Biomarkers and underlying mechanisms of
psychiatric diseases
Will discuss the DSM vs. RDoC approach to study mental health disorders
and the related biomarkers and include examples from my research.
(Reading: Cuthbert, B. N., & Insel, T. R. (2013). Toward the future of
psychiatric diagnosis: the seven pillars of RDoC. BMC medicine, 11(1), 126.)
Klára Marečková, Ph.D., M.Sc.
Burden of Diseases:
Disability-adjusted Life Years
US Burden of Disease Collaborators, 2013
Burden of Diseases:
Most Costly Conditions
The Global Economic Burden of Noncommunicable Diseases, WEF, 2011
The Most Disabling Disorders before Age 50
Sex Differences in Prevalence of
Disorders of the Brain
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
Current Diagnostics in Mental Health
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
➢ Mental disorders are disorders of brain
circuits, focus on brain-based
biomarkers, whole biosignatures
➢ Understand the etiology through
large population-based studies,
understand the mechanisms
➢ Focus on early detection and
prevention (not extreme comparisons
of patients vs. controls, study both
clinical and pre-clinical populations to
identify early biomarkers, use
standardized methods for
replication)
➢ Move towards personalized medicine
and sex-dependent therapeutics
(take into account one’s sex, genes
etc.)
Better Treatment through Better Diagnostics
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
➢ Mental disorders are disorders of brain
circuits, focus on brain-based
biomarkers, whole biosignatures
➢ Understand the etiology through
large population-based studies,
understand the mechanisms
➢ Focus on early detection and
prevention (not extreme comparisons
of patients vs. controls, study both
clinical and pre-clinical populations to
identify early biomarkers, use
standardized methods for
replication)
➢ Move towards personalized medicine
and sex-dependent therapeutics
(take into account one’s sex, genes
etc.)
Better Treatment through Better Diagnostics
Precision Medicine for Mental Disorders
• 2011 US National Academy of Sciences published major report on precision medicine
(more individualized treatment, based upon precise specification of the genetic, molecular and
cellular aspects of disease)
• For example, in oncology, analysis of genetic variants is used to predict what treatment will be
optimal
Precision Medicine for Mental Disorders
• 2011 US National Academy of Sciences published major report on precision medicine
(more individualized treatment, based upon precise specification of the genetic, molecular and
cellular aspects of disease)
• For example, in oncology, analysis of genetic variants is used to predict what treatment will be
optimal
• In contrast to other areas of medicine, the field of mental disorders research lags badly behind
the rest of medicine in moving towards precision medicine
• One syndrom such as MDD involves multiple mechanisms – HPA dysfunction, reward-seeking,
emotion regulation, modulatory neurotransmitter systems, epigenetic marks
• One mechanism (e.g. fear circuits or working memory) implicated in multiple disorders
• DSM and ICD categories do not map well onto emerging findings from genetics, systems
neuroscience and behavioral science – so it’s impossible to translate research findings into
systematic understanding of pathology and treatment directed at the mechanisms
• Biological findings that did not map on the current heterogeneous DSM categories of symptom
clusters were essentially excluded
Precision Medicine for Mental Disorders
• 2011 US National Academy of Sciences published major report on precision medicine
(more individualized treatment, based upon precise specification of the genetic, molecular and
cellular aspects of disease)
• For example, in oncology, analysis of genetic variants is used to predict what treatment will be
optimal
• In contrast to other areas of medicine, the field of mental disorders research lags badly behind
the rest of medicine in moving towards precision medicine
• One syndrom such as MDD involves multiple mechanisms – HPA dysfunction, reward-seeking,
emotion regulation, modulatory neurotransmitter systems, epigenetic marks
• One mechanism (e.g. fear circuits or working memory) implicated in multiple disorders
• DSM and ICD categories do not map well onto emerging findings from genetics, systems
neuroscience and behavioral science – so it’s impossible to translate research findings into
systematic understanding of pathology and treatment directed at the mechanisms
