PO.IM02.07 · 免疫学

Endogenous retrovirus-derived neoantigens enable a personalized cancer vaccine strategy for glioblastoma

海报缩略图:Endogenous retrovirus-derived neoantigens enable a personalized cancer vaccine strategy for glioblastoma
编号 6975 展板 3 时间 4/22 09:00–12:00 区域 Section 8 主讲 Megan Benz, BS
分会场 Novel Models of Immunotherapy Response
查看完整资料 下载 PDF 登录后可访问当前开放资料 AACR 官方页面 ↗

作者与单位

Megan Campbell Benz1, Kenan Zhang2, David M. Ashley3, Christian Garde4, Kirit Singh3, Ohrt Andersen Rasmus4, Birgitte Rønø4, Daniela Kleine-Kohlbrecher4, Stine Friis Thorsen4, Kelly M. Hotchkiss1, Jose R. Conejo-Garcia1, Mustafa Khasraw5

1Duke University School of Medicine, Durham, NC,2Duke University, Durham, NC,3Duke University Medical Center, Durham, NC,4Evaxion Biotech, Hørsholm, Denmark,5Duke Cancer Institute, Durham, NC

摘要 Abstract

Introduction/Rationale: Glioblastoma is refractory to immunotherapy because of low mutational burden and paucity of canonical neoantigens. Aberrant expression of endogenous retroviral elements (ERVs), reactivated genomic remnants of ancient viral insertions, may provide an alternative, tumor-specific antigen source suitable for personalized vaccines incorporating both mutation- and ERV-derived epitopes. Methods: Tumor and matched normal tissues from 25 glioblastoma patients, including 18 long-term survivors (>5 years) and 7 short-term survivors (<18 months), underwent whole-genome and RNA sequencing. Candidate antigens were predicted using AI-based pipelines to identify mutation-derived neoantigens and to derive ERV-encoded epitopes. AI-identified tumor-specific vaccine candidates were ranked by expression, predicted MHC class I and II binding affinity, clonality, and immunogenicity. Top candidates informed design of a personalized DNA vaccine encoding ten patient-specific sequences with a CCL19 molecular adjuvant to enhance dendritic-cell recruitment. ERV-derived immunogenicity was assessed by IFN-gamma ELISpot using peripheral blood mononuclear cells (PBMCs) from HLA-typed healthy donors stimulated with synthetic peptides. Results: High-quality mutation-derived neoantigens were rare (median 3 per patient). In contrast, all tumors expressed abundant, patient-specific ERV transcripts independent of mutation load, with several loci showing clonal, tumor-restricted expression. ERV-derived peptides demonstrated strong predicted MHC class I and II binding, and 22 of 25 tumors contained ERV epitopes meeting criteria for vaccine inclusion. These mapped predominantly to HERV-K and HERV-W families and frequently overlapped interferon-stimulated genomic loci, consistent with innate immune or epigenetic activation. Functional validation studies have demonstrated significant ERV-specific IFN-gamma responses in certain HLA allele contexts, highlighting the potential for allele-specific immunogenicity. While T-cell responses were observed in selected donor profiles, others exhibited more restricted reactivity despite validated positive controls, underscoring the importance of genetic context in immune response profiles. Additionally, our ongoing analyses are focused on elucidating ERV expression signatures that may correlate with sustained immune control by comparing long-term and short-term survivor cohorts, with the aim of identifying key immunological patterns associated with favorable outcomes. Conclusion: ERVs are consistently expressed, immunogenic, and largely tumor-restricted in glioblastoma, providing a scalable antigen source that may overcome limitations imposed by low mutational burden. These findings offer a strong proof-of-concept for personalized DNA vaccines incorporating ERV-derived neoantigens.
利益披露 Disclosure
M. C. Benz, None.. C. Garde, None.. B. Rønø, None.. S. Friis Thorsen, None.

在会议检索中打开