PRKAR1A deficiency impedes hypertrophy and reduces heart size

Abstract Protein kinase A (PKA) activity is pivotal for proper functioning of the human heart, and its dysregulation has been implicated in a variety of cardiac pathologies. PKA regulatory subunit 1α (R1α, encoded by the PRKAR1A gene) is highly expressed in the heart, and controls PKA kinase activity by sequestering PKA catalytic subunits. Patients with PRKAR1A mutations are often diagnosed with Carney complex (CNC) in early adulthood, and may die later in life from cardiac complications such as heart failure. However, it remains unknown whether PRKAR1A deficiency interferes with normal heart development. Here, we showed that left ventricular mass was reduced in young CNC patients with PRKAR1A mutations or deletions. Cardiac‐specific heterozygous ablation of PRKAR1A in mice increased cardiac PKA activity, and reduced heart weight and cardiomyocyte size without altering contractile function at 3 months of age. Silencing of PRKAR1A, or stimulation with the PKA activator forskolin completely abolished α1‐adrenergic receptor‐mediated cardiomyocyte hypertrophy. Mechanistically, depletion of PRKAR1A provoked PKA‐dependent inactivating phosphorylation of Drp1 at S637, leading to impaired mitochondrial fission. Pharmacologic inhibition of Drp1 with Mdivi 1 diminished hypertrophic growth of cardiomyocytes. In conclusion, PRKAR1A deficiency suppresses cardiomyocyte hypertrophy and impedes heart growth, likely through inhibiting Drp1‐mediated mitochondrial fission. These findings provide a potential novel mechanism for the cardiac manifestations associated with CNC.

Patients with PRKAR1A mutations are often diagnosed with Carney complex (CNC) in early adulthood, and may die later in life from cardiac complications such as heart failure. However, it remains unknown whether PRKAR1A deficiency interferes with normal heart development. Here, we showed that left ventricular mass was reduced in young CNC patients with PRKAR1A mutations or deletions. Cardiac-specific heterozygous ablation of PRKAR1A in mice increased cardiac PKA activity, and reduced heart weight and cardiomyocyte size without altering contractile function at 3 months of age. Silencing of PRKAR1A, or stimulation with the PKA activator forskolin completely abolished α1-adrenergic receptor-mediated cardiomyocyte hypertrophy. Mechanistically, depletion of PRKAR1A provoked PKA-dependent inactivating phosphorylation of Drp1 at S637, leading to impaired mitochondrial fission.
Pharmacologic inhibition of Drp1 with Mdivi 1 diminished hypertrophic growth of cardiomyocytes. In conclusion, PRKAR1A deficiency suppresses cardiomyocyte hypertrophy and impedes heart growth, likely through inhibiting Drp1-mediated mitochondrial fission. These findings provide a potential novel mechanism for the cardiac manifestations associated with CNC.

