| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Article |
Department of Pharmacology and Human Physiology, University of Bari Medical School, Bari, Italy (L.G.C.); Department of Pharmacological Sciences, University of Palermo, Palermo, Italy (L.S.); Department of Pharmacology and Physiology, University of Roma "La Sapienza," Rome, Italy (V.C.)
Abstract
Abstract I. Introduction II. Neurotransmitters and Brain Development A. Dopaminergic System B. Serotonergic System C. Noradrenergic System D. GABAergic System III. Antipsychotics and Antiepileptics A. Antipsychotic Drugs 1. Typical Antipsychotics. 2. Atypical Antipsychotics. B. Antiepileptic Drugs 1. Valproic Acid. 2. Phenytoin. 3. Phenobarbital. 4. Carbamazepine. 5. ''New'' Antiepileptics. IV. Anxiolytics and Mood Stabilizers A. Benzodiazepines B. Lithium V. Antidepressants A. Tricyclic and Atypical Antidepressants B. Monoamine Oxidase Inhibitors C. Selective Serotonin Reuptake Inhibitors D. Novel Antidepressants VI. Neuroactive Herbal Drugs A. St. John's Wort B. Ginkgo Biloba C. Kava D. Valerian VII. Future Prospects and Research Needs
The advent of psychotherapeutic drugs has enabled management of mental illness and other neurological problems such as epilepsy in the general population, without requiring hospitalization. The success of these drugs in controlling symptoms has led to their widespread use in the vulnerable population of pregnant women as well, where the potential embryotoxicity of the drugs has to be weighed against the potential problems of the maternal neurological state. This review focuses on the developmental toxicity and neurotoxicity of five broad categories of widely available psychotherapeutic drugs: the neuroleptics, the antiepileptics, the antidepressants, the anxiolytics and mood stabilizers, and a newly emerging class of nonprescription drugs, the herbal remedies. A brief review of nervous system development during gestation and following parturition in mammals is provided, with a description of the development of neurochemical pathways that may be involved in the action of the psychotherapeutic agents. A thorough discussion of animal research and human clinical studies is used to determine the risk associated with the use of each drug category. The potential risks to the fetus, as demonstrated in well described neurotoxicity studies in animals, are contrasted with the often negative findings in the still limited human studies. The potential risk for the human fetus in the continued use of these chemicals without more adequate research is also addressed. The direction of future research using psychotherapeutic drugs should more closely parallel the methodology developed in the animal laboratories, especially since these models have already been used extremely successfully in specific instances in the investigation of neurotoxic agents.
The thalidomide disaster has heightened public awareness of the deleterious effects of medicinal drugs on the developing fetus, and most women prefer not to use any medication, if at all possible, during pregnancy. It is, however, believed that a large percentage of women (up to 90% according to some estimates; Altshuler and Szuba, 1994
) take one or more drugs during pregnancy, and of these, psychoactive compounds account for at least one-third of the drugs (Table 1) (Ashton, 1991; Arnon et al., 2000). Psychiatric disorders, particularly mood and anxiety disorders, are common in women of reproductive age, and some cases may be first diagnosed during pregnancy (Altshuler and Szuba, 1994; Kuller et al., 1996). The evidence that many women develop or have recurrence of psychiatric diseases during pregnancy or lactation does not support the once hypothesized notion that emotional and psychological changes associated with maternity can confer protection against onset or relapse of such illnesses (Altshuler et al., 1996; Arnon et al., 2000). There is sufficient evidence that all psychotropic drugs readily cross the placenta to reach the fetus and may also be excreted into breast milk (Chisholm and Kuller, 1997; Arnon et al., 2000; Bar-Oz et al., 2000). Drugs in the fetus may have a higher unbound free fraction, easily penetrate into the brain, and undergo only limited hepatic and/or extrahepatic metabolism (Arnon et al., 2000; Hines and McCarver, 2002; McCarver and Hines, 2002). Thus, the decision to initiate or continue pharmacotherapy during pregnancy and puerperium requires thoughtful weighing of the potential adverse effects of embryo, fetus, or infant exposure to psychotherapeutic drugs against the risks, for both mother and offspring, of untreated mental disorders (Altshuler et al., 1996; Kuller et al., 1996; Koren et al., 1998).
|
New drugs are typically not tested before marketing in pregnant women to determine effects on the fetus, although developmental toxicology and teratology studies in animals are required (Koren et al., 1998). Typically, a general statement is made such as "Use in pregnancy is not recommended unless the potential benefits justify the potential risks to the fetus" (Koren et al., 1998). For a number of psychotherapeutic drugs, harmful effects on the developing embryo, fetus, and child are known, but for several others there is still insufficient information. Potential adverse effects on embryonic, fetal, and neonatal development induced by exposure to pharmacotherapy include classic teratogenicity as well as more subtle developmental effects. Teratogenicity is usually associated with structural abnormalities induced by exogenous compounds during organogenesis; thalidomide, which caused severe limb defects and other organ dysgenesis, or isotretinoin, which caused a wide variety of CNS1, craniofacial, and cardiovascular defects represent two examples of classic teratogens.
The fields of behavioral teratology and neurobehavioral toxicology have arisen during the past 30 years to allow researchers, particularly those using animal models of perinatal exposure to chemicals, to examine with well defined methodologies the more subtle and more long-lasting effects of such exposure. These methodologies, described in detail by various authors (Annau, 1986; Riley and Vorhees, 1986; Cuomo et al., 1996, Bignami, 1996), have proven to be extremely useful in revealing subtle postnatal effects of prenatal toxic exposures and, in conjunction with pharmacological challenges, in identifying underlying neurochemical alterations. These behavioral teratology studies have been applied in some instances to human populations, as in the case of lead (Needleman and Bellinger, 1994), but they have not yet become a required component of premarket testing of new pharmaceuticals. Given the vulnerability of the pregnant human population, it is difficult to conceive how premarket testing could be realized, but nevertheless, as the review of the literature will indicate, it is exactly the potential vulnerability of this population that needs to be addressed in the future. The types of behavioral effects most often observed in animal experiments following prenatal exposure to chemicals, in particular neurotoxic chemicals, are short- or long-term cognitive impairment, alterations in diurnal rhythms, emotional reactivity, and alterations of normal motor development. It is important to note that these behavioral effects can be seen in the offspring of treated mothers at doses that do not elicit either maternal toxicity or morphologic alterations in the neonates. Psychotherapeutic drugs, which target the CNS, are particularly prone to such neurofunctional/neurobehavioral teratogenic effects (Cuomo, 1987; Mantovani and Calamandrei, 2001). This review focuses on the effects of psychiatric drugs on the development of the fetus and the newborn. Animal and human studies are discussed, with an emphasis on structural teratogenic effects and biochemical and neurobehavioral alterations, as well as any data that may shed light on the mechanisms underlying developmental dysfunctions.
