Prognostic value and potential upstream regulator of Schwann cells in esophageal squamous cell carcinoma

Introduction

Esophageal carcinoma is a highly prevalent malignancy of the upper digestive tract with substantial mortality worldwide (1). China contributes more than half of the newly diagnosed cases each year (2). According to the National Cancer Center of China, 224,000 new cases and 187,400 deaths were reported in 2022 (3). Histologically, esophageal carcinoma is mainly classified as esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC), with ESCC accounting for >90% of cases in China. Because early-stage disease is often asymptomatic and effective screening strategies are limited, many patients present at an advanced stage. Despite multimodal therapy, including surgery, radiotherapy, and chemotherapy, the 5-year survival of ESCC remains poor (4). Tumor progression, metastatic dissemination, and postoperative recurrence constitute the principal causes of therapeutic failure in malignant diseases. Pathological features such as lymphatic invasion (LI), vascular invasion (VI), and perineural invasion (PNI) are widely recognized as indicators of aggressive tumor behavior and have been consistently associated with adverse clinical outcomes across a broad spectrum of malignancies (5-8).

The tumor microenvironment (TME) comprises malignant cells, stromal cells, infiltrating immune cells, and the microvasculature, which collectively shape tumor growth, survival, and dissemination in a non-cell-autonomous manner (9). Increasing evidence indicates that the peripheral nervous system (PNS) is an active component of the TME and contributes to cancer initiation and progression (10). Schwann cells (SCs), the principal glial cells of the PNS, are broadly classified into myelinating SCs (mSCs), which form myelin sheaths around axons, and non-myelinating Schwann cells (nmSCs), which ensheath multiple axons without forming myelin.

Neuregulins are growth and differentiation factors encoded by six genes [neuregulin 1 (NRG1)–NRG6], among which NRG1 is the most extensively studied (11). NRG1 is essential for the differentiation of neural crest cells into SC precursors and regulates precursor proliferation and migration (12). In cancer, SCs can promote tumor proliferation and metastasis through direct contact and paracrine signaling (13,14). For example, Shurin et al. reported dense nerve fiber distribution at the tumor-normal interface in melanoma, which facilitates tumor growth and dissemination (15). In tumor-associated SCs, glial fibrillary acidic protein (GFAP) is frequently upregulated and has been used as a marker of activated SCs (16). However, the contribution of SCs, particularly nmSCs, to ESCC progression remains insufficiently characterized.

Currently, research on SCs in tumors has mostly focused on a limited number of malignancies, such as neurogenic tumors and breast cancer, and their roles in promoting tumor proliferation, invasion, and metastasis have been partially elucidated (17,18). However, the infiltration characteristics, prognostic value, and regulatory mechanisms of SCs (especially nmSCs) in ESCC remain unclear. With advances in bioinformatics and the availability of large-scale public datasets containing transcriptomic profiles and clinicopathological annotations, the cellular composition and regulatory networks of the ESCC microenvironment can now be interrogated systematically. Therefore, this study for the first time clarified the prognostic significance of nmSC infiltration in ESCC and revealed the upstream teashirt zinc finger homeobox 3 (TSHZ3)-NRG1 regulatory axis, providing new insights into the neural microenvironment of ESCC. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0276/rc).

Methods

Analysis overview

In this study, we analyzed transcriptomic data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, revealing that nmSCs correlate with ESCC patient prognosis. Through comprehensive bioinformatics approaches, we identified the transcription factor TSHZ3 as a potential regulatory factor associated with nmSCs infiltration. Subsequent analysis using the Tumor Immune Estimation Resource (TIMER) database suggested NRG1 as a potential regulatory pathway. Meanwhile, we used tissue slices from ESCC patients to verify the infiltration level of nmSCs in normal esophageal mucosal tissues and tumor tissues. The expression of TSHZ3 in ESCC tumor tissues was also verified. In addition, we established a co-culture system of ESCC cells and SCs, and further validated the regulatory effect of TSHZ3 on NRG1 and its impact on SCs proliferation by knocking down TSHZ3 in ESCC cells, combined with quantitative real-time polymerase chain reaction (qRT-PCR), Western blot, and Cell Counting Kit-8 (CCK-8) assays. Finally, the immunohistochemistry (IHC) results were scored according to the Immunoreactive Score (IRS) criteria and a clinical prediction model based on nmSCs was constructed.

Data source

In this study, all data used were obtained from public databases and The Affiliated Yantai Yuhuangding Hospital of Qingdao University. The time point of extracting or obtaining the final data set for this research and analysis is September 18, 2023. Our research team has indirect access to any information that can identify specific participants. The clinical data of ESCC and transcriptomic data in FPKM format from TCGA were downloaded from the Genomic Data Commons (GDC) of the National Cancer Institute (NCI) (https://portal.gdc.cancer.gov/) (19). Using “esophageal squamous cell carcinoma” and “Homo sapiens” as keywords, the gene expression profiles of GSE69925 and GSE161533 were subsequently identified and downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo/). The GSE161533 gene expression profile includes samples from normal tissues, paratumor tissues, and tumor tissues, with clinical features encompassing demographic indicators, pathological parameters, and molecular subtypes, among others. The cell lines used in this study were the ESCC cell line (KYSE-30) and the SCs line (sNF96-2).

Single-sample gene set enrichment analysis (ssGSEA)

The signature genes of human SCs, including nmSCs and mSCs, were obtained from the Tabula Sapiens portal using OnClass (20). The infiltration levels of these cell types in samples were quantified using the ssGSEA algorithm from the GSVA R package (21), and heatmaps were generated using the pheatmap package in RStudio.

Weighted gene co-expression network analysis (WGCNA)

WGCNA was further employed to analyze the transcriptomic data of TCGA samples, identifying gene sets significantly associated with the infiltration level of target cells. As a transcriptome analysis algorithm, WGCNA can detect gene sets with highly correlated expression patterns and evaluate the associations between gene modules and clinicopathological characteristics of samples (22). The R package “WGCNA” was utilized to perform computational steps in this process. First, the top 5,000 most highly expressed genes were selected for subsequent analysis. Next, the optimal soft-thresholding power for network construction was determined, and dynamic tree-cutting algorithms were applied to establish co-expression modules represented by module eigengenes (MEs). The correlation between each ME and clinicopathological features was then calculated, with the most significantly correlated module designated as the key module to identify co-expressed gene sets strongly linked to target cell infiltration. Finally, correlation analysis using TCGA data revealed key genes exhibiting high associations with target cell infiltration, which were further validated using the GSE69925 dataset from the GEO database.

