Defining the criteria for identifying constitutional epimutations
Sloane et al. Clinical Epigenetics
Defining the criteria for identifying constitutional epimutations
Mathew A. Sloane 0
Robyn L. Ward
Luke B. Hesson 0
0 Adult Cancer Program, Lowy Cancer Research Centre and Prince of Wales Clinical School, UNSW Australia , Kensington, Sydney, New South Wales 2052 , Australia
In the January 2016 issue of Clinical Epigenetics, Quiñonez-Silva et al. (Clin Epigenetics 8:1, 2016) described a possible constitutional epimutation of the RB1 gene as a cause of hereditary predisposition to retinoblastoma. The term constitutional epimutation describes an epigenetic aberration in normal tissues that predisposes to disease. The data presented by Quiñonez-Silva et al. are interesting, but further analysis is required to demonstrate a constitutional epimutation in this family. Here, we define the criteria and describe the experimental approach necessary to identify an epigenetic aberration as a constitutional epimutation.
Constitutional epimutation; RB1; Retinoblastoma; Methylation
Basic criteria that define constitutional epimutations
An epimutation describes an epigenetic aberration that
results in the transcriptional silencing of a gene that is
normally active or the expression of a gene that is normally
inactive [
1
]. Epimutations that are widely distributed in
normal tissues and predispose to disease are known as
constitutional epimutations [
2
]. Epigenetic silencing may
occur throughout all normal tissues (soma-wide) or show
mosaicism depending on the mechanistic basis of the
epimutation. Constitutional epimutations have been
described in various genes and diseases including Lynch
syndrome (MLH1 and MSH2) [
3, 4
], fragile X syndrome
(FMR1) [5], α-thalassaemia (HBA2) [
6
] and several
imprinting disorders [
1
]. Constitutional epimutations are
confined to one allele and may occur in the absence or
presence of an in-cis genetic variant and termed primary
or secondary epimutations, respectively. Co-segregation of
an epimutation with disease provides very strong evidence
that it plays a role in disease causality; however, this is not
a defining characteristic due to their potential reversal in
the germ line and consequent non-Mendelian pattern of
inheritance [
3
].
Experimental approach for identifying constitutional epimutations
To define an epigenetic aberration as a constitutional
epimutation, the following must be demonstrated.
Firstly, promoter DNA hypermethylation must be
confined to one allele in normal tissues derived from
multiple germ layers. Peripheral blood, hair follicles
and saliva or buccal mucosa are representative
derivatives of the mesoderm, ectoderm and endoderm,
respectively, which are routinely assayed in constitutional
epimutation studies. Secondly, validation of the
presence and level of methylation should be performed
using at least two independent methods, such as
bisulphite pyrosequencing and clonal bisulphite sequencing,
to reliably eliminate confounders including cloning bias
and incomplete bisulphite conversion. Bisulphite
pyrosequencing can quantify the average level of
methylation across both alleles [
7
], whereas clonal bisulphite
sequencing across an informative (heterozygous)
promoter variant can confirm whether methylation is
allele-specific. Thirdly, the methylated allele must show
partial or complete transcriptional silencing.
Allelespecific mRNA expression levels can be accurately
quantified with cDNA pyrosequencing [
8
] using an informative
expressed variant. This usually requires extensive
knowledge of the locus through variant haplotyping in family
members. Finally, there must be co-segregation of the
methylated and transcriptionally silent allele with disease.
The study by Quiñonez-Silva et al. [
9
] has not
definitively demonstrated a constitutional epimutation of the
retinoblastoma 1 (RB1) gene in this family. Firstly, germ
line genetic testing was not performed meaning a genetic
cause of retinoblastoma cannot be ruled out. This would
normally be considered a high priority especially in light
of the strong paternal history of retinoblastoma. On the
other hand, the mother of the index case, and the origin
of the apparent epimutation, had no personal or family
history of retinoblastoma. Secondly, the authors tested
constitutional methylation of the RB1 promoter in only
one germ layer (mesoderm, peripheral blood) and with
only one method (clonal bisulphite sequencing). It is
therefore unclear whether methylation is widely
distributed in normal cells or whether the striking pattern of
methylation is reproducible using an independent method.
Methylation in melanoma cells from the index case may
represent a ‘second somatic hit’ and is not necessarily
evidence of a constitutional epimutation. Thirdly, this case is
complicated by the observation of hypermethylation on
both the maternal variant (c.[-187G; -188G]) allele and the
paternal wild-type (c.[-187T; -188T]) allele possibly related
to the fact that RB1 is imprinte (...truncated)