Systems Biology in Small Fish Models
D.L. Villeneuve
The McKim Conference on Predictive Toxicology
Sept. 25-27, 2007, Duluth, Minnesota
Systems Biology in Small Fish Models
1. What is systems biology or the “systems perspective”?
2. The systems perspective in ecotoxicology.
3. An overview of a systems-oriented small fish research program.
4. A case study
5. Conclusions
The Systems Perspective
What is a system?
A subset of the world whose behavior, and whose interaction
with the world, we believe can be sensibly described.
• The experimenter can draw a boundary around it, heuristically
• The experimenter can conduct defined perturbations within the
boundaries
• The experimenter can, by reasoning generate sensible
explanations for the changed behavior.
• Learning is assessed by how well the understanding enables
prediction of changed system behavior in response to defined
perturbations.
Kuipers, B. 1994. MIT Press, Cambridge
Brent 2004, Nature Biotech. 22: 1211-1214
The Systems Perspective
Systems biology is the study of
an organism, viewed as an
integrated and interacting
network of genes, proteins and
biochemical reactions which
give rise to life.
www.oz.net/~geoffsi/bm2003-days/bm2003-20.htm
• Systems are comprised of parts which interact.
• Interaction of these parts gives rise to "emergent properties".
• Emergent properties cannot be attributed to any single parts of the system. Irreducible.
• To understand systems, and to be able to fully understand a system's emergent properties,
systems need be studied as a whole.
http://www.systemsbiology.org/Intro_to_ISB_and_Systems_Biology/Why_Systems_Matter
The Systems Perspective
•Non-equilibrium thermodynamics (1930s-40s)
• dealt with integration quantitatively
• aimed to discover general principles >> descriptive
• established connection to molecular mechanisms
• Investigation of biological self-organization (1950s)
• examination of how structures, oscillations, and/or waves arise in a steady or
homogenous environment
• Feedback regulation in metabolism (late 1950s)
• Systems theory in biology (1960s)
• “search for general biological laws governing behavior and evolution of living matter
in a way analogous to the relation of physical laws and non-living matter”
• Metabolic control analysis (1970s)
•Approaches to characterize properties of networks of interacting chemical reactions
• Convergence with “mainstream” molecular biology – high
throughput, genome-scale, “data-rich”
Westerhoff and Palsson, 2004 Nature Biotech. 22:1249-1252
Wolkenhauer. 2001 Briefings in Bioinformatics. 2:258-270
The Systems Perspective
• Shift from reductionism to a holistic perspective
• Should reduce complexity rather than adding additional layers
of complexity
• Search for organizing principles over construction of predictive
descriptions (models) that exactly describe the evolution of a
system in space and time.
• Identification of new concepts and hypotheses that provide a
conceptual structure with logical coherence to rival chemistry
and physics.
Mesarovic et al. 2004. Syst. Biol. 1:19-27
The Systems Perspective in Ecotoxicology
• Shift from reductionism to a holistic perspective
• Historically, ecotoxicology and ecological risk
assessment was holistic in focus
• Apical/integrative endpoints: Survival, growth,
reproduction
• Reductionist in the
sense that testing of the
universe of chemicals
was the paradigm.
• Not feasible to test every chemical, let alone
every chemical mixture
• Increased emphasis on subtle, chronic
impacts (e.g., development, behavior, etc.) with
long-term population implications.
Average
Population
Size Size
Average
Population
(Proportion
of Carrying
Capacity)
(Proportion
of Carrying
Capacity)
Forecast Population Trajectories
1
1
A A
0%
0.8
0.8
0.6
0.6
0.4
0.4
B B
25%
0.2
0.2
E
C C
50%
D
D
0 >95%E 75%
0 0
5
0
5
10
10
Time (Years)
Time (Years)
15
15
20
20
The Systems Perspective in Ecotoxicology
Search for organizing principles that underlie biological
response to stressors.
Develop a conceptual structure with logical coherence that
allows us to predict, with reasonable and quantitative certainty,
integrated, ecologically-relevant impacts.
Linkage of Exposure and Effects Using Genomics,
Proteomics, and Metabolomics in Small Fish Models
• USEPA – Cincinnati, OH
• D. Bencic, M. Kostich, I. Knoebl, D. Lattier, J. Lazorchak, G. Toth, R.
