All roads lead to inflammation: Is maternal immune activation a common culprit behind environmental factors impacting offspring neural control of breathing?

Abstract

Despite numerous studies investigating how prenatal exposures impact the developing brain, there remains very little known about how these in utero exposures impact the life-sustaining function of breathing. While some exposures such as alcohol and drugs of abuse are well-known to alter respiratory function, few studies have evaluated other common maternal environmental stimuli, such as maternal infection, inhalation of diesel exhaust particles prevalent in urban areas, or obstructive sleep apnea during pregnancy, just to name a few. The goals of this review article are thus to: 1) highlight data on gestational exposures that impair respiratory function, 2) discuss what is known about the potential role of inflammation in the effects of these maternal exposures, and 3) identify less studied but potential in utero exposures that could negatively impact CNS regions important in respiratory motor control, perhaps by impacting maternal or fetal inflammation. We highlight gaps in knowledge, summarize evidence related to the possible contributions of inflammation, and discuss the need for further studies of life-long offspring respiratory function both at baseline and after respiratory challenge.

INTRODUCTION

The consequences of prenatal exposures on the health of the fetus have become better appreciated in the last decade. In utero exposures to many maternal environmental stimuli ranging from viral infections to recreational drug use including tobacco, alcohol, opioids and cocaine are now well-accepted to have long-term effects on adult disease, physiology and neurological function (Caputo et al., 2016; Fregosi and Pilarski, 2008; Scott-Goodwin et al., 2016; Song et al., 2002; Vivekanandarajah et al., 2019). While considerable evidence exists regarding the detrimental effects of other maternal exposures to stimuli such as infection or pollution during pregnancy on offspring behaviors such as cognitive function, social interaction, and anxiety (Conway and Brown, 2019; Estes and McAllister, 2016; Scola and Duong, 2017; Solek et al., 2018), there remains comparatively little known about how these exposures impact the development of the respiratory neural control system, enabling regulation of a function vital for life itself, despite the effects that many of these stimuli have on the adult respiratory system. Since nicotine exposure is arguably the best studied drug of abuse with regard to consequences on the offspring respiratory system, and it has been comprehensively reviewed elsewhere (Campos et al., 2009; Fregosi and Pilarski, 2008; Vivekanandarajah et al., 2019), we will not discuss it further here.

At birth, the infant respiratory system must be sufficiently developed in order to survive. There are many CNS regions responsible for healthy respiratory motor control including those located within the brainstem which send projections to motor neurons that innervate critical muscles involved in breathing. For example, the hypoglossal motor nucleus is located near the midline of the medulla and is responsible for innervating the upper airway and tongue muscles (Pilarski et al., 2012). The preBötzinger complex, another respiratory control region, is responsible for inspiratory rhythm generation and contains glutamatergic, glycinergic, and GABAergic inputs (Smith et al., 2013). Additionally, the preBötzinger complex comes online during rodent embryonic development at gestational day (GD)15 (mouse) and GD17 (rat), time points that coincide with the appearance of fetal breathing movements in utero (Greer and Martin-Caraballo, 2017; Smith et al., 1991; Thoby-Brisson and Greer, 2008; Thoby-Brisson et al., 2005). In humans, fetal breathing movements are detected as early as 11 weeks of gestation, suggesting that this is the time when the preBötzinger complex is initiated during human fetal development (Kotecha, 2000). Indeed, these fetal breathing movements continue to increase in prevalence until late gestation (between weeks 38–39), when the fetus spends roughly 30% of its time performing fetal breathing movements (Patrick et al., 1980). This part-time operation suggests that human respiratory neural control pathways are still maturing and fine-tuning themselves even into late gestation, creating a wide period of developmental time during which the fetal respiratory motor control system may be more susceptible to prenatal exposures. Accordingly, in humans, fetal breathing movements fluctuate following maternal meals, hypoxia, and maternal alcohol or nicotine consumption, and in rodent models, medullary nuclei including the preBötzinger complex are vulnerable to modulation by various in utero exposures such as maternal hypoxia, opioids, and stress (Fournier et al., 2013; Ji et al., 2015; Kotecha, 2000; Thoby-Brisson and Greer, 2008), all of which can influence inflammation and cytokine levels in the mothers and fetuses(Bronson and Bale, 2014; Chan et al., 2015; Gur et al., 2017; Hamilton et al., 2007; Jellema et al., 2013; Johnson et al., 2018; Neri et al., 2013; Sadowska et al., 2015; Wood and Keller-Wood, 2019).

