Introduction

Top Abstract Introduction Methods Results Discussion References

Despite extensive adverse publicity, approximately 25% of all pregnant women in the United States smoke (Bardy et al., 1993; DiFranza and Lew, 1995). The consequences for human development have been well identified in epidemiological studies: tens of thousands of spontaneous abortions and neonatal intensive care unit admissions annually and thousands of perinatal deaths and deaths from sudden infant death syndrome (DiFranza and Lew, 1995). Perhaps of greater long-term impact on society, it is now recognized that maternal smoking substantially increases the risk of learning disabilities, behavioral problems, and attention deficit/hyperactivity disorder in the offspring (Naeye, 1992; Bell and Lau, 1995; DiFranza and Lew, 1995). Although these correlations imply that maternal smoking has the potential to elicit fetal brain damage, the direct proof of a causative relationship has proven elusive because of the many covariables that operate in smoking, such as low socioeconomic status, poor prenatal care or co-abuse of other substances. These problems can be resolved through the use of animal models that incorporate exposures to the separable components present in tobacco but that obviously do not share societal covariables.

Studies of the effects of nicotine on fetal development (review, Slotkin, 1992) have confirmed that nicotine, by itself, is capable of causing fetal brain damage, as evidenced by cell loss, synaptic abnormalities and behavioral deficits. Although cigarette smoke contains thousands of bioactive compounds, the finding that nicotine is injurious to the developing brain provides a mechanistic connection between maternal smoking and adverse fetal/neonatal outcomes. In rats, fetal brain damage is demonstrable at plasma nicotine levels that simulate those found in human smokers and that do not evoke malformations or otherwise compromise general developmental parameters such as growth. Furthermore, regional targeting of adverse effects of nicotine on the developing brain follows the concentration of nicotinic cholinergic receptors, suggesting direct actions on cell development (Slotkin, 1992). Nonetheless, nicotine can exert multiple effects on the maternal-fetal unit, including alterations in cardiovascular or hormonal status, rendering it difficult to label nicotine conclusively as a neuroteratogen. However, in vitro systems have the potential to demonstrate a direct effect of nicotine on neurodevelopment. Studies in cultured PC12 cells, a cloned cell line that developmentally resembles sympathetic neurons, indicate that nicotine can inhibit DNA synthesis and neurite outgrowth, but only at concentrations several orders of magnitude above those found in smokers (Chan and Quik, 1993; Song et al., 1998). Similarly, cultured rat embryos show gross morphological disruption by nicotine exposures at very high concentrations (1 mM), sufficiently high to evoke physicochemical membrane disruption and/or gross oxidative damage (Joschko et al., 1991). The nicotine-treated embryos also evince altered development of the neuroepithelium, potentially contributing to the elevated rate of neural tube defects seen with maternal smoking (Joschko et al., 1991), but given the excessive concentrations, it is not possible to determine whether these effects are relevant to clinical exposures. It is not uncommon that the effects of nicotine on cultured cells require exposure levels above those experienced in intact systems in order to demonstrate significant alterations of tightly controlled variables such as cell replication or differentiation, even when these are known to be relevant endpoints in vivo (Konno et al., 1991; Chan and Quik, 1993; Tipton and Dabbous, 1995; Song et al., 1998). Nevertheless, a definitive demonstration that nicotine specifically affects neurodevelopment requires an in vitro system where effects can be seen at concentrations below those necessary for gross dysmorphogenesis.

