GNG7 is found to be one of the key regulators of LUAD tightly related to the prognosis through large-scale screening
To find key regulators of LUAD, we applied the DESeq2 R package to screen for DEGs in LUAD based on TCGA-LUAD datasets. The results showed that there were 7741 up-regulated and 3783 down-regulated genes among the screened 11524 DEGs (Fig. 1a). Combined with further prognostic analysis, we found that among the DEGs, nine genes (HSD17B6, PXMP4, HLF, ADGRD1, CYP17A1, ESYT3, FCAMR, C11orf16, GNG7) were significantly and positively associated with LUAD prognostic indicators including OS, DSS and PFI (Fig. 1b). Of note, although GNG7 has been reported to be differentially expressed in a variety of tumors, its roles in the initiation and progression of LUAD remain unclear. In the present study, we focused on GNG7 to explore the underlying mechanism and clinical significance in LUAD.
GNG7 expression is down-regulated in LUAD
To elucidate the expression pattern of GNG7 in cancers, we first evaluated the expression of GNG7 in 33 types of cancers by a systematic analysis based on the TCGA databases. The results showed that GNG7 expression was significantly down-regulated in 17 different tumors, including Bladder Urothelial Carcinoma (BLCA), Breast invasive carcinoma (BRCA), Colon adenocarcinoma (COAD), Lung adenocarcinoma(LUAD) et al, while it was significantly up-regulated in Cholangiocarcinoma (CHOL), Liver hepatocellular carcinoma (LIHC) and Pheochromocytoma and Paraganglioma(PCPG)(Fig. 1c). Then, the low expression of GNG7 in LUAD was further validated by using three GEO datasets (GSE32665, GSE32863, GSE43458) (Fig. 1d-f). The paired analysis got the similar results (Fig. 1g). In addition, the decreased protein level of GNG7 was also observed in LUAD using the UALCAN database (Fig. 1h). Moreover, ROC curve analysis was employed to analyze the distinguishing efficacy of GNG7 between LUAD tissue and normal tissue. The area under the curve (AUC) of GNG7 is 0.871, suggesting that GNG7 may be an ideal biomarker to distinguish LUAD from normal tissue (Fig. 1i). Together, these results indicated that GNG7 is lowly expressed in LUAD which may be a potential diagnostic marker for LUAD.
Association between clinicopathological characteristics and GNG7 expression in LUAD
To clarify the correlation between the expression of GNG7 and clinicopathological variables, we collected data from the TCGA database on 535 patients with LUAD. After data preprocessing, the relationship between gene expression profiles and clinicopathological characteristics of 513 LUAD patients was shown in the baseline data table (Table S1). The results showed that low expression of GNG7 was positively associated with high T stage, Gender (male sex), poor primary therapy outcome and high pathologic stage of LUAD, while there were no significant associations between GNG7 expression and the other clinical factors such as N stage and M stage (Fig. 2a-f). In line with these findings, the logistics regression analysis also revealed that GNG7 expression was significantly associated with T stage (OR =0.403, 95% CI: 0.274-0.589, p<0.001), Pathologic stage (OR=0.538, 95% CI: 0.377-0.766, p<0.001), Primary therapy outcome (OR=0.311, 95% CI: 0.133-0.707, p=0.006) and Gender (OR=0.502, 95% CI: 0.352-0.713, p<0.001) (Fig. 2g).
Significance of GNG7 in clinical prognosis of LUAD and clinical subgroup analysis
We utilized data from the TCGA database to investigate the prognostic significance of GNG7 in LUAD. Kaplan-Meier survival analysis based on the TCGA-LUAD dataset revealed that low expression of GNG7 was associated with poor OS (HR=0.51, 95% CI: 0.38-0.69, P<0.001), DSS (HR=0.56, 95% CI: 0.38-0.82, P=0.003) and PFI (HR=0.64, 95% CI: 0.49-0.85, P=0.002) (Fig. 3a-c). To further validate the prognostic value of GNG7 in LUAD, we utilized the PrognoScan database for further study. We included two of the GSE datasets (GSE31210 and GSE13213) in our analysis, where low GNG7 expression was significantly associated with the poorer prognosis (OS, HR=0.21, 95% CI: 0,08-0.52, Cox P=0.000748; RFS, HR=0.25, 95% CI: 0,13-0.49, Cox P=0.000069 in the GSE31210 dataset; OS, HR=0.48, 95% CI: 0,34-0.67, Cox P=0.000023 in the GSE13213 dataset) (Fig. 3d-f).
