Developmental nicotine exposure elicits multigenerational disequilibria in proBDNF proteolysis and glucocorticoid signaling in the frontal cortices, striata, and hippocampi of adolescent mice

Graphical Abstract

1. Introduction

Ten percent of women in the U.S. report smoking traditional cigarettes during pregnancy and fourteen percent divulge electronic cigarette use during pregnancy [1], [2]. Alarmingly, epidemiological research reveals that, despite the lack of corroborative evidence, the majority of individuals surveyed misperceive that electronic cigarettes represent a safer and healthier alternative to traditional cigarettes, and this misconception is most prevalent in women of reproductive age [3], [4], [5], [6], [7]. Contrary to majority opinion, the use of both traditional and electronic cigarettes during pregnancy constitutes developmental nicotine (NIC) exposure (DNE), which is associated with numerous fetal consequences including pre-mature birth, low birth weight, and Sudden Infant Death Syndrome [8], [9], [10], [11], [12]. In addition to its deleterious consequences for the newborn fetus, DNE is associated with neurodevelopmental disorders such as attention deficit hyperactivity disorder (ADHD), autism, and schizophrenia in developing children [1], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. Of further concern, the neurodevelopmental insults conferred by DNE appear transmissible across multiple offspring generations, per a recent study revealing that grand-maternal smoking (for the maternal but not paternal lineage) increases the risk for autism diagnosis in grandchildren [17].

Consistent with the epidemiological associations among DNE and various neurodevelopmental disorders, DNE hinders neurodevelopment, elicits neurodegeneration, decreases dopamine (DA) release, impairs DA turnover, and reduces tyrosine hydroxylase (TH) immunoreactivity in the frontal cortices and striata of first-generation rodent offspring, and these neurodevelopmental and neuropharmacological changes co-occur with aberrant DNE-induced neurobehavioral phenotypes such as hyperactivity and impulsivity-like behaviors [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]. In support and extension of this literature, we recently reported that DNE precipitates multigenerational nicotine preference, diurnal and nocturnal hyperactivity, aberrant circadian rhythmicity of home cage activity, and increased risk-taking behaviors that are reversibly rescued by methylphenidate and modulated by voluntary nicotine consumption [32]. In addition, we previously demonstrated that DNE elicits multigenerational alterations in nicotinic acetylcholine receptor (nAChR) and DA transporter (DAT) function in the frontal cortices and striata that are transmitted to at least second-generation adolescent DNE offspring [32]. Collectively, the multigenerational DNE-induced brain and behavioral perturbations which we recently documented [32] recapitulate key domains of multiple neurodevelopmental disorders such as ADHD, autism, and schizophrenia and are consistent with the multigenerational transmission of autism risk reported in the grandchildren of maternal smokers [17].

In addition to hyperactivity, impulsivity/risk-taking, DAT dysregulation, and other phenotypic anomalies, the children of maternal smokers and animal models thereof display downregulation of corticostriatal and hippocampal brain-derived neurotrophic factor (BDNF), a neurotrophin which acts in part via TrkB receptors and is critical for proper neurodevelopment [26], [33], [34], [35]. Notably, these DNE-induced BDNF deficits recapitulate those observed in ADHD, autism, and schizophrenia [36], [37], [38], [39]. In tandem with BDNF deficits, neurodevelopmental disorders are also associated with impaired proteolytic processing and resulting accumulation of proBDNF, the pro-apoptotic peptide precursor of BDNF which acts in part via p75NTR receptors to oppose the functions of mature BDNF [39], [40], [41], [42], [43], [44]. Therein, proBDNF proteolysis is catalyzed in part by the metalloprotease furin, the downregulation of which is documented in neurodevelopmental disorders and thought to contribute to the accumulation of proBDNF and corresponding deficiency of mature BDNF observed therein [39], [40], [41], [42], [45], [46], [47], [48], [49], [50]. Ultimately, BDNF deficits, proBDNF accumulation, and furin downregulation share analogous consequences which include disruption of neurodevelopment and synaptogenesis, impairment of dendritic arborization, altered synaptic plasticity, and aberrant synaptic pruning accompanied by learning and memory impairments, emotional dysregulation, and aberrant stress responsivity [34], [40], [44], [51], [52], [53], [54], [55], [56]. Each of these shared consequences of BDNF deficits, proBDNF accumulation, and furin dysregulation are characteristic of neurodevelopmental disorders and have been documented in DNE children and animal models [14], [36], [37], [38], [39], [41], [50], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69]. In aggregate, these findings imply that proBDNF-BDNF imbalance resulting from impaired furin-mediated proBDNF proteolysis may contribute to the neurodevelopmental insults conferred by DNE.

In conjunction with neurotrophic dysfunction, neurodevelopmentally-disordered children, the children of maternal smokers, and DNE animal models exhibit hypothalamic-pituitary-adrenal (HPA) axis dysfunction including hypocortisolemia/ hypocorticosteronemia and altered glucocorticoid (GC) receptor (GR) expression [14], [57], [58], [66], [67], [68], [69], [70], [71], [72]. Notably, HPA axis dysfunction in neurodevelopmental disorders such as ADHD and autism is mediated in part by dysregulation of the chaperone protein FK506 binding protein 5 (FKBP5), which suppresses GR activation, dimerization, and nuclear translocation and thereby attenuates GR-mediated modulation of gene expression [57], [58], [71], [72]. Similarly, HPA axis and FKBP5 dysregulation are known to disrupt neurodevelopment and synaptic plasticity), cause hyperactivity, elicit impulsivity, alter stress responsivity, and contribute to emotional dysregulation [14], [59], [71], [73], [74], [75], [76], [77], [78], [79]. These HPA axis dysfunction-related phenomena are observed in DNE children and animal models as well as neurodevelopmental disorders such as ADHD, autism, and schizophrenia [14], [36], [37], [38], [39], [41], [50], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69].

Interestingly, contemporary research reveals that BDNF and GR signaling undergo extensive reciprocal interactions which are important for glutamatergic neurotransmission, mesocorticolimbic and stress-induced neuroplasticity, and cortical dendritic morphology [80], [81], [82], [83], [84], [85]. Moreover, dysregulation of this BDNF-GR interactome may relate to neurodevelopmental disorders such as autism and neurodegenerative disorders such as Alzheimer’s disease [85], [86]. Of particular relevance to the present study, BDNF regulates GR activity via BDNF-dependent phosphorylation of GR at Ser¹⁵⁵ and Ser²⁸⁷, and activated GRs modulate Bdnf transcription as well as vesicular BDNF transport and release [87], [88], [89].

In aggregate, previous research identifies neurotrophic and HPA axis dysfunction as putative neurobiological substrates for the ADHD-, autism-, and schizophrenia-like phenotypes exhibited by first-generation DNE children and animal models. Despite these findings and our recent work demonstrating that DNE-induced neurodevelopmental disorder-like behavioral, neuropharmacological, and epigenetic anomalies are intergenerationally transmissible, no prior studies have modeled the multigenerational impacts of DNE on neurotrophic or HPA axis functionality [32]. Addressing this void in the literature, the present study evaluated the hypotheses that DNE elicits multigenerational perturbations in proBDNF/BDNF balance and GR expression and/or function reminiscent of those observed in neurodevelopmental disorders.

