The Functional and Structural Consequences of Cavernous Nerve Injury are Ameliorated by Sildenafil Citrate


  • John P. Mulhall and Alexander Müller equally contributed to this article.

John P. Mulhall, MD, Departments of Urology, Weill Medical College of Cornell University and Memorial Sloan Kettering Cancer Center, New York, NY 10021, USA. Tel: 212 746 0097 (personal line); 212-746-5653 (assistant); Fax: 212-746-0403; E-mail:


Introduction.  Radical prostatectomy (RP) is associated with erectile dysfunction (ED). A single, placebo-controlled, human study has assessed the effects of regular sildenafil use after RP and demonstrated an increased chance of preservation of preoperative erectile function.

Aim.  This study was undertaken to define the effects of such a regimen in an animal model.

Methods.  Using the cavernous nerve (CN) crush injury model, animals were divided into a number of groups: no CN injury (sham), bilateral CN injury exposed to either no sildenafil (control) or sildenafil at two doses (10 and 20 mg/kg) subcutaneously daily for three different durations (3, 10, 28 days).

Main Outcome Measures.  At these time points, CN electrical stimulation was used to assess erectile function by mean intracavernosal pressure (ICP)/mean arterial pressure (MAP) ratio. For the structural analyses, whole rat penes were harvested. Staining for Masson's trichrome was utilized to calculate the smooth muscle-collagen ratio. Immunohistochemical antibody staining was performed for endothelial (CD31 and eNOS) and neural (GAP43, NGF, and nNOS) factors and immunoblotting was performed to analyze the AKT/eNOS pathway. Terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) assay was used for the assessment of apoptotic indices and the CN architecture was evaluated by transmission electron microscopy (TEM).

Results.  Erectile function was improved with sildenafil in a time- and dose-dependent fashion with maximization of erectile function recovery occurring with daily 20 mg/kg at the 28-day time point. Sildenafil use resulted in smooth muscle-collagen ratio protection and CD31 and eNOS expression preservation. Sildenafil reduced apoptotic indices significantly compared with control. Animals exposed to sildenafil had increased phosphorylation of akt and eNOS. Tem demonstrated distinct differences in architecture between control and sildenafil groups toward an increased amount of myelinized nerve fibers.

Conclusions.  Sildenafil use in the CN crush injury model preserves erectile function that appears to be mediated predominantly through preservation of smooth muscle content and endothelial function as well as through reduction in apoptosis. Mulhall JP, Müller A, Donohue JF, Mullerad M, Kobylarz K, Paduch DA, Tal R, Li PS, Cohen-Gould L, and Scardino PT. The functional and structural consequences of cavernous nerve injury are ameliorated by sildenafil citrate. J Sex Med 2008;5:1126–1136.


Approximately 50,000 radical prostatectomies (RP) are performed each year in the United States for the treatment of prostate cancer. This operation is associated with at least transient erectile dysfunction (ED), with ED rates ranging from 20% to 90% depending upon the study reviewed [1–4]. It is postulated that the development of post-RP ED is due predominantly to a combination of temporary erectile (cavernous) nerve injury and damage to the erectile tissue secondary to neuropraxia and potentially absence of cavernosal oxygenation [5].

A single human, randomized, placebo-controlled study has been conducted examining the role of nightly sildenafil for a 36-week time period following RP [6]. This study demonstrated the ability of this regimen to increase the rate of preservation of preoperative erectile function at approximately 1 year postoperatively in the sildenafil arm compared with the placebo arm [6]. In another noncontrolled study, Schwartz et al. demonstrated that regular use of sildenafil post-RP preserved cavernosal smooth muscle content [7].

The presumed mechanism of this apparent protective effect of sildenafil was originally believed to be cavernosal oxygenation related to penile erection; however, more recently, some have postulated that sildenafil may have an endothelial or neuroprotective effect [8]. The rat cavernous nerve (CN) injury model is believed to simulate the neural injury that occurs during RP and is designed as a model to study the mechanisms of ED post-RP as well as to explore ED-minimizing strategies [9]. This study was undertaken to generate animal data in support of human studies and to explore the mechanisms by which sildenafil may preserve erectile function in this animal model.


