Generation of the Sotos syndrome deletion in mice
- First Online:
- Cite this article as:
- Migdalska, A.M., van der Weyden, L., Ismail, O. et al. Mamm Genome (2012) 23: 749. doi:10.1007/s00335-012-9416-0
Haploinsufficiency of the human 5q35 region spanning the NSD1 gene results in a rare genomic disorder known as Sotos syndrome (Sotos), with patients displaying a variety of clinical features, including pre- and postnatal overgrowth, intellectual disability, and urinary/renal abnormalities. We used chromosome engineering to generate a segmental monosomy, i.e., mice carrying a heterozygous 1.5-Mb deletion of 36 genes on mouse chromosome 13 (4732471D19Rik-B4galt7), syntenic with 5q35.2–q35.3 in humans (Df(13)Ms2Dja+/− mice). Surprisingly Df(13)Ms2Dja+/− mice were significantly smaller for their gestational age and also showed decreased postnatal growth, in contrast to Sotos patients. Df(13)Ms2Dja+/− mice did, however, display deficits in long-term memory retention and dilation of the pelvicalyceal system, which in part may model the learning difficulties and renal abnormalities observed in Sotos patients. Thus, haploinsufficiency of genes within the mouse 4732471D19Rik–B4galt7 deletion interval play important roles in growth, memory retention, and the development of the renal pelvicalyceal system.
Sotos syndrome (Sotos; MIM# 117550) is an autosomal dominant, multiple-anomaly syndrome characterized by overgrowth, a distinctive craniofacial appearance, advanced bone age, and variable learning disabilities. However, there is significant clinical heterogeneity in Sotos syndrome, with some affected individuals also showing frequent ear and chest infections, cardiac and urinary/renal defects, seizures, scoliosis, and behavioural problems (Tatton-Brown et al. 1993). The diagnosis of Sotos syndrome relied solely on clinical criteria until haploinsufficiency of the NSD1 gene (encoding a histone methyltransferase implicated in chromatin regulation) was identified as causative (Kurotaki et al. 2002). Subsequent analysis of patients clinically diagnosed with Sotos showed that haploinsufficiency of NSD1 due to intragenic NSD1 mutations, partial NSD1 deletions, or chromosomal microdeletions spanning the 5q35 region encompassing the entire NSD1 gene accounted for more than 90 % of cases (with the prevalence of intragenic NSD1 mutations and 5q35 microdeletions encompassing the NSD1 gene depending greatly on ethnic origin) (Baujat et al. 2004; Tatton-Brown and Rahman 2007; Tatton-Brown et al. 2005a, b). It has also been proposed that GPC3 mutations or 11p15 abnormalities or both may be responsible for some Sotos cases without NSD1 abnormalities (Baujat et al. 2004; Li et al. 2001; Tatton-Brown and Rahman 2007).
The major clinical features of Sotos, including overgrowth, facial abnormalities, and intellectual disabilities, are diagnosed in Sotos patients with intragenic NSD1 mutations and in Sotos individuals carrying 5q35 microdeletions (Tatton-Brown et al. 2005a, b). However, in contrast to Sotos individuals with NSD1 mutations, Sotos patients with 5q35 microdeletions tend to show less pronounced overgrowth but more profound intellectual disability (Kurotaki et al. 2003; Nagai et al. 2003; Saugier-Veber et al. 2007; Tatton-Brown et al. 2005a), and several studies have reported an increased frequency of cardiovascular and urinary/renal abnormalities in 5q35 microdeletion Sotos patients (Kurotaki et al. 2003; Nagai et al. 2003; Saugier-Veber et al. 2007). Thus, it is possible that genes other than the NSD1 gene could be dosage-sensitive and therefore responsible for the extended variability and degree of severity of phenotypes observed in Sotos patients who carry 5q35 microdeletions. Furthermore, selected patients with Sotos can have low laboratory values of the factor XII blood-clotting protein, a phenotype thought to result from the hemizygous deletion of 5q35, unmasking a functional single nucleotide variant (SNV) on the remaining nondeleted allele of the FXII gene within the common Sotos deletion interval (Kurotaki et al. 2005). The latter genetic mechanism has recently also been shown to be responsible for TAR (thrombocytopenia absent radius) syndrome where 1q21.1 deletions appear to unmask functional variants associated with the RBM8a gene (Albers et al. 2012).