2. Materials & Methods

2.1. Animals

All experimental and housing conditions for the present study were reviewed and pre-approved by the Institutional Animal Care and Utilization Committee at the University of Colorado Boulder, and conform to the guidelines for animal care and use established by the NIH and the Guide for the Care and Use of Laboratory Animals (8th Ed.). All mice were bred and maintained in the same on-site animal facility on a standard 12h light/dark cycle (lights on at 07:00) and were provided food (Envigo Teklad 2914 irradiated rodent diet, Harlan, Madison, WI) and water ad libitum. As previously described and diagrammed in Figure 1, beginning thirty days prior to mating with drug-naïve males, female C57BL/6J breeders received 0.2% saccharin (ThermoFisher, Waltham, MA) (vehicle) or 0.2% saccharin and 200 μg/mL free-base nicotine (MilliporeSigma, Burlington, MA) (DNE) in place of drinking water [32]. Solutions were replaced twice weekly. Vehicle or nicotine treatment of breeders continued until weaning of offspring at PND 21, after which point water was provided to all offspring as the sole fluid source. Randomly selected female F1 DNE offspring were subsequently mated with drug-naïve males to generate F2 DNE offspring. Notably, female F1 DNE offspring used for the breeding of F2 DNE progeny were naïve to direct (post-weaning) exposure to nicotine, and thus the transmission of DNE-induced phenotypes from F1 DNE to F2 DNE offspring occurs exclusively via the F1 maternal germline (oocytes). All experiments utilized tissues collected from both sexes of offspring at PND 45 (adolescence), and sex was included as a biological variable in all data analyses.

Figure 1. Procedural schedule for breeding as well as tissue and plasma collection.

Beginning thirty days prior (PND 60) to mating with drug-naïve male sires (PND 90), zeroth generation (F0) female C57BL/6J breeders underwent passive oral treatment with 0.2% aqueous saccharin (developmental vehicle exposure) or 0.2% aqueous saccharin containing 200 μg/mL nicotine (developmental nicotine exposure, DNE). Vehicle or nicotine treatment of F0 dams continued until weaning of first-generation (F1) developmental vehicle-exposed (F1 Veh) or developmental nicotine-exposed (F1 NIC) offspring at PND 21, whereafter water was provided as the sole fluid source for all progeny. At PND 90, randomly selected female F1 NIC mice were mated with drug-naïve male sires to generate second-generation (F2) developmental nicotine-exposed (F2 NIC) offspring. For tissue collection, whole brains were extracted from PND 45 (adolescent) offspring from each developmental exposure group (F1 Veh, F1 NIC, and F2 NIC), and bilateral frontal cortices, striata, and hippocampi were obtained by crude dissection. For plasma collection, blood samples were drawn from the left submandibular vein of PND 45 progeny from each developmental exposure group, and plasma was then isolated by centrifugation. PND, post-natal day; Veh, 0.2% aqueous saccharin; NIC, 200 μg/mL nicotine in 0.2% aqueous saccharin; F1 Veh, first-generation developmental vehicle-exposed offspring; F1 NIC, first-generation developmental nicotine-exposed offspring; F2 NIC, second-generation developmental nicotine-exposed offspring.

Importantly, the 200 μg/mL oral nicotine regimen utilized for this study yields a similar nicotine pharmacokinetic profile in C57BL/6J mice to that observed in regular smokers, elicits behavioral and neurobiological perturbations, impacts neurodevelopment, and is widely adapted across the DNE literature, thereby bolstering the generalizability of the DNE paradigm employed herein [20], [30], [32], [90], [91], [92], [93], [94]. It is also noteworthy that, upon co-housing with pre-treated dams, drug-naïve sires were permitted to consume vehicle or nicotine drinking solutions. As such, the DNE paradigm employed for the present study may be more appropriately conceptualized as parental rather than exclusively maternal DNE for first-generation DNE mice, while second-generation DNE mice were exposed to nicotine solely via the maternal germline. Another potential caveat of this study is the omission of a second-generation developmental vehicle-exposed control group. Notwithstanding this potential confound, pilot data and prior research reveal no multigenerational impacts of developmental exposure to an oral saccharin vehicle identical to that utilized for this study [32], [95], [96]. In addition, while we cannot exclude potential confounds stemming from uncontrollable environmental factors such as variations in animal care technicians, care was taken to minimize the impact of external variables via the conduction of all procedures by same experimenter (JMB) and the assignment of standardized climate-control settings for temperature, humidity, and lighting in all animal housing and tissue/plasma collection rooms. Furthermore, to minimize between-litter and between-breeder variability within each treatment group, tissues and plasma obtained from a minimum of 6 total litters from a minimum of 4 total breeder pairs were assayed for each group and experiment. There were no group differences in litter size or pup survival rates, and multivariate ANOVAs indicated no covariation with breeder, litter, or season for any outcome measure.

2.2. Tissue and Plasma Collection

Tissue collection was conducted as previously described [32]. Briefly, whole brains were collected following cervical dislocation and decapitation at PND 45, and bilateral frontal cortices, striata, and hippocampi were manually dissected in an ice-chilled glass dish. Blood samples were collected from the submandibular vein of PND 45 mice. All blood and tissue samples were collected between 07:00 – 08:00 hours, which corresponds to the trough of circadian CORT rhythms and is considered the optimal collection time for single-point comparison of basal CORT levels and GR activity [97]. Plasmas were isolated from blood samples by centrifugation.

2.3. Tissue Lysis and Subcellular Fractionation

Immediately following collection, fresh (never frozen) tissue samples were homogenized and nuclear as well as cytosolic-enriched proteins were extracted using a NE-PER kit (ThermoFisher, Waltham, MA) according to manufacturer protocol.

2.4. Determination and Standardization of Lysate Protein Concentrations

Total protein concentrations of all fractionated lysates were determined using a Pierce bicinchoninic acid (BCA) Assay Kit (ThermoFisher, Waltham, MA) according to manufacturer protocol. Samples were then diluted to a concentration of 2 μg/μL and separated into 15 μL (30 μg) aliquots that were stored at −80° C until subsequent immunoblot experiments.

