Disparate effects of MMP and TIMP modulation on coronary atherosclerosis and associated myocardial fibrosis

Matrix metalloproteinase (MMP) activity is tightly regulated by the endogenous tissue inhibitors (TIMPs), and dysregulated activity contributes to extracellular matrix remodelling. Accordingly, MMP/TIMP balance is associated with atherosclerotic plaque progression and instability, alongside adverse post-infarction cardiac fibrosis and subsequent heart failure. Here, we demonstrate that prolonged high-fat feeding of apolipoprotein (Apo)e-deficient mice triggered the development of unstable coronary artery atherosclerosis alongside evidence of myocardial infarction and progressive sudden death. Accordingly, the contribution of select MMPs and TIMPs to the progression of both interrelated pathologies was examined in Apoe-deficient mice with concomitant deletion of Mmp7, Mmp9, Mmp12, or Timp1 and relevant wild-type controls after 36-weeks high-fat feeding. Mmp7 deficiency increased incidence of sudden death, while Mmp12 deficiency promoted survival, whereas Mmp9 or Timp1 deficiency had no effect. While all mice harboured coronary disease, atherosclerotic burden was reduced in Mmp7-deficient and Mmp12-deficient mice and increased in Timp1-deficient animals, compared to relevant controls. Significant differences in cardiac fibrosis were only observed in Mmp-7-deficient mice and Timp1-deficient animals, which was associated with reduced capillary number. Adopting therapeutic strategies in Apoe-deficient mice, TIMP-2 adenoviral-overexpression or administration (delayed or throughout) of a non-selective MMP inhibitor (RS-130830) had no effect on coronary atherosclerotic burden or cardiac fibrosis. Taken together, our findings emphasise the divergent roles of MMPs on coronary plaque progression and associated post-MI cardiac fibrosis, highlighting the need for selective therapeutic approaches to target unstable atherosclerosis alongside adverse cardiac remodelling while negating detrimental adverse effects on either pathology, with targeting of MMP-12 seeming a suitable target.


Materials and methods
Animals. Mice on an Apoe −/− background were used, as a well-characterised and reproducible model of hypercholesterolaemia and spontaneous development of atherosclerosis, in comparison to Apoe +/+ mice which are not rendered hypercholesterolaemic upon high-fat feeding and are therefore considered athero-resistant 14,15 .
The housing and care of the animals and all the procedures used in these studies were performed in accordance with the ethical guidelines and regulations of the University of Bristol and the UK Home Office, with all in vivo studies approved by the Animal Welfare and Ethics Review Body (AWERB), University of Bristol. Adherence to the ARRIVE guidelines (https:// arriv eguid elines. org) for the reporting of animal in vivo experiments was also followed. Immunohistochemistry. Successive 3 μm paraffin sections were dewaxed and rehydrated. Incubation with Bloxall (Vector Labs, Cambridgeshire, UK) was used to inhibit endogenous peroxidase activity followed by 20% (v/v) goat serum in PBS. For the identification of cardiomyocytes, sections were subsequently incubated with an anti-α-actinin rabbit antibody (11,313-2-AP; Proteintech, Manchester, UK) at 2 μg/mL, followed by biotinylated anti-rabbit IgG goat antibody (A-11012; Invitrogen; Fisher Scientific, Loughborough, UK) at 20 μg/ mL, whereas capillaries were identified (endothelial cells), through incubation with biotin-conjugated Griffonia Simplicifolia Isolectin B4 (L2140; Sigma, Poole, UK) at 10 μg/mL in PBS. All sections were then incubated with horseradish peroxidase-labelled Extravidin™ (diluted 1:400 in 1% (w/v) BSA in PBS) and colour developed with SIGMAFAST™ 3,3′-Diaminobenzidine (D4293; Sigma, Poole, UK), following which nuclei were counterstained with Mayer's haematoxylin. Positive cells were counted and expressed as a percentage of the total number of nucleated cells. A negative control was always included. In this case, the primary antibody was substituted with mouse or rat IgG of the same dilution.
For dual immunohistochemistry, sections were first incubated for 30 min at room temperature with biotin-conjugated Griffonia Simplicifolia Isolectin B4 (IB4) (L2140; Sigma, Poole, UK) at 10 μg/mL, followed by horseradish peroxidase-labelled Extravidin (diluted 1:400 in 1% (w/v) BSA in PBS) and colour developed with SIGMAFAST™ 3,3′-Diaminobenzidine (D4293; Sigma, Poole, UK). Sections were then washed with PBS and incubated overnight at 4ºC with a rabbit polyclonal antibody against anti-α-actinin (11,313-2-AP; Proteintech, Manchester, UK) at 0.25 μg/mL diluted in PBS. The next day sections were incubated with goat anti-rabbit biotinylated secondary antibody (E0432; Dako, High Wycombe, UK) diluted 1:200 in 1% (w/v) BSA in PBS, followed by alkaline phosphatase-labelled Extravidin (diluted 1:400 in 1% (w/v) BSA in PBS) and colour developed with SIGMAFAST™Fast Red TR/Naphthol AS-MX (F4648; Sigma, Poole, UK). Subsequently, nuclei were counterstained with Mayer's haematoxylin (MHS1; Sigma, Poole, UK). The specificity of the immunolabelling was demonstrated by inclusion of a negative control using isotype-specific nonimmune serum or IgG. Images were acquired from a brightfield microscope using ImageProPlus image analysis software (Media Cybernetics). IB4 positive cells (brown colour) depicting capillaries were counted and expressed as the number of capillaries per × 10 magnification field. The number of α-actinin -positive cells (myocytes; red colour) in the same regions as examined for capillary density, were also counted. The reproducibility of assessments was tested for interobserver variation by using an unweighted κ test (κ = 0.75).

