We employed the TIMER (https://cistrome.shinyapps.io/timer/) (Li et al., 2017), Oncomine (https://www.oncomine.org), and GEPIA databases (http://gepia.cancer-pku.cn/) (Tang et al., 2017) to analyze the expression of UBE2C in pan-cancer, the Cancer Cell Line Encyclopedia (CCLE) tool (https://portals.broadinstitute.org/ccle/) (Ghandi et al., 2019) was employed to examine the expression of UBE2C in diverse cancer cell lines. The UALCAN tool (http://ualcan.path.uab.edu/) (Chandrashekar et al., 2017) was utilized to analyze the protein of UBE2C in diverse cancers. The expression of miRNA-140-3p was analyzed by Starbase (Li et al., 2014). The Kaplan–Meier plotter (http://kmplot.com/analysis/) (Hou et al., 2017) was employed to examine the prognosis of miRNA-140-3p in pan-cancer.

We employed the GEPIA database (http://gepia.cancer-pku.cn/) and PrognoScan database (http://dna00.bio.kyutech.ac.jp/PrognoScan/index.html) (Mizuno et al., 2009; Tang et al., 2017) to analyze the OS and RFS of UBE2C in pan-cancer; additionally, the correlation between the tumor stage and UBE2C expression was analyzed by GEPIA, and the correlation between the tumor stage and UBE2C expression was analyzed by GEPIA. Tumor stage, lymph node metastasis, and expression of miRNA-140-3p were analyzed by the UALCAN tool (http://ualcan.path.uab.edu/) (Chandrashekar et al., 2017). We also employed the prognosis tool to verify the prognosis of UBE2C in pan-cancer.

The gene mutation information of UBE2C in pan-cancer was analyzed by the cBioPortal database (https://www.cbioportal.org/) (Cerami et al., 2012).

We employed the Starbase database (http://starbase.sysu.edu.cn/) to forecast the potential miRNAs of UBE2C (Li et al., 2014), and examine the expression, prognosis, and correlation between miRNA-140-3p and UBE2C; we also used Starbase to predict the binding between miRNA, mRNA, and lncRNA.

We employed the CancerSEA database (http://biocc.hrbmu.edu.cn/CancerSEA/) to analyze the function of UBE2C in pan-cancer (Yuan et al., 2019); LinkedOmics (http://www.linkedomics.org/admin.php) was employed to analyze the KEGG pathway of UBE2C in LUAD (Vasaikar et al., 2018).

We employed the TIMER (https://cistrome.shinyapps.io/timer/) and X-Cell tools (https://xcell.ucsf.edu/) to analyze the immunological roles of UBE2C (Aran et al., 2017; Li et al., 2017), including the correlation between diverse immune cells and immune regulators. TISIDB (http://cis.hku.hk/TISIDB/) was adopted to analyze the relationship between the UBE2C expression and the 28 tumor-infiltrating lymphocytes, 45 immune stimulators, 24 immune inhibitors, 41 chemokines, 18 receptors, and 21 major histocompatibility classes (MHCs) in pan-cancer (Ru et al., 2019). The TMB and MSI scores were obtained from TCGA. A correlation analysis between the UBE2C expression and TMB or MSI was performed using Spearman’s method.

We employed the GDSC and CTRP databases to analyze the correlation between UBE2C expression and drug sensitivity (Basu et al., 2013; Yang et al., 2013).

We employed lncLocator (www.csbio.sjtu.edu.cn/bioinf/lncLocator) and CPC2 (http://cpc2.cbi.pku.edu.cn) to examine the subcellular localization and protein-coding ability of SNHG1 (Kang et al., 2017; Cao et al., 2018).

The BEAS-2B cell line was purchased from the cell bank of Kunming Institute of Zoology and cultured in BEGM media (Lonza, CC-3170). Lung cancer cell lines, including A549, H1299, and H1975, were purchased from Cobioer, Nanjing, China, with STR document; A549, H1299, and H1975 cells were all cultured in RPMI 1640 medium (Corning, Tewksbury, MA, United States) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. The siRNAs for METTL3, SNHG1, miRNA mimics NC, and miRNA-140-3p mimics were synthesized by RiboBio, and a scrambled siRNA was synthesized as a negative control. Transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions. Total RNA was collected 48 h after transfection.

