Inhibition of class I histone deacetylase activity represses matrix metalloproteinase-2 and -9 expression and preserves LV function postmyocardial infarction

Abstract

Left ventricular (LV) remodeling, after myocardial infarction (MI), can result in LV dilation and LV pump dysfunction. Post-MI induction of matrix metalloproteinases (MMPs), particularly MMP-2 and MMP-9, have been implicated as causing deleterious effects on LV and extracellular matrix remodeling in the MI region and within the initially unaffected remote zone. Histone deacetylases (HDACs) are a class of enzymes that affect the transcriptional regulation of genes during pathological conditions. We assessed the efficacy of both class I/IIb- and class I-selective HDAC inhibitors on MMP-2 and MMP-9 abundance and determined if treatment resulted in the attenuation of adverse LV and extracellular matrix remodeling and improved LV pump function post-MI. MI was surgically induced in MMP-9 promoter reporter mice and randomized for treatment with a class I/IIb HDAC inhibitor for 7 days post-MI. After MI, LV dilation, LV pump dysfunction, and activation of the MMP-9 gene promoter were significantly attenuated in mice treated with either the class I/IIb HDAC inhibitor tichostatin A or suberanilohydroxamic acid (voronistat) compared with MI-only mice. Immunohistological staining and zymographic levels of MMP-2 and MMP-9 were reduced with either tichostatin A or suberanilohydroxamic acid treatment. Class I HDAC activity was dramatically increased post-MI. Treatment with the selective class I HDAC inhibitor PD-106 reduced post-MI levels of both MMP-2 and MMP-9 and attenuated LV dilation and LV pump dysfunction post-MI, similar to class I/IIb HDAC inhibition. Taken together, these unique findings demonstrate that selective inhibition of class I HDACs may provide a novel therapeutic means to attenuate adverse LV remodeling post-MI.

the healing response within the myocardium after a myocardial infarction (MI) is complex and involves both temporal and regional changes, including inflammation, new tissue formation, and tissue remodeling (11). After coronary artery occlusion, there is an induction of bioactive peptides and cytokines, extracellular matrix (ECM) degradation, and subsequent recruitment of inflammatory cells to the site of injury (9, 16, 53). Alterations in the post-MI ECM architecture are largely attributed to changes in the expression of a number of matrix metalloproteases (MMPs) (50). Previous work has demonstrated that induction of MMP-2 gene expression can be detected by day 1 post-MI, reaches its maximum at 7 days, and then gradually decreases (43, 51). Activation of the MMP-9 promoter was detectable by 3 days, peaked by 7 days, and remained upregulated throughout the 28-day time course post-MI (43). The dramatic increases in both MMP-2 and MMP-9 have been proposed to contribute to the disruption of the cardiocyte-matrix interactive network, resulting in cardiocyte misalignment and slippage (53).

MMP-9-null mice show attenuated left ventricular (LV) dilation and improved LV function compared with wild-type mice after MI (14, 32). Loss of MMP-2 expression improves post-MI survival by both a decrease in cardiac rupture rate and better preservation of LV function by delaying and decreasing the extent of post-MI remodeling (14, 21, 37). Taken together, these studies have suggested that selective inhibition of MMPs after MI may be an effective treatment to diminish pathological remodeling and improve cardiac function after MI.

Regulation of MMP activity is complex and is controlled at several levels, including transcription, secretion, activation via proteolytic cleavage, and inhibition of activity by endogenous tissue inhibitors of metalloproteinases (TIMPs) (50). Transcriptional regulation is considered to be the rate-limiting step in MMP-9 synthesis (17, 34, 48). MMP-9 transcriptional activation is controlled by different stimuli, including growth factors that mediate through transcription factors and histone acetyltransferases (HATs) (15, 33, 42, 57). Importantly, HATs and their counterparts, histone deacetylases (HDACs), regulate gene expression not only by histone acetylation but also through the acetylation of transcription factors, coactivators, and repressors (8, 40). HDACs are grouped into four classes based on size and structure. Class I HDACs (HDAC1, HDAC2, HDAC3, and HDAC8) are ubiquitously expressed. Class II HDACs are expressed in a tissue-specific manner and are subgrouped as class IIa (HDAC4, HDAC5, HDAC7, and HDAC9) and class IIb (HDAC6 and HDAC10). Class III comprises NAD⁺-dependent deacetylases [sirtuin (SIRT)1–SIRT7]. HDAC11 is the sole member of class IV (19). HDAC inhibitors and studies using transgenic and knockout mouse models have revealed the importance of classes I and IIa in cardiac disease (38, 39). HDAC inhibitors are one of only a few classes of compounds that have been demonstrated by many studies to prevent or reverse cardiac remodeling. Both MMP-2 and MMP-9 play a major role in cardiac remodeling. Therefore, we hypothesized that HDAC inhibition would reduce MMP-2 and MMP-9 expression and attenuate the progression of post-MI LV adverse remodeling.

METHODS

Animal experiments.

