Decreasing Microtubule Detyrosination Improves Cardiac Mechanics and Sodium Channel Function in Arrhythmogenic Cardiomyopathy

Abstract

BACKGROUND:

Alterations in microtubule dynamics have been shown to affect cardiomyocyte membrane stiffness and modulate ion channels, including the cardiac sodium channel. While conditions, such as heart failure and Duchenne muscular dystrophy, are associated with increased detyrosination of microtubules and reduced sodium current, a potential role for microtubule detyrosination in arrhythmogenic cardiomyopathy has not been explored. We here investigated the impact of microtubule detyrosination on membrane stiffness, cardiac sodium channel distribution, and function in mouse and human models of arrhythmogenic cardiomyopathy.

METHODS:

Isolated ventricular cardiomyocytes from mice with cardiomyocyte-specific, tamoxifen-activated knockout of PKP2 (plakophilin-2), as well as PKP2-c.2013delC, and isogenic control human-induced pluripotent stem cell–derived-cardiomyocytes were incubated for 2 to 4 hours with compounds known to decrease microtubule detyrosination (parthenolide, 10 µmol/L; EpoY, 20 µmol/L) or vehicle (dimethyl sulfoxide). Immunocytochemistry, mechano-scanning ion conductance microscopy, patch-clamp analysis, and stochastic optical reconstruction microscopy were performed.

RESULTS:

Cardiomyocyte-specific, tamoxifen-activated knockout of PKP2 mouse cardiomyocytes displayed increased microtubule detyrosination and membrane stiffness, which were both attenuated by parthenolide treatment. Parthenolide significantly increased whole-cell sodium current density in cardiomyocyte-specific, tamoxifen-activated knockout of PKP2 mouse cardiomyocytes, with macropatch measurements demonstrating that this increase occurred both at the intercalated disc and lateral membrane. Stochastic optical reconstruction microscopy analysis revealed that parthenolide increased cardiac sodium channel cluster density at the intercalated disc of cardiomyocyte-specific, tamoxifen-activated knockout of PKP2 mouse cardiomyocytes. In contrast, parthenolide had no effect on sodium current density, cardiac sodium channel cluster size, or density in cardiomyocytes from control mice. PKP2-c.2013delC human-induced pluripotent stem cell–derived-cardiomyocytes displayed increased microtubule detyrosination and reduced sodium current density compared with isogenic control human-induced pluripotent stem cell–derived-cardiomyocytes, which were both prevented by parthenolide and EpoY.

CONCLUSIONS:

Increased microtubule detyrosination secondary to loss of PKP2 impacts cardiomyocyte (dys)function beyond the desmosome, contributing to both electrical and mechanical alterations in the setting of arrhythmogenic cardiomyopathy. Our findings identify microtubule detyrosination as a novel therapeutic target in pathophysiological conditions, such as arrhythmogenic cardiomyopathy, aimed at improving both contractile and electrical function.

WHAT IS KNOWN?

• Sodium current density is decreased in arrhythmogenic cardiomyopathy.

• Parthenolide and EpoY decrease microtubule detyrosination.

WHAT THE STUDY ADDS

• Microtubule detyrosination and membrane stiffness are increased in the setting of arrhythmogenic cardiomyopathy.

• Decreasing microtubule detyrosination by parthenolide or EpoY increases sodium current density and decreases membrane stiffness in arrhythmogenic cardiomyopathy.

• Microtubule detyrosination impacts cardiomyocyte (dys)function beyond the desmosome, contributing to both electrical and mechanical alterations in the setting of arrhythmogenic cardiomyopathy.

Arrhythmogenic right ventricular cardiomyopathy, also known as arrhythmogenic cardiomyopathy (ACM), is a genetic condition associated with an increased risk of ventricular arrhythmias and propensity to sudden cardiac death, especially in young individuals and athletes.¹ In more than half of all patients with ACM, mutations in desmosomal genes are found, including PKP2 (plakophilin-2), desmoglein-2, desmocollin-2, plakoglobin, and DSP (desmoplakin).² Desmosomes mediate cell-to-cell adhesion and are localized at the intercalated disc (ID) of cardiomyocytes. IDs are essential for the propagation of the electrical signal between cardiomyocytes; accordingly, mutations affecting the integrity of these structures have been shown to impair cardiac conduction.² PKP2 mutations represent the most common genetic alterations in patients with gene-positive ACM,³ and PKP2 deficiency has been shown to cause reduced sodium current (I_Na),^4,5 intracellular calcium dysregulation,⁶ and impaired cardiac conduction.⁷ Moreover, PKP2 interacts with proteins of the actin-myosin complex, with mutations in PKP2 inducing alterations in actomyosin organization and function, resulting in contractile dysfunction.^8,9

In pathological conditions such as heart failure (HF), hypertrophic cardiomyopathy, and Duchenne muscular dystrophy (DMD), increased membrane stiffness secondary to alterations in microtubules (more specifically, increased microtubule detyrosination) has been shown to impair cardiomyocyte contraction and relaxation.^10–16 Moreover, we have recently demonstrated that increased microtubule detyrosination also contributes to the reduced peak I_Na observed in cardiomyocytes of DMD mice.¹⁷ Hence, therapeutic interventions aimed at attenuating microtubule detyrosination may have beneficial effects on both mechanical and electrical properties of cardiomyocytes during pathophysiological conditions. To assess whether microtubule remodeling is also relevant in the setting of ACM, here, we used a well-established mouse model of ACM (cardiomyocyte-specific, tamoxifen-activated knockout of PKP2; PKP2cKO)¹⁸ and demonstrate the presence of increased microtubule detyrosination and membrane stiffness in PKP2cKO cardiomyocytes. Importantly, reduction of microtubule detyrosination by parthenolide not only decreased membrane stiffness in PKP2cKO cardiomyocytes but also significantly increased I_Na at both the ID region and lateral membrane (LM), indicating a modulatory impact of microtubule dynamics on subcellular trafficking of the cardiac sodium channel (Na_v1.5). Hence, microtubule detyrosination secondary to loss of PKP2 impacts cardiomyocyte (dys)function beyond the desmosome, contributing to both electrical and mechanical alterations in the setting of ACM. Moreover, we confirmed the presence of increased microtubule detyrosination and the beneficial effects of parthenolide and EpoY, a more specific inhibitor of microtubule detyrosination,¹⁹ in a human pluripotent stem cell–derived ACM disease model. Our findings, therefore, identify microtubule detyrosination as a novel therapeutic target in ACM, aimed at improving both contractile and electrical function.

