Vaughan & Asburys General Ophthalmology, 17th ed.

Chapter 4. Psychosis and schizophrenia

   Symptom dimensions in schizophrenia

    Clinical description of psychosis

    Schizophrenia is more than a psychosis

    Beyond positive and negative symptoms of schizophrenia

    Symptoms of schizophrenia are not necessarily unique to schizophrenia

    Brain circuits and symptom dimensions in schizophrenia

   Neurotransmitters and circuits in schizophrenia

    Dopamine

    Glutamate

   Neurodevelopment and genetics in schizophrenia

   Neuroimaging circuits in schizophrenia

    Imaging genetics and epistasis

   Summary

Psychosis is a difficult term to define and is frequently misused, not only in the media but unfortunately among mental health professionals as well. Stigma and fear surround the concept of psychosis, and sometimes the pejorative term “crazy” is used for psychosis. This chapter is not intended to list the diagnostic criteria for all the different mental disorders in which psychosis is either a defining feature or an associated feature. The reader is referred to standard reference sources such as the DSM (Diagnostic and Statistical Manual) of the American Psychiatric Association and the ICD (International Classification of Diseases) for that information. Although schizophrenia is emphasized here, we will approach psychosis as a syndrome associated with a variety of illnesses that are all targets for antipsychotic drug treatment.

Symptom dimensions in schizophrenia

Clinical description of psychosis

Psychosis is a syndrome – that is, a mixture of symptoms – that can be associated with many different psychiatric disorders, but is not a specific disorder itself in diagnostic schemes such as the DSM or ICD. At a minimum, psychosis means delusions and hallucinations. It generally also includes symptoms such as disorganized speech, disorganized behavior, and gross distortions of reality.

Therefore, psychosis can be considered to be a set of symptoms in which a person’s mental capacity, affective response, and capacity to recognize reality, communicate, and relate to others is impaired. Psychotic disorders have psychotic symptoms as their defining features; there are other disorders in which psychotic symptoms may be present, but are not necessary for the diagnosis.

Those disorders that require the presence of psychosis as a defining feature of the diagnosis include schizophrenia, substance-induced (i.e., drug-induced) psychotic disorders, schizophreniform disorder, schizoaffective disorder, delusional disorder, brief psychotic disorder, and psychotic disorder due to a general medical condition (Table 4-1). Disorders that may or may not have psychotic symptoms as associated features include mania and depression as well as several cognitive disorders such as Alzheimer’s dementia (Table 4-2).

Table 4-1 Disorders in which psychosis is a defining feature


Schizophrenia

Substance-induced (i.e., drug-induced) psychotic disorders

Schizophreniform disorder

Schizoaffective disorder

Delusional disorder

Brief psychotic disorder

Psychotic disorder due to a general medical condition


Table 4-2 Disorders in which psychosis is an associated feature


Mania

Depression

Cognitive disorders

Alzheimer’s dementia


Psychosis itself can be paranoid, disorganized/excited, or depressive. Perceptual distortions and motor disturbances can be associated with any type of psychosis. Perceptual distortions include being distressed by hallucinatory voices; hearing voices that accuse, blame, or threaten punishment; seeing visions; reporting hallucinations of touch, taste or odor; or reporting that familiar things and people seem changed. Motor disturbances are peculiar, rigid postures; overt signs of tension; inappropriate grins or giggles; peculiar repetitive gestures; talking, muttering, or mumbling to oneself; or glancing around as if hearing voices.

In paranoid psychosis, the patient has paranoid projections, hostile belligerence and grandiose expansiveness. Paranoid projection includes preoccupation with delusional beliefs; believing that people are talking about oneself; believing one is being persecuted or being conspired against; and believing people or external forces control one’s actions. Hostile belligerence is verbal expression of feelings of hostility; expressing an attitude of disdain; manifesting a hostile, sullen attitude; manifesting irritability and grouchiness; tending to blame others for problems; expressing feelings of resentment; complaining and finding fault; as well as expressing suspicion of people. Grandiose expansiveness is exhibiting an attitude of superiority; hearing voices that praise and extol; believing one has unusual powers or is a well-known personality, or that one has a divine mission.

In a disorganized/excited psychosis there is conceptual disorganization, disorientation, and excitement. Conceptual disorganization can be characterized by giving answers that are irrelevant or incoherent, drifting off the subject, using neologisms, or repeating certain words or phrases. Disorientation is not knowing where one is, the season of the year, the calendar year, or one’s own age. Excitement is expressing feelings without restraint; manifesting speech that is hurried; exhibiting an elevated mood; an attitude of superiority; dramatizing oneself or one’s symptoms; manifesting loud and boisterous speech; exhibiting overactivity or restlessness; and exhibiting excess of speech.

Depressive psychosis is characterized by psychomotor retardation, apathy, and anxious self-punishment and blame. Psychomotor retardation and apathy are manifested by slowed speech; indifference to one’s future; fixed facial expression; slowed movements; deficiencies in recent memory; blocking in speech; apathy toward oneself or one’s problems; slovenly appearance; low or whispered speech; and failure to answer questions. Anxious self-punishment and blame is the tendency to blame or condemn oneself; anxiety about specific matters; apprehensiveness regarding vague future events; an attitude of self-deprecation, manifesting as a depressed mood; expressing feelings of guilt and remorse; preoccupation with suicidal thoughts, unwanted ideas, and specific fears; and feeling unworthy or sinful.

This discussion of clusters of psychotic symptoms does not constitute diagnostic criteria for any psychotic disorder. It is given merely as a description of several types of symptoms in psychosis to give the reader an overview of the nature of behavioral disturbances associated with the various psychotic illnesses.

Schizophrenia is more than a psychosis

Although schizophrenia is the commonest and best-known psychotic illness, it is not synonymous with psychosis, but is just one of many causes of psychosis. Schizophrenia affects 1% of the population, and in the US there are over 300 000 acute schizophrenic episodes annually. Between 25% and 50% of schizophrenia patients attempt suicide, and 10% eventually succeed, contributing to a mortality rate eight times greater than that of the general population. Life expectancy of a patient with schizophrenia may be 20–30 years shorter than the general population, not only due to suicide, but in particular due to premature cardiovascular disease. Accelerated mortality from premature cardiovascular disease in patients with schizophrenia is caused not only by genetic and lifestyle factors, such as smoking, unhealthy diet, and lack of exercise leading to obesity and diabetes, but also – sorrily – from treatment with some antipsychotic drugs which themselves cause an increased incidence of obesity and diabetes, and thus increase cardiac risk. In the US, over 20% of all social security benefits are used for the care of patients with schizophrenia. The direct and indirect costs of schizophrenia in the US alone are estimated to be in the tens of billions of dollars every year.

Schizophrenia by definition is a disturbance that must last for six months or longer, including at least one month of delusions, hallucinations, disorganized speech, grossly disorganized or catatonic behavior, or negative symptoms. Positive symptoms are listed in Table 4-3 and shown in Figure 4-1. These symptoms of schizophrenia are often emphasized, since they can be dramatic, can erupt suddenly when a patient decompensates into a psychotic episode (often called a psychotic “break,” as in break from reality), and are the symptoms most effectively treated by antipsychotic medications. Delusions are one type of positive symptom, and these usually involve a misinterpretation of perceptions or experiences. The most common content of a delusion in schizophrenia is persecutory, but it may include a variety of other themes including referential (i.e., erroneously thinking that something refers to oneself), somatic, religious, or grandiose. Hallucinations are also a type of positive symptom (Table 4-3) and may occur in any sensory modality (e.g., auditory, visual, olfactory, gustatory, and tactile), but auditory hallucinations are by far the most common and characteristic hallucinations in schizophrenia. Positive symptoms generally reflect an excess of normal functions, and in addition to delusions and hallucinations may also include distortions or exaggerations in language and communication (disorganized speech), as well as in behavioral monitoring (grossly disorganized or catatonic or agitated behavior). Positive symptoms are well known because they are dramatic, are often the cause of bringing a patient to the attention of medical professionals and law enforcement, and are the major target of antipsychotic drug treatments.



Figure 4-1. Positive and negative symptoms of schizophrenia. The syndrome of schizophrenia consists of a mixture of symptoms that are commonly divided into two major categories, positive and negative. Positive symptoms, such as delusions and hallucinations, reflect the development of the symptoms of psychosis; they can be dramatic and may reflect loss of touch with reality. Negative symptoms reflect the loss of normal functions and feelings, such as losing interest in things and not being able to experience pleasure.

Table 4-3 Positive symptoms of psychosis and schizophrenia


Delusions

Hallucinations

Distortions or exaggerations in language and communication

Disorganized speech

Disorganized behavior

Catatonic behavior

Agitation


Negative symptoms are listed in Tables 4-4 and 4-5 and shown in Figure 4-1. Classically, there are at least five types of negative symptoms all starting with the letter A (Table 4-5):

·        alogia – dysfunction of communication; restrictions in the fluency and productivity of thought and speech

·        affective blunting or flattening – restrictions in the range and intensity of emotional expression

·        asociality – reduced social drive and interaction

·        anhedonia – reduced ability to experience pleasure

·        avolition – reduced desire, motivation or persistence; restrictions in the initiation of goal-directed behavior

Table 4-4 Negative symptoms of schizophrenia


Blunted affect

Emotional withdrawal

Poor rapport

Passivity

Apathetic social withdrawal

Difficulty in abstract thinking

Lack of spontaneity

Stereotyped thinking

Alogia: restrictions in fluency and productivity of thought and speech

Avolition: restrictions in initiation of goal-directed behavior

Anhedonia: lack of pleasure

Attentional impairment


Table 4-5 What are negative symptoms?


Domain

Descriptive Term

Translation

Dysfunction of communication

Alogia

Poverty of speech; e.g., talks little, uses few words

Dysfunction of affect

Affective blunting

Reduced range of emotions (perception, experience and expression); e.g., feels numb or empty inside, recalls few emotional experiences, good or bad

Dysfunction of socialization

Asociality

Reduced social drive and interaction; e.g., little sexual interest, few friends, little interest in spending time with (or little time spent with) friends

Dysfunction of capacity for pleasure

Anhedonia

Reduced ability to experience pleasure; e.g., finds previous hobbies or interests unpleasurable

Dysfunction of motivation

Avolition

Reduced desire, motivation, persistence; e.g., reduced ability to undertake and complete everyday tasks; may have poor personal hygiene


Negative symptoms in schizophrenia, such as blunted affect, emotional withdrawal, poor rapport, passivity and apathetic social withdrawal, difficulty in abstract thinking, stereotyped thinking and lack of spontaneity, commonly are considered a reduction in normal functions and are associated with long periods of hospitalization and poor social functioning. Although this reduction in normal functioning may not be as dramatic as positive symptoms, it is interesting to note that negative symptoms of schizophrenia determine whether a patient ultimately functions well or has a poor outcome. Certainly, patients will have disruptions in their ability to interact with others when their positive symptoms are out of control, but their degree of negative symptoms will largely determine whether patients with schizophrenia can live independently, maintain stable social relationships, or re-enter the workplace.

Although formal rating scales can be used to measure negative symptoms in research studies, in clinical practice it may be more practical to identify and monitor negative symptoms quickly by observation alone (Figure 4-2) or by some simple questioning (Figure 4-3). Negative symptoms are not just part of the syndrome of schizophrenia – they can also be part of a “prodrome” that begins with subsyndromal symptoms that do not meet the diagnostic criteria of schizophrenia and occur before the onset of the full syndrome of schizophrenia. Prodromal negative symptoms are important to detect and monitor over time in high-risk patients so that treatment can be initiated at the first signs of psychosis. Negative symptoms can also persist between psychotic episodes once schizophrenia has begun, and reduce social and occupational functioning in the absence of positive symptoms.



Figure 4-2. Negative symptoms identified by observation. Some negative symptoms of schizophrenia – such as reduced speech, poor grooming, and limited eye contact – can be identified solely by observing the patient.



Figure 4-3. Negative symptoms identified by questioning. Other negative symptoms of schizophrenia can be identified by simple questioning. For example, brief questioning can reveal the degree of emotional responsiveness, interest level in hobbies or pursuing life goals, and desire to initiate and maintain social contacts.

Current antipsychotic drug treatments are limited in their ability to treat negative symptoms, but psychosocial interventions along with antipsychotics can be helpful in reducing negative symptoms. There is even the possibility that instituting treatment for negative symptoms during the prodromal phase of schizophrenia may delay or prevent the onset of the illness, but this is still a matter of current research.

