Heparanase is a novel biomarker for immune infiltration and prognosis in breast cancer

Heparanase (HPSE), an endoglycosidase that cleaves heparan sulfate, regulates a variety of biological processes that promote tumor progression. In this study, we analyzed the correlation between HPSE expression and prognosis in cancer patients, using multiple databases (Oncomine, TIMER, PrognoScan, GEPIA, Kaplan–Meier plotter, miner v4.1, DAVID). HPSE expression was significantly increased in bladder, breast, lung, and stomach cancer compared to matched normal tissues. The increased HPSE expression correlated with poor prognosis and increased immune infiltration levels of B cells, CD8+ and CD4+ T cells, macrophages, neutrophils and dendritic cells in bladder and breast cancer. In breast cancer, the high HPSE expression was associated with basal-like subtypes, younger age (0-40), advanced Scarff-Bloom-Richardson grade, Nottingham Prognostic Index and p53 mutation status. In addition, using a mouse model of breast cancer, our data showed that HPSE upregulated IL-10 expression and promoted macrophage M2 polarization and T cell exhaustion. Together, our data provide a novel immunological perspective on the mechanisms underlying breast cancer progression, and indicate that HPSE may serve as a biomarker for immune infiltration and prognosis in breast cancer.


INTRODUCTION
Breast cancer is one of the most common tumors and the leading cause of cancer-related deaths in women within the globe [1]. Although improved the strategies for early diagnosis and treatment, the prognosis is still poor, mainly due to inherent aggressive behavior and lack of recognized treatment targets [2]. Therefore, there is an urgent need to develop more sensitive and specific biomarkers for the prognosis of breast cancer patients. In the last 2 decades, immunotherapy including programmed death-1 (PD-1), cytotoxic T lymphocyte associated antigen 4 (CTLA4), and programmed death ligand-1 (PD-L1) inhibitors, demonstrated major breakthroughs and became the major therapeutic approach in solid tumors, such as non-small-cell lung carcinoma (NSCLC) and malignant melanoma [3,4]. Breast cancer, harboring lots of activated tumor-infiltrating lymphocytes (TILs), is one of the most promising targets of immunotherapy among solid tumors [5][6][7][8][9][10]. Unfortunately, only a fraction of breast cancer patients responds well to immunotherapy. Since TILs serve as an independent favorable prognostic factor, and a predictive marker of chemotherapy, neoadjuvant therapy, and immunotherapy responses in breast cancer [11][12][13][14][15][16][17], identification of specific TILsassociated biomarkers may contribute to development of specific targeted immunotherapies in breast cancer.
The tumor microenvironment (TME), containing tumor cells and non-tumor cells, such as endothelial cells, immune cells, and fibroblasts [18], makes an important impact in tumor metastasis and progression [19][20][21][22][23]. Heparanase (HPSE) is the only mammalian endoglycosidase which can cleaves heparan sulfate (HS), regulates remodeling of the basement membranes AGING and extracellular matrix, as well as promotes the release of many HS-related molecules including cytokines, growth factors, and enzymes. HPSE is upregulated in many types of human tumors [24][25][26][27], and this elevation contributes to tumor angiogenesis, growth, metastasis, chemoresistance, and poor prognosis [28][29][30][31][32]. Inhibitors targeting HPSE, such as PI-88 (muparfostat), SST0001 (roneparstat), PG545 (pixatimod), and M-402 (necuparanib) have entered clinical trials. Although the role of HPSE in tumor cells has been well documented, its interaction with nontumor cells in the TME has not been sufficiently explored. Recent studies have suggested that HPSE interaction with immune cells which contains T cells, B cells, NK cells, macrophages, neutrophils, and dendritic cells, can have both pro-and anti-tumorigenic roles, depending on the setting [33]. In addition, one research has indicated that by increasing HPSE expression in the ILs, tumors can regulate gene expression of many other tumor and non-tumor cells [34]. Thus, analyzing the interaction between HPSE, breast cancer cells, and TILs might show a novel immunological perspective to understand the mechanisms of tumor progression and further improve the clinical practice in breast cancer therapy.
In this research, we analyzed HPSE expression and the role played in the prognosis of cancer patients. In addition, we investigated HPSE association with tumorinfiltrating immune cells and related immune markers in bladder and breast cancer, and analyzed the HPSE correlation with clinicopathological parameters in breast cancer.

HPSE mRNA expression in different kinds of human cancer
To analyze HPSE expression in different kinds of cancer, HPSE mRNA expression in different tumors and matched control tissues were performed using the Oncomine database. The analysis revealed a statistically increased HPSE expression in bladder, brain, CNS, breast, gastric, leukemia, lung, lymphoma, and sarcoma tumors compared to matched normal tissues. However, a decreased HPSE expression was found in colorectal, head and neck, and esophageal cancers ( Figure 1A).
Comparison of Oncomine and Timer results indicated that the HPSE expression was significantly increased in bladder, breast, lung, and stomach cancer, while it was decreased in colon, head, and neck cancer. Thus, we next analyzed the association of HPSE expression and prognosis in the above cancers.