Inflammation is a common response to stimuli ranging from foreign pathogen invasion to an acute peripheral injury. Despite the commonality of inflammation and its well-established relationship with asthma (Bel and Ten Brinke, 2017), its impact on respiratory neural control has been mostly studied in adults (reviewed in (Hocker et al., 2017)). In adults, inflammation initiated by exposure to gram negative bacterial lipopolysaccharide (LPS) (Huxtable et al., 2013; Huxtable et al., 2011) or the viral mimetic polydI:dC (Hocker and Huxtable, 2019) inhibits a well-studied form of serotonin-dependent phrenic motor neuroplasticity (phrenic long-term facilitation; pLTF) 24 hours after injection. Similarly, inflammation induced by exposure to a single night of chronic intermittent hypoxia (8 hrs, 15 episodes/hr) also blocks pLTF in adults, in a manner dependent on inflammatory pathway signaling (Huxtable et al., 2015). However, comparatively little is known about how inflammation impacts the newborn respiratory control system. Recent research indicates that a single dose of LPS administered during a critical period of postnatal CNS development (day 4) can attenuate phrenic motor facilitation following acute intermittent hypoxia in adult males (Hocker et al., 2019), indicating that a single bout of acute inflammation in early life can create life-long alterations to mechanisms underlying respiratory neural control. Although studying the effects of prenatal inflammation on the control of breathing is less well studied (Johnson et al., 2018), the consequences of prenatal inflammation on behavioral and psychological outcomes have been extensively investigated for decades.

Thus, the purpose of this review is to examine and summarize the literature on what is currently known about the effects of various maternal prenatal exposures (both intentional and unintentional) on the development of the respiratory neural control system in offspring, and to examine the possible contributions of inflammation on these endpoints. We will highlight current gaps in knowledge and identify some of many underappreciated in utero fetal exposures that could impact the development of the offspring respiratory control system. We begin by discussing offspring impairments in respiratory control caused by prenatal exposures to well-known drugs of abuse including alcohol, and illicit drugs such as opiates, cocaine and marijuana. In addition to the direct effects of the drugs themselves on fetal outcomes, we propose the possible contribution of disrupted inflammation during pregnancy in the effects of these stimuli, which may further complicate and exacerbate their physiologic effects on offspring. In contrast, we also make an argument for maternal exposures that are involuntary in nature, and that are surprisingly less well-studied in terms of their influence on offspring respiratory control and function; these include maternal exposure to infection, diesel exhaust particles in air pollution, and sleep disordered breathing, all of which have strong impacts on inflammation and on maternal respiratory function. For the purposes of this literature review, it is important to note that most in vivo studies in which drugs of abuse are experimentally modeled, the dams are often exposed both during pregnancy and postpartum. Thus, offspring exposure frequently continues throughout the critical postnatal neurodevelopmental period as well (usually via breastmilk transfer). Accordingly, although the goal of this review is to examine prenatal exposures that impact the neural control of breathing, available literature necessitates recognizing the caveat that some reported effects of fetal exposure may not be solely due to the prenatal exposure period.

ALCOHOL

Prenatal alcohol exposure has many detrimental effects on fetal development, including impacts on the heart, liver, lungs, kidney, endocrine, and nervous systems (Caputo et al., 2016; Gauthier and Brown, 2017). Certainly, epidemiological studies have shown alcohol consumption during pregnancy to be a significant risk factor for SIDS (Iyasu et al., 2002; McDonnell-Naughton et al.; O’Leary et al., 2013), with 15% of women continuing to drink into their third trimester, 6% of whom were classified as binge drinkers (Dukes et al., 2017). Epidemiologic data indicate that in utero exposure to alcohol impairs fetal, neonatal and childhood lung function (reviewed in (Gauthier and Brown, 2017)), suggesting important effects on offspring breathing. However, even though low dose and infrequent alcohol exposures can cause behavioral sequelae in rodents and humans (recently reviewed in (Noor and Milligan, 2018)), most rodent studies seem to model maternal binge drinking, where high and sporadic ethanol levels are introduced to the dams that readily pass the placenta and can be detected in fetal tissues within minutes of maternal exposure (Zorzano and Herrera, 1989); fetal alcohol levels reach the same as those in the mother within 1–2 hours (Burd et al., 2012). Additional studies on lower alcohol dose and frequency are vital to better understand how this more modest type of alcohol consumption impacts aspects of offspring physiology, especially in light of more recent studies not always accurately interpreted by the lay press indicating that “light” drinking during early pregnancy is not harmful to the offspring, and which is contrary to prevailing medical advice (Mamluk et al., 2017; McCarthy et al., 2013).

