Elevated TMC8 expression correlates with CD8⁺ T cell infiltration and favorable survival in oral squamous cell carcinoma

Introduction

Oral squamous cell carcinoma (OSCC) is a prevalent malignancy characterized by high recurrence rates and limited therapeutic options. The clinical trajectory of this disease is determined by both the intrinsic genetic alterations of neoplastic cells and their complex interactions within the tumor microenvironment (TME) (1). Extensive immunological profiling has established that the spatial distribution and quantitative density of infiltrating cytotoxic T lymphocytes fundamentally dictate the efficacy of anti-tumor responses and patient survival (2). The recruitment of these effector cells relies heavily on specific chemotactic gradients established within the local tissue bed. The precise tumor-intrinsic molecular nodes orchestrating this immune infiltration remain incompletely defined in oral malignancies.

Transmembrane channel-like 8 (TMC8) is a highly conserved membrane protein traditionally recognized for its regulatory role in zinc homeostasis and host immunity against human papillomavirus (HPV) infections (3,4). Furthermore, it has been extensively documented to modulate key intracellular signaling cascades, including TNF-α-induced apoptosis and NF-κB pathways (5,6). Emerging computational analyses of bulk transcriptomic data have implicated this protein as a prognostic biomarker across various solid tumors including head and neck carcinomas (7). Conflicting clinical correlations have been documented where elevated expression predicts poor survival in hepatocellular carcinoma but associates with favorable outcomes in other pan-cancer cohorts (8,9). A fundamental methodological limitation of these previous bulk tissue studies is the inherent amalgamation of expression profiles from malignant cells with those from the surrounding stroma and immune compartments. This ambiguity makes it impossible to determine whether the prognostic benefits associated with elevated expression reflect a tumor-suppressive cellular function or a broader microenvironmental crosstalk effect.

To address this knowledge gap, the present study integrates large-scale clinical cohorts with isolated functional cell models and system biology network analyses. By uncoupling the cell-autonomous behaviors from the complex TME, the potential role of this transmembrane protein in cancer progression is explored. The impact on malignant proliferation and matrix invasion is evaluated utilizing pure OSCC cell lines. These intrinsic cellular phenotypes are subsequently contextualized against the TME immune landscape derived from multidimensional deconvolution algorithms. Here, we further explore the potential topological associations between this transmembrane channel and downstream danger-sensing pathways through bioinformatic network analysis. This integrative approach aims to suggest how the localized malignant properties of the protein might be counterbalanced by a broader immune-modulating network, which is associated with patient survival. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0835/rc).

Methods

Data acquisition and OSCC cohort selection

Transcriptomic profiles and corresponding clinical metadata for head and neck squamous cell carcinoma were retrospectively retrieved from The Cancer Genome Atlas (TCGA) database. The original dataset inherently encompassed various anatomical subregions of the head and neck, which introduced significant spatial and biological heterogeneity. To establish a pure OSCC cohort for downstream analysis, the dataset was rigorously filtered utilizing the provided clinical site annotations. Samples originating exclusively from the oral tongue, buccal mucosa, gums, floor of mouth, retromolar trigone, and hard palate were selectively retained to eliminate confounding non-oral subregions. This stringent selection process yielded an independent analysis cohort comprising 315 tumor samples and 31 adjacent normal tissues. The raw read counts were subsequently normalized to transcripts per million to adjust for sequencing depth variations. Following data normalization and log2 transformation, the prepared expression matrix was utilized for the subsequent construction of weighted gene co-expression networks and functional enrichment evaluations.

Co-expression network construction and functional enrichment

Utilizing the normalized expression matrix, a weighted gene co-expression network analysis (WGCNA) was performed to elucidate TMC8-associated biological pathways. The top 25% most variant genes across the cohort were selected to construct a Pearson correlation matrix. A soft-thresholding power of 6 was analytically determined to ensure a scale-free network topology. The adjacency matrix was subsequently transformed into a topological overlap matrix to evaluate network interconnectedness. Average linkage hierarchical clustering coupled with a dynamic tree-cutting algorithm was then employed to delineate co-expressed gene modules.

The specific module harboring TMC8 was extracted for downstream functional annotation. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses were executed via the clusterProfiler R package to identify predominant signaling cascades. Furthermore, gene set enrichment analysis (GSEA) with 1,000 permutations was conducted to evaluate global transcriptomic shifts, with statistical significance strictly defined by a false discovery rate of less than 0.05.

Immunological microenvironment deconvolution and clustering

To further characterize the immunological landscape implied by the transcriptomic enrichment, the abundance of tumor-infiltrating immune cells was computationally quantified. The TCGA data were subjected to TPM transformation and log₂ normalization (10). Six independent deconvolution algorithms comprising xCell, TIMER, quanTIseq, MCPcounter, EPIC, and CIBERSORT were systematically applied to the expression matrix to estimate the infiltration levels of key immune populations. Subsequently, the ESTIMATE algorithm was utilized to infer the respective fractions of stromal and immune cells within the tumor architecture, yielding corresponding stromal and immune scores for each sample. Relying on these derived microenvironmental metrics, an unsupervised consensus clustering approach was performed, with the median values of the ESTIMATE scores serving as the classification threshold (11), to stratify the patient cohort into two distinct immunological phenotypes, assigned as Cluster 0 and Cluster 1. Specifically, Cluster 0 was defined as the ‘immune-inflamed phenotype’ characterized by dense immune cell infiltration, whereas Cluster 1 represented the ‘immune-desert or excluded phenotype’ (12). The spatial separation and reliability of these identified subtypes were finally validated and visualized in a two-dimensional space utilizing the t-distributed stochastic neighbor embedding dimensionality reduction algorithm.