The qRT-PCR assay was performed as documented (Jiang et al., 2018). The primer sequences are listed follows: UBE2C-F: GGA​TTT​CTG​CCT​TCC​CTG​AA, UBE2C-R: GAT​AGC​AGG​GCG​TGA​GGA​AC, METTL3-F: AGC​CTT​CTG​AAC​CAA​CAG​TCC, METTL3-R: CCG​ACC​TCG​AGA​GCG​AAA​T, SNHG1-QPCR-F-GCATCTCATAATCTATCCTGG, SNHG1-QPCR-R-CCTAGTTTTCCTCAAACTCCT, MiRNA-140-3p: TAC​CAC​AGG​GTA​GAA​CCA​CGG, β-actin-F: CTTCGCGGGCGACGAT, β-actin-R: CCA​TAG​GAA​TCC​TTC​TGA​CC. The expression quantification was obtained with the 2−ΔΔCt method.

Dual-luciferase assay was performed as documented (Xiong et al., 2021). Briefly, putative binding sites for miR-140-3p on the 3′-UTR of UBE2C were predicted by the Starbase database (http://starbase.sysu.edu.cn/). Wild-type and mutant (mut- UBE2C-3UTR or mut- -3UTR), wild-type, and mutant (mut- SNHG1 or mut- SNHG1) fragments were constructed and inserted downstream of the luciferase reporter gene in the reporter plasmid pGL3 (Promega). A549 and H1975 cells (4×10⁴ cells/well) were seeded in a six-well plate and co-transfected with a 3′-UTR UBE2C construct and miR-140-3p mimics or miR Ctrl using Lipofectamine 2000. Both firefly and Renilla luciferase expressions were measured post-transfection using the Dual-Luciferase Kit (Promega, Madison, WI, United States) according to the manufacturer’s instructions.

NSCLC cells were seeded in six-well plates overnight and then treated with actinomycin D (5 μg/ml, HY-17559, MedChemExpress) at 0, 2, 4, and 8 h. Total RNA was then isolated by TRIzol and analyzed by qPCR. The RNA expression for each group at the indicated time was calculated and normalized by β-actin.

Analysis of the UBE2C expression of pan-cancer was estimated using t-tests. For survival analysis, the HR and p values were calculated utilizing univariate Cox regression analysis. Kaplan-Meier analysis was employed to examine the survival time of patients stratified according to high or low level of the UBE2C expression. For all figures, ∗, ∗∗, and ∗∗∗ indicate p < 0.05, p < 0.01, and p < 0.001, respectively.

To examine the gene expression of UBE2C, we first used the TIMER database to analyze the levels of UBE2C in various cancers. UBE2C was significantly highly expressed in the following cancers: ACC, BLCA, CESC, CHOL, COAD, ESCA, GBM, HNSC, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, PAAD, PCPG, PRAD, READ, STAD, THCA, and UCEC (Figure 1A). We also examined the expression of UBE2C in human cancer based on GEO data from the Oncomine database. UBE2C was elevated in various cancers, including bladder, brain, CNS, cervical, colorectal, esophageal, gastric, head and neck, liver, lung, ovarian, pancreatic, and prostate cancers, as well as in lymphoma, melanoma, and sarcoma (Figure 1B). Owing to the lack of associated normal tissue data for some cancers in the TCGA databases, we next used the GEPIA database to explore UBE2C expression in human cancers, as shown in Figure 1C; a high expression of UBE2C was observed in the following cancers: ACC, BLCA, BRCA, CESC, COAD, DLBC, ESCA, GBM, HNSC, KICH, KIRC, KIRP, LAML, LGG, LIHC, LUAD, LUSC, OV, PRAD, READ, SKCM, STAD, TGCT, THCA, THYM, UCEC, and UCS. Furthermore, to determine UBE2C expression in diverse cancer cell lines, using the analysis of the CCLE database, we show that UBE2C expression was upregulated in various cancer cell lines (Figure 1D). Finally, in order to verify the above results, we used UALCAN database analysis to investigate the protein levels of UBE2C in human cancer. UBE2C protein was significantly elevated in breast, colon, and ovarian cancers, and in UCEC (Figure 1E). Overall, these results showed that UBE2C was upregulated in diverse human cancers.

