Keywords

Introduction

Gastrointestinal stromal tumors (GISTs) comprise one-fifth of soft tissue sarcomas, making them the most common sarcoma of the gastrointestinal tract [1]. The annual incidence of GIST is between 11 and 19.5 per million [2–5], and it has a prevalence of about 130 cases per million population [2–4]. For many years, GISTs were considered as smooth muscle sarcomas based on their morphology, and had been misdiagnosed as leiomyomas, leiomyosarcomas, or leiomyoblastomas. The prognosis of advanced GIST was very poor due to resistance to conventional chemotherapy and radiotherapy prior to the discovery of targeted therapies [6].

In the late 1990s, two groundbreaking discoveries had revolutionized the approach to diagnosis and treatment of GIST: (1) majority of GISTs (>95 %) were found immunohistochemically positive for the tyrosine kinase receptor KIT (also known as CD117) [7], and (2) KIT gene mutations were identified in 70–80 % of GISTs [8]. To date, KIT immunostaining and mutation screening are used as key diagnostic markers in clinical practice for GISTs, and mutant KIT is a clinically important therapeutic target in GISTs. The evolution of understanding the biology of GIST transformed it from a challenging chemotherapy-resistant disease to a model for molecular targeted therapy.

Although the initial events in GIST development are well characterized, the prognosis is clearly influenced by other genetic or epigenetic events that are still poorly understood. Aberrant microRNA expression is common in a wide range of human cancers. Accumulated evidence has shown that microRNAs are associated with clinical and pathological features in GIST, suggesting their important roles in GIST development.

This chapter gives a brief background on clinical features and biology of GIST, and provides an overview of the current knowledge on involvement of microRNAs in GIST tumorigenesis and therapeutic response.

Gastrointestinal Stromal Tumor

GISTs are thought to originate from the interstitial cells of Cajal (ICC) or their stem-like precursors [7, 8]. ICC function as pacemaker in the gastrointestinal tract that controls peristaltic contractions [7]. GISTs can be found anywhere along the gastrointestinal tract, but predominantly occur in the stomach (50–60 %) and the small intestine (30–35 %), less frequently in the colon/rectum (5 %) and esophagus (<1 %) [9]. These tumors can arise at any age, with a median age of diagnosis at 63 years [1, 9]. The tumor size varies between 2 and 30 cm at the time of diagnosis [10].

Oncogenic Mutations

The main initial event in GIST tumorigenesis is gain-of-function mutations in KIT (v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog) or PDGFRA (platelet-derived growth factor-α) genes. These genes are located on the long arm of chromosome 4 (4q12), and encode transmembrane proteins that belong to the type III tyrosine kinase receptor family.

Under normal physiological conditions, activation of KIT and PDGFRA receptors is controlled by spatial and temporal expression of their respective ligands, SCF and PDGF. Binding of these ligands to the receptors results in homodimerization, transphosphorylation of the tyrosine residues, and kinase activation that initiates signal transduction cascades promoting cell proliferation, growth, and survival [11–13]. About 75 % of GISTs harbor KIT mutations [14], whereas 10 % of GISTs harbor PDGFRA mutations [15, 16]. These mutations disrupt the autoregulatory mechanisms and cause ligand-independent constitutive activation of the encoded tyrosine kinase receptors [17], which results in aberrant cell growth and tumor formation [18]. Activation of KIT or PDGFRA stimulates several downstream signaling pathways such as mitogen-activated protein kinase (MAPK), phosphatidylinositol-3-kinase (PI3K)/AKT/mTOR and signal transducer, and activator of transcription 3 (STAT3) [19–21].

About 10–15 % of GISTs do not harbor KIT or PDGFRA mutations. These tumors display mutations in multiple cancer genes, including succinate dehydrogenase complex subunit genes (SDHA, SDHB, SDHC, and SDHD) (50 %) [22, 23], BRAF V600E substitution (13 %) [24], neurofibromin 1 (NF1) (7 %) [25, 26], and RAS family members [27]. Different signaling pathways in GIST are illustrated in Fig. 4.1.