High HPSE expression impacts prognosis in bladder and breast cancer
In order to investigate whether the HPSE expression correlates with prognosis in bladder, breast, gastric, lung, colorectal, head, and neck cancer patients, we used the PrognoScan, GEPIA, and Kaplan-Meier plotter databases to evaluate the impact of HPSE expression on survival. The relationships between HPSE expression and prognosis in different cancers using the PrognoScan database is shown in Table 1. Notably, HPSE expression significantly impacted prognosis in bladder and breast cancers. Analysis of HPSE expression and prognosis in different cancers using the GEIPA database showed that high HPSE expression was related to poor DFS and OS rates in bladder cancer ( Figure 2B). Analysis of HPSE expression and cancer prognosis using the Kaplan-Meier plotter database showed that high HPSE expression was related to poor RFS, DMFS, PPS, and OS rates in breast cancer; it was also associated with poor PPS and OS rates in stomach cancer ( Figure 2B). Therefore, it is conceivable (confirmed by at least 2 databases) that a high HPSE expression is an independent risk factor, and is associated with poor prognosis in bladder and breast cancer.

HPSE expression correlates with infiltrating immune cells in bladder and breast cancer
TILs are an independent predict factors of survival in cancer. Therefore, we analyzed whether the HPSE expression was associated with immune infiltration levels in bladder and breast cancer. We assessed the correlation of HPSE expression with immune infiltration levels in bladder and breast cancers by the TIMER database. The results showed that in bladder cancer, the HPSE expression is not related to infiltrating levels of B cells, medium correlation with infiltrating AGING levels of CD8+ T cells, weak correlation with infiltrating levels of CD4+ T cells and macrophages, and a strong correlation with infiltrating levels of neutrophils and dendritic cells (DCs). In breast cancer, the HPSE expression showed a medium correlation with infiltrating levels of B cells, CD8+ T cells, CD4+ T cells, and macrophages, and a strong correlation with infiltrating levels of neutrophils and DCs ( Figure 3).
To further explore the relationship between HPSE and the infiltrating immune cells, we investigated the association between HPSE and immune cell markers for tumor-associated macrophages (TAM), M1 macrophages, M2 macrophages, monocytes, NK cells, exhausted T cells, Tfh cells, Th1 cells, Th2 cells, Th17 cells, and Treg cells in bladder and breast cancer by the GEPIA database ( Table 2). The results showed that in bladder cancer, the HPSE expression had no correlation with Tfh cells, a weak positive correlation with TAM and Treg cells, and a weak negative correlation with M1 macrophages and Th2 cells. The HPSE expression also showed a medium positive correlation with M2 macrophages, monocytes, NK cells, exhausted T cells, and Th1 and Th17 cells in bladder cancer. In breast cancer, the HPSE expression had no correlation with Th2 cells, a weak positive correlation with M1 macrophages, NK cells, and Th17 cells, a medium positive correlation with TAM, exhausted T cells, Tfh cells, Th1 cells, and Treg cells, and a strong positive correlation with M2 macrophages and monocytes.

HPSE expression correlates with clinicopathological parameters in breast cancer
The correlation between HPSE expression and clinicopathological parameters was explored by the bc-   Figure 1C). Regarding age, the HPSE expression in 0-40 years group was significantly higher than in 40-70 and 70-96 years groups; no significant difference was found between 40-70 years group and 70-96 years group ( Figure 1D). The SBR histological grade evaluates the degree of duct formation, nucleus pleomorphism, and nuclear division count, while the NPI index stratifies patients into prognostic groups according to lymph node stage, tumor size, and tumor grade. Breast cancer patients with higher SBR grade and NPI tended to express higher levels of HPSE ( Figure 1E, 1F). Regarding P53 status, the HPSE expression in mutated group was significantly higher than in wild type group ( Figure 1G).

HPSE expression correlates with M2 macrophage polarization and IL-10 in breast cancer
HPSE target genes in human breast cancer tissues are listed in Table 3; they are discriminated by R. Using the AGING DAVID software, we found that a total of 17 GO functions (Biological Processes) were enriched ( Figure 4); this was supported by analysis of immune markers (Table 4).
Next, using Human MDA-MB-231-HPSE and MDA-MB-231-mock cells which with high and low expression of HPSE ( Figure 5A), we validated our data in a mouse model of breast cancer, using mice overexpressing HPSE and a control mock group. Because macrophages usually infiltrate at the edges of tumor tissues, we analyzed the infiltration of macrophages by immunohistochemical staining (IHC) at the tumor tissue edges. IHC staining showed that the HPSE expression correlated with an increased expression of CD163 ( Figure 5B) and VSIG4 ( Figure  5C), which are markers of M2 macrophages. Compared with MOCK group, M2 macrophages in HPSE group tended to infiltrate into tumor tissues. In addition, the HPSE expression related to an increased expression of IL-10 ( Figure 5D), which is known to induce macrophage polarization into the M2 phenotype.

