Identification of STAT1 as a biomarker for lymph node metastasis in esophageal squamous cell carcinoma

Abstract

Lymph node metastasis in esophageal squamous cell carcinoma (ESCC) is associated with poor prognosis, and screening for potential biomarkers of metastasis is significant for improving the prognosis of ESCC. After batch effect removal and integration of two ESCC transcriptome datasets (GSE157804, GSE118493), differentially expressioned gene analysis was performed. The intersection of differentially expressed genes was obtained to identify characteristic genes associated with lymph node metastasis in ESCC. GO/KEGG enrichment analyses were conducted for the characteristic genes. The top 10 pathways and pathway genes based on q-value were selected to construct a PPI network. Hub genes were identified and intersected with the characteristic genes to determine key genes. GO/KEGG enrichment, expression level analysis, and immune infiltration analysis were performed for the key genes. Immunohistochemistry was used to validate the expression levels of key genes in clinical samples. A cell model with regulated expression of key genes was established to verify their impact on the invasive and migratory capabilities of ESCC cells and epithelial-mesenchymal transition (EMT), as well as ECM‑related gene expression. A total of 542 characteristic genes associated with lymph node metastasis in ESCC were identified through differential gene analysis, which were mainly enriched in extracellular matrix-related pathways. PPI network analysis identified 63 hub genes, and the intersection with characteristic genes resulted in 20 key genes, which were also primarily enriched in extracellular matrix-related pathways. Expression level analysis showed that key genes such as STAT1, STAT2, IFIT1, and IFIT3 were highly expressed in metastatic ESCC tissues compared to primary sites and adjacent non-tumor tissues. Immune infiltration analysis revealed a significant positive correlation between IFIT3, STAT1, STAT2, and M1 macrophage polarization. Immunohistochemical analysis indicated high expression of STAT1 in ESCC patient samples with lymph node metastasis. Upregulation of STAT1 in ESCC cells enhanced cell invasion, migration, EMT levels and ECM‑related gene expression, while downregulation of STAT1 expression suppressed these processes. Consistently, an additional gain‑of‑function experiment further confirmed these promoting effects. STAT1 serves as a biomarker for lymph node metastasis in ESCC. The expression level of STAT1 is positively correlated with the degree of lymph node metastasis in ESCC clinical samples. Overexpression of STAT1 promotes the invasion, migration, EMT and ECM‑related gene expression of ESCC cells, while the downregulation of STAT1 inhibits these effects.

Introduction

Esophageal squamous cell carcinoma (ESCC) is one of the histopathological subtypes of esophageal cancer, originating from epithelial cells and usually occurring in the upper and middle esophagus. The incidence of this disease continues to increase globally (Yang et al. 2024). Despite progress in prevention, screening, early diagnosis, and treatment that has reduced mortality, the prognosis for ESCC patients remains dismal, with a five-year survival rate below 25% (Zhao et al. 2023, Zhang et al. 2023). Consequently, investigating the molecular mechanisms is crucial for improving prognosis and represents a key focus in the field.

Metastasis is a coordinated, non-random process and a principal cause of poor prognosis and cancer-related deaths (Kiri and Ryba 2024). Cancers that metastasize become systemic illnesses, requiring combined-modality therapies such as chemotherapy, targeted agents, and immunotherapy. This complexity increases treatment difficulty and the risk of treatment failure. Reported five-year survival rates for metastatic solid tumors typically fall between 5% and 30% (Kratzer et al. 2023). ESCC is characterized by its high invasiveness and metastatic potential, with lymph node (LN) metastasis being the predominant route. Furthermore, the lack of early symptoms means the majority of patients are diagnosed at a metastatic stage, making management extremely difficult (Jiang et al. 2023, Ko et al. 2023). The effectiveness of surgery is also limited; even after curative resection, about 74% of patients experience metastasis, resulting in a mere 39.8% survival rate (Li et al. 2024). This aggressive nature of ESCC may also underpin its frequent development of resistance to radiotherapy and chemotherapy, as well as its ability to evade immune surveillance (Zhao et al. 2023, Zhang et al. 2023). Aberrant expression of various genes, including miR-137, miR-375, miR-4319, and BBOX1-AS1, has been implicated in ESCC progression, such as proliferation, invasion, and stemness (Li et al. 2022, Wu et al. 2021, Hu et al. 2020, Hu et al. 2023). However, the mechanisms governing ESCC metastasis are still not well elucidated. Identifying reliable biomarkers associated with metastasis is therefore essential for developing new diagnostic and therapeutic approaches to enhance clinical outcomes.

