systems biology approaches for toxicology

SYSTEMS BIOLOGY AND NEURODEGENERATION 201

Published in 2007 John Wiley & Sons, Ltd. J. Appl. Toxicol. 2007; 27: 201–217

DOI: 10.1002/jat

JOURNAL OF APPLIED TOXICOLOGYJ. Appl. Toxicol. 2007; 27: 201–217Published online 30 January 2007 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/jat.1207

REVIEWSystems biology approaches for toxicology†

William Slikker, Jr,* Merle G. Paule, Linnzi K. M. Wright, Tucker A. Patterson and Cheng Wang

National Center for Toxicological Research, U S Food and Drug Administration, Jefferson, Arkansas, USA

Received 13 June 2006; Revised 26 October 2006; Accepted 27 October 2006

ABSTRACT: Systems biology/toxicology involves the iterative and integrative study of perturbations by chemicals and

other stressors of gene and protein expression that are linked firmly to toxicological outcome. In this review, the value

of systems biology to enhance the understanding of complex biological processes such as neurodegeneration in the devel-

oping brain is explored. Exposure of the developing mammal to NMDA (N-methyl-D-aspartate) receptor antagonists

perturbs the endogenous NMDA receptor system and results in enhanced neuronal cell death. It is proposed that continu-

ous blockade of NMDA receptors in the developing brain by NMDA antagonists such as ketamine (a dissociative

anesthetic) causes a compensatory up-regulation of NMDA receptors, which makes the neurons bearing these receptors

subsequently more vulnerable (e.g. after ketamine washout), to the excitotoxic effects of endogenous glutamate: the

up-regulation of NMDA receptors allows for the accumulation of toxic levels of intracellular Ca2+++++ under normal physio-

logical conditions. Systems biology, as applied to toxicology, provides a framework in which information can be arranged

in the form of a biological model. In our ketamine model, for example, blockade of NMDA receptor up-regulation by the

co-administration of antisense oligonucleotides that specifically target NMDA receptor NR1 subunit mRNA, dramatically

diminishes ketamine-induced cell death. Preliminary gene expression data support the role of apoptosis as a mode of action

of ketamine-induced neurotoxicity. In addition, ketamine-induced cell death is also prevented by the inhibition of NF-κκκκκB

translocation into the nucleus. This process is known to respond to changes in the redox state of the cytoplasm and has

been shown to respond to NMDA-induced cellular stress. Although comprehensive gene expression/proteomic studies and

mathematical modeling remain to be carried out, biological models have been established in an iterative manner to

allow for the confirmation of biological pathways underlying NMDA antagonist-induced cell death in the developing

nonhuman primate and rodent. Published in 2007 John Wiley & Sons, Ltd.

KEY WORDS: development; NMDA receptor antagonist; anesthetic agents; neurodegeneration; DNA repair; gene expression;

ArrayTrack

biology approach. It is the appropriate placement of these

biological modules into a proposed mechanistic flow

scheme, thus allowing for the development of integrated

computational models. However, the development of

these mathematical models often lags behind the initial

definition of the system and remains to be accomplished

for the current example.

For toxicology it is essential that quantitative correla-

tions of exposure (i.e. dose, time intervals and outcome)

be integrated into the computational model (Henry, 2003).

In addition to knowledge about the proximate toxicant

and its mechanism of action, the primary toxicological

effect or phenotypic anchor must also be utilized (Waters

et al., 2003b). At the systems biology level, quantitative

simulations can be conducted and predictions of the

model can be tested. The outcome of these iterations is

systematically incorporated back into the model to im-

prove its design and refine its predictive capabilities. The

interconnectivity of a system at this level determines its

state and extends its predictive power (Jazwinski, 2002).

The goal of systems biology is to predict the functional

outcomes of component-to-component relationships using

Introduction

Systems biology has been defined as the iterative and

integrative study of biological systems as they respond to

perturbations (Auffray et al., 2003). In this review, sys-

tems biology is explored as an approach to enhance the

understanding of complex biological processes such as

neurodegeneration in the developing nervous system.

High throughput molecular biology approaches includ-

ing genomics, proteomics and metabolomics provide the

fundamental data necessary for the building blocks of

systems biology. As these databases grow and become

linked together as integrated modules, they will provide

the intermediate components necessary for the systems

* Correspondence to: Dr William Slikker, Jr., National Center for Toxic-

ological Research/FDA, 3900 NCTR Road, Jefferson, AR 72079-9502,

USA.

E-mail: [email protected]

† This article is a U.S. Government work and is in the public domain in the

U.S.A.

202 W. SLIKKER ET AL.


DOI: 10.1002/jat

computational models that allow for the directional

and quantitative description of the complete organism in

response to environmental perturbations (Waters et al.,

2003a). Systems biology approaches can also be used as

effective tools for dissecting the mechanisms underlying

toxicological phenomena associated with exposure to

toxicants. It is the development of predictive models that

integrate responses across different organizational levels

that is the focus of this review.

The success of the systems biology approach to

solve toxicological problems lies in the establishment of

cross-disciplinary teams of scientists including toxicolo-

gists, molecular biologists, mathematicians, computational

modelers and risk assessors. The integration of rapidly

growing biological databases, including models of cells,

tissues and organs, with the use of powerful computing

systems and algorithms is necessary (Noble, 2003). These

interdisciplinary scientists are conducting systematic

experiments that account for small variations in a large

number of model components in order to determine the

overall functioning of the biological system (Auffray

et al., 2003).

One recently described developmental neurodegen-

erative study in rat pups involves the apoptotic cell

death of neurons in several brain regions following

postnatal exposure to ketamine (Ikonomidou et al.,

1999). Ketamine, a noncompetitive N-methyl-D-aspartate

(NMDA) receptor ion channel blocker, is used as a

general pediatric anesthetic. It is a nonbarbiturate,

dissociative anesthetic commonly used during short diag-

nostic and surgical procedures in infants and toddlers.

Recent data from developing rats (Ikonomidou et al.,

1999; Jevtovic-Todorovic et al., 2003; Scallet et al.,

2004; Wang et al., 2005b) suggest that anesthetic drugs

may cause dose-dependent neurodegeneration. The win-

dow of vulnerability to this effect of ketamine is re-

stricted to the period of rapid synaptogenesis (also known

as the brain growth spurt), which occurs immediately

after neurons have differentiated and migrated to their

final destinations. It is postulated that excessive suppres-

sion of neuronal activity by ketamine during the brain

growth spurt triggers neurons to commit ‘suicide’ via

apoptosis.

The developing nervous system may be more or less

susceptible to neurotoxic insult depending on the stage of

development. Because of the complexity and temporal

features of the manifestations of developmental neuro-

toxicity, this area of toxicology can benefit from a sys-

tems biology approach. The main purpose of this review

is to outline the application of the systems biology

approach to the problem of developmental neurodegener-

ation produced by ketamine and related NMDA anta-

gonists. It is proposed that the administration of ketamine

during critical developmental periods will result in a

dose-related increase in neurotoxicity (loss of neurons) by

a mechanism that involves a compensatory up-regulation

of NMDA receptor subunits. If NMDA receptor NR1

subunit antisense oligonucleotides can protect neurons

from ketamine-induced cell death, then alterations in

NMDA receptor function might be a key mechanism

underlying the enhanced neurodegeneration induced by

ketamine during development.

In order to better determine if ketamine-induced neuro-

toxicity in the developing rat has clinical relevance,

ketamine would best be examined in a nonhuman primate

model that more closely mimics the pediatric population.

The four steps of a systems biology approach reported by

Leroy Hood’s group (Auffray et al., 2003) will be dis-

cussed in this context. First, available information on the

biological system of interest will be described and a

preliminary model of how the system functions will be

formulated. Second, where possible, the genes and pro-

teins expressed in the described pathways will be defined.

However, information about genetic perturbations of

the system will generally not be available. Third, kinetic

experiments providing information across important

periods of development will be considered. Fourth, vari-

ous global datasets will be integrated to determine if

they support the model. Discrepancies will be identified

and hypotheses-driven studies will be conducted in order

to address them. Thus, data generated via iteration of

the third and fourth steps will be used to reformulate

the model in light of new data. Although mathematical

modeling is an ultimate goal of the systems biology

approach, it is not possible at this time to achieve this

computational step.

The Role of Glutamatergic Transmissionand NMDA Receptor Dysfunction inAnesthetic-induced Neurodegeneration

The amino acid L-glutamate is generally recognized as

the major excitatory neurotransmitter of the mammalian

central nervous system (CNS) and glutamate receptors

play a major role in fast excitatory synaptic transmission.

Glutamate promotes neuronal migration, differentiation

and plasticity during development and throughout life

(Komuro and Rakic, 1993). Malfunctions of the gluta-

mate system can affect neuroplasticity and cause neuronal

toxicity. In the case of anesthetic-induced neurodegen-

eration, many glutamate-regulated processes seem to be

perturbed. Abnormal neuronal development, abnormal

synaptic plasticity and neurodegeneration have been pro-

posed as mechanisms that underlie anesthetic-induced

neuronal cell loss. It is becoming clear that some of the

most important functions of the nervous system, such

as synaptic plasticity and synapse formation, critically

depend on the behavior of NMDA receptors and that

neurological damage caused by a variety of pathological

states can result from exaggerated or inappropriate activ-

ation of NMDA receptors (Choi, 1988; Olney, 1990).



