Expression of GOT2 Is Epigenetically Regulated by DNA Methylation and Correlates with Immune Infiltrates in Clear-Cell Renal Cell Carcinoma

Clear cell renal cell carcinoma (KIRC) is the most common and highly malignant pathological type of kidney cancer, characterized by a profound metabolism dysregulation. As part of aspartate biosynthesis, mitochondrial GOT2 (glutamic-oxaloacetic transaminase 2) is essential for regulating cellular energy production and biosynthesis, linking multiple pathways. Nevertheless, the expression profile and prognostic significance of GOT2 in KIRC remain unclear. This study comprehensively analyzed the transcriptional levels, epigenetic regulation, correlation with immune infiltration, and prognosis of GOT2 in KIRC using rigorous bioinformatics analysis. We discovered that the expression levels of both mRNA and protein of GOT2 were remarkably decreased in KIRC tissues in comparison with normal tissues and were also significantly related to the clinical features and prognosis of KIRC. Remarkably, low GOT2 expression was positively associated with poorer overall survival (OS) and disease-free survival (DFS). Further analysis revealed that GOT2 downregulation is driven by DNA methylation in the promoter-related CpG islands. Finally, we also shed light on the influence of GOT2 expression in immune cell infiltration, suggesting that GOT2 may be a potential prognostic marker and therapeutic target for KIRC patients.


Introduction
Renal cell carcinoma (RCC) is one of the most common malignancies of the genitourinary system, representing 3% of all malignancies in adults [1]. In the subtypes of kidney cancer, kidney renal clear cell carcinoma (KIRC) accounts for about 75% of all RCC [2][3][4]. Indeed, due to the evident complexity from both morphological and molecular points of view [5][6][7][8][9], KIRC patients have heterogeneous clinical outcomes [10,11]. Most KIRC tumors are radiotherapy and chemotherapy-resistant, and~30% of patients eventually develop metastases [3,12,13]. Thus, it is critical to identify new sensitive tumor biomarkers to advance the prognosis of KIRC.
The metabolic shift in KIRC tumors covers different pathways and specific intermediates (e.g., amino acids, aerobic glycolysis, and fatty acid metabolism) [16,24], allowing cancer cells to rapidly proliferate, and survive during nutrient depletion and hypoxia, and evade the immune system [14]. Supporting this notion, aberrant tumor growth is promoted by an enhanced supply of specific metabolites, and some of them, such as aspartate (Asp), are limiting in some tumors [25][26][27]. Asp is usually synthesized in the mitochondrial matrix through the sequential actions of MDH2 and glutamic-oxaloacetic transaminase 2 (GOT2) and then transported to the cytosol for use by GOT1 and other enzymes [28].
GOT2, situated on chromosome 16q21, is a crucial enzyme for cancer cell metabolism, (i) mediating the reversible interconversion of oxaloacetate and glutamate into aspartate and α-ketoglutarate, providing energy for tumor cells (Krebs cycle) [29]; (ii) being a key transfer enzyme in the malate-aspartate NADH shuttle activity and oxidative protection [30], maintaining glycolysis, and (iii) participating in the amino acid metabolism of tumor cells [31]. Increasing evidence has shown that dysregulation of GOT2 expression significantly influences tumor growth and the prognosis of several human neoplasms [30,[32][33][34][35][36][37][38]. However, the role of GOT2 in the development and prognosis of KIRC has not been reported. To address these issues, this study aims to evaluate the expression levels of GOT2 in KIRC and determine its epigenetic modulation, prognostic value, and correlation with tumor-infiltrating immune cells in KIRC patients through multiple databases.

