Upload
truonglien
View
225
Download
0
Embed Size (px)
Citation preview
Maternal care and Gene - Environment
Interactions Defining Development
Michael J Meaney
James McGill Professor
Dept. Psychiatry
McGill University
Douglas Hospital Research Centre
The development of an individual is an active process
of adaptation that occurs within a social and economic
context:
• To resource (food, shelter, safety) availability.
• To social interactions.
• To independence from the parent.
Early experience
Abuse
Family strife
Emotional neglect
Harsh discipline
Health Risks
Depression
Drug abuse
Anxiety
Diabetes
Heart disease
Obesity
Mechanism?
Developmental Origins of Adult Disease
Early experience
Abuse
Family strife
Emotional neglect
Harsh discipline
Health Risks
Depression
Drug abuse
Anxiety
Diabetes
Heart disease
Obesity
Individual differences
in neural and endocrine
responses to stress (defensive responses)
Stress Diathesis Models
Early experience
Abuse
Family strife
Emotional neglect
Harsh discipline
Health Risks
Depression
Drug abuse
Anxiety
Diabetes
Heart disease
Obesity
Individual differences
in neural and endocrine
responses to stress
Poverty
Early experience
Abuse
Family strife
Emotional neglect
Harsh discipline
Depression
Drug Abuse
Anxiety
Obesity
Diabetes
Heart Disease
Poverty
Effects of poverty on emotional and cognitive development are
mediated by parental factors (Conger, McLloyd, Eisenberg).
Poverty
Maternal
emotional
distressParenting
Child
behavioural
problems
Linver, Brooks-Gunn & Kohen Dev Psychol 2002; same model predicts child cortisol
levels (Lupien, McEwen, Meaney Biol Psychiatry 2000)
Nutrient Supply
Mate Quality
Violence
Infection
Population density
Parental
investment
Defensive Strategies
Foraging/Metabolism
Reproductive Strategies
Environmental Parental Developmental
signal mediation outcome
Robert Hinde: Evolution has shaped the young to use parental
signals as a „forecast‟ of the quality of the environment
into which they have been born. For most species, there is
no single, optimal phenotype.
Evolutionary biology - Maternal effects
Defense to snake predation in skink lizards
Most frequent prey
• smaller
• shorter tails
• less reactive to
snake cues
If mother has been exposed to the scent of a
predatory snake then offspring are larger, with
longer tails and ….
Control Perfume Snake0
1
2
3
4
Response to snake odours
To
ng
ue
-Flic
k
Offspring
are significantly
larger and with
longer tails
Nutrient Supply
Mate Quality
Violence
Infection
Population density
Parental
investment
Defensive Strategies
Foraging/Metabolism
Reproductive Strategies
Environmental Parental Developmental
signal mediation outcome
Robert Hinde: Evolution has shaped the young to use parental
signals as a „forecast‟ of the quality of the environment
into which they have been born. For most species, there is
no single, optimal phenotype.
Evolutionary biology - Maternal effects
How might the parent-child relations that define early family
life influence the development of individual differences in
brain development and function?
Summary
• Parental care affects the activity of
genes in the brain that regulate stress
responses, neural development and
reproduction.
• This parental effect involves a form
a “plasticity” at the level of the DNA.
Epigenetics: Any functional change in the genome that
does not involve an alteration of DNA sequence.
Every cell in your body has the same nuclear genes, but…?
Multiple phenotypes from a common genotype
Epigenetics
Any functional change to the genome that
does not involve an alteration of DNA sequence.
Epigenetics is the biological basis for gene x
environment interactions.
C T A C G T A C T C G G A A T C T C GCH
3CH
3CH
3
C T A C G T A C T C G G A A T C T C G
Genetic code is defined by the sequence of four nucleotides that
produce proteins and other molecules that serve cell function.
RNAs, Protein
Epigenetic effects refer to modifications of the chemistry of the DNA,
but not to a change of sequence. Epigenetics alters the activity of the
gene, but not its function.
C T A C G T A C T C G G A A T C T C G
CH3
CH3CH
3
• Epigenetic effects refers to modifications of the
chemistry of the DNA, but not to a change of sequence.
