Hypertrophic cardiomyopathy (HCM) is identified in about 1 in 500 people. It is the leading cause of sudden cardiac death in young adults and results in significant disability in survivors. Mutations in at least 14 genes encoding proteins of the cardiac sarcomere (with over 1,400 variants) are associated with HCM. In most cases, HCM is inherited as an autosomal dominant trait with variable penetrance (Maron et al., 2014), although the causative mutation cannot be identified in about half of the cases. This may be explained by either existence of rare sarcomeric or non-sarcomeric genetic variants, the influence of modifier genes or the presence of phenocopies (Maron et al., 2010; Lopes et al., 2015). Alterations in two genes, β-myosin heavy chain and myosin-binding protein C (MYBPC3) account for ~75% of cases where an underlying mutation has been identified (Alfares et al., 2015).

HCM has an exceedingly high prevalence in cats, affecting ~16% of the outbred population (Paige et al., 2009; Payne et al., 2015). This prevalence increases in pedigree breeds such as the Maine Coon where it is estimated at 26% (Gundler et al., 2008). A heritable component for HCM, has also been described in Ragdoll, Siberian, Sphynx, American Shorthair, Cornish Rex, Persian, European, British Shorthair, Bengal, Chartreux, and Norwegian Forest cats (Trehiou-Sechi et al., 2012; Longeri et al., 2013; März et al., 2014). Despite this, causative mutations have only been identified in Maine Coon and Ragdoll cats. In both breeds, a homozygous mutation in MYBPC3 was identified resulting in amino acid substitution—A31P and R820W in Maine Coons and Ragdolls, respectively (Meurs et al., 2005, 2007). The same R820W mutation has been identified in a human family which exhibits an almost identical phenotype to Ragdoll cats with severe left ventricular wall hypertrophy, arrhythmia, congestive heart failure and sudden cardiac death in homozygotes and only mild and very late expression in heterozygous carriers (Ripoll Vera et al., 2010; Borgeat et al., 2014). Despite analysis of candidate genes in a number of other pedigree breeds no other mutations have been identified to date (Meurs et al., 2009). There is also limited information regarding the influence of modifier genes on left ventricular (LV) hypertrophy (Borgeat et al., 2015) or the presence of phenocopies in the feline population.

Feline HCM closely resembles the comparable human condition with respect to many clinical and pathologic features, and therefore represents an important spontaneously occurring animal model (Kittleson et al., 1999; Maron and Fox, 2015). At the clinical level, feline HCM is a heterogeneous disease, both in terms of presentation and outcome.

Many cats present with signs of congestive heart failure, others die suddenly; however, the majority can have long survival times and die of non-cardiac causes (Rush et al., 2002). Prognostic indicators associated with an increased risk of cardiac death are similar to those identified in human patients and included arrhythmia, extreme left ventricular hypertrophy, reduced atrial and ventricular systolic function, regional wall hypokinesis and a restrictive diastolic filling pattern (Payne et al., 2013). That said, inter-species differences are evident, for instance feline HCM more frequently results in progressive heart failure or arterial thromboembolism as the major complication (Borgeat et al., 2014; Maron and Fox, 2015). The incidence of sudden unexpected death associated with feline HCM has not been fully evaluated in cats but may be lower than seen in human patients (Maron and Fox, 2015; Wilkie et al., 2015). In both species, myofiber disarray, intramural coronary artery disease, interstitial and replacement fibrosis are the hallmarks of HCM on histopathology (Liu et al., 1993; Wilkie et al., 2015). Furthermore, in both humans and cats HCM results in significant diastolic dysfunction, however, both species can develop end-stage cardiomyopathy as part of the natural course of disease, where pathologic remodeling includes left ventricular chamber dilation, wall thinning, and fibrosis (Cesta et al., 2005).

Despite progress in elucidating the genetic basis of HCM, there is less understanding of the molecular events that lead from mutation to disease phenotype (Lopes and Elliott, 2014). Development of hypertrophy is a late stage event in HCM, resulting from detrimental processes, which operate during the pre-hypertrophic stage of the disease (Cannon et al., 2015). Studies on septal myectomy tissue from human patients with HCM and from engineered rodent models indicate that the majority of sarcomeric mutations enhance myofilament Ca²⁺-sensitivity leading to a hypercontractile phenotype with energy deficiency and altered Ca²⁺ handling which act as the major common pathways promoting HCM (Huke and Knollmann, 2010; Marston, 2011; Song et al., 2013). Given the striking similarities between the human and feline disease outlined above, it is likely that similar common pathways initiate and drive a significant proportion of the HCM phenotype in cats. Although, contractile and electrophysiological properties of ventricular cardiomyocytes isolated from clinically normal cats and those exposed to pressure loading techniques have been assessed in detail over the last three decades (Kleiman and Houser, 1988; Weisser-Thomas et al., 2005), there remains a sparsity of information concerning the effect of HCM on contractile function in feline cardiomyocytes.

