Pericyte-Derived Dickkopf2 Regenerates Damaged Penile Neurovasculature Through an Angiopoietin-1-Tie2 Pathway

Penile erection is a neurovascular phenomenon that requires well-coordinated interactions among vascular endothelial (VE) cells, smooth muscle cells, pericytes, and neuronal cells (1,2). Erectile dysfunction (ED) affects more than half of men 40–70 years of age (3). A variety of pathological conditions, including vascular risk factors or diseases, neurological abnormalities, and hormonal disturbances, are involved in penile neurovascular dysfunction (4).

Phosphodiesterase type 5 inhibitors enhance the nitric oxide (NO)-cyclic guanosine monophosphate pathway and are currently used as a first-line therapy for ED (1). The reduced responsiveness to phosphodiesterase type 5 inhibitors in patients with neuropathy, severe angiopathy, or both, such as patients with diabetes, may be related to a decrease in the endogenous NO released from the nerve terminal and/or endothelial cells of erectile tissue (5,6). Therefore, curative therapy for advanced ED, including diabetic ED, requires a new therapeutic strategy that re-establishes structural and functional penile neurovasculature and augments endogenous NO bioactivity.

Several proteins have been effective in targeting angiogenesis of the penis in diabetic or hypercholesterolemic animal models of ED, including VE growth factor, angiopoietin (Ang)-1, and Ang4 (7–14). Targeting neural regeneration, brain-derived neurotrophic factor (BDNF) or neurotrophin-3 (NT-3) has shown some effectiveness by restoring neuronal NO synthase (nNOS)-positive neurons (15,16). However, new therapeutic agents are far from development because of incomplete effects; potential adverse effects, such as inflammation; and difficulties engineering a protein for medicine. Moreover, there is no single candidate protein that can solve the complicated underlying pathology, both angiopathy and neuropathy, in diabetic ED. Therefore, the development of a treatment modality targeting these complicated pathologies in refractory ED would be ideal.

Dickkopf2 (DKK2) is a secreted protein containing two cysteine-rich regions that act as a Wnt antagonist by binding LDL receptor–related protein 5/6 (17–19). Recently, DKK2 was found to improve recovery from hind limb ischemia and myocardial infarction by enhancing angiogenesis (20). Moreover, in a corneal angiogenesis assay, DKK2-induced capillaries had more coverage of endothelial cells by pericytes and were less leaky than VE growth factor–induced vessels (20), which suggests that DKK2 promotes mature and stable blood vessel formation. However, the role of DKK2 in the penis has not yet been explored. Therefore, we hypothesized that DKK2 may be a potential target for therapeutic angiogenesis, which ultimately leads to the restoration of physiological erection. We also determined the role of DKK2 in penile neural regeneration.

Here, we demonstrate that the overexpression of DKK2 in DKK2-Tg mice that express mouse DKK2 under the control of the endothelial cell–specific Tie2 promoter/enhancer, or in wild-type (WT) mice via local administration of DKK2 protein into the penis, rescues erectile function under diabetic conditions. This recovery was accompanied by enhanced proliferation of cavernous endothelial cells and pericytes; phosphorylation of endothelial NO synthase, restoration of the integrity of endothelial cell-cell junctions, and decreased cavernous vascular permeability; and enhanced neural regeneration through the secretion of neurotrophic factors. Transcriptome analysis of DKK2 target genes in primary cultures of mouse cavernous endothelial cells (MCECs) revealed that Ang1 expression is downregulated, and the expression of Ang2, an endogenous antagonist of Ang1, is upregulated by high glucose (HG). This effect was reversed by treatment with DKK2 protein. Using an MCEC-mouse cavernous pericyte (MCP) coculture system and an ex vivo neurite sprouting assay, we also found that DKK2-mediated juxtacrine signaling in MCECs-MCPs promotes angiogenesis and neural regeneration though an Ang1-Tie2 pathway. The recovery of erectile function mediated by DKK2 was abolished by the inhibition of Ang1-Tie2 signaling with soluble Tie2 protein (sTie2-Fc).

The primary aim of the current study was to investigate the mechanisms through which DKK2 restores diabetes-induced ED. We used DKK2-Tg mice and administered DKK2 protein into the penis of diabetic mice. Detailed mechanisms were evaluated with WT or DKK2-Tg mice and primary cultures of MCECs, MCPs, and mouse albino neuroblastoma (Neuro2A) cells. All parameters of genetically modified mice and diabetic mice were compared with those of littermate controls.