Considering that UBE2C was highly expressed in human cancer, we further explored the correlation between UBE2C expression and the pathological stages of different cancers. The results suggested that the expression of UBE2C was significantly associated with the pathological stage in ACC, BRCA, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, PAAD, TGCT, and THCA (Figures 2A–D). Given that lymph node metastasis plays a crucial role in cancer progression, we next examined the expression of UBE2C in different stages of lymph node metastasis; the results indicated that UBE2C was positively related to the lymph node metastasis of BLCA, BRCA, CESC, CHOL, COAD, ESCA, HNSC, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, MESC, PAAD, PRAD, READ, STAD, and THCA (Supplementary Figures S1A–E). Overall, these findings showed that UBE2C expression was significantly related to clinical pathological characteristics in pan-cancer types.

To explore the prognostic role of UBE2C in pan-cancer, we performed OS and RFS analysis in human cancer; the results showed that the high expression of UBE2C was not only related to poor overall survival in ACC, BRCA, KIRC, KIRP, LGG, LIHC, LUAD, MESO, PAAD, SKCM, and UVM (Figures 3A–D and Supplementary Figures S2A–D, S3A–D) but also associated with poor disease-free survival (DFS) in ACC, BRCA, KIRC, KIRP, LGG, LIHC, MESO, PAAD, PRAD, SARC, THCA, and UCEC (Supplementary Figures S2A–D). The Cox regression analysis showed that the high expression of UBE2C was related to poor progression-free survival (PFS) in ACC, KIRC, KIRP, LGG, LIHC, LUAD, MESO, PAAD, PCPG, PRAD, SARC, THCA, UCEC, and UVM. Furthermore, the high expression of UBE2C was associated with poor DSS in CESC, KIRP, LIHC, PRAD, THCA, and UCE2C (Supplementary Figures S2A–D). We previously showed that UCE2C had a high expression and correlated with prognosis in diverse cancer types. Moreover, ROC curve analysis indicated that UCE2C had the potential to act as a biomarker for the diagnosis of diverse cancer types with high sensitivity and specificity (Figures 4A–E).

Human cancer can be divided into various immunological and molecular subtypes; thus, we also examined the UBE2C expression pattern in different cancers. The results showed that UBE2C presents different expression patterns in pan-cancer (Supplementary Figures S4A–D). For example, UBE2C had a high expression in C2 and a low expression in C3. For molecular subtypes, UBE2C displayed a unique expression pattern in cancer (Supplementary Figures S5A–D). These results show that UBE2C has different expression patterns in human cancer.

TMB refers to the number of mutations of DNA in a cancer type and has emerged as a specific and sensitive biomarker of response to immune checkpoint inhibitors (Addeo et al., 2021). We also examined the correlation between UBE2C expression and TMB of human cancer; the results showed that UBE2C expression was markedly positively correlated with the TMB in ACC, PAAD, LUAD, BRCA, PRAD, SARC, LGG, KICH, STAD, BLCA, CHOL, and LUSC and negatively correlated with the TMB in COAD and THYM (Supplementary Figure S6A). MSI represents a hyper-mutable state of DNA sequences caused by the lack of DNA mismatch repair activity (Boland and Goel, 2010). We also explored the correlation between the UBE2C expression and MSI of human cancer, and the results showed that UBE2C expression was markedly positively correlated with the MSI in SARC, MESO, UVM, and ACC and negatively correlated with the MSI in SKCM and THYM (Supplementary Figure S6B).

