How do the atypical antipsychotics work?

Abstract

Understanding the action of atypical antipsychotics is useful in exploring the pathophysiology of schizophrenia and in synthesizing drugs that improve various domains of psychopathology without unwanted side effects. In animal models, atypical antipsychotic drugs appear to have a preferential action in the limbic dopaminergic system. Regionally specific action has been studied by measuring the amount of Fos protein produced in a particular brain region as a consequence of a drug's effects on the c-fos gene. Evidence suggests that the atypical and typical antipsychotic drug-induced increases in Fos levels in the nucleus accumbens are related to improvements in positive symptoms, whereas Fos increases in the prefrontal cortex, with the atypical antipsychotics only, correlate with negative symptom improvement. The extrapyramidal effects seen with typical antipsychotics are thought to be related to Fos increases in the striatonigral pathway. However, studies of Fos levels in specific brain regions reveal only the site of action, not the mode of action. The finding that atypicality is related to surmountable D₂ dopamine receptor blocking provides another venue to define and explore atypical antipsychotic drug action.

Introduction

A growing number of atypical antipsychotic drugs have emerged over the last few years. Although these new drugs are lumped together as “atypical,” their widely varying chemical structure and pharmacological actions indicate their heterogeneity. Understanding the mode of action of the atypical antipsychotic drugs is useful in exploring the pathophysiology of schizophrenia and establishing realistic goals for treating schizophrenia.

Animal models of antipsychotic action

Generally, pharmacologic tests employed in animals for predicting the antipsychotic effects of drugs in humans are based on 2 principles.

· The ability of the drug to block the action of dopamine (DA) in animals: 2 standard tests are used to assess this. When a D₂ receptor antagonist is injected into a rat, it will remain immobile for an extended period of time when placed in an unnatural position. This immobility, called catalepsy, is caused by the blockade of DA receptors within the neostriatum and is considered to be a predictor of extrapyramidal symptom (EPS) inducing potential in humans. The second test, the blockade of a conditioned avoidance reaction, is extensively employed as a predictor of antipsychotic action. After an animal has learned to respond to a sensory cue and move away from a floor to avoid electric shock, antipsychotic agents will block its ability to respond to the sensory cue presented before the shock.1

· The ability to counteract the effects of dopaminergic drugs: the administration of d-amphetamine, a dopaminergic agent, produces locomotor stimulation and stereotypy in rats. Neuroleptic agents, as well as atypical antipsychotic drugs,2,3 decrease amphetamine-induced hyperactivity.

All antipsychotic drugs are screened in animals on these tests.

Until recently, it was thought that both cataleptic potential and blocking of a conditioned avoidance reaction in animals was necessary for predicting antipsychotic action in humans. The finding that atypical antipsychotic agents such as clozapine, remoxipride, thioridazine and olanzapine block conditioned avoidance in much smaller doses than those required to produce catalepsy3,4,5 helped to clarify the meaning of the animal pharmacologic tests. Clinically, these drugs have been proven to be effective for treating psychosis, and they produce fewer EPS in patients than neuroleptic agents (i.e, typical antipsychotics). In nonhuman primates, the effective antipsychotic agents such as clozapine and sertindole do not produce dystonia.6 Thus, catalepsy in animal pharmacologic tests predicts EPS, and blockade of conditioned avoidance predicts antipsychotic potential in humans.

The observation that injections of both typical and atypical antipsychotic agents into the nucleus accumbens (NAC) antagonizes amphetamine-induced locomotor activity suggests that this nucleus may be the site of antipsychotic action. Metoclopramide, a benzamide derivative, even in doses that produce catalepsy, cannot block this effect.7 This is not a potent antipsychotic drug, but it can induce EPS. Generally, atypical antipsychotic drugs can be distinguished by a greater degree of separation of dose–response curves for antipsychotic efficacy and EPS liability than typical antipsychotic drugs.8,9,10

Recently, other animal pharmacologic tests such as prepulse inhibition and paw tests have been introduced.11 Hopefully, in future, animal models will help to predict not only antipsychotic action and EPS potential but also effectiveness on cognitive, negative and mood measures. The neuroleptic agents and the atypical antipsychotic drugs may lie on the same continuum for EPS liability, but animal models indicate that parkinsonian symptoms are not necessary for antipsychotic action and have provided a framework for developing antipsychotic drugs with no EPS.

