Epigenetic landscape of gastrointestinal stromal tumors: Mechanistic insights and therapeutic opportunities

Abstract

Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms of the gastrointestinal tract and are primarily driven by activating mutations in KIT or PDGFRA. A clinically important subset of KIT/PDGFRA wild-type GISTs harbors alterations affecting succinate dehydrogenase subunit genes, which are associated with distinct biology and epigenetic profiles. The introduction of tyrosine kinase inhibitors has transformed their management, yet resistance and recurrence remain major challenges. Increasing evidence indicates that epigenetic processes contribute to tumor initiation, progression, and therapeutic escape. Aberrant promoter methylation has been linked to silencing of tumor suppressor genes, histone modifications reshape transcriptional networks governing proliferation and apoptosis, and chromatin remodeling complexes influence lineage-specific transcription and resistance pathways. Clinical observations further demonstrate that alterations such as SETD2 loss, KDM6A downregulation, or PHH3 overexpression correlate with prognosis, while early-phase trials of histone deacetylase inhibitors illustrate therapeutic feasibility. This review synthesizes current preclinical and clinical evidence on epigenetic regulation in GIST, focusing on DNA methylation, histone modifications, and chromatin remodeling, and explores their translational implications for prognosis and therapy.

Key Words: Gastrointestinal stromal tumor; Epigenetics; DNA methylation; Histone modifications; Chromatin remodeling; Therapeutic resistance; Biomarkers

Core Tip: Gastrointestinal stromal tumors (GISTs) are driven primarily by KIT or PDGFRA mutations, yet therapeutic resistance and heterogeneous clinical behavior remain major challenges. This review highlights how epigenetic mechanisms DNA methylation, histone modifications, and chromatin remodeling shape GIST initiation, progression, and treatment response. We synthesize preclinical and clinical evidence linking epigenetic alterations to prognosis, resistance, and emerging therapeutic vulnerabilities, and discuss how chromatin-directed strategies may complement tyrosine kinase inhibition to advance precision oncology in GIST.

INTRODUCTION

Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms of the gastrointestinal tract, driven in the majority of cases by activating mutations in KIT or PDGFRA[1]. These alterations represent the earliest oncogenic events in most sporadic GISTs and establish constitutive tyrosine kinase signaling as the central pathogenic and therapeutic axis[2]. A population-based cohort study of over 23000 patients found that the incidence of GISTs has risen significantly across most digestive organs over the past two decades, while revealing substantial survival disparities that negatively affect non-Hispanic Black and Asian or Pacific Islander populations[3]. The advent of tyrosine kinase inhibitors (TKIs), beginning with imatinib, revolutionized patient outcomes and remains the cornerstone of therapy[4,5]. Notably, response to imatinib varies according to the underlying KIT mutation. Tumors harboring exon 11 mutations exhibit the highest sensitivity, whereas exon 9-mutant GISTs demonstrate lower response rates and shorter progression-free survival (PFS) with standard-dose imatinib, often requiring dose escalation or alternative TKIs[6]. This genotype-dependent therapeutic gradient further highlights the need for molecular stratification beyond histopathology. Subsequent generations of TKIs, including sunitinib, regorafenib, ripretinib, and avapritinib, have further extended treatment options, especially in resistant or refractory disease, underscoring the value of precision medicine guided by mutational profiling[7]. Despite these advances, many patients ultimately relapse, reflecting the limits of mutation-directed therapy alone. Consequently, immunotherapy has emerged as a potential alternative for TKI-refractory cases; recent meta-analytical data suggests that immune-based interventions can offer modest but clinically significant benefits, with a clinical benefit rate of approximately 42% and a median overall survival exceeding 15 months in advanced populations[8].

Unlike the majority of adult GISTs, which are characterized by oncogenic KIT or PDGFRA mutations in approximately 80% of cases, pediatric GISTs represent a distinct molecular subgroup[9,10]. Instead, these pediatric tumors are often linked to defects in the succinate dehydrogenase (SDH) complex, which affects roughly half of pediatric cases and 20%-40% of all KIT/PDGFRA wild-type (WT) GISTs[9]. Loss of function in any SDH subunit causes intracellular succinate to accumulate; this excess succinate competitively inhibits prolyl hydroxylase domain proteins that normally mark hypoxia-inducible factor (HIF) 1-alpha (HIF-1α) for oxygen-dependent degradation, allowing HIF-1α to stabilize, enter the nucleus, and form a heterodimer with HIF-1β[11]. The resulting transcriptional activation drives a pseudohypoxic program, upregulating genes involved in angiogenesis and other tumor-promoting processes. Simultaneously, succinate structurally akin to α-ketoglutarate blocks α-ketoglutarate-dependent dioxygenases, including the ten-eleven translocation (TET) enzymes critical for active DNA demethylation, thereby inducing widespread DNA hypermethylation that profoundly alters gene expression and supports tumor initiation and progression[11]. Other rare WT GISTs fall into SDH-competent categories, including those activated by alterations in the RAS/RAF signaling pathway (such as mutations in NF1, BRAF, or KRAS) and quadruple WT GISTs (lacking defects in KIT/PDGFRA, SDH subunits, or RAS/RAF components)[12,13]. GISTs arising in the context of NF1 mutations represent a biologically distinct subgroup driven by RAS-mitogen-activated protein kinase (MAPK) activation rather than receptor tyrosine kinase signaling. As most NF1-mutant tumors lack KIT or PDGFRA alterations, they are typically resistant to imatinib and other standard TKIs[10]. Surgical resection therefore remains the cornerstone of management, while targeted medical options remain limited and investigational. Quadruple WT tumors can feature uncommon gene fusions or alterations involving NTRK, FGFR1, BRAF, ALK, or fibroblast growth factor (FGF)/FGF receptor (FGFR) pathways[13]. Clinically, BRAF-mutant GISTs represent a rare but actionable subgroup, and case series and small trials indicate sensitivity to BRAF inhibition [alone or combined with mitogen-activated protein (MEK) inhibitors], which is now considered an appropriate option for unresectable or metastatic disease lacking KIT/PDGFRA mutations[14]. Compared with conventional adult GISTs, these rare variants more commonly show epithelioid cell features, a strong preference for gastric sites (especially distal stomach/antrum in SDH-deficient cases), multifocal growth, and elevated risk of lymph node metastases (particularly in SDH-deficient tumors, where lymphovascular invasion occurs in about half of cases)[9]. A defining challenge is primary resistance to imatinib in most WT GISTs, which has spurred exploration of alternative approaches, including HIF-2α inhibitors like belzutifan[15], the alkylating agent temozolomide (leveraging O-6-methylguanine-DNA methyltransferase hypermethylation)[16], and third-generation TKIs such as olverembatinib (which has shown high clinical benefit rates in early SDH-deficient data)[17].

