In Vitro Permeation Of Chromium Species Through Porcine

Anal Bioanal Chem (2006) 384: 378–384 DOI 10.1007/s00216-005-0226-z

PAPER IN FOREFRONT

Veerle Van Lierde . Cyrille C. Chéry . Nathalie Roche . Stan Monstrey . Luc Moens . Frank Vanhaecke

In vitro permeation of chromium species through porcine and human skin as determined by capillary electrophoresis–inductively coupled plasma–sector field mass spectrometry Received: 14 September 2005 / Revised: 24 October 2005 / Accepted: 31 October 2005 / Published online: 9 December 2005 # Springer-Verlag 2005

Abstract Since the species that trigger chromium allergy are not yet known, it is important to gain more of an insight into the mechanism of chromium transport through the skin and into the relationship between chromium allergy and chromium species. In vitro permeation studies with porcine and human skin were performed using a Franz static diffusion cell. Investigations attempted to elucidate (i) which Cr compounds are able to permeate through skin, (ii) the influence the Cr concentration in the donor solution has on the Cr permeation, and (iii) the effect that the time of exposure to the donor solution has on Cr permeation. Capillary electrophoresis hyphenated to inductively coupled plasma–sector field mass spectrometry (CE–ICP–SFMS) was used to separate and quantify the Cr species in the receptor fluid. 50 mmol L−1 phosphate buffer (pH 2.5) was used for CE separation, and two different electrophoretic runs were carried out (in the positive and negative modes). Pneumatic nebulization (PN)-ICP-SFMS was used in order to quantify the total amount of Cr absorbed by the skin after microwave-assisted acid digestion of the tissue. Cr(VI) was found to pass most easily through the skin. Nevertheless, Cr(VI) was also shown to be absorbed more efficiently by the skin than Cr(III), an observation attributed to a more pronounced rejection of the positively charged Cr(III) ions by the skin barrier. These results were in good agreement with in vitro permeation studies previously reported in the literature in which other analytical techniques were used. V. Van Lierde (*) . C. C. Chéry . L. Moens . F. Vanhaecke Laboratory of Analytical Chemistry, Ghent University, Proeftuinstraat 86, 9000 Ghent, Belgium e-mail: [email protected] N. Roche . S. Monstrey Department of Plastic Surgery, Ghent University Hospital, De Pintelaan 185, 9000 Ghent, Belgium C. C. Chéry NV Organon, Department of Pharmaceutics, P.O. Box 20, 5340 BH Oss, The Netherlands

Differences observed in the permeation of Cr following the application of aqueous Cr donor solutions and Cr-containing simulated sweat donor solutions are also described. Keywords Chromium . CE–ICP–MS . Porcine skin . Human skin . Simulated sweat

Introduction Chromium, and the chronic allergic contact dermatitis that it can lead to, pose a considerable health issue [1]. Occupational groups such as cement workers, chromium platers and metal workers, workers dealing with leather tanning and employees in the ceramics industry are at risk of developing chromium allergy [2–8]. Only tri- and hexavalent chromium compounds can be considered as potential haptens because all other chromium salts are unstable. Their ability to elicit allergic contact dermatitis depends mostly on the bioavailability of the chromium salts. In vitro techniques are used around the world to estimate the percutaneous absorption of potentially toxic chemicals. The penetration of chromium has been studied previously for animal as well as human skin [9–21]. The goals of this work were twofold: first to explore the potential of CE– ICP–SFMS as a low sample consumption speciation technique, and second to check the results from the in vitro chromium permeation studies using this alternative technique. To do so, in vitro permeation experiments were performed with porcine and human skin using a Franz static diffusion cell. Using this technique, we investigated (i) which Cr compounds/species are able to permeate through the skin, (ii) what influence the Cr concentration in the donor solution exerts on the Cr permeation, and (iii) what effect the exposure time has on the Cr permeation. CE–ICP–SFMS was used in order to separate and quantify chromium species in the receptor fluid, which has the advantage that only microliter amounts of sample are needed for different runs. PN–ICP–SFMS was used in order to quantify the amount of total chromium absorbed

379

by the skin after microwave-assisted acid digestion of the skin. Aqueous Cr-containing donor solutions and Cr-containing simulated sweat donor solutions were compared in terms of the ability of Cr to permeate through skin.

