Imprinting and Parent-of-Origin Expression: Allele-Specific Regulation

Most autosomal genes are expressed from both alleles, so averaging them is harmless. A few hundred genes are imprinted: one allele is silenced depending on which parent it came from, and the silencing is established by DNA methylation at imprinting control regions.

That changes experimental logic in ways that are easy to overlook. A 50% reduction in expression at an imprinted gene may be the maximum possible effect of a heterozygous perturbation. A methylation editor aimed at an imprinted promoter may be editing an allele that is already silent. And a "biallelic" expression pattern at a locus that should be monoallelic is itself a disease-relevant observation — loss of imprinting is a documented oncogenic mechanism.

This article covers the mechanism, the readouts, and the design implications. Bulk methylation measurement and validation controls are in our pillar; this piece is the allele-resolution companion.

Imprinting is established in the germline and maintained somatically, and the operational unit is the differentially methylated region (DMR):

• • Germline DMRs acquire methylation in one parental germline and are maintained thereafter. These are the primary imprints.
• • Somatic DMRs are established after fertilization, often as a consequence of a germline DMR acting on a nearby promoter or an insulator.
• • Insulator-based model. At the H19/IGF2 locus, methylation of the imprinting control region on the paternal allele blocks binding of the insulator protein CTCF, permitting an enhancer to contact IGF2. On the maternal allele, unmethylated CTCF sites form an insulator, and the enhancer instead contacts H19. The result is reciprocal expression from the two alleles.
• • Non-coding RNA model. At several loci, an imprinted long non-coding RNA is expressed from one allele and represses neighbouring genes in cis.

Two implications follow. First, imprinting is frequently regional: a single DMR controls several genes, so perturbing it can change multiple transcripts at once. Second, the readout must be allele-specific, because total expression conflates two alleles with different regulatory states.

The practical minimum for an imprinting study is a heterozygous SNP in the transcript plus a phased methylation measurement at the DMR. Without phasing, methylation data cannot be assigned to parental alleles, and the central claim — that one allele is methylated and silent — remains unproven.

Four rules that follow directly from monoallelic expression:

1. A heterozygous edit may be functionally null or functionally complete. At an imprinted locus, editing the silent allele changes nothing; editing the expressed allele changes everything. Genotype the edit and determine which allele was modified before interpreting phenotype.
2. Choose the cell background deliberately. Many commonly used lines have loss of imprinting or uniparental patterns at specific loci. Characterize the locus in your line before designing guides.
3. Expect regional effects. An edit at a germline DMR can alter several genes in the domain. Measure the domain, not one gene.
4. Distinguish methylation change from expression change. At an imprinted locus, demethylating the silent allele should permit expression; if methylation falls without expression, the locus is repressed by an additional mechanism.

Point two is the most common practical surprise: a line chosen for convenience may not be imprinted at the locus of interest at all.

Loss of imprinting — biallelic expression where monoallelic is expected — is observed in several tumour types and in developmental disorders, and it is a legitimate experimental endpoint.

How to demonstrate it credibly:

• • Allele-specific evidence, not total expression doubling. A twofold increase can arise from amplification, copy-number gain, or a transcriptional change; only allele-level data show both alleles are active.
• • Methylation at the DMR in the same sample, showing the expected pattern is absent.
• • A matched normal or isogenic control with the expected monoallelic pattern.
• • Copy-number correction, since duplication of the expressed allele mimics loss of imprinting in total-expression data.

The fourth point is where many claims fail: a duplication of the active allele produces exactly the expression signature of biallelic expression.

Practical notes for anyone building models at imprinted loci:

• • iPSC derivation can perturb imprinting. Culture conditions and reprogramming affect methylation at imprinted DMRs; characterize the loci you depend on, and prefer lines with documented imprinting status.
• • Differentiation changes the readout. Some imprinted genes are expressed only in specific lineages, so a null result in the wrong lineage means nothing.
• • Parental origin may be unknown. When it is, report expression patterns as allele-specific rather than parent-specific, unless parental origin was established.
• • Chromosome 15 and 11p15 regions are the best-characterized, and the ones where reference data are most reliable.

EDITGENE builds allele-specific models: isogenic panels with phased heterozygous SNPs, DMR methylation maps, monoallelic expression confirmed by amplicon sequencing, and epigenome-editing lines that resolve which allele was modified.

Tell us the locus and the lineage. Contact EDITGENE to request a project consultation or quote.

Imprinting is locus-, tissue- and stage-specific; the number of imprinted genes and their tissue patterns continue to be refined, and observations in one cell type should not be generalized.

Allele-specific claims require phased data. Total expression changes at imprinted loci have multiple explanations, including copy-number change, and should not be interpreted as imprinting effects without allele-level evidence.

• • Ferguson-Smith AC. Genomic imprinting: the emergence of an epigenetic paradigm. Nature Reviews Genetics. 2011;12:565–575.
• • Barlow DP, Bartolomei MS. Genomic imprinting in mammals. Cold Spring Harbor Perspectives in Biology. 2014;6:a018382.
• • Hark AT, et al. CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature. 2000;405:486–489.
• • Jirtle RL, Skinner MK. Environmental epigenomics and disease susceptibility. Nature Reviews Genetics. 2007;8:253–262.
• • Bar S, et al. Epigenetic aberrations in human pluripotent stem cells. EMBO Journal. 2019;38:e101033.