Intracavernous Delivery of a Designed Angiopoietin-1 Variant Rescues Erectile Function by Enhancing Endothelial Regeneration in the Streptozotocin-Induced Diabetic Mouse

Recombinant adenovirus-expressing FLAG-tagged COMP-Ang1 or bacterial β-gal was constructed and COMP-Ang1 recombinant protein was prepared as previously described (22).

Eight-week-old C57BL/6J mice were used in this study. The experiments were approved by the institutional animal care and use subcommittee of our university. Diabetes was induced by intraperitoneal injections of multiple low doses of STZ (50 mg/kg body wt in 0.1 mol/L citrate buffer, pH 4.5) consecutively for 5 days, as previously described (12). Eight weeks after diabetes was induced, the animals were anesthetized with ketamine (100 mg/kg) and xylazine (5 mg/kg) intramuscularly and were placed supine on a thermoregulated surgical table. The penis was exposed by use of a sterile technique. A 30-gauge insulin syringe was used to administer a single injection of ad-LacZ (2 × 10⁸ parts/20 μL) or ad–COMP-Ang1 (2 × 10⁸ parts/20 μL) and repeated injections of PBS (days −3 and 0; 20 μL) or COMP-Ang1 recombinant protein (days −3 and 0; 5.8 μg/20 μL) into the midportion of the corpus cavernosum. The incision was closed with 6-O Vicryl (polyglactin 910) sutures. We evaluated erectile function (n = 6 per group) by electrical stimulation of the cavernous nerve 2 and 4 weeks after treatment. We administered ad–COMP-Ang1 at a concentration of 2 × 10⁸ parts because we could achieve the highest erectile response at this concentration (J.-K. Ryu, W.J. Kim, S. Piao, H.-R. Jin, and J.-K. Suh, unpublished observation). STZ-induced diabetic mice that received two successive intracavernous injections of COMP-Ang1 protein showed significant recovery of erectile function 2 and 4 weeks after treatment, which was comparable to the function in the mice that received a single intracavernous injection of ad–COMP-Ang1 (Fig. 1 and Supplementary Fig. 2). On the basis of these functional results, the mice were divided into the following four groups for histologic examination, biochemical study, and vascular permeability testing: age-matched controls, STZ-induced diabetic mice without treatment, STZ-induced diabetic mice receiving repeated intracavernous injections of PBS (days −3 and 0; 20 μL), and STZ-induced diabetic mice receiving repeated intracavernous injections of COMP-Ang1 protein (days −3 and 0; 5.8 μg/20 μL). Fasting and postprandial blood glucose levels were determined with an Accu-Check blood glucose meter (Roche Diagnostics, Mannheim, Germany) before the mice were killed.

Erectile function was evaluated by electrical stimulation of the cavernous nerve. Stimulation parameters were 1–5 V at a frequency of 12 Hz, a pulse width of 1 ms, and a duration of 1 min. The ratios of maximal intracavernous pressure (ICP) and total ICP (area under the curve) to mean systolic blood pressure (MSBP) were calculated to adjust for variations in systemic blood pressure, as previously described (12).

For inhibition studies, a separate group of STZ-induced diabetic mice was given N-nitro-L-arginine methyl ester (l-NAME) (a NOS inhibitor, 75 μmol/L; Sigma-Aldrich, St. Louis, MO) or apocynin (an inhibitor of NADPH oxidase activation, 180 μmol/L; Sigma-Aldrich) in their drinking water beginning 1 day before the intracavernous injection of the COMP-Ang1 protein.

Immunohistochemistry, in situ detection of superoxide anion, and the terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay were performed as described in the Supplementary Methods.

Western blot for COMP-Ang1, eNOS, phospho-eNOS, p47^phox, inducible NOS (iNOS), and endothelial cell-cell junction proteins and cGMP determination were performed as described in the Supplementary Methods.

Plasma NO_x concentrations were determined as described in the Supplementary Methods.

The mice from each STZ-induced diabetic group and their age-matched controls were anesthetized, and the fluorescent tracer rhodamine-dextran (40 kDa, 25 mg/mL in saline; Molecular Probes, Eugene, OR) was injected into the jugular vein. After 10 min, electrical stimulation of the cavernous nerve was performed for 1 min to induce penile erection and to increase blood flow into the penis. Ten minutes after cavernous nerve stimulation, corpus cavernosum tissues were harvested and fixed in 4% paraformaldehyde for 24 h at 4°C. Frozen tissue sections (n = 4 per group) were incubated with antibody to platelet–endothelial cell adhesion molecule (PECAM)-1 (1:50; Chemicon, Temecula, CA). The intensity of fluorescence in the nonvascular area of the corpus cavernosum tissue was measured by image analysis (NIH Image J).

