Dr. Jarskog, for many years the dopamine theory of schizophrenia has been dominant in the field. What is the current status of that theory?The dopamine theory was originally based on the purported mechanism of action of antipsychotic drugs, which is dopamine receptor antagonism. Since these drugs improve psychotic symptoms in patients with schizophrenia, it seemed logical to assume that high levels of dopamine in some part of the brain are a cause of the illness. For a long time, this indirect support was the only real evidence of elevated dopamine in schizophrenia. But, with advances in imaging techniques such as SPECT and PET along with development of dopamine 2 receptor radioligands, it became possible to measure actual baseline levels of dopamine in vivo in patients, and it was found that dopamine levels in striatum were approximately twice as high in patients with schizophrenia with acute psychotic symptoms (Abi-Dargham et al., PNAS 2000; 97:8104-8109). This was found both in patients who had been previously treated as well as in first-episode neuroleptic-naïve patients, as compared to healthy controls. This was really a dramatic and seminal finding.

Does this finding change the way we think about the dopamine hypothesis?It doesn’t necessarily change our thinking, but it validates the fact that the dopamine receptor is an appropriate treatment target in schizophrenia. But another interesting finding from these studies was that acutely psychotic patients with the highest levels of baseline dopamine responded best to treatment. This suggests that there may be different types of schizophrenia. Patients can have more or less excess dopamine in the subcortical region of the brain.

So this could be the beginnings of the first example of the Holy Grail of a lab test in psychiatry where one could envision using a PET scan to determine the amount of dopamine in the caudate putamen and help you decide whether or not a particular patient is going to respond to treatment.That is right, in theory. It has some promise, but it is certainly not just around the corner.

Aside from this finding, hasn’t there been research showing that certain regions of the brain have too much dopamine and others too little dopamine in schizophrenia?Right. We believe that the subcortical regions have too much dopamine, and there is evidence–although less so–that there is too little dopamine in the cortex, especially the frontal cortex. 

So we basically have a picture of too much dopamine subcortically, which I assume might cause positive symptoms; and too little dopamine in the cortex causing cognitive impairment and negative symptoms.
Right, although negative symptoms are less well understood in terms of a mechanism.

What about the movement problems, the EPS side effects; where do these come from?That is due to blocking dopamine receptors in the dorsal striatum, the nigro-striatal pathway. There is a significant hub of neurons controlling gross and fine motor activity, coordination and initiation of movement in that region, and so when you block these D2 receptors, it causes Parkinsonian-like side effects. If you block over 80% of the available D2 receptors in the dorsal striatum you are mimicking Parkinson’s disease where symptoms such as tremor and muscle rigidity emerge after loss of about 80% of dopamine neurons.

Getting back to the basic neurobiology, in one of your papers, you wrote that, “Synaptic dysconnectivity has emerged as a core neuropathological deficit in schizophrenia.” Can you explain this?
This refers to emerging evidence that there is something wrong with the neuronal synapses in schizophrenia. In postmortem neurophathological studies, the number of dendritic spines is substantially reduced in the prefrontal cortex in schizophrenia (Glantz and Lewis, Archives of General Psychiatry 2000; 57:65-73). There is also evidence of a substantial reduction in the number of proteins specific to synaptic structure and function, such as a 30%-40% reduction of the amount of synaptophysin in the frontal cortex.

What is “synaptophysin”?
Synaptophysin is a protein that is present in almost all synapses, so it is a good marker for synaptic content. Its exact function is not understood but it probably contributes in neurotransmitter release, and a reduction of 30%-40% of synaptic content is really a substantial reduction.

I have also heard about studies showing a loss of brain volume over time in schizophrenia.
Yes. The data show that among patients who are prodromal for schizophrenia (who do not yet meet the criteria for the diagnosis), those who went on to develop psychosis had reductions in gray matter in specific areas of the cortex (Pantelis et al., Lancet 2003; 361:281-288). Those that did not become psychotic had no significant change in gray matter volume. Other studies have focused on patients who have been studied longitudinally after the onset of psychosis, and many have also shown some degree of gray matter loss (Cahn et al., Archives of General Psychiatry, 2002: 59:1002-1010).

