HIV-associated neurocognitive disorders (HAND) are seen as a result of antiretroviral therapy (ART)-related prolonged survival and increasing ages of HIV-infected individuals. Reports say that despite the success of ART, up to 80% of HIV-infected individuals are expected to develop HAND & other central nervous system (CNS) disorders in their life. Most anti-HIV drugs (except AZT) only poorly penetrate into the CNS, an HIV's sanctuary site, apparently causing various CNS disorders despite well-suppressed systemic HIV replication. Thus, the development of anti-HIV drugs that reach optimal therapeutic concentrations in the CNS is urgently needed. In this regard, fluorination is known to increase metabolic stability, delays inactivation of drugs, and elongates dosage periods because the C-F bond is highly stable. Fluorination also increases lipophilicity due to its greater hydrophobicity than the C-H bond, often increasing cell membrane penetration and oral bioavailability. We also hoped that fluorination increases CNS penetration of drugs as well. We synthesized various fluorinated PIs based on the structure of DRV and identified GRL-048 and GRL-050 in collaboration with Professor Arun K. Ghosh of Purdue University. These two compounds exerted highly potent activity against a wide spectrum of laboratory HIV-1 strains (EC50 values of 0.8 and 3 nM, respectively) and primary clinical isolates including multi-PI-resistant variants with minimal cytotoxicity. We selected HIV-1 variants with GRL-048 and GRL-050, by propagating a laboratory wild-type HIV-1NL4-3, in MT-4 cells in the presence of increasing concentrations of GRL-048 and GRL-050, and determined amino acid substitutions that emerged under the pressure of these compounds in the protease-encoding region. Finally, in view of the limited penetration of most antiviral drugs into the CNS, we evaluated the partition and distribution coefficients (logP and logD) as well as the apparent Blod Brain Barier (BBB) permeability coefficient (Papp) using an in vitro model, where we were able to demonstrate that GRL-048 and GRL-050 had a potentially enhanced penetration capability across the BBB. We also identified GRL-0739, which contains the cyclohexyl-bis-THF moiety and a sulfonamide isostere. GRL-0739 exerted potent activity against a wide spectrum of laboratory HIV-1 strains and primary clinical isolates including multi-HIV-1 PI-resistant variants with minimal cytotoxicity. In addition, GRL-0739 was active against HIV-2ROD as well as HIV-1 isolates examined. We also selected drug-resistant HIV-1 variants with GRL-0739, by propagating a laboratory wild-type HIV-1NL4-3 in MT-4 cells in the presence of increasing concentrations of GRL-0739, and determined the amino acids substitutions that emerged under the pressure of GRL-0739 in the protease-encoding region. In addition, we evaluated non-specific binding effects of physiologic human serum proteins on GRL-0739's anti-HIV-1 activity. Finally, GRL-0739 showed a favorable BBB penetration property in the in vitro BBB model. We have most recently identified GRL-0142, which contains bridged bis (bbis)-THF and two fluorine atoms in the P2 benzene moiety and exerts unprecedentedly potent activity against a variety of HIV-1 variants resistant to all PIs tested including HIV-1DRVRP51. GRL-0121 also exerts potent activity against wild-type HIV-1 and HIV-2 strains, although its activity was lesser than that of GRL-0142. Notably, the selectivity index of GRL-0142 is as great as over one million as compared to 20,063 of DRV. More intriguingly, GRL-0142 exerted unprecedentedly potent activity against all the PI-resistant HIV-1 variants. The EC50 values of GRL-0142 recorded around single-digit fM levels against HIV-1SQV-5microM, HIV-1LPV-5microM, and HIV-1IDV-5microM. In order to understand the mechanism(s) of such potent anti-HIV-1 activity, we determined whether GRL-0142 had protease dimerization inhibition activity using the FRET-based HIV-1 expression assay. The results revealed that GRL-0142 has much greater activity to block the dimerization process of HIV-1 protease by a factor of 1,000 as compared to DRV. We also expressed and purified various HIV-1 protease species including wild-type protease and crystallized them in complex with GRL-0121 and GRL-0142. We determined the binding configuration and interactions of GRL-0142 with the HIV-1 protease protein by X-ray crystallography. GRL-0142 binds at the usual active site formed by the protease dimer. GRL-0142 has many polar and non-polar interactions with HIV-1 protease. The oxygens of the bbis-THF moiety form hydrogen bond interactions with the backbone nitrogen atoms of Asp29 and Asp30, located in the S2 binding site of protease. The thiazole nitrogen forms hydrogen bond interaction with the backbone carbonyl oxygen of Asp30' located in the S2' site of protease. An amide nitrogen of GRL-0142 forms hydrogen bond interaction with the backbone carbonyl of Gly27. We have previously discussed that polar interactions of PIs with protease backbone atoms is not only responsible for higher potency against wild-type HIV-1, but also against drug-resistant HIV-1. This is because the protease backbone atoms stay invariant with amino acid substitutions for resistant protease and should be able to maintain the polar interactions with inhibitors in spite of drug resistant substitutions. The hydroxyl of GRL-0142 has hydrogen bond interactions with the catalytic aspartates Asp25 and Asp25'. The carbonyl and sulfonyl oxygens have polar interaction with Ile50 and Ile50' mediated through a bridging water molecule. This interaction with Ile50 and Ile50' located in the protease flap, through a water molecule is seen for many non-peptidic PIs. There is also a hydrogen bond from the amine nitrogen to the side chain carboxylate of Asp30' in the S2' site of protease. Besides these polar interactions, GRL-0142 has a number of van der Waals interactions with HIV-1 protease. A residue by residue analysis of the crystal structures of GRL-0142 and DRV bound with their respective proteases revealed that GRL-0142 has additional contacts with many key residues of protease. Compared to DRV, GRL-0142 has more than 90 favorable good contacts. The bbis-THF moiety of GRL-0142 has 27 and 28 good contacts with Asp29 and Asp30, respectively. The bis-THF moiety of DRV has 22 and 15 good contacts with Asp29 and Asp30, respectively. The bbis-THF group of GRL-0142 has 21 good contacts each with Ile47, and Ile50', whereas bis-THF of DRV has 16, and 3 good contacts with these residues. The P1 phenyl group of GRL-0142 has two fluorine substituents and they form strong contacts with Ile50' of the protease. Overall, the additional interactions by the bridged carbons, two fluorines, and P2' moiety of GRL-0142 is responsible for its extremely high potency against wild-type as well as mutant protease.
Showing the most recent 10 out of 59 publications
