Registration Dossier

Administrative data

Link to relevant study record(s)

Description of key information

After oral gavage to rats, Hydroquinone (HQ) is rapidly absorbed from the GI tract, followed by rapid excretion of metabolites, mainly in urine. Elimination from blood was biphasic, attributable to first-pass metabolism followed by enterohepatic cycling of a small fraction of the dose. Signs of saturation of elimination were seen at 350 mg/kg. A rapid absorption (5-10 sec) is also seen in intratracheal dosing. Despite little or no pulmonary metabolic activity, blood elimination is also rapid, and biphasic, with signs of saturation at 10 mg/kg. Dermal absorption is much slower. All three routes (oral, intratracheal, dermal) result primarily in biotransformation to HQ-glucuronides and sulfates, and only minor formation of glutathionyl (GS), cysteinyl and mercapturic acid conjugates. 95% of oral doses up to 200 mg/kg were recovered in urine and feces within 24-72 hrs. Repeated dosing with 200 mg/kg/d gave no indication of bioaccumulation. Intraperitoneal application results in bypassing the glucuronidation and sulfation detoxification pathways, and leads to higher levels of glutathionyl adducts (mono-, di-, tri-, and tetra-GS-HQ). PBPK simulations predict more GS-HQ in F344 rats compared to Sprague-Dawley, and higher levels after IP injection compared to oral dose. Compared to rats, humans have a greater capacity for hepatic detoxification, through HQ-glucuronide and mercapturic acid conjugates (N-AcCysS-HQ), limiting exposure of kidneys to nephrotoxic metabolites. Whereas deactivation steps predominate in human liver cells, bioactivation steps predominate in rat liver cells leading to higher body burdens of multi-substituted GS-HQ conjugates considered key nephrotoxic metabolites. There is evidence that toxic effects observed in the F344 rat, and genotoxic findings in IP studies in mice may be of limited biological relevance for human. PBPK models support route-to-route extrapolation using the findings in oral studies as a worst case approach for risk assessment.

Key value for chemical safety assessment

Bioaccumulation potential:
no bioaccumulation potential
Absorption rate - oral (%):
Absorption rate - dermal (%):
Absorption rate - inhalation (%):

Additional information

The toxicokinetics, metabolism and distribution of HQ have been comprehensively investigated in several key studies. The significant differences existing between animal species as well as between animals versus humans, and between strains of rats are crucial for the interpretation of the biological relevance of findings in animal studies, e.g. on repeated dose toxicity or genotoxicity, as well as for human risk assessment. Moreover, based on the findings of excretion of HQ and metabolites in urine of test persons without HQ exposure, or on background levels of HQ-derived protein-S-adducts in tissues of rats and mice, a considerable background exposure to HQ exists both in humans (Key studies: Deisinger, 1996; Deisinger et al., 1994) and in test animals (Key studies: Boatman et al., 1994, 2000a, b). This background exposure originates from dietary sources, from endogenous production, and from further uncharacterised sources.


In vivo toxicokinetic studies

Studies with single dosing

Oral administration

A toxicokinetic study, with a protocol comparable to OECD Guideline 417, was performed with male and female F344 rats with gavage application of either a single dose of 25, 50 or 350 mg/kg bw 14C-labeled HQ (LDS, MDS, HDS) or a dose of 25 mg/kg bw/d 14C-HQ after a 14-day pre-treatment with unlabeled HQ (LDR). This dose range covered both the dosages of the lifetime bioassay as well as an acutely toxic dose. After gavage application, HQ was rapidly absorbed from the gastro-intestinal tract into the blood, followed by extensive first-pass metabolism and rapid excretion primarily via the urine. Maximum blood concentrations of total 14C were attained within 20 min and 14 min after single or repeated low doses of 25 mg/kg (LDS, LDR), respectively, and within 50 min after a single dose of 350 mg/kg (HDS). Parent compound represented less than 1% of total14C in blood. Elimination of14C from blood was biphasic. Half times for the first fast α-elimination phase ranged from 0.23-0.29, 0.58-0.91, and 1.35-1.72 h for the LDS, LDR, and HDS groups, half times of the slower ß elimination phase were 2.8-10.5 h for the low dose. Saturation of elimination was indicated at the high dose. Elimination was mainly associated with the α phase which presumably represented extensive first pass metabolism, mainly by phase II biotransformation to readily excretable glucuronides and sulfates. The ß elimination phase may be attributable to enterohepatic cycling of a small fraction of the dose.

