Pan-cancer analysis of SMC2 mRNA and up-regulation of expression in LUAD
Given the role of SMC2 in a variety of human cancers, we performed a pan-cancer analysis of SMC2 mRNA levels by using the TCGA database. The results revealed that SMC2 was highly expressed in a number of cancers, including lung adenocarcinoma and uterine corpus endometrial carcinoma, as well as colon adenocarcinoma (see Fig. 1A for details). Then, we comprehensively analyzed the effects of SMC2 on overall survival (OS), DSS, and progression-free interval (PFI) in cancer patients, and the results revealed that the upregulation of SMC2 might be adverse to the prognosis of LUAD patients (Fig. 1, B to D). As a result, we chose LUAD as the targeted carcinoma in this study. Subsequently, the expression levels of SMC2 in LUAD are shown in Fig. 1, E and F. This indicates that the expression levels of SMC2 are up-regulated in LUAD tissues, regardless of the inclusion of data from the GTEx database. This trend was as well observed in paired samples (Fig. 1G). Finally, high expression of SMC2 was verified in LUAD cell lines (A549, H1975 and PC9) and human bronchial epithelial cells BEAS-2B, as well as in 8 pairs of LUAD tissues and adjacent normal tissues from our medical center (Fig. 1, H and I). Taken together, these results suggest that SMC2 is highly expressed in LUAD.
The expression of SMC at the protein level in LUAD
To begin with, in the CPTAC database of the UALCAN platform, we found that the protein expression level of SMC2 in LUAD was increased relative to normal tissues (Fig. 2A). Next, SMC2 protein expression levels in LUAD were derived from immunohistochemical staining data in the HPA database (antibody HPA071309). Consistent with the upregulation of mRNA levels, SMC2 protein expression was also significantly upregulated in LUAD tissues (Patient ID: 1907, 2003, and 4923) relative to normal tissues (Patient ID: 1470, 1678, and 2643) (Fig. 2B). Likewise, protein expression levels of SMC2 were progressively higher in pathological grades 2 and 3 compared with the normal group, although there was no significant change in grade 1 (Fig. 2C). As shown in Fig. 2, D to F, alterations in the mTOR and HIPPO pathways combined with the WNT pathway also contributed to elevated SMC2 protein expression levels.
Relationship between SMC2 and clinicopathologic features
From Fig. 3, A to I, we can observe that SMC2 was remarkably overexpressed in LUAD patients with T2-T4 stage, N1-N3 stage, M1 stage, higher pathologic stage and OS fatal event, along with in patients with DSS/PFI events. In view of these evidences, we can speculate that SMC2 may adversely influence the survival of patients.
Diagnostic and prognostic value of SMC2 in LUAD
To start with, the diagnostic effect of SMC2 in LUAD was assessed using ROC curves. As shown in Fig. 4A, the AUC value of SMC2 in distinguishing LUAD tissues from normal tissues was 0.787. The Kaplan-Meier curve revealed the effect of SMC2 expression on the tumor prognosis of patients with LUAD, the SMC2 high-expression group was correlated with poorer OS, DSS, and PFI in LUAD patients, with the risk ratios (HR) of 1.49, 1.72, and 1.51, respectively (Fig. 4, B to D) (p<0.05). In the meantime, we performed univariate (Fig, 4E) and multivariate (Fig, 4F) COX regression analyses, which further confirmed that SMC2 might be an independent prognostic factor for LUAD. In addition, according to the results of COX analysis, we further constructed Nomogram to predict the DSS of LUAD patients at 1-, 3-, and 5 years (Fig, 4G). It was reassuring that the C-index for evaluating its predictive effect was 0.701, and the calibration curve objectively demonstrated the comparatively good agreement between the predicted and actual values (Fig. 4, H to J).
Filtering mRNA potentially relevant to SMC2
We first selected the top 800 mRNAs that were closely associated with SMC2 in LUAD, and the heatmap in Fig. 5A shows the top 20 molecules. Then, differentially expressed mRNAs in LUAD were screened by volcano plots according to the filtering standard, in which 750 mRNAs were either up- or down-regulated (Fig. 5B). Besides, Venn diagrams further identified overlapping genes in the two groups (Fig. 5C). These 46 target molecules were then constructed into a PPI network through the STRING database, and the hub genes were selected based on the centrality of the nodes (Fig. 5D). In the end, Kaplan-Meier survival analysis of mRNAs (NDC80, KIFC1, SKA1, NCAPH, ESPL1, MELK, KIF11, SGO1, TOP2A, KNL1, KIF4A, TPX2, TICRR, TTK, KIF14, NCAPG) was performed to explore more about their effects on OS in LUAD patients. Interestingly, upregulation of all these molecules in LUAD exacerbated the poor prognosis of OS in patients (Fig. 6, A to P). Apparently, the detrimental effect of these mRNAs on the survival prognosis of LUAD patients is consistent with the tendency of SMC2.
Exploring the potential mechanisms of SMC2 in LUAD
Based on the 46 targets mRNA obtained from the pre-screening, we performed GO analysis and found that SMC2 may be involved in the following biological processes, including: nuclear division, mitotic sister chromatid segregation, cellular process involved in reproduction in multicellular organism, DNA replication, et al. Meanwhile, it is involved in cellular components such as condensed chromosome, centromeric region, kinetochore and mitotic spindle, and its potential molecular function is mainly to influence tubulin binding (Fig.7, A and B). In addition, we explored the potential pathway of SMC2 in LUAD using GSEA. As shown in Fig. 7, C and D, the significantly enriched KEGG pathways included cell cycle, oocyte meiosis, homologous recombination, human T-cell leukemia virus 1 infection. The REACTOME pathways included DNA damage telomere stress induced senescence, DNA replication, cell cycle checkpoints, epigenetic regulation of gene expression as well as cellular senescence. These findings may provide a reference for further studies in the future!
