Enduring effects of perinatal nicotine exposure on murine sleep in adulthood

Abstract

The long-term consequences of early life nicotine exposure are poorly defined. Approximately 8–10% of women report smoking during pregnancy, and this may promote aberrant development in the offspring. To this end, we investigated potential enduring effects of perinatal nicotine exposure on murine sleep and affective behaviors in adulthood (~13–15 wk of age) in C57Bl6j mice. Mothers received a water bottle containing 200 µg/ml nicotine bitartrate dihydrate in 2% wt/vol saccharin or pH-matched 2% saccharin with 0.2% (vol/vol) tartaric acid throughout pregnancy and before weaning. Upon reaching adulthood, offspring were tested in the open field and elevated plus maze, as well as the forced swim and sucrose anhedonia tests. Nicotine-exposed male (but not female) mice had reduced mobility in the open field, but no differences were observed in anxiety-like or depressive-like responses. Upon observing this male-specific phenotype, we further assessed sleep-wake states via wireless EEG/EMG telemetry. Following baseline recording, we assessed whether mice exposed to nicotine altered their homeostatic response to 5 h of total sleep deprivation and whether nicotine influenced responses to a powerful somnogen [i.e., lipopolysaccharides (LPS)]. Males exposed to perinatal nicotine decreased the percent time spent awake and increased time in non-rapid eye movement (NREM) sleep, without changes to REM sleep. Nicotine-exposed males also displayed exaggerated responses (increased time asleep and NREM spectral power) to sleep deprivation. Nicotine-exposed animals additionally had blunted EEG slow-wave responses to LPS administration. Together, our data suggest that perinatal nicotine exposure has long-lasting effects on normal sleep and homeostatic sleep processes into adulthood.

smoking during pregnancy is a declining but still pervasive trend. Indeed, a recent study reported that ~8.4% of women continue to smoke during pregnancy, although many of them (~20.6%) reduce their cigarette consumption by the third trimester (10). Additionally, ~25% of all people in the US show evidence of second-hand smoke exposure (22). Smoking during pregnancy has long-term consequences for the child, specifically psychiatric and social problems manifesting later in life (14). In mouse models of early-life nicotine, cigarette smoke, or e-cigarette vapor exposure, developmental abnormalities in locomotor behavior (39), depressive-like behavior (43), and cognitive function (2) have been observed extending into adolescence and adulthood. These changes may be due to deregulated cholinergic signaling during development, as nicotine is a powerful neuromodulator acting through nicotinic acetylcholine receptors (nAChRs), with off-target actions on dopaminergic transmission (7), hypocretin/orexin signaling (36), the circadian clock (9, 32), and others.

Adequate sleep and consolidated circadian rhythms play a major role in health. Indeed, poor sleep and disrupted circadian timing are associated with the development of metabolic and psychiatric disorders (23, 30). Compared with cocaine, opioids, cannabis, and alcohol use, perinatal nicotine exposure has the strongest correlation with childhood sleep problems (40). However, it is unknown whether perinatal nicotine exposure causes altered sleep in adulthood and whether this is associated with other neuropsychiatric problems. Because many populations of neurons that regulate sleep-wake states are modulated by nicotine, we hypothesized that developmental nicotine exposure confers long-lasting effects on normal sleep-wake behavior in C57BL6j mice. Specifically, we predicted that mice exposed to perinatal nicotine would demonstrate altered depressive- and anxiety-like behavior and have increased and fragmented sleep and that their responses to a homeostatic sleep and immune challenge would be exaggerated. To investigate this hypothesis, we provided nicotine (or vehicle solution) in the drinking water ad libitum to pregnant dams and then to their offspring until weaning at 21 days postbirth. Offspring were then provided ad libitum access to filtered drinking water until tissue collection in adulthood following behavioral and sleep phenotyping. Therefore, we restricted nicotine exposure to the perinatal period of conception, gestation, and preweaning (6 wk total). This protocol has been used previously to investigate how developmental exposure to nicotine alters neuronal function and behavior later in life in C57BL6j mice (25).

METHODS

Animals and perinatal nicotine administration.