Parallel research has revealed that GIST biology extends well beyond kinase mutations. Increasing attention has turned to the tumor microenvironment, angiogenic pathways, and immune regulation, which collectively shape tumor growth, therapeutic response, and resistance[18,19]. Recent studies highlight that immune cell populations including tumor-associated macrophages (TAMs), cluster of differentiation (CD) 8⁺ T cells, and natural killer (NK) cells interact with oncogenic signaling and therapeutic pressure, sometimes driving resistance or modulating prognosis[18,20]. In GIST, TAMs often polarize toward an immunosuppressive M2 phenotype under therapeutic pressure (including imatinib), promoting progression, angiogenesis, and resistance by fostering a suppressive milieu and aiding metastatic behavior[21,22]. CD8⁺ T cells, as key cytotoxic players, correlate with improved survival and lower risk when abundant, yet face functional constraints from checkpoint activation and antigen presentation defects that reduce their impact on resistant disease[21,23]. NK cells infiltrate GISTs abundantly compared to many other solid tumors, especially in gastric and PDGFRA-driven cases, contributing to better survival outcomes via interferon-γ production and antimetastatic effects, yet their efficacy is commonly limited by immunosuppressive receptor profiles (e.g., NKp30C dominance) and microenvironmental inhibitory factors[23]. In addition, proliferative markers such as Ki-67 have emerged as clinically meaningful, with imaging-based radiomics now offering non-invasive approaches to assess tumor biology and guide treatment stratification[24]. In parallel with molecular classification, clinical risk stratification remains central to the management of primary GIST and is based on integrated evaluation of tumor size, mitotic index, anatomical location, and tumor rupture status[25]. Early systems such as the National Institutes of Health (NIH)-Fletcher criteria emphasized size and mitotic activity, whereas later models, including the AFIP and modified NIH (Joensuu) classifications, incorporated tumor site and rupture as major prognostic determinants[4]. Tumor location has a strong impact on clinical behavior: Gastric GISTs generally display a more favorable prognosis, while tumors arising in the small intestine, colon, or rectum carry a substantially higher risk of recurrence and metastasis at comparable size and mitotic rate. Beyond prognostic implications, anatomical site may also influence morphologic presentation. Rectal spindle cell GISTs, in particular, can exhibit architectural and cytologic features that overlap with those seen in gastric counterparts, including areas of stromal hyalinization, focal calcification, and palisading nuclear arrangements. In malignant cases, a leiomyosarcoma-like fascicular growth pattern may be observed, occasionally complicating histopathologic interpretation[4]. These morphologic nuances necessitate careful differential diagnosis from other spindle cell neoplasms of the rectum, such as leiomyosarcoma, schwannoma, desmoid-type fibromatosis, inflammatory myofibroblastic tumor, and metastatic spindle cell melanoma. In this setting, comprehensive immunophenotyping including KIT (CD117), DOG1, CD34, smooth muscle actin, desmin, S100, and SOX10 remains essential for accurate classification[2]. Tumor rupture is now recognized as an independent high-risk feature. These location-dependent differences likely reflect underlying biological heterogeneity, including distinct mutational and epigenetic profiles[26], supporting the integration of anatomical context with molecular and epigenetic biomarkers in contemporary risk assessment.

Epigenetic changes are increasingly recognized as important contributors to GIST pathogenesis, though their full significance is only beginning to be uncovered[27]. Epigenetic mechanisms including chromatin remodeling, histone modification, and DNA methylation regulate KIT transcription, drive resistance mechanisms, and influence tumor-immune interactions[28]. As the era of personalized oncology advances, these processes warrant systematic evaluation not only as prognostic and predictive biomarkers but also as potential therapeutic targets. While the role of non-coding RNAs in GIST has been analyzed extensively elsewhere, this review focuses on these three epigenetic mechanisms, which remain comparatively underexplored[29]. While mutational profiling has transformed the clinical management of GIST, it does not fully explain inter-patient heterogeneity in clinical behavior, therapeutic resistance, or long-term outcomes[7]. Epigenetic dysregulation provides a unifying framework linking oncogenic signaling, metabolic rewiring, immune escape, and drug tolerance, yet evidence in this field remains fragmented across preclinical and clinical disciplines. This review addresses this gap by systematically integrating mechanistic studies with translational and patient-based data, highlighting how DNA methylation, histone modifications, and chromatin remodeling converge to shape GIST biology.

DNA METHYLATION IN GISTS: MECHANISTIC INSIGHTS /PRECLINICAL EVIDENCE

Global methylation changes and tumor suppressor silencing

Epigenetic deregulation in GISTs was initially recognized in studies examining repetitive sequences, where methylation serves as a safeguard for genome stability. Igarashi et al[30] profiled over one hundred GISTs and demonstrated that hypomethylation of repetitive elements such as long interspersed nuclear element-1, satellite-α, and NBL2 was consistently associated with advanced clinical stage and poor histologic features. Using bisulfite pyrosequencing and array-comparative genomic hybridization, they linked hypomethylation to increased chromosomal instability, showing that epigenetic imbalance is directly coupled to structural genomic alterations. They provided an early mechanistic basis for how global methylation erosion accelerates GIST progression by enabling genome fragility and copy number variation. Targeted analyses of promoter-specific methylation further established its functional impact on GIST biology. Bure et al[31] explored the regulation of CD34, a classical diagnostic marker, and found that its loss in subsets of GISTs was attributable to promoter hypermethylation. Reactivation with 5-aza-2’-deoxycytidine (5-aza-dC) in GIST-T1 cells confirmed the causal role of methylation in silencing CD34. They challenged the notion that immunohistochemical markers are static and suggested that their variability reflects underlying epigenetic regulation. Similarly, Geddert et al[32] analyzed CD133, a stemness-associated surface antigen. They demonstrated that promoter methylation suppressed CD133 expression, which could be restored after exposure to demethylating agents. Importantly, CD133 expression correlated with unfavorable outcomes, suggesting that its methylation status may also have prognostic significance.

The direct therapeutic implications of methylation-mediated silencing were highlighted by Yang et al[33], who established a sunitinib-resistant GIST-T1 subline through long-term drug exposure. Molecular analysis revealed that resistance was mediated by promoter hypermethylation of PTEN, resulting in silencing of this critical tumor suppressor. PTEN loss released the inhibitory constraint on the phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin cascade, thereby sustaining survival signaling despite tyrosine kinase blockade. Re-expression of PTEN reversed resistance, demonstrating that promoter methylation is not merely a biomarker but a functional determinant of drug response. These findings established a paradigm in which methylation contributes both to inherent tumor aggressiveness and to the development of acquired resistance under therapeutic pressure.

Developmental pathways and stemness regulation

As GISTs originate from interstitial cells of Cajal (ICCs), which serve as pacemaker cells of the gastrointestinal tract, the epigenetic control of developmental programs is highly relevant. Faucz et al[34] examined the role of FOXD3, a transcription factor essential for repressing KIT signaling during embryogenesis. Using zebrafish and murine models, they showed that FOXD3 deficiency led to hyperplasia of ICCs and upregulation of KIT expression, phenocopying the initiating events of GIST. In vitro silencing of FOXD3 similarly increased KIT expression in cultured cells, confirming a direct link. These findings position FOXD3 as a developmental regulator whose loss, possibly through epigenetic inactivation, promotes the expansion of ICC progenitors susceptible to transformation. Chromatin regulators further integrate into this process by linking histone modification to DNA methylation. Miao et al[35] dissected the role of EZH2, the catalytic component of polycomb repressive complex 2 (PCR2), in controlling the transcription factor PAX8. They demonstrated that EZH2-mediated H3K27me3 deposition cooperated with DNA methyltransferases to hypermethylate the PAX8 promoter, leading to its silencing. The loss of PAX8 suppressed Wnt4 transcription, impairing differentiation cues and enhancing proliferation. In GIST-T1 cells, EZH2 inhibition restored PAX8 and Wnt4 expression, while xenograft models confirmed reduced tumor growth when EZH2 was pharmacologically targeted. This mechanistic work illustrated how polycomb-mediated repression intersects with DNA methylation to reprogram developmental pathways in GIST[35]. The connection between methylation and stemness was further exemplified by Gromova et al[36], who studied endoglin (ENG/CD105), a co-receptor for transforming growth factor-β implicated in angiogenesis and self-renewal. In Kit K641E mouse models and KIT-mutant cell lines, ENG expression was upregulated, and treatment with 5-aza-dC further enhanced this effect, indicating hypomethylation-mediated control. ENG upregulation was associated with both vascular remodeling and stem-like features, suggesting that methylation deregulation can potentiate tumor aggressiveness not only by driving proliferation but also by modifying the tumor microenvironment[36].

Taken together, these studies reveal how methylation orchestrates the balance between differentiation and self-renewal in GIST. By silencing transcription factors such as FOXD3 and PAX8, and modulating stemness regulators like CD105, epigenetic changes promote a dedifferentiated state that is permissive for tumor initiation and progression.

