Effects of Prenatal Cocaine Exposureon Embryonic Expression of Sonic HedgehogMARK J. KOEBBE,1,2 JEFFREY A. GOLDEN,3 GREGG BENNETT,4 RICHARD H. FINNELL,4AND SCOTT A. MACKLER1,2,5*1Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 191042Medical Research Service, Philadelphia Veterans Administration Medical Center, Philadelphia, Pennsylvania 191043Department of Pathology, Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine,Philadelphia, Pennsylvania 191044School of Public Health, College of Veterinary Medicine, Texas A&M University, College Station, Texas 778435Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104

ABSTRACT Cocaine use by pregnant womenmay adversely affect development and behavior in theexposed infants. Sonic hedgehog (shh) is a secretedprotein that induces development of many structures inthe embryo, including dopaminergic cells in the ventralmidbrain, the limb buds, and eyes. Because prenatalcocaine exposure has been shown to adversely affectthe morphogenesis of these and other systems, thepresent study was undertaken to test the hypothesisthat maternal cocaine treatment would alter shh mRNAexpression. Cocaine HCl (60 mg/kg i.p.) was adminis-tered to pregnant mice on gestational days 6–8, thetime that immediately precedes the appearance of shh.Control dams received i.p. saline. Embryos from gesta-tional days 9–11 were examined by in situ hybridization.The temporal and spatial patterns of shh expressionwere indistinguishable between embryos from cocaine-and saline-treated dams. Examination of forebrain,midbrain, and midbody spinal cord coronal sectionsfailed to reveal any differences in the dorsoventral andmediolateral localization of shh. The distribution ofmRNA for patched (ptc), the membrane receptor forshh, was also indistinguishable between both groups.Chick embryos were next used to examine the directapplication of cocaine into the developing brain. Shhdistribution was similarly unaffected in these chickembryos. These data show that maternal cocainetreatment during early neural tube development doesnot significantly alter the expression patterns ofshh or ptc mRNA. Thus, congenital defects and be-havioral abnormalities associated with maternal co-caine use do not appear to result from altered expres-sion of the shh-ptc pathway. Teratology 59:12–19,1999. r 1999 Wiley-Liss, Inc.

Cocaine use by pregnant women has been associatedwith an elevated risk of congenital abnormalities and isthought to adversely affect neurobehavioral develop-ment in the exposed offspring (Chasnoff et al., ’89;Mackler et al., ’97). Experiments using pregnant ro-dents support the hypothesis that cocaine can act as ateratogen (El-Bizri et al., ’91; Finnell et al., ’90; Fisher

et al., ’94). Anomalies of the urogenital, cardiovascular,and skeletal structures, along with changes in the sizesof the cerebral ventricles, were seen in the exposedoffspring. Evidence for abnormal neurulation (includ-ing neural tube defects) did not occur at increasedrates, even when cocaine was administered during thetime of neural tube closure (Finnell et al., ’90). Theresults of these and other studies indicate that prenatalcocaine treatment does not cause obvious structuralabnormalities in the majority of the exposed offspring.

Studies investigating mRNA levels in mouse embryosduring neural tube closure and early central nervoussystem (CNS) development demonstrated that cocaineaffected only a limited number of neural genes (Mackleret al., ’96), suggesting that prenatal cocaine exertsspecific effects on embryonic CNS development. Inutero exposure to cocaine also altered the amounts ofcertain neurotransmitters and receptors in specificregions of the CNS in postnatal rats. These includedincreased binding sites for somatostatin (Rodriquez-Sanchez and Arilla, ’91) and b1 adrenergic ligands(Henderson et al., ’91) in the caudate-putamen, andaltered dopamine levels in the caudate-putamen(Henderson and McMillen, ’93). The majority of thesealterations returned to normal or near-normal valuesat later postnatal ages; in addition, there have not beensignificant effects observed after cocaine exposure inother neuronal molecules (de Bartolomeis et al., ’94;Henderson et al., ’91). These data indicate that theactions of prenatal cocaine exposure are limited, maynot persist during postnatal CNS development, andmay complicate experiments that examine the adverseeffects of in utero cocaine treatment on postnatalbehavior. However, studies have shown that prenatal

Grant sponsor: NIDA; Grant numbers: 07241, 00199.

*Correspondence to: Scott A. Mackler, Medical Research Service(151C), Philadelphia Veterans Administration Medical Center, Univer-sity and Woodland Aves., Philadelphia, PA 19104. E-mail:smackler@mail.med.upenn.edu

Received 12 May 1998; Accepted 25 August 1998

TERATOLOGY 59:12–19 (1999)

r 1999 WILEY-LISS, INC.

cocaine exposure does affect behavior in the postnatalperiod. Differences between offspring from cocaine- andsaline-treated rats have been observed in locomotoractivity (Hutchings et al., ’89), classical conditioning(Spear et al., ’89), and nonassociative learning (Heyseret al., ’94). The dopaminergic A9 and A10 cells in themidbrains of cocaine-exposed rats exhibited decreasedspontaneous electrical activity (Minabe et al., ’92), anddifferences in dopamine levels in the nucleus accum-bens occurred after prenatal cocaine treatment, usingmicrodialysis in juvenile and adult rats (Keller et al.,’96). Dopaminergic effects of cocaine on the mesostria-tal pathway may help to explain observed abnormali-ties in locomotor behavior.