II. Neurotransmitters and Brain Development
To gain an understanding of the short- and long-term deleterious effects resulting from any interference with brain development, one must know the nature of the interference as well as the nature of the organ at the time of insult (Rodier, 1980). From a large number of studies carried out mostly in rodents over the past 40 years, a great deal has been learned about the development of the brain (Dobbing and Sands, 1973; Dobbing, 1974; Rodier, 1980; Smart, 1991; Bayer et al., 1993). From these studies, one can infer the various stages of brain development in humans, although there are variations in the rates of brain growth among mammals, mostly dependent upon the length of gestation (Passingham, 1985; Bayer et al., 1993). Thus, the developmental ages of human and rat embryos or fetuses are comparable when major gross anatomical features and histological landmarks are similar in appearance in the two species, although their exact chronological ages are different (Bayer et al., 1993).
A first important general concept is that different parts of the central nervous system form at different stages of development; thus, there is not one critical (or safe) period, but many critical periods where exogenous compounds can exert deleterious effects. Using [3H]thymidine autoradiography, the neurogenesis of specific populations of neurons was mapped in rodent brain, and extrapolations were made to the human brain (Rodier, 1980; Bayer et al., 1993). It is beyond the scope of this review to discuss these aspects in detail, except for pointing out that different brain areas develop at different times during gestation. Additionally, within a single brain region, subpopulations of neurons develop at different rates and at different times. Production of certain neurons can occur in very short intervals (a few days), whereas longer proliferative periods exist for other neurons (Rodier, 1980). For example, in the hippocampal region, neurons in the CA1 field develop on embryonic days 17 to 20 in the rat (corresponding to gestational weeks 7.5-15 in humans), whereas dentate granule cells develop later (embryonic day 20 to postnatal day 15 in the rat, corresponding to gestational weeks 15-36 in humans) (Bayer et al., 1993). In the cerebellum, Purkinje cells develop early (embryonic days 13-15 in the rat corresponding to 5-7 weeks in humans), whereas granule cells are generated much later (postnatal days 4-19 in rats, equivalent to gestational weeks 24-40 in humans) (Bayer et al., 1993).
An additional important aspect of brain development is the so-called "brain growth spurt," a transient period of growth when the brain is growing most rapidly (Dobbing and Sands, 1973). This occurs in the first 2 postnatal weeks in the rat and in the third trimester of pregnancy and in early infancy in humans (Dobbing, 1974). One of the general features of brain growth throughout mammalian species is that adult neuronal cell number is almost accomplished (with the notable exception of cerebellar granule cells and few other neurons), before the major phase of glial multiplication begins (Dobbing, 1974). The brain growth spurt is indeed characterized by rapid proliferation of glial cells, most notably astrocytes and oligodendrocytes. In addition to axonal myelination, this period also includes synaptogenesis and definition of the brain's cytoarchitecture. Most neurotransmitter systems do indeed develop during this time frame, including synthetic and degrading enzymes, uptake systems, and receptors (Coyle, 1977; Retz et al., 1996).
A very large number of studies have been published on the development of neurotransmitter systems, mostly in rodents. In general, levels of neurotransmitters are low at birth and reach adult levels by postnatal weeks 4 to 6 (Broening and Slikker, 1998). However, notable exceptions exist; for example, high levels of GABA and acetylcholine are already present at birth (Costa, 1993; De Blas, 1993). Enzymatic systems that synthesize neurotransmitters as well as uptake systems also develop during the first 3 to 4 postnatal weeks in the rat (Costa, 1993; Broening and Slikker, 1998; Varju et al., 2001), as do most receptor systems and receptor-activated signaling pathways (Jett, 1988; Duman and Alvaro, 1993; Costa, 1998; Vallano, 1998; Rho and Storey, 2001).
Thus, the development of neurotransmitter systems that may be targeted by psychotherapeutic drugs occurs mostly postnatally in the rat and coincides with synaptogenesis and the development of neurotransmission. However, various lines of evidence suggest that neurotransmitters may have several important roles in brain development in addition to neurotransmission (Buznikov, 1984; Lauder, 1988; Emerit et al., 1992; Johnston, 1995; Retz et al., 1996; Levitt et al., 1997; Contestabile, 2000; Nguyen et al., 2001). First, the appearance of one or more key components of neurotransmission may precede synaptogenesis; for example GABA, glutamate decarboxylase (GAD), and GABAA receptors are present in embryonic neurons well before the development of GABAergic synapses (Kim et al., 1996). Second, a transient overexpression of receptors at certain developmental stages, before a decrease to adult levels, as in the case, for example, of glutamate N-methyl-D-aspartate receptors or some dopamine receptors, suggests a role for certain neurotransmitters beyond neurotransmission itself (Retz et al., 1996; Broening and Slikker, 1998). Third, coupling of receptors to signal transduction systems may also show a peculiar developmental profile, as seen, for example, in the case of muscarinic and metabotropic glutamate receptor coupling to phospholipase C (Balduini et al., 1991; Costa, 1998).
Evidence for a number of roles of neurotransmitters in brain development has emerged from a variety of animal species and experimental in vitro and in vivo models, including transgenic animals as well as studies with selective neurotoxicants. Although a complete understanding of how neurotransmitters would ultimately shape brain development in vivo has still not been achieved, these studies clearly show that these molecules, in conjunction with growth factors and cytokines, can exert profound effects on the proliferation and maturation of neuronal and glial cells. Such effects range from modulation of proliferation of neuronal stem cells, neuroblasts, and glioblasts (Azmitia, 2001; Nguyen et al., 2001; Varju et al., 2001), to regulation of migration and induction of differentiation (Retz et al., 1996; Levitt et al., 1997; Azmitia, 2001). Neurotransmitters can also act as trophic factors modulating the apoptotic processes that are known to occur at certain stages of brain development (Emerit et al., 1992; Ikonomidou et al., 2001). Such effects have been described for most neurotransmitter systems that may be affected by psychotherapeutic drugs, including serotonin (Emerit et al., 1992; Azmitia, 2001; Lesch, 2001; Nguyen et al., 2001; Okado et al., 2001; Rho and Storey, 2001), GABA (Levitt et al., 1997; Ikonomidou et al., 2001; Nguyen et al., 2001; Varju et al., 2001), dopamine (Levitt et al., 1997; Rho and Storey, 2001), norepinephrine (Duman and Alvaro, 1993; Rho and Storey, 2001), acetylcholine (Costa, 1993; Nguyen et al., 2001; Rho and Storey, 2001), and glutamate (McDonald and Johnston, 1990; Contestabile, 2000; Ikonomidou et al., 2001). The major classes of psychotherapeutic drugs target four neurotransmitter systems (dopamine, serotonin, norepinephrine, and GABA), and changes in various parameters of these systems have been reported in animals perinatally exposed to these drugs. Therefore their ontogenesis is discussed in more detail.
