Fibroblast Inward-Rectifier Potassium Current Upregulation in Profibrillatory Atrial RemodelingNovelty and Significance
Rationale: Fibroblasts are involved in cardiac arrhythmogenesis and contribute to the atrial fibrillation substrate in congestive heart failure (CHF) by generating tissue fibrosis. Fibroblasts display robust ion currents, but their functional importance is poorly understood.
Objective: To characterize atrial fibroblast inward-rectifier K+ current (IK1) remodeling in CHF and its effects on fibroblast properties.
Methods and Results: Freshly isolated left atrial fibroblasts were obtained from controls and dogs with CHF (ventricular tachypacing). Patch clamp was used to record resting membrane potential (RMP) and IK1. RMP was significantly increased by CHF (from −43.2±0.8 mV, control, to −55.5±0.9 mV). CHF upregulated IK1 (eg, at −90 mV from −1.1±0.2 to −2.7±0.5 pA/pF) and increased the expression of KCNJ2 mRNA (by 52%) and protein (by 80%). Ba2+ (300 μmol/L) decreased the RMP and suppressed the RMP difference between controls and dogs with CHF. Store-operated Ca2+ entry (Fura-2-acetoxymethyl ester) and fibroblast proliferation (flow cytometry) were enhanced by CHF. Lentivirus-mediated overexpression of KCNJ2 enhanced IK1 and hyperpolarized fibroblasts. Functional KCNJ2 suppression by lentivirus-mediated expression of a dominant negative KCNJ2 construct suppressed IK1 and depolarized RMP. Overexpression of KCNJ2 increased Ca2+ entry and fibroblast proliferation, whereas the dominant negative KCNJ2 construct had opposite effects. Fibroblast hyperpolarization to mimic CHF effects on RMP enhanced the Ca2+ entry. MicroRNA-26a, which targets KCNJ2, was downregulated in CHF fibroblasts. Knockdown of endogenous microRNA-26 to mimic CHF effects unregulated IK1.
Conclusions: CHF upregulates fibroblast KCNJ2 expression and currents, thereby hyperpolarizing RMP, increasing Ca2+ entry, and enhancing atrial fibroblast proliferation. These effects are likely mediated by microRNA-26a downregulation. Remodeling-induced fibroblast KCNJ2 expression changes may play a role in atrial fibrillation promoting fibroblast remodeling and structural/arrhythmic consequences.
Congestive heart failure (CHF) is an important cause of atrial fibrillation (AF), with structural remodeling, particularly tissue fibrosis, playing a central role.1,2 Fibroblasts are the most abundant cells in the heart.3 Fibroblast proliferation and differentiation into myofibroblasts are important contributors to arrhythmogenesis under conditions like CHF by enhancing the production of extracellular matrix proteins, such as collagen, and possibly via electric interactions with cardiomyocytes.3
Fibroblasts are known to express a wide range of ion channels, but their functional role is poorly understood.3 Cardiac fibroblasts are not electrically excitable, but they have polarized resting membrane potentials (RMPs), with average values as negative as −37 mV.4 The primary determinant of fibroblast RMP is the inward-rectifier K+ current, IK1.5 We have recently shown that voltage-gated (Kv) K+ currents in cardiac fibroblasts are remodeled in CHF.6 This study aimed to characterize changes in cardiac fibroblast IK1 in CHF and to define underlying mechanisms and potential functional significance.
Adult mongrel dogs (22–30 kg) were divided into 2 groups: controls (28 males and 2 females) and dogs with 2-week ventricular tachypacing–induced congestive heart failure (CHF; 21 males and 3 females). Dogs with CHF had unipolar pacing leads inserted fluoroscopically into the right ventricular apex, which were programmed at 240 bpm for 2 weeks.6 On study days, dogs were anesthetized with morphine (2 mg/kg subcutaneously) and α-chloralose (120 mg/kg intravenously, followed by 29.25 mg/kg per hour) and ventilated mechanically. Effective refractory periods were measured at basic cycle lengths of 150, 200, 250, 300, and 350 ms in the right-atrial appendage, with 10 basic stimuli (S1) followed by a premature extrastimulus (S2) with 5-ms decrements. AF was induced with atrial burst pacing at 50 Hz and 10 V. Mean AF duration was based on 10 AF inductions in each dog. If the mean duration of the first 5 episodes of AF was >2 minutes, AF was induced only 5 times. In 4 dogs per group, tissue sections were analyzed for fibrous-tissue content as previously described,2 by an investigator blinded to group assignment.
