Ultrastructural and immunohistochemical characterization of basal cells in three-dimensional culture models of the skin

Tissue assembly guided via substrate biophysics: applications to hepatocellular engineering

The biophysical nature of the cellular microenvironment, in combination with its biochemical properties, can critically modulate the outcome of three-dimensional (3-D) multicellular morphogenesis. This phenomenon is particularly relevant for the design of materials suitable for supporting hepatocellular cultures, where cellular morphology is known to be intimately linked to the functional output of the cells. This review summarizes recent work describing biophysical regulation of hepatocellular morphogenesis and function and focuses on the manner by which biochemical cues can concomitantly augment this responsiveness. In particular, two distinct design parameters of the substrate biophysics are examined--microtopography and mechanical compliance. Substrate microtopography, introduced in the form of increasing pore size on collagen sponges and poly(glycolic acid) (PGLA) foams, was demonstrated to restrict the evolution of cellular morphogenesis to two dimensions (subcellular and cellular void sizes) or induce 3-D cellular assembly (supercellular void size). These patterns of morphogenesis were additionally governed by the biochemical nature of the substrate and were highly correlated to resultant levels of cell function. Substrate mechanical compliance, introduced via increased chemical crosslinking of the basement membrane, Matrigel, and polyacrylamide gel substrates, also was shown to be able to induce active two-dimensional (2-D, rigid substrates) or 3-D (malleable substrates) cellular reorganization. The extent of morphogenesis and the ensuing levels of cell function were highly dependent on the biochemical nature of the cellular microenvironment, including the presence of increasing extracellular matrix (ECM) ligand and growth-factor concentrations. Collectively, these studies highlight not only the ability of substrate biophysics to control hepatocellular morphogenesis but also the ability of biochemical cues to further enhance these effects. In particular, results of these studies reveal novel means by which hepatocellular morphogenesis and assembly can be rationally manipulated leading to the strategic control of the expression of liver-specific functions for hepatic tissue-engineering applications.

Organotypic culture and surface plantation using umbilical cord epithelial cells: morphogenesis and expression of differentiation markers mimicking cutaneous epidermis

BACKGROUND

The umbilical cord epithelium (UCE) is composed of a single epithelial layer covering mucous connective tissue and it is thought to derive from the amniotic epithelium. Interestingly, UCE cells express not only simple and mucous epithelial keratins (CK8 and CK4/CK13), but also stratified epithelial keratins (CK1/10) and cornified cell envelope (CCE)-associated proteins.

OBJECTIVE

To understand the nature of UCE, UCE cells were cultured under the same conditions of organotypic culture of epidermal keratinocytes and grafted onto the back of nude mice.

METHODS

UCE cells isolated from fresh umbilical cord specimens were cultured using serum-free keratinocyte growth medium, and plated on a fibroblast-populated collagen matrix using air-liquid interface methods. UCE cells were transplanted onto the back of Balb C nu/nu mice as a thin epithelial sheet grown on a collagen matrix.

RESULTS

UCE cells formed a multi-layered stratified epithelium both in organotypic culture and surface transplantation. Regarding the expression profile of differentiation-specific proteins, such as keratins, the CCE-precursor proteins and junctional proteins, the reconstructed epithelium showed a close similarity to natural epidermis in organotypic culture.

CONCLUSION

These results suggest the possibility that UCE cells can differentiate and organize into an epidermis-like structure, when exposed to the appropriate conditions which is similar to those of cutaneous epidermis.