Non-coding RNA-associated epigenetic regulation and epitranscriptomics

Epigenetic control in GIST extends beyond protein-coding genes to encompass non-coding RNAs, which act as fine regulators of oncogenic signaling. Isosaka et al[37] identified miR-34a and miR-335 as tumor-suppressive microRNAs silenced through promoter hypermethylation in GIST-T1 cells. Restoring their expression with 5-aza-dC significantly reduced proliferation, migration, and invasion, indicating a functional role in restraining malignancy. Importantly, miR-34a directly targeted PDGFRA, linking its silencing to unchecked receptor tyrosine kinase activity[37]. Kelly et al[38] examined the 14q32 microRNA cluster, which is regulated by imprinting at the MEG3 differentially methylated region. They found that pediatric and WT GISTs displayed distinct methylation patterns at this locus, resulting in widespread dysregulation of microRNA expression. Functional studies showed that restoring microRNAs from this cluster reduced cell growth and migration, suggesting that imprinting-related methylation defects drive subtype-specific tumor behavior.

Recent investigations underscore the regulatory influence of non-coding RNAs and RNA modifications in GIST biology, particularly in driving progression, subtype-specific behavior, and resistance to TKIs[39]. Circular RNAs (circRNAs) such as those overexpressed in resistant or high-risk tumors contribute to oncogenic pathways by sponging microRNAs or stabilizing messenger RNA (mRNA), correlating with features like relapse, metastasis, and reduced survival in certain molecular subgroups[10]. MicroRNA signatures vary across GIST subtypes (e.g., SDH-deficient vs KIT/PDGFRA-driven), potentially serving as diagnostic tools or predictors of therapeutic response when combined with genomic profiling. Epitranscriptomic alterations, including dysregulated N6-methyladenosine (m6A) methylation, promote tumor growth and KIT-dependent resistance through changes in RNA stability and translation efficiency, offering novel targets for overcoming treatment failure in advanced disease[10].

At the level of long non-coding RNAs (lncRNAs), Lee et al[40] demonstrated that HOTAIR recruits PCR2 to the promoter of PCDH10, inducing DNA hypermethylation and transcriptional silencing. The loss of PCDH10, a cell adhesion molecule, facilitated invasion and metastatic traits in GIST-T1 and GIST882 cells. This finding established that lncRNAs can act as epigenetic scaffolds, guiding methylation machinery to silence tumor suppressors. Complementing this, Niinuma et al[41] found that the lncRNA MEG3 was commonly hypermethylated in GIST tissues. Treatment with a DNA methyltransferase inhibitor (DNMTi) combined with a histone deacetylase (HDAC) inhibitor reactivated MEG3 expression and induced interferon responses via a viral mimicry pathway. These results suggest that reactivation of silenced lncRNAs may have dual benefits: Restoring tumor suppressor function and engaging innate immunity.

Expanding the scope of methylation beyond DNA, Xu et al[42] investigated RNA methylation as an epitranscriptomic regulator of resistance. They identified that the RNA methyltransferase METTL3 increased m6A deposition on multidrug resistance protein 1 (MRP1) transcripts, stabilizing the mRNA and enhancing protein expression. Elevated MRP1 levels promoted drug efflux, thereby driving imatinib resistance. Resistant GIST-T1 and GIST-882 models, as well as xenografts, validated this mechanism, and inhibition of METTL3 restored drug sensitivity. Analysis of patient samples revealed that high METTL3 expression correlated with shorter PFS, demonstrating a clinically relevant link. This study broadened the concept of epigenetic regulation in GIST to include epitranscriptomic modification of RNA, with direct consequences for therapy resistance. As illustrated in Figure 1, aberrant promoter CpG hypermethylation recruits methyl-CpG-binding proteins and HDAC-containing corepressor complexes, leading to repressive histone marks (e.g., H3K27me3) and stable transcriptional silencing of tumor suppressor genes, while global hypomethylation in high-risk or SDH-deficient GISTs contributes to oncogene activation and genomic instability.

Figure 1 Epigenetic mechanisms in gastrointestinal stromal tumors. DNA methylation and histone modifications disrupt tumor suppressor genes, developmental regulators, and stemness pathways, thereby promoting genomic instability, uncontrolled proliferation, and therapeutic resistance in gastrointestinal stromal tumors. GIST: Gastrointestinal stromal tumors; LINE-1: Long interspersed nuclear element-1; CD: Cluster of differentiation; PI3K: Phosphatidylinositol 3-kinase; AKT: Protein kinase B; mTOR: Mammalian target of rapamycin; DNMT: DNA methyltransferase; ICC: Interstitial cells of Cajal; m6A: N6-methyladenosine; ncRNA: Non-coding RNA; DMR: Differentially methylated region; miRNA: MicroRNA; MRP: Multidrug resistance protein 1; IFN: Interferon.

HISTONE MODIFICATIONS IN GISTS: PRECLINICAL INSIGHTS

Histone acetylation and HDAC inhibition

Histone acetylation regulates chromatin accessibility, and its disruption in GISTs directly impacts KIT transcription and protein stability[43]. HDAC inhibitors were among the first chromatin-targeting compounds tested in this setting. Treatment of GIST cells with vorinostat (SAHA), panobinostat (LBH589), valproic acid, trichostatin A, or sodium butyrate produced dose-dependent suppression of KIT mRNA and protein, accompanied by accumulation of acetylated histone H3 and H4 within the nucleus[44]. Importantly, these inhibitors acetylated HSP90, an essential KIT chaperone, resulting in its functional inactivation and proteasomal degradation of KIT, highlighting that HDAC inhibition operates both at the level of transcriptional repression and post-translational destabilization[44]. This transcriptional repression aligns with disruption of lineage-specific factors like ETV1, which cooperates with CREB binding protein (CBP)/p300 acetyltransferases to maintain KIT super-enhancer activity in GIST (as detailed in section “chromatin remodeling and transcriptional control”). When combined with Notch1 activation, SAHA produced even stronger reductions in GIST cell viability, showing that histone deacetylation intersects with signaling networks beyond KIT itself[45]. In vivo studies confirmed these mechanistic links. Administration of panobinostat to GIST xenografts caused pronounced tumor regression associated with widespread acetylation of histone H3 and H4, demonstrating that pharmacologic acetylation shifts can be achieved in the tumor microenvironment[46]. Moreover, the combination of panobinostat with imatinib produced additive antitumor effects, supporting the hypothesis that chromatin modulation can resensitize tumors to tyrosine kinase inhibition.

Proteasome inhibition provided another unexpected entry point into histone regulation. In early studies, bortezomib was shown to induce accumulation of soluble, non-nucleosomal histone H2AX in GIST cells, a process that acted as a cytotoxic signal triggering apoptosis while simultaneously repressing KIT transcription[47]. This “soluble H2AX stress” mechanism was later confirmed and broadened with second-generation proteasome inhibitors. Rausch et al[48] demonstrated that agents such as carfilzomib, ixazomib, and delanzomib reproduced the same fundamental effect but with improved pharmacologic properties. These compounds destabilized KIT at the transcriptional level, triggered H2AX upregulation, and induced apoptosis in imatinib-sensitive as well as imatinib-resistant GIST models. Importantly, the apoptotic effect was tightly coupled to the accumulation of soluble H2AX rather than DNA damage signaling, suggesting that histone release functions as a non-canonical apoptotic trigger. The study also highlighted that proteasome inhibitors exerted dual pressure on tumor cells by simultaneously impairing KIT expression and overwhelming protein homeostasis, thereby enhancing stress signaling cascades[48]. Interestingly, soluble H2AX accumulation was not restricted to proteasome inhibition. Imatinib treatment of GIST cells also increased soluble H2AX, pointing to a convergent mechanism in which disruption of KIT activity or proteostasis culminates in chromatin destabilization[49]. This shared endpoint reinforces the concept that histone variants act as molecular sensors of oncogene dependence and stress, bridging targeted kinase inhibition and proteasome blockade under a unifying chromatin-based apoptotic program.

At the proteomic scale, acetylation landscapes in GIST were recently interrogated to understand how acetyl marks extend beyond histones to a wider regulatory network. Acetylome profiling of human GIST tissue identified nearly 3000 acetylation sites across 1319 proteins, with marked differences between low-risk and high-risk tumors[50]. Sites such as lysine 1063 on Ki67 and lysine 24 on FCHSD2 displayed strong stratification potential, suggesting that acetylation events could serve as biomarkers of proliferative capacity or altered signaling[50]. Collectively, these studies underscore that histone acetylation is not only a regulator of KIT-driven transcription but also a therapeutic vulnerability that can be manipulated to enhance apoptosis, resensitize resistant tumors, and potentially provide novel biomarkers.

