Subtype-Specific Noncoding Ribonucleic Acid Signatures in Gastric Cancer: Insights from The Cancer Genome Atlas-Based Classification

General biomarkers in EBV-associated GC

Approximately 9% of GC tumors related to EBVaGC and are identified by the presence of EBV in tumor cells, which leads to particular molecular alterations. PIK3CA mutations are a common feature of this subtype (in approximately 80% of cases) and show elevated expression of PD-L1 and PD-L2 (Baccili Cury Megid et al., 2023). These alterations activate the PI3K/AKT signaling pathway, resulting in cell survival and proliferation.

Increased expression of PD-L1 and PD-L2 is associated with high cytotoxic T lymphocyte (CTL) infiltration, that could be targeted by PD-1/PD-L1 inhibitors such as pembrolizumab. However, one of the key challenges is heterogeneous responses to immunotherapy due to variability in PD-L1 expression and tumor-infiltrating lymphocyte density (Baccili Cury Megid et al., 2023; Gong et al., 2019). Another trait that find in EBVaGC is the hypermethylation of tumor suppressor genes (TSGs) such as CDKN2A (p16) and ARID1A, which are implicated in immune evasion and carcinogenesis (Baccili Cury Megid et al., 2023).

The EBV infection cycle proceeds through latent and lytic phases. During the latent phase, small, non-coding, EBV-encoded RNAs (EBERs) are implicated in cellular transformation and proliferation. Leveraging their stable stem-loop structures, EBERs interact with cellular proteins to promote cell proliferation and immune evasion while inhibiting apoptosis (Ahmed et al., 2021). Identified EBER-interacting targets include TLR3, RPL22, La antigen, RIG-1, PKR, 2′−5′-oligoadenylate synthetase, PAX5, AUF1, IGF-1, IL-6, IL-9, and STAT3 (Ahmed et al., 2021).

EBV carcinogenesis is mediated by viral and host gene expression regulation, and the pattern of altered gene expression is determined by the latent (0, I, II, III) or lytic cycle (Ren et al., 2012). In most of the EBVaGC tumors gene expression alteration that associated with latency phases I and II, including EBER1, EBER2, and BamHI-A rightward transcripts (BARTs) (Sugiura et al., 1996). Also, some studies reported the expression of latency phase II genes such as BARF1, LMP-1, LMP-2A, and LMP-2B in EBVaGC, with EBER, EBNA1, and BARTs showing the highest expression levels during latency (Luo et al., 2005). Among the EBERs, EBER1 serves as a great target for diagnosis due to its stable and highly expression (Ahmed et al., 2021; Ahmed and Khan, 2014).

Aberrant methylation is another characteristic of EBVaGC. A significant association has been reported between EBVaGC and the CpG island methylator (CIMP) phenotype (Stanland and Luftig, 2020). Hypermethylation of p14ARF, p16INK4A (CDKN2A), and RCOR2, in the absence of concurrent MLH1 methylation, is a key differentiator between EBVaGC and MSI (Morales-Sanchez and Fuentes-Panana, 2017).

Although many gene mutations are involved in EBVaGC, certain mutations are more prevalent. For instance, mutations in PIK3CA, ARID1A, BCOR, TP53, JAK2, PD-L1, and PD-L2 are frequently reported (Yang et al., 2020). These mutations can also occur in other GC subtypes, and their prevalence is significantly higher in EBVaGC, approximately 80% for PIK3CA, 47–55% for ARID1A, and 23% for BCOR (Cancer Genome Atlas Research Network, 2014). Interestingly, TP53 mutations are relatively rare in EBVaGC, which could be due to EBV-mediated stabilization of the p53 protein (Cancer Genome Atlas Research Network, 2014).

Gene deletions also have been investigated in EBVaGC, including WWOX (16q23.1), MACROD2 (20p12.1), PTEN (10q23.31), IMMP2L (7q31.1), FAM190A (4q22.1), and PTPRD (9p24.1) (Cancer Genome Atlas Research Network, 2014; Gulley, 2015). Moreover, other less-specific alterations, such as amplification of CD44 (11p13) and HER2 (17q12), have been reported (Gulley, 2015). According to TCGA analyses, there is a substantial association between EBVaGC and upregulation of interleukin-12 (IL-12) signaling and downregulation of hypoxia-inducible factor 1-alpha (HIF-1α) transcription (Cancer Genome Atlas Research Network, 2014).

