| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,,
Departments of
* Dermatology,
Surgery,
Immunobiology,
Pathology and Interdepartmental Program in Vascular Biology and Transplantation, Yale University School of Medicine, New Haven, Connecticut, USA
1Correspondence: Department of Dermatology, Yale University School of Medicine, P.O. Box 208059, New Haven, CT 06520-8059, USA. E-mail: jeffrey.schechner{at}yale.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: transplantation • HUVEC • vascularization
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and are used to treat wounds due to burns, trauma, surgical excisions, nonhealing ulcers, and blistering diseases (2
3
4
5
6
7)
. Although these products improve wound healing, long-term engraftment has not been demonstrated (8)
. It is likely that inadequate perfusion in the post-transplantation period accounts for lack of engraftment. Whereas autologous split thickness skin grafts can become perfused in a matter of days by inosculation of preexisting graft vessels with those of the recipient, avascular skin equivalents must become perfused entirely by neovascularization from the wound bed. Under ideal circumstances, neovascularization requires 14 days or more, during which time the graft is entirely dependent on diffusion for provision of oxygen and nutrients (9
, 10)
. Since grafts are often used for recipients with compromised angiogenesis (e.g., diabetes or the aged), the time to vascularization may be even more prolonged, increasing the rate of graft failure.
Various strategies have been explored to accelerate vascularization. For example, angiogenesis can be enhanced in human skin equivalents implanted into mice by local delivery of soluble proangiogenic molecules such as vascular endothelial growth factor (VEGF) (10
, 11)
. A limitation of this approach is that vessels induced by VEGF in the absence of other factors, not all of which are known, are prone to dysfunction (12
13
14
15)
. Another promising strategy is to construct grafts that contain cultured human endothelial cells (EC). There has been recent success in forming stable human endothelium-lined, capillary-like structures in bilayered living skin equivalents in vitro (16)
that persist after transplantation onto immunodeficient mice (17)
. However, the capacity of these structures to inosculate with the recipient circulation and provide effective perfusion in vivo has not been demonstrated. Moreover, a potential limitation of the cell transplantation approach is that isolated EC may not provide sufficient information for rapid organization into a mature vascular bed containing microvessels appropriately sheathed by pericytes and smooth muscle cells.
We have improved upon simple EC transplantation by using genetic manipulation to engineer EC for enhanced survival and vascular remodeling. We have previously reported that retroviral-mediated overexpression of the anti-apoptotic gene Bcl-2 fulfills these requirements (18)
. Bcl-2 is normally up-regulated in endothelial cells after exposure to a variety of proangiogenic stimuli (19
20
21)
. Retroviral-mediated overexpression of Bcl-2 in human endothelial cells prevents involution of capillary networks formed from human EC in 3-dimensional matrices (22)
and increases the density of perfused vessels formed in these matrices in vivo (20)
. An unexpected effect of Bcl-2 overexpression is a dramatic enhancement of remodeling of synthetic human vascular beds implanted into immunodeficient mice (18)
. The effects of Bcl-2 transduction included investiture of primitive human EC-lined tubes with mouse mesenchymal cells and evolution of these sheathed tubes into structures that morphologically resemble true arterioles, capillaries, and venules.
In the current study, we first investigated whether the observed enhancement of EC survival and vascular remodeling conferred by Bcl-2 overexpression in human EC in a simple matrix could be extended to EC introduced into a true tissue matrix. Specifically, we seeded human devitalized dermis with either Bcl-2- or control EGFP-transduced HUVEC before implantation into mice. We then demonstrated that this modification augments the perfusion of functional epithelialized human skin equivalents.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and were serially cultured on gelatin-coated flasks in M199/20%FBS supplemented with glutamine, penicillin/streptomycin (Gibco), and ECGS (Calbiochem, La Jolla, CA, USA) and incubated at 37°C in 5% CO2.
