The coding gene SMARCB1 (also known as INI1, SNF5 or BAF47) is located at q11.23 on chromosome 22 and is crucial for the stability of the SWI/SNF complex, but not its structural integrity. The introduction of exogenous SMARCB1 into SMARCB1-deficient cells has been shown to activate bivalent promoters and enhancers, rescuing the expression of upregulated target genes, which were identified by Gene Ontology analysis (a widely used bioinformatics approach for annotating genes and interpreting large-scale genomics or transcriptomics data) to be related to kidney and neural development (22). As a tumor suppressor gene, oncogenesis occurs only when both SMARCB1 alleles in a cell are mutated (23). Gene sequencing has revealed that in SMARCB1-deficient tumors, chromosome 22q loss frequently extends far beyond SMARCB1 gene length, with extensive copy number loss and rearrangements, resulting in concurrent defects at multiple gene loci (24,25).

SMARCC1/2 and SMARCD1/2/3 comprise the primary framework of the BAF complex, and serve a notable role in early assembly of the SWI/SNF complex (26,27). Mutations in SMARCC1 (also known as BAF155), which inhibits fetal differentiation and maturation, facilitate the transformation of stem cells into mature cells in the human gut by augmenting CD117 and cluster cells (28). Conversely, during the resolution of inflammation, SMARCC1 undergoes polyubiquitination and proteasomal degradation, thereby inhibiting the expression of inflammatory factors (29). SMARCD2 serves a crucial role in granulocyte maturation and is a promising therapeutic target in leukemia (30,31).

SMARCA2 and SMARCA4 are mutually exclusive subunits that function as ATPases in the SWI/SNF complex, generating energy for chromatin remodeling through hydrolyzing ATP. The bromine domain recognizes acetyl-lysine residues at the N-terminus of histones and hydrolyzes high-energy phosphate bonds of ATP as an energy source (32). The SMARCA4 (BRG1) gene, located at p13.2 on chromosome 19, serves as a valuable marker for investigating rare tumors, such as small cell carcinoma of the ovary, hypercalcemic type, in younger women (<40 years old) (33,34). By contrast, mutations in SMARCA2 (BRM) are less prevalent in human tumors, most of which are reversible epigenetic alterations rather than gene deletions (35). Upon SMARCA4 inactivation, cells rely on SMARCA2 to maintain normal vital activity, a phenomenon known as synthetic lethality (36).

ARID1A/B, also referred to as BAF50a/b, is the most frequently mutated and extensively studied among these subunits, sharing a highly similar base composition and arrangement with ARID2, which exists in the SWI/SNF complex in a mutually exclusive manner. Research has shown that ARIDs contribute to proper spatial folding patterns to maintain structural stability of the complex (26). Deletion of ARID1B disordered region 1 can completely inhibit condensation of the BAF complex in vitro (37). ARID1A depletion alters genome-wide expression, promoting oncogenesis by suppressing enhancer accessibility, whereas ARID1B deficiency has negligible effects (38). The majority of ARID1A mutations result from premature termination of protein transcription due to insertion/deletion mutations (39). Point mutations impair ARID domain mutual recognition and non-specific DNA binding, thereby inhibiting ES cell differentiation without affecting AIRD1A protein expression or complex integrity (40). In addition, ARID1B loss disrupts binding to the long non-coding RNA (lncRNA) NEAT1, which is a component of paraspeckles, resulting in the loss of numerous transcription and chromatin-associated proteins, including the chromatin-modifying enzyme JMJD6, the transcription factor CCCTC-binding factor, which binds to DNA sequence-specific sites and regulates the 3D structure of chromatin, and histones (41).

In conclusion, the recruitment of the SWI/SNF complex is synergistically driven by ARID-mediated non-specific DNA binding, bromodomain interaction with histone N-termini and specific DNA sequence recognition (42–45). Binding of the SWI/SNF complex to chromatins is complex. Transcription typically starts with the recruitment of transcription factors, together with enhanced accessibility of the target DNA. A subset of these transcription factors pre-aggregate around the target nucleosome, increasing accessibility and forming a dynamic open reading frame. Additionally, this complex regulates gene expression via DNA methylation, damage repair and microRNA (miR) processing (42).