• Epigenetics alters the activity of the gene, but not its function.
C T A C G T A C T C G G A A T C T C G
CH3
CH3CH
3
• Epigenetic effects refers to modifications of the
chemistry of the DNA, but not to a change of sequence.
• Epigenetics alters the activity of the gene, but not its function.
• DNA methylation: The addition of a methyl group onto a cytosine.
C T A C G T A C T C G G A A T C T C G
CH3
CH3CH
3
• Epigenetic effects refers to modifications of the
chemistry of the DNA, but not to a change of sequence.
• Epigenetics alters the activity of the gene, but not its function.
• DNA methylation: The addition of a methyl group onto a cytosine.
• DNA methylation is chemically very stable (potentially lasting for
the life of the organism).
C T A C G T A C T C G G A A T C T C G
CH3
CH3CH
3
• Epigenetic effects refers to modifications of the
chemistry of the DNA, but not to a change of sequence.
• Epigenetics alters the activity of the gene, but not its function.
• DNA methylation: The addition of a methyl group onto a cytosine.
• DNA methylation is chemically very stable (potentially lasting for
the life of the organism).
• DNA methylation silences gene expression.
Nucleosome core particle: ribbon traces for the 146-bp DNA phosphodiester backbones
(brown and turquoise) and eight histone protein chains (Luger et al. Nature 1997).
+ -
Prevents TF
binding to DNA
TF binding involves
alteration of
chromatin structure
C T A C G T A C T C G G A A T C T C G
CH3CH
3CH3
C T A C G T A C T C G G A A T C T C G
Genetic code is defined by the sequence of four nucleotides that
produce proteins and other molecules that serve cell function.
RNAs, proteins
Epigenetic effects refer to modifications of the chemistry of the DNA,
but not to a change of sequence. Epigenetics alters the activity of the
gene, but not its function.
SIN
SIN
CH3
CH3
CH3
CH3 MeCP2
HDACCH3
CH3
MeCP2
DNA Methylation can inhibit gene expression
by blocking transcription factors binding
HDAC HDAC: Histonedeacetylase
Methylated DNA
binding protein
Naturally-occurring
variations in maternal care
Expression of specific genes
in brain regions
HPA function
Stable individual differences
in stress reactivity
Maternal licking/grooming
Source of tactile stimulation/nurturance: Enhances
activity of endocrine systems (e.g., GH/IGF) that
promote somatic growth, suppresses those
(glucocorticoids) that inhibit growth
Variations in maternal care
Fre
qu
en
cy c
ou
nt
% Licking/grooming
Low LG
Mean - 1SD
High LG
Mean + 1SD
- - - - -
HighLow
Days of
Age
Lic
kin
g/g
roo
min
g
Variations in maternal care X Days
* No differences
in time with pups
864206
8
10
12
14
16
18
High/Stress
High/Ctl
Low/Stress
Low/Ctl
% L
ickin
g/g
room
ing
Day
Gestational stress/maternal behaviour
Popeski et al: Pictures of „own‟ infant increase activity
in ventral striatum (nucleus accumbens – dopamine target)
by comparison to other infant or neutral stimuli in fMRI
studies – this effect is abolished by chronic stress.
Comparable effects in humans
Are these naturally-occurring variations
in maternal behaviour associated with
the development of individual differences
in endocrine and behavioural responses
to stress?