Clinical specimens and immunohistochemistry

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Peking University Shenzhen Hospital (approval No. 2023-177), and written informed consent was secured from all participating patients prior to surgical resection. Specifically, tissue samples were obtained from patients with OSCC from the Department of Oral and Maxillofacial Surgery, Peking University Shenzhen Hospital, between 2023 and 2024. The inclusion criteria strictly required complete clinical survival data and the absence of any preoperative chemoradiotherapy. Retrospectively collected formalin-fixed, paraffin-embedded (FFPE) tissue sections were subjected to standard deparaffinization in xylene and consecutive rehydration through a graded ethanol series. To unmask antigenic epitopes, the slides were immersed in a pH 6.0 citrate buffer solution and heated via microwave irradiation. Endogenous peroxidase activity was subsequently quenched to prevent non-specific background signals.

The pre-treated sections were incubated overnight at 4 ℃ with an anti-TMC8 primary antibody from Abcam (Cambridge, UK), utilizing a dilution factor of 1:75 under catalog number ab69859. Following strict washing procedures, the specific immunoreactivity was amplified and detected using the UltraSensitive SP Mouse/Rabbit IHC Kit manufactured by MXB Biotechnologies (Fuzhou, China). Visualization of the target protein localization was achieved through the application of a 3,3’-diaminobenzidine chromogen from the same manufacturer, followed by hematoxylin counterstaining for nuclear morphological assessment.

The spatial distribution and relative expression levels of the TMC8 protein were independently evaluated by two experienced pathologists blinded to the associated clinical metadata. The final immunoreactivity was semi-quantitatively determined by comprehensively assessing both the proportion of positively stained tumor cells and the predominant staining intensity across the examined microscopic fields.

Cell culture and siRNA transfection

Human normal oral keratinocytes (HOKs) and the targeted OSCC cell lines, specifically HN6 and HSC3, were entirely sourced from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences. The basal maintenance of these cellular models was achieved using Dulbecco’s Modified Eagle Medium, which was further enriched with 10% fetal bovine serum and a standard penicillin-streptomycin solution. All cultures were strictly propagated in a controlled, humidified incubator programmed at 37 ℃ with a constant 5% CO₂ fraction.

To investigate the downstream phenotypic consequences of TMC8 depletion, transient RNA interference methodologies were employed. HSC3 and HN6 cells were uniformly distributed into appropriate culture plates and allowed to adhere until reaching the requisite confluency. Following this preparation, sequence-specific small interfering RNA (siRNA) duplexes engineered to target TMC8, alongside corresponding non-targeting scrambled control sequences manufactured by GenePharma, were introduced into the target cells. This transient transfection procedure was efficiently mediated by the Lipofectamine 3000 reagent provided by Invitrogen, adhering rigidly to the standardized protocols delineated by the manufacturer. The suppression efficacy at the transcript level was subsequently confirmed via quantitative real-time PCR to ensure robust gene silencing prior to advancing to any phenotypic and mechanistic assays.

RNA extraction and quantitative real-time PCR

To quantify the transcriptomic alterations following targeted silencing, total cellular RNA was isolated from the cultured OSCC cell lines utilizing TRIzol reagent supplied by Takara Bio Incorporated, strictly adhering to the standard isolation protocol. Subsequent to the spectrophotometric assessment of RNA yield and purity, uniform quantities of RNA were subjected to reverse transcription to synthesize complementary DNA. This procedure employed the Evo M-MLV RT Kit Mix for qPCR manufactured by Accurate Biotechnology, executed in strict accordance with the operational protocols.

The quantitative real-time PCR assays were executed utilizing a SYBR Green-based detection system from Accurate Biotechnology on a standard real-time PCR platform. The thermocycling parameters were systematically programmed encompassing an initial denaturation phase at 95 degrees Celsius for 10 minutes, followed by 40 amplification cycles of denaturation for 5 seconds and a combined annealing and extension step at 58 degrees Celsius for 30 seconds. Melting curve analyses were routinely performed to verify the specificity of the amplification products.

The endogenous β-actin gene served as the internal normalization control to adjust for subtle variations in initial RNA input and reverse transcription efficiency. The relative messenger RNA expression level of TMC8 was mathematically determined utilizing the classical comparative threshold cycle method to validate the transient suppression efficacy prior to downstream phenotypic evaluations. The specific oligonucleotide primer sequences utilized in these molecular assays were as follows: TMC8, forward: 5'-GAACTACCCTCCCAACACG-3' and reverse: 5'-TGCTCTTGTCTCTGCCAATG-3'; β-actin, forward: 5'-AAACTGGAACGGTGAAGGTG-3' and reverse: 5'-AGTGGGGTGGCTTTTAGGAT-3'.

Cell proliferation and spatial motility assays

To assess cell viability and proliferative dynamics following targeted genetic intervention, Cell Counting Kit-8 (CCK-8) assays were systematically performed. Transfected OSCC cells were inoculated into 96-well microplates at a predetermined density. At sequential 24-hour intervals (0, 24, 48, and 72 hours), the CCK-8 chromogenic reagent was introduced into the culture system. Following a standard incubation period, the optical density was quantitatively measured at a wavelength of 450 nm utilizing a microplate spectrophotometer.

For the evaluation of two-dimensional horizontal cell motility, conventional wound healing assays were executed. Transfected cells were cultured in multi-well plates until reaching approximately 100% confluency. A standardized linear denuded zone was generated across the intact cell monolayer utilizing a sterile 200-µL pipette tip. After washing out detached cellular debris, the cultures were maintained in a serum-free basal medium to mitigate the confounding effects of cell proliferation. The progressive closure of the artificial wound was photographically documented and mathematically quantified at specific time intervals.

