Sildenafil Improves Epicenter Vascular Perfusion but not Hindlimb Functional Recovery after Contusive Spinal Cord Injury in Mice

Abstract

Nitric oxide (NO) is an important regulator of vasodilation and angiogenesis in the central nervous system (CNS). Signaling initiated by the membrane receptor CD47 antagonizes vasodilation and angiogenesis by inhibiting synthesis of cyclic guanosine monophosphate (cGMP). We recently found that deletion of CD47 led to significant functional locomotor improvements, enhanced angiogenesis, and increased epicenter microvascular perfusion in mice after moderate contusive spinal cord injury (SCI). We tested the hypothesis that improving NO/cGMP signaling within the spinal cord immediately after injury would increase microvascular perfusion, angiogenesis, and functional recovery, with an acute, 7-day administration of the cGMP phosphodiesterase 5 (PDE5) inhibitor sildenafil. PDE5 expression is localized within spinal cord microvascular endothelial cells and smooth muscle cells. While PDE5 antagonism has been shown to increase angiogenesis in a rat embolic stroke model, sildenafil had no significant effect on angiogenesis at 7 days post-injury after murine contusive SCI. Sildenafil treatment increased cGMP concentrations within the spinal cord and improved epicenter microvascular perfusion. Basso Mouse Scale (BMS) and Treadscan analyses revealed that sildenafil treatment had no functional consequence on hindlimb locomotor recovery. These data support the hypothesis that acutely improving microvascular perfusion within the injury epicenter by itself is an insufficient strategy for improving functional deficits following contusive SCI.

Introduction

Contusive spinal cord injury (SCI) leads to a host of sequelae, including loss of spinal microvasculature and ischemia. These processes contribute to cell death, axonal degeneration, and demyelination within the contusion and injury penumbra (Alexander and Popovich, 2009; Mautes et al., 2000). The onset of vascular pathology following injury is rapid and restricted to the injury epicenter (Benton et al., 2008a; Casella et al., 2006; Goodman et al., 1979; Griffiths et al., 1978; Whetstone et al., 2003). Early blood vessel dysfunction has been hypothesized to be a primary event leading to chronic histopathology following SCI (Benton et al., 2008a; Casella et al., 2002; Loy et al., 2002; Wagner et al., 1977). Microvessel shearing within the injury epicenter results in hemorrhage, the degree of which correlates with both terminal histopathology and functional deficit (Noble and Wrathall, 1989). This hemorrhage results in additional immediate damage in part caused by the further loss of blood flow due to vasoconstriction in the spared epicenter and penumbral blood vessels. Vasoconstriction contributes to the pathogenesis of an ischemic secondary injury (Conti et al., 2007; Ducker and Assenmacher, 1969). Post-traumatic ischemia in the spinal cord shows a direct linear dose-response relationship with injury severity (Tator and Fehlings, 1991).

Nitric oxide (NO) and endothelial NO synthase are important regulators of blood vessel dilation (Dawson et al., 1991). NO diffuses rapidly through tissue and across cell membranes and binds to soluble guanylate cyclase, stimulating the production of cyclic guanosine monophosphate (cGMP), which regulates several signaling pathways that affect endothelial and smooth muscle cell function. cGMP phosphodiesterases antagonize these pathways via the breakdown of cGMP. The sudden and transient reduction of endothelial NO synthase activity (Miscusi et al., 2006), and the reduction of vasodilatory signaling in contused spinal cord microvessels, may be correlated with vasoconstriction and subsequent ischemia post-injury (Anthes et al., 1996; Conti et al., 2007; Tator and Fehlings, 1991).

In addition to its vasodilatory effects, NO at low concentrations promotes vascular angiogenic cell survival, proliferation, and migration. Indeed, enhancing NO/cGMP signaling via administration of the cGMP phosphodiesterase inhibitors sildenafil or tadalafil has been shown to increase CNS angiogenesis in a rat ischemic model of embolic stroke (Ding et al., 2008; Li et al., 2007; Zhang et al., 2006).