TIMER database analysis

The TIMER database is a comprehensive tool for analyzing TCGA transcriptomic data (23). In this study, the TIMER database (https://cistrome.shinyapps.io/timer/) was used for pan-cancer analysis to validate the expression correlations between potential upstream regulatory factors and their downstream target pathways across multiple tumor types. The TIMER database was selected for the following reasons: it is based on TCGA multi-cancer transcriptomic data and provides a module for analyzing gene expression correlations. Its standardized analytical pipeline and visualized results are widely recognized, making it suitable for rapidly assessing whether a consistent association exists between two genes at the pan-cancer level. The specific analytical procedure was as follows: enter the “Correlation” module on the TIMER website, input the target gene and the candidate associated gene under the “Gene Expression” option; select all available TCGA tumor types under the “Cancer Type” option; after submission, obtain Pearson correlation scatter plots, correlation coefficients, and P values for the expression levels of the two genes across each cancer type. If a significant positive correlation is observed across multiple cancer types, it suggests a potential shared regulatory network or an upstream/downstream relationship between the two genes, providing clues for subsequent mechanistic studies. Using the above method, we validated the pan-cancer correlation between the target gene and the potential regulatory pathway.

Source of tissue samples and IHC staining

Both ESCC tumor tissues and normal esophageal mucosa samples were obtained from patients who underwent surgical resection at The Affiliated Yantai Yuhuangding Hospital of Qingdao University between 2015 and 2023. All specimens were histopathologically diagnosed as ESCC by senior pathologists. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Institutional Ethics Committee of The Affiliated Yantai Yuhuangding Hospital of Qingdao University (No. 2024-733). All patients included in this study provided written informed consent. The clinicopathological data collected included: age, sex, history of hypertension, history of diabetes, tumor markers, operation time, recurrence time, death time, pathological stage [American Joint Committee on Cancer (AJCC) 9th edition], differentiation grade, operation method, tumor location, margin distance, intravascular tumor thrombus invasion, nerve invasion and lymph node metastasis number.

IHC staining was performed to examine the expression of nmSCs markers in both ESCC tissues and paired normal esophageal mucosa, as well as to assess TSHZ3 expression in ESCC specimens. Formalin-fixed, paraffin-embedded tissue blocks were sectioned at 5 µm thickness and baked at 65 ℃ for 1 hour. Deparaffinization was carried out using xylene and gradient ethanol. Antigen retrieval was performed by boiling the sections in citrate-trisodium citrate for 4 minutes at high heat, followed by 20 minutes of sub-boiling incubation. After three phosphate-buffered saline (PBS) washes, endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Then add nmSCs characteristic protein anti-GFAP rabbit polyclonal antibody (1:50; A19058abclonial)/anti-TSHZ3 rabbit polyclonal antibody (1:200; A115667ATLAS antibody) was incubated at 4 ℃ for 16 hours. After washing with PBS 3 times, the enhancement solution and the second antibody polymer horseradish peroxidase were added in turn. The slides were placed in a 37 ℃ thermostat for 30 minutes and then washed again with PBS 3 times. Finally, 3’-diaminobenzidine (DAB) or 3-amino-9-ethylcarbazole (AEC) and hematoxylin were applied sequentially, then dehydrated with gradient ethanol and xylene, and fixed with neutral gum/glycerol gelatin.

IRS

The expression levels of GFAP and TSHZ3 proteins were evaluated by pathologists from The Affiliated Yantai Yuhuangding Hospital of Qingdao University using the IRS method. The IRS was calculated by multiplying the staining intensity score by the percentage of positive cells, yielding a total score ranging from 0 to 12 points. Staining intensity was graded on a 4-point scale: 0 points indicated no positive staining (negative); 1 point represented faint yellow/faint red staining (weakly positive); 2 points indicated brown yellow/bright red staining (positive); and 3 points denoted dark brown/deep red staining (strongly positive). The percentage of positive cells was categorized into four ranges: 1 point for <10% positive cells; 2 points for 11%-50% positive cells; 3 points for 51–80% positive cells; and 4 points for >80% positive cells (24).

Cell co-culture, transfection, and in vitro functional assays

To investigate the regulatory effect of TSHZ3 on NRG1 and its impact on SCs function, a co-culture system of ESCC cells (KYSE-30) and SCs (sNF96-2) was established. KYSE-30 cells were divided into three groups: control group (control, no treatment), empty vector group (empty, transfected with negative control siRNA), and TSHZ3 knockdown group (siRNA, transfected with siRNA targeting TSHZ3). The sequences of the TSHZ3 siRNA were as follows: sense strand, 5'-GACACGACUGUGUCGGAUATT-3'; antisense strand, 5'-UAUCCGACACAGUCGUGUCTT-3'. Transfection was performed using Lipofectamine 3000 reagent according to the manufacturer’s instructions. Forty-eight hours after transfection, KYSE-30 cells from each group were co-cultured with sNF96-2 cells at a 1:1 ratio in DMEM medium supplemented with 10% fetal bovine serum. After 48 hours of co-culture, KYSE-30 cells from each group were collected for the following assays.

qRT-PCR

Total RNA was extracted using the TRIzol method, and reverse transcription was performed to synthesize cDNA. PCR was carried out using the SYBR Green method under the following cycling conditions: initial denaturation at 95 ℃ for 30 seconds, followed by 40 cycles of denaturation at 95 ℃ for 5 seconds and annealing/extension at 60 ℃ for 30 seconds. The relative expression level of TSHZ3 mRNA was calculated using β-actin as an internal control. The primer sequences for TSHZ3 were as follows: forward, 5'-CACCTACCATCACAACCCTGCT-3'; reverse, 5'-CGACTTCCTTCTTGACCTCCAC-3'. Each experiment was performed in triplicate.

Western blot

KYSE-30 cells from each group were collected, and total protein was extracted using radioimmunoprecipitation assay (RIPA) lysis buffer containing protease inhibitors. Protein concentrations were determined using the bicinchoninic acid (BCA) method. Equal amounts of protein (30 µg) were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then wet-transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% non-fat milk at room temperature for 1 hour, followed by incubation with primary antibodies: rabbit anti-TSHZ3 (1:1,000), rabbit anti-NRG1 (1:500), and rabbit anti-β-actin (1:5,000) at 4 ℃ overnight. The next day, the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5,000) at room temperature for 1 hour, and signals were detected using enhanced chemiluminescence (ECL). Using β-actin as an internal control, the grayscale values of TSHZ3 and NRG1 protein bands were analyzed using ImageJ software, and relative expression levels were calculated. The experiment was independently repeated three times, and representative blot images are presented for each experiment.

