Schizophrenia Research Forum - A Catalyst for Creative Thinking

Live Discussion: Hippocampus in Schizophrenia Roundtable


Anthony Grace

Stephan Heckers

On Monday, 16 November 2009, Anthony Grace and Stephan Heckers led a live discussion.

The hippocampus has been a focus of increasing interest in the schizophrenia research community since the 1980s (Heckers and Konradi, 2002). A large literature has convincingly shown that hippocampal volume is smaller in schizophrenia patients (for review/meta-analysis, see Steen et al., 2006), as well as in their first-degree relatives (Boos et al., 2007). This is complemented by functional neuroimaging studies, which suggest that hippocampal activity is increased at baseline, but decreased during task performance (numerous studies, see PubMed results). A recent study reports that increased blood volume in the CA1 subfield of the hippocampus may predict the at-risk, or prodromal, stage of psychotic disorders (Schobel et al., 2009; see SRF related news story).

A much-discussed synthesis of hippocampal pathology in schizophrenia, which grew out of a 2007 Society for Neuroscience symposium, also gives the hippocampus a central role (Lisman et al., 2008; see also SRF related related meeting report). One hypothesis in this suite of models is that decreased hippocampal GABA signaling in schizophrenia disinhibits the outputs from the structure, leading to an overactive dopamine system (Lodge et al., 2009; see SRF related meeting report and slidecast by A.A. Grace), while another posits that the hippocampal GABA deficiency can be traced to faulty input from the basolateral amygdala (Berretta et al., 2009; see SRF related news story).

This discussion was exploratory, so we rely on our readers' comments to bring up the findings you find most compelling from your own and others’ research.

View Transcript of Live Discussion — Posted 18 January 2010

View Comments By:
Tobias Bast — Posted 2 November 2009
Stephan Heckers — Posted 3 November 2009
Inna Gaisler-Salomon — Posted 15 November 2009
Anthony Grace — Posted 16 November 2009
Melanie Foecking, David Cotter — Posted 16 November 2009
Paul Harrison — Posted 18 November 2009
Segundo Mesa-Castillo — Posted 23 November 2009



Transcript

Attendees/Participants

Fernando Arana
Kristan Armstrong, Department of Psychiatry, Vanderbilt University
Verian Bader, Institute for Neuropathology, Düsseldorf, Germany
Francine Benes, McLean Hospital and Harvard Medical School
Patricia Boksa, McGill University
Jill W. Brown
Andres Buonanno, NIH/NICHD
Madelaine Burley
Michael Coleman, McLean Hospital
Ke Cui, McGill University
Howard Eichenbaum, Boston University
Angela Epshtein, Managing Editor, Schizophrenia Research Forum
Melanie Foecking, Department of Psychiatry, Royal College of Surgeons in Ireland
Inna Gaisler-Salomon, Haifa University
Hugo Geerts, In Silico Biosciences
Subroto Ghose, UT Southwestern Medical Center
Todd Girard, Psychology, Ryerson University
Tony Grace, University of Pittsburgh
Sharon Graw, University of Colorado, Denver
Louise Harvey, Douglas Centre for Mental Health Research, McGill University
Stephan Heckers, Department of Psychiatry, Vanderbilt University
Hakon Heimer, Schizophrenia Research Forum
Alena Hershfeld, University of Maryland
Szatmar Horvath, Vanderbilt University
Ming-Kuei Jang, Merck & Co.
Dale Johnson, University of Houston
Anne Kirtley, Cardiff University
Christine Konradi, Department of Pharmacology, Vanderbilt University
Aileen Kroll, Treatment Advocacy Center
Witold Lipski, University of Pittsburgh
David Luck, McGill University
Lynette Mau, Hawaii
John McGrath, Johns Hopkins University
Karoly Mirnics, Vanderbilt University
Anita Must, Department of Psychology, Vanderbilt University
Andrey Potapov
Kelley Remole, Columbia University
Scott Schobel, Department of Psychiatry Columbia University
Ornella Valenti, University of Pittsburgh
Lisa Williams, Department of Psychiatry, Vanderbilt University
Neil Woodward, Department of Psychiatry, Vanderbilt University
Austin Woolard, Department of Psychiatry, Vanderbilt University
Xiaping Xie, Center for Neuroengineering, Los Angeles

Note: Transcript has been edited for clarity and accuracy.


Hakon Heimer
I would like to introduce and thank our chat leaders, Stephan and Tony. In our usual informal spirit, and because I don’t think they are strangers to this group, I won't list their lengthy achievements, but just say that Stephan Heckers is Chair of Psychiatry at Vanderbilt University, and Tony Grace is a Professor of Neuroscience, Psychiatry, and Psychology at the University of Pittsburgh. I thank Stephan and Tony for volunteering to lead this discussion on the “seahorse of the brain” in schizophrenia. I’ll now turn the floor over to Stephan. Let’s get limbic!

Stephan Heckers
Thank you for joining us today. This is a roundtable discussion on the hippocampus in schizophrenia. I would like to start our session. Tony Grace will lead the second half of the hour. I suggest that we focus on three topics in this first segment: studies of hippocampal volume, postmortem studies of hippocampal neurons, and neuroimaging studies of hippocampal function.

Tony Grace
Hi all! In the second half, I will focus more on animal research and cellular aspects, and hopefully we can end with an integrative session.

Stephan Heckers
Let's start with the first topic: Hippocampal volume change in schizophrenia. The hippocampus is smaller in schizophrenia (when compared to matched healthy control subjects). This is a very robust finding, supported by many neuroimaging studies and confirmed by several meta-analyses. What is the value of this finding?

