Gregg S Jones, Rebecca Hluhanich, Lani M Wieman, Jim Zheng, Wayne Huang, April Kinkade, Eugene J Eisenberg, Chris Yang, Damian McColl, Anita Mathias, Haolun Jin, Tomas Cihlar, Matthew Wright, Romas Geleziunas and Manuel Tsiang*
Gilead Sciences, Foster City, CA 94404, USA
Received Date: June 04, 2014; Accepted Date: July 10, 2014; Published Date: July 18, 2014
Citation: Jones GS, Hluhanich R, Wieman LM, Zheng J, Huang W, et al. (2014) Preclinical and Clinical Profile of HIV-1 Integrase Strand-transfer Inhibitor GS-9224 Compared to its Parent Compound GS-9160. J Antivir Antiretrovir 6:075-083. doi: 10.4172/jaa.10000101
Copyright: © 2014 Jones GS, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Journal of Antivirals & Antiretrovirals
In an effort to optimize the pharmacokinetic profile of GS-9160, a potent (EC50=1.2 – 4 nM) antiviral tri-cyclic
HIV-1 integrase (IN) strand transfer inhibitor (INSTI), various substitutions on the p-fluorobenzyl moiety of GS-9160
were explored. This effort led to the discovery of GS-9224, an analog containing a 5-chloro-2,4-di-fluorobenzyl
moiety. GS-9224, like its predecessor GS-9160, has potent (EC50=1.3 – 9 nM) and selective antiviral activity against
HIV-1 and acts as a bona fide integration inhibitor through elevation of 2-long terminal repeat (2-LTR) circles and
decrease of integration junctions in HIV-1 infected cells, markers of failed viral integration. Viral resistance selections
with GS-9224 yielded three mutations within the catalytic core domain of HIV-1 IN: G140S, L74M and Q148K.
When tested against a panel of HIV-1 clones engineered to express INSTI resistance mutations, the profile of
GS-9224 was comparable to GS-9160. GS-9224 exhibits improved hepatic microsome stability, better absorption
potential and a lower in vitro intestinal efflux ratio compared to GS-9160. In addition, GS-9224 displayed higher
dog plasma protein binding (~99%) than GS-9160 (~93%) suggesting a potential for protein restricted clearance.
However, the improved metabolic stability, absorption potential and serum protein binding of GS-9224 compared
to GS-9160 did not translate into an improved clinical pharmacokinetic profile. Results from a single-dose study in
human healthy subjects revealed a fast systemic clearance of GS-9224 with a terminal half-life of ~ 1 h, resulting in
a pharmacokinetic profile that would not support once-daily dosing in HIV-infected patients.
HIV-1; GS-9224; GS-9160; Strand-transfer
As an essential enzyme for HIV-1 replication, integrase (IN) is an attractive target for antiviral drug discovery. During the viral life cycle, the primary function of IN is to catalyze the insertion of reversetranscribed double stranded HIV-1 DNA into the host cell genome to establish the provirus. Inhibition of the strand transfer activity of IN as a target for therapeutic intervention is now well validated clinically [1,2]. Raltegravir (RAL) was the first IN strand-transfer inhibitor (INSTI) to be approved for clinical use and is dosed twice daily [3-5]. Elvitegravir (EVG) is co-administered with cobicistat, a new pharmacokinetic enhancer that enables once-daily dosing of EVG [6-8]. A once-daily single tablet regimen (STRIBILDTM) containing EVG, cobicistat, deoxyfluorothiacytidine (FTC) and tenofovir disoproxil fumarate (TDF) has been recently approved by the FDA for the treatment of HIV infection. In two Phase III studies, the STRIBILD regimen showed noninferior efficacy and safety in treatment-naïve HIV-infected patients when compared head-to-head with either efavirenz  or atazanavir  containing once-daily treatment regimens. Recent data from the 96-week extension analysis of both studies established the longer term durability and efficacy of the STRIBILD regimen [11,12].
Specific mutations in IN conferring cross-resistance to RAL and EVG have been identified in clinical settings [13-15]. The INSTI dolutegravir (S/GSK1349572), which is dosed once-daily without boosting, displayed a resistance profile that is distinct from those of RAL and EVG [16,17]. In phase III clinical studies, none of the subjects treated with the combination of dolutegravir and TDF/FTC for up to 48 weeks developed any treatment-associated IN mutations .
In an effort to identify additional novel INSTIs that could be dosed once-daily without pharmacokinetic enhancement we explored modifications of a tricyclic analog GS-9160, a potent and orally bioavailable compound [19-21]. Substitutions on the p-fluorobenzyl moiety of GS-9160 led to the discovery of GS-9224, which contains a 5-chloro-2,4-di-fluorobenzyl moiety . In this report, we describe detailed biological and metabolic characterization of GS-9224 including a clinical pharmacokinetic profile in healthy human volunteers.
GS-9224 and GS-9160 were synthesized at Gilead Sciences, Inc. L-870,810 was synthesized by Combi Blocks, Inc. (San Diego, CA). RAL was synthesized by Naeja Pharmaceuticals, Inc. (Edmonton, Alberta Canada). All three compounds have a purity of >95% as determined by analytical reversed phase HPLC using a Phenomenex Luna C18 column and UV detection at 214 nm and 254 nm. The structures of these three INSTIs are shown in (Figure 1).
