Characterizing the neuroimmune environment of offspring in a novel model of maternal allergic asthma and particulate matter exposure

Neurodevelopmental disorders (NDD) are a broad group of disorders where central nervous system (CNS) development and function is altered [27]. NDD, including autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD), are increasing in prevalence, with 1 in 6 children in the US experiencing a NDD in 2016 [125]. While there are some genetic origins of NDD, environmental factors such as exposure to pollution and maternal health are also recognized as risk factors [46, 100]. Mounting clinical and epidemiological data underscore the notion that environmental factors play a large role in both the underlying development of NDD and their impact on the severity of individual behavioral characteristics [52, 94, 99, 105, 114, 118, 120]. Pregnancy, in particular, marks a critical period when environmental insults have long-lasting effects on neurodevelopment.

The maternal immune system represents a key biological mechanism that links environmental factors to specific neurodevelopmental changes and later NDD diagnosis. For example, clinical studies have linked hospitalizations for bacterial and viral infections during pregnancy to an increased rate of birthing a child that will later be diagnosed with ASD [4, 14, 80, 116]. Moreover, laboratories have developed animal models that often use bacterial or viral mimics, such as lipopolysaccharide (LPS) or polyinosinic:polycytidylic acid (Poly(I:C)), respectively, to induce maternal inflammation during pregnancy and demonstrate a causal change in offspring brain and behavioral health [46]. Despite the efficacy of these models in demonstrating a link between maternal immune signaling and offspring neurodevelopment, they are limited in representing many of the common environmental sources of immune activation. With perhaps the exception of SARS-CoV-2, hospitalizations for viral infections are less frequent, and a recent meta-analysis of viral infections during pregnancy as a risk for children later being diagnosed with ASD was not supported [55]. Importantly, infections via bacterial and viral pathogens are only two of many ways in which the immune system can become activated, and researchers have begun to investigate other common environmental influences or immune conditions that can have a substantial impact on the maternal immune system and offspring brain development, notably allergic asthma.

Rates of asthma, like NDD, are currently increasing and represent a highly prevalent chronic disease that more commonly impacts ethnic/racial minority and socioeconomically disadvantaged groups in the US [23, 72]. Importantly, chances of asthma exacerbations increase during pregnancy, and the presence of asthma increases likelihood of birthing a child that is later diagnosed with ASD [1, 3, 25, 26, 33, 42, 47, 87, 88, 93]. Asthma is an inflammatory disease of the lungs characterized by bronchial hyperresponsiveness, persistent inflammation, airflow obstruction, and reduction of airflow [39, 96]. The allergens associated with asthma can vary and include, but are not limited to rodent allergen, cockroach allergen, pollen, and house dust mite [53, 72]. Given the incidence of allergies and asthma in the United States [10, 34] and the high prevalence of allergic triggers in urban environments [22, 64, 72, 106], there is pressing need to better understand the causal effects of asthma during pregnancy and the impact of specific allergic triggers on offspring neurodevelopment.

Using our mouse model of maternal allergic asthma (MAA) to initiate an immune response in pregnant mice, we have previously reported systemic elevations in interleukin-4 (IL-4) and IL-5 in dams [109, 115] that parallel clinical reports associating increased IL-4 and IL-5 in mid-pregnancy maternal serum samples of mothers with children later diagnosed with ASD [40]. Not only do the dams in our MAA model produce a similar allergic asthma cytokine profile to that observed in humans, but the offspring display characteristic ASD-like behaviors, such as decreased social interaction and increased repetitive-like behaviors [20, 109, 110]. Moreover, MAA produces transcriptional differences in microglia gene expression and neuroinflammation in both prenatal offspring and brain regions of adult mice [20, 115, 117]. These findings highlight the lasting changes that allergic asthma during pregnancy can have on offspring neurodevelopment. However, environmental insults associated with ASD can vary, and exposure to these environmental factors does not necessarily occur in isolation for humans. In fact, individuals often simultaneously encounter multiple inflammatory-inducing environmental stimulants, and less is known about the potential synergistic effects of these exposures on maternal immune response and subsequent offspring development [22, 64, 106].

Particulate matter (PM), via air pollution, is another environmental factor that is suspected to be associated with the risk of neuropsychiatric disorders/NDD such as schizophrenia, attention deficit hyperactive disorder (ADHD), and ASD [118‐120]. Not only is PM linked to neurodevelopmental disruptions when exposures occur during pregnancy [21], but it has also been tied to exacerbated asthma responses and could potentially cause new onset cases of asthma [16, 17, 44, 108] although this remains controversial. PM is a complex mixture of constituents that contains organic compounds, soot, metals and metal-oxides, nitrates, and other elements in varying quantities depending on its source [113]. As such, it is necessary to characterize individual components of PM in order to effectively regulate air quality for limiting environmental exposure to toxicants.

