Human dermal microvascular endothelial cells form vascular analogs in cultured skin substitutes after grafting to athymic mice

Abstract

Cultured skin substitutes (CSS) consisting of autologous fibroblasts and keratinocytes combined with biopolymers are an adjunctive treatment for large excised burns. CSS containing two cell types are limited by anatomical deficiencies, including lack of a vascular plexus, leading to slower vascularization after grafting than split-thickness autograft. To address this limitation, CSS were prepared containing human keratinocytes, fibroblasts, and dermal microvascular endothelial cells (HDMEC) isolated from a single skin sample. After 16 days in culture, control CSS and CSS containing HDMEC (CSS+EC) were grafted to full-thickness wounds in athymic mice. In CSS+EC in vitro, HDMEC persisted in the dermal substitutes and formed multicellular aggregates. One wk after grafting, HDMEC in CSS+EC organized into multicellular structures, some containing lumens. By 4 wk after grafting, HDMEC were found in linear and circular organizations resembling vascular analogs associated with basement membrane deposition. In some cases, colocalization of HDMEC with mouse perivascular cells was observed. The results demonstrate HDMEC transplantation in a clinically relevant cultured skin model, persistence of HDMEC after grafting, and HDMEC organization into vascular analogs in vitro and in vivo. All cells were derived from the same donor tissue, indicating the feasibility of preparing CSS containing autologous HDMEC for grafting to patients.

Figure 1

Histologies of CSS at culture day 15. Light microscopic images of plastic-embedded sections stained with toluidine blue are shown. A) Control CSS with well-organized epidermal layer and dermal layer densely packed with fibro-blasts. B) CSS + EC. Arrow indicates ring-like aggregate of cells near the dermal-epidermal junction. e, Epidermal layer; d, dermal layer. Scale bar in panel A (100 μm) is the same for both sections.

Figure 2

Localization of human endothelial cells in CSS+EC at culture day 15. Shown is immunohistochemical staining of frozen sections for the endothelial cell-specific antigen CD31. Sections were lightly counterstained with toluidine blue. A) Control CSS. B) CSS+EC. C) Normal human dermis shown as a positive control for specificity of staining. The arrows indicate regions of the CSS+EC staining positive for human CD31. e, Epidermal layer; d, dermal layer. Scale bar in panel A (100 μm) is the same for all panels.

Figure 3

Development of epidermal barrier in vitro is impaired in CSS+EC. Surface electrical capacitance (SEC, pF) of control CSS decreases with time in culture, approaching that of native human skin by 15 days. SEC values for CSS+EC also decrease with time in culture but do not approach values for either control CSS or human skin, indicating a delay or disruption of barrier formation. CSS+EC have significantly wetter surfaces than control CSS at all time points examined in vitro (P<0.001). Plotted are means ± se.

Figure 4

Histologies of CSS 4 wk after grafting to athymic mice. A) Control CSS. B) CSS+EC. C) Normal human skin. Light microscopic images of frozen sections stained with hematoxylin and eosin are shown. e, Epidermal layer; d, dermal layer. Scale bar in panel A (200 μm) is the same for all sections.

Figure 5

Localization of both human and mouse endothelial cells in CSS after grafting to athymic mice by double-immunofluorescent labeling. A–F) Sections of CSS+EC excised from mice at 1 (A, B), 2 (C, D), or 4 wk (E, F ) after grafting. G, H) Sections of control CSS excised at 4 wk after grafting. I, J) Normal human skin. A, C, E, G, I) Localization of human endothelial cell-specific CD31 (green fluorescence). B, D, F, H, J) Localization of mouse endothelial cell-specific CD31 (red fluorescence) in same sections. Arrows indicate locations of ring-like structures of aggregated human endothelial cells in CSS+EC sections. At 2 and 4 wk after grafting, some colocalization of staining for human and mouse CD31 can be seen (C–F ). Note absence of staining for human CD31 in control CSS (G) and absence of mouse CD31 staining in human skin (J), demonstrating species specificity of antibody staining. Dashed white lines indicate locations of dermalepidermal junctions. Scale bar in panel A (100 μm) is the same for all panels.

Figure 6

Deposition of basement membrane by endothelial cells in grafted CSS. A–F) Sections of CSS+EC excised from mice at 1 (A, B), 2 (C, D), or 4 wk (E, F), after grafting. G, H) Normal human skin. I, J) Control CSS at 2 wk after grafting. Sections were double-labeled with antibodies for either human (A, C, E, G) or mouse (I) CD31 (green fluorescence) and collagen type IV (B, D, F, H, J ) (red fluorescence). Arrows indicate regions of aggregated human endothelial cells in CSS+EC sections. Note increasing colocalization of collagen type IV staining and human CD31 staining in CSS+EC with increasing time after grafting. Essentially all staining for CD31 in normal human skin or in control CSS at 2 wk after grafting is colocalized with collagen type IV staining. Note also the localization of collagen type IV staining to the dermal-epidermal junction (dashed white lines in A, C, E, G, and I). Scale bar in panel A (100 μm) is the same for all panels.

Figure 7

Association of putative vascular analogs in CSS+EC with vascular smooth muscle cells at 4 wk after grafting. Smooth muscle cells were identified by immunostaining with an anti-α smooth muscle actin antibody. A, B) Section of CSS+EC double-labeled with antibodies against human CD31 (A) and α smooth muscle actin (B). C, D) Section of control CSS double-labeled with antibodies against mouse CD31 (C) and α smooth muscle actin (D). Colocalization (asterisk) of human endothelial cell CD31 staining and α smooth muscle actin staining is clearly seen in the larger human endothelial cell vascular analog (A, B). This α smooth muscle actin staining is less intense than that seen in the vessels derived from the mouse vasculature (C, D). Dashed white lines indicate locations of dermal-epidermal junctions. Scale bar in panel A (100 μm) is the same for all panels.