|Home | About | Journals | Submit | Contact Us | Français|
The development of the hippocampus and the regulation of its innervation by growth factors has been a particular interest of Karolinska Institute investigators (Ayer-LeLievre et al., 1998). The intraocular transplant preparation allowed isolation of portions of the hippocampus for detailed developmental study (Goldowitz et al., 1982). A particularly advantageous feature of the preparation was its facility for the study of adrenergic and cholinergic innervation, first from the host animal’s peripheral autonomic nervous system and, as the technique matured, from co-grafted brainstem and midbrain neurons (Olson et al., 1980; Goldowitz et al., 1984a; 1984b). Professor Lars Olson and his long-standing colleagues Professor Åke Seiger and Dr. Barry Hoffer have always been eager to have the Institute’s investigation become translated into clinical research that might improve patients’ treatment. The intraocular transplantation of human tissue included samples derived from the fetuses of woman who had schizophrenia, to study abnormalities in the growth and development of their hippocampus (Freedman et al., 1994). Basic studies of the hippocampus conducted at the Institute during this period have over the past decade led to the initiation of new treatment strategies for schizophrenia (Freedman et al., 1993; 2008). In this chapter we review the progress of using the in oculo method to understand some key features of hippocampal development and the application to the clinical setting.
A time-honored approach to the understanding the intrinsic (genetic) and extrinsic (environmental, epigenetic) mechanisms that underlie development of the CNS is to isolate the tissue of interest from its neighbors and place it into a controlled environment. At the Karolinska Inst., Dr. Lars Olson had revived a seldom-used technique, the transplantation of tissue to the anterior chamber of the eye, to bring a fresh approach to this type of experimental questions. This method had much to recommend itself over other approaches such as putting tissue under the kidney capsule or into cell culture, where either tissue was not readily visible and accessible for further manipulation or the conditions were not under physiologic conditions, respectively. The plusses of the technique were myriad including the nurturing environment of the free epithelial surface of the iris, the availability of three sources of innervation, the ability to visually monitor transplants, and finally the accessibility to manipulate the environs of the transplanted tissue. In a step-wise fashion, we examined some of the basic principles that served in the development of the area dentate of the hippocampal formation with particular attention to the afferentation of this structure.
First, we examined the developmental potential of the dentate and adjacent hippocampus (termed area dentata) isolated from its extrinsic inputs. The histotypic development of these transplants was quite impressive, forming from an initially, loosely organized neuroepithelium of the embryonic day 18 rat dentate anlagen to a structure with a well-differentiated dorsal and ventral limb (Figure 1). Perhaps more surprising was the normal appearance of the granule cell dendritic tree in most cases examined. These findings, along with the presence of normal mossy fiber projections and synapses at the light and electron microscopic level, spoke to the transplant containing all of the relevant cues important for the development of several key aspects of the dentate. Of note, in all transplants we had also included cells that gave rise to a partial regio inferior. Whether these cells are important for the relatively normal development of the dentate in oculo, presumably via the associational connections to the dentate granule cells, is something that could be critically tested by transplanting more discretely dissected regions of the dentate. Some years later we had done a lineage analysis of the dentate in the mouse (Martin et al, 2002), and it would be interesting to know if the clonal organization of the dentate was similar in oculo as in vivo. Another interesting point to examine how the transplanted dentate affects the adult neurogenic population in the sub-granular zone that has been an area of intense investigation in recent times. The in oculo system would offer a setting to explore how various experimental conditions might alter adult neurogenesis.
