HOXB8 enhances the proliferation and metastasis of colorectal cancer cells by promoting EMT via STAT3 activation

Human-derived colorectal cancer cell lines HCT-116, SW480, RKO and DLD1, normal colonic epithelial cell line NCM460, and human embryonic kidney 293T were purchased from the Chinese Academy of Sciences, Shanghai, China. HCT-116 was maintained in McCoy’s 5a Medium supplemented with 10% fetal bovine serum and 100 U/ml penicillin/streptomycin (Invitrogen, Carlsbad, CA. USA). RKO cells were maintained in DMEM with the same supplements, SW480, DLD1 and NCM460 cells were maintained in RPMI 1640 medium with the same supplements. All cells were maintained in humidified 37 °C incubator with 5% CO₂. When cell confluency reached approximately 80–100%, we harvested and passaged the cells with 0.25% trypsin for the following experiments.

HOXB8 antibody was purchased from biorbyt (San Francisco, California, USA), Zeb2 antibody was purchased from proteintech (Chicago, IIlinois, USA), CD31, Twist and N-cadherin antibody were purchased from abcam (Cambridge, MA, USA), phospho-STAT3(p-STAT3), STAT3, E-cadherin, Vimentin, Ki67 and Zeb1 antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA), p-STAT3 inhibitor (S3I-201) was purchased from Slleckchem (Houston, TX, USA), TRIzol, and SYBGreen fluorescent quantitation PCR reagent were purchased from Invitrogen (Carlsbad, CA, USA), Superscript First-Strand Synthesis Kit were purchased from Promega (Madison, Wisconsin, USA). 1% crystal violet staining solution was purchased from Sangon Biotech (Shanghai, China). Bouin’s solution was purchased from Solarbio (Beijing, China), the enhanced chemiluminescent Thermo Scientific Super-Signal West Femto were purchased from Thermo Fisher Scientific (Waltham, MA USA). Color one-step rapid mycoplasma detection kit was bought from Yeasen (Shanghai, China).

Lentiviral vectors stored in our laboratory. Total RNA of RKO was extracted using TRIzol (Invitrogen). We measured its 260 nm/280 nm absorbance value on Enzyme standard instrument, calculated the concentration and selected 260 nm/280 nm absorbance value between 1.8 and 2.0 as the template to get HOXB8 full-length cDNA by RT-PCR using Superscript First-Strand Synthesis Kit (Promega) with specific designed primers (HOXB8-F-EcoRI: 5′-CGA ATT CGC CAC CAT GAG CTC TTA TTT CG TCA AC-3′, HOXB-R-SAC II: 5′-CCG CGG CTA CTT CTT GTC GCC CTT CTG-3′). The PCR products were purified and then inserted into the EcoRI/SacII sites of lentiviral vector. The inserted fragment was confirmed by sequencing. We used Lentiviral vectors expressing EGFP as a control. Two short-hairpin RNA (shRNA) were designed targeting HOXB8:shHOXB8-1 (5′-GCTCTTATTTCGTCAACTCACTGTTCTCC-3′) and shHOXB8-2 (5′-GAGCTGGAGAAGGAGTTCCTATTTAATCC-3′). Briefly, Lenti-shRNA vector construction was done as follow: The DNA fragments containing CCAA were synthesized as the loop for shRNA, then we cloned the shRNA into human pBluescript SK (+) plasmid (pU6) with U6 promoter and inserted the U6-shRNA cassettes into an appropriate lentiviral vector. Similarly, we selected a control: Lentiviral vector carrying shRNA targeting firefly luciferase (shLuc: 5′-TGC GCT GCT GGT GCC AAC CCT ATT CT-3′). The transfer vector and the other three packaging vectors (pMD2.G, pMDL-G/P-RRE and pRSV-REV) were co-transfected into 293T cells to produce the viral particles. The concentration of each vectors was added as follow: pMD2.G 3.5 mg/10 cm, pMDL-G/P-RRE 6.5 mg/10 cm, pRSV-REV 3.5 mg/10 cm and Transfer Vector 12 mg/10 cm. After 48 h the supernatant was collected and subsequently purified using ultracentrifugation to get high-quality viral particles. The first day we seeded 5 * 10⁴ cells in 24-well plates, and the next transduced with lentivirus (multiplicity of infection, MOI = 5) supplemented with 8 μg/ml of polybrene (sigma-Aldrich Chemie, The Netherlands). Transfection efficiency was verified by western blot and Real-time Polymerase Chain Reaction.

Total RNA was purified from cells using TRIzol (Invitrogen) according to the manufacturer’s instructions. Equal amounts of RNA (500 ng) were reverse-transcribed using a cDNA synthesis kit (Promega). Diluted cDNAs (1:5 final) were subjected to qPCR analysis using SYBR Green Supermix reagent (Invitrogen). The relative amount of HOXB8 expression was normalized to an endogenous housekeeping gene GAPDH. Primer sequences: HOXB8-F: 5′-TAA GCG GCG AAT CGA GGT AT-3′; HOXB8-R: 5′-TGT TTC TCC AGC TCC TCC TG-3′. GAPDH: 5′-CCA GCC GAG CCA CATCGC TC-3′ and 5′-ATG AGC CCC AGC CTT CTC CAT-3′.

