Celastrol elicits antitumor effects by inhibiting the STAT3 pathway through ROS accumulation in non-small cell lung cancer

H460, PC-9, H520 and BEAS-2B lineage cells were obtained from the Cellular Resources Center of the SIBS (CAS, PRC). PC-9 cells were cultivated in Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher Scientific) with 10% fetal bovine serum (FBS) (TFS). H460, H520 and BEAS-2B cells were cultivated in RPMI-1640 intermediary (TFS) with 10% FBS (TFS). The environment (5% CO₂, 37 °C) was provided for cell incubation.

Cell viability and cytotoxicity in mankind NSCLC cells were identified by MTT assay. Cells (4–5 × 10³ per well) were seeded into plates with 96 wells. Then, different concentrations of celastrol were added into the plates for 48 h. Further, 25 μL MTT solution was added to the plates for 4 h before the MTT assay. Formazan crystals were dispersed in 150 µL DMSO and the optical density (OD) was identified via a Micro plate Reading device at 490 nm. The half-maximal inhibitory concentration (IC₅₀) data were identified via GraphPad Pro Prism 7.0.

The colony formation activity was assessed by the colony forming analysis. Cells were implanted into 6-well plates in 5% CO₂ atmosphere at 37 °C overnight. Celastrol (1, 2, and 4 μM) and DMSO were supplemented into the plates for 24 h. Further, 7–14 days later, when colonies were macroscopic, the plate was cleaned with phosphate-buffered saline (PBS), treated with 4% paraformaldehyde (PFA) fixation under room temperature (RT) for 15 min, then cleaned with pure water 3 times and it was dyed with Gentian violet for 10 min.

Wound-healing analysis was employed to assess the capability of cell motility. The cells were grown to about 80–90% confluence in plates with 6 wells, and afterwards, scraping of the cell mono-layer was performed at the center of the wells with a bacteria-free 10 μL pipet to create a scratched region. After that, the plates were cleaned in PBS to remove the scratched cells. Subsequently, the cells were cultivated in RPMI intermediary (without sera) with vector control or celastrol (1, 2, and 4 μM). Leica digital camera (Tokyo, Japan) was employed for microphotographs of cells that crossed the wound area.

Transwell analysis was employed to assess the cellular invasion capability. The cells were seeded into a transwell chamber at 2–4 × 10⁴ cells/chamber in conditioned medium overnight, and then they were treated with DMSO (Control) and celastrol (1, 2, and 4 μM) for 8 h. The upper intermediary was substituted by conditioned intermediary, and the lower intermediary was substituted by intermediary supplemented with 20% FBS. Afterwards, the cells were subjected to fixation in 4% PFA for fifteen min and dyed with Gentian violet for ten min. The invaded cells were photographed with a microscopy device.

Cells were subjected to ROS-sensitive dye treatment, and the production of ROS within cells was detected via the flow cell technique. Briefly, three types of cells were placed into cultivation plates with 6 wells and cultivated overnight in standard cultivation intermediary. Afterwards, the cells were stimulated with celastrol at the indicated times and concentrations. The cells were dyed with 10 μM DCFH-DA under 37 °C for 0.5 h. The ROS status was identified via the flow cell technique (FACSCalibur, BD Biosciences, USA) and fluorescence images. In certain assays, the cells were subjected to 5 mM NAC pretreatment for 60 min. In the entire assays, 10,000 living cells were assayed.

The Annexin V-FITC Programmed Cell Death Identification Kit I and Propidium Iodide (PI) were bought from BD Pharmingen (America). The human NSCLC cells were inoculated into dishes (each with 6 wells) for the purpose of growing the cells to 80% confluency in the maximum medium, and afterwards, the cells were subjected to different concentrations of celastrol (1, 2, and 4 μM) and DMSO (Control) treatment for 48 h to assess the roles of celastrol in programmed cell death. In certain assays, the cells were subjected to 5 mM NAC pretreatment for 60 min. The cells were harvested and cleaned with cold PBS for 2 times, and afterwards, they were subjected to resuspension in binding buffering solution as per the specification of the Programmed Cell Death Kit. The cells were synchronously cultivated with luciferin-labeled Annexin V and PI. Afterwards, the Annexin V-binding buffering solution was added into the mix and the fluorescence property was identified on a FACSCalibur (USA). The Flowjo program was employed to analyze the figures.

