Cyclic nucleotide signalling compartmentation by PDEs in cultured vascular smooth muscle cells

Abstract

Background and Purpose

Up‐regulation of phosphodiesterases (PDEs) is associated with several vascular diseases, and better understanding of the roles of each PDE isoform in controlling subcellular pools of cyclic nucleotides in vascular cells is needed. We investigated the respective role of PDE1, PDE5, and PDE9 in controlling intracellular cAMP and/or cGMP concentrations ([cAMP]_i, [cGMP]_i) in cultured rat aortic smooth muscle cells (RASMCs).

Experimental Approach

We used selective inhibitors of PDE1 (PF‐04471141), PDE5 (sildenafil), and PDE9 (PF‐04447943) to measure cAMP‐ and cGMP‐PDE activities with a radioenzymatic assay, in RASMC extracts. Real‐time [cAMP]_i and [cGMP]_i were recorded by Förster resonance energy transfer‐imaging in single living cells, and cell proliferation was assessed in FBS‐stimulated cells.

Key Results

PDE1, PDE5, and PDE9 represented the major cGMP‐hydrolyzing activity in RASMCs. Basal PDE1 exerted a functional role in degrading in situ the cGMP produced in response to activation of particulate GC by C‐type natriuretic peptide. In high intracellular Ca²⁺ concentrations, PDE1 also regulated the NO/soluble GC‐dependent cGMP response, as well as the β‐adrenoceptor‐mediated cAMP response. PDE5 exerted a major role in degrading cGMP produced by NO and the natriuretic peptides. PDE9 only regulated the NO‐induced [cGMP]_i increase. All three PDEs contributed differently to regulate cell proliferation under basal conditions and upon cGMP‐elevating stimuli.

Conclusions and Implications

Our data emphasize the distinct roles of PDE1, PDE5, and PDE9 in local regulation of [cAMP]_i and [cGMP]_i, in vascular smooth muscle cells, strengthening the concept of PDEs as key actors in the subcellular compartmentation of cyclic nucleotides.

What is already known

• In vascular smooth muscle cells, membrane NP receptors and soluble GC produce cGMP.

• PDE5 is described as the main cGMP hydrolyzing enzyme.

What this study adds

• PDE5 exerts a major role in degrading cGMP produced by ANP, CNP or NO.

• PDE1 controls cGMP produced by CNP, and by NO, whereas PDE9 regulates only NO‐induced cGMP.

What is the clinical significance

• Elevating cGMP concentrations represents a vasoprotective effect.

• The functional specificity of each PDE family might help identify a relevant drug target candidate.

Abbreviations

ANP

atrial natriuretic peptide

CaM

calmodulin

CFP

cyan fluorescent protein

CNP

C‐type natriuretic peptide

DEA/NO

diethylamine NONOate

FRET

Förster resonance energy transfer

MIMX

8‐methoxymethyl‐3‐isobutyl‐1‐methylxanthine

MOI

multiplicity of infection

NP

natriuretic peptide

NPR

natriuretic peptide receptor

pGC

particulate GC

RASMCs

rat aortic smooth muscle cells

sGC

soluble GC

SMC

smooth muscle cell

VSMCs

vascular smooth muscle cells

YFP

yellow fluorescent protein

1. INTRODUCTION

The cyclic nucleotides, cAMP and cGMP, are important second messengers that regulate a myriad of cellular functions. Synthesis of cAMP is catalysed by AC, a family of nine transmembrane enzymes (AC1‐9), regulated by GPCRs, and one soluble enzyme (AC10). Cyclic GMP is generated by two groups of GC, the particulate GCs (pGCs), a subfamily of catalytic receptors represented by the natriuretic peptide (NP) receptors (NPR), NPR‐A and NPR‐B, typically stimulated by atrial natriuretic peptide (ANP) and C‐type natriuretic peptide (CNP) respectively, and the soluble GC (sGC) which is directly activated by NO. Degradation of cyclic nucleotides by PDEs represents a major means to rapidly lower their cellular content and control the amplitude and duration of their responses. The 11 PDE families can be grouped into three classes according to their selectivity: the cAMP‐specific (PDE4, 7, and 8), the cGMP‐specific (PDE5, 6, and 9), and the dual hydrolytic (PDE1, 2, 3, 10, and 11) PDEs (Bender & Beavo, 2006; Keravis & Lugnier, 2012). It is now well‐accepted that PDEs participate in the subcellular compartmentation of cyclic nucleotide signalling (Conti & Beavo, 2007). This has been particularly described in cardiac myocytes where the different PDEs are expressed in distinct subcellular microdomains to prevent the diffusion of intracellular cAMP/cGMP to the entire cell (Castro, Verde, Cooper, & Fischmeister, 2006; Leroy et al., 2008; Mika, Leroy, Vandecasteele, & Fischmeister, 2012; Mongillo et al., 2004) but less investigated in vascular smooth muscle cells (VSMCs; Berisha & Nikolaev, 2017). However, there is clear evidence that cGMP signals elicited either by NO or NPs exhibit different spatiotemporal dynamics within VSMCs (Nausch, Ledoux, Bonev, Nelson, & Dostmann, 2008; Piggott et al., 2006).

PDE1, 3, 4, and 5 are described as the main functional PDE families in VSMCs (Lugnier, Schoeffter, Le Bec, Strouthou, & Stoclet, 1986), with a pattern of expression depending on the cell phenotype. Indeed, VSMCs undergo a phenotype switch from contractile to synthetic phenotype in vivo during pathological remodelling of the vascular wall (Rensen, Doevendans, & van Eys, 2007), as well as in vitro during cell culture (Belacel‐Ouari et al., 2017). PDE3 is the main cAMP‐PDE in the contractile phenotype (Hubert et al., 2014), but its activity is decreased in the synthetic phenotype, making PDE4 the predominant cAMP‐PDE family in cultured VSMCs (Dunkerley et al., 2002; Zhai et al., 2012). PDE5 is the most active cGMP‐PDE in basal condition in VSMCs (Rybalkin, Yan, Bornfeldt, & Beavo, 2003). However, under high Ca²⁺ conditions, the Ca²⁺/calmodulin (CaM)‐dependent PDE1 can become predominant. PDE1 expression is largely associated to the VSMC phenotype, with an up‐regulation of the PDE1C isoform in synthetic VSMCs contributing to their proliferative/migrative status (Cai et al., 2015; Rybalkin, Rybalkina, Beavo, & Bornfeldt, 2002). However, the characterization of the functional role of PDE1 in controlling local intracellular cyclic nucleotide signals has been hampered by the lack of truly selective PDE1 inhibitors. On the other hand, by using the well‐described PDE5 inhibitor, sildenafil, it was proposed that PDE5 controls both pGC‐ and sGC‐dependent cGMP pools in VSMCs isolated from rat or mouse aorta (Krawutschke, Koesling, & Russwurm, 2015; Nausch et al., 2008). However, high concentrations of sildenafil were used in these two studies (≥1 μM) which raises the possibility that other PDEs, such as PDE1 or PDE3, might also be inhibited.

