The compendium of matrix metalloproteinase expression in the left ventricle of mice following myocardial infarction.

Matrix metalloproteinases (MMPs) are proteolytic enzymes that break down extracellular matrix (ECM) components and have shown to be highly active in the myocardial infarction (MI) landscape. In addition to breaking down ECM products, MMPs modulate cytokine signaling and mediate leukocyte cell physiology. MMP-2, -7, -8, -9, -12, -14, and -28 are well studied as effectors of cardiac remodeling after MI. While 13 MMPs have been evaluated in the MI setting, 13 MMPs have not been investigated during cardiac remodeling. Here we measure the remaining MMPs across the MI time continuum to provide the full catalogue of MMP expression in the left ventricle after MI in mice. We found that MMP-10, -11, -16, -24, -25, and -27 increase after MI, while MMP-15, -17, -19, -21, -23b, and -26 did not change with MI. For the MMPs increased with MI, the macrophage was the predominant cell source. This work provides targets for investigation to understand the full complement of specific MMP roles in cardiac remodeling.


INTRODUCTION
In response to myocardial infarction (MI), the myocardium responds by undergoing a repair process that starts with a robust inflammatory response and ends with scar formation (11,13). Wound healing ranges from formation of a stable scar to progression to heart failure (11,13). During the inflammatory response, necrotic myocytes and damaged extracellular matrix (ECM) components from the ischemic area are enzymatically broken down. Removal is governed by proteases, in particular the matrix metalloproteinases (MMPs). ECM break-down provides the platform on which new ECM is deposited to form the infarct scar (11,13).

METHODS
Tissue collection. The samples used for immunoblotting and multiplex immunohistochemistry were previously collected from other projects and included in the mouse Heart Attack Research Tool (mHART) database and tissue bank (10). All animal procedures were performed in compliance with the Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Mississippi Medical Center. The mice used were wild-type C57BL/6J 3-to 6-mo-old mice (3 males and 3 females for each time point). MI was induced by permanently occluding the coronary artery according to the Guidelines for Experimental Models of Ischemia and Infarction and as previously described, and mice were given buprenorphine (0.5 mg/kg) before the surgery (9,20,25,37,48). Echocardiography was performed under isoflurane euthanasia using the Vevo 2100 (Visual Sonics, Toronto, ON, Canada) as previously described and outlined in the Guidelines for Measuring Cardiac Physiology in Mice (9,20,30,37). Heart rates were Ͼ400 beats/min and were not different among groups (ANOVA P ϭ 0.47). The left ventricle (LV) was sliced into three sections. The apex and base sections were separated into remote left ventricle control (LVC) and infarct (LVI; including border) zones and snap-frozen for immunoblotting. The mid-papillary section was fixed in zinc-formalin and paraffin-embedded for histological evaluation. Infarct area was measured by staining with 1% 2,3,5-triphenyltetrazolium chloride and calculating the volume percentage of LV that was infarcted.
Immunoblotting. Immunoblotting was performed according to the published guidelines (3). For first-pass assessment, the samples for each time were pooled together to allow the complete time course of remote and infarct region groups to all be run on one gel. By densitometry, MI time points that showed peak expression were selected for individual variability analysis. Total protein (10 g) samples were run on 4 -12% Criterion XT Bis-Tris precast gels (Bio-Rad, Hercules, CA) and were transferred onto Trans-Blot Turbo Transfer Pack Nitrocellulose Membranes (Bio-Rad). The membranes were stained with Peirce Reversible Protein Stain Kit for nitrocellu-lose membranes (Thermo Scientific, Rockford, IL), and blots were normalized to their total membrane stain (20). Total membrane stains are shown in Supplemental Fig. S1 (Supplemental material for this article can be found online at https://doi.org/10.6084/m9.figshare. 10271102). Membranes were blocked with Blotting Grade Blocker (Bio-Rad) in a 5.0% triphosphate buffer solution and incubated overnight with the primary antibody at 4°C against the MMP (Table  1), followed by incubation at room temperature for 1 h with appropriate secondary antibodies. MMP-9 was incubated with the secondary antibody anti-goat IgG (1:5,000 dilution, Cat. No. PI-9500; Vector Laboratories), and all other MMPs were incubated with anti-rabbit IgG (1:5,000 dilution, Cat. No. PI-1000, Vector Laboratories). The antibodies used all recognized mouse MMPs and showed specificity for that MMP. The membrane-type MMPs are all activated by furin intracellularly, and as such only the active forms were present on the membrane (21,45). Homogenized spleen and thyroid tumor samples from mice were used as positive controls. Chemiluminescence visualization was conducted by incubating the membrane with Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare) for 5 min and visualizing the blot with an ImageQuant LAS 4000 and ImageQuantTL V8.1 software (GE Healthcare).
MMP mRNA expression in MI macrophages and fibroblasts. To further validate cell source for MMPs and TIMPs, we evaluated MMP mRNA expression in previously collected transcriptomic data sets from macrophages and fibroblasts isolated from the infarct region on MI days 1, 3, and 7 and compared with day 0 no MI LV (38,39).
Statistical analysis. Data were analyzed according to the recommendations of the Statistical Considerations in Reporting Cardiovascular Research and are presented as means Ϯ SE for echocardiography and SD for all other data (28)   variables were compared using one-way ANOVA with Newman-Keuls posttest. Comparisons between two groups were made using unpaired t-test or Mann-Whitney test. MMP-25 was compared with infarct wall thinning by Pearson correlation. Statistical significance was set at P Ͻ 0.05.

