3.1 NCAPD2 expression is up-regulated in pan-cancer
Compared with normal tissues, NCAPD2 mRNA expression was significantly increased in digestive, urinary, female reproductive and respiratory system tissues (Fig. 2A, all P<0.05). Since TIMER1 does not currently include normal tissues, we analyzed the NCAPD2 mRNA expression of other tumor tissues in GEPIA2 database instead. Then, we discovered the mRNA expression of NCAPD2 was higher in various cancer types, including DLBC, LGG, OV, SARC, TGCT and THYM (Fig. 2B, all P<0.05). However, the expression of NCAPD2 in ACC, LAML and UCS was no statistical significance (Supplementary Fig. 1). Based on UALCAN portal, the protein expression of NCAPD2 was also up-regulated among the BRCA, COAD, UCEC, Clear cell RCC, LUAC, PAAD, HNSC and GBM (Fig. 2C, all P<0.05). Moreover, GSCA and GEPIA2 database showed that the expression of NCAPD2 was related with stage of the certain cancers (Fig. 2D, P<0.05). For the LUAD, ACC and TGCT, we observed a positive correlation between the NCAPD2 expression and the clinical stage (Supplementary Table 2), but not others. These data suggested that NCAPD2 was highly expressed and played an oncogene role in multiple cancers.
3.2 High NCAPD2 expression predicts poor prognosis in pan-cancer
In GEPIA2, we investigated that the OS of NCAPD2 high expression group was obviously shorter than the low one, including the ACC, LIHC, LGG, LUAD, MESO, SARC and SKCM (Fig. 3A, all P<0.05). In addition, the DFS of NCAPD2 was significantly shorter in the high expression group than in the low expression group, covering ACC, LIHC, LGG, PAAD and SARC (Fig. 3B, all P<0.05). Similarly, the expression of NCAPD2 was considerably associated with poor survival in breast cancer (OS: P=0.011, RFS: P=0.00094, DMFS: P=7.4E-06), OV (OS: P=0.0086, PFS: P=2E-04), lung cancer (OS: P=0.00053), gastric cancer (OS: P=0.013, FP: P=0.00025) and liver cancer (OS: P=3E-05, RFS: P=0.005, PFS: P=0.00079, DSS: P=8.9E-05) (Fig. 3C). These findings clearly showed that the high NCAPD2 expression was significantly associated with poor prognosis in cancer patients, which could be a important marker for all types of cancer.
3.3 NCAPD2 and related genes are involved in immune processes in pan-cancer
In order to evaluate the molecular mechanism of NCAPD2 in oncogenesis and progression, we then performed enrichment analysis. We analyzed a total of 20 NCAPD2-interacting proteins in the STRING database (Fig. 4A). In the following steps, the GEPIA2 was used to integrate the top 100 genes most similar to NCAPD2 (Supplementary Table 3). The KEGG analysis revealed that these 120 genes were primarily concentrated in ways related to the “cell cycle”, “oocyte meiosis”, “DNA replication” and “p53 signaling pathway” (Fig. 4B). The biological process (BP) analysis revealed that these genes were mainly related to “cell division”, “mitotic nuclear division”, “sister chromatid cohesion” and “hypermutation of immunoglobulin genes” (Fig. 4C). Interestingly, the somatic hypermutation of immunoglobulin genes participated in the immune response, suggesting that NCAPD2 may be related to the immune process. The findings of cellular component (CC) and molecular function (MF) were consistent with the above (Fig. 4D). In addition, the Venn diagram of the above-mentioned two groups obtained three common members, namely NCAPG, NCAPH and ASPM (Fig. 4E). NCAPG, NCAPH and ASPM have strong positive correlations with NCAPD2 expression in different cancers, especially in THYM (Fig. 4F). NCAPD2 was consistent with NCAPH, NCAPG and ASPM in the mulberry plot of data flow distribution in THYM classification with different clinicopathological parameters (Fig. 4G). Collectively, these evidences suggested that NCAPD2 might be involved in DNA repair, p53 signaling pathway, immune response and targeted ASPM protein.