Transgenic mice containing the Lacz reporter gene 3′ of the MMP-9 promoter on the CD-1 background strain were a gift from Dr. M.E. Fini and have been previously described by Mohan et al. (41). For MI experiments, coronary artery ligation was performed on 12- to 15-wk-old male wild-type and homozygous MMP-9 promoter-Lacz transgenic CD-1 mice as previously described (43). Briefly, the left anterior descending coronary artery was ligated, and MI was confirmed by LV blanching and ST segment elevation on the ECG. The class I/IIb HDAC inhibitor tichostatin A (TSA) was administered 12 h before left anterior descending coronary artery ligation. TSA (1 mg/kg) or vehicle (1% DMSO) was administered twice per day by intraperitoneal injection for the next 6 days. The class I/IIb inhibitor suberanilohydroxamic acid (SAHA; 100 mg·kg⁻¹·day⁻¹) was added to the drinking water of mice from immediately after recovery from left anterior descending coronary artery ligation until mice were euthanized. The class I inhibitor PD-106 (100 mg·kg⁻¹·day⁻¹) was administered by intraperitoneal injection immediately after ligation and once every other day for the next 6 days. Echocardiographic determinations of LV volumes and ejection fractions (40-MHz transducer, Vero 2100, Visual Sonics) were performed before MI induction and at 7 days after MI, as previously described (12). Briefly, the parasternal long-axis view of the LV was recorded, together with a recording of the surface ECG. LV volumes were determined by planimetry of the LV endocardial border at end diastole (frame with R wave) and end systole (smallest LV area in the cardiac cycle) and application of a variant of Simpson's algorithm (“method of disks”) (45, 52). Mice were then euthanized for immunohistochemical and biochemical analyses. All animal experiments were conducted under protocols approved by the Institutional Animal Care Committee of the Medical University of South Carolina in accordance with National Institutes of Health guidelines.

Heart harvest.

After terminal echocardiography while deeply anesthetized by inhalation of isoflurane, mice were weighed, and a midline sternotomy was performed to extirpate and weigh the heart. Some of the hearts were used for both biochemical and histological analysis. These hearts were sectioned along the long axis, such that the line of dissection would bisect the MI region. Hearts were also obtained from non-MI reference control mice, where MMP-9 reporter mice received either HDAC inhibitor or vehicle alone.

Western blot analysis.

Hearts were homogenized in 1% Triton X-100 lysis buffer, resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Membranes were blocked for 1 h using 5% milk in Tris-buffered saline-Tween 20 (TBST; 10 mM Tris, 0.1 M NaCl, and 0.1% Tween 20, pH 7.4). Blots were incubated with primary antibodies overnight at 4°C, washed five times (each for 5 min) with TBST, and then incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. After five washes (each for 5 min) with TBST, proteins were detected by enhanced chemiluminescence. The anti-HDAC1 antibody was from Cell Signaling (no. 2062), anti-HDAC2 antibody was from Invitrogen (no. 51-5100), and anti-HDAC3 antibody was from Abcam (ab7030-50).

Quantification of MMP-9 promoter activation.

MMP-9 transcriptional activation was assayed as a function of β-Gal staining in hearts by overnight incubation in substrate buffer (X-Gal, Sigma), as previously described (43). Hearts were photographed with the atria and right ventricles attached. The LV was separated, and the area and intensity of β-Gal staining on the LV epicardium as well as the LV epicardial area were determined from the two-dimensional photographs by digital planimetry (Amira). The area of β-Gal staining was normalized to LV epicardial area and expressed as a percentage (43).

Histology.

Hearts were dehydrated, embedded in paraffin, and sectioned at 5 μm. For MMP-2 immunolocalization, freshly dissected heart tissue was embedded in tissue freezing medium (OCT compound) in liquid nitrogen, and 15-μm-thick cryosections were prepared using a Leica cryomicrotome. For immunostaining, sections were blocked with 1% BSA (Sigma) for 1 h at room temperature and incubated overnight at 4°C with the following primary antibody: rabbit anti-MMP-2 antibody from Abcam (ab37150). After PBS rinses, sections were incubated with fluorescent-conjugated secondary antibodies (Molecular Probes) for 2 h at room temperature. Fluorescence staining was visualized using a Leica SP5 confocal microscope.

Immunohistochemical experiments on paraffin-embedded sections were performed using primary antibodies directed against MMP-9 (ab19047, Chemicon). Briefly, myocardial sections that contained the scar were blocked and then incubated overnight at 4°C with primary antibodies. After incubation with secondary antisera, MMP-9 was detected by visualization using a Leica SP5 confocal microscope. Secondary antibodies included Alexa 633-conjugated anti-chicken antibodies for β-Gal (catalog no. A21103, Molecular Probes) and Alexa 488-conjugated secondary antibodies for the other markers (catalog nos. A11006 and A11008, Molecular Probes). Negative controls for the fixed and frozen section immunohistochemistry included substitution with purified IgG (Sigma). A minimum of four hearts was used for each experimental condition. Four sections were taken from two different regions of the border region. Minimally, these sections would cover an 80-μm depth of the border zone around the infarct from each heart.