Results are expressed as means ± SE. Statistical analysis was performed by using one-way ANOVA followed by Student-Newman-Keuls post hoc tests. P < 0.05 was considered to be statistically significant.

The fasting and postprandial blood glucose concentrations of the STZ-induced diabetic mice were significantly higher than those of the control mice. Body weight was significantly lower in the STZ-induced diabetic mice than in the controls. The body weight and blood glucose levels of the STZ-induced diabetic mice did not differ significantly, regardless of the treatment given (Table 1 and Supplementary Table 1).

Exogenous COMP-Ang1 protein expression was detected by Western blot in the corpus cavernosum tissue of STZ-induced diabetic mice 3, 7, 14, and 21 days after injection of ad–COMP-Ang1. The level of exogenous COMP-Ang1 protein expression was highest 7 days after injection, and expression still was detectable at 21 days (Supplementary Fig. 1A). We also evaluated the expression of COMP-Ang1 protein in the corpus cavernosum at 1, 6, and 24 h and 3, 7, and 14 days after injection of COMP-Ang1 recombinant protein. The level of COMP-Ang1 protein peaked at the earliest time point assayed (1 h) and was detectable up to 3 days after injection (Supplementary Fig. 1A).

To localize the COMP-Ang1 protein, either expressed from ad–COMP-Ang1 or injected as a recombinant protein, immunohistochemical staining for FLAG-tagged COMP-Ang1 was performed 7 days after intracavernous injection of ad-LacZ or ad–COMP-Ang1 and 1 h after intracavernous injection of PBS or COMP-Ang1 protein.

The FLAG-tagged protein was strongly expressed in endothelial cells in both the ad–COMP-Ang1– and COMP-Ang1 protein–treated animals (Supplementary Fig. 1B).

X-gal histochemistry in fixed sagittal blocks of the whole penis 7 days after injection of ad-LacZ showed uniform staining along the penile shaft, whereas virtually no endogenous β-gal activity was observed in animals injected with the virus vehicle alone. Frozen transverse sections from the penises of diabetic mice 7 days after injection of ad-LacZ were used to determine its cellular distribution. The X-gal staining was mainly observed in endothelial cells lining the cavernous space (Supplementary Fig. 1C).

A representative intracavernous tracing after stimulation of the cavernous nerve (1–5 V, 12 Hz, 1 ms) for 1 min in age-matched control or STZ-induced diabetic mice 2 and 4 weeks after treatment is shown in Fig. 1A and Supplementary Fig. 2A, respectively. During electrical stimulation of the cavernous nerve, the ratios of maximal ICP and total ICP to MSBP were significantly lower in untreated STZ-induced diabetic mice and STZ-induced diabetic mice treated with ad-LacZ or PBS than in age-matched controls. A single intracavernous injection of the ad–COMP-Ang1 gene or repeated intracavernous injections of the COMP-Ang1 protein restored erection parameters, which reached up to 83–89% of control values, and the physiological improvement lasted up to 4 weeks after treatment (Fig. 1B and C, and Supplementary Fig. 2B and C). No detectable differences were found in MSBP among the six experimental groups (Table 1 and Supplementary Table 1). At 6 and 8 weeks after administration of either the ad–COMP-Ang1 or COMP-Ang1 protein, however, erectile function returned to baseline levels (Supplementary Fig. 3).

Immunohistochemical staining of cavernous tissue with an antibody to PECAM-1 was performed in age-matched control and STZ-induced diabetic mice 2 weeks after treatment. We found significantly lower cavernous endothelial cell content in the untreated and PBS-treated STZ-induced diabetic mice than in the control mice. Repeated intracavernous injections of the COMP-Ang1 protein completely restored cavernous endothelial content (Fig. 2A and C).

To test whether the COMP-Ang1 protein–induced increase in cavernous endothelial content resulted from endothelial cell proliferation, we assessed the number of endothelial cells staining positive for phosphohistone H3 (a nuclear protein indicative of cell proliferation). Because only a few or virtually no phosphohistone H3–positive endothelial cells were noted in diabetic mice 2 weeks after treatment with the COMP-Ang1 protein (data not shown), we further investigated whether the COMP-Ang1 protein induced endothelial cell proliferation at earlier time points. We noted significant increases in phosphohistone H3–positive endothelial cells 1, 3, and 6 h after injection of the COMP-Ang1 protein (Fig. 2B and D).

