DNA methylome perturbations: an epigenetic basis for the emergingly heritable neurodevelopmental abnormalities associated with maternal smoking and maternal nicotine exposure

Abstract

Maternal smoking during pregnancy is associated with an ensemble of neurodevelopmental consequences in children and therefore constitutes a pressing public health concern. Adding to this burden, contemporary epidemiological and especially animal model research suggests that grandmaternal smoking is similarly associated with neurodevelopmental abnormalities in grandchildren, indicative of intergenerational transmission of the neurodevelopmental impacts of maternal smoking. Probing the mechanistic bases of neurodevelopmental anomalies in the children of maternal smokers and the intergenerational transmission thereof, emerging research intimates that epigenetic changes, namely DNA methylome perturbations, are key factors. Altogether, these findings warrant future research to fully elucidate the etiology of neurodevelopmental impairments in the children and grandchildren of maternal smokers and underscore the clear potential thereof to benefit public health by informing the development and implementation of preventative measures, prophylactics, and treatments. To this end, the present review aims to encapsulate the burgeoning evidence linking maternal smoking to intergenerational epigenetic inheritance of neurodevelopmental abnormalities, to identify the strengths and weaknesses thereof, and to highlight areas of emphasis for future human and animal model research therein.

Evidence from human and animal model research indicates that maternal smoking confers a constellation of neurodevelopmental anomalies which appear to undergo intergenerational transmission via a putative epigenetic mechanism.

Introduction

Approximately 10% of women in the United States smoke traditional cigarettes during pregnancy, and 14% use electronic cigarettes during pregnancy [1–3]. The incidences of maternal smoking (MS) are higher still in certain other countries [4]. Alarmingly, epidemiological studies report that, despite an absence of supporting evidence, the majority of those surveyed misperceive that electronic cigarettes are a safer and healthier surrogate for traditional cigarettes, and this misunderstanding is most pervasive among women of reproductive age [5–9]. Contrary to majority opinion, MS of traditional or electronic cigarettes or consumption of any tobacco- or nicotine-containing product during the gestational period or during the gestational and postpartum lactational periods results in the concurrent exposure of three generations to nicotine, namely the mother herself, her unborn or newborn child, and (via exposure of her unborn or newborn child’s germline) her grandchild(ren).

MS is associated with numerous fetal consequences including premature birth, low birth weight, birth defects, and sudden infant death syndrome [10–14]. Notably, an estimated 20 to 40% of all low-birthweight births are linked to MS, and low birthweight is the strongest predictor of mortality and morbidity in neonates [15]. Furthermore, 22% of sudden infant deaths in the United States are directly attributable to MS, and the risk of sudden infant death positively covaries with number of cigarettes consumed daily [16]. Circa 1983, approximately 9% ($272 million USD) of annual expenditures on neonatal intensive care in the United States were attributable to MS [15]. In 2020, the estimated aggregate costs of MS totaled $1.16 billion USD [17].

As will be detailed in the ensuing text, MS appears to be associated with heightened risk for neurodevelopmental disorders (NDDs) including attention-deficit/hyperactivity disorder (ADHD), autism, and schizophrenia in children and possibly grandchildren as well as a plethora of discrete transdiagnostic features of NDDs such as abnormal brain and neuronal development, behavioral anomalies such as hyperactivity and impulsivity, aberrant circadian rhythmicity, predisposition to smoking and other forms of substance abuse, alterations in catecholamine and cholinergic neurotransmitter systems, neurotrophic dysfunction, hypothalamic–pituitary–adrenal (HPA) axis deregulation, and epigenetic changes such as DNA methylome perturbations. Given the associations of MS and possibly grandmaternal smoking (GMS) with NDDs and the aforementioned transdiagnostic features thereof, it is also worthwhile to highlight the economic burdens attributable to NDDs, as a portion of these costs may be indirectly related to MS. Therein, the annual economic burdens of ADHD, autism, and schizophrenia were estimated at $12.76 billion USD in 2018, $250 billion USD circa 2014, and $155 billion USD in 2013, respectively, with the economic burden of autism projected to nearly double to $460 billion USD by 2025 [18–21].

While the economic burdens of MS as well as neurodevelopmental deficits and disorders are tremendous, the human and quality of life impacts of these phenomena are immeasurably greater. As such, it is of vital importance to better understand the specific brain and behavioral consequences as well as the broader public health implications of MS and GMS. Toward this objective, the present review endeavors to synthesize the states of the human and animal model literatures on the impacts of MS and GMS (or maternal nicotine exposure and grandmaternal nicotine exposure in the case of animal model research) on neurodevelopmental outcomes as well as the emerging evidence of a mechanistic role for DNA methylome perturbations therein.

State of the human literature on the neurodevelopmental impacts of MS

Association of MS with risk for NDDs

A substantial body of research associates MS with enhanced risk for NDDs including ADHD, autism, and schizophrenia in children, and there may be a causal pathway between MS and later neurodevelopmental problems [22–36]. However, some of the associations between MS and NDDs, namely autism and schizophrenia, no longer hold once controlling for gestational age and other confounding familial variables [27–30, 37–41]. Furthermore, while the literature supporting an association between MS and ADHD is relatively robust, there are also studies which suggest that MS may not be a direct risk factor independent of familial factors for ADHD [23, 26, 31–34, 42]. These mixed findings are in line with recent debate as to whether associations between MS and offspring NDD liability are indeed causal or instead due to unmeasured confounding variables [43]. For instance, genetic liability for both smoking and NDDs exists and may link MS and child outcomes without a direct teratogenic effect of smoking per se [22, 40]. In addition, there may be other familial variables, often conceptualized as more environmental in nature (e.g. parental education and parenting), that could also contribute to (and thus confound) the relationship between MS and offspring outcomes such as NDD diagnoses. In conjunction with the possible confounding of the associations between MS and NDD diagnoses by genetic and familial variables, heterogeneity in psychiatric diagnostic classification, which stems from the fact that clinical diagnoses for NDDs and other psychiatric conditions are not objective, discrete, or standardized measurements, undermines the utility and validity of clinical diagnoses as primary outcome measures for studies assessing the associations between MS and neurodevelopmental outcomes [44]. Accordingly, it is arguably more informative and rigorous to explore the relationships of MS with specific neurodevelopmental deficits and transdiagnostic markers of NDDs rather than inherently confounded NDD diagnoses.