2.5. Immunoblotting Procedures

All immunoblots were conducted according to standard practice [98], [99]. Briefly, lysate aliquots containing 30 μg total protein were reduced and denatured via the addition of 4x Laemmli sample buffer (Bio-Rad, Hercules, CA) followed by incubation for 20 minutes at 95° C in a dry heat block. Samples were then loaded in 4-20% polyacrylamide gradient tris-glycine gels (Bio-Rad, Hercules, CA) and vertically electrophoresed for 90 minutes at 100V in denaturing tris-glycine buffer containing 0.1% SDS (Bio-Rad, Hercules, CA). Following electrophoresis, proteins were transferred to 0.45 pm Immobilon-P PVDF membrane (MilliporeSigma, Burlington, MA) by electroblotting at 100V for 60 minutes in tris-glycine buffer (Bio-Rad, Hercules, CA) containing 20% methanol (ThermoFisher, Waltham, MA). Upon completion of protein transfer, membranes were incubated for 30 minutes at room temperature (RT) in blocking solution containing 5% non-fat milk (Bio-Rad, Hercules, CA) and 3% normal donkey sera (MilliporeSigma, Burlington, MA) in tris-buffered saline (pH 7.4) (ThermoFisher, Waltham, MA) with 0.15% Tween-20 (Bio-Rad, Hercules, CA) (0.15% TBST) (ThermoFisher, Waltham, MA). Blocked membranes were then incubated overnight at 4° C with agitation in primary antibody solution prepared by diluting the appropriate primary antibody (raised in rabbit) to a concentration of 1:1000 in blocking solution. Primary antibodies used for this study were as follows: antî-tubulin (ProteinTech Cat#10094-1-AP), anti-GAPDH (ProteinTech Cat# 10494-1-AP), anti-TBP (ProteinTech Cat# 22006-1-AP), anti-BDNF (ProteinTech Cat# 28205-1-AP), anti-furin (ProteinTech Cat# 18413-1-AP), anti-glucocorticoid receptor (ProteinTech Cat# 24050-1-AP), anti-phospho-glucocorticoid receptor (Ser²⁸⁷) (MilliporeSigma Cat# ABS1582), and anti-FKBP5 (ProteinTech Cat# 14155-1-AP). Following primary antibody incubation, membranes were washed 5 x 5 minutes in 0.15% TBST with agitation. Washed membranes were next incubated (with agitation) for one hour at RT in secondary antibody solution prepared by dilution of horseradish peroxidase-conjugated donkey anti-rabbit IgG (Bio-Rad, Hercules, CA) to a concentration of 1:3000 in blocking solution. Membranes were next washed 5 x 5 minutes in 0.15% TBST (with agitation) and then incubated (without agitation) for 5 minutes in chemiluminescent substrate (Clarity Western ECL Substrate, Bio-Rad, Hercules, CA). Membrane images were captured using a FluorChem Imager (ProteinSimple, San Jose, CA).

2.6. Corticosterone ELISA

Basal plasma CORT was quantified via a colorimetric enzyme-linked immunosorbent assay (ELISA) (Arbor Assays, Ann Arbor, MI) according to manufacturer protocol.

2.7. Statistical Analyses

For immunoblot analyses, raw 8-bit image files were uploaded to ImageJ 1.52a, and background was subtracted from all images prior to densitometric measurements [39], [100], [101]. Mean grey values were measured for all target antigen and loading control bands [39], [101]. To determine relative optical density, mean grey values for target bands were divided by those for corresponding loading control bands [39], [101]. For BDNF immunoblots, relative optical density values were determined separately for proBDNF (~32 kD), dimeric BDNF (~28kD), and monomeric BDNF (~14 kD) [39]. To provide a measure of mature (dimeric + monomeric) BDNF, corresponding relative optical density values for dimeric and monomeric BDNF were summed [39]. Similarly, to provide a measure of total (pro + mature) BDNF, corresponding relative optical density values for proBDNF, dimeric BDNF, and monomeric BDNF were summed [39]. Lastly, to generate a surrogate measure of proBDNF proteolysis (ratio of proBDNF versus total BDNF), proBDNF relative optical density values were divided by corresponding total BDNF relative optical density values [39].

For GR immunoblots, corresponding relative optical density values for nuclear GRs and cytosolic GRs were summed to provide a measure of total (nuclear + cytosolic) GR abundance [102], [103]. To obtain a proxy measure for GR activation (relative nuclear GR localization), corresponding relative optical density values for nuclear GRs were divided by those for total GRs [102], [103]. Furthermore, to calculate fractional GR (Ser²⁸⁷) phosphorylation, corresponding total, nuclear, and cytosolic relative optical density values for the phospho-specific bands were divided by those for the pan-specific (total) bands. To provide biologically-relevant outcome measures, the arbitrary relative optical density values obtained for total BDNF, proBDNF, mature BDNF, total GR abundance and fractional (Ser²⁸⁷) phosphorylation, nuclear GR abundance and fractional (Ser²⁸⁷) phosphorylation, cytosolic GR abundance and fractional (Ser²⁸⁷) phosphorylation, furin, and FKBP5, as well as the values obtained for proBDNF proteolysis, relative nuclear GR localization, and fractional GR (Ser²⁸⁷) phosphorylation were used to calculate, respectively, the percentages relative optical density, percentages proBDNF proteolysis, percentages relative nuclear GR localization, and fractional GR (Ser²⁸⁷) phosphorylation versus the mean control (F1 Veh) value for each sample, in which form data were analyzed and visualized [101].

Processing of raw spectrophotometric data generated by CORT ELISAs was conducted according to manufacturer protocol. Basal plasma corticosterone levels were estimated from a four-parameter logistic regression curve defined by eight serially diluted corticosterone standards. Technical duplicates for each sample were averaged prior to subsequent statistical analyses.

For BDNF and GR immunoblots, respectively, percentages proBDNF proteolysis and percentages relative nuclear GR localization versus the mean control (F1 Veh) value for each sample were analyzed by mixed ANOVA with the between-subjects factor group (F1 Veh, F1 NIC, or F2 NIC) and the within-subjects factors region (frontal cortices, striata, or hippocampi) and measure (total BDNF, proBDNF, BDNF, or proBDNF proteolysis for BDNF immunoblots; total GR, nuclear GR, cytosolic GR, or relative nuclear GR localization for GR immunoblots). For phospho-GR (Ser²⁸⁷) immunoblots, total, nuclear, and cytosolic fractional GR (Ser²⁸⁷) phosphorylation, expressed as percentages of the mean F1 Veh (control) value, were analyzed by mixed ANOVA with the between-subjects factor group (F1 Veh, F1 NIC, or F2 NIC) and the within-subjects factors region (frontal cortices, striata, or hippocampi) and measure (total, nuclear, or cytosolic fractional GR (Ser²⁸⁷) phosphorylation). For furin and FKBP5 immunoblots, percentages relative optical density versus the mean F1 Veh (control) value were analyzed by mixed ANOVA with the between-subjects factor group (F1 Veh, F1 NIC, or F2 NIC) and the within-subjects factor region (frontal cortices, striata, or hippocampi). Basal plasma CORT data were analyzed by one-way ANOVA with the between-subjects factor group (F1 Veh, F1 NIC, or F2 NIC). Where appropriate for each ANOVA, Bonferroni’s multiple comparisons post-hoc test was applied.

Statistical analyses and data visualization were conducted using SPSS 25 (IBM Analytics, Armonk, NY) and GraphPad Prism 7.04 (GraphPad Software, La Jolla, California, USA), respectively. All datasets were screened for outliers via the ROUT test (Q=1%), and confirmed outliers were excluded from subsequent analyses where appropriate. A maximum of one outlier was excluded per group for each dataset. All data were initially analyzed by multifactoral ANOVA to determine if there were effects of sex, breeder, or litter. No main effects of or interactions with these factors were detected for any measure, and thus data were collapsed accordingly.

