Early embryogenesis in zebrafish is affected by bisphenol A exposure

The drastic advancement in industrialization and technology and the growth in human population in the past century have resulted in unprecedented environmental changes in the human history. The production of large amounts of synthetic industrial and biomedical chemicals as well as pollutants poses a risk to our ecosystem and induces negative effects on the health of wildlife and human beings. Some of the more damaging chemical contaminants are classified as endocrine disrupting chemicals (EDCs) because they can interfere with the synthesis, metabolism, and action of endogenous hormones (Phillips et al., 2008; Phillips and Foster, 2008). EDCs exert different biological effects via diverse mechanisms of actions (Judson et al., 2009; Rhind, 2008; Wigle et al., 2008). EDCs are believed to cause damages to human health and the ecological systems. With the emergence of the global problem of chemical contamination, the adverse biological effects of EDCs are gaining attention among the scientific communities, industry, governments, non-governmental organizations, and the public. There is an increasing need for the identification and quantification of all these ubiquitous chemical contaminants. The possible routes of exposure of humans to the EDCs are through the environments, consumer products, and foods (Feron et al., 2002; Mantovani et al., 2006; Poppenga, 2000; Wigle et al., 2008). To safeguard the public health, instrumental chemical analysis has been adopted globally for assessing the risk of human exposure to EDCs and their metabolites (Hotchkiss et al., 2008). Considering the severe long-term impact of EDCs on public health, a sensitive animal model is required to assess the risks of the EDCs for protecting human and ecological health.

Rapid structural and functional changes occur during the fetal life making it a vulnerable period of development. The process of development is not a simple process of unfolding the inherited genetic program, followed by the commitment of cells to specific lineages, and structural and functional differentiation in respective organs/tissues. Developmental plasticity in animals can be influenced by both genomic (epigenetic and genetic) and environmental factors, which leads to considerable changes in the developmental path for adaptive responses in the fetus (Bateson et al., 2004; Gluckman et al., 2009; Gluckman et al., 2008; Gluckman and Hanson, 2007; Gluckman et al., 2007; Gluckman et al., 2005a; Gluckman et al., 2005b; Hanson and Gluckman, 2008). To fill the information gap between exposures to EDCs and the outcomes of developmental failure, an experimental model that enables us to investigate the early developmental stages is essential. Zebrafish has been extensively used in developmental biology and has become an attractive model for chemical screening. This is a highly scalable model with a well-established genome database (Barros et al., 2008; Yeh et al., 2009; Zon and Peterson, 2005). This model has been used in general toxicology studies for decades. General toxicology studies such as identification of the median lethal concentration (LC₅₀) and end-point phenotype have been performed in zebrafish after bisphenol A (BPA) exposure (Duan et al., 2008; McCormick et al., 2011; Saili et al., 2012). Recently, next-generation sequencing technology was used to identify potential genes that are altered after BPA exposure in zebrafish embryos (Lam et al., 2011). However, the unique developmental features of zebrafish have not been used in many studies for characterization of exposure to effect. It is difficult to monitor the effects of EDCs on early development by using mammalian embryos; therefore, zebrafish, which has the ability of external fertilization, is used as an alternative model. Gibert and his colleagues used different in situ molecular markers to examine the developmental stage of otolith formation in zebrafish after BPA exposure (Gibert et al., 2011). Here, we hypothesize that the primary action of EDCs is to prevent normal development during early embryogenesis and cell fate determination (i.e. cell signaling and epigenetic modification) and thus affect normal development (i.e. cell fate determination and organogenesis), which leads to organ dysfunction. In this study, we used the zebrafish model to show that exposure of zebrafish embryos to low doses of BPA caused disturbance in dorsal/ventral patterning and segmentation, which provides a new insight on developmental toxicology of environmental pollutants.

We used the AB wild-type line in this study. The zebrafish were raised and staged as described previously (Kimmel et al., 1995). All experimental procedures on zebrafish embryos were approved by the Hong Kong Baptist University, Hong Kong Special Administrative Region.

BPA (Sigma–Aldrich, USA) was dissolved in DMSO and diluted in egg medium (E3 medium). BPA was used at a final concentration of 50 µM in all experiments, which is comparable to the concentrations used in other studies (Lam et al., 2011; Sun et al., 2009). Embryos at 1–4 cell stage were directly exposed to BPA in 2 ml of E3 medium in a 6-well plate. The embryos were grown at 28°C for the selected time points (stages), 8 hours post-fertilization (hpf) (60–75% epiboly), 14 hpf (8–10 somite), 24 hpf (prim-5), and 72 hpf (protruding mouth). Control embryos were treated with equal volume of DMSO as that in the BPA-exposed embryos.

