Long-lasting alterations to DNA methylation and ncRNAs could underlie the effects of fetal alcohol exposure in mice

Long-lasting alterations to DNA methylation and ncRNAs could underlie the effects of fetal alcohol

exposure in mice.

PMID: 23580197

Ben LauferPI: Shiva M. Singh

Contributors: Katherine Mantha, Morgan L. Kleiber, Eric J. Diehl, Sean M.F. AddisonInstitution: Western University (formerly University of Western Ontario)

1

Above and Beyond the Genome

Welcome to The Post-Genomics Era

• We’ve sequenced the human genome

• It can now be done for only a few thousand dollars!– We can sequence thousands of human genomes!– Individual people can now go see their sequence!

• The benefits we’ve seen are…

Is this it?

• Shouldn’t cracking the code of life give us more breakthroughs than we can dream of?

• Can 4 base pairs really code for complex life?

• What’s missing?

What treasure is left?

www.biocomicals.com, Alper Uzun, PhD.

5

“A mitotically (or meiotically) inheritable change in gene expression, independent of an alteration in DNA sequence”

– Berger et al. (2009) Genes Dev.

The Solution to the Post Genomics Era: Epigenetics

Epigenetic Mechanisms (Macmillan Publishers Ltd: Nature 441: 143-145. 11 May 2006)

A Few Epigenetic Mechanisms of Interest

• Histone Modifications

• MicroRNA

• DNA Cytosine Methylation

6Images from Wikimedia Commons

Redefining the Central DogmaSaletore et al. Genome Biology 2012, 13:175

Epigenetic Mechanisms

• Histone Modifications

• MicroRNA

• DNA Cytosine Methylation

8

The Histone Code

http://eukaryoticgeneexpression.weebly.com/uploads/5/6/3/0/5630004/7502951_orig.jpg

Epigenetic Mechanisms

• Histone Modifications

• MicroRNA

• DNA Cytosine Methylation

10

microRNAs

• The second level of the epigenetic landscape

– Acts at translation, as opposed to Histone modifications and DNA methylation, which act transcription

• Act as fine-tuners, rather than on and off switches like histone modifications and DNA methylation.

– Typically result in low fold changes in gene expression• However, these changes are physiologically relevant.

12

microRNA (miRNA)

http://www.wormbook.org/chapters/www_microRNA/celmicRfig5.jpg

Epigenetic Mechanisms

• Histone Modifications

• MicroRNA

• DNA Cytosine Methylation

13

DNA Methylation (and other modifications)

• Histones aren’t the only ones to enjoy modifications

• There aren’t just four base pairs anymore– Cytosine has made a few friends

http://www.atdbio.com/img/articles/epigenetic-base-modifications.png

15

DNA Cytosine Methylation

. Metivier, R. et al. Cyclical DNA methylation of a transcriptionally active promoter. Nature 452, 45–50 (2008).

Key Players of the Epigenetic Landscape

• Histone Modifications

• MicroRNA

• DNA Cytosine Methylation

16

The Epigenetic Landscape

http://cnx.org/content/m26565/latest/graphics35.jpg

It really is a landscape…

19

Waddington’s Epigenetic LandscapeWaddington, Conrad H. 1953. The Epigenetics of birds. Cambridge University Press

• A metaphor for biological development

• Cell fates are established in development by epigenetic marks much like a marble rolls down to the lowest point

• Increasing irreversibility of cell type differentiation as ridges rise between the valleys.

20

Environmental Conditions

21

Environmental Epigenetics

22

Environmentally Responsive Genome

http://learn.genetics.utah.edu/content/epigenetics/nutrition/images/pathway.jpg

23

Waddington’s Epigenetic LandscapeWaddington, Conrad H. 1953. The Epigenetics of birds. Cambridge University Press

• What happens if an obstacle gets in the way?

24

Fetal Alcohol Exposure • Leading cause of preventable birth defects and mental deficits in North

America

• FASD– Fetal Alcohol Spectrum Disorders

– 2-5% of pregnancies!

– Umbrella term for a number of physical abnormalities, behavioural and intellectual problems

– Strongest manifestation is Fetal Alcohol Syndrome (FAS)

Chudley et al. CMAJ 2005

Spectrum Disorders

Fetal Alcohol Spectrum Disorders

Fetal Alcohol Spectrum Disorders

Fetal Alcohol Syndrome (FAS)

Partial Fetal Alcohol Syndrome (PFAS)

Alcohol-Related Neurodevelopmental Disorder

Alcohol-Related Birth Defects

Fetal Alcohol Effects

Epigenetics

Where does an individual land in the spectrum?

