Identify BCAT1 plays an oncogenic role and promotes EMT in KIRC via single cell RNA-seq and experiment

Background Kidney renal clear cell carcinoma (KIRC) is a major subtype of renal cell carcinoma with poor prognosis due to its invasive and metastatic nature. Despite advances in understanding the molecular underpinnings of various cancers, the role of branched-chain amino acid transferase 1 (BCAT1) in KIRC remains underexplored. This study aims to fill this gap by investigating the oncogenic role of BCAT1 in KIRC using single-cell RNA-seq data and experimental validation. Methods Single-cell transcriptomic data GSE159115 was utilized to investigate potential biomarkers in KIRC. After screening, we used BCAT1 as a target gene and investigated its function and mechanism in KIRC through databases such as TCGA-GTEx, using genome enrichment analysis (GSEA), genome variation analysis (GSVA), gene ontology (GO) and Kyoto Encyclopedia of the Genome (KEGG). BCAT1 expression was detected in clinical tissue samples using Western Blotting (WB) and immunohistochemical (IHC) staining techniques. We established cell lines stably overexpressing and knocking down BCAT1 and performed WB, qRT-PCR, cell scratch assay and transwell assay. Results BCAT1 was highly expressed in KIRC and was associated with disease prognosis and TME. Patients with mutations in the BCAT1 gene had shorter overall survival (OS) and disease-free survival (DFS). patients with high BCAT1 expression had shorter OS, progression-free interval (PFI), and disease-specific survival (DSS). GSEA showed that BCAT1 was significantly enriched in epithelial mesenchymal transition (EMT). Bioinformatics analysis and WB and IHC staining showed that BCAT1 expression was higher in KIRC than in paracancerous tissues. In vitro experiments confirmed that BCAT1 in KIRC cells may promote EMT affecting its invasion, migration. We constructed a protein interaction network (PPI) to hypothesize proteins that may interact with BCAT1. Single-sample gene set enrichment analysis (ssGSEA) revealed the immune infiltration environment of BCAT1. Furthermore, hypomethylation of the BCAT1 promoter region in KIRC may contribute to disease progression by promoting BCAT1 expression. Conclusion BCAT1 promotes KIRC invasion and metastasis through EMT and has prognostic predictive value and potential as a biomarker. It may become a novel biomarker.

Branched-chain amino transferases 1 (BCAT1) is a cytoplasmic isoform of branched-chain amino acid transferase found in humans and a variety of organisms [1] .Branched-chain amino acids (BCAAs) include valine, leucine, and isoleucine.They serve as a nitrogen source for the formation of macromolecules such as nucleotides and are also essential for cancer cell growth.The metabolic exibility of cancer cells is determined by their ability to reprogram synthetic and catabolic pathways through altered gene expression programs and intercellular interactions in the tumor microenvironment [2,3] .In many cancers, enzymes with amino acid catabolic functions are highly expressed.These enzymes not only provide cellular energy and metabolites for anabolic processes, but also promote immune escape of cancer cells.
Epithelial-to-mesenchymal transition(EMT) affects the ability of tumor cells to invade and metastasize through multiple pathways and is an important process in the transition of malignancy from early to advanced stages, whereby fully differentiated cells lose their cell polarity and intercellular adhesion properties and instead acquire a migratory and invasive mesenchymal phenotype [43,44] .

Processing and analysis of single-cell RNA sequencing (scRNA-seq) data
The single-cell expression matrix of the KIRC single-cell dataset GSE159115 was loaded and created as "SeuratObject" via the "Seurat" package [46] , which annotated 27669 single cells with cell type based on the cell meta information therein.The 2000 high-variable genes were ltered using the "FindVariableFeatures" function and the "ScaleData" function.And based on the highly variable genes, PCA was performed on single cells using "RunPCA" function to reduce the dimensionality, check the PCA binning results and select the principal components.Subsequently, the KIRC single-cell data were clustered and analyzed using the Louvain algorithm, and the data were nonlinearly dimensionalized using the "Uniform Manifold Approximation and Projection (UMAP)" function [47] .Malignant cells from KIRC in GSE159115 were then extracted using the subset function in the Seurat package, and cell subclustering was performed using the "FindNeighbors" and "FindClusters" functions, followed by UMAP nonlinear dimensionality reduction visualization in the same manner as described above.Marker gene analysis was performed for subclustering of Malignant cells by the "FindAllMarkers" function, and Marker genes were analyzed using the wilcoxon algorithm in a selected group vs rest manner.The R package Monocle2 was used for the proposed chronological analysis of subclusters of Malignant cells [48] , and the cells were sorted using subcluster marker genes to determine the gene expression patterns among Malignant differentiation trajectories as a whole, and the differentiation trajectory-related genes were searched by the "differentialGeneTest" function.

