Introduction

Impaired wound healing is one of the major clinical problems in patients with diabetes (Jeffcoate and Harding, 2003). Poor healing of diabetic wounds is characterized by impaired angiogenesis and diminished formation of the granulation tissue. Of these, angiogenesis is considered to play a pivotal role, as it is required for successful wound repair.

In the postnatal period, neovascularization was previously thought to result only from the proliferation and migration of pre-existing, fully differentiated endothelial cells residing within the vessels, a process termed "angiogenesis" (Folkman and Shing, 1992; Risau, 1995). Recently, however, endothelial progenitor cells (EPC) have been isolated from peripheral blood cells, and have been shown to undergo incorporation into new vessels in ischemic regions (Asahara et al., 1997; Badiavas et al., 2003; Tepper et al., 2005). This is a process consistent with "vasculogenesis", a critical mechanism for establishing the primordial vascular network in the embryo (Asahara et al., 1997). Previous studies demonstrated that transplantation of ex vivo-expanded EPC improved the recovery of blood flow and capillary density in a murine hindlimb ischemia model by promoting vasculogenesis (Kalka et al., 2000). Furthermore, topical application of vascular endothelial growth factor (VEGF) protein to cutaneous wounds has been shown to increase neovascularization by mobilizing and recruiting EPC (Galiano et al., 2004). These findings suggest that not only angiogenesis but also vasculogenesis may play a major role in the process of neovascularization during adult wound healing.

Sodium N-6,2'-O-dibutyryl adenosine-3',5'-cyclic phosphate (DBcAMP) is an analog of cAMP, which promotes the production of several cytokines, including IL-6, IL-8, and transforming growth factor-, by keratinocytes and fibroblasts, as well as stimulating the growth of these cells (Zhou and Ono, 2000; Onuma et al., 2001; Takahashi et al., 2004). In Japan, an ointment that contains DBcAMP (Actosin™ ointment; Daiichi Pharmaceutical Co., Ltd, Tokyo, Japan) has been widely used to treat impaired skin ulcers. Clinical studies have shown the favorable effects of DBcAMP on impaired skin ulcers, such as diabetic foot ulcers or decubitus (Kimura et al. (Kimura Y, Suwa M, Sano T, Yamaguchi Y, Yamada A (1997) Post marketing surveillance of bucladesine (DBcAMP) ointment for chronic skin ulcers in Japan. Aust J Dermatol 38(Suppl 2): 245; Second Joint Meeting of the Wound Healing Society and European Tissue Repair Society, 1996), Miyagi et al. (Miyagi T, Nonaka S, Kimura Y, Yamaguchi Y, Yamada A (1997) Safety and efficacy of bucladesine (DBcAMP) ointment in its long-term use for patients with pressure sore in Okinawa District of Japan: an preliminary report. Aust J Dermatol 38(Suppl 2): 246; Second Joint Meeting of the Wound Healing Society and European Tissue Repair Society, 1996), Takehara et al. (Takehara K, Kawara S, Kimura Y, Yamaguchi Y, Yamada A (1997) Long-term safety and efficacy of bucladesine (DBcAMP) ointment in patients with decubitus in Hokuriku District of Japan: an interim report. Aust J Dermatol 38(Suppl 2): 246; Second Joint Meeting of the Wound Healing Society and European Tissue Repair Society, 1996)); however, the mechanism by which DBcAMP influences neovascularization during cutaneous wound healing has not been defined clearly.

In the present study, we show that in an established model of diabetes-associated delayed wound healing (Tsuboi et al., 1992), DBcAMP treatment leads to an enhanced recruitment of EPC and to significant upregulation of angiogenic growth factors, thereby significantly accelerating wound healing via processes involving both angiogenesis and vasculogenesis.

Results

DBcAMP accelerates healing and neovascularization of diabetic wounds

To investigate the effects of DBcAMP on wound closure, we created full-thickness skin wound on the dorsal skin of C57BL/KsJ-db/db (db/db) mice, and applied DBcAMP or saline every second day. On day 14 (Figure 1a), DBcAMP-treated wounds showed more than 90% epithelialization, whereas less than 50% of the wound was epithelialized in the control group. Significantly smaller wound areas were observed in DBcAMP-treated mice both on days 7 and 14 with maximal difference observed on day 14 (Figure 1b; % wound closure on day 7 in the control versus DBcAMP group: 13.311.48 vs 49.506.89%, P<0.0001; on day 14: 36.094.74 vs 96.831.61%, P<0.0001, Figure 1b). Hematoxylin and eosin staining showed thick granulation tissue and re-epithelialization in the DBcAMP-treated wounds; however, thin granulation tissue was found in the wounds of control group (Figure 1c). Several blood vessels were observed in the granulation tissue of DBcAMP-treated wound; however, vessels were scarcely observed in saline-treated wounds. The histological scores of wounds treated with DBcAMP were significantly higher than those for the control group (day 7 in the control versus DBcAMP group: 2.00.45 vs 5.20.49, P<0.0001; day 14: 4.80.37 vs 11.60.25, P<0.0001, Figure 1d).

Figure 1.

