The Use of Mouse Models to Study Epigenetics

Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press

Marnie Blewitt1 and Emma Whitelaw2

1Walter and Eliza Hall Institute, Melbourne, 3052 Victoria, Australia; 2Queensland Institute of Medical Research, Brisbane, 4006 Queensland, Australia Correspondence: [email protected]

SUMMARY

Much of what we know about the role of epigenetics in the determination of phenotype has come from studies of inbred mice. Some unusual expression patterns arising from endogenous and transgenic murine alleles, such as the Agouti coat color alleles, have allowed the study of variegation, variable expressivity, transgenerational epigenetic inheritance, parent-of-origin effects, and position effects. These phenomena have taught us much about gene silencing and the probabilistic nature of epigenetic processes. Based on some of these alleles, large-scale mutagenesis screens have broadened our knowledge of epigenetic control by identifying and characterizing novel genes involved in these processes.

Outline

1 Using mouse models to identify modiﬁers of 3 Summary and future directions epigenetic reprogramming References 2 Epigenetic phenomena in inbred mouse colonies

Editors: C. David Allis, Marie-Laure Caparros, Thomas Jenuwein, and Danny Reinberg Additional Perspectives on Epigenetics available at www.cshperspectives.org Copyright # 2013 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a017939 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939

1 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

OVERVIEW

The role of epigenetics in determining phenotype has been regulators have been performed in lower organisms, using progressed through studies of inbred mice. Under laboratory variegating phenotypes such as eye color in flies and pigmen- conditions the genome and the environment of mice are tight- tation in maize. Mammalian screens have been important ly controlled such that variance in phenotype or patterns of because certain epigenetic processes are specific to higher gene expression without a change in the underlying DNA organisms: for example, the inactivation of the second X chro- sequence is, by definition, epigenetic. Interestingly, trans- mosome in females (see Brockdorff and Turner 2014) and genes in mice appear to be particularly sensitive to epigenetic genomic imprinting (see Barlow and Bartolomei 2014). Fur- silencing, and as such provide a valuable model to study the thermore, byperformingscreens in mice, oneimmediately has underlying molecular mechanisms of epigenetic control. In the mutant mouse strains to study the effects of disruption to fact, we have come to realize that there are some endogenous epigenetic processes on phenotypes relevant to humans. Two alleles, resulting from transposon insertions, which are simi- mousemutagenesisscreenshavebeenspecificallydesignedto larly susceptible to epigenetic silencing. These alleles, termed identify genes involved in epigenetic control: the Momme and metastable epialleles, display unusual expression and inheri- the X inactivation-choice screens. The Momme mutagenesis tance patterns: for example, variegated expression in a single screen has used a GFP transgenic line that displays variegated cell type, variable expressivity between individuals, and trans- expression equivalent to position-effect variegation (PEV) in generational epigenetic inheritance. The study of these phe- Drosophila. This screen has thus far revealed .30 modifiers nomena has revealed fundamental features of epigenetic of epigenetic regulation, some known but others entirely nov- control. In some cases, these recapitulate those found in other el. Importantly, the novel players appear to be involved in complex organisms, such as position-effect variegation (PEV) mammalian specific processes, and these newly identified in Drosophila (see Elgin and Reuter 2013) and paramutation molecules expand our understanding of epigenetic control in in plants (Pikaard and Mittelsten 2014), but in other cases, the the mammalian system. phenomena are unique to mammals. Interestingly, mutation in one of the novel genes identified In addition to demonstrating many of the general features in the Momme screen has now been reported to be the under- of epigenetic control, metastable epialleles and other reporter lying cause of a rare human disease. Studies in the mouse alleles have enabled random mutagenesis screens to be per- modelsgeneratedbytheMommescreenhavebeeninstrumen- formed to find genes that are important in setting and resetting tal in helping us understand the molecular mechanisms of this epigenetic marks at these loci. Similar screens for epigenetic disease. We anticipate the same will be true in other cases.

2 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

1 USING MOUSE MODELS TO IDENTIFY as Dnmt1 (Gaudet et al. 2004), HP1-b (Festenstein et al. MODIFIERS OF EPIGENETIC REPROGRAMMING 1999), and the polycomb group protein Mel18 (Blewitt et al. 2006). Together, these findings showed that metastable Random mutagenesis screens performed in mice, which epialleles are particularly sensitive to alterations in genetic have isolated mutations in epigenetic regulators, are described in this section. First, we detail screens that were makeup, ideal for a mutagenesis screen. specifically designed for the purpose of identifying epige- A mouse line carrying a variegating green fluorescent netic regulators: the Momme and the X inactivation-choice protein (GFP) transgene directed to express in red blood screen (Sections 1.1 and 1.3). Next, we briefly explain other cells was chosen (Fig. 1) (Preis et al. 2003). The advantages of using this transgenic line are many. First, the transgene is screens that were primarily aimed at finding genes involved in embryonic development, hematopoiesis or immune reproducibly expressed in 55% of red blood cells in ho- function (Section 1.4), but have produced novel mutations mozygous animals; the reproducibility of expression be- in epigenetic regulators all the same. tween isogenic littermates makes for a clean phenotype for screening with few false positives. Second, the transgene was produced and has been maintained on an inbred (FVB/ 1.1 A Screen for Modifiers of Murine Metastable N) genetic background, which simplifies later mapping Epialleles (Mommes) of the ethylnitrosourea (ENU)-induced mutations. Third, The mutagenesis screens performed in yeast, plants, and the expression of GFP in red blood cells means that trans- flies used variegating or epigenetically controlled pheno- gene expression was able to be efficiently and sensitively types as readouts of epigenetic state: for example, mating determined at a single cell level by flow cytometry. Fourth, type switching in yeast (see Allshire and Ekwall 2014), PEV by directing expression to red blood cells, analysis could in flies (Elgin and Reuter 2013), and paramutation, RNAi, be performed relatively simply (and without killing the and RNA-directed DNA methylation in plants (Pikaard animal) using a drop of blood taken from the tail of a mouse and Mittelsten Scheid 2014). These screens have iden- at weaning. Finally, alterations in the expression of the tified epigenetic modifiers critical for the phenotype being transgene itself do not inherently alter viability of the screened, but have also been a very useful tool in unraveling offspring. key molecular features of these unusual epigenetic processes. A similar screen has been performed in mice using a 1.1.1 The Dominant Screen variegating metastable epiallele. A metastable epiallele has transcriptional activity that is less stable than expected and Males homozygous for the transgene were treated with the is associated with changes in epigenetic state (Rakyan et al. chemical mutagen N-ethyl-N-nitrosourea (ENU), which 2002). The screen was thus performed with the hope of produces point mutations throughout the genome (Rin- finding novel epigenetic modifiers, creating useful new al- chik 1991). Mature germ cells are killed by the treatment, leles of known modifiers, and helping us to understand but point mutations are produced in the spermatogonial more about the remarkable features of metastable epialleles. stem cells, so when treated males recover fertility, they can Briefly, the activity state of metastable epialleles varies be bred and their G1 offspring screened for dominant- among genetically identical individuals brought up in the acting mutations. In essence, the mice were screened for same environment called variable expressivity, and is par- alterations in transgene silencing, assuming that any such ticularly sensitive to the epigenetic state of the locus. They alterations would be attributable to mutations in genes also display variegation (i.e., different expression states whose products are important in establishing epigenetic within one tissue type). These phenomena are discussed marks (see Fig. 1 for overview of the screen). in more detail in Section 2. More than 4000 G1 offspring have been screened and 40 Several studies suggested that using a variegating meta- strains have been isolated with transgene expression more stable epiallele in a mutagenesis screen would be a good than 2 standard deviations away from the mean of nonmu- approach. First, both transgenes and endogenous metasta- tant transgenic offspring (E. Whitelaw, pers. comm.). Each ble epialleles show strain specific differences in the extent of these strains possesses heritable dominant-acting muta- of the variegation or variable expressivity, consistent with tions. This represents a dominant functional mutation rate trans-acting genetic variants altering expression of these of 1 in 100. These mutations have been named Mommes alleles (Wolff 1978; Belyaev et al. 1981; Allen et al. 1990; (Blewitt et al. 2005). The dominant mutations are called Weichmanand Chaillet 1997; Sutherland et al. 2000; Chong MommeD1-40 and details of some of these are shown in et al. 2007). Second, the extent of variegation at metastable Table 1. Identification of the mutations that segregate in epialleles is shifted by altering the dose of some proteins 20 of the 30 lines has now been reported (Chong et al. known to be involved in epigenetic reprogramming such 2007; Ashe et al. 2008; Blewitt et al. 2008; Daxinger et al.

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 3 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press

A ENU Treatment

Male recovers fertility X (10–20 wk)

GFP transgene variegated expression in red blood cells

B G0 offspring screening

SuVar Counts Counts Counts Counts GFP GFP GFP GFP Flow cytometry

Identification of a Momme

Enhancer of variegation—increase in silencing, decrease in proportion of GFP+ red blood cells Suppressor of variegation—decrease in silencing, increase in proportion of GFP+ red blood cells

Mapping the mutation C Positional cloning Fine mapping

Microsatellite markers MARKER CHR MUTANT WILD TYPE (historically) rs1238543 4 rs9872356 4 OR rs1238745 4 3.2 Mb SNP array rs1239875 4 rs1245349 4 (current) # mice 155 12 8 2 110 10 3 2

F2 offspring Mutant or wild type by GFP expression

GGGCCCAAC A T AAA AAAC D Mutation identification

Candidate gene sequencing

(historically) Wild-type Homozygous Heterozygous OR

PLNIEVPKISLLLSSHS I D* F AV FLDVSS VRGLK P LNIEVPKISLLLSSHS I N FFAV LDVSS VRGLK exome capture and deep sequencing 5′ -CCTCTCAACATTGAGGTCCCCAAAATCAGCCTCCACAGCCTCATTCTCGACTTTTCAGCAGTGTCCTTTCTTGATGTTTCTTCAGTGAGGGGCCTTAAA-3′ 3′ -GGAGAGTTGTAACTCCAGGGGTTTTAGTCGGAGGTGTCGGAGTAAGAGCTGAAAAGTCGTCACAGGAAAGAACTACAAAGAAGTCACTCCCCGGAATTT-5′ (current) 3′ -GGAGCGTTGTAACTCCAGGGGTTTTAGTCGGAGGTGTCGGAGTAAGAGT*T-5′ 3′ -GTTGTAACTCCAGGGTTTTTAGTCGGAGGTGTCGGAGTAAGAGTTGAAAA-5′ 3′ -AACTCCAGGGTTTTTCGTCGGAGGGGTCGGAGTAAGAGTTGAAAAGTCGT-5′ 5′ -ctccaggggttttagtcggaggtgtcggagtaagagttgaaaagtcgtca-3′ 3′ -CCAGGGGTTTTAGTCGGAGGTGTCGGAGTAAGAGTTGAAAAGTCGTCACA-5′ 5′ -ggggttttagtcggaggtgtcggagtaagagttgaaaagtcgtcagagga-3′ 3′ -TTTTTGGTGGGAGGTGTCGGAGTAAGAGTTGAAAAGTCGTCACAGGAAAG-5′ 3′ -TTTAGTCGGAGGTGTCGGAGTAAGAGTTGAAAAGTCGTCACAGGAAAGAA-5′ 3′ -GTCGGAGGCGTCGGACTAAGAGTTGAAAAGTCGTCACAGGAAAGAACTAC-5′ 5′ -cggaggtgtcggagtaagagttgaaaagtcgtcacaggaaagaactacaa-5′ 3′ -GGGGGGGTCGGAGTAAGAGTTGAAAAGTCGTCACAGGAAAGAACTACAAA-5′ 5′ -gaggtgtcggagtaagagatgaaaagtcgtcacaggaaagaactacaaag-3′ 3′ -GGGTCGGAGTAAGAGTTGAAAAGTCGTCACAGGAAAGAACTACAAAGAAG-5′ 5′ -tcggagtaagagttgaaaagtcgtcacaggaaagaactacaaagaagtca-3′ 3′ -GAGTAAGAGTAGAAAAGTCGTCACAGGAAAGAACTACAAAGAAGTCACTC-5′ 5′ -agagttgaaaagtcgtcacaggaaagaactacaaagaagtcactccccgg-3′ 3′ -GTTGAAAAGTCGTCACAGGAAAGAACTACAAAGAAGTCACTCCCCGGAAT-5′

Figure 1. Momme screen for modifiers of epigenetic reprogramming. (A) GFP transgenic males are treated with ENU (now the G0 generation), left to recover fertility, and then bred with GFP transgenic females to produce G0 offspring. (B) A drop of blood is taken from all G0 offspring at weaning and analyzed by flow cytometry to measure GFP expression in erythrocytes. Analysis is performed to look for variations in the extent of variegation of transgene expression. In this instance, an example of an individual with an enhancer of variegation phenotype is illustrated in the third mouse analyzed. (C) Animals with alterations in transgene variegation are backcrossed for two generations to allow mapping of the causative mutation. Mapping is performed with microsatellite markers or single-nucleotide polymorphism (SNP) arrays, and fine mapping followed up with large numbers of phenotypically mutant or wild- type animals, using additional SNPs or microsatellite markers. (D) The linked point mutation is then identified via exome capture (i.e., genomic DNA input selected using mouse exonic probes) followed by deep sequencing or by candidate gene sequencing.

