The human thyroid cancer cell lines 8505C and FTC133 were cultured at 37°C in Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher, Shanghai, China) supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher), 100 U/ml penicillin and 100 g/ml streptomycin in a humidified incubator with 5% CO₂. Antibodies against Beclin-1 (mouse, 1:1,000, SC-10086), GAPDH (mouse, 1:1,000, SC-51907), p-ERK (mouse, 1:200, SC-7383), ERK (rabbit, 1:1,000, SC153) and AKT (mouse, 1:1,000, SC-5298) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA), and the anti-p-AKT (Ser473) antibody was purchased from Abcam (rabbit, 1:200, Cambridge, Massachusetts, USA). Antibodies against cleaved-caspase-3 (rabbit, 1:1,000, #9661), p62 (rabbit, 1:1,000, #88588), LC3 (rabbit, 1:1,000, #12741), p70s6 (rabbit, 1:1,000, #2708) and p-p70s6 (rabbit, 1:500, #2211) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA). BEZ-234 was purchased from Selleck (Shanghai, China), and sorafenib was generously donated by the Bayer Company (Shanghai, China). Cells were transfected with scrambled (negative control) small interfering (si)RNA (5′-UUCUCCGAACGUGUCACGUTT-3′ and 5′-ACGUGACACGUUCGGAGAATT-3′), AKT siRNA (5′-GACGGGCACAUUAAGAUCATT-3′ and 5′-UGAUCUUAAUGUGCCCGUCTT-3′), or ATG5 siRNA (5′-GUCCAUCUAAGGAUGCAAUTT-3′ and 5′-AUUGCAUCCUUAGAUGGACTT-3′) and purchased from GenePharma (Shanghai, China). Transfection was performed using Lipofectamine™ 2000 (Invitrogen, Shanghai, China).

Nude BALB/c mice (5–6 weeks old; male; Animal Core Facility of Nanjing Medical University, Nanjing, China) were used to generate xenograft tumor models. All animal studies complied with the management rules of the National Health and Family Planning Commission of China and were approved by the Ethics Committee of Zhejiang Cancer Hospital. A suspension of 5×10⁶ 8505C or FTC133 cells in 100 µl PBS was inoculated subcutaneously into the right flanks of the mice. When the tumor sizes averaged approximately 5×5 mm, the mice were randomly divided into four groups that were matched for tumor volume, and treatment was initiated (6 mice per treatment group). Treatment groups were as follows: the vehicle control group (DMSO, ≤0.1%), the chloroquine group (CQ; Sigma-Aldrich, Shanghai, China), the sorafenib group, and the sorafenib + CQ group. Sorafenib was administered at doses of 30 mg/kg every day for 21 days. CQ was administered at doses of 60 mg/kg every day for 21 days. In the sorafenib + CQ group, sorafenib was administered 20 min after CQ administration. All drugs were administered via intraperitoneal injection. A caliper was used to assess the tumors every 4 days. After 21 days of treatment, the mice were sacrificed and the tumor tissues were harvested for study. The mice were sacrificed by anesthetizing with an intraperitoneal injection of 0.8% pentobarbital sodium (60 mg/kg), followed by cervical dislocation. All efforts were made to reduce pain experienced by the mice.

Tumor tissues or cell pellets were lysed in RIPA lysis buffer and a protease inhibitor cocktail (Sigma-Aldrich). Total protein samples were separated by 10% or 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat dry milk and incubated overnight with the various primary antibodies at 4°C. Goat anti-mouse (sc-2005) or anti-rabbit (sc-2004) antibodies were purchased from Santa Cruz Biotechnology, Inc. and diluted to 1:1,000 in 5% skim milk, Tris-HCl (pH 7.5) and 0.1% Tween-20. The immunoblots were subsequently washed and incubated with the goat anti-mouse or anti-rabbit antibody for 1 h at room temperature. The results were visualized using chemiluminescence reagent (sc-2048, Santa Cruz Biotechnology, Inc.).

