Content uploaded by Nigel Greene
Author content
All content in this area was uploaded by Nigel Greene on Sep 23, 2014
Content may be subject to copyright.
Cutaneous and Ocular Toxicology, 2009; 28(1): 1–18
REVIEW ARTICLE
A current practice for predicting ocular toxicity of
systemically delivered drugs
Chris J. Somps1, Nigel Greene1, James A. Render1, Michael D. Aleo1, Jay H. Fortner2,
James A. Dykens3, and Gareth Phillips3
1Drug Safety Research & Development, Pzer Global R & D, Groton, CT, USA, 2World Wide Comparative Medicine,
Pzer Global R & D, Groton, CT, USA, and 3Drug Safety Research & Development, Pzer Global R & D,
Sandwich, Kent, UK
Address for Correspondence: Chris J. Somps, PhD, Drug Safety Research & Development, Groton, MS 8274-1328, PGRD, Eastern Point Road, Groton, CT
06340. Tel.: 860-715-2841; Fax: 860-715-3577; E-mail: christopher.j.somps@pzer.com
(Received 1 August 2008; revised 23 October 2008; accepted 27 October 2008)
Introduction
Adverse drug-induced ocular side eects of systemi-
cally administered drugs are an important safety con-
cern facing patients, clinicians, and pharmaceutical
companies (1–3). Depending on severity, ocular side
eects can result in injury to the patient or discon-
tinuation of therapy (4), an increased risk of liability
for the clinician (5), and signicant nancial impact
to the drug developer if the drug label restricts market
availability, or if the drug is removed entirely from the
market (4). During the development of a new medi-
cine, pharmaceutical companies use animal models
to identify, characterize, and manage risk of adverse
ocular side eects in patients (6). However, the identi-
cation of an ocular safety concern during traditional
animal toxicity testing can occur after millions of
dollars and many years of research and development
have been spent on a particular drug development
program (7). Moreover, the welfare of animals used in
traditional animal toxicity testing is a concern of the
public (8). Within the pharmaceutical industry today
there is increased emphasis on reducing the amount
of traditional animal testing and improving the ability
to identify and avoid safety risks earlier in the devel-
opment of a new medicine (9). is will also avoid
costly program delays or discontinuations. In order
to achieve this improvement within our company, we
are deploying new in silico, in vitro, and in vivo tools
and approaches for identifying and characterizing
ocular toxicity risks, and are exploring ways of more
eectively applying traditional ocular safety assess-
ment tools and techniques.
For new medicines not intended for treating ocular
disease, the ideal drug is one that does not penetrate
the eye, in other words, one that has only local expo-
sure or is incapable of crossing blood–ocular barri-
ers. Currently, our understanding of a compound’s
characteristics that determine its ability to cross the
blood–ocular barriers is relatively crude and typically
ISSN 1556-9527 print/ISSN 1556-9535 online © 2009 Informa UK Ltd
DOI: 10.1080/15569520802618585
Abstract
The ability to predict ocular side eects of systemically delivered drugs is an important issue for phar-
maceutical companies. Although animal models involving standard clinical ophthalmic examinations and
postmortem microscopic examinations of eyes are still used to identify ocular issues, these methods are
being supplemented with additional in silico, in vitro, and in vivo techniques to identify potential safety
issues and assess risk. The addition of these tests to a development plan for a potential new drug provides
the opportunity to save time and money by detecting ocular issues earlier in the program. This review
summarizes a current practice for minimizing the potential for systemically administered, new medicines
to cause adverse eects in the eye.
Keywords: Eye; toxicity; drug development; drug safety; ocular toxicity
http://www.informapharmascience.com/cot
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
2 C. J. somps et al.
focused on methods for enhancing, rather than
minimizing, ocular exposure (10,11). In vitro mod-
els designed to predict in vivo ocular exposure need
further development and characterization (11). Until
the factors that determine ocular pharmacokinetics
are better understood and applied, pharmaceutical
companies will need to strategically employ more
predictive and more advanced investigative tools to
facilitate early identication, characterization, and
management of ocular safety risk of new systemic
medicines.
Early prediction of ocular safety risk begins with
the selection of the biological target and modulat-
ing agent for a particular disease. Target expression
and physiology in ocular tissues, under normal and
diseased conditions, are important factors to con-
sider when assessing the potential for ocular toxic-
ity. Obviously, if the molecular entity being targeted
by a drug is not expressed in the eye, adverse ocular
eects due to primary ocular pharmacology are not
possible. However, the prediction of ocular safety is
complicated by the potential for o-target, or unin-
tended, eects. O-target eects of a compound or
one of its metabolites are derived from its physical
and chemical properties or from its molecular inter-
action at secondary, often closely related, targets.
Knowledge of a compound and the biology and sig-
naling pathways of the intended target are critically
important to early and prospective identication of
ocular safety risk. e current and future application
of more predictive in silico tools, e.g., data mining,
structure–activity relations, and target and path-
way tools, will enhance the ability to identify ocular
toxicity early in drug development. e early use of
in silico tools will also guide the design and timing
of subsequent, more denitive in vitro and in vivo
studies.
In vitro assay systems can be used as prospective
tools to predict ocular safety risk early before the risk
has been identied in vivo. In vitro assay systems can
also be used retrospectively to gain a mechanistic
understanding of ocular safety ndings in an animal
model (9,12,13). In the case of prospective applica-
tions, in vitro tools are used to develop more informa-
tion about concerns derived from in silico investiga-
tions. For example, new gene-silencing techniques,
such as ribonucleic acid (RNA) interference (14), ena-
ble early evaluation of the eects of target modulation
in ocular tissues well before selective chemical tools
are available. In the case of retrospective applications,
in vitro tools may be used to understand the mecha-
nism of an ocular nding in an animal model or to
screen out subsequent candidate compounds that
have an increased probability of causing this nding
in vivo. Understanding the mechanism of an ocular
safety risk helps to design a safe testing program for
clinical patients.
Laboratory animals are required to denitively
identify, characterize, and manage ocular safety
risks (6,15). Ocular ndings in animals are typically
detected using routine clinical (e.g., direct or indirect
ophthalmoscopy) or histopathological examinations.
Recently, more advanced tools have been introduced
that enable improved characterization of ndings. For
example, new imaging modalities (e.g., Scheimpug
imaging, optical coherence tomography [OCT], and
laser Doppler owmetry) enable earlier detection
of drug-induced ocular eects, as well as providing
a means to quantify and track their progression or
regression. Along with standard tools such as elec-
troretinography, researchers can use these methods
to determine biomarkers of ocular side eects that
will facilitate safe use of test compounds in clini-
cal trials. For example, functional changes in visual
performance might precede structural eects, thus
allowing clinicians to safely monitor for ocular side
eects. However, depending on the dose or exposure
at which an ocular nding is detected and its severity,
such ndings may preclude further development of a
new medicine.
is review summarizes our current practice for
minimizing the potential for new systemically admin-
istered medicines to cause adverse eects in the eye.
e tools and techniques described here are not
meant to be all inclusive, but represent examples that
are currently being developed or used by the authors
today to minimize the chances of late-stage delay or
failure of new medicines because of adverse ocular
eects. Examples of predictive and investigative in
silico, in vitro, and in vivo tools will be described,
and the use of these tools to identify, character-
ize, and manage potential ocular side eects will be
discussed.
In all cases where animals are studied, experimen-
tal procedures conform to the National Institutes of
Health (NIH) Guide for the Care and Use of Laboratory
Animals and the guidelines of the Pzer Institutional
Animal Care and Use Committee. Eorts are always
made to minimize animal suering and to reduce the
number of animals used.
In silico prediction tools and strategies
In silico prediction tools and strategies include those
for predicting both target and chemotype-related
safety. Computational methods for mining and
understanding the impact of modulating the activity
of proteins and enzymes have been on the increase
for many years. ese include methods ranging from
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 3
simple literature-based inferences (16) to sophis-
ticated mathematical pathway and disease-centric
models (17,18). Understanding and predicting the
impact of target modulation can lead to better selec-
tion of targets or prioritization for monitoring and
testing in the drug development process.
Another factor that can lead to better (rational)
drug design is an understanding, or ability to predict,
the potential eects of a small molecule just from its
chemical structure (19). Novel compounds or chemi-
cal series can be prioritized even before they are syn-
thesized in the laboratory. However, the utility and
acceptance of these approaches will be limited by the
accuracy with which they can predict the occurrence
of the adverse eect. Many challenges have to be
overcome before one can accurately predict the toxic-
ity of novel chemical structures. Fundamental issues
such as the complex nature of biological systems and
the ability of chemicals to interact with a biological
system in multiple ways that could all potentially
manifest themselves as similar toxic events inevitably
mean that the problem of predicting toxicity is not
straightforward (20).
erefore, the development of computational tox-
icity prediction tools, that are either structure-based
or use computational modeling techniques on human
data, are the main approaches to be able to weed out
potentially toxic eects in humans even before hav-
ing the compound physically in hand. However, these
models need to have a good correlation with the in
vivo data, i.e., high sensitivity, as well as high spe-
cicity to avoid eliminating good compounds from
development. Ideally the structure-based in silico
model should be easy to use and easy to interpret as
these are key requirements for its usefulness.
Considering the primary pharmacology of a drug
Historically, most computational approaches for
understanding the safety eects of modulating a
specic protein or enzyme have been based upon
methods for mining the scientic literature to nd
relevant connections between a protein target and
its biological function (21). ese have ranged from
simple co-occurrence of terms in the title or abstract
of a paper to more sophisticated natural language
processing techniques that identify relationships
between the desired terms. In some cases, the use
of inference engines to predict what the eect of
modulating one protein will have on another protein
within a protein–protein interaction network has
been extrapolated to predict eects that may occur in
a complete biological system (16,22). In most cases,
these types of computational systems simply serve
to present potential hypotheses around the linkage
between protein modulation and some observed
biological eect that need to be assessed using in
vitro or in vivo experimentation.
As an example, the dierent families of cyclic
nucleotide phosphodiesterases (PDEs) are being
pursued as potential drug targets because of their
tissue-specic distribution, their critical roles in a
variety of intracellular signaling pathways and cellu-
lar functions, and their inherent drug-ability (23,24).
However, the PDE6 isoform has not been pursued
as a therapeutic target, largely because of its ocular
localization and vision-related pharmacology. PDE6
gamma is the primary regulator of cytoplasmic cyclic
guanosine monophosphate (cGMP) concentration
in rod and cone photoreceptor cells. It is activated
in response to light stimulation and hydrolyzes
intracellular cGMP. is results in the closure of
cGMP-gated ion channels and hyperpolarization of
the photoreceptor cell membrane (25,26). Precise
regulation of intracellular cGMP level is essential for
normal operation of the visual transduction cascade.