Several molecules are involved in the induction of thecomplex pattern of vertebrate CNS development. Thesonic hedgehog protein (shh) appears to be critical forthe dorsoventral organization of the neural tube (Ech-elard et al., ’93; Roelink et al., ’95) and the dermamyo-tome (Fan and Tessier-Lavigne, ’94). Shh acts along therostrocaudal extent of the neural tube, inducing differ-entiation of the ventral forebrain rostrally (Ericson etal., ’95), dopaminergic neurons in the midbrain (Hyneset al., ’95a, ’97), and motor neurons in the spinal cord(Roelink et al., ’95). Prenatal cocaine administrationdoes change postnatal striatal dopamine levels (Hender-son and McMillen, ’93) and activity of midbrain dopa-minergic neurons (Minabe et al., ’92). The presentstudy examined the hypothesis that shh expressionwould be altered in the specific regions of the mouseand chick CNS that are affected by prenatal cocainetreatment, including the dopaminergic portion of theventral midbrain. The spatial and temporal expressionof the mRNA patched (ptc), which encodes the mem-brane receptor for shh (Chen and Struhl, ’96; Goodrichet al., ’96), were also studied in the mouse to determineif mRNA levels that are in part regulated by shh arealtered by maternal cocaine use. In addition, shhdistribution was examined in the limb buds, a secondstructure thought to be adversely affected by prenatalcocaine exposure (Finnell et al., ’90).

MATERIALS AND METHODS

Drugs and chemicals

Cocaine HCl was a gift from the NIDA and wasdissolved in 0.9% NaCl. Sodium heparin, sheep serum,and levamisole were purchased from Sigma (St. Louis,MO). Digoxigenin-UTP, the anti-digoxigenin antibody(Fab fragments), 4-nitroblue tetrazolium chloride (NBT),and bromo-chloro-indolyl phosphate (BCIP) were pur-chased from Boehringer-Mannheim, Inc. (Indianapolis,IN). Affi-gel blue gel beads were purchased from Bio-Rad (Hercules, CA).

Cocaine treatment and collectionof mouse embryos

LM/Bc mice were maintained on a 12-hr/12-hr light/dark cycle in the Animal Research Facilities at TexasA&M University, and handled according to NIH guide-

lines. Female mice were examined each morning afterbeing housed with a male mouse, and if a vaginal plugwas present, then 10 PM of the previous night wasdesignated as the time of conception. Each pregnantmouse received either 60 mg/kg of cocaine HCL insaline or an equivalent volume of saline i.p. at 10 AM ongestational days (G.D.) 6–8; this dose and route ofadministration of cocaine result in the largest numberof structural defects at birth (Finnell et al., ’90). Micewere sacrificed by cervical dislocation at 10 AM on G.D.9.5, 10.5, or 11.5. The embryos were quickly removedfrom all surrounding tissues, rinsed in phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde inPBS overnight at 4°C, and then dehydrated in serialwashes of PBS plus 0.1% Tween-20 (PBT), 25%/75%PBT/methanol, 50%/50% PBT/methanol, 25%/75% PBT/methanol, and 100% methanol. The embryos werestored in 100% methanol at 220°C until later use.

Treatment of chick embryos

Chick embryos were incubated to stage 10, 11, or 12(Hamburger and Hamilton, ’51) in a humidified rotat-ing incubator. Eggs were removed from the incubatorand windowed to visualize the embryo, as previouslydescribed (Golden et al., ’95). In the first group ofembryos, a 30-gauge needle attached to a 0.5-ml tuber-culin syringe was used to make a small slit in the dorsalrhombencephalon. A single Affi-gel bead, previouslysoaked in 10 nM, 1 µM, or 100 µM cocaine diluted in 0.1M PBS was implanted into the lumen of the neural tubethrough the dorsal cut, using #55 forceps. Controlembryos were implanted with beads soaked in 1 3 PBSalone or were surgically manipulated but not implantedwith any bead. Once in the lumen of the neural tube atthe level of the rhombencephalon, the bead was movedinto the posterior prosencephalon by manipulating thebead through the wall of the neural tube with #55forceps. The eggs were sealed with clear packaging tapeand returned to a humidified incubator, but were notrocked. In a second group of embryos, cocaine (0.1 µM,10 µM, or 1 mM) was directly injected into the amnioticfluid for 2 consecutive days. Embryos from both groupswere harvested on embryonic day 4 (E4), fixed in 4%paraformaldehyde, washed three times in 1 3 PBT, anddehydrated into 100% methanol. Embryos were thenstored in 100% methanol at 220°C until in situ hybrid-ization was performed.