Methods

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Mouse Model

Experiments were performed using C57BL/6N male mice (aged 3–6 months) expressing a cardiac-specific, tamoxifen-activated deletion of PKP2 (PKP2cKO).¹⁸ Cre-negative, flox-positive, tamoxifen-injected age and sex-matched mice were used as controls. Procedures conformed with the Guide for Care and Use of Laboratory Animals of the National Institutes of Health and were approved by the New York University Institutional Animal Care and Use Committee (Parent Protocol ID: IA16-01021). Mice were euthanized, and experiments were performed 21 days after starting tamoxifen injection, a point at which an ACM phenotype becomes manifest.¹⁸

Cardiomyocyte Isolation and Drug Incubation

Isolated murine ventricular myocytes were obtained by enzymatic dissociation. For details, see the Supplemental Methods. Isolated cardiomyocytes were maintained in a Tyrode solution containing (in mM) 148 NaCl, 5.4 KCl, 1.0 MgCl₂, 1.0 CaCl₂, 0.4 NaH₂PO₄, 15 HEPES, and 5.5 glucose (pH 7.4 with NaOH) and subsequently incubated with either 10-µmol/L parthenolide (Sigma-Aldrich; dissolved in dimethyl sulfoxide [DMSO]) or DMSO (vehicle). Mouse cardiomyocytes were used for immunocytochemistry, mechano-scanning ion conductance microscopy (SICM), patch-clamp analysis, and multicolor super-resolution imaging by 2-dimensional stochastic optical reconstruction microscopy experiments between 2 and 4 hours after the start of incubation.

Human-Induced Pluripotent Stem Cell Culture, Differentiation, and Drug Incubation

Human-induced pluripotent stem cells were cultured and differentiated into cardiomyocytes (human-induced pluripotent stem cell–derived-cardiomyocyte [hiPSC-CMs]) as previously described.²⁰ We used a patient human-induced pluripotent stem cell line carrying the heterozygous mutation PKP2-c.2013delC and its isogenic PKP2-corrected control (LUMCi027-A line and LUMCi027-A-1 line²¹). Further details on hiPSC-CMs dissociation and seeding are provided in the Supplemental Methods. HiPSC-CMs were incubated with either 10-µmol/L parthenolide (Sigma-Aldrich; dissolved in DMSO), 20-µmol/L EpoY (Sigma-Aldrich; dissolved in DMSO) or DMSO (vehicle) and used for immunofluorescence assay and patch-clamp experiments between 2 and 4 hours after the start of incubation.

Confocal Microscopy

For immunocytochemistry, isolated mouse cardiomyocytes were incubated for 2 to 4 hours with either DMSO or parthenolide (10 µmol/L), plated on laminin-coated coverslips, and left to adhere for at least 30 minutes before fixation with 100% methanol.

HiPSC-CMs were cultured for 5 days to allow their recovery after dissociation and then incubated for 4 hours with either parthenolide (10 µmol/L), EpoY (20 µmol/L), or DMSO (vehicle). HiPSC-CMs were then fixed with ice-cold 100% methanol. Further details on protocols and antibodies used for immunocytochemistry are provided in the Supplemental Methods.

Images of mouse cardiomyocytes were acquired using a confocal microscope (Leica SP5) with an HCX PL APO CS 63× (NA 1.40) oil objective and spectral photo multiplier tube detector and with 488- and 633-nm lasers. For hiPSC-CMs, images were acquired with a Leica SP8 WLL confocal laser-scanning microscope using HC PL APO CS2 63x/1.40 oil objective and Z stack acquisition. Further details on imaging and quantification of α-tubulin and detyrosinated α-tubulin in mouse cardiomyocytes and hiPSC-CMs are provided in the Supplemental Methods.

Transverse Young’s Modulus Measurements by Mechano-SICM

Mechano-SICM, a modification of SICM in hopping mode,²² was used to simultaneously generate topographical (10 µm×10 µm) and Young’s modulus (YM) maps, as previously described.²³ YM is a measure of the resistance of material to deformation and is used to assess cell membrane stiffness.²³ Aerostatic constant pressure (15 kPa) was delivered to the nanopipette (200–300-nm tip diameter), which propelled the inner solution, forming a constant soft hydrojet on the tip of the pipette that can displace cellular membranes throughout the scanning procedure. In brief, the position of the z axis piezo-actuator at the moment the current was reduced by 0.7% (named topographical set-point or SP_T) was marked as the height of the sample for that point; consecutive points were then recoded to represent the topographical surface. The analysis included a topographical slope correction for the uneven membrane surface areas as described before.²⁴ Scanning solution containing (in mM) NaCl 144, HEPES 10, MgCl₂ 1, KCl 5, and pH 7.4 (NaOH) was used both as the pipette and bath solution. Images were analyzed using the SICM image viewer. Further details on YM calculation are provided in the Supplemental Methods.

Sodium Current Measurements

Data Acquisition and Analysis

Whole-cell I_Na was measured in mouse cardiomyocytes and hiPSC-CMs with the ruptured patch-clamp technique, while local I_Na generated specifically at distinct subcellular regions of mouse cardiomyocytes was measured with the macropatch cell-attached configuration. For details, see the Supplemental Methods.

Whole-Cell I_Na Measurements

I_Na in mouse cardiomyocytes and hiPSC-CMs were measured using internal and external solutions described in the Supplemental Methods. Series resistance and cell membrane capacitance were compensated for ≥80%.

