Hemodynamic stress‐induced cardiac remodelling is not modulated by ablation of phosphodiesterase 4D interacting protein

Abstract Adrenergic stimulation in the heart activates the protein kinase A (PKA), which phosphorylates key proteins involved in intracellular Ca2+ handling. PKA is held in proximity to its substrates by protein scaffolds, the A kinase anchoring proteins (AKAPs). We have previously identified the transcript of phosphodiesterase 4D interacting protein (Pde4dip; also known as myomegalin), one of the sarcomeric AKAPs, as being differentially expressed following hemodynamic overload, a condition inducing hyperadrenergic state in the heart. Here, we addressed whether PDE4DIP is involved in the adverse cardiac remodelling following hemodynamic stress. Homozygous Pde4dip knockout (KO) mice, generated by CRISPR‐Cas9 technology, and wild‐type (WT) littermates were exposed to aortocaval shunt (shunt) or transthoracic aortic constriction (TAC) to induce hemodynamic volume overload (VO) or pressure overload (PO), respectively. The mortality, cardiac structure, function and pathological cardiac remodelling were followed up after hemodynamic injuries. The PDE4DIP protein level was markedly downregulated in volume‐overloaded‐ but upregulated in pressure‐overloaded‐WT hearts. Following shunt or TAC, mortality rates were comparably increased in both genotypes. Twelve weeks after shunt or TAC, Pde4dip‐KO animals showed a similar degree of cardiac hypertrophy, dilatation and dysfunction as WT mice. Cardiomyocyte hypertrophy, myocardial fibrosis, reactivation of cardiac stress genes and downregulation of ATPase, Ca2+ transporting, cardiac muscle, slow twitch 2 transcript did not differ between WT and Pde4dip‐KO hearts following shunt or TAC. In summary, despite a differential expression of PDE4DIP protein in remodelled WT hearts, Pde4dip deficiency does not modulate adverse cardiac remodelling after hemodynamic VO or PO.


| INTRODUC TI ON
In response to hemodynamic stress, that is volume overload (VO) and pressure overload (PO), the heart undergoes molecular, structural and functional changes, collectively named cardiac remodelling. This is associated with β-adrenergic overstimulation, 1 cyclic adenosine monophosphate (cAMP) synthesis that activates protein kinase A (PKA), which phosphorylates Ca 2+ regulatory proteins to maintain cardiac output. 2,3 However, a persistent hyperadrenergic state leads to maladaptive cardiac remodelling, and ultimately to heart failure (HF) and sudden death. 4 The spatiotemporal tuning of cAMP/PKA signalling is achieved by a group of scaffolding proteins, the A-kinase anchoring proteins (AKAPs), which held PKA in proximity to its distinct substrates. 5,6 Additionally, AKAPs associate with phosphatases and phosphodiesterases, thereby providing multistep control of kinase activity to ensure proper subcellular cAMP/PKA compartmentalization. 7,8 In the heart, several AKAPs are implicated in cardiac pathophysiology. [9][10][11][12][13][14][15] Hence, regulation of AKAPs activity or expression could be a promising therapeutic strategy for treatment of heart diseases.
Phosphodiesterase 4D interacting protein (PDE4DIP), also referred to as myomegalin, is an AKAP protein that is predominantly expressed in cardiac and skeletal muscle sarcomeres in proximity to the z-disc and sarcoplasmic reticulum. [16][17][18][19][20][21] The PDE4DIP anchors PKA to cardiac myosin binding protein-C (cMyBPC) and cardiac troponin I (cTNI), and therefore facilitates PKA-mediated phosphorylation of both proteins to augment cardiac contraction under adrenergic stress. 17,22,23 Taken together, these findings suggest a possible role for PDE4DIP in cAMP/PKA compartmentalization to maintain intracellular Ca 2+ homeostasis. Therefore, it is not surprising that Pde4dip mutations are associated with familial dilated cardiomyopathy and arrhythmia. [24][25][26] However, the precise role of PDE4DIP in response to pathological stress in vivo remains unknown.
We have previously reported that Pde4dip transcript is differentially expressed in the remodelled myocardium upon hemodynamic overload. 27 Consistently, in the current study, we found that PDE4DIP protein levels are significantly decreased in the mouse heart after VO but increased after PO. However, it remains unknown whether this differential cardiac PDE4DIP expression has a functional role or is only an epiphenomenal event. To explore the role of PDE4DIP in cardiac remodelling, we generated Pde4dip knockout (Pde4dip-KO) mice using the CRISPR-Cas9 technique and exposed them to aortocaval shunt (shunt)-triggered VO or transthoracic aortic constriction (TAC)-induced PO. Our data show that PDE4DIP does not seem to be necessary for adverse cardiac remodelling following either VO or PO.

| Mice
All investigations were approved by the responsible Institutional In this study, we analysed male and female wild-type (WT) and Pde4dip-KO littermates on C57BL/6N genetic background.