Transduction of HUVEC
Stable transduction of HUVEC with a caspase-resistant form of Bcl-2 was achieved as described previously (18)
. Briefly, HUVEC were infected daily for a total of four times with a supernatant containing the packaging virus D34A, Bcl-2 in the pSG5 expression vector, and Polybrene (Gibco).
Preparation of acellular dermis
Cadaveric donor skin obtained from the Yale Skin Bank (discarded specimens) was rinsed in PBS (Gibco) with antibiotics, subjected to three rapid freeze-thaw cycles in liquid nitrogen, then incubated in PBS with antibiotics at 37°C for 2 wk, after which the epidermis was gently removed. Dermal pieces were incubated in PBS with antibiotics for 30 total days, then stored at –20°C until use.
Preparation of engineered skin equivalent
Thawed 1 cm2 pieces of acellular dermis were placed in 3 mL of KGM-2. After the dermis was rehydrated for at least 1 h at 37°C,the KGM-2 was removed. Next, 3 x 105 keratinocytes were pipetted in a 30 µL droplet of KGM-2 onto the center of the dermis. After 3 h, the graft was covered with KGM-2 and the medium was changed to fresh KGM-2 on the following day. Three days after seeding, a differentiation medium was used consisting of KGM-2, DMEM, Ham’s F-12 (Gibco), chelated FBS, cholera toxin (1x10-10 M, Calbiochem), and hydrocortisone (0.4 µg/mL; BD Biosciences, Bedford, MA, USA) plus antibiotics, with a final calcium concentration of 1.2 mM. The differentiation medium was changed every other day until addition of HUVEC 6–10 days after seeding the keratinocytes. HUVEC (8x105 per graft) were introduced to the reticular dermis via a 1cm2 cloning disk or in a 30 µL droplet in M199/20%FBS/ECGS with hydrocortisone (0.4 µg/mL), cholera toxin (1x10-10 M), epidermal growth factor (10 ng/mL; Becton Dickinson, Bedford, MA, USA), and antibiotics (droplet). After 3 h, the cloning disks were removed (if used) and more medium was added. Grafts for subcutaneous transplantation were prepared by seeding acellular dermis with HUVEC as above, omitting the use of keratinocytes. Twenty-four hours after HUVEC were seeded, all grafts were transplanted to mice.
Transplantation
Animal surgery and handling were performed according to protocols approved by the Yale IACUC. Graft sites on the backs of SCID beige CB-17 mice (Taconic, Tarrytown, NY, USA) were prepared by first removing all visible fur with a depilatory (NairTM, Carter-Wallace, New York, NY, USA). A 1 cm2 piece of mouse skin was removed to the level of fascia and a size-matched graft was sutured into the defect. The grafts were then covered with Bacitracin ointment and a waterproof sutured dressing consisting of two layers of 1.5 cm2 TelfaTM (Kendall, Mansfield, MA, USA), TegadermTM (3M, St. Paul, MN, USA), a foam bandage (Stop and Shop, Boston, MA, USA), and circumferentially wrapped DuraporeTM tape (3M). On the day of surgery and each day thereafter, 0.5–1.0 mL of MCDB with L-glutamine, antibiotics, and amphotericin (Gibco-BRL) was injected into the Telfa. The bandages were kept intact for 2 to 6 wk, at which time the mice were killed. In some experiments, 40 to 60 min prior to killing a mouse, 150 µL of a rhodamine-UEA-I conjugate (Vector Laboratories, Burlingame, CA, USA) was administered via tail vein injection. The mice were killed at times indicated in the text and the grafts were harvested. Grafts not seeded with keratinocytes were transplanted subcutaneously. A 1.5 cm incision to the level of fascia was made on the lateral abdomen. A subcutaneous pocket was created by blunt dissection into which the graft was then placed. The mouse skin was closed with surgical staples. The mice were killed and grafts harvested at the indicated times after implantation.
