Pioglitazone,a PPARγ Agonist,Upregulates the Expression of Caveolin-1 and Catalase,Essential for Thyroid Cell Homeostasis: A Clue to the Pathogenesis of Hashimoto's Thyroiditis

Abstract

Background:

Peroxisome proliferator-activated receptor γ (PPARγ) is a transcription factor that regulates the expression of multiple target genes involved in several metabolic pathways as well as in inflammation. The expression and cell localization of caveolin-1 (Cav-1), thyroperoxidase (TPO), and dual oxidase (DUOX), involved in extracellular iodination, is modulated by Th1 cytokines in human normal thyroid cells and in Hashimoto's thyroiditis (HT).

Objectives:

The objectives of this study were (i) to analyze the PPARγ protein and mRNA expression at the follicular level in HT versus controls in correlation with the one of Cav-1; (ii) to study the effects of Th1 cytokines on PPARγ and catalase expression in human thyrocyte primary cultures; and (iii) to study the effects of pioglitazone, a PPARγ agonist, on thyroxisome components (Cav-1, TPO, DUOX) and on catalase, involved in antioxidant defense.

Results:

Although the global expression of PPARγ in the whole gland of patients with HT was not modified compared with controls, there was great heterogeneity among glands and among follicles within the same thyroid. Besides normal (type 1) follicles, there were around inflammatory zones, hyperactive (type 2) follicles with high PPARγ and Cav-1 expression, and inactive (type 3) follicles which were unable to form thyroxine and did not express PPARγ or Cav-1. In human thyrocytes in primary culture, Th1 cytokines decreased PPARγ and catalase expression; pioglitazone increased Cav-1, TPO, and catalase expression.

Conclusion:

PPARγ may play a central role in normal thyroid physiology by upregulating Cav-1, essential for the organization of the thyroxisome and extracellular iodination. By upregulating catalase, PPARγ may also contribute to cell homeostasis. The inhibitory effect of Th1 cytokines on PPARγ expression may be considered as a new pathogenetic mechanism for HT, and the use of PPARγ agonists could open a new therapeutic approach.

Introduction

Peroxisome proliferator-activated receptors (PPAR) are ligand-activated nuclear transcription factors (1,2). PPARγ is present in various cell types (3) and can also be found in thyroid cells (4). The natural ligands of PPARγ can be classified into two major groups of compounds: fatty acids and phospholipids, including 15d-PGI2 and derivates of mono- and polyunsaturated fatty acids. Most of these ligands bind PPARγ with low affinity. The pharmacological agonist consists mainly of the thiazolidinedione family members: pioglitazone, rosiglitazone, ciglitazone, and troglitazone. They are synthetic analogues of 15d-PGI2 but with a very high binding affinity for PPARγ (low nanomolar range for pioglitazone) (1). PPARγ exerts its action by binding as a heterodimer with the retinoid X receptor (RXR) to PPAR response element (PPRE) to regulate gene expression (1).

PPARγ agonists are well-known modulators of the inflammatory process (3) and of the interferon (IFN)-induced chemokine secretion in normal thyroid cells and in Graves' disease (GD) (5,6). However, the Th1 and Th2 cytokine effects on PPARγ expression have not been studied in the thyroid gland. Of note, it has been shown that Th1 cytokines downregulate PPARγ, for example in adipose tissue (7,8) and in human biliary cells (9), whereas Th2 cytokines upregulate PPARγ in glial cells (10) and in human biliary cells (9).

The thyroid gland is a good model to study the effects of Th1 and Th2 cytokines. Indeed, both cellular and humoral immunity have a role in the pathogenesis of autoimmune thyroid diseases. Hashimoto's thyroiditis (HT) associated with hypothyroidism is mainly due to a cytotoxic immune reaction against thyroperoxidase (TPO) and thyroglobulin (Tg). GD is associated with hyperthyroidism due to the activation of the thyrotropin (TSH) receptor by specific stimulating autoantibodies (11). GD is considered as a Th2 dominant disease, although Th1 cytokines may prevail at the onset of its development (6).

Thyroid hormone synthesis and the maintenance of euthyroidism is a very controlled cell process (for a review, see Colin et al. (12)). It requires the transfer of iodide (I^–), circulating in the blood, into the colloid by different transporters. Dual oxidase (DUOX) generates H₂O₂, which is required at the apical border to oxidize iodide and hence to form monoiodotyrosine (MIT) and diiodotyrosine (DIT) incorporated in Tg. The combination of MIT and DIT leads to the production of triiodothyronine (T3) and thyroxine (T4), released in the circulation after endocytosis and hydrolysis of the iodinated Tg. DUOX and TPO are two essential proteins for thyroid hormone biosynthesis, which gather at the apical cellular membrane forming the thyroxisome, a complex entirely dedicated to safe and efficient thyroid hormone synthesis. Caveolin-1 (Cav-1), by creating raft-like membrane structures, is likely a required element to form the thyroxisome, as first proposed by Song et al. (13). As shown in the Cav-1 knockout mouse model, Cav-1 deficiency is associated with the cytoplasmic localization of DUOX and TPO, with the production of excessive intracytoplasmic H₂O₂, and thus with increased oxidative stress (14).