Effect of DBcAMP on full-thickness skin wound closure in genetically diabetic mice. DBcAMP or saline was applied on the wound. (a) Macroscopic appearance of wounds receiving different treatments. (b) % wound closure. DBcAMP significantly accelerated wound closure compared with the control (n=5 per group, ^*P<0.001 vs control). (c) Effect of DBcAMP on wound histology at day 14. Photomicrographs show hematoxylin and eosin-stained sections. An improved wound appearance and increased granulation tissue are observed in the DBcAMP-treated wound. The wound bed shows a thin layer of granulation tissue over adipose tissue in the control group. The wound bed shows a thick layer of granulation tissue covered with epithelium in the DBcAMP group. In higher magnifications, functional neovessels are found in the granulation tissue of DBcAMP-treated wound (arrow), but scarcely in control. (d) Histological score. The DBcAMP group had a significantly higher histological score compared with the control group (n=5 per group, ^*P<0.001 vs control).

Full figure and legend (153K)

As the healing of skin wounds requires neovascularization, we next evaluated vascularity of wound granulation tissue. Wound angiogenesis was analyzed by immunostaining of 10-m frozen sections for endothelial lineage marker CD31, to visualize neovascularization. We also evaluated the functional vessels by perfusion with Rhodamine-conjugated BS1 lectin, which is one of the established endothelial markers and stains endothelial cells lining the perfused vasculature (Ii et al., 2005). Neovascularization at the margin of saline or DBcAMP-treated wounds in diabetic mice on days 7 (Figure 2a) and 14 (Figure 2b). DBcAMP treatment significantly enhanced the number of total vessels (CD31+, green fluorescence) as well as the number of functional blood vessels (BS1 lectin+, red fluorescence). Additionally, significantly higher CD31+ BS1 lectin- cells suggested an enhanced neovascularization including recruitment of EPC into the DBcAMP-treated wounds (merge images; green fluorescence). Furthermore, DBcAMP treatment led to the formation of larger and thicker vessels growing widely in the center of the wound at day 14 (Figure 2b), whereas in the control group, only small, narrow new vessels were observed at the wound margin. Thus, DBcAMP significantly enhanced wound vascularity, as assessed by the % fluorescent area of CD31 (day 7 in the control versus DBcAMP group: 3.520.67 vs 7.021.20%, P<0.05; day 14: 7.601.17 vs 15.601.12%, P<0.002, Figure 2c).

Figure 2.

Effect of DBcAMP on the vascularity of granulation tissues at the wound margin. (a, b) Sections immunostained for CD31 (green) show neovascularization at the wound margin after 7 days (a) or 14 days (b). Labeling of functional vessels was performed by injection of Rhodamine-conjugated BS1 lectin (red) before killing. CD31+ and BS1 lectin- area (green fluorescent in merged image) indicate newly formed neovessels or infiltrating CD31+ cells, which include EPCs. At day 7, several CD31+ (and BS1 lectin-) cells are observed in the granulation tissue of DBcAMP-treated wound; however, scarcely observed in control wound (a). There are tiny and narrow new vessels at the wound margin in the control group, whereas larger and thicker vessels are growing widely in the center of the wound in the DBcAMP group at day 14. Original magnification 100. Bar=100 m. (c) Percent fluorescent area. DBcAMP significantly enhanced the vascularity of wounds (n=5 per group, ^*P<0.05 vs control, ^**P<0.002 vs control).

Full figure and legend (261K)

DBcAMP treatment increases the recruitment of EPCs into the wounds

As CD31+ staining may represent capillaries, resident mature endothelial cells as well as recruited EPCs, in the next series of experiments, we evaluated the specific contribution of circulating EPCs to wound neovascularization using an ex vivo-expanded EPC transplantation model. Ex vivo-expanded EPCs, labeled with 1,1' – dioctadecyl – 3,3,3'∧3' – tetramethylindocarbocyanine for tracking, were intravenously injected into the tail vein of mice immediately following the creation of a full-thickness dorsal skin wound in db/db mice. Seven days after wounding and EPC transplantation, mice were killed and their wounds were harvested. To visualize blood vessels in the healing wound, frozen sections were subjected to CD31 staining. Figure 3a–h shows the contribution of circulating EPC to the development of the wound vasculature on day 7. Green fluorescence indicates CD31+ capillaries and red indicates transplanted EPC, whereas double-positive cells are circulating EPC that have differentiated into vessel wall components. EPC were mainly located at the edge of the wound. EPC recruitment into the wounds of the DBcAMP group was significantly increased compared with that in the control group (12.152.21 vs 72.05.49 cells/mm² for control versus DBcAMP group, P<0.0001, Figure 3i).

Figure 3.

Contribution of EPC at the wound edge. (a–h) Representative photomicrographs of the immunostained wound edge at 7 days after creation. Red fluorescence identifies DiI-labeled EPC (a, b, e, f) and green fluorescence indicates CD31+ cells (endothelial cells) (c, g). Double-positive cells are EPC that have differentiated into cells forming the new vessels (d, h). (i) EPC recruitment into granulation tissue was significantly increased in the DBcAMP group compared with the control group (n=5 per group, ^*P<0.001 vs control). Bar=100 m.

