Improvement of Cardiomyocyte Function by a Novel Pyrimidine-Based CaMKII-Inhibitor

Abstract

Objective

Pathologically increased activity of Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) and the associated Ca²⁺-leak from the sarcoplasmic reticulum are recognized to be important novel pharmacotherapeutic targets in heart failure and cardiac arrhythmias. However, CaMKII-inhibitory compounds for therapeutic use are still lacking. We now report on the cellular and molecular effects of a novel pyrimidine-based CaMKII inhibitor developed towards clinical use.

Methods and results

Our findings demonstrate that AS105 is a high-affinity ATP-competitive CaMKII-inhibitor that by its mode of action is also effective against autophosphorylated CaMKII (in contrast to the commonly used allosteric CaMKII-inhibitor KN-93). In isolated atrial cardiomyocytes from human donors and ventricular myocytes from CaMKIIδ_C-overexpressing mice with heart failure, AS105 effectively reduced diastolic SR Ca²⁺ leak by 38% to 65% as measured by Ca²⁺-sparks or tetracaine-sensitive shift in [Ca²⁺]_i. Consistent with this, we found that AS105 suppressed arrhythmogenic spontaneous cardiomyocyte Ca²⁺-release (by 53%). Also, the ability of the SR to accumulate Ca²⁺ was enhanced by AS105, as indicated by improved post-rest potentiation of Ca²⁺-transient amplitudes and increased SR Ca²⁺-content in the murine cells. Accordingly, these cells had improved systolic Ca²⁺-transient amplitudes and contractility during basal stimulation. Importantly, CaMKII inhibition did not compromise systolic fractional Ca²⁺-release, diastolic SR Ca²⁺-reuptake via SERCA2a or Ca²⁺-extrusion via NCX.

Conclusion

AS105 is a novel, highly potent ATP-competitive CaMKII inhibitor. In vitro, it effectively reduced SR Ca²⁺-leak, thus improving SR Ca²⁺-accumulation and reducing cellular arrhythmogenic correlates, without negatively influencing excitation-contraction coupling. These findings further validate CaMKII as a key target in cardiovascular disease, implicated by genetic, allosteric inhibitors, and pseudo-substrate inhibitors.

Introduction

Heart failure is an important healthcare burden with debilitating consequences for the affected patients and high socioeconomic impact. Recent therapeutic advances for heart failure have been mainly in the field of device-based therapy, with the exception of Sacubitril, a neprilysin inhibitor approved by the FDA in 2015 in combination with valsartan. Despite these limited advances, the morbidity and mortality amongst heart failure patients remains high. In addition to pump failure, heart failure patients die from cardiac arrhythmias (sudden cardiac death) and both atrial and ventricular arrhythmias can worsen the progression of heart failure. Thus, additional strategies for therapeutic intervention in heart failure and cardiac arrhythmias are still needed.

In heart failure, neurohumoral and cytokine activation acutely serve compensatory purposes, but such chronic compensation leads to cardiac remodeling that further worsens the condition, resulting in a vicious cycle of increasingly impaired contractility and eventually myocardial failure. Pharmacotherapies such as beta blockers or ACE-inhibitors/AT2-antagonists intervene in this vicious cycle by antagonizing the compensatory stimuli.

Importantly, dysregulation of cardiac excitation-contraction couplings have been identified to be central to heart failure and cardiac arrhythmias (as reviewed in (1,2)). Specifically, diastolic Ca²⁺-leak from the sarcoplasmic reticulum (SR) through the ryanodine receptor (RyR2) is regarded as critically contributing to impaired cardiomyocyte contractility by depleting SR Ca²⁺-content, making less Ca²⁺ available for contraction (3–5). Furthermore, SR Ca²⁺-leak appears to be a mechanism for triggering atrial (6–8) as well as ventricular (9–11) arrhythmogenesis.

Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) is a multifunctional serine/threonine protein kinase that is of special interest in this cardiac pathophysiology. CaMKII is an important regulator of cardiac excitation-contraction coupling and Ca²⁺-homeostasis (as reviewed in (1,12)) by regulating the function of L-type Ca²⁺-channels, SR Ca²⁺-ATPase (SERCA2a, via phosphorylation of phospholamban (PLB) at Thr-17), and RyR2 (via phosphorylation at Ser-2814). CaMKII can furthermore activate pathways involved in hypertrophy, inflammation, and apoptosis (as reviewed in (13)). In heart failure (5,14,15) and atrial fibrillation (7,8), CaMKII is overactivated by stimuli that elevate Ca²⁺ and reactive oxygen species. Importantly, CaMKII-dependent RyR2 Ser-2814 hyper-phosphorylation has been demonstrated to induce SR Ca²⁺-leak (4,16,17). Accordingly, direct genetic or pharmacological inhibition of CaMKII as well as other disruptions of CaMKII-dependent pathways have shown promise for treating heart failure or cardiac arrhythmias (18–20). Consequently, CaMKII has become increasingly recognized as a consensus target for pharmacological intervention in heart disease. However, compounds suitable for use in humans are still lacking (21).

AS105 (Fig. 1A) is a novel CaMKII-inhibitor being advanced by Allosteros Therapeutics following computational optimization of pyrimidine-based CaMKIIδ inhibitors (21,22). We have explored the therapeutic potential of AS105 using isolated murine ventricular and human atrial cardiomyocytes. The field has largely relied on a tool inhibitor, KN-93, which is allosteric and non-ATP competitive, but is of low potency and inhibits multiple ion channels and other off-targets²¹. And this study also aims to evaluate the utility of an ATP-competitive inhibitor for cardiovascular studies of CaMKII action.

