Hypopigmenting agents: an updated review on biological, chemical and clinical aspects

Authors


*Address correspondence to Francisco Solano, e-mail: psolano@um.es

Summary

An overview of agents causing hypopigmentation in human skin is presented. The review is organized to put forward groups of biological and chemical agents. Their mechanisms of action cover (i) tyrosinase inhibition, maturation and enhancement of its degradation; (ii) Mitf inhibition; (iii) downregulation of MC1R activity; (iv) interference with melanosome maturation and transfer; (v) melanocyte loss, desquamation and chemical peeling. Tyrosinase inhibition is the most common approach to achieve skin hypopigmentation as this enzyme catalyses the rate-limiting step of pigmentation. Despite the large number of tyrosinase inhibitors in vitro, only a few are able to induce effects in clinical trials. The gap between in-vitro and in-vivo studies suggests that innovative strategies are needed for validating their efficacy and safety. Successful treatments need the combination of two or more agents acting on different mechanisms to achieve a synergistic effect. In addition to tyrosinase inhibition, other parameters related to cytotoxicity, solubility, cutaneous absorption, penetration and stability of the agents should be considered. The screening test system is also very important as keratinocytes play an active role in modulating melanogenesis within melanocytes. Mammalian skin or at least keratinocytes/melanocytes co-cultures should be preferred rather than pure melanocyte cultures or soluble tyrosinase.

Introduction

Melanin plays an important role in protecting human skin from the harmful effects of UV sun radiation and in scavenging toxic drugs and chemicals. It determines our race and phenotypic appearance. The accumulation of an abnormal melanin amount in different specific parts of the skin as more pigmented patches (melasma, freckles, ephelide, senile lentigines etc.) might become an esthetic problem yet. In Western countries, skin lighteners are applied for the prevention and treatment of irregular hyperpigmentation, such as melasma, freckles or age spots. In Asia, the use to make the skin whiter is widely extended by traditional beliefs.

In man, the variation in skin color occurs at the functional level of the epidermal melanin unit. Basically, this system relates to the density of melanocytes, the number, size and dispersion of melanosomes transferred to epidermal keratinocytes, the nature of the pigment and its degradation rate. Although hypopigmentation can be achieved not only by biological and chemical, but also by physical treatments (Jacques and McAuliffe, 1991), in this review, we deal basically with the first two types of agents. First, we present brief comments on biological agents affecting human pigmentation, and then will move to the field of the plant natural products that can be used for hypopigmenting purposes. Moreover, although the management of acquired-hyperpigmentary skin disorders is still a hard challenge for dermatologists, we tried to provide a useful description of effective depigmenting treatments. Besides in-vitro and in-vivo studies in animal models, we will focus our attention on randomized clinical trials in hyperpigmentary disorders, whose evaluation is particularly difficult because of the lack of univocal recommendations on how those clinical trials should be conducted. Despite the definition of some empirical index, such as MASI (that is calculated using the area of the face involved with hyperpigmentation) or the melasma severity scale, in fact, several studies have been realized with different approaches and the clinical efficacy of depigmenting treatments have been evaluated by taking in account different parameters, the majority of them based on subjective observations.

Biological agents that interfere with mammalian pigmentation

A number of biological effectors can reduce normal or hyperpigmentation in mammals by various means that covers melanogenesis biochemistry itself down to the destruction of the melanocyte. The mechanisms by which each effector acts are summarized in Table 1. A very good recent review has been published by Slominski et al. (2004). We would rather focus, in the following, on some updates, additions, novelties and particular mechanisms leading to hypo-or/and depigmentation.

Table 1.   Biological effectors causing hypopigmentation
MechanismBiological effector/compoundReference
Tyrosinase, TRP1, TRP2 inhibition, mostly through Mitf downregulationTGF-β1Kim et al. (2004a); Martinez-Esparza et al. (1999)
TNFαMartinez-Esparza et al. (1998)
IL1α, β, IL-6Choi et al. (2005); Hu (2000); Krasagakis et al. (1995); Swope et al. (1991, 1994)
Lysophosphatidic acidKim et al. (2004c)
C2 ceramidesKim et al. (2002a)
SphingosylphosphorylcholineKim et al. (2005b)
Simian virus 40 (SV-40)Prince et al. (2001)
Vitamin EFunasaka et al. (1999)
Dkk1Yamaguchi et al. (2004)
Calpain inhibitorsOhguchi et al. (2005)
Increased tyrosinase ubiquitinationPhospholipase D2Kageyama et al. (2004)
Fatty acidsAndo et al. (2004, 2006)
Inhibition of tyrosinase maturationAntisens ferritin light chainMaresca et al. (2006)
GlycosphingolipidsSprong et al. (2001)
Decrease of MC1R activityAgouti proteinAbdel-Malek et al. (2001); Barsh et al. (2000); Jordan and Jackson (1998); Voisey and van Daal (2002); Wolff (2003)
cAMP inhibition (anti-inflammatory agents)CorticosteroidsGallardo and Johnson (2004); Kumar and Adolph (1998); Nanda et al. (2006)
Interference with: melanosome maturation melanosome transferTGF-β1Martinez-Esparza et al. (2001)
Serine protease inhibitorsSeiberg et al. (2000a,b)
Niacinamide (vitamin B3)Hakozaki et al. (2002)
Lectins, neoglycoproteinsMinwalla et al. (2001)
Medication side effects (melanocyte loss)
 Liberation of toxic melanin synthesis intermediatesMelatonin receptor activationSlominski et al. (1989)
  Toxic fungal metabolitesMalessezia yeasts (Pityriasis Versicolor), LichenCarlson et al. (2002); De Luca et al. (1996); Podobinska et al. (2003); Thoma et al. (2005)
 Autoimmune:ImiquimodBrown et al. (2005); Fleming et al. (2004)
  Activation of Toll-like receptor-7/8 pathwayMelanoma-associated antigensAu et al. (2001); Luiten et al. (2005); Tsao et al. (2002); Yee et al. (2000)
  Melanocyte death mediated by dendritic/T cells4-tertiarybutylphenol
Triglycidyl-p-aminophenol
Jappe et al. (2005); Kroll et al. (2005)
  UnknownIFN-γCarroll et al. (1997); Yeager et al. (1992)
Alkylating agent Bisulfan
Imatinib mesilate
Hasan et al. (2003); Legros et al. (2005); Raanani et al. (2002); Sharma et al. (2005); Tsao et al. (2003)

Decrease in the intracellular levels of tyrosinase and other melanogenic enzymes

The most interesting point for this approach is Mitf (Microphthalmia-associated transcription factor; Steingrimsson et al., 2004; Vance and Goding, 2004). Mitf belongs to the basic helix-loop-helix-zip family of trabscription factors and it is crucial as it regulates both melanocyte proliferation as well as melanogenesis and is the major regulator of tyrosinase and the related enzymes (TRPs), as well as many melanosome structural proteins such as pMel17. Mutations in human MITF lead to the hypopigmentary and deafness Waardenburg syndrome type 2A. As Mitf is regulated at both the transcriptional level by the Wnt signalling pathway, as well as cAMP and IL6, and at the protein level by both p38 stress signalling and the MAP kinase pathway, potentially any compound that will affect these signalling pathways will also affect Mitf and through that melanogenesis.

Transforming growth factor (TGF-β1) plays an inhibitory role in pigment formation. It has been first observed that it is able to interfere with tyrosinase synthesis and possibly with the intracellular stability of the protein itself (Martinez-Esparza et al., 1999). Kim et al. (2004a) later explained that the same factor is able to downregulate Mitf as well, and observed that, unlike other growth factors, TGF-β1 induced a significant delay in extracellular signal regulated kinase (ERK) activation that also contribute to this mechanism. Lysophosphatidic acid (Kim et al., 2004c) and C2 ceramides are also able to respectively activate Mitf degradation or to block Mitf expression mediated by an initial effect on AKT/PKB and ERK (Kim et al., 2002a). More recently, the same authors also provided another evidence of the involvement of ERK pathway activation regulating melanogenesis by another signaling-lipid mediator, sphingosylphosphorylcholine (Kim et al., 2005b).

In addition to interleukin 1 and 6, tumor necrosis factor (TNFα) is able to decrease pigmentation by acting also on tyrosinase activity, but this seems to occur at rather high concentrations (Martinez-Esparza et al., 1998). This was not restricted to tyrosinase but also was true for TRP1 and TRP2. The authors outlined a double effect both on tyrosinase synthesis and degradation.

There are a series of papers reporting the hypopigmenting effects of fatty acids (Ando et al., 1995, 1999) acting on different steps of the melanogenic pathway. These effects are complex, as the unsaturated linoleic acid decreases tyrosinase activity, but the saturated palmitic or stearic acids increase it (Ando et al., 1995, 1998). Topical application of linolenic, linoleic and oleic acids (efficiency in decreasing order) produce a bleaching effect on guinea pig skin stimulated with UV light (Ando et al., 1998). As the number of melanosomes or the level of tyrosinase mRNA do not seem to be influenced, melanin inhibition by linoleic acid is probably due to a decrease of the amount of active tyrosinase inside the melanocytes because of a stimulation of tyrosinase ubiquitination and degradation by the proteasome (Ando et al., 1999, 2006; Kageyama et al., 2004). On contrary, those unsaturated fatty acids seem to stimulate desquamation of the skin and the stratum corneum turnover. Lee et al. (2002a,b) investigated in Korean patients with melasma the clinical efficacy of topical application of linoleic acid in combination with lincomycin and betamethasone valerate. The presence of linoleic acid in the formulation was essential for the skin lightening effect without apparent side effects.

