Chapter 51:  Neurochemistry of Schizophrenia

Schizophrenia, manic-depressive illness, psychotic depression and organic psychoses of known etiology, such as the alcoholic and senile psychoses, are the major forms of psychotic disorders

At some level of neural function, they share the ability to produce characteristic clinical features: (i) delusions or false beliefs; (ii) hallucinations or false perceptions, usually without insight into their pathological nature; and (iii) disorganization of thought, for example, incoherence. These are sometimes accompanied by bizarre behavior. Abnormalities in the structure or function of neurons are central to the various forms of psychosis. This chapter focuses on schizophrenia as an example of psychosis. It is chosen because of the challenge it represents to neurochemistry and because it is a devastating disease which affects about 1% of the population, has a 9 to 13% suicide rate and leads to annual costs, in the United States alone, for 1995, of about 65 billion dollars for medical expenses and indirect costs, such as lost income.

The current view of schizophrenia began with the pioneering clinical observations of Emil Kraepelin, a German psychiatrist, who in 1896 identified a group of psychotic patients with an early age at onset, usually at the end of the second or beginning of the third decade of life, which permanently impaired cognition and usually ended in poor outcome. The characteristic age at onset and impairment in cognition led him to designate this illness as dementia praecox. The term schizophrenia was coined by Eugen Bleuler about one decade later because of his view that many patients with the same hallucinations and delusions present in dementia praecox did not, in fact, develop severe dementia. Schizophrenia was intended to reflect the splitting of affect, or feelings, and cognition. It does not refer to split personality. Bleuler emphasized four types of symptoms: autism, ambivalence, flat affect and disturbances in volition, or will. The current view of schizophrenia is an integration of both Kraepelin's and Bleuler's views, emphasizing characteristic symptoms and cognitive disturbance in a disorder which generally comes on between the ages of 16 and 45. The pendulum has swung back to the view that cognitive impairment, which consists of deficits in attention; vigilance; working, semantic and storage memory; and executive function, is central to the illness [1]. Diagnostic criteria for schizophrenia may be found in the Diagnostic and Statistical Manual-IV of the American Psychiatric Association.

The psychopathology of schizophrenia is usually described in terms of three somewhat independent syndromes, or symptom clusters

These are positive, disorganized and negative symptoms [2]. Positive symptoms consist of the florid psychotic symptoms, mainly delusions and hallucinations. The delusions in patients with schizophrenia are usually paranoid, that is, delusions of persecution. Other characteristics are delusions of control, thoughts being inserted or removed from one's mind, grandiosity, somatic and tactile delusions and other bizarre ideas, from the perspective of normal people. Hallucinations are usually auditory in nature and may be experienced as coming from internal or external sources. Recent studies have shown abnormal temporal lobe activity in auditory sensory areas during the experience of auditory hallucinations. Disorganization as a syndrome of schizophrenia includes incoherence, illogicality, loose associations, inappropriate affect and poverty of thought content. Negative symptoms include withdrawal, impoverished emotional state, motivational difficulties, lack of energy, affective flattening, loss of spontaneity and lack of initiative. Depression and anxiety are also frequently present in schizophrenia and are independent of the three syndromes described above, which are core features of the diagnosis of schizophrenia.

Not all of these symptoms are present at any one time. They also vary in severity over time. Neurochemical studies of schizophrenia are, thus, carried out on heterogeneous populations of patients from the point of view of psychopathology. Such studies must distinguish between so-called state characteristics, such as transient increases in positive symptoms which respond to treatment, and trait characteristics, such as negative symptoms and cognitive impairment which are relatively stable. Even negative symptoms may be variable. So-called primary negative symptoms are stable and not related to positive symptoms, depression or side effects of antipsychotic drugs, while secondary negative symptoms are, as the name implies, believed to be the results of other disease processes in schizophrenia [3]. As current therapies are differentially effective in treating various components of schizophrenia (see below), it is most important to understand the neurochemical underpinnings of these components whose etiology may be only partially overlapping.

When mood symptoms are a major feature of a patient with otherwise characteristic schizophrenia, the diagnosis of schizoaffective disorder is made. Various genetic studies have established that this mixed syndrome is more closely related to schizophrenia than to mood disorders, although there is considerable support for the hypothesis that these two groups of psychoses share some genes. From the point of view of genetic and neurochemical studies, it is also important to be aware of the concept of schizophrenic spectrum disorders. These include schizoid and schizotypal personality disorders. Schizoid individuals have mainly the negative symptoms of schizophrenia without positive symptoms or disorganization, while schizotypal personality disorders manifest mild forms of positive symptoms and disorganization that do not reach the threshold of being considered psychotic.

Cognitive impairment in schizophrenia may be present to a minor extent in childhood and early adolescence but is relatively modest at that time compared with the degree of dysfunction present at the time the diagnosis of schizophrenia is made following the emergence of positive symptoms. The importance of cognitive impairment in schizophrenia is underlined by the findings that it, rather than positive symptoms, is most important for the impaired work function [4]. It is believed that during the prodromal periods prior to the emergence of delusions and hallucinations, neural abnormalities at a functional and perhaps structural level develop, producing cognitive impairment. These abnormalities are most likely quite diffuse in nature, but the functions that are predominantly localized to the frontal cortex and temporal lobe rather than the parietal lobe are the most compromised. There is also increasing evidence of a role for the basal ganglia in cognitive function. Deficits in connectivity between regions may be as important as or more important than abnormalities confined to specific brain regions. Functional magnetic resonance imaging is making it possible to study the sequence of activation of specific brain regions recruited during the performance of specific types of cognitive tasks (see Fig. 50-6) and to demonstrate whether patients with schizophrenia have the capacity to modulate the activity of various brain regions in the same fashion as normal controls. The nature of the cognitive deficits in schizophrenia varies considerably from one patient to the next, suggesting that complex neurochemical processes that evolve differently in patients with this illness are involved. Compensatory mechanisms may make it possible for some patients to perform within the normal range on specific tasks. The overall IQ is important in this regard. The neurochemistry of cognition probably involves many neurotransmitters; acetylcholine (ACh), dopamine (DA), serotonin (5-HT), GABA and glutamate, all of which have been implicated in schizophrenia, are believed to be of the greatest importance for cognition.

The cornerstone of the treatment of schizophrenia is a group of antipsychotic drugs which are of value for most forms of psychosis

The serendipitous discovery of the antipsychotic effects of chlorpromazine in 1953 by Laborit, Delay and Denicker in France and the subsequent demonstration by Arvid Carlsson of Sweden that their antipsychotic action was due to the blockade of DA receptors opened up the modern era of schizophrenia treatment and research. Chlorpromazine was followed by the introduction of many other DA receptor blockers of different chemical classes, aided by the demonstration that a specific type of DA receptor, the D2 receptor, negatively coupled to adenylyl cyclase, was their apparent target of action. These agents are called neuroleptics because they impair motor function in animals and humans. These motor effects, which are due to impairment of the extrapyramidal system, are similar to the symptoms of Parkinson's disease: rigidity, difficulty in initiating movements and tremor. In addition, antipsychotic drugs may produce dystonias and result in a very disturbing type of restlessness and agitation called akathisia. They can also produce a long-term, sometimes irreversible, impairment of motor function; because this condition is usually delayed in onset, it is called tardive dyskinesia (see Chap. 45). Tardive dyskinesia consists of abnormal involuntary movements, usually involving the facial muscles, the tongue, and sometimes the limbs and diaphragm. The rate of development of tardive dyskinesia is 4 to 5% per year of continuous exposure to neuroleptics, but for reasons unknown, less than 25% of patients will develop this side effect. Because of their ability to block D2 receptors in the anterior pituitary gland, neuroleptics also stimulate prolactin secretion, especially in females. This may produce milk secretion, termed galactorrhea.