Dopamine-containing cells can be detected in the rat brainstem by gestational day (GD) 12 or 13; soon after, they begin to sprout axons that reach the telencephalon at GD 14 (Voorn et al., 1988). Fibers arising from the substantia nigra and pars compacta can be visualized in the maturing striatum, including the primordium of nucleus accumbens, by GD 15 and 16 (Voorn et al., 1988). By GD 19, dopaminergic axons have traced out distinct pathways to most areas of dorsal and ventral striatum; at the same time, a patch-matrix type morphology can be recognized in the developing caudate putamen (Specht et al., 1981a,b). Fibers with varicosities in the striatum, already present at postnatal day (PD) 2, gradually increase through the first 2 weeks to achieve their adult feature by PD 21. Similarly, dopaminergic axons emerging from the ventral tegmental area reach the septum and the prefrontal cortex subplate at GD 16 and 17 and the cingulate cortex at GD 20 (Verney et al., 1982; Kalsbeek et al., 1988). Dopaminergic innervation to the neocortex starts resembling the adult pattern in density and distribution by PD 12. However, the adult morphology and organization are completely achieved only at the end of the first month of postnatal life (Kalsbeek et al., 1988).
Evidence that dopamine receptors are found early in brain development, before the formation of subcortical and cortical synaptic connections, suggests that dopamine, acting through its receptors, may play an important function in neural development (Todd, 1992; Castro et al., 1994; Swarzenski et al., 1994; Lidow and Wang, 1995). Convincing results suggest that, during brain maturation, dopamine may have a modulatory role in neuronal growth and in modeling neuronal and synaptic architecture (Lankford et al., 1988; Murrin and Zeng, 1990). In retinal neurons, stimulation of D1 receptors inhibits neurite outgrowth (Lankford et al., 1988; Murrin and Zeng, 1990), whereas in cortical and mesencephalic neurons activation of D2-like receptors increases the extension and branching of neurites (Todd, 1992). Moreover, stimulation of D4 receptors results in the dramatic increase in neurite length in the transfected clonal specific MN9D cell line (Swarzenski et al., 1994). Density of D1 dopamine receptors in rat striatum is approximately 10% of the adult value at birth (Jung and Bennett, 1996). During postnatal development, a steady increase in the density of both D1 and D2 dopamine receptor subtypes occurs, with a greater prevalence of D1 and D2 dopamine receptor binding sites around the time of weaning (Murrin, 1986;Gelbard et al., 1989; Murrin and Zeng, 1990; Rao et al., 1991). By the end of the second postnatal week, D1 receptor density begins to approximate the adult value (Leslie et al., 1991). Evidence also exists that D1 receptors in the prefrontal cortex achieve the adult topological pattern and density early after the birth (Leslie et al., 1991). Studies examining the ontogeny of dopamine D2 receptors have reported that significant levels of receptors are expressed by PD 3, and adult levels are reached by PD 21 (Rao et al., 1991). Forebrain dopamine D3 receptors appear to be expressed later in development than D2 receptors in the same regions. Dopamine D3 binding sites are absent at PD 3 and just detectable at PD 7 and PD 10. Appreciable D3 labeling appears in the islands of Calleja at PD 14 and in the nucleus accumbens at PD 21 (Demotes-Mainard et al., 1996; Stanwood et al., 1997). Using the polymerase chain reaction technique, it has been reported that the developmental ontogeny of D4 receptor mRNA does not correlate with the ontogeny of the D2 dopamine receptor mRNA. Indeed, the level of expression of the D4 receptor mRNA is appreciable at birth, increases to a maximum at PD 3, and declines at PD 28, whereas levels of dopamine D2 receptor mRNA are highest on PD 28 (Nair and Mishra, 1995). Information on the expression of D5 dopamine receptors in the embryonic rat brain is still insufficient; however, in the fetal primate brain, many cortical cells express D1 and D5 dopamine receptors (Lidow and Wang, 1995; Wang et al., 1997a). Dopamine D1 and D5 receptors are differently distributed, suggesting that they may play different roles in cerebral developmental processes. In particular, as D5 dopamine receptors have a higher affinity for dopamine then D1 receptors, they may be more suitable for nonsynaptic interactions, given the low level of dopamine present in the intercellular space of the fetal brain (Lidow, 1995; Lidow and Wang, 1995; Wang et al., 1997a).
Similar to dopamine, the early expression of serotonin [5-hydroxytryptamine (5-HT)] and its receptors in the developing brain has brought attention to its potential contribution in modulating neuronal developmental processes. In this context, it has been reported that parachlorophenylalanine, a 5-HT synthesis inhibitor, retarded neuronal maturation (Lauder and Krebs, 1978), and that the transient excess of serotonin during prenatal life in knock-out mice lacking monoamino-oxidase A resulted in a disrupted bamfield organization in the primary somatosensory cortex (Cases et al., 1996). In vitro experiments have shown that 5-HT regulates neuritic growth and synapse formation (Chubakov et al., 1986; Haydon et al., 1987). In undifferentiated neuroblastoma cells, high levels of 5-HT (50 µM) induce a decrease, whereas low levels (50 nM) induce an increase in the cytoplasmic tau protein (John et al., 1991). Thus, there is evidence that 5-HT plays a role in a variety of cellular processes involved in regulating metabolism, proliferation, and morphology of neurons. The fine integration of these dynamic events appears to involve multiple receptor action. Serotonergic neurons begin to sprout axons by GD 15 (Lidov and Molliver, 1982a,b; Wallace and Lauder, 1983). These axons grow rapidly, and by GD 17 serotonergic axons enter the basal forebrain, with some fibers reaching as far forward as the septum and the frontal pole of the neocortex (Lauder et al., 1982; Lidov and Molliver, 1982a; Wallace and Lauder, 1983). By GD 19 serotonergic axons have established pathways to all major divisions of the forebrain in the rat (Wallace and Lauder, 1983). Axon pathways increase in density, and terminal fields begin to appear by GD 21 (Lidov and Molliver, 1982a). Terminal field innervation continues into the postnatal developmental period and is the main feature of postnatal serotonergic development. By PD 3, elaboration of the serotonergic neuropil is underway in most cortical regions and in the hippocampus (Lidov and Molliver, 1982a; Dori et al., 1996). At this developmental stage, innervation in most brainstem regions is quite dense, and patterns of innervation have begun to resemble those present in the adult.