Eight-week-old male C57BL/6 (Orient Bio) and DKK2-Tg mice (provided by Young-Guen Kwon, Yonsei University, Republic of Korea) were used in this study. DKK2-Tg mice were backcrossed with C57BL/6 mice for at least seven generations (20). The experiments were approved by the Institutional Animal Care and Use Committee of Inha University (Assurance Number INHA 140110–267–1). Diabetes was induced by intraperitoneal injection of multiple low doses of streptozotocin (STZ) (50 mg/kg body wt in 0.1 mol/L citrate buffer, pH 4.5) for 5 consecutive days, as described previously (21). Eight weeks after diabetes was induced, the mice were anesthetized with intramuscular injections of ketamine (100 mg/kg) and xylazine (5 mg/kg) and placed in the supine position on a thermoregulated surgical table.

For the DKK2-Tg study, the mice were distributed into the following four groups: WT controls, DKK2-Tg mice, WT mice receiving STZ (50 mg/kg body wt for 5 days), and DKK2-Tg mice receiving STZ (50 mg/kg body wt for 5 days). Eight weeks after the induction of diabetes, we measured erectile function during electrical stimulation of the cavernous nerve.

For the inhibition study with sTie2-Fc, the mice were distributed into the following four groups: DKK2-Tg mice, STZ-induced WT diabetic mice, and STZ-induced DKK2-Tg diabetic mice receiving subcutaneous injection of dimeric-Fc or sTie2-Fc (4 µg/20 µL; R&D Systems). The dose of sTie2-Fc was determined based on our previous report (22). sTie2-Fc or dimeric-Fc was administered 8 weeks after the induction of diabetes. Two weeks after treatment, we measured erectile function, and then the penis was harvested for histological examination.

To test the efficacy of DKK2 protein, the mice were distributed into the following three groups: age-matched controls and STZ-induced WT diabetic mice receiving repeated intracavernous injections of PBS or DKK2 protein (days −3 and 0; 6 µg/20 µL; A&R Therapeutics). For the inhibition study with sTie2-Fc, the mice were distributed into the following four groups: age-matched controls and STZ-induced WT diabetic mice receiving repeated intracavernous injections of PBS, DKK2 protein (days −3 and 0; 6 µg/20 µL) and dimeric-Fc, or DKK2 protein (days −3 and 0; 6 µg/20 µL) and sTie2-Fc (4 µg/20 µL). sTie2-Fc or dimeric-Fc was administered immediately before the injection of DKK2 protein. We evaluated erectile function by cavernous nerve electrical stimulation 2 weeks after treatment. The penis was harvested for histological examination and biochemical study.

To examine the effect of insulin treatment on erectile function, the mice were distributed into the following three groups: age-matched controls and STZ-induced WT diabetic mice receiving repeated intraperitoneal injection of PBS or insulin (4 international units/day; Sigma-Aldrich) (23,24). Insulin treatment started 1 week after STZ injection (9 weeks of age) and continued for 9 weeks (18 weeks of age). We measured erectile function during electrical stimulation of the cavernous nerve.

Fasting and postprandial blood glucose levels were determined by an Accu-Check blood glucose meter (Roche Diagnostics) before the mice were sacrificed. We also measured glycosylated hemoglobin (HbA_1c; A1C Now System; PTS Diagnostics) and serum insulin levels by using a mouse insulin ELISA kit (Mercodia).

Human corpus cavernosum tissue samples were obtained from a 21-year-old patient with congenital penile curvature who had normal erectile function during reconstructive penile surgery and a 56-year-old patient with diabetic ED (type 2 diabetes mellitus, with a duration of 22 years; HbA_1c level 8.8% [73 mmol/mol]; BMI, 23.1 kg/m²; comorbidity hypertension; medications subcutaneous insulin and oral metformin) during penile prosthesis implantation. All tissue donors provided informed consent, and the experiments were approved by the internal review board of Inha University.

Measurement of erectile function was performed as described previously (21). Details can be found in the Supplementary Data.

The MCECs and MCPs were prepared and maintained as described previously (2,25,26). Tube formation assay, scratch wound-healing assay, transfection assay, RT-PCR, and cDNA microarray were performed as described in the Supplementary Data.

