A Comprehensive Pan-Cancer Analysis of the Tumorigenic Role of Matrix Metallopeptidase 7 (MMP7) Across Human Cancers

Growing evidence has shown the oncogenic function of matrix metallopeptidase 7 (MMP7) in various tumors. However, no systemic pan-cancer analysis on the association between MMP7 and different cancers based on big clinical data is available. TIMER2, GEPIA2, UALCAN, cBioPortal, String, Metascape, and other web databases were searched in the present study. Generally, MMP7 expression is significantly upregulated in most The Cancer Genome Atlas (TCGA) cancer types compared to the paired normal controls, yet is downregulated in tumor tissues of invasive breast carcinoma (BRCA), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC), and skin cutaneous melanoma (SKCM). MMP7 protein expression is notably higher in the primary tumor tissues of colon cancer, lung adenocarcinoma (LUAD), and uterine corpus endometrial carcinoma (UCEC) than in normal tissues and is significantly lower in the primary tumor tissues of breast cancer, clear cell renal carcinoma, and ovarian cancer. Furthermore, MMP7 expression is strongly associated with pathological stages, clinical outcomes, tumor mutational burden (TMB), and microsatellite instability (TSI). Gene amplification was detected in most TCGA cancer types. In addition, the missense mutation is the primary type of MMP7 genetic alteration in tumors. Significant positive correlations between MMP7 expression and cancer-associated fibroblasts (CAFs) have been demonstrated in most TCGA cancers. MMP7 expression was also found to be positively correlated with infiltration of dendritic cells and macrophages in some specific tumor types. Functional enrichment analysis by the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO) methods revealed that RNA processing and DNA damage checkpoints might reveal the pathogenetic mechanisms of MMP7. This pan-cancer analysis provides a clear panorama for the tumorigenic roles of MMP7 across different cancer types. Moreover, MMP7 could be a potential drug therapeutic target in such cancers.

Growing evidence has shown the oncogenic function of matrix metallopeptidase 7 (MMP7) in various tumors. However, no systemic pan-cancer analysis on the association between MMP7 and different cancers based on big clinical data is available. TIMER2, GEPIA2, UALCAN, cBioPortal, String, Metascape, and other web databases were searched in the present study. Generally, MMP7 expression is significantly upregulated in most The Cancer Genome Atlas (TCGA) cancer types compared to the paired normal controls, yet is downregulated in tumor tissues of invasive breast carcinoma (BRCA), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC), and skin cutaneous melanoma (SKCM). MMP7 protein expression is notably higher in the primary tumor tissues of colon cancer, lung adenocarcinoma (LUAD), and uterine corpus endometrial carcinoma (UCEC) than in normal tissues and is significantly lower in the primary tumor tissues of breast cancer, clear cell renal carcinoma, and ovarian cancer. Furthermore, MMP7 expression is strongly associated with pathological stages, clinical outcomes, tumor mutational burden (TMB), and microsatellite instability (TSI). Gene amplification was detected in most TCGA cancer types. In addition, the missense mutation is the primary type of MMP7 genetic alteration in tumors. Significant positive correlations between MMP7 expression and cancerassociated fibroblasts (CAFs) have been demonstrated in most TCGA cancers. MMP7 expression was also found to be positively correlated with infiltration of dendritic cells and macrophages in some specific tumor types. Functional enrichment analysis by the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO)

