CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Conclusion: We have developed novel lentiviral vectors that express Vif-resistant A3G(D128K), which potently inhibited HIV-1 replication in cell culture. Our results suggest that it may be very difficult for HIV-1 to develop resistance to A3G(D128K) while retaining activity against endogenous wild-type A3 proteins. A3G(D128K) gene therapy could provide a novel strategy for treatment and functional cure of HIV-1 infection 335 ENGINEERING HIV-RESISTANT ANTI-HIV CHIMERIC ANTIGEN RECEPTOR (CAR) T CELLS Malika Hale 1 , Taylor Mesojednik 1 , Guillermo Romano Ibarra 1 , Jaya Sahni 1 , Bryan Sands 1 , Karen Sommer 1 , Andrew M. Scharenberg 2 , David J. Rawlings 2 , Thor Wagner 2 1 Seattle Children’s Rsr Inst, Seattle, WA, USA, 2 Univ of Washington, Seattle, WA, USA Background: Advances in both gene editing and chimeric antigen receptor (CAR) technology have created new therapeutic possibilities for a variety of diseases. Broadly neutralizing monoclonal antibodies (bNAb) with specificity for different epitopes on the HIV envelope provide an opportunity to create multiple different CARs that target HIV- expressing cells. In combination with genetic disruption of CCR5 it may be possible to engineer long-lived immunity targeting the HIV reservoir. Methods: CAR were designed to target four well-defined neutralization epitopes on HIV Env: the V1/V2 epitope associated with the N160 glycan, the CD4 binding site epitope, the high-mannose-patch on V3 associated with the N332 glycan, and the membrane-proximal external region on gp41; based on bNAbs PGT145, VRC07-523, PGT128, and 10E8, respectively. Single chain variable fragments (scFv) were synthesized and cloned into a 4-1BB containing CAR construct. Primary human T cells were engineered to express the anti-HIV CARs, and tested for cell activation (cell surface expression of CD137) and specific killing of HIV-infected cells (ACH-2) in the presence of ART. CCR5-disrupted CAR T cells were generated with a CCR5-specific megaTAL nuclease to induce non-homologous end-joining recombination (NHEJ) in CCR5 or via homology-directed recombination (HDR), in which the CAR construct with flanking regions homologous to CCR5 was inserted into the CCR5 locus. CCR5-disrupted CAR T cells were tested in viral culture with a R5 virus (JR-CSF) compared to CAR T cells without CCR5 disruption. Results: Primary T cells expressing the CAR constructs were activated in the presence of HIV-infected cells, but not in the presence of HIV-uninfected cells, and specifically killed HIV-infected cells versus uninfected cells. Results with cells from three donors were statistically significant compared to cells that expressed an anti-CD19 CAR. All anti-HIV CAR T cells suppressed viral replication in culture, but after several days of viral culture CCR5-disrupted CAR T cells statistically improved viral suppression compared to CAR T cells without CCR5-disruption. Conclusion: This work demonstrates that HIV immunotherapy utilizing potent bNAb-based scFv fused to second-generation CAR signaling domains, in combination with CCR5- disruption, is feasible and effective in vitro. This strategy has the potential to induce long-term killing of HIV-infected cells in HIV-infected individuals, which might help in the effort to cure HIV. 336 VIRAL OUTCOMES AFTER ANALYTIC TREATMENT INTERRUPTIONS TO EVALUATE A FUNCTIONAL CURE Lorna Leal 1 , Constanza Lucero 1 , Montserrat Plana 1 , Nuria Climent 1 , Esteban Martinez 1 , Pedro Castro 1 , Beatriz Mothe 2 , Juan Carlos López-Bernaldo 3 , Jose M. Gatell 1 , Felipe Garcia 1 1 IDIBAPS, Barcelona, Spain, 2 IrsiCaixa Inst for AIDS Rsr, Badalona, Spain, 3 Hosp Gregorio Marañón, Madrid, Spain Background: Allowing participants to interrupt their antiretroviral treatment is still crucial to assess the efficacy of any investigational cure strategy in controlling HIV replication, but different virologic outcome vary in different analytic treatment interruptions (ATI) trials hindering the comparison between studies. Monitored Antiretroviral Pause (MAP) and restarting antiretroviral treatment (ART) at the moment of viral rebound have been proposed as safer strategies to evaluate the efficacy of interventions aimed to reduce the HIV reservoir. Methods: Here we compiled retrospective data from 9 ATI studies in chronic HIV-positive cART-suppressed individuals (n=355) and compared virologic outcomes used. The following outcomes were compared: time to viral rebound, viral load (VL) set point (defined