The Critical Role of PTEN Mutation in Cellular Process and Drug Selection of Endometrial Cancer

Background: Phosphatase and tensin homolog (PTEN) is a frequently mutated genes found in endometrial cancer (EC), making it a potential biomarker for individualized treatment opinions. In this study, a method was designed to evaluate the role of the PTEN mutation in the prognosis and drug selection of EC. We identied the potential alterations in pathways and genes related to the mechanism. Methods: cBioPortal database was used to analyze the PTEN mutation status for EC patients. Kaplan-Meier was used to analyze the prognosis of PTEN mutation in EC patients. GDSC dataset was used to identied the drugs that sensitive to cell lines with PTEN mutation. DEGs between PTEN mutation and wide type group were identied using the edgeR package. GO and KEGG analysis were carried out using the DAVID database. GSEA v3.0 were used to dig out the differences in gene mRNA levels of biological function annotation and pathways between PTEN mutation and wide type patients. PPI network of DEGs was performed using STRING and then visualized using Cytoscope software (3.7.2). Results: Our results showed that PTEN mutation was carried in 68% of EC patients. The mRNA expression level of PTEN was lower in patients with PTEN mutation than that with wide type. Prognosis analysis showed that there were favorable overall survival and progression free survival in EC patients with PTEN mutation. Moreover, it is more sensitive to AKT inhibitor (Afuresertib and AZD5363), and Mcl-1 inhibitor (MIM1) on EC cell lines with PTEN mutation than that with wide type. A total of 216 genes were identied as DEGs. GO analysis showed that DEGs signicantly enriched in chemical synaptic transmission, extracellular region, etc.. KEGG analysis suggested that DEGs signicantly enriched in categories associated with metabolic progression. GSEA analysis identied signaling pathways including fatty acid metabolism, fructose and mannose metabolism, etc.. PPI network analysis identied top 10 genes and top three clusters. Conclusions: Multiple genes and pathways may play an important role in EC patients with PTEN mutation. These results provide a potential target and therapeutic strategies for patients with PTEN mutation. Our results conrmed that mutation rate of PTEN is approximately 68% in EC patients. mRNA expression level of PTEN was lower in patients with PTEN mutation than that with wide type. Moreover, PTEN mutation have a favorable overall survival and progression free survival in EC patients. Cell lines are more sensitive to AKT inhibitor (Afuresertib and AZD5363) and Mcl-1 inhibitor (MIM1) than that with wide type in EC with PTEN mutation. Multiple genes and pathways may play an important role in EC patients with PTEN mutation. These results offer a potential target and strategies for patients with PTEN mutation.


Introduction
Endometrial cancer (EC) is one of the most common malignancies of the female reproductive system [1][2][3]. The incidence of endometria cancer is still increasing as a consequence of the populations increasing life expectancy and the overall prevalence of obesity and metabolic syndromes [4]. The number of new cases and deaths is expected to increase by 17.4 and 20.3 percent by 2025, respectively [5]. The ve years overall survival rate of EC patients is approximately 82% [6]. Mortality of EC patients is associated with poor prognosis factors. Prognostic prediction of EC patients relies on histological stage, tumor grade, depth of myometrial invasion, cervical involvement, tumor size and lymph node status, but they are not always reliable [5,6]. Recently, next-generation sequencing results reveal many genetic mutations in EC, including CTNNB1, FGFR2, HER2, P53, etc. [2,7,8]. These different mutation types result in drug resistance and poor prognosis of EC patients. Therefore, to identify potential biomarkers that effectively predict prognosis and response to drug therapy appears extremely urgent.
Phosphatase and tensin homolog (PTEN) located in 10q22.3, is one of tumor suppressor genes [9]. PTEN is a dual speci city protein and lipid phosphatase that involved in multiple cellular process and signal pathways, such as cell cycle, cell differentiation, migration and tumor invasion [10]. PTEN functions can be either dependent or independent on the negative regulation of the phosphatidylinositol-3-kinase B pathway [11]. PTEN mutation has been found in many cancers, including ovarian cancer, endometrial cancer, gastric cancer and breast cancer [12][13][14][15]. PTEN mutation may affect cancer development, cancer progression, disease prognosis and treatment strategies [16,17], and results in poor prognosis and natural resistance or sensitivity to treatment in patients. Therefore, to explore the alterations in cell process and signal pathways will provide more evidence for individualized treatment in EC patients with PTEN mutation.
In this study, we analyzed the RNA sequencing data of uterine corpus endometrial carcinoma (UCEC) based on The Cancer Genome Atlas (TCGA) to identify the key genes and pathways associated with the PTEN mutation. The Genomic of Drug Sensitivity in Cancer (GDSC) was used to evaluate the role of PTEN mutation in drug selection of EC. The aim is to uncover the potential biomarker on prognostic prediction and individualized treatment of EC patients with PTEN mutation.

