Identification of Interleukin-9 Producing Immune Cells in Endometrial Carcinoma and Establishment of a Prognostic Nomogram

Background Interleukin-9 (IL9) plays a critical role in immunity and the pathogenesis of endometrial cancer (EC), especially endometrioid EC (EEC). This study aimed to identify the IL9+ immune cell subsets and their pleiotropic functions and establish an optimized prognostic nomogram towards the promotion of personalized treatment of EEC. Methods 1,417 EC patients were involved in the present study. 143 patients from the tertiary gynecology centers in Shanghai between 2013 and 2019 were recruited, and the study protocol was approved by the Institutional Review Board (IRB) of Shanghai First Maternity and Infant Hospital. The genomic data of the other 1,274 patients were extracted from the TCGA and the MSKCC datasets, respectively. Immune and stromal scores were calculated using the ESTIMATE R tool, and the tumor infiltration of immune cells was analyzed using the TIMER platform. Metascape and GEPIA datasets were used for bioinformatic analysis. P < 0.05 was considered statistically significant. All statistical analyses were performed with GraphPad Prism and R studio. Results 552 genes that were correlated with leukocyte infiltration, lymphocyte activation, and regulation of innate immune response were up-regulated in the high immune score group. More IL9+ cell infiltration was detected in the highly and moderately differentiated EC (p = 0.04). High IL9+ lymphocyte infiltration was related to a better overall survival (p = 0.0027). IL9 positive cell clusters included ILC2s, Vδ2 γδT cells, mast cells, macrophages, and Th9 cells. Parameters such as FIGO stage, IL9 score, Vδ2 + γδT cell infiltration, classification of differentiation, and diabetes mellitus were assigned a weighted number of points in the nomogram for a specific predicted 3-, 5- and 10-year overall survival (OS). IL9–IL9R axis played a vital role in EEC, IL9R positive cell subgroups were also identified, and the related function was analyzed in the present study. Additionally, PR (Progesterone Receptor, or PGR) expression was relevant to a higher density of IL9+ lymphocyte infiltration. However, PGRMC1 (Progesterone Receptor Membrane Component 1) was negatively relevant to IL9R (p = 4.26e-8). Conclusion We observed a significant infiltration of IL9+ cells and the overrepresentation of IL-9R in tissue specimens of patients in EC cases. The nomogram incorporating the IL9 could accurately predict individualized survival probability in EEC. Additionally, this study not only established a prognostic nomogram but also assist in the firmer understanding of the relevance of the IL9-IL9R axis and IL9-producing cells in EC immunity.


INTRODUCTION
Endometrial carcinomas (ECs) are the most common gynecologic malignancies and the majority of uterine corpus cancers globally (1). Moreover, endometrioid endometrial carcinoma (EEC) is the most common subtype of endometrial cancer (EC), representing approximately 87% of all diagnosed ECs (2). EECs are usually diagnosed at early stages and are associated with a favorable prognosis (3). However, for those patients with advanced stages or tumor recurrence, the prognosis is relatively poor, with 5-year overall survival rate of approximately 20% for stage IV EEC (4,5). The routine treatments of the EC include surgery, hormonal therapy, chemotherapy, and immunotherapy. Recently, biomathematical modeling and recurrence risk estimation basing on patientrelated characters have opened a new era of care oriented towards the promotion of personalized medicine.
Additionally, it is more and more apparent that the prognosis of cancer patients is not solely determined by tumor characters, but also the circumstance, particularly the immune microenvironment, which plays an essential role in cancer biology. Interleukin-9 (IL-9) was initially defined as a T H 2-type cytokine but was reported to have pleiotropic functions, inducing the proliferation, differentiation, and effector functions of numerous immune cell subsets and plays a critical role in immunity and the pathogenesis of cancers (6,7). Most recently, much attention has been focused on a major population of T H cells that produce IL-9, namely T H 9 cells, which have been reported to have potent abilities in eradicating advanced tumors. The majority of IL-9-producing cells in cancer are T H 9 cells; however, IL-9 can also be secreted by Vd2+ gd T cells (the dominant gdT-cell subset), group 2 innate lymphoid cells (ILC2s), and some cytotoxic T cells. ILC2 is a subset of innate lymphoid cells (ILCs), its role in cancer immune response is dependent on cytokine context. ILC2s could produce IL-5, leading to eosinophil activation and an increased anti-tumor immune response in solid cancers (8). In the endometrium, Interleukin-9 (IL-9) plays a unique position in the human endometrial function and embryo implantation (9). However, the presence and role of IL9 within EEC were not thoroughly investigated. Herein, we investigated the expression of IL9, IL9R, and IL9 producing immune cells and proposed a nomogram to predict the prognosis of EEC.