Morphological analysis.
The presence and prevalence of atherosclerotic coronary arteries was quantified within myocardial cross-sections subjected to EVG histochemical staining and expressed as percentage of atherosclerotic burden, which refers to the percentage of coronary arteries containing atherosclerotic plaques per mouse heart cross-section.
Analysis of myocardial capillary density. Within Griffonia Simplicifolia Isolectin B4 (IB4) immunolabelled sections, the number of capillaries (IB4-positive endothelial cells) was counted in nine × 10 magnification fields using a computer-assisted image analysis system (ImageProPlus, Media Cybernetics), and expressed as capillary density per × 10 field of myocardium.
Analysis of myocardial myocyte density. Within anti-α-actinin immunolabelled sections, the number of myocytes was counted in nine × 10 magnification fields using a computer-assisted image analysis system (ImageProPlus, Media Cybernetics), and expressed as myocyte density per × 10 field of myocardium.

Statistical analysis.
All values are presented as mean ± standard error of the mean (SEM). When comparing group means, groups were tested for normal distribution with D' Agostino-Pearson test. For normally distributed groups, an unpaired t-test, or ordinary one-way ANOVA with Tukey's multiple comparisons post-hoc test was used. For groups that were not normally distributed, a Mann-Whitney test or Kruskal-Wallis test with www.nature.com/scientificreports/ Dunn's multiple comparisons post-hoc test was used. Survival data were compared using the Kaplan-Meier method, with the log-rank (Mantel-Cox) test for significance. Statistical significance is reported as *P < 0.05, **P < 0.01, ***P < 0.001, and †P < 0.05 is used when non-parametric test was used.

Results
Effect of MMP/TIMP modulation on sudden death in Apoe −/− mice. We have previously demonstrated that modulation of MMP expression and/or activity using the approaches detailed within this study affects atherosclerosis within the brachiocephalic artery of short-term (8-10 weeks) high-fat fed Apoe −/− mice 5,13,20 . In addition, prolonged high-fat feeding of Apoe −/− mice is associated with the development of unstable atherosclerotic plaques within the brachiocephalic artery, alongside indicators of MI and sudden death 21,22 . Accordingly, specific MMP/TIMP modulation models were developed in Apoe −/− mice in order to determine their contributory role to coronary artery atherosclerosis and associated effect on myocardial fibrosis and remodelling (summarised in Supplementary Table 1). Apoe/MMP double knockout mice presented with divergent effects on the incidence of sudden death. MMP-7 deficiency resulted in significantly lower survival after 36 weeks of high-fat feeding (Fig. 1A). Contrastingly, MMP-12 deficiency markedly reduced the frequency of sudden death in Apoe −/− mice in comparison to their wild-type controls (Fig. 1C), whereas no significant difference on survival was observed between MMP-9 double knockout and Apoe single knockout control mice (Fig. 1B). Similarly, inhibition of MMP activity with a broadspectrum MMP inhibitor, RS-130830, or TIMP-1 deficiency had no significant effect on the prevalence of sudden death (Fig. 1D,E). Conversely, helper-dependent adenovirus-mediated TIMP-2 overexpression dramatically lowered the proportion of animals suffering sudden death and promoted survival of Apoe −/− mice over the course of 18 weeks of high-fat diet (Fig. 1F).

