1. Genetic Analysis of Brain Images from 21,000 People: The ENIGMA Project
Paul M. Thompson1
on behalf of the ENIGMA Consortium2
1Professor of Neurology & Psychiatry,
David Geffen School of Medicine,
Los Angeles, CA, USA 2

2. Introduction: What is ENIGMA?
Worldwide Consortium – we pool human brain images & genome- wide scans (>500,000 common variants in your DNA)
Discover genetic variants that affect brain / disease risk
Enabled largest brain imaging studies ever performed (Nature Genetics, Apr ; 21,151 subjects)
207 co-authors, 125 institutions, >500,000 SNPs, range of brain measures (massive global collaboration; “Crowd-sourcing”)
Founded 2009; run by triumvirate of labs: Thompson (UCLA), Martin/Wright (Queensland), Franke (Netherlands); Many Working Groups

3. Why screen 21,000 brain images?
Amass a sample so vast that we can see how single-letter changes in your DNA affect the brain
Do epidemiology with images (exercise, diet, medication)
Discover genes that:
- damage the brain, affect brain wiring, cause disease (new leads in autism, Alzheimer’s disease)
- estimate our personal risk of mental decline: empowers drug trials (we do this now)
Discover new drug targets

4. What factors harm the brain. 1
What factors harm the brain? 1. Diseases, such as Alzheimer’s – several commonly carried genes boost our risk for this (ApoE4: 3x; CLU, CR1, PICALM: 10-20% more risk each)

5. What helpful or harmful factors affect the brain. 1
What helpful or harmful factors affect the brain? 1. Diseases, such as Alzheimer’s Disease – several commonly carried genes boost our risk for this (ApoE4: 3x; CLU, CR1, PICALM: 10-20% more risk)

6. Comparing drug treatments for mental illness -
Olanzapine Slows Gray Matter Loss;
Imaging Reveals Differences
Thompson/Lilly-HGDH Drug Trial/Lieberman 2008

7. Obese People have 8% more brain atrophy locally (N=432 MRI scans).
Maps show % tissue deficit per unit gain in body mass index (BMI)
1Raji et al. Brain Structure and Obesity. Human Brain Mapping, Aug

8. Geneticists discovered an “obesity gene” (FTO) – surprisingly, we were able to pick up this gene’s effect in brain images (Ho PNAS 2010)
BMI
(N=206 healthy elderly; corrected for multiple comparisons)
FTO association
(N=206 healthy elderly; corrected for multiple comparisons)

9. Alzheimer’s risk gene carriers (CLU-C) have lower fiber integrity even when young (N=398), 50 years before disease typically hits [News covered in 20 countries]
Effects occurred in multiple regions, including several known to degenerate in AD. Such regions included the corpus callosum, fornix, cingulum, SLF and ILF (Liu et al., 2009; Stricker et al., 2009). This suggests that the CLU-C related variability found here might create a local vulnerability important for disease onset.
Voxels where CLU allele C (at rs ) is associated with lower FA after adjusting
for age, sex, and kinship in 398 young adults
(68 T/T; 220 C/T; 110 C/C). FDR critical p = Left hem. on Right
Braskie et al., Journal of Neuroscience, May

10. Effect is even stronger for carriers of a
schizophrenia risk gene variant, trkA-T (N=391 people)
NTRK1 is a high affinity receptor for the neurotrophin NGF. We found that healthy T-allele carriers at rs6336 in the NTRK1 gene had lower FA broadly throughout the brain. In a recent meta-analysis, this allele was associated with a 1.64 times greater probability of developing schizophrenia in Caucasians versus the C allele . We found significant effects of NTRK1-T regions including the ILF, IFO, cingulum, and genu of the corpus callosum, which most consistently showed lower FA in schizophrenia patients versus controls in a recent meta-analysis (Ellison-Wright I, Bullmore E. Meta-analysis of diffusion tensor imaging studies in schizophrenia. Schizophr Res. 2009;108(1-3):3-10).
p values indicate where NTRK1 allele T carriers (at rs6336) have lower FA after adjusting for age, sex, and kinship in 391 young adults (31 T+; 360 T-).
FDR critical p =
b. Voxels that replicate in 2 independent halves of the sample (FDR-corrected). Left is on Right.
Braskie et al., Journal of Neuroscience, May 2012

11. We developed a polygenic test to predict your brain integrity
(7 SNPs) and rate of brain loss (empower drug trials)
Neuro-chemical genes
COMT
NTRK1
ErbB4
BDNF
Neuro-developmental genes
HFE
CLU
Neuro-degenerative risk genes
A significant fraction of variability in white matter structure of the corpus callosum (measured with DTI) is predictable from SNPs.
Kohannim O, et al. Predicting white matter integrity from multiple common genetic variants. Neuropsychopharmacology 2012, in press.

