NFATc3-dependent loss of I(to) gradient across the left ventricular wall during chronic beta adrenergic stimulation.

I to is a voltage-gated transient outward K + current that contributes to the early repolarization phase of the ventricular action potential (AP). In mice, I to is produced by a heterotetramer of pore-forming Kv4.2 and Kv4.3 subunits with associated accessory KChIP2 subunits. An interesting feature of the mammalian heart is that I to density is larger in epicardial (EPI) than in endocardial (ENDO) myocytes [ 1 – 4 ]. This transmural gradient of I to is important for normal ventricular repolarization [ 3 , 5 , 6 ]. Recent studies have examined the molecular mechanisms underlying differential I to density in ENDO and EPI cells. In the mouse ventricle, where a transmural gradient in KChIP2 and Kv4.3 is absent [ 7 ], differential Kv4.2 expression presumably underlies heterogeneous I to density [ 4 , 7 , 8 ]. Diastolic and systolic [Ca 2+ ] i also varies across the left ventricular wall; it is higher in ENDO than in EPI myocytes [ 9 , 10 ]. These differences in [Ca 2+ ] i have important implications. Indeed, it was recently demonstrated that activation of the transcription factor NFATc3 by the Ca 2+ -dependent phosphatase calcineurin transduces variations in [Ca 2+ ] I into differences in I to density in ENDO and EPI myocytes [ 11 ]. NFATc3 decreased I to by reducing Kv4 expression [ 11 , 12 ]. [Ca 2+ ] i and calcineurin/NFATc3 activity is higher in ENDO than in EPI myocytes, resulting in lower Kv4 expression and I to density in ENDO cells [ 9 , 11 ]. These data were consistent with a model in which differential patterns of [Ca 2+ ] i , calcineurin, and NFATc3 signaling contribute to regional variations in I to density in the mouse myocardium. A growing body of evidence suggests that changes in [Ca 2+ ] i and NFATc3 activity as well as loss of heterogeneous I to density are associated with the development of hypertrophy and heart failure [ 13 – 18 ]. One important mechanism involved in the regulation of [Ca 2+ ] i in ventricular myocytes under physiological and pathophysiological conditions is the β adrenergic signaling pathway. Although acute activation of β adrenergic receptors (βAR) can increase heart function, multiple studies suggest that their chronic activation causes hypertrophy, electrical remodeling, and arrhythmogenesis [ 19 , 20 ]. Indeed, chronic activation of βAR signaling activates NFATc3 and decreases I to in ventricular myocytes after myocardial infarction [ 12 ]. These changes in I to density are primarily due to changes in Kv4 expression. At present, however, whether chronic βAR signaling activation alters regional variations in calcineurin and NFAT activity as well as Kv4 expression and I to density is unknown. In this study, we tested the hypothesis that differential activation of calcineurin/NFATc3 signaling across the mouse left ventricle contributes to the dissipation of the gradient of I to density between ENDO and EPI during chronic activation of β1 adrenergic signaling. Consistent with this hypothesis, we found that chronic infusion of the βAR agonist isoproterenol (ISO) – which causes hypertrophy [ 21 – 23 ] –increased [Ca 2+ ] i , calcineurin, and NFAT transcriptional activity in EPI, but not in ENDO cells. This resulted in the dissipation of the Kv4.2 and I to gradient between EPI and ENDO cells. Unlike wild type, both NFATc3 knock out (NFATc3 −/− ) and β1 adrenergic receptor knock out (β1AR −/− ) ENDO or EPI myocytes did not undergo changes in I to density during ISO infusion. These data suggest that calcineurin and NFATc3 signaling contributes to the loss of heterogeneous Kv4 expression, and hence I to density, in the mouse left ventricle during chronic β adrenergic stimulation.

