Content uploaded by Raid G Alany
Author content
All content in this area was uploaded by Raid G Alany on Mar 02, 2015
Content may be subject to copyright.
Critical appraisal of alternative irritation models:
three decades of testing ophthalmic
pharmaceuticals
Hamdy Abdelkader†,‡, Barbara Pierscionek†, Mark Carew†, Zimei Wu§,
and Raid G. Alany†,§,*
†
Faculty of Science, Engineering and Computing, Kingston University London, Kingston upon Thames, UK,
‡
Department of Pharmaceutics, Faculty of Pharmacy, Minia University, Minia, Egypt, and
§
School of
Pharmacy, Faculty of Medical and Health Sciences, The University of Auckland, Auckland, New Zealand
*Correspondence address. School of Pharmacy and Chemistry, Kingston University London, Penrhyn Road, Kingston upon
Thames, KT1 2EE, UK. E-mail: r.alany@kingston.ac.uk, r.alany@auckland.ac.nz
Accepted 22 December 2014
Abstract
Background: Testing ocular tolerability of ocular pharmaceuticals is an
essential regulatory requirement. The current approved reference model
(gold standard) for ocular irritation testing is the Draize test. However this
method is subjective and involves using live animals, hence the need to
develop alternative in vitro and ex vivo testing strategies.
Source of data: Pubmed, Science Direct, Scopus, Google Scholar, Medline,
Current Content, Web of Science and validation reports from international
regulatory bodies; The Interagency Coordinating Committee on the Valid-
ation of Alternative Methods (ICCVAM) and European Centre for the Valid-
ation of Alternative Methods (ECVAM) were searched for in vitro alternatives.
Area of agreement: Whilst no single in vitro test can effectively replace the
Draize eye irritation test, regulatory bodies and cosmetic/pharmaceutical
industries agree that there is a need for in vitro alternatives with validated
endpoints to evaluate pharmaceutical ingredients and finished eye products.
Area of controversy: There is no single in vitro test / assay that can predict
the ocular irritation potential of mild to moderate test substances.
Area timely for developing research: This review provides a critical appraisal
of the selected in vitro and ex vivo ocular toxicity models recommended by
international regulatory bodies. These include cytotoxicity methods, biochem-
ical systems and ex vivo assays. The latter are approved by ECVAM as in vitro
alternatives for the well-known Draize test. Hen’s egg test-chorioallantoic
British Medical Bulletin, 2015, 1–13
doi: 10.1093/bmb/ldv002
© The Author 2015. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
British Medical Bulletin Advance Access published February 16, 2015
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
membrane and the isolated rabbit eye test are also accepted by regulatory
agencies in France, Germany, the Netherlands and the UK. A combination of
ex vivo assays along with histological examination of excised bovine cornea
can predict the conjunctival and corneal tolerability and cover a wider range
of ocular pharmaceutical substances.
Key words: Draize test, ocular toxicity, irritation, cornea, hen’s egg chorioallantoic membrane, bovine cornea, opacity
and permeability
Introduction
In recent decades, researchers testing ocular dosage
forms have recorded toxicological signs of ocular
tissues exposed to topically applied cosmetics and phar-
maceuticals. Ocular tissues, such as the cornea and con-
junctiva, are susceptible to injuries and adverse ocular
effects, either from the administered drug or from
pharmaceutical ingredients (excipients) used in formu-
lating ophthalmic products.
1,2
Antifungal agents, such
as amphotericin B and ketoconazole, can cause corneal
oedema and corneal abnormalities when administered
topically.
3
Excessive use of topical anaesthetics can
produce corneal lesions and ulcers.
1
Anti-inflammatory
corticosteroids have been shown to retard epithelial
corneal wound healing and induce glaucoma.
1,2,4
Ocular side effects due to excipients used in
pharmaceutical products have been reported. Many
surfactants have been studied as ocular penetration
enhancers. Non-ionic surfactants including Span 20,
Span 40, Span 85, Tween 20, Tween 40, Tween 81,
Brij 35, Brij 58, Myrj 52 and Myrj 53 were tested for
their ability to increase human corneal permeability
of fluorescein.
5
Benzalkonium chloride (BAC), a qua-
ternary ammonium cationic surfactant, is a com-
monly used preservative in ophthalmic products. BAC
has been reported to cause corneal opacification, a
decrease in corneal epithelial microvilli, conjunctival
hyperaemia (red eye) and delayed wound healing.
2
The use of surfactants in ophthalmic formulations
and ocular drug delivery is occasionally recom-
mended to increase bioavailability.
6
A preliminary requirement of any ocular formula-
tion is the lack of local adverse effects, e.g. burning,
stinging and tearing. However, this is not always the
case, especially with newly developed advanced drug
delivery systems, e.g. micellar, reverse micellar, liquid
crystalline, microemulsions, niosomes, liposomes,
organogels and coarse emulsions. Assessment of the
toxicity of ophthalmic formulations and the potential
for ocular side effects represents an essential step in
the development of new ocular formulations.
1,7
On
the regulatory side, there is relatively little guidance
from the International Conference on Harmonisation
(ICH) for non-clinical toxicity studies on ocular
drugs, including those with novel additives.
8,9
European regulatory authorities recommend an
in vivo ocular tolerance study (CPMMP/SWP/21/
00), where a single dose (20–30 µl) of the test formu-
lation is investigated in a small number of New
Zealand white rabbits (one to three) for any ocular
abnormalities that are recorded and scored.