Histochemistry
A portion of each graft was fixed in formalin and paraffin embedded for staining with hematoxylin and eosin (H+E). The other portion was snap frozen in OCT (BD Biosciences, Franklin Lakes, NJ, USA) and 4 µ cryostat sections were prepared for immunohistochemistry. Single antibody staining was performed on 4 µ-thick frozen sections or 5 µ paraffin sections using rabbit anti-human involucrin (Biomedical Technologies, Inc., Stoughton, MA, USA), anti-smooth muscle 
-actin (Novocastra Laboratories, Newcastle upon Tyne, UK), mouse anti-human CD31 (JC 70A, Dako, Carpinteria, CA, USA), mouse anti-human laminin, and anti-human collagen IV, rat anti-mouse CD31 (BD Biosciences). Biotinylated Ulex europaeus agglutinin I (UEA-1, Vector Laboratories) and BS-1 lectin (Vector Laboratories) histochemical stains were also performed. Single staining was followed by a light hematoxylin counterstain. Double staining was performed on paraffin sections using anti-smooth muscle actin and biotinylated UEA-1.
Data analysis
All quantitative data were obtained by an investigator (J.S.S. and S.C.) blinded as to specimen identity. Numbers of vessels in paired specimens were manually counted, and the total dermal area was measured using NIH image analysis software. Statistical analysis was performed using a 2-tailed paired t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). Within 24 h of seeding the grafts, a confluent layer of HUVEC adherent to the cut underside of the dermis was observed and early migration into probable preexisting vascular channels was noted (Fig. 1B
). No differences in the adhesion of EC or invasion into the grafts were appreciated between the two experimental groups (Bcl-2 EC and EGFP EC) in vitro. Grafts seeded with the Bcl-2- or EGFP-transduced EC (n=9 in each group) were then implanted subcutaneously into the mice and harvested after 30 days. The majority of grafts in both experimental groups contained numerous perfused vascular structures (Fig. 1C, D
), all of which were reactive with Ulex europaeus 1 agglutinin (UEA-1), indicating that the endothelial lining was of human origin (Fig. 1E, F
). Anti-Bcl-2 staining confirmed persistent transgene expression (Fig. 1G
). Control grafts not seeded with EC showed continued absence of UEA-1 reactivity, confirming there were no residual cutaneous vessels (Fig. 1H
, inset). Unseeded grafts showed no significant ingrowth with mouse vessels during this period (Fig. 1H
). There was a statistically significant increase in the density of vascular profiles in the Bcl-2-transduced group (55.0±21.8/mm3) compared with the EGFP-transduced controls (25.3±12.6/mm3, P=0.026, Fig. 2
). Vessels formed from Bcl-2- or EGFP-transduced cells each contained smooth muscle -actin-positive investing cells as well as reactivity with human specific antibodies directed at basement membrane components laminin and type IV collagen, features indicative of vascular maturation (Fig. 3
A–D). However, more intense anti-smooth muscle -actin reactivity with increased cellularity of the investing layers was observed in the Bcl-2-transduced group, indicating accelerated maturation (Fig. 3
C, D). These data suggest that overexpression of Bcl-2 in human EC enhances the vascularization of a tissue matrix in vivo by increasing the number of vessels and expediting maturation. Multilaminated vascular structures that resembled arterioles were observed in a limited number of grafts (n=6) containing Bcl-2- or EGFP-transduced or uninfected HUVEC, which were allowed to remain in animals for 60 days (Fig. 3E, F
), indicating the continued maturation and persistence of human vessels. In this case, potential beneficial effects of Bcl-2 transduction were not quantified because of the limited number of specimens. Collectively, these observations confirm the capability for developing complex microvessels from cultured human EC within a true tissue matrix.
