LncRNA SLC16A1-AS1 cooperates with NSUN2 to stabilize GRP78 mRNA via m5C modification in gastric cancer

Summary

Long noncoding RNAs (lncRNAs) play crucial roles in regulating chromatin dynamics and gene expression, and their dysregulation is closely linked to tumorigenesis. However, their specific functions in gastric cancer (GC) remain poorly understood. Here, we found that lncRNA solute carrier family 16 member 1 antisense RNA 1 (SLC16A1-AS1) was markedly overexpressed in GC through integrated RNA sequencing (RNA-seq) analysis and validation in clinical tissues. High SLC16A1-AS1 expression correlated with advanced cancer stage, greater invasion depth, and poorer patient prognosis. Functional assays showed that SLC16A1-AS1 overexpression promoted GC cell proliferation, whereas knockdown inhibited proliferation in vitro and in vivo. Mechanistically, SLC16A1-AS1 interacted with the 5-methylcytosine (m5C) methyltransferase NOL1/NOP2/SUN (NSUN2), enhancing m5C modification of GRP78 mRNA, which stabilized the transcript and increased GRP78 protein levels. Rescue experiments demonstrated that GRP78 overexpression reversed the proliferation-inhibitory effect of SLC16A1-AS1 depletion. These findings reveal that SLC16A1-AS1 drives GC cell proliferation via NSUN2-mediated m5C modification of GRP78 mRNA, suggesting a potential target for GC diagnosis and therapy.

Graphical abstract

Highlights

• •SLC16A1-AS1 is upregulated in GC and correlates with poor prognosis
• •SLC16A1-AS1 interacts with NSUN2 to stabilize GRP78 mRNA via m5C
• •SLC16A1-AS1 drives GC proliferation in vitro and tumor growth in vivo
• •GRP78 overexpression rescues growth defects from SLC16A1-AS1 loss

Health sciences

Introduction

Gastric cancer (GC) is estimated to cause over one million new cases annually, making it the fifth most common cancer, accounting for 5.6% of all new cancer cases. It also ranks fourth in cancer-related mortality, responsible for 7.7% of all cancer deaths.¹ GC is highly heterogeneous both molecularly and phenotypically. Despite improvements in endoscopy and advances in diagnostic and treatment strategies, the early prognosis of GC remains impractical.² Moreover, there are currently no effective biomarkers to guide chemotherapy selection for advanced GC. Therefore, elucidating the detailed molecular mechanisms driving GC initiation and progression is still essential.

Long noncoding RNAs (lncRNAs), which are transcripts longer than 200 nucleotides, play crucial roles in epigenetic regulation of GC. LncRNAs are involved in epigenetic regulation through chromatin remodeling, transcriptional activation or repression, post-transcriptional modification of mRNA, acting as molecular sponges to sequester miRNAs from their natural mRNA targets, and other mechanisms.^3,4,5 Recent advancements have identified many lncRNAs that regulate GC progression. Multiple lncRNAs have been validated as diagnostic biomarkers and prognostic predictors in GC, underscoring the urgent need to explore novel lncRNA-mediated mechanisms for GC therapy.^{6,7,8,9,10,11} For example, LINC00152 is involved in GC cell epithelial-to-mesenchymal transition (EMT), thus facilitating the migration and invasion of GC.⁶ The SNAI2-ELF3-AS1 regulates ELF3 expression at transcriptional and post-transcriptional levels and drives GC metastasis by maintaining SNAI2 overexpression.⁷ HAGLROS overexpression contributes to GC development and poor prognosis, while GAS5 acts as a tumor suppressor that inhibits tumor progression.^8,9 In this study, we identified a previously unappreciated differentially expressed lncRNA SLC16A1-AS1 (solute carrier family 16 member 1 antisense RNA 1) in GC tissues by RNA sequencing (RNA-seq). We further investigated its expression, clinical significance, and underlying mechanisms in GC. Additionally, SLC16A1-AS1 has been reported to play roles in a variety of cancers, including glioblastoma,¹² oral squamous cell carcinoma,^13,14 breast cancer,^15,16 lung cancer,^17,18 hepatocellular carcinoma,^19,20 renal cell carcinoma,²¹ bladder cancer,²² and cervical squamous cell carcinoma.²³ Despite the large amounts of reports, the specific role of SLC16A1-AS1 in GC remains unclear.

Abnormalities in epigenetic regulation, such as histone modification, DNA methylation, and RNA methylation, are key markers of tumor initiation, progression, and recurrence. 5-methylcytosine (m5C) is a well-known and conserved RNA modification, which occurs extensively across various types of eukaryotic RNA, including rRNA, lncRNA, tRNA, and mRNA.²⁴ Since the first localization of N6-methyladenosine (m6A) and its widespread presence in mRNA, more than 100 different types of RNA modifications have been discovered.²⁵ As a reversible epigenetic modification, the m5C modification of RNA affects the fate of the modified RNA molecule and plays important roles in various biological processes, including transcriptional regulation, RNA export, RNA stability, and protein synthesis.^26,27,28 RNA m5C modification is critically controlled by three types of proteins, namely the methyltransferases (writers), the demethylases (erasers), and the binding proteins (readers), respectively.²⁹ m5C in RNAs is catalyzed by NOL1/NOP2/SUN (NSUN) protein family, including NSUN1-7 and DNA methyltransferase (DNMT) homolog DNMT2.³⁰ Normal eukaryotic cells express low levels of NSUN1, NSUN2, and NSUN5 and high amounts of NSUN3, NSUN4, NSUN6, and NSUN7.³¹

The 78-kDa glucose-regulated protein (GRP78) encoded by the HSPA5 gene, also known as binding immunoglobulin protein (BiP), is a member of the heat shock protein 70 (HSP70) protein family. GRP78, as an endoplasmic reticulum (ER) chaperone, is firmly established to play a crucial role in the folding and maturation of newly synthesized membrane-bound or secretory proteins.³² Additionally, by interacting with transmembrane ER stress sensors, GRP78 plays a crucial regulatory role in the unfolded protein response (UPR). This evolutionarily conserved cellular mechanism allows cells to adapt and respond to proteotoxic stress, which is commonly encountered in oncogenic processes, metabolic disorders, and neurological conditions.^33,34,35 GRP78 has an essential role in the development of cancers, obesity, and virus infection, including the pandemic SARS-CoV-2.^36,37 Overall, GRP78 expression is positively correlated with tumor malignancy. In GC, GRP78 is recognized as a reliable and effective marker for predicting aggressive behavior and poor prognosis in patients.³⁸

In this study, we focused on the identification of a novel mechanism by which SLC16A1-AS1 promotes the proliferation of GC cells by facilitating NSUN2-dependent m5C modification of GRP78 mRNA. This finding not only elucidates a new layer of regulatory complexity in GC but also highlights potential therapeutic targets for intervention.

