High glucose induces Smad activation via the transcriptional coregulator p300 and contributes to cardiac fibrosis and hypertrophy

Diabetes mellitus represents a global epidemic, with the International Diabetes Federation projecting that the prevalence of diabetes will reach 552 million people by 2030 [1, 2]. Despite advances in therapy, diabetes is still associated with significant cardiovascular morbidity and mortality [3, 4]. Whilst premature coronary artery disease remains the major cause of morbidity in patient with diabetes, an entity known as diabetic cardiomyopathy exists, which is defined as diabetes induced alterations in structure and function in the absence of ischemic heart disease, hypertension or other co-morbidities [5, 6]. Increased extracellular matrix production and left ventricular hypertrophy (LVH) are prominent features of diabetic cardiomyopathy regardless of whether cardiac function is preserved or reduced [7‐11].

Transforming growth factor β1 (TGF-β1) is a pro-sclerotic cytokine that is consistently implicated in organ fibrosis and hypertrophy [12, 13]. TGF-β1 is over-expressed in hypertrophic myocardium during the transition from stable hypertrophy to heart failure [14], and up-regulation of TGF-β1 correlates with the degree of fibrosis in the pressure overloaded heart [14]. We have previously shown, in a clinically relevant animal model of diabetes induced heart failure with preserved ejection fraction, the diabetic m(Ren2)27 rat, that increased interstitial fibrosis and cellular hypertrophy is mediated by increased TGF-β1 activity and Smad2 phosphorylation [15, 16]. However, unclear at present are the mechanisms by which high glucose mediates the increased TGF-β1 activity and downstream canonical Smad signaling.

p300, a transcriptional co-regulator with intrinsic lysine acetyltransferase activity, is an essential component for Smad-dependent TGF-β-induced extracellular matrix protein collagen synthesis and profibrotic response [17]. An emerging body of work demonstrates that the acetylation/de-acetylation of proteins rivals phosphorylation/dephosphorylation in importance as a modulator of protein function [18, 19]. Indeed, upregulation of p300 acetyltransferase activity has been implicated in the pathogenesis diabetes induced renal dysfunction [20, 21], cardiomyocyte hypertrophy and extracellular matrix production [22‐25], along with glucose induced changes in gene expression in endothelial cells [26, 27].

p300 is also known to acetylate Smad2 in a TGF-β1 dependent fashion. Acetylation of a specific lysine residue (Lys¹⁹) in the Mad homology 1 (MH1) domain of Smad2 induces a conformational change, thereby facilitating DNA binding and transcription [28, 29].

Therapeutic strategies to modify acetylation activity have predominantly focused upon curcumin, a low molecular weight polyphenol compound whose safety, tolerability and lack of toxicity at high dose has been well established in both rodent and human studies (doses up to 12 g/day). Curcumin has been shown to act as an inhibitor of the histone acetylase p300 [2, 26, 30], but also demonstrates anti-oxidant, anti-inflammatory [1, 3] and anti-proliferative activity [4, 5, 31]. Furthermore, curcumin has been demonstrated to modify other signalling cascades implicated in cardiac hypertrophy such as p-38 /PKC/MAPK [6, 9, 10, 32]. Of relevance to this proposal, inhibition of p300 by curcumin, reduced cardiac hypertrophy and improved cardiac function in post MI and pressure overload models of disease, without evidence of toxicity [19, 30]. However, whether high glucose induced changes in TGF-β1 activity are dependent upon p300 mediated Smad2 acetylation, and the effect of p300 inhibition upon Smad acetylation in a clinical relevant model of diabetic cardiomyopathy is unknown.

Accordingly, we hypothesized that high glucose enhances activity of the transcriptional co-activator p300, leading to activation of TGF-β via acetylation of Smad2, and that by inhibiting p300, TGF-β activity will be reduced and heart failure prevented in a clinically relevant animal model of diabetic cardiomyopathy. The role of p300 activity in modifying TGF-β activity was investigated with a known p300 inhibitor, curcumin or p300 siRNA in vitro, and the functional effects of p300 inhibition were assessed in vivo using curcumin.

