Bromination Pattern of Hydroxylated Metabolites of BDE-47 Affects Their Potency to Release Calcium from Intracellular Stores in PC12 Cells

Background Brominated flame retardants, including the widely used polybrominated diphenyl ethers (PBDEs), have been detected in humans, raising concern about possible neurotoxicity. Recent research demonstrated that the hydroxylated metabolite 6-OH-BDE-47 increases neurotransmitter release by releasing calcium ions (Ca2+) from intracellular stores at much lower concentrations than its environmentally relevant parent congener BDE-47. Recently, several other hydroxylated BDE-47 metabolites, besides 6-OH-BDE-47, have been detected in human serum and cord blood. Objective and Methods To investigate the neurotoxic potential of other environmentally relevant PBDEs and their metabolites, we examined and compared the acute effects of BDE-47, BDE-49, BDE-99, BDE-100, BDE-153, and several metabolites of BDE-47—6-OH-BDE-47 (and its methoxylated analog 6-MeO-BDE-47), 6′-OH-BDE-49, 5-OH-BDE-47, 3-OH-BDE-47, and 4′-OH-BDE-49—on intracellular Ca2+ concentration ([Ca2+]i), measured using the Ca2+-responsive dye Fura-2 in neuroendocrine pheochromocytoma (PC12) cells. Results In contrast to the parent PBDEs and 6-MeO-BDE-47, all hydroxylated metabolites induced Ca2+ release from intracellular stores, although with different lowest observed effect concentrations (LOECs). The major intracellular Ca2+ sources were either endoplasmic reticulum (ER; 5-OH-BDE-47 and 6′-OH-BDE-49) or both ER and mitochondria (6-OH-BDE-47, 3-OH-BDE-47, and 4′-OH-BDE-49). When investigating fluctuations in [Ca2+]i, which is a more subtle end point, we observed lower LOECs for 6-OH-BDE-47 and 4′-OH-BDE-49, as well as for BDE-47. Conclusions The present findings demonstrate that hydroxylated metabolites of BDE-47 cause disturbance of the [Ca2+]i. Importantly, shielding of the OH group on both sides with bromine atoms and/or the ether bond to the other phenyl ring lowers the potency of hydroxylated PBDE metabolites.


Research
Brominated flame retardants are added to a wide array of consumer products. Adverse effects of these compounds on the developing nervous system give cause for concern (for review, see Costa and Giordano 2007). Longlasting neurobehavioral changes after neo natal exposure have been detected in rodents after exposure to poly brominated diphenyl ethers (PBDEs) (Branchi et al. 2003;Eriksson et al. 2001;Lilienthal et al. 2006;Rice et al. 2007;Viberg et al. 2003Viberg et al. , 2006, as well as various functional and structural alterations in the brain (Dingemans et al. 2007;Viberg 2009;Xing et al. 2009).
During development, acute effects on neuronal activity may result in long-lasting changes in proteins, brain function, and behavior. A key regulator for neuronal function is intracellular concentration of calcium ions, [Ca 2+ ] i , which regulates many cellular processes, including the release of neurotransmitters at the pre synaptic terminal (for reviews, see Clapham 2007;Laude and Simpson 2009).
Recently, not only PBDEs but also various hydroxylated metabolites of PBDEs have been found to bio accumulate in humans (Athanasiadou et al. 2008;Qiu et al. 2009). Therefore, the aim of the present study was to determine whether mono hydroxylated metabolites of the abundant BDE-47 affect [Ca 2+ ] i and to compare this with the effects of a methoxylated analog and several other environmentally relevant PBDE congeners.
Background: Brominated flame retardants, including the widely used polybrominated diphenyl ethers (PBDEs), have been detected in humans, raising concern about possible neurotoxicity. Recent research demonstrated that the hydroxylated metabolite 6-OH-BDE-47 increases neuro transmitter release by releasing calcium ions (Ca 2+ ) from intra cellular stores at much lower concentrations than its environmentally relevant parent congener BDE-47. Recently, several other hydroxylated BDE-47 metabolites, besides 6-OH-BDE-47, have been detected in human serum and cord blood.   [Dingemans et al. 2008;see also Supplemental Material (doi:10.1289/ehp.0901339)].