Differential Expression of GOT2 mRNA and Protein
Initially, pan-cancer analysis of GOT2 transcription levels was performed via the TIMER2.0 database (Tumor Immune Estimation Resource, http://timer.cistrome.org/, accessed on 1 December 2021) [39], using the differential expression module across all TCGA tumors. The statistical significance computed by the Wilcoxon test was annotated by the number of stars (* p-value < 0.05; ** p-value < 0.01; *** p-value < 0.001).
To keep the focus of our analyses on KIRC tumors, the GEPIA2 database (Gene Expression Profiling Interactive Analysis 2) (http://gepia.cancer-pku.cn/index.html, accessed on 1 December 2021) [40] was used to confirm the differential expression found in the TIMER analysis by comparing the TCGA-KIRC (523 samples) with normal kidney samples from GTEx (Genotype-Tissue Expression project, http://www.gtexportal.org/home/index.html, accessed on 1 December 2021) (100 samples). The differential threshold of log2FC was 1 and the value cutoff of 0.05.
At the protein level, we used the UALCAN platform (http://ualcan.path.uab.edu/ index.html, accessed on 1 December 2021) [41,42] to mine GOT2 expression in the highthroughput mass spectrometry data, obtained from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) of normal kidney tissues (n = 84) and primary KIRC tumors (N = 110) [43]. Integration and analysis of these data were described elsewhere [44,45]. Briefly, protein expression values (Log2 Spectral count ratio values) from CPTAC were first normalized within each sample profile, then normalized across samples. Then Z-values for each sample for GOT2 protein were calculated as standard deviations from the median across samples.
In this study, we checked the expression of GOT2 in the protein expression module of the HPA database (Human Protein Atlas, https://www.proteinatlas.org/, accessed on 1 December 2021) [46][47][48], and we analyzed the immunohistochemical results of GOT2 in tumor tissue (ID: 2176) and normal tissue (ID: 2067). The antibody used in both samples was HPA018139. All images of tissues stained by immunohistochemistry were manually annotated by a specialist, followed by verification by a second specialist. Protein expression score was based on immunohistochemical data manually scored concerning staining intensity (negative, weak, moderate, or strong) and the fraction of stained cells (<25%, 25-75% or >75%). Each combination of intensity and fractions was automatically converted into an protein expression level score as follows: negative-not detected; weak <25%-not detected; weak combined with either 25-75% or 75%-low; moderate <25%-low; moderate combined with either 25-75% or 75%-medium; strong <25%-medium, strong combined with either 25-75% or 75%-high (For more details, see https://www.proteinatlas.org/ about/assays+annotation#ihk, accessed on 1 December 2021).
The prognosis analysis was estimated by Kaplan-Meier (KM) survival curves generated by the Kaplan Meier (KM) Plotter (http://kmplot.com/analysis/, accessed on 5 December 2021) [50], GEPIA2 (http://gepia.cancer-pku.cn/index.html, accessed on 1 December 2021) [40] and HPA database (http://www.proteinatlas.org/, accessed on 5 December 2021) [46][47][48]. In this study, KIRC patients were split into high and low expression groups based on the median expression levels of GOT2, and then these two groups were compared in terms of relapse-free survival. Moreover, the hazard ratio (HR) with a 95% confidence interval (CI) and the p-value of the log-rank test were obtained. For all survival analyses, p < 0.05 was considered statistically significant.

Analysis of Immune Cell Infiltration
We calculated and compared the GOT2 gene expression contributed by different immune cell types in kidney samples (TCGA tumor/normal and GTEx normal) by the GEPIA2021 (http://gepia2021.cancer-pku.cn/, accessed on 1 December 2021) [53]. For each GTEx/KIRC-TCGA sample, we run the CIBERSORT algorithm (absolute mode) with the default parameters to obtain the absolute proportions of 22 immune cell subtypes. The

Association between GOT2 and Tumor Microenvironment Exploration
The Tumor Immune Single-cell Hub (TISCH, http://tisch.comp-genomics.org/home/, accessed on 1 April 2022) is an online single-cell RNA-seq database focused on the tumor microenvironment (TME) [54]. In our analyses, two human KIRC scRNA-seq datasets [55,56] were used to obtain the GOT2 average expression at the single-cell level. The expression of GOT2 was collapsed by the mean value. The gene expression level displayed using UMAP and violin plots was quantified by the normalized values.