4 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

Table 1. Summary of Momme mutants produced in dominant screen for modiﬁers of epigenetic reprogramming Effect on Homozygous Name variegation lethality Gene Mutation Chromosome Reference(s) Human homolog MommeD1 Suppressor Females E10, Smchd1 C Tmakes Stop Chr 17 Blewitt et al. 2005, 2008 SMCHD1, mutated some male in FSHD2 adults viable MommeD2 Suppressor E8–E9 Dnmt1 C A in Exon 25, Chr 9 Chong et al. 2007 DNMT1 Thr Lys MommeD4 Enhancer E17–E18 Smarca5 T AinExon12, Chr 8 Blewitt et al. 2005; Chong SMARCA5 Trp Arg et al. 2007 MommeD5 Enhancer E8–E9 Hdac1 7 bp Del in Exon Chr 4 Blewitt et al. 2005; HDAC1 13, Frameshift Ashe et al. 2008 MommeD6 Suppressor E6–E8 D14Abb1e T C in Exon, Chr 14 Blewitt et al. 2005; Ashe FAM208A Leu His et al. 2008; L Daxinger and E Whitelaw, pers. comm. MommeD7 Enhancer E18.5 Hbb T C in poly (A) Chr 7 Brown et al. 2013 HBB b- signal thalassaemia MommeD8 Enhancer Some viable Rlf G TExon8, Chr 4 Ashe et al. 2008; L RLF adults Cys-Phe in Zn Daxinger and E ﬁnger Whitelaw, pers. comm. MommeD9 Enhancer E6–E7 Trim28 T C at splice Chr 7 Whitelaw et al. 2010a,b TRIM28 site of Intron 13 MommeD10 Enhancer Some viable Baz1b T GinExon7, Chr 5 Ashe et al. 2008 BAZ1B Williams adults Leu Arg Syndrome MommeD11 Suppressor E14 Klf1 T AinExon3, Chr 8 E Whitelaw, pers. comm. KLF1 Anaemia Cys Stop, MommeD12 Enhancer E5–E7 eIF3h T A splice site Chr 15 Daxinger et al. 2012 EIF3H –10 bp before Exon 5 MommeD13 Suppressor E5–E8 Setdb1 A G in Exon 20. Chr 3 L Daxinger and SETDB1 GWASa for Splicing defect E Whitelaw, pers. melanoma comm. MommeD14 Suppressor Some viable Dnmt3b T Cin3′ splice Chr 2 Youngson et al. 2013 DNMT3B ICF adults site of Exon 13. Syndrome Exon 13 skipped MommeD16 Enhancer Some viable Baz1b C TExon2, Chr 5 L Daxinger and E BAZ1B Williams adults Leu Pro Whitelaw, pers. comm. Syndrome MommeD17 Suppressor Some viable Setdb1 T C in Exon 21, Chr3 L Daxinger and E SETDB1 GWAS for adults Val Ala Whitelaw, pers. comm. melanoma MommeD19 Suppressor E5–E7 Smarcc1 T G Intron 10 Chr 9 L Daxinger and E SMARCC1 Link to –Exon11 splice Whitelaw, pers. comm. colon cancer site MommeD20 Suppressor E6–E8 D14Abb1e T Cat5′ splice Chr 14 L Daxinger and E FAM208A site of Intron 1 Whitelaw, pers. comm. MommeD21 Suppressor Morc3 T AExon1, Chr 16 L Daxinger and E MORC3 Met (start Whitelaw, pers. comm. codon) Lys MommeD23 Suppressor Some viable Smchd1 A TExon12 Chr 17 L Daxinger and E SMCHD1, mutated adults Arg Stop Whitelaw, pers. comm. in FSHD2 MommeD27 Suppressor Pbrm1 A G Exon 17, Chr 14 L Daxinger and E PBRM1 Tyr-Cys Whitelaw, pers. comm. MommeD28 Enhancer Some viable Rlf A G Intron 4, Chr 4 L Daxinger and RLF adults splicing defect E Whitelaw, pers. comm. MommeD30 Enhancer E10–E12 Wiz Single base Chr 17 L Daxinger and WIZ deletion Exon 5; E Whitelaw, pers. frameshift comm. Continued

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 5 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

Table 1. Continued Effect on Homozygous Name variegation lethality Gene Mutation Chromosome Reference(s) Human homolog MommeD31 Enhancer E6–E8 Trim 28 T AExon3 Chr 7 L Daxinger and TRIM28 Cys Ser in E Whitelaw, pers. Zn Finger comm. MommeD32 Suppressor E8–E9 Dnmt1 T C in Exon 29, Chr 9 L Daxinger and DNMT1 Leu Pro in E Whitelaw, pers. BAH domain comm. MommeD33 Suppressor Male Suvar39h1 A GatExon Chr X L Daxinger and SUVAR39h1 hemizygotes 1–Intron 1 E Whitelaw, pers. live splice site comm. MommeD34 Enhancer Some adults Rlf C AExon7, Chr 4 L Daxinger and RLF viable Cys Stop, E Whitelaw, pers. null allele comm. MommeD35 Enhancer Smarca5 A GExon9, Chr 8 L Daxinger and SMARCA5 Asn Ser, E Whitelaw, pers. comm. MommeD36 Suppressor Smchd1 G A Exon 42, Chr 17 L Daxinger and SMCHD1, mutated Glu Stop E Whitelaw, pers. in FSHD2 comm. MommeD37 Enhancer Smarca5 T C Exon 13, Chr 8 L Daxinger and SMARCA5 Leu Pro E Whitelaw, pers. comm. MommeD38 Enhancer No eIF3h G AExon7, Chr 15 Daxinger et al. 2012 EIF3H homozygotes Arg Stop at 3 wk aGWAS,genome-wide association studies.

2012; Youngsonet al. 2013; Daxinger and E Whitelaw, pers. wild-type offspring was consistent with embryonic lethality comm.) and we will discuss some of these findings, along of homozygous mutants (Blewitt et al. 2005; Ashe et al. with the characterization of the strains, in sections 1.1.2– 2008). The semidominant nature of the observed pheno- 1.1.5 and 1.2. Briefly, mutations have been identified in types is consistent with epigenetic processes being dose-de- the genes coding for proteins known to play a role in epi- pendent as they are in lower organisms (Schotta et al. 2003). genetic processes: DNA methyltransferases, Dnmt1 and For each of the MommeD strains, the timing of homo- Dnmt3b; histone deacetylase, Hdac1; chromatin-remodel- zygous embryonic lethality has been established. Homozy- ing factors, Smarca5 (Snf2h), Smarcc1, Pbrm1, and Baz1b gous lethality is observed to some extent in all MommeD (WSTF); histone methyltransferases, Setdb1 and Su- pedigrees (see Table 1). For MommeD1, MommeD8, Mom- var39h1; basal transcriptional machinery, Trim28 (KAP1); meD10, MommeD14, MommeD16, MommeD17, Momme- the transcription factor Klf1; and importantly in genes not D23, MommeD28, MommeD33, and MommeD45 some previously identified to have a role in epigenetic silenc- homozygotes survive to weaning, whereas for most other ing, such as Smchd1 and Rlf (see Fig. 2 and Daxinger et strains, it appears that all homozygotes die in utero or as al. 2013). neonates. The lethality of the majority of mice homozygous for MommeD mutations indicates the importance of the encoded proteins for normal development. In particular, 1.1.2 The Importance of Mommes in Development MommeD1 and MommeD10 have provided substantial in- For each of the MommeD strains, heterozygous intercrosses sights into mammalian developmental processes and are were used to look for semidominant phenotypes and poten- discussed in Sections 1.1.2.1 and 1.1.2.2. tial homozygous embryonic lethality. In all cases, the mu- 1.1.2.1 MommeD1. MommeD1 is unique among the tations were found to be semidominant with respect to MommeDs because it displays female-specific embryonic transgene expression because either a third phenotypic class lethality; homozygous males are born in normal numbers, of transgene expression was observed at weaning (Mom- but only half survive to adulthood, whereas homozygous meD1, MommeD8, MommeD10, MommeD14, MommeD16, females die around E (embryonic day) 10.5 (Blewitt et MommeD17,andMommeD23) or the ratio of mutant to al. 2005) caused by a failure of X inactivation (Blewitt

6 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

BASAL TRANSCRIPTIONAL MACHINERY MommeD9 & 31: Trim28

Ac Ac Ac Ac

Active locus

HISTONE DEACETYLATION CHROMATIN-REMODELING COMPLEXES—opening MommeD5: HDAC1 MommeD10 & 16: Baz1b part of chromatin-remodeling complexes MommeD4, 35 & 37: Smarca5 SWI/SNF chromatin remodeler

H3 H4

H2A H2B

Naïve locus

DNA METHYLATION HISTONE METHYLATION MommeD2 & 32: Dnmt1 maintenance DNA methyltransferase MommeD13 & 17: Setdb1 H3K9 methyltransferases MommeD14: Dnmt3b de novo DNA methyltransferase MommeD33: Suv39h1 MommeD1, 23, & 36: Smchd1 may direct DNA methylation?

CHROMATIN REMODELLING COMPLEXES—closing MommeD19: Smarcc1 SWI/SNF chromatin remodeler MommeD27: Pbrm1 DNA binding unit of SWI/SNF

Me Me Me Me Me Me Me Me Me Me Me Me Me Me Me Me Me Me TRANSCRIPTION FACTOR MommeD11: Klf1 Me Me Me Me Me Me Me Me Me Me Inactive histone methylation

Ac Histone acetylation Inactive locus Me Methylated CpG residue

Figure 2. Summary of the point of action of MommeD mutants. An active, naı¨ve, and inactive chromatin template are shown, and all of the MommeDs in Table 2 are indicated; red indicates that they are suppressor mutations and so the wild-type protein represses expression, whereas green indicates that they are enhancer mutations and so the wild- type protein activates expression. et al. 2008). A nonsense mutation in a novel gene, called through study of a genetrap null allele of Smchd1. Muta- structural maintenance of chromosome hinge domain- tions in Smchd1 have been identiﬁed in a further two containing 1 (Smchd1) linked to the alteration in trans- MommeD pedigrees that display similar attributes (Table 1). gene silencing. This nonsense mutation results in non- Smchd1MommeD1/MommeD1 mutant female embryos initiate sense-mediated mRNA decay of the Smchd1 transcript. X inactivation normally, as shown by normal Xist ex- Mutation of Smchd1 was shown to cause both the altera- pression and H3K27me3 accumulation; however, they tion in transgene silencing and female-speciﬁc lethality do not acquire DNA methylation at the CpG islands of

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 7 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw genes on the inactive X, a characteristic of genes subject to X Baz1b. Study of this strain shows that Baz1b has a hitherto inactivation. Furthermore, both the embryo and the extra- uncharacterized role in craniofacial development, and sug- embryonic tissues display up-regulation of some genes nor- gests that reduction in the Baz1b protein contributes to the mally subject to X inactivation (Blewitt et al 2008). The elfin facial features of WBS patients (Ashe et al. 2008). WBS embryo proper normally undergoes random X inactivation, patients also show hypersocial, anxious personalities that whereas the extraembryonic tissues undergo paternally im- are associated with structural alterations in a number of printed X inactivation (elaborated in Brockdorff and Tur- parts of the brain (Jabbi et al. 2012). Whether haploinsuffi- ner 2014). So, these results indicate that Smchd1 is critical ciency for BAZ1B is involved in this phenotype is not for both of these X inactivation processes. Although the known. The contribution of each of the 28 genes commonly mechanisms of X inactivation have been studied for de- deleted in WBS is controversial, so the Baz1b mouse model cades, the process by which silencing occurs is still not provides atool to studysome features of this syndrome. This completely understood, so the addition of a novel player study also reveals that at least some of the WBS features have in this field opens the door to many new functional studies. an epigenetic component. Smchd1 was named because of the presence of a car- boxy-terminal SMC hinge domain, which is normally found in the canonical SMC proteins, SMC1-6. These pro- 1.1.3 Mommes Influence Expression at the Agouti teins form heterodimers that make up part of the cohesin Viable Yellow Allele and condensin complexes that are important for chromo- One of the primary aims of the Momme screen was to use some conformation during cell division and the SMC5/6 the mutants to learn more about the unusual features of complex is involved in DNA repair. It is interesting to pos- metastable epialleles. Because a variegating transgene had tulate how Smchd1 may function at the molecular level been used for the original screen, some of the MommeD during X inactivation. It may belong to a new class of pro- strains(MommeD1-5)werecrossedwithastraincarryingthe teins that bridge the gap between epigenetic control and Agouti viable yellow (Avy) allele, a single copy endogenous chromosome structure. The unbiased approach of the mu- metastable epiallele (Blewitt et al. 2005; Chong et al. 2007). tagenesis screen has enabled the identification of a novel The Avy allele is driven by an upstream retrotransposon protein involved in epigenetic gene silencing and critical and displays variable expressivity, subtle parent-of-origin for X inactivation. Because X inactivation is an epigene- effects, and transgenerational epigenetic inheritance. tic mechanism exclusive to higher organisms, this valid- More specifically, the cryptic promoter within the LTR of ates screening in the mammalian system. Very recently, the integrated IAP drives inappropriate expression of the Smchd1’s role as a tumor suppressor has been described Agouti gene (see Fig. 3A) (Michaud et al. 1994; Perry et al. (Leong et al. 2013), and it is mutated in human facioscapu- 1994). Agouti is normally expressed for just a short period lomuscular dystrophy type 2 (Lemmers et al. 2012). These in the hair growth cycle causing a switch in pigment ex- results show the broader relevance to human health of novel pression from brown or black to yellow. Normal Agouti epigenetic factors identified in screens. expression results in a subapical yellow band on a dark 1.1.2.2 MommeD10. The Momme screen has eluci- hair, which en masse results in a brown appearance of the dated a new role for the known epigenetic modifier Baz1b. coat, termed agouti. When Agouti is expressed throughout Baz1b is a mammalian bromodomain-containing pro- the hair growth cycle, an entirely yellow hair shaft is protein, which is part of two different chromatin-remodeling duced. Mice carrying the Avy, Ahvy,orAiapy alleles all pro- complexes (WINAC SWI/SNF and WICH ISWI-contain- duce some mice with a variegated coat made up of agouti- ing) bound at promoters and replication foci (Bozhenok and yellow-colored patches—a phenotype termed mottled et al. 2002; Kitagawa et al. 2003). Baz1b is mutated in the (see Fig. 3B,C). The agouti-colored patches occur because MommeD10 strain of mice, which displays subtle craniofa- sometimes the IAP LTRis epigenetically silenced and tran- cial abnormalities in the heterozygotes and more severe de- scriptional control reverts to the normal promoter. Just fects in the few surviving homozygotes (Ashe et al. 2008). like with X inactivation, the silent state is stable and heri- The observed abnormalities are reminiscent of those seen in table over hundreds of cell divisions, so one stochastic patients with Williams Beuren syndrome (WBS) or Wil- silencing event of the IAP LTR early in development can liams syndrome, and Baz1b is one of a linked group of 28 result in a patch of daughtercells, all showing normal Agou- genes that are commonly heterozygously deleted in WBS ti expression. patients. The MommeD10 point mutation causes an amino Offspring were scored for coat color and phenotyped acid substitution in a highly conserved region of Baz1b, for expression of the GFP transgene (used to infer mutant which appears to destabilize the protein product. The or wild-type status for the MommeD in question when the Baz1bMommeD10 allele is the first available mutant allele for mutation was not yet identified). In many cases, a shift was

8 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press

A The Agouti viable yellow allele B IAP

Agouti coding exons

Tissue-specific and developmentally regulated promoters

C Phenotypes of the Agouti viable yellow allele

Euchromatic state in all cells Yellow

Mosaic of euchromatic and Mottled heterochromatic states

Heterochromatic state in all cells Pseudoagouti

D Ac Ac Ac Ac H3 H4

Me Me H4K20 trimethylation Me H2A H2B Active Agouti viable yellow allele Ac Histone acetylation

Me Me Me Me Me Me Me Me Me Me Methylated CpG residue Me Me Me Me Me Me Me

Inactive Agouti viable yellow allele

Figure 3. Features of the Agouti viable yellow allele. (A) The Agouti viable yellow allele (not to scale) has an intracisternal A particle (IAP) insertion (striped box) in pseudoexon 1a (gray box), 100 kb upstream of the Agouti coding exons (black). IAP long terminal repeats (LTRs) are shown as arrowheads, and transcriptional start sites + shown as arrows. (B) An agouti wild-type mouse with an A allele has a brown coat color phenotype with a two- toned hair shaft, in which the base appears black in color, and closer to the tip shows a yellow color (see cartoon in C). This phenotype occurs because the Agouti gene, producing a yellow fur color, is only transiently expressed in the hair follicle microenvironment. The Avy allele, if expressed ubiquitously from the IAP promoter, produces a phenotype with an entirely yellow hair shaft. There is a spectrum of coat color phenotypes observed, however, because of the Avy allele ranging from completely yellow when the IAP LTR is active in all cells, to mottled due to patches of active and inactive cells, and ﬁnally animals with an agouti colored coat, called pseudoagouti, which are indistinguishable from wild-type agouti animals because of silenced IAP LTRin all cells. (Reproduced from Morgan et al. 1999.) (C) Mice with the Agouti viable yellow allele in a euchromatic state in all cells appear yellow, whereas those with the allele in a heterochromatic state in all cells appear agouti in color, termed pseudoagouti. Mice that are mosaic for the heterochromatic and euchromatic allele appear mottled. (D) Summary of the epigenetic marks found at an active or inactive Agouti viable yellow allele. The inactive allele is hypermethylated and enriched for H4K20 trimethylation. The active allele is hypomethylated and enriched for acetylated residues of histone H3 and H4 tails.