Zymography.

MMP-2 and MMP-9 zymographic levels in LV homogenates (20 μg total protein) were determined as a function of cleavage of a gelatin substrate as previously described (43). Briefly, samples were loaded on a native PAGE gel containing copolymerized protease substrate (gelatin). The presence of MMPs in the sample was determined as a function of the decrease in Coomassie brilliant blue staining of the digested gelatin at the position of the MMP.

HDAC activity.

HDAC activity was measured with a homogenous fluorescence release HDAC deacetylase assay (23). Primary tissue extracts were incubated with 4-methylcoumarin-7-amide (AMC)-K(Ac)GL-Ac substrate to assess HDAC1, HDAC2, HDAC3, HDAC6, and HDAC10 enzyme activities and AMC-K(TFA)GL-Ac substrate was used for class IIa HDAC4, HDAC5, HDAC7, HDAC8, HDAC9, and HDAC11 for 2 h at 37°C, as previously described (4, 23). Deacetylated AMC-KGL was sensitive toward lysine peptidase (1 mg/ml trypsin), and free fluorogenic AMC was generated, which was excited at 355 nm and observed at 460 nm. In brief, excess substrates were added (100 μM) together with 20 μg tissue extracts at a final volume of 25 μl and incubated at 37°C for 2 h. The assay conditions were set with excess substrate concentrations to insure linear deacetylation kinetics. For HDAC inhibitor experiments, exogenous HDAC inhibitors (typically diluted from 1,000× DMSO stock solutions using HDAC assay buffer), and corresponding volumes of DMSO were added in the control wells followed by 60 min of incubation at 37°C. Substrates were then added, and plates were returned to the 37°C incubator for 2 h. Finally, for all experiments, the tyrpsin peptidase developer was added at a final concentration of 1 mg/ml, and plates were read after 10 min at room temperature. AMC fluorescence was measured using Fluoroskan Ascent from Labsystems at exciation of 355 nm and emission of 460 nm, and background signals from predeveloped blanks were subtracted. Data were standardized using the control, and absolute deacetylated substrates were calculated based on the standard curve generated by nonacetylated AMC-KGL substrate under the same conditions.

Statistical analysis.

Changes in LV end-diastolic volume and ejection fraction were computed as a function of pre-MI values and expressed as percentages. Comparisons between groups were made using a Student's t-test (for pairwise comparisons) or ANOVA as appropriate. For two-way ANOVA, the independent blocks were the presence/absence of MI and treatment with either vehicle or TSA. Post hoc mean separation was performed using Bonferroni-adjusted pairwise comparisons. Statistical tests were performed using STATA (version 8, College Station, TX). Data are reported as means ± SE. P values of <0.05 were considered to be statistically significant.

RESULTS

Inhibition of pan-HDAC activity preserves LV function after 7 days post-MI.

In our first experimental set, mice were examined at 7 days post-MI based on previous findings that demonstrated increased MMP-2 and MMP-9 promoter activation by 1 and 3 days post-MI, respectively, which subsequently peaked at 7 days post-MI (43). LV end-diastolic volume was increased in both MI + TSA and MI groups after MI. Importantly, LV dilation was significantly lower in MI + TSA-treated mice compared with MI mice (Fig. 1, A and B). As expected, LV ejection fraction was decreased at 7 days after MI compared with baseline. However, there was a significant attenuation in the degree of LV pump function in TSA-treated mice (Fig. 1, C and D) compared with MI + vehicle mice. Post-MI survival rates of MI + TSA- versus MI + vehicle-treated mice were nearly identical.

Fig. 1.

Broad-spectrum histone deacetylase (HDAC) inhibition [tichostatin A (TSA)] preserves ventricular geometry and improves ventricular function in postmyocardial infarction (post-MI) mice. A: echocardiographic determinations of left ventricular (LV) end-diastolic volume were performed before MI induction (pre-MI) and at 7 days post-MI for the MI-only and MI group treated with TSA. LV end-diastolic volume was increased in both MI groups compared with pre-MI values. B: changes in end-diastolic volume from pre-MI values. TSA treatment attenuated the post-MI increase in LV end-diastolic volume compared with the MI-only group. C: pre-MI LV ejection fractions in mice randomized to the MI groups were similar to control values. Post-MI, LV ejection fraction was lower than pre-MI values in both MI groups, but the reduction was attenuated in the MI + TSA group. D: TSA inhibited the reduction in ejection fraction compared with the MI-only group. #P < 0.05 vs. control or pre-MI; *P < 0.05 vs. MI only.

Pan-HDAC inhibition represses MMP-2 and MMP-9 levels in the infarct zone.

Similar to previous results, MMP-9 promoter-β-Gal transgenic reporter mice had extensive β-Gal staining throughout the infarct and border regions at 7 days post-MI (Fig. 2A) (43). However, in MI + TSA-treated mice, the area of β-Gal staining was dramatically reduced compared with MI alone, and results from quantitative planimetry (Fig. 2B) demonstrated that TSA treatment reduced β-Gal staining by >2.5-fold in 7-day post-MI mice.