We therefore studied contractile and regulatory proteins in left ventricular tissue from a range of cats with and without evidence of HCM on echocardiography and/or histopathology. These included pure-bred cats such as Ragdoll with and without the homozygous R820W mutation and a number of outbred cats with unknown mutations. Feline samples were compared with myectomy samples from human patients.

We measured protein expression, troponin I and MyBP-C phosphorylation levels and isolated troponin from cat heart muscle and investigated regulation by in vitro motility assay.

We found that fundamental Ca²⁺-regulatory abnormalities in cats are virtually the same as in humans indicating this as a potentially useful animal model of hypertrophic cardiomyopathy.

This study was approved by the Royal Veterinary College (RVC) institutional ethics and welfare committee, and written owner consent for study participation was obtained. Full thickness feline myocardium was harvested from the left ventricular (LV) free wall immediately following euthanasia and placed into a pre-cooled tube on dry ice and then transferred within 30 min to a −80°C freezer or stored in liquid nitrogen at below −160°C. Samples obtained from outside the RVC were frozen at −20°C for a maximum of 2 weeks before being transferred to the RVC on dry ice for storage. Cats with HCM were euthanized either for intractable heart failure or aortic thromboembolism. Non-HCM cats were euthanized for a variety of reasons including neoplasia, acute kidney injury, and dementia. HCM was diagnosed either by echocardiography or by necropsy/histopathology or both (Figure 1). Full clinical details including treatment at the time of death were available for all feline cases in this study. Echocardiography was performed either using a GE Vivid-7 ultrasound machine with a 7.5 mHz transducer or where the cat was too unstable, a cage side exam was performed using a GE Vivid-I portable ultrasound machine (GE Systems Ltd., Hatfield, United Kingdom) with a 7.5 mHz transducer. All LV wall thickness measurements were made from 2-D recorded images. Maximal LV wall thickness in diastole was measured on the last frame before mitral valve closure on images where the mitral valve was visible and at the time point in the cardiac cycle of greatest LV internal diameter on images where the mitral valve was not visible. A leading-edge-to-trailing-edge method of measurement was used, being careful to exclude the pericardium, false tendons or papillary muscles, but including the endocardial borders (Lang et al., 2005). A diagnosis of HCM was made if the LV wall thickness exceeded 6 mm in the absence of predisposing factors such as aortic stenosis and systemic arterial hypertension (Fox et al., 1995). A record of the pattern of LV hypertrophy (global or regional) was made (Table 1). Pathological assessment consisted of gross pathology and in the majority of cases also histopathology. Briefly, a semi-quantitative scoring system was used to grade histopathologic criteria as previously described (Wilkie et al., 2015). Morphological lesions assessed included myocyte hypertrophy, myofiber disarray, fibrosis (interstitial, perivascular, replacement, subendocardial), inflammatory cell infiltrate, intramural arteriolosclerosis, myocyte degeneration, and fat infiltration. The certainty of diagnosis (HCM or non-HCM) for each sample used in the study was scored using the following criteria. ^***HCM or non-HCM heart unequivocally diagnosed by echocardiography and by pathological assessment with appropriate clinical signs. ^**HCM or non-HCM heart unequivocally diagnosed by either echocardiography or pathological assessment with appropriate clinical signs. ^*Either equivocal findings on echocardiography and or pathology or neither echocardiography nor pathology performed but the cat must have had appropriate clinical signs. Equivocal findings included, emergency cage side echocardiography which was not detailed enough to allow for accurate assessment of LV morphology and histopathological changes which were considered too minor to be diagnostic.

Human myocardial samples were obtained from patients with hypertrophic obstructive cardiomyopathy (HOCM) undergoing surgical septal myectomy for relief of left ventricular outflow tract obstruction (LVOTO). Local ethical approval was obtained from University College London Hospitals and the Brompton, Harefield, and NHLI ethics committees for collection and use of tissue samples (for clinical details, see Supplementary Table S1).

