Xp11.22 deletions encompassing CENPVL1, CENPVL2, MAGED1 and GSPT2 as a cause of syndromic X-linked intellectual disability

By searching a clinical database of over 60,000 individuals referred for array-based CNV analyses and online resources, we identified four males from three families with intellectual disability, developmental delay, hypotonia, joint hypermobility and relative macrocephaly who carried small, overlapping deletions of Xp11.22. The maximum region of overlap between their deletions spanned ~430 kb and included two pseudogenes, CENPVL1 and CENPVL2, whose functions are not known, and two protein coding genes—the G1 to S phase transition 2 gene (GSPT2) and the MAGE family member D1 gene (MAGED1). Deletions of this ~430 kb region have not been previously implicated in human disease. Duplications of GSPT2 have been documented in individuals with intellectual disability, but the phenotypic consequences of a loss of GSPT2 function have not been elucidated in humans or mouse models. Changes in MAGED1 have not been associated with intellectual disability in humans, but loss of MAGED1 function is associated with neurocognitive and neurobehavioral phenotypes in mice. In all cases, the Xp11.22 deletion was inherited from an unaffected mother. Studies performed on DNA from one of these mothers did not show evidence of skewed X-inactivation. These results suggest that deletions of an ~430 kb region on chromosome Xp11.22 that encompass CENPVL1, CENPVL2, GSPT2 and MAGED1 cause a distinct X-linked syndrome characterized by intellectual disability, developmental delay, hypotonia, joint hypermobility and relative macrocephaly. Loss of GSPT2 and/or MAGED1 function may contribute to the intellectual disability and developmental delay seen in males with these deletions.

Armeanet et al. identified a male with developmental delay and intellectual disability who carried an Xp11.22 deletion involving SHROOM4 and intellectual disability and autism has been described by Qiao [1,6,9]. Mutations of SHROOM4 have been reported to cause Stocco dos Santos X-linked mental retardation syndrome [OMIM #300579] and mutations of PHF8 cause mental retardation syndrome, X-linked, Siderius type [OMIM # 300560] [24][25][26][27][28]. Although mutations of FAM120C have not been associated with a particular human phenotype, this gene has been proposed as a positional candidate gene for autism based on its expression pattern in the brain and FAM120C's interaction with CYFIP1, an important binding partner of the fragile X mental retardation protein (FMRP) [6,29].
We describe four males (subjects 1-4) from three families who carry maternally inherited, partially overlapping Xp11.22 deletions that encompass two pseudogenes, CENPVL1, CENPVL2, whose functions are not known, and two protein coding genes-the G1 to S phase transition 2 gene (GSPT2) and the MAGE family member D1 gene (MAGED1). These deletions do not include HUWE1, KDM5C, IQSEC2, TSPYL2, SHROOM4, PHF8 or FAM120C. Subjects 1-4 have a common constellation of findings which includes developmental delay, intellectual disability, hypotonia, joint hypermobility and relative macrocephaly. Short stature and failure to thrive were also common.

Subject identification and accrual
Subjects 1 and 2 were identified by searching a clinical database of individuals referred to Baylor Genetics for copy number variant (CNV) analysis. Subjects 3 and 4 are siblings who were identified through a search of the DECIPHER database (https://decipher.sanger.ac.uk/; patient 276715). All families provided written informed consent for their inclusion in this study. This consent was obtained using consent forms approved by the Institutional Review Board of Baylor College of Medicine. The family of subject 1 has given written informed consent (as outlined in PLOS consent form) to publish his case details and photographs.

Array-based CNV analyses
Array-based CNV analyses were performed for subjects 1 and 2 on a clinical basis at Baylor Genetics. These analyses included a custom-designed array comparative genomic hybridization (aCGH) analysis using 400,000 probes, for the detection of gains or losses of genomic material, and a single nucleotide polymorphism (SNP) analysis using 60,000 probes, for the detection of uniparental disomy (UPD) and absence of heterozygosity (AOH).

X-inactivation studies
A HUMARA assay was used to determine the X-inactivation pattern in the mother of subject 1 [30]. Briefly, extracted DNA was treated with the methylation-sensitive restriction enzyme HpaII followed by PCR amplification of the CAG repeats in the AR gene. In this assay, random X chromosome inactivation is demonstrated by the generation of two PCR bands of similar intensity.