Fig. 4.1
figure 1

Signaling pathways in GIST. (a) KIT and PDGFRA signaling pathways. Mutations in KIT or PDGFRA activate MAPK, PI3K/AKT/mTOR, and STAT3 pathways. The overall percentage of specific mutation sites is given in parentheses. (b) Signaling pathways in “wild-type” GISTs. Mutations in NF1, BRAF, or RAS lead to increased MAPK signaling. Mutations in one of the SDH genes (SDHA, SDHB, SDHC, or SDHD) lead to succinate accumulation, which inhibits prolyl hydroxylase-mediated HIF1α degradation and thereby increased HIF1α-mediated transcription of VEGF and IGF. P Phosphate group (Modified from Akcakaya P, thesis for doctoral degree 2015, ISBN 978-91-7549-730-3)

Full size image

Unlike adult GISTs, pediatric GISTs (1–2 % of all GISTs) are rarely positive for KIT or PDGFRA mutations, despite expressing KIT at similar levels as adult GISTs [28]. Gene expression pattern of these tumors is also different from adult GISTs [29, 30], suggesting alternative mechanisms of KIT activation or distinct pathways in pediatric GISTs.

Chromosomal Changes in GIST

Cytogenetic studies demonstrated that about 65 % of GISTs have either monosomy of chromosome 14 or partial loss of 14q [31–33]. Loss of heterozygosity and comparative genomic hybridization studies identified two hotspot regions (14q11.2 and 14q32), pointing tumor suppressor genes at these loci might be important for GIST development [32, 34]. Several candidate genes are suggested within these regions, such as PARP2, APEX1, and NDRG2 genes at 14q11.2, SIVA [35] and microRNA clusters at 14q32 [36, 37].

Several chromosomal abnormalities have been associated with malignant behavior in GIST. Loss of the long arm of chromosome 22 is observed in approximately 50 % of GISTs and associated with malignancy [31, 33, 38]. Chromosome 9p21 deletion causes inactivation of the tumor suppressor gene CDKN2A and associated with metastatic behavior [39–42]. Gains on chromosomes 8q (including MYC), 3q (including SMARCA3) and 17q are associated with metastasis [32, 43–45].

Treatment of GIST

Surgical resection is the main therapy for localized GIST, with the goal of complete resection and avoidance of tumor rupture [46]. However, surgery is sometimes not applicable for metastatic GISTs or clinically unresectable GISTs. A small molecule tyrosine kinase receptor inhibitor, such as imatinib mesylate, is used for the treatment of advanced GISTs.

Imatinib can selectively block the enzymatic activity of both transmembrane receptor tyrosine kinases KIT and PDGFRA [47, 48]. It competes with ATP for the ATP-binding pocket located in the kinase domain, and blocks the phosphorylation of the tyrosine kinase receptors. Binding of imatinib inhibits the activation of downstream survival pathways such as PI3K-mTOR and MAPK [19], and induces cell apoptosis through BIM [49] and soluble histone H2AX [50]. In addition, imatinib reduces the expression of indoleamine 2,3-dioxygenase (IDO) [51], which is an enzyme that produces immunosuppressive metabolites. Reduction of IDO causes depletion of regulatory T cells and increase of tumor-infiltrating CD8+ T cells. Thus, imatinib stimulates an anticancer immune response by diminishing IDO-mediated immunosuppression.

The majority of GIST patients with advanced disease get a clinical benefit from imatinib treatment. Imatinib achieved disease control in 70–85 % of patients with advanced GIST, median progression-free survival increased from 8–10 months to 20–24 months, and median overall survival increased from 18–20 months to 50 months [52–54]. However, resistance to imatinib is one of the biggest obstacles in current GIST clinical practice.

Approximately 10 % of patients progress within 6 months of initial therapy, which is defined as primary resistance to imatinib [53–56]. Primary resistance shows stronger correlation with certain tumor genotypes, such as wild-type KIT or PDGFRA, KIT exon 9 mutations and PDGFRA D842V mutation [57–60]. In addition, 50–60 % of the initially responding patients develop disease progression within 2 years, regarded as secondary or acquired resistance [53–56]. The main mechanism of acquired resistance is the acquisition of secondary mutations in the kinase or loop domain of KIT or PDGFRA [61]. Several alternative mechanisms of resistance have been described. Kinase switching is one of them and several kinases have been involved in such mechanism. AXL is an oncogenic tyrosine kinase receptor that regulates the same downstream signaling pathways as KIT. Kinase switching from KIT to AXL was observed in imatinib-resistant GIST cell lines and clinical samples [62]. Besides AXL, a switch from KIT to FAK and FYN activation has also been reported in GIST cells upon acquisition of imatinib resistance, and phosphorylated FAK inhibition can re-sensitize the resistant cells to imatinib-induced cell death [63]. FAK has also been implicated in growth and survival of imatinib-resistant GIST cells [64]. In addition, gene amplification of KIT or PDGFRA was shown as a potential mechanism leading to either primary or secondary resistance [65]. Moreover, microRNAs have also been shown to play a role in imatinib resistance in GIST, as described in the following section.

