Prognostic and Immunological Roles of MMP-9 in Pan-Cancer

Background Matrix metalloproteinase-9 (MMP-9) can degrade the extracellular matrix and participate in tumor progression. The relationship between MMP-9 and immune cells has been reported in various malignant tumors. However, there is a lack of comprehensive pan-cancer studies on the relationship between MMP-9 and cancer prognosis and immune infiltration. Method We used data from TCGA and GTEx databases to comprehensively analyze the differential expression of MMP-9 in normal and cancerous tissues. Survival analysis was performed to understand the prognostic role of MMP-9 in different tumors. We then analyzed the expression of MMP-9 across different tumors and at different clinical stages. Based on the results, we assessed the correlation between MMP-9 expression and immune-associated genes and immunocytes. Finally, we calculated the tumor mutation burden (TMB) of 33 cancer types and analyzed the correlation between MMP-9 and TMB, DNA microsatellite instability, and DNA repair genes. Results MMP-9 significantly affected the prognosis and metastasis of various cancers. It was associated based on overall survival, disease-specific survival in five tumors, progression-free interval in seven tumors, and clinical stage in eight tumors, as well as with prognosis and metastasis in adrenocortical carcinoma and kidney renal clear cell carcinoma. It was also coexpressed with immune-related genes and DNA repair genes. The expression of MMP-9 was positively correlated with the markers of T cells, tumor-associated macrophages, Th1 cells, and T cell exhaustion. Furthermore, MMP-9 expression was highly correlated with macrophage M0 in 28 tumors. In addition, its expression was associated with TMB in eight cancer types and DNA microsatellite instability in six cancer types. Conclusion MMP-9 is related to immune infiltration in pan-cancer and can be used as a biomarker related to cancer prognosis and metastasis. Our findings provide prognostic molecular markers and new ideas for immunotherapy.

In recent years, the incidence of cancer and its morbidity and mortality have shown an increasing trend. Cancer is a major cause of death worldwide and is second only to cardiovascular disease. The WHO estimates that malignant tumors will become the main cause of global mortality after 2030 [17]. The tumor microenvironment (TME) influences tumor growth and development. Tumor-associated macrophages (TAMs) are macrophages that infiltrate the tumor tissue and most immune cells in the TME. Tekin et al. [18] found that macrophages release MMP-9 in pancreatic cancer. TAMs can support the proliferation, invasion, and metastasis of tumor cells. Therefore, the development of antitumor drugs that can target macrophage polarization is urgently required. Immunotherapy is highly suitable for patients with cancer because of its excellent efficacy. However, not all patients can benefit from immunotherapy and research has shown that tumor mutation burden (TMB) and DNA microsatellite instability (MSI) can be used as predictive markers for immunotherapy efficacy. TMB [19] has a good predictive value for immunotherapy in a variety of tumors. In addition, MSI [20] has been regarded as an important molecular marker for the prognosis and adjuvant treatment of colorectal cancer and other solid tumors. In view of the complexity of tumor progression, pan-cancer analysis has been widely used in cancer research and considerable progress has been made in understanding various tumor features, including cancer susceptibility variation, oncogenic pathway cooccurrence and mutual exclusion, and biological regulation network disorder [21][22][23].
MMP-9 has been found to be closely related to immunity and tumor progression; however, most studies have focused on single cancers. Here, we systematically studied MMP-9 expression and its correlation with prognosis and metastasis in 33 cancer types to help us fully understand the role of MMP-9 in tumors. A flowchart of the study is shown in Figure 1. We also analyzed the relationship between MMP-9 expression and immune cell infiltration.

Materials and Methods
2.1. Data Acquisition. Gene expression profiles, mutation data, and clinical information of 33 cancers in TCGA database were downloaded from UCSC Xena [24] (http:// xena.ucsc.edu/).The disease-specific survival (DSS) and progression-free interval (PFI) data were downloaded from TCGA Pan-Cancer (PANCAN) of UCSC Xena. After excluding cases with missing survival time data, 11,057 samples were included in the study.

Gene Expression Analysis.
We used "wilcox.test" to analyze the differential expression of MMP-9 in normal and tumor tissue samples, as well as the differential expression of MMP-9 in different cancer types in TCGA database, and drew a box diagram.
In view of the small number of normal tissue samples in TCGA database, we included data from the GTEx (genotype-tissue expression) database [25] using the "Match TCGA normal and GTEx data" option in the GEPIA2 database [26] (http://gepia2.cancer-pku.cn/#analysis) for the differential analysis to ensure more reliable results.

