G Enetic Basis for Chamber-Specific Ventricular Phenotypes in the Rat

Cardiovascular Research 57 (2003) 477–485 www.elsevier.com/locate/cardiores

G enetic basis for chamber-speciﬁc ventricular phenotypes in the rat infarct model Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 Sumeet S. Chugha,* , Stacey Whitesel a , Mark Turner b , Charles T. Roberts Jr. b , Srinivasa R. Nagallab aDivision of Cardiology, Department of Medicine, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, OR 97239, USA bDepartment of Pediatrics, Oregon Health and Science University, Portland OR, USA Received 22 May 2002; accepted 25 September 2002

Abstract Background: We, and others, have previously reported a strong correlation between increased inter-ventricular dispersion of repolarization and the occurrence of fatal arrhythmia in animal models of CHF. The existence of this and other such distinct electrophysiologic phenotypes in right (RV) vs. left ventricles (LV) could be explained by chamber-specific patterns of gene expression. Methods: We employed microarray gene profiling of 13 824 sequence-verified, nonredundant rodent cDNAs to compare myocardial gene expression in RV vs. LV of rats with surgically induced myocardial infarction (MI: n53) and in sham-operated animals (Sham: n53). Results: Significant LV infarction (3264% LV) and severe CHF were observed in all MI animals at 4 weeks. In Sham animals, 937 genes exhibited significant differential expression in RV vs. LV myocardium. In MI animals, 1158 genes exhibited significant differential expression in RV vs. LV. Of those genes exhibiting significant differential expression, only 241 were common to both Sham and MI animals. Differentially expressed genes included those involved in signal transduction, cell growth and maintenance, and apoptosis. Genes with potential roles in altered dispersion of repolarization included voltage-dependent Ca211 channel g subunit (MI 8-fold↑) and K inwardly rectifying channel subfamily J, member 10 (MI 6-fold↓). Gap junction membrane channel protein a 4 (MI 6-fold↓) and cardiac troponin I (MI 8-fold↓) were also significantly differentially expressed. Inter-ventricular comparisons revealed significantly greater alterations in gene expression vs. intra-ventricular comparisons. Conclusions: Microarray gene profiling has revealed candidate genes, some of them novel, which may account for chamber-specific ventricular electrophysiologic phenotypes, both in physiologic as well as in arrhythmogenic states such as CHF.  2003 European Society of Cardiology. Published by Elsevier Science B.V. All rights reserved.

Keywords: Arrhythmia (mechanisms); Gene expression; Heart failure; Infarction; Ion channels; Repolarization; Sudden death

1 . Introduction for arrhythmogenic states in animal models, as well as in human subjects. Vos et al. described a strong correlation The right and left cardiac ventricles have distinct between ventricular proarrhythmia and increased inter- morphologic, hemodynamic and electrophysiologic prop- ventricular dispersion of repolarization in a canine model erties. In normal mammalian hearts, these discrete prop- of chronic AV block [1]. We have reported similar observa- erties are of functional significance, and the chambers have tions in the canine model of pacing-induced heart failure a complementary role in cardiac excitation and myocardial [2]. In addition, recent studies have implicated elec- contractile function. However, disease states may result in trophysiologic changes specific to the RV in the patho- an imbalance of chamber-specific properties of the right physiology of Brugada syndrome [3,4]. and left ventricles. Such alterations are well documented Knowledge of the genetic basis of such ventricular- specific phenotypes in normal and disease states may *Corresponding author. Fax: 11-503-494-8750. E-mail address: [email protected] (S.S. Chugh). Time for primary review 23 days.

