DISCUSSION

In this study, it was reported for the first time that endothelial cells, when they are stimulated with UV irradiation, are responsible for stimulatory action upon skin pigmentation. UV-irradiated endothelial cells were clearly involved in the regulation of melanogenesis in melanocytes, and there was increased epidermal pigmentation of human skin in the presence of UV-irradiated endothelial cells. Our results indicate that endothelial cell-derived soluble factors mediate the effects of endothelial cells on melanocytes.

According to the results of this study, ET-1 at least is assumed to play a role. ET-1, which has been considered to be secreted from keratinocytes, is a well-known melanogenic factor which induces pigmentation and the tanning response upon UV irradiation (Imokawa et al., 1995; Rees, 2004; Denat et al., 2014). Previous studies reported that HUVECs and endothelial cells of other origins express ET-1 (Yanagisawa et al., 1988; Delerive et al., 1999). We confirmed these findings by showing that ET-1 is expressed by endothelial cells and that UV irradiated cells in particular produce more ET-1 than do non-stimulated cells.

These results are consistent with previous work which showed that stimulated endothelial cells could produce more ET-1 than non-stimulated cells (Delerive et al., 1999). Inhibition studies using an ET-1 antibody confirm the crucial role of soluble ET-1 in mediating endothelial cell-dependent melanocyte differentiation (Rouzaud et al., 2005). On previous studies showing increased vasculature in UV-induced hyperpigmentary skin of melasma and solar lentigo (Kim et al., 2007; Kang et al., 2010; Kang et al., 2011a), these promoting effects of UV-irradiated endothelial cells on melanogenesis are consistent with in vivo

findings. It is speculated that the interaction between melanocytes and endothelial cells during sun exposure plays a role in the activation of the melanocyte function and in the consequent cutaneous pigmentation process, resulting in the development of hyperpigmentary disorders.

This report also provides new evidence that non-UV-irradiated endothelial cells instead play an inhibitory role during skin pigmentation. In our co-culture system and in the ex vivo skin culture experiments conducted here, the presence of endothelial cells had a negative effect on melanogenesis and inhibited tyrosinase expression through MITF down-regulation.

In fact, these findings were unexpected based on our initial hypothesis in this study; however, they are consistent with epidemiological data showing that light-skinned individuals show much more prominent telangiectasia than dark-skinned persons (Berg, 1989; Halder et al., 2003; Chosidow and Cribier, 2011). Very recently, it was also reported that low levels of melanin are correlated with the development of angiogenic ocular and skin diseases (Adini et al., 2014). Indeed, human fibroblasts, unlike UV-irradiated fibroblasts, can also have an inhibitory influence on human epidermal pigmentation, although they secrete melanogenic factors such as HGF and SCF (Hedley et al., 2002; Lee et al., 2003). These findings were clearly observed in in vitro full-thickness pigmented skin models, which allow physiological interactions between the melanocyte-containing epidermis and the dermal compartment with living fibroblasts (Cario-Andre et al., 2006; Duval et al., 2014). Recently, it was shown that fibroblast-derived DKK1 is responsible for the hypopigmentation of palmo-plantar skin (Yamaguchi et al., 2009). How endothelial cells down-regulate pigmentation remains unclear, though the process may be evoked by yet unknown endothelial factors and/or an issue not

addressed in the present study.

In the preliminary phase of our study, the appropriate medium for the co-culture system using HUVECs and NHMs was established. There were dramatic influences of the batch of numerous supplements in the cell culture medium on each cell behavior. The tested media included those routinely used to culture the two cell types in a monoculture (MCDB153 and F12 for NHMs and HGM for HUVECs, respectively) as well as various mixtures (1:1 and 1:4) of F12 and HGM. Here, the medium eventually selected was the MCDB153 medium.

Because the cell viability of HUVECs was significantly decreased after 4 days of incubation in MCDB153 medium, we replaced the HUVECs with fresh ones every 2 days up to 5 days, which is mandatory to observe their effect on melanin production in NHMs (Tomita et al., 1987; Carsberg et al., 1994). The cell performance in this culture condition was closer to that obtained with HGM for HUVECs. Interesting observations included TPA, a crucial supplement for melanocyte growth, which showed a remarkable effect on HUVECs. It changed the cell morphology, creating elongated cells with an irregular cell boundary. The HGM medium itself had stimulatory effects on melanogenesis in NHMs according to our preliminary data.

Several studies have reported increased expressions of several paracrine factors, including α-MSH, ET-1, KGF, bFGF, and SCF in UV-exposed skin and hyperpigmented skin, showing the effects of these cytokines on melanocyte differentiation (Costin and Hearing, 2007; Abdel-Malek et al., 2010). However, none of these paracrine factors except 1 were significantly changed from gene expression profiling in our study. However, ET-1 may not be the only factor involved in melanocyte-endothelial cell interaction. In our study,

among other up-regulated genes identified in UV-irradiated bEnd.3 cells, tissue-type plasminogen activator (tPA) was another soluble factor (showing a 2.66-fold increase, log2 ratio: 1.41, p=0.0025). It was previously reported that UV irradiation induces tPA synthesis and plasmin activity in cultured keratinocytes (Takashima et al., 1992) and that the tPA generated by keratinocytes increases the activity of melanocytes in vitro (Maeda K, 2007).

Up-regulation of bone morphogenetic protein-2 (BMP-2) was also noted in UV-irradiated endothelial cells, although the levels were not significant (p=0.06). BMP-2 is produced by HUVECs and fibroblasts (Csiszar et al., 2005); this was shown to have a stimulatory effect on tyrosinase activity and melanocyte proliferation (Bilodeau et al., 2001). Further studies to elucidate the roles of BMP-2, tPA, and/or yet unknown factors in the interaction between endothelial cells and melanocytes are necessary.

One of the limitations of this study was that we should reproduce the consistent results with HDMECs on the melanogenesis of NHMs. Also, the additional gene expression profiling with human endothelial cells should be performed in future to find other candidate genes besides ET-1. In terms of UV irradiation, UVA irradiation should be performed additionally to evaluate the effects of different wavelength on melanogenesis.

In summary, this study provides valuable insights into the role of endothelial cells in the regulation of skin pigmentation. Our study demonstrates that endothelial cells have a negative effect on pigmentation. However, their effect differs after they are irradiated with UV, as they then positively influence the induction of melanogenesis. The UV-irradiated endothelial cells may produce secreted factors, in this case ET-1 mediates the stimulatory effect on pigmentation which may play a role in the development of hyperpigmentary

disorders. Future studies are required to identify many other regulators of melanogenesis from endothelial cells based on our gene expression profiling.