ICOSR 2009—Psychosis: Is It All From Your Hippocampus?
As part of our ongoing coverage of the 2009 International Congress on Schizophrenia Research (ICOSR), 28 March to 1 April 2009, in San Diego, California, we bring you session summaries from some of the Young Investigator travel award winners. We are indebted to James Stone of the Institute of Psychiatry in London for this report.
7 April 2009. Organizer Anthony Grace of the University of Pittsburgh introduced the Tuesday, March 31, morning session, entitled, "Hyperactivity within the hippocampal complex as a pathological basis for psychosis in schizophrenia." He said that the idea that hippocampal and frontal hyperactivity could be of central importance to the development of psychosis was gaining ground, with supportive evidence arising from diverse groups. He said he found this particularly gratifying as, only a few years ago, this idea would have almost been considered blasphemy, with most authorities claiming, rather, that hypofrontality was a defining feature of the illness.
In the first talk, Grace presented evidence that, "Hippocampal hyperactivity due to interneuron loss accounts for dopamine hyperresponsivity and diminished oscillatory activity in the MAM model of schizophrenia." (See also SRF's slide cast of Grace's talk at the 2008 SfN meeting.) He first outlined the case for an extension of the neurochemical basis of schizophrenia beyond the dopamine hypothesis, given the little evidence for a primary deficit in the dopamine system. He described the methylazoxymethanol (MAM) rodent model of schizophrenia, which Holly Moore developed when she was in his group. MAM disrupts DNA transcription and, when given to maternal rats at day 17 gestation, leads their offspring to show a number of abnormalities which closely resemble the pathology in schizophrenia. These abnormalities span three domains: anatomical (thinning of cortical limbic structures and increased cell packing density—losing neuropil, not neurons); physiological (impaired PPI, reversal learning, set shifting, and social interactions); and pharmacological (increased locomotor response to amphetamine and PCP, consistent with increased amphetamine-induced dopamine release, with this hypersensitivity arising following adolescence). Grace said that this hypersensitivity appeared to arise from a loss of the bistable “up” and “down” states of dopaminergic neurons. Neurons became unable to enter the “down” state and were no longer gated, continually responding to any input that comes in. Interestingly, a similar phenomenon is seen with PCP treated animals.
Grace then went on to describe how MAM may initiate this change. He stated that MAM leads to hyperreactivity of hippocampus thought to occur secondary to functional impairment of parvalbumin-staining GABAergic neurons (chandelier cells), leading to a hyperresponsivity of dopaminergic neurons in the VTA. However, there is no direct connection between hippocampus and VTA. He hypothesized that the effect occurs via stimulation of GABA neurons projecting from nucleus accumbens to ventral pallidum, in turn leading to reduced firing of GABA neurons from ventral pallidum to VTA. This hypothesis was supported by showing increased GABA neuron firing in nucleus accumbens following NMDA injection into hippocampus. This effect was abolished by blocking GABA or glutamate. Grace went on to add more detail to the model, highlighting the importance of burst firing of dopamine neurons in response to salient stimuli. Burst firing can only occur in dopamine neurons that are already in an active state. He demonstrated that activating the pathway from the hippocampal subiculum to VTA via nucleus accumbens and ventral pallidum led to an increase in the number of dopamine neurons in the active state, whereas a second glutamatergic path from pedunculopontine tegmentum to VTA increases the number of dopamine neurons burst firing. He likened the path from the pedunculopontine tegmentum to the “signal,” whereas the path from the hippocampal subiculum was the “gain,” the latter being activated in situations of potential danger, increasing the magnitude of the response to salient stimuli. In MAM treated animals Grace's group found that there was an increase in the number of dopamine neurons in the active state, but no increase in the rate of burst firing, suggesting the abnormalities induced by MAM primarily affected the gain of the system. They elegantly demonstrated that inactivation of the ventral subiculum reduced the response to amphetamine back to baseline levels.
Grace next discussed the role of stress in the genesis of psychosis. He showed evidence that two hours of resistant stress led to the same increase in activated dopamine cells as seen in MAM animals. He then went on to show that MAM animals had impaired latent inhibition. Lastly, he discussed the use of an α-5 subunit-selective GABA enhancing agent, α-5-containing GABA receptors being enriched in hippocampus. He showed data that this drug had no effect in control animals, but in MAM animals, it reduced the numbers of spontaneously active dopamine neurons to control levels. In summary, he said that loss of functional GABAergic interneurons in hippocampus may be a common feature of schizophrenia and MAM animals, and that this could underlie the abnormalities of gating and of hyperresponsive dopamine neurons in the illness.
