Toxicology in Vitro 28 (2014) 742–750
Contents lists available at ScienceDirect
Toxicology in Vitro journal homepage: www.elsevier.com/locate/toxinvit
KeraSkin™-VM: A novel reconstructed human epidermis model for skin irritation tests Kyoung-Mi Jung a, Su-Hyon Lee b, Won-Hee Jang a, Haeng-Sun Jung b, Yong Heo c, Young-Ho Park a, SeungJin Bae d, Kyung-Min Lim d,⇑, Seung Hyeok Seok e,⇑ a
Amorepacific Co. R&D Center, Yongin 446-729, Republic of Korea Modern Cell & Tissue Technologies Inc., Seoul 139-743, Republic of Korea Department of Occupational Health, College of Natural Sciences, Catholic University of Daegu, Daegu 712-702, Republic of Korea d College of Pharmacology, Ewha Womans University, Seoul 120-808, Republic of Korea e Department of Microbiology and Immunology, and Institute of Endemic Disease, Seoul National University Medical College, Seoul 110-799, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 31 October 2013 Accepted 27 February 2014 Available online 10 March 2014 Keywords: Skin irritation Reconstructed human skin Keraskin™-VM Sensitivity Specificity Accuracy
a b s t r a c t Several alternative in vitro methods to evaluate skin irritants have been developed recently. In July 2010, OECD officially endorsed the validated reference method (VRM) that uses reconstituted human epidermis (RhE) models as replacements for the in vivo skin irritation test. This study evaluated the KeraSkin™-VM model, a novel human epidermis model that was reconstructed with Asian skin tissue using 20 reference chemicals according to the OECD TG 439 performance standard. The test chemicals were applied to the epidermal surface side for 45 min and then rinsed, and then incubated for 42 h post-treatment. An overall accuracy of 80%, sensitivity of 90% and specificity of 70% were obtained when the results from KeraSkin™-VM were compared with UN GHS categories, which was comparable to the EpiDerm™ Skin irritation test (SIT) rates. Furthermore, KeraSkin™-VM demonstrated good performance in terms of withinlaboratory reproducibility and predictive capacity to screen skin irritants. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Since the 1980s, many in vitro tests have been developed to replace existing in vivo animal test methods (Purchase, 1997). Skin irritation tests (SIT) that use rabbits or guinea pigs cause a considerable level of discomfort or pain to animals, which has caused SIT to be a primary target for an alternative to animal test (AAT) development. Furthermore, substantial levels of species differences are being reported in the chemical-induced skin toxicity between human and animals (York et al., 1996), which raises questions about the utility of conventional SIT using animals. Many in vitro SIT methods have been developed, including Corrositex, a monolayer skin cell culture, freshly excised skin, and epidermal and full skin equivalent models (Benassi et al., 1999; Perkins et al., 1999; Poumay and Coquette, 2007; Stobbe et al.,
⇑ Corresponding authors. Address: College of Pharmacy, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, Seoul 120-808, Republic of Korea. Tel.: +82 2 3277 3003; fax: +82 2 3277 3760 (K.M. Lim). Address: Department of Microbiology and Immunology, and Institute of Endemic Disease, Seoul National University College of Medicine, Seoul 110-799, Republic of Korea. Tel.: +82 2 740 8302; fax: +82 743 0881 (S.H. Seok). E-mail addresses: [email protected] (K.-M. Lim), [email protected] (S.H. Seok). http://dx.doi.org/10.1016/j.tiv.2014.02.014 0887-2333/Ó 2014 Elsevier Ltd. All rights reserved.
2003). Among them, reconstructed human epidermis (RhE) models have advanced to an OECD-endorsed in vitro skin irritation method. RhE models that mimic the three-dimensional structure of the human epidermis were first reported by Rosdy and Clauss in 1990 (Rosdy and Clauss, 1990; Roguet et al., 1994), and a prevalidation study was conducted from 1999 to 2000 by the European Centre for the Validation of Alternative Methods (ECVAM) (Fentem et al., 2001). In vitro SIT methods that use three RhE models, EpiSkin™, EpiDerm™ SIT (EPI-200) and the SkinEthic™ RHE, were accepted by OECD TG 439 through the ECVAM validation study (Alépée et al., 2010; Kandarova et al., 2009; EC-ECVAM, 2009a, 2009b). To encourage the introduction of new reconstructed human epidermis models, OECD TG 439 describes the qualification of standard RhE models in the performance standard (TG 439 Annex I) such as general and functional integrities that include cell viability, barrier function, morphology, reproducibility and quality control (Effective Time 50). A new reconstructed human epidermis model, KeraSkin™-VM (MCTT Inc., Seoul, Korea), was developed and prepared from Korean human foreskin. The KeraSkin™-VM model closely resembles the human epidermis and exhibits similar morphology, growth characteristics, and even epidermal protein expression patterns. Moreover, the multilayered and stratified stratum corneum of
743
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