The cBioPortal
cBioPortal is an online tool for tumor genomics data analysis.We selected Kidney Renal Clear Cell Carcinoma (TCGA, Firehose Legacy) as the study dataset in the cBioPortal database, which contains 538 KIRC samples with mutations, copy number variants and mRNA expression data with survival analysis, clinical characterization, etc.

GO and KEGG pathway enrichment analyses
Through Gene Ontology (GO) analysis we obtained information on the biological process, cell composition and molecular function related to BCAT1 [49] .In addition, we used the Kyoto Encyclopedia of Genes and Genomes (KEGG) to analyze the role of BCAT1 in metabolic pathways [50] .We used the R package "ClusterPro ler" for GO and KEGG enrichment analysis of the biological functions of BCAT1 in KIRC [51] .

GSEA and GSVA
We performed the analysis using the Gene Set Enrichment Analysis (GSEA) package (4.1.0)for discovering speci c biological processes and disease-associated biological pathways or gene sets associated with BCAT1 gene differential expression in the database.
Gene Set Variation Analysis (GSVA) which is a non-parametric and unsupervised algorithm was used for Bulk RNA-seq data (TCGA) using the "gsva" algorithm in the GSVA package to quantify the hallmark of each sample from KIRC gene sets for each sample of KIRC [52] .The "limma" package also was used to analyze the differential genes between high and low enrichment scores [53] .For scRNA-seq data, the "gsva" algorithm was performed to calculate the enrichment scores of hallmark gene sets in each cell and count the differences in the enrichment scores of the hallmark gene sets between different cell types and different differentiation trajectories of malignant cells.

TISIDB
An integrated repository portal for tumor-immune system interactions (TISIDB) integrates multiple data types including PubMed database, transcriptomic and clinical data from TCGA, exomics and RNA sequencing datasets from immunotherapy patient cohorts, and we used it to analyze the relationship between BCAT1 and lymphocyte in ltration, immunomodulators, chemokines, and immune subtypes.

UALCAN
The University of Alabama at Birmingham Cancer Data Analysis Portal (UALCAN) is an interactive public database for analyzing cancer histology data.
We analyzed the relationship between BCAT1 mRNA expression levels, promoter region methylation levels, and clinicopathological KIRC information via the UALCAN website by selecting the "TCGA" section.

Construction of prognostic signature
The R packages "survival" and "survminer" were used for Kaplan-Meier prognostic analysis of the high and low BCAT1 expression groups in the TCGA-KIRC dataset.The R package "timeROC" was utilized to evaluate the AUC values of the ROC curves over time.To obtain a more accurate nomogram, the R package "rms" was used to combine pathological staging with risk scores in order to further improve the predictive e ciency of risk scores [54] .

Cell culture and specimens
The 786-o(RRID:CVCL_1051), 769-p(RRID:CVCL_1050) and 293T(RRID:CVCL_0063) cell lines used in this experiment were purchased from the cell bank of the Chinese Academy of Sciences and identi ed by the STR test.The cells were cultured at 37℃ in a 5% CO 2 humidi ed incubator using RPMI-1640 and DMEM/high glucose medium containing 10% FBS (Shanghai Yuanpei Biotechnology Co., Ltd., China).None of the cells used were contaminated with mycoplasma.
The 21 pairs of KIRC and paraneoplastic tissue specimens used in this study were obtained from surgically resected patients at the First A liated Hospital of Soochow University.Informed consent was obtained from the patients, and the ethics committee of Soochow University approved the use of the specimens.Patients had not received radiotherapy or immunotherapy prior to specimen collection, and specimens were divided after removal and stored frozen at -80℃ or in formalin at room temperature until further processing.

