Multigenerational autosomal dominant inheritance of 5p chromosomal deletions

INTRODUCTION

5p deletion syndrome (5p-) is a genetic disorder resulting from a terminal or interstitial deletion of the short arm of chromosome 5, and is among the most common deletion syndromes. When encompassing a minimal critical region, monosomy 5p may be characterized as Cri du Chat syndrome (CdCS). The incidence of CdCS ranges from approximately 1:15,000 to 1:50,000 in live-born infants, and about 1:350 in the intellectually disabled population [Cerruti Mainardi, 2006]. The hallmark features of CdCS include a high-pitched monotonous cry, low birth weight, microcephaly, hypotonia, poor growth, and developmental delay [Mainardi et al., 2001]. Specific craniofacial findings include round face, bilateral epicanthal folds, hypertelorism with a broad nasal bridge, down slanting palpebral fissures and short philtrum [Cerruti Mainardi, 2006]. CdCS is also associated with variable intellectual disability and neuropsychiatric manifestations including autistic behaviors [Church et al., 1995; Moss et al., 2008]. In addition, some individuals with CdCS may present with retinal vessel abnormalities, optic atrophy, or congenital renal and heart defects. Most patients with CdCS, however, do not have other major congenital malformations [Cerruti Mainardi, 2006].

Individuals with 5p deletion demonstrate variability in the clinical presentation. The high degree of phenotypic variability is mainly attributed to the size of the deletion, which can range from 5 to 40 Mb, and the gene content of the monosomic region [Simmons et al., 1998]. A more severe phenotype and cognitive impairment were reported to be associated with larger deletions [Wilkins et al., 1983; Cornish et al., 1999]. A number of genotype–phenotype correlation studies have been performed allowing for definition of critical regions associated with specific manifestations, such as the region for speech delay at band 5p15.3 [Church et al., 1995], the region for cat-like cry between proximal 5p15.3 and distal 5p15.2 [Overhauser et al., 1994; Gersh et al., 1995; Levy et al., 2002; Wu et al., 2005], and the region associated with dysmorphic features at band 5p15.2 [Overhauser et al., 1994; Church et al., 1995; Gersh et al., 1995; Church et al., 1997; Mainardi et al., 2001]. Multiple chromosomal locations have been implicated in intellectual disability and microcephaly, including 5p15.31 [Rossi et al., 2005], 5p15.2, 5p15.1, and 5p13p14 [Overhauser et al., 1994; Gersh et al., 1995; Mainardi et al., 2001]. Several genes on 5p, including SEMA5A (Semaphorin 5A), CTNND2 (delta-catenin), and TERT (human telomerase reverse transcriptase), have been studied in association with autism spectrum disorder, intellectual disability, schizophrenia, or altered fetal development [Medina et al., 2000; Zhang et al., 2003; Harvard et al., 2005; Vrijenhoek et al., 2008; Weiss et al., 2009].

Approximately, 90% of 5p deletions are de novo in origin, and ∼10% of cases arise from translocation events [Niebuhr, 1978; Cerruti Mainardi, 2006]. Autosomal dominant inheritance of 5p deletion has been rarely documented [Kushnick et al., 1984; Church et al., 1995; Cornish et al., 1999; Fang et al., 2008]. Here we report three multi-generational kindreds with 5p deletions, demonstrating intra- and inter-familial phenotypic heterogeneity. Each family carries a relatively small terminal deletion with breakpoints distal to 5p15.1. Phenotypic variability is manifested among the affected members of each kindred, and suggests that modifying factors contribute to the clinical presentation. On the basis of the breakpoints delineated in this study, along with correlation from the literature, our data further support the refinement of the critical region for the cat-like cry phenotype.

METHODS

Informed consent was obtained from study participants.

CLINICAL REPORTS

Family I

The proband (III-1 in Fig. 1A), was a 5 year 9 month old male with developmental delay, and autism spectrum disorder. His birth weight was 4.08 kg (∼95th centile), length 54.6 cm (>95th centile), and occipitofrontal circumference (OFC) 36.83 cm (∼98th centile). He had tracheomalacia, gastroesophageal reflux, and an unusual cat-like cry during the newborn period. His family reported problems of constipation. There was concern for developmental delay prior to the age of 3 years, as developmental milestones included rolling over and sitting unsupported at 9 months, and walking at 2 years with an uncoordinated gait. He had self-stimulatory and self-injurious behaviors, and had recently been diagnosed with autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD). He was attending school in a mainstream setting with a special education teacher. He was receiving physical, occupational, and sensory integration therapies. Previous testing for Fragile X syndrome was normal. At his initial clinic visit, his weight was in the 25th centile for age, height 16th, and OFC between 50th and 75th. He was difficult to examine due to anxiety. He had myopia and wore corrective lenses. He had a broad forehead, deep-set eyes, and preauricular pit on the left, but no other unusual findings.

