We designed and synthesized small molecule compounds as candidates for novel CCR5 inhibitors, and identified several compounds that have potent activity against wild type R5-HIV-1. GRL-117C exerted potent activity against R5-HIV-1Ba-L with a sub-nanomolar IC50 value in the MAGI assay using MAGI/CCR5 cells. The potency (IC50 values) of GRL-117C was comparable to that of MVC, as was determined by both the MAGI assay (0.6 nM vs. 0.7 nM) and the p24 assay with PBMCs (8.1 nM vs. 4.5 nM). APL demonstrated similar or slightly more potent activity than MVC, and its IC50 values were 0.2 nM and 2.6 nM for the MAGI and p24 assays, respectively. The other GRL-compounds, GRL-10007C and GRL-10018C, also demonstrated strong activity against HIV-1Ba-L in the MAGI assay (IC50: 1.4 nM and 2.9 nM, respectively). These compounds were found to be more potent as compared with the two previously published experimental CCR5 inhibitors, SCH-C and TAK-779, but were less effective than MVC and APL. Two drug-naive clinical R5-HIV-1 strains, CC1/85 cl.6 and cl.7, were also used in the assays. All the compounds tested in this study showed similar effectiveness against the CC1/85 clinical strains as compared with that against HIV-1Ba-L. We have previously observed that the IC50 value(s) of CCR5 inhibitors in MAGI assays tended to be lower as compared with those obtained via the p24 assays in PBMCs. In this study, we also observed the same trend. For example, the IC50 value of GRL-117C for the MAGI assay was 0.6 nM, but was 8.1 nM for the p24 assay (HIV-1Ba-L). We chose three compounds (GRL-117C, GRL-10007C, and GRL-10018C) for further testing. In a previous study, Trkola et al. reported the generation of an escape mutant HIV-1 for AD101 (experimental CCR5 inhibitor), and found that this mutant did not use CXCR4, but instead gained the ability to use CCR5 in an AD101-insensitive manner. Subsequently, Marozsan et al. described the generation of escape mutants under the selection pressure of VVC in vitro. Both escape mutants were fully resistant against AD101 and VVC. For the current study, AD101- and VVC-resistant HIV clones were provided by Dr. John P. Moore of Cornel University. CC101.19 (AD101-resistant) was approximately 150-fold more resistant to SCH-C (IC50: 1000 nM) as compared with its corresponding CCR5 inhibitor-sensitive viruses, CC1/85 (cl.6 and cl7, IC50: 5.2 and 6.1 nM, respectively). On the other hand, resistance against other CCR5 inhibitors, including MVC, APL, and GRL-compounds, were relatively lower in comparison; fold resistance ranged from 2.6- to 15-fold. The VVC-resistant virus (D1/85.16) also showed high resistance against SCH-C (68-fold), but remained susceptible to all other drugs to some extent (fold resistance: 3.6-fold to 12.5-fold). GRL-117C exhibited slightly decreased activity against AD101- and VVC-resistant viruses (fold resistance: 9.3- and 8.5-fold, respectively), however, its IC50 numbers remained less than 40 nM. Interestingly, GRL-10007C, which was less reactive than GRL-117C against wild type R5-HIV-1, maintained its activity against AD101- and VVC-resistant viruses, showing IC50 values of 41.1 nM (2.6-fold) and 56.9 nM (3.6-fold). This result suggested that the resistance profiles of SCH-C and its associated drugs (VVC and AD101) differ drastically from those of MVC, APL, and GRL-derivatives. GRL-10007C, which induced the least resistance in these viruses, may have a unique resistance profile among the CCR5 inhibitors tested in this study. We then wanted to determine if these compounds are effective against HIV-1s carrying MVC-resistance-associated substitutions. As shown in Figure S1 and Table S1, activity of GRL-117C was reduced when used against the highly MVC-resistant virus (HIV-1KP-5mvcR) (IC50: 686 nM). However, GRL-117C also demonstrated decreased activity against a drug-naive HIV-1 clinical strain (HIV-1KP-5pc) as compared to the laboratory HIV-1 strain (HIV-1YU2). Therefore, while the fold change of IC50 values for GRL-117C was only 4.8 between HIV-1KP-5pc and HIV-1KP-5mvcR, we hypothesized that GRL-117C may show cross-resistance with MVC, because its IC50 value against HIV-1KP-5mvcR (686 nM) was more than 10-fold greater than that of MVC (41 nM). The other derivatives, GRL-10007C and -10018C also failed to demonstrate activity against HIV-1KP-5mvcR. It is of note that the activity of TBR-652 (cenicriviroc or CVC) against HIV-1KP-5mvcR was substantially decreased as compared to that against wild type [IC50: 260 nM vs. 4.1 nM (x63-fold)], indicating that it has cross-resistance with MVC. In order to determine whether GRL derivatives block the binding of CC-chemokines to CCR5, we conducted a CC-chemokine binding inhibition assay using 125I-labeled CC-chemokines (125I-RANTES, 125I-MIP-1-alha, and 125I-MIP-1-beta and CCR5 expressing cells. All the CCR5 inhibitors tested [GRL-117C, GRL-10007C, GRL-10018C, MVC, and APL] blocked the binding of 125I MIP-1-alha to CCR5, and their EC50 values ranged from 0.1-4.3 nM. Similar results were observed for MIP-1-beta binding (EC50 range: 0.2-2.5 nM). Results demonstrated that MVC, APL, and GRL-10018C exert stronger inhibitory effects on MIP-1-alha and MIP-1-beta binding as compared with GRL-117C and GRL-10007C. In contrast, APL, GRL-117C, and GRL-10007C only moderately blocked the binding of RANTES; their EC50 values were 156, 121, and 628 nM, respectively, and binding of 125I-RANTES remained at more than 40% even in the presence of 1 microM of GRL-10007C. We have previously reported that APL does not effectively inhibit the binding and function of RANTES, even though it binds to CCR5. It is possible that GRL-10007C and GRL-117C also have similar profiles as APL in terms of their role in CC-chemokine to CCR5 binding. Three-dimensional models of human CCR5-CCR5 inhibitor complexes were defined using the crystal structure of CCR5-MVC as the template (PDB ID: 4MBS). As was previously reported by Tan et al., MVC was found to be lodged in the bottom of the largest pocket at the binding site, which were defined by residues from helices 1, 2, 3, 5, 6, and 7. It was observed that MVC forms hydrogen bonds with Tyr-37, Tyr-251, and Glu-283, and its phenyl group reaches deep into the pocket to form hydrophobic interactions with aromatic residues such as Tyr-108, Trp-248, and Tyr-251. As shown in Figure 3B, GRL-117C also binds to the same binding cavity, and similar to MVC, GRL-117C forms hydrogen bonds with Tyr-37 and Glu-283, but not with Tyr-251. Binding models of GRL-10018C with CCR5 also showed formation of hydrogen bods between GRL-10018C and Tyr-37 and Glu-283. As was described in the previous section, GRL-10018C exhibited more potent inhibitory effect on the binding of the three chemokines as compared with GRL-117C and GRL-10007C. This may be due to the fact that GRL-10018C possesses a bis-THF structure. The bulky rings occupying the upper region of the binding cavity under ECL2 cause steric hindrance with CC-chemokine when it binds to CCR5. However, the bis-THF structure does not affect the interaction between gp120 with CCR5. Moreover, as was previously reported by Tan et al., the phenyl group of MVC form hydrophobic interactions with Trp-248, thus preventing its activation-related motion. Similarly, GRL-117C and GRL-10018 also have a phenyl ring in the center of their structure that forms hydrophobic interaction with Trp-248; it is therefore possible that GRL-derivatives also prevent activation-related motion of CCR5.