Aortas were harvested from 8-week-old C57BL/6 WT mice. The aortic rings were placed in the eight-well Nunc Lab-Tek Chamber Slide System (Sigma-Aldrich) and sealed in place with an overlay of 50 µL of Matrigel. The aortic rings were cultured in medium 199 with 20 ng/mL basic fibroblast growth factor and 1% penicillin/streptomycin for 5 days. The aortic segments and sprouting cells were fixed in 4% paraformaldehyde for at least 30 min and used for immunofluorescent staining.

The mouse major pelvic ganglion (MPG) tissues were prepared and maintained as described previously (27), with minor modifications. The MPG tissues were isolated from male mice using a microscope, were transferred into sterile vials containing Hank’s balanced salt solution (Gibco), and then rinsed and washed twice in PBS. The MPG tissue was cut into small pieces, and the samples were plated on a poly-d-lysine hydrobromide–coated (Sigma-Aldrich) 12-well plate. The whole MPG tissue sample was covered with Matrigel, and the culture plate was placed on ice for 5 min prior to incubation at 37°C for 10–15 min in a 5% CO₂ atmosphere. We added 1 mL of complete Neurobasal Medium (Gibco) supplemented with 2% serum-free B-27 (Gibco) and 0.5 nmol/L GlutaMAX-I (Gibco). The dishes were then incubated at 37°C in a 5% CO₂ atmosphere. Three days after incubation, we evaluated neurite outgrowth.

To mimic an in vivo or ex vivo condition for diabetes-induced angiopathy and neuropathy, primary cultures or tissues were serum starved for 24 h and then exposed to normal glucose (NG; 5 mmol) or HG (30 mmol; Sigma-Aldrich) conditions for 2 days (MCECs, MCPs, and Neuro2A cells), 3 days (MPG tissue), or 5 days (aortic ring).

To examine the effect of pericyte-derived DKK2 on endothelial cells, the conditioned medium (CM) derived from MCPs in the presence or absence of DKK2 depletion was transferred to MCECs. In addition, CM derived from MCECs in the presence or absence of DKK2 depletion was transferred to MCPs to determine the effect of endothelial cell–derived DKK2 on pericytes. To do this, MCECs and MCPs were grown in 60-mm dishes until 80% confluence, and the medium was changed for an additional 2 days. The culture supernatants were collected and centrifuged for 10 min at 200g to remove cell debris. For DKK2 depletion, an immunoprecipitation protocol was used. Briefly, the CM was incubated with DKK2 antibody (1:200; catalog #Ab95274; Abcam) or control rabbit IgG (1:100; catalog #sc-2027; Santa Cruz Biotechnology) for 12 h at 4°C. Protein G-coupled Sepharose beads (Millipore) were added to the medium and incubated for an additional 12 h at 4°C to remove DKK2 antibody and protein. MCP or MCEC complement medium was used as a control.

We also determined the effect of CM derived from MCEC-MCP coculture on neurite sprouting from MPG tissue. To do this, MCECs and MCPs were cultured and treated under the following conditions: NG; HG + PBS; HG + DKK2 protein (200 ng/mL) + scrambled small interfering RNA (siRNA); HG + DKK2 protein (200 ng/mL) + Ang1 siRNA (200 pmol) transfection in MCECs; HG + DKK2 protein (200 ng/mL) + Ang1 siRNA (200 pmol) transfection in MCPs; and HG + DKK2 protein + Ang1 siRNA (200 pmol) transfection in both MCECs and MCPs. The culture supernatants were collected and centrifuged for 10 min at 200g to remove cell debris and then transferred to MPG tissue.

Histological examinations and BrdU labeling were performed as described in the Supplementary Data.

Western blot analysis was performed as described in the Supplementary Data.

The results are expressed as the mean ± SE. Intergroup comparisons were made by Mann-Whitney U test or Kruskal-Wallis test. P values <5% were considered significant. We used SigmaStat 3.11 software (Systat Software) for statistical analyses.

The fasting and postprandial blood glucose concentrations as well as HbA_1c levels in STZ-treated WT or DKK2-Tg diabetic mice were significantly higher than in control mice. In addition, body weight and serum insulin levels were significantly lower in the STZ-induced diabetic mice than in the controls. However, the body weight and blood glucose levels of the diabetic mice did not differ significantly regardless of treatment (Supplementary Tables 1–4).