Evidence of unstable coronary artery atherosclerotic plaques alongside signs of myocardial infarction in Apoe −/− mice.
Given the prevalence of sudden death within Apoe −/− mice alongside the effects of MMP/TIMP modulation on survival rates, hearts were histologically assessed for signs of coronary atherosclerosis and MI. Left anterior descending (LAD) coronary arteries from 36-week high-fat fed Apoe −/− mice regularly contained large occlusive atherosclerotic plaques with the presence of buried fibrous layers ( Fig. 2A), indicative of a healed plaque rupture 23 . Interestingly, analysis of serial sections, downstream of coronary arteries with signs of healed plaque rupture, revealed areas of MI (Fig. 2Bi), as surmised through loss of myocytes alongside signs of fibrosis, in comparison to healthy myocardium (Fig. 2Bii). Established atherosclerotic plaques containing buried fibrous layers were also observed within proximal coronary arteries (Fig. 2C,D), alongside occlusive atherosclerotic lesions within distal coronary arteries (Fig. 2E,F). Collectively, these observations demonstrate that prolonged high-fat feeding of Apoe −/− mice is commonly associated with the presence of unstable coronary atherosclerosis and signs of MI, supporting its role as a valuable animal model of spontaneous atherosclerosis-induced MI, and post-MI cardiac remodelling.
Plaque progression was further assessed through the quantification of coronary arteries displaying greater than 50% stenosis. In line with the findings pertaining to atherosclerotic burden, Apoe/TIMP-1 double knockout mice showed a significant increase in plaque stenosis in comparison to strain-matched Apoe single knockout controls (1.8-fold; P < 0.05; Supplemental Fig. 1e). However, deficiency of MMP-7, MMP-9, or MMP-12, MMP inhibition, and TIMP-2 overexpression did not affect the proportion of coronary plaques with > 50% stenosis (Supplemental Fig. 1).

Effect of MMP/TIMP modulation on cardiac fibrosis in atherosclerotic Apoe −/− mice. Given the
indications of MI in long-term high-fat-fed Apoe −/− mice (Fig. 2) alongside the prevalence of highly stenotic coronary atherosclerosis (Figs. 2, 3, 4), cardiac fibrosis was investigated as a surrogate marker of MI, secondary to unstable coronary artery atherosclerosis. Hearts from Apoe/MMP-7 double knockout mice displayed significantly larger areas of cardiac fibrosis in comparison to strain-matched Apoe single knockout controls (3.5-fold; P < 0.001; Fig. 5A). Conversely, deficiency of TIMP-1 in Apoe −/− mice resulted in a significantly diminished area of cardiac fibrosis compared to hearts from their respective control mice (67%; P < 0.05; Fig. 6B). However, MMP-9 or MMP-12 deficiency, broad spectrum MMP inhibition (delayed or throughout), or TIMP-2 overexpression had no significant effect on the extent of cardiac fibrosis (Figs. 5 and 6).
Effect of MMP/TIMP modulation on capillary to myocyte ratio in atherosclerotic Apoe −/− mice. In order to further investigate the role of MMPs and TIMPs to atherosclerosis-induced post-MI cardiac remodelling, the capillary to myocyte ratio was investigated as a surrogate marker of cardiac remodelling in atherosclerotic Apoe −/− mice with signs of MI and cardiac fibrosis. MMP-7 or MMP-12 deficiency, and MMP inhibition (delayed or throughout) had no effect on the capillary to myocyte ratio (Figs. 7 and 8) or on the absolute number of capillaries or myocytes (Supplemental Fig. 2). However, MMP-9 deficiency resulted in a modest, www.nature.com/scientificreports/ yet statistically significant, reduction in capillary to myocyte ratio (13%; P < 0.05; Fig. 7B). This reduction in the capillary to myocyte ratio was associated with an increase in myocyte number (1.3-fold; P < 0.01; Supplementary  Fig. 2b). Conversely, hearts from Apoe/TIMP-1 double knockout mice had significantly less capillaries (33%;  www.nature.com/scientificreports/ P < 0.05; Supplemental Fig. 2e) but comparable number of myocytes, although this failed to translate into a difference in capillary to myocyte ratio (Fig. 8B). TIMP-2 overexpression had no effect on capillary and myocyte numbers (Supplemental Fig. 1f), or their ratio (Fig. 8C).

