Exploring the Dynamic Crosstalk between the Immune System and Genetics in Gastrointestinal Stromal Tumors

Simple Summary The effect of genetic alteration on the prognosis of patients affected by GIST has been extensively demonstrated. Unfortunately, not all GISTs could benefit from targeted therapies, underlining the need to deeply understand other predictive mechanisms. The link between immune checkpoints (especially PD-L1 expression), the tumor microenvironment, and the clinical behavior of GIST with different driver mutations is under investigation and represents an intriguing research field that could lead to improved prognostication in GIST. Abstract Gastrointestinal Stromal Tumors (GISTs) represent a paradigmatic model of oncogene addiction. Despite the well-known impact of the mutational status on clinical outcomes, we need to expand our knowledge to other factors that influence behavior heterogeneity in GIST patients. A growing body of studies has revealed that the tumor microenvironment (TME), mostly populated by tumor-associated macrophages (TAMs) and lymphocytes (TILs), and stromal differentiation (SD) have a significant impact on prognosis and response to treatment. Interestingly, even though the current knowledge of the role of immune response in this setting is still limited, recent pre-clinical and clinical data have highlighted the relevance of the TME in GISTs, with possible implications for clinical practice in the near future. Moreover, the expression of immune checkpoints, such as PD-L1, PD-1, and CTLA-4, and their relationship to the clinical phenotype in GIST are emerging as potential prognostic biomarkers. Looking forward, these variables related to the underlying tumoral microenvironment in GIST, though limited to still-ongoing trials, might lead to the potential use of immunotherapy, alone or in combination with targeted therapy, in advanced TKI-refractory GISTs. This review aims to deepen understanding of the potential link between mutational status and the immune microenvironment in GIST.


Introduction
In the late 1990s, the therapeutic approach for patients with advanced gastrointestinal stromal tumors (GISTs) was dramatically revolutionized by the development of targeted therapies that completely redesigned the clinical history of this neoplasm. In this clinical setting, the administration of imatinib mesylate (IM) and several other multi-kinase inhibitors has been undoubtedly associated with improved outcomes for patients [1]. GISTs' sensitivity to targeted therapies strictly relies on the presence of pathogenic alterations Macrophages are divided into types M1 and M2. Anti-inflammatory macropha called the "M2 type", as opposed to M1 pro-inflammatory macrophages, promote an munosuppressive environment through their high expression of cytokines such as I and transforming growth factor β (TGFβ) [9].  The presence of both innate and adaptive immune cells in solid tumors and their correlation with the clinical outcome of patients have been widely demonstrated [34]. A growing body of studies has revealed that the TME of GISTs, mostly populated by TAMs and tumor-infiltrating lymphocytes (TILs) [7,36], has a significant impact on prognosis [37] and response to treatment.
Macrophages are divided into types M1 and M2. Anti-inflammatory macrophages, called the "M2 type", as opposed to M1 pro-inflammatory macrophages, promote an immunosuppressive environment through their high expression of cytokines such as IL-10 and transforming growth factor β (TGFβ) [9].
The crosstalk between tumor cells and T lymphocytes is shown in Figure 2.  In this figure, the following are represented: T-cells, M2  macrophages, dendritic cells, and natural killer cells. The M2 macrophages are characterized by  their high expression of cytokines such as IL-10 and TGFb; thanks to these cytokines, the M2 macrophages promote an immunosuppressive environment. Macrophages are divided into types M1 and M2. Anti-inflammatory macrophages, called the "M2 type", as opposed to M1 pro-inflammatory macrophages, promote an immunosuppressive environment through their high expression of cytokines such as IL-10 and transforming growth factor β (TGFβ) [9].
The crosstalk between tumor cells and T lymphocytes is shown in Figure 2. , and CD8+ lymphocytes) and higher production of immunogenic neo-antigens, leading to increased tumor regression (based on Vitiello et al.) [5].
M2 macrophages are implicated in the promotion of neoplastic spreading through the stimulation of angiogenesis, the proliferation of cancer cells, and the remodeling of extracellular matrix (ECM); in fact, this subtype of TAMs is more expressed in metastatic lesions than in primary GISTs [7,38,39]. Furthermore, metastases are enriched by a high Figure 2. The crosstalk between tumor cells and T lymphocytes in GIST. PDGFRA-mutant GISTs, compared to KIT-mutant GISTs, show an increased number of immune cells (such as NK+, CD4+, and CD8+ lymphocytes) and higher production of immunogenic neo-antigens, leading to increased tumor regression (based on Vitiello et al.) [5].
M2 macrophages and Tregs are the most represented cells, and they determine a strongly immunosuppressive TME in GISTs [7].
A small fraction of tumor-infiltrating immune cells is represented by B cells and DCs, which are usually poorly expressed or even absent in GISTs [42].
NK cells are innate immune system lymphocytes involved in immune response to tumors, and they seem to be interestingly enriched in the GIST microenvironment where they specifically target cells with a lower expression level of major histocompatibility complex 1 (MHC I), a common feature of these neoplasms. Rusakiewicz et al. [37] demonstrated that the number of NK cells, localized mainly within the fibrous stroma surrounding tumor cells [36], was significantly associated with low mitotic index in a cohort of 91 GIST patients [9]. A high level of CD3+ T and NK cells correlated with a reduced relapse rate and a more favorable prognosis in untreated metastatic GISTs [9,36,37]. CD3+ T and B cells are more concentrated in intestinal and highly proliferating GISTs as compared with those arising in the stomach and with a low proliferation index (<10%) [36]. Even though the knowledge about the TME is growing, the correlations between immune cells and other prognostic factors (tumor location, size, or mitotic index) are still controversial [43], suggesting that other factors may influence the composition of the TME.

