Transcript ppt

Behavioral Metabolomics
October 21st, 2010
By Joseph L McClay
Center for Biomarker Research and Personalized Medicine
Presentation Overview
Center for
Biomarker
Research
and
Personalized
Medicine
• The “omics” philosophy
• Metabolomics as an assay of
biological function
• Technologies (MS, NMR)
• Neurochemical metabolomics in
rodents
• Study of methamphetamine
• Summary
• Bioinformatics tools example
Hierarchies of Order
Center for
Biomarker
Research
and
Personalized
Medicine
Oltavi & Barabasi (2002) Science 298, p763
• Many omics variants:
• DNA sequence
Center for
Biomarker
Research
and
Personalized
Medicine
• GWAS
• Whole genome sequencing
• Epigenetics
• Whole genome methylation
• Gene expression (RNA)
• Expression arrays
• microRNA arrays
• Protein
• Proteomics
• Metabolites
• Metabolomics
• Metabonomics
The omics “principle”
Center for
Biomarker
Research
and
Personalized
Medicine
• Assume you know nothing
• Try to measure everything
• Is this a hypothesis-driven approach
to science?
• Advantages – new discovery
• Disadvantages – false positives, cost
Law of the Instrument
Center for
Biomarker
Research
and
Personalized
Medicine
• “If you have a hammer, everything looks
like a nail”
• Omics approaches are very technology
driven
• Technology = assays + informatics
• Pushing the limits of technology is
extraordinarily expensive
• However, there is the opportunity to break
open the complexity of biology
Metabolomics
Center for
Biomarker
Research
and
Personalized
Medicine
• Biochemistry on a large
scale
• Examination of all
endogenous
metabolites (under
1500Da) in a sample
• Several thousand in
human metabolome
• Ultimate indicators of
biological system
response
Possible applications
Center for
Biomarker
Research
and
Personalized
Medicine
• Comparison of tissue-specific
metabolic profiles
• Drug effects on metabolism
• Personalized medicine
• Developmental effects
• Metabolic disturbances in disease /
pathogenesis
• In combination with other omics
• For example, GWAS to map quantitative
trait loci for individual differences in
metabolite leves (mQTLs)
Technologies – characterizing the mixture
Center for
Biomarker
Research
and
Personalized
Medicine
Mass Spectrometry
Nuclear Magnetic
Resonance
What are the data like?
Center for
Biomarker
Research
and
Personalized
Medicine
Brain mass spec (Woods et al 2006)
Urinary 1H NMR (McClay et al 2010)
• Input is a complex
mixture of metabolites
• Integrate across
spectrum / identify
specific compounds
• Examination of relative
peak heights / integrals
or compound levels
• So, quantitative in
nature (more akin to
gene expression than
genotype data)
Methamphetamine
Center for
Biomarker
Research
and
Personalized
Medicine
• percentage of past-year MA use among persons
12+ has remained relatively stable
• Estimates ranging from 0.7% in 2002 to 0.6% in
2007
However,
admissions
to treatment
programs
have increased
dramatically
since the
mid 1990s
Rationale for a metabolomics study
of methamphetamine in mice
Center for
Biomarker
Research
and
Personalized
Medicine
• Behavioral studies and animal models are well
worked out
• While some gene expression and other studies
have been carried out, to date no metabolomics
study
• Returning to the “omics” principles outlined earlier,
do we really know all the effects of meth?