The expression of DKK2 mRNA and protein was significantly higher in primary cultures of MCPs and human brain microvascular pericytes (HBMPs) than in MCECs and human umbilical vein endothelial cells (HUVECs), respectively (Fig. 1A–D and Supplementary Fig. 1). Similarly, immunohistochemical staining of normal human or mouse erectile tissue revealed that a significant proportion of the DKK2 expression overlapped with pericytes and partially colocalized with endothelial cells (Fig. 1E–H).

Western blot analysis revealed a decrease in DKK2 expression in the penis tissue of diabetic mice in vivo and in MCECs or MCPs exposed to HG conditions in vitro (Fig. 1I–N). Immunofluorescent staining also revealed that the expression of DKK2 protein was significantly lower in cavernous tissue from patients with diabetes or diabetic mice than in the control groups (Fig. 1O–Q). These findings gave us a rationale to use DKK2 for the treatment of diabetic ED.

Eight weeks after the injection of STZ into WT or DKK2-Tg mice and the induction of diabetes, the cavernous endothelial cell and pericyte content was significantly lower in WT mice that received STZ than in untreated WT mice, whereas cavernous endothelial cell and pericyte content was relatively preserved in DKK2-Tg mice that received STZ (Fig. 2A and G). Moreover, DKK2 profoundly enhanced the proliferation of endothelial cells and pericytes under diabetic conditions both in vivo and in vitro (Fig. 2B and H and Supplementary Fig. 2). Overexpression of DKK2 significantly induced the phosphorylation of Akt and endothelial NO synthase (Supplementary Fig. 3), restored cavernous endothelial cell-cell junction proteins claudin-5 and VE-cadherin (Supplementary Fig. 4), and decreased the extravasation of oxidized LDL (Supplementary Fig. 5) under diabetic conditions compared with WT littermates that received STZ.

We further examined the role of DKK2 in MCEC and MCP monoculture or direct mixed coculture. The mixture of MCECs and MCPs formed well-organized capillary-like structures at a ratio of 3:1 (Supplementary Fig. 6). In vitro Matrigel assays revealed impaired tube formation in MCEC and MCP monoculture or coculture exposed to HG, and these impairments were completely restored by treatment with DKK2 protein (200 ng/mL) (Fig. 2C and I). DKK2 protein also promoted MCEC, MCP, HUVEC, and HBMP migration under HG conditions (Fig. 2D and J and Supplementary Fig. 7). In an ex vivo aortic ring assay, the average length and branch number of outgrowing microvessels were significantly lower in aortic segments exposed to HG than segments exposed to NG conditions. Moreover, DKK2 protein significantly enhanced the outgrowth of microvessels from aortic rings under HG conditions (Fig. 2E and K). We observed higher DKK2 expression in the sprouting front than the aortic ring (Fig. 2F).

Cavernous inducible NO synthase protein expression was significantly higher in the PBS-treated diabetic mice than in the age-matched controls. Repeated intracavernous injections of DKK2 protein (days −3 and 0; 6 µg/20 µL) significantly decreased cavernous inducible NO synthase expression in the diabetic mice (Supplementary Fig. 8A–C). We also performed immunohistochemical localization of nitrotyrosine to determine peroxynitrite generation, which is derived from NO and superoxide anion, and in situ analysis of superoxide anion production. Nitrotyrosine expression and the fluorescent products of oxidized hydroethidine in endothelial cells of the corpus cavernosum were significantly higher in WT mice that received STZ than in untreated WT mice, whereas the generation of superoxide anion and nitrotyrosine was profoundly decreased in DKK2-Tg mice that received STZ. DKK2 protein also significantly decreased cavernous reactive oxygen species production under diabetic conditions (Supplementary Fig. 8D–I).

Immunocytochemical staining revealed higher DKK2 expression in MCPs than MCECs. In contrast, after cultivation of MCECs and MCPs using an indirect noncontact coculture system, DKK2 expression was higher in MCECs than in MCPs (Fig. 3A). To confirm whether pericyte-derived DKK2 migrates into the endothelial cells, MCPs were transfected with DKK2-red fluorescent protein (RFP) DNA. We observed DKK2-RFP expression in MCECs after coculture with DKK2-RFP–transfected MCPs (Fig. 3B).