9
The standard in vivo eye test (Draize eye test) has
been around since 1940s.
10
Accordingly, 100 µl of
liquid or 100 mg of a solid test material are placed
onto the lower conjunctival sac of the albino rabbit
eye and the eyelids are closed for a fixed time period.
Ocular responses are assessed by an observer at
various time intervals for a period of up to 3 weeks
after treatment. The contralateral eye is used as an
untreated control. Three ocular tissues, namely the
cornea, iris, and conjunctiva are scored for their irri-
tation responses. A numerical score is given to differ-
ent aspects of the irritation response observed. The
maximum score is 110, with 80 points possible from
the cornea, 10 from the iris and 20 from the conjunc-
tiva. Because damage to the cornea is most critical,
its contribution to the overall score is the greatest.
This methodology has been widely challenged as it
lacks reproducibility.
11
In addition, the use of sentient
animals for ocular irritation testing has become a
major target of criticism by animal welfare organiza-
tions. Thus, there has been an increased ethical and
2H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
scientific demand to use alternatives for ocular toler-
ability rather than testing on live animals.
12
Different in vitro and ex vivo testshavebeenused
for assessing the ocular toxicity of ophthalmic dosage
forms and will be covered in this review; they are
(i) Cytotoxicity tests using cultured and isolated
cells
(ii) Biochemical systems based on synthetic pro-
teins (EYETEX
®
/IRRITECTION
®
)
(iii) Red blood cell (RBC) haemolysis assay
(iv) Slug mucosal irritation (SMI) test
(v) Hen’s egg test-chorioallantoic membrane (HET-
CAM)
(vi) Bovine corneal opacity and permeability (BCOP)
(vii) Combination of HET-CAM and BCOP
(viii) Histological documentation of treated bovine
corneas
The most appropriate in vitro toxicity tests should
target the same endpoints as the Draize rabbit eye
test in a more objective manner.
13
Corneal opacity
has been simulated by assays that measure protein
coagulation, for example the IRRITECTION
®
assay, the RBC lysis test (which can also measure
protein denaturation) and the BCOP test. Inflam-
mation is a complex process that involves, amongst
other factors, production of mediators ( prostaglan-
dins, leukotrienes, cytokines, etc.), increased micro-
vascular permeability and invasion of white blood
cells. Such a complex cascade of events may be
better modelled using organotypic tests (e.g. BCOP,
HET-CAM). The European Centre of the Valid-
ation of Alternative Methods (ECVAM) and Scien-
tific Advisory Committee (ESAC) announced the
validation of BCOP and isolated chicken eye tests in
2007. The HET-CAM and isolated rabbit eye test
are accepted by regulatory agencies in France,
Germany, the Netherlands and the UK.
Cytotoxicity tests using cultured and
isolated cells
There are a number of simple in vitro cytotoxicity
assays that measure cellular functions such as growth,
metabolic viability and plasma membrane integrity.
The recommendation from ECVAM has been for the
use of four cytotoxicity assays: the Cytosensor Micro-
physiometer (CM), neutral red uptake, RBC haemoly-
sis and fluorescein leakage (FL).
14
These in vitro tests
are recommended for drugs ranked at the bottom of
the list of potential irritants. For drugs that are likely
to be strong irritants or corrosives, organotypic tests
(BCOP, HET-CAM) have been recommended. Hence,
some assays have become more popular, others less
so,asvalidationandtestinghelpbetteridentifythe
most appropriate replacements for the Draize test.
An early comparison of in vitro test results with in
vivo scores obtained using the Draize test was con-
ducted utilizing results from a group of seven in vitro
assays.
15
The most promising correlation with in vivo
data was observed with the BCOP, HET-CAM and
the former EYETEX
®
assay (currently called IRRI-
TECTION
®
). In vitro tests that correlated less satis-
factorily with the in vivo data were the TOPKAT
comparison (in silico prediction using structure–
activity data), neutral red uptake (cytotoxicity),
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) in living dermal equivalent (cyto-
toxicity) and MICROTOX (cytotoxicity in bacteria)
assays. To date, no single in vitro test/assay has been
fully accepted as an alternative to the Draize test by
the main regulatory bodies.
14
CM
The CM test is a potentiometric assay that measures
changes in extracellular pH, due to the release of
acid equivalents from cells, as an index of metabolic
rate, e.g. glucose utilization. The CM test has been
accepted by ECVAM for the detection of non-irritants
from amongst water-soluble surfactants and water-
soluble surfactant-containing mixtures.
16
The Organ-
isation for Economic Co-operation and Development
(OECD) has issued a Test Guideline describing in
detail the CM test, its development from INVITTOX
protocol No 102 and evaluation by ECVAM,
ICCVAM, amongst others. Briefly, mouse L929 fibro-
blasts are cultured on a porous membrane/electrode
and the release of acids is measured by a potentiom-
eter. Negative control, positive control (sodium lauryl
sulphate) and test substances are applied. The MRD
50
values (the dose of the test material that induces a
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 3
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
50% decrease in metabolic rate relative to a negative
control) of these chemicals are calculated and com-
pared against cut-off values of MRD
50
for non-
irritants, mild/moderate irritants and strong irritants.
