Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The ICF syndrome (OMIM 242860) is a rare genetic disorder with a distinctive chromosomal phenotype. The acronym ICF was coined more than 20 years ago (Maraschio et al. 1988) to describe a newly identified syndrome characterized by immunodeficiency, centromeric instability, and facial anomalies. Cytogenetically ICF is distinguishable because of its variable chromosomal instability and typical structural aberrations, mostly involving the centromeric regions of chromosomes 1 and 16, and to a lesser extent chromosome 9, with the most distinctive feature being the apparent stretching or despiralization of the large blocks of juxtacentromeric—or centromere-adjacent—heterochromatin. Chromosomal anomalies in ICF patients also include homologous and nonhomologous associations of those centromeric regions, multibranched configurations involving one or more of the decondensed chromosomes, nuclear protrusions, micronuclei, and duplications or deletions of whole chromosome arms (Brown et al. 1995; Gimelli et al. 1993; Sawyer et al. 1995a; Tuck-Muller et al. 2000). Those chromosomal abnormalities are almost exclusively encountered in phytohemagglutinin-stimulated lymphocytes. In fibroblasts and Epstein–Barr virus (EBV)-transformed lymphocytes, the cytogenetic manifestations of the syndrome are limited to the occasional elongation of the heterochromatic regions (Maraschio et al. 1989).

The ICF syndrome is often fatal in childhood. The immunodeficiency is typically the cause of early fatalities, and the normal cause of death in ICF patients is infection, usually of the pulmonary or gastrointestinal tract. Although the majority of patients display only humeral immunodeficiency, a considerable number show combined immunodeficiency with additional defective cellular immunity, characterized by a reduction or inversion in the CD4+/CD8+ T-cell ratio. Thus far, only one patient lacking the immunodeficiency phenotype has been described. Facial anomalies in ICF patients are mild and include hypertelorism, low-set ears, flat nasal bridge, epicanthal folds, and macroglossia. Psychomotor and mental retardation are also often observed in this syndrome. The spectrum of phenotypic and clinical features resulting in the pronounced heterogeneity of the ICF syndrome has been comprehensively covered in a number of expert reviews (Ehrlich 2003; Ehrlich et al. 2008; Hagleitner et al. 2008).

Since its initial description, the enigmatic nature of this syndrome has continued to fascinate scientists worldwide and from different biomedical spheres with the result that, in spite of its rarity—approximately 50 cases reported so far—and the limited availability of study material, published research on various aspects of ICF has flourished over the past 20 years. Although investigations on the ICF syndrome initially seemed almost exclusively the dominion of clinical genetics, cytogenetics, and immunology, with the growing number of case reports and the complexity of the disorder becoming gradually more evident, cross-referencing and the drawing of parallelisms between specific findings in ICF and related aspects of cancer biology, epigenetics, genomics, and developmental biology have become increasingly ­frequent (Aran et al. 2011; Ehrlich 2009; Ehrlich et al. 2006; Feng and Fan 2009; Martins-Taylor et al. 2012; Toyota and Suzuki 2010; van den Brand et al. 2011). However, notwithstanding the global research efforts and the extensive range of experimental approaches deployed, the full ICF pathogenetic picture remains s­omehow elusive.

Methylation Defects and Genetic Heterogeneity

Methylation at the 5-position of cytosine within CG dinucleotides is an important epigenetic modification, critical for gene regulation and control of chromatin structure in mammalian cells. One of the hallmark molecular features of the ICF syndrome is defective DNA methylation. Most specifically, ICF patients present with significant constitutive hypomethylation at the classical satellite 2 DNA, the major constituent of the juxtacentromeric heterochromatin at 1qh and 16qh, which is normally hypermethylated in somatic cells. Chromosome 9 juxtacentromeric heterochromatin, which mainly consists of the related satellite 3, also appears to be hypomethylated, although to a lesser extent (Jeanpierre et al. 1993). A small number of other genomic regions have been shown to have significant hypomethylation in ICF syndrome, most notably the centromeric alpha satellites (Miniou et al. 1997a), Alu sequences (Miniou et al. 1997b) and the non-satellite repeats D4Z4 and NBL2 (Kondo et al. 2000). Single-copy loci showing heterogeneous hypomethylation comprise the imprinted loci D15S9, D15S63, and H19 (Schuffenhauer et al. 1995), and in female ICF cells a number of genes residing on the inactive X chromosome (Hansen et al. 2000; Bourc’his et al. 1999; Miniou et al. 1994).