TCGA data
Gene expression data and corresponding clinical information were downloaded from TCGA database (https://tcga.xenahubs.net). PTEN mutation data was obtained from cBioPortal for Cancer Genomics (https://www.cbioportal.org) [18]. The mRNA data was provided by TCGA which are public and openended, and therefore does not require the approval of a local ethics committee. Identi cation of differential expressed genes Downloaded mRNA data were analyzed to obtain differentially expressed genes (DEGs) at genes levels between wide type and PTEN mutation using Edger software package. Fold changes (|log2FC|)≥1, both P-value and FDR (false discovery rate) <0.05 were set as the cut-off value.

Function and Pathway enrichment
To determine the function of the DEGs, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment were implemented using the Database for Annotation, Visualization and Integrated Discovery (DAVID) website (https://david.ncifcrf.gov/) [19,20]. A P-value <0.05 was set as the cut-off value.

Correlation between PTEN mutation and drug response
The relation between PTEN mutation and drug response was analyzed using the Genomics of Drug Sensitivity in Cancer (GDSC) (https://www.cancerrxgene.org) [21]. For this analysis, the status of PTEN mutation in UCEC was correlated with drug sensitivity data by an analysis of variance (ANOVA). This includes coding variants and copy number alterations. We rst searched genetic feature 'PTEN' and selected the 'PTEN mut' in UCEC cell lines, and then searched for compounds with signi cant selectivity for the PTEN mutation in UCEC cell lines. Scatter plots were generated and computed via the GDSC online platform.

Gene Set Enrichment Analysis (GSEA)
The association between status of PTEN and biological processes was analyzed using GSEA v3.0 (https://www.gsea-msigdb.org/gsea/downloads.jsp). GSEA calculates a gene set Enrichment Score (ES) that estimates whether genes in pre-de ned gene set are enriched in the PTEN mutation/wide type groups. |ES|>1, p<0.05, and FDR<0.25 were considered to be statistically signi cant.

Protein-Protein Interaction Network
Protein-protein interaction (PPI) network analysis of DEGs was performed by the Search Tool for Retrieval of Interacting Genes (STRING) database (https://string-db.org/). Con dence of 0.4 was set as the minimum required interaction score. Simple tabular text output was generated from STRING, then visualized by Cytoscope software (3.7.2). We used the Cytoscape plugin Molecular Complex Detection (MCODE) plugin for screening the intersected clusters from the PPI network. The cluster determining extremities was charted, such as Kappa score (K-core) xed to ve, Degree Cutoff xed to two, Max. Depth xed to 100, and Node score Cutoff xed to 0.2, which constraints the cluster size for coexpressing networks.

Statistical analysis
All statistical analysis was conducted by R. Wilcox analysis was used to compare the mRNA levels of PTEN between PTEN mutation and wide type group. Kaplan-Meier method with log-rank test was used to analyze the prognosis of PTEN mutation in EC patients. FDR in edgeR and GSEA were adjusted for multiple testing with the Benjamini-Hochberg procedure to control FDR, respectively. P<0.05 and FDR<0.25 were considered to be statistically signi cant.

PTEN mutation in EC patients
PTEN mutation information was downloaded from cBioPortal website, and there were 347 (68%) ( Figure  1A) with PTEN mutation in UCEC patients. Mutation types are included inframe mutation, missense mutation, truncating mutation, fusion, ampli cation and deep deletion ( Figure 1B).

