Info

Memory Professor System

10x Your Memory Power

Get Instant Access

Figure 2. Correlation between negative symptoms and NAA/Cre in the DLPFC. (From Callicott et al.14).

Our second line of research has been directed at understanding the neuronal mechanisms of brain dysfunction associated with schizophrenia by studying the relationship of NAA alterations to other aspects of functional and neurochemical abnormalities. Prefrontal NAA/Cre has been shown to predict the functioning of distributed cortical circuits, as indexed by fMRI (blood oxygenation level dependent or BOLD) and PET (blood flow) activation. Bertolino et al}1 examined two cohorts of patients with schizophrenia and matched normal controls performing two different tasks that critically engage the prefrontal cortex by activating a working memory network: the Wisconsin Card Sorting test (WCST) of executive cognition and the N-Back working memory test. These subjects received a spectroscopy study at rest and a PET cerebral blood flow study during the execution of either task. In patients with schizophrenia, NAA/Cre values specifically in the DLPFC predicted the degree of blood flow activation in an extended working memory network (Figure 3). Most patients were studied after a 2 week period off medications and the patients who did the WCST performed similarly to controls, ruling out non-specific performance-related explanations of the results. Callicott et al.18 analyzed the relationship of NAA/Cre values in the DLPFC and changes in BOLD response during the execution of the N-Back with fMRI. They found an inverse correlation between NAA/Cre in the DLPFC and BOLD activation in patients with schizophrenia: the greater the apparent neuronal pathology (low NAA/Cre) the larger the change in BOLD signal during the N-Back task (Figure 4). The patients also had greater

BOLD responses than the controls, which was interpreted as lower efficiency in processing the demands of the working memory task. This pattern of inefficient prefrontal cortical processing of executive cognition and working memory is thought to be a core biological aspect of schizophrenia.19 Normal controls showed opposite relationships of performance and fMRI signal than the patients and no relationship of NAA/Cre to fMRI signal. Both patient cohorts (i.e. those who underwent the PET and the fMRI studies) had reduced NAA/Cre in the DLPFC as compared to normal values, however no correlations between reduced NAA/Cre and fMRI signal were present in the hippocampus, which also showed altered fMRI signal.18 Therefore, although the direction of the abnormality in BOLD signal measured with fMRI and CBF measured with PET is opposite, in both cases the NAA/Cre level specifically in the DLPFC predicted the degree of abnormality in activation of the working memory network in patients with schizophrenia. This suggests that the underlying biology in prefrontal cortex monitored by the NAA measures has specific functional implications for processing of prefrontally critical information.

NAA/Cre in DLPFC NAA/Cre in DLPFC

During WCST During sensorimotor control

Figure 3. Relationship of NAA/Cre in the DLPFC to blood flow measured with PET. The top panel shows areas where there was a statistically significant correlation between NAA/Cre measured in the DLPFC and blood flow measures during the execution of the WCST in 13 patients with schizophrenia. Multiple areas in a network that has been associated with working memory are highlighted. The graphs below show the correlation between NAA/Cre in the DLPFC and performance on the WCST (on the left) and on a sensorimotor control task. (From Bertolino et al.17; see color insert between pages 364 and 365.)

NAA/Cre abnormalities in the DLPFC have also been shown to predict the degree of dopamine release in the basal ganglia as measured by the displacement of C11 raclopride by amphetamine in patients with schizophrenia.20 This is an important finding because it shows that NAA/Cre ratios offer a measurement of the functional status of cortical circuitry that is relevant to dopamine transmission in the basal ganglia, the focus of much attention regarding the dopamine hypothesis of schizophrenia. These data also demonstrated for the first time in human beings evidence that prefrontal cortical function related to the activity of subcortically projecting dopamine neurons. Following the same thread of evidence, Bertolino et al.21 showed that NAA/Cre in the DLPFC correlated negatively with D2 receptor availability in the basal ganglia as measured by the binding potential for IBZM (a iodinated D2 ligand). This correlation was not found in any other region explored and was interpreted to mean that a dysfunction of DLPFC (resulting in low NAA/Cre) was related to lower tonic baseline dopamine levels in the basal ganglia (resulting in higher binding potential of IBZM, a ligand that can be displaced by endogenous dopamine).

fMRI vs. performance fMRI vs. NAA

Figure 4. Relationship of NAA/Cre in the DLPFC and fMRI BOLD signal during execution of the N-Back task.The two pictures on the top show which areas of brain have a negative correlation of BOLD signal measured with fMRI during execution of the N-Back task with performance (on the left) and NAA/Cre in the DLPFC (on the right). There is overlap between the areas that correlate with performance and NAA/Cre and all belong to the DLPFC. The graphs show the distribution of the data points for the areas shown above. fMRI signal is in arbitrary units. (From Callicott et al.18; see color insert between pages 364 and 365.)

