Runt-related transcription factor 1 promotes apoptosis and inhibits neuroblastoma progression in vitro and in vivo

Twenty diagnostic NB tumor samples were acquired from the Department of Pediatric Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. Written informed consent was signed by the guardians of the pediatric patients. Research protocols were approved by the Research Ethics Committees of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. The degree of tumor differentiation was determined by the department of pathology, Union Hospital, Tongji Medical College.

Generally, the tumor tissues were fixed in 10% paraformaldehyde and embedded in paraffin. The sections were dewaxed using xylene and rehydrated in graded alcohols. After antigen retrieval and blocking with goat serum, the sections were incubated overnight with the primary antibodies RUNX1 (A0400, ABclonal 1:100 dilution) and CD31 (ab28364, Abcam, 1:100 dilution) at 4 °C. Two pathologists assessed the tumor clinicopathological features.

Human MYCN-amplified (IMR-32, TIN-2) and non-MYCN-amplified (SK-N-AS, SH-SY5Y, and SK-N-SH) NB cell lines were obtained from Dr. Muxiang Zhou’s laboratory, Children’s Oncology/Hematology, Emory University, and cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and cultured at 37 °C in a humidified atmosphere with 5% CO2. Cisplatin and cytarabine was obtained from Selleck (Wuhan, China), both were dissolved in dimethyl sulfoxide (DMSO). The oligo sets for RUNX1 was inserted into CV186 (GeneChem Co., Ltd., Shanghai, China) and shRNAs were inserted into GV298 (Additional file 2: Table S3). Stable cells were screened by neomycin or puromycin (Invitrogen).

Cell viability under different concentrations was determined using the Cell Counting Kit-8 (Dojindo, Tokyo, Japan) according to manufacturer’s guideline. After treatment, the cells were cultured in 96-well plates. At 12, 24, 36, 48, 60, 72, 84 and 96 h, CCK8 solution (DOJINDO, Japan) was added to the 96-well plate, and cells were further incubated at 37 °C for 2 h. Absorbance was measured at 450 nm. The oligo sets primer sequences see Additional file 2: Table S1.

Apoptosis was measured by using an FITC Annexin V/propidium iodide (PI) using Apoptosis Detection Kit (BD Pharmingen, USA). Tumor cells were collected, washed twice with cold phosphate-buffered saline (PBS), resuspended with 200 μl of binding buffer at 2–5 × 10⁵ cells/ml density and stained with 2.5 μL of Annexin V-FITC and 5 μL of PI incubated at room temperature for 15 min in the dark. Apoptosis was analyzed by a flow cytometer (Becton Dickson Co.) at a wavelength of 488 nm.

Total RNA was isolated with TRIzol reagent (Invitrogen, USA), and cDNA was synthesized using a PrimeScriptTM RT Reagent Kit (Perfect Real Time, TaKaRa, Japan). Real-time PCR was performed with SYBR Green PCR Master Mix (SYBR 1 Premix Ex TaqTM II, TaKaRa, Japan) and the primer indicated in the Supplementary map. After normalization to the β-actin gene, the transcript levels of each target gene were determined according to the 2^−ΔΔCt method. The PCR and ChIP primer sequences see Additional file 2: Table S2.

Tissue or cellular protein was extracted with 1 × cell radio immunoprecipitation assay (RIPA) lysis buffer (Promega) following the provided protocol. Equal amounts of total protein were loaded onto a 12% SDS-PAGE gel, transferred onto polyvinylidene difluoride membranes (Millipore), blocked with 5% fat-free milk for 2 h at room temperature, and incubated with primary antibodies followed by horseradish peroxidase-conjugated secondary antibodies. Specific proteins were visualized with enhanced chemiluminescence detection reagent according to the manufacturer’s instructions (Pierce Biotechnology).

Cells (6× 10³) were mixed with 0.3% Nobel agar (Fisher Scientific, Pittsburgh, PA) and layered onto 0.6% solidified agar in DMEM/F12 containing 10% FBS in 6-well plates. After incubation 3 weeks, cells were fixed with 100% methanol and stained with 0.5% crystal violet dye.

Matrigel invasion assays were performed with polycarbonate filter membranes with a diameter of 6.5 mm and pore size of 8 μm (Invitrogen) using membranes coated with Matrigel matrix (BDScience, Sparks, MD). Tumor cell suspensions (1 × 10⁵ cells/well) in serum-free culture medium were added to the upper chambers and incubated for 24 h at 37 °C. The filters were fixed with methanol for 30 min and stained with 0.1% crystal violet for 20 min at room temperature.

