Corresponding author: R.J. Naylor, Veterinary Clinical Sciences, Comparative Neuromuscular Diseases Laboratory, The Royal Veterinary College, London, UK; e-mail: email@example.com
Type 1 polysaccharide storage myopathy (PSSM1), an equine glycogen storage disorder caused by a gain of function mutation (R309H) in the gene encoding glycogen synthase (GYS1), is associated with the accumulation of amylase-resistant alpha-crystalline polysaccharide inclusions within skeletal muscle. Several glycogenoses in humans have a cardiac phenotype, and reports exist of horses with PSSM and polysaccharide inclusions in cardiac muscle.
To investigate the hypothesis that horses with PSSM1 display a cardiac phenotype. Our objectives were to compare plasma cardiac troponin I (cTnI) concentration and the incidence of cardiac arrhythmias in PSSM1 homozygotes, heterozygotes, and control horses.
One hundred and twenty-five Belgian and Percheron horses under the same management were genotyped for the R309H GYS1 mutation. From these, 8 age-, breed-, and sex-matched cohorts of each genotype were identified. Plasma cTnI concentration and incidence of cardiac arrhythmias (determined by 24-hour Holter ECG) were compared between the groups.
Although some PSSM1-affected horses had mildly increased plasma cTnI concentrations, there was no significant difference in cTnI concentrations between groups. There were no significant differences in the incidence of ectopic beats, cardiac conduction intervals or mean heart rate between groups.
Conclusions and clinical importance
We found no evidence of clinically relevant cardiac myocyte injury or arrhythmias in horses with PSSM1. Additional study is required to determine whether myocardial function may be compromised in this disorder.
Type 1 polysaccharide storage myopathy (PSSM1) is a common cause of exertional rhabdomyolysis that has been described in many horses,[1, 2] with some breeds (eg, Percheron, Belgian Draft horses) having a particularly high prevalence.[1, 3] Affected horses have a gain of function, autosomal dominant, missense mutation (R309H) in the glycogen synthase gene (GYS1) and accumulation of amylase-resistant alpha-crystalline polysaccharide inclusions within skeletal muscle. Diagnosis is achieved by GYS1 R309H genotyping, using a restriction fragment length polymorphism assay after polymerase chain reaction (PCR) of DNA extracted from whole blood samples or hair roots.
Unlike the liver-specific isoform of glycogen synthase (encoded by a separate gene, GYS2), GYS1 is ubiquitously expressed, but has high expression in tissues that are heavily reliant on glycogen metabolism such as cardiac and skeletal muscle. The importance of glycogen synthase to both cardiac and skeletal muscle is apparent in human patients with homozygous null mutations in GYS1 who present with exercise intolerance and cardiomyopathy. Indeed, several of the 11 recognized glycogenoses that affect humans have a cardiac phenotype (usually cardiac failure or conduction disturbances). Typically, these are associated with cardiac polysaccharide accumulations similar to those seen in equine PSSM1.[7, 8]
Some authors have described horses with PSSM that have polysaccharide inclusions in cardiac as well as in skeletal muscle.[9, 10] Specifically, amylase-resistant inclusions were detected in cardiac muscle from 1 of 9 PSSM-affected draught horse-related breeds in 1 study,1 of 3 Quarter Horses, and in 3 of 8 Belgian and Percheron Draft horses in other studies.[10, 12] Consequently, for the current study, we hypothesized that horses with PSSM1 may have an underlying cardiac phenotype, and if so, that homozygotes would be more severely affected than heterozygotes. Our aims were to compare plasma cardiac troponin I (cTnI) concentration and the incidence of cardiac arrhythmias, between age-, breed-, and sex-matched horses that were homozygous and heterozygous for the R309H GYS1 mutation and control horses.
Materials and Methods
All studies were conducted in accordance with local university animal care and use protocols.
