The Genetic Aspects of Hypertrophic Cardiomyopathy in Cats

Transcription

1 The Genetic Aspects of Hypertrophic Cardiomyopathy in Cats Word count: Tom Schipper Student number: Supervisor: Prof. dr. Luc Peelman Supervisor: Dr. Mario van Poucke A dissertation submitted to Ghent University in partial fulfilment of the requirements for the degree of Master of Veterinary Medicine Academic year:

2 Ghent University, its employees and/or students, give no warranty that the information provided in this thesis is accurate or exhaustive, nor that the content of this thesis will not constitute or result in any infringement of third-party rights. Ghent University, its employees and/or students do not accept any liability or responsibility for any use which may be made of the content or information given in the thesis, nor for any reliance which may be placed on any advice or information provided in this thesis.

3 Acknowledgements I could not have carried out this study without the help of a large number of people. To thank them all in an orderly manner, I will group them according to the parts of the work that they contributed to. Please note that this order does not necessarily reflect the importance of their contributions to this study. Prof. dr. Luc Peelman is thanked for enthusiastically turning my suggestion for a subject into an interesting and feasible study. He is also thanked for his stimulation and guidance during the research process and his help with the writing. Dr. Mario van Poucke is thanked for his extensive help in the research process. He taught me the vast majority of the laboratory techniques applied in this study, as well as the computer processes of primer design and data analysis. He was always ready to answer any question, but also stimulated me to work independently. Linda Impe is thanked for introducing me in the laboratory and teaching me the basic techniques. She was a great help during the research process by providing any materials, advice and help I needed. Dominique Vander Donckt, Laice Royeras, Evy Beckers, Xiaoyuan Lin and Sara Lieveld are also thanked for their help with technical issues in the laboratory. Ruben Van Gansbeke is thanked for taking care of the administrative tasks associated with the laboratory work. Prof. dr. Bart Broeckx is thanked for his help with the interpretation of genomic databases and for turning this study into a PhD research proposal. All students and personnel at the laboratory of animal genetics and the laboratory of gene therapy are thanked for providing a pleasant working atmosphere. Dr. Leen Van Brantegem, dr. Veronique Saey, Norbert van de Velde and Laurien Sonck are thanked for their help in collecting heart tissue samples. Laurien Sonck is especially thanked for notifying me about interesting patients, providing histological images and answering many questions. All students that helped me at the laboratory of pathology are also thanked for their contributions. Prof. dr. Richard Ducatelle is thanked for his advice on selecting suitable patients. Prof dr. Pascale Smets is thanked for her crucial help in the final selection of suitable patients and for enabling me to take even more suitable samples from her own patients. Together with Dominique van Oort, she is thanked for providing images. The veterinarians and technicians of the general practice clinic where I did my externship are thanked for their help in sample collection. All owners who participated in this study are thanked for their contributions. Matthieu Salamone and Andy Vervaet are thanked for providing transport for sample collection. Finally, Margo Gabriëls and Wietske Dijkstra are thanked for their indispensable support and advice.

4 Table of contents List of abbreviations... 6 Summary... Error! Bookmark not defined. Samenvatting... Error! Bookmark not defined. 1. Introduction The clinical features of HCM Epidemiology Pathophysiology Clinical presentation Diagnosis Treatment Prognosis The genetics of HCM The genetics of human HCM The structure of the sarcomere The MYBPC3 gene The MYH7 gene Other sarcomeric genes Non-sarcomeric genes The genetics of feline HCM Phenocopies of HCM Pathogenesis Genotype-phenotype correlations Gene dosage Modifier genes Environmental influences Aims of this study Animals, materials and methods Patient selection DNA isolation from blood samples Spectrophotometric quality control qpcr genotyping and PCR amplification Agarose gel electrophoresis PCR product clean-up Sanger sequencing DNA isolation from tissue and qpcr genotyping RNA isolation Quality control of isolated RNA Reverse transcription... 40

5 5.12 PCR amplification of cdna Sequencing of cdna Data analysis Results Study population Transcript variant expression Variants in MYBPC3 and MYH Assessment of the pathogenicity of the found variants Discussion Transcript variants Pathogenicity of the found variants Implications of these findings Future perspectives References... 52

6 List of abbreviations α-myhc α-myosin heavy chain (protein) β-myhc β-myosin heavy chain (protein) ACE angiotensin converting enzyme (protein) ACE1 angiotensin converting enzyme (gene) ACTC1 cardiac α-actin (gene) ACTN2 α-actinin (gene) ATE arterial thrombo-embolism CHF congestive heart failure cmybp-c cardiac myosin-binding protein c (protein) CSRP3 muscle LIM protein (gene) DCM dilated cardiomyopathy ECG electrocardiography ELCv ventricular essential myosin light chain (protein) Exo-AP exonuclease Antarctic phosphatase HCM hypertrophic cardiomyopathy LVH left ventricular hypertrophy LVNC left ventricular non-compaction LVOT left ventricular outflow tract MYBPC3 cardiac myosin-binding protein c (gene) MYH6 α-myosin heavy chain (gene) MYH7 β-myosin heavy chain (gene) MYL2 ventricular regulatory myosin light chain (gene) MYL3 ventricular essential myosin light chain (gene) MYOZ2 myozenin 2 (gene) PCR polymerase chain reaction RCM restrictive cardiomyopathy RLVc ventricular regulatory myosin light chain (protein) SAM systolic anterior motion of the mitral valve TNNC1 cardiac troponin C (gene) TNNI3 cardiac troponin I (gene) TNNT2 cardiac troponin T (gene) TPM1 α-tropomyosin (gene)