• If we can better characterize the effects, we can
perhaps see pathways that could mediate the
addiction process
• Find candidate compounds for in vivo imaging
Study design
Center for
Biomarker
Research
and
Personalized
Medicine
• 8 inbred strains of mice, chosen for
maximum genetic variation
• 48 mice per strain
• Acute vehicle, 1, 3 or 10mg/kg meth
• Chronic vehicle or 3mg/kg meth for 5 days
• 1 hour behavioral assessments of
locomotor activity using automated boxes
• Followed by sacrifice, brain excision and
freezing in liquid nitrogen
• Shipment to Metabolon, RTP, NC
• GC and LC mass spectrometry
Overview schematic
Center for
Biomarker
Research
and
Personalized
Medicine
Behavioral pharmacology
Acute Behavioral Effects, Significant Outcomes
totdist
movno
1 mg/kg
3 mg/kg
10 mg/kg
Vehicle
1 mg/kg
3 mg/kg
10 mg/kg
0
Vehicle
1 mg/kg
3 mg/kg
10 mg/kg
Vehicle
1 mg/kg
3 mg/kg
Methamphetamine Dose
Methamphetamine Dose
restime
strcnt
strno
strtime
10 mg/kg
strtime
1 mg/kg
3 mg/kg
10 mg/kg
1 mg/kg
3 mg/kg
10 mg/kg
0
1 mg/kg
3 mg/kg
10 mg/kg
Vehicle
1 mg/kg
3 mg/kg
mrgdist
mrgtime
ctrdist
ctrtime
10 mg/kg
10 mg/kg
ctrtime
200
400
100
0
0
300
3 mg/kg
200
ctrdist
500
400
mrgtime
0
1 mg/kg
Methamphetamine Dose
300
Methamphetamine Dose
600
Methamphetamine Dose
400
Vehicle
Vehicle
Methamphetamine Dose
200
mrgdist
Vehicle
Methamphetamine Dose
600
Vehicle
0
0
500
50
50
500
strcnt
strno
100
1000
150
Methamphetamine Dose
100
Methamphetamine Dose
550
restime
600
Vehicle
0
0
0
50
50
movtime
100
movno
500
totdist
hactv
1000
movtime
100
1000
150
hactv
2000
Center for
Biomarker
Research
and
Personalized
Medicine
Vehicle
1 mg/kg
3 mg/kg
Methamphetamine Dose
10 mg/kg
Vehicle
1 mg/kg
3 mg/kg
Methamphetamine Dose
10 mg/kg
Vehicle
1 mg/kg
3 mg/kg
Methamphetamine Dose
10 mg/kg
Pharmacometabolomics
Center for
Biomarker
Research
and
Personalized
Medicine
• Acute vehicle, acute 3mg/kg meth
and chronic 3mg/kg meth for 5 days
• 18 mice per strain, 8 strains total
• Test for differences in metabolite
levels between groups
• 300 metabolites in total were
identified by Metabolon and tested
• False Discovery Rate control
necessary because of large number
of tests
Center for
Biomarker
Research
and
Personalized
Medicine
Acute metabolic effects
Center for
Biomarker
Research
and
Personalized
Medicine
Compound
Beta p-value q-value Primary pathway
Class
fructose
-0.282
6.3E-06
0.001 Glycolysis, gluconeogenesis
Carbohydrate
lactate
0.305
4.4E-05
0.003 pyruvate metabolism
Carbohydrate
malate
0.223
0.0001
0.006 Krebs cycle
Energy
2-hydroxyglutarate
0.129
0.0001
0.007
succinate
0.191
0.0003
0.015 Krebs cycle
Energy
tryptophan
0.156
0.0008
0.025 Tryptophan metabolism
Amino acid
fumarate
0.118
0.0009
0.027 Krebs cycle
Energy
linoleate (18:2n6)
0.272
0.0027
0.059 Long chain fatty acid
Lipid
citrate
0.078
0.0043
0.081 Krebs cycle
Energy
sorbitol
-0.224
0.0045
0.081 starch, and sucrose metabolism Carbohydrate
glycerophosphorylcholine
-0.081
0.0052
0.081 Glycerolipid metabolism
Lipid
Chronic effects – part 1
Center for
Biomarker
Research
and
Personalized
Medicine
Compound
Beta
p-value q-value Primary pathway
Class
lactate
malate
citrate
2-hydroxyglutarate
tryptophan
alanine
2-aminoadipate
3-hydroxybutyrate
urea
maltotriose
choline phosphate
gamma-aminobutyrate
ergothioneine
glycerophosphorylcholine
fructose
gamma-glutamyl alanine
serine
glucose
ribose
glycerol 3-phosphate
succinate
0.405
0.305
0.140
0.167
0.227
0.356
0.156
0.320
-0.149
0.773
0.102
0.132
0.159
-0.085
-0.177
0.227
0.067
-0.177
0.299
-0.085
0.141
1.4E-10
3.6E-10
1.3E-09
3.8E-09
6.3E-09
1.8E-06
7.0E-06
4.5E-05
8.4E-05
0.0001
0.0003
0.0003
0.0004
0.0005
0.0005
0.0006
