Opioids and the skin – where do we stand?

Authors


Paul L. Bigliardi, Department of Dermatology, Neurodermatology, CHUV, Hôpital Beaumont, BT-440, CH-1011 Lausanne, Switzerland, Tel.: +41 21 314 03 56, Fax: +41 21 314 03 78, e-mail: paul.bigliardi@chuv.ch

Abstract

Abstract:  The common ectodermal origin of the skin and nervous systems can be expected to predict likely interactions in the adult. Over the last couple of decades much progress has been made to elucidate the nature of these interactions, which provide multidirectional controls between the centrally located brain and the peripherally located skin and immune system. The opioid system is an excellent example of such an interaction and there is growing evidence that opioid receptors (OR) and their endogenous opioid agonists are functional in different skin structures, including peripheral nerve fibres, keratinocytes, melanocytes, hair follicles and immune cells. Greater knowledge of these skin-associated opioid interactions will be important for the treatment of chronic and acute pain and pruritus. Topical treatment of the skin with opioid ligands is particularly attractive as they are active with few side effects, especially if they cannot cross the blood–brain barrier. Moreover, cutaneous activation of the opioid system (e.g. by peripheral nerves, cutaneous and immune cells, especially in inflamed and damaged skin) can influence cell differentiation and apoptosis, and thus may be important for the repair of damaged skin. While many of the pieces of this intriguing puzzle remain to be found, we attempt in this review to weave a thread around available data to discuss how the peripheral opioid system may impact on different key players in skin physiology and pathology.

Opioid receptors: expression in the skin

Opioids and their receptors in the skin comprise part of the endogenous opioid system and includes peptides, such as enkephalins, endorphins, dynorphins and endomorphins, and three opioid receptors: μ-(MOR/Oprm1), δ-(KOR/Oprd1) and κ-(KOR/Oprk1) receptors (16). OR are G protein-coupled receptors (GCPRs) that mediate the effects not only of the endogenous opioid peptides but also of the exogenous opiate alkaloids like morphine.

We first reported MOR protein expression in human skin more than a decade ago, followed by further reports on MOR (2–4,6,8), DOR (7) and KOR (5) and the endogenous ligands, β-endorphin (8–11,17), enkephalins (18–21) and dynorphins (5), in the epidermis and peripheral nerve fibres. Noticeably, DOR and KOR are highly upregulated in the skin of fibromyalgia patients (22). While OR and peptide ligands are highly expressed in the CNS where they modulate pain, analgesia, emotional state, addiction, cognition, energy balance and autonomous functions (23), they are expressed at a lower density in the skin and immune cells. Using real-time PCR, we measured mRNA levels of DOR and MOR in skin epidermal keratinocytes, fibroblasts and epidermal melanocytes (EMs) (Fig. 1). Cell culture studies indicated that ectoderm-derived cells (e.g. brain cells, keratinocytes and melanocytes) express higher levels of MOR than of DOR. By contrast, mesoderm-derived fibroblasts express higher levels of DOR than of MOR. MOR expression in brain tissue is approximately 200 times higher than in EMs, 1000 times higher than in epidermal keratinocytes and 20 000 times higher than in fibroblasts. DOR expression in brain tissue is approximately 200 times higher than in EMs, 1500 times higher than in dermal fibroblasts and 2000 times higher than in keratinocytes (Fig. 1). Human skin also expresses mRNA for KOR (5).

Figure 1.

 Relative MOR and DOR mRNA expression (real-time RT-PCR) in primary human cultured foreskin melanocytes (MC), keratinocytes (KC) and fibroblasts (FC), compared with the human brain. Brain DOR number is normalized to 100 000. *P < 0.01 between DOR and MOR expression (Student’s t-test).