Looking Forward: Driver Mutations and Immune Microenvironment
The type of GIST driver mutation represents an important prognostic factor, correlating with clinical features and biological aspects of the disease [5]. The presence of deletions within c-KIT codons 557/558 is associated with a more aggressive behavior compared to other exon 11 mutations, thus resulting in shorter recurrence-free survival (RFS) for patients with resected GIST and shorter PFS for metastatic patients [44,45].
Recent evidence also suggests that the genotype can influence immune cells infiltrating the TME [43]. Surprisingly, PDGFRA-mutant GISTs showed an increased number of immune cells, compared with c-KIT-mutant GIST, and an overexpression of stimulatory cytokines (e.g., CXCL14) which additionally activate NK+, CD4+, and CD8+ cells, leading to tumor regression [5]. In particular, a study by Xiangfei S. et al. [10] demonstrated that TILs were more abundant in GISTs harboring a PDGFRA mutation. Another study by Vitiello et al. [5] showed that 75 KIT-mutant GIST patients harbored a lower number of immune cells than did PDGFRA-mutant GISTs.
Gasparotto et al. [43] studied the possible correlation between the presence and type of driver mutation and neo-antigens' immunogenic capability to bind to patient-specific HLA types. Tumor neo-antigens are short peptides that can interact with HLA molecules and be presented on the surface of tumor cells to activate T-cells' cytotoxic immunity. It turned out that GISTs carrying KIT and PDGFRA driver mutations produced more immunogenic neo-antigens compared to BRAF-or NF1-mutated GISTs and harbored a richer immune infiltrate [43]. WNT/β-catenin signaling (WNT/β-cat), RAS, and the Hedgehog (HH) pathway, usually activated in K/P WT GIST, could lead to lower tumor immune infiltration and immune evasion [43].
Not only the gene but also the type of pathogenic variant (PV) can relate to different features. D842V-mutated GISTs are more enriched in immune cells, mainly CD8+ T cells, than are non-D842V ones [10,46].
These findings suggest that PDGFRA-mutant GISTs, characterized by intrinsic resistance to standard TKIs and a better prognosis, are more immunogenic compared to genetic alterations sensitizing to common TKIs [5].