Cytokines have been shown to have an effect on the different partners of the thyroxisome. In HT, the immunological environment is mostly composed by Th1 cytokines, among which IFN-γ, interleukin (IL)-1α and β, IL-2, and tumor necrosis factor (TNF-α, TNF-β) are predominant. Th1 cytokines appear to play a key role for the inhibition of thyrocyte function in HT. Indeed, sodium–iodide symporter (NIS), TPO, and DUOX are downregulated by Th1 cytokines (15,16). Cav-1 is also dramatically downregulated in HT, resulting in a cytoplasmic mislocalization of TPO and DUOX, and intracellular iodination. It leads to pronounced oxidative stress with a high apoptosis rate (17), as also previously shown in Cav-1 knockout mice (14), and in Pendred's syndrome (18). In HT, this downregulation of Cav-1 is due, at least partly, to Th1-cytokines (IL-1α and IFN-γ) (17). The resulting oxidative stress is insufficiently compensated by the cellular antioxidant defenses and thus leads to cell apoptosis, follicular destruction, and hypothyroidism (17).

The pathways implicated in Cav-1 regulation in normal thyroid cells remain unknown. However, it has been shown that PPARγ upregulates Cav-1 gene expression in different models, for example human macrophages and mouse hepatocytes (19,20).

The aims of this study were first to analyze the PPARγ expression in HT and to correlate it with that of Cav-1 and the functional status of the follicles in term of T4 synthesis. Second, the effects of Th1 cytokines on PPARγ and catalase expression in human primary thyrocytes were analyzed. Third, the effects of pioglitazone, a PPARγ agonist, were investigated in human primary thyroid cells on thyroxisome components (Cav-1, TPO, and DUOX) and on antioxidant defense (catalase) of the thyrocytes. The understanding of the intracellular mechanism leading to thyrocyte destruction in HT and the demonstration of the beneficial impact of PPARγ agonists could open new therapeutic approaches.

Materials and Methods

Thyroid samples

Tissue samples from seven HT patients were obtained through the Biolibrary of the Université Catholique de Louvain (Brussels, Belgium). The patients underwent surgery because of suspicious nodules detected by echography, but they were finally diagnosed as HT. Thyroid samples removed in the paranodular zone of multinodular goiters, considered as normal (n = 7), were used as control samples. The surgical samples were obtained with the informed consent of patients.

The clinical data of the HT and control patients are shown in Supplementary Table S1 (Supplementary Data are available online at www.liebertpub.com/thy).

Preparation of tissue samples for light microscopy and immunohistochemistry

The thyroids were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 24 hours and embedded in paraffin. Immunostaining was performed on 5-μm-thick paraffin sections. Sections were dewaxed and rehydrated, and endogenous peroxidases were quenched with 1% H₂O₂ for 15 min. For the antigen retrieval, tissue sections were pretreated in a microwave oven in citrate buffer (pH 6.6) for one 3 min cycle at 750 W, followed by four cycles of 3 min 30 sec, each at 350 W. All sections were washed with PBS supplemented with 1% bovine serum albumin (PBS-BSA) and thereafter incubated in 5% PBS-BSA containing 2% non-immune goat serum at room temperature (RT) for 1 h. Sections were incubated at RT for 24 h with a PPARγ polyclonal rabbit antibody (1:250; SAB4502262; Sigma, Saint-Louis, MO) for 3 h with a Cav-1 polyclonal rabbit antibody (1:400; BD610059; BD Transduction Laboratory, Franklin Lakes, NJ) for 3 h with a catalase monoclonal mouse antibody (1:200; C-0979; Sigma), or overnight with a T4 monoclonal mouse antibody (1:400; MBS592059; MyBioSource, San Diego, CA).

The binding of antibodies was detected using a secondary antibody conjugated to a peroxidase labeled polymer for 1 h at RT (1:50; 32260; Pierce^® goat anti-rabbit; ThermoScientific, Rockford, IL; or 1:50; 32230; Pierce^® goat anti-mouse; ThermoScientific). The peroxidase activity was revealed with 3-3′ diaminobenzidine tetrahydrochloride (DAB; Sigma). Sections were counterstained with Mayer's hematoxylin, and rinsed and mounted in Faramount aqueous mounting medium (DakoCytomation, Carpinteria, CA). To verify the binding specificity, some sections were incubated by omitting the primary antibody.

The quantification of the immunolabeling was made by ImageJ software developed by the National Institute of Health (NIH; Bethesda, MD).

Primary cell cultures

Thyroid samples (n = 5) were obtained after surgery of multinodular goiters (Cliniques Universitaires St. Luc, Brussels, Belgium). The samples were removed in the paranodular zone, considered as normal. These tissues were isolated as previously described (21) and suspended in modified Earle's medium (BRL-Gibco, Paisley, United Kingdom), containing 5% newborn calf serum (Gibco), penicillin (100 IU/mL; Gibco), streptomycin (100 IU/mL; Gibco), glutamine (2.4 mM; Sigma), and fungizone (2.5 μg/mL; Gibco). The cells were plated in 12-well plates and cultured in a humidified atmosphere (37°C, 5% CO₂) with 1 mIU/mL of TSH (Sigma) for one week.

Cytokine effects

Cells were then incubated or not for three additional days, with 2 ng/mL of recombinant murine IL-1α and 10 ng/mL of recombinant murine IFN-γ (RD systems, Abingdon, United Kingdom) in modified Earle's medium containing 0.5% newborn calf serum, the Th1 cytokines inducing a downregulatory effect on TPO, DUOX (16), and Cav-1 (17) at these doses.

Pioglitazone effects

Cells were incubated or not for 24 h with 15 μM of pioglitazone (Sigma) in modified Earle's medium containing 0.5% newborn calf serum. The pioglitazone concentration was chosen according to previous studies by Antonelli et al. (5,6,22).