Opioids in epidermal homeostasis

The first evidence that MOR is functionally active in human skin emerged from skin organ culture experiments. Here, β-endorphin downregulated MOR in the epidermis, and upregulated the expression of TGF-β receptor II and the differentiation marker cytokeratin 16 in the human epidermis (24). In addition, the opioid agonist β-endorphin stimulated migration of human cultured keratinocytes (25). The involvement of these factors in skin disorders is evidenced by an altered epidermal MOR expression in several dermatoses: psoriasis (24), basal cell carcinoma (26) and at the margin of chronic wounds (27). Recently, we reported significant atrophy of the epidermis in MOR (6) and DOR (7) knockout (KO) mice (Fig. 2d), suggesting a controlling role for MOR on keratinocyte proliferation and differentiation. KOR-KO mice, by contrast, exhibit epidermis of normal thickness (6). DOR-KO mice showed upregulated expression of the keratinocyte differentiation marker cytokeratin 10 and their wounds healed more slowly (7). DOR-KO mice also exhibited a greater increase in epidermal thickness during wound healing (Fig. 2d), compared with wild-type mice at day 3. These results suggest a role for DOR in keratinocyte migration as well as differentiation. Still, the treatment of wounds with topical opioids has produced seemingly contradictory results, although these studies were not similar in their experimental design. Some studies reported improved healing rates after topical opioid application (28), while others reported delayed healing with effects on skin thickness and scar formation (29). Hypertrophic scars with cacaesthesia may show upregulation of different opioid receptors (30) (Fig. 2d).

Figure 2.

 Effects of opioid receptor activation on different key players in the skin and the interactions between the skin, nervous and immune system. Opioid peptides and their receptors in the epidermis are involved in the peripheral regulation of pain and itch, especially in chronic dermatitis. Opioids modulate the functions of the different immune cell types in the skin. Opioid ligands and receptors regulate melanin production in melanosomes and hair follicle pigmentation/growth and stimulate lipogenesis in sebocytes. Opioid receptors control skin differentiation in normal skin and during wound healing and scar formation.

Signalling via the OR does not appear to affect proliferation of keratinocytes and fibroblasts by MTT assay (Bigliardi P.L. & Bigliardi- Qi M, unpublished data) (18). Few studies report growth factor-like effect of opioids in skin keratinocytes and fibroblasts (31). However, we have recently shown, that β-endorphin, acting via MOR, may stimulate proliferation of melanocytes (8,9,11).

In summary, OR signalling can influence cell differentiation, migration, cytokeratin and cytokine expression in the human epidermis. Thus, this finding may have implications not only for normal skin homeostasis but also for wound healing and scar formation.

Cutaneous opioid system and pruritus

The brain receives sensory inputs from all areas of the skin surface (32). Afferent C-fibres in the skin can be modulated by neuropeptides/transmitters secreted by cutaneous cells particularly keratinocytes and immune cells distributed around these nerve endings. MOR and DOR expression have been identified on peripheral sensory nerve fibres in human skin (2,3,33,34), and we have observed changes in their structure in chronic pruritic skin (2) as well as in MOR knockout mice (6). The epidermal nerve endings of lesional skin in pruritic skin diseases, such as prurigo or chronic atopic dermatitis with hypertrophic epidermis, are thinner compared with normal skin, and their nerve endings run straight through the epidermis all the way to the stratum corneum (Fig. 2a, right). Moreover, nerve endings are thicker and less stretched in MOR knockout mice with associated atrophic epidermis compared with wild-type mice, even after induction of dry skin dermatitis with acanthosis (6).

A recent experiment used Herpes simplex virus vectors to alter MOR expression in cutaneous afferent nerve fibres. This resulted in the alteration of peripheral opioid analgesia, emphasizing the importance of the MOR system on peripheral nerves in pain (1). On the basis of our and other clinical observations and emerging results, we propose the hypothesis that topographically segregated sensory circuits of C-fibres with different afferent inputs exist in different epidermal layers and in the superficial dermis (35) (Fig. 2a). For example, keratinocytes from the granular layer, located around the non-peptidergic C-fibres (negative for SP and CGRP), can release anti-nociceptive/pruritogenic β-endorphin (36). Basal and suprabasal keratinocytes can release pro-nociceptive endothelin and may stimulate peptidergic C-fibres (positive for SP and CGPR) in the basal epidermis. These along with other observations provide further evidence of the close relationship between epidermal nerve endings and keratinocytes, and of the importance of nerve ending location in the epidermis for consequent sensation (e.g. pain, tingling or itching).

In chronic pruritic skin diseases, such as chronic, hypertrophic atopic dermatitis, lichen simplex chronicus, prurigo and in psoriasis, MOR appears to be internalized rather than expressed on the keratinocyte surface (2,24,37). This reduced expression of surface MOR can, however, be reversed by topical application of the MOR antagonist Naltrexone (38). By contrast, MOR is not downregulated in acute forms of pruritus, such as acute contact dermatitis. Tominaga et al. (5) observed a downregulation of KOR, not MOR, in the epidermis of patients with atopic dermatitis, while ultraviolet treatment of this disease downregulated MOR but restored KOR expression.

