| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Heart and Vascular Research Center, and The Department of Biomedical Engineering, MetroHealth Campus, Case Western Reserve University, Cleveland, Ohio 44109
Submitted 12 April 2004 ; accepted in final form 9 June 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
TM), effective transmural space constant (
TM), and transmural gradients of action potential duration (APD). Absolute Cx43 expression in failing myocardium, quantified by confocal immunofluorescence, was uniformly reduced (by 40 ± 3%, P < 0.01) compared with control. Relative Cx43 expression was heterogeneously distributed and lower (by 32 ± 18%, P < 0.05) in the subepicardium compared with deeper layers. Reduced Cx43 expression in heart failure was associated with significant reductions in intercellular coupling between transmural muscle layers, as evidenced by reduced TM (by 18.9 ± 4.9%) and TM (by 17.2 ± 1.4%; P < 0.01) compared with control. Heterogeneous transmural distribution of Cx43 in failing myocardium was associated with lower subepicardial TM (by 12 ± 10%) and TM (by 13 ± 7%), compared with deeper transmural layers (P < 0.05). APD dispersion was greatest in failing myocardium, and the largest transmural APD gradients were consistently found in regions exhibiting lowest relative Cx43 expression. These data demonstrate that reduced Cx43 expression produces uncoupling between transmural muscle layers leading to slowed conduction and marked dispersion of repolarization between epicardial and deeper myocardial layers. Therefore, Cx43 expression patterns can potentially contribute to an arrhythmic substrate in failing myocardium.
gap junctions; repolarization; sudden cardiac death; remodeling; coupling; ion channels
Previously, it was presumed that transmural APD gradients are induced by heterogeneities of ion channel expression. In HF, relatively selective downregulation of outward currents and an increase in inward currents of M cells could be one potential mechanism underlying the formation of steep APD gradients within the Epi-Mid interface. However, recent data (21) suggest that transmural ionic outward currents are uniformly reduced in HF, which suggests an alternative mechanism for the maintenance of heterogeneous gradients of repolarization.
Computer simulations have suggested that intercellular coupling through gap junctions is a possible mechanism responsible for maintaining electrophysiological heterogeneities between transmural muscle layers (20). Gap junctions are comprised of intercellular channels that permit the transfer of electrical current and small molecules between neighboring cells. The constituent proteins of gap junction channels, connexins, play a critical role in impulse propagation and electrical synchronization between myocytes. We previously demonstrated that connexin43 (Cx43) expression, the principal gap junction protein found in ventricular myocardium (37, 38, 40, 41), is heterogeneously distributed across the left ventricular wall in normal canine myocardium (26). Importantly, relatively reduced subepicardial Cx43 expression was associated with reduced transmural conduction velocities and the largest APD gradients within the Epi-Mid interface. One would expect that reduced Cx43 protein expression and intercellular uncoupling in HF, independent of ion channel dysfunction, can enhance transmural APD heterogeneities (39). Therefore, we hypothesized that transmural electrophysiological heterogeneities are maintained in HF as a result of regional uncoupling caused by reduced and inhomogeneous expression of the principal ventricular gap junction protein (Cx43) across the ventricular wall.
To test this hypothesis, we developed a technique for comparing, in detail, the spatial distribution of relative Cx43 protein expression to membrane voltage across the transmural wall of the pacing-induced failing canine left ventricle. We demonstrate that in HF, relatively reduced Cx43 expression produces transmural uncoupling as evidenced by reduced transmural conduction velocity and space constant, which in turn forms a substrate for transmural APD gradients. Therefore, Cx43 expression patterns can potentially contribute to arrhythmic substrates that are dependent on transmural inhomogeneities of repolarization.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Canine Model of HF
Heart failure was produced in 5 dogs by rapid pacing for 4–6 wk as described previously (1, 8, 15). Briefly, after implantation of a permanent pacemaker, the right ventricular apex of adult male mongrel dogs was paced at 240 beats/min for 4–6 wk. Electrocardiograms were monitored regularly to confirm consistent pacing capture. Echocardiograms were measured at baseline and after initial clinical symptoms of HF were evident to ensure that significant systolic contractile dysfunction (fractional shortening <30%) was present. Clinical signs of HF manifested in all dogs included anorexia, lethargy, exercise intolerance, ascites, tachypnea, and muscle wasting.
