Administration of cocaine to pregnant rabbits produces robust and long-lasting anatomical, neurochemical and behavioral alterations in their offspring. For example, exposure to cocaine following implantation [embryonic day (E) 8] through gestation (E29) produces increased length and decreased bundling of layer III and V pyramidal neuron dendrites, increased parvalbumin expression in the dendrites of interneurons, and increased GABA levels in the anterior cingulate cortex (ACC) and other dopamine-rich cortical areas. We have investigated the presence of a sensitive period of in utero exposure during specific developmental epochs prior to and during the onset of cortical development that might be capable of producing such effects. Cocaine (3 mg/kg i.v., twice daily) or saline was administered during embryonic days E16–E25 (onset and peak of corticogenesis), E8–E15 (prior to cortical plate formation), E8–E25 or E8–E29 as in earlier studies. Examination of the ACC in offspring exposed from E8 to E25 and from E16 to E25 were found to induce alterations in the development of pyramidal neurons and interneurons that are nearly identical to those induced by the complete exposure (E8–E29) paradigm. No alterations were observed following the E8–E15 cocaine exposure. These data suggest that exposure to cocaine during E16–E25, the time of peak corticogenesis, appearance of cortical monoamines and onset of D1 dopamine receptor expression, is necessary and sufficient to produce long-term effects on the organization of excitatory pyramidal neurons and inhibitory interneurons in the ACC.
Dysfunctions in neurotransmitter signaling contributes to the pathophysiology of numerous neurological disorders, which may in turn have origins in defective morphogenetic processes during brain development (Bloom, 1993; Weinberger, 1995; Levitt et al., 1997; Levitt, 1998; Stanwood and Levitt, 2001). Alterations of neurotransmitter balances in the fetus can affect critical phases of brain histogenesis (Levitt et al., 1997), and monoamines, in particular, act as morphogens in the developing mammalian CNS (Molliver, 1982; Mattson, 1988; Lauder, 1993; Whitaker-Azmitia et al., 1996). Perhaps the most striking demonstration of the important role for monoamines in pattern formation has been the characterization of mice lacking monoamine oxidase A (Cases et al., 1995, 1996). These mice fail to develop the normal barrel structures that are related to vibrissa representation in the somatosensory cortex, due to increases in serotonin levels during the sensitive period for barrel formation.
Studies of specific, drug-induced perturbations of neurotransmitter systems have also provided compelling demonstrations of the important roles for monoamines in guiding aspects of cortical development. For example, administration of low doses of cocaine (2–4 mg/kg i.v., twice daily) to pregnant rabbits during embryonic days (E) 8–29 produces specific, dosedependent and long-lasting effects on the structure and function of cortical neurons receiving dopaminergic innervation in the offspring without altering general developmental parameters [for reviews see Levitt et al. and Stanwood and Levitt (Levitt et al., 1997; Stanwood and Levitt, 2001)]. These alterations include a 30–50% lengthening of the apical dendrites of pyramidal neurons (Jones et al., 1996, 2000), an increased number of gamma-aminobutyric acid (GABA)-immunoreactive cells (Wang et al., 1995a) and increased parvalbumin-immunoreactivity in the distal dendrites of GABAergic interneurons (Wang et al., 1996a) in the anterior cingulate cortex (ACC). These changes appear to be regionally specific, in that no alterations are present in dopamine (DA)-poor cortical regions such as the visual cortex (VC). These anatomical abnormalities are accompanied by a sustained reduction in coupling of the D1 DA receptor to its G protein (Friedman et al., 1996; Jones et al., 2000) in the ACC. Functional outcomes of prenatal cocaine exposure in this model include anomalous behavior on motor and discriminative tasks (Romano et al., 1995; Romano and Harvey, 1996; Simansky and Kachelries, 1996), modifications in neuronal responsiveness in the ACC (Gabriel and Taylor, 1998) and hippocampus (Little and Teyler, 1996, 1998) and altered regulation of DA release (Wang et al., 1995b; Du et al., 1999). Other models of prenatal cocaine exposure also have revealed effects on brain structure and function (Gressens et al., 1992; Lidow, 1995, 1998; Kosofsky and Wilkins, 1998; Ronnekleiv et al., 1998; Spear et al., 1998; White et al., 1999).
