miR-149-3p Is a Potential Prognosis Biomarker and Correlated with Immune Infiltrates in Uterine Corpus Endometrial Carcinoma

Background Endocrine disruption is an important factor in the development of endometrial cancer. Expression of miR-149-3p is observed in some cancer types, while its role in uterine corpus endometrial carcinoma (UCEC) is unclear. Methods The clinical and genomic data and prognostic information on UCEC were obtained for patients from the TCGA database. The Kruskal–Wallis test, Wilcoxon signed-rank test, and logistic regression were used to analyze the relationship between clinical characteristics and miR-149-3p expression. Kaplan–Meier survival curve analysis was used to study the influence of miR-149-3p expression and miR-149-3p target genes on the prognosis of UCEC patients. The TargetScan, PicTar, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were used to determine the involvement of miR-149-3p target genes in function. Immune infiltration analysis was used to analyze the functional involvement of miR-149-3p. QRT-PCR was used to validate the expression of miR-149-3p in UCEC cell lines. Results High expression of miR-149-3p in UCEC was significantly associated with age (P < 0.001), histological type (P < 0.001), histological grade (P < 0.001), tumor invasion (P=0.014), and radiation therapy (P=0.011). High miR-149-3p expression predicted poorer overall survival (OS) (HR: 2.56; 95% CI: 1.64–4.00; P < 0.001), progression-free interval (PFI) (HR: 1.85; 95% CI: 1.29–2.65; P=0.001), and disease-specific survival (DSS) (HR: 2.33; 95% CI: 1.37–3.99; P=0.002). Low expressions of miR-149-3p target genes, including ADCYAP1R1, CGNL1, CHST3, CYGB, DNAH9, ESR1, HHIP, HIC1, HOXD11, IGF1, INMT, LSP1, MTMR10, NFIC, PLCE1, RARA, SNTN, SPRYD3, and ZBTB7A, were associated with poor OS in UCEC. MiR-149-3p may be involved in the occurrence and development of UCEC via pathways including PI3K-Akt signaling pathway, Ras signaling pathway, AGE-RAGE signaling pathway in diabetic complications, focal adhesion, and MAPK signaling pathway. miR-149-3p may inhibit the function of CD8 T cells, cytotoxic cells, eosinophils, iDC, mast cells, neutrophils, NK CD56bright cells, NK CD56dim cells, pDC, T cells, T helper cells, TFH, Th17 cells, and Treg. miR-149-3p was significantly upregulated in UCEC cell lines compared with endometriotic stromal cells. Conclusion High expression of miR-149-3p was significantly associated with poor survival in UCEC patients. It may be a promising biomarker of prognosis and response to immunotherapy for UCEC patients.