Among the commonly used neuroleptics are haloperidol, fluphenazine, molindone, thiothixene, sulpiride and thioridazine. There is little evidence that these drugs are differentially effective in schizophrenia, but they differ in potency and side effects, including motor side effects.

Because of the many negative consequences of unwanted blockade of dopaminergic function in the basal ganglia, agents which can achieve an antipsychotic action without causing motor side effects were sought from the earliest period of drug discovery in this area. The first such agent was clozapine [5]. Although chemically related to the neuroleptic loxapine, clozapine did not produce catalepsy in rodents nor did it produce acute or subacute extrapyramidal dysfunction or tardive dyskinesia in humans. Furthermore, low doses of clozapine have been shown to be tolerable to Parkinson's disease patients, who are given the drug to block psychotic symptoms produced by dopamine replacement therapy. Clozapine produces agranulocytosis in 1% of patients, which has limited its usefulness to patients who are neuroleptic-intolerant or do not respond adequately to the neuroleptics. Independent of its lack of motor side effects, cloza-pine is more effective than the neuroleptics in the control of positive and negative symptoms in patients who fail to respond adequately to neuroleptics [6]. Since the discovery of clozapine, other agents which are antipsychotic and have reduced motor side effects have been discovered: iloperidone, melperone, olanzapine, ORG 5222, quetiapine, risperidone, sertindole and ziprasidone. These agents are usually referred to as atypical antipsychotics because they cause less of the characteristic motor side effects of neuroleptics at clinically effective doses. They do not necessarily share any of the other characteristics of clozapine and differ among themselves in important ways. Clozapine, olanzapine and risperidone appear to improve cognitive function more so than typical neuroleptic drugs, although the pattern of changes produced by each drug is different. All of these agents also seem to have advantages for negative symptoms. It is debated whether these advantages are true for both primary and secondary negative symptoms or only for secondary negative symptoms. Risperidone produces marked increases in prolactin secretion, which the others do not. As will be discussed, one feature all these drugs share is a high affinity for the 5-HT_2a receptor relative to the D2 receptors [5].

Other pharmacological profiles may also produce atypical antipsychotic drugs. Remoxipride has selective antagonism for D2 receptors (see below). It was dropped from clinical use because it produced aplastic anemia. Recent studies suggest that drugs which are 5-HT_1a agonists as well as D2 antagonists may also have atypical properties.

It is generally accepted, despite the absence of evidence, that schizophrenia is a group of disorders with a common overlapping phenotype rather than a single disease entity

It has a complex mode of inheritance and variable expression. Adoption, twin and family studies carried out in the 1960s established that the vulnerability to develop schizophrenia is largely genetic. What is inherited is an increase in the risk of becoming schizophrenic rather than a gene or genes that absolutely predict the occurrence of schizophrenia. Thus, about half of monozygotic (MZ) twins are concordant for schizophrenia compared to less than 20% of dizygotic (DZ) twins, using psychosis as the phenotype. It is therefore clear that environmental as well as genetic factors are important in the development of the disorder.

The high rate of discordance between MZ twins indicates that what is inherited is a predisposition but not a certainty of developing schizophrenia. In first-degree relatives, percent lifetime expectancy to develop schizophrenia is about 10% if one parent or a sibling has schizophrenia and 45 to 50% for offspring of two schizophrenic parents. In second-degree and third-degree relatives, the expectancy drops to 3.3% and 2.4%, respectively. Furthermore, adoption studies show a lifetime prevalence of 9.4% in the adopted-away offspring of schizophrenic parents and a lifetime prevalence of 1.2% in control adoptees [8].

The distribution of schizophrenia in families indicates a complex inheritance since the risk to relatives declines markedly as the relationship becomes more distant. Studies of extended pedigrees with multigenerational schizophrenia have ruled out single dominant genes as the cause of the illness in the majority of cases. However, there may be a small proportion of cases in which a single major gene, acting either alone or in concert with multiple small genetic and environmental factors, accounts for the vulnerability to become schizophrenic. It is likely that additive effects of several genes of modest effect, termed oligogenic inheritance, or many genes of small effect, termed polygenic inheritance, are the basis for the vulnerability.

Epidemiological studies of the genetics of schizophrenia are well established and have led to numerous studies using diverse methods to identify specific genes. Many studies reporting the involvement of specific genes in schizophrenia have failed replication. One reason for this is differences in the criteria for the phenotype. Linkage results vary greatly as to whether criteria based on symptomatology are restrictive or broad, for example, excluding schizoaffective patients and/or schizophrenia spectrum patients. Additionally, so-called logarithm of odds (lod) scores must be substantially higher than 3, that is, one chance in 1,000, in a complex disorder such as schizophrenia. Early studies accepted lod scores of 2 or 3 as proof of linkage.

Current major strategies to identify the genes in schizophrenia involve genome scanning through linkage and association studies and candidate gene analysis, for example, D2, D4 and 5-HT₂ receptor genes [9]. The latter strategy ruled out the involvement of various genes that are part of the DA system, such as the genes for tyrosine hydroxylase and the D1, D2 and D4 receptors. However, several studies have suggested an association between exon 1 of the D3 receptor and schizophrenia. Despite some nonreplications, this possibility is still the subject of considerable research. Although DA is the main neurotransmitter implicated in schizophrenia, there is also keen interest in 5-HT. Associations between a T to C polymorphism at nucleotide 102 in the 5-HT_2A receptor and schizophrenia, and between a Ser-to-Tyr polymorphism at nucleotide 452 and response to clozapine have been reported in several studies [9]. Positional cloning is most likely to be successful when applied to large families containing multiply affected members, when the disease gene has a major effect, when the mode of inheritance of the phenotype is known and when there are few diagnostic errors. These features are not characteristic of schizophrenia and have led to a number of reports of linkage or association that have not been replicated. For these reasons, the currently favored strategy is based on identity by descent involving the analysis of affected siblings. Evidence of linkage is provided when affected sibling pairs share loci more often than would be expected by chance alone. This method does not require knowledge of the mode of transmission, has the power to detect genes of modest effect, is not very sensitive to misdiagnosis and can be readily applied to schizophrenia.

Using the affected sibling pair method, there have been several replications of a locus on chromosome 6p24-p22 for schizophrenia [8]. This region contains the human leukocyte antigen (HLA) locus, which has been postulated to be relevant to schizophrenia. There is limited and controversial evidence that schizophrenia may be one of a group of disorders in which expanded trinucleotide repeats of variable length are present (see Chap. 40). Advances in the molecular genetics of schizophrenia in the future may lead to identification of the group of genes that convey vulnerability to this syndrome.