I assume the patients in these studies were receiving antipsychotics?
Yes, and it is not clear whether the loss of gray matter is caused only by the underlying disease or also in part by the use of certain antipsychotics. But there is preliminary evidence from studies of childhood onset schizophrenia that some of this gray matter loss may be related to genetic factors specific to the illness.

What is the role of glutamate and GABA in the neurobiology of schizophrenia?
The initial interest in glutamate came from the observation that the drug of abuse PCP (phencyclidine) is a very strong NMDA (N-methyl-D-asparate) glutamatergic receptor antagonist, and it can temporarily cause all the symptoms of schizophrenia in an otherwise healthy individual. So this led to the theory that schizophrenia may be related to a low glutamate state.

What does glutamate normally do in the brain?
Glutamate is the main excitatory neurotransmitter and glutamatergic neurons are found throughout the brain. Normally, glutamate neurons in the cortex send axons down to dopamine neurons in the mesocortical areas, and when these neurons are stimulated, they send inputs back up to the frontal cortex. But in schizophrenia, there seems to be a decrease in both the number and activity of dopamine neurons projecting up into the frontal cortex. So why would there be such a decrease? The theory is that reduced glutamate activity in the cortex indirectly leads to this decreased dopamine activity.

And where does GABA fit into this scheme?
Glutamate neurons in the cortex also normally project onto subcortical GABA (gamma amino butyric acid) neurons. GABA is the main inhibitory neurotransmitter of the brain. Glutamate neurons normally stimulate GABA neurons that in turn inhibit dopamine neurons that extend into the striatum (the so-called mesolimbic dopamine neurons). So, if you have reduced glutamate signaling in schizophrenia, you have reduced activation of these GABA neurons, leading to less inhibition of mesolimbic dopamine neurons, allowing for more dopamine to be transmitted into the striatum, which is consistent with the subcortical hyperdopaminergic state.

That’s quite an elaborate mechanism! Can you tell us the significance?
Sure. The theory is that in schizophrenia too little glutamate in the cortex leads to less stimulation of mesolimbic GABA neurons, thereby allowing too much dopamine to be released in the mesolimbic area, which leads to positive psychotic symptoms like hallucinations and delusions. But in the cortex, the mechanism is somewhat different: too little glutamate, too little excitation of mesocortical dopamine neurons, and therefore too little dopamine released back in the cortex, accounting for cognitive deficits and possibly negative symptoms.

While this is all fairly complex, the good news is that the low glutamate hypothesis dovetails with the dopamine hypothesis.
Correct, and this has been reassuring to people and it has also spawned a great interest in modulating the glutamate system therapeutically to try to achieve symptomatic response using this mechanism. From what I have said already, it is clear that if you modulate glutamate you ultimately end up affecting dopamine as well. 

Have any glutamate stimulators been tested in schizophrenia?
There are a number of different compounds such as glycine, D serine, and D-cycloserine. These are all indirect agonists at the NMDA receptor (the major glutamate receptor). The problem with glutamate receptors is that if you directly activate the NMDA receptor, you can get excitotoxicity because you have released too much glutamate and that causes cell death. You always have to be concerned about the seizure potential of drugs that modulate glutamate. So we use indirect agonists that just kind of nudge the glutamate receptor into action, and we call them “glutamate modulators.” 

Are they effective in schizophrenia?
One of the larger studies was published last year, in which glycine and D-cycloserine were tested as an augmentation strategy for patients on antipsychotics, and that study was negative (Buchanan et al., American Journal of Psychiatry, 2007; 164:1593-1602). There was no significant effect on cognitive or negative symptoms, which were the main outcome variables. That was disappointing, but we are early in trying to target glutamate receptors. There was another study in Nature Medicine a little over a year ago that tested a metabotropic Glu2/3 receptor agonist as monotherapy, and it looked almost as good as olanzapine in acutely psychotic patients with schizophrenia (Patel et al., Nature Medicine, 2007; 13:1102-1107). So, it is a work in progress. It remains to be seen if someone who doesn’t respond to a conventional dopaminebased antipsychotic could respond to a glutamate-based treatment. That is certainly what we hope.