The behavioural phenotype in velo-cardio-facial syndrome.

03Original abstract

The concept of a ‘behavioural phenotype’ was initially coined by William Nyhan in a presidential address to the Society of Pediatric Research (Nyhan 1972). In a vivid description of the aggressive self-mutiliation he observed in patients with Lesch–Nyhan syndrome, Nyhan emphasized the role of organic factors in the development of these behaviours and referred to the behavioural phenotype as ‘behaviours which are an integral part of certain genetic disorders’. Some years later, Flint & Yule (1994) further defined the behavioural phenotype as ‘the characteristic pattern of motor, cognitive, linguistic and social abnormalities which is consistently associated with a biological disorder. . . . In some cases may constitute a psychiatric disorder’. Alternatively, the behavioural phenotype may also be defined as ‘the characteristic behavioural, psychiatric, neuropsychological and linguistic components of a genetic disorder’. Velo-cardio-facial syndrome (VCFS), the most frequent known interstitial deletion found in man, occurs in approximately 1 : 6000 live births (Botto et al. 2003). In the vast majority of individuals, VCFS is associated with chromosomal microdeletions in the q11 region of chromosome 22. The VCFS phenotype is complex with multiple congenital abnormalities affecting a number of tissues and organs. Although considerable phenotypic variability occurs, the majority of abnormalities reflect disturbances of neural crest cell development. Common physical phenotypic features include a characteristic dysmorphology, cleft palate and a high prevalence of congenital heart defects. More recently, however, a characteristic behavioural phenotype has also been described in VCFS. In this review, I will outline the evidence for a characteristic behavioural phenotype in VCFS by reviewing studies of behavioural, psychiatric, neuropsychological and linguistic disorders in children and adults with VCFS. Early studies of children with VCFS reported a stereotypic personality with poor social interaction (quantitatively and qualitatively), a bland affect with minimal facial expression and extremes of behaviour, notably disinhibited and impulsive or serious and shy (Golding-Kushner et al. 1985; Swillen et al. 1997). In a study of 15 children and adolescents, Papolos et al. (1996) reported high rates of bipolar II disorder (47%), attention deficit with hyperactivity disorder (27%) and attention deficit disorder without hyperactivity (13%). Additionally, ourselves and others have reported high rates of autistic spectrum disorders, anxiety disorders and emotional instability in VCFS children (Goldberg et al. 1993; Swillen et al. 1997; Niklasson et al. 2001; Stevens et al. 2002). In a cross-sectional study of a cohort of 60 Flemish children and adolescents with VCFS, Swillen et al. (1999a) found that, up to the age of 3 years, parents of such children most often reported somatic complaints (eating problems and constipation) and withdrawn behaviour. From early school age to adolescence, parents and teachers most frequently reported social difficulties (problematic peer relationships), social withdrawal and attention problems with no significant gender differences. Swillen et al. (1999b) have suggested that the poor social skills and social withdrawal might at least in part be a function of the adolescents’ impaired communication abilities (see below). From the age of 11 years, parents also reported increased levels of anxiety and depressed mood in adolescents with VCFS. In order to exclude the potential confounding effects of learning disability, Swillen (2001) and colleagues compared primary school children with VCFS with a group of children matched for learning disorder and speech and language impairment (children with IQ < 70 were excluded from both groups). They found that, while both groups behaved similarly with respect to problematic social interaction, poor attention and anxiety, VCFS children were more withdrawn and less aggressive compared with controls. This suggests that VCFS children may be more likely to choose internalizing behaviours compared with children with similar developmental levels. Conversely, a study by Feinstein et al. (2002) evaluated the rates and types of psychiatric disorder in children with VCFS and compared them with those found in children matched for age and cognitive ability. They reported that there were no differences in either the rate or the type of disorder found in the two groups, although both groups had greatly increased levels of psychopathology when compared with the normal population. They suggested that the explanation for this finding may be that subtle phenotypic