Clonal Evolution

Genomic Imprinting

6 min read

Genomic Imprinting

Genomic imprinting is an epigenetic phenomenon in which gene expression depends on the parent of origin — the maternal and paternal alleles of a gene are not functionally equivalent. At an imprinted locus, one allele is transcriptionally silent while the other is active, determined by parent-of-origin-specific epigenetic marks (principally DNA methylation at germline differentially methylated regions, or gDMRs) established during gametogenesis and maintained through somatic cell divisions (Falls et al., 1999; Monk et al., 2019).

The result is functional haploidy: despite having two copies of the gene, only one is expressed. This has direct consequences for cancer evolution — a single mutational or epigenetic event affecting the active allele can eliminate gene function, whereas a biallelically expressed gene requires two hits (Falls et al., 1999).

Why imprinting exists: the kinship theory

According to David Haig’s kinship theory (Haig, 2004; Wilkins & Haig, 2003), genomic imprinting evolved because of an evolutionary conflict between maternal and paternal genomes over resource allocation during development:

  • Paternally expressed genes tend to promote growth, extracting maximal maternal resources for the current offspring. The father’s other offspring are likely carried by different females, so paternal alleles have no stake in conserving this particular mother’s resources for future pregnancies.
  • Maternally expressed genes tend to restrain growth, conserving maternal resources for future offspring. The mother is equally related to all her offspring, present and future, so maternal alleles benefit from distributing resources evenly.
flowchart TD
    subgraph Maternal["Maternal Genome"]
        M1[H19 — growth restraint]
        M2[CDKN1C — cell cycle inhibitor]
        M3[IGF2R — degrades IGF2]
    end
    subgraph Paternal["Paternal Genome"]
        P1[IGF2 — growth enhancer]
        P2[KCNQ1OT1 — silences maternal alleles]
        P3[PEG1/MEST — mesoderm growth]
    end
    Maternal -->|"limits growth"| Tradeoff[Resource Allocation Tradeoff]
    Paternal -->|"promotes growth"| Tradeoff
    Tradeoff --> Fitness[Organismal Fitness]
    Tradeoff --> Cancer[When disrupted: Cancer]

This evolutionary logic predicts that imprinted genes are disproportionately involved in growth regulation — which is exactly what is observed. The same loci that mediate mother-offspring resource conflict are the loci where loss of imprinting drives tumorigenesis (Falls et al., 1999; Haig, 2004).

Mechanism

The imprinting life cycle involves three phases (Monk et al., 2019):

  1. Establishment in primordial germ cells: gDMRs acquire parent-specific methylation during gametogenesis. DNA methyltransferases DNMT3A and DNMT3B, with the cofactor DNMT3L, establish these marks.
  2. Maintenance through early embryonic reprogramming: After fertilization, global demethylation occurs, but imprinted gDMRs are protected by factors including ZFP57, TRIM28 (KAP1), DPPA3 (STELLA), and the subcortical maternal complex (SCMC). The maintenance methyltransferase DNMT1 (with cofactor UHRF1) preserves methylation at imprinted loci through DNA replication.
  3. Erasure in primordial germ cells: Imprints are erased to reset parent-of-origin marks for the next generation.
sequenceDiagram
    participant PGC as Primordial Germ Cell
    participant Gamete as Gamete
    participant Zygote as Zygote
    participant Soma as Somatic Cell
    PGC->>PGC: Erasure of old imprints
    PGC->>Gamete: Establishment (DNMT3A/3B/3L)
    Gamete->>Zygote: Fertilization
    Zygote->>Zygote: Global demethylation
    note right of Zygote: Imprinted gDMRs protected by ZFP57/TRIM28/DPPA3
    Zygote->>Soma: Maintenance (DNMT1/UHRF1)
    Soma->>Soma: Monoallelic expression maintained through cell divisions

