Abstract
Medical case reports published in the 20th century over the course of several decades show that resorcinol caused reversible adverse effects on the human thyroid gland (TG) manifested as hypothyroidism. Affected patients had ulcerating leg varicosities and underwent prolonged treatment with ointments containing high concentrations of resorcinol. In animal studies resorcinol failed to induce TG toxicity, unless pharmacokinetic/toxicokinetic (PK/TK) conditions were manipulated (e.g., injection of resorcinol in oil or application in a slow release formulation). A recently completed two-generation reproductive toxicity study in rats did not detect any adverse effects on either reproductive or TG end points (Welsch, Nemec, and Lawrence, 2008, Int. J. Toxicol. 37, this issue). Resorcinol intake via drinking water up to the palatability limit had resulted in average daily intakes (mg/kg) of 233 in F0 and F1 males and 304 (premating/gestation) or 660 (lactation) in females. Free resorcinol in blood plasma was barely detectable in a few parental animals, indicating rapid metabolism. This short review communication offers a perspective on compromised human skin barrier function as a likely cause of drastic increases in resorcinol absorption. In conjunction with multiple daily applications over many months to hyperemic, inflamed, and lesioned human skin much higher absorption was likely responsible for the reported human TG toxicity.
INTRODUCTION AND HISTORICAL BACKGROUND
Resorcinol (1,3-benzenediol; m-hydroxybenzene; m-dihydroxyphenol; CAS no. 108-46-31) has been used therapeutically in human dermatology since late in the 19th century to treat ulcerating skin lesions. Adverse thyroid gland (TG) effects have been reported in clinical case histories over a span of many decades, first in the early years of the 20th century and continuing in irregular intervals until the 1970s (e.g., Berthezene et al. 1973). Toxic side effects on the TG, manifested by reversible goiters, occurred in some patients upon prolonged multiple daily applications of ointments containing high concentrations of resorcinol (Klem 1930; Strakosch 1943; Bull and Fraser 1950; Guinet, Tourniaire, and Peyrin 1967; Berthezene et al. 1973; Pascher 1978). In the past >30 years, no additional reports regarding resorcinol-related therapy for leg ulcers with associated TG toxicity have appeared in the open literature.
In the context of the “endocrine disrupters” debate that started in the early 1990s, and based on the long history of resorcinol-induced adverse human TG side effects, the European Union’s Committee of Toxicity, Ecotoxicity and the Environment (CSTEE 2003) endorsed a report putting resorcinol on a list of chemicals with endocrine disruption potential. Resorcinol is absorbed through intact and barrier compromised skin (Boeck 1915; Klem 1930; Berthezene et al. 1973; Bontemps et al. 1995) and disrupts thyroid hormone biosynthesis by inhibiting thyroid peroxidase (Divi and Doerge 1994). However, experimentally it has proven impossible to induce TG toxicity in several laboratory animal species by conventional dosing (references in Doniach and Logothetopoulos 1953). Neither did the subcutaneous (SC) injection of resorcinol in aqueous solutions cause detectable TG toxicity in rats (Doniach and Logothetopoulos 1953) nor did high-dose (maximum tolerated dose) gavage administration in lifetime carcinogenesis bioassays conducted in mice and rats (NTP 1992).
The present resorcinol usage is described in the preceding paper (Welsch, Nemec, and Lawrence 2008). Briefly, although there is large scale industrial use in tire and rubber manufactured goods, certain resins and resin adhesives, as well as in tanning and dyeing applications, comparatively minor quantities are incorporated into low resorcinol concentration containing medicinal creams to treat acne and into cosmetic products. The recently completed two-generation reproduction study (Welsch, Nemec, and Lawrence 2008) was conducted in rats because rats are generally recognized as a valid test species with respect to human TG toxicity hazard detection. Furthermore rats are more sensitive to TG function disruption by chemicals than humans are (Choksi et al. 2003; Jahnke et al. 2004).
