Comprehensive Analysis of ANLN in Human Tumors: A Prognostic Biomarker Associated with Cancer Immunity

Background Anillin (ANLN), a ubiquitously expressed actin-binding protein, plays a critical tumor-promoting role in cell growth, migration, and cytokinesis. Numerous studies have suggested that ANLN is upregulated in many cancer types, as well as significantly associated with patient prognosis and malignant cancer characteristics. Herein, we performed an integrated pan-cancer analysis of ANLN and highlighted its underlying mechanism, which may benefit further exploration of the potential therapeutic options for cancer. Methods ANLN expression data were extracted from online databases, including TCGA, GTEx, and CCLE databases. The TIMER database was used to study the association between ANLN expression with immune checkpoint genes and immunocyte infiltration. The ScanNeo pipeline was adopted for neoantigen discovery. KEGG analysis and the STRING tool were used to elucidate the potential mechanism of ANLN in cancer development. Results ANLN is abnormally overexpressed in almost all cancer tissues compared with normal tissues. The high-ANLN expression level was positively associated with various malignant characteristics, suggesting its potential role in the immune microenvironment and poor prognosis. In addition, ANLN expression was correlated with the number of neoantigens and different phosphorylation pattern in various cancer types, revealing a functional role of genetic mutation accumulation and high phosphorylation in ANLN-mediated oncogenesis. Moreover, we found that ANLN was an important regulatory factor participating in many signaling events, especially the cell cycle and nucleocytoplasmic transport pathways. Conclusions ANLN expression is generally overexpressed in various types of cancers, and it may have an important influence on tumor progression and development. ANLN expression is significantly associated with the immune checkpoint biomarkers and tumor immunity. Together, these findings suggest that ANLN may be a predictive marker for patient prognosis across cancers.


Introduction
Due to the intricacy of tumorigenesis, an increasing number of studies have performed pan-cancer analyses of interesting genes and evaluated their relevance to the prognosis of survival and pathogenesis in cancer patients. Pan-cancer analysis was feasibly applied with the aid of many publicly available databases, such as TCGA (The Cancer Genome Atlas) and GEO (Gene Expression Omnibus), containing different datasets of gene function for diverse cancers [1][2][3].
Anillin (ANLN) is a highly evolutionarily conserved protein involved in a variety of cellular processes, involving in fitting cleavage furrows together during the entire division, especially cytokinesis [4,5]. ANLN participates in the regulation of the anaphase and telophase of the cell cycle, by connecting with myosin, RhoA, actin, and septins to help the assembly of cleavage furrow and affecting cell polarity and motor ability [6][7][8]. Existing studies have shown that ANLN is necessary for tumor growth. Our previous study revealed that knockdown of ANLN leads to the suppression of cell proliferation, migration, and invasion of hepatocellular carcinoma [9]. ANLN is highly expressed in diverse human tissues and cells [10], and it is considered an early predictor for cancer diagnosis, such as bladder urothelial carcinoma [11], colorectal cancer [12], pancreatic ductal adenocarcinoma [13], non-small cell lung cancer [14], and breast cancer [15]. Although ANLN has been suggested to be a promising new target in cancer treatment, no pan-cancer analysis of this actin-binging protein has been reported. In the present study, we provided evidence of an association between ANLN expression and various tumor types based on a combined multidatabase. The mass of data has been provided by the unprecedented bioinformatics revolution, and it can be utilized to display the whole landscape containing a variety of genes and different tumor types [16]. Pancancer analysis has been aimed at analyzing the gene function and expression, prognosis of survival, and gene mutation of tumors in many databases [17].
In this study, we utilized multiple databases to explore the function of ANLN in various human tumors by evaluating the its RNA and protein expression patterns and the associations with prognosis of survival, tumor cell immune infiltration, immune neoantigens, and immune checkpoints. We also analyzed ANLN-related gene alterations, protein phosphorylation, and signaling pathway enrichment. Through this systematic analysis, an ANLN-dominated tumorigenic molecular mechanism and immune correlation were highlighted.