Experimental Reagents and materials Potassium dichromate (purity ≥99.995% ) and chromium chloride (CrCl3·6H2O; pro analysis) were purchased from Merck (Darmstadt, Germany). Chromium nitrate (Cr(NO3)3· 9H2O; pro analysis) was obtained from UCB (Leuven, Belgium). Ultrapure water (resistivity higher than 18 MΩ cm) was obtained by purifying bidistilled water with a Milli-Q system (Millipore, Milford, MA, USA). Before use, the CE separation buffer was filtered through a 0.2 μm PTFE filter (Alltech Associates Inc., Lokeren, Belgium). Bare fused silica capillaries (internal diameter: 75 μm) were purchased from Polymicro Technologies (Phoenix, AZ, USA). Potassium hydroxide and sodium chloride (purity ≥99.5%) were from Carlo Erba (Milan, Italy). The simulated sweat was an aqueous solution of sodium chloride (5 g L−1), lactic acid (1 g L−1), urea (1 g L−1) and five amino acids (1 g L−1 each) [22–25]. The amino acids selected were threonine, methionine, serine, alanine and glycine. The pH of the simulated sweat solution was adjusted to a value of 5.5 with ammonia. The sweat solutions were incubated with Cr(III) or Cr(VI) at 37 °C for 22 hours with constant shaking (120 rpm). Urea (crystalline, purity ≥99.5%), L-alanine (purity ≥98%), DLmethionine (purity ≥99%), L-threonine (purity ≥98%) and Lserine (purity ≥99%) were purchased from Sigma (St. Louis, MO, USA). Glycine (pro analysis) and lactic acid (88%) were purchased from Vel (Leuven, Belgium). Ammonia solution (25%) was purchased from Merck (Leuven, Belgium). 50 mmol L−1 sodium phosphate buffer solution (pH 2.5) was purchased from Fluka Chemie GmbH (Buchs, Switzerland). Porcine ear skin was obtained from adult animals as waste after they had died a natural death or were killed for consumption. Human abdominal skin was supplied by the

department of plastic surgery of the Ghent University Hospital after informing the patient. The Franz diffusion cell, purchased from PermeGear Inc. (Bethlehem, PA, USA) is manufactured from glass and provides a contact area of 0.64 cm2 and a receptor volume of 5 mL. Isotonic phosphate-buffered saline (PBS) solution was used as receptor solution in the Franz diffusion cell. PBS solution was prepared by dissolving 8.00 g sodium chloride, 0.20 g potassium chloride (purity ≥99.5%, Carlo Erba, Milan, Italy), 0.20 g potassium dihydrogen orthophosphate (KH2PO4; pro analysis, UCB, Leuven, Belgium) and 1.44 g sodium phosphate dibasic dihydrate (Na2HPO4·2H2O; pro analysis, Riedel-de Haën, SigmaAldrich GmbH, Munich, Germany) in 1 L Milli-Q water. Ultrapure HNO3 (65%) was obtained via subboiling distillation of reagent grade acid, and H2O2 (pro analysis; 30%) was purchased from Merck (Darmstadt, Germany). Cobalt (1,000 mg L−1 in 5% HNO3) stock solution was purchased from Merck (Leuven, Belgium). Instrumentation A CE system (Agilent 3D, Palo Alto, CA, USA) was coupled to an ICP–SFMS element-specific detector (ELEMENT, Finnigan MAT, Bremen, Germany). The ICP– SFMS system was operated in the medium mass resolution mode (R=3,000) in order to eliminate the spectral overlap of the 40Ar12C+ ,35Cl16OH+ and 52Cr+ signals and of the 40 Ar13C+, 37Cl16O+ and 53Cr+ signals. It is equipped with a guard electrode at the ICP torch in order to achieve the highest possible sensitivity [26]. A commercially available interface (CEI-100 CETAC, Omaha, NE, USA), based on a modified microconcentric nebulizer [27] (MCN-100), which runs in the free aspiration mode at a flow rate of approximately 12 μL min−1, combined with a special, low dead volume spray chamber [28], was used to couple the CE unit to the ICP torch. The experimental conditions of the CE–ICP–SFMS set-up are given in Table 1. The CE– ICP–SFMS method described previously by Van Lierde et al. [29] was used for the speciation and detection of the

Table 1 Experimental conditions of the CE–ICP–SFMS set-up Capillary electrophoresis