CCK-8 proliferation assay

SCs from the co-culture system were seeded into 96-well plates at a density of 5×10³ cells per well, with five replicate wells per group. Cells were cultured for 0, 24, 48, and 72 hours. At each time point, 10 µL of CCK-8 reagent was added to each well, followed by incubation at 37 ℃ for 2 hours. The absorbance at 450 nm (OD₄₅₀) was then measured using a microplate reader. Relative cell proliferation activity at each time point was calculated using the 0-hour time point as a reference. The experiment was independently repeated three times.

Statistical analysis

Gene expression levels between sample groups were compared using the Mann-Whitney test, with visualization performed via the ggpubr package in RStudio. Prognostic analysis was conducted using Kaplan-Meier survival curves and log-rank tests. Correlation analysis was evaluated using Pearson’s correlation coefficient, and plots were generated using the ggstatsplot package in RStudio. For prognostic risk factor assessment in ESCC patients, Cox proportional hazards regression was employed, with statistical analysis performed using IBM SPSS Statistics 25. In univariate analysis, variables with a P value <0.05 were included in the multivariate analysis, with categorical variables converted into dummy variables. Factors with P value less than 0.05 in multivariate analysis were used to construct a clinical prediction model. The nomogram, decision curve analysis (DCA) curve, area under the curve (AUC) and forest map are drawn by rms package, riskRegression package and forestplotter package in R studio software.

Results

Patients with high nmSCs infiltration levels suffered poor prognoses

The heatmap display the infiltration levels of 30 cell types (Figure 1A). Kaplan-Meier survival analysis revealed that patients with high infiltration of nmSCs had a significantly worse prognosis (Figure 1B, P=0.04).

Figure 1 Screening of prognostic genes and Kaplan-Meier survival analysis. (A) The heatmap of 30 tumor microenvironment cell infiltrates in 95 ESCC tissue samples. (B) The relationship between nmSCs infiltration levels and prognosis in TCGA samples (P=0.04). ESCC, esophageal squamous cell carcinoma; MDSC, myeloid-derived suppressor cell; nmSC, non-myelinating Schwann cell; TCGA, The Cancer Genome Atlas.

Identification of co-expressed gene sets in nmSCs

WGCNA can identify co-expressed genes highly related to nmSCs infiltration. With the aid of the dynamic tree-cutting algorithm, the top 5,000 most variably expressed genes were clustered into 10 distinct MEs, each labeled with a unique color. Pearson correlation analysis was then performed to assess the association between MEs and clinicopathological features. As shown in Figure 2A,2B, the green module exhibited the strongest correlation with nmSCs infiltration (Cor =0.76, P=5e−19). This module contained 753 genes, which showed a highly significant association with nmSCs infiltration levels (Cor =0.88, P<1e−200; Figure 2C). These findings suggest that the green module gene list is strongly linked with nmSCs infiltration in ESCC. Further analysis of TCGA samples revealed that the expression of the transcription factor TSHZ3 was significantly correlated with nmSCs infiltration levels (P=3.78e−12, R=0.64; Figure 2D). This association was further validated in the GSE69925 dataset, where TSHZ3 expression remained highly correlated with nmSCs infiltration (P=1.14e−18, R=0.50; Figure 2E).

Figure 2 WGCNA and correlation analysis of ESCC samples. (A,B) Correlations between MEs and nmSCs infiltration levels. (C) GS of the genes contained in the green module versus module membership (nmSCs infiltration levels). (D) Correlation between TSHZ3 expression and nmSCs infiltration levels of samples from TCGA. (E) Correlation between TSHZ3 expression and nmSCs infiltration levels of samples from GSE69925. BF, Bayes factor; CI, credible interval; ESCC, esophageal squamous cell carcinoma; GS, gene significance; HDI, highest density interval; JZS, Jeffreys-Zellner-Siow prior; ME, module eigengene; nmSC, non-myelinating Schwann cell; TCGA, The Cancer Genome Atlas; TSHZ3, teashirt zinc finger homeobox 3; WGCNA, weighted gene co-expression network analysis.

TSHZ3 may promote the infiltration of nmSCs in ESCC tumors through NRG1 signaling pathway

Transcriptomic data analysis from the TCGA database demonstrated a significant correlation between TSHZ3 and NRG1 expression (Figure 3A). Further validation using the GSE161533 dataset revealed that NRG1 expression was significantly higher in tumor tissues compared to both normal and paratumor tissues, as illustrated by the violin plot (Figure 3B). Based on these findings, we propose that NRG1 may serve as a potential downstream effector in the TSHZ3-mediated regulatory pathway, promoting nmSCs infiltration within the TME of ESCC.

Figure 3 Correlation analysis of TSHZ3 and NRG1 and verification of NRG1 expression in different tissues. (A) Correlation between TSHZ3 expression and NRG1 expression of samples from TCGA. (B) Violin plot of the NRG1 expression levels in the three groups from GSE161533. *, P<0.05; ***, P<0.001. NRG1, neuregulin 1; TCGA, The Cancer Genome Atlas; TSHZ3, teashirt zinc finger homeobox 3.

TIMER database confirmed the correlation between TSHZ3 expression and NRG1 expression

Further validation through the TIMER database confirmed a strong positive correlation between TSHZ3 and NRG1 expression levels. Subsequent analysis revealed that this correlation was consistently observed across multiple tumor types, including breast cancer, lung adenocarcinoma, prostate cancer, colon cancer, and endometrial carcinoma (Figure 4).

Figure 4 Correlation between TSHZ3 expression and NRG1 expression in various tumors in TIMER database. BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; COAD, colon adenocarcinoma; ESCA, esophageal carcinoma; HNSC-HPVpos, head and neck squamous cell carcinoma; LUAD, lung adenocarcinoma; NRG1, neuregulin 1; PAAD, pancreatic adenocarcinoma; PCPG, pheochromocytoma and paraganglioma; PRAD, prostate adenocarcinoma; THCA, thyroid carcinoma; THYM, thymic carcinoma; TIMER, Tumor Immune Estimation Resource; TPM, transcript per million; TSHZ3, teashirt zinc finger homeobox 3; UCEC, uterine corpus endometrial carcinoma.