One of the advantages of the anterior eye chamber as a model system is that multiple transplants can be placed on the iris surface and the interaction between tissues can be examined. In this manner, we explored how the prior presence of a transplanted region that normally innervated the dentate in vivo, the locus coeruleus, septum, or entorhinal cortex, might alter dentate development and provide the dentate with afferent fibers. It was being appreciated in the Olson lab, from work by a student Hakan Björklund, that the cerebral cortex was very responsive to the presence of a juxtaposed locus coeruleus (Björklund et al., 1983). In distinction to the cortex, the growth trajectory and cytoarchitecture of the dentate graft were not significantly altered by mimicking the in vivo situation, even though each co-transplant supplied nerve fibers to the dentate (Figure 2). The catecholaminergic fibers of the locus coeruleus supplied the most impressive innervation of the dentate. Given a joint survival period of a few months and even a limited number of noradrenergic neurons, the locus richly blanketed the dentate with nerve fibers (Figure 3). This was not the case with the two major inputs to the dentate, the septum and the entorhinal cortex. The entorhinal innervation was histotypic (to the outer part of the dentate granule cell receptive field) but extremely limited to just the region where the transplants touched. On the other hand, input from cholinergic neurons of the septum was robust but limited to specific regions of the dentate transplant (Figure 4). This suggested a competitive interaction within the dentate neuropil. One of the major intrinsic fiber systems is the mossy fibers that are the afferents of the dentate granule cells and demonstrable with the Timm’s stain for heavy metals. The examination of neighboring sections with acetylocholinesterase histochemistry and Timm’s stain revealed a exclusionary relationship between these fiber systems (Figure 4).
This interesting relationship between dentate and septum motivated the examination of how extrinsic inputs to the transplanted dentate might modulate the extrinsic sources of dentate innervation, in particular the hyperinnervation of dentate by the locus coeruleus. To test this, triple transplants were made in the eye between locus coeruleus and dentate and either the entorhinal cortex or septum. The schematic in Figure 5 summarizes the result; while entorhinal cortex had no effect on locus innervation, the presence of a neighboring or intervening septal transplant markedly dampened the noradrenergic innervation of the dentate (Figure 6).
The in oculo system has provided unique glimpses into the regulation of nerve fiber ingrowth into the dentate, although the molecular mechanisms that serve these phenomena are yet to be discovered. More recent work in the Granholm lab has begun to molecular determinants of hippocampal development using the in oculo approach (Willis et al., 2005; Nagy et al., 1998). It would be interesting to explore if the competitive interaction between afferent systems in the transplanted hippocampal formation might underlie the “input dysfunction” that is discussed as a basis for the schizophrenic condition discussed by Freedman in the second part of this chapter. The impressive evidence for the role of the α7-nicotinic receptor in schizophrenia is intriguing relative to the in oculo findings that the septum’s cholinergic input is a determinant in the afferentation of the hippocampal formation. The in oculo paradigm would appear to have applicability to current investigations into the cellular and molecular determinants of hippocampal development and function.
Early observations of how people with schizophrenia react to their environment noted a peculiarity in the ability of persons with schizophrenia to appear unaware of the environment and yet overly responsive to it. Eugen Bleuler, who first used the term schizophrenia, also first developed the concept of an attentional dysfunction in schizophrenia (1911). The most dramatic example of the phenomenon of seeming to ignore the environment is catatonia, a rarely seen syndrome in schizophrenia today. The patient gradually stops responding to stimuli and then eventually stops moving altogether. Patients with catatonia have hyperactive electroencephalographic activity, consistent with the minds being quite active, rather than asleep or anesthetized. They respond to barbiturates with a paradoxical “awakening”, in which they resume normal movement. This response suggests that they have actively withdrawn from the world around them, which they sometimes later report was to try to block out the noise around them. When the barbiturate partially inhibits their brain’s responsiveness, they lose their withdrawal and temporarily resume normal interaction. They often report that they were fully aware, indeed acutely hyper-aware, of their surroundings during the catatonia and repeat in detail the conversations that their care-givers had about them.
During World War II, Britain called on experimental psychologists to help improve the ability of plane spotters to identify aircraft quickly and at great distance, when features might be difficult to discern. Donald Broadbent (1958) learned that there were two quite separate aspects to the task. The first was a detection problem, to be able to see the plane when it was just a dot in the clouds. The second was a classification problem, to be able to check identifying features to determine what country the aircraft that made the dot represented. People concentrating on detection were poor at classification and, vice versa, those who concentrated on classification were poor at detecting aircraft early. Broadbent hypothesized that these were two separate features of perception and that people should be trained to do each independently.