To elucidate the biological functions of UBE2C in various cancers, we utilized the LinkedOmics database to perform KEGG-enriched analysis for UBE2C in different cancer types. We found that high expression of UBE2C was mainly associated with the spliceosome, Fanconi anemia pathway, nucleotide excision repair, cellular senescence, and basal transcription factors in BRCA (Figure 5A); oxidative phosphorylation and biosynthesis of amino acids in CESC (Figure 5B); cell cycle checkpoint, DNA recombination, mRNA processing, and RNA modification in KIRC (Figure 5C); the regulation of chromosome organization and ncRNA processing in KIRP (Figure 5D); oocyte meiosis, RNA transport, and cellular senescence in LIHC (Figure 5E); regulation of cell cycle phase transition and DNA metabolism in LUAD (Figure 5F); DNA replication and protein export in LUSC (Figure 5G); and the Hedgehog signaling pathway and histidine metabolism in STAD (Figure 5H). These results indicated that UBE2C plays a crucial role in the development of different cancer types.

Considering that immune cell infiltration plays an indispensable role in cancer progression, we next explored the relationship between UBE2C expression and immune cell infiltration of different cancers. The TIMER analysis results show that UBE2C expression was significantly associated with the abundance of CD8⁺ T cells in 25 cancers, CD4⁺ T cells in 29 cancers, neutrophils in 29 cancers, DCs in 31 cancers, macrophages in 28 cancers, and B cells in 30 cancers (Figure 6A). To further verify the above results, we used the X-Cell tool to examine the correlation between UBE2C expression and immune cell infiltration in various cancers; the results suggested that UBE2C expression was positively correlated with 38 immune cell types in 29 cancers, while it was negatively correlated with 38 immune cell types in four cancers (Figure 6B). These findings indicated that UBE2C was significantly correlated with the infiltration of immune cells in human cancer.

In order to further determine the relationship between UBE2C and the tumor microenvironment (TME), we also examined the correlation between UBE2C and immune checkpoint-related genes by analysis of The Cancer Genome Atlas (TCGA) data. The results suggested that UBE2C expression was positively associated with the immune checkpoint-related genes in 31 cancers, which mainly included CD274, CTLA4, HAVCR2, LAG3, PDCD1, PDCD1LG2, SIGLEC15, and TIGIT (Supplementary Figure S7). The TISIDB tool analysis showed that UBE2C expression was positively associated with the 28 tumor-infiltrating lymphocytes, 45 immune stimulators, 24 immune inhibitors, 41 chemokines, 18 receptors, and 21 MHCs in different cancer types (Supplementary Figures S8A–F). These findings indicated that UBE2C plays an important role in the regulation of the immune response of human cancer.

The TIMER tool analysis shows that UBE2C expression was markedly correlated with the various immune cell gene markers, including B cells (CD19), CD8 + T cells (CD8A, CD8B), DCs (ITGAX, NRP1, CD1C, HLA-DAP1, HLA-DRA, HLA-DQB1, HLA-DPB1), M1 macrophages (PTGS2, IRF5, NOS2), M2 macrophages (MS4A4A, CD163), monocytes (CSF1R, CD86, KIR2DS4, KIR3DL3, KIR3DL2, KIR3DL1), natural killer cells (KIR2DL4,KIR2DL3, KIR2DL1), neutrophils (CCR7, ITGAM, CEACAM8), T cells (CD3D, CD3E, CD2), T cell exhaustion cells (CTLA4, LAG3,HAVCR2, GZMB, PDCD1), TAM cells (CCL2, IL10, CD68), follicular helper (Tfh) cells (BCL6, IL21), Th1 cells (TBX21, STAT4, STAT1, IFNG), Th2 cells (GATA3, STAT6, STAT5A), Th17 cells (STAT3, IL17A), and Treg cells (FOXP3, CCR8, STAT5B, TGFB1). These studies indicated that UBE2C plays a crucial role in the regulation of immune infiltrating cells of pan-cancer (Supplementary Tables S1, S2). We also found that UBE2C expression and different immune cell infiltrations could affect prognosis in different cancer types.

