AURKA is a prognostic potential therapeutic target in skin cutaneous melanoma modulating the tumor microenvironment, apoptosis, and hypoxia

AURKA, Aurora kinase A encoding gene, is an important signaling hub gene for mitosis. In recent years, AURKA has been implicated in the occurrence and development of several cancers. However, its relationship with the tumor microenvironment in skin cutaneous melanoma (SKCM) and the molecular mechanisms underlying its effects are still unclear. We adopted a variety of bioinformatics methods to comprehensively analyze the potential carcinogenesis of AURKA in SKCM, and constructed a prognostic nomogram model. We also dentified an inhibitor targeting AURKA and verified its therapeutic effects against SKCM using the molecular docking technology. We found that abnormally high expression of AURKA was responsible for driving the occurrence and development of SKCM, and affected various pathological factors in SKCM. In addition, AURKA was established as an independent marker of poor SKCM prognosis. We also characterized the potential mechanisms by which AURKA manifests its effects in SKCM and found that AURKA inhibits the infiltration of CD8+ T cells and promotes hypoxia by activating the TGF-β signaling pathway. At the same time, the high AURKA expression group had higher tumor stemness index and promoted cell proliferation and metastasis. Finally, the small-molecule compound ZNC97018978 targeting AURKA screened by molecular docking technology can inhibit the proliferation, invasion and metastasis of SKCM. The possible mechanism is that ZNC97018978 induces apoptosis by arresting the cell cycle, thereby inhibiting cell proliferation. AURKA is the core hub gene driving the occurrence and development of SKCM, and its expression is regulated by epigenetic modifications. AURKA can regulate the infiltration level of various immune cells in the tumor microenvironment, reshape the immunosuppressive tumor microenvironment, and apoptosis, and hypoxia. Thus, it is a prognostic biomarker and potential therapeutic target in SKCM. ZNC97018978 is an effective and safe inhibitor of AURKA in vitro; its safety and effectiveness in vivo as a potential treatment for cutaneous melanoma should be further determined.


Introduction
Skin cutaneous melanoma (SKCM) accounts for only 2% of skin cancers, yet, it is one of the deadliest skin cancers, accounting for more than 80% of skin cancer-related deaths in the world (Schadendorf et al. 2018). It is also one of the most malignant tumors endangering human health (Linares et al. 2015). In recent decades, the morbidity and mortality of SKCM have gradually increased (Siegel and Miller 2019). Although immunotherapy and targeted therapy have improved the prognosis of patients with advanced melanoma, only a few patients with SKCM can benefit from it, and there is no specific treatment for patients with metastatic melanoma (Flaherty et al. 2012;Bajor et al. 2018). Thus, it is becoming an urgent to identify new therapeutic targets and develop more effective therapeutics against SKCM.
Tumor microenvironment (TME) refers to the complex dynamic ecosystem on which tumor cells depend for survival and development, including surrounding blood vessels, immune cells, fibroblasts, bone marrow-derived inflammatory cells, various signaling molecules, and the extracellular matrix (ECM) (Hanahan and Weinberg 2011;Binnewies 1 3 et al. 2018). Immune cells and their regulation in tumor microenvironment play an important role in the occurrence and development (Pandey et al. 2022). Several studies have confirmed that the deterioration of malignant melanoma is closely related to alterations in the immune microenvironment, but no study has determined the immune cells associated with the metastasis of malignant melanoma (Kalaora et al. 2022). As for the core hub genes and TME, it is even less reported that the influence of the transfer system constitutes the interaction system.
In current study, we aimed to explore potential therapeutic targets and their biological functions in SKCM. AURKA has been identified as the core hub gene of SKCM. Studies have shown that AURKA as an oncogene is involved in the regulation of multiple key steps in the invasion and metastasis of malignant tumors (D'Assoro et al. 2014;Yan et al. 2016). It plays a key role in regulating tumor cell mobility, and is a potential therapeutic target for preventing tumor metastasis in bladder cancer , endometrial cancer (Umene et al. 2015) triple-negative breast cancer (Xu et al. 2013), nasopharyngeal carcinoma (Liu et al. 2012), and urothelial carcinoma (Scarpini et al. 2012); it also acts as a biomarker for the diagnosis, treatment, and prognosis of the aforementioned cancers. Therefore, in this study, we investigated the role of AURKA in SKCM, and characterized the mechanism by which it manifests its effect in SKCM. In addition, recent studies have shown that AURKA is overexpressed in SKCM tissues and is related to the invasion and metastasis of SKCM. Downregulation of AURKA can inhibit the proliferation and migration of SKCM cells and promote CM apoptosis and autophagy (Caputo et al. 2014;Puig-Butille et al. 2017). Our current study revealed that abnormally high expression of AURKA promotes SKCM invasion and metastasis in vitro and mediates the formation of an immunosuppressive microenvironment and regulates ferroptosis and hypoxia. Thus, it has several roles in SKCM and may be a prognostic biomarker and potential therapeutic target for SKCM patients. ZNC97018978 was found to be an effective inhibitor of AURKA by virtual drug screening. In vitro tests revealed that ZNC97018978 is a safe and effective drug for the treatment of cutaneous melanoma. Our findings provide an in-depth understanding of the function of AURKA in the occurrence and development of SKCM and unveil the role of AURKA in tumor immunotherapy.