The neurodevelopmental hypothesis suggests that the etiology of schizophrenia may involve pathological processes during brain development

This hypothesis is based on the demonstration of behavioral and cognitive disturbances in childhood and adolescence that are eventually diagnosed as schizophrenic [10]. The absence of marked neurodegenerative changes in the schizophrenic brain together with findings suggestive of cortical maldevelopment are consistent with this hypothesis.

According to this hypothesis, the etiology of schizophrenia may involve pathological processes which begin in utero or perinatally and continue to unfold until the brain approaches its adult anatomical state as a result of extensive neuronal loss and synaptic pruning during early and late adolescence. These neurodevelopmental abnormalities are proposed to lead to the activation of pathological neural circuits during adolescence or young adulthood, perhaps due to severe stress, leading to the emergence of positive or negative symptoms or both. Some cases with the phenotype of schizophrenia may be due to embryonic maldevelopment, especially of the corpus callosum and temporal lobe, for example, temporal lobe epilepsy. The emergence of evidence for cortical maldevelopment in schizophrenia and the development of several plausible animal models, which are based on neonatal lesions that produce behavioral abnormalities or altered sensitivity to dopaminergic drugs only in adolescent or adult animals [11], have made the link between maldevelopment and schizophrenia more tenable.

A consistent finding in schizophrenia is cerebral ventricular enlargement [12]. A large number of computed tomography (CT) and magnetic resonance imaging (MRI) studies indicate lateral and third ventricular enlargement and widening of cortical fissures and sulci; these are present at the onset of the illness, progress very slowly if at all and are, therefore, unrelated to the duration of illness or the treatment received. Affected MZ twins discordant for schizophrenia have larger ventricles than unaffected twins. The loss of gray matter is correlated with poor premorbid social and educational adjustment during early childhood as well as obstetric complications. These findings are not specific for schizophrenia, however, as they are also found to almost the same extent in manic-depressive illness.

Despite extensive efforts to discover a neuropathological basis for schizophrenia, no consistent characteristic lesions, at either the micro- or the macroscopic level, have yet been identified

However, various abnormalities in the temporohippocampal and frontal lobes, the two brain regions most likely to be abnormal in schizophrenia, are currently being studied. There is extensive evidence that gliosis is not present, indicating that there is no neuronal death due to traumatic, inflammatory processes or infection in schizophrenia. However, an abnormality of programmed cell death, that is, apoptosis, has not been ruled out. Increased density of neurons in the prefrontal cortex, loss of interneurons in the cingulate and prefrontal cortices and abnormalities in migration of cells from the cortical plate to the gray matter of the cortex are among the most interesting recent findings. However, they are based on a small number of samples and have not been shown to be specific for schizophrenia.

Since the mid-1980s, there has been an increasing reliance on anatomical methods to reveal neurochemical changes in postmortem studies of schizophrenia. This has been due in large part to a growing appreciation of the diversity of neurons within what were once considered by most neurochemists, and indeed most neuroscientists, to be single classes of neurons. For example, studies of the gene encoding the GABA biosynthetic enzyme glutamic acid decarboxylase (GAD), as a marker of cortical GABA cells treat these interneurons as a unitary class of cells. However, well over a dozen different types of interneurons can be distinguished morphologically, and the morphological distinctions are paralleled by physiological and neurochemical distinctions. Anatomical methods have proven very useful since they allow subpopulations of interest to be distinguished.

Recent studies have revealed various neuropathological changes in the brain, particularly in the prefrontal cortices, including the pregenual anterior cingulate cortex, and the medial temporal lobe, including the hippocampus, parahippocampal gyrus and entorhinal cortex. Unfortunately, the types of changes reported (ranging from cell loss and changes in cell density in the absence of cell loss to changes in neuronal size, position or orientation) are inconsistent. With increasingly rigorous application of quantitative, computer-assisted, neuroanatomical methods, a greater degree of consistency may emerge. It is clear that representative samples of adequate size must be studied and the comparison groups (such as bipolar disorder and major depression) be included in such studies.

Since schizophrenia typically first occurs in late adolescence or early adulthood, developmental processes are considered key to its pathogenesis. The developmental hypotheses of schizophrenia implicitly assume that changes in neurochemical and neuroanatomical markers represent the culmination of a process and that such end points may be indirect or direct.

The dopamine hypothesis has dominated schizophrenia research since the mid-1960s

This is due in large part to the fact that antipsychotic drugs block DA receptors, suggesting that overactivity of central DA systems underlies schizophrenia. However, several other hypotheses have recently been advanced or reformulated to account for the pathophysiology of schizophrenia; most pay homage to the DA hypothesis by incorporating secondary changes in central DA systems that are due to some other putative primary defect.

Several types of studies have attempted to test the hypothesis that there is a primary change in central DA function in schizophrenia. These include examination of (i) differences in concentrations of DA or its metabolites in various brain sites, (ii) the ability of chronic administration of high doses of amphetamine to induce a paranoid form of schizophrenia, (iii) the relationship between the ability of antipsychotic drugs to block D2 receptors and their specific responses and (iv) changes in other types of DA receptors. Unfortunately, the pattern of changes in the various studies is not consistent.

It is important to remember that all classification schemes for schizophrenia emphasize that there are different subtypes. In addition, a substantial body of data demonstrates age-related changes in the DA system; such changes may interact with the number of psychotic episodes and drug treatment to yield different specific symptoms of schizophrenia. It is, therefore, not unexpected that there might be comparable heterogeneity in various markers of DA function.

Dopamine and dopamine metabolite concentrations. Despite considerable effort to detect changes in concentrations of DA or its acidic metabolite homovanillic acid (HVA) in the striatum and other relevant brain areas, most findings have been negative; studies reporting changes, which in some cases are lateralized, have generally been difficult to replicate. Studies of HVA concentrations in cerebrospinal fluid (CSF) and plasma have also been disappointing. While many studies have reported increases in plasma concentrations of HVA in schizophrenia, other reports have found no differences or even decreases relative to normal controls. There is a consistent observation that neuroleptics increase CSF HVA concentrations, although the relationship of such changes to symptomatic improvement remains unclear. Part of the difficulty in obtaining reliable CSF HVA data appears to be technical in nature since in many of the studies no attempt was made to control for the contribution of peripheral sources of HVA, as can be accomplished using debrisoquin, a peripheral monoamine oxidase inhibitor, or by employing probenecid treatment to prevent blood—CSF exchange of HVA. However, even in those studies using such methods to eliminate the contribution of peripheral DA, the precise sources of CSF HVA are unclear. In one study of nonhuman primates, CSF HVA and tissue DA concentrations showed a significant correlation in only one brain area, and that accounted for only about 10% of the variance. Because the relationship between central DA function and CSF and plasma HVA concentrations is unclear, these measures do not provide strong support for a primary or major dopaminergic dysfunction in schizophrenia.

Dopamine receptors. Since the mid-1960s, striking similarities between the psychosis seen in certain subjects taking high doses of amphetamine and the symptoms of patients with paranoid schizophrenia have been noted and placed into the context of increased catecholaminergic neurotransmission. Subsequent studies emphasized the contribution of the central dopaminergic, rather than noradrenergic, systems to psychosis. The ability to treat psychosis with D2 receptor antagonists was consistent with the idea of an overactivity of central DA function.