Protein was extracted by using lysate. Conducted protein quantification with reference to BCA protein concentration assay kit. Took 30 μg of each sample for SDS-polyacrylamide gel electrophoresis and transferred to polyvinyl difluoride membranes (Millipore, Billerica, MA, USA). Blocked with 5% nonfat dry milk in tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h, 4 °C overnight incubation using primary antibodies, HOXB8 (1:1000), p-STAT3 (1:1000), STAT3 (1:1000), E-cadherin (1:1000), Vimentin (1:1000), N-cadherin (1:1000), Twist (1:1000), Zeb1 (1:1000), Zeb2 (1:1000) and internal-reference GAPDH (1:1000), used goat anti-rabbit or goat anti-mouse horseradish peroxidase (HRP)-conjugated secondary antibodies (Bio-world Technology, MN, USA) to hybridize at room temperature for 1 h, the membranes were visualized by ECL (Millipore, MA, USA) and finally conducted the imaging by using the gel imaging system.

The cells (DLD1-Control, DLD1-GFP, DLD1-HOXB8, SW480-Control, SW480-shLUC, SW480-shHOXB8-1, SW480-shHOXB8-2, HCT116-Control, HCT116-shLUC, HCT116-shHOXB8-1, HCT116-shHOXB8-2) grew to 80%–90% and cultured for 3 days continually, got the medium to color one-step rapid mycoplasma detection kit (Yeasen) according to the manufacturer’s instructions. If the cancer cells were infected by mycoplasma, the color will turn to navy blue as the positive control, otherwise, it will turn to blue purple as negative control. The result was showed in Additional file 1: Figure S1.

DLD1-Control, DLD1-GFP, DLD1-HOXB8, SW480-Control, SW480-shLUC, SW480-shHOXB8-1, SW480-shHOXB8-2, HCT116-Control, HCT116-shLUC, HCT116-shHOXB8-1, HCT116-shHOXB8-2 were inoculated into 96-well plates (2000 cells per well), at different time points (days 1, 2, 3, 4 and 5), the culture medium was removed and replaced with culture medium containing 10 μl of sterile MTT dye (5 mg/ml). After incubation at 37 °C for 4 h, the MTT solution was removed, and 150 μl dimethyl sulfoxide (DMSO) was added to each well followed by measuring the absorbance at 570 nm on an enzyme immunoassay analyzer (Bio-Rad).

DLD1-Control, DLD1-GFP, DLD1-HOXB8, SW480-Control, SW480-shLUC, SW480-shHOXB8-1, SW480-shHOXB8-2, HCT116-Control, HCT116-shLUC, HCT116-shHOXB8-1, HCT116-shHOXB8-2 were inoculated into 6-well plates (500 cells per well) respectively, cultured with CO₂ (5%) at 37 °C for 2 weeks, washed the plates using 1 × PBS three times and fixed them with 4% paraformaldehyde for 30 min, finally stained the cells using 0.1% of crystal violet (Sangon Biotech), and observed the changes in number and size of the cell colony.

DLD1-Control, DLD1-GFP, DLD1-HOXB8, SW480-Control, SW480-shLUC, SW480-shHOXB8-1, SW480-shHOXB8-2, HCT116-Control, HCT116-shLUC, HCT116-shHOXB8-1, HCT116-shHOXB8-2 were planted on 6-well plates and cultured as confluent monolayer. The cells were carefully scraped using a 20 μl pipette tip and debris was removed by washing with 1 × PBS. Cell migration was evaluated at 48 h, and observed the changes in size using an inverted microscope.

DLD1-Control, DLD1-GFP, DLD1-HOXB8, SW480-Control, SW480-shLUC, SW480-shHOXB8-1, SW480-shHOXB8-2, HCT116-Control, HCT116-shLUC, HCT116-shHOXB8-1, HCT116-shHOXB8-2 were seeded into upper chamber (Corning, Cambridge, MA, USA) with serum-free medium (2 * 10⁵ cells per well), the inserts were coated with 10 μl of 1 mg/ml Matrigel matrix l (BD Biosciences, Bedford, MD, USA) according to the manufacturer’s recommendations, in the lower chamber there was 800 μl of complete medium. Then chambers were incubated in the humidified 37 °C incubator (5% CO₂) for 24 h, removed the non-migratory cells in the upper chamber and fixed them with 4% paraformaldehyde for 30 min, after all, stained the migrated cells below with 0.1% of crystal violet (Sangon Biotech) for 20 min. The migrated cells were counted under an inverted microscope. Migration assays were similar to Matrigel invasion assay except that the transwell insert was not coated with Matrigel.

