Grading criteria of histopathological evaluation in BCOP assay by various staining methods

Abstract

This study was performed to provide information on classifying eye irritating chemicals using the BCOP assay. After the BCOP assay, bovine corneas were classified by IVIS presented in OECD test guideline 437, and three special staining methods (H&E, MT, and PAS) were performed for histopathological evaluation. Non-irritant chemicals (IVIS ≤ 3), showed intact structures. In the 3 < IVIS ≤ 55 group, epithelial cell edema was observed by H&E staining, and loose collagen bundles were confirmed by MT staining. In PAS staining, bleaching of the epithelium and reduced visibility of the basement membrane were observed. Severe irritant chemicals (IVIS > 55) showed large increases edema and nuclear condensation by H&E staining. Loose collagen bundles and vacuoles around keratocytes were also observed by MT staining. Bleaching of the epithelial layer, reduction in visibility, and thickness of the basement membrane were confirmed by PAS staining. Based on the stepwise histopathological analysis, we set the criteria and grades for histopathological evaluation and found that eye irritation was increased following the irritation degree of test chemicals. Further histopathological study will support and lead to improvements in the BCOP assay.

Introduction

The awareness of change on experimental animals has led to people to protect ethic and welfare of experimental animals [1]. In advanced countries, the new principle as 3Rs based on refinement, reproduction, and replacement was suggested to replace experimental animals [2]. Following this principle, various alternative animal tests have been studied and developed [3].

The Bovine Corneal Opacity and Permeability (BCOP) assay was listed in OECD test guideline 437 and replaces the in vivo Draize eye irritation test [4]. The BCOP assay uses corneas from slaughtered bovine and evaluates corneal opacity and permeability after treatment with chemicals [5, 6]. This assay was recommended to classify severe and non-irritant chemicals by the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM), the European Center for the Validation of Alternative Methods (ECVAM), and the Japanese Center for the Validation of Alternative Methods (JaCVAM) [7, 8].

Opacity is measured by the amount of light passing through the cornea [9], and permeability is measured using sodium fluorescein. If corneal epithelium is damaged, permeability can increase [10, 11]. The In Vitro Irritancy Score (IVIS) is calculated from the opacity and permeability. It is used to classify eye irritant chemicals. In OECD test guideline 437, chemicals with IVIS over 55 are classified as UN GHS category 1 [12], indicating severe eye damage. Chemicals below IVIS of 3 are classified as 'No category' which means they are non-irritants. However, in the BCOP assay, several chemicals (UN GHS category 1) are classified as UN GHS category 2 (2A or 2B) (eye irritation or mild irritation), and UN GHS No category chemicals were overclassified as UN GHS category 2, 2A or 2B. Therefore, the BCOP assay is not suitable for clearly classifying chemicals and additional in vitro or in vivo analyses are required to classify mild irritating chemicals [12].

In some previos studies [13–15], bovine corneas were stained with hematoxylin and eosin (H&E) staining method only, and corneal lesions were observed. OECD guidance No. 160 [16] also recommended H&E staining method, and used histopathological observation with this staining method to improve the accuracy of the BCOP assay. The bovine cornea has epithelium, stroma and endothelium. The stroma consists of collagen fibers, and Masson trichrome (MT) staining is used to observe connective tissues in the pathology [17]. H&E staining is not adequate for histopathological observation and it can not offer sufficient information to classify eye irritating chemicals. When observe a visibility of the basement membrane, Periodic acid schiff (PAS) staining had recommended in the ICE assay as another alternative eye irritation test [16].

This study was purposed, considering the corneal structure, to analyze and verify the histopathological evaluation in the BCOP assay using H&E, MT and PAS staining methods. We also tried to confirm the validity of the application for classifying eye irritating chemicals.

Materials and methods

Test chemicals and experimental equipments

The test chemicals (Table 1) used in this study were chosen according to OECD test guideline 437 and the UN GHS classification system. Following OECD test guideline 437, ethyl alcohol was used as a positive control (PC) in this study and distilled water (DW) as a negative control (NC). To measure permeability, fluorescein sodium (Sigma-Aldrich, St. Louis, Mo, USA) solution diluted with DPBS (Lonza, Cascada, MD, USA) was used. Penicillin/streptomycin (Gibco Invitrogen, 100 IU/mL penicillin and 100 μL/mL streptomycin) and hanks balanced salt solution (HBSS) were used (Sigma-Aldrich, St. Louis, MO, USA) to transport the bovine eyeballs. To measure the corneal opacity and permeability, we used an opacitometer (Testo 545, DURATEC, Germany) and spectrophotometer (Epoch, BioTek, USA). After the BCOP assay, we observed the histopathological changes in the corneas with an optical microscope (TS100, Nikon, Japan).