In the rat, whole-brain 5-HT1 serotonergic receptor density is approximately 45% of the adult value at birth, but it is only 24% in the frontal cortex (Zilles et al., 1985). Several subtypes of 5-HT1 receptors have been identified, including the 5-HT1A, 5-HT1B, and 5-HT1D receptors (Hoyer et al., 1994). The 5-HT1A receptor develops early in the CNS and is associated with secretion of S-100
from astrocytes and reduction of cAMP levels in neurons. These actions provide intracellular stability for the cytoskeleton and result in cell differentiation and cessation of proliferation (Azmitia, 2001). In cerebral cortex, 5-HT1B receptor density is 31% of the adult value at 5 days of age and reaches 65% of the adult value by 3 weeks. Whole-brain 5-HT2A serotonergic receptor density is low in the perinatal period (17% of the adult value 2 days after birth) (Bruinink et al., 1983; Roth et al., 1991) but reaches 76% of the adult value at 2 weeks, and 100% at 4 weeks of age. Thus, 5-HT2A receptors develop more slowly and are associated with glycogenolysis in astrocytes and increased calcium availability in neurons. These actions destabilize the internal cytoskeleton and result in cell proliferation, synaptogenesis, and apoptosis (Azmitia, 2001). Whole-brain 5-HT2C receptor density reaches adult values by PD 5 (Roth et al., 1991). However, 5-HT2C receptor density in cerebral cortex is only approximately 24% of the adult value at PD 1 and reaches 37% of the adult value by PD 10, 76% of the adult value by 2 weeks, and adult values by 3 weeks of age (Pranzatelli, 1993).
The development of the high-affinity 5-HT reuptake transporter has been studied by measuring [3H]5-HT uptake and [3H]paroxetine binding. Cortical [3H]5-HT uptake ranges from 19 to 22% of the adult value at birth (Kirksey and Slotkin, 1979; Huether et al., 1992) and reaches 42 to 54% of the adult value by 2 weeks of age; by 4 to 5 weeks of age, [3H]5-HT uptake in the cortex approaches adult values. Investigation of the ontogeny of the high-affinity [3H]5-HT uptake transporter by [3H]paroxetine binding discloses a somewhat faster developmental time course compared with that observed in studies using [3H]5-HT uptake. Indeed, cortical [3H]paroxetine binding to the 5-HT uptake transporter is 39% of the adult values by the end of the first postnatal week and reaches the adult value by 2 weeks of age (Pranzatelli and Martens, 1992). The widespread distribution of the 5-HT transporter during ontogeny, regulating 5-HT levels in the neuronal microenvironment, confirms the important role of serotonin in diverse physiological processes during embryonic development. If 5-HT is indeed widely expressed in embryos, teratogenic effects or more severe neurofunctional sequelae would be expected as a result of in utero exposure to agents able to inhibit its function. Although only limited evidence for detrimental effects in intact animals or in humans is currently available, effects might be seen later or be more subtle, or perturbation of the serotonin system might result in compensatory responses either in the 5-HT system or in interacting related systems. However, it has been reported that 5-HT transporter-deficient mice do not exhibit any alteration, even if there are compensatory changes in the 5-HT system, with a desensitized response to the 5-HT1A agonists despite normal levels of receptors (Wichems et al., 1997).
Noradrenergic innervation from the locus coeruleus (LC) occurs very early in the development of mammalian brain (Levitt and Moore, 1978). In the rat, noradrenergic neurons differentiate at or before GD 12 and give rise to projections shortly thereafter (Specht et al., 1981), reaching their destination before the differentiation of target neurons (Schlumpf et al., 1980). Because of the early presence of norepinephrine (NE) in the developing brain, it has been suggested that the adrenergic system regulates several aspects of pre- and postnatal brain development, including cell division, neuronal maturation, synaptogenesis, and physiological plasticity (Blue and Parnavelas, 1982; Slotkin et al., 1988, 1994). Axons arise from the noradrenergic perikarya in the developing LC at GD 14. By GD 15, these axons extend into the ventral mesencephalon and the dorsal pons to form the nascent ventral and dorsal noradrenergic bundles, respectively (Specht et al., 1981). Noradrenergic axons simultaneously innervate the medial and lateral cortex at this age; however, the dorsal cortex is not innervated until GD 19 (Levitt and Moore, 1978; Berger and Verney, 1984; Verney et al., 1984). GD 18 first identifies noradrenergic fibers identified in the hippocampus. Noradrenergic axons move laterally under the anterior commissure at GD 20 and yield the marginal and intermediate zones of the lateral frontal cortex. Medially, noradrenergic axons course through the dorsal diagonal band and rostrally to it via the developing medial forebrain bundle (Levitt and Moore, 1978; Verney et al., 1984). On arriving to the corpus callosum, these axons diverge into two axon bundles: one running above the corpus callosum penetrating the cingulate cortex, and the other coursing below the corpus callosum and entering the septum (Berger et al., 1983; Verney et al., 1984). The morphology of the noradrenergic axons begins to modify from thick, straight fibers to thin, varicose fibers by GD 20 (Berger and Verney, 1984; Verney et al., 1984). Noradrenergic fibers innervate all layers of the neocortex by PD 7 (Lidov et al., 1978; Levitt and Moore, 1979; Verney et al., 1982; Berger et al., 1983), and noradrenergic innervation to this brain area resembles that of adults by PD 14 (Levitt and Moore, 1979; Berger et al., 1983). Thus, noradrenergic innervation to the forebrain matures at an earlier age than the dopaminergic and serotoninergic innervation to the forebrain.
Whole-brain NE uptake ranges from 13 to 30% of the adult value at birth (Coyle and Axelrod, 1971; Kirksey et al., 1978) and increases rapidly, so that 70 to 100% of the adult value is reached by the end of the second postnatal week. Cortical [3H]NE uptake also develops early during postnatal life in the rat, which is compatible with the early innervation of the cortex by noradrenergic fibers. Cortical [3H]NE uptake ranges from 13 to 15% of the adult value at birth (Levitt and Moore, 1979) and reaches adult values by the end of the third postnatal week. There may also be a short-lasting overexpression of [3H]NE uptake in frontal cortex, as uptake in this brain region has been reported to exceed 200% of the adult value at 2 weeks of age (Levitt and Moore, 1979). It has been recently reported that fibroblast growth factor-2, neurotrophin-3, and transforming growth factor-1 regulate norepinephrine transporter (NET) expression in cultured neural crest cells by causing an increase in NET mRNA levels (Sieber-Blum and Ren, 2000). They also promote NET function in both neural crest cells and presumptive noradrenergic cells of the LC. The growth factors are synthetized by the neural crest cells and, therefore, are likely to have autocrine functions. NE transport regulates differentiation of noradrenergic neurons in the peripheral nervous system and the LC by promoting expression of tyrosine-hydroxylase and dopamine-beta hydroxylase. Conversely, uptake inhibitors, such as the tricyclic antidepressants and other NET inhibitors, inhibit noradrenergic differentiation in both tissues. Thus, growing evidence suggests that: 1) NET is expressed early in embryonic development; 2) NE transport is involved in regulating expression of the noradrenergic phenotype in the peripheral and central nervous system; and 3) norepinephrine uptake inhibitors can deeply disturb noradrenergic cell differentiation in the sympathetic ganglion and LC (Sieber-Blum and Ren, 2000).