* Effects hold for both males and females
fetal postnatal peripubertal adulthood
Maternal care Outcomes
GlucocorticoidReceptor
Hypothalamus
Pituitary
Adrenals
Hippocampus
Glucocorticoids
Stress
(-)
CRF
ACTH
CRF: corticotropin releasing factor. ACTH: adrenocorticotropin
Hypothalamus
Pituitary
Adrenals
Hippocampus
Glucocorticoids
(-)
CRF
ACTH
(-)
Hypothalamus
Pituitary
Adrenals
Hippocampus
Glucocorticoids
(-)
CRF
ACTH
(-)
Individual differences in glucocorticoid receptor
levels lead to altered pituitary-adrenal responses
to stress
Hypothalamus
Pituitary
Adrenals
High LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Hypothalamus
Pituitary
Adrenals
Low LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Individual differences in glucocorticoid receptor
levels lead to altered pituitary-adrenal responses
to stress
Adult offspring of High LG mothers show
more modest HPA responses to stress
Pre S10 S20 P20 P40 P60 P1200
200
400
600
800
1000High LGLow LG
Pla
sm
a A
CT
H (
pg
/ml)
Pre 10 20 40 60 80 1000
10
20
30
40
50
60
70
80High LGLow LG
Cort
icoste
rone (
µg/d
l)
Intra-hippocampal infusion of a GR antagonist completely
eliminates the maternal effect on HPA responses to stress
Hypothalamus
Pituitary
Adrenals
High LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Hypothalamus
Pituitary
Adrenals
Low LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Individual differences in glucocorticoid receptor
levels lead to altered pituitary-adrenal responses
to stress
Trai
t Anx
iety
Maternal Care (Parental Bonding Index)4030201010
20
30
40
50
60
70
80
(r = -0.54, p<.01)
Parental Care & Anxiety
Pruessner et al. J Neurosci 2004
Anxiety & HPA Responses To Stress
Pruessner et al. J Neuroscience 2004
High Low0
10
20
30
40
50
60
Sta
te A
nxi
ety
*
High Low-20
-10
0
10
20
30
40
Co
rtis
ol (
∆ µ
g/d
l/min
)
*
Ventral Striatum DA Stress Response
Right
Low High
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Rac
lop
rid
e (
∆ S
tres
s -
Bas
al)
Left
Low High
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
Rac
lop
rid
e B
d (
∆ S
tres
s -
Bas
al)
Pruessner et al. J Neurosci 2004
Inte
g. C
ort
iso
l (∆
Str
ess
-B
asal
)
PBI (Cr/M)
50403020-30
-10
10
30
50
(r= -0.63; p<.07)
Cortisol Response
PBI (Cr/M)
Dopamine Release
Rac
lop
rid
e (∆
Str
ess
-B
asal
)4030201010
-0.05
0.05
0.15
(r= -0.66; p<.05)
Hypothalamus
Pituitary
Adrenals
High LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Hypothalamus
Pituitary
Adrenals
Low LG offspring
Glucocorticoids
(-)
CRF
ACTH
(-)
Cross-fostering reveals evidence for direct, postnatal
effects of maternal care
High/High High/Low Low/High Low/Low.
.
.
.
.
GR
ir (
RO
D)
**
Cross-fostering reveals evidence for direct, postnatal
effects of maternal care on hippocampal GR expression
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
% LG
hip
po
ca
mp
al G
R m
RN
A
Within-litter variation in pup LG frequency predicts
hippocampal glucocorticoid receptor gene expression
r= 0.375, p= 0.008
How might maternal licking/grooming
regulate hippocampal glucocorticoid
receptor gene activity and HPA function?
And how do
these effects persist over the lifespan
of the offspring?
Signal
Receptor
Extracellular signalMaternal
Care
DNA
Transcription
Factors
Intracellular
signal
gene
Relevant gene - environment interaction
cAMP
NGFI-A
5-HT7 rec
5-HTTactile stimulation
(maternal LG)
GR gene
PKA
(T3)
(H-P-T)
Transcription
Factor
Summary of in vivo and in vitro studies
NGFI-A overexpression in cultured hippocampal neurons
0.0
2.0
4.0
6.0
8.0
10.0
No
rmali
zed
GR
0.0
2.0
4.0
6.0
8.0
10.0
No
rmali
zed
GR
17
0.0
1.0
2.0
3.0
No
rma
lize
d N
GF
I-A
0 1 2 3 4
0
1
2
3
No
rma
lize
d G
R
Normalized NGFI-A
Ctl EV NGFI-A Ctl EV NGFI-A Ctl EV NGFI-A
NT EV OE0
1
2
3
4
5
No
rmali
zed
GA
D1
GAD1 expression in NGFI-A
overexpression cells
Hellstrom et al., in press
What are the relevant
genomic targets?
