|Home | About | Journals | Submit | Contact Us | Français|
A genetic syndrome associated with schizophrenia, 22q11 deletion syndrome (22qDS), may represent a genetic subtype of schizophrenia (22qDS-Sz). Structural brain changes are common in schizophrenia and may involve developmental anomalies, but there are no data yet for 22qDS-Sz. The objective of this study was to assess brain structure in adults with 22qDS-Sz using magnetic resonance imaging (MRI).
Brain and arterial MRI scans of 11 adults with 22qDS-Sz (mean age = 28.4 years, SD = 6.5) were systematically assessed by a neuroradiologist for qualitative anomalies.
A high frequency of abnormalities were found: T2 white matter bright foci (BF), 90%; developmental midline anomalies, 45%; cerebral atrophy or ventricular enlargement, 54%; mild cerebellar atrophy, 36%; skull base abnormalities, 55%; and minor vascular abnormalities, 36%.
BF and skull base abnormalities, especially in association with neurodevelopmental midline abnormalities, may be distinguishing MRI features for a genetic subtype of schizophrenia involving a deletion on chromosome 22.
The term, 22q11 deletion syndrome (22qDS), refers to the congenital syndromes associated with interstitial deletions at chromosome 22q11.2, and includes velocardiofacial syndrome (VCFS), DiGeorge syndrome and conotruncal anomaly face syndrome (Bassett et al 1998; Thomas and Graham Jr. 1997). The 22qDS has a variable phenotype which involves multiple systems, including the central nervous system. Common features include congenital heart defects, palatal abnormalities, typical facial features, and central nervous system manifestations such as learning disabilities and mental retardation. Structural abnormalities of the brain and skull have also been associated with the syndrome (Altman et al 1995; Arvystas and Shprintzen 1984; Beemer et al 1986; Bingham et al 1997; Finkelstein et al 1993). In adulthood, approximately 25% of patients with 22qDS may suffer from a psychotic disorder, including schizophrenia (Murphy and Owen 1997; Pulver et al 1994; Shprintzen et al 1992). The 22qDS has been found in patients with schizophrenia (Bassett et al 1998; Chow et al 1994; Gothelf et al 1997; Karayiorgou et al 1995), and may involve as many as one in 50 patients (Karayiorgou et al 1995; Yan et al 1998). These early studies suggest that 22qDS-schizophrenia (22qDS-Sz) may be a genetic subtype of schizophrenia (Bassett et al 1998).
The most consistent structural neuroimaging finding in schizophrenia is enlargement of the lateral and third ventricles (Johnstone et al 1976; Pfefferbaum and Zipursky 1991; Shelton and Weinberger 1986), which is usually mild and overlaps with a continuum of normal anatomic variability (Daniel et al 1991). Quantitative studies using magnetic resonance imaging (MRI) show overall reductions in volumes of cortical and subcortical gray matter, but not white matter (Zipursky et al 1992, 1994). Volumetric changes in the subcortical structures (especially the basal ganglia and limbic system), cerebellar atrophy, and reduction in corpus callosum have also been described, but the findings are often inconsistent (Pearlson and Marsh 1993; Scott et al 1993; Shelton and Weinberger 1986). Reports that neurodevelopmental midline abnormalities, such as enlarged cavum septum pellucidum (CSP) or its more severe form, complete nonfusion of the septum pellucidum or cavum vergae (CV), occur more frequently in schizophrenia than in affective psychosis or normal control subjects (Degreef et al 1991, 1992; DeLisi et al 1993; Jurjus et al 1993; Kwon et al 1998; Lewis and Mezey 1985; Nopoulos et al 1997; Shioiri et al 1996; Weinberger 1987) have lent support to a neurodevelopmental model for the illness (Nasrallah et al 1990; Weinberger 1987).