Beyond positive and negative symptoms of schizophrenia

Although not recognized formally as part of the diagnostic criteria for schizophrenia, numerous studies subcategorize the symptoms of this illness into five dimensions: not just positive and negative symptoms, but also cognitive symptoms, aggressive symptoms, and affective symptoms (Figure 4-4). This is perhaps a more sophisticated, if complicated, manner of describing the symptoms of schizophrenia.



Figure 4-4. Localization of symptom domains. The different symptom domains of schizophrenia are hypothesized to be regulated by unique brain regions. Positive symptoms of schizophrenia are hypothetically modulated by malfunctioning mesolimbic circuits, while negative symptoms are hypothetically linked to malfunctioning mesocortical circuits and may also involve mesolimbic regions such as the nucleus accumbens, which is part of the brain’s reward circuitry and thus plays a role in motivation. The nucleus accumbens may also be involved in the increased rate of substance use and abuse seen in patients with schizophrenia. Affective symptoms are associated with the ventromedial prefrontal cortex, while aggressive symptoms (related to impulse control) are associated with abnormal information processing in the orbitofrontal cortex and amygdala. Cognitive symptoms are associated with problematic information processing in the dorsolateral prefrontal cortex. Although there is overlap in function among different brain regions, understanding which brain regions may be predominantly involved in specific symptoms can aid in customization of treatment to the particular symptom profile of each individual patient with schizophrenia.

Aggressive symptoms such as assaultiveness, verbally abusive behaviors, and frank violence can occur with positive symptoms such as delusions and hallucinations, and be confused with positive symptoms. Behavioral interventions may be particularly helpful to prevent violence linked to poor impulsivity by reducing provocations from the environment. Certain antipsychotic drugs such as clozapine, or very high doses of standard antipsychotic drugs, or occasionally the use of two antipsychotic drugs simultaneously, may also be useful for aggressive symptoms and violence in some patients.

It can also be difficult to separate the symptoms of formal cognitive dysfunction from the symptoms of affective dysfunction and from negative symptoms, but research is attempting to localize the specific areas of brain dysfunction for each symptom domain in schizophrenia in the hope of developing better treatments for the often-neglected negative, cognitive, and affective symptoms of schizophrenia. In particular, neuropsychological assessment batteries are being developed to quantify cognitive symptoms, in order to detect cognitive improvement after treatment with a number of novel psychotropic drugs currently being tested. Cognitive symptoms of schizophrenia are impaired attention and impaired information processing manifested as impaired verbal fluency (ability to produce spontaneous speech), problems with serial learning (of a list of items or a sequence of events), and impairment in vigilance for executive functioning (problems with sustaining and focusing attention, concentrating, prioritizing, and modulating behavior based upon social cues).

Important cognitive symptoms of schizophrenia are listed in Table 4-6. These do not include symptoms of dementia and memory disturbance more characteristic of Alzheimer’s disease, but cognitive symptoms of schizophrenia emphasize “executive dysfunction,” which includes problems representing and maintaining goals, allocating attentional resources, evaluating and monitoring performance, and utilizing these skills to solve problems. Cognitive symptoms of schizophrenia are important to recognize and monitor because they are the single strongest correlate of real-world functioning, even stronger than negative symptoms.

Table 4-6 Cognitive symptoms of schizophrenia


Problems representing and maintaining goals

Problems allocating attentional resources

Problems focusing attention

Problems sustaining attention

Problems evaluating functions

Problems monitoring performance

Problems prioritizing

Problems modulating behavior based upon social cues

Problems with serial learning

Impaired verbal fluency

Difficulty with problem solving


Symptoms of schizophrenia are not necessarily unique to schizophrenia

It is important to recognize that several illnesses other than schizophrenia can share some of the same five symptom dimensions as described here for schizophrenia and shown in Figure 4-4. Thus, disorders in addition to schizophrenia that can have positive symptoms include bipolar disorder, schizoaffective disorder, psychotic depression, Alzheimer’s disease and other organic dementias, childhood psychotic illnesses, drug-induced psychoses, and others. Negative symptoms can also occur in other disorders and can also overlap with cognitive and affective symptoms that occur in these other disorders. However, as a primary deficit state, negative symptoms are fairly unique to schizophrenia. Schizophrenia is certainly not the only disorder with cognitive symptoms. Autism, post-stroke (vascular or multi-infarct) dementia, Alzheimer’s disease, and many other organic dementias (Parkinsonian/Lewy body dementia, frontotemporal/Pick’s dementia, etc.) can also be associated with cognitive dysfunctions similar to those seen in schizophrenia.

Affective symptoms are frequently associated with schizophrenia but this does not necessarily mean that they fulfill the diagnostic criteria for a comorbid anxiety or affective disorder. Nevertheless, depressed mood, anxious mood, guilt, tension, irritability, and worry frequently accompany schizophrenia. These various symptoms are also prominent features of major depressive disorder, psychotic depression, bipolar disorder, schizoaffective disorder, organic dementias, childhood psychotic disorders, and treatment-resistant cases of depression, bipolar disorder, and schizophrenia, among others. Finally, aggressive and hostile symptoms occur in numerous other disorders, especially those with problems of impulse control. Symptoms include overt hostility, such as verbal or physical abusiveness or assault, self-injurious behaviors including suicide, and arson or other property damage. Other types of impulsiveness such as sexual acting out are also in this category of aggressive and hostile symptoms. These same symptoms are frequently associated with bipolar disorder, childhood psychosis, borderline personality disorder, antisocial personality disorder, drug abuse, Alzheimer’s and other dementias, attention deficit hyperactivity disorder, conduct disorders in children, and many others.

Brain circuits and symptom dimensions in schizophrenia

The various symptoms of schizophrenia are hypothesized to be localized in unique brain regions (Figure 4-4). Specifically, the positive symptoms of schizophrenia have long been hypothesized to be localized to malfunctioning mesolimbic circuits, especially involving the nucleus accumbens. The nucleus accumbens is considered to be part of the brain’s reward circuitry, so it is not surprising that problems with reward and motivation in schizophrenia, symptoms that can overlap with negative symptoms and lead to smoking, drug and alcohol abuse, may be linked to this brain area as well. The prefrontal cortex is considered to be a key node in the nexus of malfunctioning cerebral circuitry responsible for each of the remaining symptoms of schizophrenia: specifically, the mesocortical and ventromedial prefrontal cortex with negative symptoms and affective symptoms, the dorsolateral prefrontal cortex with cognitive symptoms, and the orbitofrontal cortex and its connections to amygdala with aggressive, impulsive symptoms (Figure 4-4).

This model is obviously oversimplified and reductionistic, because every brain area has several functions, and every function is certainly distributed to more than one brain area. Nevertheless, allocating specific symptom dimensions to unique brain areas not only assists research studies, but has both heuristic and clinical value. Specifically, every patient has unique symptoms, and unique responses to medication. In order to optimize and individualize treatment, it can be useful to consider which specific symptoms any given patient is expressing, and therefore which areas of that particular patient’s brain are hypothetically malfunctioning (Figure 4-4). Each brain area has unique neurotransmitters, receptors, enzymes, and genes that regulate it, with some overlap, but also with some unique regional differences, and knowing this can assist the clinician in choosing medications and in monitoring the effectiveness of treatment.

Neurotransmitters and circuits in schizophrenia

Dopamine

The leading hypothesis for schizophrenia is based upon the neurotransmitter dopamine. To understand the potential role of dopamine in schizophrenia, it is first important to review how dopamine is synthesized, metabolized, and regulated; and the role of dopamine receptors and the localization of key dopamine pathways in the brain.

Dopaminergic neurons

Dopaminergic neurons utilize the neurotransmitter dopamine (DA), which is synthesized in dopaminergic nerve terminals from the amino acid tyrosine after it is taken up into the neuron from the extracellular space and bloodstream by a tyrosine pump, or transporter (Figure 4-5). Tyrosine is converted into DA first by the rate-limiting enzyme tyrosine hydroxylase (TOH) and then by the enzyme DOPA decarboxylase (DDC) (Figure 4-5). DA is then taken up into synaptic vesicles by a vesicular monoamine transporter (VMAT2) and stored there until it is used during neurotransmission.



Figure 4-5. Dopamine synthesis. Tyrosine (TYR), a precursor to dopamine, is taken up into dopamine nerve terminals via a tyrosine transporter and converted into DOPA by the enzyme tyrosine hydroxylase (TOH). DOPA is then converted into dopamine (DA) by the enzyme DOPA decarboxylase (DDC). After synthesis, dopamine is packaged into synaptic vesicles via the vesicular monoamine transporter (VMAT2) and stored there until its release into the synapse during neurotransmission.

The DA neuron has a presynaptic transporter (reuptake pump) called DAT, which is unique for DA and which terminates DA’s synaptic action by whisking it out of the synapse back into the presynaptic nerve terminal; there it can be re-stored in synaptic vesicles for subsequent reuse in another neurotransmission (Figure 4-6). DATs are not present in high density at the axon terminals of all DA neurons. For example, in prefrontal cortex, DATs are relatively sparse and DA is inactivated by other mechanisms. Excess DA that escapes storage in synaptic vesicles can be destroyed within the neuron by the enzymes monoamine oxidase (MAO)-A or MAO-B, or outside the neuron by the enzyme catechol-O-methyl-transferase (COMT) (Figure 4-6). DA that diffuses away from synapses can also be transported by norepinephrine transporters (NETs) as a “false” substrate, and DA action will be terminated in this manner.



Figure 4-6. Dopamine's action is terminated. Dopamine’s action can be terminated through multiple mechanisms. Dopamine can be transported out of the synaptic cleft and back into the presynaptic neuron via the dopamine transporter (DAT), where it may be repackaged for future use. Alternatively, dopamine may be broken down extracellularly via the enzyme catechol-O-methyl-transferase (COMT). Other enzymes that break down dopamine are monoamine oxidase A (MAO-A) and monoamine oxidase B (MAO-B), which are present in mitochondria within the presynaptic neuron and in other cells such as glia.

Receptors for dopamine also regulate dopaminergic neurotransmission (Figure 4-7). The DA transporter DAT and the vesicular transporter VMAT2 are both types of receptors. A plethora of additional dopamine receptors exist, including at least five pharmacological subtypes and several more molecular isoforms. Perhaps the most extensively investigated dopamine receptor is the dopamine 2 (D2) receptor, as it is stimulated by dopamine agonists for the treatment of Parkinson’s disease, and blocked by dopamine antagonist antipsychotics for the treatment of schizophrenia. As will be discussed in greater detail in Chapter 5 on antipsychotic drugs, dopamine 1, 2, 3, and 4 receptors are all blocked by some atypical antipsychotic drugs, but it is not clear to what extent dopamine 1, 3, or 4 receptors contribute to the clinical properties of these drugs.



Figure 4-7. Dopamine receptors. Shown here are receptors for dopamine that regulate its neurotransmission. The dopamine transporter (DAT) exists presynaptically and is responsible for clearing excess dopamine out of the synapse. The vesicular monoamine transporter (VMAT2) takes dopamine up into synaptic vesicles for future neurotransmission. There is also a presynaptic dopamine D2 autoreceptor, which regulates release of dopamine from the presynaptic neuron. In addition, there are several postsynaptic receptors. These include D1, D2, D3, D4, and D5 receptors. The functions of the D2 receptors are best understood, because this is the primary binding site for virtually all antipsychotic agents as well as for dopamine agonists used to treat Parkinson’s disease.

Dopamine 2 receptors can be presynaptic, where they function as autoreceptors (Figure 4-7). Presynaptic D2 receptors thus act as “gatekeepers,” either allowing DA release when they are not occupied by DA (Figure 4-8A) or inhibiting DA release when DA builds up in the synapse and occupies these gatekeeping presynaptic autoreceptors (Figure 4-8B). Such receptors are located either on the axon terminal (Figure 4-9) or on the other end of the neuron in the somatodendritic area (Figure 4-10). In both cases, occupancy of these D2 receptors provides negative feedback input, or a braking action upon the release of dopamine from the presynaptic neuron.



Figure 4-8. Presynaptic dopamine 2 (D2) autoreceptors. Presynaptic D2 autoreceptors are “gatekeepers” for dopamine. That is, when these gatekeeping receptors are not bound by dopamine (no dopamine in the gatekeeper’s hand), they open a molecular gate, allowing dopamine release (A). However, when dopamine binds to the gatekeeping receptors (now the gatekeeper has dopamine in his hand), they close the molecular gate and prevent dopamine from being released (B).



Figure 4-9. Presynaptic dopamine-2 autoreceptors. Presynaptic D2 autoreceptors can be located on the axon terminal, as shown here. When dopamine builds up in the synapse (A), it is available to bind to the autoreceptor, which then inhibits dopamine release (B).