Donor hearts had no history of cardiac disease and normal ECG and ventricular function and were obtained when no suitable transplant recipient was found. Approval was granted by the Human Research Ethics Committees of both the University of Sydney (Protocol No. 2814) and St. Vincent's Hospital (Protocol No. H91/048/1a) for the collection and distribution of the human heart samples and by the NHS National Research Ethics Service, South West London REC3 (10/H0803/147) for study of the samples. Patients gave written consent with PIS approved by the relevant ethical committee. All samples are anonymised. The investigations conform to the principles of the Declaration of Helsinki. These samples have been previously described (Messer et al., 2009, 2016; Copeland et al., 2010; Bayliss et al., 2012).

The known cat HCM-causing mutations were identified using a PCR based assay to test for the mutation using DNA extracted from either blood pellets or buccal mucosal swabs. The tests were performed by Langford Diagnostic Laboratories, University of Bristol Faculty of Veterinary Medicine.

Whole tissue extracts were obtained for gel electrophoresis using T-Per Protein extraction reagent (Pierce, Thermo Scientific) including 1 μg/ml E64, chymostatin, and leupeptin protease inhibitors according to the manufacturer's protocol. The myofibrillar fraction of heart muscle was prepared by our standard protocol (Messer et al., 2007) and analyzed by SDS-PAGE (Bio-Rad Criterion Criterion TGX™ gradient gel 4–15%) stained with SYPRO Ruby protein stain or in Western blots using mouse anti-MyBP-C F5 (GE Healthcare) (1/1,000 dilution), anti-TnI clone 14G5 (Abcam), and mouse EA-53 anti-α-actinin antibody (Sigma) (1/2,000), visualized with ECL Western Blotting Detection Reagent (GE Healthcare).

The level of phosphorylation of troponin I (TnI) and MyBP-C in myofibrils and troponin was measured by phosphate affinity SDS-PAGE as shown in Supplementary Data 1 using the methods of Messer et al. (2009) for TnI and Copeland et al. (2010) for MyBP-C.

Troponin was isolated from cat heart muscle myofibrils, using an anti-cTnI monoclonal antibody affinity column as described by Messer et al. (2007). This yields pure and active troponin, which is usable for 3 days in an in vitro motility assay (IVMA). Where appropriate, troponin was dephosphorylated by treatment with shrimp alkaline phosphatase (Sigma, P9088) and phosphorylation levels were increased by treatment with protein kinase A (PKA) catalytic subunit (Sigma, P2645-400). Recombinant human cardiac troponin T was exchanged into cat heart troponin following the procedures and controls previously described (Messer et al., 2007, 2016; Bayliss et al., 2012; Memo et al., 2013).

Thin filaments were reconstituted with 10 nM rabbit skeletal or mouse cardiac muscle α-actin (labeled with TRITC phalloidin), tropomyosin (60 nM), and troponin (60 nM) to study Ca²⁺-regulation of filament motility by the quantitative in vitro motility assay (IVMA; Fraser and Marston, 1995; Messer et al., 2007). Thin filament movement over a bed of immobilized rabbit fast skeletal muscle heavy meromyosin (100 μg/ml) was compared in paired motility chambers in which troponin varied by a single factor (mutation, phosphorylation state, or treatment with drug). The temperature was set to 29°C. Filament movement was recorded and analyzed as previously described (Marston et al., 1996) yielding two parameters; the fraction of filaments moving and the speed of moving filaments. These parameters were measured over a range of Ca²⁺ concentrations to generate Ca²⁺-activation curves as shown previously (Memo et al., 2013; Papadaki et al., 2015). The data were fitted to the four-variable Hill equation to yield a value for EC₅₀ and n_H. EC₅₀-values from replicate experiments were analyzed by paired t-test since the distribution of EC₅₀ has been shown to be normal.

We sourced 18 cardiac samples from cats treated at the RVC and 7 from collaborating veterinary practices. Non-HCM samples were collected from patients euthanized for non-cardiac reasons. Samples were cooled (QMHA: −80°C, collaborating practices: −20°C) and stored at −80°C or in liquid nitrogen at below −160°C. All available data for the samples is collated in Table 1. We made a careful examination of the clinical records, particularly echocardiography and histology data, and scored the strength of HCM or non-HCM diagnosis for this cat population using the criteria described in the methods section. Examples of echocardiography and histological sections are shown in Figure 1 and additional echocardiography scans are shown in Supplementary Data 3 and in the Supplementary movies. All the samples were studied by gel electrophoresis and eight HCM cats and five non-HCM cats with the highest certainty of diagnosis were selected for further investigation by in vitro motility assay (IVMA).