Excretion was via the urine and faeces, primarily within the first 8 h after gavage. Typically, 87-94% of the administered dose was excreted in urine, and 0.9-2.0% in the faeces. In urine, less than 3% of the applied14C dose was excreted as parent compound, while 45-53% and 19-33% were excreted as HQ-glucuronide and HQ-sulfate besides the HQ mercapturic acid conjugate (less than 5% of14C). No marked sex-differences in the metabolism and disposition of HQ were found (Key studies: English et al., 1988, 1991, 2005).


The data of the guideline study are supported by further studies:

Very similar data were observed in male or female F344 rats (Lockhart et al., 1984; Lockhart and Fox, 1985) and male Sprague-Dawley rats (Divincenzo et al., 1979, 1984). Single gavage doses of 5 to up to 200 mg14C-labeled HQ/kg bw were rapidly absorbed followed by excretion within 24-48 hrs. About 93% of applied14C was found in urine, 1.2-3.8% in faeces, up to 0.4% in expired air, and up to 1.6 % in carcass and tissues, with highest amounts in liver (up to 0.6%) and kidneys and intestines (up to 0.04% each). In the investigated dose range, there was no dose-dependence of tissue distribution, metabolism and excretion.

Overall, the high proportion of HQ in urinary excretion in these studies indicate a high level of absorption by the oral route, and as at least part of the feces excretion could also be related to enterohepatic circulation, an absorption of 100% is thus considered for the purpose of the route-to-route extrapolation in the risk assessment.

Intratracheal administration of 14C-HQ/kg bw at doses of 0.1, 1.0 or 10 mg/kg in male Sprague-Dawley rats resulted in a very rapid absorption and systemic distribution of free HQ in arterial and venous blood, within 5 -10 seconds, to decline quickly within 12 minutes indicating a rapid metabolisation with biphasic elimination of the parent compound from blood. The absence of metabolite in the initial arterial blood samples indicated the absence of pulmonary metabolic activity. Reversibly bound HQ was also detected quickly after administration. The estimated blood half-lifes were 194 sec at 0.1 mg/kg, 74 sec at 1.0 mg/kg, and 218 sec at 10 mg/kg (Deisinger and English, 1999).

Comparing exposure of male F344 rats by oral gavage or intratracheal instillation, a dose of 50 mg14C-HQ/kg bw was rapidly absorbed from the lung and stomach with maximal blood levels observed at 2 and 7.5 min post application, respectively. Distribution equilibrium was attained within 1.5-2.5 hr, and mean half-times for distribution and elimination were similar for oral, intratracheal or intravenous application with a mean value of 19.3 min and 7.28 h, respectively. Blood clearance of HQ and metabolites was nearly complete (95%) by 2 hr, and only a small portion of14C (< 10 µg-equivalent HQ/g) was found to be associated with blood cells and plasma proteins (measured after oral application) (Supporting study: Fox et al., 1986). It seems that intraperitoneal administration might result in a slightly slower absorption, as evidenced by differences in the Tmax in veinous blood with 5 min in mice administered 75 mg/kg ip (Legathe, 1994), acompared to a Tmax achieved in less than 1 min by intratracheal administration in rats (Deisinger and English, 1999; Corley et al., 2000). Other differences include a longer blood half-life reported for the intraperitoneal injection (T1/2: 9 +/- 2 min) compared to intratracheal dosing (T1/2 < 4 min).

Studies with repeated application

Repeated dosing of 25 mg/kg bw/d did not result in tissue accumulation of HQ-derived material. About 90-95% of applied14C-HQ was recovered from excreta and cage rinsings, and less then 1% was recovered in the tissues and carcass up to 72 hrs following application (Key studies: English et al., 1988, 1991, 2005).

Following repeated application to male Sprague-Dawley rats (5 x 200 mg/kg bw/d) there was a shift in the metabolite profile in urine compared to single application. Excretion of HQ-glucuronide increased to 72.2% vs. 56.4%, while excretion of HQ-monosulfate was lower (23.2% vs. 42.3%). There was no indication of a bioaccumulation potential or an induction of monooxygenase dependent metabolism by repeated application of HQ (Supporting study: Divincenzo et al., 1979, 1984).


Based on amounts of14C recovered in tissues after single or repeated oral dosing (up to 200 mg/kg bw) there are no indications of accumulation in a specific tissue. The amounts of total14C recovered in tissues and carcass of Fischer or Sprague-Dawley rats at 24 -72 hrs following oral application ranged from 0.015 - 1.6%, with highest amounts found in liver (up to 0.6%) (Key studies: English et al., 1988, 1991, 2005; Supporting studies: Divincenzo et al., 1979, 1984; Lockhart et al., 1984; Lockhart and Fox, 1985).

