients. For reference purposes, it is interesting that patient dosing with 10, 25, and 50 Cell Culture, Viruses, and Inhibitors. Conditions for tissue culture (PBMCs, HEK 293 T, and HeLa), viruses, and drugs are described in SI Methods. mg VCV yields Cmax values of 58, 150, and 304 nM, respectively, and potent viral suppression (8). In transplant patients RAPA plasma levels range from 1 to 10 nM. In our studies, testing of Multicycle Infectivity Assay. For infection of PBMCs, cells were cultured in medium containing IL-2 (100 U/ml) and different concentrations of RAPA for RAPA/VCV at ratios of 3/10 and 3/100 revealed significant 7 days. On day 7, cells were exposed to infectious virus for 2 h. Non-adsorbed antiviral synergy, suggesting that dosing with low (1 nM) RAPA virus was removed by washing cells with PBS three times. Infected cells were plasma levels and low (58 nM) VCV plasma levels would result cultured in IL-2 medium in the presence of drugs. Unless otherwise indicated, in a RAPA/VCV ratio of 1/58, which is similar (Ͻ twofold lower) PBMCs were infected using an MOI of 0.001. Virus growth was monitored in to the 3/100 ratio shown to have considerable synergy in our culture supernatants by measuring p24 antigen levels by ELISA (National experiments. Cancer Institute) on day 4 after infection. There is a reasonable concern that the use of CCR5 antagonists in patients with R5 HIV-1 may induce viral mutations Single-Cycle Infectivity Assay. Replication-defective HIV-1 reporter viruses allowing CXCR4 coreceptor use (X4 strains). Results from both were produced by co-transfecting 2 ϫ 106 293T cells with 10 ␮g of pNL4.3- Ϫ ␮ VCV and maraviroc trials in patients infected with R5 HIV-1 env -luc3 and 10 g of Env expression vectors (pSV-JRFLenv, pCI-CC1/85, indicate that X4 variants can be detected in some patients who pCI-D1/85.16, pCI-CC101.19, or pSV-HXB2env [X4]) using the calcium phos- phate protocol, as described (17). Pseudoviruses were collected 48 h after do not respond to antagonist treatment. However, genotypic transfection, spun to remove cellular debris, and filtered to remove virus analyses suggest that such X4 variants already were present, clusters. Virus quantification was by p24 by ELISA. For infection, cells were although at an undetectable frequency, before treatment, and mixed with pseudovirus at a ratio of 10 ng p24/2 ϫ 105 cells. Luciferase activity that their presence became evident after suppression of the was measured at 72 h by luminometry using the Luciferase Assay System dominant R5 strains (36, 37). Thus, therapy with a CCR5 (Promega). Assay background signal was determined by measuring luciferase antagonist may select for pre-existing X4 strains but does not activity upon cell infection in the presence of fully inhibitory concentrations of seem to induce evolution of R5 to X4 strains. Moreover, T-20 (10 ␮M) or reverse transcriptase inhibitor efavirenz (100 nM), and typi- preliminary results show absence of immunological deteriora- cally was 30–100 luciferase units. For normalization of infectivity data, this tion in patients who have detectable X4 strains and do not background signal was subtracted from the signal obtained in the absence of T-20 or efavirenz. respond to maraviroc treatment, suggesting that more patho- genic viruses are not be selected under antagonist pressure (38, Cell–Cell Fusion Entry Assay. We used a cell–cell fusion assay that is based on 39). In any event, it is likely that a drug regimen containing cytoplasmic dye exchange mediated by HIV-1 envelope glycoprotein (49). The RAPA and VCV will be effective in preventing the emergence extent of cell fusion was quantified by flow cytometry, as described previously and replication of X4 strains, because RAPA inhibits both R5 (17). The cell–cell fusion assay and the data analysis are described in SI and X4 viral gene expression (40). Furthermore, RAPA and Methods.

20480 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810843106 Heredia et al. Downloaded by guest on September 28, 2021 ACKNOWLEDGMENTS. We thank Shawn Kuhmann and John Moore Kabat (Oregon Health Sciences University) for the HI-JC cell lines, and (Cornell Medical College, New York) for providing CC1/85, D1/85.16, Nathaniel Landau (New York University School of Medicine) for luciferase and CC101.19 viruses and constructs. We thank Emily Platt and David vectors.