ICP–SFMS

Voltage Injection

±20 kV Hydrodynamic 500 mbar·s Injection volume ca. 55 nL Bare fused silica Total length ± 65 cm 75 μm inner diameter 360 μm outer diameter 20 min KOH 1 mol L−1 20 min Milli-Q H2O 20 min separation buffer 0.5 min KOH 0.1 mol L−1 1 min Milli-Q H2O 4 min separation buffer

Capillary

Preconditioning

Before use

Before each run

Resolution Cool gas flow rate Auxiliary gas flow rate Sample gas flow rate Dwell time RF power Guard electrode Make-up liquid flow rate Signals monitored Sample time

3,000 13 L min−1 1.03 L min−1 1.18 L min−1 150 ms 1,200 W Connected 12 μL min−1 52 + 53 + Cr , Cr 10 ms

380

chromium species in the receptor fluid. Using this method implies that (i) the analysis consisted of two different electrophoretic runs in order to detect all positively and all negatively charged Cr species, and (ii) neutral chromium species were not detected, since the electro-osmotic flow inside the capillary was eliminated. This analysis did not require any sample preparation (which could increase the risk of species conversion and/or analyte losses), buffer additives or coated capillaries to eliminate interactions between the analytes and the inner surface of the CE capillary. Microwave-assisted acid digestion of the skin membranes was performed with a Milestone (Shelton, CT) MLS-1,200 mega microwave furnace. To do so, the skin membranes were dried (at 30 °C for 24 hours) after which 6 mL of HNO3 (65%) and 1 mL of H2O2 (30%) were added for digestion (amounts of skin were between 0.5 g and 1 g). The microwave digestion program for porcine and human skin is shown in Table 2. The solutions thus obtained were further diluted with Milli-Q water before analysis with PN– ICP–SFMS. Co was added as internal standard (final concentration: 50 μg L−1) in order to compensate for signal drift, instrument instability and matrix effects. External calibration was used to quantify Cr. Preparation of skin membranes Porcine ears were cleaned under cold running water before whole skin membranes were removed from the underlying cartilage. Because of the thickness of the dermis, fullthickness skin was not used. Separating the epidermis by heat [11, 30–33] was not feasible because of the coarse hair. A thin barrier membrane was obtained by slicing a section from the surface of the skin using a dermatome (Silver Dermatoom, Aesculap H. Nootens, Brussels, Belgium) at a setting of 600 μm. Hairs were cut without damaging the skin and the skin membranes were stored at −20 °C until further use and for no longer than two months. Before use, the frozen skin membranes were allowed to thaw at room temperature. All subcutaneous fat from the human abdominal skin was carefully removed and the skin surface was washed with Milli-Q water. Human skin membranes were obtained in the same way as described for porcine skin membranes and also stored at −20 °C until further use. The limited number of skin sources should be emphasized here: only three porcine skin sources and one human Table 2 Microwave digestion program for porcine and human skin

Step

Duration (min)

1 2 3 4 5 Ventilation

2 2 6 5 5 5

Power (Watt)

skin source were used in the experiments. Only results obtained with the same skin source were compared. In vitro permeation of chromium: experimental set-up Permeation experiments were performed using a Franz diffusion cell [34–36]. The Franz diffusion cell consists of a donor compartment and a receptor compartment. Skin membranes were mounted between the cell compartments with the stratum corneum towards the donor compartment, and an O-ring was used in order to position the membrane. The two cell compartments were held together with a clamp. The receptor compartment had a volume of 5 mL and was filled with PBS solution as receptor fluid. It was kept at 37 °C by circulating heated water through an external water jacket. After 30 min of equilibration of the membrane with the receptor fluid, 355 μL of the Crcontaining donor solution was introduced into the donor compartment by means of a pipette. The donor compartment was covered with parafilm to prevent evaporation of the solvent and to simulate occlusion. The receptor fluid was continuously stirred by means of a magnet bar, spinning at 400 rpm (IKAMAG RCT, IKA -Labortechnik, Staufen, Germany). For some experiments, 100 μL aliquots of receptor fluid were withdrawn through the sampling port of the receptor compartment after various time intervals and stored at −20 °C until analysis. The volume of the samples withdrawn was replaced with fresh receptor fluid (PBS solution) to keep the volume constant during the experiment.