0008-6363/03/$ – see front matter  2003 European Society of Cardiology. Published by Elsevier Science B.V. All rights reserved. PII: S0008-6363(02)00703-4 478 S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 significantly advance the understanding of ventricular MI animals. Separate microarrays were used for each arrhythmogenic mechanisms. Gene expression profiling tissue type (i.e. RV Sham, LV Sham, RV MI and LV MI). using cDNA microarrays is a useful tool for the simulta- In addition, duplicate microarrays were hybridized and neous and comprehensive evaluation of myocardial gene analyzed for each tissue type. A 10-mg amount of total expression [5]. The rat model of surgically-induced RNA underwent reverse transcription to generate myocardial infarction manifests with severe dilated car- biotinylated fluorescent cDNA probes. These probes were diomyopathy and an early incidence of fatal arrhythmia hybridized in 50% formamide, 53 Denhardt’s and 63SSC [6,7]. Having previously characterized the hemodynamic in a humidity chamber at 45 8C overnight. Streptavidin– and structural remodeling that occurs in this model during HRP and Cy5 tyramide (New England Nuclear, http:// the progression of heart failure [8], we employed mi- lifesciences.perkinelmer.com/ ) were used for signal de- croarray gene profiling to study chamber-specific ventricu- tection. Slides were scanned using a ScanArray 4000 slide lar gene expression patterns, with a focus on known/ reader (GSI Lumonics, http://www.gsilumonics.com/ ). Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 putative cardiac rhythm-related genes. Detailed descriptions of microarray procedures are available at http://medir.ohsu.edu/|geneview, under Protocols.

2 . Methods 2 .3. Data analysis 2 .1. Creation of the animal model and harvest of tissue Spot and grid parameters from array images were The Oregon Health and Science University Institutional extracted using IMAGENE 4.1 software (BioDiscovery, http:// Animal Care and Use Committee approved all aspects of www.biodiscovery.com/ ). The mean fluorescence intensi- the study, and the investigation conforms with the Guide ty of the signal pixels for each spot was corrected for local for the Care and Use of Laboratory Animals published by background fluorescence by subtracting the median fluo- the National Institutes of Health (NIH Publication No. rescence of the background pixels. Arrays were assessed 85-23, revised 1996). Surgical ligation of the left coronary for quality by visual inspection as well as determination of artery via a small left thoracotomy (Charles River Labs., detectable gene expression using control spots printed on Raleigh, NC, USA) was performed in nine male Sprague– all arrays. Sequences were determined to be detectable Dawley rats (MI), using previously described techniques using frequency signal (frequency difference significant at [8]. Three animals underwent sham surgery, i.e. left P,0.0005 by contingency table analysis). Determination thoracotomy was performed without actual ligation of the of differential expression in the detectable expressed genes left coronary artery. Harvest of tissue in all animals was was done using ARRAYSTAT (Imaging Research, http:// performed at 4 weeks following surgery. Following liberal www.imagingresearch.com/ ). The data were analyzed general anesthesia with Nembutal and heparin injection using the Proportional Model with Offsets option, which (1000 units, intra-peritoneal), the heart was excised using a log transforms the data (base 10) and screens replicate median sternotomy. In five MI animals, myocardial infarct measurements for outliers. Global differences in labeling size was determined by planimetry of serial sections taken and hybridization efficiency between arrays were normal- from each heart at 4 weeks following surgery. In the ized in a two-step process that centered data within remaining three MI animals and three Sham animals, the replicate arrays for a given condition, and then normalized RV and the infarcted and noninfarcted regions of the LV the mean data for each condition by a scaling factor. were carefully separated (inter-ventricular septum was Global scaling was used in preference to normalization by included in the LV). Tissue was flash-frozen with immedi- the signal intensity of housekeeping genes, because the ate storage at 270 8C. Gene expression comparisons were housekeeping genes that were previously used to normalize made between RV myocardium and LV noninfarcted RT-PCR expression data typically have strong signals that myocardium (remote from the infarct zone) in both Sham fall at the upper end of the distribution of signals from the and MI animals. microarray [9]. Pairwise comparisons of conditions were made using the z test. Differences were considered statisti- 2 .2. Microarray procedures cally significant if they achieved a nominal significance of P,0.05 after adjusting the cutoff P value for multiple Custom microarrays were prepared by printing 13 834 comparisons by the stepdown Bonferroni method [10]. sequence-verified rodent cDNAs (Research Genetics, Among the 13 834 genes spotted on the array, a set of 120 Huntsville, AL, USA) on silylated (superaldehyde-coated) genes were prespecified as known or potential cardiac glass slides (Telechem, Sunnyvale, CA, USA). Total RNA rhythm-related genes, based on the existing cardiac elec- was extracted using Qiagen RNeasy columns (Qiagen, trophysiology literature. In addition to analyzing all spot- Valencia, CA, USA). Comparisons were performed be- ted genes, the cardiac rhythm-related set was subjected to tween RV Sham, RV MI, LV Sham and LV MI. For each focused investigation regarding potential interrelationships comparison, RNA was pooled from three Sham or three between these genes and their function. S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 479

