Abstract

   Gastric cancer (GC) remains a prevalent and aggressive malignancy with
   a poor prognosis. This study aimed to identify diagnostic and
   prognostic biomarkers while exploring their potential functions in GC.
   A total of 598 upregulated and 506 downregulated genes were identified
   in GC patients. Among these, survival-related differentially expressed
   genes (DEGs), including ADAMTS12, F5, and VCAN, were highlighted.
   Pan-cancer analyses revealed their dysregulation across multiple tumor
   types. A novel prognostic signature, incorporating ADAMTS12 and F5,
   effectively stratified GC patients into low- and high-risk groups,
   demonstrating significant differences in overall survival and robust
   predictive performance. ADAMTS12, strongly associated with advanced
   clinical stages and poor prognosis, was validated in an independent
   cohort and exhibited promising diagnostic potential. RT-PCR and western
   blot analyses confirmed its high expression in GC tissues and cell
   lines. Functional assays further demonstrated that ADAMTS12 promotes GC
   cell proliferation and invasion. In summary, this study provides
   critical insights into the molecular landscape of GC, offering a
   potential prognostic tool and therapeutic target.

Supplementary Information

   The online version contains supplementary material available at
   10.1007/s12672-024-01724-4.

   Keywords: Gastric cancer, Biomarker, ADAMTS12, Prognostic signature

Introduction

   According to the statistics report that was submitted by GLOBOCAN 2020,
   gastric cancer (GC) is the fifth most frequent cancer in the world and
   the fourth main cause of deaths that are related to cancer [[28]1,
   [29]2]. In 2020, over one million individuals were diagnosed with
   stomach cancer, resulting in more than 750,000 deaths globally. The
   incidence of GC exhibits significant regional variation. According to
   World Health Organization (WHO) statistics, Asian
   countries—particularly China, Japan, and South Korea—are recognized as
   high-incidence regions for GC [[30]3, [31]4]. In contrast, the
   incidence of gastric cancer (GC) is relatively low in North America and
   Europe. Age and gender also play significant roles in influencing GC
   incidence. The disease is generally more prevalent among middle-aged
   and elderly individuals, particularly those aged 60 and older [[32]5].
   Currently, the tumor-node-metastasis (TNM) staging of postoperative
   specimens is a widely used clinical tool for predicting long-term
   survival outcomes in patients with gastrointestinal cancer [[33]6].
   However, research has revealed several significant limitations of the
   TNM staging method in clinical practice [[34]7]. Therefore, identifying
   novel biomarkers to predict the long-term prognosis of GC patients and
   stratify their risk is critically important. Such advancements would
   enable more precise, individualized treatment approaches, addressing a
   key challenge in GC management.

   Messenger RNA (MRNA) plays a crucial role in cancer research,
   particularly in the prediction and diagnosis of tumor prognosis
   [[35]8]. Differential mRNA expression levels serve not only as
   potential biomarkers for early cancer detection and prognosis
   prediction but also as valuable tools for guiding individualized
   treatment strategies by distinguishing high-risk patients from low-risk
   groups [[36]9]. In-depth analysis of mRNA in tumor tissues enables
   researchers to identify expression patterns linked to patient survival
   and disease progression, offering critical insights for more accurate
   prognostic assessments [[37]10, [38]11]. Moreover, mRNA research
   facilitates the identification of tumor subtypes, providing valuable
   insights into the biological characteristics of tumors and their
   clinical manifestations. Most importantly, mRNA analysis supports the
   discovery of potential therapeutic targets, laying a theoretical
   foundation for developing novel treatment strategies and driving
   progress in personalized medicine and precision therapy
   [[39]12–[40]14]. The Cancer Genome Atlas (TCGA), a comprehensive cancer
   genomics research initiative, offers an extensive repository of
   molecular data across various levels, including the genome,
   transcriptome, and epigenome, to support cancer research. The primary
   goal of TCGA is to lay the groundwork for personalized cancer treatment
   by conducting detailed molecular analyses to uncover the unique
   molecular characteristics of different cancer types [[41]15, [42]16].
   TCGA plays a broad role in cancer research, including molecular
   classification and subtype identification, discovery of cancer driver
   genes, identification of biomarkers, development of personalized
   treatment strategies, and advancement of bioinformatics tools [[43]17].
   By analyzing TCGA data, researchers can gain a deeper understanding of
   the molecular mechanisms driving cancer, providing critical insights to
   support clinical treatment and personalized medicine, thereby advancing
   the field of cancer research.

   In this study, we analyzed TCGA datasets and identified three critical
   biomarkers for GC patients. Among them, our attention focused on
   ADAMTS12. Our findings suggested ADAMTS12 may be a novel biomarker and
   therapeutic target for GC patients.

