CHAPTER 60: BIOLOGY OF PSYCHIATRIC DISORDERS

Psychiatric disorders are central nervous system diseases characterized by disturbances in emotion, cognition, motivation, and socialization. They are highly heritable, with genetic risk comprising 20–90% of disease vulnerability. As a result of their prevalence, early onset, and persistence, they contribute substantially to the burden of illness worldwide. All psychiatric disorders are broad heterogeneous syndromes that currently lack well-defined neuropathology and bona fide biologic markers. Therefore, diagnoses continue to be made solely from clinical observations using criteria in the Diagnostic and Statistical Manual of Mental Disorders (DSM) of the American Psychiatric Association, which is in its fifth edition as of 2013.

There is increasing agreement that the classification of psychiatric illnesses in DSM does not accurately reflect the underlying biology of these disorders. Uncertainties in diagnosis make it extremely difficult to study the neurobiologic and genetic basis of mental illness. This has led to the development of an alternative diagnostic scheme, termed Research Domain Criteria (RDoCs), which classifies mental illness on the basis of core abnormalities—e.g., psychosis (loss of reality) or anhedonia (decreased ability to experience pleasure), which are common symptoms of several illnesses—with the idea that such classifications will assist in defining the biologic basis of at least key symptoms. Other factors that have impeded progress in understanding mental illness include the lack of access to pathologic brain tissue except upon death and the inherent limitations of animal models for disorders defined largely by behavioral abnormalities (e.g., hallucinations, delusions, guilt, suicidality) that are inaccessible in animals.

Despite these limitations, the past decade has seen significant advances. Neuroimaging methods are beginning to provide evidence of brain pathology, genome-wide association studies and high-throughput sequencing are at last revealing genes that confer risk for severe forms of mental illness, and investigations using better validated animal models are offering new insight into the molecular, cellular, and circuit mechanisms of disease pathogenesis. There is also excitement in the potential utility of neuron-like cells induced in vitro from a patient’s peripheral tissues (e.g., fibroblasts) providing novel ways of studying disease pathophysiology and screening for new treatments. There is consequently justified optimism that the field of psychiatry will transition from behaviorally defined syndromes to true biologic disease entities and that such advances will drive the development of improved treatments and eventually cures and preventive measures. This chapter describes several examples of recent discoveries in basic neuroscience that have informed our current understanding of disease mechanisms in psychiatry.

Because the human brain can only be examined indirectly during life, genome analyses have been extremely important for obtaining molecular clues about the pathogenesis of psychiatric disorders. A wealth of new information has been made possible by recent technological developments that have permitted affordable, large-scale genome-wide association studies and fine-scale sequencing. As an example, significant progress has been made in the genetics of autism spectrum disorders (ASDs), which are a heterogeneous group of neurodevelopmental diseases that share clinical features of impaired reciprocal social communication and interaction and restricted, repetitive patterns of behavior. ASDs are highly heritable; concordance rates in monozygotic twins (~60–90%) are roughly 10-fold higher than in dizygotic twins and siblings, whereas first-degree relatives show about 50-fold increased risk compared with the general population. ASDs are also genetically heterogeneous. More than 100 known mutations account for up to 20% of cases, although none individually accounts for more than 1% (Table 60-1). It appears that most cases result from complex genetic mechanisms, including inheritance of multiple genetic risk variants and epigenetic modifications. For example, up to 10% of patients with autism have large (>500 kb) de novo copy number variations scattered across the genome, suggesting that hundreds of different genes can influence autism risk.

Amid this genetic heterogeneity, however, some common themes have emerged that inform pathogenesis of ASDs. For instance, many identified mutations are in genes that encode proteins involved in synaptic function and early transcriptional regulation (Table 60-1) and have a clear relationship to activity-dependent neural responses that can affect the development of neural systems underlying cognition and social behaviors. Mutations in these genes may be detrimental by altering the balance of excitatory versus inhibitory synaptic signaling in local and extended circuits and by altering mechanisms that control brain growth. Some mutations affect genes (e.g., PTEN, TSC1, and TSC2) that negatively regulate signaling from several types of extracellular stimuli, including those transduced by receptor tyrosine kinases. Their dysregulation can alter neuronal growth as well as synaptic development and function.