In this study, we utilized bioinformatics analyses to identify STAT1 as a potential biomarker for metastatic ESCC and confirmed its expression profile in clinical samples. We further investigated its function and demonstrated the pro-metastatic role of the Signal transducer and activator of transcription 1 (STAT1) in a cellular model.

Materials and methods

Data acquisition and analysis

The gene expression datasets GSE157804 and GSE118493 were downloaded from the Gene Expression Omnibus (GEO) database. After batch effect removal and merging, the data were divided into three groups according to sample information: primary esophageal squamous cell carcinoma (PT), metastatic esophageal squamous cell carcinoma (MT), and adjacent non-cancerous tissue (NC). Differential expression analysis was performed using the limma R software package. Differentially expressed genes (DEGs) were screened with thresholds of P < 0.05 and |log₂FC| > 0.585. Separate comparisons were conducted for PT versus NC, MT versus NC, and PT versus MT. The EVenn online tool was used to generate Venn diagrams for identifying characteristic genes across these comparisons.

Functional enrichment analysis

Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed for the functions and associated pathways of the DEGs using the cluster Profiler package of R software. GO terms and pathways with P < 0.05 were considered significant.

Protein-protein interaction (PPI) network analysis

The top 10 most significantly enriched GO terms/KEGG pathways and their associated genes were selected. A PPI network was constructed using the STRING database and visualized within Cytoscape software. Network topology parameters were calculated and analyzed using the CytoNCA plugin.

Immune infiltration analysis

Based on the core algorithm of CIBERSORT, the immune infiltration status of the GSE157804 and GSE118493 datasets was calculated using markers for 22 types of immune cells provided by the CIBERSORTx website. Correlation analysis was performed using Pearson’s correlation coefficient, and results were visualized with the ggplot2 package.

Immunohistochemistry (IHC)

Formalin-fixed, paraffin-embedded ESCC tissue samples (n = 4) were obtained from patients diagnosed with ESCC at the First Affiliated Hospital of Kunming Medical University and underwent treatment there (Table 1). This study was approved by the Ethics Committee of the First Affiliated Hospital of Kunming Medical University, and all experiments were conducted in accordance with relevant guidelines and regulations. Written informed consent was obtained from all patients. IHC staining was performed following standard protocols using a primary antibody against STAT1 (abcam, ab109320, 1:100) and a horseradish peroxidase (HRP)-conjugated secondary antibody (Goat Anti-Rabbit IgG H&L, abcam, ab6721, 1:1000). Two experienced pathologists, blinded to sample identity, independently evaluated the immunostaining. The H‑score for nuclear and/or cytoplasmic STAT1 staining was calculated by multiplying DAB intensity (0 = negative, 1 = weak, 2 = moderate, 3 = strong) by the percentage of positive tumor cells, yielding a score from 0 to 300. The formula used was: H‑score = (1 × % weak) + (2 × % moderate) + (3 × % strong). The final H‑score per sample was the average of the two evaluators’ scores.

Table 1.

Patient demographic information of paraffin sections

Sample ID	Pathological diagnosis results	Tissue type
MT1	Moderately differentiated squamous cell carcinoma with lymph node metastasis	Cancer tissue: Primary site
		Lymph node: metastatic site
MT2	Moderately to poorly differentiated squamous cell carcinoma with lymph node metastasis	Cancer tissue: Primary site
		Lymph node: metastatic site
MT3	Highly to moderately differentiated squamous cell carcinoma with lymph node metastasis	Cancer tissue: Primary site
		Lymph node: metastatic site
PT1	Radical specimens of esophageal cancer after neoadjuvant treatment, showing only carcinoma in situ without residual invasive cancer, with lymph node metastasis	Cancer tissue: Primary site
		Lymph node: metastatic site

Cell culture and transfection

KYSE-30 and KYSE-150 cells were purchased from Wuhan Procell Life Science & Technology Co., Ltd. Both cell lines were authenticated by short tandem repeat (STR) profiling, which confirmed a 100% match with the corresponding reference profiles in the ExPASy (KYSE-30, CVCL_1351) and DSMZ (KYSE-150, ACC-375) databases. The STR analysis was performed by the certified testing facility at Procell Life Science & Technology Co., Ltd Cells were cultured in a 1:1 mixture of Ham’s F-12 and RPMI-1640 media, supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin, at 37 °C in a humidified 5% CO₂ incubator.

