Melanin synthesis within melanosomes and their distribution to keratinocytes within the epidermal melanin unit determine skin pigmentation. The essential role of keratinocytes in the regulation of melanocyte growth and differentiation has been demonstrated (Donatien et al. 1993) but the regulation of keratinocyte–melanocyte interactions and the mechanism of melanosome transfer into keratinocytes are not yet fully understood.
Concerns of changes in skin color are frequently raised for medical or cosmetic reasons. Pigmentary disorders can be inherited (e.g., vitiligo, Waardenburg syndrome), acquired (e.g., postinflammatory pityriasis alba, idiophatic guttate hypomelanosis, melasma), medication related (e.g., minocycline, bleomycin, busulfan, zidovudine), and transmitted through infection (e.g., tinea versicolor). Changes in skin color are also desired for cosmetic reasons. Hyperpigmentation disorders are often treated with hydroquinones, retinoids, and tyrosinase inhibitors, but results of such treatments are sometimes disappointing (Hacker, 1996).
The protease-activated receptor 2 (PAR-2) (Nystedt et al. 1994,1995a) is a seven transmembrane G-protein-coupled receptor that is activated by a serine protease cleavage. The newly created N-terminus then activates the receptor as a tethered ligand. PAR-2 is also activated by SLIGRL, the peptide that corresponds to its new N-termini, independent of receptor cleavage (Nystedt et al. 1995b;Bohm et al. 1996). PAR-2 is expressed in skin (Santulli et al. 1995), but its biology is not yet completely understood. A role for PAR-2 activation in the inhibition of keratinocyte growth and differentiation has been suggested (Derian et al. 1997). We have recently demonstrated that the PAR-2 pathway is involved in the regulation of pigmentation (Seiberg et al. 2000).
Here we show that serine protease inhibitors that interfere with PAR-2 activation induce depigmentation by affecting melanosome transfer and distribution. Such agents may serve as an alternative treatment for depigmentation.
Materials and methods
Cells and cultures
Epidermal equivalents containing melanocytes (equivalents) (MelanoDerm) were from MatTek (Ashland, MA) and were maintained according to the manufacturer's instructions. Murine Melan-A cells (a kind gift from Dr. D. Bennett) were maintained according toBennet et al. (1987). Human HaCaT keratinocytes (a kind gift from Dr. N. Fusenig) were maintained according toBoukamp et al. (1988). Cell growth and viability were assayed using an MTS proliferation assay kit (Promega, Madison, WI) and alamarBlue (Accumed International, Chicago, IL), respectively, following the manufacturers' instructions. No SLIGRL and RWJ-50353 treatments resulted in no change in cell viability or proliferation (not shown). RWJ-50353 (Costanzo et al. 1996) and SLIGRL were dissolved in phosphate-buffered saline (PBS) for in vitro studies. All in vitro experiments were performed in triplicate and were repeated at least three times.
Melanosome isolation and transfer
Melanosome isolation was performed according toOrlow et al. (1994). Briefly, Melan-A cells were harvested in PBS supplemented with 10% (wt/vol) glucose (Sigma) and protease inhibitors (Complete, Boehringer Mannheim, Indianapolis, IN). Cell suspensions were homogenized on ice until cells were disrupted, and centrifuged at 500g, 10 min (4°C) to remove nuclei. Supernatants were supplemented with 90% (vol/vol) Percoll (Sigma) suspended in 0.25 M sucrose buffer (10 mM HEPES, 1 mM ethylenediamine tetraacetic acid, pH 7.2) to yield a final 28% Percoll concentration and were centrifuged at 10,000g, 45 min (4°C). Melanosome band was removed with a 25 gauge needle and syringe and stored at 4°C until use. Pretreated HaCaT keratinocytes were incubated with the isolated melanosomes for 2 h, followed by 10 PBS washes and Fontana-Mason (F&M) staining.