Our study investigates the effects of nicotine on brain development at the stage of neurulation in cultured rat embryos, using well-established designs (Joschko et al., 1991; Andrews et al., 1997). The mammalian central nervous system develops from a pseudostratified epithelium with highly specified patterns of mitosis and cell migration (Sauer, 1935; Fujita, 1960; Adams, 1996). During this process, it is rare to find apoptotic cells within the neuroepithelium except in highly specific locations. Accordingly, the early neuroepithelium can provide a sensitive test system for adverse effects of nicotine exposure that incorporates targeting of mitosis, cell migration and laminar organization, and cell death. Accordingly, we have examined the effects of a 48-hr exposure to nicotine in cultured rat embryos from embryonic day 9.5 through 11.5, a period corresponding to major architectural change in the neuroepithelium, and to the emergence of nicotinic cholinergic receptors (Naeff et al., 1992). A key part of our hypothesis is that, if nicotine is a specific neuroteratogen, effects will be demonstrable below the threshold for general embryonic dysmorphogenesis. Thus, we have used a range of concentrations well below that shown previously to elicit gross developmental abnormalities (Joschko et al., 1991). We have then examined targeting of cell development by assessing pyknotic and mitotic profiles as well as other evidence of cell damage.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals. All studies using animals were carried out in accordance with the declaration of Helsinki and with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health. Sprague-Dawley rats (Charles River Laboratories, Raleigh, NC) were housed in polypropylene breeding cages with heat-treated pine savings, and supplied with standard rat chow and water ad libitum, in a temperature-controlled facility with a 12-hr light:dark cycle. Each cage contained one male and two females. Vaginal swabs were examined and the sperm-positive date was taken as embryonic day zero. At embryonic day 9.5, the pregnant rats were anesthetized with ether and uteri were removed to prepare the embryos for culture.

Whole embryo culture and teratology screening. The culture procedure followed the protocols established in previous work (Andrews et al., 1997). Excised uteri were dipped immediately in 70% ethyl alcohol and 2× HBSS solution, and were placed in Waymouth's medium. Embryos, along with the decidual tissue, were removed from the uterus. The conceptus was further dissected from the decidual tissue and Reichert membrane under a stereomicroscope. Groups of four embryos were placed in 25-ml culture flasks containing 5 ml of culture medium [medium prepared from male rat serum preserved at 20°C for not more than 1 mo; heat inactivated at 55°C for 30 min, sterilized by filtration, supplemented with 50 U/ml of penicillin G and 50 µg/ml of streptomycin]. The medium was equilibrated for 3 min with 5% O₂, 5% CO₂, 90% N₂ and the flasks were placed on a rocker platform (19 oscillations/min) and incubated at 37°C. To initiate treatment, nicotine (Sigma Chemical Co., St. Louis, MO) was dissolved in dimethylsulfoxide and was added to achieve final concentrations of 1, 10 and 100 µM. An equivalent volume of dimethylsulfoxide was added to the control group. After 24 and 42 hr in culture, the medium was reequilibrated with 40% O₂, 5% CO₂, 55% N₂ for 2 min. After 48 hr of total culture time, embryos were examined under a dissecting microscope for viability by the presence of yolk sac circulation and heart beat, and for dysmorphogenesis, after which they were dissected from the membranes and processed for light microscopy. A standardized morphological scoring system (Brown and Fabro, 1981) was used to assess embryonic growth parameters as well as developmental stages.

Tissue processing and quantitative analysis. Embryos were fixed in Karnovsky's fixative, dehydrated in ascending concentrations of ethanol and embedded in media containing LX-112 resin and (2-dodecenyl-1-yl)succinic anhydride. The embryos were placed longitudinally along the long axis of plastic molds and were blocked in the same resin, sliced into 1-µm transverse sections with a Reichert Ultramicrotome, stained with 1% toluidine blue, and examined and photographed with a Zeiss Axiophot microscope.

View larger version (152K): [in this window] [in a new window]   Fig. 1.   a, Coronal section of control rat embryo, showing neural tube regions studied: alar plates of forebrain (FB), midbrain (MB) and isthmal region of hindbrain (HB); these regions of the neural tube give rise to cerebral cortex, midbrain and pontine region of the adult brain, respectively. Inset shows a camera lucida drawing of a sagittal section, with horizontal lines indicating portions of the neural tube used for quantitation. Subsequent pictures show portions of the neuroepithelium from similar regions of the neural tube. Scale bar = 250 µm. b, Neuroepithelium from a control embryo, showing closely apposed pseudostratified cells at different phases of mitosis, and their processes. The mitotic zone, which contains number of mitotic figures (MF), is located toward the lumen of the neural tube. Scale bar = 20 µm. c, Neuroepithelium from an embryo exposed to 100 µM nicotine, showing massive cell death in the hindbrain region alar plate in the region of neural tube fusion. Scale bar = 100 µm. d, Higher magnification of the necrotic zone inside the box of (c). Dying cells appear at various stages along with their debris, in the form of intra- and extracellular bodies, often engulfed by healthy cells, including those undergoing mitosis (M). Arrowheads and arrows represented pyknotic bodies and dying cells. A large nucleated phagosome (P), containing multiple dark bodies can be seen inside the epithelium. Scale bar = 20 µm.