Study Population, Data Collection, and Follow-Up
In total, data from 1,417 EC patients were analyzed in the present study (Supplementary Figure 1). 143 patients with EC from two tertiary gynecology centers in Shanghai were recruited in the present study. Inclusion criteria included: (1) patients who underwent an operation for EC from 2013 to 2019, (2) patients with final histopathological diagnosis of EC. Exclusion criteria included patients with histologically proven uterine sarcoma or other types of tumors. Tumor tissues and adjacent healthy tissues retrospectively collected from 127 consecutive patients were used for the tissue microarray (TMA, Superbiotek, Shanghai). Tissues of the other 16 patients were prospectively collected for CyTOF (cytometry by time-of-flight) analysis and immunofluorescence staining. Pathological parameters such as the quality, grading, tumor stages of the specimens were evaluated according to FIGO 2009 edition. This study was approved by the Institutional Review Board (IRB) of Shanghai First Maternity and Infant Hospital.
The cohort containing the genomic data (e.g., mRNA, mutation frequency) of 1,274 patients were extracted from the TCGA database (Firehose, Legacy, 548 patients; PanCancer atlas, 529 patients) and the MSKCC dataset (197 patients), respectively. cBioPortal platform (www.cbioportal.org) was used for bioinformatics analysis (10). In the present study, disease-free survival (DFS) was defined as the time for any recurrence. Overall survival (OS) was defined as the time for death from any cause. If the postoperative margin was negative, the operation was considered as R 0 resection. Follow-up consisted of serum tumor marker measurements every one to three months and computed tomography (CT) every six months. Complete follow-up was conducted for the entire cohort of patients.

ESTIMATE (Estimation of STromal and Immune cells in
MAlignant Tumor tissues using Expression data) algorithms calculated the immune and stromal scores (12) using Pearson's correlation coefficient. Then, these data were divided into high and low immune/stromal score groups. The selection of differentially expressed genes (DEGs) was performed by using the "limma" R package with p-value <0.05 and log fold change >1 as a filter (13).

CyTOF Staining and Barcoding
Single cells isolated from endometrial tumor tissues were washed with complete RPMI (Sigma) followed by three washes in Bariumfree PBS (Sigma) by spinning at 1,800 rpm for 3 min. Cells from each sample were stained with intercalator-103Rh to label dead cells. After one wash in the MaxPar staining buffer, living cells were fixed in Fix I Buffer followed by permeabilization. Each sample was labeled with barcodes from Cell-ID ™ 20-plex Pd barcoding kit. Barcoded samples were washed twice in the MaxPar staining buffer and pooled into one sample. Human TruStain FcX Fc receptor blocker (BioLegend) was used to block Fc receptors of cells, which were then incubated with cell-surface antibodies as listed in Supplemental Table 1 at 4°C for 30 min. After incubation, cells were washed twice in the MaxPar staining buffer and fixed as described above, followed by two washes in Perm-S buffer. Antibodies against intracellular targets were incubated with permeabilized cells in Perm-S buffer for 30 min at 4°C. At the end of the staining, cells were washed twice in MaxPar staining buffer and stored in 1 ml of MaxPar Fix and Perm Buffer containing 125 nM MaxPar Intercalator-Ir (191Ir and 193Ir) at 4°C. After 12 h, cells were washed twice in MaxPar staining buffer and stored as a pellet in MaxPar staining buffer at 4°C until analysis. To minimize the batch effect, samples were stained all in one batch then analyzed by CyTOF in two consecutive days (the day after cell staining). On the day of analysis, cells were washed twice in MaxPar water and resuspended in MaxPar water containing 10% EQ ™ four-element calibration beads followed by acquisition on CyTOF.

Immune Cell Infiltration Analysis
The infiltration of six types of immune cells (CD4+ T cells, CD8+ T cells, B cells, neutrophils, macrophages, and dendritic cells) based on RNA-Seq expression profile data was calculated by using the TIMER (Tumor IMmune Estimation Resource) algorithm (14). The correlation between IL9, IL9R, and immune cells was calculated by Spearman's correlation analysis by TIMER. The correlation coefficient >0.3 indicates a positive/negative correlation.

Enrichment Analysis
Metascape (http://metascape.org/gp/index.html) is an effective and efficient tool for experimental biologists to comprehensively analyze and interpret OMICs-based studies in the big data era. The database was used to perform the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis, which is used to predict the potential biological functions of the overlapping genes of the DEGs and target genes. Then, verification was performed by the GEPIA database (http://gepia.cancer-pku.cn) to identify hub genes (15)(16)(17)(18)(19)(20)(21)(22)(23).