Fibroblast Isolation and Culture
Atrial fibroblasts were obtained from left atria (LA) of adult mongrel dogs as previously described.7 Hearts were removed after intra-atrial injection of heparin (10 000 U) and immersed in 2 mmol/L of Ca2+-containing Tyrode solution. The left coronary artery was cannulated, and the LA tissue was perfused with 2 mmol/L of Ca2+ Tyrode solution (37°C, 100% O2), then with Ca2+-free Tyrode solution (≈10 minutes), followed by ≈60-minute perfusion with the same solution containing collagenase (≈0.48 mg/mL; CLSII, Worthington, OH) and 0.1% bovine serum albumin (Sigma). Cells were dispersed by trituration in KB (Kraftbruhe) solution (when used for electrophysiological study or sample acquisition for mRNA or protein analysis) or Medium 199 (Invitrogen) supplemented with 10% fetal bovine serum (Gibco), penicillin, and streptomycin for culture. Filtration (500-nm nanomesh) was used to remove debris, and cells were then centrifuged at 54.6g for 5 minutes to pellet cardiomyocytes. The supernatant was collected and filtered through 50-μm nanomesh and centrifuged at 314.5g for 10 minutes to concentrate fibroblasts. Freshly isolated fibroblasts were then separated; 1 aliquot was flash-frozen in liquid N2 and stored for biochemical studies, and the remaining cells were cultured on noncoated glass coverslips.7 Fibroblasts were incubated in 5% CO2/95% O2 humidified air (37°C). A medium change was performed 4 hours after plating to remove any dead cells and debris, and the medium was changed every 24 hours.
Ionic Current and RMP Recording
All in vitro recordings were obtained at 37°C. The whole-cell perforated–patch technique was used to record RMP in current-clamp mode, and tight-seal patch clamp was used to record IK1 in voltage-clamp mode. Borosilicate glass electrodes filled with pipette solution were connected to a patch-clamp amplifier (Axopatch 200A; Axon). Electrodes had tip resistances of 6 to 8 MΩ. Nystatin-free intracellular solution was placed in the tip of the pipette by capillary action (≈30 s), and then pipettes were backfilled with nystatin-containing (600 μg/mL) pipette solution. IK1 was recorded as the 300-μmol/L Ba2−–sensitive current. Tyrode solution contained (mmol/L) NaCl 136, CaCl2 1.8, KCl 5.4, MgCl2 1, NaH2PO4 0.33, dextrose 10, and HEPES 5, titrated to pH 7.3 with NaOH. The pipette solution for RMP and IK1 recording contained (mmol/L) GTP 0.1, potassium aspartate 110, KCl 20, MgCl2 1, MgATP 5, HEPES 10, sodium phosphocreatine 5, and EGTA 0.005 (pH 7.4, KOH). Junction potentials between bath and pipette solutions averaged 10.5 mV and were corrected for RMP measurements only. Currents are expressed as densities (pA/pF) to control for changes in cell size/capacitance with CHF.
One-day cultured canine atrial fibroblasts on microscope cover slips were loaded with Fura-2-acetoxymethyl ester (5 μmol/L; Invitrogen) in phenol-free M199 medium in the presence of Pluronic F-127 (20% solution in dimethylsulfoxide, 2.5 μg/mL) for 30 minutes at 36°C in a humidified incubator with 95% air/5% CO2. Cover slips were fixed in a perfusion chamber on the stage of a microscope, and fibroblasts were superfused with 1.8 mmol/L of Ca2+ Tyrode solution and maintained for ≥25 to 30 minutes at room temperature before experimental protocols to allow for deesterification of Fura-2-acetoxymethyl ester. Fura-2 was excited with dual excitation wavelengths at 340 and 380 nm, and emission was recorded at a wavelength of 510 nm. Ca2+ imaging was obtained with an IonOptix Fluorescence System mounted on an upright Nikon FN-1 microscope. To measure store-operated Ca2+ entry, cells were first exposed to Ca2+-free solution for 10 minutes. [Ca2+]o was then increased to 10 mmol/L to measure Ca2+ entry via store-operated channels activated by Ca2+-store depletion.
Cell Proliferation and Cell Cycle Analysis
Cell cycle was analyzed by flow cytometry as previously described.8 Atrial fibroblasts were seeded at 4000 cells/cm2 in T-25 culture flasks (1.0×105 cells per flask; 25 cm2 growth area) and were cultured for 3 days in M199 medium supplemented with 10% fetal bovine serum. The culture medium was replaced with lentivirus-containing medium for each group on day 2. Cells were harvested after 48-hour incubation. After trypsinization, cells were centrifuged at 314.5g for 10 minutes and washed in ice-cold PBS, then fixed overnight in 75% ethanol and stored at −20°C until assayed. For analysis, stored samples were centrifuged at 314.5g for 10 minutes and washed twice in PBS. The pelleted samples were resuspended and incubated in propidium iodide (Sigma) solution for 20 minutes at 4°C. RNase was added in the staining solution to avoid RNA contamination. The stained fibroblast population was gated with forward scatter versus side scatter plot to display the relationship of cell size versus granularity. Data were acquired using a FACScan flow cytometer (BD Biosciences, San Jose, CA), with cell counts obtained during 5 minutes of flow at 60 μL/min, to create a DNA-content frequency histogram and analyzed with Flowjo software (Tree Star Inc). A Dean–Jett–Fox model was then used to quantify cell-cycle phases, giving the percentages of cells in G0/G1, S, and G2/M. The doublet problem was resolved by a doublet discrimination gate.
