Autism spectrum disorder (ASD) affects one in 44 children in the United States (Maenner et al., 2021), and is marked by deficits in social communication, restricted interests, and repetitive behavior (American Psychiatric Association, 2013). Autistic traits are detectable between 6 and 18 months (Barbaro and Dissanayake, 2009), indicating that critical windows to genetic and environmental factors occur during prenatal and early postnatal periods. Although the high heritability of ASD suggests that genetics is a key factor (Tick et al., 2016), previous studies have estimated that non-heritable factors account for >50% of the neurobiology of ASD (Mayer et al., 2014).

Traffic-related air pollutants such as particulate matter (PM) and nitrogen dioxide (NO2) have been suggested as environmental risk factors for ASD (Dutheil et al., 2021; Volk et al., 2013). While NO2 is mainly emitted from automobile exhaust and combustion of fossil fuels (Shang et al., 2020), PM is a mixture of toxic substances with various particle sizes and chemical properties, including sulfates, nitrates, ammonia, black carbon, dust, polycyclic aromatic hydrocarbons, metallic carbon, and volatile organic compounds (Zhang et al., 2021). Both air pollutants show high annual exposure levels in South Korea and are under active regulation by the Korean government. In 2019, the annual mean PM10 and NO2 levels (42 and 52.6 μg/m3) in Seoul, the capital of South Korea, were higher than in metropolitan cities such as Los Angeles (29 and 43.2 μg/m3), Tokyo (16 and 26.3 μg/m3) and London (18 and 32 μg/m3) (airkorea.or.kr, http://www.epa.gov, http://www.kankyo.metro.tokyo.jp, http://uk-air.defra.gov.uk). Although results on the association between traffic-related air pollution and ASD have been inconsistent, previous research has suggested that PM with an aerodynamic diameter ≤10 μm (PM10) and NO2 are related to an increased risk of ASD (Flores-Pajot et al., 2016; Chen et al., 2018; Wang et al., 2021). However, research on the mechanism underlying the association between air pollution and autistic traits is scarce.

Many individuals with ASD report comorbid gastrointestinal symptoms—constipation, abdominal pain, diarrhea, gas, and vomiting (Vuong and Hsiao, 2017)—as well as deficient gut epithelium integrity and increased intestinal permeability (Emanuele et al., 2010). The gut microbiota regulates central nervous system activities through various pathways (Liu et al., 2019a), such as regulating the hypothalamic–pituitary–adrenal axis (Sudo, 2012) and producing short-chain fatty acids (SCFA) that affect brain function (Ray, 2017). A previous meta-analysis found dysbiotic microbial compositions in children with ASD (Iglesias-Vázquez et al., 2020); however, a distinct microbial signature for ASD has not been defined yet (Vuong and Hsiao, 2017).

Air pollution exposure can alter the composition of the gut microbiome (Bailey et al., 2020). Mucociliary clearage of inhaled air pollutants in the lung and contaminated food/drinking water are major routes that PM enters the gastrointestinal tract (Salim et al., 2014). PM can either support or inhibit the growth of specific microbes, causing alteration in the composition and function of the gut microbiota (Gao et al., 2017; Korpela et al., 2019; Adams et al., 2015). Moreover, PM2.5 and PM1 exposures showed negative associations with alpha diversity indices and the relative abundance of most Firmicutes, Proteobacteria, and Verrucomicrobia bacteria (Liu et al., 2019b). NO2 was associated with alternation in the gut microbiome profile in young adults, including increased Firmicutes abundance at the phylum level and Coriobacteriaceae, Ruminococcaceae, and Adidobacteriaceae abundance at the family level (Fouladi et al., 2020).

The microbiome is associated with both air pollution and autistic traits; however, this complex relationship has not been investigated yet. Furthermore, it can potentially mediate environmental risk factors in ASD (Vuong and Hsiao, 2017). The microbiota has bi-directional relationships with both genetics and environment; host genetics affect its composition and function, while environmental factors, including age, infections, diet, and xenobiotics, further shape the microbial profile (Falony et al., 2016). Moreover, early-life alterations in the microbiota can have long-term consequences for health and disease (Kumar et al., 2014). This study aimed to examine whether pre- and postnatal PM10 and NO2 exposures impact autistic traits at 6 years of age through the alteration of the gut microbiome among the children in an ongoing birth cohort. It also aimed to explore the association of PM10 and NO2 exposure (1st, 2nd, and 3rd trimesters of pregnancy; ages 2, 4, and 6 years) with autistic traits at age 6 years, the relationship of PM10 and NO2 exposure with the gut microbiome composition at age 6 years, and the association between microbiome profiles and autistic traits. Mediation analyses of statistically significant findings were also conducted to confirm the “air pollutant exposure–gut microbiome–autistic traits” pathway.