Fig. 2.

HDAC inhibition suppresses matrix metalloproteinase (MMP)-9 activation in response to MI. A: β-Gal-stained hearts from MMP-9 promoter reporter mice 7 days post-MI. TSA treatment significantly repressed the induction of MMP-9 promoter activity resulting from MI (right). B: graph showing quantitative values for areas of positive β-Gal staining, which were significantly lower in TSA-treated hearts compared with 7-day post-MI. #P < 0.05 vs. MI + vehicle.

As shown by representative zymograms (Fig. 3A), TSA treatment in hearts without MI had no detectable effect on MMP-2 or MMP-9 levels compared with no-MI + vehicle control hearts. Both MMP-2 and MMP-9 levels were dramatically increased in MI mice compared with non-MI control mice. In contrast, TSA treatment in the setting of MI reduced MMP-9 levels compared with MI + vehicle values. Specifically, MMP-2 and MMP-9 levels were increased ∼19- and 32-fold, respectively, in the MI + vehicle group compared with the no-MI group and were significantly reduced with TSA treatment (Fig. 3, B and C). The histological analysis of the MI region is shown in Fig. 4. The yellow boxes in Fig. 4, A and B, represent the border zones of MI and MI + TSA treatment, respectively. Representative confocal images of myocardial sections within the border region (high magnification of yellow boxes in Fig. 4, A and B, respectively) revealed that MMP-2 (Fig. 4, C and D, green) and MMP-9 (Fig. 4, E and F, green) immunostaining was reduced in MI + TSA hearts compared with MI + vehicle control hearts (Fig. 4). These data are consistent with the reduced MMP-2/MMP-9 levels shown in Figs. 2 and 3.

Fig. 3.

HDAC inhibition represses MMP-2 and MMP-9 levels in the infarct zone of the 7-day post-MI heart. A: representative gelatin zymogram of the infarct zone from 7-day post-MI hearts. MMP-2 and MMP-9 levels were increased over non-MI controls. TSA treatment significantly reduced post-MI MMP-9 expression and activity. B and C: gelatinographic active MMP-2 (B) and MMP-9 (C) levels expressed in arbitrary units. Active MMP-2 and MMP-9 levels were increased in the MI group (20- and 32-fold, respectively), and both were drastically reduced with TSA administration. *P < 0.05 vs. non-MI control; #P < 0.05 vs. MI + vehicle.

Fig. 4.

Immunohistochemistry reveals the reduction of MMP-2 and MMP-9 expression with TSA treatment in post-MI hearts. A and B: Masson trichrome staining of infarcted mouse hearts 7 days post-MI (A) or with TSA treatment (B). Yellow boxes in A and B represent border zones that were immunostained with MMP-2 and MMP-9. C and D: immunohistochemical analysis of active MMP-2 (green) in vehicle control-treated (C) or TSA-treated (D) 7-day post-MI hearts. E and F: MMP-9 immunolocalization in the infarct region (green) of vehicle control-treated (E) or TSA-treated (F) 7-day post-MI hearts.

To confirm that the repression of MMP-2 and MMP-9 expression was due to changes in acetylation, we used a second class I/IIb HDAC inhibitor, SAHA, which is currently approved by the Federal Drug Administration (vorinostat and zolinza) for the treatment of cutaneous T cell lymphoma (36). SAHA (100 mg·kg⁻¹·day⁻¹) treatment in mice without MI had no detectable effect on MMP-2 or MMP-9 levels (Fig. 5A). As expected, SAHA treatment dramatically reduced post-MI MMP-2 and MMP-9 levels compared with MI + vehicle values (Fig. 5, B and C). Similar to treatment with TSA, LV end-diastolic volume was increased in both MI + SAHA and MI groups at 7 days post-MI. Importantly, LV dilation was significantly lower in MI + SAHA mice compared with MI mice (Fig. 5D). As expected, LV ejection fraction was decreased at 7 days after MI compared with baseline. Again, similar to TSA treatment, there was a significant preservation of ejection fraction in MI + SAHA mice (Fig. 5E) compared with MI + vehicle mice.

Fig. 5.