RESULTS
Proof of MI was obtained using echocardiography. As shown in Fig. 1, the mice displayed LV dilation by dimensions, impaired myocyte contractility by fractional shortening, and infarct wall thinning. Infarct sizes ranged from ϳ40 to 50%, and lung mass increased, indicating the development of pulmonary edema.
Rigor and reproducibility assessment. To assess the rigor and reproducibility of our immunoblotting, both operators performed immunoblotting for MMP-8 and MMP-9, which was then compared with previous findings (17,46). MMP-8 and -9 increased following MI, with the greatest increase in expression of the pro form occurring on day 1 and the active form occurring on day 3. The interperson variation ratios for MMP-8 ( Fig. 2) and MMP-9 ( Fig. 3) were excellent. Therefore, our results were highly consistent between operators in both technique and analysis and were consistent with past reports (6,14).   These results indicate that a strong proportion of MI MMP protein expression is localized to the macrophage. These results are also consistent with MMP expression in MI neutrophils being low due to degranulation of vesicles to release the MMPs. MMP expression in MI macrophages and fibroblasts by RNAseq. Transcriptomics were previously performed on LV macrophages isolated from LV on day 0 and MI days 1, 3, and 7 (results summary in Table 3, with full individual results in Supplemental Table S1) (38). Of the 24 MMPs evaluated, 17 showed differential expression in infarct macrophages. MMP-8, -9, and -25 show prominent increases in gene expression on MI day 1, matching the immunoblotting results. For MMP-10, -11, -16, -24, and -27, immunoblotting showed increased expression, and macrophage expression showed no change or decreases in gene expression. This indicates that the increases in MMP-10, -11, -16, -24, and -27 shown by immunoblotting were not due to macrophage as a source, or there was a mismatch between macrophage gene and protein expression.
Transcriptomics were previously performed on LV fibroblasts isolated from LV on day 0 and MI days 1, 3, and 7 (results summary in Table 4, with full individual results in Supplemental Table S2) (39). Of the 24 MMPs evaluated, five showed differential expression in infarct macrophages.
MMP-16 shows prominent increases in gene expression on MI day 3 in fibroblasts, whereas immunoblotting for LV MMP-16 showed increased expression on day 1. There was no difference or no change in expression of MMP-10, -11, -16, -24, and -27 in fibroblasts. This indicates that the increases in MMP-10, -11, -16, -24, and -27 shown by immunoblotting were not due to fibroblasts as a source, or there was a mismatch between fibroblast gene and protein expression.