DOI: 10.1002/jat

Here, we discuss how the cellular aspects of altered

NMDA receptor function can explain some of the

neuropathology observed in anesthetic-induced neuro-

degeneration and how some experimental approaches

can intervene via modulation of the glutamate system.

NMDA receptors are widely distributed throughout

the CNS. The NMDA receptor is a ligand-activated ion

channel primarily composed of two families of NMDA

receptor subunits: NR1 with eight splice variants and

NR2 (A-D) (Ishii et al., 1993; Kutsuwada et al., 1992;

Monyer et al., 1992; Moriyoshi et al., 1991; Zukin

and Bennett, 1995; Durand et al., 1992; Paupard et al.,

1997). The NR1 subunit is essential for receptor function.

The functional properties of the NMDA receptor vary

throughout the CNS and the binding affinities of various

ligands for recombinant NMDA receptors depend on

subunit composition (Laurie and Seeburg, 1994; Nishi

et al., 2001; Wong et al., 2002). NR3A has at least two

splice variants (Sun et al., 1998).

NMDA receptors are involved in a variety of physio-

logical and pathological processes, including memory

and learning (Collingridge et al., 1983), neuronal deve-

lopment (D’Souza et al., 1993), epileptiform seizures,

synaptic plasticity (Meldrum and Garthwaite, 1990) and

acute neuropathologies such as stroke and other trauma-

related events (Beal, 1992). There is also evidence for

their involvement in chronic neuropathologies such as

Alzheimer’s (Cotman et al., 1989), Parkinson’s and

Huntington’s diseases (Meldrum and Garthwaite, 1990;

Greenamyre, 1993) and mental illnesses such as schizo-

phrenia and anxiety disorder (Meldrum and Garthwaite,

1990).

Modifications of synaptic efficacy are believed to play

an important role in information processing and storage

by neuronal networks. The sialic acid polymer (PSA) on

neural cell adhesion molecules (NCAM) is an important

regulator of cell surface interactions (Muller et al., 1996).

PSA-NCAM is also a neuron specific marker known to

be an NMDA-regulated molecule important in synapto-

genesis during development (Wang et al., 2005a).

The blockade of NMDA receptors is known to be

neurotoxic in some instances, but the underlying mecha-

nisms involved are unknown. The administration of

NMDA receptor antagonists such as ketamine, phencycli-

dine (PCP) and MK-801 to rats during a critical time of

development (perinatally) results in neurodegeneration

in several major brain areas (Ikonomidou et al., 1999;

Scallet et al., 2004). In 1999, Olney and co-workers

demonstrated severe widespread apoptotic degeneration

throughout the rapidly developing brain of the 7-day-old

rat pup after ketamine administration (Ikonomidou et al.,

1999). Our previous studies have also demonstrated

that repeated administration of PCP (a non-competitive

NMDA antagonist) results in a sensitized locomotor

response in rats subjected to later PCP challenge

(Johnson et al., 1998). This sensitization is associated

with apoptotic cell death and an increase in NMDA

receptor NR1 subunit mRNA and immunoreactivity of

the NMDA receptor in rat forebrain (Hanania et al.,

1999; Wang et al., 1999). The neurodegeneration and

associated deficits in prepulse inhibition were prevented

by treatment with a superoxide dismutase mimetic,

M40403 (Wang et al., 2003) suggesting an important

role of superoxide anions in NMDA antagonist-induced

apoptosis and behavioral alterations. It has also been

postulated that the excitotoxic effects of glutamate are

largely mediated by increased Ca2+ influx through acti-

vated NMDA receptors (Garthwaite and Garthwaite,

1986; Choi, 1987; Luetjens et al., 2000).

Although apoptosis may be the final result of an

excitotoxic insult in rats, the pathways leading from

mitochondrial dysfunction and ROS generation to apop-

tosis are not completely understood. The use of metallo-

porphyrins such as manganese tetrakis (4-benzoyl acid)

porphyrin has implicated O2−• in glutamate excitotoxicity

(Luetjens et al., 2000; Patel et al., 1996). However, rela-

tive to the selective nonpeptidyl superoxide dismutase

mimetic M40403, metalloporphyrins lack specificity for

O2−• (Patel et al., 1996; Salvemini et al., 1999) and are

much less potent. Recently, it has been demonstrated

(McInnis et al., 2002; Wang et al., 2004) that M40403

significantly blunts NMDA antagonist-induced cell death

over about a 10-fold range (0.3–2.5 µM), whereas the

maximal effect of manganese tetrakis (4-benzoyl acid)

porphyrin was observed at much higher concentrations of

150–200 µM (Patel et al., 1996). Thus, M40403 appears

to be a valuable tool for exploring the specific role of

O2−• in NMDA antagonist-induced apoptosis.

Our working hypotheses (Figs 1 and 2) are that:

perinatal ketamine administration produces dose-related

neurodegeneration; the exposure of developing brains to

ketamine causes selective cell death by a mechanism in

which a compensatory up-regulation of NMDA receptor

subunits is involved; and ketamine-induced neurodegenera-

tion is associated with a calcium overload via glutamater-

gic stimulation of the upregulated NMDA receptors by

endogenous glutamate that exceeds the buffering capa-

city of mitochondria and interferes with electron transport

such that reactive oxygen species (ROS) are produced.

Application of Rodent and NonhumanPrimate In Vitro and In Vivo Models inStudies of Systems Biology

Glutamate promotes certain aspects of neuronal develop-

ment, including migration, differentiation and plasticity.

During development in the rat, especially during postnatal

days 7–14, the CNS exhibits enhanced susceptibility to

the toxic effects of modulation of the NMDA receptor

system. This enhanced susceptibility has been suggested

to derive from the increased expression of specific



DOI: 10.1002/jat

Figure 1. Working model of the effects of NMDA antagonists. This figure shows how exposure of developingbrain cells to NMDA antagonists (such as ketamine or PCP) might cause a compensatory upregulation of NMDAreceptors, making cells bearing these receptors more vulnerable to the excitotoxic effects of glutamate (Slikkeret al., 2005). This figure is available in colour online at www.interscience.wiley.com/journal/jat

NMDA receptor subunits (Miyamoto et al., 2001). Dur-

ing this period, NMDA receptors are the primary media-

tors of glutamatergic fast excitatory neurotransmission in

the brain. Because of the critical role of the NMDA

receptor system in brain development, antagonism of this

system can have profound, long-lasting and detrimental

effects (Behar et al., 1999). If the stimulation of gluta-

mate release reinforces neuronal connections, then the

blockade of that stimulation by NMDA antagonists may

result in fewer and/or nonfunctional connections.

Previous in vivo studies have demonstrated that peri-

natal PCP administration results in profound behavioral

abnormalities in the adolescent rat that may be related to

enhanced apoptotic cell death of neurons in the frontal

cortex (Wang et al., 2001; 2003). These cortical deficits

may have a significant impact on the function of sub-

cortical structures such as the nucleus accumbens, which

serves as an important regulatory center by integrating

the functions of the basal ganglia and the limbic system.

The nucleus accumbens receives glutamatergic afferents

from several brain regions, in particular, the frontal cor-

tex (Groenewegen et al., 1991; Albin et al., 1992; Zahm

and Brog, 1992; O’Donnell and Grace, 1995). Medium

spiny neurons account for the majority of the neostriatal

cell population and represent a major synaptic target of

dopaminergic input to the striatum (Groves, 1983; Smith

and Bolam, 1990; Graybiel, 1990; Kotter, 1994). Thus,

alteration of the cortical input (Fig. 3) to these neurons

during development may play a significant role in medi-

ating the behavioral effects of perinatal ketamine/PCP

treatment later in life.

Ketamine is a dissociative anesthetic agent that is

widely used in pediatric medicine. While it is clear that

ketamine causes neurotoxicity in the rodent model when

given repeatedly during the brain growth-spurt period,

it is not yet known whether similar phenomena also

occur in primates. Thus, the issue of whether the neuro-

toxicity observed in young rodents has scientific and/or

regulatory relevance for the pediatric use of ketamine

relies heavily upon the confirmation of these findings in

an appropriate primate model (Haberny et al., 2002;

Wang et al., 2006). The similarity of the physiology,

pharmacology, metabolism and reproductive systems of

the nonhuman primate to that of the human, especially



DOI: 10.1002/jat

during pregnancy, make the monkey an exceptionally

good animal model for use in systems biology. No other

commonly used research animal has a functional fetal

placental unit, a hemochorial placenta, a propensity for

single births and a fetal-to-maternal weight ratio compa-

rable to that of humans. Because the brain growth spurt

period in humans and nonhuman primates extends over a

much longer period than in the rat (Fig. 4), it is difficult

to match between species the exact stage of development

during which a particular neurodevelopmental event

occurs. Clearly, matching such events between human

and nonhuman primates is much less problematic. In

addition, it is known from a variety of studies (e.g. Paule,

2001) that the monkey is an excellent model for predict-

ing the acute and chronic behavioral (functional) effects

of a variety of drugs in humans. A study on the effects

of ketamine (or other anesthetic) exposure during the

Figure 2. Proposed toxic mechanisms of NMDA antagonists. This figure depicts a working model of NMDAantagonist-induced neuroapoptosis. Prolonged activation of NMDA receptors results in an overload of intracellularCa2+ that exceeds the buffering capacity of the mitochondrion and interferes with electron transport in a mannerthat results in the production of excess superoxide anion (O2

