Long-acting capsid inhibitor protects macaques from repeat SHIV challenges

Abstract

Because no currently available vaccine can prevent HIV infection, pre-exposure prophylaxis (PrEP) with antiretrovirals (ARVs) is an important tool for combating the HIV pandemic^1,2. Long-acting ARVs promise to build on the success of current PrEP strategies, which must be taken daily, by reducing the frequency of administration³. GS-CA1 is a small-molecule HIV capsid inhibitor with picomolar antiviral potency against a broad array of HIV strains, including variants resistant to existing ARVs, and has shown long-acting therapeutic potential in a mouse model of HIV infection⁴. Here we show that a single subcutaneous administration of GS-CA1 provides long-term protection against repeated rectal simian-human immunodeficiency virus (SHIV) challenges in rhesus macaques. Whereas all control animals became infected after 15 weekly challenges, a single 300 mg kg^-1 dose of GS-CA1 provided per-exposure infection risk reduction of 97% for 24 weeks. Pharmacokinetic analysis showed a correlation between GS-CA1 plasma concentration and protection from SHIV challenges. GS-CA1 levels greater than twice the rhesus plasma protein-adjusted 95% effective concentration conferred 100% protection in this model. These proof-of-concept data support the development of capsid inhibitors as a novel long-acting PrEP strategy in humans.

Fig. 1. GS-CA1 exhibits potent antiviral activity in vitro and long-acting pharmacokinetics in rhesus macaques.

a, Representative antiviral dose–response curve for GS-CA1 in rhesus PBMCs acutely infected with SHIV-SF162P3. Data are shown as mean ± s.d. from one of seven assays (n = 3 biological replicates each). b, Plasma GS-CA1 levels measured by mass spectrometry following a single subcutaneous administration of GS-CA1 dosed at 300 mg kg⁻¹ in three male rhesus macaques and at 100 mg kg⁻¹ in two male rhesus macaques. The bottom dotted line represents the assay limit of detection (LOD; 1 nM). The top dashed lines represent one and six times the rhesus paEC₉₅ for GS-CA1 (30.2 nM and 181.2 nM, respectively). Source data

Fig. 2. Single-dose GS-CA1 confers long-term protection from repeated rectal SHIV challenge in rhesus macaques.

a, Study design. Rhesus macaques of Indian origin were treated with a single subcutaneous administration of vehicle control or GS-CA1 at week 0 followed by weekly intrarectal challenges with SHIV-SF162P3 in a dose-escalation scheme until all control animals were confirmed to be infected on week 15. b, Plasma GS-CA1 levels measured by mass spectrometry over time in rhesus macaques dosed once at 150 mg kg⁻¹ or 300 mg kg⁻¹. The data for each group are shown as mean ± s.d. from eight rhesus macaques. The bottom dotted line represents the assay LOD (1 nM). The top dashed lines represent one and six times the rhesus paEC₉₅ for GS-CA1 (30.2 nM and 181.2 nM, respectively). c, Kaplan–Meier plots showing the development of viraemia as assessed by RT–qPCR for plasma SHIV gag among rhesus macaques treated with a single subcutaneous administration of vehicle control (n = 8), GS-CA1 dosed at 150 mg kg⁻¹ (n = 8) or GS-CA1 dosed at 300 mg kg⁻¹ (n = 8). Source data

Fig. 3. Association between plasma GS-CA1 and protection from SHIV challenge.

a, Plasma levels of GS-CA1 measured by mass spectrometry among individual rhesus macaques (n = 8) treated with a 150 mg kg⁻¹ dose over time. Black circles represent time points before the first detection of viraemia. Red circles represent time points including and after the first detection of viraemia. Grey shaded boxes represent serum Env ELISA seroconversion time points. The bottom dotted line represents the assay LOD (1 nM). The bottom and top dashed lines represent one and two times the rhesus paEC₉₅, respectively. b, Same data as in a for rhesus macaques treated with GS-CA1 at a dose of 300 mg kg⁻¹. No signal above the assay LOD was observed among the eight placebo-treated control animals throughout the study. Source data

Extended Data Fig. 1. Viral loads among vehicle- and GS-CA-1-treated animals during the SHIV challenge study.

a, Plasma SHIV viral loads measured by gag RT-qPCR among vehicle control rhesus macaques (n = 8) through week 24 following 15 weekly intrarectal challenges. b, Same data as (a) among rhesus macaques (n = 8) treated with a single 150 mg/kg subcutaneous injection of GS-CA1 at week 0. c, Same data as (a) and (b) among rhesus macaques (n = 8) treated with a single 300 mg/kg subcutaneous injection of GS-CA1 at week 0. Dotted lines represent the assay LOD (200 copies per ml).

Extended Data Fig. 2. Peak and setpoint SHIV viral loads among infected rhesus macaques treated with vehicle control or GS-CA1.

a, Peak plasma SHIV viral loads measured by gag RT-qPCR among infected rhesus macaques treated with vehicle control (n = 8), GS-CA1 dosed at 150 mg/kg (n = 6), or GS-CA1 dosed at 300 mg/kg (n = 3). b, Plasma SHIV viral loads 7 weeks post-infection (wpi) among the same groups as (a). Dotted lines represent the assay LOD (200 copies per ml). Solid lines represent group medians. P values reflect Kruskal-Wallis tests.

Extended Data Fig. 3. GS-CA1-treated rhesus macaques that remain aviremic do not show immunologic evidence of SHIV infection.

a, Week 24 anti-gp140 Env ELISA comparing viremic and aviremic animals in the vehicle control and GS-CA1 groups. Dotted line represents assay LOD (1:25 endpoint titer). b, Week 19 Gag ELISPOT comparing viremic and aviremic animals in the vehicle control and GS-CA1 groups. Data are represented as spot-forming cells (SFC) per 10⁶ PBMCs. Dotted line represents the assay LOD (5 SFC per million PBMCs).

Extended data Fig. 4. Sequence analysis of study-emergent CA mutations.

a, Longitudinal CA sequence analysis of select plasma samples from viremic study animals. Samples with population-scale wild-type and V11A CA sequences are shaded grey and blue, respectively, whereas yellow boxes denote sequence failures. b, Alignment of HIV-1 (NL4.3 strain) and SHIV-SF162P3 CA amino-acid sequences. Shaded regions represent invariant residues, hashed lines denote sequence gaps, and red arrows highlight HIV-1 residues associated with GS-CA1 resistance (L56, N57, M66, Q67, K70, N74, T107). c, A single CA monomer (rendered as ribbons in cyan) within the HIV-1 CA hexamer-LEN crystal structure, with notation of the CA N-terminal and C-terminal domains (CA_NTD and CA_CTD, respectively) and a GS-CA1 molecule (carbon atoms shown as green) docked at the LEN binding site. Location of the only CA variant (V11A) transiently detected by bulk sequencing is shown relative to all HIV-1 CA residues associated with bona fide in vitro GS-CA1 and LEN resistance (highlighted as sticks).