SEARCH

SEARCH BY CITATION

Keywords:

  • Beckwith–Wiedemann syndrome;
  • genomic imprinting;
  • imprinted genes;
  • Silver–Russell syndrome

Abstract

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

Beckwith–Wiedemann syndrome (BWS) and Silver–Russell syndrome (SRS) are two congenital disorders with opposite outcomes on fetal growth, overgrowth and growth restriction, respectively. Although both disorders are heterogeneous, most cases of BWS and SRS are associated with opposite epigenetic or genetic abnormalities on 11p15.5 leading to opposite imbalances in the expression levels of imprinted genes. In this article, we review evidence implicating these genes in the developmental regulation of embryonic growth and placental function in mouse models. The emerging picture suggests that both SRS and BWS can be caused by the simultaneous and opposite deregulation of two groups of imprinted genes on 11p15.5. A detailed description of the phenotypic abnormalities associated with each syndrome must take into consideration the developmental functions of each gene involved.


SRS and BWS: same genes, opposite phenotypes

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

Silver–Russell syndrome (SRS) is a growth disorder characterized by intrauterine and postnatal growth restriction, relative macrocephaly, fifth finger clinodactyly and triangular facies. The syndrome is highly heterogeneous, with association with more than eight chromosomal regions [1]. SRS was first described as an imprinting disorder when it was found that approximately 10% of SRS patients carry a maternal uniparental disomy for chromosome 7, sometimes restricted to part of 7q [2-4]. Although no single gene defects have been described yet, the recent description of a paternal de novo deletion encompassing the imprinted gene MEST/PEG1 implicates this gene in some cases [5]. A growth-promoting role for Mest is further supported by the phenotype of knockout mice [6].

In 2005, a link between SRS and Beckwith–Wiedemann syndrome (BWS) was suggested by the description of SRS patients exhibiting genetic and epigenetic defects implicating a region on 11p15.5 previously known to be altered in most patients with BWS [7, 8]. BWS is an imprinted overgrowth disorder, characterized by somatic overgrowth (>90th percentile in length and weight at birth) and a predisposition to childhood tumors [9]. Other clinical features commonly associated with BWS include abdominal wall defects (omphalocele and umbilical hernia), ear creases or ear pits, and neonatal hypoglycemia. Most cases of BWS and SRS are sporadic although familial cases have also been described. The available molecular data suggest that the same imprinted genes, within an approximately 1 Mb region on 11p15 are implicated in both disorders.

Imprinting defects on 11p15.5

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

The BWS–SRS region contains at least eight imprinted genes clustered on 11p15.5 (Fig. 1a). The gene arrangement and most of the imprinting is conserved on distal mouse chromosome 7. The imprinting effects propagated in this region arise from two cis-acting sequences acquiring DNA methylation marks from a single parent and acting as imprinting centers (ICs).

image

Figure 1. Imprinted gene cluster in the Beckwith–Wiedemann syndrome (BWS)-SRS (Silver–Russell syndrome) region. (a) Gene organization on mouse chromosome 7 showing key genes in the region (arrows) as well as the position of the imprinting center 1 (IC1) and 2 (IC2). Maternally expressed genes are shown in red, paternally expressed genes in blue. Gene order and transcriptional orientations are conserved in the orthologous region on 11p15. (b) Epigenotypes of normal paternal (pat) and maternal (mat) chromosomes. IC1 and IC2 are exclusively marked by DNA methylation imprints on the paternal and maternal alleles, respectively (black ovals). (c–e) Epigenetic abnormalities often detected in BWS and SRS patients with the expected transcriptional imbalances in imprinted genes within the IC1- or IC2-regulated regions.