K E Y W O R D S
adenylyl cyclase, cAMP, cardiomyopathy, catecholamine, myocardial development 1 | INTRODUCTION Myocardial 3',5'-cyclic adenosine monophosphate (cAMP)-dependent protein kinase (protein kinase A, PKA) is a key regulator of heart rate and cardiac contractility, through phosphorylation of multiple calcium-handling proteins involved in excitation-contraction coupling (Zhang et al., 2019). The PKA holoenzyme consists of two regulatory and two catalytic subunits. When intracellular cAMP levels are low, the regulatory subunits bind and inhibit the catalytic subunits. Stimulation of β-adrenergic receptors (β-ARs) by the catecholamine neurohormones results in activation of the stimulatory G protein (Gαs), thereby initiating adenylyl cyclase-dependent synthesis of cAMP. The cAMP molecules then bind the PKA regulatory subunits, leading to release of the PKA catalytic subunits and subsequent kinase activation.
Mammalian cells express four isoforms of PKA regulatory subunits that are functionally nonredundant: R1α, R1β, R2α, and R2β (encoded by PRKAR1A, PRKAR1B, PRKAR2A, and PRKAR2B, respectively). Among these regulatory subunits, R1α is the most abundantly and ubiquitously expressed, and is the only isoform required for embryonic heart development (Kirschner, Yin, Jones, & Mahoney, 2009;Taylor, Ilouz, Zhang, & Kornev, 2012). In humans, inactivating mutations or deletions of the PRKAR1A gene have been associated with Carney complex (CNC), an autosomal dominant genetic disorder characterized by spotty skin pigmentation and tumors in the endocrine glands, heart (cardiac myxoma), skin, breast, and other body parts (Casey et al., 2000;Kirschner et al., 2000). CNC is typically diagnosed from childhood to adulthood, with a median age of detection at approximately 20 years (Correa, Salpea, & Stratakis, 2015). The historic adjusted average life span for these patients is 50-55 years (Correa et al., 2015). It is estimated that more than 50% of mortality related to CNC is attributed to cardiovascular complications such as heart failure (Correa et al., 2015), indicating a potential role of PRKAR1A in the maintenance of heart function and morphology. In this study, we evaluated the heart size in a cohort of young CNC patients with identified PRKAR1A mutations or deletions. Our analysis revealed that young patients with CNC had reduced left ventricular mass.
CNC is predominantly associated with PRKAR1A haploinsufficiency in patients (Casey et al., 2000;Veugelers et al., 2004). To investigate the impact of PRKAR1A deficiency on myocardial development, we generated a cardiac-specific heterozygous PRKAR1A knockout mouse line (cPRKAR1A +/-). Here, we showed that ablation of PRKAR1A in mice diminished cardiomyocyte hypertrophy and impeded physiological heart growth. In vitro studies further revealed that disruption of PRKAR1A induced PKA-dependent Drp1 inhibition, which was sufficient to repress hypertrophy.  (Correa et al., 2015;Kirschner et al., 2000).

| Magnetic resonance imaging (MRI) in CNC patients
Cardiovascular magnetic resonance steady-state free precession cine imaging was performed as described previously (Kramer et al. 2013), in a small group of CNC patients with identified PRKRA1A mutations or deletions but without prior history of cardiac myxoma. To avoid the confounders of hypertrophy that may occur with age and hypertension, the subjects selected were below or at 21 years of age. Endocardial and epicardial borders were manually drawn on volumetric short-axis slices of the left ventricle to determine the end-diastolic myocardial mass. The left ventricular mass data were then indexed to body surface area, and compared with ageand gender-matched normal references (Ven et al., 2020).

| Animals
PRKAR1A flox/+ mice (Kirschner et al., 2005) and Mlc2v-Cre +/mice (Chen et al., 1998) were backcrossed to the C57BL/6 background for at least eight generations prior to subsequent breeding. All animals were housed in the campus vivarium accredited by the American Association for Accreditation of Laboratory Animal Care. Animal studies comply with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (8th edition, 2011), and were approved by the Institutional Animal Care and Use Committee at Washington State University.

| Echocardiography
Cardiac contractile function and morphology in 3-monthold male mice was evaluated by echocardiography using the VisualSonics ultrasound imaging system VEVO 2100 equipped with a 55-MHz MS550S transducer. All mice were placed on a warmed table under anesthesia with 1.5% isoflurane for imaging. Standard short-axis M-mode views were recorded at heart rate between 450 and 600 beats per minute. Functional parameters were averaged from three consecutive contractions.

| Measurement of heart weight and cardiomyocyte cross-sectional area
Hearts from 3-month-old male mice were harvested, weighed, and immediately fixed in 4% paraformaldehyde. Heart paraffin sections were subjected to antigen retrieval using 10 mmol/L citrate buffer (pH 6.0) prior to staining with wheat germ agglutinin (W32464, Invitrogen) and mouse anticardiac troponin T (MS-295-P, Thermo Scientific, 1:100). Cardiomyocyte crosssectional area was measured from at least 250 cells in 8-9 fields per heart using the ImageJ software.