Moreover, the univariate Cox regression analysis model showed that GNG7 expression level was significantly associated with OS (HR: 0.702; 95% CI: 0.599-0.822; P<0.001) similar to T stage, N stage, M stage and Pathologic stage as well Primary therapy outcome and Residual tumor (Fig. 3g). Meanwhile, the multivariate Cox regression analysis also revealed that low expression of GNG7, similar to Primary therapy outcome and Residual tumor, was an independent risk factor for the prognosis of LUAD patients (Fig. 3h). Collectively, these results suggest that low expression of GNG7 independently predicts poor prognosis for patients with LUAD.
Given that multivariate Cox regression analysis identified low expression of GNG7 as an independent risk factor, we investigated the potential prognostic value of GNG7 in LUAD patients with different clinical subgroups. As shown in Fig. 4a-c, low expression of GNG7 was associated with poor prognosis in stage N0, including OS (HR=0.42, 95% CI: 0.27-0.67, P<0.001), DSS (HR=0.38, 95% CI: 0.21-0.69, P=0.001) and PFI (HR=0.57, 95% CI: 0.40-0.83, P=0.003). However, there was no statistically significant correlation between GNG7 expression and prognosis in the N1&N2&N3 stage (p>0.05) (Supplementary Fig. 1a-c). In addition, low GNG7 expression was significantly associated with poor prognosis in LUAD patients in M0 stage, including OS (HR=0.48, 95% CI: 0.33-0.68, p<0.001), DSS (HR=0.59, 95% CI: 0.37-0.94, p=0.026), PFI (HR=0.66, 95% CI: 0.47-0.92, P=0.015) (Fig. 4d-f). Nevertheless, no significant association was shown between GNG7 expression and prognosis in LUAD patients in the M1 stage (Supplementary Fig. 1d-f). These results suggest that low expression of GNG7 is positively associated with the poor prognosis of LUAD patients without lymph node invasion and distal metastasis.
Construction and validation of a nomogram based on the independent clinical risk factors
To provide a quantitative approach to predicting the prognosis of LUAD patients, we constructed a prognostic nomogram to predict individual survival probability based on the expression levels of GNG7 and other independent clinical risk factors (Fig. 4g). The calibration curve of the nomogram showed that the established lines of 1-, 2-, and 3-y survival highly matched the ideal line (the 45-degree line) (Fig. 4h). In addition, the C-index of the prediction model reached 0.690 (0.659-0.720), indicating that the model had a reliable potential to predict the OS of LUAD patients.
Functional enrichment and pathway analysis of GNG7-associated DEGs in LUAD
To investigate the biological functions and signaling pathways associated with GNG7, we examined the DEGs between GNG7-high and GNG7-low patients which were stratified based on the median GNG7 expression. Resultantly, 1403 mRNAs (492 upregulated and 911 downregulated), 962 lncRNAs (256 upregulated and 706 downregulated), and 21miRNAs (18 upregulated and 3 downregulated) were differently expressed in GNG7-high patients compared to GNG7-low ones (Fig. 5a, Supplementary Fig. 2a and 2c). Relative expression values of some representative DEGs between the two cohorts were shown in the form of heatmaps (Fig. 5b, Supplementary Fig. 2b and 2d). Strikingly, pathway enrichment analysis showed that the DEGs were most strongly enriched in the B cell receptor signaling pathway, T cell receptor signaling pathway and HIV infection allograft rejection which are highly related to the cellular immune response (Fig. 5c-g). These data suggested that GNG7 may play an important role in regulating the tumor immune microenvironment of LUAD.