Clone the 5‟ untranslated
region of the rat
hippocampal glucocorticoid
receptor gene
cAMP
NGFI-A
5-HT7 rec
5-HT
GR gene
PKA
Gene organization
Non-coding, regulatory
region (contains enhancers,
repressors, etc.).
Coding region
responsible for protein
synthesis.
(+ or -)
Clone the 5‟ untranslated region of the rat
hippocampal glucocorticoid receptor gene
Glucocorticoid receptor gene
(~110 kb)
1 2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’ 3’
Transfection studies with promoter-reporter constructs reveal exon
17 sequence has considerable transactivational capacity.
(McCormick et al. Mol Endo 2000)
Critical for hippocampal GR
1 2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’ 3’
DG CA1 CA2 CA30.00
0.05
0.10
0.15 High LG
Low LG
Rela
tive
OD
U
*
* **
1681 ccc
1741 ctctgctagt gtgacacact t1cg2cgcaact c3cgcagttgg 4cggg5cg6cgga ccacccctg7c
1801 ggctctgc8cg gctggctgtc accct9cgggg gctctggctg c10cgaccca11cg ggg12cgggct
1861 c13cgag14cggtt ccaagcct15cg gagtggg16cg gggg17cgggag ggagcctggg agaa
DNA sites that regulate glucocorticoid receptor gene
NGFI-A
GR Promoter 17 Sequence
1 2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’ 3’
NGFI-A binding to the GR(17) promoter
in neonates
**
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
An
tib
od
y/i
np
ut
(RO
D) 1.0
NGFI-A
cAMP
NGFI-A
5-HT7 rec
5-HT
GR gene
PKA
NGFI-A binding to Exon 17
promoter region - P4 On/Off
hi on hi off lo on lo off0
1
2
No
rmali
zed
GR
17
NGFI-A binding to GAD1 promoterin P4 on/off model
hi on hi off lo on lo off0. 00
0. 05
0. 10
0.15
No
rma
lize
d G
AD
1
NGFI-A binding to exon 17 promoter is dynamically
regulated by mother – pup interactions
…..
Exon 1 (Noncoding region) Exons 2- 9 Coding region
5’ 3’
(+)
Glucocorticoid receptor mRNA
Glucocorticoid receptor protein
Offspring of High LG mothers
17
NGFI-A
…..
Exon 1 (Noncoding region) Exons 2- 9 Coding region
5’ 3’
(+)
Glucocorticoid receptor mRNA
Glucocorticoid receptor protein
Offspring of High LG mothers
17
NGFI-A
*
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
Pup Adult
NGFI-A levels
So, while increased levels of NGFI-A can explain the increased
activity of the glucocorticoid receptor gene in the pup, it does not
explain why the difference is still observed in adult animals?
…..
Exon 1 (Noncoding region) Exons 2- 9 Coding region
5’ 3’
(+)
Glucocorticoid receptor mRNA
Glucocorticoid receptor protein
Offspring of High LG mothers
17
NGFI-A
Stable modification
of the DNA????
1681 ccc
1741 ctctgctagt gtgacacact t1cg2cgcaact c3cgcagttgg 4cggg5cg6cgga ccacccctg7c
1801 ggctctgc8cg gctggctgtc accct9cgggg gctctggctg c10cgaccca11cg ggg12cgggct
1861 c13cgag14cggtt ccaagcct15cg gagtggg16cg gggg17cgggag ggagcctggg agaa
DNA sites that regulate glucocorticoid receptor gene
NGFI-A
GR Promoter 17 Sequence
1 2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’ 3’
5’ 3’0.0
0.2
0.4
0.6
0.8
1.0
Low LG
High LG
c-M
eth
yla
tio
n
*
NGFI-A site
5‟…tgcgggggcgggg…3‟
NGFI-A
CH3CH
3
Low/Low High/High Low/High High/Low
0.0
0.2
0.4
0.6
0.8
c-m
eth
yla
tio
n
5' CpG region of NGFI-A/RE
5‟…tgcgggggcgggg…3‟
NGFI-A
CH3 CH
3
5‟…T G C G G G G G C G G G G …… 3‟
5‟…T G C G G G G G C G G G G …….3‟
CH3
CH3
5‟…T G C G G G G G C G G G G …….3‟
CH3
CH3
5‟…T G C G G G G G C G G G G …… 3‟
(no methylation)
(methylated only at 3‟; High LG)
(methylated only at 5‟)
(methylated at both 5‟ and 3‟; Low LG)
In vitro purified NGFI-A binding to ….