With further understanding of pathogenesis and the definition of specific ASD subtypes, there is reason to believe that effective therapies will be identified. Work in mouse models has already demonstrated that some autism-like behaviors can be reversed, even in fully developed adult animals, by modifying the underlying pathology; these results encourage hope for many affected individuals. Treatments that target excitation-inhibition imbalance or altered mRNA translation appear to offer early promise. For example, the genes TSC1, TSC2, and PTEN are negative regulators of signaling through the target of rapamycin complex 1 (TORC1), which regulates protein synthesis. Rapamycin, a selective inhibitor of TORC1, can reverse several behavioral and synaptic defects in mice carrying null mutations in these genes. Another example is fragile X syndrome, which is the leading cause of inherited autism and mental disability and is due to mutations in FMR1 that result in loss of the encoded fragile X mental retardation protein (FMRP). FMRP is a polyribosome-associated mRNA-binding protein that represses the translation of a subset (~5%) of all mRNAs, several of which encode proteins that comprise the postsynaptic density, including the metabotropic glutamate receptor 5 (mGluR5). Treatment of Fmr1 knockout mice with mGluR5 antagonists reduces several behavioral and morphologic abnormalities in these mice; these promising preclinical results have led to ongoing trials of mGluR5 antagonists in humans with fragile X syndrome.

The ability to catalog common genetic variants and assay them on array-based platforms and, more recently, to carry out whole-exome sequencing has allowed investigators to collect sample sizes sufficient to detect genetic risk loci for schizophrenia with genome-wide significance. Several of the identified genes are parts of molecular complexes, such as voltage-gated calcium channels (in particular, CACNA1C and CACNB2) and the postsynaptic density of excitatory synapses. As in ASDs, copy number variants, single-nucleotide polymorphisms, and small insertions and deletions are common in schizophrenia. Genes that promote risk for drug addiction have also begun to emerge from large family and population studies. The best-established susceptibility loci are regions on chromosomes 4 and 5 containing GABA_A receptor gene clusters linked to alcoholism and the CHRNA5-A3-B4 nicotinic acetylcholine receptor gene cluster on chromosome 15 associated with nicotine and alcohol addiction.

A recurrent theme that has emerged from genetic studies of psychiatric disorders is pleiotropy, namely, that many genes are associatedwith multiple psychiatric syndromes. For example, mutations in MECP2, FMR1, and TSC1 and TSC2 (see Table 60-1 for abbreviations) can cause mental retardation without ASD, others in MECP2 can cause obsessive-compulsive and attention-deficit hyperactivity disorders, some alleles of NRXN1 are associated with symptoms of both ASD and schizophrenia, and common polymorphisms in CACNA1C are strongly associated with both schizophrenia and bipolar disorder. Likewise, duplication of chromosome 16p is associated with both schizophrenia and autism, whereas DiGeorge’s (velocardiofacial) syndrome region deletions and the DISC1 (disrupted in schizophrenia 1) locus on chromosome 22 are associated with schizophrenia, autism, and bipolar disorder. The association of genes with multiple syndromes attests to the complexity of psychiatric disorders and the influence of additional factors that combine to specify the ultimate phenotype, including regulatory variants that determine cell-type specificity and timing of gene expression, protective variants, and epigenetic effects.

Studies of signal transduction have revealed numerous intracellular signaling pathways that are perturbed in psychiatric disorders, and such research has provided insight into development of new therapeutic agents. For example, lithium is a highly effective drug for bipolar disorder and competes with magnesium to inhibit numerous magnesium-dependent enzymes, including the enzyme GSK3β and several enzymes involved in phosphoinositide signaling that lead to activation of protein kinase C. These findings have led to discovery programs focused on developing GSK3β or protein kinase C inhibitors as potential novel treatments for mood disorders.