−•). The increase in O2−• turns on I-κB kinases, resulting

in the phosphorylation of I-κB serine residues, the dissociation of NF-κB proteins and their transport into thenucleus. In the nucleus, these transcription factors bind to several known genes including p53 and Bcl-XL. The con-sequences of this binding are not completely certain, but the transcription of p53 and subsequent increase in Baxis observed in several systems. With decreased Bcl-XL, increased Bax diminishes the formation of antiapoptotic Bax/Bcl-XL heterodimers in favor of pro-apoptotic Bax/Bax homodimers. These homodimers are thought to alter thepermeability of the mitochondrial membrane through which cytochrome c can leak into the cytoplasm, where itcan activate caspases that play a critical role in the ultimate demise of the cell. Several reactive oxygen species,including nitric oxide and superoxide anion (O2

−•), have been implicated in glutamate-induced neuronal death.However, little is known about the signaling pathway that mediates the postulated roles of peroxynitrite (ONOO.).Virtually any protein containing one or more tyrosyl residues can undergo nitrosation in vitro. Thus, it is importantto determine whether nitrosation of a protein alters its biological function. Much attention and debate has beendevoted to determining the precise reactive nitrogen species responsible for protein nitrosation (not the focus ofthis review). Further identification of the specific targets of endogenous nitrosation will be essential in providingclues about the associated pathological mechanisms (Slikker et al., 2005). This figure is available in colour online atwww.interscience.wiley.com/journal/jat

brain growth spurt in a monkey model would be the

most appropriate for determining whether the significant

neurodegeneration observed in the rodent model also

occurs in primates.

Both in vitro and in vivo approaches have been used

to assess the neurotoxicity associated with a wide range

of drugs at a variety of doses and exposure durations.

In vitro systems (primary culture and organotypic slice

culture) that parallel our in vivo studies were investigated

into a systems biology approach for studying the effects

of drug exposure on the developing nonhuman primate

model. Primary frontal cortical culture systems and

organotypic slice cultures, established using tissues from

rhesus monkeys, provide parallel in vitro models that

assist in evaluating the neurotoxicity of a variety of

various anesthetics at a variety of doses using a minimal

number of animals in a short period of time. Briefly,



DOI: 10.1002/jat

under deep anesthesia neonatal rhesus monkeys were

euthanized and brain tissue was collected. Several brain

areas, including the frontal cortex, were either sectioned

at 400 µm or dissociated in Ca2+- and Mg2+-free Hank’s

solution in the presence of a digestive enzyme. The brain

slices or dissociated neural cells were maintained in

culture for 5–7 days on a porous and translucent mem-

brane or on polylysine-coated dishes covered with culture

medium. For the assessment of neurotoxicity, a cytotoxi-

city detection assay (lactate dehydrogenase release, LDH),

Figure 3. Representative immunocytochemical picture of nitrotyrosine in the frontal cortex from saline- and PCP-treated rats. Protein tyrosine nitration occurs in many neurodegenerative states and is an important marker ofoxidative stress induced by peroxynitrite and possible other nitric oxide-derived oxidants. Intense nitrotyrosineimmunostaining is observed in layers II, III and IV neural cells in the PCP-treated rat (B) compared with the saline-treated rat (A). This figure is available in colour online at www.interscience.wiley.com/journal/jat

Figure 4. Periods of vulnerability to the neurotoxic effects of NMDA receptor antagonists for rats (postulated forrhesus monkeys and humans). There are two mechanisms of neurodegeneration associated with NMDA receptorantagonism: apoptosis and excitotoxicity. NMDA receptor antagonism triggers only apoptotic neurodegenerationin the rat model during synaptogenesis, a critical stage in development when synaptic connections are beingformed that begins at birth and includes the first two postnatal weeks of life (Ikonomidou et al., 1999). It is postu-lated that rhesus monkeys and humans would also be susceptible to this type of neurodegeneration duringsynaptogenesis, which spans from the third month of gestation into the second month of life for rhesus monkeys(Zecevic et al., 1989), and the third trimester of gestation into the third year of life for humans (Huttenlocher,1979). NMDA receptor antagonism triggers only excitotoxic neurodegeneration in the rat model late in adoles-cence and beyond (Farber et al., 1995), and it is postulated that rhesus monkeys and humans would be susceptibleto this type of neurodegeneration following puberty (Wright et al., 2006). This figure is available in colour onlineat www.interscience.wiley.com/journal/jat



DOI: 10.1002/jat

TUNEL (terminal dUTP nick-end labeling)-staining,

fragmented DNA detection by ELISA, and immunocyto-

chemical staining using markers of neurodegeneration

were performed according to standard laboratory protocols.

These in vitro preparations are useful for the rapid

evaluation of the neurotoxic effects of anesthetic drugs

and enable a direct study of the brain at various stages of

development. Primary (Fig. 5) and organotypic (Fig. 6)

cultures were used because these preparations maintain

important anatomical relationships and synaptic con-

nectivities, allow for the direct assessment of cell death

and are reliable models for screening and evaluating the

neurotoxicity of different anesthetic drugs. In addition,

these preparations allow for the direct application of anti-

sense oligodeoxynucleotides (ODN) that target specific

receptor genes, as well as direct enzymatic and therapeu-

tic drug treatment. This approach allows for the collec-

tion of a large amount of data from a minimal number of

subjects and allows for the investigation of cellular

mechanisms associated with ketamine-induced cell death

in a simplified primate system.

Application of a Systems BiologyApproach in Toxicological Studies ofAnesthetic-induced Neurodegeneration

The goals of the current studies on ketamine were to

determine: (1) if ketamine administration results in a

dose-related increase in neurotoxicity; (2) if ketamine

increases NMDA receptor protein expression; (3) if

NMDA receptor NR1 subunit antisense oligonucleotide

or the blockade of NF-kB translocation will reverse or

prevent ketamine-induced apoptosis, and, if so; (4) what

is its association with alterations in steady-state levels of

the Bcl-2 family of proteins?

For the in vitro experiments, 1–20 µM ketamine was

added for 24 h (on the fifth day of culture) and then

washed out with serum and glutamate-containing me-

dium, before experimental assays for cell death (ELISA,

MTT assay, LDH release and immunoblotting) were per-

formed 24 h later. Based on a Ki of 440 nM, 0.1, 1.0, 10.0

and 20.0 µM ketamine would occupy approximately 10%,

25%, 60% and 95% of the available NMDA receptors,

respectively (Wang et al., 2005b). These concentrations

were also based on clinically relevant doses, the IC50

concentration and our preliminary studies.

Application of Systems Biology Approaches atthe mRNA/DNA Level (Genomics)

Microarrays

To determine which of several candidate death genes

might be associated with anesthetic drug-induced

Figure 5. Immunofluorescence micrographs of primarymonkey frontal cortical cultures. (A) Neuron-specificstaining of cultured cells with PSA-NCAM as revealedby immunofluorescence of anti-mouse IgG conjugatedto fluorescein isothiocyanate. (B) Glia-specific stainingof cultured cells with GFAP as revealed by immuno-fluorescence of anti-rabbit IgG conjugated to rhoda-mine. (C) Hoechst 33285 nuclear staining reveals thetotal number (nuclei) of the cells in the field. Scale bar= 50 µm (Wang et al., 2006). This figure is available incolour online at www.interscience.wiley.com/journal/jat

apoptosis, gene microarray techniques were utilized. The

goal here was to determine the relationship between

anesthetic drug-induced neurotoxicity and anesthetic

drug-induced changes in gene expression. Therefore,



DOI: 10.1002/jat

gene expression profiles for different anesthetic drugs at

different time points were compared.

Tissue (in vivo and in vitro). Frozen tissue was sec-

tioned using a cryostat prior to laser capture

microdissection (LCM). Briefly, the tissue was embedded

in a mold containing optical cutting temperature (OCT)

compound. This tissue-containing cryomold was allowed

to equilibrate on dry ice and was then frozen onto the

cutting stage of the cryostat. Approximately 7–10 µm

sections were cut and immediately adhered to chilled mi-

croscope slides which were kept on dry ice until ‘same-

day’ processing for LCM or stored at −80 °C for future

processing. Keeping the tissue frozen prior to LCM

ensures that high quality mRNA is obtained from the

cells collected. Cells from the same brain region were

collected from each subject to be used as part of the sys-

tems biology study. Total RNA was isolated from these

cells and an RNA amplification kit was used to generate

labeled cRNA for microarray experiments. Briefly,

double-stranded (ds) cDNA was synthesized from the

isolated RNA prior to in vitro transcription to generate

antisense RNA, which was then used in a second round

of amplification to generate additional ds cDNA. This

ds cDNA was amplified along with cyanine 3-CTP or

cyanine 5-CTP dye in order to generate labeled antisense

cRNA.

