Dilated cardiomyopathy is a common cardiac disease in Great Dane dogs and is characterized by progressive loss of myocardial systolic function, eccentric dilation, atrial fibrillation, and congestive heart failure.1 Despite its relatively high prevalence, the etiology and molecular mechanisms underlying this disease are poorly understood. Cardiac function is partly regulated by the transcriptional activity of its constitutive cell types and study of the molecular events occurring in DCM would benefit from evaluation of tissue-wide genome expression patterns. To this end, previous studies have used oligonucleotide microarrays to contemporaneously examine the transcriptional activity of a large percentage of the canine genome,2–6 and specifically, to identify genes thought to be important in Doberman Pinschers with DCM,7 Boxers with arrhythmogenic right ventricular cardiomyopathy,8 dogs with degenerative mitral valve disease,9 as well as in humans with heart disease.10–12 To better understand the pathophysiology of DCM in Great Danes and to identify new and intriguing candidate genes for future study, we performed a second generation microarray-based analysis of >38,000 genes in Great Danes with naturally occurring end-stage DCM.
Background: Dilated cardiomyopathy (DCM) is a common cardiac disease of Great Dane dogs, yet very little is known about the underlying molecular abnormalities that contribute to disease.
Objective: Discover a set of genes that are differentially expressed in Great Dane dogs with DCM as a way to identify candidate genes for further study as well as to better understand the molecular abnormalities that underlie the disease.
Animals: Three Great Dane dogs with end-stage DCM and 3 large breed control dogs.
Methods: Prospective study. Transcriptional activity of 42,869 canine DNA sequences was determined with a canine-specific oligonucleotide microarray. Genome expression patterns of left ventricular tissue samples from affected Great Dane dogs were evaluated by measuring the relative amount of complementary RNA hybridization to the microarray probes and comparing it with expression from large breed dogs with noncardiac disease.
Results: Three hundred and twenty-three transcripts were differentially expressed (≥2-fold change). The transcript with the greatest degree of upregulation (+61.3-fold) was calstabin2 (FKBP12.6), whereas the transcript with the greatest degree of downregulation (−9.07-fold) was triadin. Calstabin2 and triadin are both regulatory components of the cardiac ryanodine receptor (RyR2) and are critical to normal intracellular Ca2+ release and excitation-contraction coupling.
Conclusion and clinical importance: Great Dane dogs with DCM demonstrate abnormal calstabin2 and triadin expression. These changes likely affect Ca2+ flux within cardiac cells and may contribute to the pathophysiology of disease. Microarray-based analysis identifies calstabin2, triadin, and RyR2 function as targets of future study.
calcium release units
Materials and Methods
The study protocol was approved by the University of Pennsylvania Institutional Animal Care and Use Committee. Full-thickness 1 cm3 tissue samples from the lateral wall of the left ventricle were obtained within 30 minutes after euthanasia from 3 client-owned Great Danes with severe end-stage DCM and from 3 large breed dogs with noncardiac disease. Great Dane dogs included a 9-year old female and two 7-year old males. All 3 dogs were diagnosed as having echocardiographic left ventricular eccentric hypertrophy13 and fractional shortening <20% in the absence of other cardiac defects (ie, severe degenerative mitral valve disease). All affected dogs were receiving conventional heart failure therapy with diuretics, digoxin or pimobendan, and ACE inhibitors at time of euthanasia. Control dogs included a 12-year-old male Labrador Retriever with gastric dilation and volvulus, a 10-year-old female Saint Bernard with intervertebral disc disease and aspiration pneumonia, and a 4-year-old male Greater Swiss Mountain Dog with lymphoma. Brief echocardiographic evaluation at the time of euthanasia was performed to exclude the presence of significant cardiac disease. Samples were transferred to RNALater,a refrigerated at 4°C for up to 24 hours, and then frozen at −80°C until further processing. Tissue processing, extraction of RNA, and general methodology of the microarray analysis has been described previously.7,9 For this study, the canine genome microarrayb was a 2nd generation oligonucleotide-based single color array that used the principle of complementary hybridization to assess expression of 42,869 probe sets, representing >38,000 canine genes. Transcript identity was achieved with the annotations provided by the chip manufacturer. In instances where annotations were not provided or incomplete, the transcript sequence was matched against a publicly available nucleotide databasec in an attempt to find a similar reported sequence. Sequences were