Thus, the aim of the present study was to investigate the role of PDE1 in cyclic nucleotide compartmentation in synthetic VSMCs. We took advantage of a new quinazoline‐based compound, PF‐04471141 (named PF1), described as a potent inhibitor of the PDE1 family (with IC₅₀ of 35–118 nM for human PDE1 subtypes) with more than 30‐ to 100‐fold selectivity over other PDE families (Humphrey et al., 2014). To monitor intracellular cyclic nucleotide concentrations, fluorescence live cell imaging was conducted in cultured rat aortic SMCs (RASMCs) using Förster resonance energy transfer (FRET)‐based sensors, an approach that has been widely used to evaluate the real‐time dynamics of cAMP‐ or cGMP‐dependent signals (Sprenger & Nikolaev, 2013). The role of PDE1 in controlling NO‐ and NP‐induced cGMP pools was compared to that of PDE5 and PDE9, a high affinity cGMP‐PDE detected in rat aorta at the mRNA level (Phillips, Long, Wilkins, & Morrell, 2005). To date, no functional studies have been reported for PDE9 in VSMCs. Here, we tested PF‐04447943 (named PF9), a compound developed as a brain permeant PDE9 inhibitor, which exhibits high affinity for rat PDE9 (K_i of 18 nM) and high selectivity for PDE9 over other PDE families, in particular over PDE1 (475‐fold) and PDE5 (830‐fold; Hutson et al., 2011).

2. METHODS

2.1. Animals

All animal care and experimental procedures conformed to the European Community guiding principles in the Care and Use of Animals (Directive 2010/63/EU of the European Parliament), the local Ethics Committee (CREEA Ile‐de‐France Sud) guidelines, and the French decree no. 2013‐118 of February 1, 2013 on the protection of animals used for scientific purposes (JORF n^o. 0032, February 7, 2013, p2199, text no. 24). Authorizations to perform animal experiments according to this decree were obtained from the French Ministry of Agriculture, Fisheries and Food (No. D‐92–283, December 13, 2012). Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010; McGrath, Drummond, McLachlan, Kilkenny, & Wainwright, 2010) and with the recommendations made by the British Journal of Pharmacology. Adult male Wistar rats (180–200 g) were purchased from Janvier Labs (Le Genest St Isle, France) and housed for a short time in our authorized animal facility (authorization N° C 92‐019‐01 valid until August 3, 2022) under standard animal housing and care conditions.

2.2. Cell isolation and culture

RASMCs were isolated as previously described (Zhai et al., 2012). In brief, rats were anaesthetized by intraperitoneal injection of pentobarbital (0.1 mg·g⁻¹; Doléthal®, Laboratoire Vétoquinol, Lure, France) and subjected to thoracotomy to isolate the aorta. The medial layer of the thoracic aorta was digested with collagenase type 2 (60 U·ml⁻¹; Worthington Biochemical Corporation, Lakewood, NJ, USA) and elastase (0.3 mg·ml⁻¹; MP Biomedicals, Solon, Ohio, USA). Harvested cells were suspended in DMEM (GIBCO, Invitrogen, Cergy Pontoise, France) containing antibiotics/antimycotic (100 U·ml⁻¹ penicillin, 100 mg·ml⁻¹ streptomycin, 0.25 mg·ml⁻¹ amphotericin B; GIBCO) and supplemented with 20% FBS (GIBCO) and then were seeded in flasks coated with collagen I (rat tail, Corning, Amsterdam, Netherlands). The medium was replaced with DMEM containing 10% FBS after 48 hr. All experiments were performed on cells at passages 2 to 6. For PDE activity assay or FRET experiments, cells were plated at a density of 3 × 10⁴ cells·cm⁻² on collagen‐coated Petri dishes or glass coverslips respectively. For cell proliferation assay, cells were plated in collagen‐coated 96‐well plates at the density of 5,000 cells per well.

2.3. cAMP‐ and cGMP‐PDE activity assays

Cells, maintained in DMEM containing 10% FBS for 48 hr after plating, were detached with 0.05% trypsin containing 0.53 mM EDTA (GIBCO), homogenized using a tissue homogenizer (Bertin Technologies) in ice‐cold lysis buffer (containing: NaCl 150 mM, HEPES 20 mM, EDTA 2 mM, NP40 0.5%, and protease inhibitor cocktails) and centrifuged at 12,000× g for 10 min at 4°C. Protein concentration of the supernatant was determined using the bicinchoninic acid protein assay, according to the manufacturer's protocol (Pierce, Thermo Fisher Scientific, Brebières, France). cAMP‐ or cGMP‐PDE activity was measured using 25 or 15 μg of RASMC proteins in the presence of 1 μM cAMP or cGMP, respectively, according to the method described by Thompson and Appleman (1971), as previously reported (Zhai et al., 2012). This method uses radio‐labelled substrates [³H]‐cAMP or [³H]‐cGMP. To evaluate the activity of specific PDE families, the assay was performed in the absence (0.05% DMSO as vehicle) or presence of PDE inhibitors: 1 μM PF1 for PDE1, 200 nM sildenafil for PDE5, 1 and 5 μM PF9 for PDE9, and 500 μM IBMX as a non‐selective PDE inhibitor. To evaluate the Ca²⁺‐dependent PDE1 activity, some experiments were performed in the presence of 800 μM CaCl₂ plus 240 nM calmodulin (CaM) or 1 mM EGTA (Kim et al., 2001). The hydrolytic activity measured in these different experimental conditions was expressed as a percentage of the total cAMP‐ or cGMP‐PDE activity defined in the absence of any drug (vehicle).

2.4. FRET imaging

We used two FRET‐based sensors for cAMP, Epac2‐camps (containing the single cAMP‐binding domain of Epac2 protein flanked by an enhanced cyan fluorescent protein [CFP] as donor and an enhanced yellow fluorescent protein [YFP] as acceptor; Nikolaev, Bunemann, Hein, Hannawacker, & Lohse, 2004) and Epac‐S^H187 (containing part of cAMP‐binding protein Epac1 sandwiched between the donor mTurquoise2 and a tandem of circular permutated Venus as acceptor; Klarenbeek, Goedhart, van Batenburg, Groenewald, & Jalink, 2015; generous gift from Dr Kees Jalink) which exhibits a higher dynamic range and signal to noise ratio, and one FRET‐based sensor for cGMP, cGi‐500 (containing the tandem cGMP‐binding domains of PKGI protein flanked by an enhanced CFP as donor and an enhanced YFP as acceptor; Russwurm et al., 2007; generous gift from Dr Michael Russwurm). When cAMP or cGMP binds to its specific binding domain, a reversible conformational change of the sensor occurs, leading to a decrease in FRET between the donor and acceptor fluorophores, resulting in an increase in the ratio of donor/acceptor emitted fluorescence intensity. This ratio will be called “CFP/YFP,” as the three sensors used in this study consist of variants of CFP and YFP, as donor and acceptor respectively.