We used the whole-mount in situ hybridization (WISH) procedure for screening on the basis of our previous study (Tse and Jiang, 2012). Briefly, BPA-exposed embryos were collected at 3 stages 60–75% epiboly, 8–10 somite (ss), and prim-5 and were fixed in 4% paraformaldehyde (PFA). Standard WISH procedure was applied using zebrafish embryos. Plasmids that were used to make antisense mRNA probes have been published previously: chd (Miller-Bertoglio et al., 1997), eng2 (Schier et al., 1996), eve1 (Joly et al., 1993), gata2 (Detrich et al., 1995), gsc (Stachel et al., 1993), krox20 (Strähle et al., 1993), myoD (Weinberg et al., 1996), pax2a (Krauss et al., 1991), and otx2 (Heisenberg et al., 1996).

BPA is one of the most common EDCs, and the chemical properties and toxicities of BPA have been reported. BPA is a selective estrogen receptor modulator (Richter et al., 2007) and can interact with thyroid hormone receptors (Moriyama et al., 2002; Zoeller et al., 2005) and peroxisome proliferator-activated receptors (Riu et al., 2011). At the physiological levels, BPA is suggested to be a factor attributed to the development of metabolic disorders in humans, such as cardiovascular diseases, obesity, and insulin resistance (Polyzos et al., 2012; vom Saal et al., 2012). A considerable number of studies in rodents have reported the negative effects of BPA on the function and development of reproductive and neuronal systems (Jašarević et al., 2011; Wolstenholme et al., 2011; Xi et al., 2011). More importantly, female mice prenatally exposed to BPA showed a decrease in fertility and fecundity (Cabaton et al., 2011) and had an adverse effect on the fertility of the male offspring (Salian et al., 2009). Furthermore, BPA administration in rodents could disturb neurons in the substantia nigra (Tando et al., 2007) and in the hippocampus (Kunz et al., 2011). In previous studies, zebrafish embryos have been exposed to BPA at concentrations similar to those used in this study, and otolith malformations (70 µM) and cardiac edema (65 µM) have been reported (Duan et al., 2008; Gibert et al., 2011). Although the general effects and the effects of BPA on development in rodents and zebrafish have been reported, important information about the effects of BPA in the initial stages of cell development remains to be addressed. To understand the mechanism underlying these effects is important because these data could reveal the fundamental cause of the observed effects; further, the data can be utilized to predict and evaluate the impact of in utero EDC exposure. In this study, we performed screening in the early stage of embryogenesis; we selected 3 critical stages, including dorsoventral (DV) patterning (60–75% epiboly), segmentation (8–10 ss), and brain development (prim-5), within 24 hpf (Tse et al., 2009; Tse et al., 2011).

DV patterning is an important developmental process in zebrafish (Schmitz and Campos-Ortega, 1994). Several zebrafish mutants have been identified on the basis of their dorsal or ventral phenotypes, which range from C5 (dorsal) to V4 (ventral) (Mullins et al., 1996; Kishimoto et al., 1997). In this study, we targeted on exposure to BPA at the early development period (within 24 hours). Follow-up examination of the effects of the exposure was performed up to 3 days after fertilization while mild dorsalization (mainly C1–C3) was observed. Dorsalization was characterized by their phenotype of shortened posterior parts during the development (Fig. 1). Because the DV patterning occurs in the early stage of embryogenesis, the effects of the BPA action can be observed by using selected in situ molecular markers (ventral markers, eve1 and gata2; dorsal markers, chd and gsc) at stage of 60–75% epiboly (Tse et al., 2009). Among various validated markers, eve1 is a zebrafish homeobox gene similar to even-skipped in Drosophila (Joly et al., 1993). eve1 is strongly expressed in the ventrolateral marginal cells. The other gene marker gata2 is a hematopoietic transcription factor gene (Detrich et al., 1995) for ventral ectoderm and hematopoietic cells in the ventral mesoderm. To trace the dorsal patterning, we used 2 dorsally expressed markers chordin (chd) and goosecoid (gsc) (Sasai et al., 1995; Stachel et al., 1993). The expression patterns of eve1 and gata2 in embryos exposed to BPA were more restricted in the ventral half of the marginal and the animal zone than that in the controls (Fig. 2A–D). On the other hand, the expression of dorsal markers chd and gsc was greater in the embryos exposed to BPA (Fig. 2E–H). We measured the angles of expressions of the markers (Fig. 2I,J). Taken together, embryos at the 60–75% epiboly stage exposed to BPA showed reduced expression levels of the ventral markers but increased expression levels of the dorsal markers.