Table 1.

Test chemicals used in BCOP assay

Number	Test chemical	CASRNa	In vivo GHS classificationb	Physical form	Treatment concentration (%)
1	Imidazole	288–32-4	Category 1	Solid	100
2	Chlorohexidine	55–55-1	Category 1	Solid	100
3	Trichloroacetic acid(30%)	76–03-9	Category 1	Liquid	20
4	2,6-Dichlorobenzyl chloride	4659–45-4	Category 2A	Liquid	20
5	Ethyl-2-methylacetoacetate	609–14-3	Category 2B	Liquid	20
6	Dimethyl sulfoxide	67–68-5	Category 2B	Liquid	20
7	Cyclopentanol	96–41-3	Category 2B	Liquid	20
8	Decamethylcyclopentasiloxane	541–02-6	Not classified	Liquid	20
9	EDTA	21,502–12-9	Not classified	Solid	100

^aChemical Abstracts Service Registry Number

^bBased on the in vivo rabbit eye test (OECD TG 405) and UN GHS (Globally Harmonized System of classification and labelling of chemicals). Category 1: Irreversible eye damages, Category 2: Classification as 2A or 2B depends on the interpretation of the UN GHS criteria for distinguish between categories

Preparation of bovine cornea

Bovine eyeballs were obtained from the Kyung-Shin slaughterhouse in Gyeongsan, Korea, and transpoted them with HBSS (4 °C) which included 1% P/S. The corneas were used within 3 h of slaughtering. After confirming the eyeballs visually, the corneas were incised including 2 to 3 mm of the sclera.

BCOP assay

We used three bovine corneas for each test chemical, and the BCOP assay about all test chemicals was proceeded once. The corneal epithelium was put on the upper chamber of the BCOP holder and fixed with screws. Both chambers were filled with medium without phenol red and incubated at 32 °C for 1 h. After the incubation, both chambers were filled with new medium without phenol red and the basal opacity was measured. Corneas with an opacity over 7 were not suitable for use. As positive and negative controls, ethyl alcohol and DW were used, respectively. The medium in the upper chamber was removed, and the corneal epitheliumwas treated with 750 μL of the test chemicals. After the treatment, we rinsed the cornea using medium with phenol red more than 3 times and medium without phenol red. Liquid chemicals were added post-incubation into medium without phenol red and incubated for 2 h. The lower chamber was filled with medium and the upper chamber with the cornea was treated with 1 mL of fluorescein sodium solution for 90 min at 32 °C. After the post-incubation, the medium in the lower chamber was extracted and 300 μL was distributed into a 96 well plate to measure the absorbance using a spectrophotometer at 490 nm.

Analysis in BCOP assay

Opacity and permeability were calculated by OECD test guideline 437. Basal opacity was calculated from the lux value of the cornea. Each lux was obtained from the holder containing medium and test chemicals. The permeability was determined by the amount of fluorescein sodium penetrated to the lower chamber. IVIS was calculated by the opacity and permeability, and the IVIS classification followed OECD test guideline 437 (Table 2).

Table 2.

In vitro irritancy score (IVIS) in BCOP assay

IVIS	UN GHS classification
≤ 3	Not classified
< 3, ≤ 55	No stand-alone prediction can be made
> 55	Category 1

Preparation of slides

After the BCOP assay, the corneal tissue was fixed in 10% formalin for 24 h. Paraffin-embedded tissue was cut into 4 μm section. We used three staining methods, H&E, MT and PAS, to observe histopathological changes in the layer of epithelial, stromal and endothelial layers. In H&E staining, nuclei were stained with alum hematoxylin. Acid alcohol (0.3%) was used to differentiate for slides. After staining with eosin, the slides were dehydrated and mounted. For the MT staining method, after Mayer's hematoxylin staining, the slides were washed and rinsed. THe slides were stained with Biebrich scarlet-acid fuchsin solution for 15 min and rinsed. The differentiated slides were put into a phosphomolybdic–phosphotungstic acid solution, until the collagen was not red. The slides were washed and transferred to a light green solution. After rinsing, they were dehydrated with 95% ethyl alcohol and absolute ethyl alcohol and washed with xylene. They were mounted with a resinous medium. In PAS staining, the slides were oxidized using periodic acid solution (0.5%), rinsed and stained with Schiff reagent. Then, the slides were rinsed and checkedfor a color change to dark pink. Mayer's hematoxylin was used as a counter stain. The slides were washed, desiccated, and mounted.