The design of the two-generation reproduction study of resorcinol adhered to test guideline criteria (Organization for Economic Cooperation Development [OECD] 416; US Environmental Protection Agency Office of Prevention, Pesticides and Toxic Substances [US EPA OPPTS 870.3800]. In view of the history of adverse TG effects in humans with dermatological disorders and the CSTEE’s specific interest in “ endocrine disrupter” ramifications, the study design exceeded regulatory guidelines. A comprehensive assessment of TG end points was performed at the end of the study. Among them were (a) qualitative and quantitative micromorphology; (b) thyroid stimulating hormone (TSH); (c) triiodothyronine (T3) and thyroxine (T4) in the parental animals of either sex; and (d) some TSH, T3, and T4 measurements in offspring. The protracted mode of resorcinol intake via drinking water is most relevant to the potential exposure of the human population-at-large from contaminated drinking water. Some of the data were first reported at the annual meeting of the Society of Toxicology (Welsch, Nemec, and Lawrence 2006) and in more detail in the preceding paper (Welsch, Nemec, and Lawrence 2008).
In summary, the two-generation reproduction study outcome allowed the following conclusions: At the maximum palatable concentration of 3000 mg resorcinol/L drinking water no adverse effects on either reproduction or TG end points were detectable. At that concentration the average daily intake in males of both generations was ∼233 mg/kg, whereas in females daily intake ranged from 304 mg/kg/day during premating and gestation to ∼660 mg/kg/day during lactation. There were no notable signs of toxicity in either the F1 or F2 parents or in their offspring. Regardless of the high daily intake of resorcinol there were no adverse TG effects, a finding that concurs with that of previously conducted animal experiments. This includes rats receiving suncutaneous (SC) injections (Doniach and Logothetopoulos 1953) and gavage bolus dosing in lifetime carcinogenesis bioassays in mice and rats (NTP 1992). The preceding studies did not include measurements of either pharmacokinetics/toxicokinetics (PK/TK) or the phamacodynamics/toxicodynamics (PD/TD) of resorcinol. However, considerations regarding rapid metabolism were included in the interpretation of the experimental observations (Doniach and Logothetopoulos 1953). Metabolic disposition studies conducted several decades later confirmed that 14C-radiolabeled resorcinol was readily absorbed from the gastrointestinal tract and efficiently metabolized. The resulting 14C-labeled products were predominantly excreted in urine as glucuronides (∼70%) and as sulfate conjugates (Kim and Matthews 1987).
RESULTS AND DISCUSSION
In pilot experiments (which were part of a full two-generation reproduction study with expanded TG end points), blood samples from the drinking water study underwent chemical analyses with a sensitive method (Welsch, Nemec, and Lawrence 2008). In a laboratory environment with 12-h light/12-h dark photoperi-ods, rats consume their daily drinking water intake in characteristic patterns (Spiteri 1982; Johnson and Johnson 1990). During darkness there are two episodes of intense drinking behavior, with the second spike of water consumption occurring during the last ∼2 h of the dark phase, which in the two-generation study was from 0400 to 0600 h. To enhance the chances of detecting free resorcinol, blood samples from F1 males and females were obtained as soon as logistically feasible once the lights were turned on at 0601 h. The blood collection time was generally within 60 min after the lights came on. Free resorcinol was detectable solely in blood of either gender in the 3000 mg/L groups. However, only 3 of 20 blood samples (10 females and 10 males) contained very low concentrations (Welsch, Nemec, and Lawrence 2008). Although these data confirmed that resorcinol was readily absorbed from the gastrointestinal as previously reported (Kim and Matthews 1987), it was also apparent that the chemical was rapidly removed from the systemic circulation. This was likely due to efficient metabolism and renal excretion of various detoxified products (Kim and Matthews 1987).
Animal Studies, Human Clinical Case Reports, and a Perspective on Absorption and Fate of Resorcinol
The outcome of the pilot PK/TK measurements (Welsch, Nemec, and Lawrence 2008) showed that about 1 h after the last burst in water consumption during the dark cycle, only 1 of 10 males and 2 of 10 females had very low free resorcinol concentrations detectable in blood. These recent results should be put into perspective and compared with earlier data of no detectable TG effects in other animal studies and with the human clinical toxicity case reports in the medical literature. The latter described TG toxicity upon prolonged use of resorcinol containing ointments among a small number of patients with skin lesions in the application area.