Figure 1. Enzymatic Characterization of AS105.

(A) Structure of AS105. (B): AS105 inhibits CaMKIIδ with low nanomolar affinity: We used a continuous spectrophotometric assay to follow the phosphorylation of a peptide substrate with recombinantly-expressed and purified CaMKIIδ at increasing AS105 concentrations. The IC50 is determined using standard non-linear regression analysis. (C): AS105 inhibition of CaMKIIδ is ATP-competitive: CaMKIIδ kinase activity is measured while varying the concentration of ATP at different AS105 concentrations. The inhibitory pattern indicates a competitive mode of inhibition against ATP. (D): Inhibition of CaMKIIδ activity by AS105 is independent of kinase autophosphorylation: Prior to the kinase activity assay CaMKIIδ was subjected to an autophosphorylation reaction in the presence of Ca²⁺/CaM and ATP (CaMKII-P) or to a control reaction lacking ATP (CaMKII) for 5 minutes at room temperature. The enzyme was then assayed for activity as normal in the presence of increasing concentrations of AS105 (pink symbols) or KN-93 (green symbols) to calculate inhibitor IC50s.

Mice with transgenic overexpression of CaMKIIδ_C develop severe heart failure associated with 50% reduction of twitch contractility and systolic Ca²⁺-transient amplitude due to over 50% reduction of SR Ca²⁺-content (16), which in turn has been attributed to about 2- to 3-fold higher SR Ca²⁺-leak (16,23) in these mice. Expression levels of SERCA2a, PLB and RyR2 are reduced in CaMKIIδ_C-TG hearts, while the expression of NCX is increased (16), in line with the heart failure phenotype.

Methods

CaMKIIδ purification and enzyme activity assays

Hexahistidine tagged human CaMKIIδ was expressed in High Five™ insect cells using a baculovirus expression system (Thermo Fisher Scientific, Waltham, MA, USA). Purification of the CaMKIIδ holoenzyme was accomplished by Ni-NTA affinity chromatography using a HisTrap HP column (GE Healthcare Life Sciences, Marlborough, MA, USA) followed by HiPrep 26/60 Sephacryl S-300 (GE Healthcare Life Sciences, Marlborough, MA, USA) size exclusion chromatography. The purified enzyme was stored at −80 °C in 25 mM Tris HCl pH 8, 300 mM KCl, 30 % glycerol (v/v), and 2 mM TCEP. The purity of the final preparation was >95% as assessed by SDS-PAGE and Coomassie Blue staining.

CaMKIIδ phosphotransferase activity was monitored using a continuous spectrophotometric assay as described previously (24). In this assay, the phosphorylation of a peptide substrate by CaMKIIδ is enzymatically coupled to the oxidation of NADH to NAD⁺, which produces a decrease in absorbance at 340 nm. The reactions were carried out in a final volume of 150 µl in 100 mM Tris HCl pH 7.5, 150 mM KCl, 0.2 mM CaCl₂, 10 mM MgCl₂, 0.1 mM ATP, 1% DMSO, 1 mM phosphoenolpyruvate, 0.28 mM NADH, 89 units/ml pyruvate kinase, 124 units/ml lactate dehydrogenase (Sigma-Aldrich), 150 µM autocamtide-3 (KKALHRQETVDAL; Biomer Technology), 60 nM CaM (EMD Millipore), and 5–10 nM CaMKIIδ. The enzyme concentration is expressed in terms of the concentration of kinase subunits and not holoenzymes. Reactions were initiated by the addition of CaM and ATP to the mix and the decrease in absorbance at 340 nm was monitored at 30 °C using a microplate spectrophotometer, SpectraMax Plus (Molecular Devices, Sunnyvale, CA, USA).

Autophosphorylation of CaMKIIδ was accomplished by incubating 37.5 nM CaMKIIδ with 450 nM CaM, 0.2 mM ATP in a buffer containing 80 nM Tris pH 7.4, 140 mM KCl, 10 mM Mg²⁺, and 0.5 mM Ca²⁺. A control reaction containing all the above components except ATP was processed in parallel. After 5 min at room temperature 20 µl of each reaction was assayed for kinase activity in the presence of increasing concentrations of AS105 or KN-93. The final components of the kinase activity assay (150 µl final volume) were 5 nM CaMKIIδ, 60 nM CaM, 150 µM autocamtide-3, 0.13 mM ATP, 92 mM Tris pH 7.4, 159 mM KCl, 11 mM Mg²⁺, 0.57 mM Ca²⁺, and 1% DMSO. In addition, 20 µl of each reaction were also identically assayed for kinase activity in the presence of 3 mM EGTA to verify the presence or absence of autophosphorylation.

Steady state initial velocity data were calculated from the slopes of the A340 curves and standard non-linear regression analysis was used to determine kinetic parameters using GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA, USA). The inhibition constant Ki was calculated using the equation Ki=IC50/(1+[ATP]/K_m). Using Michaelis-Menten analyses the K_m for ATP at the described experimental conditions was experimentally determined to be K_mATP=49 µM.

Sources of cardiac tissue

CaMKIIδ_C-overexpressing mice (25) and the respective wildtype (WT) mice were bred at the central animal research facilities of the University Medical Center Göttingen and the University Medical Center Regensburg. All investigations conformed to Directive 2010/63/EU of the European Parliament and the NIH Guide for the Care and Use of Laboratory Animals. Right atrial appendages from human patients were obtained during coronary artery bypass surgery in patients in sinus rhythm without higher degree valve dysfunction. Written consent had been given by the donors and approval had been granted by the local ethics committee. Clinical parameters of the donors are displayed in Suppl. table 1.

