Image courtesy of D. Jones.
Genomic imprinting — the epigenetic marking of a gene on the basis of parental origin, which results in monoallelic expression — is thought to control the amount of maternal resources that are allocated to the offspring between conception and weaning. Recent work by Journot and colleagues identifies a network of imprinted genes that are involved in the control of embryonic growth, and potentially in tumour development.
Although the misregulation of imprinted loci leads to developmental abnormalities in humans, biological functions are known for only a handful of imprinted genes. Motivated by this gap in our knowledge, Varrault et al. studied a maternally imprinted gene, Zac1, which encodes a zinc-finger transcription factor that induces apoptosis and cell-cycle arrest. Knocking out Zac1 resulted in fetal growth restriction in pups of Zac1-/- and Zac1+/- mice that received their knockout allele from the father (in the mother, Zac1 is inactive anyway). Interestingly, this phenotype is counterintuitive from the point of view of the function of Zac1 as a tumour suppressor gene (its loss should facilitate growth), although it is consistent with a paternally expressed imprinted gene (such genes are thought to promote fetal growth).
But Zac1 knockouts also had ossification defects and other morphological abnormalities, and many pups died postnatally as a result of lung abnormalities. To understand the underlying mechanisms, the authors performed a meta-analysis of 116 mouse microarray data sets to look for genes that are co-regulated with Zac1, assuming that genes that cluster together might share biological function. The set of genes that were most closely linked to Zac1 was enriched in imprinted genes, including H19, insulin-like growth factor 2 (Igf2), insulin-like growth factor 2 receptor (Igf2r) and others. The phenotypes of mutants in many of these genes, including Zac1, indicate that these genes form an imprinted gene network (IGN) that controls embryonic growth and differentiation.
The authors confirmed the role of Zac1 in this network; for example, they showed that loss of Zac1 led to downregulation of Igf2, cyclin-dependent kinase inhibitor 1C (Cdkn1c) and H19. In turn, overexpressing Zac1 in tissue culture resulted in the robust upregulation of some of the same genes. The effects of Zac1 on Igf2 and H19 are direct — using chromatin immunoprecipitation and an electrolytic mobility shift assay (EMSA), Varrault et al. showed that Zac1 binds directly to an E2 enhancer that lies downstream of H19 and is shared with Igf2. These results indicate that downregulation of genes such as H19 and Igf2 might account for the growth impairment of Zac1-deficient mice, albeit by an unknown mechanism.
The authors point out that the network has a scale-free architecture; for example, paternally expressed 3 (Peg3), necdin (Ndn) and guanine nucleotide binding protein, alpha stimulating (Gnas) emerge as nodes (they have high connectivity), whereas others, Igf2 and H19 among them, are connected to fewer genes. But more work is required to explore the potential implications of this observation.
The authors end on an intriguing note: global loss of imprinting has been previously implicated in tumorigenesis and shown to affect expression of many of the genes within the IGN. So, IGN might not only function in embryonic development but also in safeguarding against tumour formation.