IHC staining

IHC analysis validated GFAP expression in both normal esophageal mucosal tissues and ESCC tumor tissues, with staining scores evaluated according to the IRS scoring criteria (Figure 5A). Compared to normal esophageal mucosal tissues, GFAP was significantly overexpressed in ESCC tumor tissues (Figure 5B). Kaplan-Meier survival analysis revealed that patients with high GFAP scores had a poor prognosis (P=0.02; Figure 5C). Concurrently, we also verified the expression of TSHZ3 in ESCC tumor tissue and scored it according to IRS scoring standard (Figure 6A), and Kaplan-Meier analysis demonstrated worse prognosis in patients with high TSHZ3 expression (P=0.007; Figure 6B). IHC staining of consecutive paraffin sections showed a correlation between TSHZ3 and GFAP expression (Figure 7A), and subsequent correlation analysis of IHC scores further confirmed their significant association (R=0.44, P=7e−04; Figure 7B). These findings provide additional evidence that TSHZ3 may promote nmSCs infiltration in ESCC tumor tissues.

Figure 5 IHC staining verification of GFAP in ESCC and Kaplan-Meier survival analysis. (A) The GFAP expression in normal esophageal mucosal tissues and ESCC tissues was assessed by IHC staining and scored according to the IRS criteria. (B) In the same patient, the expression of GFAP in ESCC tumor tissue was compared with that in normal esophageal mucosa tissue. (C) Kaplan-Meier survival curve of GFAP IHC score. CI, confidence interval; ESCC, esophageal squamous cell carcinoma; GFAP, glial fibrillary acidic protein; IHC, immunohistochemistry; IRS, Immunoreactive Score.

Figure 6 IHC staining verification of TSHZ3 in ESCC and Kaplan-Meier survival analysis. (A) The TSHZ3 expression in ESCC tumor tissues was assessed by IHC staining and scored according to the IRS criteria. (B) Kaplan-Meier survival curve of TSHZ3 IHC score. CI, confidence interval; ESCC, esophageal squamous cell carcinoma; IHC, immunohistochemistry; IRS, Immunoreactive Score; TSHZ3, teashirt zinc finger homeobox 3.

Figure 7 IHC staining comparison and expression correlation analysis of GFAP and TSHZ3 in the same tissue sample. (A) Correlation between GFAP expression and TSHZ3 expression in IHC staining of continuous paraffin sections. (B) Correlation analysis of GFAP IHC score and TSHZ3 IHC score. GFAP, glial fibrillary acidic protein; IHC, immunohistochemistry; TSHZ3, teashirt zinc finger homeobox 3.

Knockdown of TSHZ3 in ESCC cells downregulated NRG1 expression and inhibited the proliferation of SCs in the co-culture system

To establish a causal link between TSHZ3 and NRG1 and to evaluate its impact on SCs function, we established a co-culture system of ESCC cells and SCs, and knocked down TSHZ3 in the ESCC cells. qRT-PCR results showed that, compared with the control and empty vector groups, the mRNA expression level of TSHZ3 in ESCC cells was significantly reduced in the TSHZ3 knockdown group (Figure 8A), indicating efficient knockdown.

Figure 8 Knockdown of TSHZ3 in ESCC downregulates NRG1 expression and inhibits SCs proliferation in a co-culture system. (A) qRT-PCR detection of TSHZ3 expression in ESCC cells from the control, empty vector, and TSHZ3 knockdown groups. (B) Representative blots from three independent experiments, with β-actin used as an internal control. (C) Western blot detection of TSHZ3 expression levels in ESCC cells. (D) Western blot detection of NRG1 expression levels in ESCC cells. (E) CCK-8 assay detection of proliferative activity of SCs in the co-culture system. ns, not significant; ****, P<0.0001. CCK-8, Cell Counting Kit-8; ESCC, esophageal squamous cell carcinoma; NC, negative control; NRG1, neuregulin 1; OD450, absorbance at 450 nm; qRT-PCR, quantitative real-time polymerase chain reaction; SC, Schwann cell; TSHZ3, teashirt zinc finger homeobox 3.

Western blotting was further performed to assess protein expression levels. Representative blots showed that the TSHZ3 protein band intensity in the TSHZ3 knockdown group was markedly weaker than that in the control and empty vector groups (Figure 8B). Quantitative analysis confirmed that TSHZ3 protein expression levels were significantly decreased in the knockdown group (Figure 8C). Concurrently, NRG1 protein expression levels were also significantly reduced in the TSHZ3 knockdown group (Figure 8D), indicating that TSHZ3 positively regulates NRG1 expression.

The CCK-8 assay was used to evaluate the proliferative activity of SCs in the co-culture system. The results showed that, compared with the control and empty vector groups, the proliferative capacity of SCs in the TSHZ3 knockdown group significantly decreased starting from 24 hours, whereas no significant difference was observed between the control and empty vector groups (Figure 8E). Collectively, these results provide direct experimental evidence that knockdown of TSHZ3 in ESCC cells downregulates NRG1 expression and inhibits the proliferative capacity of SCs in the co-culture system, further supporting the role of the TSHZ3-NRG1 axis in regulating SCs.

Clinical prediction model

Cox proportional hazards regression analysis of clinical data from The Affiliated Yantai Yuhuangding Hospital of Qingdao University samples identified nmSCs infiltration, pathological stage, lympho VI, neural invasion, and number of lymph node metastases as prognostic risk factors, with nmSCs infiltration, pathological stage, and neural invasion emerging as independent prognostic risk factors (Figure 9A). Forest plot results similarly demonstrated that nmSCs infiltration, pathological stage, and neural invasion all increased patient mortality risk (Figure 9B). Consequently, we incorporated nmSCs infiltration, pathological stage, and neural invasion as predictive variables to construct a clinical prediction model and developed a corresponding nomogram (Figure 9C), with the model demonstrating a C-index of 0.806 [95% confidence interval (CI), 0.739–0.873]. Results showed our combined model significantly outperformed the prediction model using tumor-node-metastasis (TNM) stage alone, with the AUC values of the combined predictive model being superior to those using TNM staging alone for prognostic prediction throughout the entire follow-up period (Figure 9D).