Figure 4-10. Somatodendritic dopamine-2 autoreceptors. D2 autoreceptors can also be located in the somatodendritic area, as shown here (A). When dopamine binds to the receptor here, it shuts off neuronal impulse flow in the dopamine neuron (see loss of lightning bolts in the neuron in B), and this stops further dopamine release.

Key dopamine pathways in the brain

The five dopamine pathways in the brain are shown in Figure 4-11. They include the mesolimbic dopamine pathway, the mesocortical dopamine pathway, the nigrostriatal dopamine pathway, the tuberoinfundibular dopamine pathway, and a fifth pathway that innervates the thalamus.



Figure 4-11. Five dopamine pathways in the brain. The neuroanatomy of dopamine neuronal pathways in the brain can explain the symptoms of schizophrenia as well as the therapeutic effects and side effects of antipsychotic drugs. (a) The nigrostriatal dopamine pathway, which projects from the substantia nigra to the basal ganglia or striatum, is part of the extrapyramidal nervous system and controls motor function and movement. (b) The mesolimbic dopamine pathway projects from the midbrain ventral tegmental area to the nucleus accumbens, a part of the limbic system of the brain thought to be involved in many behaviors such as pleasurable sensations, the powerful euphoria of drugs of abuse, as well as delusions and hallucinations of psychosis. (c) A pathway related to the mesolimbic dopamine pathway is the mesocortical dopamine pathway. It also projects from the midbrain ventral tegmental area but sends its axons to areas of the prefrontal cortex, where they may have a role in mediating cognitive symptoms (dorsolateral prefrontal cortex, DLPFC) and affective symptoms (ventromedial prefrontal cortex, VMPFC) of schizophrenia. (d) The fourth dopamine pathway of interest, the tuberoinfundibular dopamine pathway, projects from the hypothalamus to the anterior pituitary gland and controls prolactin secretion. (e) The fifth dopamine pathway arises from multiple sites, including the periaqueductal gray, ventral mesencephalon, hypothalamic nuclei, and lateral parabrachial nucleus, and it projects to the thalamus. Its function is not currently well known.

The dopamine hypothesis of schizophrenia: the mesolimbic dopamine pathway and positive symptoms of schizophrenia

The mesolimbic dopamine pathway projects from dopaminergic cell bodies in the ventral tegmental area of the brainstem to axon terminals in one of the limbic areas of the brain, namely the nucleus accumbens in the ventral striatum (Figure 4-11). This pathway is thought to have an important role in several emotional behaviors, including the positive symptoms of psychosis, such as delusions and hallucinations (Figure 4-12). The mesolimbic dopamine pathway also is important for motivation, pleasure, and reward.



Figure 4-12. Mesolimbic dopamine pathway. The mesolimbic dopamine pathway, which projects from the ventral tegmental area in the brainstem to the nucleus accumbens in the ventral striatum (A), is involved in regulation of emotional behaviors and is believed to be the predominant pathway regulating positive symptoms of psychosis. Specifically, hyperactivity of this pathway is believed to account for delusions and hallucinations (B).

For more than 40 years, it has been observed that diseases or drugs that increase dopamine will enhance or produce positive psychotic symptoms, whereas drugs that decrease dopamine will decrease or stop positive symptoms. For example, stimulant drugs such as amphetamine and cocaine release dopamine, and if given repetitively can cause a paranoid psychosis virtually indistinguishable from the positive symptoms of schizophrenia. Stimulant drugs are discussed in detail in subsequent chapters on treatment of attention deficit hyperactivity disorder, and on drug abuse.

All known antipsychotic drugs capable of treating positive psychotic symptoms are blockers of the dopamine D2 receptor. Antipsychotic drugs are discussed in Chapter 5. These observations have been formulated into a theory of psychosis sometimes referred to as the “dopamine hypothesis of schizophrenia.” Perhaps a more precise modern designation is the “mesolimbic dopamine hypothesis of positive symptoms of schizophrenia,” since it is believed that it is hyperactivity specifically in this particular dopamine pathway that mediates the positive symptoms of psychosis (Figure 4-13). Hyperactivity of the mesolimbic dopamine pathway hypothetically accounts for positive psychotic symptoms whether those symptoms are part of the illness of schizophrenia, or of drug-induced psychosis, or whether they are positive psychotic symptoms accompanying mania, depression, or dementia. Hyperactivity of mesolimbic dopamine neurons may also play a role in aggressive and hostile symptoms in schizophrenia and related illnesses, especially if serotonergic control of dopamine is aberrant in patients who lack impulse control. Although it is not known what causes this mesolimbic dopamine hyperactivity, current theories now state that it is the downstream consequence of dysfunction in prefrontal cortex and hippocampal glutamate activity, as will be discussed below.



Figure 4-13. Mesolimbic dopamine hypothesis. Hyperactivity of dopamine neurons in the mesolimbic dopamine pathway theoretically mediates the positive symptoms of psychosis such as delusions and hallucinations. This pathway is also involved in pleasure, reward, and reinforcing behavior, and many drugs of abuse interact here.

The mesocortical dopamine pathway and cognitive, negative, and affective symptoms of schizophrenia

Another pathway also arising from cell bodies in the ventral tegmental area, but projecting to areas of the prefrontal cortex, is known as the mesocortical dopamine pathway (Figures 4-14 and 4-15). Branches of this pathway into the dorsolateral prefrontal cortex are hypothesized to regulate cognition and executive functions (Figure 4-14), whereas branches of this pathway into the ventromedial parts of the prefrontal cortex are hypothesized to regulate emotions and affect (Figure 4-15). The exact role of the mesocortical dopamine pathway in mediating symptoms of schizophrenia is still a matter of debate, but many researchers believe that cognitive and some negative symptoms of schizophrenia may be due to a deficit of dopamine activity in mesocortical projections to dorsolateral prefrontal cortex (Figure 4-14), whereas affective and other negative symptoms of schizophrenia may be due to a deficit of dopamine activity in mesocortical projections to ventromedial prefrontal cortex (Figure 4-15).



Figure 4-14. Mesocortical pathway to dorsolateral prefrontal cortex. Another major dopaminergic pathway is the mesocortical dopamine pathway, which projects from the ventral tegmental area to the prefrontal cortex (A). Projections specifically to the dorsolateral prefrontal cortex (DLPFC) are believed to be involved in the negative and cognitive symptoms of schizophrenia. In this case, expression of these symptoms is thought to be associated with hypoactivity of this pathway (B).



Figure 4-15. Mesocortical pathway to ventromedial prefrontal cortex. Mesocortical dopamine projections specifically to the ventromedial prefrontal cortex (VMPFC) are believed to mediate negative and affective symptoms associated with schizophrenia (A). These symptoms are believed to arise from hypoactivity in this pathway (B).

The behavioral deficit state suggested by negative symptoms certainly implies underactivity or lack of proper functioning of mesocortical dopamine projections that may be the consequence of neurodevelopmental abnormalities in the NMDA (N-methyl-D-aspartate) glutamate system, described in the next section. Whatever the cause, a corollary to the original DA hypothesis of schizophrenia now incorporates theories for the cognitive, negative, and affective symptoms, and might be more precisely designated as the “mesocortical dopamine hypothesis of cognitive, negative, and affective symptoms of schizophrenia,” since it is believed that it is underactivity specifically in mesocortical projections to prefrontal cortex that mediate the cognitive, negative, and affective symptoms of schizophrenia (Figure 4-16).



Figure 4-16. Mesocortical dopamine hypothesis of negative, cognitive, and affective symptoms of schizophrenia. Hypoactivity of dopamine neurons in the mesocortical dopamine pathway theoretically mediates the cognitive, negative, and affective symptoms of schizophrenia.

Theoretically, increasing dopamine in the mesocortical dopamine pathway might improve negative, cognitive, and affective symptoms of schizophrenia. However, since there is hypothetically an excess of dopamine elsewhere in the brain – within the mesolimbic dopamine pathway – any further increase of dopamine in that pathway would actually worsen positive symptoms. Thus, this state of affairs for dopamine activity in the brain of patients with schizophrenia poses a therapeutic dilemma: how do you increase dopamine in the mesocortical pathway while simultaneously decreasing dopamine activity in the mesolimbic dopamine pathway? The extent to which atypical antipsychotics have provided a solution to this therapeutic dilemma will be discussed in Chapter 5.

Mesolimbic dopamine pathway, reward and negative symptoms

When a patient with schizophrenia loses motivation and interest, and has anhedonia and lack of pleasure, such symptoms could also implicate a deficient functioning of the mesolimbic dopamine pathway, not just deficient functioning in the mesocortical dopamine pathway. This idea is supported by observations that treating patients with antipsychotics, particularly the conventional antipsychotics, can produce a worsening of negative symptoms and a state of “neurolepsis” that looks very much like negative symptoms of schizophrenia. Since the prefrontal cortex does not have a high density of D2 receptors, this implicates possible deficient functioning within the mesolimbic dopamine system causing inadequate reward mechanisms, exhibited as behaviors such as anhedonia and drug abuse, as well as negative symptoms, exhibited as lack of rewarding social interactions, and lack of general motivation and interest. Perhaps the much higher incidence of substance abuse in schizophrenia than in healthy adults, especially of nicotine but also of stimulants and other substances of abuse, could be partially explained as an attempt to boost the function of defective mesolimbic dopaminergic pleasure centers, perhaps at the cost of activating positive symptoms.

Nigrostriatal dopamine pathway

Another key dopamine pathway in the brain is the nigrostriatal dopamine pathway, which projects from dopaminergic cell bodies in the brainstem substantia nigra via axons terminating in the basal ganglia or striatum (Figure 4-17). The nigrostriatal dopamine pathway is a part of the extrapyramidal nervous system, and controls motor movements. Deficiencies in dopamine in this pathway cause movement disorders including Parkinson’s disease, characterized by rigidity, akinesia/bradykinesia (i.e., lack of movement or slowing of movement), and tremor. Dopamine deficiency in the basal ganglia also can produce akathisia (a type of restlessness), and dystonia (twisting movements especially of the face and neck). These movement disorders can be replicated by drugs that block D2 receptors in this pathway, and this will be discussed briefly in Chapter 5.



Figure 4-17. Nigrostriatal dopamine pathway. The nigrostriatal dopamine pathway projects from the substantia nigra to the basal ganglia or striatum. It is part of the extrapyramidal nervous system and plays a key role in regulating movements. When dopamine is deficient, it can cause parkinsonism with tremor, rigidity, and akinesia/bradykinesia. When DA is in excess, it can cause hyperkinetic movements such as tics and dyskinesias. In untreated schizophrenia, activation of this pathway is believed to be “normal.”

Hyperactivity of dopamine in the nigrostriatal pathway is thought to underlie various hyperkinetic movement disorders such as chorea, dyskinesias, and tics. Chronic blockade of D2 receptors in this pathway may result in a hyperkinetic movement disorder known as neuroleptic-induced tardive dyskinesia. This will also be discussed briefly in Chapter 5. In schizophrenia, the nigrostriatal pathway in untreated patients may be relatively preserved (Figure 4-17).

Tuberoinfundibular dopamine pathway

The dopamine neurons that project from hypothalamus to anterior pituitary are part of the tuberoinfundibular dopamine pathway (Figure 4-18). Normally, these neurons are active and inhibit prolactin release. In the postpartum state, however, the activity of these dopamine neurons is decreased. Prolactin levels can therefore rise during breastfeeding so that lactation will occur. If the functioning of tuberoinfundibular dopamine neurons is disrupted by lesions or drugs, prolactin levels can also rise. Elevated prolactin levels are associated with galactorrhea (breast secretions), amenorrhea (loss of ovulation and menstrual periods), and possibly other problems such as sexual dysfunction. Such problems can occur after treatment with many antipsychotic drugs that block D2 receptors, and will be discussed further in Chapter 5. In untreated schizophrenia, the function of the tuberoinfundibular pathway may be relatively preserved (Figure 4-18).



Figure 4-18. Tuberoinfundibular dopamine pathway. The tuberoinfundibular dopamine pathway from the hypothalamus to the anterior pituitary regulates prolactin secretion into the circulation. Dopamine inhibits prolactin secretion. In untreated schizophrenia, activation of this pathway is believed to be “normal.”