The review comments offered here will now consider plausible reasons as to why clinical case publications reported that resorcinol caused human TG toxicity in patients with impaired skin barrier function. Furthermore, the failure to induce adverse TG effects in rats treated with resorcinol in aqueous solutions by multiple routes and modes of application (e.g., bolus oral intubation [gavage], SC injection, or drinking water) will be contrasted with TG toxicity that was elicited in animals upon resorcinol treatments in either oily solution or in an esterified, protracted release, formulation (Doniach and Logothetopoulos 1953).
Therapeutic Use of Resorcinol and Adverse TG Reactions in Humans
Resorcinol is a polar chemical with high water solubility. Its dermal diffusion is retarded by a lipid phase barrier of high viscosity in the stratum corneum of the skin (Roberts et al. 1978). Hydrophilic chemicals permeate the intact human skin less readily than those that are hydrophobic. The medical literature cited above indicates that the high concentration of resorcinol in formulations prepared for topical use on human skin with inflamed and ulcerating lesions appears to be related to the likelihood of induction of either acute (target organ central nervous system) or chronic (target organ TG) toxic responses. However, upon subchronic and chronic applications onto diseased human skin, even lower concentrations in ointments have caused reversible TG side effects. Those adverse actions in humans manifested themselves in chronic myxedema, a clinical condition indicative of perturbed TG homeostasis, reduced thyroid endocrine function, and hypothyroidism.
Advances in medical diagnostic methods led to a mechanistic understanding as to how resorcinol affected human TG function. It appeared likely that iodine uptake was profoundly affected (Bull and Fraser 1950). The clinical histories of several female patients who displayed classical symptoms of hypothyroidism caused by resorcinol therapy were described. The common link was that the patients had suffered from varicose vein leg ulcers for a long time. Their skin lesions had been symptomatically treated for many months or even years with ointments containing 4% to 12% resorcinol. Cessation of resorcinol therapy led to a fairly rapid reversal of the clinical symptoms of hypothyroidism. Bull and Fraser (1950) therefore cautioned against the indiscriminate application of resorcinol ointments to lesioned skin. Regardless of these explicit warnings therapeutic applications to manage chronic leg ulcers continued as did clinical case reports regarding the occurrence of hyperplastic parenchymal goiters with discrete symptoms of hypothyroidism (Pascher 1978). TG toxicity occurred with long durations of treatment, such as upon daily application of 2% resorcinol ointments on bilateral leg ulcers for more than 1 year (Guinet, Tourniaire, and Peyrin 1967). The goiters showed gradual onset over several months. The clinical symptoms dissipated when resorcinol was discontinued and thyroid hormone replacement therapy commenced. Additional case reports from female patients receiving resorcinol therapy for varicose vein leg ulcers revealed hypothyroidism after years of continuous resorcinol treatment (Berthezene et al. 1973).
The underlying disease conditions indicate that the integrity of the human skin barrier was impaired because the clinical case reports reveal that there were ulcerating skin lesions, inflammation, hyperemia, etc. Regrettably, there are no resorcinol PK data available from any of the patients affected with adverse TG side effects.
Even without objective PK data in hand, one can deduce by analogy to the dermal penetration of other drugs/chemicals that the rate of resorcinol absorption from diseased human skin, and thus the PK and PD, would be profoundly affected and very likely increased. The resorcinol concentrations in the applied ointments were variable, as were both frequency as well as duration of treatment. Furthermore, the size of the skin lesions was not described. Therefore one cannot even speculate on either daily doses or any dose-response relationships of resorcinol.