MT-2 cells were obtained from Stanford University and MT-4 cells were obtained from the NIH AIDS Research and Reference Reagent Program (Germantown, MD). SupT1 cells were obtained from the American Type Culture Collection (Rockville, MD). MT-2, MT-4 and SupT1 cells were maintained in RPMI-1640 medium (Cat # 61870- 036, Life Technologies, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Cat # SH30071.03, GE Healthcare Sciences, Logan, UT) and antibiotics. Peripheral blood monocytes (PBMCs) were isolated from the buffy coats of healthy human volunteers (Stanford Medical School Blood Center, Stanford, CA).
HIV-1 strain IIIb was obtained from Advanced Biotechnologies Inc. (Columbia, MD). The NRTI-resistant virus containing mutation K65R in reverse transcriptase (RT) was obtained from Mark Wainberg (McGill AIDS Center, Montreal, Canada). The NRTI-resistant virus RT-M184V and NNRTI-resistant viruses RT-K103N and RT-Y181C were constructed by site-directed mutagenesis . The NRTIresistant virus, RT-6TAMs (M41L, D67N, K70R, L210W, T215Y and K219Q) which confers resistance to thymidine analogs  as well as the protease (PR) inhibitor-resistant viruses, PR-I84V/L90M and PRG48V/ V82A/L90M were constructed by homologous recombination in electroporated SupT1 cells . All the above mutant viruses were constructed from an HXB2 clone.
Construction of infectious HIV-1 DNA clones with IN mutations
Mutations conferring resistance to INSTI were introduced into the infectious HIV-1 DNA clone HXB2 by site-directed mutagenesis at Gilead Sciences by our investigators . Briefly, the 15.5 Kb HXB2 plasmid DNA was cut with ApaI and SalI to generate a 3.7 Kb ApaI – SalI fragment that was gel-purified. This 3.7 Kb fragment containing both the RT and the IN coding regions was then subcloned into pUNI/V5-His B (Invitrogen, Carlsbad, CA) to generate pHXB2 (3.7). Mutations were selectively introduced into integrase using the Quick change Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) to generate the mutagenized clones, pHXB2(3.7) mut. After confirmation of the mutation(s) by sequencing, DNAs were prepared from the pHXB2(3.7) mut clones and the ApaI - SalI fragment containing the integrase mutation(s) was isolated for subcloning back into HXB2 by replacement of the wild-type 3.7 Kb ApaI – SalI fragment.
IN strand-transfer assay
Expression and purification of recombinant IN containing an N-terminal 6His-tag has been described in the materials and methods section of an earlier study conducted at Gilead Sciences by our investigators . The strand transfer assay was described in detail in the experimental procedures section of an earlier study conducted at Gilead Sciences by our investigators . Briefly, biotinylated donor DNA was bound to streptavidin coated plates. Unbound donor DNA was removed by washing, and IN was then added to each well to allow 3’-processing of the donor DNA end for 30 min. Digoxigenintagged target DNA was next added to each well, and the 3’-end-joining reaction was allowed to proceed for 30 min. The wells were washed to remove un-joined target DNA, and the chemiluminescence signal was detected following addition of horseradish peroxidase-conjugated anti-digoxigenin antibody. Test compounds were added immediately before the target DNA.
Antiviral and cytotoxicity assays
Antiviral and cytotoxicity assays in HIV-1 IIIb infected MT-2, MT-4 and SupT1 cells and PHA plus IL-2 stimulated human PBMCs have been described in the materials and methods section of an earlier study by our investigators at Gilead Sciences . Briefly, for the antiviral assay utilizing MT-2 and MT-4 cells, 50 μL of 2X test concentration (total of 9 concentrations) of 3-fold serially diluted compound in culture medium with 10% FBS was added to each well of a 96-well plate in triplicate. MT-2 cells were infected with HIV-1 IIIb virus at a multiplicity of infection (MOI) of ~0.01 for 3 hours at 37°C. Fifty microliters of infected cell suspension in culture medium with 10% FBS (~1.5 x 104 cells) was then added to each well containing 50 μL of diluted compound. The plates were then incubated at 37°C for 5 days. At the end of the incubation, 100 μL of CellTiter-GloTM Reagent (catalog # G7571, Promega Biosciences, Inc., Madison, WI) was added to each well. Cell lysis was carried out by incubating the assay plates at room temperature for 10 min followed by the quantification of chemiluminescence signal using the Envision plate reader (Perkin Elmer, Shelton, CT). For compound cytotoxicity assessment in MT-2 and MT-4 cells, a protocol identical to that of the antiviral assay was followed, except that uninfected cells were used and a higher concentration range of compound was tested.
To study the effect of serum proteins on the antiviral activity of INSTI, compounds were tested in the presence of either 35 mg/ml HSA (catalog #A-1653, Sigma, St. Louis, MO), or 1.5 mg/ml α1-AGP (catalog #G-9885, Sigma, St. Louis, MO). To assess the effect of human serum on compound potency, the assay was performed in the presence of 10%, 20%, 35% and 50% human serum from clotted whole blood (catalog #14-498E, Cambrex Bio Science, Walkersville, MD).