Three-dimensional transmigration and extracellular matrix invasion capabilities were rigorously evaluated employing modified Boyden chamber systems equipped with 8.0-µm pore size polycarbonate membrane inserts. For the migration assessments, cells suspended in a serum-free medium were seeded directly into the upper compartments. To assess the invasive potential, an identical experimental setup was utilized, with the critical exception that the upper inserts were pre-coated with Matrigel basement membrane matrix to simulate the physiological barrier. In both independent experimental configurations, the lower chambers were uniformly supplemented with complete medium containing 10% fetal bovine serum to establish a robust chemotactic gradient. Following a 24- to 48-hour incubation period, the cells successfully traversing the porous membrane were fixed, stained with a crystal violet solution, and microscopically enumerated across multiple randomly selected visual fields.

Protein interaction network and macro-omics correlation analysis

To transition from the cell-autonomous phenotypes to the broader microenvironmental crosstalk, a systems biology approach was employed to explore the predicted functional network and co-expression relationships of the target molecule. The STRING database (version 12.0) was queried to construct the topological network (13). The interaction network was specifically built centering around TMC8 and its tightly co-expressed immune-sensing hubs identified from the overarching transcriptomic analysis. A minimum required interaction score of high confidence (>0.70) was applied to filter out transient or low-probability bindings, and the disconnected nodes were excluded to maintain structural density.

To validate these topological predictions within a real-world macroscopic context, correlation analyses were performed utilizing the bulk clinical cohort. The quantitative expression profiles of key danger-sensing cascade molecules and essential T-cell recruiting chemokines were extracted. Pearson correlation coefficients were calculated to establish the statistical dependency between TMC8 and these downstream immune effectors. The resulting correlation matrix, highlighting the strength and directionality of the microenvironmental crosstalk, was subsequently visualized as a correlogram.

Statistical analysis

All quantitative data derived from the in vitro molecular and cellular assays represented the summarized results of a minimum of three independent biological replicates. Statistical evaluations and graphical visualizations were predominantly conducted utilizing the R statistical programming environment and GraphPad Prism software. Depending on the normality of the underlying data distribution, the statistical variance between two comparative groups was analyzed employing either the parametric Student’s t-test or the non-parametric Wilcoxon rank-sum test. Survival probabilities, including overall survival (OS) and progression-free intervals (PFIs), were estimated via the Kaplan-Meier method, with survival disparities formally evaluated using the log-rank test. The magnitude of associations between continuous variables was determined by computing Pearson or Spearman correlation coefficients, guided by the data distribution characteristics. A two-tailed P value of less than 0.05 was strictly designated as the threshold for statistical significance across all analytical procedures.

Results

Elevated TMC8 expression correlates with favorable prognosis in OSCC

To investigate the clinical relevance of TMC8 in OSCC, we first analyzed transcriptional profiles from TCGA database. In a rigorously filtered dataset comprising 315 OSCC tissues and 31 normal adjacent tissues (Table 1), TMC8 mRNA levels were significantly upregulated in OSCC tissues compared to normal controls (P<0.001) (Figure 1A). Notably, when stratified by clinical T classification, TMC8 expression was found to be significantly higher in early-stage lesions (T1 & T2) than in advanced-stage tumors (T3 & T4) (P=0.03) (Figure 1B), suggesting a stage-dependent expression pattern that might have clinical implications, while no significant differences were observed regarding N classification (P=0.29) or overall tumor stage (P=0.167). Consistent with the transcriptomic data, immunohistochemistry (IHC) on clinical paraffin-embedded sections corroborated that TMC8 protein exhibited intense cytoplasmic staining within the OSCC tumor nests, whereas adjacent normal epithelial tissues displayed markedly weaker immunoreactivity (Figure 1C).

Figure 1 Elevated TMC8 expression correlates with a favorable clinical prognosis despite upregulation in OSCC. (A) Scatter plot comparing TMC8 mRNA expression levels between tumor tissues (n=315) and normal adjacent tissues (n=31) in the TCGA-OSCC cohort. (B) Differential expression of TMC8 stratified by clinical T classification (T1 & T2 vs. T3 & T4). (C) Representative IHC images demonstrating TMC8 protein spatial distribution and intensity in OSCC and adjacent normal epithelial tissues. Scale bars are indicated in the images. (D,E) Kaplan-Meier survival curves for PFI (D) and OS (E) comparing patients with high versus low TMC8 expression, using the optimal cutoff value for stratification. Statistical significance was determined using the independent-samples t-test (A,B) and log-rank test (D,E). Statistical significance is denoted as *P<0.05, **P<0.01, and ***P<0.001. IHC, immunohistochemistry; OS, overall survival; OSCC, oral squamous cell carcinoma; PFI, progression-free interval; T, tumor; TCGA, The Cancer Genome Atlas; TMC8, transmembrane channel-like 8.

Despite its pronounced enrichment in malignant tissues, Kaplan-Meier survival analysis uncovered a paradoxical clinical outcome. Using the optimal cutoff value for stratification, patients in the TMC8-high group exhibited significantly prolonged OS (P=0.008) and a longer PFI (P=0.02) compared to those in the TMC8-low group (Figure 1D,1E). Together, these results suggested that while TMC8 was intrinsically upregulated in OSCC tissues, its elevated expression unexpectedly predicted a favorable clinical prognosis, suggesting a potential anti-tumor bystander effect within the TME.

TMC8 promotes OSCC cell proliferation and invasion in vitro

To ascertain the cell-autonomous functions of TMC8 independent of the host immune system, we evaluated its oncogenic potential in vitro. First, basal TMC8 expression was assessed across a panel of OSCC cell lines (HN6 and HSC3) relative to HOKs (Figure 2A). Given their robust endogenous expression and HPV-negative status, HN6 and HSC3 cells were selected for subsequent loss-of-function assays to preclude confounding viral effects. Transfection with TMC8-specific small interfering RNA (si-TMC8) achieved a significant knockdown efficiency at the mRNA level in both cell lines compared to the negative control (Figure 2B).