3. Results

The procedural paradigm employed for the breeding of mice as well as the collection of tissues and plasma is depicted in Figure 1 and detailed in the methods section. As previously described, F0 dams were pre-treated with either nicotine (200 μg/mL in 0.2% saccharin) or vehicle (0.2% saccharin) as the exclusive fluid source beginning 30 days before mating with drug-naive sires and persisting until the weaning of offspring [32]. First-generation DNE (F1 NIC) offspring were thereby exposed to vehicle and nicotine from conception through weaning, which recapitulates smoking during pregnancy and nursing, while F1 Veh mice were exposed to vehicle alone over the same period. Second-generation DNE (F2 NIC) mice are the offspring of female F1 NIC mice supplied with water alone following weaning and subsequently mated with drug-naïve males. Accordingly, F2 NIC mice were exposed to nicotine exclusively via the maternal germline (oocytes) and were therefore devoid of direct nicotine exposure. Importantly, to eliminate potential confounds arising from temporal variability in basal plasma CORT levels and GR activation, all immunoblots and ELISAs utilized tissue or plasma, respectively, collected from PND 45 (adolescent) mice within the same one-hour interval (07:00 – 08:00) spanning the approximate trough of the circadian CORT and GR activity rhythms for the light cycle employed in the present study. All experiments utilized tissue or plasma obtained from both sexes at PND 45, and sex was included as a biological variable for all statistical analyses.

3.1. DNE elicits multigenerational proBDNF proteolysis deficits in the frontal cortices and striata.

DNE is linked to deficient BDNF signaling and other neurodevelopmental insults in first-generation human and rodent offspring, and neurodevelopmentally-disordered children exhibit proBDNF/BDNF imbalance [26], [33], [34], [35], [40], [41], [42], [48], [49], [50]. In consideration of this evidence and our recent work demonstrating ADHD-like behavioral and neuropharmacological anomalies in DNE offspring and grandoffspring, we hypothesized that DNE may elicit multigenerational disruption of proBDNF proteolysis, resulting in proBDNF accumulation and BDNF deficiency similar to that observed in neurodevelopmental disorders [32]. Pursuant to this hypothesis, we conducted immunoblot analyses of total (pro + mature) BDNF abundance (Figs. 2A, 2E, and 2I, respectively), proBDNF abundance (Figs. 2B, 2F, and 2J, respectively), mature (dimeric + monomeric) BDNF abundance (Figs. 2C, 2G, and 2K, respectively), and a surrogate measure of proBDNF proteolysis (ratio of proBDNF versus total BDNF) (Figs. 2D, 2H, and 2L, respectively) in the frontal cortices, striata, and hippocampi of first- and second-generation adolescent DNE mice versus F1 Veh controls. Main effects of group (F_2,96=5.17; p=0.029) and measure (F_3,96=26.94; p=1.3e⁻¹¹), a significant group x measure interaction (F_6,96=6.10; p=0.008), a significant region x measure interaction (F_6,96=6.58; p=0.002), and a significant group x region x measure interaction (F_12,96=2.22; p=0.022) were detected for percentages relative optical densities versus the mean F1 Veh control value. There were no group differences in frontal cortical, striatal, or hippocampal total BDNF abundance. However, relative to F1 Veh mice, F1 NIC and F2 NIC mice have increased proBDNF abundance in the frontal cortices (p=0.007 and p=0.032, respectively) and striata (p=0.022 and p=0.024, respectively), reduced mature BDNF abundance in the frontal cortices (p=0.006 and p=0.022, respectively) and striata (p=0.005 and p=0.003, respectively), and reduced proBDNF proteolysis in the frontal cortices (p=0.002 and p=0.029, respectively) and striata (p=0.005 and p=0.007, respectively). No changes were detected for any of these measures in the hippocampi of F1 NIC or F2 NIC mice compared to F1 Veh controls.

Figure 2. DNE elicits multigenerational BDNF processing deficits in frontal cortex and hippocampus.

Representative Western blot images and densitometric measurements of proBDNF and mature BDNF abundance as well as BDNF processing (proBDNF proteolysis) in frontal cortex (n_F1Veh=13, n_F1NIC=10, and n_F2NIC=12), striatum (n_F1Veh=12, n_F1NIC=10, and n_F2NIC=10), and hippocampus (n_F1Veh=12, n_F1NIC=9, and n_F2NIC=9).). (A) Total BDNF abundance in frontal cortex. There were no group differences in frontal cortical total BDNF abundance. (B) proBDNF abundance in frontal cortex. F1 NIC and F2 NIC mice have increased frontal cortical proBDNF content (p=0.007 and p=0.032, respectively). (C) Mature BDNF abundance in frontal cortex. Frontal cortical mature BDNF content is decreased in F1 NIC and F2 NIC mice (p=0.006 and p=0.022, respectively). (D) proBDNF proteolysis in frontal cortex. F1 NIC and F2 mice NIC have impaired frontal cortical proBDNF processing (p=0.002 and p=0.029, respectively). (E) Total BDNF abundance in striatum. There were no group differences in striatal total BDNF abundance. (G) proBDNF abundance in striatum. F1 NIC and F2 NIC mice have increased striatal proBDNF content (p=0.022 and p=0.024, respectively). (G) Mature BDNF abundance in striatum. Striatal mature BDNF content is decreased in F1 NIC and F2 NIC mice (p=0.005 and p=0.003, respectively). (H) proBDNF proteolysis in striatum. F1 NIC and F2 mice NIC have impaired striatal proBDNF processing (p=0.005 and p=0.007, respectively). (I) Total BDNF abundance in hippocampus. There were no group differences in hippocampal total BDNF abundance. (J) proBDNF abundance in hippocampus. F1 NIC and F2 NIC mice have unaltered striatal proBDNF content (K) Mature BDNF abundance in hippocampus. Hippocampal mature BDNF content is unaltered in F1 NIC and F2 NIC mice. (L) proBDNF proteolysis in hippocampus. F1 NIC and F2 mice NIC have unaltered hippocampal proBDNF processing. Dashed lines indicate locations at which representative immunoblot images were cropped to reduce size. FCX, frontal cortices; STR, striata; HIPP, hippocampi; BDNF, brain-derived neurotrophic factor; pBDNF, proBDNF; dBDNF, dimeric BDNF; mBDNF, monomeric BDNF; B-tub, β-tubulin. All data are mean ± S.E.M. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

3.2. DNE elicits multigenerational corticostriatal furin deficits that covary with the impairment of corticostriatal proBDNF proteolysis