the KeraSkin™-VM has a barrier function that is comparable to native human skin. In this study, we introduce a new RhE model, KeraSkin™-VM model and investigated whether this model could be applied to the in vitro SIT according to OECD TG 439 performance standards and provide a foundation for a future validation study. 2. Materials and methods 2.1. Materials Twenty recommended reference chemicals that are enlisted in the OECD TG 439 performance standard were tested. Specific details about the chemical, the potential irritant information and classification, according to the GHS systems and EU, are shown in Table 1. MTT(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), D-PBS, Sodium dodecyl sulfate (SDS) and Triton X100 were purchased from Sigma–Aldrich (USA). EpiDerm™ SIT model was purchased from MatTek Corp., Ashland, MA, USA. 2.2. Human epidermal equivalent model – KeraSkin™-VM The KeraSkin™ model is a reconstructed human epidermal equivalent composed of multilayered and fully-differentiated human keratinocytes. Briefly, primary cultured human keratinocytes were seeded on a 12 mm Millicell™ (Millipore, USA) and incubated for 7 days to confluence. Next, the epithelial differentiation was induced by air–liquid interface culture for 14 days with 3T3 feeder layers. The model was delivered in 24 unit packs with all necessary culture media, a sterile nylon membrane and a quality control certificate (IC50, TEER, and histology). 2.3. Quality control criteria of KeraSkin™-VM 2.3.1. Negative control cell viability The negative control cell viability was evaluated by MTT assay. KeraSkin™-VM was exposed to D-PBS (a negative control) and then the tissue were washed with D-PBS, and further incubated with 300 lL of MTT (0.3 mg/mL) for 3 h at 37 °C in a 5% CO2 incubator. At the end of incubation, excess MTT was removed, the tissue were washed with DPBS and 2 mL of isopropanol was added to extract the reduced formazan from the tissue, which was incubated for an additional 2 h. After formazan extraction, 250 lL extraction solutions were transferred to a 96-well plate, and the absorbance
of the extracted purple dye was measured at 570 nm using an ELISA plate reader (VERSAMAXÒ, Molecular Devices, CA, USA). 2.3.2. Effective Time-50 (ET-50) with 1% Triton X-100 To evaluate barrier function, we assessed the exposure time that is required to reduce cell viability by 50% (ET50) with the application of 1% Triton X-100. KeraSkin™-VM was exposed to 1% Triton X-100 for 0, 1, 3, 5, 7, and 14 h and then an MTT assay was performed after the exposures. The extracted formazan dye absorbance was measured at 570 nm using an ELISA plate reader (Flexstation III™, Molecular Devices, USA) in MCTT Inc. The measured OD value (ODsample) was corrected by subtracting a blank OD level, and the OD values at 0 h were set as the ODnegativecontrol. The OD values were calculated into a percent viability using the following formula: % viability = [ODsample/ODnegativecontrol] 100. A calibration curve of the percent viability relative to 0 min was constructed, and the ET-50 was calculated by establishing a regression curve to a logarithmic equation. 2.3.3. Morphology and immunohistochemistry The KeraSkin™-VM model was fixed with 10% formaldehyde, embedded with paraffin, and prepared as 0.4 lm sections using a RM2255 Microtome (Leica, Germany). Immunohistochemistry was performed, using the avidin–biotin complex technique with Universal VECTASTATIN ABC kit (Vector Laboratories, USA). Paraffin sections were de-paraffinized in xylene, hydrated through a decreasing ethanol concentration grades, and endogenous peroxidase activity was quenched using 0.3% hydrogen peroxide in PBS. Non-specific binding was eliminated by incubating with diluted normal blocking serum. Sections were then serially incubated with primary antiserum diluted in buffer for 1 h, diluted biotinylated secondary antibody solution for 40 min, ABC reagent for 30 min, and peroxidase substrate solution until the desired stain intensity developed. All steps were performed at room temperature. For immunohistochemistry, we used monoclonal antibodies for Collagen type IV, CK10 from DAKO (Glostrup, Denmark), CK14 and Loricrin from abCam (Cambridge, UK), Filaggrin from Novocastra Laboratories Ltd. (Newcastle, UK), P63 and Laminin from CHEMICON International Inc. (Temecula, CA), Transglutaminase, E-Cadherin from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA) and Involucrin from NeoMarkers (Fremont, CA). Each section was counterstained with hematoxylin, mounted and the entire tissue area was examined under an Olympus DX41 microscope (Center
Table 1 Test substances. Test material
CAS #
Solid/liquid
UN GHS Cat.
VRM
Supplier
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
6940-78-9 84-66-2 86-87-3 7493-74-5 67-63-0 3446-89-7 112-61-8 5870-93-9 6259-76-3 104-55-2 112-30-1 103-95-7 111-25-1 86604-75-3 1310-58-3 629-19-6 7340-90-1 5271-27-2 111-71-7 127-18-4
L L S L L L S L L L L L L S L L L S L L
No No No No No No No No No No R38 R38 R38 R38 R38 R38 R38 R38 R38 R38
I NI NI NI NI I NI NI NI I I I I I I NI I I I I
Aldrich Sigma–aldrich Sigma Acros Sigma Aldrich Sigma SAFC Fluka SAFC Sigma–aldrich Wako Aldrich Sigma–aldrich Fluka Aldrich Acros Aldrich Fluka Sigma–aldrich
1-Bromo-4-chlorobutane Diethyl phthalate Naphthalene acetic acid Allyl phenoxy-acetate Isopropanol 4-Methyl-thio-benzaldehyde Methyl stearate Heptyl butyrate Hexyl salicylate Cinnamaldehyde 1-Decanol Cyclamen aldehyde 1-Bromohexane 2-Chloromethyl-3,5-dimethyl-4-methoxypyridine HCl 5% Potassium hydroxide Di-n-propyl disulfide Benzenethiol,5-(1,1-dimethylethyl)-2-methyl 1-Methyl-3-phenyl-1-piperazine Heptanal Tetrachloroethylene
744
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Valley, PA, USA). Hematoxylin and eosin staining was also performed for histological analysis.