7 Summary Hypertrophic cardiomyopathy (HCM) is a highly prevalent heart disease in cats. It is often occult, but poses a high risk of fatal complications. HCM in humans is a very similar disease with a genetic etiology. A large number of variants in sarcomeric genes, especially MYBPC3 and MYH7, have been associated with human HCM. Feline HCM is thought to have a similar etiology and two causative variants have been identified in MYBPC3 in the Maine Coon and Ragdoll breeds. Nevertheless, the exact etiology of feline HCM remains unknown in the majority of the cats. This study aims to identify other genetic variants that cause feline HCM. Twelve blood samples and 14 heart tissue samples from HCM-affected cats negative for the known causative variants were collected. Genomic DNA from blood and RNA from tissue converted to complementary DNA were used for Sanger sequencing of the coding regions of the MYBPC3 gene. The MYH7 gene was sequenced similarly from nine tissue samples. The pathogenicity of the found variants was evaluated by in silico prediction of amino acid substitutions, assessment of evolutionary conservation of the affected amino acids and comparison to the literature on human variants. Twenty-seven variants were identified in MYBPC3 and 24 in MYH7. None of these variants was considered to be pathogenic. The etiology of feline HCM seems to differ from that of human HCM. Evaluation of other genes is warranted for the identification of additional causative variants. Genome- and exome-wide techniques are the most suitable techniques for such studies. Samenvatting Hypertrofische cardiomyopathie (HCM) is een veel voorkomend hartaandoening bij de kat. Het is in veel gevallen occult, maar geeft een hoog risico op levensbedreigende complicaties. HCM bij de mens is een gelijkaardige ziekte met een genetische etiologie. Een groot aantal genetische varianten in de genen van de sarcomeren, met name MYBPC3 en MYH7, zijn geassocieerd met HCM bij de mens. HCM bij de kat heeft vermoedelijk een gelijkaardige etiologie en twee oorzakelijke varianten zijn gevonden in MYBPC3 bij de Maine coon en de ragdoll. Desondanks blijft de precieze etiologie van HCM onopgehelderd bij de meeste katten. Het doel van dit onderzoek is om andere varianten te vinden die HCM kunnen veroorzaken bij de kat. Twaalf bloedstalen en 24 hartweefselstalen werden verzameld van katten met HCM die negatief waren voor de bekende oorzakelijke varianten. Genomisch DNA uit bloed en complement DNA op basis van RNA uit weefsel werden gebruikt voor Sanger sequencing van de coderende regionen van het MYBPC3 gen. Sequencing van het MYH7 gen werd op dezelfde wijze uitgevoerd op 9 weefselstalen. De schadelijkheid van de gevonden varianten werd beoordeeld op basis van in silico voorspellingen van aminozuursubstituties, beoordeling van de evolutionaire conservering van de betrokken aminozuren en vergelijkingen met de literatuur over varianten bij de mens. Zevenentwintig varianten werden gevonden in het MYBPC3 gen en 24 in het MYH7 gen. Geen van deze varianten werd als schadelijk beschouwd. De etiologie van HCM bij de kat lijkt te verschillen van die bij de mens. Onderzoek van andere genen is aangewezen om meer oorzakelijke varianten te vinden. Dit onderzoek kan waarschijnlijk het best uitgevoerd worden met technieken die het gehele genoom of exoom onderzoeken. 7

8 1. Introduction Hypertrophic cardiomyopathy (HCM) is defined as hypertrophy of the left ventricular myocardium that cannot be attributed to an underlying disease or anomaly. HCM is a well-known disease in humans, but is especially common in cats. It is an often occult disease that can result in potentially fatal complications. The treatment options in cats are limited and symptomatic cats have a grave prognosis. This underlines the importance of knowledge of the etiology that can be translated into preventive measures. HCM is attributed to an underlying genetic defect in humans and it is generally accepted that this is also the case in cats. The disease-causing genetic variant, however, is usually not known in cats, as only two, breedspecific variants have been identified. A broader knowledge of the genetic aspect of HCM in cats would allow more extensive genetic screening of breeding animals. Breeders can then limit the spread of disease-causing variants in the feline population and reduce the incidence of HCM. This study aims to identify new diseasecausing variants in affected cats by analyzing candidate genes that were selected because of their importance in human HCM. 2. The clinical features of HCM 2.1 Epidemiology Among all feline heart diseases, HCM is by far the most common. In a retrospective study in two Swiss clinics, 47.5% of the cats with symptomatic primary or secondary cardiac disease was diagnosed with HCM (Riesen et al., 2007a). HCM is also highly prevalent in asymptomatic cats. Two prospective echocardiographic screenings of apparently healthy cats, one of American client-owned cats and one of British shelter cats, resulted in very similar prevalence rates of 14.6% and 14.7%, respectively (Paige et al., 2009; Payne et al., 2015b). The prevalence of the disease is age-dependent. The prevalence is 4.3% in juvenile (6-12 months) cats, while it is 29.4% in senior ( 9 years) cats (Payne et al., 2015b; Luis Fuentes and Wilkie, 2017). The mean age of diagnosis is 6 years in the general population (Riesen et al., 2007a; Abbot, 2010; Payne et al., 2015b). In certain breeds, however, younger ages are reported. Studies in Sphynx cats diagnosed HCM at median and mean ages ranging from years (Riesen et al., 2007b; Silverman et al., 2012; Trehiou-Sechi et al., 2012). Similarly, these ages were years in Maine Coons (Riesen et al., 2007b; Mary et al., 2010; Trehiou- Sechi et al., 2012) and a median age of 2.7 years was found in British Shorthairs (Granström et al., 2011). Case reports have described severe HCM in Ragdolls of 5 to 24 months of age (Lefbom et al., 2001). On the other hand, Chartreux and Persian cats have a median age of diagnosis that is similar to or higher than that of Domestic Shorthairs (Trehiou-Sechi et al., 2012). These findings might indicate that some breeds are predisposed to an early onset of HCM. Yet, it must be emphasized that these studies all included, and some even were limited to, cats presented for echocardiorgaphic screening. Asymptomatic purebred cats are often presented at young age to screen for breeding purposes, while Domestic Shorthairs are usually referred at the time that they have developed abnormal heart sounds or overt symptoms (Hägströmm et al., 2015). Furthermore, these figures are based on small study populations, as none of the studies involved more than 28 affected cats from a single breed. These biases might explain some, or possibly all, of the discrepancies in age of diagnosis. HCM is more common in male than in female cats. In large studies, the proportion of affected males ranges from 62.6% to 79.2% (Rush et al., 2002; Payne et al., 2015b). This male predisposition is found in the general population as well as within purebred families (Côté et al., 2011b). Remarkably, a study of British Shorthairs found a prevalence of 20% in males, while this was only 2.3% in females (Granström et al., 2011). Some suggest that HCM is equally common in both sexes, but that it manifests itself at a younger age and with more severe hypertrophy in males (Riesen et al., 2007b). Most cats affected by HCM are Domestic Shorthairs (Côté et al., 2011b). In retrospective studies, 68.8 to 86.9 percent of the HCM patients is classified as Domestic Shorthair or Domestic Longhair (Rush et al., 2002; Ferasin et al., 2003; Trehiou-Sechi et al., 2012), though this is only 38.1% in the study by Riesen et al. (2007a). Nevertheless, HCM is widely perceived as a disease of purebred cats. Screening studies in Maine Coons resulted in prevalence ranging from 5.6% (Riesen et al., 2007b) to 10.4% (Carlos Sampedrano et al., 2009). These figures might be underestimations, as these studies had a large proportion of females and included mostly young cats. Furthermore, studies with low prevalence rates report 8