Figure 6. Organotypic cultures prepared from 7-day-old rat pups. The brains were sectioned down the midlineand corticostriatal slices containing the anterior commissure were cut at a thickness of 400 µm. The slices weremaintained in culture for about 5–10 days on a porous and translucent membrane at the interface between themedium and CO2-enriched atmosphere. To characterize this model, (A) monoclonal anti-polysialic acid neural celladhesion molecule (neuronal specific marker) and (B) polyclonal anti- NCAM antibodies were used for theimmunostaining. To demonstrate that neurons in organotypic culture were functional, whole cell patch clamprecording was performed. The figure shows representative sodium current spikes that demonstrate the viability ofneurons in an organotypic culture. The sodium current spikes were evoked by applying a depolarizing voltagewhen the neurons were held at −60 mV. This figure is available in colour online at www.interscience.wiley.com/journal/jat

Oligonucleotide Microarrays (in vivo and in vitro). The

labeled RNA, along with labeled reference RNA, was

co-hybridized onto a rat oligo microarray containing over

22 000 sequences. The microarray slides were scanned

and the data analyzed using ArrayTrack in-house soft-

ware (Fig. 7).

An in vivo Example. Repeated administration of ketamine

produces neuronal apoptosis in neonatal rats (Ikonimdou

et al., 1999). In order to determine whether a single dose

of ketamine also produces apoptosis, 7-day-old rats were

injected subcutaneously with either water or 40 mg kg−1

ketamine. Brain tissue was harvested 1, 2 and 4 h after

the injection in order to identify significant gene expres-

sion changes in the thalamus that might be related to

acute ketamine exposure. LCM was used to collect

approximately 500 cells from the thalamus and the

microarray procedures used were identical to those

described above. Microarray analyses identified 18 sig-

nificant (fold change >1.5 and P < 0.05) gene expression

changes in the thalamus 1 h after treatment, 624 signifi-

cant gene expression changes 2 h after treatment (Fig. 8),

and 52 significant gene expression changes 4 h after treat-

ment. Of these, several genes specifically associated with

apoptosis (CYCS, ELL, PDCD8, PRKCB1 and RIPK1)

were found to be up-regulated in the thalamus following

a single injection of ketamine.



DOI: 10.1002/jat

Figure 7. Schematic of microarray procedures used in the systems biology approach (Wright et al., 2006). To deter-mine which of several candidate death genes are associated with anesthetic drug-induced apoptosis, microarraytechniques were applied to the anesthetic drug models. Laser capture microdissection was used to collect cellsfrom a specific brain region and total RNA was isolated from these cells. The first round of RNA amplification wasperformed in order to generate double-stranded cDNA. The cDNA was then amplified along with cyanine 3-CTP orcyanine 5-CTP dye in order to generate labeled antisense cRNA. The labeled antisense cRNA was hybridized onto arat oligo microarray containing over 22 000 sequences which was then scanned and the data were analysed usingArrayTrack software (NCTR). This figure is available in colour online at www.interscience.wiley.com/journal/jat

Detection of DNA Fragmentation by GelElectrophoresis (in vivo and in vitro)

Although the focus of the present work is not to abso-

lutely identify fragmented DNA as intranucleosomal, as

thought to be characteristic of apoptosis, this method may

be used later if deemed necessary. Briefly, brain regions

shown previously to be TUNEL (terminal dUTP nick-end

labeling-Assay; this assay labels broken DNA strands

often associated with apoptopic cell death)-positive will

be dissected or punched from coronal sections and then

homogenized in a DNA extraction buffer (Portera-

Cailliau et al., 1997). DNA is then phenol-chloroform

extracted, ethanol precipitated and resuspended. After

removal of RNA by RNase, the solution is incubated

overnight in a proteinase solution. DNA is then re-

extracted, precipitated, resuspended and approximately

20 µg is fractionated on an agarose gel, stained with

ethidium bromide and photographed under UV tran-

sillumination. DNA laddering in multiples of 180–200

bases is indicative of apoptosis, while a ‘smear’ of DNA

fragments of various lengths is indicative of necrosis

(Portera-Cailliau et al., 1997).

In Situ Hybridization (in vivo and in vitro)

Of particular interest are the possible mechanisms by

which ketamine might up-regulate NMDA receptors.

Surprisingly, there is little literature concerning this issue,

but it has recently been demonstrated that the distal

region of the NR1 promoter contains an active NF-kB

site and it is developmentally regulated (Liu et al., 2004).

The NMDA receptor NR1 subunit is widely distributed

throughout the brain and is the fundamental subunit

necessary for NMDA channel function. To determine

whether altered regulation of NMDA receptor subunits

promotes ketamine-induced cell death, in situ hybridiza-

tion detecting the relative densities of NMDA receptor

NR1, NR2A and NR2B subunits following anesthetic

drug administration has been performed.



DOI: 10.1002/jat

deoxy-nucleotidyl transferase. The specificity of this

probe has been previously described (Moriyoshi et al.,

1991; Monyer et al., 1992; Mori and Mishina, 1995).

Longitudinal sections (10 µM) through whole brain or

coronal sections through frontal cortical levels were cut

with a cryostat and processed for in situ hybridization

as described previously (Bartanusz et al., 1993). Analy-

sis of in situ hybridization autoradiographs was accom-

plished using hematoxylin-eosin-counterstained sections

and the negative control was prepared by adding an

excess amount (50-fold) of unlabeled probe.

For the quantification of in situ autoradiographs,

images were acquired via digital microscopy. Briefly, the

images were smoothed with a square filter to remove

noise and regions of interest (ROI) were selected using a

threshold technique that segmented the image into labeled

cells and background. The threshold was determined inter-

actively by the consensus of two trained observers (one

of whom was blind to the treatment) and then held con-

stant for all ROI within the section. The density of silver

grains associated with individual neurons was estimated

by measuring the area within a 20 µm fixed diameter

circle placed over individual neurons that exceeded the

threshold value. Background labeling was determined in

a similar fashion and subtracted from each measurement

to estimate specific labeling in each ROI. A monotonic

relationship was assumed to exist between measured

labeling and the amount of mRNA labeled with the

radioactive probe. This technique is similar to that used

by several laboratories except that a fixed size (314 µm)

rather than a variable size ROI was used (Standaert et al.,

1996; Rudolf et al., 1996).

This study demonstrated that ketamine produces sig-

nificant up-regulation in NMDA receptor NR1 subunit

mRNA expression (Fig. 9). This result is consistent with

the literature demonstrating that treatment with NMDA

antagonists produces up-regulation of the NMDA receptor

complex as measured by an increase in the Bmax of

NMDA receptor binding sites (McDonald et al., 1990;

Figure 8. Volcano plot illustrating the significant geneexpression changes in the thalamus 2 h after ketaminetreatment. Using a criterion of P < 0.05 and a fold-change threshold of 1.5 produced 624 significant geneexpression changes. All 22 000 genes contained on themicroarray are featured on the volcano plot. Featuresin red indicate ‘fully significant’ genes with P < 0.05and fold-change >1.5, features in pink indicate ‘signifi-cant P-value only’ genes with P < 0.05 and fold-change<1.5, features in yellow indicate ‘significant fold-change only’ genes with P > 0.05 and fold-change>1.5, and features in black indicate ‘non-significant’genes with P > 0.05 and fold-change <1.5. This figureis available in colour online at www.interscience.wiley.com/journal/jat

Figure 9. NR1 subunit abundance in rat brain. In situ hybridization was performed on rat brain sections (longitu-dinal and coronal) using a 35S-labeled oligonucleotide probe specific for the NMDA receptor NR1 subunit. Thefigure shows that at the frontal cortical level of a control brain, NMDA receptor NR1 subunit mRNA was promi-nent. The autoradiograph grain density for NR1 subunit mRNA was up-regulated in the ketamine-treated brain

An oligonucleotide probe complementary to the mRNA

encoding the NMDA receptor NR1 subunit was selected

on the basis of cloned cDNA sequences. The sequence of

the probe used for in situ hybridization is as follows: 5′-TTCCTCCTCCTCCTCACTGTTCACCTTGAATCG-

GCCAAAGGGACT (this corresponds to a region that is

constant across all NR1 splice variants). It was 3′ end

labeled by incubation with [35S]deoxy-ATP and terminal



DOI: 10.1002/jat

Williams et al., 1992). Chronic treatment with ethanol,

another noncompetitive NMDA antagonist (Lovinger

et al., 1989), has also been shown to up-regulate NMDA

receptor number and function both in vitro and in vivo

(Grant et al., 1990; Trevisan et al., 1994; Hu et al.,

1996). In addition, chronic ethanol stabilizes NR1 mRNA

in fetal cortical neurons in culture (Kumari and Ticku,

1998).

The ability of ketamine to enhance neurodegeneration

may be the result of a compensatory ketamine-induced

up-regulation of NMDA receptors. This up-regulation

makes neurons bearing these receptors more vulnerable,

after ketamine washout, to the excitotoxic effects of

endogenous glutamate. In fact, this hypothesis was sup-

ported by the observation that, in cells kept in a defined

serum-free media with no or very low concentrations of

glutamate for 24 h after ketamine washout, no significant

effects of ketamine were noted (as measured by an MTT

assay): 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium

bromide-assay; the MTT dye is metabolized by viable

mitochondria to a colored product that is detected

photometrically. The assay is an important indicator for

cell survival as well as mitochondrial function.