Associations of MS with discrete neurobehavioral deficits

Epidemiological studies examining the relationships between MS and various transdiagnostic features of NDDs suggest a robust association of MS with a multitude of neurodevelopmental and behavioral alterations across the lifespan, with the majority of such studies focusing on the developmental periods of infancy, childhood, and adolescence [22, 45]. Therein, MS is associated with delayed infant psychomotor and mental developmental scores as measured by the Bayley Scales of Infant Development, atypical auditory brainstem responses, composite measures of neurodevelopmental impairment, and decreased motor scores between ages 18 and 21 months [37, 46–48]. Investigations of the relationship of MS with childhood and adolescent outcomes are considerably more common than those for infants and generally report the most consistent findings. For instance, MS is associated with intellectual impairment at age 4, inattention at age 6, hyperactivity at age 8, physical aggression during early childhood as well as middle childhood and adolescence, and increased externalizing disorders such as conduct disorder and ADHD [22–23, 26, 33–34, 49–56]. Many cognitive impairments are also associated with MS including deficits in sustained attention, response inhibition, overall cognitive function and memory, receptive language, verbal learning and design memory, problem solving, speech and language, school performance, and auditory processing as well as impulsivity [57–64]. Research also reveals dose–response relationships wherein MS-related relative risks for neurodevelopmental abnormalities increase with amount smoked, results which suggest the existence of vulnerable periods of fetal development [65–67]. Several studies have also reported that, similar to the associations of neurodevelopmental deficits and disorders with a proclivity to smoke, a tendency to start smoking earlier, and a penchant to smoke more, the children of maternal smokers are more likely to smoke, initiate smoking earlier, and smoke more [1, 12, 31, 68–77]. Importantly, many of the phenotypic aberrations associated with MS in the aforementioned studies are broadly characteristic of NDDs including ADHD, autism, and schizophrenia [78–93].

Associations of MS with abnormal brain and neuronal development

MS alters the structure of the developing human nervous system in part by perturbing neuronal differentiation, is associated with long-term neurological morbidity, results in lower proportions of neurons in the developing dorsolateral prefrontal cortex, is linked to reduced brain volume, reduced cortical gray matter volume, and cortical thinning in children, and is associated with hypoactivity of the ventral striata during reward anticipation in adolescents [94–100]. Taken together, these findings link MS with indicators of broad neuroteratogenicity and developmental encephalopathy in children and are reminiscent of the decreased total brain volume in ADHD and schizophrenia as well as the diminished cortical volume and thickness and impaired neuronal differentiation in ADHD, autism, and schizophrenia [101–107].

Associations of MS with cholinergic system anomalies

Nicotinic acetylcholine receptors (nAChRs) are expressed in a neurodevelopmental stage-specific manner under tight spatiotemporal regulation and function as key mediators of neuronal differentiation, maturation, and morphology as well as synaptogenesis in the developing brain [108–112]. Moreover, cholinergic signaling mediates NDD-related phenotypes including locomotor activity and impulsivity across the lifespan [113–117]. Notably, atypical nAChR expression patterns have been documented in the brains of the children as well as the placentas of maternal smokers [111, 118–120]. These findings concur with studies linking altered nAChR expression patterns and nAChR polymorphisms to ADHD, autism, and schizophrenia [121–130].

Associations of MS with catecholamine system perturbations

The catecholamines dopamine (DA) and norepinephrine (NE) are important mediators of many neurodevelopmental processes and exert modulatory influences on an array of neurobehavioral phenotypes relevant to NDDs as well as the neurodevelopmental consequences of MS [131–134]. Infants of maternal smokers have diminished norepinephrine content in cord blood at birth, and the children of maternal smokers exhibit highly significant associations between NE transporter (NET) polymorphisms and ADHD diagnosis, cognitive-behavioral measures relevant to ADHD, and response to methylphenidate treatment [135–136]. Furthermore, MS is associated with specific subtypes of ADHD in genetically susceptible children possessing variations in the DA transporter (DAT1) and/or DA receptor D4 (DRD4) loci [137]. Similarly, hyperactivity-impulsivity and oppositional behaviors are associated with a DAT polymorphism exclusively in children that were exposed to MS [138]. These limited findings suggest an emerging role of catecholamine system perturbations and polymorphisms in the effects of MS on neurodevelopmental outcomes in children and are congruent with the altered corticostriatal NE and DA neurotransmission, NE and DA receptor expression, and DAT and NET function as well as the DA and NE receptor and DAT and NET polymorphisms linked to NDDs including ADHD, autism, and schizophrenia [139–175].

Associations of MS with neurotrophic dysfunction

Neurotrophins such as brain-derived neurotrophic factor (BDNF) play vital roles in pregnancy wherein they mediate placental growth and development as well as offspring neurodevelopment [176–178]. Aberrant BDNF signaling is implicated in NDDs and pathological features including disrupted neuronal development and synaptogenesis, impaired dendritic arborization, altered synaptic plasticity, and aberrant synaptic pruning accompanied by learning and memory impairments, emotional dysregulation, and aberrant stress responsivity [179–192]. BDNF dysfunction may facilitate neurobehavioral deficits in part via excessive inhibition of inhibitory neurons resulting in disinhibition of and reduced capacity to downregulate excitatory circuits [193–195]. Notably, research suggests that the children of maternal smokers are subject to deregulation of BDNF signaling and display abnormal BDNF concentrations in umbilical cord sera [196–197]. Furthermore, BDNF moderates associations between MS and offspring behavioral disorders [195]. This nascent literature associates MS with alterations in BDNF signaling which are linked to neurodevelopmental deficits in offspring and that are reminiscent of the atypification of BDNF signaling implicated in NDDs including ADHD, autism, and schizophrenia [179–182].

Associations of MS with HPA axis dysregulation

HPA axis dysfunction disrupts neurodevelopment and neuronal maturation, perturbs synaptic plasticity, elicits hyperactivity, impulsivity, and aberrant stress responsivity, and promotes emotional deregulation [24, 198–204]. Multiple studies have identified markers of HPA axis dysfunction in neurodevelopmentally disordered children including hypocortisolemia and altered glucocorticoid (GC) receptor (GR) expression [205–209]. Similarly, the children of maternal smokers also display hypocortisolemia and aberrant GR expression [210–213]. Taken together, these findings indicate that MS confers HPA axis dysfunction characteristic of NDDs.

Emerging intergenerational consequences of MS

Alarmingly, contemporary research associates GMS with enhanced risk for ADHD diagnosis and increased autism symptoms [42, 214]. These findings imply that the heightened vulnerabilities to neurodevelopmental deficits and disorders associated with MS may be transmissible across multiple offspring generations. However, it should be noted that the preponderance of research examining the intergenerational influences of MS has focused on outcomes related to physical health such as asthma, while only a handful of studies have assessed neurodevelopmental deficits/disorders in relation to MS. Moreover, rigor and reproducibility in intergenerational human subjects research are extremely difficult to achieve given the amount of time and other resources needed to conduct such longitudinal family study designs coupled with common hurdles related to participant recruitment and retention. In the cases of MS and GMS studies, these inherent obstacles to intergenerational human subjects research often force investigators to make various inferences in light of incomplete or insufficient longitudinal data, such as assumptions about whether or not there was additional smoke exposure at other points in the pathway between the grandmother to the grandchild, including whether the mother smoked during her pregnancy. It therefore remains indeterminate whether the associations of GMS with neurodevelopmental deficits and disorders reported thus far will persist after controlling for confounders, or whether intergenerational influences of smoking during pregnancy affect other aspects of neurodevelopment.

In retrospect, there is extensive evidence from the human literature that associates MS with a variety of neurobehavioral outcomes in children, but animal model studies have yielded much of what is known about the neurobiological effects of MS on the developing brain. Therein, while the human research associating MS with cholinergic dysregulation, catecholamine dysfunction, BDNF dysregulation, and HPA axis aberrations is relatively sparse, the findings of studies characterizing maternal nicotine exposure (MNE) animal models of MS presented hereafter provide robust evidence of NDD-like anomalies in nAChR expression and function, DA and NE neurotransmission, BDNF signaling, and glucocorticoid hormone signaling as well as glucocorticoid receptor activity in the offspring of female rodents subjected to nicotine exposure paradigms designed to recapitulate MS patterns. Furthermore, a burgeoning body of animal model literature indicates that many MNE-induced NDD-like neurobiological perturbations undergo intergenerational co-transmission along with behavioral alterations that recapitulate phenotypes exhibited by the children of maternal smokers and neurodevelopmentally disordered children alike.

