Toxicology in Vitro 28 (2014) 240–247
Contents lists available at ScienceDirect
Toxicology in Vitro journal homepage: www.elsevier.com/locate/toxinvit
A new alternative method for testing skin irritation using a human skin model: A pilot study q A. Miles a, A. Berthet b, N.B. Hopf b, M. Gilliet a, W. Raffoul c, D. Vernez b, P. Spring a,⇑ a
Dermatology Clinic, Canton Vaud University Hospital (CHUV), Lausanne, Switzerland Institute of Work and Health (IST), University of Lausanne and Geneva, Lausanne, Switzerland c Reconstructive and Plastic Surgery Clinic, Canton Vaud University Hospital (CHUV), Lausanne, Switzerland b
a r t i c l e
i n f o
Article history: Received 9 July 2013 Accepted 30 October 2013 Available online 7 November 2013 Keywords: NaCl Heptylbutyrate Hexylsalicylate Butylmethacrylate Isoproturon Bentazon DEHP Methylisothiazolinone (MI) Folpet 1-Bromohexane Methylchloroisothiazolinone (MCI/MI) Sodiumlaurylsulfate Irritation scale Draize test EpiskinÒ EpidermÒ
a b s t r a c t Background: Studies assessing skin irritation to chemicals have traditionally used laboratory animals; however, such methods are questionable regarding their relevance for humans. New in vitro methods have been validated, such as the reconstructed human epidermis (RHE) model (EpiskinÒ, EpidermÒ). The comparison (accuracy) with in vivo results such as the 4-h human patch test (HPT) is 76% at best (EpidermÒ). There is a need to develop an in vitro method that better simulates the anatomo-pathological changes encountered in vivo. Objectives: To develop an in vitro method to determine skin irritation using human viable skin through histopathology, and compare the results of 4 tested substances to the main in vitro methods and in vivo animal method (Draize test). Methodology: Human skin removed during surgery was dermatomed and mounted on an in vitro flowthrough diffusion cell system. Ten chemicals with known non-irritant (heptylbutyrate, hexylsalicylate, butylmethacrylate, isoproturon, bentazon, DEHP and methylisothiazolinone (MI)) and irritant properties (folpet, 1-bromohexane and methylchloroisothiazolinone (MCI/MI)), a negative control (sodiumchloride) and a positive control (sodiumlaurylsulphate) were applied. The skin was exposed at least for 4 h. Histopathology was performed to investigate irritation signs (spongiosis, necrosis, vacuolization). Results: We obtained 100% accuracy with the HPT model; 75% with the RHE models and 50% with the Draize test for 4 tested substances. The coefficients of variation (CV) between our three test batches were 95%), hexylsalicylate (CAS n°6259-76-3, purity >98%), butylmethacrylate (CAS n°97-88-1, purity >99%), isoproturon (CAS n°3412359-6, purity >99%, at 250 lg), bentazon (CAS n°25057-89-0, basagranÒ), bis-2-ethylhexylphtalate (DEHP, CAS n° 117-81-7), methylisothiazolinone (CAS n°2682-20-4, aqueous solution, purity >98%, 500 pm), N-trichloromethylthiophtalimide (folpet; 6.3 mg in 4.5 ml; CAS n° 133-07-3), 1-bromohexane (CAS n°111-25-1, purity >98.5%), methylchloroistothiazolinone (CAS n° 26172-55-4, purity >98%, 100/25 ppm, 250/75 ppm, 375/125 ppm, 750/250 ppm). Sodium chloride (NaCl) at 0.9% prepared by dissolving 18 g of NaCl in purified water was used as a negative control. The positive control was sodiumlaurylsulphate (SLS; CAS n°151-21-3, purity >95%). SLS was prepared at 10% in MilliQ water. All chemicals were purchased as reference standards in Sigma Aldrich (Sigma–Aldrich (Buchs, St. Gallen, and Switzerland). Except for SLS, the four chemicals were applied as neat on the skin. 2.2. Fresh human skin preparation Human abdominal full thickness skin was obtained as surgical waste from the Department of Plastic and Reconstructive Surgery at the Centre Hospitalier Universitaire Vaudois (CHUV, Lausanne, Switzerland). All human donors were Caucasian females between 35 and 48 years old and had given their full consent. They had no known previous dermatological conditions. The skin samples were de-identified for use in this study. Skin, still attached to its subcutaneous fatty layer, brought directly to our laboratory in an ice-box and rinsed with saline solution at room temperature (NaCl 0.9% at 25 ± 2 °C). The skin was dermatomed to a thickness of 0.8 mm (Acculan II, B.Braun/Aesculap, Sempach, Switzerland), allowing the conservation of the entire epidermis and superficial dermis. A series of 2.5 cm2 discs were cut out of the length of each skin sample using a scalpel and mounted in an in vitro flowthrough diffusion cell system (PermgearÒ obtained from SES Analytical System, Bechenheim, Germany). This system consisted of a series of six diffusion cells; each one was divided into a donor chamber (upper compartment) above the skin and a receptor chamber (lower compartment) below the skin, and kept together