Prognosis of PTEN mutation in EC patients
In order to effect identi cation of PTEN mutation on process and prognosis in EC patients, we compared the mRNA expression level of PTEN between PTEN mutation and wide type group, the result showed that mRNA expression of PTEN was downregulated in EC patients with PTEN mutation (P=0.049, Figure 2A). We further analyzed the relationship between PTEN status and prognosis in EC patients. The result revealed that it was a favorable overall survival (P=8.273e-4, Figure 2B) and progression free survival (P=0.715e-4, Figure 2C) in EC patients with PTEN mutation. These results indicated that PTEN mutation may in uent the progression in EC patients.

Drug selection to EC patients with PTEN mutation
We then investigated the role of PTEN mutation in EC therapy by GDSC website. The results showed that AKT inhibitor (Afuresertib and AZD5363) ( Figure 3A ,3B), Mcl-1 inhibitor (MIM1) ( Figure 3C)exhibited sensitivity for EC cell lines harboring the PTEN mutation in the GDSC website, it indicated that these compounds were potential for individualized treatment of EC patients with PTEN mutation.

Identi cation of DEGs
To identify the key genes and pathways associated with the PTEN mutation, we analyzed the differences between PTEN mutation and wide type group using edgeR software package. |log2FC|≥1, both P-value and FDR<0.05 was set as the cut-off value, and a total of 216 genes were identi ed as DEGs, these DEGs include 64 upregulated genes and 152 downregulated genes, the volcano plot of the DEGs and heatmap of top 50 upregulated genes and downregulated genes were shown in Figure 4.

Functional and Pathway enrichment
To determine the function of the DEGs, GO and KEGG pathway enrichment of all 216 DEGs were implemented using DAVID website, the GO analysis of DEGs showed signi cant enrichment in chemical synaptic transmission, extracellular region, negative regulation of endopeptidase activity, neurotransmitter transport, serine-type endopeptidase inhibitor activity, serotonin receptor signaling pathway, GABA-A receptor complex, defense response to Gram-negative bacterium, neurotransmitter: sodium symporter activity, chloride channel activity, GABA-A receptor activity, voltage-gated potassium channel activity, and neurotransmitter receptor activity ( Figure 5A). KEGG pathway analysis suggested signi cant enrichment in neuroactive ligand-receptor interaction, retinol metabolism, drug metabolism cytochrome P450, metabolism of xenobiotics by cytochrome P450, chemical carcinogenesis, GABAergic synapse, morphine addiction, retrograde endocannabinoid signaling, nicotine addiction, serotonergic synapse and cAMP signaling pathway ( Figure 5B).

GSEA
We further analyzed the in uence of PTEN mutation on cellular process. GSEA was performed to identify signaling pathways that activated in EC patients. It was found that fatty acid metabolism, fructose and mannose metabolism, p53 signaling pathway, terpenoid backbone biosynthesis and alpha linolenic acid metabolism were signi cantly enriched. These results indicated that multiple pathways and metabolism were activated which were involved in progression of EC patients with PTEN mutation (Figure 6).

Establishment of PPI network and module analysis
Finally, we constructed the PPI network of DEGs using STRING, the PPI network was involved in 186 nodes and 132 edges. The top 10 genes ranked by degree were CYG13, CT45A1, NPY, GABRA5, HTR2C, MAGEC1, PASD1, MAGEC2, SSX4 and NXF2B (Figure 7). Regions interlinked closely from the network were identi ed by cytoscape MCODE plugin in the form of clusters. It showed that the top three clusters were signi cant in the PPI network ( Figure 8).

