Sildenafil (Viagra) Aggravates the Development of Experimental Abdominal Aortic Aneurysm

Abstract

Background

cGMP‐hydrolyzing phosphodiesterase type 5 (PDE5) regulates vascular smooth muscle cell (SMC) contraction by antagonizing cGMP‐dependent protein kinase I (PKGI)–dependent SMC relaxation. SMC contractile dysfunction is implicated in the pathogenesis of aortic aneurysm. PDE5 inhibitors have been used for treating erectile dysfunction, such as drug Viagra (sildenafil). However, a few clinical cases have reported the association of Viagra usage with aortic dissection, and reduced PDE5A expression was found in human aortic aneurysm tissues. Therefore, we aimed to investigate the effect of sildenafil on experimental abdominal aortic aneurysm (AAA), the most common form of aortic aneurysm in elderly men.

Methods and Results

AAA was induced in C57BL/6J male mice by periaortic elastase in combination with blocking elastin/collagen formation via 3‐aminopropionitrile fumarate salt for 35 days. PDE5A protein levels detected by immunostaining were significantly reduced in mouse AAA. Sildenafil application in drinking water significantly aggravated aortic wall dilation and elastin degradation with pre‐existing moderate AAA. The phosphorylation level of myosin light chain 2 at Ser19, a biochemical marker of SMC contraction, was significantly reduced by sildenafil in AAA. Proximity ligation assay further revealed that the interaction between cGMP and PKGI was significantly increased by sildenafil in AAA, suggesting an elevation of PKGI activation in AAA.

Conclusions

Sildenafil treatment aggravated the degradation of elastin fibers and progression of experimental AAA by dysregulating cGMP and contractile signaling in SMCs. Our findings may raise the caution of clinical usage of Viagra in aneurysmal patients.

In the healthy aorta wall, the media layer is composed of contractile smooth muscle cells (SMCs) and elastin fibers, which maintain viscoelastic property of the aorta. Aortic aneurysm (AA) is the permanent dilatation of the aorta resulting from progressive wall degeneration by loss of medial SMC and elastic matrix degradation, ¹ which eventually leads to aortic rupture. Available evidence has suggested that SMC contractile dysfunction is associated with aberrant cellular properties implicated in aneurysm development. ¹ For example, genetic mutations of SMC‐specific contractile genes predispose humans to familial AA, ² , ³ and synthetic SMCs acquire capabilities to produce extracellular matrix proteinases and proinflammatory mediators, which contribute to vascular wall degeneration. ¹

Smooth muscle contraction is primarily dependent on the phosphorylation status of myosin light‐chain (MLC), which is regulated by the activities of myosin light‐chain kinase (MLCK) and myosin light‐chain phosphatase (MLCP). ⁴ MLCK activation leads to smooth muscle contraction, which is dependent on Ca²⁺ and calmodulin. In contrast, MLCP activation causes smooth muscle relaxation. cGMP serves as a critical mediator for SMC relaxation by lowering the intracellular Ca²⁺ concentration via multiple different mechanisms, such as suppressing Ca²⁺ influx and mobilization, promoting Ca²⁺ sequestration and extrusion. ⁴ cGMP also increases MLCP activity through activating cGMP‐dependent protein kinase I (cGKI, also known as PKGI) that phosphorylates myosin‐binding subunit of MLCP. ⁴ Thus, cGMP/PKGI antagonizes SMC contraction by facilitating MLC dephosphorylation at serine 19. ⁴

cGMP‐hydrolyzing phosphodiesterase type 5 (PDE5) is a cGMP‐specific PDE and is highly expressed in vascular SMCs. ⁵ PDE5 functions as an important regulator in SMC contraction by hydrolyzing cGMP and antagonizing PKGI‐mediated vascular relaxation. ⁵ Sildenafil is one of the PDE5 inhibitors and is well‐known for treating erectile dysfunction (Viagra) ⁵ and pulmonary hypertension (Revatio). ⁴ Sildenafil promotes SMC relaxation through amplifying the effects of the endogenous cGMP/PKGI‐dependent relaxation mechanisms. ⁶ Recent studies have indicated the potential association of cGMP‐PKG signaling with AA. For example, a gain‐of‐function mutation in the gene encoding PKGI causes thoracic AA in humans. ⁷ The use of Viagra or other PDE5 inhibitors (such as tadalafil) have been occasionally reported to be associated with aortic dissection in men. ⁸ , ⁹ , ¹⁰ PDE5A expression levels were reduced in human AA tissues. ¹¹ In the current study, we investigated the effect of sildenafil on the development of experimental mouse abdominal AA (AAA), the most common form of AA that mainly occurs in elderly men. Our findings indicate that sildenafil treatment aggravated aortic degeneration and the progression of AAA likely by dysregulating SMC contractile function.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. A complete description of Methods is provided in Data S1.