RNA purification and quantitative reverse transcription polymerase chain reaction

mRNA levels of PPARγ, cav-1, TPO, DUOX1, DUOX2, and catalase were measured by quantitative reverse transcription polymerase chain reaction (RT-PCR). These experiments were performed, as previously described (17), using the iQ5 (BioRad, Hemel Hempstead, United Kingdom) as follows: 30 cycles at 95°C for 1 min, followed by 40 cycles at 95°C for 15 sec, the appropriate annealing temperature (see Supplementary Table S2), for 45 sec, and at 81°C for 15 sec. The amplified level of the target gene was normalized against that of β-actin. Amplification of genomic DNA was avoided by using primers spanning two adjacent exons.

Western blotting

Cells were homogenized in Laemmli buffer containing a protease inhibitor cocktail (Roche, Basel, Switzerland). Western blotting and its quantification were performed, as previously described (17). All the characteristics of primary antibodies are summarized in Supplementary Table S3. All values are expressed as mean ± standard error of the mean.

Cav-1 and PPARγ immunofluorescence

Thyrocytes were cultured on cover glasses in appropriate medium. They were fixed for 10 min in 4% paraformaldehyde, rinsed with PBS, and blocked for 1 h in a PBS saponin 0.1%/BSA 5% solution at 37°C. Cells were then incubated overnight (ON) at 4°C with a Cav-1 monoclonal mouse primary antibody (1:100; SAB4200216; Sigma) diluted in PBS saponin 0.1%/BSA 1%. The co-staining was realized by adding a mix of the Cav-1 monoclonal mouse primary antibody (1:100; SAB4200216; Sigma) and the PPARγ polyclonal rabbit antibody (1:250; SAB4502262; Sigma) diluted in PBS saponin 0.1%/1% BSA. After being washed in PBS saponin 0.1%, the cells were incubated for 1 h at 37°C with an Alexa fluor (488) secondary antibody (1:300; A11029; Invitrogen, Genk, Belgium). For the co-staining, a mix was introduced containing the same Alexa fluor 488 secondary antibody (1:300; A11029; Invitrogen) and the Alexa fluor 568 secondary antibody (1:300; A11036; Invitrogen), which was then incubated for 5 min with Hoechst at a dilution of 1:5000. Cover slides were mounted in fluorescent mounting medium (DakoCytomation) for microscopic observation.

A Axio Zeiss microscope (Zeiss, Zaventem, Belgium) was used, and the quantification of the fluorescence was made by ImageJ software developed by the NIH.

Statistical analysis

RT-PCR and Western blotting results were normalized against levels of β-actin. Statistical analysis was performed using a t-test or a Mann–Whitney test when results did not pass the normality test (GraphPad, San Diego, CA). A p-value of <0.05 was considered statistically significant.

Results

Immunohistochemical detection reveals a strong follicular heterogeneity of T4 synthesis in HT

In control thyroids, the majority of the follicles contained T4 homogeneously distributed in the colloid. They were large with a flat or cuboidal epithelial layer (Fig. 1A). These normal follicles were classified as type 1.

FIG. 1.

Immunohistochemical detection of thyroxine (T4), peroxisome proliferator-activated receptor γ (PPARγ), caveolin-1 (Cav-1), and catalase in a normal thyroid gland and Hashimoto's thyroiditis (HT). In control thyroids (A), the follicles contain T4 homogeneously within the colloid. The epithelial layer is composed by flat or cuboidal cells. These follicles are referred to as a type 1 follicles. HT thyroid tissue (B) and (C) contains three types of follicles: type 1 follicles as in normal thyroid tissue, and type 2 and type 3 follicles near the inflammatory zones. Type 2 follicles are small and contain T4 in the colloid, whereas type 3 follicles are also small but without T4 in the colloid. Scale bar = 100 μm. Around the inflammatory zones of Hashimoto's glands, the type 2 follicles (2), expressing T4 (D), show a very high nuclear expression of PPARγ (E). The same follicles also show a strong expression of Cav-1 (F) and catalase (G). The expression of PPARγ, Cav-1, and catalase is low or absent in type 3 follicles (3) having lost the T4 synthesis capacity. Scale bar = 100 μm.

Hashimoto's thyroids were modified by inflammatory infiltrates, which, in some samples, invaded very large parts of the glandular parenchyma. Zones of type 1 follicles were still observed, their surfaces being variable among HT patients. At the periphery of inflammatory zones, two other types of follicles were observed: type 2 follicles, which were usually small and contained T4 in their lumina, and type 3 follicles, which were small but without T4 (Fig. 1B). In type 3 follicles located inside the inflammatory infiltrates, T4 was also observed in the cytoplasm rather than in the colloid, suggesting intracellular iodination (Fig. 1C).

Immunohistochemical detection reveals a strong follicular heterogeneity regarding PPARγ, Cav-1, and catalase expression

By analyzing the mRNA and protein PPARγ expression by RT-PCR and Western blotting, no significant differences were observed between HT samples and control glands (Supplementary Fig. S1A and B). The same applied to the quantification of the immunohistochemical labeling of PPARγ, which did not show any differences between HT and controls (Supplementary Fig. S1C). However, these results were obtained on whole samples, which, in HT, are very heterogeneous regarding the morphological and functional phenotype of the follicles that were classified in three types on histological sections.

Surrounding the inflammatory zones of Hashimoto's glands, the type 2 follicles expressing T4 (Fig. 1D) showed very high nuclear expression of PPARγ (Fig. 1E). The same follicles also showed strong expression of Cav-1 (Fig. 1F) and catalase (Fig. 1G). The expression of PPARγ, Cav-1, and catalase was low or absent in type 3 follicles, having lost the ability for T4 synthesis.