The present study focuses on defining whether there is a well-delineated sensitive period during which these long-term effects on neurodevelopment can be induced. Our data indicate that cocaine administration (3 mg/kg i.v., twice daily) from E16 to E25, corresponding to the time of peak corticogenesis and appearance of cortical monoamines in the rabbit, is necessary and sufficient to produce robust and long-lasting changes in the development of the ACC.
Materials and Methods
Proven breeder Dutch-belted rabbits from Myrtle Rabbitry (Thompson Station, TN) were housed individually in a 12 h light:dark cycle with free access to food and water. The day of breeding was designated as E0 and injections of cocaine hydrochloride (3 mg/kg, supplied by NIDA) or saline were given i.v. (through the marginal ear vein) to pregnant dams twice daily (09:00, 16:00 h) on E8–E29, E8–E25, E16–E25 or E8–E15. A total of 132 rabbits were produced for these studies in which 4–5 animals from 2–3 dams were included in each treatment group (per age, per fixation type). Cocaine and saline were administered in an injection volume of 2 ml/kg over a period of at least 30 s. Behavioral ratings were made for ~5 min after each injection as described by Murphy et al. (Murphy et al., 1995). Behavioral and physiological responses to cocaine varied somewhat across the dams, but generally included increased respiration, pupillary dilation and motor responses ranging from mild twitching to rare seizures. No weight changes were observed in pregnant dams as a result of drug treatment, and gross parameters such as kit weight and litter size were normal as previously reported (Murphy et al., 1995, 1997). Rabbits were born on E30–E31 and the day of birth was designated postnatal day 0 (P0). Kits were raised by their biological mother until killed at age P10–P12 or P20–P24 (subsequently referred to as P10 and P20, respectively). Cocaine- and saline-exposed offspring were deeply anesthetized and transcardially perfused with a saline rinse followed by 4% paraformaldehyde (pH 7.2) for anti-MAP2 and anti-parvalbumin staining or with 2% paraformaldehyde, 2% glutaraldehyde, 0.2% picric acid (pH 7.4) for anti-GABA staining.
Coronal sections were cut on a microtome or Vibratome at 50 μm and stained using previously published protocols (Wang et al., 1995a, 1996a; Jones et al., 1996). Briefly, sections were treated with Tris–glycine (pH 7.2) to reduce nonspecific labeling, blocked in 4% blotto and 0.2% Triton-X 100, and incubated overnight with monoclonal antibodies against MAP2 (microtubule-associated protein 2, 1:2500; graciously provided by Dr I. Fischer, MCP Hahnemann School of Medicine) or parvalbumin (Sigma, 1:500). Following several washes, sections were incubated in biotinylated anti-mouse IgG (Jackson, 1:1000) for 90 min. Standard avidin–biotin amplification (ABC, Vectastain) and DAB reactions were used to visualize labeled proteins. All incubations occurred at room temperature. In other experiments, sections from glutaraldehyde-perfused animals were washed in 1% sodium borohydrate, blocked in 1–2% normal goat serum and incubated overnight with a rat anti-GABA antibody (Protos, 1:1500). Following several washes, sections were incubated in biotinylated anti-rat IgG (Jackson, 1:1000) for 90 min and further reacted as above. Sections from cocaine- and saline-treated animals were always processed in parallel to minimize variability in the quality of immunostaining between groups. In order to assess the ontogenic appearance of monoaminergic projections in naive rabbits, sections were cut on a cryostat (20 μm), treated with Tris–glycine (pH 7.2) to reduce nonspecific labeling, blocked in 4% blotto and 0.2% Triton-X 100, and incubated overnight with a monoclonal antibody against tyrosine hydroxylase (TH; Sigma, 1:8000). Following several washes, sections were incubated in biotinylated anti-mouse IgG (Jackson, 1:1000) for 60 min. Enhanced avidin–biotin amplification (ABC Elite, Vectastain) and 3,3′-diaminobenzidine reactions were then used to visualize labeled TH. Negative controls in which primary antibodies were omitted revealed no specific labeling (data not shown).