Mouse Model of AAA

Ten‐week‐old wild‐type C57BL/6J male mice were treated with 0.2% 3‐aminopropionitrile fumarate salt (BAPN) (w/v) in drinking water. ¹² Periaortic application of porcine pancreas elastase or heat‐deactivated elastase (sham) were blinded performed on infrarenal abdominal aorta as previously described with slight modifications. ¹² Mice were treated with vehicle or sildenafil (100‐mg tablets; Viagra, Pfizer) prepared in drinking water to a concentration of 400 mg/L (0.6 mmol/L) as previously described, ¹³ resulting in the ingestion of ≈60 to 100 mg/kg per day, starting from the seventh day post‐surgery until harvest. ¹³ Procedures followed were in accordance with protocols approved by the University Committee on Animal Resources at the University of Rochester.

Immunofluorescence Staining and Proximity Ligation Assay

Aortic sections were stained with anti‐PDE5 polyclonal, anti‐total myosin light chain 2 (MLC) polyclonal, anti‐phospho‐myosin light chain 2 at Ser19 (pMLC) polyclonal, anti‐PKGIα/β monoclonal, anti‐calponin polyclonal, anti‐MYH11 polyclonal, anti‐α‐SMA monoclonal, anti‐LMOD1 polyclonal, anti‐F4/80 monoclonal, anti‐Mac2 monoclonal, or anti‐MMP9 polyclonal primary antibody for immunofluorescence. Proximity ligation assay was performed on aortic sections incubated with rabbit anti‐cGMP polyclonal primary antibody and mouse anti‐PKGIα/β monoclonal primary antibody using a Duolink In Situ Red Starter Kit.

Statistical Analysis

Assumptions of normality and equal variance were tested using R (version 3.6.2, https://www.R‐project.org/) by Shapiro‐Wilk test and Brown‐Forsythe test. All tests were 2‐sided, with a significance level for 2‐sided tests set at 5%. Statistical analyses and plotting were conducted using GraphPad 8 software. All data are presented as mean±SEM.

Results

PDE5A Expression is Reduced in SMCs of Mouse AAA Tissues

A previous study reported a reduction of PDE5A expression in human thoracic AA. ¹¹ Here, we aimed to examine PDE5A expression in mouse AAA tissues. Mouse AAA was induced in C57BL/6J male mice by periaortic elastase application in combination with BAPN treatment to block elastin/collagen cross‐links. ¹² Because of the limitation of AAA tissue quantity, we used immunofluorescent staining to detect PDE5A protein. We found that PDE5A immunofluorescent staining intensities were significantly reduced in SMCs of AAA medial lesion areas compared with sham (0.62±0.03 versus 1.00±0.03; P<0.0001) (Figure 1A and 1B, Figure S1), consistent with the finding in human AA tissues. Immunostaining of AAA sections with SMC markers, including calponin, myosin heavy chain 11, leiomodin 1, and α‐smooth muscle actin, revealed a dominant localization of SMC in the media lesion area (area inward of external elastic lamella) (Figure S2). These results suggest that loss of PDE5 function may be implicated in the pathogenesis of AA.

Figure 1. cGMP‐hydrolyzing phosphodiesterase type 5 (PDE5A) expression is reduced in media smooth muscle cells (SMCs) of mouse abdominal aortic aneurysm (AAA) tissues, and sildenafil aggravates development of AAA.