GOT2 mRNA and Protein Are Downregulated and Correlated with Clinicopathological Parameters in KIRC
Next, to focus our analysis on KIRC, we investigated the transcription levels of GOT2 performing a single-gene differential analysis using RNA-seq data from the TCGA database (KIRC-TCGA), compared with non-tumor tissues from the GTEx database by GEPIA2. Our results showed that the mRNA expression levels of GOT2 in KIRC tissues (n = 523) were significantly lower than in adjacent normal tissues (n = 100) ( Figure 2A). Correspondingly, in the CPTAC KIRC cohort, there was a significant downregulation of GOT2 in the tumors ( Figure 2B), consistent with the immunohistochemical (IHC) staining images from the Human Protein Atlas (HPA) ( Figure 2C). This further confirmed that the expression of GOT2 in tumor tissues was significantly lower than that in normal tissues.

Relationship between GOT2 Expression and Clinical Pathological Parameters of Patients with KIRC
We next investigated the correlation between clinical parameters and the GOT2 expression in KIRC. Data showed that GOT2 expression levels were significantly associated with stage, gender, grade, KIRC subtypes, and nodal metastasis status ( Figure 3A-G). Lastly, concerning the most commonly mutated genes in KIRC, patients harboring VHL, PBRM1, and SETD2 mutations under-expressed GOT2 ( Figure 3H-J). Thus, it is likely that GOT2 expression may serve as a potential diagnostic biomarker for KIRC patients.

Low Expression of GOT2 Is Associated with Poor Outcome in KIRC Patients
Initially, to explore the influence of GOT2 expression on KIRC outcomes, we conducted a Kaplan-Meier test and Cox regression analysis to delve into the associations with overall survival (OS) and disease-free survival (DFS). As shown in Figure 4A,B, the OS and DFS of KIRC patients with low expression of GOT2 were significantly shorter than those with high expression. At the same time, we also noticed that the low level of GOT2 protein was significantly related to the worse OS (p = 0.023) ( Figure 4C). Additionally, we investigated the relationship between GOT2 expression and clinicopathological features of KIRC patients in the Kaplan-Meier plotter database. Surprisingly, low GOT2 mRNA expression was correlated with worse OS in KIRC patients with stage 4 (HR = 0.56, p = 3.50 × 10 −2 ), grade 3 (HR = 0.53, p = 7.90 × 10 −3 ), and low mutation burden (HR = 2.28, p = 3.49 × 10 −2 ) ( Table 1). Here, the differences in the clinical characteristics suggest that the use of GOT2 as an indicator gene should be carefully combined with the patient's condition.  Bold numbers indicate a statistically significant correlation with a p-value less than 0.05. Abbreviations: CI = confidence interval.

Hypermethylation of DNA in the Promoter Region Leads to Low Expression of GOT2 in KIRC
To further explore the epigenetic mechanism underlying GOT2 underexpression, we analyzed the methylation level of seventeen probes covering the island (promoter region), N Shelf, S Shore, and Open Sea regions of GOT2, chosen through the UCSC Genome Browser (Table 2; Figure 5). Notably, the results showed that lower methylation levels for GOT2 lay on probes at the promoter (island). At the same time, most hypermethylated sites fell in the open sea, N Shelf, and S Shore regions ( Figure 6). Given that methylation of CpG sites within the gene promoter is a common mechanism in gene silencing, we next compared the methylation level of the probes that covered the GOT2 promoter between normal vs. KIRC-TCGA samples ( Figure 5). Interestingly, we found that the average methylation of all CpG sites (probes) near the TSS (transcription start site) of GOT2 was significantly higher in tumor tissues than in the normal counterpart (Aggregation, p = 0.00022) ( Figure 7A). Further analysis revealed a negative correlation between the methylation level and the mRNA of GOT2 (Aggregation: R = −0.3, p = 0.0071) ( Figure 7B), thus indicating that upregulation of DNA methylation level of CpGs island-associated promoter region may contribute to the downregulation of GOT2 in KIRC patients.