9 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw observed in the spectrum of Avy coat color phenotypes seen 1.1.4 Paternal Effect Genes in the mutant compared with the wild-type offspring. In A new and unusual phenomenon that may be the result of general, the shift in variable expressivity was consistent with incomplete erasure of epigenetic marks between genera- the alteration in expression of the GFP transgene; for ex- tions was detected when studying the coat colors of mice ample, if an increase in GFP expression was observed, then MommeD2 vy in MommeD colonies. For both Dnmt1 and Smar- mutant animals more frequently displayed an active A MommeD4 ca5 , the wild-type offspring of heterozygous mu- allele than wild-type animals. These shifts sometimes, but tant fathers have a different spectrum of Avy coat color not always, had concomitant changes in DNA methylation phenotypes when compared with wild-type offspring of at the LTRthat controls expression at the Avy locus (Blewitt wild-type parents (Chong et al. 2007). This is shown in a et al. 2005). These results show that the mutant proteins simple diagram in Figure 4. These offspring are genetically (Smchd1, Dnmt1, Snf2h, Hdac1, and others unpublished) identical and differ only in the untransmitted genotype of are important not only for transgene silencing but also for the male parent. These paternal effects may be attributable silencing at an endogenous metastable epiallele driven by to alterations in chromatin packaging or RNA populations a retrotransposon. This is reminiscent of the finding that in genetically wild-type sperm that is a result of their pro- many dominant modifiers of retrotransposon insertion al- duction in a genetically compromised male parent that leles in the fly are also dominant modifiers of PEV (Fodor then has an effect in trans on the Avy allele inherited from et al. 2010). the mother. Smarca5 and Dnmt1 are expressed in testes and Some of MommeDs(MommeD1-4) showed complex in stage VII pachytene spermatids, which is before segrega- alterations in the penetrance at Avy including sex-specific tion of homologous chromosomes to produce haploid effects. For example, female-specific effects were reported spermatids (La Salle et al. 2004; Chong et al. 2007). So, it for both MommeD1 and MommeD2. In both cases, mutant is feasible that development of sperm in a Smarca5 or females showed a shift in penetrance that reflects a greater Dnmt1 depleted environment could result in epigenetic likelihood to have an active Avy allele, compared with their alterations to the wild-type set of chromosomes. Further- littermate wild-type females (consistent with the role of more, the progeny of each spermatogonium remain con- MommeD1 and MommeD2 as suppressors of variegation), nected by cytoplasmic bridges (channels connecting the but this was not seen in the males (Blewitt et al. 2005). cytoplasms of adjacent cells), which allows transcript shar- These female-specific effects are likely to occur because of ing, providing another opportunity for wild-type haploid the involvement of the mutated proteins, subsequently sperm to have depleted levels of Smarca5 or Dnmt1. identified as Smchd1 and Dnmt1, respectively, in X inacti- Although maternal effect genes have previously been vation (Sado et al. 2000; Blewitt et al. 2008). It was proposed reported in mammals and other organisms (Zheng and that the MommeD1 and MommeD2 proteins bind the in- Liu 2012), this was the first report of paternal effect genes active X chromosome, and the inactive X acts as a sink for in mammals. Interestingly, the suppressors and enhancers such repressor proteins, leaving fewer repressor molecules of variegation identified in the Drosophila PEV screens also free to perform autosomal silencing compared with an XY displayed paternal effects (Fitch et al. 1998). Our results male cell (Blewitt et al. 2005). In support of this argument, alter the way we think about the transmission of pheno- significantly more yellow females than yellow males are typic traits from parent to offspring. The laboratory mouse seen in the Avy wild-type colony (Morgan 1999; Blewitt has provided an opportunity to observe these interesting et al. 2005). It has been known for many years that even effects. before sexual differentiation, female mammalian embryos are smaller than their male counterparts (Burgoyne et al. 1.1.5 Recessive Screen 1995; Ray et al. 1995). In combination, these findings suggest that an epigenetic difference driven by the presence or As part of the Momme mutagenesis screen, 160 G1 male absence of an inactive X chromosome may be responsible progeny with no detectable change in transgene expression for some of the phenotypic differences between the sexes. were screened for recessive mutations in modifiers of var- This idea has recently been extended using a different meta- iegation. A pedigree was produced from each founder male stable epiallele, the hCD2 variegating transgene (see Table by backcrossing the founder with four of his daughters 2), and in mouse samples that have varying sex chromo- to produce at least 32 offspring. Using this approach, some complements: XX, XY, XO, XXY (Wijchers et al. .84% of recessive modifiers of variegation would theoret- 2010). They found that several hundred autosomal genes ically be detected. Seven of the 160 pedigrees showed evi- are sensitive to X chromosome complement. It will be in- dence of a recessive mutation, comprising both enhancers teresting to look at gene expression in humans with sex and suppressors of variegation (Blewitt 2004; Vickary- chromosome aneuploidies. ous 2005). These recessive mutants generally showed more

10 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

Table 2. Table of metastable epialleles and reporter alleles Transgenerational Variable Parent-of-origin epigenetic Allele Type Phenotype Variegation expressivity effects inheritance Reference(s) Avy Endogenous; IAP Coat color, Yes Yes Yes Yes Morgan et al. insertion obesity, 1999; Wolff diabetes 1978 Aiapy Endogenous; IAP Coat color, Yes Yes Yes ? Michaud et al. insertion obesity, 1994 diabetes Ahvy Endogenous; IAP Coat color, Yes Yes Yes ? Argeson et al. insertion obesity, 1996 diabetes Axinfu Endogenous; IAP Tail kinks Yes Yes Yes Yes Rakyan et al. insertion 2003; Reed 1937; Belyaev et al. 1981 Axial defects Endogenous; Grhl2 Spina biﬁda ? Yes Yes Yes? Essien et al. up-regulation of 1990; Brouns unknown origin et al. 2011 Disorganization Endogenous; Skeletal ? Yes No ? Hummel et al. potential Gata4 abnormalities 1959; White disruption et al. 1995 mCabpIAP Endogenous; IAP mCabp ? Yes ? ? Druker et al. insertion expression 2004 cm Endogenous; IAP Defective Yes ? ? ? Porter et al. 1991 insertion Tyrosinase expression; mottled coat color Cm1OR Endogenous; IAP Defective Yes Wu et al. 1997 insertion Tyrosinase expression; mottled coat color 239B Transgene LacZ-expressing Yes Yes Yes No Kearns et al. red blood cells 2000 MTa#7 Transgene; insertion LacZ-expressing Yes Yes No Yes Sutherland et al. into L1 element red blood cells 2000 TKZ751 Transgene LacZ-expressing ? Yes Yes No Allen et al. 1990 somatic cells RSVIgmyc Transgene Transgene ? Yes Yes No Weichman and expression in Chaillet 1997 myocytes BLG transgenic Transgene; b-lactoglobulin Yes Yes ? ? Dobie et al. 1996 lines 7 and 45 centromeric expression Tyr-SV40E Transgene SV40 expression Yes Yes ? ? Bradl et al. 1991 in melanocytes hCD2-1.3b Transgene; Human CD2 Yes No No No Festenstein et al. centromeric expression in 1996 T cells GFP1 Transgene GFP-expressing Yes No No No Preis et al. 2003 red blood cells GFP3 Transgene GFP-expressing Yes No Yes No Preis et al. 2003 red blood cells

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 11 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

wt Avy/a

P < 0.005

Smarca5MommeD4/+ Avy/a Dnmt1MommeD2/+ Avy/a

Smarca5+/+ Dnmt1+/+

Genotype Phenotype

IAP

Agouti coding exons Avy Yellow

Agouti coding exons a Mottled

Figure 4. Paternal effects of Smarca5MommeD4 and Dnmt1MommeD2 on the Avy allele. (A) For the Avy progeny from a cross with a paternal wild-type mouse, the expected ratio of yellow versus mottled mice is 6:4. (B) Heterozygous males for Momme mutants of Smarca5 (left) and Dnmt1 (right) are bred with female yellow Avy heterozygotes. Offspring are scored for their coat color and genotyped for the MommeD mutations and the Avy allele. Only those offspring carrying the Avy allele and the wild-type Smarca5 or Dnmt1 allele are shown for simplicity. The spectrum of phenotypes was compared with that observed for wild-type FVB/N males bred with yellow females (see A). Wild- type offspring of mutant males have a significantly different spectrum of coat color phenotypes than offspring of wild-type males, termed a paternal effect. Both the Smarca5 and Dnmt1 wild-type offspring produced from a MommeD heterozygous father show a skew toward less mottled mice, indicating that the haploinsufficiency of these genes in the spermatogonia (diploid sperm progenitors) is sufficient to affect the epigenomic programming of the Avy locus in the next generation, and hence, the phenotype. (Adapted from Chong et al. 2007.)

subtle phenotypes than the dominant mutations, and were possible that any of the mutants identified are detected frequently sterile. Only one of the recessive mutations has because of hematopoietic defects rather than epigenetic been studied in detail (MommeR1). disruption per se. This was ruled out for MommeR1,as MommeR1 was identified as a suppressor of variegation Foxo3aMommeR1/MommeR1 animals had normal levels of re- (i.e., mutants have increased expression of the GFP trans- ticulocytes setting them apart from the Foxo3a2/2 ani- gene). MommeR1 has the interesting phenotype of prema- mals. It is not yet clear whether Foxo3a has a direct or ture ovarian failure in all female homozygotes, and ovarian indirect effect on transgene variegation, but this new line teratomas in one-sixth of these animals. A missense muta- of mice has linked premature ovarian failure with ovarian tion was identified in Forkhead box protein 3a, Foxo3a, teratomas (Youngson et al. 2011), which have previously which through complementation tests with a null allele of only been associated with late stage ovarian failure, and Foxo3a, was shown to be the causative mutation (Youngson future studies should unravel the role of the protein in et al. 2011). Foxo3a is a forkhead transcription factor, gene silencing. and so not traditionally viewed as an epigenetic modifier, although it does have a role in hematopoiesis. One possi- 1.2 ENU Mutagenesis Produces bility was that an altered blood cell compartment, in par- Hypomorphic Alleles ticular, an increased level of reticulocytes that display higher GFP transgene expression, could account for the One of the advantages of ENU mutagenesis is that it pre- elevation in transgene expression. Indeed, it is formally dominantly produces point mutations. These point mu-

12 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics tations are more likely to produce hypomorphic alleles than development. In this case, the investigators took advantage standard knockout approaches, and can be informative of a recessive phenotypic marker closely linked to Brg1— about critical residues within a protein structure. This curly whiskers (cw). cw/cw males were treated with ENU, also provides an opportunity to create an allelic series, bred with wild-type females, and G1 offspring bred with which can be useful when studying genes for which com- animals that had the cw marker in cis with the Brg1 null plete knockout is early embryonic lethal. The Momme allele (Fig. 5). 525 pedigrees were scored for the presence of screen has produced new mutations in nine known epige- mice with curly whiskers; one pedigree was found in which netic modiﬁers for which null alleles already exist (Dnmt1, no such mice were found at weaning, suggesting that a Dnmt3b, Hdac1, Trim28, Smarca5, Setdb1, Smarcc1, Pbrm1, mutation had occurred that failed to complement the and Suvar39h1; Chong et al. 2007; Ashe et al. 2008; White- Brg1 null allele. The investigators went on to identify the law et al. 2010a; Youngson et al. 2013; L Daxinger and mutation in the ATPase domain of Brg1, which uncoupled E Whitelaw, pers. comm.). The missense mutations in the ATPase activity from the chromatin-remodeling ac- Dnmt1, Trim28, and Smarcc1 appear to destabilize the pro- tivity of Brg1. The mutant Brg1 protein was unable to tein, whereas Hdac1 has a small deletion and subsequent remodel chromatin, although the ATPase activity itself frameshift mutation that alters the very carboxyl termi- was unchanged, and Brg1 still localized to chromatin and nus of the protein. The Smarca5 missense mutation, the assembled into its usual SWI/SNF complex. Homozyg- Dnmt3b splice-site mutation, and the Setdb1 missense mu- ous Brg1 hypomorphs survived until mid-gestation, allow- tation appear to behave as hypomorphic alleles. ing many more studies than could be performed using An independent ENU screen was performed to create the null mice. The observation of 1/525 pedigrees with a hypomorphic allele of brahma related gene 1, Brg1 a functional mutation in Brg1 is roughly consistent with (Bultman et al. 2005). Brg1 null mice fail to implant (Bult- the estimate of 1/700 animals with functional muta- man et al. 2000), precluding the study of later stages of tion (dominant or recessive) in a speciﬁc gene of interest

A Pedigree with no visible mutation in Brg1—mice with B Pedigree with a mutation in Brg1—no live mice with curly whiskers curly whiskers ENU ENU

x x

cw cw + + cw cw + + + + + + Brg1 Brg1 Brg1+ Brg1+ Brg1 Brg1 Brg1+ Brg1+

x x

cw + cw + cw + cw + Brg1+ Brg1+ Brg1null Brg1+ Brg1ENU Brg1+ Brg1null Brg1+

++ +cw +cw cwcw ++ cw + cw + cwcw + + null + + null Brg1 Brg1 Brg1 Brg1 Brg1 + Brg1 Brg1 + Brg1 Brg1 + Brg1 + Brg1 null Brg1 + Brg1 ENU Brg1 + Brg1ENU Brg1null

Curly whiskers No live animals with curly whiskers

Figure 5. The breeding strategy for the Brg1 hypomorph ENU screen. Only mouse chromosome 9 is shown indicating the cw locus and Brg1 loci, which are 1 cM apart. cw/cw males were treated with ENU, and bred with wild-type females. Offspring were mated with Brg1null heterozygotes, which carried the cw allele in cis with the null mutation. Resultant offspring were screened for the presence of pups with curly whiskers (in red). (A) Pups were observed with curly whiskers indicating no functional mutation in Brg1.(B) One pedigree was found in which no pups had curly whiskers indicating that a mutation had occurred that failed to complement the Brg1null allele. A hypomorphic ENU induced mutation in Brg1 was later found. (Adapted from Bultman et al. 2005.)