Broadbent saw the two different types of tasks as distinct steps in the earliest stages of perception. He reasoned that the part of the brain that makes decisions, such as which plane is which, has limited capacity and that this part of the brain needed mechanisms to prevent itself from being overwhelmed with stimuli. If the brain were always trying to process every stimulus fully, its decision-making capacity would be quickly overwhelmed and greatly slowed. One problem is how to decide prior to full processing which stimuli are important enough to be processed. Broadbent proposed that the primary filter needed to have a very simple principle. It did not need to eliminate all extraneous stimuli, but it needed to eliminate a substantial fraction of them, with little processing requirement. He proposed that if stimuli could be processed briefly to determine if they were being repeated, then the filter could simply eliminate those that were repetitive. Repeated information would generally have little significance to most goals of the organism, and hence the brain could safely ignore repeated stimuli.
Peter Venables realized that Broadbent’s model could be applied to understanding the situation of patients with schizophrenia. They were being flooded by stimuli because of the absence of a fully functional filtering mechanism. He termed their condition an input dysfunction (1964). More commonly, we use the term sensory gating deficit. The control of stimuli reaching the brain’s processing centers was lost, as Bleuler had earlier proposed. Patients might then be acutely psychotic and respond to all stimuli, including their internal thoughts, or, more chronically they might withdraw from stimuli, either voluntarily by being isolative or involuntarily through catatonia, to escape this flood of information.
We developed an electrophysiological demonstration of the hypothesized sensory gating deficit by presenting two auditory click stimuli to patients and recording their cerebral averaged evoked response using standard EEG technology. The normal response, shown by Subject A in Figure 7, is to have a larger evoked response to the first stimulus than to the second. The amplitude of the P50 wave, a positive wave occurring approximately 50 msec after the stimulus, was used to quantify the response. A lower ratio of the amplitude of the second to the first response indicates that there was significant inhibitory sensory gating. Normal subjects generally have ratios ≤ 0.40; most persons with schizophrenia exceed this value (Adler et al., 1982; Freedman et al., 1997).
This deficit in sensory gating schizophrenia has been replicated by Judd et al., 1992; Cadenhead et al., 2000; Thoma et al., 2003; Yee et al., 1998; Hong et al., 2004; Hall et al., 2007; Price et al., 2005; and Louchard-del la Chapelle et al., 2005. A meta-analysis of 45 studies was highly significant (Patterson et al., 2008). Studies that have used the beta EEG filtering parameters that we specified have generally reproduced our findings (Brochaus-Dumke et al., 2008), whereas those that use other frequencies have not (Hong et al., 2008). One of the advantages of the paradigm is that it is entirely passive and therefore it is amenable to modeling in animals, including those under anesthesia (Bickford-Wimer et al., 1990). The deficit has been modeled in animals by other groups as well (Boutros et al., 1998; Garcia-Rill et al., 2007; Javitt et al., 2000; Radek et al., 2006; Maxwell et al., 2006). The rodent model showed that one of the sites of responses that decremented rapidly to repeated stimuli was the hippocampus. There have been demonstrations of hippocampal hyperactivity in schizophrenia using other imaging paradigms (Malaspina et al., 1999).
Decreased hippocampal response to repeated stimuli in rodents depends upon the input from the midbrain, because lesion of the fimbria-fornix causes the hippocampus to become disinhibited. We studied the effects of a series of cholinergic antagonist on rats with intact innervation to establish whether the cholinergic input from the medial septal nucleus was involved and to identify the receptor type. The experiments identified low affinity nicotinic receptors, antagonized by α-bungarotoxin, as the key receptors necessary for the inhibition of hippocampal evoked response to repeated auditory stimuli (Luntz-Leybman et al., 1992). Alpha-bungarotoxin labels interneurons and these neurons can be co-labeled with antibodies for GABA in the rat hippocampus (Figure 8). The neurons are found in all hippocampal areas and various representatives can be co-labeled with cholecytokinin, somatostatin, and neuropeptide Y (Freedman et al., 1993). They are thus a subset of hippocampal interneurons in most hippocampal regions. Their prominence in CA3 stratum oriens and the dentate hilus corresponds with the increased cholinergic innervation of this region, but they are also prominent in CA1. In addition to the labeling of cell bodies and processes, there is diffuse labeling in CA2 and CA3 in particular, which may be reflective of labeling on presynaptic terminals. Postsynaptic responses of interneurons to acetylcholine are blocked by receptor antagonists like α-bungarotoxin (Alkondon et al., 1993; Frazier et al., 1998). Thus, the α-bungarotoxin-sensitive receptor emerged as a candidate mechanism for the loss of sensory gating function in schizophrenia. The gene for this receptor was subsequently cloned as Chrna7 and it was identified as the α7-nicotinic acetylcholine receptor (Sequela et al., 1993).
