antioxidants Article Rolipram Prevents the Formation of Abdominal Aortic Aneurysm (AAA) in Mice: PDE4B as a Target in AAA Saray Varona 1,2,3,† , Lídia Puertas 2,3,4,† , María Galán 2,3,4 , Mar Orriols 1,2,3 , Laia Cañes 2,3,4 , Silvia Aguiló 3,4 , Mercedes Camacho 2,3,4 , Marc Sirvent 2,5 , Vicente Andrés 2,6 , José Martínez-González 1,2,3, * and Cristina Rodríguez 2,3,4, * 1 Departmento de Patología Experimental, Instituto de Investigaciones Biomédicas de Barcelona-Consejo Superior de Investigaciones Científicas (IIBB-CSIC), 08036 Barcelona, Spain; saray.var@gmail.com (S.V.); marorriols@gmail.com (M.O.) 2 CIBER de Enfermedades Cardiovasculares, ISCIII, 28029 Madrid, Spain; lpuertas@santpau.cat (L.P.); mgalana@santpau.cat (M.G.); laia.canes@iibb.csic.es (L.C.); mcamacho@santpau.cat (M.C.); msirvent10@gmail.com (M.S.); vandres@cnic.es (V.A.) 3 Regulatory Mechanisms of Cardiovascular Remodelling Group, Instituto de Investigación Biomédica Sant Pau, 08041 Barcelona, Spain; saguilo@santpau.cat 4 Regulatory Mechanisms of Cardiovascular Remodelling Group, Institut de Recerca Hospital de la Santa Creu i Sant Pau (IRHSCSP), 08025 Barcelona, Spain 5 Angiology and Vascular Surgery Service, Hospital Universitari Germans Trias i Pujol, 08916 Badalona, Spain 6 Vascular Pathophysiology Area, Centro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC), 28029 Madrid, Spain  * Correspondence: jose.martinez@iibb.csic.es (J.M.-G.); crodriguezs@santpau.cat (C.R.)  † These authors contributed equally. Citation: Varona, S.; Puertas, L.; Galán, M.; Orriols, M.; Cañes, L.; Abstract: Abdominal aortic aneurysm (AAA) is a common life-threatening condition characterized by Aguiló, S.; Camacho, M.; Sirvent, M.; exacerbated inflammation and the generation of reactive oxygen species. Pharmacological treatments Andrés, V.; Martínez-González, J.; to slow AAA progression or to prevent its rupture remain a challenge. Targeting phosphodiesterase 4 et al. Rolipram Prevents the (PDE4) has been verified as an effective therapeutic strategy for an array of inflammatory conditions; Formation of Abdominal Aortic however, no studies have assessed yet PDE4 in AAA. Here, we used angiotensin II (AngII)-infused Aneurysm (AAA) in Mice: PDE4B as apolipoprotein E deficient mice to study the involvement of the PDE4 subfamily in aneurysmal a Target in AAA. Antioxidants 2021, disease. PDE4B but not PDE4D was upregulated in inflammatory cells from both experimental 10, 460. https://doi.org/10.3390/ antiox10030460 and human AAA. The administration of the PDE4 selective inhibitor rolipram (3 mg/kg/day) to AngII-challenged mice (1000 ng/kg bodyweight/min) protected against AAA formation, limiting the Academic Editor: Guillermo Zalba progressive increase in the aortic diameter without affecting the blood pressure. The drug strongly attenuated the rise in vascular oxidative stress (superoxide anion) induced by AngII, and decreased Received: 18 February 2021 the expression of inflammatory markers, as well as the recruitment of macrophages (MAC3+), Accepted: 11 March 2021 lymphocytes (CD3+), and neutrophils (ELANE+) into the vessel wall. Rolipram also normalized the Published: 16 March 2021 vascular MMP2 expression and MMP activity, preserving the elastin integrity and improving the vascular remodelling. These results point to PDE4B as a new therapeutic target for AAA. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in Keywords: abdominal aortic aneurysm; reactive oxygen species; PDE4B; rolipram published maps and institutional affil- iations. 1. Introduction Abdominal aortic aneurysm (AAA) is a common life-threatening degenerative condi- Copyright: © 2021 by the authors. tion characterized by the progressive weakening and dilation of the arterial wall, which Licensee MDPI, Basel, Switzerland. could ultimately lead to vascular rupture. AAA has a prevalence of 4–7% in men over the This article is an open access article age of 65 years, while aortic rupture, the more severe complication of this disease, results in distributed under the terms and extensive internal bleeding, and constitutes a major cause of sudden death among elderly conditions of the Creative Commons males [1–3]. Nowadays, the surveillance of small aneurysms and elective AAA repair is Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ the unique therapeutic approach for AAA, while the finding of effective pharmacological 4.0/). tools that limit the expansion and rupture of the aneurysms remains elusive [4,5]. The Antioxidants 2021, 10, 460. https://doi.org/10.3390/antiox10030460 https://www.mdpi.com/journal/antioxidants Antioxidants 2021, 10, 460 2 of 17 development of nonsurgical therapeutic interventions is hampered by the complexity of AAA pathogenesis, which remains only partially characterized. Chronic vascular inflammation, the proteolytic degradation of the extracellular matrix (ECM), neovascularisation, microcalcification, and vascular smooth muscle cell (VSMC) death by apoptosis are prominent features of AAA [6–9]. Among them, inflammation is critical in the onset and progression of this disease. Macrophages, lymphocytes, neutrophils, mast cells, and natural killer cells recruited into adventitia and media lead to an exacerbated production of proinflammatory cytokines and chemokines, reactive oxygen species (ROS), and metalloproteinases (MMPs), which are responsible for the proteolytic degradation of ECM [10–12]. Cyclic nucleotide phosphodiesterases (PDEs) encompass a vast family of isozymes, which catalyse the hydrolysis and inactivation of cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP), second messengers which play a major role in cardiovascular function [13–15]. In particular, by controlling cAMP degradation rates, four PDE subfamilies (PDE3, PDE4, PDE7 and PDE8) might impact the vascular function modulating endothelial permeability, monocyte/macrophage activation, and the proliferation, migration, differentiation and contraction–relaxation of VSMC [15–18]. The PDE4 subfamily, one of the best characterized subfamilies, is constituted by four mem- bers (PDE4A-D) with multiple isoform variants expressed in a wide array of tissues and involved in numerous pathophysiological processes. Targeting PDE4 has been validated as an effective therapeutic strategy for inflammatory conditions, including chronic obstruc- tive pulmonary disease (COPD), inflammatory bowel diseases (IBD), rheumatic arthritis (RA), atopic dermatitis (AD), asthma, psoriasis, and lupus [14,15]. PDE4 enzymes seem to play non-redundant functions, but both PDE4D and PDE4B have been involved in inflammation and oxidative stress [19–21], and PDE4 inhibitors have been shown to be able to reduce inflammation and oxidative stress both in vitro and in animal models of tissue damage [21–23]. Although inflammation and oxidative stress are key hallmarks of AAA, it is uncertain whether PDE4 isoforms could contribute to this human disease, thereby constituting potential targets for pharmacological interventions. Here, we demonstrate that PDE4B expression is upregulated in both human and experimental AAA, and that the inhibition of PDE4 activity reduces inflammation and oxidative stress, preventing the formation of AAA induced by angiotensin II (AngII) infusion in apolipoprotein E deficient (ApoE−/− ) mice. 2. Materials and Methods 2.1. Human Samples Human aneurysmal samples were obtained from patients undergoing the open surgery repair of AAA at the Hospital de la Santa Creu i Sant Pau (HSCSP; Barcelona, Spain). Healthy aortas were collected from multi-organ donors, as previously described [24,25]. Abdominal aorta segments were collected from patients and control donors, following strict standard operating procedures and ethical guidelines. The samples of the control subjects had no post-mortem evidence of AAA, atherosclerotic plaques, or other medi- cal conditions that affect the study. Upon collection, the samples were rapidly stored at −80 ◦ C for subsequent RNA extraction, or were processed for immunohistochemical analysis. 2.2. Animal Handling The animals were housed at the Animal Experimentation Unit (Institut de Recerca HSCSP, Barcelona, Spain) under a standard 12-h light/12-h dark cycle, in a controlled temperature and humidity environment, with ad libitum access to food and water. AAA was induced in 12-week old male apolipoprotein-E-deficient mice (ApoE−/− ; B6.129P2- Apoetm1Unc/J ) by a subcutaneous infusion of AngII, as previously described [26,27]. AngII (1000 ng/kg bodyweight (BW)/min; Sigma-Aldrich, St Louis, MO, USA) was infused via osmotic micropumps (model 1004, Alzet; Durect Corporation, Cupertino, CA, USA) for Antioxidants 2021, 10, 460 3 of 17 28 days [27–29]. In brief, the mice were anesthetized with 2% inhaled isoflurane. The anaesthetic depth was confirmed by the loss of blink reflex and/or the lack of response to foot reflex. Then, micropumps were implanted subcutaneously into the interscapular space. The recovery after surgery was carried out using aseptic techniques in a dedicated and approved surgical area. The animals were kept warm in a heating pad until they woke after surgery, and were observed carefully by the investigators throughout the post-surgery period. Analgesics were given immediately after surgery in order to prevent discomfort (buprenorphine, subcutaneously administered once a day during two days; 0.05 mg/kg). The animals were randomly distributed in three experimental groups: AngII-infused mice; AngII-infused mice receiving a daily intraperitoneal injection of the PDE4 inhibitor rolipram, starting 24 h before the micropump implantation (3 mg/kg; SelleckChem, Hous- ton, TX, USA) [30]; and saline-infused mice, used as controls. At the end of the experimental procedures, the mice were deeply anesthetized via intraperitoneal injection with ketamine (150 mg/kg) and medetomidine (1 mg/kg), sacrificed by thoracotomy, and the aortas were harvested, explored to determine the presence and severity of aneurysms, and appropri- ately processed for further analysis. 