GOT2 Expression Correlates with Immune Cell Infiltration in KIRC
Tumor-infiltrating immune cells are essential for immune response and prognosis in KIRC patients [57,58]. To determine whether GOT2 could potentially impact immune cell infiltration in KIRC, we first examined the differences in GOT2 expression across the six immune subtypes proposed by Thorsson et al. [59]. We observed that GOT2 expression was highest in patients harboring the C5 subtype (immunologically quiet) and lowest in patients exhibiting the C2 subtype (IFN-gamma dominant), indicating that GOT2 can be used as a marker for immunophenotyping of patients with clear-cell renal cell carcinoma ( Figure 8A). To better understand the role of GOT2 in the infiltration of immune cells in KIRC, we used the CIBERSORT deconvolution analysis [53] for rough correlation analysis. The immune-related signatures revealed that GOT2 was higher in CD8 + T cells, follicular helper CD4 + T (Tfh) cells, M1 and M2 Macrophages in KIRC-TCGA tumors than in normal tissues ( Figure 8B). To further expand and strengthen these results, the analysis of two independent single-cell RNA sequencing (scRNA-seq) datasets [55,56] showed that GOT2 was mainly expressed within endothelial cells, followed by proliferative T cells (Tprolif), plasmacytoid dendritic cells (pDCs), exhausted CD8 + T Cells (CD8Tex), Treg cells and conventional dendritic cells 2 (cDC2) (Figure 9A,B). These results imply that GOT2 may play an essential role in the tumor microenvironment of the clear-cell renal cell carcinoma, affecting both stroma and immune cells. Interestingly, GOT2 was broadly expressed within some clusters of immune cells (e.g., CD8ex and Tprolif) that also co-expressed some immune checkpoint inhibitors (e.g., CTLA4, TIGIT, TOX, EOMES, LAG3, PDCD1, HAVCR2, and CD96) (Figure 10), thus strongly suggesting that GOT2 is involved in the dynamic regulation of immune homeostasis and is particularly relevant to T cell functionality.  (Figure 10), thus strongly suggesting that GOT2 is involved in the dynamic regulation of immune homeostasis and is particularly relevant to T cell functionality.