In Situ Hybridization
Probe was derived from nucleotides 51–538 of D1 receptor mRNA, resulting in a 487 bp product. Double-stranded cDNA containing this sequence was first amplified from embryonic rabbit brain cDNA using custom designed primers (forward primer: CCAAGGTGACCAACTTCTTT; reverse primer: GTGGATCCTGGTGTAGGTGA) and ‘touchdown’ polymerase chain reaction (PCR) with AmpliTaq Gold (PE Biosystems): 94°C for 10 min, followed by 10 cycles at a high annealing temperature (94°C for 30 s, 60°C for 30 s, 72°C for 60 s), 10 cycles at a medium annealing temperature (94°C for 30 s, 58°C for 30 s, 72°C for 60 s) and 20 cycles at a low annealing temperature (94°C for 30 s, 56°C for 30 s, 72°C for 60 s). The product of this PCR reaction produced a single bright band on a 2% agarose gel, was ligated into a T/A plasmid cloning vector (AdvanTAge, Clontech), then transformed into competent Escherichia coli cells and plated overnight at 37°C. Colony PCR was performed on selected colonies containing the insert and the products of these reactions were restriction digested and sequenced to verify orientation and insert identity. 35S-labeled riboprobes were synthesized using the T7 Riboprobe In vitro Transcription System (Promega kit #P1460) and purified using an RNeasy kit (Qiagen #74104). A scintillation counter was used to verify the specific radioactivity and yield of the probe. During hybridization, ~8 ng of probe was used per slide in a total volume of 90 μl. All other methods were identical to those employed previously (Campbell et al., 1999). Following hybridization and washing, slides were air-dried and exposed to BioMax MR film (Kodak) for 2–3 days, then dipped in emulsion (NTB-2, Kodak) and exposed for 12 days at 4°C. Film images were scanned at high resolution and darkfield images were captured from developed slides.
In Vivo Dendritic Analysis Using 1,1′-Dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine
Minor modifications of the method described by Jones et al. (Jones et al., 2000) were used. P0–P1 rabbits were anesthetized and perfused transcardially with 4% paraformaldehyde (pH 7.2). Brains were removed from the skull, post-fixed for several days at 4°C and then blocked to reveal the ACC. Labeling of projection neurons was achieved by a 4 ms, 20 psi pulse (General Value Corporation Picrospritzer® II) of 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine (DiI)-saturated ethanol solution into the cingulum. The tissue slabs were returned back to fixative, and the containers were wrapped in aluminum foil and incubated at 37°C for 4 days. The slabs were then embedded in 18% gelatin, sectioned in the coronal plane at 100 μm using a Vibratome (Electron Microscopy Sciences OTS-3000) and collected in phosphate-buffered saline (PBS). The sections were mounted in glycerol–PBS and viewed on an Olympus AX70 confocal microscope system. Olympus Fluoview software was used to scan and view the resulting images.
Slides were coded so that the investigators were blinded to the treatment group. Sections were imaged using a Nikon G800 microscope and analyzed as previously described (Wang et al., 1995a, 1996a; Jones et al., 1996). Following GABA and parvalbumin staining, the number of immunostained cells in the ACC and VC were counted in two fields (0.15 mm2) per hemisphere per section. Four sections were analyzed from each animal in each region and 4–5 animals were included in each treatment group at each age. The dendritic distribution of parvalbumin-expressing neurons (n = 10) was measured in each field from which cell counts were obtained (60 neurons per region per animal). Quantification of parvalbumin-positive profiles was limited to layers II–III as the higher density of parvalbumin-stained cells in deeper layers made it more difficult to resolve individual processes. Significance of cell counts was determined by Mann–Whitney U-tests.