1. Little SJ, et al. (2002) Antiretroviral-drug resistance among patients recently infected 26. Willey S, et al. (2005) Inhibition of CCR5-mediated infection by diverse R5 and R5X4 HIV with HIV. N Engl J Med 347:385–394. and SIV isolates using novel small molecule inhibitors of CCR5: Effects of viral diversity, 2. Ickovics JR, Meade CS (2002) Adherence to HAART among patients with HIV: Break- target cell and receptor density. Antiviral Res 68:96–108. throughs and barriers. AIDS Care 14:309–318. 27. Ketas TJ, et al. (2007) Cell surface expression of CCR5 and other host factors influence 3. Yerly S, et al. (1999) Transmission of antiretroviral-drug-resistant HIV-1 variants. Lancet the inhibition of HIV-1 infection of human lymphocytes by CCR5 ligands. Virology 354:729–733. 364:281–290. 4. Evering TH, Markowitz M (2008) Raltegravir: An integrase inhibitor for HIV-1. Expert 28. Lee B, et al. (1998) Influence of the CCR2–V64I polymorphism on human immunode- Opinion on Investigational Drugs 17:413–422. ficiency virus type 1 coreceptor activity and on chemokine receptor function of CCR2b, 5. Strizki J (2008) Targeting HIV attachment and entry for therapy. Advances in Pharma- CCR3, CCR5, and CXCR4. J Virol 72:7450–7458. cology 56:93–120. 29. Bleul CC, et al. (1997) The HIV coreceptors CXCR4 and CCR5 are differentially expressed 6. Lazzarin A (2005) Enfuvirtide: The first HIV fusion inhibitor. Expert Opinion on Pharmacotherapy 6:453–464. and regulated on human T lymphocytes. Proc Natl Acad Sci USA 94:1925–1930. 7. (2007) FDA notifications. Maraviroc approved as a CCR5 co-receptor antagonist. AIDS 30. Gilliam BL, et al. (2007) Rapamycin reduces CCR5 mRNA levels in macaques: Potential Alert 22:103. applications in HIV-1 prevention and treatment. AIDS 21:2108–2110. 8. Gulick RM, et al. (2007) Phase 2 study of the safety and efficacy of vicriviroc, a CCR5 31. Kuhmann SE, Platt EJ, Kozak SL, Kabat D (2000) Cooperation of multiple CCR5 core- inhibitor, in HIV-1-infected, treatment-experienced patients: AIDS Clinical Trials Group ceptors is required for infections by human immunodeficiency virus type 1. J Virol 5211. J Infect Dis 196:304–312. 74:7005–7015. 9. Hladik F, et al. (1999) Coexpression of CCR5 and IL-2 in human genital but not blood 32. Westby M, et al. (2007) Reduced maximal inhibition in phenotypic susceptibility assays T cells: Implications for the ontogeny of the CCR5ϩ Th1 phenotype. J Immunol indicates that viral strains resistant to the CCR5 antagonist maraviroc utilize inhibitor- 163:2306–2313. bound receptor for entry. J Virol 81:2359–2371. 10. Lee B (1999) Quantification of CD4, CCR5, and CXCR4 levels on lymphocyte subsets, 33. Heyndrickx LT, et al. (2008) Antiviral compounds show enhanced activity in HIV-1 single dendritic cells, and differentially conditioned monocyte-derived macrophages. Proc cycle pseudovirus assays as compared to classical PBMC assays. J Virol Methods Natl Acad Sci USA 96:5215–5220. 148:166–173. 11. Reynes J, et al. (2000) CD4ϩ T cell surface CCR5 density as a determining factor of virus 34. (2008) Selzentry full prescribing information. LAB-0357–1.0. New York: Pfizer Labs, load in persons infected with human immunodeficiency virus type 1. J Infect Dis 2007. Available at: http://media.pfizer.com/files/products/uspi࿝maraviroc.pdf. Ac- 181:927–932. cessed May 29, 2008. 12. Hladik F, et al. (2005) Combined effect of CCR5-Delta32 heterozygosity and the CCR5 promoter polymorphism -2459 A/G on CCR5 expression and resistance to human 35. Ostrowski MA, et al. (1998) Expression of chemokine receptors CXCR4 and CCR5 in immunodeficiency virus type 1 transmission. J Virol 79:11677–11684. HIV-1-infected and uninfected individuals. J Immunol 161:3195–3201. 13. Reynes J, et al. (2004) Relationship between CCR5 density and viral load after discon- 36. Tsibris AM, et al. (2008) In vivo emergence of vicriviroc resistance in a human immu- tinuation of antiretroviral therapy. J Am Med Assoc 291:46. nodeficiency virus type 1 subtype C-infected subject. J Virol 82:8210–8214. 