As unsaturated fatty acids, phospholipase D2 also decreases melanogenesis through the same ubiquitin-mediated degradation of tyrosinase (Ando et al., 2004). However, the increase of phospholipase D2 activity is difficult to assess, as phorbol esters are the only well-known inducer for that enzyme, but they cannot be used for their carcinogenic effects. On contrary, it has been also described that phospholipases and arachidonic acid produced a stimulation of melanogenesis (Maeda et al., 1996), and it is possible that hyperpigmentation associated to the UV light response are mediated by those agents. If so, phospholipases and unsaturated fatty acids have not always hypopigmenting effects.

Moreover, exogenous oncogenes such as the Simian virus (SV-40) may also decrease pigmentation by inhibiting Mitf (Prince et al., 2001). Tyrosinase maturation leading to a full glycosylated and active enzyme was found to be inhibited by the expression of an antisens to ferritin light chain (Maresca et al., 2006) suggesting the involvement of the latter in pigmentation modulation: this situation can be encountered due to a particular increase in oxidative stress of melanoma cells. Glycosphingolipids are also required for sorting melanosomal proteins in the Golgi complex (Sprong et al., 2001).

α-Tocopherols have also been tested as hypopigmenting agents, although its effect is not due to a direct inhibition on tyrosinase activity. The most possible mechanism of action is to block dopaquinone and subsequent chemical oxidations in the polymerization pathway leading to the pigment. In a number of cell types, vitamin E shows antioxidant properties, inhibits lipid peroxidation but also enhances glutathione synthesis. This last effect is a plausible way to control apparent hyperpigmentation, as glutathione derivates dopaquinone to pheomelanin. Thus, as more dopaquinone is branched to pheomelanin synthesis, as less eumelanin is accumulated, leading to a decrease in skin apparent pigmentation because of the lighter color of pheomelanin. Tocopherols conjugated with phenols at position 4, such as resorcinol, seem to have a strong effect (Shimizu et al., 2001). These compounds could exert a double action, the inhibition of tyrosinase by the phenol moiety and the lipophilicity and ROS scavenger activity of the α-tocopherol moiety (Funasaka et al., 1999). However, vitamin E conjugated to ferulic acid has been efficient in facial treatments for opposite effect, stimulation of pigmentation due to a tyrosinase-independent oxidation of the ferulic moiety.

The skin on the palms of the hands and soles of the feet (palmoplantar skin) are less pigmented than the rest of the body skin. Skin transplanted from trunk epidermis gradually becomes much less pigmented. The cause of that are some transcription factors preferentially expressed by fibroblasts of the underlying dermis (Yamaguchi et al., 2004). The protein product of the gene dickkopf1 (DKK1) is a negative regulator of Wnt signaling pathway that decreases Mitf and thus inhibits melanocyte growth and pigment production. Similarly, calpain inhibitors caused marked decrease in both tyrosinase and mRNA levels in B16 cells (Ohguchi et al., 2005). Thus, treatments based on specific stimulation of DKK1 or calpain inhibition in hyperpigmented regions might be interesting new approaches to get depigmentation.

Decrease in the Melanocortin Receptor 1 (MC1R) activity

Hypopigmentation in man can result from mutations in the MC1R gene (Garcia-Borron et al., 2005; Rees, 2003), its expression and functionality of its products (Sánchez-Laorden et al., 2006). While loss-of-function mutations of melanocortin 1 receptor (MC1R) are associated with red hair phenotype in man (Schioth et al., 1999); and agouti signaling protein (ASIP) gene polymorphism may explain reduction of pigmentation in some populations (Bonilla et al., 2005), a switch from eu- to pheomelanogenesis substantially reducing pigmentation can also take place in presence of a fully functional MC1R.

ASIP regulates mammalian pigmentation (e.g. mouse hair follicles) by antagonizing the binding of αMSH to MC1R (Abdel-Malek et al., 2001; Jordan and Jackson, 1998; Voisey and van Daal, 2002; Wolff, 2003). The signaling inhibition through the receptor results from two effects: (i) a direct competition at the binding site; and (ii) a downregulation of the receptor signaling (Barsh et al., 2000). The yellow color is the result of a chronic antagonism of ASIP at the MC1R while the obese phenotype is rather the result of antagonisms at the MC3R and MC4R levels. However, despite the existence of ASIP in man, its inhibitory role of eumelanogenesis has yet to be defined.

MSH and other proopiomelanocortin-derived peptides seem not essential to ensure MC1R basal activity as suggested in POMC null mutant mice where eumelanin hair pigmentation is practically not affected (Slominski et al., 2005; Smart and Low, 2003). In fact, other cAMP inducer systems can also modulate skin pigmentation: PGE2 stimulates a protein G-coupled receptor, the steep inhibition of which may cause hypopigmentation like the one observed after intralesional or intra-articular administration of corticosteroids (Gallardo and Johnson, 2004; Kumar and Adolph, 1998; Nanda et al., 2006).

Unlike MSH and ACTH, some other hormones decrease intracellular cAMP levels, and thus interfere with melanogenesis. This is the case of androgens combined with the sex-hormone binding globulin (Tadokoro et al., 2003). However, the hypopigmenting effect is weak, and the side effects of this type of hormones made them not especially good candidates for skin lightening purposes. Some naturally occurring proteins, such as β-lactoglobulin and kappa-casein (Nakajima et al., 1996, 1997) have also been effective to produce a certain inhibition of melanogenesis in melanocyte cultures, although their relation to MSH receptor and actual effect is unknown.

Interference with melanosome maturation and transfer

The addition of TGF-β1 to melanocytes yielded significantly more stage III melanosomes even when the cells were concomitantly treated with αMSH to boost their fully melanized stage IV melanosomes (Martinez-Esparza et al., 2001). The decrease in melanosome transfer to neighboring keratinocytes may cause a skin lightening effect. Four mechanisms were investigated:

  • 1The inhibition of serine protease results in an impaired activation of protease-activated receptor 2 (PAR-2) on the keratinocyte leading to the accumulation of melanosomes within the melanocyte (Seiberg et al., 2000a). Thus, inhibition of this receptor can block the transfer among these cells and therefore the pigment dispersion to keratinocytes (Seiberg et al., 2000b). This suggests a novel mechanism for the regulation of pigmentation, mediated by the inhibition of the keratinocytes receptor PAR-2. The unique inhibitor so far tested (RWJ-50353) has not been well studied as hypopigmenting agent. Centaureidine, a flavonoid glucoside isolated from yarrow, reduces the dendrites growth and the transfer of melanosomes to keratinocytes. In addition, it diminishes the amount of tyrosinase, although it does not affect TRP1 levels (Saeki et al., 2003).
  • 2Niacinamide (Vitamin B3) was found to inhibit melanosome transfer to keratinocytes both in vitro and in man. However, the associated mechanism of action remains to be elucidated (Hakozaki et al., 2002).
  • 3In an in-vitro model, Minwalla et al. (2001) demonstrated the role of glycosylated residues on melanocyte and keratinocyte membranes as part of receptor-mediated endocytosis facilitating melanosome transfer. The authors identified and evaluated lectins and neoglycoproteins able to inhibit this transfer.

Melanocyte loss

Another mechanism leading to the hypopigmentation is inducing irreversible damage to melanocytes leading to significant decrease of their number. Melatonin receptor activation has been suggested as a peculiar biological mechanism to cause toxic melanin synthesis intermediate release and its possible involvement in vitiligo (Slominski et al., 1989). Some fungal metabolites of dermatophytes, Malessezia yeasts and lichen were found toxic to melanocytes explaining depigmented lesions on the skin (Carlson et al., 2002; De Luca et al., 1996; Podobinska et al., 2003; Thoma et al., 2005).

Also, melanocyte loss can occur as a medication side effect. In fact, depigmentation is considered as a good sign in melanoma treatment. Vaccines using melanoma-associated antigens were reported by many authors to produce depigmentation by eliciting an autoimmune response directed against malignant but also normal melanocytes (Au et al., 2001; Luiten et al., 2005; Tsao et al., 2002; Yee et al., 2000). The mechanism involves specific dendritic/T-cell activation and melanocyte recognition. The latter can also be triggered by treatment with chemical sensitizers: 4-tertiary butyl phenol (Kroll et al., 2005) and triglycidyl-p-aminophenol (Jappe et al., 2005) seem to be candidates.

Another intriguing but yet very promising approach to melanoma treatment is imiquimod, an activator of Toll-like receptor-7/8 pathway. Topical treatment of melanocytic lesions resulted in clinical responses and depigmentation (Brown et al., 2005; Fleming et al., 2004).