3. Results 3.1. Morphology of KeraSkin™-VM model
2.4. Protocol refinement Protocol for the KeraSkin™-VM skin irritation test was based on the VRM procedure adopted by OECD. Tissues were unpacked on delivery and transferred into 6 well plates filled with 0.9 mL KeraSkin™-VM culture media per well, and pre-incubated for 24 ± 2 h at 37 °C in 5% CO2. After pre-incubation, KeraSkin™-VM tissues were topically exposed to the chemicals. At least two KeraSkin™VM tissues were used for each material. For all of the tests, DPBS and sodium dodecyl sulfate (SDS) were used as the negative and positive control, respectively. Test conditions were refined, that include treatment volume, time and washing method using SDS. First, for liquid substances, the treatment volume (30 and 50 lL), nylon mesh application (Nylon net filter, Millipore), the treatment times (5, 10, 15, 30, 45 and 60 min) and three types of rinsing methods were examined. For solid chemicals, 30 or 50 lg were applied and the same volume of D-PBS was added. Treated chemical was gently spread on the tissue surface with a bulbheaded glass pipette.
KeraSkin™-VM is a new, three-dimensional human epidermal equivalent model that displays native human epidermis-like histology such as the fully-differentiated and multilayered basal layer, stratum spinosum, granular layer and stratum corneum (Fig. 1A). Full development of the stratified stratum corneum, which is critical to the skin irritation test, was also confirmed by scanning electron microscopy (Fig. 1B). Immunohistochemical staining, in addition to gross morphology, revealed similar physiological characteristics such as the expression of cytokeratin14 in the basal layer, and cytokeratin 10 in the supra-basal layer, and filaggrin in granular layer. Basal cells showed clear p63 expression which suggests that the epidermal keratinocyte stemness is preserved. In addition, E-cadherin were expressed on the cell membrane, and the cornified envelope-associated proteins like loricrin, involucrin, were detected (Fig. 1C). Histology and immunohistochemical staining of EpiDerm™ also revealed that KeraSkin™-VM has similar morphology and marker expression to EpiDerm™. 3.2. Tissue integrity and barrier function of KeraSkin™-VM model
2.5. Final protocol After 22 ± 2 h of pre-incubation, the tissue was removed from the incubator and the test materials were applied immediately. 30 lL of liquid materials were dispensed directly on the tissue surface and sterile forceps were used to tilt the insert and gently spread the liquid. A nylon mesh was placed over the tissue on the surface of the insert with care not to damage tissue layer. For solid materials, 30 lL of PBS was applied to the surface of the insert and then 30 lg of test material was placed directly onto the insert. After the materials were applied, the plates were incubated at 37 °C and 5% CO2. After 45 min, the tissues were rinsed with DPBS. Tissues were post-incubated for 42 h and then the entire medium was removed. Tissues were blotted and transferred to a 24well plate that contained MTT (0.3 mg/mL) and incubated for 3 h at 37 °C and 5% CO2. Next, tissues were rinsed with DPBS again and transferred to a new 6-well plate, prefilled with 2 mL of isopropanol. Formazan extraction was performed at room temperature for 2 h and 200 lL of formazan extract, per tissue was transferred to a 96 well plate. Optical density(OD) was measured at 570 nm using isopropanol as a blank. 2.6. Predictive ability The predictive capacity of the KeraSkin™-VM in vitro skin irritation model was evaluated by comparing the test results with known in vivo classification (UN GHS category) and VRM (Validated Reference Methods) in the OECD TG439. Test material was classified as an ‘‘irritant’’ when the cell viability was reduced to less than 50% of the negative control. Specificity represents the percentage of non-irritant reference chemicals that were correctly identified as non-irritants with this test method. Sensitivity was the percentage of irritant reference chemicals that were correctly identified irritants. Accuracy was the overall percentage of the correct classification of irritants and non-irritants. 2.7. Statistics Values are shown as mean ± SD for all of the data. The data were subjected to 2 sample t-test to determine which means were significantly different (p < 0.05). Statistical analysis was performed with Minitab software (Chicago, IL, USA).
The integrity and the barrier function of the KeraSkin™-VM models were examined by evaluating the tissue viability and resistance to exogenous stimuli. Tissue integrity (keratinocyte viability) was measured by MTT assay and effective time 50 (ET50) defined as the time necessary for 1% Triton-X 100 to reduce cell viability to 50%. Thickness and and trans-epithelial electrical resistance (TEER) were also determined. The optical density of 20 different KeraSkin™-VM lots were measured to be within the range of 0.8–1.3 (1.0 ± 0.15, Coefficient of Variation (CoV) = 15.6%), which was comparable to other commercial skin equivalent models (Fig. 2A). ET-50 values for Triton X-100 were measured between 4 and 9 h (7.1 ± 1.2 h, CoV = 16.5%), which was also similar to other skin equivalent models (Fig. 2B). Histology was used to measure the thickness of KeraSkin™-VM that was in the range of 52.7– 102.2 lm (84.6 ± 14.4 lm, CoV = 17.0%), (Fig. 2C). TEER values were all above 0.5 kX.cm™ (data not shown). 3.3. Protocol refinement A protocol for the in vitro SIT using KeraSkin™-VM was developed and SDS (1%, 2%, 5%, 10%, 20% and 30%), was used as the positive control. The target threshold cell viability was set at 50% to determine the irritant. As shown in Fig. 3A, SDS decreased the cell viability in a concentration-dependent manner and the cell viability at 20% SDS was below 20% (Fig. 3A). We then optimized the treatment volume, time and rinsing methods with 20% SDS. Two conditions were tested at various time points: 20% SDS 30 lL with mesh and 50 lL without mesh (Fig. 3B). As a result, the 30 lL treatment with mesh produced a better time-dependent effect and the cell viability dropped below 20% after 15 min. Next, we fixed the treatment condition at 30 lL with mesh and evaluated the treatment times (15, 30 and 45 min) and 3 rinsing methods (Fig. 3C). Wash method B showed the smallest variation and the best time-dependent irritation. Finally, 8 test compounds (4 non-irritants and 4 irritants) along with SDS 20% as a positive control were treated for 30 and 45 min, using the optimized condition (30 lL with a mesh and wash method B) (Fig. 3D) which revealed that a 45 min treatment produces better sensitivity. Sensitivity was estimated to be 50% and 75% for the 30 and 45 min treatment times, respectively. Table 2 describes the protocol comparisons for VRM-endorsed models and KeraSkin™-VM (see Fig. 4).