9 a considerable number of equivocal findings, usually in young cats (Riesen et al., 2007b; Gundler et al., 2008). For Ragdoll cats, a prevalence of 18.5% could be deduced from a study by Borgeat et al. (2015), though this study was not designed for prevalence screening. In British Shorthairs, a prevalence of 8.5% was found in a young and 65% female study population (Granström et al., 2011). A prevalence of 25% in Norwegian Forest cats was reported by März et al. (2015), while this was 11.4% in the study by Longeri et al., (2013). The less strict diagnostic criteria applied by März et al. are the most likely explanation for this disparity. In Sphynx cats, the prevalence is 20.2% (Chetboul et al., 2012) to 30.0% (Riesen et al., 2007b). Together with the early diagnosis in these breeds, these studies show that Ragdolls and Sphynx cats are predisposed for HCM. The possible biases in study design make it hard to conclude whether Maine Coons, British Shorthairs and Norwegian Forest cats are predisposed for HCM, compared to Domestic Shorthairs. Nevertheless, these studies still show that HCM is an important disease in these breeds. Breed predispositions are also suggested by the high proportion of purebred cats presented with HCM in general and referral practice (Smith and Dukes-McEwan, 2012). However, this might also represent a higher awareness of HCM and a higher readiness for echocardiographic examination among owners of purebred cats. 2.2 Pathophysiology Hearts affected by HCM usually have an equal or even stronger contractility than normal hearts. The main functional abnormality in feline cardiomyopathies is therefore not systolic, but diastolic dysfunction (Fox and Schober, 2015). The diastole is the period between the closing of the aortic valve and the closing of the mitral valve, during which the ventricle is filled with blood (Chung et al., 2015). It can be conceptually divided into relaxation and compliance. Relaxation is the active diminution of the contractile force built up during the systole. In cats with HCM, it is impeded by pathological changes in contractile proteins, Ca 2+ -handling and -sensitivity and energetics, as will be discussed below. The energy shortage can be worsened by myocardial ischemia, a known complication of HCM in humans that is likely also present in cats (Côté et al., 2011b; Biasato et al., 2015). The efficiency of the relaxation is further reduced by the abnormal orientation of myofibers and cardiomyocytes and the asynchronity of relaxation in a hypertrophied heart (Côté et al., 2011b). Compliance is the passive component of diastolic function. Hypertrophy in itself reduces the compliance on both the level of the individual cardiomyocyte and the ventricle as a whole (Segers and De Keulenaer, 2013). An increase in collagen (fibrosis) due to ischemia makes the ventricle more stiff (Luis Fuentes, 2002; Kershaw et al., 2015; Khor et al., 2015). Finally, the muscular hypertrophy can markedly decrease the volume of the ventricular cavity, compromising the diastolic filling (Li et al., 2015). Systolic anterior motion of the mitral valve (SAM) is a common complication of HCM. Possible causes are hypertrophy of the papillary muscles, abnormal hemodynamic patterns and abnormal valve morphology (Abbot, 2010; Schober et al., 2010; Teo et al., 2015). In SAM, the tip of the anterior valvular leaflet bends into the left ventricular outflow tract (LVOT). Together with septal hypertrophy, this obstructs the LVOT, leading to an increased work load and further damage of the ventricular muscle. Simultaneously, the bended leaflet leaves the valve insufficiently closed, resulting in regurgitation of blood to the left atrium (Li et al., 2015). Diastolic dysfunction hinders the filling of the ventricle. The stiff ventricle will only fill at a higher pressure in the left atrium (Côté et al., 2011b). Mitral regurgitation can increase the atrial pressure even more. Under the elevated pressure, the left atrium will dilate (figure 1B). An increased atrial pressure will be communicated further backwards to the pulmonary veins and capillaries. When this pressure reaches approximately 24 mmhg, fluid will seep into the lungs and cause lung edema. This condition is congestive heart failure (CHF). In contrast to dogs, left-sided CHF in cats is often accompanied by pleural effusion. The mechanism that causes this effusion is not yet known (Côté et al., 2011b). A dilated atrium predisposes for the formation of thrombi that can cause arterial thrombo-embolism (ATE). Three factors are suspected to contribute to thrombus formation. The first is stasis of blood in the dilated atrium. The second are abnormal anti-thrombotic activity and damage to the endocardium, caused by high atrial pressure and dilatation (Luis Fuentes, 2012). A third factor, hypercoagulability, might be attributed to increased fibrinogen concentrations and Factor VIII activity in cats with cardiomyopathy (Stokol et al., 2008). The atrial thrombus or a part that broke off from it can enter the circulation via the ventricle. It will eventually get stuck somewhere in the arterial system, most commonly at the aortic trifurcation. The embolization and 9