Sample Information and Pan-Cancer Expression
Analysis of ANLN. The ANLN expression data in 31 normal tissues were evaluated in the Genotype-Tissue Expression (GTEx) program and obtained from the GTEx portal. The ANLN expression data in 21 tumor cell lines were investigated via the Cancer Cell Line Encyclopedia (CCLE) database (https://sites.broadinstitute.org/ccle/). We deployed the GTEx database (https://gtexportal.org/home/) and TCGA database (https://portal.gdc.cancer.gov/) to investigate the expression matrix of ANLN and the datasets of clinical information for tumor and normal tissues, respectively. The immune infiltrating cell score data of specific cancers were obtained from the Tumor IMmune Estimation Resource (TIMER) database (https://cistrome.shinyapps.io/ timer/) to explore tumor immune infiltration analysis. The present study involved 33 cancer types which are listed in Table 1. All expression data passed log2 transformation. The UALCAN portal (http://ualcan.path.uab.edu/index .html), a website including interactive datasets used to obtain cancer-omics information, was used to explore a variety of proteomic datasets of the Clinical Proteomic Tumor Analysis Consortium (CPTAC) for performing clinical proteomics oncology [18]. By inputting "ANLN," we investigated the total degree of proteomic expression and phosphor-protein expression of ANLN in cancer and normal tissues. Six available cancer datasets, including breast cancer, ovarian cancer, colon cancer, clear cell renal cell carcinoma, uterine body endometrial carcinoma, and lung adenocarcinoma, were utilized for clinical implication analysis.

Pan-Cancer Prognosis Analysis of ANLN Expression.
We utilized the Kaplan-Meier (KM) (https://kmplot.com/ analysis/) forest plots to show the association between ANLN expression and patient prognosis, including overall survival (OS), disease-free interval (DFI), and progressionfree interval (PFI). Univariate survival analysis was used to estimate the logarithmic rank (log-rank) P value, hazard ratio (HR), and 95% confidence interval (CI).

ANLN-Related Immune Infiltration and Tumor
Microenvironment (TME) Analysis. Purity-adjusted Spearman's test was performed to evaluated the associations between ANLN expression and immune cell infiltration of six cell types, including B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages, and dendritic cells. The TIMER database was also used to visualize immune and genomics data, including infiltrating immune cell scores of 33 cancer types [19]. In addition, we assessed the following microenvironment-related scores, including the ESTIMATEScore, StromalScore, and ImmuneScore, for the respective patients' tumor samples. P value < 0.05 was considered statistically significant.

Correlation Analysis of ANLN in Immune Neoantigens
and Immune Checkpoint Genes. Neoantigens are a group of molecules that are expressed on immune cells and are generated by genetic mutations, including gene fusions, deletion mutations, and point mutations. Immune checkpoints maintain their normal immune function to intervene in the level of immune activation in vivo [20]. We used ScanNeo to separately calculate the number of neoantigens on every tumor sample cell; subsequently, the association between ANLN expression and the number of antigens was presented. The relationship between ANLN expression levels and the degree of immune checkpoint markers in various cancers was analyzed through relevant modules. P < 0:05 and R > 0:20 were considered statistically significant.
2.5. Gene Enrichment Analysis of ANLN. First, we input "ANLN" in the protein name module and "Homo sapiens" in the organism module of the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) website (https:// string-db.org/). Thereafter, we set the following primary parameters: the meaning of network edges (evidence), active interaction sources (experiments and database), minimumrequired interaction score (low confidence (0.150)), and maximum number of interactors (no more than 50 interactors in the 1st shell). Finally, we obtained the available related proteins of ANLN.
We searched for the first 100 related ANLN-targeting genes across the "Similar Gene Detection" module of Gene Expression Profiling Interactive Analysis 2 (GEPIA2) (http://gepia2.cancer-pku.cn/), which contained TCGA tumor and normal tissue datasets. We then utilized the "correlation analysis" module of GEPIA2 to obtain selected genes and a pairwise Pearson correlation analysis of ANLN. We generated the dot plot through log2 TPM, and the P value and R value were calculated. In addition, we applied the "Gene_Corr" module of TIMER2 to generate the heat map data of the selected genes, including the partial correlation and P value in the purity-adjusted Spearman's rank correlation test. Jvenn, an interactive Venn diagram viewer [21], was utilized to conduct an intersection analysis to compare the relevant and connected genes of ANLN. The enriched pathways were determined via Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses. The thresholds of analysis were |NES | >1, P value < 0.05, and FDR < 0:25. We certified pathways as substantially enriched if they satisfied the subconditions.
2.6. Statistical Analysis. The Kruskal-Wallis test was utilized to investigate ANLN expression levels in various tumor tis-sues and cancer cell lines. The difference of ANLN expression levels in tumor and normal tissues were assessed by the t-test. In survival analysis, the HR and P value were computed by the univariate Cox regression method. The relationship between ANLN expression and immunocyte infiltration level was analyzed via Spearman's correlation method. Pearson's correlation analysis was utilized to evaluate the relationship between the ANLN expression and immune infiltration scores, the number of tumor neoantigens, and immune checkpoint genes. The significance threshold of all statistical analyses was P < 0:05.

