Tumor-specific HLA/peptides (pHLA) represent attractive therapeutic targets for cancer. Two cell-based modalities can target pHLA-expressing tumors: T cell receptors (TCRs) or TCR-mimetic (TCRm) antibodies reformatted as chimeric antigen receptors (CARs). Using HLA-A2/MAGEA4230-239 as a model pHLA, we discerned the relative potency of TCR-T and CAR-T cells, informing how to best deploy these for clinical benefit. Although TCR-T cells were more sensitive at detecting low-density pHLA, TCR-T cells exerted only transient in vivo antitumor efficacy followed by tumor relapse due to deficient TCR-T cell proliferation and persistence that was associated with a more differentiated and dysfunctional phenotype. By contrast, CAR-T cells with encoded costimulatory signaling fully regressed tumors. Insufficient TCR-T cell durability was overcome by coengaging 41BB or IL-2 signaling pathways, thereby enhancing tumor control in vivo. These data establish differential activities of human TCR-T and CAR-T cells targeting the same pHLA and inform the development of optimal targeting strategies to induce durable clinical responses.

Monoclonal antibody therapies have transformed the lives of patients across a diverse range of diseases. However, antibodies can usually only access extracellular proteins, including the extracellular portions of membrane proteins that are expressed on the cell surface. In contrast, T cell receptors (TCRs) survey the entire cellular proteome when processed and presented as peptides in association with human leukocyte antigen (pHLA complexes). Antibodies that mimic TCRs by recognizing pHLA complexes have the potential to extend the reach of antibodies to this larger pool of targets and provide increased binding affinity and specificity. A major challenge in developing TCR mimetic (TCRm) antibodies is the limited sequence differences between the target pHLA complex relative to the large global repertoire of pHLA complexes. Here, we provide a comprehensive strategy for generating fully human TCRm antibodies across multiple HLA alleles, beginning with pHLA target discovery and validation and culminating in the engineering of TCRm-based chimeric antigen receptor T cells with potent antitumor activity. By incorporating mass spectrometry, bioinformatic predictions, HLA-humanized mice, antibody screening, and cryo-electron microscopy, we have established a pipeline to identify additional pHLA complex-specific antibodies with therapeutic potential.

TPS295 Background: mCRPC is characterized by an immunosuppressive tumor microenvironment with few intratumoral effector T cells, leading to low response rates to immune checkpoint inhibitors; therefore, novel immunotherapy approaches are needed. Prostate-specific membrane antigen (PSMA) is highly expressed in malignant prostate cancer cells. REGN4336 is a PSMA×CD3 bispecific antibody (bsAb) that facilitates T-cell–mediated tumor killing by bridging PSMA-expressing tumor cells with CD3+T cells, providing “signal 1” for T-cell activation. Nezastomig (REGN5678 [PSMA×CD28 bsAb]) enhances T-cell activation/proliferation by engaging the costimulatory receptor CD28 on T-cells located in proximity to PSMA, providing “signal 2.” In preclinical models, REGN4336 demonstrated dose-dependent activity against PSMA-expressing tumor cells that was enhanced in combination with cemiplimab (anti–PD-1) or nezastomig. Preclinically, nezastomig + subtherapeutic doses of REGN4336 showed similar efficacy and reduced cytokines compared with higher doses of REGN4336 monotherapy, suggesting this strategy may mitigate cytokine release syndrome (CRS). In a separate clinical trial (NCT03972657), nezastomig + cemiplimab demonstrated promising clinical activity in mCRPC, but some responders had high-grade immune-mediated AEs. REGN4336 is a novel combination partner for nezastomig. Methods: This is an open-label, Phase 1/2, first-in-human, multicenter study evaluating REGN4336 ± cemiplimab or nezastomig in pts with mCRPC (NCT05125016). Pts must have received ≥2 lines of systemic therapy for metastatic/castration-resistant disease, including a second-generation anti-androgen. Prior PSMA-targeted radioligands are permitted. Module 1 evaluates REGN4336 monotherapy with step-up dosing to mitigate CRS; Module 2 will evaluate REGN4336 + cemiplimab; Module 3 evaluates REGN4336 + nezastomig. In Modules 2 and 3, pts will receive REGN4336 step-up dosing until tolerated with Grade ≤1 CRS, followed by combination therapy. Module 1 began with REGN4336 administered SC; in addition, Modules 1–3 may evaluate REGN4336 administered IV (+ sarilumab [anti–IL-6] for CRS prophylaxis) to compare tolerability, PK, and immunogenicity across routes. Treatment continues until disease progression, intolerable AEs, or other withdrawal criteria are met. Dose escalation primary objectives: assess safety, tolerability, and PK, and determine RP2D regimens of REGN4336 ± cemiplimab or nezastomig. Dose expansion primary objectives: evaluate antitumor activity of REGN4336 ± cemiplimab or nezastomig (ORR per modified PCWG3 criteria). Exploratory objectives include PSMA-PET imaging and tissue-based biomarker analysis. This is the first study to evaluate combined ×CD3 + ×CD28 bsAb in mCRPC. As of Sept 12, 2024, 35 pts have been enrolled, including 4 in Module 3. Clinical trial information: NCT05125016 .