RASMCs were infected with an adenovirus encoding Epac2‐camps (multiplicity of infection [MOI] 500 pfu per cell), Epac‐S^H187 (MOI 200 pfu per cell), or cGi‐500 (MOI 50–60 pfu per cell) in DMEM containing 10% FBS. FRET experiments were performed 48 hr after infection. Cells were maintained in a Ringer solution during the imaging capture, as previously described (Zhai et al., 2012). All pharmacological agents were diluted in Ringer solution. The cell of interest was continuously and locally perfused with Ringer solutions with or without these agents, using a microperfusion system allowing rapid applications of these solutions (Zhai et al., 2012).

For cAMP measurements using Epac2‐camps sensor, the β‐adrenoceptor agonist isoprenaline (10 or 100 nM) was applied as a 15 s pulse. To evaluate the effect of PDE1 inhibition on the dynamics of isoprenaline response, PF1 was applied 4 min before the pulse of isoprenaline and maintained 10 min after. In some experiments, 400 nM angiotensin II (Ang II) was applied 1 min before the isoprenaline pulse to transiently increase intracellular Ca²⁺ concentration to stimulate PDE1 activity (Alvarez, Campos‐Toimil, Justiniano‐Basaran, Lugnier, & Orallo, 2006; Kim et al., 2001). For cAMP measurements using Epac‐S^H187 sensor, the application of isoprenaline (10 nM) was maintained. Ang II was then added on top of isoprenaline followed 5 min later by PF1.

For cGMP measurements, cumulative concentrations of diethylamine NONOate (DEA/NO; 1, 3, and 10 μM), ANP (0.01 and 0.1 μM), or CNP (0.01 and 0.1 μM) were tested in the absence (0.05% DMSO as vehicle) or presence of 1 μM PF1, 200 nM sildenafil, or 5 μM PF9. In some experiments, Ang II was added simultaneously to the last concentration of DEA/NO, ANP, or CNP to evaluate the contribution of PDE1 in high Ca²⁺ conditions.

FRET imaging was captured every 5 s by an inverted epifluorescence microscope connected to a camera, and average fluorescence intensities (emitted by CFP and YFP) were obtained from a region of interest corresponding to the entire cell, as previously described (Zhai et al., 2012). Data were expressed as a percentage of the CFP/YFP ratio measured just before application of the cAMP/cGMP stimulating drugs (isoprenaline, DEA/NO, ANP, or CNP, depending on the experiments). Kinetic parameters (T_peak: time to peak; T_1/2on: time to half‐peak; and τ: time from the peak to obtain 63% recovery) were determined using Microsoft® Office Excel software. τ values were determined by fitting the decrease phase of signals with the following single exponential equation: y = A*exp(−t/τdecay) + B (Leroy et al., 2008).

2.5. Cell proliferation assay

Cells were maintained in DMEM containing 10% FBS overnight and exposed to serum‐free medium (0.1% FBS) for 24 hr. Cells were then either maintained in the serum‐free medium or stimulated with 10% FBS in the absence (0.05% DMSO as vehicle) or presence of the PDE inhibitors (1 μM PF1, 200 nM sildenafil, or 5 μM PF9) and/or the cGMP agents at three different concentrations (0.1 to 10 μM SNAP; 1  to 100 nM ANP; or 1 to 100 nM CNP). In these experiments, the NO donor SNAP was preferred to DEA/NO according to its longer t _1/2 in cell culture conditions (He & Frost, 2016). Pharmacological agents were renewed after 24‐hr incubation. After 48 hr, cells were washed with PBS and cell proliferation was assessed using the water soluble tetrazolium based assay, according to the manufacturer's instructions (Roche, Germany). Briefly, cells were incubated for 4 hr with water soluble tetrazolium and absorbance was quantified with a spectrophotometer at 450 nm (after a correction at 600 nm). Data were expressed in % of the increase in the absorbance (thereafter referred to as “proliferation”) induced by 10% FBS.

2.6. Randomization and blinding

Animal were not randomized as they all were used for cell preparation, but protein extracts or cells were assigned randomly to the different treatments. Experiments and data analyses were not performed under blind conditions. However, raw data were acquired directly from the experimental procedures and then analysed through a standardized procedure that reduces any possible operator bias.

2.7. Data analysis and statistics

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology. All results were expressed as mean ± SEM of N experiments in cAMP/cGMP‐PDE activity and cell proliferation assays, or n cells in FRET imaging experiments. For PDE activity and cell proliferation assays, experiments were undertaken in duplicate to ensure the reliability of single values, but data analysis and data presentation used the mean of duplicate values. For each condition, control and treated cells were studied the same day and experiments were repeated on several days. Data normalization was undertaken to control for sources of variation of baseline parameters in each cell or assay and to allow comparison of the magnitude of drug effects in different conditions. Statistical comparisons of the different parameters were performed using Microsoft® Office Excel software for Student's t‐test or Prism 7 software (Graphpad Software, La Jolla, CA, USA) for ANOVA. When relevant, ANOVA was followed by a Dunnett's multiple comparison test (one‐way ANOVA) or Tukey's multiple comparison test (two‐way ANOVA). P value <0.05 was considered for statistical significance.

2.8. Materials

Ang II, ANP, CNP, IBMX, (−)‐isoprenaline hydrochloride, CaM, EGTA, SNAP, and sildenafil citrate were purchased from Sigma‐Aldrich (St‐Quentin‐Fallavier, France). [³H]‐cGMP (14.3 Ci·mmol⁻¹) and [³H]‐cAMP (31.3 Ci·mmol⁻¹) were supplied by PerkinElmer (Courtaboeuf, France). DEA NONOate (DEA/NO) was purchased from Cayman Chemical (Bertin Pharma, Montigny‐le‐Bretonneux, France). PF‐04471141 (PF1) and PF‐04447943 (PF9) were generous gifts from Pfizer (USA) obtained through Pfizer Compound Transfer Agreements. As all PDE inhibitor stock solutions were prepared in DMSO (Sigma‐Aldrich), control experiments were performed in the presence of equivalent concentrations of DMSO.

2.9. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander, Christopoulos et al., 2017; Alexander, Fabbro et al., 2017a; Alexander, Fabbro et al., 2017b).