• Depends on:

– Genetic Background

– Timing of Exposure

– Dosage of Exposure

– Other Epigenetic and Environmental Factors, either:• Inherited• Experienced

28

Spectrum Meets Landscape

Fetal Alcohol Syndrome (FAS)

Partial Fetal Alcohol Syndrome

Alcohol-Related Neurodevelopmental Disorder (ARND)

Alcoho

l-Rela

ted

Birth

Defe

cts (A

RBD)

Feta

l Alco

hol E

ffect

s (FA

E)

Mouse Models• Reasons for use:

– Useful for studies that would be impractical in humans

– Reach sexual maturity early (6–8 wk)

– Birth multiple offspring

– Abbreviated gestational period (18–21 days)

• Many generations can be analyzed within a relatively short timeframe (1–2 yr)

– Allows for studying long-term changes

– Allow for analysis in vivo, which is essential for epigenetic studies.

• Cell cultures do not exhibit natural epigenetic properties– See: Embryo culture and epigenetics. Velker BA, Denomme MM, Mann MR. Methods in Molecular Biology.

© Disney

Fetal Alcohol Exposure in an Animal Model

• We have shown that Fetal Alcohol Exposure (FAE) affects behaviour, learning and related genes.

• We have also recently shown that these changes are maintained for a lifetime– Even after exposure has ceased for weeks

• Are epigenetic mechanisms responsible?30

Kleiber et al. Behav. Brain. Res. 2012

Kleiber et al. Brain. Res. 2012

Epigenetic Mechanisms of Interest

• DNA Cytosine Methylation– Typically turns

expression on or off

31

Metivier, et al 2008. Nature

http://www.wormbook.org/chapters/www_microRNA/celmicRfig5.jpg

• Noncoding RNA• i.e: miRNA• Fine tuners of gene

expression• Low fold changes

Fetal Alcohol Exposure and Methylation

• FAE alters the methylation and expression of genomically imprinted (uni-parental) genes in cell cultures derived from:

– Whole embryo

– Placenta

Liu et al. 2009 Epigenetics.

32

Shukla et al. 2011 Alcohol. Clin. Exp. Res

Morison et al. 2005, Trends. Genet.

• 30% of parentally imprinted transcripts are ncRNA.

33

Imprinted ncRNA

• Key role in neurodevelopment and memory.

• Important for early life processes– and functionally important for adult

brain functions.

• Many are microRNAs

Wang et al. 2009 PLoS One

Davies et al. 2008 Adv. Exp. Med. Biol

http://4.bp.blogspot.com/_Ik_ovkt6ICg/SfjaLNt1ehI/AAAAAAAAAF4/4ZzazV4xx3Y/s320/imprinted+brain.jpg

Fetal Alcohol Exposure and miRNAs• miRNAs have been shown to be deregulated

by FAE in fetal mouse brain cell culture

• Co-incubation with folic acid prevents altered miRNA and target gene expression in mouse embryos

• Folic Acid is involved in establishing DNA methylation.

• Association between methylation and expression, but what is the mechanism behind this relationship?

Sathyan et al. 2007 J. Neurosci.

Wang et al. 2009 Hum. Reprod.

34

35

Functional Mechanisms• An alteration of methylation in a transcription factor binding

site has the potential to affect gene expression.

• CTCF binding sites in the H19/Igf2 imprinting control region show differential methylation in FASD placental tissue.

– CTCF is a highly conserved ubiquitous zinc-finger protein with multiple functions in chromatin organization and gene regulation

– It binds in a methylation sensitive manner to target sequences

– Is this the functional mechanism for the association between gene expression and DNA methylation in FASD?

Williams et al 2008. J. Exp. Med.

Filippova 2008 Curr. Top. Dev. Biol.

Haycock et al. 2009 Biol. Reprod.

36

Hypothesis

Alterations in DNA methylation and ncRNA expression are associated with life-long alterations in gene expression in the mouse brain after fetal alcohol exposure.

Continuous Preference Drinking (CPD)• Free choice

• Quantity monitored daily

• No Stress

• 70% preference for 10% EtOH

• C57BL/6J mice

• Metabolize alcohol much quicker than humans

• Blood Alcohol Concentration (BAC)

• Represents moderate drinking

• = pregnant human mother who has a drink or two every now and then.

37Young & Olney 2006 Neurobiol. Dis.

Experimental Design• Everything done downstream of your workflow is

dependent on what has happened upstream.

• Errors will amplify.

• For perspective see:

– Fundamentals of experimental design for cDNA microarrays. Churchill GA. Nature Genetics.

– Probe set algorithms: is there a rational best bet? Seo J, Hoffman EP. BMC Bioinformatics.

– Tackling the widespread and critical impact of batch effects in high-throughput data. Leek JT. Nature Reviews Genetics.

