BDNF Val66Met Polymorphism and GAD₆₇ mRNA Expression in the Prefrontal Cortex of Subjects With Schizophrenia

In the prefrontal cortex of subjects with schizophrenia, inhibitory neurotransmission appears to be altered as indicated by the decreased expression of mRNA encoding markers of γ-aminobutyric acid (GABA) neurotransmission, such as the 67-kilodalton isoform of glutamate decarboxylase (GAD₆₇), an enzyme for GABA synthesis (1, 2). This mRNA is down-regulated in the subclass of GABA neurons containing the calcium-binding protein, parvalbumin, and the mRNA encoding parvalbumin was also found to be decreased (3). These alterations are likely to contribute to prefrontal cortex dysfunction in schizophrenia, given the evidence that parvalbumin-containing GABA neurons play a critical role in the regulation of prefrontal cortex information processing during working memory tasks (4).

Several lines of evidence suggest that the signaling mediated by brain-derived neurotrophic factor (BDNF) and its receptor tyrosine kinase (TrkB) is involved in these alterations in GABA neurons in schizophrenia. First, BDNF regulates the maturation of GABA neurons in the developing cortex and influences the expression of GABA markers, including GAD₆₇ and parvalbumin (5, 6). Second, the levels of mRNA and protein for BDNF and TrkB are decreased in the prefrontal cortex of subjects with schizophrenia (7, 8). Finally, these decreases in BDNF and TrkB mRNA are significantly correlated with the decreases in GAD₆₇ mRNA levels (8).

A single nucleotide polymorphism (rs6265) in the human BDNF gene (bdnf), which produces a valine to methionine substitution at codon 66 (Val66Met) in the 5′ proregion of BDNF, was reported to reduce intracellular trafficking and activity-dependent secretion of BDNF (9). In addition, subjects with the Met66 allele showed abnormal hippocampal activity and poorer performance during an episodic memory task, consistent with a reduced functional effect of BDNF in hippocampal circuitry (9). Thus, the presence of the Met66 allele in subjects with schizophrenia would be predicted to be associated with a greater decrease of cortical GAD₆₇ mRNA expression relative to subjects with schizophrenia homozygous for the Val66 allele.

Method

The 27 subject pairs examined in this study have been described previously (8) as cohort 1 (subject pairs 1–15) and cohort 2 (subject pairs 16–27). Each pair consisted of a subject with schizophrenia and a comparison subject matched by age, gender, and postmortem interval. In order to determine BDNF Val66Met genotype, genomic DNA was isolated from cerebellar tissue of each subject using DNeasy Tissue Kit (Qiagen, Valencia, Calif.) and analyzed with Taqman 5′-exonuclease assay as described previously (9). Information regarding the PCR conditions and sequences of the primers and Taqman probes is available upon request. The expression levels of GAD₆₇ and BDNF mRNA in the prefrontal cortex of the 27 subject pairs were reported previously (8). Because the strongly negative correlation between age and prefrontal cortex BDNF expression (8, 10) would confound comparisons of mRNA expression, we compared within-pair differences in mRNA expression for two subject pair groups defined by bdnf genotype: pairs in which both subjects carried only the Val66 allele and pairs in which only the schizophrenia subject carried at least one Met66 allele.

The frequency of the Met66 allele and the proportion of subjects with this allele were compared between schizophrenia and comparison subjects using a chi-square test. To assess the impact of the Met66 allele on the change in GAD₆₇ mRNA expression, within-pair differences in BDNF and GAD₆₇ mRNA expression were subjected to a mixed-model two-way analysis of variance (ANOVA) with bdnf genotype of the subject pair (i.e., pairs in which both subjects were homozygous for the Val66 allele versus those in which only the schizophrenia subject had one or more Met66 alleles) as a between-pair factor and mRNA type (BDNF versus GAD₆₇) as a within-pair factor. Because there was a significant correlation between within-pair differences in GAD₆₇ and BDNF mRNA levels across subject pairs (8), the net effect of the Met66 allele on within-pair differences in GAD₆₇ mRNA levels, after within-pair differences in BDNF mRNA levels were controlled, was assessed by the interaction between the two factors.

Results

The proportion of subjects with the Met66 allele was significantly greater (χ²=5.3, df=1, p<0.03) among subjects with schizophrenia (six heterozygous and one homozygous subject) than among comparison subjects (one heterozygous subject) (Figure 1). The frequency of the Met66 allele in the schizophrenia group (15%) and the comparison group (2%) also significantly differed (χ²=5.9, df=1, p<0.02).