Materials and methods

Patients and collection of samples

   GC tissues and paired adjacent normal tissues were collected from 10 GC
   patients who underwent surgical resection at Qingdao Chengyang People’s
   Hospital between 2022 and 2023. None of the patients received
   preoperative treatments, including radiation or chemotherapy. To
   support future research, half of the tissues were promptly snap-frozen
   in liquid nitrogen, while the remaining half was stored at − 80 °C. The
   residual tissue portions were fixed in formalin and embedded in
   paraffin. All procedures were conducted in accordance with the
   Declaration of Helsinki, and the study was approved by Qingdao
   Chengyang People’s Hospital (Approval number: QDCY2022137). Written
   informed consent was obtained from all participants.

Cell culture and transfection

   The GC cell lines SGC7901, MKN-45, HGC-27, and AGS, along with the
   normal gastric epithelial cell line GSE-1, were purchased from the
   Chinese Academy of Sciences (Shanghai, China). All cells were cultured
   at 37 °C with 5% carbon dioxide in media supplemented with 10% fetal
   bovine serum (FBS; Gibco, Carlsbad, California, USA) and 1%
   penicillin/streptomycin (Gibco). Ribobio (Guangzhou, China) synthesized
   and engineered siRNA oligonucleotides targeting ADAMTS12 (si-ADAMTS12-1
   and si-ADAMTS12-2). Cell transfection was performed according to the
   manufacturer’s instructions (Invitrogen, USA) using Lipofectamine 3000.

Real-time PCR (RT-PCR)

   Total RNA was extracted from both the specimens and cells using TRIzol
   reagent (Invitrogen). The RNA was then reverse-transcribed into cDNA
   using the PrimeScript™ RT Reagent Kit (Takara, Dalian, China). RNA
   quantification was performed using the ABI PRISM 7900 Real-Time PCR
   System (Applied Biosystems). GAPDH was used as the internal control for
   normalization. The relative gene expression levels were determined
   using the 2 − ΔΔCt method. The PCR conditions were as follows: an
   initial 3-min denaturation at 94 °C, followed by 30 cycles of 45 s at
   94 °C, 45 s at 57 °C, and 45 s at 72 °C, with a final extension at
   72 °C for 10 min. The specific primers were as follows: ADAMTS12
   forward 5ʹ-ACTTTGGGGCGCTTTGCTAT-3ʹ and reverse
   5ʹ-GCAGGCCCTTGATAAAATGC-3′; GAPDH forward 5′-CTCAGACACCATGGGGAAGGTGA-3′
   and reverse 5′-ATGATCTTGAGGCTGTTGTCATA-3′.

Colony formation assay

   Transfected MKN-45 and HGC-27 cells were seeded into 6-well plates at a
   density of 1 × 10^3 cells per well. The plates were incubated at 37 °C
   for 2 weeks in either DMEM or RPMI 1640, each supplemented with 10%
   FBS. After incubation, cells were fixed in methanol for 30 min, stained
   with 1% crystal violet dye, and then counted. Each experiment was
   performed in triplicate.

5-Ethynyl-2′-deoxyuridine (EDU) assay

   To assess cell proliferation, the EDU assay (RiboBio, Nanjing, China)
   was performed. After 48 h of transfection, the cells were treated with
   50 μM EDU. Apollo staining and DAPI labeling were then conducted to
   visualize EDU-positive cells.

Cell invasion

   Transfected MKN-45 and HGC-27 cells (200 µl) were seeded into the upper
   chambers of Transwell inserts (Corning Inc.), which had an 8 µm pore
   size and were placed in 24-well plates. The Transwell membrane was
   pre-coated with Matrigel (BD Biosciences). Simultaneously, 600 µl of
   RPMI-1640 medium supplemented with 10% FBS was added to the lower
   chambers. The cells were incubated at 37 °C for 24 h, after which they
   were fixed and stained with 0.5% crystal violet solution
   (Sigma-Aldrich; Merck KGaA) for 15 min at room temperature. The number
   of invading cells was counted using a light microscope (Olympus
   Corporation) at ×100 magnification. Images were captured from five
   random fields and analyzed using ImageJ software (version 1.8.0.112)
   from the National Institutes of Health.

Western blot analysis

   Cells were lysed using RIPA buffer (Merck, China) for this experiment.
   Protein concentrations were normalized using the BCA assay. The lysates
   were separated by SDS-PAGE (10–15% polyacrylamide gels) and transferred
   onto PVDF membranes. The membranes were incubated overnight at 4 °C
   with primary antibodies, including anti-ADAMTS12 (1:500, Proteintech,
   USA) and anti-β-actin (1:5000, CST, USA). Afterward, the membranes were
   incubated with appropriate secondary antibodies (anti-rabbit
   antibodies, Immunoway) at room temperature (25 °C) for two hours. The
   blots were developed using ECL detection.

