Converging Effects of Three Different Endocrine Disrupters on Sox and Pou Gene Expression in Developing Rat Hippocampus: Possible Role of microRNA in Sex Differences

Abstract

Endocrine disrupting chemicals (EDCs) can impair hippocampus-dependent behaviors in rat offspring and in children. In search for key processes underlying this effect, we compared the transcriptomes of rat hippocampus on postnatal day 6 after gestational and lactational exposure to three different EDCs at doses known to impair development of learning and memory. Aroclor 1254, a commercial PCB mixture (5 mg/kg or 0.5 mg/kg), or bisphenol A (5 mg/kg or 0.5 mg/kg) were administered in chow, chlorpyrifos (3 mg/kg or 1 mg/kg) was injected subcutaneously. Male hippocampus exhibited a common effect of all three chemicals on genes involved in cell-autonomous processes, Sox6, Sox11, Pou2f2/Oct2, and Pou3f2/Brn2, all upregulated at the high dose. Additional genes of the Sox and Pou families were affected by only one or two of the chemicals. Real time RT PCR showed a comparable expression change for bisphenol A also at the lower dose. Female hippocampus exhibited much fewer genes with expression changes (almost none with false discovery rate <0.05), and none of the genes of the Sox and Pou families was affected. Since gene network analyses in male hippocampus suggested a link between Sox6 and miR-24, known to be repressed by activation of ER-alpha and to repress Sox6 in other tissues, this microRNA was measured. miR-24 was downregulated by all chemicals at the high dose in males. Values of Sox6 mRNA and miR-24 were inversely correlated in individual male hippocampus samples, supporting the hypothesis that the change in Sox6 expression resulted from an action of miR-24. In contrast, miR-24 levels remained unchanged in hippocampus of females. A sexually dimorphic response of miR-24 may thus be at the basis of the sex difference in Sox6 expression changes following exposure to the three chemicals. ER-alpha expression was also sex-dependent, but the expression changes did not parallel those of potential downstream genes such as Sox6. Sox6 is known to suppress differentiation of Parvalbumin (Pvalb)-expressing interneurons. Individual Sox6 levels (FPKM) were inversely correlated with levels of Pvalb, but not with markers of Sox6-independent interneuron subpopulations, Nos1 and 5HT3aR. Effects on interneuron development are further suggested, in males, by expression changes of Nrg1 and its receptor Erbb4, controlling interneuron migration. Our study disclosed new types of EDC-responsive morphogenetic genes, and illustrated the potential relevance of microRNAs in sexually dimorphic EDC actions.

Keywords: development; endocrine disrupter; hippocampus; microRNA; pou gene; sex difference; sox6; transcriptomics.

FIGURE 1

Epression ratio of potassium-chloride-co-transporter KCC2 (Slc12a5), neuron-specific chloride ion extruder, to sodium-potassium-chloride-co-transporter NKCC1 (Slc12a2), increasing intracellular chloride concentration, in individual hippocampus samples of male and female rat offspring at postnatal day 6. RNAseq data, ratio of FPKM (fragments per kilobase per million reads), mean ± SD and individual samples (M, male, F, female). Two way ANOVA, sex difference (p < 0001), but no treatment effects and no sex × treatment interaction.

FIGURE 2

Gene network of male rat hippocampus at postnatal day (PND) 6, with estrogen receptor alpha (Esr1). Up-regulation (red circles) or down-regulation (blue circle) of hippocampal genes exposed to Aroclor 1254 (5 mg/kg, A), bisphenol A (5 mg/kg, B), or chlorpyrifos (3 mg/kg, C) (see Table 4). Pou2f1/Oct1 and Sox11 exhibited a false discovery rate >0.05 after bisphenol A, but the effect was confirmed by real time RT PCR (Figure 3). Literature data indicate a possible inhibitory action of Esr1 on transcription of miR-24-3p, and of miR-24-3p on transcription of Sox6 (MetaCore, Shortest Paths network. TR: transcription, B: binding, M: miR binding, green: activation, red: inhibition).

FIGURE 3

mRNA levels in male hippocampus at postnatal day 6 after pre- and postnatal administration of Aroclor 1254 (ARO0.5: 0.5 mg/kg, ARO5: 5 mg/kg), bisphenol A (BPA0.5: 0.5 mg/kg, BPA5: 5 mg/kg), or chlorpyrifos (CPF1: 1 mg/kg, CPF3: 3 mg/kg). Quantitative RT PCR of hippocampus samples not used for transcriptomics. Mean ± SEM, number of samples (one/litter). Difference from control: *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 4

A. Level of miR-24-3p in hippocampus of male and female rat offspring at postnatal day 6 after pre- and postnatal administration of Aroclor 1254 (5 mg/kg, Aro5), bisphenol A (5 mg/kg, BPA5), or chlorpyrifos (3 mg/kg, CPF3). Quantitative RT PCR, individual values with mean ± S.D, *p < 0.05 different from control group. CON: controls, ARO5: Aroclor 1254, 5 mg/kg, BPA5: bisphenol A, 5 mg/kg, CPF3: chlorpyrifos 3 mg/kg. B. Relationship between miR-24-3p (% of control) and Sox6 mRNA (% of control, real time RT PCR) in individual hippocampus samples of male offspring. Inverse correlation, p = 0.0097, r = −0.4645, slope = −0.264. Green symbols: control, brown: Aroclor 1254 5 mg/kg, blue: bisphenol A 5 mg/kg, red: chlorpyrifos 3 mgkg.

FIGURE 5

(A) Estrogen receptor-alpha (ER-alpha, Esr1) in hippocampus of male and female offspring at postnatal day 6. Quantitative RT PCR, mean ± S.E.M, number of samples. ANOVA, ****p < 0.0001, ***p < 0.001, *p < 0.05. (B) Esr1 mRNA level is higher (131.8%) in male than female hippocampus. Quantitative RT PCR, mean ± S.E.M, number of samples. Unpaired t test, p < 0.05. (C) Ratio of ER-alpha (Esr1)/ER-beta (Esr2) in hippocampus of individual male and female offspring is increased in males by treatment with Aro, BPA and CPF at the higher dose level. RNAseq data, FPKM (fragments per kilobase per million reads), two-way ANOVA, treatment p = 0.0060, sex p = 0.0002, interaction treatment x sex p = 0.0113. CON: controls, ARO0.5, ARO5: Aroclor 1254 0.5 mg/kg and 5 mg/kg, BPA0.5, BPA5: bisphenol A 0.5 mg/kg and 5 mg/kg, CPF1, CPF3: chlorpyrifos 1 mg/kg and 3 mg/kg.

FIGURE 6

Relationship between Sox6 and interneuron markers expressed in individual hippocampus samples of male offspring at postnatal day 6 (RNAseq data, FPKM values). (A) Inverse correlation between expression of Parvalbumin (Pvalb) expressed in medial ganglionic eminence (MGE)-derived interneurons and Sox6, p = 0.0039, r = −0.6155, slope = −6.065. No correlation between Sox6 and levels of neuronal nitric oxide synthase (Nos1) expressed in subgroup of MGE-derived interneurons, or serotonergic ionotropic receptor Htr3a expressed in caudal ganglionic eminence-derived interneurons, both independent of Sox6. (B) Correlation between Pvalb and Sox6 with identification of treatment groups. C. Expression ratio of Pvalb/Nos1, mean ± S.E.M, N = 5-6.