• Biological findings that did not map on the current heterogeneous DSM categories of symptom
clusters were essentially excluded
• 2009 NIMH RDoC project - research classification system, NIMH’s effort to develop a precision medicine
approach for mental disorders
• New ways of classifying mental disorders based on dimensions of observable behavior and
neurobiological measures
• Neural circuits and systems are a critical factor in how the brain is organized and functions, and how
genetics and epigenetics exert their influence
• Studying full range of variation, from normal to abnormal
Morris, Bruce, & Cuthbert, 2012
! Orthogonal dimensions: Developmental aspects, Environmental aspects
https://www.ted.com/talks/thomas_insel_toward_a_new_understanding_of_mental_illness
Psychosis Depression
Maladaptive
responses
to negative
stimuli
Mood dysregulation
NIMH Research Domain Criteria (RDoC)
15 F and 16 M with PSY
14 F and 13 M with MDD
19 F and 22 M were CTRL
❖ 99 participants scanned on 3T Siemens scanner
Functional Magnetic Resonance Imaging
CPP
CPP
• a cohort of 17 000 pregnant
mothers and their offsprings, born
between 1959 – 1966, who were
followed up
➢ 1/month during the first 7 months
of pregnancy
➢ 2/month during the 8th month of
pregnancy
➢ 1/week during the 9th month and
neonatal stage
➢ at 4, 8, and 12 months
➢ at 4 and 7 years
The final assessment at the age of 7
was completed by 80 % of the initial
cohort.
CPP
❖ In 2000, Dr. Stephen Buka along with
Dr. Jill M. Goldstein, Larry J. Seidman
and Ming T. Tsuang started New
England Family Study (NEFS), aimed
to follow up the adults from CPP cohort.
❖ They successfully located cca 85% of
the CPP sample.
❖ Over 90% of the successfully located
individuals decided to participate in
further studies focused on prenatal and
early life antecedents of
neuropsychiatric, physical and
behavioral conditions such as
schizophrenia, substance use, heart
disease, learning disabilities, attention
deficit disorder, depression or suicide.
❖ Actual study sites are located in
Boston, MA (Harvard Medical School,
Harvard School of Public Health,
Massachusetts General Hospital and
Massachusetts Mental Health Center)
and Providence, RI (Brown University).
ORWH-NIMH P50 MH082679; NIMH R01s MH56956, MH074679, MH090291
Future follow-ups…In utero Birth 46-547
New England Family Studies: 50 year follow-up
N = 17,741
1959 - 1966
Childhood
Assessments
up to age 7
ADULT PHENOTYPING /
BIOMARKERS
• Clinical & Cognitive Assts.
• Structural and functional Imaging
• Hormonal Assessments
• EKG & cardiometabolic exams
• Blood and cell collection for
genetics/genomics/transcriptomics
33-45
N = 17,741
1959 - 1966
Maternal
prenatal sera
assayed for early
immune and
hormonal
biomarkers
15 F and 16 M with PSY
14 F and 13 M with MDD
19 F and 22 M were CTRL
❖ 99 participants scanned on 3T Siemens scanner
Functional Magnetic Resonance Imaging
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
+ 4 Repetitions of 90s Blocks
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
Run 1
(6 min)
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
Run 2
(6 min)
Run 3
(6 min)
+ 4 Repetitions of 90s Blocks
+ 4 Repetitions of 90s Blocks
Negative affective stimuli from IAPS
(Mild Visual Stress Task)
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
+ 4 Repetitions of 90s Blocks
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
Run 1
(6 min)
Fixation
(30s)
Negative
(30s)
Neutral
(30s)
Run 2
(6 min)
Run 3
(6 min)
+ 4 Repetitions of 90s Blocks
+ 4 Repetitions of 90s Blocks
Negative affective stimuli from IAPS
(Mild Visual Stress Task)
Time 30, 60, 90 Blood Draws
Time 15 Blood Draw
with Real-Time Hormone Response
Time 00 Blood Draw
Stress Response Circuitry
▪ Brainstem
▪ Amygdala
▪ Hypothalamus
▪ Hippocampus
▪ Ant. Cingualte
▪ Medial and Orbitofrontal Cortex
Stress Response Circuitry And The HPA Axis
Larger in the female brain, relative to cerebrum size
Larger in the male brain, relative to cerebrum size
Goldstein et al., 2001
Sex Differences in The Healthy Human Brain
Mareckova et al, Human Brain Mapping, 2016
BOLD response to negative affective stimuli
▪ n= 99 (31 PSY; 27 MDD; 41 HC, equally divided by sex, women in midcycle)
FWE p<0.05
Dysphoric mood predicted increased BOLD in HYPO
and AMYG in response to negative affective stimuli
Mareckova et al, Human Brain Mapping, 2016
R2=0.19
R2=0.07
Dysphoric mood and sex predicted
regulation of stress response
Mareckova et al, Human Brain Mapping, 2016
R2=0.23
R2=0.19
R2=not sig.