MAP2-immunostained Apical Dendrites
The ACC of P10 (Fig. 1) and P20 (Fig. 2) rabbits were examined for changes in dendritic development using MAP2 as a specific marker. MAP2-labeled dendrites in normal untreated and saline-exposed animals have very characteristic appearances. The shafts of the apical dendrites of layer III and V pyramidal neurons are straight as they exit the cell body and traverse the cortex to the pial surface in long, well-delineated bundles (Figs 1A and 2A). As we have previously reported (Jones et al., 1996), following E8–E29 cocaine exposure, MAP2-positive ACC apical dendrites appear in short, discontinuous segments because they course tortuously into the depth of the section (Fig. 1B). Similar alterations in MAP2-stained material were apparent in P10 and P20 rabbits that received cocaine from E8 to E25 and from E16 to E25, but following the E8–E15 paradigm the apical dendrites appeared normal (Figs 1D,F,H and 2). No changes were observed in the VC following any treatment regimen (data not shown).
The morphology of apical dendrites in the ACC following E16–E25 cocaine was also examined by direct labeling with DiI. The dendrites of layer II, III and V–VI neurons were visible in both saline- and cocaine-exposed neonatal rabbits. The dendrites of control rabbits were very straight and thick, as described by Stensaas (Stensaas, 1967) using Golgi impregnation and in our previous studies using DiI (Jones et al., 1996, 2000). The dendrites typically extended to the marginal zone and could be traced without altering microscope focus, indicative of outgrowth in a straight path (Fig. 3A–C). Most labeled dendrites of ACC-cocaine neurons also reached the marginal zone, but could only be traced with changes in the focal plane (Fig. 3D–I). The trajectories of E16–E25 cocaine-exposed dendrites were very similar in nature to those we have reported previously in the E8–E29 treatment paradigm.
At P10, parvalbumin-immunoreactive (IR) interneurons in the ACC are apparent in a very dense band in layers V and VI, and scattered throughout layers II and III. At P20, a greater number of parvalbumin-IR neurons are present, particularly in layers II and III. It has been reported previously that E8–E29 cocaine increases the dendritic expression of parvalbumin without altering overall parvalbumin-IR cell numbers (Wang et al., 1996a; Murphy et al., 1997). In this regard, parvalbumin immunoreactivity normally is restricted to the cell body and primary dendrites of neurons in control animals. Only the occasional secondary or tertiary dendritic profile is labeled (Fig. 4A). In contrast, the ACC interneurons of rabbits receiving E8–E29 cocaine express parvalbumin in many secondary and tertiary branches of dendrites (Fig. 4B). Qualitatively similar increases in parvalbumin-IR dendrites were observed following the E8–E25 and E16–E25 treatment regimens (Fig. 4D,F), but the E8–E15 exposure to cocaine did not produce overt changes (Fig. 4H). Quantification of the number of dendritic branches expressing parvalbumin in layers II–III of the ACC confirmed that significant increases in the number of parvalbumin-IR secondary (+55 to 100%) and tertiary (+90 to 200%) dendrites were present in animals receiving E8–E29, E8–E25 or E16–E25 cocaine (Figs 5 and 6). Quantification of dendritic expression of parvalbumin was not performed in the deeper layers, where parvalbumin-positive cells are expressed in much higher densities. A similar change was, however, clearly observable in layers V–VI. No treatment regimen produced alterations in parvalbumin expression in the VC or in parvalbumin-positive cell number in either the ACC or VC (data not shown).
GABA-IR serves as an additional component of altered ACC development as a result of in utero cocaine exposure. E8–E29 cocaine produces a long-lasting increase in GABA-IR in the ACC (Wang et al., 1995a). In P10 and P20 rabbits, GABA-IR neurons were distributed homogeneously throughout all cortical layers. The GABA-IR cell number per unit area was similar across the two ages and in each of the saline control groups. As noted for the other measures, we observed significant and equivalent increases in detectable GABA-IR neurons following E8–E29, E8–E25 and E16–E25 cocaine (+45 to 55% at P10; +30 to 35% at P20), but no changes following the E8–E15 treatment regimen (Figs 7 and 8). Also consistent with our analyses of apical dendrite and parvalbumin-IR neuron development, no changes in GABA expression were observed in the VC (data not shown).
