Cardiac remodeling and arrhythmogenesis are ameliorated by administration of Cx43 mimetic peptide Gap27 in heart failure rats.

Alterations in connexins and specifically in 43 isoform (Cx43) in the heart have been associated with a high incidence of arrhythmogenesis and sudden death in several cardiac diseases. We propose to determine salutary effect of Cx43 mimetic peptide Gap27 in the progression of heart failure. High-output heart failure was induced by volume overload using the arterio-venous fistula model (AV-Shunt) in adult male rats. Four weeks after AV-Shunt surgery, the Cx43 mimetic peptide Gap27 or scrambled peptide, were administered via osmotic minipumps (AV-ShuntGap27 or AV-ShuntScr) for 4 weeks. Cardiac volumes, arrhythmias, function and remodeling were determined at 8 weeks after AV-Shunt surgeries. At 8th week, AV-ShuntGap27 showed a marked decrease in the progression of cardiac deterioration and showed a significant improvement in cardiac functions measured by intraventricular pressure-volume loops. Furthermore, AV-ShuntGap27 showed less cardiac arrhythmogenesis and cardiac hypertrophy index compared to AV-ShuntScr. Gap27 treatment results in no change Cx43 expression in the heart of AV-Shunt rats. Our results strongly suggest that Cx43 play a pivotal role in the progression of cardiac dysfunction and arrhythmogenesis in high-output heart failure; furthermore, support the use of Cx43 mimetic peptide Gap27 as an effective therapeutic tool to reduce the progression of cardiac dysfunction in high-output heart failure.

Cx43 protein expression in the hearts of AV-Shunt rats are not affected by Gap27 administration. AV-Shunt Scr rats expressed similar protein levels of Cx43 compared to the ones observed in Sham Scr rats ( Fig. 5a,b). In addition, we found lateralization of Cx43 protein in AV-Shunt Scr rats compared to Sham rats as evidenced by immunoreactivity in the lateral membrane of the cardiomyocyte, and this was not improved by Gap27 treatment (Fig. 5c).

Discussion
In the present study, we sought to determine the effect of chronic administration of Cx43 mimetic peptide Gap27 on the progression of cardiac dysfunction and arrhythmogenesis in a high-output heart failure model induced by volume overload (i.e. AV-Shunt). The main findings of the present study are: i) Chronic Gap27 peptide administration decreases the progression of cardiac deterioration; ii) improves cardiac systolic and diastolic function; iii) reduces arrhythmia incidence and severity; and iv) reduces cardiac hypertrophy, adverse remodeling, and lung congestion in rats with high-output heart failure. Together, these findings strongly suggest that Cx43 plays a critical role in the progression of cardiac arrhythmogenesis and cardiac dysfunction in volume-overload heart failure.
Heart failure pathophysiology is characterized by a progressive decline in systolic (or contractile properties) and/or diastolic function (or passive properties) 2 . We found that 4 weeks after creation of AV-Shunt, animals display enlarged LV chambers in both systole and diastole that continues to increase up to 8 weeks post-AV-Shunt. Importantly, four weeks of chronic Gap27 peptide administration attenuated increases in systolic and diastolic chamber volumes in the AV-Shunt Gap27 group. This result suggests that Gap27 treatment has the potential to slow disease progression. Furthermore, AV-Shunt rats display both systolic and diastolic cardiac function impairment at 8 weeks post-induction, and chronic Gap27 treatment improved the contractile and passive properties of the heart in these animals. Interestingly, these effects were not exclusively observed in AV-Shunt animals since Sham rats treated with Gap27 also showed a mild improvement in systolic and diastolic cardiac function ( Table 2). One plausible explanation for the beneficial effects of Gap27 could be a decrease in the "leak" of current through Cx43 hemichannels; which in turn would promote electrical conductivity primarily through gap junctions between cardiac cells and therefore reduce the formation of arrhythmic foci 11 . We speculate that in AV-Shunt rats 23 , the generation of cardiac arrhythmias could be related to the leakage of current from cardiomyocytes through hemichannels, creating a vicious cycle that promotes arrhythmogenesis and worsens cardiac function. Nevertheless, despite that mainly Gap27 peptide is used as a Cx43 blocker, it is important to mention that this peptide could block other connexin-based channels (i.e. Cx37) and also could affects Cx43 gap junction. Importantly, it has been shown that Cx43 hemichannel formation take place in the Golgi apparatus and then are transported directly to intercalated disc in cardiomyocytes in vitro 23,24 . Furthermore, Gap27 peptide lacks the TAT sequence which allow the peptide to easily permeate the cell membrane. However, we cannot completely rule out the possibility that Gap27 affect gap junctions. Then, it is possible to speculate that Gap27 peptide might decrease current leakage between cardiomyocytes, improving cardiac cell synchronicity and reducing arrhythmia incidence.