10 concurrent vasoconstriction of collateral arteries cause ischemia of the distal tissues (Luis Fuentes, 2012). Hearts affected by HCM are prone to ventricular arrhythmias. Fibrosis can slow or unidirectionally block the electrophysiological conduction and may facilitate premature complexes (Nguyen et al., 2014). Arrhythmias can lead to a sudden change in heart rate and cardiac output. This change can result in cerebral hypoperfusion, observable as a syncope. The most severe sequela, however, is ventricular fibrillation. This chaotic myocardial activity does not result in effective blood flow and is lethal unless immediate intervention (Côté et al., 2011d). As this usually happens without a warning sign, it is described as sudden death. LV LA A B Figure 1. Pathology of HCM. (A) A heart with severe HCM, with two exceptionally large thrombi extracted from the dilated left atrium. (B) Two-dimensional echocardiographic image of a cat with HCM. The left ventricular lumen is narrowed by hypertrophy of the interventricular septum and left ventricular free wall. The left atrium is dilated. Courtesy of Dominque van Oort (A) and prof. dr. Pascale Smets (B), Cardiology Service, Department of Medicine and Clinical Biology of Small Animals, Ghent University. 2.3 Clinical presentation The clinical presentation of cats with HCM can be divided into four groups: asymptomatic, CHF, ATE and syncope or sudden death (Rush et al., 2002). The symptomatic categories are not mutually exclusive and cats presented with ATE often have concomitant CHF (Borgeat et al., 2014b). Although ATE, CHF and sudden death occur in an advanced stage of HCM after considerable pathological remodeling of the heart, few cats are diagnosed during these processes for lack of symptoms. Many affected cats are therefore only diagnosed after they have developed severe complications (Côté et al., 2011b). Asymptomatic cats might have an abnormal auscultation. SAM often causes a systolic heart murmur, as both LVOT obstruction and mitral regurgitation result in turbulence of blood. Other malformations of the mitral valve can also cause a murmur (Abbot, 2010). Nevertheless, not all cats with HCM have a heart murmur. The reported prevalence of heart murmurs in HCM-positive cats ranges from 31.3% (Paige et al., 2009) to 96.4% (Granström et al., 2011). Furthermore, benign murmurs without clinical significance are also very common in cats, so this is not a specific sign of HCM (Côté et al., 2011b). The reported specificity of heart murmur detection for HCM also varies from 42.6% (Wagner et al., 2010) to 87.4% (Paige et al., 2009). A gallop sound, characterized by a third and sometimes fourth heart tone, and can be caused by abnormal filling due to diastolic dysfunction (Côté et al., 2011b). This makes it a somewhat more specific sign of HCM, although a gallop sound is also heard in cats with other cardiomyopathies (Ferasin et al., 2003). A gallop sound is less common than a heart murmur and detected in only 2.4% of the asymptomatic cats with HCM (Payne et al., 2015b). It is more common in referral populations and one study found a gallop sound in 33.4% of the cats with HCM (Rush et al., 2002). If a cat with HCM develops arrhythmia, this can be heard as an irregular rhythm. Arrhythmia is generally less reported in cats with HCM than a gallop sound (Rush et al., 2002; Ferasin et al., 2003; Trehiou-Sechi et al., 2012; Payne et al., 2015a). CHF is the most common complication of HCM (Trehiou-Sechi et al., 2012; Spalla et al., 2016). CHF results in dyspnea, visible as heavy breathing, often with open mouth, at an increased frequency. Coughing, a common sign of lung oedema in dogs, is rarely seen in cats (Abbot, 2010). Severe CHF can lead to cyanosis of the mucosae (Payne et al., 2015a). ATE is reported to occur in 12% of the HCM-affected cats, but this might be an overestimation that does not 10

11 take into account the large number of undiagnosed, asymptomatic cats with HCM (Luis Fuentes, 2012). In more than 75% of the cases, both hind limbs are affected due to ATE of the distal aorta (Luis Fuentes, 2012; Borgeat et al., 2014b). This results in a distinctive clinical presentation with paralyzed, cold and extremely painful hind limbs that have pale or cyanotic foot pads. The femoral pulse is absent and the rectal temperature is often decreased. Similar symptoms occur in embolization of a single femoral or brachial artery. Embolization of other arteries results in less typical symptoms that will depend on the affected organ (Luis Fuentes, 2012). A syncope is a sudden and short loss of consciousness. It is witnessed by the owner in 0.9% (Trehiou-Sechi) to 5.1% (Payne et al., 2015a) of the cases. The incidence of sudden deaths in cats with HCM is not known, but is likely underreported. A two year follow-up study reported an incidence of 4.7%, suggesting that sudden death is actually a common complication of HCM (Payne et al., 2015a). 2.4 Diagnosis It is not possible to diagnose HCM by physical examination, although a clinical presentation of ATE is very suggestive. Several diagnostic techniques are available to further investigate a patient suspected of HCM Ante mortem diagnosis Radiography of the thorax is essential to evaluate the extent of lung oedema and pleural effusion in a cat in CHF, but is less applicable to diagnose HCM (Côté et al., 2011b). The hypertrophy of the left ventricle is not always visible as cardiomegaly and the size of the heart appears normal in 10-23% of the patients presented with HCM in referral clinics (Rush et al., 2002; Ferasin et al., 2003). Furthermore, cardiomegaly is a common finding in other cardiac diseases and therefore not specific for HCM (Ferasin et al., 2003; Riesen et al., 2007a). An image that is specifically associated with HCM is a "valentine" shaped heart on dorsoventral view, caused by dilatation of the left atrium. However, other cardiac diseases can also result in the same appearance and only a subset of the patients with a valentine shape hearts has HCM (Winter et al., 2015). Two biomarkers in the blood are used to assess cardiac pathology in cats. NT-proBNP (N-terminal pro-b-type natriuretic peptide) is a peptide produced by the atria as a reaction to pressure overload. Its main use is to differentiate cardiac causes of dyspnea from pulmonary causes. It has a high sensitivity in detecting severe HCM, but usually does not detect mild or moderate HCM (Hsu et al., 2009). Moreover, it detects cardiac disease in general, of which HCM is only one form (Machen et al., 2014). ctni (Cardiac troponin I) is a sarcomeric protein in the myocardium that is released into the bloodstream in case of myocardial damage. In HCM, this damage is thought to result from chronic ischemia (Langhorn et al., 2016). Nevertheless, elevated ctni levels also occur in other cardiac diseases than HCM (Wells et al., 2014). The main indication for electrocardiography is to evaluate arrhythmia. As with other secondary effects of HCM, arrhythmia is only present in a subset of the cats affected by HCM and can also be caused by other cardiac or systemic diseases. Left ventricular hypertrophy is also detectable on electrocardiography as an increased duration and amplitude of the QRS-complex (Côté et al., 2011b). This is not a sensitive test, as signs of left ventricular enlargement are seen in less than 25% of the cats with HCM (Ferasin et al., 2003; Riesen et al., 2007a). The limitations of other diagnostic techniques leave echocardiography as the gold standard to diagnose HCM in cats. The main diagnostic criterion for HCM is the thickness of the myocardium of the left ventricle. This thickness is measured by 2-dimensional echocardiography at both the interventricular septum and the left ventricular free wall. The measurement is made during the diastole, when the myocardium is at its thinnest (Côté et al., 2011b). The upper limit of myocardial thickness is now generally set at 6.0 mm for cats. However, limits of 5.5 and 5.0 mm have also been proposed and used. In addition, there is disagreement over whether the diagnosis of HCM requires an increased thickness in just one part of the interventricular septum or left ventricular free wall, or in 50% of a wall segment. Such differences have a large effect on the proportion of cats diagnosed with HCM and hamper the comparison of studies using different reference ranges (Gundler et al., 2008; Wagner et al., 2010; Côté et al., 2011b). Echocardiography can also be used to evaluate diastolic function by transmitral pulsed-wave spectral Doppler that measures the blood flow through the mitral valve. Diastolic filling occurs in two peaks, the early filling during myocardial relaxation and the atrial filling caused by atrial contraction. These phases are visualized as an E-wave and A-wave, respectively. The E/A ratio decreases in mild diastolic dysfunction, while it 11