The therapeutic landscape for EBVaGC is multifaceted, incorporating conventional chemotherapy, immunotherapy, and targeted agents. While some in vitro studies suggest chemoresistance, retrospective clinical data indicate that EBVaGC patients achieve high disease control rates (90.3–100%) and better survival with first-line platinum and fluoropyrimidine-based regimens (Corallo et al., 2020; Qiu et al., 2020; Corallo et al., 2024). Immunotherapy represents a particularly compelling strategy, underpinned by the characteristic dense lymphocyte infiltrate and frequent amplification of immune checkpoint molecules PD-L1 and PD-L2 in these tumors. Despite this strong rationale, clinical outcomes for ICIs have been variable, with overall response rates ranging from 0% to 100%. The unique genetic profile of EBVaGC also presents opportunities for targeted therapies; the high prevalence of activating PIK3CA mutations (∼80%) supports the investigation of PI3K inhibitors, while the frequent inactivation of the ARID1A tumor suppressor gene has prompted research into synthetic lethality strategies using EZH2 and PARP inhibitors (Corallo et al., 2024).

The prognostic significance of EBV positivity in GC remains a subject of discussion, though a significant body of evidence suggests a more favorable outcome. Multiple large-scale studies, including an international pooled analysis, have associated the EBV-positive subtype with improved survival and lower mortality compared with EBV-negative cases, even after adjusting for factors such as disease stage. Although some conflicting reports find no significant survival difference, a consistent observation is that within the EBVaGC cohort, the lymphoepithelioma-like carcinoma histological subtype is associated with the best prognosis (Corallo et al., 2024).

ncRNA biomarkers in EBV-GC

miRNAs, short ncRNAs (18–24 nucleotides), regulate gene expression by interacting with the 3′ untranslated region (3′ UTR) of mRNAs and causing mRNA degradation or translation inhibition. Based on the studies, EBVaGC can dysregulating miRNAs, such as downregulating pri-miR-200 that leads to impairment of E-cadherin and cell adhesion (Shinozaki et al., 2010). Also, increased expression of EBV-miR-BART14-3p showed in EBVaGC, which induce the host DNA damage response inefficiency by inhibiting the ATM protein (Lung et al., 2018).

As depicted in Figure 4, EBV-derived miRNAs play crucial roles in various oncogenic processes. These include suppressing pattern recognition receptors, modulating host cell antigen expression, inhibiting apoptosis, maintaining viral latency, and promoting cell proliferation, invasion, and metastasis (Zebardast et al., 2021). High-throughput analyses have identified several host miRNAs (hsa-miR-5787, hsa-miR-4446-3p, hsa-miR-335-3p, hsa-miR-6877-3p, and hsa-miR-1915-3p) as most involved miRNAs in EBVaGC (Jing et al., 2018). It has been suggested that hsa-miR-5787 inhibits cell proliferation by targeting eIF5; however, carcinogenesis or tumor suppression of this miRNA is not fully understood (Yoo et al., 2011). Another investigation shows that hsa-miR-5787 enhances glycolysis and chemotherapy resistance by mitochondrial oxidative phosphorylation impairment (Chen et al., 2019).

OPEN IN VIEWER

FIG. 4.

Mechanistic pathways of EBV-miRNAs in GC development. APC, adenomatous polyposis coli; BAD, the BCL2 associated agonist of cell death; BALF5, DNA polymerase catalytic subunit; BID, subgroup of the BCL‐2 family; BRUCE, baculovirus inhibitor of apoptosis repeat‐containing ubiquitin‐conjugating enzyme; BZLF1, BamHI Z fragment leftward open reading frame 1; Caprin, cytoplasmic activation/proliferation‐associated protein; DAB2, disabled‐2; EBF1, early B‐cell factor‐1; FOXP1, forkhead box protein P1; IP07, importin‐7; KLF2, Kruppel‐like factor 2; LMP, latent membrane protein; MAP3K2, mitogen‐activated protein kinase kinase; MICB, MHC class I chain‐related protein B; NDRG1, N‐myc downstream‐regulated gene 1; NF‐Κb, nuclear factor kappa‐light‐chain‐enhancer of activated B cells; NKG2D, natural killer group 2D; NLK, Nemo‐like kinase (Serine/threonine‐protein kinase); NLRP, NOD‐like receptor protein; PDL1, programmed death‐ligand 1; PTEN, phosphatase and tensin homolog; PUMA, the p53 up-regulated modulator of apoptosis; RIG, retinoic acid‐inducible gene; TOM, translocase of the outer mitochondrial membrane; TP53, tumor protein 53.

The role of hsa-miR-4446-3p in tumor biology is not well defined; however, high expression of hsa-miR-4446-3p is associated with tumor progression and targets oncogenes and TSGs such as PAX5, KMT2A, RUNX1, N-RAS, IDH2, and CEBPA (Han et al., 2023; Gaur et al., 2020). hsa-miR-335-3p is reported to have a considerable role in promoting GC proliferation, migration, invasion, and lymph node metastasis (Ma et al., 2022). The role of hsa-miR-6877-3p in tumorigenesis is unclear, but its decreased expression is associated with ovarian cancer recurrence (Sato et al., 2020). hsa-miR-1915-3p suppresses DUSP3, activating ERK1/2 signaling, and increasing cell proliferation and migration in GC (Guo et al., 2018). It is also involved in the pathogenesis of GC by suppressing antiapoptotic factor BCL-2 (Cui et al., 2019). circRNA-ABCB10 may function as a molecular sponge for hsa-miR-1915-3p, thereby promoting gastric carcinogenesis through the Rac1 axis (Liu et al., 2022a).