Whereas most animal studies modeling alcohol exposure during development focus on physical facial features, neurocognitive, or behavioral aspects of fetal alcohol spectrum disorder (FASD), developmental alcohol exposure may also be associated with functional consequences on the offspring respiratory control system. Indeed, a study of 33 children with FASD demonstrated marked sleep disturbances, a small subset of whom also had mild sleep disordered breathing (Chen et al., 2012; Goril et al., 2016), suggesting that fetal alcohol may disrupt brain regions responsible for respiratory neural control. Compared to nicotine, respiratory function following maternal alcohol consumption has been poorly studied in both humans and rats. However, a seminal study exposing rat dams to ethanol (10% v/v in drinking water) from 4 weeks pre-conception until weaning showed decreased respiratory frequency in the offspring at 3–4 weeks of age, primarily due to an increased duration of the expiratory phase (Dubois et al., 2006). Whether the greatest or most important alcohol-induced developmental detriments on respiratory function are due to the developmental effects of prenatal or postnatal alcohol exposure is still unknown.

A maternal diet consisting of 20% calories from ethanol during gestational days 7–15 in mice decreases fetal serotonergic neuron migration from the raphe as well as their maturation (Zhou et al., 2001). These serotonergic neurons project to and innervate regions responsible for respiratory control (Barker et al., 2009), suggesting a potential neurochemical mechanism for respiratory disruptions following gestational alcohol exposure. Since serotonin controls critical respiratory responses to various environmental stimuli (Pilowsky, 2014), alterations in serotonergic function can have significant impacts on respiratory neural control. In this regard, 5-HT_2AR mRNA and protein are significantly downregulated (by 50–70%) in the preBötzinger complex of postnatal day 2 rat pups that were prenatally exposed to maternal diets containing 4% or 8% alcohol (Ji et al., 2015). Brainstem slice preparations from alcohol-exposed offspring also showed significant reductions in respiratory rhythms including reduced frequency and amplitude, and a shortened inspiratory time (Ji et al., 2015), implicating deficient serotonin receptor signaling in the respiratory rhythm generator due to the effects of alcohol.

Another aspect of the healthy respiratory system is its ability to display neuroplasticity, allowing its adaptation to various environmental stimuli. One well-studied form of respiratory plasticity, long-term facilitation (LTF) occurs in both the hypoglossal and phrenic nerves in response to intermittent bouts of acute hypoxia, a form of plasticity that requires serotonin 5-HT_2A receptor-dependent signaling (Baker-Herman et al., 2004; Cao et al., 2010; Fuller and Mitchell, 2017). In brainstem slices from postnatal day 5–7 offspring whose mothers were exposed to 10% ethanol from 4 weeks pre-conception through lactation, hypoglossal LTF induced by anoxic events in vitro was absent; instead they developed LTD (Kervern et al., 2009). And as expected, 5-HT_2AR mRNA levels were similarly decreased in slices from prenatal ethanol-exposed neonates.

A role for inflammation in prenatal alcohol exposure?

In accordance with our theme of inflammation as a potential component of disrupted offspring CNS function following various maternal exposures, alcohol exposure during pregnancy increases fetal brain inflammatory cytokines such as interleukin (IL) IL-1β and IL-17 in late gestation (Pascual et al., 2017; Terasaki and Schwarz, 2016; Topper et al., 2015). Importantly, the effects of maternal ethanol consumption on offspring neural function can be significantly attenuated in the absence of the innate immune pattern recognition receptor toll-like receptor 4 (TLR4) (Pascual et al., 2017), strongly suggesting the contribution of inflammatory pathways to offspring neural deficits induced by prenatal alcohol. Despite the considerable literature on neural impairments due to fetal alcohol exposure, it has been only within the last couple of years that inflammation has begun to be investigated as a factor in alcohol-induced neural consequences (Chastain et al., 2019; Komada et al., 2017; Ren et al., 2019). To our knowledge, no studies have investigated the contributions of inflammation on alcohol-induced respiratory control and neuroplasticity deficits after either pre- or post-natal alcohol exposure. However, given the strong impairments in respiratory neuroplasticity caused by inflammation in adult animals (Huxtable et al., 2015; Huxtable et al., 2013; Huxtable et al., 2011) it will be important in future studies to determine whether mitigating alcohol-induced inflammation can restore hypoglossal and phrenic nerve neuroplasticity in alcohol-exposed offspring.

DRUGS OF ABUSE

The addictive tendencies of illicit drugs (such as heroin, cocaine and marijuana) promote their continued use into and throughout pregnancy, with approximately 6–8% of pregnant women reporting using these drugs either within the past month or during their third trimester (SAMHSA, 2018). These drugs can readily cross the placenta due to their lipophilic nature, with heroin, for example, being detected in fetal tissues within an hour of maternal use (Keegan et al., 2010). However, despite the well-known respiratory suppressive effects of opiates (Pattinson, 2008), they have long been used as analgesics, even in pregnant women. Short-term opiate use during pregnancy, in limited doses for acute pain management after surgery, is consistently reported to be safe for both the mother and the fetus (Keegan et al., 2010; Nossek et al., 2011), although long-term follow up of offspring neurological outcomes has not been extensively studied (Nossek et al., 2011).