242
A. Miles et al. / Toxicology in Vitro 28 (2014) 240–247
with a clamp. The skin sample had an area of approximately 1.77 cm2. The lower chambers were filled with saline water (0.9% NaCl (aq)) and were placed in an incubator to maintain a skin surface temperature of 32 °C. Before applying the chemicals, the system was stabilized for 15 min to allow skin samples to hydrate. The integrity of the epidermis barrier was tested by measuring transepidermal water loss (VapoMeter wireless, Delfin Technologies Ltd., Kuopio, Finland) (Netzlaff et al., 2006). Only samples with a reading below 11 g/m2/h were used; others were replaced (Pinnagoda et al., 1990). The chemicals were then applied to skin in order to analyze their potential as skin irritants. One ml of each chemical was applied on skin samples via the donor chamber: this allowed covering the entire surface of the epidermis with the chemical. No more than 3 h elapsed from removal of the skin from the patient to mounting the skin sample on the in vitro system. The system ran for 4 h. At the end of the experiment, the chemicals in the donor chamber were removed using a plastic pipette and the skin samples were washed using saline solution at room temperature to rinse away the remaining chemicals. The skin samples were then fixed in formaldehyde and sent to the Institute for Pathology, CHUV for a standard hematoxylin and eosin (H&E) stain. The experimental procedure was repeated 3 times for each chemical, using the excised viable skin from 3 different donors). Some samples were performed using a sterile 5 mm punch (Stiefel).
The concordance of this new in vitro method was evaluated in relation to its sensitivity, specificity and accuracy with the classifications given by other methods. Its reproducibility was evaluated in relation to its standard deviation (SD) and coefficient of variation (CV). Finally, exclusive histopathological assessment was performed for isoproturon (250 lg), bentazon (basagranÒ), DEHP, folpet and both MCI/MI and MI at every concentration. 2.4. Statistical analysis and definitions Descriptive statistics were performed to investigate experimental data using a specific excel formula convertor software. Sensitivity is defined as the proportion of tested substances known (classified) as irritant (by the comparative method) and tested positive by our in vitro method. Specificity is defined as the proportion of tested substances known (classified) as non-irritant (by the comparative method) and tested negative by our in vitro method. Accuracy is defined as the comparative value (in%) of our tested cases with the other methods. 3. Results
The histological slides were analyzed using a Nikon 90i microscope and 10 section images were randomly taken across each specimen (blinded). Each image was studied for changes to the epidermis (spongiosis and vacuolization of the basal layer) and especially to cell morphology. Cells were categorized as either healthy cells or damaged cells, and counted. A percentage of damaged cells was calculated for each image. Cells were considered damaged if they showed signs of vacuolization of the cytoplasm or condensation of the nuclei. After microscopic analyses of the histological slides of H&E stained skin samples (Fig. A.1), chemicals were divided into two distinct groups, according to whether they had been in contact with an irritant or a non-irritant substance. Substances were classified as irritants (R38) when the number of damaged cells produced was greater than or significantly non-different from the number of damaged cells produced by the positive control substance (SLS 10%) in the same skin-sample batch.
