Andrey A. Panteleyev Department of Dermatology College of Physicians & Surgeons Columbia University 630 West 168th Street Vanderbilt Clinic VC-1526 New York, NY 10032 USA Tel.: +1 212 305 7699 Fax: +1 212 305 7391 e-mail: email@example.com
Abstract: 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) is among the most toxic pollutants known to date that serves as a prototype for a group of halogenated hydrocarbon compounds characterized by extraordinary environmental persistence and unique ability to concentrate in animal and human tissues. TCDD can elicit a complex array of pleiotropic adverse effects in humans, although chloracne, a specific type of acne-like skin disease, is the only consistent manifestation of dioxin intoxication, thus representing a ‘hallmark’ of TCDD exposure. Chloracne is considered to be one of the most specific and sensitive biomarkers of TCDD intoxication that allows clinical and epidemiological evaluation of exposure level at threshold doses. The specific cellular and molecular mechanisms involved in pathogenesis of chloracne are still unknown. In this review, we summarize the available clinical data on chloracne and recent progress in understanding the role of the dioxin-dependent pathway in the control of gene transcription and discuss molecular and cellular events potentially involved in chloracne pathogenesis. We propose that the dioxin-induced activation of skin stem cells and a shift in differentiation commitment of their progeny may represent a major mechanism of chloracne development.
Technological progress has created great benefits for humanity, but all too often has come at the price of escalating health hazards arising from unwanted industrial by-products and chemical waste. The impact of environmental pollution on human health is a novel and formidable evolutionary challenge. Thus, understanding the molecular, cellular, clinical and epidemiological aspects of exposure to environmental pollutants is essential for our future survival.
Perhaps the most vivid example of toxic chemical agents is the halogenated hydrocarbon 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) that serves as the prototype for a large group of polyhalogenated dibenzo-para-dioxins which evoke a wide range of adverse toxic effects in humans and animals. In these tricyclic aromatic compounds, hydrogen atoms can be substituted by up to eight chlorines, thus permitting about 75 different dioxin isomers. Among them, TCDD with four chlorine atoms in lateral positions (Fig. 1) is the most biologically active isomer. Its extraordinary environmental stability and unique ability to concentrate in animal and human tissues coupled with its extreme toxicity make dioxin a ‘model’ environmental toxicant.
2,3,7,8-Tetrachlorodibenzo-p-dioxin can elicit a complex array of pleiotropic responses in humans ranging from altered immunity to reproductive defects (1). Yet, it is only the skin effects of TCDD exposure, especially chloracne, a specific type of acne-like disease, that are the most consistent manifestation of dioxin intoxication – indeed a ‘hallmark’ of TCDD exposure (2). Owing to specific features pathognomonic for dioxin skin toxicity (high correlation of severity with the level of dioxin exposure, short incubation period, persistence and specific anatomic localization), chloracne is considered to be among the most reliable biomarkers of TCDD toxicity in humans (3). Over the past 75 years, chloracne outbreaks have occurred in the setting of industrial and non-industrial accidents involving substantial numbers of people around the world. These accidents have allowed for close observation of affected individuals over extended time periods and have helped increase our understanding of the consequences of such exposure.
The clinical features of chloracne are clearly defined and thought to relate to the non-inflammatory alteration of keratinization in the pilosebaceous unit (4–9). And yet the cellular and molecular mechanisms of dioxin-induced chloracne are still unknown. Numerous questions remain unanswered. How does chloracne differ from other acneiform skin conditions? Which skin cell populations represent the primary target of dioxin activity? What is the molecular pathway for chloracne and is it a part of the classic TCDD-dependent transcriptional mechanism that requires activation of dioxin (aryl hydrocarbon) receptor (AhR) and its binding to aryl hydrocarbon receptor nuclear translocator (ARNT)?
In 1985, Tindall (9) in an excellent review of chloracne wrote: ‘Much work has been done, but considerably more work is necessary for a complete understanding of mechanisms, response and short- and long-term risks’. Now more than 20 years later, remarkably little progress has been achieved in our understanding of this disorder.
In this review, we attempt to summarize the available clinical data and recent progress in understanding the role of dioxin and its receptor (AhR) in the modulation of gene transcription (e.g. 10–13), thus intending to delineate putative dioxin-dependent molecular pathways and cellular events that are likely involved in chloracne pathogenesis.
Chloracne – an environmental disease
Chloracne is a well-recognized but poorly understood clinical entity (5–7), characterized by an acne-like eruption of comedones (blackheads and whiteheads), cysts and pustules that occurs following systemic absorption of chemical ‘chloracnegens’.
This condition was first described in German industrial workers in the late 19th century by Von Bettman (14) and later by Herxheimer (15) who suggested that it was caused by chlorine exposure and hence coined the term ‘chloracne’ based on the similarity of its clinical features with acne vulgaris. Later, chloracne was recognized in workers employed in the production of various industrial chemicals including chlorinated phenols (16,17), polychlorinated biphenyls (PCBs) (18,19) and chlorinated naphthalenes (4), while some of these compounds in pure form are not chloracnegenic or possess low chloracnegenic activity (20). Identification of the specific ‘chloracnegenic’ factor in these mixtures remained enigmatic until trace contaminants formed during their manufacture were identified and linked to the disease (9). These contaminants include polyhalogenated dibenzofurans, dibenzo-p-dioxins and azoxybenzenes. All chloracnegenic compounds are known to share certain structural features including molecular planarity and two benzene rings with halogen atoms occupying at least three of the lateral ring positions. The position of the halogen substitutions appears to be critical, as it is known that substitutions leading to molecular non-planarity dramatically diminish chloracnegenic activity (21).
In experimental studies of polychlorinated naphthalenes applied to the skin of human volunteers, chloracne occurred in a widespread distribution, not limited to the sites of application (22). Ingestion and inhalation are also effective means of chloracne induction. Based on a study of a mass intoxication by 2,3,6,7-TCDD, it was assumed that in contrast to the other cutaneous signs of exposure to polyhalogenated aromatic compounds, chloracne is always a manifestation of systemic intoxication by chemical chloracnegens and not just a cutaneous disorder (23). The role of systemic exposure in the pathogenesis of chloracne is consistent with the very low rates of dermal absorption of chloracnegens (24). Because of their highly lipophilic nature, skin exposure to chloracnegenic compounds including dioxins results in their accumulation in the stratum corneum which in effect diminishes absorption. Removal of stratum corneum dramatically increases TCDD absorption into the skin (25). Thus, the major cause of chloracne is systemic exposure.
Development of chloracne lesion. The key feature of chloracne is non-inflammatory alteration of keratinization of the pilosebaceous unit (9,26) that results in the formation of comedones, straw-coloured cysts, pustules and non-infectious abscesses that may scar (27). The patients are otherwise in good health. Pruritus is rare (28).
Chloracne usually starts as an acute erythema of the face sometimes associated with oedema (29). Erythema is followed within days by the formation of fine comedones (blackheads or whiteheads) – the most characteristic feature of chloracne (Fig. 2a). This process may involve almost every follicle in the affected area giving the involved skin a slate-gray appearance (Fig. 3a). Developing comedones start to shed hairs (Fig. 2a), while sebaceous lobules are still active, although involuting, and continue to secrete sebum (30–32).
Most investigators describing the histopathological features of chloracne have identified comedones as the primary skin lesion and suggested that squamous metaplasia of the sebaceous gland (SG) is responsible (e.g. 4,9,31).
Initially, cream-coloured cystic lesions (Fig. 3b) are less common than comedones and occur primarily on the face and neck (28,33). In long-standing lesions, non-inflammatory infundibular cysts (Fig. 3a, arrows) predominate over comedones (32) and are virtually pathognomonic of advanced chloracne (5). The cysts vary in size from 1 mm to 1 cm in diameter, and have a central orifice or pore which may not be obvious (27). In contrast to primary comedones, the pilary portion of infundibular cysts is almost always destroyed so that few or no hairs remain within the cavity. Sebaceous acini are also absent in cystic lesions (32). Involved follicles are typically surrounded by a collarette of scale (perhaps a result of infundibular hyperkeratinization) and the hair shaft is often broken and coiled. In some areas of destruction of follicular walls, a giant cell infiltrate may be present (33). These cystic lesions are virtually always sterile (32), but occasionally secondary infection of affected follicles may occur (34).