NSCLC cells were seeded in a six-well culture plate. After treatment with celastrol for 12 h, 4% PFA was used for 10 min for cell fixing, and the cells were cleaned in PBS, followed by Hoechst 33342 dyeing for 20 min. The characteristic of cell nucleus was observed by employing a fluorescence microscope (Leica, Tokyo, Japan).

Cells (1 × 10⁶ to 5 × 10⁶) were spread in petri dishes. Celastrol (1, 2, and 4 μM) was added to induce apoptosis in NSCLC cells. The cells were resuspended and lysed in 50 μL pre-treated lysis buffering solution for ten min on ice and then centrifuged for 1 min. The supernatant fraction was immediately transferred into a fresh tube. The reaction system was subsequently configured as specified in the Caspase 3 binding assay kit and maintained under 37 °C for 120 min. The absorbance was identified under 405 nm via a SpectraMax iD3 (USA).

Tissue or cellular protein lysate buffering solution was utilized to perform protein extraction. Three cancer cell lines were seeded in the six plates, treated with celastrol for 3 h, and then dissolved with a cold lytic buffer. The 10% or 12% SDS-PAGE was used for protein separation. Proteins were translocated onto a PVDF film, followed by incubation for 1.5 h with 5% skimmed milk. The film was cultivated under 4 °C with specific antisubstances. Afterwards, the membrane underwent 1 h incubation at room temperature with the corresponding secondary antibody. Image J computer software was used for determining the density of the immune response bands.

We employed the NE-PER Nuclear and Cytoplasmic Extraction Kit to isolate the nucleus and cytoplasmic proteins in H460 cells (TFS). Cells were subjected to treatment with diverse concentrations of celastrol for 12 h, and afterwards, they were stimulated with IL-6 for 0.5 h. According to the protocol, the cell lysate was obtained. Immunoblotting assay was performed to identify the related protein expression.

H460 cells were inoculated into a fluorescence cuvette, and they were treated with celastrol and/or IL-6. Then, the cells were subjected to 4% PFA fixation and permeabilization via 0.5% Triton X-100 (in PBS), and cleaned three times with PBS. Afterwards, they were subjected to 1% BSA blockade for 60 min. Cell slides were cultivated with the first antisubstance (P-STAT3) in appropriate proportions. The cells were cultivated with an anti-rabbit secondary antisubstance (Alexa Fluor® 488) on the following day and then with the DAPI stain protected from light. The cell slides were cleaned in PBS again and subsequently sealed by anti-fluorescence quenching sealing tablets to prevent fluorescence quenching until they were captured by confocal microscopy (Nikon C2).

The WMU Animal Policy and Welfare Board accepted the entire murine assays. Athymic BALB/c nude mice (16–18 g) were bought from the Vital River Experimental Animal Center (PRC). For the heterograft model, H460 cells (5 × 10⁶ cells) blended with the same volume of PBS and Matrigel in 100 μL were delivered into the hind flank of the animals. If tumor volumes were ~ 50 mm³, the animals were separated into 4 experiment groups and they were injected with 4 mg/kg napabucasin (a recognized P-STAT3 suppressor) or celastrol (2 and 4 mg/kg) intraperitoneally every other day. Tumor volume was calculated as V = (L × W × W)/2, where L and W denote the length and width, respectively. The mice were euthanized on day 15, 2 h after the final drug administration. The tumors, hearts, livers, kidneys, and lungs were collected for the histological and immunoblotting assays.

Tumor samples were subjected to fixation under RT with 4% PFA and embedded with paraffin. The thickness of the treated tissues slices was 5 μm. After that, the samples were cultivated overnight with the proper primary antisubstance under 4 °C. The signal was identified via the relevant second antisubstance. Subsequently, these sections were dyed with DAB and counterstained with hematoxylin. Imaging was performed via optical microscopy.

The hearts, lungs, kidneys, and livers of mice were treated with 4% PFA fixation and paraffin embedment. The treated tumor sample slices (5 µM) were deparaffinized and subjected to rehydration, and afterwards, they were treated with H&E dyeing. Imaging was performed via optical microscopy.

