Perinatal nicotine reduces chemosensitivity of medullary 5-HT neurons after maturation in culture.

Abstract

Perinatal exposure to nicotine produces ventilatory and chemoreflex deficits in neonatal mammals. Medullary 5-HT neurons are putative central chemoreceptors that innervate respiratory nuclei and promote ventilation, receive cholinergic input and express nicotinic acetylcholine receptors (nAChRs). Perforated patch clamp recordings were made from cultured 5-HT neurons dissociated from the medullary raphé of 0–3 day old ePet-EYFP mice. The effect of exposure to low (6 mg kg⁻¹day⁻¹) or high (60 mg kg⁻¹day⁻¹) doses of nicotine in utero (prenatal), in culture (postnatal), or both and the effect of acute nicotine exposure (10 μM), were examined on baseline firing rate (FR at 5% CO₂, pH = 7.4) and the change in FR with acidosis (9% CO₂, pH 7.2) in young (12– 21 days in vitro, DIV) and older (≥ 22 DIV) acidosis stimulated 5-HT neurons. Nicotine exposed neurons exhibited ~67% of the response to acidosis recorded in neurons given vehicle (p=0.005), with older neurons exposed to high dose prenatal and postnatal nicotine, exhibiting only 28% of that recorded in the vehicle neurons (p<0.01). In neurons exposed to low or high dose prenatal and postnatal nicotine, acute nicotine exposure led to a smaller increase in FR (~ +51% vs +168%, p=0.026) and response to acidosis (+6% vs +67%, p=0.014) compared to vehicle. These data show that exposure to nicotine during development reduces chemosensitivity of 5-HT neurons as they mature, an effect that may be related to the abnormal chemoreflexes reported in rodents exposed to nicotine in utero, and may cause a greater risk for SIDS.

Introduction

Serotonergic (5-hydroxytryptamine, 5-HT) neurons of the medullary and midbrain raphé are important for ventilatory control and arousal responses to hypercapnia (Buchanan and Richerson 2010, Darnall et al. 2016, Hodges and Richerson 2010, Hodges et al. 2008, Richerson 2004). They innervate respiratory nuclei (Connelly et al. 1989, Holtman et al. 1984, Jacobs and Azmitia 1992) and when their neurotransmitters are released respiratory output increases (Morin et al. 1990, Pena and Ramirez 2004, Peña and Ramirez 2002, Ptak et al. 2009). They are also chemosensitive, exhibiting a robust increase in firing in response to a physiologically relevant increase in CO₂ or decrease in pH in vivo and in vitro (Richerson 2004, Teran et al. 2014, Veasey et al. 1995, Wang et al. 2001).

Abnormal function or loss of medullary 5-HT neurons impairs normal ventilatory control, including responses to hypoxia and hypercapnia (Darnall et al. 2016, Hodges and Richerson 2008, Ray et al. 2011). Abnormalities in 5-HT rich brainstem regions have been consistently reported in sudden infant death syndrome (SIDS) (Bright et al. 2017, Filiano and Kinney 1992, Kinney et al. 2009, Matturri et al. 2000) for which the cause of death of infants under 1 year of age is thought to involve inability to properly respond to respiratory stress during sleep (Kinney et al. 2009). These include abnormal development and morphology of 5-HT neurons (Paterson et al. 2006), abnormal 5-HT reuptake (Narita et al. 2001, Weese-Mayer et al. 2003, Weese-Mayer et al. 2003) and reduced 5-HT receptor binding in raphé and extra raphé locations (Panigraphy et al. 2000).

The leading and most modifiable risk factor for SIDS is maternal smoking. Nicotine, readily transferred through the placenta to the fetus and to the newborn through breastfeeding (Luck et al. 1985), has been shown to induce ventilatory deficits in mammalian neonates that resemble those that are thought to occur in SIDS cases (Campos et al. 2009). These include irregular breathing and a higher incidence of apnea, as well as blunted ventilatory and arousal responses to hypoxia and hypercapnia (Eugenín et al. 2008, Hafstrom et al. 2000, Huang et al. 2010, St John and Leiter 1999, Ueda et al. 1999). Moreover, chronic fetal nicotine exposure produces 5-HT deficits such as altered 5-HT receptor binding, 5-HT release and 5-HT transporter density, and altered abundance of 5-HT terminals in various parts of the brain including the medullary raphé (Cerpa et al. 2015, Duncan et al. 2009, Muneoka et al. 2001, Muneoka et al. 1997, Slotkin et al. 2007a, Slotkin et al. 2007b, Xu et al. 2001). Whether these nicotine induced abnormalities are associated with changes in baseline function and chemosensitivity of medullary 5-HT neurons remains unclear.

We used the perforated patch clamp technique to record firing rate (FR) and the change in FR in response to a change in CO₂ from 5% to 9% (pH 7.4 to 7.2). This was then repeated during acute exposure to nicotine (10 μM). These recordings were made from 5-HT neurons cultured for at least 12 DIV after having been exposed to nicotine or vehicle in utero and/or in culture. Tissue was dissociated from the midline medulla of P0–3 mouse pups born to dams that received low (6 mg kg⁻¹day⁻¹) or high (60 mg kg⁻¹day⁻¹) doses of nicotine or vehicle starting on the 10th day of pregnancy (prenatal). After preparation of cultures, neurons were treated with low (0.05 μM) or high (0.5 μM) concentrations of nicotine or vehicle starting at 1 DIV (postnatal). Our hypothesis was that exposure to nicotine during development would alter the response to acidosis of medullary 5-HT neurons after maturation. We also examined whether nicotine exposure during development blunted the excitatory effects of acute nicotine on baseline FR or the response to acidosis. Finally, we examined whether early life exposure to nicotine was more apparent in fully mature neurons compared to juvenile neurons that had not yet fully developed a mature response to acidosis (Cerpa et al. 2017, Wang and Richerson 1999).