Using a highly purified preparation of spinal microvessels, thrombospondin-1 (TSP-1) was identified as the most significantly upregulated mRNA (58-fold) 24 h after contusive thoracic SCI (Benton et al., 2008b). TSP-1 is a potent anti-angiogenic factor and antagonist of NO signaling. Its receptor, CD47, is an inhibitor of soluble guanylate cyclase activity. Targeted deletion of CD47 led to significant functional locomotor improvements, enhanced angiogenesis, and increased epicenter microvascular perfusion in mice after SCI (Myers et al., 2011). We hypothesized that by improving NO/cGMP signaling by inhibiting the breakdown of cGMP via antagonizing PDE5, we could pharmacologically improve perfusion of the spinal microvasculature after contusive injury and reduce subsequent functional motor deficits. This hypothesis was tested by administering the PDE5-selective antagonist sildenafil to mice for 7 days post-injury and assessing changes in the spinal vasculature and hindlimb functional recovery.

Methods

Surgical procedures

All surgical intervention, care, and treatment of animals were in strict accordance with the PHS Policy on Humane Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Research Council, 1996), and University of Louisville Institutional Animal Care and Use Committee (IACUC) guidelines. All mice were obtained from Harlan (Indianapolis, IN). Young adult (6–8 weeks old, 18–21 g) female C57BL/6 mice were used in this study and anesthetized using a 1.2% avertin solution (2,2,2-tribromoethanol) administered at 240 mg/kg IP and prepared as previously described (Benton et al., 2008a). Mice received T10 laminectomies and 50-kdyn contusions using the IH impactor (Infinite Horizons Inc., Lexington, KY) as previously described (Benton et al., 2008a; Mahoney et al., 2009; Scheff et al., 2003). No difference in mortality rates between groups receiving saline or sildenafil was observed. Between groups, no statistically significant differences existed between mean displacement (450–550 μm), which is the best predictor of injury severity (Ghasemlou et al., 2005). Dose and route of administration were modeled after previous studies (Farooq et al., 2008; Pifarre et al., 2011; Puzzo et al., 2009; Zhang et al., 2006). Five minutes after injury, 5 mg/kg sildenafil citrate (PZ0003; Sigma-Aldrich, St. Louis, MO) or saline was delivered systemically by IV injection into the surgically exposed right external jugular vein. Eight hours later and twice daily for 7 days post-injury, 5 mg/kg sildenafil or saline was delivered IP with an 8-h interval separating injections, except on day 7, when sildenafil or saline was administered only once. For functional assessment, sildenafil- (n=8) and saline-treated (n=9) mice were scored by the Basso Mouse Scale (BMS; Basso et al., 2006) 1 week prior to surgery and weekly after SCI. All raters were trained and certified at the Ohio State University by Dr. Basso and were blinded to genotype/treatment group.

Treadscan analysis

Recording and analysis of sildenafil- (n=8) and saline-treated (n=9) mice placed on a driven treadmill were performed as previously described (Beare et al., 2009), except the blinded scorer set the treadmill at the best walking speed for each untrained mouse, as determined by slowly increasing the treadmill speed until the animal was consistently walking with minimal lateral/longitudinal movement (Beare et al., 2011). Locomotion was recorded at 100 frames/sec, and TreadScan software (CleverSys, Reston, VA) was used to track and quantify mean instantaneous and overall run speed, hindlimb swing time, stride length, toe spread, and rear track width. Terminal TreadScan analysis was performed at 6 weeks post-SCI.