R2=0.42
R2=not sig.
R2=0.37
Elevated cortisol response predicted lower activity
in OFC and low HYPO-AMYG connectivity
z=3.36, FWE p=0.01
In women, elevated cortisol response predicted lower
activity in mPFC and low HYPO-HIPP connectivity
z=4.39, FWE p=0.001
• Even under mild stress, females with more severe mood symptomatology
were unable to regulate arousal by inhibitory regions. This might possibly
explain why are females at higher risk to develop mood disorders.
Conclusions
• We demonstrated a transdiagnostic impact of cortisol response to mild
visual stress on brain function in the stress circuitry (independent of
diagnosis or medication).
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
➢ Mental disorders are disorders of brain
circuits, focus on brain-based
biomarkers, whole biosignatures
➢ Understand the etiology through
large population-based studies,
understand the mechanisms
➢ Focus on early detection and
prevention (not extreme comparisons
of patients vs. controls, study both
clinical and pre-clinical populations to
identify early biomarkers, use
standardized methods for
replication)
➢ Move towards personalized medicine
and sex-dependent therapeutics
(take into account one’s sex, genes
etc.)
Better Treatment through Better Diagnostics
Schizophrenia
Fetal/Birth Risk Factors Relative Risk
Low birthweight 1.6 – 3.9
Obstetric complicaitons
(preeclampsia, hypoxia)
2.4 – 2.8
Maternal malnutrition 2.0
Maternal herpes simplex 3.4 – 4.4
Maternal rubella 5.2
Maternal hypertension 2.6
Maternal diabetes 5.4
Depression
Fetal/Birth Risk Factors Relative Risk
Low birthweight 3.5
Obstetric complicaitons
(preeclampsia, hypoxia)
1.4
Maternal malnutrition 1.6
Influenza 1.7
Buka et al., 1998; 2002; Susser et al. 1996;
Sacker, 1995; Dalman et al., 1999;
Hultman et al., 1999; Brown et al. 2000;
Cannon, 2002; Goldstein et al, 2012
Developmental Risk Factors
Kinney, 1998; Sacker, 1995;
Preti, 2000; Bellingham-Young, 2003;
van Os, 1997; Machon, 1997;
Brown, 2000; Costello et a.,2007
Howerton & Bale, 2012
Sex Effects in Timing of Fetal Risk Factors
for Psychiatric Disorders
• Mid-gestation: Active period of sexual differentiation; e.g., testes begin to
secrete T
• Testosterone: Direct effects and indirect effects through aromatization into
estradiol
• Estradiol & Androgens: Major impact on neuronal growth and development
Organizaitonal Effects of Gonadal Steroids
on Fetal Brain Development
Risk for Mood Disorders
(Depression, Psychoses)
Altered Adult
HPA/HPG
Hormones
Altered Adult Stress Response
Circuitry
Prenatal maternal immune programming
of offspring adult stress response circuitry
Prenatal Stress
Maternal
Intrauterine
Disruptions
2nd, 3rd trim. (e.g.
cytokines as coactivators
of HPA
response)
Sex
Altered Fetal
HPA/HPG &
Fetal Brain
Altered Adult
HPA/HPG
Hormones
Altered Adult Stress Response
Circuitry
Prenatal maternal immune programming
of offspring adult stress response circuitry
Risk for Mood Disorders
(Depression, Psychoses)
Goldstein et al., 2014
Prenatal stress model
• We are testing hypothesis of shared etiologies to understanding sex differences
in MDD and SCZ.