There are few systematic neuroimaging studies of 22qDS, and prevalence of specific anomalies is generally unknown. There are reported minor midline defects, however, including CSP (Vataja and Elomaa 1998) or CV (Haapanen and Somer 1993), cysts of the septum pellucidum (Haapanen and Somer 1993) or cysts adjacent to the anterior horns (Mitnick et al 1994), hypoplastic corpus callosum (Conley et al 1979; McDonald-McGinn et al 1995; Ryan et al 1997), empty sella (Haapanen and Somer 1993), and small pituitary gland (Bingham et al 1997). Case studies and case series have described cerebral atrophy, enlargement of sulci and ventricles (Beemer et al 1986; Bingham et al 1997; Haapanen and Somer 1993; Ryan et al 1997), cerebellar atrophy, and small posterior fossa (Lynch et al 1995; Mitnick et al 1994; Ryan et al 1997). Severe developmental anomalies such as hydrocephalus (Bingham et al 1997; Nickel et al 1994; Ryan et al 1997), holoprosencephaly (Wraith et al 1985), and anencephaly (Strong 1968) have been reported. Focal white matter bright foci were noted in up to 30% of 22qDS patients on MRI T2 weighted images (Altman et al 1995; Bingham et al 1997; Lynch et al 1995; Mitnick et al 1994). Except for a recent case report (Vataja and Elomaa 1998), all these studies involved nonpsychotic, and mainly pediatric, patients. It is unclear how applicable these findings may be to adults with 22qDS and schizophrenia.
The current study investigated for MRI brain abnormalities in a case series of 11 adults with 22qDS and schizophrenia to begin to delineate the structural neuropathology of this genetic subtype of schizophrenia. Because of reported skull and vertebral abnormalities (Arvystas and Shprintzen 1984; Beemer et al 1986; Finkelstein et al 1993; Ming et al 1997; Ryan et al 1997), the skull base structures were assessed. In addition, cerebral and cervical arteries were also assessed using magnetic resonance angiography (MRA) because of a reported high prevalence of carotid and vertebral arterial anomalies in 22qDS (Goldberg et al 1993; MacKenzie-Stepner et al 1986; Mitnick et al 1996).
Eleven subjects (six men, five women) were referred by clinicians based on a set of 22qDS ascertainment criteria (Bassett et al 1998). All subjects provided informed consent and had MRI scanning of the brain and MRA of major cerebral and cervical blood vessels. They all had a chromosome 22q11 deletion confirmed by fluorescence-in-situ-hybridization testing using the N25 (Vysis Inc., Downers Grove, IL) probe. Subjects ranged in age from 21 to 38 years old (mean = 28.4 years, SD = 6.5). Axis I diagnoses were determined for the subjects by one of the authors (EC or AB) using SCID-IV (First et al 1997) based on a direct interview with each subject and a review of all available psychiatric records. Ten subjects met DSM-IV (American Psychiatric Association 1994) criteria for schizophrenia and one (subject 9) had schizoaffective disorder. The mean age of onset of psychotic illness was 19.9 years (SD = 4.32). Estimated IQ using Silverstein’s two-subset method (Weschler Adult Intelligence Scale-Revised vocabulary and block design subtests) (Silverstein 1982) ranged from borderline to moderate mental retardation (mean = 69.8, SD = 9.1). Four subjects had ventricular septal defects (VSD) but only two (subjects 2 and 5) had corrective surgery. All subjects had a high arched palate, except subject 3 who had a repaired cleft palate, and subject 5 who had no palatal abnormality. The detailed phenotype of nine of the subjects has been described elsewhere (Bassett et al 1998); the remaining two subjects (subjects 5 and 8) had similar physical and psychiatric findings. No subject had a history of maternal alcohol abuse during pregnancy.
The subjects were scanned using a GE Signa 1.5 Tesla MRI scanner (General Electric Company, Milwaukee) according to a research protocol. Sagittal T1 spin echo, axial proton density, and T2 weight axial images were acquired, which yielded approximately 60 sections of 3 mm thickness for each subject. All axial images were oriented in an oblique plane that passed through the anterior and posterior commissures and was perpendicular to the sagittal plane. MRA using a three-dimensional time-of-flight technique was also performed to visualize the circle of Willis and the carotid and vertebral arteries in the neck. MRA for two subjects (1 and 5) was completed only on the circle of Willis. The MR images were systematically reviewed for clinically detectable qualitative anomalies in the brain, arteries, and skull, over a 4-month period, by a research neuroradiologist (DM) blind to the physical and behavioral status of the subjects. No intra- or inter-rater reliability was measured because each MRI scan was read by a single rater once. Cortical or cerebellar atrophy and ventricular enlargement, when present, were qualitatively rated as mild, moderate, or severe according to the neuroradiologist’s own judgment. Nonfusion of the septum was rated as CSP (if incomplete) or CV (if complete). The location and number of bright foci, when present, were noted for each subject. The skull base structures, including the C1 arch, the cisterna magna, the clivus, and the odontoid, were examined for anomalies as part of the protocol. On MRA, particular attention was paid to the size and location of neck and cranial arteries and their branches.