Figure 9 Cox regression analysis and construction of clinical prediction model. (A) Cox proportional hazards regression analysis of clinical data from The Affiliated Yantai Yuhuangding Hospital of Qingdao University samples. (B) Forest plot analyzed the risk factors of patients’ survival. (C) The nomogram predicted the influence of SCs infiltration, pathological stage, and neural invasion on the patient’s survival. The numbers 0, 1, 2, 3, 6 represent the IRS for GFAP expression in Schwann cells infiltrating ESCC. The numbers 0 and 1 denote perineural invasion status: 0= absent, 1= present. (D) ROC curve showed our combined model significantly outperformed the prediction model using TNM stage alone as a single variable. AUC, area under the curve; CI, confidence interval; ESCC, esophageal squamous cell carcinoma; GFAP, glial fibrillary acidic protein; HR, hazard ratio; IHC, immunohistochemistry; IRS, Immunoreactive Score; ROC, receiver operating characteristic; SC, Schwann cell; TNM, tumor-node-metastasis.

Discussion

ESCC represents a highly aggressive malignancy with relatively poor prognosis, and its complex TME poses challenges for both diagnosis and therapeutic target selection. Currently, several biomarkers have been employed for prognostic evaluation in ESCC, including clinically established markers such as p53, EGFR, and PD-L1, as well as emerging biomarkers like SOX2 and CD44 (25,26). These molecules facilitate patient stratification based on disease aggressiveness and treatment response, making the identification of more reliable prognostic markers crucial for improving patient survival. Among various stromal components, nmSCs, the principal glial cells of the PNS, play pivotal roles in tumor biology (27). This study demonstrated that high nmSCs infiltration correlates with worse prognosis, suggesting nmSCs promote tumor progression. Mechanistically, nmSCs may enhance ESCC aggressiveness through multiple pathways: firstly, by secreting growth factors such as nerve growth factor (NGF) and glial cell line-derived neurotrophic factor (GDNF) that activate oncogenic signaling pathways (e.g., PI3K/AKT and MAPK) in cancer cells (28,29); secondly, through extracellular matrix remodeling via matrix metalloproteinases (MMPs) to facilitate tumor cell migration and invasion (30,31). Although accumulating evidence indicates that nmSCs promote proliferation, invasion, and metastasis across various malignancies (32), their regulatory mechanisms in ESCC remain unclear. In recent years, several reviews have systematically elucidated the roles of various components of the PNS in tumor initiation and progression, thereby advancing the emerging field of cancer neuroscience (33,34). Previous studies have extensively discussed the functions of sympathetic, parasympathetic, and sensory nerves within the TME, and have indicated that these neural components can directly regulate tumor cell proliferation and migration through the release of neurotransmitters (10,35,36). Furthermore, neurotransmitters (e.g., norepinephrine and acetylcholine) and neurotrophic factors have been shown to remodel the TME, thereby affecting tumor invasion and metastatic capacity. Meanwhile, the sensory nervous system can drive phenotypic changes in tumor immune responses by releasing peripheral sensory signals such as neuropeptides, thus promoting cancer progression (37,38). As an important component of the PNS, the interaction between SCs and tumors has received extensive attention in recent years. Multiple studies have summarized the pro-tumorigenic mechanisms of SCs in various malignancies and their potential as therapeutic targets (27,32). Notably, the nervous system may also exert inhibitory effects on tumor initiation and progression, thereby playing a dual role in cancer progression (39). The field of cancer neuroscience emphasizes the complex interactions between the nervous system and tumors, including tumor-induced neurogenesis, neural remodeling, and the regulation of the tumor immune microenvironment by neural signals. These findings provide new perspectives for the development of therapeutic strategies targeting nerve-tumor interactions (40,41). Therefore, in-depth dissection of the specific mechanisms of PNS components, including nmSCs, in ESCC, together with integration of the theoretical framework of cancer neuroscience, will help to identify novel prognostic markers and potential therapeutic targets.

The effects of the nervous system on tumors are highly context-dependent, and their outcomes are influenced by multiple factors. At the level of experimental models, the same type of nerve can produce diametrically opposite effects in different tumor types: sensory neurons inhibit tumor progression in melanoma but may promote it in pancreatic cancer; the vagus nerve exerts tumor-suppressive effects in pancreatic ductal adenocarcinoma and also inhibits tumor growth in colorectal cancer, whereas it may yield different results in other models (42-45). The role of sensory nerves in breast cancer is even more complex: desensitization of sensory nerves using resiniferatoxin (RTX) accelerates early tumor growth by increasing vascular leakage, whereas activation of sensory nerves with another TRPV1 agonist, Olvanil, suppresses lung and liver metastases by enhancing CD8⁺ T cell immune responses without affecting primary tumor growth (46,47). At the tissue microenvironment level, immune cell composition, vascular status, and local neuropeptide levels can all significantly influence the direction of neural regulation (44). Furthermore, the choice of intervention method is equally critical: studies using DREADD technology have shown that inhibiting sensory neuron firing promotes melanoma growth and angiogenesis, whereas overactivation produces the opposite effect (44). Notably, denervation generally leads to accelerated tumor progression, whereas neural overactivation often exerts anti-tumor effects, suggesting that a “baseline protective function” of the nervous system may be universal, and denervation simply removes this protection. Previous studies have shown that the aging phenotype of SCs (characterized by decreased proliferative capacity and delayed initiation of repair programs) may attenuate the efficacy of nerve-tumor interactions (48). However, the impact of cancer cell inoculation dose on neural regulation has not been directly investigated in the existing literature; a high inoculum of cancer cells may rapidly occupy the TME and thereby mask the modulatory effects of the nervous system. In summary, the net effect of the nervous system on tumors is not determined by a single factor, but rather arises from the synergistic interplay of multiple variables, including tumor model type, microenvironmental features, host age, and intervention approach.

Existing studies have revealed the specific mechanisms by which SCs detach from nerves and actively engage in the nervous system-cancer crosstalk during tumor progression. Deborde et al. found that nmSCs activated by cancer cells collectively organize into dynamic “tumor-activated Schwann cell tracks (TASTs)” that serve as physical pathways, exerting mechanical forces on cancer cells to enhance their migration and invasive capacity. This process depends on the activation of c-Jun, resembling the reprogramming state observed during nerve injury repair (20). Further studies have shown that PNI involves biophysical coupling between tumor cells and SCs. SCs can drive collective migration of cancer cells both within and around tumor-associated nerves (49). As the major glial cells in the PNS, SCs possess high plasticity, whereby they are remodeled by malignant cells in the TME into pro-tumorigenic factors and participate in cancer neurogenesis and multiple molecular interactions (50). Moreover, endogenous SCs can also influence tumor progression by modulating tumor angiogenesis and immune surveillance: specific ablation of SCs slows tumor growth and angiogenesis, while simultaneously enhancing anti-tumor lymphocyte infiltration and reducing immunosuppressive cells. In melanoma patients, high expression of SC-related genes is associated with better survival prognosis, suggesting that the role of SCs may be bidirectional (51). Collectively, the above findings highlight the important role of SCs (especially nmSCs) in detaching from nerves and actively participating in tumor-nerve crosstalk, providing new theoretical insights into the mechanisms underlying nmSC infiltration in ESCC and potential therapeutic targets. On this basis, multivariate Cox regression analysis in this study further confirmed that nmSC infiltration is an independent prognostic risk factor for postoperative ESCC patients [hazard ratio (HR) =1.985, P<0.001], which not only improves the accuracy of prognostic prediction but also provides guidance for personalized therapy targeting nerve-tumor interactions.

