ERRα suppression enhances the cytotoxicity of the MEK inhibitor trametinib against colon cancer cells

The human colon cells that were obtained from the State Key Laboratory of Biotherapy, West China Hospital, Sichuan University included HCT 116 (KRAS G13D), SW480 (KRAS G12 V) and SW1116 (KRASG12A) were grown in Dulbecco’s modified Eagle medium supplemented with 10% foetal bovine serum (FBS, Gibco, USA), 100 mU/mL penicillin, and 100 μg/mL streptomycin in a 5% CO₂ atmosphere at 37 °C. All the cell lines used were negative for mycoplasma. Trametinib (GSK1120212), XCT790 (HY-10426) and CCCP (HY-100941) were from Medchemexpress. Simvastatin was purchased from J&K Scientific Ltd. (Beijing, China). These agents were all dissolved in dimethyl sulfoxide (DMSO). ERRα luciferase reporter plasmid (pGMERRα-Lu) was purchased from YESEN biology (Shanghai, China). Moreover, the following primary antibodies were obtained from Abcam: UK:rabbit anti-human c-Myc mAb and rabbit anti-human cyclin D1 mAb. The following antibodies were obtained from Santa Cruz: rabbit anti-human Bax mAb, mouse anti-human ERRα mAb, mouse anti-human IDH3A mAb and mouse anti-human GAPDH mAb.

The human colon cancer tissue microarrays used in this study were prepared by Shanghai Outdo Biotech Co., Ltd. (Shanghai, China). All the patients signed informed consent forms. This study was approved by the Ethics Committee of Taizhou Hospital of Zhejiang Province.

For the cell proliferation assays, the cells were seeded in 96-well plates for 24 h and were allowed to adhere overnight in regular growth media. After treatment with the indicated drugs, the relative cell growth was measured using the Cell Counting Kit-8 (Dojingdo, Kumamoto, Japan). For the clonogenic assays, the cells were seeded into 35-mm dishes and were cultured in Dulbecco’s modified Eagle medium with 10% foetal bovine serum and 100 IUml-1 penicillin/streptomycin overnight. The cells were then treated with the drug, as indicated, in complete media for 5–6 days. The growth media with or without drug were replaced every 2 days. The remaining cells were fixed with methanol (1%) and formaldehyde (1%), stained with 0.5% crystal violet, and photographed using a digital scanner. All the experiments were performed at least three times. Representative experiments are shown.

The siRNAs against ERRα and the negative controls, the lentiviral shRNA expression vector targeting hERRα and scrambled control (pGPU6/GFP/Neo-shNC) were synthesized by GenePharma (Shanghai, China). ERRα luciferase reporter plasmid (pGMERRα-Lu) was purchased from YESEN biology (Shanghai, China) http://​www.​yeasen.​com/​index.​htm; The sequence of the shRNA/siRNA/pGMERRα-Lu sense was as follows: pGPU6/GFP/Neo-shERRα#1: 5’-CACCGTGGTGGGCATTGAGCCTCTCTACATTTCAAGAGAATGTAGAGAGGCTCAATGCCCACCATTTTTTG-3′ and pGPU6/GFP/Neo-shERRα#2: 5’-CACCGAATGCACTGGTGTCACATCTGCTGTTCAAGAGACAGCAGATGTGACACCAGTGCATTCTTTTTTG-3′; pGPU6/GFP/Neo-shNC:5'-CACCGTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAATTTTTTG-3′; siERRα: 5’-GAAUGCACUGGUGUCACAUCUGCUG-3′. The sequence of ERRα response element (32–91): GGCCTAACTGGCCGGTACCGCTAGCCTCGATAGCTTGAAGAGGTCACTGTGACCTACAACGAGCTTGAAGAGGTCACTGTGACCTACAACGGCGCGTAGA [29]; And the siRNAs, shRNAs and pGMERRα-Lu were transfected into the cells using Lipofectamine 2000 (Invitrogen/Life Sciences) according to the manufacturer’s instructions.