Whole Brain Methylation Array Analysis

39DNA Methylation

PND 70

40

• Over 6,600 genes with differences in 1 or more promoter regions

• More than half of imprinted genes in genome

• p < 0.01

• Subjected to Ingenuity Pathway Analysis

TreatedControl

Looking through the noise• What is a p-value?

• Is a p=0.01 stringent enough for data from 2.1 Million probes?

• There’s going to be some false positives!

– But do we want to get rid of all the useful data caught up in those p-values?

• Particularly relevant for spectrum disorders as they are riddled with heterogeneity since we expect large amounts of variation within the experimental group

Systems Biology

• The antithesis to a reductionist approach

• Reductionism is a philosophy that the understanding of a complex system can be achieved in full by understanding its simpler component parts.

• Systems biology, on the other hand relies on examining the entirety of cellular processes and interactions in concert .

Independent Component Analysis• Ingenuity Core Analysis

• A 1.2 fold increase in many genes of a pathway can have a potentially greater physiological impact than a 20-fold increase in a single gene.

• “Project microarray data into statistically independent components that correspond to putative biological processes, and to cluster genes according to over- or under-expression in each component.”– Further Perspective: Application of independent component analysis to microarrays. Lee SI,

Batzoglou S. Genome Biology.

45

Enriched Biological FunctionsMolecular and Cellular Functions    Name p-value # GenesCell Death 4.06E-04 - 4.97E-02 224Cellular Development 6.24E-04 - 4.52E-02 166Cellular Function and Maintenance 9.57E-04 - 4.97E-02 86Cellular Movement 4.12E-03 - 4.52E-02 41Cell Signaling 8.43E-03 - 4.97E-02 26Physiological System Development and Function    Name p-value # GenesNervous System Development and Function 3.86E-05 - 4.97E-02 273Tissue Morphology 1.64E-04 - 4.23E-02 97Behavior 1.62E-03 - 1.58E-02 24Embryonic Development 1.23E-02 - 4.23E-02 29Organismal Development 1.23E-03 - 4.23E-02 25

• Many have been previously implicated in FASD and all are highly compatible

Using MeDIP and ChIP Data

• Super Simple Stuff

• Excel table with column 1 containing gene name from upstream analysis

– Subsequent columns for metrics of interest (optional)

47

Top Affected IPA Network“Behaviour, Neurological Disease, and Psychological Disorders” (IPA Score 65)

Network Biology• The distribution of nodes (i.e: genes) in cellular

networks is highly non-uniform– Most of the nodes having only a few links to other

nodes.

• However, there are a few nodes with a very large number of links called hubs– The importance of the relationship between a system

and single gene is highlighted in these cases.

• While a network can tolerate many disruptions to its lesser-connected nodes, a similar disruption to a single hub can be catastrophic to the networks it connects.

• Further Insight:– Network biology: understanding the cell's functional organization. Barabási AL,

Oltvai ZN. Nature Reviews Genetics.

49

Top Affected IPA Network“Behaviour, Neurological Disease, and Psychological Disorders” (IPA Score 65)

40% of hub-genes promoters investigated had CTCF binding sites

Gene P

CTCF

Gene Name Protein Name

H19 N/A (ncRNA)

Gtl2 (Meg3) N/A (ncRNA)

Npy Pro-neuropeptide YAkt1 RAC-alpha serine/threonine-protein kinaseGhr Growth hormone receptor

Ntrk1 High affinity nerve growth factor receptorApoe Apolipoprotein E

Grin2c Glutamate [NMDA] receptor subunit epsilon-3

Gene Name Protein Name

AppAmyloid beta A4 protein

MbpMyelin basic protein

Atp1a2 Sodium/potassium-transporting ATPase subunit alpha-2

Grin1 G protein-regulated inducer of neurite outgrowth 1

Gene P

CTCF

CTCFBSDB: a CTCF binding site database for characterization of vertebrate genomic insulators

Bao L et al. Nucleic Acids Research 2008

51

Pten Canonical Signaling Pathway• Significantly affected canonical pathway (p=1.9E-

06)

• 54/95 molecules showed significant differential methylation in their promoters

• Controls the tempo of the process of newborn neuron integration during adult neurogenesis– including correct neuron positioning, dendritic

development, and synapse formation.Porteous et al. 2009 Neuron

52

Methylation Array Results Summary• Over 6000 genes with significant differences

in their promoters

• More than ½ of the molecules involved Pten Signaling affected

• Not a random sample– Enriched for relevant functions

• Many CTCF binding sites in important neurodevelopmental genes showed differences in methylationDNA Methylation

