Expression of ZWINT in Human Pan-Cancer
We used the TIMER2.0 database to analyze the difference in mRNA expression of ZWINT between tumors and paraneoplastic normal tissues in all TCGA projects. ZWINT expression levels were significantly elevated in most tumors, including BLCA(bladder urothelial carcinoma), BRCA(breast invasive carcinoma), CHOL(cholangiocarcinoma), COAD(colon adenocarcinoma),ESCA(esophageal carcinoma), GBM,HNSC (head and neck squamous cell carcinoma), KIRC (kidney renal clear cell carcinoma),LIHC(liver hepatocellular carcinoma),LUAD,LUSC(lung squamous cell) ,PRAD(prostate adenocarcinoma),READ (rectum adenocarcinoma), STAD (stomach adenocarcinoma) and UCEC (uterine corpus endometrial carcinoma) (p<0.001),CESC(cervical squamous cell carcinoma and endo-cervical adenocarcinoma) and KIRP(kidney renal papillary cell carcinoma)(p<0.01), PCPG(pheochromocytoma and paraganglioma) and THCA (thyroid carcinoma)(p<0.05).Interestingly, ZWINT expression was lower in KICH(Kidney Chromophobe) than in normal tissues (p<0.001) (Fig.1a).
Next, for tumors without adjacent normal tissues in the TCGA project, we used GEPIA2 to assess the difference of ZWINT expression by comparing TCGA tumor tissue and the normal tissue of GTEx data. Our study found that the expression of ZWINT remained high in DLBC, LGG, OV, SKCM, THYM and UCS tumor tissues (Fig 1b, P<0.01). However, for SARC (Sarcoma) or TGCT (Testicular Germ Cell Tumors), we did not observe significant differences (Fig S1). Overall, the mRNA expression of ZWINT was elevated in most human tumors.
ZWINT expression in different pathological stages in Pan-Cancer
We used the TCGA project to analyze the expression of ZWINT at different pathological stages in pan-cancer. In BLCA, BRCA, CESC, COAD, ESCA, HNSC, KICH,KIRC,KIRP, LUAD, LUSC, READ, STAD and UCEC, ZWINT expression levels were significantly elevated in the early stages of cancer compared to normal tissues (Fig.2). Conversely, ZWINT expression decreased in the early stages of KICH. However, there was no significant difference in ZWINT expression levels between early and late pathological stages of cancers.
Prognosis and survival analysis of ZWINT in pan-cancer
We investigated the correlation between ZWINT and OS and DFS of tumor patients in pan-cancer. In ACC (P=1.5e-08), LGG (P=0.002), LIHC (P=0.00061), LUAD (P=0.00075), MESO (P=7.9e-05), PAAD (P=0.028), SARC (P=0.0078), SKCM (P=0.0074), and THYM (P= 0.0062), high expression of ZWINT predicted poorer OS in cancers (Fig 3a). While in ACC (P=1.7e-05),BLCA (P=0.022),LIHC (P=7.8e-05),MESO (P=0.037),PRAD (P=0.0092),SARC (P=0.0022),SKCM (P=0.008),TGCT (P=0.02),THCA (P= 0.00092),and UVM (P=0.0041) cancers, high ZWINT expression was associated with poor DFS prognosis (Fig 3b). Among them, high expression of ZWINT was associated with poor OS and DFS prognosis in five cancers, namely ACC, LIHC, MESO, SARC, and SKCM. The above analysis showed that high expression of ZWINT was closely associated with poor prognosis in most tumors, indicating that ZWINT may be a promising prognostic marker for pan-cancer.
Mutation Feature of ZWINT in Pan-Cancer
We used cBioPortal to analyze the genetic mutation profile of ZWINT in pan-cancer and the correlation of certain genetic alterations with the clinical survival prognosis of patients. We observed that genetic alterations of the " Amplification " type were mainly in uterine sarcoma and cholangiocarcinoma with 2-4% alteration frequency. The" mutation" type of gene alterations were the main type of alterations in gastric adenocarcinoma, endometrial carcinoma, lung adenocarcinoma and diffuse large B-cell lymphoma with 2-4% alteration frequency (Fig S2). Next, we further analyzed the type, site, and number of cases of ZWINT gene alterations. The missense mutations of ZWINT were the predominant type of gene alterations, with missense mutations detected in a total of 72 cases. In addition, 10 cases
contained splice mutations and 10 cases carried truncating mutations(Fig 4a). In addition, we analyzed the missense mutation of ZWINT in LSCC, the alteration of I66V, and labeled I66V on the ZWINT 3D protein structure (Fig 4b). We further analyzed the correlation between ZWINT gene alterations and clinical survival prognosis of cancer patients. For example, LSCC patients with ZWINT gene alterations had a poorer prognosis in terms of DFS (P= 5.597e-3), PFS (P= 7.097e-3) and OS (P= 0.0500), but not disease-specific survival (P= 0.380), compared with patients without ZWINT alterations (Fig 4c).