For transfection, cells were seeded in six-well plates. Upon reaching appropriate confluence, they were transfected with STAT1 knockdown vectors (sh-STAT1), STAT1 overexpression plasmids (OE-STAT1), or their respective negative controls (sh-NC or OE-NC) using Lipofectamine 3000 reagent, according to the manufacturer’s instructions. Briefly, plasmids were diluted with serum-free culture medium, gently mixed, and incubated at room temperature for 5 min. Lipofectamine 3000 was diluted with 5 µL serum-free culture medium, gently mixed, and incubated at room temperature for 5 min. The vector and lipo3000 were gently mixed, incubated at room temperature for 20 min, added to the six-well plates, and incubated for 48 h in the culture incubator.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted from cells using the RNAiso Plus reagent (Takara, Cat# 9108), and Recombinant DNase I (Takara, Cat# 2270 A) was used to remove DNA contamination. Complementary DNA (cDNA) was synthesized using the PrimeScript™ RT Reagent Kit (Takara, Cat# RR037A). Quantitative PCR was performed using TB Green^® Premix Ex Taq™ II (Takara, Cat# RR820A). Relative expression was normalized to GAPDH and calculated using the 2^−ΔΔCt method. Primer information is listed in Table 2:

Table 2.

Primers sequences used for quantitative real-time PCR

Gene name	Primer sequences (5′-3′)
STAT1 forward STAT1 reverse BMP1 forward	CCCAGAACGGAGGCGAACC TGCGAATGATGTCAGGGAAAGTAAC CCACGTTTCCATCGTTCGTG
BMP1 reverse	TGCCCCTGGAGAATGTGTTC
COL1A1 forward	CAGCCGCTTCACCTACAGC
COL1A1 reverse COL3A1 forward COL3A1 reverse LOX forward LOX reverse MMP13 forward MMP13 reverse POSTN forward POSTN reverse GAPDH forward GAPDH reverse	TTTTGTATTCAATCACTGTCTTGCC GGATCAGGCCAGTGGAAATGTAAAGA CTTGCCTGTTCGATATTCAAAGACTGTT AACATCATCCTGGGTTATTC CTATTATGGCACATGGTTTC GCCATTACCAGTCTCCGAGG TACGGTTGGGAAGTTCTGGC TGCCCAGCAGTTTTGCCCAT CGTTGCTCTCCAAACCTCTA CATGTTCGTCATGGGTGTGAACC GGTCATGAGTCCTTCCACGATACC

Western blot

Cells were lysed using RIPA lysis buffer (Thermo Scientific, Cat# 89901) to extract total protein, and concentrations were determined using the BCA assay kit (Thermo Scientific, Cat# 23227). Proteins were separated by SDS-PAGE gel electrophoresis, transferred to PVDF membranes, and blocked with 5% skim milk in TBST. The membranes were then incubated overnight at 4 °C with primary antibodies against: STAT1 (abcam, ab281999, 1:1000), E-cadherin (Invitrogen, MA5-12547, 1:500), N-cadherin (Invitrogen, 33-3900, 1:250), Vimentin (abcam, ab92547, 1:1000). After washing, membranes were incubated with an HRP-conjugated secondary antibody (Goat Anti-Rabbit IgG, abcam, ab6721, 1:1000). Protein bands were visualized using a chemiluminescent substrate, and band intensities were quantified using ImageJ software.

Transwell assay

Matrigel matrix (Corning, 354234) was prepared according to the manufacturer’s instructions, and the Transwell upper chambers were seeded with serum-free cell suspensions. The lower chambers were filled with complete medium containing 10% FBS. After 24 h of incubation, the cells were fixed with 400 µL paraformaldehyde and stained with 400 µL crystal violet staining solution. After washing with PBS, the invasive cells were observed under a microscope. The number of invaded cells was quantified by ImageJ software.

Wound healing assay

Adherent cells were detached using trypsin, resuspended, and seeded in 6-well plates at a density of 2 × 10⁵ cells per well. Once the cells reached full confluence, the culture medium was removed, and the cells were washed once with PBS. Three straight lines were drawn in each well of the 6-well plate using a 200 µL pipette tip. The cells were then washed twice with PBS to remove dislodged cells and washed twice more with PBS to ensure any remaining cells were removed. The remaining cells were photographed under a microscope. The cells were then treated according to their respective groups, cultured routinely, and photographed again after 24 h. The wound area at 0 and 24 h was quantified using ImageJ software, and the wound closure rate was calculated as (initial wound area – wound area at 24 h) / initial wound area × 100%.