Driver Mutations and Immune Checkpoint Expression to Improve Prognostication in GIST
The expression of immune checkpoints and its relationship to the clinical phenotype in GIST is not understood, as it has previously been poorly evaluated.
Immune checkpoints, such as PD-L1, PD-1, and CTLA-4, by escaping immune surveillance, play a key role in tumor progression and influence the survival of patients with solid tumors [16,47]. According to the recent knowledge, despite some known limitations, immune checkpoint expression could be a potential prognostic factor and predictive biomarker of response to immune checkpoint inhibitors in patients with solid tumors [45,[48][49][50][51].
PD-1, a type I transmembrane protein of the immunoglobulin superfamily, and its ligands are key regulators in a wide spectrum of immune responses and play a critical role in autoimmunity and self-tolerance, as well as in cancer immunology [52]. PD-1 is expressed on a variety of immune cells, such as monocytes, T cells, B cells, DCs, and TILs, while PD-L1, the main ligand for PD-1, is expressed on several hematopoietic cells, especially on tumor antigen-presenting cells (APCs), and on peripheral nonhemopoietic cells. [53]. Moreover, PD-L1 can be highly expressed on DCs or on the tumor cells themselves [54]. PD-L2 is expressed on APCs and other immune and non-immune cells [53]. The PD1/PD-L1 axis is the most common immune checkpoint pathway, and it impairs T-cell proliferation and effector functions, leading to apoptosis of tumor-specific T cells [55]. Several results from the literature have shown that high expression of PD-L1 is related to poor clinical outcomes in patients with solid tumors. Previous findings showed that tumor PD-L1 expression, evaluated via immunohistochemistry (IHC) on formalin-fixed and paraffin-embedded tumor sections, was greater in GISTs than in several types of soft tissue sarcomas [56], and the PD-L1 expression level was also associated with high-risk GIST patients showing poorer outcomes in therapeutic settings [57].
Recently, circulating immune checkpoint molecules have been shown to have potential prognostic significance in metastatic GISTs [58]. High levels of plasma PD-1, PD-L1, and the butyrophilin family proteins sBTN3A1 and pan-BTN3As seem to predict a shorter PFS and a poor prognosis in patients with KIT exon11-mutated metastatic GIST treated with IM as a first-line treatment [56]. Interestingly, in the same study, lower plasma levels of soluble PD-L1 and pan-BTN3As and the absence of KIT exon 11 deletions or deletions/insertions at codons 557 and/or 558 were significant prognostic factors for a longer PFS in mGIST patients, showing different expression profiles of immune checkpoints in GISTs harboring different driver mutations [58].
Preliminary data highlight the potential role of PD-L1 expression as an independent prognostic factor also in PDGFRA-mutant GISTs [10]. The expression of PD-L1 is heterogeneous in PDGFRA-mutant GISTs, and it is inversely related to tumor size, suggesting the inhibition of tumor proliferation and a better prognosis [10]. Furthermore, PDGFRA-driven GISTs express a high concentration of indoleamine-2,3-dioxygenase (IDO), an immune checkpoint molecule that is correlated with high inflammatory cell infiltrates, CD4+ cells in particular [10].
In a recent study, Seifert et al. [59] analyzed matched tumor and blood samples from 85 patients with GIST and studied the expression of immune checkpoint molecules other than PD-1/PD-L1 using flow cytometry. Seifert et al. [59] found that immune checkpoint molecules such as lymphocyte activation gene 3 (LAG3) and T-cell immunoglobulin mucin 3 (TIM3) are upregulated on TILs in GIST tissue.
The relationship between prognosis and the quality of stromal differentiation (SD) and immune checkpoint expression in GIST has been investigated, with possible clinical practice implications for SD in the near future as a prognostic tool.
In immune checkpoint inhibitor analysis, authors focused on PD-L1 and v-domain Ig suppressor of T-cell activation (VISTA) levels, the latter being a biomarker of the tumoral microenvironment status.
Of note, an immature stroma was found to be associated with lower PD-L1 expression and VISTA, as well as a more aggressive phenotype (higher disease stage, higher tumoral grade, and higher mitosis). Like in other studies, PD-L1 expression confirmed a poorer prognostic significance, whereas VISTA positivity in immune cells was found to be protective [56].
The links between immune checkpoints, especially PD-L1 expression, the TME, and the clinical behavior of GISTs with different driver mutations are under investigation and still represent an intriguing research field.

