TMC family expression in RCCC and normal tissues
As displayed in Fig. 1A-H, TMC2, TMC6, TMC7, and TMC8 were high expressed upregulated in RCCC, while TMC1, TMC3, TMC4, and TMC5 was high expressed in normal kidney tissues. Subsequently, we used the same method to evaluate TMC gene family expression in 72 pairs of RCCC and adjacent samples. The results showed no significant difference in TMC2 expression between tumor and normal tissues, and expression of other TMCs was consistent with that before (Fig. 1I-P).
In addition, we utilized the UALCAN and GEPIA2 databases to contrast the expression levels of TMC1–TMC8 in RCCC and normal renal tissues again. The results of the two online databases are roughly the same as those obtained by R software (Figure S1). And we examined the difference in TMC family expression between normal renal cortex proximal convoluted tubule epithelial cells and clear cell renal cell carcinoma cells. Compared with normal renal cells, TMC7 and TMC8 in RCCC cells (786-O and ACHN) significantly upregulated, while the expressions of TMC1 and TMC4 were decreased (Fig. 2A-H).
To the transcription, we analyzed the immunohistochemistry (IHC) staining of normal kidney and RCCC tissues from the HPA database. A lower TMC5 expression in the tumors was discovered, as displayed in Fig. 2I. And the protein level of the TMC7 and TMC8 upregulated in RCCC compared to normal samples.
Role of the TMC family in the survival of RCCC patients
We used the online Kaplan- Meier Plotter tool to analyze the relationship between TMCs and Overall Survival (OS) in RCCC patients. As shown in Fig. 3A-H, high expression of TMC2, TMC3, TMC5, and TMC8 and low TMC7 indicated a shorter OS of RCCC patients.
In addition, we utilized R software to verify the correlation between the expression of TMC family members and the survival of RCCC patients. Our results illustrated patients with a high level of TMC2, TMC3, and TMC4 and those with a low TMC7 had a shorter OS (Fig. 3I-P). High expression of TMC2, TMC3, and TMC5 and low TMC7 negatively correlated with the Disease-Specific Survival (DSS) of RCCC patients (Figure S2A-H). And the increased expression of TMC3 and TMC5 and low TMC6 and TMC7 was negatively correlated with the Progress Free Interval (PFI) of RCCC patients (Figure S2I-P).
Univariate COX analysis and multivariate COX analysis were performed to explore the role of TMCs in the survival of RCCC patients. Results of the univariate COX regression model indicated the high levels of TMC2, TMC3, and TMC5 were risk factors (Hazard ratio > 1, p < 0.001), and high TMC7 expression was a protective factor (Hazard ratio < 1, p < 0.05). However, multivariate COX analysis showed that only TMC2 was an independent prognostic factor (Table 2).
Table 2
Univariate analysis and multivariate analysis of TMCs
Characteristics
|
Total(N)
|
Univariate analysis
|
Multivariate analysis
|
Hazard ratio(95% CI)
|
P value
|
Hazard ratio(95% CI)
|
P value
|
TMC1
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
0.797(0.590–1.076)
|
0.139
|
|
|
TMC2
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
1.696(1.250–2.300)
|
< 0.001
|
1.950(1.241–3.062)
|
0.004
|
TMC3
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
1.792(1.317–2.438)
|
< 0.001
|
1.320(0.856–2.036)
|
0.209
|
TMC4
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
1.111(0.824–1.498)
|
0.489
|
|
|
TMC5
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
1.696(1.247–2.308)
|
< 0.001
|
1.307(0.850–2.009)
|
0.222
|
TMC6
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
0.882(0.653–1.190)
|
0.410
|
|
|
TMC7
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
0.727(0.537–0.984)
|
0.039
|
0.912(0.595–1.397)
|
0.671
|
TMC8
|
539
|
|
|
|
|
Low
|
269
|
Reference
|
|
|
|
High
|
270
|
1.233(0.914–1.663)
|
0.170
|
|
|
TMC family was related to histologic grade and pathologic stage of RCCC
We divided RCCC patients into two cohorts according to the tumor's histological grade or pathological stage and investigated the TMC family expression differences using R software. As it turned out in Fig. 4A-H, the expressions of TMC1, TMC3, TMC5, TMC7, and TMC8 significantly differed between low and high histological grades of RCCC (GradeⅠ&Ⅱ vs. GradeⅢ&Ⅳ). As we expected, TMC1, TMC3, TMC5, TMC7, and TMC8 also showed the same trend between early and late pathological stages (StageⅠ&Ⅱ vs. StageⅢ&Ⅳ) as the change in histological grade cohorts(Fig. 4I-P). And then, the "pathological stage map" module of the GEPIA2 website was used to further confirm the relationship between the expression of TMCs and the pathological staging of RCCC. Compared with the R software results, the TMC1, TMC3, TMC7, and TMC8 showed significant pathological stage-specific changes except for TMC5 (Figure S3A-H).
ROC analysis was performed to comprehensively evaluate the role of TMC family expression in the prognostic outcome of RCCC patients. In the ability to predict normal and tumor outcomes, the predictive ability of TMC2, TMC3, TMC5, and TMC6 has low accuracy (AUC between 0.5 and 0.7). The ability of TMC1, TMC7, and TMC8 had a certain accuracy (AUC between 0.7 and 0.9), and TMC4 had the highest accuracy (AUC = 0.903, CI = 0.870–0.936) as showed in the Figure S3I-P.
TMC family was associated with immune infiltration in RCCC
The tumor microenvironment (TME) is composed of tumor cells, recruited cells (e.g., immune cells, vascular endothelial cells, and stromal cells), secreted products of corresponding cells (such as cytokines and chemokines), and non-cellular components of the extracellular matrix. Tumor-associated immune cell infiltration is an integral component of TME and is closely related to the prognosis and lymph node metastasis of malignant tumors.