Histone methylation and demethylation dynamics

Histone methylation, unlike acetylation, can either activate or repress transcription depending on the residue and degree of methylation. In GISTs, lysine demethylases and methyltransferases have emerged as key modulators of tumor biology. KDM4D, a demethylase that removes H3K9me3 and H3K36me3, promotes angiogenesis through direct activation of HIF1β transcription[51]. Knockdown of KDM4D suppressed endothelial recruitment and reduced tumor vascularization in xenograft models, demonstrating the role of histone demethylation in shaping the tumor microenvironment. This pathway was further linked to resistance, as upregulation of KDM4D supported imatinib-insensitive growth. Targeting this axis with miR-409-5p inhibited angiogenesis, blocked invasion, and restored sensitivity to imatinib, highlighting a tractable therapeutic link between histone demethylation and drug resistance[52].

Beyond KDM4D, inhibition of KDM4A has been shown to induce apoptosis and block proliferation across cancer models, with the compound LDD2269 providing proof of concept that demethylase inhibition is druggable[53]. Another regulator, KDM6A, functions as a tumor suppressor by erasing the repressive H3K27me3 mark at the SPARCL1 promoter. Reactivation of SPARCL1 expression reduced proliferation, invasion, and angiogenesis in vitro, while xenografts with restored KDM6A function displayed decreased liver metastasis. In patient tumors, downregulation of both KDM6A and SPARCL1 was consistently observed, suggesting clinical relevance of this pathway[54]. Among histone-modifying enzymes, SMYD2 has emerged as a critical regulator of GIST progression. SMYD2 catalyzes lysine methylation of EZH2 at residue K307, stabilizing EZH2 protein levels and reinforcing PRC2 activity[55]. The increased enzymatic activity of EZH2 leads to heightened deposition of the repressive H3K27me3 mark at the TET1 promoter, resulting in transcriptional silencing of TET1. Loss of TET1 function impairs DNA demethylation processes and diminishes activation of downstream p53-dependent apoptotic pathways, thereby promoting unchecked proliferation and survival of GIST cells[55]. This axis illustrates how histone methyltransferase activity can simultaneously stabilize oncogenic chromatin regulators and suppress tumor-suppressive demethylases, creating a feed-forward loop of epigenetic repression.

Genomic studies have also identified recurrent mutations in SETD2, the sole enzyme responsible for H3K36 trimethylation. Loss of SETD2 function reduces H3K36me3 across the genome, producing widespread alterations in chromatin architecture, transcriptional dysregulation, and impaired DNA repair. Clinically, SETD2 mutations are enriched in high-risk GISTs and associate with poor prognosis, positioning H3K36me3 deficiency as a hallmark of aggressive disease[56]. These findings highlight that histone methylation is a dynamic determinant of angiogenesis, proliferation, and risk, offering both mechanistic insight and potential therapeutic targets. As presented above, Figure 1 depicts how repressive histone modifications (e.g., H3K27me3 enrichment) reinforce DNA methylation at silenced promoters, while loss-of-function in histone-modifying enzymes (e.g., SETD2 or KDM6A) disrupts this balance, promoting oncogenic transcriptional programs and poor prognosis in GIST.

Chromatin remodeling and transcriptional control

Epigenetic remodeling in GIST extends beyond acetylation and methylation to encompass chromatin regulators and their impact on transcriptional networks. The CBP/p300 acetyltransferases emerged as pivotal co-activators maintaining oncogenic transcription[57]. These histone acetyltransferases interact with lineage-specific transcription factors such as ETV1, enhancing acetylation at KIT super-enhancers and thereby stabilizing KIT-ETV1-MAPK signaling. Pharmacologic blockade of CBP/p300 with the small-molecule inhibitor C646 caused global reductions in histone H3 acetylation at promoters of KIT-regulated genes, suppressed KIT and ETV1 transcription, and led to apoptosis in GIST cells, establishing CBP/p300 as critical sustainers of the oncogenic epigenetic circuitry[57]. This ETV1-CBP/p300 interaction sustains KIT expression through histone acetylation, consistent with HDAC inhibitor-mediated KIT suppression observed in acetylation studies (section “histone acetylation and HDAC inhibition”). Polycomb repression also intersects with GIST biology. The chromobox protein CBX7, a core PRC1 component, binds H3K27me3 and facilitates chromatin compaction[58]. Functional antagonism of CBX7 reversed this repression, reactivated the tumor suppressor locus p16INK4a, and curtailed proliferation, underscoring that polycomb-mediated histone recognition silences growth-regulatory networks. The findings point to polycomb proteins as modulators of the balance between oncogenic persistence and tumor suppressor restraint[58].

KIT inhibition itself was shown to provoke rapid epigenomic remodeling. In imatinib-sensitive GIST models, pharmacologic blockade of KIT altered patterns of histone H3 acetylation and phosphorylation within hours[59]. These shifts included decreased H3 acetylation at KIT target genes and altered phosphorylation at serine and threonine residues that regulate chromatin compaction. Such changes provide a direct mechanistic link between kinase signaling and histone states, revealing that tyrosine kinase activity dynamically governs chromatin structure and accessibility[59]. Histone phosphorylation also acts as a mitotic driver. Jin and colleagues demonstrated that PHH3, catalyzed by Aurora B kinase during mitosis, is markedly elevated in GIST cells. This modification facilitates chromatin condensation and ensures faithful chromosome segregation. Elevated PHH3 not only reflects high mitotic activity but also reinforces proliferative drive by coupling kinase signaling to chromatin mechanics. The study highlighted PHH3 as a functional mediator of cell-cycle progression and as a potential therapeutic target for proliferative blockade[60].

Finally, transcription factor stability is intimately linked to chromatin regulation. Using a connectivity map-based strategy, phenothiazines were identified as ETV1-targeting agents in GIST. Phenothiazine exposure destabilized ETV1 protein, disrupted KIT transcriptional reinforcement, and induced both apoptosis and autophagy[61]. Mechanistically, treatment activated extracellular regulated protein kinases (ERK) signaling and downstream ELK1 and EGR1 transcription factors, which promoted autophagy. Combination therapy with MEK inhibitors abrogated this adaptive autophagic response, suppressed ERK-ELK1-EGR1 signaling, and enhanced apoptosis. This work defined ETV1 destabilization as a novel therapeutic approach and illustrated how pharmacologic disruption of transcription factor chromatin interactions can suppress GIST survival[61]. Together, these findings expand the landscape of epigenetic control in GIST beyond acetylation and methylation, revealing a multilayered interplay of co-activators, polycomb repressors, histone phosphorylation, and transcription factor stability.

Epigenetic crosstalk, resistance, and therapeutic synergies

Epigenetic regulators in GIST do not act in isolation but converge with signaling pathways and drug resistance mechanisms. A striking example is the induction of BCL6 following imatinib exposure. Zeng et al[62] demonstrated that BCL6 directly recruits the NAD-dependent deacetylase SIRT1 to the TP53 promoter, producing localized histone deacetylation and transcriptional repression of p53. This chromatin remodeling silences p53-dependent apoptosis, thereby contributing to therapeutic resistance. Treatment with BI-3802 led to BCL6 degradation, dismantling the BCL6-SIRT1 complex, reinstating acetylation at the TP53 promoter, and reactivating p53 signaling. In xenograft models, this intervention synergized with imatinib and suppressed tumor growth, providing proof-of-concept that drug resistance in GIST can be mediated by histone-directed transcriptional repression and can be reversed with targeted epigenetic therapy[62]. Combination strategies further highlight the therapeutic potential of epigenetic crosstalk. Niinuma et al[41] showed that concurrent treatment of GIST cells with DNMTi and HDAC inhibitors induced pronounced epigenomic remodeling, most notably a gain of H3K4me3 at promoter regions. This activating histone mark triggered a “viral mimicry” program, characterized by expression of double-stranded RNA sensors and robust interferon signaling. The downstream immune activation potentiated tumor cell killing beyond the effects of either agent alone, demonstrating how combined epigenetic therapies can reprogram both intrinsic transcriptional states and extrinsic immune surveillance.