Based on this visual analysis, a non-irritant chemical was defined as a chemical which induced no or slight damages to skin samples while an irritant generated severe damages such as significant spongiosis even up to epidermolysis. Fig. A.2–A.8 showed skin samples exposed to a non-irritant chemical, no or slight spongiosis, a vacuolization of the basal layer and an increase in the size of certain cells were observed. A thickening of the stratum corneum (orthokeratosis) was also visible. Skin samples exposed to an irritant chemical showed important signs of alteration to the epidermis, such as marked epidermolysis (Fig. B.1), significant vacuolization of the basal layer, marked spongiosis, bloated cells and shrunken nuclei (pyknosis) with a condensation of the chromatin (Fig. B.2). For the isothiazolinones: at the highest concentration tested, 750/250 ppm MCI/MI, we see signs of irritation characterized by focal spongiosis (Fig. B.3). There are, however, no signs of significant apoptosis or suffering cells on histological sections. This irritation is to be considered as mild. At lower concentrations, we have not been able to objectify signs of irritation. MI tested at 500 ppm remained negative (Fig. A.8). No false positives or false negatives were observed with either SLS 10% (Fig. B.4) or NaCl 0.9% (Fig. A.1). Furthermore, each tested
Fig. A.1. Histopathological response to NaCl 0.9% (negative control): no irritation signs.
Fig. A.2. Histopathological response to heptylbutyrate: no irritation signs.
2.3. Histology
A. Miles et al. / Toxicology in Vitro 28 (2014) 240–247
243
Fig. A.3. Histopathological response to hexylsalicylate: no irritation signs.
Fig. A.6. Histopathological response to bentazon (basagran, pure): no irritation signs.
Fig. A.4. Histopathological response to butylmethacrylate: no irritation signs.
Fig. A.7. Histopathological response to bis-2-ethylhexylphtalate (DEHP): no irritation signs.
Fig. A.5. Histopathological response to isoproturon 250 microgrammes: no irritation signs.
chemical showed similar results on each of the three viable skin-samples batches showing good reproducibility, by their low standard deviations for counted damaged cells and, their low variation coefficients(
Contents lists available at ScienceDirect
Toxicology in Vitro journal homepage: www.elsevier.com/locate/toxinvit
A new alternative method for testing skin irritation using a human skin model: A pilot study q A. Miles a, A. Berthet b, N.B. Hopf b, M. Gilliet a, W. Raffoul c, D. Vernez b, P. Spring a,⇑ a
Dermatology Clinic, Canton Vaud University Hospital (CHUV), Lausanne, Switzerland Institute of Work and Health (IST), University of Lausanne and Geneva, Lausanne, Switzerland c Reconstructive and Plastic Surgery Clinic, Canton Vaud University Hospital (CHUV), Lausanne, Switzerland b
a r t i c l e
i n f o
Article history: Received 9 July 2013 Accepted 30 October 2013 Available online 7 November 2013 Keywords: NaCl Heptylbutyrate Hexylsalicylate Butylmethacrylate Isoproturon Bentazon DEHP Methylisothiazolinone (MI) Folpet 1-Bromohexane Methylchloroisothiazolinone (MCI/MI) Sodiumlaurylsulfate Irritation scale Draize test EpiskinÒ EpidermÒ
a b s t r a c t Background: Studies assessing skin irritation to chemicals have traditionally used laboratory animals; however, such methods are questionable regarding their relevance for humans. New in vitro methods have been validated, such as the reconstructed human epidermis (RHE) model (EpiskinÒ, EpidermÒ). The comparison (accuracy) with in vivo results such as the 4-h human patch test (HPT) is 76% at best (EpidermÒ). There is a need to develop an in vitro method that better simulates the anatomo-pathological changes encountered in vivo. Objectives: To develop an in vitro method to determine skin irritation using human viable skin through histopathology, and compare the results of 4 tested substances to the main in vitro methods and in vivo animal method (Draize test). Methodology: Human skin removed during surgery was dermatomed and mounted on an in vitro flowthrough diffusion cell system. Ten chemicals with known non-irritant (heptylbutyrate, hexylsalicylate, butylmethacrylate, isoproturon, bentazon, DEHP and methylisothiazolinone (MI)) and irritant properties (folpet, 1-bromohexane and