Suberanilohydroxamic acid (SAHA), a class I/IIb HDAC inhibitor, also represses MMP-2 and MMP-9 levels in the infarct zone of the 7-day post-MI heart. A: representative gelatin zymogram of the infarct zone from 7-day post-MI hearts treated with SAHA. Human lung cancer tissue and recombinant MMP-2 and MMP-9 were used as positive controls with only recombinant MMP-2 and MMP-9 shown (+ve control lane). The image was constructed from lanes spliced from a single zymogram. Active MMP-2 and MMP-9 were elevated in the MI group and significantly reduced with SAHA administration. B and C: gelatinographic active MMP-2 (B) and MMP-9 (C) levels expressed in arbitrary units. Active MMP-2 and MMP-9 levels were increased in the MI group (20- and 32-fold, respectively), and both were drastically reduced with SAHA administration. Selective class I and IIb HDAC inhibition (SAHA) preserved ventricular geometry and improves ventricular function in post-MI mice. D: echocardiographic determinations of LV end-diastolic volume were performed before MI induction (pre-MI) and at 7 days post-MI for the MI-only and MI group treated with SAHA. LV end-diastolic volume was increased in both MI groups compared with pre-MI values. E: pre-MI LV ejection fractions in mice randomized to the MI groups were similar to control values. Post-MI, LV ejection fraction was lower than pre-MI values in both MI groups, but the reduction was attenuated in the MI + SAHA group. #P < 0.05 vs. control or pre-MI; *P < 0.05 vs. MI + vehicle.

HDAC activity is increased in the post-MI ventricle.

Because HDAC inhibitors repress MMP-2 and MMP-9 levels in the post-MI ventricle, we performed experiments to determine which class of HDAC catalytic activity was increased. Lysates from LVs of non-MI animals and infarct zones of 7-day post-MI hearts were assayed for HDAC activity using class I/IIb tripeptide substrate and a tripeptide substrate containing a trifluoracetyl group that is readily hydrolyzed by the catalytically less active class IIa HDACs (4). Global class IIa HDAC activity was slightly increased and class I/IIb activity was nearly five times greater in the 7-day post-MI infarct zone compared with the non-MI ventricle (Fig. 6A).

Fig. 6.

HDAC activity and class I HDAC expression in the infarct zone of the 7-day post-MI heart. A: class I and II HDAC enzymatic activity was measured in 7-day post-MI mouse heart homogenates from the infarct region. Class I HDAC activity was increased at least fourfold in the MI heart compared with the control heart, whereas class II HDAC activity was not significantly altered. B: tissue homogenates from the 7-day post-MI heart were immunoblotted for HDAC1, HDAC2, HDAC3, and GAPDH. C: class I HDAC activity was quantified from 7-day post-MI mouse heart homogenates before and after pan-HDAC inhibitor (SAHA, 1 μM) and class I-specific HDAC inhibitor (PD-106, 5 μM) in vitro treatment. *P < 0.05 vs. non-MI control; #P < 0.05 vs. MI + vehicle.

We next addressed whether the difference in catalytic activity was due to differences in class I HDAC expression or regulation of activity. Western blots indicated that HDAC1, HDAC2, and HDAC3 expression were all upregulated in the infarct zone (Fig. 6B). Unfortunately, commercially available antibodies were not effective at detecting the mouse class IIb HDAC, HDAC6 (data not shown). We then used a class I-specific HDAC inhibitor to determine if HDAC1, HDAC2, and HDAC3 were the major contributors to the increased activity in the lysate from the infarct zone and that the activity reflected the expression pattern. PD-106 (Pimelic diphenylamide) is more specific for HDAC3 (at 500 nM) but also inhibits HDAC1 and HDAC2 (at >5 μmol/l) (55). As expected, SAHA inhibited the majority of the class I/IIb tripeptide substrate-generated activity (Fig. 6C). The addition of PD106 at concentrations (5 μmol/l) high enough to inhibit HDAC1, HDAC2, and HDAC3 (47) resulted in inhibition similar to what was seen with SAHA treatment.

Inhibition of HDAC1, HDAC2, and HDAC3 suppresses MMP-2 and MMP-9 upregulation and is sufficient to preserve function.

Because class I HDACs were responsible for the majority of the change in activity, we treated mice with PD-106 to determine whether class I HDACs (HDAC1, HDAC2, and HDAC3) are required for the upregulation of MMP-2 and MMP-9 in the post-MI ventricle. CD-1 mice were subjected to MI and treated with PD-106 at previously used concentrations (100 mg·kg⁻¹·day⁻¹) (47) or vehicle every other day, starting immediately post-MI, and hearts were harvested 7 days post-MI. Treatment with PD-106 dramatically repressed MMP-9 post-MI expression to levels similar to what we observed with TSA and SAHA treatment, indicating that class I HDACs mediated MMP-9 expression in the post-MI ventricle (Fig. 7, A and B). Interestingly, PD-106 treatment reduced the level of the active form of MMP-2 to a similar extent as seen with TSA and SAHA treatments (Fig. 7B). However, PD-106 did not inhibit pro-MMP-2 expression levels as dramatically as we observed for class I/IIb inhibitors TSA and SAHA (compare Figs. 3A and 5A with 7A).

Fig. 7.