2-LTR circle accumulation assay
The assay for 2-LTR circle accumulation was adapted from a previously described method . SupT1 cells were infected in bulk culture with HIV-1 IIIb at a multiplicity of infection (MOI) of 10 and at a cell density of 4 × 106 cells/ml for 3 hours at 37°C. We have confirmed in a time course experiment in our own system a previous observation that 2-LTR circles reached a peak production 24 hours post-infection . For this reason, measurement of 2-LTR circles was performed at 24 hours post-infection in our inhibition dose response studies. Both the 2-LTR junction (2-LTR circles) and the host globin gene (used to normalize for cell number) were quantified using Taqman real-time PCR . The copy number of 2-LTR circles was normalized to that of the globin gene and plotted against drug concentration for EC50 determination.
The Alu-PCR assay has been previously described . MT-2 cells were infected as a bulk culture with pseudotyped HIV-1 vector particles produced from the SODk1 2G cell line [26,27] at an m.o.i. of 10 and a cell density of 4 × 106 cells/ml for 3 hours at 37°C. The plates were incubated at 37°C for 12 hours for Late-RT product quantification and for 48 hours for Alu-PCR product quantification. The levels of Late-RT sequences and integration junctions quantified at 12 and 48 hours post-infection respectively were normalized to the level of globin gene present in the sample at these time points. Quantification by PCR was performed on the ABI Prism® 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA).
The entire curve fitting to generate EC50 and CC50 as well as error calculations were performed using Sigma Plot scientific data analysis and graphing software (Systat Software Inc., San Jose, CA).
Viral resistance selection
Resistance selections were carried out in 6-well tissue culture plates. MT-2 cells were seeded at a density of 0.5 × 106 cells per well in 5 ml of culture medium. Compounds were added at a final concentration corresponding to their antiviral EC50 or twice the EC50. HIV-1 IIIb was used at an MOI of ~ 0.01. The cultures were incubated at 37°C and split 1/2-1/3 once or twice a week depending on the growth status of the cells. Cytopathic effect (CPE) manifested as syncytia formation was used to follow progression of infections. Virus was harvested and transferred to fresh MT-2 cells in the presence of higher concentrations of compounds. Successive viral passages were obtained by repeating this procedure. The duration of each passage ranged from 10-15 days.
Population and clonal sequencing of HIV-1
Viral RNA isolated from culture supernatants was used for cDNA synthesis. Five μl of cDNA was used for PCR amplification of a 2765 base pair fragment spanning nucleotide 372 of the RT coding region to nucleotide 127 of the vpr gene and encompassing the entire IN open reading frame. The following primers were used for PCR: pLAIF3004: 5’–CAGGAAGTATACT GCATTTACC–3’ and pLAIR5768: 5’– CTAAGCCATGGAGCCAAATCC–3’ (Bio Source International, Foster City, CA). The PCR product was purified using the QIA quick PCR purification kit (Catalog #28104, QIAGEN, Valencia, CA). This PCR product was used for population sequencing of drugresistant viral pools and both DNA strands were sequenced (ELIM Biopharmaceuticals Inc, Hayward, CA). For clonal sequencing, a 1296 base pair PCR product spanning nucleotide 1608 of the RT coding region to nucleotide 412 of the vif gene was used as template. The PCR product was purified on a 1.0% agarose gel and cloned into the pCR-Blunt IITOPO plasmid using the Zero Blunt TOPO PCR Cloning Kit (Catalog #K2800-20, Invitrogen, Carlsbad, CA). Thirty transformed bacterial colonies were picked and cultured overnight in L-Broth containing 50 μg/ml kanamycin. Plasmid DNAs were isolated using a Wizard SV 96 Plasmid purification kit (Catalog #A2255, Promega, Madison WI) and sequenced through the IN region on both DNA strands (ELIM Biopharmaceuticals Inc, Hayward, CA) using the following primers: INTseq14157: 5’–ACCAGCACACAAAGGAATTGG–3’ and VIFseq2Rev5453: 5’–CCTGCTTGATATTCACACC–3’ (BioSource International, Foster City, CA).
Hepatic microsomal stability assay
Microsome (MS) suspensions were prepared by dilution of stock (in vitro Technologies, MD) with 0.05 M phosphate buffer, pH 7.4 to obtain protein concentrations of 0.5 mg/mL. Compounds were mixed with MS suspension media (pre-heated for 5 min at 37ºC) and cofactors were added to start the reaction. The final reaction mixture contained: 2 μM compound, 0.5 mg/mL microsomal protein, 1.25 mM NADPH, 3.3 mM MgCl2 and 0.05 M phosphate buffer, pH 7.4. At 0, 5, 15, 30, 45 and 60 min, 25 μL of the reaction mixture were quenched with 250 μL of internal standard (50 nM)/mobile phase B. The analytical internal standard was suspended in 0.2% formic acid/95% acetonitrile/5% water. After quenching, the plates were centrifuged at 3000 × g for 10 min. Five μL samples were analyzed by LC/MS/MS (LOQ = 0.001μM). Verapamil was used as a positive control for microsomal activity.