Persistent elevation of retinal cGMP can lead to rod
photoreceptor vesiculation, degeneration, and even-
tual cell death (27).
Close-in analogs, or isoforms, of the target also
need to be considered in early ocular safety evalua-
tion since separation of the activity of a compound
for 2 (or more) closely related targets will likely be dif-
cult. Of all the phosphodiesterases, PDE5 and PDE6
are most closely related (28,29), meaning that PDE5
inhibitors used therapeutically, such as sildenal, can
also exhibit visual side eects through a mild inhibi-
tory action on PDE6 (30).
Given the knowledge that modulation of a target’s
role or function may be directly linked to causing
an adverse eect, it is probably desirable to conduct
early screening of any compounds that are designed
to interact with it or, for that matter, designed to mod-
ulate the activity of a protein that is a close-in analog,
as in the above example.
Considering the chemical structure of a drug
Structure-based in silico prediction methods widely
used in the pharmaceutical industry can be roughly
classied into so-called “expert systems” and “data-
driven systems” (31). Expert systems try to formalize
knowledge from human experts who have assessed
the toxicity of compounds, whereas data-driven sys-
tems use computational algorithms to derive predic-
tive models from experimental data. Expert systems
are intuitively more appealing because they promise
easy access to toxicological knowledge derived from
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
4 C. J. somps et al.
human interpretation of experimental data. Some of
these programs provide reasonably high specicity of
about 80%, but they can suer from moderate sensi-
tivity, which can be in the region of 50% (32).
In contrast, data-driven systems are most com-
monly used to make predictions for compounds with
similar structures that usually manifest the toxicolog-
ical eect through the same mechanism. A number
of well-known techniques are used to establish pre-
dictive models such as partial-least squares (PLS),
recursive partitioning, support vector machines, or
neuronal nets (33). Beside the correct prediction, an
important task for the data-driven systems is the iden-
tication of chemical features that are relevant for the
observed toxicological eect. ese chemical features
may help the medicinal chemist to design a chemical
structure that has little potential for developing the
toxic eect. So far, most of the eorts in developing
computational models have been focused on pre-
dicting the risk of genetic toxicity. However, future
activities need to be directed at the prediction of more
complex safety endpoints, such as ocular toxicity.
Currently only one commercial system has tried to
predict these more complex endpoints. is system is
the Derek for Windows (DfW) application developed
and distributed by Lhasa Limited (Leeds, UK; http://
www.lhasalimited.org). DfW is a knowledge-based
expert system designed to predict the toxicity of a
chemical from its structure. Its use and application
have been described elsewhere (20,34–36).
In recent years structural alerts have been added to
the DfW knowledge base that are specically focused
on nonirritant ocular toxicity. ese alerts were devel-
oped using the knowledge base development cycle
pictured in Figure 1. e initial phase of develop-
ment involved researching public-domain literature
and books for compounds demonstrated to cause
ocular toxicity (4). e purpose was to collate and
use them to establish structure–activity relationships
(SARs) relating the biological endpoint to a chemical
substructure or chemical feature. e resulting SARs
were then implemented in the DfW knowledge base
expert system.
To date, only 14 structural alerts have been identi-
ed for ocular toxicity and have been added to the
DfW knowledge base in version 10 of the software.
e chemical classes that are found to be associated
with ocular side eects involve a range of mecha-
nisms and target sites within the eye. Examples
include 4-aminoquinolines, benzodiazepines, and
phenothiazines. 4-Aminoquinolines may cause
reversible eects when salts of the parent compound
are deposited in the cornea, but cause irreversible
photoreceptor degeneration by accumulation in the
retinal pigment epithelium (37,38). A metabolite of
benzodiazepines is the apparent cause of allergic
reactions in the eyelids or conjunctivae, blurred
vision, an ocular burning sensation, and excess
tear production (4). Typical examples of phenothi-
azines include chlorpromazine, uphenazine, and
N
N
N
N
N
N
O
N
SS
S
S
F
F
Cl
Chlorpromazine Fluphenazine Thioridazine
Figure 2. Phenothiazine compounds associated with ocular toxicity.
Figure 1. Derek for Windows (DfW) knowledge base development cycle.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 5
thioridazine (Figure 2). Chlorpromazine is reported
to cause cataract formation and to induce corneal
and lenticular pigmentary changes (39,40). e
mechanism of cataract formation may involve the
suppression of calcium eux, leading to lens protein
aggregation (41). Fluphenazine and thioridazine
are documented with pigmentary retinopathy and
maculopathy, respectively, and other phenothi-
azines may cause retinopathy, myopia, and dry eye
(41,42).
An illustration of DfW prediction of phenothi-
azine ocular toxicity is outlined in Figure 3. As
highlighted in the description of the alerts, the
scope of the alert is based on the chemical features
associated with this pharmacological class of
compounds.
e presence of a structural alert for ocular tox-
icity in a novel compound simply serves to raise
awareness for the potential of that compound to
cause toxicity. Many other factors such as metabolic
activation, tissue distribution, and clearance of the
compound may result in the absence of any observ-
able eect in vivo, so there is a clear need to conduct
early in vitro or in vivo experiments to conrm the
expression of this undesirable eect. It should also
be noted that although a single structural alert will
potentially cover many hundreds of novel chemi-
cals and 14 alerts may equate to many thousands of
structures, this is still a relatively low number com-
pared with the number of possible compounds that
could be synthesized. erefore, the absence of a
structural alert for ocular toxicity in a compound by
no means implies that the compound will not cause
toxicity in vivo.
In conclusion, in silico systems for predicting ocular
toxicity are very much in their infancy, and currently
there are few publicly available systems that could be
implemented as part of a screening paradigm. Much
work remains to be done to understand the mecha-
nisms by which compounds can cause ocular toxicity
before truly eective and predictive in silico screens
can be developed.
In vitro prediction tools and strategies
In vitro assay systems are used fairly early in the drug
development process to identify, screen, or char-
acterize potential for ocular toxicity risk. ese may
include in vitro assays using surrogate cell systems
for prediction of ocular toxicity risk. Typically, these
assay systems employ cells and tissues from dier-
ent regions of the eye (e.g., cornea, lens, or retina)
and include methods for assessing either molecular
target or chemotype safety. In the following sections,
we present examples of each of several types of assay
systems and how they can be used to predict and
characterize ocular toxicity risk. e assays presented
in this review are primarily used for internal deci-
sion making since only the 3T3 neutral red uptake
(3T3-NRU) assay has been accepted by regulatory
agencies. e challenges associated with validating
and gaining regulatory adoption have been reviewed
by others (43). Other assays that have gained some
degree of regulatory acceptance are the bovine cor-
neal opacity and permeability test method and the
isolated chicken eye test method. Both of these tests
are used in a tiered testing strategy to classify chemi-
cals as ocular corrosives or severe irritants without
further testing in animals, but are not routinely used
for assessing ocular safety risk of orally delivered
pharmaceuticals.
Figure 3. An illustration of Derek for Windows (DfW) prediction of phenothiazine ocular toxicity.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
6 C. J. somps et al.
Phototoxicity evaluation: 3T3-NRU and reactive
oxygen assays
Compounds that absorb radiation in the wavelength
range of 290–700 nm, and are present in sun-exposed
tissues such as the eye following topical or systemic
exposure, may have the potential to induce photo-
toxicity. e 3T3-NRU assay was developed and vali-
dated under the auspices of the European Centre for
the Validation of Alternative Methods (ECVAM), and
is now accepted by the European Union commission
and member states as an initial assay for all com-
pounds showing absorbance of ultraviolet (UV) and
visible light (44).
e 3T3-NRU test is based on the relative reduction
in viability of Balb/c 3T3 broblasts (clone 31) exposed
to a test chemical in the presence (+Irr) or absence
(–Irr) of ultraviolet A (UVA) light (315–400 nm) for
approximately 15 minutes at a dose of 5 J/cm2. Dye
uptake, an index of viability, is assessed 24 hours
later. e assay is an adaptation of the procedures
described by several groups (45– 47), and determines
viability by assessment of neutral red dye uptake in
balb/c 3T3 cultures exposed to the serial dilutions of
a test article compared with neutral red dye uptake in
control (compound vehicle). Additionally because of
the absence or very small amount of UVB light expo-
sure during the conduct of the 3T3 assay, related to
the cytotoxicity of UVB light to the broblast cells, this
assay is not appropriate for mainly UVB-absorbing
compounds.
e concentration of test chemical reducing cell
viability to 50% (IC50), a photo-irritancy factor (PIF),
and a mean photo eect (MPE) are determined. e
PIF is the ratio of IC50(–Irr) to IC50(+Irr). e MPE is
based on a comparison of the (+Irr) and (–Irr) dose–
response curves. A PIF < 2 or an MPE < 0.1 predicts “no
phototoxicity”; a PIF > 2 and < 5 or an MPE > 0.1 and
< 0.15 predicts “probable phototoxicity”; and a PIF > 5
or an MPE > 0.15 predicts “phototoxicity” (48).
An as-yet-unpublished survey of pharmaceu-
tical companies by the European Federation of
Pharmaceutical Industries and Associations has
revealed that an extremely high percentage of com-
pounds (> 85%) that are reported as positive in the
3T3-NRU assay are negative when tested using in vivo
animal models. Potential problems with such false-
positive ndings have been attributed to responses
at high concentrations, and the guidelines cautioned
interpretive restraint at doses > 100 g/ml (49).
Despite issues of false positives, it should be noted
that this assay predicts negatives with excellent del-
ity; no false-negative cases have been reported. is
reects the eorts made to validate the assay, and is
probably the strongest point recommending its utility.
However, given the high proportion of potential drug
candidates that absorb in the UVA region, together
with a 50% 3T3 positive hit rate for UVA-absorbing
compounds, the burden of in vivo photosafety test-
ing relative to the clinical impact of phototoxicity is
disproportionate.
To address this, phototoxicity assays that focus on
in vitro photochemistry independent of biology have
been developed. is course is reasonable, given that
photodynamic (i.e., requiring molecular oxygen as a
reactant) production of superoxide radical and sin-
glet oxygen are the proximate mediators of many, but
not all, phototoxic responses, especially Norris type
I and II reactions (50,51). To assess photodynamic
superoxide production, a tetrazolium dye can be
used, which is directly reduced by the radical to the
insoluble formazan, which can be quantied using a
spectrophotometer (52). is could be standardized
using potassium superoxide, and the eciency of the
photoreaction calculated. One advantage of this pro-
tocol is that it is readily scaled up via microtiter plates
to increase throughput if needed. Singlet oxygen pro-
duction is monitored using imidazole as a selective
acceptor that yields a trans-annular peroxide inter-
mediate, which in turn bleaches p-nitrosodimeth-
ylaniline (RNO). Monitoring RNO bleaching is also
accomplished spectrophotometrically (52).
