Expression and Gene Regulation Network of Metabolic Enzyme Phosphoglycerate Mutase Enzyme 1 in Breast Cancer Based on Data Mining

The metabolic enzyme phosphoglycerate mutase enzyme 1 (PGAM1) is a key enzyme in the glycolysis pathway, and glycolysis is closely related to cancer progression, suggesting that PGAM1 may have important functions in breast cancer. We used sequencing data from the Oncomine database and UALCAN database to analyze the expression of PGAM1 and its in ﬂ uence on the clinicopathological characteristics of breast cancer. LinkedOmics was used to identify genes related to PGAM1 expression, kinases, miRNAs, and transcription factors that were signi ﬁ cantly related to PGAM1 through GSEA. cBioPortal was used to identify the alternation frequency and form of PGAM1 in breast cancer. The expression level of PGAM1 in breast cancer was signi ﬁ cantly higher than that in normal tissues. Moreover, the expression level of PGAM1 is closely related to the molecular subtype and TP53 mutation status. The expression level of PGAM1 in HER2-positive and triple-negative tumors was signi ﬁ cantly higher than that of luminal type. The expression level of PGAM1 in TP53-mutant tumors was higher than that in non-TP53-mutant tumors. In addition, the overall survival of patients with high PGAM1 expression was signi ﬁ cantly worse than that of patients with low expression ( P = 0 : 0077 ). Through GSEA analysis, we found multiple kinases, miRNAs, and transcription factors signi ﬁ cantly related to PFKFB4. cBioPortal analysis showed that the mutation rate of PGAM1 in breast cancer was relatively low (4%), and the main form of mutation was high mRNA expression. This study suggests that PGAM1 is a potential diagnostic and prognostic marker in breast cancer. Through data mining, we revealed the potential regulatory network information of PGAM1, laying a foundation for further research on the role of PGAM1 in breast cancer.


Introduction
Breast cancer is the most common malignant tumor in women in the world and the leading cause of cancer-related deaths in women [1]. Although the survival of early breast cancer has been significantly improved, there are still some patients who subsequently relapse and metastasize. The development of various targeted drugs has prolonged the survival time of patients and has made breakthrough progress in the treatment of advanced breast cancer. However, patients with advanced breast cancer inevitably develop pri-mary or continued resistance to targeted drugs. The pathogenesis of breast cancer is extremely complex, involving processes such as cell cycle regulation and signal transduction, reflecting the function and interaction of multiple genes in multiple steps. Therefore, identifying more molecular markers for breast cancer is expected to develop more new molecular targeted therapeutic drugs.
Phosphoglycerate mutase enzyme 1 (PGAM1) is a vital glycolytic protein that catalyzes the reversible reaction of 3phosphoglycerate (3-PG) and 2-phosphoglyceride (2-PG) [2,3]. The regulatory role of PGAM1 in aggressive tumors   3 BioMed Research International has received increasing attention in recent years. To date, many studies have shown that PGAM1 is highly expressed in various tumors such as pancreatic cancer [3], oral squamous cell carcinoma [4], and hepatocellular carcinoma [5]. These studies indicate that PGAM1 is a novel oncogene that may have a regulatory role in breast cancer. However, the expression of PGAM1 in breast cancer and its effect on the prognosis are currently unclear. Therefore, we studied PGAM1 expression and mutations in breast cancer patient data in The Cancer Genome Atlas (TCGA) and various public databases. Using multidimensional analysis, we analyzed PGAM1-related genomic changes and functional networks in breast cancer. Therefore, our results may reveal new molecular targets for breast cancer diagnosis and treatment.