Figure 2 TMC8 promotes OSCC cell proliferation and invasion in vitro. (A) Basal relative mRNA expression of TMC8 across normal HOKs and three OSCC cell lines (HN6, HSC3, and CAL27) assessed by qPCR. (B) Validation of siRNA-mediated TMC8 knockdown efficiency at the mRNA level in HN6 and HSC3 cells. (C) Cell proliferation dynamics evaluated via CCK-8 assays over a 72-hour period following TMC8 knockdown. (D) Representative images of wound healing assays depicting the migratory capacity of OSCC cells at the indicated time points. Magnification: 100×. (E,F) Representative images and quantitative analysis of Transwell migration (E) and Matrigel invasion (F) assays following TMC8 knockdown. Crystal violet staining; 200×. Data are presented as the mean ± SD from three independent experiments. Statistical significance was determined using the Student’s t-test (B,E,F) and two-way ANOVA (C). ****, P<0.0001. ANOVA, analysis of variance; CCK-8, Cell Counting Kit-8; HOKs, human normal oral keratinocytes; OD, optical density; OSCC, oral squamous cell carcinoma; qPCR, quantitative polymerase chain reaction; SD, standard deviation; siRNA, small interfering RNA; TMC8, transmembrane channel-like 8.

We next performed CCK-8 assays to monitor cell growth dynamics. The proliferation rates of both HN6 and HSC3 cells were markedly attenuated following TMC8 depletion over a 72-hour period (Figure 2C). Furthermore, we investigated the spatial motility and invasive capabilities of these cells. Two-dimensional wound healing assays revealed that TMC8-silenced cells exhibited a severely impaired ability to migrate and close the denuded area over time (Figure 2D). Consistent with these observations, Transwell migration assays without Matrigel coating demonstrated that TMC8 knockdown substantially reduced the number of OSCC cells capable of migrating across the membrane (P<0.001, Figure 2E). Moreover, in Matrigel-coated Transwell assays, the capacity of OSCC cells to invade and penetrate the extracellular matrix barrier was drastically compromised in the si-TMC8 group (P<0.001, Figure 2F). Collectively, these in vitro findings suggested that TMC8 might be required for the maintenance of proliferative and invasive phenotypes in OSCC cells.

Co-expression networks link TMC8 to immune microenvironment remodeling

Given the stark contrast between the intrinsic pro-tumorigenic role of TMC8 in vitro and its favorable clinical prognosis, we hypothesized that TMC8 might be involved in remodeling the TME. To explore the underlying regulatory networks, we performed a WGCNA using data retrieved from the TCGA database based on the transcriptomic profiles, constructing a co-expression module comprising 1,538 nodes and 1,537 edges.

Functional enrichment analysis of these TMC8-associated genes revealed a profound immunological signature. Gene Ontology (GO) biological process (BP) analysis demonstrated significant enrichment in immune-activating processes, including lymphocyte migration, cellular response to cytokine stimulus, cytokine-mediated signaling pathways, cellular response to interferon-gamma, and neutrophil-mediated immunity. Furthermore, molecular function (MF) and cellular component (CC) analyses indicated enrichment in MHC class II receptor activity, actin-based cell projections, and tertiary granule membranes (Figure 3A).

Figure 3 Co-expression network and functional enrichment analysis of TMC8-associated genes. (A) Bubble plots illustrating the GO enrichment analysis of the TMC8-associated co-expression module, categorized by BP, CC, and MF. (B) Bubble plots displaying the significantly enriched pathways identified through KEGG and Reactome pathway analyses. Dot size represents the number of genes assigned to the term, and the color gradient indicates the −log₁₀ (adjusted P value). BP, biological process; CC, cellular component; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; TMC8, transmembrane channel-like 8.

Consistent with the GO results, KEGG pathway analysis identified significant enrichment in natural killer cell-mediated cytotoxicity and phagosome formation (Figure 3B). Similarly, Reactome analysis highlighted core immune cascades, prominently featuring cytokine signaling in the immune system and signaling by interleukins. To further validate this global transcriptomic shift, we conducted GSEA. The results confirmed that elevated TMC8 expression was robustly and positively correlated with multiple crucial immune gene sets (Table 2). Notably, pathways governing antigen processing and presentation [normalized enrichment score (NES) =0.784], Th1 and Th2 cell differentiation (NES =0.744), Th17 cell differentiation (NES =0.712), and natural killer cell-mediated cytotoxicity (NES =0.697) were all significantly enriched (all adjusted P<0.001). Collectively, these high-throughput transcriptomic data suggested that elevated TMC8 expression was closely associated with an immune-activated microenvironment signature, particularly involving cytokine signaling and lymphocyte recruitment pathways.

TMC8 expression is positively associated with a CD8⁺ T cell-enriched “immune-hot” microenvironment

To validate the immunological shift suggested by our transcriptomic profiling, we comprehensively assessed the correlation between TMC8 expression and the tumor immune microenvironment, focusing specifically on CD8⁺ T cell infiltration. At the single-gene level, TMC8 expression exhibited a strong positive correlation with the core CD8⁺ T cell identity marker CD8A, the cytotoxic effector molecule IFNG, the critical chemokine receptor CXCR3 governing T cell trafficking, and the classical exhaustion checkpoint PDCD1 (Figure 4A). This robust co-expression pattern indicated that TMC8 was closely associated with markers of infiltration, functional activation, and subsequent exhaustion of cytotoxic T lymphocytes within the TME.