Whole-embryo in situ hybridizations

cRNA was synthesized from linearized plasmids thatcontained either a partial-length rat shh cDNA (642base pairs, starting at nucleotide 315 of the openreading frame; kind gift of A. McMahon), a full-lengthchick shh cDNA (1.6 kB; kind gift of R. Riddle), or apartial-length mouse ptc cDNA. The ptc cDNA wascreated by PCR amplification of a 538-base pair cDNAfrom the 38 end of ptc using cDNA synthesized frommidterm mouse total RNA and the primers MPTC1(58-ttgcatctgttggcatcgg-38) and MPTC2 (58-cgggcaaaga-

PRENATAL COCAINE AND SONIC HEDGEHOG 13

cagggttttc-38), and subcloned into the Litmus 28 (NewEngland Biolabs, Beverly, MA) plasmid, and its identitywas confirmed by direct DNA sequencing. cRNA synthe-sis included the incorporation of digoxigenin-labeledUTP, and the products from each reaction were visual-ized by ethidium bromide staining after electrophoresisin an agarose gel. Both sense and antisense cRNAswere synthesized for shh and ptc. The in situ hybridiza-tions were performed as previously described (Conlonand Herrmann, ’93). Embryos were handled identicallyin groups of 4–8 in a scintillation vial, and comparablenumbers of embryos were hybridized to a sense cRNAin each experiment. The embryos were rehydrated withgraded mixtures of PBT/methanol until in 100% PBT,and then endogenous peroxidase activity was reducedby placing the embryos in 6% H2O2/PBT for 1 hr. Afterone wash in PBT containing glycine (2 mg/ml) and twowashes in PBT, the embryos were postfixed in 4%paraformaldehyde and 0.2% glutaraldehyde/PBT for 20min and then placed in a prehybridization solution(5 3 SSC (pH 4.5); 50 µg/ml yeast tRNA; 50 µg/mlheparin; 1% SDS; and 50% formamide) at 70°C for 1 hr.Hybridization proceeded overnight at 70°C in a rotat-ing oven, with ,1 µg/ml of cRNA added to freshprehybridization fluid. Posthybridization washes onday 2 consisted of three washes (30 min each) at 70°C in5 3 SSC/1% SDS/50% formamide, three washes (30 mineach) at 65°C in 2 3 SSC/50% formamide, and threewashes (5 min each) at 22°C in TBST (140 mM NaCl;2.7 mM KCl; 25 mM Tris HCl (pH 7.5); 1% Tween-20;and 2 mM levamisole). A preblock step consisted ofincubation in 10% sheep serum/TBST for 2.5 hr at22°C. The embryos were rocked gently overnight withTBST that contained anti-digoxigenin antibody (1/2,000 dilution). The anti-digoxigenin antibody had firstbeen preadsorbed to protein isolated from embryos ofsimilar ages. On day 3, after incubation with theanti-digoxigenin antibody, the embryos were washedfive times (1 hr each) at 22°C in TBST and then rockedovernight in TBST at 4°C. Detection of cRNA probes onday 4 included three washes in NTMT (100 mM NaCl;100 mM Tris HCl (pH 9.5); 50 mM MgCl2; 0.1%Tween-20; and 2 mM levamisole) and incubation for,20–30 min at 22°C with 2 ml of NTMT containing 6.75µl of NBT and 5.25 µl of BCIP. After visual inspectiondetermined that the reaction time was sufficient, theembryos were washed in PBT (pH 5.5) for 10 min at22°C, postfixed in 4% paraformaldehyde/0.2% glutaral-dehyde/PBS for 1 hr at 22°C, washed again in PBT, andcleared in glycerol.

Examination of coronal sectionsfrom mouse embryos

Individual embryos were next dehydrated in serialethanol solutions (final, 100% ethanol), embedded inparaffin, and oriented so that coronal sections could becut using a microtome. Sections 7–10 µm in thicknesswere mounted on glass slides. Slides were dipped inxylene to remove the paraffin, and examined using a

light microscope. Each embryo was examined in itsentirety, and comparisons were made between cocaine-and saline-exposed litters at specific locations.

RESULTS

The appearance of shh expression in a control, G.D.10.5 whole embryo is shown in Figure 1. The reactionproduct is easily visualized as a longitudinal streakalong the notochord and ventral region of the CNS. ShhmRNA is also localized to the distal portions of the limbbuds, endodermal structures, and the cardiovascularsystem. The specificity of the cRNA probe is demon-strated by the lack of visible staining in a similarembryo to which a sense cRNA probe was allowed tohybridize (Fig. 1).