Western blot assessment
Cells were hydrolyzed in RIPA buffer containing a mixture of protease inhibitors and phosphatase inhibitors.Protein products were separated by SDS-PAGE, transferred to nitrocellulose membranes, and cut according to the molecular weight of the target proteins.The membranes were then blocked in TBST buffer containing 3.0% BSA for 1 hour and incubated with primary antibody at 4℃ overnight.The membranes were then rinsed three times for 10 minutes each with TBST buffer and incubated with the appropriate secondary antibody for 2 hours at room temperature.Protein expression was detected using an enhanced chemiluminescence system.The expression of the target protein was normalized to the expression of GAPDH or tubulin.The primary antibodies used for WB analysis were as follows: antimouse E-cadherin antibody, anti-mouse N-cadherin antibody (BD Biosciences, USA); anti-mouse snail antibody (Cell Signaling Technology, USA); anti-mouse vimentin antibody (Santa Cruz, USA); anti-rabbit BCAT1 antibody, anti-mouse GAPDH antibody, anti-mouse tubulin antibody (Abclonal, China).The secondary antibodies were: goat anti-mouse IgG antibody, goat anti-rabbit IgG antibody (Abclonal, China).The relative expression of proteins was normalized to the expression of the same group of lysate samples were run in parallel gels to test different proteins.

RNA extraction and qRT-PCR assays
Total RNA was extracted using the TRIzol RNA protocol, and cDNA was synthesized by reverse transcription using the HiScript® III RT SuperMix for qPCR kit (Vazyme, China).BCAT1 mRNA expression was detected by qRT-PCR and normalized to β-actin.The target gene expression levels were calculated relative to β-actin.The expression level of the target gene was calculated relative to β-actin.Finally, the expression level was quanti ed using the ΔΔCt method [55] .The primers used in the experiment included: BCAT1 (Forward primer: GTGGAGTGGTCCTCAGAGTTT; Reverse primer: AGCCAGGGTGCAATGACAG); βactin (Forward primer.CATGTACGTTGCTATCCAGGC; Reverse primer: CTCCTTAATGTCACGCACGAT)

Wound-healing (cell migration) assay
The 786-o and 769-p cells were evenly spread on the well plates, and when the cells reached 95% density, the cells were gently scraped along a straight line with the tip of a sterilized pipette, with 5 lines through each well, each line kept parallel and equally spaced.Serum-free medium was added and incubated in a 37℃, 5% CO 2 incubator for 12 hours [56] .Microscopic photographs of the cells were taken at ve randomly scratched areas, and then the distance of cell migration to the scratched areas was analyzed using ImageJ software.The experiment was repeated three times.

Transwell(cell migration and invasion) assays
Transwell chambers (BD Biosciences, USA) were placed in 24-well culture plates.The surface of the chambers used in the invasion assay must be covered with a layer of matrix gel.In each upper chamber, 786-o and 769-p cells were added with 1% FBS medium, and 20% FBS medium was placed in the lower chamber as a chemoattractant.The 24-well plates were incubated in a 37℃, 5% CO 2 incubator for 24 hours, and the upper surfaces of the chambers were cleaned with a swab.The cells in the bottom layer were stained with 1% crystal violet stain, three microscopic images were taken, and the cells were counted.Results were tested in triplicate.

Statistical analysis
The analysis of all data and graphs was performed using the Sento Academic Platform software(www.xiantaozi.com) and R software (version number 3.6.3).Differences between two sample means were analyzed using the t-test, and one-way ANOVA was applied to the analysis of differences between more than two samples.We considered p values < 0.05 as statistically signi cant(*p < 0.05, **p < 0.01,***p < 0.001).