Wang,
• USEPA – Duluth, MN, and Grosse Isle, MI
• G. Ankley, E Durhan, M Kahl, K Jensen, E Makynen, D. Martinovic,
D. Miller, D. Villeneuve,
• USEPA – Athens, GA
• T. Collette, D. Ekman, J. Kenneke, T. Whitehead, Q. Teng
• USEPA-RTP, NC
• M. Breen, R. Conolly
• USEPA STAR Program
• N. Denslow (Univ. of Florida), E. Orlando, (Florida Atlantic University),
K. Watanabe (Oregon Health Sciences Univ.), M. Sepulveda (Purdue
Univ.)
• USACE – Vicksburg, MS
• E. Perkins
• Other partners
• Joint Genome Institute, DOE (Walnut Creek, CA)
• Sandia, DOE (Albuquerque, NM)
• Pacific Northwest National Laboratory (Richland, WA)
• C. Tyler (Univ. Exeter, UK)
Compartment
GABA
Dopamine
Brain
?
?
?
PACAP
Pituitary
GnRH
Neuronal
System
GnR
H
D2 R
Y2 R
GnRH
R
2
Muscimol* (+)
3
Apomorphine (+)
4
Haloperidol (-)
Y1 R
Gonadotroph
D2 R
GPa
FSHb
Blood
NPY
GAB
AB R
Activin
R
Fipronil* (-)
Y2 R
GAB
AA R
PAC1
R
1
D1 R
Follistatin
Activi
n
Chemical “Probes”
Circulating LDL,
HDL
LHb
Circulating LH, FSH
LDL R
LH R
FSH R
5
Trilostane (-)
6
Ketoconazole (-)
HDL R
Cholestero
l
Outer mitochondrial
membrane
StAR
Inner mitochondrial
membrane
Gonad
(Generalized, gonadal,
steroidogenic cell)
Activi
n
Inhibi
n
P450scc
pregnenolone
3bHSD
17α-hydroxyprogesterone
P450c1
7
20βHS
D
7
Fadrozole (-)
8
Prochloraz (-,-)
progesterone
androstenedion
e
17βHSD
17α,20β-P
(MIS)
testosterone
P450arom
P45011β.
9
11βHSD
11-ketotestosterone
Blood
estradio
l
10
Circulating Sex Steroids / Steroid
Hormone Binding Globlulin
11
Androgen / Estrogen
Responsive Tissues
(e.g. liver, fatpad, gonads)
ER
AR
12
Vinclozolin (-)
Flutamide (-)
β-trebolone (+)
Ethynyl estradiol (+)
Figure 1. Conceptual Overview of Research
Increasing Diagnostic (Screening) Utility
Increasing Ecological Relevance
Levels of
Biological
Organization
Molecular
Cellular
Organ
Individual Population
mRNA, protein,
Enzyme assays
Metabolite
profiles
Functional and
Structural change
(Pathology)
Altered reproduction Decreased numbers
or development
of animals
Well characterized
genome,
Phase 2.
low eco / regulatory
relevance
Zebrafish genomics
proteomics,
metabalomics
Computational
modeling
Systems modeling
Poorly characterized
genome
Phase 3.
Phase 1.