Antipsychotic action on specific brain sites

Proto-oncogenes, each of which serves a specific function, are normal cellular genes that cause cellular transformation. They are also called immediate early genes (IEGs) because in response to neurotransmitters and drugs, their transcription is transiently activated in neurons within minutes without new protein synthesis. These cellular IEGs include members of the related fos and jun families and zif/268. The fos family of genes has a prefix c in animals and v in viruses. The protein produced, Fos, is present in very small quantities in many neuronal cells under resting conditions. In response to a stimulus, c-fos mRNA is produced and translocated into the cytoplasm where it is translated into Fos protein. Fos protein is then translocated back to the nucleus where, in conjunction with a member of the Jun family, it binds to the activator protein-1 (AP-1) site on the promoter regions of the numerous target genes and regulates their expression. Diverse external stimuli can activate the IEGs by stimulating the production of a calmodulin-dependent protein kinase.

The proto-oncogene c-fos is a useful marker to map changes in neuronal activity.12,13 The ability of various neurotransmitters to increase c-fos expression in the central nervous system suggests that c-fos induction can occur as a consequence of synaptic activation. Fos binding to DNA changes the response of an organism for as long as a drug is given, but does not permanently change the gene. This phenomenon is called a phenocopy, in contrast to phenotype.

Neuroleptics and atypical antipsychotic drugs induce c-fos in various areas of the brain.14,15,16 This differential distribution is useful in identifying the brain regions that are targets for these drugs. Antipsychotic drugs have been reported to increase c-fos expression in the striatum, NAC, medial prefrontal cortex (PFC) and lateral septal nucleus. Recently, attention has focused on the thalamic region as well.

Striatum

The striatum is divided into the dorsolateral region, which includes caudate and putamen, and the ventromedial region, which interacts with sensorimotor and limbic association cortices. Haloperidol14 and raclopride17 induce Fos in the striatum. This effect is considered to be due to a blocking of DA receptors on striopallidal neurons.15,18,19 A number of additional findings support the possibility that Fos production is the direct result of D₂ blockade; in particular:

· coadministration of a D₂ receptor agonist prevents the induction of c-fos by haloperidol in the striatum;15

· dopaminergic agents such as cocaine and d-amphetamine decrease neuroleptic-induced striatal c-fos expression;18,20

· a close topographical relationship exists between neuronal Fos protein and D₂ receptor distribution;

· 6-hydroxydopamine (6-OHDA) lesions of the mesencephalic DA system block haloperidol-induced c-fos expression in the striatum and haloperidol- and clozapine-induced c-fos expression in the NAC.17

These results suggest that haloperidol increases the number of fos-positive neurons in the striatum by means of its D₂ receptor blocking ability.

Although evidence suggests that increased Fos production in dorsal striatum correlates with EPS, a recent study21 indicates that these findings may be applicable to short-term treatment only; Fos in the striatum was found to be downregulated by long-term treatment with clozapine and haloperidol. Haloperidol, but not clozapine, led to markedly enhanced Fos-B-like protein levels in the caudate putamen in rats. Thus, acute EPS may be related to Fos in the striatum, but long-term EPS may be related to Fos-B in caudate putamen. Whether tardive dyskinesia, a long-term EPS, is related to changes in caudate putamen is unknown.

An increase in striatal c-fos expression associated with the administration of antipsychotic agents in animals is predictive of EPS in humans, and low or no c-fos production in the striatum may be related to a low potential for producing EPS.22 The antimuscarinic agent, scopolamine, attenuates haloperidol-induced c-fos expression in the striatum. This suggests that the antimuscarinic action of clozapine may be responsible for its failure to induce c-fos in the striatum.23 However, because many other atypical antipsychotic agents that lack antimuscarinic action also fail to induce c-fos in the striatum, the antimuscarinic action of clozapine must not be the sole cause.

