Cardiomyocytes with disrupted CFTR function require CaMKII and Ca²⁺-activated Cl⁻ channel activity to maintain contraction rate

Abstract

The physiological role of the cystic fibrosis transmembrane conductance regulator (CFTR) in cardiomyocytes remains unclear. Using spontaneously beating neonatal ventricular cardiomyocytes from wild-type (WT) or CFTR knockout (KO) mice, we examined the role of CFTR in the modulation of cardiomyocyte contraction rate. Contraction rates of spontaneously beating myocytes were captured by video imaging. Real-time changes in intracellular ([Ca²⁺]_i) and protein kinase A (PKA) activity were measured by fura-2 and fluorescence resonance energy transfer, respectively. Acute inhibition of CFTR in WT cardiomyocytes using the CFTR inhibitor CFTR_inh-172 transiently inhibited the contraction rate. By contrast, cardiomyocytes from CFTR KO mice displayed normal contraction rates. Further investigation revealed that acute inhibition of CFTR activity in WT cardiomyoctyes activated L-type Ca²⁺ channels, leading to a transient increase of [Ca²⁺]_i and inhibition of PKA activity. Additionally, we found that contraction rate normalization following acute CFTR inhibition in WT cardiomyocytes or chronic deletion in cardiomyocytes from CFTR KO mice requires the activation of Ca²⁺/calmodulin-dependent kinase II (CaMKII) and Ca²⁺-activated Cl⁻ channels (CaCC) because simultaneous addition of myristoylated-autocamtide-2-related inhibitory peptide or niflumic acid and CFTR_inh-172 to WT cardiomyocytes or treatment of cardiomyoctes from CFTR KO mice with these agents caused sustained attenuation of contraction rates. Our results demonstrate that regulation of cardiomyocyte contraction involves CFTR. They also reveal that activation of CaMKII and CaCC compensates for loss of CFTR function. Increased dependence on CaMKII upon loss of CFTR function might leave cystic fibrosis patients at increased risk of heart dysfunction and disease.

Introduction

To sustain proper physiological function, the heart maintains tight control of ion channels and transporters. Four major Cl⁻ currents have been described in the heart, including the cystic fibrosis transmembrane conductance regulator (CFTR), Ca²⁺-activated Cl⁻ channels (CaCC), swelling-activated Cl⁻ channels and inwardly rectifying Cl⁻ channels. Current thinking is that CFTR, swelling-activated Cl⁻ channels and CaCC share a similar role in maintaining the resting membrane potential and minimizing action potential prolongation associated with stimulation of inward Ca²⁺ current. As a result, these Cl⁻ channels may be arrythmogenic and block hypercontraction through prevention of intracellular Ca²⁺ concentration ([Ca²⁺]_i) overload (Yamawake et al. 1992; Collier et al. 1996; Hiraoka et al. 1998).

Shortly after the identification of the CFTR gene from epithelia, a cAMP- or PKA-dependent Cl⁻ channel resembling CFTR was identified in cardiomyocytes (Nagel et al. 1992). Subsequently, expression of a cardiac-specific isoform of CFTR (cCFTR) was reported in atrial and ventricular myocytes of numerous species, including humans (Tilly et al. 1996; Warth et al. 1996; Duan et al. 1999; Gao et al. 2007). This isoform arises from alternative splicing, resulting in loss of a portion of the first intracellular loop, but otherwise retains >95% sequence homology to epithelial CFTR. The cardiac-specificisoform of CFTR behaves in a similar manner to epithelial CFTR and is activated and inhibited by the same agents that affect epithelial CFTR. Additionally, clinical mutations of CFTR, which cause the disease cystic fibrosis (CF), lie outside of this region. While no clear physiological role for CFTR in the heart has been firmly established, it has been suggested that, as a Cl⁻ channel, CFTR may be involved in the autonomic regulation of action potential duration by minimizing action potential prolongation associated with β-adrenergic stimulation (Duan et al. 2005). Studies showing that CFTR is upregulated during ischaemia (Uramoto et al. 2003) and that CFTR KO mice lose the protection of ischaemic preconditioning (Chen et al. 2004) suggest that CFTR plays an important role in cardiomyocytes. While the aforementioned studies have provided important insight into the potential roles of CFTR in the heart, we lack full understanding of the physiological role(s) of CFTR and the mechanism(s) employed to fulfil its functions.

Thus, the goal of the present study was to determine whether loss of CFTR affects ventricular cardiomyocyte function by examining the effects of CFTR loss on spontaneously beating neonatal cardiomyocytes. In addition to contraction rate measurements, we monitored real-time changes in [Ca²⁺]_i and protein kinase A (PKA) activity, and CFTR-dependent changes in Ca²⁺/calmodulin-dependent kinase II (CaMKII) activity. These experiments showed that CFTR is involved in the regulation of cardiomyocyte contraction rate. However, acute or chronic loss of CFTR was compensated for by activation of CaMKII and CaCC.

Methods

Ethical approval of animal use

FV/B mice were obtained from Charles River Laboratory (Wilmington, MA, USA) and used for generation of wild-type (WT) mice. The CFTR KO mice (strain 2515 B6.129S6-Cftr^tm1Kth/J; Zhou et al. 1994) were initially obtained from Dr Deborah Nelson at the University of Chicago (Chicago, IL, USA) and bred in-house. All animals were used within 24 h of birth. In the present study a total of 201 WT and 54 CFTR KO mice were used. Use of animals was approved by the University of Illinois at Urbana-Champaign IACUC and is in accordance with The Journal of Physiology guidelines (Drummond, 2009).