Tissue processing and immunohistochemistry

At 7 days post-injury and 4 h after the final administration of saline or sildenafil, sildenafil- (n=4) and saline-treated (n=3) mice were anesthetized and 100 μg of FITC-LEA (Lycopersicon esculentum agglutinin lectin, FL-1171; Vector Labs, Burlingame, CA) was delivered systemically by IV injection into the surgically exposed right external jugular vein, and allowed to circulate for 15 min prior to transcardial perfusion with 20 mL PBS, pH 7.4. LEA is a lectin that exhibits specificity for N-acetyl-D-glucosamine-β(1,4)N-acetyl-D-glucosamine oligomers, and when delivered intravascularly, has been widely used to identify perfused microvessels in various tissues (Jilani et al., 2003; Thurston et al., 1996). Quantitation of LEA staining 1.5 mm rostral and caudal to the contusion demonstrated no significant difference in perfusion of the spinal microvasculature by the intravital dye between groups (data not shown). Spinal cords were dissected, snap frozen, and longitudinally sectioned at 20 μm on a cryostat. The sections were thaw mounted on microscope slides (Fisher Scientific, Pittsburgh, PA) and stored at −80°C. The slides were warmed at 37°C for 20 min and fixed in ice-cold methanol for 20 min. The cords were blocked in TBS+0.1% Triton X-100, 0.5% BSA, and 10% normal donkey serum for 1 h at room temperature, and then incubated overnight at 4°C with primary antibodies in blocking buffer (monoclonal rat anti-PECAM-1, CD31, 550274, clone MEC13.3, 1:100 dilution; BD Pharmingen, San Diego, CA), monoclonal mouse anti-α smooth muscle actin (A2547, clone 1A4, 1:100 dilution; Sigma-Aldrich), polyclonal rabbit anti-laminin (L9393, 1:100 dilution; Sigma-Aldrich), polyclonal rabbit anti-GFAP (Z0334, 1:500 dilution; Dako, Carpinteria, CA), monoclonal mouse anti-CD68 (1:100 dilution; Chemicon, Billerica, MA), or polyclonal rabbit anti-PDE5A (LS-C82357/25065, 1:150 dilution; Lifespan Biosciences, Seattle, WA). The cords were subsequently incubated in secondary antibodies at room temperature for 1 h (Texas red 1:200-, FITC 1:200-, or AMCA 1:100-conjugated secondary antibody Fab fragments [all from donkey], and normal donkey serum [017-000-121]; Jackson ImmunoResearch, West Grove, PA). Negative controls included appropriate species-specific non-immune IgGs instead of primary antibodies (mouse [#08-6599] and rabbit [#08-6199] isotype control antibodies; Invitrogen, Carlsbad, CA). All images were captured with a Nikon TE 300 inverted microscope equipped with a spot CCD camera using identical exposure settings.

Image quantitation

Elements software (Nikon) was used to threshold baseline brightness and contrast identically for each image for both object and field quantitative measurements. For all assessments, the mean areas of the regions of interest (ROI) were quantified, which were centered on the contusion epicenters, and were not statistically significantly different between groups. For assessment of vascular patency, every fifth longitudinal section (11–12 20-μm sections) from each cord was stained and photographed with a 10× objective and stitched with Elements software during acquisition. The injury epicenter was defined as the domain exhibiting significant extravascular deposition and disorganization of vascular laminin immunoreactivity (Benton et al., 2008a; Whetstone et al., 2003). The injury penumbra was defined as 500 μm rostral and caudal to the epicenter (Fassbender et al., 2011; Myers et al., 2011). A scorer blinded to treatment quantified LEA-positive areas within the epicenter and penumbra using Elements software thresholding.

To approximate microvessel diameter, the Elements measurement feature “equivalent diameter” was used. It is a size feature derived from the object's area [equivalent diameter=√(4*area/π)], and determines the diameter of a circle with the same area as the measured object.

To approximate the degree of vessel patency (the vessel patency index), the sum LEA intensity and the sum area of each distinct microvessel within the lesion area was quantified by the Elements “Object Count” measurement feature. The sum LEA intensity was divided by the sum microvessel area ROI, and normalized with values obtained from equivalent area ROIs from uninjured tissue 1.5 mm rostral to the contusion epicenter. A vessel with more intense LEA staining within a smaller area will have a higher vessel patency index, which may reflect the extent of the microvessel's constriction. The Elements “Object Count” feature was used to obtain the overall number of LEA-stained microvessels within each ROI.

For assessment of platelet endothelial adhesion molecule (PECAM)-positive microvessels, every fifth longitudinal section (11–12 20-μm-thick sections) from each cord of each group was stained under identical conditions. Images were acquired with a 10× objective and stitched with Elements software during acquisition. A scorer blinded to treatment quantified the PECAM-positive area within the contusion epicenter and penumbra for each section using Elements software under identical thresholding conditions.