Histopathological evaluation

We used an optical microscope to observe histopathological changes in the corneal tissue. To observe the degree of damage, lesions, and edema induced by the test chemicals, the bovine corneas were categorized by the IVIS in OECD test guideline 437. The characteristics of the three special staining methods, were used to perform the histopathological evaluation.

Statistical analysis

SPSS 21.0 for Windows (SPSS Inc., USA) was used to perform, one-way ANOVA and Duncan's multiple range test. The significance level was set at p < 0.05.

Results

BCOP assay

All irritant test chemicals, except Nos. 5 and 6, showed higher opacity than the positive control. No. 2 showed high opacity, but its permeability was low value. All non-irritant test chemicals showed low opacity and permeability. All severe eye irritants showed an IVIS of > 55 and the non-irritants were under 3. No. 7 showed mild irritation (Category 2B) in the in vivo GHS classification, but was identified as a severe eye irritant in this assay. In addition, Nos. 4, 5, and 6 with irritation or mild irritation (UN GHS categories 2, 2A, or 2B), were classified as 'No stand-alone prediction can be made’ (Table 3). After the BCOP assay, corneal tissues were fixed in 10% formalin and the three special staining methods were used for histopathological evaluation of the corneas.

Table 3.

The results of 9 test chemicals in BCOP assay

Number	Opacity	Permeability	IVIS	In vivo GHS classificationf	BCOP classification
NC	1.2 ± 0.3ab	0.016 ± 0.007a	1.4 ± 0.3ab	Not classified	Not Classified
PC	35.6 ± 3.0d	2.166 ± 0.015e	66.9 ± 3.1e	Category 2A	Category 1
1	62.6 ± 7.9f	2.053 ± 0.089de	93.4 ± 8.6f	Category 1	Category 1
2	90.5 ± 4.9g	0.039 ± 0.003a	91.1 ± 4.9f	Category 1	Category 1
3	208.5 ± 9.5h	1.893 ± 0.251cd	236.9 ± 12.7g	Category 1	Category 1
4	13.9 ± 0.7c	0.741 ± 0.017b	25.1 ± 1.0d	Category 2A	No stand-alone prediction can be made
5	8.3 ± 1.1bc	0.016 ± 0.011a	8.6 ± 0.9bc	Category 2B	No stand-alone prediction can be made
6	5.7 ± 2.1bc	0.668 ± 0.085b	15.7 ± 3.1cd	Category 2B	No stand-alone prediction can be made
7	45.2 ± 10.2e	1.817 ± 0.306cd	72.4 ± 7.6e	Category 2B	Category 1
8	− 3.9 ± 0.7ab	0.004 ± 0.018a	− 3.8 ± 0.6ab	Not classified	Not Classified
9	2.2 ± 0.5ab	0.053 ± 0.010a	3.0 ± 0.6ab	Not classified	Not Classified

^fBased on the in vivo rabbit eye test (OECD TG 405) and using the UN GHS (Globally Harmonized System of classification and labelling of chemicals). Category 1: Irreversible eye damages, Category 2: Classification as 2A or 2B depends on the interpretation of the UN GHS criteria for distinguish between categories.Values with different letters of same column are significantly different (p < 0.05) by ANOVA and Duncan’s multiple range test

Histopathological evaluation by H&E staining

Corneas treated with non-irritant chemicals (Nos. 8, 9, and NC) showed stable corneal structures. Nos. 4, 5, and 6 (3 < IVIS ≤ 55) showed edema and loss of the squamous cell layer. However, loss of the squamous cell layer was not observed in Nos. 5 and 6. These chemicals (Nos. 4, 5 and 6) showed an increase of corneal epithelial vacuoles in the epithelium. In the irritant group, because of the loss in squamous and wing cell layers, corneal exuviation was observed. Corneal epithelial vacuoles were observed in the PC and No. 2, and severe corneal edema was observed in No. 1 compared to the other corneas. In No. 3, which showed the highest IVIS, the nuclear condensation was observed in basal cell layer (Fig. 1) (Table 4).

Fig. 1.

Histopathological evaluation in H&E staining (× 100). a Edema in squamous cell layer and epithelium layer, b Vacuoles around the epithelial cell, c Exuviation in squamous cell layer and epithelium layer, d Severe swelling of whole cornea and condensation in basal cell layer

Table 4.