MicroRNA Deregulation in GIST

MicroRNA signature of GISTs was first described by Subramanian and colleagues in 2008 [66]. The study compared microRNA profiles of 27 sarcomas with different histological types, and demonstrated that GISTs were clearly distinguished from other sarcomas based on their microRNA expressions (Table 4.1). This distinction implicates the role of microRNAs in GIST tumorigenesis and their potential applications as diagnostic markers or therapeutic targets in GIST. Compared to other sarcoma types, miR-221–222 and miR-17–92 clusters were expressed at lower level in GIST [66, 67]. These microRNAs have been shown to target the two key factors KIT and ETV1 in GIST tumorigenesis [67] (Table 4.2), suggesting that lower expression of these microRNAs in GIST could be important for the pathogenesis of this tumor type. The current known microRNAs involved in regulating key genes in GIST development and progression are shown in Fig. 4.2.

Table 4.1 MicroRNA profiling studies in GIST
Full size table
Table 4.2 Examples of aberrantly expressed microRNAs with functional role in GIST
Full size table
Fig. 4.2
figure 2

MicroRNAs involved in the regulation of GIST development, progression, and imatinib response. In brief, miR-221, miR-222, miR-494, and miR-218 directly target the KIT expression, while miR-17 and miR-20a regulates the survival factor ETV1. miR-133b and miR-137 regulate GIST progression by targeting FSCN1 and TWIST1, respectively. miR-125a-5p regulates imatinib response through the regulation of PTPN18. miR-218 also regulates imatinib response. IM Imatinib, P Phosphate group

Full size image

MicroRNAs Associated with Clinical and Pathological Features in GIST

Morphology, clinical behavior, and molecular biology of GISTs differ according to their anatomical localization [68]. Likewise, microRNA expression profiles of GISTs located in stomach are distinct from the GISTs found in small intestine [36, 37]. Notably, different sets of microRNAs associated with anatomical location were observed in different studies. For example, Haller et al. showed that gastric GISTs presented higher expressions of miR-504, miR-7-1*, miR-598, and miR-24-1*, while the intestinal GISTs had higher levels of miR-220c, miR-229, miR-370, miR-210, miR-409-3p, miR-376a, and miR-376c [37]. Choi et al. demonstrated higher expressions of miR-383, miR-136, miR-146a, and miR-409a-3p, and lower expressions of miR-124a, miR-199b, miR-451, miR-663, miR-10a, and miR-218 in the intestinal compared to gastric GISTs [36]. The discrepancy is likely due to additional factors (e.g., risk grade and mutation status) that may contribute to differences besides anatomical locations in the tumors analyzed in both studies.

Several microRNA signatures have been described in GIST progression. In terms of tumor-risk group, a number of studies revealed distinct microRNA expression patterns between the high-risk and the low-risk GISTs, and identified a number of tumor-risk associated microRNAs (Table 4.1) [36, 69–71]. In the study of Choi et al., they compared microRNA profiles of 10 high-risk and 4 low-risk GISTs, and identified 28 microRNAs to be expressed at lower level in the high-risk group [36]. Yamamoto et al. reported 24 microRNAs with lower expression in the high-risk GISTs compared to low-to-intermediate risk tumors [70]. Kelly et al. found only miR-150 to be expressed at higher level in the low-risk tumors [71], and Niinuma et al. reported higher miR-196a expression in the high-risk group [69].

Besides tumor risk, several microRNAs are associated with tumor metastasis in GISTs. For example, low expression of miR-150-3p and high expressions of miR-301a-3p and miR-196a are associated with metastasis in GIST [69, 72]. In experimental cell culture systems, two microRNAs have been evaluated for their effect on tumor progression. Overexpression of miR-137 can inhibit cell migration and regulates epithelial-to-mesenchymal transition (EMT) by targeting TWIST1 [73], and inhibition of miR-196a can suppress cell invasion in GIST cells [69].