Correlation between the expression of PPARγ and Cav-1 in the three types of follicles in HT

In normal type 1 follicles, T4 was detected in the entire colloid (Fig. 2A). PPARγ was detected in a few nuclei (Fig. 2B), and Cav-1 was located at the apical pole of the thyrocytes where normal iodination takes place (Fig. 2C).

FIG. 2.

Immunohistochemical detection of T4, PPARγ, and Cav-1 in HT. In normal type 1 follicles (1), T4 is detectable in the entire colloid (A). PPARγ is present in a few nuclei (B), and Cav-1 is located at the apical pole of thyrocytes where normal iodination takes place (C). Compared with type 1 follicles, type 2 follicles (2) have a smaller size and a higher epithelial layer, and T4 can be detected in the colloid (D). The expression of PPARγ is highly increased and present in nearly all the cell nuclei (E). Cav-1 is also highly expressed in the same follicles (F). Type 3 follicles (3) are hypofunctional. They do not contain T4 in their lumina (G) and have lower expression of PPARγ (H) and Cav-1 (I) compared with type 2 follicles. The expression of PPARγ (K) and Cav-1 (L) is absent in type 3 follicles, showing T4 in the cytoplasm rather than in the colloid (J). Although Cav-1 is not detected in the thyrocytes, it is highly expressed in endothelial cells of the numerous vessels in the inflammatory infiltrates. Scale bar = 10 μm.

Type 2 follicles, which were smaller and had a higher epithelial layer, did actively synthesize T4, which was detected in the colloid (Fig. 2D). The expression of PPARγ was highly increased and present in nearly all of the cell nuclei (Fig. 2E). Cav-1 was also highly expressed in these follicles, as shown in serial sections (Fig. 2F).

Type 3 follicles were hypofunctional. They did not contain T4 in their lumina (Fig. 2G) and had lower expression of PPARγ (Fig. 2H) and Cav-1 (Fig. 2I) compared with type 2 follicles. The expression of PPARγ (Fig. 2K) and Cav-1 (Fig. 2L) was absent in type 3 follicles, showing T4 synthesis in the cytoplasm rather than in the colloid (Fig. 2J).

PPARγ and catalase mRNA expression is decreased by IL-1α and IFN-γ (Th1 cytokines) in primary cultures of human thyrocytes

A downregulatory effect of Th1 cytokines on Cav-1 expression in human primary thyroid cells has been recently published (17). Consistent with this observation, treating cells with IL-1α and IFN-γ, the predominant Th1 cytokines in HT, strongly decreased PPARγ mRNA expression (p < 0.01 compared with controls; Fig. 3A).

FIG. 3.

Effects of Th1 cytokines on PPARγ and catalase mRNA expression in primary cultures of human thyrocytes. (A) The expression of PPARγ mRNA relative to that of β-actin signal is expressed as mean ± standard error of the mean (SEM) of two experiments (three and five independent wells for each condition). Interleukin (IL)-1α (2 ng/mL) combined with interferon (IFN)-γ (100 IU/mL) decreases PPARγ mRNA. **p < 0.01 vs. controls. (B) The expression of catalase mRNA relative to that of β-actin signal is expressed as mean ± SEM in two experiments (three and five independent wells for each condition). IL-1α (2 ng/mL) combined with IFN-γ (100 IU/mL) decreases catalase mRNA. *p < 0.05 vs. controls.

Treatment with IL-1α and IFN-γ significantly decreased catalase mRNA expression (p < 0.05 compared with controls; Fig. 3B).

Pioglitazone, a PPARγ agonist, upregulates mRNA and protein expression of Cav-1 and TPO in primary cultures of human thyrocytes, and increases DUOX1 and DUOX2 mRNA expression, without modifying DUOX protein expression

Treatment of human thyrocytes in primary cultures with 15 μM of pioglitazone significantly increased PPARγ mRNA expression (p < 0.001 compared with controls; Fig. 4A). Similar results were obtained with 10 μM of pioglitazone (data not shown). Treatment of thyrocytes with pioglitazone also increased Cav-1 mRNA and protein expression (p < 0.05 compared with controls; Fig. 4B and C). The increase of Cav-1 protein expression was confirmed by quantification of immunofluorescence (Fig. 4D). In comparison with controls (Fig. 4E), Cav-1 expression was higher after treatment with pioglitazone (Fig. 4F). The co-staining of Cav-1 and PPARγ showed that these proteins were co-expressed and increased in the same cells after pioglitazone treatment (Fig. 4H) compared with controls (Fig. 4G). PPARγ protein expression was nuclear (Fig. 4H).

FIG. 4.