Opioids and cocaine

On the other end of the spectrum, prolonged recreational use of opioids, or high dose exposures of illicit drugs during pregnancy can also cause respiratory depression in the offspring (Hultzsch and Schaefer, 2016). Respiratory depression or alterations in neural control are thought to contribute to some cases of SIDS. In this regard, in a 1984 study of infant deaths due to sudden infant death syndrome (SIDS), infant fatality was 5 times higher in offspring of mothers taking methadone and 17 times higher in mothers taking cocaine (Chasnoff et al., 1989). Accordingly, when cocaine-exposed newborns were subjected to hyperbaric hypoxia, 8 out of 9 babies from normal pregnancies aroused after 3 minutes, whereas only 6 out of 22 cocaine-exposed newborns awoke (Gingras et al., 1994; Ward et al., 1989). These data suggest that cocaine-exposed babies have a blunted hypoxic response. In addition, newborn mice prenatally exposed to cocaine showed an impaired ability to increase their ventilation in response to hypoxic challenge, an effect that was mainly due to a reduced change in total breath time compared to controls (Autret et al., 2002). While the exact mechanisms underlying these cocaine-related respiratory control deficits remain poorly understood, several possibilities have been suggested. Cocaine may act indirectly through the same mechanisms as opioids since piglets exposed prenatally to cocaine had increased met-enkephalin and β-endorphin levels (endogenous opioids) in various brainstem and medullary regions at postnatal days 2–5 and 18–22 (Liu et al., 2000; Zhang and Moss, 1995). Alternatively, the consequences of cocaine on respiratory function may involve its strong effects on inhibiting neurotransmitter transporters such as those for noradrenaline, dopamine, and serotonin (Simmler et al., 2017), leading to altered synaptic levels of neurotransmitters that play key roles in respiratory neural control (Bisgard et al., 1979; Hilaire, 2006; Viemari et al., 2004).

A role for inflammation in prenatal opiate and cocaine exposure?

Although not commonly considered, the effects of prenatal cocaine and opioids on offspring respiratory neural control may also involve enhanced neuroinflammation. For example, cocaine, by downregulating the immunosuppressive microRNA miR-124 in microglia in the striatum, indirectly results in upregulated pro-inflammatory TLR4 pathway signaling and inflammatory cytokine production (Periyasamy et al., 2018). Consistent with these pro-inflammatory effects of cocaine, users also had strongly increased serum levels of the pro-inflammatory IL-6 cytokine and greatly decreased serum levels of the anti-inflammatory cytokine IL-10 (Moreira et al., 2016). Likewise, opiate use, particularly heroin and morphine (but also methadone), cause inflammation in humans and rodent models of addiction with serum levels of multiple pro-inflammatory cytokines being increased in heroin addicts taking methadone, and in mice administered heroin or morphine (Chan et al., 2015; Kapasi et al., 2000; Mahajan et al., 2017; Neri et al., 2013; Pacifici et al., 2000; Szczytkowski and Lysle, 2008). Indeed, it is now becoming clear that multiple structurally distinct opioids like morphine can act as agonists of the inflammatory TLR4 receptor signaling complex (Grace et al., 2015; Hutchinson et al., 2011; Wang et al., 2012), perhaps mediating the activation of inflammatory signaling pathways (Grace et al., 2018; Wang et al., 2012). Although these studies have only been performed in nonpregnant human and adult rodents so far, they suggest the potential for drugs of abuse during pregnancy to induce maternal inflammation during critical periods of fetal CNS development, effects that may contribute to the offspring respiratory control deficits observed clinically. To our knowledge, inflammatory cytokines have not been evaluated in the fetal brainstem or respiratory control nuclei in the context of prenatal exposure to any drugs of abuse. This is despite dysregulated inflammation being a major consequence of administering these classes of drugs, and the significant detrimental effects that inflammation has on respiratory neural control.