Cell viability assay. To investigate possible acute effects of the PBDEs on cell viability, we used the alamarBlue (AB) and neutral red (NR) uptake assays with minor modifications [Magnani and Bettini 2000;Repetto et al. 2008;see also Supplemental Material (doi:10.1289/ehp.0901339)].
Intracellular Ca 2+ imaging. Changes in [Ca 2+ ] i were measured using the Ca 2+ -sensitive fluorescence ratio dye Fura-2 as described previously (Dingemans et al. 2007(Dingemans et al. , 2008 2+ ] i to > 175% of baseline was used to determine no observed effect concentrations (NOECs) and lowest observed effect concentrations (LOECs). A transient increase within 0-10 min from the start of exposure is referred to as an "initial increase," whereas additional increases are referred to as "late increases" (Dingemans et al. 2008). At NOEC levels, based on average and amplitude of [Ca 2+ ] i levels, increases in [Ca 2+ ] i levels to > 175% of baseline during exposure were scored as fluctuations to investigate more subtle effects on Ca 2+ homeostasis. We determined the number of cells showing fluctuations in [Ca 2+ ] i , as well as frequency, amplitude, and duration of these fluctuations.
Data analysis and statistics. All data are presented as mean ± SE from the number of cells or fluctuations in [Ca 2+ ] i . We performed statistical analyses using SPSS 16 (SPSS, Chicago, IL, USA). Categorical and continuous data were compared using Fisher's exact test and Student's t-test, respectively, paired or unpaired where applicable. Analysis of variance (ANOVA) and post hoc t-tests were performed to investigate possible doseresponse relationships. To investigate structure-activity relationships, we investigated the possible influence of hydroxylation position and/or shielding of the OH group by adjacent atomic groups on the efficacy to increase [Ca 2+ ] i . To this aim, we performed a multifactorial ANOVA using the mean increase in basal [Ca 2+ ] i induced by 20 µM of the different OH-PBDEs as the dependent variable. Hydroxylation position (ortho, meta, or para) and the presence of either one or two shielding atomic groups (phenyl ring and/or Br atom) adjacent to the OH group were used as fixed   variables ( Figure 1). A p-value < 0.05 was considered statistically significant. Hydroxylated BDE47 metabolites increase [Ca 2+ ] i . After 20 min exposure to 6-OH-BDE-47, 5-OH-BDE-47, or 4´-OH-BDE-49, the NR assay indicated a decrease in cell viability only at the 20-µM dose, with means ± SEs of 86 ± 0.3%, 86 ± 1.1%, and 93 ± 0.8% of control, respectively. We observed no significant decreases in cell viability measured by the NR assay for 6´-OH-BDE-49 or 3-OH-BDE-47. In the AB assay, exposure to 6-OH-BDE-47, 6´-OH-BDE-49, or 3-OH-BDE-47 increased the relative fluorescence intensity dose-dependently above the control level (data not shown). For 6´-OH-BDE-49, 3-OH-BDE-47, and 4´-OH-BDE-49, the LOEC for increased [Ca 2+ ] i is 20 µM. Exposure to 20 µM 6´-OH-BDE-49 results in an initial increase in 50% of the cells. We observed late increases less frequently and with lower amplitude. The shapes of the increases in [Ca 2+ ] i during exposure to 20 µM 6´-OH-BDE-49 vary widely ( Figure 4B). We observed initial increases during exposure to 20 µM 3-OH-BDE-47 in 48% of the cells. Exposure to 20 µM 4´-OH-BDE-49 resulted in a modest initial increase compared with baseline in 83% of the cells, after which we observed a larger late increase.