% Accurate NAA/Cre

Figure 4. Relationship of NAA/Cre in the DLPFC and fMRI BOLD signal during execution of the N-Back task.The two pictures on the top show which areas of brain have a negative correlation of BOLD signal measured with fMRI during execution of the N-Back task with performance (on the left) and NAA/Cre in the DLPFC (on the right). There is overlap between the areas that correlate with performance and NAA/Cre and all belong to the DLPFC. The graphs show the distribution of the data points for the areas shown above. fMRI signal is in arbitrary units. (From Callicott et al.18; see color insert between pages 364 and 365.)

The cumulative weight of the in vivo evidence from human studies has also been strengthened by data acquired in animal models of psychosis. One such model consists of producing lesions in the bilateral ventral hippocampi of newborn rats22 or primates.23 After this procedure, rats developed a number of behavioral, biological and pharmacological characteristics analogous to those seen in schizophrenia, including hyperactivity in response to amphetamine or PCP only as adults. This hyperactivity responds to neuroleptics.24 Bertolino et al.25 demonstrated that neonatal hippocampal lesions in rats lead to reduced NAA/Cre in the DLPFC, but only after the animals reached early adult life. The NAA/Cre ratios were normal in animals prior to puberty despite the hippocampal lesion already being present. Moreover, the alterations in NAA/Cre in the DLPFC of the neonatally lesioned rats25 were not due to cortical thinning and were accompanied by reductions in the concentration of the mRNA for the excitatory amino-acid receptor (EAAC1). EAAC1 is the glial glutamate transporter in the rat, critically responsible for regulation of synaptic glutamate levels, and its reduction could index alterations in glutamatergic function.

Figure 5. Relationship between DA concentration in the caudate nucleus (measured by microdialysis) and NAA/Cre in DLPFC in monkeys. (From Bertolino et al.11)

Similarly, when young monkeys received bilateral hippocampal lesions, the NAA/Cre ratios in the DLPFC of adult animals were lower than those of control monkeys, including those with virtually identical lesions produced during adult life.26 In this study it was only possible to assess adult animals and therefore the full extent of the developmental process could not be observed. Bertolino et al21 also measured striatal dopamine levels at baseline with microdialysis in these monkeys and found that it was positively correlated with NAA/Cre in DLPFC, thus confirming the suspected underlying relationship between DLPFC metabolites and dopaminergic tone in the basal ganglia of the nonhuman primate. When amphetamine was administered in the DLPFC of the monkeys (therefore mimicking a strong stimulus related response), the sign of the correlation reversed, with the animals that were lesioned neonatally releasing the greatest amount of dopamine, while having the largest reductions of NAA/Cre (Figure 5). These data, which were remarkably analogous to those observed in the amphetamine/PET experiment in patients, increase further the relevance of NAA reductions as a marker for fundamental pathophysiological alterations in schizophrenia.

The third thrust of our research has been to use NAA as a biologic phenotype to test hypotheses regarding the effect in brain of schizophrenia susceptibility genes. The first demonstration that NAA/Cre can be considered an "intermediate phenotype" for schizophrenia came from Callicott et al.,13 who showed that reductions in NAA/Cre in the medial temporal lobe were also present in healthy, unaffected siblings of patients with schizophrenia, although to a lesser extent, therefore constituting a likely indication of heritability. Egan et al22'228 have further expanded this notion and linked it to specific genes. For example, brain derived neurotrophic factor (BDNF) contributes significantly to memory performance and to other measures of hippocampal structure and function. Egan et al.21 showed that a functional variation in the gene, a substitution of a single aminoacid (a methionine for a valine) in the BDNF pro-protein, is associated with poorer performance on mnemonic tasks and abnormal hippocampal activation during fMRI. This variation was also associated with altered levels of NAA/Cre in the medial temporal lobe, as seen in figure 6. Normal controls, patients with schizophrenia and their siblings all showed variation associated with genotype, although in the siblings the evidence was not as clear-cut as for the other two groups. Egan et al21 also showed that this mutation causes changes in the intracellular trafficking of the protein, probably altering its secretion in response to neuronal signaling, thus providing a link between the phenomenological observation of memory deficits in schizophrenia and a biological mechanism of action.

Was this article helpful?

0 0

Post a comment