	Results

Top Abstract Introduction Methods Results Discussion References

Growth and morphological development. After 48 hr in culture, the control embryos had completed their neural tube fusion and displayed all the characteristic features of an 11.5-day embryo (data not shown): a well formed optic cup, optic placode and first and second visceral arches; an S-shaped heart tube with structures corresponding to the atrium, a common ventricle and bulbus cordis; and in the distal part of the heart tube, a dorsal aorta connected by arch arteries. Both fore and hind limb buds were distinct and 27 to 28 somites were visible. Only one of a total of 47 control embryos exhibited a structural anomaly (table 1). Similar results were obtained in nearly all the embryos in the nicotine exposure groups, which also showed no growth retardation. There was no significant increase in the incidence of dysmorphogenesis (Fisher's exact test), whether considered across all three nicotine groups taken together, or as individual treatment groups compared to the control group.

Neuroepithelial development. In control embryos, gross examination revealed appropriate fusion of the neural folds and a thick neuroepithelium consisting of closely apposed pseudostratified epithelium. The germinal matrix of the cerebral cortex was composed of a single zone, with the great majority of mitotic cells situated in a periventricular position (fig. 1b). The mesenchyme around the neuroepithelium contained stellate shaped cells that made contact with neighboring cells by means of cytoplasmic processes and developing small blood vessels. The nuclei of the interphase cells contained clumped, inactive heterochromatin and nucleoli with a dense granular appearance. Only occasionally, there were cells with morphological features characteristic of cell death, scattered among the far more numerous, healthy cells. These were characterized by intense staining of condensed or fragmented particles. Some of the healthy cells showed darkly stained small areas, fragments of dead cells engulfed by neuroepithelial cells.

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Quantitation of pyknotic cells in the neuroepithelium. Data represent means and S.E. obtained from four embryos in each treatment group, determined from values averaged across 10 sections (repeated measure) in each embryo, and shown as number of cells per 100 µm segment of neuroepithelium. ANOVA across all regions appears within the panel, along with subdivision by dose and region. Asterisks denote individual values that differ from the corresponding control. Healthy cells containing engulfed debris from pyknotic cells were not included in the count.

Morphology and fate of dying cells. The dying cells recognizable at the light microscopic level were marked by a darkly stained nucleus, sometimes with condensation of chromatin at one side of the nucleus, but with the other cellular constituents generally appearing relatively normal. Darkly stained (pyknotic) spherical granules of various sizes, and spherical vesicles with a dense core were also common (fig. 3, a and b). Most of the debris was found within the cytoplasm of otherwise healthy-looking cells, with few cells showing signs of cytoplasmic degeneration, suggesting a rapid engulfment of the dead cell cytoplasm. Large pyknotic granules were usually found very close to the host nucleus, often partially embedded in the healthy nucleus; smaller pyknotic granules were also found but these tended to be separated from the nucleus. Taken together, these observations indicate that "healthy" nondying cells may engulf the fragments of adjacent degenerated cells. Accordingly, the initial stages of neuroepithelial apoptosis, which were rarely observed in the embryos after 48 hr of nicotine exposure, are unlikely to be detected because of the subsequent degeneration and engulfment of debris by adjacent cells.

View larger version (151K): [in this window] [in a new window]   Fig. 3.   a, Pyknotic cells in the forebrain region of the neural tube from an embryo exposed to 100 µM nicotine. Arrowheads indicate pyknotic cells and arrows indicate cells at earlier phases of degeneration. Note the extensive intercellular spaces in the upper part of the neuroepithelium. Scale bar = 20 µm. b, Hindbrain brain neuroepithelium showing phagosomes with dark included bodies (P). Frequently, pyknotic bodies (arrowheads) and degenerating cells (arrow) are present, along with a large number of mitotic figures (MF) in the germinal zone. Scale bar = 20 µm. c, Forebrain neuroepithelium from an embryo exposed to 10 µM nicotine. Note the extensive cytoplasmic vacuolation. Pyknotic bodies are seen even in the mitotic zone (M). Scale bar = 20 µm. d, Hindbrain neuroepithelium from an embryo exposed to 10 µM nicotine. Vacuoles (arrowheads) are present in almost all cells, including those undergoing mitosis (M). Scale bar = 20 µm.