I_Na was filtered at 10 kHz and digitized at 20 kHz in mouse cardiomyocytes and filtered at 5 kHz and digitized at 40 kHz in hiPSC-CMs. Current-voltage relationships and voltage dependence of activation and inactivation ([in]activation) were determined using voltage-clamp protocols described in the relevant figures. In mouse cardiomyocyte experiments, a holding potential of −140 mV and a cycle length of 3 seconds were used, while, for I_Na measurements in hiPSC-CMs, voltage protocols had a holding potential of −120 mV and a cycle length of 5 seconds. I_Na was defined as the difference between the peak current and the current at the end of the voltage step. Current densities were calculated by dividing the current amplitude by the cell membrane capacitance. Voltage dependence of activation and inactivation curves were fitted with the Boltzmann function (y=[1+exp{(V−V_1/2)/k}]⁻1), where V_1/2 is the half-maximal voltage of (in)activation and k is the slope factor.

Macropatch I_Na Measurements

In the macropatch cell-attached configuration, I_Na recordings were conducted as described before.²⁵ The internal and external solutions used for macropatch experiments are described in the Supplemental Methods.

I_Na was filtered at 10 kHz and digitized at 20 kHz. Recordings were obtained either from the end of the cell, in the area previously occupied by the ID, or from the LM section of the cell. The voltage-clamp protocol for the determination of the current-voltage relations at the LM and ID is depicted in the insets of the respective figures. Because the recording electrode was in the extracellular space, an inverted voltage protocol was used to activate I_Na, with a cycle length of 3 seconds. Following standard reporting methods, intracellular membrane potentials and I_Na measured by macropatch were presented as an inward-directed current.

Multicolor Super-Resolution Imaging by Two-Dimensional Stochastic Optical Reconstruction Microscopy

Isolated mouse cardiomyocytes were plated on laminin-coated coverslips and left to adhere for at least 30 minutes before fixation with 100% methanol. Following immunocytochemical staining (for details, see the Supplemental Methods), coverslips were mounted on microscope glass slides with predrilled holes for imaging buffer exchange. Freshly prepared super-resolution imaging buffer consisting of 1-mg/mL glucose oxidase (Sigma, G2133), 0.02-mg/mL catalase (Sigma, C3155), 10% glucose (Sigma, G8270), and 100-mmol/L mercaptoethylamine (Fisher Scientific, 100995) in PBS (pH 8.0) was injected immediately before imaging.

Imaging

Samples were imaged using a custom-built platform based on an inverted microscopy setup (ASI-RAMM) similar to the scheme described before.²⁶ For further details on imaging, see the Supplemental Methods.

Cluster Analysis

As previously described,²⁷ reconstructed super-resolved images were processed with a smoothing filter (gaussian blur function in ImageJ, version 1.52j), adjusted for brightness and contrast, and filtered to a threshold to obtain a binary image. For the experimental data sets, regions of interest were manually drawn for each image and saved as an ImageJ region of interest file. LM clusters were defined as the clusters localized along the lateral edges of the cardiomyocytes, and ID clusters were defined as clusters localized at the cell ends (detected by N-cadherin staining). Cluster detection and parameters were obtained using the ImageJ function Analyze particles as previously described.²⁸ Because the average background cluster size in the negative controls was <7000 nm,² the minimum size/area of contiguous pixels defined as a real cluster was 7000 nm.² Signal-positive areas of smaller dimensions were considered as background and not included. GraphPad Prism 7 was used for data visualization.

Statistical Analysis

Data were analyzed using Sigma Stat 3.5 (Systat Software, Inc) and GraphPad Prism, version 8.4.3(686), for Windows (GraphPad Software). Values are shown as mean±SEM. Applied statistical tests are specified in the corresponding figure and table legends. For details on statistical analysis, see the Supplemental Methods. The level of statistical significance was set to P<0.05.

Results

Microtubule Detyrosination Is Increased in PKP2cKO Cardiomyocytes and Reduced by Parthenolide

We first explored the level of total α-tubulin and detyrosinated α-tubulin in ventricular PKP2cKO compared with control cardiomyocytes through immunocytochemistry (Figure 1A and 1B). These immunofluorescence experiments revealed that, while total α-tubulin was not affected in PKP2cKO cardiomyocytes (Figure 1C), there was a significant increase in the detyrosinated α-tubulin fraction in PKP2cKO cardiomyocytes compared with control cardiomyocytes (Figure 1D); accordingly, the ratio of detyrosinated to total α-tubulin was significantly increased in PKP2cKO cardiomyocytes (Figure 1E). In control cardiomyocytes, incubation with parthenolide had no effect on either detyrosinated or total α-tubulin levels (Figure 1C through 1E). In PKP2cKO cardiomyocytes, parthenolide did not affect total α-tubulin (Figure 1C) but significantly decreased detyrosinated α-tubulin levels (Figure 1D) and tended to decrease the detyrosinated versus total α-tubulin ratio (Figure 1E). Hence, PKP2cKO cardiomyocytes display increased levels of detyrosinated microtubules, which are attenuated by parthenolide.

Figure 1.

Microtubule (MT) detyrosination is increased in cardiomyocyte-specific, tamoxifen-activated knockout of plakophilin-2 (PKP2cKO) compared with control (CTR) cardiomyocytes (CMs) and is attenuated by parthenolide (PTL). Representative examples of confocal images showing immunolabeled total α-tubulin (red) and detyrosinated α-tubulin (green) in CTR (A) and PKP2cKO (B) CMs after incubation with either dimethyl sulfoxide (DMSO) or PTL (scale bar, 20 µm). Quantification of average total α-tubulin (C), detyrosinated α-tubulin (D), and ratio of detyrosinated α-tubulin on total α-tubulin (E) in CTR and PKP2cKO CMs after 2 to 4 hours of incubation with either DMSO or PTL (10 µmol/L). n represents the number of cells measured. CTR: N=3 mice; PKP2cKO: N=3 mice. *P<0.05 and ****P<0.0001; Kruskal-Wallis (C) and Kruskal-Wallis followed by the Dunn test for pairwise multiple comparison (D and E). Comparisons for all graphs: CTR DMSO vs PKP2cKO DMSO, CTR DMSO vs CTR PTL, and PKP2cKO DMSO vs PKP2cKO PTL.