The CM test is unique in that it measures ‘extracellular
acidification’(‘change in pH’)asaproxyofintracellu-
lar metabolism. Other aspects of cell metabolism may
also be measured using older methods such as amino
acid uptake for protein synthesis, or 5[H]-uridine
incorporation for DNA production.
RBC haemolysis
RBCs are readily available from the local abattoir,
no ethics application is required and their handling
does not require cell culture facilities. The RBC
assay has been used as a screening test in the soap
and detergent industry.
17
The principle of the test
method is based on the observation that certain
classes of chemical irritants may cause damage to
cell membranes and/or cause denaturation of
membrane proteins and other cellular proteins.
18
Such events can be correlated with the initial inflam-
matory response in tissue irritation and with
changes in protein conformation, such as those that
occur in opacification of the cornea. The biological
endpoints of cell lysis and protein denaturation
have been used singly or in combination to predict
the ocular irritation of certain classes of chemicals
(mainly surfactants).
The RBC assay is based on testing the membranoly-
tic and protein denaturation activity of test substances
employing mammalian erythrocytes isolated from
fresh blood. The percentage release of the native dye
oxyhaemoglobin (HbO
2
) as a result of lysis of fresh
blood cells is measured using UV-Vis spectrophotom-
etry (Fig. 1). The concentration of a test substance that
induces 50% lysis of the RBCs is an endpoint (H50%)
for release of haemoglobin and cell lysis. Another end-
point is spectrophotometric measuring of denatured
oxyhaemoglobin due to interaction with the test sub-
stance as an indicator of protein denaturation.
19
Fig. 1 Percentage release of HbO
2
due to lysis of fresh blood cells in response to a pharmaceutical excipient
(cyclodextrin). The response (increased optical density/colour intensity with increased concentration) is
quantified using UV-Vis spectrophotometry at 540 nm (taken from in house developed RBC assay; unpublished
data).
4H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
The RBC assay has been used to test potential
pharmaceutical excipients such as cyclodextrins
(Fig. 1) to predict the cytotoxicity of mild to mod-
erate irritants. The RBC assay has been undergoing
validation by the Validation Management Group
(VMG).
20
The test is not effective for substances of
limited aqueous solubility. Overall the test shows
good intra- and inter-laboratory reproducibility
(reliability) and represents one of the promising in
vitro alternatives to in vivo testing with a good fit
to prediction models for the assessment of acute
ocular irritation caused by surfactants and their
formulations.
Fluorescein dye leakage
The FL test is recommended by ECVAM (as above)
and also by OECD, as an alternative to the Draize
test. Madin Darby Canine Kidney (MDCK) epithelial
cells are grown on a permeable Millicell™insert.
21
The MDCK monolayer is used to model the in vivo
corneal epithelium. Test substances are then applied,
including a positive control (Brij 35) to a corneal epi-
thelium monolayer. The leakage of fluorescein from
the apical to basolateral compartments (as result of
damage to tight junctions for example) is measured.
The extent of FL is calculated from the concentration
of fluorescein in samples obtained from the basolat-
eral compartment, quantified using a spectrofluorom-
eter and compared with an appropriate blank, and a
positive control.
Neutral red uptake
Neutral red uptake (NRU) measures the viability of
cells through their ability to take up and concentrate a
dye (neutral red) into lysosomes. Cells expend energy
to protons pump into lysosomes and preserve the
acidic interior of these organelles. The neutral red
assay has many advantages, including greater sensitiv-
ity than the MTT assay.
22
The NRU may also be fol-
lowed by a sulforhodamine B (SRB) protein assay,
thereby using two assays with different endpoints on
the same plate and test chemicals. The NRU assay
with the mouse balb/c 3T3 fibroblast cell line has
been proposed as a replacement assay.
23
MTT assay
The MTT assay is one of the best known in vitro cellu-
lar viability tests. The endpoint is the function of a
mitochondrial enzyme (succinate dehydrogenase) that
converts MTT (a yellow substrate) into an insoluble
purple formazan product. Reduction in mitochondrial
function, e.g. after a toxic effect (of drug or excipient),
reduces the intensity of the MTT product, compared
with control cells. Newer substrates, with better sensi-
tivity, (e.g. XTT) or other advantages in assay
throughput (MTS, WST-1) have been introduced. The
MTT assay does not currently feature in recommenda-
tions by ECVAM or OECD. However, the short-time
exposure (STE) test uses the MTT assay in CCL-60
rabbit corneal epithelium cells,
24
as does the EpiOcu-
lar™and SkinEthic™3D models (see below). The
STE is under consideration by the OECD for the iden-
tification of category 1 chemicals.
15
Cell growth assays
Cell growth/number or cell proliferation may be
monitored by a host of assays, some of the simplest
of which measure cellular protein by traditional
assays (Lowry, Crystal Violet and Coomasie Blue).
The newer SRB protein assay is more sensitive than
other protein assays using visible dyes and is efficient
in terms of using low-cost chemicals and basic equip-
ment.
25
Cell growth assays are extensively used in
carcinogenic and anti-cancer drug testing. However,
the cornea of the eye is fully differentiated in adults
with no further cell growth (unless injured), so cell
growth assays are of limited value. Whilst animal
epithelial cells lines serve as good models for human
epithelium,
26
there is a greater move to develop
human cell lines that grow in a manner more akin to
what is found in the physiological conditions.