The most compelling of the data marshaled to support the DA hypothesis of schizophrenia is the clear relationship between antipsychotic drug efficacy and affinity for D2 receptors [13]. All known drugs with broad antipsychotic efficacy disrupt DA transmission. With the exception of reserpine, which acts by depleting vesicular stores of DA and is no longer used in the treatment of schizophrenia in Western medicine, antipyschotic drugs block the D2 receptor. An almost linear relationship between the D2 receptor affinity of an antipsychotic drug and some index of the drug's action, such as the average daily dose or plasma or CSF concentrations, has emerged. The exception to this picture has been clozapine; some of the newer atypical antipsychotic drugs also diverge modestly from this relationship. The affinity of these newer atypical antipsychotic drugs for the D2 receptor is relatively low. However, most studies have focused on the affinity of the parent compound for the D2 receptor, without considering active metabolites. Hepatic metabolism of clozapine yields active metabolites that achieve high brain concentrations; one of these, desmethylclozapine, has a higher affinity for the D2 receptor than clozapine. Nevertheless, it is unlikely that D2 receptor blockade is the sole basis for the effectiveness of these atypical agents (see below).

The ability of antipsychotic drugs to block D2 receptors has led to attempts to measure the density of DA receptors in the brains of schizophrenic subjects. The genes for five DA receptors have now been cloned and designated as D1, D2, D3, D4 and D5 [14]. The D1 and D5 receptors are positively coupled to adenylyl cyclase, with current pharmacological probes unable to distinguish between these two D1-type receptors. The two receptors have different regional patterns of expression in the brain, with the D5 receptor transcript being much more restricted in its distribution [15]. The other three DA receptors belong to the D2 family and are negatively coupled to adenylyl cyclase. Again, the CNS distributions of these receptors differ considerably, both within and across different species [15].

The D2-like receptors have been the major target of investigations aimed at uncovering changes in DA receptors in schizophrenia. The D2 receptor is the most widely distributed of the D2 family receptors and has a high affinity for typical antipsychotic drugs but lower affinity for atypical antipsychotic drugs. Atypical antipsychotic drugs, of which clozapine is the best known, have antipsychotic efficacy but lack or have a much reduced, ability to elicit extrapyramidal (parkinsonian-like) side effects (EPS) [16]. The ability of atypical antipsychotic drugs to reduce symptomatology without accompanying motor side effects may be due to their relatively low degree of occupancy, as well as reduced affinity, of striatal D2 receptors. In contrast to typical antipsychotic drugs, which result in virtually complete occupancy of dorsal striatal D2 receptors, clozapine occupancy of D2 receptors is variable across patients but almost never exceeds 70% as measured by positron emission tomography (PET) (see Chap. 54) [17]. Given the remarkable functional compensation of the striatal DA system, which may be depleted by 70% or more before giving rise to the symptoms of Parkinson's disease, the relatively low occupancy of striatal D2 receptors may be sufficient to avoid EPS. However, there are other actions of clozapine, such as the ability to treat patients who do not respond to conventional antipsychotic drugs and better efficacy in reducing certain symptoms of schizophrenia [16], which have been suggested to be due to full D2 occupancy by clozapine in cortical regions [18].

Despite a generation of studies on the involvement of D2 receptors in schizophrenia, a strong data-based argument in support of this is difficult to sustain. Most autoradiographic and in vitro assessments of D2 receptor density have not uncovered consistent evidence of an increase in the density of D2 receptors in the striatum or other brain sites. Early PET studies suggested that striatal D2 receptor density was increased in schizophrenia, but subsequent PET and single photon emission computerized tomographic (SPECT) studies using several different radioligands have failed to observe such an increase in D2 receptor density in nonmedicated schizophrenic subjects [19].

Other CNS areas examined, such as the prefrontal cortex, have also shown no dramatic changes in D2 receptor density. A recent autoradiographic study of D2 receptor-binding sites in the temporal lobe reported differences in the pattern of D2 receptor expression, with opposite directions of change in the supragranular and granular layers of temporal cortices [20].

Autoradiographic binding studies of receptors are often plagued by technical difficulties owing to the relatively long interval between death and the collection of brain tissue. The susceptibility of receptor-binding studies to long postmortem intervals can be overcome to some degree by studying receptor gene expression, since the mRNAs encoding for the receptors are more stable than the encoded proteins. However, the few studies of D2 gene expression that have been reported do not suggest a change in D2 gene expression in schizophrenia [21].

The D3 receptor has a more restricted distribution than the D2 receptor, with high expression in the ventral striatum, including nucleus accumbens, and moderate expression in some cortical regions in primate, but not rodent, species [15]. The D3 receptor exhibits considerably less affinity for clozapine than for typical antipsychotic drugs, and the precise intracellular transduction pathways through which D3 receptor occupancy is translated into intracellular events are not clear [14]. D3-binding sites in the nucleus accumbens have been reported to be increased in patients who did not receive antipsychotic drug treatment in the month prior to death [20]. In those patients receiving antipsychotic drugs within 72 hr of death, the density of D3 receptor sites was the same as seen in control subjects. These data suggest that D3 receptors are increased by the disease but paradoxically decreased by administration of DA receptor antagonists; the latter point has been confirmed in rodent studies.

In contrast to the increase in D3 receptor-binding sites reported in the ventral striatum of schizophrenics, D3 receptor mRNA levels in the cortex appear to be lower in schizophrenic subjects. However, the levels of a truncated form of the D3 receptor are not changed, suggesting that abnormal splicing of a D3 gene may lead to decreased levels of the D3 receptor with a relative accumulation of the truncated receptor [22].

The D3 receptor has several alleles. Relative abundance of one of the D3 allelic variants is associated with an increased risk for development of tardive dyskinesia after antipsychotic drug treatment.

The D4 receptor differs from the D2 and D3 receptors by displaying a very high affinity for the atypical antipsychotic drug clozapine as well as conventional neuroleptics such as haloperidol. The D4 receptor is relatively enriched in cortical regions, including the hippocampus, of primates but present in very low abundance in the striatum [15]. In view of the high affinity of the D4 receptor for clozapine, there has been considerable effort expended on studying the expression of this receptor in schizophrenia.

Despite the low density of the D4 receptor in the striatum, several groups have used the difference between the density of striatal DA receptor sites as revealed by [³H]raclopride binding, which labels D2 and D3 receptors, and [³H]YM- 09151 binding, which labels D2, D3 and D4 receptors, as an indirect index of D4 receptor density. Using this approach, most studies have found a significant increase in the density of D4-like DA receptors in the striatum [23]. The development of specific D4 receptor antagonists lagged behind the cloning of the receptor and hampered the direct assessment of D4 receptor density in schizophrenics. Several specific antagonists are now available, and studies of D4 receptor density in the schizophrenic brain are in progress.

Despite the apparent increase in D4 receptor-binding sites, as revealed by the subtractive autoradiographic method, studies using polymerase chain reaction amplification to detect D4 receptor transcripts have failed to detect any change from normal in D4 gene expression in the striatum or several cortical regions of schizophrenic subjects. More recently, however, an in situ hybridization study reported that D4 mRNA is increased in the orbitofrontal, but not other prefrontal, cortical regions [21], emphasizing that regional differences may be striking. The development of selective D4 receptor antagonists has recently culminated in clinical trials. However, the open trials reported to date have not found antipsychotic efficacy of these D4 antagonists.