14. Heredia A, et al. (2003) Rapamycin causes down-regulation of CCR5 and accumulation 37. Westby M, et al. (2006) Emergence of CXCR4-using human immunodeficiency virus of anti-HIV beta-chemokines: An approach to suppress R5 strains of HIV-1. Proc Natl type 1 (HIV-1) variants in a minority of HIV-1-infected patients following treatment Acad Sci USA 100:10411–10416. with the CCR5 antagonist maraviroc is from a pretreatment CXCR4-using virus reser- 15. Heredia A, et al. (2007) Rapamycin reduces CCR5 density levels on CD4 T cells, and this voir. J Virol 80:4909–4920. effect results in potentiation of enfuvirtide (T-20) against R5 strains of human immu- 38. Fatkenheuer G, et al. (2008) Subgroup analyses of maraviroc in previously treated R5 nodeficiency virus type 1 in vitro. Antimicrob Agents Chemother 51:2489–2496. HIV-1 infection. N Engl J Med 359:1442–1455. 16. Chou TC (1991) in Synergism and Antagonism in Chemotherapy, eds Chou TC, Rideout 39. Gulick RM, et al. (2008) Maraviroc for previously treated patients with R5 HIV-1 DC (Academic, San Diego, CA), pp 61–102. infection. N Engl J Med 359:1429–1441. 17. Abrahamyan LG, et al. (2005) The cytoplasmic tail slows the folding of human immu- 40. Roy J, Paquette JS, Fortin JF, Tremblay MJ (2002) The immunosuppressant rapamycin nodeficiency virus type 1 Env from a late prebundle configuration into the six-helix represses human immunodeficiency virus type 1 replication. Antimicrob Agents Che- bundle. J Virol 79:106–115. mother 46:3447–3455. 18. Dorr P, et al. (2005) Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti- 41. Augustine JJ, Bodziak KA, Hricik DE (2007) Use of sirolimus in solid organ transplan- human immunodeficiency virus type 1 activity. Antimicrob Agents Chemother tation. Drugs 67:369–391. 49:4721–4732. 42. Kumar MS, et al. (2005) Safety and success of kidney transplantation and concomitant 19. Strizki JM, et al. (2005) Discovery and characterization of vicriviroc (SCH 417690), a immunosuppression in HIV-positive patients. Kidney Int 67:1622–1629. CCR5 antagonist with potent activity against human immunodeficiency virus type 1. 43. Moreno A, et al. (2008) Safe use of raltegravir and sirolimus in an HIV-infected patient Antimicrob Agents Chemother 49:4911–4919. with renal impairment after orthotopic liver transplantation. AIDS 22:547–548.

20. Marozsan AJ, et al. (2005) Generation and properties of a human immunodeficiency 44. Roland ME, et al. (2008) HIV-infected liver and kidney transplant recipients: 1- and MEDICAL SCIENCES virus type 1 isolate resistant to the small molecule CCR5 inhibitor, SCH-417690 (SCH-D). 3-year outcomes. American Journal of Transplantation 8:355–365. Virology 338:182–199. 45. Cornwell PD, Ulrich RG (2007) Investigating the mechanistic basis for hepatic toxicity 21. Pugach P, et al. (2007) HIV-1 clones resistant to a small molecule CCR5 inhibitor use the induced by an experimental chemokine receptor 5 (CCR5) antagonist using a compen- inhibitor-bound form of CCR5 for entry. Virology 361:212–228. dium of gene expression profiles. Toxicologic Pathology 35:576–588. 22. Kuhmann SE, et al. (2004) Genetic and phenotypic analyses of human immunodefi- 46. Nichols WG, et al. (2008) Hepatotoxicity observed in clinical trials of aplaviroc ciency virus type 1 escape from a small-molecule CCR5 inhibitor. J Virol 78:2790–2807. (GW873140). Antimicrob Agents Chemother 52:858–865. 23. Trkola A, et al. (2002) HIV-1 escape from a small molecule, CCR5-specific entry inhibitor 47. Sulkowski MS (2008) Viral hepatitis and HIV coinfection. J Hepatol 48:353–367. does not involve CXCR4 use. Proc Natl Acad Sci USA 99:395–400. 48. Gupta SK, et al. (2005) Guidelines for the management of chronic kidney disease in 24. Platt EJ, Durnin JP, Kabat D (2005) Kinetic factors control efficiencies of cell entry, efficacies of entry inhibitors, and mechanisms of adaptation of human immunodefi- HIV-infected patients: Recommendations of the HIV Medicine Association of the ciency virus. J Virol 79:4347–4356. Infectious Diseases Society of America. Clinical Infectious Diseases 40:1559–1585. 25. Reeves JD, et al. (2002) Sensitivity of HIV-1 to entry inhibitors correlates with envelope/ 49. Melikyan GB, et al. (2000) Evidence that the transition of HIV-1 gp41 into a six-helix coreceptor affinity, receptor density, and fusion kinetics. Proc Natl Acad Sci USA bundle, not the bundle configuration, induces membrane fusion. J Cell Biol 151:413– 99:16249–16254. 423.

Heredia et al. PNAS ͉ December 23, 2008 ͉ vol. 105 ͉ no. 51 ͉ 20481 Downloaded by guest on September 28, 2021