PKP2cKO Cardiomyocytes Display Increased Membrane Stiffness That Is Attenuated by Parthenolide

A modulatory impact of microtubule dynamics and parthenolide on membrane stiffness has previously been demonstrated in the setting of HF, hypertrophic cardiomyopathy, DMD, and myocardial infarction.^10–16 We, therefore, here investigated whether membrane stiffness was altered in PKP2cKO cardiomyocytes secondary to increased microtubule detyrosination. Cell membrane stiffness was assessed by measuring transverse YM, a measure of the resistance of material to deformation,²³ using mechano-SICM (Figure 2A). At baseline, transverse YM was significantly increased in PKP2cKO cardiomyocytes compared with control cardiomyocytes (Figure 2B). Parthenolide treatment significantly decreased YM in PKP2cKO cardiomyocytes, while it had no effect on control cardiomyocytes. Hence, PKP2cKO cardiomyocytes display increased membrane stiffness, associated (at least in part) with enhanced microtubule detyrosination.

Figure 2.

Membrane stiffness is increased in cardiomyocyte-specific, tamoxifen-activated knockout of plakophilin-2 (PKP2cKO) compared with control (CTR) cardiomyocytes (CMs) and is attenuated by parthenolide (PTL). A, Representative SICM images of topography (top) and transverse Young’s modulus (YM) maps (lower) for CTR and PKP2cKO CMs after 2 to 4 hours of incubation with either dimethyl sulfoxide (DMSO) or PTL (10 µmol/L). B, Quantification of transverse YM. n represents the number of cells measured. CTR: N=6 mice; PKP2cKO: N=6 mice. **P<0.01 and ****P<0.0001; Kruskal-Wallis followed by the Dunn test for post hoc analysis. Comparisons: CTR DMSO vs PKP2cKO DMSO, CTR DMSO vs CTR PTL, and PKP2cKO DMSO vs PKP2cKO PTL.

Decreasing Microtubule Detyrosination by Parthenolide Increases I_Na at Both ID and LM of PKP2cKO Cardiomyocytes

Given our previously demonstrated link between microtubule detyrosination, parthenolide, and I_Na in a mouse of DMD,¹⁷ we next investigated the effect of parthenolide on I_Na density in control and PKP2cKO cardiomyocytes (Figure 3A through 3D). While parthenolide had no effect on I_Na in control cardiomyocytes, it significantly increased I_Na density in PKP2cKO cardiomyocytes (78% increase; Figure 3A through 3D; Tables S1 and S2). Parthenolide did not affect I_Na voltage dependence of activation and inactivation, assessed as the half-voltage of (in)activation (V_1/2) and the slope factor k in either control or PKP2cKO cardiomyocytes (Figure 3E and 3F; Tables S1 and S2). Because previous studies have demonstrated I_Na remodeling preferentially at the ID of mice with PKP2 haploinsufficiency,^4,29 we next performed macropatch measurements to assess the effect of parthenolide on I_Na specifically at the ID and LM region (Figure 4A and 4C). Interestingly, parthenolide was found to increase I_Na both at the ID and LM in PKP2cKO cardiomyocytes while not affecting I_Na in control cardiomyocytes (Figure 4B and 4D; Tables S3 and S4), indicating that microtubule detyrosination in the setting of loss of PKP2 affects I_Na throughout the entire cell and not just in the region close to the desmosome.

Figure 3.

Parthenolide (PTL) increases sodium current (I_Na) density in cardiomyocyte-specific, tamoxifen-activated knockout of plakophilin-2 (PKP2cKO) cardiomyocytes (CMs). Representative I_Na traces measured in control (CTR; A) and PKP2cKO (B) mouse CMs treated with either dimethyl sulfoxide (DMSO) or PTL. C through F, Average current-voltage (I-V) relationships (C and D) and average voltage dependence of activation and inactivation (E and F) in CTR and PKP2cKO CMs after 2 to 4 hours of incubation with either DMSO or PTL (10 µmol/L). Insets: voltage-clamp protocols. Dotted lines indicate the Boltzmann fit to the average data. n represents the number of cells measured. CTR: N=3 mice; PKP2cKO: N=4 mice. *P<0.05, 2-way ANOVA repeated measures followed by the Holm-Sidak test for post hoc analysis.

Figure 4.

Parthenolide (PTL) increases sodium current (I_Na) at both intercalated disc (ID) and lateral membrane (LM) regions of cardiomyocyte-specific, tamoxifen-activated knockout of plakophilin-2 (PKP2cKO) cardiomyocytes (CMs). Schematic representation of pipette placement for I_Na measurements performed at the ID (A) or LM (C) regions. Average current-voltage (I-V) relationships for I_Na generated at the ID (B) and LM (D) of control (CTR) and PKP2cKO CMs after 2 to 4 hours of treatment with either dimethyl sulfoxide (DMSO) or PTL (10 µmol/L). Insets: voltage-clamp protocols. n represents the number of cells measured. CTR: N=4 mice; PKP2cKO: N=4 mice. *P<0.05, 2-way repeated measures ANOVA followed by the Holm-Sidak test for post hoc analysis.

Parthenolide Increases Na_v1.5 Cluster Density at the ID of PKP2cKO Cardiomyocytes

Because parthenolide did not affect Na_v1.5 voltage dependence of activation or inactivation in PKP2cKO cardiomyocytes, we hypothesized that decreasing microtubule detyrosination enhances I_Na by increasing the abundance of Na_v1.5 protein at the sarcolemma. We, therefore, used multicolor super-resolution imaging by 2-dimensional stochastic optical reconstruction microscopy, which allows assessment of density and size of Na_v1.5 clusters within distinct subcellular domains, such as LM and ID (Figure 5A and 5B). Under control conditions, statistically significant differences in Na_v1.5 clusters between DMSO-treated control and PKP2cKO cardiomyocytes were only observed at the ID, with a larger cluster size found in PKP2cKO (Figure 5D). In control cells, parthenolide did not affect Na_v1.5 cluster size or density, regardless of the cell region (comparing control DMSO versus control parthenolide; Figure 5C through 5F). Yet, in PKP2cKO cells, parthenolide brought about a significant increase in Na_v1.5 cluster density at the ID (Figure 5C), while no significant effects of parthenolide on Na_v1.5 cluster size or density were observed at the LM (Figure 5E and 5F). Together, these findings are in line with a modulatory impact of microtubule detyrosination on subcellular Na_v1.5 distribution in the setting of PKP2 deficiency.