Tony Grace
That is certainly a robust finding, but I wonder about the etiology of this event. Is hippocampal shrinkage primary, or is it secondary to the hyperactivity leading to degeneration?

Stephan Heckers
Tony, good question. Postmortem studies cannot resolve this primary/secondary conundrum. But longitudinal imaging studies have helped.

Tony Grace
Stephan, so is it progressive? I always wondered, given the association between hippocampal hyperactivity and psychosis, if the postmortem level of loss corresponded with, e.g., the positive symptom burnout?

Scott Schobel
Stephan, it seems that, in terms of this abnormality (hippocampal brain volume reduction in schizophrenia) in the broadest sense, fulfills one clinical criteria for it being primary to schizophrenia, e.g., it's abnormal in the established illness. In terms of Tony's question (function before structure), most if not all published studies, it seems to me, either publish on functional abnormality (hyperfunctional in basal states) or structure (reduced) but do not study the two together. For this, it seems that you need longitudinal studies in high-risk states or at least in the established illness, a correlational study. Haven't seen any yet but curious if others know of data to this end. I have some data that I am working on that suggest that hyperfunction in basal states does precede volume decrease across the transition....

Tony Grace
Scott, that's great! And it gets into what I was asking, i.e., is hippocampal shrinkage and hyperactivity a chicken-and-egg thing? Or just not temporally related?

Stephan Heckers
Tony, it is very clear that the hippocampus is already smaller in first-episode patients. What is less clear is, When does the volume change occur? Some recent studies have looked at subjects who are at risk for psychosis. The findings are ambiguous.

Tony Grace
Stephan, as are all in this field! Of course, it may be that volume loss leads to hyperactivity, e.g., the neonatal hippocampal lesion model of Lipska and recent data from O'Donnell that the remaining neurons are hyperactive. Damage could lead to hyperactivity, which could engender more damage.

Anne Kirtley
Is the hippocampus uniformly smaller anterior to posterior?

Tony Grace
I am curious if someone has data related to Anne's question, Is it anterior>posterior gradient of loss?

Stephan Heckers
Anne, this has been studied repeatedly, but the findings are not clear. Some groups have found an anterior>posterior effect of volume change in schizophrenia, but several other groups have not.

Anne Kirtley
Thanks.

Neil Woodward
Stephan, how specific is volume loss to the hippocampus? Meta-analyses suggest that many brain areas are reduced in schizophrenia. Is it possible that the hippocampus is just more sensitive to a widespread deleterious effect that is in play in schizophrenia? Perhaps due to its high metabolic activity?

Tony Grace
Neil, the hippocampus is a region that is known to be sensitive to damage, but then again, this may be why it can lead to schizophrenia in susceptible individuals as well. We've thought for some time that a primary deficit may be hyperresponsivity to stress. Indeed, Eve Johnstone reported that in individuals at risk for schizophrenia, the ones who convert tend to be the ones showing hyperresponsivity to stress. On this line, we've done some work administering benzodiazepine to rats with a prenatal mitotoxin, and our preliminary work suggests we may circumvent the subsequent hyperdopaminergic state.

Stephan Heckers
Neil, you make an excellent point. Hippocampal volume change in schizophrenia occurs in the context of more widespread changes in the cortex. We need to take this into consideration as we discuss mechanisms.

Tony Grace
Stephan, I know there are lots of regional reductions, but I was under the impression that the decrease in hippocampal volume was the most robust as far as replicability across studies. Is this true?

Stephan Heckers
Tony, yes, the finding of smaller hippocampal volume has the biggest effect size, approaching one. Studying volume and activity over time is interesting, but not easy to do in human studies. Are there any data from animal experiments that speak to the relationship of activity and volume?

Tony Grace
Not yet, but it is something we are pursuing now. Just a real pain to correlate across stages. Instead, we are looking at susceptibility to stress across stages in the MAM model.

Scott Schobel
Tony, I'm wondering, in this animal work, where and when are you administering benzodiazepine? And I assume you are measuring VTA dopamine (or perhaps striatum).

Tony Grace
It was a neuroscience poster, with a small “n.” But we gave diazepam peripubertally (five days before/five days after) daily at a pretty robust dose, and measured number of DA neurons firing in the adult (which is robustly increased in all MAM-treated rats).

Stephan Heckers
Scott, what is the relationship of volume and CBV change in Alzheimer's disease? Or epilepsy? Is there a sequence? Can we learn from other "hippocampal diseases"?

Andres Buonanno
Scott, following up on Stephan's question, has reduction in hippocampal size (CA3) not been associated with predisposition to depression and post-traumatic stress disorder (PTSD)? Is there regional specificity to the loss?

Scott Schobel
Andreas, yes, reduction in hippocampal size is a characteristic of both depression and PTSD; in terms of subregions, I am not familiar with morphometric shape studies which point to particular subregions, though it seems that hippocampal size may vary according to mood state. Stephan, regarding CBV subregional dysfunction in Alzheimer's: It appears to be a basal decrease in CBV in ERC in the established illness (where volume is low) in the MCI longitudinal study going on in northern Manhattan; it appears that ERC may even be hyperfunctional prior to full-blown dementia (unpublished, should confirm with Scott Small). I think the lesson here is that there is no assumed directionality between function and structure, but it seems that perhaps excess function is not good for neuropil.

