Introduction

Prader–Willi syndrome (PWS, MIM 176270) is an imprinting disease caused by paternal deletions, maternal uniparental disomy or imprinting anomalies in the 15q11.2q13 region.1 Clinical diagnostic criteria vary with age,2 and consist chiefly of neonatal hypotonia, early-onset obesity, and developmental delay.

A PWS-like phenotype, characterised by hypotonia, obesity, acromicria and variable motor, and cognitive delays,3 has been reported in several conditions, such as maternal uniparental disomy for chromosome 14,4, 5 certain 1p36 deletions,6, 7 2p25 deletions,8 Xq21 duplications,9 Xq23q25 duplications,10 and some cases of fragile X syndrome.11, 12 However, 6q16 deletion is the most common genetic abnormality in patients exhibiting the PWS-like phenotype.

To date >30 patients with 6q deletions, encompassing the q16.2 and/or q16.3 cytogenetic sub-bands, have been reported.3, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36 However, few of them underwent molecular characterisation of their genetic abnormalities, using either chromosomal microarray analysis,3, 13, 15, 16, 17, 18, 22, 23, 25 fluorescence in situ hybridisation (FISH) analysis with bacterial artificial chromosomes (BAC) clones21 or STR analysis.24 Two publications evaluated genotype–phenotype correlations at the 6q16 locus, but included only five and three patients, respectively.13, 17 The first study identified a 4.1-Mb minimal critical region for PWS-like within the 6q16 cytogenetic band.13 Recently, obesity and PWS-like syndrome have been ascribed to loss-of-function variants in the single-minded 1 (SIM1) gene encompassed in 6q16 critical minimal region,37, 38, 39, 40 whereas a role for GRIK2 deletion in behavioural problems has been suggested.13

Here, we describe 15 new patients (including one fetus) with 6q16 deletions, including 6q16.2 and/or 6q16.3 sub-bands, investigated by chromosomal microarray analysis. Genotype–phenotype correlations were assessed. Our results confirm the major role for SIM1 haploinsufficiency in obesity and the PWS-like phenotype.

Subjects and methods

Patients

Seven French centres and one Italian centre recruited one fetus and 14 children or young adults with 6q16 deletions, encompassing the 6q16.2 and/or 6q16.3 sub-bands. Experienced geneticists examined all patients. Informed consent was obtained from all patients and/or parents for a genetic evaluation, an assessment of deletions’ parental origin and publication of clinical pictures. For the fetus, the parents provided their written informed consent to an autopsy.

Fetal examination (patient no. 1)

After termination of pregnancy, an autopsy of the fetus (patient no. 1) was performed according to protocols, including radiographs, photographs, and macroscopic and microscopic examination of all organs.41 Biometrics were compared with previously established reference values.42

Cytogenetic studies

The karyotype of the fetus was determined using in situ cultured amniocytes, following conventional procedures. For the other 14 patients, cultured peripheral lymphocytes were used.

Microarray studies were done in all 15 patients. DNA was extracted using standard procedures from cultured amniocytes (patient no. 1) or peripheral blood lymphocytes (patients no. 2–15). Patients no. 1–12 were investigated using the following oligonucleotide arrays: CGX-12 (Roche NimbelGen, Madison, WI, USA) in patient no. 1, Agilent 44 K (Agilent Technologies, Santa Clara, CA, USA) in patients no. 2–8, 10 and 11, Agilent 60 K in patient no. 9, or Agilent 180 K in patient no. 12. Patients no. 13 and 14 were evaluated using HumanHap 300 and HumanCytoSNP-12, respectively (Illumina, San Diego, CA, USA), and patient no. 15 was evaluated using Genome-Wide Human SNP Array 6.0 (Affymetrix, Santa Clara, CA, USA). Results were analysed according to Human Feb. 2009 (GRCh37/hg19) Assembly. All 15 patients have been submitted for registration in the DECIPHER database (https://decipher.sanger.ac.uk/).