We examined [Ca 2+ ] i , calcineurin, NFAT, and I to in ENDO and EPI cells from control and isoproterenol (ISO)-infused mice. We selected this animal model for multiple reasons. First , chronic ISO infusion has been shown to induce reproducible cardiac hypertrophy [ 21 , 25 ]. Second , β adrenergic signaling activates NFAT in ventricular myocytes [ 12 ]. Third , chronic β adrenergic receptor signaling has been linked to hypertrophy, heart failure, and arrhythmogenesis in humans [ 18 , 26 – 29 ]. We recorded I to from ENDO and EPI cells isolated from saline (control) and ISO-infused mice ( Fig. 1 ). Normalization of I to was performed by dividing the current amplitudes by the capacitance (an indicator of cell surface area) of the cells from which they were recorded. Analysis of these capacitance values indicated, as reported by others [ 30 ], that control ENDO cells are larger (154.2±7.7 pF, n =25; p <0.05) than EPI cells (119.5±5.6 pF, n =31). ISO infusion increased the capacitance (i.e. surface area) of EPI cells (169.5±7.8 pF, n =45) to a larger extent than ENDO (175.3±9.3 pF, n =37) cells. I to was evoked by 1 second step depolarizations from a holding potential of −90 mV to voltages ranging from −60 to +40 mV and was defined as the difference between the peak and the sustained current measured at the end of the pulse. I to was isolated from other voltage-gated Na + , Ca 2+ , and K + currents in these cells by pharmacological means as previously described [ 31 – 33 ]. Experiments were performed in the absence of ISO in the external solution. As previously reported [ 1 , 11 ], I to was larger in control EPI than in ENDO cells at most voltages examined ( p <0.05; Fig. 1A ). Indeed, at +40 mV, I to was 62.1±5.6 pA/pF and 21.6±6.1 pA/pF in EPI and ENDO cells, respectively. We found that the amplitude of I to in EPI cells (30.8±5.7 pA/pF at +40 mV) isolated from ISO-infused animals was smaller than that of control EPI cells ( n =7; p 0.05; Figs. 1A–B ). Collectively, these data suggest that chronic infusion of the βAR signaling activator ISO diminishes the transmural I to gradient by selectively decreasing this current in EPI cells. Ventricular myocytes express β1 and β2 adrenergic receptors [ 34 ]. To determine whether ISO infusion decreased I to in EPI cells through the activation of β1AR, we recorded this current in ENDO and EPI cells from control and ISO-infused β1AR knockout mice [ 35 ] (β1AR −/− ; Fig. 2 ). We found that, as in control mice, I to was larger in EPI than in ENDO cells from β1AR −/− mice at most voltages examined, suggesting that basal β1AR activity does not contribute to regional variations in I to density in the left ventricular wall. However, unlike wild type (WT) myocytes (see Fig. 1 above), ISO infusion did not decrease I to in β1AR −/− EPI or ENDO cells ( p >0.05). These data suggest that ISO decreases I to in EPI myocytes via β1 adrenergic receptor mediated signaling. We examined the molecular mechanisms underlying decreased I to density in EPI cells in ISO-infused WT mice. To do this, we determined Kv4.2 and Kv4.3 transcript levels in wild type ENDO and EPI tissue ( Fig. 3 ). As previously reported [ 36 ], Kv4.2 transcript expression levels were higher in control EPI than in ENDO (Kv4.2 ENDO/EPI=0.43±0.5, n =5). Kv4.3 transcript expressionwas similar in control and ISO-infused EPI and ENDO cells ( p >0.05). Consistent with our electrophysiological data, Kv4.2 transcript was about 55% lower ( n =5, p <0.05) in ISO EPI than in control EPI cells. Indeed, Kv4.2 transcript was similar in ISO EPI and control ENDO (Kv4.2 ENDO/EPI=1.0±0.3, n =5). In combination with the electrophysiological data above, our data indicate that sustained activation of β1 adrenergic signaling dissipates the I to gradient by selectively decreasing expression of Kv4.2 transcript in EPI cells. Activation β adrenergic signaling increases [Ca 2+ ] i in ventricular myocytes. Thus, we examined AP-evoked [Ca 2+ ] i transients in ENDO and EPI myocytes loaded with the fluorescent Ca 2+ indicator fluo-4 before and after the application of 100 nM ISO to activate βARs. Action potentials were activated in these cells via field stimulation (1 Hz; Fig. 4 ). Under control conditions, [Ca 2+ ] i transients were larger in ENDO than in EPI cells ( n =8, p< 0.05). In addition, we found that the time to 50% of the amplitude ( T 1/2 ) during the decaying phase of the [Ca 2+ ] i transient was shorter in ENDO ( T 1/2 =304±21 ms, n =12) than in EPI cells ( T 1/2 =360±13 ms, n =15; p < 0.05). Application of ISO (100 nM) increased [Ca 2+ ] i in these cells. However, we