Reconstruction of Immune Microenvironment and Signaling Pathways in Endometrioid Endometrial Adenocarcinoma During Formation of Lymphovascular Space Involvement and Lymph Node Metastasis

Background The amplification or mutation of oncogenes and escape from immune surveillance systems promote tumor metastasis. However, subtle changes in the immune microenvironment and signaling pathways are poorly understood during the formation of lymphovascular space involvement (LVSI) and lymph node (LN) metastasis of endometrioid endometrial adenocarcinoma (EEA). Patients and methods We detected tumor immunology-related signaling pathways and immunocyte subtypes according to the mRNA levels of 750 oncogenes and genes relating to the tumor microenvironment and immune response using the Nanostring PanCancer IO 360 Panel in 24 paraffin-embedded tissues of EEAs and benign gynecological diseases. Internal reference genes were used for data normalization. Results Angiogenesis and immune cell adhesion signaling pathways were activated during LVSI formation of EEA progression. However, during the development of LVSI to LN metastasis, immune system signaling pathways were significantly inhibited, including antigen presentation, cytotoxicity, lymphoid compartment, interferon signaling, and costimulatory signaling pathways. Immune-related genes (CD69, HLA-DOA, ATF3, GBP1, AP2, DTX3L, EGR1, GBP4, TAP1, EIF2AK2, MX1, ISG15, STAT1, and HLA-DRA) were significantly downregulated in EEA with LN metastasis compared to those in EEA with LVSI. Instead, hypoxia, metabolic stress, epigenetic regulation, matrix remodeling, and metastasis signaling pathways were continuously activated in LN metastasis. We also found that neutrophils, macrophages, and mast cells might be involved in LVSI formation and LN metastasis in EEA. Conclusions EEA with metastatic LNs showed significant immunosuppressive effects. Some oncogenes, matrix remodeling- and hypoxia-related genes, and neutrophil signatures showed higher expression, suggesting their potential as therapeutic targets and offering new immunotherapy strategies in EEA during LN metastasis.


INTRODUCTION
Endometrial cancer (EC) is ranked 4th and 6th in morbidity and mortality, respectively, among cancers affecting women in the United States in 2019 (1). Statistically, the mortality rate of EC slowly increased from 2012 to 2016 in China (2). The 5-year survival rates of patients with endometrial carcinoma at FIGO stages III and IV were only 57-66% and 20-26%, respectively (3,4). Endometrial carcinoma with lymphovascular space involvement (LVSI), myometrial invasion, lymph node (LN) metastasis, and high-grade cancer were associated with significantly higher recurrence rates (5). Accumulated studies have shown that LN metastasis is a strong independent prognostic factor for endometrial carcinoma recurrence (6,7). Recent studies revealed that LVSI is an independent prognostic factor for lymph node metastasis and non-locoregional recurrence in early-stage endometrial carcinoma (8,9). However, little is known about the regulation of molecular mechanisms and the tumor microenvironment for LVSI and LN metastasis in endometrial carcinoma. The immune surveillance system and cancer cells fight a seesawlike battle from occurrence to the early and advanced stages. In the early stage, the immune system produces an anti-inflammatory microenvironment to fight against cancer cells, whereas in the late stage, tumor cells escape immune surveillance, resulting in distant metastasis and recurrence (10). Antomarchi etal. (11) found that grade 1 ECs showed a strong anti-tumor immune microenvironment; however, the high-grade ECs presented immunotolerance and immunosuppression. Pakish etal. (12) showed increased infiltration of immune cells, including granzyme B+ cells, activated cytotoxic T lymphocytes, and PD-L1 + cells, in endometrial carcinomas with high microsatellite instability (MSI-H) compared to those with microsatellite stability (MSS). These studies suggested that low-grade tumors and MSIpositive endometrial carcinoma might be more sensitive to immunotherapy. However, the regulation of the immune microenvironment and signaling pathways remain poorly understood in relation to LVSI formation and LN metastasis of endometrial carcinoma.
Immune system surveillance has been shown to play a role in type I endometrioid endometrial adenocarcinoma (EEA), while it is inert in type II serous carcinoma (13,14). This study focused on the construction of the spectrum of tumor immune microenvironments of EEA during LVSI formation and LN metastasis. It was determined that immune system activation was present in EEAs with LVSI formation. However, severe immunosuppression and tolerance were observed in ECs with LN metastasis. Hypoxia, metabolic stress, epigenetic regulation, matrix remodeling, and metastasis signaling pathway-related oncogenes and neutrophil signatures showed higher expression, suggesting their potential as therapeutic targets and offering new immunotherapy strategies for LN metastasis in EEA.