3. RESULTS

3.1. Role of PDE1 in the control of cytosolic cAMP concentrations in cultured RASMCs

The contribution of PDE1 to the hydrolysis of cAMP in lysates from cultured RASMCs was evaluated using a radioenzymatic assay. Total cAMP‐PDE activity was 33.4 ± 1.6 pmol·min⁻¹·mg⁻¹ protein (N = 5) and was inhibited by the non‐selective PDE inhibitor IBMX (500 μM) by 95% (Figure 1). The PDE1 inhibitor PF1 (1 μM) only slightly reduced total cAMP‐PDE activity in the absence or presence of EGTA. However, in the presence of Ca²⁺ and CaM, total cAMP‐PDE activity was significantly increased by 20% and PF1 abolished this Ca²⁺/CaM‐stimulated cAMP‐PDE activity. This indicates that cAMP‐PDE1 activity is low in basal conditions but increased by Ca²⁺/CaM in cultured RASMCs.

Figure 1.

cAMP‐PDE1 activity in RASMCs. Total cAMP‐PDE activity was determined in the presence of 1 μM cAMP in lysates of cultured RASMCs in the absence (Ctrl) or presence of a PDE1 inhibitor (1 μM PF1), 800 μM Ca²⁺ + 240 nM CaM (Ca²⁺/CaM), a Ca²⁺ chelator (1 mM EGTA), a combination of PF1 with Ca²⁺/CaM or EGTA, or a non‐selective PDE inhibitor (500 μM IBMX). Results are expressed as percentage of the total cAMP‐PDE activity measured in Ctrl. Data are means ± SEM of five independent experiments. *P < 0.05 significantly different from Ctrl; ^# P < 0.05, significantly different as indicated; one‐way ANOVA followed by Dunnett's multiple comparison test

We then characterized the functional contribution of PDE1 in controlling the cAMP dynamics in response to stimulation with a β‐adrenoceptor agonist, by measuring changes in [cAMP]_i using the Epac2‐camps FRET sensor. A short application of the β‐adrenoceptor agonist isoprenaline (100 nM, 15 s) induced a transient increase in CFP/YFP ratio in RASMCs, reflecting a transient increase in [cAMP]_i (Figure 2a). Pre‐incubation of the cell with 1 μM PF1 did not modify this isoprenaline‐induced FRET signal (Figure 2a). The response to a lower concentration of isoprenaline (10 nM) was also insensitive to PF1 (data not shown). PF1 had also no additional effect on the response induced by isoprenaline in the presence of a PDE4 inhibitor (Supporting Information Figure S1). To physiologically activate PDE1 through an increase in intracellular Ca²⁺ concentration, we used Ang II (Alvarez, Campos‐Toimil, Justiniano‐Basaran, Lugnier, & Orallo, 2006; Kim et al., 2001). Application of 400 nM Ang II during 1 min before the 100 nM isoprenaline pulse significantly shortened the recovery phase of the response to isoprenaline (τ value decreased by 34%; Figure 2b). In the presence of PF1, this accelerating effect of Ang II was suppressed (Figure 2b). In another set of experiments, we took advantage of a recently developed generation of FRET‐based cAMP sensor, Epac‐S^H187, displaying high affinity for cAMP and higher FRET efficiency (Klarenbeek, Goedhart, van Batenburg, Groenewald, & Jalink, 2015). As shown in Figure 2c, 10 nM isoprenaline strongly increased CFP/YFP ratio by about 300%. Ang II applied on top of this isoprenaline response rapidly lowered this ratio by about one third and this Ang II‐mediated effect was largely reversed by 1 μM PF1 (Figure 2c). These data indicate that in RASMCs, PDE1 hydrolyses [cAMP]_i generated by stimulation of β‐adrenoceptors only when intracellular Ca²⁺ concentration is raised.

Figure 2.

Effect of PDE1 inhibition on β‐adrenoceptor‐induced cAMP signalling in RASMCs. (a, b) cAMP measurements were conducted in cultured RASMCs using the Epac2‐camps FRET‐based sensor, in response to a short application of isoprenaline (Iso, 100 nM, 15 s) in the absence (Ctrl; a and b) or presence of the PDE1 inhibitor (1 μM PF1, applied 4 min before isoprenaline; a), or after a 1‐min pretreatment with angiotensin II (Ang II, 400 nM; b) in the absence or presence of the PDE1 inhibitor. Upper and lower panels represent the percentage increase in CFP/YFP ratio and the corresponding kinetic parameters, respectively. Data are means ± SEM of n = 5–12 cells, as indicated in the figure. *P < 0.05, significantly different from Ctrl; ^# P < 0.05, significantly different as indicated; one‐way ANOVA followed by Dunnett's multiple comparison test. (c) cAMP measurements were conducted in cultured RASMCs using the Epac‐S^H187 FRET‐based sensor, in response to a prolonged application of isoprenaline (Iso, 10 nM). At steady state, 400 nM Ang II was applied alone for 5 min, and then in combination with 1 μM PF1. Data are means ± SEM of 15 cells. The broken line represents the mean of the continuous effect of isoprenaline maintained alone in the absence of Ang II or PF1, measured in four independent cells. Results are expressed as the percentage increase in CFP/YFP ratio

3.2. Role of PDE1 in the control of cytosolic cGMP concentrations in cultured RASMCs

Total cGMP‐PDE activity measured in lysates from cultured RASMCs was 51.1 ± 4.6 pmol·min⁻¹·mg⁻¹ protein (N = 10) and was reduced by IBMX by 94% (Figure 3). The total cGMP‐PDE activity was significantly reduced in the presence of 1 μM PF1 (by 28%) and increased in the presence of Ca²⁺/CaM (by 45%; Figure 3). PF1 strongly decreased the cGMP‐PDE activity in the presence of Ca²⁺/CaM by about twofold, to a similar level as in the presence of PF1 alone (Figure 3). EGTA alone or in combination with PF1 decreased the cGMP‐PDE activity by 28% and 44%, respectively (Figure 3).