To test the functional role of pericyte-derived DKK2 on endothelial cells, MCECs were treated with CM derived from MCPs in the presence or absence of DKK2. We observed enhanced tube formation in MCECs treated with MCP-CM compared with cells treated with complement medium for MCPs. However, MCECs treated with DKK2-depleted MCP-CM had profoundly impaired tube formation (Fig. 3C and D). Although DKK2-depleted MCEC-CM slightly decreased tube formation in MCPs compared with cells treated with DKK2 containing CM, it was not significant (Fig. 3E and F).

The expression of neurofilament, βIII tubulin, and nNOS in dorsal nerve bundle or corpus cavernosum was significantly lower in WT mice that received STZ than in untreated WT mice, whereas the neuronal cell content was completely restored in DKK2-Tg mice that received STZ (Fig. 4A and D–F). DKK2 protein also significantly enhanced neurite sprouting in an ex vivo MPG tissue culture exposed to HG (Fig. 4B and G).

Next, we asked whether the effects of DKK2 were mediated by the production of neurotrophic factors and their receptors. The cavernous expression of nerve growth factor (NGF), BDNF, and tropomyosin receptor kinase (Trk) A was significantly higher in diabetic mice receiving DKK2 protein than in PBS-treated diabetic mice and comparable to the level found in age-matched controls. We observed similar results in Neuro2A cells in vitro (Fig. 4C and H–K). The expression of TrkB and TrkC was not detectable in the penis or Neuro2A cells.

In accordance with DKK2-mediated angiogenesis and neural regeneration, the ratio of maximal intracavernous pressure (ICP) or total ICP to mean systolic blood pressure (MSBP) was significantly lower in WT mice that received STZ than in untreated WT mice, whereas erectile function was relatively preserved in DKK2-Tg mice that received STZ (Supplementary Fig. 9A, C, and D).

Two weeks after treatment, repeated intracavernous injections of DKK2 protein (days −3 and 0; 6 µg/20 µL) significantly induced the recovery of erection parameters in WT mice that received STZ (Supplementary Fig. 9B, E, and F). No detectable differences in MSBP were found among the experimental groups (Supplementary Tables 1 and 2). The dosage of DKK2 protein was determined based on the findings of our pilot study. We obtained the highest erectile function recovery at a concentration of 6 µg/20 µL (Supplementary Fig. 10).

To identify the genes regulated by DKK2, microarray analysis was performed. We selected genes for which the ratios changed more than twofold in both conditions (i.e., MCECs exposed to NG conditions compared with those exposed to HG conditions + PBS, and MCECs exposed to HG conditions + DKK2 protein compared with those exposed to HG conditions + PBS). After filtering data from 39,429 genes, 201 genes were changed more than twofold. Of these genes, only three were downregulated under HG conditions compared with NG conditions. These levels were reversed after treatment with DKK2 protein (Supplementary Tables 5 and 6). These genes included Angpt1 (Ang1) and Angpt2 (Ang2) (Fig. 5A).

We further confirmed that Ang1 mRNA and protein expression were significantly lower and Ang2 expression significantly higher in MCECs or MCPs exposed to HG conditions than in cells exposed to NG conditions. The expression of both Ang1 and Ang2 returned to baseline values after treatment with DKK2 protein (Fig. 5B–D).

Physiological erection studies revealed that inhibition of the Ang1-Tie2 pathway by sTie2-Fc (4 µg/20 µL) abolished DKK2-mediated erectile function recovery in both diabetic DKK2-Tg mice and diabetic WT mice treated with DKK2 protein (Fig. 5E–J). No detectable differences in MSBP were found among the experimental groups (Supplementary Tables 3 and 4).

Immunofluorescent staining of penis tissue revealed that the enhanced cavernous angiogenesis and neural regeneration were profoundly diminished in STZ-injected DKK2-Tg mice treated with sTie2-Fc (Fig. 5K, N, and O). Treatment of MCEC-MCP coculture or MPG tissue with sTie2-Fc also abolished the DKK2-mediated enhancement of tube formation and neurite sprouting under HG conditions (Fig. 5I, M, P, and Q).