Clinical research attributes the impairment of proBDNF proteolysis observed in neurodevelopmental disorders such as ADHD, autism, and schizophrenia to downregulation of furin, a metalloprotease which processes proBDNF to mature BDNF and thereby controls the proBDNF-BDNF ratio [39], [40], [41], [42], [45], [46], [47], [48], [49], [50]. Given the multigenerational disruption of proBDNF proteolysis detected in DNE offspring and grandoffspring, we posited that DNE may confer multigenerational downregulation of furin. Therefore, we quantified furin content, expressed as percentages relative furin optical density versus the mean F1 Veh control value, in the frontal cortices (Fig. 3A), striata (Fig. 3B), and hippocampi (Fig. 3C) of adolescent DNE offspring and grandoffspring by immunoblotting. Main effects of group (F_2,111=27.13; p=2.5e⁻¹⁰) and region (F_2,111=16.87; p=4.0e⁻⁷) and a significant group x region interaction (F_4,111=4.254; p=0.003) were detected. Compared to F1 Veh mice, F1 NIC and F2 NIC mice have reduced furin levels in the frontal cortices (p=0.004 and p=0.0005, respectively) and striata (p=0.003 and p=0.002, respectively), but not in the hippocampi. Furthermore, furin content positively co-varies with proBDNF proteolysis (r=0.79; p<0.0001, r=0.81; p<0.0001, and r=0.72; p<0.0001, respectively) in the frontal cortices (Fig. 3D), striata (Fig. 3E), and hippocampi (Fig. 3F).

Figure 3. DNE elicits multigenerational furin deficits that covary with impaired proBDNF proteolysis in frontal cortices and striata.

Representative Western blot images and densitometric measurements of furin abundance in frontal cortices (n_F1Veh=14, n_F1NIC=12, and n_F2NIC=15), striata (n_F1Veh=12, n_F1NIC=13, and n_F2NIC=14), and hippocampal (n_F1Veh=14, n_F1NIC=13, and n_F2NIC=13). (A-C) Furin abundance in frontal cortices, striata, and hippocampal. F1 NIC and F2 NIC mice have reduced (A) frontal cortical (p=0.004 and p=0.0005, respectively) and (B) striatal (p=0.003 and p=0.002, respectively) furin content, while (C) hippocampal furin content is unaltered. (D-F) Correlations among furin abundance and proBDNF proteolytic processing in frontal cortices, striata, and hippocampal. Furin abundance positively co-varies with relative nuclear GR localization in (D) frontal cortices (r=0.79; p<0.0001), (E) striata (r=0.81; p<0.0001), and (F) hippocampal (r=0.72; p<0.0001). FCX, frontal cortices; STR, striata; HIPP, hippocampi; GAPDH, glyceraldehyde 3-phosphate dehydrogenase. All data are mean ± S.E.M. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

3.3. DNE elicits multigenerational anomalies in glucocorticoid receptor (GR) activation in the frontal cortices, striata, and hippocampi

HPA axis hypoactivity such as hypocortisolemia occurs in neurodevelopmentally-disordered children and the children of maternal smokers and is thought to mediate the impulsivity displayed by these individuals [58], [57], [71]. Building upon this evidence, our previous work reveals DNE-evoked risk-taking behaviors in first- and second-generation DNE mice, and the rodent literature demonstrates altered GR expression in first-generation DNE progeny [14], [32], [66], [67], [68], [69]. In advancement of this line of research, the current study assessed the hypothesis that DNE precipitates multigenerational perturbations in GR activation. Toward these objectives, we performed immunoblot analyses of total, nuclear, and cytosolic GR abundance and GR (Ser²⁸⁷) phosphorylation as well as relative nuclear GR localization (a proxy for GR activation) in the frontal cortices, striata, and hippocampi of adolescent DNE offspring and grandoffspring.

Alterations in total GR abundance (Figs. 4A, 4E, and 4I, respectively), nuclear GR abundance (Figs. 4B, 4F, and 4J, respectively), cytosolic GR abundance (Figs. 4C, 4G, and 4K, respectively), and relative nuclear GR localization (ratio of nuclear versus total GR abundance) (Figs. 4D, 4H, and 4L, respectively) were assessed (as percentages of the mean F1 Veh control value) in the frontal cortices, striata, and hippocampi of first- and second-generation adolescent DNE mice by comparison to F1 Veh controls. Main effects of group (F_2,105=3.53; p=0.033), region (F_2,105=12.86; p=0.009), and measure (F_3,105=13.84; p=0.007), a significant group x measure interaction (F_6,105=5.49; p=0.036), a significant region x measure interaction (F_6,105=13.56; p=0.003), and a significant group x region x measure interaction F_12,105=3.44; p=0.015) were detected. Compared to F1 Veh controls, there were no alterations in total GR abundance in the frontal cortices, striata, or hippocampi of DNE mice. However, relative to F1 Veh mice, F1 NIC and F2 NIC mice have decreased nuclear GR abundance in the frontal cortices (p=0.02 and p=0.007, respectively) and striata (p=0.003 and p=0.002, respectively). In contrast to F1 Veh mice, F1 NIC and F2 NIC mice have increased cytosolic GR abundance in the frontal cortices (p=0.02 and p=0.003, respectively) and striata (p=0.013 and p=0.015, respectively). Accordingly, F1 NIC and F2 NIC mice have decreased relative nuclear GR localization in the frontal cortices (p=0.002 and p=0.001, respectively) and striata (p=0.001 and p=0.002, respectively) compared to F1 Veh mice. Contrary to the frontal cortices and striata, F2 NIC mice have increased hippocampal nuclear GR abundance (p=0.038), F1 NIC mice have decreased hippocampal cytosolic GR abundance (p=0.021), and both F1 and F2 NIC mice have increased hippocampal relative nuclear GR localization (p=0.045 and p=0.018, respectively).

Figure 4. DNE elicits multigenerational anomalies in glucocorticoid receptor (GR) activation in the frontal cortices, striata, and hippocampi.

Representative Western blot images and densitometric measurements of nuclear and cytosolic GR abundance and relative nuclear GR localization in frontal cortices (n_F1Veh=14, n_F1NIC=10, and n_F2NIC=13), striata (n_F1Veh=13, n_F1NIC=13, and n_F2NIC=13), and hippocampal (n_F1Veh=9, n_F1NIC=8, and n_F2NIC=9). (A) Total GR abundance in frontal cortices. There were no group differences in frontal cortical total GR abundance. (B) Nuclear GR abundance in frontal cortices. F1 NIC and F2 NIC mice have reduced frontal cortical nuclear GR content (p=0.02 and p=0.007, respectively). (C) Cytosolic GR abundance in frontal cortices. F1 NIC and F2 NIC mice have increased frontal cortical cytosolic GR content (p=0.02 and p=0.003, respectively). (D) Relative nuclear GR localization in frontal cortices. F1 NIC and F2 NIC mice have decreased frontal cortical relative GR nuclear localization (p=0.002 and p=0.001, respectively). (E) Total GR abundance in striata. There were no group differences in striatal total GR abundance. (F) Nuclear GR abundance in striata. F1 NIC and F2 NIC mice have reduced striatal nuclear GR content (p=0.003 and p=0.002, respectively). (G) Cytosolic GR abundance in striata. F1 NIC and F2 NIC mice have increased striatal cytosolic GR content (p=0.013 and p=0.015, respectively). (H) Relative nuclear GR localization in striata. F1 NIC and F2 NIC mice have reduced striatal relative GR nuclear localization (p=0.001 and p=0.002, respectively). (I) Total GR abundance in hippocampal. There were no group differences in hippocampal total GR abundance. (J) Nuclear GR abundance in hippocampal. F2 NIC mice have increased hippocampal nuclear GR content (p=0.038). (K) Cytosolic GR abundance in hippocampal. F1 NIC mice have reduced hippocampal cytosolic GR content (p=0.021). (L) Relative nuclear GR localization in hippocampal. F1 NIC and F2 NIC mice have increased hippocampal relative nuclear GR localization (p=0.045 and p=0.018, respectively). FCX, frontal cortices; STR, striata; HIPP, hippocampi; GR, glucocorticoid receptor; Nuc, nuclear GR; Cyt, cytosolic GR; TBP, TATA-binding protein; B-tub, β-tubulin. All data are mean ± S.E.M. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