Class I-selective HDAC inhibition also represses MMP-2 and MMP-9 levels in the infarct zone of the 7-day post-MI heart. A: representative gelatin zymogram of the infarct zone from 7-day post-MI hearts treated with the class 1 HDAC inhibitor PD-106. Human lung cancer tissue and recombinant MMP-2 and MMP-9 were used as positive controls with only recombinant MMP-2 and MMP-9 shown (+ve control lane). The image was constructed from lanes spliced from a single zymogram. Active MMP-2 and MMP-9 were elevated in the MI group and significantly reduced with PD-106 administration. B and C: gelatinographic active MMP-2 (B) and MMP-9 (C) levels were expressed in arbitrary units. Active MMP-2 and MMP-9 levels were increased in the MI group (20- and 32-fold, respectively), and both were drastically reduced with PD-106 administration. Class I-selective HDAC inhibition (PD-106) preserved ventricular geometry and improved ventricular function in post-MI mice. *P < 0.05 vs. non-MI control; #P < 0.05 vs. MI + vehicle. D: echocardiographic determinations of LV end-diastolic volume were performed before MI induction (pre-MI) and at 7 days post-MI for the MI-only group and MI group treated with PD-106. LV end-diastolic volume increased in both MI groups compared with pre-MI values. E: changes in end-diastolic volume from pre-MI values. PD-106 treatment attenuated the post-MI increase in LV end-diastolic volume compared with the MI-only group. F: pre-MI LV ejection fractions in mice randomized to the MI groups were similar to control values. Post-MI, LV ejection fraction was lower than pre-MI values in both MI groups, but the reduction was attenuated in the MI + PD-106 group. G: PD-106 inhibited the reduction in ejection fraction compared with the MI-only group. #P < 0.05 vs. control or pre-MI; *P < 0.05 vs. MI only. #P < 0.05 vs. control; *P < 0.05 vs. MI + vehicle.

Finally, we determined if treatment with the class I HDAC inhibitor PD-106 was sufficient to preserve cardiac remodeling and function. LV end-diastolic volume was increased in both MI + PD-106 and MI groups at 7 days post-MI. Importantly, LV dilation was significantly lower in MI + PD-106 mice compared with MI mice (Fig. 7, D and E). As expected, LV ejection fraction was decreased at 7 days after MI compared with baseline. There was a significant preservation of ejection fraction in MI + PD-106 mice (Fig. 7, F and G) compared with MI + vehicle mice. Importantly, both end-diastolic volume and ventricular function were at least as well preserved with PD-106 as was observed with TSA (compare Figs. 1, A and B, with 7, D and E, and compare Fig. 7, D with F and G).

DISCUSSION

One of the therapeutic goals for post-MI patients is to minimize the degree of infarct expansion and adverse remodeling while optimizing cardiac repair. Many studies have carefully documented the series of changes that occur within the myocardial ECM after MI (10, 11, 25). Prolonged ischemia results in initial cellular necrosis, which is followed by deposition of granulation tissue and, at the end stage, increased fibrosis within the ischemic region. Each of these stages in ECM remodeling is associated with time-dependent expression and activity of specific MMPs and TIMPs (35). MMP-9 directly contributes to post-MI LV remodeling (11). Importantly, post-MI LV dilation is ablated in MMP-9-deficient mice (14), and adverse LV remodeling is attenuated through pharmacological inhibition of MMPs (50). Furthermore, ablation of MMP-2 expression improves post-MI survival by better preservation of LV function by delaying and decreasing the extent of post-MI remodeling (14, 21, 37). While these transgenic studies have demonstrated that genetic ablation of either MMP-2 or MMP-9 attenuated the degree of adverse LV remodeling and dysfunction post-MI, these findings are not translatable in terms of a therapeutic target. The present study directly addressed this issue in that one of the results of HDAC inhibition is reduced expression of MMP-2 and MMP-9, which results in reduced adverse post-MI remodeling. Our data demonstrate that treatment of mice with inhibitors of class I/IIb HDACs results in repression of MMP-2 and MMP-9 expression and reduces LV dilation and demonstrates preservation of ejection fraction in the post-MI heart. Moreover, we demonstrated here, for the first time, the efficacy of using the class I-selective HDAC inhibitor PD-106 in post-MI remodeling. PD-106 treatment suppresses MMP-9 expression, partially suppresses MMP-2 expression, and results in at least similar preservation of post-MI function and attenuation of adverse remodeling to that observed with the class I/II HDAC inhibitor TSA and the class I/IIb inhibitor SAHA.