Figure 5.

Effect of parthenolide (PTL) on cardiac sodium channel (Na_v1.5) cluster density and size at the lateral membrane and intercalated disc in control (CTR) and cardiomyocyte-specific, tamoxifen-activated knockout of plakophilin-2 (PKP2cKO) mouse cardiomyocytes (CMs). Representative stochastic optical reconstruction microscopy (STORM) images showing immunolabeled Na_V1.5 (green) and N-cadherin (red) after 2 to 4 hours of incubation with either dimethyl sulfoxide (DMSO) or PTL in CTR (A) and PKP2cKO (B) CMs (scale bars: 10 µm). C through F, Quantification of Na_V1.5 cluster density and cluster size at the intercalated disc (C and D) and lateral membrane (E and F) in CTR and PKP2cKO CMs after 2 to 4 hours of incubation with either DMSO or PTL (10 µmol/L). n represents the number of cells measured. CTR: N=3 mice; PKP2cKO: N=4 mice. ***P<0.001; 1-way ANOVA followed by the Holm-Sidak test for pairwise multiple comparison (C and D), 1-way ANOVA (E), and Kruskal-Wallis followed by the Dunn test (F). Comparisons for all graphs: CTR DMSO vs PKP2cKO DMSO, CTR DMSO vs CTR PTL, and PKP2cKO DMSO vs PKP2cKO PTL.

Increased Microtubule Detyrosination and Decreased I_Na Density in ACM hiPSC-CMs Are Restored by Parthenolide

To assess the relevance of our observations in a human ACM disease model, we used hiPSC-CMs with the PKP2-c.2013delC mutation and their isogenic controls. Immunofluorescence experiments showed that total α-tubulin levels were not altered in PKP2-c.2013delC hiPSC-CMs. However, a significant increase in the detyrosinated α-tubulin fraction and the ratio of detyrosinated versus total α-tubulin was observed compared with the control. This increase was prevented by incubation with parthenolide (Figure 6A through 6E). In contrast to what we observed in control mouse cardiomyocytes, parthenolide reduced the detyrosinated α-tubulin fraction and the ratio of detyrosinated to total α-tubulin also in control hiPSC-CMs (Figure 6D and 6E). As previously shown,⁵ peak I_Na was significantly decreased in PKP2-c.2013delC compared with control hiPSC-CMs (maximal I_Na density, control: −194.9±22.1; PKP2-c.2013delC: −99.0±13.7; P=0.003), without any significant differences in gating properties (Figure 7). Different from control mouse cardiomyocytes, incubation with parthenolide also increased I_Na density in control hiPSC-CMs (by ≈40%; Figure 7A and 7C; Table S5). Parthenolide did not affect I_Na voltage dependence of inactivation in control hiPSC-CMs, yet caused a small (≈2 mV) but significant negative shift of the voltage dependence of activation (Figure 7E; Table S5). Crucially, parthenolide significantly increased peak I_Na in PKP2-c.2013delC hiPSC-CMs, doubling current density and restoring it to control levels (Figure 7B and 7D; Table S6), without impacting on voltage dependence of activation or inactivation (Figure 7F; Table S6).

Figure 6.

Microtubule (MT) detyrosination is increased in PKP2-c.2013delC human-induced pluripotent stem cell–derived-cardiomyocytes (hiPSC-CMs) vs isogenic control hiPSC-CMs (control [CTR]) and is attenuated by parthenolide (PTL). Representative examples of confocal images showing immunolabeled total α-tubulin (red) and detyrosinated α-tubulin (green) in CTR (A) and PKP2-c.2013delC hiPSC-CMs (B) after incubation with either dimethyl sulfoxide (DMSO) or PTL (nuclei in blue stained with 4′,6-diamidino-2-phenylindole [DAPI]; scale bar, 20 µm). Quantification of average total α-tubulin normalized to the number of nuclei per area (C), detyrosinated α-tubulin normalized to the number of nuclei per area (D), and ratio of detyrosinated α-tubulin to total α-tubulin (E) in CTR and PKP2-c.2013delC hiPSC-CMs after 4 hours of incubation with either DMSO or PTL (10 µmol/L). n=30 represents the number of areas analyzed from 3 independent differentiations (n=10 from 2 independent wells for each experiment). **P<0.01, ***P<0.001, and ****P<0.0001; 1-way ANOVA (C) and Kruskal-Wallis followed by the Dunn test for pairwise multiple comparison (D and E). Comparisons for all graphs: CTR DMSO vs PKP2-c.2013delC DMSO, CTR DMSO vs CTR PTL, and PKP2-c.2013delC DMSO vs PKP2-c.2013delC PTL.

Figure 7.

Parthenolide (PTL) restores sodium current (I_Na) density in PKP2-c.2013delC human-induced pluripotent stem cell–derived-cardiomyocytes (hiPSC-CMs). Representative I_Na traces measured in control (CTR; A) and PKP2-c.2013delC (B) hiPSC-CMs treated with either dimethyl sulfoxide (DMSO) or PTL. Average current-voltage (I-V) relationships (C and D) and average voltage dependence of activation and inactivation (E and F) in CTR and PKP2-c.2013delC hiPSC-CMs after 2 to 4 hours of incubation with either DMSO or PTL (10 µmol/L). Insets: voltage-clamp protocols. Dotted lines indicate the Boltzmann fit to the average data. n represents the number of cells measured. Data were collected from 4 independent differentiations for both CTR hiPSC-CMs and PKP2-c.2013delC hiPSC-CMs. *P<0.05, 2-way ANOVA repeated measures followed by the Holm-Sidak test for post hoc analysis.