For quantitation of macrophage infiltration, every fifth longitudinal section (11–12 20-μm-thick sections) from each cord of each group was immunostained under identical conditions. Images were acquired with a 4× objective and stitched with Elements software during acquisition. An area of 2 mm centered at the contusion epicenter was defined as the region of interest for quantitation of epicenter and penumbral CD68-positive cells. Thresholding was performed using identical parameters within the “Object Count” feature of Elements software to define CD68-positive cell counting criteria based on size, circularity, and intensity. A similar procedure was used for quantitation of mean PDE5A intensity. Four longitudinal sections from each cord of each group were stained under identical conditions. The “Object Count” feature was used to threshold PDE5A-positive microvessels and quantify mean fluorescence intensity of each microvessel.

cGMP enzyme immunoassay (EIA)

Uninjured mice were anesthetized and either 10 mg/kg sildenafil (n=4) or saline (n=3) was delivered systemically by IV injection into the surgically exposed right external jugular vein and allowed to circulate for 45 min. Mice were sacrificed and cords were ejected from the spinal columns using a syringe filled with ice-cold PBS and immediately frozen in liquid nitrogen. For injured mice, a protocol identical to the behavioral study was used. Moderately-contused mice were given an initial IV and twice-daily 5 mg/kg doses of sildenafil or saline for 7 days, except on day 7, when sildenafil or saline was administered only once. Spinal cord epicenters were dissected 3 mm rostral and caudal to the contusion and snap frozen. Tissue was weighed and processed for the EIA according to the manufacturer's instructions using the acetylated cGMP protocol (cyclic GMP EIA Kit [# 581021]; Cayman Chemical, Ann Arbor, MI).

Statistical analysis

For all analyses, two-tailed Student's t-tests assuming equal variance were performed. All data are expressed as mean±standard error of the mean (SEM; *p<0.05, **p<0.01).

Results

Sildenafil increases perfusion of epicenter microvasculature after SCI

We previously identified increased angiogenesis and vascular perfusion of the spinal microvasculature 7 days after SCI in CD47^−/− mice (Myers et al., 2011). As CD47 is known to mediate inhibition of NO/cGMP signaling (Isenberg et al., 2009), we administered the phosphodiesterase PDE5-specific inhibitor sildenafil for 7 days post-injury to determine the effects of enhanced NO/cGMP signaling on vascular perfusion and angiogenesis after SCI. Sildenafil treatment failed to significantly change the area of PECAM-positive microvessels within the total lesion area: the laminin-defined lesion epicenter and penumbra (0.5 mm rostral and caudal to the epicenter; Fig. 1A, B, and F). No significant difference in mean lesion area was observed between saline-treated (2.96±0.12 mm²) and sildenafil-treated (2.85±0.12 mm², p>0.5) mice. Similarly, no significant effect was observed in the area of PECAM-positive microvessels within the injury penumbra alone between saline-treated (10.1±0.4%) and sildenafil-treated (9.8±0.5%, p>0.3) mice.

FIG. 1.

Sildenafil increases perfusion of epicenter microvasculature after spinal cord injury (SCI). Fluorescein isothiocyanate-Lycopersicon esculentum agglutinin lectin (FITC-LEA)-stained area (green) was quantified 7 days post-injury (7 dpi) in saline- (A, C, and E), and sildenafil-treated (B, D, and E) longitudinal sections within the total lesion area (the epicenter as defined by extravascular deposition and disorganization of laminin [blue, delineated by the boxed regions in A and B] plus the rostral and caudal penumbra). The boxed regions in A and B are shown at higher magnification in C and D. Seven-day treatment with sildenafil had no effect on angiogenesis post-injury as quantified by platelet endothelial adhesion molecule (PECAM) (red) staining (F; scale bar=500 μm in A and B; 50 μm in C and D). Data are means±standard error of the mean (saline-treated, n=3; sildenafil-treated, n=4; **p<0.01).

Intravenous administration of the vascular patency marker LEA, however, revealed that sildenafil did significantly enhance microvascular patency within the lesion area (Fig. 1A, B, and E). No significant difference in LEA area was observed within the contusion penumbra between saline-treated (11.7±0.9%) and sildenafil-treated (12.8±0.8%, p>0.4) mice. Sildenafil treatment elicited a small but significant increase in mean equivalent diameter (12.9±0.3 μm, *p<0.05) relative to saline-treated animals (11.8±0.3 μm) of epicenter LEA-stained microvessels. No significant difference in mean equivalent diameter was found between saline-treated (12.5±0.4 μm) and sildenafil-treated (13.0±0.3 μm, p<0.1) mice of penumbral LEA-stained microvessels.