Using these assays, Onoue et al. (53) performed a
head-to-head comparison with the 3T3 assay using
39 marketed drugs as range-setting compounds, and
210 drug candidates containing 11 chemical classes.
ey found that most of the phototoxic compounds
photodynamically generated reactive oxygen species
(ROS), and there was excellent concordance between
the in vitro and in vivo models. Moreover, the 39
model compounds established thresholds for ROS-
mediated phototoxicity.
e 3T3 assay was developed to detect photoirri-
tant potential of a drug, and although it can predict
some photogenotoxic potential, it was not designed
to report the latter or photoallergic potential, both
of which occur via direct mechanisms, but also via
ROS-mediated pathology (53,54). In this way, the ROS
assay could more faithfully predict irritant, allergic,
and genotoxic responses. But it bears reiteration that
not all phototoxicity is photodynamic, so it is likely
that some combination of in vitro and in vivo assess-
ments will ultimately provide preclinical predictions
in better accord with clinical reality.
Isolated rat lens
e isolated rat lens is often used as an in vitro model
to rank order compounds for cataract risk and to study
mechanisms of cataractogenesis (55). e benets of
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 7
using the isolated lens when compared with whole
animal models include the ability to control the extra-
cellular milieu bathing the lens (e.g., drug or other
treatment conditions) and the ability to quantitatively
measure biochemical or molecular endpoints asso-
ciated with mechanisms of cataract formation (e.g.,
cholesterol biosynthesis, intracellular ion content
or transport, adenosine triphosphate [ATP] content,
glutathione reduction–oxidation [redox] status, and
calpain activation). For studying mechanisms of cata-
ract formation caused by chemicals or pharmaceutical
agents, the ability to directly expose isolated lenses to
the insulting agent signicantly reduces the time and
amount of drug needed to conduct screening of poten-
tial drug candidates for cataractogenic potential. ese
factors are critically important early in the drug devel-
opment process when quantities of drug available for
testing are typically quite small or when the amount of
time needed to produce the cataract in vivo is lengthy.
Finally, the isolated lens model enables safety assess-
ments of both intended disease targets and of unin-
tended or o-target chemotype-related eects.
Although the lens represents a large portion of
the rodent eye, it is fairly easily harvested from the
rat eye and used as an explant culture (55). However,
the development of this technique requires repeated
practice to be able to routinely harvest lenses that
are not damaged as a result of the extraction proce-
dure. Briey, following euthanasia of the rat, eyes are
removed and lenses are recovered using a posterior
approach. Extreme care should be taken in handling
the lens in order to avoid producing a mechanically
induced cataract. Cultured lenses that remain clear
for ~ 24 hours can be used to evaluate the eects of
candidate drugs on lens clarity. Lenses are typically
cultured in Medium 199 (Sigma-Aldrich, St. Louis,
MO, USA) with various supplements, and can remain
clear for periods up to 10 days. e lens explant culture
model may be used to rank order and select alterna-
tive drug candidates based on their potential to cause
opacity formation in vitro (55,56) and to investigate
mechanisms of cataract formation with various drug
candidates, such as S-(1,2-dichlorovinyl)-L-cysteine
(DCVC) (57) and ciglitazone (56).
Isolated rat lenses and candidate
drug evaluations
Lenses are treated with drugs for varying amounts of
time (hours to days), and lens clarity is used as the
phenotypic endpoint. For example, following 5 days
of exposure to 10 mM tetraethylammonium (TEA),
a broad-spectrum potassium channel blocker, a
decrease in lens clarity is clearly evident when com-
pared with concurrent control lenses (Figure 4a).
Furthermore, lens clarity can be quantied using
software (SigmaScan, Systat Software, Inc, San Jose,
CA, USA) that measures pixel intensity (Figure 4b).
is assay system can thus be used by drug designers
to identify compound structures that do not produce
a loss of lens clarity and, therefore, drugs that are
less likely to cause lens toxicity in animal models or
patients.
Isolated rat lenses have been used several times
to investigate the cataractogenic potential of 3 thia-
zolidinediones (ciglitazone, darglitazone, and eng-
litazone), DCVC, and 3 tetrahydropyrans (CJ−12,918,
CJ−13,454, and ZD2138) (55–57). Although the model
is extremely versatile and manipulable, a major limi-
tation is the increased likelihood of identifying false
positives. is is because the in vitro model allows
Figure 4. Eects of a broad-spectrum potassium channel blocker, tetraethylammonium ion (TEA), on clarity of isolated rat lenses. A)
Daily images of cultured lenses under control (left panel) and treated (10 mM TEA, right panel) conditions demonstrate a treatment-
related cloudy lens phenotype. B) Lens clarity can be quantied using software that measures pixel intensity across lens images quantify-
ing progression of lens changes.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
8 C. J. somps et al.
direct access of compounds to the lens, even though
these compounds may have limited ability to gain
access to the inner chambers of the eye and penetrate
the lens in vivo (55).
Isolated rat lenses and target evaluation
RNA interference (RNAi) is a post-transcriptional
gene silencing technology that was discovered in the
last decade (14). Within the pharmaceutical industry,
this technology is nding wide application as a potent,
selective and easy-to-use laboratory tool for exploring
the biological role(s) of potential drug targets (58). By
introducing short (19–22 bp), double-stranded RNA
(dsRNA) into the cytoplasm of cells, the expression
of specic genes can be selectively inhibited. is is
achieved by the enzymatic destruction of messenger
RNAs (mRNAs) with nucleotide sequences comple-
mentary to the short dsRNAs. us, the RNAi mecha-
nism can be used to determine the functional con-
sequences of down-regulating a specic protein. An
important advantage for target safety assessments is
that RNAi can be applied long before any good-quality
chemical matter has been developed against a particu-
lar target. Pharmaceutical researchers can thus get an
early indication of the potential risks associated with
drugging a target that may be expressed in the eye.
As an example, this technique was used to explore
the potential for cataract formation associated with the
aldehyde dehydrogense (ALDH) isoform ALDH1A1.
RNAi was used to selectively down-regulate ALDH1A1
in lens tissues in order to assess the safety risks associ-
ated with pharmacological inhibition of this enzyme
in the lens. In these studies, ecacious small interfer-
ing RNAs (siRNAs) were identied and characterized
rst in lens epithelial cell models and then applied
to isolated rat lenses. Ecacy of gene knockdown
in cell models was conrmed using mRNA (reverse
transcriptase polymerase chain reaction [RT-PCR]),
protein (immunohistochemistry), and functional (cell
viability) endpoints. Interfering RNAs were identifed
that reduced ALDH1A1 transcript in cell models by
> 70% and signicantly decreased the cell’s ability to
detoxify 4-hydroxy-2-nonenal (4-HNE), a reactive
aldehyde by-product of lipid peroxidation known to
accumulate in the lens under conditions of oxidative
stress (Figure 5). Once an eective siRNA was identi-
ed in cell-based assays, it was further tested in the
isolated rat lens culture system. Notably, in these
studies knockdown of ALDH1A1 in isolated rat lens
produced a cloudy lens phenotype when compared
with untreated lenses or lenses treated with a negative
siRNA (Figure 6). Such studies corroborate literature
reports and raise concerns for drug development pro-
grams targeting ALDH1A1 (59,60). ey also drive the
timing and ocular assessments in subsequent in vivo
study designs with early chemical matter in order to
more fully evaluate lens toxicity risk.
Ex vivo lens cholesterol biosynthesis
Inhibition of cholesterol biosynthesis has long been
recognized as a major potential mechanism of cata-
ract formation in humans and dogs, through clinical
administration of triparanol (61), and administration
of statins in dogs, respectively (62–65). A simple tech-
nique to assess the impact of drugs on lens cholesterol
synthesis has been developed (66). Rats are treated
with compound for 1 to several days. After treatment
5 lenses/treatment are pooled in 1 ml of bicarbo-
nate-supplemented Medium 199 (298 ± 2 mOsm).
Cholesterol biosynthesis and palmitate biosynthesis
are determined by the incorporation of [14C]-acetate
as previously described (66,67). Cholesterol biosyn-
thesis and palmitate biosynthesis are corrected for
recovery of [3H] internal standards and standardized
per milligram of lens protein. In the example pre-
sented here, only U18666A reduced lens cholesterol
biosynthesis after 2 days of in vivo exposure (Figure 7).
Figure 5. Eect of ribonucleic acid interference (RNAi)-
mediated knockdown of aldehyde dehydrogense (ALDH) isoform
ALDH1A1 in rat primary lens epithelial cells. A) ALDH1A1 mes-
senger RNA (mRNA) levels determined by quantitative reverse
transcriptase polymerase chain reaction (qRT-PCR) are selec-
tively reduced compared with levels in nontransfected control
cells and cells transfected with a nontargeting (negative) small
interfering RNA (siRNA). e control housekeeping gene 18S is
not aected. B) Down-regulation of the ALDH1A1 isoform in rat
lens epithelial cells reduces their capacity to detoxify the reac-
tive aldehyde 4-hydroxy-2-nonenal (4-HNE), which is reected
in the leftward shift of the cell viability curves. 4-HNE is known to
accumulate in the eye under conditions of oxidative stress.
ALDH1A1 mRNA
Cell Viability
ALDH1A1
18S
120
100
% Control
80
60
40
20
0
120
140
100
80
60
40
20
0
0246
HNE (µM)
81012
A
% Control
B
Non
transfected
Negative
siRNA
ALDH1A1
siRNA
Non transfected
Negative siRNA
ALDH1A1 siRNA
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 9
Other agents known to cause cataracts in rats through
dierent mechanisms, such as naphthalene, which
causes oxidative stress through reactive metabolite
formation, and galactose, which causes oxidative
stress through polyol accumulation, had no eect
on cholesterol biosynthesis, thereby demonstrating
the specicity of the assay. is approach has been
applied as a way to investigate the ability of numerous
pharmaceutical agents to cause cataracts in rats, and
has been used for compounds such as thiazolidinedi-
ones, CJ-12,918, and RGH-6201 (68).
In vivo assessment tools and strategies
Ocular toxicity is often detected using routine clinical
or histopathological examinations of eyes of animals
and humans. Routine clinical ophthalmic examina-
tions include visualization of structures in the ante-
rior segment (e.g., cornea, conjunctiva, lens) and the
posterior segment (e.g., retina, optic disk). Recent
advances in imaging technology have enabled more
quantitative detection, characterization, and docu-
mentation of ocular changes (e.g., Scheimpug imag-
ing, OCT, and laser Doppler owmetry). e informa-
tion gained by use of these imaging tools is facilitating
better decisions on the risk of ocular toxicity of new
drug candidates. Routine histopathological exami-
nations include thorough examination of paran-
embedded sections cut from globes and involve the
major structures of the eye.