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 13 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

(Hitotsumachi et al. 1985). This approach can be used to transcription polymerase chain reaction) followed by re- produce an allelic series without prior knowledge of the striction fragment length polymorphism analysis of the domain structures of the protein, and shows the power of X-linked gene Pctk in samples taken from females at wean- ENU-induced point mutations over null alleles produced ing (Plenge et al. 2000). Females with aberrant skewing were by homologous recombination. further interrogated by allele specific qRT-PCR analysis of Pctk, Pgk1, and Xist. Three strains were identified and two have been followed in more detail. They have been shown to 1.3 A Screen for Genes Involved in X Inactivation have one and two segregating mutations, respectively. The Choice investigators have shown that, in each case, skewing is al- Another ENU mutagenesis screen for epigenetic modifiers ready present at E6.5 and E7.5 and is similar across all was established specifically to identify autosomal factors tissues. These results suggest that the causative mutations important in X inactivation choice (Percec et al. 2002; Per- result in changes in primary choice, rather than skewing cecet al. 2003). X inactivation choice is discussed in detail in due to secondary nonrandom X inactivation (when cells Brockdorff and Turner (2014). Briefly, the choice of which die because of the choice of which X is inactivated). Fur- X chromosome to inactivate happens in the embryo proper thermore, the mutations alter random X inactivation in the shortlyafter implantation. It has been known for someyears embryo proper, but not imprinted X inactivation in the that choice is governed by the Xce locus on the X chromo- extraembryonic tissues suggesting that the role of these some; there are four different Xce alleles found in mice with factors, as least in heterozygotes, has some specificity for differing likelihoods that the X chromosome bearing that random X inactivation. Each of the mutations has been allele will become the inactive X chromosome. Xcec and linked to specific large autosomal regions. These mutants Xced are found in wild mice such as Mus castaneus; Xced have been called X inactivation autosomal factor 1, 2, and 3 is the strongest allele and the least likely to be inactivated, (Xiaf1, 2, and 3). It is anticipated that the underlying mu- whereas Xcec, Xceb, and Xcea are of decreasing strength. tations will be identified soon. Xceb and Xcea are found in various inbred Mus musculus mice. Xce heterozygotes display a predictable pattern of X 1.4 Other Screens That Have Identified Mutations inactivation skewing away from the 1:1 ratio seen in homo- in Epigenetic Modifiers zygotes and expected of a random process. The pattern of skewing, although predictable, displays considerable vari- Two ENU mutagenesis screens, performed to identify all ance; for example, Xcec/a females show on average 25% of genes within a defined chromosomal region, have identi- theircells with an active Xcea carrying chromosome, but the fied homozygous embryonic lethal mutations in epigenetic range is from 5% to 45%. modifiers. In each case, a long chromosomal deletion or Little is known about the trans-acting factors that in- inversion was used that covered the region of interest, and teract with the Xce or other elements on the X chromo- additionally covered a visible marker (the albino locus or some. It is clear that activation of Xist expression on just one the Rump White locus; Rinchik et al. 1990; Wilson et al. of the two X chromosomes is required, and it is thought 2005), which allows for simple marking of the deleted or that pairing of the two X chromosomes immediately before mutated chromosome, similar to the curly whiskers allele the initiation of X inactivation may play some role in al- described above in Section 1.2. lowing cross talk and subsequent unequal distribution of The first of these screens isolated point mutations in the trans-acting factors between the two X chromosomes. Both Embryonic ectoderm development (Eed) gene (Schumacher the chromatin insulator CTCF and pluripotency factor et al. 1996), which we now know is a core component of Oct4 play a role in this process. Elegant live-cell imaging Polycomb repressive complex 2 (PRC2) in mammals. The experiments show that X chromosome pairing is frequently null and hypomorphic alleles of Eed produced in this followed by up-regulation of a negative regulator of Xist. screen not only helped to clone the gene, but were also However, it is still unclear how the different Xce alleles the only available mutant or knockout alleles of Eed avail- influence this process (see Brockdorff and Turner 2014 able for many years. Study of these animals has informed for details). us about the role of PRC2 in embryonic development, The ENU screen performed by Percec and colleagues X inactivation, genomic imprinting, embryonic stem cell (Percec et al. 2002; Percec et al. 2003) took advantage of pluripotency, and differentiation and hematopoiesis (see the predictable skewing of X inactivation present in Xcea/c Grossniklaus and Paros 2014). and Xceb/c females. They screened for a skewing proportion In the second of these screens (Wilson et al. 2005), a more than 2 standard deviations away from the mean found mutation was identified in the Promyelocytic zinc finger in the control population by performing RT-PCR (reverse gene, Plzf (Ching et al. 2010). PLZF is a BTB/POZ domain-

14 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics containing zinc finger protein, which is involved in tran- 2.1 Variegation scriptional repression. Although this is the third mutated Variegation is the differential expression of a gene among allele of Plzf, the ENU induced mutant separates the func- cells of the same type. The variegated appearance of the tional requirements of PLZF in different biochemical path- leaves of some plants, and the coats of many dogs and ways, and the missense mutation is potentially informative cats, are good examples of variegation that we observe on about BTB domain function. a day-to-day basis. In some cases this variegation has a Two mutations in epigenetic modifiers have been iden- genetic origin (e.g., mosaic genetic states alter the color of tified in one of the ENU mutagenesis screen run at the corn kernels), but in many cases this is not the explana- Walter and Eliza Hall Institute of Medical Research, Aus- tion. Because pigmentation patterns can be easily visual- tralia. This screen was designed as a sensitized suppressor ized, coat color gene mutations have been excellent models screen, in which mice with thrombocytopenia and a hema- for the study of mosaicism in mammals. Since the 1970s, topoietic stem cell (HSC) defect are subjected to ENU mu- mouse geneticists, in particular Beatrice Mintz, became tagenesis, and screens performed to identify genes that fascinated by this phenomenon (Mintz 1970). When dis- when mutated suppress these phenotypes. This screen has cussing mottled mice she and her colleagues write (Bradl produced a point mutation in the histone acetyltransferase et al. 1991): Ep300 (Carpinelli et al. 2004) along with one in the PRC2 member Suz12 (Majewski et al. 2008). Studies on the Suz12 Their patterns are apparently due to phenoclones—phenotypi- mutant line identified a novel role of PRC2 in the restriction cally different but genetically identical clones in which the same gene must be yielding more than one kind or amount of product of HSC function (Majewski et al. 2008). in mitotic lineages of the same cell type. The phenomenon is Finally,a large-scale Australian ENUscreen, run to iden- presumably not limited to pigment cells. tify genes important in immune function, has produced At the level of gene expression, the existence of these another mutation in Eed. This screen used flow cytome- phenoclones implies that there is some element of chance try-based screening of peripheral blood cell types and num- involved in whether a gene is expressed and that epigenet- bers (Jun et al. 2003), and identified that the Eed mutant, ic control of gene expression is an inherently stochastic called Leukskywalker, had elevated leukocyte numbers. Al- process. though unpublished, this mutant is available to researchers who may be interested in such a strain. Without knowing the identity of the majority of genes 2.1.1 Variegation at Endogenous Alleles involved in epigenetic processes, it will be difficult to ful- in the Mouse ly understand the molecular basis of epigenetic control. Historically, coat color variegation was first studied in fe- Therefore, screening approaches such as those described male mice heterozygous for X-linked mutations of the mot- earlier in Section 1 will remain a valuable part of epigenetic tled locus, later identified as Atp7a. Because hemizygous research in the future. males never displayed variegation, and in fact were frequently embryonic lethal, the phenomenon was described as sex-linked variegation. The variant coat color patches in 2 EPIGENETIC PHENOMENA IN INBRED the case of the Atp7a mutations are determined by whether MOUSE COLONIES the wild-type or mutant X chromosome is subject to X Some endogenous murine alleles such as the Agouti coat inactivation in those particular cells. Analysis of the phe- color alleles can show unusual patterns of expression. These notype of these animals along with other critical studies led are termed metastable epialleles (Rakyan et al. 2002) be- Mary Lyon to propose the Lyon hypothesis of X inactiva- cause the transcriptional activity at these alleles is less stable tion, in particular, that the process is random with respect than expected, and this is associated with changes in epi- to which of the two X chromosomes is inactivated in any genetic state. This section summarizes the key features of particular cell (for further discussion see Brockdorff and the unusual expression patterns of, specifically, Agouti Turner2014). Studyof these animals has informed us about viable yellow (Avy), an allele of Axin, Axin fused (Axinfu), the timing of X inactivation in development, and the stable, and some transgenic reporters (see Table 2). The behavior heritable nature of that transcriptional silencing. of these alleles will be used to define and describe varie- Some autosomal mutations in the mouse also result in gation, variable expressivity, transgenerational epigenetic variegated expression—for example, alleles of the Agouti inheritance, parent-of-origin effects, and position effects. gene: Agouti viable yellow (Avy; Perry et al. 1994), Agouti Together, these phenomena have taught us much about intracisternal A particle yellow (Aiapy; Michaud et al. transcriptional gene silencing and the probabilistic nature 1994), and Agouti hypervariable yellow (Ahvy; Argeson et of epigenetic processes. al. 1996). Each of these alleles is characterized by a stable

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 15 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw retrotransposon insertion. Similarly, a number of alleles at Analysis of variegated transgenes has allowed more de- the albino locus, cm and cm10R, display coat color variega- tailed study of the epigenetic mechanisms contributing to tion that is the result of stable retrotransposon insertions in variegation than was technically possible with variegated cis (Porter et al. 1991; Wu et al. 1997). Importantly, these coat color alleles. One major advantage of the transgenic autosomal alleles that display variegation show that loci lines carrying transgenes directed to express in blood cells throughout the genome can display this interesting char- has been the ability to observe variegated gene expression at acteristic of stochastic transcriptional silencing previously a single cell level rather than in cell populations (Robertson thought to be restricted to genes on the X chromosome. et al. 1995; Festenstein et al. 1996; Sutherland et al. 1997; Kearns et al. 2000; Sutherland et al. 2000; Preis et al. 2003). These studies have allowed consideration of different mod- 2.1.2 Variegation at Transgenes in the Mouse els of gene expression control (e.g., different models of en- An unusually high proportion of transgenes also show var- hancer action). Importantly, they have conclusively shown iegated expression (Allen et al. 1990; Festenstein et al. 1996; that gene expression can be controlled by probabilistic Garrick et al. 1996; Weichman and Chaillet 1997; Kearns events rather than gradients of external factors (Sutherland et al. 2000; Sutherland et al. 2000). The reason why trans- et al. 1997). As techniques improve for transcriptome genes have a tendency to variegate is not fully understood. studies in single cells, it is possible that many loci through- There are several possible explanations, including high copy out the genome will be shown to display similar variegated number within transgene arrays, integration of the trans- expression patterns, which may influence cellular behavior. genes next to heterochromatin (see Sections 2.6 and 2.7), and in some cases, the foreign nature of sequences used 2.1.3 Underlying Molecular Mechanisms within the transgene sequences (for review, see Martin and Whitelaw 1996). Transgenes composed of sequences The molecular mechanisms of transgene silencing or si- with bacterial codon usage (e.g., LacZ) tend to show a lencing of endogenous metastable epialleles such as Agouti greater degree of silencing than those with mammalian co- viable yellow are still not completely understood; however, don usage (Kearns et al. 2000; Sutherland et al. 2000; Preis they show many features common to heterochromatin, et al. 2003). One possibility is that DNA sequences re- including DNA methylation, chromatin packaging alter- cognized as foreign are silenced by mechanisms normally ations including alterations in histone methylation, and reserved for integrated viruses and other invading transpos- acetylation (Fig. 3D) (Elliott et al. 1995; Garrick et al. able elements (Yoder et al. 1997), and in this light, it is 1996; Morgan et al. 1999; Sutherland et al. 2000; Blewitt interesting to note that all autosomal endogenous alleles et al. 2005; Blewitt et al. 2006; Dolinoy et al. 2010). Because that display variegation are associated with retrotransposon these alleles show many of the same features as silencing insertions (as mentioned for the Agouti and Albino alleles at other sites throughout the genome, transgenes and en- in Section 2.1.1). dogenous alleles can be used as reporter alleles to test the Transgenes have also been shown to display addition- effect of genetic or environmental alterations on epigenetic al peculiar phenotypes such as age-dependent silencing silencing. These sorts of studies have been performed using (Robertson et al. 1996) (i.e., transgene expression levels transgenic lines (Festenstein et al. 1999; Gaudet et al. 2004; reduce with age). Because the promoters of tumor suppres- Blewitt et al. 2005; Chong et al. 2007; Ashe et al. 2008; sor genes are frequently silenced and show increased DNA Whitelaw et al. 2010a; Youngsonet al. 2011), and the Agouti methylation in cancer, and the greatest risk factor for cancer viable yellow allele (Wolff 1978; Waterland and Jirtle 2003; is increasing age, age-dependent silencing of transgenes Gaudet et al. 2004; Blewitt et al. 2005; Dolinoy et al. 2006; raises the questions: Is this indicative of what occurs in Chong et al. 2007; Kaminen-Ahola et al. 2010). Some of the normal process of aging, and does it predispose to these studies were discussed in Section 1. cancer development? It has now been shown that normal human prostate tissue (Kwabi-Addo et al. 2007) and nor- 2.2 Variable Expressivity mal mouse tissues (Maegawa et al. 2010) show widespread age-dependent DNA methylation changes. Furthermore, In addition to variegation, many transgenic lines and en- two groups (Rakyan et al. 2010; Teschendorff et al. 2010) dogenous metastable epialleles display variable expressivi- have recently shown that specific DNA methylation changes ty (see Table 2), which can be defined as the differential occur in polycomb group protein target genes during aging, expression of a gene between individuals. The term was perhaps indicative of a preneoplastic state. Epigenetic de- originally used to describe a scenario, frequently observed regulation appears to be a common feature of aging, first in humans, in which patients with the same genetic aberra- suggested by studies on transgene variegation in mice. tion show variable severity of disease. In these cases, genetic