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
745
Fig. 1. Morphology of KeraSkin™-VM model. (A) Comparison between the normal human skin epidermis and the KeraSkin™-VM model by Hematoxylin and Eosin(H&E) staining; (B) scanning electron microscope image of the surface of KeraSkin™-VM model; (C) immunohistochemical staining of Keraskin™-VM for Cytokeratin 10, Cytokeratin 14, p63, E-cadherin, Loricrin, Involucrin and Filaggrin and (D) immunohistochemical staining for Cytokeratin 10, Cytokeratin 14 and p63, and H&E staining of EpiDerm™.
3.4. Predictive capacity of the in vitro skin irritation test using KeraSkin™-VM The overall in vitro KeraSkin™-VM skin irritation test performance was assessed by evaluating all 20 reference chemicals that were listed in the performance standard of OECD TG439. We also evaluated the chemicals with the EpiDerm™ SIT model which is a VRM-endorsed 3-D human epidermis model, according to the OECD TG439 guidelines. As a result, the specificity of KeraSkin™-VM was estimated to be 70% (7/10) according to the UN GHS Category (Table 4). Non-irritants, 1-bromo-4-chlorobutane, isopropanol, 4-methyl-thio-benzaldehyde and cinnamaldehyde were determined to be false positive. Sensitivity of KeraSkin™-VM was calculated to be 90% (9/10) where an irritant, 1-bromohexane was a false negative. The overall accuracy of KeraSkin™-VM was 80% (16/20). The specificity of EpiDerm™ SIT was estimated to be 60% (6/10), where nonirritants, 1-bromo-4-chlorobutane, isopropanol, 4-methyl-thiobenzaldehyde and cinnamaldehyde were determined to be false positives. Sensitivity of EpiDerm™ SIT was calculated to be 90% (9/10) where an irritant, Di-n-propyl disulfide was a false negative. The overall accuracy was 75% (15/20) of EpiDerm™ SIT which was comparable to that of the KeraSkin™-VM. When the performance of KeraSkin™-VM was re-calculated on the basis of OECD TG439 VRM, specificity, sensitivity and accuracy were 92% (11/12), 88% (7/8) and 90% (18/20), respectively. 3.5. Within-laboratory reproducibility To assess the robustness of the developed method, we evaluated the within-laboratory reproducibility. We conducted 3 runs
(each run was performed in triplicate) for all 20 reference chemicals (Fig. 5 and Table 3). All 3 runs, with the exception of 1-decanol (Chemical No. 11), were reproducible, and the 31-decanol runs were 53.0 ± 13.9%, 34.8 ± 29.3% and 32.2 ± 4.0%. Overall, withinlaboratory reproducibility was determined to be 95% (19/20). 4. Discussion In 2010, the OECD officially adopted an in vitro skin irritation test that used RhE as TG 439 (OECD TG No.439, 2010) and there are currently 3validated test methods that adhere to this test guideline, EpiSkin™, EpiDerm™ SIT (EPI-200) and the SkinEthic™ RhE model. OECD TG439 also includes performance standards for the assessment of similar and modified RhE-based test methods to encourage the introduction of new ‘‘me-too’’ RhE models and methods. Recently JaCVAM validated the LabCyte EPI-MODEL 24 SIT according to the performance standard (Katoh et al., 2009) and this model has been officially approved by the OECD as a new VRM (2013). The KeraSkin™-VM model is a reconstructed human skin model, derived from human foreskin, and consists of non-transformed human epidermal keratinocytes which have been cultured to form a highly differentiated human epidermis. Histological examination indicated that KeraSkin™-VM had similar cell layers that were comparable to native human skin that showed evidence of fullydifferentiated layers of the basal layer, stratum spinosum, granular layer, and stratified multilayer of the stratum corneum (Fig. 1A). This in vivo-likeness proved useful for in vitro SIT. Moreover, KeraSkin™-VM showed expression of pancytokeratin and cytokeratin 10 in immunohistochemical staining. Cytokeratins, which are
746
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Fig. 2. Quality control of KeraSkin™-VM model. (A) Negative control O.D values; (B) effective time 50 and (C) thickness.
Fig. 3. Protocol refinement for the skin irritation test with KeraSkin™-VM model. (A) Positive control concentration selection; (B) treatment volume; (C) treatment time and wash method and (D) confirmation of treatment time using 8 substances; Values are mean ± SD (N = 3). ⁄ represents statistically significant difference (student t-test, p < 0.05).