To monitor the trend of altered DV patterning, gata1 and pax2a were used as the markers at 8–10 somite stage. gata1 is ventrally expressed in presumptive hematopoietic cells in 2 lateral stripes (Detrich et al., 1995; Kimmel et al., 1990), while pax2a is used for marking the presumptive neural region (Krauss et al., 1991). The gata1 marker showed widening of the 2 lateral stripes of presumptive hematopoietic cells in BPA-exposed embryos (Fig. 3A,B). Additionally, pax2a staining showed a diffused expression pattern in the mid-hindbrain boundary (mhb). The otic vesicles were missing in the embryos exposed to BPA (Fig. 3C,D). On the other hand, the somite muscle widened in BPA-exposed embryos (Fig. 3E,F). The phenotype was further confirmed by using the myoD somite marker that is expressed in the dorsal mesoderm and somite muscles (Kimmel et al., 1990; Weinberg et al., 1996). Weak and abnormal myoD expression was detected in the BPA-exposed embryos (Fig. 3G,H). In addition to DV patterning, the follow-up in situ experiments illustrated the effects of BPA on somatic muscle formation in the segmentation period.

On the basis of the diffused pax2a expression in the mhb region at the 8–10 somite stage, we suspected that BPA affected brain development during the developmental process. To prove this assumption, the genetic markers krox20, otx2, and eng2b were used to monitor the brain regionalization process at the prim-5 stage (Finkelstein and Boncinelli, 1994; Joyner and Guillemot, 1994; Stuart et al., 1994). Brain regionalization is one of the fundamental processes in the early stages of vertebrate brain development, including the formation of the mhb and its adjacent brain regions (Joyner and Guillemot, 1994). The hindbrain develops into a series of rhombomeres along the anterior–posterior axis of the neural tube. Rhombomeres are believed to be involved in neuronal organization in brain development (Moens and Prince, 2002), while rhombomere 3 (r3) and 5 (r5) could be identified using the transcription factor, krox20 (Oxtoby and Jowett, 1993). In this study, BPA exposure resulted in abnormal and unorganized krox20 expression in the prim-5-stage embryos (Fig. 4A,B). The reduced size of the mhb shown by the eng2b mhb structure marker (Ekker et al., 1992; Fjose et al., 1992) was also found in the BPA-exposed embryos (Fig. 4C,D). These data were consistent with the results of decreased pax2a expression pattern at the mhb region observed in the 8–10 somite stage embryos exposed to BPA (Fig. 3D). To support this observation, an additional marker otx2 was used to confirm if BPA affects the development of midbrain structure (Mercier et al., 1995). The expression of otx2 was decreased in the BPA-exposed embryos, which indicated that the midbrain development was also affected (Fig. 4E,F). The altered percentage was consistent from the early stage to later stages, which suggested the phenotypes might due to the defect in early development. Furthermore, it should be noted that the diffused expression patterns were unlikely to be caused by the developmental delay. Collectively, BPA exposure disturbed the process of brain regionalization, which resulted in the development of abnormal rhombomeres, restricted mhb, and smaller midbrain structure.

Although our study does not provide a detailed mechanism of how BPA affects development, it strengthens our understanding about the developmental defect caused by BPA exposure. The resulting phenotype can be caused by complicated crosstalk between signaling pathways. Further studies should be performed to understand how and why the molecular markers listed above were affected. Zebrafish used in this study can act as a screening model to focus the research on the specific time point, organ, and potential signaling pathway involved in development.

In this study, the standard in situ hybridization method was used to examine the effects of EDCs on early embryogenesis. We found that BPA exposure influences DV patterning, somite formation, and brain development in zebrafish embryos. Our study showed the potential use of zebrafish for validating the effects of EDC at a particular developmental stage.

This work was supported by the Collaborative Research Fund (HKBU1/CRF/08 to C.K.C.W.), University Grants Committee, Hong Kong. The work in the W.K.F.T. laboratory is supported by the Faculty start-up fund (BIOL3840101).