16 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics heterogeneity at unlinked modifier loci (or quantitative that are epigenetically metastable in humans. These studies trait loci [QTLs]) has been presumed (and in many cases, have primarily focused on studying genetically identical found) to be the cause of the phenotypic diversity. Variable monozygotic twins (Bell and Spector 2011). Monozygotic expressivity at metastable epialleles occurs even when the twins show a high degree of discordance for many traits. In organisms are inbred (ostensibly isogenic) and reared in one case, a twin pair was identified discordant for a caudal controlled environments, suggesting that something in ad- duplication anomaly. Given the role of Axin in axis forma- dition to genetic heterogeneity and environmental factors tion (Zeng et al. 1997), AXIN1 was sequenced but no mu- is involved. tation was found (Kroes et al. 2002). Rather, the affected Mice carrying the Agouti viable yellow allele show a twin displayed DNA hypermethylation of the AXIN1 pro- spectrum of coat colors from normal agouti colored coats moter, compared with the unaffected twin (Oates et al through various degrees of mottling to a completely yellow 2006). This suggests that, at least in this case, stochastic coat (Fig. 3B,C). The spectrum of phenotypes is associated epigenetic events that occur in the context of genetic iden- with a range of Agouti expression; in the agouti mice the tity can influence human disease (Oates et al. 2006). IAP LTR has been silenced in all cells, leaving the normal Further studies have looked for differences in DNA meth- Agouti promoters and enhancers to drive expression of ylationwithin monozygotic twin pairs genome-wide. There Agouti only during a short phase of the hair-growth cycle. is evidence fora small degree of metastability in DNA meth- In the yellow colored mice, the active IAP LTR drives con- ylation between monozygotic twins (Fraga et al. 2005; Mill stitutive expression of Agouti. Because Agouti is a signaling et al. 2006; Kaminsky et al. 2009) and some evidence that molecule in other pathways, these yellow mice also show this increases with age (Fraga et al. 2005). These types of other pleiotropic effects such as diabetes and obesity. The studies provide some support for the notion that epigenetic stochastic epigenetic silencing of the IAP LTRin some mice, silencing may contribute to variable expressivity observed but not others, results in phenotypes ranging from the in humans, and possibly even to complex disease (Water- extremes of obese and diabetic with a yellow coat through land et al. 2010; Pujadas and Feinberg 2012). In addition to to indistinguishable from wild type (Fig. 3B). genome-wide association studies, epigenome-wide associ- A similar situation occurs for several other metastable ation studies are now beginning to identify loci in which epiallelesthat are also driven by IAP insertions: forexample, DNA methylation state might contribute tovariable expres- Axin fused (Axinfu) (Vasicek et al. 1997) and murine CDK5 sivity in humans (Rakyan et al. 2011). activator binding protein IAP (mCABPIAP; Druker et al. 2004). The IAP insertion at Axin is in intron 6 of the 2.3 Parent-of-Origin Effects gene. When the IAP LTR is active, it produces a truncated transcript and protein, which correlates with a kinked back- Metastable epialleles also frequently display subtle parent- bone (Rakyan et al. 2003). Animals carryingthe Axinfu allele of-origin effects (Rakyan et al. 2002). These differ from range in phenotype from having a severely kinked back- traditional parental imprinting because they are not strictly bone, most easily observed in the tail, to wild type in ap- monoallelically expressed. Rather, the allele is preferentially pearance. As for Avy, this correlates with epigenetic state and in the active state when inherited from one parent, com- activity of the fused IAP LTR(Rakyan et al. 2003). The IAP pared with the other. In cases in which transgenes display insertion at mCABPIAP was also found in intron 6 of the variegation, the degree of variegation is reduced by a small gene. In this case, no visible phenotype associated the allele amount following transmission from the mothercompared has been reported; indeed, this allele is found in all C57BL/6 with the father (Reik et al. 1987; Sapienza et al. 1987; Preis animals. Similar to Axinfu, mCABPIAP produces an aberrant et al. 2003; Williams et al. 2008). In cases in which the truncated transcript that initiates in the IAP LTR found in allele displays variable expressivity, the spectrum of pheno- intron 6, in addition to thewild-type transcript. The expres- types shifts slightly depending on the parent-of-origin. This sion of the truncated transcript varies between genetically has been observed for several transgenes (Allen et al. 1990; identical littermates and is associated with hyper or hypo- Weichmanand Chaillet 1997;Kearns et al. 2000), along with methylation of the IAP LTR (Druker et al. 2004). In the all of the Agouti coat color alleles (Wolff 1978; Duhl et al. above cases, these transcriptional differences are thought 1994;Argeson et al. 1996; Morgan et al.1999) andthe Axinfu to be the direct result of the different epigenetic marks allele (Reed 1937; Belyaev et al. 1981; Rakyan et al. 2003). found at the retrotransposable element in each case. For the Axin fused allele, the allele is 30% more likely to be These findings raise the intriguing possibility that some active (producing a kinked tail) following paternal trans- proportion of the variable penetrance observed in humans mission, compared with maternal transmission. The con- is due to stochastic epigenetic silencing rather than effects verse istrue for the Agouti viable yellowallele, which displays of QTLs. Several approaches have been taken to identify loci subtle paternal imprinting; there is a 15% greater likelihood

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 17 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

Avy/aAa/a a/a vy/a

Paternal Avy transmission Maternal Avy transmission

Phenotype

Yellow Mottled Pseudoagouti

Figure 6. Subtle parent-of-origin effects at the Agouti viable yellow allele. Pedigrees showing the proportion of yellow, mottled, and pseudoagouti animals following paternal or maternal transmission of the allele. The a/a offspring have been excluded from these pedigrees for simplicity. More yellow offspring are observed following maternal transmission of the allele indicating a lower proportion of IAP LTR silencing. This illustrates what is termed a subtle parent-of-origin effect as incomplete erasure of the IAP LTRheterochromatic environment appears to occur during male transmission of the Avy allele. of expression when the allele is inherited from the mother, sion of the Axinfu allele (Reed 1937; Belyaev et al. 1981; compared with the father (see Fig. 6). Rakyan et al. 2003). Therefore, transmission via either These small changes in expression, dependent on par- germline is capable of producing transgenerational epige- ent-of-origin, were detectable because of the sensitivity of netic inheritance. the assays used to measure transgene expression, or the The Avy allele is the most frequently studied case of dramatic effect on phenotype observed with endogenous transgenerational epigenetic inheritance. Yellow-colored metastable epialleles. The observation of this type of pa- females, which have an active Avy allele, produce 60% yel- rental effect in mice suggested that that the maternal and low-coated offspring and 40% mottled offspring. Geneti- paternal genomes might be differentially marked at more callyidenticalagouti-coloredfemales,withasilentAvyallele, loci than only those subject to traditional parental imprint- produce 40% yellow-colored, 40% mottled, and 20% ing (Pardo-Manuel de Villena et al. 2000; Ashe 2006). In- agouti-colored offspring. The epigenotype of the mother deed, very recent data using RNAseq in the embryonic and at the Avy IAP LTR alters the spectrum of epigenotypes ob- adult mouse brain suggests that .1300 genes display par- served in her offspring (see Fig. 7) (Morgan et al. 1999). ent-of-origin allelic effects (Gregg et al. 2010). Although it Using mice as a model, it has been possible to exclude is not yet clear how widespread such effects are in humans, confounding effects such as intrauterine environment or these results again show that epigenetic phenomena, first distant genetic alterations influencing epigenetic state. The suggested by studies using reporter strains of mice, may be Avy allele displays transgenerational epigenetic inheritance, true more broadly throughout the genome. even in the inbred C57BL/6 background, which greatly reduces the likelihood of a genetic contribution to this effect (Morgan et al. 1999). Morgan and coworkers also 2.4 Transgenerational Epigenetic Inheritance showed that the intrauterine environment of the yellow via the Gamete female was not responsible. Because mice expressing Agou- Perhaps the most controversial of the effects observed at ti constitutively (i.e., a yellow mouse) display pleiotropic metastable epialleles is transgenerational epigenetic inher- effects, including obesity and diabetes, it was possible itance via the gametes. This term describes any situation in that development within the uterus of such a mouse may which the epigenetic state of the parent influences the epi- alter the epigenotype and, hence, phenotype of her off- genetic state of the offspring, but cannot be attributed to spring. To rule out this possibility, zygotes were trans- cis-ortrans-acting genetic variation or mutations. This is a ferred from yellow dams into congenic females that did fascinating phenomenon that has intrigued biologists for not carry the Avy allele. The transferred embryos developed many decades. the same spectrum of coat colors (60% yellow coats, 40% Transgenerational epigenetic inheritance has been re- mottled coats) as those pups that were left in the uterus ported following maternal transmission of transgenes (Al- of their original yellow dams. This work showed that the len et al. 1990; Kearns et al. 2000; Sutherland et al. 2000), uterine environment was not responsible for the unusual maternal transmission of the Avy allele (Wolff 1978; Mor- transgenerational epigenetic inheritance observed at the gan et al. 1999), and both maternal and paternal transmis- Avy allele.

18 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

Mothers with the same genotype but different epigenotype, hence phenotype

AB a/a Avy/a a/a Avy/a

Distinct patterns of transgenerational inheritance: a different spectrum of phenotypes are observed based on the epigenotype of the mother

Phenotype

Yellow Mottled Pseudoagouti

Figure 7. Transgenerational epigenetic inheritance at the Agouti viable yellow allele. Pedigrees showing the proportion of yellow, mottled, and pseudoagouti animals following transmission of the allele from an Avy heterozygous mother with a yellow (A) or pseudoagouti (B) phenotype. The a/a offspring have been excluded from these pedigrees for simplicity. Yellow mothers, with an IAP promoter in an open chromatin configuration making the locus constitutively active in all cells, produce more yellow offspring (60%) than pseudoagouti mothers do (40%), in which the IAP promoter is in a repressive chromatin configuration and the promoter is consistently off. This difference is the visual result of transgenerational epigenetic inheritance, in which somehow the epigenetic configuration of the Agouti IAP LTR affects the ratio of phenotypes obtained in the next generation.

Epigenetic reprogramming (as discussed in Hochedlin- it is not normally observed (Blewitt et al. 2006), raising ger and Jaenisch 2014) occurs between generations during the possibility that an altered makeup of PRC1 maydecrease both primordial germ cell development and preimplan- the efficiency of epigenetic reprogramming during early tation development. The instances of transgenerational development, and implicate histone modifications in the epigenetic inheritance described above suggest that incom- process. The molecular processes underlying transgen- plete erasure of epigenetic marks between generations can erational epigenetic inheritance via the gametes remain un- sometimes occur. Working with mouse models, it is possi- clear. Many reviews have been written about this with the ble to study the transmission of epigenetic marks through most recent being Daxinger and Whitelaw (2012). We have germ cell development and embryonic development, and summarized the incomplete erasure model in Figure 8A. this has been performed for DNA methylation at the Avy allele (Blewitt et al. 2006). The removal of the vast majority 2.5 Paramutation-Like Effects in Mice of histone proteins from sperm DNA made DNA methylation an ideal candidate for the epigenetic mark passed In a distinct, but related phenomenon, described as para- between generations. mutation-like (see Pikaard and Mittelstein Scheid 2014 for The Avy (Blewitt et al. 2006) and Axinfu alleles (Rakyan paramutation in plants), RNA molecules in the sperm of et al. 2003) show similar levels of DNA methylation in ma- mice have been implicated in the epigenetic events that are ture gametes as in the somatic tissues suggesting that DNA heritable across multiple generations (Rassoulzadegan et al. methylation at the controlling IAP LTRs escape erasure 2006). The initial study involved a mutant allele of the Kit during primordial germ cell development just as the bulk locus. Kit is a receptor tyrosine kinase that is required for of IAPs do throughout the genome (Lane et al. 2003). stem cells of various types. Reduced levels of Kit are associ- However, DNA methylation was found to be completely ated with white spotting, particularly at the extremities, cleared at the Avy allele during preimplantation develop- in many mammalian species including mice, and this is ment (Blewitt et al. 2006). These studies suggest that caused by reduced numbers of melanocyte precursors dur- DNA methylation is not the inherited mark at the Avy allele, ing development. Heterozygosity for the Kittm1Alf allele is although they do not exclude the possibility that DNA associated with a white tail. The investigators found that methylation may transmit the heritable mark to other epi- following the cross of mice heterozygous for this null allele genetic modifications that are yet to be an analyzed. at the Kit locus, Kittm1Alf, with congenic wild-type litter- This study also found that haploinsufficiency for mates, someofthegeneticallywild-type offspring hadwhite Mel18, a polycomb repressive complex 1 (PRC1) compo- tails. Abnormal Kit mRNA transcripts were detected in nent, produced epigenetic inheritance in cases in which sperm. The investigators injected microRNAs known to

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 19 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

Model A: Model B: Incomplete erasure of a maternal epigenetic mark Germline-transmitted RNA reestablishes epigenetic mark

Somatic cells (2N)

Clearing and resetting of epigenetic marks

RNA molecule m m Primordial m m m m germ cells (2N) m m Establishment of Establishment of tissue-specific tissue-specific epigenetic marks epigenetic marks

m Gametes m m m (1N)

Rapid, active DNA m Fertilized egg Rapid, active DNA m demethylation of (2N) demethylation of paternal genome paternal genome Slow DNA demethylation of maternal genome

Sperm (germline)-transmitted Incomplete erasure m RNA directs reestablishment of epigenetic marks of epigenetic mark m Blastocyst (early embryo) m m m m

Epigenetic inheritance through the maternal germline Epigenetic inheritance through the paternal germline

Figure 8. Models for transgenerational epigenetic inheritance. Only the gene showing transgenerational epigenetic inheritance is shown for simplicity. “m” denotes the epigenetic mark, in which pink shows the maternal mark and blue shows the paternal mark. Epigenetic reprogramming occurs in two phases (indicated by gray arrows): ﬁrst, in primordial germ cell development and second, in preimplantation development. Model A: Incomplete erasure occurs of either a DNA methylation mark or histone mark. In a scenario in which transgenerational epigenetic inheritance is observed following maternal transmission (left gamete), some cells do not show complete removal of the epigenetic mark, and retain the inherited epigenetic mark when compared with other cells in the same blastocyst. In the instance in which there is complete erasure of epigenetic marks at a locus and transgenerational epigenetic inheritance does not occur all cells of the blastocyst would be unmarked (not shown). Model B: Germline RNA causes reestablishment of the inherited epigenetic mark. In a scenario in which transgenerational epigenetic inheritance is observed following paternal transmission, an RNA molecule is transmitted from the primordial germ cell to the mature sperm. This RNA is then transmitted to the fertilized egg and causes reestablishment of the paternal epigenetic mark in some cells that inherit the RNA molecule. target Kit mRNA into zygotes and found white tails in cases there is no evidence that these are bona ﬁde cases of the subsequent offspring. This implicates some long-lived transgenerational epigenetic inheritance via the gametes. In effect of the RNA, presumably by epigenetic alteration humans, it is impossible to adequately control for genetic of gene expression. Similar paramutation-like phenomena makeup or confounding in utero effects, or to directly ob- have been detected at some other loci in the mouse (Wagner serve a failure to clear epigenetic marks in germ cell devel- et al. 2008; Grandjean et al. 2009). Given the heritability opment and preimplantation development. Therefore, it of this effect, these results suggest that the RNA content remains to be established whether transgenerational epige- of sperm may play some role as an intermediary in trans- netic inheritance truly exists in humans. The existence of generational epigenetic inheritance. We have summarized such an effect in humans would have profound implica- that a model of transgenerational epigenetic inheritance tions for studies on the heritability of human phenotype occurs via an RNA molecule in Figure 8B. and disease, and so there is considerable interest in this Phenotypic events reminiscent of transgenerational epi- area, and it is probably only through mouse studies that genetic inheritance have been reported in human popula- we will be able to elucidate the mechanistic basis for such tions (Lumey 1992; Pembrey et al. 2006), however, in these observations.