Introduction
Uterine corpus endometrial carcinoma (UCEC) is one of the three main gynaecological malignancies, the incidence of which increases over time. Endometrial cancer is an epithelial tumor of the endometrium and a malignant tumor of the female reproductive system with a high incidence, posing a serious threat to the health of women worldwide [1,2]. It currently ranks second in the incidence of gynaecological cancers in developing countries, behind cervical cancer [3]. According to the pathogenesis and biological behavior characteristics of endometrial cancer, endometrial cancer can be divided into estrogen-dependent (type I) and nonestrogen-dependent (type II), and most endometrial cancers belong to the former, which means endometrial cancer is closely related to endocrine. e etiological mechanisms of UCEC have not been fully elucidated. Risk factors for UCEC that have been identified include unstable postmenopausal or perimenopausal oestrogen levels, obesity, infertility, diabetes, hypertension, and family history [4]. Early clinical signs in patients with UCEC include irregular vaginal bleeding in the postmenopausal or perimenopausal period, pelvic cramps, and abdominal pain. e current lack of effective early diagnosis of UCEC has resulted in many UCEC patients missing the best time for treatment, and the poor prognosis of UCEC patients, especially those with metastases after surgery and radiotherapy, is also one of the important issues to be addressed in the current treatment of UCEC [5]. Although progress has been made in early detection and treatment, a considerable number of advanced UCEC cases have been diagnosed [6,7]. With the development of gene microarray technology, increasing research data are available to help resolve the complex pathogenesis of UCEC and monitor the disease progress. erefore, systematic bioinformatic analysis of UCEC-related disease data allows for a rapid search for biomarkers that can be used for UCEC prognosis.
MicroRNAs (miRNAs) are a unique class of endogenous and small noncoding RNAs that are approximately 18 to 25 nucleotides in length. ey alter gene expression at the posttranscriptional level primarily through complete or incomplete base pairing with the 3′ untranslated region (3′UTR) of their target mRNAs. Translational repression and mRNA degradation are the 2 main pathways through which miRNAs direct gene regulation [8]. Many studies have shown that different miRNAs are abnormally expressed in different tumors and participate in tumor formation and growth as oncogenes or oncogenes [9]. ere is substantial evidence that miRNAs are stably detectable in serum and plasma and have the potential to be noninvasive biomarkers for the diagnosis and prognosis of various cancers. miRNAs offer an innovative idea for screening and detection of cancer patients [10][11][12].
In oral squamous cell carcinoma, reduced levels of miRNA-149-3p lead to malignant progression and predict poor prognosis [13]. Increased miR-149-3p expression significantly inhibited the proliferation, migration, and invasion of bladder cancer cells [14]. Only a few studies mentioned that the lnc HOXB-AS1 was upregulated in endometrial cancer and spongy miR-149-3p upregulated Wnt10b [15]. However, the expression of miR-149-3p in UCEC and its relation to clinical features has not been well studied. e present study examined the expression of miR-149-3p in UCECs using an online database and analyzed the relationship between expression levels and clinical characteristics. A survival curve was drawn to analyze the relationship between miR-149-3p expression level and overall survival (OS). Important contributions of miR-149-3p target genes to function were identified by TargetScan, PicTari, Gene Ontology (GO), and Kyoto Gene and Genome Encyclopedia (KEGG) analyses. e functionally significant involvement of miR-149-3p was analyzed by immune infiltration analysis.
QRT-PCR was used to validate the expression of miR-149-3p in UCEC cell lines. e results of this study could provide new prognostic biomarkers for UCEC.

Clinical Information.
e analysis was carried out according to references [16,17]. R (version 3.6.3) was used for statistical analysis and visualization. e R package was the base R package. e molecule was hsa-miR-149-3p (MIMAT0004609). e grouping condition was median. e disease was UCEC. Data were obtained from miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA (https://portal.gdc.cancer.gov/) UCEC project. e miR-NAseq data in RPM (reads per million mapped reads) format were log2-transformed.
e analysis was carried out according to reference [18]. R (version 3.6.3) was used for statistical analysis and visualization. ggplot2 (version 3.3.3) was used for visualization. e molecule was hsa-miR-149-3p. e disease was UCEC. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in TCGA UCEC. e miRNAseq data in RPM format were log2 transformed. Data were unfiltered. ROC Curves. e analysis was carried out according to reference [18]. R (version 3.6.3) was used for statistical analysis and visualization. e R packages were the pROC package (version 1.17.0.1) (for analysis) and the ggplot2 package (version 3.3.3) (for visualization). e molecule was hsa-miR-149-3p. e clinical variable is normal versus tumor. e disease was endometrial cancer. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data in RPM format were log2 transformed. Data are unfiltered.