Download figure to PowerPoint

IC1

The IC1 is located upstream of the maternally expressed gene H19 [10-12]. IC1 acquires a DNA methylation imprint from sperm, a process that is initiated in utero, in the developing gonads of male embryos [13]. IC1 acts as a transcriptional insulator, but only on the unmethylated maternal allele because this function is mediated via the DNA methylation-sensitive binding of the insulator factor CTCF [14, 15]. H19 and Igf2 exhibit very similar expression patterns in the developing mouse embryo and in placenta, owing to shared enhancers located downstream of H19. IC1 prevents expression of Igf2 from the maternal allele, because of the establishment of the CTCF insulator on this chromosome, but allowing Igf2 expression from the paternal allele, due to the DNA methylation mark at IC1. This epigenetic mark also spreads to the H19 promoter on the paternal allele, restricting H19 expression to the maternal chromosome. Although expressed in similar tissues, these two genes are expressed exclusively from a different parental allele, as long as the epigenetic mark at IC1 is maintained only on the paternal chromosome (Fig. 1b). Approximately 5% of BWS patients exhibit a DNA methylation mark on both alleles at IC1, an epimutation referred to as hypermethylation or gain of methylation (GOM) at IC1 (Fig. 1c). Such a defect represents an epigenotype switching, whereby the unmethylated maternal allele acquires a paternal epigenotype. One could also see this as a paternalization of IC1. A significant finding was the observation that up to 60% of SRS cases are associated with the opposite scenario, loss of DNA methylation (LOM) from the paternal IC1 allele (Fig. 1d and Table 1). This maternalization of IC1 leads to the opposite transcriptional effects from those seen in BWS. Whereas BWS is associated with an increase in IGF2 levels and decrease in H19, the opposite is predicted in SRS patients with LOM at IC1. As at least two genes are affected by the observed epigenetic alterations, it is their combined effects that are probably at play in the manifestations of these diseases.

Table 1. Frequencies of epigenetic and genetic defects in BWS and SRSa
BWSSRS
  1. BWS, Beckwith–Wiedemann syndrome; GOM, gain of DNA methylation; IC, imprinting center; LOM, loss of DNA methylation, SRS, Silver–Russell syndrome.

  2. a

    Data from Refs (22), (59), and [9].

pUPD1110–20%mUPD110%
GOM at IC15–10%LOM at IC140–60%
LOM at IC250–60%GOM at IC20%
11p rearrangements1–2%11p rearrangements4%
CDKN1C mutations5–10%mUPD710%
Unknown10–20%Unknown20–40%

IC2

At the other side of the BWS–SRS region, IC2 inherits a DNA methylation mark from the oocyte, established in growing oocytes of postnatal females [16]. IC2 is located within an intron of the Kcnq1 gene and on the unmethylated paternal allele acts as the promoter for the large non-coding RNA (lncRNA) Kcnq1ot1 [previously known as Lit1 [17, 16]]. Therefore as for IC1, IC2 performs a DNA methylation-sensitive function, although here it is the paternal allele that is active. It is the imprinted transcription of Kcnq1ot1, from the unmethylated paternal IC1, which acts as an epigenetic switch here because the function of this lncRNA is to establish a repressive chromatin structure on the paternal allele [18]. Consequently, genes located upstream or downstream of IC2 are silenced on the paternal chromosome and active only or mainly from the maternal allele. These protein-coding genes include Th (tyrosine hydoxylase), Ascl2, Kcnq1, Cdkn1c (cyclin-dependent kinase inhibitor 1c) and Phlda2 in the mouse (Fig. 1b). Characterization of a large telomeric truncation on distal mouse chromosome 7 showed that, other than Kcnq1ot1, the entire region is devoid of other paternally expressed genes required for normal development [19]. In approximately 50% of BWS patients the maternal methylation at IC2 is absent, the LOM corresponding to a paternalization of the IC2-regulated region (Table 1). The expected outcome is biallelic expression of KCNQ1OT1, with the corresponding silencing or downregulation of its target genes, such as CDKN1C and PHLDA2 (Fig. 1e). The only epigenetic defect not represented in Fig. 1 is the GOM at IC2, which could be predicted to be associated with some cases of SRS due to biallelic expression of IC2-regulated genes. To our knowledge, such epimutations have still not been described in SRS patients. However, a number of studies have established a link between SRS and an increase of IC2-regulated genes expression levels, for instance in maternally inherited duplications of the IC2 region [20, 21]. In the mouse, this is modeled by a paternally inherited deletion of the IC2, which causes a maternalization of the IC2 domain and results in growth retardation [18].