Routine preclinical ophthalmic examination
e routine ophthalmic examination is an integral
part of preclinical safety studies and allows for an
ecient and sensitive method of detecting subtle
changes in the viewable regions of the eye. e exam-
iner will often examine many animals in a single
study and needs a clear understanding of normal
ocular anatomy in order to detect ocular abnormali-
ties (69,70). All viewable ocular structures are exam-
ined because focusing on only one ocular segment
(e.g., fundus) increases the likelihood of overlook-
ing subtle changes in other ocular tissues. Although
using a single piece of equipment may be adequate
for detecting most ocular changes, several commonly
used pieces of equipment are available that will
enhance the detection of abnormalities. Typical types
of clinical ophthalmic examination include direct
ophthalmoscopy, indirect ophthalmoscopy, and slit
lamp biomicroscopy. Regardless of the instruments
used, the examiner must be trained and qualied in
their use.
A common instrument encountered in preclinical
ophthalmic examinations is the direct ophthalmo-
scope (71,72). is inexpensive and rugged instru-
ment is relatively easy to learn to use. Although it
allows for examination of the anterior and posterior
segments, it is often used as an adjunct to the indirect
ophthalmoscope. e direct ophthalmoscope is not
ideally suited for use in safety studies because the high
magnication of this equipment results in a smaller
Figure 6. Eect of ribonucleic acid interference (RNAi)-mediated
knockdown of aldehyde dehydrogense (ALDH) isoform ALDH1A1
on clarity of isolated rats lenses. A) 3 rat lenses cultured in con-
trol media for 6 days. B) 3 rat lenses cultured in control media
for 5 days following a 24-hour exposure to media containing a
nontargeting (negative) small interfering RNA (siRNA). C) 3 rat
lenses cultured in control media for 5 days following exposure to
media containing a siRNA targeting ALDH1A1. On average the 3
lenses exposed to the ALDH1A1 siRNA are less clear after 6 total
days in culture.
Figure 7. Ex vivo lens cholesterol biosynthesis in rats treated with
U18666A (2 days), naphthalene (2 days), or galactose (3 days)
was determined through the incorporation of [14C]-acetate into
cholesterol. U18666A treatment signicantly decreased (59%)
lens cholesterol biosynthesis, while no eect was observed with
naphthalene or galactose. In contrast, treatment with galactose
signicantly increased (82%) palmitate biosynthesis, while no
eect was observed with U18666A or naphthalene (78).
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
10 C. J. somps et al.
eld of view and increased time and eort to scan the
fundus. Although structures of the anterior segment
can be examined, considerable skill is required to
attain the correct depth of focus for this multilayered
evaluation. Other disadvantages of direct ophthal-
moscopy include the following: 1) units used regu-
larly in a vivarium must be professionally cleaned on
a routine basis because of the accumulation of dust,
dander, and hair in the internal lens assembly; 2)
batteries need recharging; and 3) a training period is
needed to learn to focus using the focus wheel.
Probably the most versatile method used for pre-
clinical ophthalmic examinations is indirect ophthal-
moscopy (15,71–73). is method uses a wide variety
of hand-held lenses that allow for the examination
of eyes of various sizes and provides a comfortable,
panoramic view of the fundus. Typically, a 20D lens
is used for canine and nonhuman primate examina-
tions and a 28D lens is used for rodent examinations.
e examination begins with a quick scan across the
entire fundus attempting to detect abnormalities that
require a higher magnication inspection. Although
designed for funduscopic evaluation, the light source
and optics of the indirect headset allow a high level
assessment of the anterior segment as well, for exam-
ple, the cornea, anterior chamber, and lens, before
proceeding with the funduscopic examination.
Disadvantages of indirect ophthalmoscopy include
diculty learning the technique, discomfort of the
headset with prolonged use, lack of ready portability
of some models, and expense.
e hand-held slit lamp biomicroscope (SLBM) is
a logical choice for a second tool when performing
routine ophthalmic examinations (15,71,72). is
versatile item is ideal for evaluation of the anterior
segment (i.e., cornea, anterior chamber, iris, and
lens), especially during functional testing (e.g., pupil-
lary light reex and examination of the undilated iris).
In comparison with the larger, tabletop models, the
hand-held SLBM is easily portable, allowing for rapid
and detailed examinations wherever the animals are
located. In addition to limited slit widths and open
aperture settings, many models include a cobalt lter
for use with corneal uorescein staining for the detec-
tion of corneal ulcers. Slit lamps are relatively easy to
use because of their xed focal length, but have dis-
advantages that include a relative fragility of plastic
exteriors and prism alignment in the microscope, a
need for rechargeable battery replacement, a limited
depth of visualization due to the optics, and expense.
In addition to the types of instruments, the proce-
dures used for an ophthalmic examination may vary.
Typically in the pharmaceutical industry, the oph-
thalmic examination is designed to be ecient and
thorough and to assess the ocular status of numerous
study animals, while minimizing animal stress and
discomfort. e examiner should follow a standard
procedure for each species so that each examination
is complete. is standard approach can be amended
as needed for individual protocol concerns (71,72).
Despite claims to the contrary by equipment manu-
facturers, all complete ophthalmic examinations
must be conducted through a dilated pupil. is is
accomplished by use of a mydriatic drug, typically
tropicamide, and the procedure should be clearly
specied as part of a standard operating procedure.
Of the laboratory animal species used in the phar-
maceutical industry, testing visual function is rela-
tively easy in the dog. Visual function can be initially
determined by testing the menace and pupillary light
reexes prior to mydriatic installation. In addition,
this is a good opportunity to evaluate the iris in its
natural, undilated state. ese parameters should be
evaluated on all dogs and followed by the application
of a mydriatic drug. Once eyes are fully dilated, usu-
ally within 15 to 20 minutes, the traditional ophthal-
mic examination may be conducted. e choice of
the instruments used is at the examiner’s discretion;
however, use of both indirect ophthalmoscopy and
slit lamp biomicroscopy is recommended for dogs.
During the examination, each separate compo-
nent of the anterior segment (e.g., cornea, anterior
chamber, lens, nucleus, and anterior vitreous) and
the fundus should be carefully examined. e pres-
ence of the tapetum in the canine eye provides an
excellent opportunity for retroillumination of the
anterior segment. Larger, faint opacities (e.g., in the
corneal stroma or posterior capsule) are often more
easily observed with retroillumination. e examina-
tion includes all components of the lens (i.e., anterior
lens capsule and cortex, nucleus, and posterior lens
capsule and cortex) with special attention to lenticu-
lar sutures. With sharp focusing, the precise location
of small lenticular opacities can be determined. With
appropriate dilation, most of the fundus can be exam-
ined, but the examiner must be mindful of the level of
illumination used during an examination; although
dogs with maximally dilated pupils rarely show signs
of discomfort in ambient light, they may become
uncooperative if a bright light is directed into the eye.
e ophthalmic examination in rats is very similar
to that in dogs, with a few notable exceptions. e
typical rat used in safety studies is albino, a condition
fraught with peril for the eye. Even in the standard,
low-level illumination used in rodent housing rooms,
the pupils in albino rat eyes are maximally or near
maximally constricted. Moreover, menace reexes are
dicult to obtain in the rat. ese conditions combine
to render functional testing of the rat visual appara-
tus pointless in routine ophthalmic examinations.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 11
Consequently, the rst step is often instillation of
the mydriatic drug, an approach that eliminates the
opportunity to evaluate the undilated iris.
e ophthalmic examination in nonhuman pri-
mates requires sedation, typically with ketamine,
to perform a complete and safe examination. In the
sedated monkey, a pre-examination, including the
pupillary light reex and an inspection of the undi-
lated iris, is followed immediately by mydriatic instil-
lation. When adequate dilation is achieved, generally
after 15 to 20 minutes, the complete examination,
which is similar to that described for the dog, is per-
formed. e nonhuman primate eye is unique among
the laboratory animals because of the presence of a
macula and fovea in the fundus. Since the macula
and fovea are also present in the human fundus and
are responsible for central visual acuity, an examina-
tion of these structures is mandatory (71,72,74).
Nonroutine preclinical ophthalmic examination
Increasingly, nonroutine, specialized in vivo tech-
nologies are being used for improved ocular imaging
and safety assessments. A few of these are described
below.
Optical coherence tomography
Optical coherence tomography (OCT) is a noninvasive
imaging technique used commonly in clinical oph-
thalmology practices because of the 2-dimensional,
high-resolution, cross-sectional images obtained of
the ocular posterior segment (retina, choroid, optic
nerve) (75,76). is technique, which involves meas-
uring the amplitude of backscatter light returning
from the posterior eye as a function of delay, may
have application in preclinical research.
OCT imaging is analogous to ultrasound imaging,
except that light, rather than sound, is used (76). e
OCT images represent cross-sections of a tissue that
results from the backscattering of light from structures
at dierent depths within the tissue. As light enters
a tissue, it is transmitted, absorbed, or scattered.
Transparent structures, such as the cornea, lens, and
sensory retina, generally allow light to be transmitted
to deeper structures in the eye or posterior segment.
Absorbed light is removed from the incident beam
by light absorbers, such as hemoglobin and melanin.
Light scattering occurs when there is a variation in
the types of cells present. Light may scatter in multi-
ple directions, but a strong signal occurs when light is
backscattered. Backscattering of light is highest at the
border between 2 homogeneous materials with dier-
ent indices of refraction (e.g., vitreoretinal interface).
A high amount of backscatter occurs from the nerve
ber layer, the plexiform layers, the external limiting
membrane (boundary between the inner and outer
segments of the photoreceptors), and Bruch’s mem-
brane with the choriocapillaris. Signals become weak
in structures posterior to the choriocapillaris (i.e.,
choroid and sclera) because light is attenuated as it
passes through the choriocapillaris. Other layers of
weak signal because of low backscatter are the nuclear
layers of the retina (i.e., ganglion cell layer, inner
nuclear layer, and outer nuclear layer). As a result, the
image quality of the retina may be too poor to detect
early retinal degeneration, especially in rodents. e
quality may be improved by use of high-resolution,
full-eld OCT imaging, but this involves enucleation
of the eyes (77).
Scheimpug imaging
Scheimpug imaging allows for the noninvasive
assessment of structures in the anterior segment of the
eye. Its unique optics allow the investigator to acquire
sagittal images through the cornea, anterior chamber,
and anterior portions of the lens. It can also be used
in a retroillumination mode in which light is bounced
o the back of the eye and the lens is illuminated from
behind. e Nidek EAS-1000 anterior segment analy-
sis system can be used in combination with EAS-1000
image analysis software (Nidek Co. Ltd., Aichi, Japan)
to detect and quantify ocular lesions in animal mod-
els. In a comparison with a routine slit lamp appara-
tus it was determined that the routine slit lamp is just
as sensitive for detecting drug-induced cataract as is
the Scheimpug imaging apparatus (78,79), but that
the Nidek Scheimpug system enables more ways
to image, localize, and quantify lenticular lesions
(Figure 8).