Interferon γ and the alkylating agent bisulfan treatments were reported to cause depigmentation (Carroll et al., 1997; Yeager et al., 1992). The related mechanisms remain, however, unknown. A recent interesting observation of generalized vitiligo-like depigmentation following the treatment with Imatinib mesilate has been reported and confirmed by several independent teams (Hasan et al., 2003; Legros et al., 2005; Raanani et al., 2002; Sharma et al., 2005; Tsao et al., 2003). The drug is used to treat chronic myeloid leukemia and also seems to activate pre-existing vitiligo lesions.

Chemical agents that decrease mammalian pigmentation

Chemical compounds with depigmenting activity are used in dermatology and cosmetics for a long time. As the pioneers of hypopigmentation research (Riley, 1969), a lot of work on this subject has been carried out. In the last years, a huge number of phenolic compounds have been tested as inhibitors of melanin synthesis and as photoprotectors. Naturally occurring herbal extracts, active compounds such as phenols, flavonoids, coumarins and other derivatives have gained attention as putative hypopigmenting agents. The classification based on the structure and mechanism of action has become difficult because of the high amount of products reported and the diverse systems used for testing, from just mushroom tyrosinase, mammalian tyrosinase, melanocytic cultures, co-cultures of keratinocytes and melanocytes and finally in-vivo application to animal skin. Some recent reviews have been devoted to this subject, from the basic point of view (Kim and Uyama, 2005; Seo et al., 2003) or focused to application purposes as innovative technology in dermatology (Briganti et al., 2003).

Most of these agents affect tyrosinase, but other complementary mechanisms at different levels of the melanin epidermic unit could also have a hypopigmenting effect. It should be also taken into account that the design of new hypopigmentation treatments needs not only to block skin darkening by inhibiting de novo melanogenesis, but also to reduce the already existing pigmentation. Most of the treatments are demanded because the situation of a pre-existing local hyperpigmentation exists, and the rate of basal pigment elimination is low. Treatments to increase skin desquamation are useful to this purpose, but it would be even more useful to accelerate melanosome degradation before the cells reach the stratum corneum to be finally eliminated. Melanosome and melanin degradation is a poorly studied process (reviewed by Borovansky and Elleder, 2003) and the data about this approach are scarce. Lysosomal hydrolases are able to degrade melanosome constituents except the melanin moiety. This polycyclic aromatic polymer seems to be degraded only by oxidative breakdown, and enzymes such as cytochrome P450 and especially phagosomal NADPH oxidase could be involved in at least a partial biodegradation. Any successful exploration in this field would be an invaluable contribution to hypopigmenting treatments.

The whitening effects of some pure agents or natural extracts are usually compared with the most widely used and effective agents so far used, i.e. hydroquinone, kojic or retinoids acids to provide an idea about the efficiency of these agents or extracts. We will start with these very used agents.

Tyrosinase inhibitors

Simple phenols. Hydroquinone and its derivatives

Hydroquinone (Figure 1A1) is a well-known tyrosinase inhibitor. Its bleaching properties were discovered >50 years ago, when it was observed that colored tanners wearing rubber gloves acquired discolored areas on hands and forehands. The studies on the cause of this effect pointed out to hydroquinone, an agent used in rubber synthesis. Besides tyrosinase inhibition through interaction with copper at the active site, alteration of melanosome functions, depletion of glutathione, generation of reactive oxygen species and subsequent oxidative damage of membrane lipids and proteins may play a role in the depigmenting effect of hydroquinone (Briganti et al., 2003). Due to its effectiveness, the use of this molecule for whitening skin treatments was rapidly extended. The efficacy and adverse effects of hydroquinone 4% were evaluated by Ennes et al. (2000) in a double-blind placebo-controlled trial involving 48 patients with melasma on the face. Their results indicate a total improvement of melasma in 38% of patients treated, and a partial reduction of hyperpigmentation in 57% of patients. However, its cytotoxicity and side-effects, such as permanent hypomelanosis or amelanosis were quite high so that its usage dose was limited to 2%, a concentration that has been reported to improve hypermelanosis in 14–70% of the patients (Ortonne and Passeron, 2005). It has been frequently used for many years, mixed with other compounds to increase its efficiency (Bolognia et al., 1995; Guevara and Pandya, 2001) in a number of melasma treatments, but today the human use of hydroquinone has important legal restrictions. In fact, though hydroquinone is still the gold-standard of depigmenting agents, the compound was banned in Europe in December, 2000, Eur. Com. 6 for general cosmetics purposes and formulation with this compound are available only by prescription of physicians and dermatologists.

Figure 1.

 Chemical structures of selected hypopigmenting agents. A1, A2 and A3 are three well-known tyrosinase inhibitors; B, hydroxystilbenes; C, 3 common epicatechin-gallates; D, flavonols; E, A coumarin; F, aloesin; G, kuraninone, a prenylated flavonoid; H, 18-β-glycyrrhizin, a saponin; I, Oregonin; J, Yohimbine.

The related compound hydroquinone monobenzyl ether is metabolized inside cells to yield the corresponding quinone which causes permanent depigmentation. Its use is more appropriate for patients with diffuse vitiligo (Njoo et al., 1999). In general, all diphenols leading to a reactive quinone are putative cytotoxic agents. This type of hypopigmenting agents does so by their cytotoxic action on melanocytes rather than by tyrosinase inhibition. This is the case of 4-(p-hydroxyphenyl)-2-butanone (Fukuda et al., 1998) and 4-n-butylresorcinol (Kim et al., 2005a, 2006c). The effect is in fact double as they inhibit tyrosinase but also act as cytotoxic agent as they are oxidized to their quinonic forms. Recent studies reported that topical application of 4-n-butylresorcinol 0.3% is capable of improving melasma (Researching Committee of Rucinol, 1999) and decreasing post-inflammatory hyperpigmentation following laser therapy.

Arbutin (Figure 1A1), a natural β-glycoside of hydroquinone, is commonly used. As hydroquinone, it is a good inhibitor of tyrosinase, although it does not affect its expression and synthesis of tyrosinase in human melanocyte cultures (Maeda and Fukuda, 1996). Arbutin shows a good photostability, although its decomposition is four-times higher at basic pH (approximately 9) than at acid pH (approximately 5; Couteau and Coiffard, 2000). These data were important to insure that arbutin preparations are not enriched in hydroquinone during storage before their distribution and use. Arbutin inhibits not only tyrosinase but also melanosome maturation, possibly by its reported actions on DHICA polymerase activity and the silver protein in this organelle (Chakraborty et al., 1998). Its mild effect is due to the controlled release of hydroquinone by the in-vivo hydrolysis of the glycosidic bond. To increase its efficiency, α-glucosides of arbutin have been chemically synthesized (Sugimoto et al., 2003), possibly since they are easier hydrolyzed to release hydroquinone due the higher availability of α-glycosidases. Despite the safety of arbutin as an agent to lighten skin and ameliorate hyperpigmented lesions, some reports do not confirm its effect in clinical trials (Curto et al., 1999).

Recently deoxyarbutin, synthesized by removing every hydroxyl group of arbutin, has been identified as an excellent tyrosinase inhibitor in the screening of a number of candidate compounds (Boissy et al., 2005). When applied topically to a guinea pig model, deoxyarbutin demonstrated a more sustained depigmenting effect than hydroquinone. That action of deoxyarbutin can be explained by its chemical structure: deoxysugars are characterized by a significant increase of both skin penetration ability and binding affinity for tyrosinase. Moreover, deoxyarbutin completely lacks skin irritation observed in guinea pigs following hydroquinone application, and its skin lightening effect is reversible, suggesting the absence of permanent destruction of melanocytes. In a human clinical double-blind trial topical treatment with 3% deoxyarbutin for 12 weeks provided a significant effect on overall skin lightening and moderate resolution of solar lentigenes in Caucasian subjects, suggesting that further development are needed for using deoxyarbutin as a skin lightening agent.

Kojic acid (Figure 1A2) is a good tyrosinase inhibitor, but it has some undesirable side effects. It may cause allergy (Nakagawa and Kawai, 1995) and it also has been related so some hepatic tumors in heterozygous mice deficient in p53 (Takizawa et al., 2003). Although in monotherapy kojic acid showed only a modest effectiveness, clinical trials have reported a skin lightening effect of this compound in combination with other agents (Lim, 1999). Recently, some stable derivatives have been synthesized having more efficiency because their penetration through the skin is increased. In this group, the most important ones are a derivative synthesized by joining two pyrone rings through and ethylene linkage (Lee et al., 2006) and kojyl-APPA (5-[3-aminopropyl)-phosphino-oxy]-2-(hydroxymethyl)-4H-1-pyran-4-on), tested in melanoma cells and normal human melanocytes (Kim et al., 2003).

Gentisic acid (2,5-dihydrobenzoic acid, Figure 1A3) has also been explored for use in topical cutaneous applications. It is found in gentian roots and it seemed a good inhibitor of melanogenesis (Dooley et al., 1994). Its alkyl esters have been tested as tyrosinase inhibitors in-vitro and in-cell cultures (Curto et al., 1999). Methyl gentisate appeared to be more efficient than the free acid as well as other well-known hypopigmented agents, such as hydroquinone, kojic acid, arbutin, and magnesium ascorbyl phosphate. The methyl gentisate was also less cytotoxic and mutagenic than hydroquinone, although some low adverse side-effects were reported. Thus, this compound is proposed as a good candidate for skin-lightening skin, but as far as we know, it has not been tested in skin models.