INTRODUCTION
Matrix metalloproteinases (MMPs) are a family of structuralrelated zinc-dependent endopeptidases which can degrade almost all kinds of extracellular matrix proteins, including collagens and gelatins (1,2), and can regulate the cleavage of cell surface receptors and also the release of apoptotic ligands and cell surface molecules as well (3,4). Matrix metallopeptidase 7 (MMP7), also known as matrilysin, pump-1 protease (PUMP-1), or uterine metalloproteinase, is a member of the MMP family encoded by the MMP7 gene in humans (5).
As the smallest MMP enzyme, MMP7 has the common characteristics of the MMP family, namely, degradation of casein, fibronectin, gelatin types I, III, IV, and V, and proteoglycans (2). Yet, different from most MMPs, MMP7 possesses its own unique features, which is constitutively enriched in many epithelial cell types, such as the salivary gland epithelium, gallbladder, pancreas, liver, breast, intestine, and reproductive organs (6). More and more evidence has shown that MMP7 plays a critical role in the genesis and development of tumors by reducing cell adhesion, inhibiting cancer cell apoptosis, and inducing vasculogenesis (7,8). In addition, MMP7 is also considered to be involved in tumor metastasis and inflammatory cascades reaction (9). Recently, MMP7 upregulation has been confirmed in various human cancer types, including lung, melanoma, esophagus, gallbladder, stomach, pancreas, colon, bladder, and other malignancies. Moreover, it was previously demonstrated that the downregulation of MMP7 can inhibit proliferation, migration, and invasion of cancer cells (10).
To our knowledge, however, only one pan-cancer perspective about the whole MMP family and its diagnostic/prognostic potential has been published (11). Due to the intricacy of MMP7 in tumorigenesis, it is requisite to conduct a pan-cancer analysis on the correlation between MMP7 and various human cancers based upon big clinical data. Accordingly, the present study applied The Cancer Genome Atlas (TCGA) project and the Gene Expression Omnibus (GEO) database to perform the pancancer analysis of MMP7 across all TCGA cancer types considering several aspects, such as gene/protein expression, prognostic value, genetic alteration, immune infiltration, and pathway enrichment analysis.

MATERIALS AND METHODS
TIMER2.0 TIMER2 (tumor immune estimation resource, version 2, http:// timer.comp-genomics.org/) is a website for systematical evaluation of immune infiltrates in various cancers (12). In the present study, the "Gene_DE module" was applied to explore the differential expression between tumor tissues and adjacent normal controls for MMP7 across all TCGA cancer types. In addition, "Immune-Gene module" was used to analyze the relationship between the expression of MMP7 and immune infiltrates, estimated by multiple immune deconvolution methods, including CIBERSORT, CIBERSORT-ABS, EPIC, MCPCOUNTER, QUANTISEQ, TIMER, and XCELL algorithms, across all TCGA cancers. In our study, macrophages, dendritic cells, and cancer-associated fibroblasts (MF, DCs, and CAFs, respectively) were selected. Moreover, the "Gene_Corr" module of TIMER2 was applied to evaluate the heatmap data of the correlation between MMP7 and some selected genes found in all TCGA tumors. The P-values and the correlation (cor) values were acquired using the partial Spearman's correlation test with the "Purity Adjustment" option.

GEPIA2
GEPIA2 (Gene Expression Profiling Interactive Analysis, version 2, http://gepia2.cancer-pku.cn/#index) is an updated and enhanced web server for large-scale expression profiling and interactive analysis between tumor and normal tissues from the TCGA and the GTEx (Genotype-Tissue Expression) projects (13). For certain cancer types with only a few or without normal samples, the "Expression Analysis-Box Plot" module of GEPIA2 was used to evaluate and compare the MMP7 expression between tumor tissues and their paired normal controls of the GTEx database with the setting of "Match TCGA normal and GTEx data". In addition, a pathological stage analysis across all TCGA cancers was performed using the "Expression Analysis-Pathological Stage Plot" module of GEPIA2 and yielded violin plots of MMP7 expression. Furthermore, the correlative prognostic analysis of MMP7 across all TCGA tumors was executed via the combined application of the "Survival Map" and "Survival Analysis" module of GEPIA2, which was performed using a Kaplan-Meier curve. In addition, the "Expression Analysis-Similar Genes Detection" module of GEPIA2 was used to search for the top 100 MMP7-related genes derived from all TCGA tumor tissues and corresponding normal tissues. obtained from six available datasets of tumors were explored based on the data from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset. Student's t test was used to calculate the P-value with a cutoff of 0.05 (*< 0.05; **< 0.01; ***< 0.001).
cBioPortal As an open-access web resource, cBioPortal (www.cbioportal. org) can be used to explore, visualize, and analyze multidimensional cancer genomics data (16,17). In our study, the "TCGA Pan Cancer Atlas Studies-Cancer Types Summary" module of this website was used to obtain the genetic alterations of MMP7, including the detailed alteration frequencies, mutation types, and copy number alterations (CNA) in all TCGA cancer types. Furthermore, the "Mutations" module was used to acquire the information of the mutated site for MMP7. Next, the "Comparison/Survival" module was used to produce data about the survival differences across all TCGA cancers with or without MMP7 genetic alterations via the concurrent generation of Kaplan-Meier curves and log-rank P-values.
String STRING (https://string-db.org/, version 11.0) is a web server that can predict protein-protein interactions (PPI) (18). This website was searched by inputting MMP7 and Homo sapiens with several basic settings: (1) network type-full network, (2) meaning of network edges-evidence, (3) active interaction sourcesexperiments, (4) minimum required interaction score-low confidence (0.150), and (5) max number of interactors to show no more than 50 interactors in the first shell and custom value with three maximum interactors in the second shell, obtaining 50 experimentally determined MMP7 interacting proteins and the PPI network.

