Comparative Transcriptome Analysis Identifies Desmoglein-3 as a Potential Oncogene in Oral Cancer Cells

Abstract

The role of desmoglein-3 (DSG3) in oncogenesis is unclear. This study aimed to uncover molecular mechanisms through comparative transcriptome analysis in oral cancer cells, defining potential key genes and associated biological processes related to DSG3 expression. Four mRNA libraries of oral squamous carcinoma H413 cell lines were sequenced, and 599 candidate genes exhibited differential expression between DSG3-overexpressing and matched control lines, with 12 genes highly significantly differentially expressed, including 9 upregulated and 3 downregulated. Genes with known implications in cancer, such as MMP-13, KRT84, OLFM4, GJA1, AMOT and ADAMTS1, were strongly linked to DSG3 overexpression. Gene ontology analysis indicated that the DSG3-associated candidate gene products participate in crucial cellular processes such as junction assembly, focal adhesion, extracellular matrix formation, intermediate filament organisation and keratinocyte differentiation. Validation of RNA-Seq was performed through RT-qPCR, Western blotting and immunofluorescence analyses. Furthermore, using transmission electron microscopy, we meticulously examined desmosome morphology and revealed a slightly immature desmosome structure in DSG3-overexpressing cells compared to controls. No changes in desmosome frequency and diameter were observed between the two conditions. This study underscores intricate and multifaceted alterations associated with DSG3 in oral squamous carcinoma cells, implying a potential oncogenic role of this gene in biological processes that enable cell communication, motility and survival.

Keywords: GJA1; MMP-13; RNA-Seq; bioinformatics; desmoglein-3; desmosomes; oral squamous carcinoma cells; transmission electron microscopy.

Figure 1

(a) A biclustering heatmap of the expression profile of the top 30 DEGs sorted by their adjusted p-value by plotting their log2 transformed expression values in four samples of the two conditions. (b) The volcano plot displays differentially expressed genes (DEGs) associated with DSG3. Here, the x axis represents the log2 fold change in each gene, while the y axis represents the log10 of its adjusted p-value. Each data point in the scatter plot represents a gene. Genes with an adjusted p-value less than 0.05 and a log2 fold change greater than 1 are marked as red dots, indicating upregulated genes. Blue dots represent downregulated genes, with an adjusted p-value less than 0.05 and a log2 fold change less than −1. The black dots represent the transcripts whose expression levels did not reach statistical significance between the two groups. (c) The biclustering heatmap of the expression profile of the 12 significant DEGs that are highlighted in B (also in Table S2). (d) Bar chart for the top 40 GO terms that are significantly enriched in biological processes with respect to the 12 significant DEGs based on the cutoff value. V: vector control group; D: Dsg3-overexpressing group.

Figure 2

Gene ontology (GO) enrichment analysis of the DSG3 molecular network. (a) Bar chart for the significantly enriched gene ontology (GO) terms of differentially expressed genes (DEGs), with p-value < 0.05. BP: biological processes; CC: cellular components; MF: molecular function. The y axis shows GO terms, whereas the x axis denotes the enrichment score. (b) Bubble plot of pathway analysis. The size of each circle represents the number of genes. The colour scale indicates different thresholds of the p-value.

Figure 3

Gene ontology (GO) based on over- and underrepresentation analysis of the DSG3 molecular network. The bubble plots show overrepresentation analysis results (top panel) and underrepresentation analysis results (bottom panel) of the DSG3-associated genes. The y axis shows the top 10 significantly (p < 0.05) enriched GO terms, and the x axis displays the enrichment score. BP: biological processes; CC: cellular components; MF: molecular function. The colour scale indicates different thresholds of the p-value. The numbers of candidate genes annotated with a GO term are mapped to the scatter plot by point size.

Figure 4

Validation of gene expression by RT-qPCR. (a) qPCR data indicating the relative gene expression (n = 6, pooled from three independent biological samples, mean ± SEM, * p < 0.05, **** p < 0.0001 determined by Student’s t-test). (b) Log2 ratio of D (DSG3-overexpressing) over V (vector) control samples. Except for TFPI2, which was undetectable, all other genes showed a similar trend to that of the RNA-Seq data (Figure 1).

Figure 5

Validation of the protein expression for MMP-13 and connexin-43 (Cx43). (a) Western blotting analysis of the indicated proteins. GAPDH was used as a loading control. While MMP-13 showed an increase in the DSG3-overexpressing cells (D) compared to V (vector) control, Cx43 exhibited an opposite trend with a reduction in the D cell line. Two antibodies for Cx43 raised in two different species were tested here. (b) Cx43 immunostaining in V control and D cells (green). Below is the image quantitation highlighting the pixels above a threshold. The quantitation data are displayed in the bar chart on the right (mean ± SD, **** p < 0.0001).

Figure 6

Transmission electron microscopy (TEM) of H413-D3 (DSG3-overexpressing) and H413-V control lines. Cells were grown to confluence for two days and then were treated with dispase (2.4 U/mL) for 20 min to detach the epithelial sheets. The samples were fixed immediately with glutaraldehyde overnight and stored in a fridge prior to TEM processing. (a–c) showed characteristic desmosomes with symmetrical electron-dense structures in both V control and D samples. However, those presented in D cells appeared slightly less electron-dense than those in V control cells. A cluster of desmosomes was captured in the D sample (black arrows). (d) displays two half-desmosomes presented only in one of the adjacent cells. (e) shows an immature desmosome adjacent to a short stretch of gap junction (red arrow). The enlarged images for the red dotted line boxes are displayed on the right. (f) shows a box whiskers plot of the desmosome size, which was measured in two conditions (V: n = 42; D: n = 37, Midline: median). No statistical significance (NS) was detected between them. Scale bars are 500 nm.

Figure 7

Confocal images for desmoplakin and DSG3 staining, respectively. Both proteins are predominantly distributed at the cell borders but with more enhanced and diffuse DSG3 cytoplasmic staining in D cells compared to V control. The quantitation data are shown underneath the images. Scale bars are 20 µm.