Correlation analysis between the expression of GNG7 and immune cell infiltration in LUAD
As reported, tumor-associated immune cell infiltration has a close relationship with tumor development and the prognosis of patients. Then, we utilized the ESTIMATE algorithm to assess the correlation between GNG7 and the abundance of immune cell infiltration in LUAD. The results revealed that GNG7 expression was positively correlated with the abundance of immune cell infiltration in LUAD (Fig. 6a). Specifically, further Spearman correlation analysis showed that among 24 immune cell subpopulations, GNG7 expression was positively correlated with most immune cell subsets, including Mast cell, DC, B cells, and CD8+ T cells, but negatively correlated with Th2 and Tgd cells (Fig. 6b). Consistently, the ssGSEA analysis demonstrated that the infiltration levels of most of the immune cell subsets such as Mast cells, pDCs, B cells, NK cells and CD8+ T cells were remarkably increased in LUAD patients with GNG7 high expression compared to those with GNG7 low expression (Fig. 6c). In keeping with this finding, GNG7 was significantly correlated with most immune markers of different immune cells, including CD8+T cell, B cell, Neutrophils, and Dendritic cell (Table S2). Together, these results suggest that GNG7 may contribute to the remodeling of the immune microenvironment in LUAD through promoting the infiltration of a variety of tumor-associated immune cells.
GNG7 high expression with high B cell infiltration predicts a better prognosis of LUAD patients
Of the infiltrated immune cells increased in LUAD with GNG7 high expression, B cell infiltration attracts our attention as the relatively less knowledge of this cell type in tumor immunotherapy currently. Our results showed that GNG7 was closely associated with the level of B cell infiltration in LUAD (Supplementary Fig. 3a). Specifically, the level of B cell infiltration was significantly elevated in the GNG7 high expression group compared with that in the GNG7 low expression group (Supplementary Fig. 3b). In addition, we respectively investigated the correlation of GNG7 with B cells in the tumor and normal tissues. Strikingly, GNG7 showed a strong positive correlation with the B cell marker genes CD19 and CD79A in LUAD tissues (Fig. 7a). In contrast, the correlation of GNG7 with the B cell markers CD19 and CD79A did not reach statistical significance in normal tissues (Fig. 7b). These results suggest that GNG7 expression may promote B cell infiltration in the context of LUAD. Meanwhile, through KM plot database analysis, we found that patients with high GNG7 expression tended to predict a better prognosis in the B cell enriched group but not in the B cell decreased group (Fig. 7c-d). Such finding was further corroborated by the analysis using the TIMER2.0 database, implying that high GNG7 expression corresponded to a better prognosis for LUAD patients in the context of enriched B cell infiltration. Furthermore, we found that high infiltration levels of B cells in the presence of consistent levels of GNG7 expression corresponded to a good prognosis in patients with LUAD (Fig. 7e). Taken together, it is reasonable to suggest that GNG7 may have improved patient prognosis by promoting B cell infiltration.
GNG7 dysregulation is associated with aberrant DNA methylation
Considering the importance of DNA methylation in regulating gene expression, we tested whether aberrant DNA methylation occurs in GNG7 gene in LUAD. By analyzing the data from the UALCAN database, we found that the methylation level of GNG7 was significantly higher in the tumor group compared to that in the normal group (Fig. 8a). Next, the correlation analysis based on the cBioPortal database showed that GNG7 expression was significantly negatively correlated with methylation (Fig. 8b). To further investigate the methylation of GNG7 in LUAD, we analyzed the methylation levels of different CpG sites of GNG7 in LUAD patients using the MethSurv database and presented them in the form of heat maps (Fig. 8c). The results revealed that several CpG sites of GNG7 exhibited high methylation in LUAD patient samples, including cg19477361, cg21462934, and cg27181295. Prognostic analysis showed that the above CpG sites with highly methylated levels were associated with poor prognosis in LUAD (Fig. 8d-f). These results suggest that the low expression of GNG7 in LUAD may be partly due to the methylation modification of the abovementioned CpG sites and plays a key role in tumor progression.