Thalamic dopamine pathway

Recently, a dopamine pathway that innervates the thalamus in primates has been described. It arises from multiple sites, including the periaqueductal gray matter, the ventral mesencephalon, various hypothalamic nuclei, and the lateral parabrachial nucleus (Figure 4-11). Its function is still under investigation, but it may be involved in sleep and arousal mechanisms by gating information passing through the thalamus to the cortex and other brain areas. There is no evidence at this point for abnormal functioning of this dopamine pathway in schizophrenia.

Glutamate

In recent years, the neurotransmitter glutamate has attained a key theoretical role in the hypothesized pathophysiology of schizophrenia, as well as in a number of other psychiatric disorders, including depression. It is also now a key target of novel psychopharmacologic agents for future treatments of schizophrenia and depression. In order to understand theories about glutamate in schizophrenia and other psychiatric disorders, how the malfunctioning of glutamate systems impacts dopamine systems in schizophrenia, and how glutamate systems might become important targets of new therapeutic drugs for schizophrenia, it is necessary to review the regulation of glutamate neurotransmission. Glutamate is the major excitatory neurotransmitter in the central nervous system and sometimes considered to be the “master switch” of the brain, since it can excite and turn on virtually all CNS neurons. The synthesis, metabolism, receptor regulation, and key pathways of glutamate are therefore critical to the functioning of the brain and will be reviewed here.

Glutamate synthesis

Glutamate or glutamic acid is a neurotransmitter which is an amino acid. Its predominant use is not as a neurotransmitter, but as an amino acid building block for protein biosynthesis. When used as a neurotransmitter, it is synthesized from glutamine in glia, which also assist in the recycling and regeneration of more glutamate following glutamate release during neurotransmission. When glutamate is released from synaptic vesicles stored within glutamate neurons, it interacts with receptors in the synapse and is then taken up into neighboring glia by a reuptake pump known as an excitatory amino acid transporter (EAAT) (Figure 4-19A). The presynaptic glutamate neuron and the postsynaptic site of glutamate neurotransmission may also have EAATs (not shown in the figures), but these EAATs do not appear to play as important a role in glutamate recycling and regeneration as the EAATs in glia (Figure 4-19A).



A. Glutamate is recycled and regenerated, part 1. After release of glutamate from the presynaptic neuron (1), it is taken up into glial cells via the EAAT, or excitatory amino acid transporter (2).



B. Glutamate is recycled and regenerated, part 2. Once inside the glial cell, glutamate is converted into glutamine by the enzyme glutamine synthetase (3).



C. Glutamate is recycled and regenerated, part 3. Glutamine is released from glial cells by a specific neutral amino acid transporter (glial SNAT) through the process of reverse transport (4), and then taken up by SNATs on glutamate neurons (5).



D. Glutamate is recycled and regenerated, part 4. Glutamine is converted into glutamate within the presynaptic glutamate neuron by the enzyme glutaminase (6) and taken up into synaptic vesicles by the vesicular glutamate transporter (vGluT), where it is stored for future release (7).

Figure 4-19

After reuptake into glia, glutamate is converted into glutamine inside the glia by an enzyme known as glutamine synthetase (arrow 3 in Figure 4-19B). It is possible that glutamate is not simply reused, but rather converted into glutamine, to keep it in a pool for neurotransmitter use, rather than being lost into the pool for protein synthesis. Glutamine is released from glia by reverse transport out of them by a pump or transporter known as a specific neutral amino acid transporter, (SNAT, arrow 4 in Figure 4-19C). Glutamine may also be transported out of glia by a second transporter known as a glial alanine-serine-cysteine transporter or ASC-T (not shown). When glial SNATs and ASC-Ts operate in the inward direction, they transport glutamine and other amino acids into glia. Here, they are reversed so that glutamine can get out of the glia and hop a ride into a neuron via a different type of neuronal SNAT, operating inwardly in a reuptake manner (arrow 5 in Figure 4-19C).

Once inside the neuron, glutamine is converted back into glutamate for use as a neurotransmitter by an enzyme in mitochondria called glutaminase (arrow 6 in Figure 4-19D). Glutamate is then transported into synaptic vesicles via a vesicular glutamate transporter (vGluT, arrow 7 in Figure 4-19D) where it is stored for subsequent release during neurotransmission. Once it is released, glutamate’s actions are stopped not by enzymatic breakdown, as in other neurotransmitter systems, but by removal by EAATs on neurons or glia, and the whole cycle is started again (Figures 4-19A through D).

Synthesis of glutamate cotransmitters glycine and D-serine

Glutamate systems are curious in that one of the key receptors for glutamate requires a cotransmitter in addition to glutamate in order to function. That receptor is the NMDA (N-methyl-D-aspartate) receptor, described below, and the cotransmitter is either the amino acid glycine (Figure 4-20), or another amino acid closely related to glycine, known as D-serine (Figure 4-21).



Figure 4-20. N-methyl-D-aspartate (NMDA) receptor cotransmitter glycine is produced. Glutamate’s actions at NMDA receptors are dependent in part upon the presence of a cotransmitter, either glycine or D-serine. Glycine can be derived directly from dietary amino acids and transported into glial cells either by a glycine transporter (GlyT1) or by a specific neutral amino acid transporter (SNAT). Glycine can also be produced both in glycine neurons and in glial cells. Glycine neurons provide only a small amount of the glycine at glutamate synapses, because most of the glycine released by glycine neurons is used only at glycine synapses and then taken back up into presynaptic glycine neurons via the glycine 2 transporter (GlyT2) before much glycine can diffuse to glutamate synapses. Glycine produced by glial cells plays a larger role at glutamate synapses. Glycine is produced in glial cells when the amino acid L-serine is taken up into glial cells via the L-serine transporter (L-SER-T), and then converted into glycine by the enzyme serine hydroxymethyl-transferase (SHMT). Glycine from glial cells is released into the glutamate synapse through reverse transport by the glycine 1 transporter (GlyT1). Extracellular glycine is then transported back into glial cells via a reuptake pump, namely GlyT1.



Figure 4-21. NMDA receptor cotransmitter D-serine is produced. Glutamate requires the presence of either glycine or D-serine at NMDA receptors in order to exert some of its effects there. In glial cells, the enzyme serine racemase converts L-serine into D-serine, which is then released into the glutamate synapse via reverse transport on the glial D-serine transporter (glial D-SER-T). L-Serine’s presence in glial cells is a result either of its transport there via the L-serine transporter (L-SER-T) or of its conversion into L-serine from glycine via the enzyme serine hydroxymethyl-transferase (SHMT). Once D-serine is released into the synapse, it is taken back up into the glial cell by a reuptake pump called D-SER-T. Excess D-serine within the glial cell can be destroyed by the enzyme D-amino acid oxidase (DAO), which converts D-serine into hydroxypyruvate (OH-pyruvate).

Glycine is not known to be synthesized by glutamate neurons, so glutamate neurons must get the glycine they need for their NMDA receptors either from glycine neurons or from glia (Figure 4-20). Glycine neurons release glycine, but they contribute only a small amount of glycine to glutamate synapses; glycine is unable to diffuse very far from neighboring glycine neurons because the glycine they release is taken back up into those neurons by a type of glycine reuptake pump known as the type 2 glycine transporter or GlyT2 (Figure 4-20).

Thus, neighboring glia are thought to be the source of most of the glycine available for glutamate synapses. Glycine itself can be taken up into glia as well as into glutamate neurons from the synapse by a type 1 glycine transporter or GlyT1 (Figure 4-20). Glycine can also be taken up into glia by a glial SNAT (specific neutral amino acid transporter). Glycine is not known to be stored within synaptic vesicles of glia, but as we will learn below, the companion neurotransmitter D-serine is thought possibly to be stored within some type of synaptic vesicle within glia. Glycine in the cytoplasm of glia is nevertheless somehow available for release into synapses, and it escapes from glial cells by riding outside them and into the glutamate synapse on a reversed GlyT1 transporter (Figure 4-20). Once outside, glycine can get right back into the glia by an inwardly directed GlyT1, which functions as a reuptake pump and is the main mechanism responsible for terminating the action of synaptic glycine (Figure 4-20). GlyT1 transporters are probably also located on the glutamate neuron, but any release or storage from the glutamate neuron is not well characterized (Figure 4-20). Later, in Chapter 5, we will discuss novel treatments for schizophrenia that boost glycine action, and thus glutamate action, at NMDA receptors. Such treatments are in clinical testing and include inhibitors of the key glycine transporter GlyT1, called selective glycine reuptake inhibitors or SGRIs.

Glycine can also be synthesized from the amino acid L-serine, derived from the extracellular space, bloodstream, and diet, transported into glialcells by an L-serine transporter (L-SER-T), and converted from L-serine into glycine by the glial enzyme serine hydroxymethyl-transferase (SHMT) (Figure 4-20). This enzyme works in both directions, either converting L-serine into glycine, or glycine into L-serine.

How is the cotransmitter D-serine produced? D-Serine is unusual in that it is a D-amino acid, whereas the 20 known essential amino acids are all L-amino acids, including D-serine’s mirror-image amino acid L-serine. It just so happens that D-serine has high affinity for the glycine site on NMDA receptors, and that glia are equipped with an enzyme called D-serine racemase that can convert regular L-serine into the neurotransmitting amino acid D-serine and vice versa (Figure 4-21). Thus, D-serine can be derived either from glycine or from L-serine, both of which can be transported into glia by their own transporters, and then glycine converted to L-serine by the enzyme SHMT, and finally L-serine converted into D-serine by the enzyme D-serine racemase (Figure 4-21). Interestingly, the D-serine so produced may be stored in some sort of vesicle in glia for subsequent release on a reversed glial D-serine transporter (D-SER-T) for neurotransmitting purposes at glutamate synapses containing NMDA receptors. D-Serine’s actions are not only terminated by synaptic reuptake via the inwardly acting glial D-SER-T, but also by an enzyme, D-amino acid oxidase (DAO), that converts D-serine into inactive hydroxypyruvate (Figure 4-21). Below, we will discuss how the brain makes an activator of DAO, known not surprisingly as D-amino acid oxidase activator or DAOA. The gene that makes DAOA may be one of the important regulatory genes that contribute to the genetic basis of schizophrenia, as will be explained below in the section on the neurodevelopmental hypothesis of schizophrenia.

Glutamate receptors

There are several types of glutamate receptors (Figure 4-22 and Table 4-7), including the neuronal presynaptic reuptake pump (EAAT or excitatory amino acid transporter) and the vesicular transporter for glutamate into synaptic vesicles (vGluT), both of which are types of receptors. The properties of various transporters are discussed in Chapter 2. Shown also on the presynaptic neuron as well as the postsynaptic neuron are metabotropic glutamate receptors (Figure 4-22). Metabotropic glutamate receptors are those glutamate receptors which are linked to G proteins. The properties of G-protein-linked receptors are also discussed in Chapter 2.



Figure 4-22. Glutamate receptors. Shown here are receptors for glutamate that regulate its neurotransmission. The excitatory amino acid transporter (EAAT) exists presynaptically and is responsible for clearing excess glutamate out of the synapse. The vesicular transporter for glutamate (vGluT) transports glutamate into synaptic vesicles, where it is stored until used in a future neurotransmission. Metabotropic glutamate receptors (linked to G proteins) can occur either pre- or postsynaptically. Three types of postsynaptic glutamate receptors are linked to ion channels, and are known as ligand-gated ion channels: N-methyl-D-aspartate (NMDA) receptors, α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) receptors, and kainate receptors, all named for the agonists that bind to them.

Table 4-7 Glutamate receptors



There are at least eight subtypes of metabotropic glutamate receptors, organized into three separate groups (Table 4-7). Research suggests that group II and group III metabotropic receptors can occur presynaptically, where they function as autoreceptors to block glutamate release (Figure 4-23). Drugs that stimulate these presynaptic autoreceptors as agonists may therefore reduce glutamate release and be potentially useful as anticonvulsants and mood stabilizers, and may also protect against glutamate excitotoxicity, as will be explained below. Group I metabotropic glutamate receptors may be located predominantly postsynaptically, where they hypothetically interact with other postsynaptic glutamate receptors to facilitate and strengthen responses mediated by ligand-gated ion-channel receptors for glutamate during excitatory glutamatergic neurotransmission (Figure 4-22).



Figure 4-23. Metabotropic glutamate autoreceptors. Groups II and III metabotropic glutamate receptors can exist presynaptically as autoreceptors to regulate the release of glutamate. When glutamate builds up in the synapse (A), it is available to bind to the autoreceptor, which then inhibits glutamate release (B).

