Phosphodiesterase 4D promotes angiotensin II-induced hypertension in mice via smooth muscle cell contraction

Hypertension is a common chronic disease, which leads to cardio-cerebrovascular diseases, and its prevalence is increasing. The cyclic adenosine monophosphate (cAMP)-protein kinase A (PKA) pathway participates in multiple cardiovascular diseases. Phosphodiesterase (PDE) 4 has been shown to regulate PKA activity via cAMP specific hydrolysis. However, whether PDE4-cAMP-PKA pathway influences hypertension remains unknown. Herein, we reveal that PDE4D (one of PDE4 isoforms) expression is upregulated in the aortas of experimental hypertension induced by angiotensin II (Ang II). Furthermore, knockout of Pde4d in mouse smooth muscle cells (SMCs) attenuates Ang II-induced hypertension, arterial wall media thickening, vascular fibrosis and vasocontraction. Additionally, we find that PDE4D deficiency activates PKA-AMP-activated protein kinase (AMPK) signaling pathway to inhibit myosin phosphatase targeting subunit 1 (MYPT1)-myosin light chain (MLC) phosphorylation, relieving Ang II-induced SMC contraction in vitro and in vivo. Our results also indicate that rolipram, a PDE4 inhibitor, may be a potential drug for hypertension therapy.

H ypertension is defined as an arterial systolic and diastolic blood pressure (BP) >140/90 mmHg by European Society of Cardiology/European Society of Hypertension 1 . While generally asymptomatic, hypertension is a severe risk factor for cardiovascular diseases, strokes, and kidney diseases 2 . Hypertension occurs through multiple pathogeneses, including sympathetic activation 3 , the renin-angiotensin-aldosterone system disorder 4 , inflammation 5 , and endothelial cell (EC) and smooth muscle cell (SMC) dysfunction 6,7 . Presently, most hypertension medicines have adverse effects-headaches, oedema, and hyperkalaemiawhich limit their application and lead to reduced patient compliance 8 . Besides, there are some hypertension patients who are insensitive to existing antihypertensive drugs, ultimately lead to resistant hypertension 9 . It is therefore imperative to develop potential hypertension treatments.
Phosphodiesterase (PDE), consisting of 11 subfamilies (PDE1-PDE11), is the hydrolase of cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate. Then, PDE4, consisting of four isoforms (PDE4A-D), are cAMP specific hydrolases 10 . PDE4 participates in a variety pathophysiological processes 11 , promoting SMCs' phenotypic switch and neointima formation in atherosclerosis 12 , as well as aggravating pulmonary arterial hypertension through the regulation of vascular tone and inflammatory factors 13 . In addition, PDE4 inhibitors is an effective treatment strategy for a variety of diseases, including asthma, chronic obstructive pulmonary disease, and psoriasis 14 . Exploring the role of PDE4 isoenzymes in hypertension is vital to the development of new treatment strategies.
As a second messenger, cAMP is related to cardiovascular diseases: cardiac fibrosis, abdominal aortic aneurysm, atherosclerosis, and pulmonary arterial hypertension [15][16][17][18] . In addition, the association between cAMP and these diseases was mainly established through its effector protein kinase A (PKA) 17 . However, the roles of PKA and its regulator PDE in hypertension remain unknown. Furthermore, PKA has been shown to activate AMPK 19 . AMPK inhibitor aggravated SMCs contraction and hypertension by activating MYPT1-MLC signaling pathway 20 . As known, one of the pathological processes of hypertension is vasoconstriction, and MYPT1-MLC is a classical cell contraction signaling pathway 21 . Therefore, it is hypothesized that PDE4 may affect SMCs contraction by PKA-AMPK-MYPT1-MLC pathway and thus affect hypertension.
In this study, we found that PDE4D expression was upregulated in aortic tissues of hypertensive mice. Furthermore, PDE4D, expressed in SMCs instead of ECs, contributed to hypertension development. PDE4D deficiency in SMCs and PDE4 inhibitor rolipram reduced Ang II-induced hypertension, and the protective effect of rolipram on hypertension was mainly through PDE4D in SMCs. In addition, we demonstrated that PDE4D promoted SMCs contraction and vasocontraction via PKA-AMPK-MYPT1-MLC signaling pathway.