Laser Doppler owmetry
Laser Doppler owmetry can be used to measure
blood ow in the optic nerve head (ONH) (80) in
response to compound administration. Oxygen
demand is high in the retina and ONH, and blood ow
to these tissues is governed by robust, local autoregu-
latory mechanisms (81). Toxic eects of compounds
can aect blood ow in the eye and induce retinal
damage. e Heidelberg retinal owmeter (HRF)
employs an infrared laser scan at 780 nm to detect
indirectly a Doppler shift caused by moving red blood
cells in the ONH. e HRF is designed for the human
eye and must be modied to allow focus within the
rodent eye. e data acquisition time is 2 seconds;
therefore, animal subjects require general anesthesia
for scanning.
Choice of anesthesia can have signicant impact on
this procedure. Inhalation of oxygen decreases retinal
blood ow (82); therefore, inhalant anesthetics should
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
12 C. J. somps et al.
be delivered in room air. Injectable anesthetics work
well if consistent depth can be achieved and ocular
movement halted; the authors have used ketamine
and xylazine successfully. Choice of animal model
is also critical. Albino eyes permit excessive internal
reection that interferes with the Doppler signal.
Pigmented animals work best. e ONH of rodents,
specically rats, is too small and too obscured by the
retinal vasculature to allow evaluation of this structure
alone. e author have utilizes data from the entire
scan image (256 × 64 pixels) for evaluation of rodent
retinal blood ow.
Using this technique, the author has success-
fully demonstrated decreased retinal blood ow in
response to pure oxygen breathing, as predicted by
human results (82). Further studies with a known
retinal toxicant also demonstrated decreased reti-
nal blood ow, as expected by decreased demand
(81). e technique could be applied to compounds
intended for increasing blood ow, especially in the
central nervous system.
Electroretinography
Electroretinography (ERG) is a noninvasive method
for assessing the overall health and functional status
of the retina. ere are a number of excellent reviews
describing the application of ERG in humans (83) and
various animal models (84). ERG has been used as a
tool for predicting and assessing drug-induced retinal
toxicity (85,86). e authors use full-eld, ash ERG
in preclinical animal models to assess retinal func-
tion in toxicity studies with candidate new medicines.
ese studies may be triggered by safety concerns
brought to light by in silico or in vitro studies, or initi-
ated to characterize functional correlates of ndings
observed with retinal histopathology. For example,
ERG may provide a leading biomarker of retinal eects
that precede structural changes, either temporally or
in terms of dose response, and might be used to safely
introduce a compound into human patients. Finally,
ERG may also be used to develop a preclinical model
of visual disturbance reported during clinical trials,
thus enabling investigative studies of the mechanism
of visual disturbance.
Additional nonroutine examinations
e foregoing discussion is not meant to be an
exhaustive review of ophthalmic procedures. Many
additional tests are available and may be incorpo-
rated into the ophthalmic examination to address
specic toxicity issues (87). Testing procedures for
the anterior segment are particularly abundant.
Some tests are as simple as applying a stain to the
cornea and examining with the SLBM. A variety of
ophthalmic stains may be used, including uorescein
for preocular tear lm assessment and detection of
corneal epithelial defects (88,89) or Rose Bengal stain
when trying to identify degenerating epithelial cells
as a result of inadequate precorneal tear lm (90,91).
Lacrimal gland toxicity may be determined by abnor-
mal tear production as estimated by use of Schirmer
tear test papers, especially in dogs (90). Additional
corneal evaluations include esthesiometry, which
tests the sensitivity of the cornea (92,93), pachymetry,
which measures the thickness of the cornea (94,95),
Figure 8. Scheimpug lens density analysis of changes observed
on days 7, 14, and 17 in a naphthalene-treated, pigmented rat.
Initial changes in clarity are visible by day 7 and are shown in the
upper right quadrant of each image (arrow in insets for clarity).
Note progression of the lesion as indicated in the densitometric
scan to the left of each image (see circled peak). Densities are also
quantied numerically at locations marked by an “x”. Marks over
the lesion show time dependent change in numeric values (79).
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 13
keratometry, which measures corneal curvature
(96,97), and specular microscopy for visualization of
the corneal endothelium (94,98). ese tests may be
performed after the cornea has undergone an exami-
nation with a slit lamp biomicroscope. In addition
to corneal changes that may occur with increased
intraocular pressure, changes in the ltration angle
may be evaluated through gonioscopy (99,100), and
intraocular pressure may be evaluated by use of vari-
ous types of tonometers (101,102). In the posterior
segment, uorescein angiography permits evaluation
of retinal perfusion following intravenous dye admin-
istration (103,104). ese tests are time-consuming
and not easily included in routine ophthalmic exami-
nations for safety studies. ey are more relevant
when selectively used for characterization of specic
ocular changes.
Routine microscopic evaluation of the eye
An important part of a preclinical toxicological study
is the microscopic evaluation of the eye for adverse
eects (105). For most studies, this evaluation follows
Good Laboratory Practices (GLP) and involves the
examination of paran-embedded, sagittal sections
of globes that are approximately 4 microns in thick-
ness and stained with hematoxylin and eosin. e
sections should include the bulbar and retrobulbar
optic nerve, but other extraocular tissues, except the
Harderian gland in rodents, are not routinely exam-
ined. For specialized studies, eyelids, extraocular
muscles, and lacrimal glands may be examined and
additional techniques, such as immunohistochem-
istry, plastic embedding, or transmission electron
microscopy, may be incorporated.
Microscopic evaluation of the eye requires the
following: 1) knowledge of the clinical ophthalmic
examination ndings in order to establish microscopic
correlates to clinical changes, 2) good histological
sections of globes for evaluation, 3) an understand-
ing of comparative ocular histology, 4) an awareness
of spontaneous background ndings that occur in the
eye of laboratory species, and 5) knowledge of toxico-
logical changes that may occur in ocular tissues (105).
Once clinical ophthalmic ndings are determined in
a study, the study pathologist needs to be notied
in order to ensure that the eyes will be handled in a
way at necropsy that will help ensure microscopic
correlations are achieved. Depending on the type of
clinical ocular change, the necropsy procedure, xa-
tion method, or sectioning procedure may need to be
adjusted. e clinical ophthalmic examiner has the
advantage of examining most of all ocular structures
by use of illumination and magnication, but the
pathologist only looks at sections a few microns in
thickness. As a result, the clinical examiner may detect
minute and localized changes, especially in normally
transparent structures, such as the cornea, lens, and
vitreous humor, but the pathologist may miss these
ndings unless there is a lot of care taken at the time
of trimming the globes to ensure appropriate sections
are acquired. In addition, artifacts may occur in ocular
structures during the preparation of ocular sections,
resulting in distortion of the ocular structures. ese
artifacts may mimic true pathological changes (e.g.,
changes in the primate lenses xed in Davidson’s x-
ative). Finding these changes in the lenses of control
animals that are clinically unremarkable is helpful in
interpretation. Although microscopic changes may
occur in an eye that is clinically unremarkable (e.g.,
cystic change in peripheral retina ), most microscopic
changes that occur in ocular structures are changes
that can be detected clinically.
Obtaining good histological sections of globes for
microscopic evaluation requires gentle enucleation,
adequate xation, and proper orientation for trim-
ming (106). Good enucleation techniques are essen-
tial in minimizing artifacts in the globe (e.g., retinal
detachment) and maintaining a portion of retrobul-
bar optic nerve for examination. In order to minimize
postmortem autolysis of tissues, especially the retina,
globes should be enucleated as soon as possible after
death and at the beginning of the postmortem exami-
nation. Once the enucleation has occurred, removal
of all extraocular tissue from the globe oers the fol-
lowing advantages: 1) facilitates xation of internal
ocular structures (e.g., retina), 2) helps to minimize
ocular artifacts (e.g., retinal detachment), 3) ensures
xation of the retrobulbar optic nerve, and 4) exposes
landmarks (e.g., long posterior ciliary artery) that will
be used for trimming the globe after xation.
A commonly used xative in toxicological pathol-
ogy is 10% neutral buered formalin, but this xative
is not considered to be optimal for the xation of
intraocular structures (106). erefore, it is routine
practice to use other xatives to preserve the eye for
microscopic examination, and use formalin as a post-
xation xative to maintain and rm the globe for
trimming. Since there is no ideal xative for the eye,
several dierent ocular xatives are commonly used.
For routine light microscopic examination of the
eye, enucleated globes should be submerged in an
adequate amount of xative. For some xatives (e.g.,
Davidson’s solution, modied Davidson’s solution, or
Bouin’s solution), submersion alone will be adequate
(107). For other xatives (e.g., glutaraldehyde), xa-
tion will need to be augmented by use of intravitreal
injection or the creation of a window (108). At times
standard light microscopic examination of eyes may
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
14 C. J. somps et al.
require additional techniques, such as immunohis-
tochemistry or transmission electron microscopy
and special xatives will be used (109). In addition,
systemic perfusion may be needed for the latter
technique.
Trimming globes requires knowledge of compara-
tive ocular anatomy and location of clinical ocular
ndings. For most toxicological studies, the ideal rou-
tine section of an eye will be along a standard sagittal
plane that extends through the central cornea, the
pupil, the center of the lens, the optic disk, and the
central portion of the retrobulbar optic nerve. When
clinical ocular ndings are identied, the standard
plane of section may need to be modied or addi-
tional sections may be needed to obtain a microscopic
correlate. is is especially true for funduscopic nd-
ings, since they may indicate alterations in the retina.
Sectioning of the globe of albino rats and mice along
a uniform plane may need to be aided by the use of
indelible dye, tattoo ink, or a small suture to ensure
a standard orientation (110). When a standard orien-
tation is achieved, the thickness of the layers of the
retina may be measured (111).
After globes are trimmed, they are inltrated with
a paran wax and embedded in blocks of paraf-
n for sectioning. Sections of the eye that are a few
microns thick are cut from the blocks, adhered to
glass microscopic slides, and stained. Routine sec-
tions of globes are usually stained with hematoxylin
and eosin, but since the globe consists of many dif-
ferent types of tissue, special stains may be needed
periodically to aid in the microscopic evaluation.
Special stains may include the following: 1) peri-
odic acid-Schi for staining basement membranes
(e.g., Descemet’s membrane, lenticular capsule,
basal lamina of ciliary epithelium, inner limiting
membrane of the retina), 2) trichrome stains to
help evaluate collagenous (e.g., corneal stroma and
sclera) and muscular (e.g. iridal and ciliary muscles)
structures, and 3) Bodian’s stain for axons, to men-
tion a few (112).
