VSTM2L mediates the release of extracellular vimentin to drive lymph node metastasis in gastric cancer.

Introduction
Gastric cancer (GC) is one of the most prevalent and lethal malignant tumors within the digestive system [1]. While chemoradiotherapy, surgery, and immunotherapy could improve the survival rate of the patients, many of them ultimately developed into metastatic GC [2]. However, unlike other forms of metastasis (such as hematogenous and implantation), lymph node metastasis (LNM) in GC can occur when the tumor invades the mucosal layer (T1a stage) [3]. A study involving 202,126 GC patients revealed that 5.9% of those at T1a stage exhibited LNM [4]. Furthermore, some patients also presented with GC at stages T3-4N0M0 [5]. These findings suggest that LNM in GC is influenced by specific factors and demonstrates significant individual variability.
Cancer cells mainly spread through the blood and lymphatic systems [6]. The peripheral lymphatic vessels are thin-walled capillary-like tubes without pericyte coverage, making it easier for cancer cells to penetrate them [6]. Studies have shown that approximately 95% of the vessels surrounding tumors invaded by cancer cells are lymphatic vessels [7]. Lymphatic vessels are composed of lymphatic endothelial cells (LECs), which express molecular markers including Lymphatic Vessel Endothelial Hyaluronan Receptor-1 (LYVE-1), Prospero Homeobox Protein 1, Podoplanin, Vascular Endothelial Growth Factor Receptor-3 (VEGFR-3), Neuropilin-2 [8]. Increasing evidence indicates that VEGFR-3-mediated activation of LNM is a key step in inducing lymphatic metastasis [6]. Although several molecular players and signaling pathways that facilitate lymphangiogenesis have been identified, the precise mechanisms underlying LNM is still blurry [9].
In this study, we found that VSTM2L, also known as C20orf102, is distinctively upregulated in lymph node-positive (pN+) primary GC tissues than in node-negative (pN0) tissues. Recent studies have suggested that VSTM2L may serve as a potential biomarker closely associated with cancer cell metastasis and prognosis [10, 11]. However, the biological roles and underlying molecular mechanisms of VSTM2L in GC cells remain unclear. Through immunoprecipitation coupled with mass spectrometry (IP/MS) analysis, we identified vimentin as an interacting protein of VSTM2L. Vimentin, an intermediate filament protein, is recognized for its intracellular structural properties, as well as its role in enhancing tumor malignancy through involvement in epithelial-to-mesenchymal transition (EMT) and metastasis [12]. In recent years, extracellular functions of vimentin have been proposed [13], and in this study, we demonstrated that VSTM2L promotes the release of extracellular vimentin, thus then promote VEGFR-3-mediated activation of lymphangiogenesis.

Materials and methods

Tissue samples and cell lines
Tumor tissues and matched adjacent normal tissues were collected from 62 GC patients who underwent gastrectomy at the First Affiliated Hospital of Nanjing Medical University (Nanjing, Jiangsu, China). Histopathological evaluation indicated that among the 62 GC cases, 27 were pN0 and 35 were pN+. None of the patients had received any prior treatment before surgery. All tissue specimens were immediately snap-frozen in liquid nitrogen and stored at − 80 °C until further analysis. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Nanjing Medical University, and written informed consent was obtained from all participants prior to tissue collection.
For in vitro experiments, five GC cell lines (MKN45, HGC27, AGS, MKN28, and SNU-719) and one normal human gastric mucosal epithelial cell line (GES-1) were used. MKN45, HGC27, AGS, and GES-1 were acquired from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). MKN28 and SNU-719 were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Human lymphatic endothelial cells (HLECs) were sourced from ScienCell Research Laboratories (San Diego, CA, USA).
GES-1, HGC27, MKN28, MKN45, and SNU-719 were maintained in RPMI-1640 medium (Wisent, Montreal, Canada), whereas AGS cells were cultured in F12K medium (Wisent). All media were supplemented with 10% fetal bovine serum (Wisent) and 1% penicillin/streptomycin (Wisent). HLECs were grown in endothelial cell medium containing 5% FBS. All cell lines were incubated at 37 °C in a humidified atmosphere with 5% CO₂.