Experimental Procedure

Culture preparation

Newborn mice (P0–P3) of a mixed C57BL/6 and 129 background and genetically altered to express enhanced yellow fluorescent protein under control of the enhancer region for PET1 (ePet-EYFP), which is selectively expressed in 5-HT neurons (Scott et al. 2005) (Fig 1a), were used to prepare primary cell cultures of medullary raphé tissue. Aseptic technique was used to excise the medullary raphé from newborn mice by removing a wedge of tissue from the ventromedial medulla from the pontomedullary border to the confluence of the two vertebral arteries, with longitudinal borders just lateral to the pyramidal tracts (Wang et al. 1998, Wang et al. 2001). This wedge of tissue contained the raphe pallidus, raphe magnus, raphe obscurus, parapyramidal region and tissue surrounding these nuclei. The dissected tissue was then placed in oxygenated HEPES buffered Ringer solution (in mM): 130 NaCl, 4 KCl, 1 MgCl₂, 1.5 CaCl₂, 10 HEPES, 10 Dextrose and 3 NaOH, then digested with papain solution (HEPES buffered Ringer solution with 0.5 mM EDTA, 10 U/ml papain and 0.2 mg/ml cysteine) for 30 minutes, triturated with 5% fetal bovine serum (FBS) in modified Eagles Medium (MEM) with 100 U/ml penicillin and 100 ug/ml streptomycin and plated on poly-L-ornithine and laminin coated coverslips at a density of 0.5–1×10⁵ cells ml⁻¹. Cells were allowed to attach to the coverslip for 45–50 minutes, after which 10% FBS in MEM (Nisoli et al. 1992) or 10% FBS in 54% MEM with 36% Neurobasal medium and B27 supplement (Wang, Pizzonia, et al. 1998; Wang, Tiwari, et al. 2001) was added after it had been conditioned for 2–3 weeks by glial cultures obtained from the ventromedial medulla. In our experience there is no difference in the properties of our cultures after glial conditioning over the range of 2–3 weeks. Feeding of cells occurred 4–7 days after culture preparation with a half change of Neurobasal–B27 medium containing Ara-C (1nM/ml) to reduce glial growth. Recordings were made after 12–50 DIV.

Figure 1: Identification of 5-HT neurons and recording protocol.

A) 5-HT neurons were cultured from a genetically modified mouse in which 5-HT neurons express EYFP (Scott et al, 2005). Shown is a 5-HT neuron, 1 day after preparation of culture, identifiable due to its fluorescence. Images were taken using differential interference contrast (DIC) and fluorescence microscopy. B) The responses of 5-HT neurons were recorded in 5 minute epochs of 5% and 9% CO₂ in normal aCSF and aCSF containing 10 μM (acute) nicotine. In some cases, a change in FR occurred within the first 5 minutes in acute nicotine requiring FR to be returned to baseline by decreasing current injection before pH challenges were made in nicotine.

Chronic and acute nicotine treatment of medullary 5-HT neurons

Maternal smoking during pregnancy and after birth results in both fetal and postnatal chronic exposure to nicotine. To achieve prenatal nicotine exposure of medullary 5-HT neurons in mice in utero, nicotine was given to pregnant dams at a dose of either 6 mg kg⁻¹ day⁻¹ or 60 mg kg⁻¹ day⁻¹ starting on embryonic day 10. The lower dose has been shown in rats to produce plasma nicotine concentrations equivalent to those reported in pregnant women considered to be moderate smokers (Luck et al. 1985) whereas the higher dose is required to achieve the same in mice due to their ability to metabolize nicotine 10 times faster than rats (Petersen et al. 1984, Plowchalk et al. 1992). Moreover, the larger dose is known to cause significant upregulation of (³H) nicotine binding in fetal respiratory brainstem regions (Pauly et al. 1991, Van De Kamp et al. 1994). The high dose we administered to our mice is the same as that administered by Eugenin and colleagues (2008) who recorded plasma cotinine levels of 207 ng/ml which is at the higher end of the range recorded in human amniotic fluid and fetal serum in smokers (Luck et al. 1985). To achieve in utero exposure, pregnant ePET-EYFP mice were anaesthetized with isoflurane at day 10 of gestation and an osmotic minipump (Alzet model 1002) implanted subcutaneously through an incision between the shoulder blades. The minipump delivered either 0.9% saline (vehicle) or nicotine bitartrate solution (57 or 570 mg/ml) at an infusion rate of 0.25 μl hr⁻¹ for 2 weeks. Although the steady rate of nicotine supply through use of osmotic minipumps does not reflect the episodic exposure to nicotine of smokers, the nicotine blood levels of smokers have largely been found to remain constant throughout the day (Benowitz et al. 2002). Moreover, this method of administration well simulates that of pregnant women who adopt nicotine replacement therapy (Cohen et al. 2005), and also prevents episodes of fetal hypoxia and ischemia as well as increased sympathetic nervous system activity in dams that are subjected to the stress of daily nicotine injections (Slotkin 1998). All implantations were conducted under aseptic conditions and all animals were given analgesia (0.2 mg kg⁻¹ of Flunixin) after surgery. All animal procedures were approved by the University of Iowa Institutional Animal Care and Use Committee.