27
3D cell cultures
Immortalized cells, or continuous cell lines, are
extremely useful in toxicity testing, whether for ocular
irritation, dermal irritation or for testing putative anti-
proliferative/anti-cancer drugs. Many of the assays
discussed above are performed using epithelial, or
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 5
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
non-epithelial, cells cultured on an impermeable
(often plastic) growth matrix. Epithelial cells are dis-
tinctive in having a basolateral membrane and an
apical membrane, with different ion channels, trans-
porters and receptors, which are critical for the trans-
port of ions and other molecules. Cells that are allowed
to grow and polarize on a permeable membrane (such
as a Transwell™or Millicell™) form tight junctions
and desmosomes between cells and create the barrier
function of an intact epithelium.
New developments in cell culture use permeable
supports coated with appropriate in vivo proteins for
attachment and allow cells to grow as a tissue with a
3D architecture. The 3D Vitrigel™assay uses human
corneal epithelial (HCE-T) cells grown in a collagen
matrix to represent connective tissue.
27
Transepithe-
lial electrical resistance (TEER) measurements are
taken using a handheld voltmeter, or when the support
is mounted in an Ussing chamber. Positive controls are
drugs known to induce damage to the monolayer and
reduce TEER, or to increase FL. The EpiOcular™
culture model features human keratinocytes grown
on a permeable support. An air–liquid interface is
adopted so that the cells form many layers, in common
with the corneal epithelium. Toxicity testing is by the
MTT assay.
13
The SkinEthic™model uses reconsti-
tuted human corneal epithelium. There is no air–liquid
interface, so the cell model is akin to the corneal
mucosa. Cellular viability is measured by the MTT
assay. Histological quantification is also possible, as is
the detection of lactate dehydrogenase (LDH, a large
intracellular enzyme) or cytokine release. For example,
HCE-T cells were frozen, cut into 5 μm sections using
a Cryostat, stained with haematoxylin and eosin
(H&E) and probed with antibodies for ZO-1 (tight
junctions). Visualization of the 3D cellular structures
after exposure to test chemicals allowed a toxicological
assessment of erosion of cells and disruption to tight
junctions.
27
All three models are under validation by
ECVAM.
13
Biochemical systems based
on synthetic proteins
The interaction of a test material with biological mole-
cules, such as lipids and proteins, and the subsequent
changes have been shown to contribute to irritant
effects.
28
The IRRITECTION
®
assay (In Vitro Inter-
national, Irvine, USA) is the updated protocol of the
EYETEX
®
assay. IRRITECTION
®
is a commercial
testkitthatquantifies the irritation potential of a test
material based on interaction with a synthetic protein
matrix. Small quantities of test sample when presented
to the IRRITECTION
®
reagent produce turbidity,
which is directly proportional to the irritation poten-
tial of the sample. This turbidity is produced by
changes in conformation and degree of hydration of
the IRRITECTION
®
reagent proteins which mimics
the response of the proteins in the cornea during
injury or in response to ocular irritants. Turbidity is
compared with that produced by eye irritant standards
ofknownDraizescores.TheDraizeequivalentscoreis
then determined from a calibration curve.
Despite the encouraging testing results obtained
with some hydro-alcoholic formulations tested using
this methodology, EYETEX
®
has shown poor correl-
ation with in vivo testing.
29
For example, the validity
of the EYETEX
®
for the assessment of surfactant-
based products was reported to be questionable and
contradictory.
30
In an investigation of the eye irrita-
tion potential of cosmetic product formulations, an
overall testing error of 20% comprising 18% over-
estimation and 2% underestimation was reported
and was associated with ethoxylated and propoxy-
lated surfactants or alcohol-based formulations.
30
The false results were attributed to basic physico-
chemical incompatibilities of the test material with
the EYETEX
®
reagent system. The current IRRI-
TECTION
®
assay is a new version of the former
EYETEX
®
, which has numerous assay protocols
leading sometimes to inconsistent results. There
remains a need for more data on the validity of this
test before it could be considered by regulatory agen-
cies. Furthermore, the IRRITECTION
®
test kit is
relatively expensive, compared with other in vitro
tests employing readily available excised tissues.
Slug mucosal irritation assay
The slug mucosal irritation (SMI) assay was developed
to predict the mucosal irritation of pharmaceutical
ingredients and formulations but it is yet to be
6H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
approved by regulatory agencies. It utilizes the terres-
trial slug Arion lusitanicus.
31
Test substances are
applied onto the slug body (shell-less gastropods)
and the induced mucosal secretions are measured
against positive and negative controls, then expressed
as percentage of body weight to interpret the irritation
potential of the test substance. A test substance result-
ing in a total mucus production (MP) <3% is classi-
fied as a non-irritant, 3–8% MP is a slight irritant,
8–15% MP is a moderate irritant and >15% MP is a
severe irritant.
32
In another study, tissue damage was assessed by
measuring the proteins and enzyme released from
the mucosal surface. Mild irritants result in produc-
tion of mucus with only limited effect on proteins
and enzymes. The SMI assay has been used
successfully to predict ocular irritation potential for
cosmetic and pharmaceutical formulations.
31,33
This
test has been developed and assayed in Belgium; the
Belgian A. lusitanicus was used as a test organism.
The MP is species dependent; different species can
generate different responses to test substances. The
SMI assay should be optimized and validated, if
other slug species are used.