TSHZ3 belongs to the teashirt zinc finger protein family (TSHZ1–3) and plays critical roles in embryonic development and cell differentiation. These proteins contain conserved zinc finger domains that mediate transcriptional regulation and protein-protein interactions (52). In glioblastoma, TSHZ3 functions as a tumor suppressor gene. Loss of TSHZ3 expression relieves its inhibitory effect on MMP2, thereby promoting glioma cell invasion. Exogenous overexpression of TSHZ3 significantly suppresses invasiveness and counteracts the pro-invasive effects of miR-338-5p (53). Similarly, in breast cancer, prostate cancer, lung adenocarcinoma, and colorectal cancer, low TSHZ3 expression is consistently associated with poor clinical prognosis (54-56), but their functions in ESCC remain unexplored. Through comprehensive bioinformatics analyses, our study demonstrates TSHZ3 overexpression in ESCC, suggesting its tumor-promoting role. Then identifies TSHZ3 as a biomarker positively correlated with nmSCs infiltration levels in ESCC. To our knowledge, this represents the first report establishing TSHZ3 as a cancer biomarker associated with nmSCs infiltration.

To further validate the regulatory relationship between TSHZ3 and NRG1, we established a co-culture system of ESCC cells and SCs. Functional validation was performed by knocking down TSHZ3 in ESCC cells, combined with qRT-PCR, Western blot, and CCK-8 assays. The results showed that TSHZ3 knockdown significantly reduced NRG1 protein expression levels and inhibited the proliferative capacity of SCs in the co-culture system. These data provide, for the first time, direct causal evidence for the TSHZ3-NRG1 axis, strongly supporting the hypothesis that TSHZ3 upregulates NRG1 to subsequently affect SCs function.

The present study has preliminarily demonstrated that TSHZ3 may positively regulate NRG1 expression and thereby promote nmSC infiltration; however, how TSHZ3, as a transcriptional repressor, positively regulates NRG1 remains to be further elucidated. Previous studies have shown that TSHZ3 contributes to the progression of Alzheimer’s disease by inhibiting the expression of target genes such as CASP4 (57). Based on this molecular function and the positive correlation between TSHZ3 and NRG1 expression observed in the present study, we hypothesize that TSHZ3 may upregulate NRG1 indirectly by inhibiting negative regulators of NRG1 (e.g., specific microRNAs or transcriptional repressors) rather than by directly activating its transcription. Furthermore, nmSCs are highly sensitive to NRG1 signaling: they express high levels of ErbB2/ErbB3 receptors on their surface, and NRG1 specifically binds to the ErbB2/ErbB3 heterodimer to activate downstream PI3K/AKT and MAPK signaling pathways, thereby driving nmSC proliferation and migration (58). Therefore, TSHZ3-upregulated NRG1 may specifically recruit nmSCs in a paracrine manner. Based on the existing literature and our findings, we propose for the first time a novel TSHZ3-NRG1-nmSC regulatory axis model. Future studies should generate ESCC cell lines with stable TSHZ3 knockdown or overexpression and perform co-culture migration assays with SCs to directly observe whether TSHZ3-regulated NRG1 is sufficient to drive directional SCs migration. Additionally, orthotopic xenograft models can be used to validate the effects of the TSHZ3-NRG1 axis on nmSC infiltration density, tumor growth, and PNI formation in vivo.

Our study revealed a positive correlation between TSHZ3 expression and both NRG1 expression levels and nmSCs infiltration density, with this association being further validated across multiple other malignant tumor types. The relationship between TSHZ3 and GFAP was confirmed through IHC staining. Furthermore, in vitro functional experiments directly confirmed that knockdown of TSHZ3 in ESCC cells significantly reduced NRG1 protein expression levels and inhibited the proliferative capacity of SCs in the co-culture system. These collective findings suggest that TSHZ3 likely promotes nmSCs infiltration via the NRG1 signaling pathway (Figure 10).

Figure 10 The potential interactions across TSHZ3, NRG1, nmSCs and tumor cells in the TME of ESCC. ErbB, epidermal growth factor receptor family; ESCC, esophageal squamous cell carcinoma; GDNF, glial cell line-derived neurotrophic factor; MMPs, matrix metalloproteinases; NCC, neural crest cell; NGF, nerve growth factor; nmSC, non-myelinating Schwann cell; NRG1, neuregulin 1; SCP, Schwann cell precursor; TME, tumor microenvironment; TSHZ3, teashirt zinc finger homeobox 3.

However, there are certain limitations in this study that should be acknowledged Our analysis was based on retrospective data from public databases, which may introduce selection bias. In the IHC staining experiments, the sample size was not fully sufficient, and more comprehensive biological validation experiments are required to further substantiate our conclusions.

Conclusions

By integrating data from multiple databases and employing diverse bioinformatics analytical approaches, Clinical immunohistochemical validation and in vitro co-culture functional experiments, this study identified nmSCs may serve as a significant prognostic factor in ESCC patients, with the reliability of this prognostic prediction being validated through a clinical prediction model based on IHC scoring. Further screening revealed a strong correlation between TSHZ3 expression levels and nmSCs infiltration density, Furthermore, functional experiments directly confirmed that knockdown of TSHZ3 in ESCC cells significantly downregulated NRG1 expression and inhibited the proliferation of SCs in the co-culture system. These findings support that TSHZ3 likely promotes nmSCs enrichment in tumor tissues through NRG1, thereby establishing a TSHZ3-NRG1-nmSCs regulatory network. Consequently, nmSCs serve as a predictor of poor prognosis in patients with ESCC, and we propose that TSHZ3 may facilitate nmSCs infiltration via the NRG1 signaling pathway.

