CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

534

CXCR4-Targeted CRISPR/Cas9 Lipid Nanoparticles for Excision of Latent Proviral DNA Sudipta Panja, Lubaba A. Zaman, Chen Zhang, Prasanta K. Dash, Milankumar Patel, Howard Gendelman University of Nebraska Medical Center, Omaha, NE, USA Background: A major challenge in curing HIV is eliminating latent integrated proviral DNA from CD4+ T and myeloid cells. Recent clinical studies (EBT-101) using adeno-associated viral (AAV) vectors to deliver CRISPR for HIV eradication remained unsuccessful likely due to the limitations of AAV vectors, highlighting the need for alternative delivery systems like lipid nanoparticles (LNPs). To this end, we created a lymphoid tissue-targeted LNP that includes a cyclic pentapeptide to selectively bind to CXCR4, a co-receptor that HIV uses to infect CD4+ T cells, aiming to eliminate HIV from latently infected tissue/cells reservoir more effectively. Methods: CXCR4-targeting peptide decorated LNP (T-LNP) was formulated by using microfluidic mixing. The receptor specificity was shown by CXCR4 inhibition assay. IVIS was performed to demonstrate lymphoid tissue targeting on humanized mice (hu-mice). The intact proviral DNA assay (IPDA) assay was utilized to test CRISPR gRNA delivery and viral elimination efficiency in latently HIV-infected lymphocytic cell lines (JLat 8.4, Jurkat E6), human peripheral blood lymphocytes (PBLs), and lymphoid tissue from HIV-infected, ART-treated hu-mice. Results: T-LNP and a control LNP without CXCR4 targeting (C-LNP) were tested in primary CD4+ T cells, cell lines, and hu-mice. Unlike C-LNP, T-LNP in humanized mice showed a significant shift in tissue targeting from the liver to the spleen, a primary HIV reservoir. T-LNP showed significantly higher viral excision efficiency in latently infected cells whose specificity was confirmed by a receptor inhibition assay through increasing concentrations of a CXCR4 inhibitor. T-LNP achieved nearly 80% viral DNA excision in HIV-1-infected PBLs, compared to ~7% with C-LNP. In HIV-1-infected, ART-treated humanized mice, T-LNP showed ~60% CRISPR efficacy in splenic tissue and blood, while C-LNP showed <5% efficacy, analyzed by the IPDA. Conclusions: T-LNP demonstrated lymphoid cell-targeted CRISPR/Cas9 delivery. This delivery system was effective in eliminating HIV-1 from infected CD4+ T cells. Furthermore, T-LNPs have shown superior excision efficiency compared to C-LNP in eliminating HIV-1 DNA from hu-mice infectious reservoirs. These data suggest that T-LNP has the potential to achieve a functional HIV-1 cure. The figure, table, or graphic for this abstract has been removed. Efficient Generation of Anti-HIV TRAC-CAR T Cells to Enable High Throughput Binder Screening Gabriella Kimmerly, Joseph Muldoon, Justin Eyquem, Rachel Rutishauser University of California San Francisco, San Francisco, CA, USA Background: Following recent successes using chimeric antigen receptor (CAR) T cells for various cancer and autoimmune indications, there has been renewed interest in using CAR T cells to target the reservoir of HIV-infected cells. We have previously shown that CRISPR/Cas9 targeting of CAR constructs to the TCRa constant ( TRAC ) locus enables physiologic expression of the receptor, promotes CAR T cell persistence, and enables pooled screening of hundreds of different CAR binders. Here, we demonstrate the generation of functional TRAC -targeted anti-HIV CAR T cells with binders derived from either a broadly neutralizing antibody (bNAb; PGT128) or from CD4. Methods: To produce TRAC -CAR T cells, CD3+ T cells isolated from human peripheral blood mononuclear cells were activated and electroporated with Cas9 ribonucleoprotein (RNP) targeting the first exon of the TRAC locus, followed by AAV delivery of a homology directed repair template (HDRt) encoding the CAR transgene. Edited T cells were assessed using flow cytometry to confirm both knock-out of the endogenous TCR (and CCR5 when targeted) and the knock-in rate via expression of a transduction marker (CXCR5), compared to mock- edited T cells receiving only the RNP. The in vitro killing capacity of the CAR T cells was assessed against an HIV-Env + K562 cell line expressing the mKate fluorophore, at various effector to target (E:T) ratios over 80 hours, using an Incucyte live-cell imager to track the proliferation of the fluorescent target cells. Results: We obtained high TRAC -CAR knock-in efficiency (77.4-84.3% receptor knock-in, 2 donors tested) as well as a high rate of knock-out of CCR5 , simultaneously with receptor editing (range of dual knock-out/knock-in

