Obesity has become a global health challenge reaching alarming prevalence and being a cause of chronic metabolic diseases, including type 2 diabetes (T2DM) (1–3). The etiology of obesity implicates multiple factors involving environmental, genetic, and epigenetic modification (4–7). Clinical, epidemiological, and experimental studies consistently demonstrate that abnormal maternal caloric intake or environment disturbance during the perinatal period favors the development of cardiometabolic diseases in the offspring (6, 8).

The mechanism by which an altered environment during the perinatal period induces obesity remains unknown (8). The hypothalamus is the most studied organ that integrates peripheral signals with neural circuits and nutritional signals, keeping food intake and energy expenditure (EE) in balance (9–11). The arcuate nucleus of the hypothalamus (ARC) has a diversity of neuron populations that control metabolism. Proopiomelanocortin (Pomc) and Agouti-related peptide (Agrp) co-located with neuropeptide Y (Npy) are the most studied that control energy balance. Fasting activates the orexigenic Agrp/Npy neurons inducing food intake (9–14).

In contrast, feeding activates Pomc neurons, suppressing feeding while increasing energy expenditure (15). Failure to develop Pomc and Agrp neurocircuits increases the predisposition to obesity and its comorbidities (8, 16–18). Changes in the maternal diet composition containing a significant amount of saturated fat cause obesity and glucose intolerance (19) and elicit a preference for sucrose (15). Factors other than excess nutrients consumed by the mother can alter the metabolic phenotype of her offspring. Maternal smoking, even at a low-intensity frequency during gestation, is associated with reduced fetal growth and an increased risk of preterm birth (20).

Air pollution exposure containing particulate matter with a small diameter (≤ of 2.5 μm: PM_2.5) increases the risk of obesity and T2DM (21–23). The ability of PM_2.5 to induce metabolic diseases is based mainly on inflammatory responses induced by its chemical composition, which can also include lipopolysaccharides (LPS) (24–28). LPS is a potent agonist of Toll-Like Receptor 4 (TLR4), considered a part of the etiopathology of obesity and T2DM (25, 29). TLR4 utilizes a canonical signaling cascade activating the upstream nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha (IκB), mainly by inhibitor-IκB kinase (IKK)s, which phosphorylate the IκB, inducing its degradation and releasing the Nuclear Factor Kappa B (NFκB), a transcription factor, to migrate to the cellular nucleus. NFκB in the nucleus increases the transcription of proinflammatory genes such as Tumor Necrosis Factor-alpha (Tnf alfa) (30, 31) and the expression of IKKϵ (IKK epsilon) and TANK-binding kinase 1(TBK1) (32), perpetuating the inflammatory signal. In parallel to inflammation, IKKβ (IKK beta) and IKKϵ are kinases involved in developing insulin resistance and T2DM (33–35). Therefore, the activation of TLR4/NFκB and the IKKϵ pathways are key for inflammatory responses and metabolic dysregulation in numerous tissues.

Recently we showed that short-term exposure (five days) to PM2.5 was sufficient to increase the expression of inflammatory markers such as Tlr4, Ikbke, and Tnf alpha in the hypothalamus causing leptin resistance and obesity at long-term of PM2.5-exposure (36). Other groups have shown that PM exposure activates the NFκB pathway in vivo (mice) and culture cells from different tissues (37, 38).

Due to its lower diameter, PM_2.5 can spread into tissues through the lungs and olfactory epithelium (27). A portion of PM_2.5 can also be transported to the oropharynx and ingested, reaching the gastrointestinal tract and altering gut permeability and local microbiota (31, 32). Changes in the gut morphophysiology and microbiota facilitate inflammatory responses and might be related to causing metabolic diseases such as obesity and T2DM (39–41).

Evidence suggests that perinatal exposure to PM_2.5 might alter the energy metabolism of the offspring (42–50). However, the results are divergent. It is challenging to compare these studies due to the different approaches and protocols for PM_2.5 exposures. For instance, some studies employed intratracheal instillation of liquid diesel exhaust PM_2.5 (48, 51). In contrast, other studies used natural PM_2.5 or filtered air (FA) captured by ambient chambers (49, 50, 52). Besides, the timing points elected to maternal exposures are considerably distinctive among studies, including preconception or offspring exposures in some of them (48–50), which leads to maternal inflammatory responses before conception and direct contact of animals with the particles.

Considering that the time point of maternal PM_2.5 exposure is crucial to determine the impact on the metabolic outcomes, specifically on the regulation of feeding and energy expenditure, and to tease apart the effect of preconception maternal exposure to PM_2.5, in the present study, we exposed mothers to PM_2.5 only during gestation or lactation. We had a control group that received FA simultaneously, and we kept the mothers and offspring on a chow diet. Using this paradigm, we investigated whether exposure to PM_2.5 during gestation or lactation alters the metabolic phenotype throughout the offspring’s life. More specifically, we investigated possible changes in 1) body weight, fat mass and lean mass, leptin, food intake, and energy expenditure; 2) hypothalamic neuropeptides and pro-inflammatory markers; 3) the gut microbiota diversity and composition.