Statistical Analysis
Pearson's Chi-square test for categorical variables and the Wilcoxon rank-sum test for continuous variables were used to compare various parameters in different groups divided by IL9/ IL9R expression. The Kaplan-Meier method was used to estimate OS, DFS, or PFS. Differences in survival outcomes were assessed by the log-rank test. Results were presented as hazard ratios (HRs) and 95% confidence intervals (CIs). p < 0.05 was considered statistically significant except in logistic regression (p < 0.1) for the coefficient analysis. All statistical analyses were performed with GraphPad Prism (version 8.0; GraphPad Software, La Jolla, California) and R studio (version 3.6.1; R studio, Boston, Massachusetts).

Immune/Stromal Scores, DEGs, and Enrichment Analysis of DEGs in EC
By using ESTIMATE algorithm, immune and stromal scores were calculated for 529 EC patients whose clinical data were extracted from the TCGA Pan-Cancer dataset. And immunerelated and stromal related genes were further identified (Supplemental Table 2); 552 genes were up-regulated in high immune score group and were selected for enrichment analysis. The correlation includes IFNg production, B cell activation, lymphocyte activation, and regulation of innate immune response ( Figure 1A). Both CD4 and CD8 T cells are the primary sources of IFNg. Next, the tumor infiltration with six types of immune cells was analyzed by TIMER to investigate the consistency of the enrichment analysis. These analyses above showed that EC patients with high immune scores might experience multiple immune cell infiltration and activation. In contrast, only B cells and CD8+ T cell infiltration, instead of macrophages, dendritic cells (DCs), CD4+ T cells, or neutrophil, were found relevant to the OS of EC patients ( Figure 1B).

The Component of IL-9 Producing Immune Cells in EC
Based on the Pan-cancer clinical data extracted from the TCGA dataset and TIMER platform, we found that IL9 was up-regulated in various cancer types, including EC (Supplementary Figure 2A). However, the expression of IL9 did not correlate with the infiltrating levels of any of the six types of immune cells (Supplementary Figure 2B). Through immunofluorescence staining of 143 tumor tissues of EC patients, we found that nearly all the Vd2 + gdT cells were IL9 positive. However, although most of the IL5+ cells (mostly ILC2 or T H 2) secreted IL9, they only presented as a small subgroup of IL9 positive cells in EC (Figure 2A). Figure 2B showed the gene skyline presenting the expression profiles of IL9 and its receptor IL9R in the majority of immune cell types through microarray and ULI RNA-seq data. Consistently, IL9 was enriched in ILC2, Vd2 + gdT cells, memory B cells, CD4+/CD8+ T cells, macrophages, and mast cells. Similarly, the IL9R gene was also up-regulated in B cells. Moreover, IL9R was also found upregulated in mast cells, dendritic cells, Vd2 + gdT cells, and ILC2s, with ILC2 expressing a significantly higher level of IL9R compared to other immune cells. To further investigate the patients). DEGs were generated from IL9 high and IL9 low subgroups (cutoff: FPKM 0.1, Supplementary Figure 2C). The top 10 of biological processes, cellular component, and molecular function GO terms of IL9-low (Supplementary Figure 2D) or IL9-high (Supplementary Figure 2E) related DEGs. 127 patients with ECs were then included in the current analysis. The mean age of patients at the time of diagnosis was 54 ± 8.7 years. The mean size of the tumor was 7.95 cm 2 (1-175 cm 2 ). 59% (75 patients) of the patients presented with Figo stage Ia ECC, 17.3% (22 patients) presented with stage Ib; patients presented with stages II, III, and IV were 9.4% (12 patients), 9.4% (12 patients), and 0.8% (one patient), respectively. Five patients were without staging data. The IL9+ lymphocyte count was 83.17 ± 122.1 cells/ spot(d = 1.5 mm), the Vd2 + gdT cell count was 9.85 ± 12.15 cells/spot(d = 1.5 mm).