The temporal pattern of shh expression can be seenin control embryos from G.D. 9.5 and 10.5 (Fig. 2A,C).These patterns of shh expression appeared similarbetween whole embryos from cocaine- and saline-treated mothers (Fig. 2). Shh mRNA is restricted to thesame longitudinal region of the ventral CNS, the distallimb buds, endodermal structures, and the cardiovascu-lar system. The intensity of the in situ hybridizationcolor reaction product is also similar between the twogroups of embryos. These whole-embryo in situ hybrid-ization studies indicate that maternal cocaine treat-ment does not dramatically alter the temporal andspatial patterns of shh expression (Table 1; similarpatterns were observed in G.D. 11.5 embryos, data not

Fig. 1. Pattern of shh expression in a G.D. 10.5 whole mouse embryo.A: The use of a sense cRNA probe did not result in accumulation of acolor reaction product within the whole embryo. B: Specific staining(the dark colors indicate the locations of shh expression) is visible inmultiple regions of a whole embryo handled in an identical manner,except that an antisense cRNA probe was used in place of the sensecRNA probe. Areas that are apparent include the mesencephalon (M),the floor plate region and notochord along the entire ventral spinalcord (SC), and the zone of polarizing activity (ZPA) in one limb bud.These embryos were collected from a dam that was treated with salinei.p.; the expression pattern is identical to that of previous findings(Echelard et al., ’93; Roelink et al., ’95).

14 M.J. KOEBBE ET AL.

shown). There were no observed intra- or interlitterdifferences in the numbers or gross appearance of themouse embryos from all dams.

The dorsoventral and mediolateral distributions ofshh were next studied in coronal sections along theentire rostral-caudal axis of the CNS; specific regionsare shown in Figures 3 and 4. Shh expression wasrestricted to the ventral regions of the neural tubealong the entire rostral-caudal axis in embryos fromsaline-treated litters (Figs. 3A,C,E, 4A,C,E), in agree-ment with previous studies (Echelard et al., ’93; Ro-elink et al., ’95). The pattern of shh expression wasindistinguishable in embryos collected from cocaine-treated dams (Figs. 3B,D,F, 4B,D,F).

The pattern of expression of ptc, the mRNA thatencodes the shh membrane receptor, was also examined

to determine whether or not maternal cocaine treat-ment could alter other components of the shh pathwayfor neural induction. The appearance of ptc expressionwas identical between embryos collected from saline-and cocaine-treated dams. Ptc expression overlaps thatof shh development (Goodrich et al., ’96); one example isshown in Figure 5.

Shh induces the formation of many other structures,including the zone of polarizing activity (ZPA) in theposterior portion of the developing limb bud. Maternalcocaine treatment can lead to abnormal development ofthe distal limb (El-Bizri et al., ’91; Finnell et al., ’90;Fisher et al., ’94). The location of the ZPA as indicatedby shh expression was no different in location betweensaline- and cocaine-exposed embryos (Fig. 6).

Direct administration of cocaine into the developingchick brain did not lead to changes in the pattern of shh

Fig. 2. Prenatal cocaine treatment did not change the spatial andtemporal expression of shh in G.D. 9.5 and 10.5 whole mouse embryos.Embryos whose mothers received saline i.p. injections (A, C) arecompared to embryos whose mothers were treated with cocaine i.p. onG.D. 6–8 (B, D). No differences between cocaine-exposed and controlembryos were detected at either G.D. 9.5 (A, B) or 10.5 (C, D). Thespatial pattern of shh expression along the CNS rostral-caudal axis inthe cocaine-exposed embryos is similar to that for the embryos ofsaline-treated dams, and is also the same as that described for thedeveloping mouse embryo (Echelard et al., ’93; Roelink et al., ’95).There were also no differences observed between G.D. 11.5 embryosfrom cocaine- or saline-treated dams.

TABLE 1. Total number of embryos examined for shhexpression (no. of litters)

Maternaltreatment G.D. 9.5 G.D. 10.5 G.D. 11.5

ip saline 19 (4) 10 (5) 2 (1)i.p. cocaine 27 (6) 17 (6) 5 (2)

Fig. 3. The dorsal-ventral pattern of shh expression at G.D. 9.5 is notaltered by prenatal cocaine treatment. Examination of serial coronalsections revealed that shh expression was restricted to the ventralportion of the developing CNS (the dark colors indicate the locations ofshh). Representative examples are shown from the forebrain (A, B),midbrain (C, D), and spinal cord at the level of the hindbrain-spinalcord junction (E, F). Arrows denote the ventral location of shhexpression in the CNS (A, C, E). A: Portions of the hindgut and caudalspinal cord are positioned at top right; staining is also restricted to theventral regions in these structures. The sections of the midbrain (C, D)include cuts of both the rostral (top; arrow indicates shh) and caudal(bottom, with reverse dorsal-ventral orientation; arrowheads indicateshh) regions. The notochord also expresses shh (asterisk in E). A–F:Dorsal is to the top. Scale bar 5 250 µm.