12 pseudonormalizes in moderate diastolic dysfunction and increases in severe diastolic dysfunction. Additional measurements of pulmonary venous flow are necessary to distinguish the pseudonormal pattern from a normal pattern (Côté et al., 2011b; Schober and Chetboul, 2015). All three abnormal filling patterns are common in cats with HCM, while a normal pattern is seen in only 6% of these cats (Payne et al., 2013). Another application of echocardiography is to identify secondary pathological changes. Left atrial enlargement can be assessed precisely by calculating the LA/Ao ratio, the ratio of the diameter of the atrium to that of the aorta (figure 2A). The normal LA/Ao ratio is <1.5. Spontaneous contrast, a sign that blood is in the process of clotting, or a thrombus that has already formed can be detected in the enlarged atrium (Côté et al., 2011b). SAM can be visualized as the movement of the mitral valve itself in normal 2-dimensional echocardiography, but also as mitral regurgitation, turbulence and increased outflow velocity due to obstruction in Doppler mode (figure 2B) (Schober et al., 2010). Ao LA A B Figure 2. Echocardiographic evaluation of secondary pathological changes in HCM. (A) Two-dimensional echocardiographic measurement of the diameters of the aorta and the dilated left atrium to calculate the LA/Ao ratio. (B) Assessment of SAM in Doppler mode. Mitral regurgitation is colored blue and the aortic outflow is colored red. Turbulent blood is colored green. Courtesy of prof. dr. Pascale Smets, Cardiology Service, Department of Medicine and Clinical Biology of Small Animals, Ghent University. A more advanced imaging technique is cardiac MRI. This allows a visualization of all parts of the myocardium, detection of fibrosis and an accurate estimation of the myocardial mass and volume. This technique is increasingly used in human medicine, but the high cost of the equipment makes it application in veterinary medicine not feasible (Côté et al., 2011b) Post mortem diagnosis A post mortem diagnosis can be based on macroscopic and microscopic features of the heart by necropsy and histopathology, respectively. A first parameter is the subjective impression of the heart. The left ventricle and/or atrium may seem enlarged, the atrium may contain a thrombus and the ventricle may display pale zones of fibrosis or feel stiffer (figure 3A and 3B). Amputation of the apex (figure 3C) allows for measurement of the interventricular septum and the left and right ventricular free walls (Robinson and Robinson, 2016). Because the myocardium contracts post mortem, the limits used in echocardiography during the diastole cannot be applied here (Kittleson et al., 1999). No absolute limits have been defined in the literature for myocardial thickness on necropsy, but the left ventricle is considered hypertrophic when the combined thickness of the interventricular septum and left ventricular free wall is more than four times the thickness of the right ventricular free wall (Robinson and Robinson, 2016). Another parameter is the mass of the heart. The normal mass of the feline heart is less than 20 gram, while hearts affected by HCM tend to weigh more than 20 gram (Kershaw et al., 2012). Nevertheless, there is considerable variation in cardiac mass, resulting in overlap between healthy and affected cats (Wilkie et al., 2015). To correct for the difference in body mass between cats, the mass of the heart can be expressed as the relative cardiac mass in grams heart per kilograms body mass. The normal relative cardiac mass is 4.0 g/kg (Wilkie et al., 2015). While the diagnosis based on 12