Growth factor signaling also intersects with chromatin modifications to regulate oncogenic pathways. Hayashi et al[63] reported that activation of the insulin-like growth factor-1 (IGF1) pathway inhibited glycogen synthase kinase-3β, which in turn remodeled the KITLG promoter chromatin landscape. This remodeling was marked by increased acetylation of histones H3 and H4, enrichment of the activating H3K4me3 mark, and reduction of repressive marks such as H3K9me3 and H3K27me3. Collectively, these changes enhanced KITLG transcription, reinforcing an autocrine/paracrine loop with KIT signaling and fueling GIST proliferation. Pharmacologic interference with either chromatin modifiers or IGF1 receptor disrupted this feed-forward mechanism, highlighting how membrane-to-nucleus signaling cascades impose chromatin-level control over oncogenic transcription. Finally, metabolic dysfunction also connects directly to histone regulation. Gao et al[64] uncovered that succinate dehydrogenase subunit B (SDHB) loss in GIST results in succinate accumulation, which blocks α-ketoglutarate dependent dioxygenases and stabilizes the transcription factor ZNF148. ERK-mediated phosphorylation of ZNF148 promoted its interaction with FOXM1 at the SNAIL promoter. This complex enhanced histone H3 acetylation at the promoter, driving SNAIL transcription and epithelial-mesenchymal transition (EMT). Mechanistically, in addition to promoter acetylation at SNAIL, SDHB loss was associated with increased H3K36me2 accumulation at the ZNF148 promoter, consistent with succinate-mediated inhibition of α-ketoglutarate-dependent histone demethylation, which supports transcriptional upregulation of ZNF148. ERK-driven phosphorylation of ZNF148 at Ser306 was required for stable interaction with FOXM1 and efficient SNAIL transcription, and recruitment of the histone acetyltransferase p300 emerged as a key step enabling promoter-associated histone acetylation in this axis. In patient samples, SDHB-negative tumors more frequently exhibited high ZNF148 expression and increased p-ZNF148 (Ser306), and elevated p-ZNF148 levels were associated with shorter PFS, further linking this pathway to clinically aggressive behavior. Together, these data connect mitochondrial dysfunction to a defined chromatin-dependent EMT program in SDH-deficient GIST[64]. SDH-deficient GISTs account for approximately 8% of all cases and comprise a biologically distinct subgroup that includes most pediatric GISTs, a subset of sporadic adult tumors, and syndromic forms associated with Carney triad and Carney-Stratakis syndrome. The above is illustrated in Figure 2. Together, these findings illustrate how histone modifications in GIST act as points of convergence for oncogenic signaling, metabolic derangements, and drug resistance.

Figure 2 Histone modifications in gastrointestinal stromal tumors. Histone acetylation (A) regulates KIT expression, with histone deacetylase (HDAC) inhibitors increasing acetylation, suppressing KIT transcription, destabilizing KIT protein, and restoring sensitivity to imatinib. Histone methylation (B) influences tumor behavior through KDM4D-driven angiogenesis, KDM6A loss promoting metastasis, SMYD2-EZH2-TET1 repression reducing apoptosis, and SETD2 deficiency predicting poor prognosis. Chromatin remodeling (C) sustains oncogenic transcription via CREB binding protein/p300 and ETV1, silences tumor suppressors through CBX7, drives mitosis via PHH3 phosphorylation, and enables therapeutic targeting of ETV1 with phenothiazines. Crosstalk and resistance (D) arise from imatinib-induced BCL6-SIRT1-p53 repression, reversed by BI-3802, synergistic DNA methyltransferase inhibitor + HDAC inhibitors mediated interferon activation, insulin-like growth factor-1-driven KITLG upregulation, and succinate dehydrogenase subunit B loss linking metabolism to epithelial-mesenchymal transition through ZNF148. HDAC: Histone deacetylase; GIST: Gastrointestinal stromal tumors; CBP: CREB binding protein; HIF: Hypoxia-inducible factor; IGF: Insulin-like growth factor; DNMTi: DNA methyltransferase inhibitor; IFN: Interferon.

CHROMATIN REMODELING IN GIST: PRECLINICAL EVIDENCE

Disruption of chromatin remodeling machinery represents a recurrent pathogenic mechanism in GIST. Ren et al[58] demonstrated that ARID1A, a core component of the switch/sucrose non fermentable (SWI/SNF) complex, is frequently inactivated in GIST, with knockdown models showing enhanced cell proliferation and impaired differentiation. Importantly, ARID1A loss promoted MYC-driven transcriptional activity and sensitized GIST cells to BET inhibition, highlighting a synthetic-lethal interaction between SWI/SNF deficiency and bromodomain proteins[58]. Additional remodeling-related pathways have also been implicated. Rausch et al[48] uncovered that alterations in the chromatin remodelers ATRX and DAXX are present in a subset of GIST, linking them to telomere biology and nuclear architecture. Normally, ATRX and DAXX cooperate to deposit the histone variant H3.3 at telomeric and pericentromeric chromatin, preserving nucleosome stability and preventing aberrant recombination. In GIST, loss-of-function mutations in these genes disrupted this process and triggered the alternative lengthening of telomeres (ALT) pathway, which relies on homologous recombination to sustain telomere maintenance in the absence of telomerase. ALT-positive tumors demonstrated hallmarks such as heterogeneous telomere length, ALT-associated promyelocytic leukemia bodies, and elevated TERRA RNA expression[48]. Beyond telomere regulation, ATRX/DAXX deficiency induced broader changes in chromatin compaction, derepressing repeat elements and altering transcriptional programs linked to genomic instability. These findings suggest that ATRX/DAXX-mutant GISTs form a biologically distinct subgroup with vulnerabilities to DNA damage response inhibitors and chromatin-targeting therapies. Hemming et al[65] investigated the contribution of polycomb group proteins to transcriptional repression in GIST, focusing on CBX7, a chromodomain-containing component of canonical PRC1. CBX7 recognizes trimethylated H3K27me3 placed by PRC2, thereby anchoring PRC1 to chromatin and facilitating further gene silencing via H2A ubiquitination. In GIST models, CBX7 occupancy was enriched at the CDKN2A locus, encoding p16 INK4a, a critical tumor suppressor regulating cell-cycle arrest. Pharmacological disruption of the CBX7 chromodomain using selective inhibitors (MS452 and MS351) displaced PRC1 from chromatin, reduced H2A ubiquitination, and reactivated CDKN2A transcription. This epigenetic de-repression translated into increased p16 INK4a protein levels, cell-cycle blockade, and reduced tumor growth in vitro[65]. Finally, Zeng et al[62] provided evidence that imatinib induces epigenetic rewiring through BCL6-mediated recruitment of the HDAC SIRT1 to the TP53 promoter, causing histone deacetylation, transcriptional silencing of TP53, and impaired apoptosis. Targeting BCL6 restored p53 activity and overcame resistance in xenograft models. Together, these preclinical insights establish chromatin remodeling as a central driver of GIST biology and a fertile area for therapeutic exploitation. As shown in Figure 2, chromatin remodeling complexes regulate nucleosome accessibility, facilitating the recruitment of DNA methyltransferases and histone-modifying enzymes to establish repressive states, and their dysregulation in resistant GISTs contributes to therapy escape, immune evasion, and altered transcriptional networks.

In conclusion, crosstalk between epigenetic layers is a central determinant of transcriptional regulation in GIST rather than the result of three independent processes[66]. DNA methylation and histone modifications frequently reinforce one another: Promoter CpG methylation can recruit methyl-CpG binding proteins and associated corepressor complexes containing HDACs, consolidating transcriptionally repressive chromatin states, while repressive histone marks may in turn facilitate stable DNA methylation at silenced loci[67]. Chromatin-remodeling complexes add a further regulatory layer by controlling the accessibility of DNA methyltransferases and histone-modifying enzymes to nucleosomal DNA, thereby shaping the establishment and maintenance of “locked” epigenetic programs[68]. In GIST, this integration is biologically relevant in at least two contexts highlighted in this review: (1) SDH-deficient tumors, in which succinate-mediated inhibition of α-ketoglutarate dependent enzymes simultaneously disrupts DNA demethylation and histone demethylation capacity, driving coordinated epigenomic reprogramming[39]; and (2) Therapy adaptation, where combined epigenetic targeting (e.g., DNMT and HDAC inhibition) can modulate immune-related transcriptional programs and drug sensitivity, underscoring that resistance phenotypes often arise from multi-layer epigenetic cooperation rather than isolated molecular alterations[10].