2.3. Non-Invasive Measurement of Systolic Blood Pressure The measurement of systolic blood pressure (SBP) was performed non-invasively in conscious mice prior to and throughout the experimental period using the tail-cuff plethys- mography method (CODA® tail-cuff blood pressure system; Kent Scientific Corporation; Torrington, CT, USA). The mice were previously trained for tail-cuff measurements over a period of 1 week. The measurements were recorded every week at the same time (between 9 a.m. and 11 a.m.) in order to avoid the influence of the circadian cycle, and the blood pressure values were derived from an average of at least 6 measurements per animal each day [27–29]. 2.4. Ultrasound Measurement of the Abdominal Aortic Diameter The abdominal aortic diameter was measured by ultrasound imaging before AngII infusion, and weekly until the end of the experimental protocol. In brief, the mice were anesthetized with inhaled isoflurane (2%) and laid supine onto a warming platform. Then, the precordium was shaved, and abdominal echography was carried out using a Vevo 2100 ultrasound with a 30 MHz transducer applied to the abdominal wall in order to record the abdominal aorta (VisualSonics, Toronto, ON, Canada), as previously described [24,27]. The external diameters of the abdominal aorta with values >1.5 mm were considered as aneurysm. All of the primary measurements were made from images captured on cine loops of 100 frames at the time of the study, using the software provided by the ultrasound machine. The severity of the aneurysm was established as previously described [31], based on a 4-point grading scale: type 0, no aneurysm; type I, dilated lumen in the suprarenal region of the aorta with no thrombus; type II, remodelled tissue in the suprarenal region that frequently contained thrombus; type III, a pronounced bulbous form of type II that contained thrombus; and type IV, a form in which there are multiple AAAs containing thrombus. 2.5. Basic Measurements of Cardiac Function by Echocardiography (M-mode and Doppler) The anesthetized mice (2% isoflurane inhalation) were subjected to a transthoracic echocardiography, using a Vevo 2100 ultrasound with a 30 MHz transducer (VisualSonics, Toronto, Canada) [28,29]. Two-dimensional and M-mode images were obtained in paraster- nal long-axis and short-axis views, respectively. The following parameters were recorded: heart rate (HR), end-diastolic interventricular septum thickness (IVSTd), left ventricular (LV) end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), LV end-diastolic posterior wall thickness (PWTd), and LV anterior wall thickness at the diastole (AWTd). The LV mass, LV ejection fraction (LVEF) and LV fractional shortening (LVFS) were deter- Antioxidants 2021, 10, 460 4 of 17 mined. The LV stroke volume and cardiac output (CO) were calculated according to the following: LV stroke volume (µL) = LV end-diastolic volume − LV end-systolic volume CO (mL/min) = (LV stroke volume × HR)/1000 2.6. Real-Time Polymerase Chain Reaction (PCR) The total RNA was isolated from both the mouse and human aortas using TriPure Isolation Reagent (Roche Diagnostics, Mannheim, GE). The RNA (500 ng) was reverse transcribed into cDNA using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). The quantification of the mRNA levels was carried out by real-time PCR using an ABI PRISM 7900HT sequence detection system (Applied Biosystems). Specific primers and probes (provided by the Assay-on-Demand system; Applied Biosystems) were used for the quantification of the human mRNA levels of PDE4B (Hs00277080_m1), PDE4D (Hs01579625_m1), and the reference gene, ACTB (ß-actin; HS99999903_m1). In order to analyse the mRNA levels in the mouse aorta, the following primers and probes (provided by Applied Biosystems or Integrated DNA Technologies, Coralville, IA, USA) were used: Mcp1 (Mm.PT.58.42151692), Col1a1 (Mm.PT.58.7562513), Emr1 (Mm.PT.58.11087779), Mmp2 (Mm.PT.58.9606100) and Mmp9 (Mm.PT.58.10100097). TATA-binding protein (Tbp; Mm00446973_m1) was used as endogenous control for the mouse tissues. The relative mRNA levels were determined using the 2−∆∆Ct method [32]. 2.7. Western Blotting The protein lysates were resolved using sodium dodecyl sulphate-polyacrilamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (Immobilon, Merck-Millipore; Burlington MA, USA, IPVH00010). The blots were blocked with 5% non-fat dry milk in Tris-buffered saline, 0.05% Tween 20 (TBST) buffer at room temperature for 1 h. The membranes were then incubated overnight with antibodies directed against PDE4B (NBP2-01171, Novus Biologicals, Minneapolis, MN, USA), PDE4D (ab171750, Abcam, Cambridge, UK) and ß-actin (A5441, Sigma-Aldrich). The appropriate horseradish peroxidase-conjugated secondary antibodies (Dako Products, Agilent, Santa Clara, CA, USA) and the LuminataTM Western horseradish peroxidase (HRP) Substrate (Immobilon, Merck-Millipore) were used to detect the bound antibodies. The size of the detected proteins was estimated using protein molecular-mass standards (Hyperpage Prestained Protein Marker; Bioline, Paris, France). ß-actin was used to verify the equal loading of the protein on each lane. 2.8. Histological and Immunohistochemical Analysis The arteries were fixed in 4% paraformaldehyde/0.1 M PBS (pH 7.4) for 24 h, and were embedded in paraffin. The tissue sections (5 µm-thick) were deparaffinised in xylene, rehy- drated in graded ethanol solutions, and treated with 3% hydrogen peroxide in methanol for 30 min in order to block endogenous peroxidase. The slides were incubated with 10% nor- mal serum, and endogenous avidin and biotin were blocked using a commercial kit (Vector Laboratories, Burlingame, CA, USA). Then, the samples were incubated overnight at 4 ◦ C with the following primary antibodies: MAC3 (sc-19991, Santa Cruz Biotechnology, Dallas, TX, USA), CD3 (A0452, Dako, Agilent Technologies), neutrophil elastase (M0752, Dako, Agilent Technologies), MCP1 (sc-1785, Santa Cruz Biotechnology), or PDE4B (NBP2-01171, Novus Biologicals). The samples were washed and then incubated with the appropriate biotinylated secondary antibodies for 1 h (Vector Laboratories,). After washing three times in PBS, the slides were incubated in avidin-biotin peroxidase complex (Vector Laboratories) for 30 min. The colour was developed by incubation with 3,30 -diaminobenzidine (DAB), and then the sections were counterstained with haematoxylin, dehydrated, cleared, and mounted. The samples in which the primary antibody was omitted were used as the negative controls. The histological characterization of the aortic samples was carried out by Antioxidants 2021, 10, 460 5 of 17 Masson trichrome staining, while the elastic fibre integrity was assessed by orcein staining (Casa Álvarez, Madrid, Spain) [27]. 2.9. In Situ Zymography The gelatinolytic activity was assessed in optimal cutting temperature compound (OCT)-embedded unfixed frozen tissue sections (8 µm-thick) using Quenched Fluorogenic dye-quenched (DQ)TM gelatin (D-12054, Invitrogen, Carlsband, CA, USA) as a fluorogenic substrate, as previously described [27]. Briefly, DQ-gelatin was dissolved in water at 1 mg/mL and then diluted 1:10 in 1% (w/v) low gelling temperature agarose (A9414, Sigma-Aldrich). The mixture was applied onto each section and cover-slipped. The samples were incubated at 4 ◦ C for 30 min in order to allow gelatin gelling, and were then maintained at room temperature for 24 h, protected from light. Fluorescein isothiocyanate (FITC) fluorescence was visualized using Leica TCS SP5 confocal microscopy (excitation wavelength: 495 nm; emission wavelength: 515 nm; 10 × 0.4 DRY UV objective for quantification and 40 × 0.4 OIL UV objective for illustration) and the Leica LAS AF Lite software (Leica Microsystems S.L.U, Barcelona, Spain). Z-stack images using four planes of the same region were captured (1024 × 1024; z-interval, 2.1 µm), and were stacked into a single image using a projection of intensity sum to quantify the fluorescence. The image processing and quantitative fluorescence analysis were performed using FIJI ImageJ software. Negative controls in which the samples were pre-incubated with 20 mM ethylenediaminetetraacetic acid (EDTA) before the addition of the labelled substrate were included [27]. 