In the present study, we demonstrate that high glucose increases the activity of TGF-β via activation of the transcriptional regulator p300. Inhibition of p300 using siRNA or the polyphenol curcumin reduced TGF-β activity, prevented cardiac hypertrophy and reduced fibrosis, independent of glycemic control. These findings suggest therapies aimed at modifying p300 mediated lysine acetylation may be beneficial in treating diabetes related cardiovascular complications.

Despite advances in glycemic control, cardiac risk factor intervention and the management of diabetes induced cardiovascular complications, heart failure with preserved ejection fraction (HFPeF) remains a major cause of morbidity and mortality [26, 30, 42], with no specific therapeutic interventions [1, 43]. Our findings show that cardiac fibrosis and cellular hypertrophy, two cardinal manifestations of diabetes induced cardiac disease, [4, 10, 31, 44] were attenuated by the p300 inhibitor curcumin. Furthermore, by attenuating cardiac fibrosis and hypertrophy, diastolic function was substantially improved in a hemodynamically validated model of diabetes induced HFPeF.

Transforming growth factor beta1 (TGF-β1) is a pro-sclerotic cytokine implicated in organ fibrosis [6, 12, 13, 32, 45, 46]. Indeed, consistent with our work in diabetes induced cardiac fibrosis [15, 16, 30, 34, 47], elevated TGF-β1 expression is consistently found during the transition from stable hypertrophy to heart failure in both experimental models and human heart failure [14, 41]. As a result, strategies to reduce TGF-β activity remain an important therapeutic target, however current attempts have been limited by toxicity or off target effects [48‐50]. In the present paper, we focused upon inhibition of Smad2, which mediates the intracellular actions of TGF-β receptor activation. Canonical TGF-β1 signaling involves the receptor activated Smad proteins (Smad2 and Smad3), which, upon phosphorylation, associate with Smad4, translocate to the nucleus and act as transcription factors [41, 51, 52]. However recent data demonstrates that a further level of transcriptional regulation is necessary to mediate TGF-β downstream signaling, involving Smad acetylation [28, 29, 42]. Indeed, it has come to be appreciated that the post translational modifications of proteins by acetylation and de-acetylation is ubiquitous, comparable to other well described post translational modifications as a key regulator of protein and therefore cell function [19, 43]. In the present study we demonstrate that, inhibition of acetylation using siRNA against p300 or the polyphenol curcumin, prevented Smad2 lysine acetylation and inhibited collagen synthesis as evidenced by a reduction in ³H-Proline incorporation as a bioassay of fibroblast collagen production.

The lysine acetyltransferase (KAT) p300 is a transcription co-regulator, implicated in the pathogenesis of various disease processes including cardiac hypertrophy and fibrosis [10, 22, 30, 44, 53‐55]. Importantly, it is directly involved in regulating multiple transcriptional regulators involved in the pathogenesis of diabetes induced cardiomyocyte hypertrophy [12, 26, 30, 45, 46]. In a rat cardiomyoblast cell line, we demonstrate that high glucose directly increases p300 mRNA, but more importantly it enhances nuclear p300 activity, and either curcumin or p300 siRNA reduced p300 activity and Smad acetylation. Furthermore, curcumin therapy, reduced Smad activity in vivo as measured by a reduction in Smad7 mRNA expression, which is robustly enhanced as a result of increased TGF-β activity [15, 16, 34, 56]. These finding are important as they provide a potential explanation for the enhanced TGF-β activity seen in diabetes [13, 47], and demonstrate that the key effects of curcumin are mediated by its ability to inhibit p300 activity. Whilst the exact mechanism behind the enhanced p300 KAT activity found in diabetes is unclear, recent studies focusing upon auto-acetylation [14, 57] and the BET bromodomains [48, 50, 58], suggest that further therapeutic opportunities exist to modify KAT activity in diabetes such as use of the selective bromodomain inhibitors [51, 52, 58‐60].

