Polychlorinated biphenyls (PCBs) are persistent halogenated organic pollutants that were synthesized for a host of commercial and industrial applications and products which include but are not limited to casting waxes, carbonless copy paper, paint, plastics, and inks (1). However, upon the realization of the health hazards posed from PCB exposure, production of the toxicants were halted but did not extinguish the presence of the detrimental pollutants as their lipophilic nature perpetuates their bioaccumulation and biomagnification (2). Thus, PCB exposure still occurs and health complications continue to persist (3). Environmental contamination from PCBs result from spills, leaks, and improper storage and disposal (4, 5). Once released, the environmental fates of PCBs are largely determined by their chlorination pattern (5) - as the chlorination pattern of the compound increases, so does its weight, viscosity, and lipophilicity (4). Routes of PCB exposure include inhalation, dermal contact, ingestion, and placental transfer (3, 6). Additionally, PCBs have been detected in umbilical cord and breastmilk and have been demonstrated to be transferred from mother to child during the perinatal period (6). It has been stated that the primary route of PCB exposure within the general population is primarily a result of the consumption of fatty foods such as dairy, animal products, and fish (7). Population studies demonstrate associations between coplanar PCBs and diabetes prevalence (8–10), incidence (11–13), and risk (14–16). In animal studies, such exposure has been shown to initiate and drive the development and progression of diabetes (17–23), and obesity (19, 22, 24–26). Existing evidence describing the contribution of early-life PCB exposure to diabetic-like phenotypes observed in mammalian offspring demonstrate sex-specific alterations in body composition (26) and metabolic parameters (27) as well as alterations in pro-inflammatory cytokines and hormones implicated in glucose regulation (28). However, such evidence fails to demonstrate the critical window of exposure during development that drives the long-term negative health outcomes or the genetic basis of observed phenotypes in mammalian species.

The harmful effects of coplanar PCBs, like 3, 3’, 4, 4’, 5-pentachlorobiphenyl (PCB126), are mostly elicited through its activation of the aryl hydrocarbon receptor (AhR) (29). This receptor is responsible for initiating the oxidation, reduction, and hydrolysis of xenobiotics in the detoxification pathway (30). The initial phase of reducing the toxicity of xenobiotics often results in the production of reactive oxygen species, which if not neutralized, can give rise to oxidative stress and damage (31). During the second phase of the detoxification pathway, xenobiotics and their respective metabolites are conjugated to increase their water-solubility for excretion (31). Genes involved in the second phase of detoxification are linked to the nuclear factor erythroid-2-related factor 2 (Nrf2) gene, which regulates antioxidant response by inducing the transcription of reactive oxygen species-detoxifying enzymes (32). Although the developmental role of Nrf2 in PCB-induced toxicity has been reported in zebrafish (33), these findings do not explain how Nrf2 allelic expression impacts diabetic-like phenotypes observed in mammalian offspring as a result of perinatal PCB exposure.

In the present study, we exposed Nrf2 heterozygous (Het) and Nrf2 knockout (KO) dams to PCB126 prior to and during pregnancy in an effort to delineate how both maternal and offspring Nrf2 genotype influences PCB-induced diabetic health outcomes in offspring. Dams were mated with sires of opposing Het and KO genotypes. Offspring body weight and body composition as well as glucose tolerance were monitored. Data collected from the offspring were analyzed from the perspective of offspring and maternal Nrf2 genotype and treatment, not paternal genotype. We hypothesized that PCB-induced detriments in offspring would be exacerbated by the lack of Nrf2 expression in offspring. We found that gestational PCB126 exposure precipitates adverse offspring metabolic and phenotypic responses but that offspring responses were modulated by maternal Nrf2 genotype, not offspring genotype.

Nrf2 has been identified as an integral regulator of redox homeostasis, and therefore its associated signaling pathway has been widely explored as a therapeutic target towards the prevention of a host of diseases (35, 36), including diabetes (37). Emerging data implicates the involvement of PCB exposure in the development and progression of diabetes (14–16). Unfortunately, literature insufficiently details the role of Nrf2 expression in PCB-induced diabetes. The current study focused on delineating the effects of Nrf2 allelic expression on offspring phenotypic and metabolic responses to gestational PCB126 exposure. Unpublished data from our laboratory demonstrates that PCB treatment during the nursing period did not significantly affect offspring body weight, body composition, or long-term glucose intolerance. Here, using a cross-fostering strategy, we provide additional evidence that the in utero period is the critical window that precipitates long-term negative developmental programming of altered body composition and impaired glucose tolerance in offspring. To our surprise, we found that maternal rather than offspring Nrf2 genotype impacted long-term PCB-induced developmental programming in offspring where offspring born to KO dams seemed to be more sensitive to PCB126.