The mouse orthologue of the NSD1 gene (Nsd1) is highly conserved (83 % homology at amino acid level) (Kurotaki et al. 2001); however, heterozygous Nsd1 mice do not display any gross phenotypic abnormalities (Rayasam et al. 2003). Similarly, mice carrying knockout mutations of other genes within the most frequently detected human Sotos 5q35 microdeletion interval have failed to display dominant phenotypes, suggesting that Sncb, Unc5a, Fgfr4, Mxd3, Rgs14, Slc34a1, F12, Grk6, Dbn1, and Dok3 are not dosage-sensitive and, therefore, at least individually are not likely to be responsible for the extended variability and degree of severity of clinical features observed in 5q35 microdeletion Sotos patients (Mouse Genome Informatics; http://www.informatics.jax.org). In order to allow investigation of the contribution of the 36 genes at the distal end of human chromosome 5 to the development of clinical features commonly diagnosed in Sotos patients, we used chromosome engineering to generate a new genetic mouse model of Sotos syndrome, Df(13)Ms2Dja+/−, carrying a deletion syntenic to the human chromosome 5q35.2–q35.3 region.
Materials and methods
Gene targeting in ES cells and generation of deletion mice
The 5′ Hprt MICER targeting vector MHPN55m07 (Adams et al. 2004) was linearized with KpnI and electroporated into E14Tg2a embryonic stem (ES) cells (129P2Ola), which were selected in G418 as described previously (Ramirez-Solis et al. 1995). Southern blotting was performed on BstEII-digested ES cell genomic DNA (gDNA) using a probe amplified from E14Tg2a gDNA (5′-GTC TGT TGT TAA AAG CTA AAA CCT TAG A-3′ and 5′-TGA GCT ACA GTT TGG TTC TGG TGG ATA AAC-3′) to identify correctly targeted clones. The 3′ Hprt MICER targeting vector MHPP265c24 (Adams et al. 2004) was linearized with NcoI and electroporated into MHPN55m07-targeted E14Tg2a ES cells, which were then selected in puromycin. Southern blotting was performed on SpeI-digested gDNA using a probe amplified from E14Tg2a gDNA (5′-CAG TAA TAT AGT AGA AGC ATG GTC CAT-3′ and 5′-ATG ATA CTG AAC ACA GAC AAC AGA GGC TGC T-3′). Double-targeted ES cell clones were electroporated with a Cre-expression vector and selected in hypoxanthine, aminopterin, and thymidine (HAT) medium, as described previously (Ramirez-Solis et al. 1995), to identify whether the 5′ Hprt and 3′ Hprt vectors were targeted in cis or in trans. Clones of ES cells that had undergone cis recombination were identified and confirmed by PCR (5′-AAG GGT GTT TAT TCC CCA TGG ACT AAT TAT G-3′ and 5′-CCT TCA TCA CAT CTC GAG CAA GAC GTT CAG-3′; presence of a 1.7-kb band confirmed the cis orientation). The deletion allele was designated Df(13)Ms2Dja. ES cell clones carrying the conditional deletion (prior to electroporation with Cre) were injected into C57BL/6-Tyrc-Brd blastocysts and transmitted through the germline. F1 mice carrying the conditional deletion were bred with CMV-Cre mice (Su et al. 2002) to generate heterozygous Df(13)Ms2Dja mice (Df(13)Ms2Dja+/−). These deficiency mice, Df(13)Ms2Dja+/−, were backcrossed to C57BL/6J for three to four generations and maintained on a mixed C57BL/6J-129P2Ola background. All mice were housed and experimental procedures were carried out in accordance with UK Home Office guidelines.
Comparative genomic hybridization (CGH)
CGH analysis was performed on the Agilent 244A array using tail DNA. The data were processed and analysed in R (R Foundation for Statistical Computing, Vienna, Austria).
