The Toxicologist - Society of Toxicology

cellular pathways and disease interactions, CTD was utilized with STITCH to generate
association networks between genes and proteins and to identify another intermediate
protein, ALAD. Binding assays show that a polymorphism of ALAD,
ALAD2, causes red blood cells to bind lead more tightly and previous research has
implicated ALAD as an important factor in lead susceptibility. VDR polymorphisms
are also associated with lead susceptibility. Although a correlation between
VDR B and higher blood, urine, and plasma lead levels has been reported, the
mechanism and systemic effect of the allele are not well understood. Maps of the
frequency of haplotypes in different populations for VDR B, ALAD2 and HLA-
DRB1 were obtained from GWAS and correlating biological pathways involving
MHCII, ALAD, and VDR were constructed with Genego. In combination, the relationships
identified suggest that ALAD2 and VDR B haplotypes confer an individual
with heightened lead susceptibility. Furthermore, the results suggest that elevated
levels of lead are associated with decreased expression of HLA-DRB1 and
greater risk of MS.
126 Profiling the Activity of Environmental Chemicals in
Causing Testicular Dysgenesis Syndrome Using the US EPA
Toxicity Reference Database (ToxRefDB).
M. C. Leung 1 , K. W. McLaurin 1 , J. Phuong 1 , N. S. Sipes 1 , N. C. Baker 1 ,
N. Kleinstreuer 1 , A. M. Frame 1 , R. Judson 1 , G. R. Klinefelter 2 , M. T. Martin 1
and T. B. Knudsen 1 . 1 National Center of Computational Toxicology, US EPA,
Research Triangle Park, NC; 2 Reproductive Toxicology Division, US EPA, Research
Triangle Park, NC.
Hypogonadism, cryptorchidism, hypospadias, and testicular cancer are increasingly
common male reproductive defects. Clinical and experimental evidence suggest
that these defects are associated with testicular dysgenesis syndrome (TDS). We hypothesize
an Adverse Outcome Pathway (AOP) framework for TDS starting with
disruption of cell signaling and structural targets of the fetal testis during embryogenesis,
followed by a series of key events that lead to male reproductive defects.
Since understanding chemical-endpoint associations is an approach in building the
AOP conceptual models, we mined 4209 guideline animal studies in EPA’s
ToxRefDB, which included 963 compounds tested in rats, mice, and rabbits.
Testicular neoplasia was an outcome for 60 (6.2 %) chemicals. Sperm abnormalities
were an outcome for 72 (9.8 %) chemicals tested in chronic and subchronic toxicity
studies and 44 (5.7 %) chemicals tested in prenatal developmental and/or multigenerational
fertility studies. Reductions in anogenital distance were recorded for
15 (3.5 %) chemicals tested in multigenerational studies, and 11 (1.5 %) chemicals
tested caused hypospadias or cryptorchidism in developmental or multigenerational
studies. The chemical classes showing the broadest activity across different
TDS endpoints were pesticides (amdro, benomyl, primisulfuron-methyl, tetrachlorobenzene,
and vinclozolin) and phthalates (butyl benzyl phthalate, dibutyl phthalate,
and diethylhexyl phthalate) with a lowest effect level on the order of ~300
mg/kg/day. The most potent TDS-actives (17β-estradiol, fluazinam, and tebupirimfos)
produced testicular effects below 0.1 mg/kg/day. These results suggest a hierarchical
pattern of male reproductive defects in ToxRefDB which can be used as
an AOP anchor for TDS. [This work does not reflect EPA policy].
127 Compound Toxicity Profiling and Prioritization Using Tox21
Phase I Quantitative High-Throughput Screening (qHTS)
Cytotoxicity Data.
J. Hsieh 1 , A. Sedykh 2 , R. Huang 3 , M. Xia 3 and R. R. Tice 1 . 1 Division of the
National Toxicology Program, NIEHS, NIH, Research Triangle Park, NC; 2 University
of North Carolina at Chapel Hill, Chapel Hill, NC; 3 National Center for Advancing
Translational Sciences, Bethesda, MD.
Using in vitro data to prioritize compounds for in vivo toxicity testing is a goal of
Tox21. Phase I profiled 1408 NTP compounds against 126 cell-based qHTS assays;
those that measured cytotoxicity record the ability of a chemical to adversely
perturb multiple cellular pathways in a concentration-response fashion. The curves
from 39 cytotoxicity assays (9 human cell types, 4 rodent cell types, 13 sets of identical
twin lymphoblastoid cell lines) were curated to filter noise and remove artifacts.