Alpha 7-nicotinic receptors form a five-membered homomeric ring that upon activation by agonists admits cations into the cell, including both calcium and sodium ions. The receptors are co-localized with nitric oxide synthetase, and nitric oxide appears to be one of the second messengers activated by these receptors (Figure 9; Adams et al., 1997). Blockade of the receptors increases nerve growth factor expression in the hippocampus, one of several pieces of evidence that the receptors have a role in hippocampal function beyond simple information processing (Figure 10; Freedman et al., 1993).
Chrna7 is polymorphic and has different levels of expression in various inbred mouse strains. DBA/2 mice have diminished levels of α7-nicotinic receptors, compared to C3H mice. DBA/2 mice also have diminished inhibitory sensory gating in the hippocampus (Figure 11; Stevens et al., 1996). Using polymorphic markers in Chrna7, we compared introgression of the C3H and DBA/2 variants into C3H parental strains. As expected, there was no change in expression of α7-nicotinic receptors or sensory gating function with the introgression of C3H alleles into C3H parental strains. However, the introgression of DBA/2 variants caused loss of sensory gating function and loss of expression of α7-nicotinic receptors. The receptors were not only diminished, but their distribution in the hippocampus was also changed. C3H mice have their most prominent expression in CA2 and CA3, whereas DBA/2 mice have their most prominent expression in CA1. This pattern was reproduced by introgression of the DBA/2 allele into the C3H parental background (Figure 11; Adams et al., 2001).
Schizophrenic hippocampal α7-nicotinic receptor-expressing neurons are fewer and have decreased elaboration of neuronal processes. The other markers associated with α7-nicotinic receptors in rodents are also abnormal in schizophrenia postmortem brain. There is decreased expression of cholecystokinin and somatostatin (Ferrier et al., 1983). Interneurons that contain neuropeptide Y are dystrophic, with diminished extension of processes (Iritiani et al., 2000). Interneurons that contain nitric oxide synthetase are diminished in the hippocampus and other cortical structures and appear not to be migrating from the cortical subplate (Akbarian et al., 1993). Other more general markers of interneuron integrity, GABA reuptake sites and synapsin expression, are also diminished (Simpson et al., 1989; Browning et al., 1993; Vawter et al., 2002).
Diminished postmortem expression was also found in the frontal cortex and in the nucleus reticularis thalamis (Court et al., 1999; Guan et al., 1998; Marutle et al., 2001). The nucleus reticularis thalamis does not express prominent α7-nicotinic receptors in rodents, but it does in humans and other primates, perhaps consistent with the greater role of this nucleus for inhibitory gating of sensory input to the primate neocortex (Breese et al., 1997).
Like many features of schizophrenia, sensory gating dysfunction is co-transmitted in families with the illness (Siegel et al., 1984). Therefore, a genetic linkage study of its heritability was undertaken in families with multiple cases of schizophrenia. Sites of linkage were found on several chromosomes, including chromosome 15q14. Investigators at the Salk Institute subsequently localized the human CHRNA7 at 15q14. A dinueclotide repeat sequence was isolated from a yeast artificial chromosome containing the CHRNA7 sequence, and the linkage was repeated, with a two-point LOD score for sensory gating dysfunction of 5.2, providing strong evidence for genetic linkage at this site (Figure 7; Freedman et al., 1997). The LOD score for schizophrenia itself was also positive, but less significant, because there are fewer cases of schizophrenia in the families. Many relatives have the sensory gating dysfunction but not schizophrenia. The sensory gating dysfunction is thus an example of an endophenotype, a phenotype more closely linked to the gene than the illness itself that is necessary but not sufficient to cause the illness. Because the genetic linkage was performed across the entire genome, there was no presupposition that any particular gene was involved. The convergence at chromosome 15q14 of linkage the gating dysfunction with the location of CHRNA7, the gene for a candidate mechanism for gating dysfunction, strengthened the evidence of the involvement of the α7-nicotinic receptor.