Effects on mitosis. In examining the regional distribution of pyknotic cells, we noted that, curiously, cell death was also present in the mitotic zone (fig. 3c). Darkly stained granules, indicative of engulfed pyknotic material were readily detectable in mitotic cells (fig. 1d), indicating that some of the damage to neuroepithelial cells is likely to involve the zone of proliferation. Quantitation of mitotic figures within each zone showed a strong, dose-dependent effect of nicotine on mitosis, but surprisingly, nicotine enhanced the number of mitotic figures (fig. 4). The effect differed from that seen for cell death in three respects. First, the lowest concentration of nicotine did not evoke significant changes in two of the three regions. Second, regional selectivity was not readily apparent for the mitogenic effect of nicotine, as equivalent stimulation was seen in all three regions (no significant interaction of treatment × region). Third, the effect was of much smaller magnitude, with 25 to 50% changes at the highest nicotine concentration, as opposed to 3-fold increases in cell death.

View larger version (57K): [in this window] [in a new window]   Fig. 4.   Quantitation of mitotic cells in the neuroepithelium. Data represent means and standard errors obtained from four embryos in each treatment group, determined from values averaged across 10 sections (repeated measure) in each embryo, and shown as number of cells per 100-µm segment of neuroepithelium. ANOVA across all regions appears shown within the panel, along with subdivision by dose and region. Asterisks denote individual values that differ from the corresponding control. Comparing the relative effects on mitosis vs. pyknosis, ANOVA indicates a main effect of treatment across both measures (p < .0001) but a selectively greater pyknotic effect as compared to mitotic effect (interaction of treatment × variable, p < .0001).

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our results indicate that, in concentrations below the threshold for general dysmorphogenesis, nicotine damages the developing neuroepithelium, findings that provide a basic mechanism for the long-term alterations in cell number, synaptic development and function, and behavior seen with prenatal nicotine exposure in vivo (Slotkin, 1992). Clearly, in extreme cases, such changes can also contribute to the minor increase in the incidence of neural tube defects seen in the offspring of smokers (Evans et al., 1979). Most prominently, we found evidence of concentration-dependent apoptosis and cell loss throughout the neuroepithelium, characterized by extensive vacuolation, numerous pyknotic cells and the presence of apoptotic condensations. As a result of extensive cell loss, intercellular spaces were expanded in the areas of maximal damage. It should be emphasized that these changes are selective for the neuroepithelium, as general developmental markers were unaffected even at the highest concentration of nicotine. Previous work with nicotine concentrations several orders of magnitude above those used here also found extensive neuroepithelial cell death in association with dysmorphogenesis (Joschko et al., 1991) but the two effects obviously can be differentiated from each other, with neuroepithelial cells targeted at otherwise subteratogenic concentrations. The ability of nicotine to elicit cell loss in the developing brain in vivo continues into late gestation and even into the postnatal period, as evidenced by persistently elevated expression of genes involved in apoptosis (Slotkin et al., 1997) and by further cell loss occurring after the termination of nicotine exposure (Slotkin et al., 1987b; Slotkin, 1992).