Male and female C57Bl6j mice were used throughout this study. Upon arriving at our laboratory, adult (> 8 wk old) male and female mice were group housed and allowed to acclimate to a 14:10-h light-dark cycle for 1 wk before pairing. Mice were paired (5 veh, 6 pNic pairs) and allowed to mate for a 10-day period after which the male was removed from the cage. On the first day of pairing, mice received a water bottle containing 200 µg/ml nicotine bitartrate dihydrate (Nic; calculated as free base) tartaric acid (Veh; Sigma-Aldrich, St. Louis, MO), with all solutions pH matched at 7.4 (as in Ref. 25). Solutions were refrigerated at 4°C until use, and breeding pairs received fresh solution every 2 wk. Mice had ad libitum access to one of these solutions, food (Harlan Teklad no. 7912), a cotton nestlet, and a plastic piece of housing enrichment throughout pregnancy and before weaning. At weaning, offspring were sexed, separated, and solutions were immediately replaced with normal filtered drinking water. The experimental design described above is detailed in Fig. 1. Groups numbers were as follows: males, n = 15 veh, 23 pNic; females, n = 13 veh, 16 pNic for behavioral testing; and n = 8 veh, 7 pNic males for sleep measures. All procedures were approved by The Ohio State Institutional Animal Care and Use Committee.

Fig. 1.

Experimental design. A: adult female and male C57BL6j mice were paired for 10 days to allow multiple mating opportunities. At this time, water bottles were replaced with ones containing 200 µg/ml nicotine bitartrate dihydrate in 2% wt/vol saccharin (nic) or pH-matched 2% saccharin with 0.2% tartaric acid (vol/vol) (veh). Treatment continued until weaning (total 6-wk exposure). The stud male was removed 10 days postpairing. At weaning (postnatal day 21), mother serum was collected to assess circulating cotinine concentrations, and male and female offspring were separated and singly housed. Treatment was stopped at this time and mice were given ad libitum access to normal drinking water. Seven to ten weeks later, mice underwent behavioral testing, and then a randomly chosen subset of males was chosen for EEG/EMG sleep analyses. B: nicotine-exposed mothers had higher cotinine levels in their serum at weaning, indicating that the mice were indeed being exposed to significant levels of nicotine (t₉ = 47.16, ***P < 0.0001, Student’s two-tailed t-test; n = 5 veh, 6 Nic mothers, error bars represent SE).

Cotinine assay.

Cotinine is the primary active (i.e., capable of binding nAChRs) metabolite of nicotine and has a half-life of ~55 min in mice (45). Therefore, assaying concentrations of cotinine in a biological sample provides a more temporally stable assay of nicotine exposure than nicotine, which has a half-life of only 6–7 min in mice (29). We measured maternal serum cotinine concentrations at weaning as a proxy for offspring nicotine exposure using a commercially available enzyme immunoassay kit (Calbiotech mouse/rat cotinine ELISA). Serum samples were run (in duplicate) neat and read on a plate reader (Molecular Devices, SpectraMax Plus), and cotinine concentrations were quantified via comparing unknown optical densities against a standard curve (4-parameter logistic). The intra-assay coefficient of variation was 5.18%, and the lower limit of detectability was 0.01 ng/ml, with one sample (from a veh mouse) falling below this range.

Behavioral testing.

To assess enduring behavioral effects of perinatal nicotine exposure, adult male and female offspring (≥ 9 wk of age) were subjected to several tests of anxiety- and depressive-like behavior. The following tests were completed in order: open field arena (OF; anxiety, locomotion), elevated plus maze (EPM; anxiety), forced swim test (FST; depression), and sucrose anhedonia (depression). For the EPM and FST, the light levels in the testing room were ~430 lux (broad spectrum compact fluorescent), and the light in the open field chamber was ~25 lux (measured using Traceable 3251 dual-range light meter). OF and EPM tests were completed during the same week, but FSTs occurred 1 wk later and sucrose anhedonia at least 1 wk following that to ensure any stressful effects of testing were resolved at that time. OF, EPM, and FST occurred between zeitgeber time (ZT) 6 and ZT 10 to reduce any circadian-phase related differences in behavior. Mice were allowed 30 min to acclimate to the testing room before running the assays. Males and females were tested separately, and if both sexes were tested on the same day, then the testing room was thoroughly cleaned and mice were introduced to the testing room at least an hour after the first sex was removed. Experimenters left the room during every trial.