The families of flavonoids-like agents

Flavonoids belong to the best studied group of plant polyphenols. More than 4000 members have been identified widely distributed in leaves, bark and flowers. All have phenolic and pyrane rings, so that they are benzo-γ-pyrane derivatives. They are classified into six major groups, flavanols, flavones, flavonols, flavanones, isoflavones and anthocyanidins, which differ in the conjugation of rings and the position of hydroxyl, methoxy and glycosidic groups (Kim and Uyama, 2005).

They are used in numerous natural medical treatments because a number of beneficial effects such as anticancer, anti-inflammatory, protection against UV etc. Some of these effects are based on the modulation of the immune system, but others are related to the antioxidant and ROS scavenger capacities (Yu et al., 2005) and the possibility to chelate metals at the active site of metalloenzymes. They can also have hypopigmenting effects, as they can directly inhibit tyrosinase and also can act on the distal part of the melanogenesis oxidative pathway. The main groups with hypopigmenting properties are as follows.

Hydroxystilbene derivatives. This group appears as one of the most efficient because of its high affinity for tyrosinase (Kim et al., 2002b). It includes resveratrol and other isomers, such as oxyresveratrol, gnetol (Figure 1B) and methoxylated or glucosylated derivatives (piceid-glucoside, rhapontigenin and rhaponticin). Oxyresveratrol and gnetol are more efficient tyrosinase inhibitors than resveratrol (Ohguchi et al., 2003a). The hydroxyl groups are important for inhibition, but the trans-olefin structure of the stilbene skeleton as well, as trans-resveratrol is much more potent than the cis isomer (Ohguchi et al., 2003b). In all cases, the inhibition of the enzyme is reversible. Thus, in-vivo treatments should maintain high intracellular levels of the hydroxylated stilbene inside melanocytes, one of the main problems of depigmentation treatments based on tyrosinase inhibitors. Some data (Lin et al., 2002) indicated that resveratrol is not simply a tyrosinase inhibitor, but also can reduce Mitf and tyrosinase promoter activation in B16 mouse melanoma cells, so promising a more effective effect. However, these data were not in accordance with those by Kim et al. (2002b), who reported resveratrol just inhibited tyrosinase activity as they did not observe any decrease in the amount of tyrosinase. Anyway, the effect on Mitf reported by Lin et al., 2002, as such, would reinforce the direct inhibition of melanin synthesis by resveratrol.

In a recent study with 285 different plant extracts from oriental herbs, the best one seems to be the extract from Morus alba L., which contains 2-oxyresveratrol. The extract inhibited tyrosinase activity remarkably, showed no toxicity and successfully passed a number of tests including the most sensitive ones such as eye and human skin irritation (Lee et al., 2003).

Aside natural extracts, a number of chemically resveratrol-derived compounds are also good tyrosinase inhibitors, and their hypopigmenting effect is still under investigation. Choi et al. (2002a,b) reported a more potent nitrogenated derivative (4-methoxy-benzyliden)-(3-methoxy-phenyl)-amine. In the group of resveratrol structure analogs, 4,4′-dihydroxybiphenyl (with a shorter arm linkage between the two aromatic rings) has been described as a new potent tyrosinase inhibitor in B16 mouse melanoma cells (Kim et al., 2005d). Furthermore, 4,4′-dihydroxybiphenyl has been found to exert its antimelanogenic activity through other mechanisms including the stimulation of glutathione synthesis, downregulation of the cAMP-dependent PKA signalling pathway and decreasing Mift gene expression (No et al., 2006).

The more complex but related compound 2,2′-dihydroxy-5,5′-dipropyl-biphenyl is not a direct tyrosinase inhibitor, but it is a very powerful in-vivo downregulator of melanogenesis by inhibiting tyrosinase maturation and accelerating tyrosinase degradation (Nakamura et al., 2003). Finally, and returning to natural compounds with structure resembling that of resveratrol, rosmarinic acid, rooperol and its glycosyl-derivatives could be also effective as tyrosinase inhibitors, but its effect on tyrosinase activity has not been explored only by preliminary studies (Solano et al., 2005; Theron et al., 1994).

Hydroxyflavanols conjugated to gallic acid. These compounds are isolated from green tea leaves. The most abundant are four: ECG [(-)EpiCatechin-3-O-Gallate], GCG [(-)GalloCatechin-3-O-Gallate], EGCG [(-)EpiGalloCatechin-3-O-Gallate] and EGC [(-)EpiGalloCatechin] (Figure 1C). These conjugates are much more effective than isolated gallate and catechin. The skeleton of flavon-3-ol with the bonded gallate moiety seems to be essential for the antioxidant and hypopigmenting effects, as well as to act as a free radical scavenger (No et al., 1999). GCG is the most efficient to inhibit plant catechol oxidases, but the effect on mammalian tyrosinase is still largely unknown. EGCG has also been recently described not only as a hypopigmenting but also as an antiproliferative and proapoptotic agent for human melanoma cells (Nihal et al., 2005). In addition, it has been reported that EGCG and hinokitiol (structurally not related to hydroxyflavanols) are not only tyrosinase inhibitors, but also agents that decreased Mitf production with synergistic effects on melanogenesis tested in cell-based systems (Kim et al., 2004b).

One interesting approach consisted in the formation of esters between the gallic acid and long-chain fatty acids. The fatty acid moiety improves the antioxidant capacity of the gallate ring (Kubo et al., 2000) and it adds some hypopigmenting effects per se. When the chain is longer than C10, these compounds are tyrosinase inhibitors but yet not substrates of the enzyme.

Related to catechin and gallic acid, other natural compounds with effects on pigmentation are under investigation. First, procyanidins which are polymers of catechins found in tea and fruits such as apple and grape. They are good-antioxidants and have been recently introduced as inhibitors of melanogenesis in B16 mouse melanoma cells (Shoji et al., 2005) and as agents that reduce cell viability in some melanoma cells by arresting cells in S-phase (Lozano et al., 2005). Oral administration of proanthocyanidin-rich grape seed extract has been found to be effective in lightening the UV-induced of guinea pigs skin probably through its capability of inhibiting both tyrosinase activity and oxidative stress-mediated proliferation of melanocytes (Yamakoshi et al., 2003). The efficacy of pyogenol, a standardized extract of the pine bark containing catechin, epicatechin and procyanidins, in the treatment of melasma has been successfully tested (Ni et al., 2002).

Secondly, elaters, which are molecules associated with pollen grains. The most important is the ellagic acid which is a tyrosinase inhibitor described as a preventive agent of UV stimulation of melanocytes (Shimogaki et al., 2000). In brownish guinea pigs, ellagic acid suppresses melanosynthesis without injuring melanocytes, indicating its possible utility as a safe hypopigmenting agent in cosmetic. Oral administration of Punica granatum L. extract, containing about 90% ellagic acid, effectively whitened the pigmented skin of guinea pigs (Yoshimura et al., 2005).

Flavonols. These are good inhibitors of mushroom tyrosinase (Figure 1D). The hydroxyl group at C3 and the adjacent keto group at position C4 of the flavonol ring seem to be crucial as they bind the copper ions at the tyrosinase active site. The importance of the hydroxyl group at C3 position of flavones is illustrated by apigenine. Apigenine is an anti-inflammatory agent also named 4′,5,7-trihydroxyflavone [5,7-dihydroxy-2-(4-hydroxyphenyl)-4-H-1-benzopyran-4-one], with no hydroxyl group at that position and a relatively weak hypopigmenting effect. Its analogues without hydroxyl at position C3, glycitein, daidzein and genistein, are also poor tyrosinase inhibitors.

However, the isomer 6,7,4′-trihydroxyisoflavone is a potent tyrosinase inhibitor (Chang et al., 2005), six times more than kojic acid for mushroom tyrosinase without a hydroxyl group at C3 position. To account for this strong effect, it is claimed that the hydroxyl groups at C6 and C7 positions of the isoflavone skeleton are also important for the interaction with the copper at the enzyme active site. A recent fluorescence quenching study demonstrated that the dihydroxy substitutions both in the A and B rings of flavonoids are crucial for the tyrosinase inhibitory activity (Kim et al., 2006a).

Flavonols are competitive inhibitors of mushroom tyrosinase, most of them acting at about 1 mM (Chen and Kubo, 2002). According to the same authors, quercetin (3,3′,4′,5,7-pentahydroxyflavone) is more effective than its analogues kaempferol and morin. Quercetin is present as glycosylated derivatives in onions, flowers of mexican Heteroteca inuloides and Trixis michuacana and Arnica montana. Kaempferol is found in petals of Crocus sativus (saffron). Its 3-O-glucoside does not inhibit tyrosinase (Kubo and Kinst-Hori, 1999). Mulberroside F (moracine M-6) is obtained from the Morus alba leaves, and it greatly decreases melanin formation in normal melanocytes (Lee et al., 2002a). Studies using cultured melanocytes or animal skin are very rare.

