The GEO dataset (GSE143224) (https://www.ncbi.nlm.nih.gov/) was used for differential gene expression analysis, and after batch correction, we conducted Principal Component Analysis (PCA) using the methodology outlined in a prior study [21] to identify patterns and relationships among the DEGs in LSCC and normal tissues. For volcano and heatmap analysis, genes meeting the criteria of $|$logFC$|$ $>$1 and p 0.05 in both groups were plotted. Volcano plots were generated to visualize the significance and fold-change differences of DEGs between LSCC and normal tissues, with color-coded representations of significantly upregulated and downregulated genes. Heatmap analysis illustrated gene expression patterns, offering an overview of their clustering and levels across the samples.

After screening the DEGs between normal and tumor specimens from volcano and heatmap analysis, the data was further analyzed using the online platform Metascape (Metascape Foundation, San Diego, USA) and R language (R Foundation for Statistical Computing, Vienna, Austria), and for protein-protein interaction (PPI) analysis, the proteins were analyzed using STRING.

For this experiment, we obtained 97 pairs of HNSC (head and neck squamous cell carcinoma) tissues from patients based on pre-determined inclusion and exclusion criteria to ensure the relevance and reliability of the experimental samples. Patients diagnosed with HNSC were included and excluded those who had undergone chemotherapy, radiotherapy or any other treatment, as well as individuals with a history of recurrence. All tissue samples were embedded, sliced, incubated and assessed via microscopy [22]. In the IHC experiments, specific CRNN antibodies were used to validate its protein expression in LSCC tissues. The tissue sections were subjected to antibody incubation, and staining intensities were evaluated. Subsequently, we quantitatively analyzed the images to determine the relative expression levels of CRNN in both tumor and normal tissues to obtain insights into its protein-level regulation. The detailed information regarding the tissue specimens are shown in Table 1.

Total RNA was extracted as previously described, and then reverse transcription was performed for analyzing mRNA expression. The primers of CRNN (forward primer: 5${}^{\prime }$-AAACTGTAAGCCACGGAGGG-3${}^{\prime }$; reverse primer: 5${}^{\prime }$-GTCCTCCCGGTACTGTCTCT-3${}^{\prime }$) and GAPDH (forward primer: 5${}^{\prime }$-GTCTCCTCTGACTTCAACAGCG-3${}^{\prime }$; reverse primer: 5${}^{\prime }$-ACCACCCTGTTGCTGTAGCCAA-3${}^{\prime }$) were used in this experiment. Western blotting involved protein isolation using an EBC buffer ((50 mM Tris pH 7.5, 120 mM NaCl, 0.5% NP-40)), followed by electrophoresis, membrane transfer, blocking, incubation, and visualization using an ECL (Enhanced chemiluminescence) reagent.

Kaplan-Meier curve analysis was conducted using an online platform (https://kmplot.com/analysis/). The correlation between CRNN and T-cell enrichment was analyzed following previous protocols [23].

Human laryngeal epithelial cells (HLEs) and LSCC cell lines (AMC-HN-8, Hep2, TU177, and TU686) were obtained from Procell (Wuhan, China) and cultured in DMEM supplemented with 10% FBS and Penicillin-Streptomycin. Routine screening for Mycoplasma infection was conducted using polymerase chain reaction (PCR) and fluorescence microscopy. The authenticity of human cell lines was confirmed by analyzing their short tandem repeat (STR) profiles. Lentiviral RNA/siRNA vectors were prepared following established procedures. For cell transfection, the PEI reagent was used to introduce the specified plasmids. HNSC tissues were collected from Hangzhou Hospital of Zhejiang Medical and Health Group and manipulated under the investigation of the Hangzhou Hospital of Zhejiang Medical and Health Group ethics committee. Antibodies, including CRNN (1:1000, Santa Cruz, sc-514602, Dallas, TX, USA), E-cadherin (1:1000, Santa Cruz, sc-8426, USA), N-cadherin (1:1000, Santa Cruz, sc-8424, USA), Ki67 (1:1000, Santa Cruz, sc-23900, USA), and GAPDH (1:1000, Santa Cruz, sc-47724, USA), were acquired from Santa Cruz. For the subcutaneous tumorigenesis assay involving nude mice, 4-week-old mice were injected with the designated cell lines. After 5 weeks, tumor size and weight measurements were conducted. Ethical approval was obtained from the Ethics Committee of Hangzhou Hospital of Zhejiang Medical and Health Group.

We seeded 5000 HNSC cell lines into a 96-well plate and assessed cell growth rates at 24 hours, 48 hours and 72 hours using the MTT reagent. Additionally, cell lines were seeded and stained with Edu reagent after 24 hours, followed by microscopic assessment. For cell cycle analysis via flow cytometry, the cells were collected, fixed with 70% alcohol, stained with PI (propidium iodide), and analyzed using flow cytometry.