The majority of the cats studied were outbred domestic shorthaired animals but we also studied samples from pedigree cats with HCM including Ragdoll, Maine Coon, British Short Hair, and Norwegian Forest breeds with established high prevalence of inherited HCM. The homozygous MYBPC3 R820W mutation was identified in one Ragdoll cat sample.

Figure 2 shows examples of SDS-PAGE of the myofibrils extracted from our collection of cat hearts in comparison with a human heart sample (NH). All the expected contractile proteins were identified in the cat samples in the appropriate proportions. Observation of the feline band patterns showed no signs of protein degradation (band streaking, high background or fragmented bands). H10L and H10RT are two left ventricular samples from the same cat with H10RT left at room temperature for 90 min post-excision before freezing. Comparison of the band patterns showed there was no indication of protein degradation after this treatment.

Since the only HCM-causing mutations so far identified in cats are in the MYBPC3 gene and because most MYBPC3 mutations associated with HCM in humans result in haploinsufficiency, we determined the quantity of MyBP-C relative to the constitutive protein α-actinin.

The average cMyBP-C/α-actinin ratio in each lane was normalized against the average value for the non-HCM cats. In the HCM cat samples we found that five samples (H6, H7, H8, H10, and H14) have significantly lower cMyBP-C content when compared to non-HCM samples (Figure 3A). In addition, H16 exhibits a marked reduction approaching significance in one-way ANOVA. The percentage decrease of cMyBP-C was around 20% in the majority of deficient samples (H6, H8, H10, and H16), but two samples had a much lower content, H7 (30%) and H14 (44%) (Table 2). Of the deficient samples 4 were domestic shorthair (DSH), one (H6) was a Bengal, and one (H16) was a Maine Coon (without the MYBPC3 A31P mutation). Interestingly, H5 (the Ragdoll cat genotyped to be homozygous for the MYBPC3 R820W mutation) did not exhibit a cMyBP-C deficiency, even when compared to H17 (a Ragdoll cat without the MYBPC3 R820W mutation) in a one-way T-test. In Western blots, none of the samples showed any lower molecular weight bands that could be truncated peptides (Supplementary Data 2).

Protein phosphorylation levels were determined by phosphate affinity SDS-PAGE. Cat cMyBP-C phosphorylation data is presented in Figure 3B. The overall phosphorylation of the non-HCM cat samples was 3.03 ± 0.37 mols Pi/mol (n = 4), which compares with 2.70 ± 0.07 mols Pi/mol reported in human donor heart muscle. In human myectomy samples, MyBP-C phosphorylation is substantially reduced (1.05 ± 0.13 mols Pi/mol; Messer et al., 2009).

Although some cat HCM samples, notably H6 and H17, appeared dephosphorylated, the majority of cat HCM samples had phosphorylation levels comparable with non-HCM cat samples. When all the cats were averaged, there was no significant overall reduction of cMyBP-C phosphorylation in the HCM samples (mean = 2.32 ± 0.95 mols Pi/mol, n = 17, p = 0.22) in contrast to the human myectomy samples.

Phosphorylation assessments of feline troponin I (TnI) showed similar results to MyBP-C. Certain samples appeared dephosphorylated (H2 and H10), but there was no overall significant difference between the normal and hypertrophic groups (Non-HCM: mean = 1.49 ± 0.17 mols Pi/mol TnI, n = 8; HCM: mean = 1.46 ± 0.12 mols Pi/mol TnI, n = 18, p = 0.874; Figure 3C). This is very different from human heart where donor heart muscle has a phosphorylation level similar to non-HCM cat (1.62 ± 0.06 mols Pi/mol TnI) but myectomy samples have a very low level of phosphorylation (0.26 ± 0.02 mols Pi/mol TnI) (Messer et al., 2009; Bayliss et al., 2012). The low levels of TnI or MyBP-C phosphorylation in some of the cat samples may be a consequence of the post-mortem treatment of the cat hearts although we did not detect any correlation between the time of storage at −20°C prior to −80°C freezing and the level of TnI or MyBP-C phosphorylation.

To study Ca²⁺ regulation of contraction, we extracted troponin from cat hearts using the protocols we have developed for human heart troponin. Yields were comparable to human and the troponin was fully competent at regulating thin filaments in the quantitative in vitro motility assay (IVMA). We investigated the Ragdoll (MYBPC3 R820W mutation) (H5) cat heart troponin in detail and also studied troponin from 7 other cats that had a high confidence of HCM diagnosis (see Table 1).

