Identification of new biophysical markers for pathological ventricular remodelling in tachycardia‐induced dilated cardiomyopathy

Abstract Our aim was to identify biophysical biomarkers of ventricular remodelling in tachycardia‐induced dilated cardiomyopathy (DCM). Our study includes healthy controls (N = 7) and DCM pigs (N = 10). Molecular analysis showed global myocardial metabolic abnormalities, some of them related to myocardial hibernation in failing hearts, supporting the translationality of our model to study cardiac remodelling in dilated cardiomyopathy. Histological analysis showed unorganized and agglomerated collagen accumulation in the dilated ventricles and a higher percentage of fibrosis in the right (RV) than in the left (LV) ventricle (P = .016). The Fourier Transform Infrared Spectroscopy (FTIR) 1st and 2nd indicators, which are markers of the myofiber/collagen ratio, were reduced in dilated hearts, with the 1st indicator reduced by 45% and 53% in the RV and LV, respectively, and the 2nd indicator reduced by 25% in the RV. The 3rd FTIR indicator, a marker of the carbohydrate/lipid ratio, was up‐regulated in the right and left dilated ventricles but to a greater extent in the RV (2.60‐fold vs 1.61‐fold, P = .049). Differential scanning calorimetry (DSC) showed a depression of the freezable water melting point in DCM ventricles – indicating structural changes in the tissue architecture – and lower protein stability. Our results suggest that the 1st, 2nd and 3rd FTIR indicators are useful markers of cardiac remodelling. Moreover, the 2nd and 3rd FITR indicators, which are altered to a greater extent in the right ventricle, are associated with greater fibrosis.

structural changes in the tissue architectureand lower protein stability. Our results suggest that the 1st, 2nd and 3rd FTIR indicators are useful markers of cardiac remodelling. Moreover, the 2nd and 3rd FITR indicators, which are altered to a greater extent in the right ventricle, are associated with greater fibrosis.

K E Y W O R D S
biophysical markers, cardiac remodelling, collagen, differential scanning calorimetry, fourier transform infrared spectroscopy, heart failure, myofiber

| INTRODUCTION
Non-ischemic dilated cardiomyopathy (DCM) is characterized by left ventricular (LV) dilatation and global systolic dysfunction with normal coronary arteries. 1,2 Progressive heart failure, ventricular and supraventricular arrhythmias, thromboembolisms and sudden death are the main clinical manifestations. [2][3][4] In addition, DCM constitutes the most common cause of heart failure referred for cardiac transplantation. 5 The ventricular remodelling resulting in ventricular dilatation and dysfunction has been extensively studied in vivo by means of non-invasive techniques and post-mortem in human and animal studies by histopathological and biochemical analysis. [6][7][8][9] However, the current knowledge of the mechanisms involved in the genesis of DCM is still limited. As a result, the treatments are scarce and have incomplete efficacy.
The molecular, conformational and physical characterization of the myocardium has emerged as a novel approach to study remodelling in diseased tissues. Spectrophotometric techniques, such as Fourier Transform Infrared Spectroscopy (FTIR) are powerful techniques that have been successfully applied to characterize cardiac and vascular tissues. [6][7][8][9][10] Differential scanning calorimetry (DSC) is another suitable technique for characterizing biological tissues at the mesoscale, evaluating freezable and unfreezable water 11 and assessing protein thermal stability and conformational changes in tissues. 12 DSC is particularly appropriate for evaluating the thermal stability of collagen in its purified or aggregated form, 13 directly in cardiovascular tissues 14 or in biomaterials. 15 DSC has been also successfully applied to characterize the protein components of muscle tissues, such as myosin and its subfragments, actin and sarcoplasmic proteins. [16][17][18] Nevertheless, few DSC data are available on the whole myocardium tissue, and no calorimetric data exist on tachycardia-induced DCM.
The objective of the current investigation was to identify molecular, conformational and biophysical alterations useful as biophysical markers of cardiac remodelling in non-ischemic dilated cardiomyopathy (DCM).