|
|
|
Vascularization of human skin equivalents
To test whether this methodology could be applied to a functional engineered tissue, Bcl-2-transduced HUVEC were used to produce a vascularized and an epithelialized skin equivalent. Devitalized dermis again served as the tissue matrix. In these experiments grafts were first seeded on their upper surface with keratinocytes, which were induced to stratify and differentiate by selective media exposure. The epithelialized grafts were then seeded on their underside with Bcl-2-transduced HUVEC. In vitro migration of EC into these grafts was not qualitatively different than that observed in grafts that did not contain keratinocytes (data not shown). Histologic analysis of grafts harvested 2, 4, and 6 wk after implantation into cutaneous wounds on the backs of mice revealed that the majority of grafts were epithelialized and contained numerous perfused blood vessels (Fig. 4
A–C). Antibodies directed against human involucrin were used to determine the extent of coverage with human keratinocytes. Positive reactivity for this antigen typically found in outer epidermal layers indicated the presence of a well-differentiated epidermis in the majority of grafts (Fig. 5
B). Staining with the Griffonia (Bandeiraea) simplicifolia 1 (BS-1) lectin confirmed that the murine keratinocytes did not significantly contribute to the epithelialization (Fig. 5A
). Anti-human and anti-murine specific CD31 antibodies were used to determine the identity of blood vessels. At 2 wk, grafts were primarily perfused through human endothelium-lined vessels, with murine vessels detectable only at the edges of the implant (Fig. 4D, E
). Murine vessels were similarly limited to the periphery of grafts, which were not seeded with human endothelial cells, demonstrating the lack of a significant early recipient angiogenic response (Fig. 5C
). The presence of erythrocytes within the lumena of the human endothelium-lined vessels and binding of intravenously injected fluorescently labeled UEA-1 lectin demonstrated inosculation with the murine vasculature (Fig. 6
A, B). Human endothelium-lined vessels persisted in epithelialized grafts at 4 and 6 wk, despite the progressive ingrowth of murine vessels (Fig. 4F-I
), and acquired several characteristics of mature vessels such as reactivity with antibodies directed against basement membrane components laminin and type IV collagen and investiture with smooth muscle -actin-expressing cells (Fig. 6C, E, F
). The expression of Bcl-2 in these vessels persisted for at least 6 wk (Fig. 6D
). Thus, vessels formed from seeded EC not only accelerate the development of perfusion, but persist in the stably engrafted constructs.
|
|
|
Engraftment was considered successful when the epithelium was continuous and predominantly of human origin and when there were multiple human endothelium-lined vessels per high-powered field extending through at least the bottom two-thirds of the dermis. By these criteria, 12/16 (75%) skin grafts examined 2 wk, 3/5 grafts examined 4 wk, 4/5 grafts examined 6 wk, and 3/4 examined 8 wk after implantation (73.3% of 30 grafts overall) were epithelialized with human keratinocytes and perfused through human endothelium-lined vessels. Using the same criteria, pilot studies were performed to assess the necessity of Bcl-2 transduction of endothelial cells for successful engraftment. In these experiments, only 1/6 untransduced and 2/6 EGFP transduced (examined at 2 or 4–6 wk after grafting, respectively; 25% overall) were vascularized and epithelialized (not shown). Together, these data indicate that early perfusion of functional bilayered skin equivalents can be consistently promoted through vessels formed by transplanting cultured human endothelial cells genetically modified for enhanced vascular density and maturation.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 20)
, there was a > twofold increase in the density of vessels perfusing human dermal matrix by Bcl-2 overexpression in EC. Contrary to what we had reported using a simple collagen/fibronectin matrix (18)
, EGFP-transduced cells seeded onto devitalized dermis also demonstrated some capacity to recruit investing cells, forming more mature vascular structures. This finding suggests a role of matrix composition in promoting vascular maturation, but matrix composition does not compensate for the potential benefits of Bcl-2 transduction. Since interactions between mesenchymal and endothelial cells enhance the stability of immature vessels, these observations support a theoretical advantage in early vessel stabilization as well as a benefit in overall tissue perfusion.