Stephan Heckers
Let's move on to the second topic in this first segment. The total number of hippocampal neurons is not reduced in schizophrenia (as might be inferred from the finding of smaller hippocampal volume in schizophrenia). The most consistent finding of recent postmortem studies of the hippocampus in schizophrenia is an abnormal number and function of hippocampal interneurons. What are the implications of this finding?

Tony Grace
Stephan, I have read reports that, although somata aren't changed, that the dendritic arbors are shrunken. Which may be consistent with interneuron loss and hyperactivity. Of course, do we know that interneurons are actually lost? We see decreased staining, but that may be due to absence of marker (from interneuron hypoactivity or decreased drive) rather than an actual loss.

Anne Kirtley
Stephan, if the number of hippocampal neurons is not reduced but the number of interneurons in hippocampus is reduced, then what neurons are increased in order to maintain a normal number of neurons?

Christine Konradi
Anne, maybe because interneurons make up only a small percentage?

Stephan Heckers
Anne, good question. Interneurons make up only 10 percent of all neurons, and Parv+ make up maybe 10 percent of all interneurons. So it is a subtle change, compared to all neurons. Furthermore, decreased expression of calcium binding proteins does not mean that neurons are lost.

Anne Kirtley
Thanks.

Andres Buonanno
For what it’s worth, in adult ErbB4 knockout mice there is an approximate reduction of 30 percent in PV-positive interneurons and a 50 percent reduction in γ oscillation power. Could these deficits in network function eventually lead to neuronal or processes loss?

Tony Grace
Andres, we found something very similar in our MAM rats: parvalbumin reduction and decreased evoked γ in both hippocampus and prefrontal cortex.

Christine Konradi
Tony, did you look at other markers as well, in addition to Parv?

Tony Grace
Christine, there is a decrease in GAD67 as well.

Christine Konradi
Tony, but what about calcium binding proteins and neuromodulators (CCK, CB, somatostatin...)?

Tony Grace
Christine, the thing is, two different things are being measured: cell bodies and markers. Markers could conceivably be decreased (e.g., by inactivity) without losing actual neurons.

Stephan Heckers
Andres, are there other mouse models that lead to a (selective) loss of Parv+ neurons?

Andres Buonanno
Stephan, I am not sure of other models. Also, I should emphasize that the loss is not limited to PV-positive IN, but also of NOS-positive. Interestingly, CCK basket cells were not reduced. I mentioned the PV-positive because of their contribution to γ power, and the reported reductions of γ power in patients.

Stephan Heckers
Andres, this is very interesting. How does ErbB4 drive Parv, but not CCK?

Andres Buonanno
Stephan, we think the answer is neurodevelopment. The CCK and PV-positive neurons are born in different locations. ErbB4 has been implicated in MGE development, where PV neurons are born. So the problem could be a developmental one where those neurons do not migrate properly and/or die off.

Stephan Heckers
Andres, very interesting!

Christine Konradi
Stephan, this also brings up another question: Is the primary deficit hyperactivity in the glutamate system or hypoactivity in the GABA system?

Tony Grace
Christine, If I had to guess, a little of both. There is evidence of decreased NMDA, which drives interneurons. The decrease in NMDA would lead to decreased interneuron function, and hyperactivity at AMPA receptors.

Stephan Heckers
Christine, yes, very good point. There is some compelling evidence that GABAergic dysfunction is in the context of (and even secondary to) NMDA receptor hypofunction.

Christine Konradi
Tony, this touches, of course, on the many ways by which people might get the disease.

Stephan Heckers
All, how do we interpret the finding that Parv+ neurons are decreased across many brain regions? Would it be special if it occurs in the hippocampus?

Tony Grace
Stephan, I thought it happened primarily in the hippocampus and PFC?

Stephan Heckers
Tony, Parv+ neuron changes occur in most cortical areas.

Scott Schobel
All, does the ratio of pyramidal to interneurons in hippocampus vs. cortex (10/1 vs. 6/1, I think) render hippocampus particularly vulnerable vs. cortex to interneuron deficit?

Stephan Heckers
Scott, definitely. It is one of the unique features of the hippocampus. Francine, how do you interpret the decreased number of Parv-positive neurons?

Francine Benes
Stephan, it is difficult to interpret a reduction in the numerical density of PVB+ interneurons, if there are no associated cell counting data from Nissl-stained sections. If there is a decreased density of interneurons in the hippocampus as we reported in 1998 (i.e., selectively in sectors CA3/2), then it is likely that there may be an actual loss of these cells. The reduction at that locus in our study was approximately 40-50 percent, but it was present in both schizophrenics and bipolars. In any case, those data were consistent with a loss, although it must be said that it is probably not specific to schizophrenia.

Stephan Heckers
Francine, thank you. We have some new data that show that some of the Parv+ neuron changes are in the context of cell loss, but in some regions they are not. Of interest, in CA2/3 we find clear decrease of Parv+ neurons, but no cell loss.

Francine Benes
Stephen, that is very interesting indeed. But, it is not surprising that in some subregions of the hippocampus that there might be subpopulations of interneurons that undergo a reduction in the synthesis or an increase in the degradation of their calcium binding peptides, and that this might seem like a loss of such cells, but really isn't.

Tony Grace
Francine, I think that your work on the amygdala-hippocampus really draws together the stress aspect and the interneuron loss nicely.