AAA was induced by periaortic elastase application in combination with 3‐aminopropionitrile fumarate salt (BAPN) treatment in C57BL/6J mice. A and B, PDE5 immunostaining of sham abdominal aorta (left panel) or AAA (right panel) cross‐section (A) and staining intensity of PDE5 in media lesion area of AAA, normalized to the averaged amount of staining intensity of sham at media (B). External elastic lamina (EEL) is indicated with the dashed line. Adv indicates adventitia. Sham (n=5), AAA (n=11). Data were analyzed by unpaired Student's t‐test. C, Experiment design to test the effect of sildenafil treatment on the development of pre‐existing AAA. D and E, M‐mode (D) and aorta diastole internal diameter (E) at baseline or on the seventh day post‐surgery assessed by Echo. The * indicates abdominal aorta. Dash line indicates 50% dilation compared with normal abdominal aortic diameter at baseline. n=11 mice. Data were analyzed by nonparametric paired Wilcoxon test. F, Representative infrarenal abdominal aortas. G, Quantification of the maximal abdominal aortic width. Aneurysm formation was defined as the increase in the external width of the infrarenal aorta by ≥50% compared with that in the sham/vehicle group. Sham/vehicle (n=5), sham/sildenafil (n=4), elastase/vehicle (n=17), elastase/sildenafil (n=14). Data were analyzed by parametric Welch ANOVA with Dunnett's T3 post‐hoc test. H, Viagra deteriorated elastin degradation by Van Gieson staining in AAA. I, Elastic fiber content in media area. Sham/vehicle (n=5), sham/sildenafil (n=4), elastase/vehicle (n=11), elastase/sildenafil (n=9). Data were analyzed by parametric one‐way ANOVA with Holm‐Sidak's post‐hoc test. All data are expressed as mean±SEM. Each dot represents one animal. *P<0.05, ***P<0.001, ****P<0.0001. VEH indicates vehicle.

Sildenafil Aggravates Development of AAA and Elastin Degradation

Next, we examined the effect of sildenafil on the progression and expansion of pre‐existing AAAs in mice. Mice were provided with BAPN in drinking water daily starting from 2 days before surgery until the end of the study (Figure 1C). Mice were subjected to surgery with periaortic elastase application or sham on the infrarenal abdominal aorta. Abdominal aorta size was monitored by ultrasound imaging before surgery as baseline and on the seventh day following surgery. The diastole internal diameter of the aorta by ultrasound imaging exceeded 50% dilation compared with baseline, indicating AAA occurrence (0.94±0.03 mm versus 0.51±0.01 mm; P<0.001) (Figure 1D and 1E). Mice were then randomly grouped and treated with sildenafil or vehicle daily until harvest on day 35 post‐surgery (Figure 1C). Compared with sham/vehicle controls, elastase induced remarkable dilatation in maximal aortic width, which was an increase of ≈193% (1.67±0.13 mm versus 0.57±0.004 mm; P<0.0001) (Figure 1F and 1G). Sildenafil aggravated AAA dilation (2.28±0.17 mm; P<0.05), with a 36.54% additional increase in aortic width compared with the vehicle AAA group. The images of AAA from all animals are shown in Figure S3.

Elastin degradation, as a key factor, contributes to aneurysmal wall degeneration and dilation. ¹ Elastase‐induced AAA exhibited evident degradation of elastic fibers via Van Gieson elastic staining, while sham groups showed no appreciable damage (8.62%±0.62% versus 21.58%±0.82%; P<0.0001) (Figure 1H and 1I). Sildenafil‐treated AAA manifested more severe destruction of elastin depicted by breaks and fiber thinning compared with the elastase/vehicle group (4.22%±0.56%; P<0.0001).

Effect of Sildenafil on MLC Phosphorylation in SMCs of AAA

Increased evidence has suggested that SMC contractile dysfunction is associated with aortic dissection and aneurysm. ¹ , ² , ³ SMC contraction is trigged by Ca²⁺‐dependent activation of MLCK and subsequent phosphorylation of MLC at serine 19. MLC phosphorylation (pMLC) has been frequently used as a biochemical indicator of vascular SMC contraction. Thus, we examined the effect of sildenafil on SMC contractile function by immunostaining for pMLC and total MLC in aortic tissues. We found that MLC staining was not significantly different in medial lesion areas of AAA between the elastase/vehicle and the elastase/sildenafil groups (25.51±3.12 versus 26.76±3.04; Figure 2A and 2C, Figure S4A). In adjacent sections, pMLC staining signals were significantly decreased in medial lesion areas of the AAA group compared with the sham group (26.43±1.86 versus 42.47±1.72; P<0.0001) (Figure 2B and 2D, Figure S4B). Interestingly, pMLC signals were further reduced in the sildenafil‐treated AAA group (20.15±0.95; P<0.01). Figure 2E shows the ratio of pMLC to MLC per medial lesion area for each animal. These results suggest that sildenafil may reduce aortic SMC contractility by suppressing MLC phosphorylation.