• Focus on maternal TNF-, IL-1, and IL-6, pro-inflammatory cytokines, primary
co-activators of HPA response, whose receptors are located in the sex-specific
regions of the stress-response circuitry
• Focus on the 2nd and 3rd trimestr , as the period of sexual differentiation of the
brain.
• Are disruptions in fetal hormonal programming associated with sex differences
in stress circuitry and endocrine function and development of depression and
psychoses in adulthood?
Fetal Hormonal Programming of Sex
Differences in Depression and Psychoses
Future follow-ups…In utero Birth 46-547
N = 17,741
Maternal
prenatal sera
assayed for early
immune and
hormonal
biomarkers
Childhood
Assessments
up to age 7
ADULT PHENOTYPING / BIOMARKERS
• Clinical & Cognitive Assts.
• Structural and functional Imaging
• Hormonal Assessments
• EKG & cardiometabolic exams
• Blood and cell collection for
genetics/genomics/transcriptomics
33-45
New England Family Study: 50 year follow-up
ORWH-NIMH P50 MH082679 (depression); NIMH RO1 MH56956 (psychoses)
Maternal Immune Activity and Sex-Dependent
Risk for Psychosis in The Offspring
Gilman et al., under review
Maternal Immune Activity And Sex-Dependent
Risk for MDD in The Offspring
Mean%sig.changeinHYPO
Mean%sig.changeinRAMYG
Mean%sig.changeinLHIPP
High TNFa Low TNFa High TNFa Low TNFa High TNFa Low TNFa
t(79)=1.94, p=0.05, R2=0.05 t(78)=2.04, p=0.045, R2=0.05 t(79)=2.1, p=0.04, R2=0.05
L HIPP
z=-15
Low Exposure to TNFa Prenatally Associated
with Hyperactive Subcortical Regions
Prenatal exposure to TNFa
Mean%sig.changeinHYPO
All: t(76)=-2.22, p=0.03, R2=0.06
z=-15
Low Exposure to TNFa Prenatally Predicts
High BOLD in HYPO and Low BOLD in mPFC
Prenatal exposure to TNFa:IL10
Mean%sig.changeinLHIPP
Females:
t(39)=-2.59, p=0.01, R2=0.15
Males:
t(39)=2.13, p=0.04, R2=0.11
L HIPP
z=-15
Sex Differences in The Effects of Prenatal
Exposure to TNFa:IL10 on Stress Circuitry
• Prenatal stress-immune pathways predict brain function and vulnerability to
psychiatric disorders 50 years later.
Conclusions
Novel transcriptome-based polygenic risk score for
depression predicts brain function during face
processing
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
➢ Mental disorders are disorders of brain
circuits, focus on brain-based
biomarkers, whole biosignatures
➢ Understand the etiology through
large population-based studies,
understand the mechanisms
➢ Focus on early detection and
prevention (not extreme comparisons
of patients vs. controls, study both
clinical and pre-clinical populations to
identify early biomarkers, use
standardized methods for
replication)
➢ Move towards personalized medicine
and sex-dependent therapeutics
(take into account one’s sex, genes
etc.)
Better Treatment through Better Diagnostics
Verifying these biosignatures
of dysregulated stress circuitry
in a pre-clinical sample of
typically developing young
adults from another birth
cohort,
for whom we’ve recently
collected rich neuroimaging
and biosmecimen data.
Next Steps
European Longitudinal Study
of Pregnancy and Childhood (ELSPAC)
Pediatricians
Parents/child
Teachers
Number of participating families
Number of pediatric reports according to
child’s gender and birth complications
M
F
Number of pediatric reports according to
child’s gender and birth complications
M
F
No birth complications
Birth complications
Next steps
• Recruit 120 participants (60 males, 60 females) who previously participated in the
European Longitudinal Study of Pregnancy and Childhood (ELSPAC), collect
• and link these with their longitudinal data
from pre/peri-natal period and adolescence.