Results

SLC16A1-AS1 is highly expressed in GC and correlated with poor patient prognosis

To identify essential regulatory RNAs involved in GC, we analyzed The Cancer Genome Atlas (TCGA) database and two GEO datasets (GSE58828 and GSE192468). We identified 60 overlapping upregulated lncRNAs that were differentially expressed in GC (Figure 1A; Table S1) and narrowed them down to 9 lncRNAs with reported cancer associations by additional literature review and searches in GeneCards database.

Figure 1.

SLC16A1-AS1 was elevated and linked to poor prognosis

(A) Venn diagram illustrating the overlap of differentially highly expressed lncRNAs in TCGA, GSE192468, and GSE58828 datasets, with highlighted selected candidates.

(B) Relative expression levels of SLC16A1-AS1 in various GC cell lines. The experiment was conducted in triplicate.

(C) Relative expression levels of SLC16A1-AS1 in GC tissues (n = 80) compared to adjacent normal tissues (n = 80).

(D) Relative expression levels of SLC16A1-AS1 in different clinical stages (I/II vs. III/IV).

(E) Correlation between SLC16A1-AS1 expression and invasion depth.

(F) Survival curves comparing patients with high versus low levels of SLC16A1-AS1. Error bars represent mean ± SD. ∗p < 0.05, ∗∗p < 0.01.

We then verified the expression of these 9 candidates by quantitative reverse-transcription PCR (RT-qPCR) in GC cell lines. Among them, only SLC16A1-AS1 and BANF1P2 were consistently overexpressed in all five GC cell lines compared to normal gastric epithelial cells (GES1) (Figure 1B; Figure S1). In the literature, BANF1P2 has been poorly characterized, and its expression profile in cancers is unclear; in contrast, SLC16A1-AS1 has been linked to several malignancies, including hepatocellular carcinoma, glioblastoma, and renal cell carcinoma.³⁹ The role of SLC16A1-AS1 in GC remains unexplored, which is investigated in the following experiments.

We quantified SLC16A1-AS1 expression in 80 paired GC and adjacent non-tumor tissues. SLC16A1-AS1 expression was markedly increased in most tumors (Figure 1C) and associated with aggressive clinicopathological features, including larger tumor size, deeper invasion, and advanced TNM stage (Figures 1D and 1E). Kaplan-Meier analysis revealed that high SLC16A1-AS1 expression was significantly associated with poor patient survival (Figure 1F). These findings suggest that SLC16A1-AS1 is highly expressed in GC and may play a pivotal role in tumor progression. The detail information of the GC patients enrolled in this study is listed as Table S2.

SLC16A1-AS1 promotes GC cell proliferation

To investigate the role of SLC16A1-AS1 in GC proliferation, we investigated its expression in GC cells. SLC16A1-AS1 was knocked down in HGC27 and AGS cells using small interfering RNAs (siRNAs)/short hairpin RNAs (shRNAs) and overexpressed in MKN74 cells using pcDNA3.1-SLC16A1-AS1. RT-qPCR confirmed efficient modulation of its expression (Figure 2A; Figure S2).

Figure 2.

SLC16A1-AS1 promoted malignant proliferation of GC cells

(A) Verification of the knockdown and overexpression efficiencies of SLC16A1-AS1 in GC cells.

(B–D) SLC16A1-AS1 knockdown inhibited cell viability, as shown by CCK-8 assays in HGC27 and AGS cells. Conversely, SLC16A1-AS1 overexpression promoted the phenotype, as demonstrated by CCK-8 assays in MKN74 cells.

(E) SLC16A1-AS1 knockdown reduced cell proliferation, as evidenced by EdU incorporation assays in HGC27 cells.

(F) SLC16A1-AS1 knockdown inhibited colony formation as demonstrated by colony formation assays in HGC27 cells.

(G) SLC16A1-AS1 knockdown inhibited cell proliferation as indicated by EdU incorporation assays in AGS cells.

(H) SLC16A1-AS1 knockdown inhibited colony formation as demonstrated by colony formation assays in AGS cells.

(I) SLC16A1-AS1 overexpression promoted cell proliferation as indicated by EdU incorporation in MKN74 cells.

(J) SLC16A1-AS1 overexpression promoted colony formation in MKN74 cells. All experiments were performed three times. Error bars represent mean ± SD. Scale bars, 50 μm.

CCK8 assays revealed that SLC16A1-AS1 knockdown significantly inhibited cell viability in HGC27 and AGS cells, whereas overexpression promoted viability in MKN74 cells (Figures 2B–2D). EdU assays showed similar trends, with reduced proliferation after knockdown and increased proliferation after overexpression (Figures 2E, 2G, and 2I). Colony formation assays corroborated these results, showing decreased colony numbers upon knockdown and increased colony formation upon overexpression (Figures 2F, 2H, and 2J). Collectively, these data demonstrate that SLC16A1-AS1 promotes GC cell proliferation.

NSUN2 interacts with SLC16A1-AS1

Many nuclear lncRNAs regulate gene expression by interacting with epigenetic modifiers such as histone methyltransferases,⁴⁰ DNMTs,⁴¹ or RNA methylation factors.⁴² Coupling with m5C methylation regulators has emerged as a novel mechanism by which lncRNAs influence RNA stability and translation.⁴³ We hypothesized that SLC16A1-AS1 might promote GC progression by recruiting or modulating m5C-modifying enzymes.

Nuclear-cytoplasmic fractionation showed that SLC16A1-AS1 predominantly localized to the nucleus in HGC27 and AGS cells (Figures 3A and 3B). Using the catRAPID protein-RNA interaction prediction tool, we identified NSUN2, an m5C methyltransferase, as a candidate interactor (interaction propensity score: 58.55; Z score: 2.19) (Figure 3C). RNA pull-down confirmed that biotin-labeled SLC16A1-AS1 specifically pulled down NSUN2 but not control RNA (Figures 3D and 3E). Consistently, RNA immunoprecipitation (RIP)-qPCR with an NSUN2 antibody enriched SLC16A1-AS1 compared with IgG control (Figure 3F). We next performed co-localization experiments using immunofluorescence (IF) for NSUN2 together with RNA-fluorescence in situ hybridization (FISH) for SLC16A1-AS1. The results show that both SLC16A1-AS1 and NSUN2 are predominantly localized in the nucleus, indicating their functional interaction in this compartment. (Figure 3G). Furthermore, truncation mapping showed that deletion of nucleotides 1462–1522 in SLC16A1-AS1 or amino acids 168–369 in NSUN2 abolished the interaction (Figures 3H–3J), confirming direct binding. In clinical samples, NSUN2 expression was elevated in GC tissues and positively correlated with SLC16A1-AS1 expression (Figures 3K and 3L). These findings show that SLC16A1-AS1 is a nuclear lncRNA that directly binds NSUN2 through defined interaction regions. Their nuclear co-localization and correlated expression in GC tissues indicate that the SLC16A1-AS1/NSUN2 axis plays a functional role in GC.

Figure 3.

SLC16A1-AS1 interacted with the m5C writer NSUN2

(A and B) The subcellular localization of SLC16A1-AS1 in HGC27 and AGS cells was detected by subcellular fractionation assay.