Data collection

   To investigate stomach adenocarcinoma (STAD), we collected HTSeq-Count
   data, clinicopathological information, and overall survival (OS) data
   from the TCGA database. A total of 406 patients’ clinicopathological
   characteristics were included, along with transcriptome data from 375
   STAD tissues and 359 normal tissues obtained from GTEx. Additionally,
   data from GTEx were incorporated. The clinical characteristics of the
   clinical cohort are provided in Table S1.

Data preprocessing and screening of the differentially expressed mRNAs (DEMs)

   To screen for differentially expressed genes (DEGs) in the gene
   expression profiles from the TCGA datasets, the Limma package in R
   software (version 3.5.3) was employed. To analyze the differential
   expression between GC and normal tissues from the TCGA datasets, the
   edgeR package in R was used. A significance threshold was set with a
   log2 (fold change) > 4 and an adjusted p-value of less than 0.05. The
   results were visualized using volcano plots and heatmaps to display the
   DEMs data.

Function enrichment analysis of DEMs

   Gene Ontology (GO) analysis is a widely used method in bioinformatics
   and biological research that systematically classifies and annotates
   genes and their products to better understand their functions in
   biological processes. It encompasses three main categories: molecular
   function, cellular component, and biological process. Molecular
   function describes the activities of gene products at the molecular
   level, cellular component specifies their locations within cells, and
   biological process outlines their involvement in cellular and
   organismal functions. By categorizing genes based on these functional
   categories, GO analysis provides a comprehensive approach to
   understanding gene functions. It is commonly applied to interpret
   high-throughput experimental data, such as gene expression profiles or
   proteomics data. For differentially expressed gene sets, GO analysis
   helps researchers understand their functional localization, roles in
   biological processes, and molecular functions, enabling insights into
   their roles and regulatory networks in specific biological contexts.
   KEGG analysis is another widely used method in bioinformatics and
   systems biology aimed at understanding gene and protein functions,
   particularly in the context of metabolic and signaling pathways. The
   Kyoto Encyclopedia of Genes and Genomes (KEGG) is a comprehensive
   database that provides detailed biological information, including
   insights into genomics, biochemistry, and more. In KEGG analysis,
   researchers compare gene sets of interest with data from the KEGG
   database, linking genes to specific metabolic and signaling pathways.
   This approach helps uncover the roles of gene sets in cellular
   metabolism and signal transduction, revealing how genes collaborate to
   perform essential biological functions. For gene function analysis, the
   Database for Annotation, Visualization, and Integrated Discovery
   (DAVID) version 6.8 ([44]https://david.ncifcrf.gov) was utilized. The
   DAVID analysis module is designed to transform data into biological
   insights, accelerating the study of genome-level datasets. To elucidate
   the biological significance of the differentially expressed genes
   (DEGs) identified in GC progression, we used DAVID to conduct GO
   functional and KEGG pathway enrichment analysis on the DEMs obtained
   from the Venn analysis. A false discovery rate (FDR) of less than 0.05
   was considered statistically significant.

Identification of survival-related DEMs and establishment of the prognostic
gene signature

   The differentially expressed genes (DEMs) associated with overall
   survival (OS) were identified using the TCGA STAD dataset. DEMs with a
   p-value of less than 0.05 were considered statistically significant and
   included for further analysis. To refine the number of DEMs in the
   selected panel, a Lasso-penalized Cox regression analysis was
   performed. This analysis utilized tenfold cross-validation, implemented
   through the glmnet package in R, to optimize prediction performance. To
   generate a predictive gene signature for gastric cancer patients, a
   linear combination of the regression coefficients (β) obtained from the
   Lasso Cox regression model was calculated and multiplied by the mRNA
   expression levels of the respective genes. The optimal cutoff for the
   predictive gene signature was determined using X-Tile software, which
   classified patients into high-risk and low-risk groups based on this
   cutoff. To evaluate the prognostic value of the gene signature, several
   methods were employed, including Kaplan–Meier survival analysis,
   Harrell’s concordance index (C-index), the area under the curve (AUC)
   of the receiver operating characteristic (ROC) curve, and a calibration
   plot comparing predicted and observed overall survival.