scRNA-seq and cell annotation of KIRC samples
To analyze the cellular composition of the tumor microenvironment(TME) in KIRC, the most predominant subtype in renal cell carcinoma, we extracted eight quality controlled and standardized KIRC single-cell gene expression pro les from the single-cell transcriptome sequencing dataset GSE159115 to characterize the cellular populations of human KIRC.The samples from these eight patients were quality-controlled to contain a total of 27,669 cells.To detect possible batch effects in single-cell RNA sequencing (scRNA-seq) data, we performed principal component analysis (PCA) and demonstrated that there were no signi cant batch effects for cells in different samples (Figure 1A).We used the R package "Seurat" to perform UMAP downscaling and clustering of these cells, and classi ed the cells in the KIRC tumor microenvironment into 33 clusters based on the similarities and differences in their expression patterns (Figure 1B).Further cell annotation results showed that the 33 cell clustering subgroups could be annotated into 8 classes of cells, namely CD8+ T cells, endothelial cells, epithelial cells, erythroblasts, malignant cells, monocytes/macrophages, pericytes, and plasma (Figure 1C).
To better understand the heterogeneity of malignant epithelial cells in the TME of KIRC and their potential value for prognosis and drug therapy screening, we extracted malignant epithelial cell types at the single cell level and further subdivided them into subpopulations using the "Seurat" package.The results showed that malignant epithelial cells could be classi ed into 11 subpopulations (0-10) based on different molecular markers (Figure 1D, E).This indicates that there is heterogeneity of gene expression in different clusters of malignant epithelial cells in KIRC patients, and these ndings suggest that the heterogeneity of malignant epithelial cells in the tumor microenvironment of KIRC may play a crucial role in the development of KIRC.

Malignant epithelial cell trajectory analysis and EMT-related gene screening
Gene expression heterogeneity among KIRC malignant epithelial cells may be involved in differences in the biological behavior of cancer cells, and to further identify cell subpopulations and molecular markers that are closely associated with KIRC progression, we performed an analysis of differentiation trajectories (Figure 2A).The results revealed that the 11 malignant epithelial cell subpopulations could be divided into ve differentiation stages.
We found that different subpopulations of malignant epithelial cells were differentially distributed in differentiation trajectories (Figure 2B).We then performed pseudotime analysis of malignant epithelial cells (Figure 2C, D), the purple part is the initial state of differentiation while the yellow is the terminal state of differentiation, and it can be found that cell state 4 is located at the terminal stage of differentiation.
GSVA analysis demonstrated the activation levels of 50 Hallmark pathways associated with cancer in different cell types in the TME of KIRC (Figure 2E).GSVA also demonstrated the activation levels of Hallmark pathways associated with cancer in different differentiation states of malignant epithelial cells (Figure 2F), in which transforming growth factor-β (TGF-β), PI3k-AKT-mTOR, Wnt-β-Catenin, reactive oxygen species (ROS), angiogenesis and other important signaling pathway activation levels were signi cantly upregulated .
In addition, we found that the heterogeneous genes associated with the differentiation trajectory of malignant epithelial cells showed three expression patterns after clustering (Figure 2G), and we divided three geneclusters accordingly.genecluster1 showed a high upregulation at the terminal stage of the proposed time sequence.We performed GO enrichment analysis of gencluster1 (Figure 2H), which showed that these genes were signi cantly enriched in signaling pathways related to cell adhesion, epithelial cell migration and development, such as Focal adhesion, Cell adhesion molecules, and tight junction.In addition, signaling pathways such as phosphatidylinositol-3-kinase-protein kinase B (PI3K-Akt) and mitogen-activated protein kinase (MAPK) were also signi cantly activated.
We performed GSVA analysis of the HALLMARK-EPITHELIAL-MESENCHYMAL-TRANSITION gene set using KIRC mRNA expression data from the TCGA database to obtain differentially expressed genes in samples with high versus low EMT levels (Figure 2I).1370 genes were upregulated in the high EMTenriched subgroup from the bulk RNA-seq level these genes are participating in the EMT process in KIRC.
On the other hand, we introduced proteins that had been reported to be signi cantly upregulated in KIRC in a proteogenomic study by Clark et al [57] .Genes with remarkable signi cance in scRNA-seq, bulk RNAseq, proteogenomics and prognostic analysis were jointly intersected and screened for seven key genes affecting the EMT process in KIRC including BCAT1, CTHRC1, GPX8, RUNX1, SERPINE1, SERPINH1 and TGFBI (Figure 2J).We traversed the literature and observed that BCAT1 is signi cantly upregulated during progression of various types of cancer and has emerged as an essential biomarker, but its role in KIRC is poorly understood.Correlation analysis demonstrated that BCAT1 expression was highly correlated with the HALLMARK-EPITHELIAL-MESENCHYMAL-TRANSITION gene set (r=0.562, p<0.001) (Figure 2K).Thus, we decided to further explore the role and mechanism of BCAT1 in KIRC metastasis and development.