Acknowledgments

We would like to express our sincere gratitude to all the participants in this study and extend our appreciation to the TCGA database and GEO database for providing their platforms, as well as to the contributors who uploaded their valuable datasets.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0276/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0276/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0276/prf

Funding: This work was supported by Shandong Provincial Natural Science Foundation Youth Fund (No. ZR2023QH545).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0276/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Institutional Ethics Committee of The Affiliated Yantai Yuhuangding Hospital of Qingdao University (No. 2024-733). All patients included in this study provided written informed consent.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Lander S, Lander E, Gibson MK. Esophageal Cancer: Overview, Risk Factors, and Reasons for the Rise. Curr Gastroenterol Rep 2023;25:275-9. [Crossref] [PubMed]
2. Chang J, Tan W, Ling Z, et al. Genomic analysis of oesophageal squamous-cell carcinoma identifies alcohol drinking-related mutation signature and genomic alterations. Nat Commun 2017;8:15290. [Crossref] [PubMed]
3. Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024;4:47-53. [Crossref] [PubMed]
4. Xu W, Liu Z, Bao Q, et al. Viruses, Other Pathogenic Microorganisms and Esophageal Cancer. Gastrointest Tumors 2015;2:2-13. [Crossref] [PubMed]
5. Royston D, Jackson DG. Mechanisms of lymphatic metastasis in human colorectal adenocarcinoma. J Pathol 2009;217:608-19. [Crossref] [PubMed]
6. Liebig C, Ayala G, Wilks JA, et al. Perineural invasion in cancer: a review of the literature. Cancer 2009;115:3379-91. [Crossref] [PubMed]
7. Cornwell LB, McMasters KM, Chagpar AB. The impact of lymphovascular invasion on lymph node status in patients with breast cancer. Am Surg 2011;77:874-7.
8. Chatterjee D, Katz MH, Rashid A, et al. Perineural and intraneural invasion in posttherapy pancreaticoduodenectomy specimens predicts poor prognosis in patients with pancreatic ductal adenocarcinoma. Am J Surg Pathol 2012;36:409-17. [Crossref] [PubMed]
9. Schmitt M, Greten FR. The inflammatory pathogenesis of colorectal cancer. Nat Rev Immunol 2021;21:653-67. [Crossref] [PubMed]
10. Silverman DA, Martinez VK, Dougherty PM, et al. Cancer-Associated Neurogenesis and Nerve-Cancer Cross-talk. Cancer Res 2021;81:1431-40. [Crossref] [PubMed]
11. Mei L, Nave KA. Neuregulin-ERBB signaling in the nervous system and neuropsychiatric diseases. Neuron 2014;83:27-49. [Crossref] [PubMed]
12. Raphael AR, Talbot WS. New insights into signaling during myelination in zebrafish. Curr Top Dev Biol 2011;97:1-19. [Crossref] [PubMed]
13. Zhou Y, Shurin GV, Zhong H, et al. Schwann Cells Augment Cell Spreading and Metastasis of Lung Cancer. Cancer Res 2018;78:5927-39. [Crossref] [PubMed]
14. Su D, Guo X, Huang L, et al. Tumor-neuroglia interaction promotes pancreatic cancer metastasis. Theranostics 2020;10:5029-47. [Crossref] [PubMed]
15. Shurin GV, Kruglov O, Ding F, et al. Melanoma-Induced Reprogramming of Schwann Cell Signaling Aids Tumor Growth. Cancer Res 2019;79:2736-47. [Crossref] [PubMed]
16. Zhang J, Zhao F, Wu G, et al. Functional and histological improvement of the injured spinal cord following transplantation of Schwann cells transfected with NRG1 gene. Anat Rec (Hoboken) 2010;293:1933-46. [Crossref] [PubMed]
17. Vasudevan HN, Lucas CG, Villanueva-Meyer JE, et al. Genetic Events and Signaling Mechanisms Underlying Schwann Cell Fate in Development and Cancer. Neurosurgery 2021;88:234-45. [Crossref] [PubMed]
18. Huang L, Qi G, Chen G, et al. Tumor-associated Schwann cells as new therapeutic target in non-neurological cancers. Cancer Lett 2025;624:217748. [Crossref] [PubMed]
19. Zhang Z, Hernandez K, Savage J, et al. Uniform genomic data analysis in the NCI Genomic Data Commons. Nat Commun 2021;12:1226. [Crossref] [PubMed]
20. Deborde S, Gusain L, Powers A, et al. Reprogrammed Schwann Cells Organize into Dynamic Tracks that Promote Pancreatic Cancer Invasion. Cancer Discov 2022;12:2454-73. [Crossref] [PubMed]
21. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 2013;14:7. [Crossref] [PubMed]
22. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 2008;9:559. [Crossref] [PubMed]
23. Li T, Fu J, Zeng Z, et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res 2020;48:W509-14. [Crossref] [PubMed]
24. Beyer S, Wehrmann M, Meister S, et al. Galectin-8 and -9 as prognostic factors for cervical cancer. Arch Gynecol Obstet 2022;306:1211-20. [Crossref] [PubMed]
25. Yang K, Gao F, Zhou C, et al. Immunological biomarkers and predictive model for recurrence of esophageal squamous cell carcinoma after combined immunotherapy and neoadjuvant chemotherapy. Am J Cancer Res 2024;14:4896-908. [Crossref] [PubMed]
26. Tian S, Ma R, Liu Y, et al. Clinicopathological significance of cancer stem cell marker CD44/SOX2 in esophageal squamous cell carcinoma (ESCC) patients and construction of a nomogram to predict overall survival. Transl Cancer Res 2024;13:2971-84. [Crossref] [PubMed]
27. Zhang Y, Yuan Y, Wang Y, et al. The Bridging Role of Schwann Cells in the Interaction between Tumors and the Nervous System: A Potential Target for Cancer Therapy. Mol Cancer Res 2025;23:494-502. [Crossref] [PubMed]