The father of the proband (II-3) had a medical history that included a cat-like cry during infancy, a ventricular septal defect (VSD), pharyngeal flap surgery, and constipation. He had no hearing problems. He experienced speech articulation problems and learning difficulties, particularly in mathematics, spelling, and reading, requiring special services in school. However, he was able to obtain a bachelor's degree in computer science and maintains a full time job in a related field. The mother of this proband (II-4) was healthy, but had a history of learning problems, anxiety, and attention deficit disorder (ADD). The younger brother of the proband (III-2) was developmentally normal and did not have the cat-like cry in infancy.

A paternal aunt (II-2) demonstrated a cat-like cry in infancy. She was born at 37 weeks with a birth weight of 2.34 kg (10–25th centile) and length 45.7 cm (10th centile). She had no eyebrows or fingernails at birth. She was diagnosed with hypothyroidism and a small esophagus that required surgical repair. She had no known hearing problems. She had myopia and wore glasses. She was diagnosed with speech articulation problems, learning difficulties, and hyperactivity in elementary school and required resource room help. She had a pharyngeal flap procedure in the 7th grade, which improved her articulation problems. She graduated from high school.

The paternal grandmother of the proband (I-2) was 157.5 cm tall. She has worn glasses since age of 10. She experienced multiple ear infections as a child, but had no known hearing issues. She was diagnosed with hypothyroidism, had a heart condition described as a “skip beat,” and had high cholesterol. She had premature graying of hair. She had anxiety issues. Her overall development history was not available. She graduated from high school, but had learning problems.

RESULTS

Family I

Microarray analysis using the Affymetrix SNP 6.0 platform revealed the proband (III-1 in Fig. 1A) carries an approximately 10.36 Mb deletion encompassing the distal portion of 5p (arr[hg19] 5p15.33p15.2(17,630-10,381,437)x1), which overlaps the region associated with CdCS (Fig. 2A and Fig. 3). The breakpoint of the deletion was estimated to be located between basepair positions 10,381,437 and 10,382,584 of chromosome 5 based on build 37 of the human genome sequence (hg19), within the first intron of MARCH6. FISH analysis using the LSI EGR1/D5S23, D5S721 probe confirmed the deletion in the proband (Fig. 2B) and revealed that it was paternally inherited. On the basis of the inheritance pattern, further FISH studies were conducted on at-risk family members, and revealed that the 5p deletion was also present in the proband's mildly affected paternal grandmother and moderately affected paternal aunt (Fig. 1A; FISH data not shown).

DISCUSSION

Deletions involving the short arm of chromosome 5 demonstrate variable phenotypic effect. The size and gene content of the deletion are the main determinants of the phenotypic variability associated with 5p deletion. Genotype–phenotype correlation studies have identified the critical regions for CdCS-associated clinical features, including the cat-like cry (5p15.3), speech delay (5p15.3), as well as facial dysmorphism, microcephaly and severe intellectual disability (5p15.2) [Cerruti Mainardi, 2006]. In general, patients who have larger deletions with break points in the proximal portion of 5p14 and 5p13 tend to have more significant cognitive impairment. Those who have smaller deletions with break points in band 5p15.3 exhibit less significant problems.

Most individuals with CdCS have a de novo deletion. However, familial translocations and other rearrangements have also been reported [Mainardi et al., 2001]. The multigenerational inheritance of a 5p deletion has been documented only in a few kindreds [Kushnick et al., 1984; Church et al., 1995; Cornish et al., 1999; Fang et al., 2008]. In this report, we present three familial cases of 5p deletion syndrome characterized by molecular and/or cytogenetic methodology. These terminal deletions were inherited in an autosomal dominant manner and manifest variable intra-familial phenotypes.