Marijuana

Marijuana is in a politically transitional period at the writing of this article, as it becomes legalized in many states and countries. The lipophilicity of cannabidiol (CBD) and tetrahydrocannabinol (THC), the active components of marijuana, allow them to readily cross the placenta (Grant et al., 2018; Hutchings et al., 1989) likely underlying the associations between maternal marijuana use during pregnancy and low infant birth weight, anemia, and higher incidence of admission to neonatal intensive care (Gunn et al., 2016). Thus, considerations of marijuana effects on the developing fetus with maternal use is an even more significant issue at the present time. The effects of therapeutic modulation of the endogenous cannabinoid system in early life for treating refractory pediatric epilepsy in humans has been recently reviewed, although respiratory effects were not addressed (Schonhofen et al., 2018). The cannabinoid receptor, CB1, is expressed throughout the nervous system starting as early as gestational day 16 in humans (Volkow et al., 2017), including brainstem regions responsible for respiratory control. Indeed, direct microinjection of cannabinoid derivatives into adult rat brainstem slices causes a CB1 receptor-dependent short-term depression in hypoglossal motor neuron inspiratory related activity (García-Morales et al., 2015). Although newborn mice (<2 days old) exposed to a CB1 receptor agonist in utero from gestational days 5 to 20 displayed increased tidal volumes and decreased apnea and sigh frequencies at baseline, the average apnea duration was almost twice as long (Tree et al., 2014), suggesting that the respiratory system is disrupted following cannabinoid exposure. Interestingly however, these respiratory alterations did not persist into the second week of postnatal life, although baseline and hypoxia-induced respiratory frequency was increased in postnatal day 10–12 offspring (Tree et al., 2014), suggesting that the newborn respiratory system is somehow able to compensate for the earlier deficits induced by in utero cannabinoids. Mechanisms underlying this natural adaptation are not known, but a better understanding of them may be useful to harness for therapeutic treatment.

A role for inflammation in prenatal marijuana exposure?

In contrast to the previously discussed maternal drugs, marijuana reduces peripheral inflammatory cytokines such as IL-17a, tumor necrosis factor (TNF)α, and IL-6, while increasing anti-inflammatory IL-10 cytokine levels (Karoly et al., 2018; Katz et al., 2017). Since pregnancy is classified as a very tightly regulated state of inflammation, its disruption by agents that alter this fine balance between inflammatory or anti-inflammatory activities, can be detrimental to the developing fetus (Arck and Hecher, 2013; Meyer et al., 2008; PrabhuDas et al., 2015; Romero et al., 2014; Shelton et al., 2015). To our knowledge, no studies have investigated the contributions of inflammation to the deleterious offspring effects of maternal marijuana use.

PRENATAL INFECTION

Pregnant women are more susceptible to infection (Munoz-Suano et al., 2011), with approximately one in three women having at least one infection during their pregnancy, usually bacterial in nature (Brown et al., 2009; Fang et al., 2015). Although many pathogen types can cause prenatal infection, all agents have “maternal immune activation” (MIA) in common, resulting in increased levels of maternal inflammatory cytokines. Maternal infection during pregnancy is typically associated with neurocognitive deficits in the offspring, such as schizophrenia and autism spectrum disorder (ASD), among other psychiatric disorders (Atladottir et al., 2010; Patterson, 2009; Simanek and Meier, 2015; Solek et al., 2018). Multiple retrospective studies, including one in 2010 of more than 1.6 million children showed a link between increased risk of ASD and maternal viral infection in the first or second trimesters of pregnancy (Atladottir et al., 2010). There are also associations between increased risk for schizophrenia and maternal viral infection with herpes simplex virus-2 or influenza, or the parasite Toxoplasma gondii (reviewed in (Khandaker et al., 2013)). Importantly, whereas certain cytokines such as IL-1β, IL-6 and IL-17a have been identified as key mediators of MIA (Arrode-Bruses and Bruses, 2012; Bittle and Stevens, 2018; Choi et al., 2016; Girard et al., 2010; Wu et al., 2017), the specific cause of maternal elevations in cytokines need not be pathogen- or stressor-specific in order to cause offspring neural impairments (Guma et al., 2019; Gumusoglu and Stevens, 2019). It remains debatable how efficiently maternally-produced cytokines cross the placenta to enter the fetal compartment, whether transplacental transfer is cytokine-specific, or if the fetal placenta produces its own cytokines in response to maternal cytokines (Colucci et al., 2011; Dahlgren et al., 2006; Girard and Sebire, 2016; Ponzio et al., 2007); regardless, maternal inflammation during pregnancy impacts fetal brain development.