The MRI and MRA findings are presented in Table 1 and illustrated in Figures 1 and and2.2. The most common findings (in 9 of 10 analysable scans) were T2 weighted white matter bright foci (BF). BF were found bilaterally, mainly in the frontal lobes in subcortical, cortical, and periventricular regions, in a distribution pattern that was not consistent with demylineating disease (see Figure 1, part 1 to part 3). The nine affected subjects had 2 to 63 BF, most of which were small (3 to 4 mm in diameter). Five (45%) subjects had developmental defects in midline brain structures, either CSP or CV. Four subjects had CV (see Figure 1, part 1 and part 2) and one had an enlarged CSP; there was no significant difference in median IQ between subjects with or without these midline defects (70 vs. 75; Z = −1.01; p = .31). Six subjects (54%) had detectable cerebral atrophy and/or enlargement of lateral ventricles (see Figure 1, part 1). Four (36%) subjects had cerebellar hypoplasia, one of which only involved the inferior vermis area. Three (27%) subjects had fourth ventricle enlargement. One subject had several additional structural brain anomalies, which included bilateral subcortical T2 hypointensities in the U fibers, a rare observation of signal dropout at the gray–white matter junction on MRI (see Figure 1, part 4) and signal changes consistent with mineral deposition in the basal ganglia and the thalamus (see Figure 1, part 5), likely related to known hypocalcemia in the subject. Only one subject (subject 8) had no structural brain findings on this systematic qualitative assessment.
Several anomalies of the skull base and upper vertebrae were noted. Six (55%) subjects had an enlarged C1 arch, three (27%) had a short clivus, three had a prominent cisterna magna, and two (18%) had a small odontoid. One subject had a hypoplastic skull base (see Figure 1, part 6).
On MRA, a few minor arterial anomalies were detected, none of which involved the carotid artery. Two subjects had hypoplastic A1 segments in the circle of Willis. Two other subjects had a hypoplastic or a smaller right posterior cerebral artery when compared to the left side (see Figure 2). One patient also had an atrophic right vertebral artery along its entire length.
This initial descriptive study of brain structures in a 22qDS subtype of schizophrenia found a high prevalence of several anomalies. Consistent with previous studies of schizophrenia, which have found an increased prevalence of abnormal CSP or CV in schizophrenia compared to normal control subjects [25.4% vs. 18.9% (Jurjus et al 1993); 35% vs. 13%; 30 vs. 10% (Kwon et al 1998); 12.5% vs. 1% (Nopoulos et al 1998), depending on the criteria used to define an abnormal CSP], the current study also showed a particularly high rate of these midline developmental defects in 22qDS-Sz. The high prevalence of CSP or CV in the current study is not explained by fetal alcohol syndrome, which has been associated with septum pellucidum abnormalities (Johnson et al 1996). CSP and CV have been suggested as evidence of disturbed midline development of the brain, and particularly of the limbic system, a brain area believed to be important in the development of schizophrenia (Bodensteiner and Schaefer 1990; Sarwar 1989; Schaefer et al 1994). CSP and CV may also be related to midline facial anomalies (Kjaer 1995), but the high prevalence of palatal anomalies in the current study precluded the detection of any specific relationship between palatal and septum pellucidum abnormalities.
The high prevalence of cortical atrophy, ventricular enlargement, and cerebellar atrophy in 22qDS-Sz is also consistent with findings in schizophrenia (Johnstone et al 1976; Pfefferbaum and Zipursky 1991; Shelton and Weinberger 1986). Similar findings have been reported, primarily in children, for 22qDS without schizophrenia (Bingham et al 1997; McDonald-McGinn et al 1995; Mitnick et al 1994). In schizophrenia, cortical atrophy or ventricular enlargement may be associated with a poorer outcome and increased cognitive and negative symptoms (Lieberman 1995). The relationship between psychiatric symptoms and neuropsychological functioning and the presence of cortical atrophy and ventricular enlargement will need to be investigated in future studies of 22qDS-Sz patients.