Canine Wedge Preparation
To assess the functional consequences of Cx43 expression patterns across the transmural wall, we developed a system for optically mapping action potentials from cells spanning the entire transmural wall of the arterially perfused canine wedge preparation as described previously (3, 44). Briefly, hearts were excised from 5 male mongrel dogs in HF and 8 control dogs weighing 20–25 kg. The identical control group was selected a priori for this and an earlier study (26). Wedges of myocardium measuring 
3 x 1.5 x 1 cm were dissected from the anterior or anterolateral, or posterior wall of the LV in proximity to secondary branches of the left anterior descending and circumflex coronary arteries, respectively. Wedges were perfused with oxygenated Tyrode solution composed of (in mmol/l) 129 NaCl, 25.0 NaHCO3, 0.5 MgSO4, 4.0 KCl, 5.5 dextrose, and 1.8 CaCl2. Perfusion pressure was maintained at 50–60 mmHg. Wedges were discarded if collateral arteries shunted significant flow away from the preparation as evidenced by a coronary resistance <1.2 mmHg·ml–1·min–1 (3). Preparations were completely immersed in temperature controlled (36 ± 1°C) perfusate to prevent the formation of intramyocardial temperature gradients. Wedges were stabilized against a flat imaging window by applying a gentle constant pressure via a movable piston, obviating the need for pharmacological suppression of contraction. Preparations were determined to be stable for over 4 h of perfusion, as judged by the stability (±5%) of coronary resistance, APD, and QT interval.
Transmural Optical Mapping System
Previously, we developed an optical action potential mapping system (2, 3, 18, 28) capable of resolving membrane potential changes as small as 0.5 mV with 1-ms temporal resolution from 256 sites simultaneously across the entire transmural surface of the wedge preparation. After staining with the voltage-sensitive dye 4-{
-[2(di-n-butylamino)-6-naphthyl]vinyl}pyridinium (15 µmol/l) by direct arterial perfusion for 10 min, dye was excited by 514 ± 5 nm light emitted by a 250-W tungsten-filament lamp. Fluoresced light was long-pass filtered at 610 nm and focused onto a 16 x 16 element photodiode array (C4675, Hamamatsu) through high numerical aperture photographic lenses using the tandem lens configuration (Nikon 85 mm F/1.4, 105 mm F/2.0) (18). A 768 x 493 pixel charge-coupled device video camera (model TMC-74, PULNiX; Sunnyvale, CA) was used to view and localize the mapping field relative to the anatomic features of the preparation. The anatomic reference points on the wedge were used to precisely align action potential maps with sections of tissue obtained for subsequent measurement of Cx43 distribution.
Functional differences in intercellular coupling between transmural muscle layers were assessed by measuring several electrophysiological parameters within each layer of myocardium that are each dependent on cell-to-cell coupling: APD, transmural APD gradient, conduction velocity, and rise time were assessed during steady-state pacing (2,000 ms cycle length) with the use of an optical resolution of x1.2 corresponding to a total mapping area of 14 x 14 mm. Because the transmural space constant (TM) is on the order of 1 mm, it was required to optically map with a spatial resolution of 0.33 mm (2) sequentially within each muscle layer to obtain TM as a function of transmural distance from the epicardium. These functional measurements were then compared with Cx43 expression within each layer.
Action Potential Analysis
APD was measured as the difference between depolarization and repolarization times as previously described (13, 18). APD was measured from 16 sites per muscle layer (each site equidistant from the epicardial surface) for each of 16 equally spaced transmural muscle layers. The average APD within each layer was calculated to generate a transmural APD profile from epicardium to endocardium. The derivative of the APD profile was used to determine the transmural APD gradient as a function of depth from the epicardium. We used a previously described algorithm (2) to calculate conduction velocity selectively (±15°) in the transmural direction of propagation (TM) as a function of distance from the epicardium of each of the 16 transmural muscle layers. Maximum action potential upstroke velocity (dV/dtmax) was calculated after normalizing action potential amplitudes to 100 mV.