Neurocognitive and behavioral consequences

The relationship between prenatal infection and neuropsychiatric disorders such as ASD and schizophrenia in human epidemiological studies are also supported by animal models of MIA, in which exposure of pregnant animals at specific times during pregnancy to either a viral mimetic (polyinosinic-polycytidilic acid (poly(I:C)) or the bacterial cell wall component lipopolysaccharide (LPS) are sufficient to increase levels of key MIA pro-inflammatory cytokines in maternal serum and the fetal brain, and reproduce important aspects of human neurocognitive deficits (Fernández de Cossío et al., 2017; Malkova et al., 2012; Moreno et al., 2011; Mueller et al., 2019; Schaafsma et al., 2017). Similarly, experimental maternal infection induced by poly(I:C) for example, also causes behavioral abnormalities indicative of ASD in non-human primates (Bauman et al., 2014). While we only superficially describe this very large and growing field of maternal immune activation, a recent review has very comprehensively catalogued the numerous studies in the literature describing the psychiatric outcomes of offspring in various models of MIA, while also highlighting how much has yet to be discovered regarding underlying mechanisms and pathways (Solek et al., 2018). One additional point to note here with regard to maternal inflammation is that in rodent models, administration of cytokines to the pregnant dam is alone sufficient to recapitulate offspring CNS deficits, including autism spectrum disorder (ASD), schizophrenia, and anxiety disorders, among others (Choi et al., 2016; Kim et al., 2017; Rudolph et al., 2018; Smith et al., 2007; Wu et al., 2017), indicating that any situation resulting in increased maternal cytokines has the potential to cause neural impairments in offspring.

Respiratory consequences of prenatal infection?

In stark contrast to the abundant research on the effects of MIA on offspring cognition and behavior, to our knowledge, there are few studies that have investigated the impact of maternal infection on the development or function of the offspring respiratory neural control system. There are two epidemiologic studies in humans documenting an association between maternal levels of C-reactive protein (a clinical indicator of inflammation) and maternal respiratory infection with infant wheezing, incidence of infant lower respiratory tract infection, and infant lung function (Morales et al., 2011; Van Putte-Katier et al., 2007); however, neither of these studies evaluated endpoints associated with the neural control of breathing. Nevertheless, a study by Samarasinghe et al. in 2015 began to address this gap by exposing mouse dams to a single dose of LPS at gestational day 16, then studying offspring breathing at postnatal days 7, 28, and 60. Prenatal inflammation resulted in both lower baseline ventilation and a larger ventilatory response to hypoxia in all offspring age groups. Prenatal LPS offspring also displayed a larger ventilatory response to hypoxia and hypercapnia at both day 7 and 28, although this did not persist at 60 days. Neither spontaneous nor post-sigh apneas differed between control and prenatal LPS offspring in eupneic conditions, but following hypoxic or hypercapnic challenges, spontaneous and post-sigh apnea frequency respectively, were increased in offspring exposed to prenatal LPS (Samarasinghe et al., 2015). This study is the first to identify functional changes in respiratory control in the context of maternal infection as far as we are aware. Additional investigation of underlying mechanisms is necessary given the prevalence of maternal infection and inflammation during pregnancy. We propose that MIA increases neuroinflammation within the developing fetal brainstem and other CNS regions important for respiratory neural control, disrupting mechanisms of brainstem rhythm generation or causing transmission failure to respiratory motor neurons. Since inflammation is well-known to interfere with different types of respiratory neuroplasticity (reviewed in (Hocker et al., 2017)), an impaired capacity to mount a compensatory response to respiratory challenge (i.e. neuroplasticity) may also be a contributor to these maternal inflammation-induced offspring respiratory impairments.

DIESEL EXHAUST PARTICLES

Diesel exhaust particles have become a significant exposure in today’s global world, especially in large, densely packed cities (Brook et al., 2010; Chang et al., 2018). Diesel engine emissions are well-established human genotoxins, being both mutagens and carcinogens, especially in the airway and lungs (Khalek et al., 2011). Like pathogens, they also create inflammation and increase reactive oxygen species (Steiner et al., 2016). Indeed, serum levels of pro-inflammatory cytokines such as IL-6 are increased in children living near high traffic areas since infancy (Gruzieva et al., 2017), and rats experimentally exposed to inhaled diesel exhaust particles show increased levels of IL-6, TNFα, and IL-1β in several brain regions including the cortex and midbrain (Levesque et al., 2011). The ultrafine, lipophilic particulate matter contained in diesel exhaust can pass from the lungs to the bloodstream and across the blood brain barrier (Genc et al., 2012), producing pro-inflammatory responses in the brains of exposed individuals, and even in their offspring (reviewed in (Costa et al., 2017)). Due to their high permeability, diesel exhaust particle exposure during pregnancy is directly associated with negative outcomes in offspring. Both human epidemiologic and mouse studies indicate that developmental exposure to diesel exhaust is a risk factor for developing childhood wheeze and asthma (Gregory et al., 2017; Jedrychowski et al., 2010; Manners et al., 2014; Patel et al., 2011). Notably, these studies evaluated pulmonary function, not respiratory control endpoints.