Chloracne is not associated with cutaneous inflammation (32), but in severe cases, non-infectious folliculitis can occur and the picture can resemble severe inflammatory cystic acne. This is particularly common on the back and legs (5,31). Follicular lesions in chloracne are often associated with skin thickening (32,35).
Exposure to chloracnegen action may also be associated with palmoplantar hyperkeratotic lesions of sweat gland origin (35–37). Hyperkeratinization of sweat gland duct results in acrosyringial plugging (38) similar to the plugging in the comedones of follicular origin.
Chloracne localization. The distribution of the lesions of chloracne is very characteristic. Comedones appear most frequently on the face and neck (between 90% and 100% of affected persons) and forearms (47%), forming extensive sheaths in many cases. The most commonly involved areas are below and to the outer side of the eye (the so-called malar crescent) and in the postauricular triangles. The ear lobes, suboccipital hairline and groin are often involved (Fig. 2b,c). The cysts diminish in number as the condition extends on to the cheeks, forehead, maxillary areas and sides of the neck, whereas the nose, perioral skin and supraorbital regions are usually spared. The pustular component is usually more noticeable on the neck (Fig. 3c, arrow). Comedones and cysts can appear on the shoulders, mid-portion of the back and the chest. A peculiar peppering of the skin with tenacious carbon-coloured comedones is often apparent around the umbilicus and in lower abdomen. The outer surfaces of the forearms and anterior thighs show similar although fewer comedones. The penis may be involved in a similar process, whereas the scrotum typically shows more cystic lesions (5,31,33,34). Axillary lesions are also common in chloracne patients (31,32,39). While the overall pattern is clearly acneiform, it differs from typical acne vulgaris particularly in the diminished sebum secretion and in the peculiarly dark colour of the comedones as well as the general peppered distribution in areas not prone to acne vulgaris (4,18). In patients who ingested rice oil contaminated with PCBs (Yusho, Japan), chloracne was most severe on the face, neck and genitalia with little involvement of the extremities (37).
While virtually every hair follicle (HF) in involved areas may be affected, HFs of different types are not equally susceptible to chloracne. Vellus follicles appear to be most sensitive, whereas scalp follicles are resistant (4). Predisposition of specific types of HFs to comedogenic action of chloracnegens may be relevant to specific patterns of chloracne lesion distribution over the body.
Chloracne persistence and severity (dose–response). The natural history of chloracne is highly variable. There is usually a delay of about 2–4 weeks from initial contact with the offending chemical to the onset of clinical lesions. In the case of intensive exposure, symptoms may appear within days (27).
With less severe intoxication, slow spontaneous improvement may begin immediately (33), but in general, assuming no further exposure, the skin lesions resolve within 2–3 years (28,37). And yet in some individuals, chloracne lesions may persist for as long as 15 years after all chemical exposures have ceased (34). According to one study, the mean duration of chloracne in workers accidentally exposed to by-products of 2,4,5-trichlorophenoxyacetic acid was 26 years, with some individuals remaining disfigured for more than three decades after the accident (40). Examination of the skin in a group of 288 Vietnamese veterans with a remote (17–22 years) history of exposure to the herbicide known as Agent Orange revealed persistent chloracne lesions in 11.5% of these subjects (41).
The chronicity of chloracne is still inexplicable. It may relate to the fact that chloracnegens are highly lipophilic and remain in body fat for sustained periods. The possible role of slow release of chloracnegens from fat cells in chloracne persistence was suggested based on the studies of Yusho patients in Japan in whom lesions persisted longer than 30 years showing a certain degree of correlation with blood levels of chloracnegens (42). Nevertheless, neither the extent nor the duration of chloracne necessarily correlates with the concentration and half-life of the chloracnegens in the body, thus suggesting alternative mechanisms. The ‘continuing modification of the metabolism of the pilosebaceous unit’ by the chloracnegen (6) seems to be a more likely explanation of chloracne persistence.
Recidivism of chloracne lesions despite total lack of further contact with the offending agent has also been reported (e.g. 30,43,44), but no satisfactory explanation for this phenomenon has been forthcoming.
The severity of chloracne depends on three major factors including (i) intensity and duration of exposure (‘dosage’), (ii) relative ‘chloracnegenic potency’ of the specific compound and (iii) ‘individual susceptibility’ (2). It is of interest that typical lesions of chloracne may occur in workers’ relatives who have never been in or near sources of exposure to chloracnegens. In these cases, the lesions were likely caused by contact with work clothes or tools which had been brought home, or from direct bodily contact (44,45), thus suggesting that in some circumstances, even trace amounts of chloracnegens may cause the disease.
Individual susceptibility to chloracne is quite variable. Experimental topical application of Halowax 1014, a mixture of hexa- and pentachloronaphthalenes (Bakelite Corp., New York, NY, USA), to the skin of volunteers showed that some subjects had no skin effects at all, whereas others developed severe chloracne. Older females appear to respond only weakly, or not at all, even to high concentrations of chronically applied chloracnegen (2). After an industrial explosion at a 2,4,5-trichlorophenol plant in Bolsover, Derbyshire, UK in April 1968, 79 cases of chloracne were recorded. In some workers, chloracne developed within days, whereas others showed no symptoms until between 2 and 3 months after the last known exposure. Younger men, particularly those of fair complexion, were first affected, while the reaction was most persistent in sallow-skinned men in the 25–40 age group. In some individuals, the condition was predominantly cystic, while in others there were comedones, such that not a single pore was uninvolved (34). In some patients, the condition was unchanged 11 months after cessation of exposure, while in others it was improving (28), thus suggesting that both the induction and recovery phases show individual patterns.
Studies of workers at trichlorophenol plants in Germany (46,47) and Czech Republic (48,49) revealed that most patients had severe chloracne, while only about one-third had additional evidence of systemic intoxication (neuropathy and liver damage). Of interest, only one patient with these symptoms had no chloracne. Thus, resistance to chloracne is quite rare.
Unfortunately, in most recorded cases of chloracne, it is not possible to attribute the differences in clinical effects (persistence and severity of the condition) to individual differences in sensitivity or to different levels of exposure. One of the few exceptions is a recent study of two patients with severe TCDD intoxication (50). In this case, one patient developed severe generalized chloracne, while another heavily intoxicated person (TCDD content of 26 000 pg/g of blood fat) showed only mild facial lesions, thus further confirming the role of individual susceptibility in the severity of the disease. Elucidation of the genetic factors determining susceptibility to chloracne is essential for understanding this condition.
Few detailed histological studies of chloracne have been published (4,18). According to these publications, the first signs of histological changes become apparent about 5 days after exposure to the chloracnegen and are associated with increased cell proliferation in the infundibular portion of the HF outer root sheath and in the SG duct along with a decrease in the number of sebocytes.
Ten days: there is dilatation of the infundibulum (especially in its lower portion); increased thickness in the upper outer root sheath that forms the infundibular walls; thickening of the SG duct that merges with the infundibulum and hyperplasia of the epidermis with incomplete keratinization as manifested by parakeratosis. No signs of SG necrosis or lysis are evident and normal-appearing sebocytes are present, but their number is substantially reduced resulting in diminishing of sebaceous lobules. Thus, chloracnegen-induced transformation of HF starts from its middle portion – around the bulge and SG duct.
Three to five weeks: small comedones appear. Only a few vellus follicles remain uninvolved. Anagen and catagen follicles are equally affected. Diffuse thickening of the middle outer root sheath progresses in all follicles and sebocytes disappear completely. Dilatation of the infundibulum is pronounced inferiorly and the cavity becomes filled by a comedo consisting of many layers of keratinized cells mixed with remaining sebum. Infundibular dilatation is maximal proximally, resulting in either bottle-shaped formations, with the neck near the surface, or columnar funnels along the entire length of the infundibular structure. Telogen stage hairs are present within the keratin mass; their bulb looks unaffected, suggesting that the lower portion of the HF does not undergo hyperplasia. The perifollicular dermis shows mild oedema and slight lymphocytic infiltration localizing around the SG or the proximal (lower) part of the infundibulum (18).