The mean amplitude of [Ca 2+ ] i during exposure to 20 µM of the hydroxylated metabolites was independent of the position (ortho, meta, or para) of the OH group on the PBDE molecule (ANOVA, not significant). However, OH-PBDEs in which the OH group was shielded on only one side, with either the other phenyl ring or a Br atom, induce significantly higher increases in [Ca 2+ ] i compared with OH-PBDEs in which the OH group was shielded on both sides (ANOVA, p < 0.01). However, this influence of the shielding of the OH group (on one compared with two sides) was independent of its position (ortho, meta, or para) on the PBDE molecule (ANOVA, not significant). 2+ ] i induced by 6-OH-BDE-47, 5-OH-BDE-47, and, to a lesser extent, 4´-OH-BDE-49 were reduced in Ca 2+ -free conditions, increases were still observed for all of the OH-PBDEs (Figure 4). To identify the responsible Ca 2+ stores, endoplasmic reticulum (ER) or both ER and mitochondrial Ca 2+ stores were depleted by either 1 µM thapsigargin (TG) or 1 µM TG plus 1 µM carbonyl cyanide 4-(trifluoro methoxy)phenyl hydrazone (FCCP). TG and TG/FCCP pretreatment, respectively,  When we depleted ER Ca 2+ stores in Ca 2+free conditions, the initial increase induced by 5 µM 6-OH-BDE-47 was largely diminished ( Figure 4A). Although the amplitude of the late increase was not different in Ca 2+ -free conditions, it increased after depletion of the ER Ca 2+ stores, but greatly diminished after depletion of both ER and mitochondrial Ca 2+ stores.

Hydroxylated BDE47 metabolites increase [Ca 2+ ] i by release of Ca 2+ from intra cellular stores. Although initial increases in [Ca
We also observed the variation in [Ca 2+ ] i responses during exposure to 20 µM 6´-OH-BDE-49 in Ca 2+ -free conditions ( Figure 4B). After depletion of ER Ca 2+ stores, we no longer observed either initial or late transient increases in [Ca 2+ ] i . , Ca 2+ -free saline (0 mM Ca 2+ ), Ca 2+ -free saline after pretreatment with TG (0 + TG), and Ca 2+ -free saline after pretreatment with both TG and FCCP (0 + TG + FCCP). Cells were pretreated with TG and FCCP to deplete Ca 2+ stores in ER and mitochondria, respectively. Tops and bottoms of boxes represent upper and lower quartiles, the line within the box is the median, whiskers represent lowest and highest values, and circles represent outliers (for clarity, outliers that are more than three interquartile ranges from the boxes are not shown). Data are from three to seven experiments per treatment. Numbers shown in parentheses indicate the number of cells used for data analysis; the percentages of responding cells (with increase in [Ca 2+ ] i to > 175% of baseline) are denoted above each box. Representative traces of [Ca 2+ ] i measurements of individual PC12 cells exposed to OH-PBDE for 10 min (applied as indicated by arrowheads) in external saline (containing 1.8 mM Ca 2+ ) and in Ca 2+ -free conditions are shown below. *p < 0.05, **p < 0.01, and #p < 0.001.  When the ER Ca 2+ stores were depleted, the amplitudes of both initial and late increases induced by 5 µM 5-OH-BDE-47 were largely diminished ( Figure 4C). After depletion of both ER and mitochondrial Ca 2+ stores, the late increase was further reduced.
We observed mono phasic increases during exposure to 20 µM 3-OH-BDE-47 in Ca 2+ -free conditions, with similar amplitude ( Figure 4D). After depletion of the ER Ca 2+ stores, the amplitude was significantly decreased, and even further after depletion of ER and mitochondrial Ca 2+ stores.
During exposure to 20 µM 4´-OH-BDE-49 after depletion of the ER Ca 2+ stores, the amplitude of the initial increase largely diminished ( Figure 4E). The amplitude of the late increase was not significantly changed in Ca 2+ -free conditions. After depletion of the ER Ca 2+ stores, the amplitude significantly decreased, and even further after depletion of ER and mitochondrial Ca 2+ stores. Table 1; also see Supplemental Material, Table 4

Discussion
The OH-PBDE metabolite 6-OH-BDE-47 has previously been shown to disrupt [Ca 2+ ] i in PC12 cells by releasing Ca 2+ from intra cellular stores at lower concentrations than its parent compound BDE-47 (Dingemans et al. 2008). The results presented here demon strate that other hydroxylated metabolites of BDE-47 also induce Ca 2+ release from intra cellular stores, whereas the methoxylated analog (6-MeO-BDE-47) and the parent compounds lack this effect (Figure 2). Experiments in Ca 2+ -free conditions (Figure 4) indicate that the initial increases induced by 6-OH-BDE-47, 5-OH-BDE-47, and 4´-OH-BDE-49 are partly caused by influx of extra cellular Ca 2+ . The initial and late increases induced by 6-OH-BDE-47 are caused by release of Ca 2+ from ER and mitochondria, respectively. The widely varying increase induced by 6´-OH-BDE-49 was caused primarily by release from ER Ca 2+ stores. Both initial and late increases induced by 5-OH-BDE-47 and the increase induced by 3-OH-BDE-47 are caused by release of Ca 2+ mainly from ER, but also from mitochondria. Both initial and late increases induced by 4´-OH-BDE-49 are caused by release of Ca 2+ from ER, but the late increases are also from mitochondria. When investigating fluctuations in [Ca 2+ ] i , we detected subtle effects of BDE-47 on the Ca 2+ homeostasis and observed lower LOECs for several OH-PBDEs (Table 1).