| siRNA transfection
NRCMs were transfected with siRNAs in serum-free medium 199 using HiPerfect transfection reagent (Qiagen), and H9c2 cells were transfected in serum-free Dulbecco's modified Eagle's medium using DharmaFECT 1 transfection reagent (Dharmacon) according to the manufacturers' protocols. The siRNA sequences used were as follows: PRKAR1A siRNAs (siPRKAR1A)

| Mitochondrial morphology analysis
Cells were incubated with MitoTracker Red for 30 min and visualized under a confocal laser-scanning microscope (Leica). Images acquired from three independent experiments were analyzed using the ImageJ software. To calculate the percentage of cells with elongated or fragmented mitochondria, at least 500 cells were counted. To analyze mitochondrial length, at least 15 mitochondrial particles per cell from 40 cells were measured.

| Statistical analysis
Statistical analysis was performed using GraphPad Prism 7.02. Results are expressed as mean ± SEM unless indicated otherwise. Student's unpaired t test was used to compare values between two groups except the MRI data from CNC patients, which were analyzed using Mann-Whitney test. One-way analysis of variance followed by Tukey post hoc test was used for multiple-group comparisons. A value of p < .05 was considered statistically significant.

mutations or deletions had reduced left ventricular mass
Mutations or deletions of the PRKAR1A gene cause CNC in humans (Correa et al., 2015). The major clinical manifestation of CNC in the heart is cardiac myxoma (Correa et al., 2015). To determine if CNC is associated with abnormal heart growth, we measured left ventricular mass in CNC patients with identified PRKAR1A mutations or deletions. As left ventricular mass is profoundly increased by hypertension, which is more prevalent at older ages (Buford, 2016), all included patients were at or below 21 years of age (mean age: 15.91 ± 3.30 years). When indexed to body surface area, left ventricular masses in CNC patients were significantly smaller than those in age-and gender-matched reference controls (Table 1).

PRKAR1A heterozygous knockout mice
To determine if PRKAR1A deficiency delays heart growth, we generated cardiac-specific PRKAR1A heterozygous knockout (PRKAR1A flox/+ /Mlc2v-Cre +/-, hereafter, referred to as cPRKAR1A +/-) mice using the Cre-loxP technology, by crossing the PRKAR1A flox/+ mice with the Mlc2v-Cre +/mice ( Figure 1a). Neonatal cPRKAR1A +/mice were born with the expected Mendelian ratio (Figure 1b), and developed into adulthood with similar survival rate as littermates. As the PRKAR1A flox/+ and Mlc2v-Cre +/littermates displayed comparable left ventricular mass and contractile function ( Figure  S1), only the PRKAR1A flox/+ line was used as control thereafter ( Figure 1a). Adult cPRKAR1A +/mice were indistinguishable from control mice in physical appearance, and had similar body weight as controls (Figure 1c). PRKAR1A gene recombination was only detected in cPRKAR1A +/mouse heart, but not in control hearts from the other three genotypes (Figure 1d), or noncardiac tissues in the cPRKAR1A +/mouse ( Figure 1e), indicating cardiac-specific PRKAR1A ablation. Immunoblotting analysis of tissue lysates extracted from 3-month-old cPRKAR1A +/mice further revealed a ~ 50% decrease in PKA R1α protein level in the heart, but not in the skeletal muscle, brain, or stomach (Figure 1f,g).