To simulate postnatal nicotine exposure, nicotine bitartrate solution was added to the media used to feed cultured neurons at a concentration of either 0.05 μM or 0.5 μM starting on the first DIV. These concentrations were based on measurements from earlier studies, where the nicotine blood concentration resulting from a dose of 60 mg kg^–1day⁻¹ was reported to be between 207–233 ng/ml, translating to 0.5 μM (Eugenín et al. 2008, Robinson et al. 2002). Overall 7 treatment groups were studied: 1) Vehicle, 2) 6 mg kg^–1day⁻¹ in utero, 3) 60 mg kg^–1day⁻¹ in utero, 4) 0.05 μM in culture, 5) 0.5 μM in culture, 6) 6 mg kg^–1day⁻¹ in utero + 0.05 μM in culture, and 7) 60 mg kg^–1day⁻¹ in utero + 0.5 μM in culture. Cells were obtained from a minimum of 3 animals for each group. These chronic treatments are abbreviated throughout the manuscript as low or high dose, prenatal and/or postnatal treatment. Acute nicotine treatment involved application of 10 μM nicotine bitartrate (chosen because the concentration achieves >90% of the maximum response in acutely dissociated locus coeruleus neurons) (Gallardo and Leslie 1998) added to the bath solution during the experiment.

Electrophysiological recordings

Recordings were conducted from neurons in a Plexiglas chamber mounted on an inverted microscope (Zeiss Axiovert 1000). Culture coverslips were continuously superfused at a rate of 2 ml min⁻¹ with artificial cerebrospinal fluid (aCSF) (in mM): 124 NaCl, 3 KCl, 2 MgCl₂, 10 dextrose, 1.3 NaH₂PO₄, 26 NaHCO₃, 2 CaCl₂, maintained at room temperature and aerated with 95% O₂ / 5% CO₂ to maintain pH at 7.4. After stable recordings were obtained, neurons were superfused with aCSF containing 100 μM picrotoxin, 50 μM 2-amino-5-phosphonopentanoic acid and 10 μM 6-cyano-7-nitroquinoxaline-2–3-dione to block ionotropic GABAergic and glutamatergic synaptic transmission. Experiments were conducted while pH was continuously measured with a pH microelectrode (Microelectrodes, Inc, MI-141) placed in the inflow of the recording chamber.

Recordings were made using the gramicidin perforated patch technique (Akaike 1996) to enhance seal stability and allow longer recordings, as the whole cell technique can result in loss of repetitive firing of 5-HT neurons due to rundown caused by dialysis of intracellular molecules (Richerson 1995). Patch clamp electrodes (7–12 MΩ) were made using borosilicate glass (A-M Systems, Inc., Catalog No. 593400), cleaned in acetone and fabricated using a micropipette puller (Sutter Instruments Co., Flaming/Brown Micropipette puller Model P-97). Electrodes were filled with intracellular solution (in mM): 135 potassium methanesulphonate, 10 KCl, 5 HEPES, 1 EGTA, pH =7.2, osmolarity 274±5 mOsm. Input resistance was usually ≥ 200 MΩ (seal resistance 1–8 GΩ and access resistance between 10–30 MΩ). Neurons were considered healthy when resting membrane potential was between −60 mV and −45 mV and repetitive firing was maintained between 0.5 −1.5 Hz within the first 10 minutes of recording. Recordings were amplified (Multiclamp 700B, Molecular Devices, Sunnyvale, CA, USA), low-pass filtered (10 kHz) and acquired at 10 kilosamples sec⁻¹ with a computerized data acquisition system (Digidata 1440A, Molecular Devices) using either pCLAMP 10.2 (Molecular Devices) or software custom written using Visual Basic (Microsoft, Inc).

Recordings were made in current clamp mode. Firing of neurons occurred spontaneously or was induced with depolarizing current injection to a frequency up to 1 Hz. Recordings were discontinued if a cell required constant hyperpolarization to decrease or stabilize firing and membrane potential as this usually indicates an unhealthy neuron.

Protocol

Cultured medullary 5-HT neurons were exposed to two 25 min cycles (Fig. 1b) that each included three 5-minute exposures to 5% CO₂ (baseline, pH = 7.4) interspersed with two 5-minute exposures to 9% CO₂ (acidosis, pH =7.2). For the first cycle, 5-HT neurons were recorded in normal aCSF solution to allow quantification of the effect of treatment on membrane potential and the response to acidosis. At the beginning of the 2nd cycle, 5-HT neurons were exposed to 10 μM (acute) nicotine so that the combined effects of chronic perinatal and acute nicotine on baseline FR and the response to acidosis could be quantified. Following acute nicotine exposure, the change in baseline FR was measured, and then FR was returned back to the initial value that occurred during normal aCSF baseline conditions by changing the holding current. Adjusting FR in this way allowed the effect of acidosis to be more directly compared between cycles containing normal aCSF and aCSF with 10 μM nicotine.

Data analyses

Only acidosis-stimulated (AS) neurons were used in the final analyses. These neurons were grouped based on 1) age (12–21 DIV or younger versus ≥ 22 DIV or older) (Cerpa et al., 2017) and; 2) perinatal nicotine treatment group. The two age groups were defined based on earlier work from our lab which showed a difference at 12–21 DIV versus 22– 50 DIV in the number of chemosensitive neurons (~50% vs 80% respectively) and the magnitude of the response to a decrease in pH from 7.4 to 7.2 (average increase in FR of 80% versus 200–250% respectively) (Cerpa et al. 2017, Wang and Richerson 1999). For each perinatal nicotine treatment × age group, 12–18 neurons were recorded, depending on the degree of variability in their responses.

AS neurons were defined as those that produced a reproducible and reversible increase in FR exceeding 20% in response to an increase in CO₂ from 5% to 9% or decrease in pH from 7.4 to 7.2 (as defined by Wang et al, 1998). To establish whether a 20% increase in FR occurred during exposure to acidosis, FR from all except the first minute of each 5% CO₂ exposure during aCSF was compared to FR during all except the first minute of the subsequent and preceding exposure to 9% CO₂. Acidosis-insensitive neurons did not change their FR by more than 20% in response to the same stimulus. None of the recordings from EYFP expressing neurons displayed a decrease in FR of more than 20% in response to acidosis.