Esophageal cancer with SWI/SNF mutations exhibits high aggressiveness, with a notable proportion of patients succumbing to the disease within 1 year, often due to being diagnosed with late-stage cancer (48). A 2023 study reported a pathogenic SMARCA4 mutation in 2.5% of undifferentiated carcinomas at the esophageal or gastroesophageal junction, which may be relevant to Barrett's esophagus. Notably, >50% of these patients presented with distant metastases at initial diagnosis (49). As a critical component of the tumor suppressor gene family, the loss of SMARCA4 can directly promote tumor cell proliferation. SMARCA4 mutations are primarily categorized into two classes (50). Class 1 mutations result in loss of protein function, encompassing truncation mutations, gene fusion and homozygous deletion, whereas class 2 mutations involve missense mutations. Cancers with class 1 mutations tend to be less differentiated than those with class 2 mutations, which are strongly associated with poor patient outcomes, and may co-occur with APC and CTNNB1 mutations, leading to synergistic effects (49). The mutation rates of SMARCB1, SMARCA2, SMARCA4 and ARID1A in esophageal adenocarcinoma have been reported to be 2, 9.9, 3.4 and 10.4%, respectively. Furthermore, patients with ARID1A mutation were demonstrated to have a 56% reduction in overall survival (OS) time (60.1 months vs. 26.2 months, P=0.044) compared with in those with ARID1A wild-type esophageal adenocarcinoma (51). Notably, the absence of ARID1A may be associated with high microsatellite instability (MSI-H) in tumors.

Mutations in SWI/SNF complex genes (SMARCB1, SMARCA2, SMARCA4 and ARID1A) have been identified in ~35% of patients with GC, with SMARCA2 and ARID1A accounting for 28 and 16%, respectively. Furthermore, concurrent mutations in more than two subunits are frequently detected. These patients are typically of an advanced age and exhibit susceptibility to lymphatic invasion (52). The Cancer Genome Atlas network identified SWI/SNF-deficient tumors in Epstein-Barr virus (EBV) and MSI genotypes, demonstrating effective response to chemotherapy and immunotherapy (53–55). However, in GC cases with high MLH1 or EBV deficiency, ARID1A mutations are associated with a poor prognosis (39).

Loss of ARID1A function may be caused by mutations and copy number deletions, both of which are enriched in GC with MSI and chromosomal instability (CIN), respectively. According to recent findings, PD-1 targeted therapy improves OS in patients with GC lacking ARID1A, whereas 5-fluorouracil therapy significantly benefits only those with the CIN subtype (56). This may be associated with increased activity of carcinogenic signaling pathways, which also further promote the occurrence of ARID1A-negative GC. Furthermore, ARID1A can regulate the proliferation and invasion of GC cells through non-coding RNAs, such as miR-223-3-p, miR-7641 and the lncRNA MVIH (57). ARID1A is also involved in cellular DNA damage repair to maintain genomic stability. Furthermore, when ARID1A is artificially knocked out in mice, chromosome and gene instability are increased (58).

A single center study conducted in 2022 in China suggested that elevated SMARCC1 expression may be associated with poor prognosis in patients with GC (P<0.01) (59). SMARCA4-deficient GC cases generally occur independently of EBV infection and MSI, exhibiting larger tumor volume and deeper infiltration, with a notable tendency for lymph node metastasis (60). In undifferentiated gastrointestinal carcinomas, OS and disease-free survival have been reported to be reduced in patients harboring SMARCA4 mutations (19). However, Zhang et al (61) reported that SMARCA4 negativity is not a predictor of OS.