Ethanolic extracts of Myrica rubra dried leaves have shown good depigmenting effects in-vitro and pseudo SOD activity (Matsuda et al., 1995). They contain quercetin, myricetine and some 3-O-ramnosides derivatives. In-vivo studies using these pure compounds should be made to evaluate their actual capacity as potential depigmenting agents. Methanolic extracts of Artocarpus incisus (Papua Nueva Guinea) contain hypopigmenting agents of flavone-related structure having similar effect to kojic acid. At least seven compounds were isolated, mainly isoartocarpesin [6-(3′′-methyl-1′′-butenyl)-5,7,2′,4′-tetrahydroxyflavone], and the others dihydromorin, chlorophorin, norartocarpenon, 4-prenyloxyresveratrol, artocarbene and artocarpesin (Shimizu et al., 1998). They have been assayed on the back skin of guinea pigs without any observable irritation. Promising results have been also obtained with extracts from other plant species such as Arctostaphylos (A. patula and A. viscida; Matsuda et al., 1996) and leaves of other Japanese plants, mainly the genus Prunus, P. zippelianay P. yedoensis (Matsuda et al., 1994). These extracts seem to contain not only flavonoids but also arbutin.

Flavanones. These, also named chalcones, differ from other flavonoids by the absence of the double bond at the 2–3 position of the pyrone ring. A systematic study about the influence of number and position of hydroxyl groups on the basic ring indicated the importance of the position 4 on the ring B (Nerya et al., 2004). Isoliquiritigenin chalcone and butein were the most potent tyrosinase inhibitors, although curiously also shortened the typical lag period known for the monophenolase activity of this enzyme. More recent studies (Khatib et al., 2005) indicate that the 2,4-resorcinol subunit on ring B is very important for inhibition, as changing to resorcinol substitution to position 3,5 or placing it on ring A significantly abrogates the inhibitory effect of the compounds. Nevertheless, all these studies were made on mushroom tyrosinase, and once more cell-based systems are needed to evaluate the potential hypopigmenting effects of these agents.

Hydroxycoumarins. Coumarins are lactones of phenylpropanoid acid with an H-benzopyranone nucleus. Some hydroxylated derivatives seem to be efficient to inhibit melanogenesis. These compounds also have a direct interaction with tyrosinase, although this is not the only point of affecting melanogenesis. The best one is the 7-allyl-6-hydroxy-4,4,5,8-tetramethyl-hydrocumarin (Figure 1E, Yamamura et al., 2002). It is able to stimulate glutathione synthesis and pheomelanogenesis, so that those authors proposed a combined treatment of hydrocoumarins and α-tocopherols to enhance the hypopigmenting effect by ROS scavenging mechanism. 4-Phenylcoumarins, also named neoflavonoids, and the related aurones (Okombi et al., 2006a) have been tested as tyrosinase inhibitors with moderate effects.

Aloesin. This (2-acetonyl-8-β-d-glucopyranosyl-7-hydroxy-5-methylchromone, Figure 1F) is a C-glycosylated chromone isolated from aloe plant (Piao et al., 2002). Its structure is very similar to flavonols, but the third fused ring to benzopyrane makes the affinity for the tyrosinase active site different from flavonols. Studies with aloesin and a number of chemically related chromones have pointed out that the inhibitory effect on tyrosinase is stronger than the one of arbutin and kojic acid, specially the 5-methyl-7-methoxy-2-(2′-benzyl-3′-oxobutyl)-chromone.

Further studies using mammalian tyrosinase have proven that aloesin is a competitive inhibitor of tyrosinase and it inhibits melanin production in cultured normal melanocytes (Jones et al., 2002). Aloesin penetrates the skin slowly, rendering it a promising hypopigmenting agent for cosmetic applications. Recent studies with aloe extracts indicate that it contains some compounds (isorabaichromone, feruloylaloesine and p-cumaroylaloesine) with potent antioxidant activity in cellular organelles, including mitochondria, melanosomes and microsomes. They inhibit cyclo-oxygenase 2 and TxA2 synthase. In these effects, the involvement of the caffeoyl, feruloyl and coumaroyl groups bound to the aloesin skeleton has been clearly proven. This may account for the wound healing effects of the A. vera extracts (Yagi et al., 2002). Finally, the combined treatment of aloesin and arbutin seems to show synergistic effects by respectively non-competitive and competitive tyrosinase inhibition, remaining a very low residual enzymatic activity (Choi et al., 2002a).

Prenylated flavonoids. Kuraninone (Figure 1G) is a prenylated flavonoid containing a resorcinol moiety isolated from the roots of Sophora flavecens (Son et al., 2003). According to these studies, this is a strong mushroom tyrosinase inhibitor, better than kojic acid. The lavandulyl group at the C8 position and the methoxy group at the C5 position are essential for its inhibitory effect. Sanggenon D and sophoflavescenol show considerable inhibitory activity of the mushroom enzyme, but studies on mammalian tyrosinase and melanocyte cultures are still needed. However, other studies exploring prenylated flavonoids failed to enhance tyrosinase inhibition due to the prenyl moiety (Lee et al., 2004).

Saponins and liquorice components. Liquorice is one of the most important herb in the Traditional Chinese Medicine, whose roots are widely used for the treatment of a number of diseases, demulcent, flavoring an more recently as hypopigmenting agent. The hydrophilic extracts of liquorice (Glycyrrhiza glabra Linn. and Glycyrrhiza uralensis) contain high amount (ranging: 2–14%) of saponins, mainly the anti-inflammatory glycyrrhizin frequently used to treat arthritis and rheumatism. Glycyrrhizin is composed of one glycyrrhetinic acid and two glucuronic acids (Figure 1H). Besides, these extracts contain liquiritin, a glucoside-derived flavonoid efficient for prolonged treatments of epidermic melasma (Amer and Metwalli, 2000), and isoliquiritin. The direct effect of these compounds on tyrosinase activity is not proven yet, but it has been proposed that the flavonoid ring accelerates epidermis turnover and the subsequent melanin dispersion.

Concerning the hydrophobic fraction of liquorice, the glabridine extracted of roots with ethyl acetate are in-vitro good tyrosinase inhibitors. Glabridin inhibits UVB-induced skin pigmentation (Yokota et al., 1998), although no evaluation of its efficacy as a depigmenting agent has been carried out. Recently a number of other constituents have been also identified, such as glycyrrhisoflavone and glyasperin C obtained from G. uralensis (Kim et al., 2005c) or licuraside and licochalcone A (Fu et al., 2005). Among the derivatives, the 2′-O-ethyl-glabridine is more efficient than 4′-ethyl. When both groups are ethylated, the effect is totally lost (Yokota et al., 1998). These agents are also antioxidants and anti-inflammatory able to inhibit superoxide anion production and cyclo-oxygenase activity. They do not affect cellular proliferation, and the suppression of melanogenesis in cell-based systems is moderate.

Chinese herbal medicines cocktail, which contains bletilla striata, rhizome of chuaxiong, glycyrrizae radix, pericarpium citri reticulatae and schisandrae fructus, inhibits melanin formation on the basis of its antioxidant effect. Extracted ingredients from bletilla striata showed better hypopigmenting activity than arbutin, and similar to kojic acid (Lin, 2005). The same whitening effect has been reported with other natural extracts containing antioxidants and some unidentified tyrosinase inhibitors, such as the ethanolic extract from the tropical asian palm Areca catechu L. (Lee and Choi, 1999) and gypenosides from Gynostemma pentaphylum (Solano et al., 2005).

Diarylheptanoids

Diarylheptanoids are found in Alnus hirsuta Turcz bark and others similar plants such as Agnus japonica and Myrica rubra. Oregonin (1,7-bis(3,4-dihydroxyphenyl)-heptane-3-one-5-O-β-d-xylopyranoside, Figure 1I), hirsutanonol and other diarylheptanoids inhibit tyrosinase activity and lower melanin in about 75% in B16 mouse melanoma cells, and 13–43% in MSH or forskolin stimulated cells (Cho et al., 2002). These agents are competitive inhibitors due to their 3,4-diphenol groups, although the role of the heptane chain has not been evaluated (see effect of azelaic acid later on). In turn, oregonin shows an immunomodulatory effect, specially increasing macrophage activity (Joo et al., 2002). This may have some beneficial antitumoral properties, but also it can facilitate the removal of melanin.

Miscellaneous: other possible agents with hypopigmenting effect

In addition to all agents described above, there are still a large variety of phenols, methoxyphenols, thiophenols and sulfur-containing agents that have been reported as tyrosinase inhibitors and as somehow toxic agents for melanocyte growth. Some of them have also been proposed as antimelanoma drugs, and their use as hypopigmenting agents have been reserved for particular prescriptions of physicians and dermatologists. To this group belongs 4-hydroxyanisole (Naish and Riley, 1989; Riley, 1969), N-acetyl-4-S-cysteaminylphenol (Inoue et al., 1995; Jimbow, 1991), N-2,4-acetoxyphenyl thioethyl acetamide (Jimbow et al., 1995) or N-hydroxycinnamoylphenalkylamides (Okombi et al., 2006b) to mention some illustrative examples.