In terms of survival, low expression of miR-1915 is associated with disease-free and overall survival [72], while higher miR-196a expression is associated with poorer overall survival of GIST patients [69].

MicroRNAs Associated with Chromosomal and Genetic Alterations in GIST

As previously described, loss of 14q is common in GIST [31–33]. Downregulation of multiple microRNA clusters located at chromosome 14q (i.e., 14q32.31 and 14q32.33) has been reported in GISTs with 14q loss (Table 4.1) [36, 37]. One of the microRNAs located in this region, i.e., miR-494, was shown to directly target KIT and suppress its expression, and activates downstream signaling components such as AKT and STAT3 [74]. Functionally, inhibition of miR-494 suppresses proliferation and induces apoptosis in GIST cells [74].

Given that KIT and PDGFRA are key factors involved in GIST tumorigenesis, microRNA-mediated regulation of these factors is important for GIST development. As aforementioned, miR-494, miR-221, and miR-222 have been shown to directly regulate KIT expression in GIST cells [67, 74]. Recently, miR-218 was also found directly targeting KIT, and its overexpression suppresses proliferation and invasion, and induces apoptosis in GIST-T1 cells [75]. On the other hand, PDGFRA is known to be regulated by several microRNAs in different cell types, such as miR-126 in osteoblasts [76], miR-34a in gastric cancer [77], lung cancer [78], and glioma [79], and miR-146a/146b-5p in endothelial [80] and hematopoietic cells [81]; however, no microRNA has been experimentally validated to target PDGFRA in GIST.

GISTs show differential microRNA expression patterns according to their mutation status [37]. Several microRNAs are associated with KIT or PDGFRA-mutated GISTs. For example, miR-132, miR-766, miR-652, miR-629, miR-200c, miR-342-3p, miR-185, miR-146b-5p, and miR-150 levels are higher, whereas miR-330-3p is lower in PDGFRA-mutated GISTs as compared to KIT-mutated GISTs [37]. Higher expressions of miR-221 and miR-222 were found in the wild-type tumors compared to the tumors with KIT or PDGFRA mutation [37]. Concordantly, several studies have also revealed distinct mRNA expression profiles between GISTs with KIT and PDGFRA mutations [66, 82]. These findings suggest that, despite the common pathways activated by both mutations (e.g., PI3K/AKT and MAPK) [15], differences exist in the signal transduction networks between GISTs with KIT and PDGFRA mutations. In addition, several microRNAs are differentially expressed between GISTs with a single and double KIT mutations [72], suggesting that these microRNAs may be involved in partly distinct pathways [72].

Besides KIT and PDGFRA, several microRNAs are also associated with SDHB mutation. The SDHB-mutated GISTs show several microRNAs with higher (miR-132, miR-146a, miR-193b, miR-193b*, miR-455-3p, miR-455-5p, miR-484, and miR-886-5p) and lower (miR-125b, miR-450b, miR-488*, miR-542-3p, miR-551b, miR-576-3p and miR-769-5p) expressions compared to non-SDHB-mutated tumors [71].

MicroRNAs in Imatinib Resistance in GIST

MicroRNAs are known to play a role in tyrosine kinase inhibitor resistance [83–88]. The best example is the EGFR-inhibitor resistance in lung cancer. Numerous microRNAs (e.g., miR-205, miR-374a, miR-548b, miR-30b, miR-30c, miR-221, miR-222, and miR-200 family members) have been shown to regulate EGFR-inhibitor response in lung cancer [84, 85, 89, 90]. In chronic myelogenous leukemia (CML), miR-17–19b, miR-30e, miR-203, and miR-138 have been demonstrated to modulate imatinib sensitivity, while miR-30a promotes autophagy that enhances imatinib resistance [91–99].

In GIST, only two microRNAs have been functionally determined to modulate imatinib response [72, 100], despite a number of microRNAs are associated with imatinib resistance [72]. The expression of miR-218 is lower in imatinib-resistant compared to -sensitive GIST cell lines. Overexpression of miR-218 increases imatinib-induced cell death in the imatinib-resistant GIST430 cells. On the other hand, inhibition of miR-218 expression increases cell viability and decreases apoptosis in the imatinib-sensitive GIST882 cells upon imatinib treatment. Although no target gene(s) of miR-218 was identified, the authors propose that the effect might be mediated through PI3K/AKT signaling pathway.

