Recent studies have shown that the serum IL-23 concentrations were significantly increased in HT patients in comparison with both Grave’s disease (GD) patients and healthy subjects (30). Therefore, we focused our attention on the expression of IL-23 and its receptor in the thyroid gland of HT patients. The results showed that a low level of IL-23 was expressed in the healthy control TFCs by immunohistochemistry (IHC) examination (n = 5; Figure 1). In contrast, IL-23 expression was markedly upregulated in the HT patient TFCs (n = 10; Figure 1). Additionally, IL-23R expression in the HT patient TFCs exhibited a similar pattern to the IL-23 expression, and it was also significantly increased compared with the healthy controls (n = 5; Figure 1). Cumulative data of the IHC quantification from all samples are shown in Figure 1B. These results suggest that IL-23 expression was enhanced in TFCs under autoimmune inflammatory conditions, and it might be involved in the regional autoimmunity of HT.

To determine whether TFC-derived IL-23 was induced under the HT inflammatory conditions, we applied an immortalized human TFC cell line, Nthy-ori 3-1 cells, for further experiments. The mRNA expression of IL-23p19 and IL-23R in Nthy-ori 3-1 cells were examined with an RT-PCR assay in the presence of interferon-gamma (IFN-γ), lipopolysaccharide (LPS), and tumor necrosis factor alpha (TNF-α), respectively. The results showed that the IL-23p19 expression in Nthy-ori 3-1 cells was sharply increased by IFN-γ stimulation, but not by LPS or TNF-α, compared with the control group (Figure 2A). Equally, the IL-23 protein expression levels exhibited a similar pattern to the IL-23p19 mRNA expression levels, as shown by an immunofluorescence assay (Figure 2B). Additionally, the expression of the IL-23-specific receptor, IL-23R, both in the level of mRNA and on the surface of the Nthy-ori 3-1 cells was significantly induced after treatment with IFN-γ or LPS, but not TNF-α, compared with the control (Figures 2C,D). We also found that the level of IL-23 in the culture supernatants was increased after treatment with IFN-γ compared with the control. LPS stimulation also showed a slight increase in IL-23 secretion from TFCs (Figure 2E). These results collectively suggest that the pro-inflammatory cytokine, IFN-γ, is a major inducer of TFC-derived IL-23.

As aberrant ROS accumulation is a detrimental mediator for HT (13), to determine whether the increased IL-23 expression in TFCs indeed contributes to HT pathogenesis, we utilized Nthy-ori 3-1 cells to examine the effect of IL-23 on intracellular ROS levels. After treatment with varying IL-23 concentrations for 3 h, Nthy-ori 3-1 cells were stained with DCFH-DA, and ROS accumulation was detected by both immunofluorescence and flow cytometric methods. The results showed that the ROS production in Nthy-ori 3-1 cells was induced by IL-23 in a dose-dependent manner (Figure 3). To directly confirm that the ROS regulation was dependent on the IL-23-mediated signaling pathway, a neutralizing antibody against IL-23 was pretreated in the experiment system. The results showed that antiIL-23 antibody treatment abolished the IL-23-induced ROS accumulation in Nthy-ori 3-1 cells (Figure 3). These findings suggest that IL-23 mediated the ROS accumulation in TFCs.

Under normal physiological circumstances, autophagy prevents mitochondrial ROS release by digesting dysfunctional mitochondria, which is known to activate the inflammasome (6). Therefore, we speculated that IL-23-induced ROS accumulation in TFCs may be due to autophagy inhibition. To test this hypothesis, we first measured autophagy activity in thyroid tissue from HT patients using an IHC technique. The level of the autophagy-related protein, light chain 3 (LC3B), which is incorporated into autophagosomes and serves as a well-characterized indicator of autophagy activation, was highly expressed in the healthy control TFCs. However, LC3B protein expression in TFCs was sharply decreased in HT tissues compared with the healthy controls (Figures 4A,B). Considering that the IL-23/IL-23R expression levels were significantly increased in the HT patient TFCs, we performed in vitro experiments to test whether IL-23 contributed to autophagy inhibition. Nthy-ori 3-1 cells were cultured in the presence of IL-23 (50 ng/ml) for different time points, and LC3B-I/II expressions were analyzed by western blot. As expected, the IL-23 treatment significantly inhibited LC3B-II expression in Nthy-ori 3-1 cells (Figure 4C). Furthermore, the IL-23-induced inhibition of the autophagy-related protein, LC3B-II, was reversed by the addition of a neutralizing antibody against IL-23 in Nthy-ori 3-1 cells (Figure 4D).