Although little is known about PCB-induced diabetes and obesity resulting from developmental exposure, we expected offspring body weight and composition as well as glucose tolerance would be adversely affected by gestational PCB exposure. Our anticipation was largely based on studies in which direct PCB exposure occurred. Animals directly exposed to PCBs exhibit pronounced elevations in body weight as a result of exposure (22, 38–40). Moreover, limited data detailing the effects of direct PCB exposure on body composition demonstrate conflicting results and utilize diet as a means to understand the influence of PCB exposure on fat and lean mass (20, 39). Nonetheless, PCBs have been demonstrated to alter adipocyte development (21, 24, 25, 41) and death (25), and as a result have been deemed as obesogens. Further, direct PCB exposure has been shown to induce impairments in glucose (17, 18, 21, 23, 39, 40, 42) and insulin (21, 22, 39) tolerance. Previously, in seven-week-old offspring exposed to PCB126 during pregnancy and nursing, we observed that toxicant exposure did not influence offspring body weight but did alter offspring body composition in a sex- and dose-dependent manner (26). Furthermore, our unpublished data demonstrates that exposure to PCB126 exclusively during nursing does not affect offspring body weight or body composition but does affect offspring early-life glucose tolerance in a sex-dependent manner. Mennigen et al. (43) reported that gestational exposure to Aroclor 1221, a commercial PCB mixture, did not influence offspring pre-weaning and adolescent body weights. Dioxin exposure during pregnancy was reported to not affect offspring body weight or offspring later-life glucose and insulin tolerance (44). Perinatal PCB153 exposure has been shown to alter offspring glucose homeostasis in a sex-dependent manner, where male offspring experience elevated blood glucose levels and female offspring experience elevated glucagon levels (27). Although findings regarding the perinatal influence of PCB exposure on offspring body weight differ and evidence scantly describes the effects of such exposure on offspring body composition and glucose homeostasis, our current findings support and extend our earlier work (26) and reiterate the importance of timing of exposure in regards to offspring diabetes and obesity risk. Our current results demonstrate that gestational PCB126 exposure alters offspring body weight and composition as well as impairs glucose tolerance.

Interpretations of our aforementioned findings slightly vary when taking Nrf2 genotype into consideration. Nrf2 acts as an endogenous defense whose downstream mechanisms combat and reduce excess oxidants produced from xenobiotic insults, such as PCB exposure. Specifically, Nrf2 regulates the chemical induction of a number of detoxification pathway Phase II enzymes, which include but are not limited to glutathione peroxidase, heme oxygenase 1, and NAD(P)H dehydrogenase quinone 1 (NQO1) (45, 46). While we did not measure the gene expression of Nrf2 target genes in PCB-exposed offspring of the current study, we observed pronounced elevations in the hepatic mRNA expression of NQO1 in six-week-old WT and whole-body Nrf2 Het mice directly exposed to PCB126 in a separate study. Further, NQO1 levels were significantly reduced in both male and female KO mice compared to WT mice ( Supplementary Figure 2 ). Because of the aforementioned findings in addition to the known importance of Phase II response in the detoxification of PCBs, in the current study, we hypothesized that offspring lacking Nrf2 expression would be more susceptible to the long-term health detriments associated with in utero PCB exposure. Rather, we observed that there were no significant interactions between offspring genotype and perinatal PCB exposure. Instead, we found that offspring of KO dams exposed to PCB126 during gestation are more sensitive to the long-term negative consequences of PCB exposure than offspring of Het dams that were exposed to PCB126 during gestation. PCB-exposed offspring born to KO dams had increased fat mass and significantly impaired glucose disposal whereas PCB-exposed offspring born to Het dams did not ( Figures 4 – 6 ). The developmental role of Nrf2 in PCB-induced toxicity has been previously described in zebrafish (33); however, this does not extend to rodent models. To our knowledge, the present study is the first to report the developmental role of Nrf2 in PCB-induced toxicity in rodents. Previously, mice have been utilized to delineate the developmental role of Nrf2 in neural function and/or behavior in response to pharmaceutical (47) and recreational drugs (48), antioxidants (49), as well as heavy metals (49). Diet studies have been used to better understand the role of Nrf2 expression in obesity and diabetes. Studies investigating the effects of Nrf2 deficiency on body weight and composition report divergent results. Nrf2 deficiency in animals on standard or chow fed diets have been shown to have comparable body weights (50–53) and body fat compositions (50) to WT and/or Nrf2 mice with enhanced activity, as well as lower early-life body weights when compared to their WT littermates (54). Differences in body weights reported by Pi et al. (54) are likely due to the decreased adipose tissue mass possessed by Nrf2 KO animals as they also report the ability of Nrf2 deficiency to impair adipogenesis and reduce susceptibility to diet-induced obesity. Similarly, conflicting results have been reported about the influence of Nrf2 deficiency in the presence of a high fat diet (HFD). It has been stated that deficient animals fed a HFD have lower body weight and less weight gain than WT mice (51), as well as no differences in body weight and body fat composition (50). In diabetic mice, Nrf2 deficiency worsens hyperglycemia (55). However, in the presence of a HFD, Nrf2 deficiency improves glucose tolerance when compared to WT and/or Nrf2 mice with enhanced activity (51, 53). Conversely, it has been reported that a HFD impairs glucose tolerance in both WT and KO animals, where the deficiency worsens the observed impairment (52). The findings from the aforementioned studies demonstrate the dual roles of Nrf2 in obesity and insulin resistance. While the present study did not investigate the influence of Nrf2 expression on diabetes and obesity from a dietary standpoint, PCB126 exposure could be considered a ‘second-hit’ similar to HFD feeding. Our findings demonstrate the ability of Nrf2 genotype, maternal genotype in this case, to modulate PCB-induced diabetes and obesity.