Isolation of murine cardiomyocytes

Isolation of cardiomyocytes was performed as previously described (17). Briefly, CaMKIIδ_C-overexpressing mice or the respective wildtype mice (C57BL6-J strain) were anaesthetized with isoflurane. After death by cervical dislocation, hearts were quickly excised, mounted on a Langendorff perfusion apparatus and retrogradely perfused with nominally Ca²⁺-free solution containing (in mM) 113 NaCl, 4.7 KCl, 0.6 KH₂PO₄, 0.6 Na₂HPO₄, 1.2 MgSO₄, 12 NaHCO₃, 10 KHCO₃, 10 HEPES, 30 taurine, 10 BDM, 5.5 glucose, 0.032 phenol-red for 4 min at 37°C (pH 7.4). Then, 7.5 mg/ml liberase™ (Roche diagnostics, Mannheim, Germany), trypsin 0.6%, and 0.125 mM CaCl₂ were added to the perfusion solution. Perfusion was continued for 3–4 min until the heart became flaccid. Ventricular tissue was collected in perfusion buffer supplemented with 5% bovine calf serum, cut into small pieces, and dispersed by repeatedly pipetting until no solid cardiac tissue was left. Ca²⁺-reintroduction was performed by stepwise increasing [Ca²⁺] from 0.1 to 1.7 mM. For measurements, cells were freshly plated onto superfusion chambers, which had been coated with laminin.

Isolation of human cardiomyocytes

Cardiomyocytes were isolated from human right atrial samples using the chunk-isolation technique. Samples were rinsed, cut into small pieces and incubated in a spinner flask with a solution consisting of (in mM) 100 NaCl, 10 KCl, 5 MgCl₂, 1.2 KH₂PO₄, 50 taurine, 5 MOPS, 10 BDM, 20 glucose (pH 7.2 at 37°C). After 10 min, CaCl₂ was added for a concentration of 0.02 mM. For the first digestion step, this solution was supplemented with 0.775 mg/ml collagenase type I (370 U/ml; Worthington) and also 0.4 mg/ml protease type XXIV (Sigma-Aldrich). After 45 min, the supernatant was discarded and solution containing collagenase used and the tissue was repeatedly aggregated by pipetting in addition to stirring until sufficient numbers of free cells appeared in the solution. Then, stopping solution was added to final levels of bovine calf serum (BCS) 2% and BDM to 20 mM and the supernatant with cells was centrifuged. Cells were then resuspended in storage medium containing (in mM) 30 KCl, 10 KH₂PO₄, 1 MgCl₂, 10 HEPES, 11 glucose, 20 taurine, 70 glutamic acid, 20 BDM, 2% BCS (pH 7.4 with KOH). Cells were plated onto superfusion chambers coated with laminin and left to settle for 1 h.

Incubation with AS105

2 µM AS105 (or vehicle in control groups: “Ctrl”) was included in the loading buffers for Ca²⁺-dyes for per-incubation and in the respective superfusion solutions.

Experimental solutions

For experiments using murine cardiomyocytes, normal Tyrode’s solution was used consisting of (in mM) 140 NaCl, 4 KCl, 5 HEPES, 1 MgCl₂, 10 glucose, 2 CaCl₂ (pH 7.4 with NaOH). For experiments with human myocytes, the experimental solution consisted of (mM) 4 KCl, 136 NaCl, 1.6 MgCl₂, 10 HEPES, 0.33 NaH₂PO₄, 4 NaHCO₃, 10 glucose, 2 CaCl₂ (pH 7.4 with NaOH). Loading buffers for Ca²⁺-fluorescent dye consisted of the respective experimental solution with Pluronic F-127 0.2 mg/ml in addition to the fluorescent dye. All experiments were performed at room temperature.

Investigation of EC-coupling

Ca²⁺-epifluorescence and sarcomere length measurements were performed using an epifluorescence setup (IonOptix Corp, Milton, USA) mounted to a Nikon TE2000U microscope (17). Myocytes were loaded with Fluo-4 AM (10 µM) for 15 min. Excitation was at 480±15 nm and F/F₀ of the fluorescence signal was calculated. Cells were transilluminated by red light (>650 nm) and cardiomyocyte contractility (fractional shortening) was measured in parallel to epifluorescence using a sarcomere length detection system (IonOptix Corp, Milton, USA). Cardiomyocytes were field-stimulated at 1 Hz until steady-state was achieved. Post-rest potentiation of Ca²⁺-transient amplitude was investigated following 10 s of stimulation cessation to assess the SRs ability to accumulate Ca²⁺. SR Ca²⁺-content was estimated by rapid application of 10 mM caffeine after basal stimulation. Furthermore, the time constant τ (“tau”) of the monoexponential fit of the caffeine-induced transient was calculated to estimate NCX-function. To precisely assess SERCA2a function, tau_SERCA2a was calculated from the tau of the electrically stimulated transients and tau of the caffeine-induced transients of the same cell as ${\mathit{\text{tau}}}_{\mathit{\text{SERCA}}2a}=\frac{1}{\frac{1}{{\mathit{\text{tau}}}_{\mathit{\text{paced}}}}-\frac{1}{{\mathit{\text{tau}}}_{\mathit{\text{caffeine}}}}}$ (26).