Full figure and legend (111K)

DBcAMP upregulates VEGF and SDF-1 production

To identify the mechanisms responsible for the influence of DBcAMP on EPC recruitment, we evaluated the mRNA levels for VEGF and stromal cell-derived factor-1 (SDF-1) , both of which are chemokines known to enhance EPC homing and function, in vivo. Skin tissues were harvested at days 1, 4, and 7 after wounding and assessed for VEGF and SDF-1 mRNA levels by quantitative PCR. Both VEGF (Figure 4a) and SDF-1 (Figure 4b) transcripts were upregulated by wounding in both groups. However, upregulation of VEGF and SDF-1 in DBcAMP-treated wound were significantly greater than saline-treated wound.

Figure 4.

Effect of DBcAMP on VEGF and SDF-1 expression in wound. (a, b) Quantitative RT-PCR of VEGF (a) and SDF-1 (b). Skin samples were harvested at indicated times and subjected to RT-PCR analysis. Topical application of DBcAMP (10^-2 M) upregulates VEGF expression at each time point (a). SDF-1 expression was upregulated by DBcAMP at days 4 and 7 (b) (n=5 per group, ^*P<0.05, ^**P<0.02 vs saline). (c, d) Immunofluorescence staining of macrophage marker, F4/80 and VEGF (c) or SDF-1 (d) in wound granulation tissues 7 days after wounding. Both F4/80+ cells (macrophages) and F4/80- cells (including mesenchymal cells) showed VEGF and SDF-1 expression. Bar=50 m.

Full figure and legend (340K)

Immunohistochemically, VEGF was colocalized with both F4/80+ cells, which represent macrophages, and F4/80- cells, which may include EPCs as well as mesenchymal cells such as fibroblasts (Figure 4c). SDF-1 was also stained in both F4/80+ and F4/80- cells (Figure 4d).

We chose primary cultured adult diabetic dermal fibroblasts and macrophages, which are two cell types contributing to neovascularization during wound healing by releasing angiogenic factors (Singer and Clark, 1999), for in vitro studies (Figure 5). As expected, VEGF mRNA level was downregulated in both fibroblasts and macrophages under high glucose culture condition (35 mM). DBcAMP upregulated VEGF mRNA in a dose-dependent manner, and the repression of VEGF transcript under high glucose condition was reversed by supplementation of 10^-5 M of DBcAMP (Figure 5a). The secreted VEGF protein level in supernatants harvested from cultured fibroblasts or macrophages was also increased by the addition of DBcAMP to the cultures (Figure 5b). SDF-1 transcript in macrophages, but not in the fibroblasts, was upregulated at 6 hours after the addition of DBcAMP (Figure 6a). After 24 hours, however, SDF-1 mRNA was also upregulated in fibroblasts (Figure 6a). We also evaluated the SDF-1 protein levels in supernatants harvested from cultured fibroblasts or macrophages. SDF-1 level in fibroblasts supernatant was increased by the addition of DBcAMP (Figure 6b); however, SDF-1 level in macrophages was lower than the detection limit of assay (<0.069 ng/ml, data not shown).

Figure 5.

Effect of DBcAMP on VEGF mRNA and protein expression in fibroblasts and macrophages. (a) Quantitative RT-PCR. Six hours or 24 hours after adding DBcAMP to the culture medium (0, 10^-6, 10^-5, 10^-4 M) with low (5 mM) or high (35 mM) glucose condition. VEGF transcript is decreased in high-glucose condition. DBcAMP upregulates VEGF mRNA in both fibroblasts and macrophages in a dose-dependent manner (^*P<0.05, ^**P<0.001, ^***P<0.0001). (b) ELISA for mouse VEGF. Samples are harvested 48 hours after DBcAMP treatment with low or high glucose condition. Addition of DBcAMP increased VEGF secretion by both fibroblasts and macrophages (^*P<0.05, ^**P<0.001, ^***P<0.0001). MeanSEM (n=3). Data are representative of three independent experiments.

Full figure and legend (52K)

Figure 6.

Effect of DBcAMP on SDF-1 mRNA and protein by macrophages and fibroblasts. (a) Quantitative RT-PCR. Six hours or 24 hours after adding DBcAMP to culture medium (0, 10^-4, 10^-5, 10^-6 M) under low (5 mM) or high (35 mM) glucose condition. SDF-1 mRNA is upregulated at 6 h in macrophages but not in fibroblasts. However, SDF-1 mRNA in fibroblasts is upregulated at 24 hours after adding DBcAMP. (b) ELISA for mouse SDF-1. Samples are harvested 48 hours after DBcAMP treatment with low or high glucose condition. Addition of DBcAMP increased SDF-1 secretion by both fibroblasts (^*P<0.05, ^**P<0.005). MeanSEM (n=3). Data are representative of three independent experiments.