2.10. In Situ Detection of Vascular O2 ·− Production The oxidative fluorescent dye dihydroethidium (DHE, Sigma-Aldrich) was used to detect the in situ production of superoxide anions in the aortic segments [33]. In brief, unfixed aortic segments were embedded in frozen OCT and cut into 8 µm-thick sections using a cryostat microtome. Arterial segments were equilibrated for 30 min at 37 ◦ C in Krebs-HEPES buffer, and then fresh buffer containing 2 µM DHE was topically applied onto each slide, capped with a coverslip, and incubated for 40 min in a light- protected humidified chamber at 37 ◦ C. The fluorescence was visualized using Leica TCS SP5 confocal microscopy (10.0 × 0.40 DRY UV objective for quantification, and 40 × 0.4 OIL UV objective for illustration) and the Leica LAS AF Lite software (Leica Microsystems S.L.U, Barcelona, Spain). The fluorescent images were acquired in a sequential mode, using a 561nm laser (long-wavelength excitation) to detect DHE and a 405 nm (with short- wavelength excitation) laser to detect the autofluorescence of the aorta layers. Z-stack images of four optical sections in the same plane were captured (1024 × 1024; z-interval, 2.1 µm), and were stacked into a single image using a projection of intensity sum to quantify the fluorescence. The image processing and quantitative fluorescence analysis were performed using FIJI ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA) [33]. 2.11. Statistical Analysis The results are shown as the mean ± the standard error of the mean (SEM), or as boxplots. In this latter case, the upper and lower boundaries of each box display the 75th and 25th percentiles of data, respectively; the median is indicated by a horizontal line, and the minimum and maximum values are indicated as whiskers. The significant differences were analysed using one way ANOVA, two-way ANOVA with repeated measures, or two-way ANOVA followed by Bonferroni’s post-hoc tests. When the distribution of the data failed the D’Agostino-Pearson omnibus normality test, the Mann-Whitney U test with Tukey’s post-hoc test, or the Kruskal-Wallis test with Dunn’s multiple comparison post-hoc test were applied. The differences in the AAA incidence were analysed by the χ2 test. The data were analysed with GraphPad Prism version 6.01 (GraphPad Software, San Diego, USA). The differences were considered significant at p < 0.05. two-way ANOVA followed by Bonferroni’s post-hoc tests. When the distribution of the data failed the D’Agostino-Pearson omnibus normality test, the Mann-Whitney U test with Tukey’s post-hoc test, or the Kruskal-Wallis test with Dunn’s multiple comparison post-hoc test were applied. The differences in the AAA incidence were analysed by the χ2 Antioxidants 2021, 10, 460 test. The data were analysed with GraphPad Prism version 6.01 (GraphPad Software,San 6 of 17 Diego, USA). The differences were considered significant at p < 0.05. 3. Results 3. Results 3.1. Vascular PDE4B Expression Is Enhanced in Mouse and Human AAA 3.1. Vascular PDE4B Expression Is Enhanced in Mouse and Human AAA Because PDE4B and PDE4D play a major role in inflammation and oxidative stress Because PDE4B and PDE4D play a major role in inflammation and oxidative stress [34], [34], we analysed the expression of these PDE4 isoenzymes in experimental AAA induced we analysed the expression of these PDE4 isoenzymes in experimental AAA induced by by AngII infusion in ApoE −/− mice, a preclinical model which recapitulates key aspects of AngII infusion in ApoE−/− mice, a preclinical model which recapitulates key aspects of the the human disease [26]. PDE4B was significantly expressed in mouse AAA (Figure 1). human disease [26]. PDE4B was significantly expressed in mouse AAA (Figure 1). Using Using an antibody an antibody that recognizes that recognizes the conserved the conserved C termini C termini of PDE4B, of PDE4B, a major a major bandband of 64ofkDa 64 kDa significantly significantly up-regulated up-regulated in AAA in AAA was detected was detected by Western by Western blot (Figures blot (Figure 1A and1A and S1). Figure S1). Immunostaining Immunostaining located the PD4B in inflammatory cells infiltrated in aneurysmal tissues located the PD4B in inflammatory cells infiltrated in aneurysmal tissues (Figure (Figure 1B). 1B). By By contrast, contrast, the the levels levels of of PDE4D PDE4D in in the the healthy healthy aorta aorta and and AAA AAA were were similar similar (Figure S2). These data suggest that PDE4B could play a role in AAA (Figure S2). These data suggest that PDE4B could play a role in AAA development. development. Figure Figure1.1.PDE4B PDE4Bexpression expressionisisup-regulated up-regulatedin inthe theabdominal abdominalaortas aortasfrom fromAngII-infused AngII-infusedApoE −/− ApoE−/− mice. ApoE −/mice −/− were infused with AngII (1000 ng/kg/min) or saline solution for 28 days. (A) mice. ApoE − mice were infused with AngII (1000 ng/kg/min) or saline solution for 28 days. (A) The PDE4B expression in abdominal aortas from these animals was assessed by Western blot. The The PDE4B expression in abdominal aortas from these animals was assessed by Western blot. The levels of β-actin are shown as a loading control. The boxplot on the right shows the quantification levels of of β-actin the PDE4B are shown protein levels.asThe a loading control. box extends Thethe from boxplot 25th toonthe the75th right shows theand percentile, quantification the medianof isthe PDE4B protein indicated levels. The by a horizontal line.box Theextends whiskers from the 25th represent thetomaximum the 75th percentile, and minimumand the median values (Sa- is indicated line, n = 4; by Ang a horizontal II n = 6); * pline. The < 0.01 vs.whiskers saline. (B)represent the maximum Representative PDE4Band minimum values immunostaining (Saline, in these n = 4; Ang samples. TheII narrowheads = 6); * p < 0.01 vs. saline. indicate (B) Representative the PD4B-positive PDE4B cells in immunostaining aneurysmal in these tissues. Bars: samples. 20 µm. The arrowheads indicate the PD4B-positive cells in aneurysmal tissues. Bars: 20 µm. These results prompted us to address whether PDE4B expression could be similarly These regulated inresults humanprompted us to address AAA. Aneurysmal whether samples fromPDE4B expression AAA patients could beopen undergoing similarly aor- regulated in human AAA. Aneurysmal samples from AAA patients undergoing open tic repair and non-diseased abdominal aortas from organ donors were processed. The aortic repair and non-diseased abdominal aortas from organ donors were processed. The clinical and demographic data of the patients and donors are shown in Table S1. Interestingly, in agreement with the data from experimental AAA, the aneurysmal aortas expressed higher levels of PDE4B (≈3.4-fold) than the healthy vessels (Figure 2A), mainly in inflammatory cells (Figure 2B). Conversely, the PDE4D expression remained unchanged (Figure 2C). Antioxidants 2021, 10, 460 7 of 17 Antioxidants 2021, 10, x FOR PEER REVIEW 8 of 19 Figure 2. PDE4B expression is enhanced in human AAA. (A) PDE4B expression in abdominal aorta Figure from 2. PDE4B AAA patients (AAA)expression is enhanced and healthy donors (Do), assessedinbyhuman real-timeAAA. PCR. The(A) dataPDE4B are pre- expression in abdominal sented as boxplots. The box extends from the 25th to the 75th aorta from AAA patients (AAA) and healthy donors (Do), assessed percentile, and the median is indi-by real-time PCR. The data cated by a horizontal line. The whiskers represent the maximum and minimum values (DO, n = 14; are npresented AAA as vs. = 61); * p < 0.001 boxplots. Donors. (B) The box extends Representative PDE4B from the 25th to immunostaining the 75th in healthy percentile, and the median is aorta indicated from donors and byaneurysmal a horizontaltissues line. Thepatients from AAA whiskers represent (top panels; bar: 100the µm).maximum The boxed ar-and minimum values (DO, eas are shown at a higher magnification below, and the arrowheads indicate PDE4B-positive cells n = 14; (lower AAA panels; bar: n 50= 61); µm). p < 0.001 (C)* PDE4D mRNA vs.levels Donors. in these(B) Representative PDE4B immunostaining in healthy samples. aorta from donors and aneurysmal tissues from AAA patients (top panels; bar: 100 µm). The boxed areas are shown at a higher magnification below, and the arrowheads indicate PDE4B-positive cells (lower panels; bar: 50 µm). (C) PDE4D mRNA levels in these samples. 