THERAPEUTIC AND TRANSLATIONAL LANDSCAPE

Clinical applications of DNA methylation in GIST

Methylation-defined clinical subsets: SDH-deficient tumors represent a unique clinicopathologic class characterized by a CpG island hypermethylator phenotype, first recognized through genome-wide array profiling. These studies demonstrated that SDH-deficient GISTs harbor many times more hypermethylated loci compared with KIT- or PDGFRA-mutant tumors[69]. Mechanistically, succinate accumulation within SDH-deficient cells competitively inhibits TET-family dioxygenases, impairing active DNA demethylation and resulting in a stably hypermethylated epigenome. This phenomenon illustrates how a metabolic lesion directly shapes the epigenetic landscape[70]. Within this context, promoter hypermethylation of SDHC has emerged as a recurrent and specific alteration in SDHx-WT GIST. Large multi-institutional series demonstrated its presence in more than 90% of such tumors, including many associated with Carney triad, and detected mosaic epimutations in peripheral tissues, suggesting a post-zygotic origin[71]. This observation has immediate clinical relevance: In SDHB-immunonegative, mutation-negative GISTs, SDHC methylation testing serves as a definitive diagnostic tool, delineating an epigenetic nosology distinct from germline SDHx syndromes. Clinically, SDH-deficient gastric GISTs typically arise in children and young adults, often manifest as multifocal lesions with a plexiform growth pattern, and display a tendency for lymph-node involvement. Their natural history is unpredictable, with a potential for late recurrences, and long-term surveillance is mandatory. Immunohistochemical loss of SDHB expression remains the practical first-line screen, guiding subsequent methylation analysis for diagnostic confirmation[72].

Methylation profiling has also revealed clinically relevant distinctions within KIT-mutant disease. Among exon 11 mutants, tumors carrying the Δ557-558 deletion exhibit a distinctive epigenomic signature marked by intergenic hypomethylation, recurrent copy number losses affecting the CDKN2A locus, enrichment of structural variants, and transcriptomic evidence of p53 pathway impairment. These molecular correlates align with the aggressive clinical course observed in this subgroup and separate them from other exon 11 variants. Integrated multi-omics approaches combining whole-genome sequencing, genome-wide DNA methylation profiling, and gene-expression analysis further confirm that KIT Δ557-558 tumors and extra-gastrointestinal GISTs share extensive chromosomal instability and coordinated epigenomic deregulation, supporting their classification as biologically high-risk entities[73].

Clinical applications of DNA methylation in GIST: Beyond defining molecular subsets, DNA methylation has also been explored for its clinical utility. Studies ranging from targeted promoter assays to genome-wide profiling and liquid biopsy approaches demonstrate that methylation analysis can complement mutational testing, refine prognostic assessment, and in selected contexts, inform therapeutic decisions. For clarity, the available data can be grouped into diagnostic applications (Table 1)[70-72,74-78], prognostic biomarkers (Table 2)[32,41,73,75,79-82], and predictive biomarkers (Table 3)[42,78,83-85]. The principal studies in each category are summarized in the following tables, together with their methodology, clinical role, and main findings.

Table 1 Diagnostic applications of DNA methylation in gastrointestinal stromal tumors[70-72,74-78].

Ref.	Cohort/samples	Targets	Methodology	Key findings
Casey et al[71]	59 GISTs (multi-institutional) + blood/saliva	SDHC promoter	Bisulfite sequencing	Epimutation found in approximately 94% of SDHx-WT cases, including Carney triad; mosaicism in peripheral tissues supports post-zygotic origin
Janeway et al[74]	Pediatric and WT GIST (approximately 100)	SDH subunits (SDHB, SDHC epimutations)	IHC + genetic/epigenetic testing	Established SDH-deficient GIST in children/WT adults; SDHB loss by IHC plus SDHC methylation testing defines subset
Haller et al[75]	Carney triad GISTs	SDHC promoter	Methylation assays	Recurrent SDHC hypermethylation identified as defining event in Carney triad, absent in Carney-Stratakis
Angelini et al[76]	63 tumors (KIT/PDGFRA-mutant vs SDH-deficient)	Genome-wide	450K arrays	C promoter hypermethylation represents a characteristic alteration in Carney triad but is not detected in Carney-Stratakis
Killian et al[70]	SDH-deficient vs syndromic GIST	Genome-wide	450K arrays	SDH-deficient GISTs characterized by a hypermethylator profile, with succinate accumulation blocking TET enzyme activity
Miettinen et al[72]	66/756 gastric GIST SDH-deficient	SDH status	IHC + molecular	Defined clinicopathologic spectrum of SDH-deficient GIST; IHC loss of SDHB as front-line marker
Saito et al[77]	35 GISTs	CpG island panel (p15, p16, p73, MGMT, hMLH1, MINT1/2/31)	MSP	94% tumors showed aberrant methylation; proposed CIMP-like category in GIST
Giger et al[78]	Plasma/tissue paired samples	SEPT9 (plasma methylation)	ddPCR	Circulating methylated SEPT9 showed feasibility as a liquid biopsy biomarker (AUC: 0.74-0.79)

GIST: Gastrointestinal stromal tumors; WT: Wild-type; SDH: Succinate dehydrogenase; SDHB: Succinate dehydrogenase subunit B; MGMT: O-6-methylguanine-DNA methyltransferase; IHC: Immunohistochemistry; MSP: Methylation-specific polymerase chain reaction; ddPCR: Droplet digital polymerase chain reaction; TET: Ten-eleven translocation; AUC: Area under the curve; CIMP: CpG island methylator phenotype.

Table 2 Prognostic biomarkers of DNA methylation in gastrointestinal stromal tumors[32,41,73,75,79-82].

Ref.	Cohort/samples	Targets	Methodology	Key findings
Geddert et al[32]	200 GISTs	CD133 promoter	MSP + IHC	Hypermethylation associated with reduced CD133 expression; paradoxically, CD133 expression predicted shorter DFS
Haller et al[75]	150 GISTs	SPP1 (osteopontin)	Locus-specific assays	Hypomethylation predicted poor DFS across size, site, and mitotic index
House et al[79]	38 gastric GISTs	E-cadherin (CDH1)	MSP + protein	Methylation correlated with loss of E-cadherin and independently predicted recurrence/mortality
Ricci et al[80]	Gastric GISTs	p16INK4a promoter	MSP + IHC	Hypermethylation linked to protein loss and poor outcome
Lou et al[81]	115 cases	REC8, PAX3, p16	MCAM + validation	≥ 1 gene methylated correlated with poor survival; site-specific patterns identified
Niinuma et al[41]	91 cases	MEG3	Methylation + transcript	Hypermethylation correlated with poor prognosis; restoration activated interferon responses
Ohshima et al[73]	KIT exon 11 Δ557-558	Genome-wide	WGS + methylome	Hypomethylation signature linked to CNV burden and p53 pathway inactivation; aggressive behavior
Kleijn et al[82]	28 patients	Genome-wide	EPIC methylome + CNV	CNV burden and DMRs associated with progression; added to risk models

GIST: Gastrointestinal stromal tumors; CD: Cluster of differentiation; IHC: Immunohistochemistry; MSP: Methylation-specific polymerase chain reaction; DFS: Disease-free survival; CNV: Copy number variation; WGS: Whole-genome sequencing; DMR: Differentially methylated region; EPIC: Illumina EPIC methylation array.

Table 3 Predictive biomarkers of DNA methylation in gastrointestinal stromal tumors[42,78,83-85].