Collection of clinical specimens
Thirteen pairs of human melanoma tissues and paracancerous tissues were obtained from the tissue specimen bank of the Second Affiliated Hospital of Lanzhou University. The living tissue was collected in situ immediately after ex vivo, and the paraffin-embedded formalin-fixed SKCM tissue and adjacent paracancerous tissue were subjected to immunohistochemical (IHC) staining to verify the expression of AURKA protein in SKCM tissue. This study complied with the Declaration of Helsinki protocol and was approved by the Research Ethics Committee of the Second Hospital of Lanzhou University (No. 2022A-065). All participants provided written informed consent, and the patients in this study did not receive any antitumor therapy before surgery.

Immune correlation analysis
The correlation between AURKA expression and the infiltration abundance of various immune cells in SKCM patients was determined using the TIMER algorithm (Li et al. 2020), MCP-counter algorithm (Becht et al. 2016) xCell algorithm (Aran et al. 2017), and ssGSEA algorithm (Bindea et al. 2013) of the Immuneeconv R software package (Sturm et al. 2019) The xCell algorithm (Aran et al. 2017) explored immune scoring. Subsequently, based on the ESTIMATE algorithm (Yoshihara et al. 2013) and the xCell algorithm (Aran et al. 2017), the correlations between immune score, stroma score, tumor purity score in TME, and AURKA expression in SKCM patients were analyzed. The TIMER2.0 (Li et al. 2020) (http:// timer. cistr ome. org/) online database was also used to explore the correlation of AURKA expression with the infiltration levels of four types of immunosuppressive cells, namely, CAFs, MDSCs, Tregs, M2-TAMs, to determine the effect of the infiltration abundance of immune cells on the prognosis of SKCM patients. Finally, we also analyzed the correlation between AURKA and immune markers and MHC molecules in several immune cells.

Differentially expressed gene (DEG) acquisition and functional enrichment analysis
DEGs were analyzed using the Limma package in R language. The threshold for DEG screening was set as "adj p < 0.05 and |Log2(FC)|< 1". To gain an in-depth understanding of the potential biological functions of AURKA in SKCM, we used the R language ClusterProfiler package to perform gene set enrichment analysis (GSEA). Gene set enrichment analysis (GSVA) was performed using the R software GSVA package.

Tumor stemness analysis
We used the OCLR algorithm constructed by Malta et al. to calculate the mRNA expression-based stemness (mRNAsi) index (Malta et al. 2018) and compared the mRNAsi of different groups.

Identification of the SKCM HUB gene
Based on the SKCM dataset of the cancer genome atlas (TCGA) and the GSE3189 and GSE15605 datasets of the gene expression omnibus (GEO) database, the DEGs in the tissues of SKCM patients and normal patients were analyzed, and 10 HUB genes were identified using the cyto-Hubba plugin in Cytoscape (version 3.7.2) (Doncheva and Morris 2019) The association of these ten genes with mutation, survival, and diagnosis was further analyzed.

Comprehensive analysis of AURKA in cutaneous melanoma
The data from CCLE, TCGA, GTEx, GEO and HPA (https:// www. prote inatl as. org) databases were used to analyze expression of AURKA at mRNA and protein levels in cutaneous melanoma. TCGA-SKCM data were applied to study the relationship between AURKA mRNA level and clinical characteristics of cutaneous melanoma. R package "maftools" and online databases muTarget (http:// www. mutar get. com/), cbiportal (https:// www. cBioP ortal. org/) and MethSurv 2017 (https:// biit. cs. ut. ee/ MethS urv/) were adopted and visited to analyze the mutation and methylation profiles of AURKA in cutaneous melanoma. Kaplan-Meier method was used to analyze the survival significance of AURKA in overall survival (OS) and disease-specific survival (DSS) of cutaneous melanoma patients based on TCGA-SKCM and GSE22155 data. Multivariate Cox regression analysis was performed to explore clinical factors that potentially affected the survival of cutaneous melanoma patients based on TCGA-SKCM data (Nagy and Győrffy 2021;Modhukur et al. 2018). In addition, we also performed immune correlation, functional enrichment, tumor stemness score, and hypoxia correlation analysis.

AURKA expression analysis
To further study the expression characteristics of AURKA in SKCM, we used the R software ConsensusClusterPlus package to perform consistent clustering analysis on the DEGs between the AURKA-high and -low expression groups. We also conducted survival analysis, computed immune scores, and immune examinations based on the clustering results. Furthermore, we determined point-related gene expression levels, tumor stemness index scores, and performed functional enrichment analysis.