To examine whether acute nicotine treatment altered baseline FR of a neuron, an average FR was calculated from all except the first minute of the final 5% CO₂ exposure during aCSF and compared to the average baseline FR of all except the first minute of the first exposure to 5% CO₂ during the acute nicotine cycle. Similarly, to examine whether acute nicotine exposure altered the response to acidosis, the average % change in FR in response to acidosis during acute nicotine and aCSF cycles were compared. For these comparisons a stable increase or decrease of FR by 20% during the acute nicotine cycle was considered a significant response.

Statistics

Differences in the response to acidosis (% change in FR with 9% CO₂ or pH 7.2) between treatment groups regardless of age (≥ 12 DIV) were examined by performing one-way analyses of covariance (ANCOVA) (with age as covariate) with Bonferroni posthoc comparisons. Following a significant effect of the covariate “age” further one-way analyses of variance (ANOVA) were performed for a given age group (12–12 DIV and ≥22 DIV. To examine the overall effect of treatment on the response to acidosis and the % change in baseline FR with acute nicotine exposure a 2 (±acute nicotine) × 7 (treatment groups) repeated measures ANCOVA (with age as a covariate) was performed with posthoc Bonferroni comparisons. This test was repeated for a given age group if a significant effect of the covariate “age” was found. The one-way ANCOVAs and ANOVAs described above were also performed to compare differences between treatment groups in the response to acidosis or baseline FR during acute nicotine exposure. Where normality of data was not confirmed by Levene’s test, nonparametric tests were performed. To examine differences between treatment groups in the proportion of AS neurons found in culture and the proportion of AS neurons exhibiting an increase in baseline FR with acute nicotine a Chi-Sqaure analysis was performed. Lastly a One-Way ANCOVA with treatment as a factor was also performed to examine any differences in membrane potential between treatment groups.

Results

Exposure to nicotine during development does not affect membrane potential of medullary 5-HT neurons after maturation

All 293 EYFP positive 5-HT neurons recorded in culture exhibited highly regular repetitive firing (Wang et al. 2002, Wang et al. 1998, Wang et al. 2001). Of these, 213 (73%) were acidosis-stimulated (AS) (Richerson 2004). Consistent with earlier studies (Richerson 1995, Wang et al. 2002, Wang et al. 1998, Wang et al. 2001), 5-HT neurons exhibiting a reversible ≥ 20% increase in FR in response to a decrease in pH from 7.4 to approximately 7.16 for at least 2 consecutive cycles, were considered AS; all others were considered acidosis-insensitive (none were inhibited) and were not included in the final analyses. Membrane potentials of AS neurons were between −40 mV and −60 mV, with no significant difference in membrane potential between those exposed to nicotine during development and those exposed only to vehicle (Table 1), except for older neurons in which there was a difference in membrane potential between neurons after prenatal exposure to nicotine at low dose compared to high dose (p = 0.013).

Table 1:

The mean age, membrane potentials (Vm) and currents (pA) for each treatment group within a given age group. Also shown is firing rate during the 1^st cycle of 5% CO₂. Values are mean ± SE.

Treatment group	Mean age of neurons (sample size)		Membrane Potential (mV)		Current to achieve FR 0.5 – 1Hz (pA)		FR at 5% CO2 (at pH 7.4)
	12– 21 days	≥22 days	12– 21 days	≥22 days	12–21 days	≥22 days	12–21 days	≥22 days
Vehicle	17 ± 0.4 (18)	30 ± 2.4 (18)	−44.6 ± 2.3	−46.8 ± 1.6	40.4 ± 9.4	44.0 ± 7.7	0.7 ± 0.1	0.6 ± 0.04
6 mg kg−1 day−1 pre + post	16 ± 0.5 (14)	39 ± 1.9 (12)	−43.2 ± 1.5	−45.0 ± 2.8	47.5 ± 10.4	52.1 ± 11.1	0.7 ± 0.1	0.6 ± 0.1
60 mg kg−1 day−1 pre + post	16 ± 0.9 (12)	35 ± 2.7 (12)	−42.7 ± 2.4	−43.6 ± 1.1	43.2 ± 12.3	36.5 ± 4.2	0.6 ± 0.03	0.7 ± 0.1
6 mg kg−1 day−1 prenatal	17 ± 0.5 (16)	35 ± 1.8 (15)	−45.6 ± 2.2	−52.3 ± 1.1*	37.8 ± 13.7	9.4 ± 3.3	0.7 ± 0.05	0.8 ± 0.1
60 mg kg−1 day−1 prenatal	16 ± 0.6 (14)	25 ± 0.7 (15)	−45.8 ± 2.3	−42.2 ± 3.1	38.0 ± 13.1	49.1 ± 9.2	0.7 ± 0.1	0.7 ± 0.1
6 mg kg−1 day−1 postnatal	16 ±0.7 (14)	29 ± 1.8 (16)	−42.1 ± 2.6	−44.2 ± 1.8	58.6 ± 16.8	54.4 ± 18.4	0.6 ± 0.1	0.6 ± 0.1
60 mg kg−1 day−1 postnatal	17 ± 1 (12)	31 ±1.9 (19)	−43.1 ±1.9	−44.3 ± 1.8	45.9 ± 12.8	38.3 ± 9.7	0.6 ± 0.1	0.7 ± 0.1

^(*)denotes difference to 60 mg kg⁻¹ day⁻¹ prenatal

Exposure to nicotine during development reduces chemosensitivity of medullary 5-HT neurons after maturation

The response to acidosis of all AS 5-HT neurons exposed to nicotine at any time during development (n = 178) was significantly less than that of neurons given vehicle (n = 35), (202±17% vs 301±51%, p = 0.005, F=7.97,). There was a significant effect of type of developmental nicotine exposure (treatment) on the response to acidosis (p = 0.006, F=3.1) with AS neurons exposed to the high dose prenatally and postnatally exhibiting a significantly smaller response to acidosis compared to neurons given vehicle (140±42% vs 301±51%, p = 0.039).