Todd Girard
All, since we're returning to regional specificity, although gross hippocampus effects are more robust, what is the consensus on regional effects? I thought that the postmortem results were indicative on the whole that CA1 tends to be less affected (which would be very unique to schizophrenia). In contrast, recent imaging and physiological models implicate CA1 hyperactivation most strongly. Could this be an example of the interplay between regional loss of inhibition leading to disinhibition in another region? This would seem to require another bridge beyond interneurons per se (?). Or maybe the regional effects are much less robust than I'm thinking....

Stephan Heckers
Todd, maybe. There is clearly a disconnect between studies that report CA1 changes (Schobel et al.) and those that find changes primarily in CA2/3 (Benes et al.).

Francine Benes
Well, if GABAergic activity is reduced, then this would likely lead to an increase of excitatory output from pyramidal neurons within the hippocampus. We did a whole cell recording study in our rodent model of schizophrenia based solely on postmortem findings, and in sectors CA3/2 where GABA cells are lost in the model, pyramidal neurons show a decrease of IPSCs.

Andres Buonanno
Francine, recent studies from Mody's lab would suggest that it is difficult to interpret directly the consequences of loss of inhibition in any one cell type, and unexpected results are seen at circuit level.

Stephan Heckers
Scott, how likely is it that CBV changes are present also in other CA sectors (not just CA1) and also in anterior and posterior regions?

Scott Schobel
Stephan, I think it is likely that CBV changes will be found in other hippocampal subregions with larger samples. I think the key concept here is differential vulnerability, according to the properties of the trisynaptic circuit. I would predict that since CA1 and subiculum are downstream in this circuit, changes will be maximally present in these subregions, but others may be affected as well, though less so. In terms of an anterior-posterior gradient of findings, volumetric and shape studies have evidence for both, with reductions along the anterior-posterior axis, depending upon the study, but it seems to me that there is more recent evidence for differential vulnerability in the anterior aspect. This is appealing to hypothesize b/c anterior CA1/subiculum monosynaptically connect to PFC (OFC), providing a direct link bet. Hippocampal changes and PFC dysfunction.

Tony Grace
I think one has to take into consideration that the hippocampus is organized in a sequential manner, with the final common output being in the subiculum (anterior hippocampus in humans). So hyperactivity in any of these regions could propagate. But that being said, the hippocampus tends to function more on rhythmic activity than just firing rates. And you need interneurons for rhythmic activity. So loss of interneurons would lead to disruption of rhythms (and functional loss) as well as hyperactivity, replacing gated information with higher frequency noise. This could even contribute to the prefrontal deficit via subicular-prefrontal circuits.

Stephan Heckers
Tony, do you see the anterior hippocampus as the final output station?

Tony Grace
Stephan, I see it as a major output, and the primary output to the limbic system/PFC. I wonder if hippocampal interneuron loss and hyperactivity leads to PFC interneuron loss and dysfunction, or if PFC dysfunction prevents PFC suppression of stressors, driving stress-induced hippocampal loss?

Scott Schobel
Tony, to this point (output and what comes first, PFC vs. hippocampus), I think that prodromal studies following high-risk individuals over time in longitudinal designs will help answer this question. So far, not enough evidence. Also, it seems worth noting that outcomes we are seeing in the COPE clinic are diverse and include both mood spectrum (bipolar 1, SAD bp type) and schizophrenia.

Stephan Heckers
All, what is the evidence that hippocampal pathology in schizophrenia is due to environmental stressors?

Jill W. Brown
Anyone, is there evidence of immune system dysfunction playing a role in the disease or in neuron damage?

Francine Benes
Jill, this is not a question that can be answered in a straightforward way. It is well-established that necrotic cell death is not present in schizophrenia. It is also important to note that we have not found any evidence of DNA damage in ACCx or HIPP of subjects with schizophrenia. This is a common change associated with oxidative stress. Neuronal damage can occur as a result of at least two other mechanisms. The most obvious way is through apoptosis; however, there is not compelling evidence for such a process occurring in schizophrenia. On the other hand, a decrease in interneuron numbers does not necessarily imply that GABA cells have been damaged. Rather, during early embryogenesis, when the cortex is developing, it is possible that a migration of interneuron precursor cells toward the cortical mantle may not proceed normally in schizophrenia. This could result in a reduced number of postmitotic, differentiated GABA cells at key loci such as sector CA3/2 of the hippocampus. Empiric evidence for a reduction of interneurons at this locus has come from a two-dimensional study in my laboratory (135394 Benes et al., 1998) and a 3D study from Stephan Heckers’s lab. Postmortem studies in which molecular approaches have been employed have suggested that an epigenetic mechanism involving DNMT1 in dorsolateral prefrontal cortex (see Costa et al., 2004). In the hippocampus, we have suggested that a similar mechanism is operating within GABA cells of CA3/2, but involves HDAC1 and DAXX suppression of promoter activity (Benes et al., 2007). Taken together, it seems that we should not be thinking in terms of degeneration, but rather in terms of complex cellular and molecular changes that render GABA cells dysfunctional. This perspective offers greater hope for finding specific new therapies for schizophrenia.

Tony Grace
Stephan, I don't think there is direct evidence. But Sapolsky, McEwen, and others have evidence that maintained stressors/glucocorticoids lead to hippocampal damage in terms of apical dendrite shrinkage (loss of neuropil). Stress is a major risk factor in schizophrenia, and Eve Johnstone reported that in children at genetic risk, those showing hyperresponsivity to stress tend to be those that convert.

Christine Konradi
Tony, sounds a lot like bipolar disorder to me ;-) but then again, there is a lot of overlap, I guess.

Tony Grace
Christine, I think you're right; it may be a similar etiology, with the pathophysiology dependent on timing?