Transfection
Small interfering RNAs (siRNAs) targeting VSTM2L and vimentin, along with a negative control siRNA, were obtained from GenePharma (Shanghai, China). To establish stable knockdown cell lines, short hairpin RNA (shRNA) sequences directed against VSTM2L or vimentin were subcloned into lentiviral expression vectors (GenePharma). Additionally, VSTM2L expression plasmids were constructed using a pCDNA3.1 vector. For the generation of 6×His-tagged and 3×Flag-tagged constructs, the coding sequences of His-tagged VSTM2L (His-VST) and Flag-tagged vimentin (Flag-VIM) were subcloned into a pCMV vector. Transfection was performed according to standard protocols. Stable cell lines were selected with puromycin, and transfection efficiency was confirmed by Western blot analysis. All the sequences are listed in Supplementary Table S1.

Quantitative RT‑PCR
Total RNA was extracted from cultured cells and tissues using a previously described method [14]. cDNA synthesis and quantitative PCR (qPCR) were performed according to the manufacturers’ instructions. The relative gene expression levels, normalized to the internal control GAPDH, were determined using the ΔCt method. The sequences of the qPCR primers are listed in Supplementary Table S2.

Western blot and immunohistochemistry (IHC)
Western blot and IHC analyses were carried out using previously described methods [15]. IHC staining was evaluated by calculating a staining index that incorporated both the intensity of staining and the proportion of positive GC cells, following an established scoring system. All antibody information is summarized in Supplementary Table S3.

Mouse popliteal LNM model
The popliteal LNM model was established as previously described [15]. In brief, 2 × 106 lentivirus-transduced MKN45 cells were inoculated into the footpads of 4-week-old female BALB/c nude mice. All mice were housed under specific pathogen-free conditions. Eight weeks post-injection, the mice were euthanized, and primary tumors and popliteal lymph nodes were harvested for hematoxylin and eosin (H&E) staining and IHC analysis. Lymph node volume was calculated using the formula: Volume (mm³) = length (mm) × [width (mm)] ² × 0.52.

Statistical analysis
All quantitative data are expressed as mean ± standard deviation (SD). Statistical significance was evaluated using GraphPad Prism 6 (GraphPad Software, San Diego, CA, USA). For comparisons between two groups, a two-tailed Student’s t-test was applied; for comparisons involving more than two groups, one-way ANOVA was employed. Differences were considered statistically significant at P < 0.05.

Results

VSTM2L was associated with LNM and predicted poor prognosis in GC
To identify genes associated with LNM in GC, we analyzed RNA sequencing data from The Cancer Genome Atlas Program (TCGA) database, comparing pT3-4 N+ versus pT3-4N0, and pN+ versus pN0 groups. This analysis revealed multiple differentially expressed genes (|log2FC| > 1, p< 0.01; Fig. 1A, B). We further applied logistic regression to estimate odds ratios (OR) for LNM association, which identified several risk genes (Fig. 1C). Intersection of these gene sets yielded six candidate genes (Fig. 1D), among which VSTM2L was the only risk gene (OR = 1.33, p< 0.001). Validation using TCGA and our in-house dataset (n = 62) confirmed that VSTM2L RNA expression was significantly higher in pN+ than in pN0 primary GC tissues, though no difference was observed between tumor and normal tissues (Fig. 1E, F). Consistently, immunohistochemistry and western blot analyses showed markedly elevated VSTM2L protein levels in pN + GC tissues compared to pN0 samples (Fig. 1G–I). Furthermore, VSTM2L expression positively correlated with lymphatic vessel density (Fig. 1J, K). Clinically, elevated VSTM2L expression was significantly associated with poorer overall survival (HR = 1.55, p = 0.0086) and recurrence-free survival (HR = 2.31, p = 0.014; Fig. 1L–N). Collectively, these findings establish VSTM2L as a promising predictor of LNM and unfavorable prognosis in GC.

VSTM2L promoted cell proliferation, invasion and lymphangiogenesis
Compared to the normal gastric mucosal epithelial cell line GES-1, the GC cell lines HGC27 and MKN45 demonstrated significantly elevated expression levels of VSTM2L (Fig. 2A). Consequently, we performed knockdown and overexpression of VSTM2L in the HGC27 and MKN45 cell lines to investigate its impact on cellular functions (Fig. 2B-C). It showed that knockdown of VSTM2L markedly reduced colony formation (Fig. 2D), whereas VSTM2L overexpression enhanced it (Fig. 2E; Fig. S1A). Transwell assays further demonstrated that VSTM2L depletion impaired cell invasion, while its overexpression promoted this phenotype (Fig. 2F, G; Fig. S1B). In lymphangiogenesis models, conditioned medium (CM) from VSTM2L-silenced GC cells inhibited tube formation in HLECs, whereas CM from VSTM2L-overexpressing cells strongly induced it (Fig. 2H, I; Fig. S1C). Taken together, these results indicate that VSTM2L critically promotes GC proliferation, invasion, and lymphangiogenesis.