20 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

2.6 Position Effects 2.7 Repeat-Induced Gene Silencing

The term “position-effect variegation” (PEV) was coined The number of copies of a transgene that integrate as a for a phenomenon observed in Drosophila (see Elgin and concatemer at a single integration site also influences ex- Reuter 2013) and yeast (see Grunstein and Gasser 2013; pression. Although there appears to be some interaction Allshire and Ekwall 2014). PEV has mostly been studied between the integration site and the copy number on trans- using translocations of the white gene in flies. Expression of gene expression (Williams et al. 2008), in general, extremely the white gene was found to be dependent on its location high copy number transgenes show low or no expression with respect to pericentric heterochromatin. If the white per transgene copy. Garrick and colleagues (Garrick et al. gene (which produces red pigment in the eye) neighbored 1998) investigated this effect using Cre-mediated deletion heterochromatin, then the white gene expression was fre- of two very low-expression, high-copy-number (.100) quently variegated, producing a mosaic red and white eye. a-globin-driven LacZ transgenes. When these transgenes This was attributed to incomplete spreading of neighbor- were reduced to five copies and one copy, respectively, ing heterochromatin through the white gene (Henikoff the proportion of expressing red blood cells jumped from 1990). This interesting phenotype has been studied in clas- ,1% to .50% with a concurrent decrease in DNA meth- sic mutagenesis screens to identify key components in- ylation and opening of the chromatin structure at the volved in silencing and spreading of heterochromatin transgene. This finding was the first evidence that repeat- (Schotta et al. 2003). Transgenes in mice are particularly induced gene silencing, already reported in plants and flies susceptible to PEV, as described in Section 2.1; variegating (Henikoff 1998), exists in mammals. transgenes have been located in repeats (Sutherland et al. Does repeat-induced gene silencing influence the ex- 2000), near centromeres (Dobie et al. 1996), and near telo- pression of endogenous genes? There are many repeat- meres (Zhuma et al. 1999; Pedram et al. 2006; Gao et al. ed gene families in mammals (e.g., rRNA genes, histone 2007), all of which are heterochromatic regions. The im- genes). Although often arranged as tandem repeats, similar portance of this juxtaposition to heterochromatin is not to those seen at transgene arrays, these do not appear to be known. Nevertheless, these studies show that heterochro- subject to repeat-induced gene silencing. Instead, there are matin has the capacity to spread in mammals, as it does in specialist forms of epigenetic control that exist in these Drosophila. families (Keverne 2009). Interestingly, the D segment genes If the aim of making a transgenic mouse is to achieve of the immunoglobulin heavy chain genes are also arranged predictable expression of the gene under the control of a as tandem repeats and show many of the hallmarks of specific promoter, the integration site dependence of trans- regions subject to repeat-induced gene silencing found in gene expression is clearly a problem. Several groups have lower organisms and at centromeric heterochromatin, such used transgenesis to identify the sequence elements re- as antisense transcription and dimethylation of histone 3 quired to produce integration-site-independent transgene at lysine 9 (H3K9me2; Chakraborty et al. 2007). The in- expression (e.g., for the b-globin gene). In this case, there vestigators propose that the D segments may be subject to is interest in finding the controlling elements for gene ther- repeat-induced silencing. apy purposes. The sequence elements found to protect the For transgene silencing and silencing at centromeres transgene from PEV are known as locus control regions and telomeres, the repeats are clustered. In other circum- (LCRs). LCRs perform this function by somehow ensuring stances, such as for endogenous retroviruses or retrotrans- an open chromatin structure throughout the transgene, in posons, the repeats are spread throughout the genome with a cell lineage-dependent manner (Kioussis and Festenstein single copies in thousands of different locations. These 1997). There have been two main theories about how LCRs dispersed repeats are frequently silenced throughout devel- function: first, by direct contact between the LCR and the opment, presumably to protect the genome from the mu- promoter, or second, by LCRs producing a general opening tagenic effects that result from active mobile elements, and of chromatin in the region. With the development of new additionally to protect neighboring regions from the strong chromosome capture techniques (Carter et al. 2002), there promoters in these retroelements. The silencing in this is now some consensus that chromatin looping enables the case must occur in trans either by physical contact (e.g., b-globin LCR to interact with promoters. A recent study looping) or by diffusible elements such as RNA. When has found that such long-range DNA interactions are not this silencing is inefficient, it is possible for readthrough detectable in every cell, and so at least for the b-globin transcription to occur just as it does at the discussed en- locus, long-range interactions that occur in only a subset dogenous metastable epialleles (e.g., Avy, Axinfu, and of cells could be responsible for variegation (Noordermeer mCabpIAP). The IAPs found at these metastable epialleles et al. 2011). are all of a young class, in evolutionary terms, which may

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 21 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw explain their different behavior compared with the rest of Other disease? the complement of dispersed repeats. Most ancient dis- Genotype Behavior persed retroviral repeats have accumulated mutations that Body weight make them nonfunctional. Environment Stress response Embryonic development Stochastic 2.8 The Effects of Environment on Metastable epigenetic variation Craniofacial development Epialleles Figure 9. Contribution of stochastic epigenetic variation, genotype, The finding that expression at metastable epialleles can be and the environment to phenotype. Genotype, the environment influenced by the environment has stimulated much re- (sometimes by altering epigenotype), and stochastic epigenetic var- search. In the case of Avy, the decision about whether the iation can alter the phenotype of the mouse. Known and measurable locus will be “on” or “off” is made in the early postimplan- differences that are a consequence of altered epigenetic control in- tation embryo (Blewitt et al 2006), and the percentage of clude body weight, behavior, stress responses, craniofacial development, and embryonic development. yellow pups in a litter has been shown to decrease as a result of exposure of the female parent to a diet rich in methyl groups (folic acid, betaine, vitamin B12; Wolff et al. 1998; Cooney et al. 2002; Waterland and Jirtle 2003), genistein ground. Indeed, 80% of the variance in body weight of (Dolinoy et al. 2006), or ethanol (Kaminen-Ahola et al. inbred mice has been attributed to intangible variation 2010). It is interesting to compare these findings with those rather than genetic or environmental differences (Gartner made decades ago at the variegating wmottled locus in Dro- 1990). Gartner refers to this as “the third component”, the sophila. Increasing the temperature during development first and second being genetics and environment, respec- has been shown to suppress variegation (reduce silencing) tively. Other examples include incomplete penetrance of a at this locus (see Elgin and Reuter 2013). Another metasta- phenotype associated with a particular genotype in inbred ble epiallele, Axinfu, has also been shown to be sensitive to mice (Biben et al. 2000), and transmission ratio distortion maternal diet (Waterland et al. 2006). These findings em- owing to lethality of some but not all animals of a particular phasize the plasticity of the epigenome during early devel- genotype in inbred mice (Carpinelli et al. 2004; Blewitt et al. opment, providing an opportunity for the environment to 2005). Furthermore, regressing embryos are observed in permanently influence an individual’s phenotype. As such, isogenic lines of mice at far higher frequency than expected and because it will be possible to document the epigenome because of de novo lethal mutations, and independent of in a particular tissue of an individual in fine detail, the uterine position. All of these are examples of intangible epigenome may provide us with the capacity to infercertain variation at a tissue or organism level. The concept of in- facts about an individual’s past and better predict disease tangible variation has emerged bystudying inbred strains of risk for that individual (see Fig. 9). However, it remains mice. unclear how much of the epigenome in any particular tis- What is the molecular basis for this intangible variation sue will turn out to be both susceptible to environmental or phenotypic noise? Given the stochastic nature of gene events and then stable for life, both of which are required for expression that is responsible for the variable expressivity the epigenome to have predictive value. at a cellular level observed at many transgenes and endogenous metastable epialleles, one plausible explanation is that similar stochastic patterns of gene expression in early 2.9 “Intangible Variation” development subsequently maintained byepigenetic mech- Intangible variation, or developmental noise, is defined as anisms is responsible for many cases of developmental noise the variation in phenotype that is unable to be explained by (Blewitt et al. 2004). In addition to what is known about contributions from genotype and the environment alone metastable epiallele expression, there is nowample evidence (Falconer 1989). This type of variation is most easily ob- from single-cell analyses of gene expression in mammalian served when both genotype and environment are invariant cells that gene expression has a stochastic element and in- (e.g., in inbred strains of mice [or other laboratory organ- herent noise (e.g., Raj et al. 2006). Furthermore, increased isms] housed in standard conditions). Clearly, variable ex- variation at the cellular level in gene expression (Chi and pressivity can be considered a case of intangible variation. Bernstein 2009) and at the level of the whole organism in Intangible variation is more widespread than the few quantitative traits is observed when epigenetic modifiers are known instances of metastable epialleles. In inbred strains depleted (Whitelawet al. 2010b). This implicates epigenetic of mice, many quantitative traits conform to a bell-shaped mechanisms in buffering transcriptional noise, and conse- curve despite the fixed environment and genetic back- quent phenotypic noise (see Fig. 10).

22 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

Wild-type animals show a from studies in the laboratory mouse. This has been possi- normal distribution of ble fora numberof reasons: mice are relativelyeasy to rear in phenotypes for measurable quantitative traits controlled environments, they have been maintained as inbred strains for decades, facilities such as the Jackson Lab- oratories have maintained thousands of different inbred strains for distribution to the research community, the fully Increased intangible annotated genome sequence of the C57BL/6 mouse has variation in animals with lower levels of been available to all, and perhaps most importantly, their epigenetic modifiers genomes can be manipulated by transgenesis, homologous

Proportion of individuals recombination, and ENU mutagenesis. Geneticists using mice to understand and model hu- Mean man health and disease did not fully embrace the field of Phenotypic measure epigenetics until the 1990s. Although gene silencing was Figure 10. Intangible variation is increased in animals with decreased being studied before this in the particular cases of X inac- levels of epigenetic modifiers. Measurable phenotypes in wild-type tivation and parental imprinting, epigenetic processes at animals conform to a normal distribution (bell-shaped curve) autosomal genes were not generally considered even among around a mean measurement (black line). Animals with lower levels of epigenetic modifiers display increased variance in such measurable the developmental biologists. There were reports in the phenotypes (red line). literature of the unusual behavior of some autosomal alleles (e.g., Agouti viable yellow and Axin fused)—that is, that The outcomes of stochastic gene expression and devel- they displayed variable phenotypes in inbred strains, but opmental noise can be both positive and negative. Noise these were considered quirky. Wenow realize that epigenet- can result in cell death or embryonic lethality, but the nat- ic processes play an integral part of developmental pro- ural variance produced by stochastic gene expression can cesses in mammals and much of this has emerged from also be beneficial by providing plasticity (Eldar and Elowitz studying mice with mutations in the genes involved in 2010). In development, it is commonly believed that exter- laying down and reprogramming these marks. ENU mu- nal cues signal to homogeneous populations to begin dif- tagenesis screens have identified the genes and, possibly ferentiation. However, it is also possible that probabilistic more importantly, have provided models that can be used events lead to differences in gene expression that result in to study the consequences of disruption of these processes phenotypic differences in what are considered homoge- to phenotype. neous cell populations (Pujadas and Feinberg 2012). In Studies in mice have stimulated those studying human the hematopoietic system, it has been shown that het- health and disease to embrace the field of epigenetics and to erogeneity exists in clonal stem cell populations; impor- apply this knowledge to their specific areas of interest. tantly, this heterogeneity is not just for the expression Some human diseases are now considered to be, at least of an individual gene, but rather represents differences in part, the consequence of mutations in modifiers of epi- that exist throughout the genome. In this case, the noise genetic reprogramming (e.g., Williams syndrome and ICF in gene expression appears to help determine lineage choice syndrome). One of the benefits of using the mouse to study (Chang et al. 2008). Similar noise-dependent cell fate or epigenetic control is that one can study human-like pheno- developmental patterning decisions seem to exist in several types, such as behavior, memory, and learning. Colonies scenarios—for example, neural crest cell fate (Shah et al. of animals provide opportunities to study reproductive fit- 1996) and inner cell mass lineage choice (Morris et al. 2010; ness, germline epigenetic reprogramming, and the role Yamanaka et al. 2010). What we have learned about the of the epigenome in gene–environment interactions and stochastic mechanisms of gene expression from the study we can look forward to much progress in this area in the of variegating transgenes and metastable epialleles appears near future. to hold true elsewhere in the genome, and may have important roles in development and differentiation (Pujadas and Feinberg 2012). REFERENCES ∗ Reference is also in this collection.

3 SUMMARY AND FUTURE DIRECTIONS Allen ND, Norris ML, Surani MA. 1990. Epigenetic control of transgene Over the last 20 years we have learned much about the expression and imprinting by genotype-speciﬁc modiﬁers. Cell 61: 853–861. epigenetic processes that are integral to development and ∗ Allshire R, Ekwall K. 2014. Epigenetics in Saccharomyces pombe. Cold differentiation in mammals, and much of this has emerged Spring Harb Perspect Biol doi: 10.1101/cshperspect.a018770.

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 23 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