| Disruption of PRKAR1A induced PKA activation without altering cardiac function
In agreement with previous findings that homozygous ablation of PRKAR1A causes PKA activation (Amieux et al., 2002;Yin et al., 2008), measurement of PKA activity using a phospho-PKA substrate (RRXS*/T*) antibody revealed enhanced phosphorylation of some, but not all PKA substrates in the cPRKAR1A +/heart ( Figure 2a). Moreover, transfection of neonatal rat cardiomyocytes (NRCMs) with PRKAR1Aspecific siRNAs resulted in a similar decrease (~50%) in the level of R1α protein, and markedly increased the levels of phosphorylated PKA substrates (Figure 2b). Robust increases in phospho-PKA substrate levels were also detected after incubation with forskolin, an adenylyl cyclase agonist known to induce extensive PKA activation (Figure 2b). At 3 months of age, cPRKAR1A +/and control mice displayed comparable left ventricular ejection fraction (EF, Figure 2c Figure 2k,l). Together, these results suggested that basal contractile function was maintained in cPRKAR1A +/mice.  (d) DNA isolated from control or cPRKAR1A +/heart (ventricle) was used for genotyping PCR. PRKAR1A deletion band was only detected in the cPRKAR1A +/heart. (e) DNA isolated from heart (ventricle), skeletal muscle, brain, and stomach of the cPRKAR1A +/mouse was used for genotyping PCR. PRKAR1A deletion band was only detected in the heart, but not in other tissues of the cPRKAR1A +/mouse. (f,g) Heart protein lysates from 3-month-old male mice were subjected to immunoblotting (n = 4 per group). Values are mean ± SEM and analyzed using two-tailed Student's t test. ** p < .01

| PRKAR1A deficiency impeded hypertrophic growth during heart development
In line with a trend toward decreased posterior wall thickness and ventricular diameter, echocardiography revealed a significant reduction in the left ventricular mass in cPRKAR1A +/mice ( Figure 3a). Moreover, PRKAR1A ablation significantly decreased heart weight/body weight ratio and heart weight/ tibia length ratio at 3 months of age (Figure 3b,c). Increases in heart weight after birth are achieved predominantly through an increase in cell size (i.e., cardiomyocyte hypertrophy) (Ahuja, Sdek, & MacLellan, 2007). To determine if PRKAR1A deletion hinders cardiac hypertrophic growth, heart sections of cPRKAR1A +/mice and littermates were subjected to wheat germ agglutinin (WGA) staining. As shown in Figure 3d, heterozygous deletion of PRKAR1A dramatically reduced cardiomyocyte cross-sectional area, indicating diminished cardiomyocyte hypertrophy. To determine if PRKAR1A is necessary for α1-adrenergic receptor-mediated hypertrophy, NRCMs were transfected with PRKAR1A siRNAs prior to incubation with phenylephrine (PE), a selective α1-adrenergic receptor agonist. While PE stimulation dramatically increased myocyte surface area in control cells, PRKAR1A-depleted cells were largely irresponsive to PE treatment (Figure 4a,b), suggesting that silencing of PRKAR1A abrogated hypertrophic growth of cardiomyocytes. To further investigate the role of PKA in hypertrophy, NRCMs were treated with PE in the presence of the PKA activator forskolin. Stimulation with forskolin completely abolished PE-induced increase in myocyte size (Figure 4c,d). Intriguingly, treatment with forskolin alone slightly but significantly reduced cell surface area (Figure 4c,d).
These results suggested that activation of PKA antagonized α1-adrenergic receptor-mediated cardiomyocyte hypertrophy.

| Silencing of PRKAR1A resulted in mitochondrial elongation
Initiation of cardiomyocyte hypertrophy requires mitochondrial fission, which generates fragmented mitochondria (Pennanen et al., 2014). Therefore, we examined mitochondrial morphology in PRKAR1A-depleted H9c2 myoblasts. Knockdown of PRKAR1A dramatically reduced the number of cells with fragmented mitochondria, and robustly increased those with elongated mitochondria (Figure 5a,b). Moreover, mitochondrial length was significantly increased in PRKAR1A-depleted cells (Figure 5c).