Diesel exhaust is one common source of PM that has been identified as a risk factor for ASD [57, 103, 118‐121]. Combustion-derived diesel exhaust includes soot particles containing elemental carbon and iron, which is the most common transition metal found in PM [127]. Iron in PM can occur from fuel additives and as a product of normal engine wear [74, 113], and iron-soot (IS) exposure has previously been shown to cause oxidative lung injury, induce inflammation of the lungs, and can cross the blood–brain barrier through nasal inhalation [48, 126, 127]. The links between PM and worsening of asthma symptoms and increased risk of neuropsychiatric disorders in offspring make PM exposure during pregnancy an important area that merits investigation.

There are many human studies and animal models investigating the links between gestational exposure to PM and neurodevelopmental outcomes in offspring [1, 2, 11, 118‐120]. In addition, many studies have investigated the impacts of asthma as a risk factor for ASD, including our own [3, 20, 25, 26, 33, 42, 47, 68, 87, 88, 93, 109, 110, 115, 117]. However, these investigations consider asthma or PM only as independent risk factors for developing neuropsychiatric disorders. Despite the links between allergic asthma with ASD and the link between PM and ASD, there are no studies investigating the neuroimmune outcome on offspring of these environmental factors in conjunction. This is an apparent oversight given that PM exposure can worsen symptoms of asthma [44], and the source of environmental factors contributing to ASD are likely multifaceted.

In this study, we hypothesized that when mice are exposed to MAA and ultrafine iron-soot (UIS) particles during pregnancy (MAA–UIS), the neuroimmune outcomes will show heightened inflammation in offspring compared to that of MAA or UIS exposure alone. We also suspected that, because microglia can respond to changes in the neuroimmune environment through proliferation and are suspected to be associated with ASD behaviors [28, 51], we would see signs of functional differences in the frontal cortex and hippocampus—two brain regions implicated as being impacted developmentally in neuropsychiatric disorders such as ASD [6, 31, 51, 102]. Using our established mouse paradigm of maternal aerosol exposure to study offspring outcomes, we demonstrate that MAA and UIS alone, as well as MAA and UIS combined, alter the neuroimmune profile in the brains of offspring that is sustained from adolescence into early adulthood.

Among the many well established environmental factors that can impact fetal neurodevelopment, asthma and air pollution represent two major sources of immune stimulation that are on the rise, making them a significant concern for pregnant individuals. Based on previous studies of maternal immune activation with asthma, and PM exposure during pregnancy, we hypothesized that the combination of these two environmental stimuli would result in an exacerbated neuroimmune response in offspring. Although the appearance of an additive or synergistic effect of MAA and UIS exposure combined was limited, we did identify increases in cytokine concentrations across all treatment groups in the cortex and hippocampus that may suggest converging pathways for each insult/exposure. However, we also observed potentially suppressive activities that may suggest activation of competing or diverging pathways. Importantly, some of changes appear to be sustained across treatment groups from adolescence into early adulthood in the cortex, demonstrating lasting impacts of these gestational exposures on the neuroimmune environment later in life. Although we identified increases in cytokines in the hippocampus within all treatment groups at P15, these did not remain elevated into early adulthood. Overall, these data show that MAA and UIS environmental stimuli can result in an altered neuroimmune environment that persists from juvenile to adult timepoints.

The allergic response in the lungs of individuals with asthma is characterized by an influx of immune cells, such as neutrophils, mast cells, macrophages, and T-helper (T_H)2 cells. Our mouse model of MAA previously showed elevated IL-4, IL-5, IL-17, and IFNγ in the lung and peripheral blood of mice exposed to aerosolized OVA during pregnancy [20, 110, 115]. These increases in maternal serum cytokines result in neuro-immune signaling changes in the fetal brain during in utero development [115]. Our present data extend these findings by revealing evidence of increases in cortical and hippocampal cytokines in juvenile mice of MAA dams. In the cortex, for example, MAA increased IL-1β, IL-2, IL-17, IL-10, and IL-7. In the hippocampus, we identified IL-1β, IL-2, IL-7, IL-13, IL-15, and RANTES as being elevated in P15 offspring of MAA-exposed dams. These observed increases in MAA compared to PBS-AIR controls demonstrate the independent neurodevelopmental impact of allergic inflammation during pregnancy on offspring neuroinflammation. In addition to these findings in the MAA-only treatment group, we also identified suppressive effects of combining MAA with UIS in the cortex in juvenile offspring, notably, IL-10 an anti-inflammatory cytokine, whose suppression may suggest a lack of immune regulation and shift towards more neuroinflammation. Moreover, increases in cytokines as a result of UIS exposure alone were also identified in the hippocampus of juvenile offspring, with elevated IFNγ, IL-1β, IL-7, and IL-12(p40). To the authors’ knowledge, investigations into these neurobiological outcomes in offspring under UIS exposure during gestation have not been previously reported, highlighting the novelty of our model and findings.