Immunohistochemistry (IHC) was performed on all colon cancer samples and the tissues of xenograft tumour using biotin-streptavidin HRP detection systems. Paraffin-embedded tissue sections were collected. After deparaffinization with xylene and dehydration in a graded alcohol series, the tissue sections were subjected to antigen retrieval by microwaving in sodium citrate buffer for 10 min and then inhibiting endogenous peroxidase activity. After nonspecific binding was blocked, the slides were incubated with ERRα (1:100) and IDH3A (1:200) antibody (Santa Cruz Biotechnology, CA, USA); c-Myc and Cyclin D1 antibody (1:200; Abcam, Cambridge, UK,) in phosphate-buffered saline (PBS) overnight at 4 °C in a humidified container. Biotinylated secondary antibodies (Zhongshan Golden Bridge Biotechnology Co. Ltd., China) were then used according to the manufacturer’s recommendations. The sections were incubated with HRP-streptavidin conjugates appropriate for detecting ERRα; IDH3A; c-Myc and Cyclin D1. The brown color indicative of peroxidase activity was developed by incubation with 0.1% 3,3-diaminobenzidine (Zhongshan Golden Bridge Biotechnology Co. Ltd. China) in distilled water for 1–3 min at room temperature. The appropriate positive and negative controls were included in each IHC assay.

A dual luciferase reporter gene assay was performed by using a multifunctional microplate reader (Synergy H1, BioTek, Vermont, USA) and Dual-Luciferase® Report Assay System kit (TransGen Biotech, China). The following procedures were used: Luciferase Reaction buffer II was mixed with thawed Luciferase Reaction substrate II, placed in a centrifuge tube pre-wrapped in foil, and stored at − 80 °C. Thawing took place at room temperature in a dark environment. Stop & Glo buffer was thawed at room temperature and added to 50 x Stop & Glo substrate to prepare a 1 x Stop & Glo Reagent. The cell culture medium was discarded and the cells were washed twice with PBS. Any residual liquid was also removed before 100 uL of 1 x CLB lysis buffer (5 x CLB was diluted to 1× CLB with sterile water) was added into each well. The cells were lysed by shaking on a shaker for 15 min after which 20 uL of the cell lysate was drawn and added to a 96-well opaque detection plate. A total of 100 uL of LARII was the added quickly to the wells containing lysate and gently mixed. Cell lysate was detected on the multi-function microplate reader. Parameters were 10s of reading and 2–3 s delays. The firefly luciferase activity value (F) was measured in relative luminometer units (RFUs). After F was measured, the 96-well plate was immediately taken out of the multifunctional microplate reader and a 100 uL of 1 x Stop & Glo Reagent was added to each well and mixed evenly. The multifunctional microplate reader was used to measure RLUs of renilla luciferase activity (R) over 10s read periods and 2–3 s delays. The relative transcriptional activity of the promoter region was determined by the F/R ratio.

The cell migration assay was performed using a BD BioCoat™ Matrigel™ Invasion Chamber (BD Biosciences, San Jose, CA. The cells were photographed and counted in three random microscopic fields under a 10× objective to calculate the number of cells that migrated. The graph was plotted for the number of cells that invaded per microscopic field.

The cell migration ability was assessed by a scratch wound assay. The transfected cells were cultured in 6-well plates. When the cells reached 90% confluence, a scratch wound was created using a pipette tip. The wound edges were photographed with a Nikon Eclipse TE 2000-U (Nikon, Japan), and the scratch widths were analysed using ImageJ software (NIH). Three trials were used for each condition.

The cells were lysed in RIPA buffer (150 mM NaCl, 1% NP-40, 50 mM Tris-HCl, PH 7.4, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, 1 mM deoxycholic acid and 1 mM EDTA) with protease inhibitors and phosphatase inhibitors (Calbiochem, Darmstadt, Germany). The protein concentration was determined by the Bradford protein assay kit (BioRad). The proteins were separated by SDS-PAGE and were immunoblotted and transferred to polyvinyl difluoride (PVDF) membranes (Millipore) according to standard protocols. Finally, we used the BioRad semidry transfer system to analyse the expression of the proteins, including ERRa, c-Myc, cyclin D1, Bax, and GAPDH.