(C) Interaction between SLC16A1-AS1 and NSUN2 was predicted using the RNA-protein interaction prediction tool available at http://service.tartaglialab.com/page/catrapid_group.

(D and E) RNA pull-down assay was performed to verify the interaction between SLC16A1-AS1 and NSUN2 (100 kDa) in HEK293T cells by silver staining and western blotting.

(F) RIP assay was conducted to validate the interaction between SLC16A1-AS1 and NSUN2 using an NSUN2 antibody.

(G) IF and RNA-FISH were performed to evaluate the co-localization of NSUN2 and SLC16A1-AS1.

(H) Truncated forms of NSUN2 (deleted 168–369 aa) and SLC16A1-AS1 (deleted 1,462–1,522 bp) were cloned into their respective plasmids.

(I and J) RIP assay was performed to detect the enrichment of SLC16A1-AS1 in vector control, SLC16A1-AS1 (I) or NSUN2 (J) overexpression (OE), and SLC16A1-AS1 or NSUN2 OE-mutant samples using NSUN2 antibody.

(K) The relative expression level of NSUN2 was assessed in GC tissues (n = 80) compared to adjacent normal tissues (n = 80), showing significantly higher NSUN2 expression in GC tissues.

(L) The correlation between SLC16A1-AS1 and NSUN2 was assessed in human GC tissues. All experiments were performed three times. Error bars represent mean ± SD. Scale bars, 50 μm.

SLC16A1-AS1 facilitates NSUN2-mediated m5C modification of GRP78 mRNA

To uncover downstream targets, we performed RNA-seq after SLC16A1-AS1 knockdown and RNA bisulfite sequencing (Bis-seq) to map m5C sites in GC tissues (Figure 4A). Kyoto Encyclopedia of Genes and Genes (KEGG) analysis revealed “Protein processing in the endoplasmic reticulum” as a top enriched pathway (Figure 4B). Integration of RNA-seq and m5C data identified GRP78 as the only gene positively correlated with SLC16A1-AS1 and harboring m5C sites (Figures 4C and 4D).

Figure 4.

SLC16A1-AS1 regulated NSUN2-mediated m5C modification of GRP78 mRNA

(A) Experimental flowchart for analysis of RNA-seq and m5C Bis-seq.

(B) Knockdown of SLC16A1-AS1 in AGS cells was performed for RNA-seq and KEGG pathway analysis.

(C) HSPA5 (GRP78) was screened by m5C Bis-seq and KEGG pathway enrichment analysis.

(D) The m5C peaks of GRP78 were visualized using IGV from m5C Bis-seq data.

(E) The western blot experiment verified the knockdown efficiency of NSUN2 sgRNAs by CRISPR technology.

(F) The m5C-MeRIP assay assessed the enrichment of GRP78 in HEK293T cells with NSUN2 knockdown compared to control.

(G) The m5C-MeRIP assay evaluated the enrichment of GRP78 in HEK293T cells with SLC16A1-AS1 knockdown compared to control.

(H) Knockdown of NSUN2 resulted in decreased protein expression of GRP78.

(I and J) GRP78 mRNA stability was assessed by RT-qPCR in AGS cells treated with actinomycin D (ActD) following SLC16A1-AS1 knockdown, with or without NSUN2 transfection.

(K) GRP78 mRNA expression was measured by RT-qPCR in HGC27 and AGS cells with SLC16A1-AS1 knockdown, and in MKN74 cells with SLC16A1-AS1 OE.

(L) Protein expression of GRP78 was analyzed by western blot following SLC16A1-AS1 knockdown or OE.

(M) The efficiency of SLC16A1-AS1 knockdown or overexpression was validated. All experiments were performed three times. Error bars represent mean ± SD.

Given NSUN2’s role as an m5C “writer”^26,44,45, we used methylated RIP-qPCR (MeRIP-qPCR) to assess GRP78 mRNA methylation. GRP78 mRNA was enriched by an m5C antibody, and NSUN2 knockdown reduced this enrichment (Figures 4E and 4F). Similarly, SLC16A1-AS1 knockdown decreased GRP78 mRNA m5C enrichment (Figure 4G). Functionally, NSUN2 knockdown reduced GRP78 protein levels (Figure 4H). Actinomycin D chase assays showed that SLC16A1-AS1 knockdown accelerated GRP78 mRNA decay, which was rescued by NSUN2 overexpression (Figures 4I and 4J; Figure S3). SLC16A1-AS1 knockdown reduced, and overexpression increased, GRP78 expression at both mRNA and protein levels (Figures 4K–4M). These data suggest that SLC16A1-AS1 interacts with NSUN2 to enhance m5C modification of GRP78 mRNA, stabilizing it and promoting GC malignancy.

GRP78 overexpression rescued the inhibitory effect of SLC16A1-AS1 knockdown

To further validate GRP78 as a downstream effector of SLC16A1-AS1, we supplemented GRP78 utilizing HGC27 and AGS cells in which SLC16A1-AS1 expression was silenced (Figures 5A and 5B). CCK-8, EdU incorporation, and colony formation assays demonstrated that GRP78 overexpression significantly restored the proliferative capabilities of both HGC27 and AGS cells (Figures 5C–5L). These results underscore the pivotal role of GRP78 in mediating the effects of SLC16A1-AS1 on cellular proliferation.

Figure 5.

Overexpression of GRP78 rescued the phenotype of GC cells induced by SLC16A1-AS1 silencing

(A and B) GRP78 protein expression was assessed by western blot in HGC27 and AGS cells treated with siSLC16A1-AS1 plus vector or siSLC16A1-AS1 plus GRP78-OE.

(C and D) Cell viabilities were rescued by GRP78 OE, as indicated by CCK8 assays in HGC27 and AGS cells treated with siSLC16A1-AS1 plus GRP78-OE.

(E–H) EdU incorporation assays demonstrated that GRP78 OE rescued cell proliferation in HGC27 and AGS cells treated with siSLC16A1-AS1.

(I–L) Colony formation was rescued by GRP78 OE in HGC27 and AGS cells treated with shSLC16A1-AS1 plus GRP78-OE. Three independent experiments were performed. Error bars represent mean ± SD. Scale bars, 50 μm.

SLC16A1-AS1 promotes the tumorigenesis of GC xenografts through GRP78-mediated regulation

To validate the oncogenic role of SLC16A1-AS1 in GC progression, we established the xenograft tumor model in nude mice. The in vivo results corroborated in vitro observations, demonstrating that SLC16A1-AS1 knockdown significantly inhibited tumor growth, as evidenced by reduced tumor volume and weight compared to the control group (Figures 6A–6C). Notably, GRP78 overexpression not only enhanced tumor growth parameters but also effectively rescued the tumor-suppressive effects induced by SLC16A1-AS1 knockdown.

Figure 6.