The above results suggested that UBE2C may act as an oncogene in cancer progression; therefore, we further explored the correlation between UBE2C expression and different drug sensitivities in different cancer cell lines from the GDSC and CTRP databases. The results indicated that the expression of UBE2C was positively correlated with the drug sensitivity of NPK76-II-72-1, vorinostat, WZ3105, YM201636, GSK690693, KIN001-102, AR-42, QL-X-138, GSK1070916, PHA-793887, TAK-715, and BX-912 and was markedly negatively correlated with the drug sensitivity of trametinib, RDEA119, selumetinib, 17-AAG, PD-0325901, docetaxel, CI-1040, AZ628, (5Z)-7-oxozeaenol, bleomycin (50 μM), and bortezomib in the GDSC database (Figure 7A; Supplementary Table S3). In the CTRP database, we observed that UBE2C expression was positively correlated with the drug sensitivity of GSK-J4, belinostat, BRD-A86708339, tivantinib, panobinostat, tacedinaline, ciclopirox, LRRK2-IN-1, PX-12, vorinostat, SR-II-138A, skepinone-L, BI-2536, PAC-1, PL-DI, and cerulenin and negatively associated with the drug sensitivity of BRD-K99006945 dasatinib, trametinib, selumetinib, PD318088, lovastatin, CAY10576, GDC-0879, simvastatin, IC-87114, and fluvastatin (Figure 7B; Supplementary Table S4). In summary, these results suggested that UBE2C was significantly related to drug sensitivity in different cancer cell lines and may be a promising therapeutic target for cancer patients.

MiRNAs play a crucial role in the regulation of mRNA expression; according to the ceRNA theory, miRNA expression should be negatively correlated with mRNA expression, as UBE2C shows high expression in pan-cancer; therefore, regulatory miRNA expression should be lower in cancer. We next utilized diverse public databases to predict the upstream miRNAs of UBE2C. We obtained four miRNAs (hsa-miR-17-5p, hsa-miR-889-3p, hsa-miR-513b-5p, and miR-140-3p) that intersected with UBE2C expression. Among these miRNAs, only miR-140-3p was significantly decreased in human cancers (Figure 8A). Next, we analyzed the expression and prognosis of miR-140-3p in pan-cancer; the results showed that miR-140-3p had a low expression in various cancers. In contrast, a high expression of miR-140-3p was observed in KIRC and LIHC. Furthermore, a lower expression of miR-140-3p was correlated with poor prognosis in KIRP, BLCA, PRAD, STAD, LUAD, KIRC, READ, THYM, and CESC (Figures 8B,C). Further studies showed that miR-140-3p expression was significantly associated with the occurrence of lymph node metastasis in different cancers (Supplementary Figures S9A–E). The ROC curve analysis showed that miR-140-3p had the potential to act as a prognostic marker in the diagnosis of different human cancers (Supplementary Figures S10A–C). We also performed a pan-cancer examination of the correlation between miR-140-3p and UBE2C; the results showed that miR-140-3p was negatively correlated with the expression of UBE2C in human cancer (Figure 9J). To verify the above hypothesis, we used NSCLC cell lines. We found that UBE2C was highly expressed and miR-140-3p downregulated in NSCLC cell lines (Figures 8D,E). We also found that UBE2C could be influenced by the expression of a miR-140-3p mimic, whereby the overexpression of miR-140-3p could significantly reduce the expression of UBE2C in NSCLC cells, including A549 and H1975 cells (Figures 8F–H). Additionally, luciferase reporter assays showed that the overexpression of miR-140-3p significantly repressed luciferase activity in NSCLC cells transfected with the wild-type (UBE2C-3UTR WT) reporter plasmid, whereas no obvious inhibition was observed in cells transfected with the mutant reporter plasmid (UBE2C-3UTR MUT) (Figures 8I,J). Consequently, these results suggested that miR-140-3p could bind to the UBE2C 3′UTR promoter region in NSCLC.