6-week-old female nude mice were obtained from the Shanghai Slac Laboratory Animal Co. Ltd and kept in Wenzhou medical university animal lab for 1-week accordance with the requirements of the German Animal Welfare Act (reference number: G11.559). HCT116-shHOXB8-1 and its correspond control cells (2 * 10⁶/0.2 ml) were injected subcutaneously into the right flank of 7 mice respectively. Tumor size was measured every 3 days to calculate the respective tumor volume using the following formula: volume = (width² * length)/2. Mice were sacrificed at 4 weeks. To evaluate long-distance metastasis, HCT116-shHOXB8-1 and its correspond control cells (3 * 10⁶/0.2 ml) were injected into 7 nude mice by way of tail vein to imitate long-distance metastasis. Mice were sacrificed at 7 weeks, and the metastases were fixed by Bouin’s solution (Solarbio) and confirmed by H&E staining.

Immunohistochemistry was performed using Ki-67, CD31 and HOXB8 antibodies. Tissue sections were deparaffinized in xylene and rehydrated with ethanol, then blocked with 10% normal goat serum in 1 × PBS (pH 7.5) followed with incubation with primary antibody overnight at 4 °C. Tissue sections were stained with biotinylated secondary antibody (Vector lab, Burlingame, CA) for 0.5 h at 37 °C. The peroxidase reaction was developed with diaminobenzidine (DAB kit, Vector lab) and the slides were counterstained with hematoxylin. Representative images were captured with a Leica DM4000B microscope (Jena, Germany).

SPSS17.0 (IBM, Armonk, NY, USA) and Prism 5.0 software (GraphPad Software, San Diego, CA, USA) were used in the experiment. Each set of experimental data was represented as “mean ± SD”, student’s t tests were applied to analyze the statistical significance between groups. P < 0.05 was accepted for statistical significance.

To test whether HOXB8 expression correlates with CRC, multiple microarray datasets of CRC from The Cancer Genome Atlas (TCGA) were analyzed, and we found that the HOXB8 expression level was significantly higher in CRC tissues than in normal tissues (Fig. 1a). Furthermore, to determine the expression level of HOXB8 in CRC cell lines, RT-PCT analysis was performed. Compared to a normal colorectal mucosa cell line, NCM460, HOXB8 mRNA expression was up-regulated in CRC cell lines such as SW480, HCT116, SW620, RKO and DLD-1 (Fig. 1b). In addition, the level of HOXB8 protein, evaluated by western blot, was also up-regulated in these CRC cell lines (Fig. 1c). Taken together, these results suggested that HOXB8 might be involved in CRC.

While successfully constructed the overexpression of HOXB8 in DLD1 cell (Fig. 2a), we performed MTT and colony formation assays to evaluate the effects of HOXB8 on the proliferation of CRC cells, and we also evaluate the effect of HOXB8 on the migration and invasion of CRC cells by performing wound healing and transwell assays. The results showed that overexpression of HOXB8 promoted cellular proliferation (Fig. 2b, c) and the migration and invasion of CRC cells (Fig. 2d–f). These results indicated that HOXB8 promotes the proliferation and invasion of CRC cells in vitro.

To further verify the biological function of HOXB8 in CRC cell lines, HOXB8 was silenced in the cell lines HCT116-shHOXB8-1/-2 and SW480-shHOXB8-1/-2 (Fig. 3a, b). Knockdown of HOXB8 significantly inhibited the proliferation in both cell lines (Fig. 3c–f) and also their migration and invasion (Fig. 3g–n).

Next, we aimed to explore the function of HOXB8 in tumor growth in vivo. HCT116-shHOXB8-1 and its control were subcutaneously injected into the right flanks of nude mice. Silencing of HOXB8 inhibited tumor growth (Fig. 4a–c), all these tumors were further verified by H&E staining (Fig. 4d). Immunohistochemical (IHC) analysis showed that the proportion of Ki-67-positive cells and CD31-stained vessels was lower in HOXB8-knockdown tumors than in control tumors (Fig. 4e). To further verify the function of HOXB8 in distant metastasis in vivo, HCT116-shHOXB8-1 and its control were injected into the tail veins of nude mice. Indeed, knockdown of HOXB8 decreased not only lung metastasis but also lymph node metastasis (Fig. 5a–c), these metastasis tumors were verified by H&E staining.

Based on the above results, HOXB8 markedly enhanced the motility of CRC cells especially the proliferation and invasion ability. Additionally, STAT3 is known to play an oncogenic role in many malignancies including CRC. In western blot, we found that HOXB8 overexpression decreased the expression of E-cadherin and increased the expression of Vimentin, N-cadherin, Twist, Zeb1 and Zeb2, and p-STAT3 levels were positive correlated with HOXB8 levels, despite stable expression of STAT3 (Figs. 6, 7). p-STAT3 inhibitor (S3I-201) was used to treat DLD1-HOXB8 for 12 h, the expression of p-STAT3 was reduced compared to untreated DLD1-GFP, and EMT was reversed (Fig. 8) what’s more, STAT3 inhibitor also affectted the mobility of HOXB8-overexpression cells (Additional file 2: Figure S2).