Prognostic Estimation in EEC Patients
Next, the correlation of the number of IL9+ cells with the expression of the biomarkers was evaluated, respectively ( Figure 3A). And we elucidated that up-regulated Progesterone Receptor (PR) expression(p = 0.04) and high/ moderate EC differentiation(p = 0.04) were relevant to more IL9+ lymphocyte infiltration. To detect the connection of IL9+ lymphocyte infiltration and survival, we performed a univariate survival analysis of the IL9+ lymphocytes and Vd2 + gdT cells for 127 patients after a median follow-up of 81.8 months (range, 25-124 months, specimens from 27 patients were not sufficient for staining or met the exclusion criteria). Specifically, high IL9+ lymphocyte infiltration (IL9 hi , IL9+ cells ≥100 cells/spot, 0.6 cells/mm 2 ) was positively correlated with a better overall survival (p = 0.0027) ( Figure 3B). Vd2 + gdT cells were reported to be the main resource of IL9 in plenty of solid tumors and also one of the resources of IL9 in EC; however, the present study revealed the opposite that less than high Vd2 + gdT cell infiltration (Vd2 lo , Vd2 + gdT cell<18 cells/spot, 0.1 cells/mm 2 ) was related to a better OS (p = 0.0221, HR 7.61) ( Figure 3C). It might be because that in EC, Vd2+gdT cell presented 0.15-0.86% of immune cells, T H 9 cells are 3.6-14 times as many as Vd2 + gdT cells in the present study. Next, all of these covariates were included in the multivariable logistic regression model. Briefly, on multivariable analysis, FIGO stage (Estimate Std. −7.582, Pr(>|t|) = 0.000006), IL9 score (Estimate Std. 8.917, Pr(>|t|) = 0.026), classification of differentiation (Estimate Std. 2.865, Pr(>|t|) = 0.067), and diabetes mellitus (Estimate Std. 9.561, Pr(>|t|) = 0.07) emerged as predictors for the outcome of interest. To predict the overall survival of patients with EC, a prognostic nomogram was established through the Cox regression model analysis. Though the infiltrating Vd2 + gdT cells did not present as a predictor, it was still considered for the development of the nomogram given its clinical importance in the univariate survival analysis. Each factor in the nomogram was assigned a weighted number of points, and the sum of points for each patient was in accordance with a specific predicted 3-, 5-and 10-year OS ( Figure 3D). We also validated the nomograms by using the concordance index (C-index) to assess the predictive accuracy of the nomograms. For the internal validation of the nomogram, the concordance index for our survival model was 0.754(SD = 0.148, p = 0.0006, Figure 3E), which showed a good agreement between the nomogram-predicted survival and actual survival.

IL9R Expression Enriched in Endometrium and EECs
The expression of IL9R positively correlated with the infiltrating levels of B cells, DC cells, CD8+ T cells, and CD4+T cells ( Figure  4A). Concerning IL9R RNA expression, it is enriched in normal female tissues, especially in the endometrium ( Figure 4B). Concerning protein expression in cancer, IL9R significantly enriched in endometrial cancer compared to renal cancer, breast cancer, and other malignancies ( Figure 4C). Specifically, although in a few cases, IL9R is only expressed in the tumor cells (Figures 4N,  O). However, mostly, IL9R was not only detected in tumor cells but also infiltrated lymphocytes (Figures 4D-M). Thus, we focused on the IL9R RNA expression in immune cells by analyzing Monaco and HPA datasets and found that IL9R up-regulated in Naïve-B cells, T H 2 cells, Treg cells, eosinophils, and neutrophils ( Figures 4P-R). Firstly, IL9R RNA was found not relevant to the DFS or OS of endometrial cancer based on a small sample size of data(n = 168) from the TCGA dataset ( Figures 5A, B). However, in a larger cohort of 541 patients, patients with high IL9R expression showed significantly better OS (p < 0.05, Figure 5C). Similar to IL9, IL9R RNA was also up-regulated in endometrial cancer tissues (UCEC) compared to adjacent normal tissues (n = 265, p < 0.05, Figure 5D). Specifically, the expression level of IL9R RNA was significantly downregulated in Stage IV tumors compared to the early stages (p = 0.0129, Figure 5E).

Pathway and Process Enrichment Analysis of IL9R
As IL9R was enriched in B cells, T cells, mast cells, eosinophil, and neutrophils, pathway and process enrichment analysis had been carried out in different cell types with the following ontology sources: KEGG Pathway, GO Biological Processes, Reactome Gene Sets, Canonical Pathways and CORUM. The top 20 clusters with their enriched representative terms were shown in Figure 5F. Specifically, the IL9R related genes were associated with several pathways such as oxidative phosphorylation and cell cycle of B cells, RUNX1 pathways, lymphocyte differentiation and activation, Jak-STAT signaling pathway of T cells, and CD28 costimulation of eosinophils, and leukocyte activation of mast cells and so on. To further capture the relationships between the terms, a subset of enriched terms of T cells and mast cells had been selected and rendered as a network plot, where terms with a similarity >0.3 were connected by edges. The network was visualized using Cytoscape, where each node represented an enriched term.
were identified in Figure 6B; the involved network included T cell and B cell activation, T H 17 differentiation, leukocyte migration, IL12 pathway, and IFNg production ( Figure 6E). The expression of PGR and PGRMC1 positively correlated with the infiltrating levels of CD8+T cells (Figures 6C, D).