Neurocognitive and behavioral consequences

In human epidemiologic studies, exposure of pregnant women to diesel exhaust particles correlates well with the increased incidence of offspring ASD (Roberts et al., 2013; Volk et al., 2013a) and other neuropsychiatric disorders such as anxiety, depression, and attention deficits in children (Perera et al., 2012; Roberts et al., 2013; Volk et al., 2013b). Likewise, mice exposed to diesel exhaust prenatally from conception at embryonic day 0 until weaning at postnatal day 21 also exhibit several behavioral impairments characteristic of ASD (Chang et al., 2018), suggesting that prenatal exposure to diesel exhaust particles in rodents during pregnancy causes similar neural deficits as maternal infection during pregnancy.

Prenatal diesel exhaust particle exposure has also been shown to contribute to other pathologies through two hit models as demonstrated in several studies by Bolton et al. The first group of studies explored the effects of a high fat diet in adult mice that were exposed in utero (gestational days 9–17) to maternal diesel particles (the first hit). At embryonic day 18, the offspring brains of mothers exposed to diesel particles showed increased expression of several inflammatory cytokines including IL-1β and IL-6 (Bolton et al., 2012). In adulthood, the offspring were fed a high fat diet as their second hit, and they demonstrated greater microglial activation within the hypothalamus, amygdala, dentate gyrus, and CA1 (but not CA3) region of the hippocampus (Bolton et al., 2012). In a separate study, pregnant mice were exposed both to diesel particles (every three days from gestational days 2–17) as well as maternal stress induced by limiting the available nesting material (from gestational days 14–19). The combination of both prenatal exposures increased anxiety in adult offspring, but cognitive ability was only impaired in males (Bolton et al., 2013). Interestingly, while the upregulation of fetal brain IL-1β mRNA levels were equal between male and female offspring, the potent anti-inflammatory cytokine IL-10 was decreased in the male offspring brain while it was increased in the female offspring brain (Bolton et al., 2013), potentially indirectly reducing overall inflammation in the female brain.

Respiratory consequences of prenatal diesel exhaust exposure?

While there are no human studies exploring functional offspring respiratory neural control outcomes after prenatal diesel exhaust exposure to our knowledge, based on data reported in the brain, it is feasible that diesel particles may also impact respiratory neuron function given the effects of diesel exhaust on both serotonin and dopamine levels in other brain regions. For example, levels of serotonin, dopamine, and their metabolites were increased within the amygdala of prenatally-exposed offspring at 3 weeks of age (Yokota et al., 2013); increased serotonin levels were also detected in the dorsal raphe nucleus of male offspring at 6 weeks of age (Yokota et al., 2016), a region which projects to respiratory control nuclei. It is possible that increased serotonin levels could create obstructive apneas due to serotonin-mediated depression of inspiratory motor output to the genioglossus (Morin et al., 1994) causing airway collapse in offspring prenatally exposed to diesel exhaust particles. Regardless, even if prenatal diesel exposure alone does not cause detectable respiratory impairments, it may serve as one hit of a two-hit model, that could potentially exacerbate respiratory consequences caused by other prenatal or later life exposures that ultimately synergize to disrupt the respiratory system. Better understanding the consequences of prenatal diesel exhaust exposure on long-term offspring respiratory function will be an important area for future study.

SLEEP DISORDERED BREATHING

An often overlooked and highly underdiagnosed maternal exposure during pregnancy is obstructive sleep apnea (OSA), the most common type of sleep disordered breathing in which periodic airway obstructions cause intermittent hypoxia (IH), hypercapnia and sleep fragmentation. OSA is estimated to occur in approximately 5–32% of pregnancies depending on other comorbidities such as obesity and diabetes (Antony et al., 2014; Robertson et al., 2019). OSA is a potent inflammatory stimulus in adults and children, increasing plasma indicators of inflammation including C-reactive protein, IL-1β, IL-6, IL-17, IL-23, and TNFα cytokine levels (Huang et al., 2016; Kong et al., 2018; Said et al., 2017). Increased maternal serum cytokines is an essential aspect of maternal immune activation. Indeed, an increase in maternal serum IL-6 or IL-17a levels alone is sufficient to alter offspring brain development and behavior in various models of MIA (Choi et al., 2016; Smith et al., 2007). Thus, given the similarity in cytokines that are upregulated in OSA as well as in various models of MIA, we propose that maternal sleep apnea may be another previously unconsidered, yet highly prevalent, model of MIA. In support of this idea, a study of 12-month-old children whose mothers had sleep disordered breathing during pregnancy showed impaired social development (Tauman et al., 2015), similar to children whose mothers had infections during their pregnancies described above. However, an important caveat in this regard is that most studies reporting elevated cytokines in individuals with OSA were tested in non-pregnant (and typically male) populations. Given that the maternal immune response during pregnancy is finely controlled to prevent rejection of the allogeneic fetus (Munoz-Suano et al., 2011), it is not clear whether pregnant women with OSA will also have the predicted increase in circulating cytokines; this is an important question that remains to be addressed.