Effect of pioglitazone on PPARγ mRNA expression and Cav-1 mRNA and protein expression in primary cultures of human thyrocytes. (A) The expression of PPARγ mRNA relative to that of β-actin is expressed as mean ± SEM of three representative experiments (five independent wells for each condition). Treatment with 15 μM pioglitazone increases PPARγ mRNA expression. **p < 0.01 vs. controls. The expression of Cav-1 mRNA (B) and protein (C) relative to that of β-actin is expressed as mean ± SEM of three quantitative reverse transcription real-time polymerase chain reaction (RT-qPCR) experiments (five independent wells for each condition in each experiment) and four Western blotting experiments (three independent wells for each condition in each experiment). A representative blot is shown, and lines originating from two different parts of the same gel were aligned. Cav-1 mRNA and protein expression is significantly increased in cells treated with 15 μM of pioglitazone compared with controls. *p < 0.05 vs. controls. Cav-1 protein expression was detected by immunofluorescence, and the fluorescence was quantified on multiple representative parts of the cell culture by ImageJ software and then expressed as mean ± SEM of two experiments (D). Treatment with 15 μM of pioglitazone increases the expression of Cav-1 protein expression (F) compared with controls (E). Nuclei were stained by Hoechst. **p < 0.01 vs. controls. Scale bars = 50 μm. In the presence of pioglitazone (H), both PPARγ and Cav-1 expression are increased compared with controls (G). Cav-1 was detected in the cytoplasm and PPARγ in the nucleus, within the same cells. Nuclei were stained by Hoechst. Scale bars = 50 μm.

mRNA (Fig. 5A) and protein (Fig. 5B) expression of TPO was significantly increased in human thyrocytes incubated with pioglitazone compared with controls (p < 0.01 and <0.05, respectively).

FIG. 5.

Effects of pioglitazone on TPO, DUOX 1–2 mRNA, and protein expression in primary cultures of human thyrocytes. The expression of TPO mRNA (A) and protein (B) relative to that of β-actin is expressed as mean ± SEM of three RT-qPCR experiments (five independent wells for each condition in each experiment) or three Western blotting experiments (three independent wells for each condition in each experiment). A representative blot is shown, and lines originating from two different parts of the same gel were aligned. TPO mRNA and protein expression is significantly increased in cells treated with 15 μM of pioglitazone compared with controls. *p < 0.05 vs. controls; **p < 0.01 vs. controls. The relative expression of DUOX1 mRNA (C), DUOX2 mRNA (D), and DUOX protein (E) relative to that of β-actin is expressed as means ± SEM in three RT-qPCR experiments (five independent wells for each condition in each experiment) or four Western blotting experiments (three independent wells for each condition in each experiment). A representative blot is shown, and lines originating from two different parts of the same gel were aligned. DUOX1 mRNA expression is significantly increased in cells treated with 15 μM of pioglitazone compared with controls, as well as DUOX2 mRNA. The DUOX protein expression is not significantly increased in the treated cells. *p < 0.05 vs. controls; **p < 0.01 vs. controls.

In addition, treatment with pioglitazone also significantly increased DUOX1 (Fig. 5C) as well as DUOX2 mRNA expression (Fig. 5D; p < 0.05 and <0.01 compared with controls, respectively). The total DUOX protein showed a non-significant tendency to increase upon PPARγ agonist treatment (Fig. 5E).

Pioglitazone, a PPARγ agonist, upregulates mRNA and protein expression of catalase in primary cultures of human thyrocytes

mRNA (Fig. 6A) and protein (Fig. 6B) expression of catalase was significantly increased in human primary thyrocytes incubated with pioglitazone compared with controls (p < 0.05).

FIG. 6.

Effects of pioglitazone on catalase mRNA and protein expression in primary cultures of human thyrocytes. The expression of catalase mRNA (A) and protein (B) relative to that of β-actin is expressed as mean ± SEM in three RT-qPCR experiments (five independent wells for each condition in each experiment) or two Western blotting experiments (three independent wells for each condition in each experiment). A representative blot is shown. Catalase mRNA and protein expression is significantly increased in cells treated with 15 μM of pioglitazone compared with controls. *p < 0.05 vs. controls.

Discussion

Pioglitazone and rosiglitazone—synthetic PPARγ ligands—have been shown to have a beneficial effect on the development of Th1-autoimmune thyroiditis by inhibiting the IFN-induced secretion of chemokines by thyrocytes (6,23). This study also demonstrate a beneficial effect of pioglitazone on Cav-1 expression and thus on the integrity of the thyroxisome, which is essential for thyroid hormone synthesis. It is indeed the first time that pioglitazone is shown to upregulate Cav-1 mRNA and protein expression in primary cultures of human thyroid cells, as already demonstrated in human macrophages (19) and mouse hepatocytes (20).

In the thyroid cell, Cav-1 has an important functional role. In Cav-1 knockout mice (14), the loss of Cav-1 leads to the mislocalization of TPO and DUOX, to intracellular iodination and intracellular H₂O₂ production associated to an oxidative stress, cell apoptosis, and finally hypothyroidism. The association between the loss of Cav-1 and the development of hypothyroidism has also been demonstrated in Pendred's syndrome (18) and in HT (17). The present study also shows that pioglitazone upregulates the mRNA and protein expression of Cav-1 and TPO, as well as DUOX mRNA expression. This further supports that Cav-1, TPO, and DUOX are important elements of the thyroxisome, which is key for maintaining thyroid cell homeostasis and euthyroidism. By increasing Cav-1 expression, pioglitazone may thus be facilitating normal thyroid hormone synthesis.

Disruption of the thyroxisome is a key event in the development of HT and is induced by Th1 cytokines (IFN-γ and IL-1α), which downregulate Cav-1, TPO, and DUOX (17). It has been previously observed that Th1 cytokines inhibit NIS mRNA expression (15,24) and thyroid hormone release (25). IL-1 and IFN-γ also downregulate TPO, DUOX, and Tg expression in human thyrocytes (16,26,27).