It is widely accepted that KOR signalling suppresses itch, while MOR signalling can stimulate itch (39–41). However, if KOR signalling were truly itch suppressive, one would expect more rather than less scratching in KOR-KO mice. However, KOR- and MOR-KO mice do not differ in their scratching behaviour after induction of dry skin dermatitis (6). An obvious explanation of this finding is not clear. It may be that the pharmacological effects of putative agonists or antagonists in in vitro binding assays do not reflect the in vivo situation. Alternatively, it could be that the different in vivo models used are not directly comparable. Moreover, KOR agonists have been clinically tested on acute forms of itching, while KOR- and MOR-KO mice studies have used a chronic form of itching (i.e. a dry skin itch model). Topically applied Naltrexone (an OR antagonist) works only on chronic rather than acute atopic dermatitis (38), suggesting a different pathogenic basis for acute versus chronic itching. Acute itching is mostly due to histamine, prostaglandins, leucotriens, tryptase, CGRP, SP or KOR agonists, while MOR seems to be important in chronic forms of pruritus. In addition, it seems that the ligands are responsible not only for the final intensity and quality of the sensation, but also for the quantity of the specific receptors on keratinocytes and cutaneous nerve endings and the interaction between agonists and different types of opioid receptors. There may therefore be some benefit in defining an epidermal-neuronal unit in a similar way to the epidermal-melanin unit, where each nerve ending interacts with different keratinocytes around the neuron.

Opioids in melanocytes

Epidermal melanocytes

While a direct association between the β-endorphin/MOR system and pigmentation has only recently been shown experimentally (8,9,11), the older literature contains several intriguing observations suggesting such an involvement. For example, β-LPH, the immediate precursor of β-endorphin and β-MSH, not only stimulates melanogenesis in sheep, but serum levels of β-LPH are elevated in generalized hyperpigmentation in humans. Moreover, plasma levels of β-endorphin are higher in patients with vitiligo than in unaffected controls, and β-endorphin expression is higher in patients with lesional versus uninvolved skin in this disorder (42).

Recently, we reported that the β-endorphin/MOR system is prominently expressed in human EMs in situ and in vitro, where the peptide and its receptor appear to be closely associated with the melanin-producing melanosome (8) (Fig. 3a). Normal EMs respond to β-endorphin with increased melanogenesis, providing direct evidence that the β-endorphin/MOR system is functionally active in these neural crest-derived cells (8) (Fig. 3a). The β-endorphin/MOR-associated effects here were of magnitude similar to those reported for α-MSH and ACTH (42). Furthermore, β-endorphin exerts potent dendritogenic effects in EMs (Fig. 3a), and so can facilitate active transfer of melanin to recipient keratinocytes (Fig. 3a).

Figure 3.

 (a) Expression and function of the β-endorphin in human skin in situ and in epidermal melanocytes (EMs) in vitro at the protein and gene levels. (i) β-END (red) strongly co-localized (yellow) with gp100-positive EMs (arrows). (ii) μ-Opiate R (red) strongly co-localized (yellow) with gp100-positive EMs (arrows). (iii, iv) A marked increase in cell dendricity was seen 72 h after β-END stimulation in EMs. In addition, a 433-bp product specific to μ-opiate R and a 260-bp product specific to POMC were detected in these cultured cells under baseline condition. (v) A visible increase in melanogenesis was evident in EM pellets of the stimulated EMs. (b) Expression and function of the β-endorphin in the human hair bulb in situ and in hair follicle melanocytes and in vitro at the protein and gene levels. (i) β-END expression (red) was strongly co-localized (yellow) with gp100-positive melanocytes (green) located in the bulb and ORS of anagen hair follicles. (ii) μ-Opiate R (red) was expressed strongly by a minor subpopulation of melanocytes (green) located in the proximal/peripheral matrix region (arrow) and in the ORS of anagen hair follicles, but was absent from melanocytes in the melanogenic zone (green). Both a 260-bp product specific to POMC and a 433-bp product specific to μ-opiate were detected in cultured hair follicle keratinocytes (i1 and ii1) and hair follicle melanocytes (i2 and ii2). (iii) Primary cultures of human follicular melanocytes expressed β-END and μ-opiate receptor peptides (lower four paired fluorescence and bright-field panels). In pigmented bulbar melanocytes expression of β-END and μ-opiate R were confined to the peri-nuclear region (arrow). In amelanotic hair follicle melanocytes, β-END and μ-opiate R were present throughout the cell body. (iii and iv upper phase contrast panels) A marked increase in cell dendricity was seen 72 h after β-END stimulation in hair follicle melanocytes. (v) A visible increase in melanogenesis was evident in pellets of β-END-stimulated hair follicle melanocyte.