Tetracaine experiments

Tetracaine-experiments to measure SR Ca²⁺-leak were performed according to the method of Shannon et al. (27). Na⁺- and Ca²⁺-free bath solution was prepared by replacing Na⁺ in NT with Li⁺. Tetracaine 1 mM was added to this solution to prepare the tetracaine solution. The fractional shift in diastolic fluorescence upon tetracaine (an allosteric blocker of ryanodine receptors) under 0Na⁺/0Ca²⁺ conditions was measured, followed by caffeine application to calculate leak-load-relationship. [Ca²⁺]-values were calculated based on the previously reported diastolic Ca²⁺-concentration of 96 nM for the CaMKIIδ_C-TG mouse (16), assuming a Kd for Fluo-4 of 1100 nM in cardiomyocytes (28) using the equation $[{\mathit{\text{Ca}}}^{2+}]=\frac{\raisebox{1ex}{$F$}\!\left/ \!\raisebox{-1ex}{$F_0$}\right.\ast {\mathit{\text{Kd}}}_{\mathit{\text{Fluo}}}}{\left(\raisebox{1ex}{${\mathit{\text{Kd}}}_{\mathit{\text{Fluo}}}$}\!\left/ \!\raisebox{-1ex}{${\mathit{\text{diast Ca}}}^{2+}$}\right.\right)-(F/F_0)+1}$ (29).

Ca²⁺-spark measurements

Cardiomyocytes were incubated with 10 µM Fluo-3 AM for 15 min and experiments were started after washing out the loading buffer for 5 min with experimental solution. Fluorescence measurements for Ca²⁺-sparks were performed with a laser scanning confocal microscope (LSM 5 Pascal, Zeiss, Germany). Fluo-3 was excited at 488 nm and emitted fluorescence was collected through a 505 nm long-pass emission filter. Fluorescence images were recorded in line-scan mode with 512 pixels per line (width of each scan line 38.4 µm), pixel time 0.64 µs. Ca²⁺-sparks were detected and quantified using the ImageJ (Wayne Rasband, National Institutes of Health, Bethesda, MD) plugin Sparkmaster (30) with visual confirmation of sparks detected. Ca²⁺-spark frequency (CaSpF) was calculated from this and normalized to scanned myocyte width and scanning interval. Ca²⁺-spark size (CaSpS, calculated as: amplitude * width * duration) was added for all sparks within a cell to calculate the SR Ca²⁺-leak for this cell. Only cells displaying Ca²⁺-sparks were included in the statistics. In addition to Ca²⁺-sparks, we also compared the occurrence of spontaneous global intracellular Ca²⁺-release events (“SCaEs”, (8,31)) in these cells in order to evaluate cellular antiarrhythmic effects of AS105. For measurements in wildtype murine cardiomyocytes, recordings were done using a Zeiss LSM 700 (Zeiss, Germany), width of each line 35.5 µm, and sparks evaluated as described above.

Patch clamp measurement of L-type Ca²⁺-currents

Patch-clamp experiments were performed using ventricular cardiomyocytes isolated from CaMKIIδ_C-TG mice using an EPC-10 amplifier and Patchmaster software (HEKA, Germany). The L-type Ca²⁺-current (I_Ca,L) was recorded at room temperature (21 °C) using the ruptured-patch whole-cell-patch clamp technique in voltage-clamp mode. After rupture, 2 minutes was allowed for equilibration of intracellular solution and cytosol before starting recordings. Fast and slow capacitance as well as series resistance were compensated for using the built in functions of Patchmaster. Pipettes were pulled to resistances of 2–3 MΩ and filled with Na⁺- and K⁺-free intracellular solution consisting of (in mM) 90 Cs-methanesulfonate, 20 CsCl, 10 HEPES, 4 MgATP, 0.4 Tris-GTP, 3 CaCl₂, and 10 EGTA (pH 7.2 at 21 °C). Myocytes were superfused with K⁺-free external solution consisting of (in nM) 120 tetraethyl ammonium chloride, 10 CsCl, 10 HEPES, 10 glucose, 1 MgCl₂ and 2 CaCl₂ (pH 7.4 at 21 °C). From a holding potential of −80 mV, cells were briefly depolarized to −40 mV for 50 ms to ensure inactivation of Na⁺-currents, then clamped to test potentials between −30 mV and +40 mV for 200 ms in 10-mV steps, increasing in intervals of 1 s to establish current-voltage (IV) relationship. Measured currents were normalized to membrane capacitance (pA/pF).

Results

Characterization of AS105, an inhibitor of CaMKIIδ

The small molecule inhibitor AS105 (Fig. 1A) was identified during optimization of pyrimidine-based CaMKIIδ inhibitors (21,22). We synthesized AS105 and characterized its inhibitory properties using recombinantly expressed and purified CaMKIIδ, the CaMKII isoform that is predominantly expressed in cardiac tissue (32). In vitro, CaMKIIδ phosphorylation assays using peptide substrates revealed that AS105 is quite potent, with a measured IC50 of 8 nM (Fig. 1B). Computational modeling using crystal structures of CaMKIIδ predicts AS105 docks in the ATP binding pocket, a deep cleft located between the kinase N-terminal and C-terminal lobes. Drug binding would thus be expected to block ATP from binding at this site. To validate this kinetically, we carried out ATP competition experiments by measuring the inhibitory capacity of AS105 at increasing ATP concentrations. As predicted by our computational modeling, these kinetic analyses showed that AS105 is competitive against ATP (Fig. 1C). Given this information, the inhibition constant (Ki) of AS105 for CaMKIIδ was calculated, Ki = 3 nM, suggesting that even at the intracellular millimolar ATP concentrations present in cardiomyocytes, AS105 would be an effective inhibitor of CaMKIIδ.