In addition to independent effects of MAA or UIS treatment on cytokine concentrations in the cortex and hippocampus of juvenile offspring, several cytokines did not respond further to the combined effects of MAA and UIS. Most notably, we observed similar elevations in the cortex of IL-1ɑ, IL-1β, IL-2, and IL-17 in dual-exposed MAA–UIS offspring suggesting a potential ceiling effect produced by either environmental insult alone, or a maximal response through converging signaling pathways. IL-1β, IL-13, and IL-17 were also trending or increased in the MAA and UIS single treatment groups as well as the MAA–UIS treatment group at P15, and they remained elevated into the P35 timepoint in the cortex. Our data suggest a sustained elevation in these cytokines from P15 to P35 as a result of both MAA and UIS that have the potential for long-lasting impacts on neurodevelopment in the cortex of these offspring.

Consistent with the pleiotropic nature of cytokines in the central nervous system (CNS), IL-1β, IL-2, and IL-17 have all been identified as having neurotrophic properties [5, 13, 29, 58, 66, 69, 78, 92] (Rochman et al. 2013). Indeed, high concentrations (500 ng/mL) of IL-1β can have neurotoxic effects on neurons when exposed for 3–5 days [92], and IL-17 is detected at high levels in the CNS in multiple sclerosis and associated with the neuroinflammatory pathology of the disease [58, 66]. Compared to these neurotoxic concentrations, our data represent moderate increases in cytokines with less than 2.5-fold increases in treatment groups compared to PBS-AIR controls, and may not represent overt inflammation per se. However, these smaller changes in brain cytokine levels during the juvenile period may be biologically significant given their alternative functions in promoting neuronal survival and neurodevelopment. For example, IL-1β acts as a chemokine that guides neurite outgrowth in cortical neurons [69], and IL-17 acts in initiating the release of brain-derived neurotrophic factor (BDNF), glia-derived neurotrophic factor (GDNF), and nerve growth factors (NGF) that are associated with neuronal cell survival and repair [78]. Taken from this view, these cytokines, which are generally considered overtly inflammatory, may be having a more subtle impact on the neuroarchitecture of offspring brains than models finding dramatic increases in concentration of IL-1β and IL-17.

Further demonstrating the neuropoietic nature of these cytokines within the CNS and adding to the idea that the moderate increases observed in this model may be altering the neuropatterning of the offspring brains, IL-2 has been found to have neurotrophic properties and is necessary for proper cytoarchitecture in development [8, 8, 9, 9]. Both IL-2 and IL-13 are among several cytokines that are known to decrease in concentration at P14 under homeostatic conditions, and this developmental timepoint in mice is characterized by a high degree of synaptogenesis and pruning [37, 81]. In contrast, our model, which investigated cytokine concentrations at P15, still within this window of high synaptogenesis, showed increased IL-2, representing a shift in homeostatic load. Taken together, it may be that these sustained moderate increases in cytokines of the cortex are changing the trajectory of cortical development and promoting altered connectivity in the cortex linked to behavioral changes such as decreased social interaction and repetitive behaviors previously identified in our model [19, 109]. This phenomenon of altered connectivity has also been implicated in the core behaviors associated with ASD, specifically the social deficits and restrictive and repetitive behaviors [24, 73, 77].

Similar to our findings in the cortex, we also identified elevations in several cytokines, most notably IFNγ and IL-7, at P15 in the hippocampus of MAA and/or UIS groups compared to control offspring. IFNγ receptors are present on both neurons and glia [90]. In the hippocampus, IFNγ appears to play a role in synaptogenesis [65]. Some investigators have found that overexpression resulted in increased neurogenesis in the dentate gyrus, and because of its neuromodulatory effects, it has been suggested that this may impact cognition and social behavior as a result [7, 35, 36]. Additionally, IL-7 promotes survival and neurite outgrowth in hippocampal neurons [70, 76]. Considering the effects of IL-7 and IFNγ, and that the hippocampus is a major neurogenic niche in the developing brain, future studies may benefit from investigating the potential for hippocampal overgrowth in offspring brains in response to UIS or allergic asthma inflammation during pregnancy. Indeed, this phenomenon of hippocampal overgrowth has been identified in cases of ASD [43, 86, 104] and has been implicated in the deficits associated with emotion perception and sensory processing in ASD individuals [6, 43]. Curiously, the observed increases in hippocampal cytokine concentrations at P15 were not observed in any treatment group of P35 offspring compared to PBS-AIR controls. Although we can only speculate about these findings, it may be that these changes resolve during adolescence when additional brain maturation may be present to compensate for developmental overgrowth much in the same way that volumetric increases in the hippocampus of ASD individuals do not persist into adulthood [43, 104]. Although IL-7 and IFNγ are only two examples of cytokines that were elevated across the treatment groups, they illustrate the broader findings that treatment with MAA, UIS, or MAA–UIS can alter the hippocampal neuroimmune environment.