Monosomy for a contiguous stretch of genes at the distal portion of chromosome 5p is associated with CdCS. Several genes in the region including SEMA5A, CTNND2, and TERT are of particular interest because their products play major roles during embryonic and neuronal development [Simmons et al., 1998; Medina et al., 2000; Zhang et al., 2003]. The SEMA5A gene encodes a membrane protein with a semaphorin domain and several thrombospondin type-1 repeats, and acts as a signaling molecule that regulates axonal guidance and neuronal migration during development. Haploinsufficiency of SEMA5A may contribute to the neurodevelopmental phenotype observed in CdCS [Simmons et al., 1998], as recent evidence suggests, this gene may play a role in autism susceptibility [Weiss et al., 2009]. The CTNND2 gene is expressed early during neuronal development, and encodes an adherens protein known to control neuronal cell motility. Monosomy of the CTNND2 gene has been demonstrated in association with severe intellectual disability and developmental delay in CdCS [Medina et al., 2000].

Among the three families presented here, the copy number of SEMA5A and CTNND2 was variable, due to deletion size and location. Family II, harboring the largest 5p deletion among the kindreds at 13.52 Mb, demonstrated monosomy for both the SEMA5A and CTNND2 genes. In contrast, in family I, the 10.36 Mb loss results in monosomy for SEMA5A and disomy for CTNND2 as the deletion breakpoint occurs between the two genes. Finally, the relatively small 4.79 Mb deletion in Family III does not encompass either gene, resulting in a disomic state for SEMA5A and CTNND2 (Fig. 3). The presence of CTNND2 in a two-copy state may explain the milder cognitive impairment observed in Families I and III as compared to individuals who harbor larger 5p deletions. The literature supports this observation on the basis of historic CdCS studies, as well as more recent observations, in that duplication of 5′ CTNND2 and deletion of the 3′ portion of the gene in a CdCS patient is hypothesized to result in amelioration of the cognitive phenotype [Sardina et al., 2014]. Small exonic deletions of CTNND2 have also been reported in individuals with low normal IQ and learning problems with or without autistic features or developmental delay [Asadollahi et al., 2014]. In addition, a recent study identified that deleterious variants in the CYFIP1, DLG1, PLXNA3, and CTNND2 genes are enriched in severely affected patients in female-enriched multiplex families (FEMFs) with severe autism and demonstrated the loss-of-function effect of CTNND2 in autism and neurodevelopment by in vivo and in vitro functional analyses [Turner et al., 2015].

Interestingly, PLXNA3, with enriched variants in FEMFs, encodes a receptor for SEMA5A [Turner et al., 2015], which is located distal to CTNND2 (Fig. 3). The SEMA5A locus has been identified to be among candidate genomic regions associated with autism risk by several Genome-wide association studies (GWAS) [Weiss et al., 2009; Prandini et al., 2012]. SEMA5A expression is significantly decreased in the brain of individuals with autism [Melin et al., 2006]. Moreover, SEMA5A expression quantitative trait loci significantly overlap rare autism-specific CNVs and known autism candidate genes, suggesting that SEMA5A acts as a common downstream effector for known autism CNVs and genes [Cheng et al., 2013]. The presence of autistic behaviors in Families I and II with monosomy of SEMA5A, while not in Family III with normal copy number of the gene, provides additional evidence to support the hypothesis that SEMA5A is an autism susceptibility locus and that aberrant modulation of SEMA5A expression plays an important role in pathophysiology of autism.

The telomerase reverse transcriptase (TERT) gene is localized at 5p15.33 and encodes the rate-limiting component of telomerase complex essential for telomere length maintenance and sustained cell proliferation [Zhang et al., 2003]. It has been demonstrated that individuals with CdCS who are haploinsufficient for TERT have shorter telomere length compared to age-matched unaffected individuals [Zhang et al., 2003]. Insufficient or suboptimal expression due to monosomy of TERT in our three familial cases is likely to play a contributory role in associated clinical phenotypes. Consistent with this, the affected paternal grandmother (I-2; Fig. 1A) in Family I manifested premature graying of hair, which has been reported in other individuals affected by CdCS [Kushnick et al., 1984].

