Background:  Gene therapy with long antisense RNA effectively blocks HIV replication in vitro. The lymphoid trafficking, antiviral effects and safety of multiple infusions of lentiviral vector transduced cells expressing env antisense are unknown.

Methods:  In an ongoing phase I/II trial, HIV-1-infected subjects fully suppressed on ART received over 14 weeks until 6 infusions each of 10¹⁰ autologous CD4 T cells transduced ex vivo with a lentiviral vector encoding a 937nt env antisense construct (VRX496). At week 18, eligible subjects underwent scheduled treatment interruption to evaluate effects on viral load and CD4 count during viremia. Rectal biopsies were obtained at baseline, after the third and sixth infusions, and 6 weeks post-scheduled treatment interruption. Intraepithelial lymphocytes were isolated and CD4 cells quantitated by flow cytometry. Vector copies per CD4 cell were measured in Intraepithelial lymphocytes and blood. Average viral load at 10 and 14 weeks post scheduled treatment interruption was compared to the subject’s pre treatment set point when available. Patients were monitored for vector-derived replication competent lentivirus and immunogenicity, and T cell receptor (TCR) diversity.

Results:  Of the total, 12 subjects completed infusions and 7 underwent scheduled treatment interruption. Vector was detected in Intraepithelial lymphocytes from 8 of 10 patients. VRX496-transduced T cells persisted following scheduled treatment interruption in 4 and 6 of 7 evaluable subjects in gut and blood, respectively. Vector copy number per CD4 T cell in Intraepithelial lymphocytes and blood were strongly correlated (p = 0.0004). CD4 cells increased concomitant with dosing in the blood but not the gut. Following scheduled treatment interruption, all subjects had rebound viremia and decreased blood CD4 counts, whereas mucosal CD4 cells were lower than blood at baseline and did not decrease following scheduled treatment interruption. Viral load set points decreased by 0.27 to 1.03 log in 3 patients; 4 patients had low nadir CD4 and discontinued scheduled treatment interruption prior to establishing set points. Half of patients developed antibodies against vesicular stomatitis virus-G (VSV-G), which did not correlate with adverse events or loss of vector modified cells. No TCR oligoclonality or replication competent lentivirus was detected, and there were no safety concerns associated with multiple infusions.

Conclusions:  Gene-modified CD4 T cells ex vivo traffic to gut lymphoid tissue in most patients, persist following treatment interruption, and are correlated with persistence in blood. Multiple infusions of lentiviral vector modified cells increase CD4 counts prior to scheduled treatment interruption and may decrease viral load set-point post-scheduled treatment interruption. The protocol has been amended and repowered to further evaluate this aspect.