Figure 3.

cGMP‐PDE activity in RASMCs. Cyclic GMP‐PDE activity was determined in the presence of 1 μM cGMP in lysates of cultured RASMCs in the absence (Ctrl) or presence of a PDE1 inhibitor (1 μM PF1), 800 μM Ca²⁺ + 240 nM CaM (Ca²⁺/CaM), a Ca²⁺ chelator (1 mM EGTA), a combination of PF1 with Ca²⁺/CaM or EGTA, a PDE5 inhibitor (200 nM sildenafil), a PDE9 inhibitor (PF9 at 1 or 5 μM), or a non‐selective PDE inhibitor (500 μM IBMX). Results are expressed as percentage of the total cGMP‐PDE activity measured in Ctrl. Data are means ± SEM of six independent experiments except for sildenafil condition (N = 4; statistical analysis was not undertaken on this group). *P < 0.05, significantly different from Ctrl; ^# P < 0.05, significantly different as indicated; one‐way ANOVA followed by Dunnett's multiple comparison test

We then evaluated the functional role of PDE1 in controlling [cGMP]_i by using the FRET‐based cGMP sensor cGi‐500. In cultured RASMCs, increasing concentrations of the NO donor, DEA/NO (1, 3, and 10 μM), elicited a concentration‐dependent increase in CFP/YFP, reflecting the raise in cGMP production by sGC (Figure 4a); 1 μM PF1 did not modify the amplitude of DEA/NO‐induced cGMP responses nor the recovery phase (Figure 4a). The NPs, ANP and CNP, induced a low FRET response at 0.01 μM but markedly increased the FRET ratio at 0.1 μM (Figure 4b–c). PF1 had no effect on the dynamics of ANP‐induced cGMP responses (Figure 4b) but markedly increased the amplitude of the CNP (0.1 μM)‐mediated response (by twofold), without altering the kinetics of its recovery phase (Figure 4c).

Figure 4.

Effect of PDE1 inhibition on soluble or particulate GC‐induced cGMP signalling in RASMCs. Cyclic GMP measurements were conducted in cultured RASMCs using the cGi‐500 FRET‐based sensor, in response to increasing concentrations of an NO donor (DEA/NO; 1, 3, and 10 μM; a), ANP (0.01 and 0.1 μM; b), or CNP (0.01 and 0.1 μM; c), applied as indicated in the figure, in the absence (Ctrl) or presence of a PDE1 inhibitor (1 μM PF1, maintained throughout the experiment). Upper and lower panels represent the percentage increase in CFP/YFP ratio and the corresponding kinetic parameter of the recovery phase, respectively. Data are means ± SEM of 14–25 independent cells as indicated in the figure. ^§ P < 0.05, significantly different as indicated; ANOVA for repeated measures

To examine the role of PDE1 in high [Ca²⁺]_i conditions, 400 nM Ang II was added simultaneously to the higher concentration of DEA/NO or NPs (Figure 5). Ang II transiently decreased the FRET signal to 10 μM DEA/NO, delayed the onset of its response and significantly accelerated its recovery phase (τ value decreased by 48% by Ang II compared to Ctrl condition). In the presence of Ang II, cell treatment with 1 μM PF1 suppressed the transient decrease of cGMP signals induced by Ang II and significantly delayed the recovery phase of the DEA/NO‐mediated cGMP response (τ value increased by 39% in the presence of PF1; Figure 5a). Ang II did not significantly affect the dynamics of the cGMP responses induced by ANP or CNP. PF1, in the presence of Ang II, did not alter the dynamics of the response to 0.1 μM ANP (Figure 5b) but significantly increased the maximum response to 0.1 μM CNP (Figure 5c). However, this potentiating effect of PF1 was similar to that observed in the absence of Ang II (twofold increase in FRET signal amplitude in the presence or absence of Ang II, respectively; Figures 4c and 5c).

Figure 5.

Effect of PDE1 inhibition in the presence of Ang II on soluble or particulate GC‐induced cGMP signalling in RASMCs. Cyclic GMP measurements were conducted in cultured RASMCs using the cGi‐500 FRET‐based sensor, in response to increasing concentrations of an NO donor (DEA/NO; 1, 3, and 10 μM; a), or ANP (0.01 and 0.1 μM; b) or CNP (0.01 and 0.1 μM; c), applied as indicated in the figure, in the absence (Ctrl) or presence of a PDE1 inhibitor (1 μM PF1, maintained throughout the experiment). Angiotensin II (Ang II, 400 nM) was applied simultaneously to the last concentration of DEA/NO (a), ANP (b), or CNP (c) and maintained until the end of experiment. Upper and lower panels represent the percentage increase in CFP/YFP ratio and the corresponding kinetic parameter of the recovery phase, respectively. Data are means ± SEM of 19–25 independent cells as indicated in the figure. ^§ P < 0.05, significantly different as indicated; ANOVA for repeated measures. *P < 0.05, significantly different from Ctrl; Student's t‐test

3.3. Role of PDE5 in the control of cytosolic cGMP concentrations in cultured RASMCs

As shown in Figure 3, 200 nM sildenafil, a PDE5 inhibitor, reduced the total cGMP‐PDE activity by 55% in cultured RASMCs (statistical analysis was not undertaken on this data set as N = 4). In cGMP‐FRET imaging, PDE5 inhibition by sildenafil significantly potentiated the response to DEA/NO by increasing the maximum response to 10 μM DEA/NO (from 61% to 81% compared to Ctrl) and by markedly prolonging its recovery phase (τ value increased twofold; Figure 6a). Sildenafil dramatically elevated the peak of the cGMP response induced by 0.1 μM ANP, without altering the kinetic τ parameter (Figure 6b). Similarly, sildenafil strongly increased the amplitude of the cGMP signal elicited by 0.1 μM CNP, without altering its decline phase (Figure 6c). These data indicate a major contribution of PDE5 in controlling [cGMP]_i stimulated by various signals in cultured RASMCs.

Figure 6.

Effect of PDE5 inhibition on soluble or particulate GC‐induced cGMP signalling in RASMCs. Cyclic GMP measurements were conducted in cultured RASMCs using the cGi‐500 FRET‐based sensor, in response to increasing concentrations of an NO donor (DEA/NO; 1, 3, and 10 μM; a), ANP (0.01 and 0.1 μM; b), or CNP (0.01 and 0.1 μM; c), applied as indicated in the figure, in the absence (Ctrl) or presence of a PDE5 inhibitor (200 nM sildenafil, Sil; maintained throughout the experiment). Upper and lower panels represent the percentage increase in CFP/YFP ratio and the corresponding kinetic parameter of the recovery phase, respectively. Data are means ± SEM of 13–23 independent cells as indicated in the figure. ^§ P < 0.05, significantly different as indicated; ANOVA for repeated measures. *P < 0.05, significantly different from Ctrl; Student's t‐test

3.4. Role of PDE9 in the control of cytosolic cGMP concentrations in cultured RASMCs

The expression of PDE9 protein was evaluated by western blot, suggesting the presence of different PDE9 isoforms in cultured RASMCs (Supporting Information Figure S2), as reported in rodent brain tissue by Patel et al. (2018). The PDE9 inhibitor, PF9, did not significantly lower the total cGMP‐PDE activity in cultured RASMCs at 1 μM but strongly reduced it (by 37%) at 5 μM (Figure 3). Hence, the functional role of PDE9 in the control of [cGMP]_i was assessed using 5 μM PF9 in real‐time FRET experiments. As shown in Figure 7a, PF9 significantly modified the cGMP response to cumulative DEA/NO concentrations, by increasing its maximum amplitude (by twofold at 3 μM DEA/NO) and delaying its recovery phase (τ value increased by 35%). By contrast, 5 μM PF9 affected neither the ANP‐ nor the CNP‐elicited cGMP responses (Figure 7b–c). This suggests a specific role of PDE9 in controlling the sGC‐related pool of cGMP in cultured RASMCs.