NMDA (N-methyl-D-aspartate), AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid), and kainate receptors for glutamate, named after the agonists that selectively bind to them, are all members of the ligand-gated ion-channel family of receptors (Figure 4-22 and Table 4-7). These ligand-gated ion channels are also known as ionotropic receptors and ion-channel-linked receptors. The properties of ligand-gated ion channels are discussed in Chapter 3. They tend to be postsynaptic and work together to modulate excitatory postsynaptic neurotransmission triggered by glutamate. Specifically, AMPA and kainate receptors may mediate fast, excitatory neurotransmission, allowing sodium to enter the neuron to depolarize it (Figure 4-24). NMDA receptors in the resting state are normally blocked by magnesium, which plugs a calcium channel (Figure 4-25). NMDA receptors are an interesting type of “coincidence detector” that can open to let calcium into the neuron to trigger postsynaptic actions from glutamate neurotransmission only when three things occur at the same time: glutamate occupies its binding site on the NMDA receptor, glycine or D-serine binds to its site on the NMDA receptor, and depolarization occurs, allowing the magnesium plug to be removed (Figures 4-25 and 4-26). Some of the many important signals by NMDA receptors that are activated when NMDA calcium channels are opened include long-term potentiation and synaptic plasticity, as will be explained later in this chapter.



Figure 4-24. Glutamate at AMPA and kainate receptors. Unlike NMDA receptors, AMPA and kainate receptors require only glutamate to bind in order for the channel to open. This leads to fast excitatory neurotransmission and membrane depolarization. Sustained binding of the agonist glutamate will lead to receptor desensitization, causing the channel to close and be transiently unresponsive to agonist.



Figure 4-25. Magnesium as a negative allosteric modulator. Magnesium is a negative allosteric modulator (NAM) at NMDA glutamate receptors. Opening of NMDA glutamate receptors requires the presence of both glutamate and glycine, each of which binds to a different site on the receptor. When magnesium is also bound and the membrane is not depolarized, it prevents the effects of glutamate and glycine and thus does not allow the ion channel to open. In order for the channel to open, depolarization must remove magnesium while both glutamate and glycine are bound to their sites on the ligand-gated ion-channel complex.



Figure 4-26. Signal propagation via glutamate receptors. (A) On the left is an AMPA receptor with its sodium channel in the resting state, allowing minimal sodium to enter the cell in exchange for potassium. On the right is an NMDA receptor with magnesium blocking the calcium channel and glycine bound to its site. (B) When glutamate arrives, it binds to the AMPA receptor, causing the sodium channel to open, thus increasing the flow of sodium into the dendrite and of potassium out of the dendrite. This causes the membrane to depolarize and triggers a postsynaptic nerve impulse. (C) Depolarization of the membrane removes magnesium from the calcium channel. This, coupled with glutamate binding to the NMDA receptor in the presence of glycine, causes the NMDA receptor to open and allow calcium influx. Calcium influx through NMDA receptors contributes to long-term potentiation, a phenomenon that may be involved in long-term learning, synaptogenesis, and other neuronal functions.

Key glutamate pathways in the brain

Glutamate is a ubiquitous excitatory neurotransmitter that seems to be able to excite nearly any neuron in the brain. That is why it is sometimes called the “master switch.” Nevertheless, there are about a half-dozen specific glutamatergic pathways that are of particular relevance to psychopharmacology, and especially to the pathophysiology of schizophrenia (Figure 4-27). They are:

(a) Cortico-brainstem

(b) Cortico-striatal

(c) Hippocampal-striatal

(d) Thalamo-cortical

(e) Cortico-thalamic

(f) Cortico-cortical (direct)

(g) Cortico-cortical (indirect)

(a) Cortico-brainstem glutamate pathways. A very important descending glutamatergic pathway projects from cortical pyramidal neurons to brainstem neurotransmitter centers, including the raphe for serotonin, the ventral tegmental area (VTA) and substantia nigra for dopamine, and the locus coeruleus for norepinephrine (pathway a in Figure 4-27). This pathway is the cortico-brainstem glutamate pathway, and it is a key regulator of neurotransmitter release. Direct innervation of monoamine neurons in the brainstem by these excitatory cortico-brainstem glutamate neurons stimulates neurotransmitter release, whereas indirect innervation of monoamine neurons by these excitatory cortico-glutamate neurons via GABA interneurons in the brainstem blocksneurotransmitter release.

(b) Cortico-striatal glutamate pathways. A second descending glutamatergic output from cortical pyramidal neurons projects to the striatal complex (pathway b in Figure 4-27). This pathway is known as the cortico-striatal glutamate pathway when it projects to the dorsal striatum, or the cortico-accumbens glutamate pathway when it projects to a specific area of the ventral striatum known as the nucleus accumbens. In either case, these descending glutamate pathways terminate on GABA neurons destined for a relay station in another part of the striatal complex called the globus pallidus.

(c) Hippocampal-accumbens glutamate pathway. Another key glutamate pathway projects from the hippocampus to the nucleus accumbens and is known as the hippocampal-accumbens glutamate pathway (c in Figure 4-27). Specific theories link this particular pathway to schizophrenia (see below). Like the cortico-striatal and cortico-accumbens glutamate pathways (b in Figure 4-27), the hippocampal glutamate projection to the nucleus accumbens also terminates on GABA neurons there that in turn project to a relay station in the globus pallidus.

(d) Thalamo-cortical glutamate pathway. This pathway (d in Figure 4-27) brings information from the thalamus back into the cortex, often to process sensory information.

(e) Cortico-thalamic glutamate pathway. A fifth glutamate pathway, known as the cortico-thalamic glutamate pathway, projects directly back to the thalamus (pathway e in Figure 4-27), where it may direct the manner in which neurons react to sensory information.

(f) Direct cortico-cortical glutamate pathways. Finally, a complex of many cortico-cortical glutamate pathways are present within the cortex (pathways f and g in Figure 4-27). On the one hand, pyramidal neurons can excite each other within the cerebral cortex via direct synaptic input from their own neurotransmitter glutamate (f in Figure 4-27).

(g) Indirect cortico-cortical glutamate pathways. On the other hand, one pyramidal neuron can inhibit another via indirect input, namely via interneurons that release GABA (g in Figure 4-27).



Figure 4-27. Glutamate pathways in the brain. Although glutamate can have actions at virtually all neurons in the brain, there are key glutamate pathways particularly relevant to schizophrenia. (a) The cortico-brainstem glutamate projection is a descending pathway that projects from cortical pyramidal neurons in the prefrontal cortex to brainstem neurotransmitter centers (raphe, locus coeruleus, ventral tegmental area, substantia nigra) and regulates neurotransmitter release. (b) Another descending glutamatergic pathway projects from the prefrontal cortex to the striatum (cortico-striatal glutamate pathway) and to the nucleus accumbens (cortico-accumbens glutamate pathway), and constitutes the “cortico-striatal” portion of cortico-striato-thalamic loops. (c) There is also a glutamatergic projection from the ventral hippocampus to the nucleus accumbens. (d) Thalamo-cortical glutamate pathways are pathways that ascend from the thalamus and innervate pyramidal neurons in the cortex. (e) Cortico-thalamic glutamate pathways descend from the prefrontal cortex to the thalamus. (f) Intracortical pyramidal neurons can communicate directly with each other via the neurotransmitter glutamate; these pathways are known as cortico-cortical glutamatergic pathways. (g) Intracortical pyramidal neurons can also communicate via GABAergic interneurons.

The NMDA hypofunction hypothesis of schizophrenia: ketamine and phencyclidine

A major current hypothesis for the cause of schizophrenia proposes that glutamate activity at NMDA receptors is hypofunctional due to abnormalities in the formation of glutamatergic NMDA synapses during neurodevelopment. This so-called “NMDA receptor hypofunction hypothesis of schizophrenia” arises in part from observations that when NMDA receptors are made hypofunctional by means of the NMDA receptor antagonists PCP (phencyclidine) or ketamine (Figure 4-28), this produces a psychotic condition in normal humans very similar to symptoms of schizophrenia. Hypothetically, genetic abnormalities also make NMDA receptors and their synapses hypofunctional in order to cause schizophrenia itself. Amphetamine, which releases dopamine, also produces a psychotic condition of delusions and hallucinations in normal humans similar to the positive symptoms of schizophrenia. What is so attractive about the NMDA receptor hypofunction hypothesis of schizophrenia is that unlike amphetamine, which activates only positive symptoms, PCP and ketamine also mimic the cognitive, negative, and affective symptoms of schizophrenia such as social withdrawal and executive dysfunction. Another attractive aspect of the NMDA hypofunction hypothesis is that it can also explain the dopamine hypothesis of schizophrenia, namely, as a downstream consequence of hypofunctioning NMDA receptors.



Figure 4-28. Site of action of PCP and ketamine. The anesthetic ketamine binds to the open channel conformation of the N-methyl-D-aspartate (NMDA) receptor. Specifically, it binds to a site within the calcium channel of this receptor, which is often termed the PCP site because it is also where phencyclidine (PCP) binds. Blockade of NMDA receptors may prevent the excitatory actions of glutamate.

The NMDA hypofunction hypothesis of schizophrenia: faulty NMDA synapses on GABA interneurons in prefrontal cortex

Although NMDA receptors and synapses are ubiquitous throughout the brain, and PCP or ketamine block all of them, a current leading theory of schizophrenia suggests that schizophrenia may be caused by neurodevelopmental abnormalities in the formation of glutamate synapses at a specific site: namely, at certain GABA interneurons in the cerebral cortex (see g in Figure 4-27, and also box 1 in both Figure 4-29A and Figure 4-29B). Something appears to be wrong with the genetic programming of those particular GABA interneurons that can be identified in prefrontal cortex as containing a calcium binding protein called parvalbumin (Figure 4-29B). These parvalbumin-containing GABA interneurons appear to be faulty postsynaptic partners to incoming glutamate input from pyramidal neurons in prefrontal cortex, and to form defective NMDA receptor containing synaptic connections with incoming pyramidal neurons (Figure 4-29B, box 1; compare Figure 4-29A, box 1). Thus, they have hypofunctioning NMDA receptors on their dendrites, defective synapses between the glutamate neuronal axons and the GABA interneuronal dendrites, and thus faulty glutamatergic information coming in to the GABA interneuron (Figure 4-29B, box 1). This so-called “dysconnectivity” may be genetically programmed from a variety of faulty genes that all converge on the formation of these particular NMDA synapses.



A. Hypothetical site of glutamate dysfunction in schizophrenia, part 1. Shown here is a close-up of cortical pyramidal neurons communicating via GABAergic interneurons. (1) Glutamate is released from an intracortical pyramidal neuron and binds to an NMDA receptor on a GABAergic interneuron. (2) GABA is then released from the interneuron and binds to GABA receptors of the α2 subtype that are located on the axon of another glutamate pyramidal neuron. (3) This inhibits the pyramidal neuron, thus reducing the release of downstream glutamate.



B. Hypothetical site of glutamate dysfunction in schizophrenia, part 2. Shown here is a close-up of cortical pyramidal neurons communicating via GABAergic interneurons in the presence of hypofunctional NMDA receptors. (1) Glutamate is released from an intracortical pyramidal neuron. However, the NMDA receptor to which it binds is hypofunctional, preventing glutamate from exerting its full effects via the NMDA receptor. (2) This prevents GABA release from the interneuron; thus, stimulation of α2 GABA receptors on the axon of another glutamate neuron does not occur. (3) When GABA does not bind to the α2 GABA receptors on its axon, the pyramidal neuron is no longer inhibited. Instead, it is disinhibited and overactive, releasing excessive glutamate downstream.

Figure 4-29

Parvalbumin-containing GABA interneurons in the prefrontal cortex of patients with schizophrenia have other problems as a consequence of this dysconnectivity, in that they also have deficits in the enzyme that makes their own neurotransmitter GABA (namely, decreased activity of GAD67 (glutamic acid decarboxylase)), causing a compensatory increase in the postsynaptic amount of α2-subunit-containing GABAA receptors in the postsynaptic axon initial segment of the pyramidal neurons they innervate (Figure 4-29B, box 2; compare Figure 4-29A, box 2).

What are the consequences of the hypothetical dysconnectivity of glutamate with these particular GABA interneurons? When parvalbumin-containing GABA interneurons fail to function properly, they do not adequately inhibit key glutamatergic pyramidal neurons in the prefrontal cortex, causing those glutamate neurons to become hyperactive (Figure 4-29B box 3; compare Figure 4-29A box 3). This hypothetically disrupts the functioning of downstream neurons, especially dopamine neurons (Figures 4-30B4-31B, and 4-32B, explained below). So, one sick synapse in a neuronal circuit can affect the whole circuit, from GABA interneuron and the glutamate neurons it innervates, to downstream dopamine neurons and beyond.