found that ISO increased [Ca 2+ ] i to a larger extent in EPI than in ENDO cells. Indeed, although there was a tendency for [Ca 2+ ] i transients to be larger in ISO than in control (i.e. no ISO) ENDO cells, this difference was not statistically significant. Note also that the amplitude of the [Ca 2+ ] i transient in control ENDO, ISO ENDO, and ISO EPI cells was similar ( p >0.05). ISO also increased ( p <0.05) the rate of decay of the [Ca 2+ ] i transient in EPI ( T 1/2 =227±15, n =9) and ENDO myocytes ( T 1/2 =248 ±26 ms, n =12). Note, that in the presence of ISO, the T 1/2 of the [Ca 2+ ] i of ENDO and EPI cells was similar ( p >0.05). These data suggest that β adrenergic signaling activation increases [Ca 2+ ] i in the left ventricular wall predominantly by increasing [Ca 2+ ] i in EPI cells. We investigated the mechanisms underlying the larger increase in the amplitude of the [Ca 2+ ] i transient in EPI than in ENDO myocytes during activation of βAR signaling. Western blot analyses were used to determine expression β1ARs and specific PKA subunits (R1 and II α) as well as the level of phospholamban protein that is phosphorylated at the PKA-specific serine 16 (Ser16 P -PLB) using a phospho-specific antibody [ 37 ] in ENDO and EPI ( Fig. 5 ). Only hearts from wild type animals not infused (i.e. without pumps) with ISO were used in these studies. We found that β1AR protein expression was higher in ENDO than in EPI tissue ( Fig. 5A ; n =5 hearts, p 0.05). We also investigated whether the higher β1AR protein expression in ENDO than in EPI was associated with higher PKA phosphorylation of phospholamban in ENDO than in EPI. As noted above, we used a phospho-specific antibody recognizing Ser16 P -PLB to determine the levels of this protein in ENDO and EPI tissue from hearts perfused using a Langendorff system with saline (i.e. control) or 100 nM ISO for 5 min ( Fig. 5D ). Consistent with the β1AR protein data described above, we found that basal Ser16 P -PLB levels were higher in ENDO than in EPI under control conditions (i.e. saline perfusion; p <0.05), suggesting that basal PKA activity is higher in ENDO than in EPI myocytes. In vivo perfusion with ISO (100 nM) increased Ser16 P -PLB levels in ENDO and EPI ( n =6 hearts, p <0.05). We investigated whether the differences in β1AR expression and PKA-dependent phosphorylation of PLB described above were linked to regional variations in L-type Ca 2+ currents ( I Ca ) and SR Ca 2+ load in EPI and ENDO under control conditions and during activation of βAR signaling ( Fig. 6 ). I Ca was evoked by a 200 ms voltage step to potentials ranging from −30 to +50 mV from the holding potential of −40 mV. As previously reported [ 9 ], under control conditions, the amplitude of I Ca was similar in EPI and ENDO cells ( Fig. 6A ). However, in agreement with our β1AR protein data described above, ISO (100 nM) increased I Ca to a larger extent (≈30%) in ENDO than in EPI cells ( n =6, p <0.05). Consistent with the Ser16 P -PLB data and a previous study [ 9 ] by our group, the amplitude of the [Ca 2+ ] i transient evoked by the application of 20 mM caffeine (i.e. SR Ca 2+ load) was higher in ENDO than in EPI cells under control conditions ( Fig. 6B ). As shown above, acute application of ISO (100 nM), increased the amplitude of the evoked [Ca 2+ ] i transient in EPI cells to a larger extent than in ENDO cells. Note, however, that activation of βAR signaling with ISO increased SR Ca 2+ load in EPI cells (1.52±0.15-fold, n =9) to a larger extent than in ENDO cells (1.04±0.08-fold, n =9, p <0.05). These data suggest that activation of β1AR signaling induces a larger increase in the [Ca 2+ ] i transient in EPI than in ENDO, at least in part, by increasing SR Ca 2+ load to a larger extent in EPI than in ENDO myocytes. In adult ventricular myocytes, changes in [Ca 2+ ] i can alter the expression of Kv4 potassium channels via activation of the calci-neurin-NFATc3 transcriptional signaling pathway [ 11 , 12 ]. Thus, we examined whether chronic ISO infusion activated this signaling pathway in ENDO or EPI cells. First, we determined calcineurin activity in ENDO and EPI cells from control and ISO-infused mice. Consistent with our [Ca 2+ ] i data, calcineurin activity was higher in ENDO than in EPI from control animals ( Fig. 7A ). Calcineurin activity was higher in ISO-infused than in control EPI cells ( p <0.05). Indeed, calcineurin activity was similar in EPI and in ENDO from ISO-infused mice ( p 0.05). Next, we examined NFAT transcriptional activity in control and ISO-infused ENDO and EPI tissue ( Fig. 7B ). In