PRENATAL COCAINE AND SONIC HEDGEHOG 15

expression (Fig. 7). No abnormalities were observed in15 embryos (developmental stages 10–12) that had acocaine-soaked bead placed into the brain, althoughtwo embryos did look smaller compared to the otherembryos. Amniotic fluid injections of cocaine for 2 daysalso did not alter shh expression in 8 of 9 embryos(developmental stages 10–12); one embryo in this groupdid exhibit a small expansion in the distribution of shhcompared to control embryos (all embryos that wereexposed to cocaine via amniotic fluid injections were ofsimilar sizes).

DISCUSSION

The present study demonstrates that treatment of apregnant mouse with a teratogenic dose of cocaine priorto and throughout the period of neural tube closuredoes not affect the spatial and temporal patterns of shhexpression in the CNS of LM/Bc embryos. Gross defectsin CNS organization have not been consistently re-ported in newborn rats and mice whose mothers weretreated with cocaine during pregnancy (Finnell et al.,

’90; Minabe et al., ’92), suggesting that if prenatalcocaine exposure does disrupt brain development, itmay do so by causing subtle changes in neuronalorganization. Sonic hedgehog, a critical inductive sig-nal for the identity of ventral neurons along the rostro-caudal extent of the CNS, represents one molecule that

Fig. 4. The dorsal-ventral pattern of shh expression at G.D. 10.5 isnot altered by prenatal cocaine treatment. Similar representativecoronal sections are shown for G.D. 10.5 embryos after maternaltreatment with either saline (A,C,E) or cocaine (B, D, F). There are noobvious differences between cocaine-exposed and control embryos inthe dorsal-ventral distribution of shh in the forebrain (A, B), midbrain(C, D), or spinal cord (E, F) at the level of the hindbrain-spinal cordjunction. A–F: Dorsal is to the top. Orientation and symbols are thesame as in Figure 3. Scale bar 5 500 µm (A–D) or 250 µm (E, F).

Fig. 5. Prenatal cocaine treatment also did not alter expression of theshh receptor ptc at G.D. 9.5. Whole embryos that were stained witheither an antisense shh probe (A) or an antisense patched probe (B)are shown together. The spatial expression of ptc at this age overlapswith that of shh (Goodrich et al., ’96), which is similar to the presentfindings. Both of these embryos were collected from litters of damsthat had been treated with the same dose of cocaine on G.D. 6–8 (60mg/kg, i.p.). The results obtained with the ptc probe are the same aswith embryos collected from saline-treated mice (not shown).

Fig. 6. Prenatal cocaine treatment did not affect shh expression in theZPA of the developing limb bud. Shh is expressed in the ZPA ofdeveloping limb buds. Shh was restricted to the posterior region of thedeveloping limb bud in both saline- and cocaine-exposed embryos, asviewed in the whole embryo (A, saline; B, cocaine). Coronal sectionsfrom these embryos were made perpendicular to the plane of the limbbuds, as shown in A and B. Shh expression was limited to a small zonein these limb buds (C, saline; D, cocaine). Scale bar 5 250 µm.

16 M.J. KOEBBE ET AL.

could conceivably be affected by cocaine and result inabnormal CNS development. We hypothesized that theventral midbrain would be one specific structure todemonstrate cocaine-associated changes in shh expres-sion, because prenatal cocaine treatment did alterdopaminergic cell function (Keller et al., ’96; Minabe etal., ’92). Detailed examination using in situ hybridiza-tion studies (Figs. 2–5) failed to support this hypoth-esis; the expression of shh mRNA was unchangedthrough midgestation of the developing mouse, despiteteratogenic doses of cocaine. Expression of ptc mRNAalso did not appear to be altered by the same scheduleof maternal cocaine treatment.

Other structures that can be affected by maternalcocaine treatment include the eyes, limbs, urogenitalorgans, and heart (El-Bizri et al., ’91; Finnell et al., ’90;Fisher et al., ’94; Silva-Araujo et al., ’96). Expression ofshh and related genes regulates pattern formation inthese same structures (Ekker et al., ’95; Levin et al.,’95). Examination of the zone of polarizing activity inthe developing limb buds (Fig. 6) did not reveal anyalterations in the pattern of shh mRNA expression afterprenatal cocaine treatment. These results suggest thatextra-CNS expression of shh mRNA is also not affectedby maternal cocaine treatment. Approximately 5–12%of the exposed embryos would, however, be expected tocontain cardiovascular, genitourinary, or limb malforma-tions (Finnell et al., ’90). No abnormalities were ob-

served in the pattern of shh expression in more than 45examined embryos before the time when these congeni-tal malformations could be observed (Finnell et al., ’90).