Ni-Komatsu et al. (2005) have identified a new family of four novel tyrosinase inhibitors by screening a tagged-triazine library using a triethyleneglicol linker at one of the triazine scaffold site. The chemical genetic screening was performed in cultured melan-a cells. These inhibitors offer several advantages. First, the method of screening, as the triazine scaffold provides flexibility in the modification of the library to get more agents. Secondly, triazine scaffold shows structural similarity to purine and pyrimidine, giving place to a new family on non-phenolic inhibitors. Thirdly, the affinity of these agents to tyrosinase is quite high, with IC50 in the 10 micromolar range, which is better than hydroquinone or kojic acid. Finally, they are not toxic and the treated cells do not show morphological changes. Although they do not affect tyrosinase and TRPs activities or processing, dismissing possible multiple effects unless they are mixed with other agents, they are promising novel agents for the treatment of hypopigmentation.

Besides these agents, H2O2 might be also effective because it is able to carry out the chemical degradation of melanin in vitro and a direct bleaching effect on this pigment. H2O2 causes a temporal repression of Mitf (Jiménez-Cervantes et al., 2001). A comparison of normal melanocytes with melanocytes from vitiligo patients shows lower catalase activity and higher vitamin E and ubiquinone levels in the last ones (Maresca et al., 1997), suggesting that an imbalance in the antioxidant status and increase in the intracellular peroxide levels could be involved in vitiligo. Anti-oxidants like green tea extracts and quercetin have been recently proposed against hydrogen peroxide-induced oxidative stress (Jeong et al., 2005). However, the use of H2O2 in the treatment of hyperpigmented skin is difficult to assess and its final effect on the late steps of melanogenesis uncertain. First of all, it is difficult to reach concentrations of hydrogen peroxide high enough to degrade melanin by oxidative attack, and low concentrations of H2O2 are generated in situ during melanogenesis (Mastore et al., 2005) and other photochemical reactions (Elleder and Borovansky, 2001) without any appreciable bleaching effect. Secondly, H2O2 induces apoptosis, but the mechanisms are so complex that this stress is considered both oxidative and reductive at the same time (Pervaiz and Clement, 2002). Thirdly, it is not possible to mix H2O2 with antioxidants in combined treatments. Thus, its use in practice has some severe disadvantages.

Ascorbic acid is used as an antioxidant because its capacity to reduce back o-dopaquinone to dopa, thus avoiding melanin formation. However, ascorbic acid has other adverse affects as it can induce a large increase of free radicals with traces of metal ions by the Fenton reaction. To improve its stability, skin absorption and hypopigmenting effect, some ascorbate esters, such as the magnesium ascorbyl-2-phosphate, has been synthesized (Kameyama et al., 1996). Daily application of a cream containing 10% of this ascorbic acid derivative has been found to produce a significant lightening effect in patients with melasma. Ascorbic acid iontophoresis has been demonstrated as an effective depigmenting strategy in a randomized, double-blind, placebo-controlled trial (Huh et al., 2003).

Some drugs such as the antihypertensive captopril (Espin and Wichers, 2001), the antithyroid methimazole (Andrawis and Khan, 1996), the antifungal miconazole (Mun et al., 2004) and the fungal metabolite terrein (Park et al., 2004) have been reported as tyrosinase inhibitors. The first two drugs were described as inhibitors of mushroom tyrosinase, but their effects on mammalian tyrosinase are still not directly tested. Both molecules have relatively small molecular weight and contain thiol groups, making very likely the interaction with copper ions at the tyrosinase active site and also the trapping of o-dopaquinone and other melanin precursors to form colorless conjugates (Olander et al., 1983). Methimazole has also some effects on melanocyte morphology (Kasraee, 2002). Miconazole reduces tyrosinase expression in B16 cells, suppressing the effect of αMSH or forskolin.

Finally, some agents act on non-melanocortin hormonal receptors. Yohimbine (Figure 1J) was described as a good hypopigmenting agent. It is obtained from yohimbe bark, an African tree with aphrodisian effects. 12–48 h exposure of melanocytes to 0.1 mM decreases tyrosinase levels 30–80% (Fuller et al., 2000). The underlying mechanism is unknown, but a direct effect at the tyrosinase active site is ruled out as cell-based systems are required to obtain hypopigmentaion. Although this compound is under clinical investigation as a safe hypopigmenting agent, collateral effects should be considered before its use.

Histamine increases the levels of tyrosinase mediated cAMP via PKA. The release of histamine has been related with the stimulation of melanogenesis provoked by UV light, especially eumelanogenesis (Lassalle et al., 2003), whereas cAMP and NO stimulate both, eu- and pheomelanogenesis. Thus, a possible approach to minimize skin pigmentation might be to block histamine action. The antagonist of H2 receptors famotidine blocks the action, but antagonists of H1 and H3 do not have any effect (Yoshida et al., 2000). The agonists dimaprit and nordimaprit enhance tyrosinase degradation (McEwan and Parson, 1991). Nordimaprit derivatives with bulky groups stimulate tyrosinase activity (Fechner et al., 1993). So, the effect of histamine-related agents is difficult to predict and their use to treat hypo- and hyperpigmentation is so far unreliable. In turn, the use of all these compounds is doubtful considering their side effects in the control of inflammation.

Tyrosinase inhibition without interaction with its active site

Although it is correct to assume that inhibition of tyrosinase activity by phenols mimicking the substrates and other copper- chelating agents is the most effective mean to decrease skin pigmentation, other approaches exist to design more effective hypopigmenting strategies. Some of them are also related to tyrosinase, but focused on the alteration of the tyrosinase processing, tyrosinase gene expression or on increasing the rate of tyrosinase degradation. Some of the agents affecting tyrosinase expression are also tyrosinase inhibitors, such as resveratrol, and those effects have been briefly mentioned in the former section. In other cases, downregulation of tyrosinase expression without a direct tyrosinase inhibition by interaction with the active site has been reported for a number of non-physiological agents, such as the fungal metabolite terrein, which promotes a sustained ERK and subsequently Mift downregulation (Park et al., 2004), for the alkaloid piperlonguminine (Kim et al., 2006b) and for 3,4-dihydroxyacetophenone (Kim et al., 2006c).

To inhibit the processing and maturation of tyrosinase, the most direct way to act at this level is disturbing the N-glycosylation system in the Golgi. d-pantetein-S-sulfonate has been used for this purpose in human normal and malignant melanocytes. Tunicamycin and other antibiotics also act on this process. However, the inhibition of N-glycosylation is quite unspecific and associated with a number of side effects rendering their use unsuitable as hypopigmenting agents.

2,2′-dihydroxy-5,5′-dipropyl-biphenyl has also been reported as a valuable putative whitening agent because of its double effect, inhibition of maturation and acceleration of tyrosinase degradation (Nakamura et al., 2003). Inulavosine, isolated from Inula nervosa roots, it also seems to promote lysosomal tyrosinase degradation (Motokawa, 2003). On contrary, glycosphingolipids are required for sorting melanosomal proteins in the Golgi complex (Sprong et al., 2001) so those drugs that affect the formation of these compounds block the protein transport of tyrosinase to melanosomes and indices a strong loss of pigmentation. Similarly, several steroid drugs, such as progesterone and U18666A (an androsterone derivative) provoke tyrosinase accumulation in an artefact compartment related to late endosomes/lysosomes. This leads to a strong decrease of melanogenesis (Hall et al., 2003). The use of these drugs, isolated or combined with other agents is still unexplored.

Lipophilic compounds with antioxidant properties

Besides tyrosinase activity, expression and stability, other alternative approaches use antioxidants to inhibit the chemical reactions leading to melanin formation, change the type of melanin formed or interfere with the distribution of pigment and melanosome transfer. The main group of agents used as hypopigmenting-aimed treatments are lipophilic compounds with antioxidant properties because of the involvement of quinones and oxidative reactions in the polymerization of melanogenic intermediates to the final melanin product. Brief comments on these agents, their properties and mechanism of action as follows:

Lipoic (also named thiooctic or 6,8-dithiol-octanoic acid) and its reduced form dihydrolipoic acid seem to block the expression of Mitf (Lin et al., 2002). Importantly, lipoic acid also seems to inhibit the activation of NF-κB transcription factor (Saliou et al., 1999). So, its effect is not only restricted to its antioxidant properties and even dopaquinone trapping (Tsuji-Naito et al., 2006), but also extends to the modulation of melanogenic enzyme expression, making lipoic acid one of the main candidates able to decrease Mitf levels and subsequently those of melanogenic enzymes. Although no clinical trials have proved its effectiveness, lipoic acid is included in commercially available topical formulations.

All-trans-retinoic acid has a complex behavior although it is used as hypopigmenting agent. In fact, some antioxidants used in cosmetics show weak effects on pigmentation, and the reported data are sometimes contradictory depending on the conditions of assay. This agent is able to activate Mitf expression through PKC and stimulate melanocyte differentiation (Lin et al., 2002). However, it also reduces the number of cells since it triggers apoptosis. In summary, it might increase pigmentation when the levels of melanin are very low or melanocytes are scarce and inactive, but it has opposite effects in melanocytes previously exposed to UV.