Isolation and culture of neonatal mouse cardiomyocytes

Ventricular myocytes were cultured from neonatal mice using a similar protocol to that previously reported (Xiang et al. 2005). Newborn mice were killed by decapitation with surgical scissors. Decapitation was used so that isolated myocytes would not be affected by anaesthetics or altered environment (redox and pH (CO₂)). This procedure is justified to yield viable myocytes after isolation, because both sedation and anaesthetics significantly reduce the survival rate ofisolated heart cells from neonates. Decapitation of neonates avoiding prior anaesthesia was done in compliance with the Report of the AVMA Panel on Euthanasia JACMA, Col 218, No. 5, 1 March 2001, pp. 682–683 (http://www.avma.org/resources/euthanasia.pdf) and the NIH ARAC Guidelines for the Euthanasia of Mouse and Rat Fetuses and Neonates (http://oacu.od.nih.gov/ARAC/euthmous.htm). Neonatal mouse cardiac tissue devoid of atria were exposed to enzymatic digestion. Subsequently, isolated cells were pipetted onto 35 mm Petri dishes precoated with 1.5% gelatin type A or laminin. Cells were plated at high and low density for measurement of contraction rate or real-time imaging experiments, respectively. High-density plating allowed cells to form a syncytium, whereas low-density plating favoured individual quiescent myocytes. Cells were incubated at 37°C for 2–3 days until they displayed appropriate myocyte morphology or formed a syncytium for contraction rate measurements. Culture media (containing serum and 20 mm Hepes for buffering) was replaced daily and at least 1 h prior to experiments.

Measurement of contraction rate

Measurement of myocyte contraction rate was performed as previously reported (Xiang et al. 2002). The 35 mm Petri dishes were placed in a temperature- and CO₂-regulated apparatus. Using a CCD camera (Photometrics, Tucson, AZ, USA) and MetaMorph imaging software (Molecular Devices, Downington, PA, USA), 5 s video images were obtained every 2 min. Following acquisition, a specific point within the field was selected and analysed for movement in each x, y and z plane. Movements in the x and y plane were calculated and number of deflections noted as number of contractions. Contraction velocity measurements were also recorded. Each data point represents a single observation from different experiments so that only one data set was obtained from each experiment. Each drug was tested using multiple isolations to ensure that the observed responses were not particular to a single isolation.

Measurement of [Ca²⁺]_i

Quiescent ventricular myocytes were cultured at low density as described in ‘Isolation and culture of neonatal mouse cardiomyocytes’. Upon formation of myocyte morphology (24–48 h after culture), culture media was removed and cells were washed with 1× Dulbecco's phosphate-buffered solution (DPBS), after which cells were loaded with 5 μm fura-2 AM dissolved in 0.01% pluronic F-127 plus 0.1% DMSO in DPBS at room temperature for 30 min. Following loading, cells were washed with 1× DPBS or Ca²⁺-free DPBS [with EGTA (500 μm) to remove residual free Ca²⁺] and left untreated for 30 min to remove any unloaded dye and allow for complete de-esterification of fura-2 AM into fura-2, its active form. For Ca²⁺-free experiments, experiments were performed in Ca²⁺-free DPBS with EGTA. Changes in [Ca²⁺]_i were monitored by fura-2 excitation at 340 and 380 nm and emission at 500 nm. Since the ratio of the light emitted at 340 nm excitation to that emitted at 380 nm excitation (340/380 ratio) is a direct index of the level of [Ca²⁺]_i, data are expressed as the 340/380 ratio following subtraction of background fluorescence.

For monitoring of Ca²⁺ transients, we used fura-2 in a similar manner to that described above. In order to capture full transients, myocytes were plated at a density that allowed spontaneous contraction, but facilitated visualization of a single layer of myocytes. Myocytes cultured in this manner retained their spontaneous beating at regular intervals, but at a substantially reduced rate.

Measurement of real-time PKA or cAMP activity via fluorescence resonance energy transfer (FRET)

Following 24 h in culture, neonatal cardiomyocytes were infected with an adenovirus containing AKAR2.2 (PKA) or ICUE3 (cAMP; 100 multiplicity of infection (MOI), 24 h) in a similar manner to that previously published (Soto et al. 2009). Between 24 and 48 h postinfection cells were washed and stored in DPBS for FRET recordings. Dual emission ratio imaging was acquired for 200 s at 20 s intervals. Images in both channels were subjected to background subtraction, and ratios of yellow fluorescent protein (YFP) to cyan fluorescent protein (CFP) were calculated at different time points and then normalized to baseline signals. For AKAR2.2, phosphorylation of the PKA consensus site leads to an increase in the YFP/CFP ratio (Soto et al. 2009), whereas the binding of cAMP to the Epac site of ICUE3 causes a decrease in FRET (DiPilato et al. 2004).