One hundred and twenty-five Belgian and Percheron horses maintained at pasture at an Auburn University research facility were genotyped. In brief, DNA was extracted from stored whole blood in EDTA using Nucleon's BACC kit1 according to the manufacturer's instructions. Genotyping of the entire herd, performed using a restriction fragment length polymorphism assay, identified 9 homozygotes (H/H), 45 heterozygotes (HR), and 71 unaffected (RR) horses. From these horses, 3 cohorts containing 8 age-, breed-, and sex-matched horses of each genotype were chosen for further cardiac evaluation. Where a horse with an identical signalment match was not available, age was prioritized, followed by breed and lastly sex. All horses were not in current work, had 24-hour turn-out, and were fed 2 kg/day of a supplementary concentrate diet and alfalfa as forage. The concentrate diet was designed to provide 22 MJ/day per horse, and contained 149 g/kg crude protein, 45 g/kg fat, 143 g/kg fiber, and 165 g/kg starch.
Venous blood samples were collected by direct jugular venipuncture into plain and heparinized Vacutainers2 from 22 of the study horses at rest. Additional blood samples were collected from 12 of these horses, 4 and 24 hours after 20 minutes of submaximal free exercise (trot and canter). The samples were centrifuged at 4°C, and serum or plasma separated and stored at −80°C. Plasma cTnI concentrations were measured using UniCel Dxl 800 Access 3 immunoassay. Serum creatine kinase (CK) and aspartate transferase (AST) activities were determined using a Roche Cobas C311 analyzer.4
All horses had routine physical examination and cardiac auscultation performed. Subsequently, on a separate occasion, Holter electrocardiographic (ECG) examinations over 24 hours were performed on all horses during stall confinement using a Lifecard CF system.5 Electrocardiographic data were analyzed digitally6 for ectopic beats, mean QRS width, QT and PR intervals, and heart rate.
The normally distributed (Kolmogorov–Smirnov test) age, breed, and sex distributions within groups were compared using 1-way analysis of variance (ANOVA) and chi-squared tests to confirm appropriate matching of groups. Subsequently, 1-way ANOVA or Kruskal–Wallis tests (depending on data normality) were used to compare data between genotyped groups (where suitable, non-normally distributed data was transformed before analysis). Subsequently, data from homozygous and heterozygous horses were combined to enable comparison of data from all horses with the GYS1 mutation and unaffected controls. The correlation between plasma cTnI concentration and serum muscle enzyme activity was assessed using Pearson's correlation coefficient. All analyses were performed using statistical software7 and differences were considered statistically significant when P < .05.
The homozygote group (HH) consisted of 3 Percheron and 5 Belgian Draft horses, including 3 mares and 5 geldings, with a mean age (±SD) of 9.6 (±2.6) years; the heterozygotes (HR) consisted of 1 Percheron and 7 Belgians, 5 mares, and 3 geldings, with a mean age of 9.8 (±2.4) years, and the control group (RR) consisted of 3 Percheron and 5 Belgian horses, including 3 mares and 5 geldings with a mean age of 9.9 (±2.7) years). There was no significant difference among the age (P = .997), breed (P = .11), or sex (P = .47) of the 3 groups.
None of the horses examined had any signs of cardiac disease on physical examination. Although occasional horses had plasma cTnI concentrations that were above the reference ranges, there was no significant difference between the resting or 24-hour postexercise plasma cTnI concentrations of the 3 groups of horses (Fig 1). No association was identified between cTnI concentration and resting CK (P = .73) or AST activity (P = .96), or between cTnI concentration and 4-hour postexercise CK (P = .52) or 24-hour postexercise AST activity (P = .42) (Fig 2).
Occasional ventricular (VPC) and atrial premature complexes (APC) were present on electrocardiograms of 10 and 7 horses, respectively, 6 of these horses experiencing both forms of ectopic beats. However, there were no significant differences in the incidence of ectopic beats among the different groups (Table 1). No other forms of cardiac arrhythmia were identified in any horse. There were no significant differences in PR, QRS, or QT duration or mean heart rate among the groups (Table 1).
Table 1. Mean (±SD) incidence of ectopic beats, heart rate, and cardiac conduction intervals calculated from 24-hour Holter electrocardiographic recordings for matched groups
Atrial Premature Complexes in 24 hours
Venticular Premature Complexes in 24 hours
Mean Heart Rate bpm
QT Interval seconds
QRS Interval seconds
PR Interval seconds
HH, homozygotes; HR, heterozygotes; RR, control horses.
n = 8 horses for each group.