3.2. Treatment of ApoE−/− Mice with Rolipram Prevents AAA Formation Induced by AngII Infusion We sought to determine whether the up-regulation of PDE4B might play an active role in AAA pathophysiology. For this purpose, rolipram, a PDE4 selective inhibitor, was administered to AngII-infused ApoE−/− mice, and their aneurysm progression was monitored by ultrasonography throughout the experimental period. Interestingly, rolipram protected against AAA formation in this model (Figure 3). The macroscopic visualization of isolated aortas at the end of the experimental procedure showed that the abdominal aortic dilation induced by AngII was ameliorated by rolipram (Figure 3A). The weekly ultrasound monitoring of the suprarenal aortic diameter, which estimates the timeline of aneurysm development, revealed an early trend towards a lower vascular dilation in rolipram-treated mice, which was statistically significant after 2 weeks of AngII infu- sion (Figure 3B–D). This effect was associated with a significant reduction of aneurysm incidence. While 77% of AngII-infused mice developed aneurysms, this percentage de- creased to 30% after rolipram treatment (p < 0.02) (Figure 3E). Likewise, less severe forms of AngII-induced aneurysm were developed in the mice receiving rolipram (Figure 3F). Masson trichrome staining evidenced the striking disturbance of the vascular morphology exhibited by AngII-challenged mice, and how rolipram improved the vascular phenotype (Figure 3G). Furthermore, this drug attenuated the compensatory increase in aortic colla- gen I expression induced by AngII (Figure 3H). This drug affected neither the enhanced blood pressure levels (Figure S3) or cardiac hypertrophy (and the compensatory increase of systolic function) evoked by AngII (Table S2), excluding any influence of hemodynamic Antioxidants 2021, 10, 460 8 of 17 Antioxidants 2021, 10, x FOR PEER REVIEW or cardiac effects on the observed responses. Likewise, bodyweight 9 of 19 was not significantly modified by PDE4 inhibition (Table S3). Figure 3. Rolipram ameliorates AAA development in AngII-infused ApoE−/− mice. ApoE−/− mice −/− mice. ApoE−/− mice were infused with Figure 3. Rolipram ameliorates were infused withAAA development saline solution (Sal) orinAngII AngII-infused (1000 ng/kg/min)ApoEfor 28 days. The mice challenged saline solution (Sal)with or AngII (1000 AngII were ng/kg/min) treated for 28 days. or not with rolipram (ROL, 3The mice challenged mg/kg/day). with macroscopic (A) Representative AngII were treated or not with images of the aortas from each experimental group. (B) Time-course analysis of the abdominal rolipram (ROL, 3 mg/kg/day). (A) Representative macroscopic images of the aortas from each experimental group. (B) aortic diameter by ecochardiography. The data are mean ± SEM (saline-infused mice, n = 9; AngII- Time-course analysisinfused of theanimals, abdominal aortic diameter n = 13; AngII-infused by ecochardiography. mice treated with rolipram, n = 10).The data*vs. p < 0.05: are mean ± SEM (saline-infused saline- infused animals, mice, n = 9; AngII-infused mice; #vs. AngII-challenged mice. (C) Quantitative n = 13; AngII-infused mice treated analysis withofrolipram, the aortic diameter n = 10).atpthe < 0.05: * vs. saline-infused end of the experimental period (28 days). The results are expressed as boxplots. The box extends mice; # vs. AngII-challenged mice. (C) Quantitative analysis of the aortic diameter at the end from the 25th to the 75th percentile, and the median is indicated by a horizontal line. The whiskers of the experimental period (28 days). The results are expressed represent the maximum as boxplots. and minimum The box extends values. from p < 0.05: *vs. the 25thmice saline-infused to the 75th at the samepercentile, time; # and the median is vs. AngII-challenged mice. (D) Representative ultrasonographic frames after 28 days of AngII in- indicated by a horizontal line. The whiskers represent the maximum and minimum values. p < 0.05: * vs. saline-infused fusion. Transverse and longitudinal images taken at the level of the suprarenal aorta are shown. mice at the same time; The #aortic vs. diameter AngII-challenged mice. is indicated with (D) line. a yellow Representative (E) Incidence ofultrasonographic frames after 28 days of AngII AAA in each experimental infusion. Transversegroup. p < 0.05: * vs. saline-infused and longitudinal images taken mice;at#vs. theAngII-challenged mice. (F) The aorta level of the suprarenal aneurysmareseverity shown. The aortic diameter is based on the Manning scale. (G) Masson trichrome staining of the abdominal aortas from each indicated with a yellow line. (E) Incidence of AAA in each experimental group. p < 0.05: * vs. saline-infused mice; #vs. AngII-challenged mice. (F) The aneurysm severity based on the Manning scale. (G) Masson trichrome staining of the abdominal aortas from each group. Representative images are shown. Bars: 200 µm. (H) The Col1A1 mRNA levels assessed in these samples by real-time PCR. The data are presented as box and whisker plots. The box extends from the 25th to the 75th percentile, and the median is indicated by a horizontal line. The whiskers represent the maximum and minimum values (saline-infused mice n = 8; AngII-challenged mice, n = 12; AngII-challenged mice treated with rolipram, n = 8). p < 0.05; * vs. saline-infused mice; # vs. AngII-challenged mice. these samples by real-time PCR. The data are presented as box and whisker plots. The box extends from the 25th to the 75th percentile, and the median is indicated by a horizontal line. The whiskers represent the maximum and minimum values (saline-infused mice n = 8; AngII-challenged mice, n = 12; AngII-challenged mice treated with rolipram, n = 8). p < 0.05; * vs. saline-infused mice; # vs. Antioxidants AngII-challenged 2021, 10, 460 mice. 9 of 17 3.3. Rolipram Ameliorates AngII-Induced Inflammation and Oxidative Stress Rolipram exhibits immunomodulatory 3.3. Rolipram and anti-inflammatory Ameliorates AngII-Induced Inflammation andproperties [19,34,35]. Oxidative Stress Therefore, we aimed to determine whether this drug limits vascular inflammation Rolipram exhibits immunomodulatory and anti-inflammatory properties and [19,34,35]. oxidative stress, which are characteristics Therefore, of aneurysmal we aimed to determine whether disease. Thelimits this drug immunohistochem- vascular inflammation and ical analysis evidenced thatstress, oxidative rolipram whichmarkedly reduced are characteristics the infiltration of aneurysmal of The disease. inflammatory immunohistochemical cells evoked by AngII, resulting in less accumulation of macrophages (MAC3+ of analysis evidenced that rolipram markedly reduced the infiltration inflammatory cells; Fig- cells evoked by AngII, resulting in less accumulation of macrophages ure 4A), lymphocytes (CD3+ cells; Figure 4B), and neutrophils (neutrophil elastase+ cells; (MAC3+ cells; Figure 4A), lymphocytes (CD3+ cells; Figure 4B), and neutrophils (neutrophil elastase+ cells; Figure 4C) Figure 4C) into the aortic wall. Accordingly, rolipram ameliorated the vascular expression into the aortic wall. Accordingly, rolipram ameliorated the vascular expression of Emr-1 (a of Emr-1 (a macrophage marker) (Figure 5A), as well as that of monocyte chemoattractant macrophage marker) (Figure 5A), as well as that of monocyte chemoattractant protein-1 protein-1 (Mcp1/Ccl2) (Figure (Figure (Mcp1/Ccl2) 5B,C). Interestingly, the increase 5B,C). Interestingly, in superoxide the increase in superoxideanion levels anion levels detected detected in the aorta from AngII-infused mice by DHE staining was totally suppressed in the aorta from AngII-infused mice by DHE staining was totally suppressed byby rolipram rolipram treatmenttreatment (Figure (Figure 6). 6). Figure 4. Inflammatory cell infiltration in the abdominal aorta from AngII-infused ApoE−/− mice is attenuated by rolipram. Figure 4. Inflammatory cell infiltration in the abdominal aorta from AngII-infused ApoE−/− mice is ApoE−/− mice were infused with saline solution or AngII (1000 ng/kg/min) for 28 days. The mice challenged with AngII attenuated by rolipram. ApoE mice were infused with saline solution or AngII (1000 ng/kg/min) −/− were treated or not with rolipram (3 mg/kg/day). (A–C) Representative images corresponding to macrophage ((A) MAC3), for 28 days. The mice challenged with AngII were treated or not with rolipram (3 mg/kg/day). (A– lymphocyte ((B) CD3) andimages C) Representative neutrophil ((C) neutrophil corresponding elastase, ELANE) to macrophage infiltration, (A; MAC3), analysed by lymphocyte (B;immunohistochemistry. CD3) and The arrowheads indicate positive cells (Bars: 50 µm). The boxplots on the right show neutrophil (C; neutrophil elastase, ELANE) infiltration, analysed by immunohistochemistry. the quantification of the The positive cells for Mac-3, arrowheads indicate positive cells (Bars: 50 µm). The boxplots on the right show the quantification and CD3, and the neutrophil elastase per aortic section. The box extends from the 25th to the 75th percentile, the median of theispositive indicated by afor cells horizontal Mac-3, CD3,line. The andwhiskers represent the neutrophil the maximum elastase andsection. per aortic minimum Thevalues (saline-infused box extends n = 5;the mice, from AngII-challenged 25th to the 75th mice, n = 8; AngII-challenged percentile, and the medianmice treated