Timing of Appearance of Dopaminergic Circuitry Components in the ACC
We next examined when two components of dopaminergic circuitry, specifically TH-IR fibers and D1 receptors, are expressed in the embryonic rabbit ACC. A moderate-to-high density of TH-positive fibers was observed at E22 in both the ACC and striatum (Fig. 9B,C). Similarly, intrinsic cells of the ACC and striatum express D1 receptor mRNA at this stage of development (Fig. 10B,C). Interestingly, D1 receptor mRNA is expressed at higher levels in the ventrolaterally located, presumbably postmitotic, striatal cells than in the proliferative zone of the ganglionic eminence (Fig. 10B) at E22, whereas TH immunoreactivity is equally dense in these regions (Fig. 9B). Neither TH protein nor D1 receptor mRNA were observable in any forebrain region at E15 (Figs 9A and 10A). The absence of dopaminergic afferents and postsynaptic receptors through E15 is consistent with the lack of alterations of the E8–E15 cocaine exposure on the development of ACC neurons.
Effects of Cocaine on Cortical Development
The organization of MAP2-labeled apical dendrites, and parvalbumin and GABA expression in the ACC were consistently and robustly altered in offspring of animals exposed to cocaine for long (E8–E29), intermediate (E8–E25) or restricted (E16–E25) durations. These data indicate that in utero exposure to cocaine from E16 to E25, the period of peak neuronal differentiation in the rabbit cortex, is sufficient to produce unequivocal and long-lasting changes in the organization of the ACC. It is particularly striking that the relatively short treatment regimen from E16 to E25 produced alterations that did not reverse postnatally after prolonged removal of exposure to the drug. We have not yet tested whether the changes induced by E16–E25 exposure persist into adulthood, but following E8–E29 cocaine, these changes persist for the life of the animal and in some cases become even more dramatic as the rabbits age (Wang et al., 1995a, 1996a; Jones et al., 1996). Cocaine administration from E8 to E15, on the other hand, did not produce alterations in the development of dendrites or interneurons. These findings suggest that cocaine exposure after E15 is required to induce deficits in cortical development in this animal model. The data from these different developmental periods indicate that it is unlikely that duration of exposure to cocaine is a critical determinant itself, because the E8–E15 and E16–E25 periods are similar in length and the restricted E16–E25 period produced effects of the same magnitude as the full E8–E29 or intermediate E8–E25 paradigms.
Previous morphometric analyses have shown no effects of prenatal cocaine exposure on cortical thickness, neuron number, neuron size or neuronal density in this model (Wang et al., 1995a; Jones et al., 1996). There also are no observable changes in the development of axons or astroglia in the ACC (Jones et al., 1996; Wang et al., 1996b). Thus, changes in the number of parvalbumin-IR dendrites cannot be attributed to changes in neuronal number. The increase in GABA-IR cells also is probably due to increased GABA levels within existing cells, rather than an increase in the total number of interneurons. This interpretation is supported by the lack of change in parvalbumin-IR cell number following in utero cocaine. Recent studies in rodents have demonstrated that many GABAergic interneurons of the cerebral cortex are generated in the ganglionic eminence of the ventral telencephalon and migrate tangentially to reach the cerebral cortex (Anderson et al., 1997; Lavdas et al., 1999). These cells exhibit many phenotypic properties of neurons as they migrate, including the formation of axons and dendrites and the synthesis of neurotransmitter. It will be interesting to examine at what stage of a rather extensive embryonic migration the increased GABA levels appear following in utero cocaine.