Patients and Specimens
Paraffin sections from 24 patients with EEA and benign gynecological disease were obtained from the pathology department of Peking University People's Hospital. LVSI and LN metastasis were important pathological progress indicators of EEA. Cases were separated into 4 groups of 6 cases: Normal control, LVSI-LN-, LVSI+LN-, and LVSI+LN+. The clinicopathological data of 18 EEAs are listed in Table 1. LVSI and LN metastasis of endometrial carcinoma were re-identified by a senior pathologist. The study was approved by the ethics committee of Peking University People's Hospital (2019PHB031-01).

Bioinformatics and Statistical Analysis
The raw gene expression values were normalized to those of housekeeping genes using the NanoString nSolver 4.0 software (nanoString, Seattle, WA) and log2 transformation. Differentially expressed genes, along with false discovery rate (FDR) corrected p-values was screened out for signaling pathway analysis according to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Signaling pathway scores were clustered by unsupervised hierarchical clustering, and the 1-euclidean distance was used as the similarity measure. The P-values were adjusted using the Benjamini-Yekutieli (BY) false discovery rate and the Bonferroni correction. The false discovery rate was limited to ≤5% using P-values. Furthermore, the absolute and relative abundance of an immune cell subtype was estimated by simply taking the average log2 expression of the characteristic genes. Along with nanoString n-Solver software, one-way ANONA were used for statistical analysis. P-value for significance was set at p<0.05. Scatter plots charts were obtained using Graphpad Prism 8.

Immunohistochemical Staining
Three-mm-thick slices stained with H&E or immunohistochemical (IHC) were obtained from formalin-fixed paraffin-embedded (FFPE) EEAs tissues, which were the same wax blocks as the previous nanostring analysis. IHC staining of FFPE slides was performed using monoclonal mAbs against BIRC5(ab76424), CD68(ab213363) and CD163(ab213612) (Abcam, Cambridge, UK), with 1:500 working dilution. Paraffin sections are first dewaxed in steps from xylene to different concentrations of alcohol, and finally placed in tap water for washing. The slices were treated with heated citric acid repair fluid for antigen repair. After incubation with 3% H2O2 for 10 min, endogenous peroxidase was removed. Goat serum was used for blocking for 30 min. The first antibody was incubated overnight at 4 degrees. The next day, the second antibody labeled with horseradish peroxidase was incubated for 30 min. In the middle of each step, 1 x PBS should be used for cleaning at 5 min * 3 times. DAB was used for IHC staining observed under the microscope.
Scoring for BIRC5, CD68, and CD163 was evaluated by percentage of cells stained in tumor and stromal tissue compartments by a pathologist.

Heatmap of Signaling Pathway and Differentially Expressed Genes (DEGs) Between EEAs and Benign Gynecological Lesions
HE staining from 24 patients with EEA and benign gynecological disease was showed in Figure 1. Cases were separated into 4 groups of 6 cases: Normal control (Non-cancer), LVSI-LN-, LVSI+LN-, and LVSI+LN+. The clinicopathological data of 18 EEAs are listed in Table 1. First of all, we analyzed the signal pathways and differentially expressed genes between 6 cases of normal control group and 18 cases of EEAs. The data showed that MAPK, Hedgehog signaling, Wnt signaling were significantly suppressed in EEAs. The other 22 signaling pathways, including cell proliferation, DNA damage repair and so on, are greatly activated in EEAs (Figure 2A). Differentially expressed genes (DEGs) between EEAs and normal control were showed in Figure 2B.  Table 2.