Once the histological sections of globes are pre-
pared, they are examined for microscopic changes.
e pathologist must be aware of the histological
appearance of ocular structures in the dierent
laboratory species in order to accurately identify true
changes in ocular histology (113,114). For example,
the macula is not present in the eye of a dog or rat.
e microscopic examination involves evaluation of
the cornea (e.g., epithelium, stroma, and endothe-
lium), anterior chamber, iris and ciliary body (e.g.,
epithelium, stroma, and muscle), ltration angle,
choroid, sclera, lens, retina (e.g., sensory retina
and retinal pigment epithelium), vitreous body and
optic nerve (e.g., optic disk, bulbar optic nerve, and
retrobulbar optic nerve). In addition to the sagittal
section of the bulbar optic nerve, the examination of
the optic nerve should include a cross-section of the
retrobulbar optic nerve (dog, monkey) or intracranial
optic nerve (rat). is provides an assessment of the
retinal ganglion cells by looking at their axons in the
optic nerve.
After identifying and localizing the microscopic
changes, the changes are characterized as toxicologi-
cally meaningful and those that are not toxicologically
meaningful (e.g., spontaneous background ndings). As
with the identication of clinical ophthalmic ndings,
descriptive terms from an approved lexicon are used
to characterize the ndings. Spontaneous background
ndings may include mineralized foci in the corneal
stroma of rats, mononuclear cell inltrates in the uvea
of monkeys, or multifocal retinal changes in certain
Figure 9. Histological appearance of retina stained with hema-
toxylin and eosin from Long-Evans rats used in toxicity study
involving N-nitroso-N-methylurea (MNU). A) Normal retina of a
control rat with all layers clearly visible, including the photore-
ceptor layer (white star). B) Retina with diuse loss of photore-
ceptors (white star) from a rat 3 days following a single 60-mg/kg
dose of MNU (124).
A
B
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 15
strains of rats, to name a few (115– 117). Toxicologically
meaningful ndings may be as subtle as minimal swell-
ing of lens bers (118) or as obvious as marked loss of
photoreceptors due to apoptosis as with N-methyl-N-
nitrosourea toxicity in rats (Figure 9) (119).
Once an ocular nding is identied by routine,
light microscopic examination, it may be further
characterized by the use of immunohistochemis-
try (e.g., oxidative stress of photoreceptors) (120),
plastic embedding (e.g., loss of axons in optic nerve
stained with paraphenylenediamine) (121), or
transmission electron microscopy (e.g., ultrastruc-
tural evaluation of mitochondria in the optic
nerve) (122), to mention a few. e identication of
changes in some ocular tissues may require the use
of techniques not used in standard, GLP, toxicologi-
cal studies (87). For example, the identication of
retinal ganglion cell loss involves the use of retinal
at mounts (123).
Summary
e ability to predict ocular side eects of systemi-
cally delivered drugs is an important issue for phar-
maceutical companies. Although animal models
involving standard clinical ophthalmic examina-
tions and postmortem microscopic examinations
of eyes are still used to identify ocular issues, these
methods may be supplemented with additional in
silico, in vitro, and in vivo techniques. Predictions
of ocular side eects begin with in silico compu-
tational methods for mining data that pertain to
the modulation of candidate targets and related
pathways. Structure-based in silico prediction
systems are derived either from human experts or
from computational algorithms. Although in silico
systems for predicting ocular toxicity are still being
developed, they may provide valuable information
in a screening paradigm, allowing for more suitable
targets and compounds to be chosen. In vitro assays
are also used early in drug developmental programs
to identify potential ocular toxicity. ese methods
may use surrogate cell systems to predict the risk
of ocular toxicity. e methods often include the
use of tissue or cell cultures specic to a portion of
the eye (e.g., lens). Although ocular ndings occa-
sionally occur in regulatory GLP studies involving
animals, animals may be used in early non-GLP
studies to monitor or characterize an ocular nd-
ing. In these studies, the routine ophthalmic and
ocular microscopic methods may be supplemented
with techniques that help characterize the nd-
ing (e.g., Scheimpug imaging) or may be used to
monitor the nding in clinical studies (e.g., ERG),
thus providing a method of ensuring clinical safety.
e addition of these tests to a development plan
for a potential drug provides the opportunity to save
time and money by detecting ocular issues earlier in
the program.
Acknowledgments
e authors appreciate and thank Damir Simic,
Colleen Doshna, and Dan Baltrukonis for help in the
conduct of experiments and data collection.
Declaration of interest: e authors report no conicts
of interest.
References
1. Moorthy RS, Valluri S. Ocular toxicity associated with systemic
drug therapy. Curr Opin Ophthalmol 1999; 10(6):438–446.
2. Fraunfelder FW. Ocular adverse drug reactions. Expert Opin
Drug Saf 2003; 2(4):411–420.
3. Santaella RM, Fraunfelder FW. Ocular adverse eects asso-
ciated with systemic medications: recognition and manage-
ment. Drugs 2007; 67(1):75–93.
4. Fraunfelder FW, ed. Drug-Induced Ocular Side Eects. 5th
ed; Oxford, UK: Butterworth-Heinemann. 2001.
5. Brick DC. Medication errors result in costly claims for oph-
thalmologists. Surv Ophthalmol 1995; 40(3):232–236.
6. Smith CG, O’Donnell JT, eds. e Process of New Drug
Discovery and Development. 2 ed; New York: Informa
Healthcare. 2006.
7. Illanes O, Anderson S, Niesman M, Zwick L, Jessen BA.
Retinal and peripheral nerve toxicity induced by the admin-
istration of a pan-cyclin dependent kinase (cdk) inhibitor in
mice. Toxicol Pathol 2006; 34(3):243–248.
8. Mateld M. Animal experimentation: the continuing debate.
Nat Rev Drug Discov 2002; 1(2):149–152.
9. Kramer JA, Sagartz JE, Morris DL. e application of dis-
covery toxicology and pathology towards the design of safer
pharmaceutical lead candidates. Nat Rev Drug Discov 2007;
6(8):636–649.
10. Cunha-Vaz JG. e blood-ocular barriers: past, present, and
future. Doc Ophthalmol 1997; 93(1–2):149–157.
11. Hornof M, Toropainen E, Urtti A. Cell culture models of the
ocular barriers. Eur J Pharm Biopharm 2005; 60(2):207–225.
12. Davila JC, Rodriguez RJ, Melchert RB, Acosta D. Predictive
value of in vitro model systems in toxicology. Annu Rev
Pharmacol Toxicol 1998; 38(1):63–96.
13. Kruszewski F. Human cells as in vitro alternatives for ocular
toxicity studies. Comments Toxicol 1998; 6(3):221–233.
14. Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello
CC. Potent and specic genetic interference by double-
stranded RNA in Caenorhabditis elegans. Nature 1998;
391(6669):806–811.
15. Peier RL, McCary B, Bee W, Wegener A, Geller AM, Boyes
WK. Contemporary methods in ocular toxicology. Toxicol
Mech Methods 2000; 10(1):17–39.
16. Loging W, Harland L, Williams-Jones B. High-throughput
electronic biology: mining information for drug discovery.
Nat Rev Drug Discov 2007; 6(3):220–230.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
16 C. J. somps et al.
17. Pollard J, Butte AJ, Hoberman S, Joshi M, Levy J, Pappo J. A
computational model to dene the molecular causes of type 2
diabetes mellitus. Diabetes Technol er 2005; 7(2):323–336.
18. Musante CJ, Lewis AK, Hall K. Small- and large-scale biosim-
ulation applied to drug discovery and development. Drug
Discov Today 2002; 7(20 suppl):S192–S196.
19. Taylor JB, Triggle DJ. eds. Comprehensive Medicinal II;
New York: Elsevier Science. 2006.
20. Greene N. Computational models to predict toxicity. In:
Triggle DJ. ed. Comprehensive Medicinal Chemistry II. 1 ed;
Oxford: Elsevier Ltd. 2007:909–932.
21. Hale R. Text mining: getting more value from literature
resources. Drug Discov Today 2005; 10(6):337–379.
22. Elliston K. Breaking through the cognitive barriers that
impede critical path research. Am Biotechnol Lab 2004;
22(12):28–30.
23. Bender AT, Beavo JA. Cyclic nucleotide phosphodiesterases :
molecular regulation to clinical use. Pharmacol Rev 2006;
58(3):488–520.
24. Lugnier C. Cyclic nucleotide phosphodiesterase (PDE)
superfamily: a new target for the development of specic
therapeutic agents. Pharmacol er 2006; 109(3):366–398.
25. Chabre M, Deterre P. Molecular mechanism of visual trans-
duction. Eur J Biochem 1989; 179(2):255–266.
26. Yartz S, Hurley JB. Transduction mechanisms of verte-
brate and invertebrate photoreceptors. J Biol Chem 1994;
269(20):14329–14332.
27. Lolley RN, Farber DB, Rayborn ME, Hollyeld JG. Cyclic
GMP accumulation causes degeneration of photorecep-
tor cells: simulation of an inherited disease. Science 1977;
196(4290):664–666.
28. Kameni Tcheudji JF, Lebeau L, Virmaux N, Maftei CG,
Cote RH, Lugnier C, Schultz P. Molecular organization of
bovine rod cGMP-phosphodiesterase 6. J Mol Biol 2001;
310(4):781–791.
29. Cote RH. Characteristics of photoreceptor PDE (PDE6):
similarities and dierences to PDE5. Int J Impot Res 2004;
16(suppl 1):S28–S33.
30. Marmor MF, Kessler R. Sildenal (Viagra) and ophthalmol-
ogy. Surv Ophthalmol 1999; 44(2):153–162.
31. Muster W, Breidenbach A, Fischer H, Kirchner S, Muller L,
Pahler A. Computational toxicology in drug development.
Drug Discov Today 2008; 13(7–8):303–310.
32. Egan WJ, Zlokarnik G, Grootenhuis PDJ. In silico prediction
of drug safety: despite progress there is abundant room for
improvement. Drug Discov Today Technol 2004; 1(4):381–387.
33. Chen YZ, Yap CW, Li H. Current QSAR techniques for toxi-
cology. In: Ekins, S. ed. Computational Toxicology: Risk
Assessment for Pharmaceutical and Environmental Chemicals.
New York: John Wiley & Sons. 2007:217–238.
34. Ridings JE, Barratt MD, Cary R, Earnshaw CG, Eggington CE,
Ellis MK, Judson PN, Langowski JJ, Marchant CA, Payne MP,
Watson WP, Yih TD. Computer prediction of possible toxic
action from chemical structure: an update on the DEREK
system. Toxicology 1996; 106(1–3):267–279.
35. Greene N, Judson PN, Langowski JJ, Marchant CA.
Knowledge-based expert systems for toxicity and metabo-
lism prediction: DEREK, StAR and METEOR. SAR QSAR
Environ Res 1999; 10(2–3):299–314.