SMARCA4 serves a crucial role in the development and homeostasis of the mouse gut by regulating the proliferation and differentiation of duodenal stem cells, as well as villi and crypt formation (62). The absence of SMARCA4 can lead to intestinal inflammatory disorders, triggering autoimmune enteritis (63). Furthermore, evidence has shown that the SWI/SNF complex has a crucial role in the differentiation of intestinal pluripotent ES cells (64). Undifferentiated tumors of the small intestine originating from SWI/SNF mutations are rare and must be distinguished from other tissue-derived tumors during diagnosis (65).

Gene sequencing has identified ARID1A mutations in ~17% of hepatocellular carcinoma (HCC) cases and ARID2 mutations in ~6% of cases (66). These tumors are typically characterized by their large size, tendency to induce refractory ascites and increased probability of recurrence. This phenomenon is attributed to the pronounced Warburg effect in ARID1A-deficient HCC, which enables tumor cells to sustain high metabolic activity, even in glucose-depleted environments (67). Deletion of the tumor suppressor gene ARID1A in hepatocytes can result in the upregulation of CD133 and EPCAM (hepatocyte stem cell markers) and the loss of PROS1 and DCXR (hepatocellular-specific markers), consequently leading to the reacquisition of self-renewal and differentiation potential (68). Furthermore, the tumor driver mTORC1 degrades the ARID1A protein in the liver, impairing chromosome accessibility and activating carcinogenic signaling pathways (such as YAP), thereby promoting HCC (69). Inactivation of ARID1A disrupts chromatin spatial structure, interrupts the normal transcription of cancer genes, and increases the aggressiveness of HCC by inhibiting the BRG1-RAD21 axis (70). Increased release of pro-inflammatory factors, such as TNF-α and IL-6, also contributes to ARID1A-deficient liver steatosis and HCC development. After the depletion of ARID1A in endothelial cells, the concentration of ANG2 (a critical angiogenic factor) is markedly increased, which can promote cell proliferation and migration, providing an enhanced nutritional foundation for HCC growth (71). Additionally, ARID1A serves a role in the regulation of the tumor microenvironment, and its loss may augment the responsiveness of HCC to immunotherapy (72).

Furthermore, ARID2 expression is positively associated with HCC prognosis. It inhibits tumor cell metastasis in HCC by recruiting DNMT1, suppressing epithelial-mesenchymal transition and accelerating the DNA methylation of Snail (73). Under normal conditions, the nucleotide excision repair process of DNA damage requires the participation of ARID2 to r counteract the effects of carcinogens, such as ultraviolet light and chemical substances, thereby reducing cellular variation and apoptosis (74).

The SWI/SNF complex acts as tumor suppressor in humans and its reduced expression promotes carcinogenesis through multiple mechanisms. However, one study challenged this perspective. A significant increase in SMARCB1 protein levels was reported to be observed in patients with HCC and was associated with advanced Tumor-Node-Metastasis (TNM) classification (75). The downstream target gene NUP210, which encodes the nuclear pore complex, facilitates the interaction between SMARCB1 and chromatin. SMARCB1 also sustains HCC via NUP210-mediated cholesterol homeostasis and metabolism of exogenous substances. These findings indicated that the SWI/SNF complex may have a dual effect on tumor inhibition and carcinogenesis in the liver. Overexpression of ARID1A can promote tumor initiation; however, in established liver tumors, reduced ARID1A leads to tumor progression and metastasis due to reduced expression of inhibitory factors (76).

SWI/SNF mutations are present in ~50% of gallbladder cancer and BDC, second only to endometrial cancer (77). It has been reported that the SWI/SNF complex subunits ACTL6A and SMARCA4 facilitate TGF-β-induced LINC00313 gene expression, thereby modulating the Wnt signaling pathway and indirectly promoting the growth of cholangiocarcinoma in vivo (78). However, research on the SWI/SNF complex in BDC remains limited. A 2017 study linked PBRM1 (encoding BAF180, a variant subunit) to advanced BDC (79). Previous studies have suggested that KRAS mutations are one of the main pathogenic pathways of (80,81). In addition, mutations of chromatin remodeling genes (ARID1A or PBRM) have frequently been observed in cholangiocarcinoma via next generation sequencing (82,83). However, ARID1A and KRAS co-mutations have been reported to be absent in intra- or extrahepatic BDC, indicating a novel tumorigenesis mechanism (84). Guo et al (85) demonstrated that the TGF-β-Smad4 pathway may mediate the loss of resistance to biliary injury, resulting in uncontrolled epithelial proliferation in the presence of ARID1A mutations.