LncRNAs by interacting with DNA, RNA, and proteins could play a regulatory role in tumor initiation, progression, and metastasis (Xie et al., 2019). Abnormal expression of lncRNAs in EBV-infected cells is linked to the pathogenesis of EBVaGC (Zhang et al., 2020). For instance, SNHG8 is significantly upregulated in EBV-positive compared with EBV-negative GC cell lines and tissues (Huang et al., 2016). SNHG8 regulates various genes (e.g., BNLF2a, LF3, BHRF1, and BHLF1) and host targets (EIF4A2, TRIM28, PLD3, NAP1L1, TRPM7, and RPL18A), collectively resulting in oncogenesis (Huang et al., 2016). Silencing SNHG8 with shRNA in vitro and in vivo arrests the G0/G1 phase of the cell cycle, which indicates its oncogenic role and highlights its potential as a therapeutic target (Liu et al., 2018).

RNU12, a host-derived lncRNA, is involved in tumor suppression, and EBV infection modulates its expression in EBVaGC (Huang et al., 2016). Although the precise mechanism remains unclear, it may inhibit tumor progression by silencing miR-575 and regulating the proapoptotic factor BLID (Wang et al., 2023).

Upregulation of lncRNA H19 in EBVaGC is associated with increased tumor proliferation (Yang et al., 2012). H19, which is activated by c-Myc, suppresses the TP53/BAX/BCL2 apoptotic pathway and induces tumor progression (Liu et al., 2016). Decreasing plasma level of H19 after tumor resection shows a considerable biomarker potential for EBVaGC diagnosis (Yörüker et al., 2018). The role of H19 in EMT, metastasis, and poor prognosis make it a more interesting biomarker for GC (Li et al., 2019; Liu et al., 2022b).

MIR143HG is another lncRNA that although its role in EBVaGC is not clear, emerging evidence suggests its association with gastric tumorigenesis. MIR143HG probably acts as a tumor suppressor by inhibiting the Wnt/β-catenin and MAPK pathways and modulating HOXB7 methylation, which its expression alters in EBVaGC (Gou et al., 2023; Zhu et al., 2024; Nishikawa et al., 2017).

General biomarkers in MSI-GC

Microsatellites are short, repetitive DNA sequences found throughout the coding and noncoding regions of the genome. Their repetitive structure makes them prone to replication errors, such as insertions and deletions. These errors are normally corrected by the DNA MMR system, which is essential for maintaining GS (Puliga et al., 2021).

MSI-H-GC (about 22% of cases) have a favorable prognosis, especially in early stages. MSI-H-GC produce high levels of TMB and neoantigen due to MMR deficiency, which enhances immunogenicity and improves response to ICIs such as pembrolizumab. Hence, these tumors have good potential for immunotherapy (Ajani and Pedersen, 2024).

The relationship between clinical features and MSI-GC is still undefined; however, research is ongoing to identify prognostic and predictive markers to better classify MSI-GC cases. MMR gene mutations frequently result in missense or frameshift mutations affecting critical cell cycle regulators, apoptosis, genomic integrity, and mitotic progression, as illustrated in Figure 5 (Hudler, 2012; Cancer Genome Atlas Research Network, 2014; Cancer Genome Atlas Research Network, 2014; Puliga et al., 2021). The altered expression of MMR proteins (hMLH1, hPMS2, hMSH2, hMSH6), often due to promoter hypermethylation or coding region mutations, causes MMR deficiency (Ma et al., 2005). Compared with MSS tumors, MSI is a significant risk factor in GC and can increase mutation rates up to 1000-fold (Dudley et al., 2016).

OPEN IN VIEWER

FIG. 5.

Functional loss of MMR proteins causes mutations in oncogenes and tumor suppressor genes. MMR, malfunction of mismatch repair.

The assessment of MSI is thorough comparing microsatellite sequences in normal and tumor tissues. Five loci including BAT-25, BAT-26, D17S250, D5S346, and D2S123 by The Bethesda panel analyzes and based on that samples divides into MSI-H (instability at ≥2 loci), MSI-L (instability at 1 locus), or MSS (no change) (Murphy et al., 2006; Corso et al., 2011; Smith et al., 2023).

“Hotspots” are specific genomic regions with a high frequency of mutations, recombination, or structural rearrangements, playing a vital role in gene regulation and disease pathogenesis (Edsinger and Moroz, 2024). EGFR, ERBB2, ERBB3, and PIK3CA have been identified as MSI-associated hotspots in cancers and especially GC (Ghojazadeh et al., 2022). A meta-analysis also showed that HER2 expression is absent in more than 97% of benign MSI-H tumors (Zhang et al., 2024). Notably, mutations in tumor suppressor genes such as TP53, which are common in many cancers, are less frequent in MSI-H tumors. Unlike microsatellite stability (MSS) tumors, MSI tumors exhibit high expression of immune ligands such as CTLA-4, IDO, LAG-3, and PD-L1 (Llosa et al., 2015).