Statistical analysis

All experiments were independently repeated three times, and statistical analyses were performed using GraphPad Prism 9. Quantitative data conforming to a normal distribution are presented as mean ± standard deviation (mean ± SD). Independent t-tests were used for comparisons of means between two groups, while one-way analysis of variance (ANOVA) was used for comparisons among multiple groups. A p value of less than 0.05 was considered statistically significant.

Results

Differential gene analysis

Differential expression analysis identified 5,137 DEGs in primary ESCC tissues compared to adjacent non-tumor tissues (Fig. 1A). Comparison of metastatic ESCC tissues with adjacent non-tumor tissues identified 1,087 DEGs (Fig. 1B), while the comparison between metastatic and primary ESCC tissues yielded 678 DEGs (Fig. 1C). A Venn diagram generated using the EVenn online tool identified 542 characteristic genes common to metastatic ESCC samples (Fig. 1D).

Fig. 1.

Differential gene analysis. A Volcano plot of differentially expressed genes in PT versus NC. B Volcano plot of differentially expressed genes in MT versus NC. C Volcano plot of differentially expressed genes in PT versus MT. D Venn diagram identifying characteristic genes of metastatic esophageal squamous cell carcinoma

GO/KEGG enrichment analysis of characteristic genes and key gene screening

GO enrichment analysis revealed that the characteristic genes were primarily enriched in biological processes (BP) related to external encapsulating structure organization, extracellular matrix organization, extracellular structure organization, collagen fibril organization, and collagen metabolic process. Cellular components (CC) included the collagen-containing extracellular matrix, external side of plasma membrane, and collagen trimer, while molecular functions (MF) were enriched in extracellular matrix structural constituent and various peptidase activities (Fig. 2A). KEGG pathway analysis indicated significant enrichment in Protein digestion and absorption, extracellular matrix (ECM)-receptor interaction, Transcriptional misregulation in cancer (Fig. 2B). Given the strong ECM association, a PPI network was constructed using genes from the top 10 most significant GO/KEGG terms (Fig. 2C). We screened for the top 10% of hub genes based on degree coefficient, obtaining 63 hub genes. By creating a Venn diagram using the EVenn online tool, we identified the intersection of hub genes and characteristic genes, resulting in 20 key genes for further analysis (Fig. 2D).

Fig. 2.

GO/KEGG enrichment analysis of hub genes and key gene screening. A GO enrichment terms of characteristic genes. B KEGG enrichment terms of characteristic genes. C PPI network of top 10 GO term-related genes and KEGG genes. D Identification of key genes through intersection of hub genes and characteristic genes

Analysis of key genes

GO and KEGG analysis of the 20 key genes further confirmed their strong association with ECM-related processes and structures. The GO analysis showed these key genes were primarily enriched in BP such as collagen fibril organization, extracellular matrix organization, extracellular structure organization, and external encapsulating structure organization; CC such as fibrillar collagen trimer, banded collagen fibril, complex of collagen trimers, collagen trimer, and collagen-containing extracellular matrix; and MF such as platelet-derived growth factor binding, extracellular matrix structural constituent, extracellular matrix structural constituent conferring tensile strength, and growth factor binding (Fig. 3A). The KEGG enrichment results indicated that the key genes were primarily enriched in signaling pathways such as Protein digestion and absorption, Relaxin signaling pathway, Cytoskeleton in muscle cells, Proteoglycans in cancer, and Toll-like receptor signaling pathway (Fig. 3B). Next, the expression levels of key genes in three types of tissues were validated. Among them, COL1A1, STAT1, STAT2, COL3A1, IFIT1, IFIT3, RSAD2, TLR2, TGFB1, ISG15, BMP1, COL5A1, and OAS2 were all expressed at higher levels in the MT group compared to the NC and PT groups (Fig. 4A), although some of these genes may not show statistically significant differences. Immune infiltration analysis revealed a significant positive correlation between IFIT3, STAT1, STAT2 and M1 polarization of macrophages (Fig. 4B). Since STAT1 exhibited the highest expression in MT tissues and correlated with M1 polarization of macrophages, it was chosen as the main subject of observation.

