BUB1 predicts poor prognosis and immune status in liver hepatocellular carcinoma

Accurate assessment of the tumour immune microenvironment promotes individualized immunotherapy regimens and screens dominant populations suitable for immunotherapy. Therefore, potential molecular markers were investigated to make an overall assessment of the immune microenvironment status of liver hepatocellular carcinoma (LIHC). In this study, a total of 121 differentially expressed genes (DEGs) were identified, and DEGs were enriched in the epithelial–mesenchymal transition, hypoxia, myogenesis, and p53 pathways. A total of 20 hub genes were selected and a strong correlation was identified between these hub genes and prognosis. The expression of budding uninhibited by benzimidazoles 1 (BUB1) was found to be upregulated in LIHC and was strongly related to immune cells and immune checkpoint molecule expression. Immunohistochemistry (IHC) indicated that BUB1 expression was higher in LIHC tissues than in normal liver tissues. BUB1 knockdown resulted in reduced proliferation and vertical migration ability of LIHC cells, and reduced the expression of phospho‐SMAD family member 2 and phospho‐SMAD family member 3 proteins. IHC showed that BUB1 expression was accompanied by immune cell infiltration into LIHC tissues. These results suggest that BUB1 may serve as a potential prognostic biomarker for LIHC and as an indicator of its immune status.

*With equal contribution.
Liver hepatocellular carcinoma (LIHC) is the seventh most common malignancy and ranks second in terms of global cancer-associated mortality [1]. Thus, it poses a significant threat to human health on a global scale [2]. Approximately 780,000 individuals are diagnosed with LIHC every year, and a substantial proportion of these individuals are diagnosed with late-stage tumours [3]. Despite significant progression in the treatment of LIHC recently, the survival probability of these patients is still below 20%, and their prognosis is unsatisfactory [4]. This prognosis status of patients is partly dependent on the surrounding microenvironment during tumour occurrence and development. The surrounding immune microenvironment can enable LIHC to escape immune recognition and attack by the body [5]. The activation of inhibitory receptors and ligands, challenges associated with antigen presentation, and accumulation of immunosuppressive cell populations constitute the immunosuppressive microenvironment, thus promoting LIHC progression [6,7]. Therefore, research on genes related to tumour immune infiltration may aid the assessment of the immune microenvironment of LIHC, thereby facilitating patient stratification for immunotherapy and the development of personalized treatment approaches.
Budding uninhibited by benzimidazoles 1 (BUB1) is a kinase ensuring segregational fidelity in daughter cells during chromosomal segregation [8]. It is encoded by the BUB1 gene in humans. BUB1 can be broadly segregated into three regions, that is, a catalytic threonine or serine kinase domain, a scaffold for protein recruitment, and a tetratricopeptide repeat part [9]. It is recognized as a tumour promoter in gastric cancer [10]. BUB1 was reported to be linked to forkhead box O3 in pancreatic cancer, which promotes pancreatic cancer development [11]. The unfavourable prognostic effect of BUB1 expression in LIHCidentified by bioinformatic analysishave been corroborated [12]. Though BUB1 activity has been explored, the mechanisms by which BUB1 affects the prognosis and immune microenvironment in LIHC remain unclear. In this study, we first identified differentially expressed genes (DEGs) from the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA) databases and then identified the hub genes to further screen the immune-related gene BUB1. We observed that BUB1 was associated with prognostic and immune roles in LIHC, thereby highlighting its relevance as a marker for LIHC immunotherapy.

Data pre-processing and DEGs identification
The Affy package in R [13] was used to analyse original data. After correcting for the inter-batch difference, original data were subjected to perform background correction, quality control, and standardization processing. The limma package in R [14] was applied to filter DEGs from the data set. The interception criteria for DEGs in the above database were |log fold change (FC)| ≥1 and p < 0.05 after adjustment. The results were visualized as a Venn diagram.
Gene ontology analysis, kyoto encyclopaedia of genes and genomes analysis, and gene set enrichment analysis of DEGs Gene ontology (GO) is one of the most frequently used tools for gene annotation, often employed for largescale gene annotation. Kyoto encyclopaedia of genes and genomes (KEGG) was used to identify pathways associated with DEGs. The distribution trends of genes and phenotypes were evaluated using gene set enrichment analysis (GSEA). GO analysis, KEGG analysis, and GSEA were conducted using the clusterProfiler package in R [15].