Enkephalin in the striatum

Enkephalin may be involved in regulating c-fos in the striatum and thereby affecting EPS. Haloperidol administration increases enkephalin mRNA in the striatal neurons,24 whereas clozapine causes a small increase. Moreover, after antipsychotic drug administration, c-fos expression and enkephalin increases occur in the same cells,14 suggesting that c-fos may be located in enkephalinergic neurons and that enkephalin may be a target for fos expression.

Nucleus accumbens

The NAC has 2 divisions: the shell and the core. The shell is allied with limbic circuits and the core, with the extrapyramidal system. All neuroleptic drugs increase c-fos in both the shell and the core of this nucleus.14,15 Atypical antipsychotic drugs, however, increase c-fos in the shell but not in the core. Wan et al25 compared the effects of the intraperitoneal administration of haloperidol, chlorpromazine, thioridazine, clozapine, raclopride, risperidone and ritanserin on Fos expression in the rat brain. Ritanserin did not increase Fos in any region, indicating 5-HT₂ serotonin receptor antagonism alone is not responsible for the observed effects of antipsychotic drugs. The single shared effect of all the antipsychotic preparations was the increased Fos-like immunoreactivity in the NAC. Metoclopramide, a drug that produces EPS without antipsychotic effects, induces Fos in the core but not in the shell, suggesting that Fos expression in the shell of the NAC may be related to clinical improvement of positive symptoms of schizophrenia.

All antipsychotic agents that have been tested act on the shell of the NAC and improve positive symptoms, but only the atypical antipsychotic agents act on negative symptoms. Accordingly, c-fos expression in the shell of the NAC is probably not associated with improvement of negative symptoms.

Neurotensin

Perikaryal neurotensin immunoreactivity is largely absent in the rat striatum except after striatal DA depletion or D₂ receptor blockade. After D₂ receptor blockade, neurotensin immunoreactivity occurs in 2 subpopulations of striatal neurons.26 One subpopulation, located mainly in the rostral, dorsomedial and ventromedial aspects of striatum, comprises moderate-to-large sized cells. These exhibit intense neurotensin immunoreactivity but rarely display Fos immunoreactivity.27,28 A second subpopulation, predominantly in the patch and matrix compartments in the lower quadrant of the striatum, is prominent after reserpine administration and shows very light neurotensin immunoreactivity.28

Haloperidol, chlorpromazine, pimozide and trifluoperazine29 induce Fos that is frequently colocalized with neurotensin in the dorsolateral striatum30 and the NAC.31,32 In contrast, clozapine has been found to increase neurotensin concentrations in the NAC only.33,34,35 With long-term treatment, haloperidol, but not clozapine, increases neurotensin in the globus pallidus.36 With sertindole, the effect in the NAC is dose dependent, and at higher doses, selectivity is lost.37 Intraventricular injections of neurotensin antagonize apomorphine-induced locomotor activity but not stereotypy in rats38 and selectively increase the number of spontaneously active neurons in A₁₀ but not A₉ areas,39 both of which are indicative of its antipsychotic activity without significant EPS.

Neuronal Fos-like immunoreactivity and neurotensin/neuromedin N mRNA are expressed in the same population of striatal neurons after a single injection of haloperidol.40 This finding represents an essential anatomical link between c-fos and neurotensin. Fos-like immunoreactivity was seen in neurons within 2 hours, and increases in neurotensin were noted 7 hours after a haloperidol injection.14,41 This suggests that Fos may be necessary for the induction of the neuromedin gene. Injection of an antisense oligonucleotide to c-fos blocks the haloperidol-induced increase in Fos and subsequent mRNA for neurotensin.30 Thus, Fos appears to be a second messenger in the induction of endogenous neuropeptides. The practical therapeutic use of neurotensin in psychosis should therefore be explored.