One of the unique properties of CaMKII family of proteins is their ability to become “autonomous” or Ca²⁺-independent after autophosphorylation on Thr287 (Thr286 in CaMKIIα) (reviewed in (33)). Phosphorylation at this site prevents the enzyme from reverting to its initial autoinhibited state and renders the kinase constitutively active regardless of Ca²⁺/CaM levels in the cell, until dephosphorylated by phosphatases (reviewed in (33)). Oxidation, GlcNAcylation and S-nitrosylation, are additional recently described CaMKIIδ post-translational modifications that functionally mimic the effects of autophosphorylation [34–36].

The widely used CaMKII tool inhibitor KN-93 is sensitive to the autophosphorylation status of CaMKII, inhibiting only the unphosphorylated form of the kinase (37). KN-93 is thought to interfere with the process CaMKII activation and it is classified as a “non ATP-competitive” or “allosteric” inhibitor. To further explore AS105 mechanism of action, we compared the inhibitory properties of AS105 and KN-93 towards unphosphorylated and autophosphorylated CaMKIIδ (Fig. 1D). Prior to the kinase activity assay we subjected CaMKIIδ to an autophosphorylation reaction in the presence of Ca²⁺/CaM and ATP (CaMKII-P) or to a control reaction lacking ATP (CaMKII) for 5 minutes at room temperature. We then obtained inhibitor IC50s by assaying the autophosphorylation or control reactions for activity as normal in the presence of increasing concentrations of AS105 or KN-93. As autophosphorylated CaMKII is Ca²⁺/CaM independent, the Ca²⁺ chelator EGTA was included in the CaMKII-P kinase activity reactions to ensure only autophosphorylated kinase is being followed. For CaMKII-P reactions, inclusion of EGTA in the kinase assays did not alter the IC50s calculated from them (Fig. 1D). Control reactions that lacked ATP in the autophosphorylation reaction and that were subsequently assayed for kinase activity in the presence of EGTA showed no measurable activity, verifying that no autophosphorylation had occurred and that, as previously demonstrated, CaMKII needs Ca²⁺/CaM for activity unless autophosphorylation on Thr287 has occurred (Fig. 1D).

The results shown in Fig. 1D reveal that, in agreement with previous observations, KN-93 is unable to inhibit CaMKIIδ that has been subjected to autophosphorylation prior to kinase activity measurements, i.e. it blocks activation of the kinase but does not block catalytic activity. In contrast, AS105 inhibits unphosphorylated and autophosphorylated kinase equally well (IC50s of 8 and 7 nM respectively); it is not sensitive to the autophosphorylation state of the enzyme. In the absence of autophosphorylation, KN-93 inhibits CaMKIIδ with an IC50 value of 1.4 µM. Taken together, the high potency of AS105 (as demonstrated in vitro), coupled with its potential to inhibit multiple forms of CaMKIIδ, make it a promising lead for optimization for clinical use.

Functional experiments

Effects of AS105 (2 µM) on isolated cardiomyocyte Ca²⁺-handling and SR Ca²⁺-leak were compared against vehicle as described below.

AS105 effects on excitation-contraction coupling in murine cardiomyocytes

Effects of AS105 on excitation-contraction coupling were investigated in ventricular cardiomyocytes from CaMKIIδ_C-overexpressing mice. We found that Ca²⁺-transient amplitudes (mean data and original Ca²⁺-transients in Fig. 2A) and, consequently, systolic cardiomyocyte contractility (mean data and original single cell twitches in Fig. 2B) were significantly improved by AS105 during basal electrical stimulation. (Ca²⁺-transients amplitude: AS105: 3.44±0.06 F/F₀ vs Ctrl: 3.23±0.06 F/F₀, n=94 vs n=92, p<0.05; twitch amplitude: AS105: 4.66±0.19% fractional cell shortening vs Ctrl: 4.04±0.17%, n=93 vs n=92, p<0.05.) Underlying Ca²⁺-transients, SR Ca²⁺-content was also improved upon AS105 treatment, as assessed by the amplitudes of caffeine-induced Ca²⁺-transients (illustrative original registrations and mean values in Fig. 2C, caffeine-induced transient amplitude AS105: 5.49±0.19 F/F₀ vs Ctrl: 4.88±0.14 F/F₀, n=78 vs n=66, p<0.05). We found that fractional release of Ca²⁺ from the SR during basal stimulation (calculated by comparing the amplitude of electrical- to caffeine-induced transients) was not altered by AS105, suggesting that the improvements of Ca²⁺-transient can be attributed to the improved SR Ca²⁺-loading instead of just a larger fraction of stored Ca²⁺ being released from the SR. Ca²⁺-released from the SR during systole is recovered via SERCA2a. SERCA2a-function was not altered by AS105 (Fig. 2E, AS105: tau_SERCA2a 0.330±0.009 s vs Ctrl: 0.339±0.009 s, n=78 vs n=66). (For cells without caffeine-data, SERCA2a-function was assessed by time to 80% decay of electrically evoked transients (rt80%). This also was not influenced by AS105 (data not shown).) The function of NCX was assessed during caffeine-application. AS105 did not influence NCX activity (Fig. 2F, AS105: tau 4.76±0.45 s vs Ctrl: 4.67±0.47 s, n=78 vs. n=66). Thus, we could exclude that the improvement of SR Ca²⁺-loading upon AS105 were due to enhancing SR Ca²⁺-reuptake (i.e. activation of SERCA2a) or reducing cardiomyocyte Ca²⁺-extrusion (i.e. inhibition of NCX).

Figure 2. Effect of AS105 on Excitation Contraction Coupling.