Percutaneous Absorption of Resorcinol in Human Volunteers
One (comparatively speaking more recent) study explored the safety of resorcinol-containing acne medications in four male human volunteers who had intact, healthy skin. A solution containing 2% resorcinol was applied twice a day to three of the subjects on about 30% of their total body skin area. The experimental design aimed at exceeding maximal acne treatment levels at least 10-fold. The daily doses were 60 times higher than during typical acne treatments (Yeung et al. 1983). The data collected included back calculations from total urinary elimination of resorcinol metabolites and flux rates through healthy human skin (0.37 μg/cm2/h). Those rates were compared to in vitro measurements of skin penetration through healthy skin samples from surgical excision specimens. The flux rates (0.86 μg/cm2/h) were compatible with those calculated when the exposure concentrations of resorcinol/cm2 of skin were determined (Yeung et al. 1983).
In the same study, absorption and metabolic disposition were measured with exposures lasting up to 4 weeks. Blood samples were collected after 1, 2, 3, and 4 weeks and analyzed for free and esterified resorcinol (glucuronidation and sulfation products). No free resorcinol was detectable at all in samples undergoing either HPLC or GC/MS analyses. The functional status of the TG was measured via TSH, T3, and T4 determinations and remained unchanged during resorcinol exposure. The authors concluded that their studies, using highly exaggerated exposure conditions, demonstrate the safety of topical ointments since in general use, as regards over-the-counter availability, acne ointments are formulated and administered at <2% resorcinol content with explicit use guidelines (Yeung et al. 1983). Human skin absorption from other cosmetic formulations through intact skin is negligible (Cosmetology Report 1993). Both studies conducted on healthy human skin indicate very low absorption rates.
Human Skin Barrier Functions
Experimental and clinical observations indicate that the permeability characteristics of intact normal human skin are very different from those of diseased skin with impaired barrier function. Inflammation and associated hyperemia profoundly enhance drug absorption (Cohen and Rice 2001; Maibach and Patrick 2001). One elegantly designed study has demonstrated how the percutaneous penetration of topically applied drugs is drastically increased when the human skin barrier function is experimentally impaired (Benfeldt, Serup, and Menne 1999). Quantitative data regarding percutaneous salicylic acid (SA) penetration were obtained under well-controlled conditions. The experimental set-up allowed using each of eighteen human subjects as its own control. The test drug, SA, was the compound that was judged to be comparable to resorcinol in its keratolytic properties soon after resorcinol was first introduced into medicinal use (Strakosch 1943). The penetration kinetics of SA under various barrier disruption intensities were compared with a dermal microdialysis technique. The measurements allowed to compare the in vivo PK of SA and relate the data to a noninvasive graded disruption of human skin barrier functions (Benfeldt, Serup, and Menne 1999). Three different intensities of skin barrier perturbation were created by test region specific treatments in separate small areas of the forearm in the same subject. The extent of barrier impairment reached from mild (acetone treatment) to moderate (1% detergent treatment with sodium lauryl sulphate) to severe (mechanical tape stripping combined with 2% detergent). Compared with normal, healthy human skin, the mean SA penetration increase was 2.2-fold in acetone-treated skin, 46-fold in mild dermatitis, and ∼150-fold in severe dermatitis (Benfeldt, Serup, and Menne 1999).
By analogy, one can deduce that in the reported clinical case reports of human TG toxicity skin barrier impairment caused by inflammation and hyperemia is highly likely to have increased the absorption of resorcinol. It appears reasonable to assume that increased skin absorption caused by the combination of barrier impairment and multiple daily administrations of resorcinol-containing ointments created PK and PD conditions that were responsible for significant increases in resorcinol absorption. Thus free resorcinol reached the TG via the systemic circulation in concentrations sufficient to gradually cause TG toxicity in a concentration-over-time of exposure-related fashion.