Medial prefrontal cortex

The prelimbic, dorsal and anterior cingulate, and medial precentral cortices collectively make up the PFC.42 These areas receive afferents from A₁₀ DA neurons in the ventral tegmentum of the midbrain.43 The infralimbic region also receives the same afferents as well those from the thalamic paraventricular nucleus.44 Stimulation of the dorsomedial thalamic nucleus increases the firing rate of PFC pyramidal cells.45,46 These medial thalamic projections to the frontal cortex are glutamatergic.46,47

Unlike haloperidol, clozapine increases Fos-positive neurons in the PFC. The Fos is selectively restricted to the pyramidal cells in the deeper layers of the ventral aspects of the rat PFC, including infralimbic and prelimbic cortex, but not the medial central cortex. Other atypical antipsychotic drugs (but not neuroleptics and remoxipride) also increase Fos in the PFC.48 Guo et al49 used brain lesion, pharmacologic and immunohistochemical techniques to explore the receptor mechanisms by which clozapine increases c-fos expression in the PFC. Reduction of the serotonin, norepinephrine or DA content by selective lesions in the brain did not decrease clozapine-induced c-fos expression in the rat PFC, suggesting it is unlikely that these mechanisms are involved in the clozapine-induced changes in c-fos expression in the PFC.17,49 An alternate explanation for this action has been put forward — that c-fos expression in the PFC is the drug-induced downstream effect of GABA afferents from the paraventricular nucleus of the thalamus to the PFC. Clozapine may exert its action on the PFC indirectly through afferent input from the medial thalamus, hippocampus, entorhinal cortex and basolateral amygdala.

The most likely site of action of atypical antipsychotic drugs on negative symptoms of schizophrenia is the PFC.17 This is based on the fact that atypical antipsychotic drugs that improve negative symptoms increase c-fos expression in the PFC in animal pharmacologic tests. However, the effectiveness of atypical drugs on primary negative symptoms has not been conclusively proven. A number of interpretations are possible in this regard — that c-fos expression in the PFC is:

· coincidental only, and it is the absence of EPS with atypical drug administration that is responsible for the apparent improvement of negative symptoms;

· an indirect effect of drug action elsewhere;

· directly correlated with improvement of negative symptoms. This assumption is supported by the findings that medial sulcal atrophy in the PFC decreases the effectiveness of clozapine on negative symptoms. Neuropsychologic testing and neuroimaging findings also point to a frontal lobe defect.50,51 Patients with schizophrenia have a defect in smooth eye tracking which is associated with the medial PFC. Because clozapine acts on the PFC and not on the medial PFC, it does not alleviate the defect in smooth eye tracking.

Lateral septal nucleus

Clozapine greatly increases the number of Fos immunoreactive neurons in the lateral septal nucleus as well.52 Haloperidol and not raclopride produced a modest increase in c-fos expression in the lateral septal nucleus, indicating that fos production may not be related to D₂ receptor blockade. Apart from risperidone, all of the antipsychotic drugs tested to date elevate Fos-like immunoreactivity in the lateral septal nucleus, suggesting that this limbic structure is important for antipsychotic activity.52 In contrast to the PFC, the lateral septum contains abundant 5-HT_1A receptors and very few 5-HT₂ receptors.53 Neither clozapine nor haloperidol appear to block 5-HT_1A. Therefore, the increase of Fos in the septal area may be due to mechanisms unrelated to 5-HT and DA.

Thalamic paraventricular nucleus

The paraventricular nucleus of the thalamus is in a pivotal location between the reticular formation and forebrain DA system and may serve an important role in the pathophysiology of schizophrenia. Clozapine administration increases Fos in the thalamic paraventricular nucleus, which provides glutamatergic projections to the PFC and the NAC; these projections may be associated with the observed c-fos induction in the PFC.17,54 Comparable doses of raclopride, sulpiride, remoxapride and haloperidol did not induce Fos in this region, whereas loxapine and very high doses of haloperidol induced a modest increase. Also, a D₁ antagonist did not induce Fos or alter clozapine-induced Fos.54 There is electrophysiological evidence that iontophoretically applied clozapine, but not haloperidol, can reduce the inhibitory actions of a 5-HT receptor agonist on pyramidal cells in the PFC.55

General pattern of c-fos distribution

Haloperidol administration induces c-fos expression in the NAC, lateral septal nucleus and striatum, and clozapine increases c-fos-positive neurons in the all the above-mentioned areas as well as the medial PFC.16,17,56,57 This suggests that the typical and atypical antipsychotic drugs evoke different patterns of neuronal activity and that the different clinical profiles of drugs may be related to the regionally different effects. D₂ receptor antagonism is not sufficient to explain the unique pattern produced by clozapine because it is common for both drug groups.58 The exact receptor mechanisms underlying antipsychotic drug-induced c-fos expression in the medial PFC and other regions remain unknown.