Ref.	Cohort/samples	Targets	Methodology	Key findings
Giger et al[78]	SDH-deficient cohorts	MGMT promoter	MSP	MGMT methylation enriched in SDH-deficient GIST; suggested biomarker for temozolomide sensitivity
Ricci et al[83]	74 cases	MGMT promoter	MSP	Confirmed MGMT methylation in SDH-deficient/WT tumors (67% vs 15%); advocated alkylator trials with stratification
Italiano et al[84]	Imatinib-resistant cases	IGF2/IGF1R deregulation	Epigenetic expression analysis	Epigenetic activation of IGF axis implicated in imatinib resistance
Urbini et al[85]	Quadruple WT vs SDH-deficient	FGF4	Genomic vs methylation	In SDH-deficient, FGF4 induced by methylation; in WT, due to duplication. Suggested FGFR inhibitors
Xu et al[42]	53 cases (tissue)	METTL3-MRP1 (RNA methylation)	m6A + functional assays	High METTL3 expression linked to shorter PFS in resistant patients; RNA methylation mechanism of imatinib resistance

MRP1: Multidrug resistance protein 1; GIST: Gastrointestinal stromal tumors; WT: Wild-type; SDH: Succinate dehydrogenase; MGMT: O-6-methylguanine-DNA methyltransferase; IGF2: Insulin-like growth factor 2; IGF1R: Insulin-like growth factor 1 receptor; FGF: Fibroblast growth factor; m6A: N6-methyladenosine; MSP: Methylation-specific polymerase chain reaction; PFS: Progression-free survival; FGFR: Fibroblast growth factor receptor.

Clinical applications of histone modifications in GIST

The earliest clinical attempts to translate histone modulation into therapy in GIST were tested in phase I trials. Bauer et al[86] conducted a dose-escalation study of the pan-HDAC inhibitor panobinostat in combination with imatinib in patients with advanced, imatinib-refractory GIST. They established the maximum tolerated dose of panobinostat and demonstrated on-target activity, as histone H3 acetylation increased in peripheral blood mononuclear cells. Although objective tumor shrinkage was not observed, several patients experienced prolonged disease stabilization, indicating biological activity and supporting the feasibility of HDAC inhibition in resistant GIST[86]. In another trial, Deming et al[87] evaluated vorinostat combined with the proteasome inhibitor bortezomib in patients with refractory solid tumors, including a subset with GIST. The regimen was generally tolerable, with fatigue and thrombocytopenia as the most frequent adverse events, and provided modest clinical benefit in the form of stable disease. Importantly, histone acetylation changes were confirmed pharmacodynamically, reinforcing proof of mechanism in patients[87].

Beyond these early trials, a growing number of patient-based investigations have explored histone modifications as diagnostic, prognostic, or predictive biomarkers in GIST. These studies, summarized in Table 4, rely on tumor cohorts and link histone alterations with clinical outcomes.

Table 4 Histone modifications as diagnostic, prognostic, and predictive tools in gastrointestinal stromal tumors[41,50,52,54-56,60-62,64].

Ref.	Patient cohort/clinical role	Methodology	Key findings
Jin et al[60]	119 primary GISTs/diagnostic	IHC, digital quantification	PHH3 counts correlated strongly with mitotic index and NIH risk category improved reproducibility compared with conventional mitotic figures
Huang et al[56]	134 patients (discovery: 18 tumors from 14 patients; prevalence validation: 125 tumors from 120 patients; 89 high-risk/metastatic, 34 low/intermediate-risk)/prognostic	Exome + targeted resequencing (KIT/PDGFRA/SETD2); H3K36me3 IHC; 450K methylation array; RNA-sequencing	SETD2 mutations in 11.2% (10/89) of high-risk and 0% (0/34) of low/intermediate-risk GISTs gastric SETD2-mutant tumors show hypomethylated heterochromatin and reduced H3K36me3; associated with shorter relapse-free survival
Shen et al[54]	48 GISTs, paired tumor vs adjacent tissue/diagnostic/prognostic	IHC, ChIP-qPCR	KDM6A and SPARCL1 correlated with mitotic index, risk classification, and metastatic recurrence. Diagnostic ROC analysis: AUC: 0.698 (KDM6A), 0.834 (SPARCL1), 0.858 (EZH2). Low KDM6A linked to higher metastasis rates and reduced survival
Ji et al[55]	46 GIST tumors with matched adjacent tissue; stratified by NIH risk (low/intermediate/high)/diagnostic/prognostic	IHC, molecular assays	Tumors showed SMYD2 and EZH2 upregulation with H3K27me3 increase and TET1 downregulation; high-risk cases enriched for this profile
Niinuma et al[41]	131 primary GISTs, validation in 44 tumors/prognostic	Methylation assays, ChIP	MEG3 hypermethylation associated with higher recurrence risk (HR = 2.3; 95%CI: 1.2-4.6) altered H3K4me3 linked to poor recurrence-free survival
Qiu et al[52]	26 paired GIST tumors vs adjacent tissues/prognostic	IHC, tissue analysis	miR-409-5p expression is decreased in GIST tumors compared to adjacent normal tissues. Lower miR-409-5p levels correlate with higher GIST risk grade
Gao et al[64]	67 WT GIST patients/prognostic	IHC, tissue studies	SDHB deficiency: Higher ZNF148 expression, significantly shorter PFS (median approximately 104 months vs not reached). High ZNF148-pS306: Inferior PFS. SDHB deficiency = independent predictor of shorter PFS
Wang et al[50]	9 GISTs (3 low-, 3 moderate-, 3 high-risk)/prognostic	Mass spectrometry	Identified 2904 acetyl sites on 1319 proteins (quantified 2548 sites on 1169 proteins) vs low risk, 42 sites (38 proteins) up and 48 sites (44 proteins) down in high/moderate risk. Largest shifts: Ki67 K1063Ac (increase), FCHSD2 K24Ac (decrease)
Yen et al[61]	Public microarray datasets (including GIST samples with exon 11 mutations) compared to other sarcoma subtypes; validation on patient-derived GIST tissues/translational	Connectivity map + tissue validation	ETV1, along with KIT, DOG1, and PKCθ, was significantly overexpressed in GIST patient samples compared to multiple sarcoma subtypes; phenothiazines (trifluoperazine, thioridazine) emerged as candidate ETV1-targeting agents with potential for biomarker-guided drug repositioning
Zeng et al[62]	Twelve matched pairs of GIST resections from the same patients, taken pre-imatinib and after adjuvant imatinib with subsequent relapse/predictive	IHC, paired tissue analysis	IHC on paired samples revealed that BCL6 expression was consistently elevated after imatinib treatment in the 12 patient pairs, linking its upregulation with resistance and recurrence, and suggesting a rationale for combining BCL6 inhibition with imatinib therapy

GIST: Gastrointestinal stromal tumors; WT: Wild-type; NIH: National Institutes of Health; IHC: Immunohistochemistry; ChIP: Chromatin immunoprecipitation; qPCR: Quantitative polymerase chain reaction; ROC: Receiver operating characteristic; AUC: Area under the curve; HR: Hazard ratio; CI: Confidence interval; SDHB: Succinate dehydrogenase subunit B; PFS: Progression-free survival; PKC: Protein kinase C.

Therapeutic implications of chromatin remodeling in GIST

Clinical profiling links specific chromatin programs to distinct treatment needs. In KIT exon 11 deletion tumors that include codons 557-558, whole-genome and methylome data from a surgical cohort showed pervasive structural variation, intergenic hypomethylation, frequent 9p/22q losses (with recurrent CDKN2A involvement), and transcriptomic signatures consistent with p53 pathway inactivation and chromosomal instability features concentrated in the high-risk subgroup[73]. Although survival differences did not reach significance, these patients trended toward poorer outcomes, aligning the molecular profile with clinical aggressiveness[73]. Notably, they demonstrated that Hsp90 inhibition has already documented PFS benefit in advanced GIST and is approved in Japan. Given the elevated hypoxia signature observed in KIT 557-558 cases, they highlight a rationale to explore whether hypoxia-linked biology can enrich for benefit from Hsp90-directed regimens in this genotype[73].