Previous studies in horses with PSSM1 were focused on the histological and ultrastructural findings observed in skeletal muscle of affected animals.[13, 14] In the present study, our aim was to investigate cardiac phenotype in affected horses, given the presence of cardiac signs in glycogenoses of humans[15-17] and reports of PSSM-affected horses with cardiac pathology[10-12] and sudden death. We did not, however, find evidence for ongoing cardiomyocyte injury or intermittent arrhythmias in affected horses.
Plasma cTnI concentration is a sensitive marker of cardiac damage in horses with increases reflecting cardiomyocyte injury caused by disease, metabolic derangements, and ischemia.[18-20] Some human athletes with rhabdomyolysis of undetermined etiology have increased cTnI and T concentrations.[21, 22] Furthermore, researchers have suggested that increased cardiac troponin concentrations are a useful screening test for subclinical cardiac involvement in human patients with Pompe disease (a deficiency of lysosomal acid glucosidase that results in abnormal polysaccharide inclusions within striated muscle fibers). Recently, increased cTnT, but not cTnI concentrations, were identified in patients with skeletal muscle pathology and no identifiable cardiac disease, suggesting that cTnT may be less specific for myocardial damage than previously suggested. As affected horses in our study were susceptible to rhabdomyolysis and had substantial skeletal muscle pathology, we considered it possible that any increases in cTnI could conceivably reflect skeletal rather than cardiac-muscle damage (either because of expression of the cardiac-specific isoform in diseased skeletal muscle or nonspecificity of the assay in horses), but we did not observe any association between CK or AST activity and cTnI concentration at rest or after 20 minutes of submaximal exercise.
Although there was no difference in cTnI concentrations among the genotype groups, 2 individuals, homozygous for the GYS1 mutation had increased concentrations of cTnI either pre- or postexercise. The effect of exercise on cTnI concentration in horses has been evaluated previously. Short bursts of high intensity exercise resulted in minimal cTnI increases in Thoroughbred racehorses in 1 study, with all values < 0.035 ng/mL. In contrast, endurance exercise over 60 and 120 km resulted in much greater increases in cTnI in both clinically normal horses (mean, 0.13 ng/mL) and in those withdrawn from the event (mean, 0.31 ng/mL). In the present study, horses performed 20 minutes of submaximal exercise, which (on the basis of previous data) seems unlikely to have precipitated cardiac damage per se. Indeed, cTnI concentrations for all control horses except 1 remained below the laboratory upper reference range (0.04 ng/mL). Consequently, we believe that the increased concentrations recorded in some PSSM1-affected individuals may reflect low grade underlying cardiomyocyte damage, perhaps associated with amylase-resistant polysaccharide inclusions within the myocardium. In horses with monensin toxicosis and overt clinical signs of cardiac disease, much greater cTnI concentrations (up to 3.5 and 70 ng/mL) have been reported.[19, 20] However, subclinical monensin toxicosis has been associated with cTnI concentrations up to 0.10 ng/mL, which is of a similar magnitude to the increases reported, herein, in some animals.
There are individual reports of sudden (perhaps cardiac-related) death in Cob Normand Draft horses with PSSM but, in the present study, we found no evidence of disturbed conduction or arrhythmogenesis in PSSM1-affected horses. In contrast, atrioventricular conduction defects have been reported in mice with a glycogen storage cardiomyopathy associated with over-expression of AMP-activated protein kinase (a model of Wolff Parkinson White syndrome). Similarly, human patients with juvenile onset Pompe disease may have a shortened PR interval that is presumed to result from enhanced atrioventricular conduction secondary to accumulation of amylase-resistant polysaccharide within cardiac myocytes.[17, 27] These electrocardiographic changes resolve after patient supplementation with the deficient lysosomal acid alpha glucosidase enzyme.16 The incidence of cardiac arrhythmias in other glycogenoses of humans is variable. One study documented ventricular tachycardia in patients with type 3 glycogen storage disease (deficiency of glycogen debranching enzyme), whereas another study failed to identify any cardiac abnormalities in the same disorder. Such differences may result from variation in disease severity and other factors, such as age, diet, athleticism, or recent exertion. In the current study, age, diet, and management were standardized, but it remains to be determined whether PSSM1 predisposes horses to cardiac arrhythmias at exercise. Future study, with assessment of exercising electrocardiograms, may be indicated. From similar diseases in humans, one might expect that faster conduction at the AV node would be manifested as a shortened PR interval, potentially predisposing these horses to the onset of supraventricular tachycardia. However, despite rare reports of sudden death in some horses with PSSM, the authors remain unaware of cardiac disease being an overt problem in athletic horses with PSSM1.