Gene expression profiling has revealed MSI-GC markers, such as HLA-DRA, HLA-DRB5, HLA-DQA1, HLA-DMA, FAS, JAK2, and CASP8, which are implicated in the prognosis and mechanism of the disease (Hang et al., 2018). The retinoic acid receptor beta gene is associated with good prognosis due to its inverse relationship with MSI and immune infiltration, and could be a biomarker and therapeutic target (Cai et al., 2024). This role is maintained despite its association with proliferation, invasion, and EMT.

ncRNA profile in MSI-GC

Regardless of extensive research on MSI-GC, its association with ncRNAs is still not clear. According to the ceRNA hypothesis, lncRNAs such as RP11-999E24.3, LINC00106, HULC, and RP11-314B1.2 could affecting the prognosis of GC through cross-regulated gene expression by competing for shared miRNAs (Qi et al., 2020).

A comprehensive analysis of the ceRNA network in MSI-GC identified 60 lncRNAs. Of these, 43 were found to be associated with MSI and were subsequently investigated for their prognostic value.MIR99AHG was recognized as the center of a subnetwork in which let-7f-5p played a protective role and miR-125a-5p played a risky role (Zhang et al., 2023b).

MIR99AHG shows a remarkable correlation with immune markers within the tumor microenvironment (TME) of MSI-GC, suggesting its potential as a biomarker for diagnosis, targeted therapy, and immunomodulation (Zhang et al., 2023b).

The heterogeneity of ncRNAs in MSI-H tumors implies the role of additional molecular factors in immune evasion. Even though MSI-H tumors have high immunogenicity, the response to ICIs varies across patients; thus, there is a need for complementary markers such as T-cell receptor (TCR) diversity, TME composition, and cytokine signatures to predict prognosis and treatment outcomes (Zhang et al., 2023a; Cancer Genome Atlas Research Network, 2014).

General biomarkers in CIN-GC

The CIN is commonly observed in various cancers and generally characterized by the frequent gain or loss of whole chromosomes or large chromosomal segments (Maleki and Rocken, 2017). CIN triggers carcinogenesis through inducing loss of heterozygosity or amplification of oncogenes, resulting in marked tumor heterogeneity (Maleki and Rocken, 2017; Bakhoum and Landau, 2017). This subtype is the most common molecular variant in GC (about 50% of cases) and is associated with considerable copy number alterations in genes such as ERBB2, FGFR2, MET, and KRAS, all of which serve as potential therapeutic targets (Danishevich et al., 2023).

Despite extensive research, the precise cause of CIN remains unknown. Some hypotheses suggest that it is owing to defects in oncogenes such as RAS and TSGs such as TP53; however; mutations in these genes have further been reported in tumors with stable karyotypes (Maleki and Rocken, 2017).

TP53, a key regulator of DNA damage and cell cycle, is predominantly mutated in the CIN subtype and related to drug resistance and poor prognosis (Nemtsova et al., 2023). Based on TCGA data, 71% of CIN tumors contain TP53 mutations, regularly linked to an intestinal phenotype (Zhang et al., 2022). These mutations typically occur in the DNA-binding domain, interfere with the regulation of target genes, and sometimes enhance carcinogenesis through a dominant-negative effect. TP53 mutations also prevent its degradation by ubiquitin ligases, leading to the accumulation of the mutant protein and increasing tumorigenesis (Nemtsova et al., 2023).

CCNB1 (Cyclin B1), an essential mediator of the G2/M transition in the cell cycle, is commonly overexpressed in the CIN-GC. Its presence in the nucleus and cytoplasm is contributed to CIN, poor prognosis, and reduced survival, establishing it as a potential therapeutic target (Nemtsova et al., 2023).

CDK1, vital for entry into mitosis, is overexpressed in CIN-GC and is related to tumor progression. It is implicated in advanced cancer by phosphorylating hTERT. Inhibition of CDK1 causes cell cycle arrest and reduced proliferation, particularly in cells with CDKN2A mutations. CDKN2A inhibitors such as PD-0332991 and dinaciclib have exhibited favorable antitumor effects (Nemtsova et al., 2023).

Several genes regulating cell division exert essential influences on CIN-GC. AURKA, a mitosis-essential kinase at 20q13, contributes to CIN by interfering with mitotic mechanisms (Mesic et al., 2016). This enzyme is involved in GC progression by phosphorylating targets such as histone H3, NDEL1, TACC3, and TPX2, and interacting with β-catenin and TP53 (Nemtsova et al., 2023).