Sildenafil significantly lowered the vessel patency index of microvessels within the epicenter (0.96±0.06, **p<0.01) relative to saline-treated animals (1.22±0.05). No significant difference in the vessel patency index was found within penumbral microvessels between animals treated with saline (0.95±0.08) or sildenafil (0.93±0.06, p>0.2). Sildenafil also significantly increased the mean number of distinct, LEA-stained microvessels (387±27, *p<0.05) within the epicenter relative to saline-treated animals (299±24). To confirm that the dose and route of administration antagonizes PDE5 activity, cGMP levels within the spinal cord were measured. Forty-five minutes after delivery, sildenafil significantly increased cGMP levels within the uninjured spinal cord (black bars in Fig. 2). Sildenafil also significantly increased cGMP levels within the contused spinal cord after a 7-day, twice-daily IP dosing regimen (grey bars). The fold increase in cGMP between saline- and sildenafil-treated groups was slightly higher in contused spinal cord tissue (2.3×) versus uninjured spinal cord (1.9×).

FIG. 2.

Sildenafil significantly increases spinal cord cyclic guanosine monophosphate (cGMP). Intravenous injection of 10 mg/kg sildenafil significantly increased cGMP levels after 45 min in uninjured (black bars) spinal cord homogenates as assessed by enzyme immunoassay (saline-treated, n=3; sildenafil-treated, n=4; *p<0.05). Intravenous injection and twice-daily IP injection of 5 mg/kg sildenafil for 7 days following moderate contusive injury similarly increased cGMP levels in contusion epicenters (grey bars). Data are means±standard error of the mean (saline-treated, n=5; sildenafil-treated, n=5; *p<0.05).

PDE5 is expressed in smooth muscle and endothelial cells of the spinal microvasculature

PDE5A expression within the spinal cord was primarily localized to vascular smooth muscle cells, as detected by co-localization with α-actin, of larger microvessels with a mean vessel diameter of 5 μm or larger (Fig. 3B), with little or no detectable expression in smaller vessels. These larger vessels may express PDE5A within endothelial cells as well, as detected by co-localization with the endothelial marker PECAM (Fig. 3A). No significant increase in the number of microvessels expressing PDE5A or the intensity of PDE5A expression within spinal microvessels was detected 7 days post-injury within the epicenter or injury penumbra (Fig. 3C and D).

FIG. 3.

Phosphodiesterase 5A (PDE5A) localizes to both endothelial and smooth muscle cells of the spinal microvasculature. PDE5A staining (red) co-localizes with markers for both endothelial cells (platelet endothelial adhesion molecule [PECAM], blue, A) and smooth muscle cells/pericytes (α-actin, blue, B) in larger spinal cord microvessels (diameter>5 μm), but not in smaller capillaries (scale bar=100 μm). XZ and YZ confocal planes are shown to confirm co-localization. (C and D) Sildenafil does not change microvascular localization or expression of PDE5A (red) 7 days post-injury as assessed by mean immunofluorescence intensity (scale bar=50 μm). Data are means±standard error of the mean (saline-treated, n=3; sildenafil-treated, n=4; LEA, Lycopersicon esculentum agglutinin).

Sildenafil had no effect on CD68⁺ macrophage infiltration after moderate contusive SCI

Inhibition of cGMP phosphodiesterase activity after CNS damage has been implicated in reducing inflammation and macrophage recruitment and infiltration to the site of injury (Pifarre et al., 2010,2011). We assessed the effects of 7-day treatment of sildenafil on the early inflammatory response post-contusion. Sildenafil had no significant effect on CD68⁺ macrophage infiltration into the contusion epicenter and penumbra 7 days post-injury (Fig. 4A, B, and C).

FIG. 4.

Sildenafil treatment has no effect on macrophage infiltration after moderate contusive spinal cord injury (SCI). CD68-positive cells (red) were quantified 7 days post-injury in saline-treated (A and C) and sildenafil-treated (B and C) longitudinal sections within the lesion epicenter and penumbral areas (Lycopersicon esculentum agglutinin [LEA], green; glial fibrillary acidic protein [GFAP], blue; scale bar=1000 μm in A and B). Data are means±standard error of the mean (saline-treated, n=3; sildenafil-treated, n=4; dpi, days post-injury).