To further confirm the inhibitory effect of IL-23 in autophagy, we applied an autophagy tandem reporter, mRFP-GFP-LC3, to infect Nthy-ori 3-1 cells, and the autophagic flux status was measured. The lysosomal environment can quench GFP fluorescence but does not affect mRFP fluorescence; therefore, the autophagosome appears as a yellow dot (the result of merged GFP and mRFP fluorescence) in the cytoplasm, and the autolysosome appears as a red dot in the lysosome (31, 32). As shown in Figures 4E,F, IL-23 treatment significantly reduced both the yellow dots that represent autophagosomes (P < 0.01) and the red dots that represent autolysosomes (P < 0.001). However, when the neutralizing antibody against IL-23 was added in the culture medium, the fluorescent puncta formation was significantly reversed in the TFCs. Taken together, these findings indicate that IL-23 inhibited autophagy activity, which might contribute to the ROS accumulation in the TFCs of HT patients.

To determine the molecular mechanism by which IL-23 inhibited the autophagy activity, we tested the activation of AKT/mTOR pathway, which is reported to contribute to the autophagosome formation (33). The results showed that IL-23 was a strong inducer of AKT/mTOR signaling pathway activation, exhibiting significantly increased phosphorylation of Akt (Ser473), mTOR (Ser2448), S6K (Thr389), and S6 (Ser235/236) in Nthy-ori 3-1 cells (Figure 5A). Apart from AKT/mTOR activation, IL-23 also promoted NF-κB p65 (Ser536) phosphorylation (an indicator of NF-κB activation), and Stat3 (Ser727) activation (a well-characterized responder to IL-23 stimulation) (Figure 5B). Additionally, neutralizing antibody treatment against IL-23 abolished the IL-23-induced AKT, S6k, p65, and Stat3 (Figure 5C) activation in Nthy-ori 3-1 cells. But the addition of Stattic, a pharmacological inhibitor of Stat3 (Ser727), did not rescue IL-23-induced suppression of LC3B-II protein (Figure 5D).

To determine whether the IL-23-induced AKT/mTOR activation contributes to autophagy inhibition, we took advantage of rapamycin, a pharmacological inhibitor of mTOR that positively regulates autophagy, to examine the autophagy activation. Nthy-ori 3-1 were pretreated with or without rapamycin and then stimulated with IL-23, and the phospho-S6 and LC3B-II expression levels were detected by western blot. A significant decrease in the phosphorylated S6 level in the presence of rapamycin was observed, which was accompanied with increased LC3B-II protein levels (Figure 6A), suggesting that IL-23-induced mTOR activation suppressed autophagy activation.

It was reported that mTOR can regulate NF-κB activity in cancer cells (34), therefore, we sought to determine whether IL-23-induced mTOR is the upstream signaling to regulate NF-κB activation in TFCs. The rapamycin treatment significantly decreased the level of phospho-NF-κB p65 protein in the IL-23-treated cells (Figure 6B), suggesting that mTOR functioned as an upstream regulator to mediate NF-κB p65 activation in IL-23-stimulated TFCs. Additionally, IKK-16, a pharmacological inhibitor of NF-κB, significantly rescued IL-23-induced suppression of the LC3B-II protein (Figures 6C,D). To further confirm the involvement of mTOR and NF-κB in autophagy, we used an autophagy tandem reporter, mRFP-GFP-LC3, to infect Nthy-ori 3-1 cells, and the autophagic flux status was analyzed. The autophagosome (yellow puncta) and autolysosome (red puncta) numbers in the cells were significantly increased after mTOR (rapamycin) or NF-κB (IKK-16) inhibitors treatments in the presence of IL-23 (Figures 6E,F; P < 0.001). These findings suggest that IL-23-induced autophagy inhibition is attributed to the activation of AKT/mTOR/NF-κB signaling pathway activation.

To further confirm that the IL-23-induced AKT/mTOR/NF-κB signaling pathway contributes to ROS accumulation in TFCs, the extent of ROS accumulation was examined in Nthy-ori 3-1 cells that were treated with IL-23 in the presence or absence of rapamycin or IKK-16. Expectedly, we found that the addition of rapamycin or IKK-16 both significantly inhibited IL-23-induced ROS accumulation in TFCs (Figures 6G,H). Collectively, we confirmed that IL-23-mediated mTOR/NF-κB activation contributed to autophagy inhibition and ROS accumulation in human TFCs.

To further confirm that autophagy inhibition and ROS accumulation in TFCs of HT patients were due to the activation of the AKT/mTOR/NF-κB signaling pathway, we evaluated the activation status of these signaling proteins in HT patient tissues by IHC. The results showed that the phosphorylated AKT, mTOR, and its downstream target S6 expression levels were significantly upregulated in the TFCs of HT patient tissues compared with the healthy controls, whereas the total AKT, mTOR, and S6 expression levels in the TFCs were at the similar levels between the healthy controls and HT patients (Figures 7A–C). Additionally, the phosphorylated NF-κB p65 levels were also increased in the TFCs of HT tissues compared with the healthy controls (Figure 7D). These data suggest that the AKT/mTOR/NF-κB signaling pathway was activated in the thyroid tissue of HT patients, which exhibited similar results as the IL-23-induced TFCs in vitro.