Moreover, this in vitro study demonstrates that Th1 cytokines downregulate PPARγ expression in human thyroid cells. A similar effect has already been observed in adipose tissue (7), but it is the first time that it has been reported in the thyroid gland. In HT, the suppressive effect of IFN-γ and IL-1α on PPARγ could be mediated by early growth response gene 1 (Egr-1), at least in part, as demonstrated in osteoarthritis chondrocytes (28). It could also be mediated by an excessive production of reactive oxygen species (ROS) by the inflammatory infiltrate surrounding the HT follicles and inside the thyrocytes. In addition to reactive nitrogen species (RNS), such as nitric oxide (NO) (29), ROS are also thought to act as intracellular mediators of Th1 cytokine-induced effects in thyroid cells (26).

There are multiple sources of ROS in thyroid cells, including free radicals, O₂ ^ċ–, OH^ċ–, peroxyl (RO₂ ^ċċ), alkoxyl (RO^ċ), and several non-radical molecules such as H₂O₂. H₂O₂, essential for thyroid hormone synthesis, is part of the physiological oxidative load. Although Th1 cytokines downregulate DUOX expression and therefore the production of extracellular H₂O₂ (16,17,26,27), the Th1-induced disruption of the thyroxisome leads to cytoplasmic localization of DUOX and excessive H₂O₂ production in the cytoplasm (17). The superoxide radical (O2^ċ–) also plays an important role in the modulation of thyroid cell function, as shown by an increase of NOX2 mRNA and protein expression in response to the Th1 cytokines, in vitro in human thyroid cells and in vivo in HT (30).

To face the overproduction of RNS/ROS, the thyroid cells have strong antioxidant defenses of various types. Peroxiredoxins and catalase allow intracellular H₂O₂ to be detoxified: their expression is increased in HT, but is insufficient to compensate for the oxidative stress (17). The present data also show a decrease of catalase mRNA expression in the thyroid cells exposed to Th1 cytokines. Moreover, it is shown that pioglitazone increases catalase mRNA expression allowing the overproduction of intracellular H₂O₂ to be prevented and avoiding oxidative stress and cell apoptosis. The regulation of catalase by PPARγ has previously been demonstrated in human adipocytes (31), and it has also been shown that PPARγ activation is cardioprotective to H₂O₂-induced stress in the newborn heart by the upregulation of catalase (32). PPARγ ligands have a protective effect on renal ischemia/reperfusion injury by reducing iNOS (33). They also suppress NADPH oxidase (34) and restore superoxide dismutase in endothelial cells (35).

The extrapolation of these in vitro results, showing a downregulatory effect by Th1 cytokines on PPARγ expression, to the in vivo situation in HT is at first not obvious. Indeed, global quantification of PPARγ mRNA and protein expression showed no significant decrease in whole HT glands compared to controls. This could be explained by the great heterogeneity among Hashimoto's patients regarding inflammatory extent and by the fact that the inflammatory cells did express PPARγ (36). Moreover, within a single gland, the follicles are very heterogeneous from both a morphological and a functional point of view. This is already the case in normal thyroid tissue where active follicles and hypofunctionning follicles coexist (37). This follicular heterogeneity was even more accentuated in HT (17), and was very well demonstrated in the present study when considering PPARγ expression. Indeed, at the periphery of the inflammatory infiltrates, hyperactive type 2 follicles and inactive type 3 follicles were observed. In type 2 follicles, PPARγ expression was very high, and this could indicate a protective mechanism to face the Th1 cytokines produced by the inflammatory cells. In the same way, PPARγ could increase the expression of the thyroxisome components, mainly Cav-1, to maintain a normal apical iodination. PPARγ could also increase the antioxidant defenses of the follicular cells as shown by the strong catalase expression in the type 2 active follicles. At a higher degree of inflammation, Th1 cytokines completely impair the PPARγ expression as well as the one of Cav-1 and catalase, leading to a loss of normal hormonal synthesis. In some follicles, there are also signs of intracellular iodination, associated with the absence of PPARγ and Cav-1. The proportion of these inactive type 3 follicles will determine the level of hypothyroidism.

In conclusion, PPARγ is decreased by Th1 cytokines as shown by the present in vitro and in vivo results. This downregulation leads to a decreased expression of Cav-1, a key element of the thyroxisome, and alters the adequate targeting of TPO and DUOX at the apical membrane required to perform extracellular iodination. This opens up a new perspective in the physiopathology of HT, which might be due to a decrease of PPARγ by Th1 cytokines.

Activation of PPARγ by pioglitazone upregulates Cav-1 expression. This allows proper localization of TPO and DUOX at the apical membrane of the thyrocytes and efficient T4 formation in the colloid. Pioglitazone also upregulates catalase mRNA expression, allowing efficient protection against oxidative stress, which could be due to intracellular H₂O₂ accumulation in absence of Cav-1 (Fig. 7).

FIG. 7.

Schematic representation of the potential role of PPARγ in normal thyroid physiology and in the Th1-induced HT. PPARγ is thought to play a central role in normal thyroid physiology by upregulating Cav-1, essential for thyroxisome organization and extracellular iodination. By upregulating catalase expression, PPARγ also maintains cell homeostasis and prevents oxidative stress. The inhibitory effect of Th1 cytokines on PPARγ expression may be considered as a new pathogenic mechanism for HT. A downregulation of PPARγ leads to a disruption of the thyroxisome and a downregulation of catalase. The disruption of the thyroxisome is associated with intracellular iodination and an increased oxidative stress in the thyrocytes. If the thyrocytes cannot counteract the increased oxidative stress because of a decreased antioxidant defense, this can result in cell damage.

Footnotes

Acknowledgments

The authors wish to thank Marc de Bournonville for his technical help and Professors F. Miot and X. De Deken (IRIBHM, ULB, Brussels) for their collaboration and gifts of DUOX and TPO antibodies.

