Transcriptomic pathway and benchmark dose analysis of Bisphenol A, Bisphenol S, Bisphenol F, and 3,3',5,5'-Tetrabromobisphenol A in H9 human embryonic stem cells

Highlights

• In vitro treatment of, estrogen receptor negative, H9 human embryonic stem cells (hESC; WA09) with Bisphenols (BP) A, F, S and Tetrabromo BPA (TBBPA) resulted in perturbed gene expression.

• All bisphenols affected a large number of genes in a dose-dependent manner.

• BPA, BPF, and BPS exhibited similar potencies and TBBPA, the least structurally similar bisphenol, exhibited much lower potency.

• This study suggests potential impacts of these chemicals on embryonic stem cells.

Abstract

Bisphenol A (BPA) is a chemical used in the manufacturing of plastics to which human exposure is ubiquitous. Numerous studies have linked BPA exposure to many adverse health outcomes prompting the replacement of BPA with various analogues including bisphenol-F (BPF) and bisphenol S (BPS). Other bisphenols are used in various consumer applications, such as 3,3′,5,5’-Tetrabromobisphenol A (TBBPA), which is used as a flame retardant. Few studies to date have examined the effects of BPA and its analogues in stem cells to explore potential developmental impacts. Here we used transcriptomics to investigate similarities and differences of BPA and three of its analogues in the estrogen receptor negative, human embryonic stem cell line H9 (WA09). H9 cells were exposed to increasing concentrations of the bisphenols and analyzed using RNA-sequencing. Our data indicate that BPA, BPF, and BPS have similar potencies in inducing transcriptional changes and perturb many of the same pathways. TBBPA, the least structurally similar bisphenol of the group, exhibited much lower potency. All bisphenols robustly impacted gene expression in these cells, albeit at concentrations well above those observed in estrogen-positive cells. Overall, we provide a foundational data set against which to explore the transcriptional similarities of other bisphenols in embryonic stem cells, which may be used to assess the suitability of chemical grouping for read-across and for preliminary potency evaluation.

Introduction

Bisphenol A (BPA) is an industrial chemical used in the manufacture of polycarbonate plastic and is present in many consumer products including can linings, epoxy resins in dental sealants, and thermal paper receipts (Cao et al., 2010; Cooper et al., 2011; Fleisch et al., 2010). Consequently, human exposure to BPA is ubiquitous. The US National Health and Nutrition Examination Survey (NHANES III) and the Canadian Health Measures Survey, found detectable levels of BPA in 90–93% of the urine samples analyzed in the population in ages six years and older (LaKind and Naiman, 2015; LaKind et al., 2019; Findlay and Kohen, 2015). Further, the levels of BPA in urine span a large range of concentrations, with levels of exposure demonstrating demographic associations to gender, race/ethnicity, environmental, consumer, and lifestyle factors (Calafat et al., 2005; Calafat et al., 2008; Huang et al., 2018). BPA has also been detected in human serum, milk, and placenta at concentrations ranging from 0.2 to 20 ng/g and exceeding 100 ng/g in one study of placental tissue (Vandenberg et al., 2007). Furthermore, BPA has been detected in maternal serum and amniotic fluid during fetal development (Shekhar et al., 2017). Human exposure to BPA is widespread with no reduction in human exposure for over a decade (LaKind et al., 2019).

Many animal studies report developmental effects of BPA. For example, in utero and early postnatal exposure to BPA is associated with adverse effects in offspring such as impaired brain development and function, including changes in dopamine responsiveness in neurons and astrocytes (Kunz et al., 2011; Miyatake et al., 2006). Others have shown that prenatal exposure to BPA can delay the development of cortical neurons in the prenatal stage during embryonic development and induce lasting epigenetic disruption in the brain (Kumar and Thakur, 2017). Public concern regarding BPA's prevalence and potential toxicity, including endocrine disruptor effects (Prins et al., 2019), has prompted industry to replace BPA with other structural analogs such as bisphenol S (BPS) and bisphenol F (BPF) (Fig. 1A) (Rosenmai et al., 2014). Other bisphenols, such as 3,3′,5,5′-tetrabromobisphenol A (TBBPA), are used as flame retardants in the electronics industry (Colnot et al., 2014). In humans, BPS and BPF have been detected in urine at concentrations comparable to BPA (Karrer et al., 2020). TBBPA has also been detected in human urine at similar or higher concentrations than BPA (Ho et al., 2017).