KCNJ2 Overexpression and Dominant Negative Constructs
Lentiviral constructs were used, carrying wild-type or dominant negative KCNJ2 cDNAs or pWPI-plasmid negative control expressing green fluorescent protein (GFP) only, as previously reported.9 Lentivirus preparation was performed as previously described.10 LA fibroblasts from control dogs were isolated and grown in T75 flasks. At near confluence (3–4 days of culture), fibroblasts were trypsinized, counted, and plated in 12-well plates at 4.4×104 cells per well. After 4 to 6 hours of recovery, cells were transduced with lentiviral vectors at 50 multiplicity of infection. After overnight incubation (≈15 hours), cells were washed 3× with 10% fetal bovine serum–containing medium. After an additional 48 hours of culture, electrophysiological studies, Ca2+ imaging, or flow cytometry was performed.
For overexpression, sense and antisense oligonucleotides were synthesized by Invitrogen, and the double-stranded RNA was created by annealing. For knockdown, the anti-miR-26a oligonucleotide (AMO-26a) with locked nucleic acid chemistry was synthesized by Exiqon. Scrambled oligonucleotides with locked nucleic acid were used as negative controls (AMO-NC). LA fibroblasts of dogs in primary culture were transfected with AMO-26a (10 nmol/L) or AMO-NC (10 nmol/L) with Lipofectamine 2000 (Invitrogen). LA fibroblasts from control dogs were isolated and grown in T75 flasks. At near confluence (3–4 days of culture), fibroblasts were trypsinized, counted, and plated in 12-well plates at 4.4×104 cells per well. After 4 to 6 hours of recovery, cells were exposed to AMO-26a, AMO-NC, or vehicle in Lipofectamine. After overnight incubation, cells were washed 3× with 10% fetal bovine serum medium. After an additional 48 hours of transfection, cells were used for patch-clamp studies.
Taqman Real-Time Quantitative Reverse Transcription Polymerase Chain Reaction
Freshly isolated dog fibroblasts were resuspended in a lysis buffer, and RNA was isolated with Nucleospin RNA II (Macherey Nagel), including DNase treatment to prevent genomic contamination. Messenger RNAs were reverse-transcribed with the High-Capacity Reverse Transcription Kit (Applied Biosystems). Quantitative polymerase chain reaction was performed with TaqMan probes and primers from Applied Biosystems for housekeeping genes HPRT, β2-microglobulin, and G6PD, as well as for KCNJ2, collagen-1, collagen-3, fibronectin-1, fibrillin-1, and α-smooth muscle actin (α-SMA). SyBr green primers were used to quantify KCNJ12 and KCNJ4. The geometric mean expression of HPRT, β2-microglobulin, and G6PD was used for normalization. Quantitative polymerase chain reaction reactions were performed with Taqman Gene Expression Master Mix (Applied Biosystems). Reactions were run on a Stratagene MX3000. Relative gene expression values were calculated by the 2−ΔCt method.
Freshly isolated and cultured fibroblasts were rinsed with PBS and fixed for 10 minutes with 1:1 acetone:methanol at −20°C, and then cells were blocked for 1 hour with 5% bovine serum albumin at room temperature. The fibroblasts were incubated with mouse anti-α-SMA (1/500; Sigma), goat antivimentin (1/500; Santa Cruz), followed by donkey antimouse IgG-Alexa Fluor 555 (1/500; Invitrogen), donkey anti-rabbit IgG-Alexa Fluor 488 (1/500; Invitrogen), and TOPRO-3 iodide (1/1000; Invitrogen). Fluorescent images were obtained with an Olympus Fluoview FV1000 inverted confocal microscope.
Protein was extracted, quantified, and processed as we previously described.8 Freshly isolated fibroblasts were lysed in detergent-based buffer (150 mmol/L NaCl, 20 mmol/L Tris–HCl, pH 7.4, 1 mmol/L EDTA, 1 mmol/L EGTA, 1% Nonidet P-40, 1% Triton X-100, 1 mmol/L NaF, 1 mmol/L Na3VO4, and protease inhibitors). Protein samples were separated by gel electrophoresis and transferred to polyvinylidene difluoride membranes. Membranes were blocked and incubated with mouse anti-α-SMA (1/1000; Sigma), goat antivimentin (1/1000; Santa Cruz), anti-Kir2.1 (1:200; Neuromab), and GAPDH (1/10 000; Fitzgerald) antibodies. Secondary antibodies conjugated to horseradish peroxidase were used for detection via chemiluminescence.