Fathead minnow
molecular markers
metabolomics
Fathead minnow 21 d reproduction test
high eco / regulatory
relevance
→’s Depict the flow of information
Population
modeling
A Case Study with Fadrozole
N
N
N
N
CN
CN
Aromatase
7
Ovary
A
B
6
*C*
4
IC50 = 6.93 +/- 0.80 μM
*D*
3
*E*
2
1
0
0.1
1
10
100
fadrozole (uM)
5000
Brain
A
4500
*B*
4000
3500
fmol/h/mg
fmol/h/mg
5
3000
*C*
2500
2000
IC50=8.82 +/- 1.58 μM
*D*
1500
1000
*E*
500
*F*
0
0.01
0.1
1
10
fadrozole (uM)
100
1000
N
A Case Study with Fadrozole
N
CN
Cholesterol
CYP11A
CYP17
(hydroxylase)
Pregnenolone
CYP17
(lyase)
17a-OH-Pregnenolone
3b-HSD
Progesterone
DHEA
3b-HSD
3b-HSD
CYP17
(hydroxylase)
CYP21
17a-OH-Progesterone
CYP21
CYP17
(lyase)
20b-HSD
CYP19
Androstenedione
17b-HSD
estrone
17b-HSD
CYP19
11-deoxycorticosterone
CYP11B1
Testosterone
11-deoxycortisol
corticosterone
CYP11B2
CYP11B2
17α20β-dihydroxy-4pregnen-3-one
CYP11B1
11β-OHTestosterone
11βHSD
aldosterone
cortisol
11-Ketotestosterone
17b-estradiol
Androstenedione
17b-HSD
N
estrone
17b-HSD
CYP19
Testosterone
17b-estradiol
24 h-E2
CN
25oC,
Ex vivo E2 production
(ng/ml*12 h)
0.4
12 h
A
0.35
0.3
B
0.25
0.2
C
0.15
0.1
0.05
0
0
RIA
Ex vivo T production
(ng/ml*12 h)
N
CYP19
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
24 h-T 3
30
Fadrozole (ug/L)
B
AB
A
0
3
30
VTG (mg/ml)
Fadrozole, Female Vtg
18
16
14
12
10
8
6
4
2
0
a
Female 24 h
b
b
z
0
6
Fadrozole (ug/L-nominal)
60
Brain
sGnRH
neurons
cGnRH
neurons
dopaminergic
neurons
sGnRH release
cGnRH release
dopamine release
GnRH
FSH
gonadotrophs
LH
gonadotrophs
FSH
LH
Blood supply
Blood supply
Pituitary
Testis
Ovaries
Leydig
cells
steroids
Liver
Sertoli
cells
Spermatocytes
Sperm mat.
Steroid metab.
Theca
cells
Granulosa
cells
steroids
vitellogenesis
Oocytes
oocyte mat.
Figure Key
state transition
Gonadotropin expression
Steroid metabolismand secretion
transcriptional activatio
translational activation
transcription inhibition
dissociation
association
genes
GnRH release
mRNA
protein
activated
protein
receptor
Vitellogenesis and
Oocyte maturation
Steroidogenesis
Cholesterol transport
pheno
type
Figure Key
state transition
transcriptional activatio
translational activation
transcription inhibition
dissociation
association
genes
mRNA
protein
activated
protein
receptor
pheno
type
N
N
Inner mitochondrial membrane
CN
Fadrozole
exposure
Ex vivo E2 production
(relative to control)
10
post-exposure
0
3
30
#
#
*
1
*
*
*
#
#
#
0.1
24 h
48 h
4d
9d
Time point
exposure
10
8d
10 d
12 d
16 d
post-exposure
#
Evidence for
compensatory/feedback
response to fadrozole.
Ex vivo T production
(relative to control)
#
#
#
#
#
*
#
1
0
3
30
24 h
48 h
4d
0.1
8d
9d
Time point
10 d
12 d
16 d
Aromatase A (CYP19A) gene expression
exposure
18000
post-exposure
CYP19A mRNA
(copies/ng total RNA)
16000
c
14000
12000
10000
0
8000
3
30
b
6000
4000
2000
b
b
a a
b
a
a
a
a
b
b
a
ab
b
a a
0
1
4
8
9
Time point (d)
12
16
Pituitary
LHβ
LHR
Cholesterol
FSHβ
StAR
Inner mitochondrial
membrane
Cyt b5
Cholesterol
CYP11A
Pregnenolone
CYP17
(hydroxylase)
CYP17
(lyase)
17a-OH-Pregnenolone
3b-HSD
Progesterone
FSHR
G-protein mediated
Signal transduction
DHEA
3b-HSD
3b-HSD
CYP17
(hydroxylase)
17a-OH-Progesterone
CYP17
(lyase)
20b-HSD
CYP19
Androstenedione
17b-HSD
17b-HSD
CYP19
Testosterone
17α20β-dihydroxy-4pregnen-3-one
CYP11B1
11β-OHTestosterone
11βHSD
Fig. 1
estrone
11-Ketotestosterone
17b-estradiol
P450scc (CYP11A) gene expression
exposure
P450scc mRNA
(transcript abundance relative to
controls)
2.5
2
post-exposure
b
b
1.5
a
a
ab
a
b
ab
a
a
1
b
a
a
a
a
a
ab
3
a
0.5
0
1
4
8
9
Time point (d)
12
0
16
30
FSHR gene expression
exposure
450
post-exposure
0
3
30
400
FSHR mRNA
(copies/ng total RNA)
c
350
b
300
a
b
200
150
c
b
250
a
b
b
a
a
a
a
a
100
b
ab
a
a
50
0
1
4
8
9
Exposure duration (d)
12
16
Despite evidence for compensation at the molecular and
biochemical level, 21 d exposure to fadrozole causes
adverse apical effect on reproduction.