An increasing body of data indicates that 
1-adrenergic receptors are present in cultured rat cortical neurons at an early development stage. The number of 1 clusters gradually increases on both cell bodies and neuronal processes in the culture environment from day 0 to day 20. Interestingly, it has been shown that their expression is developmentally regulated and that both neuronal activity and receptor occupancy influence receptor expression; however, neuronal activity dominates over receptor occupancy in the regulation of receptor expression (Wang et al., 1997b). The receptors are mainly expressed on the cell body in the early stages of the cortical cultures and later along neuronal processes. Moreover, receptor binding studies using [3H]prazosin to label 1-adrenergic receptors in the rat cerebral cortex have shown a progressive increase in receptor density during postnatal development (Schoepp and Rutledge, 1985; Slotkin et al., 1990), with 54% of adult value by 2 weeks and full adult value by 3 weeks of age (Schoepp and Rutledge, 1985; Slotkin et al., 1990). The 2A-adrenoceptor subtype is widely expressed during periods of neuronal migration and differentiation throughout the developing brain; both 2A receptor mRNA and protein expression are strongly expressed by GD 19 and GD 20, respectively (Winzer-Serhan et al., 1997a,1997b). The increased expression occurs in the cortical plate and intermediate and subventricular zones, corresponding to tiers of migrating and differentiating neurons. This transient up-regulation of 2A-adrenoceptors is restricted to the lateral neocortex. At GD 20 functional 2A-adrenoceptors are also detected in deep layers of lateral neocortex. During the first week of postnatal development, the expression of 2A receptor mRNA and protein changes markedly, giving rise to a more mature pattern of anatomical distribution. The temporal and spatial distribution of 2A-adrenoceptors in developing neocortex is consistent with expression of functional proteins on migrating and differentiating layer IV to II neurons, suggesting that these receptors may mediate a neurotrophic effect of NE during fetal cortical development (Winzer-Serhan and Leslie, 1999). 2B Receptor mRNA is transiently expressed in the developing vascular plexus during the time of neovascularization in the brain. Additionally, developmentally regulated expression is also detected in the caudate putamen and in the cerebellum, in a pattern which parallels the expression of 2A- and 2C-adrenoceptors in these structures (Winzer-Serhan and Leslie, 1997). Furthermore, the developmental pattern of 2C-adrenoceptor mRNA and protein expression is in marked contrast to the early and transient expression of that of 2A-receptor. There is no widespread expression of 2C-adrenoceptor mRNA or protein in the fetal brain. Expression occurs during the postnatal period, after the major period of neuronal migration and differentiation, and is largely restricted to areas in which there is expression in the adult (Winzer-Serhan et al., 1997a,b). These findings suggest that 2C-adrenoceptors do not play a relevant role in regulating developmental processes. This assumption is supported by the fact that 2C-adrenoceptor-deficient mice do not exhibit any apparent behavioral or morphological defects (Link et al., 1995). Density of -adrenoceptors in whole brain is low at birth (only 14% of the adult value) (Erdtsieck-Ernste et al., 1991), and adult values are reached by 3 weeks of age. A similar pattern of development also occurs in the cortex (Pittman et al., 1980; Lorton et al., 1988). In the adult, the 1-adrenoceptor subtype represents approximately 80% of the total -adrenoceptors present in the cortex. This relative proportion is also present during postnatal ontogeny, except for the perinatal period, when 2 receptors provide a greater percentage to the total (Pittman et al., 1980; Erdtsieck-Ernste et al., 1991).
-Aminobutyric acid (GABA) is one of the earliest substances to appear in the mammalian developing brain (Lauder et al., 1986; Miranda-Contreras et al., 1998). Three types of GABA receptors have been currently identified in the CNS: the GABAA and GABAC, which are both ionotropic, and the metabotropic GABAB receptor (Chebib and Johnston, 1999). Furthermore, six different subunit families have been recognized to constitute CNS GABAA receptors: 1-6, 1-3, 1-3, 
, 
, and 
, whereas three additional subunits, 
1-3, have been distinguished as part of the GABAC receptor (Bormann and Feigenspan, 1995; Luddens et al., 1995).
The early appearance of GABA and its receptors during embryonic brain development, long before the onset of inhibitory synaptogenesis, led to the suggestion that it may play a maturative role before work as neurotransmitter (Lauder, 1993). If GABA serves critical developmental roles, then interference in this early functioning could influence the normal course of brain development. In the rat, GABA has been detectable at ED 12 in axons running through the brainstem (Lauder et al., 1986). In the spinal cord, mRNA encoding both isoforms GAD65 and GAD67 of the synthesizing enzyme glutamate decarboxylase have been detected at ED 11 (Somogyi et al., 1995). At ED 13 GABA-immunoreactive fibers have been found to project from spinal cord and brainstem toward midbrain and diencephalons (Lauder et al., 1986), whereas at ED14 GABAergic cells have been identified in the lateral cortex and by ED 16 in the basal forebrain and all regions of the primitive cortex. Such cortical cells are located in the marginal and intermediate zones as well as in the subplate (Van Eden et al., 1989). Additional cells are also located in the ventricular and subventricular zones of the neocortex at ED 16 to 17. GABA neurons remain in the subplate during most part of pregnancy and spread over cortical plate by ED 18 (Cobas et al., 1991). During corticogenesis GABA neurons are located in the marginal zone and subplate to contact migrating neurons and affect cortical afferent development (Meinecke and Rakic, 1992; Lauder, 1993). By birth in rats, a well expanded axonal plexus is recognizable around maturating hippocampal cells (Lubbers and Frotscher, 1988). Similarly, in the postnatal cerebellum GAD-immunoreactive fibers encompass differentiating granule cells (Lauder, 1993), suggesting that GABA may play a role in granule cell maturation both in cerebellum and hippocampus.