with is indicated by arolipram, n = line. horizontal 6). InThe (A–C), p < 0.05: * vs. whiskers saline-infused mice; # vs. AngII-challenged mice. represent the maximum and minimum values (saline-infused mice, n = 5; AngII-challenged mice, n = 8; AngII-challenged mice treated with rolipram, n = 6). In (A–C), p < 0.05: * vs. saline-infused mice; # vs. AngII-challenged mice. 0, x FOR PEER REVIEW 11 of 19 Antioxidants 2021, 10, 460 10 of 17 Figure 5. Rolipram limits the inflammatory response triggered by AngII in ApoE−/− mice. ApoE−/− mice were infused withFigure 5. Rolipram or AngIIlimits the inflammatory response triggered byAngII AngII intreated ApoEor −/− mice. ApoE−/− saline solution (1000 ng/kg/min) for 28 days. Mice challenged with were not with rolipram mice were (A,B) (3 mg/kg/day). infused with (A) The Emr1 saline solution and Mcp1 or AngII (B) mRNA (1000 levels were ng/kg/min) for 28 days. analysed by real-time Mice challenged PCR (saline-infused mice n = 8; with AngII were treated or not with rolipram (3 mg/kg/day). A–B) The Emr1 (A) and Mcp1 AngII-challenged mice, n = 12; AngII-challenged mice treated with rolipram, n = 9). (C) The immunohistochemical analysis (B) mRNA of MCP1 levels in expression were analysedaortas the abdominal by real-time from eachPCR (saline-infused experimental mice n = 8;images group. Representative AngII-challenged are shown (Bar:mice, 50 µm;n = 12; AngII-challenged saline-infused mice treated mice n = 5; AngII-challenged with mice, n = rolipram, n = 9). (C) 8; AngII-challenged miceThe immunohistochemical treated with rolipram, n = 6). Theanalysis data are of MCP1 presented as boxexpression and whiskerinplots. the abdominal aortas The box extends fromfrom each the 25th experimental to the group. 75th percentile, and theRepresentative median is indicated im- by a horizontal line. The whiskers represent the maximum and minimum values. In (A–C), p < 0.05: ages are shown (Bar: 50 µm; saline-infused mice n = 5; AngII-challenged mice, n = 8; AngII-chal- * vs. saline-infused mice; # vs.lenged AngII-challenged mice.with rolipram, n = 6). The data are presented as box and whisker plots. The mice treated box extends from the 25th to the 75th percentile, and the median is indicated by a horizontal line. The whiskers represent the maximum and minimum values. In (A–C), p < 0.05: * vs. saline-infused mice; # vs. AngII-challenged mice. Antioxidants 2021,10, Antioxidants2021, 10,460 x FOR PEER REVIEW 11 12ofof17 19 Figure 6. Rolipram reduces the increased production of vascular superoxide anions induced by AngII Figure 6. Rolipram reduces the increased production of vascular superoxide anions induced by in ApoEin−/ApoE AngII − mice. The ApoE−/− mice were infused with saline solution or AngII (1000 ng/kg/min) −/− mice. The ApoE−/− mice were infused with saline solution or AngII (1000 for 28 days. The ng/kg/min) mice for 28 challenged days. The mice with AngII were challenged treated with AngIIor not treated were with rolipram (3 mg/kg/day). or not with rolipram (3Their vascular superoxide mg/kg/day). anion production Their vascular superoxide was assessed anion by DHEwas production staining in the assessed byabdominal DHE stainingaorta insections the of each experimental abdominal group. aorta sections Representative of each experimental images group.ofRepresentative the DHE staining are shown images (Bars:staining of the DHE 50 µm). are data The shownare(Bars: 50 µm). presented Theand as box data are presented whisker as box extends plots. The and whisker from plots. The the 25th tobox extends the 75th from percentile, the 25th and medianto the 75th percentile, is indicated and median by a horizontal line.isThe indicated by arepresent whiskers horizontal theline. The whiskers maximum repre- and minimum sent the values maximum and (saline-infused minimum mice values (saline-infused n = 6; AngII-challenged mice mice, n = 8;nAngII-challenged = 6; AngII-challenged mice, n with mice treated = 8; AngII-challenged mice treated with rolipram, n = 7). p < 0.01: * vs. saline-infused mice; # vs. AngII- rolipram, n = 7). p < 0.01: * vs. saline-infused mice; # vs. AngII-challenged mice. challenged mice. 3.4. Rolipram Attenuates the Destructive Vascular Remodelling Triggered by AngII, and Limits MMP Expression 3.4. Rolipram and Activity Attenuates the Destructive Vascular Remodelling Triggered by AngII, and Limits MMP WeExpression next sought andtoActivity determine whether rolipram could impact on AngII-induced vascu- lar remodelling. We next soughtImportantly, rolipram to determine preserved whether the elastin rolipram integrity, could impact onwhich was severely AngII-induced vas- disrupted by AngII-infusion. cular remodelling. Orcein Importantly, stainingpreserved rolipram shows that PDE4 the inhibition elastin reduced integrity, whichthewashigh se- number of elastin by verely disrupted fibre breaks detected AngII-infusion. in challenged Orcein mice (Figure staining shows that PDE47A,B). In conjunction inhibition reduced with the extensive the high number of disruption of elastic elastin fibre breakslamina evoked detected by AngII, mice in challenged a marked up-regulation (Figure of 7A,B). In con- both the Mmp-2 and Mmp-9 mRNA levels was observed. Rolipram partially junction with the extensive disruption of elastic lamina evoked by AngII, a marked up- abrogated the AngII-mediated regulation of both theinduction Mmp-2 and of Mmp-2 Mmp-9 mRNAexpression (Figure levels 8A), while was observed. a trendpartially Rolipram for the normalization of the Mmp-9 mRNA levels was detected, although the abrogated the AngII-mediated induction of Mmp-2 expression (Figure 8A), while a trenddifferences did not reach for thestatistical significance normalization (Figure mRNA of the Mmp-9 8B). Interestingly, in situ zymography levels was detected, although the in differences the aortic sections did not revealed that thesignificance reach statistical PDE4B inhibitor (Figureabrogated the enhanced 8B). Interestingly, MMP in situ activity induced zymography in the by AngII (Figure 8C). aortic sections revealed that the PDE4B inhibitor abrogated the enhanced MMP activity induced by AngII (Figure 8C). Antioxidants 2021, 10, 460 12 of 17 Antioxidants 2021, 10, x FOR PEER REVIEW 13 of 19 Figure Figure 7. Rolipram 7. Rolipram attenuates theattenuates the induced vascular remodelling vascular remodelling by AngII in ApoE−/− mice.induced ApoE−/− by AngII in ApoE−/− mice. mice were infused with saline solution or AngII (1000 ng/kg/min) for 28 days. The mice challenged ApoE − / − mice orwere infused with saline with AngII were treated not with rolipram (3 mg/kg/day). (A)solution orimages Representative AngII (1000 ng/kg/min) for 28 days. The showing the structure of the elastin fibres in the abdominal aortas after orcein staining. The arrowheads mice challenged with AngII were treated or not with rolipram indicate elastin ruptures (Bars: 100 µm). (B) The boxplot shows the quantitative analysis of the (3 mg/kg/day). (A) Representative images number showing of ruptures the structure in the elastin ofsection. fibers per aortic the elastin fibres The box extends inthethe from 25thabdominal to the aortas after orcein staining. 75th percentile, and the median is indicated by a horizontal line. The whiskers represent the maxi- Theand mum arrowheads minimum valuesindicate elastin (saline-infused mice n ruptures (Bars:mice, = 6; AngII-challenged 100nµm). (B) The boxplot shows the quantitative = 12; AngII-chal- lenged mice treated with rolipram, n = 7). p < 0.05: * vs. saline-infused mice; # vs. AngII-challenged analysis of the number of ruptures in the elastin fibers per aortic section. The box extends from mice. the 25th to the 75th percentile, and the median is indicated by a horizontal line. The whiskers represent the maximum and minimum values (saline-infused mice n = 6; AngII-challenged mice, n = 12; AngII-challenged mice treated with rolipram, n = 7). p < 0.05: * vs. saline-infused mice; # vs. Antioxidants 2021, 10, x FOR PEER REVIEW 14 of 19 AngII-challenged mice. Figure Figure8. Rolipram attenuates the 8. Rolipram increase in MMP attenuates theactivity increasetriggered by AngII activity in MMP triggered by AngII in ApoE−/− mice. in ApoE−/− mice. ApoE−/− mice were infused with saline solution or AngII (1000 ng/kg/min) for 28 days. The mice − / − ApoE with mice were infused with saline solution(A,B) or AngII (1000 ng/kg/min) for 28 days. The mice challenged AngII were treated or not with rolipram (3 mg/kg/day). The Mmp2 (A) and Mmp9 (B) mRNA levels were analysed by real-time PCR. The data are presented as box and challenged with AngII were treated or not with rolipram whisker plots. The box extends from the 25th to the 75th percentile, and the median is indicated by (3 mg/kg/day). (A,B) The Mmp2 (A) a and Mmp9 horizontal (B)whiskers line. The mRNA levels represent the were maximum analysed and minimum byvalues real-time PCR.mice (saline-infused The data are presented as box and n = 8; AngII-challenged mice, n = 12; AngII-challenged mice treated with rolipram, n = 9). (C) The whisker MMP activityplots. analysedThe by in box extends from situ zymography the 25th in the abdominal to the aortas from 75th percentile, and the median is indicated by a each experimental group. Representive images are shown (Bars: 50 µm). The boxplot on the right shows the quantifi- horizontal line. The whiskers represent the maximum and cation of the MMP activity (saline-infused mice n = 5; AngII-challenged mice, n = 6; AngII-chal- minimum values (saline-infused mice n = 8; AngII-challenged mice, n = 12; AngII-challenged mice AngII- lenged mice treated with rolipram, n = 6). In A to C, p < 0.05: * vs. saline-infused mice; # vs. treated with rolipram, n = 9). (C) The challenged mice. MMP activity analysed by in situ zymography in the abdominal aortas from each experimental group. 