The chemical structure of celastrol is shown in Fig. 1A. For the purpose of evaluating the suppressive role of celastrol, cellular activity was evaluated via the MTT analysis. Celastrol significantly inhibited the activity of H460, PC-9, and H520 cells, with IC₅₀ values of 1.288 µM, 2.486 µM, and 1.225 µM, respectively (Fig. 1B). The cell counting trypan blue assay showed that the growth of H460 and PC-9 cells was significantly inhibited by celastrol in a concentration dependent manner (Additional file 1: Fig. S1A). Besides, the colony forming analysis revealed that celastrol could markedly inhibit the colony forming capability of NSCLC cells (Fig. 1C). Additionally, for the purpose of determining whether the treatment with celastrol was related to oncocyte migration and invasion, variations in migratory and invasive potency were identified via the wound healing assay and transwell assay; in contrast to controls, the migratory and invasive potency of these three lineage cells was dose-dependently inhibited by celastrol (Fig. 1D and E, and Additional file 1: Fig. S1B). At the same time, we conducted the same experiment in normal human cells. The results showed that celastrol had no obvious inhibitory effect on cell viability. To sum up, these results revealed that celastrol could validly suppress the viability, migration, and invasion of mankind NSCLC cells.

To assess whether celastrol can trigger programmed cell death, 3 lineage cells, H460, PC-9, and H520 were subjected to celastrol treatment at 3 diverse concentrations for 24 h, followed by staining with Annexin V FITC and PI, and the proportion of programmed cell death was identified via the flow cell technique. Celastrol markedly triggered programmed cell death in NSCLC cells (Fig. 2A). The outcomes of Hoechst 33342 dyeing showed that the celastrol-treated cells showed a potent blue fluorescent result and obvious apoptosis bodies. It was further confirmed that celastrol could induce apoptosis (Fig. 2B). Here, we also found that the activity of caspase 3 was significantly increased, as determined by the caspase activity assay, and the pro-apoptotic effect of celastrol was confirmed (Fig. 2C). Overall, celastrol markedly promoted apoptosis of NSCLC cells.

The production of ROS will cause programmed apoptosis. Promoting apoptosis by increasing ROS in tumor cells is one of the mechanisms of action of many chemotherapeutic agents [26, 27]. For the purpose of clarifying whether celastrol-triggered programmed cell death occurred due to ROS generation, we exposed H460, PC-9, and H520 cells to 4 µM celastrol and identified ROS via DCFH-DA dyeing. The numerical results herein revealed that celastrol remarkably elevated the ROS levels in NSCLC cell lines in a time-dependent manner (Fig. 3A, B and Additional file 2: Fig. S2A, B). Celastrol displayed dosage-reliant triggering of ROS in H460, PC-9, and H520 cells. As anticipated, the ROS signal was decreased in cells subjected to oxidation preventive NAC pretreatment (Fig. 3C, D and Additional file 2: Fig. S2C, D). ROS fluorescence images also showed that the intracellular green fluorescence increased rapidly with the time period of celastrol treatment (Fig. 3E and Additional file 2: Fig. S2E, F). Similarly, the intracellular green fluorescence accumulated with increasing drug concentration, and pretreatment with NAC could eliminate this phenomenon (Fig. 3F and Additional file 2: Fig. S2G, H). Afterwards, NAC was utilized to identify the participation of ROS in celastrol-triggered programmed cell death in NSCLC cells. Flow cytometry revealed that celastrol-triggered programmed cell death in H460 and PC-9, and H520 cells was reversed by 5 mM NAC pretreatment (Fig. 3G). These results were confirmed by the levels of apoptosis proteins, BAX and BCL-2 (Fig. 3H). Overall, the results indicated that an enhanced ROS level is the critical modulator of celastrol-triggered programmed cell death in NSCLC cells.