Notably, these three genes account for only a proportion of the phenotype observed in 5p deletion syndrome, and concomitant loss of other genes in the region is certain to play an important role in the clinical features as well. Broader study of gene function and structure in the 5p region is necessary to further elucidate genotype–phenotype correlations. Intriguingly, on the basis of the breakpoints defined in this study, relative to phenotype, our data further support the narrowing of the cat-like cry critical region. The proband in Family III has dysmorphic features, developmental delay, microcephaly, and poor growth, but does not exhibit a cat-like cry. Therefore, the microdeletion in this family does not appear to include the critical region for the cat-like cry in CdCS [Wu et al., 2005]. Recently, Elmakky et al. [2014] reported a small terminal deletion (arr[hg19] 5p15.33p15.32(22,178-5,539,182)x1) due to an unbalanced 5;15 chromosomal translocation causing atypical CdCS phenotypes including cat-like cry [Elmakky et al., 2014]. On the basis of the breakpoints from Family III in our study, along with the proband reported by Elmakky, we narrow down the critical region for cat-like cry to a 733.4 Kb region at 5p15.32 (4,805,799–5,539,182) encompassing only three genes, LINC01020 (a long noncoding RNA gene), ADAMTS16, and ICE1. The developmental role of these two protein-coding genes remains largely unknown, and both are potential candidates for the cat-like cry phenotype associated with CdCS. ADAMTS16 belongs to the ADAMTS (a disintegrin and metalloproteinase domain with thrombospondin motifs) gene family, encoding secreted proteinases. The ICE1 product is one protein component of little elongator complex ELL, which regulates the transcription of small nuclear RNAs (snRNA) [Smith et al., 2011]. Interestingly, a zebrafish ICE1 mutant (ice1^hi3655dTg) demonstrated abnormal pharyngeal arches (Phenotype Annotation (1994–2006) (2006) Mutant Data Curated from Older Literature. ZFIN Historical Data). It is notable that pharyngeal anomalies, which could contribute to high-pitch voice, are frequently observed in patients diagnosed with Cri-du Chat. Since ICE1 regulates transcription and may be more susceptible to haploinsufficiency, the potential involvement of ICE1 in mammalian larynx formation warrants further investigation.

We identified seven reports of familial 5p terminal deletions in the literature (summarized in Fig. 3 and Table I). Among them, only one was characterized by high-resolution microarray technology with a deletion of 10.5 Mb [Fang et al., 2008]. The breakpoint was estimated to be located within the ropporin 1-like gene (ROPN1L). This deletion is similar to that identified in Family I of our study. In the Fang et al. study, the affected members showed variable clinical features. Two affected females had moderate intellectual disability, as well as psychoses (auditory hallucinations, paranoia, delusional thinking, and self- laughing), rarely associated with CdCS and not present in the other three affected male family members nor in Family I of this study. The remaining cases reported in the literature were characterized by conventional chromosome and FISH analysis and therefore the size and breakpoints could not be precisely determined [Church et al., 1995; Cornish et al., 1999; Kushnick et al., 1984]. The breakpoints were estimated to be located within band 5p15.3 for five cases, and band 5p15.1 for the case reported by Kushnick et al., 1984 (Table I and Fig. 3). All affected members of the family reported by Cornish et al. [1999] presented high-pitched cat-like cry, but variable learning disability, psychomotor development, and intelligence quotient (IQ) ranging from 70 to 105. Four cases from Church et al. [1995] are atypical without cat-like cry and facial dysmorphism, and only presented mild phenotypes with developmental delay and low-normal intelligence with a variable expressivity among affected family members. The 5p deletion with the breakpoint within 5p15.1 in the family of Kushnick et al. [1984] is the largest among all familial cases summarized in Table I. Both the proband and the mother had cat-like cry in infancy. The proband exhibited hypotonia and several facial dysmorphism. The mother had mild intellectual disability, several dysmorphic features, and premature graying of hair.

5p deletion syndrome is a well-described clinical entity, but multigenerational inheritance has been rarely documented. We present a detailed clinical and cytogenetic/molecular characterization of three kindreds with familial 5p deletions. Intra-familial phenotypic variability suggests that additional modifying elements including genetic and environmental factors may impact the clinical manifestations observed in patients with 5p deletions. In this line of thought, it is notable that in families I and II neuropsychiatric and developmental concerns were documented in both the maternal and paternal parent of the proband. It is also of interest that within each multigenerational family, the proband was an affected child diagnosed in the newborn or early childhood period. This occurrence may be associated with a more severe presentation in early development, and/or may be attributed to improved access to healthcare and increased knowledge among healthcare providers allowing for early referral and diagnosis of this genetic disorder. In addition, the results from these clinical data further support refinement of the critical region for cat-like cry to a ∼700 Kb region at 5p15.32 encompassing only two coding genes, ADAMTS16 and ICE1. Over time, the use of high-resolution technology such as SNP arrays and next-generation sequencing techniques to define breakpoints at basepair resolution will continue to aid in the delineation of genotype–phenotype correlations in 5p deletion syndrome.