Metascape
Metascape (http://metascape.org) is a reliable, productive, and intuitive resource for gene annotation and gene list enrichment analysis (19). In this study, a gene list, including the top 100 MMP7-correlated genes from the tool of GEPIA2 and the 50 experimentally determined MMP7 interacting genes from String were inputted. Next, H. sapiens was inputted as species in step 2, and the "Express Analysis" module was performed in step 3. Finally, the gene list analysis report was exported for the further Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and gene ontology (GO) enrichment analysis.

Other Web Servers
OSppc ("https://bioinfo.henu.edu.cn/Protein/OSppc.html") is a web server named Online consensus Survival analysis web server based on Proteome of Pan-cancers, including TCGA, RPPAs (reverse-phase protein arrays) data, and CPTAC mass spectrometry data. LOGpc ("https://bioinfo.henu.edu.cn/ DatabaseList.jsp"), named Long-term Outcome and Gene Expression Profiling Database of pan-cancers, which is a web server encompassing 209 expression datasets, provides different survival terms for distinct cancers. In this study, OSppc was used for pan-cancer differential expression and prognosis analysis at the protein level and LOGpc with multiple data sources including TCGA was utilized for pan-cancer prognosis analysis as well.
TMB and TSI were performed via the "Pan-Cancer mutational analysis" module on a Chinese website ("https:// www.aclbi.com/static/index.html#/"). The proportional Venn diagram, KEGG pathway, and GO enrichment (based on the data derived from the tool of metascape) dot bubbles and histograms were obtained from another Chinese bioinformatic website ("http://www.bioinformatics.com.cn/"). Both websites are online platforms for data analysis and visualization.

Immunohistochemistry Staining and Analysis
The present study obtained ethics approval from the Ethics Committee of Affiliated People's Hospital, Jiangsu University. The validation of MMP7 expression was performed via immunohistochemical staining according to the manufacturer's instructions. Briefly, formalin-fixed paraffin-embedded tissue sections were subject to heat induced epitope retrieval at pH 6.0 for 20 min heating step with the citric buffer. MMP7 antibody was obtained from Proteintech (10374-2-AP, 1:1000 dilution), followed by HRP-labeled secondary antibody (1:5000) and DAB coloring. Images were scanned and analyzed using the Image J software (v1.51n), allowing the quantitation of mean density (integrated optical density (IOD)/area) of the positive staining. GraphPad Prism 6 software was used for the statistical analysis and the results were considered to be significant when P value ≤ 0.05, indicated with * in the figure legends (*≤ 0.05; **< 0.01; ***< 0.001; ****< 0.0001).