Functional GABAA receptors are detectable on mitotically active precursor elements in the neocortical proliferative zone (LoTurco et al., 1991; Owens et al., 1999). They display a major affinity for GABA and appear rather insensitive to receptor desensitization processes (Owens et al., 1999). Such functional dissimilarities in GABAA receptor functioning in precursor elements and postmitotic neurons might derive from the discrepancy in subunit composition (Araki et al., 1992; Poulter et al., 1993). Indeed, in the embryonic cortical plate, where postmitotic neurons are predominantly located, 3/2-3 and 3 subunits have been found to predominate (Ma and Barker, 1995).
Whereas in the adult brain GABAA receptor activation has been related to the mediation of synaptic inhibition, in immature neurons this causes marked membrane depolarization that can induce action potential discharge (Ben-Ari et al., 1989; Owens et al., 1996, 1999; Dammerman et al., 2000; Gao and van den Pol, 2001). The relatively higher intracellular chloride concentration is considered responsible for the observed response (Ben-Ari et al., 1989; Chen et al., 1996; Rohrbough and Spitzer, 1996). With further development, chloride concentration declines so that the effect of GABA becomes increasingly inhibitory (Owens et al., 1996). During the period of active neurogenesis and until about the first postnatal week, the activation of GABAA receptors has been shown to induce membrane depolarization and a rise in cytosolic Ca2+ (Cherubini et al., 1991; LoTurco et al., 1995; Owens et al., 1996, 1999). The activation of voltage-dependent Ca2+ channels occurring during depolarization has been thought to contribute to the elevation in intracellular Ca2+ (Leinekugel et al., 1995; Ben-Ari, 2002). These findings have suggested that one potential consequence of GABAA receptor in maturing neurons is the activation of Ca2+-dependent second messenger pathways (Cherubini et al., 1991), which in turn can affect a variety of processes, including proliferation, synaptogenesis, and circuit modeling.
With regard to GABAB receptors, immunohistochemical studies have shown that both R1 and R2 subunits are present in the embryonic cortex, and GABAB receptor activation can influence the movement of immature cortical neurons. However, functional GABAB receptor-mediated postsynaptic responses have been reported that do not occur in the neocortex until after the second postnatal week (Luhmann and Prince, 1991), although it has been observed that presynaptic receptor activation occurs by the first week (Fukuda et al., 1993).
Experimental evidence has shown that GABA triggers signals in proliferating cells located in the telencephalic ventricular zone, functioning as a modulator of cell proliferation (LoTurco et al., 1995; Owens et al., 1999; Haydar et al., 2000). Investigating [3H]thymidine or bromodeoxyuridine incorporation in cells derived from the ED 16 to 19 cortex has demonstrated that GABA can influence DNA synthesis in proliferating cells (LoTurco et al., 1995). Moreover, GABA has been reported to prevent exit from the cell cycle and to reduce cell cycle duration of cells from the ventricular zone of embryonic cortex. Studies on the migratory responses have disclosed that GABA may stimulate directed migration (chemotaxis) of cells derived from ED 18 ventricular and subventricular zone, whereas facilitates chemokinesis (random motility) of more mature neurones derived from the cortical plate-subplate regions (Behar et al., 1996, 2000, 2001). Interestingly it has also been documented that GABAC and GABAB receptor activation in rats is able to promote migration out of the ventriculare zone and the intermediate zone, respectively, whereas GABAA receptor activation could provide a "stop signal," once cells have reached the cortical plate (Behar et al., 2000). The relevance of these results requires further evaluation, since factors others than GABA have been found implicated in the arrest of cell migration at the cortical plate (Dulabon et al., 2000; Supèr et al., 2000). Selectively antagonizing GABAC and GABAB receptor activation has resulted only in a delay but not in a complete arrest of migration, suggesting that, although GABA-mediated signaling could promote neuronal migration, it is not absolutely crucial for this process, since its absence may be physiologically compensated (Behar et al., 2000). The GABA-mediated migratory signals have been reported to act through Ca2+ transients that affect cell movements by altering the dynamics of cytoskeletal remodeling (Gomez and Spitzer, 1999). On the other hand, similar investigations in immature neurons from embryonic mouse brain have indicated that N- methyl-D-aspartate-type glutamate receptor, rather than GABA receptor, activation seems to affect migration, suggesting that a discrepancy exists between these two rodent species regarding the nature of signals moderating neuronal migration in immature brain (Varju et al., 2001; Owens and Kriegstein, 2002).
Current data seem to validate the hypothesis that in embryonic brain GABA acts by accelerating neuronal maturation and promoting formation of functional synapses (Varju et al., 2001). The transformation of a growth cone to a synaptic element implicates the maturation of the biochemical machinery of neurotransmission; this transition may be affected in part by changes in subunit composition of GABAA receptors (Maric et al., 1997; Owens et al., 1999) and probably involves switches in the expression of components implicated in GABA synthesis, storage, and release (Somogyi et al., 1995). GABA has been reported to enhance the density of intracellular organelle in rat cerebellar granule cells (Hansen et al., 1987), including the Golgi apparatus, rough endoplasmic reticulum, microtubules, and coated vesicles, and may stimulate metabolic activity of neurons. It has been also described that GABA up-regulates the expression of specific GABAA receptor subunits (1 and 2), and promotes the synthesis of a number of neuron-specific proteins, including neuron-specific enolase and neural cell adhesion molecules (Belhage et al., 1998). In cultured embryonic hippocampal and neocortical neurons, GABAA receptor activation has been shown to stimulate neurite outgrowth and maturation of GABA interneurons (Barbin et al., 1993; Marty et al., 1996; Maric et al., 2001).
III. Antipsychotics and Antiepileptics
The annual incidence of psychosis in pregnant women has been reported to be 7.1 cases per 100,000 (Nurnberg, 1989), and epidemiological studies indicate that psychotic women neither recover nor require decreased doses of maintenance drugs during pregnancy (Trixler and Tényi, 1997). In addition, reducing or discontinuing medications in psychotic pregnant women responsive to treatment may result in a raised individual risk of relapse (Casiano and Hawkins, 1987). Thus, emotional and somatic changes occurring during gestation or puerperium do not protect from the development or the recurrence of psychosis. Although physicians often become hesitant when recommending antipsychotic drugs during pregnancy because of their potential fetotoxicity, avoiding medications is not usually possible. Indeed, withholding treatment for such mentally ill patients carries potentially serious consequences, as untreated psychosis may adversely influence the course of pregnancy (Kris, 1961). In addition, a recent study suggests that emotional stress during organogenesis can cause congenital malformations, particularly those of cranial neural crest (Hansen et al., 2000). Thus, in all cases, clinicians must carefully weigh the risk of fetal exposure to antipsychotic medications against the potential adverse effects to both mother and fetus of untreated mental illness. The two major groups of antipsychotic drugs are the typical neuroleptics, such as phenothiazines, thioxanthenes, and butyrophenones; and the atypical neuroleptics, such as clozapine, risperidone, olanzapine, ziprasidone, and quetiapine.