6
10
4
8
Fadrozole (ug/L)
*
2
*
Vtg (mg/ml)
0
20
Cumulative Number of Eggs
(Thousands)
E2 (ng/ml)
8
Control
2
10
6
50
*
*
*
4
2
10
0
*
-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0
*
0
Control
2
10
2 4
Exposure (d)
*
50
Fadrozole (µg/l)
• Under what conditions would compensation be successful?
• Under what conditions does it fail to prevent adverse effect?
• Under what conditions does it contribute to or exacerbate the
initial impact of the stressor?
6
8 10 12 14 16 18 20
A Case Study with Fadrozole
• Supervised – low content analyses
• Unsupervised – high content analyses
High-content analyses
N
N
Fadrozole
Gonads
CN
Brain
Liver
High desnity
oligonucleotide
microarrays
Enriched Gene Ontology Categories
Brain
Cholesterol biosynthesis
Cholesterol transport
Bile acid synthesis
Cytoskeletal components
Oxidative phosphorylation
Coagulation/wound
response
Immune response
Ion transport
Cell adhesion
Reproduction
Nucleotide biosynthesis
Liver
Ribosomal proteins
Ribosomal RNAs
Translation
Oocytes/oogenesis
Iron transport/plasma
proteins
Immune response and
inflammation
Cytoskeletal components
Ovary
Ribosomal proteins
Ribosomal RNAs
Extracellular matrix
Connective tissue
Coagulation/wound
response
Cytoskeletal components
Oxidative phosphorylation
Ion transport
Embryonic development
Cell adhesion
Oogenesis
Reverse engineering gene regulatory architecture
Different reverse engineering algorithms.
Dynamic Bayesian Network (Yu et al 2004, Zhao et al 2006 )
Boolean (Hashimoto et al 2004)
Information theory (Margolin et al 2006)
Network Approach
•
•
•
•
Start with complex system
Determine players and links in network
Calculate network statistics
Use network statistics to make predictions about network
function
Robustness and fragility in ecosystems and gene networks
is related to network architecture (sensitivity or
insensitivity to network perturbation)
In general, more interactions = more plasticity
Biological systems – generally many nodes with relatively
few links, and relatively few nodes with many links (critical
regulatory nodes)
Csete & Doyle 2004 Trends in Biotechnol. 22:446-450
Zhao et al. 2006. BMC Bioinformatics 7:386
Csete & Doyle 2002. Science. 295:1664-1669
CONCLUSIONS
• Consideration of interactions and relationships is at the heart of the
systems perspective. Systems models serve as a critical translator
between initiating events (modeled by QSAR) and biological outcome.
• Systems models need not be infinitely detailed. Resolution required
defined by the questions to be answered.
• Toxicity pathways linking exposure to adverse effect include the direct
effect of the chemical and a wide variety of indirect secondary,
tertiary….and potentially stochastic effects that influence the apical
outcome.
• Molecular and biochemical responses to stressors are both dose and
time-dependent, and best represented in three dimensions. Important
consideration for biomarker-based bioassay and high-throughput
screening.
CONCLUSIONS
• Modern “omics” tools facilitate unprecedented scale and detail in the
descriptive analysis of biological systems. The greater challenge is in
defining relationships and the general principles that govern them.
•Overall architecture of the systems or networks can reveal important
attributes related to function.
•Hypothesis driven and unsupervised analysis of biological systems or
networks can reveal critical regulatory nodes that integrate signals and
drive component function or phenotype.
• The ultimate goal of systems ecotoxicology is discovery of
generalized or universal principles that govern biological responses to
stressors.