Survival and Clinical
Analysis. The expression of MMP-9 was extracted from the gene expression profile data, and the samples were divided into high-and low-expression groups according to the median MMP-9 expression. We used the Kaplan-Meier method to analyze the survival information and "survival" [27] and "survminer" to draw the survival curve. We also performed COX analysis of the survival data, and the R package "forestplot" was used to visualize the results.

mRNA Expression Levels of MMP-9 in Different Types of Human Cancers.
To determine the differences in the expression levels of MMP-9 in various human cancers, we examined the MMP-9 expression levels using the RNA-seq data of multiple malignancies from TCGA database. The differential expression of MMP-9 between tumor and adjacent normal tissues across tumor types is shown in Figure 2(a). Except for tumors without normal tissue data, MMP-9 expression was significantly higher in tumor samples than in normal samples.
To further examine the prognostic potential of MMP-9 in different cancers, we evaluated the PFI of the 33 cancer types. Higher MMP-9 expression levels were associated with    These results indicate that high MMP-9 expression might be a risk factor for poor prognosis in ACC, BLCA, KIRC, LIHC, UVM, THCA, and GBM, while low MMP-9 expression might be a risk factor for poor prognosis in DLBC, UCEC, SKCM, and CESC.

Relationship between MMP-9 Expression and the Clinical
Stage. Next, we analyzed the expression of MMP-9 in relation to the tumor stage in the 33 cancer types and found that it was closely related to the clinical stage in eight tumors (Figures 6(a)-6(h)). MMP-9 was differentially expresses according to the clinical stage and was specifically positively correlated with the tumor stage in ACC, BLCA, and KIRC, in which MMP-9 expression increased with tumor progression. These results suggest that MMP-9 expression has the potential to influence cancer prognosis by affecting lymph node metastasis. These results suggest that MMP-9 is involved in promoting cancer progression or metastasis.

Correlation between MMP-9 Expression and Immune
Cell Infiltration. Many studies have shown that MMP-9 is related to immune cells [33,34]. Therefore, we evaluated the correlation between MMP-9 and immune cell infiltration in 33 tumors. Through survival analysis and clinical correlation analysis, we found that MMP-9 was related to poor prognosis and metastasis in ACC and KIRC. DLBC was used as the control group. The correlation between the expression level of MMP-9 and six types of infiltrating immune cells in ACC, KIRC, and DLBC is shown in Figures 7(a)-7(c). The expression of MMP-9 was positively correlated with the infiltration of B cells, CD8+ cells, CD4+ cells, and macrophages in ACC and KIRC, while it was mostly negatively correlated in DLBC. In addition, our results indicated a marked correlation between MMP-9 expression and the macrophage M0 in 28 cancer types (Table 1). MMP-9 was positively correlated with the macrophage M1 in four tumors (Figures 8(a)-8(d)). The levels of infiltrating macrophage M2 were positively correlated with MMP-9 expression in HNSC, CESC, and COAD (Figures 8(e)-8(g)) and negatively correlated in SKCM, LIHC, and THCA (Figures 8(h)-8(j)). In addition, TIMER2.0 analysis showed that MMP-9 had a strong positive correlation with macrophages (Figure 7(d)). These results showed that high MMP-9 expression was positively correlated with immune cell infiltration.

Correlation between the MMP-9 Expression Level and
Immune Cell Markers. The TME [35] can affect survival and tumor metastasis. We performed immune cell marker gene coexpression analyses in ACC, KIRC, and DLBC and found that the expression of MMP-9 was mainly positively correlated with the expression levels of most marker sets of T cells, TAMs, M2 macrophages, Th1 cells, and T cell exhaustion, especially in ACC (Table 2), while no such correlation was observed in DLBC.

Coexpression of DNA Repair
Genes with MMP-9 and GSEA. To better understand the potential mechanism of MMP-9 expression in cancers, we analyzed its expression in ACC, KIRC, and DLBC using GSEA. The results showed that MMP-9 was mainly enriched in immune-related pathways in KIRC, such as immune response regulating cell surface receptor signaling and regulation of immune effector process (Figure 9(c)), and in pathways related to gene silencing and RNA modification in ACC and DLBC (Figures 9(a)  and 9(b)). We further used RNA sequence data from TCGA database to evaluate the correlation between MMP-9 and five DNA repair genes and found that MMP-9 was associated with multiple DNA repair genes in various tumors (Figure 9(d)). More specifically, MMP-9 was moderately positively correlated with MSH2 in ACC and negatively correlated with EPCAM and PMS2 in KIRC. In addition, Table 1: Correlation analysis between MMP-9 and macrophage M0 of TCGA database (the P values are indicated as * P < 0:05, * * P < 0:01, and * * * P < 0:001).