Improving therapies for relapsed/refractory multiple myeloma (RRMM) remains an unmet medical need, as despite improvements in treatment management, there is still no curative therapy and relapse remains a persistent problem. The emergence of T cell targeted immunotherapies, including BCMAxCD3 bispecific antibodies (bsAbs), has marked a significant advancement in patient outcomes. Among these, the BCMAxCD3 bispecific antibody, linvoseltamab, has emerged as a potential best-in-class option for RRMM; however, deeper and more durable responses in this patient population are still required. In RRMM, linvoseltamab activates T cells by engaging the T cell receptor (TCR)/CD3 complex in the presence of BCMA-expressing MM tumor cells leading to tumor cell killing. However, for optimal activation, T cells require engagement of costimulatory receptor signaling (e.g., CD28) in addition to the TCR/CD3 complex. CD28-targeting bsAbs, including REGN5668 (MUC16xCD28) and REGN5837 (CD22xCD28), are an emerging class of therapeutics capable of providing targeted costimulation and enhancing T cell effector function in combination with CD3-targeted bsAbs. In MM, CD38 is a cell-surface, multifunctional ectoenzyme that is highly expressed on MM blasts, making it a good candidate for a CD28-targeted bsAb. Thus, here we examine whether providing CD28 costimulatory signaling via a novel CD38xCD28 bsAb in the context of BCMAxCD3 bsAb therapy (linvoseltamab) can drive enhanced antitumor activity in preclinical models. We explored combining a novel CD38xCD28 bsAb antibody, REGN7945, with linvoseltamab. This combination potently enhanced linvoseltamab's anti-tumor activity by delivering a critical costimulatory signal to CD28-expressing T cells, augmenting T cell activation, cytokine production, and cytotoxicity in vitro. We demonstrated strong enhancement of linvoseltamab's anti-tumor activity with REGN7945 in multiple in vivo murine xenograft MM models. Further, in immunocompetent mice genetically engineered to express human CD3, CD28, BCMA, and CD38, we demonstrated potent combinatorial efficacy of linvoseltamab and REGN7945 to treat syngeneic murine tumors. Based on these promising preclinical results, a Phase 1 clinical trial is planned to evaluate the safety, tolerability, and preliminary efficacy of the CD38xCD28 bsAb REGN7945 and the BCMAxCD3 bsAb linvoseltamab in patients with RRMM. This innovative approach, leveraging the combinatorial potential of CD38xCD28 costimulation with linvoseltamab's targeted cytolytic activity, represents a novel therapeutic strategy with the potential to improve RRMM treatment outcomes.