34
HET-CAM
The CAM is the vascularized respiratory membrane
that surrounds an embryonic bird developing in the
egg. It is a complete tissue including veins, arteries
and capillaries. The CAM has been recognized as a
possible model for predicting the irritant effect of
Fig. 2 HET-CAM vascular responses used to score controls and test substances; (A) no response, (B) hyperaemia, (C)
haemorrhage and (D) clotting, adopted from Abdelkader et al.
38
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 7
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
chemicals on the conjunctiva.
35
Amodification of
the CAM test, called the HET-CAM has been
reported.
36,37
The HET-CAM is a rapid, sensitive and
inexpensive test, whereby fresh fertilized hen’seggs
are incubated at 37°C ± 0.5 and relative humidity of
62.5 ± 7.5%. On Day 3 of incubation, the shell is
broken and the egg is grown in a purpose-made
chamber; the actual testing is undertaken on Days 9
or 10 of development.
37
The test substance is instilled
onto the membrane, and after 20 s the membrane is
irrigated with water. The membrane is then scored for
irritant effects including hyperaemia, haemorrhage
and coagulation/clotting at 0.5, 2 and 5 min (Fig. 2).
The HET-CAM reduces animal suffering, is
simpler, faster and less expensive than the Draize test
and employs a less subjective and more reproducible
scoring system.
35,39–41
It has been successfully used to
assess the irritation potential of promising pharma-
ceutical systems including microemulsion, liquid
crystalline, o/w emulsion,
35
niosomes,
38
gels
42
and
more recently in situ gelling films/inserts.
43
Accord-
ingly, negative and positive controls comprised 1-cm
Whatman filter paper discs. These were immersed
in phosphate buffer saline (PBS) (negative control)
and 1% w/v BAC solution in PBS (positive control)
and placed onto the surface of CAM. Test ocular
films were moistened by dipping into PBS and then
placed on the surface of the CAM. Conventional
inflammatory responses of hyperaemia, haemorrhage
and clotting/coagulation of the CAM blood vessels
and capillaries were recorded and scored over 5 min
similar to the standard HET-CAM test.
43
Until recently, no single test has completely
replaced the Draize eye irritation test. However, the
HET-CAM test has been accepted by regulatory
bodies in many European countries, e.g. The UK,
France, Germany and the Netherlands to be a full
replacement for animal testing for severe irritants.
20,31
Conversely, the HET-CAM has limitations for testing
water-insoluble substances and substances that stain
and obscure the visibility of the CAM.
20
Isolated organs
Isolated, enucleated eyes (IEEs) from bovine,
44
porcine,
45
chicken
45
and rabbit
46
were used as potential
models for alternative tests to Draize. The results
obtained from the IEEs test correlated well with in
vivo data. IEEs do not require the use of live animals
but they use animal by-products including redundant
animal tissues for other experiments. More advanta-
geously, IEE tests do not require ethics approval. The
main argument against the acceptance of isolated
rabbit eyes is their limited supply from domestic
sources. Therefore, laboratory rabbits are still
needed as eye donors.
47
The BCOP assay is a good example on IEEs
assay. The BCOP assay involves an assessment of
Fig. 3 Corneal manifestations used to score the test substances
in the BCOP test; unstained (left), fluorescein-stained (right),
reused with permission from Abdelkader et al.
38
8H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
corneal opacity and disruption of the corneal barrier
(as assessed by the passage of a fluorescent dye) after
exposure to the test material (Fig. 3). Despite the fact
that corneal injury represents only one endpoint of
in vivo eye irritation testing, damage to the cornea is
so critical that it contributes 80 out of the total 110
points allocated in the Draize test. The BCOP assay
is used across the cosmetic and pharmaceutical indus-
tries to evaluate the ocular irritation potential of sur-
factants, pharmaceutical intermediates and finished
products.
29,44,48
The scientific Advisory Committee of the European
Centre for the Validation of Alternative Methods
(ECVAM) announced the validation of two in vitro
assays: the BCOP and the isolated Chicken Eye (ICE)
test. These two tests are an acceptable replacement of
the Draize test for severely irritant substances; for
mild irritants, animals are still required.
20,31
Combined HET-CAM and BCOP tests
There is a general agreement that the HET-CAM test
serves as a validated ocular model for conjunctival
irritation testing, because it responds to irritant sub-
stances with an inflammatory reaction similar to that
produced by conjunctival tissue.
20,39,49
However,
corneal transparency is critical for normal vision.
50
Consideration must be given to the integrity of the
cornea when using the developed ocular formulation,
hence, the drive to establish the safety of the test
material on both the cornea and the conjunctiva.
51
Combined isolated enucleated eyes and HET-CAM
tests were developed, and these combined tests were
simultaneously compared with the Draize test. A
good correlation was obtained with a broad range of
chemical substances.
51
Conjunctival and corneal irri-
tation of ocular niosomes and discomes and their
Fig. 4 Light micrographs of H&E stained bovine corneas showing full-depth normal corneal layers at two different magnifications
(Aand B); severely injured cornea with epithelial detachment and stromal vacuolization (C) and delayed coagulation and
chromatin formation (D). Micrographs are for in house developed sections (unpublished data).