State of the animal model literature on the neurodevelopmental impacts of MS

Paradigms for animal modeling of the intergenerational impacts of MS

The information that animal models can provide to human research is significant [215]. For example, animal models enable intergenerational study designs that incorporate a specific controlled dose of a specific drug over a defined interval in a controlled setting. This experimental precision affords valuable insights into the exposure-, dosage-, and developmental stage-specific impacts of prenatal exposures on discrete behavioral phenotypes and accompanying biological changes at high resolution.

Translational validity is of the utmost importance in the design of animal models of the children and grandchildren of maternal smokers. To this end, a consensus paradigm for modeling the intergenerational impacts of MS utilizes a nicotine (or tobacco smoke) exposure regimen informed by the findings of epidemiological research that the majority of maternal smokers initiate regular smoking well in advance of conception and continue smoking throughout pregnancy and nursing [216, 217]. Therein, nicotine or tobacco smoke exposure of founder (F0) female dams (MNE) commences prior to mating with drug-naïve sires and continues until weaning of first-generation (F1) offspring. By this design, the F1 generation embryo/fetus and the maternal germ cells are subjected to direct nicotine or tobacco smoke exposure. Subsequent breeding of F1 generation progeny that were maternally exposed to nicotine or tobacco smoke during development (which model the children of maternal smokers and are referred to as MNE animals hereafter) with drug-naïve mates yields second-generation (F2) offspring that were exposed to nicotine or tobacco smoke exclusively via the maternal germline (which model the grandchildren of maternal smokers and are referred to as GMNE animals hereafter). Successive breeding of GMNE (F2) progeny with drug-naïve mates yields third-generation (F3) offspring that are entirely devoid of direct exposure to nicotine or tobacco smoke (which model the great grandchildren of maternal smokers). Reiteratively, MNE animals model the children of maternal smokers, GMNE animals model the grandchildren of maternal smokers, and F3 progeny model the great grandchildren of maternal smokers. As such, GMNE animal models provide insight into the intergenerational or multigenerational impacts of MS, while F3 generation progeny yield insight into the transgenerational effects of MS.

Importantly, despite its utility for modeling neurodevelopmental outcomes in the children and grandchildren of maternal smokers, the current MNE and GMNE animal model literature is limited in many facets. Perhaps most notably, whereas in the human literature the prevalences of traditional and electronic cigarette consumption among maternal smokers are comparable, an overwhelming majority of the MNE (F1) animal model studies referenced in the manuscript utilized pure nicotine exposure paradigms (which are more comparable to electronic than traditional cigarettes) rather than tobacco smoke exposure paradigms (which are more comparable to traditional than electronic cigarettes), and all GMNE (F2) animal model studies exclusively utilized nicotine exposure paradigms, a scenario that ultimately hinders the translatability of findings from animal model studies to the human condition. Owing further to this imbalance in the animal model literature, limited insights can be derived by evaluation of data from maternal tobacco smoke exposure animal model studies independent of data from MNE animal model studies. Nevertheless, an amalgamation of MNE (but not GMNE) animal model investigations which utilized different exposure paradigms and routes of administration including combustible tobacco inhalation chambers, aerosolized/vaporized nicotine apparatuses, oral nicotine delivery, nicotine injections (e.g. intraperitoneal or subcutaneous), and nicotine infusions (e.g. osmotic minipumps) have reported multiple NDD-related phenotypes that are common across study designs.

Intergenerational behavioral impacts of MNE

Consistent with behavioral anomalies documented in the children of maternal smokers, animal model research demonstrates that MNE mice exhibit hyperactivity, aberrant rhythmicity of activity, impulsivity-like and risk-taking behaviors, and inattention [50–52, 218–226]. These results recapitulate behavioral phenotypes shared among ADHD, autism, and schizophrenia patients [78–93]. Moreover, MS animal model studies also reveal reduced nicotine sensitivity, increased nicotine self-administration, and enhanced oral nicotine consumption and preference [219, 224, 227]. These findings are concordant with the proclivity to smoke, earlier initiation of smoking, and tendency to smoke more documented in the children of maternal smokers and linked to both clinical diagnoses and discrete symptoms (in the absence of a diagnosis) of ADHD, autism, and schizophrenia [1, 12, 31, 68–77]. Notably, contemporary research indicates that MNE-evoked hyperactivity is transgenerationally transmissible, while perturbed rhythmicity of activity, risk-taking behaviors, and enhanced oral nicotine consumption and preference appear to be intergenerationally transmissible [224]. These findings are consistent with human studies linking GMS to ADHD and autism in grandchildren [42, 214]. Altogether, the findings of animal model studies indicate that MNE elicits an array of behavioral anomalies that are consistent with NDDs including ADHD, autism, and schizophrenia and that appear to be transmissible to subsequent generations.

Interestingly, treatment with methylphenidate, a first-line medication for NDDs such as ADHD, rescues NDD-like phenotypes including hyperactivity, altered rhythmicity of activity, and risk-taking behaviors in both MNE and GMNE rodents [222–224, 228]. These findings support the validity of MNE and GMNE rodents as models for the intergenerationally transmissible impacts of MS. Similarly, voluntary oral nicotine consumption is elevated in MNE and GMNE rodents and exerts normalizing effects on hyperactivity and aberrant rhythmicity of activity that span both generations [224]. These results support the self-medication hypothesis for the comorbidity of smoking in NDDs, which postulates that the increased prevalence of smoking in NDDs such as ADHD and schizophrenia is attributable to therapeutic-like effects of nicotine [73, 224, 229–232].

Impacts of MNE on brain and neuronal development

MNE and GMNE rodents display several markers of impaired brain and neuronal development including premature neuronal differentiation, atypical neuronal morphology, neuronal apoptosis, disrupted neuronal maturation, aberrant neuronal migration, and reduced cortical volume and thickness [221–222, 233–239]. These results mirror the reduced total brain volume, decreased cortical volume, thickness, and neuronal content, and aberrant neurodifferentiation shared between neurodevelopmentally disordered children and the children of maternal smokers [94–107]. Notably, whether these developmental encephalopathies and indicators of neuroteratogenicity exhibited by MS animal models are intergenerationally transmissible remains indeterminate.

