Blocking TNF to Potentiate the ICI-dependent Immune Awakening in Melanoma

Background and originality of the project with regards to the state of the art

Despite the tremendous breakthrough immune checkpoint inhibitors (ICI) such as anti-PD-1 and anti-CTLA-4 brought for the treatment of advanced melanoma patients, 60% of them do not respond or relapse within 5 years following treatment initiation. Severe immune-related adverse events (irAEs) are triggered upon ICI therapy, such as colitis, which can be treated with inhibitors (TNFi) of the Tumor Necrosis Factor alpha (TNF) such as infliximab. The pharmacodynamic impact of TNFi on the immune and clinical responses remains to be clarified.

Although TNF has been identified as a soluble factor able to trigger tumor necrosis in mice, the chronic production of TNF in the tumor microenvironment was shown to contribute to cancer progression. The investigators and others have shown this cancer promoting property relies on multiple mechanisms, including immune escape. For instance, the investigators showed that TNF triggers activation-induced cell death (AICD) in CD8 T cells, thereby limiting the infiltration of CD8 T cells in mouse melanoma tumors. In mouse models of melanoma, breast and colon cancers, TNFi enhance ICI (anti-PD-1, anti-CTLA-4) efficacy and reduce the severity of irAEs such as colitis. Hence, combining TNF-blocking antibodies to ICI promotes therapy efficacy and tolerance in vivo.

Investigators' unpublished data show all treatments promoted the differentiation of circulating T cells towards an effector memory phenotype (data not shown) between baseline (before treatment induction, week 0 [W0]) and week 6 (W6) post-treatment induction. They also significantly 1- Background and originality of the project with regards to the state of the art Despite the tremendous breakthrough immune checkpoint inhibitors (ICI) such as anti-PD-1 and anti-CTLA-4 brought for the treatment of advanced melanoma patients, 60% of them do not respond or relapse within 5 years following treatment initiation. Severe immune-related adverse events (irAEs) are triggered upon ICI therapy, such as colitis, which can be treated with inhibitors (TNFi) of the Tumor Necrosis Factor alpha (TNF) such as infliximab. The pharmacodynamic impact of TNFi on the immune and clinical responses remains to be clarified. Although TNF has been identified as a soluble factor able to trigger tumor necrosis in mice, the chronic production of TNF in the tumor microenvironment was shown to contribute to cancer progression. The investigatorsand others have shown this cancer promoting property relies on multiple mechanisms, including immune escape. For instance, the investigators showed that TNF triggers activation-induced cell death (AICD) in CD8 T cells, thereby limiting the infiltration of CD8 T cells in mouse melanoma tumors. In mouse models of melanoma, breast and colon cancers, TNFi enhance ICI (anti-PD-1, anti-CTLA-4) efficacy and reduce the severity of irAEs such as colitis. Hence, combining TNF-blocking antibodies to ICI promotes therapy efficacy and tolerance in vivo. Based on preclinical studies, the investigators implemented two clinical trials (TICIMEL and MELANFalpha) in advanced melanoma patients (stage IIIc/IV) treated with ICI [anti-PD-1, nivolumab (nivo) + anti-CTLA-4, ipilimumab (ipi)] in combination (TICIMEL) or not (MELANFalpha) with TNFi [certolizumab (certo) or infliximab (inflix)]. Whereas inflix is a bivalent anti-TNF chimeric IgG1 monoclonal antibody, certo is a monovalent PEGylated Fab' fragment of a humanized anti-TNF monoclonal. All patients (i.e., 92) have been enrolled between 2018 and 2021. TICIMEL (n=32) is a phase 1b clinical trial aiming at evaluating the safety and tolerability of co-administering ipi, nivo and TNFi (certo or inflix). Investigators' first data, obtained from 14 patients, indicate that both tritherapies are safe in humans, with a promising high response rate in the certo cohort. MELANFalpha (n=60) is a pilot study aiming at discovering predictive biomarkers of response/resistance to ICI. It notably shows high TNF levels are detected upon ipi/nivo treatment in plasmas from non-responders. Investigators' unpublished data show all treatments promoted the differentiation of circulating T cells towards an effector memory phenotype between baseline (before treatment induction, week 0 [W0]) and week 6 (W6) post-treatment induction. They also significantly increased the proportion of Th1 T cells in patients' blood. However, those effects were significantly amplified in the certolizumab cohort as compared to the ipi/nivo one. Accordingly, the investigators observed increased IFN-gamma plasma concentrations in the certolizumab cohort. Proportions of Th2, Th17 and Th1/17 T cells were not significantly affected by the three treatment regimens. Moreover, TNFi combination with ICI was associated with increased proportions of circulating MART-1-specific CD8 T cells. These cells mainly exhibited a central/effector memory phenotype and show increased expression of CXCR3 (Fig. 1D-F). The investigators performed CiteSeq single cell RNA sequencing (scRNAseq) on peripheral blood mononuclear cells (PBMCs) obtained before (W0) and at W6 from 4 advanced melanoma patients enrolled in each of the 3 cohorts. Signs of activation and maturation of immune responses were obvious in all patients from TICIMEL, but one, at W6. This indicates TNFi unlikely compromise the immune boost observed in patients treated with ipi/nivo. Moreover, the investigators show in a mouse melanoma model (B16Ova) whereby treatment conditions mimic that of TICIMEL and MELANFalpha (alphaPD-1+alphaCTLA-4+/- TNFi [anti-mouse TNF, clone XT3.11]), that the tritherapy increased the frequency of total tumor regressions and overall survival, as compared to the bitherapy. Mice, cured upon tritherapy, developed anti-melanoma memory responses, as evidenced by a rechallenge experiment. Understanding how TNF blockade increases the awakening of patients' anti-melanoma immune responses upon ICI and how the properties of TNFi may differentially modulate these responses is pivotal to:

• (i) get a better understanding of the cellular and molecular pathways, which need to be engaged to overcome some of the mechanisms responsible for ICI resistance.
• (ii) identify new and potentially better strategies to promote these responses.
• (iii) identify patients who might benefit the most from such therapies.

Hypothesis, main objective(s) and endpoint(s) of the project

Considering investigators' recently published and unpublished data, the investigators hypothesize that TNF blockade may promote ICI efficacy in patients. This effect may however be affected by the structural properties of the TNF blocking agent used. To evaluate this hypothesis, the investigators will compare the clinicobiological responses of patients from the TICIMEL (ipi+nivo+certo or inflix) and MELANFalpha (ipi+nivo, without TNFi) clinical trials. To clarify the molecular mechanisms by which certo and inflix differently impact on immune responses, the investigators will perform ex vivo and in vivo experiments in co-culture systems as well as in a unique mouse melanoma model, respectively.

Project plan describing the methodology and work to be performed To reach the primary and secondary objectives, the proposal is divided into three main WPs.

WP1: To decipher how certolizumab and infliximab influence the ICI-dependent systemic immune responses in advanced melanoma patients.

WP2: To analyse the impact TNFi have on ICI-dependent anti-melanoma immune responses in tumors in advanced melanoma patients. The investigators already performed immunohistochemistry (IHC) on formalin-fixed paraffin-embedded (FFPE) melanoma biopsies from patients enrolled in the first part of TICIMEL to evaluate the tumor-infiltrating leukocytes before and along therapy (CD3, CD8, CD4, FOXP3, CD68).

WP3: To analyse the impact TNFi have on ICI-dependent anti-melanoma immune responses in co-culture experiments and in mice.