The cells were collected in Trizol (Invitrogen, USA) for total RNA extraction as the manufacturer’s protocol instructed. Retrotranscription was performed with the Reverse Transcriptase M-MLV (Takara, Japan). The RT-PCR reactions were performed with a SYBR Premix Ex Taq™ kit (Takara,  Japan) on the iQ5 Real-Time PCR detection system (BioRad, Hercules, USA). The primers used were as follows: ERRa, forward: CACTATGGTGTGGCATCCTGT, reverse: CGTCTCCGCTTGGTGATCTC; IDH3A, forward: AGCCGGTCACCCATCTATGAA, reverse: CytC, forward: CAGTGCCACACCGTTGAAAA reverse: TGCATCGGTTATTTCACACTCC; cyclin D1, forward: GCTTCTGGTGAACAAGCTC, reverse: GTGGGCGTGCAGGCCAGACC; and c-Myc, forward: CAGCTGCTTAGACGCTGGATT, reverse: GTAGAAATACGGCTGCACCGA. The data were analysed using the 2^−ΔΔCT method.

A flow-based Annexin V assay was used to measure cell apoptosis after treatment with the drugs. Briefly, the cells were treated with DMSO, trametinib, simvastatin, and trametinib plus simvastatin for 24 h. We used the Annexin V, FITC Apoptosis detection kit (Dojindo Molecular Technologies, Japan) to assess cell apoptosis. The cells were washed in PBS, resuspended in 500 μl of ANX-V binding buffer and were then stained with 5 μl of Annexin-V-fluorescein isothiocyanate (FITC) for 15 min on ice in the dark, according to the manufacturer’s instructions. Subsequent to the staining, the cells were incubated with 10 μl of propidium iodide (PI) for 5 min on ice in the dark. The analyses were performed using a Navios flow cytometer (Beckman Coulter).

The drug interaction between simvastatin and trametinib was determined by the combination index (CI) value. The CI was evaluated by CompuSyn software (ComboSyn, Inc., Paramus, NJ), using the method proposed by Chou et al. [30]. CI valued < 1, =1, and > 1 indicate synergism, additive and antagonism effects, respectively.

To investigate the expression of ERRα in colon cancer tissues, we randomly selected 12 pairs of colon cancer tissue samples for western blot analysis. The results showed that the expression of ERRα was higher in the carcinomatous tissues than in the distal normal tissues (Fig. 1a). Next, we detected ERRα expression by IHC from the pathological tissues of 66 colon cancer patients who had undergone tumour resection. As indicated in Fig. 1b, the expression levels of ERRα were significantly higher in colon tumour tissues than that in distal normal tissues. Unlike in normal tissues, colon tumour tissues also showed positive staining for ERRα in the nucleus. In the normal and cancer tissues, the mean immunoreactivity scores were 0.363 and 4.867 respectively. For ERRα, most of the tumour tissues immunoreactivity scores were 4–7(moderately staining) and 8–12 (highly staining) and the percentages are 41 and 29%, respectively, while majority of the adjacent normal tissues had a score of 0–3 (lowly staining), and the percentage is 98% (Fig. 1c). Then, we also studied the effect of ERRα suppression on the malignant phenotypes of colon cancer cells. The results revealed that the cellular growth and colony information were strongly inhibited in the HCT116 and SW480 cells with shERRα#1 or shERRα#2 transfection compared with the cells transfected with the control shRNA (Figs. 1d-e, 2e-f). To verify whether shERRα performs its inhibition function properly, we constructed a luciferase assay reporter system by transfected the ERRα luciferase reporter plasmid into SW480 cells. The luciferase activity of pGMERRα-Lu significantly decreased in the cells transfected with the shERRα#1 and shERRα#2 (Fig. 1f). We also evaluated whether the inhibition of ERRα activity by a ligand modulated cell proliferation and colony formation in colon cancer cells. XCT790, a potent and specific inverse agonist of ERRα, was used in this further study. The CCK8 assay showed that XCT790 treatment dramatically inhibited colon cell growth (Fig. 1g) and colony formation (Fig. 1h-i). A western blot analysis was used to test the effect of XCT790 treatment on the protein level of ERRα. As expected, colon cancer cells treated with XCT790 showed a reduced level of ERRα compared to the vector control (Additional file 1: Figure S1a). Moreover, XCT790 treatment decreased the expression of genes encoding hyperplasia proteins, including c-Myc and cyclin D1 (Fig. 1j). Then, we also found that the colon cancer cells transfected with si-ERRα displayed less migrated cells compared to the vector control in the transwell assay and wound healing assay (Additional file 1: Figure S1d-g). Collectively, the results suggest that ERRα is involved in the regulation of proliferation and migration of colon cancer cells and plays a role as an oncogene in colon cancer.