Intergenerational impacts of MNE on the cholinergic system

Animal model research indicates that the neuroteratogenic effects of MNE are attributable in part to MNE-evoked atemporal nAChR stimulation early in development [233–237]. Additional studies reveal alterations in both the expression and function of a variety of nAChR subtypes throughout the fetal, adolescent, and adult corticostriatothalamic circuitry of MNE rodents, suggesting that the impacts of MNE on corticostriatothalamic cholinergic signaling persist across the lifespan [238–247]. These findings mimic the perturbed nAChR expression patterns reported in the placenta and the children of maternal smokers, recapitulate the misconnectivity and cholinergic anomalies documented within the corticostriatothalamic neurocircuitry of neurodevelopmentally disordered individuals, and support the implication of corticostriatothalamic and cholinergic dysfunction in NDD-related phenotypes including locomotor hyperactivity, inattention, and impulsivity [111, 113, 115, 118–130, 248–258]. In furtherance of these findings, GMNE animal models exhibit an array of abnormalities in both the expression and function of multiple nAChR subpopulations within the frontal cortices, striata, and thalami, implying that MNE-induced alterations in the corticostriatothalamic cholinergic system are intergenerationally transmissible [224].

Intergenerational impacts of MNE on the catecholamine system

Concordant with the corticostriatal catecholamine system dysfunction and DA and NE receptor and transporter polymorphisms implicated in NDDs as well as the catecholamine system perturbations and polymorphisms implicated in the effects of MS on neurodevelopmental outcomes in children, MNE and GMNE animals exhibit a variety of anomalies in DA and NE signaling [135–175, 222, 224, 228, 233, 259–262]. Therein, adolescent and adult MNE rodents display decreased DA synthesis and release, altered DAT and NET as well as DA receptor expression, impaired DAT function, and impaired DA turnover within the corticostriatal circuitry [222, 224, 228, 233, 259–262]. In support and expansion of these findings, additional research demonstrates intergenerational transmission of dopaminergic and DAT dysfunction in animal models of the children and grandchildren of maternal smokers [139–165, 224]. These results are again consistent with the catecholamine system abnormalities documented in ADHD, autism, and schizophrenia and thus further support the notion that MNE-evoked NDD-like phenotypes are heritable.

Intergenerational impacts of MNE on neurotrophic signaling

Comparable to the perturbations of neurotrophic signaling documented in ADHD, autism, and schizophrenia patients and the implication of BDNF deregulation in neurodevelopmental deficits in the children of maternal smokers, animal model studies reveal aberrant BDNF signaling within the fetal, adolescent, and adult mesocorticolimbic circuitry as well as the juvenile brainstem of MNE rodents [179–183, 195–197, 259, 253–267]. Indicative of the intergenerational transmissibility thereof, recent research reveals that MNE and GMNE rodents exhibit deficiencies in frontal cortical and striatal BDNF content coupled with proportional accumulations of proBDNF, the proapoptotic peptide precursor of BDNF which functions in opposition to BDNF both during development and across the lifespan [185, 268–269]. Furthermore, the accumulations of proBDNF detected in MNE and GMNE rodents co-occur and are inversely correlated with downregulation of furin, a metalloprotease which catalyzes proteolytic conversion of proBDNF to BDNF and is downregulated in NDDs and thought to mediate proBDNF-BDNF imbalance therein [182, 184, 270–277]. These findings suggest that MNE and GMNE induce intergenerational transmission of NDD-like proBDNF-BDNF imbalance which appears to be driven by impairment of furin-mediated proBDNF proteolysis stemming from furin downregulation [268]. Collectively, these results provide strong evidence of neurotrophic dysfunction in MNE and GMNE animal models of the children and grandchildren of maternal smokers, respectively.

Intergenerational impacts of MNE on HPA axis function

Analogous to the HPA axis dysfunction such as hypocortisolemia and deficient GC signaling documented in neurodevelopmentally disordered children and the children of maternal smokers, MNE and GMNE rodents exhibit hypocorticosteronemia, aberrant GR expression in the frontal cortices and striata, and perturbed GR activity in the frontal cortices, striata, and hippocampi, suggesting that MNE-induced NDD-like HPA axis dysfunction is intergenerationally transmissible [24, 198, 208–213, 268, 278–281].

Overview of DNA methylation and the environmental epigenetics of MS

DNA methylation is a mechanism for the epigenetic regulation of gene expression wherein methyl moieties are covalently bound to nucleotide bases within DNA molecules. In turn, these methylated nucleotides can directly modulate the transcriptional accessibility of DNA sequences and/or indirectly influence chromatin architecture via interactions with histone modifiers.

Environmental epigenetics is a burgeoning field of study examining how environmental exposures such as MS affect epigenetic processes such as DNA methylation, particularly during critical and sensitive periods of development (i.e. the prenatal period) [282]. Therein, during the prenatal period methylation patterns of the germline and somatic cell lineages are beginning to be established [283–285]. These epigenetic marks are involved in modulating functional pathways and are key for the health and development of the embryo. Direct exposure to toxins, such as MS, is associated with changes in offspring methylation patterns in a variety of tissues [286–287]. Thus, prenatal smoke exposure at critical points of fetal development may exert important and persistent effects on DNA methylation which in turn influence gene expression and phenotypes later in life. As such, DNA methylation profiles may have utility as biomarkers of MS, as estimators of risk for behavioral deficits and disease, and as therapeutic targets [288–290]. In conjunction with its critical roles in brain and behavioral phenotypes during development and across the lifespan, the DNA methylome also mediates intergenerational phenotypic transmission via the germline.

As will be expounded in the subsequent passages, a plurality of the MS-associated as well as MNE- and GMNE-induced behavioral and neurobiological alterations reported in the aforementioned studies co-occurs with DNA methylome perturbations. Considering the implication of epigenetic changes in the induction by environmental exposures as well as the intergenerational transmission of a broad spectrum of brain and behavioral phenotypes, this co-transmission of DNA methylome and neurobehavioral anomalies in the children of maternal smokers as well as MNE and GMNE animal models suggests a putative mechanism wherein DNA methylome changes are substrates for the induction and intergenerational transmission of MS-associated as well as MNE- and GMNE-induced neurodevelopmental abnormalities. While the present transmittal focuses on DNA methylome alterations as epigenetic bases for these neurodevelopmental abnormalities, it is important to note that such phenotypes are likely also mediated by impacts on other epigenetic processes such as histone modifications.

State of the human literature on the DNA methylomic impacts of MS

DNA methylation mediates neurodevelopment, neuroplasticity, and synaptogenesis, neurodegenerative processes, learning, memory, and cognition, anxiety, stress responsivity, and emotional regulation, locomotor activity and impulsivity/risk-taking, the circadian rhythmicity of behavior, and cholinergic, dopaminergic, neurotrophic, and HPA axis signaling [179–180, 198, 208–209, 235, 272, 291–318]. As such, DNA methylome alterations may contribute to neurobehavioral deficits including the hyperactivity, impulsivity, inattention, and circadian anomalies, the disruption of brain and neuronal development, the malfunction of cholinergic and catecholamine systems, and the deregulation of BDNF and HPA axis signaling in neurodevelopmentally disordered individuals as well as the children of maternal smokers and/or animal models thereof. In support of this possibility, NDDs including ADHD, autism, and schizophrenia are associated with DNA methylome alterations [313, 319–332]. For instance, DNA methylation levels at birth are inversely related to symptom severity in pediatric ADHD patients, and DAT1 promoter methylation mediates hyperactivity and impulsivity in ADHD [236, 303, 319]. As will be detailed in the ensuing text, MS is also associated with DNA methylome perturbations, which may constitute a mechanistic basis for the increased incidences of neurodevelopmental deficits and disorders in the children and possibly the grandchildren of maternal smokers [98, 302, 333–338].

