The Synthetic Curcumin Analogue GO-Y030 Effectively Suppresses the Development of Pressure Overload-induced Heart Failure in Mice

Curcumin is a naturally occurring p300-histone acetyltransferase (p300-HAT) inhibitor that suppresses cardiomyocyte hypertrophy and the development of heart failure in experimental animal models. To enhance the therapeutic potential of curcumin against heart failure, we produced a series of synthetic curcumin analogues and investigated their inhibitory activity against p300-HAT. The compound with the strongest activity was further evaluated to determine its effects on cardiomyocyte hypertrophy and pressure overload-induced heart failure in mice. We synthesised five synthetic curcumin analogues and found that a compound we have named GO-Y030 most strongly inhibited p300-HAT activity. Furthermore, 1 μM GO-Y030, in a manner equivalent to 10 µM curcumin, suppressed phenylephrine-induced hypertrophic responses in cultured cardiomyocytes. In mice undergoing transverse aortic constriction surgery, administration of GO-Y030 at a mere 1% of an equivalently-effective dose of curcumin significantly attenuated cardiac hypertrophy and systolic dysfunction. In addition, this low dose of GO-Y030 almost completely blocked histone H3K9 acetylation and eliminated left ventricular fibrosis. A low dose of the synthetic curcumin analogue GO-Y030 effectively inhibits p300-HAT activity and markedly suppresses the development of heart failure in mice.

All types of heart disease finally lead to the development of heart failure, which is a leading cause of death worldwide. The situation has been described as an epidemic and persists despite the use of established treatment options for heart failure 1,2 . Therefore, innovative treatment strategies are urgently required. If conditions such as hypertension or myocardial infarction continue to stress the heart, it eventually leads to the failure of systolic function. Functional heart failure is closely associated with pathological cardiomyocyte hypertrophy 3,4 . Thus, controlling cardiomyocyte hypertrophy is a major target of heart failure treatment.
It is now well understood that the histone acetyltransferase (HAT) activity of the transcriptional coactivator p300 in the nucleus of cardiac myocytes plays a key role in pathological myocyte hypertrophy and heart failure 5,6 . When the heart is subjected to hypertrophic stress, neurohumoral factors regulating the renin-angiotensin system and the sympathetic nervous system are activated. These factors bind with receptors on the surface of cardiomyocytes and then activate various signalling pathways, eventually reaching the nucleus and activating p300. p300 acetylates hypertrophy-responsive transcription factors such as MEF2 and GATA4, as well as histones, thus causing pathological myocyte hypertrophy by facilitating the transcription of hypertrophy-associated genes 5,6 . In GO-Y030 at a dose 1/10th that of curcumin significantly suppressed PE-induced hypertrophic responses in cardiomyocytes. To investigate whether GO-Y030 suppresses PE-induced cardiomyocyte hypertrophy by inhibiting p300-HAT activity, primary cultured cardiomyocytes were used. The cells were stimulated with PE in the presence or absence of curcumin or GO-Y030 for 48 h. Histone fractions isolated by acid extraction from these cells were subjected to western blotting to assess acetylated histone H3K9 and total histone H3 levels. The results showed that the acetylation level of histone H3K9 was increased by PE compared with the control. The increase was significantly suppressed by low doses of GO-Y030 (0.3, 1 μM), whereas higher doses of curcumin (3, 10 μM) were required to have the same effect (Fig. 3a,b). Next, to investigate the effect of GO-Y030 on the transcriptional activity of hypertrophic response genes, a quantified PCR analysis of ANF and BNP mRNA levels was performed. The analysis revealed that 1 µM GO-Y030 and 10 µM curcumin significantly suppressed the PE-induced increase in the mRNA levels of ANF and BNP (Fig. 3c,d). In addition, immunofluorescence staining with anti-MHC antibodies revealed that both 0.3 and 1 μM GO-Y030 significantly suppressed the PE-induced increase in the surface area of the cells, whereas 10 µM curcumin was required to have the same effect (Fig. 3e,f). These results suggest that GO-Y030 suppressed PE-induced hypertrophic responses in cardiomyocytes more strongly than curcumin.