Animal Groupings and Sildenafil Administration

Sprague-Dawley rats, initially weighing 250–300 g, were randomly divided into four groups: (i) sham (no CN crush, no sildenafil); (ii) control (C; bilateral CN crush, no sildenafil); and two treatment groups (bilateral CN crush, sildenafil sc): (iii) sildenafil 10 mg/kg (S10); and (iv) sildenafil 20 mg/kg (S20) subcutaneously daily commencing day of CN crush until 24 hours prior to sacrifice. Within each of the four groups, three time subgroups were analyzed, 3 days, 10 days, and 28 days after CN crush. Ten animals were analyzed for the following groups: sham, C28, S20-28; and five animals in each additional subgroup for a total number of 65 animals. The animals were cared for and housed under strict guidelines established by the Cornell University Institutional Animal Care and Use Committee guidelines.

CN Injury

For the initial surgery, the animals were anesthetized using 4% isoflurane and placed in the supine position. Through a lower midline incision, the major pelvic ganglion (MPG) lying on the dorsal prostate and the CN emanating form the ganglion were identified using a Zeiss operating microscope. For the CN crush injury, 5 mm distal to the MPG, a #7 Dumont hemostat was applied to the CN for 30 seconds, removed for 30 seconds and then reapplied for a further 30 seconds.

CN Stimulation

At either 3, 10 or 28 days after CN crush, the animals were anesthetized for the second, nonsurvival surgery. The left internal carotid artery was cannulated with heparinized polyethylene-50 tubing, connected to a pressure transducer and an amplifier unit (Harvard Apparatus, Holliston, MA, USA) recording the mean arterial pressure (MAP). The amplifier was attached to a data acquisition module (DI-190, Dataq Instruments, Akron, OH, USA). Measurement of the intracavernosal pressure (ICP) was achieved by inserting a 24-gauge needle into the corporal body. After CN identification, a stainless steel bipolar electrode with parallel hooks (1 mm apart) was placed around the nerve, just distal to the ganglion but proximal to where the nerve had been crushed. The electrode cable was linked to a Grass S48 stimulator (Quincy, MA, USA), and stimulated (parameters of 1.5 mA, 20 Hz, pulse width 5 milliseconds, at 7.5 volts) for 60 seconds each. Both nerves were stimulated and the maximal ICP generated with the corresponding MAP (ICP/MAP ratio) was recorded and presented as percentage.

Tissue Harvesting

Structural analyses were performed only on the 28-day group specimens. At the completion of CN stimulation and icp/map ratio recording, the penis was detached from its attachments to the ischiopubic rami and transected. On a side table, using a Zeiss operating microscope, the penis was microdissected so that all extra-tunical tissue was removed. The whole penis was divided into segments for embedding in paraffin (for terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) and immunohistochemistry) and for snap freezing (for immunohistochemistry and molecular analyses).


Following routine dehydration and paraffin embedding, tissue samples were cut into 5–7 µm sections from the mid-shaft of the penis mounted on slides and dried. Then the tissue slides, show ing the cross-section of the corpora cavernosa were deparaffinized and rehydrated and stained for Masson's trichrome to assess smooth muscle-collagen ratio utilizing Image J histomorphometry software version 1.33 National Institutes of Health. Five nonoverlapping tissue sections on each slide were reviewed, and two slides from two different animals per group were randomly selected. Thus, in each group (sham, control, and treatment), a total of 20 sections were reviewed for histomorphometry.