Measurement of phospholamban phosphorylation

Neonatal cardiomyocytes, pretreated with 20 μm CFTR_inh-172 for various times, were chilled, washed, and harvested in lysis buffer [20 mm Hepes pH 7.5, 2 mm EDTA pH 8.0, 150 mm NaCl, 10% glycerol, 0.6% NP₄O, 20 mm Na₄P₂O₇, 50 mm NaF, 2.5 mm Na₃VO₄ and Protease Inhibitor Single-Use Cocktail (Thermo Scientific, Pierce, Rockford, IL, USA)]. The lysates were clarified at 13,200g for 20 min at 4°C. The samples were resolved on an SDS-PAGE gel. Phospholamban phosphorylation was detected with a phospho-threonine 17 antibody (PLB-Thr17, Badrilla Ltd, Leeds, UK). Additionally, total phospholamban was probed with an antiphospholamban antibody (PLB, Badrilla Ltd) and γ-tubulin with an anti-γ-tubulin antibody (Sigma, St Louis, MO, USA). Primary antibodies were visualized with IRDye 680CW goat antimouse or with IRDye 800CW goat antirabbit secondary antibodies using an Odyssey scanner (Li-cor biosciences, Lincoln, NE, USA). The signals obtained from PLB-Thr17 were corrected and plotted as the increase over internal control (γ-tubulin).

Chemicals

Niflumic acid, thapsigargin, nifedipine and myristoylated-PKA inhibitor fragment 14–22 (PKI) were obtained from Sigma-Aldrich (St Louis, MO, USA). KB-R7943 came from Tocris Bioscience (Ellisville, MO, USA). BayK-8644 was purchased from Calbiochem (San Diego, CA, USA), while CFTR_inh-172 was obtained from Calbiochem or the Cystic Fibrosis Foundation Therapeutics compound collection (Rosalind Franklin University of Medicine and Science, Chicago, IL, USA). Myristoylated-autocamtide-2-related inhibitory peptide (AIP) was obtained from Biomol (Plymouth Meeting, PA, USA). Fura-2 AM was purchased from Invitrogen (Carlsbad, CA, USA). Ryanodine came from Alomone Labs (Jerusalem, Israel). Niflumic acid, CFTR_inh-172, ryanodine, thapsigargin, and fura-2 AM were dissolved in DMSO, while PKI and AIP were dissolved in water. All drugs were made as 1:1000 stock solutions and diluted on the day of the experiment.

Statistical analysis

For presentation of changes in contraction rate, [Ca²⁺]_i or PKA activity, peak responses were subtracted from their respective baseline values. For percentage baseline measurements, the last measured contraction rates were divided by baseline values. Normalization was performed by GraphPad Prism software (GraphPad Software, La Jolla, CA, USA) by setting the first baseline measurement to 100% and the value of 0 to 0%. For comparison between two groups of data, Student's unpaired t test was used to determine significance. For multiple comparisons, one-way ANOVA was performed. Time course curves were compared by two-way ANOVA. Results were considered significant at P < 0.05 and are denoted accordingly.

Results

Effect of pharmacological inhibition and genetic deletion of CFTR on cardiomyocyte contraction rate

To examine the potential role of CFTR in the regulation of ventricular cardiomyocyte contraction rate, we used syncytial neonatal mouse ventricular myocytes, which form a single spontaneously beating unit. In control experiments (DMSO or H₂O), we found that cardiomyocytes held a steady rhythm of beating over time (Figs 1, 4 and 5). To determine the acute effect of CFTR inhibition on cardiomyocyte contraction rate, we used the selective CFTR inhibitor, CFTR_inh-172 (Ma et al. 2002). Application of CFTR_inh-172 (10 and 20 μm) resulted in a transient, but significant (P < 0.05), dose-dependent inhibition of contraction rate thatsubsequently returned to baseline levels (Fig. 1). Addition of CFTR_inh-172 (20 μm) to cardiomyocytes from CFTR KO myocytes failed to significantly inhibit the contraction rate when compared with baseline or DMSO (P > 0.05), demonstrating specificity of action of CFTR_inh-172 and the ability of CFTR KO cardiomyocytes to function normally without CFTR. Examination of deflection magnitude and contraction velocity, indirect indicators of contractility, showed no significant changes upon addition of CFTR_inh-172 (P > 0.05; data not shown). As seen from the representative trace in Fig. 1C, CFTR_inh-172predominantly increased the interval between contractions. Thus, CFTR is involved in the physiological regulation of cardiomyocyte contraction rate; however, both acute and chronic loss of CFTR activity can be compensated for.

Figure 1. Pharmacological inhibition, but not genetic deletion, of CFTR attenuates cardiomyocyte contraction rate.

A, syncytial WT ventricular myocytes were treated with DMSO (n≥ 6) or CFTR_inh-172 (10 μm or 20 μm; n= 7 each) for 20 min following a 10 min baseline period. Separately, the effect of CFTR_inh-172 (20 μm; n= 5) on ventricular myocytes obtained from CFTR KO mice was also examined. *P < 0.05; **P < 0.01; ***P < 0.001 vs. DMSO or CFTR KO mice by two-way ANOVA. B, comparison of the maximal inhibitory effect of CFTR_inh-172 on WT and CFTR KO cardiomyocyte contraction rate. ***P < 0.001 vs. DMSO by one-way ANOVA; +++P < 0.001 vs. 10 μm CFTR_inh-172 by Student's unpaired t test. C, normalized representative trace of spontaneously beating WT ventricular myocyte with original traces prior to and after CFTR_inh-172 (20 μm) addition. Deflection in the y-axis is expressed as arbitrary units (a.u.).

Figure 4. Calcium/calmodulin-dependent kinase II activity is required to maintain normal contraction rates following pharmacological inhibition or genetic deletion of CFTR.