Our study found that TMC2, TMC6, TMC7, and TMC8 upregulated, while TMC1, TMC3, TMC4, and TMC5 downregulated in RCCC tissues. It seemed unreasonable that RCCC patients with high TMC2, TMC3, TMC4, TMC5, and TMC8 and those with low TMC6 and TMC7 had a shorter survival time. But RNA-seq is a bulk-seq in which a variety of cells intermingled. Genes that RNA-seq highly expresses in tumor tissue are likely to be molecules in immune cells, depending on the immune infiltration of the tumor.
According to the available data, we calculated 64 immune cells in RCCC using the "xCELL" algorithm using R software. Interestingly, the expression of TMC2, TMC5, TMC6, and TMC8 positively correlated with the number of B cells, while TMC7 negatively correlated with B cells. TMC4 and TMC6 negatively correlated with the degree of CD4 + T cells. The expression of TMC6 and TMC8 positively correlated with the number of CD8 + T cells, while TMC1, TMC5, and TMC7 negatively correlated with CD8 + T cells. The expression of TMC3, TMC5, TMC6, and TMC8 positively correlated with the number of NKT cells, while TMC7 negatively correlated with NKT cells. TMC6 and TMC8 positively correlated with the degree of macrophages, while TMC1, TMC2, and TMC7 negatively correlated with macrophages. The expression of TMC6 and TMC8 positively correlated with the number of dendritic cells, while TMC7 negatively correlated with dendritic cells (Fig. 5A). Next, we utilized the “ssGSEA” algorithm and TIMER2.0 to further verify the influence of TMC family genes on the immune state of RCCC and obtained roughly the same results (Fig. 5B-I, Fig. 6A-H).
The two major non-tumor components of TME, considered predictive biomarkers in cancer patients, are immune and stromal cells. A higher score, as calculated using ImmuneScore or StromalScore, indicates a more significant proportion of immune or stromal components in the TME. ESTIMATEScore is the sum of ImmuneScore and StromalScore, and thus represents the combined ratio of these two components in the TME. The results with the R package “estimate” hint that the high expression of TMC3, TMC6, and TMC8 positively correlated with the degree of immune infiltration in RCCC (Fig. 6I-P).
We also conducted the co-expression analysis to explore the association between TMC family expression and immune checkpoint genes in RCCC using the R package “circlize” and TIMER2.0 web. In RCCC patients, the levels of TMC2, TMC6, TMC7, and TMC8 were positively correlated with PDCD1, CD274, CD276, CTLA-4, TIGIT, LAG3, HAVCR2, IL-2, CLCL9, TLR4, ENTPD1, ICOS, CCL5, and CD28, whereas the association between TMC4 and these immune checkpoints was negative (Fig. 7, Figure S4 and Figure S5).
Then, we evaluated the biomarker relevance of TMCs by comparing them with standardized biomarkers based on their predictive power of response outcomes and OS of ICB sub-cohorts in RCCC patients. In Fig. 8, the AUC of TMC6 and TMC8 was greater than 0.5 in 2 of 3 RCCC ICB sub-cohorts (“Custom”). This result implied that TMC6 and TMC8 were more vital as predictive immunological biomarkers than other TMCs.
Correlation analysis among TMC family members
A gene family is a group of genes with significant similarities encoding similar protein products, usually grouped by functional sets. In this part, we explored the relationship among TMC family members using co-expression analysis and the STRING 11.5 database. The mRNA expression of TMC1, TMC2, TMC3, TMC4, and TMC5 was positively correlated with each other, as were the level of TMC6, TMC7, and TMC8. However, the mRNA levels of TMC1-5 and TMC6-8 were negatively correlated (Fig. 9A). The interaction between TMC family proteins was analyzed using the STRING 11.5 database, and the TMC family was closely co-expressed (Fig. 9B).
Functional and pathway enrichment analysis of the TMC family in RCCC
Finally, we utilized the CAMOIP web to analyze the differences in cancer-related function and pathway between the TMCs-high and -low groups by the GSEA method and to predict TMCs-related phenotypes and signaling pathways (Fig. 9C-J).
Interestingly, “ribosome” ranked first in the Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of TMC1, TMC2, and TMC3. Except for TMC5 and TMC6, the other TMCs were all involved in “oxidative phosphorylation.” Except for TMC8, the other TMCs were all engaged in “retinol metabolism” and “complement and coagulation cascades.” It was worth noting that the top ten KEGG analysis results of TMC6 included “allograft rejection,” “primary immunodeficiency,” “intestinal immune network for IgA production,” “Th1 and Th2 cell differentiation”, “complement and coagulation cascades” and “Th17 cell differentiation”. And the top ten KEGG enrichment analysis results of TMC8 included “graft-versus-host disease,” “allograft rejection,” “viral protein interaction with cytokine and cytokine receptor,” “primary immunodeficiency,” “cytokine-cytokine receptor interaction,” “chemokine signaling pathway,” “Th1 and Th2 cell differentiation” and “Th17 cell differentiation”.
Gene Ontology (GO) biological process analysis revealed that almost all TMCs (except TMC4) were involved in the immune response. It was especially true for TMC5, TMC6, and TMC7, since the top ten results of their GO-BP analysis were all related to immunity. GO cellular component analysis showed that most TMCs were involved in the composition of “immunoglobulin complexes” apart from TMC2 and TMC4. Besides, TMC6 and TMC8 were also related to “T cell receptor complexes.” GO molecular function analysis demonstrated that TMC3, TMC7, and TMC8 were associated with “antigen-binding” and “immunoglobulin receptor binding.”
The results concordantly showed the inseparable relationship between the TMC family and immune cell infiltration.