these experiments, we used a transgenic mouse in which luciferase gene expression is driven by multiple NFAT binding elements [ 38 ]. We quantified NFAT activity by using RT-PCR to measure luciferase transcript in ENDO and EPI from saline (control) or ISO-infused mice. Consistent with the calcineurin data above, we measured higher luciferase transcript levels (i.e. NFAT activity) in control ENDO than EPI. Note, however, that while ISO infusion increased NFAT activity (i.e. luciferase expression) in EPI, it failed to do so in ENDO cells. These data suggest that chronic activation of βAR adrenergic signaling decreases regional differences in calcineurin and NFAT activity across the left ventricular wall, by selectively increasing the activity of this signaling pathway in EPI cells. We tested the hypothesis that NFATc3 is required for down-regulation of I to in EPI cells during chronic ISO infusion. To test this hypothesis, we examined I to in ENDO and EPI cells from NFATc3 null (NFATc3 −/− ) mice infused with saline (control) or ISO ( Fig. 8 ). As previously reported [ 11 ], and unlike wild type mice, I to is similar in NFATc3 −/− EPI (56.91±4.21 pA/pF at +40 mV) and ENDO cells (57.61±3.73 pA/pF at +40 mV; p >0.05). Consistent with our hypothesis, EPI myocytes showed no decrease in I to following chronic ISO (49.33±7.64 pA/pF) in comparison to control (55.61±5.73 pA/pF). Likewise, chronic ISO infusion did not decrease I to in NFATc3 −/− ENDO cells (at +40 mV 56.91±4.21 pA/pF in control cells vs. 54.21±9.44 pA/pF in ISO-infused cells). These data suggest that NFATc3 is required for I to downregulation in EPI cells during chronic βAR signaling activation.

Our data suggest that there are important variations in βAR signaling across the mouse left ventricular wall and that these differences contribute to heterogeneous changes in I to density during chronic β1 adrenergic signaling activation in EPI and ENDO. We found significant differences in [Ca 2+ ] i signaling across the left ventricular wall under control conditions and during activation of βAR signaling. Not only is the amplitude of the AP-evoked [Ca 2+ ] i transient larger in ENDO than in EPI myocytes [ 9 ], but it also increases to a larger extent in EPI than in ENDO myocytes during βAR signaling activation. Our data suggest a mechanism for these differences in [Ca 2+ ] i in ENDO and EPI myocytes under control conditions and during βAR signaling. β1AR protein expression and Ser16 P -PLB levels (an indicator of PKA activity) are higher in ENDO than in EPI. Furthermore, SR Ca 2+ load is higher in ENDO than in EPI cells under control conditions. Together, these data suggest that basal PKA activity is higher in ENDO than in EPI cells, which increases PLB phosphorylation and hence SERCA pump activity. Higher SERCA pump activity likely contributes to higher SR Ca 2+ load and SR Ca 2+ release during EC coupling in ENDO than in EPI cells. As expected, we found that I Ca increases in ENDO and EPI during ISO treatment. However, if as noted above, basal PKA activity is higher in ENDO than in EPI why is I Ca amplitude similar in ENDO and EPI [ 9 ]? Although our data do not provide an answer to this difficult question, one intriguing possibility is that the activity of a protein phosphatase that opposes PKA is higher near Ca 2+ channels in ENDO than in EPI, which would increase the threshold for PKA-dependent modulation of I Ca in ENDO compared to EPI. Two recent studies support this hypothesis. Calcineurin activity is higher in ENDO than in EPI [ 11 ]. Furthermore, calcineurin opposes PKA actions on I Ca : calcineurin inhibition increases I Ca in ventricular myocytes [ 39 ]. Future experiments should examine the mechanisms underlying differential modulation of I Ca in ENDO and EPI by PKA. Activation of βAR signaling increased I Ca and Ser16 P -PLB levels in ENDO and EPI. Yet, activation of this signaling pathway increased SR Ca 2+ load in EPI, but not in ENDO myocytes. Consistent with this, acute application of ISO evoked a larger increase in the amplitude of the transient in EPI than in ENDO cells. Indeed, the relatively smaller increase in the [Ca 2+ ] i transient observed in ENDO cells is likely due to the larger increase in I Ca induced by ISO in these cells than in EPI cells. Higher β1AR expression in ENDO than in EPI could contribute to this differential effect of ISO on I Ca . On the basis of these findings, we propose a model for differential [Ca 2+ ] i signaling in ENDO and EPI during βAR signaling. In this model, basal SR Ca 2+ load in ENDO cells is high – compared to EPI – and at