Immunoprecipitation followed by western blotting was then performed to investigate whether GO-Y030 suppresses cardiomyocyte hypertrophy by inhibiting the p300/GATA4 pathway, as curcumin does. Nuclear protein was extracted from the cardiomyocytes, and immunoprecipitation using anti-p300 antibodies and subsequent western blotting were performed. The results showed that the PE-induced increase in the interaction between p300 and GATA4 was suppressed by 1 µM GO-Y030 to the same extent as by 10 µM curcumin (Fig. 3g).  GO-Y030 at a dose 1/10th that of curcumin significantly suppressed PE-induced hypertrophic responses in cardiomyocytes. Primary cultured cardiomyocytes were treated with 3 or 10 μM curcumin, or with 0.3 or 1 μM GO-Y030 and were stimulated with 30 μM PE. (a) Histone fractions isolated from these cells were subjected to western blotting using anti-acetyl-histone H3 (Lys9) antibodies and anti-histone H3 antibodies. Full-length blots are presented in Supplementary Fig. S6. (b) The levels of acetylated histone H3K9 and total histone H3 were quantified. The data are presented as the mean ± SEM of three individual experiments. (c,d) Total RNA was extracted from the cells, and quantitative PCR was performed for ANF (c), BNP (d), and 18S. The data are presented as the mean ± SEM of three individual experiments. (e) Immunofluorescence staining was performed using anti-MHC antibodies and Alexa Fluor 555-conjugated anti-mouse IgG. Scale bar: 20 μm. (f) The surface area of these cells was measured using ImageJ software. All data are presented as the mean ± SEM of three individual experiments. (g) Nuclear extracts prepared from primary cultured cardiomyocytes were subjected to western blotting with anti-p300 antibodies, anti-GATA4 antibodies, and GO-Y030 at a dose 1/100th that of curcumin significantly suppressed TAC-induced cardiac hypertrophy and systolic dysfunction in vivo. Because GO-Y030 strongly suppressed cardiomyocyte hypertrophy, it is likely that it also prevents pathological cardiac hypertrophy and the development of heart failure. To test this hypothesis, C57BL/6 J male mice were subjected to TAC (or to a sham operation as a control). The TAC mice were then randomly assigned to daily oral treatment with the vehicle, 1 or 50 mg/kg of curcumin, or 0.1 or 0.5 mg/kg of GO-Y030. Cardiac function was assessed by echocardiography 6 weeks after the operation. The results indicated that LVPWT, IVSd, LVIDs, and LVMI, which are parameters of cardiac hypertrophy, were increased by TAC, and that 50 mg/kg of curcumin significantly suppressed these changes. Surprisingly, 0.5 mg/kg of GO-Y030 at a dose 1/100th that of curcumin also suppressed the changes. GO-Y030 at a dose 1/100th that of curcumin also prevented the TAC-induced decrease in FS, which is a parameter of cardiac function, to the same extent as 50 mg/kg of curcumin ( Fig. 4 and Table 1).
GO-Y030 at a dose 1/100th that of curcumin significantly suppressed TAC-induced hypertrophic responses in mouse heart. After the echocardiography assessment, the hearts were isolated ( Fig. 5a) and investigated for the effect of GO-Y030 on hypertrophic responses. The heart weight and body weight of all the mice were measured, and the ratio of heart weight to body weight (HW/BW) was calculated. The results showed that 0.5 mg/kg of GO-Y030 prevented a TAC-induced increase in the HW/BW ratio to the same extent as 50 mg/kg of curcumin (Fig. 5b). Next, a histological analysis was performed by staining the heart tissues with WGA, followed by measuring the cross-sectional areas. The results showed that the two parameters were significantly increased by TAC operation and were significantly suppressed by 50 mg/kg of curcumin and 0.5 mg/kg of GO-Y030 (Fig. 5c,d). Finally, the mRNA levels of ANF and BNP in the LV were investigated by quantitative PCR analysis. As shown in Fig. 5e,f, the mRNA levels of these hypertrophic genes were significantly increased by TAC. Curcumin of 50 mg/kg significantly suppressed these increases; however, 0.5 mg/kg of GO-Y030 also suppressed them to the same extent as the 50 mg/kg of curcumin.