To quantify the activity of each compound in these assays, we calculated a
weighted version of Area Under the Curve (wAUC) intended to capture potency
and efficacy simultaneously, as well as the conventional half-maximal activity concentration
(AC50) value. The average wAUC or AC50 values across the 39 cytotoxicity
assays were used to rank compounds and the rankings were compared, for
880 compounds, with available rat acute oral toxicity data. The compounds were
categorized as toxic or non-toxic based on various “Globally Harmonized System of
Classification and Labelling of Chemicals” (GHS) acute toxicity thresholds (50,
26 SOT 2013 ANNUAL MEETING
300, 500, 2000 mg/kg). The wAUC approach consistently prioritized more toxic
compounds at the early stage (up 1% non-toxic compounds of dataset) at all
thresholds. The best receiver operating characteristic (ROC) enrichment value at
1% and AUC value are 14.3 [95% CI=1.8-26.8] and 0.73 [95% CI=0.65-0.80],
respectively, at a threshold of 50 mg/kg. Also, some toxic compounds (e.g., actinomycin
D, colchicine, daunomycin) had higher ranks based on wAUC due to their
ill-fitted curves in some of the cell lines, resulting in missing AC50 values. We conclude
that the wAUC provides an additional and useful metric to estimate the toxicity
potential of compounds for Tox21 qHTS assays. Based on this metric, compounds
screened in Tox21 can be prioritized for more extensive testing.
128 Predicting Cellular Dynamics and Key Events in
Developmental Toxicity with a Multicellular Systems Model.
T. B. Knudsen 1 , M. R. Rountree 1 , S. Hunter 2 , N. C. Baker 3 , R. Spencer 3 ,
R. S. DeWoskin 4 and W. Setzer 1 . 1 NCCT, US EPA, Research Triangle Park, NC;
2 NHEERL, US EPA, Research Triangle Park, NC; 3 Lockheed Martin, Research
Triangle Park, NC; 4 NCEA, US EPA, Research Triangle Park, NC.
Computer simulation of cellular networks is one possible solution for modeling key
events in developmental toxicology. We constructed a multicellular agent-based
model (ABM) of early limb-bud development in CompuCell3D (www.compucell3d.org/).
The model simulates key cellular behaviors (mitosis, apoptosis, adhesion,
migration, chemotaxis, shape, secretion), organizing centers (AER, ZPA) and
signals (FGFs, SHH, BMPs, RA). It effectively emulates hindlimb-bud development
during a 42h period in mouse (Theiler stages 16-19) and 160h in human
(Carnegie stages 13-16). The ABM reflects biological variability across parallel simulations
for spatio-temporal expression of biochemical gradients and cell behaviors,
ultimately manifesting in trajectories of outgrowth. To evaluate the model as a tool
for predictive toxicology, we selected 5-Fluorouracil (5FU) as a prototype. 5FU perturbed
13 of 650 ToxCast assays based on AC50s (or LECs) at or below 15 uM.
5FU effects observed in the assays were disruption of stem cell (mES) growth and
differentiation, suppression of TGFb1 signaling and mitochondrial density, p53-induction,
mitotic arrest, reduced cell proliferation and increased cell death.
Challenging the ABM with concentration-response data derived from mES cell
number produced a dose-dependent wave of mitotic arrest and apoptosis, disrupting
outgrowth. Varying the dose and time of exposure localized the primary key
event to arrest of SHH-expressing cells and their geometric relationships to cells expressing
GREM1, a BMP antagonist maintained by SHH signals. Different outcomes
emerged when perturbation of the SHH/GREM1/BMP loop was switched
between mitotic arrest and excessive apoptosis, indicating the importance of considering
both cellular consequences together. These findings support the application
of multi-cellular ABMs as tools to translate cellular dynamics into simulation
of emergent (higher order) tissue effects for predictive toxicology. [This abstract
does not necessarily reflect EPA policy.]
129 Chemical Structure-Based In Silico Phototoxicity Prediction:
An Approach from a Combination of Photochemical
Properties.
Y. Haranosono, S. Nemoto, M. Kurata and H. Sakaki. Senju Pharmaceutical Co.,
Ltd., Kobe, Japan.
Some photochemical properties are essential factor for prediction of phototoxicity,
since phototoxicity is caused by photo-activation of the compounds. Highest
Occupied Molecular Orbital – Lowest Unoccupied Molecular Orbital Gap (HLG)
is a photochemical property of needful energy for photo-activation, and HLG was
reported to be related with phototoxicity based on in vitro 3T3 Neutral Red Uptake
assay (3T3 NRU assay). However there are few reports to predict of phototoxicity
using photochemical property including HLG. In this research, we established the
stepwise approach using Maximum-Conjugated-π-Electron- Number (PEN MC ) of
the compounds in addition to HLG for in vitro phototoxicity prediction. HLG and
PEN MC were calculated by ChemDraw ® and Chem3D ® for total 64 compounds
which were known the results of 3T3 NRU assay (32 positive and 32 negative). As
step 1, we set the cut lines of HLG as follows; the compounds that have over 8.0 of
HLG were determined as negative, and the compounds that have less than 5.2 of
HLG were determined as positive. On the other side, the compounds that have
HLG from 5.2 to 8.0 showed no predicting performance to phototoxicity, and then
we defined this range of HLG as gray zone. As step 2, we employed PEN MC for the
gray zone compounds. We found that PEN MC also indicated correlation to in vitro
phototoxicity, and therefore set the cut lines as follows; the compounds that have
12 and more than of PEN MC were determined as positive, and 11 or less of PEN MC
were determined as negative. This stepwise approach for phototoxicity prediction