| Transthoracic aortic constriction
Surgery was done using a minimally invasive technique as described previously. 27 Briefly, 10-week-old WT and Pde4dip-KO mice were anaesthetised using intraperitoneal injections of a mixture of xylazine and ketamine. A 27-gauge needle was tied against the aorta using a 5-0 non-absorbable suture. Sham animals underwent the same procedure except banding of the aorta.

| Aortocaval shunt
Surgery was done as described previously. 28 In brief, 10 week-old mice were anaesthetised using isoflurane, and a mid-line laparotomy was performed to expose the abdominal aorta and inferior vena cava between the renal arteries and the iliac bifurcation. A 23-gauge needle was inserted into the exposed aorta at a 45 deg angle and pushed through to the inferior vena cava, creating the shunt. Cyanoacrylate (Pattex) was used to seal the puncture. Sham animals underwent the same procedure except for the creation of the shunt.

| Echocardiography
The mice were anaesthetised using 1% isoflurane, and echocardiography was performed using Vevo2100 Imaging Software 3.1.0 (VisualSonics). During this procedure, the body temperature was maintained within physiological range (36°C-37.5°C) using a heating pad, and heart rates were kept consistent between experimental groups at 450-600 b.p.m. Electrocardiogram monitoring was obtained using hind limb electrodes. The left ventricle (LV) geometry and systolic function were assessed by using standard 2D parasternal short axis views in accordance with recommendations where available. 29 Relative wall thickness (RWT) = (LV anterior wall thickness at diastole + LV posterior wall thickness at diastole) divided by LV end-diastolic diameter (LVAWTd+LVPWTd/ LVEDD).
The speckle tracking echocardiography was performed as described. 30 Several tracking points were placed on the endocardial and epicardial border in parasternal long-axis views. These were used as a guide for border delineation and subsequent frame-byframe tracking throughout the cardiac cycle. The software automatically divides the LV into six segments: two basal, two mid and two apical segments and calculate parameters of deformation (strain, strain rate) and parameters of motion (displacement and velocity), separately for each segment as well as overall mean values. The presented data are the averages from these six different values per heart. The peak longitudinal strain rate during early LV filling, termed as the reverse longitudinal strain rate and the radial diastolic peak velocity were quantified to assess the diastolic function. 31

| Quantitative real-time polymerase chain reaction
The DNA-free RNAs were isolated from the left ventricles using RNeasy kit and the RNase-free DNAse Set (Qiagen), and the RNAs concentration was measured by NANO 2000 (Termal Scientific).
The cDNA synthesis was done using the iScript cDNA synthesis kit (Bio-Rad Laboratories). Real-time PCR was performed using Bio-Rad iQ-Cycler. Transcripts were amplified using SYBR green fluorescent dye and calculated with the delta-delta C t method using Gapdh as denominator. Primers included for following genes:

| Immunoblotting
Harvested left ventricle were homogenized in RIPA buffer

| Statistical analysis
Statistical analyses were carried out using Prism software version

| PDE4DIP protein is differentially expressed after pathological hemodynamic overload
We assessed PDE4DIP protein level in WT murine remodelled hearts after shunt and TAC, well-established models of VO-and PO-induced cardiac remodelling, respectively. As shown in Figure 1, PDE4DIP protein levels were differentially expressed in an opposite direction, consistent with our previous microarray data. 27 Compared with sham control groups, average PDE4DIP protein was significantly reduced by ≈43% in shunt hearts (p < 0.05) ( Figure 1A) but increased by ≈50% in the TAC hearts (p = 0.06) ( Figure 1B). These data suggest a potential role for PDE4DIP in myocardial response to hemodynamic stress.

| Generation of Pde4dip-KO mice
To get insights into the PDE4DIP function in the heart, we introduced a genomic deletion in mice by CRISPR/Cas9. Genome editing generated a 11 bp deletion located in the coding region of the exon 4 of the Pde4dip gene ( Figure 1C), resulting in a frameshift mutation, thereby disrupting the reading frame of Pde4dip. Heterozygous Pde4dip +/− founders were identified by Sanger sequencing of PCR products covering the edited region ( Figure 1D). The heterozygous mice from F1 generation were intercrossed and the genotyping of the offspring was performed by digestion of the PCR products with the restriction enzymes, NciI or HpaII; both restriction sites are absent in the KO allele. Thus, the WT allele (210 bp) produces two bands of 130 and 80 bp with either of the two enzymes, whereas the KO allele is not cut and therefore gives a 210 bp PCR fragment ( Figure 1E). Quantitative

RT-PCR analysis of Pde4dip-KO heart showed a very low Pde4dip
transcript level, probably due to unstable Pde4dip mRNAs with the frameshift deletion ( Figure 1F). Western blot analysis clearly showed no detectable PDE4DIP protein in the Pde4dip-KO mice ( Figure 1G).

| No appreciable structural or functional changes in Pde4dip-KO mice at baseline
The Pde4dip-KO mice were born at expected Mendelian ratios (Table S1)

| Survival, cardiac structure and function are comparable between WT and Pde4dip-KO after shunt
We subjected WT and Pde4dip-KO littermates to chronic VO by shunt.
The stroke volume (SV) was significantly, but similarly, increased at 1-week post-shunt in both genotypes, indicating comparable VO ( Figure 3A). The relative wall thickness (RWT) was significantly, but comparably, lower at 1-week post-shunt ( Figure 3B), indicating similar eccentric hypertrophy. The mortality rates up to 12 weeks after shunt did not differ between Pde4dip-KO and WT mice ( Figure 3C).  decreased in both genotypes after shunt ( Figure 3G,H and Table 1).