(A&B): Ca²⁺-transients and contractility are improved by AS105 during basal electrical stimulation: Ca²⁺-transients (Fluo-4 F/F₀) and contractions (fractional sarcomere length shortening) were recorded in isolated cardiomyocytes from mice with heart failure due to CaMKIIδ_C-overexpression, as illustrated by original registrations (A). Mean values (B) show improved Ca²⁺-transient amplitudes and contractility upon AS105 during basal 1 Hz stimulation. n= indicates cells/mice. (C&D): AS105 improves SR Ca²⁺-content: SR Ca²⁺-content was measured by measuring caffeine-induced Ca²⁺-release. As shown by original registrations and mean values (C), AS105 improved SR Ca²⁺-content. Caffeine experiments also allowed to investigate systolic fractional SR Ca²⁺-release during basal stimulation, which was not altered by AS105 (D). n= indicates cells/mice. (E&F): AS105 does not impair diastolic SERCA2a or NCX activity: SERCA2a function was calculated from decay constants of electrically stimulated and caffeine-induced Ca²⁺-transients (E). This showed that SERCA2a activity was not altered by AS105. NCX-function was measured during caffeine exposure and was not affected by AS105 (F). n= indicates cells/mice.

AS105 effects on SR Ca²⁺-leak

Thus, we assessed post-rest potentiation of Ca²⁺-transient amplitudes during 10 s of stimulation cessation in the murine cardiomyocytes as an integrative measure of the ability of the SR to accumulate Ca²⁺. We found clearly enhanced post-rest-potentiation in AS105 treated cardiomyocytes, as illustrated in original registrations and mean values in Fig. 3A (AS105: 33.9±1.9% increase of the transient F/F₀ after the pause vs Ctrl: 26.9±1.5%, n=39 vs. n=46, p<0.05), indicating an improved ability to accumulate Ca²⁺ in the SR, which explains the increased SR Ca²⁺-content under basal stimulation.

Figure 3. AS105 effects on SR Ca²⁺-leak.

(A): AS105 improves SR Ca²⁺-accumulation: Post-rest increase of Ca²⁺-transient amplitude was investigated to assess the ability of the SR to accumulate Ca²⁺, which was enhanced by AS105. n= indicates cells/mice. (B–D): AS105 reduces SR Ca²⁺-leak in ventricular murine cardiomyocytes: As illustrated in original recordings (B), tetracaine-sensitive shift of [Ca²⁺]_i was measured to assess SR Ca²⁺-leak in ventricular cardiomyocytes from mice with heart failure due to CaMKIIδ_C-overexpression. AS105 reduced SR Ca²⁺-leak (C) by strongly attenuating SR leak/load-relationship (D). n= indicates cells/mice. (E–H): AS105 reduces SR Ca²⁺-leak in human atrial cardiomyocytes: Effects of AS105 on SR Ca²⁺-leak in human atrial cardiomyocytes were assessed by measuring Ca²⁺-sparks (E). (Original line scan recordings 20% contrast enhanced for better reproduction.) AS105 reduced both the size of individual Ca²⁺-sparks (F) as well as the frequency of their occurrence (G), resulting in a pronounced reduction of SR Ca²⁺-leak (H). n= indicates sparks/patients (F) or cells/patients (G&H). (I): AS105 suppresses arrhythmogenic cellular events: The occurrence of arrhythmogenic spontaneous global Ca²⁺-release events (original recording (I)) in human atrial cardiomyocytes was assessed during confocal line scans. AS105 suppressed the occurrence of these events by ~50%. n= indicates cells/patients.

We assessed AS105 effects on SR Ca²⁺-leak first using tetracaine experiments (27) in ventricular cardiomyocytes from CaMKIIδ_C-overexpressing mice. As illustrated in Fig. 3B, the tetracaine-induced diastolic [Ca²⁺]-shift was clearly lower in AS105-treated cells (Fig. 3C), indicating lower SR Ca²⁺-leak. ([Ca²⁺]-shift upon AS105: 16.46±1.58 nM vs. 23.25±2.20 nM, n=28 vs. n=24, p<0.05. For raw fluorescence signal AS105: 16.05±1.57% reduction of fluorescence vs. Ctrl: 22.81±2.21%, n=28 vs. n=24, p<0.05.) Importantly, AS105 effected a smaller SR Ca²⁺-leak despite the higher SR Ca²⁺-loading in the AS105-treated cells, as the leak/load-relationship was reduced by AS105 by ~65% (Fig. 3D, AS105: 1.59±0.42% vs. Ctrl: 4.55±1.18%, n=14 vs. n=15, p<0.05; for raw fluorescence AS105: 2.63±0.44% vs. Ctrl: 6.05±1.21%, n=14 vs. n=15, p<0.05).

To validate our key finding that AS105 suppresses SR Ca²⁺-leak for human cells and to do so using another method for assessing Ca²⁺-leak, we measured Ca²⁺-sparks in right atrial cardiomyocytes from human donors. As shown in Fig. 3E, we found that AS105 reduced both the size of individual Ca²⁺-sparks (Fig. 3F, AS105: 329.6±26.4 nF/F₀*m*s vs Ctrl: 454.0±34.5 nF/F₀*m*s, n=293 vs n=359 sparks, p<0.05) as well as the frequency of Ca²⁺-spark occurrence (Fig. 3G, AS105: 1.89±0.17 sparks/100µm/s vs Ctrl: 2.44±0.21 sparks/100µm/s, n=59 vs n=62 cells, p<0.05). This results in a ~44% reduction of total SR Ca²⁺-leak per cell in the human atrial cardiomyocytes (Fig. 3H, AS105: 6.24±1.18 mF/F₀ vs. Ctrl: 11.07±2.08 mF/F₀, n=59 vs. n=62, p<0.05).