Initial investigations on the DNA methylation status in ICF were prompted by the observation that chromosome rearrangements seen in ICF cells were reminiscent of the chromosomal changes induced by the undermethylating agent 5-­azacytidine (examples in Fig. 6.1). Key cytological evidence implicating DNA hypomethylation in the formation of pericentromeric anomalies in ICF was provided by Ji et al. (1997), who confirmed in a pro-B-cell line the preferential formation at a very high frequency of pericentric rearrangements of chromosome 1—identical to the diagnostic chromosomal aberrations in ICF syndrome—on treatment with the DNA methylation inhibitors 5-azadeoxycytidine and 5-azacytidine.

Fig. 6.1
figure 00061

Demethylation by 5-azacytidine induces ICF (immunodeficiency, centromere instability, and facial anomalies syndrome)-like ‘stretching’ of the juxtacentromeric heterochromatin in normal B cells

Full size image

At the molecular level, the methylation pattern of classical satellite DNA in ICF patients, which mimics an embryonic, undermethylated pattern, was initially shown by restriction endonuclease analysis with methyl-sensitive enzymes (Jeanpierre et al. 1993). Subsequently, Hassan et al. (2001) managed to quantify the extent of the abnormal methylation in ICF by using a bisulfite conversion-based method and showing that the cytosine methylation level of the satellite 2 DNA sequences was almost 70 % in normal lymphoblasts and fibroblasts, compared with only 20 % in ICF cells. In line with previous findings, a paper recently published by the Esteller group (Heyn et al. 2012), reporting whole-genome bisulfite DNA sequencing of an ICF patient, showed that despite a global loss of DNA methylation, the shape of genetic features, such as promoters or CpG islands, is conserved, and the most profound changes occur in inactive heterochromatic regions, satellite repeats, and transposons.

The ICF syndrome—the transmission of which is compatible with an autosomal recessive mode of inheritance—was initially linked to chromosome 20q11.2 by homozygosity mapping (Wijmenga et al. 1998). Subsequently, the DNA methyltransferase 3B gene (DNMT3B) was mapped by positional cloning at this chromosomal location and mutations in this gene identified as responsible for the methylation defects observed in ICF (Hansen et al. 1999; Xu et al. 1999). Along with DNMT3A, DNMT3B acts to methylate cytosine residues de novo. The de novo methyltransferases DNMT3A and DNMT3B are essential for normal development, and Dnmt3b −/− homozygous mouse embryos present with multiple developmental defects and fail to develop to term, although development appears normal up to E9.5 (Okano et al. 1999).

Dnmt3b colocalizes with Dnmt3a to the pericentromeric heterochromatic regions in murine embryonic stem cells (Bachman et al. 2001). In mammalian cells, DNMT3B has been found to interact with constitutive centromere protein CENP-C to modulate DNA methylation and the histone code at centromeric regions, and its loss has been shown to lead to elevated chromosome misalignment and segregation defects during mitosis and increased transcription of centromeric repeats (Gopalakrishnan et al. 2009). DNMT3B has also been found to associate in heterochromatic regions with HDAC1 and HDAC2, the ATP-dependent chromatin remodeling enzyme hSNF2H, and two components of the histone H3 lysine 9 methylation machinery, namely, HP1 proteins and Suv39h1 (Geiman et al. 2004). In summary, DNMT3B appears to interact with four chromatin-associated enzymatic activities common to transcriptionally repressed, heterochromatic regions of the genome: DNA methyltransferase, histone deacetylase, ATPase, and histone methylase activities.