Figure 4 TMC8 expression is positively associated with a CD8⁺ T cell-enriched “immune-hot” microenvironment. (A) Scatter plots illustrating strong positive correlations between TMC8 expression and key immune markers: CD8A (core identity), IFNG (cytotoxic effector), CXCR3 (chemokine receptor), and PDCD1 (exhaustion checkpoint) in the TCGA-OSCC cohort. (B) Cross-validation of the correlation between TMC8 expression and CD8⁺ T cell infiltration scores calculated via six independent deconvolution algorithms (xCell, TIMER, quanTIseq, MCPcounter, EPIC, and CIBERSORT). (C) t-SNE dimensionality reduction visually corroborating the separation of the cohort into distinct “immune-hot” (Cluster 0) and “immune-cold” (Cluster 1) immunological phenotypes based on ESTIMATE metrics. The x-axis represents t-SNE1, and the y-axis represents t-SNE2. (D) Box plots comparing the expression levels of selected cytotoxic T cell markers and immune checkpoints between Cluster 0 (hot) and Cluster 1 (cold). (E) Box plot demonstrating the significant upregulation of TMC8 within the highly infiltrated immune-hot cluster (Cluster 0). Statistical significance for group comparisons was determined using the Wilcoxon rank-sum test. Statistical significance is denoted as *P<0.05, **P<0.01, ***P<0.001, and ****P<0.0001. OSCC, oral squamous cell carcinoma; TCGA, The Cancer Genome Atlas; TMC8, transmembrane channel-like 8; t-SNE, t-distributed stochastic neighbor embedding.

To ensure these findings were biologically robust and not algorithm-dependent, we cross-validated the relationship using six independent immunological deconvolution algorithms, specifically xCell, TIMER, quanTIseq, MCPcounter, EPIC, and CIBERSORT. Strikingly, all six algorithms consistently demonstrated a highly significant positive correlation between continuous TMC8 expression levels and CD8⁺ T cell infiltration scores (Figure 4B).

Moving beyond linear correlations, we stratified the OSCC cohort into distinct spatial immunological phenotypes based on ESTIMATE algorithm-derived scores. Unsupervised clustering successfully delineated the cohort into a highly infiltrated “immune-hot” subtype designated as Cluster 0, and an “immune-cold” subtype designated as Cluster 1. This separation was visually corroborated by t-SNE dimensionality reduction (Figure 4C). Consistent with this clustering, the expression levels of key cytotoxic T cell markers were significantly elevated in Cluster 0 (Figure 4D). Most crucially, TMC8 itself was markedly upregulated in this immune-hot cluster compared to the immune-cold cluster (P<0.001, Figure 4E). Collectively, this multi-dimensional analysis demonstrated that elevated TMC8 expression was a robust feature of a CD8⁺ T cell-enriched, inflamed microenvironment in OSCC.

Co-expression analysis of TMC8 with the cGAS-STING and chemokine networks

Having observed in Figure 4 that elevated TMC8 expression was associated with a CD8⁺ T cell-enriched microenvironment, we sought to explore the potential macro-molecular networks correlated with this immunological shift. A fundamental limitation of reductionist in vitro models was their inability to capture the multicellular crosstalk inherent to the tumor ecosystem. Therefore, to evaluate these associations within a real-world macroscopic context, correlation matrices were constructed utilizing the bulk TCGA-OSCC clinical cohort.

In stark contrast to isolated cellular behaviors, the unsupervised correlogram revealed a profound, macroenvironmental co-expression module (Figure 5A). Specifically, TMC8 expression exhibited highly significant and robust positive correlations with core T-cell recruiting chemokines, prominently CXCL10 (r=0.43, P<0.001) and CCL5 (r=0.51, P<0.001). Furthermore, TMC8 showed positive correlations with components of the upstream innate danger-sensing machinery, including TMEM173 (STING) (r=0.29, P<0.001), TBK1 (r=0.16, P<0.01), and IRF3 (r=0.22, P<0.001), alongside inflammatory effectors such as MX1 (r=0.36, P<0.001).

Figure 5 Topological network mapping and macroscopic clinical correlation of TMC8 within the tumor immune ecosystem. (A) Unsupervised correlogram illustrating the Pearson correlation matrix between TMC8, core components of the innate danger-sensing machinery (e.g., TMEM173, TBK1, IRF3), and downstream inflammatory/chemotactic effectors (e.g., CXCL10, IL1B, MX1, CCL5) within the bulk TCGA-OSCC clinical cohort. The color gradient represents the strength and directionality of the Pearson correlation coefficient (r), ranging from negative (blue) to robust positive (red) associations. (B) PPI network constructed utilizing the STRING database (version 12.0), demonstrating the dense topological interconnectivity among the identified TMC8-associated immune nodes. The network highlights a predicted functional interaction module among these immune-associated proteins. Disconnected nodes were strictly excluded to maintain structural density, and the minimum required interaction score was set to high confidence (>0.70). Statistical significance is denoted as *P<0.05, **P<0.01, and ***P<0.001. OSCC, oral squamous cell carcinoma; PPI, protein-protein interaction; r, Pearson correlation coefficient; TCGA, The Cancer Genome Atlas; TMC8, transmembrane channel-like 8.

To determine whether these clinical correlations reflected a coordinated biological module rather than stochastic co-expression, we mapped these molecules onto a protein-protein interaction (PPI) network utilizing the STRING database (Figure 5B). The topological mapping confirmed that these downstream innate immune sensors and chemokine effectors formed a highly dense, structurally interconnected communication hub.

Collectively, this multi-dimensional macro-omics integration delineated a highly specific ecological paradigm: within the complex OSCC microenvironment, TMC8 did not act in isolation. Rather, its expression network was associated with the STING-TBK1 innate sensing axis and chemokine networks based on database predictions. This suggested a hypothesis that the correlation between TMC8 and cytotoxic T cell recruitment might involve a multicellular microenvironmental crosstalk, potentially neutralizing its cell-autonomous malignancy.