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 9
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
Table 1 Summary of the in vitro and ex vivo alternatives to the Draize test
Irritation test/assay Evaluation endpoints Regulatory status Irritation
potential range
Chemical classes/formulations Reference
CM Cytotoxicity test based on changes in
extracellular pH as an index of
metabolic rate
Entered into evaluation by
ECVAM/OECD issued test
guidelines
Non-irritant to
irritant
Surfactants and surfactant-based
formulations
20,21
RBC assay Cytotoxicity based on haemolysis and
release of oxyhaemoglobin from RBC
and denaturation of oxyhaemoglobin
Ongoing validation by ECVAM
and ESAC
Mild irritant to
moderate
irritant
Surfactants, surfactant-based
formulations and
cyclodextrins
18
Fluorescein dye
leakage
Cytotoxicity based on permeability of
fluorescein through a monolayer of
cultured epithelial cells
Ongoing validation by ECVAM
and ESAC
Mild to
moderate
Surfactants and surfactant-based
formulations
21
Neutral red uptake Cytotoxicity based on inhibition of uptake
of neutral red dye by a monolayer of
cultured cells
Not validated yet Mild irritant Surfactants and surfactant-based
formulations
20
STE test Cytotoxicity test based on measuring cell
viability via MTT assay
Not validated yet Mild to severe
irritant
Surfactants, surfactant-based
formulations, colour
excipients
20
Biochemical assay
based on
commercial
synthetic proteins
Based on protein denaturation and corneal
damage
Poor correlation to the Draize test
and not subject to
recommended by ECVAM
Mild to
moderate
Surfactants and surfactant-based
formulations
20
Slug mucosal
irritation test
Based on measuring excreted mucus,
proteins and other enzymes by the slug
Submitted by ECVAM for
validation
Mild to severe
irritant
Surfactants, surfactant-based
formulations and colour
excipients
20
HET-CAM Based on vascular effects (hyperaemia,
haemolysis and coagulation) due to
conjunctival damage
Accepted by regulatory agencies in
France, Germany, the
Netherlands and UK/ongoing
validation by ECVAM
Mild to severe
irritant
Surfactants, surfactant-based
formulations, microemulsion,
in situ gels and hydrophilic
inserts/films
38,43
BCOP Based on corneal opacity and permeability.
Histopathological changes are highly
recommended to be incorporated for
endpoint evaluation
Validated by ECVAM Moderate to
severe
irritant
Surfactants, solid materials and
surfactant-based formulations
38
10 H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
ingredients were successfully evaluated by combined
HET-CAM and BCOP assays.
38
Histopathological documentation of the
bovine corneas post-BCOP testing
The main limitation of the BCOP test is that it is not
sufficiently sensitive for discriminating between mild
to moderate irritants. Further, predicting ocular irri-
tation using opacity and permeability endpoints is
challenging when the test substances produce a
delayed reaction by interacting with nucleic acids
and mitochondrial proteins, instead of immediate
loss of epithelial integrity. Histological examination
of the cornea after treatment with test substances
(Fig. 4) can provide a more comprehensive assess-
ment of the depth of injury and cellular damage to
the three principle layers of the cornea.
52,53
More-
over, histological examinations can be used object-
ively to quantify changes in stromal thickness, stromal
vacuolization, nuclei coagulation and delayed chro-
matin formation (Fig. 4).
38
Irritant-induced corneal
opacity results from disruption of the well-ordered
collagen fibres matrix of the stroma and is always
associated with stromal oedema. Measuring the
stromal thickness has been employed to serve as
a quantitative measure of the degree of corneal
opacity.
38
Epithelial defects and stromal oedema are two
common causes of ocular irritation after topical
application of test substances.
54
Cell loss, vacuoliza-
tion (presence of vacuoles or holes), pyknosis (nuclei
coagulation), saponification (due to the effect of
alkali) and separation of cells from Bowman’s
membrane are characteristic lesions observed in the
epithelium in response to physically or chemically
induced trauma. Histological scores should be corre-
lated with opacity and permeability scores obtained
using the BCOP test, preferably in conjunction with
HET-CAM test results. It is accepted that the greater
the stromal oedema, the greater the corneal opacity
found.
50
A summary of all the aforementioned tests in-
cluding their endpoints, regulatory status and chem-
ical classes/formulations tested are summarized in
Table 1.
Conclusion
Three decades of research has failed to yield an
acceptable alternative to the standard Draize eye test.
However, many in vitro alternatives to the standard
rabbit (Draize) test have been employed by formula-
tion scientists and pharmaceutical/cosmetic compan-
ies for screening new eye products. These tests and
assays could alleviate animal suffering by employing
immortal cell lines, non-human blood or excised
tissues from local abattoirs, and are more objective
endpoints compared with the Draize test. Assessment
of ocular damage from cell lysis, protein denatur-
ation, corrosion, saponification and cytotoxicity for
a wide range of chemical substances has been
attempted using these in vitro alternatives. However,
none of these tests can effectively monitor corneal
innervation or corneal sensitivity, nor can they dis-
criminate between substances that cause mild to
moderate irritation, or between substances that exert
a delayed toxic effect at the sub-cellular level. Com-
bining histological examination with conjunctival
responses from HET-CAM, and corneal effects from
the BCOP assay, seems the most logical approach for
preliminary in vitro ocular irritation testing.
Conflict of Interest statement
The authors report no conflict of interest.
References
1. Basu PK. Toxic effects of drugs on the corneal epithe-
lium: a review. Cutan Ocul Toxicol 1984;2:205–27.
2. Li J, Tripathi RC, Tripathi BJ. Drug-induced ocular dis-
orders. Drug Saf 2008;31:127–41.