Materials and Methods
2.1. Oncomine Analysis. Oncomine (http://www.oncomine .org) has 715 gene expression datasets and data from 86,733 cancer tissues and normal tissues [6]. It is currently a widely used bioinformatics data analysis platform. Oncomine integrated RNA-seq and DNA-seq data from GEO database, TCGA database, and published literature. After logging in Oncomine, we can see a search box and filter on the left side of the webpage. The filter catalog is divided into several parts, including "Primary Filters", "Sample Filters", "Dataset Filters", and "Concept Filters". Primary Filters can be used to select analysis types, datasets, data sources, cancer types, etc. For example, in order to analyze the expression of PGAM1 in breast cancer, selecting "Breast cancer", "Cancer vs. Normal", and "Clinical Specimen" in the data filter is to know the expression level of PGAM1 in breast cancer tissues and normal tissues. We analyzed the mRNA expression of PGAM1 in breast cancer in multiple cohorts in the Oncomine 4.5 database, including TCGA breast cancer, Curtis breast cancer, Ma Breast 4, and Farmer Breast datasets. The difference in the expression of PGAM1 in breast cancer tissue and the corresponding normal tissue was evaluated, and the difference related to P < 0:05 was considered significant. Fold change was used for differential expression analysis.
2.2. ENCORI Analysis. ENCORI database (http://starbase .sysu.edu.cn/index.php) is an open platform for studying miRNA-ncRNA, miRNA-mRNA, ncRNA-RNA, RNA-RNA, RBP-ncRNA, and RBP-in CLIP-seq, degradome-seq, and RNA-RNA interaction group data. Combining gene expression data of 32 cancers derived from 10882 RNA-seq and 10546 miRNA-seq data, ENCORI allows researchers to perform pan-cancer analysis of RNA-RNA and RBP-RNA interactions. ENCORI also provides a platform for survival and differential expression analysis of miRNA, lncRNA, pseudogenes, and mRNA. In this study, ENCORI database was used to further verify the expression level of PGAM1 in breast cancer compared with normal tissues. At the same time, the ENCORI database was used to evaluate the effect     Positively correlated significant genes 2.3. UALCAN Analysis. UALCAN (http://ualcan.path.uab .edu) includes RNA-seq data of 31 cancer types in the TCGA database and has corresponding clinical pathological characteristics [7]. It is mainly based on the relevant cancer data in the TCGA database for biomarker identification, expression difference analysis, survival analysis, etc. The analysis platform can be used to analyze the relationship between a single gene and cancer stage, tumor grade, age, or other clinicopathological characteristics. This study used the UALCAN platform to analyze the association between PGAM1 and the clinicopathological characteristics of breast cancer, including molecular subtype, stage, and TP53 mutation status. For example, in the "Analysis" module, we enter "PGAM1" in the input box, select the cancer type as "breast cancer", and finally click the "Explore" button. For the input genes, relevant analyzable items will be jumped out, as well as information links of genes in other databases.

LinkedOmics
Analysis. The LinkedOmics database (http://www.linkedomics.org/login.php) is a web-based platform that can be used to analyze 32 TCGA cancer-related cubes [8]. It usually includes five steps, namely, "Select Cancer Cohort", "Select Search Dataset", "Select Search Dataset Attribute", "Select Target Dataset", and "Select Statistical Method". During the analysis, we selected "TCGA_BRCA", "HiSeq RNA", "PGAM1", "RNAseq", and "Pearson Correlation test", respectively. In this study, the LinkFinder module was used to analyze the differentially expressed genes related to PGAM1 in the breast cancer cohort of the TCGA database, and the genes Negatively correlated significant genes   The results were displayed in volcano maps and heat maps. We used LinkFinder module gene set enrichment analysis (GSEA) to perform GO (cell component (CC), biological process (BP), and molecular function (MF)) analysis of differentially related genes, KEGG signal pathway analysis, kinase target enrichment, miRNA target enrichment, and transcription factor target enrichment.
2.5. GeneMANIA Analysis. GeneMANIA (http://www .genemania.org) was used to construct a protein-protein interaction (PPI) network, generate hypotheses about the function of PGAM1 gene, and analyze protein networks that may interact with PGAM1 protein. The PPI network constructed by GeneMANIA is based on many publicly available large-scale biological datasets to find related genes. These protein-protein interactions include physical interactions, coexpression, predicted potential effects, cross-linking between signaling pathways, and colocalized expression in cells. But GeneMANIA does not provide quantitative values between related proteins. These proteins include possible direct or indirect interactions with PGAM1 protein, similar to coexpression, colocalization, or direct binding.