We were surprised to find that the influence of maternal Nrf2 genotype on offspring metabolic and phenotypic responses to gestational PCB exposure supersedes that of the offspring Nrf2 genotype. To better understand this finding, we conducted a literature search on the effects of maternal genotype on offspring response respective and irrespective to environmental contaminant exposure. Obtained search results were not relevant and further demonstrate the need to investigate the influence of the maternal gene-toxicant exposure interactions on offspring response. Thus, we propose that our future studies examining this phenomenon will measure PCB126 parent compound levels in dams throughout pregnancy, in fetuses, and in offspring throughout the lifecycle with the hopes of understanding more about placental PCB transfer respective to Nrf2 genotypes and its short- and long-term effects on offspring response. Our preliminary findings in non-pregnant females with varying Nrf2 genotypes (WT, Het, and KO) showed no differences in sera and liver PCB126 parent compound measurements (data not shown). These data suggest that genotype does not influence PCB126 levels; however, this does not account for potential hormonal differences, placental effects, or hepatic Nrf2 expression changes a result of pregnancy. Thus, we propose the need to examine the placental transport of PCB126 parent compound in pregnant females with varying Nrf2 genotypes to assess the influence of placental oxidative stress and inflammation on offspring responses as a result of gestational PCB exposure. Ahmed et al. (28) reported that PCB126 gestational exposure disrupts placental tissues. Moreover, the current study would have benefited from an analysis of the glucose tolerance of Het and KO dams prior to and during pregnancy +/- PCB exposure. Further, one could argue that our experiments should have included WT comparisons or a Het/Het breeding scheme. We chose not to go this route because it would allow for us to look for maternal genotype differences, and the Het/Het breeding scheme would have produced many mice that were unnecessary. Our rationale for this study design is based on our preliminary analyses; we performed PCB exposure experiments in non-pregnant female WT and whole-body Het and KO mice and found that (1) PCB126 parent compound levels within the sera and livers of WT and Het animals were not significantly different, (2) AUC for blood glucose measurements did not differ between WT and Het animals ( Supplementary Figure 1 ). It is important to note that our decision to eliminate WT animals from this study did not influence our ability to obtain sufficient data to answer our research question. Furthermore, the present study was designed to assess PCB-induced detriments in offspring using body composition analyses and glucose tolerance testing. This study could have benefited from employing insulin tolerance tests and serum measurements in addition to glucose stimulated insulin secretion assays to further validate the observed detrimental metabolic outcomes in offspring resulting from gestational PCB exposure. Lastly, future experiments should include a prolonged fasting period before glucose tolerance testing as well as utilize intraperitoneal administration of the glucose bolus to respectively tease out differences between groups and eliminate any potential effects of incretin hormones on insulin function and glucose homeostasis. In conclusion, this study demonstrates that in utero PCB exposure caused long-lasting alterations in offspring body weight, body composition, and glucose tolerance. This result reaffirms the importance of the timing of exposure and suggests that maternal interventions during the gestational period should be explored to lessen the negative effects of toxicant-induced health complications in offspring. Further, data from the current study show the influence of the maternal genotype on offspring response to early-life toxicant exposure. Although the developmental role of Nrf2 in PCB-toxicity is poorly understood, our work demonstrates that maternal allelic expression should not be ignored.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by University of Kentucky Institutional Animal Care and Use Committee (IACUC).

BR, SN, and KP contributed to the conception and design of this study. BR, SN, OT, AA, FS, CC, SH, CH, NM, and KP contributed to data collection and/or interpretation of the data collected. BR SN, OT, AA, FS, CC, SH, CH, NM, HS, and KP contributed to the drafting of the manuscript and/or revising critically for important intellectual content. All authors have approved the manuscript and agree to be accountable for the content of the work.

This study and core services were supported by US NIH grants (National Institute of General Medical Sciences, P20GM103527, and the National Institute of Environmental Health Sciences, P42ES007380 and P30ES026529). The Southern Regional Education Board Doctoral Scholars Program supported BR. CH was supported by the SURES (Summer Undergraduate Research in Environmental Health Sciences) program funded by NIEHS R25ES027684. NM was supported by Berea College Office of Internships & Career Development.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.