The corporal tissue of the rat penis was immunohistochemically stained for endothelial (CD31 and eNOS) and neuronal (nNOS, GAP-43, and NGF) factors. Following tissue processing, tissue sections were incubated with biotinylated secondary antibodies for 16 minutes, at room temperature. Secondary antibodies were used at a dilution of 1:2,000 (Vectastain ABC Kit [Mouse IgG] cat # PK-6102, Vectastain ABC Kit [Rabbit IgG] cat # PK-6101, Vectastain ABC Kit [Goat IgG] PK-6105). After secondary antibody incubation, immunohistochemical detection was carried out with Ventana Medical Systems' DAB Detection Kit following the manufacturer's instructions. Nonspecific signal was blocked with Blocker D; endogenous peroxidase was inhibited with Inhibitor D; streptavidin-HRP substrate was used to conjugate DAB to the biotinylated secondary antibody; tissue sections were incubated with DAB; and DAB signal was enhanced with Copper D enhancer (all provided with DAB Detection Kit, Ventana Medical Systems).

All antibodies used were manufactured by Abcam Inc., Cambridge, MA, USA. (CD31, ab7389; eNOS, ab3520; phospho-eNOS [S116], ab3867; AKT, ab8805; phospho-AKT [S473], ab11901; NGF, ab6198; nNOS, ab1376; GAP-43, ab11136). eNOS positive control: human endothelial cell lysate (BD Transduction Laboratories); eNOS negative control: Jurkat cells treated with LY294002 extract and untreated bovine aortic endothelial cell (BAEC) lysate. Beta-actin was used as a positive loading control for polyacrylamide gel electrophoresis (PAGE), and for normalization in the Image J analysis. Controls for signal specificity were sections run through immunohistochemistry without the primary antibody. Quantification of staining density was identical to the technique used for the Masson's trichrome staining analysis.


Quantification of staining density was identical to the technique used for the Masson's trichrome staining analysis. Control for fluorescence specificity was sections run through immunofluorescence without the primary antibody. This also ensured that any positive signals were not due to autofluorescence.


Deparaffinized tissue sections had cell nuclei exposed by treatment with recombinant proteinase K, followed by citrate/Triton X buffer on ice. TUNEL positive cells were detected by using In Situ Cell Death Detection Kit (Roche Applied Science). The percentage of cells stained with the TUNEL method to total number of cells stained with DAPI was recorded in percentage as the apoptotic index (AI). One slide from each of two animals were processed and using an Axio Plan Microscope each slide was analyzed by counting cells in five nonoverlapping zones of the whole-mounted corpora cavernosa section at a 40× magnification. Positive control (apoptosis positive): DNase-treated sham penis sections run through TUNEL staining procedure. Negative control (no apoptosis): sham penis section run through TUNEL staining procedure.


The 1 mm3 cubes of tissue were lysed in tissue homogenization buffer (Complete Protease Inhibitor Cocktail Tablets, Roche Applied Science), and phosphatase inhibitor cocktail (Phosphatase Inhibitor Cocktail Set 1, Calbiochem). Homogenates were centrifuged at 14,000 g for 15 minutes, and supernatants recovered. Homogenates were denatured by boiling in sodium dodecyl sulfate (SDS) sample buffer, and total proteins were separated by SDS-PAGE on 4–12% gradient gels (Invitrogen). Separated proteins were transferred to polyvinylidene difluoride membranes (Pall Life Sciences), and immunoblotted overnight at 4°C. The antibodies used are listed above. The blots were washed three times, 10 minutes each, at room temperature with TBST/3% milk. The blots were incubated with goat antirabbit IgG-HRP or goat antimouse IgG-HRP conjugate secondary antibodies diluted 1:5,000 in TBST/5% milk, 1 hour, room temperature. p-eNOS positive control: human endothelial cell lysate (BD Transduction Laboratories); p-eNOS negative control: Jurkat cells treated with LY294002 extract and untreated BAEC lysate; Akt/p-Akt positive control: untreated Jurkat cell lysate. p-Akt Negative Control: Jurkat cells treated with LY294002 extract.

Protein signal was detected by enhanced chemiluminescence (Chemicon International). Semiquantitative protein expression levels were determined by computer graphical analysis using Image J version 1.33 (NIH) and graphical plots were generated with spss software. Immunoblot protein expression bands were normalized to their corresponding beta-actin bands for graphical analysis.