The effects of epidermal keratinocytes and dermal fibroblasts on the formation of cutaneous basement membrane in three-dimensional culture systems

The cutaneous basement membrane (BM) plays an important role in normal and pathological conditions. However, few studies have addressed the formation of the cutaneous BM using three-dimensional culture systems. In this study, to elucidate the effects of human epidermal keratinocytes and dermal fibroblasts on the formation of the cutaneous BM, keratinocytes were cultured on several dermal substrates in the presence or absence of fibroblasts at the air-liquid interface. After 2 weeks of culture, immunohistochemical stainings for the components of the BM and electron microscopic studies of the BM zone (BMZ) were performed. In cultures of keratinocytes alone on dead reticular dermis or collagen gel without fibroblasts, beta4 integrin chain, laminin, type IV and VII collagens were all expressed. However, ultrastructurally, BMZ was not formed. In cultures of keratinocytes on fibroblast-populated collagen matrix, laminin, and type IV and VII collagens were expressed more strongly than in the absence of fibroblasts. In addition, elements of the BMZ such as hemidesmosomes, lamina lucida, lamina densa and anchoring fibrils were formed, although it was still incomplete. In the culture of keratinocytes alone on de-epidermized dermis (DED) (surface up), beta4 integrin chain, laminin, and type IV and VII collagens were strongly expressed. Also, the BMZ appeared similar to that in normal skin. In cocultures of keratinocytes and fibroblasts on DED or cultures of keratinocytes on DED combined with fibroblast-populated collagen matrix, type IV collagen was expressed more strongly than in cultures of keratinocytes alone. Ultrastructurally, similar findings to those of cultures of keratinocytes alone on DED were seen. Interestingly, when keratinocytes and fibroblasts were cocultured on DED, some fibroblasts were seen in the upper dermis as a result of migration into the dermis through partial loss of the lamina densa. These results show that keratinocytes produce most of the components of the BM such as laminin, and type IV and VII collagens. In addition, fibroblasts stimulate the expression of the components of the BM and the formation of a BMZ, suggesting that fibroblasts may produce laminin, and type IV and VII collagens or influence the effects of keratinocytes on the formation of the BM through a keratinocyte-fibroblast interaction.

Fibroblasts and ascorbate regulate epidermalization in reconstructed human epidermis

Skin equivalent model provides a new investigating system to study the role of extracellular matrix and dermal factors such as collagen, basement membrane components and fibroblasts (Fb) which contribute to cell-cell and cell-matrix interactions. Although basement membrane factors is known to play an important role in epidermal differentiation and epidermal-matrix adhesion, comparative effects of these extracellular matrix and dermal factors on the reconstruction of epidermis are little known. In this study, we investigated effects of type I collagen (Coll I), type IV collagen plus laminin (LAM) coated Coll I (Coll IV+LAM), and human Fb enriched Coll I (Coll I+Fb) on epidermal reconstruction. When human keratinocytes were cultured on three different gels containing Coll I, Coll IV+LAM and Coll I+Fb, basal keratinocytes were cuboidal and perpendicular to the dermo-epidermal junction only in the gel containing Coll I+Fb. Proliferation marker expression was prominent and differentiation marker expression was similar with those of normal skin in the gel containing Coll I+Fb than in the other gel models. Since ascorbate is suspected to exert an effect as a modulator of proliferation and differentiation in keratinocytes, we tested the effects of ascorbate on human epidermis reconstruction. When 25 microg/ml ascorbate was added, disordered arrangement of epidermis was disappeared and differentiation marker expression was similar with its expression in normal skin. These data indicate that human Fb and a modulator of proliferation and differentiation such as ascorbate are essential for epidermalization in reconstructed epidermis.

Effects of collagen IV and laminin on the reconstruction of human oral mucosa

To investigate the effects of basement membrane proteins on the reconstruction of mucosa equivalent, oral mucosa substitute were cultured on (1) type I collagen gels, (2) type IV collagen-coated type I collagen gels, (3) laminin-coated type I collagen gels, and (4) type I collagen gels containing both type IV collagen and laminin. H/E and PAS staining showed that the characteristics of the oral mucosa were preserved under all the experimental conditions. However, the basal keratinocytes appeared cuboidal when the type I collagen gels were coated with type IV collagen plus laminin. The expression of the differentiation markers was similar, but weak staining of filaggrin, K13, and involucrin was observed with the type IV collagen plus laminin coating. Furthermore, electron microscopy revealed that the size of the basal keratinocytes was relatively small and uniform when both type IV collagen and laminin were used. These findings suggested that these two major basement membrane proteins are important in the process of differentiation in mucosal keratinocytes.