Cardiac arrhythmias are a strong predictor of morbidity and mortality in heart failure patients 6,25 , and play an especially prominent role in sudden cardiac death in this population 26,27 . In agreement with our previous findings, we observed a high incidence of cardiac arrhythmias in AV-Shunt rats 18,20 . AV-Shunt rats had a high arrhythmia score compared to Sham rats and chronic treatment with Gap27 markedly reduced the incidence and severity of the arrhythmic events. Recent studies by González et al. 16 showed a similar salutary effect of Cx43 blockade on adrenergic-induced cardiac arrhythmias in a Duchenne muscular dystrophy model 16 . In addition, it has been shown that transplantation of Cx43-expressing embryonic cardiomyocytes (Cx43-eCMs) in myocardial infarction zone, markedly protects the heart from ventricular tachycardia induced by electrical pacing 13 . Together, these results suggest that controlling/improving current flow through gap junction channels instead of hemichannels in the failing heart reduces arrhythmia incidence.
Connexins are key proteins that allow the conduction of the electrical impulse between cardiomyocytes 10,12 . It has been described that one of the mechanisms related to cardiac arrhythmogenesis in various cardiovascular pathologies and animal models of heart diseases is the uncoupling of gap junctions and connexin hemichannel  www.nature.com/scientificreports www.nature.com/scientificreports/ formation [12][13][14][15][16]28 . Some studies showed that minutes after the induction of ischemia in isolated rat hearts, a progressive increase in electrical uncoupling occurs, which is correlated with the formation of hemichannel structures 29 . Furthermore, alterations in the expression of Cx43 protein are associated with the generation of cardiomyopathies 13,30 . Knock-out mice lacking Cx43 protein die early, while partial knock-out mice, which express <50% of the protein, showed low survival due to cardiac dysfunction 30 . In contrast, overexpression of Cx43 in cardiac cells improves conductivity in infarcted hearts 31 . Here we showed no differences in the expression levels of Cx43 in cardiac tissue from AV-Shunt rats compared to Sham rats and that Gap27 treatment improve cardiac arrhythmogenesis in AV-Shunt rats without changing Cx43 protein expression. In addition, we found a higher lateralization of Cx43 in cardiomyocytes from AV-Shunt animals which was not affected by Gap27. Recently, it has been proposed that Gap27 may work better on Cx43-S368 phosphorylated form 32 . Cx43-S368 is phosphorylated by PKC 33 . Under normal conditions in neonatal rat hearts, Cx43 display low phosphorylation in S368 34 . However, after 42 days of subcutaneous injections of isoprotenerol (a sympatho-mimetic agent), the phosphorylation of Cx43 and the expression of PKC in the left ventricle were increased, which was correlated with Cx43 lateralization 35 . These results suggest that increasing cardiac sympathetic drive enhances PKC activity and Cx43-S368 phosphorylation which may contribute to Cx43 lateralization in the heart. Then, is plausible to speculate that in HF conditions, a known condition where sympathoexcitation take place, cardiac Cx43 lateralization depends on S368 phosphorylation. Further studies are needed to completely address Cx43 phosphorylation status and the beneficial effect, if any, of administration of more specific mimetic peptide targeted against Cx43-S368 in HF pathophysiology.