Histology, image analysis, and electron microscopy
Sections from swine biopsies and equivalents were stained with hematoxylin and eosin (not shown) or F&M, and monolayer cultures were stained with F&M (Sheenan & Hrapckak, 1980). F&M detects silver nitrate reducing molecules, which in skin identifies primarily melanin. At least three sections per equivalent or biopsy, three equivalents per experiment, were processed. At least 100 cells were used for each melanosome transfer group. Each experiment was repeated three times. Data are presented as the average of all experiments, with standard deviation (SigmaPlot 5.0, SPSS Science, Chicago, IL). F&M- or dihydroxyphenylalanine-stained sections, intact equivalents, or monolayers were used for image analysis. Empire Images database 1.1 was used on a Gateway 2000 P5-100 computer (Media Cybernetics, Silver Springs, MD). Image Pro plus version 4.0 was used for capturing images and image analysis. Parameters measured were the surface area of silver deposits and the tissue area or cell number, and the ratio of silver deposits per area or per cell was calculated. A value of one was assigned to untreated controls, and values of treatment groups were normalized to their relevant controls. Statistical analysis was done using SigmaStat 2.0 (SPSS Science) software. In all experiments there was no difference between PBS-treated equivalents and untreated controls (not shown). Electron microscopy was performed using standard protocols, as described byPiekos (1989).
Animals
Yucatan swine were from Charles River (Maine). Swine were housed in appropriately sized cages in an environmentally controlled room with a 12 h light, 12 h dark photoperiod and were supplied with food and water ad libitum. Animal care was based on the Guide for the Care and Use of Laboratory Animals, NIH Publication 85–23. Test compound or vehicle (ethanol:propylene glycol mix, 70:30 wt/wt) was applied topically, twice a day, 5 d per wk, for 8-9 wk. Treatments of individual swine were always arranged in a head to tail order on one side, and in a tail to head order on the other side of the animal. Color measurements of treated sites and nearby untreated regions were taken before the start of treatments and every 2 wk, using a Minolta Chromameter model CR300 (Osaka, Japan). Biopsies were taken using standard techniques. Swine experiments were repeated with at least three individual swine. The reversal of depigmentation effect was repeated twice.
Results
PAR-2 affects pigmentation via modulation of melanosome uptake
Our previous studies suggest that the PAR-2 pathway affects pigmentation via keratinocyte–melanocyte interactions (Seiberg et al. 2000). As shown in Figure 1, treatment of multilayered equivalents with the PAR-2 peptide agonist SLIGRL induces pigmentation in individual melanocytes. Treatment with RWJ-50353 (Costanzo et al. 1996), a serine protease inhibitor that affects PAR-2 activation, results in decreased pigmentation. As PAR-2 is expressed in keratinocytes but not in melanocytes and as keratinocyte-melanocyte contact is required for the PAR-2 effect on pigmentation (Seiberg et al. 2000), we tested the possible role of the PAR-2 pathway in melanosome uptake. Melanosomes isolated from Melan-A cells (according toOrlow et al. 1994) were incubated with HaCaT keratinocytes that were pretreated for 2 d with SLIGRL or RWJ-50353. As shown in Figure 2(a)(b)(c), an increase in uptake of melanosomes by the keratinocytes was observed following SLIGRL treatment. Ingestion of melanosomes was inhibited following RWJ-50353 treatment. Image analysis of the melanin area within the keratinocytes Figure 2d showed a 2.2-fold increase in melanosome uptake following PAR-2 activation, and an about 80% decrease following RWJ-50353 treatment (p = 0.006 and 0.002, respectively, t test). These results suggest that the keratinocyte PAR-2 is involved in melanosome uptake.
Figure 1.
The effect of RWJ-50353 and SLIGRL on pigmentation in individual melanocytes. Equivalents were treated with SLIGRL (10 
M) and with RWJ-50353 (0.1 M) for 3 d, followed by F&M staining of histologic sections. Melanocytes were from a Hispanic donor. Images of individual melanocytes are shown. Left panels, untreated control; middle panels, RWJ-50353; right panels, SLIGRL. Scale bar: 10 M.
Figure 2.
The PAR-2 pathway affects melanosome ingestion by keratinocytes. Melanosomes were isolated from Melan-A cells according toOrlow et al. (1994). Images of HaCaT keratinocytes untreated (a) or treated with SLIGRL (10 M, (b)) or RWJ-50353 (10 M, (c)) for 2 d, followed by a 2h incubation with the isolated melanosomes, extensive wash, and F&M staining. Scale bar: 10 M. (d) Melanin area per cell, quantified by image analysis.