Several physiological and pharmacological studies in rodents have noted that the actions of DA and certain D2-like agonists, such as, quinpirole, cannot be fully accounted for by binding to one of the five cloned DA receptors. This seems to be particularly true in cortical regions, such as the prefrontal and entorhinal cortices. The genes encoding some D2-like DA receptors, including those present in the renal medulla and brown adipose tissue of rodents, have not been cloned. Molecular biological studies of striatal DA receptors have been interpreted to suggest that the reported striatal D4 receptor differs from the cloned D4 receptor. While it remains to be determined to what degree this site corresponds to the D4 receptor, there are sufficient data to suggest that other D2-like DA receptors may be identified and found to contribute to the pathophysiology of schizophrenia [24].

Less attention has focused on potential changes in D1 receptors than on D2-like receptors, primarily due to the high correlation between antipsychotic efficacy and D2, but not D1, receptor affinities. Those data that have been reported do not offer much support for the involvement of D1 receptors in schizophrenia. A relatively large number of antipsychotic drugs lack appreciable affinity for the D1 receptor in vivo. Moreover, selective D1 antagonists, all of which target D5 receptors as well, have not been shown in clinical studies to have antipsychotic effects. In addition, there does not appear to be any change in the density of D1 receptor-binding sites in the striatum. However, a recent PET study of D1 receptor occupancy in schizophrenic patients has suggested that D1, but not D2, receptors are decreased in the prefrontal cortex; this report awaits independent confirmation.

Prefrontal cortical dopaminergic hypoactivity. The past decade has marked a turn from the view that hyperactivity of central DA systems is the major defect in schizophrenia. In addition, recent technical advances have allowed investigators to probe the dynamic function of the DA system rather than relying on static measures. Studies on the neurochemical basis of schizophrenia have focused on uncovering increases in concentrations of DA or its metabolites or concentrations of DA receptors. However, awareness of trans-synaptic regulation of striatal DA systems led investigators to consider the possibility of an increase in striatal DA function being secondary to, or accompanied by, a decrease in functional DA tone in cortical regions that project onto the striatum.

A series of studies a generation ago reported that rats sustaining experimental depletions of DA in the prefrontal cortex showed increases in dopaminergic function in the striatal complex. Although it has proven difficult to replicate exactly the original reports, subsequent studies have consistently found that depletion of prefrontal cortical DA leads to an enhanced responsiveness of striatal DA systems in response to various challenges; these effects are observed primarily in the nucleus accumbens [25]. Among the challenges that evoke such changes in striatal DA function are stress, psychostimulant drugs and DA receptor antagonists [25]. Behavioral studies also have shown that changes in DA in the prefrontal cortex lead to behavioral changes that can be linked to the cognitive deficits present in schizophrenia. Interestingly, several behavioral studies have suggested that deviation in either direction from some ideal level of DA function in the prefrontal cortex may be deleterious.

Most animal studies examining corticostriatal relationships after cortical DA depletions have used indirect measures of trans-synaptic changes in downstream targets, such as the striatum, as the dependent measure. However, there are few direct data in human studies that indicate a decrease in DA levels in the prefrontal cortex. As noted earlier, one report has suggested that D1 receptor density may be decreased in the prefrontal cortex. In addition, recent immunohistochemical data have suggested a decrease in the density of both tyrosine hydroxylase— and DA transporter—immunoreactive axons in the prefrontal cortex of schizophrenic subjects [26].

The concept of concurrent cortical hypodopaminergic state and subcortical hyperdopaminergic tone has profound implications for our current understanding of schizophrenia and the treatment of the disorder [25,27]. As noted earlier, symptomatology in schizophrenia can be viewed as positive or negative. Positive symptoms, such as hallucinations and delusions, are typically sensitive to antipsychotic drug treatment, whereas negative symptoms, such as motivational difficulties, impoverished emotional state and withdrawal, are difficult to treat with conventional antipsychotic drugs but in many cases respond relatively well to cloza-pine [28]. Negative symptoms have been linked to a decrease in cortical DA tone, while positive symptoms are thought to reflect excessive dopaminergic tone in subcortical sites such as the nucleus accumbens [25,27]. The ability of clozapine to target negative symptoms may be related to its ability to sharply increase extracellular DA levels in the prefrontal cortex [29] or to its relatively high 5-HT₂:D2 receptor affinity [5].

Changes in evoked dopamine release. Studies of changes in receptor density or concentrations of DA or DA metabolites are static measures, providing a snapshot of the status of DA systems at a given point in time. However, many systems exhibit normal baseline function but are dysfunctional when tested under conditions that perturb the system. Until recently it has not been feasible to study dynamic neurochemical changes in the schizophrenic brain.

The availability of contemporary in vivo imaging techniques has led to recent studies specifically aimed at uncovering abnormalities in the dynamic function of central DA systems (see Chap. 54). By monitoring D2 receptor occupancy with a radioligand that is readily displaced from the DA receptor by endogenous DA, it is possible to monitor changes in release of DA. Two recent studies, using [¹¹C]raclopride and [¹²³I]iodobenzamide for PET and SPECT, respectively, have found that amphetamine-evoked DA release is increased in the striatum of schizophrenic subjects [30,31]. These data strongly suggest a change in phasic DA release under demand conditions. Since several stimuli, including stress as well as psychostimulants, can transiently increase DA release in the forebrain, examination of the effects of behavioral or pharmacological challenges to the DA system will be a major focus of investigation of neurobiological studies of schizophrenia.

Serotonin has been implicated in a variety of behaviors and somatic functions which are disturbed in schizophrenia

These include hallucinations, cognition, sensory gating, mood, aggression, sexual drive, appetite, motor activity, pain threshold, endocrine function and sleep [32]. It also has an important role in neurodevelopment. Many of these functions are clearly relevant to the etiology of positive and negative symptoms and of the cognitive impairment, which constitute the core abnormalities of schizophrenia. The biochemical and anatomical complexity and diversity of the serotonergic system and its extensive interactions with multiple neurotransmitters provide the physiological substrate for the ability of 5-HT to influence all of these behaviors. There are 14 known 5-HT receptor subtypes, two presynaptic and 12 postsynaptic, that mediate the multiple actions of 5-HT. Serotonin has potent influences on dopaminergic and glutamatergic neurotransmission via its action on 5-HT_1A, 5-HT_2A and 5-HT₃ receptors, in particular. These interactions occur at the level of DA neurons in the ventral tegmentum and substantia nigra and of 5-HT neurons in the medial and dorsal raphe, as well as at various projection fields of these nuclei. Serotonin may have an overall inhibitory effect on dopaminergic neurotransmission in the basal ganglia and prefrontal cortex. However, stimulatory effects have also been reported, depending on the nature of the influences on dopaminergic neurotransmission (Chap. 13).