Abstract Chimeric antigen receptor (CAR) T cell therapies redirect a patient's T-cells to recognize tumor antigens, and are an effective clinical strategy in the treatment of several hematological malignancies. However, targeting of solid tumors remains challenging, in part due to features of the solid tumor microenvironment that impair the expansion, persistence, and function of CAR T cells. To date, most studies employ xenograft models, thus limiting a mechanistic understanding of CAR T biology in the context of an intact immune system. To model the resistance of solid tumors to CAR T therapy and dissect how endogenous immune populations impact this process, we have generated an immunocompetent mouse model of CAR T therapy. Murine CARs recognizing human CD20 have been engineered with an extracellular scFv domain and intracellular CD3z signaling and 4-1BB costimulatory domains. Using flow cytometry and live-cell imaging assays, we demonstrate potent, antigen-dependent in vitro cytotoxicity and cytokine release by huCD20-targeting CAR T cells against solid (MC38) or heme (EL4) tumor cells engineered to express cell-surface huCD20. Moreover, huCD20 CAR T cells exhibit in vivo efficacy against both EL4-huCD20 and MC38-huCD20 tumor models, while CAR T that recognize an irrelevant antigen do not affect tumor growth. These results highlight a platform to explore resistance mechanisms and potential immune cell interactions that may impair CAR T therapy directed against solid tumors.

Supplementary Materials and Methods

Proteasomal degradation inhibitors, Bortezomib and Lactacystin had little or no effect on PRLR turnover.

In vitro Potency of HER2 ADC, PRLR ADC and HER2xPRLRbs ADC in a Panel of Breast Cancer Cell Lines

Abstract Recombinant IL-2 has been used for the treatment of metastatic melanoma and renal cell carcinoma, and induced complete, durable tumor regression in some patients. ​However, its broader use in cancer immunotherapy has been limited by severe toxicity. Here, we describe the development of REGN10597, a PD-1-targeted, receptor-masked IL-2 immunocytokine with attenuated systemic IL-2 activity but maintained capacity to engage endogenous IL-2Ra on PD-1+ T cells. We will share preclinical data on in vitro characterization as well as in vivo efficacy and toxicity studies of REGN10597. Citation Format: Jiaxi Wu, Nicolin Bloch, Aaron Chang, Ramandeep Bhavsar, Qingqing Wang, Supriya Patel, Vidur Garg, Hassan Shakil, Michael Amatulli, Drew Dudgeon, Yuetian Yan, Shunhai Wang, Willy Ramos, Pamela Krueger, Kristin Vazzana, Alison Crawford, Jacqueline Idun, Corinne Decker, David DiLillo, Samira Chandwani, Tammy Huang, Jessica Kirshner, Chungguang Guo, Lynn Macdonald, Erica Ullman, Aynur Hermann, William Olson, Samuel Davis, John Lin, Eric Smith, Tong Zhang. REGN10597: A PD-1-targeted, receptor-masked wild type IL-2 with improved therapeutic window for cancer immunotherapy [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr IA012.

HER2xPRLR bispecific antibodies with different HER2 and PRLR arms induced similar rates of HER2 degradation in T47D cells.

Supplementary Data from Immuno-PET Monitoring of Lymphocytes Using the CD8-Specific Antibody REGN5054

<p>Legends for Supplementary Figures S1-S12</p>

Legends for Supplementary Tables S1 and S2

Additional details of methods covered in the main materials and methods section

PRLR degradation was not significantly affected by PRLR antibody or Prolactin (PRL).