Patched, a product of the ptc gene, is thought to beone multiple-pass transmembrane protein that acts asa receptor for sonic hedgehog (Chen and Struhl, ’96).Secretion of sonic hedgehog activates transcription ofptc and also opposes the function of the patched proteinin nearby cells, a coordinated mechanism that helpsto regulate pattern formation during development(Goodrich et al., ’96). Prenatal treatment of pregnantmice with cocaine did not affect the distribution of ptc(Fig. 5). This result indicates that teratogenic doses ofcocaine administered during neurulation did not alterexpression of ptc, and demonstrates a lack of an effect ofcocaine on one mRNA, ptc, that has its expressionregulated by shh. Even if prenatal cocaine producedchanges in shh expression that were not detected by thepresent study, they did not result in changes in ptcexpression. This lack of an effect on ptc expressionsupports the finding that prenatal cocaine does notresult in significant functional changes in shh levels.

Cocaine also failed to adversely affect shh expressionin the chick embryo (Fig. 7). These experiments wereperformed in order to determine whether or not thepresence of cocaine within the embryonic CNS altersmidbrain development. Cocaine was directly adminis-tered into either the amniotic fluid or the telencepha-

Fig. 7. Direct exposure of the developing chick embryo to cocaine did not alter shh expression in the CNS.A: Shh expression in an untreated chick whole embryo, viewed from above. The color reaction product isvisible in the ventral diencephalon (arrowhead), the zona limitans interthalamica (arrow), and the ZPA ofa limb bud (asterisk). B: Similar structures in the developing brain of a chick embryo that underwentimplantation of a cocaine-soaked bead (double arrow) demonstrated an identical pattern of shhexpression. Arrow and arrowhead in B same as in A.

PRENATAL COCAINE AND SONIC HEDGEHOG 17

lon, avoiding both concerns of placental transfer andthe indirect effects mediated by cocaine on uterine andplacental blood flow that are known to occur in thepregnant mammal (Rama-Sastry, ’95). Cocaine wasadded at relatively high concentrations, and in a smallnumber of chick embryos structural defects were ob-served. However, even in these abnormal embryos thepattern of shh expression in the CNS appeared normal.Again, the presence of cocaine in the developing CNSdoes not appear to affect shh expression.

The schedule for cocaine administration in the pre-sent study was selected for two reasons. Expression ofshh mRNA begins in the neural tube at G.D. 8–9 in themouse, is present for the next 4–5 G.D., and is no longerdetectable late in pregnancy (Echelard et al., ’93; Hyneset al., ’95b; Roelink et al., ’95). The peak of shhexpression in the midbrain occurs during G.D. 9–11.Treatment of the dam immediately prior to the onset ofshh expression and throughout the early stages ofneural tube development seemed to be the optimal timeto analyze the effects of cocaine on shh mRNA expres-sion. In addition, this dose and timing of cocaineadministration resulted in the greatest number ofcongenital malformations without adversely affectingmaternal health (Finnell et al., ’90). Studies of cocaine’sgestational consequences have used other routes ofdrug administration. Intraperitoneal injections maydirectly affect the embryos by damage to the uterineartery or the uterus, but we did not observe any signs ofsuch damage in the examined dams at time of sacrifice.The effects of cocaine on embryonic development mayresult from altered uterine blood flow and/or transportof cocaine across the placenta. Different routes ofcocaine administration (i.p., s.c., i.v., or oral) maycontribute to these effects because they all lead to thepresence of cocaine in the systemic circulation.

Earlier studies that examined the effects of prenatalcocaine on development used many different schedulesfor maternal cocaine administration, and these variedschedules may have contributed to discrepancies infetal or postnatal outcomes. For example, the majorityof biochemical or behavioral studies that examinedpostnatal animals administered cocaine throughoutmost of the pregnancy. Prenatal treatment with cocaineat lower doses or after neural tube closure would seemless likely to alter shh expression; however, thesepossibilities were not addressed in the present study.Daily treatment of pregnant rhesus monkeys withcocaine (q.i.d. from weeks 3–4 of pregnancy) did resultin specific biochemical differences in midbrain dopamin-ergic cells (Ronnekleiv and Naylor, ’95; Fang et al., ’97).One difference between these experiments and thepresent study is the timing and duration of cocaineadministration relative to neural tube formation andshh expression.