Results and discussion Experiments were performed in order to evaluate the effect of (i) the oxidation state of Cr, (ii) the Cr concentration in the donor solution and (iii) the exposure time on the in vitro permeation of Cr through porcine and human skin. Since sweat plays a significant role in the elicitation of chromium dermatitis [12], both Cr-containing aqueous solutions and Cr-incubated simulated sweat were applied onto the porcine and human skin. Detection limit of Cr with CE–ICP–SFMS Table 3 shows the limit of detection (LOD, 3s criterion for ten 30 s background sections) for Cr in (i) 0.14 mol L−1 HNO3, and (ii) PBS solution obtained with CE–ICP– Table 3 Limits of detection for Cr with CE–ICP–SFMS Cr-species Matrix

250 0 250 400 650 0

Cr(III) Cr(III) Cr(VI) Cr(VI)

0.14 mol L−1 HNO3 PBS 0.14 mol L−1 HNO3 PBS

LOD LOD (μg Cr cm−2 skin) (μg Cr L−1) 0.06 0.09 0.05 0.09

8 12 6 11

381

SFMS under the experimental conditions described in Table 1. Potassium dichromate was used as Cr(VI) standard and chromium nitrate as Cr(III) standard. The LOD of Cr is slightly higher for PBS solution compared to that for 0.14 mol L−1 HNO3, probably due to matrix effects caused by the presence of alkali metals (sodium and potassium). Determination of Cr species in the receptor fluid using CE–ICP–SFMS after in vitro exposure of porcine and human skin to Cr The application of CE–ICP–SFMS to the detection and quantification of the Cr species in the receptor fluid has an important advantage: only microliters of sample are necessary for several runs. This implies that only microliters (≥10 μL) had to be withdrawn from the receptor compartment during the in vitro permeation experiments. After filtration, samples were introduced into the CE–ICP– SFMS system without any further sample preparation. External calibration was used to quantify the Cr species in the receptor fluid; the Cr standards were prepared in PBS solution in order to overcome matrix effects (Table 3). For a typical electropherogram of sample separation, the authors refer the reader to [29]. Effect of oxidation state Table 4 shows the amount of Cr in the receptor fluid after exposing the skin tissue to three different chromium salts (CrCl3·6H2O, Cr(NO3)3·9H2O, K2Cr2O7) for 168 hours. After 168 hours, no Cr was detected in the receptor fluid after skin exposure to the Cr(III) compounds, CrCl3·6H2O and Cr(NO3)3·9H2O. After skin exposure to K2Cr2O7, a detectable amount of Cr had permeated through the skin. No species conversion of K2Cr2O7 was observed. These results show that Cr(VI) is able to diffuse through the skin barrier when occlusion is simulated, and hence, that the bioavailability of Cr(VI) differs significantly from that of Cr(III). From this, it is clear that Cr(VI) can pass more easily through the skin barrier than Cr(III), which is in Table 4 Amount of Cr in receptor fluid after exposure of (a) porcine and (b) human skin to different chromium salts for 168 hours Chromium salt (0.034 mol Cr L−1)

(a) Receptor fluid (b) Receptor fluid (μg permeated (μg permeated Cr cm−2 skin) Cr cm−2 skin)

Aqueous solutions CrCl3·6H2O
0.17

0.13


agreement with the results from the in vitro permeation experiments performed with human skin by Gammelgaard et al. [37]. We discuss this further later in the manuscript. Different species of Cr are present in simulated sweat [29]. From Table 4, it is clear that, of these Cr species, only Cr(VI) was observed to pass through the skin membrane; in other words, the LOD does not allow the possible permeation of the other Cr species present in simulated sweat to be observed. Effect of concentration From Table 4, it is clear that detectable amounts of Cr were only observed in the receptor fluid after exposure of the skin to K2Cr2O7. Therefore, only K2Cr2O7 solutions were applied onto skin samples in order to evaluate the influence of the Cr concentration in the donor solution on the in vitro permeation of Cr. The amount of Cr in the receptor fluid after exposure of (a) porcine and (b) human skin to different concentrations of K2Cr2O7 for 168 hours is shown in Table 5. After exposure of the skin to a donor solution containing 0.25% K2Cr2O7, no Cr was detected in the receptor fluid after 168 hours. A dichromate peak was observed from a concentration of 0.5% K2Cr2O7 applied onto the skin. However, increasing the Cr(VI) concentration in the donor solution does not result in a proportional increase in the content of Cr(VI) in the receptor fluid. Comparison of the quantitative results showed that less Cr passed through the skin when the Cr was present in simulated sweat than for aqueous Cr-containing solutions. This may be explained by the ability of sweat to partially reduce Cr(VI) into Cr(III) [29], which has been shown to pass through the skin with much more difficulty than Cr(VI), if at all (Table 4). Besides, Cr species formed during the incubation of simulated sweat with Cr [29] may not be able to diffuse through the skin barrier. Both factors result in a smaller fraction of unbound Cr(VI) that is able to pass the skin barrier and be detected in the receptor fluid. Table 5 Amount of Cr in receptor fluid after exposure of (a) porcine and (b) human skin to different concentrations of K2Cr2O7 for 168 hours Concentration K2Cr2O7 (%)