Sildenafil has no effect on hindlimb locomotor functional recovery after moderate contusive SCI

We assessed the effects of 7-day treatment of sildenafil on the functional recovery from murine contusive SCI using the BMS. Sildenafil had no significant effect on motor scores up to 6 weeks following moderate contusion (Fig. 5A), and no significant differences existed between groups in their BMS subscores (Fig. 5B). TreadScan analysis (Beare et al., 2009) revealed that sildenafil treatment had no significant effect on mean instantaneous run speed (106.4±6.9 mm/sec) relative to saline control (108.5±5.1 mm/sec, p>0.8) mice placed on a treadmill 6 weeks after injury. Sildenafil treatment similarly had no significant effect on overall run speeds (85.9±7.7 mm/sec) relative to saline controls (93.6±5.0 mm/sec, p>0.4). Sildenafil treatment had no significant effect on hindlimb swing time, stride length, toe spread, and rear track width (Fig. 5C–F).

FIG. 5.

Sildenafil has no effect on hindlimb locomotor functional recovery after moderate contusive spinal cord injury (SCI). Twice-daily 5 mg/kg sildenafil for 7 days post-injury failed to improve functional recovery from moderate contusion after 6 weeks as assessed by the Basso Motor Scale (BMS, A) and subscore analysis (B). Sildenafil (black bars) failed to significantly improve hindlimb swing time (C), stride length (D), toe spread (E), or track width (F), relative to saline-treated mice (gray bars) as determined by TreadScan analysis. Data are means±standard error of the mean (saline-treated, n=9; sildenafil-treated, n=8).

Discussion

The central hypothesis of this study was that by improving cGMP signaling by inhibiting cGMP breakdown via the phosphodiesterase antagonist sildenafil, we could pharmacologically improve angiogenesis and perfusion of the spinal microvasculature following murine contusive injury, and reduce subsequent functional motor deficits.

Several studies have proposed that antagonizing cGMP phosphodiesterases modestly improves functional recovery following rat ischemic CNS injury. These studies differ substantially in methodology. For example, the models of ischemic injury, spinal ischemia (Kiymaz et al., 2008), focal cerebral ischemia (Zhang et al., 2006), and middle cerebral artery occlusion (MCAO; Menniti et al., 2009) all differed. The phosphodiesterase antagonist investigated also differed: sildenafil (Kiymaz et al., 2008; Zhang et al., 2006) or PF-5 (Menniti et al., 2009). The duration and route of drug administration varied: 1 day, IP (Kiymaz et al., 2008), 6 days, oral (Zhang et al., 2006), or 7 days, subcutaneous (Menniti et al., 2009). Finally, different methods of neurological assessment were utilized: the Tarlov classification (Kiymaz et al., 2008), body-swing and limb-placement tests (Menniti et al., 2009), or foot-fault and adhesive removal tests (Zhang et al., 2006).

With respect to SCI, our study assessed locomotor behavioral recovery for a full 6 weeks after injury, avoiding the complications that early spinal shock can have on more acute assessments. Indeed, while acute neurological improvements following sildenafil administration have been reported in a spinal ischemia model (Kiymaz et al., 2008), no significant acute functional improvements following tadalafil administration could be found acutely after trauma in a weight drop model of SCI (Serarslan et al., 2010). Additionally, our rationale for a more chronic duration of treatment was to assess the potential benefits of improving angiogenesis after injury, as observed after embolic stroke (Ding et al., 2008; Li et al., 2007; Zhang et al., 2006), in addition to the potential vasodilatory protective effects of a more acute administration. While we observed no increase in angiogenesis following sildenafil treatment 7 days post-injury, we cannot exclude the possibility that sildenafil had an earlier but unsustainable effect on epicenter or penumbral microvasculature following injury.