Author Disclosure Statement

The authors have nothing to disclose.

References

Hamblin

, Chang

, Fan

, Zhang

, Chen

. 2009. PPARs and the cardiovascular system. Antioxid Redox Signal, 11:1415–1452.

Rogue

, Spire

, Brun

, Claude

, Guillouzo

. 2010. Gene expression changes induced by PPAR gamma agonists in animal and human liver. PPAR Res, 2010:325183.

Mandard

, Patsouris

. 2013. Nuclear control of the inflammatory response in mammals by peroxisome proliferator-activated receptors. PPAR Res, 2013:613864.

Poncin

, Gerard

, Boucquey

, Senou

, Calderon

, Knoops

, Lengele

, Many

, Colin

. 2008. Oxidative stress in the thyroid gland: from harmlessness to hazard depending on the iodine content. Endocrinology, 149:424–433.

Antonelli

, Ferrari

, Frascerra

, Corrado

, Pupilli

, Bernini

, Benvenga

, Ferrannini

, Fallahi

. 2011. Peroxisome proliferator-activated receptor alpha agonists modulate Th1 and Th2 chemokine secretion in normal thyrocytes and Graves' disease. Exp Cell Res, 317:1527–1533.

Antonelli

, Ferrari

, Mancusi

, Mazzi

, Pupilli

, Centanni

, Ferri

, Ferrannini

, Fallahi

. 2013. Interferon-alpha, -beta and -gamma induce CXCL11 secretion in human thyrocytes: modulation by peroxisome proliferator-activated receptor gamma agonists. Immunobiology, 218:690–695.

Mracek

, Cannon

, Houstek

. 2004. IL-1 and LPS but not IL-6 inhibit differentiation and downregulate PPAR gamma in brown adipocytes. Cytokine, 26:9–15.

Waite

, Floyd

, Arbour-Reily

, Stephens

. 2001. Interferon-gamma-induced regulation of peroxisome proliferator-activated receptor gamma and STATs in adipocytes. J Biol Chem, 276:7062–7068.

Harada

, Isse

, Kamihira

, Shimoda

, Nakanuma

. 2005. Th1 cytokine-induced downregulation of PPARgamma in human biliary cells relates to cholangitis in primary biliary cirrhosis. Hepatology, 41:1329–1338.

10.

Paintlia

, Paintlia

, Singh

. 2006. IL-4-induced peroxisome proliferator-activated receptor gamma activation inhibits NF-kappaB trans activation in central nervous system (CNS) glial cells and protects oligodendrocyte progenitors under neuroinflammatory disease conditions: implication for CNS-demyelinating diseases. J Immunol, 176:4385–4398.

11.

Michels

, Eisenbarth

. 2010. Immunologic endocrine disorders. J Allergy Clin Immunol, 125:S226–237.

12.

Colin

, Denef

, Lengele

, Many

, Gerard

. 2013. Recent insights into the cell biology of thyroid angiofollicular units. Endocr Rev, 34:209–238.

13.

Song

, Driessens

, Costa

, De Deken

, Detours

, Corvilain

, Maenhaut

, Miot

, Van Sande

, Many

, Dumont

. 2007. Roles of hydrogen peroxide in thyroid physiology and disease. J Clin Endocrinol Metab, 92:3764–3773.

14.

Senou

, Costa

, Massart

, Thimmesch

, Khalifa

, Poncin

, Boucquey

, Gerard

, Audinot

, Dessy

, Ruf

, Feron

, Devuyst

, Guiot

, Dumont

, Van Sande

, Many

. 2009. Role of caveolin-1 in thyroid phenotype, cell homeostasis, and hormone synthesis: in vivo study of caveolin-1 knockout mice. Am J Physiol Endocrinol Metab, 297:E438–451.

15.

Ajjan

, Watson

, Findlay

, Metcalfe

, Crisp

, Ludgate

, Weetman

. 1998. The sodium iodide symporter gene and its regulation by cytokines found in autoimmunity. J Endocrinol, 158:351–358.

16.

Gerard

, Boucquey

, van den Hove

, Colin

. 2006. Expression of TPO and ThOXs in human thyrocytes is downregulated by IL-1alpha/IFN-gamma, an effect partially mediated by nitric oxide. Am J Physiol Endocrinol Metab, 291:E242–253.

17.

Marique

, Van Regemorter

, Gerard

, Craps

, Senou

, Marbaix

, Rahier

, Daumerie

, Mourad

, Lengele

, Colin

, Many

. 2014. The expression of dual oxidase, thyroid peroxidase, and caveolin-1 differs according to the type of immune response (TH1/TH2) involved in thyroid autoimmune disorders. J Clin Endocrinol Metab, 99:1722–1732.

18.

Senou

, Khalifa

, Thimmesch

, Jouret

, Devuyst

, Col

, Audinot

, Lipnik

, Moreno

, Van Sande

, Dumont

, Many

, Colin

, Gerard

. 2010. A coherent organization of differentiation proteins is required to maintain an appropriate thyroid function in the Pendred thyroid. J Clin Endocrinol Metab, 95:4021–4030.

19.

Llaverias

, Vazquez-Carrera

, Sanchez

, Noe

, Ciudad

, Laguna

, Alegret

. 2004. Rosiglitazone upregulates caveolin-1 expression in THP-1 cells through a PPAR-dependent mechanism. J Lipid Res, 45:2015–2024.

20.