Aberrant Expression of MMP7 in Different Cancers
TIMER2 was used to determine the expression of MMP7 across various TCGA tumors. As illustrated in Figure 1A, the expression of MMP7 in the tumor tissues of cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal papillary cell carcinoma (KIRP), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), and thyroid carcinoma (THCA) was significantly higher than in the normal tissues (P < 0,001). A remarkably higher expression of MMP7 in the tumor tissue of skin cutaneous melanoma (SKCM) was found compared to metastatic tissue (P < 0.01). It is intriguing that the expression level of MMP7 in the tumor tissues was found to be considerably lower compared to the control tissues in breast invasive carcinoma (BRCA), two pathological types of kidney malignant diseases, namely kidney chromophobe (KICH) and kidney renal clear cell carcinoma (KIRC) with P < 0.001, and liver hepatocellular carcinoma (LIHC) with P < 0.05.
For certain cancer types without or with limited normal tissues shown in Figure 1A, the normal tissues of the GTEx dataset were involved as controls by using GEPIA2. The results indicate that the expression of MMP7 in the tumor tissues of lymphoid neoplasm diffuse large B-cell lymphoma (DLBC) and ovarian serous cystadenocarcinoma (OV) was considerably higher than the normal controls ( Figure 1B, P < 0.05), while lower MMP7 expression was detected in SKCM tumor tissue when compared with normal tissue ( Figure 1B, P < 0.05). No significant difference in MMP7 expression between the tumor tissues and normal tissues in the remaining TCGA cancer types (adrenocortical carcinoma [ACC], bladder urothelial carcinoma [BLCA], and others) was determined based on whatever TCGA dataset or TCGA plus GTEx dataset was selected ( Figures 1A, B).
The results of the CPTAC dataset revealed that the protein expression level of MMP7 is notably higher in the primary tumor tissues of colon cancer, LUAD, and uterine corpus endometrial carcinoma (UCEC) than in the normal tissues ( Figure 1C, P < 0.001). The protein expression level of MMP7 was also found to be significantly lower in the primary tumor tissues of breast cancer ( Figure 1C, P < 0.05) and clear cell RCC ( Figure 1C, P < 0.001) when compared with normal tissues. Yet, MMP7 expression level was significantly lower in ovarian cancer tumor tissues ( Figure 1C, P < 0.01) when compared with normal tissue, which is opposite to the result shown in Figure 1B. To further explore pan-cancer differential expression of MMP7 at the protein level, OSppc for pan-cancer differential expression was applied. The results were almost consistent with what is illustrated in Figure 1C. Besides, the OSppc database also showed that there was no significant difference in protein expression of MMP7 between the tumor tissues and normal tissues in GBM (P = 0.1003, figure not shown).
Moreover, immunohistochemistry staining was performed using nine pairs of matched cancerous and para-cancerous tissues in STAD and COAD, respectively, to verify the expression of MMP7. As illustrated in Figure 2, the expression of MMP7 was significantly higher in STAD tumor tissue ( Figure 2B) when compared with the para-cancerous tissue ( Figure 2A) (P < 0.01, Figure 2C), findings that are consistent with online data from TCGA and CPTAC datasets. Similar results were obtained from the matched cancerous tissue ( Figure 2E) and para-cancerous tissue ( Figure 2D) in COAD (P < 0.001, Figure 2F). Furthermore, the clinicopathological information of the above-mentioned STAD and COAD patients is summarized in Supplementary Table 1.
By using the "Expression Analysis-Pathological Stage Plot" module of GEPIA2, a significant association between MMP7 expression and the different pathological stages in several cancer types ( Figure 1D

