NOTCH Pathway Genes in Ovarian Cancer: Clinical Significance and Associations with Immune Cell Infiltration

Background : Activation of the NOTCH signaling pathway is associated with tumorigenesis. The aim of this study was to investigate NOTCH pathway gene functions and regulatory mechanisms in ovarian cancer (OC). Methods : We conducted a bioinformatics analysis of publicly available datasets in order to identify potential NOTCH -related mechanisms, associated genes, biological pathways, and their relation to immune function. Results : Significant differential expression of the NOTCH pathway genes DLL1 , DLL3 , DLL4 , HES1 , HEY1 , JAG1 , NOTCH2 , NOTCH3


Introduction
Ovarian cancer (OC) has the highest mortality of all gynecologic cancers [1].The median age of patients diagnosed with OC is approximately 60 years, with most women diagnosed at advanced stages and with metastases [2].OC is characterized by high levels of recurrence and metastasis, drug resistance, and mortality [3].Despite advances in surgical techniques and the application of combination chemotherapy, the 5-year survival rate for patients with advanced OC is only 40-45% [4].Although a great deal of research has been carried out on the pathogenesis of OC, little is known about its regulation and the underlying mechanisms.Molecular and genomic aberrations in OC data has accumulated thanks to the advent of large, multi-cohort genomic mapping libraries like The Cancer Genome Atlas (TCGA) [5,6].However, how the data relates to each other, OC disease state, body processes, and other factors has not been well chacterized.This characterization can provide clues as to mechanisms behind OC, which can lead to devel-oping better treatments.Also, this work can lead to development of biomarkers for early detection of OC.The lack of good techniques for early detection contributes significantly to the low survival rate of OC.
The NOTCH signaling pathway is a highly conserved ligand-receptor signaling pathway involved in various aspects of cancer biology, including cancer stemness, angiogenesis, epithelial-mesenchymal transition (EMT), tumor immunity, and drug resistance [7,8].The most interesting members of the NOTCH pathway are the NOTCH receptors (NOTCH1, NOTCH2, NOTCH3, NOTCH4), the serrated typical NOTCH ligands Jagged 1 (JAG1) and Jagged 2 (JAG2), the Delta-like typical NOTCH ligands (DLL1, DLL3, DLL4), and the typical downstream genes known as hair and division enhancer 1 (HES1) and hair/division enhancer-associated YRPW patterned protein 1 (HEY1) [8,9].There is evidence that NOTCH signaling plays a pleiotropic role in cancer, with dysregulation leading to reduced proliferation and apopto-sis of cancer cells [10,11].mRNA expression levels for the four types of NOTCH receptor (NOTCH1-4) have different prognostic significance in OC patients [12].The NOTCH2/NOTCH3/DLL3/MAML1/ADAM17 signaling network has also been associated with OC [13].Genes that are associated with NOTCH pathway genes can be identified by bioinformatic analysis.Overall, the mechanism and function of the NOTCH pathway in OC remains unclear, and its mechanism of action has still to be clarified.
Cancer results from the co-evolution of malignant cells and the tumor microenvironment (TME) [14].Bidirectional interactions between the TME and cancer cells regulate its progression, development, metastasis, invasion, and resistance to therapy [15].Immune infiltration is a critical part of the TME and plays a central role in OC progression [16,17].It is important to realize the immune system plays a significant role in the development and progression of OC, and that its dysregulation can result in immune escape and resistance to therapy [18].Studies of immune cells in the TME of OC have focused on T cells, DCs, MDSCs, macrophages, NK cells, γδ T cells, and B cells [19].Therefore, identification of the genes that regulate the immune response is essential for understanding how OC cells evade or even suppress the antitumor immune response.To date, activated NOTCH signaling pathways in the TME have been widely reported in various cancer types [8,14,20].However, so far there have been few studies on the correlations between NOTCH members and immune responses in OC, or on the associated mechanisms.
In the present, study we conducted a comprehensive bioinformatics analysis of public datasets to investigate potential NOTCH-related mechanisms, genes, biological pathways, and their relation to immune function.This study looked toidentify valuable candidate biomarkers or targets for the treatment of OC.