Genetic alteration of BCAT1 gene in KIRC
We used the cBioPortal tool to analyze mutations in the BCAT1 gene in KIRC using the Kidney Renal Clear Cell Carcinoma (TCGA, Firehose Legacy) dataset, which contains 538 KIRC samples (Figure 3A).
We found that the overall genetic variation rate of BCAT1 in KIRC was 5%, with alterations at the mRNA level being the predominant form (Figure 3B).There was one somatic mutation site in the amino acid sequence of BCAT1 where missense mutations occurred.Analysis revealed no signi cant differences in BCAT1 mRNA expression levels between CNV copy number ampli cation and deletion, nor between mutation and non-mutation (Figure 3C, D).However, we found that methylation of the promoter region of BCAT1 was negatively correlated with BCAT1 mRNA expression (r=-0.17,p<0.05) (Figure 3E).
We further investigated the effect of BCAT1 gene mutation on patient prognosis.The results showed that overall survival (OS) and disease-free survival (DFS) were signi cantly decreased in the mutated group compared with the non-mutated group of KIRC patients (Figure 3F, G).Taken together, these data suggest that mutations in BCAT1 may have an impact on tumor progression and thus patient prognosis.

Expression levels of BCAT1 in KIRC and prognosis of patients
We obtained the structure of the protein encoded by BCAT1 gene from The Human Protein Atlas database (Figure 4A).We identi ed BCAT1 expression levels in KIRC tumor tissues and paraneoplastic tissues in the TCGA database and found that BCAT1 expression was signi cantly elevated in tumor tissues (Figure 4B).We also found elevated levels of BCAT1 expression in multiple KIRC cell lines through the cBioPortal tool (Figure 4C).
We performed Western Blot on tumor tissue and paracancerous tissue from 21 pairs of clinical specimens (Figure 4D) and immunohistochemical staining on 20 of them (Figure 4E), all of which showed signi cant overexpression of BCAT1 in KIRC.This validated the results in the database.
The ROC curve we plotted as well revealed that BCAT1 was used for single gene diagnosis of KIRC with high accuracy (AUC=0.806,95% CI: 0.756-0.856)(Figure 4F).We used the Kaplan-Meier method to analyze the prognostic value of BCAT1.The results (Figure 4G) showed that among KIRC patients, the BCAT1 high expression group had a shorter progression free interval (PFI) compared to patients in the low expression group, and OS and disease-speci c survival (DSS) were similarly signi cantly lower.
By performing analysis through the UALCAN database, we obtained the expression of BCAT1 in different molecular subtypes, different genders, different grades, different N-stages and different tumor stages of KIRC.The results (Figure 4H) showed that among the different molecular subtypes, it was signi cantly more expressed in the ccB molecular subtype than in the ccA subtype.And the expression was higher in male patients than in females.Among the tumor grades, the highest BCAT1 expression was found in KIRC of Grade 4. BCAT1 expression was higher in patients with lymph node metastasis.There was a slight difference in BCAT1 expression in patients with different stages of KIRC.

BCAT1 co-expression network construction and enrichment analysis by GSEA
We constructed a protein-protein interaction (PPI) network to nd BCAT1 co-expressed or interacting genes.Using the GeneMANIA database, we identi ed the top 20 interacting genes (Figure 5A).And in the STRING database, we visualized the top 50 interacting genes (Figure 5B).The PPI network covers seven types of linkage types: physical interactions, co-expression, predicted, co-localization, genetic interactions, pathway, and shared protein domains.it stands to reason that BCAT2 is closely related to BCAT1, and these genes may play a joint role with BCAT1 in the development of KIRC.
Enrichment analysis was performed by GSEA software, and cell adhesion-related pathways such as Focal Adhesion, N-cadherin, PI3K-AKT and WNT signaling pathways were enriched in KIRC.All these pathways are closely associated with EMT.Further suggesting that we BCAT1 may be involved in the regulation of EMT in KIRC thereby affecting its metastasis (Figure 5C).