A, syncytial WT ventricular myocytes were treated with H₂O (n= 5) or AIP (1 μm, n= 6) for 30 min following a 10 min baseline period. B, separately, the effect of simultaneous addition of AIP plus CFTR_inh-172 (20 μm; n= 8) for 20 min was examined and compared with responses by CFTR_inh-172 alone (from Fig. 1A). *P < 0.05; ***P < 0.001 vs. CFTR_inh-172 alone by two-way ANOVA. C, WT cardiomyocytes were exposed to CFTR_inh-172 (20 μm) for various lengths of time (n= 3–5 each) and then harvested to examine PLB-Thr17 phosphorylation. All results were quantified and normalized. A representative blot is shown. *P < 0.05 vs. no treatment by Student's unpaired t test. D, cardiomyocytes from CFTR KO mice (n= 8) were treated with AIP (1 μm) for 30 min following baseline measurements. Contraction rates were compared with those of cardiomyocytes from CFTR KO mice shown in Fig. 1A. *P < 0.05; **P < 0.01 vs. cardiomyocytes from CFTR KO alone by two-way ANOVA. E, a comparison of changes in contraction rate due to AIP in WT cardiomyocytes (n= 6), WT cardiomyocytes plus CFTR_inh-172 (n= 8) and cardiomyocytes from CFTR KO mice (n= 8). **P < 0.01; ***P < 0.001 vs. WT by Student's unpaired t test.

Figure 5. Calcium-activated Cl⁻ channel plays an important role in cardiomyocyte contraction rate following CFTR inhibition or KO.

A, syncytial WT ventricular myocytes were treated with DMSO (n= 6) or niflumic acid (100 μm; n= 7) for 30 min following a 10 min baseline period. B, separately, the effect of simultaneous addition of niflumic acid (100 μm) plus CFTR_inh-172 (20 μm; n= 8) for 20 min was examined and compared with the effect of CFTR_inh-172 alone (from Fig. 1A). C, ventricular myocytes from CFTR KO mice (n= 8) were treated with niflumic acid (100 μm) for 20 min after baseline measurements. Contraction rates were compared with those of the cardiomyocytes from the CFTR KO mice shown in Fig. 1A. *P < 0.05; **P < 0.01; ***P < 0.001 vs. cardiomyocytes from CFTR KO alone by two-way ANOVA. D, a comparison of changes in contraction rate due to niflumic acid in WT cardiomyocytes (n= 7), WT cardiomyocytes plus CFTR_inh-172 (n= 8) and cardiomyocytes from CFTR KO mice (n= 8). ***P < 0.001 vs. WT by Student's unpaired t test.

Effect of CFTR_inh-172 on [Ca²⁺]_i and real-time PKA activity

The contraction rate of spontaneously beating cardiomyocytes is dependent on changes in [Ca²⁺]_i and PKA activity (Vinogradova et al. 2006). Therefore, to understand the mechanism whereby CFTR blockade transiently inhibits cardiomyocyte contraction rate, we investigated whether CFTR_inh-172 alters [Ca²⁺]_i and/or PKA activity. We first examined [Ca²⁺]_i by monitoring changes in fura-2 fluorescence, which is a direct index of [Ca²⁺]_i, in cardiomyocytes from WT and CFTR KO mice. In both quiescent and spontaneously beating WT myocytes, CFTR_inh-172 (20 μm) significantly increased [Ca²⁺]_i and [Ca²⁺]_i transients when compared with DMSO-treated myocytes (P < 0.001; Fig. 2). In contrast, in cardiomyocytes from CFTR KO mice CFTR_inh-172 failed to elicit a significant increase in [Ca²⁺]_i compared with DMSO (P > 0.05). In spontaneously beating myocytes, CFTR_inh-172 increased both systolic and diastolic Ca²⁺ (Fig. 2B). To understand the source of this [Ca²⁺]_i increase, we performed additional experiments targeting extracellular and/or intracellular Ca²⁺ stores and the responsible channels/transporters. Since spontaneously beating cardiomyocytes are sensitive to pharmacological modulation of Ca²⁺ channels/transporters (data not shown), we used quiescent cardiomyocytes for these experiments. Removal of extracellular Ca²⁺ completely prevented the CFTR_inh-172-induced increase in [Ca²⁺]_i (Fig. 2C). In contrast, inhibition of sarcoplasmic reticulum Ca²⁺ release by pretreatment with thapsigargin (1 μm) and ryanodine (10 μm) prior to CFTR_inh-172 did not have any significant influence on [Ca²⁺]_i changes (Fig. 2C). While many routes of Ca²⁺ entry in cardiomyocytes exist, L-type Ca²⁺ channels and the Na⁺–Ca²⁺ exchanger are the most prominent. Therefore, we examined whether one or both of these proteins are involved in mediating the CFTR_inh-172-induced increases in [Ca²⁺]_i. Pretreatment with the L-type Ca²⁺ channel inhibitor nifedipine (10 μm), but not the Na⁺–Ca²⁺ exchange inhibitor KB-R7943 (10 μm), completely prevented CFTR_inh-172-induced increases in [Ca²⁺]_i (Fig. 2C). Thus, CFTR inhibition causes L-type Ca²⁺ channel activation, resulting in a transient increase in [Ca²⁺]_i.

Figure 2. Inhibition of CFTR increases [Ca²⁺]_i via L-type Ca²⁺ channels.