Analysis of Differential Expression of ANLN in Pan-
Cancer Tissues and Normal Tissues. The present study performed a systematic analysis of the oncogenic influence of ANLN expression in human. As shown in Figure 1(a), we utilized the GTEx database to determine the expression pattern of ANLN in different tissues. We subsequently evaluated ANLN expression levels in 21 different human cancer cell lines from the CCLE database ( Figure 1(b)). The results from TCGA database displayed the discrepancy of ANLN expression levels in tumor and adjacent normal tissues in individual cancer samples, thereby revealing ANLN expression specificity (Figure 1(c)). Due to the insufficient sample size of normal tissues in TCGA, we combined the data of normal and tumor tissues of GTEx and TCGA databases to evaluate the differential expression of ANLN in 27 cancer types (Figure 1(d)). The above analyses have shown that ANLN is abnormally overexpressed in almost all cancer tissues compared with normal tissues.

Pan-Cancer Analysis of ANLN Protein Phosphorylation.
The phosphorylation-dephosphorylation cascade is regarded as an essential condition in oncogenesis. Several types of ANLN protein phosphorylation, including phosphorylation of serine and threonine residues, were identified via the CPTAC dataset. We subsequently compared the different phosphorylation features of ANLN in normal and tumor tissues by utilizing the CPTAC dataset. The ANLN phosphorylation data in various tumor types were shown in Figure 4.

Pan-Cancer Association Analysis of ANLN Expression
and Tumor Immune Infiltration. The degree of tumor immune infiltration of different immune cells is significantly associated with cancer initiation, progression, and metastasis [22]. TME composed of the external matrix, various cells, and relevant factors play an important role across cancers and affect the diagnosis, clinical treatment sensitivity, and survival of cancer patients. Therefore, we explored the association of ANLN expression with the infiltration level of immune cells in various tumors, particularly in COAD, LGG, and KIRC. We found that the ANLN expression was closely correlated with CD8 + T cells ( LGG. Additionally, the association of ANLN expression with various developmental processes in TME was analyzed. We used the estimate package in R to calculate the StromalScore and ImmuneScore in tumor samples.  Thus, ANLN expression was negatively related to the Immu-neScore in STAD but positively related to the ImmuneScore in THCA and KIRC. In addition, the StromalScores of LUSC and UCEC were negatively associated with ANLN expression, but in THCA, it showed a positive association.

Correlation Analysis of ANLN Expression with Immune
Neoantigens and Immune Checkpoints across Cancers. A series of molecules expressed in immune cells are regarded as immune checkpoints that are essential in regulating autoimmunity and can mediate the level of immune activation [23]. In tumor cells, mutated genes encode nascent antigens, which are considered cancer neoantigens. Exploiting the immune activity of tumor neoantigens promotes the synthesis of neoantigen vaccines [24]. As shown in Figure 6 3.6. Enrichment Analysis of ANLN-Related Partners. We next investigated ANLN-interacting proteins and the genes associated with ANLN expression for a series of pathway enrichment analyses to observe the molecular mechanism of ANLN in the tumor progression. Using the STRING tool, we obtained 50 proteins that interacted with ANLN and the relevant protein network was displayed in Figure 7(a). We analyzed the first 100 genes that correlated with ANLN expression in cancer expression datasets of TCGA by using the GEPIA2 tool. We showed that ANLN expression was positively associated with the DEPDC1 (R = 0:6), KIF14 (R = 0:6), KIF23 (R = 0:62), RACGAP1 (R = 0:6), and CKAP2L (R = 0:65) genes (all P < 0:001) (Figure 7(b)). In addition, ANLN was closely associated with the above five genes in various tumor types, as shown in the corresponding heat map (Figure 7(c)). A Venn diagram showed that RAC-GAP was the intersection gene of the above two datasets (Figure 7(d)). Finally, we classified the tumor samples into subgroups according to different ANLN expression levels to observe its influence on cancers. We identified the enriched biological processes and signaling pathways in these two groups by KEGG analysis (Figure 7(e)). We generated the visual pathway map based on the permutation of   Oxidative Medicine and Cellular Longevity the NES score. The pathway map showed that the cell cycle, nucleocytoplasmic transport, and Fanconi anaemia pathway were positively related to ANLN expression. In contrast, complement and coagulation cascades, pancreatic secretion, and chemical carcinogenesis-DNA adducts were significantly negatively related to ANLN expression (Figure 7(e)).