5‟…T G C G G G G G C G G G G …… 3‟
5‟…T G C G G G G G C G G G G …….3‟
CH3
CH3
5‟…T G C G G G G G C G G G G …….3‟
CH3
CH3
5‟…T G C G G G G G C G G G G …… 3‟
(no methylation)
(methylated only at 3‟; High LG)
(methylated only at 5‟)
(methylated at both 5‟ and 3‟; Low LG)
NGFI-A binding to ….
(High LG)
(Low LG)
*
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
Pup Adult
NGFI-A levels
So while levels of NGFI-A are similar in animals reared
by High or Low licking/grooming mothers, the NGFI-A
site in the adult offspring of Low LG mothers is methylated
and therefore cannot interact with NGFI-A.
But….
…..
Exon 1 (Noncoding region) Exons 2- 9 Coding region
5’ 3’
(+ or -)
Glucocorticoid receptor mRNA
Glucocorticoid receptor protein
Offspring of Low LG mothers
17
NGFI-A
CH3
1.5
1.0
0.5
0.0
Adult NGFI-A
…..
Exon 1 (Noncoding region) Exons 2- 9 Coding region
5’ 3’
(+ or -)
Glucocorticoid receptor mRNA
Glucocorticoid receptor protein
Offspring of High LG mothers
17
NGFI-A
1.5
1.0
0.5
0.0
Adult NGFI-A
0.1
0.3
0.5
0.7
0.9
LowHigh
Exo
n 1
7le
vels
In vivo binding of NGFI-A to the glucocorticoid
receptor regulatory site in adult animals as a function
of maternal care.
Maternal Care
Glucocorticoid receptor
gene expression
CRF expression
Individual differences in stress reactivity
GC feedback
sensitivity
Differential methylation of GR promoter
… differences in NGFI-A
binding.
High/Ctl High/TSA Low/Ctl Low/TSA0.0
0.2
0.4
0.6
0.8
1.0
Rela
tive
RO
D
*
Weaver et al. Nature Neuroscience 2004
C G
Ac
TFTF
Ac
Histone Acetylation
C G
Ac
TFTF
Ac
High/Veh High/TSA Low/Veh Low/TSA
0.0
0.2
0.4
0.6
0.8
1.0
An
tib
od
y/I
np
ut
*
Transcription Factor (NGFI-A) Binding
CmG
MeCP
2
mSINHDAC
HDAC
TF
TF
Ac
Ac
X
C G
Ac
TFTF
Ac
0
0.1
0.2
0.3
0.4
Vehicle 100 ng/ml TSA
GR
IR
(R
OD
-IO
D)
*
Low LG
High LG
Glucocorticoid Receptor
(Hippocampus)
Low/Veh Low/TSA High/Veh High/TSA
Basal Stress Stress + 40 Stress + 60 Stress + 900
10
20
30
40
50
Co
rtic
os
tero
ne
(
µg
/dl)
Time (min)
Low/Veh
Low/TSA
High/Veh
High/TSA
Glucocorticoid stress response
Weaver et al. Nature Neuroscience 2004
GGAGCCTGGG AGAA CCAAGCCTCG GAGTGGGCG GGGGCGGGAGCH
3CH
3
GGAGCCTGGG AGAA CCAAGCCTCG GAGTGGGCG GGGGCGGGAGCH
3
Offspring of Low LG mothers
Offspring of High LG mothers
Lower levels of GR in hippocampus -
Increased stress response
High levels of GR in hippocampus -
Modest stress response
cAMP
NGFI-A
5-HT7 rec
5-HTTactile stimulation
(maternal LG)
GR gene
PKA
(T3)
(H-P-T)
Do these „maternal‟ signals alter the methylation
of the exon 17 GR promoter?