Discussion
An imbalance between MMP and TIMP levels facilitates ECM remodelling during the development, progression, and rupture of atherosclerotic plaques, and during subsequent cardiac fibrosis and remodelling secondary to MI 1,2 . Consequently, this study utilised genetic deficiency of select MMPs/TIMPs alongside strategies to retard MMP activity to assess the contribution of specific MMPs and TIMPs to the development of coronary atherosclerosis and related cardiac fibrosis and remodelling in hypercholesterolaemic Apoe −/− mice. The principal findings indicate divergent roles for MMPs and TIMPs in the development of coronary atherosclerosis and

MMP-7.
We demonstrate that Apoe/MMP-7 double knockout mice presented with reduced atherosclerotic burden and increased cardiac fibrosis, alongside a heightened incidence of sudden death after 36-weeks of highfat feeding, suggesting that MMP-7 has an adverse role in atherosclerosis, whereas, after an MI, MMP-7 activity Atherosclerotic burden (%) is presented as average plaque burden/coronary artery/mouse. Statistical significance is reported as ***P < 0.001 using unpaired Students t-test. Scale bar represents 1 mm and is applicable to all panels. In all cases, n = 10/group. www.nature.com/scientificreports/ prevents aberrant fibrosis. The indicated adverse role for MMP-7 in atherosclerosis progression, and potential plaque vulnerability is in accordance with a previous study assessing brachiocephalic artery atherosclerosis in mice 5 , alongside reports of increased macrophage foam cell MMP-7 expression in rupture-prone human carotid plaques 27 , and an association in plasma MMP-7 levels with plaque phenotype and mortality rates 28 . Mechanistically, MMP-7 can cleave N-cadherin from the cell surface of VSMCs, promoting their apoptosis 29 , which may clarify the increased plaque VSMC content observed in MMP-7 knockout mice 5 , and support the association of increased MMP-7 levels with VSMC apoptosis and fibrous cap thinning in murine and human plaques 27,29 . Conversely, our findings imply MMP-7 imparts a protective role during the post-MI cardiac remodelling process. During post-MI remodelling, the healing scar tissue predominantly consists of cardiac fibroblasts and marginating macrophages within the infarct border region, with macrophages and cardiomyocytes (within remote regions) both shown to express comparable MMP-7 levels after MI 30 . However, in an LAD coronary artery ligation mouse model, MMP-7 deficiency was associated with increased macrophage accumulation after MI 30 ; Figure 4. Effect of MMP inhibition and TIMP modulation on coronary artery atherosclerosis in apolipoprotein E-deficient (Apoe -/-) mice. Quantification and representative images of coronary atherosclerotic burden as assessed in EVG-stained hearts from (A) 36-week high-fat fed Apoe -/mice treated with or without a broadspectrum MMP inhibitor, (B) 36-week high-fat fed Apoe -/mice deficient for TIMP-1 in comparison to Apoe single knockout controls, and (C) 18-week high-fat fed Apoe -/mice with administration of a TIMP-2 helperdependent (HD) adenovirus or an empty HD as control. Atherosclerotic burden (%) is presented as average plaque burden/coronary artery/mouse. Statistical significance is reported as **P < 0.01 using unpaired Students t-test. Scale bar represents 1 mm and is applicable to all panels. In A, n = 6-8/group; in B, n = 7-9/group; in C, n = 7-9/group. www.nature.com/scientificreports/ although effects on invasion and regression were not explored to yield mechanistic insight into the heightened macrophage accumulation. Unfortunately, although the aforementioned study demonstrated that MMP-7 deficiency had no effect on infarct size, it did not directly assess cardiac fibrosis 30 . In the context of fibrosis, despite its multiple ECM substrates MMP-7 has been characterised as pathologic mediator of fibrosis expansion in the lungs 31 and liver 32,33 . However, in support of our observations, administration of a selective MMP-7 inhibitor suppressed cardiac fibrosis in a mouse model of uraemia-induced cardiac remodelling 34 . We therefore propose that macrophage accumulation and their accompanying release of MMP-7 within advanced plaques and during the early phase of post-MI remodelling perturb mesenchymal cell growth and survival (VSMCs in plaques and cardiac fibroblasts within the myocardium), and their associated release of ECM proteins, such as collagens.
Therefore, we speculate that while inhibiting MMP-7 activity may maintain the VSMC-and ECM-rich fibrous cap of plaques -protecting them from ensuing rupture, targeting MMP-7 during the early stages of post-MI cardiac remodelling would increase cardiac fibrosis due to loss of ECM/collagen turnover, predisposing to heart failure. As such, therapeutic strategies to blunt MMP-7 activity in unstable plaques or increase its expression after an MI would require approaches permitting localised delivery and action to limit the potential adverse effects. Cardiac fibrosis is presented as average fibrillar collagen content (blue) in µm 2 /heart cross-section/mouse. Statistical significance is reported as ***P < 0.001 using unpaired Students t-test. Scale bar represents 1 mm and is applicable to all panels. In all cases, n = 10/group. www.nature.com/scientificreports/