0.0008
0.0012
0.0015
0.0016
0.0018
Glycolysis, gluconeogenesis
Carbohydrate
Krebs cycle
Energy
Krebs cycle
Energy
Tryptophan metabolism
Amino acid
Alanine metabolism
Amino acid
Lysine metabolism
Amino acid
Ketone bodies
Lipid
Urea, arginine metabolism
Amino acid
Sucrose metabolism
Carbohydrate
Glycerolipid metabolism
Lipid
Glutamate metabolism
Amino acid
Glycerolipid metabolism
Lipid
Sucrose metabolism
Carbohydrate
1.3E-07
1.6E-07
3.7E-07
8.3E-07
1.1E-06
0.0003
0.0007
0.003
0.005
0.007
0.013
0.015
0.015
0.019
0.019
0.022
0.025
0.035
0.040
0.041
0.044
Gamma-glutamyl
Glycine, serine and threonine Amino acid
Glycolysis, gluconeogenesis
Carbohydrate
Nucleotide sugars
Carbohydrate
Glycerolipid metabolism
Lipid
Krebs cycle
Energy
Chronic effects – Part II
Center for
Biomarker
Research
and
Personalized
Medicine
Compound
Beta
p-value q-value Primary pathway
Class
uridine
0.283
0.002
0.058
adenosine 5'diphosphoribose
1.356
0.003
0.059
nicotinamide
1.120
0.003
0.059
Nucleotide
Cofactors and
Nicotinamide metabolism vitamins
Cofactors and
Nicotinamide metabolism vitamins
guanosine 5'- monophosphate
0.822
0.003
0.065
Purine metabolism
Nucleotide
glutamine
-0.076
0.003
0.065
Glutamate metabolism
dehydroascorbate
0.648
0.004
0.071
Ascorbate metabolism
Amino acid
Cofactors and
vitamins
ribulose 5-phosphate
1.210
0.004
0.077
Nucleotide sugars
Carbohydrate
phenylalanine
0.091
0.004
0.081
Tyrosine metabolism
Amino acid
maltose
0.692
0.005
0.081
Sucrose metabolism
Carbohydrate
cysteine
1.277
0.005
0.081
Cysteine metabolism
Amino acid
butyrylcarnitine
-0.127
0.005
0.081
Fatty acid metabolism
Lipid
pipecolate
-0.118
0.006
0.086
Lysine metabolism
Amino acid
inosine
0.848
0.006
0.095
Purine metabolism
Nucleotide
Pyrimidine metabolism
Alternate parameterization
Center for
Biomarker
Research
and
Personalized
Medicine
• Between group comparison shows the extensive
metabolic disruption due to meth administration
• However, does not disaggregate acute from chronic
meth effects.
• For this we need a 2nd parameterization:
Intercept (a) represents the “simplest” condition--acute vehicle (av).
Parameter 1 (d1) captures marginal effect of acute meth over acute
vehicle. Parameter 2 (d2) captures marginal effect of chronic vehicle
injection over “just” acute meth. Parameter 3 (d3) captures marginal
effect of chronic meth over chronic vehicle injection + acute meth. We
include with a random intercept to account for clustering within strain
(u0).
Metabolite level = a + b1*d1 + b2*d2 + b3*d3 + u0 + e
Results – alternative
parameterization
Center for
Biomarker
Research
and
Personalized
Medicine
parm
Compound
d1
2-hydroxyglutarate
d3
ergothioneine
d3
choline phosphate
Beta
p-value
q-value
0.129 1.20E-04 0.048
0.19
3.00E-04 0.069
0.118 3.50E-04 0.069
Primary pathway
Class
Citric acid cycle
energy
Dietary
Ceramide signaling phospholipid
Behavioral Metabolomics
Center for
Biomarker
Research
and
Personalized
Medicine
Sensitization:
•Essentially is an increase in response to the
same dose of drug after repeated exposure
•We are measuring locomotor activity
•In locomotor terms, sensitization means that mice
will move around more after their dose of drug on
the last day, as compared to the first day
•However, the automated boxes measure
locomotor activity in several ways
•Around 20 locomotor activity variables are
collected
Factor analysis
4 factors: horizontal/total movement, vertical
movement, center/margin time, stereotypy
Center for
Biomarker
Research
and
Personalized
Medicine
Factor
Factor1
Factor2
Factor3
Factor4
Factor5
Factor6
Factor7
Factor8
Factor9
Factor10
Factor11
Factor12
Variance Difference Proportion Cumulative