Skeletal muscle is composed of 3 different fiber types, which differ in their metabolic properties, in particular, their propensity for beta oxidation of lipid and anaerobic glycolysis. Type 1 skeletal muscle fibers have more mitochondria and therefore oxidative capacity, and typically rely on lipid metabolism for ATP generation. In contrast, type 2a (mixed oxidative-glycolytic) and type 2x (glycolytic) fibers are more heavily reliant on the metabolism of glycogen to provide ATP, possess fewer mitochondria, and are more susceptible to fatigue. As such, it is not surprising that it is the type 2 muscle fibers that are predominantly affected by the gain of function GYS1 mutation in PSSM1 and that pathology is most severe in muscles with high proportions of glycolytic (type 2x) fibers. The formation of less highly branched polysaccharide may act to decrease the amount of metabolic substrate available for energy production in type 2 skeletal myocytes and the resulting intracellular accumulations also may disrupt contractile function.
Cardiac myocytes are similar to type 1 skeletal muscle fibers, in that oxidation of lipids is predominantly responsible for generation of ATP under aerobic conditions. In contrast, conducting cells are more dependent on carbohydrate metabolism, and therefore may be more susceptible to the presence of the GYS1 mutation in PSSM1 than working ventricular myocytes. Despite this, the mechanism of pre-excitation in patients with AMP kinase mutations appears to be a consequence of disruption of the annulus fibrosus by vacuole-laden ventricular myocytes rather than by metabolic abnormalities in conduction cells. This allows atrial impulses to bypass the atrioventricular node and activate the ventricular myocardium independently, causing a shortened PR interval.
PSSM in horses frequently is associated with poor performance in the absence of clinical signs of rhabdomyolysis, and although this may be a reflection of subclinical muscle pathology, it also may have a cardiac component. In this study, it was not possible to evaluate the myocardium histologically to determine the presence or absence of amylase-resistant polysaccharide inclusions within the myocardium of affected animals. Unfortunately, we were unable to perform echocardiographic assessment of myocardial function because of the size and conformation of the horses, but any cardiac dysfunction, if present, likely remains subclinical in most horses because no animals had overt signs of cardiac disease on physical examination. Nevertheless, additional study, including resting echocardiography, postexercise echocardiography, and dobutamine stress testing or some combination of these, may be helpful in identifying any underlying myocardial dysfunction in PSSM1.
The pathophysiology of PSSM1 remains unclear. Among potential disease mechanisms, both metabolic compromise and direct physical disruption caused by accumulating glycogen and polysaccharide are possibilities. In this study, we found no conclusive evidence supporting the presence of an overt cardiac phenotype in horses with PSSM1, although the presence of an occasional horse with increased cTnI concentrations suggests that subclinical cardiac disease may exist in some individuals and warrants additional investigation. Given that exercise seems to limit severity of the skeletal muscle phenotype in affected animals, 1 hypothesis that may warrant additional investigation is that the constant work performed by the beating myocardium decreases the susceptibility of cardiomyocytes to accumulation of polysaccharide and with it, the cardiac complications seen in other glycogenoses. Additional investigation into pathophysiology may assist in directing investigation of novel treatments in this common equine disorder.
The authors thank colleagues at Auburn University for their kind assistance in collecting clinical samples. Authors disclose no conflict of interest.
GE Healthcare, Buckinghamshire, UK
Becton, Dickinson and Company, Oxford, UK
UniCel Beckman Coulter, Miami, FL
Roche Diagnostic Corporation, Indianapolis, IN
Spacelabs Healthcare, Issaquah, WA
Laboratory Corporation of America, Ambulatory monitoring services, Burlington, NC