Ex vivo culture of keratinocytes on papillary and reticular dermal layers remodels skin explants differently: towards improved wound care

The interactions between dermal cells and the extracellular matrix (ECM) are vital in maintaining the integrity of skin and would healing [14]. Structurally, the ECM is composed of the fibrous proteins (i.e., collagens, elastin, and laminin) that give skin its robust tensile strength alongside elasticity, and the highly hydrated proteoglycans made of glycosaminoglycan (GAG) chains branching off linear protein cores [10, 14]. The variation in the ECM and cell-integrin expression across the regions of the skin distinguishes various environments that recruit and modulate the properties and activities of different cell types [14]. In this regard, the fibroblasts and the ECM which they produce has been shown to be characteristically different between the papillary and the reticular dermal layers [6].

In light of the diverse ECM profile, studies in wound healing and dermal substitute materials have focused at recapitulating normal wound healing and skin [3]. In this regard, the use of in vitro monolayer cultures of human skin cells to understand cellular mechanism involved in wound healing is limited by the absence of the multidimensional 3D environment physico-chemical cues [1]. While it is well accepted that in vivo studies in humans offer a conclusive testimonial of the cell-ECM crosstalk that is efficacious for clinical applications, such studies present several practical, ethical, and moral concerns. As such, in vivo studies in animal models have received increased attention as they allow for exhaustive analysis of the cellular and ECM interaction in living tissue. However, the clinical efficacy of using animal models, such as murine and porcine, which have different skin and healing profile to that of humans remains unclear [1]. As such, in vitro studies using organotypic and 3D human skin equivalents seeded with either keratinocytes or fibroblasts or both are widely used [5, 8] to bridge the gap between in vitro and in vivo methods. However, although these models provide a multidimensional environment around cells, the matrices are simplified with limited fidelity to human skin in vivo [5, 8].

On the other hand, the use of human full- and partial-thickness ex vivo culture allows the study of particular cell activities within an intact skin ECM. Indeed, human skin ex vivo culture systems have been employed in several dermatological and wound healing studies [5, 8]. In this study, to elucidate the effects of human epidermal keratinocytes on the ECM features in the dermis, skin explants were cultured (dermal-side up) and seeded with keratinocytes at the air–liquid interface, and histological and immunohistochemical staining for the cells and components of the ECM were performed.

Seeded keratinocytes infiltrated the reticular and papillary layers differently (Fig. 1). Since enzymatic removal of the epidermis from skin explants can potentially adulterate the ECM and cells, skin explants were cultured with the epidermis still attached (Fig. 1c). However, in line with findings by others [9, 13], the epidermis showed spongiosis-like features with marked detachment of the spinous layer in all explants after 7 days in culture (Fig. 1d–i). No differences across all samples were observed after 3 days in culture (data not shown). This was also observed in explants cultured epidermal-side up (data not shown). However, the dermal layers remained intact for 14 days with loosening of the exposed surface after 7 days (Fig. 1f, g) and poor reproducibility beyond 21 days partly due to coiling up of the explant, which enclosed the exposed surface (data not shown). However, human dermal explants cultured at a liquid–air interface in vitro have been shown by others to remain viable after 7 days [12] and in some cases 75 days [11]. While no keratinocytes were detected on explant surfaces after 7 days (Fig. 1j), increased keratinocyte infiltration in both papillary and reticular samples was eventually observed after 14 days, suggesting proliferation (Fig. 1h, i). However, while seeded keratinocytes appeared to form clusters in the papillary layer with shallower tissue infiltration (Fig. 1k), their spread was homogeneous and deeper in the reticular dermis with a characteristic intact single-cell-layer coverage of the surface (Fig. 1i, l). This varied keratinocyte distribution may be due to differences in the physical features of the two dermal layers, though others have demonstrated variation in ECM biochemistry, including that which affects keratinocyte proliferation and migration (i.e., hyaluronan and other proteins) as well as ECM remodelling, across dermal layers [11].

Indeed, after seeding of keratinocytes, the ECM features in the reticular and papillary dermal layers showed notable differences (Fig. 2). While the normal structure and distribution of ECM features such as elastin, collagen, reticular fibres, ground substance, and GAGs remained unchanged in control samples after 14 days in culture (Fig. 2a, b), newly deposited ground substance and or GAGs (light blue/green stain) were observed around seeded keratinocytes in both the papillary (Fig. 2c) and reticular samples (Fig. 2d). The localisation of the new ground substance and GAG positive stain, typically found in the papillary layer, as well as col VII, a basement membrane component (Fig. 2e, f), around seeded keratinocytes within the reticular dermal layer suggest the involvement of seeded keratinocytes in ECM remodelling. Indeed, the presence of keratinocytes has been shown to lead to the production of keratinocyte growth factors and GAGs (i.e., hyaluronan) by fibroblasts that promote keratinocyte migration [2, 4, 7], ECM production, and re-epithelialisation in wound healing through an interplay of complex autocrine and paracrine pathways. It has been shown by others that keratinocyte-derived IL-1α and β control fibroblast expression of keratinocyte growth factor (KGF/KGF7), which, in turn, enhances keratinocyte expression of tumour necrosis factor-α (TNF-α) resulting in collagen synthesis by fibroblasts as well as expression of MMPs and TIMPs, and synthesis of ECM components by both keratinocytes and fibroblasts [2, 7]. Moreover, in the immunostain for laminin (Fig. 2g–i) and intracellular gap junction connections Cx26 and Cx43 (Fig. 2j, k), more new laminin and Cx26-positive stains were detected in the papillary than in the reticular tissue sections around seeded keratinocytes. However, significantly more Cx43-positive stain (p < 0.05) was detected in the reticular tissue sections than in the papillary (Fig. 2 l). Cx26 and Cx43 are known to directly affect the proliferation and migration of keratinocytes and fibroblast [7, 15] and the observed differential expression of these connexins suggests differences in cellular activities.

Since there was no evidence of native keratinocyte migrating from epidermis to the exposed dermal surface in this work, the positive stain for CK14 along with newly deposited col VII and GAGs within and on the exposed dermal layers around keratinocytes suggests that the skin explants supported the production of robust layer-seeded keratinocytes. However, the differences in the nature of spread of keratinocytes, ground substance, and basement membrane-associated proteins (col VII and laminin) suggest differences in physico-chemical interaction between the seeded keratinocytes and features of the dermal layers. Furthermore, the difference in Cx26 and Cx43 expression between the dermal layers suggests possible difference in intercellular interaction and possible fibroblast involvement. Indeed, the role of these connexins in the regulation of fibroblast and keratinocyte proliferation, migration, and activity in both normal and diseased skin are well established [2].

Although the cellular and ECM differences in skin layers are well established, their clinical impact in wound interventions remains unclear. These findings in ex vivo skin explant cultures are in line with suggestions that the varying dermal layer features (ECM and potentially fibroblast) may play key roles in re-epithelialisation and should perhaps be considered in the design and development of biomaterials for wound applications as well as in wound healing interventions such as autologous keratinocyte grafting [15]. Indeed, the differential expression of these ECM features in the two dermal layers around newly seeded keratinocytes suggests potential imperative features and mechanisms in wound healing interventions. However, detailed and specific analysis of the cellular pathways, proliferation, and activities involved are necessary for conclusive clinical-efficacy characterisation.