PND 70

Whole Brain ncRNA Expression Array Analysis

53

Gene Expression

PND 70

microRNA

54

miRNA Array Heatmapsp < 0.05FC 1.2

Treated

Treated

Treated

Treated

Control Control

Control Control

Laufer et al. Disease Models & Mechanisms. 2013

55

ncRNA Venn Diagram

p < 0.05FC 1.2

1

Laufer et al. Disease Models & Mechanisms. 2013

56

Imprinted Noncoding RNA Clusters

• Localized to the brain

• Only 3 clusters in mouse genome– Sfmbt2 (Chr 2), Snrpn-Ube3a (Chr 7), Dlk1-Dio3

(Chr 12)– 20% of altered miRNAs in all exposure paradigms

• Associated with FASD related endophenotypes

Cluster of Interest: Snrpn-Ube3a (Chr 7)

57

III5htr2cPre-mRNA

I II III IV Va Vb VI

Receptor with a stronger serotonin response

Inclusion of exon Vb without mRNA editing during alternative splicing

snoRNA binds to mRNA

H/MBII-52(SNORD115)

• Showed significant up-regulation in:• all 4 treatment paradigms

• CPD and injections• both array types

• miRNA and gene

Laufer et al. Disease Models & Mechanisms. 2013

58

PND 70

microRNA

miRNA and Gene Expression Results Summary

• Global Expression changes

• Individual miRNAs unique to treatment paradigm

• 20% of affected miRNAs belong to imprinted clusters

• H/MBII-52 only ncRNA (and gene) affected in all paradigms and arrays

Gene Expression

Whole Brain Bioinformatic Analysis

59

Gene ExpressionmicroRNADNA Methylation

PND 70

Creating a miRNA Target Filter

What types of Data can be used?

• Any Gene Expression Data!

• Most arrays and RNA-Seq technologies assay miRNAs

• Just create two separate files:

1. microRNA Expression

2. Gene Expression

Importing and formatting miRNA Expression

microRNA Analysis

Adding Gene Expression

Expression Pairing

High Quality Regulatory Relationships

Filtering

End Result!

Wasn’t that Easy?

• I really enjoyed the user interface of this program.

• It makes accessing a large annotated database of genetic information a breeze.

• Letting you get back to the biology!

+ The Power of Ad-Hoc

IPA miRNA Target Filter

71

miRNA ID miRNA Fold Change Gene ID Gene Fold

ChangeConfidence of

Interaction

mir-369-5p -1.336 Pten 1.377 High

mir-25 -1.224 Pten 1.377 High

mir-495 -1.232 Pten 1.377 High

mir-152 1.208 Otx2 -1.27 High

mir-1224 1.528 Nmnat1 -1.237 Moderate

mir-431 1.366 Nmnat1 -1.237 Moderate

mir-743a 1.341 Nmnat1 -1.237 Moderate

mir-17* 1.451 Slitrk2 -1.202 High

mir-200a* 1.178 Slitrk2 -1.202 Moderate

• 34 genes identified with reverse pairwise relationships to predicted miRNAs

• 4 are highly compatible with FASD:

72

miRNA Target Filter Gene of Interest• A novel role for Pten in FASD:

– Pten is a lipid phosphatase that suppresses Akt activation.

– Akt:

• Regulates neuronal development, including morphogenesis, dendritic development, synapse formation, and synaptic plasticity.

• Showed a gain of methylation at a predicated CTCF binding site in its promoter.

– More than half of the Pten signaling genes were significantly enriched for on the methylation arrays

Porteous et al. 2009 Neuron

73

Other Target Genes of Interest• Otx2:

– Expressed in the brain and involved in mood disorders– Identified in our previous study on long-term brain gene

expression changes in FASD.

• Nmnat1:– Protects against axonal degeneration following mechanical or

toxic insults by delaying axonal degeneration.

• Slitrk2:– Significant expression is detected only in the adult brain. – Uniquely expressed in immature neurons– Inhibitory effect on neurite outgrowth.

Kleiber et al. Brain. Res. 2012

Summary of Results

74

Mir-369

PtenImpaired behaviour, learning and memory

Chr 12 ICR

Mir-152 Otx2 Mood Disorders

Mir-25

Mir-495

Mir-1224Mir- 743a

Nmnat1Reduced ability to protect axonsMir-431

Chr 12 ICR

Mir-25 Promoter

Mir-431 Promoter

Nmnat1 PromoterLaufer et al. Disease Models & Mechanisms. 2013

75

Proposed Molecular Cascade

Fetal EthanolExposure

Methylation ncRNA Gene Expression Endophenotypes

Take Home Message

• Don’t drink when you’re pregnant.

• There’s no safe time or safe amount.

• This knowledge should be spread by informed experts (like you!)

• These findings should be considered a public health guideline for future generations.

For more info on FASD see: http://www.nofas.org/

Resources• researchgate.net and academia.edu

• Concise Summaries of Headline Epigenetic Literature (some written by yours truly)– http://epigenie.com/

79

et al.

Questions?

About the presenter:VisitBenLaufer.com