Prognostic Analysis of MMP7 Across All TCGA Cancers
To evaluate the prognostic value of differential expression of MMP7 in all TCGA tumors, the correlation between MMP7 expression and survival data was determined via GEPIA2. As presented in Figure 3A, high expression of MMP7 was significantly associated with poorer overall survival (OS) in KIRC (P = 0.031), acute myeloid leukemia (LAML) (P = 0.00035), brain lower grade glioma (LGG) with P = 0.0086, and LIHC (P = 0.012). As for the disease-free survival (DFS) analysis, low expression of MMP7 was found indicating remarkably better prognosis in KIRC (P = 0.0027), LGG (P = 0.0025), sarcoma (SARC) (P = 0.044), and thymoma (THYM) with P = 0.0023 ( Figure 3B). Additionally, a meta-analysis (Cox regression analysis) was performed to assess the prognostic value of the MMP7 in different cancer types based on the online platform "https://xenabrowser.net/datapages/". As shown in Supplementary Figure S1, high expression of MMP7 correlated remarkably with poor OS in LIHC (P < 0.001), PAAD (P = 0.03), KIRC (P = 0.03), LAML (P < 0.001), and LGG (P = 0.022), findings that are almost consistent with the results in Figure 2A.
Moreover, OSppc for pan-cancer prognosis analysis was used to evaluate the prognostic role of MMP7 in different cancer types at the protein level. As shown in Supplementary Figure S2, lower protein expression of MMP7 was significantly associated with better OS in KIRC (P = 0.0444), which was parallel with the prognostic value of MMP gene expression in KIRC ( Figure 3A).
At the same time, LOGpc with multiple data sources including TCGA was utilized for pan-cancer prognosis analysis of MMP7. It was demonstrated that higher expression of MMP7 was associated with better OS in DLBC (Supplementary Figure  S3A PAAD (K) (GSE62452, P = 0.0334). Additionally, the relationship between differential expression of MMP7 and the potential clinical significance across all the TCGA cancer types is summarized in Supplementary Table S2.

Genetic Alteration Analysis of MMP7
Genetic alteration analysis of MMP7 was performed using cBioPortal. The results demonstrated that cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) had the highest alteration frequencies involving MMP7, which was found to be around 8% with "amplification" as the elementary alteration type ( Figure 4A). Additionally, the "amplification" type of copy number alterations (CNA) was the main type found in several cancers, including HNSC (~5% frequency), OV (~5% frequency), BLCA (~4% frequency), and SARC (~3% frequency). Moreover, all uterine carcinosarcoma (UCS), ESCA, mesothelioma (MESO), and KIRC cases with genetic alterations had amplification of MMP7, with frequencies around 4%, 3%, 1%, and < 1%, respectively ( Figure 4A). All testicular germ cell tumors (TGCT) with a~1.5% frequency, prostate adenocarcinoma (PRAD) with a~1% frequency, and KIRP (< 1% frequency) cases with genetic alterations were found to hold the copy number deletion of MMP7 as their alteration type ( Figure 4A). Figure 4B illustrates the detailed types, sites, and case numbers for the genetic MMP7 alterations and also showed that the cardinal type of genetic MMP7 alteration is the missense mutation. Furthermore, the F65L mutation located at the peptidoglycan binding domain (PG_binding_1 domain containing 31-82 aa) was associated with the maximum cases with MMP7 mutations, which was detected in one case of CESC, two cases of UCEC, and one case of GBM ( Figure 4B).
Furthermore, the potential correlation between the genetic alteration of MMP7 and the prognosis across all TCGA cancer types was determined. It was found that OV cases with altered MMP7 indicated better clinical outcomes in disease-specific survival (P = 0.0318) but not OS (P = 0.056), DFS (P = 0.248), and progression-free survival (P = 0.124) when compared with the unaltered group ( Figure 4C). Figure 4D reveals that UCEC cases with unaltered MMP7 demonstrated poorer outcome with respect to DFS (P = 0.0348) and disease-specific survival (P = 0.0385) but not OS (P = 0.126) and progression-free survival (P = 0.202) than the unaltered group. For the other TCGA cancers (except KIRC with limited cases and only one case of censoring data in the altered MMP7 group), no significant difference in the prognostic data between the altered and unaltered MMP7 groups was detected (data not shown).
Additionally, the association between the expression of MMP7 and tumor mutational burden/microsatellite instability (TMB/MSI) was explored via the open-access platform, "http://www. bioinformatics.com.cn/", based on the R software v4.0.3. As illustrated in Figure 5A, a positive association between MMP7