In situ hybridization analysis does not reliably quan-tify amounts of mRNA molecules. One limitation of thepresent study is that small differences in shh and ptcmRNA levels may have occurred in different cellular

populations between cocaine- and saline-exposed em-bryos. If such changes in abundance did occur and werenot detected in whole embryos, then a more detailedstudy of the dorsoventral and mediolateral distributionof shh should be expected to demonstrate differences.Examination of many coronal sections did not revealany abnormal patterns of distribution (Figs. 3, 4),supporting the conclusion that prenatal cocaine doesnot significantly affect shh mRNA expression. Prelimi-nary experiments, using Northern blots and RNAseprotection assays, have similarly not revealed consis-tent and significant changes in shh levels in the wholemouse embryo after maternal cocaine treatment (Tsueiand Mackler, unpublished observations). The presentresults do not indicate whether or not changes in shhprotein levels may have occurred. However, in situhybridization permits the study of mRNA distributionin the whole embryo. Western blots would need to beperformed to determine if changes in protein levelsoccur in response to maternal cocaine treatment, in theabsence of significant changes in mRNA levels.

A strong connection between prenatal cocaine use,altered CNS development, and abnormal behavior inhumans has been difficult to establish, although initialreports raised appropriate concerns (Chasnoff et al.,’89). Potential confounding variables present in descrip-tions of cocaine-exposed human infants, including theabuse of other drugs and poor prenatal health care,may have limited the validity of these observations.Outcomes of children exposed to cocaine while in uterowho are entering school are now being reported. Long-term effects of prenatal cocaine exposure on develop-ment and behavior that are solely attributable tococaine, in humans and in animal studies, have notbeen easy to demonstrate (Mackler et al., ’97). In thepresent study, the lack of an effect of cocaine on thepatterns of shh mRNA expression, a critical inducer ofboth CNS and non-CNS tissues during embryonicdevelopment, supports the hypothesis that prenatalcocaine exposure does not dramatically alter develop-ment in the majority of mouse embryos. Other, moresubtle alterations may occur in the CNS after exposureto cocaine that result in postnatal behavioral changes.However, the majority of children appear to be able torecover from the effects of prenatal cocaine exposure;the observed subtle behavioral deficits may not remainevident with global measures of cognition (Hurt et al.,’96).

ACKNOWLEDGMENTS

We thank C.P. O’Brien for helpful discussions, andthe Philadelphia VAMC medical media service for helpwith the figures.

LITERATURE CITEDChasnoff I, Griffith D, MacGregor S, Dirkes K, Burns K. 1989.

Temporal patterns of cocaine use during pregnancy. JAMA261:1741–1744.

Chen Y, Struhl G. 1996. Dual roles for patched in sequestering andtransducing hedgehog. Cell 87:553–563.

18 M.J. KOEBBE ET AL.

Conlon R, Herrmann B. 1993. Detection of messenger RNA by in situhybridization to postimplantation embryo whole mounts. MethodsEnzymol 225:373–383.

de Bartolomeis A, Austin MC, Goodwin GA, Spear LP, Pickar D,Crawley JN. 1994. Dopaminergic and peptidergic mRNA levels injuvenile rat brain after prenatal cocaine treatment. Mol Brain Res21:321–330.

Echelard Y, Epstein DJ, St-Jacques B, Shen L, Mohler J, McMahonJA, McMahon AP. 1993. Sonic hedgehog, a member of a family ofputative signaling molecules, is implicated in the regulation of CNSpolarity. Cell 75:1417–1430.

Ekker SC, Ungar AR, Greenstein P, von Kessler DP, Porter JA, MoonRT, Beachy PA. 1995. Patterning activities of vertebrate hedgehogproteins in the developing eye and brain. Curr Biol 5:944–955.

El-Bizri H, Guest I, Varma D. 1991. Effects of cocaine on rat embryodevelopment in vivo and in cultures. Pediatr Res 29:187.

Ericson J, Muhr J, Placzek M, Lints T, Jessell TM, Edlund T. 1995.Sonic hedgehog induces the differentiation of ventral forebrainneurons: a common signal for ventral patterning within the neuraltube. Cell 81:747–756.

Fan CM, Tessier-Lavigne M. 1994. Patterning of mammalian somitesby surface ectoderm and notochord: Evidence for sclerotome induc-tion by a hedgehog homolog. Cell 79:1175–1186.

Fang Y, Janowsky A, Ronnekleiv OK. 1997. Cocaine exposure in fetalrhesus monkey: Consequences for dopamine D1- and D2-like recep-tor binding densities. Brain Res Dev Brain Res 104:163–174.

Finnell R, Toloyan S, Van Waes M, Kalivas P. 1990. Preliminaryevidence for a cocaine-induced embryopathy in mice. Toxicol ApplPharmacol 103:228–237.

Fisher JE, Potturi RB, Collins M, Resnick E, Zimmerman EF. 1994.Cocaine-induced embryonic cardiovascular disruption in mice. Tera-tology 49:182–191.

Golden J, Fields-Berry S, Cepko C. 1995. Construction and character-ization of a highly complex retroviral library for lineage analysis.Proc Natl Acad Sci USA 92:5704–5708.