Importantly, associated with its effects on SR Ca²⁺-leak, AS105 also effectively suppressed arrhythmogenic spontaneous global Ca²⁺-release events (“SCaEs”) in these human cardiomyocytes (shown in Fig. 3I), which were reduced by 51% (AS105: 0.082±0.011 events/cell/s vs Ctrl: 0.168±0.019 events/cell/s, n=255 vs n=262 cells, p<0.05). The reduced SCAEs resulted from fewer cardiomyocytes with any events and a reduced frequency of events in arrhythmic cells. Only 24.3% of cardiomyocytes showed any events with AS105 as compared to 35.5% of cells under control conditions (p<0.05), while AS105 reduced the frequency of events in these arrhythmogenic cells to 0.34±0.03 events/(arrhythmic cell)/s as compared to 0.47±0.04 in Ctrl (p<0.05). AS105 thus appears to both reduce the size of potential arrhythmia triggers (the number of cells capable to trigger arrhythmias) as well as the frequency at which these triggers arise.

AS105 effects on the L-type Ca²⁺-current

We tested whether the effects of AS105 on Ca²⁺-transient amplitude could be due to enhancing the L-type Ca²⁺-current (thus increasing the Ca²⁺-influx into the myocyte) by performing patch clamp experiments in the CaMKIIδ_C-TG cardiomyocytes. As shown in Fig. 4, we found that AS105 did not significantly affect the L-type Ca²⁺-current (peak current at 0 mV: AS105: −7.55±0.56 pA/pF vs Ctrl: −8.64±0.69 pA/pF, n=17 vs n=17 cells).

Figure 4. AS105 effects on L-type Ca²⁺-current.

(A&B): AS105 does not affect L-type Ca²⁺-current: IV-relationship of the L-type Ca²⁺-current in CaMKIIδ_C-TG cardiomyocytes was assessed by patch clamp recordings in voltage-clamp mode and was not affected by AS105 (A), as illustrated in original recordings (B). n= indicates cells/mice.

AS105 effects in wildtype cardiomyocytes

To assess whether the positive inotropic effects of AS105 in the CaMKIIδ_C-TG cells can indeed be attributed to the reduction of the pathologically high SR Ca²⁺-leak in these cells, we investigated the effects of the compound on SR Ca²⁺-leak and excitation-contraction coupling in cardiomyocytes from the corresponding wildtype mice, which also show some basal Ca²⁺-leak (16).

As shown in Suppl. Fig. 1A&B, AS105 could, to a slight but significant degree, further suppress the already low SR Ca²⁺-leak also in these healthy cells (AS105: 4.22±0.81 mF/F₀ vs Ctrl: 2.15±0.34 mF/F₀, n=50 vs n=53 cells, p<0.05). This minor reduction of the normal low SR Ca²⁺-leak, however, could not further increase SR Ca²⁺-content in these cells (Suppl. Fig. 1C: 4.73±0.16 F/F₀ vs Ctrl: 5.01±0.33 F/F₀, n=13 vs n=15). As fractional systolic SR Ca²⁺-release was not affected by AS105 (Suppl. Fig. 1D: AS105: 69.22±3.87% vs Ctrl: 73.58±2.39%, n=13 vs n=15), AS105 in these healthy cardiomyocytes (in contrast to the heart failing CaMKIIδ_C-TG cells) did not increase the amplitude of systolic Ca²⁺-transients (Suppl. Fig. 1E: AS105: 3.26±0.09 F/F₀ vs Ctrl: 3.43±0.11F/F₀, n=68 vs n=69 cells) or the resulting cellular contractility (Suppl. Fig. 1F: 3.88±0.28% systolic fractional cell shortening vs Ctrl: 4.22±0.28%, n=68 vs n=69).

While F/F₀ cannot be compared between WT and CaMKIIδ_C-TG, since diastolic Ca²⁺-levels are different between these genotypes, Ca²⁺-transient amplitudes have been calculated based on these diastolic Ca²⁺-levels (16). Performing this analysis for our current data confirms normal systolic Ca²⁺-transient amplitudes in our WT, but severely diminished ones in the CaMKIIδ_C-TG cardiomyocytes (TG: 257.6±6.0 nmol/L vs WT: 481.0±9.1 nmol/L, n= 69 vs n=92, p<0.05) due to much lesser SR Ca²⁺-content in the CaMKIIδ_C-TG cells (caffeine-induced transients TG: 450.4±18.8 nmol/L vs. WT: 2119.9±574.2 nmol/L n=15 vs n=66). Thus, the ability of AS105 to improve SR Ca²⁺-content and systolic Ca²⁺-release could depend on the presence of pathologically high SR Ca²⁺-leak with consecutively diminished SR Ca²⁺-loading, while it is not effective in this regard under physiologic conditions.

Discussion

Inhibition of CaMKII has been suggested as a novel therapeutic approach in heart failure and cardiac arrhythmias. Increased CaMKII activity in these conditions has been linked to disturbances in excitation-contraction coupling, especially increased Ca²⁺-leak from the SR. While currently available CaMKII-inhibitory compounds are not suitable for clinical use, new small molecule CaMKII-inhibitors are under development that could fulfill that promise (21). We now report on AS105, a novel small molecule designed to be an ATP-competitive CaMKII-inhibitor that is one of a series of new inhibitors being developed. Our investigations confirmed that indeed, the CaMKII-inhibition by AS105 occurs via ATP-competition. We could also demonstrate that in this manner, unlike the often-used KN-93, AS105 is equally effective against phosphorylated and unphosphorylated CaMKII. This is of relevance in so far, as recent research suggests important roles for other modes of CaMKII activation besides the canonical Ca²⁺/CaM dependent pathway, i.e. by oxidation (34) (at Met-281/282) and by nitrosylation (36) (presumably at Cys-290). Our investigations showed AS105 to have high potency with an in vitro IC50 in the low nanomolar range.