These observations underscore the presence and functional relevance of increased microtubule detyrosination in a human ACM model.

EpoY Decreases Microtubule Detyrosination and Increases I_Na Density in ACM hiPSC-CMs

Because parthenolide may also act on other pathways in addition to reducing microtubule detyrosination,^30,31 we also assessed the effects of EpoY, an inhibitor of VASH (vasohibin) and SVBP (small VASH-binding protein), the enzyme complex responsible for α-tubulin detyrosination¹⁹ and, therefore, a more specific pharmacological inhibitor of microtubule detyrosination. Similar to parthenolide, EpoY significantly reduced the detyrosinated α-tubulin fraction and the ratio of detyrosinated to total α-tubulin in PKP2-c.2013delC hiPSC-CMs but did not change total α-tubulin levels (Figure 8A and 8B). Finally, EpoY significantly increased I_Na density without affecting Na_v1.5 voltage dependence of (in)activation (Figure 8C and 8D; Table S7), confirming a direct mechanistic link between microtubule detyrosination and I_Na.

Figure 8.

EpoY decreases microtubule (MT) detyrosination and restores sodium current (I_Na) density in PKP2-c.2013delC human-induced pluripotent stem cell–derived-cardiomyocytes (hiPSC-CMs). A, Representative examples of confocal images showing immunolabeled total α-tubulin (red) and detyrosinated α-tubulin (green) in PKP2-c.2013delC hiPSC-CMs after incubation with either dimethyl sulfoxide (DMSO) or EpoY (nuclei in blue stained with 4′,6-diamidino-2-phenylindole [DAPI]; scale bar, 50 µm). B, Quantification of average total α-tubulin normalized to the number of nuclei per area, detyrosinated α-tubulin normalized to the number of nuclei per area, and ratio of detyrosinated α-tubulin to total α-tubulin in PKP2-c.2013delC hiPSC-CMs after 4 hours of incubation with either DMSO or EpoY (20 µmol/L). n=28 represents the number of areas analyzed from 3 independent differentiations (n=8–10 from 2 independent wells for each experiment). ****P<0.0001; Mann-Whitney U test for detyrosinated α-tubulin and for total α-tubulin, unpaired Student t test for the ratio of detyrosinated α-tubulin to total α-tubulin. C, Representative I_Na traces measured in PKP2-c.2013delC treated with either DMSO or EpoY. D, Average current-voltage (I-V) relationships (left) and average voltage dependence of activation and inactivation (right) in PKP2-c.2013delC hiPSC-CMs after 2 to 4 hours of incubation with either DMSO or EpoY (20 µmol/L). Insets: voltage-clamp protocols. Dotted lines indicate the Boltzmann fit to the average data. n represents the number of cells measured. Data were collected from 4 independent differentiations. *P<0.05, 2-way ANOVA repeated measures followed by the Holm-Sidak test for post hoc analysis.

Discussion

Increased microtubule detyrosination is known to cause mechanical resistance in the heart and compromise contraction in the setting of different pathological conditions, such as HF, DMD, and hypertrophic cardiomyopathy.^{13,14,32–34} In addition, we recently showed that microtubule detyrosination is associated with decreased Na_v1.5 membrane localization and reduced I_Na in DMD cardiomyocytes, and this can be rescued by parthenolide.¹⁷ Several models of ACM, including PKP2 deficiency, have demonstrated decreased I_Na,^4,5,29 in addition to functional and structural remodeling, including contractile dysfunction.^5,8,9,35,36 We, therefore, here investigated the potential modulatory role of microtubule dynamics in the setting of ACM. Our findings demonstrate the presence of increased microtubule detyrosination in PKP2-deficient mouse cardiomyocytes and PKP2-c.2013delC hiPSC-CMs. In addition, we observed increased membrane stiffness in PKP2-deficient cardiomyocytes, which was attenuated by pharmacological reduction of microtubule detyrosination. The latter also restored I_Na in PKP2-c.2013delC hiPSC-CMs and increased I_Na and Na_v1.5 membrane localization within 2 subcellular domains of PKP2-deficient cardiomyocytes (ID and LM), indicating a modulatory impact of microtubule dynamics on Na_v1.5 trafficking and subcellular distribution. Overall, our results demonstrate that microtubule detyrosination secondary to loss of PKP2 impacts cardiomyocyte (dys)function beyond the desmosome, contributing to both electrical and mechanical alterations in the setting of ACM.

The tyrosination/detyrosination cycle is a posttranslational modification of microtubules mediated by 2 enzymes: the VASH and SVBP complex, responsible for the removal of tyrosine from microtubules (ie, detyrosination) and TTL (tubulin tyrosine ligase), which adds back tyrosine onto microtubules.³⁷ Tyrosinated tubulin is associated with more dynamic microtubules,³⁷ while tubulin detyrosination increases microtubule stability,³⁸ leading to membrane stiffness and impaired cardiomyocyte contraction.¹⁵ As such, depletion of VASH, and, therefore, of detyrosinated microtubules, improves contraction and relaxation in failing cardiomyocytes.¹¹ The compound parthenolide, a sesquiterpene lactone originating from the feverfew plant, is mainly known for its anticancer and anti-inflammatory properties^30,31 but has also been shown to decrease the detyrosinated fraction of microtubules.^10,12–14 The mechanism by which parthenolide achieves the latter is not completely understood. It has been shown that parthenolide does not act directly on the VASH-SVBP complex but, instead, binds to tubulin and covalently modifies it, thereby destabilizing microtubules.¹⁹ Nevertheless, we recently demonstrated that both parthenolide and EpoY, an inhibitor of the VASH-SVBP complex and, therefore, a more specific pharmacological blocker of microtubule detyrosination,¹⁹ were able to reduce microtubule detyrosination and increase I_Na in a mouse model of DMD cardiomyocytes.¹⁷ Furthermore, we now show that EpoY significantly decreased detyrosinated α-tubulin levels and increased I_Na density in PKP2-c.2013delC hiPSC-CMs, confirming a direct mechanistic link. Hence, while the mechanisms by which parthenolide and EpoY decrease microtubule detyrosination may be different, the end result of both interventions showed similar beneficial effects.