Results
Phosphodiesterase 4D (PDE4D) expression is upregulated in angiotensin (Ang) II-induced hypertensive mice. We first established a hypertensive model in wild-type (WT, C57BL/6J) male mice ( Supplementary Fig. 1a, b). To initially investigate PDE4 expression after hypertension, we evaluated mRNA levels of each PDE4 isoform (Pde4a-d) in control and hypertensive mice aortas. The results revealed a increase in Pde4d mRNA level of hypertensive mice (Fig. 1a). The western blot and immunohistochemical Data are expressed as mean ± standard error of mean (SEM). Two-tailed Student's t test was performed to compare differences between two groups. *p < 0.05, ***p < 0.001. L lumen.
staining tests showed that PDE4D expression was increased in hypertensive mice (Fig. 1b-e). However, there was no change in other PDE4 isoforms ( Fig. 1a and Supplementary Fig. 2a, b). Together, these findings indicate that PDE4D expression is elevated in Ang II-induced hypertensive mice aortas.
We then induced hypertension in two knockout mice groups. Using the tail-cuff method, we measured BP on the first day and every other day during Ang II infusion. After 2 weeks, we harvested the aorta tissues (Fig. 2a) Fig. 2d, e). Hematoxylin and eosin (H&E) staining revealed that Ang II-induced vessel wall media thickening, which was reduced in Pde4d SMC−/− Ang II infused mice (Fig. 2f, g). In addition, masson-trichrome staining demonstrated that SMC Pde4d deficiency reversed Ang II-induced vascular fibrosis (Fig. 2h, i). These results indicate that PDE4D in SMCs, but not in ECs, contribute to Ang II-induced mice hypertension.
PDE4D promotes SMCs contraction via the PKA-AMPK-MYPT1-MLC signaling pathway in vitro. To determine PDE4D's role in regulating SMC contraction, we evaluated its impact on rat aorta smooth muscle cells (RASMCs) in vitro. First, we verified that PDE4D protein expression but not PDE4A-C was elevated by 5.37-fold in RASMCs after Ang II stimulation (100 nM, 24 h; Fig. 4a, b and Supplementary Fig. 4a, b). Consistently, we found that it was Pde4d upregulation instead of other PDE4 isoforms in RASMCs treated with Ang II in mRNA level ( Supplementary Fig. 4c). Immunofluorescence staining also showed that PDE4D was increased by Ang II in vitro (Supplementary Fig. 4d). Besides, PDE4 activity was increased by 2.34fold with Ang II stimulation ( Supplementary Fig. 4e). These results indicated that PDE4D was upregulated by Ang II in vitro. Then, we introduced PDE4D small-interfering RNA (siRNA) to validate whether Ang II induces SMCs contraction via PDE4D. After PDE4D siRNA administration, only PDE4D expression, instead of other PDE4 isoforms, was reduced in RASMCs' mRNA and protein levels ( Supplementary Fig. 5a-c). PDE4D was reduced by PDE4D siRNA through immunofluorescence staining ( Supplementary Fig. 5d). Indeed, PDE4 activity was also suppressed when PDE4D was knockdown ( Supplementary Fig. 5e). We next performed a collagen gel cell contraction assay to explore RASMCs contraction, which revealed that while Ang II promoted RASMC contraction, the addition of PDE4D siRNA inhibited it (si-control + Ang II vs. si-PDE4D + Ang II: 72.68% ± 2.11% vs. 58.65% ± 1.76%) (Fig. 4c, d). in SMCs contributes to Ang II-induced mouse hypertension. a Scheme of hypertensive mice inducement. b SBP and c DBP were measured in Pde4d flox/flox and Pde4d EC−/− mice, with or without Ang II treatment: n = 5 in Pde4d flox/flox + saline group, n = 8 in Pde4d flox/flox + Ang II group, n = 5 in Pde4d EC−/− + saline group, and n = 8 in Pde4d EC−/− + Ang II group. d SBP and e DBP were measured in Pde4d flox/flox and Pde4d SMC−/− mice, with or without Ang II treatment: n = 5 in Pde4d flox/flox + saline group, n = 10 in Pde4d flox/flox + Ang II group, n = 5 in Pde4d SMC−/− + saline group, and n = 8 in Pde4d SMC−/− + Ang II group. Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. ***p < 0.001 for Pde4d flox/flox + Ang II group vs. Pde4d flox/flox + saline group, and ### p < 0.001 for Pde4d SMC−/− + Ang II group vs. Pde4d flox/flox + Ang II group. f Representative H&E staining under the indicated experimental conditions. g Measurement of arterial wall media thickness including all mice in f. h Representative masson-trichrome staining under the indicated experimental conditions. i Quantification of the positively stained area to the aortic wall area including all mice in h. Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. **p < 0.01, ***p < 0.001. L lumen. . Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. **p < 0.01 and ***p < 0.001 for Pde4d flox/flox + Ang II group vs. Pde4d flox/flox + saline group. # p < 0.05, ## p < 0.01, and ### p < 0.001 for Pde4d SMC−/− + Ang II group vs. Pde4d flox/flox + Ang II group.
