Hyperbaric Oxygen and Bone Marrow–Derived Endothelial Progenitor Cells in Diabetic Wound Healing

Etiology of Chronic Diabetic Wounds

The pathophysiology of diabetic lower extremity ulcerations and delayed healing has been well described. Contributing factors include progressive development of a sensory, vasomotor, and autonomic neuropathy leading to loss of protective sensation; deformity that increases plantar foot pressure; and alterations in autoregulation of dermal blood flow. Most significantly, diabetics show earlier development and progression of lower extremity peripheral arterial occlusive disease with a predilection for the trifurcation level of vessels just distal to the knee. In addition, the tissue microcirculation is severely diseased (microangiopathy), even in patients with patent proximal vessels. Impaired host response to infection and other cellular dysfunctions also contribute to the refractory nature of wound healing.

Approximately 20% of diabetic lower extremity ulcers have arterial flow insufficiency as their primary etiology, approximately 50% will have primary diabetic neuropathy, and approximately 30% will have both conditions. ^3,4 Even after correction of large blood vessel dysfunction (revascularization), only about 47% of patients will heal in a span of 20 weeks with standardized treatment, including glycemic control, débridement of necrotic tissue, control of infection, use of moist dressings, and protection from pressure or trauma related to ambulation. ²¹ The probability of healing decreases progressively among patients whose wounds are older than 2 months or larger than 2 cm ² or who have a full–skin thickness ulcer with exposed tendons or deeper tissues. ⁹ Beyond standard treatment practices, the topical application of becaplermin (platelet-derived growth factor BB) and platelet releasate have been shown to increase healing. ^22–24 Bioengineered human skin or dermis equivalents have also shown benefits. ²⁵ These therapeutic modalities have received approval from the US Food and Drug Administration only to treat wounds that do not extend to fascia, muscle, or bone in individuals with adequate arterial flow. The effectiveness of HBO₂ as an adjuvant therapy for the treatment of lower extremity ulcerations in diabetics has been supported in six randomized trials and evaluations by a growing number of independent, evidence-based reviews. ^{16–20,26–34} Although, in general, “an angiogenesis effect” has been implicated, there is only a paucity of information on its basic mechanisms of action.

Vasculogenesis and Angiogenesis in Wound Healing

Normal wound healing proceeds through an orderly sequence of steps, involving removal of necrotic debris and infection, resolution of inflammation, repair of the connective tissue matrix, angiogenesis, and resurfacing. Problem or chronic wounds are those that have failed to follow this sequence and do not achieve a sustained anatomic and functional result. ⁶ The hypoxic nature of all acute wounds has been demonstrated, but when hypoxia is pathologically increased, wound healing is impaired and the rate of wound infection increases. ^{9,21–23,35,36} Local oxygen tensions in the vicinity of the wound are approximately half the values observed in normal, nonwounded tissue. ³⁶ Fibroblast recruitment, collagen deposition, angiogenesis, and intracellular leukocyte bacterial killing are oxygen-sensitive responses involved with normal wound healing. ^37,38 It is now well established that an essential part of normal healing is the formation of new vessels within the provisional wound matrix that is referred to as granulation tissue formation. Neovascularization of wound granulation tissue occurs by the processes of angiogenesis and/or vasculogenesis. ³⁹ Angiogenesis refers to the process by which resident endothelial cells of the wound's adjacent mature vascular network proliferate, migrate, and remodel into neovessels that grow into the initially avascular wound tissue aided by mature stromal cells, such as fibroblasts. Vasculogenesis is a de novo process by which EPCs recruited to the wound differentiate into endothelial cells and give rise to a replacement vascular network. ^39–41 We have identified a critical role for EPCs in ischemic wound healing. ⁴²

It was once believed that vasculogenesis occurred only during embryonic life; however, EPCs have been identified in peripheral blood in adults and increased with traumatic or surgical wounds. ⁴³ EPC mobilization can be stimulated by peripheral ischemia, vigorous exercise, chemotherapeutic agents, and hematopoietic growth factors. ^44–50 EPCs also can be obtained by direct bone marrow harvesting and ex vivo manipulations. ⁵¹ Hematopoietic progenitor cells are typically obtained for the purpose of bone marrow transplantation by administration of chemotherapeutic agents and growth factors. ⁵² Using these agents to obtain EPCs for purposes such as wound healing or therapeutic neovascularization has been considered, but generalized application is thwarted owing to risks such as acute arterial thrombosis, angina, sepsis, and death. ^52–54 Nevertheless, some experimental data suggest that parenteral administration of granulocyte colony-stimulating factor can reduce the rate of amputation and other surgical procedures in patients with diabetic foot infections. ⁵⁵ EPCs given to animals can be incorporated into the foci of neovascularization in surgically induced ischemic organs and limbs. ^44,46,56 Transplantation of autologous bone marrow cells and purified EPCs into ischemic areas, including the wounds of diabetic animals, has been shown to improve perfusion and accelerate healing. ^{47,51,56–58}

