Neurodegenerative diseases from systems medicine perspectiveemreguney.net/doc/ms_neuro.pdf · Network medicine is an emerging field that aims to use interactome-based analyzes to

Neurodegenerative diseases from systemsmedicine perspective

Emre Güney, PhD

Hospital del Mar Research Institute (IMIM)& Pompeu Fabra University(UPF)

MSc on Neuroscience - UAM

February 26th, 2018

Systems medicine

(Hampel et al., 2018, Pharm Rev)

2

Systems-level representation of the cellular network

True disease Module

Observablediseasemodule

Missing link

Unknown disease protein

False disease association

Observable InteractomeComplete Interactome

Node deletion Loss offunction}Link deletion

Modification of theinteraction strength:Loss or gain of functon

disease causingperturbations

cb

1

10

100

0 100 200 300 400 500 600 700 800 900 1000 1100Num

ber

of d

isea

ses

Number of associated genes

0

10

20

30

40

50

60

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Num

ber

of d

isea

ses

Fraction of nodes in observable module

e f

1

10

20

10 Nc 100

Total number of disease genes

Obs

erve

d m

odul

e si

ze

peroxisomaldisorders

rheumatoidarthritis

Percolation prediction

multiplesclerosis

Network completeness

p 100

1

c

Rela

tive

mod

ule

size

Current interactomeLarge module: observableSmall module: fragmented

Low coverageNo

observable modules

Completenetwork

Fragmented network

p’c p”c

Disease modules as local perturbations

Large M

odule

Small

Module

Complete interactomeAll modulesobservable

GWAS

Multiple sclerosisgenes

OMIM

Signalling

Complexes

Kinase - Substrate

Metabolic

Literature

Regulatory

Yeast two-hybrid

GWAS & OMIM

Other disease genes

Molecular interactions

Gene with multiple disease associations

OMIMImmunologic deficiencysyndromesHematologic diseasesBlood protein disorders

GWASConnective tissue diseasesAutoimmune diseasesJoint diseasesMusculoskeletal diseasesRheumatoid arthritis

Signaling, Complexes, Literature, RegulatoryInteraction with multiple lines of evidences

AKT1

CD40

HLA-B

HLA-C

STAT3

TAP2

NFKBIZ

IL2RA

TNFRSF1A

EHMT2PTK2

IL7R

MAPK1

Observable module for Multiple sclerosis

Building the Interactome

Gradual fragmentation of the interactome Estimating the critical module size

d

a

Menche et al., Figure 1

(Menche et al., 2015, Science)

Recent advances in *omics have given rise to rich data sets:Genome / Transcriptome / Proteome / Epigenome / Exosome /

Metabolome / Microbiome / ...

3

Network representationNetworks (in biology) are formally defined as Graphs (in math &computer science)

G(V, E)

V: Vertices (nodes)

E: Edges (interactions) connecting vertices

In a biological network, vertices can be of di�erent types:proteins

genes

metabolites

...And edges correspond to various types of interactions between them 4

Interactome: The cellular map

(Barabási and Oltvai, 2004, Nat Rev Genet)

Proteins “talk to each other” by physically interacting with eachother

These interactions are essential for performing biologicalprocesses

The network of interactions between proteins: Interactome5

How to generate such a map for the cell?

6

Interactome generation: Protein interaction data

Protein-protein interaction (PPI) data is spread across variousrepositories

BioGrid

MIPSDIP

KEGG

Reactome

7

Interactome generation: Integrating protein interaction data

(Garcia-Garcia et al., 2010, BMC Bioinformatics)8

Interactome visualization and analysis

Cytoscape: A tool to visualize and analyze biological networks(Cline et al., 2007, Nat Protoc)

9

Network analysis for understanding a system

Percentage of your

Facebook friends

from the same country(Ugander et al., 2011, arXiv:1111.4503)

∼85%

image from relenet.com

10

Network analysis for understanding a system

Percentage of your Facebook friends from the same country(Ugander et al., 2011, arXiv:1111.4503)

∼85%

image from relenet.com10

How can one use interactome to extract biologically meaningfulinformation?

