The Toxicologist - Society of Toxicology

using area under the concentration curve (AUC) in brain and blood for triadimefon
and triadimenol as dosimetrics. All dosimetric-based HEDs were above the oral reference
dose of 0.11 μmol triadimefon/kg/day.
902 A DERMAL ABSORPTION MODEL BASED ON
SUCCESSIVE PARTITIONING THROUGH A NON-
HOMOGENOUS STRATUM CORNEUM LIPID MATRIX.
D. Van Der Merwe 1 , P. Schumm 2 and C. M. Scoglio 2 . 1 Diagnostic
Medicine/Pathobiology, Kansas State University, Manhattan, KS and 2 Department of
Electrical and Computer Engineering, Kansas State University, Manhattan, KS.
Models of dermal absorption typically treat the stratum corneum lipids as a homogenous
medium through which solutes diffuse according to Fick’s first law of
diffusion. This approach does not explain non-first order diffusion observed experimentally
when the dose rate varies. Successive partitioning through a non-homogenous
stratum corneum lipid matrix offers an alternative approach that can simulate
non-linear dermal absorption patterns. It is based on a conceptual model of the
stratum corneum that includes varied local physical-chemical characteristics and
dynamic solute effects within the lipid matrix. In this approach the rate of solute
exchange between distinct lipid domains become rate limiting, rather than the rate
of diffusion through a relatively voluminous, homogenous medium. The approach
is demonstrated using a large, sparse and regular network model where nodes have
variable characteristics, including limitations on node capacities for solutes that
simulate variable solubility limits in different regions within the stratum corneum
lipid matrix. Rates of solute movement from node to node is determined by the
rates of solute movement within nodes, as characterized by diffusivity, and the rates
of exchange between nodes, as characterized by partitioning coefficients. The network
outputs produce absorption to dose relationships that can be characterized
using power equations, similar to equations used to describe absorption to dose relationships
in experimental dermal absorption data.
903 THE UTILITY OF THE MINI-PIG IN P38 MAP KINASE
INHIBITOR TESTING.
J. Schützsack 1 , A. Gibbs 2 , J. Parish 2 and K. Gill 2 . 1 LEO Pharmacology A/S, DK-
2750 Ballerup, Denmark and 2 Covance Laboratories Ltd., Harrogate, United
Kingdom. Sponsor: D. Everett.
P38 map kinase inhibitors are investigated as a target for many types of inflammatory
diseases due to their inhibitory action on pro inflammatory cytokines such as
TNFα and IL-1β and other inflammatory mediators. The repeat dose toxicity of a
P38 MAPKi was evaluated in rats and minipigs. The minipig was chosen as the
non-rodent species as there were no major qualitative or quantitative differences between
man and minipig in an in vitro metabolite profiling study. Toxicity included
lymphoid proliferation in rats and GI inflammation in minipigs characterised by
increased WBC and acute phase reactants and confirmed histologically. A number
of target related toxicities have previously been reported clinically and preclinically
e.g. hepatic, cardiac, CNS and skin toxicities but were not observed in these studies.
Rats were much less sensitive towards the adverse effects of the compound and
tolerated doses 10-100-fold higher than minipigs although the reason for the difference
in sensitivity is not known. Previously Davis (SOT 2008) reported that
species specific toxicities have been observed with kinase inhibitors and the dog is
uniquely sensitive to p38 MAPK toxicity due to over expression of p38 in B lymphocytes
leading to acute lymphoid necrosis and colonic haemorrhage; thus the
non-human primate (NHP) has traditionally been selected as the second species.
Despite the minipig being more sensitive than the rat, it was considered to be much
less sensitive than the dog and comparable to the NHP and human. These data
demonstrate that the minipig is an excellent alternative to the primate for the toxicological
evaluation of p38 MAPK inhibitors and thus in line with the directive on
animal experimentation, 86/609/EEC aimed at reducing the number of NHP. The
minipig does not share the problems encountered with the use of the dog due to
over-expression of p38 and demonstrated excellent concordance with the human.
Also, the minipig was shown to be a reasonable predictor of human AUC 0-24
and
C max
concentrations.
904 ZINC FINGER NUCLEASE-MEDIATED CREATION OF
RODENT KNOCKOUT MODELS ON TOXICOLOGY.
I. D. Carbery 1 , X. Cui 1 , A. Harrington 2 , L. Liaw 2 and E. Weinstein 1 . 1 Sigma
Advanced Genetic Engineering Labs, Sigma-Aldrich, St. Louis, MO and 2 Transgenic
Core, Maine Medical Center, Scarborough, ME. Sponsor: I. Grossi.
Animal models are an essential element of drug development and toxicity assessment.