The open field chamber consisted of a 40 × 40 cm transparent acrylic box flanked by stacked grids of intersecting infrared beam emitter/detectors (as in Ref. 6). The bottom grid detected motion in the X/Y direction while a second (raised) grid detected rearing behavior. Mice were placed into the open field (which was loosely covered with fresh bedding material) for 10 min, and data were segmented into 2-min bins from the start time for analyses. Each chamber was cleaned with 70% EtOH followed by soapy water between each mouse to limit olfactory carry-over between trials. Total locomotor activity and central tendency were scored automatically using Photobeam Activity System software (San Diego Instruments, San Diego, CA). Males were tested on a separate day from females to prevent olfactory cues from altering locomotor behavior. More time spent in the center of the arena indicates a less anxious phenotype.

The EPM test consisted of putting mice onto a raised platform of perpendicular open and closed arms (raised ~1.5 m off the ground) for 5 min. Mice were always placed in the center facing the closed arm to start the test, and the whole apparatus was cleaned with 70% EtOH followed by soapy water after each trial. The following variables were scored via videotape analysis by a condition-blind observer using Observer XT 8.0 software (Noldus Information Technology, Leesburg, VA): time in center (time to enter either arm), time spent in open arms, time spent in closed arms, number of closed, and number of open arm entries. Upon entering a closed or open arm, mice were scored as remaining on that arm until they entered (>one-half body in different arm) a different arm of the maze. Increased time spent in the open arms indicates a less anxious phenotype.

Mice were tested for behavioral despair in the FST. Mice were placed into a clear 3,000- m container filled with ~2,000 ml room temperature water (~22°C, measured by thermometer) for 6 min, during which their swimming and floating behavior was monitored via videotape. After the test, mice were removed, patted dry, and returned to their home cage. Containers were emptied and cleaned between each test. Videotapes were analyzed for time spent swimming (i.e., vigorous swimming or scratching at the sides of container) and time spent immobile (i.e., minimal movement or only movements necessary to keep afloat). Increased immobility time is interpreted as a learned-helplessness behavior indicating a depressive-like response (4).

Sucrose anhedonia was completed during the first 6 h of the active phase (i.e., ZT 14–20). Fifteen-milliliter water bottles were filled with a 2% sucrose solution or normal drinking water, weighed, and presented during this time to all mice for 2 consecutive days (on the second day, water bottles were placed on the side opposite to the position on the first day to control for a side bias). Mice did not have access to food at this time. After the 6-h test, the water bottles were removed under dim red light (≤5 lux) and weighed again to assess amount consumed. Preference for the sucrose solution was calculated by comparing the amount of sucrose drank against the total amount of liquid drank. Reduced sucrose preference indicates a depressive-like phenotype (34). After the final behavioral task, a subset of male mice (randomly chosen, equally distributed across treatment groups) were fitted with EEG/EMG biotelemeters for sleep analyses (described below).

Telemeter implantation.

F20-EET wireless transmitters [Data Sciences International (DSI), St. Paul, MN] were implanted as previously described previously (5) by an experienced surgeon using aseptic technique (N. Zhang). Mice were deeply anesthetized with isoflurane vapors (3% induction, 1.5% maintenance) and affixed into a stereotaxic apparatus. A midline incision from the posterior margin of the eyes to midway between the scapulae was made. The skull was exposed and cleaned, and two stainless steel screws (00–96 × 1/16; Plastics One, Roanoke VA) that would serve as cortical electrodes were inserted through the skull to contact the dura mater. One screw was situated 1 mm anterior to bregma and 1 mm lateral from the sagittal suture, while the second screw was placed 2 mm posterior to bregma and 2 mm contralateral from the sagittal suture. The transmitter itself was inserted into as subcutaneous pocket along the back of the animal, and another set of leads was inserted into the cervical trapezius muscles for EMG measurement. After the screws were secured with dental cement, the animal was sutured and administered buprenorphine (0.05 mg/kg sc) to provide postoperative analgesia. Postoperative warmth was provided by heating pad; animals were monitored until mobile and then returned to the recording cabinet. A further 5 days of analgesia was given via ibuprofen administration in the drinking water. Sleep recordings did not start until 2 wk following transmitter implantation to ensure that the trauma of surgery did not itself alter sleep. EEG/EMG data, temperature, and activity data were collected into a computer running Ponemah Software v6.3 (DSI).