Application of Antisense Oligonucleotides

Previous experiments demonstrated that antisense oligo-

nucleotides specific for NMDA receptors were able to

prevent receptor up-regulation in response to physiolo-

gical or pharmacological challenges (Wahlestedt et al.,

1993; Roberts et al., 1998; Lai et al., 2000; Wang et al.,

2005a; 2005b). A similar strategy was used to test the

hypothesis that the up-regulation of the NR1 subunit was

necessary for ketamine-induced cell death. Antisense

technology can be used to block the function of specific

genes in neural cells. Antisense RNA is complementary

to the normally expressed RNA and presumably inhibits

the expression of normal RNA strands. When mRNA

forms a duplex with a complementary antisense RNA

sequence, translation is blocked because duplex RNA is

quickly degraded by ribonucleases in the neural cells.

Therefore, antisense oligonucleotides can provide funda-

mental data to help understand the mechanisms that

underlie ketamine-induced cell death.

To determine whether the administration of antisense

oligonucleotides targeted to specific NMDA receptor sub-

units can block the up-regulation of the NMDA receptor

NR1 subunit mRNAs and proteins, antisense oligo-

nucleotides targeted to the NR1 subunit were used in

forebrain cultures. The experimental protocol was as

follows. Control cultures were maintained in normal

medium for 7 days. Experimental cultures were main-

tained in normal medium for 5 days, then treated for 24 h

with ketamine (1, 10 or 20 µM) alone, ketamine (10 µM)

plus 2 µM antisense oligonucleotide for NR1, or ketamine

(10 µM) plus sense oligonucleotide for NR1 or antisense

alone. The experimental cultures were then washed

with serum-containing media to remove ketamine and

maintained for another 24 h. All cells were harvested on

day 7.

Based on search results from GeneCards™, the human

NMDA receptor NR1 gene has 28919 bases. Comple-

mentary DNA encoding the key subunit of the human

NMDA receptor (NR1) was isolated from a human brain

cDNA library. The human NR1 (hNR1) cDNA encodes

an open reading frame of approximately 2.7 kb that

shares high homology with the rat NR1 subunit and the

mouse zeta 1 subunit (Planells-Cases et al., 1993; Karp

et al., 1993). Therefore, the cDNA clone hNR1 codes for

a human brain NMDA receptor subunit cognate to the

rodent and murine brain NR1 subunits. The gene for the

NR1 subunit consists of 21 exons and can be alterna-

tively spliced, producing transcript variants differing

in the C-terminus. Cell-specific factors are thought to

control the expression of these different isoforms, which

possibly contributes to its functional diversity. An 18-mer

antisense oligonucleotide corresponds to nucleotides 4–21

(5′-CAGCAGGTGCAT-GGTGCT) of the NR1 subunit

mRNA, which directly follows the translation initiation

codon. The sequence of the antisense oligonucleotide was

chosen to target the 5′ coding region of the NR1 mRNA,

and this sequence has proven effective and specific in

previous studies (Wahlestedt et al., 1993; Roberts et al.,

1998; Lai et al., 2000; Wang et al., 2005a; 2005b).

The MTT assay was used here to monitor the amount

of total cell loss (both necrotic and apoptotic). Figure 10

Figure 10. Quantitative analyses of ketamine-inducedneurotoxicity and the protective effect of NR1 anti-sense oligonucleotide as assessed using MTT assays.Each condition was assessed at least in triplicate andexperiments were independently repeated three times.Data are presented as mean ± SD. A probability of* P < 0.05 was considered significant (one-way ANOVAwith Holm-Sidak test). K, ketamine; NR1-AS, NR1 anti-sense oligonucleotides; NR1-SE, NR1 sense oligonucleo-tides (Wang et al., 2006)



DOI: 10.1002/jat

shows that ketamine at 10 and 20 µM resulted in a sub-

stantial decrease (~75%) in mitochondrial metabolism

of MTT. Co-administration of 10 µM ketamine with

NR1 antisense oligonucleotide also blocked the reduc-

tion of PSA-NCAM expression induced by ketamine (not

shown).

Nuclear Factor-kB (NF-kB) Signaling in Ketamine-induced Neuronal Cell Death (ElectrophoreticMobility Shift Assay)

The transcription factor NF-kB is known to respond to

changes in the redox state of the cytoplasm and has been

shown to translocate into the nucleus in response to

NMDA-induced cellular stress (Ko et al., 1998). NF-kB

is normally sequestered in the cytoplasm, bound to the

regulatory protein IkB. In response to a wide range of

stimuli, including oxidative stress, IkB is phosphorylated

on serine residues by the enzyme IkB kinase. The net

result is the release of the NF-kB dimer, which is then

free to translocate into the nucleus.

The possible relationship between ketamine-induced

cell death and potential intracellular mediators (NF-kB

translocation) was examined using an electrophoretic

mobility shift assay. Nuclear protein extracts from control

and treated cultures were tested in this assay using a 32P-

labeled oligonucleotide containing the classical NF-kB

consensus sequence. In this study, cultures were exposed

to ketamine for 24 h, and then ketamine was washed out

and replaced with serum-containing media for another

24 h. SN-50, an inhibitor of NF-kB translocation (2.5 µM)

and SN-50-Control [2.5 µM; (SN-50-Control has a chem-

ical and molecular structure similar to SN-50 but its

functional residues are replaced)] were co-administered

with ketamine. The medium was then removed, and the

attached cells were washed with PBS. Nuclear extracts

were prepared according to published methods (Dignam

et al., 1983; Osborn et al., 1989) with some modifica-

tions. Briefly, cells were homogenized in buffer and the

lysate was microcentrifuged to collect nuclei. Nuclear

protein was extracted by suspending the nuclei in extrac-

tion buffer, subjecting them to centrifugation and divid-

ing the supernatant into aliquots.

Double-stranded DNA containing the sequence corre-

sponding to the classical NF-kB consensus site was end-

labeled with [γ-32P]ATP. Unincorporated nucleotides

were removed and binding reactions were effected. Gels

were dried and subjected to autoradiography and the

radiographs were analysed using an automatic image

analysis system.

NF-kB translocation appears to be a necessary step in

the cell death induced by PCP (McInnis et al., 2002),

cyanide and excitotoxic stimuli (Shou et al., 2000). In

the present example, ketamine produced a remarkable

increase in translocation of NF-kB into the nucleus.

This result is consistent with our previous report (Wang

et al., 2001) that perinatal phencyclidine (PCP; non-

competitive NMDA receptor antagonist) administration

results in enhanced cortical apoptosis, increased nuclear

translocation of NF-kB and a reduced Bcl-XL/Bax

ratio. The protection against ketamine-induced cortical

neuron cell death and the decreased PSA-NCAM im-

munostaining noted in the presence of SN-50 suggests

that there is a causal relationship between these events.

There is evidence in the literature suggesting that the

transcriptional regulation of target genes by NF-kB

can be tissue-specific and possibly gene-specific within

a given cell type. The ability of SN-50 to prevent

ketamine-induced cell death demonstrates that NF-

kB is crucial to these processes. However, it remains

to be determined whether ketamine-induced NF-kB

translocation is specifically responsible for activation

of apoptotic or necrotic pathways – or both – leading

to cell death. Figure 11 shows a representative gel in

which the nuclear protein from treated cultures retards

the migration of the labeled NF-kB binding sequence.

Ketamine (10 µM) caused an approximately 40% increase

in the density of this band relative to control (Fig. 11a,b).

SN-50 dose-dependently protected against the neuro-

toxic effect of ketamine as indicated by the MTT assay

(Fig. 11c).

Oxidative Stress and Mitochondrial DNA RepairSystem

Reactive oxygen species (ROS) are extremely reactive

and interact with DNA and other biomolecules. Living

cells have defensive systems that include antioxidant

molecules. When the balance between antioxidants and

oxidants is upset, ROS can generate lesions in the DNA

such as base modifications and chain breaks.

Mitochondria are cellular organelles responsible for

energy supply in the form of ATP. They are also essen-

tial for a variety of processes such as development, aging,

disease and cell death (either apoptotic or necrotic). Tem-

porary or sustained loss of mitochondrial function can

have a major impact on the integrity of cellular defenses

and repair processes and may result in a decreased cap-

acity to mount an appropriate stress response (Toescu

et al., 2000). Oxidative damage can cause formation of

8-hydroxyguanine, often termed 8-oxo-dG. The produc-

tion of 8-hydroxyguanine is almost exclusively elicited

by oxidative stress, with the main attack site by oxidative

radicals occurring at the N7-C8 bond. 8-Oxoguanine is a

modified guanine whose formation is induced by endo-

genous or exogenous ROS. 8-Oxoguanine is used as a

biomarker of oxidative DNA damage. DNA polymerases

preferentially insert adenine (instead of thymine) opposite

8-hydroxyguanine; therefore, in the absence of repair,

these oxidative damage adducts can lead to G to T tran-

sitions. Antioxidants act in vitro and in vivo to decrease

oxidative damage. The 8-hydroxyguanine lesion causes



DOI: 10.1002/jat

Figure 11. Inhibition of ketamine-induced nuclear translocation of NF-kB by SN-50 and NR1 antisense oligo-nucleotides in primary cell culture of PND3 (monkey) frontal cortex cells (Wang et al., 2006). Top panel (a) shows arepresentative electrophoretic mobility shift assay where each lane corresponds to the conditions shown in themiddle panel (b). The bottom panel (c) indicates that SN-50 protected against the neurotoxic effects of ketamineas indicated by the MTT assay in a dose-dependent manner. No significant protective effects were observed with0.3 or 0.6 µM but protection began at 1.25 µM and was maximal at 2.5 µM. Panel (b) and (c) show the data as mean± SD from three identical experiments. A probability of * P < 0.05 was considered significant (one-way ANOVAwith the Holm-Sidak test)

mutational frequencies of 1%–5% and is one of the most

abundant oxidative lesions (Halliwell, 2001). Apurinic

endonuclease (APE) and DNA glycosylase (OGG-1) are

two major classes of DNA repair enzymes that remove

bases with oxidative lesions.