considered matched and the transcript identity recorded if the E-score was ≤1 × 10−20. To qualify for analysis, a transcript had to be evident in the DCM or control tissue and probe signal intensity significantly different by use of a 2-tailed Student's t-test with P<.05. Differential expression between DCM and control tissue was defined as ≥2-fold change in expression found by both MAS5 and GCRMA analysis.14 Using such a stringent criteria allowed us to avoid the false positives obtained with use of MAS5 alone as well as to include the ability of GCRMA to account for nonspecific hybridization-related background signal.15 Transcripts with a negative fold change were downregulated in DCM tissue, whereas transcripts with a positive fold change were upregulated in DCM tissue, compared with control tissues. To further explore the findings of the microarray-based results, differential gene expression for the 2 transcripts of greatest interest was evaluated by use of a real-time reverse transcriptase-quantitative polymerase chain reaction (RT-qPCR) assay as described previously.7,9 For this study, GAPDH was used as the housekeeping gene and relative quantification was performed by the ΔΔCt method. Primer sequences were as follows: calstabin2-F: 5′-CCCGATGTGGCGTATGGA-3′, calstabin2-R:5′-TCAAAGATGAGGGTGGCATTG-3′; triadin-F: 5′-TCCAGTGTGTCTACTTGGA-3′, triadin-R: 5′-TGCGGGAGTGACAGGAAACT-3′. Mean relative RT-qPCR expression between experimental groups was determined by a Student's t-test on log transformed data.
Three hundred and twenty-three transcripts were differentially expressed in Great Dane myocardium versus control. Two hundred and ninety-nine transcripts were upregulated (Supplemental Table S1). The transcript with the greatest degree of upregulation (+61.3-fold) coded for FK506-binding protein 1B isoform a, also known as FKBP12.6 or calstabin2.16,17 Additional upregulated transcripts with well-described cardiac functions included titin (+8.96-fold), myosin light chain (+5.84-fold), collagen III α1 chain (+5.03-fold), tropomyosin (+4.47-fold), collagen I α1 chain (+4.35-fold), myosin heavy polypeptide 7B (+3.11-fold), and calmodulin protein kinase (+2.40-fold). Twenty-four transcripts were downregulated (Supplemental Table S2). The transcript with the greatest degree of downregulation (−9.07-fold) was triadin. None of the other 23 downregulated transcripts were genes with well-described cardiac functions. MIAME-compliant18 data from the oligonucleotide microarray analysis were deposited in the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) for public access (accession, GSE16049). Using RT-qPCR, mean expression of triadin in relative units was lower in Great Danes (0.00074; range, 0.00003–0.0002) versus in control (0.34; range, 0.010–1.00; Fig 1A), although this change was not statistically significant (P=.053). Mean RT-qPCR expression of calstabin2 in relative units was significantly greater in Great Danes (58.7; range 8.5–128.2) versus in control (0.676; range, 0.120–1.00; P=.015; Fig 1B).
Microarray-based analysis has been previously used to identify target genes of interest in both humans10–12 and dogs4–9 with heart disease, as well in dogs with a variety of noncardiac diseases.19–22 The simultaneous evaluation of thousands of genes identifies a subset of genes that are differentially expressed between experimental groups. This subset of genes can then be examined for clues as to the underlying molecular mechanisms that contribute the disease process. In our study of >38,000 genes in Great Danes with DCM, we find it striking that the 2 genes with the greatest amount of differential regulation are both essential components of the cardiac ryanodine receptor (RyR2). RyR2 is a homotetrameric Ca2+ release channel that spans the sarcoplasmic reticulum (SR), and is responsible for Ca2+ release from the SR into the cytosol. SR Ca2+ release is triggered by opening of sarcolemmal L-type Ca2+ channels, and the subsequent flux of Ca2+ from the SR permits binding of Ca2+ to the cardiac sarcomere and initiation of systolic contraction. Abnormalities in RyR2 function are associated with development of arrhythmias, loss of excitation-contraction coupling, and myocardial failure.23,24 These changes have been regarded as being associated with the phenotype of arrhythmogenic right ventricular cardiomyopathy in Boxer dogs.8 RyR2 function is closely regulated by a host of complementary molecules, including triadin, calstabin2, junctin, sorcin, calsequestrin, and phosphatases 1 and 2A.25 Together with RyR2 and the colocalized sarcolemmal L-type calcium channels, these molecules form so-called calcium release units (CRU), which act in concert to synchronously release Ca2+ at the beginning of systole.26 Poor function or coordination of CRU results in reduced systolic Ca2+ release, diastolic Ca2+ leakage, formation of cardiac arrhythmias, and reduced contractility.27 Thus, regulation of RyR2 is critical to proper myocardial function.