To further determine whether the electrophysiological changes measured in HF could be attributed to an effect of uncoupling, the relatively selective intercellular uncoupler carbenoxolone (34) was administered to normal arterially perfused canine left ventricular (LV) wedges (n = 6). Wedges were paced at 2,000 ms, and APD gradients were quantified from optical maps of wedges with and without carbenoxolone (20–100 µM).
Effective Space Constant Measurement
Previously, we developed a method for measuring the effective cardiac space constant with high-resolution optical mapping as a means for estimating intercellular resistivity in intact tissues (2, 26). Briefly, a subthreshold unipolar stimulus was applied to the myocardial surface through a platinum electrode. The extent of the membrane voltage (Vm) decay from the site of injection of the subthreshold stimulus was measured optically. It has been demonstrated that this approach could be applied to the canine wedge preparation, as the space constant in the transmural direction is independent of electrical loading caused by variations of fiber angle orthogonal to the transmural surface (i.e., rotational anisotropy) (46). The decay of Vm in the transmural direction was fit to an exponential to obtain the effective TM. To ensure reproducibility, TM was measured from three separate sites within each muscle layer and averaged to determine the average TM for each layer. This was repeated every 2 mm from epicardium to endocardium.
Histological Analysis
After each preparation was optically mapped, the wedges were fixed in 10% buffered formalin for 24 h and embedded in paraffin. Sections were cut from the most superficial regions of the transmural surface (corresponding to the optically mapped region) and stained with hematoxylin and eosin stain, as described previously (44). Tissue sections were imaged with a Nikon Eclipse E600 microscope for offline assessment of the extracellular matrix. Inflammatory cells, fiber separation, and deposition of collagen were qualitatively assessed at positions where electrophysiological parameters were measured. In addition, myocyte size was measured along the transmural axis from 50 myocytes within each layer of myocardium where Cx43 was also quantified.
Relative Transmural Cx43 Quantification by Immunofluorescence
Immunofluorescence analysis of Cx43 was performed as previously described (4, 26). Briefly, paraffinized tissue layers were sectioned at a thickness of 5 µm and mounted on gelatin-coated slides. Sections were deparaffinized, placed in citrate buffer, and boiled in a microwave oven for 10 min. The sections were incubated overnight with anti-Cx43 antibodies (Zymed, diluted 1:400) and then incubated with CY3-conjugated goat anti-rabbit IgG (Zymed diluted 1:800) before being examined by laser scanning confocal microscopy (x40 oil immersion lens, airy 1 pinhole). The degree of confocality was kept constant (depth of focus 
602 nm) for each experiment to minimize overlap of Cx43 label. Each layer was analyzed from eight fields to obtain an average Cx43 quantity within a layer. To eliminate artifactual quantification of Cx43 protein, samples were discarded when the imaging plane was not parallel to the long axis of the fibers, as judged by the length-to-width ratios of myocytes <4 (26). Relative Cx43 quantity was defined as the proportion of total myocardial tissue area occupied by Cx43 immunoflourescent signal, as described previously (9). To compare Cx43 expression across experiments, each measurement of Cx43 was normalized to the mean of the total Cx43 measured for an entire experiment.
Statistical Analysis
All HF data were compared with previously published control measurements made with identical techniques. Statistical analysis of the data was performed using a Student's t-test for paired data or a single-factor ANOVA. A P < 0.05 was considered statistically significant. All values are reported as means ± SD, unless otherwise noted.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Subepicardial, midmyocardial, and subendocardial Cx43 expression for normal and HF ventricles is compared in representative immunofluorescence images shown in Fig. 1. Within each transmural layer, Cx43 localizes to the longitudinal ends of individual myocytes in both normal and HF, consistent with normal cellular gap junction localization reported previously (22). This pattern was observed across all experiments. Overall Cx43 signal from two representative experiments (Fig. 1) is reduced in HF compared with normal. Importantly, subepicardial Cx43 related signal in HF is reduced compared with midmyocardial and subendocardial Cx43 signal, consistent with previous findings in normal LV myocardium (26).
|
|
Transmural conduction velocity.