747
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Table 2 Protocol parameters of 4 models. The RhE methods had very similar protocols and all used a post-incubation period of 42 h. Variations mainly concern, (A) pre-incubation time and volume, (B) application of test substances, (C) post-incubation and (D) MTT assay. Model Procedure
EpiSkin™
EpiDerm™ SIT
Skinethic RHE™
KeraSkin™-VM
(A) Pre-incubation Incubation time Medium volume
18–24 h 2 mL
18–24 h 0.9 mL
Contents lists available at ScienceDirect
Toxicology in Vitro journal homepage: www.elsevier.com/locate/toxinvit
KeraSkin™-VM: A novel reconstructed human epidermis model for skin irritation tests Kyoung-Mi Jung a, Su-Hyon Lee b, Won-Hee Jang a, Haeng-Sun Jung b, Yong Heo c, Young-Ho Park a, SeungJin Bae d, Kyung-Min Lim d,⇑, Seung Hyeok Seok e,⇑ a
Amorepacific Co. R&D Center, Yongin 446-729, Republic of Korea Modern Cell & Tissue Technologies Inc., Seoul 139-743, Republic of Korea Department of Occupational Health, College of Natural Sciences, Catholic University of Daegu, Daegu 712-702, Republic of Korea d College of Pharmacology, Ewha Womans University, Seoul 120-808, Republic of Korea e Department of Microbiology and Immunology, and Institute of Endemic Disease, Seoul National University Medical College, Seoul 110-799, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 31 October 2013 Accepted 27 February 2014 Available online 10 March 2014 Keywords: Skin irritation Reconstructed human skin Keraskin™-VM Sensitivity Specificity Accuracy
a b s t r a c t Several alternative in vitro methods to evaluate skin irritants have been developed recently. In July 2010, OECD officially endorsed the validated reference method (VRM) that uses reconstituted human epidermis (RhE) models as replacements for the in vivo skin irritation test. This study evaluated the KeraSkin™-VM model, a novel human epidermis model that was reconstructed with Asian skin tissue using 20 reference chemicals according to the OECD TG 439 performance standard. The test chemicals were applied to the epidermal surface side for 45 min and then rinsed, and then incubated for 42 h post-treatment. An overall accuracy of 80%, sensitivity of 90% and specificity of 70% were obtained when the results from KeraSkin™-VM were compared with UN GHS categories, which was comparable to the EpiDerm™ Skin irritation test (SIT) rates. Furthermore, KeraSkin™-VM demonstrated good performance in terms of withinlaboratory reproducibility and predictive capacity to screen skin irritants. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Since the 1980s, many in vitro tests have been developed to replace existing in vivo animal test methods (Purchase, 1997). Skin irritation tests (SIT) that use rabbits or guinea pigs cause a considerable level of discomfort or pain to animals, which has caused SIT to be a primary target for an alternative to animal test (AAT) development. Furthermore, substantial levels of species differences are being reported in the chemical-induced skin toxicity between human and animals (York et al., 1996), which raises questions about the utility of conventional SIT using animals. Many in vitro SIT methods have been developed, including Corrositex, a monolayer skin cell culture, freshly excised skin, and epidermal and full skin equivalent models (Benassi et al., 1999; Perkins et al., 1999; Poumay and Coquette, 2007; Stobbe et al.,
⇑ Corresponding authors. Address: College of Pharmacy, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, Seoul 120-808, Republic of Korea. Tel.: +82 2 3277 3003; fax: +82 2 3277 3760 (K.M. Lim). Address: Department of Microbiology and Immunology, and Institute of Endemic Disease, Seoul National University College of Medicine, Seoul 110-799, Republic of Korea. Tel.: +82 2 740 8302; fax: +82 743 0881 (S.H. Seok). E-mail addresses: [email protected] (K.-M. Lim), [email protected] (S.H. Seok). http://dx.doi.org/10.1016/j.tiv.2014.02.014 0887-2333/Ó 2014 Elsevier Ltd. All rights reserved.
2003). Among them, reconstructed human epidermis (RhE) models have advanced to an OECD-endorsed in vitro skin irritation method. RhE models that mimic the three-dimensional structure of the human epidermis were first reported by Rosdy and Clauss in 1990 (Rosdy and Clauss, 1990; Roguet et al., 1994), and a prevalidation study was conducted from 1999 to 2000 by the European Centre for the Validation of Alternative Methods (ECVAM) (Fentem et al., 2001). In vitro SIT methods that use three RhE models, EpiSkin™, EpiDerm™ SIT (EPI-200) and the SkinEthic™ RHE, were accepted by OECD TG 439 through the ECVAM validation study (Alépée et al., 2010; Kandarova et al., 2009; EC-ECVAM, 2009a, 2009b). To encourage the introduction of new reconstructed human epidermis models, OECD TG 439 describes the qualification of standard RhE models in the performance standard (TG 439 Annex I) such as general and functional integrities that include cell viability, barrier function, morphology, reproducibility and quality control (Effective Time 50). A new reconstructed human epidermis model, KeraSkin™-VM (MCTT Inc., Seoul, Korea), was developed and prepared from Korean human foreskin. The KeraSkin™-VM model closely resembles the human epidermis and exhibits similar morphology, growth characteristics, and even epidermal protein expression patterns. Moreover, the multilayered and stratified stratum corneum of
743
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