Construction of protein-protein interaction network and hub gene identification
The interaction between proteins encoded by the DEGs was evaluated by the STRING tool [16]. Then, Cytoscape software (v3.7.1) [17] was a platform for visualizing the protein-protein interaction (PPI) information. The top 20 genes were selected as hub genes using the cytoHubba plug-in [18].

Prognostic validation of the hub genes
Hub genes strongly correlated with prognosis were screened out by univariate and LASSO Cox regression analysis. The Survminer package (https://github.com/ kassambara/survminer) and the survival package in R (https://github.com/therneau/surviva) were then used for mapping analysis. The association between genes and clinicopathological characteristics was analysed by logistic regression and receiver operating characteristic curve analyses. The selection criterion was p < 0.05.

Expression and verification of hub genes
The Gene Expression Profiling Interactive Analysis database [19] was used to analyse hub gene expression and related transforming growth factor b (TGFb) signalling. The online tool TIMER [20] was used to explore the correlation between hub genes and tumour-infiltrating immune cells.

Tissue specimens and ethics
A total of 42 paraffin-embedded and formalin-fixed LIHC tissue samples were collected from Lanzhou University Second Hospital. All patients had undergone curative resection surgeries. The study protocol was approved by the Institutional Ethics Committee of Lanzhou University Second Hospital. Written informed consent was obtained from all patients. Disease-free survival (DFS) was the period between the day of surgery and the date of recurrence.
Wherein a score of 0 indicated <5% positive cells; a score of 1 indicated 6%-20% positive cells; a score of 2 indicated 21%-50% positive cells; and a score of three indicated >50% positive cells.

Transfection
To silence the expression of endogenous BUB1 in human LIHC cell lines, 2.5 µg short hairpin RNA (shRNA) plasmids (Genechem, Shanghai, China) and 5 lL Lipofectamine 6000 (Beyotime, Shanghai, China) were mixed in 500 lL DMEM (Hyclone, Waltham, MA, USA) medium for 5 min. After incubation for 5 min, the mixture was added to cells that had reached a confluence of 60%-70% of the 6-well plate (Corning, New York, NY, USA). The cells were used for the further experiment after transfection for 24 h. shRNA sequences were as follows: shBUB1, Cell counting kit-8 assay

Migration assays
For migration assays, 1 9 10 5 HUH7 and SKP1 cells were independently resuspended in the medium (1% foetal bovine serum) (ThermoFisher) and seeded in the upper chamber of a Transwell chamber (Corning) placed in 24well culture plates (Corning). Five hundred microlitres of medium (10% foetal bovine serum) were added to the bottom chamber. After incubation at 37°C for 24 h, a cotton swab was used to remove the non-migrating cells on the membrane located at the interface of the upper and lower chambers, and the migrated cells were fixed, stained, counted, and imaged. Six areas from each membrane were imaged from three independent wells.

Statistical analysis
The clusterProfiler package in R (Yu et al., 2012) and GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA) were used for data analysis. Student's t-tests were used to compare data between groups. Significance was defined as p < 0.05.
The correlation between these 20 hub genes and the poor prognosis of LIHC patients was analysed via Kaplan-Meier analysis (p < 0.05) (Fig. 4). The 20 hub genes were then subjected to LASSO-Cox regression analysis to calculate the correlation coefficients. After cross-validation (Fig. 5A), hub genes strongly related to survival were finally identified that contained BUB1, CENPF, and PRC1 (Fig. 5B).