DISCUSSION
HPSE, the only mammalian endoglycosidase that cleaves HS, contributes to tumor angiogenesis, growth, metastasis, chemoresistance, and poor prognosis in multiple tumors [24][25][26][27][28][29][30][31][32]. In this research, we demonstrate that the HPSE expression is obviously increased in bladder, breast, lung, and stomach cancer, but decreased in colon, head, and neck cancer. In addition, our data show that the high HPSE expression is an independent risk factor in bladder and breast cancer, indicating that HPSE can be used as a prognostic biomarker for bladder and breast cancer. These findings are supported by previous studies demonstrating that high HPSE gene and protein levels are associated with bladder cancer invasion and metastasis [35]. Analysis of HPSE in urine from 282 individuals showed that the urine HPSE levels were elevated during bladder cancer progression [36]. HPSE mRNA levels in bladder cancer tissues related to tumor stage, histological grade, size, number, recurrence and lymph node metastasis [37]. Inhibition of HPSE expression suppressed invasion, migration and adhesion of bladder cancer cells [38]. In addition, HPSE overexpression accelerated the obesity-associated breast cancer progression [39]. However, the HPSE expression in metastatic lesions does not always reflect the expression in primary tumors. In stage I breast cancer patients, a strong HPSE staining was associated with shorter overall survival rates [40]. Tumor growth, vascularization and recurrence were significantly reduced by inhibition the procoagulant activity of HPSE [41]. Furthermore, inhibition of HPSE expression reduced tumor metastasis by reducing extracellular regulated protein kinase (ERK) and phosphorylation of protein kinase B (Akt) [42].
The immune cells in TME play a critical role in tumor progression, and are recognized as an independent predict factor of cancer survival. Moreover, the immune cells-associated HPSE has important pro-tumorigenic and anti-tumorigenic functions [33]. A previous study has indicated that the increased HPSE expression in TILs can regulate gene expression in other tumor and non-tumor cells [34], suggesting that there might be an HPSE-TME crosstalk that can affect occurrence,  TAM, Tumor-associated macrophage; Th, T helper cell; Tfh, Follicular helper T cell; Treg, Regulatory T cell; R value of Spearman's correlation, the absolute value of R, 0-0.09 means no correlation, 0.1-0.3 means weak correlation, 0.3-0.5 means medium correlation, 0.5-1.0 means strong correlation, + means positive correlation, -means negative correlation; MEAN means the average value of relate genes and markers of one specific immune cells with statistical significance. P-value < 0.01 was considered statistical difference, * P-value < 0.01; ** P-value < 0.001; *** P-value < 0.0001. TAM, Tumor-associated macrophage; Th, T helper cell. R value of Spearman's correlation, the absolute value of R, 0-0.09 means no correlation, 0.1-0.3 means weak correlation, 0.3-0.5 means medium correlation, 0.5-1.0 means strong correlation, + means positive correlation, -means negative correlation. P-value < 0.01 was considered statistical difference, * P-value < 0.01; ** P-value < 0.001; *** P-value < 0.0001.

development, and fate of tumors. Our study proved the correlation between HPSE expression and tumorinfiltrating immune cells, and supported the intimate correlation between HPSE expression and immune cell infiltration in bladder and breast cancer.
TAMs are the most plentiful immune cells within the TME [43,44]. According to their function, macrophages can be divided into M1 and M2 phenotypes. The M1 phenotype is characterized by proinflammatory cytokines (IL-1β, TNFα, INFα, IL-6, IL-12, IL-16) and tumoricidal activity. In contrast, the M2 phenotype is characterized by anti-inflammatory cytokines (IL-4, IL-10, IL-13, TGF-β) that promote angiogenesis, tissue remodeling, and repair [45,46]. Therefore, macrophages play a dual role in tumor growth: on one hand, initiating immune responses against transformed cells; and on the other hand, AGING promoting tumor angiogenesis and growth [47][48][49][50] Macrophages are often found both in primary and metastatic tumors, and contribute to tumor progression [51,52]. In HPSE knockout mice, macrophages decreased the levels of pro-inflammatory cytokines (TNF-α, IL-1β, CXCL2, and IL6), and impaired infiltration into lung tumors, which were smaller than tumors in wild type animals [53]. In a mouse model of pancreatic cancer, HPSE overexpression was associated with increased macrophage expression of M2 cytokines IL-6, IL-10, CCL-2, VEGF, and macrophage scavenger receptor-2 (MSR-2), increased tumor size, and increased levels of tumor-infiltrating macrophages [54]. In addition, increased HPSE expression in epithelium has been associated with inflammation and inflammation-associated tumorigenesis, such as inflammatory bowel disease (IBD) [55], pancreatitis [56], and esophageal carcinoma [57]. HPSE has been also suggested as a key regulator of the tumor microenvironment [53] and tumor progression [58]. These findings show that HPSE, regardless of its cellular source, promotes tumorigenesis.