Systemic Bioavailability of Resorcinol in Relationship to Route and Mode of Exposure
The interpretation offered above to explain the human TG toxicity observed in clinical case reports is supported by animal experiments. It required special resorcinol formulations to create PK/PD and TK/TD exposure conditions that allowed TG toxicity to manifest itself. To wit, resorcinol dosing, by either oral bolus administrations (NTP 1992) or by SC injections in aqueous solution, causes no TG toxicity in rats (Doniach and Logothetopoulos 1953), rabbits, or mice (see original references in Doniach and Logothetopoulos 1953). However, when resorcinol is dissolved in oil and injected SC to rats twice daily for variable lengths of time up to 69 days, TG fresh weights increase and TG histology reveals hyperplasia. Other rat experiments with resorcinol diacetate, a slow-release formulation of resorcinol, also causes adverse TG effects. Although, regrettably, no concurrent PK measurements were conducted, it was concluded that rapid metabolism of resorcinol was responsible for the failure to elicit TG toxicity (Doniach and Logothetopoulos 1953). Thus, it appears that without applying methods affecting the PK of resorcinol, the free resorcinol concentrations that remain bioavailable for systemic distribution are neither high enough nor sufficiently high over time to reach thyrotoxic concentrations.
Measurements of the PK of 14C-labeled resorcinol following SC injections in aqueous solution by several dosing regimens and sample collection times revealed that free resorcinol is not detectable by HPLC and that radioactivity is rapidly eliminated in the urine. Final confirmation of the molecular structures of the metabolites following enzymatic hydrolysis with glucuronidase was performed by gas chromatography/mass spectrometry (GC/MS) (Merker et al. 1982). Other studies regarding the metabolic disposition of radioactive resorcinol, given to rats of either gender by oral bolus dosing at doses of 112 or 225 mg/kg, show that the chemical is readily absorbed and efficiently converted to water-soluble products, which are primarily eliminated via urine (Kim and Matthews 1987). In both sexes, > 90% of the radioactivity is excreted via the kidneys in 24 h. The most prominent disposition pathways involve glucuronide (70%) and sulfate conjugates (Kim and Matthews 1987). It is reasonable to anticipate that route and mode of entry profoundly affect PK/TK of a chemical. In the recently completed two-generation reproduction study (Welsch, Nemec, and Lawrence 2008), resorcinol was ingested via the drinking water, leading to a protracted daily intake related to a well-defined pattern of drinking behavior. Thus, in spite of very high total daily resorcinol doses, the internal peak resorcinol concentration delivered to potential target organs, such as the central nervous system or the TG, should be lower compared to oral bolus or SC injection dosing. Relatively speaking, the dose is low at any given time compared to single-bolus administrations in aqueous solutions. In fact, the pilot bioanalytical measurements (Welsch, Nemec, and Lawrence 2008) confirmed that free resorcinol was generally no longer detectable about 1 h after the last water consumption. It appears likely that efficient metabolism, possibly by first pass through the liver, removed free resorcinol. Therefore, the TG was not exposed to thyrotoxic resorcinol concentrations. Overall, this outcome adds to the weight of the evidence that PK/TK, metabolism, and bioavailability of free resorcinol for delivery to the critical target organ are key determinants for TG toxicity manifestations.
CONCLUSIONS
No experimental studies similar to those conducted by Benfeldt, Serup, and Menne (1999) with SA have been performed with resorcinol in human subjects. However, the findings of the SA penetration measurements appear to be highly relevant to resorcinol. The SA data provide quantitative PK confirmation for the long-held insight that barrier impaired skin is more permeable to drugs and chemicals than intact skin is. The SA skin penetration observations may be generally applicable to hydrophilic agents. Skin absorption differences may thus be profound and penetration may be up to ∼150-fold higher in severely barrier function impaired human skin. One can conclude by analogy to SA that the human myxedema and hypothyroidism occurring in patients with skin ulcers were caused by high rates of absorption of resorcinol through the barrier impaired skin. The disease conditions allowed sufficiently high concentrations of free resorcinol to reach the human TG and exert thyrotoxic effects. On the other hand, rapid metabolism in animals, and rats in particular, precludes resorcinol from reaching TG toxic concentrations. Even if high doses are administered by bolus, the PK and rapid metabolism of resorcinol do not allow resorcinol concentrations in the systemic circulation to reach levels that can cause TG toxicity.