Azelaic acid (non-adienoic acid) was originally described in the Istituto Dermatologico San Gallicano as a hypopigmenting agent (Nazzaro-Porro et al., 1979). Other dicarboxylic acids between C8–C14 had also weak hypopigmenting effects. This was first justified on the basis of a competitive inhibition of tyrosinase (Breathnach et al., 1979), but later was attributed to their ROS scavenger capacity and alteration of the mitochondrial metabolism (Passi et al., 1991a,b). The last mechanism is more probable than a direct inhibition of tyrosinase activity, although there are still some papers considering fatty acids as blocking agents at the tyrosinase active site. In-vitro studies on mushroom and mammalian tyrosinase do not clearly support this hypothesis.

Azelaic acid has been reported to be effective in hypermelanosis caused by physical and chemical agents, as well as others skin disorders characterized by abnormal proliferation of melanocytes. Topical azelaic acid (15–20%) is efficacious in treating melasma and post-inflammatory hyperpigmentation (Breathnach, 1996). Rigoni et al. (1989) reported a mean reduction in melasma intensity of 51% in 37 patients treated (two applications/day) with 20% azelaic acid base cream for 6 months. The best clinical results were obtained in a randomized, double-blind comparative study in dark skinned women, demonstrating that azelaic acid 20% is more effective than hydroquinone 2% and equivalent in efficacy with 4% hydroquinone in lightening of lesions and concomitant reduction of lesion size (Verallo-Rowell et al., 1989). In other double-blind study involving 329 women, 20% azelaic acid was shown to be as effective as hydroquinone 4%, without the latter's undesirable side effects (Balina and Graupe, 1991). Azelaic acid may arrest the progression of lentigo maligna to melanoma and even induce its disappearance (Nazzaro-Porro et al., 1989). It has no relevant side effects and is not able to induce depigmentation on normally pigmented skin, freckles, senile lentigines and nevi, suggesting its selective antiproliferative and cytotoxic action on abnormal melanocytes. The only problem of melasma treatment with azelaic acid is that its therapeutical response is rather slow. In order to improve this aspect combined therapies with other agents, particularly compounds capable of accelerating desquamation have been evaluated. Azelaic acid 20% plus tretinoin 0.05 or 0.1% has been shown to be more effective than azelaic acid alone. In an open-label randomized study of 50 patients, 24 weeks of treatment with azelaic acid 20% and azelaic acid 20% plus tretinoin 0.05% resulted in good results in 5.3% and 34.8% of patients respectively (Zaumseil and Graupe, 1995). Sarkar et al. (2002) have also studied sequential therapy of the topical steroid clobetasol propionate and azelaic acid. Ninety-seven percent and 90% of patients had excellent responses with azelaic acid plus steroid and azelaic acid, respectively.

A chemically synthesized fatty acid, trans-4-aminoethylcyclohexanecarboxylic acid (trans-AMCHA) suppresses, at least partially, the increase in melanogenesis caused by arachidonic acid and UV light exposure (Maeda and Naganuma, 1998). These effects could be mediated by the inhibition of prostaglandin synthesis as PGs seem to stimulate melanin formation (Dutkiewicz et al., 2000). Once more all reports do not agree on this point and PGs are also found to inhibit melanogenesis (Kashiwagi et al., 2002). Discrepancies are probably due to complex mechanisms of PGs action.

An important problem: some tyrosinase inhibitors and antioxidants stimulate melanin formation

In spite of the widely described reports using phenolic antioxidants and free radicals scavengers as hypopigmenting agents, an increasing number of recent data indicate that the indiscriminated use of these compounds should be considered with caution. An opposite effect seems to be possible in some cases, and unexpected hyperpigmentation can occur. The complete understanding of the different mechanisms of action of all antioxidants may help identifying of true hypopigmenting agents for its use in cosmetics and dermatology.

Although in-vitro quercetin is a tyrosinase inhibitor (Kubo and Kinst-Hori, 1999), this flavone has been described as a strong inductor of melanogenesis in normal and malignant human melanocytes (Nagata et al., 2004) and in reconstituted three-dimensional human epidermis models (Takeyama et al., 2004). This stimulation was blocked by actinomycin D and cycloheximide, indicating that the effect involves both transcriptional and translational processes. In epidermal models, accumulation of mature melanosomes (stages III and IV) and extension of the melanocytic dendrites to adjacent keratinocytes were visible in the basal layer of the epidermis after 7 days treatment with quercetin, pointing out that flavone acts as stimulator at different steps of melanogenesis, and the direct in-vitro inhibition of tyrosinase or its antioxidant effect are moderate in comparison with the whole visible in-vivo effects. Quercetin has been also described as an unspecific kinase inhibitor (McGovern and Stoichet, 2003), but its role in the substantial promotion of melanogenesis is still unknown. Finally, some recent reports indicate that flavonoids increase free radicals instead of scavenging them (Matsuo et al., 2005), so that flavonoids could be cytotoxic for normal human cells and their related adverse side-effects are hindering its use in cosmetic applications.

Like quercetin, there are other recent reports of putative hypopigmenting agents with totally opposite effects. Thus, glycyrrhizin stimulates melanogenesis in B16 melanoma cells (Yung et al., 2001). The effect is likely to occur at the transcriptional level, as suggested by mRNA increases in a dose-dependent manner reported in this study. Expression of TRP2 mRNA is also increased, but no significant change was observed with TRP1. Interestingly, similar concentrations of glycyrrhetinic acid are not effective, indicating that glycoside moiety of glycyrrhizin is essential for the stimulatory effect on melanogenesis. In addition, several coumarins of seven umberiferae plants also exhibited potent melanogenesis stimulation with significant enhancement of cell proliferation in a dose-dependent manner (Matsuda et al., 2005).

Finally, some treatments based on an antioxidant effect, such as the Chinese traditional medicine-1 (aqueous extracts of the plant Astragalus membraneus Bunge) activates Mitf and therefore pigmentation (Lin et al., 2002). Similarly, Epimedium koreanum is an important oriental herbal medicine. Its aerial part is used to stimulate and potentiate some hormones, to cure impotence and loss of memory, whereas the roots are used for the treatment of asthma and menstrual disorders. Its main effective antioxidant ingredient, ikarisoside A, stimulates melanin formation in a dose-dependent manner in B16 mouse malignant melanocytes (Chun et al., 2001).

Stimulation of the desquamation

Desquamation accelerates the loss of melanin by the peeling of the stratum corneum cells. To this regard, the most important agents are retinoids, but there are other agents used to this purpose.

Retinoids

Retinoic acid was first described as an agent that produces a derangement of the melanosomal structure and an arrest of the melanosome maturation at stages I and II (Orlow et al., 1990). In general, retinoids are hypopigmenting agents as they induce dispersion of keratinocyte pigment granules, interference with pigment transfer, acceleration of epidermal turnover, reduction cohesiveness of corneocytes and induction of desquamation (Nair et al., 1993). Topical application tretinoin (all-trans-retinoic acid) has been reported to improve melasma and other hyperpigmentary disorders in a high percentage of patients with some side effects, such as erythema and peeling in the area of application, as well as post-inflammatory hyperpigmentation. Tretinoin 0.1% monotherapy reduced the average MASI score by 32% in African-American patients with melasma (Kimbrough-Green et al., 1994). Another similar study in Caucasian women indicated that 68% of tretinoin-treated patients were clinically rated as improved or much improved, compared with 5% in the vehicle group (Griffiths et al., 1993). However, tretinoin is slow to work (24 weeks or more). Tarazotene, an acetylenic topical retinoid, improves the irregular hyperpigmentation associated with photoaging and lightening of the pigmented spots (Sefton et al., 2000). Isotretinoin 0.05% showed no significant positive results in 30 Thai melasma patients (Leenutaphong et al., 1999). The efficacy and the safety of β-carotene topical application in melasma were evaluated by Kar (2002). β-carotene in nanothalospheres, which are special vectors able to deliver intact β-catotene into the intracellular space of melanocytes, showed good skin lightening action (50–74% improvement) with minimal side effects in 38.5–55.1% of treated patients.

Adapalene is a naphthoic acid derivative with potent retinoid activity; it controls cell proliferation and differentiation and has significant anti-inflammatory action. In a randomized clinical trial, the efficacy of adapalene gel 0.1% was found to be comparable to that of tretinoin 0.05% cream in the treatment of melasma and solar lentigines, but patients using adapalene showed fewer side-effects and greater acceptability (Dogra et al., 2002).

Chemical peels

The effect of pure retinoids is too slow and other depigmenting chemical agents have combined to improve efficacy. Pigmented spot, such as freckles or actinic lentigines, melasma and post-inflammatory hypermelanosis may be removed by the peeling of corneocytes and epidermal upper layer keratinocytes. Superficial and medium-depth peels with trichloroacetic acid (TCA), α-hydroxy acids, such as lactic acid, glycolic acid and salicylic acid, and tretinoin may be used mainly in fair-skinned individuals even if some clinical trials have proved their efficacy also in dark-skinned patients. This rejuvenation action has also been attributed to fatty acids and liquiritin. To reduce the chemical irritation which can occur during the peel, glycolic, tartaric or salicylic acids have been used for in gel mixtures. A clinical trial on TCA peels reported 40% complete regression and 50% partial regression of lentigines (Cotelessa et al., 1999). Following pretreatment with topical tretinoin and localized TCA, a peel consisting of 50% salicylic acid, methyl salicylate, and croton oil proved efficacious in the treatment of solar lentigines, even after one application (Swinehart, 1992). Lactic acid supplemented with ascorbic acid determined a general skin whitening effect in subjects with medium to dark skin (Smith, 1999). Salicylic acid is effective in inducing skin desquamation in dark-skinned individuals (Grimes, 1999), possibly acting also by a non-competitive inhibition of tyrosinase (Kubo et al., 1994).

