A blank canvas might be a simple piece of white linen to most. But that same piece of fabric evokes myriad patterns of hue and color from an artist. Could skin have the same effect on melanocytes, the pigment-producing cells that daub the human body with melanin? New findings from HMS researchers suggest that the human skin epithelium is more than a passive canvas. Janice Brissette, associate professor of dermatology at Massachusetts General Hospital, and colleagues have discovered that it is the epithelial cells, the pigment-receiving cells in the skin, that instruct the melanocytes where to apply their paint. The results, reported in the Sept. 7 Cell, suggest that the human skin is one giant template for color, a perspective that completely revises the understanding of skin pigmentation.
Skin and hair color are the result of an intricate relationship between cutaneous epithelial cells and pigment-producing melanocytes. The latter synthesize melanin in specialized organelles, the melanosomes, and then ship these pigment factories along an elaborate dendritic network to either the epidermal keratinocytes or the epithelial cells that form the hair shaft. Together, the melanocytes, their dendritic networks, and the multiple cells they contact represent a single pigmentary unit.
Scientists have puzzled over how this unit is built and maintained. “For years people have focused on melanocytes as the proactive partners and viewed the epithelial cells as simply reactive, but our work indicates that it is the epithelial cells that tell the melanocytes what to do,” said Brissette. She compares the human skin to a child’s coloring book. “The pigment recipient cells—a subset of the skin’s epithelial cells—form a drawing that outlines where the pigment should be placed. Then the pigment donor cells—the melanocytes—melanize the recipients and color in that picture.”
Brissette and colleagues made this discovery when studying the morphogenic role of the transcription factor Foxn1. Loss of this protein causes the well-known nude mouse phenotype. In the nude mouse, hairs become brittle and fall off, exposing the pink skin underneath; mice, unlike humans, have no melanin in the epidermis beneath their coat—all the pigmentation goes to their coat hairs.
Brissette, together with joint first authors on the paper, Lorin Weiner, an instructor in dermatology, and research fellow Rong Han, both at MGH, noticed that when they forced expression of Foxn1 in mouse keratinocytes, the animals turned black at sites where there is typically no hair, such as the paws, footpads, and snout. When they examined the animals’ skin, they found that in both hairy and hairless regions, the epidermis was extensively melanized. Given that Foxn1 expression in these transgenic animals is driven by the keratin 5 promoter, which is only active in epithelial cells and not melanocytes, the simplest explanation for the skin color change was that the keratinocytes can actively recruit the pigment donors.
To test that idea, the researchers examined the skin microscopically. While the epidermis in normal mice loses pigment donor cells rapidly after birth, the researchers found that melanocytes were retained in the epidermis of the Foxn1 transgenic animals. Furthermore, melanocyte recruitment seemed heavily dependent on Foxn1. As the transgenic animals aged, the number of Foxn1-expressing cells gradually declined along with the number of melanocytes in the epidermis. When the researchers stained skin sections for the melanocyte marker Tyrp1, they found it closely colocalized with Foxn1-expressing cells (see figure).
Interestingly, the gain in epidermal pigmentation in these transgenic animals did not come at the expense of coat color, indicating that the additional Foxn1 had no effect on melanocyte migration into hair follicles. Instead, the transgene broadened the brush strokes of the melanocytes so they formed pigmentary units with keratinocytes as well. In effect, Foxn1 made the mouse skin more humanlike. In fact, when the researchers examined human epidermis, they found that half the pigment recipients expressed the transcription factor—human epidermis has two major classes of pigmented cells. The mouse and human data suggest that Foxn1 is a major regulator of pigmentation in mammals.
How does expression of Foxn1 in pigment recipient cells lead to recruitment of pigment donors? Since it is unlikely that the intracellular transcription factor is released by pigment recipients, it must trigger some other signals that communicate with the melanocytes. Unfortunately, while many downstream genes are known to be regulated by Foxn1, its direct targets have never been identified. But the researchers wondered if fibroblast growth factor-2 (Fgf-2) may be involved since it is known to have potent actions on melanocytes, including a proliferative effect, though it is not normally produced by these cells.
To probe the Foxn1–Fgf-2 relationship, Weiner, Han, and research fellow Jian Li examined the effect of Foxn1 on primary keratinocytes. When the researchers infected the cells with a Foxn1-expressing adenovirus, they found that Fgf-2 was induced within eight hours, an indication that the cytokine may be a primary target. This idea was supported by chromatin immunoprecipitation experiments, which showed that Foxn1 can bind directly to the Fgf-2 gene at both a conserved intron and just upstream of the core promoter.
But even though Foxn1 might turn on the Fgf-2 gene in keratinocytes, is the cytokine the trigger that stimulates melanocytes? To address this, the researchers blocked Fgf-2 activity in the Foxn1 transgenic mice and found that inhibiting Fgf-2 neutralized the effect of the Foxn1 transgene.
Though this is the first evidence that epithelial cells have some control over their own pigmentation, the finding makes a lot of sense, suggested Brissette. In both human and mouse embryos, the epidermis is colonized with melanoblasts, but as mice mature, these become localized in the hair follicles. “So with just that simple observation, you can see that two different epidermal environments elicit different melanoblast responses,” she said.
Epithelial control over melanization also explains some of the nude mouse phenotype. In the normal mouse hair follicle, there are seven cell types that can potentially be pigmented, but only two are melanized and only one, the cortical cells, normally expresses Foxn1. In the absence of the transcription factor, the cortical cells lose pigmentation while the medullary cells retain it. The result is that the coat hairs are fairer than usual. “This also demonstrates the precision with which epithelial cells control the targeting of pigment,” said Brissette. Her lab is now focused on signaling in human skin to determine if the Foxn1–Fgf-2 relationship is conserved.
In addition to explaining why melanocytes are so discerning about what cells to stain, the finding could also have far-reaching consequences for the study of human disease. In melanoma, for example, melanocytes become independent of epithelial cells and begin making their own growth factors, including Fgf-2, Brissette explained. And in vitiligo, an autoimmune disease, melanocytes are lost. “People have been thinking about such disorders from just the melanocytes’ perspective,” Brissette said, “but if we focus on the counterparts, the epithelial cells, and start looking at the interactions between cells, maybe some useful therapies may emerge.”