DA receptors

Although the relation between D₂ receptor blocking and antipsychotic potential is well established, the relation is not linear. In vitro studies provide some support for the finding that blocking over 80% of the receptors produces EPS and thus affects relative antipsychotic efficiency.59 Conversely, D₂ receptor occupancy of less than 80% produces atypical features, including low EPS.60 In a single-photon emission tomography (SPET) study employing a D₂ receptor ligand in EPS-free patients,61 olanzapine displayed a level of binding in the brain between that of haloperidol and clozapine. In striatal D₂ receptor binding, olanzapine was similar to clozapine.61 These results argue that moderate D₂ receptor antagonism is associated with fewer EPS. The looseness of attachment of an antipsychotic drug to the D₂ receptor seems to differentiate typical and atypical drugs.62,63,64 Typical drugs are firmly attached and atypical drugs are loosely attached to D₂ receptors, and thus, the latter do not induce EPS.

The actions of antipsychotic drugs on D₁, D₃ and D₄ receptors have also been explored. D₃ receptors are expressed mainly in the olfactory tubercle, NAC, island of Calleja and hypothalamus.65,66,67 D₄68 receptors have a unique distribution, with the highest levels in the PFC, and D₃ receptors are distributed in the NAC, major island of Calleja and lateral septal nucleus. Quinpirole, which has approximately equal antagonist affinities for D₃ and D₄ receptors, induced a significant decrease in clozapine-induced c-fos expression in PFC, NAC, major island of Calleja and lateral septal nucleus. In contrast, the more selective D₃ receptor agonist 7-OH-DPAT69,70 significantly reduced clozapine-induced increases in c-fos expression (Fos protein levels) in the major island of Calleja, NAC and lateral septal nucleus, with no effect on the Fos level in the PFC. These data suggest that the action on D₃ receptors may mediate the c-fos effects in the NAC, major island of Calleja and lateral septal nucleus, whereas D₄ receptors may be responsible for the action in PFC.

Based on the selective antagonism of clozapine at D₄ compared with D₂ receptors, expectations that D₄ receptor blockers would improve schizophrenia ran high.71 Roth et al72 assessed the affinities of 13 atypical and 12 typical antipsychotic drugs on D₄ receptors and examined D₄/D₂ ratios. Many atypical drugs (e.g., melperone, quetiapine, fluperlapine) had very little D₄ receptor blocking activity, and many typical (e.g., loxapine, chlorpromazine, fluphenazine, mesoridazine, thioridazine, trifluoperazine) and some atypical drugs (i.e., olanzapine, clozapine, risperidone, zotepine, tiospirone) drugs had high D₄ affinities. Interestingly, the ratios of D₂/D₄ did not differentiate the typical and atypical agents but 5-HT_2A/D₂ ratios did. Moreover, a selective D₄ receptor antagonist, L-745,870, was found to be ineffective in the treatment of schizophrenia in a placebo-controlled study.73

The role of D₁ receptors in the pathophysiology of schizophrenia is also unclear. Repeated administration of cocaine and amphetamine produces long-term behavioural changes as well as addiction. Cocaine enhances extracellular DA by blocking reuptake, and amphetamine does the same by increasing DA release. The effects of psychomotor stimulants occur in striatal neurons that selectively express the D₁ receptor subtype.74,75,76,77 Amphetamine or cocaine administration in the D₁-receptor-deficient mouse77 does not induce Fos-like activity, and pure D₁ receptor antagonists have been reported to produce catalepsy in rats78,79 and dystonia in monkeys primed with typical neuroleptics.80 These findings suggest that D₁ receptors may have a role in producing EPS or that D₂ receptors may act through D₁ receptors to produce EPS.