4.Representive Discussion images are shown (Bars: 50 µm). The boxplot on the right shows the quantification of theThere MMP is currently an imperative need for the identification of effective pharmacolog- activity (saline-infused mice n = 5; AngII-challenged mice, n = 6; AngII-challenged mice ical therapies which allow the stabilization or regression of established AAA; however, treated the with insufficient rolipram, n comprehension of =the6). complex to C, p < 0.05: *mechanisms In A pathophysiological vs. saline-infused underly- mice; # vs. AngII-challenged ing this mice. disease is a serious pitfall for this purpose. In the search for new pharmacological targets in this disease, we have reported, for the first time, that vascular PDE4B expression is upregulated in both human and experimental AAA. Furthermore, the pharmacological inhibition of PDE4 attenuates vascular inflammation and oxidative stress, preserving vas- cular integrity, thereby preventing aneurysm progression and reducing AAA incidence and severity. The PDE4 family encompasses a group of cAMP-specific PDEs encoded by four dif- ferent genes, PDE4A–D, with each of them generating long and short variants due to the Antioxidants 2021, 10, 460 13 of 17 4. Discussion There is currently an imperative need for the identification of effective pharmacological therapies which allow the stabilization or regression of established AAA; however, the insufficient comprehension of the complex pathophysiological mechanisms underlying this disease is a serious pitfall for this purpose. In the search for new pharmacological targets in this disease, we have reported, for the first time, that vascular PDE4B expression is upregulated in both human and experimental AAA. Furthermore, the pharmacological inhibition of PDE4 attenuates vascular inflammation and oxidative stress, preserving vascular integrity, thereby preventing aneurysm progression and reducing AAA incidence and severity. The PDE4 family encompasses a group of cAMP-specific PDEs encoded by four dif- ferent genes, PDE4A–D, with each of them generating long and short variants due to the use of alternate promoters or differential splicing [13]. Similarly to other PDEs, the PDE4 subtypes differ in their regulation, tissue-expression, subcellular distribution, and protein interaction. In fact, selective knockdown evidences that each subtype undertakes specific and non-redundant biological functions [36]. In particular, PDE4B, which is the predominant subtype expressed in inflammatory cells, critically influences immune and inflammatory responses, and constitutes an appealing target for the development of new therapeutic agents against a wide range of disorders, specifically, inflammatory and autoim- mune pathologies [19,34]. However, despite the facts that AAA is a chronic inflammatory disease, and innate or adaptive immune responses have emerged as promising targets to limit human AAA [4], the potential of PDE4 enzymes as therapeutic targets for this disorder has not been previously addressed. As a first approach to test this, we analysed PDE4B expression in aneurysmal aorta from AngII-infused ApoE−/− mice. The accumulating evidence supports the supposition that the overactivation of the renin–angiotensin system (RAS) plays a critical role in the onset and progression of human AAA [37]. Consistently, the infusion of the main RAS effector (AngII) into ApoE−/− mice induces the formation of AAA which resemble those found in humans, characterized by exacerbated inflammation, oxidative stress, and ECM disorganization, as we and others have reported [24,26,27,31]. Interestingly, we showed that, while PDE4B was hardly detected in the vascular wall from the control animals (saline-infused ApoE−/− mice), it was markedly increased in AAA from the AngII-challenged mice located in inflammatory cells. In agreement with this, PDE4B, undetectable in healthy human aorta, was also significantly upregulated in human AAA, exhibiting a strong immunostaining in the inflammatory cells. Instead, the PDE4D expression was similar in healthy and diseased aorta, either in humans or in the murine model. These data suggest that PDE4B could be involved in aneurysm pathogenesis, while PDE4D might have a minor role in this disease. Although aging is an important risk factor in the pathogenesis of AAA, it should be highlighted that, in our study, there were no significant differences in age between the donors and the AAA patients (p = 0.114; Mann Whitney U test). Therefore, it seems that we can rule out that differences in age account for the increased expression of PDE4B observed in the human aneurysmal samples. Regardless of whether or not age modulates PDE4B in the vasculature, therapeutic strategies impact- ing on cAMP levels have been proposed as novel approaches for age-related vascular diseases [38]. In view of the potential contribution of the up-regulation of PDE4B to AAA pathogene- sis, we evaluated the impact of PDE4 inhibition on the formation of AAA induced by AngII. We found that the systemic administration of rolipram, the prototypical selective PDE4 inhibitor, to AngII-challenged mice limited the growth of the aortic diameter, as assessed by ultrasonography, while reducing the incidence and severity of aneurysm. Preclinical studies analyzing the consequences of PDE4 blockade on AAA expansion had not been previously carried out, although the inhibition of PDE4 activity limits the progression of cerebral aneurysms [39], a disorder which shares some pathological characteristics with AAA, such as ECM proteolysis and inflammation. It has been previously reported that rolipram could impact on vascular contractility, inducing vascular relaxation in specific Antioxidants 2021, 10, 460 14 of 17 vascular beds [40], whereas the consequences of PDE4 inhibition on cardiac function and hypertrophy remains unclear [41,42]. Despite this potential vasodilator effect, in our experimental conditions, rolipram did not affect blood pressure levels, and the cardiac hypertrophy and systolic function in AngII-infused mice remained unchanged. Therefore, we ruled out that the beneficial action of rolipram on AAA could be attributable to blood pressure- or cardiac-dependent mechanisms. The attenuation of the AngII-mediated increase of the aortic diameter observed in mice treated with rolipram results from combined anti-inflamatory and anti-oxidant effects, which provide an overall vasoprotection and th eimprovement of vascular remodelling. Rolipram significantly attenuated the oxidative stress induced by AngII, in agreement with previous studies showing antioxidant effects derived from PDE4 inhibition [21–23]. Remarkably, the rolipram-mediated inhibition of AngII-induced AAA was also associ- ated with a potent anti-inflammatory effect, as it strongly reduced the local inflammatory infiltrate and the expression of inflammatory markers. It has been extensively reported that PDE4 inhibition exerts vasculoprotective actions against inflammation by inhibiting leukocyte–endothelial cell interactions, attenuating the platelet-mediated neutrophil re- cruitment at the sites of vascular injury, and improving the endothelial barrier integrity in different pathological settings [43–46]. Furthermore, the decrease of MMP expression and the consequent reduction of its activity, likely secondary to the decrease in MMP-producing inflammatory cells reported here, could account, at least in part, for the beneficial effect exerted by this drug on vascular integrity. Indeed, rolipram preserved the disruption of elastin fibres which is characteristic of the aneurysmal process. The impact of PDE4 inhibitors on MMP expression and/or activity has been previously reported, mostly in the context of pulmonary diseases, in which these drugs abolish the enhanced elastolytic activity driven by inflammatory stimulus [47] and attenuate MMP-2 and/or MMP-9 lev- els [48,49]. Taken together, therefore, our results and those described above suggest that the selective blockade of PDE4 activity could be a promising therapeutic strategy for a myriad of vascular pathologies with an inflammatory component, from AAA to atherosclerosis restenosis and venous thrombosis [43–46]. In