the KeraSkin™-VM has a barrier function that is comparable to native human skin. In this study, we introduce a new RhE model, KeraSkin™-VM model and investigated whether this model could be applied to the in vitro SIT according to OECD TG 439 performance standards and provide a foundation for a future validation study. 2. Materials and methods 2.1. Materials Twenty recommended reference chemicals that are enlisted in the OECD TG 439 performance standard were tested. Specific details about the chemical, the potential irritant information and classification, according to the GHS systems and EU, are shown in Table 1. MTT(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), D-PBS, Sodium dodecyl sulfate (SDS) and Triton X100 were purchased from Sigma–Aldrich (USA). EpiDerm™ SIT model was purchased from MatTek Corp., Ashland, MA, USA. 2.2. Human epidermal equivalent model – KeraSkin™-VM The KeraSkin™ model is a reconstructed human epidermal equivalent composed of multilayered and fully-differentiated human keratinocytes. Briefly, primary cultured human keratinocytes were seeded on a 12 mm Millicell™ (Millipore, USA) and incubated for 7 days to confluence. Next, the epithelial differentiation was induced by air–liquid interface culture for 14 days with 3T3 feeder layers. The model was delivered in 24 unit packs with all necessary culture media, a sterile nylon membrane and a quality control certificate (IC50, TEER, and histology). 2.3. Quality control criteria of KeraSkin™-VM 2.3.1. Negative control cell viability The negative control cell viability was evaluated by MTT assay. KeraSkin™-VM was exposed to D-PBS (a negative control) and then the tissue were washed with D-PBS, and further incubated with 300 lL of MTT (0.3 mg/mL) for 3 h at 37 °C in a 5% CO2 incubator. At the end of incubation, excess MTT was removed, the tissue were washed with DPBS and 2 mL of isopropanol was added to extract the reduced formazan from the tissue, which was incubated for an additional 2 h. After formazan extraction, 250 lL extraction solutions were transferred to a 96-well plate, and the absorbance
of the extracted purple dye was measured at 570 nm using an ELISA plate reader (VERSAMAXÒ, Molecular Devices, CA, USA). 2.3.2. Effective Time-50 (ET-50) with 1% Triton X-100 To evaluate barrier function, we assessed the exposure time that is required to reduce cell viability by 50% (ET50) with the application of 1% Triton X-100. KeraSkin™-VM was exposed to 1% Triton X-100 for 0, 1, 3, 5, 7, and 14 h and then an MTT assay was performed after the exposures. The extracted formazan dye absorbance was measured at 570 nm using an ELISA plate reader (Flexstation III™, Molecular Devices, USA) in MCTT Inc. The measured OD value (ODsample) was corrected by subtracting a blank OD level, and the OD values at 0 h were set as the ODnegativecontrol. The OD values were calculated into a percent viability using the following formula: % viability = [ODsample/ODnegativecontrol] 100. A calibration curve of the percent viability relative to 0 min was constructed, and the ET-50 was calculated by establishing a regression curve to a logarithmic equation. 2.3.3. Morphology and immunohistochemistry The KeraSkin™-VM model was fixed with 10% formaldehyde, embedded with paraffin, and prepared as 0.4 lm sections using a RM2255 Microtome (Leica, Germany). Immunohistochemistry was performed, using the avidin–biotin complex technique with Universal VECTASTATIN ABC kit (Vector Laboratories, USA). Paraffin sections were de-paraffinized in xylene, hydrated through a decreasing ethanol concentration grades, and endogenous peroxidase activity was quenched using 0.3% hydrogen peroxide in PBS. Non-specific binding was eliminated by incubating with diluted normal blocking serum. Sections were then serially incubated with primary antiserum diluted in buffer for 1 h, diluted biotinylated secondary antibody solution for 40 min, ABC reagent for 30 min, and peroxidase substrate solution until the desired stain intensity developed. All steps were performed at room temperature. For immunohistochemistry, we used monoclonal antibodies for Collagen type IV, CK10 from DAKO (Glostrup, Denmark), CK14 and Loricrin from abCam (Cambridge, UK), Filaggrin from Novocastra Laboratories Ltd. (Newcastle, UK), P63 and Laminin from CHEMICON International Inc. (Temecula, CA), Transglutaminase, E-Cadherin from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA) and Involucrin from NeoMarkers (Fremont, CA). Each section was counterstained with hematoxylin, mounted and the entire tissue area was examined under an Olympus DX41 microscope (Center
Table 1 Test substances. Test material
CAS #
Solid/liquid
UN GHS Cat.
VRM
Supplier
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
6940-78-9 84-66-2 86-87-3 7493-74-5 67-63-0 3446-89-7 112-61-8 5870-93-9 6259-76-3 104-55-2 112-30-1 103-95-7 111-25-1 86604-75-3 1310-58-3 629-19-6 7340-90-1 5271-27-2 111-71-7 127-18-4
L L S L L L S L L L L L L S L L L S L L
No No No No No No No No No No R38 R38 R38 R38 R38 R38 R38 R38 R38 R38
I NI NI NI NI I NI NI NI I I I I I I NI I I I I
Aldrich Sigma–aldrich Sigma Acros Sigma Aldrich Sigma SAFC Fluka SAFC Sigma–aldrich Wako Aldrich Sigma–aldrich Fluka Aldrich Acros Aldrich Fluka Sigma–aldrich
1-Bromo-4-chlorobutane Diethyl phthalate Naphthalene acetic acid Allyl phenoxy-acetate Isopropanol 4-Methyl-thio-benzaldehyde Methyl stearate Heptyl butyrate Hexyl salicylate Cinnamaldehyde 1-Decanol Cyclamen aldehyde 1-Bromohexane 2-Chloromethyl-3,5-dimethyl-4-methoxypyridine HCl 5% Potassium hydroxide Di-n-propyl disulfide Benzenethiol,5-(1,1-dimethylethyl)-2-methyl 1-Methyl-3-phenyl-1-piperazine Heptanal Tetrachloroethylene
744
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Valley, PA, USA). Hematoxylin and eosin staining was also performed for histological analysis.