Although this discussion focused on epimutations in BWS and SRS, different kinds of rearrangements, deletions and duplications implicating 11p15 have also been identified in these syndromes [reviewed in [22]]. In most cases, however, the effect on up- or downregulation of specific imprinted genes is similar to that expected for epimutations. One of the main conclusions from these studies is that in most cases, BWS and SRS implicate direct transcriptional imbalances in more than one imprinted gene. The notable exception to this conclusion is the finding of maternally inherited CDKN1C mutations in BWS. To understand the pathologies of BWS and SRS, one must therefore know about the specific developmental functions of genes abnormally expressed in each disease. For this, the mouse has provided the best developmental genetic model system.

IC1-regulated genes

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

Igf2

Igf2 codes for a growth factor highly expressed in the fetus and in all components of the placenta exclusively from the paternal allele [23-25]. In mice, excess expression of Igf2 causes overgrowth of both the fetus and placenta, intestinal, cardiac and adrenal defects [26]. Conversely, targeted disruption of paternally transmitted Igf2 causes pups and placentae to be about half the size of wild-type littermates [27, 23].

Igf2 has a placental-specific transcript (Igf2 P0) that is transcribed only in the trophoblast labyrinth of the mouse placenta [28]. Even though P0 contributes to only 10% of the placenta's total Igf2 expression, P0 mutant placentae are growth restricted [29]. In addition, P0 mutant pups are also growth restricted weighing approximately 70% of wild-type littermates at birth. This shows how the placenta, acting as the fetomaternal interface of resource allocation, plays a crucial role in embryonic growth [29]. A similar P0 promoter was also shown to drive imprinted expression of IGF2 in humans, although its expression is not restricted to the placenta [30].

H19

H19 is a maternally expressed ncRNA that is transcribed by RNAPII, polyadenylated, spliced, capped and exported to the cytoplasm [31, 32]. H19 is abundantly expressed in the developing embryo in mesoderm- and endoderm-derived tissues and is an abundant transcript in the placenta. It is completely repressed after birth except in the heart and skeletal muscle of adults [33, 34]. Until recently, the physiological function of H19 remained somewhat enigmatic although several results suggest it has tumor suppressor effects [35, 36]. For instance, the loss of H19 expression is also associated with Wilms' tumors seen in BWS [37, 38]. Transgenic mice overexpressing H19 have no reported abnormal phenotype [39-41]. Despite these results, other studies support a developmental role for the ncRNA. For instance, a targeted deletion of H19 was found to rescue at least partially the embryonic phenotype of parthenogenetic embryos [42], which normally express a double dose of H19 [11].

A new understanding of the developmental role of H19 came from the observation that the first exon of H19 harbors a highly conserved pre-microRNA (miR-675) sequence that gets processed into two mature microRNAs (miR-675-5p and miR 675-3p) [43, 44]. The high conservation of this miRNA across therians suggests it may have an important function and possibly be entirely responsible for H19's functionality [45]. Although the H19 transcript is highly expressed in the embryo, the mature miR-675 is barely detectable there, suggesting that its processing is under developmental regulation [33]. Furthermore, despite its broad and abundant presence, H19 knockout mice present with only a slight overgrowth phenotype (8%) but are otherwise viable and fertile [46]. The overgrowth is thought to be attributed to the overexpression of Igf2 in these embryos, which becomes activated on the maternal allele with an expression level of 25% that of the paternal allele. Recent evidence suggests that miR-675-3p can regulate Igf2 function by targeting its receptor Igf1r [33]. Five other imprinted genes in the embryo, Gnas, Cdkn1c, Dlk1, Rtl1 and Igf2r, four of which are on different chromosomes than H19, were also found to be upregulated in the H19 knockout and brought back to normal levels with the introduction of the H19 transgene. This suggests that H19 may be involved in the control of a network of imprinted genes in the embryo in trans, possibly via the regulated formation of mature miR-675 [40]. This could be related to the finding that roughly 10% of SRS and 20–25% of BWS patients exhibit LOM at imprinted loci on chromosomes other than 11p15.5, in addition to IC1 and/or IC2 defects [47].