Mutations of DNMT3B in ICF syndrome are heterogeneous. Analysis of 14 patients revealed 11 different mutations, including eight different missense mutations, two nonsense mutations, and a splice site mutation (Wijmenga et al. 2000). Missense mutations of DNMT3B in ICF patients occur in or near the catalytic domain of the DNMT3B protein, presumably affecting the normal activity of the enzyme. Nonsense mutations giving rise to a truncated protein always occur as compound heterozygous mutations, highlighting that the DNMT3B protein is essential for life. Most recently, a murine model for ICF syndrome has been engineered by generating Dnmt3b mutations in mice (Ueda et al. 2006). Homozygous mice carrying two missense alleles of Dnmt3b show many ICF-like characteristics, and exhibit low body weight, cranial facial anomalies, and T-cell death by apoptosis. They also show hypomethylation of heterochromatin repeat DNA.

Mutations in the DNMT3B gene account for approximately 60 % of ICF cases. Wijmenga and collaborators were the first to report a number of ICF patients not carrying mutations in the DNMT3B gene (Wijmenga et al. 2000). Similar findings were subsequently published by other research groups (Jiang et al. 2005; Kubota et al. 2004). Jiang et al. (2005) showed that the patients carrying a mutation in the DNMT3B gene had alpha satellite methylation patterns comparable to control samples. In contrast, the patients who did not carry mutations in DNMT3B exhibited hypomethylation of the alpha satellite DNA as well as classical satellite. These findings led to the proposal of the existence of two distinct types of ICF syndrome with different genetic and epigenetic characteristics, namely, a type 1, in which patients display mutations in the DNMT3B gene but have normal alpha satellite methylation, and a type 2, characterized by normal DNMT3B and hypomethylation of alpha satellite DNA. There appears to be no genotype–phenotype correlation between patients with and without DNMT3B mutations (Hagleitner et al. 2008).

In 2011, by means of homozygosity mapping, whole-exome sequencing and Sanger sequencing on 11 ICF type 2 cases, de Greef and collaborators identified in some of the patients mutations in ZBTB24, the zinc-finger- and BTB (bric-a-bric, tramtrack, broad complex)-domain-containing 24, a gene belonging to a large family of transcriptional factors including some members with prominent regulatory roles in hematopoietic development and malignancy (de Greef et al. 2011). ZBTB24 is ubiquitously expressed and its expression appears to be coregulated with DNMT3B during B-cell differentiation. A deletion in the ZBTB24 gene, resulting in a loss of function of the corresponding protein, was also reported in a consanguineous Lebanese ICF type 2 family with three affected brothers (Chouery et al. 2012). In a DNMT3B mutant patient, whole-genome bisulfite DNA sequencing showed that ZBTB24 harbored a hypermethylated DMR (differentially methylated region) in its promoter, suggesting that inactivating hypermethylation of this gene might contribute to the type 1 disease phenotype (Heyn et al. 2012).

Notwithstanding a great deal of progress with the understanding of the genetic and methylation defects that characterize the ICF syndrome at the molecular level, the causal relationships between DNMT3B and/or ZBTB24 mutations, methylation abnormalities, and the range of ICF phenotypes still needs to be fully clarified.

Original Interphase Studies

Once it became clear that the ICF distinctive ‘chromosomal phenotype’ comprising anomalous mitotic configurations involving mainly chromosomes 1 and 16, and to a lesser extent chromosome 9, combined with an abnormal methylation pattern of the classical satellite DNA, was the hallmark feature of the ICF syndrome, there followed several attempts at investigating the ICF genetic and cytogenetic traits in interphase as well as in metaphase. Of course, by that time, the introduction of the fluorescence in situ hybridization (FISH) technique to the diagnostic scenario had allowed the reinvention of classical cytogenetics into molecular cytogenetics, opening up the possibility to visualize specific DNA sequences within the nuclear context and allowing the identification of individual chromosomes and the study chromosomal abnormalities in interphase (Volpi and Bridger 2008).