Clampfit 9.2 (Axon), GraphPad Prism 4.0, and Origin 5.0 were used for data analysis. All data are expressed as mean±SEM. Multiple group statistical comparisons were obtained by 2-way ANOVA, and individual group mean differences were evaluated by Student t tests with Bonferroni correction. A 2-tailed P<0.05 was considered statistically significant.
Properties of the Model
CHF significantly increased right-atrial effective refractory period at all basic cycle lengths (Online Figure IA). Mean AF duration increased substantially in dogs with CHF (Online Figure IB). CHF reduced arterial pressures and increased filling pressures (Online Figure IC–IE). Atrial histopathology confirmed the presence of fibrosis in dogs with CHF (Online Figure II), and quantitative polymerase chain reaction confirmed enhanced extracellular matrix gene expression (Online Figure IIIA and IIIB).
CHF-Induced Changes in Fibroblast Phenotype
Bright field microscopic images of freshly isolated fibroblasts from each group are shown in Figure 1A, with CHF fibroblasts being systematically larger (Figure 1B–1D). Immunoflorescence suggested that CHF fibroblasts exhibited enhanced vimentin and α-SMA expression (Figure 1E), an impression confirmed by image quantification (Figure 1F and 1G). Western blot analysis further supported CHF-induced fibroblast α-SMA and vimentin upregulation (Figure 1H and 1I). Gene expression mRNA levels corresponding to the extracellular matrix proteins collagen-1, collagen-3, fibronectin-1, and fibrillin-1 were all greater in freshly isolated CHF fibroblasts versus control fibroblasts (Online Figure IIIA and IIIB). These observations indicate that CHF alters atrial fibroblast phenotype.
Changes in IK1 and RMP
Figure 2A and 2B show examples of IK1 recordings from freshly isolated fibroblasts obtained with the ramp protocol shown in the inset. The current was strongly suppressed by 300 μmol/L Ba2+. Overall data (Figure 2C) indicate significantly larger IK1 in CHF. CHF significantly increased the expression of KCNJ2 (Kir2.1) at both mRNA (by 52%) and protein (by 80%) levels (Figure 2D and 2E). KCNJ12 (Kir2.2) mRNA expression (Online Figure IIIC) was 2 orders of magnitude less than that of KCNJ2 (Online Figure IIID) and was not affected by CHF. KCNJ4 (Kir2.3) was undetectable. CHF significantly increased RMP (from −43.2±0.8 mV, control, to −55.5±0.9 mV; P<0.001; Figure 2F). Ba2+ (300 μmol/L) significantly reduced RMP and greatly attenuated the RMP differences between control and CHF fibroblasts, suggesting that they were due to IK1 upregulation in CHF (Figure 2F). Similar effects were seen with a 10-fold larger Ba2+ concentration (3 mmol/L; Figure 2G). Of note, despite the statistically significant reduction in RMP with Ba2+, the fibroblasts maintained a negative RMP, indicating a contribution from conductances other than IK1.
To further verify the role of IK1 in fibroblast RMP changes with CHF, we performed the studies shown in Online Figure IV. Consistent with expected behavior, the current conductance (and particularly the inward current component) was greatly enhanced (Online Figure IVA and IVB) by increasing extracellular K+ concentration ([K+]o) to 75 mmol/L (equimolar substitution for Na+). In addition to enhancing conductance (Online Figure IVC and IVD), increased [K+]o shifted the reversal potential positively, from −53.4±3.7 mV and −64.3.7±3.8 mV with 5.4 mmol/L [K+]o in controls and dogs with CHF, respectively, to −14.9±1.1 mV (P<0.001) and −15.1±1.1 mV (P<0.001) in 75 mmol/L [K+]o. The RMP was significantly reduced by 75 mmol/L [K+]o for both control and CHF conditions (Online Figure IVE and IVF) and elevating [K+]o largely eliminated the RMP differences between controls and dogs with CHF, with values averaging −30.9±1.5 mV and −32.5±1.6 mV, respectively, in 75 mmol/L [K+]o (P=not significant).
Changes in Fibroblast Ca2+ Entry
Figure 3A and 3B show store-dependent Ca2+ entry data from control and CHF fibroblasts, respectively. Cells in short-term (20 hours) culture were first exposed to nominally Ca2+-free extracellular solution to deplete Ca2+ stores, and then Ca2+ entry was observed on increasing [Ca2+]o to 10 mmol/L (Figure 3A and 3B). Resting [Ca2+]i (Figure 3C) and store-dependent Ca2+ entry (Figure 3D) were greater in CHF cells. We were unable to use Ba2+ as a probe to inhibit IK1 and study the role of IK1 differences in Ca2+ entry because Ba2+ interacts directly with Fura-2.11 However, increasing [K+]o to reduce RMP substantially suppressed store-dependent Ca2+ entry under both control and CHF conditions and greatly reduced the difference between control and CHF values (Figure 3D). These results suggest that RMP is a significant determinant of fibroblast Ca2+ entry and that RMP differences due to IK1 remodeling may contribute to the increased Ca2+ entry caused by CHF.