These in vivo observations are confirmed by in vitro determinations of protein binding and tissue/blood partition coefficients in tissues of male F344 and Sprague-Dawley rats. Partition coefficients were similar for different tissues (range 0.62 – 0.85). In tissues, free HQ amounted to ca. 70 – 90%, irreversible and reversible protein binding of HQ to ca. 2-3% and 10-27%, respectively. In plasma, the corresponding values were ca. 40% free HQ, ca. 3% irreversibly bound, and ca. 57% reversibly bound HQ. Irreversible binding of14C-HQ was highest in the erythrocyte fraction of whole blood (ca. 17–20%), possibly due to interactions with hemoglobin (Key studies: Hill et al., 1996; Fiorica et al., 1996; Corley et al., 2000).

In vivo and in vitro metabolism studies

Oral application

The dominant metabolites excreted in urine of rats are the HQ-glucuronide and HQ-sulfate besides the HQ mercapturic acid conjugate. The latter indicates the intermediate formation of a sulfhydryl reactive metabolite and subsequent conjugation with glutathione (Key studies: English et al., 1988, 1991, 2005; Supporting studies: Divincenzo et al., 1979, 1984; Lockhart et al., 1984; Lockhart and Fox, 1985) (see also attached document with metabolism scheme adopted from WHO, 1994). In plasma, a glutathione conjugate (2-(5-glutathionyl)HQ) was identified as minor metabolite besides the HQ-glucuronide (Supporting study: Fox et al., 1986). A possible sex difference in metabolism was indicated by urinary excretion profiles (values of male vs. female F344 rats) with 61.8-63.9% vs. 43.4-49.0% as HQ-glucuronide, 19.1-26.2% vs. 28.8-35.9% as HQ-sulfate, and 0.6-1.4% vs. 2.1-3.0% as free HQ (Lockhart et al., 1984; Lockhart and Fox, 1985).

Sulfhydryl-bound HQ was quantitated in kidneys, livers, spleens and 24-hour urine in groups of male and female F344 rats and of male Sprague-Dawley rats dosed with 0, 25, 50 or 100 mg HQ/kg bw/d by single or 6 weeks (5 d/w) gavage or by single intraperitoneal injection (only done with SD rats). 24 hrs after the last administration, tissues were assayed for the presence of mono-, di- (2,3-di-, 2,5-di-, 2,6-di-), tri- and tetra-substituted HQ-protein-S-adducts.

A gender- and species specifity was detected with regard to total protein adduct levels in kidneys, blood and liver being higher in female F344 rats vs. male F344 or SD rats. In all examined tissues, the amount of total protein-S adducts increased with increasing dose. The relative proportions of the individual adducts were found to be strongly dependent on tissue type reflecting tissue-specific metabolism pathways. In blood protein primarily mono-adducts were detected. High levels of mono-protein-S-adducts in liver (at least 72% of total adducts) suggest that the liver is primarily responsible for the conversion of parent HQ to metabolites derived from oxidation of benzoquinone (BQ). Despite high levels of total protein adducts (up to 1003 pmol/mg protein) there were no signs of hepatotoxicity. In contrast, in the kidneys the high proportion of tri- and tetra-substituted adducts (at least 60%) suggests further oxidation and conjugation steps taking place in this organ. Alternatively, extra-renal conversion of glutathionyl adducts, initially formed in the liver and possibly in the blood, may undergo further bioconversion within the kidney. In urinary protein, only di- and tri-protein-S-adducts were detected with 2-fold total adduct levels in male F344 rats compared to SD rats (about 450 vs. 230 pmol/mg protein).

Elevated protein-S-adduct levels in kidneys from female F344 rats versus levels in male F344 or SD rats are consistent with the differences in the acute nephrotoxicity of HQ seen in these strains and genders of rats in former studies. In addition, increased levels of the tri- and tetra-substituted protein-S-adducts found in kidneys are consistent with the increased nephrotoxicity of the di- and tri-glutathione conjugates of HQ, which may represent the proximate renal toxicants. However, after repeated oral application of HQ there was no correlation between significantly elevated protein-S-adduct levels and measured biochemical indices of nephrotoxicity (see IUCLID-Section 7.9.3) (Key studies: Boatman et al., 1994, 2000a, b).