Transmission Electron Microscopy (TEM)

The CN from sham, control, and sildenafil-treated animals were harvested, excising segments from 2 mm proximal to 2 mm distal to the zone of crush injury. The nerves were fixed in 4% Glutaraldehyde in 0.5 M cacodylate buffer at 4°C for 2 hours. The tissue was fixed in 1% Osmium Tetroxide and then dehydrated in graded ethanols. The tissue was infiltrated in graded resins, finally embedded in fresh 100% resin. Ultrathin sections (70 nM) were cut with a diamond knife and stained with 5% Uranyl Acetate and Lead Citrate. The resultant stained sections were viewed on a Hitachi H-600 transmission electron microscope run at 75 kV by a fully trained expert in TEM (LCG). Five nonoverlapping zones were reviewed on one slide from each of two animals in each group.

Statistical Analysis

The mean averages were built for maximum ICP/MAP ratios, smooth muscle-collagen ratios, antibody staining, and AIs for each group, and reported as the mean ± standard deviation. Individual pairwise comparison between groups was analyzed with independent two-tailed t-tests. Results were considered statistically significant if the P value was less than 0.05.


Erectile Hemodynamics (Figure 1)

Figure 1.

Functional results. The mean maximum intracavernosal pressure (ICP) divided by the corresponding mean arterial pressure (MAP), reported as ICP/MAP ratio as a percentage for sham, control, and both treatment groups S10 and S20 at different time points. The mean value plus the 95% confidence interval for each group is reported in the text. *Significantly higher compared with all other groups (P < 0.001). #Significantly improved compared with corresponding C group (P < 0.05). Significantly improved compared to S10 at 28 days.

The mean ICP/MAP ratios in the sham group was 70 ± 6% and remained significantly higher compared with all other groups (P < 0.01). However, daily subcutaneous sildenafil commencing the day of CN crush injury, administered at 10 mg/kg (S10) and 20 mg/kg (S20) showed an improvement in the mean ICP/MAP ratio compared with control at all three time points. The control group (C) demonstrated improvement in erectile function over time with the ICP/MAP ratio rising from 18 ± 10% at 3 days to 31 ± 13% at 10 days, remaining relatively unchanged at 28 days (32 ± 8%). This time-dependent increase in ICP/MAP ratios from 3 to 28 days was seen in every group including control, S10 and S20 (P < 0.05). The S10 group showed higher ICP/MAP ratios at all time points compared with control. While not statistically significant at all time points, there was a trend to significance. The best ICP/MAP ration in the S10 group (38 ± 6%) was achieved at 28 days (C28: 32 ± 8%, = 0.08). Of note, the administration of 20 mg/kg sc commencing the day of CN crush resulted in significantly improved ICP/MAP values compared with control even after 3 days (28 ± 11% vs. 18 ± 10%; = 0.047). Further improvement in a dose- and time-dependent manner was demonstrated by the S20 group, displaying the highest ICP/MAP ratio recovery (46 ± 14%) at 28 days, which was highly significant compared with C28 (32 ± 8%; = 0.001). A significant ICP/MAP ratio difference was seen at 28 days between S10 and S20 (38 ± 6% vs. 46 ± 14%, = 0.02).

Structural Analysis

Smooth muscle-collagen ratio assessment, immunohistochemistry, TUNEL assay, and immunoblotting were performed exclusively on 28 day-penile tissue samples from all four groups (sham, control, S10, and S20) for comparison.

Smooth Muscle-Collagen Ratios

The staining with Masson's trichrome in the sham group revealed a smooth muscle-collagen ratio of 16 ± 1.3%, the highest smooth muscle-collagen ratio of all animals. This result was significantly higher compared with 2.5 ± 0.2% for the control group (P < 0.05) and remained superior to both treatment groups as well (< 0.05). The S20 group reached 7 ± 0.6% significantly greater than the control group (< 0.05), and displayed a clear trend toward improvement compared with the S10 group (3 ± 0.2%; P = 0.05).