28. Kim HS, Kim JY, Song CL, et al. Directly induced human Schwann cell precursors as a valuable source of Schwann cells. Stem Cell Res Ther 2020;11:257. [Crossref] [PubMed]
29. Tian Z, Ou G, Su M, et al. TIMP1 derived from pancreatic cancer cells stimulates Schwann cells and promotes the occurrence of perineural invasion. Cancer Lett 2022;546:215863. [Crossref] [PubMed]
30. Ferdoushi A, Li X, Griffin N, et al. Schwann Cell Stimulation of Pancreatic Cancer Cells: A Proteomic Analysis. Front Oncol 2020;10:1601. [Crossref] [PubMed]
31. Liu H, Kato Y, Erzinger SA, et al. The role of MMP-1 in breast cancer growth and metastasis to the brain in a xenograft model. BMC Cancer 2012;12:583. [Crossref] [PubMed]
32. Deborde S, Wong RJ. The Role of Schwann Cells in Cancer. Adv Biol (Weinh) 2022;6:e2200089. [Crossref] [PubMed]
33. Vermeer PD, Restaino AC, Barr JL, et al. Nerves at Play: The Peripheral Nervous System in Extracranial Malignancies. Cancer Discov 2025;15:52-68. [Crossref] [PubMed]
34. Jones G, Anderson JL, Nguyen PTT, et al. Novel approaches to clinical trial design in cancer neuroscience. Neuron 2025;113:2791-813. [Crossref] [PubMed]
35. Amit M, Eichwald T, Roger A, et al. Neuro-immune cross-talk in cancer. Nat Rev Cancer 2025;25:573-89. [Crossref] [PubMed]
36. Wang X, Fan Y, Wang Q, et al. Tumor-infiltrating nerves: unraveling the role of cancer neuroscience in tumorigenesis, disease progression, and emerging therapies. Discov Oncol 2025;16:1209. [Crossref] [PubMed]
37. Sur D, Zeng Y, Kobayashi H, et al. Entangled cellular and molecular relationships at the sensory neuron-cancer interface. Neuron 2025;113:2760-90. [Crossref] [PubMed]
38. Martel Matos AA, Scheff NN. Sensory neurotransmission and pain in solid tumor progression. Trends Cancer 2025;11:309-20. [Crossref] [PubMed]
39. Zhang Y, Liao Q, Wen X, et al. Hijacking of the nervous system in cancer: mechanism and therapeutic targets. Mol Cancer 2025;24:44. [Crossref] [PubMed]
40. Thiel V, Sur D, Picoli CC, et al. Next-gen tools in cancer neuroscience. Cell Rep 2025;44:116258. [Crossref] [PubMed]
41. Khanmammadova N, Islam S, Sharma P, et al. Neuro-immune interactions and immuno-oncology. Trends Cancer 2023;9:636-49. [Crossref] [PubMed]
42. Renz BW, Tanaka T, Sunagawa M, et al. Cholinergic Signaling via Muscarinic Receptors Directly and Indirectly Suppresses Pancreatic Tumorigenesis and Cancer Stemness. Cancer Discov 2018;8:1458-73. [Crossref] [PubMed]
43. Prazeres PHDM, Leonel C, Silva WN, et al. Ablation of sensory nerves favours melanoma progression. J Cell Mol Med 2020;24:9574-89. [Crossref] [PubMed]
44. Costa PAC, Silva WN, Prazeres PHDM, et al. Chemogenetic modulation of sensory neurons reveals their regulating role in melanoma progression. Acta Neuropathol Commun 2021;9:183. [Crossref] [PubMed]
45. Dubeykovskaya Z, Si Y, Chen X, et al. Neural innervation stimulates splenic TFF2 to arrest myeloid cell expansion and cancer. Nat Commun 2016;7:10517. [Crossref] [PubMed]
46. Bencze N, Schvarcz C, Kriszta G, et al. Desensitization of Capsaicin-Sensitive Afferents Accelerates Early Tumor Growth via Increased Vascular Leakage in a Murine Model of Triple Negative Breast Cancer. Front Oncol 2021;11:685297. [Crossref] [PubMed]
47. Erin N, Akman M, Aliyev E, et al. Olvanil activates sensory nerve fibers, increases T cell response and decreases metastasis of breast carcinoma. Life Sci 2022;291:120305. [Crossref] [PubMed]
48. Painter MW, Brosius Lutz A, Cheng YC, et al. Diminished Schwann cell repair responses underlie age-associated impaired axonal regeneration. Neuron 2014;83:331-43. [Crossref] [PubMed]
49. Amit M, Maitra A. The Boring Schwann Cells: Tumor Me-TAST-asis along Nerves. Cancer Discov 2022;12:2240-3. [Crossref] [PubMed]
50. Yurteri Ü, Çifcibaşı K, Friess H, et al. Schwann Cells in Peripheral Cancers: Bystanders or Promoters? Adv Biol (Weinh) 2022;6:e2200033. [Crossref] [PubMed]
51. Rocha BGS, Picoli CC, Gonçalves BOP, et al. Tissue-resident glial cells associate with tumoral vasculature and promote cancer progression. Angiogenesis 2023;26:129-66. [Crossref] [PubMed]
52. Kesdiren E, Martens H, Brand F, et al. Heterozygous variants in the teashirt zinc finger homeobox 3 (TSHZ3) gene in human congenital anomalies of the kidney and urinary tract. Eur J Hum Genet 2025;33:44-55. [Crossref] [PubMed]
53. Li Y, Huang Y, Qi Z, et al. MiR-338-5p Promotes Glioma Cell Invasion by Regulating TSHZ3 and MMP2. Cell Mol Neurobiol 2018;38:669-77. [Crossref] [PubMed]
54. Yamamoto M, Cid E, Bru S, et al. Rare and frequent promoter methylation, respectively, of TSHZ2 and 3 genes that are both downregulated in expression in breast and prostate cancers. PLoS One 2011;6:e17149. [Crossref] [PubMed]
55. Zhang X, Liu Y, Peng BZ, et al. The transcription factor TSHZ3 promotes tumor immunosuppression and inhibits metastasis in lung adenocarcinoma. Front Immunol 2025;16:1519815. [Crossref] [PubMed]
56. Zhou Y, Wang S, Yin X, et al. TSHZ3 functions as a tumor suppressor by DNA methylation in colorectal cancer. Clin Res Hepatol Gastroenterol 2021;45:101725. [Crossref] [PubMed]
57. Kajiwara Y, Akram A, Katsel P, et al. FE65 binds Teashirt, inhibiting expression of the primate-specific caspase-4. PLoS One 2009;4:e5071. [Crossref] [PubMed]
58. Piovesana R, Pisano A, Loreti S, et al. Notch Signal Mediates the Cross-Interaction between M2 Muscarinic Acetylcholine Receptor and Neuregulin/ErbB Pathway: Effects on Schwann Cell Proliferation. Biomolecules 2022;12:239. [Crossref] [PubMed]