A total of 150 μl of Matrigel (BD Biosciences) solution per well was added to a 48-well plate and a gel was formed at 37 °C for 1 h. Human umbilical vein endothelial cells (HUVECs) were serum-starved by culturing in serum-free culture medium for 24 h, suspended at a density of 9 × 10³ cells/well and added to the Matrigel-coated wells.

The mitochondrial membrane potential was determined using tetraethylbenzimidazolylcarbocyanine iodide (JC-1) dye (KeyGEN BioTECH, China). Tumor cells were cultured in a confocal dish, and then, JC-1 was stained with 1 × incubation buffer for 20 min at 37 °C in the dark. Next, cells were washed twice with PBS, and 1 × incubation buffer was added. Fluorescence emission was observed under a confocal microscope; green fluorescence was observed using excitation and emission wavelengths of 498 and 553 nm, respectively, and red fluorescence was observed using excitation and emission wavelengths of 570 and 600 nm, respectively.

Human BIRC5, CSF2RB and NFKBIA promoters and their truncated forms were amplified from genomic DNA by PCR and subcloned into pGL3-Basic (Promega). We used a luminometer to measure luciferase activity.

A ChIP assay was performed according to the manufacturer’s instructions for an EZ-ChIP kit (Beyotime, China). Real-time qPCR was performed as previously described.

Female BALB/c nude mice were purchased Beijing Vital River Laboratory Animal Technology (Beijing, China), and experimental protocols were approved by the Animal Care Committee of Tongji Medical. For in vivo tumor growth, 1-month-old male nude mice (n = 5 per group) were injected subcutaneously into the right upper back with 1 × 10⁷ tumor cells stably transfected with the indicated vectors. Thirty days later, the mice were sacrificed and examined as described above. The experimental metastasis studies were performed by tail vein injection of tumor cells (2 × 10⁶ per mouse, n = 5 per group) into nude mice as previously described. Forty-five-day-old mice were sacrificed and examined as described above.

All experiments were performed at least three times, and the data are indicated as the mean ± SEM or median. The cutoff values were calculated by the average expression of the gene. Unpaired t-test were used to compare two groups, and multiple group comparisons were analyzed by variance (*P < 0.05, **P < 0.01, and ***P < 0.001 were regarded as significant differences).

To determine the crucial role of transcription factors in apoptosis, we analyzed the mining of public microarray datasets of KEGG [11] and ChIP-X [12] (Fig. 1a). Based on the number of target genes, we found HNF4A and RUNX1 as a potential transcription factors that might regulate apoptosis. In order to explore the significance of HNF4A and RUNX1 in NB, Kaplan–Meier curves of 498 patients (GSE49710) showed that high HNF4A and RUNX1 levels were correlated with favorable overall survival in NB patients (p = 0.053 and 0.019, respectively, Fig. 1b). Based on the significant of survival, we chose RUNX1 for the further examination. In addition, mining of public datasets (GSE49710) showed that RUNX1 levels were negatively correlated with MYCN amplification (P < 0.0001), progression (P < 0.0001), and advanced stage (P = 0.0014, Fig. 1c) by International Neuroblastoma Staging System (INSS). Notably, in public datasets (GSE49710), there was no correlation between RUNX1 and MYCN (R = − 0.071, P = 0.13, Supplementary Fig. S1a), which indicate RUNX1, not HNF4A, might serve as an independent prognostic factor for favorable outcome of NB. Moreover, we measured the level of RUNX1 in 5 ganglioneuroma (GN, as control) tissues, 5 well differentiated (WD) NB tissues, 5 poor differentiated (PD) NB tissues, and 5 undifferentiated (UD) NB tissues. As shown in Fig. 1d and Fig. 1e, western blot analyses and immunohistochemical staining revealed that RUNX1 was highly expressed in GN tumor and WD tumor tissues relative to the PD and UD NB tissues. Furthermore, RUNX1 expression was observed in NB cell lines, SH-SY5Y, SK-N-SH, SK-N-AS, IMR-32 and TN-2 (Fig. 1f). Collectively, these data demonstrate that RUNX1 is an independent prognostic marker for NB apoptosis and progression.