Six to twelve weeks: large comedones are apparent in all follicles from which only a few hair shafts emerge indicating hair shedding. Two types of pustules may occur: (i) deep within the dermis producing the gross swelling in which the comedo is visible centrally or (ii) superficial in which the comedo is much less prominent (Fig. 3c, arrow). In deep pustules, thinning of the outer root sheath is common, eventually resulting in rupture with extrusion of the keratin/lipid mass into the dermis, followed by a foreign-body reaction. These pustules may communicate with one another forming a carbuncle-like pattern. Healing is associated with irregular depressed scars (Fig. 3d).
Differential diagnosis: chloracne versus acne vulgaris
Based on the resemblance of the comedones and cysts that characterize chloracne to those seen in patients with severe acne vulgaris, chloracne was originally thought to represent an acne vulgaris-like disease. It is now clear that the pathogenesis of these disorders is distinctive (6,33).
Differentiating chloracne from acne vulgaris requires consideration of the clinical features, anatomic localization, age of onset and exposure history (51). The differential features of acne vulgaris and chloracne are summarized in Table 1.
Table 1. Differential features of acne vulgaris and chloracne
Age group affected
Adolescence and early adulthood
Any age group
Face including the nose, upper back and chest
Retroauricular and malar areas, axillae, groin, extremities; nose is spared
Inflammatory lesions are common
Inflammation is very rare (only as a secondary effect after cyst rupture)
Decreased; xerosis as a common associated condition
General hyperplasia of the infundibulum and sebaceous gland duct; significant thickening of the upper follicle
Biochemistry and microbiology
Biochemistry of comedones
More free fatty acids, triglycerides and total triglyceride pool
More squalene, wax ester and cholesterol
Androgen dependent; testosterone and dihydrotestosterone stimulate sebum production
Antiandrogenic; the role of androgens in chloracne is unknown
Propionibacterium acnes and Propionibacterium granulosum in sebaceous gland duct and hair canal
The anatomic distribution of chloracne is distinctive (5,31). Chloracne also differs from acne vulgaris in that it can occur in any age group, including prepubertal children (8). Epidemiological studies show no evidence that adolescent acne is a predisposing factor to developing chloracne (5). Indeed, virtually everyone with sufficient exposure to chloracnegens will develop chloracne (while the severity of the condition may be quite different, as we mentioned earlier).
Chloracne lesions rarely manifest inflammation (33), while in acne vulgaris inflammation is a common feature. The characteristic inflammation of acne vulgaris may relate to sebaceous lipids, their metabolites and by-products of the Propionibacterium acnes which are known irritants (52). P. acnes colonization is an essential feature of acne vulgaris lesions, but they are totally absent in chloracne which may explain the non-inflammatory nature of this condition (6,33).
Analysis of skin surface lipids using thin-layer chromatography reveals that in chloracne squalene, wax esters and cholesterol are increased, whereas free fatty acids, triglycerides and the total triglyceride pool are dramatically decreased (6). Skin surface cholesterol originates predominantly from keratinous structures (53) and because keratinized comedones are typically far more numerous in chloracne when compared with acne vulgaris, it is not surprising that elevated cholesterol is characteristic of chloracne. These alterations in skin surface lipids suggest that chloracnegens distort metabolic pathways in the pilosebaceous unit (6).
In chloracne patients, the skin surface is not oily. SGs are often smaller than usual or absent and sebum production is dramatically reduced (29). Thus, while the pathogenesis of acne vulgaris lesions involves excessive sebum excretion, chloracne is usually associated with cutaneous xerosis. This difference may account for the relative ineffectiveness of standard therapeutic agents for acne vulgaris in chloracne patients (33).
Elevated sebum secretion is a sine qua non for acne vulgaris and strongly correlates with the severity of the condition (54). Sebum secretion is androgen dependent, thus suggesting the involvement of androgens in pathogenesis of acne vulgaris. Both testosterone and its metabolite dihydrotestosterone (DHT) can stimulate sebum production (55,56). While the effect of TCDD and other chloracnegens on serum androgen concentration is controversial (e.g. 57–61), they have been shown to impede androgen action both in vitro and in vivo (62), probably by interfering with androgen receptors and/or by blocking androgen-induced cell proliferation through suppression of cyclin D1 and modulation of p21. TCDD mimics the action of DHT, but when TCDD and DHT are applied together, TCDD completely blocks DHT activity (62). Thus, chloracnegens appear to suppress androgen effects which is consistent with the suppression of sebogenesis seen in chloracne patients.
The similarities between chloracne and acne vulgaris have led to the use of agents effective in treating the latter in patients with the former. Topical retinoids appear to be effective comedolytic agents in acne vulgaris (6,65). Some reports have suggested that retinoids are effective in the treatment of chloracne as well (e.g. 63,64). However, a large-scale experimental study has shown that oral administration of vitamin A prior to topical application of chloracnegens to the skin of healthy volunteers neither prevents the development of chloracne nor alters the time of onset, the clinical appearance, progression of the reaction and the histology of the affected sites (4). Additional studies further confirmed that moderate and severe forms of chloracne are not sensitive to vitamin A therapy (31,33,66,67).
High doses of orally administrated corticosteroids have also proven ineffective in severe chloracne (35,66). Furthermore, it was shown that the use of steroids in the early stages of chloracne development tended to aggravate the condition (34). Despite these negative results, retinoid/corticosteroid treatment is still used to treat chloracne. Unsurprisingly, in most cases, chloracne is refractory to treatment with the usual measures employed in acne vulgaris (e.g. 33,39,44). Of these, only surgical modalities such as incision and drainage of comedones and cysts, dermabrasion and light electrodesiccation may be helpful in some chloracne patients (68).
Thus, chloracne appears to be resistant to all tested forms of treatment (34,66) and none of the occasional positive reports is supported by placebo-controlled double-blind studies. Indeed, it seems likely that the statement of Dr Linda Birnbaum (69)‘…chloracne is not curable; the only treatment we know of is time’ is perfectly true. Therefore, the only way to manage chloracne is to prevent it by eliminating any possibility of exposure to chloracnegens. Once exposure does occur, affected individuals should be removed from exposure sites and efforts made to mobilize accumulated chloracnegens from the body. Recent research by groups at the University of Vienna (70) and at the University of Western Australia (71) showed that the rate of elimination of chloracnegens can be effectively increased using a synthetic dietary fat substitute known as olestra. Natural fats consist of a triglyceride molecule with three attached fatty acids, whereas olestra is synthesized using a sucrose backbone which can bind up to eight fatty acids. As a result, there is no gastro-intestinal absorption of olestra. Chloracnegens are highly soluble in olestra, thus accelerating their fecal excretion (70). Experimental studies with mice gavaged with the 14C-labelled chloracnegen hexachlorobenzene show that combined dietary olestra and caloric restriction causes a 30-fold increase in the rate of excretion of labelled compound, relative to an ad lib diet or reduced caloric intake alone (72).
Animal models for chloracne
The interspecies differences in toxic effects of chloracnegen exposure (e.g. toxicity, thymic atrophy, etc.) are quantitative rather than qualitative (73). Surprisingly, the skin effects of these agents vary across species. Specific chloracne-like skin changes are characteristic only for rabbits (ears only) (4,74–76), primates (77) and laboratory mice carrying a mutation in the hairless (hr) gene (78). Some specific changes in the skin due to the action of dioxin-like compounds occur in cows and horses (51). Other laboratory, farm or wild animals, including extremely dioxin-sensitive species like guinea pigs, manifest no skin changes. No good explanation exists for these inter-species differences.
Rabbit ear. Application of halogenated hydrocarbons to the inner side of the rabbit ear as a laboratory model for chloracne was first proposed by Kimmig and Schulz (74). The treatment results in the formation of numerous comedones mimicking human chloracne (74). As in the human disease, within 5–7 days after starting daily topical application of chloracnegens, follicular keratosis and hypertrophy of SG duct epithelium associated with marked shrinkage of SG acini occur (79). Gradual recovery of the pilosebaceous apparatus follows cessation of treatment (4).
Rabbit ear model is exquisitely sensitive to dioxin1 and is widely used for the evaluation of chloracnegenic potency of chemical mixtures and as a detection system for the presence of dioxin in contaminated soil and at commercial workplaces (26,80,81). Surprisingly, this model has never been used for studies of cellular and molecular mechanisms of chloracne development.