DISCUSSION
The goal of this study was to evaluate the MMP family members that had not been previously examined in the MI LV. We validated our immunoblotting approach using MMP-8 and MMP-9 as controls and evaluated MMP- MMP-8 and -9 were evaluated as positive control experiments, as both have previously been shown to increase after MI (6). We showed excellent reproducibility both between the  duplicate evaluations (performed by A. R. Kaminski and E. T. Moore) and with the literature. In our study, MMP-10 and -11 increased on MI day 1 in the infarcted zone. Although MMP-10 is not constitutively expressed in healthy tissue, it is heavily secreted by macrophages during acute inflammation in damaged or infected tissue (36). During infection with Pseudomonas aeruginosa, MMP-10 is responsible for mitigating the inflammatory actions of M1 macrophages as well as activating M2 anti-inflammatory macrophages (36). In cardiac tissue, TIMP-4 regulates MMP-10 during the process of LV remodeling, and in end-stage heart failure patients myocardial samples revealed a positive correlation between upregulation of MMP-10 and LV dilation, indicating that MMP-10 influenced ECM structure (47). Inhibition of MMP-10, accordingly, may help to prevent LV dilation. Substrates of MMP-11 include collagen IV and VI, fibronectin, ␣2-macroglobulin, and insulin-like growth factor-binding protein 1 (1). MMP-11 is regulated in M2 macrophages compared more so than M1 macrophages (14). Immunohistochemistry showed no difference in MMP-11 localized in macrophages and a decrease in neutrophils, whereas total MMP-11 increased, suggesting that neutrophils and macrophages are not the major source of MMP-11 in MI. High MMP-11 activity on MI day 1 in M1 macrophages suggests that MMP-11 may be involved in pro-inflammatory signaling or ECM turnover, which are hallmarks of the inflammatory phase. MMP-16, also known as MT3-MMP, had highest expression on MI day 1 in both the pro and active forms, which suggests that it may have an active role in early myocardial remodeling. MMP-16 may have an indirect role through its activation of MMP-9 and, more importantly, MMP-2 (31,49). MMP-2 cleaves collagen I, IV, V, VII, and X, laminin, aggrecan, fibronectin, and tenascin (4). This widespread degradative effect upon multiple extracellular matrix substrates suggests that MMP-16 alone may have a more influential role in directing LV remodeling than previously predicted.
MMP-24, also known as MT5-MMP, is a membrane-type MMP expressed in the nervous system (2). Importantly, MMP-24 cleaves MMP-2, which degrades type IV collagen to break down basement membranes (4,32). MMP-24 has been evaluated as a potential therapy for Alzheimer's disease due to regulatory effects on amyloid precursor protein metabolism (2). Because MMP-24 was highest on day 1 after MI, its role to activate MMP-2 is likely important in cardiac remodeling (15).
MMP-25, also known as MT6-MMP, is a membrane-type MMP expressed in lung, spleen, and leukocytes (43). MMP-25 substrates include fibrin, fibronectin, collagen I, and collagen IV (12). In the context of MI, pro-inflammatory N1 neutrophils contribute to LV wall thinning after MI by generating large amounts of MMP-12 and MMP-25 (34). MMP-25 showed highest expression on MI day 7 and independently tracked with infarct wall thinning on MI day 7. Therefore, MMP-25 may be relevant to neutrophil prolongation of pro-inflammation.
MMP-27 has a unique COOH-terminal extension, causing it to be retained in the endoplasmic reticulum (8  phages in the cycling human endometrium (7). CD163 and CD206 are expressed by M2 macrophages (7). In our study, MMP-27 was increased in its active form (58 kDa) on MI day 3, when macrophages are undergoing a metabolic shift and upregulating genes associated with M2 macrophages (38). Whether or not MMP-27 plays a role in phenotypic shift of macrophages has not been examined. By immunohistochemistry, MMP-11, -16, -24, -25, and -27 increased with MI. These results are consistent with the immunoblotting results. Whereas MMP-10 increased by immunoblotting, a similar increase by immunohistochemistry was not observed. By transcriptomics, the macrophage was not a predominant source of MMP-10, whereas cardiac fibroblasts showed decreased expression with MI. This indicates that the immunoblotting results likely reflect the combination of effects across cell types. Transcriptomics, immunoblotting, and immunohistochemistry combined confirmed increased MMP-11, -16, -24, and -25 in the MI LV and indicated that macrophages were the predominant source of these MMPs.
Further studies on the MMPs that were differentially expressed are needed to better understand the roles of each MMP as well as their functional relationships to each other. There is a need to follow up on the MMPs elevated after MI to determine their specific cause and effect relationship in the left ventricle. MMPs have a number of roles in the myocardium, including processing of extracellular matrix substrates, inflammatory mediators, and growth factors, and the implications of these findings need to be determined (11,18,24). In particular, the role of the substrate fragments generated by these MMP should be investigated, as substrate proteolysis is known to release active biopeptides that influence MI remodeling (29). In addition to clarifying mechanisms of action of these MMPs in the mouse myocardium after MI, evaluating whether these MMPs also are elevated in humans after MI would provide a translational link to our results.