Seven subsequent independent studies have shown that the 15q13 CHRNA7 locus is also linked or associated with schizophrenia (Kaufmann et al., 1998; Gejman et al., 2001; Liu et al., 2001; Riley et al., 2000; Stassen et al., 2002; Tsuang et al., 2001; Xu et al., 2001). Ethnic groups include Europeans, European-Americans, Bantus, African-Americans, Han Chinese, and Azoreans. The range of phenotypes with which CHRNA7 is associated also strengthens the finding. Association between CHRNA7 and P50 sensory gating has been replicated (Raux et al., 2003). The NIMH Consortium on the Genetics of Schizophrenia (COGS) found association between CHRNA7 and prepulse inhibition (Greenwood et al., 2008), and association has also been found to episodic memory dysfunction in schizophrenia (Dempster et al., 2006). Nevertheless, like all genetic findings in schizophrenia, these findings have not been universally replicated. Association study of 14 candidate genes by the NIMH Molecular Genetics Collaboration using Genome Wide Association (GWAS) standards found none with genome-wide significance. However, the intron 4 CHRNA7 marker that showed positive signals—rs10438342 G allele, relative risk 1.10, P = 0.043 (Sanders et al., 2008)—is similarly positive in another large GWAS study, relative risk 1.09, P = 0.017, Stefansson et al., 2008). In the Stefannson et al. study and a similar by Stone et al. (2008), there were rare individuals with schizophrenia who had deletions of a small region of chromosome 15 that contains CHRNA7. Many of these deletions were de novo, appearing in the affected individual but not in either parent. These cases also had no family history of schizophrenia. In most individuals with schizophrenia, however, the coding region of the gene is intact and polymorphisms, if any, have been found in the promoter region of the gene (Leonard et al., 2002). Thus, consistent with the postmortem expression, there appears to be diminished expression of a normal receptor.
A straightforward therapeutic strategy to overcome the inhibitory gating deficit is to stimulate α7-nicotinic receptors. Nicotine transiently activates α7-nicotinic receptors and transiently enhances P50 sensory gating in schizophrenia (Adler et al., 1993). However, the effect rapidly dissipates before it produces any meaningful clinical effect. A more suitable drug would be an agonist less likely to cause tachyphylaxis, the loss of response to repeated doses. Tachyphylaxis is a classical characteristic of nicotinic receptors and reflects receptor desensitization among other neuronal responses. DMXB-A (3-2,4 dimethoxybenzylidene anabaseine) is a selective α7-nicotinic receptor agonist (de Fiebre et al., 1995). It has 40% bioavailability in plasma after oral administration and no obvious severe side effects. DMXB-A, administered intra-gastrically, increased the inhibition in the DBA/2 mouse model of the sensory gating deficit in schizophrenia with less tachyphylaxis than nicotine (Stevens et al., 1998). We completed stability studies and preclinical toxicology in two species to obtain an FDA Investigational New Drug exemption for its clinical testing and performed a Phase 1 study that showed that it increased P50 inhibition in schizophrenia (Olincy et al., 2006).
There was significant improvement in the Phase 1 study in the neuropsychological performance of the patients, particularly their ability to pay attention during repetitive tasks, which is one of the purposes of sensory gating—to exclude distraction and thereby protect concentration. In a Phase 2 trial with one month’s administration, DMXB-A also improved attention, as well as the negative symptoms alogia and anhedonia, which are often resistant to dopamine receptor antagonist antipsychotic drugs (Freedman et al., 2008).