Associations of MS with DNA methylome perturbations in placental tissue

Both epigenome-wide association studies (EWAS) and gene-specific methylation studies have yielded significant associations between MS and placental methylation patterns. Summarily, DNA methylomic studies of placental tissue associate MS with aberrant global DNA methylation patterns as well as differential methylation of specific genes involved in brain and neuronal development, neuronal differentiation, synaptogenesis, and proliferation of various cell types within the central nervous system [335]. Candidate gene studies yielded associations between MS and methylation of genes involved in the metabolism of the potentially carcinogenic compounds found in cigarette smoke such as AHRR (CYP1A1) as well as the serotonin and glucocorticoid receptor subtypes HTR2A and NR3C1, respectively (Table 1) [339–341]. Notably, MS is associated with altered placental methylation of the NR3C1 promoter which mediates concomitant hypocortisolemia detected in corresponding salivary samples [341–342].

Table 1.

Glossary of gene symbols, gene names, and encoded protein functions for selected genes exhibiting differential methylation signatures in placental tissues, cord blood, and whole blood. Encoded protein functions were excerpted from Entrez Gene summaries

Symbol	Name	Encoded Protein Function
AHRR	Aryl Hydrocarbon Receptor Repressor	Implicated in the regulation of cell growth and differentiation. Exhibits nuclear receptor and sequence-specific double-stranded DNA binding activities. Functions as a feedback modulator of AhR by repressing AhR-dependent gene expression.
NR3C1	Nuclear Receptor Subfamily 3 Group C Member 1	Functions as a transcription factor that binds to glucocorticoid response elements as well as a regulator of other transcription factors. Inactive form is found in the cytoplasm but translocates to the nucleus upon activation by glucocorticoid ligands. Involved in inflammation, proliferation, and differentiation.
DNMT3A	DNA Methyltransferase 3a	De novo DNA methyltransferase responsible for the establishment of DNA methylation patterns in embryos, the methylation of most imprinted loci in germ cells, and the establishment of DNA methylation patterns during development. Can actively repress transcription through the recruitment of HDAC activity.
MBD3	Methyl-CpG Binding Domain Protein 3	Acts as transcriptional repressor and plays a role in gene silencing. Does not bind to DNA by itself. Binds to DNA with a preference for sites containing methylated CpG dinucleotides (in vitro). Binds to a lesser degree DNA containing unmethylated CpG dinucleotides. Recruits histone deacetylases and DNMTs.
SLC6A3	Solute Carrier Family 6 (Neurotransmitter Transporter, Dopamine), Member 3	Member of the sodium- and chloride-dependent neurotransmitter transporter family. Terminates the action of dopamine by its high affinity sodium-dependent reuptake into presynaptic terminals. Variants associated with idiopathic epilepsy, attention-deficit hyperactivity disorder, dependence on alcohol and cocaine, susceptibility to Parkinson’s disease, and protection against nicotine dependence.
DRD1	Dopamine Receptor D1	Most abundant dopamine receptor in the central nervous system. Stimulates adenylyl cyclase and activates cyclic AMP-dependent protein kinases. Regulates neuronal growth and development and mediates some behavioral responses.
MAOB	Monoamine Oxidase B	Catalyzes the oxidative deamination of biogenic and xenobiotic amines and has important functions in the metabolism of neuroactive and vasoactive amines in the central nervous system and peripheral tissues. MAOB preferentially degrades benzylamine and phenylethylamine. Involved in dopamine metabolic processes.
CHRNA5	Cholinergic Receptor Nicotinic Alpha 5 Subunit	Member of the superfamily of ligand-gated ion channels that mediate fast signal transmission at synapses. Contribute to acetylcholine receptor activity. Involved in regulation of synaptic vesicle exocytosis and response to nicotine. Localizes to acetylcholine-gated channel complexes and dopaminergic synapses.
CHRNB4	Cholinergic Receptor Nicotinic Beta 4 Subunit	Exhibits acetylcholine receptor activity and acetylcholine-gated cation-selective channel activity. Involved in signal transduction and synaptic transmission. Localizes to acetylcholine-gated channel complex.
BDNF	Brain-Derived Neurotrophic Factor	During development, promotes the survival and differentiation of selected neuronal populations of the peripheral and central nervous systems. Participates in axonal growth, pathfinding and in the modulation of dendritic growth and morphology. Major regulator of synaptic transmission and plasticity at adult synapses in many regions of the CNS.
GDNF	Glial Cell Derived Neurotrophic Factor	Highly conserved neurotrophic factor involved in negative regulation of apoptotic processes; positive regulation of gene expression, and nervous system development, and survival and differentiation of dopaminergic neurons
FURIN	Furin (Paired Basic Amino Acid Cleaving Enzyme)	Calcium-dependent serine endoprotease that proteolytically activates different proprotein substrates traversing the secretory pathway. Involved in regulation of protein turnover and metabolism as well as regulation of transforming growth factor beta1 production and proteolytic conversion of proBDNF to mature BDNF.
DLGAP2	DLG Associated Protein 2	Molecular adaptor activity and protein domain specific binding activity. Involved in regulation of postsynaptic neurotransmitter receptor activity. Localizes to dendritic spines, glutamatergic synapses, and postsynaptic density. Knock-out confers hyperactivity, aggression, impaired reversal learning, decreased dendritic spine density and mESPC amplitude, synaptopathies, and enhanced paired-pulse ratio.
NRP2	Neuropilin 2	Exhibits semaphorin receptor activity. Localizes to axon; glutamatergic synapse; and integral component of postsynaptic membrane. Involved in nervous system development, neural crest cell migration, and axonal guidance.

Associations of MS with DNA methylome perturbations in cord blood

Cord blood is a surrogate tissue that is widely utilized for EWAS of the effects of environmental exposures on the epigenome. Notably, the cord blood DNA methylome is shared between mother and child and in some instances may relate to phenotypes shared between maternal smokers and their children. However, it warrants consideration that the cord blood DNA methylome has not been shown to substantially relate to the DNA methylomes of specific cells and cell types within brain regions implicated in NDDs.

Studies examining cord blood from the children of maternal smokers intimate that MS confers global DNA hypomethylation [343]. The Pregnancy and Childhood Epigenetics Consortium (PACE) conducted a meta-analysis across 13 birth cohorts across the USA and Europe (N = 6685) examining the association between MS and newborn cord blood methylation at over 450 000 CpG sites [337]. Results suggested that 6073 CpGs were considered statistically significant when using FDR correction, while 568 CpGs remained significantly associated after Bonferroni correction for multiple testing [337]. Fewer statistically significant findings were found for any MS than for sustained MS, suggesting that comprehensive measurement of MS is an important consideration, with more sustained smoking over the entire pregnancy offering more power when compared to any smoking (yes/no) for studying epigenetic effects of MS [337].