as the mean VL between weeks 8 and 12, if the difference was < 0.5 log10) and VL at a predefined time following ATI (w1, w2, w3, w4, w5, w6, w8, w10, w12, w24 and w48) (in both cases absolute value and Δ with pre-ART), time to reach set point, time to reach a certain threshold, peak, time to peak and area under the curve (AUC). We also performed a sensitivity analysis in which those who discontinued the ATI before week 12 were assigned a set point VL equal to their last known VL. Results: Time to viral rebound was strongly and inversely correlated with VL set-point (absolute and Δ, and in sensitive analysis), VL at all time points after ATI, VL peak and AUC (p<0.0001 for all comparisons). Only 3.5% of individuals in whom VL rebounded after ATI presented a sustained VL drop to <1,000 copies/ml thereafter (3 cases to undetectable level, all with peak viremia <10,000 copies/ml). VL remained undetectable at set-point, w24 and w48 after ATI in a 3.3%, 1.9% and 0% of subjects. A significant correlation was observed between pre-ART VL and ATI VL set point, w24, w48 and peak viremia (p<0.0001). ATI VL remained significantly lower than pre-ART VL in all time-points [mean (SD) Δ set-point was -0.36 (0.88), p<0.001], except at peak viremia. 10% and 25% of patients had a Δ set-point > 1log and > 0.5 log, respectively. Conclusion: No individual presenting a viremia above 10,000 copies/ml during ATI was able to control VL thereafter. Time to VL rebound correlates with most of virologic outcomes used in previous ATI studies, supporting its use as a valid virologic outcome in shorter, monitored antiretroviral pause for HIV cure studies. 337 A BISPECIFIC APPROACH FOR TARGETING IMRS IN HIV-1 LATENCY Debbie S. Ruelas 1 , Yanyan Zheng 1 , Fahimeh Raoufi 1 , Laurence Fayadat-Dilman 1 , Yuan Li 2 , Julie Strizki 2 , Richard Barnard 2 , Mohammad Tabrizifard 1 , Daniel Gorman 1 1 Merck & Co, Palo Alto, CA, USA, 2 Merck & Co, Inc, West Point, PA, USA Background: In cART suppressed individuals, antigen exposure, microbial translocation and increasing immune activation lead to persistent T cell dysfunction. A select group of immune modulatory receptors (IMRs) are upregulated on the surface of exhausted CD8s, and on CD4 T cells that harbor latent HIV. Blockade of IMR pathways has been shown to reverse immune cell dysfunction and lower the threshold for reactivation of latently infected cells in vitro. We hypothesize that simultaneous engagement of multiple IMR targets, via bispecific antibodies, will enhance anti-HIV efficacy compared to single or combination antibodies. Methods: We examined the fate, and binding orientation of bispecific antibodies that target certain IMRs (PD-1, Lag3, and TIGIT) known to play a role in HIV infection. Cells that separately express specific IMRs were generated and used to measure the ability of bispecific antibodies to simultaneously bind two separate cells (trans-binding orientation). Conversely, we used an Amnis-based approach to measure the ability of bispecific antibodies to simultaneously bind two separate epitopes on the same cell (cis-binding orientation). Furthermore, we examined whether bispecific antibodies are capable of lowering the reactivation threshold of latently infected CD4s in a primary model of latency. Finally, we characterized IMR expression of HIV-specific CD8 T cells from infected individuals to examine whether bispecific antibodies are capable of reversing dysfunction. Results: We found that bispecific antibodies are capable of binding in both cis- and trans-orientation. Bivalent and one-armed antibodies completely lacked the ability to bind two separate molecules on distinct cells whereas bispecific molecules were capable of binding in a trans-orientation with a frequency of 25% in some cases. We observed that exogenous PD-L1 is capable of blocking reactivation in a primary CD4 T cell model of latency. And, found that the IMR, Lag-3, is increased on the CD8 T cells of HIV-infected individuals compared to uninfected donors. Conclusion: Binding orientation influences an antibody’s functional potential. Our data demonstrates that bispecific anti-IMR antibodies are capable of both cis- and trans- binding. Cis-binding may enable efficient internalization of an antibody-receptor complex, thus increasing therapeutic action. These studies may inform the future design of bispecific antibodies and provide rationale for an IMR-targeting bispecific antibody approach to treat HIV infection.

Poster and Themed Discussion Abstracts

CROI 2017 132