methylchloroisothiazolinone (MCI/MI)), a negative control (sodiumchloride) and a positive control (sodiumlaurylsulphate) were applied. The skin was exposed at least for 4 h. Histopathology was performed to investigate irritation signs (spongiosis, necrosis, vacuolization). Results: We obtained 100% accuracy with the HPT model; 75% with the RHE models and 50% with the Draize test for 4 tested substances. The coefficients of variation (CV) between our three test batches were 95%), hexylsalicylate (CAS n°6259-76-3, purity >98%), butylmethacrylate (CAS n°97-88-1, purity >99%), isoproturon (CAS n°3412359-6, purity >99%, at 250 lg), bentazon (CAS n°25057-89-0, basagranÒ), bis-2-ethylhexylphtalate (DEHP, CAS n° 117-81-7), methylisothiazolinone (CAS n°2682-20-4, aqueous solution, purity >98%, 500 pm), N-trichloromethylthiophtalimide (folpet; 6.3 mg in 4.5 ml; CAS n° 133-07-3), 1-bromohexane (CAS n°111-25-1, purity >98.5%), methylchloroistothiazolinone (CAS n° 26172-55-4, purity >98%, 100/25 ppm, 250/75 ppm, 375/125 ppm, 750/250 ppm). Sodium chloride (NaCl) at 0.9% prepared by dissolving 18 g of NaCl in purified water was used as a negative control. The positive control was sodiumlaurylsulphate (SLS; CAS n°151-21-3, purity >95%). SLS was prepared at 10% in MilliQ water. All chemicals were purchased as reference standards in Sigma Aldrich (Sigma–Aldrich (Buchs, St. Gallen, and Switzerland). Except for SLS, the four chemicals were applied as neat on the skin. 2.2. Fresh human skin preparation Human abdominal full thickness skin was obtained as surgical waste from the Department of Plastic and Reconstructive Surgery at the Centre Hospitalier Universitaire Vaudois (CHUV, Lausanne, Switzerland). All human donors were Caucasian females between 35 and 48 years old and had given their full consent. They had no known previous dermatological conditions. The skin samples were de-identified for use in this study. Skin, still attached to its subcutaneous fatty layer, brought directly to our laboratory in an ice-box and rinsed with saline solution at room temperature (NaCl 0.9% at 25 ± 2 °C). The skin was dermatomed to a thickness of 0.8 mm (Acculan II, B.Braun/Aesculap, Sempach, Switzerland), allowing the conservation of the entire epidermis and superficial dermis. A series of 2.5 cm2 discs were cut out of the length of each skin sample using a scalpel and mounted in an in vitro flowthrough diffusion cell system (PermgearÒ obtained from SES Analytical System, Bechenheim, Germany). This system consisted of a series of six diffusion cells; each one was divided into a donor chamber (upper compartment) above the skin and a receptor chamber (lower compartment) below the skin, and kept together
242
A. Miles et al. / Toxicology in Vitro 28 (2014) 240–247
with a clamp. The skin sample had an area of approximately 1.77 cm2. The lower chambers were filled with saline water (0.9% NaCl (aq)) and were placed in an incubator to maintain a skin surface temperature of 32 °C. Before applying the chemicals, the system was stabilized for 15 min to allow skin samples to hydrate. The integrity of the epidermis barrier was tested by measuring transepidermal water loss (VapoMeter wireless, Delfin Technologies Ltd., Kuopio, Finland) (Netzlaff et al., 2006). Only samples with a reading below 11 g/m2/h were used; others were replaced (Pinnagoda et al., 1990). The chemicals were then applied to skin in order to analyze their potential as skin irritants. One ml of each chemical was applied on skin samples via the donor chamber: this allowed covering the entire surface of the epidermis with the chemical. No more than 3 h elapsed from removal of the skin from the patient to mounting the skin sample on the in vitro system. The system ran for 4 h. At the end of the experiment, the chemicals in the donor chamber were removed using a plastic pipette and the skin samples were washed using saline solution at room temperature to rinse away the remaining chemicals. The skin samples were then fixed in formaldehyde and sent to the Institute for Pathology, CHUV for a standard hematoxylin and eosin (H&E) stain. The experimental procedure was repeated 3 times for each chemical, using the excised viable skin from 3 different donors). Some samples were performed using a sterile 5 mm punch (Stiefel).