Scott Schobel of Columbia University presented a talk entitled "CA1 subfield, subiculum and orbitofrontal cortex are differentially targeted in schizophrenia." He initially discussed the features of any region proposed as a primary site of brain dysfunction in psychosis: alterations should be present in established psychotic illness, as well as in the prodrome, and they should be selectively associated with the clinical features of the illness. He suggested hippocampus fit these criteria, with reduced hippocampal volume being reported in first-episode psychosis, and with the CA1 and subiculum implicated in reduced volume, and hippocampus having shown inappropriate activation during passive viewing of faces in patients with schizophrenia. He discussed the possibility of using different imaging modalities to measure hippocampal sub-fields and showed that cerebral blood volume (CBV), using a gadolinium scan subtraction method normalized to sagittal sinus, had the highest resolution, and that it also correlated directly with PET-FDG measures. He showed that CBV measures in different brain regions could be used to investigate connectivity.
Schobel hypothesized that the normal relationship between hippocampus and its projection structures should be enhanced in schizophrenia. CBV measures were performed in subjects with attenuated psychotic symptoms and first-episode patients, and showed increased CBV in CA1 and orbitofrontal cortex, and reduced CBV in dorsolateral prefrontal cortex, as well as a trend for increase in subiculum and basolateral amygdala in patients with schizophrenia compared to controls. For all these measures, subjects with attenuated symptoms had intermediate CBV compared to patients and controls. He went on to show that CA1 CBV in prodromal subjects predicted transition to psychosis with a positive predictive value of 75 percent and a negative predictive value of 80 percent. Furthermore, CA1 CBV covaried with delusional severity, PANSS positive subscale, as well as with social withdrawal and avolition. Schobel stated that CBV was unlikely to be driven by antipsychotic drug treatment, as in rats there was no association of CBV with risperidone use, and there was no difference between patients on and off antipsychotic drugs. Unlike CBV, hippocampal volume did not predict transition to psychosis, but with transition, hippocampal volume was found to be reduced, primarily in CA1 and subiculum. CA1 hypermetabolism appears static over time, but predicts subsequent hippocampal volume loss. He said that further work needs to be done, as the molecular mechanism underlying increased hippocampal CBV is still unknown, and the relationship to dopaminergic sensitivity, as assessed with PET, has yet to be investigated.
Belinda Garner of the ORYGEN Research Center in Melbourne, Australia, also talked about hippocampus in her presentation, "Hyper-responsivity to stress in the HPA axis during early psychosis." She noted that stress had been implicated in psychosis, with increased emotional reactivity to life stress. The HPA axis interacts with neurotransmitter systems and has a role in neural plasticity. HPA activity has been reported to be abnormally elevated in psychosis, particularly in the early phase and in unmedicated first-episode patients, with increased ACTH and cortisol, and increased pituitary volume in those prodromal subjects who underwent transition compared to those who did not. There was a correlation between pituitary volume and time to transition, with subjects with larger pituitary volume having a shorter latency to transition. Garner also investigated daily life hassles and stressful life events, showing that daily hassles correlated with BPRS and cortisol levels, and that cortisol levels correlated with the Hamilton Anxiety Scale. She also presented data on pituitary volumes and antipsychotic treatment response, finding that subjects with larger pituitary volumes responded less well to quetiapine. Improvements in psychotic and negative symptoms (BPRS and SANSS) correlated with reductions in cortisol blood levels. Garner said that these results supported the value of programs aimed at reducing stress in individuals with prodromal symptoms, and in patients with schizophrenia.
In the last talk, Carol Tamminga of the University of Texas, Southwestern, discussed alterations in hippocampal function in schizophrenia correlated with impairments in declarative and relational memory. She emphasized the role of the two main divisions of the hippocampus, the anterior hippocampus being involved in associative functions, and the posterior hippocampus being termed the sensory hippocampus. She discussed the unidirectional passage of information from dentate gyrus to CA3 to CA1, but also mentioned that recurrent connections, from CA3 to dentate gyrus, also exist. She also described the distinct functions of these subregions, with CA3 being required for pattern completion, and dentate gyrus being involved in pattern identification. Tamminga showed that regional cerebral blood flow was increased in anterior hippocampus in unmedicated patients with schizophrenia by around 8 percent, this effect being lost in patients on medication. She described some preliminary data (n = 2) showing increased CBV (using the MRI gadolinium subtraction method) in anterior hippocampus CA3 in patients with schizophrenia off medication, but not in patients on medication. Furthermore, there was reduced task-relevant BOLD response to a relational memory task in the same region. She suggested that this was caused by a higher resting state, meaning that the increase in blood flow during the task was relatively less, with BOLD calculated as change from baseline. She suggested that, in view of the region's unique function, increased CA3 perfusion might be expected to lead to increased pattern identification, which could conceivably lead to some of the symptoms of schizophrenia, with delusions potentially arising from the tendency to see patterns where there are none.—James Stone.