Discussion
In recent years, cancer has become a more prominent public health problem. The number of patients with cancers in the world has exceeded 90 million [25]. To better understand the underlying molecular mechanisms of cancer and to devise better treatments, many studies have performed pan-cancer analysis to detect new prognostic and diagnostic biomarkers. Previous studies had focused on the association between ANLN expression and the malignant progression of tumors, demonstrating that ANLN is upregulated in various cancers. Herein, we first presented a systematic and comprehensive analysis of ANLN in human carcinoma. We analyzed ANLN expression in 33 different cancers using public databases including TCGA, CPTAC, and GEO. We then summarized the gene expression, protein phosphorylation, and enrichment analysis of ANLN-related partners as ANLN is a unique scaffolding protein that has been demonstrated to be an important factor in cell division, and it has a close relationship with major cytoskeletal structures. ANLN is usually degraded after mitosis and remains in the nucleus during mitosis. Overexpression of ANLN may disrupt the normal regulatory mechanism by influencing the actin-myosin cytoskeleton in events other than cytokinesis, thus directly promoting cancer progression [5]. Numerous studies have implied that ANLN is overexpressed in a variety of human cancers. By analyzing ANLN expression levels using the GTEx portal, CCLE database, and TCGA database, we confirmed that ANLN was overexpressed in various cancers compared with adjacent normal tissues and we also showed that ANLN played an important role as a prognostic marker of various cancers, agreeing with those previous reports on lung cancer [26], pancreatic cancer [13], and breast cancer [15]. However, the upstream factors regulating the expression of ANLN and why ANLN is consistently upregulated in various cancer types are undefined and await further investigations.
ANLN phosphorylation status plays a vital role in cytokinesis. Previous studies have shown that ANLN is an essential cytoskeletal protein that combines the actomyosin ring with the membrane and its organizer, RhoA [5,8]. ANLN is highly phosphorylated during mitosis [27]. Furthermore, ANLN phosphorylation is positively involved in its cellular distribution and function during cytokinesis, such as meiotic division [28,29]. In the present study, we investigated several types of ANLN protein phosphorylation and compared the phosphorylation levels of ANLN between the main tumor and normal tissues using the CPTAC database. Our study showed that the phosphorylation level of ANLN was higher in some cancers, such as ovarian cancer, lung cancer,    BTLA  CD200  TNFRSF14  NRP1  LAIR1  TNFSF4  CD244  LAG3  ICOS  CD40LG  CTLA4  CD48  CD28  CD200R1  HAVCR2  ADORA2A  CD276  KIR3DL1  CD80  RDCD1  LGALS9  CD160  TNFSF14  IDO2  ICOSLG  TMIGD2  VTCN1  IDO1  PDCD1LG2  HHLA2  BTNL2  CD70  TNFSF9  TNFRSF8  CD27  TNFRSF25  VSIR  TNFRSF4  CD40  TNFRSF18  TNFSF15  TIGIT  CD274  CD86  CD44  TNFRSF9 Pearson's rho    and breast cancer. It has been reported that phosphorylation regulates the interaction between ANLN and the equatorial membrane, allowing ANLN to recruit Rho and its regulators upstream and downstream in time to ensure that the cells divide successfully [27]. Further investigations are required to determine the molecular mechanisms of the protein phosphorylation level of ANLN in tumors.
In recent years, scientists have focused on the TME. Immunity in tumorigenesis has an important influence on cancer therapy. The immune microenvironment, which is considered the "seventh marker feature" of cancer [30], includes T cells, dendritic cells, B cells, macrophages, and neutrophils. However, in the incipient stage of tumors, the TME, known as the inflammatory tumor microenvironment made up of tumor cells, inflammatory cells, and the stroma around the tumor, has a vital influence on tumor progression, angiogenesis, and genomic instability [31]. Consequently, we explored the roles that ANLN plays in the TME in human tumors. Our research revealed that the ANLN expression level was strongly associated with six infiltrating immune cells, including CD4 + T cells, dendritic cells, B cells, macrophages, CD8 + T cells, and neutrophils, in KIRC, COAD, and LGG. We further assessed the correlation between the expression levels of ANLN and TME in various cancers and found that it was negatively associated with the ImmuneScore in STAD but positively associated with the ImmuneScore in THCA and KIRC. The StromalScores in LUSC and UCEC were negatively correlated with ANLN expression levels, but the Stro-malScores in THCA were positively correlated with ANLN expression levels. These results were consistent with those of previous studies. For instance, ANLN depletion