Sleep deprivation and LPS administration.

After 2 days of baseline sleep recording, mice were subjected to a total sleep deprivation protocol (via gentle handling) for 5 h beginning at the start of the inactive phase (ZT 0), when sleep pressure was high. Biopotentials and mice were monitored for signs of sleep, and if necessary, the experimenter would tap on the cage or disturb the nest to rouse the mouse. Mice were allowed 18 h of undisturbed recovery sleep following sleep deprivation. Thirty-two hours following sleep deprivation, at ZT 13 (1 h before lights off), mice were injected (intraperitoneally) with 0.5 mg/kg lipopolysaccharides (Escherichia coli serotype 0127:B8; Sigma) in PBS. Sleep was monitored for the following 24 h.

Sleep analyses.

Sleep was analyzed using NeuroScore Software v3.2 (DSI). In brief, raw biopotentials were imported, bandpass filtered (EEG: 0.3–25 Hz; EMG: 25–50 Hz), and analyzed in 10-s epochs by a blinded observer using the Neuroscore Mouse Sleep scoring module (signal processed as in Ref. 12). The delta band was set to 0.5–4 Hz while theta was set to 6–9 Hz. Wake was defined as having variable high-frequency EEG with high EMG activity; NREM sleep was defined as having low-frequency, high-voltage EEG with low EMG activity; REM sleep was characterized by predominant theta frequencies in the EEG (>2.5 theta/delta ratio) with low EMG activity. Two-hour binned data were used for analyses (i.e., averages of the preceding 2 h).

For spectral analyses, segments of baseline or recovery sleep following sleep deprivation were subjected to Fourier transformation (Hamming signal window; normalized to no. of samples in spectrum), and periodogram data were cleaned of artifacts and organized by vigilance state. Spectral power in each vigilance state was plotted from 0.5 to 25 Hz with a ~0.5-Hz frequency window. For data normalization, spectral power during baseline was subtracted from data generated following sleep deprivation at the same time of day.

Statistics.

Two-tailed t-tests were used to compare group means. Vigilance state day/night averages were compared using two-way ANOVAs. Pre- and post-LPS spectral comparisons were completed using a two-way ANOVA with time point (pre- or post-LPS) and treatment (veh vs. pNic) as independent variables followed by Sidak’s multiple comparisons test. Statistical analyses were completed using SPSS Version 23 (IBM, Armonk, NY) or GraphPad Prism 6 (GraphPad Software, La Jolla, CA). A P ≤ 0.05 was considered statistically significant.

RESULTS

To verify that mice exposed to nicotine perinatally received similar amounts of the drug, serum from the mothers was collected at the time of weaning and cotinine concentrations were determined. Maternal levels of cotinine were 0.2 ± 0.402 ng/ml for vehicle-treated and 11.08 ± 0.364 ng/ml for nicotine-treated mice (t = 47.16, P < 0.0001) (Fig. 1). These cotinine values indicate that mice were receiving a stable level of nicotine, and cotinine concentrations were consistent with previous reports of light-smoking or second-hand smoke during pregnancy (35, 44). Additionally, body mass trajectories, spleen masses, and body and tail lengths in adulthood were equivalent between groups, suggesting that pNic treatment did not impart gross developmental abnormalities (data not shown).

Behavioral phenotype in adulthood.

Male (but not female) mice that received nicotine perinatally had significantly reduced locomotor behavior in the OF test (Fig. 2). This was not associated with anxiety-like behavior, as central tendency in the open field was unchanged between groups. Additionally, no behavioral changes on the EPM were detected, indicating that perinatal nicotine exposure did not confer long-lasting changes in anxiety-like behavior in either male or female offspring (Fig. 3). No changes were detected in sucrose preference (2%) on either testing day or in floating behavior in the FST, indicating that perinatal nicotine did not confer changes in depressive-like behavior into adulthood (Fig. 3).

Fig. 2.