Discussion

The present study reveals a distinct phenotypic divergence regarding the biological function of TMC8 in OSCC. In isolated in vitro models devoid of stromal or immunological components, TMC8 was shown to promote malignant phenotypes, including tumor cell proliferation and spatial invasion. Conversely, within the complex tissue architecture of clinical specimens, elevated TMC8 expression correlates with dense CD8⁺ T cell infiltration and significantly favorable survival outcomes. This apparent contradiction suggests that the prognostic value of TMC8 might not be solely dependent on its isolated cellular effects, but rather linked to its association with broader ecological shifts within the TME (2,14,15). Such functional dichotomy underscores the necessity of evaluating putative oncogenes within the context of host immunity, where intrinsic cellular fitness may be clinically neutralized by concurrent microenvironmental vulnerabilities.

siRNA-mediated knockdown of TMC8 effectively suppressed the proliferative capacity and spatial invasiveness of HPV-negative OSCC cells in the current in vitro assays. This cell-autonomous pro-tumorigenic role aligns with previous findings in hepatocellular carcinoma, where TMC8 was identified as a tumor promoter whose overexpression correlates with aggressive disease progression and poor clinical survival (8). However, isolated culture models inherently strip the neoplastic epithelium of its native immunological context. While the intrinsic pro-tumorigenic properties of TMC8 may dominate clinical outcomes in certain immunologically cold malignancies, the paradoxical survival trajectories observed in our highly expressing cohort highlight the complexity of its biological function in vivo. Such prognostic discrepancies underscore that clinical outcomes are frequently influenced by host-tumor interactions rather than isolated cellular phenotypes, prompting an exploration into the comprehensive immunological landscape of the TME (16).

To further evaluate this immunological landscape, an unsupervised clustering approach relying on ESTIMATE-derived microenvironmental metrics was performed. This purely data-driven methodology delineated the clinical cohort into distinct ‘immune-inflamed’ and ‘immune-desert’ phenotypes. TMC8 was highly enriched within the immune-inflamed phenotype. To cross-validate this finding and mitigate potential algorithm-dependent biases, six independent immunological deconvolution platforms were systematically applied (17,18), all of which consistently demonstrated a highly significant positive correlation between TMC8 expression and the estimated infiltration levels of CD8⁺ T lymphocytes. Extensive clinical evidence establishes that the density of such cytotoxic T cells within the tumor contexture serves as a valuable independent predictor of favorable clinical outcomes across solid tumors (2,19), including oral and head and neck squamous cell carcinomas (20,21). Consequently, the enrichment of TMC8 within this highly infiltrated microenvironment provides a potential explanation for its protective prognostic value and helps to understand the divergence between its in vitro cellular functions and clinical presentations.

Notably, our transcriptomic analysis revealed that TMC8 was significantly upregulated in early-stage lesions (T1 & T2) and correlated with a favorable prognosis. This observation aligns with recent findings that early-stage OSCCs possess an inherent, spatially defined immune defense mechanism driven by intense T lymphocyte infiltration (22). However, to fully understand this early immune activation, it is imperative to consider the spatial heterogeneity within the tumor immune microenvironment. As recently highlighted by advanced single-cell and spatial transcriptomic studies, specific tumor subpopulations often co-localize with distinct stromal elements at the leading edge to dictate immune evasion in squamous cell carcinomas (23,24). Indeed, the diverse local tissue architecture, particularly the spatial compartmentalization of stromal subsets and immunosuppressive cells like myeloid-derived suppressor cells (MDSCs), acts as a critical barrier against effective T-cell infiltration (25,26). Conversely, the formation of specific intraepithelial immune niches directly governs the dynamic transition between immune-hot and immune-cold phenotypes (27). This spatial complexity is further compounded by substantial compartment-specific immune checkpoint expression differences between intraepithelial and stromal regions, as recently evidenced in head and neck cancers (28,29). Therefore, we postulate that the elevated TMC8 expression in early stages strategically drives robust multicellular crosstalk within these specific spatial compartments. This localized signaling likely orchestrates an inflamed, immune-hot architecture that effectively restricts tumor progression before widespread stromal immunosuppression can be established.

A fundamental limitation of reductionist in vitro models is their inherent inability to capture the complex intercellular communication defining the native tumor ecosystem. While isolated cellular assays demonstrate the pro-tumorigenic potential of TMC8 in vitro, they cannot explain the clinical correlation with immune infiltration. By integrating macroscopic clinical cohorts with bioinformatic network mapping, the present study proposes a potential multicellular crosstalk network. Crucially, the observed correlations between TMC8 and the cGAS-STING axis do not primarily reflect tumor-cell intrinsic signaling. Bioinformatic predictions and co-expression analyses suggest that TMC8 might be associated with the upstream innate danger-sensing machinery, specifically the cGAS-STING-TBK1 axis, and subsequent inflammatory networks. We postulate that the high-burden proliferative pressure conferred by TMC8 in vitro readily induces endogenous replication stress and genomic instability. Irreversibly damaged dsDNA fragments are subsequently released as immunogenic damage-associated molecular patterns (DAMPs) (30-32). Through microenvironmental signal propagation within the cGAS-STING network, these DAMPs can be taken up by interstitial myeloid cells, thereby driving profound, broader tissue-level inflammatory remodeling and the secretion of CXCL10/CCL5 (33). The activation of this specific innate sensing axis is extensively documented to be paramount for driving tumor-intrinsic immunogenicity and overcoming local immunosuppression (34-37). The statistically significant clinical co-expression of these danger-sensing hubs with critical T-cell recruiting chemokines, prominently CXCL10 and CCL5, suggests that the recruitment of cytotoxic lymphocytes might involve multicellular microenvironmental crosstalk rather than solely relying on direct, cell-autonomous secretion from the neoplastic epithelium (23,38). Ultimately, as the tumor-intrinsic regulation of the microenvironment emerges as a critical determinant of anti-tumor immunity (39,40), the elevated TMC8 expression, while associated with cell-autonomous malignant phenotypes in vitro, might be concurrently linked to a broader immune-activating cascade in vivo. This potential interaction could counterbalance the localized pro-tumorigenic effects, providing a plausible hypothesis for the observed survival benefit in OSCC.