Correlation Analysis Between ZWINT Expression and Immune Infiltration
Tumor-infiltrating immune cells are the main component of TME and play an important role in tumorigenesis, progression, and metastasis. To explore whether ZWINT is involved in the pan-cancer immune infiltration process, we evaluated the association between ZWINT expression and tumor purity. We retrieved the association between ZWINT and immune cells infiltration such as CD8+ T cells, CD4+ T cells, B cells, neutrophils, macrophages, tumor-associated fibroblasts and monocytes in individual tumors using the TIMER2.0 database. We found that ZWINT was positively correlated with infiltration values of Myeloid-derived suppressor cells (MDSCs) and common myeloid progenitor. In contrast, ZWINT expression was negatively correlated with the infiltration values of endothelial cells, hematopoietic stem cells, and tumor-associated fibroblasts in most tumors. (Fig 5a-c). ZWINT and immune infiltration of T cell CD4+ Th2 cells were positively correlated. ZWINT and immune infiltration of T cell CD4+ central memory, T cell CD4+ effector cells were negatively correlated in most tumors, except for THYM. ZWINT expression levels were positively correlated with the infiltration values of CD8+cells in KIRC and THYM and negatively correlated with PAAD(Fig S3).
ZWINT-Related Gene Enrichment Analysis and Function Analysis
We used STRING, GEPIA2, Metascape, and CancerSEA tools to analyze the molecular mechanisms of ZWINT in tumorigenesis development and the ZWINT-related functional states of cancers. First, we selected 10 ZWINT-binding proteins from protein-protein interaction network analysis and analyzed the correlation of ZWINT with 10 proteins in all cancers of TCGA project(Fig6a). Our analysis showed that the expression of CASC5,
BUB1B, NDC80, ZW10, NSL1, DSN1, BUB1 and NUF2 were positively correlated with ZWINT in most cancer types(Fig6b). Then, we used the GEPIA2 tool to obtain the top 100 genes associated with ZWINT expression. Through intersection analysis of the abovementioned two groups, we obtained 4 co-members, namely NDC80,DSN1,BUB1 and NUF2(Fig6c). Next, we explored the correlation of ZWINT with NDC80 and DSN1 in several tumors. In GBM, the correlation between ZWINT and NDC80 was 0.879 (p=2.47e-50), and the correlation between ZWINT and DSN1 was 0.723 (p=5.01e-26). In LGG, the correlation between ZWINT and NDC80 was 0.87 (p=6.37e-160), and the correlation between ZWINT and DSN1 was 0.692 (p=1.25e-74). In LIHC, the correlation between ZWINT and NDC80 was 0.906 (p=6.33e-140), and the correlation between ZWINT and DSN1 was 0.817 (p=3.35e-90). In LUAD, the correlation between ZWINT and NDC80 was 0.853 (p=3.15e-147), and the correlation between ZWINT and DSN1 was 0.712 (p=1.12e-80) (Fig6d). Next, we used the Metascape tool to analyze ZWINT-related pathway and biological process based on ZWINT-correlated genes, which is selected from the top 100 genes and is correlated greater than 0.7 with ZWINT. Heat maps and networks colored by clusters were used to visualize the gene enrichment results. GO enrichment analysis showed that these genes above are mainly associated with biological processes such as “mitotic cell cycle”, “regulation of cell cycle process”, “regulation of mitotic nuclear division” ,“spindle elongation” and “chromosome localization” (Fig 6e).
Finally, we used single cell sequencing from CancerSEA to analyze the correlation between ZWINT and 17 cancer functional states. Our results showed that in most tumors, ZWINT was positively correlated with cell cycle, DNA damage, DNA repair and proliferation, and negatively correlated with inflammation and quiescence(Fig7a). Next, we analyzed the top three cancers correlated with cell cycle: BRCA, LUAD, and AML. Specifically, in BRCA, the four cancer functional states with the highest correlation with ZWINT were cell cycle (cor:0.77,p<0.001), DNA repair (cor:0.70,p<0.001), DNA damage (cor:0.62,p<0.001), and proliferation (cor:0.50,p< 0.001). In LUAD, in addition to the above four states, ZWINT was negatively correlated with angiogenesis (cor:-0.33,p<0.001). The functional state of tumors associated with ZWINT is mainly in DNA repair (cor:0.64,p<0.001), cell cycle (cor:0.63,p<0.001), invasion (cor:0.49,p<0.001), inflammation (cor:- 0.57,p<0.001), and angiogenesis (cor:-0.41,p<0.01)in AML(Fig7b).