Argeson AC, Nelson KK, Siracusa LD. 1996. Molecular basis of the pleio- supraphysiological production of platelets in the absence of thrombo- tropic phenotype of mice carrying the hypervariable yellow (Ahvy) poietin signaling. Proc Natl Acad Sci 101: 6553–6558. mutation at the agouti locus. Genetics 142: 557–567. Carter D, Chakalova L, Osborne CS, Dai YF, Fraser P. 2002. Long-range Ashe A, Morgan DK, Whitelaw NC, Bruxner TJ, Vickaryous NK, Cox LL, chromatin regulatory interactions in vivo. Nat Genet 32: 623–626. Butterfield NC, Wicking C, Blewitt ME, Wilkins SJ, et al. 2008. A Chakraborty T, Chowdhury D, Keyes A, Jani A, Subrahmanyam R, Iva- genome-wide screen for modifiers of transgene variegation identifies nova I, Sen R. 2007. Repeat organization and epigenetic regulation of genes with critical roles in development. Genome Biol 9: R182. the DH-Cmu domain of the immunoglobulin heavy-chain gene locus. Ashe A, Whitelaw E. 2006. Imprinting: The hidden genome. In The Mol Cell 27: 842–850. implicit genome (ed. Caporale LH), pp. 282–293. Oxford University Chang HH, Hemberg M, Barahona M, Ingber DE, Huang S. 2008. Tran- Press, Oxford. scriptome-wide noise controls lineage choice in mammalian progen- ∗ Barlow DP, Bartolomei MS. 2014. Genomic imprinting in mammals. itor cells. Nature 453: 544–547. Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a018382. Chi AS, Bernstein BE. 2009. Developmental biology. Pluripotent chro- Bell JT, Spector TD. 2011. A twin approach to unraveling epigenetics. matin state. Science 323: 220–221. Trends Genet 27: 116–125. Ching YH, Wilson LA, Schimenti JC. 2010. An allele separating skeletal Belyaev DK, Ruvinsky AO, Borodin PM. 1981. Inheritance of alternative patterning and spermatogonial renewal functions of PLZF. BMC Dev states of the fused gene in mice. JHered72: 107–112. Biol 10: 33. Biben C, WeberR, Kesteven S, Stanley E, McDonald L, Elliott DA, Barnett Chong S, Vickaryous N, Ashe A, Zamudio N, Youngson N, Hemley S, L, Koentgen F, Robb L, Feneley M, et al. 2000. Cardiac septal and Stopka T, Skoultchi A, Matthews J, Scott HS, et al. 2007. Modifiers of valvular dysmorphogenesis in mice heterozygous for mutations in epigenetic reprogramming show paternal effects in the mouse. Nat the homeobox gene Nkx2-5. Circ Res 87: 888–895. Genet 39: 614–622. Blewitt ME. 2004. “Epigenetic reprogramming in the mouse.” PhD the- Cooney CA, Dave AA, Wolff GL. 2002. Maternal methyl supplements in sis, University of Sydney, Australia. mice affect epigenetic variation and DNA methylation of offspring. J Blewitt ME, Chong S, Whitelaw E. 2004. How the mouse got its spots. Nutr 132: 2393S–2400S. Trends Genet 20: 550–554. Daxinger L, Whitelaw E. 2012. Understanding transgenerational epige- Blewitt ME, Vickaryous NK, Hemley SJ, Ashe A, Bruxner TJ, Preis JI, netic inheritance via the gametes in mammals. Nat Rev Genet 13: Arkell R, Whitelaw E. 2005. An N-ethyl-N-nitrosourea screen for genes 153–162. involved in variegation in the mouse. Proc Natl Acad Sci 102: Daxinger L, Oey H, Apedaile A, Sutton J, Ashe A, Whitelaw E. 2012. 7629–7634. A forward genetic screen identifies eukaryotic translation initiation Blewitt ME, Vickaryous NK, Paldi A, Koseki H, Whitelaw E. 2006. Dy- factor 3, subunit H (eIF3h), as an enhancer of variegation in the namic reprogramming of DNA methylation at an epigenetically sen- mouse. G3 2: 1393–1396. sitive allele in mice. PLoS Genet 2: e49. Daxinger L, Harten SK, Oey H, Epp T, Isbel L, Huang E, Whitelaw N, Blewitt ME, Gendrel AV, Pang Z, Sparrow DB, Whitelaw N, Craig JM, Apedaile A, Sorolla A, Yong J, et al. 2013. An ENU mutagenesis screen Apedaile A, Hilton DJ, Dunwoodie SL, Brockdorff N, et al. 2008. identifies novel and known genes involved in epigenetic processes in SmcHD1, containing a structural-maintenance-of-chromosomes the mouse. Genome Biol 14: R96. hinge domain, has a critical role in X inactivation. Nat Genet 40: Dobie KW,Lee M, Fantes JA, Graham E, Clark AJ, Springbett A, Lathe R, 663–669. McClenaghan M. 1996. Variegated transgene expression in mouse Bozhenok L, Wade PA, Varga-Weisz P. 2002. WSTF-ISWI chromatin mammary gland is determined by the transgene integration locus. remodeling complex targets heterochromatic replication foci. EMBO Proc Natl Acad Sci 93: 6659–6664. J 21: 2231–2241. Dolinoy DC, Weidman JR, Waterland RA, Jirtle RL. 2006. Maternal Bradl M, Larue L, Mintz B. 1991. Clonal coat color variation due to a genistein alters coat color and protects Avy mouse offspring from obe- transforming gene expressed in melanocytes of transgenic mice. Proc sity by modifying the fetal epigenome. Environ Health Perspect 114: Natl Acad Sci 88: 6447–6451. 567–572. ∗ Brockdorff N, Turner B. 2014. Dosage compensation in mammals. Cold Dolinoy DC, Weinhouse C, Jones TR, Rozek LS, Jirtle RL. 2010. Variable Spring Harb Perspect Biol doi: 10.1101/cshperspect.a019406. histone modifications at the Avy metastable epiallele. Epigenetics 5: Brouns MR, De Castro SC, Terwindt-Rouwenhorst EA, Massa V,Hekking 637–644. JW, Hirst CS, Savery D, Munts C, Partridge D, Lamers W, et al. 2011. Druker R, Bruxner TJ, Lehrbach NJ, Whitelaw E. 2004. Complex patterns Over-expression of Grhl2 causes spina bifida in the Axial defects mu- of transcription at the insertion site of a retrotransposon in the mouse. tant mouse. Hum Mol Genet 20: 1536–1546. Nucleic Acids Res 32: 5800–5808. Brown FC, Scott N, Rank G, Collinge JE, VadolasJ, Vickaryous N, White- Duhl DM, Vrieling H, Miller KA, WolffGL, Barsh GS. 1994. Neomorphic law N, Whitelaw E, Kile BT, Jane SM, Curtis DJ. 2013. ENU mutagen- agouti mutations in obese yellow mice. Nat Genet 8: 59–65. esis identifies the first mouse mutants reproducing human b- Eldar A, Elowitz MB. 2010. Functional roles for noise in genetic circuits. thalassemia at the genomic level. Blood Cells Mol Dis 50: 86–92. Nature 467: 167–173. Bultman S, Gebuhr T, Yee D, La Mantia C, Nicholson J, Gilliam A, ∗ Elgin SCR, Reuter G. 2013. Position-effect variegation, heterochromatin Randazzo F, Metzger D, Chambon P, Crabtree G, et al. 2000. A Brg1 formation, and gene silencing in Drosophila. Cold Spring Harb Perspect null mutation in the mouse reveals functional differences among 5: a17780. mammalian SWI/SNF complexes. Mol Cell 6: 1287–1295. Elliott JI, Festenstein R, Tolaini M, Kioussis D. 1995. Random activation Bultman SJ, Gebuhr TC, Magnuson T. 2005. A Brg1 mutation that un- of a transgene under the control of a hybrid hCD2 locus control couples ATPase activity from chromatin remodeling reveals an essen- region/Ig enhancer regulatory element. EMBO J 14: 575–584. tial role for SWI/SNF-related complexes in b-globin expression and Essien FB, Haviland MB, Naidoff AE. 1990. Expression of a new mutation erythroid development. Genes Dev 19: 2849–2861. (Axd) causing axial defects in mice correlates with maternal pheno- Burgoyne PS, Thornhill AR, Boudrean SK, Darling SM, Bishop CE, type and age. Teratology 42: 183–194. Evans EP.1995. The genetic basis of XX-XYdifferences present before Falconer DS. 1989. Introduction to quantitative genetics, 3rd ed. Wiley, gonadal sex differentiation in the mouse. Philos Trans R Soc Lond B Biol New York. Sci 350: 253–260 [discussion 260–261]. Festenstein R, Tolaini M, Corbella P, Mamalaki C, Parrington J, Fox M, Carpinelli MR, Hilton DJ, Metcalf D, Antonchuk JL, Hyland CD, Miliou A, Jones M, Kioussis D. 1996. Locus control region function Mifsud SL, Di Rago L, Hilton AA, Willson TA, Roberts AW, et al. and heterochromatin-induced position effect variegation. Science 271: 2004. Suppressor screen in Mpl2/2 mice: c-Myb mutation causes 1123–1125.

24 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

Festenstein R, Sharghi-Namini S, Fox M, Roderick K, Tolaini M, Norton Kaminen-Ahola N, Ahola A, Maga M, Mallitt KA, Fahey P, Cox TC, T, Saveliev A, Kioussis D, Singh P. 1999. Heterochromatin protein 1 Whitelaw E, Chong S. 2010. Maternal ethanol consumption alters modifies mammalian PEV in a dose- and chromosomal-context-de- the epigenotype and the phenotype of offspring in a mouse model. pendent manner. Nat Genet 23: 457–461. PLoS Genet 6: e1000811. Fitch KR, YasudaGK, Owens KN, Wakimoto BT.1998. Paternal effects in Kaminsky ZA, Tang T, Wang SC, Ptak C, Oh GH, Wong AH, Feldcamp Drosophila: Implications for mechanisms of early development. Curr LA, Virtanen C, Halfvarson J, Tysk C, et al. 2009. DNA methylation Top Dev Biol 38: 1–34. profiles in monozygotic and dizygotic twins. Nat Genet 41: 240–245. Fodor BD, Shukeir N, Reuter G, Jenuwein T. 2010. Mammalian Kearns M, Preis J, McDonald M, Morris C, Whitelaw E. 2000. Complex Su(var) genes in chromatin control. Annu Rev Cell Dev Biol 26: patterns of inheritance of an imprinted murine transgene suggest 471–501. incomplete germline erasure. Nucleic Acids Res 28: 3301–3309. Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML, Heine- Keverne B. 2009. Monoallelic gene expression and mammalian evolu- Suner D, Cigudosa JC, Urioste M, Benitez J, et al. 2005. Epigenetic tion. Bioessays 31: 1318–1326. differences arise during the lifetime of monozygotic twins. Proc Natl Kioussis D, Festenstein R. 1997. Locus control regions: Overcoming het- Acad Sci 102: 10604–10609. erochromatin-induced gene inactivation in mammals. Curr Opin Ge- Gao Q, Reynolds GE, Innes L, Pedram M, Jones E, Junabi M, Gao DW, net Dev 7: 614–619. Ricoul M, Sabatier L, VanBrocklin H, et al. 2007. Telomerictransgenes Kitagawa H, Fujiki R, Yoshimura K, Mezaki Y, Uematsu Y, Matsui D, are silenced in adult mouse tissues and embryo fibroblasts but are Ogawa S, Unno K, Okubo M, Tokita A, et al. 2003. The chromatin- expressed in embryonic stem cells. Stem Cells 25: 3085–3092. remodeling complex WINAC targets a nuclear receptor to promoters Garrick D, Sutherland H, Robertson G, Whitelaw E. 1996. Variegated and is impaired in Williams syndrome. Cell 113: 905–917. expression of a globin transgene correlates with chromatin accessibil- Kroes HY, Takahashi M, Zijlstra RJ, Baert JA, Kooi KA, Hofstra RM, van ity but not methylation status. Nucleic Acids Res 24: 4902–4909. Essen AJ. 2002. Twocases of the caudal duplication anomaly including Garrick D, Fiering S, Martin DI, Whitelaw E. 1998. Repeat-induced gene a discordant monozygotic twin. Am J Med Genet 112: 390–393. silencing in mammals. Nat Genet 18: 56–59. Kwabi-Addo B, Chung W,Shen L, Ittmann M, Wheeler T,Jelinek J, Issa JP. Gartner K. 1990. A third component causing random variability beside 2007. Age-related DNA methylation changes in normal human pros- environment and genotype. A reason for the limited success of a 30 tate tissues. Clin Cancer Res 13: 3796–3802. year long effort to standardize laboratory animals? Lab Anim 24: Lane N, Dean W, Erhardt S, Hajkova P,Surani A, Walter J, Reik W. 2003. 71–77. Resistance of IAPs to methylation reprogramming may provide a Gaudet F, Rideout WM 3rd, Meissner A, Dausman J, Leonhardt H, mechanism for epigenetic inheritance in the mouse. Genesis 35: Jaenisch R. 2004. Dnmt1 expression in pre- and postimplantation 88–93. embryogenesis and the maintenance of IAP silencing. Mol Cell Biol La Salle S, Mertineit C, Taketo T,Moens PB, Bestor TH, Trasler JM. 2004. 24: 1640–1648. Windows for sex-specific methylation marked by DNA methyltrans- Grandjean V, Gounon P, Wagner N, Martin L, Wagner KD, Bernex F, ferase expression profiles in mouse germ cells. Dev Biol 268: 403–415. Cuzin F, Rassoulzadegan M. 2009. The miR-124-Sox9 paramutation: Lemmers RJ, Tawil R, Petek LM, Balog J, Block GJ, Santen GW, Amell RNA-mediated epigenetic control of embryonic and adult growth. AM, van der Vliet PJ, Almomani R, Straasheijm KR, et al. 2012. Di- Development 136: 3647–3655. genic inheritance of an SMCHD1 mutation and an FSHD-permissive Gregg C, Zhang J, Weissbourd B, Luo S, Schroth GP, Haig D, Dulac C. D4Z4 allele causes facioscapulohumeral muscular dystrophy type 2. 2010. High-resolution analysis of parent-of-origin allelic expression in Nat Genet 44: 1370–1374. the mouse brain. Science 329: 643–648. Leong HS, Chen K, Hu Y, Lee S, Corbin J, Pakusch M, Murphy JM, ∗ Grossniklaus U, Paros R. 2014. Transcriptional silencing by Polycomb Majewski IJ, Smyth GK, Alexander WS, et al. 2013. Epigenetic regu- group proteins. Cold Spring Harb Perspect Biol doi: 10.1101/csh lator Smchd1 functions as a tumor suppressor. Cancer Res 73: 1591– perspect.a019331. 1599. ∗ Grunstein M, Gasser SM. 2013. Epigenetics in Saccharomyces cerevisiae. Lumey LH. 1992. Decreased birthweights in infants after maternal in Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a017491. utero exposure to the Dutch famine of 1944–1945. Paediatr Perinat Henikoff S. 1990. Position-effect variegation after 60 years. Trends Genet Epidemiol 6: 240–253. 6: 422–426. Maegawa S, Hinkal G, Kim HS, Shen L, Zhang L, Zhang J, Zhang N, Liang Henikoff S. 1998. Conspiracy of silence among repeated transgenes. Bio- S, Donehower LA, Issa JP. 2010. Widespread and tissue specific age- essays 20: 532–535. related DNA methylation changes in mice. Genome Res 20: 332–340. Hitotsumachi S, Carpenter DA, Russell WL. 1985. Dose-repetition in- Majewski IJ, Blewitt ME, de Graaf CA, McManus EJ, Bahlo M, Hilton AA, creases the mutagenic effectiveness of N-ethyl-N-nitrosourea in Hyland CD, Smyth GK, Corbin JE, Metcalf D, et al. 2008. Polycomb mouse spermatogonia. Proc Natl Acad Sci 82: 6619–6621. repressive complex 2 (PRC2) restricts hematopoietic stem cell activity. ∗ Hochedlinger K, Jaenisch R. 2014. Induced pluvipotency and epige- PLoS Biol 6: e93. netic reprogramming. Cold Spring Harb Perspect Biol doi: 10.1101/ Martin DI, Whitelaw E. 1996. The vagaries of variegating transgenes. cshperspect.a019448. Bioessays 18: 919–923. Hummel KP. 1959. Developmental anomalies in mice resulting from Michaud EJ, van Vugt MJ, Bultman SJ, Sweet HO, Davisson MT,Woychik action of the gene, disorganization, a semi-dominant lethal. Pediatrics RP. 1994. Differential expression of a new dominant agouti allele 23: 212–221. (Aiapy) is correlated with methylation state and is influenced by pa- Jabbi M, Kippenhan JS, Kohn P, Marenco S, Mervis CB, Morris CA, rental lineage. Genes Dev 8: 1463–1472. Meyer-Lindenberg A, Berman KF.2012. The Williams syndrome chro- Mill J, Dempster E, Caspi A, Williams B, Moffitt T,Craig I. 2006. Evidence mosome 7q11.23 hemideletion confers hypersocial, anxious person- for monozygotic twin (MZ) discordance in methylation level at two ality coupled with altered insula structure and function. Proc Natl CpG sites in the promoter region of the catechol-O-methyltransferase Acad Sci 109: E860–E866. (COMT) gene. Am J Med Genet B Neuropsychiatr Genet 141B: 421– Jun JE, Wilson LE, Vinuesa CG, Lesage S, Blery M, Miosge LA, Cook MC, 425. Kucharska EM, Hara H, Penninger JM, et al. 2003. Identifying the Mintz B. 1970. Clonal expression in allophenic mice. Symp Int Soc Cell MAGUK protein Carma-1 as a central regulator of humoral immune Biol 9: 15. responses and atopy by genome-wide mouse mutagenesis. Immunity Morgan HD. 1999. “Epigenetic inheritance at the murine agouti locus.” 18: 751–762. PhD thesis, University of Sydney, Australia.