Screening of small-molecule inhibitors for AURKA
Crystal structure of AURKA protein was retrieved from the PDB database (https:// www. rcsb. org/) and optimized using the Prime module of the Schrodinger20152 software to obtain the three-dimensional structure of biological activity. The SiteMap module was used to predict small-molecule ligand binding sites on AURKA, and the Virtual Screening Workflow tool of the ZINC database (www. zinc15. docki ng. org) was applied to screen out small-molecule ligands.

Flow cytometry
A875 cells (1.0 × 10 5 cells/mL) were cultured in a flask. The culture medium was aspirated once the cells were attached. ZNC97018978 of different concentrations was then added. After 48 h, the cells were collected, washed once with precooled PBS, and resuspended. The suspension (1 mL) was centrifuged at 1,000 rpm for 5 min. The supernatant was discarded. The remaining fraction was fixed with pre-cooled 70% ethanol, washed with PBS and centrifuged. Then, the cells were suspended by 100 μL RNase A solution, followed by water bath at 37 ℃ for 30 min. Propidium iodide (PI) staining buffer (400 μL) was finally added to stain the cells at 4 ℃ away from light for 30 min. Cell cycle was then assessed by flow cytometry.

Transwell assay
Logarithmically growing cells were harvested and prepared into a single-cell suspension. The cell suspension was added to the upper chamber of a Transwell chamber pre-coated with Matrigel (BD, 356234), and the DMEM with 10% FBS was added to the lower chamber. After routine culture for 48 h, cells that invaded to the lower chamber were collected, fixed with 4% paraformaldehyde for 30 min, stained with Giemsa Staining Solution (Solarbio, G1015) for 30 min, and finally observed microscopically.

Cell transfection
One day before siRNA knockdown target gene transfection, 0.5-2 xio 5 cells were seeded in a petri dish, and 500 dishes of medium without antibody and serum were added to make the cell density reach 50% during transfection. Take an appropriate amount of Lipo3000 and serum-free DMEM, and stir at room temperature for 5 min. Take appropriate amount of siRNA and serum-free DMEM and mix well. Gently mix the two mixtures above (to avoid fragmentation in the RNA) and let stand at room temperature for 20 min to form a fluorescent mixture. Add the mixed solution to the well (about 400, 1) of the medium containing the cells and mix gently. Incubate for 4-6 h in a 37 °C incubator, replacing the medium with complete serum containing medium. After 6 h of transfection, the transfection efficiency can be detected: scratch assay.
At first, a "十 (ten in Chinese)" marker was made on the backside of a six-well plate. Two lines (spaced by 0.5-1 cm) were drawn in each well. Logarithmically growing A875 cells were seeded into the six-well plate at (5-6) × 10 5 cells/ well overnight until 100% coverage. Subsequently, two intersecting lines parallel to the existing lines were made using a 10 µL peptide. PBS washing was performed to remove floating cells. Serum-free medium was added, and the cells were then cultured at 37 ℃ with 5% CO 2 in an incubator. The results at 0, 6, 12 and 24 h were observed by electron microscopy and images were captured.

Immunohistochemistry (IHC)
The xenograft tumors were paraffin-embedded, sectioned, deparaffinized and hydrated to water. PBS washing was performed for three times (5 min per). Antigen retrieval was obtained via heating with sodium citrate. 3% H 2 O 2 was added at room temperature to block endogenous peroxidase activity, followed by addition of normal goat serum. Primary antibodies were added overnight at 4 ℃. On the following day, secondary antibodies were added for 2 h at room temperature. Finally, the specimens were successively treated with DAB, hematoxylin, gradient alcohol and xylene. Microscopic observation was then performed. The results were assessed and scored by two pathologists who were blinded to the sample source. The staining intensity was analyzed using the semi-quantitative integration method.

Statistical analysis
After log2 transformation, all gene expression data were normalized. Wilcox tests were used to compare the differences between the two groups, and for Kaplan-Meier curves, p values and hazard ratios (HR) with 95% confidence intervals (CI) were obtained by logrank test and univariate Cox proportional hazards regression. Univariate and multivariate cox regression analysis results Forest maps were implemented by the &quot;forestplot&quot; R package, and Spearman or Pearson correlation analysis was performed between the two variables. The above results were achieved with the R (v4.0.3) packages ggplot2 and pheatmap. p < 0.05 was considered significant.