Control of hepatocyte function on collagen foams: sizing matrix pores toward selective induction of 2-D and 3-D cellular morphogenesis

While microporous biopolymer matrices are being widely tested as cell culture substrates in hepatic tissue engineering, the microstructural basis for their control of cell differentiation is not well understood. In this paper, we studied the role of void size of collagen foams in directing the induction of liver-specific differentiated morphology and secretory activities of cultured rat hepatocytes. Hepatocytes cultured on collagen foams with subcellular sized pore diameters of 10 microm assumed a compact, cuboidal cell morphology, rapidly achieving monolayer coverage, and secreted albumin at the rate of 40 +/- 8 pg/cell/d. Increasing the pore size to 18 microm elicited a distinctly spread cellular phenotype, with discontinuous surface coverage, and albumin secretion rates declined precipitiously to 3.6 +/- 0.8 pg/cell/d. However, when collagen foams with an even higher average void size of 82 microm were used, hepatocytes exhibited high degree of spreading within an extensive three-dimensional cellular network, and exhibited high albumin secretory activity (26 +/- 0.6 pg/cell/d). The effect of void geometry on cellular ultrastructral polarity was further analyzed for the three void size configurations employed. The distribution of the cell-cell adhesion protein, E-cadherin, was primarily restricted to cell-cell contacts on the 10 microm foams, but was found to be depolarized to all membrane regions in cells cultured on the 18 and 82 microm foams. Vinculin-enriched focal adhesions were found to be peripherally clustered on cells cultured on 10 microm foams, but were found to redistribute to the entire ventral surface of cells cultured on the 18 and 82 microm foams. Overall, we demonstrate the significance of designing pore sizes of highly adhesive substrates like collagen toward selective cell morphogenesis in 2-D or 3-D. Subcellular and supercellular ranges of pore size promote hepatocellular differentiation by limiting 2-D cell spreading or effecting 3-D intercellular contacts, while intermediate range of pore sizes repress differentiation by promoting 2-D cell spreading.

The requirement for basement membrane antigens in the production of human epidermal/dermal composites in vitro

The importance of a dermal element when providing permanent wound cover for skin loss has become evident as the shortcomings of pure epidermal grafts are recognized. We are developing a skin composite formed from sterilized human de-epidermized acellular dermis, keratinocytes and fibroblasts with the ultimate aim of using this composite to cover full-thickness excised burn wounds. These composites can be prepared with or without basement membrane (BM) antigens initially present on the dermis. This study investigates the importance of retaining BM antigens on the dermis to the production and appearance of these composites in vitro. Skin composites prepared from dermis with BM antigens either present or absent initially were studied throughout 3 weeks. Composites with BM antigens present initially were significantly better than those initially lacking BM antigens in: (i) the degree of epithelial cell attachment to the underlying dermis (hemidesmosomes were seen only in the former); (ii) the morphology of the epithelial layer; (iii) the consistent presence of collagen IV and laminin and the increasing expression of tenascin; and (iv) the amount of soluble collagen IV and fibronectin detected in the conditioned media. We conclude that an initial BM antigen template is vital in this skin composite model for the attachment and differentiation of the epithelial layer and for the subsequent remodelling of the BM in vitro.