FISH was performed using chromosomal preparations according to standard protocols to confirm the 6q deletions characterised by microarray.43 Probes were prepared from bacterial artificial chromosomes BAC using rolling circle amplification followed by nick translation labelling. The absence of parental deletion was checked in 14 cases, the exception being patient no. 4.

Parental origin study

Microsatellites and SNP array analysis were performed in nine patients (patients no. 2, 3, 5, 7, 8, 11, 13, 14, and 15) to investigate the parental origin of the imbalance. We either selected microsatellites at the common deleted region of the UCSC Genome Browser microsatellite or designed simple repeat tracks and primers using the NCBI Primer-BLAST program (D6S1671, D6S475, D6S2079, D6S20CA, D6S15AAT, D6S21TA, and D6S18GT). After PCR, fragment analysis was performed on an ABI 3730 XL DNA sequencing analyser and processed using GeneMapper 3.7 software (Applied Biosystems, Foster City, CA, USA). For the patient no. 15, parental origin study was performed analysing a total of 16 informative SNPs selected from 1008 SNPs located in the deleted region.

Supplementary Table S1 in the Supplementary Information lists the primers used for each microsatellite.

Results

Clinical and fetopathological data

Patient no. 1

Patient no. 1 was a male fetus at 35 weeks of gestation (WG), who was the product of the first pregnancy of unrelated parents. The mother has unilateral hearing loss and the maternal grandmother has a bilateral hearing loss. A high-risk maternal screening test for Down syndrome prompted karyotype determination on amniotic fluid cells, which showed a 6q14-q16 deletion. Pyelectasis was seen on sonogram at 23 WG. The parents requested termination of pregnancy at 35 WG. Foot length was under the 5th centile and weight was 2140 g (5th centile). The pyelectasis was confirmed. The facial gestalt consisted of a short straight forehead, marked suborbital folds, a broad nasal bridge, prominent philtrum with a thin upper lip, micrognathia, and abnormally overfolded helices with a small horizontal fold along the upper edge (Figure 1a).

Figure 1
figure 1

Fetopathological study of patient no. 1. (a) Facial features: short straight forehead, marked suborbital folds, broad nasal bridge, prominent philtrum, thin upper lip, micrognathia, and abnormally hemmed ears with a small horizontal fold along the upper edge of the helix. (b) Radiographs of the feet: bilateral calcaneal fragmentation and hypermineralisation. (c) Sagittal section through the brain: internal capsule dysmorphism with fusion of anterior caudate nucleus and putamen (black arrows). (d) Cerebral white matter containing ectopic neurons (black arrows). (e) Cerebellar grey matter containing ectopic neurons (black arrows). (f) Sagittal section through the cerebellum showing focal neuronal ectopia.

Full size image

The radiographic skeletal survey showed delayed bone maturation relative to gestational age, absence of ossification of the distal femoral epiphyses, hypoplasia of the sixth cervical vertebral body, sternal dysplasia, bilateral brachymesophalangia of the fifth digits, and bilateral calcaneal fragmentation with increased mineralisation (Figure 1b).

Microscopic examination of the brain evidenced fusion of the anterior caudate nucleus and putamen (Figure 1c), multiple ectopic neurons in the white matter (Figure 1d) and ectopic Purkinje cells in the internal granular layer of the cerebellum (Figure 1e). Two large heterotopias were identified in the white matter of the paravermis (Figure 1f).

Patients no. 2–15

All 14 patients had developmental delay with variable degrees of cognitive deficiency. Table 1 lists the main clinical data and Figure 2 shows photographs of several patients.

Table 1 Clinical description of each of our 15 patients
Full size table
Figure 2
figure 2

Photographs of four study patients. Note the round face, full cheeks, bulbous nose, and a prominent philtrum in patient no. 2; horizontal eyebrows and a prominent philtrum in patient no. 5; round face and full cheeks in patient no. 7; and a triangular face shape in patient no. 14.