Maraschio and collaborators published the first study on the interphase b­ehavior of the centromeric heterochromatin of chromosomes 1 and 16 in ICF lymphocytes by means of nonisotopic in situ hybridization, reporting involvement of those chromosomal regions in increased somatic pairing, nuclear protrusions, and micronuclei (Maraschio et al. 1992). Miniou and collaborators carried out the first cytological investigation of DNA methylation defects in ICF using 5-methylcytosine monoclonal antibody to evaluate the methyl content of nuclei and micronuclei (Miniou et al. 1994). They also used alpha and classical satellite probes in single- and dual-color FISH to visualize heterochromatin stretching in interphase. Sawyer and collaborators applied FISH with ‘painting’ probes for chromosomes 1 and 16 to document the progression of centromeric instability from simple heterochromatin decondensation to the subsequent formation of multibranched configurations, and finally nuclear projections and micronuclei involving both chromosomes 1 and 16 (Sawyer et al. 1995b). Similar interphase cytogenetic investigations were carried out by Stacey et al. (1995). Subsequent FISH experiments using the satellite 2-related probe pHuR 195 confirmed that chromosome fusion in the ICF syndrome occurs only at regions of decondensed centromere-adjacent heterochromatin, and that the alpha satellite repeats, the main component of centromeres, always remain outside the regions of multiradiate chromosome fusions (Sumner et al. 1998).

Gisselsson and collaborators showed, in four ICF lymphoblastoid cell lines, an increased colocalization of the hypomethylated 1qh and 16qh sequences in interphase, abnormal looping of pericentromeric DNA sequences at metaphase, formation of bridges at anaphase, chromosome 1 and 16 fragmentation at the telophase–interphase transition, and, in apoptotic cells, micronuclei with overrepresentation of ­chromosome 1 and 16 material. Their results suggested that 1qh–16qh associations in interphase can lead to disturbances of mitotic segregation, resulting in micronucleus formation and sometimes apoptosis (Gisselsson et al. 2005). An association between satellite 2 demethylation induced by 5-azacytidine with missegregation of chromosomes 1 and 16 was shown in human somatic cells by Prada et al. (2012).

Overall, the original interphase cytogenetic studies were mostly concerned with tracking the succession of cytological events leading to the distinctive karyotypic features of the ICF syndrome. Accordingly, the main focus of those research endeavors was on the dynamics of chromosomal instability and references to the possible pathogenetic significance of interphase chromosome organization were rare and strictly speculative. However, those initial ‘nuclear’ observations were fundamental to prepare the experimental and conceptual ground for subsequent studies on the epigenetic impact of nuclear architecture and interphase chromosome organization in the ICF syndrome.

Gene Expression and the Regulation Conundrum

A significant landmark in the sequence of research undertakings aimed at clarifying the connection between methylation defects and gene expression in ICF was a paper published by Ehrlich and collaborators in 2001 (Ehrlich et al. 2001). By means of an oligonucleotide microarray containing approximately 5,600 different human genes, gene expression was analyzed in lymphoblastoid cell lines from six ICF patients carrying different DNMT3B mutations. A total of 45 genes were shown to have significant differences in expression in ICF versus control cells. Of the 32 genes described in detail in the paper, a substantial number appears to play a part in lymphogenesis, signal transduction, and apoptosis. Nine genes in particular are implicated in lymphoid cell differentiation or function after V(D)J recombination and are involved in signal transduction or transcriptional control. Despite the differences in gene expression between ICF and control lymphoblastoid cell lines, no differences in methylation at the promoters or 5′-transcribed regions were observed in a sample of genes analyzed by COBRA analysis, leading the authors to suggest that the hypomethylation of classical satellite DNA sequences in ICF syndrome might result in a decreased ‘heterochromatinization’ that affects the regulation of genes elsewhere in the genome.