AURKA is an important core hub gene of SKCM
We identified the top ten hub genes associated with SKCM ( Fig. 1A-E, G) and found that their mutation frequency was low (Fig. 1F, Supplementary Fig. 1A), however, missense mutations were the main polymorphisms and were mainly expressed as C > T and C > A mutation ( Supplementary  Fig. 1B, C). Copy number variation (CNV) occurred in all ten hub genes, except in AURKA and UBE2C, which were Hete. Amp, and the remaining eight hub genes showed Hete. Amp and Hete. Del ( Supplementary Fig. 1E). Correlation analysis revealed that of the ten hub genes, KIF20A mRNA H, I Univariate Cox regression analysis of the overall survival and disease-specific survival of the top ten hub genes. J Correlation analysis of the top ten hub genes expression level was negatively correlated with CNV, while the mRNA expression levels of the remaining nine genes were positively correlated with CNV, of which CDKN3 and AURKA had the strongest correlation ( Supplementary  Fig. 1D). At the same time, univariate COX regression and KM analysis revealed that AURKA affected the prognosis of SKCM patients (Fig. 1H, I, Supplementary Fig. 1F-J). The ROC diagnostic curve found that the AURKA and UBF2C AUC values were the highest at 0.9 and above (Supplementary Fig. 2), Correlation analysis found that the correlation coefficients between AURKA and the other nine hub genes were all above 0.7 except for PBK (0.65) and KIF20A (0.69) (Fig. 1J). In conclusion, analysis of the role of the top ten hub genes in SKCM revealed that AURKA mainly affected the progression and deterioration of SKCM patients.

Clinical significance of AURKA
To precisely understand the role of AURKA in the development of SKCM, we analyzed the expression, clinicopathological characteristics, prognosis, mutation, copy number variation, and methylation of AURKA in SKCM, and found that AURKA mRNA and protein expression was high in the SKCM cell lines of the CCLE database, GTEx database, SKCM-TCGA database, as well as the GSE3189 and GSE46517 datasets of the GEO database and those of the HPA database (p < 0.05) ( Fig. 2A-F). The mRNA expression of AURKA was investigated in 33 tumors in TCGA and revealed that the mRNA level of AURKA in the other 31 tumors was higher than that in the corresponding normal tissues except for TGCT and PCPG Fig. 2I. Further IHC analysis revealed that AURKA expression in SKCM tissues was significantly higher than that in adjacent normal tissues (Fig. 2G, H), and high AURKA expression levels were significantly correlated with T stage, radiation therapy, disease-specific survival (DSS), PFI event, melanoma ulceration, overall survival (OS), and gender in the TCGA-SKCM dataset. (Fig. 3L). In addition, KM analysis revealed that AURKA, which was highly expressed in both the GSE22155 dataset and the TCGA-SKCM dataset, was a risk factor associated with a poor prognosis in SKCM patients ( Fig. 2J-L). Furthermore, high expression of AURKA is a risk factor associated with a poor prognosis with multiple clinicopathological factors in SKCM patients ( Fig. 2N-O). Multivariate Cox regression analysis showed that AURKA expression, T stage, N stage, and pathological stage were independent factors affecting the prognosis of SKCM (Fig. 2M). Further analysis revealed that missense mutations were most prevalent in AURKA, with a low mutation frequency of 0.43%. In SKCM patients, the most significant mutation in the AURKA-high and -low expression groups was TTN mutation with a mutation frequency of 72%. The TP53 mutation most affected AURKA expression (Fig. 3A, B, D).
However, AURKA mutation did not affect SKCM prognosis and level of immune cell infiltration in SKCM patients (Fig. 3C, E). However, there was a significant positive correlation between the expression of AURKA and copy number variation (p < 0.05), and AURKA expression affected the level of immune cell infiltration ( Supplementary Fig. 1D,  E, Fig. 3F). AURKA expression was significantly associated with the methylation level of AURKA was negatively correlated (Fig. 3G) and affected the prognosis of patients (Fig. 3H) in both SKCM and the other 38 tumors. AURKA expression was strongly positively correlated with three key enzymes regulating DNA methylation, namely DNMT1, DNMT3A, and DNMT3B0 (Fig. 3I, J). These results suggest that AURKA may mediate tumorigenesis by regulating the methylation status in SKCM. Taken together, these findings strongly suggest that AURKA is an oncogene closely associated with the progression and metastatic deterioration of SKCM and thus, serves as a potential therapeutic target and prognostic biomarker of SKCM.

Identification of SKCM molecular typing based on AURKA expression
A total of 54 DEGs, including 52 upregulated genes and 2 downregulated genes, were identified in AURKA-high and -low expression groups (Fig. 4C). According to the CDF curve, the relative change of the CDF Delta area curve, and the heat map of the consistent clustering results, the optimal number of clusters was determined to be 2 (K value = 2) (Fig. 4A, B). Therefore, all SKCM patients were divided into 2 groups, namely, AURKA type I (274 cases, 58.0%) and AURKA type II (196 cases, 42.0%) (Fig. 4B). KM survival analysis showed that compared with the AURKA type II group, AURKA type I patients had significantly improved prognosis, OS, DSS, and disease-free survival (Fig. 4J, K). The tumor stemness index was higher in the AURKA type II than that of the AURKA type I group (Fig. 4I).