Correlation Heat Map
Correlations between all pairs of genes in the NOTCH pathway were assessed using Pearson's correlation coefficient [24].R (version 3.6.3)and ggplot2 (version 3.3.3)were used for the analysis [24].

Association of NOTCH Pathway Gene Expression with the Clinical Features of TCGA-OC
R (version 3.6.3)and Basic R package were used for this analysis [24].The grouping condition was the median.

Univariate and Multivariate Cox Regression Analysis
R (version 3.6.3)and survival package (version 3.2-10) were used for regression analysis [24].Cox regression was used as the statistical method [29,30].PFI was used for the type of prognosis.Included variables were FIGO stage, primary therapy outcome, race, age, histologic grade, anatomic neoplasm subdivision, venous invasion, lymphatic invasion, tumor residual, tumor status, and 11 NOTCH pathway genes [23].Prognostic information from the reference was provided as supplemental data [31].

Correlation Analysis for Genes Associated with NOTCH Pathway Genes
R (version 3.6.3)and stat base package (version 3.6.3)were used for this analysis [24].
Pearson correlation analysis was used to test for relationships between genes involved in the NOTCH pathway.A significant negative correlation was found between the expression of DLL3 and NOTCH2, while the expression levels of most other NOTCH pathway genes showed a significant positive correlation with each other (Fig. 4).

Relationship between NOTCH Pathway Gene Expression and the Clinical Characteristics of TCGA OC Patients
The clinical characteristics and gene expression data for 379 OC tumor samples were downloaded from the TCGA database (Supplementary Table 1).DLL4 and NOTCH4 expression were associated with FIGO stage (p < 0.001 and p = 0.036, respectively), while NOTCH3 expression was associated with race (p = 0.039) and age (p = 0.044).The expression levels of other NOTCH pathway   genes (DLL1, DLL3, HES1, HEY1, JAG1, JAG2, NOTCH1, and NOTCH2) did not correlate significantly with the clinical data of OC patients.
GO and KEGG enrichment analyses were conducted on the NOTCH pathway and related genes (102 genes in total) (Supplementary Table 2).As shown in Fig. 10 and Supplementary Table 3, the top five biological processes identified were the regulation of NOTCH signaling pathway, NOTCH signaling pathway, negative regulation of NOTCH signaling pathway, NOTCH signaling involved in heart development, and positive regulation of NOTCH signaling pathway.The top five cytological components were the mitochondrial inner membrane, polysomal ribosome, cytosolic ribosome, mitochondrial proton-transporting ATP synthase complex, and cytosolic part.The top five molecular functions were NOTCH binding, structural constituent of ribosome, electron transfer activity, cytochrome-c oxidase activity, and heme-copper terminal oxidase activity.As shown in Fig. 11 and Supplementary Table 3, the top five pathways identified were the NOTCH signaling pathway, breast cancer, endocrine resistance, Th1 and Th2 cell differentiation, and oxidative phosphorylation.