Full figure and legend (41K)

DBcAMP indirectly promotes EPC migration

To assess the migratory activity of EPC, 10 ng/ml of VEGF or SDF-1 was used as a chemotactic agent. The migratory response of EPC exposed to conditioned medium (CM) from fibroblasts or macrophages treated with different doses of DBcAMP was measured by a modified Boyden chamber migration assay (Figure 7). The number of migrated EPC treated with high-glucose-conditioned CM were significantly decreased compared with low-glucose-conditioned CM; however, both types of CM induced a statistically significant increase of EPC migration when DBcAMP was included in the medium. Enhanced migration in response to DBcAMP+ CM from fibroblasts was significantly attenuated in the presence of anti-VEGF- and anti-SDF-1-blocking antibodies (Figure 7a). On the other hand, the effect of DBcAMP+CM from macrophages was attenuated only by blockade of VEGF, but not of SDF-1 (Figure 7b).

Figure 7.

DBcAMP indirectly induces EPC migration. (a, b) The migratory response of EPC to CM from cultured fibroblasts (a) or macrophages (b) treated with different doses of DBcAMP was measured with a modified Boyden chamber migration assay. The results shown are obtained with condition media from 12 hours cultures. EPC migration was decreased in high-glucose-conditioned (35 mM) CM compared to low-glucose-conditioned (5 mM) CM. DBcAMP significantly induced the migratory activity of EPC in a dose-dependent manner. These effects were inhibited by neutralization of VEGF or SDF-1 in fibroblasts, and VEGF in macrophages (^*P<0.05, ^**P<0.005, ^***P<0.0001). MeanSEM (n=3). Data are representative of three independent experiments.

Full figure and legend (30K)

Discussion

DBcAMP induces various biological effects, including cell proliferation, differentiation, and migration. Previous reports indicate that these effects are mediated by a modulation in mitogen-activated protein kinase activity through protein kinase A-dependent pathway (Cheng et al., 1998; Pueyo et al., 1998; Zhou and Ono, 2000; Onuma et al., 2001; Takahashi et al., 2004).

As delayed wound healing in diabetics is caused by cellular dysfunction and impairment of growth factor production (Lerman et al., 2003), favorable effects of DBcAMP on diabetic wound healing are suggested. In this study, topical administration of DBcAMP to diabetic wound significantly promoted granulation tissue formation and re-epithelialization compared with the control. These findings experimentally endorse the stimulatory effects of DBcAMP on impaired wound healing.

The healing of skin wounds requires neovascularization. Previous studies have shown that neovascularization is impaired in diabetes (Jeffcoate and Harding, 2003; Lerman et al., 2003). Our observations of enhanced vascularity in the DBcAMP-treated diabetic wounds and subsequent accelerated wound healing suggest that both increased neovascularization and re-epithelialization as a target for DBcAMP effects. The present study also shows that topical application of DBcAMP to diabetic wounds increases the recruitment of EPC in the granulation tissue and that EPC primarily localize at the border of the granulation tissue, a site where blood supply is the most insufficient, and thus neovascularization most required. Recent studies have shown that recruitment of EPC in regenerating vasculature is in part mediated by a hypoxic gradient via the enhanced expression of proangiogenic factors like SDF-1 or VEGF (Ceradini et al., 2004; Tepper et al., 2005), which, subsequently, mobilize EPC from the bone marrow and homed to the ischemic sites. Under hypoxic conditions such as in wounds, EPC are stimulated to form organized cell clusters, which then form cord-like vascular structures. These vascular cords undergo canalization and connect to existing vessels (Tepper et al., 2005). These reports are consistent with our present findings in DBcAMP-treated wounds. Only a few EPC were found in the granulation tissue of control wounds. Previous studies have also shown that tissue VEGF levels are decreased in diabetics (Lerman et al., 2003), and EPC from diabetics show functional impairment (Tepper et al., 2002). Our data of delayed wound healing in diabetic mice is thus in accordance with these studies. Homing of circulating EPC to the ischemic tissues is a critical feature in EPC-mediated neovascularization. As VEGF and SDF-1 are known chemoattractant for EPCs, we evaluated the expression pattern of these cytokines in response to DBcAMP. Our data indicating an increased expression of both these cytokines, both at mRNA and protein (histochemically) level and enhanced EPC homing to the DBcAMP-treated diabetic wounds, thus suggest that DBcAMP-induced augmentation of VEGF and SDF-1 provide critical signals for EPC homing and their participation in the process of wound healing. Several types of cells are known to produce VEGF and SDF-1 in response to stimuli encountered in tissue microenvironment. Our data showing that the protein expression of VEGF and SDF-1 in the wounds was colocalized in monocyte/macrophage lineage cells as well as in cells other than monocyte/macrophages thus support this notion. As mentioned above, VEGF and SDF-1 (which are chemokines for EPC) are both upregulated by hypoxia or tissue injury, after which EPC are mobilized and recruited to the sites of regeneration where these factors are being produced. It should be noted, however, that other factors such as inflammatory cytokines produced in response to injury might also enhance the expression of these angiogenic cytokines. We therefore confirmed the specific contribution of DBcAMP in the upregulation of these cytokines, in vitro. Studies confirmed that in macrophages and fibroblasts, DBcAMP treatment not only attenuated the high glucose-mediated inhibition of VEGF an SDF-1 mRNA expression but also enhanced their expression compared to control cells, suggesting that mechanisms of enhanced neovascularization and wound healing in response to DBcAMP involve at least, in part, enhanced expression of these cytokines. Other studies have also shown that transcription of VEGF or other cytokines, which depend on the cAMP/protein kinase A pathway, is usually upregulated from 4 to 24 hours after stimulation with DBcAMP (Cheng et al., 1998; Pueyo et al., 1998). Accordingly, the stimulatory effects of DBcAMP on the production of VEGF and SDF-1 by fibroblasts might likely be mediated by a cAMP/protein kinase A-dependent pathway. In ischemic tissues, SDF-1 expression usually shows a vascular and perivascular distribution, being mainly localized to the endothelial cells, whereas little expression is reported in smooth muscle cells, pericytes, and surrounding stromal cells (Ceradini et al., 2004). However, very recently, it has been reported that SDF-1 produced from dermal fibroblasts increases keratinocyte proliferation (Florin et al., 2005). Furthermore, as we mentioned above, several types of cells including fibroblasts express SDF-1 in wound granulation tissue. Taken together, it indicates that stromal cells including fibroblasts might also be potential sources of SDF-1 when wound healing is promoted by application of DBcAMP. It is noteworthy that in our studies mRNA levels of SDF-1 in macrophages were relatively low and SDF-1 protein expression could only be detected by immunofluorescence and not as secreted proteins in ELISA. These data suggest the possibility that macrophages are low producer of SDF-1 and that they mostly act as the responder cells as they do express SDF-1 receptor CXCR4 (Katschke et al., 2001), and contribute to the SDF-1-induced VEGF expression. It still, however, remains a possibility that macrophages might have a potential to produce SDF-1, as the behavior of peritoneal macrophages, which we used in vitro studies, may not completely resemble the biology of the wound site. We also evaluated the mRNA expression of PDGF, also a known EPC chemoattractant (Keswani et al., 2004); however, DBcAMP did not influence PDGF mRNA levels either in fibroblasts or in macrophages (data not shown).