PKA-dependent phosphorylation and inactivation of the mitochondrial fission protein Drp1
Mitochondria are highly dynamic organelles that undergo continuous fusion and fission, and the balance between these two opposing processes determines the final mitochondrial morphology. In particular, mitochondrial fission is driven by dynamin-related protein 1 (Drp1), a dynamin-like GTPase that is recruited to mitochondria upon activation (Fonseca, Sanchez-Guerrero, Milosevic, & Raimundo, 2019).

| DISCUSSION
Cardiac complications account for more than half of all deaths in patients with CNC, a familial multiple neoplasia syndrome caused by mutations of PRKAR1A (Correa et al., 2015). However, the impact of CNC on heart development remains unknown. In this study, we demonstrated that left ventricular mass was reduced in young CNC patients. Our animal studies further revealed that cardiac-specific ablation of PRKAR1A impeded myocardial hypertrophic growth during development. Disruption of PRKAR1A provoked PKA-dependent inhibition of the mitochondrial fission protein Drp1, a critical molecule necessary for initiation of cardiomyocyte hypertrophy (Figure 7c). Our results uncovered the effect of the naturally occurring PRKAR1A haploinsufficiency on heart morphology for the first time, and provided potential mechanistic explanations.
The reduction in heart mass in our cPRKAR1A +/mice was associated with a decrease in cardiomyocyte cross-sectional area, suggesting that deletion of PRKAR1A hampered cardiomyocyte hypertrophy. It has been shown that cardiac-specific homozygous ablation of PRKAR1A induces thinning of the ventricular walls at embryonic day 11.5, due to a significant reduction in cardiomyocyte proliferation (Yin et al., 2008). As the Mlc2v-Cre +/line used in our study induces significant recombination as early as embryonic day 8.3 (Davis, Maillet, Miano, & Molkentin, 2012), the decreased heart weight in adult cPRKAR1A +/mice may also be attributed to diminished cardiomyocyte proliferation during early embryonic development. Interestingly, global PRKAR1A null embryos are smaller in size (Amieux et al., 2002), suggesting that PRKAR1A is likely universally required for growth of various tissues.
Accumulating evidence indicates that PKA catalytic activity is necessary for the phenotypes caused by PRKAR1A deficiency. In response to cAMP stimulation, tumors with PRKAR1A mutations collected from CNC patients exhibit higher total PKA activity than tumors from non-CNC patients (Kirschner et al., 2000). Moreover, constitutive F I G U R E 5 Silencing of PRKAR1A resulted in mitochondrial elongation. (a) H9c2 myoblasts were transfected with control (siGFP) or PRKAR1A siRNA (siPRKAR1A), followed by staining with MitoTracker red to label mitochondria (Red) and DAPI to label nuclei (blue). Boxed areas are enlarged on the left. Scale bar = 20µm; (b) Percentage of H9c2 cells with fragmented or elongated mitochondria. At least 500 cells were counted (n = 3). Two-way ANOVA with Sidak test. (c) Mitochondrial length. At least 15 mitochondrial particles of 40 cells were measured (n = 3). Two-tailed Student's t test. ** p < .01, *** p < .001 PKA activation has been detected in cardiac myxoma, a characteristic manifestation of CNC (Tseng et al., 2017). Homozygous deletion of PRKAR1A increases basal PKA activity, resulting in heart defects and embryonic lethality (Amieux et al., 2002;Yin et al., 2008). Interestingly, loss of PRKACA (encoding PKA catalytic subunit α) partially rescues the cardiac phenotype in global PRKAR1A null embryos (Amieux et al., 2002). Moreover, survival of the cardiac-specific homozygous PRKAR1A knockout mice is extended to 18 weeks by PRKACA haploinsufficiency, but is only prolonged by 1 day by haploinsufficiency of PRKACB (encoding PKA catalytic subunit β) (Yin, Pringle, Jones, Kelly, & Kirschner, 2011). These results suggest that PRKAR1A deficiency predominantly activates the PKA catalytic subunit α. In this study, PRKAR1A deficiency modestly increased phosphorylation of certain PKA substrates in adult hearts (Figure 2a). PKA activity might be higher during early development of the cPRKAR1A +/hearts, as previously shown for the homozygous PRKAR1A knockout heart (Amieux et al., 2002;Yin et al., 2008). Indeed, prolonged PKA activation in adult mice can cause degradation of the PKA catalytic subunit, resulting in decreased total PKA activity (Hemmings, 1986). Moreover, PKA has been shown to activate phosphodiesterases, which mediate cAMP degradation and eventually restricts PKA activity (Liu et al., 2012).