3 . Results (Sham or MI), only 241 genes were common to both Sham and MI animals. A compressed view of the cluster analysis 3 .1. Evidence of myocardial infarction and significant of all expressed genes showing relative differential expres- heart failure sion in Sham RV, Sham LV, MI RV and MI LV is shown in Fig. 2A. Detailed gene expression data for all cDNAs One of the nine MI animals died within 48 h of surgery. profiled on the microarray are available in tabular form at At 4 weeks, all MI animals had signs of heart failure, with http://medir.ohsu.edu/|geneview, under Rat Myocardial evidence of significant 4-chamber dilated cardiomyopathy Infarction Model: Gene Expression Data. when tissue was harvested. In five separate MI animals that underwent planimetry, the average infarct size was 3264% 3 .4. Functional classification of global gene expression of LV myocardium. The differentially expressed genes represented several Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 3 .2. Global gene expression patterns of RV vs. LV in categories of broad functions. Fig. 3 shows the distribution Sham animals of differentially expressed genes based on their functional classification. From the available literature, a function was Of a total of 13 824 cDNA sequences profiled on the ascribable for 151 (16%) differentially expressed genes in arrays, 12 567 were determined to be detectable for this the Sham comparisons and for 233 genes (20%) in the MI comparison. The majority of profiled genes were not comparisons of RV vs. LV. significantly differentially expressed in RV vs. LV (92.5%, Fig. 1). However, among the remaining genes, 5.8% had 3 .5. Alterations in myocardial gene expression of increased expression in RV vs. LV, and 1.7% genes had individual chambers decreased expression in RV vs. LV. As a result of myocardial infarction, expression of 3 .3. Global gene expression patterns of RV vs. LV in MI profiled genes was altered significantly in MI RV vs. Sham animals RV for 67 genes. When MI LV and Sham LV were compared, there was significant differential expression of In MI animals, a total of 11 639 cDNA sequences were 24 genes. Only 45 of these 91 genes were found to overlap expressed at detectable levels. As in Sham animals, the with those differentially expressed in RV vs. LV com- majority of genes in MI animals (90%) were also ex- parisons. pressed similarly in RV vs. LV. Of the remaining genes, 3.6% were significantly upregulated in RV vs. LVand 6.4% 3 .6. Differential expression of known/putative cardiac were significantly downregulated (Fig. 1). Among the total rhythm-related genes genes that were significantly differentially expressed In Sham animals, of 120 prespecified genes profiled, significant differential expression in RV vs. LV was observed for 11 genes. In MI animals, there was significant differential expression in RV vs. LV for 17 of 120 genes. The majority of these genes did not overlap. However, of the 28 genes (11117), five genes overlapped and were differentially expressed in a similar manner for RV vs. LV in both groups. Fig. 2B illustrates the relative differential expression of selected genes (58 out of 120) in Sham RV, Sham LV, MI RV and MI LV. Table 1 provides details of the extent of relative upregulation or downregulation in RV vs. LV (fold change), as well as the specific function of selected genes from this subgroup.