The first hypothesis of an involvement of 5-HT in the etiology of schizophrenia was based on the psychotomimetic effect of lysergic acid diethylamide (LSD), which was found to be an antagonist at brain 5-HT receptors. This led to the hypothesis that decreased serotonergic activity was related to the positive symptoms of schizophrenia [32]. It was subsequently found that this effect of LSD was due to its 5-HT_2A agonist properties, not 5-HT antagonism, since the ability to produce visual hallucinations of a large number of indolealkylamine drugs with agonist activity is highly correlated with their affinity for this receptor [32]. Furthermore, the primary effect of LSD in humans is visual hallucinations, which are relatively rare in schizophrenia. None of the other core symptoms of schizophrenia or the cognitive dysfunction described previously are reliably produced by LSD and related drugs. Decreasing serotonergic activity with the 5-HT synthesis inhibitor parachlorophenylalanine does not exacerbate schizophrenia or produce psychosis in normals [32,33]. As discussed below, some studies indicate decreased 5-HT_2A receptor density in the cortex of schizophrenics.

The 5-HT deficiency hypothesis was followed by the hypothesis that endogenous production of psychotomimetic indoleamines, such as N,N-dimethyltryptamine or psilocybin, might be etiological in schizophrenia. However, the concentrations of such compounds in the brain, plasma and urine of patients with schizophrenia are not increased nor has an increase in the N-methyltransferase enzyme required for the synthesis of some of these compounds been found [32]. However, the decrease in the density of 5-HT_2A receptors in schizophrenia is consistent with downregulation secondary to an increase in stimulation of this receptor by 5-HT or related compounds.

The major hypothesis concerning the role of 5-HT in schizophrenia is based, in part, on the interest in the role of 5-HT in the mechanism of action of drugs such as clozapine, olanzapine, risperidone, ziprasidone, sertindole and M 100907, formerly MDL 100907. As previously mentioned, these drugs are called atypical antipsychotic drugs because they produce no (clozapine) or diminished motor side effects at clinically effective doses. Furthermore, they have varying advantages for treating the cognitive dysfunction and positive and negative symptoms of patients with schizophrenia. Clozapine is effective in treating the positive symptoms of up to 60% of those patients with schizophrenia whose positive symptoms fail to respond to typical neuroleptic drugs. The basis for this advantage has been the subject of intensive research interest. One feature which they share is potent 5-HT_2A receptor antagonism relative to weak D2 receptor antagonism, which contrasts with the typical neuroleptic drugs [34]. The selective 5-HT_2A antagonist M 100907 has been reported to be effective for both positive and negative symptoms in early clinical trials; these results need to be confirmed by further controlled study. There is considerable evidence that 5-HT_2A receptor antagonism can modulate dopaminergic and glutamatergic neurotransmission. This is consistent with the hypothesis that there is increased 5-HT_2A-mediated neurotransmission in schizo-phrenia which may be normalized by these agents [28]. This, in turn, may contribute to decreased glutamatergic activity via the recently discovered 5-HT_2A receptors on the apical dendrites of pyramidal neurons in the frontal cortex, which project to the mesolimbic regions implicated in schizophrenia. They may also contribute to the decreased dopaminergic activity in the prefrontal cortex which has been suggested to be a factor in the etiology of negative symptoms. There are also strong influences of 5-HT on the cholinergic system which may be relevant to the cognitive impairment in schizophrenia. Clozapine has been shown to increase the extracellular concentration of ACh in the frontal cortex of rats [35]. This may explain its ability to improve cognition because it is also a potent antimuscarinic agent which would be expected to lead to impaired memory function.

The serotonin system and its interactions. No major differences in CSF concentrations of 5-hydroxyindoleacetic acid (5-HIAA), the major metabolite of 5-HT, between schizophrenic patients and controls have been reported. However, a significant inverse relationship between CSF 5-HIAA concentrations and brain atrophy, as determined by CT, has been found [32]. Consistent with the hypothesis that the interaction between 5-HT and DA is most relevant to schizophrenia, neither CSF HVA nor 5-HIAA concentrations alone correlated with ventricular brain ratio, a measure of white matter loss, which is often increased in schizophrenia, whereas the ratio of HVA to 5-HIAA was negatively correlated with ventricular brain ratio [32,35]. This suggests that structural brain abnormalities in schizophrenia might be associated with a relative increase in serotonergic, compared to dopaminergic, activity.

The 5-HT hypothesis of schizophrenia predicts that patients should have abnormal responses to serotonergic challenge agents. Thus, it has been shown that the temperature and hormone response to MK-212, a 5-HT_2A/2C agonist, and MCPP, a 5-HT_2C agonist, are blunted in unmedicated schizophrenic patients. These results are consistent with the hypothesis that 5-HT_2A/2C responsivity is diminished in schizophrenia and the postmortem data of decreased 5-HT_2A receptor densities in schizophrenia [36]. Enhanced response to the 5-HT_1A agonist ipsapirone has also been reported.

A number of studies have found a decrease in 5-HT_2A receptor binding in various cortical areas [36]. Decreased 5-HT_2A receptor mRNA was also found in the dorsolateral prefrontal, superior temporal, anterior cingulate and striate cortices of patients with schizophrenia. These decreases could be due to downregulation of 5-HT_2A receptors. Since 5-HT_2A receptor density is decreased by 5-HT_2A receptor stimulation, this may be the result of increased 5-HT_2A receptor activity. However, the possibility that antipsychotic drug treatment is the cause of these decreases in 5-HT_2A receptor density has not been fully excluded.

There is also evidence for an increase in the density of 5-HT_1A receptors in the prefrontal, orbital and temporal cortex and in the hippocampus in postmortem specimens in patients with schizophrenia [36]. 5-HT_1A receptors are located pre- and postsynaptically on glutamatergic neurons. Increased glutamate-uptake sites, kainate, N-methyl-d aspartate (NMDA) and 5-HT receptors have been found in the same individuals, leading to the suggestion that the abnormalities in these two neurotransmitters are linked. It was suggested that 5-HT_1A modulation of glutamatergic activity might be abnormal in schizophrenia [37].

5-HT_1A agonists and 5-HT_2A antagonists usually have synergistic, but sometimes opposite, effects on a variety of biochemical systems [38]. It is noteworthy that both decreased 5-HT_2A and increased 5-HT_1A receptor-binding sites have been reported in the dorsolateral prefrontal cortex in schizophrenia. These two abnormalities together might produce synergistic influences on cortical processes, for example, glutamate outflow to subcortical areas such as the nucleus accumbens. An imbalance in the 5-HT_1A to 5-HT_2A receptor ratio could also contribute to abnormalities in the function of cortical association pathways.

Derangements in central nervous system GABA systems have long been suspected in schizophrenia

Although several reports have suggested that CSF concentrations of GABA are decreased in schizophrenia, there are as many negative reports as positive ones. Early reports suggested that the activity of GAD is decreased in several brain regions in schizophrenia and that high-affinity GABA uptake is altered in medial temporal lobe structures. However, these studies have also been difficult to replicate.

The dual role of GABA as a transmitter and as an integral part of intermediary metabolism have clouded the significance of changes in GABA concentration in post-mortem samples from schizophrenia. Despite long-standing speculation concerning the role of GABA in schizophrenia, conflicting biochemical data in various postmortem studies have dampened enthusiasm for its importance in the pathophysiology of the disease. However, recent anatomical studies, focusing on cortical areas, have led to a resurgence in interest in the role of GABA in the pathophysiology of schizophrenia. The differ-ent types of cortical GABAergic interneurons have dictated an emphasis on chemical neuroanatomy of the GABAergic system rather than neurochemical studies of GABA synthesis and metabolism. Specifically, reports of a lamina-specific decrease in the number of small (presumptive) interneurons in the pregenual an-terior cingulate and prefrontal cortices of schizophrenic patients have been most influential [39]. In addition, several reports have suggested an upregulation of benzodiazepine binding in the cortex in a regionally specific fashion [40]; these reports fit well with clinical literature documenting the benefits of benzodiazepine augmentation of conventional neuroleptic treatment.