Employing the well-characterized PKP2cKO mouse model and a human ACM model, we here, for the first time, demonstrate increased microtubule detyrosination and membrane stiffness in the setting of ACM. The mechanisms underlying the observed alterations in microtubule dynamics remain to be fully elucidated. Altered expression levels of proteins regulating microtubule stability have been reported in the setting of dilated, ischemic, and hypertrophic cardiomyopathy.³⁷ In addition, conditions such as DMD, hypertrophic cardiomyopathy, and HF share many pathological features, including intracellular calcium dysregulation, mitochondrial dysfunction, and increased reactive oxygen species production^{6,10–15,39,40} but to what extent these are the cause or consequence of microtubule detyrosination is not clear. It has been shown that intracellular calcium (Ca²⁺) is required for efficient microtubule detyrosination, and Ca2⁺-dependent proteases, CAPN1 (calpain 1) and CAPN2 (calpain 2), can regulate VASH-mediated microtubule detyrosination⁴¹; hence, the changes in intracellular Ca²⁺previously observed in PKP2cKO mice may have contributed.^18,36,42 Interestingly, Kim et al⁴² demonstrated that this Ca²⁺ dysregulation actually precedes the development of cardiomyopathy in PKP2cKO mice, consequent to an increased Ca²⁺ influx secondary to increased membrane permeability through connexin-43 hemichannels. They also observed widening of the intercellular space, reduced cluster size of phosphorylated PKC (protein kinase C), and separation of the latter from the adhesive complex at the cell end. These observations are consistent with previous studies demonstrating dislodgment of PKC from the cell junction after loss of mechanical junction integrity. PKP2 has been shown to regulate PKC signaling by recruiting it to DSP; as a consequence, loss of PKP2 disrupts the DSP-PKP2-PKC complex, leaving PKC free to phosphorylate other substrates.⁴³ One potential substrate includes TTL: PKC-mediated phosphorylation of TTL inhibits its activity, thereby preventing tubulin tyrosination.⁴⁴ Hence, the increased free PKC induced by loss of PKP2 would be expected to increase TTL phosphorylation and consequently increase microtubule detyrosination. Finally, PKP2 deficiency leads to dissociation of DSP from adherens junctions,⁴⁵ and DSP loss has been shown to alter microtubule dynamics,⁴⁶ providing an additional potential mechanism underlying the observed increased microtubule detyrosination in PKP2cKO mouse cardiomyocytes and PKP2-c.2013delC hiPSC-CMs.

Independent of the exact mechanisms involved, increased microtubule detyrosination clearly has detrimental consequences, as evidenced by the fact that PKP2cKO cardiomyocytes displayed increased membrane stiffness, which was attenuated by parthenolide. This increased membrane stiffness, and the increased buckling effect expected to result from the less flexible detyrosinated microtubules running along the long axis of the cells,¹² may oppose the increased sarcomeric shortening that results from prolonged calcium transients.^35,36 It is reasonable to assume that these opposing forces can create a mechanical stress, furthered by the lack of proper anchoring at the ID, damaging the integrity of the cell as a whole. In addition to restoring membrane stiffness, parthenolide was able to almost double I_Na in PKP2cKO mouse cardiomyocytes (while not affecting the voltage dependence of activation or inactivation) and increase Na_v1.5 on the membrane. Although it is becoming increasingly clear that the microtubule network plays a key role in the trafficking of Na_v1.5 in cardiomyocytes, the exact mechanisms remain to be explored. Our previous work demonstrated that Taxol (which increases microtubule detyrosination) decreased I_Na,⁴⁷ while reducing microtubule detyrosination in a mouse model of DMD cardiomyocytes increased I_Na,¹⁷ underscoring the modulatory impact of microtubule dynamics on Na_v1.5. It has been demonstrated that trafficking of Na_v1.5 by kinesins and dyneins, the motor proteins responsible for its transport on microtubules in the cell, is faster on tyrosinated microtubules.^48,49 Moreover, microtubule plus (+)-end tracking proteins that are important for the delivery of Na_v1.5 to the membrane, such as EB1 (end-binding protein 1), CLIP-170 (cytoplasmic linker protein 170), and CAP-Gly (cytoskeleton-associated protein glycine-rich),^50,51 may also play a role. In fact, the interaction between EB1 and CLIP-170 appears to be facilitated on tyrosinated microtubules,^52–55 while CAP-Gly has been shown to be more present on tyrosinated microtubules,⁵² indicating that tyrosinated microtubules could facilitate the release of Na_v1.5 at its membrane destination. Hence, multiple factors may modulate the observed impact of microtubule detyrosination on I_Na. While parthenolide did not affect I_Na or Na_v1.5 subcellular distribution in control mouse cardiomyocytes, it did affect both microtubule detyrosination and I_Na in (isogenic) control hiPSC-CMs; possible explanations for this observation include species differences and relative immaturity of hiPSC-CMs. Importantly, PKP2-c.2013delC hiPSC-CMs displayed not only increased microtubule detyrosination but also reduced I_Na compared with (isogenic) control hiPSC-CMs, which were both restored to control levels by parthenolide and EpoY. These observations confirm the functional relevance of microtubule remodeling in a human ACM model.