Representative transmural conduction maps during point stimulation of the endocardium in normal and failing myocardium are shown in Fig. 3, left (top and bottom, respectively). Overall, conduction across the transmural surface was slower in HF, as evidenced by the increased number of isochrone lines in the bottom panel as well as greater separation between action potential upstrokes compared with normal myocardium (Fig. 3, right). For all experiments, overall TM was significantly reduced from 39.5 ± 6.7 cm/s in normal to 31.9 ± 4.2 cm/s in HF. Importantly, TM was slowest in the subepicardium of control and HF myocardium, as evidenced by localized crowding of activation isochrones (Fig. 3, left) and discrete gaps between subepicardial action potential upstrokes (Fig. 3, upstrokes D and E). It is also evident from transmural contour maps (Fig. 3, left) that there is some variation in conduction patterns within each transmural muscle layer. In general, when TM was averaged across each transmural muscle layer in failing myocardium, a clear and highly reproducible TM profile was evident for all experiments (Fig. 2B). Failing myocardium exhibited uniformly and significantly slower TM than control myocardium (by 18.9 ± 4.9%) for every transmural layer. Importantly, TM was slowest (by 12 ± 10%) in the subepicardium of failing myocardium compared with deeper layers.
|
TM slowing in the subepicardial layers, the rate of rise of action potential upstrokes recorded from subepicardium was fastest (Fig. 3, right). Specifically, dVm/dtmax (horizontal bars) in HF measured from subepicardial cells was larger than midmyocardial and subendocardial values (by 58 ± 14%). This finding indicates that TM slowing in the subepicardium was attributable to a localized reduction of intercellular coupling in subepicardial layers and not depressed excitability.
Effective TM.
To assess intercellular coupling more directly, the effective TM was measured from multiple transmural layers spanning the ventricular wall. In general, TM was uniformly reduced (by 20 ± 2%, P < 0.01) in failing compared with normal myocardium (Fig. 2C). TM in HF had a highly reproducible transmural profile that closely corresponded to Cx43 expression and TM profiles. Importantly, subepicardial TM in HF was significantly shorter (by 13.2 ± 5.8%) than TM in deeper myocardial layers, further reaffirming that intercellular coupling remained impaired in the subepicardial layer during HF.
Histological analysis. Electrotonic spread of current is not only affected by gap junctions but also by the extracellular matrix and myocyte dimensions. Representative histological sections of transmural tissue from the subepicardium and midmyocardium of failing myocardium are shown in Fig. 4, left. The dimensions of myocytes in the transmural direction (open bars) are similar between different layers of tissue. Across all HF experiments (Fig. 4, right) the average transmural cell dimensions for epicardial, mid-myocardial, and endocardial myocytes was similar (12.4 ± 3.2, 12.4 ± 2.0, and 12.2 ± 2.1 µm, respectively). In addition, there was no gross evidence of invasion of inflammatory cells or fiber separation in portions of HF myocardium where Cx43 expression and electrophysiology was measured. There were also no significant differences in myocyte size between control and HF. Taken together, these data indicate that regional uncoupling between the epicardial and deeper transmural layers cannot be explained by the regional variations of cell dimension or changes in the extracellular matrix.