NOTCH Pathway Gene Expression in OC and Correlation with Immune Cells
As shown in Fig. 12, NOTCH pathway gene expression was correlated with tumor-infiltrating immune cells (TIIC) in OC.DLL1 gene expression correlated positively with eosinophils, iDC, macrophages, mast cells, NK cells, Tcm, Tem, TFH and Tgd, and negatively with aDC, cytotoxic cells, T cells, Th1 cells, and Th17 cells.DLL3 expression correlated positively with Th2 cells, and negatively with CD8 T cells, cytotoxic cells, DC, iDC, macrophages, neutrophils, NK CD56bright cells, NK CD56dim cells, pDC, T cells, Tcm, Tem, Th1 cells and Th17 cells.DLL4 expression correlated positively with eosinophils, iDC, macrophages, mast cells, neutrophils, NK cells, pDC, T helper cells, Tcm, Tem, TFH and Tgd, and negatively with
Estrogen promotes the differentiation of ovarian multi-ciliated cells (MCC) by reducing DLL1 expression via the estrogen receptor β [36].Sequential combination of cisplatin/eugenol targets the NOTCH-Hes1 pathway and eliminates drug-resistant cancer stem cells [37].OC progression and resistance have been linked to JAG1, an oncogene, and hence targeting the GATA1/JAG1/NOTCH pathway might represent a new treatment strategy [38].JAG1 inhibits OC cell growth and may act by suppressing the NOTCH1 signaling pathway [39].JAG1/NOTCH3 interactions constitute a juxtaposed secretory loop that promotes the proliferation and dissemination of OC cells in the peritoneal cavity [40].Thus, BMP9-NOTCH1 signaling could also be a new therapeutic target axis for OC treatment [41].MiR-34b inhibits proliferation and epithelial-mesenchymal transition in OC by targeting NOTCH2 [42].Activation of NOTCH3-mediated signaling by IL-8 secreted from cancerassociated fibroblasts and cancer cells promotes stem cell sexuality in OC [43].In the present study, genes associated with the NOTCH pathway were mainly enriched in five signaling pathways.These were the NOTCH signaling pathway, breast cancer, endocrine resistance, Th1 and Th2 cell differentiation, and oxidative phosphorylation.In conclusion, these findings suggest the NOTCH pathway could regulate OC progression by modulating endocrine resistance, Th1 and Th2 cell differentiation, and oxidative phosphorylation.Therefore, targeting of the cross-talk network rather than a single pathway may be a more effective treatment strategy.
The tumor infiltration of immune cells and the immune evasion by cancer cells play key roles in tumor progression [9].JAG2+ TANs are closely associated with the IL-8-driven immune evasion microenvironment, thus providing a potential therapeutic target for enhancing the immunity against OC [44].Glycolytic targeting of microRNA and EZH2 in OC has been reported to lead to effector T cell dysfunction [45].In the present study, expression of the DLL1, HES1 and NOTCH3 genes was negatively associated with T cell infiltration, whereas expression of the DLL4, HEY1, JAG1, JAG2, NOTCH1, NOTCH2 and NOTCH4 genes was positively associated with T cells.In contrast, DLL3 expression was positively correlated with Th2 cells and negatively correlated with other types of T cells.The use of adaptive T cells, such as chimeric antigen receptor T cell therapy, may therefore be a promising new paradigm for the treatment of OC.A better understanding of NOTCH signaling could improve therapeutic approaches.
In summary, we evaluated the expression levels, mutations, and immune correlations of the NOTCH pathway in OC in order to identify potential biomarkers and targets.In addition to improving clinical decision making, the results contribute to a deeper understanding of the complex involvement of the NOTCH pathway in OC.Future studies should investigate the role of NOTCH signaling in OC in vitro and in vivo, as none were performed in the present study.

Conclusions
Activation of NOTCH plays an important role in mediating the development and progression of OC through multiple pathways, including the regulation of immune cells, endocrine resistance, Th1 and Th2 cell differentiation, and oxidative phosphorylation.Downregulation of JAG2 and NOTCH1 expression were associated with significantly worse PFI in OC patients.JAG2 and NOTCH1 may be potentially useful biomarkers for the treatment of OC, or as therapeutic targets.

Fig. 2 .
Fig. 2. The distribution of genetic alterations in NOTCH pathway genes in OC cases based on the cancer type summary in cBioPortal.

Fig. 4 .
Fig. 4. Correlations between the expression level of genes in the NOTCH pathway.

Fig. 7 .
Fig. 7. Forest plots showing the results of Cox regression analysis for NOTCH pathway genes and other clinical features as predictors of progression free interval (PFI) in OC patients.(A) Univariate regression analysis.(B) Multivariate regression analysis.