HSNC cell lines were seeded into 6-well plates and subjected to a wound assay using 10 µL tips to create wounds. Images were captured at the initial time point (0 h) and after 2 days (48 h) to compare the wound area. For the transwell assay, cells were seeded into the upper chamber with Matrigel treatment, and after 24 hours, the count of migrated cells was analyzed.

Statistical analyses were conducted using the student’s t-test for continuous variables and the chi-square test for categorical variables. Results are expressed as means $\pm$ standard error of the mean (SEM) from three independent experiments. Statistical significance was set at a p-value of 0.05 or lower.

Initially, bioinformatics methods were employed to analyze DEGs in LSCC using data from GSE143224. As illustrated in Fig. 1A,B, significant differences in gene expression were observed between the tumor and non-malignant groups. Subsequently, we utilized the R software “limma” package for further screening and analysis of the DEGs and the results are shown in the volcano and heatmap plots of Fig. 1C,D. Among these DEGs, there were 36 upregulated genes and 65 downregulated genes, and they were primarily associated with processes such as epidermis development, cell differentiation, cell adhesion and glutathione metabolism (Fig. 2A,B). The bubble plot also highlighted the significant roles of DEGs in regulating epidermis development and glutathione metabolism (Fig. 2C). Moreover, among the DEGs, CRNN was identified as a critical interaction gene (Fig. 2D).

Fig. 1.

Analysis of DEGs in NSC from GEO dataset. (A,B) Batch correction analysis of DEGs between tumor and non-malignant. (C,D) Volcano plot and heatmap analysis of DEGs of HNSC. DEGs, differentially expressed genes; GEO, Gene Expression Omnibus.

Fig. 2.

Enrichment analysis of DEGs and PPI analysis. (A–C) GO-KEGG analysis of DEGs by metascape and R language. (D) Protein-protein interaction measured by STRING. GO-KEGG, gene ontology-kyoto encyclopedia of genes and genomes; PPI, protein-protein interaction.

Next, we further analyzed the expression of CRNN in HNSC (LSCC is a subtype of HNSC) (Fig. 3A,B), and the results indicated downregulation of CRNN expression in tumor samples compared to normal samples in HNSC. Moreover, CRNN expression was also decreased in LSCC compared with the normal specimens from GSE143224 (Fig. 3C). Additionally, we validated the downregulation of CRNN mRNA in LSCC patients’ tissues (Table 1) in comparison to matched adjacent normal tissues (Fig. 3D). The expression of CRNN protein was found to be consistently lower in the tumor group compared to the normal group, as indicated by both IHC (Fig. 3E) and western blotting (Fig. 3F). Survival analysis using Kaplan-Meier plots revealed that patients with high CRNN expression had longer survival rates compared to those in the low CRNN expression group (Fig. 3G,H). These findings suggest a positive correlation between CRNN downregulation and poor prognosis in LSCC.

Fig. 3.

Dowregulation of CRNN expression in LSCC. (A) Column plot showing CRNN expression in HNSC based on TCGA data. *** represents the Tumor group compared to the Normal group. (B) GEPIA online analysis illustrating CRNN expression in HNSC. * represents the tumor group compared to the normal group. (C) Box plot demonstrating CRNN expression in tumor and normal tissues. *** represents the tumor group compared to the normal group. (D) CRNN expression in 97 pairs of HNSC tissue specimens. *** represents the Tumor group compared to the Normal group. (E,F) Analysis of CRNN expression in tissue specimens using IHC and western blotting, respectively. (G) Kaplan-Meier Plotter online analysis assessing the relationship between CRNN and patient survival. (H) Kaplan-Meier curve evaluating the impact of CRNN expression on patient survival rates. IHC, Immunohistochemistry; HR, hazard rate.

We further investigated the relationship between CRNN expression and the tumor microenvironment. As shown in Fig. 4A, a significantly positive correlation was observed between CRNN expression and CD8+ T cells, while CRNN expression displayed a negative association with tumor purity and macrophage infiltration. Additionally, CRNN exhibited positive correlations with mast cells, B cells and neutrophils, and conversely, it displayed negative correlations with NK cells and macrophages (Fig. 4B). Furthermore, CRNN expression was positively correlated with T-cell enrichment (Fig. 4C). These findings collectively indicate that CRNN is closely associated with the immune profile of tumor cells. However, the underlying mechanisms through which CRNN regulates immune responses in LSCC warrant further investigation.

Fig. 4.