SLC61A1-AS1 knockdown inhibited the growth of GC xenograft tumors in nude mice, while GRP78 overexpression alleviated this effect

(A) Comparison of xenograft sizes across experimental groups: shSLC16A1-AS1 vs. shNC; vector vs. GFP-GRP78; shSLC16A1-AS1+vector vs. shSLC16A1-AS1+GFP-GRP78.

(B and C) The volume and weight of xenografts in the experimental groups after injection of GC cells in male nude mice.

(D) IHC with Ki-67 and H&E staining of xenograft tumor sections from both the experimental and control groups to assess tumor cell proliferation. H&E staining of tumor tissue exhibits the following histopathological features: nuclear atypia and marked basophilia (black arrows), loosely arranged cells and occasional mitotic figures (blue arrows), extensive necrosis (red arrows), characterized by nuclear fragmentation/lysis and eosinophilic homogeneous cytoplasmic staining, and peripheral fibrous encapsulation, with connective tissue infiltrating the parenchyma in an intersecting pattern (green arrows).

(E–G) Efficiencies of SLC16A1-AS1 knockdown and GRP78 overexpression. Error bars represent mean ± SD. Scale bars, 50 μm.

To further characterize the proliferative capacity of the xenograft tumors, we performed immunohistochemical (IHC) analysis of Ki-67, a well-established proliferation marker. Tumors from SLC16A1-AS1 knockdown mice exhibited markedly reduced Ki-67 expression compared with controls, while GRP78 overexpression significantly increased Ki-67 positivity and reversed the proliferation inhibition caused by SLC16A1-AS1 depletion (Figure 6D, up). Furthermore, H&E staining revealed that the tumor tissue of the shNC group exhibited nuclear atypia and strong basophilia, characterized by nuclear fragmentation or dissolution and eosinophilic homogeneous staining of the cytoplasm. These pathological abnormalities were alleviated in the SLC16A1-AS1 knockdown group but were exacerbated when GRP78 was overexpressed (Figure 6D, lower). The efficiencies of SLC16A1-AS1 knockdown and GRP78 overexpression were confirmed as shown in Figures 6E–6G.

These comprehensive in vivo findings demonstrate that SLC16A1-AS1 plays a crucial role in promoting GC tumorigenesis, potentially through regulation of the GRP78-mediated proliferation pathway. The rescue experiments further substantiate the functional relationship between SLC16A1-AS1 and GRP78 in driving GC progression.

Discussion

Emerging evidence has suggested the significant roles of lncRNAs in development, progression, metastasis, and chemoresistance of GC.^46,47,48 The present study investigated the functional role and molecular mechanisms underlying the oncogenic potential of SLC16A1-AS1 in GC. Our findings, derived from a range of experimental approaches both in vitro and in vivo, underscore the significance of SLC16A1-AS1 in GC progression. SLC16A1-AS1 was consistently found to be highly expressed in GC tissues compared with adjacent non-tumorous tissues, consistent with previous reports implicating its upregulation in various cancers.³⁹ This overexpression was found to be correlated with aggressive clinicopathological features, including larger tumor size, deeper invasion, and advanced TNM staging, suggesting its potential as a biomarker for disease severity and prognosis (Figure 1). These observations highlight the clinical relevance of SLC16A1-AS1 in GC. Although SLC16A1-AS1 is frequently up-regulated and acts oncogenically in the majority of malignancies, context-specific cues from the tumor microenvironment, distinct genetic backgrounds, and divergent epigenetic rewiring can switch its role to a tumor-suppressive one in select breast cancer subtypes.¹⁶ These observations underscore the functional plasticity of SLC16A1-AS1 and highlight the need to evaluate its biological role within precise cellular and clinical contexts.

Previous studies have shown that lncRNAs exert their biological functions through diverse mechanisms, including interactions with DNA, RNA, and proteins. LncRNAs are involved in various cancer-related pathological processes, such as transcriptional regulation, mRNA splicing, RNA modification, mRNA degradation, and translation.^40,46,49 For example, FGD5-AS1 promoted GC progression by stabilizing the YBX1 protein level.⁴⁹ The present study demonstrated that SLC16A1-AS1 promoted GC cell proliferation both in vitro and in vivo. Knockdown of SLC16A1-AS1 inhibited cell viability, proliferation, and colony formation in both HGC27 and AGS cells, while its overexpression enhanced these malignant phenotypes in MKN74 cells. Furthermore, in the xenograft mouse model, deficiency of SLC16A1-AS1 led to reduced tumor growth and decreased Ki-67 expression, further substantiating its role in tumorigenesis (Figures 2 and 6).

Mechanistically, we identified NSUN2 as an interacting partner of SLC16A1-AS1. NSUN2 is known for its role as an RNA methyltransferase that catalyzes m5C modification on RNA molecules, thereby influencing RNA export, stability, and translation efficiency.^26,50,51 Our data showed that NSUN2 was highly expressed in GC (Figure 3K), consistent with findings from other studies.⁵² Furthermore, our data demonstrated a physical interaction between SLC16A1-AS1 and NSUN2, which was validated through RNA pull-down, RIP-qPCR assays, and IF-FISH (Figures 3C–3J). Importantly, we observed a positive correlation between expression levels of SLC16A1-AS1 and NSUN2 in GC tissues (Figure 3L), suggesting a coordinated regulation. This finding is consistent with the role of many lncRNAs that interact with proteins to regulate gene expression. For instance, HOTAIR binds to the polycomb repressive complex 2 (PRC2), leading to H3K27 trimethylation and the repression of HOX loci gene expression, which reprograms chromatin states to promote cancer metastasis.⁵³ Concurrently, HOTAIR recruits LSD1 (lysine-specific demethylase 1), which demethylates H3K4, a mark associated with active transcription. This dual recruitment results in gene silencing, particularly at the HOXD locus.⁵⁴ Thus, we conclude that SLC16A1-AS1 binds to NSUN2 and modulates its function on RNA modification.

Further investigations using m5C Bis-seq, RNA-seq, and subsequent validation revealed that SLC16A1-AS1 facilitated NSUN2-mediated m5C modification of GRP78 mRNA. GRP78, a key chaperone protein involved in the UPR, is crucial for maintaining protein homeostasis in the ER and is implicated in cancer cell survival and proliferation.^36,37 Our study showed that SLC16A1-AS1 knockdown reduced m5C modification of GRP78 mRNA, destabilizing it and leading to decreased GRP78 expression at both mRNA and protein levels (Figures 4I–4M). Furthermore, the reduction of proliferation caused by SLC16A1-AS1 knockdown could be rescued by GRP78 overexpression, highlighting GRP78 as a key downstream effector of SLC16A1-AS1 in promoting GC cell proliferation (Figures 5 and 6). These findings suggest that the interaction between SLC16A1-AS1 and NSUN2 is crucial for the methylation and stabilization of GRP78 mRNA in GC. Specifically, NSUN2, as a methyltransferase, appears to be involved in modifying the m5C methylation marks on GRP78 mRNA, with this interaction and subsequent methylation changes potentially influencing the development and progression of GC (schematic diagram of the mechanism shown in Figure 7).

Figure 7.