We also examined the cellular sources of Ang1 that may play a major role in DKK2-mediated angiogenesis and neural regeneration. Compared with MCECs transfected with Ang1 siRNA, the transfection of Ang1 siRNA into MCPs or both MCECs and MCPs profoundly diminished the DKK2-mediated enhancement of tube formation in mixed MCEC-MCP coculture exposed to HG (Supplementary Fig. 11A and B). We also determined whether CM derived from MCECs or MCPs cultivated under the aforementioned conditions affects neural regeneration. Similar to the results of the tube formation assays, DKK2-mediated neurite sprouting was more impaired in MPG tissue treated with CM derived from Ang1 siRNA-transfected MCPs than in the tissue treated with CM derived from Ang1 siRNA-transfected MCECs (Supplementary Fig. 11C and D). These findings suggest that pericyte-derived Ang1 plays a crucial role in DKK2-mediated angiogenesis and neural regeneration.

We finally examined whether insulin treatment rescues erectile function under diabetic conditions. The treatment of diabetic mice with insulin did not restore erectile function (Supplementary Fig. 12A–F). Insulin treatment also failed to restore cavernous DKK2 expression in the diabetic mice (Supplementary Fig. 12G and H). Metabolic and physiological variables, including body weight, blood glucose concentrations, and systemic blood pressure, are summarized in Supplementary Table 7.

Here, we investigated whether DKK2 plays a role as a positive regulator of angiogenesis and neural regeneration, exerting beneficial effects in diabetic ED. DKK2-mediated interactions between pericytes and endothelial cells promoted angiogenesis and neural regeneration through an Ang1-Tie2 pathway, rescuing erectile function in diabetic animals. The detailed mechanisms of action by which DKK2 restores erectile function are illustrated in Fig. 6.

To test whether DKK2 induces cavernous angiogenesis under pathological conditions, we immunohistochemically evaluated the expression of CD31 and platelet-derived growth factor receptor-β (PDGFRβ). Similar to the results of previous studies in STZ-induced diabetic rats (28,29) and mice (2,9,11), the cavernous endothelial cell and pericyte area was significantly smaller in WT diabetic mice than in control mice. BrdU labeling revealed increased endothelial cell and pericyte proliferation, and these cellular contents were relatively well preserved in STZ-treated DKK2-Tg mice. DKK2 protein also promoted tube formation, proliferation, and migration by endothelial cells and pericytes, and enhanced microvessel sprouting from the aortic ring under HG conditions.

Endothelial cell-cell junctions serve as a barrier by regulating paracellular permeability and play a crucial role in vascular formation, the vascular network, and remodeling of blood vessels (30). DM promotes LDL oxidation and extravasation, which in turn induces vascular inflammatory responses and endothelial cell apoptosis (31). We recently revealed impaired cavernous endothelial cell-cell junctions and increased cavernous endothelial permeability to oxidized LDL in diabetic mice (32). In the current study, cavernous endothelial cell-cell junction proteins were well preserved, and less oxidized LDL was extravasated in STZ-injected DKK2-Tg mice than in STZ-injected WT littermates. The restoration of endothelial cell-cell junctions and enhanced pericyte coverage on endothelial cells by DKK2 may be attributable to a decrease in cavernous vascular permeability.

The interaction between endothelial cells and pericytes plays a crucial role in blood vessel formation and vascular maturation (33). The close anatomical relationship between endothelial cells and pericytes implicates paracrine or juxtacrine signaling (i.e., endothelial cell-pericyte signaling). Using human and mouse erectile tissues in vivo, and endothelial cells (MCECs and HUVECs) and pericytes (MCPs and HBMPs) in vitro, we found that pericytes are the major source of DKK2 expression. In indirect noncontact MCEC and DKK2-RFP–transfected MCP coculture experiments, we confirmed that pericyte-derived DKK2 migrates into endothelial cells. Moreover, we found profound tube formation impairment in MCECs treated with DKK2-depleted MCP-CM, whereas the treatment of MCPs with DKK2-depleted MCEC-CM did not significantly affect tube formation. These findings suggest that pericyte-derived DKK2 may play a crucial role in DKK2-mediated angiogenesis.

The number of functional nNOS-positive neurons is critical for physiological penile erection (34). In the current study, DKK2-Tg mice that received STZ had preserved nNOS, neurofilament, and βIII tubulin expression in the corpus cavernosum and dorsal nerve bundle. Moreover, direct administration of DKK2 protein or DKK2 protein–treated CM derived from MCEC-MCP coculture profoundly enhanced neurite sprouting in MPG tissue under HG conditions. In penis tissue from WT diabetic mice and Neuro2A cells exposed to HG, the expression of NGF, BDNF, and TrkA was greatly restored by treatment with DKK2 protein. Given that the activation of Wnt signaling has neuroprotective effects (35,36), the Wnt signaling antagonist DKK2 may exert its neurotrophic effects independent of the Wnt pathway.