6.40
2.04
0.36
0.36
4.36
2.13
0.25
0.61
2.23
0.57
0.13
0.74
1.66
0.24
0.09
0.83
1.42
0.60
0.08
0.91
0.82
0.53
0.05
0.96
0.29
0.10
0.02
0.97
0.18
0.01
0.01
0.98
0.17
0.11
0.01
0.99
0.06
0.01
0.00
1.00
0.06
0.05
0.00
1.00
0.01
0.00
0.00
1.00
Variable
Factor1
Factor2
Factor3
Factor4
-------------+-------------------------------------------------------------------------------rtime_sens~p
0.08
0.34
0.04
-0.08
rmovno_sen~p
-0.01
0.98
0.05
0.00
ractv_sens~p
-0.01
0.96
0.04
-0.01
ctrtime_se~p
-0.06
-0.06
-0.99
0.05
ctrdist_se~p
0.84
0.03
-0.23
0.17
mrgtime_se~p
0.07
0.06
0.99
-0.05
mrgdist_se~p
0.89
0.01
0.34
0.00
strtime_se~p
0.58
-0.19
-0.19
0.33
strno_sens~p
0.33
-0.09
-0.03
0.80
strcnt_sen~p
0.83
-0.10
-0.06
0.19
vtime_sens~p
0.03
0.67
0.02
-0.10
vmovno_sen~p
0.02
0.97
0.06
-0.03
vactv_sens~p
-0.02
0.95
0.02
-0.06
restime_se~p
-0.98
-0.02
0.00
-0.09
movtime_se~p
0.98
0.02
0.00
0.09
movno_sens~p
0.29
-0.02
-0.12
0.82
totdist_se~p
0.97
0.02
0.16
0.07
hactv_sens~p
0.90
-0.02
0.05
0.34
Create BLUPs for each animal for sensitization, i.e.
increase in horizontal movement over course of study
Results – metabolomics
analysis of sensitization
Center for
Biomarker
Research
and
Personalized
Medicine
compound
beta
p-value q-value primary pathway
malonylcarnitine
0.817
7.74E-05
0.014
serine
-0.084
1.54E-04
0.014
homocarnosine
-0.155
1.56E-04
0.014
Amino acid conj
ergothioneine
0.2
2.64E-03
0.177
Unknown
histamine
0.777
5.00E-03
0.238
NADH
0.167
6.20E-03
0.238
Amino acid conj
class
Lipid / Energy
Serine threonine met Amino acid
Peptide
N/A
Histidine metabolism Amino acid
Nicotinamide /
energy
Cofactors and
vitamins
In this analysis, we are correlating individual differences in the levels
of specific metabolites with individual differences in sensitization to
methamphetamine.
Summary
Center for
Biomarker
Research
and
Personalized
Medicine
• Metabolomics analysis can yield insights
into the metabolic sequelae of drug
administration
• In this study, we observed extensive and
dramatic alterations to neurochemistry
following meth administration
• Among specific findings were changes to
glutamine / alanine-related metabolites and
choline phosphate following chronic
adminsitration
• Associations with sensitization implicated
histamine and homocarnosine
Summary (contd)
Center for
Biomarker
Research
and
Personalized
Medicine
• Previous studies have implicated GABA,
histamine, phospholipids etc in relation to
stimulant drug abuse / administration
• This first attempt at neurochemical /
behavioral metabolomics appears
promising
• Much additional work to be done
• Application to other drug / behavior
pairings (e.g. PPI and antipsychotics)
Center for
Biomarker
Research
and
Personalized
Medicine
Many statistical development
opportunities
For example, identify
subsets of metabolites
whose concentrations
are always coupled.
Use that to define test
statistic:
– Multivariate
– Eliminates some of
the dynamics
Acknowledgements
Center for
Biomarker
Research
and
Personalized
Medicine
CBRPM, School of Pharmacy
Edwin van den Oord
Daniel Adkins
Shaunna Clark
Renan Souza
Department of Pharmacology and Toxicology
Patrick Beardsley
Rob Vann
Sarah Vunck
Angela Batman (now at Pfizer UK)
Funding: NIDA
http://www.pharmacy.vcu.edu/biomarker/
Databases
Center for
Biomarker
Research
and
Personalized
Medicine
• What does my metabolite do?
• Choline Phosphate
• Gamma-glutamyl alanine
• Search databases:
• Reactome
• KEGG – Kyoto Encyclopedia of Genes
and Genomes
• BioSystems @ NCBI
Web sites
Center for
Biomarker
Research
and
Personalized
Medicine
• www.reactome.org
• http://www.genome.jp/kegg/
• www.ncbi.nlm.nih.gov/biosystems/