13 macroscopic features is confident in severe cases of HCM, such as shown in figure 3, it can be equivocal in milder cases. In these case, information from ante mortem examinations and the circumstances of death together with histopathology can be used to confirm or refute a diagnosis of HCM. C A B Figure 3. Macroscopic features of severe HCM on necropsy. (A) Generally enlarged and deformed appearance of the heart. (B) Opening of the dilated left atrium reveals a large thrombus. (C) Cross-section view of the ventricles after amputation of the apex. Both the interventricular septum and the left ventricular free wall are hypertrophic, narrowing the lumen of the left ventricle. This heart weighed 40 grams after removal of the thrombi. Courtesy of Dominique van Oort, Cardiology Service, Department of Medicine and Clinical Biology of Small Animals, Ghent University The cardinal features of HCM on histopathology are hypertrophy and disarray of the cardiomyocytes. The cardiomyocytes are larger and may display signs of degeneration. They are arranged in a disordered manner, instead of the normal parallel arrangement, as shown in figure 4 (Robinson and Robinson, 2016; Miller and Gal, 2017). Fibrosis can be seen in the interstitium and in zones of necrosis due to ischemia. Ischemic zones are also characterized by arteriosclerosis, where the small blood vessels are narrower due to smooth muscle and connective tissue proliferation in their walls (Biasato et al., 2015; Miller and Gal., 2017). Figure 4. Histological image of the myocardium of a cat with HCM. The cardiomyocytes show a marked disarray in the center of the image. Hematoxylin and Eosin stain, 20x magnification. Courtesy of Laurien Sonck, Laboratory of Pathology, Department of Pathology, Bacteriology and Avian diseases, Ghent University Differential diagnosis The identification of hypertrophy of the left ventricular myocardium does not yet mean a definitive diagnosis of HCM. There are several other diseases that can increase the mass of the myocardium and these are grouped under the term left ventricular hypertrophy (LVH). It is not possible to distinguish HCM from all forms of LVH by echocardiography alone, but there can be some indications. Severe hypertrophy (>7.0 mm) is usually only present in HCM. SAM is also far less common or even nonexistent in LVH and therefore very suggestive of HCM (Abbot, 2010; Côté et al., 2011b). Nevertheless, further examinations are necessary to exclude LVH, especially in older cats. 13

14 A very common form of LVH is caused by hyperthyroidism, the most common endocrine disease in cats, especially older cats. Measurement of total serum thyroxine is a readily available diagnostic test that is often conclusive. The enlarged thyroid gland is palpable in over 90% of the hyperthyroid cats, but also in some euthyroid cats (Peterson, 2013). Enlarged thyroid glands can also be detected on necropsy. A second common form of LVH results from systemic hypertension. Hypertension also affects mostly older cats and is usually secondary to chronic kidney disease. The diagnosis can be made by Doppler sphygmomanometry, where a pressure 170 mmhg means hypertension, while mmhg is equivocal (Taylor et al., 2017). There are no strict pathologic criteria to diagnose hypertension post mortem. Hypertension can be suspected in cats with target organ damage to the brain, eye and/or kidney (Kohnken et al., 2017). Excluding hypertension also excludes most cases of hyperaldosteronism, an endocrine disease that can cause severe hypertension, but also directly affects the heart. This disease is relatively rare and typically also causes severe muscle weakness (Feldman, 2015). A last endocrine disorder that can cause LVH is acromegaly. The hypertrophy can be severe and SAM can be present (Myers et al., 2014). Acromegaly is rare and is usually accompanied by symptoms of diabetes mellitus and enlargement of the mandible, forehead and pawns (Reusch, 2015). LVH can also be caused by ventricular pressure overload due to aortic stenosis. The stenosis itself or the secondary increase in ejection speed can be detected on echocardiography and the stenosis can be seen on necropsy (Côté et al., 2011a; Robinson and Robinson, 2016). Infiltration of the myocardium by leukocytes in case of myocarditis or neoplastic cells can also give the impression of hypertrophy. These processes often, but not always, change the echogenicity of the myocardium, allowing them to be detected on echocardiography (Côté et al., 2011b; Côté et al., 2011c). On post mortem examination, histopathology can identify the infiltrative cells (Robinson and Robinson, 2016). 2.5 Treatment Asymptomatic cats are often treated with antagonists for the β1-adrenergic receptor, usually atenolol. This can alleviate SAM, but there is no evidence that atenolol treatment increases survival (Schober et al., 2013). Cats that are at risk for ATE because of severe atrial dilatation, a thrombus or spontaneous contrast on echocardiography can be prescribed anticoagulants, usually acetylsalicylic acid or clopidogrel (Côté et al., 2011b). Cats with acute CHF can be treated with oxygen supplementation combined with diuretics and thoracocentesis to diminish lung oedema and pleural effusion, respectively. Chronic CHF can be treated with a combination of diuretics and ACE-inhibitors, supported by a low-salt diet (Côté et al., 2011b). For cats presented with ATE, pain management with opoid analgetics is the primary treatment. Anticoagulants can prevent extension of the thrombus or the formation of new thrombi, but do not dissolve the existing thrombus. This is also not desirable, as sudden reperfusion of ischemic tissues can bring lifethreatening metabolites into the systemic circulation (Luis Fuentes, 2012). Further treatment is supportive in hope of spontaneous recovery. None of these therapies addresses the underlying pathological remodeling of the heart. They therefore do not reverse the disease process and can only reduce the risk or the extent of complications. 2.6 Prognosis The outcome of asymptomatic HCM is highly variable. Many cats never develop atrial dilatation or arrhythmia and therefore remain asymptomatic until they ultimately die from another cause. In a considerable proportion of the asymptomatic cats, however, the disease progresses and leads to fatal complications. A five year follow-up of asymptomatic cats with HCM, mainly Domestic Shorthairs, reported a cardiac death for 22.2% of the cats (Schober et al., 2013). Maine Coons and Ragdolls have a higher risk of death than Domestic Shorthairs (Payne et al., 2010). The prognosis of CHF caused by HCM in cats is usually more favorable than CHF resulting from other cardiomyopathies. Nevertheless, survival is highly variable, ranging from a few months to more than 1.5 year (Côté et al., 2011b). Only 12% of the cats presented in general practice with ATE survives the first 7 days. This is largely due to euthanasia, which is often chosen by the owner because of the severe pain and low chances of recovery (Borgeat et al., 2014b). The chance of recovery is good if only one limb if affected and still has some motor 14