| Generation of a pig model of tachycardiainduced DCM and experimental procedure
This study includes seventeen female domestic swine (Landrace-Large White cross) weighing 54 AE 3 kg: seven control healthy animals (control group) and ten animals with tachycardia-induced NIDCM (DCM group). The study protocol was approved by the Animal Care and Use to obtain an accurate calculation of the parameters. The heart rhythm and QRS width complex in the ECG were calculated. In the echocardiogram, the LV ejection fraction (LVEF) and the end-diastolic (ED) and end-systolic (ES) LV dimensions were measured. The femoral vein and artery were cannulated, and two catheters (Millar Instruments, Inc., Houston, TX, USA) were placed into the right (RV) and the left (LV) ventricles for measurement of the LV and RV pressures. Subsequently, a mid-thoracotomy was performed. Before death, a bolus of pentobarbital was administered. The heart was excised, and samples from the RV and LV were immediately frozen at À80°C for molecular, biochemical and biophysical characterization. For the histological analysis, the myocardial samples were fixed, cryopreserved in 30% sucrose in phosphate-buffered saline, embedded in Tissue-Tek O.C.T. (Sakura â ), and snap-frozen in liquid nitrogen-cooled isopentane.

| Human control samples
Human autopsy hearts (n = 3) were obtained at Department of Pathology, Hospital Santa Creu i Sant Pau from deceased patients who died of non-cardiac causes. Hearts were weighed, measured and samples from RV and LV ventricles were excised and frozen at À80°C for lipid characterization. The project was approved by the local Ethics Committee of Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, and conducted in accordance with the guidelines of the Declaration of Helsinki. All patients gave written informed consent that was obtained according to our institutional guidelines.

| Myocardial fibrosis
It was assessed at the molecular level [analysis of the collagen type I and type III mRNA expression and protein levels] and by histological Masson's trichromic staining. For the latter, interstitial collagen deposition was assessed as the percentage of blue staining of 5 myocardial areas of 10 different immunohistochemical sections per heart using the ImageJ software.

| Determination of cardiomyocyte number
The amount of cardiomyocytes was calculated as the sum of nuclei observed in 5 fields at 409 of 10 different immunohistochemical sections.

| Determination of cardiomyocyte size
Sections of both venticles were stained with haematoxylin/Eosin. All the slides were analysed using an Olympus VANOX AHBT3 microscope and were photodocumented using a Sony DXC-S500 camera.
Longitudinal and transversal diameter of the cardiomyocytes were measured in at least 5 microscopic uniformly distributed fields at 409 of 10 different immunohistochemical sections per heart.

| Western blot analysis
The protein levels of collagen and HSP70 were determined by Western blot analysis. Equivalent amounts of total protein were electrophoresed under reducing conditions on SDS-polyacrylamide gels. The samples were electrotransferred to nitrocellulose membranes, which were then saturated at room temperature for 1 hours in TTBS (20 mm Tris-HCl, pH 7.5, 500 mm NaCl, 0.01% Tween 20, and 5% non-fat milk). Western blot analyses were performed with specific monoclonal antibodies against collagen type III (Abcam, clone FH-7A ab6310) and heat shock proteins 70 (HSP70; Abcam, ab47455) and their corresponding secondary antibodies (1:10,000 dilution; Dako). Equal protein loading in each lane was verified staining filters with Pounceau and also by incubating the blots with monoclonal antibodies against Troponin T (Thermo Scientific, clone 13-11, cardiac isoform Ab-1).

| Myocardial lipid content
Myocardial-pulverized tissue (7 mg) from the porcine RV and LV was weighed and homogenized in 0.1 mol/L NaOH. The protein content of the extracts was quantified by a Pierce BCA Protein Assay (Thermo Fisher Scientific, Waltham, MA) to normalize the lipid content. Lipid extraction was performed as previously described. [19][20][21][22] The lipid extract was concentrated by evaporating the organic solvent under a N 2 stream to prevent lipid oxidation. and cardiolipin (CL) was applied to each of these types of plates plate type, respectively, as standards. The spots corresponding to CE, TG, FC, PE, PC, SM and CL were quantified by densitometry using a ChemiDoc system and Quantity-One software (Bio-Rad, Hercules, CA).

