EACR26-0667

Auto-antibodies neutralizing IFNα improve immune checkpoint blockade in patients with melanoma

A. Eberhardt1, K. Saker2, P. Franc2, M. Dufeu3, S. Paul4, J. Caramel3, A. Puel5, S. Dalle1, T. Walzer2, S. Trouillet-Assant2, Working group/consortium6
1Cancer Research Centre of Lyon, Hospices Civils de Lyon, Team "Cancer Cell Plasticity in Melanoma", Dermatology Unit, Lyon, France
2CIRI Centre International de Recherche en Infectiologie, Lyon, France
3Cancer Research Centre of Lyon, Team "Cancer Cell Plasticity in Melanoma", Lyon, France
4CIRI Centre International de Recherche en Infectiologie, Team GIMAP, Lyon, France
5Necker Hospital for Sick Children, Laboratory of Human Genetics of Infectious Diseases, Paris, France
6ANTI-CANCER Study Group
Introduction:

Auto-antibodies neutralizing type-I IFNs (AAN-I-IFN) were recently uncovered as enriched in critically ill COVID-19 patients, where they act as major negative regulators of anti-viral immune response. IFN-I also play a central role in the regulation of anti-tumor immune response, however, the presence and functional impact of AAN-I-IFN has never been explored in tumoral context. Here, we evaluated the presence of circulating AAN-I-IFN in a large pan-cancer patient cohort and evaluated their functional impact on patient outcome.

Material and method:

We evaluated the frequency and neutralizing activity of circulating AAN-I-IFN in a cohort of 1,966 cancer patients across several tumor types (melanoma, n=453; glioma, n=270; colon cancer, n=267; diffuse large B-cell lymphoma, n=249; lung cancer, n=238; follicular lymphoma, n=107; Hodgkin lymphoma, n=96, stomach cancer, n=90; breast cancer, n=16; others, n=180) by ELISA and using an ISRE-luciferase neutralization assay in the presence of varying concentrations of IFN-Is. Response to immune checkpoint blockade (ICB) was analyzed according to relapse-free survival of patients in n=270 melanoma patients. For mechanistic analyses, mice were inoculated with Yumm1.7 ICB-resistant melanoma cell line and immunized to generate AAN-I-IFN prior treatment with anti-PD1. Spatial transcriptomic (Xenium, Human 5k panel) was performed on FFPE tumor samples from AAN-I-IFN-positive patients and -negative matched controls.

Result and discussion:

We found an overall frequency of 0.61% of cancer patients positive for AAN-I-IFN, with similar frequencies across all cancer types. Consistent with previous observations in healthy individuals, AAN-I-IFN were significantly increased across ageing in cancer patients. No differences in the frequency of AAN-I-IFN was observed between cancer patients and healthy subjects (0.617% vs 0.963%, p-value=0.23). Regarding the impact of AAN-I-IFN on melanoma patients’ clinical response to ICB, we observed a significant enrichment of long-duration responses to ICB (>4 years) in patients with AAN-I-IFN. In vivo experiments using an ICB-resistant melanoma model confirmed that only AAN-I-IFN-positive mice showed strong reduction of tumor growth in response to ICB. Finally, spatial transcriptomic analysis revealed a significant enrichment of tumor-infiltrating effector memory CD8⁺ T-cells in patients with AAN-I-IFN as compared to negative matched controls.

Conclusion:

Together, our results highlight that AAN-I-IFN potentiate ICB efficacy in melanoma patients, and uncover a previously underestimated role of auto-antibodies directed against cytokines in the regulation of anti-tumor immune response in cancer patients.