Karoly Mirnics
I agree with Tony, no direct evidence. The fact that we can mimic certain deficits with environment does not mean that you do not need a genetic predisposition and GxE interplay to develop the disease-related pathology.

Tony Grace
Karoly, that is worth emphasizing. There is still more sporadic than genetically linked schizophrenia from my understanding; it’s just easier to look at genetically linked.

Karoly Mirnics
Good point, Tony. I think that the whole point of 50 percent genetics and 50 percent environment is faulty at the level of individual patient; the GxE contribution is likely to be very different from case to case.

Neil Woodward
Heritability of hippocampus volume is rather low (.40-.60) compared to cortex (.70-.95), suggesting that genetic effects might be more modest in hippocampus. On the other hand, the hippocampus may be more sensitive to the interface between environment and genetics.

Francine Benes
Stress generally affects the CA1 subregion preferentially. However, kainic acid tends to produce excitotoxic injury in sectors CA3/2 at the earliest time point, although it eventually extends to CA1 as well. The expression of kainate receptor subunits is particularly abnormal in sectors CA3/2 of schizophrenia and bipolar disorders, so this receptor system may play an important role in at least some subregional pathology.

Hakon Heimer
Tony, about 10 minutes to go (officially, though we have no real time limit on the room). Do you have any additional threads that you'd like to get going?

Tony Grace
Hakon, I think we're really hitting on some important points now!

Stephan Heckers
Francine, this is important. There might be more than one mechanism for hippocampal pathology in schizophrenia, some leading to CA1 changes and some to CA2/3 changes.

Francine Benes
I'm sorry, but I must go. It has been a very interesting discussion.

Tony Grace
Francine, but have these studies really extended into the subiculum? It seems that many of the hippocampologists stop just dorsal to subiculum.

Francine Benes
Tony, you are quite correct in pointing this out. Microscopic studies or molecular studies that employ laser microdissection are generally very labor intensive, making it necessary to set a point at which the analyses end. The hippocampal sectors most broadly studied are sectors CA1 and CA3. They both have a distinctive cyto-architectural arrangement that makes it possible to study specific aspects of the trisynaptic pathway. Layers rich in GABAergic interneurons (i.e., stratum oriens and stratum radiatum) are particularly well delineated, making it relatively straightforward for researchers to investigate these neuronal subpopulations. In the subiculum, the projection neurons and interneurons are mixed together, as they are in neocortical layers. The subiculum lacks a laminar arrangement with clear implications for connectivity with other brain regions.

Tony Grace
We have results showing a powerful drive of subiculum from amygdala and from the locus coeruleus to the subiculum, although the NE drive may be indirect through other regions.

Stephan Heckers
Tony, you are right. There are very few studies of the subiculum! How do the subiculum changes affect DA function (in VTA or accumbens)?

Tony Grace
We showed in our paper that driving the subiculum increases the population activity (i.e., number of neurons firing) of VTA DA neurons, and this causes an increase in the gain of the DA system. This is also responsible for amphetamine hyperactivity, since if we inactivate the subiculum of MAM rats, the hyperresponsivity disappears. We think this is why hippocampal hyperactivity correlates with psychosis. In fact, stress and the MAM model of schizophrenia seem to work via common pathways, increased hippocampal (i.e., subiculum) influence over the VTA.

Stephan Heckers
Tony, this is very important. If this can be confirmed with imaging studies (looking at DA release), it would make for a very compelling story.

Neil Woodward
Tony, how close is the correspondence between hippocampus dysfunction and DA hyperactivity? Is it restricted to ventral striatum?

Tony Grace
Neil, there is a correspondence between hippocampal hyperactivity and psychosis in human schizophrenia. But we showed that activation of the subiculum will drive VTA DA neuron population activity, increasing its response to stimuli. As far as parts of the striatum, the hippocampus drives the ventral striatum, which drives the VTA, but Haber's work suggests that this will drive DA neurons that project both to ventral striatum and to integrative striatum, which is where Laruelle sees the biggest change in schizophrenia DA release.

Neil Woodward
Thanks, Tony!

Inna Gaisler-Salomon
All, are there functional imaging studies pointing to CA2/3 dysfunction, or only postmortem?

Stephan Heckers
Inna, I am not aware of selective CA2/3 findings in imaging studies. But there are limits to the resolution of our methods....

Scott Schobel
Inna, I agree with Stephan but need more studies with larger numbers to definitively say. I can say that we have extended our schizophrenia sample and still see CA1 specificity with n = 32 patients and n = 24 controls comparable in age/gender. Med-free progressors we have followed longitudinally show increases in subiculum, CA1 is static from baseline to follow-up (relatively elevated), CA3 and DG appear to slightly increase (all unpublished but presenting at ACNP).

Stephan Heckers
I also want to introduce the topic of hippocampal function in schizophrenia. Neuroimaging studies have revealed abnormal hippocampal activity at rest and decreased recruitment during the performance of cognitive tasks in schizophrenia. What is the implication of hippocampal dysfunction for models of cognition and psychosis in schizophrenia?

Howard Eichenbaum
Stephan, the findings on loss of inhibitory cells in CA2/3 and potentially consequent hyperactivity is similar to a situation we observe in place cells in aged, memory-impaired rats. It may be that hyperactivity of principal cells leads to increased pattern completion rather than pattern separation by neural ensembles, and that could lead to poor encoding of new information and instead a tendency to see connections between loosely related memories and incoming information. This is a lot of speculation....