The effect of developmental nicotine exposure on the response to acidosis was also dependent on age (p = 0.011, F=6.56) with older but not younger neurons exhibiting a significantly lower response (p = 0.002, F= 3.77 and p=0.532, F=0.85 respectively). Pairwise comparisons between exposure groups in older neurons also showed that the response to acidosis was significantly smaller in neurons given high dose nicotine prenatally and postnatally (n = 12, 85±13%, p = 0.006) or prenatally (n = 17, 107±13%, p = 0.007) compared to vehicle (n =17, 308±71%) (Fig 2a). Notably, in mice given the high dose prenatally and postnatally the percentage of AS neurons in culture was only 55% of the total number of neurons recorded, which was 22% less than that reported for neurons given vehicle (Table 2). On the other hand, those given high dose prenatally exhibited a 30% increase in AS neurons compared to those given vehicle. Comparisons between treatment groups for the proportion of neurons that were AS did not yield a dose-dependent difference between neurons exposed to nicotine compared to vehicle.

Figure 2: Prenatal exposure to nicotine reduces the response of mature neurons to acidosis.

A) Box (1^st and 3^rd quartiles, median) and whisker (5^th and 95^th percentiles) plots show the % change in firing rate (FR) in response to 9% CO₂ (pH 7.2) of 12–21 and ≥ 22 day old neurons exposed to vehicle or nicotine during the prenatal and/or postnatal period. Cross represents the mean. Statistical significance denoted by (*) is only shown for differences to vehicle. One-way ANOVA analysis revealed a significant effect of nicotine treatment (p=0.002) in older neurons with those given 60 mg kg⁻¹ day⁻¹ of nicotine prenatally and both pre and postnatally exhibiting a significantly smaller response to acidosis compared to neurons given vehicle (p = 0.007 and 0.006 respectively). B) Examples of the response to acidosis of neurons given vehicle, low dose nicotine or high dose nicotine both prenatally and postnatally.

Table 2:

Proportion of (A) AS neurons in culture and (B) AS neurons increasing baseline FR with Acute nicotine treatment.

	(A)		(B)
	Proportion of AS neurons		Proportion of AS neurons excited by Acute nicotine
	12–21 days	≥22 days	≥12 days
Vehicle	0.67	0.77	0.69
6 mg kg−1 day−1 pre + post	0.70	0.65	0.56
60 mg kg−1 day−1 pre + post	0.70	0.55	0.31*
6 mg kg−1 day−1 prenatal	0.80	0.75	0.77
60 mg kg−1 day−1 prenatal	0.68	0.89	0.50
6 mg kg−1 day−1 postnatal	0.74	0.76	0.67
60 mg kg−1 day−1 postnatal	0.64	0.90	0.71

^(*)denotes difference to Vehicle treated neurons.

Direct comparison between age groups showed that there was no difference in the overall response to acidosis between older (n=17) and younger (n=18) neurons given vehicle (308±71% and 295±75% respectively, p = 0.847, F=0.037), but in neurons given nicotine throughout development the response of all older neurons (n=91) was significantly smaller than that of all younger neurons (n=86) (154±13% and 254±31% respectively, p = 0.02, F=9.99, Fig 2a). These data indicate that nicotine exposure during development reduces the response to acidosis after neurons begin to mature, possibly by impairing further development of the response to acidosis. Figure 2b shows examples of the response to acidosis in older neurons given vehicle, compared to prenatal and postnatal exposure to low or high dose nicotine.

Nicotine exposure during development reduces the response of 5-HT neurons to acute nicotine

AS 5-HT neurons responded to acute nicotine (10 μM) with a significant increase in baseline FR (+118±13%, p = 0.016, F=5.90). 28 of 213 neurons exhibited a decrease (>−20%) in baseline FR (−42±3%), 92 exhibited no change (<±20%) and 172 exhibited an increase (+183±17%).

There was no interaction with the covariate age (p = 0.411, F=0.68), but there was an interaction between the acute effect of nicotine on baseline FR and developmental exposure to nicotine (p = 0.026, F=2.46). Pairwise comparisons for FR before and during acute exposure to nicotine showed that significant increases in FR occurred in all groups (p<0.001) except those given prenatal and postnatal nicotine at low (n=27) or high (n=26) doses (p = 0.101 and 0.098, respectively, Fig 3a). When the response to acute nicotine was calculated as the percent change in FR, there was again an overall effect of nicotine treatment (p=0.038, F=2.27) with significant differences in the responses of these 2 groups (low dose: n = 27, +50±17, p = 0.015; high dose: n = 26, +52±19, p = 0.017) compared to vehicle (n = 35, +168±40%) (Fig. 3b). Figure 3c shows examples of the effect of acute nicotine on FR in older neurons given vehicle, low dose or high dose prenatal and postnatal nicotine.

Figure 3: Prenatal exposure to nicotine reduces the effect of acute nicotine on 5-HT neurons.

Box (1^st and 3^rd quartiles, median) and whisker (5^th and 95^th percentiles) plots show A) the baseline firing rate (FR at 5% CO₂) for neurons within each treatment group (vehicle, pre/postnatal nicotine) during aCSF (black) and aCSF with 10 μM (acute) nicotine (grey) and B) the % change in baseline FR in response to acute nicotine. Cross represents the mean and closed circles are outliers. A) # denotes statistically significant differences between aCSF and acute nicotine for a given treatment group. Repeated measures analysis revealed an interaction between perinatal nicotine treatment and the baseline response to 10 μM nicotine (p = 0.026). Posthoc Bonferroni tests showed that exposure to acute nicotine elicited an average increase in baseline FR of all treatment groups except those given 6 and 60 mg kg⁻¹ day⁻¹ pre and postnatally (p =0.101 and 0.098 respectively). B) Statistical significance denoted by (*) is only shown for differences to vehicle. One-way ANOVA analysis revealed a significant effect of nicotine treatment on the % change in baseline FR (p=0.038) with exposure to acute nicotine with groups given 60 mg kg⁻¹ day⁻¹ of nicotine prenatally and both pre and postnatally exhibiting a significantly smaller increase in baseline FR (p = 0.015 and p = 0.017 respectively) compared to vehicle. C) Examples of the change in FR in response to acute nicotine of neurons given vehicle, low dose or high dose prenatal and postnatal nicotine.