The number of GAD55 mRNA-expressing neurons in the dorsolateral prefrontal cortex in schizophrenia have been reported to be decreased, but a corresponding decrease in the number of small (presumptive) interneurons in the same cortical region was not observed [41]. Moreover, the number of prefrontal cortical neurons expressing NADPH diaphorase, which is a marker for nitric oxide synthase—containing cortical neurons representing a small subset of GABAergic interneurons, was sharply decreased in prefrontal cortical gray matter but markedly increased in the underlying white matter [42]. This observation suggests a failure of specific interneurons to migrate to their normal cortical targets. Finally, recent data suggest that expression of the GABA transporter GAT-1 is decreased in GABAergic terminals in the prefrontal cortex.

The difficulties in obtaining consistent data concerning the numbers of interneurons as reflected by measurements of cell size or markers for all interneurons, such as GAD protein or transcripts, may be due to these markers identifying all of the several different types of cortical interneurons. Such global measures would hamper the ability to detect a consistent significant change in a discrete subpopulation of interneurons. Accordingly, attention has recently shifted to studies of markers of subpopulations of interneurons. Cortical interneurons can be categorized on the basis of expression of three different calcium-binding proteins: parvalbumin, calbindin and calretinin. Their distributions in the cortices are largely nonoverlapping, and there are morphological and physiological distinctions that are correlated with particular calcium-binding, protein-containing interneurons. Most studies have reported no changes in the number, density or laminar distribution of parvalbumin-containing prefrontal cortical interneurons; there also does not appear to be a change in the number of calretinin-containing interneurons. However, a single report has indicated an increase in the density of calbindin-containing interneurons. Despite the utility of defining interneurons on the basis of calcium-binding proteins, it should be recognized that each of the three classes of interneurons divided in this fashion can be further subdivided. Thus, further studies will be required to define changes in specific interneurons.

Any changes in GABAergic systems in schizophrenia, particularly changes observed in the cortex, may be a direct manifestation of some dysfunction of GABA neurons, such as derangement in the normal development and migration of GABA neurons. Alternatively, GABA dysfunction may be secondary to primary changes in afferents that regulate GABA neurons. Dopaminergic axons in the cortex synapse with both interneurons and pyramidal cells in the cortex and appear to excite certain GABAergic interneurons. If there is a decrease in the dopaminergic innervation of the prefrontal cortex in schizophrenia, it follows that certain changes in interneurons, such as a decrease in GAD67 mRNA, may be secondary to a primary dysfunction of the cortical DA innervation. Obviously, in studies of postmortem tissue aimed at uncovering possible GABA interactions with DA, it will be necessary to exclude any contribution of antipsychotic drugs. 5-HT axons synapse with and drive cortical interneurons, some of which express the 5-HT_2A receptor, and may, therefore, also trans-synaptically regulate interneurons.

Excitatory amino acids may play a role in the pathophysiology of schizophrenia

It has been two decades since the initial report of a decrease in CSF concentrations of glutamate and the elaboration of an hypothesis that schizophrenia may be due to decreased glutamatergic tone. During the period since publication of this paper, there has been an explosion of interest in the role of excitatory amino acids in neurodegeneration. More recently, two series of studies, one clinical and one preclinical, have led to an intense interest in derangements of excitatory amino acid systems as contributors to the pathogenesis of schizophrenia.

Several clinical studies have reported that the NMDA receptor antagonists phencyclidine (PCP) and, to a lesser extent, ketamine can evoke mental changes that resemble the negative and positive symptoms of schizophrenia in normal subjects and that exacerbate these symptoms in schizophrenics [43]. The reported ability to elicit both positive and negative symptoms is different from the response to chronic high-dose amphetamine administration, which produces mainly positive symptoms. However, it must be noted that the negative symptoms produced by PCP and ketamine, such as withdrawal and lack of motivations, are secondary to the disorganization and depersonalization produced by these drugs. Moreover, positive symptoms produced by amphetamines are rarely the classic paranoia and auditory hallucinations of schizophrenia. More interesting is the cognitive impairment produced by ketamine and PCP. A counterpoint to the findings that NMDA antagonists can result in a schizophreniform psychosis has been the attempt to pharmacologically modify a presumed hypoglutamatergic tone, as suggested by the PCP and ketamine studies, by administering drugs that act at the glycine modulatory site of the NMDA receptor. Several reports indicate that administration of glycine and d-cycloserine, usually as neuroleptic augmentation strategies, improves symptomatology in schizophrenic subjects [44].

Concurrent with studies examining the consequences of NMDA receptor antagonists in humans were studies of the effects of these antagonists in laboratory animals. Acute administration to adult, but not neonatal, rats of competitive and noncompetitive NMDA antagonists results in a circumscribed loss of neurons but does not induce gliosis; the pattern of neuronal, but not glial, involvement is the hallmark of pathological studies of schizophrenia. NMDA antagonist-elicited neurotoxicity can be blocked by administration of clozapine and other atypical antipsychotic drugs [45]. The number of neurons that undergo degeneration after NMDA receptor antagonist treatment is quite small, and the areas most impacted are the posterior cingulate and retrosplenial cortices, with a very limited degeneration in other cortical regions, including the hippocampus and amygdala. Unfortunately, this distribution of neurodegenerative changes does not mirror the pathological changes that have been reported in schizophrenia. A recent study has followed dopaminergic markers and cognitive function in nonhuman primates subchronically treated with PCP and reported persistent decreases in DA utilization in the prefrontal cortices paired with specific cognitive deficits; the latter were reversed by administration of clozapine [46].

The endogenous processes that might lead to NMDA receptor-dependent degeneration are not known. Among the potential endogenous ligands that may occupy the NMDA receptor or regulate the function of NMDA receptors is N-acetylaspartyl glutamate (NAAG), which is found in neurons and antagonizes NMDA receptor-mediated events. NAAG is metabolized to glutamate and N-acetylaspartate by N-acetyl α-linked acidic dipeptidase (NAALADase). NAAG concentrations have been reported to be increased in the hippocampus of schizophrenic subjects, and NAALADase activity is decreased in the hippocampus and prefrontal cortex [47]. Either an increase in NAAG by blocking NMDA receptor-mediated events or the associated decrease in glutamate and aspartate would lead to the glutamatergic hypofunction.

There are a variety of complex interactions between excitatory amino acid systems and central DA systems. Since neuroleptics markedly alter various indices of glutamatergic function, it is difficult to eliminate their contribution to various postmortem measurements of amino acids. The use of tissue obtained from nonschizophrenic patients who are maintained on antipsychotic drugs is one strategy, but even this strategy cannot eliminate an interaction between medication and diagnosis. The difficulty in obtaining tissue from untreated patients or patients in whom neuroleptic treatment had been discontinued for a reasonable period of time prior to death continues to make it difficult to dissect state- and trait-dependent differences in excitatory amino acid metabolism.