Intraperitoneal application

HQ (198 mg/kg bw) was administered by intraperitoneal injection to Sprague-Dawley rats after pre-treatment with avicin to inhibit gamma-glutamyl transpeptidase. Five S-conjugates were identified in bile (2.8% of the total dose), and one S-conjugate in urine (1.5% of the dose). The major biliary S-conjugate identified was 2-glutathion-S-ylhydroquinone [2-(GSyl)HQ] (2.3% of the dose), besides 2,5-diglutathion-S-ylhydroquinone [2,5-(diGSyl)HQ], 2,6-diglutathion-S-ylhydroquinone [2,6-(diGSyl)HQ], 2,3,5-triglutathion-S-ylhydroquinone [2,3,5-(triGSyl)HQ] and 2-(cystein-S-glycyl)hydroquinone. 2-(N-acetylcystein-S-yl)hydroquinone was the only urinary thioether metabolite identified. 2-(GSyl)HQ, 2,5-(diGSyl)HQ, or 2,6-(diGSyl)HQ were formed in vitro by the NADPH-dependent oxidation of HQ in the presence of rat liver microsomes and glutathione. The quantity of S-conjugates excreted in urine and bile appeared sufficient to propose a role of such metabolites in HQ-mediated nephrotoxicity and nephro-carcinogenicity (Key study: Hill et al., 1993).

Administration of HQ to male Sprague-Dawley rats at a single dose of 100 mg HQ/kg bw produced significantly higher (5-10 fold) levels of total protein-S-adducts in most tissues (e.g. blood, kidney, spleen) when comparing intraperitoneal to oral (gavage) application. The greatest increases were seen in blood and spleen. However, the relative proportions of individual adducts were not changed by the route of application. Obviously, i.p. application results in bypassing of the detoxification pathways of glucuronidation and sulfation, which are present in the gut and liver, leading to a higher production of glutathionyl adducts (mono-, di-, tri-, and tetra-GS-HQ adducts) as active precursors for protein binding. Otherwise, as adduct levels in livers were only up to 2-fold higher after i.p. vs. oral application, extrahepatic activation is likely to account for the majority of the protein binding seen in kidneys, spleen, and blood (Key study: Boatman et al., 1994, 2000a, b; Corley et al., 2000).

As a consequence, a different spectrum of metabolites is available after intraperitoneal application than after oral application presumably with a higher amount of reactive metabolites, which is of crucial influence on the outcome of toxicity testing, as e.g. in vivo genotoxicity assays (see Section 7.6.2). Consequently, the biological relevance of findings from animal assays with intraperitoneal application is highly questionable for the human exposure situation.

Intratracheal administration

Intratracheal administration of14C-HQ/kg bw at doses of 0.1, 1.0 or 10 mg/kg in male Sprague-Dawley rats resulted in a majority of glucuronide acid conjugates of HQ at all dose levels, which were detected starting at 45 seconds post-dosing. Very small amounts of HQ-sulfate conjugates were detected, also after 45 seconds, and were higher in the high dose group. Minor metabolites detected sporadically included mono-glutathionyl-HQ, 2,3 -diglutathionyl-HQ, 2,3,5,6 -tetraglutathionyl-HQ, or 2,3,5 -triglutathionyl-HQ. The absence of metabolites in the initial arterial blood supported the limited pulmonary metabolic activity, but systemic metabolism still occured quite rapidly, with detection of HQ-glucuronide consisting of 93% of the blood14C within 12 minutes post-dosing at the doses 0.1 or 1.0 mg/kg. However signs of dose-dependent metabolic saturation were observed for the dose 10 mg/kg (Deisinger and English, 1999). The results from the intratracheal administration support a similar theoritical absorption capacity by the oral and pulmonary routes. Although there is no evidence of local pulmonary metabolic activity, the rapid appearance of HQ-glucuronide conjugates supports a rapid systemic metabolisation, with little or no production of glutathione conjugates. Results of the intratracheal studies provide supporting information for potential inhalation exposure, although due to the large particle size distribution of standard HQ production only a minor fraction of airborne particles would be expected to reach the alveolar region.

In vitro studies

In in vitro assays, species differences in the activities of glucuronidation and of acetylation were observed. The activities of these detoxifying metabolic steps were found to be lowest in male F344 rats compared to other rodent species or strains, or to humans.

Liver microsomes of male F344 rats exhibited a lower initial rate of HQ glucuronidation than those of human donors (N=8) and male B6C3F1 mice with values of 0.077, 0.101-0.223 and 0.218 nmol/mg protein/min, respectively (Supporting study: Seaton et al., 1995).

The mean specific acetylase activities in the microsomal fraction of freshly isolated rodent kidneys were ordered as follows: male B6C3F1 mice > male SD rat > female F344 rat > male F344 rat. No acetylase activity was detected in the cytosolic fraction. Acetylase activity was found to be significantly higher in kidney than in liver. Repeated dosing of HQ at 50 mg/kg/day for 10 consecutive days to rats did not induce or repress the activities of N-acetylase or N-deacetylase in kidney, and the strain and sex specific differences observed after single application were confirmed. Consequently, the distribution of CysHQ/N-AcCysHQ is controlled largely by the acetylase activity in the kidneys of rodents. The significantly lower rate of acetylation in the kidneys of male F344 rat compared to other rodent species, suggests that the male F344 rat has an inferior capacity for detoxification of CysHQ, which favours formation of products of the alternative competitive oxidative pathway with formation of protein-S-adducts (see data of Boatman et al., 2000). This, in turn, may contribute to the chronic nephrotoxic effects of HQ observed in the male F344 rat and explain the absence of these effects in female F344 and male SD rats (Key study: Barber et al., 2006).