The observations that tricyclic antidepressants (e.g., imipramine) inhibit serotonin and/or norepinephrine reuptake and that monoamine oxidase inhibitors (e.g., tranylcypromine) are effective antidepressants initially led to the view that depression is caused by a deficiency of these monoamines. However, this hypothesis has not been substantiated. A cardinal feature of these drugs is that long-term administration is needed for their antidepressant effects. This means that their short-term actions, namely promotion of serotonin or norepinephrine function, are not per se antidepressant but rather induce a cascade of adaptations in the brain that underlie their clinical effects. The nature of these therapeutic drug-induced adaptations has not been identified with certainty. One theory holds that, in a subset of depressed patients who display upregulation of the hypothalamic-pituitary-adrenal (HPA) axis characterized by increased secretion of corticotropin-releasing factor (CRF) and glucocorticoids, excessive glucocorticoids cause atrophy of hippocampal neurons, which is associated with reduced hippocampal volumes seen clinically. Chronic antidepressant administration might reverse this atrophy by increasing brain-derived neurotrophic factor (BDNF) in hippocampus. A role for stress-induced decreases in the generation of newly born hippocampal granule cell neurons, and its reversal by antidepressants through BDNF and other growth factors, has also been suggested.

A major advance in recent years has been the identification of several rapidly acting antidepressants with non–monoamine-based mechanisms of action. The best established is ketamine, a noncompetitive antagonist of N-methyl-D-aspartate (NMDA) glutamate receptors, which exerts rapid (hours) and robust antidepressant effects in severely depressed patients who have not responded to other treatments. Ketamine, which at higher doses is psychotomimetic and anesthetic, exerts these antidepressant effects at low doses with minimal side effects. However, the response to ketamine is transient, which has led to several approaches to maintain treatment response, such as repeated ketamine delivery. The mechanism underlying ketamine’s antidepressant action is not known, but its striking clinical efficacy has stimulated animal research on the role of glutamate neurotransmission and synaptic plasticity in key limbic regions. Recent evidence supports a role for TORC1 activation because administration of rapamycin blocks the antidepressant-like effects of ketamine in animal models. Mechanisms by which ketamine activates TORC1 are currently an active area of investigation.

A major goal in the field of drug abuse has been to identify neuroadaptive mechanisms that lead from recreational use to addiction. Such research has determined that repeated intake of abused drugs induces specific changes in cellular signal transduction, leading to changes in synaptic strength (long-term potentiation or depression) and neuronal structure (altered dendritic branching or cell soma size) within the brain’s reward circuitry. These modifications are mediated in part by changes in gene expression, achieved by drug regulation of transcription factors (e.g., CREB [cAMP response element-binding protein] and ΔFosB [a Fos family protein]) and their target genes. Such alterations in gene expression are associated with lasting alterations in epigenetic modifications, including histone acetylation and methylation and DNA methylation. These adaptations provide opportunities for developing treatments targeted to drug-addicted individuals. The fact that the spectrum of these adaptations partly differs depending on the particular addictive substance used creates opportunities for treatments that are specific for different classes of addictive drugs and that may, therefore, be less likely to disturb basic mechanisms that govern normal motivation and reward.

Increasingly, causal relationships are being established between individual molecular and cellular adaptations and specific behavioral abnormalities that characterize the addicted state. For example, acute activation of μ-opioid receptors by morphine or other opiates activates G_i/o proteins, leading to inhibition of adenylyl cyclase, resulting in reduced cyclic AMP (cAMP) production, protein kinase A (PKA) activation, and activation of the transcription factor CREB. Repeated administration of these drugs (Fig. 60-1) evokes a homeostatic response involving upregulation of adenylyl cyclases and PKA and increased activation of CREB. Such upregulation of cAMP-CREB signaling has been identified in the locus coeruleus, periaqueductal gray, ventral tegmental area (VTA), nucleus accumbens, and several other CNS regions, and contributes to opiate craving and signs of opiate withdrawal. The fact that endogenous opioid peptides do not produce tolerance and dependence, while morphine and heroin do, may relate to the observation that, unlike endogenous opioids, morphine and heroin are weak inducers of μ-opioid receptor desensitization and endocytosis. Therefore, these drugs cause prolonged receptor activation and inhibition of adenylyl cyclases, which provides a powerful stimulus for the upregulation of cAMP-CREB signaling that characterizes the opiate-dependent state.