Cavernous phospho-eNOS (Ser1177) expression, as assessed by immunohistochemical staining, was significantly lower in the untreated and PBS-treated STZ-induced diabetic mice than in the age-matched controls at 2 weeks after treatment. Repeated intracavernous injections of the COMP-Ang1 protein significantly increased endogenous eNOS phosphorylation in the STZ-induced diabetic mice, which was comparable to that in age-matched controls (Fig. 3A and B). Similar to the results of immunohistochemical staining, immunoblot analysis of the corpus cavernosum tissues revealed an increase in cavernous phospho-eNOS expression in STZ-induced diabetic mice 2 weeks after treatment with the COMP-Ang1 protein, whereas no changes were noted in the total amount of eNOS in control or STZ-induced diabetic mice regardless of treatment (Fig. 3C).

Cavernous cGMP decreased significantly in untreated and PBS-treated STZ-induced diabetic mice compared with that in the age-matched controls at 2 weeks after treatment. Intracavernous administration of the COMP-Ang1 protein significantly increased cGMP concentrations compared with those in untreated or PBS-treated STZ-induced diabetic mice (Fig. 3D).

We performed Western blots to evaluate the cavernous tissue expression of p47^phox and iNOS in age-matched control and STZ-induced diabetic mice 2 weeks after treatment. Cavernous p47^phox and iNOS protein expression was significantly higher in the untreated and PBS-treated STZ-induced diabetic mice than in the age-matched controls. Repeated intracavernous injections of the COMP-Ang1 protein significantly decreased cavernous p47^phox and iNOS expression in the STZ-induced diabetic mice (Supplementary Fig. 4).

In the in situ analysis of superoxide anion production, the fluorescent products of oxidized hydroethidine in both the cavernous endothelium and the smooth muscle were significantly higher in the untreated and PBS-treated STZ-induced diabetic groups than in the control group. Intracavernous administration of the COMP-Ang1 protein significantly decreased superoxide anion production in the cavernous endothelial cells and smooth-muscle cells of STZ-induced diabetic mice but not to the level found in the age-matched controls (Fig. 4).

We also performed immunohistochemical localization of nitrotyrosine to determine peroxynitrite generation, which is derived from NO and superoxide anion. Nitrotyrosine expression in endothelial cells and smooth-muscle cells of the corpus cavernosum was higher in untreated and PBS-treated STZ-induced diabetic groups than in the control group. Repeated intracavernous injections of the COMP-Ang1 protein significantly decreased nitrotyrosine generation in the cavernous endothelial cells and smooth-muscle cells of the STZ-induced diabetic mice, which was comparable to that in age-matched controls (Fig. 5).

Double labeling of cavernous tissue with TUNEL and the antibody to PECAM-1 or smooth-muscle α-actin showed that the number of apoptotic cells in cavernous endothelial cells and smooth-muscle cells was significantly greater in the untreated and PBS-treated STZ-induced diabetic groups than in the control group. Repeated intracavernous injections of the COMP-Ang1 protein significantly decreased apoptosis in the cavernous endothelial cells and smooth-muscle cells of the STZ-induced diabetic mice, although not to the level found in the age-matched controls (Fig. 6).

We performed Western blots to evaluate the cavernous tissue expression of the adherens junction protein (vascular endothelial [VE]-cadherin) and tight junction proteins (zonular occludens-1 [ZO-1], occludin, and claudin-5) in age-matched control and STZ-induced diabetic mice 2 weeks after treatment. The cavernous expression of VE-cadherin, ZO-1, occludin, and claudin-5 protein was significantly lower in the untreated and PBS-treated STZ-induced diabetic mice than in the age-matched controls. Repeated intracavernous injections of COMP-Ang1 protein completely restored the expression of endothelial cell-cell junction proteins in the diabetic mice (Fig. 7A–E).