Similarly, proteins that perform similar functions interact with eachother (guilt-by-association principle)

11

Interactome analysis: Guilt-by-association

True disease Module

Observablediseasemodule

Missing link

Unknown disease protein

False disease association

Observable InteractomeComplete Interactome

Node deletion Loss offunction}Link deletion

Modification of theinteraction strength:Loss or gain of functon

disease causingperturbations

cb

1

10

100

0 100 200 300 400 500 600 700 800 900 1000 1100Num

ber

of d

isea

ses

Number of associated genes

0

10

20

30

40

50

60

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Num

ber

of d

isea

ses

Fraction of nodes in observable module

e f

1

10

20

10 Nc 100

Total number of disease genes

Obs

erve

d m

odul

e si

ze

peroxisomaldisorders

rheumatoidarthritis

Percolation prediction

multiplesclerosis

Network completeness

p 100

1

c

Rela

tive

mod

ule

size

Current interactomeLarge module: observableSmall module: fragmented

Low coverageNo

observable modules

Completenetwork

Fragmented network

p’c p”c

Disease modules as local perturbations

Large M

odule

Small

Module

Complete interactomeAll modulesobservable

GWAS

Multiple sclerosisgenes

OMIM

Signalling

Complexes

Kinase - Substrate

Metabolic

Literature

Regulatory

Yeast two-hybrid

GWAS & OMIM

Other disease genes

Molecular interactions

Gene with multiple disease associations

OMIMImmunologic deficiencysyndromesHematologic diseasesBlood protein disorders

GWASConnective tissue diseasesAutoimmune diseasesJoint diseasesMusculoskeletal diseasesRheumatoid arthritis

Signaling, Complexes, Literature, RegulatoryInteraction with multiple lines of evidences

AKT1

CD40

HLA-B

HLA-C

STAT3

TAP2

NFKBIZ

IL2RA

TNFRSF1A

EHMT2PTK2

IL7R

MAPK1

Observable module for Multiple sclerosis

Building the Interactome

Gradual fragmentation of the interactome Estimating the critical module size

d

a

Menche et al., Figure 1

(Menche et al., 2015, Science)

Proteins that perform similar functions or involved in similardiseases interact with each other (guilt-by-association)

12

Neurodegenerative diseases: A systems-level perspective

13

Alzheimer’s Disease (AD)

Proteins of genes known to be involved in AD (orange hexagons,sources: OMIM, DisGeNET, GUILDify2) 14

Alzheimer’s Disease (AD)

Most relevant functions among AD-related genes15

Parkinson Disease (PD)

Proteins of genes known to be involved in PD16

Parkinson Disease (PD)

Most relevant functions among PD-related genes17

Huntington Disease (HD)

Proteins of genes known to be involved in HD 18

Huntington Disease (HD)

Most relevant functions among HD-related genes19

Amyotrophic Lateral Sclerosis (ALS)

Proteins of genes known to be involved in ALS 20

Amyotrophic Lateral Sclerosis (ALS)

Most relevant functions among ALS-related genes21

Shared functional componentsacross neurodegenerative diseases

22

Alzheimer’s Disease and Parkinson Disease

23

Alzheimer’s Disease and Parkinson Disease

Perez-Rubio et al., 2017, Sci RepAguirre-Plans et al., submi�ed

24

Alzheimer’s Disease and Huntington Disease

25

Alzheimer’s Disease and Huntington Disease

26

Alzheimer’s Disease and Amyotrophic Lateral Sclerosis

27

Alzheimer’s Disease and Amyotrophic Lateral Sclerosis

28

From interactome to diseasome, the network of diseasesneurodegenerative cardiovascular

neoplasms

glucose metabolism

respiratory

digestive system

Hyperglycemia

BreastneoplasmsAsthma

Colorectalneoplasms

Vitiligo

Type 2 diabetes

Prostaticneoplasms

Gastrointestinaldiseases

Diabeticnephropathy

Drug-inducedliver injury

Liver cirrhosis

Non-alcoholicfatty liverdisease

Pulmonaryemphysema

Sepsis

Squamous cellcarcinoma

Non-small celllung carcinoma

Glioma

Pancreaticneoplasms

Epithelialovarian cancer

Colon neoplasmsLeukemia

Liver neoplasms

Diabeticretinopathy

Chronicobstructivepulmonary

disease

Acute kidneyinjury

Chronic kidneydisease

Vasculardiseases

Pulmonaryfibrosis

Middle cerebralartery infarction

IschemiaAtherosclerosis

Systemic lupuserythematosus

nephritisHuntington

disease

Amyotrophiclateral sclerosis

Alzheimer'sdisease

Parkinsondisease

Diabeticcardiomyopathies

Goh et al., 2007, PNASCuadrado et al., to appear in Pharm Rev

29

Neurodegenerative diseases: Challenges and opportunities

30

We still do not completely understand the disease pathology

a cb

d

Proximal Distant

Proximity (zc)