MMP-9.
In this study, Apoe/MMP-9 double knockout mice had no significant changes in terms of occurrence of sudden death or coronary artery atherosclerosis. The previously reported contribution of MMP-9 to the development of atherosclerotic lesions in hypercholesterolaemic mouse models varies, depending on the vascular bed assessed. An adverse role was attributed to MMP-9 in the formation of aortic lesions 35 , while it was deemed protective against plaque development within the brachiocephalic artery 5 , through enhanced plaque stability 5,36 via promotion of VSMC proliferation, migration, and intimal growth [37][38][39] .
Regarding cardiac remodelling, Apoe/MMP-9 double knockout mice displayed similar cardiac fibrosis to their relevant control mice, with a modest reduction in the capillary to myocyte ratio. A similar lack of effect on myocardial collagen content was reported in two previous studies using the LAD coronary artery ligation-induced MI mouse model alongside MMP-9 deficiency 40 or transgenic overexpression of MMP-9 41 . However, other murine studies deploying global deficiency or macrophage-restricted overexpression of MMP-9 within the LAD ligation-induced MI model have associated MMP-9 expression with reduced cardiac fibrosis post-MI, alongside diminished macrophage accumulation and left ventricular enlargement, but enhanced neovascularization 42,43 . Figure 6. Effect of MMP inhibition and TIMP modulation on cardiac fibrosis in apolipoprotein E-deficient (Apoe -/-) mice. Quantification and representative images of cardiac fibrosis as assessed in WGA-stained hearts from (A) 36-week high-fat fed Apoe -/mice treated with or without a broad-spectrum MMP inhibitor, (B) 36-week high-fat fed Apoe -/mice deficient for TIMP-1 in comparison to Apoe single knockout controls, and (C) 18-week high-fat fed Apoe -/mice with administration of a TIMP-2 helper-dependent (HD) adenovirus or an empty HD as control. Cardiac fibrosis is presented as average fibrillar collagen content (green) in µm 2 / heart cross-section/mouse. Statistical significance is reported as †P < 0.05 using Mann-Whitney test. Scale bar represents 1 mm and is applicable to all panels. In A, n = 7-9/group; in B, n = 10/group; in C, n = 4-5/group. www.nature.com/scientificreports/ We observed no effect of MMP-9 deficiency on capillary number, with the reduction in the capillary to myocyte ratio driven through an increase in myocyte number. However, we did not explore if the increased myocyte number afforded from MMP-9 deficiency is through effects on myocyte survival or proliferation. In support of the former, autophagy promotes survival of myocytes during the early inflammatory response post-MI 44 , and MMP-9 inhibition increased autophagy in rat hearts with post-MI chronic heart failure 45 .