Primates. Primate skin provides the most relevant model for human chloracne. In rhesus macaques, systemic administration of a single dose of dioxin between 30 and 70 μg/kg of body weight results in cysts and comedones typical of chloracne. According to McConnell and Moore (77), the cysts represent transformed SGs that have undergone squamous metaplasia and keratinization. Excessive growth of cysts results in their rupture with a secondary inflammatory response. Besides chloracne-like lesions, other skin effects of dioxin intoxication have been found in macaques including alopecia, loss of nail plates and significant hyperplastic and metaplastic alterations in lipid-secreting Meibomian glands (77).
Hairless mice. The hairless (hr/hr) mouse is the only known laboratory strain that responds to dioxin treatment with symptoms strikingly similar to human chloracne (78,82–84). Hr mouse mutants show disintegration of HFs and total hair loss during the third week of postnatal life. As a result, hr mouse skin contains two unique structural remnants of the HF: (i) the utriculi, epidermis-associated comedo-like sacs derived from the HF infundibulum and (ii) closed dermal cysts. The epithelium of the utriculi in hr/hr skin is actively keratinizing, whereas the epithelium of dermal cysts displays only a low level of keratinization (85,86). Dioxin treatment induces rapid involution of utriculi and adjacent SGs without any signs of squamous transformation of sebocytes. Even at the latest stages of SG involution after the complete loss of SG structure, the remaining single sebocytes appear normal (83,87) similar to findings in human chloracne (4). Thus, dioxin-induced SG involution appears to be due to cessation of sebocyte replenishment rather than altered sebocyte differentiation.
In contrast to the utricular and SG epithelia, the dermal cyst epithelium shows hyperkeratinization after TCDD treatment (83,88,89). The resultant structures and their growth dynamic resemble the keratinized cysts characteristic of chloracne in dioxin-exposed individuals (5,33). Under TCDD action, dermal cyst epithelium normally expressing HF-specific keratin 17 (90,91) starts to express epidermis-specific keratin 1 along with total loss of keratin 17 positivity. These findings suggest that TCDD treatment results in the conversion of dermal cyst keratinocyte differentiation programme from the ‘follicular’ to the ‘epidermal’ type (83). TCDD-induced suppression of follicular and sebaceous differentiation in hr mouse skin is reversible. Cessation of TCDD treatment is followed by gradual restoration of functional SGs (87).
These data suggest that dioxin-induced chloracne-like changes in hr mice most likely reflect an imbalance in the determination of epithelial-precursor-cell fate, rather than a primary pathogenetic event in the SG or dermal cyst.
Horses and cows. In both these species, exposure to chloracnegens (chlorinated naphthalenes) results in extensive hyperkeratosis (or ‘X disease’), abnormal growth of hooves and (only in horses) Meibomian gland abnormalities (92,93). Some of these symptoms resemble certain effects of dioxin intoxication in humans and primates, but no clinical or experimental studies have been performed in farm animals ever since and whether they are able to develop chloracne-like lesions under dioxin exposure is still not known. In 1971, accidental intoxication of horses with dioxin-containing chemicals in Missouri resulted in alopecia (22 horses) and skin ulcers (16 horses). Half of the animals died (94,95).
Dioxin – the most potent environmental chloracnegen
Dioxins became infamous as an occupational cause of chloracne after a trichlorophenol reactor explosion in 1949 in Nitro, West Virginia (22). Later, it was suggested that dioxins may be the major chloracnegenic factor in various polyhalogenated chemical formulations (96) and soon thereafter, a series of studies were undertaken to elucidate their chemical and toxic properties (74,97). But it was not until the 1970s that dioxins were studied in detail when extreme toxicity of TCDD became apparent (20) and large numbers of people were exposed to its action in several industrial accidents around the world. This was also abetted by the extensive exposure of American veterans to TCDD-containing defoliants including Agent Orange during the Vietnam War (e.g. 98,99).
Dioxin fate in the environment
Chemically, TCDD (C12H4O2Cl4; molecular weight 322 g/mol) is very stable and resistant to photolysis, chemical oxidants and heat up to 800°C. It is highly soluble in organic solvents (especially in o-dichlorobenzene) and practically insoluble in water. Its half-life in soil is about 1.5 years (100), but in river and lake sediments, it may remain for decades (101) and is a source of enduring exposure for aquatic organisms. One of the most critical chemical features of TCDD is its extraordinary lipophilicity (102) and associated ability to concentrate in terrestrial and aquatic food chains reaching very high concentrations in the tissues of birds of prey, seals and humans (103–105).
2,3,7,8-Tetrachlorodibenzo-p-dioxin is formed as a by-product during the combustion of chlorine-containing materials and during high-temperature reactions of industrial organic synthesis utilizing chlorine (106). While some believe that environmental TCDD originated from forest fires and home furnaces (107), TCDD was first detected in stratified lake and swamp sediments dating to the 1930s (108,109), a time when mass production of multiple polyhalogenated aromatic compounds was just beginning. These findings support the largely industrial origin of the TCDD in the environment.
The initial source of TCDD was the production of polychlorophenol-based wood preservatives. Later, during WW II, the first herbicides with hormone-like action were developed in the USA based on 2,4-dichlor (2,4-D) and 2,4,5-trichlorphenoxyacetic (2,4,5-T) acids. In the 1960s, the production of 2,4-D and 2,4,5-T increased substantially and soon reached its peak due to the extensive use of these herbicides during the Vietnam War. Between 1962 and 1971, about 15 million gallons of a mixture containing both 2,4-D and 2,4,5-T known as Agent Orange was used as a defoliant in Vietnam. The average concentration of TCDD in Agent Orange was about 3 μg/ml (110). The total amount of TCDD released into the environment during the Vietnam War was estimated to be about 500 kg (99).
By the 1970s, the production of phenoxyacetic acids and other halogenated hydrocarbons such as DDT and the PCBs was banned in the United States and in most European countries, but the environmental release of TCDD continued nonetheless. For example, the total TCDD deposition flux to the Great Lakes in 1996 was estimated at about 57 g (111). Although this may seem like a very small amount, 57 g of dioxin is equivalent to 1 year of the World Health Organization's defined Tolerable Daily Intake of TCDD for approximately one billion people.2
Currently, the most significant sources of dioxin contamination are combustion of municipal wastes; chemical industrial processes; metal smelting; the use of motor oil with chlorine-containing additives; paper bleaching; and continuing production of polychlorophenols, PCBs and chlorophenoxyacetic acid-based herbicides in some developing countries (107,112).
Routes of human exposure, metabolism and bioaccumulation
The major route of dioxin intake is dietary (up to 95%) and the average daily consumption of TCDD is estimated to be in the range of 30–200 pg in Europe (113) and 50 pg in the USA (114) resulting in average content in human tissues about 4–6 p.p.t. (115). Dioxin half-life in the human body is estimated to be 7–11 years (116,117).
Due to increasingly stringent regulations, dioxin emissions in the US decreased by ≈80% between 1987 and 1995 (118). At the same time, some individuals may be at unique risk to dioxin exposure. For example, in a study conducted in Sweden, it was found that people regularly eating fish in the diet showed a fourfold higher concentration of TCDD in blood serum compared with people eating fish occasionally (119). Workers in the textile, paper, leather industry, weed control, capacitor production and waste incineration may also be at high risk (110).
Measurement of TCDD content in adipose tissue of the general population in the USA (more than 800 people) showed that tissue TCDD residues increase with age (120). Levels in males significantly exceed those in females (121). Breast-feeding may be a significant source of TCDD exposure for newborns (122).
Dioxin toxicity: interspecies differences
2,3,7,8-Tetrachlorodibenzo-p-dioxin is among the most toxic substances known, at least for laboratory animals (20,123). At the same time, dioxin toxicity varies substantially in different species (Table 2). These differences impede the direct extrapolation of observations in laboratory animals to humans. Nevertheless, based on taxonomic consanguinity, it was proposed that dioxin lethal dose, 50% (LD50) for humans is comparable with the toxicity level identified in monkeys (70–280 μg/kg of body weight) (77), thus greatly exceeding the human toxicity of other poisons including sarin, cyanides, strychnine and curare.