AURKA function annotation
To further study the functional role of AURKA in SKCM, we first analyzed the skin melanoma single-cell sequencing datasets GSE72056 and GSE81383 based on the can-cerSEA online database (Yuan et al. 2019) and found that the expression level of AURKA was correlated with the cell cycle, proliferation, invasion, EMT and DNA damage (Fig. 4D, E). Abnormally high expression of AURKA can promote the malignant development of melanoma. We then performed GSEA enrichment analysis on patients from TCGA skin melanoma AURKA-high and -low expression groups. It was found that the cell cycle, DNA damage and repair, mitosis, glycolysis, matrix, and oncogenic activation pathways such as the p53 and TGF-β signaling pathways, were significantly positively correlated with the AURKA-high expression group. It was negatively correlated with immune activation signaling pathways such as DC cells and MHC molecules (Fig. 4G). Therefore, AURKA may be related to the occurrence and development of tumors. Analysis of the GASElite online database  (Liu et al. 2018) revealed 54 AURKA-related genes mainly related to apoptosis, cell cycle, DNA damage and repair, and EMT activation (Fig. 4F). Further GSEA functional enrichment analysis to explore potential mechanisms between the two subtypes of AURKA revealed that the AURKA type II group was significantly enriched in the cell cycle, DNA damage, matrix, and oncogenic activation pathways such as p53 and TGF-β signaling pathways, indicating that AURKA and its related genes may be related to tumorigenesis. The AURKA type I was significantly enriched in immune activation-related pathways, including MHC molecules, DC cells, TCL cells, and IL-2 activation signaling pathways, further suggesting that AURKA promotes the occurrence and development of skin melanoma (Fig. 4H). The effect of AURKA on melanoma was further assessed by analyzing the effect of AURKA expression on the function of melanoma cells using the small-molecule compound ZNC97018978 targeting AURKA. The results revealed that a decrease in the expression level of AURKA in the A875 melanoma cell line also significantly reduced the proliferation and migration ability of the cells (Fig. 9B-D; Transwell experiment) ( Fig. 9E; Scratch test). This corroborates that AURKA may be involved in the cell cycle, proliferation, apoptosis, and immune regulation during the occurrence, development, and metastasis of SKCM.

AURKA reshapes the tumor immunosuppressive microenvironment
Recent studies have demonstrated that AURKA, which is highly expressed in SKCM, inhibits effector T cell-mediated cytotoxicity (Punt et al. 2021). At the same time, the functional annotation results indicated that AURKA influences the TME, therefore, we further investigated the correlation between AURKA and TME. Analysis of the TMER2.0 online database revealed that the increased infiltration levels of CD8+, CD4 + T cells, B cells, and DC cells in SKCM patients improved the prognosis of patients, and the increased infiltration level of MDSCs was an adverse risk factor for SKCM patients (Supplementary Fig. 3A-P). It is H Functional enrichment analysis between AURKA type I and AURKA type II group patients. I Tumor stemness index score between the AURKA type I and AURKA type II group patients. J-K Kaplan-Meier curve showing survival between the AURKA type I and AURKA type II groups suggested that the infiltration abundance of immune cells in TME is closely related to tumor prognosis. Therefore, the differences in the infiltration abundance of various immune cells in patients in the AURKA-high and -low expression groups and the AURKA types I and II groups were analyzed. The results showed that DC cells, CD8+, CD+ 4, B cells, and other anti-tumor immune cells were significantly enriched in the AURKA-low expression and AURKA type I groups, while immunosuppressive cells were significantly enriched in with the AURKA-high expression and AURKA type II groups (Figs. 5H, 6G), confirming that AURKA has an immune-suppressing effect. Using the ESTIMATE algorithm and the Xcell algorithm, we determined that AURKA was negatively correlated with all three scoring forms, and Immune correlation analysis of AURKA in SKCM. A-D Analysis of the correlation between AURKA expression and immune cell infiltration abundance using the TIME, MCP-counter, xCell, and ssGSEA algorithms. E Correlation of AURKA with immune cell marker genes. F Analysis of the correlation between AURKA expres-sion and stromal score, immune score, and ESTIMATE score using the ESTIMATE algorithm. G The correlation between AURKA and the infiltration level of four kinds of immunosuppressive cells in SKCM patients. H Differences between immune-infiltrating cells between the AURKA type I and AURKA type II group patients Fig. 6 Correlation between AURKA expression and CD8+ T cell infiltration level and T cell characteristic genes, MHC molecular analysis. A, B The MHC signaling pathway was significantly enriched in the AURKA-low expression and AURKA type I groups. C, D GSEA analysis of the correlation between AURKA expression and CD8+ T cell infiltration level. E, F Molecular correlation analysis of AURKA and MHC. F, G Differential expression of MHC molecules in the AURKA-high and -low expression groups and between the AURKA types I and II groups. G Immune scores in the AURKA-high and -low groups. H, I Differential expression of MHC molecules in the AURKA-high and -low expression groups and between the AURKA types I and II groups. J GASElite database analysis of the correlation between AURKA expression and CD8+ T cell infiltration level. K TIMER2.0 database analysis of the correlation between AURKA expression and CD8+ T cell infiltration level. L, M Difference in the expression of T cell characteristic genes between the AURKA-high and -low expression groups and AURKA type I and type II groups. N Correlation between AURKA expression and the expression of T cell signature genes positively correlated with tumor purity (Fig. 5F). Analysis using the Xcell algorithm revealed that the immune score in the AURKA-low expression and AURKA type I groups were higher than that in the AURKA-high expression and AURKA type II groups (Figs. 5H, 6G), which further confirmed the tumor-promoting function of AURKA in TME. Finally, three algorithms in the Immuneeconv package of R language combined with the ssGSEA algorithm and the online database TMER2.0 were used to further verify the relationship between AURKA and various tumor immune cells; we found that NK, T, DC, M1, activated CD4 T cells, activated CD8 T cells, and effector memory CD8 T cells were inversely correlated with the infiltration abundance of tumor-promoting immune cells (MODS, Treg, CAF, Th2, M2) Abundance was positively correlated (Fig. 5A-D, G). At the same time, anti-tumor immune cell marker genes was negatively correlated (Fig. 5E), further supporting the role of AURKA in promoting tumor immune escape.