(a) Receptor fluid (b) Receptor fluid (μg permeated (μg permeated Cr cm−2 skin) Cr cm−2 skin)

Aqueous solutions 0.25 < LOD 0.5 0.18 1.0 0.19 2.5 0.21 5.0 0.24 Incubated in simulated sweat 0.25 < LOD 0.5 0.12 1.0 0.15 2.5 0.16 5.0 0.17

0.17 0.21 0.23 0.25

< LOD 0.18 0.20 0.23 0.25

0.13 0.15 0.17 0.20

< LOD 0.10 0.12 0.15 0.16

382

Effect of exposure time In Table 6, the levels of Cr in the receptor fluid after exposure of (a) porcine and (b) human skin to (i) 0.5% K2Cr2O7 in H2O and (ii) 1.0% K2Cr2O7 in simulated sweat are shown for different exposure times. For this experiment, a 100 μL aliquot was withdrawn daily from the receptor compartment (and replaced with fresh PBS solution) through the sampling port of the diffusion cell. From Table 6, increased Cr transport through the skin was observed with exposure time, resulting in detectable amounts of Cr in the receptor fluid from an exposure time of 72 hours onwards for aqueous Cr-containing donor solutions applied onto porcine skin and from 96 hours onwards for (i) aqueous Cr-containing donor solutions applied onto human skin and (ii) Cr-incubated simulated sweat donor solutions applied onto porcine and human skin. Again, a comparison of the quantitative results showed that, during the entire experiment, slightly more Cr was observed in the receptor fluid for aqueous Cr-containing donor solutions than for Cr-incubated simulated sweat. Determination of total Cr in porcine and human skin by means of microwave-assisted acid digestion PN–ICP–SFMS after in vitro exposure of skin to Cr After the experiments described in the previous paragraph had been performed, the skin membrane was rinsed with plenty of Milli-Q water and stored at −20 °C until further analysis. The total amount of Cr which was absorbed by the skin (or remained in the skin) was determined quantitatively using ICP–SFMS with the standard sample introduction Table 6 Amount of Cr in receptor fluid after exposure of (a) porcine and (b) human skin to (i) 0.5% K2Cr2O7 in H2O and (ii) 1.0% K2Cr2O7 in simulated sweat for different exposure times Exposure time (hours)

(a) Receptor (b) Receptor fluid (μg permeated fluid Cr cm−2 skin) (μg permeated Cr cm−2 skin)

Aqueous solutions (i) 24 < LOD 48 < LOD 72 0.10 0.11 96 0.12 0.13 120 0.14 0.15 144 0.15 0.16 168 0.18 0.17 Incubated in simulated sweat (ii) 24 < LOD 48 < LOD 72 < LOD 96 0.09 0.10 120 0.11 0.12 144 0.13 0.14