15 function, but poor in the more common case embolization of both hind limbs and decreased temperature. Even in the case of survival, affected cats still have a severe underlying heart disease and many die of CHF or recurrent ATE (Luis Fuentes, 2012). 3. The genetics of HCM 3.1 The genetics of human HCM Pathological descriptions of HCM-compatible findings in humans that died suddenly date back for centuries, but the description by Teare (1958) is viewed as the first modern recognition of HCM as a distinct disease (Maron et al., 2014; Kittleson et al, 2015). The human disease shows a similar morphology and histopathology as its feline counterpart and therefore, both variants of HCM are viewed as essentially the same disease (Maron and Fox, 2015). As in cats, human HCM is associated with a wide range of clinical expressions, ranging from death in adolescence to a normal life span without complications. HCM is less frequent in humans than in cats, but is still the most common inherited heart disease with a prevalence of at least 1 in 500 (Maron et al., 2014). However, taking into account subtle cases identified by novel imaging techniques and genetically affected, but phenotypically negative patients, HCM even affects 1 in 200 humans (Semsarian et al., 2015). There are some differences in complications between the species: ATE is rare in humans, while sudden death seems less common and atrial fibrillation far less common in cats compared to humans (Maron and Fox, 2015). Nevertheless, the many similarities to the human disease, combined with its high incidence and fast progression, make feline HCM the ideal animal model for human HCM (Freeman et al., 2017). This relationship could also be exploited in reverse: the extensive knowledge on the genetics of HCM in humans can provide clues for the investigation of the etiology of HCM in cats. In 1960, human HCM was first described as a familial disease. The pedigree of the affected family was compatible with an autosomal dominant pattern of inheritance (Hollman et al., 1960). This description was soon followed by reports of the same pattern of familial sudden cardiac death in other families (Kittleson et al., 2015). The autosomal dominant pattern is still recognized today, but with a variable expression and incomplete and age-dependent penetrance (Maron et al., 2014; Sabatar-Molina et al., 2018). De novo variants as a cause of HCM are uncommon and X-linked or recessive inheritance patterns are even rarer (Maron et al., 2014; Roma-Rodrigues and Fernandes, 2014; Sabatar-Molina et al., 2018). The first causative genetic variant was identified in the β-myosin heavy chain gene (MYH7) in 1990, soon to be followed by other nonsynonymous variants in the same gene in other families. However, many affected families did not carry such a variant, prompting more linkage studies to associate other genes with the disease. Most of the genes found in this process code for sarcomeric proteins (Seidman and Seidman, 2011; Kittleson et al., 2015). The cardiac myosin binding protein-c gene (MYBPC3) and MYH7 are now established as the most important genes, containing 70% of the known causative variants. Other established sarcomeric genes involved in HCM are cardiac troponin T and I (TNNT2 and TNNI3), ventricular essential and regulatory myosin light chain (MYL2 and MYL3), α-tropomyosin (TPM1) and cardiac α-actin (ACTC1) genes (Maron et al., 2012; Ho et al., 2015; Sabater-Molina et al., 2018). Together, these eight genes contain 99% of the pathogenic or likely pathogenic variants found in patients that were referred for HCM testing (Welsh et al., 2017a). Over 30 other genes have been proposed to be involved in HCM, but the evidence here is more contested. For only 5 of these genes, the evidence is considered to be strong (Maron et al., 2012; Walsh et al., 2017a). Sarcomeric genes are associated with other cardiomyopathies. Dilated cardiomyopathy (DCM) causes systolic dysfunction and dilation of the ventricle, while restrictive cardiomyopathy (RCM) causes diastolic dysfunction with normal ventricular dimensions. Left ventricular non-compaction (LVNC) is characterized by ventricular trabeculae and sometimes systolic dysfunction (Elliot et al., 2008). Most sarcomeric genes are associated with both DCM and RCM, while only ACTC1, MYH7 and TNNT2 are associated with HCM (Teo et al., 2015). The genetic picture of HCM is further complicated by the large number of different variants that have been associated with this disease. Only for the genes with convincing evidence, this number is over 1400 (Maron et al., 2014). The majority of these variants is limited to a single or only a few families, the so-called private variants (Ho et al., 2015). Some variants, however, are found in multiple families in a specific geographic region and seem to originate from a common ancestor. These founder variants can affect thousands of people (Seidman and Seidman, 2011). The most extreme example is a 25 bp deletion in MYBPC3 that has arisen years ago and is now present in approximately 4% of the individuals originating from the Indian subcontinent, affecting million people (Dhandapany et al., 2009; Kittleson et al., 2015). Other founder 15

16 variants were discovered in Finland, France, Iceland, Italy, Japan, The Netherlands, South Africa and the United States (Seidman and Seidman, 2011; Carrier et al., 2015). Finally, the genes involved in HCM contain some 'mutation hot spots' where the same variant has arisen repeatedly in unrelated families (Ho et al., 2015; Sabatar-Molina et al., 2018). The development and increasing use of high-throughput sequencing techniques has shed a new light on the genetic background of HCM. Sequencing of healthy persons revealed a large amount of variation in the sarcomeric genes, including variants that were considered to be pathogenic. Approximately 6.5% of the variants thought to cause HCM occur in controls at a frequency that is incompatible with the estimated prevalence of HCM, casting doubts on their pathogenicity (Walsh et al., 2017b). Other variants are rare in both the HCM-affected and healthy population and their pathogenicity remains unclear. These variants of uncertain significance make up a large part of the variants that were formerly classified as disease-causing. Despite the extensive knowledge of the genetics of HCM in humans, about 35% of the cases remains unexplained in genetic terms. The chances for identifying a causative variant are somewhat higher with a family history of HCM (Maron et al., 2014). 3.2 The structure of the sarcomere Myofibrils are the main component of skeletal and cardiac myocytes. They consist of contractile units, the sarcomeres, that are arranged in series. The sarcomeres contain an M disk in the middle and are delineated by Z disks. Thick filaments extend from the M disk to both ends of the sarcomere, while thin filaments extend towards its middle from the Z disks. These filaments partially overlap and interact with each other. When the Ca 2+ -concentration in a myocyte increases, the thick filaments slide along the thin ones, narrowing the sarcomere. The cumulative narrowing of the sarcomeres of a muscle results in macroscopic contraction (Nelson and Cox, 2008b). The main component of thick filaments is the protein myosin. Myosin is a hexamere that consists of two heavy chains, two essential light chains and two regulatory light chains. The myosin binding protein-c is bound to myosin, but is not considered to be a part of the thick filaments. The giant protein titin spans the distance between the Z disk and M disk and is also connected to the thick filaments. The thin filaments contain actin, tropomyosin and troponin. The main component is filamentous actin (F-actin), a polymer of globular actin (G-actin) subunits. Tropomyosin forms a long strand that is bound to actin. The troponin complex is bound to tropomyosin and consists of troponin C, I and T (figure 5). The interaction between the thick and thin filaments is established by the binding of myosin to actin (Nelson and Cox, 2008b). Figure 5. The major sarcomeric proteins involved in HCM, labeled with the names of their respective genes. The thick filament is colored green, the thin filament is colored pink. From: Seidman and Seidman, The MYBPC3 gene The MYBPC3 gene codes for the cardiac isoform of myosin binding protein-c (cmybp-c). In humans and mice, this isoform is only expressed in cardiomyocytes. Two other isoforms are known, which are confined to skeletal muscles (Knöll, 2012). Compared to its skeletal muscle counterparts, the cardiac isoform contains extra signaling domains with three phosphorylation sites (Carrier et al., 2015). Variants in the MYBPC3 gene are not only associated with HCM, but also with other cardiomyopathies, i.e. DCM, RCM and LVNC (Carrier et al., 2015). 16