Discussion
PTEN is a multifunctional tumor suppressor gene, which is mutated in large number of cancers at different frequency. PTEN alterations are more frequent in glioblastoma (30%-60%) [22,23], breast cancer(11%-15%) [24,25], prostate cancer (17%-21%) [26,27]. PTEN mutation is thought to be an early step in the development and progression of EC patients [28,29]. It is reported that PTEN mutation is in 30-50% of EC cases [30] [31]. We found that 68% of EC patients harbored PTEN mutation in the present study. These ndings suggested that PTEN is strongly involved in the development and progression of EC patients.
The role of PTEN mutation in prediction of prognosis varies in different cancer types. Previous study reported that PTEN mutation have a signi cantly poorer prognosis of survival and disease recurrence in clear cell renal cell carcinoma (ccRCC) patients [17]. Recent research showed that low expression of PTEN is of limited value in predicting the prognosis of patients with ccRCC for overall survival and progression free survival, but it is signi cantly associated with an unfavorable disease-speci c survival [32]. However, it is also reported that PTEN mutation was correlated with a favorable prognosis for EC patients [9]. In addition, it is indicated that PTEN mutation only outside exons 5-7 was signi cantly associated with better survival [33]. Consistently, our results showed that PTEN mutation have a favorable overall survival and progression free survival in EC patients.
PTEN mutation contributes to natural resistance or sensitivity to chemotherapy. It is found that somatic mutations of PTEN were associated with resistance to immune checkpoint inhibitors in patients with glioblastomas [34]. Moreover, PTEN loss contributes to hyperactivation of the PI3K pathway and to drug resistance in breast cancer [35], and PTEN deletion is closely related to acquired resistance to EGFR-TKIs [36]. Conversely, it was found that GSK690693, a AKT inhibitor, selectively inhibited on ccRCC with PTEN mutation [17]. Our results revealed that PTEN mutation are more sensitive to AKT inhibitor (Afuresertib and AZD5363) in EC cell lines than that in wide type, also Mcl-1 inhibitor (MIM1) does. These results provide evidence for the potential individualized therapy of EC patients with PTEN mutation.
PTEN, a dual speci city protein and lipid phosphatase, is inactivated by PTEN enzymatic activity which leads to induction of cell proliferation and inhibition of cell death, causing cancer development and progression [35]. PTEN can be secreted into the extracellular environment and works as a tumor suppressor in a cell non-autonomous manner [36]. Role of PTEN has recently been established in metabolic reprogramming, including drug metabolism and lipid metabolism. PTEN mutation may affect the expression or activity of cytochrome P450, suggesting the potential role of PTEN in drug response or resistance. Consistently, our results showed that DEGs related to extracellular region, neurotransmitter transport, GABA-A receptor activity, drug metabolism-cytochrome P450, metabolism of xenobiotics by cytochrome P450, chemical carcinogenesis, GABAergic synapse, serotonergic synapse and cAMP signaling pathway in EC patients with PTEN mutation. These results further explain why are more sensitive to AKT inhibitor (Afuresertib and AZD5363) and Mcl-1 inhibitor (MIM1) in EC cell lines with PTEN mutation than that with wide type.
GSEA analysis suggests that PTEN mutation is signi cantly associated with multiple cancers related fatty acid metabolism, fructose and mannose metabolism, p53 signaling pathway, terpenoid backbone biosynthesis and alpha linolenic acid metabolism. Features of cancer cell reprograming are increased in glucose uptake and fermentation of glucose to lactate. Evidence suggested that inactivation of PTEN leads to fatty acid accumulation in cancer cells [37]. Furtherly, loss of PTEN resulting in p53 increase can counteract tumor growth by activating an effective fail-safe mechanism of cellular senescence in prostate cells [38]. Therefore, it is suggested that PTEN mutation is associated with cellular progression and signal pathways.
Overall, our results provide a potential therapeutic targets and strategies for EC patients with PTEN mutation. However, our results should be veri ed by molecular biology experiments even though we investigated the role of PTEN mutation in cellular process and in progression of EC patients by bioinformatic analysis, Additionally, we only analyzed the role of PTEN mutation in drug selection of EC cell lines, it also should be validated by animal experiments, clinical researches and cohort studies.

Conclusions
Our results con rmed that mutation rate of PTEN is approximately 68% in EC patients. mRNA expression level of PTEN was lower in patients with PTEN mutation than that with wide type. Moreover, PTEN mutation have a favorable overall survival and progression free survival in EC patients. Cell lines are more sensitive to AKT inhibitor (Afuresertib and AZD5363) and Mcl-1 inhibitor (MIM1) than that with wide type in EC with PTEN mutation. Multiple genes and pathways may play an important role in EC patients with PTEN mutation. These results offer a potential target and strategies for patients with PTEN mutation.

Declarations
The data that support the ndings of this study are available from TCGA database (https://portal.gdc.cancer.gov/) and GDSC database (https://www.cancerrxgene.org), which are all publicly available.

Ethics approval and consent to participate
This study does not involve animal and/or human tissue/individual data/participants, there is no ethics related issues. No permissions were required to use any repository data involved in the present study.

Consent for publication
Not Applicable.