Both the relative activities of GGT and N-acetylase/N-deacetylase influence the level of the nephrotoxic metabolite 2,3,5-(tricysteinyl-S-yl)HQ. This critical metabolite is easier to oxidize to the corresponding quinone than 2,3,5-(triGSyl)HQ or the quinol-mercapturic acids. In vitro, rats exhibited the highest rates of GGT-mediated hydrolysis and transpeptidation of 2,3,5-(triGSyl)HQ, and the highest sensitivity to to 2,3,5-(triGSyl)HQ-mediated nephrotoxicity compared to mice, hamsters and guinea pigs. The HQ cysteine conjugate, which is the product of GGT-mediated hydrolysis, exhibits a balance between N-acetylation and N-deacetylation, which is shifted to the formation of the N-acetylconjugate in rats. In this species, the ratio of hydrolysis/transpeptidation was about 0.34, the ratio of N-deacetylation/N-acetylation was about 0.25. In the guinea pig, which is the only other species susceptible to the nephrotoxic effects of 2,3,5-(triGSyl)HQ, the balance of N-acetylation and N-deacetylation favoured the formation of the cysteine conjugate while GGT activity is lower than in rats and mice. (Supporting study: Lau et al., 1995).


In vivo studies

Considerable background exposure of humans to HQ exists from dietary sources (e.g. 214 to 620 µg total HQ ingested from a single meal) as well as from drinking coffee and tea, as the occurrence of natural HQ and its glucopyranoside conjugate, arbutin, is whidespread in common foods. Background exposure from other uncharacterised sources was estimated to account to 2770 µg/d based on urinary excretion of total HQ, while occupational or food sources contribute only 7-18% of the total exposure. A very small additional exposure to HQ may result from cigarette smoking.

The large increases in the plasma concentrations and urinary excretion rates of total HQ, following diets with high HQ content, indicate that arbutin is extensively absorbed from the gastro-intestinal tract and becomes bioavailable as HQ, however, with a delay compared to absorption of a dose of free HQ. Hydrolysis of arbutin to free HQ probably occurs in the acidic milieu of the stomach. HQ absorbed from dietary sources appeared to be rapidly conjugated and was not detectable as free HQ in plasma or urine (Key studies: Deisinger, 1996; Deisinger et al., 1994).

In an experimental study, a single human male volunteer (41 yrs old) was orally dosed with 275 mg HQ (ca. 4 mg/kg bw) by self-ingestion of the dosing solution through a drinking straw. The time course of plasma levels and urinary excretion of HQ was monitored for 24 hrs. A 20-min delay in absorption from the gastro-intestinal tract was found as the plasma level of total HQ increased from a background level of ca. 0.01 mg/L at 15 min p.a. to a maximum level of ca. 0.10 mg/L at 30 min p.a.. HQ was rapidly eliminated from plasma as the peak level decreased to background level within 2 hr p.a.. Urinary excretion of metabolites was nearly complete after ca. 6-8 hrs for HQ sulfate and HQ glutathione conjugates, and after ca. 12 hrs for HQ glucuronide being the main metabolite. Cumulative excretion of HQ conjugates in urine within 24 hrs accounted to ca. 180 mg equivalents as HQ glucuronide, ca. 120 mg equivalents as HQ sulfate, and ca. 11 mg equivalents as HQ glutathione conjugates. As the total excretion of 312.5 mg HQ-equivalents was higher than the applied dose of 275 mg HQ, these data indicate a background excretion of about 40 mg HQ-equivalents within 24 hrs originating from dietary or endogenous sources (Key study: Corley et al., 1998, 2000). The data from a single volunteer were confirmed by further in vitro investigations on toxicokinetics of HQ in human liver and kidney cells and by an extended PBPK model (Key study: Poet et al., 2004).

In a lethally intoxicated human (suicidal case with an unknown oral dose of photographic developer solution) treated with extensive gastric lavage and forced diuresis, at a timepoint at least 30 hrs after intake, the highest concentrations of unchanged hydroquinone were found in urine (3.39 mg/L), liver (457 ng/g) and kidney (212 ng/g) followed by spleen (124.5 ng/g), and brain (62.5 ng/g) (no quantitative data on metabolites or total HQ) (Supporting study: Saito et al., 1994). Both, saturation of HQ metabolism at the high ingested dose and medical treatment with forced diuresis are of influence on observed tissue and urinary levels of free HQ in this case.