Stephan Heckers
Howard, which stage of memory formation would be more affected in a CA2/3 model and which one in a CA1 model?

Howard Eichenbaum
CA2/3 would lead to the pattern separation disorder, CA1 might lead to a disorder in temporal context processing, but the two would surely interact.

Stephan Heckers
Howard, this is very helpful. Does this lead to strong prediction for cognitive studies in patients?

Howard Eichenbaum
Craig Stark's work in distinguishing CA3 from CA1 is impressive; this high-resolution work might address that issue.

Stephan Heckers
Howard, agreed.

Neil Woodward
Tony, Stephan, what are the implications for neuroimaging with regard to hippocampus rhythmic activity? This is very important for cognitive imaging studies because many patients have trouble performing the tasks we use. We end up trying to figure out hippocampus dysfunction in patients with relatively good cognitive performance—a minority of patients. Could resting state connectivity shed light on hippocampus dysfunction?

Stephan Heckers
Neil, great point.

Hakon Heimer
All, we're about five minutes from the time when some of you may have to go to lunch, dinner, breakfast (depending on location). This would be a good time for sweeping conclusion statements or suggestions for research priorities.

Tony Grace
Hippocampus = important.

Stephan Heckers
Hakon, the next phase of the hippocampus + schizophrenia research will move to an integration of our structural and functional findings.

Tony Grace
And longitudinally as well, this will help to segregate cause and effect.

Scott Schobel
Thank you all for the interesting discussion. Need to go. Kind regards, Scott. P.S. Testing hippocampal basal subregional function with cognitive paradigms seems like a priority in addition to symptoms.

Stephan Heckers
Hakon, thank you for hosting this roundtable. It was fun!

Hakon Heimer
Stephan, you're welcome!

Tony Grace
Hakon, yes, a lot of good information and ideas!

Andres Buonanno
Thank you all for a wonderful chat, especially to Tony and Stephen. Very stimulating!

Angela Epshtein
Thank you, all. Bye.

Hakon Heimer
Thanks to Stephan and Tony for leading the discussion! ::wine

Howard Eichenbaum
Indeed thanks Tony and Stephan. Very helpful!

Tony Grace
We should get together over some beverages of our choice at ACNP to talk about this further!

Stephan Heckers
Thanks to all!::big grin.

Karoly Mirnics
Thanks, guys, it was fun.

Tony Grace
Bye, everyone, and thanks for participating! ::beer

Melanie Foecking
Thank you, all, especially Stephan and Tony. Great discussion. Until next time, hopefully.

Stephan Heckers
Tony, cheers ::beer ::beer ::beer.

Tony Grace
And thanks to you, Stephan!

Hakon Heimer
That's a lot of beer. Don't drink and drive!

Stephan Heckers
Hakon, virtual beer, what a great concept!

Hakon Heimer
You don't gain weight, so your mile time will not go down!

Karoly Mirnics
I am not a fan of such virtual reality. Prefer the genuine stuff. :-)

Anne Kirtley
Thank you for organizing this, SRF. Will help with the thesis!

Hakon Heimer
Bye, all.

Comments on Online Discussion
Comment by:  Tobias Bast
Submitted 2 November 2009
Posted 2 November 2009

Perhaps the issue of functional differentiation along the longitudinal axis of the hippocampus may be considered?

Neuroanatomical and functional evidence (e.g., Bannerman et al., 2004; Bast, 2007; Bast and Feldon, 2003; Bast et al., 2009; Moser and Moser, 1998; Small, 2002) suggests that the septal to intermediate hippocampus, interacting with the dorsolateral band of the entorhinal cortex, plays a key role in the rapid encoding and subsequent retrieval of accurate visuospatial information; in contrast, the temporal to intermediate hippocampus is linked to behavioral control processes (including sensorimotor, emotional, motivational, and executive functions) by way of connectivity to prefrontal cortex and subcortical sites and the ascending dopamine systems.

Accordingly, the septo-temporal level of hippocampal overactivity may be a critical determinant of the behavioral dysfunction resulting from hippocampal overactivity. The recent study by Schobel et al. (2009), showing that overactivity of mainly anterior (i.e., temporal to intermediate) hippocampus (CA1) is associated with psychotic symptoms, offers support for this view.

References:

Bannerman DM, Rawlins JN, McHugh SB, Deacon RM, Yee BK, Bast T, Zhang W-N, Pothuizen HHJ, Feldon J. Regional dissociations within the hippocampus--memory and anxiety. Neurosci Biobehav Rev. 2004;28:273-83. Abstract

Bast T. Toward an integrative perspective on hippocampal function: from the rapid encoding of experience to adaptive behavior. Rev Neurosci. 2007;18:253-81. Abstract

Bast T, Feldon J. Hippocampal modulation of sensorimotor processes. Prog Neurobiol. 2003;70:319-45. Abstract

Bast T, Wilson IA, Witter MP, Morris RGM. From rapid place learning to behavioral performance – a key role for the intermediate hippocampus. PLoS Biol. 2009;7(4):e1000089. Abstract

Moser MB, Moser EI. Functional differentiation in the hippocampus. Hippocampus. 1998;8:608-19. Abstract

Schobel SA, Lewandoski NM, Corcoran CM, Moore H, Brown T, Malaspina D, Small SA. Differential targeting of the CA1 subfield of the hippocampal formation by schizophrenia and related psychotic disorders. Arch Gen Psych. 2009;66:938-46. Abstract

Small SA. The longitudinal axis of the hippocampal formation: its anatomy, circuitry, and role in cognitive function. Rev Neurosci. 2002;13:183-94. Abstract

View all comments by Tobias BastComment by:  Stephan Heckers, SRF Advisor
Submitted 3 November 2009
Posted 3 November 2009

This is a very good suggestion. We will review the evidence for regionally selective changes in the hippocampus in schizophrenia, including CA1 versus CA2/3 versus CA4 and anterior versus posterior hippocampus.