Goodrich LV, Johnson RL, Milenkovic L, McMahon JA, Scott MP. 1996.Conservation of the hedgehog/patched signaling pathway from fliesto mice: Induction of a mouse patched gene by Hedgehog. Genes Dev10:301–312.

Hamburger V, Hamilton H. 1951. A series of normal stages in thedevelopment of the chick embryo. J Morphol 88:49–91.

Henderson M, McMillen B. 1993. Changes in dopamine, serotonin,and their metabolites in discrete brain areas of rat offspring after inutero exposure to cocaine or related drugs. Teratology 48:421–430.

Henderson M, McConnaughey M, McMillen B. 1991. Long-termconsequences of prenatal exposure to cocaine or related drugs:Effects on rat brain monoaminergic receptors. Brain Res Bull26:941–945.

Heyser CJ, McKinzie DL, Athalie F, Spear NE, Spear LP. 1994. Effectsof prenatal exposure to cocaine on heart rate and nonassociativelearning and retention in infant rats. Teratology 49:470–478.

Hurt H, Brodsky NL, Braitman LE, Giannetta J. 1996. Inner citychildren with in utero cocaine exposure do not differ from controls on

Wechsler Preschool and Primary Scales of Intelligence-Revised(WPPSI-R). Pediatr Res 49:268.

Hutchings D, Fico T, Dow-Edwards D. 1989. Prenatal cocaine: Mater-nal toxicity, fetal effects, and locomotor activity in rat offspring.Neurotoxicol Teratol 11:65–69.

Hynes M, Porter JA, Chiang C, Chang D, Tessier-Lavange M, BeachyPA, Rosenthal A. 1995a. Induction of midbrain dopaminergic neu-rons by sonic hedgehog. Neuron 15:35–44.

Hynes M, Poulsen K, Tessier-Lavigne M, Rosenthal A. 1995b. Controlof neuronal diversity by the floor plate: Contact-mediated inductionof midbrain dopaminergic neurons. Cell 80:95–101.

Hynes M, Stone D, Dowd M, Pitts-Meek S, Goddard A, Gurney A,Rosenthal A. 1997. Control of cell pattern in the neural tube by thezinc finger transcription factor and oncogene Gli-1. Neuron 19:15–26.

Keller RW, Johnson KS, Snder-Keller AM, Carlson JN, Glick SD. 1996.Effects of prenatal cocaine exposure on the mesocorticolimbic dopa-mine system: An in vivo microdialysis study in the rat. Brain Res742:71–79.

Levin M, Johnson R, Stern C, Kuehn M, Tabin C. 1995. A molecularpathway determining left-right asymmetry in chick embryogenesis.Cell 82:803–814.

Mackler S, Bennett G, Tsuei V, Finnell R. 1996. Cocaine selectivelyalters neurotransmitter receptor mRNAs in mouse embryos. ReprodToxicol 10:37–42.

Mackler S, Finnell R, Vorhees C, Hurt H. 1997. The effects of prenatalcocaine on CNS development. Natl Inst Drug Abuse Res Monogr Ser178:27–29.

Minabe Y, Ashby CR, Heyser C, Spear LP, Wang RX. 1992. The effectsof prenatal cocaine exposure on spontaneously active midbraindopamine neurons in adult male offspring: an electrophysiologicalstudy. Brain Res 586:152–156.

Rama-Sastry B. 1995. Cocaine addiction, placental function, and fetalgrowth. Placental toxicology. CRC Press. p 134–159.

Rodriquez-Sanchez M, Arilla E. 1991. Changes in striatal somatosta-tin receptors in pups after cocaine administration to pregnant andnursing dams. Neurosci Lett 134:37–40.

Roelink H, Porter JA, Chiang C, Tanabe Y, Chang DT, Beachy PA,Jessell TM. 1995. Floor plate and motor neuron induction bydifferent concentrations of the amino-terminal cleavage product ofsonic hedgehog autoproteolysis. Cell 81:445–455.

Ronnekleiv OK, Naylor BR. 1995. Chronic cocaine exposure in thefetal rhesus monkey: Consequences for early development of dopa-mine neurons. J Neurosci 15:7330–7343.

Silva-Araujo A, Silva MC, Simon A, Nguyen-Legros J, Ali SF, TavaresMA. 1996. The effects of prenatal exposure to cocaine on thedopaminergic cells in the rat retina. An immunocytochemical andneurochemical study. Exp Eye Res 62:697–708.

Spear LP, Kirstein CL, Bell J, Yoottanasumpun V, Greebaum R,O’Shea J, Hoffman H, Spear NE. 1989. Effects of prenatal cocaineexposure on behavior during the early postnatal period. Neurotoxi-col Teratol 11:57–63.

PRENATAL COCAINE AND SONIC HEDGEHOG 19