On both immunohistochemistry and immunofluorescence, the area of staining for CD31 was significantly reduced in the control group (22 ± 4%) compared with the sham group (42 ± 12%, P = 0.02). However, in both sildenafil groups, the expression of CD31 was preserved in comparison with the control group, with 33 ± 10% (< 0.01) and 29 ± 6% (< 0.01) in the S10 and S20 groups, respectively (Figure 2). There were no statistically significant differences between the sham and either sildenafil group (> 0.05). On immunohistochemistry, eNOS expression in the sham group was 15 ± 2% with a significant reduction noted in the control group (12 ± 1.5%, < 0.01). The S10 group reached 15 ± 4% and showed a trend toward eNOS staining preservation (P = 0.09) compared with C. However, eNOS staining in the S20 group was 17 ± 3%, significantly improved compared with C (< 0.001).

Figure 2.

CD31 staining. At 28 days after cavernous nerve injury, the control group demonstrated a lower density of CD31 staining compared with sham on both immunohistochemistry (upper panel) and immunofluorescence (lower panel).

Using immunohistochemistry, analyses of the neuronal antibodies NGF, nNOS, and GAP-43 were conducted. Interestingly, the analysis of GAP-43 expression, using image analysis software, showed significantly increased staining in the animals that underwent nerve crush alone (control 9.5 ± 1.5%) when compared with the sham group (6 ± 2%, < 0.001). Both sildenafil treatment groups also had increased GAP-43 staining levels compared with sham animals (P < 0.05). At 10 ± 3%, the S10 group was equivalent in GAP-43 staining to the C group (P = 0.6), while the S20 group value was 13 ± 4%, significantly increased GAP-43 expression level compared with C (= 0.03) and a trend toward higher intensity compared with S10 (= 0.06). The expression of NGF was similar for the C and the S20 group (9 ± 8% and 8.5 ± 3%, P = 0.8). But for both of these groups, staining levels were higher compared with sham animals (4 ± 2.5%), which was significant for the S20 group (P = 0.002), but not for the control group because of a large standard deviation in this group (P = 0.07). The nNOS staining levels were at a comparable level for the control and both treatment groups ranging between 7% and 9% and remained statistically lower compared with 14% in the sham group (P < 0.05).

Apoptosis Analysis

At 28 days after CN crush injury, the S10 and S20 groups demonstrated significant reduction in apoptosis within the corporal tissue with a mean AI of 24 ± 8% and 21 ± 4%, respectively compared with an AI of 63 ± 4% in the control group (P < 0.001) ( Figures 3 and 4). For comparison, the AI value in the sham group was 10 ± 7% (P < 0.001).

Figure 3.

TUNEL assay. Using the TUNEL assay, the distribution of fluorescent apoptotic cells in the corpora cavernosa was photomicrographically illustrated at 40× magnification. The control group revealed the highest and the sham group the lowest rate of apoptosis. Both S10 and S20 groups displayed a significant reduction in the number of apoptotic cells compared with the control group. White arrows indicate apoptotic cell nuclei (TUNEL positive cells).

Figure 4.

Apoptotic indices (AI). AI presented as ratio of apoptotic nuclei to total number of nuclei counted showed a significantly reduced AI percentage for both sildenafil groups S10 and S20 compared with the highly elevated AI percentage in the control group (P < 0.01). However, both treatment groups demonstrated higher AI compared with sham (P < 0.05).

AKT/eNOS Immunoblotting (Figure 5)

Figure 5.

Immunoblotting for eNOS/AKT. Both treatment groups, S10 and S20, demonstrated greater activation (phosphorylation) of AKT and eNOS compared with the control group.

On immunoblotting, sildenafil treatment resulted in a significant increase in expression of phosphorylated AKT (S473) (Figure 5), with S10 and S20 resulting in a 7.5- and 16-fold increased expression (results expressed as fold-increased expression compared with nonphosphorylated forms). Similarly, statistically significant differences were demonstrated for S10 and S20 groups, for expression of phosphorylated eNOS (S1177) compared with the nonphosphorylated forms, with 5.5- and 8.5-fold increases demonstrated, respectively.