View all comments by Stephan HeckersComment by:  Inna Gaisler-Salomon
Submitted 15 November 2009
Posted 15 November 2009

As correctly pointed out by Tobias Bast, the hippocampus should no longer be viewed as a homogenous unit, and its subregions are increasingly positioning themselves as individual, yet interconnected, structures. The question that I find fascinating is, What is it about a particular subregion or tract that differentiates it from the others and makes it more vulnerable to schizophrenia-related symptoms?

A study by Schobel et al. from Scott Small’s group at Columbia, recently published in the Archives of Psychiatry (Schobel et al., 2009), was highly informative in singling out the CA1 subregion, hyperactive at baseline in both symptomatic schizophrenia patients and in prodromal subjects. The CA1 subregion is unique in its role as an integrator of contextual information, and in its ties to VTA neurons (Lisman and Otmakhova, 2001).

In Stephen Rayport’s lab at Columbia, we have recently applied the unique imaging technique used by Schobel et al. in humans to mice (Gaisler-Salomon et al., 2009a), and found that mice genetically engineered to express less glutaminase, the enzyme responsible for recycling glutamate from glutamine in the brain, have a hypoactive CA1 and a markedly reduced response—both behavioral and neurochemical—to amphetamine (Gaisler-Salomon et al., 2009b). By comparison, a hyperactive hippocampus in MAM-treated animals, a valid and widely used animal model of schizophrenia, was found to be accountable for an increased behavioral response to amphetamine and for enhanced DA neuronal population activity in the VTA (Lodge and Grace, 2007).

In integrating these findings, it seems plausible that dysregulated hippocampal activity, particularly in CA1, is responsible for the ill-functioning DA system in schizophrenia. Since glutaminase-deficient mice display phenotypes that are inverse to what was observed in MAM-treated animals and other schizophrenia animal models, and since hippocampal glutamate was reduced in our mice, our recent findings may provide clues to the question of “what’s so special in CA1/”—perhaps it is the high dependence of this structure on intact, stable glutamate provision (from the glutamate-glutamine recycling pathway or other sources) that makes it special, but also more vulnerable to schizophrenia-related pathology.

References:

Gaisler-Salomon I, Schobel SA, Small SA, Rayport S (2009a) How high-resolution basal-state functional imaging can guide the development of new pharmacotherapies for schizophrenia. Schizophr Bull 35:1037-1044. Abstract

Gaisler-Salomon I, Miller GM, Chuhma N, Lee S, Zhang H, Ghoddoussi F, Lewandowski N, Fairhurst S, Wang Y, Conjard-Duplany A, Masson J, Balsam P, Hen R, Arancio O, Galloway MP, Moore HM, Small SA, Rayport S (2009b) Glutaminase-deficient mice display hippocampal hypoactivity, insensitivity to pro-psychotic drugs and potentiated latent inhibition: relevance to schizophrenia. Neuropsychopharmacology 34:2305-2322. Abstract

Lisman JE, Otmakhova NA (2001) Storage, recall, and novelty detection of sequences by the hippocampus: elaborating on the SOCRATIC model to account for normal and aberrant effects of dopamine. Hippocampus 11:551-568. Abstract

Lodge DJ, Grace AA (2007) Aberrant hippocampal activity underlies the dopamine dysregulation in an animal model of schizophrenia. J Neurosci 27:11424-11430. Abstract

Schobel SA, Lewandowski NM, Corcoran CM, Moore H, Brown T, Malaspina D, Small SA (2009) Differential targeting of the CA1 subfield of the hippocampal formation by schizophrenia and related psychotic disorders. Arch Gen Psychiatry 66:938-946. Abstract

View all comments by Inna Gaisler-SalomonComment by:  Anthony Grace, SRF Advisor (Disclosure)
Submitted 16 November 2009
Posted 16 November 2009

Actually, these are all good points. But as with most complex systems, the parts do not operate in isolation. If one looks at both the anatomical and functional connectivity of the hippocampus, I have come to the opinion (right or wrong) that the anatomical connectivity can be interpreted on the basis of the functions. I.e., the more dorsal systems (in the rat) or more CA1-like are involved with locating one's position in space, and have the appropriate sensory inputs for such a function. But as one moves ventrally (in the rat, or anterior in the human), there are more limbic inputs, which means a translation from "where am I" to "what is the emotional salience of where I am?" This latter concept in my opinion fits in well with what we term as context.

View all comments by Anthony GraceComment by:  Melanie FoeckingDavid Cotter
Submitted 16 November 2009
Posted 16 November 2009

In addition to the really interesting findings in CA1, I would like to contribute recent unpublished (in preparation) findings we obtained from a proteomics investigation. We followed the hypothesis that the different subregions have distinct roles in regulation of hippocampal circuitry, and alterations within them are likely to contribute in a primary way to the clinical presentation of schizophrenia.