Transcriptome analysis showed that Ang1 and Ang2 are target genes of DKK2. Inhibition of the Ang1-Tie2 pathway with sTie2-Fc diminished DKK2-mediated angiogenesis and neural regeneration in STZ-injected DKK2-Tg mice and MCECs or MPG tissue exposed to HG. Pretreatment with sTie2-Fc also abolished DKK2-mediated erectile function recovery in both STZ-injected DKK2-Tg mice and STZ-injected WT mice treated with DKK2 protein. These findings suggest that the Ang1-Tie2 pathway is crucial for DKK2-mediated restoration of the penile neurovascular structure and erectile function. Although the Ang1-Tie2 pathway is well known to play an important role in generating a stable and functional vasculature (37), its role in the nervous system is largely unknown. However, Ang1 has been reported to promote neurite outgrowth in dorsal root ganglion cells positive for Tie2 receptor through the transactivation of TrkA receptor (38).

For clinical applications, a Good Laboratory Practice preclinical study of DKK2 protein has begun. An interim report revealed no toxicity in C.B-17 SCID mice (i.e., no unscheduled death, clinical signs, changes in body weight, or abnormal necropsy findings) after a single intravenous injection of DKK2 protein at a dosage of up to 90 mg/kg body wt (Supporting Data 1 in the Supplementary Data). Also, no tumorigenesis related to DKK2 protein was noted in breast cancer cell lines (MDA-MB-231, MCF-7) or a lung cancer cell line (A549) in vitro (Supporting Data 2 in the Supplementary Data) or in vivo after implantation of a lung cancer cell line (A549) in nude mice (Supporting Data 3 in the Supplementary Data). It was reported that DKK2-Tg mice exhibited increased retinal blood vessel formation during the developmental period (20). DKK2-Tg mice also showed increased blood vessel density and decreased hemorrhage in the oxygen-induced retinopathy model (39), suggesting the clinical utility of DKK2 in diabetic retinopathy.

Our study has some limitations. First, this study did not explain how DKK2 regulates the expression of Ang1 and Ang2. Second, the mechanisms by which the Ang1-Tie2 pathway is involved in neural regeneration need to be documented in detail. Finally, we measured systemic blood pressure by use of a noninvasive tail-cuff system, not by direct carotid artery cannulation, because of concerns about the viability of animals and the reliability of the ICP results with significant bleeding during the carotid artery cannulation. Systemic blood pressure and ICP were not measured simultaneously because vibrations occurring from electrical stimulation of the cavernous nerve can impede the accurate measurement of systemic blood pressure by use of the tail-cuff system. Thus, we measured systemic blood pressure immediately before electrical stimulation of the cavernous nerve.

Taken together, our findings reveal a unique function of DKK2 in reprogramming damaged erectile tissue toward neurovascular repair through an Ang1-Tie2 pathway. Dual angiogenic and neurotrophic effects of DKK2, especially in the form of locally injectable protein, may provide a paradigm shift in the development of novel therapeutics not only for ED, but also for other ischemic vascular diseases or neurological disorders.

Funding. This work was supported by a grant from the Korean Health Technology R&D Project; the Ministry of Health & Welfare, Republic of Korea (grants A110076 [to J.-K.S.] and H15C0508 [to J.-K.R. and H.M.K.]); a Medical Research Center Grant (grant 2014R1A5A2009392 [to J.-K.R. and H.J.P.]); and the National Research Foundation of Korea (NRF), funded by the Korean government (Ministry of Science, ICT and Future Planning, grant 2016R1A2B2010087 [to J.-K.R.]).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. G.N.Y. and H.-R.J. designed and performed the experiments and wrote the manuscript. M.-J.C., A.L., K.G., N.N.M., J.O., M.-H.K., and K.-M.S. performed the experiments. H.J.P. analyzed and critically discussed the data. H.M.K. and Y.-G.K. contributed essential reagents. J.-K.R. and J.-K.S. designed the experiments, supervised the project, and wrote the manuscript. J.-K.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.