A, quiescent WT ventricular myocytes were loaded with fura-2 to monitor changes in [Ca²⁺]_i. Following a 3 min baseline period, either CFTR_inh-172 (20 μm, n= 32) or DMSO (n= 19) was added, and subsequent effects were measured for 20 min. B, representative trace (n= 12) of spontaneously beating cardiomyocyte that was loaded with fura-2 to measure Ca²⁺ transients prior to and after CFTR_inh-172 (20 μm) addition. Separate experiments with DMSO showed no change in diastolic or systolic Ca²⁺ (data not shown; n= 4). C, quiescent WT cardiomyocytes were bathed in Ca²⁺-free solution (n= 25), pretreated with ryanodine (10 μm) and thapsigargin (1 μm; n= 26), KB-R7943 (10 μm; n= 23), nifedipine (10 μm; n= 26) or KB-R7943 and nifedipine (n= 23) for 30 min prior to [Ca²⁺]_i measurements. Additionally, in myocytes from CFTR KO mice, fura-2 fluorescence was monitored upon CFTR_inh-172 (20 μm; n= 23) or DMSO (n= 24) addition. ***P < 0.001 vs. CFTR_inh-172 by Student's unpaired t test; NS, not significant.

Subsequently, we sought to determine the effect of CFTR_inh-172 on PKA activity by using FRET to monitor PKA activity in real time. Compared with DMSO, CFTR_inh-172 (20 μm) significantly inhibited PKA activity, in both magnitude and duration (P < 0.001; Fig. 3A). This inhibition was transient and recovered to baseline levels within 20 min, resembling the pattern of inhibition observed on contraction rates. The CFTR_inh-172-induced inhibition of PKA was not the result of upstream alterations in adenylyl cyclase activity, since separate experiments measuring cAMP showed that CFTR_inh-172 caused no change in cAMP levels (DMSO 0.17 ± 0.02 vs. CFTR_inh-172 0.20 ± 0.01 –ΔYFP/CFP, P > 0.05 vs. DMSO). To understand how CFTR inhibition leads to a decrease in PKA activity we examined whether there was a link between the [Ca²⁺]_i change and PKA activity. In these experiments, inhibition of the L-type Ca²⁺ channel with nifedipine pretreatment (10 μm) completelyprevented CFTR_inh-172-induced attenuation of PKA activity (Fig. 3B), indicating that activation of L-type Ca²⁺ channels was responsible for the CFTR_inh-172-induced inhibition of PKA activity. Similarly, direct activation of L-type Ca²⁺ channels by BayK-8644 (1 μm) caused a significant decrease in PKA activity (Fig. 3B), providing further evidence that L-type Ca²⁺ channel activation can negatively affect PKA activity. To verify that inhibition of PKA activity leads directly to functional attenuation of contraction rate, we examined the effect of myristoylated PKI (20 μm), a selective membrane-permeable peptididic inhibitor of PKA, on cardiomyocyte contraction rate. In agreement with previous reports (Vinogradova et al. 2006; Soto et al. 2009), we found that PKA inhibition caused a significant decrease in contraction rate (P < 0.001; Fig. 3C). As seen in both FRET and contraction rate experiments, CFTR_inh-172 did not cause a further decrease in PKA activity or spontaneous contractions when combined with PKI, which elicited a more severe depression in PKA activity and spontaneous contractions compared with CFTR_inh-172 (Fig. 3C). In separate experiments, we found that direct activation of L-type Ca²⁺ channels by BayK-8644 (1 μm) also led to a significant inhibition of contraction rate (H₂O 11 ± 3 vs. BayK-8644 −133 ± 5 Δbeats min⁻¹, P < 0.001 vs. H₂O, n= 6). Taken together, our data indicate that CFTR blockage activates L-type Ca²⁺ channels, causing a transient increase in [Ca²⁺]_i and hence, a transient inhibition of PKA activity, thereby leading to a decrease in contraction rate.

Figure 3. CFTR_inh-172-induced activation of L-type Ca²⁺ channels causes inhibition of PKA and contraction rate.

A, quiescent WT ventricular myocytes were infected with AKAR2.2 to monitor real-time changes in PKA activity. Following a 5 min baseline period, CFTR_inh-172 (20 μm; n= 24) or DMSO (n= 18) was added, and subsequent effects were measured for 20 min. B, WT cardiomyocytes were pretreated with nifedipine (10 μm; n= 24) for 30 min prior to CFTR_inh-172 addition. Additionally, PKA activity following CFTR_inh-172 (20 μm; n= 19) was monitored in myocytes from CFTR KO mice. In separate experiments, the effects of BayK-8644 (1 μm; n= 15), PKI (20 μm; n= 14) or myristoylated PKI plus CFTR_inh-172 (n= 14) on PKA activity in WT cardiomyocytes were measured. *P < 0.05; **P < 0.01 vs. CFTR_inh-172; ++P < 0.01 vs. DMSO by Student's unpaired t test. C, syncytial WT ventricular myocytes were treated with PKI (20 μm; n= 8) for 30 min following a 10 min baseline period. Inset shows a comparison of syncytial WT cardiomyocytes treated with PKI alone or with PKI plus CFTR_inh-172 (20 μm; n= 7).