To explore the function of RUNX1 in NB, we further investigated the effects of overexpression or knockdown of RUNX1 on cell proliferation, migration, invasion and tumorigenesis in NB cell lines (SH-SY5Y and SK-N-SH). Stable transfection of RUNX1 led to its overexpression in SH-SY5Y and SK-N-SH, while two independent short hairpin RNAs (shRNAs), sh-RUNX1#1 and sh-RUNX1#2, were used to deplete RUNX1 in the SH-SY5Y and SK-N-SH cell lines (Fig. 2a). The subsequent finding from CCK-8 (Fig. 2b), soft agar (Fig. 2c) and matrigel invasion (Fig. 2d) revealed that SH-SY5Y and SK-N-SH cells transfected with RUNX1 showed a decreased in cell growth, viability, invasion and migration. However, silencing of RUNX1 had opposite results with the aforementioned factors. Next, tube formation assays indicated that overexpression or silencing of RUNX1 respectively decreased and facilitated tube formation of endothelial cells, than those transfected by mock or scramble shRNA. (Fig. 2e). Taken together, these data show that RUNX1 plays a major role in regulating cell growth, proliferation, aggressiveness and tumorigenesis in NB cells.

To test the potential predictive role of RUNX1 in NB therapy, we first examined the direct effect of RUNX1 on NB cell apoptosis. Flow cytometric assays (Fig. 3a) and JC1 staining (Fig. 3b) were used to determine the apoptosis of NB cells that were stably transfected with RUNX1, sh-RUNX1#1, or sh-RUNX1#2. The number of apoptotic cells was significantly increased in NB cells with stable overexpression of RUNX1 compared with cells mock transfected; conversely, RUNX1 knockdown decreased apoptosis compared with sh-Scb transfected. We further examined apoptosis-related cell signaling events by real-time quantitative (Fig. 3c) and Western blot analyses (Fig. 3e). The results indicated that the overexpression of RUNX1 significantly decreased the levels of Bcl-2 but increased the transcription and expression levels of Bax and cleaved caspase-3 in both SH-SY5Y and SK-N-SH cells. The ratio of the Bcl-2/Bax was decreased and increased by expression or knockdown of RUNX1 in NB cells. Analysis of the ratio of the Bcl-2/Bax show that RUNX1 induced proapoptotic potential (Fig. 3d and Fig. 3f), and efficiently activated the downstream caspase-3 and apoptosis. Thus, these results revealed that the overexpression of RUNX1 promoted apoptosis in SH-SY5Y and SK-N-SH cells.

To determine how RUNX1 transcription factor regulates apoptosis, we first analyzed the publicly available datasets of 498 NB cases (GSE49710) and 649 NB cases (GSE45547) derived from the Gene Expression omnibus (GEO) which correlated with RUNX1 and the apoptosis gene of RUNX1 derived from the KEGG [13] datasets and ChIP-X (Fig. 4a). Based on the three overlapping analyses, we focused on 7 genes (BIRC5, TUBA1A, FOS, NFKBIA, TNFSF10, CSF2RB, PARP4) regulated by RUNX1 that were associated with apoptosis and analyzed the expression of these 7 genes in 498 patients (GSE49710, Fig. 4b). We found that the correlation between RUNX1 and TUBA1A was inconsistent in the two databases, so it was removed from further analyses (Supplementary Fig. S2a-g). Next, we examined the mRNA and protein levels of BIRC5, CSF2RB, NFKBIA, FOS, PARP4 and TNFSF10 following stable transfection with RUNX1, sh-RUNX1#1 or sh-RUNX1#2 in NB cells (Fig. 4 c and d). The results indicated that stable overexpression of RUNX1, increased the expression of CSF2RB and NFKBIA, decreased the expression of BIRC5. However, stable knockdown of RUNX1 had opposite results with the aforementioned factors. Meanwhile, the expression of FOS, PARP4 and TNFSF10 was not affected by ectopic overexpression or knockdown of RUNX1. These findings indicated that transcription factor RUNX1 facilitated the apoptosis in NB through regulating the expression of CSF2RB, NFKBIA and BIRC5.