3.4. Glucocorticoid receptor (Ser²⁸⁷) phosphorylation is unaltered in the frontal cortices, striata, and hippocampi of DNE mice

Given the findings of the present study that DNE precipitates multigenerational alterations in BDNF signaling coupled with deregulation of GR activity, we were intrigued by the notion that BDNF regulates GR activity via BDNF-dependent GR (Ser²⁸⁷) and GR (Ser¹⁵⁵) phosphorylation [87]. Accordingly, we inferred that DNE-induced disruption of BDNF signaling may perturb GR (Ser²⁸⁷) phosphorylation and thereby alter GR activity. To examine this possibility, we quantified (by immunoblotting) total (nuclear + cytosolic) GR (Ser²⁸⁷) phosphorylation (Figs. 5A, 5D, and 5G, respectively), nuclear GR (Ser²⁸⁷) phosphorylation (Figs. 5B, 5E, and 5H, respectively), and cytosolic GR (Ser²⁸⁷) phosphorylation (Figs. 5C, 5F, and 5I, respectively) in the frontal cortices, striata, and hippocampi of adolescent DNE offspring, DNE grandoffspring, and F1 Veh control mice and analyzed these data as percentages of the mean F1 Veh control value. Compared to F1 Veh controls, there were no alterations in total, nuclear, or cytosolic GR (Ser²⁸⁷) phosphorylation in the frontal cortices, striata, or hippocampi of first- or second-generation DNE mice.

Figure 5. Glucocorticoid receptor (Ser²⁸⁷) phosphorylation is unaltered in the frontal cortices, striata, and hippocampi of ONE mice.

Representative Western blot images and densitometric measurements of total, nuclear, and cytosolic GR (Ser²⁸⁷) phosphorylation in frontal cortex (n_F1veh=8, n_F1NIC=8, and n_F2NIC=8), striatum (n_F1Veh=8, n_F1NIC=8, and n_F2NIC=8), and hippocampus (n_F1veh=8, n_F1NIC=8, and n_F2NIC=8). (A) Total GR (Ser²⁸⁷) phosphorylation in frontal cortex. There were no group differences in frontal cortical total GR (Ser²⁸⁷) phosphorylation. (B) Nuclear GR (Ser²⁸⁷) phosphorylation in frontal cortex. There were no group differences in frontal cortical nuclear GR (Ser²⁸⁷) phosphorylation. (C) Cytosolic GR (Ser²⁸⁷) phosphorylation. There were no group differences in frontal cortical cytosolic GR (Ser²⁸⁷) phosphorylation. (D) Total GR (Ser²⁸⁷) phosphorylation in striatum. There were no group differences in striatal total GR (Ser²⁸⁷) phosphorylation. (E) Nuclear GR (Ser²⁸⁷) phosphorylation in striatum. There were no group differences in striatal nuclear GR (Ser²⁸⁷) phosphorylation. (F) Cytosolic GR (Ser²⁸⁷) phosphorylation in striatum. There were no group differences in striatal cytosolic GR (Ser²⁸⁷) phosphorylation. (G) Total GR (Ser²⁸⁷) phosphorylation in hippocampus. There were no group differences in hippocampal total GR (Ser²⁸⁷) phosphorylation. (H) Nuclear GR (Ser²⁸⁷) phosphorylation in hippocampus. There were no group differences in hippocampal nuclear GR (Ser²⁸⁷) phosphorylation. (I) Cytosolic GR (Ser²⁸⁷) phosphorylation in hippocampus. There were no group differences in hippocampal cytosolic GR (Ser²⁸⁷) phosphorylation. FCX, frontal cortices; STR, striata; HIPP, hippocampi; GR, glucocorticoid receptor; pS287, phosphoglucocorticoid receptor (Ser²⁸⁷); TBP, TATA-binding protein; B-tub, β-tubulin. All data are mean ± S.E.M.

3.5. FKBP5 content is unaltered in the frontal cortices, striata, or hippocampi of DNE mice but negatively covaries with relative nuclear GR localization across all groups.

Overexpression of the GR chaperone FKBP5 is implicated in the HPA axis downregulation, GR hypoactivity, and impulsivity observed in neurodevelopmental disorders [58], [57], [71]. Considering this evidence and given that first- and second-generation DNE mice exhibit impulsivity-like risk-taking behaviors accompanied by corticostriatal GR hypoactivity, we posited that DNE may elicit multigenerational upregulation of FKBP5 [32]. Addressing this hypothesis, the present study conducted immunoblot analyses to probe for multigenerational DNE-induced alterations in FKBP5 expression in the frontal cortices, striata, and hippocampi of adolescent F1 NIC and F2 NIC mice by comparison to F1 Veh controls. To this end, alterations in FKBP5 content in the frontal cortices (Fig. 6A), striata (Fig. 6B), and hippocampi (Fig. 6C), presented as percentages of the mean F1 Veh control value, were analyzed. No main effects of or interactions with group or region were detected for FKBP5 abundance. However, FKBP5 abundance negatively covaries with relative nuclear GR localization (r=−0.33; p<0.031, r=−0.58; p<0.0004, and r=−0.81; p<0.0001, respectively) in the frontal cortices (Fig. 6D), striata (Fig. 6E), and hippocampi (Fig. 6F) within and across groups.

Figure 6. FKBP5 content is unaltered in the frontal cortices, striata, or hippocampal of DNE mice but negatively covaries with relative nuclear GR localization across all groups.

Representative Western blot images and densitometric measurements of FKBP5 abundance in frontal cortices (n_F1Veh=13, n_F1NIC=12, and n_F2NIC=14), striata (n_F1Veh=11, n_F1NIC=11, and n_F2NIC=12), and hippocampal (n_F1Veh=11, n_F1NIC=11, and n_F2NIC=13). (A-C) FKBP5 abundance in frontal cortices, striata, and hippocampal. FKBP5 content is unaltered in (A) frontal cortices, (B) striata, and (C) hippocampal of F1 NIC and F2 NIC mice. (D-F) Correlations among FKBP5 abundance and relative nuclear GR localization in frontal cortices, striata, and hippocampal. FKBP5 abundance negatively co-varies with relative nuclear GR localization in (D) frontal cortices (r=−0.33; p<0.031), (E) striata (r=−0.58; p<0.0004), and (F) hippocampal (r=−0.81; p<0.0001). FCX, frontal cortices; STR, striata; HIPP, hippocampi; FKBP5, FK506 binding protein 5; GAPDH, glyceraldehyde 3-phosphate dehydrogenase. All data are mean ± S.E.M.