As β-endorphin and MOR are expressed by both EMs and keratinocytes, autocrine and paracrine action appears highly likely in the regulation of melanocyte physiology (Fig. 3a). Interestingly, H2O2-mediated oxidation of epidermal β-endorphin in vitiligo (due to methionine residue oxidation) results in a loss of its melanogenic properties (43). Moreover, dysfunction of the sympathetic nerves (including release of opioids from nerve endings) may play a role in segmental-type vitiligo. Despite this, the positive correlation between β-endorphin expression and melanocyte differentiation (i.e. pigmentation and dendricity) suggests that β-endorphin may indeed be involved in the modulation of melanocyte differentiation through autocrine control. Expression of MOR also correlates positively with melanocyte differentiation in vitro, and MOR may be up-regulated by its ligand β-endorphin, in a manner similar to MC-1R upon binding to α-MSH (42).

One of the perplexing features of β-endorphin involvement in melanocyte biology is the finding that while melanocortins induce melanogenesis by activating predominately the cAMP second messenger system (42), β-endorphin (like other opioids) inhibits this pathway. MOR agonists are known instead to activate protein kinase C (PKC) and PKC-β-dependent pathways, and importantly these pathways have recently been implicated in the regulation of melanogenesis (44). Moreover, MITF, a transcription factor for the tyrosinase gene, can also regulate PKC-β transcription (42).

Hair follicle melanocytes

Melanocytes of the hair follicle can be distinguished from their epidermal cousins by their much deeper location in the skin (i.e. beyond direct UVA/B stimulation) and by their tight linkage to the hair growth cycle (45). β-Endorphin expression is high in the melanogenically active melanocytes of the anagen VI hair bulb (11) (Fig. 3b) but low in rare non-dendritic melanocytes in catagen follicles. Thereafter, it increases again at telogen, at least in those catagen-surviving or ‘reservoir’ melanocytes distributed in the telogen secondary germ. The striking upregulation of β-endorphin expression in melanogenically active bulbar melanocytes during anagen VI suggests that these cells may be an important local source of β-endorphin at this stage of the hair cycle (Fig. 3b), and may play an important role in the regulation/maintenance of human hair follicle pigmentation (Fig. 2c). This pattern of β-endorphin expression appears to be species specific as it appears to be largely confined to the sebaceous glands throughout the murine hair growth cycle (42). These data join an increasing list of significant differences in the regulation of human and murine hair growth and pigmentation (45).

μ-Opioid receptor expression also exhibits hair cycle-dependent fluctuations in human skin (11), although unlike β-endorphin this receptor is detected only in a small subpopulation of weakly melanogenic melanocytes located in the proximal/peripheral anagen IV hair bulb matrix (Fig. 3b) and in amelanotic melanocytes located in the outer root sheath (ORS). The presence of β-endorphin but the absence of MOR in melanogenically active bulbar melanocytes is striking and perplexing. It could be simply artifactual, i.e. significant occupancy of MOR by β-endorphin permits only low levels of receptor availability to the anti-MOR antibody (i.e. ligand–receptor saturation). Alternatively, MOR expression itself may be downregulated, as it appears to be the case in catagen and telogen hair follicles. Given that β-endorphin and MOR are co-expressed only in the relatively undifferentiated melanocytes of the ORS and the proximal/peripheral matrix, these melanocytes may be especially susceptible to MOR signalling. For example, β-endorphin/MOR signalling may be involved in the migration of immature melanocytes into the hair matrix and their subsequent differentiation into pigment-producing cells. In this regard, recent studies have shown that exogenously supplied β-endorphin stimulates the migration of cultured human foreskin keratinocytes (42).