Males exposed to perinatal nicotine have reduced locomotion in a novel open field environment. A: males showed reduced total movement over the course of a 10-min open field test (10 min total: t₃₈ = 2.828, P = 0.0282); however, they did not show anxiety-like behavior as central tendency was unchanged between treatment groups. B: this locomotor deficit in a novel environment was not evident in female mice, which had no discernable change in the open field between treatment groups (males: n = 15 veh, 25 pNic, females: n = 16 veh, 13 pNic, error bars represent SE, **P < 0.01, *P < 0.05 Student’s two-tailed t-test).

Fig. 3.

Perinatal nicotine exposure does not alter behavior on the elevated plus maze, sucrose preference, or forced-swim tests in adulthood. A and D: total time spent in the open and closed arms of the elevated plus maze in the 5-min test (max denoted by dotted line), total number of entries into either open or closed arm, and percent time spent on the open arm. B and E: sucrose preferences on 2 consecutive days of testing. C and F: total swimming and float durations in the 6-min forced swim test (dotted line denotes maximum) (males, n = 15 veh, 23 pNic; females, n = 13 veh, 16 pNic; error bars represent SE).

Altered sleep in male mice exposed to perinatal nicotine.

After observing altered locomotor behavior in a novel open field environment, we chose to evaluate the sleep phenotype in male offspring. A randomly selected subset of male mice, equally distributed across treatment groups was chosen for these tests. After recovery from EEG/EMG telemeter implantation surgery (2 wk), the baseline phenotype was examined for 2 days. No changes were observed in locomotor activity or subcutaneous temperature values over time, and nicotine-exposed mice showed normal EEG traces (Fig. 4A). Mice exposed to perinatal nicotine reduced wakefulness and increased NREM sleep during the inactive phase (Fig. 4, D and E). Despite increased time spent in NREM sleep, the delta (0.5–4 Hz) component of NREM sleep, an index of prior sleep pressure, showed normal patterns of activity over time, and whole day analyses of vigilance-state specific EEG spectra were unchanged between groups (Fig. 5).

Fig. 4.

Perinatal nicotine exposure has long-lasting effects on sleep-wake states into adulthood in male mice. A: representative EEG/EMG traces from vehicle (left) and (right) perinatal nicotine-exposed (right) mice. No differences were detected in spontaneous locomotor activity (B) or subcutaneous temperature rhythms (C). However, changes in the amount of time spent (D) awake (two-way ANOVA main effect of time: F_3,39 = 69, P < 0.0001; main effect of treatment: F_1,13 = 8.349, P = 0.0127; day 1 t₁₃ = 2.406, P = 0.032; night 2 t₁₃ = 2.463, P = 0.029), in non-rapid eye movement (NREM) sleep (two-way ANOVA main effect of time: F_3,39 = 68.83, P < 0.0001; main effect of treatment: F_1,13 = 6.849, P = 0.0213; day 1 t₁₃ = 2.219, P = 0.045; night 2 t₁₃ = 2.518, P = 0.026) but not in REM sleep were observed (n = 7–8/group, error bars represent SE, *P < 0.05 by Student’s two-tailed t-test).

Fig. 5.

Perinatal nicotine does not alter NREM sleep microstructure or the EEG spectra within wakefulness, NREM or REM sleep. A: NREM delta power over time as a percent total EEG power within each 2-h bin. Power spectra of the 2 baseline days for wakefulness (B), NREM sleep (C), and REM sleep (D) were not different between groups (n = 8 veh, 7 pNic; error bars represent SE).

To further investigate the sleep homeostat, mice underwent a 5-h period of total sleep deprivation, and sleep time and NREM spectral components were analyzed during different portions of the recovery period (18 and 6 h, respectively). Mice exposed to perinatal nicotine slept more during the recovery period and had a stronger increase (compared with their own baseline ZT-matched recording period) in NREM spectral power, specifically in the low delta (0.5–1 Hz) range (Fig. 6). This suggests that perinatal nicotine altered the homeostatic response to sleep loss in adulthood.

Fig. 6.

Perinatal nicotine exposure alters the homeostatic response to sleep deprivation. A: representative EEG spectrograms from a vehicle and nicotine-treated mouse during and following 5-h sleep deprivation. Total wake time following sleep deprivation (B) and (C) total wake and sleep times during the 18-h recovery period (C) (wake t₁₃ = 2.251, P = 0.042; sleep t₁₃ = 2.248, P = 0.0426) normalized (to NREM power during the same time period of baseline recording) NREM spectral power during the first 6 h of recovery sleep (D) [zeitgeber time (ZT) 5–11; whole spectrum average t₁₀₀ = 2.635, P = 0.0097; n = 8 veh, 7 pNic, error bars represent SE, *P < 0.05, **P < 0.01 by Student’s two-tailed t-test].