3.6. DNE elicits multigenerational deficits in basal plasma corticosterone levels

To further assess the intergenerational transmissibility of DNE-evoked HPA axis anomalies, we next evaluated our presumption that downregulation of CORT signaling may contribute to the corticostriatal GR hypoactivity displayed by first- and second-generation DNE mice. To this end, basal plasma CORT levels were quantified by ELISA. Therein, a main effect of group (F_2,35=7.46; p=0.002) was detected for basal plasma CORT levels. F1 NIC and F2 NIC mice have deficient basal plasma CORT levels (p=0.0035 and p=0.014, respectively) relative to F1 Veh controls.

4. Discussion

Toward a more thorough understanding of the neurobiological substrates mediating the association of DNE with neurodevelopmental disorders in children and grandchildren, the current study affords novel insight into the heretofore uncharacterized multigenerational impacts of DNE on frontal cortical, striatal, and hippocampal proBDNF/BDNF expression and balance, furin expression, GR expression and activation, and FKBP5 expression along with basal plasma CORT levels in adolescent mice. A visual distillation of the findings reported herein is provided in Figure 8.

Figure 8. Synopsis of immunoblot results.

Dual-gradient heatmaps depicting the percentage differences relative to F1 Veh control mice in each outcome measure (vertical axis) and brain region (horizontal axis) for F1 NIC (left) and F2 NIC (right) mice. A percentage difference of zero indicates no difference (depicted in gray) relative to F1 Veh control mice, whereas percentage differences of −50% and 50% indicate a 50% decrease (depicted in red) and a 50% increase (depicted in green), respectively, relative to F1 Veh control mice. F1 NIC, first-generation developmental nicotine-exposed adolescent offspring; F2 NIC, second-generation developmental nicotine-exposed adolescent offspring. FCX, frontal cortices; STR, striata; HIPP, hippocampi; BDNF, brain-derived neurotrophic factor; pBDNF, proBDNF; dBDNF, dimeric BDNF; mBDNF, monomeric BDNF; GR, glucocorticoid receptor; pGR, phospho-glucocorticoid receptor (Ser²⁸⁷); FKBP5, FK506 binding protein 5; ^p<0.05; +p<0.01; ‡p<0.001.

The neurotrophin BDNF and its pro-apoptotic precursor proBDNF reciprocally regulate myriad neurodevelopmental, neurodegenerative, neuroinflammatory, synaptodendritic, and neurobehavioral processes, and proBDNF-BDNF imbalance is implicated in the etiology of ADHD, autism, and schizophrenia [36], [37], [38], [39], [43], [44], [104], [105], [104], [105], [106], [107], [108], [109], [110], [111]. Similarly, BDNF deficits have been documented in DNE children and in the frontal cortices and striata of DNE progeny [26], [33], [34], [35], [36], [37]. Based on this evidence, we tested the hypothesis that DNE may shift the proBDNF-BDNF balance toward the accumulation of proBDNF and the deficiency of BDNF. While total BDNF was unaltered, we detected an excess of proBDNF and a proportional dearth of BDNF in the frontal cortices and striata of DNE offspring and grandoffspring. These data are the first to demonstrate that DNE-induced corticostriatal proBDNF-BDNF imbalance is intergenerationally transmissible. Moreover, these findings recapitulate the BDNF deficiency and proBDNF accumulation purported to underpin neurodevelopmental disruption and aberrant neurodegeneration, respectively, in ADHD, autism, and schizophrenia [32], [34], [36], [37], [38], [39], [40], [44], [51], [52], [53], [54], [55], [56]. Preclinical evidence indicates that BDNF and proBDNF dysregulation confer nearly identical ensembles of neurobehavioral abnormalities including synaptoplastic alterations, atypical dendritic morphology, neuroinflammation, neurodegeneration, anxiety- and depression-like behaviors, and cognitive and memory disturbances that are characteristic of neurodevelopmental disorders. Consequently, the results of this study intimate that neurodevelopmental disorders as well as our prior findings of multigenerational neurodevelopmental disorder-like phenotypes in DNE progeny may constitute either a composite effect or an emergent property of co-occurring BDNF deficits and proBDNF accumulation [32], [104], [105], [106], [107], [108], [109], [110], [111]. Likewise, given that BDNF and proBDNF modulate extensively overlapping clusters of neurodevelopmental, neurodegenerative, neurofunctional, neuroinflammatory, synaptoplastic, and behavioral processes, the multigenerational DNE-induced proBDNF-BDNF imbalances reported herein imply a “dual-hit” mechanism whereby the co-aberration of BDNF and proBDNF signaling may additively or synergistically disrupt neurodevelopment, facilitate neurodegeneration, elicit neuroinflammation, alter synaptoplasticity and dendritic spine density, and perturb behavior to an increased extent relative to selective downregulation of BDNF or upregulation of proBDNF alone [43], [44], [104], [105], [106], [107], [108], [109], [110], [111].

From a translational perspective, the findings of this study identify BDNF deficiency and proBDNF accumulation as potential therapeutic targets for the attenuation of DNE-evoked neurobehavioral insults, whereby pharmacological rebalancing of proBDNF and BDNF signaling could mitigate multigenerational phenotypic anomalies attributable to DNE. Since direct pharmacological rescue of BDNF deficits is not presently viable, future studies could evaluate the safety and efficacy of anti-proBDNF antibody therapies, which have been shown to equilibrate aberrant proBDNF signaling, upregulate BDNF, restore neuronal morphology, and normalize aberrant stress-, anxiety-, and depression-related behaviors [104], [105], [112]. Therein, considering that BDNF and proBDNF bilaterally regulate neuronal structure and function as well as behavior, memory, and cognition, the equilibration of proBDNF signaling via anti-proBDNF antibody treatment should exert therapeutic effects on such neurobehavioral phenotypes stemming from both BDNF deficits, proBDNF accumulation, and interactions thereof [104], [105], [106], [107], [108], [109], [110], [111], [112].