GO-Y030 at a dose 1/100th that of curcumin significantly suppressed TAC-induced cardiac fibrosis.
To confirm that the changes in cardiac fibrosis were indeed caused by the TAC operation and then ameliorated by the curcumin and GO-Y030 treatments, heart tissues were stained with MT, and areas of perivascular fibrogenesis were measured. The results showed that fibrotic areas were significantly increased by TAC operation and that this change was suppressed by 50 mg/kg of curcumin and 0.5 mg/kg of GO-Y030 (Fig. 6a,b). Next, the mRNA levels of genes associated with fibrosis in the mouse hearts were measured. Similar to the results of the histological analysis, collagen 1a1, collagen 3a1, and fibronectin levels were increased after the TAC operation, and 50 mg/kg of curcumin and 0.5 mg/kg of GO-Y030 significantly suppressed these increases ( Fig. 6c-e).

GO-Y030 at a dose 1/100th that of curcumin significantly suppressed TAC-induced increases in histone acetylation.
To determine whether a low dose of GO-Y030 can suppress TAC-induced histone acetylation, western blotting was performed using histone fractions from the hearts. The acetylation of histone H3K9 was also enhanced by TAC; however, 0.5 mg/kg of GO-Y030 suppressed the increase in this acetylation to the same extent as 50 mg/kg of curcumin ( Fig. 7a,b). This suggests that the improvement in heart failure induced by GO-Y030 is attributable to its strong inhibition of p300-HAT activity.
GO-Y030 toxicity was not observed in the TAC mice at a concentration of 0.5 mg/kg. Finally, the toxicity of GO-Y030 in mice was investigated. Blood was collected from the mice 6 weeks after the TAC operation. Markers of liver function, including alanine transferase, aspartate transaminase, and total bilirubin, and markers of renal function, including creatinine and blood urea nitrogen, in the blood were tested. There were no differences in any parameters among the three groups: TAC + vehicle, curcumin, or GO-Y030 ( Supplementary  Fig. S1a-e). Next, the weights of the liver and kidney were corrected for body weight. No significant differences in liver weight to body weight or kidney weight to body weight were found among the four groups ( Supplementary  Fig. S1f,g). In addition, the liver and kidney tissues were stained with HE and periodic acid-Schiff staining, respectively, and were observed by microscopy. After the administration of curcumin or GO-Y030, no disruption of hepatic lobules was observed, and there were no aberrations in the glomerulus or mesangial matrix ( Supplementary Fig. S2a,b).

Discussion
This study demonstrates that GO-Y030 inhibits p300-HAT activity and PE-induced hypertrophy in cultured cardiomyocytes at 1/10th the dose of curcumin. More importantly, the oral administration of GO-Y030 ameliorated TAC-induced cardiac hypertrophy and heart failure at 1/100th the dose of curcumin. The high effectiveness of GO-Y030 at a low dose may overcome the problem of the very large dose of curcumin that would be needed in anti-β-actin antibodies. Full-length blots are presented in Supplementary Fig. S7. (h) The nuclear extracts were immunoprecipitated with goat anti-GATA4 polyclonal antibodies and subjected to western blotting with anti-p300 antibodies, anti-acetyl-lysine antibodies, and anti-GATA4 antibodies. The original images are presented in Supplementary Fig. S8. (i) The levels of acetylated GATA4 were quantified. The data are presented as the mean ± SEM of three individual experiments. (2020) 10:7172 | https://doi.org/10.1038/s41598-020-64207-w www.nature.com/scientificreports www.nature.com/scientificreports/ clinical settings due to its low bioavailability. This may enable the development of an effective new treatment for heart failure.