3. Foster CS, Lass JH, Moran-Wallace K, et al. Ocular tox-
icity of topical antifungal agents. Arch Ophthalmol
1981;99:1081–4.
4. Basu PK, Avaria M, Jankie R. Effect of hydrocortisone
on the mobilization of leucocytes on the corneal wounds.
Br J Ophthalmol 1981;65:694–8.
5. Marsh RJ, Maurice DM. The influence of non-ionic
detergents and other surfactants on human corneal per-
meability. Exp Eye Res 1971;11:43–8.
6. Jiao J. Polyoxyethylated nonionic surfactants and their
applications in topical ocular drug delivery. Adv Drug
Deliv Rev 2008;60:1663–73.
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 11
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
7. Lang JC, Roehrs RE, Rodeheaver DP, et al. Design and
evaluation of ophthalmic pharmaceutical products. In:
Banker DGS, Rhodes CT (eds). Modern Pharmaceutics.
New York: Marcel Dekker, 2002, 626–717.
8. Avalos J, Jacobs A, Wilkin JK. Toxicity testing for ocular
drug products. In: Green K (ed). Advances in Ocular
Toxicology. New York: Plenum, 1997, 261–8.
9. Short BG. Safety evaluation of ocular drug delivery for-
mulations: techniques and practical considerations.
Toxicol Pathol 2008;36:49–62.
10. Draize JH, Woodward G, Calvery HO. Methods for the
study of irritation and toxicity of substances applied top-
ically to the skin and mucus membranes. J Pharmacol
Experimental Ther 1944;82:377–90.
11. Rougier A, Cottin M, de Silva O, et al. In vitro methods:
their relevance and complementarity in ocular safety
assessment. Lens Eye Toxic Res 1992;9:229–45.
12. Bruner LH. Ocular irritation. In: Frazier JM (ed). In Vitro
Toxicity Testing: Applications to Safety Evaluation.
New York, USA: CRC Press, 1992, 149–90.
13. Barile F. Validating and troubleshooting ocular in vitro
toxicology tests. J Pharmacol Toxicol Methods 2010;
61:136–45.
14. Scott L, Eskes C, Hoffmann S, et al. A proposed eye irri-
tation testing to reduce and replace in vivo studies using
bottom-up and top-down approaches. Toxicol In Vitro
2010;24:1–9.
15. Sina JF, Galer DM, Sussman RG, et al. A collaborative
evaluation of seven alternatives to the Draize eye irrita-
tion test using pharmaceutical intermediates. Fundam
Appl Toxicol 1995;26:20–31.
16. Hartung T, Bruner L, Curren R, et al. First alternative
method validated by a retrospective weight-of-evidence
approach to replace the Draize eye test for the identifica-
tion of non-irritant substances for a defined applicability
domain. ALTEX 2010;27:43–51.
17. Pape WJW, Hoppe U. Standardization of an in vitro
red blood cell test for evaluating the acute cytotoxic
potential of tensides. Arzneimittelforsch 1990;40:
498–502.
18. Alves EN, Presgrave Rde F, Presgrave OA, et al. A
reassessment of the in vitro RBC haemolysis assay with
defibrinated sheep blood for the determination of the
ocular irritation potential of cosmetic products: compari-
son with the in vivo Draize rabbit test. Altern Lab Anim
2008;36:275–84.
19. Mehling A, Kleber M, Hensen H. Comparative studies
on the ocular and dermal irritation potential of surfac-
tants. Food Chem Toxicol 2007;45:747–58.
20. McNamee P, Hibatallah J, Costabel-Farkas M, et al. A
tiered approach to the use of alternatives to animal
testing for the safety assessment of cosmetics: eye irrita-
tion. Reg Toxicol Pharmacol 2009;54:197–209.
21. OECD, Fluorescein leakage test method for identifying
ocular corrosives and severe irritants, Section 4 in OECD
guideline for the testing of chemicals 2012, 16.
22. Repetto G, del Peso A, Zurita JL. Neutral red uptake
assay for the estimation of cell viability/cytotoxicity.Nat
Protoc 2008;3:1125–31.
23. Ahn JH, Eum KH, Lee M. Assessment of the dermal and
ocular irritation potential of lomefloxacin by using in
vitro methods. Toxicol Res 2010;26:9–14.
24. Takahashi Y, Hayashi T, Watanabe S, et al. Inter-
laboratory study of short time exposure (STE) test for
predicting eye irritation potential of chemicals and cor-
respondence to globally harmonized system (GHS) classi-
fication. J Toxicol Sci 2009;34:611–26.
25. Vichai V, Kirtikara K. Sulforhodamine B colorimetric
assay for cytotoxicity screening. Nat Protoc 2006;1:
1112–6.
26. Reichl S. Cell culture models of the human cornea—a
comparative evaluation of their usefulness to determine
ocular drug absorption in-vitro. J Pharm Pharmacol
2008;60:299–307.
27. Yamaguchi H, Kojima H, Takezawa T. Vitrigel-eye irri-
tancy test method using HCE-T cells. Toxicol Sci
2013;135:347–55.
28. Spielmann H. Ocular irritation. In: Gomez-Lechon C
(ed). In Vitro Methods in Pharmaceutical Research.
London: Academic Press, 1997, 265–87.
29. Grover P, Jamil K. Alternatives methods for dermal and
ocular animal safety testing of chemicals. PINAS 2001;
B67:1–20.
