Pan-cancer analysis of MMP1 expression and its prognostic value
Herein, we first analyzed the expression of MMP1 in 35 common cancer types to explore its possible roles in pan-cancer. Compared with that in the normal samples, the expression of MMP1 was significantly upregulated in 16 cancer types, namely, BRCA, BLCA, CESC, CHOL, COAD, ESCA, GBM, HNSC, LUAD, LUSC, PAAD, PCPG, READ, STAD, THCA, and UCEC while the expression of MMP1 was significantly downregulated in KICH and KIRP (Fig. 1A). Afterward, GEPIA databases were used to validate the expression of MMP1 in pan-cancer. MMP1 expression levels in BRCA, BLCA, CESC, CHOL, COAD, DLBC, ESCA, HNSC, LUAD, LUSC, PAAD, READ, SKCM, STAD, UCEC, and UCS were upregulated compared with those in the corresponding normal controls (all p < 0.05, Fig. 1B).
Next, the KM-plotter was used to analyze overall survival (OS) according to the expression of MMP1 in pan-cancer. The results showed that a higher expression of MMP1 was associated with significantly worse OS in BRCA, CESC, KIRC, KIRP, LIHC, LUAD, PDA, PCPG, SARC, STAD, THYM, THCA, and UCEC patients (all p < 0.05, Fig. 2A). However, there were no significant differences in terms of OS in BLCA, ESCA, ESCC, HNSC, LUSC, OV, READ, and TGCT patients (all p > 0.05).
Furthermore, a higher expression of MMP1 was significantly associated with worse relapse-free survival (RFS) and distant metastasis-free survival (DMFS) in BRCA patients (p < 1e-16 and 3.2e-08, respectively) (Fig. 2B, C). These results demonstrated that MMP1 expression may serve as an unfavorable prognostic marker for patients with BRCA. Subsequently, the TCGA data validated that MMP1 was upregulated and associated with poor RFS in BRCA patients (Fig. 3A, B). Collectively, these data demonstrated that MMP1 was upregulated in BRCA, thereby indicating that MMP1 might act as a crucial regulator in the carcinogenesis of BRCA.
Development and evaluation of MMP1 expression-correlated clinicopathologic nomogram
The immunohistochemical results of BRCA confirmed the expression of MMP1 in 60 BRCA patients from the second affiliated hospital of Nanhua university (Fig. 3C). We also conducted a heat map to explore correlations between low and high MMP1 expression with clinicopathologic characteristics, such as gender, age, clinical stage, T, N, and M stage, and found that MMP1 expression was significantly associated with T stage and N stage (Figure 3D).
In addition, multivariate Cox regression analyses were carried out to elucidate whether or not MMP1 was an independent prognostic indicator of BRCA. The results of which illuminated that age, AJCC stage, and MMP1 were independent prognostic predictors in the multivariate Cox analysis (all p < 0.01). Based on the results above, a clinicopathologic nomogram with an optimal concordance index (C-index, 0.76) that incorporated MMP1 of high expression to another two clinical characteristics containing age and AJCC stage was developed to predict individual OS of 1, 3, and 5 years (Figure 4A, B). The calibration plot was portrayed to confirm the satisfactory predictive discrimination of the nomogram and was discovered to be closer to the ideal curve (Figure 4C), which presented the perfect stability of the nomogram. Briefly, the efficiency of the prognostic nomogram was clarified from multiple aspects.
Analysis of biological pathways between high and low expression of MMP1 in BC patients
GSEA was conducted to decipher the primarily enriched signaling pathways and biological functions between the high- and low-MMP1 groups. As shown in Figures 5A and B, results using the KEGG database demonstrated that cell cycle, olfactory transduction, pentose and glucuronate interconversions, and porphyrin and chlorophyll metabolism were fundamentally enriched in the high-MMP1 group.
GO analysis and KEGG pathway enrichments were performed to explore the underlying interplay of these valid DEGs. As shown in Fig. 5C, 10 Biological Processes were enriched, such as cerebellum, metencephalon, and hindbrain development. Ten molecular functions including receptor and signaling receptor ligand activities were related. Moreover, 10 cellular components regarding cornified envelope were found.
Regarding the KEGG pathway analysis as shown in Fig. 5D, high expression of MMP1-related pathways were enriched, such as the PPAR signaling pathway, IL-17 signaling pathway, and protein digestion and absorption, which revealed potential mechanisms and pathways activated during tumor progression. Immune processes and regulations were significantly enriched in both GO and KEGG analyses.