, Matsusue

, Kashireddy

, Cao

, Yeldandi

, Rao

, Gonzalez

, Reddy

. 2003. Adipocyte-specific gene expression and adipogenic steatosis in the mouse liver due to peroxisome proliferator-activated receptor gamma1 (PPARgamma1) overexpression. J Biol Chem, 278:498–505.

21.

Nilsson

, Husmark

, Nilsson

, Tisell

, Ericson

. 1996. Primary culture of human thyrocytes in Transwell bicameral chamber: thyrotropin promotes polarization and epithelial barrier function. Eur J Endocrinol, 135:469–480.

22.

Antonelli

, Ferrari

, Fallahi

, Berti

, Materazzi

, Minuto

, Giannini

, Marchetti

, Barani

, Basolo

, Ferrannini

, Miccoli

. 2009. Thiazolidinediones and antiblastics in primary human anaplastic thyroid cancer cells. Clin Endocrinol, 70:946–953.

23.

Fallahi

, Ferrari

, Corrado

, Giuggioli

, Ferri

, Antonelli

. 2014. Targeting chemokine (C-X-C motif) receptor 3 in thyroid autoimmunity. Recent Pat Endocr Meta Immune Drug Discov, 8:95–101.

24.

Pang

, Hershman

, Smith

, Pekary

, Sugawara

. 1990. The mechanism of action of tumour necrosis factor-alpha and interleukin 1 on FRTL-5 rat thyroid cells. Acta Endocrinol, 123:203–210.

25.

Sato

, Satoh

, Shizume

, Ozawa

, Han

, Imamura

, Tsushima

, Demura

, Kanaji

, Ito

, et al. 1990. Inhibition of 125I organification and thyroid hormone release by interleukin-1, tumor necrosis factor-alpha, and interferon-gamma in human thyrocytes in suspension culture. J Clin Endocrinol Metab, 70:1735–1743.

26.

Poncin

, Colin

, Decallonne

, Clinckspooor

, Many

, Denef

, Gerard

. 2010. N-acetylcysteine and 15deoxy-Δ12,14-prostaglandin J2 exert a protective effect against autoimmune thyroid destruction in vivo but not against interleukin-1α/interferon γ-induced inhibitory effects in thyrocytes in vitro . Am J Pathol, 177:219–228.

27.

Poncin

, Lengele

, Colin

, Gerard

. 2008. Differential interactions between Th1/Th2, Th1/Th3, and Th2/Th3 cytokines in the regulation of thyroperoxidase and dual oxidase expression, and of thyroglobulin secretion in thyrocytes in vitro. Endocrinology, 149:1534–1542.

28.

Nebbaki

, El Mansouri

, Afif

, Kapoor

, Benderdour

, Duval

, Pelletier

, Martel-Pelletier

, Fahmi

. 2012. Egr-1 contributes to IL-1-mediated down-regulation of peroxisome proliferator-activated receptor gamma expression in human osteoarthritic chondrocytes. Arthritis Res Ther, 14:R69.

29.

van den Hove

, Stoenoiu

, Croizet

, Couvreur

, Courtoy

, Devuyst

, Colin

. 2002. Nitric oxide is involved in interleukin-1alpha-induced cytotoxicity in polarised human thyrocytes. J Endocrinol, 173:177–185.

30.

Colin

, Poncin

, Leveque

, Gallez

, Gerard

. 2014. Differential regulation of the production of reactive oxygen species in Th1 cytokine-treated thyroid cells. Thyroid, 24:441–452.

31.

Okuno

, Matsuda

, Miyata

, Fukuhara

, Komuro

, Shimabukuro

, Shimomura

. 2010. Human catalase gene is regulated by peroxisome proliferator activated receptor-gamma through a response element distinct from that of mouse. Endocrine J, 57:303–309.

32.

Chen

, Jin

, Crawford

, Cheng

, Saafir

, Wagner

, Yuan

, Ding

. 2012. Cardioprotection from oxidative stress in the newborn heart by activation of PPARgamma is mediated by catalase. Free Rad Biol Med, 53:208–215.

33.

Betz

, Schneider

, Kress

, Schick

, Wanner

, Sauvant

. 2012. Rosiglitazone affects nitric oxide synthases and improves renal outcome in a rat model of severe ischemia/reperfusion injury. PPAR Res, 2012:219319.

34.

Hwang

, Kleinhenz

, Lassegue

, Griendling

, Dikalov

, Hart

. 2005. Peroxisome proliferator-activated receptor-gamma ligands regulate endothelial membrane superoxide production. Am J Physiol Cell Physiol, 288:C899–905.

35.

Verreth

, De Keyzer

, Davey

, Geeraert

, Mertens

, Herregods

, Smith

, Desjardins

, Balligand

, Holvoet

. 2007. Rosuvastatin restores superoxide dismutase expression and inhibits accumulation of oxidized LDL in the aortic arch of obese dyslipidemic mice. Br J Pharmacol, 151:347–355.

36.

Dantas

, Pereira

, de Melo Rego

, da Rocha

Jr , Pitta Ida

, Marques

, Duarte

, Pitta

. 2015. The role of PPAR gamma in systemic sclerosis. PPAR Res, 2015:124624.

37.

Gerard

, Many

, Daumerie

, Costagliola

, Miot

, DeVijlder

, Colin

, Denef

. 2002. Structural changes in the angiofollicular units between active and hypofunctioning follicles align with differences in the epithelial expression of newly discovered proteins involved in iodine transport and organification. J Clin Endocrinol Metab, 87:1291–1299.