AURKA inhibits CD8+ T cell infiltration by activating the TGF-β signaling pathway
As a key component of adaptive immunity, CD8+ T cells have become the crux of immunotherapy in recent years owing to their specificity for antigen recognition and powerful tumor killing (Zhang and Zhang 2020). AURKA was found to be inversely correlated with CD8+ T cell infiltration based on GSEA enrichment analysis (Fig. 6C, D), indicating that it is immunosuppressive. To verify the results of GSEA enrichment analysis, we used the TMER2.0 and GASElite online databases for further analysis and found that AURKA expression was negatively correlated with CD8+ T cell infiltration (Fig. 6J, K), confirming that AURKA promotes the formation of immunosuppressive TME. We next explored the correlation and differential expression of T cell signature genes and AURKA and found that multiple T cell signature genes in patients in the AURKA-high expression and AURKA type II groups were significantly lower than those in the AURKA-low expression and AURKA type I groups, and multiple T cell signature genes were found to be negatively correlated with AURKA ( Fig. 6L-N). In addition, AURKA was negatively correlated with antigen presentation and processing based on GSEA enrichment and correlation analysis (Fig. 6A, B), MHC-related genes were negatively correlated and significantly expressed in the AURKA-low expression group and AURKA type I groups (Fig. 6E, F, H,  I),which further confirmed that AURKA can promote the formation of immunosuppressive TME.
We then used GSEA to explore the molecular mechanism of AURKA affecting T cell infiltration and found that genes related to the TGF-β signaling pathway were highly expressed in AURKA and significantly enriched in AURKA type II patients (Fig. 7F-I); AURKA and TGF-β signaling is positively correlated. In addition, to verify the effect of AURKA on the TGF-β signaling pathway, we conducted a correlation analysis between AURKA and important genes of the TGF-β signaling pathway (TGFβR1, ACVR2B, SMAD2, SMAD4, ACVR1C, TGFβ1), and found that AURKA was positively correlated with important genes of TGF-β signaling pathway (Fig. 7C, D), and was highly expressed in AURKA type I and AURKA type II patients (Fig. 7A, B), suggesting that AURKA inhibited CD8+ by activating the TGF-β signaling pathway and T cell infiltration. To further verify this phenomenon, the expression of six important genes in the TGF-β signaling pathway was investigated using small interfering RNA (siRNA) under the condition of low AURKA expression. The expression of five important genes of the TGF-β signaling pathway (TGFβR1, SMAD2, SMAD4, ACVR1C, TGFβ1) was decreased under low AURKA expression (Fig. 7J), confirming that AURKA activates the TGF-β signaling pathway to inhibit T cell infiltration.