<p>Supplementary Figure S1: Correlation of REGN421 serum concentrations and anti-HT1080 tumor activity Supplementary Figure S2: The potent anti-tumor activity of targeting Dll4 in ovarian xenograft models is dependent on blocking stromal Dll4 Supplementary Figure S3: Dll4 antibody REGN1035 treatment induces abnormal tumor vessels in ovarian cancer xenograft models Supplementary Figure S4: Dll4 expression in ovarian cancer xenograft models is restricted to the tumor vasculature Supplementary Figure S5: Notch receptor expression in ovarian cancer xenograft models Supplementary Figure S6: Notch signaling in TOV-112D cells is mediated predominantly by Notch1 Supplementary Figure S7: Combined blockade of Dll4 and VEGF reverses liver vascular changes induced by Dll4 blockade alone Supplementary Figure S8: Dll4 is expressed in the endothelium of adult heart and liver Supplementary Figure S9: Notch1 signaling activity in heart and liver is dependent on Dll4 Supplementary Figure S10: Pericyte coverage in ovarian xenograft tumors is unaltered in response to anti-Dll4 antibody treatment Supplementary Figure S11: Direct contact between endothelial and tumor cells in TOV-112D tumors Supplementary Figure S12: Direct contact between endothelial and tumor cells in A2780 tumors</p>

In T47D/HER2 cells, HER2xPRLR bsAb1 augments HER2 ADCinduced cell cycle arrest in G2M phase.

Supplementary Table S1: Dll4 Ab REGN421 binds human and monkey Dll4 with high affinity Supplementary Table S2: Combined blockade of Dll4 and VEGF signaling results in enhanced inhibition of A2780 ovarian tumors

The clinical success of cancer immunotherapy, including engineered T cell therapies, has revolutionized treatment paradigms and patient outcomes. While hematological tumors have benefited most from cell therapy approaches, the treatment of solid tumors remains a challenge in part due to the limited availability of abundant and tumor-specific cell-surface antigens. Peptides derived from intracellular tumor-specific proteins, such as cancer-testis antigens (CTAs), that are presented via HLA (pHLA) enable a therapeutic opportunity to target tumors while sparing normal tissue. Surface-accessible pHLA complexes may be targeted with engineered T cell receptors or TCR mimetic antibodies reformatted to chimeric antigen receptors (CARs). Using MAGE-A4 as a model CTA, we compared engineered human TCR- and CAR-T cells head-to-head to understand how to best deploy these modalities. To this end, we generated fully-human, HLA-A2/MAGE-A4(230-239)-specific TCR and CARs harboring CD28/CD3z or 41BB/CD3z signaling domains. TCR and CAR-T cells demonstrated similar robust on-target reactivity, cytokine release, and target cell lysis in vitro. In vivo, each of these modalities showed potent, dose-dependent anti-tumor efficacy against human xenograft tumors expressing low, endogenous levels (~500 cell-surface copies) of the MAGE-A4 peptide. However, differences emerged when we examined the in vivo kinetics and durability of tumor suppression. MAGE-A4 CD28/z CAR-T demonstrated the most rapid and potent tumor clearance, while the 41BB/z CAR-T showed delayed but ultimately complete efficacy. TCR-T cells induced tumor regressions during the first 2 weeks of treatment, but this response was transient and followed by tumor relapse. These differential responses correlated with early, modest accumulation of CD28/z CAR-T in line with fast tumor clearance. 41BB/z CAR-T showed a remarkable ~800-fold expansion in the tumor versus limited in vivo TCR-T proliferation. Mechanistically, the MAGE-A4 TCR induced strong CD3 proximal signaling associated with a greater induction of T cell dysfunction markers and limited cytotoxic potential in vitro. However, stimulating 41BB signaling pathways in the MAGE-A4 TCR T cells augmented long-term cytotoxicity. These data demonstrate that tumor-specific pHLA complexes can be potently targeted by both TCR and CAR-T cells, and that co-stimulatory signaling is necessary to mediate durable anti-tumor activity. Citation Format: Corinne E. Decker, Jacqueline Idun, Katja Mohrs, Thomas Meagher, Kevin Bray, Iryna Petriv, Jonathon Golas, Timothy Helms, Dharani Ajithdoss, Gavin Thurston, John Lin, Jessica R. Kirshner, David J. DiLillo. TCR-T and CAR-T cells targeting HLA-A2/MAGEA4 demonstrate differential tumor control, reflecting co-stimulatory signaling requirements [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1780.