We demonstrated that maternal exposure to air pollution affects the offspring’s energy metabolism. To our surprise, these findings were time-dependent throughout the offspring’s life and gender-specific.

Birth weight is a parameter for estimating intrauterine growth (65, 66). Our study found that PM_2.5 exposure during gestation was associated with a tendency to increase the offspring’s birth weight independently of the gender, which is compatible with increased fetus growth. Males in the PM_2.5/FA group displayed increased body weight from weaning until young adulthood. However, this difference diminished at 30 weeks of age. Nonetheless, the females in the PM_2.5/FA group did not display higher body weight at weaning. However, the females were heavy at a young age (8 weeks) and kept higher adiposity until later in life (30 weeks). This finding suggests that the effect of PM_2.5 during gestation was long-lasting only in females. Several studies demonstrated that PM_2.5 exposure during gestation affects body weight. However, some studies showed the opposite of us, a reduction in birth body weight (49, 50). Even though these mice evolved with increased body weight in adulthood (49, 50), another study showed an unchanged offspring birth weight and a decrease in adulthood (48).

The difference between our results might be due to different time points of maternal exposure to PM_2.5. Differently from us, other studies (48–50) exposed the females during gestation and seven weeks before. The three studies mentioned above suggested maternal preconception exposure might be more harmful than during gestation (48–50). Thus, the mild phenotype we observed in our study might be partially due to limited time points of exposure. The inconsistency among studies may also rely on the route of administration of PM_2.5 and the source of PM_2.5. Some studies employed intratracheal liquid instillation (48). Other studies (49, 50) and ours used concentrated PM_2.5 in specialized chambers. The sources of pollutants might also contribute to different results since the composition of PM_2.5 depends on the emission sources.

Even though other studies found lower birth weight after maternal PM_2.5 exposure, the outcome was increased body weight at a younger age, at least for male offspring. This result agrees with our data that showed higher body weight in male offspring whose mothers were exposed to PM_2.5 either during gestation or lactation. Together these data suggest that changes in birth weight, whether low or high, induce weight gain in young male adulthood. In female offspring, body weight at weaning was similar among the three groups. However, they displayed an increase in body weight at a younger age. Adipose tissue delivers leptin balancing with triglyceride stores (67, 68). Thus, circulating leptin is known to define body energy stores (67–69). We found increased leptin levels in the same groups with elevated body weight in younger offspring. Later in life, only females displayed high leptin levels and gonadal fat mass. Concurrently the increased leptin levels suggested increased fat mass.

Our data showed an elevation of food intake followed by increased levels of the orexigenic Agrp and Npy in males from the PM_2.5/FA group compared to FA/FA and FA/PM_2.5. The elevated Agrp and Npy expression in the hypothalamus might count for the increased food intake we observed in this group. In the arcuate nucleus, the central melanocortin system comprises Pomc, Agrp, and melanocortin receptors (MC4Rs). Human mutations in the MC4R trigger obesity (70), and AgRP neurons are the natural antagonist of these receptors favoring increased food intake (8). Changes in maternal nutrition alter the innervation of Pomc and Agrp in arcuate to PVN. The critical window to compromise these projections is during lactation, not gestation. This occurred since rodent projections are developed three weeks after the offspring is born. In humans, the development of those projections occurs in utero (8, 15–18, 71). Lippert et al. (2020) demonstrated that maternal exposure to HFD during lactation revealed sexually dimorphic expression of dopamine phenotypes, males showing hyperlocomotion, and females displaying a high preference for sucrose (15). Another study exhibited that maternal HFD exposure seven days before and during gestation and lactation predisposes the offspring to develop insulin resistance and obesity by activating white adipose tissue inflammatory pathways (72).

In contrast, maternal undernutrition also increases the risk of the offspring developing metabolic diseases such as obesity (73). In male offspring from the FA/PM_2.5 group, despite displaying elevated body weight at weaning, the body weight was similar to the FA/FA group at eight weeks of age. The leptin levels followed decreased body weight, suggesting a decreased fat mass (67, 68) in this group.

Air pollution exposure, as well as changes in nutritional diet, have an impact on energy balance. In our study, female and male offspring from FA/PM_2.5 group displayed higher RER. This result indicates changes in substrate utilization, burning more fat than carbohydrates. The level of O₂ consumption and Ucp1 expression in BAT was not uniform in our data. Another study that exposed the mothers to PM_2.5 observed decreased Ucp1 expression in BAT in the group PM_2.5-exposed during lactation (48). This result was accompanied by increased adiposity (48). In males, O₂ consumption was higher in offspring from FA/PM_2.5 group than FA/FA, but Ucp1 expression showed a tendency to increase but was not significant. Interestingly, this offspring (from the FA/PM_2.5 group) displayed lower body weight and leptin levels than PM_2.5/FA group, suggesting that higher O₂ consumption might contribute to decreasing body mass in this group. Together with a mild increase in food intake looks like maternal exposure only during lactation had a lower effect on the EE of male offspring. Females from the PM_2.5/FA group also displayed higher O₂ consumption and Ucp1 in the BAT than FA/FA. However, despite higher EE, PM_2.5/FA females presented higher body weight and leptin levels. Thus, in this group, the increased food intake and the RER might contribute to keeping the higher body weight than the EE.