fact, this could explain the low rate of cardio- vascular events reported in patients with chronic obstructive pulmonary disease receiving the selective PDE4 inhibitor roflumilast [50]. Therefore, our study shows the preventive effect of rolipram treatment in the forma- tion of experimental AAA. Further studies should be conducted in order to assess the potential of this drug to slow or arrest the growth of pre-established AAA, an experi- mental approach which more closely resembles the medical condition of AAA patients. Finally, it should be noted that rolipram shows high affinity for PDE4B; however, it exhibits side-effects, such as emesis, due to the inhibition of the PDE4D subtype. Owing to their potential anti-inflammatory activity, there is an active search of novel and selective PDE4B inhibitors [19,51] which avoid these PDE4D-related side-effects. Although several candi- dates have been developed, they have neither been tested in chronic preclinical studies (except for inhaled administration against respiratory diseases) nor incorporated into clinical practice. We cannot exclude that the inhibition of other PDE4 enzymes besides PDE4B might also account for the rolipram-mediated inhibition of aneurysm formation reported here, in particular PDE4D, which represents the major PDE4 variant in VSMC, and the second one in inflammatory cells behind PDE4B [18]. However, in view of the specific up-regulation of PDE4B in both human AAA and in a preclinical model, it could be hypothesized that selective PDE4B inhibitors might benefit AAA patients, and would constitute a promising therapeutic avenue for this disease. In conclusion, inflammatory cell depletion or the blockade of cytokine production/ activity has consistently shown beneficial effects preventing the formation of experimental aneurysms [52,53]. Our results provide an additional proof-of-concept supporting the targeting of inflammation/oxidative stress in general, and PDE4 in particular, as a thera- peutic strategy for AAA. Novel specific PDE4B inhibitors would constitute a promising therapeutic avenue to fight against AAA. Antioxidants 2021, 10, 460 15 of 17 Supplementary Materials: The following are available online at https://www.mdpi.com/2076-392 1/10/3/460/s1, Figure S1: Original blots shown in Figure 1; Figure S2: PDE4D expression analysis in abdominal aortas from Ang II-infused ApoE−/− mice; Figure S3: Blood pressure values in saline or AngII-infused ApoE−/− mice treated or not with rolipram. Table S1: Patients’ and donors’ clinical features; Table S2: Systolic function in saline or AngII-infused ApoE−/− mice treated or not with rolipram; Table S3: Bodyweight values. Author Contributions: C.R. and J.M.-G. designed and supervised the study, interpreted the data, and wrote the manuscript. S.V., L.P., M.G., M.O., V.A., S.A. and L.C. participated in the in vivo studies, and in the analyses and interpretation of data. S.V., L.P. performed the immunohistochemical, Western blot, and expression analysis, and analysed and interpreted the data. M.S. and M.C. participated in the human sample and demographic data collection. All of the authors were involved in writing the paper. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by Instituto de Salud Carlos III (ISCIII) [PI18/0919], the Spanish Ministerio de Ciencia e Innovación (MICINN) [RTI2018-094727-B-100 and PID2019-108489RB-100], the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR) [2017-SGR-00333], Sociedad Española de Arteriosclerosis [Beca FEA 2020 Investigación Básica], and Sociedad Española de Cardi- ología [SEC/FEC-INV-BAS 20/005]. The study was co-founded by Fondo Europeo de Desarrollo Regional (FEDER), “A way to make Europe”. L.C. was supported by a FI (AGAUR). The CNIC is supported by MICINN, ISCIII, and the Pro-CNIC Foundation. Institutional Review Board Statement: The use of human discarded tissue was performed in ac- cordance with the Declaration of Helsinki of 1975, revised in 2013, and approved by the HSCSP Ethics Committee (IIBSP-NET-2019-40). The animal handling and disposal were performed in accor- dance with the principles and guidelines established by the Spanish Policy for Animal Protection RD53/2013 and the European Union Directive 2010/63/UE. All of the procedures were approved by the local ethical committee (Ref 10750; Law 5/June 21, 1995; Generalitat de Catalunya). Informed Consent Statement: Written informed consent was obtained from all of the patients or legal representatives. Data Availability Statement: The data presented in this study are available on request from the corresponding author. Acknowledgments: We thank Joan Repullés for the technical assistance and Cristina González for the management of the mouse colonies. Conflicts of Interest: The authors declare no conflict of interest. References 1. Nordon, I.M.; Hinchliffe, R.J.; Loftus, I.M.; Thompson, M.M. Pathophysiology and epidemiology of abdominal aortic aneurysms. Nat. Rev. Cardiol. 2011, 8, 92–102. [CrossRef] 2. Kent, K.C. Clinical practice. Abdominal aortic aneurysms. N. Engl. J. Med. 2014, 371, 2101–2108. [CrossRef] 3. GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: A systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015, 385, 117–171. [CrossRef] 4. Golledge, J. Abdominal aortic aneurysm: Update on pathogenesis and medical treatments. Nat. Rev. Cardiol. 2019, 16, 225–242. [CrossRef] 5. Chaikof, E.L.; Dalman, R.L.; Eskandari, M.K.; Jackson, B.M.; Lee, W.A.; Mansour, M.A.; Mastracci, T.M.; Mell, M.; Murad, M.H.; Nguyen, L.L.; et al. The Society for Vascular Surgery practice guidelines on the care of patients with an abdominal aortic aneurysm. J. Vasc. Surg. 2018, 67, 2–77.e2. [CrossRef] [PubMed] 6. Kuivaniemi, H.; Ryer, E.J.; Elmore, J.R.; Tromp, G. Understanding the pathogenesis of abdominal aortic aneurysms. Expert Rev. Cardiovasc. Ther. 2015, 13, 975–987. [CrossRef] [PubMed] 7. Golledge, J.; Muller, J.; Daugherty, A.; Norman, P. Abdominal aortic aneurysm: Pathogenesis and implications for management. Arterioscler. Thromb. Vasc. Biol. 2006, 26, 2605–2613. [CrossRef] 8. Camacho, M.; Dilmé, J.; Solà-Villà, D.; Rodríguez, C.; Bellmunt, S.; Siguero, L.; Alcolea, S.; Romero, J.M.; Escudero, J.R.; Martínez- González, J.; et al. Microvascular COX-2/mPGES-1/EP-4 axis in human abdominal aortic aneurysm. J. Lipid. Res. 2013, 54, 3506–3515. [CrossRef] 9. Torres-Fonseca, M.; Galan, M.; Martinez-Lopez, D.; Cañes, L.; Roldan-Montero, R.; Alonso, J.; Reyero-Postigo, T.; Orriols, M.; Mendez-Barbero, N.; Sirvent, M.; et al. Pathophisiology of abdominal aortic aneurysm: Biomarkers and novel therapeutic targets. Clin. Investig. Arterioscler. 2019, 31, 166–177. Antioxidants 2021, 10, 460 16 of 17 10. Furusho, A.; Aoki, H.; Ohno-Urabe, S.; Nishihara, M.; Hirakata, S.; Nishida, N.; Ito, S.; Hayashi, M.; Imaizumi, T.; Hiromatsu, S.; et al. Involvement of B cells, immunoglobulins, and syk in the pathogenesis of abdominal aortic aneurysm. J. Am. Heart Assoc. 2018, 7, e007750. [CrossRef] 11. Zhou, Y.; Wu, W.; Lindholt, J.S.; Sukhova, G.K.; Libby, P.; Yu, X.; Shi, G.P. Regulatory T cells in human and angiotensin II-induced mouse abdominal aortic aneurysms. Cardiovasc. Res. 2015, 107, 98–107. [CrossRef] [PubMed] 12. Tsuruda, T.; Kato, J.; Hatakeyama, K.; Kojima, K.; Yano, M.; Yano, Y.; Nakamura, K.; Nakamura-Uchiyama, F.; Matsushima, Y.; Imamura, T.; et al. Adventitial mast cells contribute to pathogenesis in the progression of abdominal aortic aneurysm. Circ. Res. 2008, 102, 1368–1377. [CrossRef] [PubMed] 13. Omori, K.; Kotera, J. Overview of PDEs and their regulation. Circ. Res. 2007, 100, 309–327. [CrossRef] [PubMed] 14. Keravis, T.; Lugnier, C. Cyclic nucleotide phosphodiesterase (PDE) isozymes as targets of the intracellular signalling network: Benefits of PDE inhibitors in various diseases and perspectives for future therapeutic developments. Br. J. Pharmacol. 2012, 165, 1288–1305. [CrossRef] [PubMed] 15. Bender, A.T.; Beavo, J.A. Cyclic nucleotide phosphodiesterases: Molecular regulation to clinical use. Pharmacol. Rev. 2006, 58, 488–520. [CrossRef] [PubMed] 16. Bobin, P.; Belacel-Ouari, M.; Bedioune, I.; Zhang, L.; Leroy, J.; Leblais, V.; Fischmeister, R.; Vandecasteele, G. Cyclic nucleotide phosphodiesterases in heart and vessels: A therapeutic perspective. Arch. Cardiovasc. Dis. 2016, 109, 431–443. [CrossRef] 17. Fertig, B.A.; Baillie, G.S. PDE4-mediated cAMP signalling. J. Cardiovasc. Dev. Dis. 2018, 5, 8. [CrossRef] [PubMed] 18. Houslay, M.D.; Baillie, G.S.; Maurice, D.H. cAMP-Specific phosphodiesterase-4 enzymes in the cardiovascular system: A molecular toolbox for generating compartmentalized cAMP signaling. Circ. Res. 2007, 100, 950–966. [CrossRef] 19. Azam, M.A.; Tripuraneni, N.S. Selective phosphodiesterase 4B inhibitors: A review. Sci. Pharm. 