Interestingly, contrary to our findings in PKP2-c.2013delC hiPSC-CMs, we did not observe a baseline difference in I_Na between PKP2cKO and control mouse cardiomyocytes. There are several potential explanations for these observations, including differences in genetic defect (deletion versus mutation), species (mouse versus human), model (cultured hiPSC-CMs versus intact heart comprising also other cell types), and level of maturity (hiPSC-CMs versus adult mouse). In addition, (compensatory) alterations in gene expression, which may have counterbalanced I_Na reduction in PKP2cKO cardiomyocytes, may not occur in hiPSC-CMs. The unaltered baseline I_Na in PKP2cKO cardiomyocytes is also different from our previous observations in mice with heterozygous PKP2 deletion,⁴ which may be consequent to the fact that PKP2 deficiency is present from the embryonic stage in the latter but induced at the adult stage in the former. In cardiomyocytes from PKP2 heterozygous mice, we previously observed a large subpopulation of Na_v1.5 clusters held back, a distance away from the ID region at the cell end (marked by the location of the N-cadherin plaque).²⁸ We speculated that an impediment to complete the last steps of microtubule-mediated trafficking and insertion caused this clustering of Na_v1.5 at a distance from the membrane and hence, reduced I_Na. In the PKP2cKO cardiomyocytes investigated in the current study, this impediment may not be strong enough to retain all Na_v1.5 clusters away from the membrane, resulting in a current density not different from that observed in control cardiomyocytes. As such, the increased Na_v1.5 cluster size observed in the ID region of PKP2cKO cardiomyocytes under control conditions may reflect a combination of channels that have already reached the membrane (maintaining I_Na) and channels-in-waiting retained at a distance from the ID. Upon exposure to parthenolide, the consequent increased microtubule dynamics then facilitate membrane delivery of this retained pool of channels, thereby increasing Na_v1.5 cluster density and I_Na at the ID region. At the LM of PKP2cKO cardiomyocytes, parthenolide increased I_Na yet had no significant effects on Na_v1.5 cluster size or density. While subdomain-specific differences in microtubule-mediated trafficking and modulatory proteins may play a role here, methodological factors should also be considered. Given the fact that the impact of parthenolide on I_Na in this subdomain was relatively small compared with the ID, it is possible that the stochastic optical reconstruction microscopy measurements were not able to detect significant differences in Na_v1.5. Morphological differences between the LM and ID, including a larger volume in the former, may also have contributed. In any case, our observation that parthenolide increased I_Na at both the ID and LM underscores the fact that altered microtubule dynamics in the setting of ACM have functional implications beyond the desmosome. The latter is also of particular relevance for the observed microtubule detyrosination-dependent membrane stiffness observed in PKP2cKO cardiomyocytes.

Regarding arrhythmogenesis, it is our current view (also based on our previous observations^18,36,42) that ventricular arrhythmias in PKP2-deficient hearts are primarily consequent to intracellular calcium dysregulation (acting as a trigger), altered microarchitecture due to developing fibrosis (a substrate), and the preservation of enough excitable volume to support multiple fragmented waves, hence presenting the 3 conditions necessary for fibrillatory activity (trigger, substrate, and excitable volume). In this context, reduced I_Na would be proarrhythmic, and hence, its restoration would be potentially beneficial. Yet, even when I_Na remains unchanged (as in PKP2cKO cardiomyocytes), one can speculate that an intervention that increases I_Na without altering kinetics (eg, by favoring trafficking) would have a potential antiarrhythmic effect, as it would increase the electrotonic influence of an activating site, favoring conduction and, as such, minimizing excitable gaps that could permit reentrant activity.

In conclusion, we demonstrate that increased microtubule detyrosination secondary to loss of PKP2 impacts cardiomyocyte (dys)function beyond the desmosome, contributing to both electrical and mechanical alterations in the setting of ACM. Because clinical management of patients with ACM is aimed at preventing HF, as well as arrhythmias and sudden cardiac death, a therapeutic approach targeting electrical and structural remodeling would be highly advantageous, in particular because these are interrelated and may exacerbate each other. As such, our present findings identify microtubule detyrosination as a potential novel therapeutic strategy in ACM, aimed at preventing both contractile and electrical dysfunction.

ARTICLE INFORMATION

Acknowledgments

The authors thank Sagun Jonchhe, Dipika Gupta, and Shaun Christie from the Department of Biochemistry and Molecular Pharmacology, NYU School of Medicine, for expert technical support in performing stochastic optical reconstruction microscopy experiments.

Disclosures

None.

Supplemental Material

Supplemental Methods

Tables S1–S7

References 56–59

Supplementary Material

hae-19-e014354-s001.pdf ^{(392KB, pdf)}

Funding Statement

This work was supported by the Foundation Leducq (grant 17CVD02 to Drs Remme and Delmar); the Netherlands CardioVascular Research Initiative (CVON): the Dutch Heart Foundation, Dutch Federation of University Medical Centers, the Netherlands Organisation for Health Research and Development, the Royal Netherlands Academy of Sciences (grants CVON-eDETECT 2015-12 and CVON-PREDICT2 2018-30 to Dr Remme); the Dutch Heart Foundation (grant PREVENT 01-003-2023-0439 to Dr Remme); the Netherlands Organisation for Health Research and Development (ZonMw; Innovational Research Incentives Scheme Vidi grant 91714371 to Dr Remme and PSIDER project 10250022110004 to Drs Bellin and Meraviglia); the Novo Nordisk Foundation Center for Stem Cell Medicine (Novo Nordisk Foundation grant NNF21CC0073729 to Dr Bellin); European Union’s Horizon 2020 Research and Innovation Programme under European Research Council (grant ERC-CoG Mini-HEART 101001746 to Dr Bellin); the National Institutes of Health (grants RO1-HL148609 and R35-HL160840 to Dr Delmar and grants 1R35GM134947-01, 1R01AI153040-01, and 1P01CA247773-01/549 to Dr Rothenberg); the British Heart Foundation (grant RG/F/22/110081 to Drs Gorelik and Sanchez-Alonso, and grant FS/IBSRF/23/25188 to Dr Sanchez-Alonso); and the British Heart Foundation Imperial Center of Research Excellence Award (grant RE/18/4/34215 to Dr Swiatlowska). Genomics of Unexplained Cardiac Arrest (GenUCA): the British Heart Foundation, the German Centre for Cardiovascular Research, and the Dutch Heart Foundation (grant 2020B001 to Dr Remme).