< LOD < LOD < LOD 0.14 0.15 0.17 0.18 < LOD < LOD < LOD 0.09 0.11 0.12

system (pneumatic nebulization, PN) after microwaveassisted acid digestion of the skin membranes. Quantification was carried out by means of external calibration using an aqueous Cr standard solution. Effect of oxidation state The amount of Cr in porcine and human skin after exposure of the skin to different Cr salts for 168 hours is given in Table 7. Detectable amounts of Cr were found in the skin after exposure of the skin to all of the applied Cr compounds. The amount of Cr found in the skin was, however, substantially higher after the application of K2Cr2O7 onto the skin, which implied a larger accumulation of Cr in the skin after exposure of the skin to Cr(VI). These results are in agreement with the results from previous in vitro permeation experiments [12, 38–40]. However, these results cannot be compared directly since different skin sources and different exposure times were used. Comparison of the quantitative results shows that less Cr is bound to the skin when the Cr salts are incubated in simulated sweat before application onto the skin. For Cr(III), this may be explained by the formation of Cr species [29] that were not observed to (i) pass through and (ii) bind to the skin; similarly for Cr(VI), since a fraction of Cr(VI) is reduced to Cr(III) after which the same Cr(III) species are formed as for the incubation of Cr(III) with simulated sweat [29]. Despite the stronger binding of Cr(III) to skin proteins [12], the amount of Cr found in the skin after exposure to Cr(III) compounds is lower than that seen after exposure to Cr(VI). Consequently, the smaller permeation of Cr(III) through the skin is probably due to greater rejection of the positively charged Cr(III) ions by the skin barrier [37]. Effect of concentration Table 8 shows the amount of Cr in porcine and human skin after exposure of the skin to different concentrations of CrCl3·6H2O and K2Cr2O7 for 168 hours. From this, it was clear that increasing the concentration of K2Cr2O7 in the donor solution (aqueous solutions) from 0.5% to 2.5% results in an increase (although not a Table 7 Amount of Cr in (a) porcine skin and (b) human skin after exposure of the skin to different Cr salts for 168 hours Chromium salt (0.034 mol·L−1Cr) Aqueous solutions CrCl3·6H2O Cr(NO3)3·9H2O K2Cr2O7 Incubated in simulated sweat CrCl3·6H2O Cr(NO3)3·9H2O K2Cr2O7

(a) (μg Cr cm−2 skin)

(b) (μg Cr cm−2 skin)

2.0 4.4 103

2.4 6.2 115

12.4 15.3 121

1.1 2.3 73

1.5 3.9 91

2.3 4.3 87

383 Table 8 Amount of Cr in (a) porcine and (b) human skin after exposure of the skin to different concentrations of CrCl3·6H2O and K2Cr2O7 for 168 hours Chromium salt

Aqueous solutions CrCl3·6H2O CrCl3·6H2O K2Cr2O7 K2Cr2O7 K2Cr2O7 Incubated in simulated sweat CrCl3·6H2O CrCl3·6H2O K2Cr2O7 K2Cr2O7 K2Cr2O7

Applied (a) (μg Cr concentration (%) cm−2 skin)

the aqueous CrCl3·6H2O donor solution, despite the formation of Cr species during incubation in simulated sweat [29] that are apparently unable to bind to the skin.

(b) (μg Cr cm−2 skin)

Effect of exposure time 0.9 (0.034 mol L−1 Cr) 4.5 (0.17 mol L−1 Cr) 0.5 (0.034 mol L−1 Cr) 2.5 (0.17 mol L−1 Cr) 5.0 (0.34 mol L−1 Cr)

0.9 (0.034 mol L−1 Cr) 4.5 (0.17 mol L−1 Cr) 0.5 (0.034 mol L−1 Cr) 2.5 (0.17 mol L−1 Cr) 5.0 (0.34 mol L−1 Cr)

2.0

2.4

2.1

2.4

103

115

121

129

137

159

126

136

162

1.1

1.5

2.3

1.5

2.1

4.2

73

91

87

120

141

95

126

144

121

proportional increase) in the total Cr concentration in the skin. Applying a concentration of K2Cr2O7 in the donor solution further, up to 5.0%, did not result in much more of an increase in the total Cr concentration in the skin. A possible explanation for this may be the limited binding capacity of the skin; saturation of the Cr-binding sites may occur. These results are in agreement with those from previous in vitro permeation experiments [15, 21, 41]. After exposure of the skin to CrCl3·6H2O, the level of Cr in the skin was of the same order of magnitude after application of either 0.9% or 4.5% CrCl3·6H2O. Increasing the Cr(III) concentration in the donor solution did not result in a considerable increase in the Cr concentration in the skin. These results are in agreement with the literature, where in vitro binding studies with soluble skin proteins were performed [42], and where an increased binding of Cr was observed with increasing chromate concentration in the donor solution, whereas no or little change was observed in the binding of Cr from donor solutions with different Cr(III) concentrations. This was interpreted as saturation of the binding sites of the skin proteins for Cr(III). After exposure of the skin to CrCl3·6H2O (incubated in simulated sweat), little difference was observed in the Cr levels in the skin after applications of either a 0.9% or a 4.5% CrCl3·6H2O solution. A comparison of the quantitative results (Table 8) indicates that the amount of Cr which absorbed in the skin after application of CrCl3·6H2O, incubated in simulated sweat, was not much lower than for

