Abstract NG04: Immune evasion mechanisms in melanoma resistant to targeted- and/or immunotherapy
Willy Hugo,Lu Sun,Marco Piva,Shirley Lomeli,Chunying Song,Antoni Ribas,Roger S. Lo
DOI: https://doi.org/10.1158/1538-7445.am2018-ng04
2018-07-01
Immunology
Abstract:Abstract Despite recent advances in mutation- targeted and immune therapies for melanoma, prolonged patient survival is still hampered by therapeutic resistance. For MAPKi, acquired resistance is common. For example, a majority of V600E/K mutant BRAF metastatic melanomas treated with MAPK inhibitor (MAPKi) will acquire resistance within 1-2 years of the treatment initiation, despite BRAF inhibitors (BRAFi) and their combination with MEK inhibitors (BRAFi+MEKi) being remarkably active initially. Our work and others' have shown that immune modulation by MAPKi may contribute to the in vivo anti-tumor activities of this oncogene-targeted approach. Importantly, my own recent study has shown that around half of all acquired or late MAPKi resistant melanoma display profound CD8 T-cell depletion and exhaustion, which were tightly associated with reduced antigen-presentation. Immunotherapies using antibodies against the T cell checkpoints, CTLA-4 and PD-1, have been shown to enhance the ability of endogenous T cells to destroy malignant cells in a variety of cancers, including melanoma. When successful, these therapies can result durable tumor control. However, the majority of patients do not respond to such therapies. Multiple studies including ours have delineated multiple factors limiting the efficacy of single agent anti-CTLA-4 or anti-PD-1 treatments, including the lack of tumor infiltration by T cells, insensitivity to IFN-gamma signaling and the activation of immune-suppressive processes represented by Innate anti-PD-1 RESistance (IPRES) signatures. Functionally, IPRES signatures include the downstream genes of potent immune-suppressive pathways and wound-healing processes linked to suppression and exclusion of T cells (e.g., angiogenesis, extracellular matrix reorganization, mesenchymal/invasive transition). Unsupervised gene signature enrichment clustering revealed four subclasses with combinations of high/low enrichment scores of interferon (IFN)/T-cell inflammation and of IPRES related signatures in both our dataset and the TCGA melanoma. The cellular processes represented by the four subsets significantly resemble those induced during wound healing where: 1) the IFN-high stage matches the early, acute inflammation stage and 2) the IPRES-high stage matches the stage of neovascularization, extracellular matrix remodeling and suppression of inflammation. Importantly, the patients from different subclasses have statistically different overall survival where the IFN-high/IPRES-low group showed the best survival and IFN-low/IPRES-low group was the worst. In murine mutant BRAF melanoma YUMM1.7, regressing/responding tumors on the 6th day of BRAFi treatment showed IFN-high/IPRES-low pattern with a significant infiltration of T cells and other immune cell populations into the tumor. The residual tumors on the 15th day showed IFN-high/IPRES-high pattern and showed further immune infiltration. By the time of acquired resistance to BRAFi, however, all tumors were in the IFN-low/ IPRES-high stage and there was no or minimal immune infiltration in the tumor. We observed that MAPKi-induced an IFN-high/IPRES-high state in the residual melanoma tumors from patients undergoing MAPKi treatment. Given the immune suppressive nature of IPRES, we hypothesized that its activation may cause the observed intratumoral immune (T-cell, in particular) dysfunction and depletion. Indeed, combination of MAPKi with immune checkpoint blockade (ICB) therapy targeting the PD-1 only works when the ICB is administered before the IFN-high/IPRES-high stage. Also intriguing is the finding that IPRES was induced in vitro along with a tumor immune phenotype switch MAPKi resistant melanoma cell lines, suggesting, a tumor cell-intrinsic process. We found that IPRES signatures include potent immune-suppressive pathways like the VEGF and TGFB pathways. TGFB signaling is known to regulate