Additionally, AMPK activation inhibits phosphorylation of myosin phosphatase targeting subunit 1 (MYPT1) and myosin light chain (MLC), consequently attenuating SMCs contraction 20 . Therefore, we hypothesized that PDE4D may further increase MYPT1 and MLC phosphorylation by suppressing AMPK activation, promoting Ang II-induced SMCs contraction. Indeed, Ang II increased MYPT1 and MLC phosphorylation, whereas PDE4D siRNA suppressed MYPT1 and MLC phosphorylation (Fig. 4f, h, i). Moreover, the effect of PDE4D siRNA on reducing MYPT1 and MLC phosphorylation was largely reversed by PKI ( Supplementary Fig. 6a, c, d). We also used AMPK inhibitor (Compound C, 20 μM, 2 h) for further validation. The inhibitory effect of PDE4D siRNA on pMYPT1 and pMLC was blocked by  Fig. 6e-g). These results suggest that PDE4D promotes SMCs contraction via inhibition of PKA activity and AMPK phosphorylation, and conversely promotes MYPT1 and MLC phosphorylation.
PDE4D promotes vasocontraction through the PKA-AMPK-MYPT1-MLC signaling pathway in Ang II-induced mice hypertension. To further validate the mechanism identified in vitro above, we detected PKA activity and AMPK, MYPT1, and MLC phosphorylation in mice aorta tissues. Consistently, Ang II infusion reduced PKA activity in mice aortas, and this reduction was reversed in Pde4d SMC−/− Ang II mice (Fig. 5a). Western blot also exhibited that Ang II infusion reduced AMPK phosphorylation in the aorta, an effect which was recovered in Pde4d SMC−/− mice (Fig. 5b, c). Ang II infusion also increased MYPT1 and MLC phosphorylation, and again Pde4d SMC−/− mice exhibited reduced Ang II-induced MYPT1 and MLC phosphorylation (Fig. 5b, d, e). These results suggest that PDE4D promotes vasocontraction, and thus contributes to Ang II-induced hypertension in mice, through the PKA-AMPK-MYPT1-MLC signaling pathway.

Discussion
In this study, we observed upregulated PDE4D expression in hypertensive mice aortas, which showed that PDE4D contributes to hypertension. Furthermore, via EC-and SMC-specific Pde4d knockout hypertensive mice, these models revealed a causal association between SMC Pde4d and vasocontraction in hypertension. To further elucidate this association, we investigated a potential mechanism for PDE4D involvement in SMC contraction and hypertension development, and identified the PKA-AMPK-MYPT1-MLC signaling pathway to be a likely candidate. Importantly, we demonstrated that rolipram, a pan PDE4 inhibitor, relieved Ang II-induced hypertension mainly by inhibiting PDE4D in SMCs, which suggested that PDE4D might represent a potential therapeutic hypertension target (Fig. 8).