Caco-2 cells were grown to confluence on 24 well PET (polyethylene-terephthalate) plates (BD Biosciences, San Jose, CA). Donor HBSS buffer from Invitrogen was supplemented with 10 mM HEPES and 15 mM glucose with pH adjusted to pH 6.5. Receiver HBSS buffer was supplemented with 1% BSA with pH adjusted to pH 7.4. After equilibration with transport buffer, trans-epithelial electrical resistance was measured to confirm membrane integrity. The assay was initiated by the addition of compound containing buffer to the donor compartment. Samples of 100 μL were removed from the receiver compartment at 1 and 2 h post compound addition. The removed samples were immediately mixed with 400 μL of 0.1% formic acid in acetonitrile to precipitate protein and stabilize test compounds. Cells were dosed on either the apical or baso-lateral side to determine the forward (A to B) and reverse (B to A) permeability. Permeability through the membrane and non-specific binding was measured through a cell free trans-well. To test for non-specific binding and compound instability, the total amount of compound was quantified at the end of the experiment and compared to the material present in the original dosing solution to determine % recovery. Samples were analyzed by LC/MS/MS.
Plasma protein binding assay
Equilibrium dialysis was conducted at 37°C using Beagle dog and human plasma containing either GS-9224 or GS-9160 at the final concentrations of 2 μM. Based on a pilot time-to-equilibrium study and stability data, a 30-hour dialysis time was chosen for plasma protein binding determinations. Prior to the study, dialysis membranes were soaked for approximately one hour in 0.133 M phosphate buffer, pH 7.4. Plasma samples containing compound (1 mL) and compoundfree phosphate buffer (1 mL) were placed into opposite sides of the assembled dialysis cells. The dialysis cells were rotated slowly in a 37°C water bath for 30 h. Plasma samples were next drained into preweighed polypropylene tubes containing 1 mL of phosphate buffer. Buffer samples were drained into pre-weighed tubes containing 1.0 mL of blank plasma of the matching species. Post-dialysis plasma and buffer weights were measured and recorded for calculations.
Pharmacokinetic profiling in healthy human subjects
Ten healthy subjects were randomized to receive a single dose of GS-9224 (n=8) or matching placebo (n=2) following an overnight fast. Doses of 50 and 100 mg were administered in orange juice on day 1 and day 9 to the same patients (including a rinse of the dispensing container) to a total liquid volume of 240 mL with dosing within 15 minutes of study drug preparation. Following study drug administration, subjects were restricted from water intake until after the 2-hour blood draw and from food intake until after collection of the 4-hour blood draw. Serial blood samples were collected at times: 0 (predose), 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5, 4, 5, 6, 8, 10, 12, 18, 24, 28, 32, 36, 40, 44, and 48 hours postdose. In subjects that received GS-9224 plasma concentrations were measured at all-time points; protein binding of GS-9224 was assessed at the 1-hour time point.
Assessment of adverse events and concomitant medications continued throughout the duration of the study. Laboratory analysis, serial ECGs, and physical examinations were performed at screening, at defined intervals during the confinement periods, and at the followup evaluation 7 days after the last dose of study drug administration.
The study protocol and informed consent document(s) were reviewed and approved by the study center’s Institutional Review Board and subjects were provided written informed consent before study participation. The studies were carried out in accordance with the clinical research guidelines established by the Basic Principles defined in the U.S. 21 CFR Part 312.20 and the principles enunciated in the latest version of the Declaration of Helsinki.
GS-9224 exhibits potent antiviral activity and is synergistic with HIV antiviral drugs
In an effort to improve the pharmacokinetic profile of previously published INSTI GS-9160 [19-21], various substitutions on the p-fluorobenzyl moiety of GS-9160 were explored leading to the discovery of the close analog GS-9224 containing a 5-chloro-2,4- fluorobenzyl moiety (Figure 1) .
The antiviral activity of GS-9224 was evaluated in HIV-1 infected MT-2 and MT-4 cell lines as well as in primary human PBMCs (Table 1). The anti-HIV activity of GS-9224 was similar in all three cell types with an EC50 range of 1.3 to 9 nM. This potency is within a range similar to that observed for RAL and slightly lower than the parent analog GS-9160. To assess the impact of human serum protein binding on the antiviral potency of GS-9224 and to provide a target trough concentration in humans to achieve efficacy, the antiviral activity of GS-9224 was determined in the presence of human serum and the human serum components including human serum albumin (HSA) and α1-acid glycoprotein (α1-AGP) (Table 2). The EC50 of GS-9224 in 100% human serum was extrapolated from the EC50 values determined in the presence of 10, 20, 35 and 50% human serum. The combined effect of both serum proteins, HSA and α1-AGP increased the EC50 of GS-9224 by ~14.6-fold in MT-2 cells, while the EC50 values of GS- 9160 and RAL were increased by 10.8- and 3-fold, respectively. Whole human serum had an effect similar to the combined serum proteins in MT-2 cells and increased the EC50 of GS-9224 by 12.3-fold. In human PBMCs, the combined effect of HSA and a1-AGP increased the EC50 of GS-9224 by 8.6-fold, a similar shift in the EC50 values as that observed for GS-9160 and RAL (10.8- and 7.3-fold, respectively).