30. Balls M, Botham P, Bruner L, et al. The EC/HO inter-
national validation study on alternatives to the Draize
eye irritation test. Toxicol In Vitro 1995;9:871–929.
31. Adriaens E, Bytheway H, De Wever B, et al. Successful
prevalidation of the slug mucosal irritation test to assess
the eye irritation potency of chemicals. Toxicol In Vitro
2008;22:1285–96.
32. Lenoir J, Claerhout I, Kestelyn P, et al. The slug mucosal
irritation (SMI) assay: development of a screening tool
for the evaluation of ocular discomfort caused by sham-
poos. Toxicol In Vitro 2011;25:1919–25.
33. Khutoryanskaya OV, Morrison PW, Seilkhanov SK,
et al. Hydrogen-bonded complexes and blends of poly
(acrylic acid) and methylcellulose: nanoparticles and
mucoadhesive films for ocular delivery of riboflavin.
Macromol Biosci 2014;14:225–34.
34. Dhondt MM, Adriaens E, Pinceel J, et al. Slug species- and
population-specific effects on the end points of the Slug
Mucosal Irritation test. Toxicol In Vitro 2006;20:448–57.
12 H. Abdelkader et al., 2015, Vol. 0, No. 0
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
35. Leighton J, Nassauer J, Tachao R. The chick embryo in
toxicology: an alternative to the rabbit eye. Food Chem
Toxicol 1985;23:293–8.
36. Auerbach R, Kubai L, Knighton D, et al. A simple pro-
cedure for the long-term cultivation of chicken embryos.
Dev Biol 1974;41:391–4.
37. Luepke NP. Hen’s egg chorioallantoic membrane test
for irritation potential. Food Chem Toxicol 1985;23:
287–91.
38. Abdelkader H, Wu Z, Al-Kassas R, et al. Conjunctival
and corneal tolerability assessment of ocular naltrexone
niosomes and their ingredients on the hen’s egg chorio-
allantoic membrane and excised bovine cornea models.
Int J Pharm 2012;432:1–10.
39. Alany RG, Rades T, Nicoll J, et al. W/O microemul-
sions for ocular delivery: Evaluation of ocular irritation
and precorneal retention. J Control Rel 2006;111:
145–152.
40. Bagley DM, Rizvi PY, Kong BM, et al. Evaluation of the
vascular components of the chorioallantoic membrane
assay as a model for eye irritation potential: II. Cutan
Ocul Toxicol 1991;10:105–13.
41. Budai P, Lehel J, Tavaszi J, et al. HET-CAM test for
determining the possible eye irritancy of pesticides. Acta
Vet Hung 2010;58:369–77.
42. Ilva DR, Colin RG, Alany RG. Comparison of ion-
activated in situ gelling systems for ocular drug delivery.
Part 2: Precorneal retention and in vivo pharmacody-
namic study. Int J Pharm 2011;411:78–85.
43. Abdelkader H, Pierscionek B, Alany RG. Novel in
situ gelling ocular films for the opioid growth
factor-receptor antagonist-naltrexone hydrochloride:
fabrication, mechanical properties, mucoadhesion,
tolerability and stability studies.IntJPharm
2014;477:631–42.
44. Muir CK. Opacity of bovine cornea in vitro induced by
surfactants and industrial chemicals compared with
ocular irritancy in vivo. Toxicol Lett 1985;24:157–62.
45. Burton AB, York M, Lawrence RS. The in vitro assess-
ment of severe eye irritants. Food Cosmet Toxicol
1981;19:471–80.
46. Scaife MC. An in vitro cytotoxicity test to predict the
ocular irritation potential of detergents and detergent
products. Food Chem Toxicol 1985;23:253–8.
47. Prinsen MK, Koëter HB. Justification of the enucleated
eye test with eyes of slaughterhouse animals as an alter-
native to the Draize eye irritation test with rabbits. Food
Chem Toxicol 1993;31:69–76.
48. Gautheron P, Dukic M, Alix D, et al. Bovine corneal
opacity and permeability test: an in vitro assay of ocular
irritancy. Fundam Appl Toxicol 1992;18:442–9.
49. Bruner LH, Kain DJ, Roberts DA, et al. Evaluation of
seven in vitro alternatives for ocular safety testing.
Fundam Appl Toxicol 1991;17:136–49.
50. Nishida T, Cornea. In: Krachmer JH, et al (eds). Cornea:
Fundamentals, Diagnosis and Management. Philadel-
phia: Elsevier, 2005;3–21.
51. Weterings PJ, Vanerp YH. Validation of the Becam
assay: an eye irritancy screening test. In: Goldberg AM
(ed). Alternative Methods in Toxicology. New York:
Mary Ann Liebert, Inc, 1987, 515–21.
52. Curren DR, Evans MG, Raabe HA, et al. Correlation of
histopathology, opacity and permeability of bovine
corneas exposed in vitro to known ocular irritants.Vet
Pathol 2000;37:557.
53. Curren RD, Harbell JW. In vitro alternatives for ocular
irritation. Environ Health Perspect 1998;106:485–92.
54. Harbell JW, Raabe HA, Evans MG, et al. Histopathology
associated with opacity and permeability changes in
bovine corneas in vitro.TheToxicologist1999;48:336–7.
Critical appraisal of alternative irritation models, 2015, Vol. 0, No. 0 13
by guest on February 17, 2015http://bmb.oxfordjournals.org/Downloaded from
