Prediction of mood disorders and imaging phenotypes from genomic data

Prediction of mood disorders and imaging phenotypes from genomic data: findings in individuals at familial risk of mood disorder.

Dr Jess Sussmann

Wellcome Trust Research Fellow University of Edinburgh

No disclosures

Introduction • Mood disorders (BD and MDD), cross over of risk

• Characterised by deficits in emotion regulation

• Heritable with complex genetic architecture.

• Currently limited understanding of mechanisms and no biomarkers to guide diagnosis

• Key goal of neuroimaging is to identify objective neurobiological markers of illness in order to:

• Enhance understanding of aetiology

• Increase diagnostic precision

• Identify markers of risk

NEUROBIOLOGY

Langan & McDonald Molecular Psychiatry 2009:14:833; Phillips & Swartz American Journal of Psychiatry 2014

In bipolar: “Disruption of higher order cognitive control in

association with increased responsiveness of brain regions

involved in emotional regulation”

• Are abnormalities reported in patients secondary consequence of

illness, medication and current mood/state effects?

• What is the neurodevelopmental trajectory? Are they seen:

• early on in the disorder?

• prior to illness?

• unaffected relatives

• related to genetic risk?

Longitudinal familial high risk

studies

SCOTTISH BIPOLAR FAMILY STUDY

Study design

HC

well

HR

well

HR

MDD

HR

BD

HR

other

HC

ill

HC (no info)

HR (no info)

154

114 30 2 - 8

87 39 4 2 (22) 77

111

123

5

11

7

(35)

Baseline: imaging,

behaviour,

clinical

Visit 2: imaging,

behaviour,

clinical

Visit 3: behaviour,

clinical

Healthy controls Bipolar High Risk

RESULTS

BASELINE FINDINGS: Controls versus High Risk groups (unaffected)

Individual genetic risk loci

Cumulative genetic loading

PREDICTORS of illness: can those who become ill be distinguished

from those who remain well from BASELINE data

BASELINE FINDINGS

BASELINE: White matter integrity

Sprooten et al Biological Psychiatry 2011, 70: 350

corpus callosum, anterior limb internal capsule*, inferior* and superior longitudinal fasiculi,

uncinate fasciculus*

Between Group (Controls v HR) DECREASES IN HR

-ve correlation with mood instability

BASELINE Functional imaging

Whalley et al Biological Psychiatry 2011, 70: 343

Between Group (Controls v HR) INCREASES IN HR

+ve correlation with mood instability

Baseline summary

• Neuroimaging differences seen in bipolar high risk individuals versus controls at baseline:

• decreased white matter integrity in connections between emotion processing and emotion regulation networks with increased responsivity of limbic regions, incl striatum (mesolimbic)

Qu: •Do any of these differences relate specifically to genetic risk? •Do any abnormalities predict illness?

GENETIC IMAGING: individual loci

Genetic imaging: individual risk loci

DGKH

Diacylglycerol kinase eta: Identified in two independent GWAS as susceptibility gene for BD (Baum et al, 2008). DGKH involved in pathway through which lithium is thought to exert its therapeutic effects

Anterior cingulate: Greater activation in limbic regions

in risk carriers

Whalley et al Neuropychopharm 2012, 37: 919

DGKH

GENETIC IMAGING: cumulative loading

Cumulative risk scores

Evidence that a large number of common low penetrant causal variants contribute to overall risk The cumulative effect of these variants can be estimated using polygenic profiling Polygenic risk score (PGRS)* consists of the weighted sum of associated alleles derived from an independent dataset: eg Psychiatric Genomics Consortium

* Ripke et al Mol Psych 2012, 18: 497

PGRS: white matter

Whalley et al Biological Psychiatry 2013, 74: 280

Widespread decreases in integrity w increasing risk (controlling for group):

Peak in superior longitudinal fasciculus, inf longitudinal fasc, inferior fronto-

occip fasc.

PGRS: functional imaging

Whalley et al Translational Psychiatry 2012 Jul 3;2:e130.

Increases in activation w increasing risk (controlling for group):

Peak in anterior cingulate and amygdala

Genetic imaging summary

• increased activation of limbic regions (anterior cingulate and medial temporal lobe) in association with increased risk

• decreases in white matter integrity in widespread regions including cortico-limbic connections involved in emotional regulation

•Do any abnormalities predict illness?

PREDICTORS OF ILLNESS

Structural findings

Decreased cortical thickness at baseline in PHG in HR who develop MDD

Martina Papmeyer (PhD thesis)

Functional findings

Whalley et al PlosONE 2013 March 3:e57357

Conclusions RISK ASSOCIATED:

• increased responsivity of limbic structures (esp amygdala

and anterior cingulate) and decreased white matter

integrity in

• unaffected relatives

• AND in relation to genetic risk (loci and PGRS)

PREDCITIVE MARKERS:

Early findings suggests

• thinning of cortical regions in medial temporal lobe

• increased activation of insula/inferior frontal

- Increased responsivity in limbic regions reflecting genetic

predisposition, in conjunction with prefrontal cortical

abnormalities results in illness

Langan & McDonald Molecular Psychiatry 2009:14:833; Phillips & Swartz American Journal of Psychiatry 2014

In bipolar: “Disruption of higher order cognitive control in

association with increased responsiveness of brain regions

involved in emotional regulation”

Future directions • effective connectivity techniques to determine

direction of interactions within emotional circuit

• biological pathway focussed polygenic risk scores: neurobiol (synaptogenesis)

biological (Ca2+, glu rec signalling)

pharmacological (5HT DA)

• computerised pattern classification to assess

predictive markers

• new wave of data collection commencing summer

2014; further clinical assessment (remain MDD or convert to BD?)

Acknowledgements

Andrew McIntosh

Heather Whalley

Jeremy Hall

Stephen Lawrie

Liana Romaniuk

Eve Johnstone

Tiffany Stewart

Alix McDonald

Lee Murphy

Wellcome Trust

Health Foundation

Royal Society

Brain and Behaviour Research

Foundation

Scottish Mental Health Research

Network

Sackler Foundation

BRIC (Brain Research Imaging

Centre)

SINAPSE

+ thanks and acknowledgements to the patients and their families for taking part