In Table 9, the amount of Cr in porcine skin after exposure of the skin to 0.5% K2Cr2O7 for 48 and 168 hours is shown. From Table 6, an increase in Cr transport through the skin with exposure time was observed. From Table 9, it is clear that there is also an accumulation of Cr in the skin. The long exposure time needed to observe a detectable amount of Cr in the receptor fluid (Table 6) and the high Cr concentrations that were found in the skin (Table 9) suggest that binding of Cr in the skin interferes with diffusion of Cr through the skin. The amount of Cr in the skin and in the receptor fluid increases with the duration of the exposure time. As the process of Cr permeation continues, the Cr concentration in the skin increases and increasing amounts of Cr are able to permeate further through the skin to the receptor fluid. Contribution of sweat to the elicitation of chromium allergy Since it is not known which Cr species are responsible for the elicitation of chromium allergy, the impact of the presence of sweat on eliciting chromium contact dermatitis cannot be fully understood. The Cr ions cannot elicit an immune response by themselves but must be chemically linked to proteins in order to elicit a T-cell response [43]. Cr(VI) is absorbed more easily through the skin and it can cross cell membranes more readily than Cr(III) salts [44]. Unlike Cr(VI), Cr(III) has a strong protein binding capacity [45]. It has been speculated that Cr(III) is the actual hapten, and that Cr(VI), after entering the skin cells, is reduced to Cr(III), which in its turn binds onto intracellular proteins, thereby forming immunogenic complexes [46]. Sweat is able to reduce a fraction of Cr(VI) into Cr(III) and, in turn, Cr(III) can partially form complexes with sweat and/or leather compounds [29]. These both result in less residual unbound Cr(VI). It is clear that the presence of sweat contributes to the elicitation of Cr allergy (more specifically, leather allergy) due to its ability to extract Cr from leather [47]. However, it remains uncertain whether sweat contributes in a negative or a positive way to the elicitation of Table 9 Amount of Cr in porcine skin after exposure of the skin to 0.5% K2Cr2O7 for 48 and 168 hours Exposure time (hours) Aqueous solutions 48 168 Incubated in simulated sweat 48 168

Porcine skin (μg Cr cm−2 skin)

6.5 103

9.7 115

3.1 73

4.2 91

384

chromium allergy, since it is not known which Cr oxidation state is the main one responsible for the sensitization.

Conclusions This study describes the use of CE–ICP–SFMS for Cr speciation and quantification during in vitro permeation experiments. An important advantage of CE–ICP–SFMS is that only microliter amounts of sample are necessary for different runs. The number of Cr species observed in the experiments is low, but a separation technique (such as CE) is necessary in order to check for the absence or presence of species conversion. PN–ICP–SFMS was used in order to determine the total amount of Cr found in porcine and human skin after exposure of the skin to Cr-containing solutions. The results obtained with CE–ICP–SFMS were in good agreement with those from previous works, which suggests the possibility of applying this technique in biomedical research. From the in vitro permeation experiments, it was clear that good agreement existed between the results obtained with porcine skin and human skin. This once again confirmed the suitability of porcine skin as a model for human skin. It was also observed that Cr(III) salts have a strong affinity for the skin, whereas the affinity of Cr(VI) for the skin is less pronounced since it passes more easily through it. The fact that Cr(VI) is more strongly absorbed by the skin than Cr(III) could not be explained by the higher affinity of Cr(III) for the skin, but can be explained by a greater rejection of the positively charged Cr(III) ions by the skin barrier. Water and simulated sweat were compared as media for the K2Cr2O7 donor solutions. Higher concentrations of chromate were needed in order to obtain the same amount of permeation through the skin when simulated sweat was used as solvent. This could be explained by the ability of sweat to reduce a fraction of Cr(VI) to Cr(III), since Cr(III)—and probably also the Cr(III) species formed by subsequent complexation— passes through the skin less easily than Cr(VI). Acknowledgements The authors would like to thank the Fund for Scientific Research—Flanders (Belgium) for financial support (research project G.0037.01). Veerle Van Lierde is a Research Assistant of the Fund for Scientific Research—Flanders. The authors would also like to thank Prof. Dr. Jean-Marie Naeyaert and Dr. Barbara Boone for their voluntary contributions to this project.