Consistent with prior reports examining both MS-exposed offspring and adult smokers, the top finding of the PACE meta-analysis of MS and newborn cord blood methylation was for AHRR (aryl hydrocarbon receptor repressor), a gene that is involved in the detoxification of chemicals found in tobacco smoke (i.e. the xenobiotic response) and which acts as a feedback inhibitor of the aryl hydrocarbon receptor (AHR) [333, 337, 344–349]. Accordingly, AHRR demethylation and resultant upregulation may constitute an adaptive cellular response to the accumulation of chemicals found in cigarette smoke [345]. Furthermore, the methylation levels of several AHRR CpGs in cord blood of infants of maternal smokers are inversely related to cotinine levels, suggestive of a dose–response relationship between MS and AHRR demethylation [333]. Critically, GMS is also associated with AHRR demethylation in cord blood [349], findings which are rendered even more remarkable in light of the extensive involvement of AHR in phenotypes shared among neurodevelopmentally disordered individuals, the children of maternal smokers, and MNE and GMNE rodents. Therein, AHRs are expressed under tight spatiotemporal control within the developing brain, broadly regulate neuronal development, neuronal morphology, and synaptic maturation, neuronal migration and dendritic outgrowth, and neuronal differentiation, modify locomotor activity and the circadian rhythmicity thereof, mediate cholinergic influences on neurodevelopment by regulating acetylcholinesterase activity across various neurodevelopmental stages and processes, modulate DA synthesis by regulating tyrosine hydroxylase expression and thus indirectly regulates NE synthesis, influence BDNF expression in cortical neurons, and tune HPA axis function and the circadian rhythmicity thereof by regulating the activation of glucocorticoid receptors as well as the synthesis and release of cortisol [350–365]. In addition, AHR is subject to polymorphisms associated with increased severity of autism symptoms, and AHRR has expansive intragenic CpG islands which exhibit differential methylation that is dose-dependently associated with exacerbation of symptom severity in ADHD [366–368].

Interestingly, the aforementioned findings of decreased AHRR methylation in cord blood have not been replicated in placental tissue or saliva cells (345); however, this is not surprising given that AHRR is expressed at low levels in the placenta, while cord blood cells show high interindividual variation [345, 369]. Taken together, these results highlight the importance of examining the DNA methylomic impacts of MS across tissue types and considering differences in normative gene expression therein.

Additional evidence from studies of cord blood from maternal smokers reveals aberrant locus-specific DNA methylation signatures at CpGs located within genes (Table 1) including DNMT3A, DNMT1, and MBD3, SLC6A3 (DAT) and MAOB, CHRNA1 and CHRNA5, GDNF, NGFR, and FURIN, DLGAP2 (also known as SAPAP2), and NRP2 that encode proteins with functions (Table 1) relevant to phenotypes characteristic of the children of maternal smokers, neurodevelopmentally disordered children, and animal models thereof [337, 370–373]. For instance, DLGAP2 is known to mediate synaptic organization and neuronal signaling and is associated with both schizophrenia (370) and autism (371, 372), while NRP2 is expressed in neural crest cells, appears to be necessary for proper neuronal migration, and is associated with autism [337, 370–373]. Moreover, the differential methylation of SLC6A3 as well as FURIN are consistent with the DAT dysfunction as well as the furin downregulation and related proBDNF-BDNF imbalance, respectively, documented in MNE and GMNE animal models.

Complementing the locus-specific DNA methylome perturbations identified in the children and grandchildren of maternal smokers, functional annotation enrichment analyses have provided valuable insight into the overarching impacts of MS on the cord blood DNA methylome. As exemplified in Table 2, research reveals differential methylation signatures at CpG sites mapped to loci enriched for multiple functional annotations relevant to the neurodevelopmental impacts of MS as well as the neuropathologies of ADHD, autism, and schizophrenia [337]. Therein, MS is associated with differential methylation of gene sets enriched for GO biological process annotations such as behavior (GO:0007610), brain development (GO:0007420), regulation of nervous system development (GO:0051960), regulation of membrane potential (GO:0042391), neurotrophin signaling pathway (GO:0038179), response to steroid hormone (GO:0048545), and sequence-specific DNA binding (GO:0043565) [337]. These findings relate MS to DNA methylome alterations which disproportionately impact genes involved in a broad spectrum of biological processes with particular relevance to phenotypic alterations exhibited by the children of maternal smokers and animal models thereof as well as the pathosymptomatology of NDDs.

Table 2.

Exemplary enriched functional annotations of differentially methylated genes in cord blood from maternal smokers and their relation to neurodevelopmental disorder-like phenotypes in the children of maternal smokers and animal models thereof. GO, gene ontology; DMGs, differentially methylated genes. [337]

GO Identifier:	GO Biological Process Annotations of Differentially Methylated Genes in MS Cord Blood	n (DMGs)	Related Phenotypes Associated with MS in Humans and Animal Models
0007610	Behavior	33	Neurobehavioral Deficits
0007611	Learning or memory	14	Neurobehavioral Deficits
0050890	Cognition	15	Neurobehavioral Deficits
0050795	Regulation of behavior	17	Neurobehavioral Deficits
0040017	Positive regulation of locomotion	36	Neurobehavioral Deficits
0048511	Rhythmic process	12	Neurobehavioral Deficits
0007420	Brain development	47	Developmental Encephalopathy
0030900	Forebrain development	29	Developmental Encephalopathy
0021987	Cerebral cortex development	9	Developmental Encephalopathy
0021761	Limbic system development	8	Developmental Encephalopathy
0021766	Hippocampus development	6	Developmental Encephalopathy
0030902	Hindbrain development	11	Developmental Encephalopathy
0051960	Regulation of nervous system development	47	Neuroteratogenicity
0050767	Regulation of neurogenesis	44	Neuroteratogenicity
0045664	Regulation of neuron differentiation	39	Neuroteratogenicity
0043005	Neuron projection	30	Neuroteratogenicity
0001764	Neuron migration	10	Neuroteratogenicity
0016358	Dendrite development	13	Neuroteratogenicity
0050770	Regulation of axonogenesis	13	Neuroteratogenicity
0070997	Neuron death	20	Neuroteratogenicity
0030522	Intracellular receptor signaling pathway	27	Cholinergic and Catecholine Dysfunction
0034765	Regulation of ion transmembrane transport	18	Cholinergic and Catecholine Dysfunction
0042391	Regulation of membrane potential	20	Cholinergic and Catecholine Dysfunction
0034765	Regulation of ion transmembrane transport	18	Cholinergic and Catecholine Dysfunction
0019199	Transmembrane receptor protein kinase activity	20	Cholinergic and Catecholine Dysfunction
0048015	Phosphatidylinositol-mediated signaling	28	Cholinergic and Catecholine Dysfunction
0038179	Neurotrophin signaling pathway	34	Neurotrophic Dysfunction
0048011	Neurotrophin TRK receptor signaling pathway	34	Neurotrophic Dysfunction
1,901,653	Cellular response to peptide	28	Neurotrophic Dysfunction
0019838	Growth factor binding	17	Neurotrophic Dysfunction
0090287	Regulation of cellular response to growth factor stimulus	18	Neurotrophic Dysfunction
0046879	Hormone secretion	21	HPA Axis Dysregulation
0009914	Hormone transport	21	HPA Axis Dysregulation
0048545	Response to steroid hormone	19	HPA Axis Dysregulation
0048732	Gland development	19	HPA Axis Dysregulation
0080135	Regulation of cellular response to stress	33	HPA Axis Dysregulation
0051169	Nuclear transport	37	HPA Axis Dysregulation
0043565	Sequence-specific DNA binding	61	DNA Methylome Perturbations
0000975	Regulatory region DNA binding	61	DNA Methylome Perturbations
0003714	Transcription corepressor activity	29	DNA Methylome Perturbations
0003713	Transcription coactivator activity	33	DNA Methylome Perturbations
0000785	Chromatin	29	DNA Methylome Perturbations
0001047	Core promoter binding	11	DNA Methylome Perturbations
0017053	Transcriptional repressor complex	11	DNA Methylome Perturbations