Statistical analysis

   A one-way analysis of variance (ANOVA) was used to compare gene
   expression patterns between normal gastric tissues and GC samples. The
   Pearson chi-square test was employed to examine categorical data. To
   assess overall survival across different groups, Kaplan–Meier (K–M)
   survival analysis was conducted, along with a two-sided log-rank test.
   A two-tailed p-value of less than 0.05 was considered statistically
   significant.

Results

Identification of DEMs in GC and functional enrichment analysis

   First, we analyzed TCGA datasets to identify DEGs between GC samples
   and normal samples. Based on criteria of LogFC > 4 and p < 0.05, we
   identified 598 upregulated genes and 506 downregulated genes in GC
   samples (Fig. [45]1A and B). Next, we performed functional enrichment
   analysis to explore the potential functions of these genes. KEGG
   analysis revealed that the 598 upregulated genes were primarily
   enriched in the p53 signaling pathway, viral protein interaction with
   cytokines and cytokine receptors, viral myocarditis, and the NF-kappa B
   signaling pathway. Conversely, the 506 downregulated genes were mainly
   associated with retrograde endocannabinoid signaling, the cAMP
   signaling pathway, protein digestion and absorption, prion diseases,
   and Parkinson’s disease. Furthermore, GO analysis showed that the
   upregulated genes were predominantly involved in the regulation of
   mitotic sister chromatid separation, nuclear division, and sister
   chromatid segregation, while the downregulated genes were primarily
   linked to sodium ion homeostasis, sodium ion export across the plasma
   membrane, response to muscle activity, and organic acid catabolic
   processes (Fig. [46]1C).

Fig. 1.

   [47]Fig. 1
   [48]Open in a new tab

   Differential expression analysis and Functional Enrichment in GC. A and
   B Heatmap and Volcano plots illustrating the differential expression of
   genes between GC samples and normal samples. Genes with LogFC > 4 and
   p < 0.05 are highlighted in red (upregulated) and blue
   (down-regulated). The differential expression analysis was performed
   using the limma package in R, with Benjamini–Hochberg correction for
   multiple testing. C Functional Enrichment Analysis of DEMs, including
   KEGG pathway analysis and GO) analysis. The upregulated genes are
   associated with pathways such as p53 signaling, viral interactions, and
   sister chromatid segregation, while down-regulated genes are linked to
   cAMP signaling and sodium ion homeostasis

Identification of survival-related DEMs in GC patients

   We then integrated sequencing data from GC patients in the TCGA
   datasets with corresponding clinical information to further identify
   survival-related DEGs in GC patients. Using a threshold of p < 0.01, we
   performed Kaplan–Meier (K–M) survival analysis on the 598 upregulated
   genes and 506 downregulated genes and identified 10 survival-related
   DEGs in GC patients, including CLRN3, BCL11B, PDSS1, PRSS3, MYB,
   ADAMTS12, EPCAM, MAP7, F5, and VCAN (Fig. [49]2A and B). Given the
   consistency between expression patterns and clinical significance, we
   focused on ADAMTS12, F5, and VCAN. The expression of these genes was
   significantly elevated in GC specimens and was associated with poor
   prognosis (Fig. [50]2C). Additionally, we performed pan-cancer analysis
   and found that ADAMTS12 (Fig. [51]3A), VCAN (Fig. [52]3B), and F5
   (Fig. [53]3C) exhibited dysregulated expression in various types of
   tumors, underscoring their important roles in tumor progression.

Fig. 2.

   [54]Fig. 2
   [55]Open in a new tab

   Kaplan–Meier survival analysis of DEMs in GC. A and B Kaplan–Meier
   survival curve presenting the overall survival of GC patients based on
   the expression levels of identified DEMs. The log-rank test was used to
   compare survival curves between groups, with a significance level set
   at p < 0.01. C Forest plot illustrating hazard ratios for the
   identified 10 survival-related DEMs in GC patients, including CLRN3,
   BCL11B, PDSS1, PRSS3, MYB, ADAMTS12, EPCAM, MAP7, F5, and VCAN

Fig. 3.

   [56]Fig. 3
   [57]Open in a new tab

   Pan-Cancer Analysis of A ADAMTS12, B VCAN, and C F5 in various tumor
   types. Statistical significance was assessed using one-way ANOVA for
   comparisons between multiple groups, with post hoc Tukey’s test applied
   to determine pairwise differences. A p-value < 0.05 was considered
   statistically significant

The prognostic value of ADAMTS12, F5 and VCAN in pan-cancer analysis

   To further explore the clinical significance of ADAMTS12, F5, and VCAN
   in tumors, we conducted a pan-cancer analysis. As shown in Figs. [58]4A
   and B, we observed that ADAMTS12 expression was significantly
   associated with overall survival (OS) and disease-free survival (DFS)
   in several types of tumors, including KIPAN, BLCA, KIRP, ACC, MESO,
   PAAD, and UVM. Additionally, the prognostic value of VCAN and F5 in
   pan-cancer analysis is shown in Figures S1A and S1B, and Figures S2A
   and S2B. The expression of both VCAN and F5 was also found to be
   associated with the prognosis of multiple tumor types.