We aimed to characterize the differential protein expression in each of the hippocampal subregions in schizophrenia (n = 20) and bipolar disorder (n = 20) compared to control samples (n = 20). We used laser-assisted microdissection, and Difference-in-Gel-Electrophoresis to enrich for these tissues and to compare protein profiles. Samples were grouped according to the different disease/control , and we found 213 spots to be differentially expressed between disease groups in the different hippocampal subregions. Extensive statistical analysis was undertaken to correct for possible confound (pH of the brains, postmortem interval, and drug treatments).

Differential expression in the subregions was observed as follows:
CA4: 32 protein spots (SCZ), 30 protein spots (BPD).
Dentate: 23 spots (SCZ), 26 spots (BPD)
CA1: 33 spots (SCZ), 34 spots (BPD).
CA2/3: 53 spots (SCZ), 113 spots (BPD)

The number of differentially expressed spots is higher for CA2/3 (chi square; p <0.001), implicating this region particularly in schizophrenia.

Identification of differentially expressed proteins by mass spectrometry revealed proteins that are implicated in a range of different processes, e.g., cytoskeletal, synaptic, and metabolic changes.

In the CA2/3 hippocampal subregion, changes were observed in proteins which are indicative of clathrin-mediated endocytosis, cell migration, and apoptosis.

View all comments by Melanie Foecking
View all comments by David CotterComment by:  Paul Harrison
Submitted 18 November 2009
Posted 18 November 2009

Sorry I missed the discussion, and I look forward to seeing the transcript.

The Schobel et al. study is fascinating and rightly speaks to heterogeneity within the hippocampus in schizophrenia (which likely applies also to cell populations, strata, anteorposterior gradients, etc.). Two oldish findings may be relevant to the issue.

First, we measured total (polyadenylated) mRNA content in the different hippocampal subfields in (chronic) schizophrenia and control subjects (Harrison et al., 1997). This index provides a measure of “total” gene expression and may be seen as a molecular correlate of cellular activity. Although not significant (in a small sample), there was an increase in mRNA signal in CA1 in schizophrenia, in contrast to the unchanged or decreased levels in other hippocampal and parahippocampal regions. This finding supports a relative “hypermetabolism,” and a difference, between CA1 and other subfields, in the disorder. (In passing, hyperactivity of CA1 neurons underlying paranoid schizophrenia was hypothesized in 1992; Krieckhaus et al., 1992.)

Second, a while ago we summarized the quantitative studies of hippocampal glutamate receptors and found evidence that, in terms of the magnitude of changes, CA1 was “less” affected than other subfields (Harrison et al., 2003). We drew a similar conclusion for studies of synaptic proteins (Harrison and Eastwood, 2001).

The findings together suggest that the heterogeneity of hippocampal involvement in schizophrenia is itself heterogeneous, depending on what is being measured, and how it is affected by the usual range of clinical and other variables.

References:

Harrison PJ, Eastwood SL (2001) Neuropathological studies of synaptic connectivity in the hippocampal formation in schizophrenia. Hippocampus 11: 508-519. Abstract

Harrison PJ, Burnet PWJ, Falkai P, Bogerts B, Eastwood SL (1997) Gene expression and neuronal activity in schizophrenia: a study of polyadenylated mRNA in the hippocampal formation and cerebral cortex. Schizophrenia Research 26: 93-102. Abstract

Harrison PJ, Law AJ, Eastwood SL (2003) Glutamate receptors and transporters in the hippocampus in schizophrenia. Annals of the New York Academy of Sciences 1003: 94-101. Abstract

Krieckhaus EE, Donahoe JW, Morgan MA (1992) Paranoid schizophrenia may be caused by dopamine hyperactivity of CA1 hippocampus. Biological Psychiatry 31: 560-570. Abstract

View all comments by Paul HarrisonComment by:  Segundo Mesa-Castillo
Submitted 23 November 2009
Posted 23 November 2009

Use of new imaging technology has enabled researchers to attempt to identify regions of interest that might be related to schizophrenia. Thus far, results obtained (Shenton et al., 2001; Joyal et al., 2003; Bogerts, 1999; Mesa, 2001) indicate that the left temporal lobe is among the regions of interest; in particular, the medial structures appear to have a decrease in the volume of the amygdala, hippocampus, and parahippocampal gyrus. This suggests that exploration at the cellular level by means of other techniques such as electron microscopy could glean useful data. There is one postmortem study (Mesa, 2005) in which electron microscopic techniques were successfully implemented in studying three out of four of the regions of interest in schizophrenia research. Those areas included the primary auditory cortex, the amygdala, and the hippocampus of the left cerebral hemisphere.

References:

Shenton ME, Dickey CC, Frumin M, McCarley RW. A review of MRI findings in schizophrenia. Schizophr Res 2001; 49: 1–52. Abstract

Joyal CC, Laakso MP, Tiihonen J, Syvalahti E, Vilkman H, Laakso A, et al. Schizophrenia: A Volumetric Magnetic Resonance Imaging Study in First Episope Neuroleptic-Naive Patients. Biological Psychiatry 2003; 54: 1302-4. Abstract

Bogerts B. The neuropathology of schizophrenic diseases: historical aspects and present knowledge. Eur Arch Psychiatry Clin Neurosci. 1999; 249 Suppl 4:2-13. Abstract

Mesa CS. Regiones de Interés del Cerebro en la Esquizofrenia. Estudios Post-mortem con Microscopía Electrónica de la Amígdala, el Hipocampo y la Corteza Auditiva Primaria. WPA Educational Section, Junio 2005.

Mesa CS. An ultrastructural study of the temporal lobe and peripheral blood in schizophrenic patients. Rev Neurol 2001; 33: 619-23.

View all comments by Segundo Mesa-Castillo