Calcium/calmodulin-dependent kinase II in the recovery of CFTR_inh-172-induced contraction rate depression

Protein kinase A and CaMKII regulate overlapping signalling cascades, leading to complementary regulation of proteins (i.e. phospholamban) involved in cardiomyocyte contraction (Wang et al. 2004). We therefore sought to determine whether CFTR inhibition alters CaMKII activity. We first examined the effect of CaMKII inhibition on cardiomyocyte contraction rate. In cardiomyocytes from WT mice, addition of AIP (1 μm), a selective CaMKII inhibitor, alone reproducibly caused a minor, statistically insignificant (P > 0.05) decrease in contraction rate compared with control cardiomyocytes (Fig. 4A). However, in separate experiments, when we added CFTR_inh-172 (20 μm) together with AIP, cardiomyocyte contraction rate was not able to recover to baseline in a manner comparable to CFTR_inh-172 alone (Fig. 4B). To explore further the role of CaMKII, we measured the effect of CFTR_inh-172 on PLB-Thr17 phosphorylation, a specific CaMKII phosphorylation site on phospholamban. In agreement with our functional data, CFTR_inh-172 (20 μm) caused a significant increase in PLB-Thr17 phosphorylation (Fig. 4C). Although this phosphorylation occurred very rapidly (within 2 min), it was transient, as phosphorylation returned to near-baseline levels within 4 min (Fig. 4C). Thus, while CFTR_inh-172 induces a transient increase in [Ca²⁺]_i that causes a decrease in contraction rate, it also activates CaMKII, which facilitates recovery from CFTR inhibition. To determine whether CaMKII is also important in conditions of chronic CFTR loss, we examined the effects of CaMKII inhibition on cardiomyocytes from CFTR KO mice. Figure 4D and E demonstrates that inhibition of CaMKII by AIP profoundly attenuated the contraction rate of cardiomyocytes from CFTR KO mice and that this effect was significantly greater than the action of AIP on the contraction rate of WT cardiomyocytes. Therefore, although cardiomyocytes from CFTR KO mice produce normal contraction rates, they have increased dependence on CaMKII activity to maintain these rates.

Role of CaCC in the regulation of cardiomyocyte contraction rate

Subsequently, we sought to determine how CaMKII activation facilitates an increase in contraction rate for recovery from CFTR inhibition. Since PLB-Thr17 phosphorylation is transient (Fig. 4C), we hypothesized that contraction rate recovery is not purely a result of altered PLB phosphorylation. Loss of CFTR activity is known to lead to increased CaCC activity as a compensatory Cl⁻-conductive pathway (Grubb et al. 1994; Hartzell et al. 2005). With our observations that CFTR inhibition increases [Ca²⁺]_i and CaMKII activity, we sought to determine whether activation of CaCC might compensate for loss of CFTR activity. To investigate the role of CaCC in the regulation of cardiomyocyte contraction rate, we used niflumic acid (100 μm) to inhibit CaCC directly. Addition of niflumic acid alone resulted in a rapid, but transient, decrease in contraction rate, which recovered to baseline (Fig. 5A). Thus, inhibition of either CFTR or CaCC resulted in similar functional effects on cardiomyocyte contraction rate, indicating that, in ventricular cardiomyocytes, CFTR and CaCC may have complementary roles. To determine whether CaCC activity is necessary for recovery from CFTR inhibition, we performed experiments in which we simultaneously applied niflumic acid (100 μm) and CFTR_inh-172 (20 μm) to cardiomyocytes from WT and CFTR KO mice. We found that inhibition of both CFTR and CaCC in WT cardiomyocytes strongly abrogated the contraction rate andprevented full recovery to baseline (Fig. 5B). Examination of CaCC inhibition in cardiomyocytes from CFTR KO mice also showed a significant inhibitory effect of niflumic acid on cardiomyocyte contraction rate (Fig. 5C). Like AIP (Fig. 4E), niflumic acid-induced inhibition of contraction rate in cardiomyocytes from CFTR KO mice was significantly enhanced when compared with that in cardiomyocytes from WT mice (Fig. 5D). This result indicates an increased reliance on CaCC. Thus, CaCC is important for recovery from CFTR inhibition and constitutes an important downstream effector of CaMKII activation.

Discussion

Since the first electrophysiological recording of CFTR in the heart (Nagel et al. 1992), there has been little dispute regarding it being responsible for the cardiac PKA-stimulated Cl⁻ channel. While there have been conflicting reports about CFTR expression and/or activity in humans (Warth et al. 1996; Berul et al. 1997; Du et al. 2000), studies have shown that CFTR is expressed, and electrophysiological recordings have been made from rats (Uramoto et al. 2003), guinea-pigs (Hume et al. 1994), rabbits (Hart et al. 1996), swine (Gao et al. 2007), dogs (Hume et al. 1994), primates (Warth et al. 1996) and, most pertinent to our study, mice (Lader et al. 2000). Despite these studies, questions remain regarding the role of CFTR in the heart. Thus, the primary goal of our study was to determine whether CFTR plays a physiological role in the regulation of heart function.