According to an Australian study that performed whole gene sequencing, ~14% of pancreatic ductal adenocarcinoma (PDAC) cases exhibit mutations in the SWI/SNF complex (ARID1A, SMARCA4 and PBRM1) (86). ARID1A serves a crucial role in maintaining pancreatic cellular homeostasis. Accordingly, ARID1A deletion induces acinar cell damage and transformation into the pancreatic duct and mucus accumulation, potentially leading to PDAC (87). Meanwhile, the anticancer activity of T cells in the tumor microenvironment is diminished, along with inhibition of the IFN signaling pathway, resulting in immunosuppression and weakening of the antitumor immune response (71). ARID1A knockout in mice has been shown to increase vimentin expression and epithelial-mesenchymal transition markers, resulting in poorly differentiated or undifferentiated pancreatic tumors. These tumors resemble rhabdoid cells, exhibiting highly aggressive behavior, and have severe adverse effects on prognosis (88,89). Furthermore, ARID1A silencing irreversibly accelerates PDAC formation (70,90). Notably, the SWI/SNF complex serves a crucial role in the development of islet cells. The deletion of ATPase subunits, such as SMARCA2 and SMARCA4, in endocrine progenitor cells disrupts glucose homeostasis disorder in mice after birth, leading to severe impaired glucose tolerance, hyperglycemia and hypoinsulinemia (91).

SWI/SNF mutations have been observed in CRC associated with ulcerative colitis, with an ARID1A mutation rate of ~44% and a SMARCA4 mutation rate of ~17% (92). Dysfunction of autophagy in the intestinal epithelium with SMARCA4 deletion leads to the accumulation of intracellular reactive oxygen species, potentially compromising the intestinal mucosal barrier, thereby triggering intestinal inflammation and tumor formation (93). Mutated SMARCA4 promotes CRC proliferation through multiple mechanisms (94–96). By contrast, SMARCA4 has been reported to be crucial for sustaining tumor stem cell activity (97). Notably, knockout of SMARCA4 in mice has been shown to inhibit colonic tumor growth and induce apoptosis (98). SMARCB1 exhibits a similar effect (99), as its loss induces CRC cell differentiation and inhibits neoplasm formation. Chen et al (100) revealed that bromodomain containing 9 (BRD9), a newly discovered SWI/SNF complex subunit, is significantly enriched in cancer tissues, including colon and lung cancer. Furthermore, tumor progression could be inhibited by Wnt/β-catenin signaling pathway following BRD9 deletion. Research on animal models has demonstrated that inactivation of the enhancer H3K27ac and dysregulation of the APC/β-catenin signaling pathway in infiltrating adenocarcinoma cells with ARID1A deletion is crucial for the development of CRC (3,101).

Loss of the SWI/SNF complex is rare in CRC, occurring primarily in cases involving inflammatory bowel disease and the MSI genotype. SMARCA5 is associated with an imbalance of T helper 17/regulatory T cells in intestinal tissues, contributing to the promotion and exacerbation of colon inflammation, which may be associated with an increased risk of CRC (102,103). In general, CRC with SWI/SNF mutations exhibits higher malignancy and an advanced TNM stage, and the prognosis worsens with mismatch repair defects and the presence of the BRAF V600E mutation (104). Notably, ARID1A loss is closely associated with lymphatic infiltration, recurrence and metastasis (105). Nevertheless, patients with stage IV CRC who are ARID1A-negative have been demonstrated to exhibit improved OS, compared with those patients without an ARID1A mutation, which is potentially attributable to the robust response of ARID1A mutations to immune checkpoint inhibitors (106).