Full size image

Cytogenetic and molecular results

Table 2 reports the cytogenetic abnormalities. Microarray analyses showed overlapping 6q deletions, extending from 92 138 719 bp to 108 227 875 bp (hg19). Supplementary Table S2 lists the genes included in the deletions. Except for patients no. 14 and 15, all patients had deletions that included the SIM1 gene. Minimal deletion size across patients ranged from 1.73 to 14 Mb.

Table 2 Cytogenetic characterisation of the 6q16 deletions in our 15 patients
Full size table

Discussion

Our results obtained in the largest reported series of patients with 6q16 deletion, including 6q16.2 and/or 6q16.3 sub-bands, support a strong association between this chromosomal abnormality and a distinct phenotype reminiscent of PWS.

The 6q16.2q16.3 chromosomal region is not polymorphic: the Database of Genomic Variants (http://dgv.tcag.ca/) contains no large copy number variation in this region in healthy individuals, and all reported 6q16 deletions occurred de novo in symptomatic patients. The region contains no low copy repeats or recurrent breakpoints.

An imprinting effect in 6q16 deletions was hypothesised by Faivre et al14 based on the paternal origin of a de novo 6q16 deletion in a patient with PWS-like. The authors speculated that the phenotype might be ascribable to the haploinsufficiency of paternally expressed genes located in the deleted region. Other observations support this hypothesis.18, 19 In our series, only two of nine deletions in patients, for whom parental-origin data were obtained, were located in the maternal chromosome, which is consistent with the ratio reported previously for interstitial deletions at any site.44 In another study, de novo imbalances not mediated by low copy repeats were significantly more often of paternal than of maternal origin.45 Thus, to date, although there is no strong evidence supporting an imprinting mechanism in the 6q16 region, a parent-of-origin effect cannot be excluded, as none of the three maternally derived deletions, which were currently reported, (patient no. 11, 14, and case 4 from Bonaglia et al report)13 was associated with PWS-like features.

Learning disabilities, behavioural disorders, and obesity are common in 6q16 deletions (Figure 3 and Table 3). Our observations narrow the minimal critical region for PWS-like phenotype (obesity, developmental delay with or without hypotonia and/or short extremities) to a 1-Mb region within the previously reported 4.1-Mb minimal region,13 from nt 100 382 250 bp to nt 101 346 495 bp on Human Feb. 2009 (GRCh37/hg19) Assembly. This region contains the SIM1, MCHR2, and ASCC3 genes. SIM1 encodes a transcription factor that mediates hypothalamic paraventricular nucleus development. In mice, postnatally induced Sim1 deficiency causes hyperphagic obesity, and Sim1 overexpression partially corrects the obesity by normalising food intake.46, 47 Sim1 neuron ablation in adult mice induces hyperphagic obesity.48 In humans, SIM1 disruption due to an apparently balanced translocation caused severe obesity and hyperphagia in a girl.49 Obesity was a feature in several patients with 6q16 deletion and SIM1 deficiency (Figure 3 and Table 3). Loss-of-function variants in SIM1 may cause human obesity with or without PWS-like features.37, 38, 39, 40 However, in our study, SIM1 deletions in patients no. 5, 11, 12, and 13 were not associated with obesity. Thus, although SIM1 may have a critical role in regulating body weight, SIM1 deletion is not sufficient to develop obesity. In patients no. 12 and 13, the impact of other associated chromosomal abnormalities cannot be excluded. In particular, patient no. 13 had an additional 16p11.2 duplication that might have protected against obesity, as this copy number variation is associated with a low body mass index.50 On the contrary, patient no. 15 is obese, despite having a deletion that does not encompass SIM1. Another obese patient with a 6q16 deletion, sparing SIM1, was previously reported, but no gene is known in the overlapping deleted region in these two patients (our patient no. 15 and the patient no. 11 of Rosenfeld and collaborators study).17 A position effect cannot be excluded, although none of the known SIM1 enhancer sequences is deleted in these two patients.51