For our studies, we used neonatal ventricular cardiomyocytes. Unlike adult ventricular myocytes, neonatal ventricular myocytes spontaneously beat as a single syncytial unit when cultured appropriately. This property allowed us to examine myocyte function without exogenous perturbation. While it may be tempting to compare these cells to adult pacemaker or atrial cells, which also spontaneously contract, it is more likely that neonatal ventricular myocytes are analogous to adult ventricular myocytes. In a comparison of their electrophysiological properties, Nuss & Marban (1994) showed that mouse neonatal and adult ventricular myocytes express similar ion channels and have comparable action potential durations. In previous exploratory studies, we found that CFTR blockade in paced adult mouse ventricular myocytes transiently inhibited contractility in a manner similar to the effects observed in thepresent study using neonatal mouse ventricular cardiomyocytes (Dr. Dagoberto Soto, University of Illinois, Urbana-Champaign, unpublished results). While the exact role of CFTR in adult ventricular cardiomyocytes remains to be determined, we have shown that CFTR plays an active role in the regulation of heart function, as measured by CFTR_inh-172-induced depression of neonatal cardiomyocyte contraction rate. These results are in agreement with cardiomyocyte simulation studies, which showed that decreased CFTR current density lengthens action potential duration (Kuzumoto et al. 2008). In our study, we found that inhibition of CFTR causes a transient increase in [Ca²⁺]_i due to activation of L-type Ca²⁺ channels, whose importance in action potential duration and contraction regulation are well established. We did not dissect out the mechanism whereby inhibition of CFTR activates L-type Ca²⁺ channels, but a potential mechanism is through alterations in membrane potential (see Fig. 6). While the effects of CFTR inhibition on neonatal cardiomyocyte membrane potential have not been examined, the simulation studies of Kuzumoto et al. (2008) predict that decreased CFTR activity causes increased depolarization during action potential repolarization. Additionally, Xu et al. found that inhibition of CaCCs prolongs early action potential repolarization (Xu et al. 2002). Our findings that CFTR_inh-172 and niflumic acid have similar effects on contraction rate suggest that CFTR may also affect membrane potential and action potential duration. However, further studies are required to explore this possibility.

Figure 6. Proposed model for the effect of CFTR disruption on cardiomyocyte signalling and ion transport.

Acute inhibition of CFTR by CFTR_inh-172 (1) probably causes an increase in membrane potential (V_m), which leads to increased activation of L-type Ca²⁺ channels (LTCC) (2). Increased LTCC activation raises intracellular Ca²⁺ (3), resulting in inhibition of PKA and activation of CaMKII (4). Inhibition of PKA causes a decrease in contraction rate. Activation of CaMKII stimulates Ca²⁺-activated Cl⁻ channels (CaCC) (5), which probably serves to restore changes in membrane potential and decrease LTCC hyperactivity (6). This decrease in [Ca²⁺]_i probably removes the inhibition on PKA activity, thus restoring contraction rate.

In addition to alterations in [Ca²⁺]_i, we have also shown that CFTR inhibition leads to a transient decrease in PKA activity, as measured by A-kinase activity receptor (AKAR) phosphorylation. This study, along with others (Vinogradova et al. 2006; Soto et al. 2009), has shown that PKA is an important regulator of contraction. We did not attempt to determine the exact mechanism whereby CFTR blockade leads to PKA inhibition. However, we did show that this inhibition was caused by L-type Ca²⁺ channel-mediated Ca²⁺ influx. Cardiac myocytes contain Ca²⁺-sensitive adenylyl cyclases (AC V and AC VI) that can be inhibited by increases in [Ca²⁺]_i (Willoughby & Cooper, 2007). However, our finding that cAMP remains unchanged following CFTR inhibition makes this unlikely. An additional possibility is that Ca²⁺-sensitive phosphatases, such as calcineurin, are involved in the inhibition we observed. It has been shown that calcineurin antagonizes PKA activity through dephosphorylation of PKA effectors (Santana et al. 2002) and even directly inhibits PKA itself (Orie et al. 2009). The lack of effect on cAMP also suggests that CFTR inhibition probably does not alter hyperpolarization-activated cyclic nucleotide-gated channels (HCN), which are known to be involved in contraction rate regulation (Z.M.S., unpublished results; Er et al. 2003), although we cannot rule out an indirect effect of membrane potential alteration. Altogether, we have documented how inhibition of CFTR activity can alter ion channel and intracellular signalling activities to affect cardiomyocyte function.

Importantly, we found that CFTR-dependent depression of contraction rate was transient and was restored to normal rates. Additionally, we found that cardiomyocytes from CFTR KO mice displayed normal contraction rates, further indicating that CFTR activity or expression is not critical to maintaining cardiomyocyte function. This fact is not surprising given that mutations in CFTR do not lead to early embryonic death in CF patients or mice. The ability of cardiomyocytes to compensate for transient or long-term loss of CFTR does not discount its significance, but rather underscores the adaptability of the heart to maintain its vital function. It is well accepted that gene knockout causes adaptation, but we have shown that short-term and chronic inhibition of CFTR leads to compensatory changes in cardiac myocyte function. Our data demonstrate that CaMKII and CaCC activity are necessary for recovery from acute CFTR inhibition in WT cardiomyocytes and maintenance of normal contraction rates in cardiomyocytes from CFTR KO mice. Interplay between CFTR and CaCC has been noted in several tissues that express both channels (Hartzell et al. 2005). In airway epithelia, loss of CFTR, either in CF patients or in CFTR KO mice, leads to compensatory increases in CaCC activity (Grubb et al. 1994). Additionally, in bovine pulmonary artery endothelial cells (Wei et al. 2001), Xenopus oocytes (Kunzelmann et al. 1997) and mouse parotid acinar cells (Perez-Cornejo & Arreola, 2004), expression of CFTR leads to a negative regulation of CaCC. While niflumic acid is a well-documented inhibitor of CaCC (Hartzell et al. 2005), it is not specific and can also affect CFTR (Scott-Ward et al. 2004) and K⁺ channels (Wang et al. 1997; Zhou et al. 2007). Although we cannot rule out effects on non-CFTR targets, the effects of niflumic acid on the contraction rate of cardiomyocytes from CFTR KO mice make us confident that our results are not due to additional effects on CFTR. This, coupled with the involvement of [Ca²⁺]_i and CaMKII, both of which activate CaCC, strengthens our view about the importance of CaCC during loss of CFTR activity. However, since Ca²⁺ and CaMKII also affect many other ion channels and kinases important in contraction, further studies are required.