Prenatal insults have been shown to have lifelong impacts on microglia function and are suspected to play a prominent role in NDD [32, 38, 50, 59, 112, 117]. In ASD, some postmortem studies have identified differences in microglia density and morphology in brains of individuals [60, 82, 83]. In our previous study of MAA, we found DNA methylation differences in adult microglia, and several of these changes occurred in regulatory genes that are shared among some ASD individuals [117]. Given these findings, we sought to examine the density of microglia in the P15 brains of our MAA and UIS exposure model. We observed a significant increase in microglia density within the hippocampus of offspring exposed to MAA–UIS, but these increases were not present in the frontal cortex. One plausible explanation for why these increases were only observed in the hippocampus may be due to the higher density of microglia known to be present in the hippocampus. This higher density of microglia is thought to make the hippocampus more vulnerable to inflammation [18, 101], and disruptions in the dentate gyrus have been linked to NDD [15, 124]. Our findings of increased microglia density in the hippocampus of MAA–UIS offspring mirror data from another maternal immune activation model that utilizes the viral mimic poly(I:C). Specifically, Juckel et al. [56] reported an increase in microglia density in the hippocampus but not the cortex of offspring born from immune-activated dams. Similarly, another study of maternal immune activation using LPS stimulation showed an increase in microglia density in the hippocampus [30]. While it is difficult to make conclusions about microglial function based on density data alone, our observed difference in the hippocampus in combination with similar reports from other maternal immune activation models [30, 56], suggest that asthma allergy and PM mediated immune activation during pregnancy can result in a deviation from homeostatic activity in the offspring hippocampus. Of interest, IFNγ and IL-12 were increased in MIA-UIS in the hippocampus, cytokines that are often associated with a T_H1 response and M1 macrophage/microglia skewing. These cytokines were not elevated in the cortex. Moreover, in the cortex and hippocampus UIS is associated with increased IL-13 that may balance the signals from IL-12 and IFNγ as seen in peripheral responses. IL-13 is often considered anti-inflammatory in the CNS, with some studies pointing to a neuroprotective impact in CNS diseases and injuries [45, 63, 75, 91]. Although speculative, the results may suggest combined inflammatory M1/T_H1 associated cytokines IFNγ and IL-12, could influence the MAA–UIS microglia density results while these are protected in the cortex or under UIS conditions through IL-13 production. More investigation is needed on the combination of cytokines released and action on microglia cells in the context of environmental exposures.

Although we did not collect maternal data in this preliminary study, data from previous MAA studies demonstrate increased systemic inflammation characteristic of an allergic asthma response, specifically with increased IL-4, IL-5, and IL-13 [19, 110, 115], suggesting the potential for a similar response in dams of MAA in the current model. Speculation about the systemic impacts of UIS on the maternal immune system, however, is difficult. Many studies of PM exposure suggest IL-6, IL-8, and TNFα as the main cytokines upregulated in response to PM exposure [41, 67, 79, 89, 111]. This difference in cytokine response highlights the potential reason we see differences in the impact between MAA and UIS in our model. However, models of PM can vary widely in the size of PM and composition [79], making speculation about the maternal response in the UIS groups, and the potential role this plays in offspring neurodevelopment, difficult. This variation in PM studies underscores the need for future investigations to identify the maternal cytokine milieu in this model.

While our findings do not necessarily demonstrate an additive effect of MAA and UIS with regard to the cytokines we investigated, we did see a synergistic impact of MAA–UIS on microglia density in the hippocampus. These findings demonstrate the potential for additive effects of maternal asthma exposure when coupled with PM exposure. Independently, studies have shown in both humans and animal models that PM exposure during pregnancy can increase the susceptibility of offspring developing asthma [49, 84, 123]. This increased susceptibility of asthma in offspring is also seen in children of asthmatic mothers [71, 85] (Mattes et al. 2013), suggesting the potential for systemic immune dysfunction when these two stimuli are combined during pregnancy. The findings in this unique model of MAA and UIS exposure highlight the importance of investigating the impact of these closely linked and prevalent environmental factors.

Our data add to our previous studies on the impact of MAA on fetal brain development, showing here that this model impacts region-specific cytokine concentrations in both the juvenile and adolescent periods. To the investigators’ knowledge, this was the first study to assess the impact of ultrafine iron-soot exposure during gestation on offspring neurobiology. Taken together, these data highlight the importance of understanding the impact that common environmental stimuli can have on fetal development, and the potential for these stimuli to have long-lasting changes in offspring.