Overexpression of BCAT1 promotes EMT in KIRC and facilitates cell migration and invasion
We transiently transfected BCAT1 overexpression into 786-o and 769-p cells, and then examined the mRNA and protein expression levels of BCAT1 by qRT-PCR and Western blot.As shown in Figure 6A, we detected the expression of EMT-related markers by Western blot, and the expression levels of N-cadherin and Snail were signi cantly increased, and the expression levels of E-cadherin were signi cantly decreased.
We then performed stable BCAT1 overexpression in 786-o and 769-p cell lines using pCDH-BCAT1-3x Flag.We similarly validated the overexpression effect (Figure 6B).
The 786-o and 769-p cells overexpressing BCAT1 had enhanced migratory ability (Figure 6C), and the Transwell assays showed that these cells had signi cantly enhanced migratory and invasive abilities (Figure 6D).

Knockdown of BCAT1 inhibits EMT in KIRC and suppresses cell migration and invasion
In addition, we transiently knocked down BCAT1 in 786-o and 769-p using siRNA.We veri ed the effect at the mRNA level using qRT-PCR and at the protein level using Western blot (Figure 7A).After BCAT1 was successfully knocked down, the protein expression levels of EMT-related markers changed, with a signi cant decrease in N-cadherin and Snail expression levels and a signi cant increase in E-cadherin expression levels.After constructing BCAT1 stably knocked-down cells by the above method, scratch healing assays showed a signi cant decrease in the migratory ability of these cells (Figure 7B).transwell assays also showed that BCAT1 knockdown signi cantly inhibited the migratory and invasive ability of the cells (Figure 7C).

Correlation of BCAT1 with immune cells
Using the TISIDB database, we investigated the correlation between BCAT1 expression and various immune cell in ltrations based on pan-cancer samples (Figure 8A), and found that BCAT1 was closely associated with various immune cell in ltrations in most cancers.BCAT1 was differentially expressed in different immune subtypes of KIRC, with the highest expression in the c5 immune subtype (Figure 8B).In 534 KIRC samples, the in ltration abundance of most immune-associated cells showed a positive correlation with the mRNA expression level of BCAT1 (Figure 8C), including Th1 cells, Th2 cells, regulatory T cells (Treg), natural killer (NK) cells, and macrophages.These results suggest that BCAT1 expression may be involved in the regulation of immune in ltration in the KIRC tumor microenvironment.
To further determine the relationship between BCAT1 expression and immune cell in ltration in KIRC, we used the ssGSEA algorithm to enrich the gene set of 24 characteristic immune cell markers (Figure 8D) and analyzed the correlation between BCAT1 expression and the degree of enrichment of common immune cells in KIRC, in which macrophages and Th2 cells were signi cantly enriched with increased BCAT1 expression enrichment, while the opposite was true for cytotoxic cells and the CD56bright subpopulation of NK cells (Figure 8E).
We also compared the differences in immune cell in ltration between the high and low BCAT1 mRNA expression groups, which showed signi cant differences in macrophages, Th2 cells, Th1 cells, T cells, and the CD56bright subpopulation of NK cells (Figure 8F).In addition, radargrams visualised the correlation of immune activation checkpoints, immune suppression checkpoints and major histocompatibility complex (MHC) molecules with BCAT1 expression levels (Figure 8G,H,I).

Univariate analysis of the prognostic value of BCAT1 and the prognostic model
We performed univariate analysis on the KIRC samples from the TCGA database.The results showed that BCAT1 had signi cant prognostic predictive signi cance on DSS, OS and PFI, as well as pathological stage, TMN stage, pathological grading, immunotherapy outcome and serum calcium (Figure 9A,B,C).
We also included pathological grading, age, pathological staging, immunotherapy outcome, and gender combined with BCAT1 mRNA expression levels to establish a risk prognosis model using multifactorial regression analysis and plotted a nomogram, and the higher the score, the worse the prognosis as measured by the nomogram.The calibration curve showed that the nomogram has good performance in predicting prognosis (Figure 9D).It can be used to comprehensively predict the prognosis of KIRC patients.

Correlation of BCAT1 promoter region methylation with KIRC clinical factors and BCAT1 mRNA level expression
As mentioned above, the methylation level of BCAT1 promoter region was signi cantly and negatively correlated with the mRNA expression of BCAT1, and hypomethylation of BCAT1 promoter region might promote the expression of BCAT1.Therefore, we further analyzed the correlation between the methylation levels of BCAT1 promoter region and clinical factors by UALCAN (Figure 10A).The analysis showed signi cant differences in the methylation level of this promoter region between different ages, grades and stages, while there were no signi cant differences between different genders and different N stages.
Overall, the methylation level of BCAT1 promoter region was signi cantly lower in KIRC tumor tissues than in normal kidney tissues.Subsequently, our further analysis showed that the promoter region methylation levels of cg07259733, cg21500300, cg12371924, cg19008597, cg07537523, cg04011247, cg10764357, and cg13980808 were signi cantly negatively correlated with BCAT1 mRNA expression (Figure 10B).