Representative TEM pictures, conducted on CN cross-section samples at the area of the prior nerve crush, from the control and S20 groups at all three time points (3, 10, and 28 days) are shown in Figure 6. Compared with sham (not shown), in the control group at 3 days after CN crush, severe disruption and loss of architecture was seen which underwent reorganization over time. A time-dependent neural regeneration is seen within the control as well as for the S20 group. However, at all three time points, the S20 group showed less disorganization of neuronal integrity. In particular, the amount of myelinated nerve fibers in the sildenafil group was higher at all time points but most pronounced at 28 days. These descriptive findings of the improved CN nerve preservation with daily sildenafil administration are consistent with the functional data measuring the maximal ICP/MAP ratio during electrical CN stimulation.

Figure 6.

Transmission electron microscopy. Documenting all three time points (3, 10, and 28 days) after cavernous nerve (CN) crush injury the neural architecture is presented for the control group (upper panel) and the S20 treatment group (lower panel). A time-dependent improvement in neural arhcitecture is seen in both groups. However, the S20 groups showed more improved CN architecture with a greater density of myelinated nerve fibers at all three time points, most pronounced at 28 days.


The rat model of CN injury was first described by Quinlan in 1989 [10] and has been used extensively to investigate techniques and interventions to prevent ED following RP. We documented in prior experiments the use of this animal model in the assessment of functional and structural changes of erectile tissue after CN injury conducted using various techniques [9]. Our findings in this current series of experiments showed a highly significant reduction in ICP/MAP ratios in animals after CN injury compared with the sham group. The reduced ICP/MAP ratios were supported by our findings of a significant reduced smooth muscle-collagen ratio and a significant increased apoptosis rate in the control group compared with sham animals. Previous work in rat models, utilizing neurotomy as the mechanism of injury, has demonstrated increased apoptosis within the corpora following CN injury [11]. User et al. similarly confirmed increased apoptosis, especially subjacent to the tunica albuginea involving smooth muscle cells and hypothesized that damage to the subtunical smooth muscle cells prevents compression of the perforating subtunical veins resulting in veno-occlusive dysfunction (VOD) and subsequent failure of recovery of erectile function [12].

The daily use of sildenafil citrate commencing on the day of CN crush injury improved ICP/MAP ratio compared with the control group at all three time points (3, 10, and 28 days). The 28-day time point was chosen as this is generally believed to represent the 2-year time point in a human, and after RP we expect erectile function recovery by this time point. The administration of daily 10 mg/kg sildenafil displayed a promising trend in erectile function recovery, but never reached statistical significance (P > 0.05). However, the S20 group showed a clear, time-dependent increase in erectile function preservation. Impressive is our documentation that the S20 group resulted in significant improvement of the functional data compared with control even after only 3 days (P = 0.047). We believe that the improvement in erectile function as early as 3 days after CN crush is related to apoptosis prevention even at this early stage. We have documented significant apoptosis prevention using sildenafil at 28 days and we believe that apoptosis reduction is at play at the third post-CN crush day. Furthermore, it has been well documented that apoptosis occurs within 24 hours of CN injury [12].