To detect and localize oxidative DNA damage associated

with ketamine administration, double-immunofluorescence

analyses were performed using monoclonal anti-8-oxoguanine

and polyclonal anti-NCAM (neural cell adhesion mole-

cule) antibodies. Our in vitro studies with ketamine (pre-

liminary data) indicate that 8-oxoguanine is primarily

localized in neurons and glial cells (in nuclei as well as

mitochondria). Ketamine (10 µM) was found to remark-

ably up-regulate 8-oxoguanine expression within 6 h

after its removal.

Application of the Systems Biology Approach atthe Protein Level

Western Blot Analysis (In vitro and In vivo)

To determine the effect of co-administration of NMDA

receptor antisense oligonucleotides on anesthetic drug-

induced neurotoxicity and to determine whether the ad-

ministration of antisense oligonucleotides targeted to the

NR1 NMDA receptor subunit blocks steady-state protein

levels, an antisense oligonucleotide was used. To deter-

mine whether decreased PSA-NCAM expression is asso-

ciated with ketamine-induced loss of cortical neurons or

to local NMDA receptor blockade (as determined by

polysialyl transferase activity), Western blot analyses of



DOI: 10.1002/jat

PSA-NCAM/NR1 protein/Actin ratios and the protective

effects of the antisense oligonucleotides were examined.

Western blot analysis was used for quantification of

protein expression levels. Cells were lysed in buffer con-

taining a protease inhibitor cocktail. The homogenate was

centrifuged and the supernatant was collected for assay.

The protein concentrations were determined and equal

amounts of protein were loaded on each lane of a SDS-

polyacrylamide gel and then transferred to a polyviny-

lidene difluoride membrane. The blots were probed with

anti-PSA-NCAM (monoclonal), anti-NR1 (monoclonal)

and anti-actin (monoclonal). Immunoblot analyses were

performed with horseradish peroxidase-conjugated anti-

mouse and anti-rabbit IgG using enhanced chemilu-

minescence Western blotting detection reagents. The

bands corresponding to PSA-NCAM, NR1 and β-actin

were scanned and densitometrically analysed using

an automatic image analysis system. These quantitative

analyses were normalized to β-actin (after stripping) and

expressed as mean ± SEM. One-way ANOVAs were

used to compare levels of each protein among different

treatment groups.

Co-administration of NR1 antisense oligonucleotide

was able to almost completely block the neuronal cell

death induced by ketamine (Fig. 12). Our data indicate

that ketamine markedly up-regulated NMDA receptor

NR1 subunit protein. Co-administration of antisense

oligonucleotide specifically prevented NR1 up-regulation

and blocked the reduction of PSA-NCAM expression

induced by ketamine.

Immunocytochemistry for Staining Neurons andGlia (In vitro and In vivo)

A mouse monoclonal antibody to polysialic acid neural

cell adhesion molecule (PSA-NCAM) was used to iden-

tify neurons, and a rabbit polyclonal antibody to glial

fibrillary acidic protein (GFAP) was used to identify glial

cells in monkey frontal cortical cell cultures. An indirect

immunofluorescence technique was used to visualize

immunoreactivities. The cells were incubated with pri-

mary antibodies and the bound antibodies were revealed

using fluorescein isothiocyanate-conjugated sheep anti-

mouse IgG second antibody or rhodamine-conjugated

sheep anti-rabbit IgG second antibody. The cells

were examined with a light microscope equipped with

epifluorescence.

In control cultures PSA-NCAM immunoreactivity

(Fig. 13A) was intense on the surface of cell bodies and

processes of neurons. Figure 13B shows that ketamine

treatment dramatically diminished PSA-NCAM immunore-

activity. PSA-NCAM expression was weak and character-

ized by typical residue pieces, fragmentation and neuronal

shrinking. Figure 13C shows that ketamine treatment

resulted in a significant down-regulation of PSA-NCAM

and that application of SN-50 effectively protected

Figure 12. Western blot analysis of the effect ofketamine and NR1 antisense oligonucleotide on theregulation of NMDA receptor NR1 subunit protein andPSA-NCAM expression primary cell culture of PND3(monkey) frontal cortex cells (a). β -actin was used as aloading control and was not affected by ketamine. Thedata from three independent experiments were quan-tified using densitometry and expressed as the ratio ofPSA-NCAM to actin and NR1 to actin (b). Statisticalcomparisons consisted of a one-way ANOVA withthe Holm-Sidak test. A probability of * P < 0.05 wasconsidered significant (Wang et al., 2006)

neurons from ketamine-induced cell death: the inactive

control peptide for SN50 was ineffective. These data

indicate that ketamine-induced neurotoxicity in neonatal

monkey frontal cortical cultures manifests primarily as

neuronal cell loss.

Conclusion

This review article provides an overview of our efforts

to apply a systems biology approach to understand a

particular toxicological problem: ketamine-induced neuro-

degeneration in developing nonhuman primates and

rodents, using both in vitro and in vivo models. Systems

biology, as adopted for toxicology, is referred to as sys-

tems toxicology and involves the study of system pertur-

bations caused by chemicals or stressors. By monitoring

alterations in gene, protein and cell signal expression that

are linked firmly to toxicological outcomes, it is hoped to

define the affected system(s) in an integrative manner.



DOI: 10.1002/jat

Although not yet fully delineated, the working model for

NMDA antagonist (e.g. ketamine)-induced neurodegenera-

tion during development involves the modulation of nor-

mally occurring brain sculpting mechanisms that control

CNS development. Exposure of the developing mammal

to ketamine or other NMDA antagonists perturbs the

endogenous NMDA receptor system and results in

enhanced neuronal cell death (Haberny et al., 2002). Up-

regulation of the NR1 subunit of the NMDA receptor by

NMDA antagonists appears to play a critical role in the

subsequent cell death that can occur even in the absence

of the antagonist. Blockade of this NMDA receptor up-

regulation dramatically diminishes the cascade of events

that lead to cell death. Ketamine produces an increase

in NR1 subunit expression and co-administration of NR1

antisense oligonucleotides specifically prevents the syn-

thesis of NR1 protein and blocks the neuronal loss

induced by ketamine. Inhibition of NF-kB nuclear trans-

location by SN50 also prevents NMDA antagonist-

induced neuronal cell death. Iterative perturbations of the

model systems have allowed for refinement of the general

model and reinforced the selective pathways that repre-

sent the whole of the data.

Although many more studies are needed in order to

perfect a quantitative model, some general pathways have

been identified using carefully selected agents as directed

by a systems biology approach. Further elucidation of the

precise developmental stages at which susceptibility to

NMDA receptor antagonists induce neurodegeneration

will come with future experiments, as will important

dose-response information. Use of the same phenotypic

anchor (neuronal cell death) across experiments and

models will facilitate comparisons and generalizations

and eventually provide the opportunity for a thorough

description of the processes underlying the adverse events

described.

References

Albin R, Makowiec RL, Holligsworth ZR, Dure IVLS, Penny JB,Young AB. 1992. Excitatory amino acid binding sites in the basalganglia of the rat: quantitative autoradiographic study. Neuroscience

46: 135–148.Auffray C, Imbeaud S, Roux-Rouquie M, Hood L. 2003. From func-

tional genomics to systems biology: concepts and practices. C. R.

Biol. 326: 879–892.Bartanusz V, Jezova D, Bertini LT, Tilders FJH, Aubry JM, Kiss JZ.

1993. Stress-induced increase in vasopressin and corticotrophin-releasing factor expression in hypophysiotrophic paraventricularneurons. Endocrinology 132: 895–902.

Beal MF. 1992. Mechanisms of excitotoxicity in neurologic diseases.FASEB J. 6: 3338–3344.

Behar TN, Scott CA, Greene LC, Wen X, Smith SV, Maric D, Liu Q-Y, Colton CA, Baker IJ. 1999. Glutamate acting at NMDA receptorsstimulates embryonic cortical neuronal migration. J. Neurosci. 19:4449–4461.

Choi DW. 1987. Ionic dependence of glutamate neurotoxicity.J. Neurosci. 7: 369–379.

Choi DW. 1988. Glutamate neurotoxicity and diseases of the nervoussystem. Neuron 1: 623–634.

Collingridge GL, Kehl SJ, McLennan H. 1983. Excitatory aminoacids in synaptic transmission in the Schaffer collateral-commissuralpathway of the rat hippocampus. J. Physiol. 334: 33–46.

Figure 13. Effect of ketamine and SN-50 on the decrease in PSA-NCAM expression in monkey frontal corticalcultures (Wang et al., 2006). PSA-NCAM immunoreactivity was intense in the control culture (A) and diminished inthe ketamine-treated culture (B). Scale bar, 50 µm. Densitometry measurements were used to calculate a ratio ofPSA-NCAM to actin in each lane for each of three independent experiments and the data are shown as the mean± SD of those ratios (C). SN-50 (2.5 µM) effectively prevented the reduction of PSA-NCAM induced by ketamine.No protective effect was observed from the inactive control peptide for SN-50 (2.5 µM). This figure is available incolour online at www.interscience.wiley.com/journal/jat



DOI: 10.1002/jat

Cotman CW, Geddes JW, Bridges RJ, Monaghan DT. 1989. N-methyl-D-aspartate receptors and Alzheimer’s disease. Neurobiol. Aging 10:603–605.