Respiratory consequences of prenatal OSA?

To our knowledge, there are few studies that have investigated the capacity of gestational OSA exposure to alter offspring respiratory control in rodent models. While not a comprehensive model of OSA due to the lack of hypercapnia and sleep fragmentation, intermittent hypoxia alone is sufficient to recapitulate many of the morbidities observed in OSA. Although postnatal offspring exposed to GIH starting at embryonic day 0.5 until birth showed no alterations in diaphragm muscle contractility or endurance properties (McDonald et al., 2016), rat offspring exposed to GIH from embryonic day 5 through birth had higher ventilatory responses, effects which were maintained into adulthood (4 months of age) (Gozal et al., 2003). These data suggest that GIH may alter the neural control of breathing in a long-lasting manner. To better study respiratory neural activity in GIH offspring, and the role of inflammation in this context, brainstem-spinal cord preparations were made from neonatal offspring exposed to GIH throughout pregnancy starting at gestational day 10. Preparations from GIH offspring displayed an increase in burst rhythm regularity and tended to have a lower burst rise time compared to controls (Johnson et al., 2018), potentially suggesting that the GIH offspring brain has disrupted respiratory rhythm and pattern generating neural circuitry. In the context of a neonatal challenge with LPS, the typical reductions in baseline respiratory frequency that are observed with inflammation, failed to occur in GIH offspring preparations (Johnson et al., 2018). Based on the work of Jacono and colleagues in which respiratory rhythm regularity reflects a pathologic loss of a dynamic respiratory control system essential for allowing successful adaptation in the context of challenge or disease (Jacono, 2013; Young et al., 2019), we hypothesize that gestational sleep apnea exposure during pregnancy may make the adult offspring respiratory control system less adaptable to challenge. In this context, we predict that prenatal GIH will have long-term negative consequences on offspring respiratory function.

CONCLUSION AND FUTURE RESEARCH NEEDS

Impacts to the offspring respiratory control system have been studied following comparatively few types of inflammatory prenatal exposures. As depicted in Figure 1, each of these exposures has the capacity to alter aspects of fetal brain development, many of which lead to functional CNS deficits in the offspring later in life. Whereas maternal exposure to certain factors during pregnancy (like opioids and cocaine) are known to impact neurotransmitter receptor signaling pathways that modulate breathing, maternal immune activation caused by infection or diesel exhaust for example, exerts neural consequences via initiation of inflammation. We suggest that the potential for maternal exposures that may appear innocuous upon first inspection, or that may cause effects that are subtle enough to go unnoticed in normal physiological circumstances (like diesel pollution), may later manifest themselves in the context of an impaired response to challenge, especially in the respiratory system which must be able to rapidly respond and adapt to changing demands. Finally, we posit that many of these prenatal stimuli that are poorly studied with regard to respiratory control have at least one factor in common with the other more well-studied prenatal stimuli - - - maternal inflammation. It is ironic to note that the best-studied forms of prenatal exposure on the respiratory control system are in response to what may be considered otherwise “optional” or voluntary maternal exposures, such as cigarette smoking and illicit drug use. Conversely, maternal exposures that are all but unavoidable, such as maternal infection, exposure to diesel exhaust particles due to city-dwelling, and development of gestational sleep apnea are aspects of our modern world that have to date, been virtually overlooked with regard to offspring neural control of respiration. The growing incidence of breathing disorders such as asthma and sleep apnea in children and adults, together with the frequency and prevalence with which these less studied maternal exposures occur, underscores the need to better understand the influence of these gestational exposures on development of the offspring respiratory control system.

Figure 1. Multiple maternal exposures impact the developing brain in utero and lead to functional neural deficits in her offspring.

Many of the maternal stimuli known to have consequences on offspring respiratory control involve the presence of peripheral inflammation, a factor seldom considered in the context of alcohol or use of drugs of abuse. Conversely, maternal factors whose activities are well-known to involve inflammation in offspring neurocognitive deficits, respiratory function is rarely evaluated. Here we propose that sleep apnea during pregnancy increases maternal peripheral pro-inflammatory cytokines similar to established models of maternal immune activation. These cytokines negatively impact development of the fetal brain ultimately leading to impaired offspring respiratory system function later in life.

Highlights.

• Respiratory endpoints are seldom studied following the maternal exposures for which inflammation is a known mediator of offspring neural impairments (including maternal infection and diesel pollution).

• We suggest that maternal inflammation is a common point of convergence among prevalent prenatal exposures including alcohol, opioids, marijuana, and gestational sleep apnea.

• The need for further research to better understand the influence of prevalent but unstudied prenatal exposures on offspring breathing disorders in adulthood is highlighted.