Immune Cell Infiltration of MMP7 in TCGA Cancer Types
The primary role of MMP7 is to facilitate degradation of the extracellular matrix (ECM) (2), which constitutes the major components of the tumor microenvironment (TME) together with blood vessels, signaling molecules, immune cells, and fibroblasts (20,21). TME and cancer cells and cross-talk between them play a critical role in cancer development and progression (22). Hence, the potential correlation between the expression of MMP7 and the infiltration levels of CAFs/tumor- infiltrating immune cells across all TCGA cancers using EPIC, MCPCOUNTER, XECLL, TIDE, CIBERSORT, CIBERSORT-ABS, and QUANTISEQ methods was explored. As shown in Figure 6A, a considerably positive correlation between the expression of MMP7 and the infiltration of CAFs was indicated in BRCA-LumA, CESC, COAD, HNSC, human papillomavirus (HNSC-HPV), KIRC, KIRP, LIHC, LUAD, LUSC, OV, TGCT, and THCA based on all EPIC, MCPCOUNTER, XECLL, and TIDE algorithms. Figure 6B illustrates the specific scattergrams for the above-mentioned tumor types derived from the MCPCOUNTER algorithm, which was applied as an example in the present study (data not shown for EPIC, XECLL, and TIDE algorithms).
Furthermore, a significantly positive correlation between the expression of MMP7 and the infiltration of dendritic cells (DCs) in KIRP and PRAD, between MMP7 expression and infiltration of activated-DC in BLCA, BRCA-basal, THCA, and THYM, between MMP7 expression and resting-DC infiltration for CHOL, LGG, LIHC, LUAD, PRAD, SKCM-Primary, and THCA was found (Figure 7).

Functional Enrichment Analysis of MMP7-Related Genes
To explore the potential molecular mechanisms of MMP7 in tumorigenesis, MMP7-interacting genes and MMP7-correlated genes were screened out using the STRING and GEPIA2 tools, respectively, for the pathway and process enrichment analysis. Figure 8A illustrates a total of 50 experimentally determined MMP7 interacting genes in the PPI network. Next, the top 100 MMP7-correlated genes, among which five genes (E74 Like ETS Transcription values were listed. The maximum positive correlation coefficients were chosen to determine the corresponding heatmap, indicating a positive correlation between MMP7 expression and the five genes described above in a large proportion of all the TCGA cancers ( Figure 8B). Furthermore, the above two groups of genes were performed for the intersection analysis, and one common gene, Secreted Phosphoprotein 1 (SPP1), was obtained and presented in a proportional Venn diagram ( Figure 8C). The above two groups of gene datasets were further utilized to perform KEGG and GO enrichment analysis using Metascape and another open platform, "http://www.bio-informatics.com.cn/". The dot bubble of the KEGG enrichment analysis showed all of the potential pathways implicated in the effects of MMP7 on tumorigenesis in which "RNA degradation" might be a particular pathway with the most significant P-value and the maximal gene counts ( Figure 8D). The GO (molecular function) enrichment analysis data of Figure 8E listed the top 10 cellular biological functions ranked by their P-values, including poly(A)-specific ribonuclease, endopeptidase inhibitor, peptidase inhibitor, endopeptidase regulator, metalloendopeptidase inhibitor, 3'-5'exoribonuclease, peptidase regulator, exoribonuclease (producing 5'-phosphomonoesters), exoribonuclease, and enzyme inhibitor activities. In addition, typical enriched GO terms consisting of three categories, namely, biological process (blue), molecular function (green), and cellular component (red) are demonstrated in Figure 8F. Each category showed the top 10 GO terms according to the P-values.