Mechanism of SLC16A1-AS1 in promoting GC proliferation via NSUN2-mediated m5C modification of GRP78 mRNA

In GC cells, SLC16A1-AS1 interacts with NSUN2 (an RNA methyltransferase) to form an RNA-protein complex, which enhances m5C methylation of GRP78 mRNA. Elevated SLC16A1-AS1 expression increases m5C modification levels on GRP78 mRNA, thereby stabilizing GRP78 transcripts and upregulating GRP78 expression. Consequently, this regulatory axis drives GC cell proliferation, highlighting a critical oncogenic role of the SLC16A1-AS1/NSUN2/GRP78 pathway in tumor progression (created with BioRender).

The malignant proliferation of GC cells depends on aberrant lactate metabolism. MCT1, encoded by SLC16A1, maintains intracellular pH balance and supports tumor metabolic reprogramming by mediating lactate efflux, thereby facilitating GC proliferation and invasion.^55,56,57 Our study shows that SLC16A1-AS1 promotes GC proliferation through the NSUN2/GRP78 axis, and previous research indicates that it can act as a co-activator of E2F1 to upregulate SLC16A1/MCT1.²² Thus, SLC16A1-AS1 may enhance GC malignancy through a dual mechanism, promoting lactate metabolism via MCT1 and stabilizing GRP78 to alleviate ER stress.

Recent studies have reported that SLC16A1/MCT1 overexpression correlates with recurrence, poor prognosis, and deeper invasion in GC,⁵⁵ while SLC16A1 silencing reduces lactate efflux and invasion, effects partially reversed by SLC16A1-AS1 overexpression.⁵⁷ Furthermore, crosstalk between MCT1-mediated metabolism and GRP78-regulated ER stress may contribute to tumor microenvironment formation and immune escape. Together, these findings reinforce that SLC16A1/MCT1 and SLC16A1-AS1 synergistically drive GC malignancy.

Overall, our findings provide novel insights into the oncogenic role of SLC16A1-AS1 in GC through its regulation of NSUN2-mediated m5C modification of GRP78 mRNA. This study not only advances our understanding of the molecular mechanisms driving GC progression but also highlights potential therapeutic targets for combating this aggressive malignancy. Future studies could explore clinical utility of targeting the SLC16A1-AS1/NSUN2/GRP78 axis in GC treatment strategies, aiming to improve patient outcomes.

Limitations of the study

Despite the meaningful findings in this study, there are some limitations that warrant consideration. First, the clinical sample size is small (80 pairs) and from a single center, lacking population diversity and comprehensive clinical outcome data. Second, the mechanistic research is incomplete, with unclear details of SLC16A1-AS1/NSUN2 interaction, unelucidated GRP78 downstream pathways, and no exploration of crosstalk with the tumor microenvironment. Third, experimental models are limited, including only a few GC cell lines and immunodeficient mouse subcutaneous xenografts that fail to simulate natural tumor progression. Further investigations will be carried out based on the findings of this study.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jinfei Chen (jinfeichen@sohu.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •RNA-seq data: the dataset generated during this study is available in the NCBI Gene Expression Omnibus (GEO) repository under accession number GEO: GSE311114 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE311114).
• •Bis-seq data: the dataset generated during this study is available in the GEO under accession number GEO: GSE310529 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE310529).
• •This paper does not report original code.
• •Western blot data: all original, uncropped western blot images are available as supplemental information.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.

Acknowledgments

All authors thank Dr. Shuwen Cheng for the annotation and analysis of the GEO and TCGA datasets and thank Cloud-Seq Biotech Co. Ltd. (Shanghai, China) for the RNA-BisSeq service and subsequent bioinformatics analysis. We also thank the Special Fund for the Discipline Cluster of Oncology of Wenzhou Medical University (z1-2023001), the National Natural Science Foundation of China (nos. 32270649, 82272701, 82273458, 82070804, and 81972626), and the Start-up Fund for the Recruited Talents of the First Affiliated Hospital of Wenzhou Medical University (2021QD025) for financial support.

Author contributions

B.Z., W. Zhao, and Y.J. performed most of experiments and statistical analysis. W.S. and Yong Li conducted bioinformatics analysis. Q.Z., Yue Li, J.W., and D.C. collected clinical tissue and performed the correlated analysis. K.R. and W. Zhang guided experimental techniques. Y.C. and L.Y. analyzed the clinical data. J.C., F.Y., and X.J.W. supervised the study, designed the experiments, and wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
NSUN2	proteintech	Cat#20854-1-AP; RRID: AB_10693629
GAPDH	proteintech	Cat#60004-1-Ig; RRID: AB_2107436
GRP78	proteintech	Cat#11587-1-AP; RRID: AB_2119855
β-actin	proteintech	Cat#66009-1-Ig; RRID: AB_2687938
GFP	abcam	Cat#ab290; RRID: AB_303395
Critical commercial assays
PrimeScript™ RT Reagent Kit (Perfect Real Time)	Takara	Cat# RR037A
Cell Counting Kit-8	Beyotime	Cat# C0038
BeyoClick™ EdU Cell Proliferation Kit	Beyotime	Cat# C0075S
the PARIS Kit	Invitrogen	Cat# AM1921
Fluorescent In Situ Hybridization Kit	RiboBio	Cat# C10910
in vitro transcription kit	Thermo Fisher Scientific	Cat# AM1333
BCA assay kit	Thermo Fisher Scientific	Cat#A55864
ClonExpress Ultra One Step Cloning Kit V2	Vazyme	Cat#C116-01
NEBNext® rRNA Depletion Kit	NEB	Cat#E7850L
Biological samples
Human GC tissues	Nanjing First Hospital	This paper
Chemicals, peptides, and recombinant proteins
Trizol reagent	Invitrogen	Cat# 15596026
Fetal bovine serum (FBS)	Gibco	Cat# A5670701
RPMI-DMEM	Gibco	Cat#6125276
RPMI-1640	Gibco	Cat#6125258
DMEM/F12	Gibco	Cat#6125206
Opti-MEM	Gibco	Cat#31985070
PBS	Gibco	Cat#70011044
Lipofectamine 2000	Invitrogen	Cat#11668019
Pierce™ Streptavidin Magnetic Beads	Thermo Fisher Scientific	Cat# 88817
4% paraformaldehyde (PFA)	Beyotime	Cat#P0099
Triton X-100	Beyotime	Cat#ST1723
Proteinase K	Sigma	Cat#P2308
Polybrene	Sigma	Cat#TR-1003
Deposited data
RNA-seq data	GEO:GSE311114	https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE311114
Bis-seq data	GEO:GSE310529	https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE310529
Experimental models: Cell lines
HEK-293T	Chinese Academy of Sciences
GES1	Chinese Academy of Sciences
AGS	Chinese Academy of Sciences
HGC27	Chinese Academy of Sciences
MKN28	Chinese Academy of Sciences
KMN74	Chinese Academy of Sciences
SNU1	Chinese Academy of Sciences
Experimental models: Organisms/strains
BALB/c nude mice (BALB/c-nu/nu)	Nanjing Medical University
Oligonucleotides
siSLC16A1-AS1-1:	RiboBio	CCCTGTCTGACGAGCTCTA
siSLC16A1-AS1-2:	RiboBio	CTCAGGAAAATGCCGGACT
siSLC16A1-AS1-3:	RiboBio	GATTTCAAGCACACCAACA
Recombinant DNA
shSLC16A1-AS1-1:CCGGAGAACAAGAA GGCCCAGTTTGCTCGAGCAAACTGG GCCTTCTTGTTCTTTTTTG
shSLC16A1-AS1-2:CCGGTAGTGCAGCC AGTCCGCTAAACTCGAGTTTAGCG GACTGGCTGCACTATTTTTG
NSUN2 (forward: 5’- CACCGGCTGCGGA TTGCAACACGCG-3’, reverse: 5’- AAACC GCGTGTTGCAATCCGCAGCC-3’),
NSUN2 (forward: 5’- CACCGGGATGCC TGGAATCACACAG-3’, reverse: 5’- AA ACCTGTGTGATTCCAGGCATCCC-3’),
NSUN2 (forward: 5’- CACCGGAAGAAA AGGCAGCTCTACA-3’, reverse: 5’- AAACTGTAGAGCTGCCTTTTCTTCC-3’),
Software and algorithms
GraphPad Prism version 10.4.0	GraphPad Software
Image J	NIH
SPSS 26.0	IBM SPSS