34
BioMed Research International 2.6. cBioPortal Analysis. cBioPortal is an open-access network analysis database that can be used for comprehensive exploration of genomics data of a variety of cancers [9]. The cBioPortal website currently stores DNA copy number data (assuming discrete values for each gene, such as "deep deletion" or "amplification" and log 2 levels), mRNA and microRNA expression data, nonsynonymous mutations, protein levels and phosphoprotein levels (RPPA) data, DNA methylation data, and limited clinical data. cBioPortal greatly reduces the access barriers between complex genomic data and cancer researchers. It facilitates quick, intuitive, and high-quality access to molecular profiles and clinical prognostic correlations of large-scale cancer genome projects. In this study, we used cBioPortal to analyze the frequency and type of PGAM1 gene alternation in 817 breast cancer specimens in TCGA data.

Cancer Cell Line Encyclopedia (CCLE) Analysis. Cancer Cell
Line Encyclopedia database (https://portals .broadinstitute.org/ccle) is an online database jointly developed by Broad Institute and Novartis Research Foundation. At present, the database has collected and visualized the genetic information of more than 1100 cell lines, including copy number and mRNA expression (RNAseq).

PGAM1 Expression in Breast Cancer.
We initially evaluated PGAM1 mRNA levels in multiple breast cancer studies from TCGA and Gene Expression Omnibus (GEO). The analysis flowchart was shown in Supplementary Figure 1. The data in TCGA database showed that the mRNA expression of PGMA1 in breast cancer tissues was significantly higher than that in normal tissues (Figure 1(a), P < 0:001). The Curtis Breast (Figure 1(b), P < 0:001) and Ma Breast 4 (Figure 1(c), P = 0:016) databases further validated that the expression level of PGAM1 in breast cancer tissue was significantly higher than that in breast tissue. Therefore, PGAM1 expression can be used as a potential diagnostic indicator for breast cancer. In addition, we also observed that the expression of PGAM1 in the basal-like subtype was significantly higher than that of the luminal-like subtype (Figure 1(d), P = 0:045). In another database, we also found that the expression level of PGAM1 in breast cancer was significantly higher than that in normal tissues (Figure 1(e), P < 0:001).In addition, the overall survival of patients with high PGAM1 expression was significantly worse than that of patients with low expression (Figure 1(f), P = 0:0077). Finally, we evaluated (y)   Figure 2).
We then analyzed the relationship between the expression level of PGAM1 and the clinicopathological characteristics of patients, and found that PGAM1 was significantly related to molecular subtypes and TP53 mutation status (Figure 2). The expression of PGAM1 in HER2-positive and triple-negative tumors was significantly higher than that in luminal breast cancer (Figure 2(b), P < 0:001). The expression level of PGAM1 in TP53-mutant breast tumors was significantly higher than that in non-TP53-mutant tumors (Figure 2(d), P < 0:001). However, there was no significant difference in the expression level of PGAM1 between tumor tissues of different N and TNM stages.