Permanently reduced receptor-stimulated coupling of D1 receptors to their G proteins is another well-characterized abnormality following E8–E29 cocaine exposure, with an onset as early as E22 (Friedman et al., 1996; Jones et al., 2000). D1-like receptor activation decreases process outgrowth by cortical neurons (Reinoso et al., 1996; Todd, 1992), and thus a loss of D1 receptor signaling may underlie the increased dendritic growth following cocaine. In fact, neurons isolated from embryonic medial frontal cortex (including the ACC) from cocaine-exposed rabbit embryos show greater process extension than saline-matched control neurons in vitro (Jones et al., 2000). Our data show that significant levels of D1 receptor transcript and monoaminergic innervation are not present at E15, but are readily detectable at E22 in the rabbit cortex. These data are consistent with the finding that D1 receptor protein cannot be immunoprecipitated until after E15 (Jones et al., 2000). This timing of expression also is consistent with the finding that deleterious effects of cocaine exposure on the development of DA-rich cortical areas do not occur until D1 receptors are expressed.
The effects of gestational cocaine exposure in the rabbit are evident embryonically, with observable increases in dendritic length, abnormal in vitro growth properties and reduced D1 receptor coupling expressed as early as E21–E22 (Jones et al., 2000). The current study demonstrates that the ACC is not responsive in a deleterious fashion to in utero cocaine until after E15, suggesting that the onset of the modulation of developmental influences of monoaminergic systems by cocaine occurs rapidly, between E16 and E21. In addition, regimen duration is unlikely to be the key factor because relatively short periods of cocaine administration in postnatal or adult animals, and the early E8–E15 in utero scenario, do not produce permanent alterations. Rather, it appears that the precise developmental timing of prenatal cocaine exposure is crucial in determining the long-term effects on cortical structure and function.
The results of this study suggest that in utero cocaine modulates a coordinated developmental epoch consisting of differentiation of cortical neurons, innervation by DA fibers and expression of postsynaptic receptors. In this regard, D1 receptors normally become functionally coupled to their G protein soon after the onset of their expression (Sales et al., 1989). We hypothesize that cocaine-induced elevation of DA levels during this sensitive period, in which D1 receptor signaling pathways are being activated for the first time, produces rapid overexcitation of the system and permanent loss of D1–Gs coupling. Loss of normal D1 receptor-mediated influences on neuronal development and excitability might then lead to the plethora of anatomical and functional changes seen in the rabbit (Levitt et al., 1997; Levitt, 1998). For example, D1 receptors are expressed on pyramidal neurons and parvalbumin-positive GABAergic interneurons in the cortex (Bergson et al., 1995; Muly et al., 1998), the cellular elements in which we observe robust changes in response to in utero cocaine. Cortical regions receiving prominent DA input regulate arousal, memory, cognition and attention (Benes, 1993; Devinsky et al., 1995; Williams and Goldman-Rakic, 1995; Goldman-Rakic, 1998; Granon et al., 2000). Thus, deficits in D1 receptor signaling may contribute to the attentional and arousal deficits observed in the rabbit model (Romano et al., 1995; Romano and Harvey, 1996; Gabriel and Taylor, 1998), and in human children exposed to cocaine prenatally (Delaney-Black et al., 1996; Karmel and Gardner, 1996; Gingras and O'Donnell, 1998; Mayes et al., 1998; Richardson, 1998).
Relation to Other Animal Models of Prenatal Cocaine Exposure
A multitude of distinct patterns of deficits, ranging from global to specific, have been reported in animal models of prenatal cocaine exposure (Dow-Edwards et al., 1990; Gressens et al., 1992; Chen et al., 1993; Lidow, 1995, 1998; Ronnekleiv et al., 1995, 1998; Vorhees et al., 1995; Levitt et al., 1997; Kosofsky and Wilkins, 1998; Levitt, 1998; Spear et al., 1998; White et al., 1999). This widespread variability probably reflects the use of different species and administration paradigms (Dow-Edwards, 1996; Levitt, 1998). At high dosages, cocaine produces widespread deleterious effects on brain development. For example, oral administration of 10 mg/kg to pregnant rhesus monkeys for 62 days during pregnancy (E40–E102) produces profound and long-lasting abnormalities in lamination and glial differentiation throughout the cerebral cortex (Lidow, 1995, 1998). Similar findings have been observed following 20 mg/kg cocaine injected s.c. into pregnant mice (Gressens et al., 1992; Kosofsky and Wilkins, 1998).