Discussion
Kidney cancer has the highest mortality rate of the three major urologic tumors.Renal cell carcinoma accounts for more than 90% of renal cancer cases.The mortality rate of renal cell carcinoma is slowly decreasing worldwide, bene ting from the development of treatment modalities and increased frequency of medical screening [41] .KIRC, the predominant type of renal cell carcinoma, has a well-established immunotherapy system [58,59] , but the search for new targets or valuable biomarkers is of great value for both the treatment and prognosis of KIRC.
In the present study, we rst pro led KIRC at the single-cell level using publicly available bioinformatics data.Based on information from the GSE159155 database, heterogeneity in the differentiation trajectory of KIRC malignant epithelial cells was revealed, with high activation of key signaling pathways such as EMT, TGF-β, Notch, IL2-STAT5, WNT-β-catenin, ROS, and angiogenesis in cells at the end of the differentiation trajectory.GO analysis revealed that high expression of terminal differentiation genes was enriched in EMT and cell adhesion-related pathways such as PI3K-AKT and MAPK.Combining prognostic information, proteogenomic data, bulk rna-seq and scRNA-seq data, we screened the target gene of interest, BCAT1.BCAT1, as a member of the branched-chain amino acid transferases, is an important component of amino acid metabolism.BCAT1 is highly expressed in a variety of cancers, promotes cancer cell migration and invasion, and has received much attention in recent years.
BCAT1 has been reported to be highly expressed in a variety of cancers and to be associated with poor progression and prognosis.Renal cancer has seldom been linked to BCAT1.Only one researcher found that BCAT1 was co-expressed in hypertension and renal cell carcinoma by bioinformatic analysis and may be a key gene in hypertension-associated renal cell carcinoma [60] .
The pro-carcinogenic mechanism of BCAT1 varies in different tumors.BCAT1 may drive the ammoni cation of BCAA in the circulation, causing these cancer cells to accumulate BCAA [8] , bind to Sestrin2, activate the mTORC1 signaling pathway, phosphorylate downstream effector molecules, and regulate important cellular biological processes [61] .It has been reported that EZH2 inactivation and oncogenic NRAS mutations together activate BCAT1, enhance BCAA metabolism and mTOR signaling, and promote the development of myeloid leukemia [9] .Luo et al. also reported that BCAT1 induces mTORmediated autophagy and reduces the sensitivity of hepatocellular carcinoma cells to cisplatin [15] .In gastric cancer, BCAT1 acts as an oncogene by activating the PI3K/AKT/mTOR signaling pathway to promote proliferation, invasion, and angiogenesis [62] .BCAT1 activates the mTORC1 pathway to play a pro-oncogenic role in breast and endometrial cancers, while BCAT1 inhibits mitochondrial reactive oxygen species (ROS) production in breast cancer cells [27,28,33] In the present study, we found that BCAT1 was enriched in pathways such as PI3K/AKT and WNT in KIRC by GSEA pathway enrichment analysis.Several reports have shown that EMT can be regulated by regulating PI3K/AKT/mTOR or WNT signaling pathways [63][64][65][66][67] , thereby affecting tumor invasion and metastasis.
Increased BCAA levels promote tumor growth, while increased BCAA catabolism also promotes tumor growth.In glioma cells, BCAA degradation by BCAT1 and increased glutamate formation provide the necessary nitrogen source for the synthesis of some non-essential amino acids (aspartate, serine, and alanine) as well as nucleotides, leading to increased proliferation, migration, and invasiveness of tumor cells in vitro [12] .In contrast, in acute myeloid leukemia, glioma, and lung cancer, massive expression of BCAT1 reduces the level of its substrate α-KG and promotes tumor progression [4,11,12,37] .This may be due to the fact that a signi cant decrease in a-KG affects the activity of α-KG-dependent dioxygenases, which play an important role in hypoxic response and epigenetics.