Analysis of GO and KEGG Pathways of Breast Cancer
Coexpressed Genes Related to PGAM1. The functional module of LinkedOmics was used to analyze the mRNA sequencing data of 526 breast cancer patients in TCGA. As shown in the volcano map (Figure 3(a)), 1925 genes (dark red dots) showed a significant positive correlation with PGAM1, while 1427 genes (dark green dots) showed a significant negative correlation (false discovery rate ðFDRÞ < 0:01). The heat maps, respectively, show 50 important gene sets that are positively and negatively correlated with PGAM1 (Figures 3(b)  and 3(c)). Among them, ACTR1A, PGD, RRM2, NUP93, and SLC25A5 are the top 5 genes with the most significant positive correlation, while RBM16, ZMAT1, SF1, FLJ21062, and DMTF1 are the top 5 genes with the most significant negative correlation. Similarly, we analyzed the mRNA levels of ACTR1A, PGD, RRM2, NUP93, and SLC25A5 through the CCLE database (Supplementary Figure 3). In addition, the mRNA expression levels of RBM16, ZMAT1, SF1, and DMTF1 were also analyzed by the CCLE database (Supplementary Figure 4). These results indicate that PGAM1 may have an extensive gene regulatory network.
Biological process analysis showed that the biological functions of differentially expressed genes positively related to PGAM1 focused on antigen processing and presentation, chromosome separation, granulocyte activation, mitotic cell cycle phase transition, ribonucleotide metabolism, and mitochondrial transport (Figure 4(a)). Cell component analysis showed that differentially expressed genes related to PGAM1 played a structural role in the inner mitochondrial membrane, the intrinsic components of the organelle membrane, and condensed chromosomes (Figure 4(b)). Molecular function analysis showed that the biological functions of differentially expressed genes positively related to PGAM1 were concentrated in electron transfer activity, nucleotidyl transferase activity, nuclease activity, etc. (Figure 4(c)). KEGG analysis showed that differentially expressed genes positively related to PGAM1 were significantly enriched in phagocytosis, glutathione metabolism, Toll-like receptor signal transduction pathways, lysosomes, and human cytomegalovirus infection pathways (Figure 4(d)).

Frequency and Type of PGAM1 Gene Alternation in
Breast Cancer. Then, based on the sequencing data from breast cancer patients in the TCGA database, we used cBio-Portal to determine the type and frequency of PGAM1 changes in breast cancer. Of the 817 (4.2%) breast cancer patients, 34 had PGAM1 gene alternation ( Figure 6). These changes include mRNA upregulation in 22 cases (16%), downregulation in 11 cases (10%), and 1 case (5%) of deep deletion. Therefore, upregulation of mRNA is the most common type of PGAM1 alteration in breast cancer.