We further investigated the effects of AS105 on isolated cardiomyocytes to delineate some of the roles of CaMKII on dysregulation of Ca²⁺-homeostasis. The first focus of our investigations was on AS105 effects (2 µM) on SR Ca²⁺-leak, which we investigated in both ventricular cardiomyocytes from CaMKIIδ_C-overexpressing mice with heart failure as well as in human atrial cardiomyocytes. Both methodologies employed by us to measure SR Ca²⁺-leak, tetracaine-sensitive Ca²⁺-leak and Ca²⁺-spark evaluation, indicate that AS105 effectively reduced SR Ca²⁺-leak – by 44 % (in the atrial cardiomyocytes) to 65 % (in the failing ventricular cells). Mechanistically, we could show that this was directly due to reduction of RyR2 “leakiness”, as we found leak/load relationship to be improved upon AS105. Diastolic Ca²⁺-leak from the SR is regarded as an arrhythmogenic mechanism both in ventricular tachyarrhythmias (9–11) as well as in atrial fibrillation (6–8) by inducing spontaneous electrical activity (i.e. providing an arrhythmogenic trigger due to delayed afterdepolarizations (8)). We could show that for AS105, the reduction of diastolic Ca²⁺-leak was indeed associated with a reduction of cellular arrhythmogenic correlates, indicating that AS105 might have antiarrhythmic effects. Ca²⁺-leak from the SR has been shown to deplete SR Ca²⁺-content (3–5), leading to lower amounts of Ca²⁺ being available for systolic release and impaired contractility. Our results indicate that, indeed, the reduction of SR Ca²⁺-leak by AS105 resulted in an enhanced capability to accumulate Ca²⁺ in the SR: In the failing ventricular cardiomyocytes, AS105 significantly improved post-rest-potentiation of Ca²⁺-transient amplitudes, a measure for the ability of the SR to store and release Ca²⁺ that been found to be severely impaired in heart failure (38). Accordingly, AS105 improved SR Ca²⁺-loading in these cells and thus also enhanced systolic Ca²⁺-release and systolic contractility of these cardiomyocytes upon basal electrical stimulation. In cardiomyocytes from healthy WT mice with low SR Ca²⁺-leak, however, AS105 did not affect systolic Ca²⁺-transients or cellular contractility. (This is in line with previous findings using another CaMKII-inhibitor also in healthy cardiomyocytes, where a reduction of SR Ca²⁺-leak did not increase SR Ca²⁺-content (39). In addition, we could show that the even lower SR Ca²⁺-leak in CaMKIIδ-knockout mice also did not affect SR Ca²⁺-load (17). It is possible that a certain threshold exists below which SR Ca²⁺-leak does not affect SR Ca²⁺-content e.g. due compensation by SR Ca²⁺-reuptake via SERCA2a. As such, the positive inotropic effects of AS105 appear to depend on it being able to lower pathologic SR Ca²⁺-leak (such as in failing hearts or in atrial fibrillation), while the compound has no direct positive inotropic effect. Thus, AS105 could also have therapeutic potential in the context of heart failure. Noteworthy, while the reduction of diastolic Ca²⁺-leak suggests a decreased Ca²⁺-sensitivity of the RyR2, this seems to be overcome by the increased Ca²⁺-content such that on balance, systolic fractional Ca²⁺-release is not different. From our measurements of L-type Ca²⁺-currents, we can also exclude potential effects of AS105 on this current, which might have affected SR Ca²⁺-loading or systolic Ca²⁺-transient amplitudes. Of note, CaMKII inhibition by AS105 had no discernible negative effects on cardiomyocyte excitation-contraction coupling in our experiments. Importantly, while SR Ca²⁺-reuptake could theoretically be compromised by CaMKII-inhibitors due to interference with regulation SERCA2a by PLB, AS105 in the concentration investigated by us had no effect on diastolic Ca²⁺-transient decay and calculated SERCA2a function. This is in line with previous findings by us showing that even upon 87% reduction of total CaMKII activity due to ablation of CaMKIIδ, frequency dependent acceleration of relaxation, an important physiologic adaptive mechanism, is not impaired (17). However, other results from this mouse model of reduced CaMKII-activity suggest that under specific pathologic conditions such as acidosis, reduction of CaMKII might have effects – a caveat towards CaMKII-inhibitors in general that will have to be closely monitored during further investigations leading to clinical use of such drugs.

Highlights.

• AS105 is a novel, pyrimidine-based ATP-competitive CaMKII-inhibitor.

• It is in this manner also effective against autophosphorylated CaMKII (as opposed to KN-93).

• AS105 does not negatively affect basal excitation-contraction-coupling in cardiomyocytes.

• In cardiomyocytes from CaMKIIδ_C-overexpressing mice with heart failure, AS105 reduces SR Ca²⁺-leak, thus improving SR Ca²⁺-accumulation, leading to improved systolic function.

• AS105 also reduces SR Ca²⁺-leak in human atrial cardiomyocytes, suppressing arrhythmogenic single cell Ca²⁺-events.