DISCUSSION AND CONCLUSIONS
MMP7 is a well-established and important participant in regulating various pathophysiological processes, including wound healing, aging, bone growth, remodeling, and signal transduction pathways that control inflammation, cell growth, and angiogenesis (7,(23)(24)(25). Besides, MMP7 has been recognized as an oncogenic factor that manages the oncogenesis and progression of multiple tumors by mediating differentiation, proliferation, invasion, and metastasis of cancer cells (10). In this study, a comprehensive pan-cancer analysis of MMP7 across all the TCGA cancers to uncover the potential roles and underlying mechanism of MMP7 in the occurrence, development, and clinical outcomes of different cancers is presented.
Consistent with previous studies, heterogeneous expression of MMP7 across all TCGA cancers with downregulation in BRCA, KIRC, KICH, and LIHC and upregulation in most of the other cancers was obtained (10,11), while low expression of MMP7 was obtained in SKCM tumor tissue, a finding that was opposite to that in a previous study (26). Larger sample sizes and the setting of "Match TCGA normal and GTEx data" in this study might contribute to the different results. Due to the limitation of the CPTAC dataset, only MMP7 protein expression in breast cancer, clear cell RCC, colon cancer, LUAD, ovarian cancer, and UCEC was obtained. Parallel to the MMP7 gene expression level, the MMP7 protein was downregulated in breast cancer and clear cell RCC. Abundant expression of MMP7 protein was detected in colon cancer, LUAD, and UCEC in this study, a finding that was similar to those from previous studies (27)(28)(29)(30). Interestingly, the results concerning the overexpression of MMP7 gene in ovarian cancer shown in Figure 1B was not reproduced at the protein level as shown in Figure 1C. Previous studies have also demonstrated that gene expression analysis of MMP7 was inconsistent with its protein expression (31). The opposite results illustrated in Figures 1B, C were most likely based on the different databases and inconsistent sample sizes. MMP7 exhibits elevated expression in multiple cancer types, including OV (31,32). On that account, the expression of MMP7 in OV at both the gene and protein levels needs to be confirmed in a larger series in further studies. The correlation between the expression of MMP7 and the pathological stages in different cancer types was evaluated first. As the tumor progressed, MMP7 expression increased, suggesting that MMP7 plays an essential role in the tumorigenesis and development of KICH, KIRC, PAAD, THCA, SKCM, and OV. KIRC, LAML, LGG, and LIHC patients with high expression of MMP7 were significantly correlated with worse OS; KIRC, LGG, SARC, and THYM patients with high expression of MMP7 were remarkably associated with worse DFS. LOGpc analysis with multiple data sources also proved that high expression of MMP7 indicated poor OS in STAD, KIRP, LIHC, LUCA, and PAAD. Overexpression of MMP7 is always correlated with poor prognosis, which has been demonstrated in many previous studies for gastric cancer (33,34), pancreatic cancer (35,36), colorectal cancer (37), bladder cancer (38), ovarian cancer (32), and others. In the present study, different results addressing the prognostic value of MMP7 expression in various cancer types were obtained, which might have stemmed from larger sample sizes and multiple datasets. Overall, these results suggest that MMP7 could function as an adverse prognostic factor for different cancers.
Defined as a copy number increase of a restricted region of a chromosome arm, gene amplification is a typical alteration in cancer (39,40). Moreover, gene amplification is considered to be a major cause of tumorigenesis (40). In our study, gene amplification was detected in most TCGA cancer types and presented the highest alteration frequency. According to the Pfam database, the protein sequence of MMP7_Human (consists of 267 amino acids) has two major domains, PG_binding_1 and Peptise_M10. It was shown that the primary type of genetic alteration of MMP7 in tumors is the missense mutation in which F65L mutation located at PG_binding_1 domain has the maximum number of cases containing the MMP7 mutation. Furthermore, the F65L mutation appears to be capable of inducing a frame shift mutation of the MMP7 gene, causing the translation from the amino acid F (Phenylalanine) to L (Leucine) at the 65th site of MMP7 protein sequence, which could be a risk factor of oncogenesis in CESC, UCEC, and GBM. However, this finding still needs further investigation. We then focused on the prognostic value of genetic MMP7 alterations in different cancers. Altered MMP7 is correlated with better disease-specific survival in OV yet is associated with worse DFS and disease-specific survival in UCEC, findings that suggest that genetic alterations of MMP7 may play a significant role in different cancers. Nevertheless, extremely limited studies showing the relationship between the genetic alteration of MMP7 and clinical outcomes in various cancers have been published.