While NMDA receptors have figured most prominently in hypotheses of the pathogenesis of schizophrenia, several studies have reported changes in various non-NMDA, α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) and kainic acid receptors as well. Among the most consistent findings are a decrease in AMPA and kainate receptor subunit mRNAs in neurons of the medial temporal lobe, especially the hippocampus. In addition, kainic acid receptor binding has been reported to be decreased in the prefrontal cortex of schizophrenic patients.

Neuropeptides that function as neurotransmitters may play a role in schizophrenia

Neuropeptides are usually colocalized with conventional transmitters; in many cases, two or more peptides and a classical transmitter may be present in certain cells (see Chap. 18). Thus, it is not surprising that neuropeptides have been the focus of considerable attention in studies of psychoses such as schizophrenia. Of the large number of known neuroactive peptides, neurotensin, cholecystokinin (CCK), neuropeptide Y (NPY) and somatostatin have been implicated in schizophrenia, to date.

Neurotensin is a tridecapeptide found in a large number of brain regions that are typically associated with schizophrenia, including medial temporal lobe structures, the prefrontal cortices, basal ganglia structures and the amygdala. In addition, neurotensin in the rodent is colocalized with DA in ventral tegmental area neurons that project to the prefrontal cortex and ventral striatum; colocalization of neurotensin and DA is much more restricted in midbrain neurons of primate species, including humans.

Several studies have reported decreased CSF neurotensin concentrations in schizo-phrenic patients, although there are negative reports as well. It has been suggested that decreased concentrations of neurotensin define those patients with prominent negative symptoms, and antipsychotic drug treatment tends to normalize CSF concentrations of neurotensin [48]. However, neuroleptics have minimal effect on negative symptoms. There have been few studies of neurotensin receptors, of which two have been cloned and at least one additional neurotensin receptor postulated on the basis of physiological and pharmacological data. [¹²⁵I] Neurotensin binding has been reported to be decreased in the entorhinal cortex of schizophrenic patients.

Part of the continued interest in neurotensin in schizophrenia is due to a large body of literature that has found that central administration results in changes very similar to those observed with atypical antipsychotic drugs [49]. Moreover, neurotensin and DA are colocalized in certain midbrain neurons, and in the striatum neurotensin expression in medium spiny neurons is regulated by antipsychotic drugs through a D2 receptor mechanism. However, it has been difficult to study the role that neurotensin may play in antipsychotic drug actions because of a paucity of pharmacological tools. The development of neurotensin agonists and antagonists that enter the CNS should help this situation.

Cholecystokinin (CCK) is colocalized with virtually all midbrain DA neurons in the rat and is present in neurons in other regions (including the entorhinal and prefrontal cortices, hippocampus and the amygdala) that have been suggested to undergo structural changes in schizophrenia. A fairly large number of clinical trials of the effects of CCK or CCK analogues, such as ceruletide, in schizophrenia have been performed. However, despite an initial flush of optimism, the great majority of placebo-controlled studies reported no significant effect of CCK, either alone or as an adjunct to conventional antipsychotic drugs. Similarly, controlled studies of small numbers of patients on the CCK antagonist proglumide failed to show efficacy.

Despite the overall lack of success in clinical trials, certain preclinical and postmortem studies suggest that CCK may play a role in the pathophysiology of schizophrenia. CCK is present in most, if not all, midbrain DA neurons of the rat, and CCK-B receptor antagonists have electrophysiological effects on these DA neurons that are strikingly similar to those of antipsychotic drugs.

There have been very few postmortem studies of CCK gene expression or CCK receptors in schizophrenia. One interesting series of studies noted that although midbrain DA neurons of the rat typically contain CCK, studies of human substantia nigra have failed to observe CCK mRNA-containing cells in this region. Strikingly, however, CCK mRNA-containing substantia nigra neurons were relatively abundant in samples from most schizophrenic subjects; this increase in CCK gene expression does not appear to be due to neuroleptic treatment. A single study has reported the number of CCK mRNA-containing cells in the entorhinal cortex to be decreased in schizophrenia.

Neuropeptide Y (NPY) and somatostatin concentrations have been found to be decreased, in several postmortem studies of schizophrenia, in the cortex [50], where these peptides are found in interneurons. In addition, decreased CSF concentrations of NPY have been reported. Cortical concentrations of NPY and somatostatin are markedly decreased in Alzheimer's disease and related dementias [50], and these decreases in schizophrenia are most often seen in studies of elderly patients or those with a marked cognitive decline. More studies are clearly warranted to define the contribution of changes in NPY and somatostatin to the cognitive dysfunction that is present in schizophrenia.

Acetylcholine has also been suggested to play a role in schizophrenia

This is based on the observation that extrapyramidal side effects elicited by treatment with antipsychotic drugs are often treated by administration of anticholinergic drugs. The rationale for anticholinergic treatment to reduce EPS is based on the fact that ACh is the transmitter of striatal interneurons that impinge on medium spiny neurons. Because of the intricate interrelationship between striatal DA and ACh, early hypotheses emphasized an imbalance between DA and ACh in tardive dyskinesia and schizophrenia.

Early studies focused on measures of ACh function in the striatum and reported that choline acetyltransferase (ChAT) activity was decreased in the nucleus accumbens of schizophrenics. Early studies also reported changes in acetylcholinesterase (AChE) activity in red blood cells and the nucleus accumbens of schizophrenic patients. There have been no controlled studies of such changes in unmedicated schizophrenics.

Several contemporary studies have implicated the prefrontal cortex as a potential site at which cholinergic dysfunction is manifested. This focus was dictated in part by the cognitive dysfunction seen in schizophrenia. Both ChAT and AChE activity are decreased in the cortex of schizophrenic patients who do not meet the pathological criteria for Alzheimer's disease. In addition to these neurochemical studies, recent postmortem studies have reported an increase in the number of pedunculopontine tegmental nucleus cells containing NAPDH diaphorase, which is present in pontine cholinergic cells, in schizophrenic subjects [51]. However, the level of pontine ChAT protein as measured by immunoblotting was decreased. The observation of an increase in the number of NADPH diaphorase neurons may be due to the use of this marker rather than a specific cholinergic marker such as ChAT. Nonetheless, in view of the function of these pontine reticular formation cholinergic neurons in attention and cognition, it will be important to unravel specific changes in the pontine cholinergic neurons. Recent studies have found that clozapine can increase extracellular concentrations of ACh in rat prefrontal cortex.

Schizophrenic subjects and unaffected relatives show a decrease in the P50 auditory evoked response to the second of two sequential tone presentations. Freedman et al. [52] found that the P50 deficit was linked to a locus on the long arm of chromosome 15. This appears to be the locus of the α7 nicotinic receptor subunit, which forms homomeric functional nicotinic ACh receptors that are characterized by their ability to bind α-bungarotoxin. Since deficits in the P50 evoked response had previously been linked to deficits in bungarotoxin binding in the interneurons of the hippocampus and since nicotine normalizes the deficit in sensory gating, deficits in α7 nicotinic ACh receptors on interneurons of the hippocampal complex have been suggested to be a predisposing factor to the development of schizophrenia.