VSTM2L upregulated the levels of vimentin and focal adhesion-related proteins
To elucidate the molecular mechanism of VSTM2L, we first assessed its impact on EMT, a hallmark of metastasis. Surprisingly, silencing VSTM2L did not alter the expression of classic EMT markers such as E-cadherin, N-cadherin, Slug, Snail, MMP9, or MMP2 (Fig. 3A). However, vimentin, a key EMT regulator, was consistently downregulated upon VSTM2L knockdown and upregulated upon its overexpression (Fig. 3A, C). Given vimentin’s established role in invasion via focal adhesion remodeling, we examined key adhesion components [16]. We found that integrin β1, p-FAK, RhoA, vinculin, and filamin A were downregulated following VSTM2L silencing (Fig. 3B) and increased upon its overexpression (Fig. 3C). Furthermore, VSTM2L knockdown induced depolymerization of F-actin, a cytoskeletal component whose dynamics are regulated by vimentin [17] (Fig. 3D). Gene set enrichment analysis indicated that VSTM2L-associated genes were enriched in distinct biological processes in pN+ versus normal samples (Fig. 3E, F). These results collectively suggested that VSTM2L promotes malignancy by regulating vimentin. Notably, while VSTM2L did not affect vimentin mRNA levels (Fig. 3G; Fig. S2A), it significantly shortened vimentin protein half-life upon knockdown (Fig. 3H-I; Fig. S2B-C). Immunofluorescence confirmed the reduction in vimentin protein without altered intracellular localization after VSTM2L silencing (Fig. 3J; Fig. S2D).

VSTM2L interacted with vimentin
To uncover the functional mechanism of VSTM2L in GC, we expressed His-tagged VSTM2L in MKN45 cells and performed co-immunoprecipitation (co-IP) with an anti-His antibody, followed by silver staining and mass spectrometry of the precipitates (Fig. 4A). MS analysis identified vimentin as a high-confidence interacting partner of VSTM2L (Fig. 4B; Supplementary Table S4). This interaction was confirmed by Western blot, which showed that exogenous His-VSTM2L specifically bound to endogenous vimentin (Fig. 4C, D). Reciprocal co-IP assays further validated the association, demonstrating that both endogenous vimentin and exogenous Flag-tagged vimentin interact with VSTM2L (Fig. 4C–E). To determine if this binding is direct, we conducted an in vitro GST pull-down assay. Purified His-VSTM2L specifically bound to immobilized GST-vimentin, but not to GST alone, confirming a direct protein-protein interaction (Fig. 4F). Additionally, immunofluorescence microscopy revealed substantial co-localization of VSTM2L and vimentin within GC tissues and cells (Fig. 4G, H; Fig. S2E). Collectively, these results demonstrate that VSTM2L binds directly to vimentin and enhances its protein stability.

VSTM2L promoted vimentin secretion in GC
Immunohistochemistry and Western blot analyses consistently demonstrated that vimentin expression was significantly higher in pN+ tumors compared to pN0 tissues, as well as in cancerous versus adjacent non-tumorous tissues (Fig. 5A–C). Correlation analyses further revealed positive associations among vimentin expression, VSTM2L levels, and lymphatic vessel density (Fig. 5D, E). Clinically, elevated vimentin expression correlated with poorer overall survival (HR = 1.45, p = 0.025; Fig. 5F). Beyond its intracellular roles, extracellular vimentin (ECV) has been implicated in receptor signaling and microenvironment remodeling [13]. Hence, we further evaluated whether VSTM2L participate in vimentin secretion. First, conditioned medium from GC cells transduced with VSTM2L-targeting siRNA was collected. The secreted total proteins and extracellular vimentin were subsequently analyzed using Ponceau Red staining and Western blot, respectively. As illustrated in Fig. 5G, ECV levels were significantly reduced in the VSTM2L-silencing groups compared to the control groups. Conversely, overexpression of VSTM2L led to an increase in ECV levels (Fig. 5H). Consistent with these findings, knockdown of VSTM2L resulted in a decrease in vimentin abundance within the cell supernatants as detected by ELISA assays, while overexpression of VSTM2L produced an opposing effect (Fig. 5I, J). Previous studies have indicated that vimentin is mainly secreted extracellularly through the type III unconventional protein secretion (UPS) pathways [18]. Consequently, we further investigated whether vimentin in GC cells is also secreted through a similar mechanism. Utilizing previously established methods and optimized concentrations of inhibitors for assessing endothelial cell secretion of ECV, we evaluated the effects of monensin, nigericin, and neomycin [19]. Treatments known to impact the efficacy of the UPS pathway (specifically neomycin and nigericin), as opposed to those affecting Golgi-dependent transport such as monensin, resulted in a decreased concentration of ECVs in the culture medium when compared to controls (Fig. 5K). These results suggest that VSTM2L promotes vimentin secretion in GC cells, likely through a UPS-dependent mechanism.