Our in vivo and in vitro studies revealed that depletion of PRKAR1A attenuated cardiomyocyte hypertrophy, likely due to PKA activation. Indeed, PKA is known to inhibit cardiomyocyte hypertrophy through phosphorylation of histone deacetylase 4 (HDAC4) and HDAC5, which then accumulate in the nuclei and repress myocyte enhancer factor 2 (MEF2)-dependent transcription of genes involved in cell differentiation and growth (Backs et al., 2011;Ha et al., 2010). Particularly, adrenergic stimulation-induced expression of fetal genes, such as the classical hypertrophy markers atrial natriuretic peptide (ANP) and β-myosin heavy chain (β-MHC), is augmented by the PKA inhibitor F I G U R E 6 Knockdown of PRKAR1A provoked PKA-dependent phosphorylation and inactivation of the mitochondrial fission protein Drp1. (a,b) NRCMs were transfected with control (siGFP) or PRKAR1A siRNA (siPRKAR1A) for 48h, then treated with PKA inhibitor (H89, 5µM) or vehicle for 4h. Cell lysates were immunoblotted using indicated antibodies with β-actin as a loading control. (a) Representative images; (b) Quantitation of phospho-Drp1 levels. Two-way ANOVA with Sidak test. * p < .05; (c) NRCMs transfected with siGFP or siPRKAR1A were subjected to immunofluorescent staining for Drp1 (green) and the mitochondrial marker COX IV (red). Mitochondrial and cytosolic localized Drp1 appears yellow and green, respectively, in the merged image. Scale bar = 10µm. (d) Manders' overlap coefficients for COX IV and Drp1. Light gray bars indicate COX IV-associated Drp1, and dark gray bars indicate Drp1-associated COX IV. At least 45 cells were evaluated (n = 4). Twotailed Student's t test. *** p < .001 H89, but is diminished by the PKA activators forskolin or 8-CPT-6-Phe-cAMP (Patrizio et al., 2008). Deficiency of phosphodiesterase 1C (PDE1C), a major phosphodiesterase that hydrolyzes cAMP in human myocardium, attenuates pathological cardiomyocyte hypertrophy in vitro and in vivo, in a PKA-dependent manner (Knight et al., 2016). Conversely, inhibition of PKA with PKI restores hyperglycemia-induced hypertrophy (Cheng et al., 2019). It is noteworthy that persistent PKA activation in a cardiac-specific PKA catalytic subunit α transgenic mice results in cardiomyocyte hypertrophy and ventricular dilation (Antos et al., 2001). The discrepancy may be due to chronic exposure of the transgenic hearts to supraphysiological levels of PKA activity (~2.4-to 8-fold increase vs. controls) (Antos et al., 2001). In cardiomyocytes, cAMP and PKA activity are not uniformly distributed but are confined primarily within the plasma membrane, sarcoplasmic reticulum, and myofilament compartments (Ghigo & Mika, 2019). Different extracellular stimuli induce differential activation of PKA within specific compartments (Liu et al., 2012). Therefore, aggravated hypertrophy of the PKA transgenic hearts may also be caused by nonselective activation of PKA in additional subcellular locales.
Our clinical findings have limitations. Firstly, left ventricular mass in CNC patients was compared with pediatric normal references, which can be variable depending on the methodology used (Buechel, Kaiser, Jackson, Schmitz, & Kellenberger, 2009;Ven et al., 2020). An alternative control could potentially be age-and gender-matched healthy siblings of the CNC patients. Secondly, our sample size is relatively small partly because CNC clinical cases are rare. These findings in patients need to be confirmed by larger clinical studies.
In conclusion, this study reveals that PRKAR1A deficiency delays heart growth during development. Mechanistically, PRKAR1A deficiency inhibits cardiomyocyte hypertrophy, likely through PKA-dependent inhibition of Drp1. As PRKAR1A haploinsufficiency in humans causes CNC, our study may provide novel mechanistic insights regarding the cardiac mortality associated with this hereditary genetic disorder.