Fluorescent in situ hybridization (FISH)
Chromosome spreads of activated splenocytes and ES cells were generated and labelled as described previously (Migdalska et al. 2012; Robertson 1987). Mouse BAC clones (http://bacpac.chori.org/mmouse24.htm) were located inside (RP23-99C7) or outside (RP24-204D5) the deletion region and labelled with DIG-dUTP or biotin-dUTP, respectively.
Tissues were fixed in 10 % neutral buffered formalin for 24 h and embedded in paraffin; then 5-mm sections were cut and stained with haematoxylin and eosin.
Body weight and length analysis
Embryos were weighed at day 15.5 of gestation (E15.5), and mice were weighed at 1 day, or 10, 28, or 52 weeks of age. The body length measurements were taken at 10 and 28 weeks of age.
Dual-energy X-ray absorptiometry (DEXA)
Mice at 10 or 28 weeks of age were analysed on a PIXImus II Densitometer (GE Medical Systems, Buckingshire, UK) as described previously (Migdalska et al. 2012). Data were statistically analysed using the two-tailed Student’s t-test.
Mice at 28 and 52 weeks of age were euthanized and assessed using the MX20 Faxitron system (Faxitron Bioptics, LLC, Lincolnshire, IL). Two lateral X-ray images were acquired: a whole-body image and a lower abdominal image. The latter images were analysed to check for the presence of stones in the kidneys.
Mice were group-housed from weaning age (3 weeks) and behavioural tests performed at 22–28 weeks of age. All mice were prehandled for 1–2 min every day for 4 days prior to testing for habituation. Two tests were performed, specifically a social recognition test and an olfactory/vomeronasal function test, both as previously described (Migdalska et al. 2012).
Generation of Df(13)Ms2Dja+/− mice carrying a 1.5-Mb deletion of the 4732471D19Rik–B4galt7 region
Df(13)Ms2Dja+/− embryos are small for their gestational age
Adult Df(13)Ms2Dja+/− mice are smaller and lighter than their littermates
Df(13)Ms2Dja+/− mice fed on a normal-fat diet were lighter and shorter than their wild-type littermates at 28 weeks (males, Fig. 2c, d; females, data not shown). Detailed analysis of the body composition of these mice revealed a statistically significant decrease in the lean mass, but not fat mass, of Df(13)Ms2Dja+/− mice (males, Fig. 2e, f; females, data not shown).
In addition, to determine whether clinical features of Sotos patients, including macrocephaly, facial abnormalities, hypotonia, seizures, and/or scoliosis, could be observed in adult Df(13)Ms2Dja+/− mice, they were subjected to dysmorphology and X-ray analysis. However, Df(13)Ms2Dja+/− mice were viable, fertile, and did not show any overt skeletal or craniofacial phenotypes (data not shown).
Behavioural phenotyping of adult Df(13)Ms2Dja+/− mice
Histopathological analysis of Df(13)Ms2Dja+/− mice reveals dilation of the renal pelvicalyceal system
We generated and phenotypically characterized the heterozygous deficiency Df(13)Ms2Dja mouse—a genetic model of Sotos syndrome covering a 1.5-Mb region syntenic to the telomeric part of human chromosome 5. This region contains 36 genes, which are conserved between human and mouse (Fig. 1). Df(13)Ms2Dja+/− mice were viable, fertile, and indistinguishable from wild-type littermate controls in terms of head and facial morphology, muscle tone, and curvature of their spine. Thus, Df(13)Ms2Dja+/− mice do not recapitulate the macrocephaly, facial abnormalities, hypotonia, or scoliosis observed in some Sotos patients (Tatton-Brown and Rahman 2007). However, it is difficult to reliably model facial abnormalities in mice (Tobin et al. 2008). Sotos individuals have facial abnormalities that include a high and broad forehead (the head is said to resemble an inverted pear), frontotemporal hair sparsity, malar flushing, down-slanting palpebral fissures, and a pointed chin, dysmorphology that is probably hard to observe in mice.