To assess whether these mice showed normal EEG responses to a bacterial immune challenge, they were injected (intraperitoneally) with 0.5 m/kg LPS 1 h before the start of the active (dark) phase (~ZT 13), and EEG/EMG biopotentials were monitored for the following 24 h. Mice that did not receive perinatal nicotine showed normal increases in EEG delta frequencies during the hours following LPS administration, while those exposed to perinatal nicotine had a largely blunted response (Fig. 7).

Fig. 7.

Perinatal nicotine alters the EEG slow-wave response to intraperitoneal LPS. A: representative EEG spectrograms from a veh (top) and pNic (bottom) treated mouse during 10 h following intraperitoneal LPS (0.5 mg/kg) injection (denoted with arrow). In the veh-treated mouse, an obvious frequency decline into the delta (0.5–4 Hz) range is evident following LPS, where the same pattern is absent in the mouse give pNic. B: summed (10 min binned) data over time showing the delta band over time as a percent of total spectral power before and following LPS. C: pNic prevented LPS-induced increases in EEG delta power, as shown by trajectory plots pre- and post-LPS injection. A significant difference was detected in the veh mice but not in pNic-treated animals (two-way ANOVA main effect: of pre- vs. post-LPS F_1,13 = 8.798, P = 0.0109; Sidak’s multiple comparisons post hoc test: veh-pre vs. veh-post-LPS t = 3.446, adjusted P = 0.0087; n = 8 veh, 7 pNic, error bars represent SE, **P < 0.01).

DISCUSSION

Together, our data are consistent with an enduring effect of perinatal nicotine exposure on sleep into adulthood. Previous reports have detailed the trajectory of sleep abnormalities in nicotine-exposed rats during the early postnatal period (16), but these previous studies were never continued into adulthood to determine whether developmental compensation would occur. Indeed, a similar increase in NREM sleep during the postnatal period was reported in rats exposed to nicotine prenatally (16). This increased NREM sleep was associated with upregulation of specific nAChR mRNAs in brain regions regulating vigilance states, including the α₄- and α₇-subunits. In another study using various doses and administration regimens of nicotine, acute administration resulted in enhanced arousal and reduced sleep, while repeated dosing resulted in increased sleep (38). In addition, a parallel human polysomnography study on infants born to smoking mothers demonstrated reduced cortical arousal up to 5–6 mo following birth in comparison to infants born to nonsmoking mothers (37), which may highlight species similarities in the developmental response to nicotine. Similarly, prenatal nicotine exposure delayed arousal and waking in both sleeping human infants and sleeping lambs (15, 18)

Despite increased NREM sleep at baseline, and reduced locomotor activity in the open field, the spectral components of the EEG were similar between groups. It was only after a sleep-deprivation challenge that changes were observed in the quality and microstructure of sleep. Indeed, following sleep deprivation, mice exposed to perinatal nicotine showed enhanced NREM spectral power [specifically the low-delta range (0.5–1 Hz)] during the recovery period compared with vehicle-exposed mice (normalized to their own baseline data; Fig. 6). β₂-Containing nAChRs (which also often include the α₄-subunit) contribute to the arousal (and homeostatic rebounding) effects of acute nicotine exposure, as mice lacking β2 nAChRs do not show homeostatic responses to nicotine-induced arousal (26). Developmental exposure to nicotine could impart long-lasting changes in nAChR expression and thus signaling, which may contribute to how these mice differentially respond to a homeostatic sleep challenge.