In vitro studies

In an in vitro study with microsomal fractions of livers from eight human donors (6 males, 2 females, ages 17-49 yrs), initial rates of glucuronidation of HQ, which is the main metabolic pathway, were in the range of 0.101-0.281 nmol/mg protein/min. The interindividual variability between human donors (6 males, 2 females, ages 17-49 yrs) was found to be about 3fold (Supporting study: Seaton et al., 1995).

The high variability in the activity of N-acetylase in human kidney tissue in vitro correlated with the time of collection of the specimen (mean value at 1 h post mortem 5.79 ± 2.60 nmol/min/mg protein) pointing to an apparent instability of this enzyme (Barber et al., 2006). Based on the work of Poet et al. (2004), liver acetylation may be quantitatively more important than kidney acetylation in overall elimination of HQ-S-conjugates in humans.


A PBPK model (Key study: Corley et al., 2000), which was structured to simulate HQ kinetics and the total formation of key metabolites in F344 rats, SD rats and human, was developed based on toxicokinetic data from different primary sources. The validity of this PBPK model was checked with experimental data from animal studies of English et al. (1988, 1991, 2005), Hill et al. (1993), and DiVincenzo et al. (1984), Boatman et al., (2000) and with human experimental data published by Corley et al. (2000). The model was found to successfully describe a variety of rat and human data from these studies arriving at the following conclusions.

With regard to the formation of the different metabolites in rats, sulfate conjugates account for approximately one half to two thirds of the glucuronide conjugates. For a human volunteer, glucuronide and sulfate conjugation of HQ was reasonably well simulated by the PBPK model. However, the model significantly underpredicted the amounts of GSH conjugates in humans which presumably originated from other sources (e.g. diet or endogenous metabolism).

The model was used to identify possible mechanisms of strain-specific toxicity between male F344 rats and Sprague-Dawley rats by comparing the internal dose surrogate for renal toxicity (total flux through the oxidation pathway leading to formation of BQ and HQ-GSH). There is a dose-related shift from less than 3 - 6% between HQ doses of 25 mg/kg bw to ca. 350 mg/kg bw in the formation of HQ-GSH. The predicted formation of GSH metabolites is about 1.7 fold higher in F344 rats compared to SD rats up to dose levels of 500 mg/kg bw after single oral gavage. This strain difference in internal dosimetry is consistent with the acute nephrotoxicity differences observed between these rat strains after HQ-exposure (data of Boatman et al., 1996). The first-pass effect of glucuronide and sulfate conjugation on the internal dosimetry of GSH metabolites is evident when comparing the intraperitoneal versus the oral route, especially at lower doses (e.g. 5-fold and 1.3-fold increases of GSH metabolites at 25 and 150 mg/kg bw, respectively, for i.p. vs. gavage application). The predicted increase in glutathione conjugation was consistent with increased protein adduct levels observed in rats (data of Boatman et al., 2000).

Since the publication in 2000 this PBPK model has been supplemented with toxicokinetic data on HQ-GHS dependent metabolic pathways (Poet et al., 2004; Barber and English, 2006) using an iterative, nested, in vitro metabolism study design coupled with pharmacokinetic modelling.

Apparent metabolic rate constants for the conversion of hydroquinone through several subsequent metabolic steps to the mono-glutathione conjugate and subsequent detoxification via mercapturic acid formation were investigated in hepatocytes from male F344 rats and from human donors (six males, ages 48-72 yrs) in vitro, and were combined with the PBPK model developed by Corley et al. (2000) to make comparative statements on the toxicokinetics of HQ in the liver of humans versus rats (Key study: Poet et al., 2004).

Glucuronide conjugation, which is the initial favoured step in the metabolism of HQ, has a significantly higher, about 4-fold capacity (Vmax) and intrinsic clearance (Vmax/Km) in human compared to rat liver cells. Additionally, in human hepatocytes, an overall higher capacity for detoxification was indicated by the higher relative rates of metabolism of the mono-glutathione conjugate GS-HQ to CysS-HQ byγ-glutamyltranspeptidase (GGT) and dipeptidase. With regard to the further metabolism of the CysS-HQ conjugate, N-acetylation producing the mercapturic acid conjugate (N-AcCysS-HQ) predominates in human hepatocytes and is favoured over deacetylation as reverse reaction. To saturate this pathway, higher concentrations of HQ-Cys would be needed in humans than in rats. Thus the intrahepatic mercapturic acid pathway may effectively detoxify HQ-S-conjugates and limit exposure of the human kidney to both glutathione and cysteine conjugates of HQ. Otherwise, the rat may be at greater risk for producing the nephrotoxic di-glutathione conjugates of HQ since the intrinsic clearance (Vmax/Km) for production of diGS-HQ was significantly (6-fold) higher in rat than in human hepatocytes. Additionally, the production of the putative proximal toxicant, HQ-triSG, can be expected to be much higher in F344 liver than in human liver. An estimation of the total amounts of di- and higher GS-HQ conjugates produced after oral absorption predicted considerably higher amounts in rats compared to humans, especially at the low dose range of 0.1 to 10 mg/kg bw (about 21 to 14-fold difference compared to about 6-fold difference at 1000 mg/kg bw). In contrast, predicted concentrations of CysS-HQ are considerably higher in humans and the difference between humans and rats increases with the dose from 100- to 200-fold in the dose range 0.1 to 1000 mg/kg bw.