The study of interconnected brain circuits that drive behavior has been greatly advanced through newer methods in brain imagining that have documented abnormalities in neural function and connectivity in psychiatric disorders. The past decade has also witnessed the development of revolutionary new techniques—optogenetics and designer receptors and ligands—that provide unprecedented temporal and spatial control of neural circuits and permit detection of neural activity in real time in awake, behaving animals.

Positron emission tomography (PET), diffusion tensor imaging (DTI), and functional magnetic resonance imaging (fMRI) have identified neural circuits that contribute to psychiatric disorders, for example, defining the neural circuitry of mood within the brain’s limbic system (Fig. 60-2). Integral to this system are the nucleus accumbens (important also for brain reward—see below), amygdala, hippocampus, and regions of prefrontal cortex. Recent optogenetic research in animals, where the activity of specific types of neurons in defined circuits can be controlled with light, has confirmed the importance of this limbic circuitry in controlling depression-related behavioral abnormalities. Given that many symptoms of depression (so-called neurovegetative symptoms) involve physiologic functions, a key role for the hypothalamus is also presumed. A subset of depressed individuals shows a small reduction in hippocampal size, as noted above. In addition, brain imaging investigations have revealed increased activation of the amygdala by negative stimuli and reduced activation of the nucleus accumbens by rewarding stimuli. There is also evidence for altered activity in prefrontal cortex, such as hyperactivity of subgenual area 25 in anterior cingulate cortex. Such findings have led to trials of deep brain stimulation (DBS) of either the nucleus accumbens or subgenual area 25, which appears to be therapeutic in some severely depressed individuals.

In schizophrenia, structural and functional imaging studies have identified a 3% loss of brain volume, most of which is in gray matter. This loss is progressive, and cortical gray matter appears to be particularly affected over time. The temporal lobes, particularly the left superior temporal gyrus, Heschl gyrus, and planum temporale, are often the most severely affected. The rate of loss in these regions as well as in frontal and parietal lobes appears to be greatest early in the course of the disease. Functional imaging studies provide evidence of reduced metabolic (presumably neural) activity in the dorsolateral prefrontal cortex at rest and when performing tests of executive function, including working memory. There is also evidence for impaired structural and task-related functional connectivity, mainly in frontal and temporal lobes.

These neuroimaging findings in schizophrenia have been confirmed in pathologic studies that show enlargement of the ventricular system and reduction of cortical and subcortical gray matter in frontal and temporal lobes and in the limbic system. The reduction in cortical thickness is associated with increased cell packing density and reduced neuropil (defined as axons, dendrites, and glial cell processes) without an apparent change in neuronal cell number. Specific classes of interneurons in prefrontal cortex consistently show reduced expression of the gene encoding the enzyme glutamic acid decarboxylase 1 (GAD1), which synthesizes γ-aminobutyric acid (GABA), the principal inhibitory neurotransmitter in the brain. Neuregulin 1 (NRG1), a member of the epidermal growth factor (EGF) family of growth factors, and its receptor ERBB4, have been implicated in schizophrenia, and they serve important roles in the maturation of GABAergic interneurons in cerebral cortex; loss of NRG1-ERBB4 in mice leads to a reduced neuropil, thus phenocopying a pathologic finding in schizophrenia. These findings are consistent with one working hypothesis of schizophrenia as a developmental neurodegenerative disorder due in part to loss of cortical interneurons in frontal and temporal lobes.