Discussion
Breast cancer is the most common malignant tumor in women worldwide. Although the prognosis of breast cancer has been significantly improved in recent years, most breast cancers still inevitably recur and metastasize even after comprehensive treatment. Currently, there is an urgent clinical need to identify more molecular markers for accurate diagnosis and prognosis prediction of breast cancer. As we all know, even when there is sufficient oxygen, cancer cells still metabolize glucose for energy through glycolysis. Tumortargeted glycolysis is considered a potential therapeutic strategy for tumor targeting. PGAM1 is a key enzyme for glycolysis, and a large number of studies suggest that PGAM1 is closely related to the malignant progression of tumors [10][11][12][13]. However, whether PGAM1 has a regulatory role in breast cancer remains unclear. In order to gain a deeper understanding of the diagnostic and prognostic value of PGAM1 in breast cancer, we conducted a bioinformatics analysis of the public sequencing database on the effect of PGAM1 on the clinicopathological characteristics and prognosis of breast cancer, and discussed the correlation of PGAM1 in breast cancer. The regulatory network laid the foundation for further elucidating the mechanism of PGAM1 in breast cancer.
We first investigated the correlation between the expression level of PGMA1 and the clinicopathological characteristics of breast cancer patients. Analysis of transcript sequencing data from 137 clinical samples in the TCGA database found that the level of PGAM1 mRNA in breast cancer was significantly higher than that in normal breast tissue. The Curtis Breast and Ma Breast 4 databases further validated that the expression level of PGAM1 in breast cancer tissue 36 BioMed Research International was significantly higher than that in breast tissue. These results strongly suggest that PGAM1 can be used as a diagnostic biomarker for breast cancer. Subsequently, we also found that PGAM1 was closely related to aggressive clinicopathological features. For example, we found that the expression level of PGAM1 in HER2-positive and triple-negative subtypes was higher than that of luminal subtypes. Interestingly, we also found that the expression of PGAM1 in TP53-mutant tumors was significantly higher than that in non-TP53-mutant breast cancers. As a tumor suppressor gene, TP53 is the most common mutant gene in breast cancer [14]. TP53 mutation can lead to cell cycle arrest, apoptosis, metabolism, DNA repair, and cell senescence of breast cancer cells [15,16]. In addition, TP53 mutations are also associated with drug resistance and poor prognosis of breast cancer [17,18]. Therefore, PGAM1 is likely to exert its cancerpromoting effect through TP53-dependent signaling pathways, but the specific mechanism is worthy of further study. Finally, survival analysis further confirmed that the expression of PGAM1 was significantly related to the overall survival (OS) of patients. This shows that PGAM1 can not only be used as a biomarker for breast cancer diagnosis but also as a biomarker for predicting the prognosis of breast cancer.
We used GSEA to perform enrichment analysis on PGAM1 and obtained multiple kinases, miRNAs, and transcription factors that are significantly associated with it.
Our results indicate that the functional network of PGAM1 is mainly involved in metabolism, biological regulation, response to stimuli, and cell proliferation. We found that PGAM1 in breast cancer was related to kinase networks including CDK1, PLK1, and CHEK1. These kinases could participate in the regulation of mitosis, DNA damage, and intracellular signal transduction. CDK1 is an essential cyclin-dependent kinase, which regulates the cell cycle by interacting with specific cell cycle regulator cyclins. CDK1 expression is elevated in a variety of cancers, often leading to unrestricted proliferation of malignant tumor cells [19][20][21]. PLK1 is a serine threonine kinase that plays a vital role in centrosome maturation, mitotic chromosome separation, and mitosis, and is closely related to the occurrence and development of malignant tumors [22][23][24]. As a kinase, CHEK1 also plays an important role in tumor progression by regulating the cell cycle [25,26].
Abnormalities in these pathways are closely related to tumor progression. In addition, we also identified several miRNAs related to PGAM1. These small RNAs could participate in the posttranscriptional regulation of gene expression and then affect tumor progression. A large number of miR-NAs have been reported to be related to tumor proliferation, apoptosis, cell cycle, invasion, metastasis, drug resistance, and angiogenesis. In fact, miR-382, miR-488, and miR-453 can be used as diagnostic and prognostic markers for malignant tumors.
Considering the significant correlation between PGAM1 and the prognosis of patients, PGAM1 may play an important regulatory role in the malignant progression of breast cancer. We then explored the statistically relevant genes in the expression of PGAM1. Correlation analysis suggests that PGAM1 is positively correlated with the expression of some oncogenes and is a key regulator that affects cancer proliferation, invasion, metastasis, and patient survival. These results further suggest that PGAM1 can promote the malignant progression of breast cancer by interacting with other oncogenes.
Our research inevitably has some limitations. First of all, our results rely on online databases, but the database provides less information about patient treatment options, so we cannot explore whether treatment has an impact on gene expression. Secondly, although our results indicate that PGAM1 may be a potential diagnostic and prognostic marker for breast cancer. We have not conducted in vivo and in vitro studies to verify whether PGAM1 is a true oncogene. The results of this study still need to be verified by a large number of experiments. Thirdly, the specific role and molecular mechanism of PGAM1 in breast cancer have not been thoroughly explored in this study, and more molecular biology studies are needed to further explore the molecular details of PGAM1. Fourthly, GeneMANIA does not provide quantitative values between related proteins. Finally, the impact of PGAM1 on survival was mainly based on univariate analysis, without adjusting for confounding factors such as age.
In conclusion, this study provides preliminary evidence for PGAM1 as a potential marker for the prognosis of breast cancer. At the same time, our results indicate that PGAM1 is significantly associated with several tumor-associated kinases (such as CDK1 and PLK1), miRNA (such as miRNA-382), and transcription factors (such as RSRFC4) in breast cancer tissues. However, the analysis results of this study based on data mining still need to be verified by more studies, including functional tests and molecular mechanisms, which will help to further clarify the regulatory role of PGAM1 in breast cancer.

Data Availability
All data generated or analyzed in this study are included in this article.