CHRNA7 has roles beyond adult brain dysfunction in schizophrenia. Expression of α7-nicotinic acetylcholine receptors begins in neuroblasts and is maximal n both developing rodents and humans, when levels are over four times adult values (Figure 12; Adams et al., 2002; Court et al., 1997). Alpha7-nicotinic acetylcholine receptors are expressed in CA1 near birth at particularly high levels (Adams et al., 2003). Their activation facilitates maturation of glutamate and GABA synapses (Aramakis et al., 2000; Liu et al., 2006). DBA/2 mice not only have diminished α7-nicotinic acetylcholine receptors, but they also have diminished levels of kainate and AMPA-type glutamate receptors, compared to NMDA-type receptors, a pattern also found in schizophrenia (Zilles et al., 2000; Porter et al., 1997). Thus, the developmental role of α7-nicotinic acetylcholine receptors suggests an influence on the ultimate functioning of the brain, and its dysfunction in schizophrenia, that may be more important than its role in sensory gating in adult life. Cholinergic innervation develops near birth and therefore during much of brain development, acetylcholine is not available to activate these receptors. However, free choline in high concentrations, as are found in amniotic fluid, can activate α7-nicotinic acetylcholine receptors (Ilcol et al,, 2002; Alkondon et al., 1997; Frazier et al., 1998). Choline has already been shown to be essential for normal hippocampal development (Meck et al., 1989; Albright et al., 1999).
Just as we have stimulated α7-nicotinic acetylcholine receptors in schizophrenic adults with DMXB-A to improve brain function, it is possible that stimulation of α7-nicotinic acetylcholine receptors, during gestation may overcome the adverse effects of receptor deficiency on neuronal development. Unfortunately, the endogenous ligand choline is often deficient in pregnant women because of inadequate nutrition, genetic abnormalities in the synthetic enzyme phenethanolamine methyl transferase, or sequestration of choline in the maternal liver because of stress (Gossell-Williams et al., 2005; Liu et al., 2007). We therefore attempted dietary choline supplementation. Figure 13 shows the effects in the DBA/2 mouse model. Supplementation of mouse dams during pregnancy and nursing resulted in offspring who had improved hippocampal sensory gating in adulthood, compared to those who received normal mouse diets. Neither group received supplementation after weaning (Stevens et al., 2008). Based on these data, the Food and Drug Administration has granted us an Investigational New Drug exemption to assess choline supplementation in a double-blind, placebo controlled trial in humans. Choline, administered as phosphatidylcholine, has already shown positive effects on cognition in the offspring of supplemented women (Zeisel, 2006). The purpose of the test is to determine if it will also ameliorate the sensory gating deficit and other endophenotypes associated with schizophrenia, to decrease the eventual risk of developing the illness.
Hippocampal development and function are a focus for many neurobiologists concerned about the brain dysfunction in human illness. Basic studies at the Karolinska Institute have shown the innervation of the hippocampus by cholinergic, adrenergic, and glutamatergic afferents is an exquisitely controlled phenomenon. Early intervention in the fetal period to improve brain development and prevent future mental illness is a possible translational outcome of this research.
The authors wish to acknowledge Drs. Lars Olson, Ake Seiger, and Ingrid Stromberg. The in oculo work would not have been possible without the major contribution from the technical staff at the Histologen Inst. that started with the incomparable technical expertise of Ingrid in putting embryonic tissue into the anterior eye chamber, the histological expertise of Hullan, Carina and assistance from Waldtraut and Maude.
NOTE: This article is contributed for the special issue “Half a Century of Monoamine Research: A Tribute to Nils-Åke Hillarp and His Students”
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Robert Freedman, Chair, Dept Psychiatry, University of Colorado at Denver and Health Sciences Center, 13001 E. 17th Pl. Campus Box F546, Aurora, CO 800045, USA.
Dan Goldowitz, Centre for Molecular Medicine and Therapeutics, Child and Family Research Inst, Dept Medical Genetics, University of British Columbia, Office: 980 W 28th Ave, Rm 2026 - Lab: 2032, Mailing: 950 West 28th Avenue, Vancouver, BC V5Z 4H4, CANADA.