Total RNA was isolated using an RNA Prep Pure kit (Tiangen, Beijing, China), according to the manufacturer's instructions. The synthesis of cDNA was performed with HiScript II Q RT Super Mix for qPCR (+gDNA wiper) (Vazyme Biotech Co., Nanjing, China). qPCR was performed with an AceQ qPCR SYBR Green Master Mix (without ROX) (Vazyme Biotech Co.). GAPDH was used as an internal control. Primer sequences used in the qPCR analysis were purchased from GenePharma (Shanghai, China), and the sequences were as follows: AKT, 5′-CCAACACCTTCATCATCC-3′ and 5′-CTCCTCCTCCTGCTTCTT-3′; ATG5, 5′-GCAGATGGACAGTTGCACACA-3′ and 5′-TTTCCCCATCTTCAGGATCAA-3′; and GAPDH, 5′-CGGAGTCAACGGATTTGGTCGTAT-3′ and 5′-AGCCTTCTCCATGGTGGTGAAGAC-3′. Melt curve was made to determine the optimal condition (95°C 15 sec, 60°C 30 sec, 95°C 15 sec). The PCR protocol is as follows: denaturation 95 °C for 5 min, then 40 amplification cycles (95°C for 5 sec and 60°C for 31 sec, at a ramp-rate of 1.6°C/sec).

8505C and FTC133 cells (2×10⁵/well) were seeded into a 6-well plate. After treatment with ATG5Ri mimic, sorafenib (10 µΜ), BEZ-235 (0.8 µM) or combined sorafenib + BEZ-235, the cells were harvested and analyzed for apoptosis by Annexin V and propidium iodide staining, using an FITC Annexin V Apoptosis Detection kit (Life Technologies, Waltham, MA USA), according to the manufacturer's instructions, and a flow cytometer (FACS, Beckman Coulter, Miami, FL, USA).

Paraffin-embedded mouse tumor samples were sliced into 5-µm sections and mounted on microscope slides. Slides were incubated at 37°C overnight for deparaffinization. The slides were dipped in dimethylbenzene three times (10 min/dip). The slides were then immediately dipped in 100, 95, 90, 80, 70 and 50% alcohol continuously for 2 min at each percentage (%) of alcohol. Antigen retrieval was performed by heating for 15 min in 0.01 M sodium citrate buffer (pH 6.0). The slides were washed in TBS three times (5 min/wash), treated with 3% H₂O₂ and blocked with 5% BSA for 20 min, then incubated with 50 µl cleaved-caspase-3 (1:200 dilution) at 4°C overnight. Subsequently, the slides were washed, and incubated with a mouse anti-rabbit antibody for 1 h at room temperature. A DAB horseradish peroxidase color development kit (ZSGB-BIO, Beijing, China) was used to detect positive staining. The slides were also counterstained with hematoxylin for 25 sec, washed under running water for 3 min, then washed with distilled water for 1 min followed by dipping in 50, 70, 80, 90 and 100% alcohol continuously for 1 min each. Finally, the slides were dipped in dimethylbenzene for 3 min, air-dried and mounted with Permount™ Mounting Medium (eBioscience, San Diego, CA, USA).

8505C and FTC133 cells were plated at a density of 1×10⁴ cells/well in 96-well microtiter plates, and each plate was incubated for 24 h at 37°C in 5% CO₂. Following treatment, the absorbance of the contents of each well was assessed at 470 nm following the use of a CCK-8 kit (Beyotime, Guangzhou, China).

Data are presented as the mean ± standard error of the mean. The results were analyzed using One-way ANOVA, and multiple comparison using Student-Newman-Keuls (SNK) and least significant difference (LSD) tests. P<0.05 was considered to indicate a statistically significant difference.