Endothelial cell-cell junctions serve as a barrier by regulating paracellular permeability (30–33). Thus, we examined cavernous endothelial permeability in age-matched control and STZ-induced diabetic mice 2 weeks after treatment. We used 40 kDa rhodamine-dextran because we found that significant leakage of the 40-kDa fluorescent tracer occurred across the cavernous endothelium in STZ-induced diabetic mice, whereas control mice showed minimal leakage (J.-K. Ryu, W.J. Kim, S. Piao, H.-R. Jin, and J.-K. Suh, unpublished observation). We noted in a significant increase in rhodamine-dextran fluorescence in the nonvascular area of the corpus cavernosum tissue in untreated and PBS-treated mice compared with that in control mice. Repeated intracavernous injections of COMP-Ang1 protein significantly decreased cavernous endothelial permeability in the STZ-induced diabetic mice, which was similar to the level found in age-matched controls (Fig. 7F and G).

We determined whether NOS or NADPH oxidase participated in COMP-Ang1–induced cavernous angiogenesis and subsequent restoration of the erectile response in l-NAME– or apocynin-treated STZ-induced diabetic mice. Histologic and physiological erection studies indicated that l-NAME–treated STZ-induced diabetic mice that received COMP-Ang1 protein failed to restore cavernous endothelial content or recover erectile function (Supplementary Fig. 5). In contrast, treatment with apocynin did not block the COMP-Ang1 protein–induced enhancement of cavernous angiogenesis and restoration of erectile function. Instead, a partial increase in cavernous endothelial content or improvement in erectile function was noted in the apocynin-treated STZ-induced diabetic mice, but it did not reach the level of the COMP-Ang1 treatment group. Apocynin significantly decreased superoxide anion production in the diabetic corpus cavernosum tissue (Supplementary Fig. 5). We observed a significant reduction in plasma NO_x levels in l-NAME–treated diabetic animals (Supplementary Fig. 6). The administration of l-NAME or apocynin did not influence the body weight and blood glucose levels in the STZ-induced diabetic mice (data not shown). These findings suggest that the recovery of erectile function promoted by COMP-Ang1–induced cavernous angiogenesis depends on NOS but not on NADPH oxidase.

We showed here that two successive injections of COMP-Ang1 recombinant protein into the corpus cavernosum of STZ-induced diabetic mice significantly increased cavernous endothelial cell proliferation, eNOS phosphorylation (Ser1177), and cGMP expression and decreased the production of ROS, such as superoxide anion and peroxynitrite. These changes restored erectile function in STZ-induced diabetic mice. Intracavernous injection of COMP-Ang1 protein also induced complete recovery of endothelial cell-cell junction proteins and decreased cavernous endothelial permeability in STZ-induced diabetic mice (Supplementary Fig. 7).

To determine whether the improved erectile response was related to an increase in the number of endothelial cells, we assessed the expression of PECAM-1 by immunohistochemical analysis. In agreement with the findings of previous studies in STZ-induced diabetic rats (12,34,35), we found a significant decrease in the cavernous endothelial area in untreated and PBS-treated STZ-induced diabetic mice compared with that in control mice. COMP-Ang1 protein induced complete recovery of the PECAM-1–positive endothelial area in the corpus cavernosum. Immunohistochemical examination of phosphohistone H3 expression revealed that the COMP-Ang1–induced increase in cavernous endothelial content was the result of endothelial proliferation, which is consistent with a recent report (22).

In this study, COMP-Ang1 failed to induce cavernous angiogenesis and restore erectile function in l-NAME–treated diabetic mice, suggesting that the recovery of erectile function mediated by COMP-Ang1–induced cavernous angiogenesis depends on NOS, similar to results of a previous study (36) showing that Ang1-induced angiogenesis was abolished or reduced in the presence of the NOS inhibitor l-NAME or in Nos3^−/− mice. The NO/cGMP system is the principal mediator of penile erection (37). In the current study, eNOS phosphorylation (Ser1177) was significantly decreased in both untreated and PBS-treated STZ-induced diabetic mice compared with the control group. A previous study reported that eNOS is inactivated in the penis of STZ-induced diabetic animals by O-linked N-acetylglucosamine modification, specifically at Ser1177 (38), and by increased ROS generation (12). In the current study, endogenous eNOS phosphorylation was completely restored in the STZ-induced diabetic mice by the intracavernous injection of COMP-Ang1 protein. Thus, regeneration of cavernous endothelial cells and a decrease in the generation of superoxide anion and peroxynitrite in endothelial cells by COMP-Ang1 may account for the restoration of eNOS activity. Intracavernous administration of COMP-Ang1 protein also significantly increased the concentration of cGMP, although not to the level found in the age-matched controls. The increased formation of cGMP resulting from recovery of NO bioavailability is clear evidence of improved erectile response to cavernous nerve stimulation in the diabetic mice.