Our data showing the ability of conditioned media from fibroblasts or macrophages treated with DBcAMP to chemoattract EPC and attenuation of this process when neutralizing antibodies to VEGF and/or SDF-1 were included in culture, which further suggests DBcAMP-induced VEGF/SDF-1 as a mediator of EPC function in wound healing. However, a direct influence of DBcAMP on EPC is still unclear. Recent studies have shown that cAMP mediates SDF-1-dependent migration by upregulating surface expression of CXCR4 (the receptor for SDF-1) in several types of cells (Cole et al., 1999; Kury et al., 2003). Such findings suggest that other potential cytokines may also mediate DBcAMP effects in wound healing.

In conclusion, DBcAMP accelerates wound healing in diabetes, at least partly by upregulating angiogenic cytokines VEGF and SDF-1, thereby promoting EPC recruitment and vasculogenesis in wound granulation tissue. These observations may partly explain a mechanism by which DBcAMP stimulates the clinical healing of skin ulcers, especially in patients with impaired diabetic ulcer. Furthermore, our data suggest that vascular remodeling induced by DBcAMP might have therapeutic potential in not only skin wound healing but also ischemic disorders such as myocardial or limb ischemia.

Materials and Methods

Animals

Genetically diabetic db/db mice were obtained from Clea Japan Inc. (Japan). All procedures were performed in accordance with the guidelines of the Animal Care and Use Committees of Kyoto Prefectural University of Medicine.

Creation of wounds

Mice were between 8 and 12 weeks old at the time of study. The animals were housed in individual cages, and wounds were created as described previously (Greenhalgh et al., 1990) Briefly, after induction of deep anesthesia by intraperitoneal injection of sodium pentobarbital (160 mg/kg intraperitoneal), full-thickness excisional skin wounds were made on the backs of the mice using a 6-mm skin biopsy punch. Each wound was covered with a semipermeable polyurethane dressing (OpSite^®, Smith and Nephew, Massillon, OH). Twenty microliters of DBcAMP diluted with saline (10^-2 M) or saline was applied every second day by injection through the Opsite with a 27-G needle and was allowed to spread over the wound bed. The application of DBcAMP or saline was continued up to the day of wound harvest.

Monitoring of wound healing

A total of five db/db mice were examined at each time point. Wound healing was monitored on days 1, 7, and 14 after wound creation. Photographic images were analyzed using NIH Image software by tracing the wound margin with a high-resolution mouse and calculating the pixel area. Then, wound area data were compared using the paired Student's t-test.

Histological score

Wounds were resected after 14 days and a histological score was assigned by examination in a blinded manner according to the method described previously (Greenhalgh et al., 1990). Briefly, each specimen was given a score between 1 and 12 as follows: 1–3, none to minimal cell accumulation and granulation tissue or epithelial migration; 4–6, thin, immature granulation dominated by inflammatory cells but with few fibroblasts, capillaries, or collagen deposition and minimal epithelial migration; 7–9, moderately thick granulation tissue, ranging from domination by inflammatory cells to more fibroblasts and collagen deposition; and 10–12, thick, vascular granulation tissue dominated by fibroblasts and extensive collagen deposition.