Correlation of MMP1 expression with immune cell infiltration in BRCA
TME cells constitute a vital element of tumor tissue. Increasing evidence has elucidated their clinicopathological significance in predicting outcomes and therapeutic efficacy. Therefore, the association of MMP1 expression with TME score in BRCA was analyzed. As shown in Fig. 6A, the TME score of the high-MMP1 group was significantly higher than that of the low-MMP1 group in BRCA. Furthermore, this difference was due to the differences in immune and stromal scores.
MMP1 plays an important role in the immune system; thus, we explored the correlation between MMP1 expression and immune cell infiltration. As shown in Fig. 6B, the immune infiltrating cells that positively correlated with MMP1 expression included macrophages M0 (p < 0.001), activated T cell CD4 memory (p < 0.001), activated dendritic cells (p < 0.001), neutrophils (p < 0.001), macrophages M1 (p < 0.001), regulatory T cells (Tregs) (p = 0.014), and activated NK cells (p = 0.039), whereas the immune infiltrating cells that negatively correlated with MMP1 expression included resting mast cells (p < 0.001), naive B cells (p < 0.001), T cell CD8 (p < 0.001), monocytes (p = 0.007), and gamma delta T cells (p = 0.037) in BRCA. Afterward, we analyzed the differences in the infiltration of various immune cells within high- and low-MMP1 groups. As shown in Fig. 6C, the infiltration levels of macrophages M0, macrophages M1, activated T cell CD4 memory, and activated dendritic cells in the high-MMP1 group were significantly higher than those in the low-MMP1 group. Alternatively, the infiltration level of T cell CD8 and resting mast cells in the high-MMP1 group was significantly lower than that in the low-MMP1 group. Our results demonstrated that MMP1 together with immune cells is involved in the progression of BRCA. According to the essential connection between TMB and immune checkpoints, it is necessary to investigate the relation between TMB and MMP1 expression. The correlation curve from Fig. 6D showed a significant positive correlation between TMB and MMP1 expression (R = 1, p < 2.2e-16). Finally, immunofluorescent staining results of BRCA confirmed the above results (Fig. 7A).
Correlation between MMP1 and immune checkpoint gene expression in BRCA
It is well-known that immunotherapy plays an important role in BRCA, and the expression of immune checkpoint genes can not only predict prognosis but also predict the response to immunotherapy. Therefore, the correlation between MMP1 expression and immune checkpoint gene expression was analyzed. As shown in Fig. 7B, MMP1 was positively correlated with almost all immune checkpoint genes, including TNFSF4, CD28, CD86, CTLA4, LAIR1, LAG3, TNFSF14, ICOS, CD40, CD160, TNFSF9, CD276, IDO1, LGALS9, PDCD1LG2, CD274, TIGIT, TNFRSF25, TNFSF15, KIR3DL1, TNFRSF14, TNFRSF8, CD80, CD48, PDCD1, CD70, TNFRSF9, NRP1, HAVCR2, and TNFRSF4, whereas it was negatively correlated with TNFSF14 and CD160. Furthermore, a positive correlation was found between MMP1 and most immune checkpoint gene expression, indicating that MMP1 can be used as a biomarker to predict immunotherapy response.
Clinical drug sensitivity analysis and immunotherapy efficacy evaluation of prognostic characteristics of MMP1
Currently, the research on drug treatment of BRCA is a hot topic and has received much attention. We performed a drug sensitivity analysis within high- and low-MMP1 groups to explore the relationship between MMP1 and drug sensitivity. As shown in Fig. 8A, the sensitivity of 20 drugs related to BRCA treatment was screened out. Among them, the IC50 of AFP464 in patients with high-MMP1 expression was lower than that in patients with low-MMP1 expression, indicating a better efficacy in the high-MMP1 expression group. Therefore, MMP1 may be involved in the resistance to these drugs in BRCA. The machine learning-based score (IPS) of four subtypes (CTLA-4_neg_PD-1_neg, CTLA-4_pos_PD-1_pos, CTLA-4_pos_PD-1_neg, and CTLA-4_neg_PD-1_pos) was calculated to predict the responses to ICI. Based on the expression of MMP1, all BRCA patients do not have an evident difference in responding to immunotherapy (Fig. 8B).