Correlation of miR-149-3p Expression with Clinical
Features. Correlation analysis of gene expression with clinical characteristics was carried out according to reference [16]. R (version 3.6.3) was used for statistical analysis and visualization. ggplot2 (version 3.3.3) was used for visualization. e molecule was hsa-miR-149-3p. Clinical variables were age, histological type, tumor invasion, and radiation therapy. e disease was UCEC. Data were miRNAseq data and clinical data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data in RPM format were log2 transformed. Data filtering conditions included removing controls/normal (not all items have controls/ normal) and keeping clinical information available.
Logistics analysis was carried out according to references [18,19]. R (version 3.6.3) was used for statistical analysis and visualization. R package was mainly a basic package. e statistical method was a dichotomous logistic model. e independent variable was hsa-miR-149-3p. e type of independent variable was low high dichotomous. e disease was UCEC. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miR-NAseq data in RPM format were log2 transformed. Data filtering conditions include removal of control/normal (not all items have control/normal) and retention of clinical information.
e Kaplan-Meier method analysis was carried out according to references [16,17]. R (version 3.6.3) was used for statistical analysis and visualization. e survminer package (version 0.4.9) was used for visualization. e survival package (version 3.2-10) was used for statistical analysis of survival data. e molecule was hsa-miR-149-3p. Subgroups were 0-50 and 50-100. Prognostic types were OS, PFI, and DSS. e disease was UCEC. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data in RPM format were log2 transformed. Supplementary data were prognostic data from reference [20]. Data filtering conditions included removal of control/normal (not all items had control/normal) and retention of clinical information.
COX regression analysis was carried out according to references [19][20][21]. R (version 3.6.3) was used for statistical analysis and visualization. e survival package (version 3.2-10) was used for statistical analysis of survival data. Statistical method was the Cox regression module. e prognosis type was OS. Included variables were clinical stage, primary therapy outcome, age, weight, height, BMI, histological type, residual tumor, histologic grade, tumor invasion (%), menopause status, hormones therapy, diabetes, radiation therapy, surgical approach, and hsa-miR-149-3p. e disease was UCEC. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data in RPM format were log2 transformed. Supplementary data were prognostic data from reference [20]. Data filtering included removal of control/normal (not all items have control/normal) and retention of clinical information.
Forest plot: software was R (version 3.6.3). R package was the ggplot2 package.
Nomogram plot analysis was carried out according to literature [19,20]. R package was rms package and survival package. e prognosis type was OS. Included variables were clinical stage, primary therapy outcome, age, histological type, residual tumor, tumor invasion, radiation therapy, and miR-149-3p.

Predicted
Putative Targets of miR-149-3p. miR-149-3p targets were obtained from Database TargetScan, miRanda, TarBase, miRTarBase, miR2Disease, miRecords, and miR-Walk [22][23][24][25]. UCEC mRNA expression media files are downloaded from the website (https://bioinfo.life.hust.edu. cn/miR_path/download.html). e UCEC prognosis-related genes were analyzed according to reference [16]. R (version 3.6.3) was used for statistical analysis and visualization. e survminer package (version 0.4.9) was used for visualization. e survival package (version 3.2-10) was used for statistical analysis of survival data. e statistical method was Cox. e subgroups were 0-50 versus 50-100. e prognosis type was OS. e disease was UCEC. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data were log2 transformed in RPM format. Supplementary data were prognostic data from reference [20]. Data filtering criteria included removal of control/normal (not all items had control/normal) and retention of clinical information. e common genes among miR-149-3p target genes, UCEC-downregulated genes, and UCEC prognosis-related genes were analyzed according to the Venn diagram.

GO and KEGG Analyses of miR-149-3p Targets.
e Database for Annotation, Visualization, and Integrated Discovery (DAVID) can provide a comprehensive set of functional annotation tools to facilitate understanding of the biological significance behind a large number of genes. GO and KEGG analyses were performed on the targets of miR-149-3p using the DAVID database (https://david.ncifcrf. gov/) [26][27][28]. GO and KEGG enrichment pathways (adjusted P value less than 0.05) were considered significant categories.

