A novel BMI-1 inhibitor QW24 for the treatment of stem-like colorectal cancer

Human colorectal cancer cell lines HCT116, HT29, HCT8, HCT15, LS174T, LoVo, mouse colon cancer cell line CT26 and normal cell lines NCM460, L02 and HAF were obtained from the American Type Culture Collection (ATCC). HUVEC were obtained from ScienCell Research Laboratories. HCT116, HT29, HCT8, HCT15, LoVo and CT26 were cultured in RPMI-1640, LS174T was maintained in DMEM, and HUVEC was cultured with ECM. All the media were obtained from Gibco. Medium was supplemented with 10% FBS, penicillin (100 units/ml) and streptomycin (100 units/ml). BALB/c nude mice and BALB/c mice were from the National Rodent Laboratory Animal Resources, Shanghai Branch of China. All animal experimental protocols were approved by the Animal Investigation Committee of the Institute of Biomedical Sciences, East China Normal University.

SRB cell proliferation assay were performed as previously described [30]. Cells were seeded in 96-well plates (3000 cells/well) and incubated with indicated concentrations of QW24 or other small molecular compounds. After incubation for indicated time, the cells were then fixed with trichloroacetic acid, stained with SRB (Sigma Aldrich, Argentina, Cat# S1402), and analyzed for percent of survival on 96-well plate reader. The IC50 (half maximal inhibitory concentration) value was calculated using GraphPad software.

Cells were lysed in cell lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate and 1% protease inhibitor cocktails) and boiled for 10 min with loading buffer (2% SDS, 10% glycerol, 10% β-Mercaptoethanol, Bromphenol Blue and Tris-HCl, pH 6.8). Lysates were fractionated on polyacrylamide gels and transferred to nitrocellulose. The blots were probed with specific antibodies followed by secondary antibody, then membranes were examined by the LI-COR Odyssey infrared imaging system (LI-COR Biotechnology, Lincoln NE). BMI-1 antibody (Catalog #ab14389) was from Abcam, H2A antibody (Catalog #12349) and Ub-H2A antibody (Catalog #8240) were from Cell Signaling Technology. β-actin antibody (Catalog #A5441) and LC3 antibody were from Sigma. The secondary antibody was conjugated with IRDye 680/800 (Catalog #926–32,221, #926–32,210, Millennium Science).

Cell colony formation assay was performed as described previously [30, 31]. Colorectal cancer cells HCT116, HT29 and HCT8 were seeded 2000 per well in 6-well plate and allowed to grow for 24 h. Then cells were treated with indicated concentrations of PTC-209 or QW24 for 7 days. Colonies were fixed with 4% paraformaldehyde for 30 min, stained with 0.1% crystal violet for 5 min. Colonies were visualized under a microscope, and all the fields were imaged and counted. Colony formation as a percentage of vehicle control for each cell line is presented.

Sphere formation assay was performed as reported previously [18, 32]. HCT116 cells were cultured in 96-well clear flat bottom ultra-low attachment microplate (catalog #3474, Corning) with serum-free medium (SFM) (DMEM/F12) supplemented with 20 ng/ml of basic fibroblast growth factor (FGF) and epidermal growth factor (EGF) (PeproTech), 5 μg/ml of insulin (Sigma–Aldrich), 0.4% bovine serum albumin (BSA, Invitrogen), and 2% B27 (Invitrogen). Culture medium was replaced or supplemented with additional growth factors twice a week.

For in vitro LDAs, QW24 pretreated HCT116 cells were dissociated into single cells and seeded in 96-well plates at indicated cell doses. For each cell dose, at least 8 wells were seeded with cells, and for the lower cell doses (10 cells or single cell per well), 12 wells were plated. After 1 week incubation, wells containing spheres were scored, and the number of positive wells was used to calculate the frequency of sphere-forming units using the ELDA software (http://​bioinf.​wehi.​edu.​au/​software/​elda/​index.​html) provided by the Walter and Eliza Hall Institute [33].