Table 2. Dioxin LD50 values for different animal species (263 with additions)
LD50 (μg/kg body weight, oral)
Dioxin toxicity in animals has some peculiar features. One of them is its lengthy latency. Laboratory rodents die within 3 weeks of a lethal dose, whereas in larger animals the delay is greater, for example, up to 6 weeks in monkeys (124). Then, except for animals with severe liver necrosis (chicken and rabbits), the immediate cause of death cannot be attributed to a specific organ or system pathology and remains unclear for most tested animal species. Finally, the total dose of TCDD required to evoke a pathological response is less if the dose is administered over time when compared with a single dose.
Clinical symptoms of TCDD toxic action
In laboratory animals, TCDD causes hyperplasia of epithelial tissues including the skin, bile duct, digestive and genito-urinary tracts. It also results in generalized oedema of internal organs; hepatomegaly; thymus, bone marrow and testis atrophy; and immunosuppression (125–127). Exposed animals exhibit a wasting syndrome that resembles starvation or anorexia (128,129). It was found that TCDD is an active promoter of two-step carcinogenesis in rodent liver (130,131) and skin (132,133). Dioxin also induces embryotoxic and teratogenic effects in mice such as cleft palate and abnormalities of kidney development (128).
The human toxicity of TCDD is poorly understood. Some reports describe a broad spectrum of effects including emphysema, myocardial degeneration, toxic nephritis, hypertension, peripheral oedema, anorexia, gastritis, bursitis and intolerance to cold (30,43,134). A wide range of neurological symptoms have also been reported, including peripheral neuropathy, paraesthesias, headaches, vertigo, coordination disturbances, loss of libido, easy fatigability and emotional instability (135,136). The carcinogenic and teratogenic effects of TCDD documented in animals have not been verified in humans despite its designation by the International Agency for Research on Cancer as a class 1 carcinogen, meaning a ‘known human carcinogen’. At the same time, no case of lethal human poisoning with TCDD has been reported and most data on its human toxicity are speculative. The only specific, sensitive, consistent and long-lasting sign of human TCDD intoxication is chloracne (1,9,137).
The threat of TCDD for human health is beyond any doubt, but the level of its danger is still undefined (138). Further insight into cellular and molecular mechanisms of TCDD action in target tissues would provide better understanding of the potential risk of TCDD exposure.
Dioxin-induced molecular pathways
Canonical (AhR/ARNT) pathway. Many, if not all, of the effects of TCDD in humans and animals appear to be mediated by the AhR, a member of the basic-helix-loop-helix (bHLH) period-ARNT-single-minded (Per-ARNT-Sim) (PAS) family of dimeric transcription factors (139,140). This receptor was identified as a factor essential for cytochrome P4501A1-mediated induction of several liver enzymes by TCDD treatment (141). The subunit composition of the AhR took years to resolve, due to the low levels of receptor expression (e.g. 100–1000 fmol/mg protein) combined with its inherent instability (142). Nevertheless, extensive work from many laboratories has yielded a comprehensive model for TCDD–AhR signalling (Fig. 4).
According to this model, the unliganded AhR exists in the cytosol in a tetrameric complex, consisting of the AhR ligand-binding subunit, a dimer of the 90 kDa heat shock protein (Hsp90) (143,144), and the hepatitis B virus X-associated protein 2 (XAP2), an immunophilin-related protein (145). XAP2 stabilizes the AhR and plays a role in sequestering the non-liganded AhR in the cytoplasm by an as-yet uncharacterized mechanism (146). Hsp90 functions as a molecular chaperone and plays a key role in the proper folding and stability of the non-liganded AhR, thus maintaining a high-affinity ligand-binding conformation for the receptor and preventing nuclear translocation and/or dimerization (147–149). Another factor required for the stability of non-liganded AhR is p23 that is thought to stabilize the latent receptor–hsp90 heterocomplex, thus playing the role of a co-chaperone (150–152).
After passive diffusion into target cells, TCDD binds hsp90/XAP2/p23-associated AhR, thereby initiating receptor complex conformational transformation which results in a receptor species with increased affinity for DNA and a much slower rate of ligand dissociation (153). This event is associated with nuclear translocation of the activated complex during which hsp90 and the other proteins dissociate from the TCDD/AhR complex and are exchanged for ARNT. Heterodimerization of AhR with ARNT – another helix-loop-helix PAS protein which is a central dimerization partner in the family of PAS transcription factors (154) – results in the formation of a transcriptionally active dimer. The AhR–ARNT heterodimer recognizes a specific DNA sequence 5′-GCGTGNA/T-3′′, called the dioxin response element, located in the promoter region of certain genes, thus resulting in their transactivation (155). One such gene is CYP1A1, encoding the cytochrome P4501A1 protein, which has proven to be a useful biomarker of exposure to dioxin-type chemicals in many animal species (156). Using CYP1A1 as a model system, it has been shown that AhR/ARNT-induced transactivation depends on the activity of multiple coactivators that bind AhR and/or ARNT and provide a bridge between heterodimer and general transcription factors located in the promoter of the regulated gene. These cofactors include Sp1, nuclear coactivators (Ncoa1,2,3), coactivator receptor interacting protein (RIP)-140, retinoblastoma protein, Myb binding protein (Mybbp1a), promyelocytic leukemia (PML) and others (reviewed in Ref. 11).
In addition to Cyp1A1 induction, the AhR/ARNT heterodimer also mediates other biological effects of TCDD including teratogenesis, tumor promotion and immunosuppression mainly affecting skin, liver and the nervous system (132,139,157). The downstream target genes for AhR/ARNT heterodimers mediating these effects are mostly unknown.
Although originally thought to have a relatively narrow structural specificity (158), the AhR is now known to recognize a broad array of chemical structures, including non-aromatic and non-halogenated compounds (140). Nonetheless, it is paradoxical that the endogenous ligand for AhR is still undefined and its physiological functions other than xenobiotic metabolism still remain to be elucidated.
Substantial insight into AhR function was obtained by generating two Ahr ‘null’ mouse strains using a gene targeting strategy (159,160). Despite some minor phenotypic differences, these strains exhibit strikingly similar features including decreased liver size, subtle hepatic portal fibrosis, decreased constitutive expression of cytochrome P4501A2 and decreased body size during the first 4 weeks of life. Ahr-null mice are resistant to TCDD-induced toxicity, thus confirming the role of the Ahr gene in mediating TCDD-dependent pathways (161) and yet some TCDD effects seem to be Ahr independent (162,163). Most studies of Ahr-null mice have focused on their response to dioxins. However, these knockout experiments also reveal that the Ahr plays important roles during early development in such tissues as the liver, eye and kidney (164,165). Six- to 12-month-old Ahr-null mice also have progressive, focal lesions on the dorsal skin associated with alopecia and ulceration (164). The skin phenotype in Arnt-null mice is consistent with the finding that AhR is expressed in epidermal keratinocytes in a differentiation-associated manner (166).
The studies with recombinant inbred mice revealed four allelic variants of Ahr among laboratory strains. The highest susceptibility to toxic effects of TCDD segregates with the Ahrb1 allele, which occurs in the prototype ‘responsive’ strain, C57BL/6. ‘Non-responsive’ mice, such as the prototype DBA/2 strain, represent the Ahrd allele with a 10-fold lower affinity for TCDD compared with Ahrb1 allele and are much less sensitive to toxic and biochemical effects of TCDD action (167). Functional polymorphism was also shown for human AhR gene (168). Potential association of this polymorphism with individual differences in susceptibility to chloracne is still unknown.
Alternative molecular pathways of dioxin toxicity. The AhR–ARNT transcriptional activation pathway is now considered to be the primary mediator of TCDD biological action (reviewed in Refs 158,169). Nevertheless, none of the genes currently known as targets for the AhR/ARNT dimer can explain the pleiotropic effects of dioxin toxicity, including the skin symptoms. Thus, while there is strong experimental evidence for initial TCDD binding to the AhR, several alternative explanations exist for the subsequent biochemical events which lead to the pleiotropic toxic effects of dioxin (170).
