Dissection of Single-Cell Landscapes for the Development of Chimeric Antigen Receptor T Cells in Hodgkin Lymphoma

Abstract

The success of targeted immunotherapies for hematological malignancies has heralded their potential as salvage therapies as well as in earlier treatment lines (Cappell & Kochenderfer, 2023). While conventional chemotherapy-based treatments can achieve long-term survival in up to 90 % of treated patients with classic Hodgkin lymphoma (cHL), these therapies are associated with treatment-related comorbidities, calling for more tailored and specific approaches (Schaapveld et al., 2015; Shanbhag & Ambinder, 2018). While targeted treatments, especially immunotherapies are taking oncology by storm, the utility in cHL is so far limited to CD30 and PD-1-targeting strategies and there is a clear lack of drugable relevant target structures in this disease. This can be partly attributed to technical difficulties of analyzing the malignant Hodgkin-Reed-Sternberg (HRS) cells specifically. Capitalizing on our previous work using large scale data mining to inform target discovery, we hypothesized that combining different analytical methods with large single-cell RNA-Sequencing (scRNA-Seq) datasets would permit selective target definition with functional relevance to the disease and thereby allow the development of novel immunotherapeutic strategies. Leveraging microarray profiles of laser-dissected HRS cells and a scRNA-Seq cohort of cHL patients (total of n = 44 primary samples; n = 34 cHL samples; n = 10 RLN (reactive lymph node) control samples), we screened for novel target antigens highly expressed on HRS cells with functional relevance in the tumor microenvironement (TME) of cHL. Unbiased in silico analyses revealed CD80, CD86 and PD-L1 as most suitable candidate target antigens with CD86 showing the highest expression on HRS cells. ScRNA-Seq analyses unveiled a shift of the CD80-CD86-CTLA-4-CD28 towards the immunosuppressive CTLA-4 axis in the TME of cHL compared to RLN controls. In advanced cell culture models, including iPSC-derived organoid models, blockage of CD86 lead to the decreased expression of PD-1 and CTLA-4 and an overall reversal of the exhaustive phenotype of cHL-associated T cells. High protein expression of CD86 on HRS cells and in the TME (cHL-infiltrating tumor-associated macrophages (cHL-TAM), B cells) was confirmed in different validation cohorts including relapsed and refractory cHL (r/r cHL) patients by conventional immunohistochemistry and multiplexed immunofluorescence (n = 34 cHL patients). Following target identification, CAR T cells redirected against CD86 were developed and the functionality of these CAR T cells was investigated in preclinical models both in vitro and in vivo. Anti-CD86 CAR T cells effectively deplete cHL-TAM and are highly effective in various in vitro and in vivo models of cHL, including models of CD30-negative disease. Given the fundamental role of the CD80-CD86-CTLA-4-CD28 axis in the generation of the adaptive immune response, detailed toxicity assessments were carried out leveraging murine surrogate anti-CD86 CAR T cells, with similar binding and activation thresholds as their human counterpart. These anti-mCD86 CAR T cells did not cause toxicities in lymphodepleted, immunocompetent mice. In addition, the impact of anti-CD86-directed immunotherapies (e.g. anti-CD86-blocking antibodies, anti-mCD86 CAR T cells) on bacterial host defense and formation of antigen-specific adaptive immunity was investigated in syngeic mouse models. Anti-CD86 immunotherapy did not lead to enhanced bacteremia in a model of gram-negative sepsis, while preclinical vaccination models revealed a mildy reduced formation of antigen-specific T cell development in mice. In summary, we provide a framework for unbiased, multi-dimensional target screening and highlight the functional relevance of the immunosuppressive CD86-CTLA-4 axis in cHL. CD86-directed immunotherapy could reverse the exhaustive phenotype of cHL-associated T cells, while demonstrating strong treatment efficacy in xenograft mouse models. Importantly, elaborate toxicity assessments of anti-CD86-targeted immunotherapies utilizing syngenic mouse models did not reveal measureable toxicity in mice. Overall, our data emphasizes the vast translational potential of CD86-targeted immunotherapies in cHL and provide a strong rationale for further clinical investigations.