Ultrastructural analysis of the RWJ-50353 effect
Equivalents treated with RWJ-50353 were analyzed for melanosome formation and distribution using electron microscopy. As shown in Figure 3, melanosomes in melanocytes of the treated samples Figure 3b were less mature and increased in number, relative to untreated controls Figure 3a. Dendrites containing mature melanosomes were identified within treated keratinocytes Figure 3c in higher numbers (nine in 27 fields) than in untreated controls (one in 25 fields). These data suggest abnormal melanosome formation and slow or impaired melanosome transfer into keratinocytes in the treated samples.
Figure 3.
Electron microscopy analysis of the RWJ-50353 effect on equivalents. Representative melanosomes as identified in (a) control and (b) RWJ-50353 (100 M) treated equivalents. (c) A melanocyte dendrite, containing melanosomes, inside an RWJ-50353-treated keratinocyte. Such structures could not be easily identified in untreated controls. Scale bar: (a, b) 0.1 M; (c) 0.5 M.
A dose-dependent in vivo whitening effect of RWJ-50353
Pigmented Yucatan swine were topically treated with RWJ-50353 twice daily for 8 or 9 wk. Skin color was measured by chromameter before the start and throughout the treatment period. A visible whitening effect was observed starting at the fourth week (highest dose). By 8 wk all RWJ-50353-treated sites exhibited whitening, with a saturation effect for the two highest doses Figure 4a. Chromameter measurements throughout this study Figure 4b showed that the saturation effect is time and concentration dependent. No signs of irritation were observed during the course of treatment.
Figure 4.
RWJ-50353-induced depigmentation in vivo. Yucatan swine were treated with vehicle (a), and 10 M (b), 50 M (c), and 250 M (d) of RWJ-50353 for 8 wk. (a) Picture of the swine (both sides) after 8 wk of treatment. (b) Chromameter measurements of skin color (L* scale, 0=black, 100=white) during the treatment phase. Both a dose response and a time response are observed.
Histologic analysis of skin biopsies taken at the eighth week of treatment further confirmed the depigmenting effect of RWJ-50353 Figure 5a, b, c, and d. Reduced pigment deposition is observed in RWJ-50353-treated skin throughout the basal layer, as well as suprabasally. No other changes were observed in the treated sections, skin architecture was intact, and no inflammatory infiltrate was detected. Image analysis of F&M-stained sections Figure 5e demonstrated the dose-response lightening effect of RWJ-50353, with statistically significant changes at 50 and 250 M (p =0.003, p <0.001, respectively, t test).
Figure 5.
Histologic analysis of swine skin samples treated with RWJ-50353. Swine were treated as described in Figure 4 and biopsies were taken at the eighth week of treatment. F&M staining of skin sections revealed a dose-dependent reduction in melanin deposition at treated sites: (a) vehicle; (b)–(d) 10 M, 50 M, and 250 M of RWJ-50353. Scale bar:20 M. (e) Relative pigmentation, calculated as melanin area per epidermis area and normalized to untreated controls, obtained by image analysis of F&M-stained skin sections.
The lightening effect of RWJ-50353 was reversible. Yucatan swine treated for 8 wk were followed visually with no treatment for a further 4 wk. Darkening of the depigmented sites was visible by the fourth week. Histologic analysis Figure 6a, b, c, d and e showed a gradual increase in pigment production and distribution even before the visual observation of re-pigmentation. Image analysis of F&M-stained sections Figure 6f quantified this gradual re-pigmentation. Statistical analysis (t test) showed significant lightening (p=0.003) following 8 wk of RWJ-50353 treatment (relative to control), and a slow increase in re-pigmentation weekly, which became statistically significant (relative to the 8 wk treatment time point) only after 4 wk (p =0.003).
Figure 6.