Immunofluorescence staining

For fluorescence analysis of wound vascularity, before being killed animals were perfused with Rhodamine-conjugated BS1 lectin (Sigma, St Louis, MO) 7 or 14 days after creation of wounds. The animals were then perfused with phosphate-buffered saline and vasculature was in vivo fixed by perfusing with 4% paraformaldehyde. The wounds were then harvested from the dorsum of the animals and sectioned. Sections were stained with rat anti-CD31 antibody (1:100) (BD Biosciences, San Jose, CA). Green fluorescence was generated by labeling with FITC–streptavidin (Vector Laboratories, Burlingame, CA) and biotinylated anti-rat antibody (Vector Laboratories). Wound vascularity was analyzed by calculating the % fluorescent area (CD31), as described previously (Jacobi et al., 2002). Immunofluorescence double staining of F4/80 and VEGF or SDF-1 were performed on frozen sections (6-m thick). Sections were stained with rat anti-F4/80 antibody (1:250) (Caltag Labo-ratories, Burlingame, CA) and rabbit anti-VEGF antibody (1:400) (Santa Cruz) or rabbit anti-SDF-1 antibody (1:200) (Torrey Pines Biolabs, Houston, TX). Normal rabbit IgG (Santa Cruz) and normal rat IgG (BD Biosciences) were used as isotype controls. Green fluorescence was generated by labeling with FITC–streptavidin (Vector Laboratories) and biotinylated anti-rat antibody (Vector Laboratories). Red fluorescence was generated with Cy3-conjugated anti-rabbit antibody (Vector Laboratories).

Transplantation of ex vivo-expanded EPC

Just after wounding, 1 10⁶ EPC taken from db/db mice were injected intravenously. To evaluate the accumulation of EPC at the wound site, the cells were labeled with 1,1' – dioctadecyl – 3,3,3',3' – tetramethylindocarbocyanine – labeled acetylated low-density lipoprotein (Biomedical Technologies, Stoughton, MA), as described previously (Yamaguchi et al., 2003). Wounds were harvested on day 7 and embedded for cutting of frozen sections, which were stained with an anti-CD31 antibody as described above. A total of 20 different granulation tissue fields (four sections from each animal) were selected, and the CD31+ (DiI-labeled) EPCs were counted at 100 magnification.

Cell culture

Fibroblasts were isolated from db/db mouse. After removal of epidermal keratinocytes, fibroblasts were digested from the dermal skin layer using 5% collagenase for 30 minutes. Culture was performed in DMEM (Gibco, Grand Island, NY) containing 10% fetal bovine serum (FBS). The explants were grown to confluence and were passaged. After the first passage, fibroblasts were grown in low-glucose-containing media (5 mM) or high-glucose-containing media (35 mM) (D-glucose; Sigma, St Louis, MO). As a control for the osmotic pressure of the high glucose concentrations, D-mannitol (Sigma, St Louis, MO) was added in low-glucose-containing media (5 mM D-glucose plus 30 mM D-mannitol). Fibroblasts were cultured in hyperglycemic medium for 20 days before each experiment. Fibroblast cultures used in this experiment ranged from five to 10 passages.

Ex vivo expansion of EPC was performed as described before (Mallat et al., 2002). In brief, bone marrow cells obtained by flushing the tibia and femur of db/db mouse were plated on rat plasma vitronectin-coated (Sigma, St Louis, MO) culture dishes and maintained in endothelial cell basal medium-2 (Cambrex, East Rutherford, NJ) supplemented with 5% FBS, human VEGF-A, human fibroblast growth factor-2, human epidermal growth factor, insulin-like growth factor-1, and ascorbic acid. After 4 days of culture, non-adherent cells were removed by washing, new medium was added, and culture was continued until day 7.

Macrophages were isolated from the peritoneal cavity of db/db mouse using thioglycollate, as described previously (Stein and Gordon, 1991). Briefly, 4 days after the intraperitoneal injection of 3 ml of thioglycollate, macrophages were harvested from the peritoneal cavity and cultured in RPMI 1640 medium (Cambrex) containing 10% FBS with low-glucose-containing-media (5 mM D-glucose plus 30 mM D-mannitol) or high-glucose-containing-media (35 mM D-glucose) for 4 days before each experiment.

Quantitative real-time RT-PCR

Skin samples were harvested 0, 1, 4, and 7 days after surgery and homogenized in RNA-Stat (Tel-Test Inc., Friendswood, TX). RNA was isolated according to the manufacturer's instructions.

For in vitro studies, fibroblasts and macrophages were cultured in the absence or presence of various doses of DBcAMP and in low or high glucose culture conditions. Cells were harvested after 6 or 24 hours and RNA was extracted using RNA-Stat according to the manufacturer's instructions. Total RNA was reverse transcribed using Taqman Mutiscribe RT Kit (Biosystems, Foster City, CA) and amplification was performed on the Lightcycler (Roche, Indianapolis, IN) with the following primers and probes: VEGF: forward – 5'-CATCTTCAAGCCGTCCTGTGT-3'; reverse, 5'-CAGGGCTTCATCGTTACAGCA-3' and FAM-CCGCTGATGCGCTGTGCAGG-BHQ. SDF-1 – forward, 5'-CCTCCAAACGCATGCTTCA-3'; reverse, 5'-CCTTCCATTGCAGCATTGGT-3' and FAM-CTGACTTCCGCTTCTCACCTCTGTAGCCT-TAMRA. 18S – forward, 5'-CGGGTCGGGAGTGGGT-3'; reverse, 5'-GAAACGGCTACCACATCCAAG-3' and FAM-TTTGCGCGCCTGCTGCCTT-BHQ. The relative level of expression of the target gene mRNAs was calculated by the comparative C_T method, with normalization for the control gene, 18s. Differences of C_T values were calculated for each mRNA by taking the mean value from duplicate reactions and subtracting the mean value of 18s RNA. The fold change in the expression of each target gene by treated cells relative to control cells was calculated as: relative expression=2^C_T.