In KIT/PDGFRA-WT disease, clinical series underscore that these tumors are less responsive to imatinib and often depend on FGFR signaling through FGF4 overexpression. Two patient subsets converge on this axis but via different chromatin mechanisms: SDH-deficient GISTs use methylation-mediated insulator disruption, whereas “quadruple WT” GISTs use focal 11q13 duplications that reposition FGF4 near the ANO1 super-enhancer both identified directly in primary tumors[85]. Therapeutically, this molecular stratification supports considering FGFR-targeted strategies (selective FGFR inhibitors or multi-kinase agents with FGFR activity) in WT GISTs, while regorafenib approved in later lines and with some FGFR activity remains an option under current standards[85]. In addition, paired patient specimens obtained before and after imatinib exposure showed increased BCL6 expression post-therapy, providing clinical evidence of an adaptive, chromatin-linked resistance program that represses TP53; while the combination concept is supported preclinically, these patient tissue data specifically flag BCL6 as a clinically observable resistance biomarker to motivate trials of BCL6-modulating combinations with imatinib[62].

DISCUSSION

Across GIST subtypes, chromatin remodeling provides a mechanistic bridge from genotype to clinical behavior and offers tractable nodes for therapy. In KIT exon 11 Δ557-558 GISTs, the combination of intergenic hypomethylation, structural alterations, and 9p/22q chromosomal losses leads to suppression of p53 signaling, providing a mechanistic basis for their aggressive behavior and highlighting potential translational avenues[73]. First, the hypoxia-linked expression signature reported in this group provides a rationale to test whether Hsp90 inhibition (already clinically active in GIST) preferentially benefits patients whose tumors exhibit high hypoxia/p53-inactivation scores[88]. Second, the high frequency of CDKN2A alteration aligns with preclinical evidence that relieving polycomb repression (e.g., targeting CBX7 to reactivate p16 INK4a) can enforce G1 arrest[65], nominating cell-cycle checkpoint restoration as a complementary strategy in this genotype. In WT GISTs, aberrant activation of the FGF4-FGFR pathway can emerge through two epigenetic or structural routes: Loss of insulator function via methylation in SDH-deficient tumors, and focal genomic duplication in quadruple WT cases[85]. This duality matters clinically: It justifies a biomarker-led, FGFR-oriented approach in KIT/PDGFRA-WT disease while clarifying that “FGF4-high” tumors may not share the same epigenetic etiology[85]. In parallel, chromatin-active cytotoxics and degraders broaden combination options beyond kinase sequencing. Proteasome inhibitors suppress KIT transcription and trigger histone-linked apoptosis across TKI-resistant models[48]. SWI/SNF deficiency creates dependency on BET activity[58] and BCL6 induction after imatinib in paired patient samples marks a therapy-induced, SIRT1-dependent silencing of TP53 that can be pharmacologically undone[62]. Collectively, these data support a pragmatic clinical roadmap: (1) Incorporate methylation/structural readouts (e.g., Δ557-558 status, CDKN2A loss, FGF4 mechanism) into risk and trial stratification; (2) Match chromatin vulnerabilities to combinations FGFR blockade in FGF4-driven WT disease[85], Hsp90 inhibitors in hypoxia-high Δ557-558 tumors[48], proteasome or BET inhibitors in defined chromatin contexts[48,58]; and (3) Intercept adaptive resistance with BCL6-directed agents to restore p53-mediated apoptosis alongside KIT inhibition[62]. In sum, chromatin remodeling is both a classifier and a lever in GIST one that can refine prognosis, reveal drug sensitivities, and enable rational combinations to extend the durability of response. In parallel, identification of predictive biomarkers including specific molecular alterations, non-coding RNA signatures, SLITRK3 expression, aberrant DNA methylation patterns, and radiomics-derived phenotypes will be essential for optimizing individualized regimens and improving long-term outcomes.

While the accumulating evidence underscores the central role of non-coding RNAs in GIST pathogenesis and resistance, this review focuses primarily on chromatin remodeling, while other epigenetic layers such as non-coding RNAs (microRNAs, lncRNAs, circRNAs) and epitranscriptomic regulation are only briefly mentioned despite their clear relevance to GIST biology. Much of the available evidence derives from small, single-institution cohorts or preclinical systems, limiting generalizability and raising the possibility of inflated effect sizes. Heterogeneity in assay platforms, endpoints, and clinical definitions further complicates comparison across studies. Mechanistic findings such as polycomb de-repression, BET sensitivity in SWI/SNF loss, or BCL6-SIRT1-TP53 repression remain largely preclinical and require clinical validation. Finally, proposed therapeutic strategies based on chromatin biology are still exploratory, and their true clinical impact will depend on prospective, molecularly stratified trials.

As mentioned above, SDH-deficient GISTs frequently stem from germline or somatic disruptions in SDHx subunit genes (SDHA, SDHB, SDHC, SDHD) or epigenetic mechanisms such as SDHC promoter hypermethylation, often signaling an underlying hereditary predisposition to paraganglioma-pheochromocytoma syndrome and related malignancies[39]. Germline testing of normal tissue is essential to identify pathogenic variants, facilitating cascade screening for at-risk relatives who may carry elevated lifetime risks of paragangliomas, pheochromocytomas, renal cell carcinomas, pituitary adenomas, and other SDHx-linked tumors, even in the absence of personal or strong family history[39,89]. Penetrance for clinically manifest tumors remains relatively low in non-proband carriers (e.g., approximately 10%-23% by age 60-70 years for SDHB/SDHC/SDHD variants, with even lower estimates for SDHA), highlighting that many carriers will never develop disease, yet probands often present at younger ages (median approximately 28 years) with potentially aggressive or metastatic phenotypes[89,90]. Due to current limitations in routine SDHA full-gene sequencing and methylation analysis, all individuals diagnosed with SDH-deficient GIST should undergo lifelong monitoring for additional SDHx-related tumors regardless of germline results, using standardized protocols that include baseline and periodic imaging (e.g., magnetic resonance imaging neck/thorax/abdomen/pelvis every 3-5 years) alongside clinical and biochemical surveillance to balance early detection benefits against burden and anxiety[39,89]. This unified approach, ideally coordinated in specialized centers, addresses regional variability in testing recommendations and supports family-based risk stratification while recognizing the distinct challenges in paediatric cases, where SDHx-linked PPGL may manifest early with catecholamine-related complications or syndromic features[91,92].

Despite growing interest in epigenetic regulation in GISTs, several limitations currently constrain clinical translation. First, most mechanistic insights derive from in vitro models or small retrospective patient cohorts, and robust validation in large, prospective, molecularly annotated datasets remains limited[93,94]. Second, epigenetic alterations in GIST often coexist with strong oncogenic drivers and metabolic reprogramming, making it challenging to document causality from secondary adaptive changes[95]. Third, tumor heterogeneity across molecular subtypes including KIT/PDGFRA-mutant, SDH-deficient, NF1-associated, and quadruple WT tumors complicates the development of universally applicable epigenetic biomarkers or therapeutic strategies[96]. Future research should prioritize integrative multi-omics approaches combining whole-genome sequencing, DNA methylation profiling, chromatin accessibility mapping, transcriptomics, and spatial immune characterization to resolve subtype-specific epigenetic dependencies[97-99]. Functional studies linking defined epigenetic states to immune evasion, drug tolerance, and metastatic behavior will be critical to identify clinically actionable vulnerabilities[100]. In parallel, biomarker-driven clinical trials evaluating rational combinations of epigenetic modulators with TKIs or immunotherapies are required to determine therapeutic benefit beyond preclinical observations. Finally, standardized epigenetic assays and analytical pipelines will be essential to enable reproducibility and cross-study comparison.

CONCLUSION

Epigenetic deregulation, particularly through histone modifications and chromatin remodeling, has emerged as a fundamental driver of GIST biology, shaping transcriptional networks, therapeutic resistance, and clinical behavior. Preclinical studies highlight multiple druggable vulnerabilities, and early clinical investigations provide proof of feasibility for targeting histone and chromatin regulators. Integrating these insights into patient care will require systematic validation in larger, well-annotated cohorts and prospective trials that embed molecular stratification. Ultimately, chromatin-directed strategies may expand the therapeutic armamentarium in GIST, offering opportunities to overcome resistance and refine precision oncology approaches.