Several limitations of the present study must be acknowledged. Primarily, the reliance on macroscopic multi-omics and isolated cellular monocultures restricts our ability to dynamically visualize the proposed multicellular crosstalk in vivo. The absence of an immunocompetent murine model prevents the direct functional validation of the association between TMC8 and T lymphocyte recruitment. Nevertheless, our multidimensional network analysis utilizing a real-world clinical cohort provides valuable clinical context supporting the potential ecological role of TMC8, which requires further experimental validation. Furthermore, while we have topologically linked TMC8 to the cGAS-STING-TBK1 axis and downstream chemokine networks, the precise intracellular triggers bridging this transmembrane protein to innate immune sensing remain to be elucidated. Although our current model strongly implicates the aforementioned replication stress and genomic instability, determining whether TMC8 is linked to the immune microenvironment primarily through these stress-induced or via alternative inflammatory cascades constitutes the principal focus of our future investigations.

Conclusions

In conclusion, the present study highlights a potential functional divergence regarding the role of TMC8 in OSCC. At the cellular level, TMC8 promotes malignant proliferation and invasion in vitro. However, within the macroscopic tumor ecosystem, the elevated expression of this transmembrane protein is associated with a broad innate immune-sensing network, potentially correlating with a multicellular crosstalk that is linked to the dense infiltration of cytotoxic CD8⁺ T lymphocytes. This potential immune-activating consequence might counterbalance the intrinsic pro-tumorigenic properties of TMC8, ultimately providing a hypothesis for the unexpected clinical trajectory and survival benefit. These findings highlight the necessity of evaluating putative oncogenic drivers within their native immunological context, thereby providing a novel molecular perspective for microenvironmental stratification and therapeutic assessment in oral malignancies.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0835/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0835/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0835/prf

Funding: This work was supported by the Guangdong Provincial Medical Science and Technology Research Fund (No. A2024488), the Basic Research Program of Shenzhen Innovation Council (No. JCYJ20250604183723030), the Guangdong Provincial Basic and Applied Basic Research Fund Provincial Enterprise Joint Fund Project (No. 2022A1515220002), the Shenzhen Clinical Medical Research Center for Oral Diseases (No. 20210617170745001-SCRC202201001), the Sanming Project of Medicine in Shenzhen, Oral and Maxillofacial Surgery Team, Professor Yu Guangyan, Peking University Hospital of Stomatology (No. SZSM202111012), the Shenzhen Fund for Guangdong Provincial High-level Clinical Key Specialties (No. SZGSP008), and the National Natural Science Foundation of China (No. 82273379).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0835/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. The study was approved by the Institutional Review Board of Peking University Shenzhen Hospital (approval No. 2023-177), and written informed consent was secured from all participating patients prior to surgical resection.