Figure 7.

Effect of PDE9 inhibition on soluble or particulate GC‐induced cGMP signalling in RASMCs. Cyclic GMP measurements were conducted in cultured RASMCs using the cGi‐500 FRET‐based sensor, in response to increasing concentrations of an NO donor (DEA/NO; 1, 3, and 10 μM; a), ANP (0.01 and 0.1 μM; b), or CNP (0.01 and 0.1 μM; c), applied as indicated in the figure, in the absence (Ctrl) or presence of a PDE9 inhibitor (5 μM PF9; maintained throughout the experiment). Upper and lower panels represent the percentage increase in CFP/YFP ratio and the corresponding kinetic parameter of the recovery phase, respectively. Data are means ± SEM of 12–44 independent cells as indicated in the figure. ^§ P < 0.05, significantly different as indicated; ANOVA for repeated measures. *P < 0.05, significantly different from Ctrl; Student's t‐test

3.5. Role of PDE1, PDE5, and PDE9 in the control of RASMCs proliferation

We then evaluated the role of the cGMP/PDE pathway in regulating cell proliferation. A 48‐hr treatment of cultured RASMCs in the presence of 10% FBS significantly increased cell proliferation by about 4.1 ± 1.0‐fold compared to cells treated with 0.1% FBS (N = 10). Regarding the effect of PDE inhibitors, the FBS‐stimulated proliferation was significantly reduced by 4.8 ± 1.9% (N = 8) in the presence of 1 μM PF1 and by 7.9 ± 0.7% (N = 7) in the presence of 5 μM PF9, whereas 200 nM sildenafil had no significant effect. We then assessed cell proliferation in response to sGC or pGC activation. As shown in Figure 8a, the NO donor SNAP decreased the FBS‐stimulated cell proliferation, with a reduction of 15.0 ± 6.2% at 10 μM (N = 6). The effect of SNAP was not significantly potentiated by PDE inhibitors (Figure 8a). Increasing concentrations of ANP from 1 to 100 nM did not modify the FBS‐stimulated cell proliferation (Figure 8b). Cell treatment with PF1 or PF9 had no effect on ANP response. By contrast, sildenafil potentiated the global response to ANP, eliciting a decrease in cell proliferation by about 20% (N = 7; Figure 8b). As shown in Figure 8c, CNP appears to decrease the FBS‐stimulated cell proliferation only at the highest concentration tested (100 nM). None of the three PDE inhibitors tested significantly changed the global effect of CNP (Figure 8c).

Figure 8.

Effect of soluble or particulate GC activation on RASMC proliferation in the absence or presence of PDE1, PDE5, or PDE9 inhibition. RASMC proliferation was measured after a 48‐hr stimulation with 10% FBS, in the absence or presence of an NO donor (SNAP; 0.1, 1, and 10 μM; a), ANP (1, 10, and 100 nM; b), or CNP (1, 10, and 100 nM; c), in the absence (Ctrl) or presence of a PDE inhibitor (1 μM PF1, 200 nM sildenafil, Sil, or 5 μM PF9). Data are expressed in % of the cell proliferation induced by 10% FBS. Data are means ± SEM of 6 (a) or 7 (b and c) independent experiments. ^§ P < 0.05, significant effect of sildenafil (Sil); two‐way ANOVA, followed by Tukey's multiple comparison test

4. DISCUSSION

The major findings of the present series of experiments are that, under short‐term pharmacological modulation of cyclic nucleotide/PDE pathways in cultured RASMCs, we observed that (a) basal PDE1 activity has no influence on [cAMP]_i but selectively controls [cGMP]_i produced by CNP; (b) PDE1 is turned on by Ang II and then becomes active on the β‐adrenoceptor‐dependent [cAMP]_i pool as well as on the NO‐dependent [cGMP]_i pool; (c) PDE5 is promiscuous and controls [cGMP]_i independently of the nature of the activated GC; (d) the role of PDE9 is limited to the regulation of the NO‐elicited [cGMP]_i pool. Under long‐term modulation of the cGMP/PDE pathway, PDE1 and PDE9 inhibition induces small but significant reductions in the proliferative effect of FBS in the absence of exogenous cGMP manipulation, whereas PDE5 controls FBS‐stimulated proliferation upon cGMP stimuli.

Evaluation of the functional role of PDE1 in intact tissues was rather limited by the lack of potent and selective inhibitors. In cultured RASMCs, we previously observed that another non‐selective PDE inhibitor, 8‐methoxymethyl‐3‐isobutyl‐1‐methylxanthine (MIMX; 50 μM) had no effect on the β‐adrenoceptor‐induced increase in [cAMP]_i but potentiated the effect of a PDE4 inhibitor (Zhai et al., 2012). However, the role of PDE1 inhibition in this effect of MIMX could only be suspected given the high concentration used, possibly affecting other cAMP‐PDEs. Furthermore, MIMX is not a suitable agent to analyse cGMP signalling, as MIMX inhibits PDE5 and PDE1 in the same range of concentration (Gonçalves et al., 2009). Another common PDE1 inhibitor, vinpocetine, used at a concentration around its IC₅₀ for PDE1 (Gonçalves et al., 2009), was shown to have no effects on the total cellular cGMP accumulation induced by ANP in subcultured RASMCs (Kim et al., 2001) or on cGMP signals induced by NO in primary SMCs isolated from mouse aorta (Krawutschke et al., 2015), suggesting that basal PDE1 activity had little effect on [cGMP]_i in these conditions.

We tested a new PDE1 inhibitor, PF1 at 1 μM, a concentration chosen to be 10 times higher than its IC₅₀ for PDE1, 30 times lower than its IC₅₀ for PDE3, PDE4, or PDE9, and 10 times lower than its IC₅₀ for PDE5 (Humphrey et al., 2014). In this condition, PF1 totally abolished the Ca²⁺/CaM‐stimulated cAMP‐ and cGMP‐PDE activities measured in the lysates of the RASMCs, strongly supporting inhibition of PDE1 by PF1. Furthermore, the fact that PF1 exhibited a different pattern of functional responses to that observed with sildenafil or PF9 supports a selective action of PF1 towards PDE1 over the other PDEs (see below). Overall, PF1 appears to be a useful tool to assess PDE1 function. We observed that in cultured RASMCs, PF1 had no effect on basal cAMP‐PDE activity but decreased basal cGMP‐PDE activity by almost 28%, suggesting that basal PDE1 activity preferentially hydrolyses cGMP. This basal cGMP‐PDE1 activity might be related to the PDE1A subtype, as PDE1A and PDE1B are preferential cGMP‐hydrolyzing enzymes, compared to PDE1C (Bender & Beavo, 2006) and as PDE1A but not PDE1B mRNA expression has been detected in these cells (Zhai et al., 2012). This is consistent with data obtained in subcultured RASMCs showing that PDE1A shRNA significantly increased total cGMP intracellular levels, without affecting cAMP levels (Nagel et al., 2006). Moreover, we observed that the Ca²⁺/CaM‐stimulated PDE1 hydrolyzed both cAMP and cGMP in cultured RASMCs. Contribution of PDE1A and PDE1C to this stimulated hydrolytic activity is worth considering (Cai et al., 2011; Kim et al., 2001; Zhai et al., 2012).