Serotonin receptors

Atypical antipsychotics block serotonin 5-HT₂ receptors.58,81,82,83 When the ratio of 5-HT₂ to D₂ receptor blocking is greater than 1, atypical antipsychotic action such as therapeutic effects on negative symptoms and few EPS are noted. Although 5-HT₂84 receptor blockade and atypical profile correlate, the exact mechanisms by which 5-HT2 blocking improves negative symptoms and induces fewer EPS are unclear. It will be recalled that positive symptoms are associated with a hyperdopaminergic state in the limbic lobe, which is rich in dopaminergic innervation. Serotonin inhibits DA release, and in the limbic lobe, with high 5-HT₂ and low D₂ receptor density, D₂ receptor blocking action prevails and positive symptoms are controlled. Negative symptoms, on the other hand, are associated with a hypodopaminergic state in the frontal lobe, which is rich in serotonergic innervation. Here, with the rich distribution of 5-HT₂ and sparse distribution of D₂ receptors,85 serotonin inhibits DA release and the hypodopaminergic state of the frontal lobe becomes normal, thereby improving negative symptoms. Thus, the differential distribution of D₂ and 5-HT₂ receptors explains why atypical antipsychotic drugs exert opposite action on DA transmission in the frontal and limbic lobes.

Given that α ₁ adrenergic86,87 and α ₂ adrenergic88 receptor blockade has also been proposed to be involved in the mechanism of action of the new antipsychotics, “multireceptor” action may actually explain the unique clinical profile of atypical antipsychotic agents (Table 1). Clozapine, which acts on many different receptor types, has been proven to be the clinically most effective drug with the least EPS. Olanzapine also exhibits multireceptor action, whereas risperidone and sertindole have predominantly D₂ and 5-HT₂ receptor blocking effects. Any subtle differences in the action of the 2 groups of drugs remain to be elucidated.

Table 1

Electrophysiological activity

Individual DA cells show both tonic and phasic activity. Spikes that represent the firing of the single neuron form the background noise or tonic activity against which bursts of 5–20 spikes/s form the signals or phasic activity. When phencyclidine, which mimics schizophrenia, is injected into an animal, DA neurons in the ventral tegmental area show increased tonic activity and decreased phasic burst activity. This suggests that the positive symptoms of schizophrenia may be due to increased tonic activity, bringing a low signal-to-noise ratio. Atypical antipsychotic drugs act on dorsal and ventral tegmental areas differently; they downregulate the firing rates in the ventral but not the dorsal tegmental area, and this has 2 important effects: the hyperdopaminergic state in the limbic area is normalized, and the extrapyramidal system is not affected (i.e., no EPS). Typical antipsychotics increase the actively firing cells in both A₉ and A₁₀, whereas atypical antipsychotics affect the A₁₀ areas only.89 Long-term administration of typical antipsychotic drugs decreases the firing of both A₉ and A₁₀ areas due to depolarization block. Atypical drugs quetiapine,90 olanzapine,4 sertindole91 induce depolarization block in A₁₀ only, and thus, the antipsychotic effect is produced without EPS.92,93

Conclusions

With knowledge of the intracellular and genetic action of antipsychotic drugs, the site of action of these drugs is better understood. Evidence suggests that D₂ receptors are related to improvement in positive symptoms, and such an action may reside in the shell region of the NAC. Similarly, data indicate that the medial PFC may mediate antipsychotic effects on negative symptoms. However, other possibilities should also be entertained; other regions may be involved as well, and these effects may be mediated by some not-yet-defined afferents. The brain regions associated with cognitive symptoms remain to be delineated.

Although c-fos genes provide some structural basis for the symptoms of schizophrenia, the action on the receptors are not clear. D₂, D₃ or D₄ receptors, or all of them in combination, may be involved in the alleviation of positive symptoms, and negative symptoms may be due to a hypodopaminergic state in the frontal lobe; 5-HT receptors may play a role in this regard. On the other hand, it is likely that the selective action of the new drugs on the ventral and not dorsal tegmental nucleus and the effect on the paraventricular nucleus of thalamus may also play a part in improving negative symptoms.

Robertson and colleagues52 proposed a new definition of atypicality. They suggest that if the number of Fos-positive cells in the NAC is greater than the number of Fos-positive cells in the striatum, the drug should be considered atypical. This index allows atypicality to be quantified, but its usefulness requires further validation. In addition, the concept that the ability of the drug to bind to D₂ receptors in a reversible and surmountable fashion may also contribute to atypicality should be studied further.94