Dignam JD, Lebovitz RM, Roeder RG. 1983. Accurate transcriptioninitiation by RNA polymerase II in a soluble extract from isolatedmammalian nuclei. Nucleic Acids Res. 11: 1475–1489.

D’Souza SW, McConnell SE, Slater P, Barson AJ. 1993. Glycinesite of the excitatory amino acid N-methyl-D-aspartate receptor inneonate and adult brain. Arch. Dis. Child. 69: 212–215.

Durand GM, Gregor P, Zheng X, Bennett MV, Uhl GR, Zukin RS.1992. Cloning of an apparent splice variant of the rat NMDAreceptor NMDAR1 with altered sensitivity to polyamines andactivators of protein kinase C. Proc. Natl Acad. Sci. USA 89: 9359–9363.

Farber NB, Wozniak DF, Price MT, Labruyere J, Huss J, St. Peter H,Olney JW. 1995. Age-specific neurotoxicity in the rat associated withNMDA receptor blockade: potential relevance to schizophrenia? Biol.

Psychiatry 38: 783–787.Garthwaite G, Garthwaite J. 1986. Neurotoxicity of excitatory amino

acid receptor agonists in rat cerebellar slices: dependence on calciumconcentration. Neurosci. Lett. 66: 93–198.

Grant KA, Valverius P, Hudspith M, Tabakoff B. 1990. Ethanol with-drawal seizures and the NMDA complex. Eur. J. Pharmacol. 176:289–296.

Graybiel AM. 1990. Neurotransmitters and neuromodulators in the basalganglia. Trends Neurosci. 13: 244–254.

Greenamyre JT. 1993. Glutamate-dopamine interactions in the basalganglia: Relationship to Parkinson’s disease. J. Neural Transm. 91:255–269.

Groenewegen HJ, Berendse HW, Meredith GE, Haber SN, Voorn P,Wotters JG, Lohman AHM. 1991. Functional anatomy of the ventrallimbic system-innervated striatum. In The Mesolimbic Dopamine

System: From Motivation to Action, Willner P, Scheel-Kruger J (eds).John Wiley & Sons: London, 19–59.

Groves PM. 1983. A theory of the functional organization of theneostriatum and the neostriatal control of voluntary movement. Brain

Res. 286: 109–132.Haberny KA, Paule MG, Scallet AC, Sistare FD, Lester DS, Hanig JP,

Slikker W Jr. 2002. Ontogeny of the N-methyl-D-aspartate (NMDA)receptor system and susceptibility of neurotoxicity. Toxicol. Sci. 68:9–17.

Halliwell B. 2001. Role of free radicals in the neurodegenerativediseases: therapeutic implications for antioxidant treatment. Drugs

Aging 18: 685–716.Hanania T, Hillman GR, Johnson KM. 1999. Augmentation of

locomotor activity by chronic phencyclidine is associated with anincrease in striatal NMDA receptor function and an upregulation ofthe NR1 receptor subunit. Synapse 31: 229–239.

Henry CJ. 2003. Evolution of toxicology for risk assessment. Int. J.

Toxicol. 22: 3–7.Hu X-J, Follesa P, Ticku MK. 1996. Chronic ethanol treatment pro-

duces a selective upregulation of the NMDA receptor subunit geneexpression in mammalian cultured cortical neurons. Brain Res. Mol.

Brain Res. 36: 211–218.Huttenlocher PR. 1979. Synaptic density in human frontal cortex-

developmental changes and effects of aging. Brain Res. 163: 195–205.

Ikonomidou C, Bosch F, Miksa M, Bittigau P, Vockler J, Dikranian K,Tenkova TI, Stefovska V, Turski L, Olney JW. 1999. Blockade ofNMDA receptors and apoptotic neurodegeneration in the developingbrain. Science 283: 70–74.

Ishii T, Moriyoshi K, Sugihara H, Sakurada K, Kadotani H, Yokoi M,Akazawa C, Shigemoto R, Mizuno N, Masu M, Nakanishi S. 1993.Molecular characterization of the family of the N-methyl-D-aspartatereceptor subunits. J. Biol. Chem. 268: 2836–2843.

Jazwinski SM. 2002. Biological aging research today: potential, peeves,and problems. Exp. Gerontol. 37: 1141–1146.

Jevtovic-Todorovic V, Hartman RE, Izumi Y, Benshoff ND, DikranianK, Zorumski CF, Olney JW, Wozniak DF. 2003. Early exposure tocommon anesthetic agents causes widespread neurodegeneration inthe developing rat brain and persistent learning deficits. J. Neurosci.

23: 876–882.Johnson KM, Phillips M, Wang C, Kevetter GA. 1998. Chronic

phencyclidine induces behavioral sensitization and apoptotic cell

death in the olfactory and piriform cortex. J. Neurosci. Res. 52: 709–722.

Karp SJ, Masu M, Eki T, Ozawa K, Nakanishi S. 1993. Molecular clon-ing and chromosomal localization of the key subunit of the humanN-methyl-D-aspartate receptor. J. Biol. Chem. 268: 3728–3733.

Ko HW, Park KY, Kim H, Han PL, Gwag BJ, Choi EJ. 1998. Ca2+-mediated activation of c-Jun N-terminal kinase and nuclear factor kBby NMDA in cortical cell cultures. J. Neurochem. 71: 1390–1395.

Komuro H, Rakic P. 1993. Modulation of neuronal migration byNMDA receptors. Science 260: 95–97.

Kotter R. 1994. Postsynaptic integration of glutamatergic and dopa-minergic signals in the striatum. Prog. Neurobiol. 44: 163–196.

Kumari M, Ticku MK. 1998. Ethanol and regulation of the NMDAreceptor subunits in fetal cortical neurons. J. Neurochem. 70: 1467–1473.

Kutsuwada T, Kashiwabuchi N, Mori H, Sakimura D, Kushiya E, ArakiK, Meguro H, Masaki H, Kumanishi G, Arakawa M, Mishina M.1992. Molecular diversity of the NMDA receptor channel. Nature

358: 36–41.Lai SK, Wong CK, Yang MS, Yang KK. 2000. Changes in expression

of N-methyl-D-aspartate receptor subunits in the rat neostriatumafter a single dose of antisense oligonucleotide specific for N-methyl-D-aspartate receptor 1 subunit. Neuroscience 98: 493–500.

Laurie DJ, Seeburg PH. 1994. Regional and developmental heterogene-ity in splicing of the rat brain NMDAR1 mRNA. J. Neurosci. 14:3180–3194.

Liu A, Hoffman PW, Lu W, Bai G. 2004. NF-kappaB site interactswith Sp factors and up-regulates the NR1 promoter during neuronaldifferentiation. J. Biol. Chem. 88: 564–575.

Lovinger DM, White G, Weight FF. 1989. Ethanol inhibits NMDA-activated ion current in hippocampal neurons. Science 243: 1721–1724.

Luetjens CM, Bui NT, Sengpiel B, Munstermann G, Poppe M,Krohn AJ, Bauerbach E, Krieglstein, Prehn JH. 2000. Delayedmitochondrial dysfunction in excitotoxic neuron death: cytochrome crelease and secondary increase in superoxide production. J. Neurosci.

20: 5715–5723.Malis CD, Bonventre JV. 1985. Mechanism for calcium potentiation of

oxygen free radical injury to renal mitochondria. J. Biol. Chem. 261:14201–14208.

McDonald JW, Silverstein FS, Johnston MV. 1990. MK-801 pretreat-ment enhances N-methyl-D-aspartate brain injury and increases brainN-methyl-D-aspartate recognition site binding in rats. Neuroscience

38: 103–113.McInnis J, Wang C, Anastasio N, Hultman M, Ye Y, Salvemini D,

Johnson KM. 2002. The role of superoxide and nuclear factor-kappaB signaling in N-methyl-D-aspartate-induced necrosis andapoptosis. J. Pharmacol. Exp. Ther. 301: 478–487.

Meldrum B, Garthwaite J. 1990. Excitatory amino acid neurotoxicityand neurodegenerative disease. Trends Pharmacol. Sci. 11: 379–387.

Miyamoto K, Nakanishi H, Moriguchi S, Fukuyama N, Eto K,Wakamiya J, Murao K, Arimura K, Osame M. 2001. Involvement ofenhanced sensitivity of N-methyl-D-aspartate receptors in vulnerabil-ity of developing cortical neurons to methylmercury neurotoxicity.Brain Res. 901: 252–258.

Monyer H, Sprengel R, Schoepfer R, Herb A, Higuchi M, Lomeli H,Burnashev N, Sakmann B, Seeburg PH. 1992. Heteromeric NMDAreceptors: molecular and functional distinction of subtypes. Science

256: 1217–1221.Mori H, Mishina M. 1995. Structure and function of the NMDA

receptor channel. Neuropharmacology 34: 1219–1237.Moriyoshi K, Masu M, Ishii T, Shigemoto R, Mizuno N, Nakanishi S.