Palliat

ive

Non-pa

lliativ

e

Off-lab

elDA

ILY

ME

D L

abel

s

FDA Reports

Palliat

ive

Non-pa

lliativ

e

Off-lab

el

Relative Efficacy (%)

Proxim

al

Distan

t

Relative Efficacy (%)

All drugs are distant

All drugs are proximal

Rat

io o

f pro

xim

al d

rugs

(%)Proximal drugs

Distant drugs

N/A

N/A

N/A

N/A

N/A

HIGH RELATIVE EFFICACY

LOW RELATIVE EFFICACY

Guney et al., 2016, Nat Comm31

Genetic hetereogeneity among patients

Schork, 2015, Nature32

Not all disease-associated genes are altered in a given patient

a

1

10

100

1000

10000

0 0.2 0.4 0.6 0.8 1

1

10

100

1000

0 0.2 0.4 0.6 0.8 1

0

0.2

0.4

0.6

0.8

1

fract

ion

of p

airs

with

sig

nific

ant o

verla

p

0.001

0.01

0.1

1

1

10

100

1000

10000

0.001

0.01

0.1

1

Asthma

Huntin

gton

Parkins

on

Pairw

ise

Jacc

ard

sim

ilarit

yp < 10�36

Asthma

Huntin

gton

Parkins

on

p < 10�36

Frac

tion

of p

airs

with

sign

ifica

nt o

verla

p

Asthma

Huntin

gton

Parkins

on

p < 10�77 p < 10�18 p < 10�23 p < 10�16

Num

ber o

f per

turb

ed g

enes

Fraction of subjects in which gene is perturbed

Num

ber o

f gro

up-w

ise

DE

gene

s

Fraction of subjects in which gene is perturbed Asthma

Huntin

gton

Parkins

on

p < 10�4 p < 10�5 p < 10�3b

dc

fe

Frac

tion

of g

roup

-wis

e D

E ge

nes

Num

ber o

f com

mon

DE

gene

s

control

control

Bonferroni

raw p-value

7.4%3.6% 3.7%

20%9.5%

29%

FKBP5

PTGESRPS6KA2

HIF3A CD163FAM107A

PER3ATP6V1C2TFCP2L1

PDK4

FKBP5

PTGESRPS6KA2

THRANR1D1

CD163FAM107A

PER3ATP6V1C2

TFCP2L1PDK4

controlasthma

controlasthma

The overlap between perturbed genes of two individuals with thesame disease

is low (< 30%), suggesting high heterogeneity at thetranscription level

is higher than the overlap between the PeePs of healthy subjectsMenche et al., 2017, Npj Sys Bio & App

33

Several resources for systems medicine and interactome analysis

Resource URLDisease-gene information

DisGeNET www.disgenet.orgInteractome generation

BIANA (web server) sbi.imim.es/BIANA.phpInteractome visualization and analysis

Cytoscape www.cytoscape.orgGUILDify web server sbi.imim.es/GUILDify2.php

Transcriptomic analysisPEPPER R package github.com/emreg00/pepper

34

Concluding remarks

Interactome, the network of interactions between proteins,provides a framework for understanding and characterizingbiological processes

Network medicine is an emerging field that aims to useinteractome-based analyzes to identify genes associated todiseases and develop novel therapeutics

The biological mechanisms underlying neurodegenerativediseases are highly complex, o�en involving perturbations onmultiple genes

These genes are involved in diverse biological functions, some ofwhich are common across di�erent neurodegenerative diseases

Genetic hetereogeneity among patients both poses a challengeand opportunity in developing personalized treatments 35

Acknowledgements

UPFBaldo OlivaJavier GarcíaJoaquim Aguirre-PlansCarlota Rubio-Perez

MUHarald H.H.W. Schmidt

CeMMJörg Menche

NEUAlbert-László Barabási

emreguney.net

36