A single randomized, placebo-controlled study in men post-RP has been conducted examining the role of nightly sildenafil for 6 months after surgery [6]. This analysis demonstrated the ability of this regimen to increase the rate of preservation of preoperative erectile function (based on validated inventory assessments) at 48 weeks in the sildenafil arm compared with the placebo arm (27% vs. 4%, respectively) [6]. In another controlled study (without a placebo group), Schwartz et al. gave either 50 mg or 100 mg of sildenafil on alternate nights for 6 months after RP [7]. Patients had a percutaneous biopsy before RP and after 6 months of sildenafil treatment. Comparing the smooth muscle content of the posttreatment biopsy to the preoperative biopsy, no change was demonstrated in the 50 mg group but an increase in smooth muscle content was noted in the 100 mg group. In our study, we used doses of 10 mg/kg and 20 mg/kg of sildenafil, given at induction of anesthesia and continued daily. The dosing of 10 mg/kg sildenafil sc leads to maximum plasma concentrations comparable to 100 mg Viagra taken orally by men (approximately 50 nM; Data on file, Pfizer Laboratories, Sandwich, UK). Because of a higher clearance of sildenafil in rodents (T1/2 0.4–1.3 hours) and because no measureable sildenafil exists in rat serum 10 hours after a single 10 mg/kg sc dose, we chose a second dosing group of 20 mg/kg sc to increase the duration of exposure [13]. Some authorities have suggested that the functional data we present in this article are representative of a sildenafil effect at 30 hours postinjection because of a depot build-up of sildenafil subcutaneously. While Musicki et al. have shown this to be the case, their series of experiments used dosing in a three-times-a-day fashion, at 20 mg/kg sc for 3 weeks [14]. In young animals, a single injection gave 53 nM free sildenafil, 26 nM after 10–18-hour washout, 13 nM after a 3-day washout and 1 nM after 1-week washout. In another study, more representative of our model, Behr-Roussel et al, in the rat, used 20 mg/kg sc once daily for 6–8 weeks with a 1-week washout [15]. No serum level of sildenafil was measured at this time point. Thus, when 20 mg/kg sc is given two or three times per day, there appears to be accumulation in the plasma (clearance is saturated, as well as the potential local site depot development). We believe the likelihood of sildenafil forming subcutaneous depots or saturating clearance after a single 20 mg/kg dose of sildenafil is far less likely as, after a single 20 MG/kg sc injection 24-hour coverage is not achieved and the drug has been fully eliminated by the next injection.

Our findings of a decreased smooth muscle-collagen ratio after CN injury are consistent with other literature [16–19]. Functionally, RP is associated with loss and collagenization of corporal smooth muscle tissue resulting in fibrosis and VOD development. Ferrini et al. documented in a recent study the prevention of VOD in aged rats by long-term orally administration of sildenafil for 45 days [17]. In our series of experiments, the administration of sildenafil 20 mg/kg sc resulted in prevention of smooth muscle collagenization compared with animals that underwent CN injury alone. Recent evidence suggests that bilateral cavernous neurotomy-induced hypoxia and decreased smooth muscle: collagen ratio within the erectile tissue can be restored by the daily administration of tadalafil for 3 months [18]. Also the third available phosphodiesterase type 5 (PDE5) inhibitor vardenafil appears to prevent fibrosis occurring in the penis of a rat exposed to CN injury [19]. Thus, this may be a class-effect. However, Ghofrani et al. demonstrated that in the pulmonary artery, oxygenation was maximized with sildenafil at physiologic doses compared with vardenafil and tadalafil at physiologic and supra-physiologic doses [20]. Thus, not all PDE5 inhibitors are necessarily identical in their cellular and tissue effects and caution should be used in comparing one PDE5 inhibitor's effect with another on preservation of cavernosal tissue integrity without direct comparison experiments.

The mechanism of action of sildenafil is through PDE5 inhibition resulting in the accumulation of intracellular cyclic guanosine monophosphate. While the effects on vascular smooth muscle relaxation are well appreciated, other effects on apoptosis and endothelial integrity are less well appreciated. Recently, Salloum et al. demonstrated that both sildenafil and vardenafil reduced the area of cardiac necrosis in a rabbit model of cardiac ischaemia-reperfusion [21,22]. In a follow-up study, Das et al. using mouse cardiac myocyte cells exposed to hypoxia and reoxygenation showed that sildenafil-treated cells demonstrated less necrosis and apoptosis than control cells [23]. The authors also showed increased expression of eNOS in the sildenafil-treated groups, along with an increase in the ratio of the antiapoptotic protein Bcl-2 compared with the pro-apoptotic protein Bax.