Global expression profiling of three patients with DNMT3B mutations was published by Jin et al. (2008). This study generated an impressive list of nearly 800 genes appearing to be up- or downregulated in ICF cells, a large number of which are involved with immune function, development, and neurogenesis. Detailed DNA methylation mapping on a subset of deregulated genes revealed that a fraction of them are typically methylated at low levels in normal cells and lose their methylation in ICF cells. By chromatin immune precipitation the authors also showed that histone modification patterns at affected promoters were dramatically altered, demonstrating for the first time that loss of DNMT3B function in ICF can lead to significant hypomethylation of nonrepetitive regions of the genome and alterations in the histone code. They also observed loss of binding of the SUZ12 component of the PRC2 polycomb repression complex and DNMT3B to derepressed genes, including a number of homeobox genes critical to the immune system and craniofacial development.

Heterochromatic genes have also been reported to undergo epigenetic changes and escape silencing in ICF (Brun et al. 2011); however, the precise mechanism driving their transcription is unclear. Indeed, although DNA hypomethylation was found in all heterochromatic genes and in all patients analyzed, gene expression was restricted to some genes and every patient had his or her own group of activated genes. Surprisingly, heterochromatic genes—either active or inactive—appeared to be associated with histone modifications typical of inactive chromatin. Hypomethylation of subtelomeric regions, associated with aberrant transcription and advanced replication timing of these sequences, has also been described in ICF (Yehezkel et al. 2008).

All together, these expression studies have succeeded in shedding important light on the extent and nature of gene deregulation in ICF resulting from the combined effect of DNA hypomethylation and alterations in the histone code. However, even allowing for prediction of alterations in relevant downstream targets of misexpressed transcriptional regulators, it is obvious that the full ICF epigenetic picture is not yet complete. In summary, on the basis of what has been published so far, and taking also in consideration recent evidence obtained by whole-genome bisulfite DNA sequencing of a DNMT3B mutant patient showing that, despite global loss of genomic DNA methylation, transcriptional active loci and rRNA repeats escape hypomethylation and overall the shape of genetic features such as promoters and CpG islands is conserved (Heyn et al. 2012), it is apparent that the relationship between methylation defects, altered gene expression, and clinical and phenotypic features in ICF needs further investigation.

The Nuclear Architecture Perspective

Over the years a number of hypotheses advocating an epigenetic model for the ICF syndrome have been advanced by different research groups. In particular, possible mechanisms linking altered gene expression to the hypomethylation of pericentromeric heterochromatin have been postulated, generally envisaging inappropriate release or recruitment of regulatory complexes by the hypomethylated satellite DNA, affecting the transacting regulatory properties of the heterochromatin (Hassan et al. 2001; Hansen et al. 1999; Xu et al. 1999; Ehrlich et al. 2001; Bickmore and van der Maarel 2003).

Although so far only a handful of investigations have attempted to test those hypotheses, some interesting findings have already emerged. For example, two publications have independently reported repositioning of the chromosome 1 juxtacentromeric heterochromatin domain to a more interior location within the nuclear volume in ICF lymphocytes (Dupont et al. 2012; Jefferson et al. 2010). More specifically, microscope observations on a cell-by-cell basis, and comparative measurements of the distance between chromosome 1 juxtacentromeric heterochromatin and the nuclear rim carried out in our laboratory, have revealed that the extent of heterochromatin association with the extreme nuclear periphery is reduced in ICF cells (both type 1 and type 2), suggesting its specific repositioning to a more internal location within the nuclear space (Jefferson et al. 2010). Dupont and collaborators have independently confirmed repositioning of the chromosome 1 juxtacentromeric heterochromatin toward the nuclear interior by means of three-dimensional (3D) FISH and confocal microscopy analysis on type 2 patient cells (Dupont et al. 2012). Based on the differential distribution of early and late-replicating chromatin within the nuclear space (Ferreira et al. 1997; Sadoni et al. 1999), these findings on the relocation of the juxtacentromeric heterochromatin away from the extreme nuclear periphery to a more internal position appear to be in line with the previously reported data on the advanced replication of the hypomethylated satellite 2 in ICF (Hassan et al. 2001).