Fig. 4.

   [59]Fig. 4
   [60]Open in a new tab

   Pan-cancer analysis revealed the prognostic impact of ADAMTS12
   expression. A Association of ADAMTS12 expression with overall survival
   in various tumor types, including KIPAN, BLCA and KIRP. B Correlation
   of ADAMTS12 expression with disease-free survival across multiple tumor
   types. Kaplan–Meier survival curves were generated, and differences in
   survival were assessed using the log-rank test, with a significance
   level set at p < 0.05

Construction of a new gene prognostic signature for GC patients

   The above analyses revealed that ADAMTS12, F5, and VCAN were closely
   associated with the survival outcomes of GC patients. Therefore, we
   performed LASSO regression to construct a prognostic signature for GC.
   The coefficients of ADAMTS12, F5, and VCAN are presented in
   Fig. [61]5A. The model achieved the best fit when two of these
   genes—ADAMTS12, F5, and VCAN—were included (Fig. [62]5B). The risk
   score model classified 370 GC tumor samples into low- or high-risk
   subgroups based on a median threshold, and a heatmap of F5 and VCAN
   expression in GC was also generated (Fig. [63]5C). Furthermore, overall
   survival analysis showed that the high-risk group had significantly
   poorer survival outcomes (Fig. [64]5D). To assess the prognostic
   performance of the model, receiver operating characteristic (ROC)
   analysis, accounting for time, was performed. The area under the ROC
   curve (AUC) values for one-year, three-year, and five-year overall
   survival were 0.595, 0.619, and 0.739, respectively (Fig. [65]5E).

Fig. 5.

   [66]Fig. 5
   [67]Open in a new tab

   Construction and evaluation of a gene prognostic signature for GC. A
   Coefficient plot of ADAMTS12, F5, and VCAN in the LASSO regression
   model. The optimal penalty parameter (lambda) was determined using
   tenfold cross-validation to prevent overfitting. B Selection of the
   optimal number of genes for the prognostic signature using LASSO
   regression, with the minimum lambda indicated. C Heatmap displaying the
   expression levels of F5 and VCAN in GC tumor samples classified into
   low- and high-risk subgroups based on the risk score model, with
   significant differences analyzed using Student’s t-test (p < 0.05). D
   Kaplan–Meier survival analysis comparing overall survival between low-
   and high-risk groups defined by the prognostic signature, with the
   log-rank test used to assess differences in survival (p < 0.05). E ROC
   analysis assessing the prognostic model’s performance over one, three,
   and five years, with area under the curve (AUC) values calculated to
   evaluate the diagnostic accuracy

High expression of ADAMTS12 in GC and its association with clinical variables

   We focused on ADAMTS12, one of the three survival-related genes, as its
   expression and function in GC have not been previously documented.
   Compared to non-tumor specimens, GC tissues exhibited a significant
   increase in ADAMTS12 expression (Fig. [68]6A). Additionally, we
   observed that high ADAMTS12 expression was associated with the age of
   GC patients (Fig. [69]6B), but it did not correlate with gender
   (Fig. [70]6C) or tumor grade (Fig. [71]6D). Furthermore, high ADAMTS12
   expression was positively associated with advanced clinical stage
   (Fig. [72]6E) and T stage (Fig. [73]6F), but it was not correlated with
   M stage (Fig. [74]6G) or N stage (Fig. [75]6H). The association between
   ADAMTS12 expression and various clinical factors is shown in the
   heatmap (Fig. [76]6I). Additionally, we evaluated the prognostic
   significance of ADAMTS12 expression in GC patients using Cox regression
   analyses. The results of the multivariate analysis, shown in
   Table [77]1, demonstrate that clinical stage, age, and ADAMTS12
   expression independently predict poor prognosis in GC patients.

Fig. 6.

   [78]Fig. 6
   [79]Open in a new tab

   Association of ADAMTS12 expression with clinical variables in GC. A
   Expression levels of ADAMTS12 in GC specimens compared with non-tumor
   specimens from TCGA datasets. Statistical significance was determined
   using the Student’s t-test (p < 0.05). B–H Correlation between ADAMTS12
   expression and several clinical factors of GC patients, assessed using
   Spearman or Pearson correlation coefficients, depending on the data
   distribution, with p-values reported for each correlation. I Heatmap
   illustrating the association between ADAMTS12 expression and various
   clinical factors in GC, with a two-sided p-value < 0.05 indicating
   statistical significance in the associations

Table 1.