As mentioned previously, H19 has shown tumor suppressor properties; however, it was recently shown that miR-675 on its own can suppress cell proliferation in culture [33]. Loss of miR-675 was suggested to be responsible for the 32% placental overgrowth at E18.5 seen in H19 knockout mice and that it may function to restrict placental growth in the last half of gestation. This miR is expressed most highly in the labyrinth portion of the placenta from E11.5 with expression increasing to birth, peaking at E19.5. The mouse data therefore provides a new framework to understand trans effects that the H19 transcript might have on growth regulation. If confirmed in human, the function of miR-675 on the regulation of Igf2 and Igf1r would have profound impact on the understanding of the etiology of BWS and SRS. Of note in this respect is the report of a small duplication of H19 and IC1, but not IGF2, which causes SRS upon maternal transmission [48].

Another ncRNA antisense to H19, named 91H, has been discovered in mouse and human [49]. Transcription of 91H is not imprinted and even though it was shown to affect Igf2 expression in trans [50], its potential role in BWS or SRS has not been explored.

IC2-regulated genes

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

Ascl2

Ascl2 codes for a maternally expressed transcription factor that plays a crucial role in placental development [51, 52]. It is expressed highly in the ectoplacental cone (EPC) and weaker in the extraembryonic ectoderm, which contribute to the spongiotrophoblast and labyrinth layers of the mouse placenta, respectively. After about day 12 of gestation, Ascl2 expression becomes patchy and begins to decline in the mature placenta [53].

Knocking out Ascl2 on the maternal allele has such a dramatic effect on the trophoblast that the embryo does not survive past mid-gestation. The knockout placenta has a defect in the polyploid giant cell layer, which becomes expanded in the mutant, and in the labyrinth layer, which is underdeveloped and does not have the highly vascularized appearance of a wild-type labyrinth. In addition, the Ascl2-null placentae completely lack a spongiotrophoblast. This phenotype suggests that Ascl2 is essential for proper differentiation of the trophoblast [52].

The DNA sequence in-between the IC1- and IC2-regulated domains contains short repeats and retrotransposons along with the gene Th. A deletion spanning this region between the most distal gene controlled by IC1 (Ins2) and the most proximal gene controlled by IC2 (Ascl2) does not disrupt imprinting in the region. However, maternal inheritance of this deletion results in growth retarded pups [54]. It was found that the deletion causes a twofold downregulation of Ascl2 (at E9.5) from the maternal allele. As in the Ascl2-null, the placentae of these Ascl2 hypomorphs display defects in all three placental layers. The placenta have a greatly reduced spongiotrophoblast layer, an expanded giant cell layer, a disorganized and more highly vascularized labyrinth with an increase in trilaminar trophoblast layer cell types, and no glycogen cells [55]. ASCL2 is not imprinted in humans [56-58], so its role in the etiology of BWS or SRS may be questioned, however, the hypomorphic Ascl2 mouse model demonstrates how abnormal dosage of this gene can have serious phenotypic effects on placental function and embryonic growth. Whether the human ortholog also shows haploinsufficiency, possibly in cases of growth restriction, remains to be determined.

Cdkn1c

Mutations in the maternally expressed gene CDKN1C account for approximately 25% of familial cases of BWS and 5–10% of sporadic cases [59]. CDKN1C negatively regulates the cell cycle by targeting cyclin-dependant kinases involved in the G1/S transition, and overexpression of Cdkn1c leads to G1 arrest [60]. Cdkn1c is expressed in all major organs during the development [61]. Cdkn1c is an important regulator of growth and development in mice. In excess, it causes growth retardation from at least E13.5 with no catch up [62]. Furthermore, as Cdkn1c levels increase in these mice, their size decreases demonstrating this gene's dosage sensitivity [62].