The depigmenting effect of RWJ-50353 is reversible. Yucatan swine were treated for 8 wk with 250 M of RWJ-50353, and followed without treatment for a further 4 wk. Biopsies were taken before the start of treatment (a), after 8 wk (completion of treatment phase, b), and at the ninth, tenth, and twelfth weeks (c–e, 1–4 wk after treatment was terminated). Scale bar: 12 M. F&M-stained sections revealed re-pigmentation after treatment had been stopped, with no irritation or other side-effects. Differences in epidermal thickness result from the different sites of biopsies, and have no correlation to treatments. (f) Image analysis of F&M-stained sections was used to quantify pigmentation, relative to the untreated control.
Ultrastructural analysis of RWJ-50353-treated skin
Skin samples from Yucatan swine treated with RWJ-50353 for 8 wk were analyzed by electron microscopy. Melanosomes within keratinocytes of treated sites were 30%-40% smaller and less pigmented compared with controls Figure 7a, b and c. Only single melanosomes, and no melanosome complexes, were observed in either the control or the treated sites. Moreover, the distribution of melanosomes within the treated skin was abnormal. Melanosomes were detected mainly at the epidermal-dermal border, compared to a more random distribution in the untreated controls Figure 7d, e).
Figure 7.
Electron micrpscopy analysis of Yucatan swine skin treated with RWJ-50353. (a) A representative melanosome inside a keratinocyte of untreated swine skin. (b), (c) Representative melanosomes inside keratinocytes of RWJ-50353 (250 M) treated swine skin are smaller and less pigmented than control ones. (d) A random distribution of melanosomes in control swine epidermis. (e) Melanosomes of RWJ-50353 (250 M) treated swine skin are detected mainly at the epidermal–dermal border (marked). Scale bar: (a-c) 0.05 M; (d, e) 0.8 M.
Discussion
Various pigmentary disorders and cosmetic applications require the use of depigmenting agents. Currently available topical agents used for the reduction of pigmentation include tyrosinase inhibitors and melanocyte-cytotoxic agents (reviewed inJimbow & Jimbow, 1998). Although advances have been made, there is currently a need for safer, more effective, and less irritating depigmenting therapies. Basic understanding of the regulation of pigment production and distribution could aid in the identification of alternative depigmenting agents.
The process of melanogenesis is well studied. Melanin is produced within melanosomes, which later migrate into the melanocyte's dendrite tips using myosin V filaments (Wei et al. 1997) and a dynein ''motor'' (Ogawa et al. 1987). The regulation of pigment production by keratinocyte–melanocyte interactions and the subsequent transfer and distribution of pigment into keratinocytes, however, are not well understood. Several mechanisms for melanosome transfer from the dendrite tips into the keratinocytes have been suggested, including phagocytosis, release of melanosomes into intercellular spaces followed by endocytosis, direct inoculation (''injection''), and keratinocyte-melanocyte membrane fusion. No molecular mechanism has been identified for melanosome transfer (reviewed inYamamoto & Bhawan, 1994;Jimbow & Sugiyama, 1998).
As PAR-2 modulation affects pigmentation only when a keratinocyte-melanocyte contact is established (Seiberg et al. 2000), we looked at the effect of the PAR-2 pathway on melanosome uptake by keratinocytes. Here we show the first molecular mechanism involved in melanosome uptake, as PAR-2 affecting agents regulate the ingestion of melanosomes by keratinocytes in culture. Preliminary data not presented here show that PAR-2 affects keratinocyte ingestion of microspheres and Escherichia coli particles, suggesting a role for PAR-2 in keratinocyte phagocytosis. The keratinocyte receptor PAR-2 could therefore be a part of the regulatory mechanism of skin pigmentation. Synthetic compounds that affect the PAR-2 pathway are shown to modulate melanosome ingestion. RWJ-50353, a serine protease inhibitor that reduced melanosome uptake in culture, is shown to have a dose-dependent depigmenting activity in vivo, with no irritation or other side-effects.