Although its role in actively regulating gene expression remains unproven, the nuclear periphery is generally considered a transcriptionally silent ‘address’ within the nuclear volume, characterized by poor gene density (Croft et al. 1999; Shopland et al. 2006; Tanabe et al. 2002) and a high concentration of nontranscribed sequences (Scheuermann et al. 2004). Interestingly, a case of functional repositioning of the chromosome 1 juxtacentromeric heterochromatin within the nuclear volume had been reported previously by Barki-Celli et al. (2005), who found that the 1q12 heterochromatin domain location within the nuclear volume was different when resting B cells were compared to B cells treated with trichostatin A (TSA), a histone deacetylase inhibitor, with the induced hyperacetylation shown to cause a significant reduction in the percentage of 1q12 signals associated with the nuclear periphery, probably by interfering with the establishment of ‘epigenetic codes’ required for localizing 1q12 to the nuclear membrane, as suggested by the authors.

Long-range heterochromatin–gene associations have gained recognition as a potential novel epigenetic mechanism of gene regulation in mammalian cells. So far, a correlation between gene silencing and localization to transcriptionally repressive heterochromatic compartments has been reported in mouse cycling lymphocytes (Brown et al. 1997, 1999; Grogan et al. 2001), human and mouse erythroid cells (Francastel et al. 1999, 2001; Schubeler et al. 2000), and retinoblastoma cells (Bartova et al. 2002). Replacement of the endogenous β-globin locus control region (LCR) with an ectopic regulatory element, the IgH LCR, was shown to induce looping of the globin locus out of its chromosome territory and bring it closer to the repressive centromeric heterochromatin compartment by Ragoczy et al. (2003). A link between centromeric recruitment and establishment of allelic exclusion at the immunoglobulin heavy-chain gene in mouse B cells was also reported (Roldan et al. 2005). In many instances, uncertainty remains whether nonrandom positioning and association of genes with specific nuclear compartments are the cause or the consequence of gene function, although the general consensus points toward an interactive model in which the functional potential of a locus facilitates its association with a nuclear compartment, which in turn influences the functional properties of the locus (Misteli 2004). Accordingly, recruitment to heterochromatin domains may help to stabilize the inactive state, rather than actively promoting silencing.

In our laboratory, we were interested in understanding if the decondensation of the juxtacentromeric heterochromatin, as observed in metaphase, and general chromosomal instability reported in ICF patients, corresponded to changes in the three-­dimensional properties of the heterochromatin in interphase, our working hypothesis being that disruption in the heterochromatin spatial configuration might interfere with transcriptional silencing and be indirectly responsible for the changes in gene expression accounting for the symptoms of ICF. Accordingly, as well as the heterochromatin intranuclear positioning, we analyzed and compared, in patients and controls, the large-scale organization of the juxtacentromeric heterochromatin of chromosome 1 and the intranuclear positioning of a set of genes—four from chromosome 1 and one as a control from chromosome 6, on which we had concurrently conducted expression and methylation analysis—and their colocalization with the juxtacentromeric heterochromatin of chromosome 1. Examples of our findings are presented in Figs. 6.2 and 6.3. In contrast to the consensual view that decondensation and stretching of the juxtacentromeric heterochromatic blocks—similar to that observed in ICF cells in metaphase—could also be expected in interphase, we found that in ICF nuclei the juxtacentromeric heterochromatin appears more compactly shaped. We also found that for two of the loci analyzed—CNN3 and RGS1—the extent of gene–heterochromatin colocalization was significantly reduced in ICF cells, in which these genes are overexpressed but present no changes in promoter methylation (Jefferson et al. 2010).