Fibroblast Proliferation and Differentiation
To study CHF-induced changes in fibroblast proliferation and differentiation, along with the potential contribution of IK1 remodeling, we had to perform experiments with short-term (3 days) cultured cells. We first verified that 3-day culture does not alter fibroblast IK1 or RMP (Online Figure V). We then collected fibroblasts for proliferation analysis by flow cytometry. Figure 4A and 4B show representative DNA-content histograms and Dean–Jett–Fox model fitting of control and CHF atrial fibroblasts (G0: resting phase; G1 phase: increased size and ready for DNA synthesis; G2/M phase: cells with doubled DNA content in premitotic and mitotic phases). CHF increased the total cell count and cell content in the G2/M phase (Figure 4C and 4D). Mean total cell counts are shown in Figure 4C and percentages in each phase in Figure 4D. CHF significantly increased the percentage of cells in the G2/M phase (Figure 4D), indicating increased proliferation. Cultured CHF fibroblasts also showed properties indicating greater myofibroblast differentiation versus control fibroblasts, including altered cell morphology and greater expression of vimentin and α-SMA protein (Online Figure VI).
IK1 Regulates RMP, Ca2+ Entry, and Proliferation
The experiments shown in Figure 3 suggest that IK1 may contribute to the control of atrial fibroblast Ca2+ entry. To explore the functional contributions of IK1 more directly, we used a gene transfer approach to vary the current in fibroblasts. A dominant negative KCNJ2 construct (KCNJ2-DN) with the GYG motif replaced by a triple-alanine (AAA) sequence was used to suppress endogenous KCNJ2 current. Wild-type KCNJ2 overexpression (KCNJ2-OE) was used to enhance the current. KCNJ2-DN and wild-type KCNJ2 were packed into a lentivirus vector containing GFP. A control virus was also prepared that contained only GFP inserted into the lentiviral vector. Infected cells were identified by green fluorescence.
Figure 5A shows original recordings of IK1 in fibroblasts infected with controls (lentivirus carrying GFP alone), KCNJ2-DN, and KCNJ2-OE constructs, before and after the addition of 300 μmol/L of Ba2+ to the superfusate. Currents were reduced by KCNJ2-DN and increased by KCN2-OE and strongly suppressed by Ba2+. Figure 5B shows current–voltage relationships, with an inset showing control and KCNJ2-DN currents on an expanded current scale for clearer resolution. KCNJ2-OE greatly increased IK1, whereas KCNJ2-DN strongly reduced IK1 density. Compatible with a role of IK1 in fibroblast differentiation, cell capacitance increased with KCNJ2-OE (Figure 5C). KCNJ2-OE substantially hyperpolarized the RMP, and consistent with the effects of Ba2+ shown in Figure 2F and 2G, KCNJ2-DN significantly reduced RMP. Exposure to Ba2+ eliminated the RMP differences among constructs (Figure 5D and 5E) with full effects seen at 300 μmol/L, consistent with the notion that the RMP differences are caused by changes in IK1.
We then went on to use the gene transfer approach to confirm directly the ability of IK1 to regulate atrial fibroblast Ca2+ entry and proliferation. KCNJ2-DN decreased, and KCNJ2-OE increased, the resting Ca2+ level (Figure 6A) and store-operated Ca2+ entry (Figure 6B). Increased [K+]o attenuated Ca2+ entry in the presence of the lentiviral-GFP control vector and KCNJ2-OE (Figure 6B), but as expected given the virtual elimination of IK1 produced by KCNJ2-DN, elevated [K+]o did not alter Ca2+ entry in the presence of KCNJ2-KD. Fibroblast proliferation indices were enhanced by KCNJ2-OE and suppressed by KCNJ2-DN (Figures 6C and 6D). These results confirm the role of IK1 in governing fibroblast Ca2+ entry and proliferation.
Role of RMP in Controlling Fibroblast Ca2+ Entry
The most obvious way in which changes in IK1 could affect fibroblast Ca2+ entry is through resulting changes in RMP and the voltage gradient driving Ca2+ into the cell. To test directly the effect of RMP on fibroblast Ca2+ entry, we studied store-operated Ca2+ entry in fibroblasts under voltage-clamp conditions, with voltages set to approximate the RMP of control fibroblasts (−40 mV) and CHF fibroblasts (−55 mV). Figure 7A shows [Ca2+]i recordings in one cell, obtained at a holding potential of −40 mV (left) and then in the same cell at −55 mV (right). Hyperpolarizing the fibroblast increased the Ca2+-transient amplitude. Figure 7B shows the mean Ca2+-transient amplitude in cells in which we were able to study Ca2+ entry under stable conditions under both voltages (order randomized in different cells). RMP had a highly significant effect on Ca2+-transient amplitude.