Analysis of EKG waves can provide an indirect measure of cardiac electrical conductivity. In our experiments, we found several EKG abnormalities in AV-Shunt rats, including increases in P wave duration, the P-R interval, and the QRS interval. Importantly, chronic Gap27 administration normalized the observed EKG abnormalities in AV-Shunt rats. Prolonged P wave duration is a marker of electromechanical dysfunction which has been closely associated with an increase in atrial chamber size, atrial fibrillation, and death in cardiovascular diseases 36,37 . Also, it is thought that prolonged P wave duration reflects the presence of arrhythmogenic substrates both in the atrium and in the ventricle. In support of this notion, a linear relationship between the fibrotic index and the duration of the P wave has been observed 38 . The P-R interval denotes the time from the beginning of atrial depolarization to the onset of ventricular depolarization 39 . A prolonged P-R interval can be attributed to delayed conduction through the AV node, through the atrial tissue, or through the Purkinje system. More importantly, a prolonged P-R interval is associated with a higher risk of atrial fibrillation and mortality 39 . Importantly, our findings showed that AV-Shunt rats displayed increases in P-R interval duration compared to Sham rats and that Cx43 mimetic peptide Gap27 significantly reduces PR interval duration. However, our results did not discard the effect of Gap27 on Cx43 gap junction formation in the intercalated disc, which could contribute to decrease P-R duration in AV-Shunt animals. www.nature.com/scientificreports www.nature.com/scientificreports/ Patients with heart failure often present with a prolonged QRS interval 40,41 . Prolonged QRS interval is indicative of delayed electrical conduction, which can result in both intra-and interventricular mechanical desynchrony. This desynchrony leads to inefficient contraction and an increase in energy demand ultimately leading to structural remodeling of the heart [42][43][44] . Cardiac hypertrophy in conjunction with structural remodeling and reduction of cardiac efficiency results in an increase in wall stress and electrical remodeling like Cx43 re-localization from the intercalated disc to the lateral membrane. Together these perpetuate the desynchrony and remodeling cycle contributing to deterioration of cardiac function in heart failure 45 . Accordingly, ventricular desynchrony in heart failure patients is associated with poor outcomes. Interestingly, heart failure patients who have successfully responded to resynchronization therapy, to reverse desynchrony, shown signs of reverse remodeling (i.e. LVEDV reductions) along with improve survival 43,46 . We found that Gap27 treatment reduces LVEDV, hypertrophic index, and collagen content in the hearts of AV-Shunt rats. Therefore, Gap27 treatment may improve cardiac synchrony in AV-Shunt rats. However, we did not specifically test this hypothesis in this study. Moreover, it has been shown that Cx43 expression is not restricted only to cardiomyocytes in the heart but also is expressed in fibroblast 47 . Therefore, is it possible that our results regarding cardiac remodeling could be associated to effects on cardiomyocytes and also to fibroblast in the heart. Our data suggests that Gap27 treatment improves cardiac electrical conductivity in both the atria and the ventricle. Future studies should address whether Gap27 treatment improves cardiac synchrony.
In summary, we found that chronic treatment with Gap27 prevents the progression of cardiac dysfunction, improves electrical and cardiac remodeling and reduces arrhythmogenesis in high-output heart failure. Our results suggest that Gap27 may have salutary effects in the setting of heart failure as a therapeutic tool for the treatment of cardiac arrhythmogenesis and cardiac function impairment.

Materials and Methods
High-output heart failure model. Adult male Sprague-Dawley rats (320 ± 9 g) (n = 24) underwent arterio-venous shunt (AV-Shunt) surgery to induce overload volume to the heart. This protocol was performed under sterile conditions, keeping the animal anaesthetized with isoflurane (5% for induction, 1.5% for maintenance in O 2 ). The postoperative management consisted in the administration of 1 mg/kg of ketoprofen (s.c.), 5 mg/kg of enrofloxacin (s.c.) and 5 ml of sterile physiological saline (i.p.), together with the topical administration of 2% lidocaine hydrochloride. Rats with Sham surgery (n = 16) were handled under the same conditions. At 4th week post-shunt, the degree of cardiac failure was evaluated by M-mode echocardiography on an anaesthetized animal under isoflurane (5% for induction, 1.5% for maintenance in O 2 ). Rats that fulfil following criteria were considered in this study: i) Ejection fraction (EF) ≥ 50; ii) End diastolic volume (EDV) and iii) stroke volume (SV) ≥ mean+2 standard deviation (SD) fold changes relative to Sham Scr rats of its experimental series [18][19][20] . Animals were handled in accordance with standards set by the National Institutes of Health guidelines for the Care and Use of Laboratory Animals. All experimental protocols were approved by the Ethics Committee of the Pontificia Universidad Católica de Chile (protocol ID #170710022).