3. Results 3.1. Morphology of KeraSkin™-VM model
2.4. Protocol refinement Protocol for the KeraSkin™-VM skin irritation test was based on the VRM procedure adopted by OECD. Tissues were unpacked on delivery and transferred into 6 well plates filled with 0.9 mL KeraSkin™-VM culture media per well, and pre-incubated for 24 ± 2 h at 37 °C in 5% CO2. After pre-incubation, KeraSkin™-VM tissues were topically exposed to the chemicals. At least two KeraSkin™VM tissues were used for each material. For all of the tests, DPBS and sodium dodecyl sulfate (SDS) were used as the negative and positive control, respectively. Test conditions were refined, that include treatment volume, time and washing method using SDS. First, for liquid substances, the treatment volume (30 and 50 lL), nylon mesh application (Nylon net filter, Millipore), the treatment times (5, 10, 15, 30, 45 and 60 min) and three types of rinsing methods were examined. For solid chemicals, 30 or 50 lg were applied and the same volume of D-PBS was added. Treated chemical was gently spread on the tissue surface with a bulbheaded glass pipette.
KeraSkin™-VM is a new, three-dimensional human epidermal equivalent model that displays native human epidermis-like histology such as the fully-differentiated and multilayered basal layer, stratum spinosum, granular layer and stratum corneum (Fig. 1A). Full development of the stratified stratum corneum, which is critical to the skin irritation test, was also confirmed by scanning electron microscopy (Fig. 1B). Immunohistochemical staining, in addition to gross morphology, revealed similar physiological characteristics such as the expression of cytokeratin14 in the basal layer, and cytokeratin 10 in the supra-basal layer, and filaggrin in granular layer. Basal cells showed clear p63 expression which suggests that the epidermal keratinocyte stemness is preserved. In addition, E-cadherin were expressed on the cell membrane, and the cornified envelope-associated proteins like loricrin, involucrin, were detected (Fig. 1C). Histology and immunohistochemical staining of EpiDerm™ also revealed that KeraSkin™-VM has similar morphology and marker expression to EpiDerm™. 3.2. Tissue integrity and barrier function of KeraSkin™-VM model
2.5. Final protocol After 22 ± 2 h of pre-incubation, the tissue was removed from the incubator and the test materials were applied immediately. 30 lL of liquid materials were dispensed directly on the tissue surface and sterile forceps were used to tilt the insert and gently spread the liquid. A nylon mesh was placed over the tissue on the surface of the insert with care not to damage tissue layer. For solid materials, 30 lL of PBS was applied to the surface of the insert and then 30 lg of test material was placed directly onto the insert. After the materials were applied, the plates were incubated at 37 °C and 5% CO2. After 45 min, the tissues were rinsed with DPBS. Tissues were post-incubated for 42 h and then the entire medium was removed. Tissues were blotted and transferred to a 24well plate that contained MTT (0.3 mg/mL) and incubated for 3 h at 37 °C and 5% CO2. Next, tissues were rinsed with DPBS again and transferred to a new 6-well plate, prefilled with 2 mL of isopropanol. Formazan extraction was performed at room temperature for 2 h and 200 lL of formazan extract, per tissue was transferred to a 96 well plate. Optical density(OD) was measured at 570 nm using isopropanol as a blank. 2.6. Predictive ability The predictive capacity of the KeraSkin™-VM in vitro skin irritation model was evaluated by comparing the test results with known in vivo classification (UN GHS category) and VRM (Validated Reference Methods) in the OECD TG439. Test material was classified as an ‘‘irritant’’ when the cell viability was reduced to less than 50% of the negative control. Specificity represents the percentage of non-irritant reference chemicals that were correctly identified as non-irritants with this test method. Sensitivity was the percentage of irritant reference chemicals that were correctly identified irritants. Accuracy was the overall percentage of the correct classification of irritants and non-irritants. 2.7. Statistics Values are shown as mean ± SD for all of the data. The data were subjected to 2 sample t-test to determine which means were significantly different (p < 0.05). Statistical analysis was performed with Minitab software (Chicago, IL, USA).
The integrity and the barrier function of the KeraSkin™-VM models were examined by evaluating the tissue viability and resistance to exogenous stimuli. Tissue integrity (keratinocyte viability) was measured by MTT assay and effective time 50 (ET50) defined as the time necessary for 1% Triton-X 100 to reduce cell viability to 50%. Thickness and and trans-epithelial electrical resistance (TEER) were also determined. The optical density of 20 different KeraSkin™-VM lots were measured to be within the range of 0.8–1.3 (1.0 ± 0.15, Coefficient of Variation (CoV) = 15.6%), which was comparable to other commercial skin equivalent models (Fig. 2A). ET-50 values for Triton X-100 were measured between 4 and 9 h (7.1 ± 1.2 h, CoV = 16.5%), which was also similar to other skin equivalent models (Fig. 2B). Histology was used to measure the thickness of KeraSkin™-VM that was in the range of 52.7– 102.2 lm (84.6 ± 14.4 lm, CoV = 17.0%), (Fig. 2C). TEER values were all above 0.5 kX.cm™ (data not shown). 3.3. Protocol refinement A protocol for the in vitro SIT using KeraSkin™-VM was developed and SDS (1%, 2%, 5%, 10%, 20% and 30%), was used as the positive control. The target threshold cell viability was set at 50% to determine the irritant. As shown in Fig. 3A, SDS decreased the cell viability in a concentration-dependent manner and the cell viability at 20% SDS was below 20% (Fig. 3A). We then optimized the treatment volume, time and rinsing methods with 20% SDS. Two conditions were tested at various time points: 20% SDS 30 lL with mesh and 50 lL without mesh (Fig. 3B). As a result, the 30 lL treatment with mesh produced a better time-dependent effect and the cell viability dropped below 20% after 15 min. Next, we fixed the treatment condition at 30 lL with mesh and evaluated the treatment times (15, 30 and 45 min) and 3 rinsing methods (Fig. 3C). Wash method B showed the smallest variation and the best time-dependent irritation. Finally, 8 test compounds (4 non-irritants and 4 irritants) along with SDS 20% as a positive control were treated for 30 and 45 min, using the optimized condition (30 lL with a mesh and wash method B) (Fig. 3D) which revealed that a 45 min treatment produces better sensitivity. Sensitivity was estimated to be 50% and 75% for the 30 and 45 min treatment times, respectively. Table 2 describes the protocol comparisons for VRM-endorsed models and KeraSkin™-VM (see Fig. 4).