2014, 82, 453–481. [CrossRef] [PubMed] 20. Houslay, M.D.; Adams, D.R. PDE4 cAMP phosphodiesterases: Modular enzymes that orchestrate signalling cross-talk, desensiti- zation and compartmentalization. Biochem. J. 2003, 370, 1–18. [CrossRef] 21. Xu, B.; Qin, Y.; Li, D.; Cai, N.; Wu, J.; Jiang, L.; Jie, L.; Zhou, Z.; Xu, J.; Wang, H. Inhibition of PDE4 protects neurons against oxygen-glucose deprivation-induced endoplasmic reticulum stress through activation of the Nrf-2/HO-1 pathway. Redox Biol. 2020, 28, 101342. [CrossRef] 22. Zhong, J.; Yu, H.; Huang, C.; Zhong, Q.; Chen, Y.; Xie, J.; Zhou, Z.; Xu, J.; Wang, H. Inhibition of phosphodiesterase 4 by FCPR16 protects SH-SY5Y cells against MPP+-induced decline of mitochondrial membrane potential and oxidative stress. Redox Biol. 2018, 16, 47–58. [CrossRef] 23. Bhat, A.; Tan, V.; Heng, B.; Lovejoy, D.B.; Sakharkar, M.K.; Essa, M.M.; Chidambaram, S.B.; Guillemin, G.J. Roflumilast, a cAMP-specific phosphodiesterase-4 inhibitor, reduces oxidative stress and improves synapse functions in human cortical neurons exposed to the excitotoxin quinolinic acid. ACS Chem. Neurosci. 2020, 11, 4405–4415. [CrossRef] 24. Orriols, M.; Varona, S.; Martí-Pàmies, I.; Galán, M.; Guadall, A.; Escudero, J.R.; Martín-Ventura, J.L.; Camacho, M.; Vila, L.; Martínez-González, J.; et al. Down-regulation of Fibulin-5 is associated with aortic dilation: Role of inflammation and epigenetics. Cardiovasc. Res. 2016, 110, 431–442. [CrossRef] [PubMed] 25. Alonso, J.; Galán, M.; Martí-Pàmies, I.; Romero, J.M.; Camacho, M.; Rodríguez, C.; Martínez-González, J. NOR-1/NR4A3 regulates the cellular inhibitor of apoptosis 2 (cIAP2) in vascular cells: Role in the survival response to hypoxic stress. Sci. Rep. 2016, 6, 34056. [CrossRef] 26. Daugherty, A.; Manning, M.W.; Cassis, L.A. Angiotensin II promotes atherosclerotic lesions and aneurysms in apolipoprotein E-deficient mice. J. Clin. Investig. 2000, 105, 1605–1612. [CrossRef] [PubMed] 27. Cañes, L.; Martí-Pàmies, I.; Ballester-Servera, C.; Alonso, J.; Serrano, E.; Briones, A.M.; Rodríguez, C.; Martínez-González, J. High neuron derived orphan receptor-1 (NOR-1) expression strengthens the vascular wall response to angiotensin II leading to aneurysm formation in mice. Hypertension 2021, 77, 557–570. [CrossRef] 28. Cañes, L.; Martí-Pàmies, I.; Ballester-Servera, C.; Herraiz-Martínez, A.; Alonso, J.; Galán, M.; Nistal, J.F.; Muniesa, P.; Osada, J.; Hove-Madsen, L.; et al. Neuron-derived orphan receptor-1 modulates cardiac gene expression and exacerbates angiotensin II-induced cardiac hypertrophy. Clin. Sci. 2020, 134, 359–377. [CrossRef] 29. Galán, M.; Varona, S.; Guadall, A.; Orriols, M.; Navas, M.; Aguiló, S.; de Diego, A.; Navarro, M.A.; García-Dorado, D.; Rodríguez- Sinovas, A.; et al. Lysyl oxidase overexpression accelerates cardiac remodeling and aggravates angiotensin II-induced hypertrophy. FASEB J. 2017, 31, 3787–3799. [CrossRef] [PubMed] 30. Xiao, L.; O’Callaghan, J.P.; O’Donnell, J.M. Effects of repeated treatment with phosphodiesterase-4 inhibitors on cAMP signaling, hippocampal cell proliferation, and behavior in the forced-swim test. J. Pharmacol. Exp. Ther. 2011, 338, 641–647. [CrossRef] 31. Manning, M.W.; Cassis, L.A.; Daugherty, A. Differential effects of doxycycline, a broadspectrum matrix metalloproteinase inhibitor, on angiotensin II-induced atherosclerosis and abdominal aortic aneurysms. Arterioscler. Thromb. Vasc. Biol. 2003, 23, 483–488. [CrossRef] [PubMed] 32. Jover, E.; Silvente, A.; Marín, F.; Martínez-González, J.; Orriols, M.; Martinez, C.M.; Puche, C.M.; Valdés, M.; Rodriguez, C.; Hernández-Romero, D. Inhibition of enzymes involved in collagen cross-linking reduces vascular smooth muscle cell calcification. FASEB J. 2018, 32, 4459–4469. [CrossRef] [PubMed] Antioxidants 2021, 10, 460 17 of 17 33. Martínez-Revelles, S.; García-Redondo, A.B.; Avendaño, M.S.; Varona, S.; Palao, T.; Orriols, M.; Roque, F.R.; Fortuño, A.; Touyz, R.M.; Martínez-González, J.; et al. Lysyl oxidase induces vascular oxidative stress and contributes to arterial stiffness and abnormal elastin structure in hypertension: Role of p38MAPK. Antioxid. Redox Signal. 2017, 27, 379–397. [CrossRef] 34. Wang, P.; Ohleth, K.M.; Wu, P.; Billah, M.M.; Egan, R.W. Phosphodiesterase 4B2 is the predominant phosphodiesterase species and undergoes differential regulation of gene expression in human monocytes and neutrophils. Mol. Pharmacol. 1999, 56, 170–174. [CrossRef] 35. Castro, A.; Jerez, M.J.; Gil, C.; Martinez, A. Cyclic nucleotide phosphodiesterases and their role in immunomodulatory responses: Advances in the development of specific phosphodiesterase inhibitors. Med. Res. Rev. 2005, 25, 229–244. [CrossRef] 36. Richter, W.; Menniti, F.S.; Zhang, H.T.; Conti, M. PDE4 as a target for cognition enhancement. Expert. Opin. Ther. Targets 2013, 17, 1011–1027. [CrossRef] 37. Hackam, D.G.; Thiruchelvam, D.; Redelmeier, D.A. Angiotensin-converting enzyme inhibitors and aortic rupture: A population- based case-control study. Lancet 2006, 368, 659–665. [CrossRef] 38. Kiss, T.; Giles, C.B.; Tarantini, S.; Yabluchanskiy, A.; Balasubramanian, P.; Gautam, T.; Csipo, T.; Nyúl-Tóth, Á.; Lipecz, A.; Szabo, C.; et al. Nicotinamide mononucleotide (NMN) supplementation promotes anti-aging miRNA expression profile in the aorta of aged mice, predicting epigenetic rejuvenation and anti-atherogenic effects. Geroscience 2019, 41, 419–439. [CrossRef] 39. Yagi, K.; Tada, Y.; Kitazato, K.T.; Tamura, T.; Satomi, J.; Nagahiro, S. Ibudilast inhibits cerebral aneurysms by down-regulating inflammation-related molecules in the vascular wall of rats. Neurosurgery 2010, 66, 551–559. [CrossRef] [PubMed] 40. Komas, N.; Lugnier, C.; Stoclet, J.C. Endothelium-dependent and independent relaxation of the rat aorta by cyclic nucleotide phosphodiesterase inhibitors. Br. J. Pharmacol. 1991, 104, 495–503. [CrossRef] [PubMed] 41. Rao, Y.J.; Xi, L. Pivotal effects of phosphodiesterase inhibitors on myocyte contractility and viability in normal and ischemic hearts. Acta Pharmacol. Sin. 2009, 30, 1–24. [CrossRef] [PubMed] 42. Eschenhagen, T. PDE4 in the human heart—Major player or little helper? Br. J. Pharmacol. 2013, 169, 524–527. [CrossRef] [PubMed] 43. Totani, L.; Piccoli, A.; Dell’Elba, G.; Concetta, A.; Di Santo, A.; Martelli, N.; Federico, L.; Pamuklar, Z.; Smyth, S.S.; Evangelista, V. Phosphodiesterase type 4 blockade prevents platelet-mediated neutrophil recruitment at the site of vascular injury. Arterioscler. Thromb. Vasc. Biol. 2014, 34, 1689–1696. [CrossRef] [PubMed] 44. Sanz, M.J.; Cortijo, J.; Taha, M.A.; Cerdá-Nicolás, M.; Schatton, E.; Burgbacher, B.; Klar, J.; Tenor, H.; Schudt, C.; Issekutz, A.C.; et al. Roflumilast inhibits leukocyte-endothelial cell interactions, expression of adhesion molecules and microvascular permeability. Br. J. Pharmacol. 2007, 152, 481–492. [CrossRef] [PubMed] 45. Souza, D.G.; Cassali, G.D.; Poole, S.; Teixeira, M.M. Effects of inhibition of PDE4 and TNF-alpha on local and remote injuries following ischaemia and reperfusion injury. Br. J. Pharmacol. 2001, 134, 985–994. [CrossRef] [PubMed] 46. Baumer, Y.; McCurdy, S.; Weatherby, T.M.; Mehta, N.N.; Halbherr, S.; Halbherr, P.; Yamazaki, N.; Boisvert, W.A. Hyperlipidemia- induced cholesterol crystal production by endothelial cells promotes atherogenesis. Nat. Commun. 2017, 8, 1129. [CrossRef] [PubMed] 47. Johnson, F.J.; Reynolds, L.J.; Toward, T.J. Elastolytic activity and alveolar epithelial type-1 cell damage after chronic LPS inhalation: Effects of dexamethasone and rolipram. Toxicol. Appl. Pharmacol. 2005, 207, 257–265. [CrossRef] 48. Growcott, E.J.; Spink, K.G.; Ren, X.; Afzal, S.; Banner, K.H.; Wharton, J. Phosphodiesterase type 4 expression and anti-proliferative effects in human pulmonary artery smooth muscle cells. Respir. Res. 2006, 7, 9. [CrossRef] [PubMed] 49. Martin-Chouly, C.A.; Astier, A.; Jacob, C.; Pruniaux, M.P.; Bertrand, C.; Lagente, V. Modulation of matrix metalloproteinase production from human lung fibroblasts by type 4 phosphodiesterase inhibitors. Life Sci. 2004, 75, 823–840. [CrossRef] 50. White, W.B.; Cooke, G.E.; Kowey, P.R.; Calverley, P.M.; Bredenbröker, D.; Goehring, U.M.; Zhu, H.; Lakkis, H.; Mosberg, H.; Rowe, P.; et al. Cardiovascular safety in patients receiving roflumilast for the treatment of COPD. Chest 2013, 144, 758–765. [CrossRef] 51. Lourenço, E.M.G.; Fernandes, J.M.; Carvalho, V.F.; Grougnet, R.; Martins, M.A.; Jordão, A.K.; Zucolotto, S.M.; Barbosa, E.G. Identification of a selective PDE4B inhibitor from bryophyllum pinnatum by target fishing study and in vitro evaluation of quercetin 3-O-α-L-arabinopyranosyl-(1→2)-O-α-L-rhamnopyranoside. Front. Pharmacol. 2020, 10, 1582. [CrossRef] [PubMed] 52. Eliason, J.L.; Hannawa, K.K.; Ailawadi, G.; Sinha, I.; Ford, J.W.; Deogracias, M.P.; Roelofs, K.J.; Woodrum, D.T.; Ennis, T.L.; Henke, P.K.; et al. Neutrophil depletion inhibits experimental abdominal aortic aneurysm formation. Circulation 2005, 112, 232–240. [CrossRef] [PubMed] 53. Sun, J.; Sukhova, G.K.; Yang, M.; Wolters, P.J.; MacFarlane, L.A.; Libby, P.; Sun, C.; Zhang, Y.; Liu, J.; Ennis, T.L.; et al. Mast cells modulate the pathogenesis of elastase-induced abdominal aortic aneurysms in mice. J. Clin. Investig. 2007, 117, 3359–3368. [CrossRef] [PubMed]