Parallel in vivo and in vitro transcriptomics analysis reveals calcium and zinc signalling in the brain as sensitive targets of HBCD neurotoxicity

Abstract

Hexabromocyclododecane (HBCD) is a brominated flame retardant (BFR) that accumulates in humans and affects the nervous system. To elucidate the mechanisms of HBCD neurotoxicity, we used transcriptomic profiling in brains of female mice exposed through their diet to HBCD (199 mg/kg body weight per day) for 28 days and compared with those of neuronal N2A and NSC-19 cell lines exposed to 1 or 2 µM HBCD. Similar pathways and functions were affected both in vivo and in vitro, including Ca²⁺ and Zn²⁺ signalling, glutamatergic neuron activity, apoptosis, and oxidative stress. Release of cytosolic free Zn²⁺ by HBCD was confirmed in N2A cells. This Zn²⁺ release was partially quenched by the antioxidant N-acetyl cysteine indicating that, in accordance with transcriptomic analysis, free radical formation is involved in HBCD toxicity. To investigate the effects of HBCD in excitable cells, we isolated mouse hippocampal neurons and monitored Ca²⁺ signalling triggered by extracellular glutamate or zinc, which are co-released pre-synaptically to trigger postsynaptic signalling. In control cells application of zinc or glutamate triggered a rapid rise of intracellular [Ca²⁺]. Treatment of the cultures with 1 µM of HBCD was sufficient to reduce the glutamate-dependent Ca²⁺ signal by 50%. The effect of HBCD on zinc-dependent Ca²⁺ signalling was even more pronounced, resulting in the reduction of the Ca²⁺ signal with 86% inhibition at 1 µM HBCD. Our results show that low concentrations of HBCD affect neural signalling in mouse brain acting through dysregulation of Ca²⁺ and Zn²⁺ homeostasis.

Keywords: Androgen; BFR; Dihydrotestosterone; Glutamate; GnRH; Neurotoxicity; Oestrogen; Prolactin; Transcriptomics.

Fig. 1

Viability of N2A and NSC19 cells exposed to HBCD: a N2A and b NSC19 cells were incubated for 48 h with a geometric series of concentration between 1.56 and 50 µM of HBCD and viability was measured with the MTT assay. Three independent experiments were performed using eight replicates in each, and the average between replicates and experiments are reported (n = 3) with error bars showing the standard deviation. The regression curves were fitted in SigmaPlot (Version 12), using the “Exponential Decay, Single, 3 Parameter” model. The cellular viability is express in percentage of the DMSO control. A caspase-3 assay was used as indicator of apoptosis in the a N2A cell line and b in the NSC19 cell line exposed to HBCD. The two cell cultures were incubated for 24 h with four different concentrations of HBCD (1, 2, 4 or 8 µM). A caspase-3 fluorescence assay was then performed and the fluorescence measured every 15 min in a microplate reader with 485 nm excitation and 530 nm emission, and is here expressed in arbitrary units. Three independent experiments were performed using eight replicates in each, and the average between replicates and experiments are reported (n = 3). Error bars denote standard deviation and an asterisk represents a p < 0.05

Fig. 2

Venn diagram of genes called significantly regulated (q < 0.2) in microarray analysis of mouse brain, N2A cells and NSC19 cells exposed to HBCD. Mice were exposed to HBCD via the diet for 28 days at a dose of 199 mg/kg body weight per day, resulting in a HBCD concentration in the brain of 40 µmol/kg (wet weight). N2A and NSC19 cells were exposed to either 1 or 2 µM of HBCD for 24 h and the numerals shown are the combined numbers of unique genes called significant in the two conditions. The full lists of differentially regulated genes, fold-change data and associated p- and FDR can be found in Supplementary Table 2. The Venn diagram was constructed using Venny (http://bioinfogp.cnb.csic.es/tools/venny/index.html)

Fig. 3

Pathway analysis of genes called significantly regulated (q < 0.2) in microarray analysis of mouse brain, N2A cells and NSC19 cells exposed to HBCD. Mice were exposed to HBCD via the diet for 28 days at a dose of 199 mg/kg body weight per day, resulting in a HBCD concentration in the brain of 40 µmol/kg (wet weight). N2A and NSC19 cells were exposed to either 1 or 2 µM of HBCD for 24 h. Lists of significantly regulated genes from each condition were subjected to pathway analysis, using the IPA software, and statistical significance of overrepresentation of genes in different “Canonical Pathways” and “Diseases & Biofunctions” are shown as a heat-map along with predicted “Upstream regulators”. Only selected pathways that had significant enrichment in mouse brain plus two cell culture conditions are shown. The full data-set including genes in each pathway is presented in Supplementary Tables 4A–D

Fig. 4

Intracellular [Zn²⁺] in N2A cells exposed to HBCD for 24 h. N2A cells were exposed to HBCD with or without other treatments for 24 h and intracellular Zn²⁺ release was detected using the Zn²⁺-specific fluorescent probe, Zinpyr-1. a N2A cells were incubated for 24 h with 1 µM HBCD with or without 2 h pre-incubation with the antioxidant, NAC, or the zinc chelator, DEDTC. b The N2A cell line was incubated for 24 h with 1 µM HBCD with or without 2 h pre-incubation with the NO synthase blocker, L-NAME or with DEDTC. The experiment was carried out three times with 16 technical replicates in each experiment (n = 3). Data are expressed as fold-change of fluorescence measured in the control. The bars represent the averages of values from the three experiments and error bars show the standard deviation. An asterisk and a hash sign indicate a statistical differences from untreated control cells and cells treated with 1 µM HBCD, respectively (ANOVA, p < 0.05)

Fig. 5

HBCD effects on glutamate- and zinc-dependent Ca²⁺ transients. Primary hippocampal mixed neuronal cultures were exposed to 1 or 2 µM of HBCD for 24 h. Cells were loaded with Fura-2 AM and intracellular Ca²⁺ responses to (a, b) glutamate (200 µM, 30 s) or (c, d) Zn²⁺ (200 µM, 30 s) were monitored. A representative Ca²⁺ response (a, c) and the averaged initial rates of fluorescence change (b, d) are shown. Note that the Ca²⁺ response is monitored after a short lag time following the application of the ligand, likely due to the timing of the perfusion system. The initial rate of the response is dependent on the net influx of Ca²⁺ into the cytoplasm and represents the activity of GPR39. Inserts to b show representative DAPI staining of neuronal cultures, seeded with the same number of cells, following control (DMSO alone) or HBCD treatment at the indicated concentration. Inserts to c show: left panel: neurons loaded with the Zn²⁺ sensitive dye, Fluozin-3 AM, treated with Zn²⁺ (200 µM, 30 s) as marked by the bar; right panel: Neurons loaded with Fura-2 AM, treated with Thapsigargin (200 nM) and ATP (5 µM) as marked, and subsequent to the depletion of Ca²⁺ stores and recovery of the fluorescent signal to baseline, Zn²⁺ (200 µM, 30 s) was applied as marked. The bars represent the arithmetic mean of three independent experiments, each consisting of at least 7 averaged replicates for each condition (n = 3), **p < 0.01