|Compound||Anti-HIV in MT2a||Cytotoxicity in MT2a||Anti-HIV in MT4a||Cytotoxicity in MT4a||Anti-HIVin PBMCb||Cytotoxicityin PBMCb|
|EC50 (nM)||CC50(nM)||EC50 (nM)||CC50 (nM)||EC50 (nM)||CC50 (nM)|
|GS-9224||1.3 ± 1.1||1600 ± 600||9.4 ± 0.5||700 ± 300||5 ± 6||5850 ± 900|
|GS-9160||1.2 ± 0.5||4200 ± 1000||4 ± 3||800 ± 400||1.2 ± 0.9||3880 ± 1700|
|RAL||7.5 ± 1||>100,000||8 ± 2||58200 ± 7300||2.6 ± 1.2||210,500 ± 90,000|
aMean ± SD of at least 4 independent experiments.
b Mean ± SD of at least 3 different donors
Table 1: Antiviral Activity of GS-9224 in Various HIV-1 Infected Human Cell Types.
|Compound||Anti-HIV in MT2 Cells||Anti-HIV in Human PBMCs|
|10% FBSa,b||10% FBS + HSA + α1-AGPb||100% HSd||10% FBSa,c||10% FBS + HSA + α1-AGPc|
|EC50 (nM)||EC50 (nM)||EC95 (nM)e||EC50 (nM)||EC95 (nM)e||EC50 (nM)||EC50 (nM)||EC95 (nM)e|
|GS-9224||1.3 ±1.1||19 ±5||118 ±28||16||100||2.2 ±0.1||19 ±6||121 ±39|
|GS-9160||1.2 ±0.5||13 ±6||76 ±35||12||75||1.2 ±0.9||13 ±13||80 ±79|
|RAL||7.5 ±1||23 ±5||162 ±35||–||–||2.6 ±1.2||19 ±14||121 ±93|
aThe EC50 for Control (10% FBS) and (10% FBS + HSA + a1-AGP ) were run in parallel.
b Mean ±SD of at least 3 independent experiments.
cMean ±SD of at least 3 donors.
d Mean of at least 2 independent experiments.
e The values of EC95 were calculated from EC50
Table 2: Effect of Serum Proteins on the Antiviral Activity of GS-9224.
Similar to GS-9160, GS-9224 showed synergistic antiviral effect in HIV-1-infected MT-2 cells when combined with approved drugs from different antiretroviral classes including PIs (lopinavir, atazanavir, and nelfinavir), NNRTIs (efavirenz) and N(t)RTIs (tenofovir, zidovudine, emtricitabine, and lamivudine) (Table S1).
GS-9224 inhibits HIV-1 integrase function and integration in infected cells
The inhibitory effect of GS-9224 on IN enzymatic activity was evaluated in a biochemical assay that measures the activity of IN in an in vitro integration reaction mediating the strand transfer of viral donor DNA unto host or target DNA. GS-9224 produced inhibitory activity (IC50 = 55 nM) that was comparable to that of GS-9160 (IC50 = 29 nM) and RAL (IC50 = 34 nM) (Table 3). An elevation in the levels of circular HIV DNA species containing 1- or 2-long terminal repeats (1-LTR and 2-LTR circles) following the treatment of infected cells with a small molecule can provide evidence that HIV integration has been inhibited . Treatment of HIV-1 infected SupT1 cells with GS-9224 caused a ~3.9-fold increase in 2-LTR circles. The effect was dose-dependent with an EC50 of 9.5 nM (Table 3). Similar results were obtained with the parental analog of GS-9224, GS-9160 and the clinically validated integrase strand-transfer inhibitor, RAL. This result provided initial evidence that GS-9160 can block HIV-1 integration in a cell.
|Compound||Strand Transfera||2LTR Circles in SupT1b||Cytotoxicity in SupT1c||Late RT in MT-2b||Alu-LTR in MT-2b||Cytotoxicity in MT-2c|
|IC50 (nM)||MaximumFold-Change||EC50 (nM)||CC50 (nM)||EC50 (nM)||EC50 (nM)||CC50 (nM)|
|GS-9224||55 ± 10||3.9 ↑||9.5||2600||>100||2.8||1600|
|GS-9160||29 ± 3||2.3 ↑||11||970||>100||1.8||4200|
|RAL||34 ± 5||2.4 ↑||12||>100,000||>200||3.6||>100,000|
aMean ± SE of at least 4 independent experiments using IN enzyme.
b Mean of at least 2 independent experiments in HIV-1 infected SupT1 or MT-2 cells.
cMean of at least 4 independent experiments in uninfected SupT1 or MT-2 cells.
Table 3: Inhibition of Integrase Enzymatic Strand Transfer Activity and Integration in HIV-1 Infected SupT1 and MT-2 Cells.
Another method to assess whether HIV-1 integration is inhibited in infected cells is by direct measurement of integration junctions between HIV proviral DNA and the host chromosomal DNA . Detection by PCR of nucleic acid products containing Alu-repeat sequences and portions of HIV-1 DNA represents a marker of successful HIV-1 integration. These products typically peak at 48 hours post-infection. In the presence of GS-9224, these products decreased in a dose-dependent manner with an EC50 of 2.8 nM, an activity comparable to that observed with GS-9160 and RAL in the same assay (Table 3). To ensure that this reduced integration was not attributable to an impairment of the upstream process of reverse transcription, accumulation of late RT products was assessed in the presence of GS-9224. GS-9224, like the other two strand-transfer inhibitors, GS-9160 and RAL did not affect the accumulation of late reverse transcription products which were in sharp contrast to the inhibitory effect noted with the NNRTI efavirenz (Table 3). This result provides unequivocal evidence that GS-9224 acts as an authentic inhibitor of integration in HIV-1 infected cells.