Signaling Pathway Characteristics of EEA During LVSI Formation and LN Metastasis
Signaling pathways involved in LVSI formation and LN metastasis compared to LVSI-LN-group were shown in Figure 3. We found significantly different distributions of signaling pathways among the three groups (LVSI-LN-, LVSI +LN-, and LVSI+LN+) in EEA progression. Persistent activation of cell proliferation and PI3K-AKT signaling pathways was observed in the LVSI+LN-group LVSI+LN+ group compared to the LVSI-LN-group without metastasis in EEA progression ( Figures 3A, B). Hedgehog signaling pathways were downregulated whereas the angiogenesis, immune cell adhesion, apoptosis, DNA damage repair, and JAK-STAT signaling pathways were upregulated in EEA with LVSI compared to those in the LVSI-LN-group without metastasis ( Figures 3C-H). Hypoxia, metabolic stress, epigenetic regulation, matrix remodeling, and metastasis signaling pathways were upregulated in EEA with LN metastasis compared to the LVSI-LN-group without metastasis ( Figures  3I-M). However, antigen presentation, cytotoxicity, lymphoid compartment, interferon, and costimulatory signaling pathways were all downregulated in EEA with LN metastasis compared to EEA with LVSI ( Figures 3N-Q). Cytokine and chemokine, myeloid compartment, autophagy, TGF-b, MAPK, NF-kB, Notch, and Wnt signaling pathways showed no significant changes in the absence of LVSI, LVSI formation, and LN metastasis during EEA progression (Supplementary Figure 1).

Distribution Characteristics of Immunocyte Subsets in EEA During LVSI Formation and LN Metastasis
The results revealed that immunocyte subtypes and genes relating to immune surveillance and immune escape, such as NK cells, CD8 T cells, Treg cells, DCs, exhausted CD8, TIL count, PD-L1, and CTLA4, showed no significant differences in expression in the absence of LVSI or with LVSI and LN metastasis in EEA progression ( Supplementary Figures 2 and 3). However, total neutrophils were upregulated in the LVSI+LN+ group compared to the LVSI-LN-group (P<0.05) ( Figure 5B). Total macrophages also showed an upward trend, though not statistically significant. Whereas mast cells showed a downward trend, although there was no statistical significance (P=0.093) (Figures 5A, C). The ratio of neutrophils in TIL was up-regulated in LVSI+LN+ group, compared to LVSI+LN-and LVSI-LN-group ( Figures 5E, F). The ratio of macrophages in TIL showed an upward trend in LVSI+LN+ group, compared to LVSI+LN-group and LVSI-LNgroup ( Figure 5D).
We further analyzed gene expression relating to macrophages, neutrophils, and mast cells. Macrophage-related gene CD68 was upregulated in the LVSI+LN group compared to the LVSI-LNgroup. However, CSFIR and CCL2 were downregulated in the LVSI+LN+ group compared to the LVSI+LN-group ( Figure 6A). Neutrophil-related gene FCAR was upregulated in the LVSI+LNgroup and LVSI+LN+ group compared to the LVSI-LN-group. Neutrophil-related gene CXCL1 were upregulated in the LVSI +LN+ group compared to those in the LVSI+LN-group ( Figure  6B). Among mast cell-related genes, tumor necrosis factor (TNF) was upregulated in the LVSI+LN-group compared to the LVSI-LN-group. Moreover, the mast cell-related genes CAP3, HDC, and MS4A2 were downregulated in the LVSI+LN+ group compared to those in the LVSI-LN-group ( Figure 6C).

HE Staining and Immunohistochemical Staining Confirmed the Expression of Hub Genes and Immunocytes
Compared to the LVSI-LN-group without metastasis, BIRC5 related to cell proliferation was in top 20 gene in LVSI+LNgroup and LVSI+LN+ group. So, we further verified the BIRC5 in situ protein expression in EEAs tissues. We observed that the brown black positive granules were expressed in the LVSI+LN-group and LVSI+LN+ group, but were absent in the LVSI-LN-group tissues without LVSI and LN metastasis (Figure 7). The total number of neutrophils was significantly increased in the LVSI+ LN+ group