Hypertension is well known to be a complex syndrome involving multiple organs, tissues, and cells 32,33 . Among the cell types associated with hypertension, PDE4D is also expressed in fibroblasts 34,35 . Adventitial fibroblasts, another major component of blood vessels, are the primary cause of collagen deposition and aortic stiffening in hypertension 36 . In this study, we observed vascular collagen deposition in Ang II infusion mice. Accordingly, it should not be discounted that PDE4D could further contribute to the development of hypertension by interfering with collagen production in fibroblasts, a possibility warranting future investigation.
Hypertension is commonly associated with inflammation 37,38 , and the cell types involved in inflammation (T lymphocytes 39 , B lymphocytes 40,41 , dendritic cells, monocytes, and macrophages 42 ) are all known to promote hypertension. PDE4D has been shown to interact with cytokines, regulate the function of inflammatory cells, and aggravate the inflammatory response [43][44][45] . While PDE4D's role in the inflammation response is outside the scope of this study, our findings, along with the body of literature evidence, suggest that PDE4D could also contribute to hypertension via inflammation regulation. Although, further study would be needed to validate this supposition.
While collagen deposition and inflammation may be potential additional mechanisms, we demonstrated a link between PDE4D and the PKA-AMPK signaling pathway. cAMP is known to be involved in signal transduction through PKA regulation 46 . Recently, researchers have found that PKA phosphorylates AMPKα at Thr-172 through the widely expressed tumor suppressor liver kinase B1, ultimately activating AMPK 19,30 . AMPK activity has been linked to numerous cardiovascular diseases, including hypertension, atherosclerosis, and heart failure 20,47-49 . Crucially, AMPK activation lowers BP and suppresses SMC contractility by inhibiting the MYPT1-MLC signaling pathway 20 . Consistent with previous reports, our results exhibited that PDE4D upregulated MYPT1 and MLC phosphorylation by inhibiting the PKA-AMPK signaling pathway, inducing SMCs contraction and thereby, hypertension.
In conclusion, our study provided that PDE4D in SMCs aggravated Ang II-induced hypertension. We identified the mechanism by which PDE4D affected SMCs contraction via in vitro and in vivo experimental models and verified those results through several molecular biology approaches. In addition, rolipram alleviated hypertension mainly through PDE4D in SMCs. This study elucidated PDE4D as a potential target for the treatment of hypertension and, potentially, other cardiovascular diseases. To test the effect of rolipram on hypertension, WT, Pde4d flox/flox and Pde4d SMC−/− mice at 8 weeks old were infused with Ang II (490 ng kg −1 min −1 ) and subcutaneously implanted with osmotic pumps for 14 days. In total, 0.375 mg ml −1 rolipram (PDE4 inhibitor, 8 ml kg −1 day −1 , dissolved in ethyl alcohol; Sigma-Aldrich, Cat#: R6520) was administered orally, via gavage, daily for 14 days. WT mice were divided into four groups randomly: Rolipram (−) mice infused with saline (n = 5) or Ang II (n = 9), and Rolipram (+) mice infused with saline (n = 6) or Ang II (n = 10). Pde4d flox/flox and Pde4d SMC−/− mice were divided into four groups randomly: Pde4d flox/flox + Ang II Fig. 6 Effect of rolipram on Ang II-induced hypertension in mice. a Scheme of hypertensive mice treated with vehicle or rolipram. Rolipram (3 mg kg −1 day −1 ) was orally administered daily for 14 days. b SBP and c DBP were measured in wild-type (WT) mice with or without Ang II/rolipram treatment: n = 5 in Rolipram (−) + saline group, n = 9 in Rolipram (−) + Ang II group, n = 6 in Rolipram (+) + saline group, and n = 10 in Rolipram (+) + Ang II group. Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. *p < 0.05, ***p < 0.001 for Rolipram (−) + Ang II group vs. Rolipram (−) + saline group, and # p < 0.05, ### p < 0.001 for Rolipram (+) + Ang II group vs. Rolipram (−) + Ang II group. d Representative H&E staining under the indicated experimental conditions. e Measurement of arterial wall media thickness including all mice in d. f Representative masson-trichrome staining under the indicated experimental conditions. g Quantification of the positively stained area to the aortic wall area including all mice in f. Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001 in e, g. Concentration-response curves for h PE and i Ang II induced vasocontraction of mesenteric resistance artery in WT mice with or without Ang II/rolipram treatment including all mice. Data are expressed as mean ± SEM. Two-way ANOVA with Bonferroni's post hoc test was performed to compare the difference between the multiple groups. *p < 0.05, **p < 0.01, ***p < 0.001 for Rolipram (−) + Ang II group vs. Rolipram (−) + saline group group. # p < 0.05, ## p < 0.01, ### p < 0.001 for Rolipram (+) + Ang II group vs. Rolipram (−) + Ang II group. L lumen.