The penises from STZ-induced diabetic mice had higher superoxide anion and peroxynitrite levels than did those from the controls. Endothelial cells generate ROS, which plays a crucial role in physiological and pathological conditions. Low levels of ROS function as signaling molecules to mediate endothelial cell proliferation and migration in response to hypoxia, ischemia, and angiogenic factors, such as VEGF-A and Ang1 (39,40). A previous study (41) in human umbilical vein endothelial cells reported that Ang1 triggers the production of superoxide anion by activating NADPH oxidase and that this generation of superoxide anions promotes endothelial cell migration. Furthermore, endothelial NADPH oxidase–derived ROS plays a critical role in Ang1-induced angiogenesis in porcine coronary artery endothelial cells, and targeted deletion of the NADPH oxidase subunit p47^phox gene suppresses Ang1-induced angiogenesis (41), which suggests that ROS plays an important role in promoting Ang1-mediated angiogenesis in physiological conditions. In contrast, high levels of ROS in pathological conditions induce apoptosis and cell death (42). In pathological conditions in which ROS is already generated at high levels, as in the current study, COMP-Ang1 profoundly reduces the generation of superoxide anion and peroxynitrite in endothelial and smooth-muscle cells, possibly by inhibiting the expression of cavernous p47^phox and iNOS, respectively (Supplementary Fig. 7). In agreement with these findings, intracavernous administration of COMP-Ang1 protein significantly reduced the number of apoptotic cells in the cavernous endothelium and smooth muscle in STZ-induced diabetic mice. In contrast to previous findings in physiological conditions in vitro (41), inhibition of NADPH oxidase by apocynin did not block COMP-Ang1–mediated cavernous angiogenesis in STZ-induced diabetic mice in vivo. Instead, we observed partial recovery of cavernous endothelial content and erectile function in STZ-induced diabetic mice treated with apocynin, although not to the level found in the COMP-Ang1–treated group.

In the current study, cavernous endothelial permeability to a 40-kDa fluorescent tracer was increased in STZ-induced diabetic mice concurrently with decreased expression of endothelial cell-cell junction proteins, including VE-cadherin, ZO-1, occludin, and claudin-5. Similar to our findings, several investigators (30–33,43,44) also reported an increase in vascular permeability in STZ-induced diabetic retinal or cerebral microvasculature coincident with a decrease in endothelial junction proteins. These findings suggest that alterations in endothelial junction proteins in diabetes may be a mechanism by which cavernous endothelial permeability is increased. Intracavernous injection of the COMP-Ang1 protein significantly decreased cavernous endothelial permeability in STZ-induced diabetic mice by restoring endothelial cell-cell junction proteins. Specific activation of Tie2 in cell-cell contact in the quiescent vasculature also is known to be a mechanism by which Ang1 reduces endothelial permeability and promotes vessel stabilization (29).

Because of concerns about the utility of viral vector–mediated gene therapy for non–life-threatening disease, such as erectile dysfunction, it is necessary to evaluate whether an angiogenic factor can restore erectile function when it is given as a recombinant protein. Compared with that in STZ-induced diabetic animals that received a single intracavernous injection of ad–COMP-Ang1, the similar restoration of erectile function after two successive injections of COMP-Ang1 protein is noteworthy. From a clinical standpoint, therapeutic proteins are safer and more feasible for treating humans than is gene therapy.

In summary, intracavernous delivery of the recombinant COMP-Ang1 protein successfully restored the endogenous NO-cGMP pathway through the regeneration of healthy and nonleaky cavernous endothelium and inhibition of ROS-mediated cavernous endothelial cell apoptosis. Furthermore, these changes resulted in physiologically relevant changes in erectile function. These findings support the concept of angiogenic factor therapy as a curative therapy for erectile dysfunction. Cavernous endothelial regeneration by use of angiogenic factor protein is a novel therapeutic strategy for treating erectile dysfunction resulting from diabetes.

This work was supported by a Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MEST) (R0A-2007-000-20018-0 to J.-K.S.) and by a grant from the Korea Healthcare technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A084511 to J.-K.S.).

H.-R.J. wrote the manuscript and researched data. W.J.K. researched data and contributed to discussion. J.S.S., S.P., M.J.C., M.T., S.H.S., and G.N.Y. researched data. G.Y.K. contributed to discussion. J.-K.R. wrote the manuscript and reviewed and edited the manuscript. J.-K.S. reviewed and edited the manuscript.

The authors thank Jennifer Holmes, Medical Editing Services Co., for help in preparing the manuscript.