Mammals inherit two sets of chromosomes, one from each parent, and therefore possess two copies of each gene. For the majority of these genes, both alleles are expressed or repressed, depending upon the cell type. However, a little less than 1% of mammalian genes are imprinted, which means these are monoallelically expressed in a parent-of-origin (PofO)-specific manner. Since the discovery of genomic imprinting (GI) in the 80s, this field of biology was observed from different angles to better understand the originality of this mode of inheritance (Figure 1).

In brief, (i) many works have tackled the origin and the dynamics of acquisition of this process across the development of mammals from pre-implantation to post-fertilization and beyond (Monk et al., 2019), (ii) other groups have focused on both the conservation and specificity of these mechanisms across phylogeny, developmental stages and tissues (Monk, 2015; Patten et al., 2016; Edwards et al., 2019), (iii) a growing body of research has determined the key role played by imprinted genes in biological functions, physiological processes and diseases in humans (Peters, 2014; Tucci et al., 2019), (iv) thanks to the exceptional progress of knowledge in epigenetics over the last 15 years, significant advances were made on the molecular mechanisms at play through the study of GI as an example of epigenetic regulation. These achievements have benefited from the growing variety, volume and availability of omics data. Current and future applications include sequence-based analyses encompassing the many types of molecules and interactions involved in GI to get a more comprehensive and accurate view of such an epigenetic phenomenon.

In the present review, we offer a quick overview of the first three aspects mentioned above, completed by a more developed part on the molecular mechanisms involved in GI through epigenetic marks, noncoding RNAs and chromatin organization. Building on this, we introduce why and how novel sequencing technologies and multi-omics approaches will help tackle the study of GI genome-wise.

The identification of the first imprinted genes (Barlow et al., 1991; Rachmilewitz et al., 1992) sparked initial efforts towards elucidating the mechanisms of imprint establishment, maintenance and erasure (Morgan et al., 2005; Ferguson-Smith and Bourchis, 2018; Monk et al., 2019). In primordial germ cells (PGC), the genome undergoes extensive DNA demethylation, including the removal of existing previous parent-specific imprints. New imprints are acquired at later stages of gametogenesis, according to the sex of the embryo, with a sex-specific timeline. In sperm, imprint establishment starts before birth and is completed in perinatal period, whereas in the female germline imprints are acquired after birth, during oocyte growth (Lucifero et al., 2004; Kato et al., 2007). Those germline imprints, known as primary imprints, are left on specific regions called Imprinting Control Regions (ICRs), which are the site of key imprinting regulations. DNMT3A and its cofactor DNMT3L are the main genes involved in the de novo methylation activity in both germlines (Bourc’his et al., 2001; Kato et al., 2007). KDM1B, which encodes a lysine demethylase almost exclusively expressed in growing oocytes, is critical for establishing several maternal imprints during oogenesis, as well as non-histone transcriptional regulators, including ZFP57 and NLRP2 among others (Begemann et al., 2018; Ferguson-Smith and Bourchis, 2018). ZFP57 has further post-fertilization role, when imprinted regions of paternal and maternal germline withstand a wave of genome-wide demethylation followed by a wave of de novo methylation. In addition, DNMT1 is crucial to maintain the methylation imprints in the preimplantation embryo (Hirasawa et al., 2008). Other key regulators of the maintenance of GI include in particular DPPA3 (Nakamura et al., 2007), CTCF (Engel et al., 2006) and components of the nucleosome remodeling and histone deacetylation (NuRD) complex, such as MBD3 (Reese et al., 2007) and MTA2 (Ma et al., 2010).

Approximately 200 imprinted genes have been documented to date in humans and mice. In other species, a few dozen loci have been experimentally validated at most, such as in rats and pigs with 14 and 45 imprinted genes identified to date, respectively (http://www.geneimprint.com, last accessed January 2022). Several studies suggested that imprinted genes were less conserved across mammals than initially thought (Monk et al., 2006; Khatib et al., 2007). Genes of the Kcnq1 cluster found to be imprinted in the mouse placenta are not in humans (Monk et al., 2006), while the opposite was shown for L3MBTL (Li et al., 2005). In a more complex way, IGF2R was imprinted in the mouse but exhibited a polymorphic, variable imprinting pattern in humans (Xu et al., 1993). These findings suggest that GI differs between mammals and displays species-specific regulation patterns, raising questions on the conservation of ICRs across species. Interestingly, the placenta is the tissue with the most imprinting discrepancies between the mouse and humans (Monk, 2015). Genome-wide analyses showed most of the imprinted clusters with differentially methylated regions (DMRs) in the human placenta are not differentially methylated in the mouse placenta (Miri et al., 2013; Court et al., 2014), which suggests that widespread differences have occurred during imprinting evolution. Additionally, many imprinted genes exhibit brain-specific functions and expression patterns. A textbook case is UBE3A, which shows a biallelic expression in most tissues but a maternal expression profile within certain neuronal subtypes (Albrecht et al., 1997). In a more complex way, IGF2 is paternally expressed in the subgranular zone of the hippocampus, acting as an autocrine factor, but biallelically expressed in the subventricular area, displaying a paracrine role (Ferrón et al., 2015). Such functionally important mechanisms of transcriptional dosage control highlight the shape-shifting nature of GI across cells and tissues.