36. Greene N. Computer systems for the prediction of toxicity:
an update. Adv Drug Deliv Rev 2002; 54(3):417–431.
37. Hobbs HE, Sorsby A, Freedman A. Retinopathy following
chloroquine therapy. Lancet 1959; 2(7101):478–480.
38. Maguire A. Amodiaquine hydrochloride. Corneal deposits
and pigmented palate and nails after treatment of chronic
discoid lupus erythematosus. Lancet 1962; 1(7231)667.
39. Isaac NE, Walker AM, Jick H, Gorman M. Exposure to phe-
nothiazine drugs and risk of cataract. Arch Ophthalmol 1991;
109(2):256–260.
40. Subashini K, Rao VA. Chlorpromazine-induced cata-
ract and corneal pigmentation. Indian J Pharmacol 2004;
36(5):323–324.
41. Jaanus SD. Ocular side eects of selected systemic drugs.
Optom Clin 1992; 2(4):73–96.
42. Power WJ, Travers SP, Mooney DJ. Welding arc maculopathy
and uphenazine. Br J Ophthalmol 1991; 75(7):433–435.
43. Ubels JL, Clousing DP. In vitro alternatives to the use
of animals in ocular toxicology testing. Ocul Surf 2005;
3(3):126–142.
44. EMEA Report. Note for guidance on photosafety testing.
CPMP/SWP/398/01. 2002. http://www.emea.europa.eu/
pdfs/human/swp/039801en.pdf. accessed June 27, 2002.
45. Borenfreund E, Puerner JA. Toxicity determined in vitro
by morphological alterations and neutral red absorption.
Toxicol Lett 1985; 24(2–3):119–124.
46. Nilsson R, Maurer T, Redmond N. A standard protocol for
phototoxicity testing. Results from an interlaboratory study.
Contact Dermatitis 1993; 28(5):285–290.
47. Spielmann H, Balls M, Dupuis J, Pape WJ, Pechovitch G,
Silva Od, Holzhuetter HG, Clothier R, Desolle P, Gerberick F,
Liebsch M, Lovell WW, Maurer T, Pfannenbecker U, Potthast
JM, Csato M, Sladowski D, Steiling W, Brantom P. e
International EU/COLIPA In Vitro Phototoxicity Validation
Study: results of phase II (blind trial). Part 1: the 3T3 NRU
phototoxicity test. Toxicol In Vitro 1998; 12(3):305–327.
48. Peters B, Holzhutter HG. In vitro phototoxicity testing: devel-
opment and validation of a new concentration response
analysis software and biostatistical analyses related to the
use of various prediction models. Altern Lab Anim 2002;
30(4):415–432.
49. Spielmann H, Muller L, Averbeck D, Balls M, Brendler-
Schwaab S, Castell JV, Curren R, de Silva O, Gibbs NK,
Liebsch M, Lovell WW, Merk HF, Nash JF, Neumann NJ,
Pape WJ, Ulrich P, Vohr HW. e second ECVAM workshop
on phototoxicity testing. e report and recommendations
of ECVAM workshop 42. Altern Lab Anim 2000;
28(6):777–814.
50. Epstein JH, Wintroub BU. Photosensitivity due to drugs.
Drugs 1985; 30(1):42–57.
51. Epstein JH. Phototoxicity and photoallergy. Semin Cutan
Med Surg 1999; 18(4):274–284.
52. Onoue S, Tsuda Y. Analytical studies on the prediction of
photosensitive/phototoxic potential of pharmaceutical sub-
stances. Pharm Res 2006; 23(1):156–164.
53. Onoue S, Kawamura K, Igarashi N, Zhou Y, Fujikawa M,
Yamada H, Tsuda Y, Seto Y, Yamada S. Reactive oxygen spe-
cies assay-based risk assessment of drug-induced phototox-
icity: classication criteria and application to drug candi-
dates. J Pharm Biomed Anal 2008; 47(4–5):967–972.
54. Lynch AM, Robinson SA, Wilcox P, Smith MD, Kleinman M,
Jiang K, Rees RW. Cycloheximide and disulfoton are posi-
tive in the photoclastogencity assay but do not absorb UV
irradiation: another example of pseudophotoclastogenicity?
Mutagenesis 2008; 23(2):111–118.
55. Aleo MD, Avery MJ, Beierschmitt WP, Drupa CA, Fortner JH,
Kaplan AH, Navetta KA, Shepard RM, Walsh CM. e use of
explant lens culture to assess cataractogenic potential. Ann
N Y Acad Sci 2000; 919: 71–187.
56. Aleo MD, Doshna CM, Navetta KA. Ciglitazone-induced
lenticular opacities in rats: in vivo and whole lens explant
culture evaluation. J Pharmacol Exp er 2005;
312(3):1027–1033.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
Ocular toxicity of systemically delivered drugs 17
57. Walsh Clang CM, Aleo MD. Mechanistic analysis of S-(1,2-
dichlorovinyl)-L-cysteine-induced cataractogenesis in vitro.
Toxicol Appl Pharmacol 1997; 146(1):144–155.
58. Pauls E, Este JA. RNA interference as a tool for target valida-
tion. Drug Discov Today Technol 2004; 1(2):135–140.
59. Choudhary S, Xiao T, Vergara LA, Srivastava S, Nees D,
Piatigorsky J, Ansari NH. Role of aldehyde dehydrogenase
isozymes in the defense of rat lens and human lens epithe-
lial cells against oxidative stress. Invest Ophthalmol Vis Sci
2005; 46(1):259–267.
60. Lassen N, Bateman JB, Estey T, Kuszak JR, Nees DW,
Piatigorsky J, Duester G, Day BJ, Huang J, Hines LM,
Vasiliou V. Multiple and additive functions of ALDH3A1 and
ALDH1A1: cataract phenotype and ocular oxidative damage
in Aldh3a1(-/-)/Aldh1a1(-/-) knock-out mice. J Biol Chem
2007; 282(35):25668–25676.
61. Kirby TJ. Cataracts produced by triparanol (MER-29). Trans
Am Ophthalmol Soc 1967; 65:494–543.
62. MacDonald JS, Gerson RJ, Kornbrust DJ, Kloss MW,
Prahalada S, Berry PH, Alberts AW, Bokelman DL.
Preclinical evaluation of lovastatin. Am J Cardiol 1988;
62(15):16J–27J.
63. Gerson RJ, MacDonald JS, Alberts AW, Kornbrust DJ,
Majka JA, Stubbs RJ, Bokelman DL. Animal safety and toxi-
cology of simvastatin and related hydroxy-methylglutaryl-
coenzyme A reductase inhibitors. Am J Med 1989; 87(4A):
28S–38S.
64. Gerson RJ, MacDonald JS, Alberts AW, Chen J, Yudkovitz
JB, Greenspan MD, Rubin LF, Bokelman DL. On the etiol-
ogy of subcapsular lenticular opacities produced in dogs
receiving HMG-CoA reductase inhibitors. Exp Eye Res 1990;
50(1):65–78.
65. Gerson RJ, Allen HL, Lankas GR, MacDonald JS, Alberts AW,
Bokelman DL. e toxicity of a uorinated-biphenyl HMG-
CoA reductase inhibitor in beagle dogs. Fundam Appl Toxicol
1991; 16(2):320–329.
66. Mosley ST, Kalinowski SS, Schafer BL, Tanaka RD. Tissue-
selective acute eects of inhibitors of 3-hydroxy-3-methylgl-
utaryl coenzyme A reductase on cholesterol biosynthesis in
lens. J Lipid Res 1989; 30(9):1411–1420.
67. de Vries AC, Vermeer MA, Hendriks AL, Bloemendal H,
Cohen LH. Biosynthetic capacity of the human lens upon
aging. Exp Eye Res 1991; 53(4):519–524.
68. Aleo MD. Lenticular toxicity of systemic therapeutics in the rat:
case studies and mechanism-based screening approaches.
Toxicologist CD Ocial J Soc Toxicol 2008; 102(S-1).
69. Heywood R, Gopinath C. Morphological assessment of vis-
ual dysfunction. Toxicol Pathol 1990; 18(1 pt 2):204–217.
70. Schiavo DM. e use of laboratory animals in toxicology: an
ophthalmoscopic assessment. Toxicol Pathol 1990; 18(1 pt
2):222–223.
71. Rubin LF. Current methods of eye examination in vivo. In:
Hockwin O, Green K, Rubin LF. eds; Manual of ototoxicity
testing of drugs. New York: Gustav Fischer Verlag. 1992,
pp. 81–108.
72. Kuiper B, Boeve MH, Jansen T, Roelofs-van Emden ME,
uring JW, Wijnands MV. Ophthalmologic examination
in systemic toxicity studies: an overview. Lab Anim 1997;
31(2):177–183.
73. Schiavo DM. e use of laboratory animals in toxicology:
an ophthalmoscopic assessment. Toxicol Pathol 1990; 18:
222–223.
74. Guymer RH, Chiu AW, Lim L, Baird PN. HMG CoA
reductase inhibitors (statins): do they have a role in age-
related macular degeneration? Surv Ophthalmol 2005;
50(2):194–206.
75. omas D, Duguid G. Optical coherence tomography—a
review of the principles and contemporary uses in retinal
investigation. Eye 2004; 18(6):561–570.
76. Schuman JS, Puliato CA, Fujimoto JG. Optical Coherence
Tomography of Ocular Diseases. 2nd ed; orofare, NJ:
Slack. 2004; pp. 3–53.
77. Grieve K, Paques M, Dubois A, Sahel J, Boccara C, Le
Gargasson JF. Ocular tissue imaging using ultrahigh-res-
olution, full-eld optical coherence tomography. Invest
Ophthalmol Vis Sci 2004; 45(11):4126–4131.
78. Baltrukonis D, Fortner JH, Ryan AM, Somps CJ, Aleo MD,
Verdugo ME. Comparisons between U18666A, naphtha-
lene, and galactose on lens cholesterol biosynthesis and
evaluation of cataract progression by Scheimpug and
slit lamp exams. Association for Research in Vision and
Ophthalmology (ARVO), May 4–9, 2003 in Fort Lauderdale,
FL, USA. 44ARVO E-Abstract 3489; 2003.
79. Baltrukonis D, Fortner JH, Doshna CM, Render JA, Verdugo
ME, Somps CJ. Evaluation of lens transparency using
Scheimpug photography and slit lamp biomicroscopy
in the CJ-12,918 cataract model in the albino rat. Invest
Ophthalmol Vis Sci 2004; 45(suppl 1, Apr):U171.
80. Petrig BL, Riva CE, Hayreh SS. Laser Doppler owmetry
and optic nerve head blood ow. Am J Ophthalmol 1999;
127(4):413–425.
81. Harris A, Ciulla TA, Chung HS, Martin B. Regulation of
retinal and optic nerve blood ow. Arch Ophthalmol 1998;
116(11):1491–1495.