DISCUSSION
T cell-mediated adaptive immunity plays an active role in the antitumor process. DCs, macrophages, and B cells present tumor antigens to CD4+ helper T cells. In turn, they cooperate to induce CD8+ T cells and B cells as well as activate NK cells and macrophages. These pathways ultimately eliminate tumor cells through the CD8+ cytotoxic T cell-dependent apoptosis pathway. Moreover, an imbalance in T cell-mediated responses plays a negative role in cancer progression, leading to an immunosuppressive environment and tumor escape (15)(16)(17). High TIL levels occur in anti-tumor immune response and indicate a good prognosis in gastric cancer and breast cancer (18,19). Moreover, regulation of lymph node function by NK cells is associated with prognosis in patients with stage II colon cancer (20). Study have shown that the average fluorescence intensity of CD8 staining in peripheral blood of patients with grade 3 EEC is lower than that of healthy donors.
Cytotoxic T cells with decreased CD8 expression were positively correlated with EEAs (21). Endometrial tumor microenvironment reduces the recruitment of NK cells and changes the phenotype and function of NK cells (22). The profiles of immune infiltration NK cells, DCs and CD8+T cells showed associated with patients survival in TCGA uterine cancer cohort (23). This study showed that immunocytes (CD8, NK cells and DC) tended to recede in the advanced stage of LN metastasis compared to those in the LVSI+LNgroup, leading to immunotolerance or immunosuppression. However, due to the limited sample size, these immune cells did not show significant difference. The obtained expression profile of immune microenvironment regulation during LN metastasis in endometrial carcinoma was consistent with the results of previous studies. In this study, the ratio of macrophages in TILs showed an upward trend in the LVSI+LN+ group compared to that in the LVSI+LN-and LVSI-LN-groups. Various studies have shown that tumor-associated macrophages (TAMs) in endometrial carcinoma are associated with increases in LVSI and LN metastasis and a poor survival outcome (24). TAMs are polarized M2 macrophages, which secrete cytokines, chemokines, and growth factors to promote EC development and inhibit the anti-tumor immune system. CTHRC1 increases the recruitment of M2-like macrophages, prompting myometrial invasion in endometrial carcinoma by regulating the integrin-Akt signaling pathway (25). TAM reduced the ERa expression in  EC cells via HOXB13 by increasing CXCL8 secretion, thus promoting EC invasion and metastasis (26). Invasive macrophages induce ERa expression through epigenetic mechanism mediated by IL17A, which makes endometrial cancer cells sensitive to estrogen (27). For a long time, the function of neutrophils in the tumor microenvironment has been ignored. Recently, accumulating studies have shown that neutrophils are involved in promoting tumor progression and metastasis (28,29). Wculek & Malanchi (30) showed that neutrophils are involved in lung metastasis in mouse breast cancer models, whereas drug-or gene-mediated suppression of the leukotriene-generating enzyme arachidonate 5-lipoxygenase abolished neutrophil-associated pro-metastatic functions. During the formation of the tumor metastasis microenvironment, increased secretion of CXCL1 and CXCL2 by endothelial cells and megakaryocytes promotes the release of neutrophils into circulation via the regulation of CXCR2 signaling. Neutrophil-derived MMP9 is more likely to activate and participate in pre-tumor functions (31). Neutrophils produce rEEAtive oxygen species, which cause DNA damage, genomic instability, and gene mutations in precancerous epithelial cells, thus promoting the carcinogenic transformation (32). For the first time, this study found that neutrophils showed a rising trend when LN metastasis develop in EC. Compared to the LVSI-LN-group, VEGFB and MMP9 showed higher expression in the LVSI+LN-group. VEGFB, MMP9, and neutrophils may cooperatively regulate angiogenesis in the primary metastasis of endometrial carcinoma. Moreover, CXCL1 mRNA expression was upregulated in the LVSI+LN+ group compared to that in the LVSI+LN-group, suggesting that the high expression of chemokine CXCL1 may be involved in the recruitment of neutrophils to lymph vessels during the progression of lymph node metastasis in EEA.
Studies have also found that mast cells play a dual role in regulating cancer progression (33,34). First, mast cells participate in the internal and acquired immune process, promote DC migration and maturation, and interact with T and B cells. Additionally, they can promote the release of inflammatory factors such as TNF-a, MIP, and MCP and participate in the formation of an anti-tumor inflammatory  group compared to those of the LN-negative group (36). Conversely, another study found that mast cell density was increased in metastatic LNs in breast cancer (37). In this study, we found downregulated expression of mast cell-related genes in both LVSI formation and LN metastasis, suggesting that mast cells play a negative role in regulating immune response tolerance and LN metastasis in EEA.
In conclusion, this study found that ECs with metastatic LNs showed significant immunosuppressive effects. Some oncogenes as well as genes relating to matrix remodeling, hypoxia, macrophages, and neutrophil signatures showed higher expression, suggesting their potential for use as therapeutic targets and offering new immunotherapy strategies in EEA during LN metastasis.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the ethics committee of Peking University People's hospital. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

AUTHOR CONTRIBUTIONS
YC chiefly wrote this manuscript and collected clinical data and samples. XZ was primarily responsible for pathological diagnosis. ZW gave advice on the selection of clinical samples. JW was mainly responsible for project design. All authors contributed to the article and approved the submitted version.