differences in behaviour found in VCFS may not be identified using standard symptom inventories. As the first recognized cohort of children and adolescents with VCFS was followed up into adulthood, high rates of major psychiatric disorder were observed. Shprintzen et al. (1992) suggested that more than 10% had developed psychiatric disorders that mostly resembled chronic schizophrenia with paranoid delusions, although operational criteria were not used. Subsequently, in a small study of VCFS adults (n = 14), 11 (79%) were found to have a psychiatric diagnosis of which four (29%) had DSM-111-R schizophrenia or schizoaffective disorder (Pulver et al. 1994). In a further study of 10 VCFS adults, Papolos et al. (1996) reported that four (40%) of their sample had psychotic symptoms, two (20%) had schizoaffective disorder, three (30%) had bipolar 1 disorder while three (30%) had bipolar 11 disorder. In a large study, Murphy et al. (1999) found that 18 (42%) of a sample of 50 VCFS adults had a major psychiatric disorder; 15 (30%) had a psychotic disorder with 12 (24%) fulfilling DSM-IV criteria for schizophrenia while a further six (12%) had major depression without psychotic features. We found that individuals with schizophrenia had fewer negative symptoms (symptoms such as apathy, loss of initiative and social withdrawal) and a relatively later age of onset (mean age = 26 years) compared with nondeleted controls. Using different ascertainment strategies, however, Bassett et al. (1998) reported a relatively early age of onset (mean age = 19 years) in their sample of 10 individuals with VCFS and schizophrenia. More recently, Bassett et al. (2003) reported no significant differences in age of onset, positive or negative symptoms or global functioning in a sample of 16 individuals with VCFS and schizophrenia compared with 46 adults with schizophrenia without VCFS. However, in an accompanying editorial, Kendler (2003) suggested that this sample of VCFS individuals with schizophrenia were probably not a representative sample as most were selected because they were first diagnosed with schizophrenia and only later shown to have the deletion. In addition, as the two groups were not well matched for IQ, this lack of matching may also have confounded their results. Clearly, larger and more representative samples using properly IQ-matched controls are required to resolve the differing results between these studies. While borderline or mild learning disability is one of the most consistently reported features in VCFS, severe learning disability is rare. In a study of 37 VCFS children, Swillen et al. (1997) reported a wide variability in intelligence ranging from moderate learning disability to average intelligence with a mean full-scale IQ (FSIQ) of approximately 70. Forty-five per cent of individuals (n = 17) had a learning disability (IQ < 70), the vast majority (82%) of which was mild (IQ 50–69). Similarly, Moss et al. (1999) reported that the mean FSIQ of their sample of 33 children and adults was 71, with 17 (52%) of their sample demonstrating learning disability. There were no differences in mean FSIQ measures between children with or without congenital heart disease or palate abnormalities. However, VCFS individuals with a familial deletion were found to have a lower mean FSIQ than individuals with a de novo deletion (Swillen et al. 1997; Gerdes et al. 1999). A specific neuropsychological profile has also been described in children with VCFS with verbal IQ (VIQ) exceeding performance IQ (PIQ) on tests of general intellectual functioning (Swillen et al. 1997; Gerdes et al. 1999; Moss et al. 1999; Woodin et al. 2001). However, in one of the few studies of VCFS adults, we failed to find a consistent VIQ > PIQ discrepancy in a sample of 19 VCFS individuals (Henry et al. 2002). In fact, six (31.5%) of the sample showed a significant discrepancy between VIQ and PIQ (11 points or more) in favour of PIQ while only three (16%) had a significant discrepancy in favour of VIQ. These findings may suggest differential development of skills in VCFS where, compared with non-VCFS individuals, visual-spatial abilities develop at a slower rate while verbal abilities remain stable. This hypothesis is supported by a preliminary study of 26 children and adolescents (6–16 years old) with VCFS by Campbell et al. (2002), where we found that FSIQ decreased with age and that this decline was entirely attributable to decreases in PIQ over time. No such differences occurred among the control group of ‘typically’ developing siblings. Several studies have suggested that VCFS children perform better at reading and spelling than arithmetic (Swillen et al. 1999b; Wang et al. 2000). Moss et al. (1999) reported that both orally