VSTM2L promoted LNM in a vimentin-dependent manner
To define the functional contribution of vimentin, we knocked it down in GC cells using siRNA (Fig. S3A). Loss of vimentin significantly suppressed GC cell proliferation and invasion, as shown by colony formation and Transwell assays (Fig. S3B, C). It also impaired the ability of GC cells to stimulate migration, sprouting, and tube formation in HLECs (Fig. S4A, B), supporting a critical role for vimentin in lymphangiogenesis. Accordingly, treatment with recombinant extracellular vimentin (rVIM) enhanced these lymphangiogenic capacities in HLECs (Fig. S4C, D). To determine whether VSTM2L acts through vimentin, we performed rescue experiments. Silencing vimentin abolished the pro-tumorigenic effects of VSTM2L overexpression on GC cell proliferation and invasion (Fig. S5A, B). Similarly, rVIM supplementation reversed the inhibitory effect of VSTM2L knockdown on HLEC tube formation and migration (Fig. S5C, D). We further validated these findings in an in vivo popliteal LNM model (Fig. 6A). Mice injected with VSTM2L-overexpressing MKN45 cells developed larger popliteal lymph nodes, an effect that was rescued by concurrent vimentin knockdown (Fig. 6B–C). H&E staining confirmed metastatic lesions in lymph nodes, demonstrating that vimentin silencing abrogated the pro-metastatic effect of VSTM2L (Fig. 6D). Immunohistochemistry of footpad tumors further showed that VSTM2L upregulated Ki67, platelet endothelial cell adhesion molecule-1 (CD31), LYVE-1, and vimentin, but these effects were reversed upon vimentin knockdown (Fig. 6E–I). Taken together, these data establish that VSTM2L promotes GC proliferation, invasion, and lymphangiogenesis by enhancing vimentin expression and secretion.

Extracellular vimentin activated the VEGFR-3 signaling in HLEC
VEGFR-3, predominantly expressed on LECs, is a central regulator of developmental lymphangiogenesis. Its ligands, VEGF-C and VEGF-D, often secreted by cancer cells, bind and activate VEGFR-3, initiating downstream signaling cascades such as PI3K/AKT, ERK1/2, and JNK, which collectively promote LEC proliferation, migration, and survival [6] (Fig. 7A). Based on our observation that rVIM stimulates lymphangiogenesis and HLEC migration, we hypothesized that extracellular vimentin modulates VEGFR-3 expression and/or activation. Supporting this, treatment of HLECs with rVIM significantly enhanced phosphorylation of VEGFR-3, ERK1/2, and AKT, comparable to the effect of VEGF (Fig. 7B). Immunofluorescence further revealed that both VEGF and vimentin promoted VEGFR-3 expression and its surface localization in HLECs (Fig. 7C). Importantly, conditioned medium from VSTM2L-silenced GC cells reduced phosphorylation of these signaling effectors, an effect that was substantially reversed by rVIM supplementation (Fig. 7D, E). Together, these findings establish extracellular vimentin as a functional modulator of the VEGFR-3 signaling axis in lymphangiogenesis.