Cells were treated with the indicated concentration of QW24 for indicated time, and total RNA was extracted using TRIzol (Takara, Japan) according to the manufacturer’s instructions. 1 μg of total RNA was used for cDNA synthesis using a cDNA reverse transcription kit (Takara, Japan). Real-time PCR was performed in triplicates using gene-specific primers on a Stratagene Mx3005P PCR system (Agilent Technologies) machine. The mRNA expression levels were normalized to β-actin. The specific primer for BMI-1 and β-actin are BMI-1-F: TGGAGAACTGGAAAGTGACTCTGG; BMI-1-R: AAGAAGATTGGTGGTTACCGCTG; β-actin-F: TGAAGTGTGACGTGGACATC; β-actin-R: CATACTCCTGCTTGCTGATC. All analysis was performed using the Microsoft Excel and GraphPad Prism 6 software.

After treatment with different concentrations of QW24 or PTC-209, cells were trypsinized, washed with PBS and stained with CD44-PE (Catalog #561861, BD Biosciences) antibody and CD133-APC (Catalog #130–113–668, Miltenyi Biotec) antibody for 15 min at 4 °C in staining buffer (PBS + 0.5% BSA) in the dark. The stained cells were analyzed using BD LSRII flow cytometry (BD Biosciences).

Transwell migration assay was performed as described previously [34]. Cell migration assays were performed using a 24-well chamber containing a membrane of 8 μm pore size (Millipore). HCT116, HT29 and CT26 cells were firstly cultured in serum-free medium for 12 h. Then, 100 μl cell suspension with 4 × 10⁴ cells were added into the upper chamber, and the bottom chamber was filled with 600 μl culture medium containing 10% FBS. After incubation for 12 h at 37 °C, the cells were fixed with 4% paraformaldehyde and stained using 0.1% crystal violet. The non-invaded cells on the upper membrane surface were removed with a cotton tip. Three replicates were performed for each group and the numbers of invaded cells were counted in 6 randomly selected high-power fields under a microscope (Leica, Heidelberg, Germany).

BMI-1 construct was cloned into the lentiviral vector pTSiN, and virus production was performed as described previously [35]. 293 T cells were co-transfected with pTSiN empty vector or pTSiN-BMI-1 vector along with packing and envelop plasmids by Lipofectamine 2000 according to the manufacturer’s instructions. At 2 days post-transfection, HCT116 cells were transduced by culturing with a 1:0.5 or 1:2 mixture of fresh medium and virus supernatant with polybrene (4 μg/ml final concentration) for 72 h. Then the infected cells were treated with indicated concentration of PTC-209 or QW24.

As reported previously [36, 37], HCT116 cells (3 × 10⁶ cells/mouse) were subcutaneously injected into male BALB/c-nude mice (6–8 weeks of age). After the volume of tumor nodules reached approximately 100 mm³, the mice were randomly assigned to the indicated groups and the mice were injected intraperitoneally (i.p.) with DMSO vehicle (n = 10), PTC-209 30 mg/kg (n = 10), QW24 15 mg/kg (n = 10), QW24 30 mg/kg (n = 10) for 24 days, with the measurement of bodyweight and the tumor dimensions. The tumor volume was calculated using the following equation: tumor volume = length × width × width × 0.52 [38].

For colorectal cancer liver metastasis study, male BALB/c mice (6–8 weeks of age) were anesthetized using 150 mg/kg 2, 2, 2-tribromethanol plus 350 mg/kg tert-amyl alcohol and then 1 × 10⁶ CT26-luciferase colon cancer cells were surgically injected into the spleens as described previously [39]. One week after injection, mice were administrated with splenectomy and mice were randomized into three groups. Then mice were intraperitoneally (i.p) injected with PTC-209 30 mg/kg (n = 9), QW24 30 mg/kg (n = 10) every day for 12 days, with an equivalent volume of dimethyl sulfoxide (DMSO) was injected in vehicle group (n = 9). Hepatic metastases were monitored every 3 days using the IVIS Imaging System (Xenogen Corporation, Alameda, CA). Images and measurements of bioluminescent signals were acquired and analyzed using Living Image and Xenogen software [30]. For survival study, we used the same animal model as descripted previously in this section. The mice were treated with DMSO (n = 10), PTC-209 30 mg/kg (n = 10) and QW24 30 mg/kg (n = 10) every day for 21 days, and the survival of the mice were recorded to 14 weeks and Kaplan-Meier survival curve was analyzed.