Periventricular and deep white matter BF are nonspecific abnormalities, sometimes found on MRI, which have no definitive anatomic correlates (Marsh et al 1996). BF may be associated with aging, demylineating disease, or cerebrovascular disease (Awad et al 1986; Brown et al 1992; Coffey et al 1987; Fazekas et al 1988). Higher than expected rates of BF have been reported in bipolar disorder and in major depression in the elderly (Brown et al 1992; Coffey et al 1988; Miller et al 1989; Swayze et al 1990). In schizophrenia, rates of BF were reported to be 5% to 10% of patients (Brown et al 1992, 1995), and were comparable to rates in control subjects (7% to 14% (Brown et al 1992, 1995). An increased prevalence of BF reported in congenital rubella patients with schizophrenia (Lane et al 1996; Lim et al 1995) was felt to be related to ischemic changes associated with congenital rubella but not with schizophrenia (Lane et al 1996). Ischemic changes may be responsible for the increased BF in 22qDS. However, none of the subjects in the current study had known cardiovascular risk factors for BF, such as hypertension, cerebrovascular accidents, vasculitis, or diabetes, and only subjects 2 and 5 had had surgical repair for congenital cardiac defects. Furthermore, the previously reported prevalence of BF in 22qDS patients was only 27% (Mitnick et al 1994), about one-third of the rate in the current study. This suggests that BF may be a more common feature in 22qDS then previously thought, and may be particularly prevalent in 22qDS-Sz.
In contrast to brain anomalies, vascular anomalies are rarely reported in schizophrenia (Aleem and Knesvich 1987; Remington and Jeffries 1984). They were also uncommon in this sample of adults with 22qDS-Sz, and less common than previously reported in children with 22qDS scanned prior to corrective palatal surgery (Finkelstein et al 1993; Goldberg et al 1993; Mitnick et al 1996). This may be due to self-correction of positional anomalies of soft tissue during development. For example, medial displacement of the carotid arteries in children may be reducible simply by hyperextension of the neck (Mitnick et al 1996). Alternatively, vascular anomalies in 22qDS may be especially associated with major palatal involvement, which appears to be uncommon in 22qDS-Sz adults (Bassett et al 1998).
The skull base anomalies were unexpected findings. They have not been previously reported in 22qDS or schizophrenia. However, they are consistent with other skeletal abnormalities and platybasia previously described in 22qDS (Arvystas and Shprintzen 1984; Finkelstein et al 1993; Ming et al 1997; Ryan et al 1997). One subject had a hypoplastic skull base, raising the question of structural stability of the cervical spine. The subject had no cervical or neurologic complaints. This type of radiologic observation, however, may be important, as unstable cervical vertebrate and skull base structures may lead to compression of the brain stem and are potentially dangerous. Further skeletal studies are needed to fully assess the clinical significance of these skull base abnormalities in 22qDS.
This descriptive study represented an initial MRI investigation on adults with 22qDS-Sz. It is limited by small sample size and lack of a control group, which prevent the determination of statistical significance of the findings of the current study. However, the results for this putative genetic subtype of schizophrenia indicate some overlap of structural brain anomalies with those consistently found in schizophrenia. This may be related to abnormal development of neural crest cells, which has been suggested as the pathogenesis of schizophrenia (Bogerts 1993) and for 22qDS (Thomas and Frias 1987; Van Mierop and Kutsche 1984). Thus, a shared pathogenetic process of abnormal neural crest cell development may account for the high association between 22qDS and neuropsychiatric disorders (Chow et al 1994). Therefore, 22qDS-Sz may provide a neurodevelopmental model of schizophrenia with a known etiology, suitable for further investigations into the pathogenesis of schizophrenia. Results from this study also suggest that MRI abnormalities likely to be more prevalent in 22qDS-Sz, such as BF and skull base abnormalities, may be features that could help clinicians and researchers identify patients with schizophrenia or related disorders who may be at increased risk of having 22qDS.
This study was supported by grants from the National Alliance for Research on Schizophrenia and Depression, the Scottish Rite Schizophrenia Research Program, and the Clarke Institute of Psychiatry Foundation Research Fund.