Results
By searching a clinical database of >60,000 individuals referred for CNV analyses, we identified two males (subjects 1-2) that were hemizygous for relatively small (< 2 Mb) Xp11.22 deletions involving CENPVL1, CENPVL2, MAGED1 and GSPT2. A search of the DECIPHER database (https://decipher.sanger.ac.uk/), revealed two brothers (subjects 3 and 4) who carried partially overlapping Xp11.22 deletions that also involved these genes. All of these individuals inherited their deletions from their asymptomatic mothers. The molecular findings and clinical phenotypes of these subjects are summarized below and in Fig 1 and Table 1.
Subject 1 is a 3-year-8-month-old male of Northern European descent who carries an~1 Mb Xp11.22 deletion (minimum deletion chrX:50,847,688-51,773,705; maximum deletion chrX:50,789,912-51,786,912; hg19) that he inherited from his asymptomatic mother (Fig 2). This deletion was confirmed using a series of PCR primers designed to amplify unique regions of Xp11.22 (S1 Table). Studies performed on the maternal sample did not show evidence of skewed X-inactivation. His asymptomatic brother was not found to carry this deletion.
Subject 1 was conceived naturally by non-consanguineous parents. His family history was notable for two previous miscarriages. He was born at 39 weeks gestation via cesarean section due to a failure to progress. His birth weight was 3.675 kg (75 th centile) and his length as was 54 cm (99 th centile).
Over time, he was noted to have global developmental delay. He sat by himself at 1 year of age. He could walk with the aid of a walker at 22 months of age. He learned to crawl at 2 years of age and walked independently at 2 years 3 months of age. He transferred objects from one hand to another at 8-10 months of age, fed himself with his hands at 15 months of age and held a spoon at 2.5 years of age. He began to coo at 18 months of age and by 3 years of age  could say some phrases and could echo words, but was not speaking in sentences. A sedated auditory brainstem response (ABR) evaluation and a brain MRI performed at approximately 21 months of age were normal. He is currently enrolled in a developmental preschool program and receives speech and physical/occupational therapy. Other medical problems encountered by subject 1 include: congenital muscular torticollis which resolved spontaneously by 2 years of age, laryngomalacia for which he underwent supraglottoplasty at 6 months of age, a right-sided cryptorchidism and a right-sided inguinal hernia which were corrected surgically at 10 months of age, gastroesophageal reflux disease that was treated medically until he was 2 years 5 months of age, and food allergies diagnosed at 1-year of age when he had anaphylaxis after eating eggs.
At 3 years 8 months of age, his height was 95.1 cm (11 th centile), his weight was 13.3 kg (7 th centile), and his occipital frontal circumference (OFC) was 51.5 cm (85 th centile). On exam he was found to have relatively small testes, a hypoplastic scrotum, a sandal gap, bilateral medial malleolar displacement and joint hypermobility with a Beighton hypermobility score of 8/9 with hip hypermobility not being tested due to patient's inability to follow commands [33]. Dysmorphic features were not noted (Fig 3).
Subject 2 is a 7-year-9-month-old male of Asian Indian descent who carries an Xp11.  Xp11.22 deletions as a cause of syndromic X-linked intellectual disability that involves two genes-ACOT9 and SAT1 [OMIM # 300862; 313020]. Both of these changes were inherited from his asymptomatic mother.
ACOT9 and SAT1 have not been implicated in human disease. ACOT9 encodes Acyl-CoA thioesterase 9, a protein that hydrolyze Coenzyme A esters [34]. Since only a part of ACOT9 is duplicated, this could lead to a loss of ACOT9 function. We note, however, that ACOT9 has a very low probability of loss-of-function intolerance in the ExAC database (http://exac.broadinstitute.org/; pLI = 0.0) [35]. Hence is it unlikely that haploinsufficiency of this gene would be deleterious. The entire SAT1 gene is duplicated. This gene encodes spermidine/spermine N1-acetyltransferase, a rate-limiting enzyme in the catabolic pathway of polyamine metabolism [36]. Overexpression of SAT1 in mice leads to permanent hair loss at the age of 3 to 4 weeks and female infertility due to ovarian hypofunction and hypoplastic uteri [37]. Taken together, these findings suggest that the Xp22.11 gain carried by subject 2 is unlikely to be causative. Subject 2 was conceived naturally by non-consanguineous parents. Pregnancy was complicated by poor maternal weight gain and concern for poor fetal growth starting at 33 weeks gestation. He was born via normal vaginal delivery at 39 weeks gestation. At birth, he weighed 2.95 kg (20 th centile) was 52 cm long (87 th centile) and had an OFC of 33.5 cm (22 nd centile). He was noted to have central hypotonia and was difficult to feed.
At one month of age he was diagnosed with gastroesophageal reflux disease for which he continues to be followed by a gastroenterologist. Over time, he was noted to have developmental delay. He sat at 8 months of age, pulled to a stand at 17 months of age and began walking at 24 months of age. He had foot and ankle weakness/instability requiring the use of ankle foot orthoses (AFOs). Weakness in his hands was manifest in difficulty in writing letters at school.