A critical issue addressed by our findings is the underlying mechanism by which nicotine affects brain development. In vivo models of prenatal nicotine exposure are invariably confounded by the multiple effects of the drug on the maternal-fetal unit, including alterations in fetal nutritional status and cardiovascular function, largely attributable to the vasoconstrictor actions of nicotine (Slotkin, 1992). However, the use of an in vitro system obviates the participation of uteroplacental circulation; furthermore, within the embryo itself at this developmental stage, the vasculature is neither adequately developed nor functionally innervated, so that nicotine cannot evoke altered perfusion. The yolk sac is the main nutritive organ during organogenesis and we found no abnormalities in this structure in any of the nicotine-treated embryos. Indeed, instead of a generalized insult to the developing embryo, nicotine appears to target selective cell populations within the neuroepithelium, as evidenced by the regional selectivity of cell death (hindbrain > forebrain > midbrain). Although it is not possible at this time to demarcate the reasons for the vulnerability of particular cells, the potential role of the specific cellular elements with which nicotine interacts, nicotinic cholinergic receptors, stands out prominently. Nicotinic receptors are first detectable by the end of the period of nicotine exposure used here (Naeff et al., 1992). In the fetal brain, nicotine induces the formation of its own receptors (Slotkin et al., 1987a), so it is possible that the receptors are present in even higher concentration in the nicotine group. Studies conducted at later stages of development have shown that the regional selectivity for nicotine-induced apoptosis and cell loss are determined primarily by the nicotinic receptor concentration and that receptor stimulation evokes immediate evidence of cell damage (Slotkin et al., 1987a; Smith et al., 1991; Slotkin, 1992). Similarly, in vivo exposures at developmental stages before the emergence of the receptors do not evoke signs of cell damage or apoptosis (Slotkin et al., 1993, 1997), nor does nicotine elicit mitotic abnormalities in neuronal cell lines that lack nicotinic receptors (Song et al., 1998). With embryonic exposure, we found the greatest concentration of pyknotic cells in the hindbrain, the region in which receptors emerge first (Naeff et al., 1992). Furthermore, we found lower concentrations of pyknotic cells in the mitotic zones, the areas showing lowest receptor concentrations (Naeff et al., 1992). The similarities of distributions of receptors and pyknotic cells, along with the ability of low concentrations of nicotine to elicit cell damage, all suggest that the emergence and distribution of nicotinic receptors are critical in determining the pattern of neuroepithelial damage. A definitive proof of this hypothesis will require detailed receptor mapping in both control and nicotine-exposed embryos. However, earlier work indicates that the point at which nicotinic receptors can be detected and quantified may lag significantly behind the actual emergence of the receptor-positive phenotype (Naeff et al., 1992; Broide et al., 1995; Ostermann et al., 1995). Unfortunately, a simpler, alternative approach using nicotinic receptor blocking agents does not provide an unequivocal tool with which to demonstrate receptor-mediated mechanisms of nicotine on neuroepithelial development: cholinergic input is required for proper assembly of cortical cytoarchitecture (Hohmann et al., 1988; Navarro et al., 1989; Slotkin, 1992) and cholinergic antagonists themselves interfere with cell replication in the developing brain (Whitney et al., 1995). A receptor-mediated mechanism does not rule out additional participation of other processes; e.g., the E11.5 embryo possesses metabolic enzymes capable of forming cotinine, which is also found in the fetus with maternal smoking. To date, no information is available concerning the potential developmental neurotoxicity of cotinine or other nicotine metabolites.