Fig. 3.

GO/KEGG enrichment analysis of key genes. A GO enrichment terms of key genes. B KEGG enrichment terms of key genes

Fig. 4.

Expression levels of key genes and immune infiltration analysis. A Expression levels of key genes in NC, PT, and MT samples. B Immune infiltration analysis results of key genes

Validation in clinical samples

To verify the expression of STAT1 in clinical samples, we collected cancer tissue and lymph node tissue samples (metastatic site) (Lymph node) from patients diagnosed and treated in our hospital for immunohistochemical analysis. Patient MT1 was clinically diagnosed with moderately differentiated ESCC with lymph node metastasis. Patient MT2 was clinically diagnosed with moderately to poorly differentiated ESCC with lymph node metastasis. Patient MT3 was clinically diagnosed with highly to moderately differentiated ESCC with lymph node metastasis. Patient PT1 underwent neoadjuvant therapy and achieved a radical cure, with only carcinoma in situ observed and no residual invasive cancer or lymph node metastasis. Immunohistochemical staining was quantified using the H-Score method. As shown in Fig. 5, the H-Score was significantly higher in the MT group than in the PT group. Among all samples, MT2 (moderately to poorly differentiated ESCC with metastasis) showed the highest H-Score for STAT1, while PT1 (post-neoadjuvant therapy with no residual cancer) showed the lowest. These results suggest that the STAT1 positivity is associated with the clinical phenotype of cancer cell metastasis and indicating that STAT1 may be a marker and crucial regulatory molecule of ESCC metastasis.

Fig. 5.

Immunohistochemical validation of STAT1 expression in primary (Cancer) and metastatic (Lymph node) tissues from ESCC patients (MT1, MT2, MT3, PT1). Representative images are shown at 10× and 40× magnification as indicated. Semi-quantitative H-scores were calculated based on staining intensity and percentage of positive tumor cells. n = 4 patients, data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

STAT1 promotes invasion, migration, EMT, and ECM remodeling in ESCC cells

We investigated the expression levels of STAT1 in ESCC cell lines. We found that STAT1 expression was relatively low in the well-differentiated KYSE-30 cells and relatively high in the poorly differentiated KYSE-150 cells (Fig. 6A). Next, we modulated STAT1 expression levels in KYSE-30 and KYSE-150 cells, and the results showed that transfection with OE-STAT1 increased STAT1 mRNA and protein levels in KYSE-30 cells, while sh-STAT1 reduced them in KYSE-150 cells (Fig. 6B and C). Transwell assay results showed that STAT1 overexpression promoted the invasive capability of KYSE-30 cells, while STAT1 knockdown impaired this ability (Fig. 6D). Wound healing assay results showed that STAT1 overexpression enhanced the migratory capability of KYSE-30 cells, while STAT1 knockdown reduced the migratory capability of KYSE-150 cells (Fig. 6E). Western blot analysis also showed that STAT1 overexpression in KYSE-30 cells increased the mesenchymal markers vimentin and N-cadherin and decreased the epithelial marker E-cadherin. Conversely, STAT1 knockdown in KYSE-150 cells produced the opposite effect (Fig. 6F). Our GO/KEGG analyses of characteristic and key genes revealed strong enrichment in ECM‑related pathways. We therefore hypothesized that STAT1 may promote ESCC cell invasion and migration by remodeling the ECM. Accordingly, we examined the mRNA expression of a panel of ECM‑related genes (BMP1, COL1A1, COL3A1, LOX, MMP13, and POSTN) after STAT1 modulation in ESCC cells. As shown in Fig. 6G, overexpression of STAT1 in KYSE‑30 cells significantly upregulated all six ECM‑related genes compared with control or OE‑NC groups. Conversely, knockdown of STAT1 in KYSE‑150 cells markedly reduced their mRNA levels. Taken together, these results indicate that STAT1 promotes invasion, migration, and epithelial-mesenchymal transition (EMT) in ESCC cells. Additionally, STAT1 modulates ECM‑related gene expression, which may contribute to its pro‑invasive and pro‑migratory effects.

Fig. 6.