When comparisons were made between treatment groups for the proportion of AS neurons that responded to acute nicotine by increasing FR by more than 20%, a significant difference was found between neurons given vehicle, in which 24 out of 35 neurons (69%) increased FR in response to acute nicotine, and those given high dose prenatal and postnatal nicotine, in which only 8 out of 26 (31%) increased FR in response to acute nicotine (Table 2, p = 0.005). These data indicate that developmental exposure to nicotine reduces sensitivity of medullary 5-HT neurons to the excitatory effects of acute nicotine and that this effect is greater when exposure has occurred throughout development.

Nicotine exposure during development reduces the enhancement of chemosensitivity caused by acute nicotine

The response to acidosis was determined during acute exposure to nicotine in 197 of 213 AS 5-HT neurons. For all neurons as a group, we found no significant effect of acute nicotine on the response to acidosis (p = 0.102, F= 2.71) and no interaction with the covariate “age” (p = 0.067, F=3.4). However, there was an interaction with developmental nicotine treatment (p = 0.014, F= 2.76). Pairwise comparisons within a given treatment group for differences in the response to acidosis during control conditions compared to acute nicotine exposure showed a significantly larger response to acidosis during acute nicotine exposure only in the vehicle group (287±48 vs 350±72, p = 0.013, Fig 4). However, no significant difference was found between treatment groups in the effect of acute nicotine exposure on the response to acidosis (Kruskal – Wallis, p = 0.091).

Figure 4: Prenatal + postnatal exposure to nicotine reduces the effect of acute nicotine on the response to acidosis of 5-HT neurons.

Box (1^st and 3^rd quartiles, median) and whisker (5^th and 95^th percentiles) plots show the % change in firing rate (FR) in response to 9% CO₂ (pH 7.2) of each individual neuron within a treatment group (vehicle, pre/postnatal nicotine) in aCSF (black) and aCSF with 10μM nicotine (grey). Cross represents the mean and closed circles are outliers. # denotes differences between aCSF and acute nicotine for a given treatment group. Repeated measures analysis found no significant effect of acute nicotine on the response to acidosis (p = 0.102), but there was an interaction between perinatal nicotine exposure and acute nicotine on the response to acidosis (p=0.014) with Bonferroni posthoc tests showing that only neurons given vehicle had a significantly greater response to acidosis during a brief acute exposure to nicotine (p = 0.013).

Discussion

The primary goal of the current study was to examine the effect of nicotine exposure during early development on chemosensitivity of cultured medullary 5-HT neurons after they mature, and on the effect of acute nicotine treatment on baseline FR and the response to acidosis. Cell culture was chosen to carry out this work because it enables highly stable patch clamp recordings from older 5-HT neurons. In culture these neurons develop over a time course that closely mimics that seen in vivo (Cerpa et al. 2017, Wang and Richerson 1999). Moreover, many cellular properties are maintained in cell culture, such as neurotransmitter identity (e.g. tryptophan hydroxylase and 5-HT content), and baseline electrophysiological properties (e.g. spike width, afterhyperpolarization and chemosensitivity) (Wang et al. 2001). Finally, with this technique the effects of nicotine on chemosensitive 5-HT neurons could be examined without the use of anesthesia which is known to mask chemosensitivity of medullary 5-HT neurons (Massey et al. 2015, Massey and Richerson 2017), and without effects of trauma that occur when using brain slices or acute dissociation (Richerson and Messer 1995). The data revealed that chemosensitivity of 5-HT neurons was reduced following perinatal nicotine exposure, particularly in more mature neurons exposed prenatally to the higher dose of 60 mg kg⁻¹ day⁻¹. We also found that perinatal exposure to nicotine blunted the excitatory effect of acute nicotine on baseline FR and chemosensitivity of cultured medullary 5-HT neurons.

Nicotine exposure during development alters function of 5-HT neurons (Cerpa et al. 2015, Slotkin et al. 2007a, Xu et al. 2001). In the current study we confirm that a portion of neurons affected in the medullary raphé are intrinsic chemosensitive 5-HT neurons. Further, we show that nicotine’s effects extend to the chemosensitivity of these neurons, significantly reducing their response to acidosis (50% lower than vehicle).

Chemosensitive 5-HT neurons in both culture and slices begin to exhibit a response to acidosis, albeit a modest one, 1 week after birth with the response gradually increasing and reaching full maturity by about postnatal day 22 (Cerpa et al. 2015, Wang and Richerson 1999). Despite this, the majority of nicotine’s effects were observed in neurons exposed in utero (± culture) indicating that the fetal period comprises a critical period for the development of chemosensitivity of 5-HT neurons. In rodents 5-HT neurons of the medullary raphé have been identified as early as embryonic day 12 and begin 5-HT synthesis and axonal growth shortly after (Wallace and Lauder 1983). This period coincides with the initial appearance of nAChRs (mouse) (Broide et al. 2019) with evidence from the human fetus showing colocalization of 5-HT neurons and (α4) nAChR subunits in the medulla as early as 15 gestational weeks (Duncan et al. 2008a). Prolonged exposure to nicotine in utero has been associated with a decrease in 5-HT terminals and evidence of 5-HT nerve damage in the brainstem (Slotkin et al. 2005, Xu et al. 2001) and specifically in the raphé obscurus (Cerpa et al. 2015), effects that are consistent with nicotine’s ability to alter the normal trajectory of cell development (Roy et al. 1998). These changes have also been associated with increased 5-HT turnover, binding, receptor expression and cell signalling, likely to offset synaptic loss (Cerpa et al. 2015, King et al. 1991, Muneoka et al. 1997, Slotkin et al. 2005, Xu et al. 2001). It is likely that 5-HT neurons of the current study have been similarly affected by prenatal (± postnatal) nicotine exposure ultimately leading to altered development of chemosensitivity. Consistent with this is the prominence of nicotine’s effects on older rather than younger neurons in which the response to acidosis is expected to be fully developed and mature (Cerpa et al, 2017, Wang and Richerson, 1999). Evidence for a potential loss of chemosensitive 5-HT neurons was also observed in older (when chemosensitive neurons are more numerous) versus younger neurons given the higher dose pre and postnatally, although this result was not significant.