By studying the ultrastructural changes in melanosomes and their transfer following RWJ-50353 treatment, we identified the accumulation of melanosomes within treated melanocytes, with an increase in early stages and empty melanosomes. We also identified an increase in melanosome-containing dendrites within treated keratinocytes. Although we cannot rule out other mechanisms, we suggest that neither melanosome formation and function nor dendrite penetration into the keratinocytes are directly affected by the drug. We propose that RWJ-50353-treated keratinocytes are unable to actively take or receive melanosomes from the presenting dendrites. This keratinocyte ''inability'' leads to the accumulation of melanosomes in the melanocytes, and could possibly turn on a negative feedback mechanism that slows pigment production. Our earlier data document reduced TRP-1 expression following RWJ-50353 treatment (Seiberg et al. 2000). Such a negative feedback mechanism could explain the reduced TRP-1 expression, as TRP-1 is a major melanosomal glycoprotein. Following RWJ-50353 treatment the melanocytes contain more melanosomes than required for homeostasis, and therefore melanosome production, and TRP-1 synthesis, are reduced. The increase in pigment deposition within melanocytes following PAR-2 activation cannot be mechanistically explained by melanosome trafficking only, and requires further study.
In vivo ultrastructural studies revealed an abnormal distribution of melanosomes in RWJ-50353-treated swine skin. The polarity of this distribution, at the dermal-epidermal border, could provide a clue to extracellular matrix components or adhesion molecules involved in melanosome transfer. It is not likely that melanosomes were released into the intercellular space, and were not taken immediately by keratinocytes, as such a mechanism should result in melanosome accumulation and skin darkening, which were not observed. Therefore, we assume that these melanosomes were later translocated into the keratinocyte using either a PAR-2-independent mechanism or a re-activated PAR-2. As in vivo we could not completely inhibit melanogenesis or pigment transfer with RWJ-50353 (up to 10mM, twice daily treatment), we suggest that either the keratinocyte PAR-2 is not the only mechanism for melanosome transfer, or that PAR-2 is re-activated when RWJ-50353 levels are reduced. It is important to note that in RWJ-50353-treated skin the transferred melanosomes are of poor quality, reflecting changes in melanosome formation and/or melanogenesis prior to their transfer. These changes agree with our suggested negative feedback mechanism, by responding to accumulation of nontransferred melanosomes with reduced new pigment production. Although the drug indirectly affected melanosome quality, it had no effect on the mode of melanosome transfer. As expected for dark skinned individuals, the dark skinned swine melanosomes were always transferred singly, and melanosome complexes were never observed, regardless of the treatment.
The epidermal-melanin unit, a functional unit that produces and distributes melanin (reviewed inOrtonne, 1995), is shown to have a role in the regulation of pigmentation. The keratinocyte PAR-2 is involved in the regulation of melanosome transfer, and therefore affects skin pigmentation. Modulation of the PAR-2 pathway with serine protease inhibitors such as RWJ-50353 could offer an alternative to depigmenting therapies.
References
| 1. | Bennett DC, Cooper PJ & Hart IR. A line of non-tumorigenic mouse melanocytes, syngeneic with the B16 melanoma and requiring a tumour promoter for growth. Int J Cancer (1987) 39: 414–418. | PubMed | ISI | ChemPort | |
| 2. | Bohm SK, Kong WY & Bromme D et al. Molecular cloning, expression and potential functions of the human proteinase activated receptor 2. Biochem J (1996) 314: 1009–1016. | PubMed | ISI | |
| 3. | Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A & Fusenig NE. Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol (1988) 106: 761–771. | Article | PubMed | ISI | ChemPort | |