Nthy-ori 3-1, a human thyroid follicular epithelial cell line from the European Collection of Animal Cell Cultures, was cultivated in RPMI-1640 (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco) in the presence of 5% CO₂ and 37°C. The patients in this study were recruited from the Hospital Affiliated of Jiangsu University. Thyroid glands were obtained from 10 patients with HT (10 females, aged 34–61; mean age 48.1 ± 2.9 years) who underwent a thyroidectomy. An HT diagnosis was made based on clinical evaluations and Japanese guidelines as described previously (3, 42). All patients were not taking any anti-cytokine therapy and other drugs affecting thyroid function at the time of sampling. Thyroid tissues from five patients with a simple goiter (five females, aged 38–59; mean age 50.6 ± 3.8 years) were used as controls based on clinical evaluations and laboratory findings. All samples were taken in accordance with the regulations and approval of the Institutional Review Board of the Affiliated Hospital of Jiangsu University. In all cases, written informed consent was obtained, and the study was approved by the Ethics Committee of the Affiliated Hospital of Jiangsu University and conducted in accordance with the Declaration of Helsinki guidelines.

Primary antibodies, such as rabbit anti-LC3B; rabbit anti-AKT, rabbit anti-p-AKT (Ser473), rabbit anti-mTOR, rabbit anti-p-mTOR (Ser2448), rabbit anti-p70S6K, rabbit anti-p-p70S6K (Thr389), mouse anti-S6, rabbit anti-p-S6 (Ser235/236), rabbit anti-NF-κB p65, rabbit anti-p-NF-κB p65 (Ser536), rabbit anti-Stat3, rabbit anti-p-Stat3 (Ser727), and rabbit anti-β-actin were obtained from Cell Signaling Technology (USA). HRP-labeled goat anti-rabbit IgG, HRP-labeled goat anti-mouse IgG, PE-labeled goat anti-rabbit IgG secondary antibodies, anti-IL-23, and anti-IL-23R were from Santa Cruz (USA). IFN-γ and TNF-α were from PeproTech (USA). Recombinant human IL-23 protein and human IL-23p19 antibody, as well as the IL-23 Quantikine ELISA kit were from R&D systems (USA). The ROS assay kit was from Beyotime Biotech (China). Rapamycin and LPS were purchased from Sigma-Aldrich (USA). IL-23R was from Abcam. Stattic was from Merck Millipore (USA), and the IKK-16 was from Selleck Chemicals (USA).

Thyroid tissues were obtained from the Department of Pathology of The Hospital Affiliated to Jiangsu University. Tissue sections were performed as follows: the samples were fixed in 10% neutralized formalin, embedded in paraffin, sectioned into 4-µm sections, and mounted on slides. After deparaffinization and rehydration, antigen retrieval was performed by boiling the samples in 10 mmol/L citrate buffer (pH 6.0) for 10 min, and the slides were washed with phosphate-buffered saline, blocked with 5% normal goat serum, and then incubated with a primary antibody overnight at 4°C. After three washes with PBS, the sections were treated with the corresponding streptavidin peroxidase-conjugated secondary antibody (Maixin Biotechnology Co., Ltd.). Tissue sections were then counterstained with 3,3′-diaminobenzidine and hematoxylin and observed under a light microscope. The results of quantitative analyses of all samples with the Image-Pro plus 6.0 software (Version X; Media Cybernetics, USA) are presented graphically.

Total RNA was extracted from Nthy-ori 3-1 cells with Trizol reagent (Invitrogen, 15596026, USA) according to the manufacturer’s instructions after treatments with IFN-γ, LPS, or TNF-α for 6 h. RNA was eluted with RNase-free water, and the RNA concentrations were determined using a BioMate 3S analyzer (Thermo Fisher Scientific, USA). RT-PCR was performed using the RNA PCR Kit ver. 3.0 (Takara Biotechnology, RR019B, China). Using random hexamers and oligo dT as a primer, 1 µg of total RNA was reverse transcribed in a 20-µl volume with the PrimeScript^® RT reagent Kit (Perfect Real Time). The qRT-PCR was performed with SYBR^® Premix Ex Taq™ (Tli RNaseH Plus) in a real-time PCR Mx3000PTM System (Genetimes Technology, China). For mRNA detection, the following primers were utilized: IL-23p19 (forward: 5′-cccaaggactcagggacaac-3′; reverse: 5′-agagaaggctcccctgtgaa-3′); GAPDH (forward: 5′-aggtgaaggtcggagtcaac-3′; reverse: 5′-gggtggaatcatattggaaca-3′). GAPDH was used as the internal control. The 2^−ΔΔCt method was used to calculate the relative gene expression levels.