The results of the structure-activity relationship study determined by the in vitro p300-HAT assay revealed that GO-Y030 inhibited p300-HAT activity most strongly among all the analogues and that it had p300-HAT inhibitory activity that was approximately nine times higher than that of curcumin. There are two key structural differences between curcumin and GO-Y030 that potentially explain this difference in their degree of p300-HAT inhibitory activity: the differences in their ketone structures and the differences in the structure of their functional groups. Evidence from this and other studies suggests that the decisive difference is in their functional groups. In terms of ketone structure, both curcumin and GO-Y030 are α, β-unsaturated ketones. Both have been reported to act as Michael acceptors; the α, β-unsaturated ketone structure of curcumin has been reported to bind covalently to p300, and it appears very likely that that of GO-Y030 does as well [16][17][18][19][20] . However, there is a key difference in the structures of the two ketones: curcumin is an α, β-unsaturated β-diketone, whereas GO-Y030 is an α, β-unsaturated monoketone. The present study found that the analogue GO-Y022, which is also a monoketone, had the same degree of p300-HAT inhibitory activity as curcumin. This finding suggests that there is no difference in binding affinity in the Michael reaction between the α, β-unsaturated β-diketone structure of curcumin and the α, β-unsaturated monoketone structure of the analogues. This view is additionally supported by the fact that the analogue GO-Y041, which does not have an unsaturated ketone structure, did not inhibit p300-HAT activity, most likely because it did not bind to p300 due to its lack of a Michael acceptor.  www.nature.com/scientificreports www.nature.com/scientificreports/ Unlike the differences in ketone structures, the differences in the position and type of the functional groups attached to the aromatic rings appear to strongly affect the degree of p300-HAT inhibitory activity. GO-Y030, which has two methoxymethoxy groups (3 and 5) on each aromatic ring, had much greater inhibitory activity www.nature.com/scientificreports www.nature.com/scientificreports/ than GO-Y022, which has a methoxy group (3) and a hydroxyl group (4) on each ring. Moreover, the fact that GO-Y078, which has three methoxy groups (3, 4 and 5) on one ring and two methoxy groups (3′ and 5′) and a hydroxyl group (4′) on another, inhibited p300-HAT activity more strongly than GO-Y031, which has two methoxy groups (3 and 5) and a methoxymethoxy group (4) on each ring. This finding suggests that the addition of a large functional group, such as the methoxymethoxy group at position 4, reduces binding affinity to p300. It has previously been reported that the functional groups on the aromatic rings of curcumin are involved in hydrogen bonding to p300 19 . Furthermore, we previously reported that the methoxy groups at position 3 did not affect p300-HAT inhibitory activity 21 . Taken together, these findings suggest that the addition of a methoxy group or a methoxymethoxy group at position 5 is important for binding affinity to p300 via the hydrogen bond. Thus, the greater p300-HAT inhibitory activity of GO-Y030 compared with that of curcumin can be attributed to the differences in the functional groups of the two compounds, rather than to the differences in their ketone structures.
Our experiment with cultured cardiomyocytes revealed that curcumin inhibited cardiomyocyte hypertrophy and the acetylation of histones at a concentration of 10 μM, whereas GO-Y030 did so at 1 μM, i.e., 1/10th the concentration of curcumin. The results of both the in vitro HAT assay and the cultured cardiomyocyte experiment showed that GO-Y030 had the same degree of inhibitory effect as curcumin at 1/10th the concentration; therefore, as the cell permeability of GO-Y030 and curcumin in cardiomyocytes was roughly the same, and as GO-Y030 was a more potent inhibitor of p300-HAT activity than curcumin in the cardiomyocyte nucleus, it can be concluded that GO-Y030 suppressed cultured cardiomyocyte hypertrophy at a lower concentration than curcumin.