3.4 Alterations of NCAPD2 gene are associated with development and progression of pan-cancer
It is widely acknowledged that genomic alteration is almost associated with tumorigenesis. In UCEC, SKCM and COAD, NCAPD2 alteration showed comparatively high mutation level, and the high amplification level of NCAPD2 in TGCT, UCS, OV (Fig. 5A). There were 153 missense, 23 truncating, 9 splice, and 3 fusion mutations between amino acids 0 and 1401. Among R1241S/C/H was the most frequent mutation site and discovered in STAD (n=1), LGG (n=1) and SKCM (n=2) (Fig. 5B). Compared with patients with NCAPD2 alterations, the LUSC patients without altered NCAPD2 had better prognosis in PFS (P=0.0242) and DFS (P=0.0144). Both ESCA (P=0.0316) and LGG (P=5.877E-3) patients without NCAPD2 alterations had longer survival than with NCAPD2 alterations (Fig. 5C). Our findings indicated that NCAPD2 expression was significantly positively associated with MSI, LOH, MATH and ploidy in most tumors (Fig. 5D). Through the intersection analysis of the above four groups, we obtained 15 cancers in which NCAPD2 was changed in different genetic heterogeneity (Fig. 5E). Consistent with these observations, NCAPD2 DNA methylation was significantly negatively correlated with NCAPD2 gene expression in most cancers (Fig. 5F). The above results indicated that the gene alteration of NCAPD2 might promote the occurrence and development of tumors.
3.5 NCAPD2 expression is associated with cancer stemness and drug sensitivity in pan-cancer
Several studies have recently indicated that cancer stemness index plays a key role in tumor pathology and shows potential drug targets for anticancer therapies. We observed a positive correlation between NCAPD2 and cancer stemness: DNAss and RNAss (Fig. 6A). Consistently, NCAPD2 expression had a significantly positive correlation with HRD score in most cancers, which evaluated as predictors of response to neoadjuvant platinum-based therapy (Fig. 6B). Our studies showed that NCAPD2 expression had significant positive associations with TMB in most cancers (Fig. 6C). According to the data of GDSC and CTRP, the relationship between NCAPD2 expression level and drug sensitivity was analyzed. Our studies found that pevonedistat, topotecan , vorinostat , navitoclax and other drugs were highly sensitive to NCAPD2. In contrast, NCAPD2 was resistant to drugs such as trametinib and selumetinib (Fig. 6D). Using CellMiner, the expression of NCAPD2 in ICG-001 and Volitinib was negatively correlated with IC50, and positively with clofarabine and gemcitabine (Fig. 6E). We speculated that these indicators may benefit from predicting immunotherapies and NCAPD2 might be a specific molecular target for the study of targeted drugs.
3.6 NCAPD2 is correlated with immune infiltration in LUAD and LUSC
On the basis of the above analysis, we speculated that NCAPD2 was related to different immune cell types in TME. The results showed that NCAPD2 was positively correlated with activated CD4+ T cells, Th2 and memory B cells (Fig. 7A). Since the pathologic types of the tumors were mostly squamous carcinoma and adenocarcinoma, we selected LUSC and LUAD as following studies. Further analysis showed that the relationship between NCAPD2 expression level and the infiltration of CAFs and CD4+T cells were positive in LUAD and LUSC, but negative in DCs (Fig. 7B). To further examine the association between NCAPD2 expression and immunomodulators, we identified that NCAPD2 expression was positive correlation linked with CD276 (also known as B7 homolog 3 protein) in LUAD and LUSC (r>0, P<0.05) (Fig. 7C). For major histocompatibility complex (MHC), there were negative correlations with NCAPD2 expression in HLA-DPB1 (r<0, P<0.05) (Fig. 7D). The above results confirmed that NCAPD2 may influence antitumor immunity by controlling TME composition, especially in LUAD and LUSC.
3.7 NCAPD2 is a potential biomarker of immunotherapy in LUAD and LUSC
IHC staining from HPA confirmed that NCAPD2 protein was up-regulated in LUAD and LUSC tissues (Fig. 8A). The diagnostic value was assessed using the ROC curve analysis in LUSC, the area under curve (AUC) was 0.897 at 1 year, 0.675 at 3 years and 0.781 at 5 years. Furthermore, the AUC in this risk signature was 0.642, 0.752 and 0.716 for predicting 1 year, 3 years and 5 years survival, respectively in LUAD (Fig. 8B). Our further study found that the expression level of NCAPD2 was closely correlated with immune subtypes in LUAD and LUCS (Fig. 8C). With the Pearson method, the NCAPD2 mRNA expression was negatively correlated with ESTIMATEScore, ImmuneScore and StromalScore in LUAD and LUSC (Fig. 8D, all P<0.05). Compared with normal somatic copy number alterations (SCNA), the tumor infiltration levels were higher in DCs with different SCNA of NCAPD2 in LUAD and LUSC, and arm level gain frequently occurred in most TIICs (Fig. 8E, P<0.05). These findings suggested that NCAPD2 could be an immune related biomarker on the efficacy of immunotherapy and a prognostic marker for LUAD and LUSC.