GS-9224 is active against drug-resistant mutants of HIV-1
The antiviral activity of GS-9224 was determined against a panel of drug-resistant mutants of HIV-1. The panel included mutants that were resistant to N(t)RTIs (K65R, M184V and TAMs), NNRTIs (K103N and Y181C) and PIs (I84V/L90M and G48V/V82A/L90M). GS-9224, like GS-9160 and RAL, retained activity against all NRTI-, NNRTI- and PI-resistant mutants tested (Table S2). The antiviral activity of GS-9224 against these drug-resistant mutants was comparable to its activity against the wild-type reference virus HIV-1 IIIb (Table 1).
Activity of GS-9224 against viruses with mutations conferring resistance to other INSTIs
Resistance mutations previously selected with other INSTIs [14,29-42] were introduced into an HIV-1 infectious molecular clone to determine if they confer cross-resistance to GS-9224 (Table 4). The T66I and L74M mutant viruses showed no resistance against GS- 9224, GS-9160 or RAL. E92Q displayed comparable resistance to GS- 9224 and GS-9160 (~20-fold) but lower resistance to RAL (~6-fold). Similarly, N155H conferred higher levels of resistance to GS-9224 and GS-9160 (~60–100-fold) compared to RAL (~24-fold). In contrast, Q148K appears to have a complementary resistance pattern to that of E92Q and N155H, displaying a higher level of resistance to RAL (>100- fold). E138K alone does not contribute any resistance to GS-9224 and the other INSTIs, but it greatly enhanced the resistance to all INSTIs tested when combined with Q148K. In summary, the resistance pattern of GS-9224 is similar to that of GS-9160, but somewhat distinct from that of RAL.
|Compoundsj||EC50 (nM)||Fold-Change in EC50 Relative to HIV-1 IIIb|
|HIV-1 IIIb||T66Ia||L74Mb||E92Qc||E92Vd||E138Ke||Q148Kf||V151Ag||N155Hh||N155Si||L74M/ E92V||E92V/ V151A||E138K/Q148K||E92V/G140S/V151A|
a Selected by L-708,906 (30, 32), S-1360 (31, 33) and EVG (37).
b Selected by GS-9224 (this report), L-708,906 (32), S-1360 (33), L-870,810 (42) and GS-9160 (19).
c Selected by EVG (31) and L-870,810 (42).
d Selected by GS-9160 (19).
e Selected by S-1360 (33).
f Selected by GS-9224 (this report), S-1360 (31) and RAL (14, 39).
g Selected by GS-278012, an analog of GS-9160 (19).
h Developed in SHIV (89.6P) infected Rhesus macaques treated with L-870,812 (34).
i Selected by S-1360 (31).
j The EC50 and fold-change represent the mean of a least two independent experiments performed in triplicate.
Table 4: Resistance Profile of Mutant Viruses Conferring Resistance to INSTIs.
in vitro resistance selection in the presence of GS-9224
Serial passaging of HIV-1 IIIb virus in MT-2 cells in the presence of increasing concentrations of GS-9224 (Figure 2) led to the emergence of a virus pool displaying a 4.6-fold resistance to GS-9224 by passage 4 and >100-fold resistance by passage 6 (Table 5). The viruses selected with GS-9224 displayed similar levels of cross-resistance to RAL both at passage P4 and P6, but remained fully sensitive to compounds from other antiretroviral classes.
Figure 2: Progress of resistance selection and and emergence of resistance mutations selected in vitro in the presence of GS-9224. The EC50 and CC50 values of GS- 9224 in MT-2 cells are indicated by the red and blue lines respectively. Each closed black circle represents one passage of the infected cell culture. Viral passages 3, 4, 5, 6 and 7 are denoted as P3, P4, P5, P6 and P7 and represent a transfer of cell-free virus supernatant to fresh uninfected cell culture with a doubling of the drug concentration. The culture supernatant from the last infected cell culture before the increase of inhibitor concentration (red circle) was used for population sequencing and the mutations identified in the integrase region are indicated. In addition, P4 and P6 viral pools were used for clonal sequencing to determine the percentage of clones containing each identified mutation. P4 virus: L74M (62%), G140S (62%); P6 virus: L74M (100%), G140S (100%), Q148K (93.3%).
|Compound||EC50 , nM (fold change relative to HIV-1 IIIb|
|HIV-1 IIIba||GS-9224 P4 Virusb||GS-9224 P6 Virusb|
|TFV||3.8 ± 0.8||3 ± 0.4 (0.9)||5.9 ± 0.9 (1.6)|
|EFV||0.7 ± 0.1||1.3 ± 0.2 (1.9)||1.5 ± 0.3 (2.2)|
|LPV||13.0 ± 4.0||27.0 ± 6.0 (2.1)||20.9 ± 0.8 (1.6)|
|RAL||3.5 ± 0.6||11.0 ± 1.2 (3.1)||237.0 ± 126.0 (>100)|
|GS-9224||4.0 ± 2.0||18.6 ± 2.7 (4.6)||>1000 (>100)|
a Mean ± SD of at least 6 independent experiments.
b Mean ± SD of at least 3 independent experiments.