Due to the structural similarities of BPF, BPS, and TBBPA to BPA (Fig. 1A), there is concern regarding developmental toxicity and neurotoxicity for these BPA analogs, similarly to BPA. Toxicological studies for some of the replacement analogs have shown similar toxicities to BPA (Hercog et al., 2019; Kojima et al., 2019; Pelch et al., 2019; Ahmed and Atlas, 2016). In addition, NHANES data demonstrate associations between BPA, BPF, and BPS exposure and obesity in children and adolescents and show that BPF exposure is positively associated with a higher risk of obesity observed in boys (Liu et al., 2019). Several studies show that some analogs are as or more potent than BPA for several toxicity endpoints. For example, BPS is a more potent inducer of adipogenic genes than BPA in 3 T3-L1 mouse embryonic fibroblasts (Ahmed and Atlas, 2016). Further, BPS induced a distinct transcriptomic profile but a similar phenotype to BPA in differentiating primary human stem cells (Boucher et al., 2016a). Few studies have examined or compared the potencies of BPA analogs for developmental toxicity in vitro or in vivo within the same study (Moreman et al., 2017). Using transcriptomics, one recent study examined the effect of BPS, BPF, and BPA in differentiating mouse embryonic stem cells and found that the analogues were at least as toxic as BPA (Yin et al., 2019). However, the effect of these chemicals on undifferentiated human stem cells is not known.

With thousands of new industrial chemicals introduced each year, there is a critical need to reduce reliance on animal testing and to develop highly predictive human-based biological models for safety assessments. Regulatory agencies face challenges in keeping pace with the evaluation of new chemical entities and substitutions for potentially toxic chemicals currently in use in consumer products and industrial applications. This is of particular concern with chemicals like BPA and its analogs for which there is widespread human exposure (Pelch et al., 2019). There is a clear need to develop methodologies to rapidly and comprehensively assess the potential toxicity of structurally similar chemicals to assess whether structural or functional analogs exert similar effects and whether they can or cannot be grouped under a single chemical profile. Global transcriptional profiling enables a direct comparison of the broad biological response networks to test articles, providing an effective approach to address this gap. Specifically, transcriptomic analyses provide a means to determine similarities and differences in gene expression induced by exposure to structurally related chemicals and to assess relative potencies using traditional benchmark concentration (BMC) analyses.

In this study we used RNA-sequencing to characterize the concentration-response for global transcriptomic alterations induced by BPA, BPF, BPS, and TBBPA in H9 (WA09) human embryonic stem (hES) cells. H9 cells were initially established from the inner cell mass of a blastocyst, a structure formed during the earliest steps of development of fertilized human embryos that develops into the fetus (Thomson et al., 1998). This cell line offers the unique ability to study effects of chemicals on cells that retain the fundamental self-renewal and pluripotency (“stemness”) properties of stem cells, are at the origin of human development, and offer a genomic landscape distinct from that of mature somatic cells routinely used in toxicology (Thomson et al., 1998). In vivo, stem cells enable embryogenesis, tissue repair, and self-renewal (Pazianos et al., 2003). Therefore, the use of H9 hES cells offers the ability to potentially understand chemical effects on early human embryonic development and the genomic pathways that regulate “stemness” in an in vitro setting. Since bisphenols have induced adverse developmental effects in animal studies (Iwamuro et al., 2003; Kabuto et al., 2004; Liu et al., 2018; Zhou et al., 2011), analysis of bisphenol toxicity to hES cells is of relevance to potential human health effects. Further, the embryonic nature of H9 cells is such that they do not express estrogen receptors, and thus these cells provide a model for identifying non-estrogen receptor-dependent effects of bisphenols.

To assess the relative biological activity and potency of BPA and structural analogs BPS, BPF, and TBBPA, we examined the stem cell transcriptome in H9 hES cells following exposure to these chemicals at concentrations that covered greater than three orders of magnitude (1 μM to >100 μM). Overall, this study provides a broad overview of the patterns of gene expression changes induced by BPA and the structural analogs BPS, BPF, and TBBPA on the transcriptome of hES cells and demonstrates how these data can be applied to derive BMC information useful for comparing relative potencies of these structurally related chemicals.