Fig. 6.2
figure 00062

The intranuclear positioning and large-scale organization of the juxtacentromeric heterochromatin is altered in ICF B cells. Chromosome 1 heterochromatin is visualized by two-­dimensional (2D) fluorescence in situ hybridization (FISH) with the classical satellite probes D1Z1 (red) in nuclei from two controls (a, c), an ICF type 1 patient (b), and an ICF type 2 patient (d). Chromosome 16 heterochromatin is visualized by cohybridization with D16Z3 (green) (a, b)

Full size image
Fig. 6.3
figure 00063

The extent of interphase gene–heterochromatin colocalization for the two loci RGS1 (green) (a, b) and CNN3 (green) (c, d) is significantly reduced in ICF cells (b, d), in which they are significantly overexpressed, when compared to respective controls (a, c). Chromosome 1 heterochromatin is visualized in red by hybridization with D1Z1 in all four images

Full size image

Our combined findings on the altered large-scale organization and intranuclear positioning of chromosome 1 juxtacentromeric heterochromatin in ICF are particularly significant in the light of the mounting experimental evidence suggesting chromosome band 1q12, which corresponds to the juxtacentromeric heterochromatin of chromosome 1, to be the core of a nuclear domain with functional significance. Earlier investigations showed physical association of this genomic region with the human polycomb group complex, a repressor of the homeotic gene expression (Saurin et al. 1998), and also with the oncogenic transcriptional regulator TLX1/HOX11 in leukemic T cells (Heidari et al. 2006). Most relevantly, in ICF cells the subcellular distribution of HP1 proteins is altered and the 1qh satellite DNA is associated in G2 with a giant HP1-PML nuclear body (Luciani et al. 2005, 2006).

A large-scale effect of DNA hypomethylation on the organization of chromosome ‘territories’ within the nuclear volume in ICF was proposed by Matarazzo et al. (2007) on the basis of their findings on repositioning of misexpressed genes located in the pseudo-autosomal region 2 (PAR2) of the X and Y chromosomes. As DNA hypomethylation appears to cause changes in the interphase organization of the inactive chromosome X territory that extends far beyond the genes which are immediately subject to hypomethylation and transcriptional activation, they suggest that it is plausible that elsewhere in the ICF genome, altered chromosome territory organization encompassing extended regions around hypomethylated sequences may allow for the inappropriate transcriptional activation of relocated genes if the right transcription factor environment is available.

Finally, DNA replication appears to be globally affected in ICF cells. Gartler and collaborators (Hansen et al. 2000) indicated advanced replication time as a major determinant of escape from silencing for genes on the inactive X and Y chromosomes they had identified in ICF syndrome. They also showed that satellite 2 replication in ICF is advanced compared with that in normal cells (Hassan et al. 2001). Yehezkel and collaborators reported that hypomethylation of subtelomeric regions was associated in ICF with advanced telomere replication timing (Yehezkel et al. 2008). More recently, the De Sario group reported that in ICF cells heterochromatic genes replicate earlier in the S-phase, global replication fork speed is higher, and S-phase is shorter, and suggested that these replication defects may result from chromatin changes which modify DNA accessibility to the replication machinery (Lana et al. 2012).

From what has been published so far on different aspects of interphase ­organization in ICF cells, a distinctive ICF ‘nuclear phenotype’ has begun to emerge in which both local and global changes in the chromatin large-scale organization and positioning, together with changes in chromatin proteins distribution, appear mechanistically linked to the deregulation or impairment of fundamental nuclear functions, including chromosome condensation and segregation, DNA replication, and transcription.

Conclusion

The complexity of ICF, in particular the combination of phenotypic variability and genetic heterogeneity that characterizes it, has intrigued geneticists and cell biologists since this syndrome was initially described more than 20 years ago. Indeed, irrespective of its rare occurrence, ICF has been the object of numerous and diverse studies that have yielded interesting insights into its pathogenesis and prompted a substantial amount of speculation on the relationship between methylation defects and the ICF phenotypic spectrum. Up to now only a small number of investigations have attempted to study the ICF syndrome from a nuclear architecture perspective. However, findings that have emerged so far appear to support—cautiously but consistently—a role for large-scale chromatin organization in the interphase nucleus as a structural and functional intermediary between DNA hypomethylation and altered gene expression in ICF, suggesting the existence of additional layers of epigenetic complexity in this syndrome and warranting further investigations.