Id Description
Our previous bioinformatics and experimental studies demonstrated that HPSE promotes malignant progression, angiogenesis and metastasis in breast cancer by enhancing the crosstalk between tumor cells and platelet [65]. In this study, multi-database integration analysis indicated that high HPSE expression contributes to macrophage M2 polarization and T cell exhaustion, thus promoting tumor growth. These findings were corroborated also by IHC staining of breast tumor tissues; however, since the nude mouse model is not suitable for T cell analysis, we focused on M2 macrophage polarization. The results indicated that the protein levels of IL-10, CD163, and VSIG4 were significantly increased in breast tumor tissues in mice overexpressing HPSE, indicating that HPSE might promote macrophage M2 polarization (CD163, VSIG4) by upregulating IL-10. Together, our results show that HPSE may serve as a novel biomarker for immune infiltration and prognosis in breast cancer.

PrognoScan database analysis
The PrognoScan (http://www.abren.net/PrognoScan/) is an online database used to evaluate the biological relationship between gene expression and prognostic contains overall survival (OS), relapse-free survival (RFS), distant metastasis-free survival (DMFS), disease-specific survival (DSS), and disease-free survival (DFS) in various types of cancers [68], and provide corresponding p-value, hazard ratio (HR), and 95% confidence intervals. Therefore, it has been used to analyze the correlation between the expression of HPSE and survival in different cancers with the adjusted cox p-value < 0.05.

GEPIA database analysis
Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancer-pku.cn/index.html) is used to perform survival analysis (OS and DFS) depended on RNA sequencing data from TCGA database [69]. And in our study, was used to analyze correlation between HPSE mRNA expression and survival in various types of cancers; HR and log-rank p-values were provided, and the threshold was p-value < 0.05.

Kaplan-Meier plotter database analysis
Kaplan-Meier plotter database (http://kmplot.com/ analysis/) is used for analyzing gene association with OS, RFS, DMFS, and post progression survival (PPS) in breast, ovarian, lung and gastric cancer [70]. And in our study, was used to identify the correlation between HPSE mRNA expression and survival in the above four cancer types. The HR and log-rank p-values were provided, and the threshold was p-value < 0.05.

HPSE and infiltrating immune cells and markers
TIMER database analysis TIMER (https://cistrome.shinyapps.io/timer/) using deconvolution statistical method to analyze the infiltration levels of immune cells including B cells, CD4+ T cells, CD8+ T cells, neutrophils, macrophages, and dendritic cells (DCs) based on gene expression profiles [71]. And in our study, was used to analyze the correlation between HPSE expression and the above immune infiltrating cells in bladder and breast cancers, and provided partial correlation coefficients; the threshold was p-value < 0.01. The absolute value of R, 0-0.09 meant no correlation, 0.1-0.3 meant weak correlation, 0.3-0.5 meant medium correlation, and 0.5-1.0 meant strong correlation.

GEPIA database analysis
We used GEPIA database to identify the correlation between HPSE and related genes and markers in immune cells. the threshold was p-value < 0.01; R, 0-0.09 meant no correlation, 0.1-0.3 meant weak correlation, 0.3-0.5 meant medium correlation, and 0.5-1.0 meant strong correlation.

GO functional annotation analysis
The database for annotation, visualization, and integrated discovery (DAVID) v.6.8 (https://david. ncifcrf.gov) [74] was used to perform GO [75] functional annotation analyses (positive target genes of HPSE). The background list parameter was human genome, and the threshold was p-value < 0.05.

In vivo studies
Human MDA-MB-231-HPSE and MDA-MB-231-mock cells, which represented breast cancer cell lines with high and low expression of HPSE, were handseled by Dr. Israel Vlodavsky, the GFP inserts were performed in Glyconovo Technologies Co., Ltd, and the HPSE expression have been detected between MDA-MB-231-HPSE and MDA-MB-231-mock cells by Western blot analysis, and the specific operation method were described previously [65].
Cell culture, animal care, and establishment of a nude mouse model were described previously [65].
Histochemical score (H-score) of staining was calculated by multiplying these two variables.