   Univariate and multivariate analysis of overall survival in GC patients
   Characteristics Total (N) Univariate analysis Multivariate analysis
   Hazard ratio (95% CI) p value Hazard ratio (95% CI) p value
   Pathologic stage 347
    Stage I & stage II 160 Reference Reference
    Stage III & stage IV 187 1.947 (1.358–2.793) < 0.001 1.947
   (1.357–2.793) < 0.001
   Age 367
    ≤ 65 163 Reference Reference
    > 65 204 1.620 (1.154–2.276) 0.005 1.673 (1.176–2.381) 0.004
   Gender 370
    Female 133 Reference
    Male 237 1.267 (0.891–1.804) 0.188
   ADAMTS12 370
    Low 184 Reference Reference
    High 186 1.544 (1.108–2.151) 0.010 1.583 (1.122–2.233) 0.009
   [80]Open in a new tab

ADAMTS12 expression was increased in GC using our cohorts

   We then collected 10 pairs of GC specimens and corresponding normal
   tissues to evaluate the expression of ADAMTS12 in GC. RT-PCR results
   revealed a significant increase in ADAMTS12 expression in GC specimens
   compared to normal tissues (Fig. [81]7A). Subsequent ROC analysis
   confirmed the strong diagnostic value of ADAMTS12 in distinguishing GC
   specimens from normal tissues, with an AUC of 0.84 (Fig. [82]7B).
   Additionally, RT-PCR and western blot analyses of several GC cell lines
   showed a marked upregulation of ADAMTS12 expression in four different
   GC cell lines (Fig. [83]7C and D). Importantly, these findings were
   consistent with the results from TCGA datasets. The elevated expression
   of ADAMTS12 in both clinical specimens and cell lines highlights its
   potential as a diagnostic marker for GC.

Fig. 7.

   [84]Fig. 7
   [85]Open in a new tab

   A RT-PCR analysis comparing ADAMTS12 expression in 10 pairs of GC
   specimens and normal specimens, revealing a distinct increase in GC
   specimens. Statistical significance was determined using the paired
   Student’s t-test (p < 0.05). B ROC analysis confirming ADAMTS12 as a
   robust diagnostic marker for distinguishing GC specimens from normal
   specimens, with an AUC of 0.84, calculated using the trapezoidal rule.
   C RT-PCR results showing ADAMTS12 expression in various GC cell lines
   compared to normal controls, analyzed using one-way ANOVA followed by
   Tukey’s post hoc test (p < 0.05). D Western blot results demonstrating
   significantly elevated ADAMTS12 expression in various GC cell lines
   compared to normal controls, with quantification performed using ImageJ
   software and analyzed by one-way ANOVA followed by Tukey’s post hoc
   test (p < 0.05)

Knockdown of ADAMTS12 expression suppressed the proliferation and invasion of
GC cells

   To elucidate the functional role of ADAMTS12 in GC progression, we
   began by employing siRNA-mediated knockdown of ADAMTS12 expression in
   MKN-45 and HGC-27 cells. The efficiency of si-ADAMTS12 transfection was
   confirmed through comprehensive analyses, including RT-PCR and western
   blot assays, as shown in Fig. [86]8A. Subsequent functional assays,
   such as the Colony Formation Assay and EdU assays, revealed a
   significant suppression of cellular proliferation in both MKN-45 and
   HGC-27 cells following ADAMTS12 knockdown (Fig. [87]8B and C).
   Additionally, Transwell assays demonstrated a marked reduction in cell
   invasion capacity upon downregulation of ADAMTS12 expression in both
   cell lines (Fig. [88]8D). These findings collectively suggest that
   ADAMTS12 plays a critical role in promoting cell proliferation and
   invasion in GC, highlighting its potential as a therapeutic target for
   inhibiting GC progression.

Fig. 8.

