¹Department of Dermatology, Stanford University, Stanford, California, USA

Correspondence: Dr Marvin Karasek, Department of Dermatology, Stanford University, 269 Campus Drive, CCSR-2, Room 2155, Stanford, California 94305-5168, USA. E-mail: marvek@stanford.edu

Abstract

Specific cell markers for selective isolation and growth of cells from the lymphatic and vascular systems have increased our understanding of the structure–function relationships of both systems. Using these markers, a subset of blood vessel endothelial cells (BECs) with the properties of lymphatic endothelial cells (LECs) has been identified in psoriasis and eczema. The differentiation potential of BECs and LECs in vivo and in vitro is reviewed.

Of increasing interest in the pathology of inflammatory and fibrotic skin disease is a role for transformation of both blood vessel endothelial cells (BECs) and lymphatic endothelial cells (LECs) into new cell types. A transition during which epithelioid cells lose their polarity and cohesiveness and transform into spindle-shaped cells that are more fibroblast- or myofibroblast-like is referred to as epithelial-to-myofibroblast transition (EMT). This commentary reviews evidence for EMT in postembryonic BECs and LECs.

In BEC culture, dermal microvascular endothelial cells isolated from either neonatal or adult skin maintain a typical epithelial cobblestone appearance and closely duplicate the morphology and ultrastructure of microvascular cells in intact blood vessels. In the presence of factors that maintain intracellular levels of cyclic AMP, BECs continue to express BEC-specific markers over extended periods in culture. When exposed transiently to inflammatory cytokines (interleukin-1 (IL-1), tumor necrosis factor- (TNF-)) or to phorbol esters, BECs lose tight junctions, increase permeability, and convert reversibly to a more spindle-shaped morphology similar to that observed in an inflamed, activated endothelium. In contrast to short-term exposure to inflammatory cytokines, long-term stimulation with the same inflammatory mediators permanently transforms BECs to spindle-shaped cells. In permanently transformed BECs, markers that identify BECs (factor VIII, type IV, collagen, CD 31) are lost (Romero et al., 1997) and matrix proteins normally synthesized by myofibroblasts (-smooth muscle actin, type I collagen) are expressed (Chaudhuri et al., 2007). A balance between the activation of protein kinase A and the inhibition of protein kinase C has been shown to play a key role in the regulation of EMT in vitro (Bokhari et al., 2005).

Earlier studies by Gröger et al. (2004) demonstrated that exposure of BECs to IL-3 induces a new set of transformed BECs that is different from that following exposure to IL-1. BECs acquire the LEC-specific markers prox-1 and podoplanin and continue to express CD 31. In these earlier studies a constant exposure to IL-3 was required to maintain a LEC-like phenotype. In vivo IL-3 stimulates expression of the major histocompatibility complex class II and cytokine production and may play a role in lymphocyte trafficking in inflammatory skin disease.

In this issue Gröger et al. (2007) extend their earlier studies of BECs in vitro and describe a new subset of BECs, both in psoriasis and in eczema, that expresses markers found in LECs. They ask whether this subset of BECs was the type of differentiation of BECs that they had described previously in vitro (Gröger et al., 2004) or the consequence of an unusual cytokine activation of BECs. A similar subset of BECs was not observed in urticaria in which inflammation is not prominent. To address this question, the investigators compared prodoplanin and LYVE-1 expression both in cytokine-simulated LEC and BEC cell cultures and in organ cultures of normal adult skin. Organ cultures of skin more closely simulate conditions approximating those observed in vivo and minimize changes in cell culture that may not completely represent those in vivo. As shown by Gröger et al. (2007) in this issue, a similar subset of BEC-expressing LEC markers was observed in BECs in organ culture. Of interest was the observation that in BEC cell culture, IL-3 and TNF- induced prodoplanin and LYVE-1 in all BECs, not only in a subset of cells. These differences, although not major, are most likely another example of the important influences of tissue environment, culture medium, and passage number on gene expression—factors that must be evaluated when using cell culture as an in vitro model of the vascular system.

In involved psoriatic skin, multiple studies have shown that the number of CD8⁺ T cells is increased in the papillary dermis. This population expresses proinflammatory cytokines, and in the presence of these mediators a decrease in tight gap junctions would be expected. In agreement with in vitro observations, a loss in tight gap junctions has been described at the ultrastructural level in the psoriatic vasculature. As shown in confluent BEC monolayers (Bokhari et al., 2005), a marked increase in permeability occurs following exposure to inflammatory cytokines and further demonstrates a close relationship among inflammation, gap junctions, and EMT. It would also be of interest to determine whether uninvolved skin of patients with psoriasis, particularly from those with an active Koebner response, would show an increased expression of LEC markers in the same set of BECs observed in vivo following exposure to IL-3 and TNF-. Methods to isolate endothelial cells from small biopsies of normal and pathologic skin samples have been described and would be suitable for both functional and molecular profile analysis.

In another inflammatory disorder (scleroderma), there is a strong correlation among inflammation, the disappearance of blood vessels, and the appearance of myofibroblasts. A mechanism to explain these changes is not known at present, but, based on the evidence for the transformation of BECs into myofibroblasts by inflammatory cytokines in vitro (Chaudhuri et al., 2007) and by drugs that inhibit these changes (Karasek, 2007), the transformation of microvascular endothelial cells into myofibroblasts in vivo may be one way to interpret the pathologic changes observed in vivo. EMT in BECs isolated from patients with scleroderma would be of interest in light of the strong potential for endothelial-to-myofibroblast transformation induced by inflammatory cytokines.

In Kaposi's sarcoma (KS) there is additional evidence for a major role for EMT in the genesis of spindle cells, both in vivo and in vitro. In vivo human herpesvirus 8 (HHV8) has been detected in all forms of KS, as well as in atypical endothelial cells, prior to spindle-cell development (Kennedy et al., 1998). In vitro HHV8 infection of normal BECs results in the rapid conversion of normal BECs to spindle-shaped cells that are morphologically similar to the spindle-shaped cells induced by cytokines (Moses et al., 1999).

Finally, we might ask whether podoplanin, a transmembrane glycoprotein that co-localizes with members of the ERM (ezrin, radixin, moesin) protein family, plays a more important role in BEC and LEC physiology and pathology than recognized previously. Podoplanin is upregulated in several types of experimental and human tumors, including squamous-cell carcinoma of the skin (Schacht et al., 2005). In other epithelial cell types (e.g., canine kidney epithelial cells), podoplanin induces EMT via activation of Rho A (Martin-Villar et al., 2006). Because tight gap junctions are necessary to maintain BEC homeostasis, upregulation of podoplanin expression by inflammatory cytokines in BEC and binding to ERM may activate kinases central to regulation of EMT in BECs and LECs. Additional studies of the effect of podoplanin on BEC and LEC structure and function would be of interest.