1. Typical Antipsychotics.
Typical neuroleptic agents act primarily by blocking brain dopamine receptors, and radioligand binding assays indicate that antipsychotic potency is highly correlated with affinity for D2 dopamine receptors. These findings have been further strengthened by positron emission tomography data, which show that the effectiveness of typical neuroleptics is associated with an occupancy of 80% of D2 dopamine receptors, whereas higher occupancy rates may be associated with more adverse effects, without greater effectiveness (Baldessarini and Tarazi, 2001). Although many standard neuroleptics, in particular thioxanthenes and phenothiazines, bind with relatively high affinity to other subtypes of dopamine receptors, as well as to other receptors, it appears that the antipsychotic effects of classic neuroleptics require D2 receptor blockade, followed by a decreased dopaminergic activity.
About one third of the agents commonly used for the treatment of psychosis may exert teratogenic effects in laboratory animals. Several compounds can cause cleft palate in mice, without overt teratogenic activity in other species. These comprise, among others, fluphenazine, haloperidol, trifluperidol, and thioridazine (Vichi et al., 1968; Vichi, 1969; Szabo and Brent, 1974). Studies addressing phenothiazine teratogenicity in rats have yielded conflicting results (Jelinek et al., 1967; Clark et al., 1970; Beall, 1972; Singh and Padmanabam, 1978), whereas haloperidol causes increased incidence of fetal resorption, delayed delivery, and neonatal death at doses 2- to 10-fold higher than the maximum doses used in humans (Dollery, 1999).
Because of their small molecular size and relative lipophilicity, neuroleptics are assumed to readily cross the placenta and to enter fetal circulation (Pacifici and Nottoli, 1995). Data on fetal outcome for psychotic women treated with neuroleptics during gestation are limited, so the potential risks of antipsychotic exposure during pregnancy are still not fully known. Moreover, accumulating evidence suggests that children born to psychotic mothers may exhibit increased risks of abnormalities not related to neuroleptic exposure. Indeed, previous studies, comparing pregnant psychotic women with or without exposure to phenothiazines during gestation, reported the rate of fetal damage to be similar in both groups, but approximately twice that observed in the general population, suggesting that maternal psychiatric disease may constitute itself a risk factor for fetal anomalies (Sobel, 1960; Rieder et al., 1975). The mechanism underlying the increased risk related to the mental illness remains not comprehended, but investigators have pointed out that psychotic women often smoke, misuse other substances, are socioeconomically disadvantaged, and have poor compliance with prenatal care (Bennedsen, 1998).
Results on reproductive effects associated with gestational use of neuroleptics are conflicting, since many findings derive from retrospective and prospective studies, whose accuracy has been often questioned for inadequate attention to potential confounding variables, such as diagnosis and severity of the illness, dosage, maternal age, and exposure to other medications, as well as to alcohol and illicit drugs. Indeed, in most studies the majority of patients received phenothiazines for the treatment of insomnia, vomiting, and anxiety, and not for the therapy of psychiatric disorders, so that medications were probably not taken at the schedule and dosage usually administered to psychotic subjects. In one study in which the risk of abnormalities after in utero exposure was similar to that of unexposed children (Milkovich and Van der Berg, 1976), a re-examination of findings, using a longer follow-up time, demonstrated a trend toward an increased rate of malformations in infants exposed in utero to phenothiazines in weeks 4 to 10 of pregnancy (Edlund and Craig, 1984). Another study revealed an association between gestational phenothiazine exposure (including exposure during the first trimester) and an enhanced rate of birth defects (Rumeau-Rouquette et al., 1977). More specifically, this study, correlating the different outcomes with the chemical structure of phenothiazines used during the pregnancy, has provided evidence that phenothiazines with three carbon aliphatic side chains (e.g., chlorpromazine) were associated with a higher rate of malformations, whereas those with two carbon side chains were not. A large prospective study that analyzed data published between 1963 and 1995 on the effects of prenatal exposure to neuroleptics reported that gestational use of low-potency neuroleptics may confer a significant, albeit small, increase in the likelihood of poor outcome (Altshuler et al., 1996). Unfortunately, the nature of this meta-analysis excluded the possibility to assess differential risks associated with individual phenothiazines (Pinkofsky et al., 1997). Few studies have investigated fetal outcome for pregnant women treated with high-potency neuroleptics, and most data refer to the effect of prenatal haloperidol. Although two early case reports describing limb malformations raised concerns regarding first-trimester exposure to haloperidol (McCullar and Heggeness, 1975; Dieulungard et al., 1996), several studies failed to demonstrate an increased teratogenic risk with this drug (Van Waes and Van de Velde, 1969; Hanson and Oakley, 1975). Nevertheless, similar to low-potency antipsychotics, in the large majority of these studies women were given haloperidol in association with other medications, thus not allowing definite conclusions.
In recent years, increasing attention has been given to the more subtle, nonstructural alterations produced by drugs given prenatally. Such changes involving motor ability, emotionality, and learning and memory capability constitute a sensitive tool for detecting subtle damage to the functioning of the central nervous system induced by exposure to medications at sensitive phases and at dose levels frequently below those commonly associated with manifest signs of neurotoxicity (Cuomo, 1987). These alterations in behavior seem to be partly due to drug-induced changes in the developmental pattern of specific neurotransmitter systems.
Most studies examining the influence of prenatal antipsychotic exposure have focused on various aspects of the dopaminergic system. In the majority of these investigations, haloperidol has been used as prototype of this class of drugs. In utero haloperidol exposure has been found to decrease cell proliferation in the forebrain (Blackhouse et al., 1982; Patel and Lewis, 1988) and to affect the expression of DNA polymerase in the mesencephalon and forebrain (Castro et al., 1990). Additionally, gestational haloperidol exposure induces a reduction of nerve growth factor receptors and mRNA in neonate rat forebrain (Alberch et al., 1991), suggesting that prenatal haloperidol exposure may have a critical impact on forebrain development. This assumption has been confirmed by electrophysiological investigations demonstrating that in 2-week-old pups prenatal haloperidol caused a significant decline in the number of spontaneously active midbrain dopamine neurons. However, whether the decrease in the number of active cells may result from a physical loss rather than from a functional change in the threshold of spontaneous activity is still uncertain (Zhang et al., 1996). Prenatal dopamine receptor occupancy was demonstrated to be a critical factor in controlling the development of forebrain target cells through selective changes in the expression of plasticity-related genes, whereas the expression of other genes, including several proto-oncogenes, was unaffected (Castro et al., 1994). Moreover, the widespread distribution of c-fos gene expression in the fetal rat brain following dopamine D1 receptor stimulation is in contrast to the response of c-fos occurring in adult rats (Shearman et al., 1997), when D1 receptor activation induces c-fos gene expression after depletion of dopamine. Denervating lesions of dopaminergic projections reduced D1 postsynaptic receptor expression in the immature nervous system and up-regulated D1 receptors in mature rodents (LaHoste and Marshall, 1994). This indicates that removal of dopaminergic innervation or receptor blockade in immature brain results in a paradoxical change, rather than in compensatory overexpression of dopamine receptors, as observed in adult rats.