EGFR plays a critical role in the regulation of cell proliferation and differentiation, and EGF is a crucial ligand of EGFR [31]. Here, we found that EGF up-regulated the expressions of ERRα, p-ERK and c-Myc in the HCT116, SW480 and SW1116 cell lines by western blot analysis (Fig. 2a, Additional file 2: Figure S2a). The functional studies revealed that the activated EGFR signalling also promoted cell proliferation, as demonstrated by the Cell Counting Kit-8 assay and colony formation. Further studies indicated that the inhibition of ERRα by shERRα#2, si-ERRα or XCT790 completely reversed the EGF treatment-induced cell proliferation (Fig. 2b-c, e-f, h-i, Additional file 2: Figure S2b) and the expressions of ERRα and c-Myc (Fig. 2g, j, Additional file 2: Figure S2c). In addition, qPCR analysis indicated that shERRα entirely reversed the up-regulation of ERRα, IDH3A [28] and CytC (ERRα’s downstream target) [32] induced by EGF treatment (Additional file 3: Figure S6a), and the up-regulation of luciferase activity of pGMERRα-Lu induced by EGF treatment was also significantly reversed in SW480 cells transfected with the shERRα#2 (Fig. 2d). Together, the above-mentioned data indicated that activated EGFR signalling acts by increasing ERRα to promote the proliferation and survival of colon cancer cells.

RAF-MEK-ERK (mitogen-activated protein kinase (MAPK) pathway) signalling is frequently activated in human cancers, resulting in an increase in cellular proliferation [23]. Trametinib, a specific MEK inhibitor, is clinically used in melanoma (Additional file 2: Figure S2d). Here, we found that trametinib inhibited cell growth (Fig. 3a) and decreased the expression of ERRα and its downstream target IDH3A (Fig. 3b). Further investigations indicated that trametinib partially reversed the elevated cell number and colony formation induced by EGF stimulation (Fig. 3c-e). QPCR analysis indicated that trametinib did not entirely reverse the up-regulation of ERRα, IDH3A and CytC induced by EGF treatment compared with the trametinib treatment alone (Additional file 3: Figure S6b), and western blot analysis also demonstrated that trametinib did not entirely reverse the up-regulation of ERRα, c-Myc and cyclin D1 induced by EGF treatment (Fig. 3f, Additional file 2: Figure S2e).

Therefore, we combined trametinib with shERRα, si-ERRα and XCT790 to investigate whether ERRα suppression enhances the cytotoxicity of trametinib against colon cancer. Our results showed that the combination was more effective at restraining cell proliferation (Fig. 3g, l, Additional file 2: Figure S2f) and colony formation (Fig. 3h-i, m-n). And the luciferase activity of pGMERRα-Lu was more significantly decreased in SW480 cells treated by combined trametinib and shERRα#2 (Fig. 3k). Western blot analysis also demonstrated that the combination inhibited ERRα, IDH3A, c-Myc and cyclin D1 more thoroughly compared to the single treatment (Fig. 3j, o, Additional file 2: Figure S2g-i). Furthermore, qPCR analyses also showed a substantial reduction in the ERRα and its downstream target genes IDH3A and CytC in the combination group (Additional file 3: Figure S6c). Although trametinib is an effective drug that suppresses the growth of colon cancer cells, it did not achieve adequate cytotoxicity and inhibit the over-expression of ERRα induced by EGF. This implies that the co-inhibition of ERRα and MEK achieved more efficiency.

