Since our first report of darunavir (DRV) in 2003, we continued optimization based on the structure of DRV, seeking novel protease inhibitors (PIs) that are more potent against a variety of existing multi-PI-resistant HIV-1 variants with greater safety, do not permit or substantially delay the emergence of HIV-1 variants resistant to the very PIs, and favorably penetrate into the CNS, and identified GRL-142. GRL-142 contains newly generated pharmacophores such as an unprecedented 6-5-5 ring-fused crown-like tetrahydropyranofuran as the P2-ligand (Crn-THF), P1-bis-fluoro-benzyl (bis-Fbz), and P2'-cyclopropyl-amino-benzothiazole (Cp-Abt). We examined the activity of five PIs including GRL-121 and -142 against seven resistant HIV-1 variants, which we had previously selected in vitro with each of the seven FDA-approved PIs (invitroHIVPIRs). Most of the seven variants were significantly less susceptible to two PIs, lopinavir (LPV) and ATV, that have presently been relatively well used in clinics. DRV also failed to effectively block most of the seven variants with IC50 value fold-differences ranging from 2- to 86-fold. However, GRL-121 showed extremely potent activity against all the seven variants examined, presenting IC50 values ranging 0.0018 to 0.13 nM. The activity of GRL-121 against all the seven variants was significantly more potent than that against cHIVNL4-3WT. Surprisingly, GRL-142 showed even more potent activity against the seven variants with IC50 values of 0.0000019 nM (1.9 fM) to 0.015 nM. The activity of GRL-142 against the variants was also significantly more potent than that against cHIVNL4-3WT. We, furthermore, examined the activity of the five PIs against six recombinant infectious clinical HIV-1 variants (rCLHIVs) that are highly resistant to all the currently available PIs including DRV. GRL-121 exerted highly potent activity with IC50 values of 0.028 to 12 nM, while GRL-142 again showed even more potent activity with IC50 values of 0.0052 to 0.69 nM. Moreover, we carried out assays of DRV, GRL-121, and GRL-142 against 18 recombinant HIV-1 variants carrying a single amino acid substitution known to be associated to HIV-1 resistance to various PIs. DRV was effective against all the recombinant clones with IC50 values of 0.29 to 4.7 nM. GRL-121 was again found to be highly potent against all the variants examined with IC50 values of 0.015 to 350 pM. GRL-142 was even more potent against the variants as compared to GRL-121. It is noted that GRL-142 was extremely potent against three recombinant HIV-1 variants (cHIVNL4-3V32I, cHIVNL4-3G48V, and cHIVNL4-3I50V) with IC50 values of 12, 36, and 93 attomolar (aM), respectively. We determined the structural interactions of GRL-142 with wild-type HIV-1 protease (PRWT) using X-ray crystallography. GRL-142 binds in the active site of PRWT in two distinct conformations (related by 180-degrees rotation) with relative occupancies of 0.53 and 0.47. The structural description is derived from the interactions of the major conformation of GRL-142 with PRWT. GRL-142 has a Crn-THF as the P2-ligand moiety and a Cp-Abt as the P2' ligand moiety. Both groups make critical interactions with amino acids spanning distinct regions of PRWT active site. The Crn-THF moiety has two oxygen atoms and they both form hydrogen bonding interactions with the backbone amides of D29 and D30. The thiazole nitrogen makes a hydrogen bond interaction with the backbone NH of D30'. The P2' amino group forms polar interactions with the sidechain carboxylate of D30'. The carbonyl and sulfonyl oxygens have polar interactions with the PR flap residues I50 and I50' through a bridging water molecule. The transition state mimic hydroxyl group forms polar interactions with the catalytic aspartates D25 and D25'. There is another hydrogen bond interaction from the amide nitrogen of the carbamate moiety to the backbone carbonyl oxygen of G27. Many of these polar interactions with HIV protease are also seen in DRV complexed with PRWT. The P1-phenyl moiety of GRL-142 has two fluorine atoms that form critical interactions with PRWT that are not formed by DRV. The crystal structure indicates that fluorine substitution causes very favorable halogen interactions within the S1-pocket of PRWT. The high electronegativity of fluorine (3.98) versus hydrogen (2.20) leads to highly polarized halogen bond interactions with the backbone (F-H-C and F-C=O) of the G49 and I50, respectively. Contribution of fluorine atoms to the binding affinity of GRL-142 is not limited only by backbone interactions, but they also form close contacts with the H-atoms of C-gamma and C-delta of P81'. Overall, the two fluorine atoms establish a halogen bond bridge within the S1-pocket by interacting with residues of both subunits of PRWT. Moreover, direct interactions of GRL-142 with both R8' and I50 highly likely contribute significant additional polar interactions compared to other PIs. R8' and the PR's flap residues are associated with PR structure, functions, dynamics, substrate binding, and activity of PIs. We also compared the differences in vdW contacts of GRL-142 with V32I amino acid substitution, a key substitution for HIV-1's acquisition of PI-resistance. As shown in Fig. 3D, GRL-142 has good vdW contacts with wild-type V32; however, vdW interactions proved to be much better with mutated I32, especially the interactions of GRL-142 with I32'. Our simulations showed that I32 and I32' have an improved vdW interaction energy (1-kcal/mol) with GRL-142 compared to V32 and V32'. These features should at least partly explain the significantly greater activity of GRL-142 to HIV-1 variants containing the V32I substitution in their PR. Persistent HIV-1 replication and inflammation in the CNS, which can occur even in patients receiving cART with an undetectable plasma viral load, is most likely responsible for HAND. Hence, we finally quantified GRL-142 concentrations in plasma, cerebrospinal fluid (CSF), and brain of rats (n=2) and compared those figures with those of DRV obtained under the same conditions. When DRV was perorally administered at a dose of 5 mg/kg together with RTV (8.33 mg/kg), the Cmax was achieved around 90 min after the PO administration. The DRV concentrations determined in 15 and 90 min after the peroral administration turned out to be 0.595 micro and 0.847 microM in plasma; 0.00100 microM and 0.00116 micro in CSF; and 0.0110 micro and 0.0157 micro in brain, respectively. The plasma concentrations of DRV were much greater than the IC50 value (3.2 nM); however, concentrations in CSF were lower than the IC50 value and those in brain were slightly above the DRV IC50 value but still substantially lower than the DRV IC95 value of DRV (0.3 microM), suggesting that DRV likely fail blocking the replication of HIV-1 in the CNS. By contrast, when GRL-142 was perorally administered at a dose of 5 mg/kg together with RTV (8.33 mg/kg), the Cmax was achieved around 360 min after the PO administration. Plasma samples collected 60 and 360 min after the PO administration contained 0.189 microM GRL-142 and 0.974 micro, respectively. CSF samples contain below detection levels and 0.000532 microM); brain contained 0.00724 microM and 0.0326 micro in 60 and 360 min, respectively. Since the IC50 and IC95 values of GRL-142 are 19 pM and 0.28 nM, GRL-142 concentrations in brain are calculated to be 1,882-fold greater than the IC50 value of GRL-142 and 114-fold greater than IC95 of GRL-142, while 562-fold lower than CC50 of GRL-142. These data strongly suggest that GRL-142 would potently block the infection and replication of HIV-1 in the brain.
Showing the most recent 10 out of 59 publications