Prenatal haloperidol treatment did not alter levels of dopamine and its metabolites in the basal ganglia of 1- to 58-day-old rats (Rosengarten et al., 1983; Williams et al., 1992). However, gestational exposure to haloperidol has been shown to reduce the number of postsynaptic dopamine receptors (Rosengarten and Friedhoff, 1979; Miller and Friedhoff, 1986; Scalzo et al., 1989), although other reports did not confirm these findings (Madsen et al., 1981; Moon, 1984; Schmidt and Lee, 1991). Moreover, depending on the timing of exposure, developmental treatment with haloperidol can affect the response of rat offspring to pharmacological challenges to the dopaminergic system (Rosengarten and Friedhoff, 1979; Spear et al., 1980; Cuomo et al., 1985; Scalzo and Spear, 1985). In particular, prolonged prenatal exposure of rats to haloperidol has been found to significantly influence their behavioral responsiveness to a dopamine receptor agonist such as apomorphine at 60 days of age. The intensity of stereotyped behaviors as well as the effects on locomotor activity elicited by apomorphine in haloperidol-pretreated animals have been shown to be markedly attenuated when compared with controls (Rosengarten and Friedhoff, 1979; Cuomo et al., 1985). These data, indicating a behavioral subsensitivity of the dopaminergic system of haloperidol-exposed rats to pharmacological stimulation, parallel neurochemical results showing a decrease in [3H]spiroperidol binding in the striatum of rats born to mothers treated with haloperidol during gestation (Rosengarten and Friedhoff, 1979). On the other hand, the prolonged administration of haloperidol during the first 3 weeks of postnatal life has been reported to produce, in 60-day-old rats, an opposite response pattern (behavioral supersensitivity to apomorphine) which again correlates with neurochemical data (increased [3H]spiroperidol binding in the striatum) (Rosengarten and Friedhoff, 1979). Furthermore, a challenge dose of haloperidol induces smaller increases in dopamine turnover in adult rats treated with this neuroleptic during early postnatal life (Cuomo et al., 1981). Although increased central dopamine receptor sensitivity to apomorphine as well as an attenuated response to a challenge of haloperidol on dopamine turnover after prolonged haloperidol treatment also occur in adult rats, there is no evidence that these changes persist up to 40 days after the last administration of this neuroleptic agent (Cuomo et al., 1983b). These findings suggest that the particular period of developmental administration of haloperidol plays a critical role in causing enduring neurofunctional changes (Cuomo et al., 1983b) and that compensatory mechanisms occurring in response to a prolonged treatment during development are markedly different from those occurring during adulthood.
Additional studies have shown that prolonged postnatal exposure to haloperidol alters the ultrasonic emission elicited by the removal of rat pups from their nest (Cagiano et al., 1986). This response is a reliable indicator of emotional reactivity during development. In particular, neonatal administration of this neuroleptic agent produced a significant decrease in the rate of calling, an increase in the duration of calls, and a decrease in the minimum and maximum frequency of calls. There is evidence that dopamine plays an important role in the regulation of sexual behavior in rats (Gessa and Tagliamonte, 1975), and dopaminergic mechanisms are thought to underlie sexual dysfunctions produced by the administration of some neuroleptics (Buffum, 1982; Segraves, 1982). In this regard, Hull et al. (1984) have shown that the prolonged administration of haloperidol to pregnant rats, at a relatively high dose (2.5 mg/kg), impairs the sexual behavior of male offspring. Indeed, haloperidol-exposed animals had significantly fewer ejaculations than controls. Since ultrasonic calls emitted by male rats during mating activity seem to be a sensitive indicator of their sexual motivation (Barfield et al., 1979), the aim of other studies was to investigate the influence of prenatal or early postnatal exposure to a low dose of haloperidol (0.5 mg/kg), which itself does not affect sexual behavior in the offspring, on both preejaculatory and postejaculatory ultrasonic vocalizations. The results showed that the latency of emission of the first precopulatory 50-kHz call was not influenced by the early postnatal haloperidol treatment. Conversely, the period of the 22-kHz call emission was shorter in haloperidol-treated animals than in controls (Cuomo et al., 1991). The comparison of the results of this study with those of previous experiments (Cagiano et al., 1988) further confirms that the behavioral consequences of developmental treatments with this dopamine antagonist are critically dependent upon the period of administration. In fact, the latency of emission of the first precopulatory 50-kHz ultrasound as well as the duration of the period of the 22-kHz postejaculatory call emission were significantly increased by prenatal exposure to haloperidol (Cagiano et al., 1988; Cuomo et al., 1990). Since it has been shown that D2 dopamine receptors are involved in the emission of 22-kHz postejaculatory calls (Cagiano et al., 1989), the selective alterations in this ultrasonic parameter produced by prenatal and neonatal treatment with haloperidol may be due to interactions with the development of this receptor subtype. Finally, the finding that developmental administration of haloperidol to two inbred strains of mice (C57 BL/6J and DBA/2J), who display an opposite behavioral reactivity to stimulation of dopamine receptors (Sansone et al., 1981), caused distinct behavioral changes in their offspring (Cuomo et al., 1984) and also suggests that pharmacogenetic determinants play a role in the behavioral consequences of developmental exposure to this neuroleptic.
In contrast to the abundance of animal studies, human studies focusing on the potential neurobehavioral sequelae of prenatal exposure to typical antipsychotics are very limited. In the absence of a continuous follow-up, it is not possible to evaluate the neurobehavioral effects in children born to mothers treated with antipsychotics while pregnant. However, after a follow-up to the age of 5 years, two studies by Edlund and Craig (1984) and Kris (1965) did not report significant differences in behavioral and intellectual functioning in children with and without histories of prenatal exposure to typical neuroleptics. Moreover, in a case-control study, Stika et al. (1990) failed to document any difference between infants receiving and not receiving medications when evaluating school behavior and proficiency of 68 children gestationally exposed to typical neuroleptics. Thus, the limited data in humans sharply diverge from those provided by animal studies; in animals gestational exposure to antipsychotics has