Ironically, the mode of cocaine administration in animal studies that most closely models the human pharmacokinetics is under-represented — i.v. delivery (Dow-Edwards, 1996; Levitt, 1998). The rabbit model of i.v. cocaine use thus provides an opportunity to examine specific changes in ontogeny at low doses. There are no alterations in maternal weight, litter size, pup weight (at birth and throughout postnatal development), cortical thickness, laminar width and cell number following doses up to 4 mg/kg/injection (Murphy et al., 1995; Wang et al., 1995a). Thus far, our analysis indicates that deficits are induced only in regions of the cortex that receive prominent dopaminergic input (Wang et al., 1995a, 1996a; Jones et al., 1996, 2000), suggesting that cocaine-induced alterations of DA transmission produce the effects. Furthermore, our demonstration that cocaine-induced changes in neuronal growth are initiated around mid-gestation (Jones et al., 2000) has greatly abated the concern that problems in maternal-infant care account for the observed developmental changes. The current demonstration that significant effects on the development of the ACC are produced by exposure to a low dose of cocaine for only one-third of gestation provides further support for the specificity and lack of general teratology in the rabbit model. We have not examined whether cocaine exposure after this embryonic period is capable of producing effects on cortical development, but data from rodent models suggest that some aspects of CNS function can be regulated by neonatal cocaine exposure (Anderson-Brown et al., 1990; Frick and Dow-Edwards, 1995; Seidler et al., 1995).
Relation to Prenatal Cocaine Exposure in Humans
Initial reports of the impact of prenatal cocaine exposure on newborns were very confusing, because some reports suggested gross physical malformations, others observed specific but subtle deficits in cognitive and emotional development, and yet others observed no effects whatsoever [for reviews see Gingras et al., Hawley and Mayes et al. (Gingras et al., 1992; Hawley, 1994; Mayes et al., 1998)]. More recent and better-controlled studies have shown clearly that the offspring of relatively modest cocaine abusers show measurable functional deficits, particularly in cognitive and attentional domains (Delaney-Black et al., 1996; Karmel and Gardner, 1996; Gingras and O'Donnell, 1998; Mayes et al., 1998; Richardson, 1998; Scher et al., 2000).
The current data support the hypothesis that cocaine exposure during a short but key period of gestation is required to produce measurable CNS deficits in offspring. Using time of origin as the criterion, this time period corresponds to the second trimester in humans (Sidman and Rakic, 1973) and mid-gestation in the rhesus monkey (Rakic, 1977). However, it is clear that low-to-moderate cocaine usage during the first trimester by pregnant women produces cognitive and physiological deficits in children (Richardson, 1998; Scher et al., 2000). Pronounced effects on developing brain circuitry have also been observed in both a primate model in which cocaine is administered from day 20 to day 60 of a 165-day gestation (Ronnekleiv et al., 1995; 1998) and in rabbits receiving 30 mg/kg cocaine s.c. from day E7 to day E15 (Weese-Mayer et al., 1993). Therefore, the current results should not be interpreted as indicating that cocaine usage by women early in their pregnancy does not harm the development of the fetus. Rather, the parameters of cortical development that we measure appear to be most susceptible to cocaine exposure during this sensitive period of development.
This work was supported by DA 11165 to P.L. and a PhRMA Foundation fellowship to G.D.S. We thank Dr Frank Middleton for assistance in the isolation of the rabbit D1 receptor cDNA and in situ hybridization studies.
Address correspondence to Gregg D. Stanwood, PhD, Department of Neurobiology, University of Pittsburgh School of Medicine, E1440 Biomedical Science Tower, Pittsburgh, PA 15261, USA. Email: firstname.lastname@example.org.