As a central conclusion of the publication of Poet et al. (2004), the balance between bioactivation, potentially through pathways producing multi-conjugated glutathione metabolites, and deactivation, potentially through the GGT pathway with formation of mercapturic acid, is likely to be crucial to the species- and gender-specificity of toxic effects observed in test animals. Furthermore, the observation that deactivation steps predominate in human liver cells, and bioactivation steps predominate in rat liver cells are decisive for human risk assessment. Body burdens of higher substituted glutathione HQ conjugates, which are considered to be the key nephrotoxic metabolites, will be much higher in rats than in humans. In the case of HQ there is considerable evidence from the species-specific metabolism and toxicokinetics that toxic effects observed in test animals at a certain dose level, especially in the F344 rat, are of questionable biological relevance for a comparable human exposure situation (Key study: Poet et al., 2004).

Discussion on absorption rate:

Experimental in vivo and in vitro studies on dermal absorption through rat skin

In a dermal absorption study according to OECD Guideline 427, groups of male and female F344 rats were exposed to 14C-labelled HQ (5.4% aqueous solution) at doses of 25 or 150 mg/kg bw applied to a skin area of ca. 2 cm2, corresponding to dermal exposure concentrations of 12.5 and 75 mg/cm2 skin surface. Most of the applied 14C-HQ (61-71%) was recovered from the skin surface at 24 h. Cumulative recoveries of 14C by 168 h after application amounted from 7.8 - 12.8% of the material in urine, 3.8 - 8.9% in the chamber rinsings, and 1.7 - 3.7% in faeces. From 0.14 - 2.2% was found in the excised skin at the exposure site, and 2.6 - 12.9% was associated with the tissues and the carcass. Total recoveries were in the range of 90.1 - 94.6%. Blood levels were found to be below or near the limit of quantification (ca. 0.01% of applied dose or 0.06 µg equivalents HQ/g blood) except for 25 mg/kg bw female rats at the 0.5 hr timepoint (0.47 µg equivalents HQ/g blood) and 150 mg/kg bw female rats at the 1 hr timepoint (1.13 µg equivalents HQ/g blood). In urine, 4 - 8% of the dose accounted to HQ glucuronide, < 2% to HQ sulfate, and 0.4 - 1.5% to unchanged HQ. Mercapturic acid conjugate and benzoquinone were occasionally detected.

The low percent of applied dose recovered from urine and faeces (9 - 16%), and the low amount observed in the blood indicate that HQ is not readily absorbed percutaneously even at doses producing dermal irritation. Systemic absorption after dermal exposure was estimated to account to 11% based on a comparison of urinary excretion after oral and dermal absorption. This value may represent a slight overestimation of the actual dermal absorption as due to technical difficulties test substance leaked out from the dermal application device resulting in possible oral uptake or direct contamination of urine (Key study: English et al., 1988; English and Deisinger, 2005).

In vitro, dermal absorption of HQ was found to be slow based on flux rates of 14C-labelled HQ through full-thickness skin of male F344 rats measured with Franz-type diffusion cells with a 5% aqueous HQ donor solution. Skin integrity was not affected by contact with HQ. The permeability constant was 2.26 cm/h x 10- 5, and the flux rate was 1.09 µg/cm2/h (Key study: Barber et al., 1993, 1995).