Imprinted genes relevant to cancer

GeneExpressed AlleleFunctionCancer Relevance
IGF2PaternalFetal growth factorLOI → biallelic expression in >20 tumor types; 70% of Wilms’ tumors
H19MaternalNoncoding RNA, growth restraintSilenced in many cancers; context-dependent tumor suppressor/oncogene
CDKN1C (p57^KIP2^)MaternalCyclin-dependent kinase inhibitorMutated in BWS; epigenetic silencing in tumors
IGF2RMaternal (polymorphic in humans)Degrades IGF2; activates TGF-βMutated in 60% of HCC, 30% of breast cancers
DLK1PaternalInhibits adipocyte differentiationInvolved in growth regulation
GNAS (Gs-α)Maternal (tissue-specific)G-protein signalingTissue-specific imprinting creates tissue-specific vulnerabilities

Relevance to clonal evolution

Genomic imprinting alters the parameters of somatic evolution in specific, predictable ways:

  1. Accelerated driver acquisition. Imprinted tumor suppressor genes require only one hit (mutation or silencing of the single active allele) instead of two. The effective driver mutation rate is higher at imprinted loci than at biallelically expressed loci — a factor that should be incorporated into branching-process-model parameterization.

  2. Epigenetic drivers. Loss of imprinting at growth-promoting loci (e.g., IGF2 LOI) increases gene dosage without any DNA sequence change. This is a faster-than-mutation path to a fitness advantage, operating at epigenetic rather than genetic timescales.

  3. Coordinated multi-gene effects. The clustering of imprinted genes under shared imprinting centre control means a single epigenetic event (IC methylation change, UPD, mitotic recombination) can simultaneously dysregulate multiple oncogenes and tumor suppressors — a compound driver event.

  4. Tissue-specific vulnerability. Tissue-specific imprinting (e.g., GNAS is biallelic in most tissues but monoallelic maternal in renal proximal tubules, thyroid, and pituitary) means vulnerability to single-hit inactivation is also tissue-specific, contributing to tissue tropism of certain cancers.

  5. Inherited epigenetic risk. Polymorphic imprinting — where imprinting status varies between individuals (IGF2R, WT1) — means some individuals are constitutionally more susceptible to losing tumor suppressor function at these loci, analogous to a germline heterozygous mutation but arising epigenetically (Falls et al., 1999).

  6. Asexual evolution and the exposure of recessive mutations. Cancer cells reproduce asexually — no recombination, no segregation, no independent assortment (McGranahan & Swanton, 2017). In a sexual population, a recessive deleterious allele can hide behind a functional copy. In an asexual population at an imprinted locus, there is no hiding. The single active allele is exposed. A mutation that silences it has an immediate fitness consequence. The effective driver mutation rate at imprinted tumor suppressor loci is higher than at biallelically expressed loci not because the mutation rate differs, but because selection sees the mutation immediately. This parameterization concern feeds into the branching-process-model: the Bozic-Nowak framework’s δ = d/b ratio assumes diploid genetics with two-hit kinetics at tumor suppressor loci. Imprinted loci violate that assumption — they operate under single-hit kinetics.

  7. The tetraploid-imprinting intersection. whole-genome-duplication creates extra gene copies that buffer against deleterious mutations. A chromosomal loss that kills a diploid cell becomes survivable in a tetraploid one. Imprinting creates the opposite condition — functional haploidy that makes a single loss consequential. These are opposing forces on the same parameter: how many mutational or epigenetic events does it take to alter gene function? In a tetraploid cell at an imprinted locus, the interaction depends on which alleles are active. If the imprinted locus retains its parent-of-origin silencing after WGD, the cell may still have only one functional copy despite carrying four genomic copies. The interpretative context — which Haig (2004) frames as the genome being read, not executed — determines what a mutation at that locus means. A deletion that is silent in one ploidy-imprinting context becomes a driver in another. This is context-dependent interpretation of the genetic text applied to genome architecture.

Graph View

Backlinks