Expression of BUB1, CENPF, and PRC1 in LIHC
The median expression of BUB1, CENPF, and PRC1 in tumour and normal liver samples is shown in Fig. 6. The expression of BUB1 (Fig. 6A), CENPF (Fig. 6B), and PRC1 (Fig. 6C) was found to be upregulated in LIHC as well as in other cancer types when compared to that in control samples. BUB1 (Fig. 6A), CENPF (Fig. 6B), and PRC1 (Fig. 6C) were upregulated in stages I-III of LIHC and downregulated in stage IV (Fig. 6). Multivariate and univariate Cox analyses suggested that BUB1 expression was an independent prognostic factor of the overall survival of LIHC patients ( Table 2, p < 0.05).

DISCUSSION
There is clear evidence that the immunosuppressive microenvironment of LIHC facilitates immune escape and tolerance through a variety of mechanisms; this feature makes LIHC a challenging candidate for immunotherapy [23]. Emerging evidence suggests that counteracting these immunosuppressive mechanisms may significantly alter the clinical outcomes of LIHC, and will aid in conducting in-depth investigations on the correlation between the occurrence and prognosis of LIHC and immune genes [24]. Currentlyaided by bioinformatic analysiscancer immunotherapy [25] is ushering in a new epoch of anti-cancer therapies. In this study, the gene expression profiles in LIHC and associated pathological mechanisms were investigated using bioinformatics tools. We identified BUB1 as being closely related to tumour cell proliferation, metastasis, and immunity in LIHC. We found that research on its mechanism of action in LIHC is currently scarce; therefore, we performed a comprehensive analysis of its expression, clinical relevance, potential prognostic, and immunologic value in LIHC.
BUB1 is verified as a key protein in mitosis, and it has been recognized as an oncogene in gastric and breast cancer [10,26]. In this study, BUB1 expression was found to be upregulated in LIHC tissue samples. Further, BUB1 was significantly associated with the clinical stage of LIHC. Survival analyses indicated that BUB1 was associated with reduced survival. Further, BUB1 knockdown resulted in reduced proliferation and vertical migration of LIHC cells. Therefore, we speculated that BUB1 may act as an oncogene, thereby promoting LIHC progression.
Common pathways that are deregulated in LIHC include the cell cycle, immune response, DNA replication, DNA repair, p53, and TGFb signalling. Some studies have shown that TGFb signalling has a vital influence on the progression of LIHC [27] through the regulation of immune cell survival, development, proliferation, and differentiation [28,29]. TGFb signalling reduces CD4+/CD8+ T cell, natural killer cell, and macrophage counts. It also promotes the generation of regulatory T cells by upregulating FoxP3 via the SMAD2/SMAD3 pathway [30]. Recently, combinatorial therapy with anti-PD-L1 immune checkpoint inhibitors (ICIs) and systemic TGFbR1 inhibitors (e.g. galunisertib) or anti-TGFb antibodies was shown to promote the infiltration of T cells into the tumour core, thereby inducing tumour regression and anti-tumour immunity [31]. BUB1, as a kinase, is reported to have a vital impact on TGFb-dependent signalling [21]. There is an established fact that BUB1 is the engine of TGFb signalling [32]. Because of the immunosuppressive effect of TGFb-dependent signalling in LIHC and the role of BUB1 in TGFb signalling, we speculated that BUB1 might be a key molecule in the TGFb-based immune regulation in LIHC. Our results verified the hypothesis that BUB1 expression was related to immune cell counts and immune checkpoint molecule expression in LIHC. A previous study also suggested BUB1 as a new indicator of the ICI response in lung cancer [33]. Thus, we predicted that BUB1 might also be an important gene for ICIs therapy and might provide insights that enable us to identify novel targets for immune therapies for LIHC. We confirmed that BUB1 was associated with immune cell infiltration in LIHC, highlighting its influence on LIHC immune status. In a subsequent study, we will verify the involvement of BUB1 in immune regulation in vivo.

CONCLUSION
In summary, BUB1 expression may predict poor prognosis and immune status of LIHC. Thisstudy provides a theoretical basis and a novel insight for further exploring the specific regulatory mechanisms associated with BUB1 in LIHC.
We would like to express our gratitude to Editage (www. editage.cn) for English language editing. This work was supported by the Key Talents Project of Gansu Province