Beyond the Tumor: The Immune-Modulating Effects of Imatinib
Along with the direct inhibition of oncogenic signaling, IM can modulate tumorinfiltrating immune cells, enhancing an immuno-stimulatory microenvironment through different mechanisms [15] (Figure 3).

Beyond the Tumor: The Immune-Modulating Effects of Imatinib
Along with the direct inhibition of oncogenic signaling, IM can modulate tumor-infiltrating immune cells, enhancing an immuno-stimulatory microenvironment through different mechanisms [15] (Figure 3). The immune-stimulating effect of IM is mediated by a reduction in the level of IDO, an immune checkpoint molecule able to inhibit T cells [60].
Through IDO blockade, IM increases the number of intratumor CD8+ T cells and reduces Tregs, leading to stimulation of the immune response against tumor cells [16,62].
c-KIT is also expressed on the surface of immune cells (i.e., mast cells), and its pathway, usually upregulated by activating driver mutation, plays an important role in the recruitment of innate immune cells (DCs, NK, CD8+, and CD4+ T cells) and the regulation of immune-suppressive cells (Tregs) [63].
Several recent lines of evidence suggest direct correlations between the TIL counts in cancer tissue, the ratio of CD8+ effector T cells to Tregs, and a favorable prognosis in various malignancies [64,65].
Inhibition of KIT signaling by IM is important for the downregulation of IDO. In fact, high levels of IDO and a low ratio of CD8+/Treg cells have been correlated to primary or acquired resistance to target agents [16,37]. The inhibition of c-KIT with TKIs alone or associated with specific antibodies can decrease the number of immune-suppressor cells and enhance antitumor immune function [60,61].
Based on this evidence, an open-label, multi-center, single-arm, phase 2 trial (NCT03291054), started in 2017, is testing the overall response rate (ORR) of patients with advanced GIST treated with epacadostat, an IDO inhibitor, in combination with pembrolizumab after failing on at least two TKI regimens. The immune-stimulating effect of IM is mediated by a reduction in the level of IDO, an immune checkpoint molecule able to inhibit T cells [60].
Through IDO blockade, IM increases the number of intratumor CD8+ T cells and reduces Tregs, leading to stimulation of the immune response against tumor cells [16,62].
c-KIT is also expressed on the surface of immune cells (i.e., mast cells), and its pathway, usually upregulated by activating driver mutation, plays an important role in the recruitment of innate immune cells (DCs, NK, CD8+, and CD4+ T cells) and the regulation of immune-suppressive cells (Tregs) [63].
Several recent lines of evidence suggest direct correlations between the TIL counts in cancer tissue, the ratio of CD8+ effector T cells to Tregs, and a favorable prognosis in various malignancies [64,65].
Inhibition of KIT signaling by IM is important for the downregulation of IDO. In fact, high levels of IDO and a low ratio of CD8+/Treg cells have been correlated to primary or acquired resistance to target agents [16,37]. The inhibition of c-KIT with TKIs alone or associated with specific antibodies can decrease the number of immune-suppressor cells and enhance antitumor immune function [60,61].
Based on this evidence, an open-label, multi-center, single-arm, phase 2 trial (NCT03291054), started in 2017, is testing the overall response rate (ORR) of patients with advanced GIST treated with epacadostat, an IDO inhibitor, in combination with pembrolizumab after failing on at least two TKI regimens.
IM additionally supports KIT-dependent crosstalk between DCs and NK cells, resulting in the production of immune-stimulating interferon-gamma (IFN-γ). Enhanced IFN-γ production by NK cells has been reported in patients affected by metastatic GIST after IM treatment. Patients with a high number of activated NK cells after IM therapy have a good prognosis [36] and can be identified as "immunologic responders" because of their better response to cytokine-based immunotherapy [37,66]. In a mouse model, Kats et al. proved an increasing production of INF-γ and a reduction in tumor size after treatment with anti-KIT CAR T cells [67].
Cytotoxic T-lymphocyte-associated antigen 4 (CTLA4) is an inhibitory immune-checkpoint receptor expressed on the surface of activated CD8+ T cells and on CD4+ T cells, which are implicated in the down-modulation of T helper and upregulation of Treg immune-suppressive activity [47]. Preclinical data have shown that CTLA-4 blockade in GIST-bearing mice can lead to the accumulation of CD8+T cells with enhanced INF-γ production [68].
This mechanism may explain the beneficial effect of combining IM with CTLA-4 blockade, observed by Balachandran et al. [60]. CTLA-4 blockade synergizes with TKIs in mouse models, leading to the study of this combination in humans.
In a clinical trial, all eight patients with a stage III/IV GIST treated with combined INF-α and IM achieved a complete response (CR) [69]. Before the administration of INF-α and IM, INF-γ was barely detectable; after 4 weeks of treatment, TILs increased in number and produced a high level of INF-γ [69].