Although pyknotic figures were less prominent in the ventricular zone, where neurons and glia are generated, we nevertheless found pyknotic cells in close association with the mitotic region of the alar plate; there was also a significant number of mitotic cells with engulfed remnants of damaged cells, implying either that mitotic cells are among those being damaged, or that cells immediately surrounding the mitotic zone are targets for nicotine. However, the cells in this region also displayed an elevated rate of mitosis. Obviously, given the enhancement of cell death, the induction of mitotic figures by nicotine could represent compensatory mitogenesis. However, two features argue against this explanation. First, the concentration-response relationships differ for cell death and mitogenesis: cell death is enhanced 3-fold, whereas mitosis is stimulated only 25 to 50% by nicotine exposure; cell death continues to increase as the concentration is raised above 10 µM, whereas mitosis is stimulated maximally by 10 µM nicotine. Perhaps more importantly, however, the regional selectivity seen for cell death (hindbrain > forebrain > midbrain) is not shared by enhanced mitosis, which instead shows a virtually identical stimulation in all three regions. These data suggest that nicotine exerts two distinct effects, one evoking cell death, and the other evoking mitosis, with two separable concentration relationships and discrete cell targets. Indeed, there is ample prior evidence for nicotine to act as a direct mitogen in a variety of isolated cell lines (Quik et al., 1994; Maritz and Thomas, 1995), with totally independent actions mediating cell damage or death (Konno et al., 1991; Tomek et al., 1994; Tipton and Dabbous, 1995). For the developing brain, it is particularly perplexing that, in late gestation or after birth, nicotine elicits mitotic arrest through the very same nicotinic receptors that we presume mediate promotional effects on cell division in the neural tube stage (Slotkin, 1992). The actions of nicotine are thus very strongly dependent on the developmental context in which exposure occurs. This is not unexpected, as neurotransmitters, acting as trophic substances, typically elicit biphasic developmental actions, promoting cell division at early stages but fostering the switch from replication to differentiation later in development (Buznikov et al., 1970; Lauder, 1985; Slotkin et al., 1987c, 1988a,b; Whitaker-Azmitia, 1991). Accordingly, acetylcholine and by extension nicotine, both influence cell formation and cytoarchitectural organization of the brain through a combination of promotional and inhibitory actions (Hohmann et al., 1988; Navarro et al., 1989; Slotkin, 1992). The difference is that, in the setting of maternal smoking or nicotine administration, nicotine exposure elicits the effects prematurely and with excessive intensity relative to the normal developmental signal. Such overstimulation is likely to enhance and discoordinate naturally occurring processes in the modeling of the developing brain, including mitosis and apoptosis. Regardless of the actual mechanism underlying enhanced mitotic figures in the nicotine-exposed embryos, it is evident that this effect does not offset the much larger cell loss caused by nicotine, as shown by the much greater magnitude of cell death found here and by the eventual deficits in the total number of cortical cells seen with in vivo exposure models (Slotkin et al., 1987b; Slotkin, 1992). Indeed, accumulated mitotic figures could even represent mitotic arrest at S-phase, rather than enhanced cell division.

In our study, we were able to detect significant effects of nicotine on neuroepithelial development at 1 µM, a concentration well below that for dysmorphogenesis (Joschko et al., 1991) or for cell damage in culture systems other than developing embryos (Konno et al., 1991; Tomek et al., 1994; Tipton and Dabbous, 1995). The average venous plasma concentration of nicotine in smokers or transdermal patch users is typically only 0.1 to 0.3 µM (Lichtensteiger et al., 1988), but arterial peak values in smokers are two to three times higher. Effects obtained in rat embryos at 1 µM nicotine thus represent exposures well within the 10-fold intraspecies extrapolation uncertainty factor adopted as a standard for extending the No Observable Adverse Effect Level from rats to humans (Faustman and Omenn, 1996). A further consideration is the fact that human exposures occupy the entire period of gestation and involve additive or synergistic effects with other tobacco components, whereas we examined a brief exposure period to nicotine alone; these factors are likely to enhance the probability of fetal neural damage with maternal smoking.

There are important ramifications if these findings can be extended to human pregnancy. First, the use of an in vitro model to identify nicotine itself as a neuroteratogen provides a specific mechanism connecting maternal smoking with adverse neurobehavioral outcomes, disproving once and for all that the ancillary factors in smoking are solely responsible for the damage. Second, the direct role of nicotine implies that caution must be exercised in the utilization of nicotine replacement therapy for smoking cessation. Issues such as total dose and delivery formulation are likely to be critical in avoiding adverse effects. Although, with acute exposures, the placental barrier can be expected to protect the fetus from nicotine to some extent, smokers tend to maintain a steady-state plasma level that enables equilibration of the fetal compartment with maternal plasma. Similarly, continuous exposure paradigms such as transdermal nicotine patches in man, or osmotic minipump infusions in rats, maintain a constant drug level in mother and fetus (Lichtensteiger et al., 1988), and the exposure paradigm used in our study is obviously most akin to continuous infusions. Thus, in recommending nicotine replacement for smoking cessation in pregnancy, continuous exposure paradigms (transdermal patch) may be less desirable than episodic dose delivery (nicotine-containing chewing gum, nicotine inhaler). At the very least, removal of the patch at night-time should allow for a daily decay of plasma levels as is seen with smoking; although we did not model fluctuating exposures in rat embryos, it would be worthwhile to determine if episodic recovery periods could reduce the net adverse effects of nicotine on brain development. Finally, if nicotinic receptors underlie the loss of neural cells, then the ontogenetic emergence of these receptors can be used to identify the critical period and specific sites for which damage is most probable.