In line with these biochemical studies, we showed that PDE1 controlled some cGMP signals under basal conditions in living cells, but its activity on cAMP signals induced by β‐adrenoceptor stimulation was increased in the presence of Ang II. Ang II is a vasoconstrictor agent, increasing intracellular Ca²⁺ concentrations in cultured RASMCs allowing stimulation of PDE1 (Kim et al., 2001). Indeed, we observed that Ang II altered the cAMP response to β‐adrenoceptor stimulation, by accelerating the decay phase of the cAMP‐FRET signal elicited by a short application of isoprenaline or by decreasing the amplitude of the signal elicited by a maintained application of isoprenaline. PF1 reversed most of these Ang II‐mediated effects, revealing a role of the stimulated form of PDE1 in regulating [cAMP]_i concentrations in cultured RASMCs. The fact that PF1 only partly blocked the effect of Ang II on top of the isoprenaline response might be due to a partial inhibition of PDE1. However, this was unlikely as PF1 totally blocked the Ca²⁺/CaM‐activated PDE in the cAMP‐PDE activity assay. Another possible explanation is that Ang II decreases [cAMP]_i through a PDE1‐independent mechanism, for example, by reducing its production through Ca²⁺ inhibition of ACs, such as AC5 and AC6 (Dessauer et al., 2017; Ostrom et al., 2002; von Hayn et al., 2010) or by increasing its degradation through another PDE, such as PDE4 via CaMKII‐mediated stimulation as shown in cardiomyocytes (Mika, Richter, & Conti, 2015). Moreover, Ang II treatment was shown to increase the activities of PDE4 and PDE1 and the protein expression levels of PDE4A, PDE1A, and PDE1C (Mokni et al., 2010). However, our results clearly indicate that the role of PDE1 in controlling cAMP pools is low in cultured RASMCs but is intensified under high [Ca²⁺]_i conditions.

To our knowledge, our study is the first to demonstrate a functional role of PDE1 in controlling cGMP dynamics in cultured VSMCs. The lack of effect of PDE1 on cGMP signals reported by Krawutschke et al. (2015) might be explained either by the difference in PDE1 inhibitor (i.e., vinpocetine) or by the difference in VSMCs origin (i.e., primary SMCs isolated from mouse aorta). Interestingly, we observed that basal and Ca²⁺‐stimulated PDE1 exerted distinct roles on cGMP signals depending on the stimulus. Under basal conditions, PF1 appeared to preferentially affect the amplitude of the CNP‐mediated cGMP response, whereas under high [Ca²⁺]_i conditions, PF1 slowed the degradation phase of cGMP produced by DEA‐NO. One possible reason for this difference would be that the NO donor produced a rise in [cGMP]_i which was below the affinity of PDE1 for cGMP, so that basal PDE1 would be inactive. However, this is unlikely as DEA‐NO elicited higher maximal FRET ratios than CNP. More likely, the presence of different PDE1 isoforms, characterized by different affinities for their substrate as well as for Ca²⁺/CaM (Bender & Beavo, 2006), contributes to this difference by delimiting distinct intracellular PDE1 signalling domains. As discussed above, a PDE1A isoform, active in basal conditions, would locally control the CNP/NPR‐B‐dependent cGMP pool, in a restricting submembrane domain where PDE1A would be either maximally active or insensitive to the Ca²⁺ signalling generated by Ang II. By contrast, the NO/sGC/cGMP pool would be regulated by a cytosolic PDE1 isoform requiring a Ca²⁺ stimulus. It is noteworthy that we did not observe any effect of PDE1 inhibition on the ANP/NPR‐A‐dependent cGMP pool. This might appear contradictory to the reported inhibitory effect of vinpocetine on the Ang II potentiating effect on the ANP‐mediated cGMP response in cultured RASMCs (Kim et al., 2001). However, in this study, cGMP accumulation was measured by using a cGMP RIA on the cell lysate, disturbing all the subcellular microdomains. We believe that using intact living cells constitutes a more relevant tool to analyse the functional compartmentation of cellular signalling.

PDE5 is described as the main cGMP‐hydrolyzing PDE in VSMCs. Consistent with this description, sildenafil used at a low concentration to retain selectivity for PDE5, inhibited >50% of total cGMP‐PDE activity and strongly potentiated the cGMP responses elicited by the NO donor, ANP, and CNP in cultured RASMCs. This is in accordance with two other studies, using higher sildenafil concentrations, reporting that PDE5 controlled both the sGC‐ and the NPR‐dependent cGMP signals in mouse and rat primary aortic SMCs (Krawutschke et al., 2015; Nausch et al., 2008). Thus, PDE5 appears to be a key player in regulating cGMP pools in both contractile and synthetic VSMCs. The slower kinetics of the recovery phase of the FRET signal elicited by the NPs compared to that obtained with the NO donor is also in line with these two studies showing more sustained cGMP signals after pGC compared to sGC activation. In addition to cGMP degradation by PDEs, cGMP export out of the cell could also account for this difference (Krawutschke et al., 2015).

Finally, our study identified a new component of vascular cGMP signalling. While recent data has highlighted the potential benefits of targeting PDE9 to treat heart failure, the role of PDE9 in the vascular system remains unknown (Lee et al., 2015). Here, we showed that 5 μM PF9 reduced the total cGMP‐PDE activity by 37% in cultured RASMCs. When adding together PDE1, PDE5 and PDE9 activities, the contribution of PDE9 to the total cGMP‐PDE activity appears to be overestimated, suggesting that PF9 loses its PDE9 selectivity in this biochemical in vitro assay, acting possibly on PDE1 or PDE5. In FRET imaging, PF9 selectively altered the cGMP response to DEA/NO without altering the cGMP signals induced by ANP and CNP. This pattern of PF9 effect is clearly different from that observed with PF1 or sildenafil, strongly supporting that in intact cells, 5 μM PF9 selectively acts on PDE9, in line with the 20 times lower potency of PF9 to inhibit PDE9‐mediated functional effects in intact cells, compared to its potency to inhibit PDE9 activity in a cell‐free system (Hutson et al., 2011). In cardiomyocytes, PDE9 was shown to regulate NP‐ rather than NO‐stimulated cGMP concentrations (Lee et al., 2015). In RASMCs, PF9 potentiated the NO‐related cGMP response and acted at lower concentrations of DEA‐NO compared to sildenafil, in agreement with the higher cGMP affinity of PDE9 than PDE5 (Bender & Beavo, 2006). Thus, in RASMCs, the sGC‐cGMP pool would be regulated by PDE5 and PDE9, acting together.