|
|
TM (by 18 ± 5%), similar to the reduction observed in HF (Fig. 6A). In addition, carbenoxolone significantly reduced the average transmural conduction velocity from control (by 23.8 ± 5.3%) to a similar extent observed in HF (Fig. 6B). The transmural APD profile, measured from a representative experiment before and after carbenoxolone perfusion, is shown in Fig. 7A. During carbenoxolone administration, subepicardial APD decreased, whereas mid- through endocardial APD increased compared with control (Fig. 7A), thereby increasing total APD dispersion to similar values measured in HF. For all experiments, uncoupling significantly increased total transmural APD dispersion by 54 ± 40% (Fig. 7B, left). Moreover, increased dispersion of repolarization was explained specifically by amplification of APD gradients at the Epi-Mid interface (Fig. 7B, left), whereas uncoupling did not substantially affect APD gradients within the Mid zone or the Mid-Endo zone as evidenced by no significant change in the APD profile in Fig. 7A. Therefore, selective uncoupling significantly enhances transmural APD gradients in a pattern very similar to those seen in HF, suggesting relative intercellular uncoupling across the transmural wall can explain amplification and maintenance of APD gradients seen in HF.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
TM and TM, particularly in subepicardial regions where relative Cx43 expression was lowest. Importantly, HF amplified APD gradients to levels exceeding those required for the development of conduction block and reentry (19 and 1), and did so at the Epi-Mid interface, precisely where relative Cx43 expression and intercellular coupling was most diminished. These data suggest that Cx43 expression patterns resulting from disease-induced remodeling of gap junctions can potentially contribute to arrhythmic substrates that are dependent on transmural dispersions of repolarization. HF-induced gap junction remodeling alters transmural intercellular coupling. Cardiac gap junctions play a major role in impulse propagation and have also been implicated in arrhythmogenesis. Previous reports (11, 25, 40) have demonstrated that Cx43 is reduced in a variety of heart failure models. Similarly, in the present study relative Cx43 quantity in failing myocardium was reduced by 40% in every transmural muscle layer compared with control, corresponding well with previous reports (11) of a 50% reduction of Cx43 protein expression and 40% reduction of Cx43 mRNA in human congestive heart failure. However, the significance of Cx43 expression patterns on electrophysiological function and arrhythmogenesis was poorly understood. We developed a system for concurrent high-resolution mapping and relative Cx43 quantification, which allowed, for the first time, a detailed analysis of the effect of HF-induced gap junction remodeling on electrophysiological function. We found that subepicardial Cx43 expression was reduced by 30 ± 18% compared with deeper layers in failing LV, similar to the reduction of Cx43 expression in the subepicardium previously observed in normal canine (26) and murine (42) ventricular myocardium. However, because overall Cx43 expression was markedly reduced in HF, the reduction of Cx43 in subepicardial regions accentuated uncoupling between EPI and deeper myocardial layers.
Because protein expression does not necessarily correlate with protein function, it was important to compare the distribution of Cx43 expression to functional indices of intercellular coupling. In the present study, Cx43 reduction in HF was associated with a 12% reduction in overall TM. Moreover, discrete conduction slowing in failing myocardium was consistently found in subepicardial regions where relative Cx43 expression was lowest. The magnitude of relative conduction slowing and Cx43 remodeling between HF, control, and transmural muscle layers was consistent with theoretical predictions (29). Moreover, regional TM slowing could not be attributed to depressed excitability because conduction slowing was associated with enhanced rather than depressed dV/dtmax of the action potential upstroke (Fig. 3) (32). These data suggest that Cx43 remodeling in HF causes functional uncoupling between muscle layers, particularly in the Epi-Mid interface.
To further exclude the possibility that regional conduction changes were explained by altered excitability rather than uncoupling, we used TM as an index of coupling across the transmural wall (2). TM was reduced in every transmural muscle layer during HF (Fig. 2C). The finding that TM was particularly reduced in subepicardial compared with deeper myocardial layers in HF is consistent with previous findings of decreased intercellular coupling in normal subepicardium (26, 44). It is important to note that TM is dependent on both intra- and extracellular resistivities. Therefore, it is possible that fibrosis and invasion of inflammatory cells associated with HF (6) could also influence TM. Because the model of HF utilized in this study does not produce significant changes in the extracellular matrix common to human HF (14), remodeling of extracellular resistances is unlikely to explain changes in TM observed in the present study. It has also been demonstrated that changes in myocyte geometry can influence functional indexes of coupling, such as conduction (31). While there were no significant differences in myocyte dimensions between control and HF, decreased myocyte cross-sectional area has been reported previously in a similar model of pacing-induced HF (33). Importantly, the rapid pacing model of HF subjects the myocardium to increased metabolic demand as well as significant alterations in mechanical stress, both of which are capable of remodeling myocardium. Therefore, the lack of correspondence between studies is most likely due to model differences.