Preparation of CM

Fibroblasts or macrophages were incubated in 35 mm dishes with 10% FBS/DMEM or 10% FBS/RPMI 1640, respectively, until subconfluent, and were washed in phosphate-buffered saline. Then low- or high-glucose-conditioned DMEM or RPMI 1640 containing 10^-4, 10^-5, or 10^-6 M DBcAMP without FBS was added to the dishes, and supernatants were collected after culture for 48 hours. Normal low-glucose-conditioned culture medium and supernatant from cultures without DBcAMP were used as the controls. The CM thus obtained was subjected to ELISA and was used for the EPC migration assay.

ELISA for VEGF

ELISAs for VEGF and SDF-1 were performed using commercially available from R&D Systems (Minneapolis, Minnesota, USA) following the manufacturer's protocol.

Migration assay

EPC migration was evaluated using a modified Boyden's chamber assay, essentially as described previously, using VEGF and/or SDF-1 as chemoattractants. (Schratzberger et al., 2000). To neutralize the bioactivity of VEGF or SDF-1, CM were preincubated with anti-mouse VEGF antibody (0.2 g/ml) (R&D Systems) or anti-mouse SDF-1 antibody (100 g/ml) (R&D Systems) for 30 minutes at room temperature. Migration was evaluated by counting the mean number of migrating cells in five high-power fields (original magnification 40) per chamber.

Statistical analysis

All results are presented as the meanSEM. Statistical comparisons between two groups were performed by Student's t-test and analysis of variance was used for serial analyses, with P<0.05 being considered to indicate statistical significance. All of the in vitro experiments were repeated at least three times.

Conflict of Interest

The authors state no conflict of interest.