Linking the NMDA hypofunction hypothesis of schizophrenia with the dopamine hypothesis of schizophrenia: positive symptoms

A complex set of interactions allows glutamate to determine dopamine release. Most relevant to schizophrenia are the glutamate pathways that regulate the mesolimbic and mesocortical dopamine pathways shown in Figures 4-11through 4-16. Cortico-brainstem glutamate pathways regulate the output of glutamate from the cortex to the brainstem neurotransmitter center known as the ventral tegmental area (VTA) for both the mesolimbic dopamine projection (pathway a in Figure 4-27 and in Figure 4-30A) and for the mesocortical dopamine projections (pathway a in Figure 4-27 and in Figure 4-32A).

First, we will discuss the glutamate regulation of mesolimbic dopamine neurons (Figure 4-30). It appears that the cortico-brainstem glutamate neurons that innervate only the dopamine neurons projecting from the VTA to the nucleus accumbens – i.e., the mesolimbic dopamine pathway – directly innervate those specific dopamine neurons (Figure 4-30A), and thus stimulate them. You can imagine what would happen if these upstream glutamate neurons were too active (Figures 4-29B and 4-30B): they would cause hyperactivity of the downstream mesolimbic dopamine neurons (Figure 4-30B). This is exactly what is hypothesized to be happening in schizophrenia. The dopamine hyperactivity of these downstream mesolimbic dopamine neurons is associated with the positive symptoms of schizophrenia but is actually caused hypothetically by dysconnectivity in upstream glutamate neurons, namely, defective and hypofunctional neurodevelopmental glutamate innervation of parvalbumin-containing GABA interneurons at NMDA receptor-containing synapses (Figure 4-29B and 4-30B).



Figure 4-30. NMDA receptor hypofunction and positive symptoms of schizophrenia, part 1. (A) The cortical brainstem glutamate projection communicates with the mesolimbic dopamine pathway in the ventral tegmental area (VTA) to regulate dopamine release in the nucleus accumbens. (B) If NMDA receptors on cortical GABA interneurons are hypoactive, then the cortical brainstem pathway to the VTA will be overactivated, leading to excessive release of glutamate in the VTA. This will lead to excessive stimulation of the mesolimbic dopamine pathway and thus excessive dopamine release in the nucleus accumbens. This is the theoretical biological basis for the mesolimbic dopamine hyperactivity thought to be associated with the positive symptoms of psychosis.

It is also possible that the dysconnectivity of upstream glutamate neurons in the hippocampus contributes to downstream mesolimbic dopamine hyperactivity via a four-neuron circuit (Figure 4-31A). That circuit consists of (1) the dysconnected and defective hippocampal parvalbumin-containing GABA interneuron, going to (2) the hippocampal glutamate neuron projecting to the nucleus accumbens; then that neuron projecting to two GABA spiny neurons in sequence, (3) the first GABA spiny neuron going from nucleus accumbens to globus pallidus, and finally (4) the second GABA spiny neuron going from globus pallidus to VTA (Figure 4-31A). Loss of adequate glutamate function at parvalbumin-containing GABA interneurons in the hippocampus could lead to hyperactive glutamate output from glutamate neurons that project by this circuit to the mesolimbic dopamine neurons in the VTA, with consequential dopamine hyperactivity and positive symptoms of schizophrenia (Figure 4-31B). Stimulating two GABA neurons in sequence has the net effect of disinhibition (inhibition of inhibition) at the VTA, the same result as direct stimulation (which was illustrated for the prefrontal cortex in Figure 4-30A). The bottom line is that excessive upstream glutamate output from either the prefrontal cortex or the hippocampus may contribute to downstream dopamine hyperactivity and positive symptoms of schizophrenia.



Figure 4-31. NMDA receptor hypofunction and positive symptoms of schizophrenia, part 2. Hypofunctional NMDA receptors at glutamatergic synapses in the ventral hippocampus can also contribute to mesolimbic dopamine hyperactivity. (A) Glutamate released in the ventral hippocampus binds to NMDA receptors on a GABAergic interneuron, stimulating the release of GABA. The GABA binds at receptors on a pyramidal glutamate neuron that projects to the nucleus accumbens; this inhibits glutamate release there. The relative absence of glutamate in the nucleus accumbens allows for normal activation of a GABAergic neuron projecting to the globus pallidus, which in turn allows for normal activation of a GABAergic neuron projecting to the ventral tegmental area (VTA). This leads to normal activation of the mesolimbic dopamine pathway from the VTA to the nucleus accumbens. (B) If NMDA receptors on ventral hippocampal GABA interneurons are hypoactive, then the glutamatergic pathway to the nucleus accumbens will be overactivated, leading to excessive release of glutamate in the nucleus accumbens. This will lead to excessive stimulation of GABAergic neurons projecting to the globus pallidus, which in turn will inhibit release of GABA from the globus pallidus into the VTA. This will lead to disinhibition of the mesolimbic dopamine pathway and thus excessive dopamine release in the nucleus accumbens.

Linking the NMDA hypofunction hypothesis of schizophrenia with the dopamine hypothesis of schizophrenia: negative symptoms

Next, we will discuss the glutamate regulation of mesocortical dopamine neurons (Figure 4-32). It appears that different cortico-brainstem glutamate neurons regulate those unique dopamine neurons in the VTA that project only to the prefrontal cortex – the mesocortical dopamine pathway (Figure 4-32A) – than regulate those dopamine neurons in the VTA that project to the nucleus accumbens as the mesolimbic dopamine pathway (Figure 4-30A). Thus, different populations of glutamate neurons regulate the different populations of dopamine neurons. Cortico-brainstem glutamate neurons destined to regulate mesocortical dopamine neurons in the VTA do not directly innervate them (Figure 4-32A) as do the cortico-brainstem glutamate neurons destined to regulate mesolimbic dopamine neurons in the VTA (Figure 4-30A). Instead, the glutamate neurons regulating mesocortical dopamine neurons do so by indirectly innervating an inhibitory GABA interneuron that itself innervates the mesocortical dopamine neurons (Figure 4-32A). Thus, activation of these particular glutamate neurons leads first to activation of GABA interneurons, which then inhibit mesocortical dopamine neurons (Figure 4-32A). You can imagine what would happen if these glutamate neurons were too active (Figure 4-29B and 4-32B): hypoactivity of the mesocortical dopamine neurons (Figure 4-31B). This is exactly what is hypothesized to be occurring in schizophrenia. The dopamine hypoactivity of these mesocortical dopamine neurons is associated with the negative and cognitive symptoms of schizophrenia. It is hypothetically caused by the same upstream dysconnectivity of glutamate with GABA interneurons that causes the hyperactivity of mesolimbic dopamine neurons, namely, the neurodevelopmental abnormality in glutamate innervation of parvalbumin-containing GABA interneurons at their NMDA synapses (Figures 4-29B and 4-30B). Only in this case, it is affecting a different population of glutamate neurons in the prefrontal cortex and with different downstream consequences: namely, production of negative and cognitive symptoms of schizophrenia rather than positive symptoms.



Figure 4-32. NMDA receptor hypofunction and negative symptoms of schizophrenia. (A) The cortical brainstem glutamate projection communicates with the mesocortical dopamine pathway in the ventral tegmental area (VTA) via pyramidal interneurons, thus regulating dopamine release in the prefrontal cortex. (B) If NMDA receptors on cortical GABA interneurons are hypoactive, then the cortical brainstem pathway to the VTA will be overactivated, leading to excessive release of glutamate in the VTA. This will lead to excessive stimulation of the brainstem pyramidal neurons, which in turn leads to inhibition of mesocortical dopamine neurons. This reduces dopamine release in the prefrontal cortex and is the theoretical biological basis for the negative symptoms of psychosis.

Different populations of cortico-brainstem glutamate projections thus regulate the release of dopamine from both the mesocortical and the mesolimbic dopamine projections, although it appears that this regulation is the opposite for the glutamate neurons that regulate the mesolimbic dopamine pathway compared with the glutamate neurons that regulate the mesocortical dopamine pathway (compare Figures 4-30A and 4-32A), all due to the presence or absence of a GABA interneuron in the VTA.

Neurodevelopment and genetics in schizophrenia

What causes schizophrenia? Nature (i.e., genetics) or nurture (i.e., the environment or epigenetics)? The modern answer seems to be: both. Modern theories of schizophrenia no longer propose that a single gene causes schizophrenia (pure nature) (Figure 4-33) any more than a bad mother can cause schizophrenia (pure nurture). Instead, it seems more likely that schizophrenia is a “conspiracy” among many genes and many environmental stressors to cause abnormal development of brain connections throughout life. In fact, not only is there no single gene for schizophrenia (or for any other major psychiatric disorder) (Figure 4-33), there is no single gene for any specific psychiatric symptoms, behaviors, personalities, or temperaments (Figure 4-34). Genes do not code for mental illnesses or for psychiatric symptoms. Instead, genes code for proteins (Figure 4-35). Today, mental illnesses are thought to be linked in part to someone inheriting an entire portfolio of many genes that carry risk for a mental illness, especially in combination, and set the stage for a mental illness, but do not cause mental illness per se (Figure 4-36). In schizophrenia, multiple risk genes each hypothetically code for a subtle molecular abnormality (Figure 4-35), any one of which alone may be clinically silent until stress from the environment puts a load on these defective genes, and also causes even normal genes to be expressed when they should be silenced, or silenced when they should be expressed (Figure 4-36), a process called epigenetics (discussed briefly in Chapter 1Figure 1-30).



Figure 4-33. Classic theory of inherited disease. According to the classic theory of inherited disease, a single abnormal gene can cause a mental illness. That is, an abnormal gene would produce an abnormal gene product, which, in turn, would lead to neuronal malfunction that directly causes a mental illness. However, no such gene has been identified, and there is no longer any expectation that such a discovery might be made. This is indicated by the red cross-out sign over this theory.



Figure 4-34. Symptom endophenotype model. Another theory, the symptom endophenotype model, posits that, rather than genes causing mental illness, genes instead cause individual symptoms, behaviors, personalities, or temperaments. Thus, an abnormal gene encoding for a symptom, behavior, or trait would cause neuronal malfunction leading to that symptom, behavior, or trait. However, no genes for personality or behavior have been identified, and there is no longer any expectation that such a discovery might be made – as indicated by the red cross-out sign over this theory.



Figure 4-35. Subtle molecular abnormalities. Genes do not directly encode mental illnesses, behaviors, or personalities. Instead, they encode proteins. In some cases, genes may produce genetically altered proteins that code for subtle molecular abnormalities, which in turn may be linked to the development of psychiatric symptoms. Thus, a gene may code for an abnormality in the neurodevelopmental process or in the synthesis or activity of enzymes, transporters, receptors, components of signal transduction, synaptic plasticity machinery, and other neuronal components. Each subtle molecular abnormality may convey risk for the development of mental illness, rather than directly causing a mental illness.



Figure 4-36. Stress–diathesis model of schizophrenia. Schizophrenia may occur as the result of both genetic (nature) and epigenetic (nurture) factors. That is, an individual with multiple genetic risk factors, combined with multiple stressors causing epigenetic changes, may not have sufficient backup mechanisms to compensate for inefficient information processing within a genetically “biased” circuit. The circuit may be unsuccessfully compensated by overactivation, or it may break down and not activate at all. In either case, the abnormal biological endophenotype would be associated with an abnormal behavioral phenotype, and thus with psychiatric symptoms such as hallucinations, delusions, and thought disorder. Such abnormal circuit activation would be potentially detectable with functional brain scanning, and psychiatric symptoms would be manifest on clinical interview.

Thus, mental illnesses are due not only to genes that are abnormal in their DNA and in the function of the proteins they code, but also to normal genes that make normal functioning proteins but are activated or silenced at the wrong times by the environment (nature and nurture, Figure 4-36). In the case of schizophrenia, the problem seems to be “dysconnectivity” of neurons, particularly in hippocampus and prefrontal cortex, and particularly at glutamate synapses with NMDA receptors that become hypofunctional. Stress, traumatic experiences, learning, sensory experiences, sleep deprivation, toxins, and drugs are all examples of how normal genes, such as those that regulate the formation and removal of synapses, are turned on and off by the environment (Figure 4-36). Cannabis use is a particularly malicious environmental stressor to those vulnerable to schizophrenia. These are all examples of the notion of “experience-dependent” development of synaptic connections, something that is hypothetically abnormal in schizophrenia, both from the experiences that the patient may have and from the genes that respond to these experiences. Thus, in schizophrenia an individual not only hypothetically inherits many abnormal genes, which may converge on the formation of NMDA-receptor-containing glutamate synapses, but theoretically also has notable experiences from a stressful environment that cause the abnormal expression or abnormal silencing of perfectly normal genes, in just the necessary sequence to cause this illness (Figure 4-36).