MicroRNA-26 Regulation of Atrial Fibroblast IK1
The larger increase in the KCNJ2 protein than mRNA in CHF points to mediation by microRNA. We have previously shown that miR-26 targets IK1 and that its downregulation in cardiomyocytes from animals with sustained AF governs cardiomyocyte IK1 enhancement.12 We therefore considered the possibility that miR-26 regulation may contribute to the atrial fibroblast IK1 enhancement that we observed in CHF. Expression of the miR-26a isoform was decreased in freshly isolated LA fibroblasts from dogs with CHF, whereas miR-26b was unaffected (Figure 8A). AMO-26a transfection into atrial fibroblasts with lipofectamine effectively suppressed miR-26a (Figure 8B, left). To exclude nonspecific effects, we examined miR-21 expression, which was unaffected by AMO-26a (Figure 8B, right). We then looked at the result of knocking down miR-26a, to mimic its downregulation in CHF, on IK1 in atrial fibroblasts. Figure 8C shows original recordings from a fibroblast exposed to lipofectamine alone, a fibroblast exposed to a scrambled-control oligonucleotide (AMO-NC), and a fibroblast transfected with AMO-26a. IK1 was clearly larger after AMO-26a exposure, as indicated by the mean current–voltage data in Figure 8D. Finally, we examined the effect of miR-26a knockdown on atrial fibroblast RMP and noted substantial hyperpolarization (Figure 8E).
In this study, we analyzed the consequences of CHF-induced IK1 upregulation in fibroblasts on fibroblast function, noting hyperpolarized RMP, enhanced Ca2+ entry, and increased proliferation indices. The mechanistic role of IK1 changes was supported by genetically modifying IK1 through KCNJ2-OE and knockdown, and the potential contribution of hyperpolarization to CHF-induced fibroblast Ca2+ entry increases was demonstrated by simultaneous voltage clamp and Ca2+ microfluorometry. CHF-induced miR-26a downregulation was implicated as the mechanism of KCNJ2/ IK1 upregulation.
Functional Role of Ion Channels in Cardiac Fibroblasts
Although the presence of ion channels in cardiac fibroblasts is well established,3 their functional role is less clear. Ca2+ entry via nonselective cation channels of the transient receptor potential family plays a role in fibroblast proliferation, differentiation, and extracellular matrix protein secretion.8,13 There is evidence that this action is mediated via Ca2+-dependent activation of extracellular signal–related protein kinases14 and contributes to the AF-related arrhythmogenic substrate.8,13 The function of fibroblast K+ channels is less clear. Previous work has indicated that Kir channels contribute to RMP determination. In cell coculture systems, myofibroblasts can be shown to couple to cardiomyocytes and alter their electrophysiological properties, inducing a variety of arrhythmogenic mechanisms.14,15 Although the importance of fibroblast–cardiomyocyte coupling in vivo is still controversial, mathematical modeling work suggests that it may account for complex fractionated electrogram properties in fibrotic tissues.16 We have recently evaluated the effects of fibroblast ion channel remodeling on the potential electric and arrhythmogenic interactions between coupled fibroblasts and cardiomyocytes, finding that if fibroblasts were well coupled to cardiomyocytes, fibroblast Kv current downregulation would suppress the AF-substrate, whereas Kir current upregulation would enhance it.17 We have also obtained evidence for a profibrotic role of Kv current downregulation, although the underlying mechanism is unclear.6
Control of Fibroblast Ca2+ Entry and Function by Kir Currents
This study is the first of which we are aware to show the control of fibroblast Ca2+ entry and proliferation by Kir2.1 current. Several lines of evidence converged to clarify the role of fibroblast IK1. IK1 block with Ba2+ or K+ driving force reduction by elevating [K+]o reduced the RMP and elevated [K+]o reduced fibroblast store-operated Ca2+ entry. Functional KCNJ2 knockdown with dominant negative overexpression reduced atrial fibroblast IK1, RMP, Ca2+ influx, and proliferative activity, whereas KCNJ2-OE had the opposite effects. The role of RMP in mediating IK1 effects on Ca2+ entry was directly supported by experiments showing that hyperpolarization of voltage-clamped fibroblasts enhanced fibroblast store-operated Ca2+ entry.