Correlation between CRNN and TME. (A) Scatter diagram illustrating the correlation between infiltration and CRNN expression. (B,C) Bubble plot and scatter plot displaying the relationship between CRNN and T cell enrichment. TME, tumor microenvironment.

To confirm the reduced expression of CRNN in LSCC, we compared CRNN expression between LSCC cell lines and human laryngeal epithelial cells. As shown in Fig. 5A, the expression of CRNN was significantly lower in AMC-HN-8, Hep2, TU177 and TU686 compared to HLEs. To investigate the functional role of CRNN in LSCC, we utilized CRNN overexpression and shCRNN plasmids to create CRNN-OE and KD cell lines in AMC-HN-8 and TU177, respectively (Fig. 5B). As illustrated in Fig. 5C, CRNN overexpression led to a decrease in cell growth in AMC-HN-8 and TU177 cell lines, while CRNN knockdown resulted in increased cell growth in these cell lines. Cell proliferation, assessed using Edu staining in the same cell lines, yielded similar changes (Fig. 5D), further confirming the role of CRNN in cell proliferation regulation. Moreover, CRNN overexpression inhibited cell cycle progression, while CRNN deficiency increased the proportion of cells in the G2 phase compared to control group cell lines (Fig. 5E). These findings collectively indicate that CRNN negatively regulates cell viability and cell cycle progression in LSCC cells.

Fig. 5.

CRNN inhibits cell growth in LSCC cells. (A) Western blotting showing the expression of CRNN in LSCC cell lines. (B) Expression of CRNN in TU177 and AMC-HN-8 cells after transfection with the indicated plasmids, as measured by western blotting. (C) Cell growth of CRNN-OE or knockdown TU177 and AMC-HN-8 cells assessed by MTT assay. (D) Edu assay was conducted to determine cell viability in AMC-HN-8 and TU177 cells with the indicated plasmids transduction. (E) Flow cytometry assay was performed to assess cell cycle in CRNN-OE or KD TU177 and AMC-HN-8 cells. Data represent three independent experiments. *p 0.01. ***p 0.001. #p 0.05. ##p 0.01. ###p 0.001. * indicates CRNN group compared to NC group. # indicates shCRNN-1# group compared to Scramble group.

Subsequently, we assessed the impact of CRNN on cell migration and invasion using the same cell lines and found that CRNN overexpression led to reduced cell migration in both AMC-HN-8 and TU177 cell lines, while CRNN knockdown had the opposite effect, promoting cell migration in these cell lines (Fig. 6A). Additionally, CRNN overexpression inhibited cell invasion, whereas CRNN depletion increased cell invasion in both AMC-HN-8 and TU177 cell lines (Fig. 6B). Furthermore, CRNN overexpression increased E-cadherin levels and decreased N-cadherin levels in AMC-HN-8 and TU177 cell lines, whereas CRNN deficiency had the opposite effects (Fig. 6C), indicating that CRNN negatively regulates cell migration and invasion in LSCC cells.

Fig. 6.

CRNN reduces cell migration in LSCC cells. (A) Assessment of the impact of CRNN-OE or deficiency on cell migration in TU177 and AMC-HN-8 cell lines using a wound-healing assay. (B) Transwell assay measuring cell migration in TU177 and AMC-HN-8 cell lines. (C) Western blotting analysis of N-cadherin and E-cadherin expression. Data represent three independent experiments. *p 0.01. **p 0.01. ***p 0.001. #p 0.05. ##p 0.01. ###p 0.001. * represents the CRNN group compared to the NC group. # represents the shCRNN-1# group compared to the Scramble group.

Similarly, we introduced AMC-HN-8 cells transduced with CRNN plasmids into subcutaneous injections in nude mice. As shown in Fig. 7A, CRNN overexpression resulted in reduced tumor growth in nude mice, whereas knockdown of CRNN induced tumor growth. Furthermore, in tumors from nude mice with CRNN overexpression, we observed a decrease in the expression of Ki67 and N-cadherin, while E-cadherin expression increased (Fig. 7B). Conversely, CRNN deficiency led to an increase in Ki67 and N-cadherin expression, accompanied by reduced E-cadherin expression in tumors. Collectively, our data strongly suggest that CRNN negatively regulates tumor growth in vivo.

Fig. 7.

CRNN overexpression suppresses tumor growth in nude mice. (A) Nude mice were divided into four groups: NC, CRNN, scramble, and shCRNN-1#. Images of tumors in each group, as well as tumor volume and weight, are presented in (A). (B) IHC analysis of tumors with matched staining intensity from nude mice was conducted, as shown in Fig. 3E. **p 0.01. ***p 0.001. # represents the p 0.05. ###p 0.001. * represents the CRNN group compared to the NC group. # represents the shCRNN-1# group compared to the Scramble group.