Experimental model and study participant details

Ethics approval and consent to participate

The study involved 80 human participants, aged between 42 and 88 years, the detail clinical information in Table S2, with surgical treatment dates between 2013 and 2015 at Nanjing First Hospital of Jiangsu Province in China. Tumor tissues and adjacent non-cancerous tissues were collected during surgery and immediately preserved in liquid nitrogen for rapid freezing, then storaged at -80°C refrigerator. Patients were diagnosed by pathological biopsy, and none had received local or systemic treatment prior to surgery. All procedures performed in studies involving human participants were in accordance with the ethical standards of Nanjing medical University(No.2019-726). Written informed consent was obtained from all patients prior to sample collection.

Animals

Immune-deficient, 4-week-old BALB/c male nude mice were acquired from Nanjing Medical University and maintained under rigorous specific pathogen-free (SPF) conditions. Stable AGS cell lines were generated 3 groups as follows: (1) shSLC16A1-AS1 and shNC controls; (2) GFP-GRP78 and empty vector transfectants; (3) AGS-shSLC16A1-AS1 cells transfected with either vector or GFP-GRP78. Subsequently, 4×106 cells of these 6 cell variants were injected subcutaneously into each flank of the mice (n=6). Tumor growth was monitored closely, with tumor size and volume recorded 3 times weekly. Tumor sizes were accurately measured using a caliper, and volumes were calculated employing the formula: π/6 × length (longest diameter) × width2 (corresponding perpendicular diameter). After 3 weeks of injections, the animals were humanely euthanized, and tumors were excised. A segment of the excised tumor tissue was preserved in PFA to facilitate immunohistochemical (IHC) staining for Ki67 expression and hematoxylin-eosin (HE) staining. Meanwhile, the remaining tumor tissue was carefully stored at −80°C for future analysis, including RT-qPCR or Western blot assays. All animal procedures were approved by the laboratory animal management and ethics committee of Nanjing Medical University (Approval No: IACUC-1807020) and conducted in accordance with institutional guidelines for animal care and use in research.

Cell lines

Human gastric cancer cell lines (AGS, HGC27, MKN28, MKN74 and SNU1), the normal gastric epithelium cell line GES1 and human embryonic kidney (HEK) 293T cells were obtained from the Chinese Academy of Sciences Committee on Type Culture Collection Cell Bank (Shanghai, China), cell lines were verified by STR analysis. SNU1, GES1 and HEK293T were grown in RPMI-DMEM medium (Gibco, Life Technologies, California, USA) supplemented with 10% fetal bovine serum and penicillin (100 UI/ml)/streptomycin (100 mg/ml) (Gibco, life technologies, California, USA), MKN74, MKN28 and HGC27 was cultured in RPMI-1640 with the same condition. AGS was cultured in DMEM/F12-basic with the same condition. Those cells were maintained in an incubator at 37°C with 5% CO2. All cell lines were routinely tested for mycoplasma using MycoAlert TM (Lonza).

Method details

Bioinformatics analysis of GEO and TCGA databases

RNA sequencing (RNA-seq) count data of stomach adenocarcinoma (STAD) were obtained from The Cancer Genome Atlas (TCGA). Differential expression analysis was performed in R using the DESeq2 package, with significant genes defined as those with q < 0.05 and an absolute log2 fold change >1.0.

For the GEO datasets GSE192468 and GSE58828, expression matrices were downloaded and processed in R. Probe-to-gene annotation was performed using the AnnoProbe package, and duplicate genes were consolidated by averaging their expression values. To account for group heterogeneity, normalization was conducted with the limma package, which applies linear modeling. Differentially expressed genes were identified using the same criteria (q < 0.05 and log2FC >1.0). LncRNA annotation in all datasets was based on the GENCODE comprehensive gene annotation (GTF file). All analyses were performed in R.

Plasmid construction, transfection of siRNA and plasmid

SLC16A1-AS1 shRNA-1 (shSLC16A1-AS1) and 2 were synthesized by Sangon Biotech (Shanghai, China) and constructed in pLKO.1 vector. Scrambled shRNA was cloned into the same vector, as a negative control (shNC). The SLC16A1-AS1 and NSUN2 cDNA were incorporated into pcDNA3.1 vectors to create constructs capable of expressing SLC16A1-AS1 and NSUN2 in mammalian cell lines. Additionally, the pCMV-EGFP-GRP78-neo plasmid was procured from Miaoling Biology (Wuhan, China). Truncated versions of SLC16A1-AS1 and NSUN2 constructs were produced using the ClonExpress Ultra One Step Cloning Kit V2 (Vazyme, C116-01) and subsequently verified through sequencing services provided by Sangon Biotech.

For the purpose of transfections, both plasmid DNA (48-hour expression) and siRNA (24-hour expression) were administered utilizing Lipofectamine 2000 (Invitrogen), following the manufacturer's instructions. Specifically, the siRNAs employed were targeted towards SLC16A1-AS1 or served as non-targeting control siRNA sourced from RiboBio (Guangzhou, China). In experiments involving double transfections, either of expression vectors or siRNA, an equal cumulative amount of DNA and/or siRNA was transfected into each experimental sample. This was achieved by supplementing with empty vectors or control siRNA to maintain consistency across controls and single-transfection setups. All sequences of siRNA, shRNA and primers are listed in Table S1.