Tumors or mice tissue specimens were excised after sacrificing mice and specimens were immediately fixed in 10% neutral buffered formaldehyde for 24 h, progressively dehydrated in solutions containing an increasing percentage of ethanol (75, 85, 95 and 100%, v/v), and embedded into paraffin blocks. Sections were cut from the paraffin blocks and IHC was carried out using anti–Ki-67 (1:250; Catalog #ab15580, Abcam), anti-PCNA (1:250; Catalog #ab18197, Abcam) and anti-BMI-1 (1250; Catalog #ab14389, Abcam) as primary antibodies. For hematoxylin-eosin (H&E) staining, samples were stained with H&E to indicate nucleus and cytoplasm, respectively.

In the process of cancer development and progression, the basic characteristic of cancer is that the tumor cells proliferate immortally [6]. Inhibiting cancer cells proliferation would be the primary standard for small molecule compounds screening. To this end, we used colorectal cancer stem-like cell lines HCT116 and HT29 [40, 41] for screening more than 500 synthetic compounds (Additional file 1: Figure S1A). HCT116 cell line was identified as cancer stem-like cells with the high expression of CSCs markers CD133 and CD44 [42, 43]. Moreover, HCT116 was identified as stem-like by immunostaining analyses for the differentiation marker KRT20 [41]. Furthermore, it was characterized that the HCT116 cell line does not express CDX1, which regulates intestine-specific gene expression and enterocyte differentiation [44]. Thus, HCT116 has no ability of differentiation and contains mainly CSCs [44]. As colorectal cancer stem-like cell, HCT116 was widely used in various studies [45‐48]. HT29 cell line was identified as cancer stem-like cells with the high expression of CSCs markers CD44 and CD24. CD44⁺ CD24⁺ cells are enriched in the HT29 cell line, which is clonogenic in vitro and can initiate tumors in vivo [44]. As colorectal cancer stem-like cell, HT29 was widely used in various studies [49‐52]. Both HCT116 and HT29 are spheroid-derived cells that have high self-renewal capacity [53].

We detected the influence of numerous compounds to colorectal cancer cells proliferation of HCT116 and HT29 by SRB assay and determined the IC50 (half maximal inhibitory concentration) of the compounds (Additional file 1: Figure S1A). The previously reported BMI-1 inhibitor PTC-209 [18] has the IC50 of 1.67 μM in HCT116 cells and 1.36 μM in HT29 cells based on our results (Additional file 1: Figure S1B). Thus, we selected the four compounds QW07, QW30, QW10 and QW24 (chemical structure shown in Additional file 1: Figure S1C and Fig. 1a) as candidates whose IC50s were less than 1 μM in HCT116 and HT29 cells (Additional file 1: Figure S1B), suggesting they had potent activity comparable to PTC-209. As reported [18], BMI-1, the primary regulator of colorectal CICs, could be the critical target for colorectal cancer therapy. Therefore, cells were treated with the candidate compounds for 12 h respectively and then we detected their effects on the protein level of BMI-1 by western blotting. Finally, we achieved the only compound QW24 that dramatically suppressed cancer cells proliferation and down-regulated BMI-1 simultaneously. By contrast, the other three compounds have no obvious effect on BMI-1 protein level (Additional file 1: Figure S1D). To verify the results further, we increased the treatment concentrations of compounds and extended the treatment time to 24 h. The results showed that QW24 decreased the protein level of BMI-1 dramatically, however, QW30 with lower IC50 had no effect even if the treatment concentration was increased to 10 μM (Additional file 1: Figure S1E). These results together confirm that QW24 potently inhibits the colorectal cancer stem-like cells proliferation and decreases the protein level of BMI-1 significantly.