presented mathematical reasoning and ‘rote’ calculation skills appeared equally impaired compared with reading comprehension and single-word reading. In addition, reading comprehension and single-word reading yielded similar results. However, other studies of verbal academic skills report better performance on single-word reading than reading comprehension (Woodin et al. 2001). In order to exclude the potential confounding effects of learning disability, Moss et al. (1999) excluded individuals with an FSIQ lower than 70 from the analysis. The same psychoeducational patterns, that is, VIQ > PIQ, reading and spelling > maths abilities, appeared in this more homogenous group as for the whole group, suggesting that this profile is indeed characteristic of a 22q11 deletion rather than a consequence of a more general learning disability. Campbell & Swillen (2004) have suggested that these deficits in nonverbal and mathematical abilities may be explained by deficits in spatial cognition. In the only study yet performed, we utilized the visual object and space perception battery (Rapport et al. 1998) in a group of 19 adults with VCFS (Henry et al. 2002). We found that VCFS adults had object- perceptual impairments, specifically on the subtests of silhouettes (recognition of objects from unusual views) and object decision (object recognition) compared with an age- and IQ-matched control group. Previously, it has been suggested that ‘rote’ (repetitive) verbal learning and memory are an area of strength for individuals with VCFS while visual- spatial memory and complex verbal memory such as story recall are impaired (Swillen et al. 1999b). More recently, Bearden et al. (2001) examained more specific memory abilities in 29 children with VCFS. The VCFS group demonstrated selective deficits in visual-spatial immediate and delayed memory when compared with ‘rote’ verbal memory. In addition, the VCFS group performed significantly better on an object memory test compared with a pure visual spatial memory task indicating a dissociation between these two domains. Wang et al. (2000) also reported that, in their sample of 36 VCFS children, the majority (67%) of children appeared to be better at number recall than spatial memory, although 20% did show an opposite pattern of performance with higher scores on spatial memory than number recall. Both scores correlated significantly with the children's performance on mathematics subtests. However, we investigated spatial working memory and strategy formation in VCFS adults using the ‘Executive Golf’ task and found no significant differences between the VCFS group and IQ-matched controls (Henry et al. 2002). In a recent study, we examined 13 VCFS adults with schizophrenia, and 15 VCFS adults without a history of psychosis using comprehensive neuropsychological assessments (van Amelsvoort et al. 2004). We found that, compared with the nonpsychotic group matched for age, gender and IQ, VCFS individuals with schizophrenia had specific impairments of strategy formation, spatial working memory and attention. We suggested that these deficits may reflect developmental brain abnormalities in frontal brain regions which may be associated with the development of schizophrenia in VCFS individuals. Delayed speech and language development is one of the most consistent features in VCFS and a major concern to parents of children with VCFS (Golding-Kushner et al. 1985; Gerdes et al. 1999). In the first descriptive study of language in VCFS children, Golding-Kushner et al. (1985) reported delays in comprehension and use of vocabulary and syntax in early childhood and increasing difficulty with abstract reasoning as language demands increased with age. Expressive language and speech skills tended to be disproportionately low in comparison with receptive skills, and both were lower than expected based on tests of cognitive development. Pragmatic skills, including use of language and social communication, were among the most impaired communication skills in all age groups. The youngest children, including those with good speech intelligibility, tended to rely on nonverbal communication at home. The older children (11–18 years) used language that was terse and concrete, and missed nuances of meaning typically gleaned from interpretation of one's communication partner's tone of voice, facial expression and choice of words. While early development in VCFS is often characterized by mild delay in most areas, expressive language is often specifically delayed compared with other milestones (Golding-Kushner 2004). Many children with VCFS are essentially nonverbal through 30 months of age (Scherer et al. 1999; D’Antonio et al. 2001), but show dramatic improvement between 3 and 4 