In light of the apparent multigenerational impairment of proBDNF proteolysis in DNE mice, we hypothesized that DNE may downregulate the metalloprotease furin, which processes proBDNF to BDNF and is deficient in myriad neurodevelopmental disorders [39], [40], [41], [42], [45], [46], [47], [48], [49], [50]. In support of this hypothesis, results indicated deficient furin content in the frontal cortices and striata of F1 and F2 NIC mice, the same regions wherein impaired proBDNF proteolysis was detected. Lending further support to the presumptive contribution of furin deficits to impaired proBDNF proteolysis in DNE mice, furin abundance positively covaried with proBDNF proteolysis in the frontal cortices, striata, and hippocampi across all groups. To the best of our knowledge, these findings are the first to suggest that DNE elicits multigenerational disruptions of furin-mediated proBDNF proteolysis which recapitulate observations in neurodevelopmental disorders [39], [40], [41], [42], [48], [49], [50]. From a clinical perspective, these findings identify furin deficits as a potential therapeutic target for normalization of proBDNF proteolysis and associated neurodevelopmental insults. To this end, and given that furin expression is negatively regulated by microRNA-24 (miR24), future research could assess the utility of anti-miR24 oligonucleotide therapy as a tool for enhancing furin expression and thereby normalizing proBDNF proteolysis in DNE offspring and grandoffspring [113].

Epidemiological research demonstrates HPA axis hypoactivity, namely hypocortisolemia and deficient GC signaling, in pediatric ADHD patients and DNE children alike [58], [70]. Similarly, DNE is known to alter GR expression in the frontal cortices and striata of first-generation rodent progeny [66]. Building upon this research, the current study examined total, cytosolic and nuclear GR abundance and a proxy measure for GR activation in the frontal cortices, striata, and hippocampi of first- and second-generation adolescent DNE mice. While total GR content was unaltered, reductions in nuclear GR abundance accompanied by proportional increases in cytosolic GR content were identified in the frontal cortices and striata of first- and second-generation DNE mice. The associated decreases in relative nuclear GR localization are indicative of corticostriatal GR hypoactivity. These findings recapitulate the HPA axis hypoactivity reported in neurodevelopmentally-disordered children, the children of maternal smokers, and first-generation DNE rodents, and further suggest that these phenomena are transmitted to at least second-generation DNE offspring [58], [66], [70]. Conversely, the present study also demonstrated elevated nuclear GR accumulation in tandem with reduced cytosolic GR content in the hippocampi of first- and second-generation DNE mice, demonstrative of hippocampal GR hyperactivity. Importantly, as the hippocampi mediate negative feedback on the HPA axis and GC signaling, the apparent multigenerational GR hyperactivity detected in the hippocampi would be expected given the GR hypoactivity observed in the frontal cortices and striata of adolescent DNE offspring and grandoffspring [114], [115]. In aggregate, these data are the first to demonstrate that DNE evokes multigenerational GR dysfunction analogous to the HPA axis deregulation observed in ADHD, autism, and schizophrenia as well as DNE children [57], [58], [70], [71]. By extension, these findings imply that corticostriatal GR hypoactivity and/or hippocampal GR hyperactivity may underlie the impulsivity observed in neurodevelopmental disorders, DNE children, and rodent models thereof.

In light of recent research demonstrating that BDNF regulates GR activity via BDNF-dependent GR (Ser²⁸⁷) and GR (Ser¹⁵⁵) phosphorylation, we posited that DNE may disrupt GR activity in part via alterations in BDNF-dependent GR (Ser²⁸⁷) phosphorylation [87]. Contrary to this hypothesis, no alterations in total, nuclear, or cytosolic GR (Ser²⁸⁷) phosphorylation were detected in the frontal cortices, striata, or hippocampi of first- or second-generation adolescent DNE mice. This result implies that DNE-induced multigenerational GR hypoactivity does not result from aberrant BDNF-dependent GR (Ser²⁸⁷) phosphorylation. However, a cognate role for aberrant BDNF-dependent GR (Ser¹⁵⁵) phosphorylation in DNE-evoked multigenerational GR hypoactivity remains possible, as we were unable to assess GR (Ser¹⁵⁵) phosphorylation due to the unavailability of an antibody suitable for this application.

An alternative explanation for the multigenerational GR hypoactivity in the frontal cortices and striata of DNE mice is DNE-mediated upregulation of FKBP5, a chaperone protein that buffers the activation, dimerization, and nuclear translocation of cytosolic GRs and contributes to HPA axis dysregulation in neurodevelopmental disorders such as ADHD, autism, and schizophrenia [57], [58], [71]. Contrary to this hypothesis, no DNE-evoked alterations in FKBP5 content were observed in any brain region assayed. However, correlational analyses indicated that, as expected given its biological function, FKBP5 abundance negatively covaries with relative nuclear GR localization in the frontal cortices, striata, and hippocampi both within and across all groups of mice. Although the results of the current study do not implicate FKBP5 in the GR hypoactivity observed in DNE offspring and grandoffspring, the substanital positive correlation identified between GR activation and FKBP5 expression warrants future research to assess whether blockade of FKBP5 via the novel inhibitor SAFit2 normalizes GR activity in DNE progeny and, if so, whether this treatment reverses any behavioral anomalies displayed by these mice [116].

Another potential explanation for DNE-induced multigenerational corticostriatal GR hypoactivity and hippocampal GR hyperactivity is hypocorticosteronemia. Accordingly, we next evaluated basal plasma CORT levels in DNE offspring and grandoffspring. Results are the first to demonstrate that DNE indeed elicits multigenerational hypocorticosteronemia. Notably, this hypocorticosteronemia is reminiscent of the hypocortisolemia characteristic of neurodevelopmentally-disordered and DNE children [58], [70]. Moreover, as DNE does not perturb GR phosphorylation at a BDNF-regulated residue (Ser²⁸⁷) or alter FKBP5 expression, DNE-evoked hypocorticosteronemia may primarily underlie the multigenerational corticostriatal GR hypoactivity identified herein and, by extension, may mediate the behavioral aberrations that we previously reported in DNE mice [32].

Cumulatively, the novel findings reported herein reveal that DNE elicits multigenerational proBDNF-BDNF imbalance which correlates with furin downregulation in the frontal cortices and striata, corticostriatal GR hypoactivity accompanied by hippocampal GR hyperactivity, and basal plasma CORT deficits. By extension, DNE-induced multigenerational downregulation of furin appears to disrupt proBDNF proteolysis and thereby shift the proBDNF-BDNF equilibrium toward the accumulation of proBDNF. Moreover, while the causal relationship between hypocorticosteronemia and imbalanced corticostriatal-hippocampal GR activity in DNE offspring and grandoffspring remains indeterminate, these findings nevertheless imply that DNE aberrantly enhances HPA axis negative feedback in an FKBP5-independent manner. Taken together, the multigenerational neurotrophic and neuroendocrine perturbations engendered by DNE recapitulate core features of neurodevelopmental disorders including but not limited to ADHD, autism, and schizophrenia. However, further research is warranted to elucidate the mechanistic bases of these findings. To this end, future rodent studies should explore epigenetic processes whereby DNE may facilitate multigenerational transmission of neurodevelopmental disorder-like phenotypes.

Figure 7. DNE elicits multigenerational deficits in basal plasma corticosterone levels.

Basal plasma corticosterone levels measured by enzyme-linked immunosorbent assay (n_F1Veh=14, n_F1NIC=12, and n_F2NIC=12). F1 NIC and F2 NIC mice have deficient basal corticosterone levels (p=0.0035 and p=0.014, respectively). All data are mean ± S.E.M. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.