Our data which show a reduced effect of acute nicotine treatment on baseline FR and the response to acidosis support that impaired cholinergic modulation is present in our culture. In the current study acute nicotine exposure either increased or decreased baseline firing of 5-HT neurons. These effects have been previously reported in 5-HT neurons of the dorsal raphé (Li et al. 1998, Mihailescu et al. 2001) as well as extramedullary raphé locations (Cheeta et al. 2000, Cordero-Erausquin and Changeux 2001) and are mediated through stimulation of nAChRs. Prolonged exposure to nicotine has been associated with abnormal cholinergic modulation attributed to desensitization and/or upregulation of nAChRs (Marks et al. 1993, Pauly et al. 1991, Wang and Sun 2005). In fact, chronic exposure to the higher dose used in the current study (60 mg kg⁻¹ day⁻¹) has been shown in mice to elicit significant upregulation of [3H⁻] nicotine binding (Van De Kamp et al. 1994) a classic response to nAChR desensitization (Wang and Sun 2005). It is therefore possible that the effects of this dose on both the response to acidosis and acute nicotine treatment may reflect similar changes in our culture.

Upregulation of nAChRs has been reported in serotonergic brain regions following fetal, postnatal and adolescent exposure (Miao et al. 1998, Slotkin 2002, Slotkin et al. 1987, Trauth et al. 1999) including in the raphe obscurus, but not the raphé pallidus, of the baboon fetus following prenatal exposure (Duncan et al. 2009). Our cultures contain neurons from the raphé obscurus, the raphé pallidus and the raphé magnus. However, it is not possible to examine whether nicotine has selective effects on specific subsets of 5-HT neurons in dissociated culture. The raphé obscurus however has recently been shown to belong to a subset of medullary 5-HT neurons whose primary role is to influence respiratory motor output whereas those in the raphé magnus influence chemosensory input and processing (Brust et al, 2014). Thus, it may be that specific targeting by nicotine of the raphé obscurus shown in earlier work (Cerpa et al. 2017, Duncan et al. 2009) may reflect this characteristic of this subset of neurons. Alternatively, it may reflect a role of the raphé obscurus in respiratory rhythm generation, which involves cholinergic modulation by α4-containing nAChRs (Shao and Feldman, 2002; Shao et al, 2008). α4- containing nAChRs have also been found in the raphé magnus (Cucciaro and Commons 2003), however further work conducted using other approaches will be necessary to determine whether the effects of perinatal nicotine exposure are selective for specific subsets of medullary 5-HT neurons.

Nicotinic upregulation following prenatal nicotine exposure has also been linked to a downregulation and partial replacement by nicotine of muscarinic control (Coddou et al. 2009, Eugenín et al. 2008). However, this occurrence which has been observed in brainstem-spinal cord and slice preparations alters the modulation of but does not necessarily reduce the response to acidification (Coddou et al. 2009, Eugenín et al. 2008). Although our findings differ from those of these earlier studies, we cannot rule out the possibility that this may be a potential mechanism of prenatal nicotine exposure in our culture. The contradictory findings between these earlier studies and our own likely reflect differences in methodology. Unlike the earlier studies which used slices or brainstem-spinal cord preparations containing the pre-inspiratory parafacial group as well as the preBotzinger complex, in the current study the use of dissociated culture ensured elimination of most synaptic input into the medullary raphé and isolation of 5-HT neurons (Wang et al, 1998) such that their intrinsic properties could be studied. Furthermore, we studied the effects of perinatal nicotine exposure in neurons aged over 12 days whereas these earlier studies focused on animals that were under a week old. It is therefore likely that our data mostly represent long term direct effects of perinatal nicotine exposure on 5-HT neurons rather than an effect on a chemosensitive site which they innervate.

A very likely mechanism for the reduced response to acidosis of nicotine exposed neurons may involve an increase in 5-HT1a receptors leading to increased autoinhibition of these neurons. Stimulation of 5-HT1a receptors inhibits firing of 5-HT neurons (Hajós et al. 1999, Sprouse and Aghajanian 1987) and application of 5-HT1a agonists in the medullary raphé decreases the ventilatory response to CO₂ in unanaesthatised piglets (Messier et al. 2004) and rats (Taylor et al. 2005). Through this mechanism prolonged nicotine exposure has been shown to inhibit 5-HT neurons in the adult and fetal brainstem in areas receiving 5-HT innervation (Slotkin et al. 2005, Xu et al. 2001), the dorsal raphé of adult rats (Sperling and Commons 2011) and in the raphé obscurus of the medullary raphé (Cerpa et al. 2015, Duncan et al. 2009), known to play a major role in respiratory rhythm modulation (DePuy et al. 2011, Ptak et al. 2009). We can not say whether these effects extend to the response to acidosis however Cerpa and colleagues (2015) recently showed that in the raphé obscurus of week old mice following prenatal nicotine exposure, increased 5-HT1a was associated with reduced cfos immunoreactivity during hypercapnia.