Because of the low risk of inflammatory hyperpigmentation, glycolic acid peels can be used safely in dark-skinned individuals. 1% tretinoin peel has been suggested as a well-tolerated and effective chemical peel for melasma as well as 70% glycolic acid (Khunger et al., 2004). According to Kligman's opinion, however, 1% of tretinoin could act as modulator of gene expression, leading to acceleration of epidermal cell turnover, rather than as peeling agent (Kligman, 2004). A synergistic skin-lightening effect can be obtained by the association between retinoids and chemical peels, such as 0.1% retinaldehyde and 6.4% of glycolic respectively (Kasraee et al., 2005).

Combination therapies

Skin lightening can be achieved by interfering with the different steps of melanogenesis pathway. Thus, the association of depigmenting agents with different mechanisms of action is a useful strategy to improve clinical efficacy, reducing the duration of therapy and the risk of adverse effects. Kligman's formula, which includes 5% hydroquinone, 0.1% tretinoin, and 0.1% dexamethasone, is the most used combination treatment (Kligman and Willis, 1975). Besides the synergistic therapeutical effect, dexamethasone decreases the irritative effects of hydroquinone and produces mild depigmentation, through an antimetabolic action (Menter, 2004), whereas tretinoin abrogates the epidermal atrophy that can occur with corticosteroids.

Modifications to the original formula have been studied. A stabilized triple combination of hydroquinone 4%, tretinoin 0.05% and fluocinolone acetonide 0.01% has been evaluated in two multicenter, randomized double-blind, controlled trails including patients with moderate to severe melasma and Fitzpatrick skin types I–IV (Taylor et al., 2003). After 8 weeks of treatment, a relevant clinical efficacy was observed in >77% of patients across all racial/ethnic groups. Long-term safety of the triple combination was confirmed in a 12-month extension of the previously reported 8-week trial (Torok et al., 2005). A combination treatment containing monomethyl ether of 2% hydroquinone (4-hydroxyanisole) and tretinoin 0.01% improved actinic lentigo (Fleischer et al., 2000; Ortonne et al., 2004). The addition of 2% kojic acid in a gel containing 10% glycolic acid plus 2% hydroquinone determined complete clearance of diffuse melasma in 60% percent of patients compared to 47 % of those receiving gel without kojic acid (Lim, 1999). The efficacy of a topical cream, with kojic acid, phytic acid and buthyl methoxydibenzoylmethane, in treating facial melasma in all Fitzpatrick skin types has been assessed in a double-blind controlled comparative study (Levy et al., 2005).

Sarkar et al. (2002) have demonstrated that serial glycolic acid peels can significantly improve the clinical efficacy of a modified Kligman's formula. Improvement of pigmented spots has been observed in Asian women treated with daily application of 10% glycolic acid and 2% hydroquinone in association with 70% glycolic acid peels every 3 weeks (Lim and Than, 1997).

Final remarks

Point 1: tyrosinase inhibition is still the number one key

Tyrosinase inhibition has so far been the most common approach, but only a few of these tyrosinase inhibitors have practical application for several reasons. A successful strategy should take into account not only criteria based on the efficiency of tyrosinase inhibition, but also other parameters related to cytotoxicity, solubility, effective cutaneous absorption, penetration and, stability. Concerning melanogenic enzymes, it is quite surprising to see the poor impact TRPs alteration on the pigmentation status of melanocytes. Taking into account the perspective of >15 years since that TRPs were added to the enzymatic regulation scheme of mammalian melanogenesis, these complementary enzymes to tyrosinase proved not to be key tools to control apparent pigmentation in melanocytes. Tyrosinase is the rate-limiting enzyme and the cornerstone to control pigmentation in human skin.

Point 2: new innovative combined approaches are needed

Despite the large number of molecules showing depigmenting properties in vitro, only someone of them are able to induce an effective hypopigmenting effect measurable in clinical trials. This gap between in-vitro and in-vivo studies suggests that a new strategy is needed for discovering new depigmenting agents and validating their efficacy and safety. Successful treatments need combination of two or more agents acting on different mechanisms to produce a synergistic hypopigmenting effect. These approaches should take in account of the following aspects: (i) synergistic effects of combined therapies; (ii) stability of whitening formulation; (iii) toxicity and skin penetration; and (iv) definition of markers and targets for evaluating depigmenting properties in vitro and in vivo.

As previously described, association of depigmenting molecules is able to determine a synergistic whitening effect with a reduction of adverse effects. The improvement of lightening effect of combined therapies is certainly due to the possibility to interfere with different steps of melanogenesis pathway but it is limited by the instability of compounds inside the formulation and skin penetration of the mixtures. For example, the equimolar mixture of hydroquinone and kojic acid has a synergistic effect and complexation of hydroquinone with β-cyclodextrin, leads to a further improvement of efficacy of this topical formulation (Ferioli et al., 2001). All-trans retinoic acid plus 4-hydroxyanisole produced a moderate hypopigmentation with minimal skin irritation in Yucatan swine skin with melasma or after UV-treatments whereas agents alone did not produce significant effects (Nair et al., 1993). 4,4′-dihydroxybiphenyl deserves special comments as it is not irritant, its skin penetration is good and its action is multifactorial. It causes tyrosinase inhibition, downregulation of the cAMP-dependent PKA signalling pathway, repression of Mitf gene expression, enhancement of glutathione synthesis and antioxidant (No et al., 2006). Provided all these effects are confirmed, it might be one of the most potent melanogenic inhibitors described so far.

Point 3: transdermal agent delivery as a success limiting factor

One of the main problems for achieving efficient depigmentation is to choose the system of transdermal delivery and drug formulation to ensure that useful agent concentrations will actually reach the melanocyte. For example, out of five agents used to deliver gentisic acid, one (Duro-Tak 87-2510) was the most effective, and the addition of 1% dodecylamine improved skin permeation (Bian et al., 2003).

Another example is the use of liposomes as carriers for many lipophilic drugs. Liposomes containing N-butyldeoxynojirimycin composed by cephalins and cholesterol esters have achieved the inhibition of tyrosinase maturation in endoplasmic reticulum by blocking N-glycosylation (Costin et al., 2002). Encapsulation of tyrosinase inhibitors or agents blocking different steps of melanogenesis and melanin dispersion could also be considered to improve agent delivery. Thus, α-Lipoic acid and some unsaturated fatty acids are quite appropriate, as they may decrease the expression of the melanogenic enzymes and increase the degradation of tyrosinase. As such, they can be used to prepare liposomes and emulsions. Some commercial preparations not only combine lipoic and unsaturated fatty acids, but also add antioxidants, e.g. ascorbyl esters and some organic acid to increase peeling.

Another problem for a correct evaluation of depigmenting agents is the evidence that pigmentation processes is regulated by several genes and proteins contained in melanosomes. Thus one of the reasons of failure of the many molecules characterized in vitro as good inhibitors of melanin synthesis might be the poor bioavailability and penetration into melanosomes.

Point 4: testing system

For in-vitro testing studies, the use of mammalian-derived tyrosinase should be considered much better than mushroom tyrosinase. For cell-based assays, it is obvious that mammalian skin or at least keratinocytes/melanocytes co-cultures should be preferred as testing models rather than only pure melanocyte cultures. First, the affinity of inhibitors for mammalian tyrosinase is commonly lower than for mushroom tyrosinase. Secondly, ‘in vitro’ tyrosinase inhibitors such as quercetin, apigenine or glycyrrhicin can stimulate melanogenesis in cell-based systems. Thirdly, keratinocytes play an active role in modulating the melanogenesis in melanocytes (Hirobe, 2005). For instance, arbutin is a more potent inhibitor of melanin formation on melanocytes co-cultured with keratinocytes than on pure melanocyte cultures (Yoon and Hearing, 2003). It has been recently reported that extracts of Lepidium apetalum decrease skin pigmentation in brown guinea pigs and cultures of human melanoma cells, only if these were carried out in a keratinocyte-conditioned medium (Choi et al., 2005). This is due to a mechanism involving IL-6-mediated downregulation of Mitf rather than a direct inhibition of tyrosinase activity. Keratinocytes are essential for the production of IL-6 and the effect of Mift to decrease tyrosinase mRNA, protein expression and finally melanin synthesis in melanocytes.

It is also noteworthy to point out that hypopigmentation strategies aim not only to inhibit melanin formation de novo but also to lower hyperpigmentation. Strategies to address the degradation of the preformed melanin are mainly achieved by desquamation or possibly by bleaching degradation with endogenous peroxides.

Finally, in order to evaluate clinical trials on depigmenting treatments without any ‘bias’, the definition and standardization of guidelines shared by all the scientific community are needed. Critical points, such as patient selection, disease improvement and treatment safety evaluation, should be regulated to obtain a clear comparison of effectiveness of new depigmenting agents as well as combination with already known skin lighteners.

Ancillary