The majority of Cdkn1c knockout mice die around the time of birth. They display many characteristics of BWS including kidney dysplasia, placentomegaly and developmental defects in cleft, abdomen and endochondrial bone, but do not display the classic symptom of overgrowth at birth seen in humans [63-65]. However, Cdkn1c knockout mice were recently shown to be overgrown in utero with a slowing growth trajectory that leaves them normal in size at birth. Maternal knockouts are 15% heavier at E15.5 and 8% heavier at E18.5 [63].

Cdkn1c knockout placentae display severe defects in the labyrinth layer. They have areas of collagen deposition and pooled maternal blood that worsen through the course of gestation [63]. The hemorrhaging of maternal blood suggests there is a breakdown in the separation between maternal blood and fetal circulation, mediated by the trilaminar trophoblast layer which is composed of sinusoidal trophoblast giant cells and syncitiotrophoblast layers I and II [66, 63]. Indeed, knockout placentae have depleted sinusoidal trophoblast giant cells suggesting that CDKN1C is responsible for the differentiation of these trophoblast cells [63]. The knockout labyrinth exhibits reduced vascularization and reduced diameter of mutant fetal capillaries [67]. Mature glycogen content is depleted, but immature cells of the glycogen cell lineage are still formed, suggesting that Cdkn1c is needed for the full differentiation of these cells. As stated previously, placentomegaly is apparent in these mutant mice, however, the overgrowth is specifically in the trophoblast cells, suggesting that lack of Cdkn1c interferes with trophoblast differentiation leading to their over-proliferation [67]. The size of the embryo correlates with the degree of impairment of the placenta, again demonstrating the placenta's powerful influence on the growth of the embryo proper [67].

A mouse model that both overexpresses Igf2 and lacks expression of Cdkn1c shows an increase in the severity of several aspects of the BWS phenotype. These mice display dramatic placentomegaly with placentae weighing 190% of wild-type littermates. Their placentae have a highly disorganized labyrinth, fibrin cysts, apoptotic cells and large accumulations of red blood cells. Kidney dysplasia was also notably more pronounced. These mice also show macroglossia, a hallmark of BWS in humans [68].

Cdkn1c is also unique among the IC2-regulated genes in terms of the regulation of its own imprinting by DNA methylation. Although the promoter region of Cdkn1c is not methylated in sperm [69, 70], the paternal allele gradually acquires a DNA methylation mark following implantation [69, 71, 70]. As this mark is not directly inherited from the mature gametes, it constitute a somatic differentially methylated region (DMR) acquired downstream of the IC2/Kcnq1ot1 silencing pathway [72]. Indeed, transcription of Kcnq1ot1 from IC2 is essential for the acquisition of this epigenetic mark at Cdkn1c [69]. Even if it is not gametic in origin, this mark is nevertheless essential to maintain the silent state of the paternal allele, because Cdkn1c becomes biallelically expressed in Dnmt1 null embryos, which fail to maintain DNA methylation marks, including at IC2 [69, 73, 71]. Because the lncRNA Kcnq1ot1 is also biallelically expressed in these mutant embryos [71], the acquisition and maintenance of the somatic DMR at Cdkn1c appears essential for its silencing. These results suggest that Cdkn1c levels can be controlled by epigenetic mechanisms independent of IC2 function. In support for such a model, mice deficient for the SNF2/helicase family of chromatin remodeler HELLS/LSH fail to acquire and/or maintain the somatic DMR at Cdkn1c. This leads to a biallelic expression of Cdkn1c even if the gametic imprint at IC2 is not affected in these mutant animals [74]. A recent study based on the genome-wide analysis of CpG methylation levels in human placentae from normal pregnancies as well as triploidies and complete hydatidiform moles led to the identification of a paternal DNA methylation imprint on the CDKN1C gene [75]. On the basis of the mouse results, this would warrant more detailed analyses of CDKN1C methylation in BWS and SRS cases of unknown etiology.