Epigenetically, the pattern of methylation that occurs in the promoter region of BCAT1 is highly correlated with disease onset and progression.In colorectal cancer, researchers found signi cantly higher levels of BCAT1 methylation in circulating tumor DNA (ctDNA) than in normal controls [19,20,[68][69][70][71] .In a related study in ovarian cancer, BCAT1 was signi cantly hypermethylated in low malignant potential (LMP) and high grade (HG) plasmacytoid epithelial ovarian tumors [32] .And investigators studying adverse outcomes in non-alcoholic fatty liver disease showed that high BCAT1 expression and hypomethylation predicted an increased incidence of adverse outcomes such as hepatocellular carcinoma (HCC) [72] .Other investigators have reported increased BCAT1 expression and decreased BCAT1 promoter methylation levels in most hepatocellular carcinomas [73] .
In our study, the results of cBioPortal analysis showed that the methylation level of the promoter region of BCAT1 was signi cantly and negatively correlated with the mRNA expression of BCAT1.And by using the UALCAN tool, we found that the methylation level of BCAT1 was decreased in KIRC tumor tissues.Hypomethylation of BCAT1 was also highly correlated with factors such as tumor progression and advanced age in various clinical features.
The cBioPortal analysis showed that BCAT1 gene mutation caused a signi cant decrease in OS and DFS in KIRC patients, making the prognosis worse.In contrast, among KIRC patients in the TCGA database, we plotted Kaplan-Meier curves showing shorter OS, PFI and DSS in the BCAT1 high-expression group compared with the low-expression group.roc curves showed excellent prognostic predictive value of BCAT1.
In our analysis of TCGA data using UALCAN, we found that BCAT1 was highly expressed in the tumor tissues of KIRC patients and that patients with high BCAT1 expression had higher tumor grades and more lymph node metastases.BCAT1 expression also differed between sexes and molecular subtypes.cBioPortal also showed that BCAT1 expression was upregulated in KIRC cell lines.We collected clinical tissue samples and veri ed the high expression of BCAT1 in KIRC by Western blot and immunohistochemical staining.
Then, we veri ed the relationship between BCAT1 and EMT pathway in KIRC by in vitro experiments.eMT was promoted in KIRC cells after BCAT1 overexpression, and both migration and invasion abilities were signi cantly enhanced.eMT was inhibited in KIRC cells after BCAT1 silencing, and migration and invasion abilities were decreased.
We constructed the PPI network.The 20 and 50 genes that most signi cantly interacted with BCAT1 were identi ed using GeneMANIA and STRING databases, respectively.
In this study, we also analyzed the role of BCAT1 in the immune microenvironment.High BCAT1 expression was highly correlated with the activation of several immune-related cells in tumors.The immune system became more active in BCAT1 mutant cancers.We also found that BCAT1 expression was strongly associated with increased abundance of Treg cells, Th2 cells, and other cells.ssGSEA analysis showed the correlation between high BCAT1 expression in KIRC and the degree of enrichment of common immune cells, which was positively correlated with the enrichment of macrophages, Treg, and Th2 cells, while cytotoxic cells and the CD56bright subpopulation of NK cells were negatively correlated.
Treg cells have been reported to suppress aberrant immune responses against autoantigens and antitumor immune responses.Large in ltration of Treg cells in tumor tissue is usually associated with poor prognosis [74] .There is a drift in the Th1/Th2 balance in a variety of cancers, often to a Th2-dominant state, which may be associated with immune escape [75] .M2-type macrophages have also been reported to be involved in immune escape of tumor cells [76] .In contrast, NK cells and cytotoxic cells are thought to kill cancer cells [77,78] .
Our study has some limitations, but it also provides clues for in-depth study of speci c mechanisms.

Conclusions
To conclude, BCAT1 promotes the invasion and metastasis of KIRC through EMT and has prognostic predictive value with potential as a biomarker.Furthermore, some clues related to the signalling pathway and immune microenvironment have been identi ed.This study presents a novel direction and ideas to investigate the mechanisms underlying the occurrence and development of KIRC.

Figure 1 Annotation
Figure 1

(
B) Visualization by showing the top 50 interacting genes.(C) GSEA results showing the pathway of BCAT1 enrichment in KIRC.

Figure 10 Factors
Figure 10