years of age (Shprintzen 2000; Solot et al. 2001). By school age, expressive language and speech improve, perhaps as a result of intervention, but specific language impairment persists (Solot et al. 2001). In addition, higher order receptive language skills involving abstract thinking remain poorly developed, affecting both communication and academic skills. Several studies have sought to determine the cause of language delays and disorders in children and adults with VCFS. Sensorineural hearing loss was reported in approximately 10% of patients with VCFS and conductive hearing loss in about 45% (Digilio et al. 1999). However, the nature of the language disturbance in VCFS is different from patterns of those with hearing loss as a primary aetiology, and very few children with VCFS seem to have speech or language problems directly attributable to hearing problems (Golding-Kushner 2004). How specific is the pattern of communication disorder seen in VCFS individuals? To exclude the potential confounding effects of a more general learning disability, Glaser et al. (2002) compared 27 VCFS individuals aged 6–19 years with age-, gender- and IQ-matched controls with idiopathic developmental delay (DD). Receptive language scores were significantly below expressive language scores for VCFS group, but not for the DD group. This was especially true for children over age 9 years, with whom tasks of reasoning were included. These findings suggest that VCFS individuals have a specific speech and language disorder not attributable to a more general learning disability. Furthermore, VCFS individuals with a maternally derived deletion scored significantly lower on receptive tasks compared with VCFS individuals with a paternally derived deletion (Eliez et al. 2001; Glaser et al. 2002). The available evidence suggests that there is a characteristic behavioural phenotype in VCFS with specific behavioural, psychiatric, neuropsychological and linguistic disorders observed in affected individuals. Why is this important? First, the observation of a specific behavioural phenotype in VCFS has important implications for clinical practice. The high rates of behavioural, psychiatric, neuropsychological and linguistic disorders in affected individuals suggest that clinical geneticists and other physicians treating affected individuals need to consider the risk of such diagnoses in VCFS individuals and should seek specialist referral where appropriate. Second, identification of a specific behavioural phenotype in people with a known chromosome 22q11 deletion suggests that haploinsufficiency of a neurodevelopmental gene or genes mapping to chromosome 22q11 is associated with the high rates of behavioural, psychiatric, neuropsychological and linguistic disorders seen in VCFS individuals. In particular, the high rates of schizophrenia seen in VCFS individuals (Pulver et al. 1994; Murphy et al. 1999), the increased prevalence of VCFS in people with schizophrenia (Karayiorgou et al. 1995; Gothelf et al. 1997; Bassett et al. 1998) and the results of molecular genetic studies of schizophrenia (Harrison & Owen 2003; Lewis et al. 2003) suggest that a gene or genes associated with schizophrenia maps to chromosome 22q11. Third, identification of the genetic determinants of the behavioural phenotype will increase our understanding of the underlying pathophysiology of these disorders in VCFS and in the wider population and will facilitate future genetic epidemiological studies on the effects of gene–environment interactions. Fourth, neuroimaging studies of the behavioural phenotype in VCFS will provide novel insights into the neurobiology of these disorders in VCFS and in the wider population. In a preliminary study of VCFS individuals with and without schizophrenia, we recently reported that VCFS adults with schizophrenia have significant reductions in whole brain volumes, particularly involving white matter (van Amelsvoort et al. 2004). The clinical heterogeneity of disorders such as schizophrenia and attention deficit hyperactivity disorder undoubtedly reflects an underlying genetic heterogeneity. The study of disorders with a known genetic aetiology such as VCFS serves as a model to understand the genetics and pathogenesis of behavioural, psychiatric, neuropsychological and linguistic disorders in the wider population (Murphy & Owen 2001). Recent experience with other genetic disorders – for example, the association between Alzheimer's disease and Down's syndrome – suggests that this approach might be one of the most successful for understanding the genetics and underlying pathophysiology of complex diseases (Murphy 2002).

Journal of intellectual disability research : JIDR, 2004 · doi:10.1111/j.1365-2788.2004.00620.x