β-Endorphin/MOR are also detected in primary HFM cultures consisting of both terminally differentiated bulbar melanocytes (melanogenically active) and of amelanotic melanocytes (Fig. 3b). However, expression is reduced/lost from the highly melanized dendrites of the former cells, suggesting that β-endorphin expression could be associated with early stages in melanosome biogenesis. This was supported by immuno-electron microscopy studies that revealed a close association of β-endorphin and MOR expression primarily in less mature stage II and III melanosomes, which represent the early stages of active melanogenesis (8). The positive correlation between β-endorphin/MOR expression and melanocyte differentiation status in culture may suggest autocrine control of melanocyte function.

Exogenously supplied β-endorphin stimulates increased melanogenesis and proliferation in follicular melanocyte cultures (8) to an extent comparable with those observed with β-endorphin, α-MSH and ACTH stimulation on EM (42). This suggests that the proliferative potential of opiate receptor signalling may need to be broadened. β-Endorphin has also potent dendritogenic effects in HFM cultures (Fig. 3b).

Opioids in hair follicles

Slominski et al. have reported β-endorphin immunoreactivity in the keratinocytes of the hair follicles in the scalp but not in truncal skin (42). We found β-endorphin expression to be only weakly detectable in the hair growth-inductive follicular dermal papilla (FP) during anagen (Fig. 3b). However, β-endorphin immunoreactivity becomes more readily detectable in the FP during catagen and telogen, as well as in the hair follicle epithelium. This reduction in FP β-endorphin levels in anagen is contrary to the usual pattern of hair cycle-dependent fluctuations in growth factors (46). By contrast, the FP mesenchyme and follicular epithelium consistently shows prominent MOR expression throughout the hair cycle. Thus, the presence of the β-endorphin and MOR in keratinocytes and FP fibroblasts suggest that paracrine effects may also regulate follicular melanocyte behaviour during catagen and telogen. It is interesting to note that β-endorphin expression is readily detected in isolated FP cells under in vitro conditions, in contra-distinction to anagen FP cells in situ. This could be due to their differentiation/proliferative status in vitro. A possible role for MOR signalling in hair growth may be inferred from recent evidence showing that topical application of a selective PKC inhibitor, bisindolylmaleimide, reduces hair length (presumably via anagen truncation) and calibre (presumably via alteration of keratinocyte proliferation/ differentiation) in guinea pigs and mice in vivo (47).

Opioids in sebocytes

The human sebaceous gland is also a target for POMC peptides including β-endorphin (48). Its high-affinity receptor, MOR, can be detected in the peripherally located sebocytes of the human sebaceous gland in situ, while MOR, but not DOR, has been reported to be expressed in the human sebocyte cell line SZ95 (46). β-Endorphin significantly suppressed epidermal growth factor-induced growth of these sebocytes when grown in a chemically defined calcium-rich medium. β-Endorphin also enhanced lipogenesis and was as potent as linoleic acid in increasing C16:0, C16:1, C18:0, C18:1 and C18:2 fatty acid levels in these cells.

Role of opioid receptors in skin immune cells

There are still few studies on the role of opioids on immune cells in the skin (see later). However, their potential importance can be inferred from their impact on immune cells located at other sites in the body. Opioids can modulate immunity either directly through receptors expressed on the immune cells themselves, or indirectly via receptors in the nervous system. Numerous studies over the last 30 years have reported on the immune actions of the opioids, and all immune cell types (e.g. neutrophils, monocytes/macrophages/dendritic cells, B and T lymphocytes) have been described as targets (Fig. 2b). Multiple approaches have been employed to characterize opioid receptors expressed by immune cells, including RT-PCR, radioligand binding with general opioid, MOR-, DOR- and KOR-selective ligands, as well as functional assays in vitro and in vivo. Altogether, the literature shows that the expression of opioid receptors on immune cells is dependent on their maturation and activation state (49,50). DOR and KOR have been demonstrated more easily than MOR. DOR is upregulated in T lymphocytes upon activation in vitro (50) and in vivo (51). Opioid receptor function is regulated at the transcriptional level, and IL-1, IL-4, IL-6 and TNF-α can upregulate MOR (52) in a NF-κB-dependent manner (53). NF-κB and other factors also regulate KOR and DOR transcription (53). Overall, KOR activation decreases the inflammatory response by downregulating several cytokines and chemokines (54). By contrast, MOR activation may induce a pro-inflammatory response (54), although anti-inflammatory effects of MOR signalling were also reported at the intestinal barrier in experimental colitis, in hepatic inflammation (55) or in the skin after incision wounding (56).