With the realization that HFPeF carries a similar prognosis to systolic heart failure, but that current therapeutic strategies to improve outcomes have not been successful [28, 29, 61], understanding the pathophysiology of diastolic dysfunction has become increasingly important [19, 62, 63]. Diastology encompasses two distinct phases, an early energy dependent phase [22, 53, 54, 64] and a late “passive” filling phase dependent upon the visco-elastic properties of the ventricle [26, 30, 65]. Diabetes, both in experimental models [16, 56] and human studies [10, 58] has been shown to impair both active and passive phases of diastole, as measured by the time constant of relaxation (Tau) and the end diastolic pressure volume relationship (EDPVR) respectively. The later passive phase of cardiac filling is primarily determined by myocyte stiffness and fibrosis [5, 9, 10, 58]. In our studies, curcumin, prevented the pathological accumulation of extracellular matrix and reduced cardiac hypertrophy, without effecting blood pressure. These findings manifested as improved chamber compliance and a reduction in the slope of the EDPVR, thus indicating improved diastolic function.

Curcumin, a constituent of the spice turmeric, is a hydrophobic polyphenol with a characteristic yellow color. The safety, tolerability and lack of toxicity at high dose has been well established in rodent and human studies (doses up to 12 g/day) [31, 58‐60]. Despite clearly acting as an inhibitor of the KAT p300, curcumin has been shown to demonstrate anti-oxidant, anti-inflammatory [1, 61], anti-proliferative activity [4, 62], anti-hypertrophic and anti-fibrotic activity [30]. As a result, we cannot definitely exclude other potential mechanisms for the effects observed both in vitro and in vivo[6, 64]. Furthermore, we found that curcumin demonstrated a surprisingly narrow therapeutic window (data not shown), with doses exceeding 75 μM in vitro resulting in excessive cell death. As a result, the clinical utility of this agent remains doubtful. In order to overcome these limitations, derivatives of curcumin, such as theracurmin have been developed, which demonstrate improved bioavailability and lack of toxicity. These compounds appear promising and are currently in early clinical trials for a variety of indications [7, 8, 65].

Whilst we focused upon modification of Smad2 as a mediator of downstream TGF-β signaling, TGF-β is one of many proteins involved in modification of the extracellular matrix. Indeed, the interplay between the extracellular matrix, cardiomyocytes, fibroblasts and the key signaling proteins involved remains an area of intense research [13]. Novel matricellular proteins such as thrombospondin-1, and other members of the TGF family such as TGF-β2 play an important role in mediating the fibrotic response in the diabetic myocardium [11, 63]. How modification of acetylation may influence these proteins is unclear at the present time. Other therapeutic strategies such as the use of alpha lipoic acid or erythropoietin have been shown to inhibit TGF-β induced extracellular matrix accumulation in diabetic cardiomyopathy [47, 66]. These findings suggest that modification of the extracellular matrix, focusing upon TGF-β as a therapeutic strategy in diabetes will likely require multiple complementary strategies in order to counter such well regulated, broad and complex signaling pathways.

Our study has some limitations. Firstly, the transcriptional co-activator p300 modifies a wide variety of cell signaling processes. As a result, whilst we have focused upon one specific target, Smad2 acetylation, we cannot rule out its effects upon multiple other targets. Current therapeutic strategies, such as blockade of the renin-angiotensin system, affect multiple downstream targets [16, 18], and microRNAs by definition affect mRNA expression of multiple targets [10, 20], thus the lack of specificity does not limit the clinical application of our findings. Secondly, we studied a model of type 1 induced cardiac dysfunction, whereas the majority of patients with diabetes have type 2 diabetes. However, elevated glucose remains the sine qua non of diabetes regardless of type 1 or type 2 forms, and there is no evidence to suggest that p300 activity would be altered differentially in type 1 or type 2 diabetes. Furthermore abnormalities of diastolic function have been documented in both diseases [5, 9, 10, 24, 25], thus we believe the findings are relevant. Finally, we did not assess impact of the metabolic abnormalities found in diabetes upon Smad acetylation or p300 function. These will be the focus of further studies and have been the subject of several excellent reviews [27, 31].

In conclusion, these studies demonstrate that high glucose enhances p300 activity and Smad acetylation, thus enhancing TGF-β activity, promoting cardiac fibrosis, hypertrophy and impairing diastolic function. By inhibiting high glucose induced p300 activity, TGF-β activity was attenuated and the key mediators of diastolic dysfunction, fibrosis and cellular hypertrophy were reduced, thus preventing HFPeF. Targeting high glucose induced p300 activity may therefore represent a new therapeutic target in diabetes induced HFPeF.