Rats are preferred over mice for their closer resemblance to human anatomy
and physiology., Additionally, their larger size allows for multiple sample collections
in pharmacokinetics studies. Until now, knockout rats were not readily available
due to limitations in embryonic stem cell technology. Here we report on the creation
of a suite of rat knockouts on toxicology targets and the first knockout mouse
(Mdr1a -/-) generated using the zinc finger nuclease (ZFN) technology. ZFNs are
injected into one-cell rodent embryos to introduce site-specific modifications on
the chromosome, leading to the disruption of gene function (1). Similar to creating
a transgene, knockout founders can be obtained and analyzed at a high rate in 6-8
weeks of time from the time of injection. For example, 77% of live births from
mouse embryos injected with Mdr1a ZFNs were founders, carrying deletions ranging
from 7 bp to 695 bp around the target site, whereas for Mdr1a knockout rats,
12.5% live births were founders with deletion of 6 bp - 21 bp. In addition to
Mdr1a, we are in the process of creating rats with Bcrp, PXR, Mrp1 and Mrp2 disrupted,
respectively. Detailed genotypes and injection statistics of each model will
be reported. Homozygous animals will be phenotyped. Furthermore, we have recently
generated a p53 knockout rat model that has potential applications in genotoxicity.
References: 1. Geurts et al (2009). Knockout rats via embryo microinjection
of Zinc-finger nucleases.
Science. Vol. 325, 433.
905 GSH-DEPLETED ERYTHROCYTE RAT MODEL OF
DRUG-INDUCED HEMOLYTIC ANEMIA.
J. M. McMillan, R. Mosley and D. C. McMillan. Pharmacology and Experimental
Neurosciences, University of Nebraska Medical Center, Omaha, NE.
We describe studies to develop a primaquine-sensitive hemolytic anemia animal
model using male Sprague-Dawley rats. To increase the sensitivity of normal rat
erythrocytes to pro-oxidant hemolytic agents, donor rat red blood cells (RBCs)
were isolated and treated in vitro with a sufficient amount (2.65 mM) of diethyl
maleate (DEM) to deplete GSH by >95%. GSH-normal and GSH-depleted RBCs
were stained with the fluorescent cell-tracking dyes DiI and DiO, respectivley,
mixed together in an equal ratio, and returned to the circulation of syngeneic recipient
rats. Experiments were performed with this rat model using dapsone (0-120
mg/kg) and primaquine (0-50 mg/kg/day x 5 days). Treatment of the rats with dapsone
provoked a dose-dependent increase in the rate of removal of the dye-tagged
RBCs, and decreased the ratio of DiO:DiI RBCs, indicating that the GSH-depleted
RBCs were removed preferentially over the GSH-normal RBCs. In contrast,
primaquine treatment did not affect RBC survival as determined by lack of detectable
change in the GSH-depleted RBC:GSH-normal RBC ratio. However,
75% of the rats in the primaquine high dose group (50 mg/kg/day) died during the
course of the experiment. Overall, the data support the proof of concept; however,
rats may not be the optimal species to use in pursuit of this animal model. This research
is supported by a sub-contract (08-04-072) from the University of
Mississippi (USAMRMC Award # W81XWH-07-2-0095).
906 EVALUATION OF SCORE METHODS FOR THE
PREDICTION OF DRUG-INDUCED LIVER INJURY IN
HUMANS BY USING CHIMERIC PXB-MICE ® WITH
HIGHLY HUMANIZED LIVER.
S. Nagatsuka 1 , D. Hynes 1 , S. Ninomiya 1 , M. Kakuni 2 , C. Tateno-Mukaidani 2 ,
T. Shimada 2 and Y. Yamazoe 3 . 1 ADME & Toxicology Research Institute, Sekisui
Medical Co., Ltd., Tokai-mura, Ibaraki, Japan, 2 PhoenixBio Co., Ltd., Higashi-
Hiroshima, Hiroshima, Japan and 3 Graduate School of Pharmaceutical Sciences,
Tohoku University, Sendai, Miyagi, Japan.
Of the leading causes of withdrawal of a newly developed drug, drug-induced liver
injury (DILI) takes a significant position. Toxicogenomic approaches have been imployed
using in vivo (experimental animals, mainly rats) or in vitro (human hepatocytes
or hepatoma cells) systems in order to find of promising biomarkers for
DILI. We have used chimeric PXB-mice ® , in which more than 70% of hepatic
parenchymal cells are replaced by human hepatocytes, for the toxicogenomic analyses
of hepatotoxicants. This animal model, which mimics human-type drug metabolism,
has a potential to bridge the gap between rodent-type and human-type livers
and to explain the difference of in vivo and in vitro response of human hepatocytes
against hepatotoxicants. By using 20 different hepatotoxicants (acetaminophen,
amiodarone, diclofenac, d-penicillamine, flutamide, erythromycin, valproate,
sulindac, indomethacin, perhexilene, methyldopa, amitriptyline, tamoxifen, acetylsalicylic
acid, methotrexate, demeclocycline, hydrazine, hydroxyurea, imipramine,
orotic acid) and seven non-hepatotoxicants, we have analyzed changes in hepatic
gene expression in rats and PXB-mice ® . These drugs were orally administered to
rats and PXB-mice ® thee-times daily at high doses (ca. 20% of reported LD 50
), followed
by hepatic total RNA preparation and gene expression analyses using
oligonucleotide microarray chips. Several maker gene candidates, which specifically
SOT 2010 ANNUAL MEETING 193