Experimental in vivo and in vitro studies on dermal absorption through human skin

Dermal absorption of 14C-labelled HQ, as 2% HQ in a cream base, was studied after open application to the forehead or forearm skin of human male volunteers (14 volunteers in total with an age from 18 - 80 yrs). Exposures to 0.1 mg/cm2 skin surface, amounting to total doses of 0.5 or 2.5 mg/person, lasted to up to 24 h. For 24 h treatments, loss of test substance due to e.g. rubbing during sleep is possible. Radioactivity was determined in skin surface washes (cotton balls or swabs) after 0, 1, 3, 6 or 24 h of exposure and in excreted urine at 8 h and at 24 h intervals up to 96 h. 14C-activity located in the stratum corneum was measured by 10 repeated tape strippings at the indicated timepoints. HQ quickly disappeared from the exposed skin surface, filling and passing through the stratum corneum, into the body circulation, and being excreted in the urine. The amount absorbed within 6 h accounted to ca. 45% of the applied dose with 16% located in the stratum corneum. The absorption process was found to be continuous and the peak plasma concentration appeared at 4 h after start of exposure. Excretion in urine was predominantly as glucuronide conjugates of HQ besides a small amount of HQ and benzoquinone. Based on urinary excretion a total flux of 1.0 and 1.9 µg/cm2/h was calculated for human forearm or forehead skin, respectively (Key study: Wester et al., 1988).

These data are supported by a further in vivo skin absorption study. During a 24 h dermal exposure of six human volunteers to a 2 mg dose of 14C-HQ in a 95% ethanolic vehicle, HQ was readily absorbed from human forehead skin (exposure area 16 cm2). Peak elimination was observed within the first 8 to 24 h (up to 40% of the dose), and elimination was completed within 5 days. Based on the amount recovered in urine, at least 57% of the applied dose of 2 mg was absorbed systemically (Supporting study: Buck et al., 1988). A flux rate of 2 µg/cm2/h can be estimated for human forehead skin based on a dermal dose of 0.125 mg/cm2 with 40% eliminated in urine within 24 h.

In an in vitro skin absorption study, human skin samples (1 cm2) were exposed to 14C-HQ, applied as a 2% HQ cream at a dose of 0.2 mg/cm2, in a flow-through diffusion cell device. After a total treatment time of 24 h, 34% of HQ had permeated through the skin, and 9.3% was located in the skin. In the skin, the amount of HQ was 3 times higher than the amount of benzoquinone. Possible metabolism in skin had no significant effect on permeation profiles, on the flux of ca. 2.9 µg/cm2/h, and on the lag time of skin absorption of 8 h. This indicates that it took 8 h for HQ to penetrate through human skin into the receptor fluid. The flux rate of 2.9 µg/cm2/h observed in vitro was about two-fold higher than the flux rates of 1.0 and 1.9 µg/cm2/h observed in vivo for forearm and forehead skin. This difference was discussed to be a reflection of the doubled dose in vitro of 0.2 mg/cm2 compared to 0.1 mg/cm2 in the in vivo assay (Key study: Wester et al., 1988).

In a further in vitro study, a permeability constant of 0.933 cm/h x 10- 5and a flux rate of 0.524 µg/cm2/h through human stratum corneum was measured with Franz-type diffusion cells using a 5% aqueous HQ solution as donor fluid. Integrity of the used skin membranes was measured via permeation of tritiated water on the days prior to and following the permeation assay with HQ. Skin integrity was not affected by contact with HQ (Key study: Barber et al., 1993, 1995).

Comparison of dermal absorptions in humans and rats

Principally, dermal absorption of HQ was found to be slow. At a low dermal exposure concentration of 0.1 mg/cm2 skin surface, the flux rate of HQ through human skin is in the range of 1-2 µg/cm2/h both in vivo and in vitro (vehicle: 95% ethanol or cream base) (Key studies: Wester et al., 1988; Buck et al., 1988). In parallel in vitro assays with a 5% aqueous solution of HQ and identical test conditions, the flux rates through human stratum corneum and full thickness rat skin were similar (0.524 µg/cm2/h vs. 1.09 µg/cm2/h; key study: Barber et al., 1993, 1995) indicating that there is no significant difference in dermal absorption between human and rat skin.

In in vivo experimental studies with human volunteers, low total doses of 2 mg/person, corresponding to ca. 0.1 mg/cm2, seem to be readily absorbed, with up to 60% of the dose appearing in urine (Key studies: Wester et al., 1988; Buck et al., 1988). In contrast, in in vivo studies with male F344 rats, systemic absorption accounted to only about 10% at distinctly higher total doses of 25 and 150 mg/kg, corresponding to dermal exposure concentrations of 12.5 and 75 mg/cm2 (Key studies: English et al., 1988; English and Deisinger, 2005). Consequently, at toxicologically relevant doses oral absorption in rats was found to be about 9 to 10-fold higher than dermal absorption, which is also reflected in thresholds for toxic effects after acute and repeated dosing.

As the low flux rate is the limiting parameter, dermal absorption of HQ through human skin is expected to be decreased to values similar to rats at higher, toxicologically relevant doses of HQ (range about > 10 mg/cm2). Consequently, a dermal absorption rate of 10% is considered to be justified for human risk assessment.