4 . Discussion

In both Sham and MI animals, a significant proportion of genes profiled using microarrays were expressed differentially in RV compared to LV. In Shams, 5.8% showed Fig. 1. Chamber-specific gene expression. (A) Data from Sham animals increased expression in RV (compared to LV) and 1.7% and (B) from MI animals. The Venn diagram shows relative number of showed decreased expression. In MI animals, these trends genes upregulated in RV vs. LV in both groups of animals. were reversed, and a smaller proportion of genes were 480 S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021

Fig. 2. (A) This cluster analysis (GeneMaths) provides an overall comparison of absolute gene expression (of all genes spotted on the array), between RV Sham, LV Sham, RV MI and LV MI (columns 1–4 from left to right, respectively). The color scale (green–yellow–orange–red) denotes a quantification of gene expression, green indicating least and red the greatest degree of gene expression. (B) This cluster analysis provides a means of finding potential similarities in gene expression trends or cross talk between 58 of the 120 preselected known/putative cardiac rhythm-related genes. The displayed data are based on differential expression of these genes in RV Sham, LV Sham, RV MI and LV MI in columns 1–4 from left to right, respectively. The top of the figure shows the color scale for increasing relative gene expression from a scale of 1–4 (green→red). Parentheses to the left of the figure relate families of genes that cluster together based on extent of gene expression, with a scale from 60 to 100 (100 represents strongest cluster relationship). S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 481 Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021

Fig. 2. (continued) 482 S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021

Fig. 3. Functional classification of genes differentially expressed in RV vs. LV. Data in pie diagrams is expressed as a percentage of total genes. upregulated in the RV compared to LV (3.6% upregulated served in the intact normal and infarcted canine heart [12], vs. 6.4% downregulated). Of 120 preselected known/puta- as well as right and left Purkinje strands [13]. More tive cardiac rhythm-related genes, 11 exhibited significant recently, Antzelevitch et al. reported significant differences differential expression in RV compared to LV in Shams, in the magnitude of the monophasic action potential notch and 17 in MI animals, respectively. However, when altered in right vs. left canine ventricular epicardium, which were myocardial gene expression was evaluated in individual attributable to relative downregulation of the Ca21 -acti- 1 chambers (MI RV vs. Sham RV; MI LVvs. Sham LV), only vated outward K current (Ito) in RV vs. LV [14]. Our data 91 genes (0.01% of total genes profiled) were found to indicate the existence of significant inter-ventricular ge- exhibit significant differential expression. netic diversity in the normal heart, which may account for The presence of electrophysiologic heterogeneity be- some of the unique electrophysiologic properties of each tween RV and LV has been demonstrated in the normal chamber. For instance, genes affecting Ca21 regulation and heart. Using an in situ canine model of digitalis toxicity, transport that were differentially expressed in RV vs. LV Damato et al. first observed quantitative differences be- included Ca21 -transporting ATPase (Atp2a2, upregulated tween automatic pacemaker rates of right and left bundle 7-fold in RV Sham) and calmodulin 2 (Calm2, upregulated branch systems [11]. Subsequently, findings of functional 9-fold in RV Sham). dissimilarity between right and left ventricles were ob- In cardiac disease states, inter-ventricular elec-