Therapeutic Mechanisms for Tissue Hyperoxia Induced by HBO₂

If periwound tissue is normally perfused, a steep oxygen gradient from the periphery to the hypoxic wound center supports a normal wound healing response. ^35,59 Eradicating this gradient will arrest capillary growth. Data pertaining to radiation necrosis suggest that in the setting of microangiopathy, the O₂ gradient across a wound is inappropriately flat. ⁶⁰ HBO₂ will raise arterial O₂ tension to several thousand Torr and tissue tension to ≅300 Torr. ⁶¹ It has been hypothesized that HBO₂ creates a steep gradient, and this increases migration of macrophages into the center of the wound and release of angiogenic growth factors. ⁶⁰ This mechanism remains unproven, however, and some data argue that the O₂ gradient per se is not necessarily an element of refractory wounds. ⁶²

Production of superoxide radical (O₂ ^.) is increased in many organs by hyperoxia. ³⁴ In recent years, we have also found that HBO₂ increases nitric oxide ( ^. NO) production by stimulating several isoforms of nitric oxide synthase (NOS). ^31,36,63 HBO₂ appears to increase ^. NO-containing substances in both the pulmonary and systemic vascular beds. ¹⁸ Humans exposed to 100% O₂ versus air have ≅ 38% more ^. NO in their exhaled breath. ^16,64 Studies from our laboratory indicate that HBO₂ stimulates healing by two mechanisms: synthesis of growth factors within wounds and bone marrow EPC mobilization. ^65,66 The data suggest that both of these processes may be related to augmentation of free radical synthesis by HBO₂.

HBO₂ increases synthesis of several growth factors that could stimulate angiogenesis and production of granulation tissue. Vascular endothelial growth factor (VEGF)-A is the most specific growth factor for neovascularization, and it is increased in experimental wounds by HBO₂. ^67–70 HBO₂ will also stimulate synthesis of basic fibroblast growth factor and transforming growth factor β₁ by human dermal fibroblasts and angiopoietin 2 by human umbilical vein endothelial cells, and it up-regulates platelet-derived growth factor receptor in experimental wounds. ^19,67–70 Whereas hypoxia-inducible factor 1 (HIF-1) is the major mediator of VEGF transcription under hypoxic conditions, VEGF synthesis by macrophages is also stimulated by oxidants and by ^. NO. ^{35,37,38,59,68} Intriguingly, hypoxic stabilization of HIF-1α requires a functional electron transport chain and production of reactive oxygen species. ^68,70,71 Many downstream effects of VEGF are also stimulated via ^. NO, so elevated ^. NO synthesis by HBO₂ may improve VEGF function. ^65,72,73 VEGF and angiopoietin, as well as stroma-derived growth factor 1, influence homing and EPC differentiation to endothelial cells. ^{40,50,57,71,74,75} Therefore, stimulation of growth factor synthesis within wounds may enhance angiogenesis and increase vasculogenesis by improving EPC recruitment and function. This latter potential mechanism is currently under intense investigation in our laboratory.

The bone marrow is the major reservoir for adult organ-specific stem cells, including hematopoietic stem cells and endothelial progenitors. ^44–46,76 Many chemokines or cytokines trigger EPC release via induction of metalloproteinase 9 (MMP-9) in bone marrow, and ^. NO was recently linked to this process. ^50,72,74,77 Using VEGF-A as a proximal stimulus, Aicher and colleagues demonstrated that endothelial NOS becomes activated in bone marrow stroma; ^. NO then S-nitrosylates and activates MMP-9, which releases the stem cell active cytokine, soluble kit ligand (stem cell factor [SCF]). ⁷² This agent shifts endothelial progenitor and hematopoietic stem cells from a quiescent to the proliferative niche and stimulates EPC mobilization to the peripheral blood. ^45–48,71 We found that over a course of 20 HBO₂ treatments, circulating EPCs increased fivefold in a group of patients undergoing prophylactic HBO₂ in anticipation of surgery to reduce their risk of osteoradionecrosis (Figures 1 and 2). ⁶⁶ We have also found that HBO₂ will stimulate EPC recruitment and augment healing of ischemic wounds in mice. ⁶⁵ Our studies in mice demonstrated that HBO₂ increases synthesis of ^. NO in bone marrow, increases the concentration of circulating SCF, and stimulates EPC mobilization. ^65,66 If NOS was inhibited, all of the HBO₂-mediated events were blocked. ^65,66 Therefore, HBO₂ appears to stimulate EPC mobilization by a ^. NO-dependent mechanism. In summary, we believe the sequence of events is as follows: HBO₂ → NOS → ^. NO → nitrosylation of MMP-9 → cleavage of membrane-bound SCF → SCF prompts EPC proliferation and mobilization → EPC released into peripheral circulation.