Chronic Gap27 peptide administration. At fourth week post-shunt a second surgery for the implantation subcutaneous of an osmotic mini-pump (2.5 µl/hr, 2ML4, ALZET) containing the Cx43 mimetic peptide Gap27 (Sham Gap27 and AV-Shunt Gap27 ) (ANASPEC, amino acid sequence SRPTEKTIFII) or Scrambled Gap27 (Sham Scr and AV-Shunt Scr ) (ANASPEC, amino acid sequence RQLIITSFIPT) at a concentration of 1 µg/kg diluted in sterile physiological saline were implanted for 28 days 48 . The surgery was performed under general anesthesia with isoflurane (5% for induction, 1.5% for maintenance flush in O 2 ) and a postoperative management consisted in the administration of 1 mg/kg of ketoprofen (s.c.), 5 mg/kg of enrofloxacin (s.c.) and 5 ml of sterile physiological saline solution (i.p.), together with the topical administration of 2% lidocaine hydrochloride. www.nature.com/scientificreports www.nature.com/scientificreports/ Invasive hemodynamic parameters. Rats were anaesthetized with a mixture of 800 mg/kg urethane (SIGMA, USA) and 40 mg/kg α-chloralose (SIGMA, USA). A transducer catheter (SPR-869, MILLAR INSTRUMENTS) was introduced into the left ventricle through the right carotid. Cardiac left ventricle pressure-volume (PV) loops were recorded at steady state. Measurements were calibrated by injecting a hypertonic saline bolus (30% wt/vol NaCl) to determine conductance, and relative volume units were converted absolute using the cuvette calibration method. The LV end-of-systole volume (LVESV), LV end-diastolic volume (LVEDV), LV end-systole pressure (LVESP), LV end-diastolic pressure (LVEDP), ejection volume (SV), cardiac output (CO) and ejection fraction (EF) were obtained from 10-15 representative loops. Volume at pressure 0 and End-Systolic Pressure-Volume Relationship (ESPVR) were used as indicators of diastolic and systolic function, respectively. These analyses were calculated by a mathematical algorithm from a single beat obtained from pressure-volume curves according to Klotz et al. 21 for V-0 and Takeushi et al. 22 for ESPVR. All recordings were sampled to 1 kHz and were analyzed using LABCHART7 Pro v7.2 software (ADINSTRUMENTS) 18,19,21,22,49 . Arrhythmia Scoring and electrocardiography (EKG) analysis. 2-lead EKG was recorded in anaesthetized rats with a mixture of 800 mg/kg urethane (SIGMA, USA) and 40 mg/kg α-chloralose (SIGMA, USA). The incidence of arrhythmias was determined as arrhythmic events in one hour of recording. The arrhythmia score was calculated according to scores previously used in Curtis et al. (1988Curtis et al. ( -2013 50,51 and Miller et al. 51 , based on the kind and duration of the arrhythmic event, and using the definitions established in the Lambeth Conventions II 52 . A ventricle premature beat (VPB) was defined as a ventricular electrical complex (complete electrical event: QRS, RS, QRST or RST) that is different in shape (voltage and/or duration, i.e., height and/or width) from the preceding (non-VPB) ventricular complex. Ventricle Tachycardia (VT) was defined as a sequence of a less 4 VPB consecutive. Ventricle Fibrillation (VF) was defined as a sequence of a minimum of 4 consecutive ventricular complexes without intervening diastolic pauses, which peak-peak interval and the height vary without a pattern. In addition, for waves analysis, the EKG signal was processing with the Stationary Wavelet Transform Denoising 1D tool from MATLAB software. Then, the signal denoised was analyzed with EKG plugin from LABCHART 7 Software (ADINSTRUMENTS). Parameters such as the P wave, PR interval, QRS interval, QTc (Bazzett), T wave, RR interval were calculated. All recordings were sampled to 1 kHz 49-51 . Histological analysis. After rats were euthanized, hearts were removed and arrested in diastole with KCl 1 M (SIGMA, USA). Immediately, hearts were fixed with 4% paraformaldehyde and were stored in PBS + 0.1% azide until embedded in paraffin. Transversal sections of 5 µm thickness were stained with modified Masson's trichrome stain (without hematoxylin) (DIAPATH, Italy) and picrosirius red (DIAPATH, Italy). The deposition of interstitial collagen was quantified by using IMAGEJ software (NIH Software, USA) in images acquired at 40×(NIKON ECLIPSE E400). Per cent fibrosis was expressed as the ratio of the fibrotic area into total left ventricle area.