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. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Bruni D, Angell HK, Galon J. The immune contexture and Immunoscore in cancer prognosis and therapeutic efficacy. Nat Rev Cancer 2020;20:662-80. [Crossref] [PubMed]
3. Kurima K, Yang Y, Sorber K, et al. Characterization of the transmembrane channel-like (TMC) gene family: functional clues from hearing loss and epidermodysplasia verruciformis. Genomics 2003;82:300-8. [Crossref] [PubMed]
4. Lazarczyk M, Pons C, Mendoza JA, et al. Regulation of cellular zinc balance as a potential mechanism of EVER-mediated protection against pathogenesis by cutaneous oncogenic human papillomaviruses. J Exp Med 2008;205:35-42. [Crossref] [PubMed]
5. Gaud G, Guillemot D, Jacob Y, et al. EVER2 protein binds TRADD to promote TNF-α-induced apoptosis. Cell Death Dis 2013;4:e499. [Crossref] [PubMed]
6. Vuillier F, Gaud G, Guillemot D, et al. Loss of the HPV-infection resistance EVER2 protein impairs NF-κB signaling pathways in keratinocytes. PLoS One 2014;9:e89479. [Crossref] [PubMed]
7. Lin B, Wang S, Yao Y, et al. Comprehensive co-expression analysis reveals TMC8 as a prognostic immune-associated gene in head and neck squamous cancer. Oncol Lett 2021;22:498. [Crossref] [PubMed]
8. Lu P, Ding Q, Ding S, et al. Transmembrane channel-like protein 8 as a potential biomarker for poor prognosis of hepatocellular carcinoma. Mol Clin Oncol 2017;7:244-8. [Crossref] [PubMed]
9. Song J, Tang Y, Luo X, et al. Pan-Cancer Analysis Reveals the Signature of TMC Family of Genes as a Promising Biomarker for Prognosis and Immunotherapeutic Response. Front Immunol 2021;12:715508. [Crossref] [PubMed]
10. Zhang S, Zhang W, Zhang J. Comprehensive analysis of immune cell infiltration and significant genes in head and neck squamous cell carcinoma. Oral Oncol 2022;126:105755. [Crossref] [PubMed]
11. Duan F, Wang W, Zhai W, et al. A novel diagnostic model for predicting immune microenvironment subclass based on costimulatory molecules in lung squamous carcinoma. Front Genet 2022;13:1078790. [Crossref] [PubMed]
12. Xu Q, Chen S, Hu Y, et al. Landscape of Immune Microenvironment Under Immune Cell Infiltration Pattern in Breast Cancer. Front Immunol 2021;12:711433. [Crossref] [PubMed]
13. Szklarczyk D, Gable AL, Nastou KC, et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res 2021;49:D605-12. [Crossref] [PubMed]
14. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144:646-74. [Crossref] [PubMed]
15. Pitt JM, Marabelle A, Eggermont A, et al. Targeting the tumor microenvironment: removing obstruction to anticancer immune responses and immunotherapy. Ann Oncol 2016;27:1482-92. [Crossref] [PubMed]
16. Binnewies M, Roberts EW, Kersten K, et al. Understanding the tumor immune microenvironment (TIME) for effective therapy. Nat Med 2018;24:541-50. [Crossref] [PubMed]
17. Li B, Severson E, Pignon JC, et al. Comprehensive analyses of tumor immunity: implications for cancer immunotherapy. Genome Biol 2016;17:174. [Crossref] [PubMed]
18. Aran D, Hu Z, Butte AJ. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol 2017;18:220. [Crossref] [PubMed]
19. Galon J, Costes A, Sanchez-Cabo F, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 2006;313:1960-4. [Crossref] [PubMed]
20. So YK, Byeon SJ, Ku BM, et al. An increase of CD8(+) T cell infiltration following recurrence is a good prognosticator in HNSCC. Sci Rep 2020;10:20059. [Crossref] [PubMed]
21. Zhao Y, Chen D, Yin J, et al. Comprehensive Analysis of Tumor Immune Microenvironment Characteristics for the Prognostic Prediction and Immunotherapy of Oral Squamous Cell Carcinoma. Front Genet 2022;13:788580. [Crossref] [PubMed]
22. Liao W, Lu J, Xu Y, et al. The Role of Infiltrated T Lymphocyte in Oral Squamous Cell Carcinoma: Insights into Clinicopathological Characteristics and Prognosis. J Inflamm Res 2024;17:2195-204. [Crossref] [PubMed]
23. Zumwalt TJ, Arnold M, Goel A, et al. Active secretion of CXCL10 and CCL5 from colorectal cancer microenvironments associates with GranzymeB+ CD8+ T-cell infiltration. Oncotarget 2015;6:2981-91. [Crossref] [PubMed]
24. Ji AL, Rubin AJ, Thrane K, et al. Multimodal Analysis of Composition and Spatial Architecture in Human Squamous Cell Carcinoma. Cell 2020;182:497-514.e22. [Crossref] [PubMed]
25. Wu SZ, Roden DL, Wang C, et al. Stromal cell diversity associated with immune evasion in human triple-negative breast cancer. EMBO J 2020;39:e104063. [Crossref] [PubMed]
26. Li F, Han Y, Ou F, et al. Spatial heterogeneity of MDSCs mediated by ANXA1-FPRs signaling drives immune suppression in OSCC progression. Nat Commun 2026;17:2535. [Crossref] [PubMed]
27. Phanthunane C, Wijers R, De Herdt MJ, et al. Intratumoral Niches of B Cells and Follicular Helper T Cells, and the Absence of Regulatory T Cells, Associate with Longer Survival in Early-Stage Oral Tongue Cancer Patients. Cancers (Basel) 2022;14:4298. [Crossref] [PubMed]
28. Noda Y, Atsumi N, Nakaya T, et al. High-Sensitivity PD-L1 Staining Using Clone 73-10 Antibody and Spatial Transcriptomics for Precise Expression Analysis in Non-Tumorous, Intraepithelial Neoplasia, and Squamous Cell Carcinoma of Head and Neck. Head Neck Pathol 2025;19:65. [Crossref] [PubMed]
29. Noda Y, Yagi M, Tsuta K. Spatial Transcriptome Analysis of B7-H4 in Head and Neck Squamous Cell Carcinoma: A Novel Therapeutic Target for Anti-Immune Checkpoint Inhibitors. Head Neck Pathol 2025;19:78. [Crossref] [PubMed]
30. Chen M, Linstra R, van Vugt MATM. Genomic instability, inflammatory signaling and response to cancer immunotherapy. Biochim Biophys Acta Rev Cancer 2022;1877:188661. [Crossref] [PubMed]
31. Li Z, Chen H, Wang C, et al. Augmenting cGAS/STING-mediated innate immunity by copper-based nanoplatform for synergistic cancer immunotherapy via inducing irreversible DNA damage. J Control Release 2025;386:114075. [Crossref] [PubMed]
32. Du H, Xiao N, Zhang S, et al. Suppression of TREX1 deficiency-induced cellular senescence and interferonopathies by inhibition of DNA damage response. iScience 2023;26:107090. [Crossref] [PubMed]
33. Mekers VE, Kho VM, Ansems M, et al. cGAS/cGAMP/STING signal propagation in the tumor microenvironment: Key role for myeloid cells in antitumor immunity. Radiother Oncol 2022;174:158-67. [Crossref] [PubMed]
34. Lu C, Guan J, Lu S, et al. DNA Sensing in Mismatch Repair-Deficient Tumor Cells Is Essential for Anti-tumor Immunity. Cancer Cell 2021;39:96-108.e6. [Crossref] [PubMed]
35. Wang H, Hu S, Chen X, et al. cGAS is essential for the antitumor effect of immune checkpoint blockade. Proc Natl Acad Sci U S A 2017;114:1637-42. [Crossref] [PubMed]
36. Ablasser A, Chen ZJ. cGAS in action: Expanding roles in immunity and inflammation. Science 2019;363:eaat8657. [Crossref] [PubMed]
37. Tani T, Mathsyaraja H, Campisi M, et al. TREX1 Inactivation Unleashes Cancer Cell STING-Interferon Signaling and Promotes Antitumor Immunity. Cancer Discov 2024;14:752-65. [Crossref] [PubMed]
38. House IG, Savas P, Lai J, et al. Macrophage-Derived CXCL9 and CXCL10 Are Required for Antitumor Immune Responses Following Immune Checkpoint Blockade. Clin Cancer Res 2020;26:487-504. [Crossref] [PubMed]
39. Kalbasi A, Ribas A. Tumour-intrinsic resistance to immune checkpoint blockade. Nat Rev Immunol 2020;20:25-39. [Crossref] [PubMed]
40. Zhou Z, Huang S, Fan F, et al. The multiple faces of cGAS-STING in antitumor immunity: prospects and challenges. Med Rev (2021) 2024;4:173-91. [Crossref] [PubMed]