Some believe that dioxin-induced toxicity is mediated by activation of the epidermal growth factor receptor (EGFR), based on the observation that dioxin effects resemble those of epidermal growth factor (EGF) such as early eye opening and incisor eruption, loss of body weight (‘wasting syndrome’), epithelial hyperplasia and retardation of hair growth (171–173). According to this concept, dioxin binding to AhR activates the protein tyrosine kinase pp60src (c-Src), which amplifies EGF signalling (174). This dioxin effect is thought to be AhR mediated but ARNT independent (175,176). It seems plausible that some elements of TCDD toxicity, particularly those related to the growth factor-like syndrome, could be mediated by AhR-induced activation of c-Src independently of ARNT (177). In this scheme, activated c-Src kinase is proposed to trigger various cellular responses, particularly those involved in the growth factor signal transduction pathway (178,179). The precise mechanism of dioxin signalling transduction by AhR to activate c-Src is not known.
Downstream molecular targets of TCDD-dependent pathways. While the initial molecular events evoked by TCDD are relatively well understood, the downstream effects and molecular targets of TCDD activity are not well defined. During the last decade, numerous studies have been performed using differential display in human and mouse hepatoma cells treated with TCDD or its antagonists (e.g. 180,181). These studies identified two major functional pathways: toxic and stress related as the main targets of TCDD activity, findings consistent with the classic (AhR/ARNT) mechanism of TCDD toxicity.
The first pathway includes a battery of phase I and phase II xenobiotic metabolizing enzymes, such as CYP1A1, CYP1A2 and CYP1B1 (encoding cytochromes P4501A1, -1A2 and -1B1, respectively), glutathione S-transferase, UDP-glucuronosyltransferase and NADPH:quinone oxidoreductase (157,182). All of these genes are induced by TCDD and other chloracnegens via AhR/ARNT and play an essential role in both metabolic activation and detoxification of TCDD and other xenobiotics including polycyclic aromatic hydrocarbons (183). Whether induction of P4501A1 and other detoxifying enzymes is causally linked to dioxin toxic effects remains unclear.
Another functional TCDD-dependent downstream pathway includes genes independent of detoxification and implicated in the control of stress response and developmental processes. Among them, plasminogen activator inhibitor-2 (PAI-2) (184,185) and a wide range of growth factors and their receptors (186) have been identified, thus further suggesting that AhR may have physiological roles independent of xenobiotic metabolism (187).
2,3,7,8-Tetrachlorodibenzo-p-dioxin-dependent growth factors include: (i) EGF which shows a significant decrease of binding capacity in the hepatic plasma membrane due to TCDD action (188); (ii) EGFR that is dramatically upregulated in palatal epithelium of TCDD-exposed mouse newborns resulting in a high incidence of cleft palate (189,190); (iii) transforming growth factor (TGF)-α which is increased in TCDD-treated primary human keratinocytes (186), but is actively downregulated by TCDD during mouse palatogenesis in vivo (190); (iv) TGF-β2 which is strongly suppressed by TCDD in vitro (191,192), but in vivo (embryogenesis) is up- or downregulated by TCDD depending on the state of fetal development (193); (v) TGF-β3 which is upregulated by TCDD in a dose-dependent manner in vitro (187); (iv) interleukin-1β (IL-1β) which is positively regulated by TCDD (184) and possesses a prominent antagonism on TCDD-induced effects (194); (7) tumor necrosis factor (TNF) which seems to mimic TCDD action (195,196) and is upregulated by TCDD (123,197), thus suggesting that acute toxicity of TCDD may be mediated by TNF (198).
The role of TCDD in cell cycle regulation through transactivation of p27 and jun-B as well as in apoptosis control through Bax transactivation is also established (157,199,200).
The studies of TCDD-dependent expression in vivo are quite limited and have been performed in only a few organs and tissues including liver, lung, thymus, spleen and the reproductive system of laboratory animals (e.g. 201). Despite their limited number, these studies have revealed several genes not previously shown by in vitro experiments. The major targets of TCDD activity in vivo were not only cytokines and growth factors like EGFR, TGF-β and different interleukins (201,202) but also genes implicated in apoptotic and angiogenic pathways. Among TCDD-regulated apoptotic genes are Fas ligand, caspases, several genes of Bcl-2 family and TNF-related apoptosis-inducing ligand, while angiogenesis-related genes include vascular endothelial growth factor (VEGF), the plasminogen activator cascade (PAI-1, PAI-2, PLAU and PLAUR) and angiogenin (201). In the reproductive system, dioxin-inducible factor-3 was identified as one of the major TCDD targets (203).
In summary, TCDD exposure results in altered expression of many growth factors, cytokines, apoptotic and angiogenic factors that could provide explanations for TCDD effects on proliferation and differentiation signals in the skin. At the same time, little evidence is available to link these expressional changes to the specific TCDD-induced skin pathology.
TCDD effects on epithelial cells in culture. The establishment of adequate in vitro experimental model is of paramount importance for elucidating the mechanisms of dioxin action in skin. A survey of 23 transformed human epithelial cell lines yielded none that could be used as a model to study TCDD-induced chloracne. All examined cell lines displayed no changes in morphology, viability or growth patterns following TCDD exposure despite profound upregulation of xenobiotic-metabolizing enzymes like CYP1A1 (204).
Studies of TCDD-treated primary human keratinocytes cultured in monolayer have shown equivocal results (205–208). No difference in sensitivity to dioxin was also found between primary keratinocytes isolated from HRS/J haired and hairless (hr/hr) mouse skin (209) that manifest strikingly different sensitivity to TCDD in vivo. Thus, traditional monolayer cultures lacking the structural and functional complexity of intact skin appear to be incapable of answering questions pertaining to chloracne. In contrast, by using a three-dimensional epithelial organotypic culture system, it was shown that 8 days’ postplating, dioxin-treated cultures possess a well-developed cornified layer that was absent in controls (210). Dioxin-treated cultures displayed a more flattened morphology specific for suprabasal keratinocytes, earlier onset of filaggrin expression and evidence of more active keratinization. Involucrin expression was evident in many basal keratinocytes, a feature never observed in control cultures (210). Thus, TCDD exerts profound effects on keratinocyte differentiation in vitro, but these effects require conditions or factors present in vivo or in organotypic culture and lacking in monolayer cultures. The potential role of such ‘epicellular factors’ in the maintenance of skin homeostasis is supported by a strong line of evidence. For example, studies using co-cultures of keratinocytes and dermal fibroblasts have clearly demonstrated active production of mutually inductive growth factors that impact proliferation in both epithelial and mesenchymal cells and modulate expression of structural proteins and epithelial architecture (211,212). The pivotal role of epithelial–mesenchymal communication in the control of HF cycling and hair growth is now an axiom (213,214).
The role of AhR in TCDD-induced changes in organotypic cultures of human skin keratinocytes is not well defined. CYP1A1 mRNA expression can be induced in an AhR-dependent manner in the absence of any xenobiotic, simply by altering the adhesive characteristics of the cells. A similar response can be achieved by suspending keratinocytes in semisolid medium in the absence of a xenobiotic (215). Suspension ablates both cell–cell and cell–substrate contacts and is a technique used to mimic the process of keratinocyte differentiation in which basal cells must modify their adhesive characteristics in order to migrate upward. Suspension culture triggers a variety of changes in gene expression parallel to those observed during keratinocyte differentiation. The novel observation that genes classically associated with xenobiotic-metabolizing enzymes can be induced in the absence of a xenobiotic simply by modifying cellular adhesive characteristics is critical because it suggests that AhR/ARNT signal transduction participates not only in detoxification but also in normal physiological pathways. This dual role for AhR signal transduction in skin was originally proposed soon after AhR was identified as the mediator of TCDD toxicity (216).
Dioxin effects on sebocytes are of particular interest because sebocyte differentiation is severely affected following dioxin exposure in vivo (31,87). It is known that TCDD abrogates adipogenesis in primary mouse embryo fibroblasts (217) and affects lipid metabolism in laboratory rodents and in humans (218,219). Nevertheless, the effects of TCDD on sebocyte differentiation in human skin remain to be defined.
Molecular targets for TCDD activity in the skin. The data on potential molecular targets of TCDD activity in skin in vivo are very limited. The exposure of mouse fetuses to TCDD caused premature expression of filaggrin and increased expression of involucrin at E16, thus suggesting that an AhR-dependent pathway is a modulator of terminal differentiation in fetal mouse skin (220). Topical application of TCDD to hr mice results in ectopic expression of epidermal-specific keratin 1 and loss of HF-specific keratin 17 in dermal cyst epithelia, suggesting a toxin-induced conversion of their differentiation programme from the ‘follicular’ to the ‘epidermal’ type (83).