To investigate the direct effects of RUNX1 on BIRC5, NFKBIA and CSF2RB transcription and expression in NB cell lines, we used JASPAR (http://​jaspar.​genereg.​net) to first identify the binding profile of the transcription factor RUNX1 [14, 15] in vertebrates (Fig. 5a). Subsequently, we used performed chromatin immunoprecipitation (ChIP) and quantitative PCR (qPCR) assays for promoter-binding as well as a dual-luciferase assay for promoter-activity assays. ChIP and qPCR assays indicated that overexpression of RUNX1 facilitated the binding of BIRC5, NFKBIA and CSF2RB promoter regions in NB cells; conversely, stable knockdown of RUNX1 decreased the binding of BIRC5, NFKBIA and CSF2RB promoter regions (Fig. 5b). Dual-luciferase assays revealed that the activity of BIRC5 was significantly decreased in NB cells with stable overexpression and increased by silencing of RUNX1, whereas NFKBIA and CSF2RB showed opposite results (Fig. 5c). To further verify the above results, we analyzed the survival of BIRC5, NFKBIA and CSF2RB expressing patients. Kaplan-Meier curves of 498 (GSE49710) NB patients revealed a difference in patient survival between the high and low BIRC5, NFKBIA and CSF2RB expression groups (Fig. 5d). Patients with low BIRC5, high NFKBIA and CSF2RB expression had higher survival probability. Collectively, these results indicated that RUNX1 directly interacted with the binding site within the BIRC5, NFKBIA and CSF2RB promoters to alter their transcription and protein levels in NB cells.

We next investigated the effects of RUNX1 overexpression on tumor growth, metastasis, and angiogenesis and apoptosis in vivo. Consistent with previous findings, the growth rate as judged by the in vivo fluorescence of the fluorescent SH-SY5Y xenografted tumors as well as their final tumor-weights were significantly decreased when stably overexpressing RUNX1. Consistently suppression of RUNX1 expression by stable transfection of sh-RUNX1#1 resulted in a significant increase in the growth and weight of xenograft tumors (Fig. 6 a and b). Stable transfection of RUNX1 or sh-RUNX1#1 into SH-SY5Y cells resulted in a significant decrease or increase in Ki-67 proliferation index and CD31-positive microvessels levels of subcutaneous.

xenograft tumors in nude mice (Fig. 6c). Western blot assays (Fig. 6d) indicated that stable transfection of RUNX1 into SH-SY5Y cells resulted in a significant increase in the levels of CSF2RB, NFKBIA, Bax and cleaved caspase-3 and decrease in the levels of BIRC5 and Bcl-2 of subcutaneous xenograft tumors in nude mice. However, silencing of RUNX1 had opposite results with the aforementioned factors. The ratio of the Bcl-2/Bax was decreased and increased by expression or knockdown of RUNX1 in vivo (n = 5, per group). In addition, in vivo metastasis assays involving athymic nude mice infused systemically (via tail vein) with RUNX1-overexpressing fluorescent SH-SY5Y cells resulted in a significant decreased in lung metastatic colonies, while the RUNX1-knocked down cells increased these lung metastases (Fig. 6e). Taken together, these results indicated that RUNX1 suppresses NB cell growth, metastasis, and angiogenesis and promotes apoptosis in vivo.

It has been reported that RUNX1 can be upregulated in blood tumors by cytotoxic drugs [16]. We hypothesized that RUNX1 could be upregulated by cytotoxic drugs in NB. The CCK-8 assay showed that different concentrations of cytarabine and cisplatin affected the growth of SH-SY5Y cells after 24 h and IC₅₀ values 2 μM and 15 μM, respectively (Fig. 7a). Meanwhile, both cytarabine and cisplatin treatment of SH-SY5Y cells in a time-dependent manner (Fig. 7b). Treatment with cytarabine and cisplatin (2 μM and 15 μM, respectively) significantly increased the level of RUNX1 in time-dependent manner in SH-SY5Y cells. RUNX1 was obviously increased with cisplatin treatment compared with cytarabine treatment in SH-SY5Y cells (Fig. 7c). To further analyze whether cisplatin regulates the apoptosis via RUNX1, we performed the rescue experiments. Western blot (Fig. 7d), flow cytometry assays (Fig. 7e) and JC-1 (Fig. 7f) staining assays showed that the treatment of SH-SY5Y cells with cisplatin resulted in increased the ratio of Bcl-2/Bax and cleaved caspase-3, the late and early apoptosis and decreased mitochondrial depolarization, respectively compared with control group (DMSO + sh-Scb). Meanwhile, knockdown of RUNX1 abolished the change in apoptosis treated with cisplatin. Collectively, our data show that cisplatin could promote apoptosis through promoting RUNX1 expression in NB cells.