Cancer type
Cor  3.7. Correlation between the MMP-9 Expression Level and TMB, MSI, and Immune Checkpoint Genes. TMB and MSI are important for immunotherapy response. Here, we calculated the TMB of each tumor sample and analyzed the correlation between MMP-9 and TMB in 33 tumors. MMP-9 was positively correlated with TMB in six tumors, including ACC, BRCA, COAD, brain lower grade glioma (LGG), OV, and UCEC, and negatively correlated with HNSC and LUSC (Figure 10(a)). Next, we analyzed the correlation between MSI and MMP-9 levels. MSI was positively correlated with MMP-9 in COAD and sarcoma, whereas it was negatively correlated in four tumors (Figure 10(b)). In addition, most immune checkpoint genes were coexpressed with MMP-9, especially PDCD1 and CTL4, which are the targets of immune checkpoint inhibitors.

Discussion
MMP-9 can degrade the extracellular matrix components and promote tumor invasion and metastasis. The high expression of MMP-9 is closely related to the development, invasion, and metastasis in many cancers. Here, we found that MMP-9 promotes cancer development and progression in some cancers, suggesting that MMP-9 expression can be used to predict metastasis, especially in kidney cancer. In addition, correlation analysis showed that the expression of MMP-9 was correlated with different levels of immune infiltration and immunological markers. Finally, we evaluated the relationship between MMP-9 expression and TMB and MSI. The results showed that MMP-9 may be used as a biomarker for pan-cancer prognosis.
In this study, we obtained the expression levels of MMP-9 and the prognosis and relevant indices of 33 cancer types from TCGA database. Differential expression of MMP-9 in cancer and normal tissues was observed in all cancers, with MMP-9 being overexpressed in tumor tissue across cancer types. This suggested that dysregulated or excessive MMP-9 could cause tumorigenesis. As for the survival analysis, higher expression levels of MMP-9 were correlated with poorer prognosis in patients with ACC, BLCA, KIRC, and LIHC. In contrast, high levels of MMP-9 were favorable for the prognosis of lymphoma. The results indicated that MMP-9 promotes bladder and cervical cancer invasion and metastasis. MMP-9 is a potential prognostic biomarker for various cancers, including lung, ovarian, pancreatic, and breast cancers [11,16]. However, in our study, analysis based on three survival indicators showed that high MMP-9 expression was associated with poor prognosis in ACC and KIRC. The correlation between MMP-9 and renal cancers has not been reported in previous studies. In addition, our analysis of OS, DSS, and PFI showed that high expression of MMP-9 is a protective factor in DLBC; however, this has not been observed in previous studies. MMP-9 promotes metastasis via ECM decomposition [36]. The expression of MMP-9 was related to the clinical stage in eight tumors, suggesting that MMP-9 may be involved in tumor metastasis. In addition, MMP-9 increased with the progression of cancer in three types of urological tumors. These results suggest that MMP-9 may be used as an indicator of prognosis and metastasis in pan-cancer.
Furthermore, we found that MMP-9 expression was correlated with immune infiltration levels in multiple cancer types, especially ACC and KIRC. It was positively correlated with the infiltration of B cells, CD8+ cells, CD4+ cells, and macrophages in ACC and KIRC, while it was mostly negatively correlated in DLBC. This suggests that MMP-9 may lead to poor prognosis by participating in tumor immune infiltration. Moreover, MMP-9 expression levels were mainly positively correlated with immune cell markers. Notably, in ACC, MMP-9 was moderately correlated with four Th1 marker genes (TBX21, STAT4, STAT1, and IFNG), suggesting that it may be involved in Th1 differentiation. Th1 cells induce the activation of macrophages, NK cells, B cells, and CD8+ T cells [37]. Concurrently, we also found that MMP-9 was moderately correlated with the immune markers of CD8+ T cells (CD8A and CD8B) and T cells (CD3D, CD3E, and CD2). These results suggest that MMP-9 may promote cell-mediated inflammatory   BRCA  CESC  CHOL  COAD  DLBC  ESCA  GBM  HNSC  KICH  KIRC  KIRP  LAML  LGG  LIHC  LUAD  LUSC  MESO  OV  PAAD  PCPG  PRAD  READ  SARC  SKCM  STAD  TGCT  THCA  THYM  UCEC