GO-Y030, at a dose 1/100th that of curcumin, ameliorated cardiac hypertrophy and cardiac dysfunction in a pressure overload model. Additionally, it inhibited perivascular fibrosis and hypertrophy of individual cardiomyocytes, and it inhibited histone acetylation. This reveals that, relative to curcumin, GO-Y030 had a 10-fold greater effect in vivo than in cultured cardiomyocytes. There are several possible explanations for this difference in effect. First, it is possible that GO-Y030 acts on other cells in addition to cardiomyocytes and therefore additively suppresses the development of heart failure. Cardiac fibrosis is closely related to the progression of heart failure 22,23 , and we have found that GO-Y030 suppressed the fibrotic response at a lower concentration than curcumin in primary cultured cardiac fibroblasts (data not shown). Thus, it can be assumed that GO-Y030 had a much greater effect in vivo than in cultured cardiomyocytes because it affected both cardiomyocytes and cardiac fibroblasts. Second, it is possible that GO-Y030 has greater bioavailability than curcumin. It has been reported that the low bioavailability of curcumin is due to hydrophobicity, poor absorption, and rapid metabolism 24 . Differences in the structure of curcumin and GO-Y030 may affect these factors, potentially improving the bioavailability of GO-Y030 over that of curcumin. The present study also suggests that GO-Y030 is safe, finding no toxicity for the 0.5 mg/kg dose at six weeks. This confirms the results of previous studies, which found that mice had no adverse reactions to the administration of feed containing 0.1% GO-Y030 for 2 months 25,26 . Further investigation of the bioavailability and possible side effects of GO-Y030 is required to assess its potential clinical application.
The present study has also demonstrated that 1 µM GO-Y030 suppressed the activation of the p300/GATA4 signaling pathway in cardiomyocyte hypertrophy as well as 10 µM curcumin did. p300 controls gene expression by acetylating histone and transcriptional factors; it also regulates transcriptional factors by forming a transcriptional complex as a scaffold protein 27,28 . We previously reported that curcumin suppressed the development of heart failure not only by inhibiting the acetylation of histone and GATA4 but also by interrupting the formation of the p300/GATA4 complex 12 . In this study, GO-Y030 also inhibited both the acetylation of histone and GATA4 and the interaction between p300 and GATA4. These results suggest that GO-Y030 suppressed the development of heart failure via the same mechanisms as curcumin.
In summary, this study has demonstrated that the curcumin analogue GO-Y030 effectively inhibits p300-HAT activity and ameliorates TAC-induced progression of cardiac hypertrophy and heart failure at much lower doses than the natural compound curcumin. These findings suggest that GO-Y030 may have a therapeutic effectiveness in the clinical setting equal to that of curcumin at a much lower dosage. This would be a great benefit both to elderly patients who have difficulty swallowing a large volume of oral medications and to patients under fluid restriction conditions who must keep their water intake low when taking oral medications. Further studies are expected to apply this novel drug to heart failure patients in clinical settings.

Materials and Methods
Materials. Curcumin was purchased from Nagara Science Corporation (Gifu, Japan). The curcumin analogues GO-Y022, GO-Y030, GO-Y031, GO-Y041, and GO-Y078 were synthesised as described previously 25 . These compounds were dissolved in dimethyl sulfoxide and stored at −20 °C.
In vitro p300-HAT assay. The in vitro p300-HAT assay was performed as described previously 21 . In brief, 5 μg of core histones from calf thymus (Worthington, USA) was incubated in HAT buffer with purified p300-HAT recombinant domain in the presence of curcumin or GO-Y030 at room temperature for 30 min. The reactions were initiated by adding acetyl-CoA to each sample and then incubating the sample for 30 min. All samples were subjected to SDS polyacrylamide gel electrophoresis (SDS-PAGE) followed by western blotting with rabbit polyclonal anti-acetyl-histone H3 (Lys9) antibodies and rabbit polyclonal anti-histone H3 antibodies (Cell Signalling Technology, USA). The 50% inhibitory concentration (IC 50 ) was calculated from the concentration-response curve.   . (a,b) (a) Histone fractions from the mouse hearts were subjected to western blotting to assess acetylated histone H3K9 and total histone H3 levels. Full-length blots are presented in Supplementary  Fig. S9. (b) The levels of acetylated histone H3K9 and total histone H3 were quantified. All data are presented as the mean ± SEM (n = 4).