DISCUSSIONS
Endometrioid endometrial carcinomas are frequently estrogendependent tumors with heterogeneous prognosis (24,25). Though biomathematical tools have provided an excellent chance for the promotion of targeted therapies and personalized treatments, a tiny fraction of them is sufficient to yield excellent prognostic results for EECs. This might be owing to the heterogeneity of EEC and the comprehensive characterization of the tumor microenvironment. This inspired us that endometrial immune profile and cytokines might be new parameters for prognostic prediction (26,27). Investigators have shown that the immune score could predict survival in cancer patients (28,29). Specifically, cytokines and chemokines have been highlighted owing to their capability of active or suppress immune cascades in the immune microenvironment (30,31). The common cytokine receptor g chain (gc) family of cytokines includes interleukin-2 (IL-2), IL-4, IL-7, IL-9, IL-15, and IL-21. This set of cytokines exhibits broad pleiotropic actions on the immune system, bridging both the innate and adaptive immune systems. IL-9 was first discovered around the 1990s as a late T cell growth factor as well as a mast cell growth factor. IL-9 is produced predominantly by helper T cells such as T H 2 and T H 9 cells; it could also be provided by ILC2 cells, mast cells, and NK T cells. It usually functions on T and B cells, gdT cells, eosinophils, neutrophils, and mast cells (32) through the activation of a JAK/STAT pathway and plays a critical role in immunity and the pathogenesis of cancer. IL9 yields different responses depending on the cancer type. This cytokine not only has been shown to exhibit anti-tumor activity but also has been presented as a tolerogenic cytokine in most solid cancers to promote T regulatory cell function (33,34). However, it has not been extensively studied in endometrial carcinoma. In the present study, we found that the high density of tissue-resident IL9+ cells was associated with a better prognosis. Because Vd2 T cells were considered as a major source of IL-9 (35), we also evaluated the infiltration of Vd2 T cells in endometrial cancer. We found that elevated Vd2 T cell infiltration was relevant to worse overall survival, and multivariable analysis showed an attenuated diagnostic value of this cell type. CyTOF and immunofluorescence were performed to explain the discrepancy. The results revealed that IL-9 was also produced by Vd2 T cell, ILC2, mast cells, eosinophils, M2 macrophages, T H 9, and NK/NKT cells in the endometrial cancer tissues. Intriguingly, IL9 positive immune cells were also ESR positive according to CyTOF analysis. Besides, IL9R, the receptor of IL9, was up-regulated in the endometrium tissue and endometrial cancer tissues as well. In the present study, IL-9Ra is not only detected on T cells and B cells but also on other hematopoietic cells such as eosinophils, neutrophils, mast cells, and ILC2s. Recent studies have shown that sex hormone levels regulate tissue-resident populations of some tissue-resident immune cells such as ILC2s in homeostasis (36). Ovariectomy of ERa−/− females and orchidectomy of ERa−/− males restored ILC2 numbers and function to WT levels. For advanced stages and recurrence of endometrial cancer, hormone treatment using progesterone could slower the growth of endometrial cancer cells and govern the immune microenvironment. However, the association between endometrioid resident immune cells and estrogen or progesterone remains unclear (37). The present study provided an initial idea of the association between IL9R and PGRMC1.

CONCLUSIONS
In conclusion, we observed a significant infiltration of IL9+ cells and the overrepresentation of IL-9R in tissue specimens of patients in EC cases. Our proposed nomogram, based on FIGO classification, IL9 score, and gdT score, can classify patients into subgroups with different prognosis and will help facilitate personalized strategies for EC patients. We also elucidated that IL9 was relevant to the PR/PGR expression level, IL9R was negatively relevant to PGRMC1. Additionally, this study not only established a prognostic nomogram but also assisted in the firmer understanding of the relevance of the IL9-IL9R axis and IL9-producing cells in EC immunity.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Institutional Review Board (IRB) of Shanghai First Maternity and Infant Hospital. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
HT and HF conceived the project, designed the study, and drafted the manuscript. HT, X-PW, and HF designed the study, wrote and revised the manuscript, and approved the final submission. HT, HF, and X-PW revised the manuscript and approved the final submission. XH, M-FW, Y-FS, X-LW, and Y-RL were involved in the design of the study. All authors contributed to the article and approved the submitted version.