References

1. Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T et al. (1997) Isolation of putative progenitor endothelial cells for angiogenesis. Science 275:964–967 | Article | PubMed | ISI | ChemPort |
2. Badiavas EV, Abedi M, Butmarc J, Falanga V, Quesenberry P (2003) Participation of bone marrow derived cells in cutaneous wound healing. J Cell Physiol 196:245–250 | Article | PubMed | ISI | ChemPort |
3. Ceradini DJ, Kulkarni AR, Callaghan MJ, Tepper OM, Bastidas N, Kleinman ME et al. (2004) Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nat Med 10:858–864 | Article | PubMed | ISI | ChemPort |
4. Cheng T, Cao W, Wen R, Steinberg RH, LaVail MM (1998) Prostaglandin E₂ induces vascular endothelial growth factor and basic fibroblast growth factor mRNA expression in cultured rat Muller cells. Invest Ophthalmol Vis Sci 39:581–591 | PubMed | ISI | ChemPort |
5. Cole SW, Jamieson BD, Zack JA (1999) cAMP up-regulates cell surface expression of lymphocyte CXCR4: implications for chemotaxis and HIV-1 infection. J Immunol 162:1392–1400 | PubMed | ISI | ChemPort |
6. Florin L, Maas-Szabowski N, Werner S, Szabowski A, Angel P (2005) Increased keratinocyte proliferation by JUN-dependent expression of PTN and SDF-1 in fibroblasts. J Cell Sci 118:1981–1989 | Article | PubMed | ISI | ChemPort |
7. Folkman J, Shing Y (1992) Angiogenesis. J Biol Chem 267:10931–10934 | PubMed | ISI | ChemPort |
8. Galiano RD, Tepper OM, Pelo CR, Bhatt KA, Callaghan M, Bastidas N et al. (2004) Topical vascular endothelial growth factor accelerates diabetic wound healing through increased angiogenesis and by mobilizing and recruiting bone marrow-derived cells. Am J Pathol 164:1935–1947 | PubMed | ISI | ChemPort |
9. Greenhalgh DG, Sprugel KH, Murray MJ, Ross R (1990) PDGF and FGF stimulate wound healing in the genetically diabetic mouse. Am J Pathol 136:1235–1246 | PubMed | ISI | ChemPort |
10. Ii M, Nishimura H, Kusano KF, Qin G, Yoon YS, Wecker A et al. (2005) Neuronal nitric oxide synthase mediates statin-induced restoration of vasa nervosum and reversal of diabetic neuropathy. Circulation 112:93–102 | Article | PubMed | ISI | ChemPort |
11. Jacobi J, Jang JJ, Sundram U, Dayoub H, Fajardo LF, Cooke JP (2002) Nicotine accelerates angiogenesis and wound healing in genetically diabetic mice. Am J Pathol 161:97–104 | PubMed | ISI | ChemPort |
12. Jeffcoate WJ, Harding KG (2003) Diabetic foot ulcers. Lancet 361:1545–1551 | Article | PubMed | ISI |
13. Kalka C, Masuda H, Takahashi T, Kalka-Moll WM, Silve M, Kearney M et al. (2000) Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic neovascularization. Proc Natl Acad Sci USA 97:3422–3427 | Article | PubMed | ChemPort |
14. Katschke KJ Jr, Rottman JB, Rush JH, Qin S, WuL, LaRosa G et al. (2001) Differential expression of chemokine receptors on peripheral blood, synovial fluid, and synovial tissue monocytes/macrophages in rheumatoid arthritis. Arthritis Rheum 44:1022–1032 | Article | PubMed | ISI | ChemPort |
15. Keswani SG, Katz AB, Lim FY, Zoltick P, Radu A, Alaee D et al. (2004) Adenoviral mediated gene transfer of PDGF-B enhances wound healing in type I and type II diabetic wounds. Wound Repair Regen 12:497–504 | Article | PubMed | ISI |
16. Kury P, Koller H, Hamacher M, Cornely C, Hasse B, Muller HW (2003) Cyclic AMP and tumor necrosis factor-alpha regulate CXCR4 gene expression in Schwann cells. Mol Cell Neurosci 24:1–9 | PubMed | ISI | ChemPort |
17. Lerman OZ, Galiano RD, Armour M, Levine JP, Gurtner GC (2003) Cellular dysfunction in the diabetic fibroblast: impairment of migration, vascular endothelial growth factor production, and response to hypoxia. Am J Pathol 162:303–312 | PubMed | ISI | ChemPort |
18. Mallat Z, Silvestre JS, Le Ricousse-Roussanne S, Lecomte-Raclet L, Corbaz A, Clergue M et al. (2002) Interleukin-18/interleukin-18 binding protein signaling modulates ischemia-induced neovascularization in mice hindlimb. Circ Res 91:441–448 | Article | PubMed | ISI | ChemPort |
19. Onuma H, Mastui C, Morohashi M (2001) Quantitative analysis of the proliferation of epidermal cells using a human skin organ culture system and the effect of DbcAMP using markers of proliferation (BrdU, Ki-67, PCNA). Arch Dermatol Res 293:133–138 | Article | PubMed | ISI | ChemPort |
20. Pueyo ME, Chen Y, D'Angelo G, Michel JB (1998) Regulation of vascular endothelial growth factor expression by cAMP in rat aortic smooth muscle cells. Exp Cell Res 238:354–358 | Article | PubMed | ISI | ChemPort |
21. Risau W (1995) Differentiation of endothelium. FASEB J 9:926–933 | PubMed | ISI | ChemPort |
22. Schratzberger P, Schratzberger G, Silver M, Curry C, Kearney M, Magner M et al. (2000) Favorable effect of VEGF gene transfer on ischemic peripheral neuropathy. Nat Med 6:405–413 | Article | PubMed | ISI | ChemPort |
23. Singer AJ, Clark RA (1999) Cutaneous wound healing. N Engl J Med 2 341:738–746 | ChemPort |
24. Stein M, Gordon S (1991) Regulation of tumor necrosis factor (TNF) release by murine peritoneal macrophages: role of cell stimulation and specific phagocytic plasma membrane receptors. Eur J Immunol 21:431–437 | PubMed | ISI | ChemPort |
25. Takahashi H, Honma M, Miyauchi Y, Nakamura S, Yamamoto AI, Iizuka H (2004) Cyclic AMP differentially regulates cell proliferation of normal human keratinocytes through ERK activation depending on the expression pattern of B-Raf. Arch Dermatol Res 296:74–82 | Article | PubMed | ISI | ChemPort |
26. Tepper OM, Capla JM, Galiano RD, Ceradini DJ, Callaghan MJ, Kleinman ME et al. (2005) Adult vasculogenesis occurs through in situ recruitment, proliferation, and tubulization of circulating bone marrow-derived cells. Blood 105:1068–1077 | PubMed | ISI | ChemPort |
27. Tepper OM, Galiano RD, Capla JM, Kalka C, Gagne PJ, Jacobowitz GR et al. (2002) Human endothelial progenitor cells from type II diabetics exhibit impaired proliferation, adhesion, and incorporation into vascular structures. Circulation 106:2781–2786 | Article | PubMed | ISI |
28. Tsuboi R, Shi CM, Rifkin DB, Ogawa H (1992) A wound healing model using healing-impaired diabetic mice. J Dermatol 19:673–675 | PubMed | ChemPort |
29. Yamaguchi J, Kusano KF, Masuo O, Kawamoto A, Silver M, Murasawa S et al. (2003) Stromal cell-derived factor-1 effects on ex vivo expanded endothelial progenitor cell recruitment for ischemic neovascularization. Circulation 107:1322–1328 | Article | PubMed | ISI | ChemPort |
30. Zhou LJ, Ono I (2000) Stimulatory effects of dibutyryl cyclic adenosine monophosphate on cytokine production by keratinocytes and fibroblasts. Br J Dermatol 143:506–512 | Article | PubMed | ISI | ChemPort |

Acknowledgments

The DBcAMP used in these experiments was kindly donated by Daiichi Pharmaceutical Co., Ltd (Tokyo, Japan)