28
BioMed Research International responses by participating in Th1 differentiation and T cell activation. Th1 cells regulate macrophage function at multiple levels. In addition, MMP-9 was associated with macrophage immune marker genes. More specifically, MMP-9 expression was positively correlated with IL-10 (a TAM marker), which is often associated with tumor immune evasion. The markers of M2 macrophages were moderately correlated with MMP-9 expression in tumors, suggesting that MMP-9 may be involved in the differentiation of macrophages. Most importantly, in ACC, MMP-9 expression was strongly correlated with most markers of T cell exhaustion, including TGFB1, PDCD1, CTLA4, LAG3, and GZMB. T cell exhaustion is one of the main causes of immune dysfunction that leads to a poor prognosis [38]. This suggests that MMP-9 may be the cause of poor prognosis in patients with ACC. At the same time, T cell exhaustion is also one of the reasons for poor immunotherapy response. In contrast, TAMs are important cellular components of the TME [39] and imbalance of M1/M2 plays a key role in tumor progression, immune escape, and drug resistance [40]. Therefore, the development of antineoplastic drugs that target macrophage polarization is important. Tekin et al. [18] found that M0 macrophages secrete MMP-9 in the early stages of pancreatic cancer development, which promotes tumor progression. This is consistent with the findings of our study. In addition, we found that MMP-9 was highly positively correlated with M0 macrophage levels in 27 types of tumors. Although research has shown that M2 macrophages can alter miR-149-5p to increase the expression of MMP-9 in liver cancer [41], in our study, MMP-9 and M2 macrophages were negatively correlated in LIHC. These results indicated that MMP-9 is involved in the recruitment and activation of immune cells and that MMP-9 inhibition may be another approach for tumor immunotherapy based on macrophages.

BioMed Research International
In ACC, MMP-9 was highly correlated with the markers of T cell exhaustion, which can be reversed by PD-1 inhibitors. A recent study [42] identified TMB as a marker for evaluating the therapeutic effect of PD-1 inhibitors. Therefore, we analyzed the relationship between MMP-9 expression and TMB expression. Our results also showed that MMP-9 has a significant positive correlation with TMB in a variety of cancers. This suggests that in these cancers, patients with high MMP-9 expression may be more suitable for immunosuppressive therapy. Furthermore, MSI plays an important role in the diagnosis, prognosis, and treatment of multiple tumors, especially colon cancer [43]. Our results showed that MSI is positively correlated with MMP-9 in COAD. In brief, patients with high MMP-9 expression may be more suitable for immunotherapy.
Immune checkpoints are closely related to tumor immune escape. Hence, we analyzed the relationship between the expression of MMP-9 and certain common immune checkpoint genes. The results showed that MMP-9 was significantly associated with immune checkpoints in most tumor types. This may be related to the poor prognosis of some tumors in the survival analysis. Another study [44] indicated that inhibition of MMP-2/MMP-9 improves the efficacy of PD-1 or CTLA4 blockade in the treatment of primary and metastatic tumors.
Monferran et al. [45] reported that the DNA repair protein Ku interacts with MMP-9 at the cell membrane of highly invasive hematopoietic cells. Our results also showed that MMP-9 was correlated with various DNA repair genes. These findings may help in understanding the role of MMP-9 in gene expression and gene repair. The GSEA results also suggested that MMP-9 participates in immune regulation. This is consistent with the results of our previous analysis. This suggests that MMP-9 is a potential target for immunotherapy.
Although we comprehensively analyzed MMP-9 expression in 33 tumors, many deficiencies exist in our study. First, our data source was relatively single and simple as we used mainly TCGA database data. Second, our findings require further validation in the clinical setting. Third, although we found that the expression of MMP-9 is related to immune cell infiltration and survival, we could not prove its causal relationship, and hence, its prognostic value needs to be further studied.
In conclusion, MMP-9 can be used as a pan-cancer prognostic biomarker involving immune infiltration, especially in kidney cancer. These findings may contribute to clinical decision-making and cancer immunotherapy.