Phlda2

The maternally expressed gene Phlda2 encodes a cytoplasmic protein with a pleckstrin homology domain [76, 77]. Unlike the potent growth regulators Cdkn1c and Igf2, Phlda2 is expressed at low levels in the embryo. It is, however, expressed highly in the extraembryonic tissue, where it is expressed in the extraembryonic ectoderm and EPC and becomes restricted to the labyrinth by E10.5 [78]. In mouse, normal levels of Phlda2 expression are needed for proper placental development. Conceptuses lacking a maternal copy of Phlda2 have placentae that are about 25% overgrown, detectable by E12.5. Both the labyrinth and spongiotrophoblast of these mutant placentae are expanded laterally, with the spongiotrophoblast also expanded in thickness [78]. The knockout mice have increased glycogen cell content in the spongiotrophoblast but this does not confer a growth advantage to the embryo. Interestingly, loss of Phlda2 partially rescues growth retarded Igf2-null placentae, indicating it can act independently of the Igf2 pathway [78].

Phlda2 overexpression in mice due to loss of imprinting (LOI) causes placental growth retardation [79]. LOI at IC2 by deletion of the Kcnq1ot1 promoter leads to the overexpression of several maternally expressed genes and both placental and fetal growth retardation, however, the overexpression of Phlda2 alone contributes substantially to the placental growth retardation seen in the IC2 LOI phenotype but not to the fetal growth retardation. The dosage-sensitive function of Phlda2 in placental development clearly argues for an important role of this imprinted gene in both BWS and SRS cases implicating IC2 epimutations or rearrangements.

IC functions

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

There are a number of possible explanations for the apparent antagonism between the IC1- and IC2-regulated genes. First, some of the implicated gene products may play functionally antagonistic roles in the regulation of embryonic growth. Such a model has been proposed for the action of IGF2 and CDKN1C in at least some embryonic tissues [68]. Second, perhaps some of the imprinted genes involved can negatively regulate the expression of at least another imprinted gene with opposite growth regulation effect. Such a model is supported by in vitro and in vivo data for the direct or indirect effects of IGF2 levels on the expression of Cdkn1c [80]. With the discovery of miRNAs not only within H19 (the miR-675 discussed above), but also within Igf2 [81], new possibilities for such interactions can now be explored experimentally. In the case of SRS, similar considerations could also be invoked to explain how aberrant expression of different imprinted genes on human chromosomes 7 and 11 can lead to a similar growth restriction phenotype.

Alternatively, epigenetic defects involving one IC could affect expression in the other domain. This would manifest itself as an apparent dominance of one center over the other. Although the issue of expression levels or spatio-temporal regulation of expression patterns has not been studied in details, data from the mouse model clearly support the independent function of each IC with regard to allele-specific expression. Transgenic experiments have shown that both IC1 and IC2 can function independently as ICs when inserted at ectopic sites in the genome [39, 82]. Furthermore targeted deletions of IC1 and IC2 only affect the imprinting of genes under their regulation, with no evidence of cross-talk across centers [73, 18]. In fact imprinting at IC1 is faithfully maintained on a truncated chromosome 7 in which the entire IC2 domain is deleted [19].

Other developmental considerations

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

The study of the IC1 and IC2 imprinted regions in mouse has given us much insight into the genes underlying the BWS and SRS phenotypes; however, there are still a number of questions to be answered in humans. First, the possible role of antisense transcripts in these phenotypes needs to be clarified. While two ncRNAs have been discovered at the H19 locus in mouse that are conserved in humans, miR-675 and 91H, an additional antisense transcript called H19 opposite tumor suppressor was found in human [83]. This RNA has the potential to code for a nucleolar protein not conserved in the mouse. This gene, which could be part of 91H, acts as a tumor suppressor. While, maternal duplication of the entire IGF2-IC1-H19 region has no phenotype, duplications of H19 alone can lead to SRS, possibly due to increased expression of miR-675 or of one more of the antisense transcripts in this region. The H19 antisense transcripts may also be involved in the subset of patients with BWS that develop Wilms tumor. Furthermore, hypomethylation of IC1 is not associated with reduced levels of IGF2 expression levels in cultured fibroblasts [84] or in blood serum of SRS patients [85], further raising the possibility that the primary effect may be on one or more H19 antisense genes or the miR-675.