Our data suggest that a 7-day treatment with the PDE5 antagonist sildenafil is insufficient by itself to improve functional recovery from a moderate contusive SCI in mice. No differences were observed in BMS scores or subscores, or in hindlimb swing time, stride length, toe spread, and rear track width, gait characteristics that allow TreadScan software to successfully differentiate mild from moderate contusion injuries (Beare et al., 2009). A 2-week treatment with the PDE5 antagonist vardenafil also failed to improve functional recovery from an MCAO ischemic injury in mice (Royl et al., 2009), raising the possibility that the CNS tissue protection or regeneration observed after PDE5 inhibition in rats may be a species-specific effect. Alternatively, the degree and extent of tissue damage seen following a contusive injury may be too severe to observe behavioral improvements following vasodilatory therapeutic treatment relative to less severe injury models. Indeed, a similar dose and duration of sildenafil treatment had significant behavioral improvements in a murine model of multiple sclerosis (Pifarre et al., 2011). We cannot exclude the possibility that higher doses of sildenafil may improve recovery from contusive SCI. While sildenafil has the strongest affinity for PDE5, higher doses of sildenafil approach the median inhibitory concentration (IC₅₀) for cyclic adenosine monophosphate (cAMP) phosphodiesterases expressed in CNS-localized arteries (Farooq et al., 2008). We therefore avoided higher concentrations to prevent confounding interpretation of the antagonistic effects of sildenafil on cGMP versus cAMP metabolism.

Sildenafil administration increases cGMP levels within the mouse spinal cord. The localization of PDE5 expression within spinal cord blood vessels supports the hypothesis that antagonism of PDE5 by sildenafil has the effect of enhancing NO/cGMP signaling, primarily within the vasculature of the spinal cord. Sildenafil treatment increased vascular patency within the injury epicenter, but not within the penumbra rostral and caudal to the injury, as the mean LEA area, mean equivalent diameter, and vessel patency index, all improve within the epicenter following treatment. Following moderate contusive injury in the mouse, endothelial NO synthase mRNA expression significantly increases acutely after 24 h within the epicenter, but not rostral or caudal to the epicenter (GSE5296; Edgar et al., 2002). The potential increase in NO production after 24 h and subsequent activation of soluble guanylate cyclase within the epicenter may contribute to the effects of sildenafil and raise overall cGMP levels. While the overall levels of cGMP were higher in uninjured tissue, the fold increase in cGMP levels following sildenafil treatment in the injured spinal cord epicenter was slightly higher, consistent with this hypothesis. While sildenafil treatment appears to increase the mean equivalent diameter within the epicenter microvasculature, further investigation is required to ascertain whether the effects of sildenafil on vascular perfusion status occurs primarily within these smaller-diameter vessels, or within larger smooth muscle cell-containing vessels that feed the microvasculature.

Following spinal cord injury, NO levels exhibit a biphasic increase (Nakahara et al., 2002), with an immediate response within 30 min post-injury, and a second, delayed wave of increase, presumably due to inducible nitric oxide synthase (iNOS) activation and the post-traumatic inflammatory response (Conti et al., 2007). Although sildenafil treatment decreases microglial activation/macrophage infiltration in a murine model of multiple sclerosis (Pifarre et al., 2011), sildenafil treatment had no effect on microglial/macrophage staining following contusive SCI. Despite this, sildenafil could potentially be exacerbating the inflammatory response and oxidative damage generated by high NO levels post-injury. NO levels increase after sildenafil (Lee et al., 2010) or tadalafil administration (Serarslan et al., 2010) under some pathological conditions, but sildenafil decreases NO metabolites, cytokine expression, and macrophage recruitment in others (Handra et al., 2011; Kim et al., 2008; Wang et al., 2009). Future experiments will require assessment of the effects of sildenafil on spinal cord oxidative damage, neuronal apoptosis, and leukocyte infiltration following injury, to ascertain whether these responses mask the potentially beneficial effects of vasodilation. As sham groups with or without sildenafil were not included in this study, we cannot definitively define the degree of perfusion restored by sildenafil relative to uninjured animals. However, Benton and associates (2008a) showed that at 7 days post-SCI, epicenter perfusion was roughly 20% of sham values.