Table 1 Details of selected cardiac-rhythm related genes Gene name Symbol SHAM RV/LV MI RV/LV Biologic process Ca21 -ATPase (cardiac sarcoplasmic reticulum) Atp2a2 7.12* 2.36 Ca21 ion transport, cation transport ATPase, Na11 /K transporting, b3 polypeptide Atp1b3 0.75 0.07* Monovalent inorganic cation transport Ca21 channel b1 subunit Cacnb1 0.34 0.19* Ca21 transport Ca21 channel, voltage-dependent, g subunit 6 Cacng6 0.92 8.18* Ca21 transport Cl2 channel 3 Clcn3 1.09 0.18* Cl2 transport Gap junction membrane channel protein a 4 Gja4 1.02 0.17* Cell communication, cell–cell signaling Matrix metalloproteinase 24 Mmp24 12.66* 2.35 Extra-cellular matrix degradation K1 channel, subfamily K, member 3 Kcnk3 6.5* 4.05 K1 transport K1 inwardly-rectifying channel, subfamily J, member 10 Kcnj10 0.6 0.17* K1 transport Procollagen, type I, a 1 Col1a1 0.25 0.16* Cell adhesion Procollagen, type I, a 2 Col1a2 2.54 4.70* Cell adhesion Procollagen, type IV, a 1 Col4a1 0.42 0.20* Cell adhesion Troponin I, cardiac Tnni3 0.69 0.13* Cytoskeletal organization, muscle contraction and development Numbers represent fold-changes in gene expression RV vs. LV in Sham and MI animals. *, Statistically significant differential gene expression for the specified comparison, i.e. Sham RV vs. LV or MI RV vs. LV. S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 483 trophysiologic heterogeneity has been associated with related to extra-cellular discontinuities in cardiac muscle, arrhythmogenicity. In a canine model of chronic AV-block, include gap junction membrane channel protein a4 (Gja4; Vos et al. observed a strong correlation between acquired downregulated 6-fold in MI RV vs. LV), matrix metallop- torsade-de-pointes arrhythmia and inter-ventricular disper- roteinase 24 (Mmp24; upregulated 13-fold in Sham RV vs. sion of repolarization [1]. They proposed that inter-ven- LV but no significant differential expression in MI), tricular dispersion of repolarization be added to bradycar- procollagen, type I, a1 (Col1a1; downregulated 6-fold in dia, prolonged repolarization and early afterdepolarizations MI RV vs. LV) and procollagen, type I, a2 (Col1a2; as an important factor in the initiation of this arrhythmia. upregulated 5-fold in MI RV vs. LV). Also, in the present We, and others, have reported similar findings in canine study, when Mmp 24 expression was compared in Sham models of tachycardia-induced dilated cardiomyopathy and LV and MI LV, there was upregulation in MI LV (Fig. 2B). fatal arrhythmia [2,15]. With ibutilide administration, we Upregulation of matrix metalloproteinase subtypes was observed an increased dispersion of left-right ventricular previously shown to occur in the rat MI model, but has Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 action potential duration in CHF, which correlated with only been evaluated in LV myocardium [25]. increased incidence of polymorphic ventricular tachycardia LV hypertrophy has long been associated with an [2]. In the present study, surgically induced MI and increased propensity for genesis of ventricular arrhythmia progression to heart failure resulted in significant altera- [26,27], but the molecular pathways leading to fatal tions in genetic expression in both LV and RV. While arrhythmia remain uncertain. In a mouse model of aortic significant overlap between the chambers continues to stenosis and hypertrophy, the expression of the cardiac exist, the trend for chamber-specific genetic expression sarcoplasmic reticulum Ca21 -ATPase gene (Atp2a2; en- reverses in heart failure. There is a shift in chamber- coding the protein isoforms SERCA2a and SERCA2b) specific gene expression toward relative downregulation of may be a determinant of mortality. Lorell et al. observed a genes in the RV compared to the LV. Gene expression significantly lower mortality in transgenic animals express- profiling identified several potential candidate genes that ing SERCA2a gene compared to wild-type mice [28]. In may account for chamber-specific alterations in repolariza- patients with heart failure who undergo reverse remodel- tion following myocardial infarction. These include the ing, SERCA2A can exhibit differential expression in RV potassium channel, subfamily K, member 3 (Kcnk3; vs. LV. Barbone et al.reported that structural reverse-re- significantly upregulated in RV in Sham, but not MI), modeling and hemodynamic benefit from left-ventricular Ca21 channel b1 subunit (Cacnb1; downregulated 5-fold assist devices correlated with increased differential expres- in RV MI, not Sham) and voltage-dependent Ca21 channel, sion of normalized SERCA2A content in RV vs. LV [29]. g subunit 6 (Cacng6; upregulated 8-fold in RV MI, not These observations bear an interesting similarity to the Sham). A previous study in the rat infarct model detected findings in the present study, with Atp2a2 upregulation differential expression of .200 genes (out of approximate- (7-fold) observed in RV vs. LV of Sham animals, but no ly 4000 genes profiled) following MI [16]. However, this chamber-specific expression in MI animals. analysis was restricted to the LV/interventricular septum, and comparisons with RV myocardial gene expression were not performed. Interestingly, in the present study, 5 . Limitations while myocardial infarction caused altered gene expression within individual chambers, the number of genes affected The technique of cDNA microarray gene profiling has were small (n591; 0.01% of total) compared to inter- evolved to universal acceptance [5,30–33]. However, chamber gene expression (Sham RV vs. LV 7.5% of total; expression data, i.e. mRNA levels, may not accurately MI RV vs. LV 10% of total). reflect protein levels, since translational control and post- Discontinuous electrical propagation resulting from al- translation processing may occur [25]. In turn, the expres- tered distribution of gap junctions is an important factor in sion of a protein may not always have a physiological or promoting reentrant ventricular arrhythmias in cardiac pathological consequence. However, the identification of disease states [17,18]. Significant alterations in the number candidate genes on the basis of quantification of their and arrangement of gap junctions have been observed in expression level and the subsequent application of this ischemic heart disease [19–22]. Nakajima et al. recently knowledge to disease gene identification and transgenic reported differential deposition of collagen in atrium vs. technology is a logical step toward realization of the new ventricle in transgenic mice with elevated levels of active genes-to-mechanisms paradigm [34–36]. The present in- myocardial transforming growth factor-b1 [23]. In the rat vestigation examined gene expression at a single time- MI model, Cleutjens et al. observed differential collagen point (4 weeks) following myocardial infarction. It is remodeling in RV vs. LV with respect to time after likely that the nature and extent of gene expression vary infarction, as well as qualitative and quantitative differ- depending on the time duration at which it is evaluated. In ences [24]. Our data confirm the existence of such the present study, we included the interventricular septum chamber-specific differences between the two ventricles, within the LV analysis. This was based on the findings of a especially in MI animals. Candidate genes potentially previously published microarray study in a similar model, 484 S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 which did not find significant changes in gene expression experimental myocardial infarction: beneficial effects of long-term in interventricular septum vs. LV free wall [16]. However, therapy with captopril. Circulation 1985;72(2):406–412. [7] Opitz CF, Mitchell GF, Pfeffer MA, Pfeffer JM. Arrhythmias and others have identified region-specific gene expression in death after coronary artery occlusion in the rat. Continuous tele- comparisons of LV free wall vs. interventricular septum metric ECG monitoring in conscious, untethered rats. Circulation [37–39]. Finally, Ito current is an important contributor to 1995;92(2):253–261. repolarization duration, but expression of encoding genes [8] Anand IS, Liu D, Chugh SS et al. Isolated myocyte contractile was not evaluated with the microarrays employed in the function is normal in postinfarct remodeled rat heart with systolic dysfunction. Circulation 1997;96(11):3974–3984. present study. [9] Lee PD, Sladek R, Greenwood CM, Hudson TJ. Control genes and variability: absence of ubiquitous reference transcripts in diverse mammalian expression studies. Genome Res 2002;12(2):292–297. 6 . Conclusions [10] Hochberg Y. A sharper Bonferroni procedure for multiple tests of significance. Biometrika 1988;75:800–802. Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 [11] Damato AN, Lau SH, Bobb GA. Digitalis-induced bundle-branch We have identified significant differences in genetic ventricular tachycardia studied by electrode catheter recordings of expression between right and left ventricles, which may the specialized conducting tissues of the dog. Circ Res explain some of the distinct morphologic and physiologic 1971;28(1):16–22. properties of the two chambers in the normal rat heart. In [12] Hope RR, Scherlag BJ, El-Sherif N, Lazzara R. Hierarchy of ventricular pacemakers. Circ Res 1976;39(6):883–888. our model, myocardial infarction resulted in further altera- [13] Reiser J, Anderson GJ. Differences in automaticity between Purkinje tion of chamber-specific gene expression patterns. Inter- strands from right and left dog ventricle. Am J Physiol chamber differences in gene expression were observed to 1980;239(2):H247–251. be significantly greater than intra-chamber alterations, [14] Di Diego JM, Sun ZQ, Antzelevitch C. Ito and action potential notch underscoring the potential importance of understanding are smaller in left vs. right canine ventricular epicardium. Am J Physiol 1996;271(2 Pt 2):H548–561. interactions in right vs. left ventricular pathophysiology. [15] Chen YJ, Hsieh MH, Chiou CW, Chen SA. Electropharmacologic Detailed characterization of this inventory of candidate characteristics of ventricular proarrhythmia induced by ibutilide. J genes, which includes a group of known/putative cardiac Cardiovasc Pharmacol 1999;34(2):237–247. rhythm-related genes, may be useful in understanding the [16] Stanton LW, Garrard LJ, Damm D et al. Altered patterns of gene role of chamber-specific electrophysiologic properties in expression in response to myocardial infarction. Circ Res 2000;86(9):939–945. normal as well as arrhythmogenic states such as heart [17] Spach MS, Boineau JP. Microfibrosis produces electrical load failure. variations due to loss of side-to-side cell connections: a major mechanism of structural heart disease arrhythmias. Pacing Clin Electrophysiol 1997;20(2 Pt 2):397–413. [18] Spooner PM, Joyner RW, Jalife J. Discontinuous conduction in the A cknowledgements heart. Armonk, NY: Futura, 1997. [19] Peters NS, Green CR, Poole-Wilson PA, Severs NJ. Reduced content The authors greatly appreciate the technical assistance of connexin43 gap junctions in ventricular myocardium from provided by Patrick Pattee, Matthew Rodland and Jean hypertrophied and ischemic human hearts. Circulation O’Malley. 1993;88(3):864–875. [20] Peters NS. New insights into myocardial arrhythmogenesis: dis- This investigation was funded, in part, by a grant from tribution of gap-junctional coupling in normal, ischaemic and the Collins Medical Trust to SSC. hypertrophied human hearts. Clin Sci (Colch) 1996;90(6):447–452. [21] Severs NJ. Pathophysiology of gap junctions in heart disease. J Cardiovasc Electrophysiol 1994;5(5):462–475. [22] Smith JH, Green CR, Peters NS, Rothery S, Severs NJ. Altered R eferences patterns of gap junction distribution in ischemic heart disease. An immunohistochemical study of human myocardium using laser [1] Verduyn SC, Vos MA, van der Zande J, van der Hulst FF, Wellens scanning confocal microscopy. Am J Pathol 1991;139(4):801–821. HJ. Role of interventricular dispersion of repolarization in acquired [23] Nakajima H, Nakajima HO, Salcher O et al. Atrial but not torsade-de-pointes arrhythmias: reversal by magnesium. Cardiovasc ventricular fibrosis in mice expressing a mutant transforming growth Res 1997;34(3):453–463. factor-b1 transgene in the heart. Circ Res 2000;86(5):571–579. [2] Chugh SS, Johnson SB, Packer DL. Altered response to ibutilide in [24] Cleutjens JP, Verluyten MJ, Smiths JF, Daemen MJ. Collagen a heart failure model. Cardiovasc Res 2001;49(1):94–102. remodeling after myocardial infarction in the rat heart. Am J Pathol [3] Yan GX, Antzelevitch C. Cellular basis for the Brugada syndrome 1995;147(2):325–338. and other mechanisms of arrhythmogenesis associated with ST- [25] Peterson JT, Li H, Dillon L, Bryant JW. Evolution of matrix segment elevation. Circulation 1999;100(15):1660–1666. metalloprotease and tissue inhibitor expression during heart failure [4] Takagi M, Aihara N, Kuribayashi S et al. Localized right ventricular progression in the infarcted rat. Cardiovasc Res 2000;46(2):307– morphological abnormalities detected by electron-beam computed 315. tomography represent arrhythmogenic substrates in patients with the [26] Kohya T, Kimura S, Myerburg RJ, Bassett AL. Susceptibility of Brugada syndrome. Eur Heart J 2001;22(12):1032–1041. hypertrophied rat hearts to ventricular fibrillation during acute [5] Brazma A, Hingamp P, Quackenbush J et al. Minimum information ischemia. J Mol Cell Cardiol 1988;20(2):159–168. about a microarray experiment (MIAME)-toward standards for [27] Cameron JS, Myerburg RJ, Wong SS et al. Electrophysiologic microarray data. Nat Genet 2001;29(4):365–371. consequences of chronic experimentally induced left ventricular [6] Pfeffer MA, Pfeffer JM, Steinberg C, Finn P. Survival after an pressure overload. J Am Coll Cardiol 1983;2(3):481–487. S.S. Chugh et al. / Cardiovascular Research 57 (2003) 477–485 485