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
745
Fig. 1. Morphology of KeraSkin™-VM model. (A) Comparison between the normal human skin epidermis and the KeraSkin™-VM model by Hematoxylin and Eosin(H&E) staining; (B) scanning electron microscope image of the surface of KeraSkin™-VM model; (C) immunohistochemical staining of Keraskin™-VM for Cytokeratin 10, Cytokeratin 14, p63, E-cadherin, Loricrin, Involucrin and Filaggrin and (D) immunohistochemical staining for Cytokeratin 10, Cytokeratin 14 and p63, and H&E staining of EpiDerm™.
3.4. Predictive capacity of the in vitro skin irritation test using KeraSkin™-VM The overall in vitro KeraSkin™-VM skin irritation test performance was assessed by evaluating all 20 reference chemicals that were listed in the performance standard of OECD TG439. We also evaluated the chemicals with the EpiDerm™ SIT model which is a VRM-endorsed 3-D human epidermis model, according to the OECD TG439 guidelines. As a result, the specificity of KeraSkin™-VM was estimated to be 70% (7/10) according to the UN GHS Category (Table 4). Non-irritants, 1-bromo-4-chlorobutane, isopropanol, 4-methyl-thio-benzaldehyde and cinnamaldehyde were determined to be false positive. Sensitivity of KeraSkin™-VM was calculated to be 90% (9/10) where an irritant, 1-bromohexane was a false negative. The overall accuracy of KeraSkin™-VM was 80% (16/20). The specificity of EpiDerm™ SIT was estimated to be 60% (6/10), where nonirritants, 1-bromo-4-chlorobutane, isopropanol, 4-methyl-thiobenzaldehyde and cinnamaldehyde were determined to be false positives. Sensitivity of EpiDerm™ SIT was calculated to be 90% (9/10) where an irritant, Di-n-propyl disulfide was a false negative. The overall accuracy was 75% (15/20) of EpiDerm™ SIT which was comparable to that of the KeraSkin™-VM. When the performance of KeraSkin™-VM was re-calculated on the basis of OECD TG439 VRM, specificity, sensitivity and accuracy were 92% (11/12), 88% (7/8) and 90% (18/20), respectively. 3.5. Within-laboratory reproducibility To assess the robustness of the developed method, we evaluated the within-laboratory reproducibility. We conducted 3 runs
(each run was performed in triplicate) for all 20 reference chemicals (Fig. 5 and Table 3). All 3 runs, with the exception of 1-decanol (Chemical No. 11), were reproducible, and the 31-decanol runs were 53.0 ± 13.9%, 34.8 ± 29.3% and 32.2 ± 4.0%. Overall, withinlaboratory reproducibility was determined to be 95% (19/20). 4. Discussion In 2010, the OECD officially adopted an in vitro skin irritation test that used RhE as TG 439 (OECD TG No.439, 2010) and there are currently 3validated test methods that adhere to this test guideline, EpiSkin™, EpiDerm™ SIT (EPI-200) and the SkinEthic™ RhE model. OECD TG439 also includes performance standards for the assessment of similar and modified RhE-based test methods to encourage the introduction of new ‘‘me-too’’ RhE models and methods. Recently JaCVAM validated the LabCyte EPI-MODEL 24 SIT according to the performance standard (Katoh et al., 2009) and this model has been officially approved by the OECD as a new VRM (2013). The KeraSkin™-VM model is a reconstructed human skin model, derived from human foreskin, and consists of non-transformed human epidermal keratinocytes which have been cultured to form a highly differentiated human epidermis. Histological examination indicated that KeraSkin™-VM had similar cell layers that were comparable to native human skin that showed evidence of fullydifferentiated layers of the basal layer, stratum spinosum, granular layer, and stratified multilayer of the stratum corneum (Fig. 1A). This in vivo-likeness proved useful for in vitro SIT. Moreover, KeraSkin™-VM showed expression of pancytokeratin and cytokeratin 10 in immunohistochemical staining. Cytokeratins, which are
746
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Fig. 2. Quality control of KeraSkin™-VM model. (A) Negative control O.D values; (B) effective time 50 and (C) thickness.
Fig. 3. Protocol refinement for the skin irritation test with KeraSkin™-VM model. (A) Positive control concentration selection; (B) treatment volume; (C) treatment time and wash method and (D) confirmation of treatment time using 8 substances; Values are mean ± SD (N = 3). ⁄ represents statistically significant difference (student t-test, p < 0.05).
747
K.-M. Jung et al. / Toxicology in Vitro 28 (2014) 742–750
Table 2 Protocol parameters of 4 models. The RhE methods had very similar protocols and all used a post-incubation period of 42 h. Variations mainly concern, (A) pre-incubation time and volume, (B) application of test substances, (C) post-incubation and (D) MTT assay. Model Procedure
EpiSkin™
EpiDerm™ SIT
Skinethic RHE™
KeraSkin™-VM
(A) Pre-incubation Incubation time Medium volume
18–24 h 2 mL
18–24 h 0.9 mL