1991. Molecular cloning and characterization of the rat NMDAreceptor. Nature 354: 31–37.

Muller D, Wang C, Skibo G, Toni N, Cremer H, Calaora V, RougonG, Kiss JZ. 1996. PSA-NCAM is required for activity-inducedsynaptic plasticity. Neuron 17: 413–422.

Nishi M, Hinds H, Lu HP, Kawata M, Hayashi Y. 2001. Motoneuron-specific expression of NR3B, a novel NMDA-type glutamate receptorsubunit that works in a dominant-negative manner. J. Neurosci. 21:RC185.

Noble D. 2003. The future: putting Humpty-Dumpty together again.Biochem. Soc. Trans. 31: 156–158.

O’Donnell P, Grace AA. 1995. Synaptic interactions among excitatory



DOI: 10.1002/jat

afferents to nucleus accumbens neurons: hippocampal gating ofprefrontal cortical input. J. Neurosci. 15: 3622–3639.

Olney JW. 1990. Excitotoxic amino acids and neuropsychiatric disor-ders. Annu. Rev. Pharmacol. Toxicol. 30: 47–71.

Osborn L, Kunkel A, Nabel GJ. 1989. Tumor necrosis factor alpha andinterleukin 1 stimulate the human immunodeficiency virus enhancerby activation of nuclear factor kappa B. Proc. Natl Acad. Sci. USA

86: 2336–2340.Patel M, Day BJ, Crapo JD, Fridovich I, McNamara JO. 1996. Require-

ment for superoxide in excitotoxic cell death. Neuron 16: 345–355.Paule MG. 2001. Validation of a behavioral test battery for monkey. In:

Methods of behavioral analysis in neuroscience, Buccafusco JJ (ed).CRC Press LLC: Boca Raton, FL; 281–294.

Paupard MC, Friedman LK, Zukin RS. 1997. Developmental regulationand cell-specific expression of NMDA receptor splice variants in rathippocampus. Neuroscience 79: 399–409.

Planells-Cases R, Sun W, Ferrer-Montiel AV, Montal M. 1993.Molecular cloning, functional expression, and pharmacological char-acterization of an N-methyl-D-aspartate receptor subunit from humanbrain. Proc. Natl Acad. Sci. USA 90: 5057–5061.

Portera-Cailliau C, Price DL, Martin LJ. 1997. Non-NMDA andNMDA receptor-mediated excitotoxic neuronal deaths in adult brainare morphologically distinct: Further evidence for an apoptosis-necrosis continuum. J. Comp. Neurol. 378: 88–104.

Roberts E, Meredith MA, Ramoa AS. 1998. Suppression of NMDAreceptor function using antisense DNA blocks ocular dominance plas-ticity while preserving visual responses. J. Neurosci. 80: 1021–1032.

Rudolf GD, Cronin CA, Landwehrmeyer GB, Standaert DG, PenneyJB, Young AB. 1996. Expression of N-methyl-D-aspartate in theprefrontal cortex of the rat. Neuroscience 93: 417–427.

Salvemini D, Wang ZQ, Zweier JL, Samouilov A, Macarthur H, MiskoTP, Currie MG, Cuzzocrea S, Sikorski JA, Riley DP. 1999. Anonpeptidyl mimic of superoxide dismutase with therapeutic activityin rats. Science 286: 209–210.

Scallet A, Schmued LC, Slikker W, Grunberg N, Faustino PJ, Davis H,Lester D, Pine PS, Sistare F, Hanig JP. 2004. Developmentalneurotoxicity of ketamine: morphometric confirmation, exposureparameters, and multiple fluorescent labeling of apoptotic neurons.Toxicol. Sci. 81: 364–370.

Shou Y, Gunasekar PG, Borowitz JL, Isom GE. 2000. Cyanide-inducedapoptosis involves oxidative-stress-activated NF-kappaB in corticalneurons. Toxicol. Appl. Pharmacol. 164: 196–205.

Slikker W, Xu Z, Wang C. 2005. Application of a systems biologyapproach to developmental neurotoxicology. Reprod. Toxicol. 19:305–319.

Smith AD, Bolam JP. 1990. The neuronal network of the basal gangliaas revealed by the study of synaptic connections of identified neu-rons. Trends Neurosci. 13: 259–265.

Standaert DG, Landwehrmeyer GB, Kerner JA, Penney JBJ, Young AB.1996. Expression of NMDAR2 glutamate receptor subunit mRNAin neurochemically identified interneurons in the rat neostriatum,neocortex and hippocampus. Brain Res. Mol. Brain Res. 42: 89–102.

Sun L, Margolis FL, Shipley MT, Lidow MS. 1998. Identification of along variant of mRNA encoding the NR3 subunit of the NMDAreceptor: its regional distribution and developmental expression in therat brain. FEBS Lett. 441: 392–396.

Toescu EC, Myronova N, Verkhratsky A. 2000. Age-related structuraland functional changes of brain mitochondria. Cell Calcium 28: 329–338.

Trevisan L, Fitzgerald LW, Brose N, Gasic GP, Heinemann SF, DumanRS, Nestler EJ. 1994. Chronic ingestion of ethanol up-regulatesNMDAR1 receptor subunit immunoreactivity in rat hippocampus.J. Neurochem. 62: 1635–1638.

Wahlestedt C, Golanov E, Yamamoto S, Yee F, Ericson H, Yoo H,Inturrisi CE, Reis DJ. 1993. Antisense oligodeoxynucleotides toNMDA-R1 receptor channel protect cortical neurons from excito-toxicity and reduce focal ischaemic infarctions. Nature 363: 260–263.

Wang C, Anastasio N, Popov V, LeDay A, Johnson KM. 2004. Block-ade of N-methyl-D-aspartate receptors by phencyclidine causes theloss of corticostriatal neurons. Neuroscience 125: 473–483.

Wang C, Fridley J, Johnson KM. 2005a. The role of NMDA receptorupregulation in phencyclidine-induced cortical apoptosis in organo-typic culture. Biochem. Pharmacol. 69: 1373–1383.

Wang C, McInnis J, Ross-Sanchez M, Shinnick-Gallagher P, Wiley JL,Johnson KM. 2001. Long-term behavioral and neurodegenerativeeffects of perinatal phencyclidine administration: implications forschizophrenia. Neuroscience 107: 535–550.

Wang C, McInnis J, West JB, Bao J, Anastasio N, Guidry JA, Ye Y,Salvemini D, Johnson KM. 2003. Blockade of phencyclidine-inducedcortical apoptosis and deficits in prepulse inhibition by M40403, asuperoxide dismutase mimetic. J. Pharmacol. Exp. Ther. 304: 266–271.

Wang C, Sadovova N, Fu X, Scallet A, Hanig J, Slikker W. 2005b. Therole of NMDA receptors in ketamine-induced apoptosis in ratforebrain culture. Neuroscience 132: 967–977.

Wang C, Sadovova N, Hotchkiss C, Fu X, Scallet AC, Patterson TA,Hanig J, Paule MG, Slikker Jr. W. 2006. Blockade of N-methyl-D-aspartate receptors by ketamine produces loss of postnatal day 3monkey frontal cortical neurons in culture. Toxicol. Sci. 91: 192–201.

Wang C, Showalter VM, Jillman GR, Johnson KM. 1999. Chronicphencyclidine increases NMDA receptor NR1 subunit mRNA in ratforebrain. J. Neurosci. Res. 55: 7762–7769.

Waters M, Boorman G, Bushel P, Cunningham M, Irwin R, Merrick A,Olden K, Paules R, Selkirk J, Stasiewicz S, Weis B, Van Houten B,Walker N, Tennant R. 2003a. Systems toxicology and the chemicaleffects in biological systems (CEBS) knowledge base. EHP Toxi-

cogenomics 111: 15–28.Waters MD, Olden K, Tennant RW. 2003b. Toxicogenomic approach

for assessing toxicant-related disease. Mutat. Res. 544: 415–424.Williams K, Dichter MA, Molinoff PB. 1992. Up-regulation of N-

methyl-D-aspartate receptors on cultured cortical neurons after expo-sure to antagonists. Mol. Pharmacol. 42: 147–151.

Wong HK, Liu XB, Matos MF, Chan SF, Perex-Otano I, Boysen M,Cui J, Nakanishi N, Trimmer JS, Jones EG, Lipton SA, Sucher NJ.2002. Temporal and regional expression of NMDA receptor subunitNR3A in the mammalian brain. J. Comp. Neurol. 450: 303–317.

Wright LKM, Popke EJ, Allen RR, Pearson EC, Hammond TG, PauleMG. 2006. Effect of chronic NMDA receptor and/or sodium channelblockade on learning behavior. Neurotoxicol. Teratol. (submitted).

Zahm DS, Brog JS. 1992. On the significance of subterritories in the‘accumbens’ part of the rat ventral striatum. Neuroscience 50: 751–767.

Zecevic N, Bourgeois JP, Rakic P. 1989. Changes in synaptic densityin motor cortex of rhesus monkey during fetal and postnatal life.Brain Res. Dev. Brain Res. 50: 11–32.

Zukin RS, Bennett MV. 1995. Alternatively spliced isoforms of theNMDAR1 receptor subunit. Trends Neurosci. 18: 306–313.

systems biology approaches for toxicology

Documents