Immunostaining. Cardiac sections of 20 µm thick were cut in a cryostat at −20 °C. Tissue were incubated with blocking solution with 5% gelatin from cold-water fish skin, 0.5% Triton X-100 and Phosphate Buffer Saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ). Then were incubated with primary rabbit antibody against Connexin-43 (71-0700, THERMOFISHER) at a 1:100 dilution in blocking solution, followed by incubation with the secondary anti-rabbit antibody conjugated with Alexafluor-488 at a 1:1,000 and Alexa Fluor-594 phalloidin at a 1:1,000 diluted in blocking solution. Slices were mounted in Vectashield with DAPI (VECTOR Laboratories). The immunostaining was visualized in a high-resolution fluorescent microscope (LEICA, Germany) and images were acquired at 40×. Western blot. Left ventricle tissue samples were homogenized in RIPA buffer (SIGMA, USA) supplemented with 1% protease inhibitor (SIGMA, USA) plus 1% of phosphatases inhibitors (PHOSPHOSTOP SIGMA, USA), then centrifuged at 12,000 rpm for 30 min at 4 °C and the supernatant was stored at −80 °C. The protein concentration was determined by a BCA assay (PIERCE), following the manufacturer's specifications. 50 ug of proteins were separated by SDS-PAGE under reducing conditions on gels of 12% of acrylamide and electrophoretically transferred to PVDF membranes (MILLIPORE). The electrophoretic run was performed in the Tris-glycine buffer for 2 hours at 100 V constant voltage and the transfer was performed in cold, for 2 hours at 300 mA constant. The membrane was blocked with 5% non-fat milk in Tris Saline Buffer (20 mM Tris base, pH 7.2, 0.3 M NaCl) with 0.05% tween20. The membrane was immunoblotted with the primary antibodies against Connexin-43 (71-0700, THERMOFISHER) and GADPH (sc-32233, SANTACRUZ) at a 1:1000 and 1:2,000 dilution in 5% milk in TBS, respectively. Next, were incubated with secondary antibody conjugated with HRP (Anti-rabbit: 074-1506, KPL; Anti-mouse: 074-1806, KPL) at a dilution of 1:2000 in blocking solution plus 0.1% tween20. Proteins were visualized using ECL Western Blotting Detection kit (PIERCE). Each membrane was exposed to high sensitivity film (BIOMAX MR Film of Carestream, KODAK) and was revealed with development and fixation solution (DENTUS, AGFA, USA). Finally, the films were scanned, and the images were analyzed with the software image studio lite version 5.2 (LI-COR BIOSCIENCE, USA). The relative amount of proteins was calculated as the intensity ratio of the band of the protein of Cx43 versus the intensity of the band corresponding to GADPH.

Statistical analysis.
The results were shown as mean ± standard error of the mean (SEM). Differences between groups were assessed with one or two-way ANOVA, accordingly to matrix data, followed by Sidak´s posthoc comparisons for parametric distribution. While for non-parametric data, was analyzed with Kruskal-Wallis test followed by Dunn post hoc test. Differences between two groups were assessed t-test for parametric distribution or Mann-Whitney test for nonparametric distribution. A probability value of p < 0.05 was considered statistically significant. All tests were performed by GRAPHPAD PRISM software (La Jolla, CA, USA, version 7.0).