82. Riva CE, Grunwald JE, Sinclair SH. Laser Doppler veloci-
metry study of the eect of pure oxygen breathing on
retinal blood ow. Invest Ophthalmol Vis Sci 1983;
24(1):47–51.
83. Fishman GA, Birch DG, Holder GE, Brigell MG.
Electrophysiologic Testing in Disorders of the Retina, Optic
Nerve, and Visual Pathway. 2nd ed; New York: Oxford
University Press. 2001.
84. Narfstrom K, Ekesten B, Rosolen SG, Spiess BM, Percicot CL,
Ofri R. Guidelines for clinical electroretinography in the dog.
Doc Ophthalmol 2002; 105(2):83–92.
85. Rosolen SG, Kolomiets B, Varela O, Picaud S. Retinal elec-
trophysiology for toxicology studies: applications and limits
of ERG in animals and ex vivo recordings. Exp Toxicol Pathol
2008; 60(1):17–32.
86. Bee WH. Standardized electroretinography in primates:
a non-invasive preclinical tool for predicting ocular side
eects in humans. Curr Opin Drug Discov Devel 2001;
4(1):81–91.
87. Hockwin O, Green K, Rubin LF, eds. Manual of Ototoxicity
Testing of Drugs. New York: Gustav Fischer Verlag. 1992:
109–413.
88. Williams KK, Shepard AR, Rice RL, McCartney MD, Wax
MB, Hiddemen JW. Corneal wound healing in New Zealand
white rabbits following anterior keratectomy and treatment
with moxioxacin ophthalmic solution 0.5% or gatioxacin
ophthalmic solution 0.3%. J Ocul Pharmacol er 2007;
23(6):517–525.
89. Viau S, Maire MA, Pasquis B, Gregoire S, Fourgeux C, Acar
N, Bretillon L, Creuzot-Garcher CP, Jore C. Time course
of ocular surface and lacrimal gland changes in a new sco-
polamine-induced dry eye model. Graefes Arch Clin Exp
Ophthalmol 2008; 246(6):857–867.
90. Barabino S, Chen W, Dana MR. Tear lm and ocular surface
tests in animal models of dry eye: uses and limitations. Exp
Eye Res 2004; 79(5):613–621.
91. Xiong C, Chen D, Liu J, Liu B, Li N, Zhou Y, Liang X, Ma P, Ye
C, Ge J, Wang Z. A rabbit dry eye model induced by topical
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
18 C. J. somps et al.
medication of a preservative benzalkonium chloride. Invest
Ophthalmol Vis Sci 2008; 49(5):1850–1856.
92. Liu Y, Gan L, Carlsson DJ, Fagerholm P, Lagali N, Watsky MA,
Munger R, Hodge WG, Priest D, Grith M. A simple, cross-
linked collagen tissue substitute for corneal implantation.
Invest Ophthalmol Vis Sci 2006; 47(5):1869–1875.
93. Kaps S, Richter M, Spiess BM. Corneal esthesiometry in the
healthy horse. Vet Ophthalmol 2003; 6(2):151–155.
94. Ollivier FJ, Brooks DE, Komaromy AM, Kallberg ME, Andrew
SE, Sapp HL, Sherwood MB, Dawson WW. Corneal thick-
ness and endothelial cell density measured by non-contact
specular microscopy and pachymetry in Rhesus macaques
(Macaca mulatta) with laser-induced ocular hypertension.
Exp Eye Res 2003; 76(6):671–677.
95. Kowalski RP, Romanowski EG, Mah FS, Yates KA, Gordon YJ.
Intracameral Vigamox (moxioxacin 0.5%) is non-toxic and
eective in preventing endophthalmitis in a rabbit model.
Am J Ophthalmol 2005; 140(3):497–504.
96. Zhou X, Qu J, Xie R, Wang R, Jiang L, Zhao H, Wen J, Lu F.
Normal development of refractive state and ocular dimen-
sions in guinea pigs. Vision Res 2006; 46(18):2815–2823.
97. Fernandes A, Bradley DV, Tigges M, Tigges J, Herndon JG.
Ocular measurements throughout the adult life span of rhesus
monkeys. Invest Ophthalmol Vis Sci 2003; 44(6):2373–2380.
98. Kiuchi T, Tatsuzaki H, Hommura S, Oshika T. Specular
microscopic study of X-ray-irradiated rabbit cornea. Eye
2004; 18(9):929–934.
99. Smith RS, Korb D, John SW. A goniolens for clinical monitor-
ing of the mouse iridocorneal angle and optic nerve. Mol Vis
2002; 8:26–31.
100. Johnstone MA . e aqueous outow system as a mechani-
cal pump: evidence from examination of tissue and aque-
ous movement in human and non-human primates. J
Glaucoma 2004; 13(5):421–438.
101. Gelatt KN, Mackay EO. Eect of dierent dose schedules
of bimatoprost on intraocular pressure and pupil size in
the glaucomatous beagle. J Ocul Pharmacol er 2002;
18(6):525–534.
102. Giurida S, Bucolo C, Drago F. Topical application of a nitric
oxide synthase inhibitor reduces intraocular pressure in
rabbits with experimental glaucoma. J Ocul Pharmacol er
2003; 19(6):527–534.
103. Le Meur G, Weber M, Pereon Y, Mendes-Madeira A, Nivard
D, Deschamps JY, Moullier P, Rolling F. Postsurgical assess-
ment and long-term safety of recombinant adeno-associated
virus-mediated gene transfer into the retinas of dogs and pri-
mates. Arch Ophthalmol 2005; 123(4):500–506.
104. Julien S, Kreppel F, Beck S, Heiduschka P, Brito V, Schnichels
S, Kochanek S, Schraermeyer U. A reproducible and quan-
tiable model of choroidal neovascularization induced by
VEGF A165 after subretinal adenoviral gene transfer in the
rabbit. Mol Vis 2008; 14:1358–1372.
105. Weisse I. Microscopic examination of the eye. In: Hockwin O,
Green K, Rubin LF, eds. Manual of Ocular Testing of Drugs.
New York: Gustav Fischer Verlag. 1992; 137–163.
106. Saunders LZ, Rubin LF. Ophthalmic Pathology of Animals.
An Atlas and Reference Book. Basel, Germany: S. Karger. 1975;
pp. 244–253.
107. Latendresse JR, Warbrittion AR, Jonassen H. Creasy DM.
Fixation of testes and eyes using a modied Davidson’s
uid: comparison with Bouin’s uid and conventional
Davidson’s uid. Toxicol Pathol 2002; 30(4):524–533.
108. Saby JA, Sigler RE, Klaus S. Comparison of xatives for his-
tologic evaluation of the canine eye. J Histotechnol 1991;
14:251–255.
109. Smith RS, Zabaleta A, Simon WMJ, Bechtold LS, Ikeda S,
Relyea MJ, Sundberg JP, Kao WW-Y, Liu C-Y. General and
special histopathology. In: Smith R, ed. Systemic Evaluation
of the Mouse Eye: Anatomy, Pathology, and Biomethods. New
York: CRC Press. 2002:265–281.
110. Smith RS, Nishina PM, Ikeda S, Jewett P, Zabaleta A, Simon
WMJ. Interpretation of ocular pathology in genetically engi-
neered and spontaneous mutant mice. In: eds. Ward JM,
Mahler JF, Maronpot RR, Sundberg JP, Frederickson RM,
Pathology of Genetically Engineered Mice. Ames, IA: Iowa
Sate University Press. 2000:217–231.
111. Michon JJ, Li ZL, Shioura N, Anderson RJ, Tso MO. A com-
parative study of methods of photoreceptor morphometry.
Invest Ophthalmol Vis Sci 1991; 32(2):280–284.
112. Lee WR. ed. Ophthalmic Histopathology. 2nd ed; London,
UK: Springer-Verlag. 2002; pp. 19.
113. Millichamp NJ. Species specicity. Factors aecting the
interpretation of species dierences in toxic responses of
ocular tissues. In: Chiou GCY. ed. Ophthalmic Toxicology.
New York: Raven Press. 1992:59–82.
114. Rubin LF. Comparative anatomy of the eye. In: eds. Hockwin
O, Green K, Rubin LF. . Manual of Ocular Testing of Drugs.
New York: Gustav Fischer Verlag. 1992:219–244.
115. Taradach C, Greaves P. Spontaneous eye lesions in labora-
tory animals: incidence in relation to age. Crit Rev Toxicol
1984; 12(2):121–147.
116. Sinha DP, Cartwright ME, Johnson RC. Incidental mono-
nuclear cell inltrate in the uvea of cynomolgus monkeys.
Toxicol Pathol 2006; 34(2):148–151.
117. Hubert MF, Gillet JP, Durand-Cavagna G. Spontaneous
retinal changes in Sprague Dawley rats. Lab Anim Sci 1994;
44(6):561–567.
118. Render JA, Carlton WW. Induced cataracts, lens, rat. In:
eds. Jones TC, Mohr U, Hunt RD. Eye and Ear. New York:
Springer-Verlag. 1991:63–73.
119. Kiuchi K, Yoshizawa K, Shikata N, Matsumura M, Tsubura A.
Nicotinamide prevents N-methyl-N-nitrosourea-induced
photoreceptor cell apoptosis in Sprague-Dawley rats and
C57BL mice. Exp Eye Res 2002; 74(3):383–392.
120. Tanito M, Kaidzu S, Anderson RE. Delayed loss of
cone and remaining rod photoreceptor cells due to
impairment of choroidal circulation after acute light
exposure in rats. Invest Ophthalmol Vis Sci 2007;
48(4):1864–1872.
121. Sadun AA, Smith LE, Kenyon KR. Paraphenylenediamine:
a new method for tracing human visual pathways. J
Neuropathol Exp Neurol 1983; 42(2):200–206.
122. Mutsaers SE, Carroll WM. Focal accumulation of intra-ax-
onal mitochondria in demyelination of the cat optic nerve.
Acta Neuropathol 1998; 96(2):139–143.
123. Silva-Araujo A, Silva MC, Abreu-Dias P, Tavares MA. Eects
of prenatal cocaine exposure in the retinal ganglion cell
layer of the rat. A morphometric analysis. Mol Neurobiol
1995; 11(1–3):87–97.
124. Liu C-N, Simic D, Baltrukonis D, Doshna CM, Render JA,
Fortner JH, Verdugo ME, Weisenburger WP, Somps CJ. Early
MNU-induced retinal toxicity detected by electroretinogra-
phy and visual acuity in funduscopically normal rats. Invest
Ophthalmol Vis Sci 2004; 45(suppl 1, Apr): U341.
Cutaneous and Ocular Toxicology Downloaded from informahealthcare.com by Pfizer Ltd on 10/09/12
For personal use only.