MMP-12.
Coronary arteries from Apoe/MMP-12 double knockout mice displayed significantly reduced atherosclerosis which was associated with a decreased incidence of sudden death over the course of thirty-six weeks of high-fat feeding. Previous studies in animal models and humans support a pro-atherosclerotic role for MMP-12. Rabbit studies utilising macrophage-specific overexpression of MMP-12 demonstrated that plaque development and progression were accelerated 46,47 . Similarly, global deletion of MMP-12 or administration of an MMP-12 specific inhibitor in Apoe-deficient mice, blunted brachiocephalic artery atherosclerotic plaque formation and progression, respectively 5,48 . Finally, macrophage foam cell expression of MMP-12 in human carotid plaques is associated with symptomatic severity and predicts future adverse cardiovascular events 49 . Our findings strengthen the call for therapeutic clinical trials targeting MMP-12 for the prevention of atherosclerosis. Capillary to myocyte ratio is presented as average/heart cross-section/mouse. Statistical significance is reported as *P < 0.05 using unpaired Students t-test. Scale bar represents 1 mm and is applicable to all panels. In A, n = 9/group; in B and C, n = 10/group. www.nature.com/scientificreports/ Despite the protective effect on coronary artery atherosclerotic burden, MMP-12 deficiency had no effect on cardiac fibrosis or capillary to myocyte ratio after thirty-six weeks of high-fat feeding. Previous findings have shown MMP-12 levels are rapidly elevated within the hearts of mice post-MI (through LAD ligation), predominantly due to the inflammatory response, particularly neutrophils 50,51 . Within the same model, MMP-12 inhibition did not affect infarct size or mRNA expression of fibrillar collagens (Col1a1 and Col3a1), although cardiac fibrosis was not evaluated 50 .
While the contrasting roles for MMP-12 in atherosclerosis and post-MI remodelling may be ascribed to different cell sources and substrates, and chronic versus acute inflammatory responses, respectively, hypercholesterolaemia and underlying atherosclerosis may also explain the divergent observations. High-fat feeding of Apoe-deficient mice and ensuing hypercholesterolaemia is associated with sustained monocytosis, contributing to macrophage accrual within developing atherosclerotic plaques 52 . Accordingly, in our study the coronary vessels and myocardium have been exposed to heightened inflammation for months before the formation of coronary Figure 8. Effect of MMP inhibition and TIMP modulation on capillary to myocyte ratio in atherosclerotic apolipoprotein E-deficient (Apoe -/-) mice. Quantification and representative images of capillary to myocyte ratio as assessed in isolectin B4-immunohistochemically stained hearts from (A) 36-week high-fat fed Apoe -/mice treated with or without a broad-spectrum MMP inhibitor, (B) 36-week high-fat fed Apoe -/mice deficient for TIMP-1 in comparison to Apoe single knockout controls, and (C) 18-week high-fat fed Apoe -/mice with administration of a TIMP-2 helper-dependent (HD) adenovirus or an empty HD as control. Capillary to myocyte ratio is presented as average/heart cross-section/mouse. Scale bar represents 1 mm and is applicable to all panels. In A and B, n = 10/group; in C, n = 7-9/group. www.nature.com/scientificreports/ atherosclerosis and associated cardiac changes, whereas hypercholesterolaemia and monocytosis are absent within wild-type mice before LAD ligation-induced MI. This difference between our atherosclerosis-induced cardiac fibrosis model and the LAD ligation model may explain the observed cardiac-related discrepancies.