AURKA participates in the regulation of hypoxia
The functional enrichment analysis of GSEA and GSVA revealed that hypoxia-related signaling pathways were significantly enriched in the AURKA-high expression group (Fig. 8A, B). In recent years, hypoxia has been found to play an important role in the tumor microenvironment (Scharping et al. 2021). Therefore, we further explored the correlation between AURKA and hypoxia, as well as between hypoxia and the tumor microenvironment. We found that AURKA was positively correlated with 14 hypoxia-related genes, especially in the AURKA-high expression group and AURKA subtype II; CDKN3 was highly correlated with AURKA (Fig. 8C, E-G). Correlation analysis of 15 hypoxia-related genes and AURKA in pan-cancer revealed that many hypoxia-related key genes were positively correlated with AURKA in various tumors, especially CDKN3 was significantly positively correlated with AURKA in 40 tumors (Fig. 8D). It was revealed that the abnormally high expression of AURKA mediates hypoxia. Correlation analysis of immune cell infiltration levels in TME of 15 hypoxia-related genes based on the TMER and EPIC algorithms revealed that hypoxia characteristic genes played a role in inhibiting immune cell infiltration (Fig. 8H, I), suggesting that hypoxia remodels the tumor immunosuppressive microenvironment and promotes tumor immune escape. To further verify these findings, we analyzed the correlation between hypoxia signature genes and T cell signature genes and MHC molecules and found that hypoxia signature genes were negatively correlated with T cell signature genes and MHC molecules ( Fig. 8J, K). Therefore, hypoxia remodels the tumor immunosuppressive microenvironment and promotes tumor immune escape. In conclusion, AURKA mediates hypoxic processes to reshape the tumor immunosuppressive microenvironment and promotes tumor immune escape.

Screening of small-molecule inhibitors of AURKA and their effects on skin melanoma
We screened the top 100 potential small-molecule inhibitors targeting AURKA, and preliminarily evaluated the toxic effects of the top 10 small-molecule drugs on A875 cells by MTT test. We found that the ZNC97018978 molecule has better biological activity and can significantly inhibit AURKA in A875 cells. proliferation (Fig. 9A). The effect of ZNC97018978 on the proliferation and migration of A875 melanoma cells was then analyzed. MTT results revealed that ZNC97018978 had strong cytotoxicity to A875 and could significantly inhibit the proliferation of A875 cells in a dose-and time-dependent manner (Fig. 9G). We also of 15 hypoxia signature genes with AURKA in SKCM patients. F, G Differential expression of 15 hypoxia signature genes in AURKAhigh and -low expression groups and between AURKA type I and type II. H, I TMER and EPIC algorithm analysis of the correlation between 15 hypoxia signature genes and immune cell infiltration levels. J, K Correlation of 15 hypoxia signature genes with MHC molecules and T cell signature genes evaluated the effect of ZNC97018978 on the A875 cell cycle and apoptosis and found that ZNC97018978 led to A875 cell arrest in the G2/M phase and promoted apoptosis (Fig. 9H, K). In addition, we also detected the expression of AURKA, Caspase-3, Bax, Bcl-2, and CDK1 in A875 cells after ZNC97018978 treatment and found that compared with the control group, the expression of AURKA, Bcl-2, and CDK1 in A875 cells in the ZNC97018978-treated group was significantly lower, and the of Bax and Caspase-3 was significantly higher (Fig. 9I, J, L).