   [89]Fig. 8
   [90]Open in a new tab

   Suppression of GC cell proliferation and invasion upon knockdown of
   ADAMTS12 expression. A RT-PCR and western blot results confirming the
   efficient knockdown of ADAMTS12 expression in MKN-45 and HGC-27 cells
   using si-ADAMTS12, with statistical significance assessed by the paired
   Student’s t-test (p < 0.05). B and C Colony formation assays and EdU
   assays illustrating a significant reduction in the proliferation of
   MKN-45 and HGC-27 cells upon ADAMTS12 knockdown, analyzed by one-way
   ANOVA followed by Tukey’s post hoc test (p < 0.05). D Transwell assay
   demonstrating inhibited invasion of MKN-45 and HGC-27 cells following
   knockdown of ADAMTS12 expression, with statistical significance
   evaluated using the Mann–Whitney U test (p < 0.05)

Discussion

   In clinical practice, the diagnosis and prognosis prediction of GC
   patients primarily rely on various methods [[91]18]. Initially, doctors
   gather preliminary information by thoroughly understanding the
   patient’s symptoms and conducting a physical examination [[92]19].
   Subsequently, imaging examinations, such as gastroscopy and X-ray,
   provide detailed information about the tumor’s structure. Blood
   biomarker tests, including CA 19-9 and CEA, serve as complementary
   tools for diagnosis and predicting disease progression [[93]20,
   [94]21]. Additionally, pathological examinations involve obtaining
   tissue specimens through biopsies to determine histological types and
   assess lymph node involvement. Molecular biomarker testing, including
   HER2 and EGFR, provides deeper molecular insights that aid in
   formulating personalized treatment plans [[95]22, [96]23]. The TNM
   staging system is employed to stage cancer, evaluating the severity of
   the disease and predicting prognosis [[97]24, [98]25].

   Targeted treatment strategies, such as anti-HER2 therapy, focus on
   specific molecular markers for treatment. However, these methods also
   have limitations. Gastroscopy is invasive and may not be suitable for
   all patients. Imaging examinations have limited resolution and may miss
   early or small lesions. Blood biomarker specificity is often
   constrained, and levels may be elevated in other diseases. Pathological
   examinations require tissue sampling, which can limit access to deep or
   hard-to-reach areas. Molecular biomarker analysis involves complex
   techniques and higher costs. The TNM staging system does not account
   for tumor molecular characteristics, and targeted treatments are only
   effective for patients with specific target markers.

   MRNA has diverse potential applications in tumor diagnosis and
   prognosis prediction. First, through comprehensive molecular analysis,
   distinct mRNA expression patterns associated with specific cancer types
   or subtypes can be identified in tumor tissues, providing potential
   biomarkers for early cancer diagnosis and subtype classification.
   Second, by analyzing mRNA expression levels, a more accurate assessment
   of cancer prognosis can be made, offering valuable insights for the
   development of personalized treatment plans by healthcare professionals
   [[99]26]. Additionally, mRNA analysis aids in determining the molecular
   subtypes of tumors, providing essential information for understanding
   the biological heterogeneity of cancer [[100]27, [101]28]. In this
   study, we analyzed transcriptome data from GC patients in the TCGA
   datasets and identified 598 upregulated genes and 506 downregulated
   genes. Furthermore, functional enrichment analysis revealed that the
   differentially expressed genes are associated with activated biological
   processes such as the p53 signaling pathway, immune responses, and
   viral infections. In contrast, biological processes related to cell
   signaling, neural system functions, and metabolism appear to be
   suppressed. These findings provide valuable insights for further
   exploring the relationship between gene expression changes and specific
   physiological conditions or disease states.

   We then identified survival-related differentially expressed genes
   (DEMs) in GC patients by integrating TCGA sequencing data with clinical
   information. Ten DEMs, including CLRN3, BCL11B, PDSS1, PRSS3, MYB,
   ADAMTS12, EPCAM, MAP7, F5, and VCAN, were found to be significantly
   associated with survival outcomes in GC patients. Among these,
   ADAMTS12, F5, and VCAN stood out due to their consistent upregulation
   in GC specimens, which correlated with poor prognosis. Pan-cancer
   analysis further highlighted the dysregulated expression of ADAMTS12,
   VCAN, and F5 across multiple tumor types, emphasizing their potential
   roles in tumorigenesis. In this analysis, ADAMTS12, F5, and VCAN were
   associated with both overall and disease-free survival in various
   cancers, reinforcing their prognostic significance. The upregulation of
   these genes in GC and their consistent association with poor prognosis
   underscores their clinical relevance in GC and beyond. Additionally, we
   developed a prognostic model using F5 and VCAN, which demonstrated a
   strong ability to predict clinical outcomes in GC patients. Our
   findings not only provide potential prognostic biomarkers for GC but
   also highlight the broader implications of ADAMTS12, F5, and VCAN in
   cancer biology.