RT-qPCR

Total RNA was extracted from tissues and cells using TRIzol reagent (Invitrogen) and quantified with a NanoDrop ND-1000 Spectrophotometer (Agilent, CA, USA). Complementary DNA (cDNA) was synthesized using the PrimeScript™ RT reagent Kit (Takara Bio USA, Inc., Mountain View, CA, USA). Quantitative real-time PCR (qPCR) was performed using the LightCycler® 480 system (Roche, Basel, Switzerland) with GAPDH as an internal control. Primer sequences were designed using NCBI Primer-BLAST and synthesized by Sangon Biotech (Shanghai). All primers are shown in Table S3.

Cell viability assessment

Cell viability was quantitatively evaluated utilizing the Cell Counting Kit-8 (CCK-8, Beyotime). Specifically, 3×10³ cells per well were seeded into a 96-well plate and cultured for varying durations ranging from 24 to 96 hours. Following the specified incubation periods, the absorbance values were recorded at a wavelength of 450 nm using a spectrophotometer. Based on these absorbance readings, a cell activity curve was plotted to visually represent the temporal changes in cell viability.

Colony formation assay

To assess the colony-forming ability of the cells, 1000 cells per well were seeded into 6-well plates and maintained in a culture medium supplemented with 10% FBS for two weeks. After the incubation period, colonies were fixed with 4% paraformaldehyde (PFA) for 30 minutes, followed by staining with 0.1% crystal violet for an additional 30 minutes. The number of visible colonies was then enumerated using an inverted microscope, providing a quantitative measure of colony formation.

EdU incorporation assay

To evaluate the proliferative capacity of cells, the EdU incorporation assay was performed using the BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 555, adhering strictly to the manufacturer's protocol. Briefly, 1000 cells per well were seeded into 96-well plates and allowed to grow until they reached 40-60% confluence. Subsequently, the cells were exposed to 50μM EdU solution for 2 hours at 37°C in a humidified atmosphere containing 5% CO2. Following EdU labelling, the cells were washed twice with PBS and fixed with 4% PFA for 30 minutes. Cell membranes were then permeabilized using 0.5% Triton X-100 for 10 minutes at room temperature. Alexa Fluor 555 dye was added to the cells for 2 hours, enabling the visualization of EdU-positive cells (appearing red) under an Olympus FSX100 microscope (Olympus, Tokyo, Japan). Additionally, Hoechst 33342 staining was performed to identify the nuclear region (appearing blue), facilitating the accurate counting of EdU-incorporating cells.

RNA bisulfite-seq (Bis-seq) library construction and sequencing

High throughput Bis-seq was provided by Cloudseq Biotech Inc. (Shanghai, China). Briefly, 2 μg of total RNA for each sample was rRNA depleted using NEBNext® rRNA Depletion Kit (New England Biolabs, Inc., Massachusetts, USA). rRNA-depleted RNA was bisulfite converted and purified using the EZ RNA methylation Kit (Zymo Research). RNA libraries were then constructed with TruSeq Stranded Total RNA Library Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. The library quality was evaluated with BioAnalyzer 2100 system (Agilent Technologies, Inc., USA). Library sequencing was performed on an Illumina Hiseq instrument with 150 bp paired end reads.

Subcellular fractionation and localization of SLC16A1-AS1

Subcellular fractionation to isolate cytoplasmic and nuclear fractions was conducted using the PARIS Kit (Invitrogen, Carlsbad, CA, USA), strictly adhering to the manufacturer's instructions. This process aimed to determine the cellular localization of LncSLC16A1-AS1. All experimental steps were carefully executed in accordance with the provided protocol. Following the fractionation, RNA was extracted separately from the nucleus and cytoplasm. RT-qPCR analysis was then performed on these RNA samples to quantify the levels of SLC16A1-AS1, along with the housekeeping genes GAPDH (as a cytoplasmic marker) and U6 snRNA (as a nuclear marker). This analysis allowed for the assessment of the subcellular distribution of SLC16A1-AS1 relative to these reference markers.

Co-localization of IF and RNA-FISH

The localization of SLC16A1-AS1 was performed by FISH assay (RiboBio Fluorescent In Situ Hybridization Kit) according to the manufacturer’s protocol. After probe hybridization, incubate with the NSUN2 antibody overnight at 4°C. Three washes with PBS, the corresponding fluorescence-labelled secondary antibodies (1:200; Cell Signaling Technology) incubated for 1 h at room tempreture. The nuclei were counterstained by DAPI (Beyotime). The immunofluorescence signals were captured using Leica confocal laser scanning microscope (CLSM). Biological replicates were counted three times.

In vitro transcription and RNA pulldown assay

The gene fragment was amplified by PCR using the SLC16A1-AS1 plasmid as the template, generating the DNA template for in vitro transcription (primer sequences are provided in the Table S2). The PCR product was purified by agarose gel electrophoresis. Biotin-labeled RNA was synthesized using a T7 in vitro transcription kit. For RNA secondary structure formation, 3 μg of RNA was denatured at 85°C for 2 min in a PCR thermocycler, then immediately placed on ice for 5 min. Subsequently, 10×RNA structure buffer (10 mM Tris pH 7.0, 100 mM KCl, 10 mM MgCl₂) was added to facilitate proper RNA folding.

The biotin-labeled RNA was incubated with Pierce™ Streptavidin Magnetic Beads (Thermo Scientific; Cat# 88817) under rotating conditions at room temperature for 1 hour. Following RNA-bead conjugation, the RNA-bound beads were incubated with whole-cell lysate at 4°C with rotation for 1 hour. After incubation, the beads were washed three times with ice-cold wash buffer (20mM Tris (pH 7.5), 10mM NaCl, 0.1% Tween-20) to remove nonspecifically bound proteins. The RNA-associated proteins were then eluted and pulled down with the coated beads. Finally, the captured proteins were analyzed by Silver staining and Western blot.

Silver staining

Following electrophoresis, the SDS-PAGE gel was fixed in 40% ethanol/10% acetic acid with agitation for 30 minutes to immobilize proteins, followed by three 5 minutes ultrapure water washes to remove residuals. The gel was then sensitized in 0.02% sodium thiosulfate (1 min) to enhance silver binding capacity, subsequently incubated in 0.1% silver nitrate/0.02% formaldehyde (20 min, dark) for silver-protein complex formation. Protein bands were developed in 2% sodium carbonate/0.04% formaldehyde solution with gentle agitation (1-5 minutes until optimal band visibility), with the reaction terminated by 1% acetic acid treatment before final water rinses for preservation.

Western blot analysis

Cells were harvested, washed with ice-cold PBS, and lysed in RIPA buffer containing proteinase inhibitors for 30 minutes. The lysates were centrifuged to remove debris, and protein concentrations were determined using a BCA assay kit (Pierce, Rockford, IL). Equal amounts of protein (20 μg) were loaded onto SDS-PAGE gels and subsequently transferred to PVDF membranes. The membranes were blocked in PBS containing 5% nonfat milk and 0.05% Tween-20, then incubated overnight with primary antibodies against GRP78 (Proteintech, 1:6000), NSUN2 (Proteintech, 1:1000) and GAPDH (Proteintech, 1:20000) at 4°C. Next, the membranes were incubated with HRP-conjugated secondary antibodies at room temperature. After washing, immunoreactive bands were visualized using the Chemistar™ High-sig ECL Western Blot Substrate (Tanon), yielding clear and distinct bands on the membrane.