E .
.
.
.
.
.
HighLow
C -
meth
yla
tion
5' CpG region of NGFI-A/RE
A period of methylation of the 5’ CpG site (E20 to PND 1)
followed by a demethylation (PND 1 to PND 6).
Hippocampal GR(17) methylation
5‟…tgcgggggcgggg…3‟
NGFI-A
CH3
E .
.
.
.
.
.
C -
meth
yla
tio
n
5' CpG region of NGFI-A/RE
Demethylation corresponds to period of differences
in maternal LG
- - - - -
HighLow
Days of Age
Lic
kin
g/g
roo
min
g
cAMP
NGFI-A
5-HT7 receptor
5-HTTactile stimulation
(maternal LG)
GR gene
PKA CBP
Summary of in vivo and in vitro mechanisms
NGFI-A (egr-1, zif-268, krox-24)
High LG high NGFI-A levels
5‟…tgcgggggcgggg…3‟CH
3CH
3
Acetyl
Lys9
Lys9
NGFI-A CBP
CBP (CREB binding protein)
is increased in High LG
offspring - CBP is a
histone acetyltransferase
C T A C G T A C T C G G A A T C T C G
CH3
CH3
CH3
DNA methylation serves as an interface between
the dynamic environment and the fixed genome
DNA methylation serves to imprint social factors,
such as maternal behavior, upon the offspring‟s genome.
Do comparable processes occur in humans?
• Post-mortem studies of hippocampus.
• Samples from suicide victims/controls.
• QSBB (Gustavo Turecki) - forensic
phenotyping.
• Human exon 1F promoter (Turner &
Muller, J Molec Endo, 2005)
Hippocampal samples from humans
• Human brain bank (suicide victims vs controls).
• All suicide victims (and none of the controls)
experienced verified abuse in childhood.
Human glucocorticoid receptor gene
A B C D E F G H I J 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’3’
Human glucocorticoid receptor gene
A B C D E F G H I J 2 3 4 5 6 7 8 9
Variable exon 1 region Constant region
5’3’
Control Suicide0.0
0.5
1.0
1.5GRtotal
GR
mR
NA
Control Suicide0.0
0.3
0.6
0.9
1.2 GR1F
GR
1-F
vari
an
ts
McGowan et al. PlosOne 2008, Nature Neuroscience 2009
Suicide vs abuse - GR expression
GRtotal
GR
mR
NA
/GA
PD
H (
log
co
nc
.)
Control Suicide Suicide
- Abuse + Abuse
0.0
0.4
0.8
1.2
*
GR
-1F
mR
NA
/GA
PD
H (
log
co
nc
.)
Control Suicide Suicide
- Abuse + Abuse
0.0
0.5
1.0
1.5
*
McGowan et al. PlosOne 2008, Nature Neuroscience 2009
GR
1-F
Cp
G M
eth
yla
tio
n (
%)
Control Suicide Suicide
- Abuse + Abuse
0
20
40
60*
0
20
40
60
30 31 32
CpG sites
Suicide vs abuse - CpG methylation
McGowan et al. PlosOne 2008, Nature Neuroscience 2009
Control NGFI-A expression
Co-transfection studies (NGFI-A vector w/
human GR exon 1F-luciferase construct)
0
2000
4000
6000
8000
10000
12000
14000
Lu
cif
era
se
ac
tivit
y
mPlasmid nmPlasmid Antisense
McGowan et al. PlosOne 2008, Nature Neuroscience 2009
GR1F promoter methylation levels in WBC of PTSD patients and controls (new subjects)
CpG sites
Meth
ylation leve
ls (0
/ 0)
0
10
20
30
40
50
PTSD (n=8)
Control (n=4)
s2 s3
s1
4
s1
3
s2
4
s3
4s4
s3
0
s3
1
s5
s3
5
s1
5
s2
5s6
s2
6
s1
6s7
s3
7
s1
7
s2
7
s1
8
s3
8
s8
s2
9
s2
0
s1
2
s1
0s1
1
s1
9
s2
1s2
2
s2
3
s2
8
s3
6
s3
3
s3
2