Associations of MS with DNA methylome perturbations in offspring whole blood

Comparable to the aforementioned findings in cord blood from maternal smokers, studies in whole blood from the children of maternal smokers show differential methylation across multiple CpG sites mapped to noteworthy specific genes (Table 1) such as AHRR, CHRNB4, SYTL1, DRD1, and BDNF [196, 374–378]. These studies also mapped differential methylation patterns in whole blood from the children of maternal smokers to gene sets enriched for functional annotations of numerous NDD-related processes including axonogenesis (GO:0050373), axon guidance (GO:0007411), neuron projection guidance (GO:0097485), and neuron projection development (GO:0031175) as well as locomotion (GO:0040011). Noteworthy pathway annotations of said data included dopamine receptor mediated signaling pathway (P05912), axon guidance mediated by semaphorins (P00007), and nicotine pharmacodynamics pathway (P06587) [196, 374–378]. Building upon this research, another in whole blood from the children of maternal smokers concluded that MS-associated DNA methylome alterations mediate the effect of MS on schizophrenia related outcomes [379].

Associations of MS with DNA methylome perturbations in other tissues

Minimal research has examined the effects of MS on DNA methylation patterns in tissues other than placental cells, umbilical cord blood, and whole blood. Nevertheless, research characterizing leukocyte samples from the children of maternal smokers demonstrates associations between MS and global DNA hypomethylation as well as demethylation of repetitive elements [380, 381].

A study examining postmortem fetal brain tissue from maternal smokers indicates that MS perturbs DNA methylation in the developing fetal brain and confers co-occurring deficits in prefrontal cortical neuronal content [98]. Taken together, these results suggest that exposures experienced throughout the course of life (from the prenatal period to adulthood) and accompanying epigenetic changes may be one important biological pathway by which MS affects later behavior and developmental outcomes.

Cumulatively, human research efforts to date demonstrate that MS confers a constellation of DNA methylome anomalies which are congruent with findings in NDDs. However, notwithstanding the aforementioned study associating GMS with differential AHRR methylation in grandchildren, there is a paucity of human research on the intergenerational effects of MS on DNA methylation in relation to both gene expression and behavior in children and especially grandchildren.

State of the animal model literature on the DNA methylomic impacts of MS

The available animal model research on the intergenerational as well as germline epigenomic impacts of nicotine is disproportionately more expansive for paternal versus MNE. For instance, several studies have traced intergenerational and transgenerational transmission of global and locus-specific DNA methylome anomalies from paternal sperm to offspring and grandoffspring brain that co-occur with a variety of NDD-like behavioral and neurobiological perturbations [382–388]. Moreover, the breadth and volume of human research characterizing the DNA methylomic impacts of MS far exceed that of MNE and GMNE animal model research. The direction of this disparity is quite unusual considering that in most instances the animal model research on a given topic tends to mature/progress more rapidly than corresponding human research, as is the case for the comparative states of the human and animal model literatures on the neurobiological and neurobehavioral impacts of MS. Despite the paucity of previous research investigating heritable DNA methylome perturbations in MNE and GMNE animal models, a modicum of pioneering studies in this realm have produced compelling and insightful results that warrant postliminary research.

MNE and GMNE rodents exhibit DNA methylome perturbations that co-occur with and in some cases have been shown to directly mediate or correlate with phenotypes encompassed by transdiagnostic features of NDDs including hyperactivity, risk-taking behaviors, anomalous circadian rhythmicity of activity, enhanced nicotine consumption and preference, aberrant DA system and nAChR function, BDNF deregulation and proBDNF-BDNF imbalance, and hypocorticosteronemia as well as dysregulation of GR activity [183, 259, 264, 278, 389–390]. The de novo DNA methylase DNMT3A and the DNA demethylase TET2 reciprocally regulate global and locus-specific DNA methylation patterns [391–392]. As such, downregulation of DNMT3A and/or upregulation of TET2 could underlie the abovementioned multigenerational global DNA methylome deficits and by extension numerous other phenotypic aberrations documented in MNE and GMNE rodents. In support of this possibility, research demonstrates the co-occurrence of global DNA hypomethylation in the frontal cortices and striata of MNE and GMNE mice, downregulation of DNMT3A in the frontal cortices, striata, and hippocampi of MNE mice and in the frontal cortices and striata of GMNE mice, and downregulation DNMT3B in the hippocampi of MNE mice [393–394]. Considering the absence of detectable changes in TET2 expression in MNE and GMNE mice, corticostriatal DNMT3A deficits may primarily mediate the intergenerational transmission of corticostriatal global DNA hypomethylation conferred by MNE [224]. Taken together with the results of human studies associating MS with differential methylation of DNMT1 and DNMT3A, these data imply that MS could elicit broad DNA methylome perturbations in part by interfering with key mediators of the establishment and maintenance of methylation marks across the epigenome.

In light of research implicating nicotine-evoked upregulation of ovarian nAChRs in the reduced successful pregnancy rates, decreased oocyte viability, and increased oocyte abnormalities documented in female smokers, it has recently been suggested that the intergenerational impacts of MNE may be facilitated in part by nicotine-evoked, embryonic ovarian nAChR-mediated DNA methylome perturbations in oocytes [395–399]. This hypothesis is supported by contemporary research demonstrating that oocytes from MNE mice exhibit global DNA hypomethylation, diminished expression of multiple DNMT subtypes, and altered chromatin configuration as well as compromised integrity and viability [400]. Taken together with the aforementioned findings of global DNA hypomethylation and DNMT3A downregulation in the frontal cortices and striata of MNE and GMNE mice, these data demonstrate DNA methylome perturbations in the female germline of MNE mice which mimic those reported in the frontal cortices and striata of GMNE mice [224]. By extension, these results suggest a plausible mechanism for the intergenerational impacts of MNE wherein MNE induces overexpression of ovarian nAChRs in MNE embryos, which in turn facilitates DNMT-mediated DNA methylome alterations in the oocytes (germline) thereof that are subsequently inherited by and manifest in the brains of GMNE progeny. Suggestive of a similar MNE-evoked intergenerational epigenetic cascade in the male lineage, recent research reveals impaired maturation of gonocytes to spermatogonia in male MNE mice [401]. These findings imply that the male germline is also susceptible to and may mediate the intergenerational transmission of the DNA methylomic impacts of MNE. Altogether, the aforementioned evidence suggests that intergenerational transmission of DNA methylome perturbations via the female and possibly the male germline may be a primary mechanism underlying the heritability of NDD-like brain and behavioral phenotypes documented in animal models of the children and grandchildren of maternal smokers.