| Vibrational characterization
One portion (5 mg) of myocardial tissue was freeze-dried and used for vibrational characterization. Fourier transform infrared spectroscopy/attenuated total reflectance (FTIR/ATR) spectra of the freeze-dried tissues were acquired using a Nicolet 5700  Spectra were then subjected to ATR and baseline corrections and normalized in the amide II region. These spectra were next used in calculation of integrated bands intensities and their ratios. To quantify these various components, the areas of the different absorption bands were computed from the individual spectrum of each tissue, and the appropriate ratio of areas was used according to the literature data. 23,24 Second derivatives and Fourier self-deconvolution (using k = 1.7 and half-width = 13.5/ cm) were used to enhance the chemical information present in overlapping infrared absorption bands of spectra. All spectra processing was performed using Omnic 8.0. The spectra presented for each group were calculated by averaging the spectra of all samples within each group.

| Differential scanning calorimetry
Calorimetric analyses were performed using a DSC Pyris calorimeter (Perkin Elmer, Waltham, MA). The calorimeter was calibrated using Hg and In as standards, resulting in a temperature accuracy of AE0.1°C and an enthalpy accuracy of AE0.2 J/g. Fresh samples, 5-10 mg in weight, were set into hermetic aluminium pans and equilibrated at the initial temperature for 5 minutes before cooling to À100°C at 10°C/min. Then, the thermograms were recorded during the heating at 10°C/minutes until reaching 90°C. Once DSC measurements were performed, the pans were reweighted to check that they had been correctly sealed. A second series of experiments were performed on freeze-dried samples; in this case, freeze-dried segmented parts of myocardium tissues (2-3 mg) were set into standard aluminium pans and equilibrated at 20°C before recording thermograms during heating at 10°C/minutes until reaching 200°C. A detailed description of the procedure to calculate the amount of total, freezable and unfreezable water in hydrated proteins and tissues has been included in the material and methods section of the Supporting information.

| Statistical analysis
Continuous variables are shown as the mean AE standard deviation (SD). Variables were compared between the groups using Student 0 s t-test for independent samples and one-way ANOVA, followed by Tukey's post hoc test, for comparison between each subgroup. Differences were considered to be statistically significant when P < .05.

| Phenotype of pacing-induced heart failure in DCM pigs
Animals in the DCM group underwent 23 AE 2 days of RV rapid pacing (detailed in Figure 1), and all developed dilated cardiomyopathy according to cardiac function parameters. As shown in  (Figure 2A), suggesting that this mechanism of cardiac protection is activated in DCM pigs. In addition, glycogen storage seems to be a key feature in the protection of hibernated myocytes. 28,29 Therefore, we mea-  Figure S1). According to the scarce MAC387 staining, inflammation seems to be almost absent in porcine myocardial samples ( Figure 3A).

| Pattern of FTIR bands in the myocardium of DCM pigs
A detailed Table (Table S1) and description of the FTIR bands detected in porcine myocardial control tissue can be found in the Supporting Information. As shown in Figure 4A-C, the major protein absorption bands reported in Table S1 for the control ventricles were conserved at the same wave numbers in the averaged

| FTIR indicators of myofibrillar and extracellular matrix proteins altered in DCM ventricles
Two distinct FTIR indicators of the myofiber/collagen ratio, the 1st As shown in Figure S2B, the amounts of different protein secondary structures were mostly preserved in dilated ventricles. We only observed a slight increase in the band 1632/cm, which is assigned to b sheets structures, in the dilated samples.

| Neutral lipid content (cholesteryl ester, free cholesterol, triglycerides, fatty acids and phospholipids) alterations in dilated ventricles
Quantitative analysis of neutral lipids by thin layer chromatography (TLC) after lipid extraction showed that in control pigs, the RV contained much higher triglyceride (TG; Figure 6A) and cholesteryl ester (CE) levels ( Figure 6B) than the LV. This strong difference in TG and CE between control RV and LV was  Figure 6A and B). Accordingly, there were no differences in the total lipid content measured by TLC ( Figure 6C) or FTIR indicators ( Figure 6D) between the dilated ventricles. The phospholipid pattern of the myocardium (described in detailed in the Result section of the Supporting Information) did not show differences between ventricles or groups ( Figure S3).