Figure 2. Effect of sildenafil on myosin light chain 2 (MLC) phosphorylation in media smooth muscle cells (SMCs) of abdominal aortic aneurysm (AAA).

A, Immunostaining of total MLC2 in cross‐sections of abdominal aorta. B, Immunostaining of MLC2 phosphorylated at Ser19 (pMLC) in adjacent sections. C, Staining intensity of MLC in sham aorta media or media lesion area of AAA. Data were analyzed by parametric one‐way ANOVA with Holm‐Sidak's post‐hoc test. D, Staining intensity of pMLC in sham aorta media or media lesion area of AAA. Data were analyzed by parametric one‐way ANOVA with Holm‐Sidak's post‐hoc test. E, Ratio of pMLC staining intensity per media lesion area to that of MLC for each animal. The elastase/vehicle group was compared with the elastase/sildenafil group by nonparametric Mann–Whitney test. C through E, Sham/vehicle (n=5), sham/sildenafil (n=4), elastase/vehicle (n=11), elastase/sildenafil (n=9). All data are expressed mean±SEM. Each dot represents one animal. *P<0.05, **P<0.01, ****P<0.0001. Adv indicates adventitia; a.u., arbitrary unit; EEL, external elastic lamina; ns, not significant; and VEH, vehicle.

Effect of Sildenafil on PKGI Activation in SMCs of AAA

We hypothesize that sildenafil reduces MLC phosphorylation through PKGI‐mediated dephosphorylation of MLC. cGMP binds to allosteric sites in the PKGI regulatory domain and activates PKGI. ⁴ To determine the effect of sildenafil on PKGI activation, cGMP and PKGI binding was detected by proximity ligation assay (PLA) (Figure 3C through 3E, Figure S5A). Compared with sham groups, elastase AAA showed increased PLA signals in media lesion areas (1.66±0.16 versus 1.00±0.06; P<0.05). The PLA signal was further increased in the sildenafil‐treated AAA group (2.96±0.27; P<0.0001), indicating increased cGMP and PKGI binding upon sildenafil treatment. Immunofluorescence staining showed that PKGI expression in AAA was not significantly changed by sildenafil (0.81±0.06 versus 0.79±0.03; Figure 3A and 3B, Figure S5B). These results suggest that PKG activation is upregulated by sildenafil treatment in aortic SMCs of AAA.

Figure 3. Effect of sildenafil on cGMP‐dependent protein kinase I (PKGI) activation in media smooth muscle cells (SMCs) of abdominal aortic aneurysm (AAA).

A, Immunostaining of PKGI in cross‐sections of abdominal aorta. B, Staining intensity of PKGI in the media lesion area of AAA, normalized to the averaged amount of staining intensity of sham at media. Data were analyzed by parametric one‐way ANOVA with Holm‐Sidak's post‐hoc test. C, Proximity ligation assay (PLA) principal paradigm. D, PLA detected cGMP and PKGI interaction in cross‐sections of abdominal aorta. E, Number of PLA dots were normalized to total cell number in the media lesion area. Data were analyzed by parametric one‐way ANOVA with Holm‐Sidak's post‐hoc test. B and E, Sham/vehicle (n=5), sham/sildenafil (n=4), elastase/vehicle (n=11), elastase/sildenafil (n=9). F, Proposed model: chronic sildenafil treatment causes contractile dysfunction of SMCs via activation of cGMP‐PKG signaling, which leads to the decrease of aortic contractility, increase of aortic wall stress, and subsequent deterioration of AAA development with pre‐existing damage. All data are expressed as mean±SEM. Each dot represents one animal. *P<0.05, ****P<0.0001. Adv indicates adventitia; EEL, external elastic lamina; ns, not significant; PDE5, cGMP‐hydrolyzing phosphodiesterase type 5; pMLC, phospho‐myosin light chain 2 at Ser19; and VEH, vehicle.