Total results of histopathological evaluation for 9 test chemicals

Test chemical no. (IVIS)	Characters of histopathological observation
	H&E staining method	MT staining method	PAS staining method
NC (1.4 ± 0.3ab)	No lesion in corneal epithelium	No lesion in corneal stroma	No lesion in corneal basement membrane
PC (66.9 ± 3.1e)	Little corneal epithelial vacuoles	Loose stromal layer	No lesion in corneal basement membrane
1 (93.4 ± 8.6f)	Loss and exuviation of squamous cell layer	Loose stromal layer and increase of vacuoles around keratocyte	Clear change was not showed
2 (91.1 ± 4.9f)	Loss and exuviation of squamous cell layer	Increase of stromal layer distance	Increase of basement membrane bleaching
3 (236.9 ± 12.7g)	Sever edema in epithelium layer	Extremely loose stromal layer	Reduction of visibility in basement membrane
4 (25.1 ± 1.0d)	Increase of corneal epithelial vacuoles and swelling	Little increase of vacuoles around keratocyte and loose stromal layer	Little bleached basement membrane
5 (8.6 ± 0.9bc)	Increase of corneal epithelial vacuoles and swelling	Little increase of vacuoles around keratocyte and loose stromal layer	Little bleached basement membrane
6 (15.7 ± 3.1cd)	Increase of corneal epithelial vacuoles	Increase of stromal layer distance	Little bleached basement membrane and increase of vacuoles
7 (72.4 ± 7.6e)	Vacuoles around the basement membrane	Increase of stromal layer distance	Reduction of visibility and increase vacuoles in basement membrane
8 (− 3.8 ± 0.6ab)	Stable corneal structure	Stable morphological sturcture in stroma	Stable structure in basement membrane
9 (3.0 ± 0.6ab)	Stable corneal structure	Stable morphological sturcture in stroma	Stable structure in basement membrane

Values with different superscript letters of IVIS are significantly different (p < 0.05) by ANOVA and Duncan’s multiple range test

Histopathological evaluation by MT stainin

Collagen bundles in the stroma of corneas exposed to non-irritating chemicals showed robust, not loose morphological structures, and no vacuoles were observed surrounding the keratocytes. Test chemicals (3 < IVIS ≤ 55) had loosened collagen bundles and showed vacuoles around the keratocytes. In the severe irritant chemicals, histopathological changes in the stroma were confirmed. The vacuoles around the keratocytes and widened distance of collagen layer were increased as the irritation increased (Fig. 2) (Table 4).

Fig. 2.

Histopathological evaluation in MT staining (× 100). a Loosened and widened collagen bundles in stroma, b Increase vacoules around keratocyte and loosened collagen bundles (or collagen layers) in stroma

Histopathological evaluation by PAS staining

We found deep purple color in the epithelium, which was a PAS positive reaction with glycogen, suggested no damage in the epithelium and basement membrane. Corneal epithelial layers (3 < IVIS ≤ 55) showed bleaching, swelling and vacuolation. For example, No. 3 showed a reduction of visibility in the basement membrane, and increased bleaching in the epithelium. All corneal epithelial layers were bleached by severe irritants. In the epithelium, the basement membrane showed a significant decrease in visibility compared to the other corneas, which had a low IVIS. The visibility of the basement membrane decreased as irritation increased, but there was no morphological change (Fig. 3) (Table 4).

Fig. 3.

Histopathological evaluation in PAS staining (× 100). a Edema and vacuoles in epithelium layer, b Reduction of visibility in basement membrane and increase of bleaching in epithelium layer, c Reduction of visibility and thickness in the basement membrane

Grading criteria of histopathological evaluation

The criteria for grading the histopathological evaluation were made based on the staining results (Table 5). When no lesions and robust corneal structures were observed, a histopathological grade of 0 was assigned (Non-irritant). If a lesion in the epithelium was observed, the irritant was classified as grade 1 (Mild irritant). When damage to the epithelium and stroma was observed, the irritant was classified as grade 2 (Moderate). Lesions on the whole cornea, indicated a grade 3 irritant (Severe irritant).

Table 5.