We also observed that maternal exposure to PM_2.5 during pregnancy reduces the gut microbiota diversity and changes the composition independently of the offspring’s gender. However, to our surprise, the body weight, fat mass, and leptin levels in males lost the difference among the three groups. On the contrary, we observed an increased fat mass and leptin levels without changes in body weight in the female PM_2.5/FA group than in FA/FA.

Evidence demonstrated that exposure to air pollution affects gastrointestinal health (74, 75); however, the results are conflicting. Kim et al. (2014) showed that exposure to air pollution reduces gut microbiota biodiversity (74). In contrast, Mutlu et al. (2019) suggested that PM_2.5 exposure increases gut microbiota alpha diversity (39, 75). Also, several studies (76–78) highlighted the influence of air pollution on obesity and T2DM through gut microbiota. Our study observed that maternal PM_2.5 exposure displayed altered offspring gut microbiota composition, characterized by changes in bacterial diversity and different amounts of some taxon. Similar to other studies (78–80), our results showed that the male and female offspring PM_2.5/FA group displayed decreased alpha diversity (Shannon or Simpson index) compared to the FA/FA and the FA/PM_2.5 groups.

In contrast, Liu et al. (2020) described an increase in alpha diversity associated with maternal PM_2.5 exposure during the gestational period (77). In the same study, despite an increase in alpha diversity, beta diversity analysis revealed an apparent dissociation between the control and PM_2.5 exposed group (77). Our study also observed an evident dissociation among the three groups. We observed differences in the abundance of some taxon among the groups. Female PM_2.5/FA offspring showed lower proportions of Bacteroidetes phylum and higher proportions of Verrucomicrobia compared to female FA/FA or FA/PM_2.5. Like females, male PM_2.5/FA offspring showed higher ratios of Verrucomicrobia phylum but lower proportions of Firmicutes phylum compared to male FA/FA or FA/PM_2.5. In our study, more significant amounts of Akkermansia (Verrucomicrobia phylum) were observed in the male and female offspring from the PM_2.5/FA group compared to FA/FA group. Akkermansia has been associated with improved glucose homeostasis and weight loss (41, 77, 81, 82). It is known that genetic and environmental factors can change microbiota composition, and if the individual is obese, the changes could be more dramatic. It is also essential to mention that the gut microbiota composition can variate within an individual (41, 83).

Serino et al. (2012) described that an increased Allistipes genus could be a protective factor to type 2 diabetes in an HFD scenario (84). A study by our group (83) demonstrated that probiotic administration in DIO animals could increase these genera. Unexpectedly, our study observed increased amounts of Allistipes in females FA/PM_2.5 compared to FA/FA and PM_2.5/FA compared to FA/PM_2.5. This result could also explain the less harmful effect of maternal exposure to PM_2.5.

Thingholm et al. (2019) observed a significant association between some individual microbial genera with obesity, pointing to the decrease in Akkermansia, Oscillibacter, and Allistipes (85). There were similarities in our results with this mentioned study. We also observed increased amounts of Oscillibacter in males FA/FA and FA/PM_2.5 compared to PM_2.5/FA. Still, our data differed slightly from Thingholm et al. (2019) when analyzing Akkermansia and Allistipes, as already discussed above (85).

However, some studies demonstrate that excessive mucin degradation can favor the access of pathogens to the mucosa (86, 87). Polluting particles can affect the protective characteristics of the mucus in the intestinal tract, changing the permeability and composition of the microbiota and impairing health (88). Although the offspring have not been subject to pollution, maternal exposure during gestation can result in placental epigenetic modification and fetus reprogramming (89).

The current study has some limitations regarding data collection time points. We measured food intake at 13-14 weeks and energy expenditure at 18-19 weeks. Nevertheless, the body weight at 13 weeks kept a similar pattern as 8 weeks in both genders, not representing a critical interpretation problem. In the current study, we evaluated gut microbiota composition only from mice at 30 weeks. Surprisingly, male but not female offspring lost the difference in their body weight and serum leptin levels. On the contrary, in another study (48), only male offspring but not female offspring from mothers exposed to CAP displayed increased adiposity in adulthood. However, in this other study, all mothers were exposed during pre-conception (48).

In summary, our data support the hypothesis that exposure to air pollution during gestation is more harmful to energy balance than exposure during lactation, independent of gender. Male offspring had an unfavorable metabolic phenotype at a young age. However, at an older age, only females kept more adiposity. Ultimately, our data highlights the importance of controlling air pollution, especially during gestation.

The original contributions presented in the study are publicly available. This data can be found here: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA894691. In addition, we included some raw data in the Supplementary Material . Further inquiries can be directed to the corresponding author.

PP, MV, AS conceived the study, designed experiments, and interpreted the results. OZ, CC, VY, AS, GC, CF, SN, and PP performed all experiments. PP wrote the manuscript. OZ, CC, Y-BK, AS, MV, PS, and MS helped with discussions and edited the manuscript. All authors contributed to the article and approved the submitted version.