| 4. | Costanzo MJ, Maryanoff BE & Hecker LR et al. Potent thrombin inhibitors that probe the S1' subsite: tripeptide transition state analogues based on a heterocycle-activated carbonyl group 1. J Med Chem (1996) 39: 3039–3043. | Article | PubMed | ISI | ChemPort | |
| 5. | Derian CK, Eckardt AJ & Andrade-Gordon P. Differential regulation of human keratinocyte growth and differentiation by a novel family of protease-activated receptors. Cell Growth Diff (1997) 8: 743–749. | PubMed | ISI | ChemPort | |
| 6. | Donatien P, Surleve-Bazeille JE, Thody AJ & Taieb A. Growth and differentiation of normal human melanocytes in a TPA-free, cholera toxin-free, low serum medium and influence of keratinocytes. Arch Dermatol Res (1993) 285: 385–392. | Article | PubMed | ISI | ChemPort | |
| 7. | Hacker SM. Common disorders of pigmentation: when are more than cosmetic. cover-ups required?. Postgrad Med (1996) 99: 177–186. | PubMed | ISI | ChemPort | |
| 8. | Jimbow K & Jimbow M. Chemical, pharmacologic and physical agents causing hypomelanoses. In: Nordlund JJ, Boissy RE, Hearing VJ, King RA, Ortonne J-P, eds.The Pigmentary System (1998) Oxford: Oxford University Press pp 621–627. |
| 9. | Jimbow K & Sugiyama S. Melanosomal translocation and transfer. In: Nordlund JJ, Boissy RE, Hearing VJ, King RA, Ortonne J-P, eds.The Pigmentary System (1998) Oxford: Oxford University Press pp 107–114. |
| 10. | Nystedt S, Emilsson K, Larsson AK, Strombeck B & Sundelin J. Molecular cloning and functional expression of the gene encoding the human proteinase-activated receptor 2. Eur J Biochem (1995a) 232: 84–89. | Article | PubMed | ISI | ChemPort | |
| 11. | Nystedt S, Emilsson K, Wahlestedt C & Sundelin J. Molecular cloning of a potential proteinase activated receptor. Proc Natl Acad Sci USA (1994) 91: 9208–9212. | PubMed | ChemPort | |
| 12. | Nystedt S, Larsson AK, Aberg H & Sundelin J. The mouse proteinase-activated receptor-2 cDNA and gene. Molecular cloning and functional expression. J Biol Chem (1995b) 270: 5950–5955. | Article | PubMed | ISI | ChemPort | |
| 13. | Ogawa K, Hosoya H & Yokota E et al. Melanoma dynein: evidence that dynein is a general ''motor'' for microtubule-associated cell motilities. Eur J Cell Biol (1987) 43: 3–9. | PubMed | ISI | ChemPort | |
| 14. | Orlow SJ, Zhou BK, Chakraborty AK, Drucker M, Pifko-Hirst S & Pawelek JM. High molecular weight forms of tyrosinase and the tyrosinase-related proteins: evidence for a melanogenic complex. J Invest Dermatol (1994) 103: 196–201. | Article | PubMed | ISI | ChemPort | |
| 15. | Ortonne J-P. Melanins and melanocytes, form and function. Clin Drug Invest (1995) 10 Suppl.2: 1–16. |
| 16. | Piekos WB. Temporal separation of rhabdom shrinkage and MVB formation in the light-adapting crayfish retina. J Exp Zool (1989) 250: 17–21. | Article | ISI | |
| 17. | Santulli RJ, Derian CK & Darrow AL et al. Evidence for the presence of a protease-activated receptor distinct from thrombin receptor in human keratinocytes. Proc Natl Acad Sci (1995) 92: 9151–9155. | PubMed | ChemPort | |
| 18. | Seiberg M, Paine C, Sharlow E, Costanzo M, Andrade-Gordon P, Eisinger M & Shapiro S. PAR-2 regulates pigmentation via keratinocyte–melanocyte interactions. Exp Cell Res (2000) 254: 25–32 10.1006/excr.1999.4692. | Article | PubMed | ISI | ChemPort | |
| 19. | Sheenan DC & Hrapckak BB. eds. Theory and Practice of Histo-Technology. (1980) St Louis: CV Mosby pp 223–277. |
| 20. | Wei Q, Wu X & Hammer JA, 3rd. The predominant defect in dilute melanocytes is in melanosome distribution and not cell shape, supporting a role for myosin V in melanosome transport. J Muscle Res Cell Motil (1997) 18: 517–527. | Article | PubMed | ISI | ChemPort | |
| 21. | Yamamoto O & Bhawan J. Three models of melanosome transfer in caucasian facial skin: hypothesis based on an ultrastructural study. Pig Cell Res (1994) 7: 158–169. | ISI | ChemPort | |
Acknowledgments
We would like to thank D. Burtis, A. Elsawaf, A. Johnson, F. Liebel, and J. Pote for technical assistance and K. Martin for fruitful discussions throughout this study. Special thanks to Dr. D. Bennett for Melan-A cells and Dr. N. Fusenig for HaCaT keratinocytes.