17 cmybp-c is not essential for life and development, as knock-out mice for this protein have a normal viability and life span (Marston et al., 2012). Nevertheless, these mice show a marked left ventricular hypertrophy and impaired diastolic function, resembling the HCM phenotype (Harris et al., 2002; Carrier et al., 2004). The function of cmybp-c is dependent on its interactions with other sarcomeric proteins, as it shows no enzymatic activity and no relevant affinity to sugars or small ions like calcium (Pfuhl and Gautel, 2012). cmybp-c is not essential for sarcomeric structure, as the organization of the sarcomeres is normal in both knockout mice and human patients carrying a substitution variant in MYBPC3 (Harris et al., 2002; Pfuhl and Gautel, 2012). The function of the constitutive binding of the C-terminus to myosin and titin is wellestablished. This binding serves as anchoring for the smybp-c protein (Pfuhl and Gautel. 2012). The details of the variable interactions of the N-terminus with actin and/or myosin are not fully understood (Pfuhl and Gautel, 2012; Moss et al., 2015). These interactions are thought to be dependent on phosphorylation and to modify the interactions between actin and myosin. Contraction of the cardiac myofilaments is increased in the presence of phosphorylated cmybp-c in comparison to dephosphorylated cmybp-c. According to a commonly postulated underlying mechanism, the dephosphorylated cmybp-c binds to myosin with its N- terminus and inhibits the formation of cross-bridges by limiting the mobility of the myosin head domains. The phosphorylated cmybp-c, in contrast, binds to actin and interacts with tropomyosin, thereby increasing the sensitivity of the thin filaments to Ca 2+ (Carrier et al., 2015; Moss et al., 2015). The relationship between the degree of phosphorylation and cross-bridge kinetics is continuous rather than discrete. The phosphorylation is controlled by multiple kinases, which are part of different physiological pathways (Moss et al., 2015). An example is β-adrenergic stimulation of the heart, that activates campdependent protein kinase. Phosphorylation of cmybp-c by this kinase contributes to an enhancement in cardiac contractility and relaxation, enabling an increase in cardiac output (Carrier et al., 2015). These characteristics suggest that cmybp-c plays a central role in the regulation of cardiac function. This corresponds with the observation that knock-out mice possess basic cardiac contractile function, but lack the normal hemodynamic responses (Sadayappan and De Tombe, 2012). In 45% of the patients were a causative variant is identified, it is located in the MYBPC3 gene (Ho et al., 2015). This amounts to more than 500 different variants (Roma-Rodrigues and Fernandes, 2014). Both truncating and substitution variants have been described in MYBPC3. The pathogenicity of these variants seems to be rather similar, although some state that substitution variants cause an earlier onset of disease than truncating variants (Sabater-Molina et al., 2018). A possible pathophysiological mechanism by which variants in the MYPC3 gene cause HCM is an increase in calcium sensitivity of the cardiomyocytes. Cardiomyocytes extracted from human HCM-patients or mice that carry a disease-causing variant in MYBPC3 have a higher Ca 2+ -sensitivity than those from wild-type controls. This effect, however, is hard to distinguish from secondary changes in phosphorylation patterns, which are common in HCM patients and also affect calcium sensitivity. A second line of evidence for this hypothesis comes from the fact that skinned cardiomyocytes become more sensitive to calcium when cmybp-c is physiochemically extracted (Marston et al., 2012). 3.4 The MYH7 gene The MYH7 gene codes for the β-myosin heavy chain (β-myhc) protein. Twelve sarcomeric myosin heavy chain isoforms are recognized in mammals. Two of these, α-myhc (encoded by the MYH6 gene) and β-myhc, are expressed in the heart. α-myhc is the dominant form in the atria and β-myhc is the dominant form in the ventricles in humans and possibly in most other mammals (Mascarello et al., 2016). In rats, however, α-myhc is the most common form in the ventricles. α-myhc might also be important in feline ventricles, as suggested by the response on thyroid hormone stimulation. Thyroid hormone increases the synthesis of αmyhc, but not β-myhc, and causes marked left ventricular hypertrophy in rats and cats, while this is limited in humans (Dillmann, 2002; Côté et al., 2011e). In addition to the heart, MYH7 is also expressed in most skeletal muscles in both humans and cats (Mascarello et al., 2016). Variants in MYH7 are associated with several hereditary pathologies in both tissues in humans. These usually, but not exclusively, show an autosomal dominant pattern of inheritance. Liang distal myopathy and myosin storage myopathy, two diseases of skeletal muscles, are caused by variants that reside mainly in exons and 37-40, respectively. The main symptom of these diseases is muscle weakness, but muscular (pseudo)hypertrophy can also be present. Concurrent cardiomyopathy has been described for both diseases, 17