At 33 months of age his developmental age was estimated to be 17 months. At 3 years 4 months of age he had less than a 100-word vocabulary and was not using phrases. More recent assessments have shown a statistically significant difference between verbal skills at 76 (5 th centile) and nonverbal reasoning 92 (30 th centile) as assessed by the Differential Ability Scales-2 nd Edition. At 6 years of age, his adaptive behavior, social-emotional, and cognitive scores, as determined by the Developmental Profile-3, were quite low at 2%, 2%, and 0.1% respectively. He was originally enrolled in a developmental preschool but ultimately transitioned into mainstream classes. He is currently in the second grade but continues to receive occupational and speech therapy. A brain MRI obtained at three years of age was normal.
At 7 years 9 months of age his height was 119.2 cm (10 th centile), his weight was 18.5 kg (<1 st centile, Z score = -2.40) and his OFC was 54.5 cm (96 th centile). He had diffuse, profound hypotonia, was still using AFOs, had pes planus, and had bilateral medial malleolar displacement. He was noted to have joint hypermobility and facial weakness with a tented mouth but did not have dysmorphic features. He had normal deep tendon reflexes. He made excellent eye contact but struggled to execute simple commands and had very limited expressive language.
Subjects 3 and 4 are full brothers who both carry~1.5 Mb Xp11.22 deletions (chrX:51,357,052-52,838,176, hg19) that they inherited from their asymptomatic mother. Subject 3 is described in the DECIPHER database as patient 276715. Subject 3 and 4 were conceived naturally by non-consanguineous parents. They have an older, paternal half-brother who has no medical problems. A deceased maternal great uncle was reported to be "handicapped" but further details regarding his condition are not available.
Subject 3 is a 6-year-old male who was born at 42 weeks gestation. His birth weight was 3.3 kg (50 th centile). He started walking at 27 months of age and talked at 36 months of age. He has mild hypermetropia and intermittent exotropia, hypotonia, joint hypermobility and arthralgias. He does not have dysmorphic features. At 6 years of age, his height was at the 0.4 th centile, his weight at the 2 nd centile, and his OFC at the 50 th centile. Subject 4 is a 4-year-old male who was born at 36 3/7 weeks of gestation by emergency cesarean secretion. His birth weight was 2.7 kg (39 th centile). He started talking at 24 months of age and walking at 30 months of age. He had exotropia and was treated for amblyopia. He has epicanthal folds but does not have additional dysmorphic features. He has a single capillary nevus. He has pes planus for which he uses orthotics. At 4 years of age his height was at the 2 nd centile, his weight at the 25 th centile and his OFC at the 75 th centile.
It is currently unclear whether deletion of the~430 kb critical region on Xp11.22 defined by subject 1-4 constitutes a contiguous gene deletion syndrome, in which the associated constellation of phenotypes is due to the loss of two or more genes, or if the loss of a single gene within the critical region is sufficient to cause all of these phenotypes.
Loss of CENPVL1 and CENPVL2 as a potential contributor to the phenotypes seen in subjects 1-4 CENPVL1, CENPVL2 are pseudogenes of the centromere protein V gene (CENPV) located on chromosome 17p11.2 [OMIM #608139]. CENPV is a kinetochore protein whose deficiency in cells leads to defects in chromosome alignment in metaphase, lagging chromosomes in anaphase, failure of cytokinesis, and apoptotic cell death [38]. CENPV also plays a role in directional cell motility and acts as a scaffolding molecule that links microtubules and Src family kinases [39].
Although it is unclear whether CENPVL1, CENPVL2 are required for normal development, a growing body of evidence suggests that pseudogenes may perform a variety of biological functions including the regulation of their parent genes [40][41][42][43][44]. Hence, it is possible that loss of CENPVL1 and CENPVL2 may contribute to the phenotypes of subjects 1-4.
Loss of GSPT2 as a potential contributor to intellectual disability and developmental delay GSPT2 is widely expressed with relative abundance in brain [45]. Whibley et al. suggested that GSPT2 was a candidate gene for X-linked intellectual disability after identifying a maternally inherited tandem duplication of GSPT2 in two brothers with intellectual disability [46]. The effects of GSPT2 deficiency on CNS function have not been demonstrated in humans or studied in mice. However, GSPT2 has a high probability of loss-of-function intolerance (pLI = 0.9) with no loss-of-function variants being seen among controls catalogued in the ExAC database even though 7.3 loss-of-function variants were expected [35]. GSPT2 encodes a GTPase that belongs to the GTP-binding elongation factor family. GSPT2 complexes with eukaryotic peptide chain release factor 1 to mediate translation termination and prevent the translation of an extended protein product (readthrough) [47,48]. This suggests that loss of GSPT2 may lead to the production of abnormally elongated proteins which might, in turn, be detrimental to cells. Since GSPT2 is more abundantly expressed in the brain, its loss may have greater effects on these cells. Xiao et al. have shown that GSPT2 also interacts with BIRC5 (also known as survivin), an antiapoptotic protein that plays an essential role in early brain development [49,50]. They suggest that GSPT2 and BIRC5 may function cooperatively in nuclear processes. However, ablation of Birc5 in the neuronal precursor cells of mouse embryos starting at E10.5 leads to reduced brain size associated with severe, multifocal apoptosis in the cerebrum, cerebellum, and brainstem-a phenotype not seen in our subjects who have relative macrocephaly and, in the case of subjects 1 and 2, normal brain MRIs [49,50].