Sensitivity of cross-reacting antihuman antibodies in formalin-fixed porcine skin: including antibodies to proliferation antigens and cytokeratins with specificity in the skin

Although no animal is a perfect skin model for the study of toxicological and therapeutic agents, structurally the pig may be superior to even non-human primates. Because our work involves effects of toxicological and therapeutic agents on the skin, we wanted to identify stains which may prove useful as well as determine cross-reactivity of some newer antihuman antibodies. We performed a battery of formalin-fixed skin from weanling pigs and minipigs. The battery of antibodies included LCA, CD3, OPD-4, CD34, UCHL-1, L-26, KP-1, MAC-387, Factor XIIIa, Leu-7, S-100 protein, HMB-45, GFAP, synaptophysin, neurofilament protein, ubiquitin, vimentin, type IV collagen, laminin, fibronectin, Factor VIII related antigen, Desmin-M, smooth muscle actin, cytokeratin 7, cytokeratin 20, AEI/AE3, CAM 5.2, EMA, GCDFP, Ki-67, and PCNA. Immunohistochemical stains for CD3, Leu-7, S-100 protein, type IV collagen, laminin, Factor VIII related antigen, GFAP, synaptophysin, neurofilament protein, ubiquitin, smooth muscle actin, vimentin, Desmin-M, cytokeratin 7, cytokeratin 20, AE1/AE3, CAM 5.2, Ki-67 and PCNA showed consistent cross-reactivity. In formalin-fixed tissue, only antibodies to lymphoreticular cells showed poor cross-reactivity. A high percentage of the remaining antibodies did show good cross-reactivity but with some interesting similarities and differences in specificity.

Hemidesmosomes show abnormal association with the keratin filament network in junctional forms of epidermolysis bullosa

Junctional epidermolysis bullosa is a group of hereditary bullous disorders resulting from defects in several hemidesmosome-anchoring filament components. Because hemidesmosomes are involved not only in keratinocyte-extracellular matrix adherence, but also in normal anchorage of keratin intermediate filaments to the basal keratinocyte membrane, we questioned whether this intracellular function of hemidesmosomes was also perturbed in junctional epidermolysis bullosa. We used quantitative electron microscopic methods to assess certain morphologic features of hemidesmosome-keratin intermediate filaments interactions in skin from normal subjects (n = 11) and from patients with different forms of junctional epidermolysis bullosa (n = 13). In addition, skin from patients with autosomal recessive epidermolysis bullosa simplex with plectin defects (n = 3) or with autosomal recessive dystrophic epidermolysis bullosa (n = 4) were included as controls. Values were expressed as a percentage of the total number of hemidesmosomes counted. In normal skin 83.3% +/- 3.3 (SEM) hemidesmosomes were associated with keratin intermediate filaments and 90.1% +/- 1.9 had inner plaques. In Herlitz junctional epidermolysis bullosa (laminin 5 abnormalities, n = 4) these values were reduced to 45.3% +/- 11.5 (p < 0.001; analysis of variance) and 50.3% +/- 12.8 (p < 0.001), respectively. In junctional epidermolysis bullosa with pyloric atresia (alpha6beta4 abnormalities, n = 3) the values were also reduced [41.8% +/- 7.0 (p < 0.001) and 44.5% +/- 5.7 (p < 0.001), respectively]. In the non-Herlitz group (laminin 5 mutations, n = 3) the counts were 66.7% +/- 7.1 (p > 0.05) and 70.5% +/- 8.5 (p < 0.05), and in skin from patients with bullous pemphigoid antigen 2 mutations (n = 3) the counts were 54.3% +/- 13.8 (p < 0.01) and 57.1% +/- 13.9 (p < 0.01). In epidermolysis bullosa simplex associated with plectin mutations the values were 31.9% +/- 8.9 (p < 0.001) for keratin intermediate filaments association and 39.9% +/- 7.1 (p < 0.001) for inner plaques. Findings in recessive dystrophic epidermolysis bullosa patients' skin were indistinguishable from normal control skin with inner plaques (90.5% +/- 2.5) and keratin intermediate filaments attachment (86.3% +/- 2.1). These findings suggest that the molecular abnormalities underlying different forms of junctional epidermolysis bullosa appear to affect certain critical intracellular functions of hemidesmosomes, such as the normal connections with keratin intermediate filaments. This may have important implications for the maintenance of basal keratinocyte integrity and resilience in junctional epidermolysis bullosa.