| FTIR indicators of proteoglycans and polysaccharides altered in DCM ventricles
Proteoglycans mainly contribute to the 1079/cm band and the specific band at 1226/cm, overlapping with protein amide III (Table S1).
Glycogen and other polysaccharides contribute to the FTIR spectrum of the myocardium in the 1200-1000/cm region ( Figure 4C). The First, we have validated our in vivo pig model as a translational model of dilated cardiomyopathy. It has been previously reported that heart failure is associated with global myocardial metabolism abnormalities typical of myocardial hibernation in different in vivo pig models. 33,34 Therefore, we tested several molecular candidates of myocardial hibernation in our in vivo model. We have measured protein levels of HSP70, heat shock proteins ubiquitously expressed that play a role in protein folding and exert protective effects. The high HSP70 levels that we have found in the dilated ventricles of our pig model indicate that this protective mechanism found in different in vivo models of heart failure also is present in our in vivo model. Previous studies also reported that glycogen storage is a key feature in the protection of hibernated myocytes. 28,29 Our results show that glycogen phosphorylase, mainly involved in the glycogen degradation pathway, was significantly down-regulated in the both, RV but also in LV. These results indicate that, in our in vivo model, myocardial hybernation is not a restricted phenomena limited to the pacing site or pacing ventricle but a more extensive process affecting the whole heart, as previously described in other in vivo models of heart failure. 33,34 Like in these in vivo models, global myocardial metabolism alterations are compatible with differences in regional contractility. Indeed, we found crucial mechanical differences between right and left ventricles in our in vivo model. In addition, we have found other characteristics in the heart of our pigs, such as Our model showed a high degree of fibrosis in both dilated ventricles, although the extent was greater in the right ventricle. The extended fibrosis in the RV is associated with a greater reduction of 2nd FTIR indicator (myofiber/collagen ratio) and a significant increase in FTIR 3rd indicator (carbohydrate/lipid ratio). The 3rd indicator augmentation seems to be related, at least in part, to the decrease in triglyceride (TG) content specifically in the right dilated ventricles. Our study revealed higher TG and CE contents in the RV than in the LV of control animals. This phenotype is consistent with the higher effort and energetic consumption needs of a healthy LV.
In DCM pigs, these differences in the neutral lipid content between ventricles are lost because of decrease in the RV, concomitant with an increase in the LV. Different research groups including ours have reported an increased TG myocardial accumulation in several cardiomyopathies, including dilated, 19 ischemic 20-22 and diabetic cardiomyopathy. [40][41][42] In this particular pacing-induced DCM model, the TG increase occurs in the LV, while the opposite occurs in the RV.
This finding suggests important differences in the progression of the RV and LV from a healthy to a pathological state, at least in terms of lipid accumulation. In line with our results, previous studies have shown that tachycardia decreases the TG content of the RV in a rat model of dilated cardiomyopathy. 43 The authors propose that pacing places a greater burden on the RV and that this justifies higher TG mobilization. Additionally, in accordance with our results, a previous study using a Syrian hamster model reported that cardiomyopathy progresses with an increase in ECM and a decrease in cellular lipids. 24 In addition to the global myocardial metabolism abnormalities (typical of myocardial hibernation) and to the alterations in TG and CE content in the hearts of DCM pigs, we also observed a reduction in creatine levels (quantified by the specific band at 1304/cm, 44 Table S1). Taken together, these results suggest that structural remodelling in this pig model is closely associated with metabolic derangements. These results support the translational nature of our pig model because, in humans, dilated cardiomyopathy occurs with a progressive reduction in creatin. 45,46 Moreover, a tight relation between alterations in creatin levels and the severity of heart failure estimated by ejection fraction has been previously reported in humans. 46,47 In conclusion, our work identifies the FTIR 1st and 2nd indicators

| CLINICAL IMPACT
Our results, obtained in a high translational model using novel research techniques, provide new key biophysical markers of pathological ventricular remodelling that are useful for the characterization of dilated cardiomyopathy. Interestingly, these biophysical markers showed significant differences between the right and the left ventricles, indicating ventricle-specific remodelling alterations. In addition, the study supports the concept that ultrastructural alterations persist in tachycardiomyopathy-induced cardiomyopathy, as some clinical studies have suggested.

CONFLI CT OF INTEREST
None.