Overall, our study underlines the specific roles of PDE1, PDE5, and PDE9 for the local control of [cGMP]_i in synthetic RASMCs, indicating a co‐operative, rather than a redundant role of each PDEs in this process, as illustrated in Figure 9. However, this specific role was not obvious in a more downstream process, that is, the control of cell proliferation. Several issues have to be considered regarding these two cellular responses: (a) the duration of the assay, based on an acute pharmacological treatment (“minutes” in FRET imaging) compared to a long‐term treatment (“48 hr” for cell proliferation) where compensatory mechanisms might be activated; (b) the nature of the measured response, namely, cGMP concentration, as a unique marker, compared to cell proliferation, a complex downstream response involving several signalling pathways, including cGMP‐dependent targets, such as PKG or also PDE3 responsible for cAMP elevation, or even cGMP‐independent targets. Thereby, PDE1C has been shown to regulate SMC proliferation through a critical cAMP/PKA‐dependent internalization of PDGFRβ (Cai et al., 2015). Thus, our data showing an absence of PF1 potentiating effect on cGMP‐dependent anti‐proliferative response, as well as the inhibition of FBS‐stimulated proliferation by PF1, are consistent with this study. Regarding PDE5, we confirmed its role in controlling RASMC proliferation by potentiating NPR‐A‐mediated responses. By itself, sildenafil did not reduce the FBS‐stimulated proliferation, whereas it was shown to reduce it modestly in PDGF‐BB‐stimulated human arterial SMCs (Wilson, Guo, Umana, & Maurice, 2017). Finally, the absence of significant effect of sildenafil on CNP‐mediated response could be explained by the reported contribution of NPR‐C, the third subtype of NPR which does not possess GC activity, to the anti‐proliferative effect of CNP in VSMCs (Khambata, Panayiotou, & Hobbs, 2011). Thus, we admit that several parameters are critical in the design of such functional experiments and might interfere with the data, among which the nature of the cell proliferation stimulus (i.e., FBS as an overall autocrine/paracrine mitogen stimulus that might have limited the anti‐proliferative effect of our treatments, compared to a defined mitogen), the concentration of the cGMP‐dependent stimulus, or the nature of the functional assay. However, our results show that the cGMP pools from different origins differentially regulate cell proliferation in VSMCs, so that further studies would be required to fully characterize the specific role of PDEs in controlling vascular cell proliferation. This might have clinical relevance as the pathological proliferative state of VSMCs is associated with vascular diseases, such as atherosclerosis and restenosis (Owens, Kumar, & Wamhoff, 2004) and is partly related to up‐regulation of PDEs (Cai et al., 2015; Nagel et al., 2006). Thus, PDE inhibitors may represent potential therapeutic drugs to counter this state, and a better understanding of the functional specificity of each PDE family in controlling cGMP pools might help identifying a relevant drug target candidate.

Figure 9.

Regulation of distinct cyclic nucleotides pools by PDE1, PDE5, and PDE9 in cultured RASMCs. The β‐adrenoceptor‐stimulated cAMP pool is controlled by the Ca²⁺‐activated PDE1. The NO/sGC‐dependent cGMP pool is regulated in concert by the Ca²⁺‐activated PDE1, PDE5, and PDE9. The ANP/NPR‐A‐dependent cGMP pool is under the control of PDE5, whereas the CNP/NPR‐B‐dependent cGMP pool is under the control of both PDE5 and basal PDE1. Note that other PDEs, not addressed in the current study, may also contribute

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

L.Z., G.V., R.F., and V.L. conceived and designed the experiments. L.Z., K.B., and B.M. performed the experiments. L.Z., K.B., and V.L. analysed the data. L.Z., R.F., and V.L. wrote the paper. All authors critically reviewed the content and approved the final version of the manuscript.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis and Animal Experimentation, and as recommended by funding agencies, publishers and other organisations engaged with supporting research.

Supporting information

Figure S1. Effect of PDE1 inhibition on β‐AR‐induced cAMP signaling in the presence of PDE4 inhibitor in RASMCs. Cyclic AMP measurements were conducted in cultured RASMCs using the Epac2‐camps FRET‐based sensor, in response to a short application of isoproterenol (Iso, 10 nM, 15 s) in the presence of the PDE4 inhibitor (10 μM Ro‐20‐1724, Ro, applied 4 min before Iso), and in the absence or presence of the PDE1 inhibitor (1 μM PF1, applied 4 min before Iso). Upper and lower panels represent the mean variation of the percentage increase in CFP/YFP ratio and the corresponding kinetic parameters, respectively. Data are mean ± SEM of 8 cells.

Figure S2. Expression of PDE9 protein in cultured RASMCs and rodent brain tissue. Western blotting was performed as previously reported (Hubert et al., 2014). Protein extracts were prepared from RASMCs cultured at passages 4 (P4), 5 (P5) or 6 (P6), as well as from mouse and rat brain tissues (as control tissue for PDE9 expression), in standard loading buffer under reducing conditions. Following primary antibodies were used: a rabbit anti‐PDE9 (1/1000; #13128–5, kind gift from Dr. L. Jaffe, University of Connecticut Health, USA) and a rabbit anti‐GAPDH, as loading control (1/1000; #5174, Cell Signaling Technology, Leiden, The Netherlands). A donkey anti‐rabbit IgG‐HRP was used as secondary antibody (1:10000; sc‐2313, Santa Cruz Biotechnology, Dallas, TX, USA). This representative immunoblot shows several bands, a clear band between 55 and 70 kDa (putatively corresponding to PDE9A6 and PDE9A13 isoforms, according to Jaffe's details and literature) in mouse and rat brain extracts, and with a lower intensity in cultured RASMC extracts, as well as additional bands at higher molecular weight that might refer to the 3 novel isoforms PDE9X‐100, PDE9X‐120 and PDE9X‐175, recently described by Patel et al. (2018). The matching GAPDH signal is also shown.

Zhang L, Bouadjel K, Manoury B, Vandecasteele G, Fischmeister R, Leblais V. Cyclic nucleotide signalling compartmentation by PDEs in cultured vascular smooth muscle cells. Br J Pharmacol. 2019;176:1780–1792. 10.1111/bph.14651