To determine whether FDA-approved inhibitors exist that block the activity of ERRα, we performed a literature review and found that statins and bisphosphonates inhibit the activity of ERRα by blocking its cholesterol modification. Cholesterol is identified as the first functional endogenous ERRα ligand, and it increases the transcriptional activity of ERRα, while statins lower intracellular sterol levels, thus attenuating ERRα transactivation [28]. In addition, we found that simvastatin decreased the expression of its downstream target IDH3A and proliferation-related genes, such as c-Myc and cyclin D1, in the HCT116 and SW480 cell lines (Fig. 4b). The functional studies revealed that simvastatin inhibited the proliferation and colony formation of colon cancer cells (Fig. 4a, c-d, Additional file 4: Figure S3a-c). Moreover, consistent with the effect of trametinib, simvastatin partially reversed the EGF treatment-induced proliferation (Fig. 4c-d, Additional file 4: Figure S3b-c). Taken together, our results indicated that simvastatin decreased the transcriptional activity of ERRα and inhibited tumour growth in colon cancer.

The above-mentioned data indicated that simvastatin might strengthen the antitumour efficiency of trametinib by inhibiting the activity of ERRα. Further CCK8 assays indicated that simvastatin significantly enhanced the cytotoxicity of trametinib in the HCT 116 and SW480 cells (Fig. 5a). The colony formation assays revealed that simvastatin, combined with trametinib, inhibited cell survival more significantly than simvastatin or trametinib alone in the two colon cancer cell lines (Fig. 5b-c). Moreover, the flow cytometry assays showed that this combination also produced a combined activity, with regard to cell apoptosis in colon cancer cells (Fig. 5g-h). Western blot analysis demonstrated that simvastatin synergised with trametinib and dramatically reduced the expression of the IDH3A, the proliferation-related genes c-Myc and cyclin D1, and incresesd the pro-apoptotic gene Bax (Fig. 5e). Next, quantitative real-time PCR showed that the combination therapy strongly decreased the mRNA expression of ERRα and its downstream targets IDH3A, c-Myc, and cyclin D1 compared with the single drug in the HCT116 cells (Fig. 5d), and the similar results were also found in SW480 cells (Additional file 3: Figure S6d). Furthermore, the luciferase activity of pGMERRα-Lu was more potently decreased in SW480 cells when combined trametinib and simvastatin (Fig. 5f). To investigate the combined effects, we calculated the combination index (CI) values and the Fa values using CompuSyn software (ComboSyn, Inc., Paramus, NJ, USA). According to the method proposed by Chou et al., the combination index (CI) values of < 1, =1, > 1, indicate synergistic, additive and antagonistic effects, respectively[30] . The combination index (CI) values were 0.03 and 0.19 in the HCT116 and sw480 cells, respectively, indicating that the combined therapy produced a synergistic effect in the two cell lines (Additional file 5: Figure S4a).

To investigate the combined effect in vivo, we implanted HCT116 tumours in nude mice, and they were assigned to the following four groups: untreated control, trametinib, simvastatin, or a combination of trametinib and simvastatin. The combination group showed a statistically significant reduction in tumour volume and weight compared with the vehicle-treated controls or the monotherapy groups in the HCT116 xenografts (Fig. 5i-j). Next, we detected ERRα, IDH3A, c-Myc and Cyclin D1 expression by immunostaining pathological tissue sections of xenograft tumour. As indicated in Fig. 5k-l, the overall protein expression levels of ERRα, IDH3A, c-Myc and Cyclin D1 were significantly weaker in combination group. Furthermore, a western blot was preformed to investigate the expression of proliferative proteins in the lysate from the xenografts. In contrast to the monotherapy groups, a combination of trametinib and simvastatin significantly down-regulated the expressions of c-Myc and cyclin D1 (Additional file 5: Figure S4b). Altogether, our findings unveiled that trametinib, combined with simvastatin, produced synthetic lethality in vitro and in vivo.