Immune Infiltration Analysis by ssGSEA.
e analysis was performed according to reference [18]. R (version 3.6.3) was used for statistical analysis and visualization. e R package was the GSVA package (version 1.34.0) [29]. e immunoinfiltration algorithm was ssGSEA (built-in algorithm of the GSVA package). e molecule was hsa-miR-149-3p. Immune cells were aDC (activated DC), B cells, CD8 T cells, cytotoxic cells, DC, eosinophils, iDC (immature DC), macrophages, mast cells, neutrophils, NK CD56bright cells, NK CD56dim cells, NK cells, pDC (plasmacytoid DC), T cells, T helper cells, Tcm (T central memory), Tem (T effector memory), Tfh (T follicular helper), Tgd (T gamma delta), 1 cells, 17 cells, 2 cells, and Treg. e disease International Journal of Endocrinology was endometrial cancer. Data were miRNAseq data from level 3 BCGSC miRNA Profiling in the TCGA UCEC project. e miRNAseq data in RPM format were log2 transformed. e data filtering condition was to remove control/normal (not all items had control/normal). Markers for 24 immune cells were obtained from reference [30].

Statistical Analysis.
Statistical analysis was carried out according to reference [18]. All statistical analyses were performed using R (v.3.6.3). e Wilcoxon rank-sum test, chi-square test, and Fisher exact test were used to analyze the relationship between clinical characteristics and miR-149-3p. P values less than 0.05 were considered statistically significant.

Clinical Characteristics.
A total of 546 patients were analyzed in the present study (Table 1)

Discussion
e development of type I endometrial cancer is associated with continuous estrogen stimulation of the endometrium without progestin antagonism. e endometrium lacks progesterone antagonism, and continuous stimulation by  estrogen will result in a prolonged state of hyperproliferation, which will further develop into endometrial cancer. MiRNAs regulate the cell cycle and cell differentiation and migration, which may also act as tumor suppressor genes or oncogenes during tumorigenesis and tumor development [32]. MiRNAs are also involved in the development of various cancers and are described in detail in UCEC. MiR-181c affects the growth of estrogen-dependent endometrial cancer cells by targeting PTEN, which may be an effective target for endometrial cancer therapy [33]. MiR-320a exerts antitumor effects on endometrial cancer by regulating IGF-1R, which can be used as a target for endometrial cancer gene therapy [34]. Downregulated miR-29b expression is associated with poor prognosis in endometrial cancer (EC) and contributes to the evaluation of EC prognosis [35]. miRNA-205 has potential clinical utility as a prognostic marker for endometrial cancer [36]. erefore, detection of miR-149-3p is an ideal candidate for improving the early detection of UCEC.
Immune cell infiltration has emerged as a new indicator of prognosis for patients with different types of solid tumors [43]. ere is heterogeneity in the immune response between tumors [44]. Immune infiltration in UCEC is currently a hot topic, and an understanding of immune infiltration will

Conclusion
miR-149-3p was highly expressed in UCECs and correlated with poorer OS compared with normal tissues. e high miR-149-3p expression in UCEC was significantly associated with age (P < 0.001), histological type (P < 0.001), histological grade (P < 0.001), tumor invasion (P � 0.014), and radiation therapy (P � 0.011). miR-149-3p may be involved in the occurrence and development of UCEC via pathways, including PI3K-Akt signaling pathway, Ras signaling pathway, AGE-RAGE signaling pathway in diabetic complications, focal adhesion, and MAPK signaling pathway. Expression of miR-149-3p was correlated with immune infiltration in UCEC. is study partially elucidates the role of miR-149-3p in UCEC and provides a promising biomarker for prognosis and immunotherapy response in UCEC patients.

Data Availability
All data generated or analyzed during this study are included in this article. e datasets generated in this study are available from TCGA that provides free resources.

Ethical Approval
TCGA is a public database. Ethical permission was obtained for patients participating in the database. Users can download relevant data for free to conduct research and publish relevant articles. Our research is based on opensource data, so there are no ethical issues.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Xiaoyuan Lu and Buze Chen conceived and designed the experiments; Li Jing, Sicong Liu, and Haihong Wang