We detected no or mild effects on cell viability for the investigated PBDEs and MeO/ OH-PBDEs, indicating that the observed effects on [Ca 2+ ] i were not confounded by cyto toxicity. The AB assay, which is based on mitochondrial activity, appeared to be less useful in determining cell viability because for 6-OH-BDE-47, 6´-OH-BDE-49, and 3-OH-BDE-47 the relative fluorescence intensity increased dose dependently above the control level, suggesting induction of mitochondrial activity. This may be related to mitochondrial uncoupling, which was previously demonstrated for 6-OH-BDE-47 in isolated zebrafish mitochondria (van Boxtel et al. 2008).
The low basal [Ca 2+ ] i of PC12 cells is maintained by the removal of Ca 2+ ions by the plasma membrane Ca 2+ ATPases and Na 2+ -Ca 2+ exchanger (Duman et al. 2008;for review, see Westerink 2006). Additionally, Ca 2+ can be sequestered into organelles, mainly ER and mitochondria. Both 3-OH-BDE-47 and 4´-OH-BDE-49 induce increases in [Ca 2+ ] i even after depletion of both ER and mitochondria by TG and FCCP. Ca 2+ has also been shown to accumulate in endosomes, lysosomes, secretory granules, the Golgi apparatus, and nucleus (for review, see Laude and Simpson 2009). The Golgi apparatus stores Ca 2+ via sarco plasmic/ER Ca 2+ ATPase pumps (Missiaen et al. 2007), which are inhibited by TG. Therefore, it is unlikely that release of Ca 2+ from the Golgi apparatus caused the additional increase, whereas this remains unclear for the other mentioned organelles.
At concentrations not affecting the average and amplitude of increases in [Ca 2+ ] i , BDE-47, 6-OH-BDE-47, and 4´-OH-BDE-49 caused an increase in the frequency, amplitude, and/or duration of fluctuations in [Ca 2+ ] i . These subtle effects on Ca 2+ homeo stasis resulted in lower NOECs for most of these brominated flame retardants, particularly BDE-47 (Table 1). Ca 2+ signals vary from micro domains to globally across the cell and from milli seconds to many hours (Laude and Simpson 2009 Moody and Bosma 2005). Therefore, the observed effects of OH-PBDEs at low concentrations can be of relevance for the development of the nervous system. None of the parent PBDEs except BDE-47 showed any effects on Ca 2+ homeostasis in PC12 cells. Nonetheless, neuro toxic effects of  have been detected at different biological levels (Costa and Giordano 2007). Because of the lack of effects by parent PBDEs in the present study, no effects of bromination pattern could be investigated. We confirmed that the activity of the OH-PBDEs depended on the presence of the OH group, because no effects were observed during exposure to the methoxylated analog of 6-OH-BDE-47. The higher activity of 6-OH-BDE-47 compared with its methoxylated analog is in line with other studies, mostly on endocrine effects (Cantón et al. 2008;Kojima et al 2009;van Boxtel et al. 2008). The mean amplitude of increases in [Ca 2+ ] i could not be related to the location of the OH group on the PBDE mole cule. However, it appeared that when the OH group was shielded on both sides by either the other phenyl ring and/or Br atoms (as in 6´-OH-BDE-49 and 3-OH-BDE-47), the OH-PBDE increased [Ca 2+ ] i less than when the OH group was less shielded (as in 6-OH-BDE-47, 5-OH-BDE-47, and 4´-OH-BDE-49). Also, the OH-PBDEs with only one shielded side of the OH group induced release of Ca 2+ from ER at the lowest concentrations (6-OH-BDE-47 and 5-OH-BDE-47) or with the highest amplitude in Ca 2+ -free conditions (4´-OH-BDE-49). Thus, the toxicity of OH-PBDEs appears attenuated by shielding of the OH group on both sides by either the other phenyl ring and/or Br atoms.
