Polychlorinated biphenyls 105 and 118 form thyroid hormone receptor agonists after cytochrome P4501A1 activation in rat pituitary GH3 cells

Abstract

Background: Polychlorinated biphenyls (PCBs) may interfere with thyroid hormone (TH) signaling by reducing TH levels in blood, by exerting direct effects on TH receptors (TRs), or both.

Objective: Our objective was to identify individual PCBs that directly affect TH signaling by acting on the TR.

Methods: We administered a mixture of six PCB congeners based on their ortho substitution pattern, including PCBs 77 and 126 (non-ortho), PCBs 105 and 118 (mono-ortho), and PCBs 138 and 153 (di-ortho), to pregnant Sprague-Dawley rats from gestational days (G) 6 to 16. This mixture, or various combinations of the components, was also evaluated in a transient transfection system using GH3 cells.

Results: The mixture reduced serum TH levels in pregnant rats on G16 but simultaneously up-regulated the expression of malic enzyme in liver. It also functioned as a TR agonist in vitro; however, none of the individual PCB congeners comprising this mixture were active in this system. Using the aryl hydrocarbon receptor (AhR) antagonist alpha-naphthoflavone, and the cytochrome P450 (CYP)1A1 antagonist ellipticine, we show that the effect of the mixture on the thyroid hormone response element required AhR and CYP1A1.

Conclusions: We propose that PCB 126 induces CYP1A1 through the AhR in GH3 cells, and that CYP1A1 activates PCB 105 and/or 118 to a form a compound that acts as a TR agonist. These data suggest that some tissues may be especially vulnerable to PCBs interfering directly with TH signaling due to their capacity to express CYP1A1 in response to coplanar PCBs (or other dioxin-like molecules) if sufficient mono-ortho PCBs are present.

Keywords: AhR; CYP1A1; PCB metabolism; endocrine disruption; thyroid hormone.

Figure 1

Chemical structures of PCB congeners that constitute the mixture used in the animal studies as described in the text. (A) Non-ortho PCB congeners: 3,3’,4,4’-tetrachlorobiphenyl, PCB 77, and 3,3’,4,4’,5-pentachlorobiphenyl, PCB 126. (B) Mono-ortho PCB congeners: 2,3,3’,4,4’-pentachlorobiphenyl, PCB 105, and 2,3’,4,4’,5-pentachlorobiphenyl, PCB 118. (C) Di-ortho PCB congeners: 2,2’,3,4,4’,5’-hexachlorobiphenyl, PCB 138, and 2,2’,4,4’,5,5’-hexachlorobiphenyl, PCB 153.

Figure 2

Effect of PCB treatment on serum total T₄ (A) and ME mRNA in liver (B) of pregnant Sprague-Dawley rats at the time of sacrifice on G16. Error bars represent mean ± SE (A) or mean ± SE ME/β-actin (B). Numbers of animals in each group are as follows: (A) Control, 6; Dose 1, 8; Dose 2, 8. (B) Control, 5; Dose 1, 6; Dose 2, 7. See “Materials and Methods” for treatment details. **p < 0.01, significantly different from control group using the Bonferroni t-test after one-way ANOVA.

Figure 3

Effects of 1 × 10⁻⁷ M T₃ (A), 10 μM PCB Mix 6 (B), or individual PCB congeners (see Table 4 for concentrations) (C) treatments on relative luciferase activity in GH3 cells. Error bars represent mean ± SE of relative luciferase activity normalized to control wells. All treatments were performed in triplicate, and the final results obtained from three separate experiments. Values are reported as percent control for the purpose of illustration. **p < 0.01, significantly different from control (vehicle treatment) group using a Student t-test.

Figure 4

Role of AhR in PCB 126-induction of cytochrome P450 genes. Cells were treated with 10μM PCB 126 in the presence or absence of 1 x 10⁻⁶ M of the AhR antagonist, α-NF. The levels of CYP1A1 (A) and CYP1B1 (B) mRNAs were measured by real-time PCR. PCB 126 significantly increased CYP1A1 (A) and CYP1B1 (B) mRNAs and α-NF abrogated these effects. Error bars represent mean ± SE CYP1A1/β-actin (A) or CYP1B1/β-actin (B) mRNAs and are expressed as fold induction over vehicle alone (DMSO). ^ap < 0.01, significantly different from control group using the Bonferroni t-test after one-way ANOVA. ^bp < 0.01, ^cp < 0.05, significantly different from PCB 126 –treated group using the Bonferroni t-test after one-way ANOVA.

Figure 5

Effects of cytochrome P450 antagonsits on T₃-induced (A) and PCB Mix 6–induced (B) relative luciferase activity in GH3 cells. “–” indicates no treatment; “+” indicates treatment with compound shown left of row. α-NF and TMS were used at concentrations of 10⁻⁶ M; ellipticine was used at 10⁻⁷ M. Error bars represent mean ± SE of relative luciferase activity reported as percent control for the purpose of illustration. ^ap < 0.01, significantly different from control group using the Bonferroni t-test after one-way ANOVA. ^bp < 0.01, significantly different from PCB Mix 6–treated group using the Bonferroni t-test after one-way ANOVA.

Figure 6

Effects of PCB congener combinations on expression of CYP1A1 mRNA (A) and TR-mediated relative luciferase activity (B) in GH3 cells. (A) PCB concentrations are described in Table 4; CYP1A1 mRNA was measured by real-time PCR. Only treatment groups that included PCB 126 significantly increased CYP1A1 expression. (B) PCB congener concentrations are described in Table 4. Error bars represent mean ± SE of relative luciferase activity and values are reported as percent control for the purpose of illustration. An increase in relative luciferase activity was observed when cells were PCBs 126, 105, and 118. **p < 0.01 [significantly different from corresponding PCB combination group not treated with PCB 126 (A) or group treated with PCBs 118 and 105 (B) using the Bonferroni t-test after two-way ANOVA].