Although this functional role has never before been described in fibroblasts, there is supportive evidence from previous work in endothelial cells. Bradykinin-induced changes in bovine endothelial cell cytosolic Ca2+ are consistent with an influx mechanism directly related to the Ca2+ electrochemical gradient.18 Nitric oxide synthesis and proliferation of umbilical cord endothelial cells induced by basic fibroblast growth factor seem to depend on inward-rectifier K+ current augmentation.19 Finally, hyperpolarization increases cytoplasmic [Ca2+] in arteriolar endothelial cells.20
Novel Elements and Potential Significance
Cardiac fibrosis is an important contributor to cardiac dysfunction and arrhythmogenesis, and it is a particularly significant contributor to the substrate that allows enhanced AF maintenance in CHF.21 Fibroblasts play a central role in the fibrotic process.22 Here, we addressed a novel regulatory aspect of fibroblast physiology, functional control by IK1, along with the remodeling of IK1 and its contribution to altered fibroblast function in a clinically relevant fibrotic paradigm: the CHF-induced atrial profibrillatory substrate.1,2 We report for the first time that CHF-related atrial IK1 upregulation and consequent fibroblast hyperpolarization enhance fibroblast Ca2+ entry and cell proliferation. This work identifies a novel participant in the profibrotic response, with potential implications for the development of novel therapeutic interventions. Ion channels are targets for new antiarrhythmic agents.23 Our study shows that in addition to altering cardiac electrophysiology, interventions that target IK1 may affect cardiac structural remodeling, particularly because the principal IK1 subunit, KCNJ2/Kir2.1, is common to both cardiomyocytes and fibroblasts. The predominance of KCNJ2 in cardiac IK1 is well recognized.24 This study indicates that KCNJ2 is similarly predominant in fibroblasts: DN-KCNJ2 almost completely eliminated fibroblast IK1 (Figure 5B, inset). IK1 blockers are being developed as potential antiarrhythmic molecules,25,26 based on their ability to inhibit cardiomyocyte IK1 and destabilize AF-maintaining rotors.26,27 An additional potentially interesting consequence, based on the work reported here, might be the suppression of atrial fibrosis. The converse may also hold. Fibroblasts engineered to overexpress Kir2.1, Nav1.5, and connexin-43 subunits rescue normal propagation and decrease arrhythmia complexity in cocultured cardiomyocyte–fibroblast monolayers.28 A risk of applying this approach therapeutically might be the profibrotic consequences of increased IK1.
MicroRNAs are significant control molecules in cardiac remodeling.29 We have previously shown that miR-26 downregulation contributes to AF-promoting remodeling by upregulating cardiomyocyte IK1.12 Here, we have identified an additional potential profibrillatory consequence of disease-related miR-26 downregulation: fibroblast activation via fibroblast IK1 upregulation consequent to removal of miR-26 induced negative regulation of the KCNJ2 gene.
Fibroblast hyperpolarization significantly increased store-operated Ca2+ entry by ≈30% (Figure 7B); however, CHF fibroblasts showed an ≈70% increase in Ca2+ entry (Figure 3D). Thus, the hyperpolarization caused by IK1 upregulation is likely not the only factor increasing Ca2+ entry in CHF fibroblasts. In addition to KCNJ2, miR-26 controls the expression of the gene encoding TRPC3 subunits.8 TRPC3 subunit upregulation caused by miR-26 downregulation in AF enhances fibroblast Ca2+ entry.8 Thus, TRPC3 expression changes caused by CHF-induced miR-26a downregulation likely also contributed to the increased fibroblast Ca2+ entry observed in CHF fibroblasts in this study.
We performed experiments in isolated fibroblasts to evaluate their cell biology in detail. Analysis of fibroblast properties in situ is greatly complicated by their small size and a dearth of specific probes. Paracrine effects in vivo could significantly alter fibroblast behavior and were not analyzed here. In addition, in this study we measured the fibrous-tissue content only in the LA appendage (Online Figure II). The distribution of fibrosis may not be uniform in atria.
Fibroblast proliferation might be affected by lentiviral infection. We therefore verified cell counts on culture in control fibroblasts versus lentivirus-GFP infected fibroblasts and found no significant differences (Online Figure VII).
We used Ba2+ as one of several tools to compare the contribution of IK1 in CHF fibroblasts with that in control. Ba2+ can affect a variety of K+ currents, and the possibility of nonspecific effects requires caution in the interpretation of data. Evidence against any significant nonspecific effects of Ba2+ in our system is provided by the results of KCNJ2 knockdown on the response to Ba2+ (Online Figure VIII). Ba2+ had no significant effect on currents once KCNJ2/ IK1 was knocked down, indicating the absence of any significant effect on other currents under our recording conditions.
We thank Nathalie L’Heureux, Chantal St-Cyr, and Audrey Bernard for technical assistance and France Thériault for secretarial help with the article.
Sources of Funding
This work was supported by the Canadian Institutes of Health Research (44365, 6957) and the Heart and Stroke Foundation of Canada and the Fondation Leducq (European-North American Atrial Fibrillation Research Alliance; 07CVD03).
In December 2014, the average time from submission to first decision for all original research papers submitted to Circulation Research was 14.47 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.116.305326/-/DC1.
- Nonstandard Abbreviations and Acronyms
- α-smooth muscle actin
- atrial fibrillation
- anti-miR oligonucleotide
- congestive heart failure
- green fluorescent protein
- dominant negative KCNJ2 construct
- KCNJ2 overexpression
- left atrial
- resting membrane potential
- Received September 23, 2014.
- Revision received January 16, 2015.