a level similar to one observed during activation of βAR signaling in EPI and ENDO cells. PKA-dependent phosphorylation of PLB at serine 16 does not translate into an increase in SR Ca 2+ load in EPI cells. Although the exact mechanisms underlying this are unclear, it is intriguing to speculate that in ENDO cells activation of βAR signaling increases SERCA pump activity and SR Ca 2+ so that there is no net change in SR Ca 2+ load. In this model, SR Ca 2+ load and release during βAR signaling is similar in EPI and ENDO cells. Our observation that the amplitude of the [Ca 2+ ] i transient in ISO EPI and ENDO cells is similar suggests this assumption is reasonable. Our data indicate that chronic activation of βAR signaling has important consequences on [Ca 2+ ] i and I to across the left ventricular wall. Not only did it dissipate transmural differences in [Ca 2+ ] i , but it did also eliminate the I to gradient across the left ventricular wall, presumably by selectively decreasing Kv4.2 expression in EPI cells. We also found that NFATc3 is required for I to downregulation during sustained activation of βAR signaling. These findings are consistent with recent studies indicating that the calcineurin/NFATc3 signaling pathway regulates the expression of K + channels in heart under physiological and pathophysiological conditions [ 11 , 12 ]. Indeed, a gradient in [Ca 2+ ] i and calcineurin/NFATc3 signaling has been suggested to underlie differential Kv4 expression across the mouse left ventricular free wall. Our data suggest that chronic ISO infusion dissipates this Kv4.2 and I to gradient by increasing [Ca 2+ ] i as well as calcineurin and NFAT activity in EPI, but not in ENDO cells. NFATc3 downregulates Kv4.2 and Kv4.3 channel expression in adult ventricular myocytes [ 11 , 12 ]. This observation raises an important question: why does chronic ISO infusion only result in the downregulation of Kv4.2 in EPI myocytes? A recent study by our group examining the mechanisms by which NFATc3 regulates Kv4 expression in control EPI and ENDO suggests an answer to this conundrum. In this study [ 11 ], we found that Kv4.2 and Kv4.3 genes, both of which have putative NFAT binding sites in their promoters, have different thresholds for NFAT-dependent suppression. At relatively low levels of NFAT activity (i.e., similar to those in control EPI cells), expression of Kv.4.2 and Kv4.3 is high. Increasing NFAT activity by about 1.6-fold (to about the same level observed in ENDO cells) downregulated Kv4.2, but not Kv4.3 expression in mouse ventricular myocytes. Higher elevations (presumably 3-fold) in NFAT activity are required for downregulation (60%) of both Kv4.2 and Kv4.3 genes [ 11 , 12 ]. Thus, the level of NFAT activity observed in ISO-infused EPI (i.e. similar to control ENDO) is sufficient to downregulate Kv4.2, but not Kv4.3 expression. The experiments in this study were performed in mice to take advantage of available genetically engineered animals. Unlike mice and rats, I to in canine and human hearts is produced by Kv4.3 [ 40 ]. A recent study suggests that NFATc3 modulates Kv4.3 expression and hence I to in canine ventricular myocytes [ 41 ]. Thus, it is intriguing to speculate that chronic activation of β1AR could decrease I to density via an NFATc3-dependent downregulation of Kv4.3 expression in larger mammals. Note, however, that the signaling pathways underlying regional variations of Kv4.3 expression in canine are still unclear. Experiments need to be performed to establish the relationship between β1AR, calcineurin/NFAT, and Kv4.3 expression in human and canine hearts. To conclude, our data clearly indicate regional variations in β1AR signaling in the left ventricular free wall. Furthermore, our findings support the view that β1AR signaling and calcineurin/NFATc3 signaling play a critical role in the modulation of I to across the left ventricular wall and hence modulates the ventricular excitability under physiological pathophysiological conditions.

Myocytes were obtained from the left ventricle ENDO and EPI of wild type (WT), NFATc3 −/− , β1AR −/− , and NFAT-luc as previously described [ 24 ]. Electrophysiological signals were recorded using an Axopatch 200B. For [Ca 2+ ] i measurements cells, were loaded with the acetomethylester version of the fluorescent Ca 2+ indicator fluo-4. RT-PCR and Western blot analyses were performed as described elsewhere [ 11 , 12 ]. Calcineurin activity was quantified using a commercially available kit (Promega). Data are presented as mean±SEM. An expanded Materials and methods section can be found online.