64.1%-74.1%). While both the PGT128 and CD4-derived binder CAR T cells demonstrated in vitro the capacity to kill HIV-Env + target cells compared to mock-edited T cells ( TRAC -ko) over multiple E:T ratios, the PGT128 CAR T cells exhibited greater control than the CD4 CAR T cells at an E:T ratio of 1:4 (Figure 1). Conclusions: Successful generation of TRAC -CAR T cells that can recognize and kill Env-expressing cells provides proof-of-concept for functional anti-HIV TRAC -CAR T cells, with the capacity to screen hundreds of binders both in vitro and in vivo in a pooled format due to the largely monoallelic expression of the TRAC locus. We hope to leverage this design to identify optimal engineered T cell receptors for recognizing and killing HIV-infected cells.

Poster Abstracts

536

Infusion of Autologous CCR5-Deleted CD4 Cells Does Not Change the HIV Reservoir: The TRAILBLAZER RCT Carl J. Fichtenbaum 1 , Ronald J. Bosch 2 , Hanna Mar 2 , Ashish A. Sharma 3 , Joumana Zeidan 3 , Steven G. Deeks 4 , Jeffrey M. Jacobson 5 , George Yendewa 5 , Jane Reese 5 , Joseph Puleo 6 , Rebeka Bordi 3 , Marthe Pauwels 3 , Gary Lee 7 , Michael M. Lederman 5 , Rafick P. Sekaly 3 1 University of Cincinnati, Cincinnati, OH, USA, 2 Harvard TH Chan School of Public Health, Boston, MA, USA, 3 Emory University, Atlanta, GA, USA, 4 University of California San Francisco, San Francisco, CA, USA, 5 Case Western Reserve University, Cleveland, OH, USA, 6 Brown University, Providence, RI, USA, 7 Lyell Immunopharma, Inc, South San Francisco, CA, USA Background: TRAILBLAZER is a multi-center blinded, randomized trial of infusion of autologous zinc-finger endonuclease-modified CCR5-edited CD4+ lymphocytes (MC) compared to autologous infusion of unmodified CD4+ lymphocytes (UC). We hypothesized that abrogating CCR5 expression in CD4+ memory T cells would result in accelerated clearance of the replication competent HIV reservoir. Methods: We enrolled PWH age 18-70 years on stable antiretroviral therapy (ART) with a CD4 count >350 c/mm 3 and HIV-1 RNA < 50 copies/mL for ≥ 48 weeks. Participants were randomized (2:1) to receive a single intravenous infusion of 0.5 to 4x10 10 ex vivo expanded autologous CD4+ memory T cells that have been transduced with a zinc finger nuclease designed to knock out the CCR5 gene or unmodified CD4+ T cells. All participants received cyclophosphamide conditioning (1 gm/M 2 ). The frequency of cells harboring intact HIV was estimated using the intact proviral DNA assay (IPDA, Accelevir) at entry, week 48 and 96. Results: 30 participants were enrolled (19 MC, 11 UC). The median age was 55 years (min, max: 28, 69 years); all were male sex at birth; 30% non-white race; median duration of ART 14 years; and a median CD4 count 650 cells/mm 3 (357,1105 cells/mm 3 ). There were no significant unanticipated side effects from the cyclophosphamide conditioning or study product infusion, and no Grade ≥3 treatment-related adverse events. Four participants had unevaluable IPDA data due to amplicon signal failure (all in CCR5-modified arm), 1 participant had undetectable IPDA at all time points, and 1 participant prematurely discontinued the study for reasons unrelated to the study. In the remaining 24 participants (N=13 MC; N=11 UC), IPDA had minimal changes over time (Figure), with mean changes between -0.1 and +0.1 log10 in both treatment arms. There was no evidence the CCR5-modified arm had greater changes in IPDA than the unmodified arm (p=0.48 week 96, primary endpoint). Mean changes in total provirus were between -0.08 to -0.05 log10, with no differences in the treatment arms (p=1.0 at week 96). Conclusions: In this blinded randomized controlled trial of re-infusion of autologous CCR5-gene edited CD4+ memory T cells, there was no significant reduction in HIV reservoir as measured by intact or total HIV proviruses using the IPDA assay at weeks 48 or 96. It will be important to determine the threshold of CCR5-edited memory CD4+ T cells and tissue HIV reservoir to understand these results.

535

CROI 2025 139