The numbers in parentheses represent the fold-change over wild-type virus.
Table 5: Resistance Phenotype of GS-9224 Selected Viral Pools.
GS-9224 Selected a different pattern of INSTI resistance mutations than GS-9160
Population sequencing of GS-9224-selected viral passages revealed the successive emergence of mutations G140S, L74M and Q148K in the catalytic core domain of HIV-1 IN (Figure 2). Clonal sequencing of passage 4 revealed that G140S and L74M are each present in 62% of the clones, suggesting that at least 24% of the clones carried both mutations. Both G140S and L74M were present in 100% of the clones sequenced at passage 6 while Q148K, a third mutation which appeared in passage 6 was present in 93.3% of the clones. The S24G mutation which was detected by population sequencing of passage 5 virus is a known natural polymorphism of IN that is not associated with resistance to INSTIs . In fact, by passage 6, this mutation was no longer detectable. Since the level of resistance increased from 4.6-fold to >100-fold from passages 4 to 6 with the appearance of the additional mutation Q148K, we wished to determine whether combining IN mutations L74M, G140S and Q148K in a molecular clone of HIV- 1 would recapitulate resistance to GS-9224. These mutations were introduced either individually or together into an infectious molecular clone of HIV-1. Interestingly, as individual mutations, Q148K conferred 42-fold resistance to GS-9224, L74M alone had no effect and G140S alone conferred 6-fold resistance (Table 6). When L74M and G140S were combined, the double mutant conferred 11-fold resistance, suggesting that L74M may potentiate resistance conferred by G140S, a property of L74M that we previously reported with regard to the E92V mutation . Notably, the triple mutant L74M/G140S/Q148K conferred >100-fold resistance to GS-9224, a level consistent with that observed for passage 6 virus (Tables 5 and 6). All combinations of the tested mutations displayed comparable resistance levels to GS- 9160 and GS-9224, but different levels of resistance against RAL. In particular, L74M did not potentiate G140S resistance against RAL. In addition, while the triple mutant conferred a higher level of resistance than Q148K alone against GS-9224 and GS-9160, it conferred a lower level of resistance than Q148K alone against RAL.
|Compoundsd||EC50 (nM)||Fold-Change in EC50 Relative to HIV-1 IIIb|
a Selected by GS-9224 (this report), L-708,906 (32), S-1360 (33), L-870,810 (42) and GS-9160 (19).
b Selected by GS-9224 (this report) and L-Chicoric acid (29).
c Selected by GS-9224 (this report), S-1360 (31) and RAL (14, 39).
d The EC50 and fold-change represent the mean of a least two independent experiments performed in triplicate.
Table 6: Resistance Profile of Mutant Viruses Selected by GS-9224
GS-9224 exhibits improved in vitro ADME properties compared to GS-9160
GS-9224 displayed a longer half-life than GS-9160 in human and dog hepatic microsomal extracts (Table 7). In addition, GS-9224 did not demonstrate any significant inhibition or induction of human CYP450 isoforms within clinically achievable range of concentrations (data not shown). When permeability was measured across a monolayer of Caco- 2 cells, GS-9224 displayed higher permeability than GS-9160 in both apical to baso-lateral and baso-lateral to apical directions with a lower efflux ratio than GS-9160 or RAL (Table 7). Finally, the assessment of plasma protein binding by equilibrium dialysis against dog and human plasma showed that GS-9224 is more highly protein bound than GS- 9160 both in dog plasma (~99.5% vs. ~93.5%) and in human plasma (~97.7% vs. ~89.6%) (Table 7).
|Compound||Liver Microsomal StabilityaT1/2 (min)||Cellular Permeability in Caco-2 Cellsb||Plasma ProteinBindingc (% Bound at 2 µM)|
|Dog||Human||Papp A→B (cm/sec × 10-6)||Papp A→B (cm/sec × 10-6)||EffluxRatio||Dog||Human|
aMean of 13 independent experiments.
b Mean of 2 independent experiments.
cMean of 3 determinations using equilibrium dialysis.
Table 7: In Vitro ADME Profiling of GS-9224 Relative to GS-9160 and RAL.
Similar to GS-9160, GS-9224 is not metabolized by CYP450 enzymes in microsomes or hepatocytes but undergoes significant metabolic modification via glucuronidation. A study was carried out to determine the specific uridine diphosphate transferase (UGT) isoforms responsible for the metabolism of GS-9224. The results showed that UGT 1A1 and 1A3 are the main UGT isoforms responsible for glucuronidation of both GS-9160 and GS-9224 (Table 8). The characterization of GS-9224 metabolites formed systemically both in vitro and in vivo suggests the formation of a single phenolic glucuronide metabolite of GS-9224 (Figure 3).
|UGT Isoform||GucuronidationRatea,b (pmol/min/mg)|
a Mean of duplicate test.
bThe Limit of rate detection is 0.02 pmol/min/mg
Table 8: Rate of Glucuronidation of GS-9224 by Major UGT Isoforms.