TIMP-1.
Atherosclerotic plaque burden and degree of stenosis were increased within the coronary arteries of Apoe/TIMP-1 double knockout mice, mirroring previous observations within the aortic root 53 and consistent with the reported athero-protective effect afforded from systemic TIMP-1 overexpression in Apoe-deficient mice 54 . Yet, a comparable study revealed no beneficial effects of TIMP-1 overexpression on aortic atherosclerosis 55 . Similarly, TIMP-1 deficiency was shown to exert no change in atherosclerotic plaque size within the aortic root 56 or brachiocephalic arteries 12 of Apoe-deficient mice. These discrepancies suggest that TIMP-1 levels, and by inference inhibition of MMP activity, may differentially affect atherosclerotic plaque growth depending on the vascular bed and stage of atherosclerosis, with influence on plaque composition more relevant than plaque size in advanced plaques 24 . Interestingly, the increased coronary plaque burden observed in Apoe/TIMP-1 double knockout mice was associated with a marked reduction in cardiac fibrosis. Comparable findings have been reported in the LAD ligation model 14-days post-MI 57 , implying the increased MMP activity provided from TIMP-1 deficiency promotes adverse myocardial remodelling, accelerating subsequent heart failure. Although, clinical evidence revealed elevated TIMP-1 expression in patients with extended cardiac fibrosis due to chronic pressure overload or heart failure 58,59 , which may reflect compensatory upregulation of TIMP-1 or the complex role of fibrosis post-MI. We also observed decreased capillary number within the hearts of Apoe/TIMP-1 double knockout mice, reflecting adverse myocardial wound healing due to perturbed neovascularisation. Indeed, fibroblast-derived TIMP-1 can prevent MMP-dependent cleavage and generation of anti-angiogenic collagen IV fragments 60 . Although, previous reports suggest that blocking TIMP-1 increases angiogenesis in an in vivo sponge model of angiogenesis 61 . These findings collectively suggest that TIMP-1 is protective against the development of coronary atherosclerosis, while aggravating cardiac fibrosis and affecting the remodelling process.

TIMP-2. Studies assessing brachiocephalic artery atherosclerosis have consistently demonstrated an athero-
protective role for TIMP-2 12,13 . However, TIMP-2 overexpression did not alter coronary artery plaque burden or degree of stenosis in the current study. This may be due to the reduced prevalence of stenotic plaques within this sub-study as mice were high-fat-fed for 18-weeks only, with potentially longer-term feeding required to detect an effect on plaque burden, as observed in MMP-7 and MMP-12-deficient mice.
Despite evidence demonstrating that TIMP-2 is upregulated in chronic pressure-overloaded human hearts 59 and post-MI in the mouse LAD ligation model, our findings did not establish a link between TIMP-2 overexpression and cardiac collagen density. This lack of effect may be associated with the short duration of high-fat feeding and associated diminished coronary atherosclerosis, resulting in less adverse cardiac events as seen in the longer-term experiments. Indeed, loss of TIMP-2 has been shown to hasten adverse post-MI remodelling, as evidenced through decreased collagen content alongside greater infarct size, both attributed to increased inflammation and associated MMP-14 expression and activity 62 . Although TIMP-2 overexpression did not impart any protective effects in atherosclerosis or cardiac fibrosis, longer-term studies would be appropriate to evaluate its therapeutic potential to combat the progression of both diseases. In addition, assessing the effect of TIMP-2 deficiency in long-term high-fat fed Apoe-deficient mice may also provide useful in the pathophysiological assessment of both pathologies.

Conclusions
This study highlights the divergent roles of specific MMPs and TIMPs in hypercholesterolaemia-induced coronary artery atherosclerotic plaque formation alongside cardiac fibrosis and remodelling, secondary to MI. Overall, the data suggest that nuanced and selective therapeutic approaches, that target specific MMPs or TIMPs during a defined therapeutic window, are necessary for treatments of both atherosclerosis and adverse post-infarction remodelling to avoid opposing deleterious effects on each pathology. The absence of adverse consequences on cardiac fibrosis with MMP-12 deficiency, supports targeting of MMP-12 for the treatment of unstable atherosclerosis. Finally, the concurrent spontaneous development of advanced coronary atherosclerosis and MI-associated cardiac fibrosis support use of the prolonged high-fat fed Apoe-deficient mouse as an alternative to the LAD occlusion model of nonperfused MI and models of reperfused MI 63 .

Data availability
The data that support the findings of this study are included within the manuscript or available from the corresponding author upon reasonable request. www.nature.com/scientificreports/