T-cell inflammation and promote T-cell tolerance by downregulating TCR signaling. TGFB ligands (TGFB1/2/3 and ACTIVIN-A/B) are normally secreted by immune cells including T-cells, M2-macrophages and endothelial cells during the inflammation resolution phase of wound healing. In our data and the TCGA melanoma, we observed higher mRNA and protein expression of TGFB ligands in the IPRES-high melanoma subgroups and in tumors treated with MAPKi. We postulate that high levels of TGFB ligands, from tumor and stromal sources, could counteract the T-cell anti-tumor immunity. Of interest, a study has shown some efficacy of targeting the TGFB pathway by fresolimumab (a human monoclonal antibody against TGFB1-3) in melanoma (Clinical Trial ID: NCT0035646017). Another study currently combines nivolumab (anti-PD-1) with galunisertib (a TGFBR type I/II small molecule inhibitor) in multiple solid tumors (NCT02423343). In relation to the VEGF pathway, we observed an anti-correlation between the mRNA levels of VEGFA and CD8A in both our MAPKi and the TCGA melanoma cohorts. We also observed that VEGFA mRNA and its protein level were expressed higher in the non-IFN-/T-cell inflamed subclasses. Several publications had shown that tumors that were non-responsive to anti-PD-1 have higher VEGFA levels associated with either PTEN loss or beta-catenin activation. Tumor-secreted VEGFA was reported to inhibit T-cell diapedesis from the vasculature, and thus infiltration into the tumor bed. Our observations are thus in line with such a role of VEGFA in blocking T-cell homing into the tumor. Consistently, a recent report showed that anti-VEGFA antibody pre-treatment followed by co-treatment with aPD-L1 improved treatment responses among renal cell carcinoma (RCC) patients. The authors reported increased tumor-specific T-cell trafficking into the tumor with the dual treatment. An earlier study combining anti-CTLA-4 with anti-VEGFA also showed similar result. Thus, anti-VEGFA combined with anti-PD-1/L1 may have a role in addressing non-T-cell inflamed melanoma with innate/acquired aPD-1 resistance or acquired MAPKi-resistance. Thus, given the clinical availability of both TGFB and VEGFA antagonists, we are testing the efficacy of combinatorial MAPKi and anti-TGFB/VEGFA to improve T-cell mediated, antitumor immunity through suppression of IPRES. Finally, our recent study showed that MAPKi-treated, regressing melanoma already evolve epigenetic alterations, temporally far ahead of any clinical evidence of disease progression. However, it was not clear if the observed transition into an IPRES-high, immune-suppressive state in these melanoma has any epigenetic basis. In parallel studies in cell lines, we characterized two groups of melanomas derived from continuous treatment with BRAFi or BRAFi+MEKi: MAPK-redundant (Rr) vs. MAPK-reactivated (Ra). The Ra lines could be resensitized to additional MAPK inhibitors while the Rr lines could not (due to MAPK-independent growth and surival). Intriguingly, only the Rr-lines showed enrichments of the IPRES signatures. Importantly, we noted that chromatin enhancer regions (marked by significant H3K27ac peaks) of the Rr lines were correlated with more invasive melanomas. The same invasive melanoma enhancer profile can also be observed in the patient-derived MAPKi-on-treatment tumors. By analyzing the actual mRNA expression changes of the genes proximal to the invasive melanoma specific H3K27ac peaks, we found that the upregulated genes were enriched with IPRES-related processes. This implies that the tumor cell-intrinsic up-expression of IPRES-related genes is regulated at the epigenetic level. In summary, we anticipate that our study will lead to novel insights into the determinants and/or bottlenecks of effective and durable anti-tumor responses by tumor-specific T-cells in order to pave the way to a new combination therapies in human melanoma patients. Citation Format: Willy Hugo, Lu Sun, Marco Piva, Shirley Lomeli, Chunying Song, Antoni Ribas, Roger S. Lo. Immune evasion mechanisms in melanoma resistant to targeted- and/or immunotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr NG04.
immunology