To investigate the effect of QW24 on the proliferation of colorectal cancer cells, we used several human colorectal cancer cell lines, including LS174T, HCT116, HT29, HCT8, HCT15, LoVo, and mouse colon cancer cell line CT26, with SRB assay. The results showed that QW24 inhibited the proliferation of colorectal cancer cell lines dramatically with concentration dependency, and the IC50 values for all cell lines were around 1.0 μM (Fig. 1b and c). To explore the toxicity of QW24 on normal cells, the IC50 of QW24 for four normal cell lines was determined, including human normal liver cell L02, human skin fibroblast cell HAF, human normal colon epithelium cell NCM460 and human umbilical vein endothelial cell HUVEC. The IC50 values for normal cells were significant higher than cancer cells (Fig. 1c, Fig. 1d and Additional file 1: Figure S1F), indicating that QW24 has lower toxicity to normal cells. Moreover, to compare the inhibitory effects of QW24 with PTC-209 on cancer stem-like cell growth, the stem-like cells HCT116 were treated with QW24 or PTC-209 with same concentration and time. The results showed that QW24 had comparable anti-cancer activity as PTC-209, and QW24 even had more significant effect of inhibiting cancer cell proliferation than PTC-209 (Fig. 1e). The other colorectal cancer lines were also treated with PTC-209 or QW24 and the results also showed that QW24 inhibited the cell proliferation more significantly than PTC-209 (Additional file 2: Figure S2B), as well as the results of colony formation assay (Additional file 2: Figure S2A). Moreover, QW24 down-regulated BMI-1 expression more obviously than PTC-209 with the treatment concentrations of 2 μM and 4 μΜ (Fig. 1f).

We determined the protein levels of BMI-1 in different cancer cells and normal cells by western blotting (Additional file 3: Figure S3A). The expression levels of BMI-1 were lower in normal cells compared to cancer cells, which was consistent with the result that QW24 has lower toxicity to normal cells. QW24 impacts the colony formation and cell viability of colorectal cancer cells by specifically decreasing BMI-1, while normal cells with minimal BMI-1 expression are much less affected. The specificity of QW24 towards BMI-1 was demonstrated by the fact that normal cells expressing low levels of BMI-1 were nonresponsive while colorectal cancer cells expressing higher BMI-1 showed a dose-dependent decrease in cell viability upon treatment with QW24. Furthermore, the Cancer Genome Atlas (TCGA) database showed BMI-1 was significantly overexpressed in colorectal tumors compared with normal tissues in patients (Additional file 3: Figure S3B), and the cases with higher expression levels of BMI-1 showed worse survival rates than cases with lower expression levels (Additional file 3: Figure S3C).

As mentioned above, BMI-1 is one of the key components in PRC1 complex, which induces the ubiquitination of H2AK119. To further evaluate the effect of QW24 on this pathway, we selected three representative human colorectal cancer cell lines (HCT116, HT29 and LS174T) that were more sensitive (with lower IC50) to QW24 (Fig. 1c). Ub-H2A and H2A level were determined with the treatment of QW24 at 2 μM and 4 μΜ. The result showed that QW24 downregulated BMI-1 and its downstream Ub-H2A, but had no obvious influence on H2A (Fig. 1g). To further confirm the inhibitory effect of QW24 on cell growth through down-regulating BMI-1, HCT116 cells were transduced with lentivirus overexpressing BMI-1 and empty vector lentivirus as control. As shown in Fig. 1h and i, BMI-1 overexpressed cells had reduced sensitivity to the growth inhibitory effects of QW24, with the similar tendency as PTC-209. Taken together, these results indicate that QW24 significantly inhibits colorectal cancer stem-like cell growth by reducing BMI-1 protein level.