The grade criteria for histopathological evaluation

IVIS	H&E staining method	MT staining method	PAS staining method		Histopathological grade
	Epithelium	Stroma	Epithelium	Basement membrane
≤ 3	–	–	–	–	0 (Non-irritant)
> 3,  ≤ 25	Edema	No lesion	Edema	No lesion	1 (Mild)
> 25,  ≤ 55	Edema: up to wing cell Exuviation Vacuoles	Loosen collagen bundle Vacuoles around keratocyte	Bleaching	Decrease of visibility	2 (Moderate)
> 55	Edema: up to basal cell Exuviation Vacuoles Nuclear condensation	Increase of loosen collagen bundle Increase of vacuoles around keratocyte	Increase of bleaching	Decrease of visibility Decrease of thickness	3 (Severe)

Discussion

The BCOP and ICE assays are alternative animal tests for the in vivo Draize test and aid in classifying eye-irritating chemicals [18]. The BCOP assay calculates the IVIS from the opacity and permeability of a treated cornea [1, 19]. Test chemicals with an IVIS of less than 3 are classified as non-irritants, and over 55 are classified as severe irritant in OECD test guideline 437. However, an IVIS of > 3 but ≤ 55 in the BCOP assay shows the limitation of the assay, which can-not classify the irritations like UN GHS categories 2, 2A, or 2B for some chemicals. To complement the BCOP assay, various studieshave been conducted, and histopathological evaluation is recommended [20, 21]. The bovine cornea is comprised of epithelium, stroma, and endothelium, which are necessary to classify the degree of eye irritation as mild, moderate, or severe irritation [22]. H&E staining allows for the observation of tissue structures and characteristics of the cytoplasm, nucleus and extracellular matrix [23, 24]. PAS staining makes it easier to observe the basement membrane compared to H&E staining. The corneal stroma is composed of collagen fibers. For collagen fibers, triple or four-fold staining such as with van Gieson, Mallory, or Masson trichrome (MT) staining are performed. MT staining has quickly replaced the classical technique of H&E staining. For the above reasons, H&E staining is not enough to observe the histopathological changes in the bovine cornea. To suggest information about eye irritation prediction of test chemicals, we conducted histopathological evaluations using H&E, MT, and PAS staining methods to assess the bovine corneal structure.

In this study, we performed the BCOP assay and various special staining methods (H&E, MT, and PAS) to confirm the histopathological evaluation in bovine cornea. Corneas treated with non-irritant chemicals showed stable morphology and no lesions in all staining methods. When we observed the stroma, the collagen bundles were showed connected strongly, so we found that the test chemicals did not pass through the stroma. PAS staining produced as positive reaction with glycogen in the cornea [25]. Because of this reaction, superior epithelium and basement membrane in the cornea were stained a dark purple color when treated with non-irritant chemicals.

Severe irritants affected wing and basal cells and produced severe edema and lesions. Similar to some previous studies, in this study, we found that severe irritants such as imidazole (No. 1) and 30% trichloroacetic acid (No. 3) increased corneal thickness [15, 26]. The reason for this was corneal protein condensation, which increased opacity and thickness [13, 27]. The wing cell undergoes re-epithelialization during wound healing, and the basement membrane produces new epithelial cells for of epithelial wound recovery [28]. Loose collagen and vacuoles around the keratocytes were observed in the stroma by MT staining, and visibility of the basement membrane was highly decreased, whereas epithelium bleaching was increased by PAS staining. From these results, we concluded the test chemicals penetrated the corneal epithelium and induced structural changes of the cornea and damage in the stroma. These lesions confirmed the potential of severe eye irritation [13]. Moreover, another study evaluated eye irritation by swelling and the hole size of the keratocyte vacuoles [22].

We suggested histopathological evaluation gading criteria as 0 (Non-irritant), 1 (Mild), 2 (Moderate) and 3 (Severe) using various special staining methods. The depth and degree of corneal damage have been reported as top-down irritation reaction by corneal structure [22]. For this reason, corneal morphological changes will progress from the squamous epithelium to the endothelium. In this study, histopathological evaluation was performed with stepwise analysis according to corneal structure and the range of corneal damage, and it could provide a information for the IVIS classification of eye irritant chemicals like UN GHS categories 2, 2A, or 2B.

We expect that the histopathological results of this study will suggest other applications of special staining after the BCOP assay, and additional histopathological evaluation of corneal thickness, the degree of damage, and lesions, will help to anticipate and classify test chemical irritation. The histopathological grade criteria in this study will support the evaluation and classification of irritation after the BCOP assay. Furthermore, it will increase the accuracy of the BCOP assay results, and continuous histopathological evaluation studies will further improve these alternative eye irritation tests.

Compliance with ethical standards

Conflict of interest

The authors have no conflict of interest to disclose.