Loss of MAGED1 as a potential contributor to intellectual disability and developmental delay A search of the ExAC database suggests that there is a paucity of loss-of-function variants in MAGED1 with only one loss-of-function being observed and 16.4 being predicted. This gives MAGED1 a high probability of loss-of-function intolerance (pLI = 0.98) [35]. Although it is unclear how loss of MAGED1 function affects humans, experiments performed in mice suggest that loss of MAGED1 may result in neurocognitive and/or neurobehavioral abnormalities.
Bertrand et al. demonstrated that MAGED1 is highly expressed throughout the brain particularly in the striatum, the hippocampus, the incipient dentate gyrus, the thalamus, the hypothalamus, the olfactory bulb and in the neuroepithelium and subventricular zone of the cerebral cortex [51]. Subsequently, Yang et al. used behavioral tests and electrophysiological recording to show that 5-6-month-old Maged1-null mice displayed reduced basal synaptic transmission, pronounced hippocampal dysfunction, impaired spatial learning, and a deficit in long-term potentiation induction [52]. They also showed that loss of MAGED1 resulted in a reduced dendritic spine density and a reduction in the number of synapses in the hippocampi of Maged1-null mice. They concluded that MAGED1 plays a role in synaptic transmission and hippocampus-dependent learning and memory function.
Dombret et al. demonstrated that a lack of MAGED1 function in male mice results in deficits in social interactions [53]. Specifically, these mice also had reduced ultrasonic vocalization responses to females in estrus, reduced direct social interactions with normal social preference, reduced exploratory behavior, impaired social memory or no preference for social novelty, increased self-grooming and increased anxiety. Dombret et al. went on to shown that although the hypothalami of Maged1-null males were normally formed, their hypothalamic mature oxytocin levels were lower than in wild-type males, possibly due to defective oxytocin post-translational processing. Interestingly, an acute subcutaneous oxytocin injection rescued the impaired social recognition of MAGED1-deficient male mice. They concluded that a lack of MAGED1 function could play a role in autism.
Mouri et al. demonstrated depression-like behavior in Maged1-null mice such as decreased social exploratory behavior, decreased social interaction, increased immobility time during forced swimming and tail suspension, and a decrease in sucrose preference [54]. They found that acute and chronic administration of sertraline and imipramine reversed all or part of the depression-like behavior in these mice. They concluded that a lack of MAGED1 function could play a role in the development of depressive behaviors.
The molecular mechanisms by which MAGED1 affects brain development and function have not been fully elucidated. However, Sullivan et al. demonstrated that MAGED1 binds and positively regulates the transcriptional activity of a subset of transcription factors in the basic helix-loop-helix (bHLH) PER-ARNT-SIM (PAS) family [55]. This subset includes SIM1, NPAS4 and ARNT2 whose deficiencies have been shown to affect brain development/function [55][56][57][58].
Phenotypes seen in MAGED1-deficient mice which are not currently evident in subjects 1-4 MAGED1-deficient mice develop progressive obesity associated with hyperphagia and reduced motor activity [53]. Obesity is not seen subjects 1-4. In contrast, short stature and failure to thrive are common. Maged1-deficient mice also have a significant osteoporotic phenotype with a marked decrease in bone density and deterioration of trabecular architecture [53]. The bone densities of subjects 1-4 have not been measured, but none have a history of recurrent or pathogenic fractures. It is possible that these phenotypes will become apparent in subjects 1-4 over time. Alternatively, these phenotypes may be absent due to inherent differences between mice and humans or due to the compensatory effects of the reduced function of other genes in the Xp11.22 critical region.

Conclusions
These results suggest that deletions of an~430 kb region on chromosome Xp11.22 that encompass CENPVL1, CENPVL2, GSPT2 and MAGED1 cause a distinct X-linked syndrome characterized by intellectual disability, developmental delay, hypotonia, joint hypermobility and relative macrocephaly. Loss of GSPT2 and/or MAGED1 function may contribute to the intellectual disability and developmental delay seen in males with these deletions.
Supporting information S1 Table.