At pM concentrations of TSP-1, the receptor CD47 is a highly potent inhibitor of NO signaling by inhibiting soluble guanylate cyclase production of cGMP (Isenberg et al., 2006). Mice with targeted deletion of CD47 exhibit enhanced epicenter vascularity and increased microvascular patency following contusive SCI (Myers et al., 2011). Although abrogation of the inhibition of NO signaling likely contributes to the observed improvements in hindlimb functional recovery after injury in these mice, our results from increasing NO/cGMP signaling with sildenafil suggests that, by itself, increasing vasodilation and microvascular patency after SCI is insufficient to improve hindlimb motor function. CD47 is a known receptor for signal regulatory protein (SIRP)α, and CD47^−/− mice exhibit defects in neutrophil diapedesis across the vascular endothelium and reduced myeloperoxidase activity after contusion (Myers et al., 2011), suggesting that these defects in the acute, destructive inflammatory responses after SCI may contribute more significantly to the observed functional improvements. Alternatively, at least in platelets, CD47 has been proposed to inhibit NO/cGMP signaling at the level of cGMP-dependent protein kinase (cGKI) activity (Butt et al., 1994; Isenberg et al., 2005), downstream of cGMP and phosphodiesterase activity. If CD47 is significantly contributing to the functional deficits seen following SCI via its antagonism of NO/cGMP-mediated vasodilation, PDE5 antagonism via sildenafil treatment would only partially relieve the CD47-mediated inhibition of NO/cGMP signaling. This may also explain the lack of effect on angiogenesis or vessel sparing following sildenafil treatment.

Our results indicate that the acute enhanced perfusion of the spinal vasculature 7 days post-injury fails to improve behavioral function acutely (7–14 days). As microvascular perfusion status was not assessed before 7 days post-injury, we cannot distinguish between the possibilities that the lack of acute functional improvement is due to a lack of acute enhanced perfusion, or to a detrimental, sildenafil-mediated decrease in mean arterial blood pressure. However, 10 mg/kg sildenafil fails to significantly affect blood pressure up to 2 h after administration (Li et al., 2007). As we did not assess vascular perfusion status of the spinal cord microvasculature beyond 7 days post injury, we cannot definitively conclude that the lack of chronic functional improvement is due to a lack of enhanced perfusion, or a lack of effect of perfusion status on hindlimb recovery at this time point post-injury. Pharmacokinetic analyses in mice suggest that by 4 h following IP administration of 50 mg/kg sildenafil, most of the drug has been eliminated from the CNS and plasma (Puzzo et al., 2009). Furthermore, many of the effects of contusive injury on the murine spinal vasculature, such as angiogenesis, blood–spinal cord barrier permeability, and overall vascularity occur acutely, with some of these effects being restored to uninjured control levels by 21 days post-injury (Benton et al., 2008a; Whetstone et al., 2003). We therefore speculate that our drug regimen would not improve vascular perfusion at later time points, nor would we anticipate that enhanced vascular perfusion at this late time point would affect behavioral recovery.

In conclusion, improving epicenter vascular patency following contusive SCI is, by itself, an insufficient strategy for improving hindlimb functional recovery. Rescue of the spinal microvasculature within the injury epicenter and penumbra remains a largely unexplored yet promising therapeutic avenue to facilitate tissue sparing and functional recovery following contusive SCI. An important advantage of targeting vascular responses is the ability to administer treatments intravenously, which is rapid, widely accessible, and clinically highly relevant. Combinatorial therapeutic strategies that include enhancing vascular patency and reducing vasospasm, in addition to promoting endothelial cell survival and angiogenesis (Han et al., 2010), should continue to be explored. Assuming that no deleterious effects prevent maintenance of mean arterial pressure, using vasodilators to enhance vascular perfusion after SCI may facilitate drug delivery to injured tissues, and could be part of a beneficial therapeutic strategy.

Acknowledgments

This work was supported by grants NS045734 and RR15576, the Kentucky Spinal Cord and Head Injury Research Trust, Norton Healthcare, and the Commonwealth of Kentucky Research Challenge for Excellence Trust Fund (to T.H. and S.R.W.). We thank Christine Yarberry for surgical assistance, Kim Fentress for animal care, and Johnny Morehouse for BMS and TreadScan analyses. We thank Dr. Richard Benton for his critical comments on the manuscript.

Author Disclosure Statement

No competing financial interests exist.