Single-dose pharmacokinetics in healthy human subjects
After administration of single doses of GS-9224, 50 or 100 mg as an oral solution in the fasted state, maximum mean (CV%) concentrations of 110 ng/mL (51%) and 215 ng/mL (41%) were achieved at the 0.75 hour nominal time point, thereafter falling to <1.0 ng/mL in 8/8 (100%) and 7/8 (87.5%) subjects by the 12 hour time point for the 50 mg and 100 mg dose group, respectively. These data indicated rapid plasma elimination of GS-9224, consistent with a ~1.0 hour terminal elimination half-life that was incompatible with a goal of once-daily dosing. The free fraction (%fp) of GS-9224 in plasma samples collected after administration of 50 mg and 100 mg GS-9224 was 7.66 ± 1.25% and 9.44 ± 3.86% (mean ± SD), respectively.
Single doses of GS-9224 50 mg and 100 mg were safe and well tolerated in healthy, adult, male and female subjects. No discontinuations due to adverse events or serious adverse events occurred. All adverse events were mild with the exception of moderate diarrhea, vomiting, and headache in 1 subject. Headache, upper abdominal pain, diarrhea, and stomach discomfort (1 report of each) were each considered by the investigator to be related to study treatment.
GS-9224 is an analog of GS-9160, a previously reported investigational INSTI . GS-9224 was designed in an effort to optimize the pharmacokinetic profile of GS-9160 while retaining its antiviral potency. The medicinal chemistry optimization focused primarily on the p-fluorobenzyl moiety of GS-9160. Profiling of various analogs led to the identification of GS-9224 that differs structurally from GS-9160 by the addition of one chlorine atom and one fluorine atom on the p-fluorobenzyl moiety (Figure 1) .
Similar to GS-9160, GS-9224 exhibits potent and selective antiviral activity against HIV-1 and acts as an efficacious inhibitor of the IN strand transfer activity in an in vitro biochemical enzymatic assay. The elevation of 2-LTR circles in HIV-infected cells treated with GS- 9224 was indicative of the inhibition of proviral DNA integration. In addition, GS-9224 profoundly reduced the number of integration events in treated cells as measured by Alu-LTR quantification of integration junctions without alteration of total viral DNA synthesis measured by the late-RT signal, demonstrating that GS-9224 is an authentic inhibitor of HIV-1 integration
GS-9224 showed synergistic antiviral effect when combined with approved HIV drugs from different antiretroviral classes and retained antiviral potency against NRTI-, NNRTI- and PI-resistant HIV-1 mutants, an observation consistent with its established mechanism of action. Characterizing the activity against a series of viruses expressing various INSTI resistance mutations revealed a similar resistance profile for GS-9224 and GS-9160. In comparison with RAL, both compounds showed lower degree of resistance against clinically relevant mutation Q148K, but higher level of resistance against E92Q and N155H mutants observed in patients treated with either RAL or EVG [44,45].
in vitro passaging of HIV-1 in the presence of gradually increasing concentrations of GS-9224 resulted in the selection of an IN mutant variant with >100-fold resistance to the selecting inhibitor and a similar cross-resistance to other INSTIs including RAL and the parent inhibitor GS-9160. Clonal sequencing of the resistant variant revealed the emergence of G140S and L74M mutations followed by the Q148K mutation. It is interesting to note that Q148K emerged as a primary mutation in patients treated with RAL who experienced virologic failure [3,4], but in the selection with GS-9224, Q148K emerged only after L74M and G140S. Thus, it appears that Q148K is a secondary mutation that augments resistance of L74M/G140S against GS-9224. In contrast, Q148K displays a higher degree of resistance to RAL as a primary single mutation compared to the triple mutant L74M/G140S/ Q148K selected by GS-9224.
The additional chlorine and fluorine atoms on the fluorobenzyl moiety of GS-9160 conferred to GS-9224 a moderately higher hepatic microsome stability, better absorption potential, lower efflux ratio and higher serum protein binding compared to GS-9160 (Table 7).
In healthy human subjects, GS-9224 rapidly achieved maximum plasma concentrations within one hour of administration as a solution in the fasted state. Thereafter, concentrations rapidly declined with a terminal elimination half-life of ~1.0 hour. These human pharmacokinetic results were inconsistent with the target pharmacokinetic profile to provide sustained, therapeutic concentrations over a 24 hour dosing interval. As such, from a clinical pharmacology perspective, GS-9224 was deemed not suitable for QD dosing. In these clinically derived samples, the mean free fraction of GS-9224 was 7.5 to 10%, indicating lower in vivo protein binding than that estimated from in vitro plasma equilibrium dialysis for both dogs and humans. This could, at least in part, contribute to the fast systemic elimination of GS-9224 in humans. Single doses of GS-9224 at 50 mg and 100 mg were safe and well tolerated in healthy, adult, male and female subjects
In conclusion, GS-9224 has retained the antiviral potency, mechanism of action and resistance profile of the predecessor compound, GS-9160. In addition, it exhibited improved in vitro ADME profile including higher metabolic stability, absorption potential and serum protein binding. However, these differential properties did not translate into an improved pharmacokinetic profile in healthy human volunteers. Results from the assessment of the singledose pharmacokinetics revealed a profile of GS-9224 that would not be compatible with the once-daily oral dosing of GS-9224 due to faster than expected systemic elimination of the drug.
The authors thank Richard Pastor, Sammy Metobo, Rachael Lansdown and Salman Jabri for synthesizing.