To investigate whether autophagy influences the therapeutic efficacy of sorafenib in vivo, 5–6-week-old male BALB/c nude mice were used as an animal model. CQ is a potential autophagy inhibitor that has been used to regulate autophagy in vivo and in vitro (19,20). Mice were randomly divided into control, sorafenib (30 mg/kg/day), CQ (60 mg/kg/day), or sorafenib + CQ groups (n=6/group) at day 7 after subcutaneous injection of 5×10⁶ cancer cells. Tumor volumes were monitored every 4 days, and the mice were sacrificed after 21 days of treatment for isolation of tumor tissues (Fig. 1A and B). Compared with the tumor sizes in the control group (1,280.93±245.27 mm³ in 8505C xenograft mice; 2,176.57±343.82 mm³ in FTC133 xenograft mice), the tumor volumes significantly decreased after administration of CQ or sorafenib alone (602.36±101.18 and 698.61±54.53 mm³ in 8505C xenograft mice, respectively; 600.42±107.33 and 697.07±171.34 mm³ in FTC133 xenograft mice, respectively) (Fig. 1C and D), while the combined treatment of CQ + sorafenib significantly suppressed tumor volumes compared with CQ or sorafenib treatments alone in the 8505C and FTC133 xenograft mice (358.99±86.54 and 344.36±84.36 mm³, respectively) (Fig. 1C and D). There were no significant differences between CQ and sorafenib treatment groups in 8505C or FTC133 xenograft mice.

8505C and FTC133 xenograft mice were sacrificed and the tumor tissues were isolated for further studies after 21 days of treatment. The proteins LC3 and p62 were used as autophagy markers for monitoring autophagic flux and were examined by western blotting of lysed tumor tissues. The expression level of LC3II/GAPDH increased significantly in the sorafenib treatment group compared with the control group, while the expression level of p62 was reduced in the sorafenib treatment group compared with the control group (Fig. 2A-C). To confirm the effect of sorafenib on autophagy in thyroid cancer, we treated 8505C and FTC133 cell lines with various concentrations of sorafenib in vitro for 24 h. With an increase in sorafenib concentration, LC3II/GAPDH levels increased significantly compared with the control. When treated with 10 µΜ sorafenib, p62 protein levels were obviously reduced compared with the control (Fig. 2D-F). These results indicated that sorafenib increased autophagic flux in 8505C and FTC133 xenograft mouse tumors and in vitro experiments. In CQ-treated tumor tissues, the expression level of LC3II/GAPDH increased significantly compared with the control. However, the expression level of p62 in the CQ treatment group increased significantly compared with the control group, which appeared to be contrary to the sorafenib group (Fig. 2A-C). These results indicated that autophagic flux was inhibited by CQ.

Caspase-3 is activated in apoptotic cells by both extrinsic and intrinsic pathways (21,22). We used caspase-3 activation to assess apoptosis since caspase-3 is an early marker for apoptosis (23). To assess the effect of sorafenib and CQ on tumor apoptosis, we used IHC analysis of cleaved caspase-3 in different treatment groups of mice. As expected, in the sorafenib- and CQ-treated groups, the level of cleaved caspase-3 increased compared with the control, suggesting CQ or sorafenib induced apoptosis in 8505C and FTC133 xenograft mouse tumors. In the CQ + sorafenib combined treatment group, a further increase of cleaved caspase-3 was observed compared with CQ- or sorafenib-treated groups in both 8505C and FTC133 xenograft mice (Fig. 2G). Our results demonstrated an activation of caspase-3 in 8505C and FTC133 xenograft mouse tumors, suggesting that caspase-mediated apoptosis occurs following sorafenib and CQ treatment.

Tumor tissues were used to analyze the target-suppressing effects of sorafenib. As shown in Fig. 3A-D, compared with the control groups, the expression of the ERK protein was not significantly different in the sorafenib treatment groups of 8505C and FTC133 xenograft mice. However, phosphorylation levels of ERK were significantly inhibited in both 8505C and FTC133 xenograft mice that were treated with sorafenib. Furthermore, we observed that the phosphorylation levels of AKT and p70s6k were significantly reduced in the sorafenib treatment groups compared with control groups. The results indicated that CQ exerted similar effects to sorafenib. The phosphorylation levels of ERK, AKT and p70s6k were significantly reduced after CQ treatment. Furthermore, in 8505C and FTC133 xenograft mice treated with combined CQ + sorafenib, the dephosphorylation effect was enhanced compared with that achieved by either drug alone (Fig. 3A-D).