The Clinical
Relevance of the Multifaceted Role of the Immune System: Immunotherapy for GIST Patients?
Immune checkpoint inhibitors (such as anti-CTLA4, anti-PD-1/PD-L1, anti-TIM-3, or anti-LAG3 antibodies) and IDO inhibitors could become a potential future strategy to improve the effects of targeted therapy in GIST [16,17]. In fact, despite the efficacy of tyrosine kinase inhibition, patients with metastatic GIST develop resistance to target therapy and tumor progression. To date, knowledge of the efficacy of immunotherapy in this setting is limited, and few clinical trials were designed (Table 1) during the last decade.

Combination Therapy with TKIs and ICIs
The study of immunotherapy in GISTs is evolving. In the randomized unblinded phase II trial NCT02880020, patients with advanced/metastatic refractory GIST were enrolled and randomized 1:1 to receive either nivolumab (240 mg Q2 wks) or nivolumab (240 mg Q2 wks) plus ipilimumab (1 mg/kg Q6 wks) for up to 2 years. In the nivolumab arm, 7/15 (46.7%) patients had stable disease (SD) with a median PFS of 8 weeks, while in the nivolumab + ipilimumab arm, 2/12 (16.7%) had SD, and the median PFS was 9 weeks [81].
In a phase I trial (NCT01738139), no responses were seen among 35 KIT-mutant GIST patients treated with the combination of ipilimumab (3 mg/kg) and IM (400 mg orally twice daily). Only one patient with a wild-type gastric GIST showed stable disease and continued to receive the treatment for 16 months [82]. In a phase I study (NCT01643278), 20 patients with unresectable/metastatic GIST were enrolled and received ipilimumab plus dasatinib. All patients had primary or secondary KIT resistance mutations or primary PDGFRA mutations. Most patients featured rapid disease progression according to the RECIST criteria, and few (7/13) had partial responses according to the Choi criteria, with a median PFS of less than 3 months. Only one patient, whose GIST harbored a PDGFRA exon 18 D842V mutation, remained in the trial for about 13.9 months [83].
The phase 1b/2 trial NCT03609424 is studying the efficacy of IM plus PDR001 (spartalizumab), an anti-PD-1 antibody, in advanced GIST after the failure of standard TKI therapies including IM, sunitinib, and regorafenib.
Two clinical trials are evaluating the efficacy of avelumab in GISTs. In REGOMUNE (NCT03475953), avelumab is administered in association with regorafenib in multiple solid tumors. The AXAGIST study (NCT04258956), a phase II single-arm trial, is testing the antitumor activity of avelumab in combination with axitinib in patients with unresectable/metastatic GIST after progression on second-or third-line treatment.
Hopefully, the results of these ongoing trials will soon provide new treatment frontiers.

Conclusions
Tumor mutational status is biologically and clinically important in many types of tumors [84] and has made GIST a paradigmatic model of oncogene addiction. GISTs are composed of many different genetic subtypes. Despite the relevance of mutational status, GISTs represent a heterogeneous genetic and clinical subgroup, showing variable clinical outcomes even in patients showing the same KIT or PDGFRA mutation.
The current research paradigm in oncology is shifting to the immune system and the TME [85], and recent literature data highlighted the potential role of the TME in GISTs as well, beyond the known impact of the mutational status. Therefore, deciphering the key activities of tumoral microenvironment components, and how they are influenced by each other, may be the answer to clinical heterogeneity in GIST, going further than the known paradigm model of oncogene addiction.