Several studies have reported a GABAergic, glutamatergic and noradrenergic involvement in the modulation of 5-HT neurons (Haddjeri et al. 2000, Pan et al. 1989, Vandermaelen and Aghajanian 1983) and in facilitating nicotine’s effects on these neurons (Chang et al. 2011, Hernández-Vázquez et al. 2014, Li et al. 1998, Mihailescu et al. 2002). In the dorsal raphé, acute nicotine increases firing of 5-HT neurons through release of norepinephrine (Li et al. 1998) and glutamate (Chang et al. 2011) and both stimulates and inhibits firing through the release of GABA (Mihailescu et al. 2002), with the last being mediated through stimulation of α7 containing nAChRs (Hernández-Vázquez et al. 2014). While it is possible that some of the observed effects of acute and perhaps perinatal nicotine on 5-HT neurons may be attributed to such events, these effects are likely minimal. This is attributed to our combined use of appropriate inhibitors and acutely dissociated culture in which 5-HT neurons have been shown to maintain their response to acidosis – supporting that this response is intrinsic (Corcoran et al. 2009), and in which the usual noradrenergic input into the medullary raphé is mostly eliminated (Vandermaelen and Aghajanian 1983, Wang et al. 2001). In line with this is a recent finding by Cerpa and colleagues (2015) which showed that the blunting effect of prenatal nicotine exposure on firing rate of 3-day old raphé obscurus 5-HT neurons was present regardless of the absence or presence of GABAergic and glutamatergic blockers.

While highly speculative, the reduced chemosensitivity of neurons treated perinatally with nicotine may also involve nicotinic effects on pH sensitive calcium – activated nonselective cation (CAN) currents (Cerpa et al. 2015, Massey et al. 2013, Richerson 2004). This current which in cultured medullary 5-HT neurons is inhibited by acidosis over the pH range 7.2 – 7.6 (Tiwari et al. 2000) is activated through acute cholinergic receptor activation in other areas of the brain including the cortex and hippocampus even in the presence of glutamatergic and GABAergic blockers (Yoshida and Hasselmo 2009, Yoshida et al. 2012). Whether similar cholinergic modulation of CAN currents occurs in the medullary raphé and whether these can be altered by perinatal nicotine exposure to subsequently affect the intrinsic chemosensitivity of 5-HT neurons requires further investigation. Ultimately further work is required to pinpoint the exact mechanisms by which altered cholinergic modulation of medullary 5-HT neurons can affect their intrinsic chemosensitivity.

The novel finding that perinatal nicotine exposure diminishes in vitro chemosensitivity of 5-HT neurons is important. The magnitude of the response to acidosis of medullary 5-HT neurons (~ 3–4 fold increase in response to a change in pH from 7.4 to 7.2) has been shown to be consistent across datasets (Wang et al. 2002, Wang et al. 1998, Wang et al. 2001) and of a magnitude close to the robust response of the respiratory network in vivo (Fencl et al. 1966, Pappenheimer et al. 1965). Therefore, a reduction in chemosensitivity of these neurons such as that reported here, could significantly reduce the response in vivo. Previously, loss of 5-HT neurons has been shown to result in a diminished ventilatory response to CO₂. Targeted chemical lesions and specific chemical inhibition of medullary 5-HT neurons diminish the ventilatory response to CO₂ by up to 50% (Dias et al. 2007, Messier et al. 2004, Nattie et al. 2004, Taylor et al. 2005). Further, genetic loss of function or deletion of 5-HT neurons has the same effect both in vivo and in vitro (Hodges et al. 2008, Li and Nattie 2008, Ray et al. 2011).

Reduced chemosensitivity of medullary 5-HT neurons is relevant to SIDS pathophysiology. Pathological features of SIDS include brainstem 5-HT abnormalities (Paterson et al. 2006), with cholinergic abnormalities also reported in infants whose mothers smoked during pregnancy (Duncan et al. 2008b). A physiological feature reported in some SIDS cases is abnormal ventilatory control during sleep (Kahn et al. 1992, Kato et al. 2001, Schechtman et al. 1991) including an impaired ventilatory response to CO₂ (Shannon et al. 1977). Chronic perinatal nicotine exposure (Coddou et al. 2009, Eugenín et al. 2008) and loss of 5-HT neurons (Hodges et al. 2008, Li and Nattie 2008, Ray et al. 2011) each blunt the ventilatory response to CO₂. More recently it was shown that chronic perinatal nicotine exposure can blunt the hypercapnic ventilatory response through loss of medullary 5-HT neurons (Cerpa et al. 2015). While we acknowledge that the conditions under which our results were obtained do not replicate those in vivo and in particular those under which a human fetus may be exposed to nicotine in utero, our data do suggest that a potential mechanism by which chronic perinatal nicotine exposure may blunt the neonatal response to CO₂ response is by impairing chemosensitivity of medullary 5-HT neurons. Future work which would help increase our understanding of this mechanism might involve examining the impact of perinatal nicotine exposure on the culture and examining whether these changes persist in other preparations where synaptic networks remain intact or in vivo.

We have provided novel evidence of a direct action of chronic perinatal nicotine exposure on acidosis-sensitive medullary 5-HT neurons. Prenatal and/or postnatal nicotine treatment, particularly with a dose known to elicit fetal nAChR upregulation, reduced pH chemosensitivity of medullary 5-HT neurons. Further, 5-HT neurons exposed to this dose exhibited nicotinic deficits that affected function both in the absence and presence of acidosis. Reduction in sensitivity of these neurons to acidosis would be expected to reduce the hypercapnic ventilatory response, which has been proposed to be involved in the pathophysiology of SIDS. Our data therefore provide novel insight into a potential mechanism by which chronic perinatal nicotine exposure may increase the risk of SIDS through impaired function of chemosensitive medullary 5-HT neurons that contribute to the ventilatory response to CO₂.