AURKB, a crucial mitotic kinase, is implicated in CIN-GC by regulating chromosome condensation, the mitotic spindle, and cytokinesis. Its abnormal expression stimulates the G1/S transition by activating CCND1 via histone H3 phosphorylation (Nemtsova et al., 2023). Inhibition of AURKB reduces CCND1 levels, suppresses proliferation, and induces G2/M arrest (Wang et al., 2020).

HER2, located on chromosome 17, is overexpressed in 7–34% of GC cases, particularly in the CIN subtype. This increased expression is associated with activation of the PI3K and MAPK pathways, tumor growth, a poor prognosis, and metastasis. Drugs such as trastuzumab, which target HER2, improve clinical response and survival in HER2-positive cases (Kanayama et al., 2018; Shitara et al., 2020).

Despite advances, the genes that drive CIN-GC remain unclear, as many mutations are linked to other subtypes. CDH1, CTNNA1, and RHOA mutations are more common in GS-GC, and KRAS, PIK3CA, and ARID1A are more common in CIN-GC or EBVaGC (Cancer Genome Atlas Research Network, 2014). Current investigations have linked copy number changes and aneuploidy to particular mutations and clinical factors such as metastasis and treatment efficacy (Smith et al., 2023).

The clinical features of CIN-GC are well understood, allowing for the prediction of tumor progression and the selection of more aggressive treatments. This subtype is associated with a poor prognosis and shorter survival (Díaz del Arco et al., 2018; Tsai et al., 2020), and more than 40% of cases have high PD-L1 expression (Kim et al., 2016). Checkpoint inhibitors such as avelumab, nivolumab, and pembrolizumab have shown promising results in treating CIN-GC (Kang et al., 2017).

In the CIN subtype, defined by genomic disruption, aneuploidy, and recurrent mutations in the TP53 gene, there is no distinct signature of promoter hypermethylation of tumor suppressor genes. This subtype falls into the CIMP-low category (low methylation phenotype), and its carcinogenesis pathway is driven primarily by genetic events. In fact, the main epigenetic event associated with this subtype is its opposite phenomenon, global hypomethylation. This overall decrease in methylation, particularly in repetitive regions of the genome, can promote structural instability of chromosomes and directly contribute to the CIN that is the main characteristic of this subtype (Zeng et al., 2017; Toshima et al., 2023; Usui et al., 2021).

The key genomic regions of interest include (1) long interspersed nuclear elements (LINE-1), which comprise approximately 17% of the human genome. Due to their abundance, LINE-1 methylation serves as a reliable surrogate for global DNA methylation, and studies have shown that LINE-1 hypomethylation in GC is associated with poorer prognosis and increased genomic instability. (2) Alu elements are among the most prevalent repetitive sequences in the genome, and their hypomethylation can disrupt heterochromatin organization, promote chromatin relaxation, and compromise genome stability by increasing DNA vulnerability and impairing repair efficiency. (3) Satellite DNA consists of highly repetitive sequences concentrated in centromeric and pericentromeric regions, with satellite alpha (Satα) being one of the most important types. Since centromeres play a crucial role in accurate chromosome segregation, hypomethylation in these regions can impair centromere function, increase segregation errors, and ultimately lead to aneuploidy, which is a defining feature of the CIN subtype (Ichida et al., 2018; Abe et al., 2022).

ncRNA profile in CIN-GC

There is a lack of significant evidence for the exact role of ncRNAs in CIN-GC, but some studies have suggested their involvement in the development of CIN. miR-22 (Jafarzadeh-Samani et al., 2017), miR-26a (Qiu et al., 2017), miR-28 (Jeddi et al., 2019), and miR-186 (Cao et al., 2016) have been identified in GC, but their precise role in CIN-GC remains undefined. miR-22 inhibits MDC1 and PTEN at the S-phase checkpoint, resulting in chromosomal repair impairment and CIN (Lou et al., 2006). Also, in colon cancer, HER2 activates AKT and miR-22 leads to aneuploidy and CIN by reducing homologous recombination. miR-22 directly suppresses MDC1, disrupting DNA repair and exacerbating CIN. The transcription factor AP-4 regulates this pathway by activating MDC1 and repressing miR-22; consequently, the loss of its function promotes defective DNA repair and increased CIN (Mohapatra et al., 2023). miR-26a has a dual role in cancer. Overexpression of miR-26a causes centrosome and multipolar spindle defects, inducing CIN (Mohapatra et al., 2023; Castellano et al., 2017). This miRNA impairs mitotic checkpoints by repressing CHFR and YWHAE, and restoration of these genes corrects the defects (Mohapatra et al., 2023). miR-28 promotes CIN by inhibiting the translation of MAD2, a crucial mitotic checkpoint protein. miR-28 amplification is associated with loss of the Von Hippel-Lindau (VHL) tumor suppressor, and anti-miR-28 therapy could be beneficial in VHL-deficient cancers. However, due to the sensitivity of MAD2 function, therapeutic targeting requires high precision (Mohapatra et al., 2023). miR-186 amplification is associated with increased CIN, particularly in the presence of arsenite, and disrupts cell division and DNA repair. In HaCaT cells, increased miR-186 induces bipartite, dicentric, and ring chromosomes (Sage et al., 2017). Although the possibility of BUB1 repression by miR-186 is has been suggested, direct evidence is lacking, and the role of BUB1 in CIN remains complex and controversial (Fujibayashi et al., 2020; Wu et al., 2019).