While Cx43 quantity can influence intercellular coupling, the spatial distribution of Cx43 protein around myocytes may also play a significant role in cell-to-cell coupling. Previously, Cx43 lateralization around cardiac myocyte was reported in human dilated (17) and hypertrophic (30) cardiomyopathies, both of which are associated with fibrosis. In contrast, the present study demonstrates that Cx43 in HF primarily localizes to the transverse-oriented intercalated disks of myocytes in all transmural layers, which was the pattern observed in normal (24) and failing (11) LV myocardium devoid of fibrosis. This raises the interesting possibility that Cx43 redistribution around myocytes may be influenced by remodeling of the extracellular matrix. Importantly, the lack of Cx43 lateralization in this model of HF strongly suggests that functional differences in cell-to-cell coupling were attributable to reduced Cx43 quantity rather than altered Cx43 localization around myocytes.
Although somewhat controversial, intercellular coupling may also depend on the state of Cx43 phosphorylation (5) and/or the expression of Cx45, which was shown to be upregulated in HF by some (43) but not others (11). The very close fit of relative Cx43 quantity and conduction velocity measurements to theoretical predictions (29) suggests that compensatory expression of other connexins, if present, may only have minor functional consequences in this model of HF. However, the importance of factors influencing gap junction conductance warrants further study as experimental models continue to explore the relationship between human HF and arrhythmogenesis.
HF-induced gap junction remodeling causes arrhythmogenic gradients of repolarization. Previously, we found that HF is characterized by the development of large gradients of repolarization, specifically at the Epi-Mid interface, which underlied the mechanism of conduction block and reentrant arrhythmogenesis in this model of HF (1). In the present study we hypothesized that arrhythmogenic repolarization gradients are maintained across the transmural wall as a result of regional uncoupling caused by HF-induced gap junction remodeling. One would expect that transmural APD gradients arise, on the one hand, from intrinsic differences in ion channel expression and function between myocytes that span the ventricular wall. On the other hand, intercellular coupling between transmural layers is expected to attenuate transmural APD gradients. Therefore, uncoupling between transmural muscle layers should amplify gradients of repolarization, even in the absence of any HF-induced changes in ion channel expression and function (39). Because HF is a complicated disease characterized by ion channel and gap junction remodeling, the question remains whether intercellular uncoupling alone in the absence of ion channel remodeling is sufficient to create APD gradients equal to or greater than the arrhythmogenic gradients observed in HF. We found that uncoupling with carbenoxolone caused transmural APD gradients to increase significantly and to specifically localize within the Epi-Mid interface, which is precisely where APD gradients were maximal in HF. These findings suggest that uncoupling induced by Cx43 remodeling may explain amplification of transmural repolarization gradients observed in failing myocardium (1). A potential limitation to this analysis is the fact that, in addition to uncoupling, carbenoxolone may have other effects presumably due to its detergent like activity and metabolic inhibition (16, 34).
While HF-induced gap junction remodeling may explain arrhythmogenic gradients of repolarization, it is important to consider the possibility that regional remodeling of ion channels could also play a role. However, using a model of HF identical to this study, Li et al. (21) found that the intrinsic heterogeneity in APD between myocytes that span the transmural wall was not increased in HF. This suggests that transmural repolarization gradients characteristic of HF cannot be explained by heterogeneous remodeling of ion channels.
The present study suggests that regional gap junction remodeling across the failing left ventricular wall is an important mechanism for maintaining pathophysiological inhomogeneities of repolarization in syncytial preparations. It is also important to emphasize that intercellular uncoupling by HF slowed conduction (Fig. 2). Conduction slowing itself is arrhythmogenic and has been shown to be an important independent prognosticator of risk in patients with severe left ventricular dysfunction (45). Therefore, HF-induced gap junction remodeling may have a multitude of electrophysiological consequences, which promote arrhythmogenesis. These results suggest the interesting possibility that gap junctions may be important targets for future therapies aimed at reducing the risk of arrhythmias in HF (12).
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
P < 0.05) in HF myocardium compared with control. Subepicardial Cx43 expression is significantly reduced (*P < 0.05) compared with deeper layers for both control and HF myocardium. Transmural conduction velocity (
). Epicardial cell types have shortest APD compared with mid- and endocardial cell types. Uncoupling by carbenoxolone shortens epicardial APD, while increasing mid and endocardial APD (
). B, left: overall transmural APD dispersion increases for all experiments during HF and carbenoxolone administration. Right, Epi-Mid APD gradients are significantly increased in HF and during carbenoxolone administration compared with control (*P < 0.05).