Type 1 diabetes (T1D) is one of the most common chronic metabolic disorders in children and adolescents, with an estimated worldwide incidence of 128,900 new cases and corresponding prevalence of nearly 1,100,000 existing cases in the under 20 year age group (Patterson et al., 2019). Genetic factors are known to play a significant role in the development of autoimmune disorders, including T1D in humans, potentially in combination with environmental factors (Bach, 2001; Rook, 2012; Regnell and Lernmark, 2013). In recent decades, age at diagnosis of T1D has declined (Group, 2008; Harjutsalo et al., 2013), suggesting that the impact of environmental factors during early development has increased. Factors such as maternal infections or infections in early life, deficiency of specific nutrients during pregnancy and/or early childhood, have been associated with risk of T1D in observational studies (Stene and Gale, 2013). Other proposed risk factors of T1D include alterations in gut microbiota (Chatenoud et al., 2012) and exposure to environmental chemicals during development (Bodin et al., 2015). However, the impact of environmental factors on disease development is not well characterized. While epidemiological studies have identified specific associations between environmental factors and disease risk, causal relationships require also controlled, experimental in vitro and in vivo exposure models.
Environmental chemicals, particularly persistent organic pollutants (POPs) including per- and polyfluoroalkyl substances (PFAS), organochlorine pesticides (OCPs), polychlorinated biphenyls (PCBs) and brominated flame retardants (BFRs), such as hexabromocyclododecane (HBCD) and polybrominated diphenyl ethers (PBDEs), are commonly detected in humans as demonstrated by the European Human BioMonitoring data dashboard (HBM4EU, 2020). Although humans are exposed to complex, man-made chemical mixtures via several sources and routes, knowledge of adverse effects caused by such chemicals is mainly based on studies focusing on a single or few chemicals at a time. However, combined exposure to multiple chemicals may cause synergistic or additive adverse health effects, even if the levels of all substances included in the chemical exposure mixture are below their individual safety thresholds (JRC, 2018).
Humans are exposed to PFAS, PCBs and BFRs through food (Harrad et al., 2004; Hites et al., 2004; Karrman et al., 2006; Haug et al., 2011), via inhalation (Casey et al., 1999; Mandalakis et al., 2008) and indoor dust (Wilford et al., 2004; Hazrati and Harrad, 2006; Hwang et al., 2008), as well as through breastfeeding (Karrman et al., 2007). After intake and absorption, perfluorinated compounds predominantly associate with proteins and occur at the highest concentrations in blood, liver and kidneys (Karrman et al., 2006; Lau, 2012), while the main site of PCB and BFR accumulation is adipose tissue (Mullerova and Kopecky, 2007). During the early development in humans, POPs can pass the placental barrier, reach and deposit in fetal tissues and organs (Porpora et al., 2013; Li et al., 2020a). They are, in general, found in fetal organs at concentrations lower than in maternal serum, but similar to placental levels (Mamsen et al., 2019).
In living organisms, many POPs are known to cause adverse health impacts. For example, PFAS exposure has been associated with endocrine disruption (Lopez-Espinosa et al., 2012), increased cholesterol levels (Li et al., 2020b), altered metabolic processes in fat and liver (Bassler et al., 2019), and has been linked with autoimmune disorders such as celiac disease (Sinisalu et al., 2020) and type 1 diabetes (McGlinchey et al., 2020). Several PCB congeners have been associated with adverse developmental (Mol et al., 2002), endocrine (Vasiliu et al., 2006; Cao et al., 2008; Chen et al., 2008) and immunological (Heilmann et al., 2006; Lee et al., 2007; Park et al., 2008) effects. Adverse health consequences of exposure to BFRs have also been observed, including thyroid disorders (Stapleton et al., 2011; Zota et al., 2011), diabetes (Lee et al., 2010) and neurodevelopmental (Gascon et al., 2011) health effects.
The number of exposure studies investigating autoimmune disease progression in animal models are limited. We have previously reported that exposure to a single PFAS, perfluoroundecanoic acid (PFUnDA), caused a non-monotonic dose-response effect on the immune system in NOD mice, where low and intermediate exposures resulted in a delay of onset of autoimmune diabetes, while exposure to very high levels accelerated insulitis development (Bodin et al., 2016). Further, exposure of the NOD mice to a POP mixture containing 7 PCBs, 9 OCPs, 7 BFRs and 6 PFAS (Berntsen et al., 2017) resulted in a trend of decreased peritoneal macrophage phagocytosis (Berntsen et al., 2018). Additionally, an early event of insulitis development, namely a reduced number of F4/80-tissue resident macrophages, was also apparent in the histological analysis of pancreas in these mice. We have previously shown that immunotoxic effects of a POP mixture in in vitro studies of both rodent and human blood-derived macrophages were mainly caused by the PFAS in the mixture (Berntsen et al., 2018).
Human studies investigating early life exposure to POPs and their role in T1D development show inconsistent results. A cross-sectional study including 820 patients with T1D showed a significant association between high PFAS levels in blood serum and lower risk of T1D development in children and adults (Conway et al., 2016). Conversely, elevated PFOS levels were reported in young patients newly diagnosed with T1D (Predieri et al., 2015), while another birth-cohort study found no evidence that fetal and early life exposure to POP (including 14 PFAS) was a significant risk factor for later T1D development (Salo et al., 2019). Other studies suggest that PFAS exposure is associated with hyperglycemia, serum HDL cholesterol and increased blood insulin (Lin et al., 2009), may lead to immunotoxicity (Borg et al., 2013), and modulates neonatal serum phospholipids associated with increased risk of T1D (McGlinchey et al., 2020). Overall, the molecular mechanisms underlying the effects of exposure are still not well characterized.
Herein we investigated metabolic alterations in NOD mice following pre- and postnatal exposure to a mixture of persistent organic pollutants, in which the concentrations of POPs were based on estimated daily intake levels reported in the Nordic population (Berntsen et al., 2017). Since PFAS was the most potent chemical group in the POP mixture regarding macrophage phagocytosis (Berntsen et al., 2018), we focused on possible effects of PFAS on the metabolome.