QW24 had significant effect on suppressing colorectal cancer stem-like cell growth in vitro by down-regulating BMI-1 protein. However, it was still unclear about how QW24 reduced the protein level of BMI-1. As reported [18], PTC-209 down-regulated BMI-1 by decreasing the transcription of BMI-1. Thus, we firstly determined the effect of QW24 on the transcription of BMI-1. The protein level of BMI-1 was decreased noticeably (Fig. 2a and b), whereas there was no significant change with the mRNA level of BMI-1 at same treatment condition (Fig. 2c-f). These results suggested that QW24 might decrease the stability of BMI-1 protein. Then we treated cells with cycloheximide (CHX), which blocked the protein synthesis in eukaryote cells, to detect the stability of BMI-1 when treated with QW24. As shown in Fig. 2g, compared with CHX treatment alone, the protein level of BMI-1 decreased more significantly with the co-treatment of CHX and QW24, indicating QW24 shortened the lifetime of BMI-1 protein and decreased the protein stability. However, PTC-209 had no obvious effect on the BMI-1 protein stability. Thus, QW24 has different mechanism with PTC-209 for down-regulating BMI-1 that QW24 decreases the stability of BMI-1 protein by inducing protein degradation.

The main protein degradation pathways in eukaryote cells are ubiquitin-proteasome pathway and autophagy-lysosome pathway [54]. In ubiquitin-proteasome pathway, cellular proteins that are destined for degradation are marked by the covalent attachment of multiple ubiquitin molecules, which provide a recognition signal for the proteasome [55]. In autophagy-lysosome pathway, targeted proteins are sequestered by phagophores that mature into autophagosome, and then delivered into lysosomes for degradation by autophagosome-lysosome fusion. Chloroquine, the lysosomal inhibitor, blocks autophagy mainly by impairing autophagosome fusion with lysosomes [56]. Firstly, the protease inhibitor MG132 [55] was used to detect whether BMI-1 protein was degraded by ubiquitin-proteasome pathway. The results showed that MG132 treatment did not prevent the decreasing of BMI-1 protein level, indicating the decreasing effect was not rescued (Fig. 3a). However, the treatment with lysosome inhibitor Chloroquine significantly rescued QW24 induced down-regulation of BMI-1 protein level (Fig. 3a). Therefore, we concluded that QW24 degraded BMI-1 protein by the autophagy-lysosome pathway. To validate this conclusion further, the autophagy maker LC3 was detected. The western blotting results showed that the level of LC3 II increased in dose-dependent and time-dependent manners with QW24 treatment (Fig. 3b-d), indicating QW24 induced the down-regulation of BMI-1 through autophagy-lysosome pathway. In addition, we transfected GFP-LC3 plasmid to HCT116 cells and there were autophagosome formed with QW24 treatment (Fig. 3e and f), which provided more evidence for the activation of autophagy-lysosome pathway. Overall, these results indicated that QW24 induced BMI-1 protein degradation through autophagy-lysosome pathway.