References 1. Halbert AR, Gebauer KA, Wall LM (1992) Contact Dermatitis 27:214–219 2. Nriagu JO, Nieboer E (eds)(1988) Chromium in the natural and human environment. Wiley, New York 3. Polak L (1983) Immunology of chromium. In: Burrows D (ed) Chromium:metabolism and toxicity. CRC, Boca Raton, FL

4. Avnstorp C (1992) Acta Derm-Venereol 179:1–22 5. Estlander T, Jolanki R, Kanerva L (2000) Contact Dermatitis 43:114 6. Peltonen L, Fraki J (1983) Contact Dermatitits 9:190–194 7. Kvitko E (2001) Contact Dermatitis 45:256 8. Kanerva L, Jolanki R, Estlander T, Alanko K, Savela A (2000) Am J Contact Dermat 11:155–160 9. Bronaugh RL (1995) Toxicol Method 5:265–273 10. Bronaugh RL (1995) Toxicol Method 5:275–281 11. Dick IP, Scott RC (1992) J Pharm Pharmacol 44:640–645 12. Samitz MH, Katz S, Shrager JD (1967) J Invest Dermatol 48:514–520 13. Bronaugh RL, Stewart RF, Congdon ER, Giles AL (1982) Toxicol App Pharmacol 62:474–480 14. Bronaugh RL, Stewart RF, Congdon ER (1982) Toxicol App Pharmacol 62:481–488 15. Wahlberg JE, Skog E (1965) Arch Dermat 92:315–318 16. Mali JWH, Van Kooten WJ, Van Neer FCJ (1963) J Invest Dermatol 41:111–122 17. Fujii M, Yamanouchi S, Hori N, Iwanaga N, Kawaguchi N, Matsumoto M (1997) Biol Pharm Bull 20:249–254 18. Bronaugh RL, Franz TJ (1986) Brit J Dermatol 115:1–11 19. Hood HL, Bronaugh RL (1999) J Cosmetic Sci 50:63–64 20. Baranowska-Dutkiewicz B (1981) Arch Toxicol 47:47–50 21. Liden S, Lundberg E (1979) J Invest Dermatol 72:42–45 22. Hansen MB, Johansen JD, Menné T (2003) Contact Dermatitis 49:206–212 23. Flint GN (1998) Contact Dermatitis 39:213–221 24. Nygren O, Wahlberg JE (1998) Analyst 123:935–937 25. Flint GN, Carter SV, Fairman B (1998) Contact Dermatitis 39:315–316 26. Appelblad PK, Rodushkin I, Baxter DC (2000) J Anal Atom Spectrom 15:359–364 27. Vanhaecke F, Van Holderbeke M, Moens L, Dams R (1996) J Anal Atom Spectrom 11:543–548 28. Schaumlöffel D, Prange A (1998) Patent No 198 41 288.6 29. Van Lierde V, Chéry CC, Moens L, Vanhaecke F (2005) Electrophoresis 26:1703–1711 30. Bhatti AS, Scott RC, Dyer A (1988) J Pharm Pharmacol 40 (Suppl):45P 31. Baumberger PJ, Suntzeff V, Cowdry EV (1942) J Natl Cancer Inst 2:413–423 32. Scheuplein RJ (1965) J Invest Dermatol 45:334–346 33. Bronaugh RL, Congdon ER, Scheuplein RJ (1981) J Invest Dermatol 76:94–96 34. Friend DR (1992) J Control Release 18:235–248 35. Montagna W, Yun JS (1964) J Invest Dermatol 44:11–21 36. Marcarian H, Calhoun ML (1966) Am J Vet Res 27:765–772 37. Gammelgaard B, Fullerton A, Avnstorp C, Menné T (1992) Contact Dermatitis 27:302–310 38. Mali JWH, Van Kooten WJ, Van Neer CJ (1963) J Invest Dermatol 41:111–122 39. Wahlberg JE, Skog E (1965) Arch Dermatol 92:315–318 40. Spruit D, Van Neer FCJ (1966) Dermatologica 132:179–182 41. Wahlberg JE (1968) Dermatologica 137:17–25 42. Samitz MH, Katz SA, Scheiner DM, Gross PR (1969) Acta Derm-Venereol 49:142–146 43. Sinigaglia F (1994) J Invest Dermatol 102:398–401 44. Bradberry SM, Vale JA (1999) J Toxicol Clin Toxicol 37:195– 200 45. Rytter M, Haustein UF (1982) Brit J Dermatol 106:161–168 46. Polak L, Turk JL, Frey JR (1973) Prog Allergy 17:145–226 47. Samitz MH, Gross S (1960) J Occup Med 2:12–14