[28] Ito K, Yan X, Feng X et al. Transgenic expression of sarcoplasmic model human diseases. Proc Nat Acad Sci 1998;95(19):11371– reticulum Ca21 atpase modifies the transition from hypertrophy to 11376. early heart failure. Circ Res 2001;89(5):422–429. [35] Priori SG, Barhanin J, Hauer RNW et al. Genetic and molecular [29] Barbone A, Holmes JW, Heerdt PM et al. Comparison of right and basis of cardiac arrhythmias: impact on clinical management Part left ventricular responses to left ventricular assist device support in III. Circulation 1999;99(5):674–681. patients with severe heart failure: a primary role of mechanical [36] Spach MS. Mechanisms of the dynamics of reentry in a fibrillating unloading underlying reverse remodeling. Circulation myocardium: developing a genes-to-rotors paradigm. Circ Res 2001;104(6):670–675. 2001;88(8):753–755. [30] Geraci MW, Moore M, Gesell T et al. Gene expression patterns in [37] Tschope C, Heringer-Walther S, Koch M, Spillmann F, Wendorf M, the lungs of patients with primary pulmonary hypertension: a gene Hauke D et al. Myocardial bradykinin B2-receptor expression at microarray analysis. Circ Res 2001;88(6):555–562. different time points after induction of myocardial infarction. J [31] Lee CK, Weindruch R, Prolla TA. Gene-expression profile of the Hypertens 2000;18(2):223–228. ageing brain in mice. Nat Genet 2000;25(3):294–297. [38] Wickenden AD, Jegla TJ, Kaprielian R, Backx PH. Regional 1 [32] Yang J, Moravec CS, Sussman MA et al. Decreased SLIM1 contributions of Kv1.4, Kv4.2, and Kv4.3 to transient outward K Downloaded from https://academic.oup.com/cardiovascres/article/57/2/477/307422 by guest on 01 October 2021 expression and increased gelsolin expression in failing human hearts current in rat ventricle. Am J Physiol 1999;276(5 Pt 2):H1599– measured by high-density oligonucleotide arrays. Circulation H1607. 2000;102(25):3046–3052. [39] Sandmann S, Yu M, Kaschina E, Blume A, Bouzinova E, Aalkjaer C [33] Schena M, Shalon D, Davis RW, Brown PO. Quantitative moni- et al. Differential effects of angiotensin AT1 and AT2 receptors on toring of gene expression patterns with a complementary DNA the expression, translation and function of the Na11 –H exchanger microarray. Science 1995;270(5235):467–470. and Na12 –HCO3 symporter in the rat heart after myocardial [34] Lee P, Morley G, Huang Q et al. Conditional lineage ablation to infarction. J Am Coll Cardiol 2001;37(8):2154–2165.