Abstract IA21: Immune mechanisms in gastrointestinal stromal tumor

Abstract

Abstract Gastrointestinal stromal tumor (GIST) is the most common sarcoma subtype and typically contains an activating mutation in KIT or, less frequently, PDGFRA. Imatinib mesylate inhibits the KIT and PDGFRA oncoproteins and achieves a partial response or stable disease in over 80% of patients with advanced GIST. Unfortunately, disease progression typically occurs within 2 years, usually due to a secondary KIT mutation. Moreover, while adjuvant imatinib prolongs recurrence-free survival, it does not appear to increase the chance of cure following surgery. Thus, novel therapeutic strategies are needed for GIST, which has led us over the last 8 years to investigate the immune response. In a genetically engineered mouse model of GIST with a Kit deletion, we discovered that imatinib works partly by inhibiting tumor cell production of an immunosuppressive enzyme called indoleamine 2,3-dioxygenase (IDO), which catabolizes L-tryptophan. The downstream tryptophan metabolites are immunosuppressive and promote the development and activation of regulatory T cells (Treg) and suppress effector T cells. Consequently, by decreasing IDO production, imatinib can enhance the antitumor immune response. IDO expression was also found in human GIST specimens. While imatinib alone increased the T-cell immune response, we studied the effects of additional T-cell activation. Anti-CTLA-4 increased the antitumor efficacy of imatinib in the mouse model, which led to an NCI trial (NCT01643278) that tested the combination of ipilimumab and dasatanib in advanced, refractory GIST. Disappointingly, there were no responses by RECIST criteria, although dasatinib has limited activity in resistant GIST and concomitant tyrosine kinase inhibition was necessary in the mouse model. Nevertheless, the importance of intratumoral T cells was also shown by Zitvogel, who found that intratumoral CD8 T cells detected by immunohistochemistry correlated with survival in human GIST. Recently, we analyzed 106 human GISTs and found that intratumoral T cells often had a high-level expression of programmed death 1 (PD-1), and there was tumor expression of its ligand PD-L1. In the mouse model, anti-PD-1 or anti-PD-L1 alone lacked efficacy, but each increased the antitumor effects of imatinib. We also performed a detailed analysis of tumor-associated macrophages (TAMs) in 57 human GISTs. Macrophages are generally categorized as classically (M1) or alternatively (M2) activated, although these terms really just define the extremes of a spectrum. M1 macrophages can lyse tumor cells, secrete IL-12 and TNF, and present antigen to T cells. Meanwhile, M2 macrophages promote tumor growth by secreting IL-10 and supporting angiogenesis. Surprisingly, we found that TAMs in GIST were M1 at baseline, became M2 after imatinib therapy, and reverted to M1 after imatinib resistance developed. Accordingly, transcriptome analysis of purified TAMs revealed that there were no genes expressed more than two-fold between untreated and resistant GISTs. Because the TAMs were M1, it may be preferable to pursue strategies to activate TAMs in GIST, instead of depleting them. Recently, we have found in the GIST mouse model that anti-CD40 therapy increased the antitumor response to imatinib. It is now clear that there is a natural antitumor immune response to GIST, given the presence of M1 TAMs and intratumoral T cells expressing PD-1. However, other aspects of the immune response to GIST remain undefined. Manipulating the immune environment in GIST has potential therapeutic benefit. Several immunotherapy trials are opening in GIST, including those incorporating IDO inhibition and anti-PD-1. Citation Format: Ronald DeMatteo. Immune mechanisms in gastrointestinal stromal tumor [abstract]. In: Proceedings of the AACR Conference on Advances in Sarcomas: From Basic Science to Clinical Translation; May 16-19, 2017; Philadelphia, PA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(2_Suppl):Abstract nr IA21.