Triadin is a transmembrane protein that is closely bound to RyR2, and together with its related protein, junctin, acts as a sensor of Ca2+ concentration within the SR.28,29 Triadin exists as several different isoforms, with a 32 kDa isoform (triadin-1), being most prevalent in cardiac muscle.30 The triadin molecule possesses a long C-terminal domain that extends from the RyR2 complex into the lumen of the SR. This C-terminal domain is rich in motifs of repeating lysine-glutamic acid residues that compete with Ca2+ for binding onto calsequestrin. Studies of triadin in vitro offer insight into the complex interactions among triadin, calsequestrin, and RyR2. When SR Ca2+ concentration is low, triadin is primarily bound to calsequestrin and stimulation of RyR2 is inhibited.31 When SR Ca2+ concentration is high, triadin-calsequestrin binding is lessened and the probability of RyR2 opening is increased.31 When triadin-calsequestin binding is inhibited through use of a triadin analogue, RyR2 channels are more likely to be chronically open, resulting in diastolic Ca2+ leak and diminished systolic Ca2+ flux.32 Thus, the combined interactions between Ca2+, triadin, junctin, and calsequestrin help initiate and coordinate RyR2 opening during systole and closing during diastole. In general, reduced triadin-calsequestrin interaction results in hyperactive RyR2, arrhythmogenic Ca2+ leakage, and poor excitation contraction coupling.31,32 Interestingly, overexpression of triadin in vitro also results in hyperactive RyR2, altered excitation contraction coupling, and spontaneous arrhythmogenic membrane depolarizations.31,33 In the intact animal, triadin, along with junctin, calsequestrin, and L-type Ca2+ channels, modulates RyR2 function in an even more complex and incompletely understood manner. The fact that decreased triadin expression may alter expression of junctin and calsequestrin, as well as disrupting function of the entire CRU by preventing the colocalization of RyR2 and sarcolemmal calcium channels, makes interpretation of triadin function across different experimental models difficult.34 In a small study of humans with heart failure, triadin expression was decreased by 22% and was believed to contribute to abnormal excitation contraction coupling.28 While the nature of interactions between triadin, RyR2, and the CRU have not been fully described, there is general agreement that abnormal triadin expression can result in altered RyR2 activity, altered excitation contraction coupling, and reduced myocardial function.28,29,32,34 Thus, it is possible that decreased triadin expression is important in the development of myocardial failure and arrhythmias in Great Danes with DCM.