The best evidence that the environment is involved in schizophrenia is that only half of identical twins of patients with schizophrenia also have schizophrenia. Having identical genes is thus not enough to cause schizophrenia, but presumably epigenetics is also in play such that the affected twin not only expresses some abnormal genes that the unaffected twin might not express, but also expresses some normal genes at the wrong time and silences other normal genes at the wrong time, and together these factors cause schizophrenia in one twin but not the other.

The best evidence for the role of dysconnectivity genes in schizophrenia is the convergence of evidence implicating multiple genes that regulate not only neuronal connectivity in general, but glutamate synapse formation and removal in particular (Table 4-8). This includes dysbindinneuregulinErbB4, and DISC1, among others (Figures 4-36 and 4-37). Dysbindin, also known as dystrobrevin binding protein 1, is involved in the formation of synaptic structures and regulation of the activity of the vesicular transporter for glutamate, vGluT. Neuregulin is involved in neuronal migration, genesis of glial cells, and subsequent myelination of neurons by glia. Neuregulin also activates an ErbB4 signaling system that is co-localized with NMDA receptors. These ErbB4 receptors also interact with the postsynaptic density of glutamate synapses, and may be involved in mediating the neuroplasticity triggered by NMDA receptors. Both dysbindin and neuregulin impact the formation and function of the postsynaptic density, a set of proteins that interacts with the postsynaptic membrane to provide both structural and functional regulatory elements for neurotransmission and for NMDA receptors. DISC1 (disrupted in schizophrenia 1) is aptly named for a disrupted gene linked to schizophrenia that makes a protein involved in neurogenesis, neuronal migration, and dendritic organization, and also affects the transport of synaptic vesicles into presynaptic glutamate nerve terminals and regulates cAMP signaling, which would affect the functions of glutamate neurotransmission mediated by metabotropic glutamate receptors.



Figure 4-37. Overview of neurodevelopment. The process of brain development is shown here. After conception, stem cells differentiate into immature neurons. Those that are selected migrate and then differentiate into different types of neurons, after which synaptogenesis occurs. Most neurogenesis, neuronal selection, and neuronal migration occur before birth, although new neurons can form in some brain areas even in adults. After birth, differentiation and myelination of neurons as well as synaptogenesis continue throughout a lifetime. Brain restructuring also occurs throughout life, but is most active during childhood and adolescence in a process known as competitive elimination. Key genes involved in the process of neurodevelopment include DISC1 (disrupted in schizophrenia 1), ErbB4, neuregulin (NRG), dysbindin, regulator of G protein signaling 4 (RGS4), D-amino acid oxidase activator (DAOA), and genes for AMPA.

Table 4-8 Susceptibility genes for schizophrenia


Genes for

Dysbindin (dystrobrevin binding protein 1 or DTNBP1)

Neuregulin (NRG1)

DISC1 (disrupted in schizophrenia 1)

DAOA (D-amino acid oxidase activator; G72/G30)

DAAO (D-amino acid oxidase)

RGS4 (regulator of G protein signaling 4)

COMT (catechol-O-methyl-transferase)

CHRNA7 (α7-nicotinic cholinergic receptor)

GAD1 (glutamic acid decarboxylase 1)

GRM3 (mGluR3)

PPP3CC

PRODH2

AKT1

ErbB4

FEZ1

MUTED

MRDS1 (OFCC1)

BDNF (brain-derived neurotrophic factor)

Nur77

MAO-A (monoamine oxidase A)

Spinophylin

Calcyon

Tyrosine hydroxylase

Dopamine D2 receptor (D2R)

Dopamine D3 receptor (D3R)


Dysbindin, DISC1, and neuregulin all affect normal synapse formation. They all affect NMDA receptor number by altering NMDA receptor trafficking to the postsynaptic membrane, NMDA receptor tethering within that membrane, and NMDA receptor endocytosis that cycles receptors out of the postsynaptic membrane to remove them. Thus, it is easy to see how multiple genetic or epigenetic abnormalities in the expression of these particular genes could lead to dysconnectivity of glutamate neurons in schizophrenia (Figures 4-37 and 4-38).



Figure 4-38. Multiple susceptibility genes converge on NMDA synapses in schizophrenia. There is a powerful convergence of susceptibility genes for schizophrenia upon the connectivity, synaptogenesis, and neurotransmission at glutamate synapses, and specifically at NMDA receptors. Susceptibility genes shown here include those that affect various neurotransmitters involved in modulating NMDA receptors, namely glutamate, γ-aminobutyric acid (GABA), acetylcholine (ACh), dopamine (DA), and serotonin (5HT). That is, abnormalities in genes for various neurotransmitters that regulate NMDA receptors could have additional downstream actions on glutamate functioning at NMDA receptors. Thus, genes that regulate these other neurotransmitters may also constitute susceptibility genes for schizophrenia. The idea is that any of these susceptibility genes could conspire to cause NMDA receptor hypofunction, which would lead to abnormal long-term potentiation (LTP), abnormal synaptic plasticity and connectivity, inadequate synaptic strength, and/or dysregulation of α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) receptors. Any combination of sufficient genetic risk factors with sufficient stress or environmental risk will result in the susceptibility for schizophrenia to become manifest as the disease of schizophrenia with the presence of full syndrome symptoms.

Other risk genes implicate specific proteins that directly regulate glutamate synapses and, if abnormally expressed, could add to the misery of a disconnected and dysfunctional NMDA glutamate synapse (Figure 4-38). For example, the gene for DAOA (D-amino acid oxidase activator) codes for a protein that activates the enzyme DAO (D-amino acid oxidase). DAO degrades the co-transmitter D-serine that acts at glutamate synapses and at NMDA receptors. DAOA activates this DAO enzyme, so abnormalities in the gene for DAOA would be expected to alter the metabolism of D-serine. This in turn would alter glutamate neurotransmission at NMDA receptors. Another schizophrenia susceptibility gene active directly at glutamate synapses is RSG4 (regulator of G-protein signaling), and this gene product also impacts metabotropic glutamate receptor signaling through the G-protein-coupled signal transduction system.

Normally, when glutamate synapses are active, their NMDA receptors trigger an electrical phenomenon known as long-term potentiation (LTP). With the help of dysbindin, DISC1, and neuregulin, LTP leads to structural and functional changes of the synapse that make neurotransmission more efficient, sometimes called “strengthening” of synapses (Figure 4-39). This includes increasing the number of AMPA receptors. AMPA receptors are important for mediating excitatory neurotransmission and depolarization at glutamate synapses. Thus, more AMPA receptors can mean a “strengthened” synapse. Synaptic connections that are frequently used develop recurrent LTP and consequential robust neuroplastic influences, thus strengthening them according to the old saying “nerves that fire together wire together.” However, if something is wrong with the genes that regulate synaptic strengthening, it is possible that this causes less effective use of these synapses, makes the NMDA receptors hypoactive, leading to ineffective LTP and fewer AMPA receptors trafficking into the postsynaptic neuron (Figure 4-39). Such a synapse would be “weak,” theoretically causing inefficient information processing in its circuit and possibly also causing symptoms of schizophrenia. The strengthening or weakening of a glutamate synapse is an example of “activity-dependent” or “use-dependent” or “experience-dependent” regulation of NMDA receptors and functionality at glutamate synapses. This not only occurs when these synapses first form, but continues throughout life as a sort of ongoing remodeling in response to what experiences the individual has, and thus how much that synapse is used or neglected. Abnormalities in these continuing dynamics at NMDA receptors and glutamate synapses may explain why the course of schizophrenia is progressive and changes over time for most patients, namely, from an asymptomatic period, to a prodrome, to a first break psychosis with robust treatment responsiveness, to multiple psychotic episodes with declining treatment responsiveness, to a state of pervasive negative and cognitive symptoms without recovery.



Figure 4-39. Neurodevelopmental hypothesis of schizophrenia. Dysbindin, DISC1 (disrupted in schizophrenia 1), and neuregulin are all involved in “strengthening” of glutamate synapses. Under normal circumstances, N-methyl-D-aspartate (NMDA) receptors in active glutamate synapses trigger long-term potentiation (LTP), which leads to structural and functional changes of the synapse to make it more efficient, or “strengthened.” In particular, this process leads to an increased number of α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid (AMPA) receptors, which are important for mediating glutamatergic neurotransmission. Normal synaptic strengthening means that the synapse will survive during competitive elimination. If the genes that regulate strengthening of glutamate synapses are abnormal, then this could cause hypofunctioning of NMDA receptors, with a resultant decrease in LTP and fewer AMPA receptors. This abnormal synaptic strengthening and dysconnectivity would lead to weak synapses that would not survive competitive elimination. This would theoretically lead to increased risk of developing schizophrenia, and these abnormal synapses could mediate the symptoms of schizophrenia.

Another important aspect of synaptic strength is that it likely determines whether a given synapse is eliminated or maintained. Specifically, “strong” synapses with efficient NMDA neurotransmission and many AMPA receptors survive whereas “weak” synapses with few AMPA receptors may be targets for elimination (Figure 4-39). This normally shapes the brain’s circuits so that the most critical synapses are not only strengthened but also survive the ongoing selection process, keeping the most efficient and most frequently utilized synapses, while eliminating inefficient and rarely utilized synapses. However, if critical synapses are not adequately strengthened in schizophrenia, it could lead to their wrongful elimination, causing dysconnectivity that disrupts information flow from circuits now deprived of synaptic connections where communication needs to be efficient (Figure 4-39). Competitive elimination of “weak” but critical synapses during adolescence could even explain why schizophrenia has onset at this time. Normally, almost half of the brain’s synapses are eliminated in adolescence (Figure 4-40). In adulthood, you may lose (and replace elsewhere) about 7% of the synapses in your cortex every week! If abnormalities in genes for dysbindin, neuregulin, and/or DISC1 lead to the lack of critical synapses being strengthened, these critical synapses may be mistakenly eliminated during adolescence with disastrous consequences, namely the onset of symptoms of schizophrenia. Also, it may be that the dysconnectivity of abnormal glutamate synapses present from birth is masked by the presence of many additional weak connections prior to adolescence, acting with exuberance to compensate for defective glutamate connectivity, and with that compensation destroyed by the normal competitive elimination of synapses in adolescence, schizophrenia emerges.



Figure 4-40. Synapse formation by age. Synapses are formed at a furious rate between birth and age 6. Competitive elimination and restructuring of synapses peaks during pubescence and adolescence, leaving about half to two-thirds of the synapses present in childhood to survive into adulthood.

Neuroimaging circuits in schizophrenia

Imaging brain circuits in patients with schizophrenia with functional magnetic resonance imaging (fMRI) uncovers abnormal information processing in brain areas linked to cognition and emotion. Modern psychiatric research techniques can put a “load” on brain circuits, and thereby perform a type of psychiatric “stress test” while visualizing the activity of the brain circuits. fMRI brain scanners can most readily detect the activity of neurons near the surface of the brain, which are mostly pyramidal neurons in the cortex, although some deeper gray-matter areas such as the striatum and amygdala can also be imaged. The activity of neurons so visualized in the cortex is the first leg of various brain circuits, especially glutamate neurons, as shown in Figure 4-27, acting within feedback loops from the cortex to the striatal complex, with that information relayed to the thalamus via a GABA neuron, then back to the cortex again via another glutamate neuron as cortico-striato-thalamo-cortical or CSTC feedback loops. Brain circuits like this are information-processing “engines” that are activated by various tasks or loads placed upon them, and seeing them light up may be literally watching the brain “think.”

Function in the brain is topographical, meaning different brain circuits process different kinds of information. For example, the dorsolateral prefrontal cortex (DLPFC) is thought to be most closely linked to cognitive functioning such as problem solving, whereas the ventromedial prefrontal cortex (VMPFC) – along with the amygdala – is thought to be most closely linked to emotional functioning, such as mood. Neurons in various brain areas “stressed” with an information-processing “load” literally light up a specific brain area that can be visualized with current neuroimaging techniques. Thus, doing a calculation can light up the DLPFC and seeing a sad face can activate the VMPFC and amygdala.