BP measurement by tail-cuff plethysmography. SBP and DBP in mice were measured using the CODA non-invasive BP system (Kent Scientific Co., Torrington, CT, USA) according to the manufacturer's instructions 51 . Each mouse was gently placed in a sizeable holder and allowed to acclimate for 5 min. The tail was then threaded through the occlusion cuff and the sensor cuff, which was then attached to the controller. For each measurement, five values of SBP and DBP were recorded for each mouse and their mean values were used as the final result. Histological and immunohistochemical analysis. Mouse aorta segments were cut at the thoracic aorta, embedded vertically with OCT compound (SAKURA, Cat#:4538), and then stored at −80°C. Ten to 15 serial frozen sections containing the entire vascular lumen were sectioned using a freezing microtome (Leica CM1860), and then fixed with 4% paraformaldehyde. Frozen sections (6 μm) were stained by immunohistochemical staining of PDE4D (1:100, Abcam, Cat#: ab14613) by the 3-amino-9-ethylcarbazole staining method, H&E (Solarbio, Cat#: G1120), and masson-trichrome staining (Servicebio, Cat#: G1006). Images were photographed using a Leica optical microscope (Leica Microsystems, Germany) and the integrated optical density (IOD) values of positive staining analyzed using Image-Pro Plus software (Media Cybernetics, USA). For statistical analysis, 5 images per mouse of each group were randomly selected. For PDE4D immunohistochemical staining, the content of PDE4D was quantified as the ratio of positively stained area to the total cross-sectional area of the aortic wall. For masson-trichrome staining, the degree of vascular fibrosis was quantified as the ratio of the positively stained area to the total cross-sectional area of the aortic wall. For H&E staining, arterial wall media thickness was measured using Nikon NIS-Elements image analysis software (Nikon Instruments Inc., Japan).
PDE4 activity assay. PDE4 activity in RASMCs was detected via a PDE4 activity assay kit (Abcam, Cat#: ab139460) according to the manufacturer's instructions. PDE4 activity was inhibited with rolipram (20 uM) during the test. Total PDE activity assay and PDE activity assay after inhibition of PDE4 were performed for each sample, and PDE4 specific activity was calculated by subtracting inhibitory activity from total activity. Absorbance was measured at OD = 620 nm via a multimode microplate reader (BioTek Synergy™ HTX, BioTek Instrument, Inc., Winooski, USA). The relative activity of PDE4 was quantified as the ratio of the active PDE4 content to the sample total protein content.
Cell contraction assay. RASMCs contraction was detected using a Cell Contraction Assay Kit (Cell Biolabs, Inc., San Diego, CA, USA, Cat#: CBA-201) according to the manufacturer's instructions 53 . RASMCs were treated with siRNA for 48 h with or without Ang II for 24 h, and then cultured in collagen gel for 48 h to develop mechanical load. The surface image of the collagen gel was captured via digital camera, and analyzed using Image-Pro Plus software (Media Cybernetics, USA). The percentage of contraction was the ratio of gel contracted surface area to the dish bottom.
Statistics and reproducibility. Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software Inc., La Jolla, CA). Data are expressed as means ± standard error of mean. Two-tailed Student's t test was performed to compare differences between two groups from at least three independent experiments. Oneway ANOVA or two-way ANOVA with Bonferroni's post hoc test was performed to compare differences between multiple groups, using at least three independent experiments. p value < 0.05 was considered statistically significant.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Raw data of genotyping and western blot are provided in Supplementary Fig. 7. Source data underlying the graphs are provided in Supplementary Data 1. Other relevant data are available from the corresponding author upon request.