Work in rodent and nonhuman primate models of addiction has established the brain’s reward regions as key neural substrates for the acute actions of drugs of abuse and for addiction induced by repeated drug administration (Fig. 60-2). Midbrain dopamine neurons in the VTA function normally as rheostats of reward: they are activated by natural rewards (food, sex, social interaction) or even by the expectation of such rewards, and many are suppressed by the absence of an expected reward or by aversive stimuli. These neurons thereby transmit crucial survival signals to the rest of the limbic brain to promote reward-related behavior, including motor responses to seek and obtain the rewards (nucleus accumbens), memories of reward-related cues (amygdala, hippocampus), and executive control of obtaining rewards (prefrontal cortex).

Drugs of abuse alter neurotransmission through initial actions at different classes of ion channels, neurotransmitter receptors, or neurotransmitter transporters (Table 60-2). Studies in animal models have demonstrated that although the initial targets differ, the actions of these drugs converge on the brain’s reward circuitry by promoting dopamine neurotransmission in the nucleus accumbens and other limbic targets of the VTA. In addition, some drugs promote activation of opioid and cannabinoid receptors, which modulate this reward circuitry. By these mechanisms, drugs of abuse produce powerful rewarding signals, which, after repeated drug administration, corrupt a vulnerable brain’s reward circuitry in ways that promote addiction. Three major pathologic adaptations have been described. First, drugs produce tolerance and dependence in reward circuits, which promote escalating drug intake and a negative emotional state during drug withdrawal that promotes relapse. Second, sensitization to the rewarding effects of the drugs and associated cues is seen during prolonged abstinence and also triggers relapse. Third, executive function is impaired in such a way as to increase impulsivity and compulsivity, both of which promote relapse.

Imaging studies in humans confirm that addictive drugs, as well as craving for them, activate the brain’s reward circuitry. In addition, patients who abuse alcohol or psychostimulants show reduced gray matter in the prefrontal cortex as well as reduced activity in anterior cingulate and orbitofrontal cortex during tasks of attention and inhibitory control. It is thought that damage to these cortical areas contributes to addiction by impairing decision-making and increasing impulsivity.

There is increasing evidence for the involvement of inflammatory mechanisms in a subset of depressed patients. These individuals display elevated blood levels of interleukin 6 (IL-6), tumor necrosis factor α (TNF-α), and other cytokines. Moreover, rodents exposed to chronic stress exhibit similar increases in peripheral cytokines, and peripheral or central delivery of those cytokines to normal rodents increases their susceptibility to chronic stress. These findings have led to the novel idea of using peripheral cytokines as biomarkers of a subtype of depression and the potential utility of developing new antidepressants that oppose cytokine action.

Recent evidence has also linked proinflammatory signaling in the brain to addiction, particularly to alcohol. Human alcoholism is associated with impaired innate immunity, increases in circulating proinflammatory cytokines, and an increase in brain levels of the cytokine monocyte chemotactic protein-1 (MCP-1, also referred to as CCL2). Many of these cytokines are produced by astrocytes and microglia, and by neurons under certain pathologic conditions, where they play important roles in modifying neuronal function and plasticity. For example MCP-1 modulates the release of certain neurotransmitters and, when administered into the VTA, increases neuronal excitability, promotes dopamine release, and increases locomotor activity. Recent gene expression array studies of alcohol drinking in mice have identified a network of regulated cytokines in brain, and a role in regulation of alcohol consumption has been recently validated for several of them, including IL-6. A major focus of current research is to define the site and mechanism by which proinflammatory cytokines impair brain function to elicit a depressive episode or promote drug abuse.

This brief narrative illustrates the substantial progress that is being made in understanding the genetic and neurobiologic basis of mental illness. It is anticipated that biologic measures will be used increasingly to diagnosis and subtype a psychiatric disorder and that targeted therapeutics will become available to treat them.