To our knowledge, this is the first study to examine LPS-induced changes to the EEG in mice developmentally exposed to nicotine. These mice had a blunted slow-wave (i.e., delta; 0.5–4 Hz) response to intraperitoneal LPS administration compared with mice that did not receive nicotine during the perinatal period. This delta-frequency-enhancing effect of LPS is largely driven by proinflammatory cytokines (primarily IL-1β, TNF-α, and IL-6) produced in response to Toll-like receptor (specifically TLR-4) activation on circulating leukocytes (28). Therefore, the EEG-modulatory properties of LPS are directly dependent on the immune response. Nicotine is a well-known immunosuppressive agent and can inhibit IL-1β production in circulating monocytes (33), as well as alter splenocyte proliferation and cytokine production in a time- and dose-dependent manner (19). This immunosuppressive effect seems to be governed by the α₇-nAChR (42), and fundamentally dictates the cytokine response to immunogenic challenge. Recent work also indicates that IL-1β production can be reduced by nicotine acting through α₇-, α₉-, and/or α₁₀-containing nAChRs (21). It is likely that the differential responses to LPS that we observed were due to impaired or reduced cytokine production in mice developmentally exposed to nicotine, perhaps due to altered expression of specific nAChR subtypes. Indeed, gestational exposure to nicotine or nicotine + ethanol suppresses splenocyte responsiveness to the immunogenic molecules concanavalin A or LPS well into adulthood (3).

How could developmental nicotine exposure alter sleep-wake states in adulthood? Many neural populations that contribute to the initiation, stabilization, and structural organization of vigilance state switching are directly modulated by nicotine or are subject to secondary effects of the drug (11). Indeed, nAChRs are widely distributed throughout the brain and are expressed on wake-modulating neural populations such as cholinergic neurons in the basal forebrain (27), hypocretin/orexin neurons (α₄β₂- and α₇-nAChRs) in the lateral hypothalamus (17, 24, 46), and dopaminergic neurons in the ventral tegmental area (8). Aberrant activation of these receptors during critical developmental windows could impart long-lasting changes on the excitability of these neurons or alter their response to endogenous cholinergic signaling later in life. Developmental nicotine exposure produces changes in dendritic structure that are regulated by epigenetic modifications (25). Subsequent changes in synaptic function due to prenatal nicotine exposure may also provide a potential mechanism for altered circuits regulating sleep. Additionally, pre- and early postnatal exposure to nicotine alters central chemoreception pathways, which play a powerful role in the regulation of breathing (13). Many neuronal populations that play a role in chemosensation also regulate arousal and sleep, providing a link among nicotine exposure, ventilation deficits, and sleep abnormalities (20). An alternative explanation involves the effects of nicotine on maternal behavior, as saline-exposed offspring that were cross-fostered with nicotine-exposed mothers show behavioral changes later in life that are likely due to changes in maternal care (31, 41). Indeed, altered maternal behavior is associated with long-lasting effects on sleep in human infants (1).

Perspectives and Significance

In summary, our data demonstrate that perinatal nicotine has enduring effects on murine sleep extending into adulthood. Specifically, our data indicate that mice exposed to perinatal nicotine were in a state of hypersomnia, showed enhanced responses to a homeostatic sleep challenge, and blunted their EEG responses to the immunogenic somnogen LPS as adults. The causes of these changes may reflect persistent alterations to nAChR signaling during development, altered immune system development and homeostatic control of sleep, or an altered maternal care environment. A limitation of this study is the lack of sleep data in adult female mice exposed to perinatal nicotine. We chose to focus on males for these analyses given the initial behavioral phenotype observed in the OF arena (Fig. 2). Future studies should aim to include females in subsequent sleep analyses to assess whether these results are generalizable to both sexes or whether a sexual dimorphism in these traits becomes evident. Although beyond the scope of this study, future experiments should tease apart these mechanisms to gain a holistic understanding of the factors that contribute to long-lasting changes in sleep induced by perinatal nicotine.

GRANTS

These studies were funded by an OSU Honors Arts and Sciences Undergraduate Research Scholarship and an OSU School of Behavioral Sciences Undergraduate Research Grant (both to R. F. Don).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.C.B., R.F.D., R.T.B., and R.J.N. conceived and designed research; J.C.B., R.F.D., and N.Z. performed experiments; J.C.B. analyzed data; J.C.B., R.F.D., R.T.B., and R.J.N. interpreted results of experiments; J.C.B. prepared figures; J.C.B. drafted manuscript; J.C.B., R.F.D., N.Z., R.T.B., and R.J.N. edited and revised manuscript; J.C.B., R.F.D., N.Z., R.T.B., and R.J.N. approved final version of manuscript.