The second microRNA is miR-125a-5p, which was found at higher expression levels in the imatinib-resistant than the -sensitive GISTs [72]. Overexpression of miR-125a-5p increases cell viability in the single KIT-mutated GIST882 cells upon imatinib treatment. However overexpression or suppression of miR-125a-5p in the double KIT-mutated GIST48 cells has no effect on imatinib response, suggesting that microRNA-mediated regulation is an alternative resistance mechanism to secondary KIT mutations in GIST. Ectopic expression of miR-125a-5p suppresses its target gene PTPN18 expression and silencing of PTPN18 increases cell viability in GIST882 cells upon imatinib treatment. The authors also observed an increased expression of miR-125a-5p and a decreased expression of PTPN18 in the imatinib-resistant subclone of GIST882 cells as compared to its sensitive counterpart, providing the functional evidence of miR-125a-5p-mediated regulation in imatinib resistance. PTPN18 is a member of the PEST domain containing protein-tyrosine phosphatase superfamily, which has been shown to dephosphorylate the phosphotyrosine residues of several tyrosine kinases, such as HER2 and SRC [101, 102]. Takahashi et al. recently demonstrated that altered phosphorylation of tyrosine kinases is an alternative mechanism of imatinib resistance in GIST [63]. Further studies have yet to determine whether the tyrosine kinases described by Takahashi et al. could be the substrate(s) of PTPN18.

Clinical Implications of microRNAs in GIST

MicroRNA expression profiles can distinguish GISTs from other sarcomas, and distinct microRNA expression signatures are associated with clinical, molecular, and histopathological features of GIST. These findings suggest a promising role for microRNAs as diagnostic and prognostic indicators in GIST.

Given their relatively higher stability in clinical samples and robust expression patterns, microRNAs have been suggested to have a greater utility as biomarkers in comparison to mRNAs [103]. Importantly, microRNAs can be released into the body fluids through microvesicles, which gives them a potential value as noninvasive biomarkers [104, 105]. Future studies evaluating the potential of circulating microRNAs as response markers for treatment or as reflective markers of GIST biological outcome would have a clinical benefit. However, there are some obstacles for circulating microRNAs, e.g., identification of an appropriate endogenous control and fluctuations in microRNA expression caused by diet, infection, treatment, trauma, or other factors [106].

Inhibition of KIT and PDGFRA by imatinib is the key therapeutic approach for advanced GISTs beside surgery. However, imatinib resistance is one of the biggest challenges in current GIST clinical practice. Post-transcriptional inhibition of oncogenes by microRNA mimics and activation of tumor suppressor genes by microRNA inhibitors are currently under investigation for their potential as therapeutic agents in cancer. KIT-targeting microRNA mimics (e.g., miR-221, miR-222, miR-494) [67, 74] may be used directly to target GIST cells to enhance the effect of imatinib for the purpose of overcoming resistance. Likewise, microRNA mimics/inhibitors for microRNAs specific to imatinib resistance, metastasis, risk grade or survival may be used for therapeutic purposes. Off-target effects and delivery of these molecules to specific GIST tissues/cells remain as the biggest challenges. Several strategies have been developed for delivery of microRNA-based therapeutics, including the use of nanoparticles, liposomes, antibodies and nucleic acid structure modifications [107].

Conclusion

In the last 20 years, growing knowledge of GIST molecular biology has revolutionized the clinical management of this disease, from a treatment-resistant uncontrolled disease to the development of targeted therapies. Despite tyrosine kinase inhibitors improve the outcome of the majority of patients, they fail to provide a permanent cure and resistant clones are observed in most of the initially responding tumors.

Development of alternative treatment strategies is needed in order to overcome resistance to ATP-competitive kinase inhibitors. Complete understanding of molecular biology in GIST development, progression, and treatment response is necessary to establish a ground for developing effective combinational therapies with a goal of not only to temporarily control the disease, but also to permanently eradicate all tumor cells.

MicroRNAs have been shown to play a role not only in the GIST tumorigenesis, but also in the stratification of patients at risk of developing the disease or therapy response. Although this research area is still relatively understudied, the work reported in the last 3 years is indicative of the excitement in this area. Ongoing and future studies will illuminate effectiveness and safeness of microRNAs as novel agents for GIST treatment and their predictive value as novel biomarkers. This will hopefully turn GIST from a model of targeted therapies that control the disease progression to a model of complete cancer cure.