Supernatant IL-23 levels were measured by a quantitative enzyme immunoassay technique with the IL-23 Quantikine ELISA Kit according to the manufacture’s instructions (R&D System) after treatments with IFN-γ, LPS, or TNF-α for 24 h. A microplate reader (Thermo Fisher Scientific) capable of measuring absorbances at 450 and 570 nm was used to measure the color intensity that developed in each well. All assays were done in duplicate. The detection limit of the assay was 2.7 pg/mL.

The intracellular ROS level was monitored using a fluorescent probe, 2,7-dichlorohydrofluorescein diacetate (DCFH-DA) (Beyotime Biotechnology, China), which is a specific probe for hydrogen peroxide. Briefly, cells were seeded at 2 × 10⁵ cells/well in a six-well plate overnight and treated with indicated concentrations of IL-23 in the presence or absence of anti-IL-23 (5 µg/mL), rapamycin (10 nM), or IKK-16 (5 µM) pre-treatment, followed by incubation with a final concentration of 10 µM DCFH-DA for 30 min at 37°C. The cells were then collected, washed, and resuspended three times in RPMI-1640 without FBS. The DCF fluorescence of 1 × 10⁵ cells was detected and analyzed immediately using flow cytometry (FACS Calibur; BD Biosciences, USA) with an excitation wavelength of 488 nm and an emission wavelength of 525 nm. The intracellular ROS levels by flow cytometry analysis are presented by percentages of positive cells (% positive cells) and mean fluorescence intensity. Additionally, the DCF fluorescence of the ROS production from the Nthy-ori 3-1 cells was assessed using a fluorescence microscope (Olympus, Japan).

Nthy-ori 3-1 cells were cultivated in 24-well plates and treated with IFN-γ, LPS, or TNF-α for 24 h to test for IL-23 or IL-23R. First, the plates were washed twice with PBS, fixed in 4% paraformaldehyde at room temperature for 20 min, and permeabilized, and then non-specific receptors were blocked with 0.1% Triton-X-100 and 5% BSA in PBS at room temperature for 30 min. Next, the cells were incubated with primary antibody to detect IL-23 and IL-23R overnight at 4°C. They were then incubated with a secondary PE-labeled antibody at room temperature for 1 h. In all samples, DAPI (Beyotime Biotechnology) was added to visualize DNA in the cell nucleus. These stained cells were observed with a fluorescence microscope.

Nthy-ori 3-1 cells were infected with adenoviral-expressing mRFP-GFP-LC3 (Hanbio Biotechnology, China) at a multiplicity of infection of 200 for approximately 2 h. To quantify the number of puncta, mRFP-GFP-LC3B-transfected cells were seeded in a culture plate with 24 wells for 24 h before the treatment. Images were then recorded by fluorescence microscopy. The red puncta that overlay with the green ones and appear yellow in merged images are indicators of autophagosomes, while the free red puncta that do not overlay with the green ones and appear red in merged images are indicative of autolysosomes. The number of puncta per cell was determined using Image-Pro plus 6.0. More than 10 cells were analyzed for each condition.

Briefly, whole cell lysates, cytoplasm, and nuclear lysates were prepared with a protein extraction kit (Merck Millipore), and protein concentrations were detected using a Biomate 3s (Thermo Fisher Scientific). First, 5 µg of protein was subjected to electrophoresis on a 10–15% SDS-PAGE gel and then transferred onto a polyvinylidene difluoride membrane (Merck Millipore) by electrophoresis. After blocking for 1 h in 5% BSA, the membranes were incubated with antibodies against proteins or β-actin (standard controls) followed by HRP-conjugated secondary antibodies. The signals were detected using a Pierce ECL-plus substrate (Thermo Fisher Scientific) and scanned with a Fluor Chem FC3 camera system (Protein-Simple, USA). The images were analyzed with the Alpha View software (AIC, USA), and quantitative analyses are presented graphically.

Statistical analyses were performed using the Graphpad Prism 5 software (GraphPad Software, Inc., USA). The descriptive data are expressed as the mean ± SEM, and numerical data between two groups were compared using the homogeneity of variance or Wilcoxon Mann–Whitney tests, as appropriate. Differences in the mean values of various groups were analyzed using one-way ANOVA with the Tukey–Kramer multiple comparison test or the Kruskal–Wallis H test with the Nemenyi test. P < 0.05 was considered statistically significant.