Discussion
AURKA is a cell cycle-regulated kinase, which has been confirmed to be abnormally expressed in a variety of tumor tissues and is associated with tumor cell proliferation, invasion, metastasis, drug resistance, and clinical prognosis (Goos et al. 2013). In this study, we identified AURKA as a core driving hub gene in the occurrence and development of melanoma, we established AURKA as an independent prognostic factor of SKCM, and the nomogram developed based on AURKA expression also showed good prognostic prediction. We also found that AURKA can promote melanoma cell proliferation, invasion, and metastasis, as well as, remodel the tumor immunosuppressive microenvironment and participate in the treatment response of melanoma. Based on molecular docking technology and in vitro and in vivo experiments, we found that the small-molecule compound ZNC97018978 targeting AURKA has obvious anti-SKCM activity. These findings suggest that abnormally high expression of AURKA may serve as a potential tumor target and prognostic marker for the identification and treatment of melanoma patients with poor prognoses.
Mutations in genes in normal cells promote the progression of normal cells to tumors through stages of proliferation and dysplasia, and ultimately, metastatic tumors (Garnis et al. 2004). Therefore, genetic analysis of known oncogenes can provide further insights into the roles of these genes in tumor progression (Hahn and Weinberg 2002). In the present study, we found that AURKA exhibited a lower mutation rate, and missense mutation was the most common type of mutation. We also found that 1041 genes had somatic mutations in the AURKA-high and -low expression group in SKCM patients, of which the frequency of TTN mutation was the highest at 72%, and the BRAF mutation also had a remarkable role in the pathogenesis of SKCM. Furthermore, 103 gene mutations were found to affect the expression of AURKA, of which the TP53 mutation was the most significant. We also found that the expression level of AURKA was significantly correlated with copy number variation (p < 0.05), and affected the level of immune cell infiltration. These findings provided a new perspective for the analysis of variations in AURKA.
Recruiting tumor-activated CD8+ T cells is the focus of immunotherapy. Dendritic cells (DC) play a central role in the uptake and presentation of cancer antigens, and the initiation and activation of CD8+ T cell anti-tumor immunity (Jhunjhunwala and Hammer 2021). In contrast, CD4+ T cells secrete a variety of cytokines with direct effector functions and activate other immune cells such as B cells and CD8 T cells (Rakhra et al. 2010).In the present study, we also observed that AURKA inhibited the immune-infiltrating levels of anti-tumor immune cells (DC, B cells, CD4 + T cells, and CD8+ T cells) in the TME, while increasing the infiltration level of tumor immune escape-promoting cells (MDOS,). Our findings are consistent with that of previous reports (Punt et al. 2021;Vilgelm et al. 2016;Gide et al. 2018) suggesting that AURKA can inhibit the increased level of anti-tumor immune cell infiltration in the TME. We postulate that the formation of the immunosuppressive microenvironment is promoted for the following. First, AURKA inhibits the function of dendritic cells by activating the TGF-β signaling pathway (Keirsse et al. 2017) and negatively regulates T lymphocytes (Korn et al. 2009). It also inhibits the transcription of cytotoxicity-related genes such as perforin, granzyme A, and granzyme B (Chen et al. 2005; Seoane and Gomis 2017) which can significantly inhibit the adaptive immune surveillance ability of T lymphocytes and play a role in promoting tumorigenesis (Chen et al. 2005;Seoane and Gomis 2017;Li and Flavell 2008). Second, AURKA inhibits the cytotoxic function of CD8+ T cells through the ferroptosis pathway, thereby weakening the anti-tumor function of CD8+ T cells; AURKA mediates the development of a hypoxic microenvironment, thereby remodeling the immunosuppressive microenvironment. Therefore, our findings suggest that AURKA may play an immunosuppressive role by remodeling the tumor immunosuppressive microenvironment and impairing tumor immune surveillance function, however, the mechanism underlying its immunosuppressive effect is not known and must be further studied.
Hypermethylation of promoter regions often leads to silencing or inactivation of tumor suppressor genes in cancer cells Mehdi and Rabbani 2021). In the present study, we found that AURKA overexpression may be associated with AURKA hypomethylation. Interestingly, AURKA methylation was associated with SKCM prognosis, and patients with AURKA hypermethylation had lower overall survival, which was consistent with the prognosis of mRNA expression of this gene. Moreover, the expression of AURKA in 39 kinds of tumors was found to be closely related to the expression of DNMT1, DNMT3A, and DNMT3B, and AURKA methylation was found to promote the increased expression of EMT immune-related factors. Therefore, our findings suggest a role of AURKA methylation in SKCM and can be further investigated.
Since AURKA is the core hub gene in the occurrence and malignant development of cutaneous melanoma, it is closely related to the stage, treatment, and prognosis of cutaneous melanoma. This finding has important clinical implications, as it reveals that targeting AURKA may benefit not only patients with early-stage primary melanoma but also patients with advanced-stage melanoma. In recent years, several attempts have been made to develop AURKA inhibitors, some of which have entered the clinical trial stage. However, these AURKA inhibitors that are currently under investigation have limitations, such as poor selectivity, which can cause serious adverse reactions in patients (neutropenia, hypertension, hepatotoxicity, somnolence, fatigue, etc.) (Souza 1864; Panicker et al. 2019). Therefore, the development of efficient AURKA inhibitors with a clear mechanism of action, strong specificity, and less toxic side effects is essential. Therefore, we used molecular docking technology to screen for small-molecule inhibitors that target AURKA and identified ZNC97018978 from the ZINC database that can bind AURKA and block its expression. In vitro experiments confirmed that ZNC97018978 could significantly inhibit the proliferation and invasion of A875 cells and promote G2/M phase arrest. In vivo experiments further showed that ZNC97018978 can inhibit the growth of CM xenografts without obvious toxic and side effects on other organs. Furthermore, MTT assay and flow cytometry confirmed that after ZNC97018978 treatment, the proliferation rate of A875 cells decreased significantly, the expressions of proapoptotic Bax and Caspase 3 increased, and that of antiapoptotic Bcl-2 and CDK1 decreased. It is suggested that the anti-melanoma activity of ZNC97018978 may induce apoptosis by blocking the cell cycle, and thereby, inhibit cell proliferation.
This study improves our understanding of the relationship between AURKA and SKCM, but some limitations remain.
First, although we investigated the correlation between AURKA and immune infiltration in SKCM patients, interpretation of immune analyses of different SKCM subgroups was lacking. Second, most analyses in this study were based on AURKA mRNA levels. A deeper analysis based on protein level will further improve the validity of the data. Finally, the effect of ZNC97018978 on skin melanoma needs to be tested in vivo.
In conclusion, our results suggest that AURKA can serve as a potential therapeutic target for SKCM, and the role of epigenetic regulation in AURKA function. Thy also suggests that AURKA likely regulates TME immune cell infiltration in SKCM patients as well as via ferroptosis regulation. Our study may help clarify the role of AURKA in the development of SKCM, and offer insights into developing more precise and personalized immunity against AURKA in the future.