Recently, it was shown that constitutive activation of AhR pathway is sufficient to trigger inflammatory skin lesions in mice with activation of IL-18 and chemokine C–C motif ligand 20 as the primary mechanism (221). As these effects were induced by ligand-independent upregulation of AhR expression, the possible role of TCDD in the control of these genes in vivo remains unknown.
Mechanisms of chloracne
Cellular targets of dioxin activity in the skin
The major effect induced by TCDD in different skin structures (Table 3) is alteration of cellularity that results in prominent hyperplastic or hypoplastic responses. Thus, different epithelial structures in the skin respond to TCDD differentially: epidermis and infundibulum (which is characterized by epidermal-like type of differentiation) undergo prominent hyperplasia; sebaceous and sweat glands lose their secretory activity and are replaced by keratinizing cells, while the lower portion of the follicle (hair bulb) gradually involutes. Each cellular target of dioxin activity in skin possesses one common feature – they all represent endpoints of distinctive differentiation programs characteristic of HF pluripotent stem cells or their immediate progeny – transit-amplifying cells. The balance between these differentiation pathways is tightly regulated by still poorly understood signal transduction pathways (e.g. 222–224). Hypoplasia of some skin epithelial structures along with prominent hyperplasia in others in response to dioxin suggests that the pathogenesis of TCDD-induced chloracne may involve alteration of stem cell homeostasis and lineage commitment. The following lines of evidence support this concept.
Table 3. Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) upon different epithelial structures in human and mouse skins
Infundibulum and sebaceous gland (SG) duct (upper hair follicle)
Hypoplasia, complete disappearance
Acrosyringial plugging; the palmoplantar hyperkeratotic lesions; hyperhidrosis
1Chloracne is a persistent process that remains years after chemical exposure (67). The involvement of long-lived stem cells in chloracne pathogenesis would provide an exhaustive explanation for the chronic course of this condition. The involvement of stem cells is also consistent with the delayed onset of chloracne after exposure to the inducing chemicals.
2The first signs of TCDD-induced hyperproliferation in human and mouse skin are visible in the lowermost part of the permanent portion of the HF (4,83,225) – the region where HF stem cells are located (226). Other TCDD-responsive cell compartments in skin including epidermis, SG duct, sweat gland duct and lower HF outer root sheath all harbor populations of cells with high proliferative potential – early transit-amplifying cells (227–229).
3It is a common view that chloracne is associated with alterations of sebocyte differentiation and the gradual transformation of these cells into keratinizing cells which results in squamous metaplasia of SGs (e.g. 30). But neither human nor animal (hr mice and rabbit ear) studies provide any morphological evidence for the transformation of sebocytes to keratinocytes. Instead, in chloracnegen-exposed skin, sebocytes differentiate normally. But once they complete their holocrine cycle, they are not replenished and are gradually replaced by keratinizing cells coming from the SG duct or outer root sheath (4,29,45,225).
4Dermal cysts, a specific feature of hr mouse skin, originate from cells representing remnants of the HF bulge. They express HF-specific keratins and are capable of sebaceous differentiation (86). TCDD treatment induces loss of HF-specific keratin 17 and ectopic expression of epidermal keratin 1 in dermal cyst epithelium, thus suggesting a switch from a pilosebaceous to an epidermal differentiation pattern in precursor cells (83). The resulting abrogation of sebocyte replenishment and their replacement with immigrating keratinocytes in hr mouse skin are identical to the process of comedo formation in human chloracne.
5The TCDD-dependent molecular pathway is relevant to haematopoietic stem cell (HSC) homeostasis and lineage commitment. Studies in C57BL/6J mice (230) show that HSCs are significantly elevated following TCDD treatment and demonstrate a marked shift in lineage commitment with reduced capacity of bone marrow to generate pro-T lymphocytes. AhR-dependent TCDD-induced proliferation of HSC (exit from stem cell compartment) and the shift in their differentiation lineage commitment are a major cause of TCDD-induced thymic atrophy (230,231) that occurs in many mammalian species (232). The cellular kinetics of TCDD-induced thymus atrophy is highly reminiscent of TCDD-induced transformation in utricle epithelium and SG of hr mice (83) and in pilosebaceous units in chloracne patients and in rabbit ear skin (4).
6It was suggested that AhR, like other bHLH-PAS family members, plays a role in the formation or maintenance of three-dimensional tissue structure in skin (210). On the other hand, three-dimensional structure of the epidermis and HF is an outcome of sophisticated sequence of epithelial–mesenchymal communications which in turn is a key mechanism of HF stem cell control. Nevertheless, direct evidence of AhR/ARNT pathway implication in epithelial–mesenchymal signalling is still lacking.
Based on the above evidence, we propose that chloracnegen-induced transformation of the pilosebaceous unit is driven by activation and accelerated exit of cells from the stem cell compartment coupled with a shift from a pilosebaceous differentiation pattern to an epidermal one. This may result in imbalance in early multipotent cells’ commitment and their preferential differentiation along an epidermal lineage with consequent diminution of SG and lower HF portion along with prominent epidermal/infundibular hyperplasia and hyperkeratinization. This model is highly consistent with the pattern of morphological changes in the skin of both chloracne patients (32,35) and animal models of chloracne (4,83,225). Hyperplasia of the infundibulum along with the switch of its content from semiliquid sebum to solid keratin could explain infundibulum dilation and comedo development. Similar mechanism is likely involved in TCDD-induced transformation of eccrine sweat glands which are also known to be a target of chloracnegen activity (35).
In summary, the multipotent HF stem cell concept (226,228) provides a comprehensive basis for understanding TCDD-induced changes in human and hr mouse skin and can help explain the histopathological differences between chloracne and acne vulgaris: while the latter is associated with exaggerated sebogenesis, chloracne is associated with redirection of pilosebaceous differentiation pathways to epidermal type, thus resulting in exaggerated keratinization and a switch from sebaceous to keratinized comedones.
Putative molecular mechanisms of chloracne development
The potential role of AhR/ARNT interactions in pathogenesis of TCDD-induced skin lesions. Despite clear evidence documenting the involvement of AhR/ARNT dimerization in TCDD toxicity in such target tissues as lungs, liver and thymus, the role of this dimer in mediating skin effects of TCDD action is still unclear.
First, the testing of different chloracnegens in hr mouse skin revealed no clear correlation between the level of chloracnegen affinity to the Ahr and the skin response. All tested compounds that did not bind the Ahr did not exhibit any skin effects. Nevertheless, some compounds with high affinity to the Ahr had little or no effect on hr mouse skin (233). Then, TCDD effects on potent regulators of epithelial homeostasis– the oncogenes c-fos and c-jun – are thought to be Ahr independent (163), suggesting that the skin effects of TCDD, at least in part, may also be Ahr independent.
On the other hand, in hr mice, the susceptibility to TCDD-induced epidermal hyperplasia strongly segregates with the AhRb1 allele (233) suggests direct involvement of AhR in the development of skin effects of TCDD toxicity at least in hr skin. The high level of AhR expression and its direct correlation with keratinocyte differentiation in human epidermis (166) support the possible involvement of AhR in chloracne pathogenesis also in humans. Yet, clear proof of this assumption is still lacking.
The crucial role of ARNT in several regulatory pathways implicated in response to the environmental insult (234–236) suggests that ARNT may be a key element of skin-environment communication, yet very little is known about the function(s) of ARNT in the skin (237). It was shown that ARNT is essential for such effects of TCDD action as thymic involution (238) and induction of xenobiotic-metabolizing enzymes including CYP1A1, CYP1A2 and UDP-glucuronosyltransferases (239), but its potential role in pathogenesis of TCDD-induced skin effects was never assessed. Nevertheless, essential role of many AhR/ARNT-dependent genes in the maintenance of skin homeostasis indirectly suggests that AhR/ARNT heterodimerization may be a prerequisite of chloracne pathogenesis as well.