Beclin-1 is a key autophagy-related gene that takes part in the initial process of autophagy (24). Beclin-1 protein expression was increased in xenograft mice treated with CQ or sorafenib, while combined CQ + sorafenib treatment increased the expression of Beclin-1 protein more than CQ or sorafenib treatment alone (Fig. 3A and E).

The MAPK pathway is a well-known target of sorafenib. However, the AKT/mTOR pathway was inhibited by sorafenib in the present study. The AKT/mTOR pathway is a major pathway that regulates the process of autophagy (25,26). Next, we demonstrated the role of the AKT/mTOR pathway in sorafenib-treated thyroid cell lines. To construct an AKT gene-silencing cell model, we transfected siRNAs into 8505C and FTC133 cell lines. The silencing effect was examined by western blotting. The protein levels of p-AKT were reduced significantly in AKT siRNA-transfected cell line models (27). The LC3 and p62 proteins were examined again for the evaluation of autophagy levels. Following silencing of AKT by siRNA, LC3II increased and p62 decreased compared with the control group (Fig. 4A-C). AKT silencing in combination with sorafenib treatment significantly enhanced LC3II expression levels and reduced p62 compared with AKT silencing or sorafenib treatment alone (Fig. 4A-C). BEZ-235, a dual, ATP-competitive PI3K and mTOR inhibitor, was selected to demonstrate the role of the AKT/mTOR pathway in sorafenib-induced autophagy. In a previous study, 0.8 µM BEZ-235 inhibited AKT/mTOR pathway activity and inhibited 8505C and FTC133 cell line proliferation (27). Thus, 8505C and FTC133 cell lines were treated with 0.8 µM BEZ-235 for 24 h, and western blotting was used to evaluate the expression levels of LC3II and p62. LC3II expression increased and p62 expression decreased compared with the control group in both 8505C and FTC133 cell lines (data not shown). Combined treatment with sorafenib and BEZ-235 significantly enhanced LC3II expression and reduced p62 expression (data not shown). The AKT/mTOR pathway negatively regulates autophagy and apoptosis (17). In the present study, 8505C and FTC133 cell lines treated with AKT siRNA exhibited increased cleaved capase-3 expression, and combined treatment of AKT siRNA with sorafenib enhanced cleaved capase-3 expression compared with either treatment alone (Fig. 4D and E).

In our results, sorafenib treatment or AKT/mTOR pathway suppression induced autophagy in thyroid cancer cell lines. As autophagy plays a dual role in cancer treatment, we examined the role of autophagy in sorafenib treatment of thyroid cancer (11,12). ATG5 is a key gene in the process of autophagy that can be silenced to demonstrate the role of autophagy in different treatments (6). Three different siRNAs (ATG5Ri) were transfected into 8505C and FTC133 cell lines, and RT-qPCR was used to evaluate their silencing effect. ATG5 siRNA2 appeared to have strong silencing effects and was selected for use in further experiments (data not shown).

Treatment with sorafenib or ATG5Ri increased the apoptosis of 8505C and FTC133 cell lines compared with the control group (apoptosis rate, 4.9 and 5.6%); the apoptosis rates after sorafenib or ATG5Ri treatment were 24.7 and 25.2%, respectively, in 8505C cells (Fig. 5A and B), and 53.2 and 27.3%, respectively, in FTC133 cells (Fig. 5C and D). Combined treatment with sorafenib and ATG5Ri increased apoptosis rates more than either treatment alone in both cell lines (42.8% in 8505C cells; 69.7% in FTC133 cells; Fig. 5). Silencing of ATG5 significantly increased the apoptosis rate after combined treatment of sorafenib and BEZ235 in FTC133 cells (72.6 vs. 68.1%; Fig. 5). Silencing of ATG5 increased the anti-proliferative effect of sorafenib and combined treatment of sorafenib and BEZ-235 in both 8505C and FTC133 cell lines (Fig. 6).