This is the first study to demonstrate that adult subjects with 22qDS and schizophrenia differ from normal controls on quantitative measures of brain tissue and fluid volumes. Subjects with 22qDS-SZ had strikingly larger lateral ventricles together with smaller gray matter volumes. The deficits in gray matter volume were most prominent in the frontal and temporal lobes. These findings are consistent with findings in other patients with schizophrenia (Gur et al 2000a
; Gur et al 2000b
; Harvey et al 1993
; Lim et al 1996
; Sullivan et al 1998
; Zipursky et al 1992
). These results raise the possibility that genetic abnormalities underlying 22qDS may provide clues to the pathogenesis of the brain abnormalities associated with schizophrenia.
Patients with schizophrenia show substantial variability in measures of ventricular and gray matter volumes as well as considerable overlap with normal controls. It has been tempting in the past to explain this variability as being the result of the etiologic heterogeneity of schizophrenia; however, all subjects with schizophrenia in this study had confirmed chromosome 22q11.2 deletions. It could be that the variability in brain structures observed reflects variability in the size and precise location of the deletions, although there are no reported correlations between the extent of the 22q11.2 deletion and the severity or spectrum of the 22qDS phenotype (Demczuk and Aurias 1995
; Lindsay et al 1995
). Alternatively, patients with similar genetic deletions and a shared etiology for their schizophrenia may still demonstrate substantial variability in measures of brain tissue and fluid volumes as a result of heterogeneity in the expression of the genetic abnormality. This is consistent with the recent report by Pearlson et al (1998)
of substantial variability in brain measures in individuals with Down’s syndrome and the general variability of the 22qDS phenotype (Bassett and Chow 1999
; Cohen et al 1999
; Demczuk and Aurias 1995
; Lindsay et al 1995
Our findings of increased ventricular volume and decreased total gray matter volume are consistent with a study of children and adolescents with 22qDS by Eliez et al (2000)
. Both the current study and the Eliez et al (2000)
study noted decreased left parietal gray matter volume. A recent study by Kates et al (2001)
also found decreased left parietal lobe volume in 22qDS children, but the deficit was in white matter. There are several other findings reported in studies of children with 22qDS that were not detected in the current study of adults with 22qDS: decreased total brain volume, decreased total white matter volume, increased adjusted frontal lobe volume, and normal temporal lobe volume (Eliez et al 2000
; Kates et al 2001
). These differences are unlikely to be due to the image processing techniques, as all three studies used the same image processing software (Kates et al 1999
). A number of factors likely account for these differences.
The current study, the Kates et al study (Kates et al 2001
), and the Eliez et al study (Eliez et al 2000
) involved small samples: 14, 10, and 15 22qDS subjects, respectively. The varying difference in the decreases in total brain volume observed 6% in the current study, 8.5% in the Kates et al (2001)
study (both not statistically different), and 11% in the Eliez et al (2000)
study (statistically significant) may therefore reflect sampling errors commonly associated with small sample sizes.
As in the general population, children with 22qDS may have a different pattern of volumetric brain measures than adults with the syndrome. Cross-sectional studies of healthy subjects indicate that cortical gray matter volumes decrease from childhood to adulthood, especially in the frontal lobes, and white matter volumes increase with age (Jernigan and Tallal 1990
; Pfefferbaum et al 1994
; Sowell et al 1999
). A longitudinal MRI study confirmed significant decreases in total cortical, frontal, and parietal gray matter in normal adolescents over a 4-year period (Rapoport et al 1999
). Gray matter volume decreases may therefore only become apparent during or after adolescence. Similarly, white matter deficits that were due to delayed neurodevelopment in 22qDS may disappear by adulthood.