Animal experiments. Male
Immunofluorescent staining and measurement of cardiomyocyte surface area. Immunofluorescent staining of the cultured cardiomyocytes was performed as described previously 12,29 .
The cardiomyocytes were cultured in glass chamber slides (Nalge Nunc International, USA) and were stained with anti-myosin heavy chain (MHC) antibodies (Leica Biosystems, Germany) and Alexa Fluor 555-conjugated anti-mouse IgG (Invitrogen) using the indirect immunoperoxidase method. Hoechst 33258 (Dojinjo, Kumamoto, Japan) was used for nuclear staining. Fifty cardiomyocytes were randomly selected from each group, and the surface area of these cells was measured using ImageJ software (version l.33 u).
Transverse aortic constriction (TAC). C57BL/6J male mice (8 weeks old) were anaesthetised with 1.0-1.5% isoflurane and their limbs were anchored. While the mice were connected to a ventilator (0.1-0.3 mL/min, 150 times/min), the pleura was incised to the second rib, and the aortic arch was ligated using a 7-0 nylon suture ligature with a 27-gauge needle. The needle was then promptly removed, and the intercostal muscle and skin were sutured using 5-0 nylon suture ligature. A sham operation was performed with the same surgical procedure, except that the suture around the aortic arch was not tied. Drug treatments. One day after the operation, the 86 mice that had undergone the TAC operation were randomly assigned to five groups: vehicle (1% gum Arabic, n = 18), 1 mg/kg curcumin (n = 17), 50 mg/kg curcumin (n = 15), 0.1 mg/kg GO-Y030 (n = 18), and 0.5 mg/kg GO-Y030 (n = 18). The compounds were dissolved with 1% gum Arabic and administrated to the mice orally by gastric gavage once a day for 6 weeks. The sham mice (n = 14) were treated with the vehicle orally.
Echocardiography. Echocardiography was performed using a 10-12 MHz probe and a Sonos 5500 Ultrasound System (Philips, The Netherlands) as described previously 12,31 . Interventricular septum thickness at end-diastole (IVSd), left ventricular internal diameter end-diastole (LVIDd), left ventricular internal diameter end-systole (LVIDs), and left ventricular posterior wall thickness (LVPWT) were obtained from M-mode recordings. Fractional shortening (FS) and LV mass were calculated as (LVIDd − LVIDs)/LVIDd × 100 (%) and 1.055 [(IVSd + LVIDd + LVPWT) 3 − (LVIDd) 3 ], respectively. LV mass index (LVMI) is represented as the ratio of LV mass to body weight. Histological analysis. The mice were euthanised, and their hearts were isolated and cut into two transverse slices at the mid-level of the papillary muscles. The samples were fixed with 10% formalin and embedded in paraffin. They were then stained with FITC-conjugated wheat germ agglutinin (WGA) and Masson trichrome (MT). The perivascular fibrotic area was measured as described previously 12,31,32 . The sections were deparaffinised and incubated with FITC-conjugated WGA (Sigma-Aldrich) diluted 1:100 (10 μg/mL) in 1% BSA/PBS for 60 min while being protected from light. From each group, several sections were randomly photographed with a fluorescence microscope (LSM 510 META, Zeiss, Japan), and the surface areas of 50 cardiomyocytes were measured using ImageJ software. MT-stained perivascular sections were photographed using an Eclipse 80i microscope (Nikon, Japan). The areas of perivascular fibrogenesis were measured using ImageJ software, and the resulting value divided by the total area of the photograph (12,288 pixels) was regarded as the relative vascularised fibrosis area. The entire heart was imaged with a Leica TL5000 Ergo microscope (Leica Microsystems, Japan), and the area of interstitial fibrosis was measured using ImageJ software. The results are represented as the relative fibrosis area (% of total myocardial area).