Over the last 10 years, the study of OR- and opioid peptide-KO mice has facilitated elucidation of the role of the opioid system in vivo (16) with regard to changes in immunity and skin physiology. MOR-KO mice exhibited a decrease in morphine-induced immune alterations (57,58), reduction of acute fever and neuroinflammation induced by lipopolysaccharide (59), chronic stress-induced lymphocyte apoptosis, and allergic airway responses (60,61). In one study, MOR deletion reduced skin mast cell content but did not alter skin CD4+ T lymphocytes (6). Thymocyte negative selection, as assayed in vitro, is altered in DOR-knockout mice, although adult mice display normal thymocyte and splenocyte distributions (62,63). Finally, KOR-KO mice have fewer thymocytes, more splenocytes, as well as an increased antibody response (64). β-Endorphin-deficient mice exhibit augmented cytokine production (65), while studies with preproenkephalin-KO mice provided evidence that this opioid is a Th2 cytokine (66).

Human dermal antigen-presenting cells express enkephalin (67) and β-endorphin regulates cytokine secretion by epidermal dendritic (Langerhans) cells (12). KOR activation suppressed the ability of dendritic cells to induce T-lymphocyte proliferation, whereas it did not alter their antigen uptake or phenotypic maturation (13). DOR also appears as an important modulator of dendritic cell function, as shown by receptor upregulation during dendritic cell maturation and enkephalin-induced chemotaxis (68).

Opioids are an important link between inflammatory responses and pain sensation. At the inflammation site, leukocytes secrete inflammatory mediators that induce pain. Opioid peptides that bind to opioid receptors on peripheral nerve terminals, by contrast, produce analgesia (69,70). There is a reciprocal cross talk between chemokines and opioid peptides: so that the production of cytokines and chemokines by cells migrating to the inflamed tissue can be regulated by both chemokines and opioid peptides. Opioid peptide release can be regulated by chemokines (14,54). Interestingly, at the receptor level, heterologous cross-desensitization between opioid receptors and chemokine receptors (71) as well as the induction of opioid peptide release from neutrophils by chemokines (70) have been demonstrated. Therefore, the pattern of chemokines present at the inflammation site can alter leukocyte responses to opioids produced locally and hence regulate further inflammatory mediator release. Recently, morphine has been shown to change local cytokine expression and neutrophil infiltration in a skin incision model: when given acutely before incision, morphine was found to reduce cytokine expression (56). By contrast, chronic morphine administration before incision increased the expression of these cytokines (72). The distinct roles of each opioid peptide and their cognate receptors on the immune cell populations present in the skin, both under physiological or pathological situations remain to be fully explored.

Conclusions

Figure 2 shows our current understanding of the many roles the OR system may play in human skin and how this system may impact the different key players in the skin, i.e. keratinocytes, melanocytes, hair follicles, sebaceous glands, different immune cell types and the peripheral nerve fibres. The multidimensional nature of this highly interactive system highlights how this cutaneous neuro-immune-endocrine system can affect skin physiology and pathology, and so how this has implications for skin treatment. Many questions remain to be answered. How is cutaneous cell differentiation modulated by OR signalling? What is the role of opioids in the interactions between cutaneous cell populations and peripheral nerves? Novel approaches such as knock-down or over-expression in different skin cells or specific knockout of the receptor in conditional KO mice models could help us separate central from peripheral effects of opioids on the skin. New therapies for skin diseases like psoriasis or atopic dermatitis, for wound healing and for pain and itch management will emerge when we better understand how these systems interact in the skin.

Acknowledgements

Procter & Gamble for supporting the POMC-related melanocyte research in D.J. Tobin’s laboratory. The work by C. Gaveriaux-Ruff was supported by ‘Université de Strasbourg’, CNRS and INSERM. Bigliardi’s work was supported by the University of Lausanne. We also thank M. Clavel from the University Hospital Lausanne, CEMCAV for the design work.

Ancillary