Fig. 5 summarizes the potential mechanisms of AhR and ARNT involvement in the pathogenesis of chloracne. The downstream targets of AhR/ARNT activity include the complex of xenobiotic-metabolizing enzymes, PAI-2, IL-1β and other growth factors and cytokines (184,240) that play a role in the maintenance of skin homeostasis. TCDD-activated AhR may also compete for ARNT with other ARNT-binding factors (e.g. hypoxia inducible factor (Hif) 1α), thus reducing its availability for other interactions, not directly dependent on AhR activity. This pathway may suppress erythropoietin, VEGF, platelet-derived growth factor and glycolytic enzymes (155). The binding of TCDD with AhR may compete with as-yet unidentified endogenous AhR ligand. The ARNT-independent direct activation of the protein kinase cascade by TCDD/AhR complex (177) may also represent a molecular mechanism of chloracne induction because its downstream targets (e.g. EGFR, c-Myc, IL-1β, PAI-2, p53, MMP1, etc.) are known to be involved in regulating epithelial cell proliferation and differentiation (241–245).
Stem cell recruitment and lineage commitment as potential targets of TCDD activity in the skin. Among all existing attempts to provide an explanation for the puzzling features of chloracne, we wish to highlight the notion by Cunliffe et al. (6) who proposed that chloracne is determined by ‘the continuing modification of the metabolism of the pilosebaceous unit’ by the chloracnegen. Although this proposition was made long before the advent of the concept of HF stem cell, it invokes the existence of long-lived, stable cell population in the HF which can propagate the impact of the chloracnegen over several hair cycles (which normally last in humans about 3–5 years). Only stem cells fulfill these requirements. In itself, Cunliffe's prescient recognition of stable, long-living modification of the HF as the major cause of chloracne sheds little light on the pathomechanism of this disease. However, in conjunction with the HF stem cell concept, it provides a basis for a totally new approach to further explore the mechanisms of TCDD action upon skin.
According to the current concept, all epithelial structures in mammalian skin are maintained throughout the adult life by stem cells that not only generate daughter cells that undergo terminal differentiation along several pathways specific for every skin epithelial lineage but also self-renew (228,246). During the last decade, solid evidence has been obtained pointing at the bulge region of the HF as the primary niche for the skin stem cell population (247,248). There is reason to believe that the nature and location of stem cells may differ between HF types. Stem cells are normally quiescent or slow cycling [over 14 months in the mouse; (249)], but in response to certain stimuli like wounding or still mysterious hair cycle-inducing signals, they can be recruited into proliferation. Another set of still poorly understood signals determines the specific pattern of differentiation for resulting transit-amplifying cells. As a result, epidermal, SG or hair matrix progenitor cells are replenished by direct progeny of HF stem cell population (228).
As we mentioned before, pathogenesis of TCDD-induced chloracne suggests the following stem cell-associated events (Fig. 6): (i) recruitment of stem cells into cycling and their active exit into transit-amplifying cell compartment that probably determine hyperplasia in the middle portion of chloracne follicles; (ii) elevated level of stem cell renewal that maintains the sufficient number of stem cells even under their active exit into proliferating compartment, thus probably determining the persistence of chloracne; (iii) the shift in the fate of transit-amplifying cells – differentiation along epidermal pathway at the expense of HF and sebaceous differentiation.
Recently, it was shown that transient activation of a nuclear protooncogene c-Myc in mouse skin drives keratinocytes from the stem to the transit-amplifying compartment and stimulates differentiation along the epidermal and sebaceous lineage at the expense of hair-type differentiation (222,223,250). It was also found that the increased proliferation in the basal layer was not dependent on continuous activation of transgene. ‘Thus, transient activation of c-myc could result in sustained effects on the behaviour of keratinocytes’ (222,223). This model of c-Myc effects on HF stem cells is also supported by the finding that ablation of β-catenin, genetically upstream of c-Myc, prevents cycling of HFs by forcing stem cells into an epidermal fate (224).
Thus, the general scheme of c-Myc-dependent changes in the kinetics of HF stem cells fits the main features of TCDD-induced skin phenotype, while some noticeable differences also exist. For example, deregulation of c-Myc expression stimulates transit-amplifying cells to adopt epidermal and sebaceous fates in the expense of HF differentiation (222,250), although TCDD action stimulates epidermal differentiation at the expense of both, HF and sebaceous lineages. This difference suggests that if c-Myc pathway is a target of TCDD activity in the skin, its deregulation is not the only mechanism of chloracne pathogenesis. This proposition is also supported by partial, not complete, similarity between skin phenotype in transgenic mice overexpressing c-Myc (222) and chloracne-like symptoms in laboratory animals (83).
Currently, little is known about possible interactions between AhR/ARNT and c-Myc pathways. c-Myc is tightly regulated by growth factors and has been shown to be inhibited by the TGF-β2 signalling (251,252) which in turn is strongly suppressed by TCDD in vitro (191,192), suggesting potential positive influence of TCDD upon c-Myc expression. On the other hand, TCDD is a potent suppressor of c-Myc-induced apoptosis in Alb/c-Myc transgenic mice (253). In the liver of TCDD-treated guinea pigs, c-Myc DNA binding is downregulated (254). Treatment of 3T3-L1 preadipocyte cells in vitro (pretreated with TCDD) with antisense c-Myc oligonucleotides blocked the toxic effects of TCDD (255).
Another possible link between TCDD/AhR and c-Myc may be represented by c-Src/β-catenin pathway that plays an essential role in HF and skin biology (256). Although c-Src, a well-known target of TCDD action (177), does not directly participate in c-Myc phosphorylation, it phosphorylates tyrosine residues of β-catenin, leading to its extensive degradation (257). β-catenin, in turn, is a known upstream modulator of c-Myc expression (258–260). The role of β-catenin as a linker in the c-Src/c-myc signal transduction pathway in epithelial tissues is also confirmed by a significant increase in the tyrosine phosphorylation state of β-catenin in renal epithelial cells under c-Src activation (261).
Thus, existing data strongly suggest that c-Myc/Max and TCDD/AhR signalling pathways are linked together, but mechanisms of their interaction are still unclear.
According to our model of chloracne pathogenesis, TCDD not only induces the recruitment of stem cells into proliferation but also stimulates stem cell renewal, thus driving chloracne persistence. Although some genes involved in Wnt signalling, adhesion and transcriptional regulation have been proposed as possible candidates to play a role in self-renewal of HF stem cells (262), the mechanisms of this regulation remain largely unknown. Therefore, the possible role of TCDD/AhR pathway in the maintenance of stem cell population also remains obscure.
Thus, based on the data presented above and taking into account the scheme for the role of c-Myc in skin stem cell control (263), we propose that the primary mechanism of chloracne may be represented by TCDD-induced stimulation of keratinocytes to exit from the stem to the transit-amplifying cell compartment, resulting in hyperplasia. This exit is asymmetrical, with differentiation of resultant cells mainly along the epidermal lineage at the expense of pilosebaceous differentiation. We further suggest the AhR/c-Src/c-Myc cascade as the likely dioxin-induced signal transduction pathway leading to chloracne (Fig. 6). Predisposition of specific types of HFs to chloracne may relate to differential dynamics of stem cell activity in various types of follicles.
It is our hope that this model will provide a blueprint for future studies and that step-by-step testing of the assertions put forth in our review will lead to the disclosure of molecular and cellular pathways of chloracne development.
Chloracne is the most specific and sensitive biomarker of dioxin intoxication that allows clinical and epidemiological evaluation of exposure level at the threshold dose for humans. This is an extremely important fact in evaluating large groups of people exposed to unknown chemicals, or where the magnitude of exposure cannot be determined. Thus, the elucidation of cellular and molecular mechanisms of this clinical condition is essential for understanding and controlling the impact of polyhalogenated environmental toxicants on the general population and on specific risk groups like chemical workers and military forces.
Discovery of the genes directly implicated in the TCDD-dependent pathway of skin toxic response would have far-reaching implications for the development of new methods of chloracne differential diagnosis and therapy. This will also provide a solid basis for the identification of the genetic basis of differential human susceptibility to chloracnegen action and improvement of occupational health standards.
Furthermore, the studies of the role of AhR-dependent regulatory pathways in skin biology may open a new avenue in understanding the basic mechanisms maintaining skin stem cell homeostasis.
The minimal comedogenic topical dose of dioxin for the rabbit ear was estimated at 80 ng (80 × 10−9 g) (19,78).
The World Health Organization has determined that the Tolerable Daily Intake of dioxin is 1–4 pg/kg of total body weight per day.