Calcium/calmodulin-dependent kinase II has been identified as an important intracellular signalling molecule in cardiac myocytes involved in regulating numerous cellular processes, including activation of ion channels, modulation of action potential, hypertrophy and apoptosis (Anderson, 2007). In our study, we found that CaMKII is important in the recovery of cardiomyocyte contraction rate caused by inhibition of CFTR. Its role in this process appears to be, in part, due to its ability to activate CaCC, since inhibition of CaCC or CaMKII prevented recovery from CFTR blockade. Calcium/calmodulin-dependent kinase II has also been shown to decrease the inactivation rate of L-type Ca²⁺ channels (Guo & Duff, 2006), which may aid in contraction rate recovery. Another important finding is that, while cardiomyocytes from CFTR KO mice display normal baseline contraction rates, in the presence of CaMKII inhibition these rates were severely diminished. In contrast, CaMKII inhibition in WT mice showed only minimal attenuation of baseline contraction rate. These data indicate that cardiomyocytes from CFTR KO mice have an increased dependence on CaMKII. These findings are of particular interest, since cardiac myocytes with increased CaMKII expression or activity are prone to pathological cardiac conditions. Calcium/calmodulin-dependent kinase II has been linked to heart failure; hearts from mice with heart failure have increased expression of CaMKIIδ, and overexpression of CaMKII causes heart failure in mice (Hoch et al. 1999; Maier, 2005). In addition, increased CaMKII activity may lead to prolonged action potential duration and cause long QT syndrome by increasing late-inactivating or slowly inactivating Na⁺ currents (Bers, 2008). At this time, we cannot say whether the increased dependence on CaMKII is due to long-term loss of CFTR expression or function. However, it would be intriguing to determine whether CF mutations, which disrupt CFTR processing and trafficking (e.g. ΔF508; Kopito, 1999) or perturb CFTR channel gating (e.g. G551D; Li et al. 1996) also display increased dependence on CaMKII activity. Thus, our findings and the observations that CFTR is downregulated in ventricular tissue from heart failure patients (Solbach et al. 2008) and that heart damage occurs in some CF patients (Wiebicke et al. 1993; Zebrak et al. 2000) warrant future investigation of the role of CaMKII in CF hearts.

There have been scattered reports of cardiac abnormalities in CF patients, including alterations in left ventricular function, tachycardia, cardiomyopathy, myocardial fibrosis and necrosis (Ambrosi et al. 1993; Wiebicke et al. 1993; Florea et al. 2000; Zebrak et al. 2000). More recently, it was found that 51% of CF patients with severe lung infection had significantly decreased ejection fraction during exercise (Nash et al. 2007). It is unclear whether these findings are due to primary cardiac defects or are secondary effects from altered pulmonary function. Additionally, with the involvement of CFTR in smooth muscle contraction and vascular reactivity (Robert et al. 2005), vascular effects on the heart must also be considered. The physiological relevance of CFTR in the heart requires investigation, since recent advances in patient care have extended the average lifespan of CF patients to the mid-30s, with some individuals with milder mutations achieving normal lifespans (70–80 years; http://www.cff.org). Therefore, it is likely that as therapeutic advances continue to be made and individuals live longer, more active lives, CF patients will enter the age range where cardiovascular abnormalities become apparent and may exhibit significant clinical cardiac symptoms. A more complete understanding of the role of CFTR in the heart may lead to the identification of potential cardiac defects in CF patients and thus, increase the quality of care for these patients.

Our present study presents findings that further our understanding regarding the physiological role of CFTR in cardiomyocytes. Importantly, we found that CFTR is involved in the regulation of cardiomyocyte contraction. Acute loss of CFTR activity leads to a reduction in the spontaneous beating of neonatal ventricular myocytes through alterations in [Ca²⁺]_i and PKA. However, CaMKII and CaCC activity compensate for this loss and restore myocyte function. While cardiomyocytes from CFTR KO mice can maintain normal spontaneous contractions, they exhibit increased dependence on both CaMKII and CaCC.

Glossary

Abbreviations

AIP

myristoylated-autocamtide-2-related inhibitory peptide

AKAR2.2

A-kinase associated receptor 2.2

CaCC

Ca²⁺-activated Cl⁻ channel

CaMKII

Ca²⁺/calmodulin-dependent kinase II

cCFTR

cardiac isoform of CFTR

CF

cystic fibrosis

CFTR

cystic fibrosis transmembrane conductance regulator

DPBS

Dulbecco's phosphate-buffered solution

FRET

fluorescence resonance energy transfer

ICUE3

indicator of cAMP using Epac 3

KO

knockout

PKA

protein kinase A

PKI

myristoylated-PKA inhibitor fragment 14–22

PLB-Thr17

threonine 17 of phospholamban

WT

wild-type

Author contributions

Z.M.S. was responsible for the conception, design and performance of all experiments. He also drafted and finalized the manuscript. V.DeA. performed the CaMKII Western blot experiments and contributed to the manuscript preparation and finalization. Y.X. and P.M.B. contributed to the conception and design of experiments, intellectual content and finalization of the manuscript. All experiments were performed at the University of Illinois at Urbana-Champaign.