Although their role is complex and not fully understood, HDACs clearly play an intricate part in the molecular regulation of pathological cardiac remodeling. HDACs are grouped into four classes based on size and structure. Class I HDACs (HDAC1, HDAC2, HDAC3, and HDAC8) are ubiquitously expressed. Class II HDACs are expressed in a tissue-specific manner and are subgrouped as class IIa (HDAC4, HDAC5, HDAC7, and HDAC9) and class IIb (HDAC6 and HDAC10). Class III comprises NAD⁺-dependent deacetylases (SIRT1–SIRT7). HDAC 11 is the sole member of class IV (19). Several in vitro and in vivo studies have demonstrated that pan-HDAC inhibitors are efficacious in halting and even reversing pathological cardiac hypertrophy. TSA blocks agonist-induced hypertrophy in cultured rat neonatal (2) and adult (8) cardiomyocytes. HDAC inhibitors have been shown to block pathological cardiac hypertrophy induced by continuous infusion of isoproterenol (29) or ANG II and also importantly in pressure overload created by transverse aortic constriction (27, 28). TSA treatment can even reverse established cardiac hypertrophy in mice subjected to transverse aortic constriction (27). Long-term treatment with TSA or another pan-HDAC inhibitor, scriptaid, preserved both cardiomyocyte cross-sectional area and significantly improved ventricular function, which was maintained with treatment for up to 9 wk (28). In addition, TSA treatment inhibited cardiac myocyte autophagy in a pressure overload model of cardiac hypertrophy (5). Mice treated with valproic acid, a weak HDAC inhibitor with many off-target effects (20), showed significantly attenuated cardiomyocyte hypertrophy and collagen formation in the remote and border zones and improved LV fractional shortening 4 wk post-MI. HDAC inhibitors have also been shown to be protective in models of myocardial ischemia-reperfusion. More than one study has demonstrated a significant decrease in infarct size and myocardial cell death with class I/IIb HDAC inhibitor treatment (18, 60). We recently demonstrated that class I HDAC inhibition conferred protective benefits to the ischemic reperfused heart and that inhibition of HDAC6, a class IIb HDAC, provided no beneficial effects (3).

Interestingly, a significant number of the genes upregulated in heart disease are repressed by treatment with HDAC inhibitors. In addition to MMP-2 and MMP-9, these include β-myosin heavy chain, atrial natriuretic factor, brain natriuretic peptide, and Na⁺/Ca²⁺ exchanger (NCX)1 (8, 39). Key transcription factors of Ncx1 and brain natriuretic peptide, NK2 homeobox 5 (Nkx2.5) and YY1, respectively, interact with HDACs, and these interactions have been implicated to play a critical role in the regulation of these genes during both normal and pathological conditions (8, 51). Deacetylation of Nkx2.5 results in the release of HDAC1, HDAC2, and HDAC5, recruitment of HAT p300, and upregulation on Ncx1 in cardiac hypertrophy (8). One possibility is that HDACs may regulate MMP-2 and MMP-9 expression by directly deacetylating one of the transcription factors binding to MMP-2 or MMP-9 promoters. NF-κB, activator protein-1, Sp1, and Ets, all play an important role in MMP-9 upregulation in response to ischemia, and they are all acetylated (33, 56, 61). In this respect, it is interesting that NF-кB acetylation is important for HDAC inhibitor-induced cardioprotection in an in vitro mouse model of ischemia-reperfusion (59). On the other hand, HDACs could affect MMP-2 and MMP-9 expression indirectly, possibly by regulating the levels of inflammatory cytokines expressed and released in the infarct and border zones. We have shown that TSA treatment represses the induction of retinal TNF-α levels in rats subjected to retinal ischemia (13). Another study (24) has shown that long-term treatment with SAHA reduced circulating levels of multiple inflammatory cytokines, including TNF-α, IL-1β, and IL-6, in the rat DOCA-salt model of hypertensive cardiomyopathy. These decreases in inflammatory cytokines are correlated with reduced cardiac hypertrophy and fibrosis. Importantly, post-MI inflammatory cells are potent regulators of the wound healing environment, and neutrophils and macrophages are major sources of MMP-9. Neutrophils and macrophages are recruited in the first few minutes post-MI and release their stored MMP-9. Much of the increase in MMP-9 seen in the days after the infarct is the result of newly synthesized MMP-9. Class I HDACs could play a direct role in regulating MMP-9 transcription or possibly influence macrophage polarization from proinflammatory M1 to anti-inflammatory M2 macrophages, which express much lower levels of MMP-9. Clearly, the inhibition of HDAC activity affects multiple pathways, which collectively mediate the protective responses, and resolution of these important questions awaits future investigation.

Several recent studies have begun to address the possible changes in HDAC expression and catalytic activity in heart disease. Global HDAC activity is increased in hypertrophic spontaneously hypertensive rat hearts (6), and sequential immunoprecipitation-HDAC activity assays have demonstrated an increase in HDAC2 but not HDAC1 activity in pressure overload hypertrophy (26). McKinsey and colleagues (31) demonstrated that there is a consistent increase in class IIb HDAC6 expression and activity in the right ventricle of rat models of chronic hypertension. Granger et al. (18) also showed an increase in HDAC activity in a mouse model of ischemia-reperfusion. Ours is the first work to examine HDAC expression levels and catalytic activity after MI. We demonstrated a dramatic increase (fivefold) in class I HDAC activity and a slight increase in class IIa HDAC activity in the infarct zone. It is important to remember that the difference in HDAC expression pattern in the normal myocardium versus the post-MI infarct zone may be due to completely different cell populations. Post-MI wound repair involves several complex phases, including the recruitment of inflammatory cells and fibroblasts to the ischemic zone as well as the proliferation and maturation of fibroblasts, myofibroblasts, and macrophages, which replace the resident myocytes, fibroblasts, and endothelial cells lost after ischemia. This may account in part for the reason that HDAC catalytic activities in the infarct zone were relatively higher than what has been previously reported in rodent models of hypertensive heart disease (31). Other reasons may include that the rat hypoxia model results in significantly less ventricular remodeling compared with what occurs post-MI. Certainly, one cannot assume that changes in HDAC expression and global activity are the only clear requisite to demonstrating how a particular HDAC participates in cardiac pathology. Class I and II HDACs are found in unique complexes. The activity, location, and target of each of these unique HDAC complexes may be independently regulated. Further work is needed to examine these subtle but more physiologically relevant changes in specific HDAC complex activity.