In addition to miRNAs, some lncRNAs such as TERRA, cenRNA, enhancer RNAs, PCAT2, GUARDIN, Ginir, CCAT2, and NORAD are involved in the progression of CIN and various cancers (Mohapatra et al., 2023); however, their exact role in CIN-GC is still not elucidated.

TERRA maintains chromosomal integrity by regulating telomere dynamics. This lncRNA regulates chromatin structure by suppressing telomerase activity and interacting with telomere proteins. Excessive overexpression or depletion of TERRA disrupts telomere homeostasis and induces CIN; overexpression produces telomere shortening and DNA damage, while depletion promotes telomere lengthening and tumorigenesis (Yang et al., 2022; Sagie et al., 2017; Mohapatra et al., 2023). Hence, TERRA is a potential therapeutic target that requires fine-tuning.

CenRNAs constitute a subset of lncRNAs that are characterized by interaction with centromeric proteins such as CENP-A. Elimination of CenRNAs leads to loss of CENP-A and chromosome segregation instability, contributing to CIN. Furthermore, amplification of CENP-A in cancer cells results in the aberrant localization of centromeric proteins and exacerbates genomic instability (Mohapatra et al., 2023; Shrestha et al., 2017).

Ginir is an lncRNA that induces CIN via promoting DNA breaks, mitotic abnormalities, and centrosome duplication. This outcome is mediated via disrupting the interaction between the proteins Cep112 and BRCA1, which modulates centrosome function. The role of Ginir in promoting genomic instability reveals its carcinogenic properties (Panda et al., 2018).

GUARDIN is a p53-induced lncRNA that supports telomere stability and DNA repair through the preservation of TRF2 expression by inhibiting miR-23a and interacting with BRCA1/BARD1. GUARDIN suppression disrupts associated pathways, promotes chromosomal fusion, and impairs DNA repair, resulting in CIN (Yoshino et al., 2021; Hu et al., 2018).

Colon cancer-associated transcript 2 (CCAT2) transcript (at 8q24) is overexpressed in GI cancers, particularly colon cancer, and is associated with poor prognosis (Deng et al., 2023). This lncRNA contributes to tumor growth, metaphase defects, and polyploidy through the upregulation of MYC and regulation of BOP1 (Ling et al., 2013). It also disrupts chromosome-microtubule associations by activating AURKB, thereby enhancing CIN and tumorigenesis (Chen et al., 2020).

NORAD is another lncRNA that is activated following DNA damage and is a vital factor in maintaining GS. The deletion of NORAD triggers mitotic errors such as anaphase bridges and mitotic slippage, consequently resulting in CIN. NORAD is implicated in the maintenance of DNA repair through inhibiting PUMILIO proteins (particularly PUM2). The loss of function in NORAD causes chromosome segregation defects; however, this dysfunction can be corrected by circular RNA containing PUMILIO response elements (Lee et al., 2016; Elguindy and Mendell, 2021).

Some investigations indicate the involvement of NORAD in GC, showing that its inhibition inducing apoptosis and repressing cell proliferation (Raei et al., 2022). Additionally, another study has associated six lncRNAs, including LINC02678, HOXA10-AS, RHOXF1-AS1, AC010789.1, LINC01150, and TGFB2-AS1, with genomic instability in GC (Sun et al., 2021). Despite the potential role of these ncRNAs in the development of CIN, further research is required to elucidate their exact role in CIN-GC.

General biomarkers in GS-GC

GS-GC encompasses approximately 20% of GC cases and is predominantly marked by diffuse, poorly adherent carcinomas. Another feature of GS-GC is mutations in EMT-related genes, such as RHOA and CDH1 mutated forms, which are linked to tumor invasion and peritoneal dissemination (Cancer Genome Atlas Research Network, 2014). RHOA mutations promote cell migration, cytoskeletal remodeling, and therapy resistance. Furthermore, rearrangements involving genes such as CLDN18 with ARHGAP26 or ARHGAP6 are common in this type of tumor, consistent with the diffuse phenotype according to the Lauren classification (Röcken, 2023).