OPEN IN VIEWER

Figure 1.

Mean CD34⁺ population in the blood of humans before and after hyperbaric oxygen (HBO₂) treatments. Data are the fraction of CD34⁺ cells within the gated population using leukocytes obtained from 26 patients before and after their first, tenth, and twentieth HBO₂ treatment. *Repeated measures one-way analysis of variance, p < .05 versus the pre-HBO₂ first treatment value.

OPEN IN VIEWER

Figure 2.

Colony-forming cells (CFC) in the blood of humans before and after hyperbaric oxygen (HBO₂) treatments. Data are the colonies counted after a 14-day incubation. *t-test performed on each data set before and after the first treatment, p = .036; before and after the tenth treatment, p = .041; and before and after the twentieth treatment, p = .049.

HBO₂ Increases Circulating EPCs In Humans

In initial human studies on the mobilization of EPCs by HBO₂, blood was obtained from 26 patients before and after their first, tenth, and twentieth HBO₂ treatments for osteoradionecrosis prophylaxis (the standard preoperative course of therapy is 20 treatments). Blood leukocytes were harvested and analyzed for the presence of EPCs based on flow cytometry and colony-forming cells (CFCs). Control human subjects included patients breathing 100% O₂ but not pressurized and pressurized attendants not breathing 100% O₂ in whom tissue-level hyperoxia is not achieved. ⁶⁶ In HBO₂-treated patients (but not controls), the CD34⁺ population in blood (see Figure 1) and the number of CFCs in peripheral blood were significantly increased in response to each exposure to HBO₂ (see Figure 2). ⁶⁶ Of note, we did not find elevations in CFCs prior to the tenth and twentieth treatments (see Figure 2), although the numbers of CD34⁺ cells were elevated (see Figure 1). This suggested that only cells recently mobilized by HBO₂ exhibit an increased propensity to grow and form colonies, a subject that is currently being actively investigated in our laboratory.

To specifically address whether the increased CFCs were coming from the fraction of cells expressing CD34, the circulating monocyte population of nine patients before and after their twentieth HBO₂ treatment was isolated and fractionated using paramagnetic polystyrene beads coated with antibody to CD34. In the CD34⁺ fraction, prior to treatment, there were 13 ± 0.3 colonies, and after HBO₂treatment, 23 ± 3 colonies grew (p < .05), whereas in the CD34⁻ fraction, 12 ± 0.7 colonies grew prior to treatment and 13 ± 0.6 (not significant) grew after HBO₂treatment. ⁶⁶ These data indicate that the CD34⁺cell fraction (that contains the EPC pool) was the one that exhibited improved growth potential in response to hyperoxia. We were also interested in characterizing these cell-surface antigenic phenotypes of the cell colonies. Colonies were harvested, washed, and stained with antibodies. Figure 3 shows a typical flow cytometry scatter plot demonstrating that a large number of these cells express chemokine receptor 4 (CXCR4) and vascular endothelial growth factor receptor-2 (VEGFR-2), which are highly specific markers for EPCs. ⁷¹ As CXCR4 is required for EPCs homing to sites of injury or ischemia, ⁷¹ and both CXCR4 and VEGFR-2 are coexpressed on EPCs, ⁷⁵ our findings suggest that the cells mobilized by HBO₂ are EPCs and may be functional in improving neovascularization.

OPEN IN VIEWER

Figure 3.

Expression of surface markers on the progeny of mobilized CD34⁺ cells. The figure shows the results from 50,000 cells. APC = allophycocyanin; CXCR4 = chemokine receptor 4; FITC = fluorescein isothiocyante; VEGF R2 = vascular endothelial growth factor receptor 2.