Our strategy for producing these vascularized skin equivalents was designed to be clinically applicable; therefore, matrix and cellular components were selected that could reasonable be adapted for such a utilization. Although a variety of complex synthetic matrices have been developed and successfully applied to forming bilayered skin equivalents (1
, 18
, 24
, 25)
, which, in some cases, support the survival of human endothelial cells (16
, 17)
, we chose to use acellular dermis. The devitalization process removes all immunoreactivity while largely maintaining critical basement membrane components that allow epidermal integrin-mediated cellular attachment and polarization (26
, 27)
. It has been purported that largely due to the presence of elastic fibers, this approach better replicates the mechanical properties of skin than synthetic matrixes (28
, 29)
and facilitates vascularization, at least in part through population of preexisting vascular channels with host endothelial cells (26
, 30)
. Furthermore, devitalized dermis is already widely used as a biological dressing to aid the healing of wounds due to surgery, burns, and chronic venous stasis.
The optimal source of endothelial cells for vascularizing skin equivalents remains unresolved. Other investigators have used human dermal microvascular endothelial cells (HDMEC) (17)
and HUVEC (16)
for such purposes. To limit potential immunoreactivity, ideally, autologous cells should be used; but because expansion in culture is necessary to produce adequate numbers of EC for perfusing the grafts, the resultant delayed availability limits this approach to nontraumatic wounds. Furthermore, in the case of autologous HDMEC, a painful and often scarring secondary procedure is required to harvest adequate numbers of cells from adults. It has recently been shown that allograft rejection of human fetal skin grafts is delayed compared with neonatal skin (31)
. The relative immunogenicity of allogeneic HUVEC compared with nonfetal EC in skin grafts is an unresolved issue that is under investigation.
A potential concern of our approach is whether there are adverse effects of retroviral-mediated overexpression of the survival gene Bcl-2. As described previously, a replication-deficient retroviral vector was used to avoid production of infectious virus, and Bcl-2 does not induce malignant transformation of human endothelial cells (18)
. The possibility of interaction of Bcl-2 with transforming genes is under investigation in a more stringent tumorigenesis model based on using SV40 immortalized EC (32)
; to date, exogenous Bcl-2 at the expression levels achieved by retroviral transduction, does not appear to increase carcinogenicity (unpublished results). In the event that retrovirally mediated overexpression of Bcl-2 proves to be unsafe for human applications, nontransduced cells can be used. Preliminary evaluation of bilayered grafts seeded with control EGFP- or nontransduced EC proved that such a strategy is possible but less reliable, indicating that further development of the methodology is likely to be necessary for this approach to be practical for clinical applications. An elucidation of the changes produced by Bcl-2 on the phenotype of the EC may provide important information for optimizing use of untransduced EC.
In summary, we report for the first time that human skin equivalents can be constructed that develop perfusion in vivo through vessels derived from cultured HUVEC. Furthermore, Bcl-2 transduction of EC incorporated into grafts enhanced vascular maturation and perfusion and likely improved the reliability of engraftment. Skin equivalents seeded with these modified EC become perfused prior to murine neovascularization. Therefore, this strategy for perfusing skin equivalents may minimize the need for ingrowth of recipient vessels, which is likely to improve graft performance, particularly when applied to patients with impaired capacity for angiogenesis such as diabetics and the elderly.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication June 11, 2003. Accepted for publication August 6, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. E. Niklason, A. T. Yeh, E. A. Calle, Y. Bai, A. Valentin, and J. D. Humphrey Regenerative Medicine Special Feature: Enabling tools for engineering collagenous tissues integrating bioreactors, intravital imaging, and biomechanical modeling PNAS, February 23, 2010; 107(8): 3335 - 3339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. R. Shepherd, D. R. Enis, F. Wang, Y. Suarez, J. S. Pober, and J. S. Schechner Vascularization and engraftment of a human skin substitute using circulating progenitor cell-derived endothelial cells FASEB J, August 1, 2006; 20(10): 1739 - 1741. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. R. Enis, B. R. Shepherd, Y. Wang, A. Qasim, C. M. Shanahan, P. L. Weissberg, M. Kashgarian, J. S. Pober, and J. S. Schechner Induction, differentiation, and remodeling of blood vessels after transplantation of Bcl-2-transduced endothelial cells PNAS, January 11, 2005; 102(2): 425 - 430. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