12. Brain measures are a good target for genetic analysis – may be easier to find genes that promote disease
difficult
easier

13. The way you do this is relatively simple, obtain a large group of subjects and for each person you measure a phenotype.
You then obtain DNA, find a specific location in the genome and determine the letter or genotype of each person. You then attempt to see if there is a relationship between the genotype and the phenotype. In this toy example, you can see there’s a clear additive effect of the A allele on intracranial volume. The more A alleles you have, the bigger your intracranial volume.

14. Finding Genetic Variants Influencing Brain Structure
CTAGTCAGCGCT
CTAGTCAGCGCT
Intracranial Volume …
CTAGTAAGCGCT
The way you do this is relatively simple, obtain a large group of subjects and for each person you measure a phenotype.
You then obtain DNA, find a specific location in the genome and determine the letter or genotype of each person. You then attempt to see if there is a relationship between the genotype and the phenotype. In this toy example, you can see there’s a clear additive effect of the A allele on intracranial volume. The more A alleles you have, the bigger your intracranial volume.
CTAGTAAGCGCT
CTAGTCAGCGCT
C/C
A/C
A/A
SNP
Phenotype
Genotype
Association

15. GRIN2b is over-represented in AD
First Genome-Wide Screens of Brain Images ( )
GRIN2b genetic variant is associated with 2.8% temporal lobe volume deficit;
The NMDA-type glutamate receptor is a target of memantine therapy; survives
genome-wide significance correction; detected with GWAS in N=742 subjects
GRIN2b is over-represented in AD
- could be considered an Alzheimer’s disease risk gene
- needs replication
Jason L. Stein1, Xue Hua PhD1, Jonathan H. Morra PhD1, Suh Lee1, April J. Ho1, Alex D. Leow MD PhD1,2, Arthur W. Toga PhD1, Jae Hoon Sul3, Hyun Min Kang4, Eleazar Eskin PhD3,5, Andrew J. Saykin PsyD6, Li Shen PhD6, Tatiana Foroud PhD7, Nathan Pankratz7, Matthew J. Huentelman PhD8, David W. Craig PhD8, Jill D. Gerber8, April Allen8, Jason J. Corneveaux8, Dietrich A. Stephan8, Jennifer Webster8, Bryan M. DeChairo PhD9, Steven G. Potkin MD10, Clifford R. Jack Jr MD11, Michael W. Weiner MD12,13, Paul M. Thompson PhD1,*, and the ADNI (2010). Genome-Wide Analysis Reveals Novel Genes Influencing Temporal Lobe Structure with Relevance to Neurodegeneration in Alzheimer's Disease, NeuroImage 2010.

16. Effect was later replicated in a younger cohort (Kohannim 2011)
GRIN2b (glutamate receptor) genetic variant associates with brain volume
in these regions; TT carriers have 2.8% more temporal lobe atrophy
Effect was later
replicated in
a younger cohort
(Kohannim 2011)
Jason L. Stein1, Xue Hua PhD1, Jonathan H. Morra PhD1, Suh Lee1, April J. Ho1, Alex D. Leow MD PhD1,2, Arthur W. Toga PhD1, Jae Hoon Sul3, Hyun Min Kang4, Eleazar Eskin PhD3,5, Andrew J. Saykin PsyD6, Li Shen PhD6, Tatiana Foroud PhD7, Nathan Pankratz7, Matthew J. Huentelman PhD8, David W. Craig PhD8, Jill D. Gerber8, April Allen8, Jason J. Corneveaux8, Dietrich A. Stephan8, Jennifer Webster8, Bryan M. DeChairo PhD9, Steven G. Potkin MD10, Clifford R. Jack Jr MD11, Michael W. Weiner MD12,13, Paul M. Thompson PhD1,*, and the ADNI (2010). Genome-Wide Analysis Reveals Novel Genes Influencing Temporal Lobe Structure with Relevance to Neurodegeneration in Alzheimer's Disease, NeuroImage, 2010.