Autologous transplantation of urothelium into demucosalized gastrointestinal segments: evidence for epithelialization and differentiation of in vitro expanded and transplanted urothelial cells

PURPOSE

Our study established a technique for in vitro expansion and subsequent transplantation of autologous urothelial cells into vascularized seromuscular segments from stomach and colon in sheep. The proof of proliferation and differentiation of the transplanted urothelium in the absence of resident urothelium is considered to be a prerequisite for use of this technique in bladder augmentation.

MATERIALS AND METHODS

Autologous sheep urothelial cells were expanded in vitro and grown on collagen membranes for sheet grafting. Using a vital stain, viability and confluency status of the urothelial graft were determined before transplantation into demucosalized segments isolated from the sheep stomach and colon gastrointestinal pouches. The gastrointestinal segments were sewn up and remained in the abdomen as small pouches stiched to the abdominal wall. Take and differentiation of transplanted cells within the pouch were assessed two and three weeks later using histological and immunohistological means.

RESULTS

Urothelial cells grew well on collagen membranes. A confluency status > 40% and co-culturing with 3T3 feeder cells favored successful transplantation. Two weeks after transplantation a multilayered urothelial-like epithelium was found to line the lumen of the pouch. The epithelium was characterized by a distinct urothelium-typical distribution of basal and luminal keratins and the expression of the umbrella cell-specific marker uroplakin III. Moreover, the epithelium had an underlying basal lamina which focally contained collagen type IV.

CONCLUSIONS

The data indicate that in vitro expanded urothelial cells are capable of epithelializing demucosalized gastrointestinal segments forming a genuine, differentiated "neo" urothelium.

Three-dimensional cultures of keratinocytes and an application to in vitro-amyloid production of cutaneous amyloidosis

Some three-dimensional culture models of the skin were reviewed including our systems using a collagen dermal substitute and a matrix dermal substitute. No obvious junctional structures, such as hemidesmosomes and the lamina densa, were formed between the basal keratinocytes and the dermal substitutes, when the cytoplasmic membrane of the basal keratinocytes directly faced the collagenous materials. On the other hand, when the cytoplasmic membrane of the basal keratinocytes faced the preformed basement membrane, the type IV collagen film, or the extracellular matrix gel, an organized interaction occurred between the basal keratinocytes and the dermal substitute through hemidesmosomes and a rudimentary lamina densa. Keratinocyte differentiation in the suprabasal layers seemed to be closely related to such a basal cell differentiation. Our preliminary examination of the experimental amyloid production by the epidermal sheet from the lesional skin of patients with primary localized cutaneous amyloidosis suggested that the basal cells in the transplanted lesional epidermis produced amyloid fibrils in our in vitro culture model. This is another use of the three-dimensional culture models of the skin in addition to the application of the systems to wound treatment.

Differences in dermal analogs influence subsequent pigmentation, epidermal differentiation, basement membrane, and rete ridge formation of transplanted composite skin grafts

This study evaluated the in vitro and in vivo function of composite skin equivalents based on two different dermal analogs. Keratinocytes derived from the same dark-skinned neonatal foreskins were seeded onto both acellular human dermis and fibroblast-contracted collagen gels. Each type of composite graft readily formed an epithelium in vitro. However, the undulating surface of the acellular dermis acted as a template and organized the seeded keratinocytes into a rete ridge-like pattern, whereas the smooth surface of the fibroblast-contracted collagen gels generated an epithelium with a linear basal layer. Moreover, when acellular dermis was used, the composite grafts demonstrated enhanced melanocyte proliferation. When transplanted to athymic mice, both composite grafts formed a fully differentiated human epidermis, but repigmentation of the grafts when acellular dermis was used was more extensive and only the epidermis on the fibroblast-contracted collagen gels showed signs of hyperproliferation at 6 weeks after grafting. These results demonstrate that the type of dermal analog incorporated into a composite skin graft can influence the subsequent functionality of the skin substitute.