MRI, fMRI, rs-fMRI Cortisol, Sex hormones, Genes Questionnaires etc.
• Basic demographics data, medications, drug use
• Depression and Anxiety-related questionnaires (BDI, MFQ, STAI, POMS)
• Neuroimaging data (T1, T2 FLAIR, task fMRI, rs fMRI, DTI)
• Physiological data (ECG, breathing, skin conductance; during task fMRI and
rs fMRI)
• Anthropometric & cardio data (blood pressure, heart rate, BMI,
bioimpedance, subcut. fat)
• Hormonal data (salivary cortisol, testosterone, estrogen, progesterone),
taken by passive drool between 8-9 am, females in late follicular phase (days
10-15 from onset of menstruation), 37% on oral contraception, all analyzed
by ELISA kits
• Buccal swaps (will be analysed with genetic and epigenetic chips)
• Longitudinal ELSPAC data (questionnaires from MDs, parents, teachers,
child; prenatal till 19)
VULDE data overview (n=131)
▪ 131 participants recruited from the ELSPAC cohort
▪ 61 males, 70 females
▪ All 23 or 24 years old when VULDE testing (MRI, saliva etc.)
▪ All White Caucasians of European origin, from South Moravia,
Czech Republic
▪ Females tested during the high estrogen phase (menstrual
cycle day 10-15)
▪ 26 out of the 70 females were on oral contraception
▪ Majority of them are university students
Demographics
• Can we observe structural and functional changes among VULDE
participants who had more depressive symptoms?
• Could we use the altered structure and function in these regions as
a potential early biomarker of depressive symptomatology in young
adults?
• Would this biomarker differ among males and females?
• Would it be related to prenatal stress or depression during
adolescence?
VULDE research questions
Design of the negative affect fMRI task
(International Affective Picture System)
Jacobs et al., 2015
ROI Direction
of the sig.
relationsh
ip
R2
PAG M>F 0.03
HYPO - L
AMYG - R
AMYG M>F 0.03
L HIPP M > F 0.04
R HIPP M>F 0.06
L sgACC M>F 0.12
R sgACC M>F 0.08
L dpgACC - R
dpgACC M>F 0.04
L OFC - R
OFC - L
mPFC M>F 0.10
R mPFC M>F 0.07
➢ VULDE IAPS fMRI data processed using the
same pipeline as in the Mareckova et al (HBM,
2016) paper on NEFS
➢ extracted BOLD response from the Negative >
Netural contrast in the hypothesized ROIs
(Mareckova et al, HBM, 2016).
➢ MANOVA showed a significant ROI*Sex
interaction (F=4.71, p<0.0001).
➢ Posthoc analyses showed that men had greater
BOLD to negative affective stimuli than women in
PAG, R AMYG, ACC, HIPP, mPFC (viz table on
the left).
Sex differences in BOLD response to negative affect
Sex differences in BOLD response to negative affect
✓ Goldstein et al (2005) paper demonstrating sex differences in the physiology of
the brain but not in the subjective feelings.
✓ Goldstein et al (2010) paper demonstrating larger BOLD response in the stress
circuitry in men vs. mid-cycle women.
Sex differences in BOLD response to negative affect
✓ VULDE women scanned on
days 10-15; NEFS women
scanned on days 10-14
✓ Both studies demonstrated
women’s lower BOLD
response to negative affective
stimuli in the subcortical
arousal regions (PAG, AMYG)
as well as HIPP and cortical
regions that control the arousal
(ACC, mPFC).
✓ Both studies found the largest
effect sizes in mPFC.
Goldstein et al., Journal of Neuroscience, 2010
✓ Goldstein et al (2005) paper demonstrating sex differences in the physiology of
the brain but not in the subjective feelings.
✓ Goldstein et al (2010) paper demonstrating larger BOLD response in the stress
circuitry in men vs. mid-cycle women.
