Abstract 3020: Human leukocyte antigen (HLA) typing of a broad panel of cancer patient-derived xenograft (PDX) models for immune therapies

Abstract

Background In immuno-oncology research, an appropriate microenvironment for both human tumor cells and human immune cells is an important but challenging factor for successful development of relevant preclinical models. Humanized patient-derived xenograft (PDX) models can support establishment of such an environment based on matching human leukocyte antigen (HLA) profiles of PDX and compatible human immune cell populations. In this study, we determined the individual HLA types of a broad panel of 291 established PDX models from 18 different tumor entities. Furthermore, we performed comprehensive HLA matching analyses of all PDX models and 13 peripheral blood mononuclear cell (PBMC) donors to enable personalized studies with HLA-matched PDX. Methods The HLA profile of each PDX was determined comprising HLA class I, II and non-class types in 4-digit resolution. Matching analyses were performed according to donor-recipient HLA matching criteria recommended by the Blood and Marrow Transplant Clinical Trials Network (BMT CTN). For comparative HLA profile analyses, allele frequencies of 8862 German stem cell donors provided by The Allele Frequency Net Database were downloaded. Results The individual HLA profiles of all PDX models were generated comprising HLA-A, -B and -C (class I), HLA-DQA1, -DQB1, -DRB1, -DPA1 and -DPB1 (class II) as well as HLA-H (non-class) types. Comparative analyses of HLA profiles revealed that obtained frequencies of PDX HLA alleles are comparable with frequencies of a representative population of 8862 German stem cell donors. HLA profile matching analyses performed for all PDX models and PBMC donors resulted in 15 matches regarding PDX from head and neck, lung, mammary and pancreatic carcinoma as well as from endometrial cancer, glioma, lymphoma, melanoma and sarcoma. Dependence of immune therapy efficacy on matched HLA profiles was shown exemplarily for a head and neck squamous cell cancer (HNSCC) PDX. The HNSCC PDX was humanized with PBMC from 5 donors representing different degrees of HLA matching (33 - 100 %). Treatment with the checkpoint inhibitor Nivolumab caused a more powerful tumor growth inhibition in matched HLA profiles compared to non-matched profiles. This indicated a relation between antitumoral immune therapy effect and HLA matching in preclinical models. Conclusions We generated a comprehensive HLA profile portfolio containing information of a broad panel of PDX models. Matched HLA profiles of PDX models and PBMC donors enable personalized, preclinical immune-oncology studies to support the development of novel therapeutic strategies. Citation Format: Theresia Conrad, Annika Wulf-Goldenberg, Maria Stecklum, Michael Becker, Konrad Klinghammer, Jens Hoffmann. Human leukocyte antigen (HLA) typing of a broad panel of cancer patient-derived xenograft (PDX) models for immune therapies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 3020.