Figure 4A shows an example of the Ca²⁺ titration curves of HCM cat H5 compared to a non-HCM cat, N4; both had a naturally high level of troponin phosphorylation (Figure 3C). The left shifted Ca²⁺ curves of both sliding speed and fraction motile and the calculated lower EC₅₀ indicates that the HCM cat has a 1.5–2.0 fold higher Ca²⁺-sensitivity than non-HCM cat troponin. This was confirmed with a total of 6 H5 troponin preparations, compared with 2 different non-HCM troponins (mean non-HCM EC₅₀ = 0.094 ± 0.005 μM, mean H5 EC₅₀ = 0.045 ± 0.004, mean EC₅₀ non-HCM/EC₅₀ H5 = 2.19 ± 0.27, p = 0.008).

Ca²⁺-sensitivity of HCM cat troponin was tested in seven further cats, including two that showed MyBP-C haploinsufficiency (H7 and H14), all of which showed a higher Ca²⁺-sensitivity than non-HCM troponin in paired (non-HCM + HCM) assays combining fraction motile and sliding speed at similarly high levels of troponin phosphorylation (mean non-HCM EC₅₀ = 0.090 ± 0.005 μM, mean all HCM troponin EC₅₀ = 0.046 ± 0.003, mean EC₅₀ non-HCM/EC₅₀ HCM = 2.05 ± 0.13, n = 7, p < 0.0001), see Figure 5 and Table 3. There were no significant differences in n_H and maximum sliding speed between non-HCM and HCM cat troponin.

In man and mouse, the Ca²⁺-sensitivity of normal heart muscle troponin is modulated by TnI phosphorylation at Ser 22 and 23 by PKA. This was also true for the non-HCM cat heart samples. Phosphorylation levels were manipulated by dephosphorylating highly phosphorylated troponin and treating unphosphorylated troponin with PKA+ATP. Figure 6A shows the Ca²⁺-activation curve for a typical non-HCM cat troponin in the phosphorylated and unphosphorylated states; there is an ~2-fold increase in EC₅₀ upon phosphorylation. The same pattern of results was found with all five cat hearts tested (see Figure 7 and Table 2).

We have frequently found that the Ca²⁺-sensitivity of troponin extracted from the heart muscle of patients with HCM obtained from myectomy operations is not modulated by phosphorylation (Supplementary Table S3; Bayliss et al., 2012; Messer et al., 2016). This also occurred in our HCM cat troponin samples; Figure 6B illustrates this for cat H5. The insensitivity of Ca²⁺-sensitivity to the level of TnI phosphorylation was found for every HCM cat troponin, irrespective of the causative mutation (Figure 7 and Table 3).

In studies using human and mouse heart muscle we found that coupling could be restored to HCM thin filaments by adding small molecules such as Epigallocatechin-3-gallate (EGCG) and Silybin (Bayliss et al., 2012; Papadaki et al., 2015; Messer et al., 2016). We tested whether recoupling could be observed in cat heart troponin.

When 100 μM EGCG was added to troponin from H5 cat heart it restored coupling in reconstituted thin filaments measured by IVMA (Figure 8A). EGCG did not affect the EC₅₀ of unphosphorylated H5 troponin but it increased the EC₅₀ of phosphorylated H5 troponin, thus restoring EC₅₀ to the values present in the non-HCM cat. The change in EC₅₀ on phosphorylation in the presence of EGCG was 2.21 ± 0.16 fold in four separate assays with H5. EGCG also recoupled troponin from H13 and H14 cat hearts by a similar amount (see Table 3). Figure 8B shows the restoration of EC₅₀ to normal by EGCG with H13.

In previous studies of HCM in human myectomy samples we observed that it was possible to restore coupling by replacing troponin T (TnT) with wild-type (human) recombinant TnT. We tested whether this also occurred with HCM cat troponin. When we replaced TnT in troponin from the Ragdoll MYBPC3 R820W H5 cat heart with wild-type (human) recombinant TnT by an exchange reaction we found that the modulation of Ca²⁺-sensitivity by TnI phosphorylation was restored (Supplement S3 and Table 3). The change in EC₅₀ on phosphorylation was 1.61 ± 0.01 fold, comparable with the same experiment in HCM human heart troponin (1.85 ± 0.13 fold) and with the phosphorylation effect on EC₅₀ in non-HCM cat (Figure 6).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.