We performed a pan-cancer analysis to determine the upstream lncRNAs of miR-140-3p. We employed the Starbase and lncRNABase tools to predict potential lncRNAs that bind to miR-140-3p. We identified four possible lncRNAs, including TUG1, SNHG1, and MIR663AHG. Of these lncRNAs, only SNHG1 was significantly upregulated in various human cancers (Supplementary Figure S11A) and a high expression of SNHG1 was associated with a good prognosis in THYM, STAD, READ, OV, LUAD, HNSC, ESCA, BLCA, and LGG, but was associated with a poor prognosis in UCEC, SRAC, LIHC, KIRP, KIRC, ACC, and STAD, while the expression of other lncRNA showed no significant differences in other cancer types (Supplementary Figures S11B–E). ROC curve analysis showed that SNHG1 had the potential to act as a biomarker for the diagnosis of ACC, BLCA, CHOL, COAD, ESCA, GBM, HNSC, KIRC, KIRP, LAML, LGG, NSCLC, OV, PRAD, READ, and SKCM (Supplementary Figures S12A–D).

Based on the ceRNA theory, lncRNA should be negatively correlated with miR-140-3p and positively correlated with UBE2C expression, a condition that was fully satisfied by the expression of SNHG1. We found that SNHG1 exhibited a significant negative correlation with the expression of miR-140-3p (Figures 9A,J) and a positive correlation with the expression of UBE2C in the pan-cancer analysis. We performed the subcellular localization of SNHG1 and potential coding analysis using different public databases. The subcellular localization analysis using the lncLocator tool indicated that SNHG1 was mainly located in the cytoplasm of cells (Figure 9B). We also analyzed the coding potential of SNHG1 using a coding potential calculator, and the results showed that SNHG1 did not possess the protein-coding ability (Figure 9C).

To verify the above hypothesis, we used NSCLC cell lines to evaluate SNHG1 expression. SNHG1 was highly expressed in NSCLC cell lines (Figure 9D). Furthermore, target-binding sites of miR-140-3p on SNHG1 were predicted using Starbase (Figure 9E). In addition, luciferase reporter assays showed that the overexpression of miR-140-3p significantly repressed luciferase activity in NSCLC cells transfected with the wild-type (SNHG1-WT) reporter plasmid, whereas no obvious inhibition was observed in cells transfected with the mutant reporter plasmid (SNHG1-MUT) (Figures 9F,G). Finally, we found that UBE2C could be influenced by changes in SNHG1 expression: SNHG1 knockdown significantly reduced UBE2C expression in NSCLC cells, including A549 and H1975 cells (Figures 9H,I). Overall, these results partially demonstrate that SNHG1 may act as a potential sponge for miR-140-3p and modulates the expression of UBE2C in NSCLC cells.

Recently, emerging evidence has demonstrated that m6A modification in ncRNAs is extremely widespread and functionally modulates the eukaryotic transcriptome to influence RNA splicing, export, localization, and stability (Zhang et al., 2017; Coker et al., 2019; Wang et al., 2019; Wu et al., 2019a). To study whether m6A modifications were present on SNHG1, we first used the SRAMP database to predict potential m6A methylation sites in the transcript of SNHG1, and several m6A methylation sites in the transcript of SNHG1 were detected (Figure 10A). To determine the relationship between the m6A methylation of RNA and SNHG1 expression, we analyzed the correlation between the levels of m6A methyltransferase and m6A demethylases; the results demonstrated that SNHG1 was significantly positively or negatively correlated with the expression of METTL3, ALKBH5, and FTO, respectively (Figure 10B). A previous study also demonstrated that METTL3 could maintain the stability of lncRNA and promoted NSCLC drug resistance and metastasis (Jin et al., 2019b). We analyzed the GEO data set and found that METTL3 knockdown significantly reduced the SNHG1 expression levels (Figure 10C).

We hypothesized that m6A modification allows maintenance of SNHG1 RNA stability and was upregulated in NSCLC. To verify the above hypothesis, we used the NSCLC cell lines to perform siRNA-mediated silencing of METTL3, a core component of the m6A methylase complex (Zeng et al., 2020). Depletion of METTL3 significantly reduced the expression of SNHG1 in NSCLC (Figures 10D,E). We then examined the loss of SNHG1 RNA after blocking new RNA synthesis with actinomycin D. The results showed that SNHG1 presented reduced RNA stability after knockdown of METTL3 (Figures 10F,G). These findings raise the possibility that the modification of m6A in SNHG1 could improve the stability of the transcripts, which may partially explain the significantly increased regulation of SNHG1 in NSCLC.