In diffuse-type of GS-GC, overexpression of CLDN18.2 is linked to invasion and increased lymphatic metastasis (Rohde et al., 2019). These changes reflect EMT features, such as decreased cell adhesion in GS-GC, and may facilitate peritoneal metastasis. The reduced expression of CLDN18.2 in HER2-overexpressing tumors suggests a cross-regulation between these two pathways (Moentenich et al., 2020). Targeted therapies such as zolbetuximab in combination with chemotherapy have shown promising results in improving survival (Sahin et al., 2021), highlighting CLDN18.2 as a potential therapeutic target.

Bioinformatics analyses and gene expression studies have suggested crucial signaling pathways in GS-GC development, including H19-mediated Rb-E2F1 signaling and CDK-β-catenin activity. These pathways are activated in the GS subtype, indicating their role in oncogenesis and progression. (Ling et al., 2020). CDH1, CTNNA1, HIST1H1C, RHOA, and TGFBR2 are among the most frequently altered genes in GS-GC patients.

In the GS subtype, epigenetic changes occur in a targeted and focal manner. The most prominent and characteristic event is hypermethylation of the promoter of the tumor suppressor gene CDH1, which encodes the protein E-cadherin. This hypermethylation often acts as a “second hit” alongside genetic mutations in the same gene, leading to its complete silencing, disruption of cell integrity, and promotion of an aggressive phenotype. In contrast, there is evidence of hypomethylation of genes related to the Wnt signaling pathway in this subtype. This suggests that the epigenetic landscape of GS is a combination of targeted silencing of tumor suppressor pathways through hypermethylation and activation of oncogenic pathways through hypomethylation (Toshima et al., 2023; Usui et al., 2021; Cancer Genome Atlas Research Network, 2014).

A study of methylation status across 23 genetic loci in 98 GC samples showed that GS tumors have a distinct epigenetic profile. Unsupervised clustering placed most GS tumors in a group with lower overall methylation levels, and inter-variable clustering identified Wnt-associated loci with the lowest mean methylation score in GS tumors. In particular, SFRP2 and APC were significantly hypomethylated, with a mean score of 1.3 compared with MSI (2.7), EBV (2.1), and CIN (2.4) tumors. This pattern was most pronounced in signet-ring cell carcinoma, which is enriched in the GS subtype (Canale et al., 2020; Toshima et al., 2023). While hypomethylation of proto-oncogenes such as c-MYC and HRAS occurs in gastric carcinogenesis more broadly, the specific hypomethylation of SFRP2 and APC defines the GS subtype and suggests Wnt pathway dysregulation through a distinct epigenetic mechanism.

Among the key genes mentioned above, Claudin 18.2 (CLDN18.2) has been extensively studied as a crucial therapeutic target for GS-GC. The primary treatment targeting CLDN18.2 is the chimeric monoclonal antibody zolbetuximab, which was evaluated in the Phase II FAST trial. This study demonstrated that adding zolbetuximab to first-line EOX chemotherapy, consisting of epirubicin, oxaliplatin, and capecitabine, significantly improved both progression-free survival (hazard ratio [HR] 0.44) and overall survival (HR 0.55) in patients with advanced esophagogastric cancer expressing moderate to strong levels of CLDN18.2. The survival benefit was even greater in the subgroup with the highest CLDN18.2 expression (Röcken, 2023).

Despite conflicting findings on association of CLDN18.2 with clinical features and survival, its restricted expression in normal tissue and consistent tumor presence make it a specific therapeutic target. The FAST trial’s positive results with zolbetuximab confirm CLDN18.2’s potential, with efforts underway to standardize its detection and scoring (Sahin et al., 2021).

ncRNA profile in GS-GC

One of the main features of GS-GC is the EMT process, and ncRNAs associated with EMT could serve a critical role in clarifying its molecular mechanism. For instance, specific miRNAs, such as miR-26a, suppress EZH2, increase E-cadherin, and prevent EMT and invasion (Ma et al., 2016). In contrast, downregulation of miR-200c in GC elevates ZEB1 expression, leading to E-cadherin suppression and promotion of EMT (Kurata et al., 2018). Loss of miR-101 also disrupts E-cadherin function by upregulating EZH2 (Carvalho et al., 2012). However, the effects of miRNAs vary depending on the tumor type; for example, miR-9 enhances EMT by suppressing E-cadherin in esophageal cancer (Song et al., 2014); however, in melanoma, it increases E-cadherin by suppressing SNAIL1 (Liu et al., 2012).

Studies have suggested that reduction in expression of miR-128 and miR-148a in GC cells is mediated by H. pylori, which also increases MMP-3 and MMP-7 (SGC-7901). These enzymes degrade E-cadherin, thereby promoting cancer cell migration and invasion (Costa et al., 2016). Activation of the MMP/E-cadherin pathway in H. pylori-infected cells is linked to downregulation of these miRNAs (Yang et al., 2018).