Second, while, SRS often involves changes in IC1, there are very few examples of an alteration in IC2 alone in SRS. It is not clear whether this is because such mutations do not occur, or because such mutations do not result in a clinical diagnosis of SRS. Furthermore, virtually all cases of paternal uniparental disomy (UPD) 11p15.5 leading to BWS and IC1 methylation changes leading to SRS or BWS are mosaic. It is presumed that non-mosaic paternal UPD is lethal due to the aberrant imprinting of genes involved in early implantation and it may be that non-mosaic loss of maternally expressed genes in the IC2 region does not result in a viable pregnancy. In the mouse, non-mosaic maternal and paternal UPD for the distal chromosome 7 imprinted domain are both embryonic lethal [86]. The paternal UPD for distal 7 embryos die at mid-gestation and exhibit a placental phenotype similar to the one seen in Ascl2-null embryos [52, 87]. Expression of Ascl2 from a transgene is able to rescue this phenotype; however, the rescued embryos still fail to develop to term in most cases. Most of the phenotypic traits of these embryos are consistent with Igf2 overexpression and loss of Cdkn1c and Phlda2 [88].While ASCL2 has not been found to be imprinted in humans so far, it remains possible that it displays lineage restricted imprinting in early trophoblast cells responsible for implantation.

In addition, phenotypic correlations with a specific underlying mechanism is challenging given that a large number of patients with BWS and SRS exhibit LOM at multiple imprinted regions in the genome [47, 89]. Furthermore, in patients with LOI at IC1, the degree to which methylation is lost/or gained is variable from case to case. In other words, there is generally mosaicism and not complete LOM in such cases. In humans, studies are generally limited to blood and it is thus impossible to know the extent of methylation changes in other somatic tissues. Interestingly, one study reported that among the patients identified with multiple LOM defects, four showed LOM at both IC1 and IC2, of which three had SRS and one had BWS [47]. The authors suggested that the relative degree of LOM at the two loci may have determined the resulting phenotype.

Conclusion

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

With the discovery that SRS is often associated with epigenetic and genetic defects mapping to 11p15.5, it became clear that SRS and BWS represent related phenotypes caused by opposite imbalances in the levels of imprinted genes expressed from the IC1 and IC2 domains. Together with the Angelman and Prader–Willi syndromes, this therefore represents the second example of such opposite imprinted disorders [90]. Studies on the orthologous BWS–SRS region in the mouse have shown that several imprinted genes in the region show dosage-sensitive functions with important roles in the regulation of placental function and embryonic growth. Notably, in addition to IGF2 and CDKN1C, further work will be required to establish the roles played by the H19-encoded miR-675 and the placental regulator PHLDA2 in the SRS and BWS syndromes. Further considerations for a molecular description of the phenotypes observed include the possibility of mosaicism as well as the presence of multiple epimutations, affecting different imprinted regions in the genome.

Acknowledgements

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References

Work in the laboratories of L. L. and W. P. R. was supported by operating research grants from the Canadian Institutes of Health Research (CIHR grants MOP-119357 and MOP-106430, respectively). L. L. was supported by a Canada Research Chair. W. P. R. receives salary support from the Child & Family Research Institute.

References

  1. Top of page
  2. Abstract
  3. SRS and BWS: same genes, opposite phenotypes
  4. Imprinting defects on 11p15.5
  5. IC1-regulated genes
  6. IC2-regulated genes
  7. IC functions
  8. Other developmental considerations
  9. Conclusion
  10. Acknowledgements
  11. References