Effects of STAT1 on invasion, migration, EMT, and ECM remodeling in ESCC cells. A STAT1 expression levels in poorly differentiated (KYSE-150) and well-differentiated (KYSE-30) esophageal cancer cells. B Exogenous expression-regulated STAT1 levels detected by RT-qPCR. C Exogenous expression-regulated STAT1 levels detected by Western Blot. D Effect of differential STAT1 expression on invasion capacity of esophageal cancer cells. Representative images and quantification are shown. E Effect of differential STAT1 expression on migration capacity of esophageal cancer cells. Representative images and wound closure rates are shown. F Effect of differential STAT1 expression on EMT-related protein levels in esophageal cancer cells. G Effect of differential STAT1 expression on ECM‑related gene mRNA levels of esophageal cancer cells. All experiments were independently repeated three times (n = 3). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Gain-of-function validation confirms STAT1-mediated invasion, migration, EMT, and ECM remodeling in KYSE-150 cells

To further validate the positive regulatory role of STAT1, we further overexpressed STAT1 in these cells to determine whether additional STAT1 could further enhance their malignant phenotypes. The efficiency of STAT1 overexpression was confirmed by RT‑qPCR and Western blot (Fig. 7A and B). Transwell assay showed that STAT1 overexpression significantly enhanced the invasive capability of KYSE-150 cells (Fig. 7C). Wound healing assay demonstrated that STAT1 overexpression promoted cell migration, as evidenced by a faster wound closure rate (Fig. 7D). Western blot analysis of EMT markers revealed that STAT1 overexpression in KYSE-150 cells increased the mesenchymal markers vimentin and N-cadherin, while decreasing the epithelial marker E-cadherin (Fig. 7E). Furthermore, RT‑qPCR analysis showed that STAT1 overexpression upregulated the mRNA levels of ECM‑related genes (COL1A1, COL3A1, MMP13, POSTN, LOX, and BMP1) in KYSE-150 cells (Fig. 7F). Collectively, these additional gain-of-function experiments further support that STAT1 promotes invasion, migration, EMT, and ECM‑related gene expression in ESCC cells.

Fig. 7.

Gain-of-function validation confirms STAT1-mediated invasion, migration, EMT, and ECM remodeling in KYSE-150 cells. A STAT1 overexpression efficiency detected by RT-qPCR. B STAT1 overexpression efficiency detected by Western blot. C Effect of STAT1 overexpression on invasion capacity of KYSE-150 cells. Representative images and quantification are shown. D Effect of STAT1 overexpression on migration capacity of KYSE-150 cells. Representative images and wound closure rates are shown. E Effect of STAT1 overexpression on EMT-related protein levels. F Effect of STAT1 overexpression on ECM‑related gene mRNA levels. All experiments were independently repeated three times (n = 3). Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Discussion

In this study, we identified STAT1 as a central regulator and potential biomarker for lymph node metastasis in ESCC through an integrated approach combining bioinformatics, clinical validation, and functional experiments. Our findings not only establish a correlation between STAT1 expression and metastatic propensity but also demonstrate its functional role in driving the invasive and migratory capabilities of ESCC cells.

Our study began with a bioinformatic signature of characteristic genes associated with lymph node metastasis in ESCC. Intriguingly, these genes were predominantly enriched in pathways related to the ECM. The ECM, a major component of the tumor microenvironment (TME), normally acts as a physical barrier against metastasis. However, tumors can remodel the ECM to acquire pro-tumorigenic functions (Anderson and Simon 2020; Goenka et al. 2023). Components like fibronectin, collagen I, and tenascin-C interact with integrins to potently influence cell migration, proliferation, and signaling (Peng et al. 2024; Liu et al. 2023). Our analysis thus reaffirms that ECM dysregulation is a cornerstone of metastatic ESCC.

Among the characteristic genes, we identified STAT1 as a potential key regulator. STAT1, a member of the STAT family, transduces signals from cytokines and growth factors, and plays multifaceted roles in tumor progression and therapy resistance across various cancers (Li et al. 2023). For instance, in gastric cancer, it promotes proliferation, migration, and invasion (Zhang et al. 2023). In prostate cancer, CDKL3-mediated ubiquitination stabilizes STAT1, exhibiting pro-tumor effects (Jiang et al. 2023). Moreover, in hepatocellular carcinoma, HKDC1 facilitates immune evasion via the STAT1/PD-L1 axis in a CD8⁺ T cell-dependent manner (Zhang et al. 2024). Although STAT1 has been less studied in ESCC, its upregulation has been associated with MAGE-C3-mediated PD-L1 induction and enhanced lung metastasis (Wu et al. 2021), which aligns with our findings. Our study further reveals that metastasis-associated genes are notably enriched in ECM-related pathways. While STAT1 is not a canonical ECM component, it is known to regulate genes involved in ECM remodeling, as demonstrated by our finding that STAT1 modulates the expression of multiple ECM‑related genes. This suggests that the pro-metastatic functions of STAT1 observed in our study may be mediated, at least partially, through alteration in the ECM landscape, thereby facilitating tumor cell invasion and migration.