Calstabin2 is another important component of the RyR2 regulatory complex.25 Each of the 4 RyR2 tetramers contains a binding site for 1 calstabin2 molecule. The putative function of calstabin2 is to stabilize RyR2 in the closed configuration during diastole.24 Previous reports suggest that hyperphosphorylation of RyR2 leads to dissociation of calstabin2 and diastolic Ca2+ leakage from the SR into the cytosol35–37; however, these results are highly controversial.38–40 We have previously shown that decreased calstabin2 expression and protein is associated with “leaky” RyR2 in Boxer dogs with arrhythmogenic right ventricular cardiomyopathy.8 Interestingly, the results of the present study indicate markedly increased transcription of calstabin2 in Great Danes with DCM. Previous studies have primarily concentrated on downregulated calstabin2 as a potential cause of arrhythmias and reduced contractility in the setting of arrhythmogenic right ventricular cardiomyopathy or catecholaminergic polymorphic ventricular tachycardia,24,25,41–45 and studies investigating the effects of upregulated calstabin2 expression are relatively few. In mice, overexpression conferred a protective effect against inducible ventricular tachycardia presumably through reduction of diastolic SR Ca2+ leak.46 These results are consistent with in vitro studies demonstrating that overexpression of calstabin247 or stabilization of calstabin2 binding to RyR248 reduces spontaneous arrhythmogenic Ca2+ release in cardiac myocytes. Importantly, even after stimulation with high levels of catecholamines (a trigger for formation of malignant arrhythmias) the protective effect of increased calstabin2 expression remains evident.46 The finding of increased calstabin2 and simultaneously decreased triadin in Great Danes with DCM is intriguing. Boxer dog arrhythmogenic right ventricular cardiomyopathy and Great Dane DCM demonstrate opposite changes in calstabin2 expression. We find it intriguing that the characteristic phenotype of disease in these breeds agrees with the putative consequence of altered expression. Boxer dogs exhibit calstabin2 downregulation and a high degree of malignant arrhythmias and sudden death, while Great Danes exhibit calstabin2 upregulation and in the experience of the authors and others1,49–51 demonstrate a relatively low incidence of ventricular arrhythmias and sudden death. Triadin downregulation in Great Danes is similarly intriguing in that in the experience of the authors and others,1,49–51 DCM in giant breeds is associated with a high incidence of myocardial failure. It is tempting to speculate that increased calstabin2 expression is a secondary and compensatory mechanism in response to hyperactive RyR2 function (possibly because of triadin deficiency) in Great Danes. Further studies investigating the complex interactions between RyR2, its associated regulatory proteins, and Ca2+ flux are needed.
There are several important limitations to our study. Only a small number of dogs were investigated, reflecting the difficulty in obtaining tissue samples from client-owned dogs in a timely fashion posteuthanasia. In our microarray analysis, we accounted for the low patient numbers by utilizing 2 analysis algorithms and insisting that transcripts be found differentially expressed by both.15 We believe that low patient numbers affected our ability to find a significant difference (P=.053) in triadin expression by RT-qPCR, and further studies involving a larger cohort of dogs are needed. Our control group, while consisting of large breed dogs, was not comprised of Great Danes, and important breed differences may exist, although it is unlikely that the magnitude of fold-change seen in our study were purely due to breed. All 3 Great Danes were receiving cardiac medications and the influence of these drugs on triadin and calstabin2 expression is unknown. We note that this limitation is similarly encountered in expression studies involving humans.10,11 There are limitations specifically associated with microarray-based studies.52,53 The canine microarray used in our study contains a substantial number of genes; however, it is not utterly inclusive of the entire canine genome, and the knowledge gained from the study may be incomplete. Microarray experiments yield a tremendous amount of data, and discovering connections between the hundreds of differentially expressed genes can be difficult. Moreover, the exact functions of many of the genes present on the microarray are unknown. We only sampled a single area of the left ventricle, and important regional differences in expression of cardiac proteins may exist.54 Despite these limitations, we find it remarkable that of >38,000 genes analyzed, the 2 genes with the greatest degree of differential regulation are critical to proper RyR2 function. We caution that the changes found in our study may represent a secondary finding that is caused by the heart failure phenotype, effect of medical treatment, transcriptional abnormalities of other cardiac genes, or other unknown causes. Further molecular and genetics studies are required to determine the association between triadin and calstabin2 abnormalities on the development and progression of DCM in Great Danes. Finally, evaluation of expression patterns remains several steps removed from the end activity of the translated protein, and further studies evaluating triadin and calstabin2 protein content, RyR2 function, and Ca2+ flux are required.
In conclusion, oligonucleotide microarray analysis indicates significant downregulation of triadin and upregulation of calstabin2. These molecules are critically important for proper functioning of RyR2 and the CRU. To the authors' knowledge, this study is the first to report a molecular abnormality in Great Danes with DCM. Alternations in RyR2 function may play an important role in Great Danes with DCM and further study into the molecular mechanisms involving triadin and calstabin2 is warranted.
aAmbion/Applied Biosystems, Austin, TX
bGeneChip Canine Genome 2.0 Array, Affymetrix Inc, Santa Clara, CA
cBasic Local Alignment Search Tool (BLASTn), National Center for Biotechnology Information, Bethesda, MD. Available at: http://www.ncbi.nlm.nih.gov/BLAST/. Accessed December 3, 2007.
Supported by a grant from the American Kennel Club-Canine Health Foundation.
The authors thank Marcy Kuentzel and John Tine for their assistance.