RNA immunoprecipitation (RIP) assay

Cells were first harvested and then re-suspended in 1 ml of lysis buffer that contained a protease inhibitor cocktail and RNase inhibitor to ensure the integrity and stability of the RNA during the extraction process. Following this, the cell lysate was centrifuged at 13,000 rpm for 10 minutes at 4°C to separate the supernatant containing the target RNA-protein complexes from cellular debris. The supernatant was incubated with 30∼40 μl of Protein A/G-Sepharose beads (Genescript) and 2 μg of primary antibodies for 4 hours at 4°C. After incubation, the beads were washed once with ice-cold 1×PBS and subsequently washed three times with lysis buffer. Next, the beads containing the RNA-protein complexes were subjected to RNA extraction using Trizol reagent (Invitrogen Life Technologies) in the presence of Proteinase K (Sigma). Finally, the purified RNA was subjected to RT-qPCR analysis to quantify the levels of the target RNA that had been immunoprecipitated.

MeRIP-qPCR

The total RNA was extracted utilizing the Trizol method and subsequently dissolved to achieve a concentration of at least 1000 ng/μl. RNA fragmentation and subsequent product purification was adhered to the NEBNext® Magnetic RNA Fragmentation Module Protocol. The reaction mixture was prepared by combining 22 μl of fragmented RNA, 2 μl of 3M Sodium Acetate (NaAc), 2 μl of 5 mg/ml Linear Acrylamide, and 60 μl of 100% Ethanol, and incubated at -80°C for 1 hour. It was then centrifuged at 14,000 rpm for 25 minutes at 4°C. The supernatant was discarded, and the pellet was washed twice with 300 μl of 75% ethanol. The pellet was dried thoroughly and subsequently dissolved in DEPC-treated water.

Protein-A/G magnetic beads 25 μl were washed twice with 500 μl of PBS. Then, 1 ml of PBS and 3 μg of anti-m5C were added to the washed magnetic beads. The mixture was incubated with shaking at 4°C overnight. Following this, the RNA-IP buffer (containing 150 mM LiCl, 0.5% NP-40, and 10 mM Tris-HCl, pH 7.5) was used to wash the beads three times. RNA samples were then incubated with the magnetic beads in 500 μl of RNA-IP buffer at 4°C for 4 hours. After incubation, the beads were subjected to a series of washes: once with RNA-IP buffer, once with RNA-IP buffer supplemented with 150 mM LiCl, once with RNA-IP buffer containing 300 mM LiCl, and finally, once more with plain RNA-IP buffer. To isolate the RNA, the IP system was briefly centrifuged and then placed on a magnetic rack in an ice bath for 1 minute. After carefully removing the supernatant, the pellet was resuspended in 1 ml of Trizol reagent for RNA extraction. The purified RNA was subsequently subjected to RT-qPCR analysis to assess RNA enrichment by m5C antibody.

RNA sequencing (RNA-seq) analysis

Total RNA was extracted from AGS-siNC and AGS-siSLC16A1-AS1 cells using TRIzol, following the standard protocol. Sequencing and initial analysis of SLC16A1-AS1 knockdown and control cell lines were conducted by NovelBio (Shanghai, China), with subsequent analyses performed using R. Duplicate gene reads with extremely low counts were filtered out. Differential gene expression (DGE) analysis was carried out using the DESeq2 package in R. Differentially expressed genes were then subjected to KEGG pathway analysis. Approximately one-sixth of the clustered pathways were identified as potentially most relevant to gastric cancer progression. The normalized expression matrix obtained from the DGE results was visualized using the Heatmap package in R.

Constructing NSUN2 knockdown cell lines using CRISPR/Cas9 technology

To select candidate sgRNAs for NSUN2 genome editing, an online CRISPR tool, GuidePro v2.1.0, developed by Dr. Xu’s Lab at MD Anderson Cancer Center (https://bioinformatics.mdanderson.org/apps/GuidePro), was utilized. The pLentiCRISPRv.2 plasmid, created by Feng Zhang (Addgene Cat# 52961), was digested with BsmBI. Three sgRNA oligonucleotides were cloned into the plasmid following a standard protocol. For transient transfections and lentivirus-mediated DNA transfer, HEK293T cells were transfected using PEI. 48 hours post-transfection, the virus-containing medium was collected and filtered. HEK-293T cells were then infected with this virus-containing medium supplemented with 8 μg/ml Polybrene (Sigma, USA) for 24 hours, after which the medium was replaced with fresh culture medium containing 1.5 μg/ml puromycin to select cells that harbored the sgRNAs of interest. Three days post-selection, cells were lysed for Western blot analysis of NSUN2 expression, to evaluate NSUN2 knockdown efficiency. sgRNA sequences are shown in Table S3.

mRNA stability assay

To assess mRNA stability, AGS stable SLC16A1-AS1-knockdown cells and their corresponding control cells were plated in 6-well plates and treated with Actinomycin D (ActD) (5 μg/mL, GLPBIO) to block transcription. Cells were harvested at 0-, 4-, 8-, and 12-hour post-treatment, and total RNA was extracted using TRIzol reagent. For the NSUN2 rescue experiment, AGS-shSLC16A1-AS1 cells (shRNA-1&2) were transfected with vector or Flag-NSUN2 and then treated with ActD. Cells were harvested at 0-, 3-, 6-, and 9-hour post-treatment to purify total RNA. Equal amounts of RNA were reverse-transcribed into cDNA, and GRP78 mRNA levels were quantified by RT-qPCR. The relative mRNA abundance at each time point was normalized to the t = 0-hour baseline and analyzed for decay kinetics.

Quantification and statistical analysis

Data were analyzed using SPSS 22.0 software. Survival curves were generated using Kaplan-Meier plots and compared using log-rank tests. Differences between groups were evaluated using paired Student’s t-test and two-way ANOVA. The Chi-square test was used to analyze pathological features associated with SLC16A1-AS1 expression in GC. All quantitative experiments were conducted in triplicate, and all quantitative data are presented as the mean ± standard deviation (SD). A significance level of P<0.05 (∗) was considered significant, while P<0.01 (∗∗), P<0.001 (∗∗∗), and P<0.0001 (∗∗∗∗) were considered highly significant.

Additional resources

All other items: any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Published: January 17, 2026

Supplemental information

Table S2. Clinicopathological characteristics of gastric cancer patients included in the study

This table summarizes the detailed clinicopathological information of all enrolled gastric cancer patients, including basic demographic data (age and gender, with gender coded as 1 = Male and 2 = Female), tumor-related pathological features (pathological type, tumor differentiation grade, and anatomical lesion location), and tumor staging information (TNM stage classification and overall clinical stage). Additionally, it provides key follow-up data such as patient survival status (coded as 1 = Deceased and 2 = Alive) and survival time (followed up until September 2016).