It has been reported that self-renewal could be the novel target for colorectal cancer therapy [18]. Sphere-formation assay, which recognizes and cultures stem cells, evaluates the abilities of self-renewal and differentiation for CICs in vitro on single cellular level [32]. Limiting dilution analysis (LDA) [33], which can calculate the frequency of CICs among cancer cells, is crucial for studying CICs. HCT116, the colorectal cancer stem-like cell line, was used in this study. As shown in Fig. 4a-d, QW24 suppressed the sphere-initiating cell frequency of HCT116 significantly, indicating QW24 inhibited the self-renewal ability of CICs. When cells were treated with the concentrations of 2 μM and 4 μM, the frequency of CICs decreased from 1/5 to 1/26 and 1/149 respectively (Fig. 4d). In addition, both the numbers of spheres in each well and the diameter of sphere decreased in dose-dependent manner with the treatment of QW24 (Fig. 4a, b and c). To further confirm the effect of QW24 on CICs, we used the colorectal CSCs makers CD133 and CD44 by the method of fluorescence-activated cell sorting (FACS). We found that with the treatment of QW24, the number of CICs decreased dose-dependently (Fig. 4e). Since death from colorectal cancer frequently results from tumor metastases [57] and at least 50% of patients develop metastases [58], we next explored the effect of QW24 on cancer cell migration in vitro. We firstly found that the QW24 treated cells had an obvious change of cellular morphology after 8-h treatment (Fig. 4f). Then we evaluated the effect of QW24 on colorectal cancer cells migration by transwell assay, and the result showed that QW24 inhibited the migration of colorectal cancer cells dramatically (Fig. 4g). These results together confirm that QW24 suppresses the self-renewal of colorectal CICs and cancer cell migration in vitro. Meanwhile, we determined the effect of QW24 on the proliferation of colorectal cancer cells with the same treatment condition as in transwell migration assay. HCT116, HT29 and CT26 cells were treated with QW24 at the concentrations of 0, 1, 2, 4 μM for 12 h, and the results showed that QW24 has no significant effect on cell viability at the same treatment condition as in transwell assay (Additional file 4: Figure S4A), which suggested that the reason that resulted in less cells translocating from the upper well to the down well was not due to the effect of QW24 on cell proliferation.

The above results demonstrated that QW24 had significant effects on inhibiting colorectal cancer cell proliferation, migration and self-renewal of colorectal CICs in vitro. To evaluate the therapeutic effect of QW24 on colorectal cancer in vivo, we performed the subcutaneous xenograft model in mice. The cancer stem-like cell line HCT116 was implanted subcutaneously to male BALB/c-nude mouse as described in Methods. As shown in Fig. 5a, b and c, both QW24 and PTC-209 inhibited tumor growth significantly. Notably, QW24 had a more significant tumor-inhibitory effect than PTC-209 at the same dose of 30 mg/kg/day. No obvious toxicity was noted in the animals during therapy experiments as assessed by mean body weight (Fig. 5d). Moreover, there was no obvious organ damage based on H&E staining results (Additional file 5: Figure S5A), and the mice had no abnormal behavior or side effect during treatment period. In addition, the tumor tissues were analyzed by western blotting and immunohistochemistry (IHC). Compared with the blank control group, the expression of BMI-1 in tumors was reduced in QW24 treated group (Fig. 5e). The proliferation maker Ki67 and PCNA decreased in QW24 treated group (Fig. 5f). These results confirm the inhibitory effects of QW24 on BMI-1 expression and tumor growth in vivo, and indicate anti-BMI-1 strategy could be a potential clinical therapy for colorectal cancer.

As mentioned above, the tumor metastasis is the major reason for cancer lethal. The liver is the most common site of metastasis in colorectal cancer patients because of its anatomical situation regarding its portal circulation and the incidence of liver metastasis in colorectal cancer patients is about 35% [59]. Then we would explore the effect of QW24 on colorectal cancer metastasis in vivo and the liver metastasis model was performed as described in Methods. The results showed that QW24 administration suppressed the hepatic metastases effectively, and had a preferable effect than PTC-209 in vivo (Fig. 6a, c, d and e). The number of the tumor nodules decreased dramatically within QW24 treated mice (Fig. 6c and d). No obvious toxicity was observed with the measurement of mice body weight (Fig. 6b). Moreover, we used the same liver metastases model with new mice to determine the survival rate after the i.p. injection treatment with DMSO vehicle, PTC-209, and QW24 daily for 21 days and the survival of the mice were recorded to 14 weeks. The survival of mice was significantly prolonged with QW24 treatment, compared with vehicle control and PTC-209 administration group (Fig. 6f). These results established that QW24 prevents the hepatic metastases of colorectal cancer and improves the mice survival effectively.