HBO₂ Increases Circulating EPCs in Mice via NOS

In our initial studies searching for the mechanisms for EPC mobilization with HBO_2,we also evaluated EPCs (assessed as cells that coexpressed CD34 and stem cell antigen 1) in the peripheral blood of mice. We subsequently demonstrated (with a number of specific EPC markers in both diabetic and nondiabetic mice models) that HBO₂ induces EPC release via a ^. NO-mediated mechanism. ^65,66,78 In preliminary studies, mice were exposed to HBO₂ for 90 minutes at up to 2.8 ATA and also to a pressure control, 2.8 ATA pressure, using a gas containing 7.5% O₂ (so that O₂ partial pressure was the same as ambient air, 0.21 ATA O₂). In selected studies, mice were pretreated with intraperitoneal l-nitroarginine methyl ester (l-NAME, to inhibit NOS, 40 mg/kg administered 2 hours prior to pressurization). In anesthetized mice, blood was obtained by aortic puncture and bone marrow was harvested by clipping the ends off a tibia and flushing the marrow cavity with 1 mL phosphate-buffered saline. Exposure to 2.8 ATA O₂ (but not 100% O₂ at ambient pressure or 2.8 ATA pressure at a total O₂ partial pressure of only 0.21 ATA O₂) increased EPC mobilization immediately after treatment, as well as 16 hours post-treatment, and with repeated treatments (Figure 4). ^65,66

OPEN IN VIEWER

Figure 4.

Mean CD34⁺/stem cell antigen 1 (Sca-1)⁺ cells in blood from mice undergoing hyperbaric oxygen (HBO₂) treatment (Tx). From left to right, the bars show the control conditions (air; l-nitroarginine methyl ester [l-NAME]; oxygen/no hyperbarics; hyperbarics-air; and endothelial nitric oxide synthase [eNOS] knock out [KO]-air); the groups with blood collected immediately after HBO₂ (without and with l-NAME pretreatment); the groups with blood collected 16 hours after a single HBO₂treatment (eNOS KO, wild-type mice, wild-type mice with l-NAME pretreatment); and those with blood collected 16 hours after two consecutive HBO₂ treatments (without and with l-NAME pretreatment), respectively. *One-way analysis of variance, p < .05 versus the control group (n per group indicated in parenthesis). ATA = atmospheres absolute.

There is a precedence for rapid mobilization of stem cells from bone marrow, but most emigration is believed to occur after a period of cell proliferation within the marrow niche. ^48,49 We found that the number of EPCs peaked at 16 hours after mice were exposed to 2.8 ATA 100% O₂, and if mice were exposed to 2.8 ATA 100% O₂ for 90 minutes on 2 successive days, the number increased even more (see Figure 4). There was no additional increase in peripheral blood EPCs if mice were exposed to more than two HBO₂ treatments. The leukocyte count in peripheral blood and bone marrow did not increase in response to HBO₂ (Table 1), but there was a significant elevation in CFCs in both blood and bone marrow, indicating the activation of these cells to grow as colonies with the hyperoxia stimulus.

EPC mobilization did not occur in mice lacking genes for eNOS (see Figure 4). If wild-type mice were injected before HBO₂ with the nonspecific NOS inhibitor l-NAME, EPC mobilization also did not occur (see Figure 4). Soluble kit ligand (skit or SCF) was significantly elevated in the peripheral blood of HBO₂-exposed mice, and this elevation was also blocked by l-NAME (see Table 1). These early findings, as well as other subsequent extensive data from our laboratory, support the view that HBO₂ mobilizes EPCs by increasing bone marrow ^. NO synthesis. ^65,66,78

HBO₂ Increases ^.NO Concentration in Bone Marrow

In early studies, we hypothesized that HBO₂ augments stem cell mobilization because it stimulates ^. NO synthesis in the bone marrow, as ^. NO is known to play a central role with stem cell release from bone marrow. ⁷² Anesthetized mice had a small hole drilled into a femur, where a ^. NO-specific microelectrode was placed. ⁶⁵

Mice were placed in a hyperbaric chamber where, while breathing air, baseline measurements were obtained. The chamber was then flushed with 100% O₂ and compressed to 2.4 ATA O₂. The results (Figure 5) demonstrated a pronounced elevation of ^. NO synthesis. If mice were pretreated with l-NAME (40 mg/kg intraperitoneally), no ^. NO signal was detected. ⁶⁵

OPEN IN VIEWER

Figure 5.

Hyperbaric oxygen stimulates nitric oxide (^.NO) synthesis in mouse bone marrow. Data show mean and standard error (SE), depicted by the gray shading; dark vertical (hatched) shading shows the time required to pressurize the mice to 2.4 ATA after flushing the chamber with 100% O₂.