Discussion
KIRC is characterized by profound metabolic reprogramming that involves multiple pathways [14,15]. Current evidence suggests that changes in the supply of specific metabolites, such as aspartate, which is essential for nucleotide and protein synthesis in proliferating cells and maintains the reducing potential [28,60,61], can function as opportunistic fuel sources for high proliferation and tumor growth [25,27,62]. As part of the malate-aspartate shuttle, mitochondrial GOT2 generates aspartate from oxaloacetate and glutamate [63]. Additionally, this enzyme is involved in energy transduction, specifically amino acid metabolism and the urea and TCA cycles. Thus far, there is no available information about the detailed roles of GOT2 in KIRC. Herein, we elucidated the most comprehensive insights into understanding the epigenetic regulation and the potential association of GOT2 with the clinical and immunity of KIRC.
Based on a pan-cancer perspective, we initially demonstrated that GOT2 is differentially expressed in 18 tumor types, thus potentially being a therapeutic target. Further interrogating KIRC, we showed that the GOT2 mRNA and protein levels were markedly decreased in KIRC patients than in normal tissues. Besides, we observed that this transaminase was markedly lower as the pathological stage increased and was also strongly impacted by other clinicopathological characteristics, which conferred a worse outcome. Our findings are consistent with Zhao et al. [64], who also reported the decreased expression and prognostic value of GOT2 in hepatocellular carcinoma (HCC). The results from a recent study support a scenario in which in VHL-deficient KIRC, but not in non-clear renal cell carcinomas (NCRCC), the simultaneous suppression of GOT1 and GOT2 is HIF1α-dependent, which impairs oxidative and reductive aspartate biogenesis [61]. Hence, to compensate for the low levels of aspartate seen in the KIRC [65], glutamine metabolism has a dominant role in sustaining KIRC growth [66]. This conceivably explains the down-expression of GOT2 in KIRC patients harboring VHL, PBRM1, and SETD2 mutations seen in our study, thus suggesting that GOT2 repression represents a specific metabolic feature of KIRC.
DNA methylation of specific CpG sites in the promoter region is tightly linked with transcription repression. In the last few years, its role in carcinogenesis has been of considerable interest [67][68][69]. It is currently well known that KIRC is characterized by many epigenome aberrations [70,71]. Furthermore, many studies have pointed out the occurrence of a pattern known as CpG island methylator phenotype (CIMP) in 20% of KIRC [71,72]. However, no study has previously been carried out to analyze the role of DNA methylation in GOT2 expression in KIRC. Here, for the first time, we provided evidence that the methylation of the GOT2 promoter was increased in KIRC patients compared to normal samples. Additionally, the correlation analysis results revealed that promoter methylation was negatively correlated with the regulation of gene expression. According to these results, it can be speculated that the DNA hypermethylation in the promoter-associated CpG islands may be one of the mechanisms leading to GOT2 down-expression in KIRC. However, additional efforts are necessary to determine the potential impact of additional events, such as chromatin structural modifications, miRNAs, and the influence of metabolites on patients exhibiting GOT2 promoter hypermethylation.
In addition, another innovative aspect of this study clarified the significant correlations between GOT2 expression and various tumor-infiltrating immune cells in KIRC. Previous studies have found that T cells and macrophages represented the dominant populations in most KIRC cases [73,74], consistent with our findings, which indicates that GOT2 expression was more likely to affect the tumor infiltration of subtypes of T cells, especially CD8 + T cells and follicular helper CD4 + T (Tfh) cells, and M1 and M2 macrophages compared to normal renal tissue. Our deep-dissection of individual cell subsets from scRNA-seq data revealed that GOT2 was broadly expressed within exhausted CD8 + T Cells (CD8Tex) and in the proliferative T cells (Tprolif). Unlike all solid tumors, high tumor-infiltrating CD8 + T-cells predicted poor overall survival and inferior therapeutic responsiveness in patients with KIRC [75][76][77]. However, a comprehensive characterization of immune cells from KIRC patients using scRNA-seq along with T-cell-receptor (TCR) sequencing revealed that CD8 + T-cells exhibited four distinct groups that may represent transcriptional states upon tumor infiltration with distinct prognostic significance: two of them were associated with a PD-1 + TIM-3 + exhausted subcluster, one with a proliferative subcluster, and a fourth with the higher levels of cytokine signaling [73]. Moreover, a correlation observed between increased clusters with the signature CD8_6 (CD8 + T-cells) and TAM_3 (macrophages) showed a better prognosis. In another study, a first-in-class CAR T-cell therapy co-expressing GOT2 enhanced T-cell metabolic function for treating GPC3-positive solid tumors, supporting the progress of a future first-in-human trial in subjects with GPC3-positive tumors [78]. Considering this context, we argue that GOT2 is likely to play distinct roles at different stages of T-cell exhaustion and might potentially be modulated by the spectrum of changes in TME conditions of KIRC patients, including tumor metabolism, hypoxia, nutrient restriction, and exhaustion driven by chronic stimulation, thus strengthening the potential application of synergic modulation of the GOT2 and T cell exhaustion markers in nonresponsive KIRC patients to boost antitumor and immune responses.
In conclusion, using a series of rigorous bioinformatics analyses, we showed that the mRNA and the expression levels of the GOT2 protein were significantly decreased in KIRC patients compared to normal ones. This low expression was positively associated with clinicopathological features, culminating in poor clinical outcomes for KIRC patients. Notably, we provide the first mechanism insights into the epigenetic-mediated regulation of GOT2, which is driven by the DNA methylation in the promoter-related CpG islands. Finally, we also shed light on the influence of GOT2 expression in immune cell infiltration, suggesting that GOT2 may be a potential prognostic marker and therapeutic target for KIRC patients.