Within-pair percentage changes in BDNF and GAD₆₇ mRNA expression levels for each pair are shown in Figure 1. In pair 20, the schizophrenia subject showed a 310% increase in BDNF mRNA expression level. As reported previously, this abnormally high expression of BDNF mRNA (more than seven standard deviations above the comparison subject mean) appeared to be due to a tricyclic antidepressant overdose (8). We excluded pair 20 from the analysis. Pair 15 was also excluded because the schizophrenia and comparison subjects both had one Met66 allele and one Val66 allele.

The mean within-pair percentage changes in BDNF and GAD₆₇ mRNA expression levels in the schizophrenia subjects were –15.0% and –9.9%, respectively, for the five subject pairs with a Met66 allele in the schizophrenia subject, and –38.5% and –31.2%, respectively, for the 20 pairs homozygous for the Val66 allele (Figure 2). Two-way ANOVA failed to reveal a significant interaction between pair genotype and mRNA type (F=0.22, df=1, 23, p=0.88), indicating that the presence of a Met66 allele in schizophrenia subjects did not contribute to the decreased level of GAD₆₇ mRNA expression. The ANOVA did reveal a tendency for pair genotype to affect within-pair differences in mRNA expression (F=3.76, df=1, 23, p<0.07), suggesting that BDNF and GAD₆₇ mRNA expression changes might be smaller in the subject pairs with the Met allele relative to those without a Met allele.

Discussion

Although the primary focus of this study was to assess the effect of bdnf genotype on GAD₆₇ mRNA expression, we unexpectedly encountered a significantly lower frequency of the bdnf Met66 allele in the comparison group (2%) than in the schizophrenia group (15%). However, previous case/control studies using larger populations reported much higher frequencies (19%–22%) of the Met66 allele in comparison subjects that were similar to (9) or even higher than (11) the frequency found in schizophrenia subjects. Considering the limited size of our cohort, it seems likely that the low frequency of the Met66 allele in the comparison group represents a chance finding.

Because BDNF has been shown to induce GAD₆₇ expression in cortical neurons in vitro and in vivo (5, 6), the decreased expression of GAD₆₇ mRNA in the prefrontal cortex of subjects with schizophrenia might reflect a reduced availability of BDNF. BDNF signaling in the cortex is determined not only by its synthesis in neurons but also by its release from them. Thus, GAD₆₇ expression levels might be modified by the BDNF Val66Met genotype, given that the Val to Met substitution in the 5′ prodomain of pro-BDNF peptide reportedly causes abnormal intracellular trafficking and a subsequent reduction of activity-dependent secretion of BDNF (9). Thus, we expected that the magnitude of decrease in GAD₆₇ mRNA would be greater in schizophrenia subjects with the Met allele after the level of BDNF mRNA expression was controlled. However, no evidence for such an effect was found, suggesting that the possible deficit in activity-dependent release of BDNF caused by the presence of the Met allele does not contribute to the decrease in GAD₆₇ mRNA expression in subjects with schizophrenia. It is interesting that the decreases in both GAD₆₇ and BDNF mRNA expression in the prefrontal cortex appear to be less severe in schizophrenia subjects with the Met66 allele than in those without the Met66 allele. Further studies are needed to both confirm this observation and to illuminate potential mechanisms underlying the association between the presence of the Met allele and a tempering of the deficit in GAD₆₇ and BDNF mRNA expression in subjects with schizophrenia. These findings, together with our previous observation of no change in cortical GAD₆₇ mRNA expression in mice with a neuron-specific inducible knockout of bdnf(8), suggest that the reduced availability of BDNF may not be a primary cause of decreased GAD₆₇ expression in the prefrontal cortex of subjects with schizophrenia. In contrast, other factors such as the availability of TrkB (8) or allelic variants in the GAD₆₇ gene (12) may be more important.

Received May 18, 2005; revision received Aug. 17, 2005; accepted Sept. 19, 2005. From the Department of Psychiatry and the Department of Neuroscience, University of Pittsburgh. Address correspondence and reprint requests to Dr. Lewis, Department of Psychiatry, University of Pittsburgh, W1650 Biomedical Science Tower, 3811 O’Hara St., Pittsburgh, PA 15213; [email protected] (e-mail).Supported by a Young Investigator Award from the National Alliance for Research on Schizophrenia and Depression’s Essel Foundation (Dr. Hashimoto) and by NIMH grants MH-43784 and MH-45156 (Dr. Lewis).The authors thank Dr. Daniel Weinberger (NIMH) for providing technical information regarding human BDNF Val66Met genotyping.