All 22qDS subjects in the current study suffered from schizophrenia or schizoaffective disorder. The psychiatric status of children and adolescents with 22qDS in the two previous studies (Eliez et al 2000
; Kates et al 2001
) was unspecified. It is unlikely, however, that at an average age of 10.5 years (Eliez et al 2000
) or 10.1 years (Kates et al 2001
), any of the subjects had schizophrenia. A recent study of childhood onset schizophrenia suggests that total cortical gray matter reduction over time may be greater in schizophrenia than controls, and temporal lobe gray matter volume reduction over time may be associated with schizophrenia (Rapoport et al 1999
). Cross-sectional data in the current study of 22qDS adults () suggested that gray matter volumes might be declining more rapidly with age in the patient group relative to the control group. Even though a greater age-dependent decline in total gray matter volume in 22qDS-SZ individuals relative to controls was not confirmed in a regression analysis, there was a significant age effect on regional (right occipital) gray matter volume reduction in 22qDS-SZ. Interestingly, an age-related decline in temporal gray matter volumes was recently reported in a cross-sectional study of 22qDS children (Eliez et al 2001
). Taken together, these two findings may provide some supportive evidence of an abnormal or deteriorating neurodevelopmental course in 22qDS, which may or may not be modified by the expression of schizophrenia. Only a longitudinal investigation of individuals with 22qDS including repeated MRI assessments from childhood to adulthood could properly address the issue as to whether the degree and pattern of reductions in gray matter volumes are different in 22qDS subjects who eventually develop schizophrenia than in those who do not.
While we have described significant quantitative differences in ventricular and gray matter volumes in patients with 22qDS and schizophrenia in the current study, it would be very important to know whether those patients with 22qDS who do not go on to develop schizophrenia differ from this group in measures of brain structure. This would be critical in establishing a link between the type and magnitude of the brain abnormalities and the development of schizophrenia. Patients with 22qDS are believed to have a 25% risk of developing schizophrenia (Murphy et al 1999
). If structural brain abnormalities are more common in those who develop schizophrenia, then qualitative and quantitative evaluation of MRI brain scans may have potential value in assisting in the identification of individuals with 22qDS who are most likely to develop schizophrenia. This could be of significant clinical importance in developing strategies for the possible prevention or early treatment of schizophrenia in individuals with 22qDS.
The current study has several limitations: 1) As mentioned above, the sample size for the 22qDS-SZ subjects and matched controls was small, and some true differences may therefore have been missed due to low power; however, the fact that several significant differences were observed indicates that effect sizes for brain structural abnormalities in this subtype of schizophrenia are relatively large. 2) Although other studies have noted changes in the size of cerebellum and corpus callosum in children with 22qDS (Altman et al 1995
; Eliez et al 2000
; Mitnick et al 1994
; Usiskin et al 1999
), these structures were not assessed in the present study. 3) We did not control for differences in IQ between the two groups in this study. It could be argued that some of the findings may be due to the lower IQ of the 22qDS-SZ subjects. IQ appears to correlate with total brain volumes in adults (Pennington et al 2000
); however, total brain volumes did not differ significantly between the two groups in the current study, and our findings changed little after correction for total brain volumes. Also, lower intellectual functioning is a common feature of 22qDS (Bassett et al 1998
; Golding-Kushner et al 1985
; Swillen et al 1997
); therefore, controlling for IQ may remove the precise effects that are of interest. 4) Findings in the current study may not apply to adults with 22qDS who do not have a psychotic disorder. Such individuals would make an interesting comparison group in future studies to help determine which findings are related to schizophrenia and which to 22qDS.
In summary, this quantitative MRI study of adults with 22qDS-SZ demonstrated differences in brain structures relative to normal controls that are consistent with previous findings in schizophrenia. This work suggests that 22qDS-SZ is associated with a number of brain structural abnormalities, which can be identified using quantitative methods. As only some individuals with 22qDS develop schizophrenia, it may be that particular brain differences underlie the development of schizophrenia in those affected patients. The results also suggest that 22qDS may provide a valuable model for investigating the relationship between genetically determined abnormalities in brain development and the expression of schizophrenia.