Increased ROS levels have been found to trigger programmed cell death via various downstream pathways, such as ER stress and mitochondrion-related functional disorder [28]. Therefore, our team examined ER stress-related proteins, such as ATF4, P-eIF2α, and eIF2α, to determine if ROS levels elevated via celastrol triggered ER stress in NSCLC cells. These biomarkers of ER stress were rapidly increased in H460, PC-9, and H520 cells exposed to celastrol and this triggering occurred as early as 1 h after the treatment (Fig. 4A). Celastrol also dosage-dependently stimulated the production of ER stress-associated proteins, and after NAC pretreatment of NSCLC cells for 1 h, this induction was prevented (Fig. 4B). These research findings provide conclusive evidence that celastrol triggers ROS; thus, inducing ER stress and causing programmed cell death in NSCLC cells.

Previous studies have reported that ROS regulate a variety of cell signal paths, and elevated ROS levels can inhibit the STAT3 pathway [21]. As a transcriptional factor, STAT3 is pivotal for oncocyte growth and development [29]. For this reason, our team reckoned whether celastrol could influence the IL-6/STAT3 signaling pathway in NSCLC cells. Cytokine IL-6 is the main inflammatory mediator and stimulator of STAT3, which can block programmed cell death and facilitate cancer survival. Therefore, our team used 20 ng/ml IL-6 and different concentrations of celastrol to treat H460, PC-9, and H520 cells. The result showed that the stimulation of STAT3 triggered by IL-6 was dosage-dependently reversed by celastrol (Fig. 5A). In addition, as per the immunofluorescence dyeing analysis, IL-6 therapy induced the cytoplasm-nucleus P-STAT3 transfer, which could be inhibited by celastrol (Fig. 5B). We used western blotting to detect the protein expression in the cytoplasm and nuclear extracts. These results further confirmed that the nuclear translocation of P-STAT3 was significantly inhibited by celastrol (Fig. 5C). We found that P-STAT3 in H460, PC-9, and H520 cells was markedly reduced with celastrol treatment; however, under the same treatment conditions, the expression of STAT3 did not change significantly. Interestingly, the inhibitory effect of celastrol-induced P-STAT3 was reversed after NAC pretreatment for 1 h in the three cell lines (Fig. 5D). Finally, we investigated the expression of ER stress related proteins after IL-6 and celastrol co-treatment, The experimental results are shown in the Additional file 2: Fig. S2I, The level of ER stress in IL-6 and celastrol co-treated group was higher than that in IL-6 treated group, but lower than that in celastrol treatment group, this suggests that IL-6 stimulated STAT3 phosphorylation can reduce the ER stress level caused by celastrol. This suggests that celastrol inhibits the STAT3 pathway by accumulating intracellular ROS. To sum up, these outcomes revealed that celastrol suppressed the IL-6/STAT3 pathway via ROS production in NSCLC cells.

To further evaluate the anti-cancer efficiency of celastrol in vivo, we established and employed the cell-derived xenograft NSCLC model, and the H460 NSCLC cell line was selected for model establishment. The STAT3 suppressor napabucasin was utilized as the positive control. The mean volumes of tumors in the celastrol group (2 mg/kg or 4 mg/kg) were smaller compared to those in the control group and napabucasin (4 mg/kg) group (Fig. 6A), and the image of the tumors excised from nude mice after seven sessions of treatment with celastrol or napabucasin showed similar results (Fig. 6B). Consistent with the above data, the results shown in Fig. 6C indicated great reduction in the tumor weight after treatment with celastrol. Importantly, the body weight of mice did not decrease obviously (Fig. 6D), and the image of HE staining of vital organs (Heart, liver, kidney, and lung) did not present apparent toxicities (Fig. 6E). Thus, celastrol showed relative safety in vivo. As presented in Fig. 6F, the immunohistochemical results of NSCLC cells showed that, after injection of celastrol, the expression of P-STAT3 was decreased as the P-eif2α proteins increased (Fig. 6F). Western blotting showed that treatment with celastrol decreased the level of P-STAT3 and triggered the expression of apoptotic protein BAX; the BCL-2 level was inhibited. Meanwhile, ER stress induced by celastrol in the in vivo NSCLC model was also verified; the levels of ATF4 and P-eIF2α proteins were increased (Fig. 6G). Altogether, these results suggested that celastrol suppresses cancer development and cellular proliferative activity in vivo.