MiR-148a reduces the invasive properties of GC cells through inhibiting DNMT1 and targeting SMAD2, thereby inhibiting EMT (Wang et al., 2013). Similarly, miR-29b/c, by suppressing DNMT3A, causes CDH1 silencing and promotes metastasis. Downregulation of miR-204 in GC reportedly results in E-cadherin levels elevation and a decrease in EMT markers such as N-cadherin, vimentin, and TWIST through inhibiting TGFBR2 and modulating the TGF-β pathway (Bure et al., 2019).

MiR-5003-3p, miR-217, and miR-9, targeting the 3′ UTR of the CDH1 gene, reduce E-cadherin expression and thereby increase migration, invasion, and EMT in GC. Additionally, miR-544a, by reducing CDH1 and AXIN2, induces EMT through activation of the Wnt pathway and the translocation of β-catenin into the nucleus; this miRNA is proposed as a prognostic marker and candidate as a therapeutic target in metastatic GC (Bure et al., 2019; Yanaka et al., 2015).

Among the miRNAs regulating EMT in GC, miR-376a is associated with advanced disease stages and poor prognosis. miR-381 acts as a tumor suppressor, which inhibits TMEM16A and the TGF-β pathway. miR-128 is downregulated in paclitaxel-resistant cells. Conversely, expression of miRNAs that increase E-cadherin can reverse EMT, reduce invasion, and increase sensitivity to apoptosis, suggesting their vital role in overcoming drug resistance (Bure et al., 2019).

HOX antisense intergenic RNA (HOTAIR) is one of the first lncRNAs identified as an inducer of cancer progression, including GC. By engaging the PRC2 complex, HOTAIR induces epigenetic silencing of miR-34, which activates HGF/c-MET/SNAIL cascades and drives EMT (Liu et al., 2015). Furthermore, HOTAIR promotes the methylation of histone H3K27 in the E-cadherin promoter region, thereby reducing its expression and promoting tumor invasion. Knockdown of HOTAIR reduces PRC2 activity, elevates acetylation, reduces H3K27 methylation, and suppresses EMT (Song et al., 2019).

The lncRNA XLOC_010235 is significantly upregulated in GC tissues and affects EMT and tumor progression by inhibiting SNAIL1. Amplification of XLOC_010235 leads to the elevation of N-cadherin, vimentin, and MMP2/MMP9, while decreasing E-cadherin levels (Bure et al., 2019). Similarly, the lncRNA ZFAS1 is upregulated in tissue, serum, and exosomes of GC patients and promotes EMT through ZEB1 activation and ZEB2 stabilization. Exosomes containing ZFAS1 contribute to metastasis. Inhibition of ZFAS1 reduces invasive markers and elevates E-cadherin levels (Pan et al., 2017; Zhou et al., 2016).

Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is an oncogenic lncRNA, and its amplification in GC leads to the upregulation of E-cadherin and EMT modulation through N-cadherin, ZEB1, and SNAIL. FRLnc1 also promotes cell migration by activating FOXM1 and EMT genes such as TGFβ−1 and TWIST (Bure et al., 2019).

LINC00978 overexpression in GC correlates with cancer progression by promoting EMT. Inhibition of LINC00978 results in increased E-cadherin, suppression of EMT factors such as SLUG and TWIST1, decreased TGF-β/SMAD2, and decreased cell invasion and migration (Fu et al., 2018).

Urothelial carcinoma-associated 1 (UCA1) is amplified in GC and is associated with invasion and metastasis. UCA1 upregulation is observed in response to TGFβ−1. Suppression of UCA1 increases the levels of E-cadherin and ZO-1 and reduces vimentin and SNAIL levels, suggesting its role in promoting EMT (Zhang and Huang, 2018; Ramli et al., 2021).

Taurine upregulated gene 1 (TUG1) overexpression in GC is linked to tumor growth. This lncRNA interacts with the PRC2 complex and represses cell cycle inhibitors such as P21, P15, P16, and P57, leading to increased cell proliferation. TUG1 inhibition reduces GC cell proliferation in vitro and in vivo (Zhang et al., 2016).

LncRNAs commonly modulate this process through sponging EMT-regulating miRNAs. For example, Linc00152, by sponging miR-193b-3p and increasing ETS1 expression, inhibits cell proliferation, migration, and invasion in GC. Also, the lncRNA XIST, as a miR-101 sponge, modulates EZH2 expression in GC cells (Bure et al., 2019).

LncRNA-ATB is implicated in the EMT of GC via the TGF-β/miR-200/ZEB axis. Increased expression of lncRNA-ATB upregulates ZEB1, suppresses miR-200, and induces EMT. Downregulation of lncRNA-ATB upregulates E-cadherin and downregulates mesenchymal markers (ZEB1 and N-cadherin). LncRNA SNHG6 is also upregulated in GC and leads to invasion and metastasis. SNHG6 inactivation increases p21, inhibiting EMT by downregulating EZH2 and activating the c-Jun N-terminal kinases (JNK) pathway. Acting as a ceRNA, SNHG6 sponges miR-101-3p and enhances ZEB1 to increase invasive properties (Bure et al., 2019).