The positive correlation between STAT1 and M1 macrophage polarization suggests a potential association between STAT1 expression and M1-type immune responses within the tumor microenvironment, which may fuel metastasis. Previous studies in breast cancer have demonstrated that STAT1 can be activated by IFN-γ (Totten et al. 2021), a key inflammatory cytokine commonly produced by M1 macrophages. Given this, it is plausible that STAT1 in ESCC cells may similarly be modulated by cytokines derived from M1 macrophages. Based on these findings, we hypothesize the possible existence of a feed-forward loop: M1-derived signals activate STAT1 in ESCC cells, thereby enhancing their malignant phenotype, which might in turn influence the polarization state of tumor-associated macrophages. This potential pro-metastatic alliance within the TME warrants further investigation, and if confirmed, targeting this loop may offer a novel therapeutic strategy to disrupt the metastatic cascade.

Critically, our clinical and functional data solidify the pro-metastatic role of STAT1 in ESCC. Clinical validation confirmed that STAT1 expression was significantly elevated in metastatic ESCC samples compared to the non-metastatic post-neoadjuvant sample,, with the most pronounced upregulation observed in the patient with the highest grade of malignancy (MT2, moderately to poorly differentiated). Moreover, cellular models demonstrated that STAT1 overexpression enhanced the invasive, migratory, and EMT phenotypes of ESCC cells, while its knockdown attenuated these effects. Previous reports have demonstrated that the ERK/STAT1 axis regulates ECM deposition and tumor cell apoptosis (Duan et al. 2023). The connection between STAT1 and ECM remodeling is further supported by our bioinformatic enrichment analysis, which revealed that STAT1-associated key genes were predominantly enriched in ECM-related pathways. Consistently, we found that STAT1 modulates the mRNA expression of multiple ECM-associated genes. These findings further support the notion that STAT1 contributes to the ECM-rich microenvironment and facilitates ESCC metastasis.

This study has limitations. Our strategy of selecting only the top 10 pathways for PPI network construction may have introduced potential bias, as this may exclude less enriched but biologically relevant pathways. This limitation should be considered when interpreting our findings. Additionally, although we verified STAT1’s pro-metastatic role, the specific molecular mechanisms have not been investigated, such as its direct target genes and whether it directly or indirectly regulates EMT transcription factors (e.g., Snail, Slug, Twist). The current sample size is only 4 cases for immunohistochemical validation; therefore, larger clinical cohorts are needed to validate STAT1’s utility as a reliable biomarker for assessing lymph node metastasis in ESCC. Our future research will focus on elucidating these mechanisms and validating our findings in larger cohorts.

Conclusion

In summary, through integrated bioinformatics and experimental validation, we identified STAT1 as a potential biomarker for lymph node metastasis in ESCC. STAT1 expression was positively correlated with the degree of lymph node metastasis in ESCC clinical samples. Functionally, STAT1 enhanced the invasion, migration, and EMT capabilities of ESCC cells, whereas its suppression impaired these malignant phenotypes. These findings position STAT1 as a promising candidate for further investigation into ESCC metastasis mechanisms and potential therapeutic strategies.

Author contributions

Guoli Lv and Haoyu Li: Conceptualization, Methodology, Formal analysis, Data curation, Writing- Original draft preparation. Youming Lei: Visualization, Investigation, Software. Jungao Peng: Investigation, Methodology, Software. Fanghao Liu and Wei Zhao: Writing -Review & Editing, Supervision, Project administration, Funding acquisition.

Funding

Sub-project of the Special Fund for Clinical Medical Research Centers, Yunnan Clinical Medical Research Center for Aging Diseases: Investigating the Differences in the Immune Microenvironment of Primary and Metastatic Sites in Elderly Patients with Esophageal Cancer and the Mechanisms Associated with Esophageal Cancer Metastasis; Grant Number: 202102AA310069.

Data availability

This article can be available from the corresponding author on reasonable request.

Declarations

Conflict of interest

The authors declare that they have no conflicts of interest.