T3 Induces Both Markers of Maturation and Aging in Pancreatic β-Cells

Abstract

Previously, we showed that thyroid hormone (TH) triiodothyronine (T₃) enhanced β-cell functional maturation through induction of Mafa High levels of T₃ have been linked to decreased life span in mammals and low levels to lengthened life span, suggesting a relationship between TH and aging. Here, we show that T₃ increased p16^Ink4a (a β-cell senescence marker and effector) mRNA in rodent and human β-cells. The kinetics of Mafa and p16^Ink4a induction suggested both genes as targets of TH via TH receptors (THRs) binding to specific response elements. Using specific agonists CO23 and GC1, we showed that p16^Ink4a expression was controlled by THRA and Mafa by THRB. Using chromatin immunoprecipitation and a transient transfection yielding biotinylated THRB1 or THRA isoforms to achieve specificity, we determined that THRA isoform bound to p16^Ink4a , whereas THRB1 bound to Mafa but not to p16^Ink4a On a cellular level, T₃ treatment accelerated cell senescence as shown by increased number of β-cells with acidic β-galactosidase activity. Our data show that T₃ can simultaneously induce both maturation (Mafa) and aging (p16^Ink4a ) effectors and that these dichotomous effects are mediated through different THR isoforms. These findings may be important for further improving stem cell differentiation protocols to produce functional β-cells for replacement therapies in diabetes.

Figure 1

T₃ induces p16^Ink4a expression in rodent islets/cell lines and hESCs differentiated toward islets. A: Mouse islets and MIN6 cells had increased p16^Ink4a mRNA after 4 days and 24–48 h of T₃ exposures, respectively. C57Bl/6J 16-week-old mice, n = 4; MIN6, n = 5 independent samples in duplicate. B: Rat islets and INS1 cells had T₃-induced p16^Ink4a mRNA expression after 4 days and 24–48 h culture in the presence of T₃. Adult rat islets, n = 3 independent samples; INS1, n = 4 independent samples. Differentiated hESCs exposed to T₃ during the last step of differentiation were analyzed by qPCR for expression of maturation and senescence markers. T₃ increased both MAFA mRNA (C) and senescence marker P16^INK4A mRNA (D). E: There was a linear correlation between the induction of MAFA and P16^INK4A in matched samples from hESCs. n = 5–11 samples from two to four individual experiments. Mean ± SEM. *P < 0.05 with respect to baseline. R and R² values are shown for linear correlations.

Figure 2

Kinetics of T₃ induction of Mafa and p16^Ink4a suggest that they are both targets of TH. INS1 were incubated in the presence of T₃ for different durations and different doses of T₃ at 24 h. A: Kinetics of T₃-induced Mafa induction. B: Kinetics of T₃-induced p16^Ink4a transcription. INS1 cells, n = 3–10 individual samples. Mean ± SEM. *P < 0.05 with respect to control.

Figure 3

Potential TREs were identified in the Cdkn2a gene and experimentally tested with ChIP. A: Potential TREs based on shown consensus sequence were identified in the Cdkn2a (p16^Ink4a) gene using AliBaba and TRANSFAC/MATCH. B: ChIP confirmed the binding of THR to site 5 in the Cdkn2a (p16^Ink4a) gene (MIN6 cells). *P < 0.03 with respect to IgG. n = 13 immunoprecipitation reactions from four individual experiments. bp, base pairs; RPII, RNA polymerase II.

Figure 4

T₃ induction of Mafa and p16^Ink4a is differentially mediated by specific Thr agonists in P7 rat islets. P7 rat islets were treated in vitro with 150 pmol/L of either THRA agonist CO23 or THRB agonist GC1 in RPMI 1640 (11.1 mmol/L glucose plus 10% CS-FBS). A: Mafa transcription was significantly increased by THRB agonist GC1. B: p16^Ink4a transcription was upregulated by THRA agonist CO23. n = 3 independent experiments. Mean ± SEM. *P < 0.05 with respect to control.

Figure 5

Biotinylated receptor isoforms show specificity of binding to Mafa and Cdkn2a TREs. Double transfection of MIN6 cells with BirA and either Blrp-TEV-THRA or Blrp-TEV-THRB1 plasmids, followed by streptavidin ChIP, tested specific binding of THRA and THRB1 to potential TREs in the Cdkn2a and Mafa genes. Site 5 in the Cdkn2a gene was enriched with THRA (A) but not with THRB1 (B), indicating the binding of the THRA isoform. Neither site 2 nor site 3 in in the Mafa gene was enriched with THRA (C) but site 2 was with THRB1 (D), indicating the specific binding of THRB1 to site 2 and not site 3. n = 6–8 samples from three to four independent experiments (MIN6 cells). Mean ± SEM. *P < 0.05 with respect to control.

Figure 6

Differential THR isoform induction of Mafa and p16^Ink4a shown by siRNA. siRNA knockdown of THR isoform transcription (40%) was specific (siThra [A] and siThrb [B].) C: Basal levels of p16Ink4a transcription were significantly decreased with siThra. D: Basal Mafa RNA levels were unchanged with either siThra or siThrb. E: β-Cell aging marker Igf1r levels were significantly decreased with both siThra and siThrb. Results in MIN6 cells are shown. n = 3 experiments, each in triplicate. Mean ± SEM. *P < 0.05 respect to scrambled siRNA (siSc).

Figure 7

T₃ induced cellular senescence in mouse islets and hESC treated with GC1 increased insulin secretion at 16.8 mmol/L glucose and absolute insulin content and decreased P16^INK4A levels. A: In vitro T₃ treatment of islets isolated from 5-month-old C57Bl/6J mice increased their senescence-associated β-gal activity, a marker of cellular senescence. Data presented for FACS sorted β-gal⁺ cells from individual experiments of islets pooled from 10 mice. THRB agonist GC1 substituted for T₃ in stage 5 of the differentiation protocol of hESC cells resulted in decreased P16^INK4A mRNA with no change of MAFA mRNA (B) and increased insulin secretion in response to 16.8 mmol/L glucose (C) and insulin content (D). Absolute values of insulin secretion per experiment (averaged two to three samples): 0.02, 0.04, and 0.06 pg Ins/ng DNA for controls and 0.04, 0.04, and 0.06 pg Ins/ng DNA for GC1-treated cells. Absolute insulin content: controls, average 5.4 pg Ins/ng DNA (1.02, 4.84, 10.70); GC1-treated, average 13.7 pg Ins/ng DNA (4, 15.47, 18.32). n = 8–9 samples, three independent experiments. *P < 0.003 in A; *P < 0.04 in B and D.