leads to familial acute respiratory distress syndrome (ARDS) [32]. In addition, IFN-γ induces upregulation of ANLN transcription in human epithelial cells [33]. Together, these results suggested that ANLN interacts with immune cells and this correlation warrants further in-depth studies. At present, little evidence of ANLN expression associated with immunity is provided in cancers. In the present study, we investigated the various roles of ANLN in modulating tumor immunology in some tumors. The expression levels of ANLN correlated with more than 40 checkpoints in various tumors, and they were positively associated with LIHC, THCA, and KIRC. By analyzing the relationship between ANLN expression and the quantity of neoantigens, we observed that ANLN expression was positively correlated with LUAD, BRCA, STAD, UCEC, and HNSC, suggesting that ANLN may enhance immunity in some tumor types. Recent findings have shown that the role of ANLN is significant in the immune reaction via association between ANLN and KDR, which may serve as a potential predictor of breast cancer survival [34].
To analyze the molecular mechanism of the ANLN in the tumor progression, we further explored the function of ANLN in cancers by analyzing the enrichment of ANLNrelevant genes and proteins. Indeed, published works by others have reported expressional dysregulation of many signaling pathways after depletion of ANLN expression in breast, pancreatic, and bladder cancer cell lines [11,13,35]. For example, molecular pathways related to epithelial and epidermal cell differentiation are activated after ANLN knockout in breast cancer cells [35]; ECM adhesion and actin cytoskeleton pathways are dysregulated in pancreatic cells, which is consistent with the decreased cell motility following ANLN depletion in these cells [13]; changes of other noncanonical pathways like Toll-like receptor and cytokine signaling are also observed in bladder cancer cells [11]. Different from the abovementioned studies, our analyses showed that the cell cycle, nucleocytoplasmic transport, and Fanconi anaemia pathways were significantly positively correlated with ANLN expression. In addition, ANLN expression was strongly associated with that of DEPDC, KIF14, KIF23, RACGAP1, and CKAP2L genes across cancers. Previous reports indicated that ANLN could promote tumor progression through its well-studied function as a scaffold protein in regulating cytokinesis. However, recent studies have provided novel evidence that ANLN may serve as a nuclear regulator for transcriptional reprogramming and control intercellular adhesion and cellular motility to alter the phenotypic plasticity of cancer cells, thus exerting its tumor-promoting functions [36,37]. In line with the above description, our result also demonstrated that RAC-GAP1 was a highly relevant gene to ANLN expression across cancers by the intersection analysis of STRING and GEPIA2 datasets. RACGAP1 has been identified as a prominent activator of tumor invasion and metastasis in various cancers, including uterine carcinosarcoma [38], breast cancer [39], and ovarian cancer [40]. It could be speculated that ANLN induces the motility of cancer cells through regulating RAC-GAP1 expression in an undefined manner. Similar to the novel regulator function of ANLN in cell motility, recent studies also revealed the emerging role of ANLN in regulating cancer stemness, a phenotypic hallmark which shared many characteristics with cellular metastasis. In breast cancer, ANLN upregulation markedly enhances the selfrenewal potential of epithelial-type MCF10AneoT cells, whereas loss of ANLN decreases stem/progenitor properties of mesenchymal-type MDA-MB-231 and BT549 breast cancer cells [35]. In addition, ANLN was used as a biomarker to monitor the late dividing retinal precursors and stem cells when fused with the eGFP reporter in an in vivo zebrafish model [41]. Though the underlying mechanism remains unknown, ANLN-regulating cancer stemness could be associated with expressional regulation of a stem cell-specific transcriptional network [35,42]. Since ANLN expression was highly correlated with the nucleocytoplasmic transport pathway, the activity of many transcriptional regulators related to stemness thus could possibly be altered under ANLN differential expression. However, further investigations are needed to verify the hypothesis.

Conclusion
In conclusion, the present study demonstrated that ANLN expression is closely related to poor prognosis in various cancers. The high ANLN expression level and elevated phosphorylation of ANLN may be potential factors for tumor progression. ANLN expression is significantly associated with the immune checkpoint biomarkers and tumor immunity. Together, these findings suggest that ANLN may be a predictive marker for patient prognosis across cancers.