Figure 3
figure 3

Schematic alignment of 6q16 deletions obtained using Database of genomic Variants (DGV) Custom Tracks tool (http://dgv.tcag.ca/gb2/gbrowse/dgv2_hg19/). (a) Representation of molecularly defined 6q16 deletions encompassing 6q16.2 and/or 6q16.3 sub-bands, reported here (red bars) or previously (grey bars). Previously reported deletions were characterised by DNA microarray,3, 13, 15, 16, 17, 18, 22, 23, 25 FISH analysis using BAC clones,21 or STR analysis.24 *Fetal case, **overweight, ***only perinatal data were available. (b) Enlargement of the minimal critical region defined by PWS-like patients, excluding patient no. 15 from the present series and Subject 11 from Rosenfeld et al series. The region contains three genes: MCHR2, SIM1, and ASCC3.

Full size image
Table 3 Summarised description of our patients and published ones
Full size table

Some patients with SIM1 loss-of-function variants have cognitive impairments and/or behavioural disorders.38, 39, 40 However, none of those described to date had a history of neonatal hypotonia or feeding difficulties early in life.38 Recently, a statistically significant association was demonstrated between the SIM1 SNP rs3734354 (Pro352Thr) and language impairment.52 SIM1 loss-of-function is possibly responsible for neurobehavioural disorders. The penetrance and severity of neurobehavioural disorders in patients with SIM1 loss-of-function variants seem to be lower than of those of obesity. Thus, the very high penetrance of cognitive impairment and behavioural disorders in patients with 6q16 deletions is probably due to haploinsufficiency of other genes in the same region. GRIK2 abnormalities may be associated with autistic-like behaviour in patients with 6q16 deletion.13 In our study, three of eight patients having behavioural disorders (patient no. 3, 13, and 14) were not deleted for GRIK2. Although a position effect on GRIK2 cannot be excluded, an alternative possibility is the involvement of other genes in the deleted region. MCHR2 encodes a melanin-concentrating hormone receptor expressed in the brain, and may contribute to regulate body weight in rodents.53 In humans, a SNP of this gene may exert a moderate effect on food-intake abnormalities.54 Genome-wide association studies identified MCHR2 as a putative risk factor for bipolar affective disorders.55 None of our patients had psychiatric diagnoses, but 80% exhibited MCHR2 haploinsufficiency and displayed behavioural features (emotional instability, fits of anger, aggressiveness, hyperphagia). Patient no. 14, who had severe autistic traits and profound intellectual disability, carried the smallest 6q16 deletion in this patient series, encompassing only the MCHR2 gene within the minimal region described here. The functions of the other deleted genes in our patients would not seem to bear any obvious relationship to their phenotype (Supplementary Table S2).

Patient no. 1 is the second prenatally diagnosed reported case of molecularly characterised 6q16 deletion, but the first one with fetopathological examination.24 Autopsy finding included abnormalities in neuronal migration and grey nuclei. However, he had a large 6q16 deletion (14 Mb) that encompassed several developmental genes, including EPHA7.56 In addition, none of the central nervous system abnormalities observed in this patient has been found by brain-imaging studies in previously reported cases of 6q16 deletion. In one study, various brain malformations were found in 65% of patients with 6q16 deletions.17 Of the seven patients who underwent cerebral magnetic resonance imaging in our study, only one had ventriculomegaly, and none had neuronal migration abnormalities.

To conclude, 6q16 deletion syndrome is a contiguous gene-deletion syndrome, in which SIM1 haploinsufficiency probably explains the incomplete penetrance of the obesity phenotype. Our clinical observations support a role in human neurodevelopment for other genes located in the 6q16 region. Further research on how these genes impact brain development and behaviour, together with the identification of additional individuals carrying 6q16 abnormalities, will improve our understanding of how loss of these genes may contribute to the genesis of neurodevelopmental diseases.