Deficiency mice fed on a normal-fat diet did, however, show significantly reduced growth at all time points measured (embryonic day 15.5–52 weeks of age; Fig. 2). The presence of reduced postnatal growth in Df(13)Ms2Dja+/− mice contrasts with human data, as Sotos patients with intragenic NSD1 mutations and those with 5q35 microdeletions encompassing the NSD1 gene exhibit overgrowth (Tatton-Brown and Rahman 2007). However, the overgrowth phenotype observed in Sotos patients with 5q35 microdeletions is not very pronounced, even at the early stages of development (Kurotaki et al. 2003; Nagai et al. 2003; Saugier-Veber et al. 2007; Tatton-Brown and Rahman 2007). Collectively, these data suggest divergent roles for Nsd1/NSD1 in regulating growth in mouse and man.
Df(13)Ms2Dja+/− mice showed deficits in long-term memory retention in a socially relevant testing paradigm (Fig. 3). In contrast to other artificial novel object discrimination assays, the mechanism of long-term social recognition in mice is multisensory, requiring the integration of both volatile and nonvolatile olfactory and pheromone cues from conspecifics (Noack et al. 2010). Moreover, the sensitivity of the assay is enhanced by its ethological relevance; it does not rely on prior conditioning and instead exploits an innate olfactory-mediated behavioural response of rodents (Engelmann et al. 2011). Social recognition is hippocampus-dependent (Kogan et al. 2000) and is well established for learning and memory testing (Engelmann et al. 2011; Richter et al. 2005) and for assessing cognitive impairment in mice (Mitsui et al. 2009). Thus, Df(13)Ms2Dja+/− mice may potentially model the intellectual/learning disability observed in Sotos syndrome patients (Tatton-Brown et al. 2005b). Given that the 4732471D19Rik-B4galt7 region spans 36 genes, it is possible that other genes within the interval might, in addition to Nsd1, contribute to the cognitive phenotype observed in Df(13)Ms2Dja+/− mice.
Histopathological examination of tissues from Df(13)Ms2Dja+/- mice revealed no signs of advanced bone age or cardiovascular anomalies, although these abnormalities are observed only in a subset of Sotos patients carrying 5q35 microdeletions (Nagai et al. 2003; Saugier-Veber et al. 2007; Tatton-Brown and Rahman 2007). However, kidney abnormalities were observed in our deficiency animals, specifically dilation of the pelvicalyceal system (Fig. 4). Anomalies of the urinary/renal system, including vesicoureteric reflux, hydronephrosis, and small kidneys, are diagnosed in some (~10 %) Sotos patients carrying 5q35 microdeletions (Kurotaki et al. 2003; Nagai et al. 2003; Saugier-Veber et al. 2007; Tatton-Brown and Rahman 2007). Hydronephrosis is characterized by dilation of the renal pelvis and calyces, so Df(13)Ms2Dja+/− mice seem to recapitulate key features of this anomaly. To date, no mice carrying knockout mutations in genes mapped within the 5q35 interval have been shown to exhibit dilation of the pelvicalyceal system, suggesting that Sncb, Unc5a, Nsd1, Fgfr4, Mxd3, Rgs14, Slc34a1, F12, Grk6, Dbn1, and Dok3 are unlikely, at least individually, to contribute to this phenotype in Df(13)Ms2Dja+/− mice (Mouse Genome Informatics, http://www.informatics.jax.org). However, it is possible that haploinsufficiency of other genes in the interval may contribute with Nsd1 to the renal organ system anomalies observed in Sotos patients and Df(13)Ms2Dja+/− mice. Synergy between genes within copy number variable regions has recently been demonstrated for cardiac development anomalies associated with Down syndrome (Grossman et al. 2011) and postulated for other genomic disorders (Lupski et al. 2011).
We show here that haploinsufficiency of gene(s) in the 4732471D19Rik-B4galt7 region of MMU13, syntenic to human 5q35.2–q35.3, results in pre- and postnatal undergrowth, cognitive impairment, and dilation of the renal pelvicalyceal system in Df(13)Ms2Dja+/− mice. Thus, Df(13)Ms2Dja+/− mice have contributed new insights into the role that genes within the Sotos syndrome deletion interval play in Sotos-associated intellectual disability, growth, and development of the urinary/renal system.
We thank the Research Support Facility of the Sanger Institute for taking care of the animals. DJA was supported by Cancer Research UK and the Wellcome Trust, and MJA was supported by Cancer Research UK.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.