SMC2 genetic alteration in LUAD
The COSMIC website shows the distribution of different types of SMC2 mutations in cancer (Fig. 8, A and B). Missense substitutions were the most common type of mutation, following by synonymous substitutions, and G>A, C>T and A>G were the most common substitution mutations. Next, the SMC2 gene mutations in LUAD were analyzed by three datasets in the cBioPortal database. The results indicated that SMC2 was genetic alteration in 2.4% of LUAD patients (Fig. 8C). As shown in Fig. 8D, we can visualize the specific types and frequency of gene mutations. SMC2 mRNA expression was elevated in the shallow deletion group compared with the diploid group, and interestingly, the expression of gain groups was also increased (Fig. 8E). In the LUAD samples, 23 SMC2 missense mutation sites were shown, one of which was K879N, suggesting that this is one of the protein activation hotspots, as well as two SMC2 truncating mutation sites (Fig. 8F).
Association between SMC2 expression and immune characteristics in LUAD
To further understand the role of SMC2 in LUAD, we first analyzed the correlation between SMC2 and TILs. Based on the ssGSEA algorithm, we found that SMC2 was negatively correlated with CD8+ T cells, Th 17 cells, mast cells, dendritic cells (DC), T follicular helper cells (TFH), B cells, etc (Fig. 9, A and B). On the other hand, using the TIMER algorithm and the TIDE algorithm in the TIMER database, we found that SMC2 was negatively and positively correlated with activated B cells and MDSC, respectively (Fig. 9C). Notably, when activated B cells were enriched, the OS of LUAD patients was significantly better, however, when MDSC was enriched, the OS of LUAD patients was dramatically worse (Fig. 9, D and E). Taking the above information together, it is reasonable to hypothesize that the cancer-promoting effects of SMC2 may be related to a certain extent to the low enrichment of B cells as well as the high enrichment of MDSC. In Fig. 9F, the arm-level deletion of SMC2 copy number was associated with reduced abundance of CD8+ T cells, CD4+ T cells, neutrophils, and dendritic cells when compared with the diploid/normal state. These results indicate that copy number changes of SMC2 in LUAD may be an element that regulates the immune microenvironment.
Immune checkpoints play a vital role in the immune microenvironment of LUAD, and these directly regulate the resident's anti-tumor immune response(Chi et al. 2021). Therefore, in this case, we next analyzed the association between SMC2 and immunoinhibitor (Fig. 10A). Interestingly, SMC2 was significantly positively correlated with CD274 (r = 0.298, p < 6.26e-12), PDCD1LG2 (r = 0.226, p < 2.29e-07), TGFBR1 (r = 0.208, p < 1.85e-06), and LAG3 (r = 0.181, p < 3.54e-05) (Fig. 10B). We also analyzed beside the association between SMC2 and immunostimulator (Fig. 10D). It is also intriguing that SMC2 was markedly negatively correlated with TNFSF13 (r = -0.471, p < 2.2e-16), TMEM173 (r = -0.404, p < 2.2e-16), TNFRSF14 (r = -0.399, p < 2.2e-16), and HHLA2 (r = -0.282, p < 7.53e-11) (Fig. 10C).
Chemokines and chemokine receptors are essential for tumor infiltration by immune cells(Li et al. 2015). Therefore, we analyzed the correlation between SMC2 expression levels and immune cell chemokines and receptors in LUAD using the TISIDB database. Heatmap results showed that several chemokines and receptors were significantly correlated with SMC2 expression in LUAD (Fig. 11, A and D). Next, we concretely analyzed the correlation between SMC2 expression and chemokines/receptors. The results showed that SMC2 expression was negatively correlated with CCL14 (r = -0.378, p < 2.2e-16), CCL17 (r = -0.363, p < 5.74e-18), CXCL16 (r = -0.347, p < 5.66e-16), CX3CL1 (r = -0.268, p < 6.93e-10), CX3CR1 (r = -0.31, p < 7.89e-13), CCR6 (r = -0.246, p < 1.56e-08), CCR7 (r = -0.197, p < 6.6e-06), and CXCR5 (r = -0.14, p =0.00146) (Fig. 11, B and C), and these results revealed that the SMC2 gene may play an influential role in tumor immune. In addition, we analyzed the relationship between SMC2 and LUAD immune subtypes. As shown in Fig. 11E, SMC2 was highly expressed in type C2 (IFN-γ-dominant) and type C4 (lymphocyte-depleted), while it was least expressed in type C3 (inflammatory). This implies that the expression of SMC2 is directly related to the immune microenvironment of LUAD. Finally, The Tumor Immune Dysfunction and Rejection (TIDE) algorithm is extensively used to predict cancer immunotherapy response, with higher TIDE scores being associated with poorer immunotherapy outcomes (Jiang et al. 2018). TIDE scores were dramatically higher in those with high SMC2 expression compared to those with low expression, which resulted in a notably lower rate of response to their immunotherapy (Fig. 11F).