- Accepted January 21, 2015.
- © 2015 American Heart Association, Inc.
- Andrade J,
- Khairy P,
- Dobrev D,
- Nattel S
- Li D,
- Fareh S,
- Leung TK,
- Nattel S
- Yue L,
- Xie J,
- Nattel S
- Kamkin A,
- Kiseleva I,
- Isenberg G
- Chilton L,
- Ohya S,
- Freed D,
- George E,
- Drobic V,
- Shibukawa Y,
- Maccannell KA,
- Imaizumi Y,
- Clark RB,
- Dixon IM,
- Giles WR
- Wu CT,
- Qi XY,
- Huang H,
- Naud P,
- Dawson K,
- Yeh YH,
- Harada M,
- Kuo CT,
- Nattel S
- Harada M,
- Luo X,
- Qi XY,
- et al
- Thomas J,
- Epshtein Y,
- Chopra A,
- Ordog B,
- Ghassemi M,
- Christman JW,
- Nattel S,
- Cook JL,
- Levitan I
- Dawson K,
- Wakili R,
- Ordog B,
- Clauss S,
- Chen Y,
- Iwasaki Y,
- Voigt N,
- Qi XY,
- Sinner MF,
- Dobrev D,
- Kaab S,
- Nattel S
- Du J,
- Xie J,
- Zhang Z,
- Tsujikawa H,
- Fusco D,
- Silverman D,
- Liang B,
- Yue L
- Rohr S
- Ashihara T,
- Haraguchi R,
- Nakazawa K,
- Namba T,
- Ikeda T,
- Nakazawa Y,
- Ozawa T,
- Ito M,
- Horie M,
- Trayanova NA
- Laskey RE,
- Adams DJ,
- Johns A,
- Rubanyi GM,
- van Breemen C
- Scharbrodt W,
- Kuhlmann CR,
- Wu Y,
- Schaefer CA,
- Most AK,
- Backenköhler U,
- Neumann T,
- Tillmanns H,
- Waldecker B,
- Erdogan A,
- Wiecha J
- Souders CA,
- Bowers SL,
- Baudino TA
- Takanari H,
- Nalos L,
- Stary-Weinzinger A,
- de Git KC,
- Varkevisser R,
- Linder T,
- Houtman MJ,
- Peschar M,
- de Boer TP,
- Tidwell RR,
- Rook MB,
- Vos MA,
- van der Heyden MA
- Filgueiras-Rama D,
- Martins RP,
- Mironov S,
- Yamazaki M,
- Calvo CJ,
- Ennis SR,
- Bandaru K,
- Noujaim SF,
- Kalifa J,
- Berenfeld O,
- Jalife J
- Kumarswamy R,
- Thum T
Novelty and Significance
What Is Known?
Cardiac fibroblasts play a central role in tissue fibrosis, which are an important contributor to a variety of arrhythmias, including atrial fibrillation.
Cardiac fibroblasts possess a range of ion channels, but the functional role of fibroblast ion channels is poorly understood.
Congestive heart failure (CHF) causes prominent atrial fibrosis, and is, therefore, a major risk factor for the arrhythmia.
What New Information Does This Article Contribute?
CHF leads to the upregulation of the background inward-rectifier potassium current (IK1) in atrial fibroblasts, likely by downregulating a microRNA (miR-26) that targets IK1.
An increase in atrial fibroblast IK1 hyperpolarizes the cell membrane, enhancing Ca2+ entry by increasing the driving force for transmembrane Ca2+ movement.
These findings define a new pathway for CHF-induced atrial fibrosis involving an increase in IK1, leading to hyperpolarization and enhanced Ca2+ entry and resulting in fibroblast activation.
CHF is an important clinical risk factor for atrial fibrillation, the commonest sustained arrhythmia and a major source of morbidity and mortality. Although CHF-induced atrial fibrosis is thought to contribute to the atrial fibrillation substrate, the mechanisms leading to this fibrosis are poorly understood. Moreover, the functional roles of ion channels in fibroblasts, the cells that produce fibrosis, are not well understood. Here, we studied changes in the background inward-rectifier potassium current (IK1) in atrial fibroblasts from dogs with CHF induced by ventricular tachypacing. We found significant upregulation of IK1, which hyperpolarized the resting membrane potential of the fibroblasts. This hyperpolarization enhanced Ca2+ entry, known to be an important fibroblast-activating mechanism, causing fibroblasts to proliferate. We also found that IK1 upregulation is caused by CHF-induced decreases in atrial fibroblast expression of a microRNA, miR-26, that targets the gene (KCNJ2) encoding IK1. These findings reveal a new function of fibroblast potassium channels, ie, the control of fibroblast activation, and show that these channels can mediate a pathological arrhythmia–promoting response. Further elucidation of this signaling pathway would provide novel insights into the mechanisms controlling atrial fibrillation and might allow for the development of new therapeutic approaches.