Several animal studies confirmed the genera tion of hydroxylated metabolites of PBDEs in vivo (Hakk et al. 2009;Malmberg et al. 2005;Marsh et al. 2006), and that OH-PBDEs were also formed in human liver cells exposed to BDE-99 (Stapleton et al. 2009). Interestingly, marine organisms have also been shown to produce hydroxylated and methoxylated metabolites (Hakk and Letcher 2003).
Only very recently, the occurrence and accumulation of hydroxylated metabolites were confirmed in humans (Athanasiadou et al. 2008), with total OH-PBDE serum concentrations up to 120 pmol/g lipids. Another study also detected OH-PBDEs in U.S. fetal serum samples and confirmed the bioaccumulation of these metabolites (Qiu et al. 2009). Moreover, they demon strated that concentrations of OH-PBDEs were similar or sometimes even higher than the concentration of PBDEs. Fetal total OH-PBDE serum concentrations ranged from 2.01 to 899.1 ng/g lipids (median, 21.96 ng/g lipids). The most abundant BDE-47 metabolites found in fetal blood were 5-OH-BDE-47 and 6-OH-BDE-47 (Qiu et al. 2009). It is concerning that these metabo lites caused an increase of [Ca 2+ ] i at much lower concentrations than the other metabolites of BDE-47 investigated.
All five OH-PBDEs investigated in the present study caused Ca 2+ release from intracellular Ca 2+ stores, although with different LOECs. Likely, depending on the position, the OH group and adjacent phenyl ether and/ or Br atoms, other hydroxylated metabolites of tetra and pentaPBDEs have a similar effect on cellular calcium homeostasis. The median (21.96 ng/g lipids) and highest (899.1 ng/g lipids) concentrations of total OH-PBDEs observed in fetal plasma correspond to approximately 0.4 and 17.4 nM, respectively, in blood (calculated with average physiologic parameters). Thus, the highest concentration observed in human blood is only two orders of magnitude lower than the LOEC for Ca 2+ release from intra cellular stores by OH-PBDEs (1 µM). Moreover, the LOEC for increased Ca 2+ fluctuations is even lower (0.2 µM), meaning that the margin of exposure is insufficient in some individual exposure situations. Also, because OH-PBDEs are not associated with lipids-as are the parent PBDEs-but have a high affinity for plasma proteins (Verreault et al. 2005), the estimated blood concentration calculated from exposure values at a lipid-weight-adjusted basis (nanograms per gram lipids) may be underestimated. However, the LOEC (1 µM) used to calculate the margin of exposure is higher for other metabolites, and it remains to be determined whether the observed effects on fluctuations in [Ca 2+ ] i could result in functional or even adverse effects in vivo.
Because exposure to organo halogen compounds within the time frame of rapid brain development can result in behavioral defects in mice (for review, see Costa and Giordano 2007), it is concerning that children are exposed to these environmental pollutants prenatally and postnatally. Moreover, several studies have observed inter actions between environmental pollutants to enhance (neuro) toxicity. Additive and synergistic neuro toxic effects of polychlorinated biphenyls (PCBs) and PBDEs have been detected in vivo  and in vitro (Gao et al. 2009). Concern about possible effects on the developing brain arises from the fact that an increase in [Ca 2+ ] i by release from intra cellular stores appears to be a common mechanism for OH-PBDEs and ortho-PCBs (for review, see Mariussen and Fonnum 2006). Therefore, a possible additive effect of these environmental pollutants with respect to increases in cytosolic [Ca 2+ ] i , which not only is a trigger for neurotransmitter release but affects many cellular processes (Clapham 2007), is not unlikely.
Because very high concentrations of PBDEs are occasionally measured in humans, the voluntary and legislative measures to reduce the release of PBDEs into the environ ment appear justified. Also, hydroxylated metabolites of PBDEs, which were recently found to bioaccumulate in humans (Athanasiadou et al. 2008;Qiu et al. 2009), either from man-made PBDEs or of natural origin, are currently not taken into account in regulatory human risk assessment. The results presented here reveal a structure-activity relation ship for metabolites of PBDEs (more shielding of the OH group reduces the potency of OH-PBDEs) and reinforce that oxidative metabolism should be included in human risk assessment of persistent organic pollutants.