• We tried to replicate the clinical findings on the effects of sex and depressive
symptomatology on brain function during negative affect (Goldstein et al, 2005;
Goldstein et al, 2010; Mareckova et al, 2016) in typically developing young adults
from a prenatal birth cohort.
• We conclude that men demonstrated higher BOLD response to negative affect than
women and that these sex differences were particularly pronounced in individuals
with more depressive symptomatology.
• These findings suggest that during mild stress, men might recruit the regulatory
regions to a greater extent than women, possibly explaining why are women
more vulnerable to depression than men.
• The fact that there were no sex differences in the demographics or the mood-related
variables suggests that vulnerability to mood disorders might be detectable at
the level of the brain well before it might manifest in the behavior.
Conclusions
Baseline cortisol in clinical samples (NEFS)
F(2,47)=3.59, p=0.04, R2=0.13
Cortisol in men Cortisol in women
F(2,44)=3.45, p=0.04, R2=0.14
CTRL MDD SCZ CTRL MDD SCZ
Hormonal biomarker
of vulnerability to depression?
Cortisol & Anxiety trait in men
p=0.75
Hormonal biomarker
of vulnerability to depression?
Cortisol & Anxiety trait in women
t(69)=-2.14, p=0.036, R2=0.06
Cortisol & Anxiety trait in men
p=0.75
OC
Free
t(128)=-2.46, p=0.02, R2=0.05 t(128)=-2.07, p=0.04, R2=0.03
STAI-T & Overall GM STAI-T & Overall GM/brain size
Volume of Grey Matter (GM)
• MANOVA: STAI-T: F=5.28, p=0.02; Lobe: F=97.96, p<0.001; STAI-T*Lobe: F=4.79, p=0.003
Volume of Grey Matter in Different Lobes
t(129)=-2.56
p=0.01
R2=0.05
t(129)=-1.91
p=0.06
R2=0.03
t(129)=-1.86
p=0.07
R2=0.03
t(129)=-1.79
p=0.08
R2=0.02
Frontal Lobe Parietal Lobe
Temporal Lobe Occipital Lobe
• MANOVA: STAI-T: F=5.28, p=0.02; Lobe: F=97.96, p<0.001; STAI-T*Lobe: F=4.79, p=0.003
Volume of Grey Matter in Different Lobes
Cortical Volume = Product of Cortical
Thickness & Area
Cortical Volume = Product of Cortical
Thickness & Area
p=0.93
Overall cortical thickness & STAI-T
Cortical Volume = Product of Cortical
Thickness & Area
p=0.93
Overall cortical thickness & STAI-T Overall cortical area & STAI-T
t(128)=-1.98, p=0.049, R2=0.03
Cortical Volume = Product of Cortical
Thickness & Area
Mareckova et al, Cerebral Cortex, 2018
Prenatal stress and gray matter volume in young
adulthood
Prenatal stress and mood in young adulthood
Jensen et al., 2015
Metaanalysis of hypometabolic ROIs
in depressed patients vs. healthy controls
Mareckova et al, Cerebral Cortex, 2018
Prenatal stress and GM volume in young adulthood
Mareckova et al, Scientific Reports, 2018
Perinatal stress and human hippocampal volume:
Findings from typically developing young adults
Mareckova et al, Scientific Reports, 2018
Perinatal stress and human hippocampal volume:
Findings from typically developing young adults
o Diagnosis limited to symptoms
(but symptoms are late manifestations
of brain disorders)
o Etiology unknown
(treatments for symptoms not cures)
o Detection late; for most disorders,
prevention not well developed
➢ Mental disorders are disorders of brain
circuits, focus on brain-based
biomarkers, whole biosignatures
➢ Understand the etiology through
large population-based studies,
understand the mechanisms
➢ Focus on early detection and
prevention (not extreme comparisons
of patients vs. controls, study both
clinical and pre-clinical populations to
identify early biomarkers, use
standardized methods for
replication)
➢ Move towards personalized medicine
and sex-dependent therapeutics
(take into account one’s sex, genes
etc.)
Better Treatment through Better Diagnostics
Thank you!