Discussion
The TNM staging system, based on tumor invasion depth (T), lymph node metastasis (N), and distant metastasis (M), is fundamental to diagnosing, treating, and prognosticating GC [20]. The presence of LNM signifies enhanced tumor invasiveness and metastatic potential, portending a poorer prognosis and higher recurrence risk [21]. Moreover, LNM status is critical in managing early GC, determining whether a patient undergoes radical gastrectomy with D2 lymphadenectomy or qualifies for endoscopic resection [22]. While numerous diagnostic models using clinicopathological or imaging features have been developed to evaluate LNM, the molecular drivers remain incompletely characterized [23, 24]. Here, we report that VSTM2L expression is significantly elevated in GC patients with LNM. Functional assays demonstrated that VSTM2L enhances tumor proliferation, invasion, and LNM. Mechanistically, we found that VSTM2L binds to vimentin, promoting its expression and secretion, which in turn stimulates LNM.
Emerging evidence has documented oncogenic roles of VSTM2L across various cancers. In ovarian cancer, VSTM2L contributes to anoikis resistance and serves as a potential biomarker for metastasis and prognosis [11]. In prostate cancer, VSTM2L localizes primarily to mitochondria and functions as a key regulator of ferroptosis, correlating positively with disease progression [10]. In GC, in silico analyses have suggested VSTM2L as a component of prognostic risk models [25]. Notably, VSTM2L has been identified as a secreted protein that interacts with Humanin, and its soluble form has been detected in the blood of patients with cholangiocarcinoma [26, 27]. Building on this foundation, our study elucidates the function and molecular mechanism of intracellular VSTM2L in promoting GC lymphatic metastasis. Nevertheless, it remains to be determined whether VSTM2L can be secreted from GC cells or directly influences HLECs.
In this study, we demonstrated that VSTM2L promotes LNM in GC in a vimentin-dependent manner. Vimentin, a widely recognized marker of mesenchymal phenotype, has been consistently associated with enhanced tumor growth, invasion, and unfavorable prognosis in GC [28, 29]. Consistent with this, our data showed that vimentin expression was significantly elevated in GC patients with LNM. Through IP/MS analysis and Western blot validation, we further established that VSTM2L physically interacts with vimentin and enhances its protein stability. The regulation of vimentin is known to involve multiple mechanisms, particularly post-translational modifications such as O-GlcNAcylation, phosphorylation, and ubiquitination [30]. For instance, Chen et al. reported that Ladinin 1 competitively binds vimentin and disrupts its interaction with the E3 ubiquitin ligase MAEA, thereby reducing K48-linked ubiquitination and stabilizing vimentin in GC cells [31]. Similarly, Jang et al. revealed that Polo-like kinase 1-mediated phosphorylation of vimentin activates TGF-β signaling, promoting metastasis and immune suppression via Programmed Death-Ligand 1 upregulation in lung adenocarcinoma [32]. In light of these findings, we hypothesize that the binding of VSTM2L to vimentin may influence its post-translational modification landscape, particularly ubiquitination, to enhance vimentin stability and drive metastatic progression.
Although most studies have primarily focused on intracellular vimentin, there is a growing recognition of its role as an extracellular protein. Vimentin can be found exposed at the cell surface in an oligomeric form or secreted into the extracellular environment in both soluble and vesicle-bound forms [13]. Despite these observations, the contribution of ECV to cancer progression remains incompletely defined. Studies indicate that tumor endothelial cell-derived vimentin promotes tumor angiogenesis, immune infiltration, and immunosuppression, partly by mimicking Vascular endothelial growth factor activity [18]. Moreover, vimentin-positive extracellular vesicles have been shown to induce EMT through the delivery of specific protein cargo that facilitates cellular remodeling [33]. Clinically, vimentin is detectable in the serum of cancer patients. In GC, circulating tumor cells expressing surface vimentin are associated with poor treatment response and reduced survival [34]. Similarly, in gastrointestinal stromal tumors, macrophage-like circulating tumor cells positive for surface vimentin have emerged as novel and robust biomarkers for assessing metastatic risk [35]. Given its pleiotropic nature and context-dependent functionality, ECV represents a promising molecular target for theragnostic strategies. Supporting this, anti-ECV antibodies have been shown to inhibit angiogenesis in vitro and in vivo [18]. Encouragingly, a veterinary study in dogs with spontaneous bladder cancer demonstrated that vaccination targeting ECV elicited promising therapeutic responses, further underscoring its translational potential [18].

Conclusions
Thus, this study highlighted the VSTM2L-vimentin pathway as a key driver of LNM in GC. Our findings elucidated a mechanism whereby VSTM2L stabilizes vimentin and facilitates its secretion, enabling the activation of pro-lymphangiogenic VEGFR-3 signaling. Therefore, targeting this axis, specifically VSTM2L or extracellular vimentin, represents a viable and promising therapeutic strategy to inhibit LNM and improve patient outcomes.

Supplementary Information
Below is the link to the electronic supplementary material.