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 25 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press M. Blewitt and E. Whitelaw

Morgan HD, Sutherland HG, Martin DI, Whitelaw E. 1999. Epigenetic Reed SC. 1937. The inheritance and expression of fused, a new mutation inheritance at the agouti locus in the mouse. Nat Genet 23: 314–318. in the house mouse. Genetics 22: 1–13. Morris SA, TeoRT,Li H, Robson P,Glover DM, Zernicka-Goetz M. 2010. Reik W, Collick A, Norris ML, Barton SC, Surani MA. 1987. Genomic Origin and formation of the first two distinct cell types of the inner cell imprinting determines methylation of parental alleles in transgenic mass in the mouse embryo. Proc Natl Acad Sci 107: 6364–6369. mice. Nature 328: 248–251. Noordermeer D, de Wit E, Klous P,van de Werken H, Simonis M, Lopez- Rinchik EM. 1991. Chemical mutagenesis and fine-structure functional Jones M, Eussen B, de Klein A, Singer RH, de Laat W.2011. Variegated analysis of the mouse genome. Trends Genet 7: 15–21. gene expression caused by cell-specific long-range DNA interactions. Rinchik EM, Carpenter DA, Selby PB. 1990. A strategy for fine-structure Nat Cell Biol 13: 944–951. functional analysis of a 6- to 11-centimorgan region of mouse chro- Oates NA, van Vliet J, Duffy DL, Kroes HY, Martin NG, Boomsma DI, mosome 7 by high-efficiency mutagenesis. Proc Natl Acad Sci 87: Campbell M, Coulthard MG, Whitelaw E, Chong S. 2006. Increased 896–900. DNA methylation at the AXIN1 gene in a monozygotic twin from a Robertson G, Garrick D, Wu W, Kearns M, Martin D, Whitelaw E. 1995. pair discordant for a caudal duplication anomaly. Am J Hum Genet 79: Position-dependent variegation of globin transgene expression in 155–162. mice. Proc Natl Acad Sci 92: 5371–5375. Pardo-Manuel de Villena F, de la Casa-Esperon E, Sapienza C. 2000. Robertson G, Garrick D, Wilson M, Martin DI, Whitelaw E. 1996. Age- Natural selection and the function of genome imprinting: Beyond dependent silencing of globin transgenes in the mouse. Nucleic Acids the silenced minority. Trends Genet 16: 573–579. Res 24: 1465–1471. Pedram M, Sprung CN, Gao Q, Lo AW, Reynolds GE, Murnane JP.2006. Sado T, Fenner MH, Tan SS, Tam P, Shioda T, Li E. 2000. X inactivation Telomere position effect and silencing of transgenes near telomeres in in the mouse embryo deficient for Dnmt1: Distinct effect of hypo- the mouse. Mol Cell Biol 26: 1865–1878. methylation on imprinted and random X inactivation. Dev Biol 225: Pembrey ME, Bygren LO, Kaati G, Edvinsson S, Northstone K, Sjostrom 294–303. M, Golding J. 2006. Sex-specific, male-line transgenerational respons- Sapienza C, Peterson AC, Rossant J, Balling R. 1987. Degree of methyla- es in humans. Eur J Hum Genet 14: 159–166. tion of transgenes is dependent on gamete of origin. Nature 328: Percec I, Plenge RM, Nadeau JH, Bartolomei MS, Willard HF. 2002. 251–254. Autosomal dominant mutations affecting X inactivation choice in Schotta G, Ebert A, Dorn R, Reuter G. 2003. Position-effect variegation the mouse. Science 296: 1136–1139. and the genetic dissection of chromatin regulation in Drosophila. Semin Cell Dev Biol 14: 67–75. Percec I, Thorvaldsen JL, Plenge RM, Krapp CJ, Nadeau JH, Willard HF, Bartolomei MS. 2003. An N-ethyl-N-nitrosourea mutagenesis screen Schumacher A, Faust C, Magnuson T. 1996. Positional cloning of a global regulator of anterior–posterior patterning in mice. Nature 384: for epigenetic mutations in the mouse. Genetics 164: 1481–1494. 648. Perry WL, Copeland NG, Jenkins NA. 1994. The molecular basis for Shah NM, Groves AK, Anderson DJ. 1996. Alternative neural crest cell dominant yellow agouti coat color mutations. Bioessays 16: 705–707. fates are instructively promoted by TGFb superfamily members. Cell ∗ Pikaard C, Mittelsten Scheid O. 2014. Epigenetic regulation in plants. 85: 331–343. Cold Spring Harb Perspect Biol doi: 10.1101/cshperspect.a019315. Sutherland HG, Martin DI, Whitelaw E. 1997. A globin enhancer acts by Plenge RM, Percec I, Nadeau JH, Willard HF. 2000. Expression-based increasing the proportion of erythrocytes expressing a linked trans- assay of an X-linked gene to examine effects of the X-controlling gene. Mol Cell Biol 17: 1607–1614. element (Xce) locus. Mamm Genome 11: 405–408. Sutherland HG, Kearns M, Morgan HD, Headley AP, Morris C, Martin Porter S, Larue L, Mintz B. 1991. Mosaicism of tyrosinase-locus tran- DI, Whitelaw E. 2000. Reactivation of heritably silenced gene expres- scription and chromatin structure in dark vs. light melanocyte clones sion in mice. Mamm Genome 11: 347–355. of homozygous chinchilla-mottled mice. Dev Genet 12: 393–402. TeschendorffAE, Menon U, Gentry-Maharaj A, Ramus SJ, Weisenberger Preis JI, Downes M, Oates NA, Rasko JE, Whitelaw E. 2003. Sensitive flow DJ, Shen H, Campan M, Noushmehr H, Bell CG, Maxwell AP, et al. cytometric analysis reveals a novel type of parent-of-origin effect in the 2010. Age-dependent DNA methylation of genes that are suppressed mouse genome. Curr Biol 13: 955–959. in stem cells is a hallmark of cancer. Genome Res 20: 440–446. Pujadas E, Feinberg AP.2012. Regulated noise in the epigenetic landscape Vasicek TJ, Zeng L, Guan XJ, Zhang T, Costantini F, Tilghman SM. 1997. of development and disease. Cell 148: 1123–1131. Two dominant mutations in the mouse fused gene are the result of Raj A, Peskin CS, Tranchina D, Vargas DY, Tyagi S. 2006. Stochastic transposon insertions. Genetics 147: 777–786. mRNA synthesis in mammalian cells. PLoS Biol 4: e309. Vickaryous NK. 2005. “An ENU mutagenesis screen to identify modifiers Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E. 2002. Metastable of gene silencing in the mouse.” PhD thesis, University of Sydney, epialleles in mammals. Trends Genet 18: 348–351. Australia. Rakyan VK, Chong S, Champ ME, Cuthbert PC, Morgan HD, Luu KV, Wagner KD, Wagner N, Ghanbarian H, Grandjean V, Gounon P, Cuzin Whitelaw E. 2003. Transgenerational inheritance of epigenetic states at F, Rassoulzadegan M. 2008. RNA induction and inheritance of epige- Fu the murine Axin allele occurs after maternal and paternal transmis- netic cardiac hypertrophy in the mouse. Dev Cell 14: 962–969. sion. Proc Natl Acad Sci 100: 2538–2543. Waterland RA, Jirtle RL. 2003. Transposable elements: Targets for early Rakyan VK, Down TA,Maslau S, Andrew T,YangTP,Beyan H, Whittaker nutritional effects on epigenetic gene regulation. Mol Cell Biol 23: P, McCann OT, Finer S, Valdes AM, et al. 2010. Human aging-associ- 5293–5300. ated DNA hypermethylation occurs preferentially at bivalent chroma- Waterland RA, Dolinoy DC, Lin JR, Smith CA, Shi X, Tahiliani KG. 2006. tin domains. Genome Res 20: 434–439. Maternal methyl supplements increase offspring DNA methylation at Rakyan VK, Down TA, Balding DJ, Beck S. 2011. Epigenome-wide as- Axin fused. Genesis 44: 401–406. sociation studies for common human diseases. Nat Rev Genet 12: Waterland RA, Kellermayer R, Laritsky E, Rayco-Solon P, Harris RA, 529–541. Travisano M, Zhang W, Torskaya MS, Zhang J, Shen L, et al. 2010. Rassoulzadegan M, Grandjean V,Gounon P,Vincent S, Gillot I, Cuzin F. Season of conception in rural Gambia affects DNA methylation at 2006. RNA-mediated non-mendelian inheritance of an epigenetic putative human metastable epialleles. PLoS Genet 6: e1001252. change in the mouse. Nature 441: 469–474. Weichman K, Chaillet JR. 1997. Phenotypic variation in a genetically Ray PF,Conaghan J, Winston RM, Handyside AH. 1995. Increased num- identical population of mice. Mol Cell Biol 17: 5269–5274. ber of cells and metabolic activity in male human preimplantation White RA, Dowler LL, Pasztor LM, Gaston LL, Adkison LR, Angeloni SV, embryos following in vitro fertilization. J Reprod Fertil 104: 165–171. Wilson DB. 1995. Assignment of the transcription factor GATA4 gene

26 Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press The Use of Mouse Models to Study Epigenetics

to human chromosome 8 and mouse chromosome 14: Gata4 is a Wu M, Rinchik EM, Wilkinson E, Johnson DK. 1997. Inherited somatic candidate gene for Ds (disorganization). Genomics 27: 20–26. mosaicism caused by an intracisternal A particle insertion in the Whitelaw NC, Chong S, Morgan DK, Nestor C, Bruxner TJ, Ashe A, mouse tyrosinase gene. Proc Natl Acad Sci 94: 890–894. Lambley E, Meehan R, Whitelaw E. 2010a. Reduced levels of two YamanakaY,Lanner F,Rossant J. 2010. FGF signal-dependent segregation modiﬁers of epigenetic gene silencing, Dnmt3a and Trim28, cause of primitive endoderm and epiblast in the mouse blastocyst. Develop- increased phenotypic noise. Genome Biol 11: R111. ment 137: 715–724. Whitelaw NC, Chong S, Whitelaw E. 2010b. Tuning in to noise: Epige- Yoder JA, Walsh CP, Bestor TH. 1997. Cytosine methylation and the netics and intangible variation. Dev Cell 19: 649–650. ecology of intragenomic parasites. Trends Genet 13: 335–340. Wijchers PJ, YandimC, Panousopoulou E, Ahmad M, Harker N, Saveliev Youngson NA, Vickaryous N, van der Horst A, Epp T, Harten S, A, Burgoyne PS, Festenstein R. 2010. Sexual dimorphism in mamma- Fleming JS, Khanna KK, de Kretser DM, Whitelaw E. 2011. A lian autosomal gene regulation is determined not only by Sry but by missense mutation in the transcription factor Foxo3a causes terato- sex chromosome complement as well. Dev Cell 19: 477–484. mas and oocyte abnormalities in mice. Mamm Genome 22: 235– Williams A, Harker N, Ktistaki E, Veiga-Fernandes H, Roderick K, Tolaini 248. M, Norton T,Williams K, Kioussis D. 2008. Position effect variegation Youngson NA, Epp T, Roberts AR, Daxinger L, Ashe A, Huang E, Lester and imprinting of transgenes in lymphocytes. Nucleic Acids Res 36: KL, Harten SK, Kay GF,Cox T, et al. 2013. No evidence for cumulative 2320–2329. effects in a Dnmt3b hypomorph across multiple generations. Mamm Genome 24: 206–217. Wilson L, Ching YH, Farias M, Hartford SA, Howell G, Shao H, Bucan M, Zeng L, Fagotto F, Zhang T, Hsu W, Vasicek TJ, Perry WL 3rd, Lee Schimenti JC. 2005. Random mutagenesis of proximal mouse chro- JJ, Tilghman SM, Gumbiner BM, Costantini F.1997. The mouse Fused mosome 5 uncovers predominantly embryonic lethal mutations. Ge- locus encodes Axin, an inhibitor of the Wnt signaling pathway that nome Res 15: 1095–1105. regulates embryonic axis formation. Cell 90: 181–192. Wolff GL. 1978. Inﬂuence of maternal phenotype on metabolic dif- Zheng W, Liu K. 2012. Maternal control of mouse preimplantation de- ferentiation of agouti locus mutants in the mouse. Genetics 88: velopment. Results Probl Cell Differ 55: 115–139. 529–539. Zhuma T, Tyrrell R, Sekkali B, Skavdis G, Saveliev A, Tolaini M, Roderick Wolff GL, Kodell RL, Moore SR, Cooney CA. 1998. Maternal epigenetics vy K, Norton T, Smerdon S, Sedgwick S, et al. 1999. Human HMG box and methyl supplements affect agouti gene expression in A /a mice. transcription factor HBP1: A role in hCD2 LCR function. EMBO J 18: FASEB J 12: 949–957. 6396–6406.

Cite this article as Cold Spring Harb Perspect Biol 2013;5:a017939 27 Downloaded from http://cshperspectives.cshlp.org/ on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press

The Use of Mouse Models to Study Epigenetics

Marnie Blewitt and Emma Whitelaw

Cold Spring Harb Perspect Biol 2013; doi: 10.1101/cshperspect.a017939

Subject Collection Epigenetics

Metabolic Signaling to Chromatin Epigenetic Determinants of Cancer Shelley L. Berger and Paolo Sassone-Corsi Stephen B. Baylin and Peter A. Jones Histone and DNA Modifications as Regulators of Maintenance of Epigenetic Information Neuronal Development and Function Geneviève Almouzni and Howard Cedar Stavros Lomvardas and Tom Maniatis Histone Modifications and Cancer A Structural Perspective on Readout of Epigenetic James E. Audia and Robert M. Campbell Histone and DNA Methylation Marks Dinshaw J. Patel Epigenetics and Human Disease The Necessity of Chromatin: A View in Huda Y. Zoghbi and Arthur L. Beaudet Perspective Vincenzo Pirrotta Induced Pluripotency and Epigenetic Germline and Pluripotent Stem Cells Reprogramming Wolf Reik and M. Azim Surani Konrad Hochedlinger and Rudolf Jaenisch Long-Range Chromatin Interactions Comprehensive Catalog of Currently Documented Job Dekker and Tom Misteli Histone Modifications Yingming Zhao and Benjamin A. Garcia RNAi and Heterochromatin Assembly Epigenetic Regulation of Chromatin States in Robert Martienssen and Danesh Moazed Schizosaccharomyces pombe Robin C. Allshire and Karl Ekwall Dosage Compensation in Drosophila Histone Variants and Epigenetics John C. Lucchesi and Mitzi I. Kuroda Steven Henikoff and M. Mitchell Smith

For additional articles in this collection, see http://cshperspectives.cshlp.org/cgi/collection/