In addition, we observed that HBO₂daily treatments (2.4 ATA 100% O₂ for 90 minutes) improved hindlimb perfusion by laser Doppler imaging after femoral ligation or excision (Figure 6) and increased the bone marrow–derived EPCs within the incisional wounds 3 days postwounding (with four HBO₂treatments) (Figure 7). By day 8 after wounding, ischemic excisional punch biopsy wounds treated daily with HBO₂healed faster (Figure 8). These wound healing improvements were not observed in mice that received intraperitoneal treatment with l-NAME (40 mg/kg) prior to HBO₂, indicating that the improvement in wound healing is mediated by ^. NO. ⁶⁵ In addition, bone marrow and blood obtained from mice treated daily by HBO₂showed significant increases in the EPC progenitor marker VEGFR-2 messenger ribonucleic acid (mRNA) by semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) (Figure 9). ⁶⁵ Figure 9 shows data collected from 30 friend virus B (FVB) mice that were treated with HBO₂ (100% O₂ at 2.4 ATA for 90 minutes, five daily treatments). Each day, animals were harvested, and blood (retro-orbital bleeding) and bone marrow (flushing of tibias and femurs) were collected (three mice for each time point had bone marrow collected and five mice for each time point had blood collected). VEGFR-2 (kdr) mRNA was semiquantitated by RT-PCR. Although VEGFR-2 is also present in mature endothelial cells, the acute rise of the marker in bone marrow and blood strongly suggested an HBO₂-mediated vasculogenic response. Altogether, the data presented above (see Figures 1 to 9 and Table 1) indicate that EPCs mobilized by HBO₂ are likely functional and that HBO₂ mediates the effects on EPCs via ^. NO mechanisms.

OPEN IN VIEWER

Figure 6.

Ischemic hindlimb blood flow by laser Doppler flux (LDF) is significantly improved in friend virus B (FVB) mice by daily hyperbaric oxygen (HBO₂) treatment. Mice underwent femoral ligation, and LDF was measured at days 3 and 7 of HBO₂ treatment (n = 7 mice per group).

OPEN IN VIEWER

Figure 7.

Bone marrow–derived endothelial progenitor cells (BMD EPC) are significantly increased in incisional wounds with daily hyperbaric oxygen (HBO₂) treatment. BMD EPCs were tracked to wounds using the friend virus B (FVB)/Tie-2-LacZ chimeric mice. Wound biopsies were obtained after four HBO₂ daily treatments; HBO₂was started on the day of wounding (n = 3 mice per group). HPF = high-power fields.

OPEN IN VIEWER

Figure 8.

Ischemic excisional wounds close significantly faster with hyperbaric oxygen (HBO₂). Friend virus B (FBV) mice underwent femoral ligation and excisional hindlimb wounds and were treated daily with HBO₂for 8 days, starting on the day of wounding. Wound closure was digitally monitored, and the surface area was calculated using Image J software (National Institutes of Health).

OPEN IN VIEWER

Figure 9.

Reverse transcriptase polymerase chain reaction (RT-PCR) for bone marrow–derived endothelial progenitor cell marker vascular endothelial growth factor receptor 2; kinase insert domain receptor (Kdr) after hyperbaric oxygen (HBO₂) treatment in friend virus B (FVB) mice blood (A) and bone marrow (B). FVB mice underwent HBO₂ treatments, and each day blood and bone marrow were collected for VEGFR-2 RT-PCR; n = mice per time point; the x-axis indicates the number of daily HBO₂ treatments; and the y-axis shows the relative fold increase in VEGFR-2 messenger ribonucleic acid (p < .05, starting at two treatments for blood and three treatments for bone marrow) mkdr = messenger RNA for kinase insert domain receptor.

Summary

Our data indicate that exposure to HBO₂ will mobilize EPCs from the bone marrow in humans and mice and that this appears to be a ^. NO-dependent process. In ischemic and diabetic murine wound models, we have demonstrated that HBO₂ will hasten wound healing and influence the distribution of progenitor cells within the wound microenvironment. Although further study is required, preliminary data indicate that similar mechanisms may be at work in human subjects exposed to HBO₂. It is our current working hypothesis that HBO₂ hastens wound healing because of concurrent events at the bone marrow and wound site. HBO₂ enhances EPC mobilization, mobilized cells exhibit better growth potential, progenitor cell recruitment to ischemic and diabetic wounds is enhanced, and wounds heal faster. HBO₂ may be the only currently described safe, clinically useful tool to enhance EPC biology, with minimal, if any, side effects. These novel findings should be further investigated in humans and detailed mechanisms elucidated in animal models. Information on specific molecular mechanisms may then lead to exploration of synergistic pharmacologic interactions that could enhance the response to hyperoxia treatments.