Studies in patients with schizophrenia suggest that they cannot adequately recruit the hippocampus during memory recall, even though output from the hippocampus seems to be high in the first place. Also, patients with schizophrenia do not seem to be able to appropriately activate the DLPFC during a working memory task (compare Figures 4-41A and 4-41B), with decreased recruitment correlated with worsening cognitive symptoms (Figure 4-41C). Actually, the results are somewhat inconsistent across studies, and it appears that prefrontal cortical or hippocampal dysfunction in schizophrenia is likely to be more complicated than just “up” (hyperactivation) or “down” (hypoactivation), but might be better characterized as “out of tune.” According to this concept, either too much or too little activation of neuronal activity in the prefrontal cortex is suboptimal and can potentially be symptomatic, just as a guitar string is out of tune whether it has too much or too little tension on it.



Figure 4-41. n-back test in schizophrenia. (A) Functional neuroimaging studies have suggested that information processing in schizophrenia is abnormal in certain brain regions. Information processing during cognitive tasks has been evaluated using the n-back test. In the 0-back variant of the test, participants view a number on a screen and then indicate what the number was. In the 1-back test, participants are shown a stimulus but do not respond; after viewing the second stimulus, the participant then pushes a button corresponding to the first stimulus. The n can be any number, with higher numbers associated with greater difficulty. Performing the n-back test results in activation of the dorsolateral prefrontal cortex (DLPFC). The degree of activation indicates how efficient the information processing is in DLPFC, with both overactivation and hypoactivation associated with inefficient information processing. (B) Patients with schizophrenia exhibit inefficient information processing during cognitive challenges such as the n-back test. To perform near normal, these individuals must recruit greater neuronal resources, initially resulting in hyperactivation of the dorsolateral prefrontal cortex (DLPFC). Under increased cognitive load, however, schizophrenia patients do not appropriately engage and sustain the DLPFC, with resultant hypoactivation. (C) The degree of DLPFC activity, as measured by functional neuroimaging, correlates with the number of cognitive symptoms that a patient exhibits.

How can circuits in schizophrenia be both hyperactive and hypoactive? Patients with schizophrenia appear to utilize greater prefrontal resources when performing cognitive tasks and yet achieve lower accuracy because they have cognitive impairment despite their best efforts. To perform near normal, patients with schizophrenia engage the DLPFC, but do so inefficiently, recruiting greater neural resources and hyperactivating the DLPFC. When performing poorly, schizophrenia patients do not appropriately engage and sustain the DLPFC, and thus show hypoactivation. Thus, DLPFC circuits in schizophrenia patients can either be underactive and hypofrontal or overactive and inefficient.

It is interesting to note that unaffected siblings of patients with schizophrenia may have the very same inefficient information processing in DLPFC that schizophrenia patients have. Although unaffected siblings of schizophrenia patients might have some mild degree of cognitive impairment, they do not share the full syndrome of schizophrenia; however, neuroimaging reveals that they may share the same inefficient DLPFC functioning while performing cognitive tasks that characterizes their sibling with schizophrenia. The unaffected siblings of a schizophrenia patient may thus share some of the susceptibility genes for schizophrenia with their affected sibling, but not enough of these risk genes to have the full syndrome of schizophrenia itself. Functional neuroimaging also has the potential of unmasking inefficient information processing in clinically silent presymptomatic patients destined to progress to the full schizophrenia syndrome, but much further research is required to see if this will become clinically useful.

Schizophrenia has also long been recognized as having impairments in the ability to identify and accurately interpret emotions from overt sources, including facial expressions. This may be due to inefficient information processing within the VMPFC and amygdala and can be measured by imaging the response of the amygdala to emotional input, especially from facial expressions. The amygdala is normally activated by looking at scary, threatening faces, or by assessing how happy or sad a face may be and while attempting to match emotions to faces (Figure 4-42). Whereas healthy controls may activate the amygdala in response to scary or fearful or emotionally charged faces (Figure 4-42A), patients with schizophrenia may not (Figure 4-42B). This may represent distortion of reality as well as impairment in recognizing negative emotions and in decoding negative emotions in schizophrenia. Failure to mount the “normal” emotional response to a scary face can also represent an inability to interpret social cues and lead to distortions in judgment and reasoning in schizophrenia. Thus, these negative and affective symptoms of schizophrenia may be due in part to lack of emotional processing under circumstances when this should be occurring.



Figure 4-42. Fearful stimuli and schizophrenia. (A) Normally, exposure to an emotional stimulus, such as a scary face, causes hyperactivation in the amygdala. (B) Patients with schizophrenia often have impairments in the ability to identify and interpret emotional stimuli. The underlying neurobiological explanation for this may be inefficient information processing within the ventral system. In this example, the amygdala is not appropriately engaged during exposure to an emotional stimulus.

On the other hand, a neutral face or neutral stimulus may provoke little activation of the amygdala in a healthy person (Figure 4-43A), yet an over-reaction in a patient with schizophrenia (Figure 4-43B), who may mistakenly judge people negatively or conclude wrongly that another holds strong unfavorable impressions of them or may even be threatening them. Activating emotional processing in the amygdala when it is inappropriate may accompany the symptom of paranoia, and lead to impaired interpersonal functioning including problems in social communication. Thus, patients with schizophrenia may exhibit deficits in recognizing emotions that may be manifested either as positive or negative symptoms of this disorder. The underlying biological endophenotype of amygdala activation (or lack of activation) can be assessed with neuroimaging whether the patient is experiencing these symptoms or not. Looking at the efficiency of emotional information processing may help clinicians identify and understand emotional symptoms that are difficult for patients with schizophrenia to express.



Figure 4-43. Neutral stimuli and schizophrenia. (A) Normally, exposure to a neutral stimulus, such as a neutral face, causes little activation of the amygdala. (B) Schizophrenia patients may mistakenly judge others as threatening, with associated inappropriate hyperactivation of the amygdala.

Imaging genetics and epistasis

Not only can the impact of a mental illness such as schizophrenia be imaged today, so can the impact of certain genes. That is, single genes can alter the efficiency of information processing in everyone, and as such may endow risk for mental illness, but not cause mental illness by themselves, as discussed above (Figure 4-36). Thus, individuals with the gene for catechol-O-methyl-transferase (COMT) that has higher enzyme activity (called Val, for the amino acid valine substituted at a critical site) have lower dopamine levels in DLPFC, and thus less efficient information processing there compared to individuals with the gene for COMT with lower enzyme activity and higher dopamine levels (called Met) (Figure 4-44). This difference in neuronal “effort” is usually not apparent as cognitive difficulties in the normal population, but may serve as one of many risk factors for schizophrenia (Table 4-8 and Figure 4-38).



Figure 4-44. Risk genes and efficiency of information processing in schizophrenia. Several genes have been identified that may confer risk of inefficient information processing in schizophrenia. The single genes with the greatest risk may be the catechol-O-methyl-transferase Val allele (COMT-Val) and the methylene tetrahydrofolate reductase T allele (MTHFR-T). Risk may be even greater for individuals with multiple risk genes, and in particular for patients with both the COMT-Val allele and another risk gene. Perhaps the greatest risk has been seen for individuals who carry the neuregulin 1 (NRG1), ErbB4, and AKT risk genes.

Don’t have T with Val

The effects of two or more risk genes working together to increase the risk of schizophrenia can now be demonstrated with neuroimaging that shows how certain risk genes “conspire” to decrease the efficiency of information processing in the DLPFC during a cognitive load in schizophrenia (Figure 4-44). That is, the Val variant of COMT by itself may or may not consistently alter DLPFC activity during a working memory test in schizophrenia compared to individuals with the Met variant of COMT, but when combined with another genetic variant of another enzyme that independently reduces the availability of dopamine in prefrontal cortex, there is a more robust increase in DLPFC activity during a working memory load (Figure 4-44). That variant is the T form of the enzyme MTHFR (methylene tetrahydrofolate reductase), involved in the formation of dopamine and in regulating the activity of COMT at the level of gene expression. There can be reduced efficiency of information processing during a working memory load with the T form of MTHFR alone, but combined with the Val form of COMT, the change in information processing is more robust, thus demonstrating the epistatic interaction of MTHFR-T with COMT-Val (Figure 4-44). So, you may want to have tea with Valerie, but you don’t necessarily want to have T with Val, or your cognitive difficulties may be epistatically worsened if you have schizophrenia.

Be careful of having your alphabet soup

Another example of epistasis is how the efficiency of information processing in DLPFC gets worse if you combine two or three of several of the schizophrenia risk genes that have horrible-sounding names from various letters of the alphabet: NRG1 (neuregulin), ErbB4, and AKT. We have already mentioned above that NRG1 is involved in neuronal migration and synapse formation, and can impact the formation and function of the postsynaptic density, a set of proteins that interacts with the postsynaptic membrane to provide both structural and functional regulatory elements for neurotransmission and for NMDA receptors. NRG1 also affects NMDA receptor number by altering NMDA receptor trafficking to the postsynaptic membrane, NMDA receptor tethering within that membrane, and NMDA receptor endocytosis that cycles receptors out of the postsynaptic membrane to remove them. NRG1 directly interacts with ErbB4, and activates the ErbB4 signaling system that is co-localized with NMDA receptors and interacts with the postsynaptic density of glutamate synapses and is involved in mediating the neuroplasticity triggered by NMDA receptors. Thus, it is easy to see how a genetic abnormality in the NRG1 protein could readily interact in an undesirable way with a genetic abnormality in its receptor ErbB4, to cause problems with information processing and to increase the risk for schizophrenia. In fact, when patients with schizophrenia have the risk genes both for NRG1 and for ErbB4, but not when they have just one of them, their information processing in DLPFC is abnormal (Figure 4-44).

How about AKT? AKT is a kinase enzyme, part of the intracellular signal transduction cascades discussed in Chapter 1 (Figures 1-91-161-181-28). This particular kinase specifically interacts with β-arrestin 2 and GSK-3 during signal transduction and has been shown to regulate neuronal cell size and cell survival, and to increase synaptic plasticity. Malfunctioning of AKT can lead to overactivation of GSK-3 and lack of internalization of dopamine D2 receptors, contributing to dopamine hyperactivity and the risk of schizophrenia. By itself or in combination with the risk gene for either NRG1 or ErbB4, the risk gene for AKT does not appear to compromise the efficiency of information processing in the DLPFC of patients with schizophrenia (Figure 4-44). However, when combined with both the risk genes for NRG1 and ErbB4, the risk gene for AKT further increases the activation and inefficiency of information processing in DLPFC of patients with schizophrenia who have both the risk genes for NRG1 plus ErbB4, a further example of epistasis, but here among three risk genes (Figure 4-44). Thus, it is apparent that the more risk genes, the more the bias towards circuit breakdown in schizophrenia, and the more likely the illness will manifest, as shown in the hypothetical interactions of Figures 4-36 and 4-38.

Summary

This chapter has provided a clinical description of psychosis, with special emphasis on the psychotic illness schizophrenia. We have explained the dopamine hypothesis of schizophrenia, and the related NMDA receptor hypofunction hypothesis of schizophrenia, which are the major hypotheses for explaining the mechanism for the symptoms of schizophrenia. The major dopamine pathways and the major glutamate pathways in the brain have all been described. Overactivity of the mesolimbic dopamine system may mediate the positive symptoms of psychosis and may be linked to hypofunctioning NMDA glutamate receptors in parvalbumin-containing GABA interneurons in the prefrontal cortex and hippocampus. Underactivity of the mesocortical dopamine system may mediate the negative, cognitive, and affective symptoms of schizophrenia and could also be linked to hypofunctioning NMDA receptors at different GABA interneurons.

The synthesis, metabolism, reuptake, and receptors for both dopamine and glutamate are described in this chapter. Dopamine D2 receptors are targets of all known antipsychotic drugs. NMDA glutamate receptors require interaction not only with the neurotransmitter glutamate, but also with the cotransmitters glycine or D-serine. Dysconnectivity of NMDA-receptor-containing synapses caused by genetic and environmental/epigenetic influences is a major hypothesis for the cause of schizophrenia, including its upstream glutamate hyperactivity and NMDA receptor hypofunction, as well as its downstream increases in mesolimbic dopamine but decreases in mesocortical dopamine. A whole host of susceptibility genes that regulate neuronal connectivity and synapse formation also increase the risk for schizophrenia and converge upon the NMDA-receptor-containing glutamate synapse as a hypothetical central biological flaw in schizophrenia. Malfunctioning neural circuits can be imaged in schizophrenic patients, including those in the dorsolateral prefrontal cortex that are linked to cognitive symptoms, and those in the amygdala that are linked to symptoms of emotional dysregulation. The effects of risk genes on the efficient functioning of brain circuits can also be imaged, including the epistatic interaction of two or more risk genes.