Echocardiographic analysis revealed that both WT and
The global longitudinal systolic strain rate (systolic parameter), radial diastolic peak velocity and peak reverse longitudinal strain rate (diastolic parameters) were mildly but comparably deteriorated in both genotypes at 12 weeks after shunt ( Figure 3I,J; Table 1).

Consistent with biventricular VO induced by shunt, morphomet-
ric analysis at 12 weeks post-shunt revealed significant LV hypertrophy, as evidenced by increased ratio of the LVW to TL, and marked right ventricle (RV) hypertrophy, as judged by increased Fulton index, was no difference between WT and Pde4dip-KO hearts ( Table 1). LungW was comparably increased after shunt, suggesting congestive HF in both genotypes (Table 1). Pathologically elevated central venous pressure, characterized by hepatic congestion (increased liverW-to-TL), splenic congestion (increased spleenW-to-TL) and peripheral oedema (increased total bodyW-to-TL), was evident after shunt, suggesting similar RV dysfunction in both genotypes (Table 1).

Histological analysis revealed no differences between WT
and Pde4dip-KO hearts at 12 weeks after shunt ( Figure 4A-C).
Cardiomyocyte hypertrophy did not differ between shunt-operated WT and Pde4dip-KO hearts, as evidenced by similar increased cross surface areas ( Figure 4A,B). Shunt hearts showed a tendency towards increased myocardial fibrosis, with no difference between genotypes ( Figure 4A,C). Molecular analysis showed similar upregulation of cardiac stress transcripts, natriuretic peptide type A (Nppa) and natriuretic peptide type B (Nppb) and downregulation of ATPase,

phorylation of troponin-I and PLN did not differ between WT and
Pde4dip-KO hearts in sham or shunt groups ( Figure 4E and Figure S1).
Overall, our data illustrate that both WT and Pde4dip-KO mice experienced comparable adverse cardiac remodelling after shunt.

| Similar pathological cardiac remodelling in WT and Pde4dip-KO mice following TAC
Although both are hemodynamic stresses, PO and VO induce different functional and molecular adaptations in cardiac hypertrophy, causing morphologically distinct types of cardiac remodelling. 27,28 We therefore extended our study and exposed WT and Pde4dip-KO  Figure 5C).  Table 2). The global longitudinal systolic strain rate was significantly reduced in both genotypes at 12 weeks post-TAC, confirming systolic impairment ( Table 2). Furthermore, the diastolic parameters, radial diastolic peak velocity and peak  reverse longitudinal strain rate, were markedly decreased in both genotypes at 12 weeks post-TAC, indicating a comparable diastolic dysfunction ( Figure 5I,J and Table 2).

Quantification of cardiomyocyte hypertrophy revealed no differences in myocyte surface areas in TAC-operated WT and
Pde4dip-KO mice ( Figure 6A,B). Picrosirius red staining showed similar myocardial fibrosis between WT and Pde4dip-KO hearts after TAC ( Figure 6A,C). TAC-induced cardiac hypertrophy was mainly due to LV hypertrophy, as reflected by marked increased in LVW-to-TL ratio ( Table 2). LungW-to-TL ratio was markedly increased in both genotypes after TAC ( Table 2). WT and Pde4dip-KO animals showed a comparable upregulation of Nppa and Nppb and downregulation of Atp2a2 transcripts post-TAC ( Figure 6D).
Taken together, our data show that both WT and Pde4dip-KO mice experienced equal cardiac dysfunction with signs of manifested HF after TAC.

| DISCUSS ION
PDE4DIP interacts with a cAMP-related PDE4D and PKA and recruits cMyBPC and cTNI at the sarcomere. Its sarcomeric localization is therefore compatible with a mechanism that would augment β-adrenergic-mediated phosphorylation of cMyBPC and cTnI to enhance cardiac contraction upon adrenergic stimulation. 16,17,22,23 We therefore propose that PDE4DIP, beside its structural role as scaffold, could be functionally involved in maintenance of intracellular We showed here that CRISPR-mediated targeted deletion of Pde4dip does not alter the adverse cardiac remodelling upon pathological hemodynamic stress. We therefore conclude that murine PDE4DIP is not required for cardiac hypertrophy, fibrosis or contractile dysfunction, at least in response to 12 weeks of hemodynamic challenges. It is likely that our current study was not continued long In summary, although pathological hemodynamic overload markedly alters PDE4DIP protein levels in WT hearts, there was no difference in adverse cardiac remodelling between WT and Pde4dip-KO mice after exposure to VO or PO. Thus, our results do not support a major role of PDE4DIP on myocardial structure and function in healthy heart or following hemodynamic stress.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available in the supplementary material of this article.