17. Caudate association peak in PDE8B gene, replicates in 2nd young cohort (N=1198 people total)
Same Gene implicated
in Autosomal Dominant
Striatal Degeneration
-Phosphodiesterase = key protein in the dopamine signaling cascade

18. Replication through collaboration
>200 scientists, 12 countries; must have DNA and MRI scans
Many new members joining

19. Meta-Analysis – each site uploads its genome-wide scans - see if 500,000 genetic variants affect brain volume or brain wiring - each site’s “vote” depends on how many subjects they assessed

20. After Replication: N = 21,151; P = 6.7 x 10-16
4 Nature Genetics papers (April ) - largest brain imaging studies in the world
Chromosome 12 variant boosts brain volume by 9cc, IQ by 1.3 points
HP volume variant ~ 3 years of aging
FOREST
PLOT
Hippocampal Volume SNP – equivalent to ~3 years of aging; also HMGA-C carriers had 9cc bigger brains, +1.3 IQ points
Interestingly, when we were doing our association study in ~10,000 people we learned about another large consortium also conducting associations to hippocampal volume. In a giant consortium game of Go Fish, we sent them a list of our top hits. They sent us a list of their top hits. Remarkably, we both had the same SNP in the same direction as the most associated. This was incredible validation of our findings, showing for the first time truly significant association to brain structure phenotypes at a genome-wide level.
After Replication: N = 21,151; P = 6.7 x 10-16
(Stein authors, Nature Genetics, 2012)

25. Genome-Wide Screen of the Human Connectome discovers an Alzheimer risk gene (ENIGMA-DTI)
Discovery sample – Young Adults
Replicated sample – ADNI
Jahanshad/Thompson, under review

26. Autism Risk Gene linked to Differences in Brain Wiring CNTNAP2-CC Carriers have different networks Circles show hubs with different eccentricity (a measure of isolation; N=328 people)
To track brain changes, during brain development and disease, and in drug trials, we are developing powerful tensor-based morphometry approaches to detect brain structure differences, using mathematics adapted from continuum mechanics, i.e. the analysis of elastic and fluid deformations. We also develop non- linear image registration approaches, which fluidly deform 3D brain images to match each other. The structural differences between HIV/AIDS patients' brains and healthy normal subjects' brains are detected here (red colors; 3rd image) by computing flow fields that deform one brain scan onto another, measuring brain anatomical differences in 3D. The same approach can map brain growth (1st image) or Alzheimer's Disease (2nd image). The applied deformation vector fields, with billions of degrees of freedom, are designed to align images in exquisite detail by maximizing the mutual information between their intensity distributions. We are exploring other information-theoretic approaches for image registration based on generalized divergences between probability distributions, such as the Jensen-Renyi divergence and Kullback-Leibler divergence on material density functions. These deformation fields can be analyzed statistically with Lie group methods, and log-Euclidean metrics, that operate on the manifold of symmetric matrices. Related work is developing nonlinear registration methods for alignment, denoising and statistical analysis of diffusion tensor images (see below). We are also studying the statistics of deformation tensor fields, using the Hencky strain and other continuum-mechanical measures. Applications include mapping patterns of brain growth in children, and detecting drug effects on the brain. We are also developing new methods for the registration, segmentation, and analysis of DTI data (diffusion tensor images; shown in the right column here), using Riemannian manifold methods, tensor denoising, and information theory. One project investigates genetic influences on fiber architecture in the brain, using DTI data from twins
E. Dennis
et al. 2012

27. Acknowledgments *Jason Stein, UCLA
*Sarah Medland, QIMR, Brisbane, Australia
*Alejandro Arias Vasquez,
Netherlands
*Derrek Hibar, UCLA
Working Groups:
ENIGMA1
ENIGMA2
ENIGMA-DTI
ENIGMA-PGC
ENIGMA-MOUSE
NIH, European +
Australian Funding
Agencies
(Crowd-Sourcing)