Although this is the first work showing that HDAC1, HDAC2, and/or HDAC3 alone are required for MMP-2 and MMP-9 induction and contribute to adverse ECM remodeling in the post-MI ventricle, a recent study (54) has demonstrated that TSA treatment reduced MMP-2 and MMP-9 levels in esophageal squamous cell carcinoma. A second study (49) has shown that HDAC10 suppresses cervical cancer metastasis through inhibition of MMP-2 and MMP-9. These studies and ours highlight how the functions of HDACs are diverse and often class or isoform dependent. Inhibition of class I HDACs had the same effect on MMP-9 expression as pan-HDAC inhibitors; therefore, class I HDACs regulate MMP-9 expression in the post-MI ventricle. On the other hand, the class I inhibitor had less of an effect in repressing MMP-2 expression than pan-HDAC inhibitors. Therefore, our data support that although class I HDACs play a role in MMP-2 expression, other HDACs (possibly HDAC6) are involved as well. Interestingly, treatment with the class I inhibitor PD-106 reduced the level of the active form of MMP-2 but did not inhibit pro-MMP-2 expression as much as TSA or SAHA. Pro-MMP-2 forms a tight complex with TIMP-2, which is essential for MMP-2 activation by membrane type 1 (MT1-)MMP (44). Therefore, class I HDACs may play a role in MMP-2 maturation via affecting the MMP-2-TIMP-2 and MT1-MMP complex or regulating the expression of MT1-MMP or TIMP-2.

Our work complements that of recent studies highlighting the role of class I HDACs in regulating pathological cardiac remodeling and growth in response to pressure overload. We have previously demonstrated that HDAC1 and HDAC2 are recruited by the transcription factor Nkx2.5 to Ncx1. Deacetylation of Nkx2.5 results in the release of HDAC1/2, recruitment of HAT p300, and upregulation on Ncx1 in cardiac hypertrophy (8). McKinsey and colleagues (7) demonstrated that inhibition of class I HDACs suppressed hypoxia-induced cardiopulmonary remodeling and blunted hypoxia-induced expression of IL-1b, IL-2, and cytokine-induced neutrophil chemoattractant-2 protein in rats exposed to hypobaric hypoxia for 3 wk.

In summary, the present study provides an additional proof of concept for a potential therapeutic benefit in the acute use of class I-specific HDAC inhibitors: to prevent adverse remodeling and preserve cardiac function immediately after MI. Further work is necessary to elucidate which class I HDAC (HDAC1, HDAC2, and/or HDAC3) mediates the changes in MMP-2 and MMP-9 expression and the mechanism of regulation. Future studies also need to address whether the attenuated LV dilation and improved LV function seen in this study persist in the weeks and months after HDAC inhibitor treatment. Finally, we need to determine if isoform-specific HDAC inhibitors provide synergistic therapeutic benefit in combination with currently used therapeutic agents. These findings characterize novel functions for class I HDACs in post-MI remodeling and show that class I-selective HDAC inhibitors hold clear therapeutic promise in ischemic heart disease.

GRANTS

This work was supported by an American Heart Association (AHA) grant-in-aid, National Institutes of Health (NIH) Grants R01-HL-095696 and UL1-TR-000062, a merit award from the Veterans Affairs Health Administration, and a gift from Mr. and Mrs. Robert Tarr (to D. R. Menick). C. B. Kern was supported by NIH Grants R01-HL-121382 and 5-P20-RR-016434 and an AHA Scientist Development Grant. F. G. Spinale was supported by NIH Grants HL-057952 and HL-059165 and a merit award from the Veterans Affairs Health Administration. D. Kimbrough was supported by NIH Postdoctoral Fellowship Grant T32-HL-07260 and an AHA Predoctoral Fellowship.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: S.K.M. and D.R.M. conception and design of research; S.K.M., C.B.K., D.K., B.A., H.K., W.T.R., R.K.P., and J.C.C. performed experiments; S.K.M., C.B.K., H.K., F.G.S., R.M., and D.R.M. analyzed data; S.K.M., D.K., R.M., and D.R.M. interpreted results of experiments; S.K.M., C.B.K., D.K., and R.M. prepared figures; S.K.M., C.B.K., F.G.S., R.M., and D.R.M. edited and revised manuscript; S.K.M. and D.R.M. approved final version of manuscript; D.R.M. drafted manuscript.