It is clear that TMB is a promising biomarker for prediction of sensitivity to immune checkpoint inhibitor treatments in several cancer types, especially in cancers with MSI (41,42). The potential correlation between the expression of MMP7 and TMB/MSI across various cancers in the present study was assessed. According to the results, MMP7 might be a potential target for cancer immune therapy in LGG, ACC, BRCA, LUAD, LUSC, STAD, CESC, UCEC, LIHC, HNSC, COAD, TGCT, DLBC, and CHOL. However, the influence of the expression of MMP7 on TMB/MSI has rarely been reported previously. Accumulating evidence indicates that CAFs, tumor-associated immune cells, and ECMs in TME can interact with cancer cells thus playing essential roles in tumor growth, invasion, antitumor immunotherapy, and prognosis (10,(43)(44)(45). Hence, the correlation between MMP7 expression and cancer immunity was assessed. It was shown that the expression of MMP7 positively correlated with the infiltration of CAFs in most TCGA cancer types and was positively associated with the infiltration of DC/activated-DC/resting DCs and MF/M0/M1 in some specific tumor types, indicating that MMP7 may also reflect the immune status in different cancers. Of note, MMP7 expression positively correlated with M0 and M1 yet negatively correlated with M2 in BRCA, implying that MMP7 overexpression may cause a shift toward an M1-like phenotype. Previous studies have shown that tumor-associated macrophages in TME predominately resemble the M2-polarized cells and correlate with poor prognosis (46,47). Thus, MMP7 expression may contribute to better clinical outcomes. In-depth investigations are still needed to further investigate the association between the MMP7 expression and tumor-associated macrophage infiltration and the underlying molecular mechanisms.
Nowadays, with the development of computational biology and bioinformatics, multiple computational methods, especially KEGG pathways and GO terms, have been widely applied to describe specific pathways and biological processes in different cancers (48,49). In the present study, 50 MMP7-interacting genes and 100 MMP7-correlated genes were first obtained using STRING and GEPIA2, respectively. Functional enrichment analysis using KEGG pathways and GO terms methods with the 150 genes was then performed. It was demonstrated that these genes may play important roles in the "RNA degradation", "poly(A)-specific ribonuclease activity", "cell-cycle checkpoint", and other pathways. Previous studies have shown that RNA degradation is a process during the cell cycle that regulates RNAdependent mechanisms, with ribonucleases (endonucleases and exonucleases) along with other enzymes and cofactors that are also involved (50,51). Moreover, both altered RNA processing and DNA damage checkpoints are considered to be strongly linked to pathogenesis of human cancers (52)(53)(54). Taken together, these data suggest that RNA processing and DNA damage checkpoints might implicate the functional mechanisms of MMP7 and that MMP7 might be a potential therapeutic drug target in different cancers.
To summarize, our study systemically analyzed MMP7 in a pancancer manner, evaluating the potential association of the expression of MMP7 with the clinical outcome (Supplementary  Table S2), pathological stages, TMB, MSI, and CAFs/immune cell infiltration in various cancer types, which will assist in providing a clearer panorama for the roles of MMP7 across human cancers.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Affiliated People's Hospital, Jiangsu University. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
NM and YL undertook the data analysis and wrote the manuscript. LX, PJ, and XB designed the study and edited the manuscript. JD and YW aided in the methodology and data duration. XZ, FY, YZ, and JZ assisted in the conception and data representation. NM and LX contributed to the funding. All of the authors reviewed the data and analysis, and read and approved the manuscript.

FUNDING
Database mining, data collection, and manuscript preparation of this work was supported by the funding from the Innovation and Entrepreneurship Plan of Jiangsu Province (No. SCBS202101) and the Affiliated People's Hospital of Jiangsu University (No. KFB2020005) from LX. Figure editing and layout and language editing service of this work were supported by the Medical and Health Guidance Project of Xiamen City (3502Z20214ZD1219) from NM.