Discussion

In the present study, we demonstrated experimentally that chronic sildenafil treatment aggravated the progression of pre‐existing aortic elastin degradation and AAA dilatation induced by periaortic elastase in mice. Our finding is consistent with clinical case reports of aortic dissections in a few patients after abuse of PDE5 inhibitors sildenafil or tadalafil. ⁸ , ⁹ , ¹⁰ Our finding of PDE5A downregulation in mouse AAA is also in line with the report of PDE5A reduction in human aortic tissue obtained from Marfan, tricuspid, and bicuspid thoracic aneurysm samples, ¹¹ suggesting that PDE5A downregulation may contribute to the pathogenesis of aortic dissection and/or aneurysm. Although the causes of AA in the thoracic and abdominal aorta are different, they share many common pathological features such as elastic fiber degeneration and SMC dysfunction and loss. ¹ Given that PDE5 inhibitor–induced aortic dissections occurred more often in the thoracic parts of human aortas, the role of PDE5 inhibitors in the experimental model of thoracic AA also deserves to be investigated in the future. PDE5 inhibitors have been clinically used in patients with erectile dysfunction or pulmonary hypertension, and these patients often have cardiovascular diseases. Thus, caution may be taken when using PDE5 inhibitors in patients with, in particular, a predisposition to aortic diseases or cardiovascular risk factors.

We also observed an increase in cGMP binding to PKGI in AAA with sildenafil, a critical step for PKG activation. This result suggests an enhanced PKGI activation in mouse AAA by sildenafil, which is consistent with the finding that a gain‐of‐function mutation in human PKGI contributes to thoracic AA and aortic dissection. ⁷ Activation of cGMP‐PKG signaling is well‐known to suppress SMC contraction by decreasing MLC phosphorylation. ⁴ We indeed observed less pMLC in AAA tissues with sildenafil treatment, suggesting a reduction of SMC contractile function. Therefore, it is possible that chronic sildenafil treatment causes contractile dysfunction, which leads to the decrease of aortic wall elasticity (viscoelasticity), increase of aortic wall stress, and deterioration of AAA development in the abdominal aortic loci with pre‐existing damage (Figure 3F). Increasing evidence has suggested that SMC contractile dysfunction contributes to AA and aortic dissections. For example, human genetics studies have revealed that genetic mutations of SMC contractile genes, such as ACTA2 (encoding SMC‐specific alpha‐actin, α‐SMA) ² and MYH11 (encoding SMC‐specific myosin heavy chain), ³ result in familiar thoracic AA or dissections. A recent study has also reported a decrease of maximum contraction in SMCs ≈30% from patients with sporadic AAA compared with normal SMCs. ¹⁴ Thus, maintaining the normal aortic SMC contractile function is believed to be important for reducing wall stress in response to the pulsatile blood flow and high pressure in aortas.

Previous studies that investigated the contribution of angiotensin II (Ang II)–induced hypertension on AAA have indicated that AAA induction by Ang II is independent of the blood pressure–elevating effects of Ang II. For example, apolipoprotein E knockout mice infused with chronic Ang II increased mean arterial pressure and formed AAA, whereas administration of hydralazine lowered systolic blood pressure but did not prevent AAA formation. ¹⁵ Therefore, the effect of PDE5 inhibition on AAA is not likely dependent on its effect on blood pressure regulation. It is worth further examination of the effect of Viagra on Ang II–induced AAA in the future.

Multiple cell types participate in AAA development at different stages. Immunostaining of macrophage markers in sham or AAA (Figure S6) showed that F4/80‐ or Mac2‐positive cells are significantly increased in AAA tissues, while no significant difference was detected with sildenafil treatment. This suggests that the impact of macrophage infiltration by sildenafil, if any, is not likely a major factor in this AAA model, particularly at the late stage. We do not exclude the possibility of Viagra to affect macrophage function in other AAA models. It also remains unclear whether PDE5 inhibition affects endothelial function in AAA, which warrants exploration in the future with endothelial‐specific PDE5A knockout or transgenic mice.

Sources of Funding

This work was funded by National Institutes of Health (HL134910 and HL154318) to C. Y. and the American Heart Association (20PRE35210148) to C.Z.

Disclosures

None.

Supporting information

Data S1. Supplemental Materials and Methods

Figure S1–S6