The discovery of the crucial roles of imprinted genes came from uniparental mouse embryos and then from human imprinting disorders (Peters, 2014). Experimentally-produced uniparental embryos show lethality due to aberrant GI patterns in several species including cattle, sheep and pig (Lagutina et al., 2004; Zacchini et al., 2011; Sembon et al., 2012), which is in line with pioneering studies showing development arrest due to a lack of embryonic or extraembryonic tissues in the mouse (Barton et al., 1984; McGrath and Solter, 1984). Paternally-expressed IGF2 is a well-studied example of imprinted gene that positively regulates fetal growth (DeChiara et al., 1991; Ferguson-Smith et al., 1991). Oppositely, maternally-expressed GRB10 acts as an essential growth restrictor (Shiura et al., 2009). It has been proposed that many imprinted genes contributing to growth control pathways are coordinately regulated in multiple tissues within an imprinted gene network (Varrault et al., 2006). As suggested through the contribution of GI to growth-related phenotypes, imprinting dysregulation has been identified in a set of 13 so-called imprinting disorders harbouring convergent patterns of molecular alterations and clinical features (Eggermann et al., 2021). Imprinted genes also have a long-known and important role in the development of the mammalian brain and in adult behaviour, which is illustrated by the contributions of PEG1 and PEG3 to maternal behaviour (Ho-Shing and Dulac, 2019; Tucci et al., 2019).

Beside an accurate understanding of the molecular regulation of the different imprinting regions, acquiring a global overview of the imprinted gene network remains crucial to better apprehend their major roles genome-wise. In this context, recent developments in omics (including genomics, transcriptomics, epigenomics and chromatin structure analyses, see Figure 2) should provide more and more comprehensive insights on the role of GI in complex traits in mammals (O’Doherty et al., 2015) and human disorders (Monk et al., 2019).

Review of sequence-based technological developments shows a transition taking place along two transversal axes, from bulk to single-cell approaches and from short to long reads. Such an evolution carries many promises in the context of GI studies, especially as imprinted sites host very diverse elements and are subjected to various regulatory features. It is therefore the right time to characterize in depth imprintomes and understated regulations across loci, stages, cell types and species, which will lead to a better mechanistic understanding of GI. The incorporation of C-technologies as part of multi-omics integrative approaches could in particular reveal imprinted interactomes (Naveh et al., 2021).

Analyses of coding sequences remain an essential gateway to increase our understanding of GI, as the identification of new imprinted genes leads to various further studies. Affordable genome-wide data acquisition benefiting from pedigree-based designs can be implemented across phylogenetic clades, thereby helping to address large sets of questions related to imprinting evolution. These include a better understanding of the early evolutionary history and diversification of GI. A closer look at GI in certain taxa and tissues through integrative omics approaches could for example help clarify the constraints applied to imprinted clusters, the mechanisms that enabled the acquisition of DMRs and the role of transposable elements in the evolution of mammalian development (Bogutz et al., 2019; Hanin and Ferguson-Smith, 2020; Hanna and Kelsey, 2021; Senft and Macfarlan, 2021). Comprehensive genome scans for imprinted genes in species with little or no previous evidence for GI bring important information, since it promotes the understanding of both GI evolution and related phenomena such as methylation reprogramming and allele-specific expression, which regulate key biological processes in vertebrates (Frésard et al., 2014; Zhuo et al., 2017; Skvortsova et al., 2019). A recent epigenome comparison across five placental mammals notably showed striking species-specific features, with distinct GI mechanisms between humans, nonrodents and rodents (Lu et al., 2021).

All this highlights the great interest of studying how GI may influence phenotypes across mammals. Some strategies are emerging to identify the impact of very subtle changes related to GI on intermediate molecular phenotypes (Greenwald et al., 2019; Lindsly et al., 2021). At a higher phenotypic level, we know from familial and association studies that GI contributes to complex phenotypes, including syndromic disorders (Eggermann et al., 2021), cancer (Goovaerts et al., 2018) and several other developmental phenotypes in both humans (Workalemahu et al., 2020) and other mammals (Freking et al., 2002; Van Laere et al., 2003). Investigating multi-scale GI-phenotype relationships could provide insights on unusual patterns of missing heritability, with the potential for many applications. Genomic prediction in domestic animals could for example benefit from explicitly modeling GI for some economically important phenotypes (O’Doherty et al., 2015; Hu et al., 2016; O’Brien and Wolf, 2019). In cancer, evasion of growth suppression is mediated through many imprinted loci (Stampone et al., 2018; Lecerf et al., 2019; Sutton et al., 2019). Studies on experimental models or patient tissues would be helpful to further document the contribution of dysregulated imprinting patterns to cancer evolution (Lozano-Ureña et al., 2021; Taguchi et al., 2021). More generally, imprinted clusters host key genes offering a gateway to larger epigenomic studies. We therefore believe that current developments in sequencing technologies are essential to significant advances in the characterization of such unusual modes of trait transmission.