The concordance of this new in vitro method was evaluated in relation to its sensitivity, specificity and accuracy with the classifications given by other methods. Its reproducibility was evaluated in relation to its standard deviation (SD) and coefficient of variation (CV). Finally, exclusive histopathological assessment was performed for isoproturon (250 lg), bentazon (basagranÒ), DEHP, folpet and both MCI/MI and MI at every concentration. 2.4. Statistical analysis and definitions Descriptive statistics were performed to investigate experimental data using a specific excel formula convertor software. Sensitivity is defined as the proportion of tested substances known (classified) as irritant (by the comparative method) and tested positive by our in vitro method. Specificity is defined as the proportion of tested substances known (classified) as non-irritant (by the comparative method) and tested negative by our in vitro method. Accuracy is defined as the comparative value (in%) of our tested cases with the other methods. 3. Results
The histological slides were analyzed using a Nikon 90i microscope and 10 section images were randomly taken across each specimen (blinded). Each image was studied for changes to the epidermis (spongiosis and vacuolization of the basal layer) and especially to cell morphology. Cells were categorized as either healthy cells or damaged cells, and counted. A percentage of damaged cells was calculated for each image. Cells were considered damaged if they showed signs of vacuolization of the cytoplasm or condensation of the nuclei. After microscopic analyses of the histological slides of H&E stained skin samples (Fig. A.1), chemicals were divided into two distinct groups, according to whether they had been in contact with an irritant or a non-irritant substance. Substances were classified as irritants (R38) when the number of damaged cells produced was greater than or significantly non-different from the number of damaged cells produced by the positive control substance (SLS 10%) in the same skin-sample batch.
Based on this visual analysis, a non-irritant chemical was defined as a chemical which induced no or slight damages to skin samples while an irritant generated severe damages such as significant spongiosis even up to epidermolysis. Fig. A.2–A.8 showed skin samples exposed to a non-irritant chemical, no or slight spongiosis, a vacuolization of the basal layer and an increase in the size of certain cells were observed. A thickening of the stratum corneum (orthokeratosis) was also visible. Skin samples exposed to an irritant chemical showed important signs of alteration to the epidermis, such as marked epidermolysis (Fig. B.1), significant vacuolization of the basal layer, marked spongiosis, bloated cells and shrunken nuclei (pyknosis) with a condensation of the chromatin (Fig. B.2). For the isothiazolinones: at the highest concentration tested, 750/250 ppm MCI/MI, we see signs of irritation characterized by focal spongiosis (Fig. B.3). There are, however, no signs of significant apoptosis or suffering cells on histological sections. This irritation is to be considered as mild. At lower concentrations, we have not been able to objectify signs of irritation. MI tested at 500 ppm remained negative (Fig. A.8). No false positives or false negatives were observed with either SLS 10% (Fig. B.4) or NaCl 0.9% (Fig. A.1). Furthermore, each tested
Fig. A.1. Histopathological response to NaCl 0.9% (negative control): no irritation signs.
Fig. A.2. Histopathological response to heptylbutyrate: no irritation signs.
2.3. Histology
A. Miles et al. / Toxicology in Vitro 28 (2014) 240–247
243
Fig. A.3. Histopathological response to hexylsalicylate: no irritation signs.
Fig. A.6. Histopathological response to bentazon (basagran, pure): no irritation signs.
Fig. A.4. Histopathological response to butylmethacrylate: no irritation signs.
Fig. A.7. Histopathological response to bis-2-ethylhexylphtalate (DEHP): no irritation signs.
Fig. A.5. Histopathological response to isoproturon 250 microgrammes: no irritation signs.
chemical showed similar results on each of the three viable skin-samples batches showing good reproducibility, by their low standard deviations for counted damaged cells and, their low variation coefficients(