In our study, the immunohistochemical staining for endothelial factors including the antibodies for CD31 and eNOS appears to indicate an endothelial protective effect with sildenafil use. Both staining levels were significantly reduced by CN injury compared with sham animals (P < 0.05). Thus, daily sildenafil treatment resulted in a restoration of CD31 and eNOS staining levels to that of sham animals (no statistical differences between sham and either treatment group). Furthermore, these findings have been supported by our immunoblotting results documenting a significant increase in activated (phosphorylated) AKT and eNOS compared with nonphosphorylated forms in both sildenafil treatment groups. Our results support the observations of Musicki et al. reporting on an extended erectile response after CN stimulation with a prolonged detumescence time and an increased expression of phosphorylated eNOS and Akt in aged rats (19 months old) after long-term administration of sildenafil for 3 weeks [24]. They concluded that patients with impaired erectile function may benefit from a long-term use of sildenafil by enhancing Akt-dependent eNOS phosphorylation. While the increased expression of phosphorylated eNOS and Akt was extremely high; however, our data were reproducible between animals and staining was quantified by an experienced morphometrics analyst.

Of note, the immunohistochemistry for the neural factors GAP43 and NGF provided similar staining ratios, specifically, the lowest staining density for the sham animals and significantly higher levels in all CN injury groups whether treated or untreated. Speculatively, this might be explained through ongoing neuronal regenerative processes with higher activation of GAP43 and NGF after CN damage. Our findings regarding the immunohistochemical nNOS staining pattern in the corporal tissue did not show an alteration with sildenafil treatment after CN injury compared with control animals and all groups continued to be reduced compared with sham rats. It remains unclear whether the nNOS recovery at the CN terminals followed the same behavior. However, the TEM results shed light on the local neural preservation after CN damage. After CN damage alone (untreated), a neural reorganization was observed over time continuing until 28 days. With sildenafil treatment (20 mg/kg), an improvement in neural organization and greater density of myelin sheaths were seen compared with the control group. Mechanistically, we postulate that sildenafil use in this model results in a mild neuroprotective rather than a neuroregenerative effect. This is supported by the TEM data as well as the fact that the functional recovery curve in all three groups is the same but in the S20 group never nadirs as low as the other two groups even at 3 days after CN crush. Finally, this is supported by two studies looking at the use of PDE5 inhibitors in animal models of stroke demonstrating significant functional recovery in the sildenafil- and tadalafil-treated animals compared with untreated animals.


We have demonstrated, in the CN crush injury model, that sildenafil citrate preserves erectile function and acts by preserving smooth muscle content, endothelial integrity and reducing apoptosis. At this time, based on TEM data, while there are clear differences in CN architecture between treated and untreated animals, it is unclear if these changes represent a genuine neuroregenerative effect. The pharmacological effect on erectile function recovery is dose- and time-dependent. Furthermore, it appears that the endothelial protection afforded by sildenafil is mediated, at least in part, by phosphorylation of AKT and eNOS. The clinical implications of these data are obvious.


This study was funded in part by the Dow Wallace Fund and a research grant from Pfizer Inc.

Conflict of Interest: This study was funded in part by the Dow Wallace Fund and a research grant from Pfizer Inc.

Statement of Authorship

Category 1

  • (a)Conception and Design
  • John P. Mulhall

  • (b)Acquisition of Data
  • John P. Mulhall; Alexander Müller; John F. Donohue; Michael Mullerad; Raanan Tal; Keith Kobylarz; Darius A. Paduch; Leona Cohen-Gould

  • (c)Analysis and Interpretation of Data
  • John P. Mulhall; Alexander Müller; Michael Mullerad; Leona Cohen-Gould

Category 2

  • (a)Drafting the Article
  • Alexander Müller; Michael Mullerad; John F. Donohue; John P. Mulhall

  • (b)Revising It for Intellectual Content
  • John P. Mulhall; Michael Mullerad; Alexander Müller

Category 3

  • (a)Final Approval of the Completed Article
  • John P. Mulhall; Alexander Müller