Epigenetic factors beyond global and locus-specific DNA methylation patterns, deregulation of DNA methyltransferases, and the germline transmissibility thereof may also be involved in the heritability of phenotypic aberrations elicited by maternal exposure to nicotine and other drugs of abuse [402–403]. Two such factors are methyl-CpG-binding protein-2 (MeCP2) and histone deacetylase 2 (HDAC2), the former of which (MeCP2) binds to methylated DNA and recruits the latter (HDAC2) as part of a cascade that links the DNA methylome and chromatinome and ultimately leads to heterochromatization and epigenetic silencing of methylated loci [404–405]. As such, it is noteworthy that downregulation of both MeCP2 and HDAC2 co-occurs with global DNA hypomethylation and DNMT3A downregulation in both MNE and GMNE mice [393]. These findings are especially compelling in light of research linking alterations in MeCP2 and HDAC2 expression to disruption of neuronal development, synaptogenesis, and synaptoplasticity, perturbation of BDNF signaling, learning and memory deficits, aberrant stress responsivity, and HPA axis dysregulation [296, 405–421].

Cumulatively, the findings of animal model research to date imply that MNE pervasively alters the DNA methylome and consequently the broader epigenome in developing progeny and the germline thereof, thereby contributing to the induction and intergenerational transmission of myriad NDD-like phenotypic aberrations. While paternal nicotine exposure is rather robustly linked to transgenerational brain and behavioral consequences attributable in part to epigenomic alterations, at present there is insufficient evidence to assess the transgenerational transmissibility of phenotypic aberrations conferred by MNE.

Discussion

Interpretations and implications

As visualized in Figure 1, human and animal model research characterizing the children and grandchildren of maternal smokers has, to date, yielded varying strengths of evidence for associations of MS and GMS with transdiagnostic features of NDDs including developmental encephalopathies, markers of neuroteratogenicity, cognitive-behavioral deficits, catecholamine and cholinergic system anomalies, neurotrophic dysfunction, HPA axis dysregulation, and global as well as locus-specific DNA methylome perturbations. Summarily, evidence linking MS to an array of NDD-related neurobehavioral alterations is strong across both the human and animal model literatures, but the preponderance of evidence extending these findings to GMS is derived from GMNE animal models. Moreover, while a few human studies tentatively associate MS but not GMS with various NDD-related neurobiological anomalies, relatively thorough characterizations of neurobiological anomalies in MNE and GMNE animal models yield considerable evidence of multigenerational transmission of MNE-induced nAChR, dopaminergic, and DAT dysfunction, BDNF-proBDNF imbalance, and HPA axis downregulation including hypocorticosteronemia and GR hypoactivity. The human literature affords substantial evidence associating both global and locus-specific DNA methylome alterations with MS but almost entirely lacks evidence indicating whether GMS similarly impacts the DNA methylome. Conversely, the animal model literature provides significant evidence that MNE precipitates multigenerational global DNA methylome deficits but lacks sufficient epigenomic resolution to evaluate the intergenerational transmissibility of locus-specific DNA methylome anomalies identified in the human literature on MS.

Figure 1.

Survey of human and animal model research on the intergenerational impacts of maternal smoking. Dual-gradient heatmaps depicting arbitrary measures of the strength of evidence (based on quantity of studies) for various categories of neurodevelopmental anomalies from human studies of maternal smoking (MS) and grandmaternal smoking (GMS) as well as animal model studies of maternal nicotine exposure (MNE) and grandmaternal nicotine exposure (GMNE).

Despite the abovementioned weaknesses, the cumulative evidence from human and animal model studies links MS and to a lesser extent GMS to an ensemble of brain and behavioral phenotypes consistent with transdiagnostic features of NDDs and suggests a putative intergenerational epigenetic mechanism thereof wherein nicotine from MS aberrantly stimulates embryonic neuronal and embryonic ovarian nAChRs and thereby elicits self-propagating biological and epigenomic alterations in the developing brain and primordial germ cells, respectively, which in turn precipitate an array of neurodevelopmental deficits in offspring and enable germline epigenetic transmission of such phenotypes to grandoffspring.

Future directions

Toward the resolution of a pressing void in the human literature, future human studies should endeavor to delineate whether and to what extent maternal and GMS are associated with NDD-related neurobiological perturbations such as cholinergic and catecholamine system dysfunction, disruption of neurotrophic signaling, and HPA axis hypofunction. Furthermore, given the implication of DNA methylome anomalies in the NDD-related brain and behavioral phenotypes exhibited by the children of maternal smokers, multigenerational human studies should be conducted to assess the transmissibility of MS-related DNA methylomic alterations to the grandchildren of maternal smokers. In addition to evaluating the intergenerational impacts of MS on the DNA methylome, longitudinal human studies capable of measuring within-individual changes in DNA methylation in a variety of tissues over time should be conducted to garner complementary insights into the intragenerational plasticity of DNA methylation [422–423]. Importantly, these and other future human studies would benefit greatly from considering and directly evaluating the differential impacts of exposure windows, dosage, genotype, and other often-overlooked variables.

On the other hand, future animal model research is warranted to address the need for high-resolution, epigenome-wide profiling of locus-specific differential DNA methylation signatures across discrete brain regions and distinct cell types, the results of which will elucidate the gene-, cell type-, and brain regional-specificity of MNE-evoked multigenerational epigenetic alterations and the relevance to NDDs thereof. Subsequent animal model studies should also assess the transgenerational transmissibility of NDD-related phenotypic aberrations conferred by MNE and the possible roles of DNA methylome perturbations therein.

It is important to acknowledge that little is known about how animal research in the intergenerational arena translates to the human condition or about how human and animal studies can best account for complex human variables such as polysubstance use, variable dosing, differing modes of exposure (for instance traditional versus electronic cigarette consumption), and environmental as well as genetic risk factors. As such, in addition to advancing conventional epidemiological and animal model lines of inquiry, future research should implement translational epigenomic study designs that coalesce human subject and animal model experiments in an effort to better address key questions concerning the intergenerational transmissibility of MS-related epigenomic perturbations and consequent transcriptomic alterations as well as the relationships thereof with NDD-like brain and behavioral phenotypes. Notably, translational epigenomic research is also better equipped to inform the development of novel treatments and prophylactics intended to partially ameliorate or even prevent the intergenerational neurodevelopmental consequences of MS.

From a public health perspective, prevention and cessation of MS are of the utmost importance. Nevertheless, for maternal smokers who are unwilling or unable to quit, emerging evidence suggests that dietary supplementation may confer modest mitigation of or protection against the impacts of MS on the DNA methylome and, by extension, the brain and behavior. For instance, maternal vitamin C supplementation has shown partial efficacy toward the prevention of MS-related DNA methylome perturbations in a randomized clinical trial, and there is reasonable evidence that maternal and offspring folate supplementation may confer similar benefits [424–428]. However, these findings are preliminary and more research in this domain is necessary. As such, care should be taken not to misinform maternal smokers about the possible benefits of dietary interventions. Ultimately, abstinence is the only means of preventing the deleterious intergenerational impacts of MS.

Author contributions

JB, LY, VK, and JS contributed to the preparation and editing of the manuscript.

Conflict of interest

JB, LY, VK, and JS declare no potential or actual conflicts of interest.

Data availability

No new data were generated or analyzed in support of this manuscript.