   Among ADAMTS12, F5, and VCAN, our attention focused on ADAMTS12.
   ADAMTS12 is an enzyme protein belonging to the metalloproteinase
   family, playing a crucial role in the extracellular matrix [[102]29].
   The term “A Disintegrin-like” in its name signifies its association
   with the ADAM family (A Disintegrin and Metalloproteinase), while
   “Thrombospondin Motifs” refers to the presence of domains similar to
   those found in thrombospondin. The protein encoded by ADAMTS12 exhibits
   metalloproteinase activity, allowing it to cleave and degrade protein
   molecules within the extracellular matrix [[103]30, [104]31]. In both
   physiological and pathological conditions, ADAMTS12 plays a crucial
   role in various biological processes, including cell migration, tissue
   repair, angiogenesis, and the metabolism of collagen and proteoglycans.
   Its regulated activity is closely linked to the onset and progression
   of several diseases, including cancer, arthritis, and cardiovascular
   disorders [[105]32–[106]34]. In the realm of cancer research,
   alterations in the expression levels of ADAMTS12 are often observed,
   correlating with aspects of tumor initiation, progression, and
   prognosis. Through mechanisms influencing extracellular matrix
   remodeling, cell–cell interactions, and signaling pathway regulation,
   ADAMTS12 likely plays a role in the biological processes of tumors
   [[107]35–[108]37]. In addition, Chen et al. reported that analyzed the
   expression of ADAMTS12 in GC tissues and its involvement in the
   glycolysis pathway using bioinformatics and experimental methods. The
   findings demonstrated that ADAMTS12 promotes the proliferation and
   glycolysis of GC cells, while metformin can inhibit these effects. In
   vivo experiments further confirmed that metformin suppresses the
   proliferation and glycolysis of GC cells via ADAMTS12, suggesting that
   ADAMTS12 could be a potential target for metformin therapy in GC
   [[109]38]. Moreover, Hou et al. found that ADAMTS12 expression was
   significantly elevated in gastric cancer samples and correlated with
   poor prognosis. Gene Set Enrichment Analysis (GSEA) indicated that high
   ADAMTS12 expression was associated with cancer and immune-related
   pathways, while low expression was linked to the oxidative
   phosphorylation pathway. Additionally, CIBERSORT analysis revealed a
   positive correlation between ADAMTS12 expression and macrophage
   infiltration, along with a negative correlation with T follicular
   helper cells, suggesting its role in modulating the tumor
   microenvironment and metabolic reprogramming in gastric cancer
   [[110]39]. In this study, we also reported that ADAMTS12 was highly
   expressed in GC specimens and predicted a poor prognosis. Importantly,
   we performed RT-PCR and western blot, confirming the expression of
   ADAMTS12 was distinctly increased in GC specimens, which was consistent
   with the above findings. Our findings suggested ADAMTS12 as a novel
   prognostic and diagnostic biomarker for GC patients. Moreover, we
   performed in vitro assays and found that knockdown of ADAMTS12
   distinctly suppressed the proliferation and invasion of GC cells,
   suggesting it as a tumor promotor in GC.

   There are several limitations that should be acknowledged in this
   study. Firstly, although our analysis extensively utilized
   bioinformatics tools, including functional enrichment, survival
   analysis, and the construction of a prognostic signature, the absence
   of in vivo validation is a notable limitation. In vivo experiments are
   essential for confirming the functional roles of genes like ADAMTS12,
   F5, and VCAN in GC progression. Secondly, while the clinical
   information provided by the TCGA datasets offers valuable insights, it
   is important to recognize the inherent heterogeneity among GC patients.
   Variations in clinical characteristics, treatment regimens, and
   response patterns among individuals may affect the robustness and
   generalizability of our findings in clinical settings. Thirdly, the
   prognostic signature developed using LASSO regression, while offering a
   predictive model, introduces a certain level of complexity. The limited
   number of genes selected (VCAN and F5) may not fully capture the
   multifaceted nature of GC prognosis. Furthermore, the predictive
   performance of the model should be validated in independent cohorts to
   confirm its reliability.

Conclusion

   In this study, we analyzed TCGA datasets to identify differentially
   expressed genes (DEGs) in GC, which are involved in multiple key
   pathways and biological processes. By integrating sequencing data with
   clinical information from GC patients, we identified survival-related
   DEGs, with a particular focus on ADAMTS12, F5, and VCAN. The high
   expression of these genes is closely associated with poor prognosis,
   and the prognostic signature we constructed demonstrates strong
   predictive performance at multiple time points. Further investigation
   of ADAMTS12 confirmed its significant role in GC, with its elevated
   expression correlating closely with clinical and pathological features.
   In conclusion, our findings not only provide insights into the
   molecular characteristics of GC but also offer a novel perspective for
   prognosis assessment based on gene signatures. However, further
   validation and independent studies are needed to confirm the
   reliability of these results for clinical applications.

Supplementary Information

   [111]Supplementary Material 1.^ (479.5KB, docx)

Acknowledgements