NF-kappa B mediated Up-regulation of CCCTC-binding factor in pediatric acute lymphoblastic leukemia

Our previous genome-wide microarray analysis of 100 Chinese pediatric ALL cases [15, 16] indicated that CTCF is up-regulated in leukemia cells (Figure 1A). To validate this finding, we performed qRT-PCR analysis of 10 paired cDNA samples (n = 20) to determine the transcriptional levels of CTCF. Each paired sample was obtained from the same patient at the time of new diagnosis (ND) and complete remission (CR). CTCF mRNA was elevated in the ND samples compared with the CR samples (Figure 1B and Table 1, fold change 2.05, p = 0.000, paired samples t-test), consistent with the bioinformatics analysis (Figure 1A).

CTCF protein levels were measured by Western blot in samples from 28 patients (n = 52), including 8 unpaired samples (n = 8, 4 ND and 4 CR), 16 ND-CR paired samples (n = 32), and 4 ND-CR-RE matched samples (n = 12) (Table 2). One bone marrow (BM) sample from a patient with immune thrombocytopenic purpura (ITP) was selected as a negative control. CTCF was uniformly expressed at high levels in the ND samples and reduced to normal levels upon CR (Figure 2A, 2B, and 2C). Given the outcome differences among patients with varying cytogenetic abnormalities, we assessed paired samples from different subtypes of ALL, including t(12;21) (TEL-AML1), t(1;19) (E2A-PBX1), t(9;22) (BCR-ABL) (data not shown), and other B-ALL (with no translocation). Similar results were observed among the different subtypes (Figure 2B), indicating that CTCF expression patterns are independent of the cytogenetic subtypes.

To investigate the expression features of CTCF in relapsed patients, samples were collected from 4 relapsed ALL patients. Interestingly, CTCF expression levels increased again after disease relapse (Figure 2C), suggesting that CTCF might be a sensitive biomarker that is predictive of relapse.

In addition to the clinical samples, we further determined the expression features of CTCF in various human lymphoblastic leukemia cell lines. Two B-lineage ALL (B-ALL) cell lines Nalm-6 and Reh, and one T-lineage ALL (T-ALL) cell line Jurkat were evaluated. The Nalm-6 cell line contains no fusion gene, whereas the Reh cell line carries the TEL-AML1 fusion gene. As shown in Figure 2D, no differences of CTCF mRNA levels were observed among three types of leukemia cells (p = 0.831, one-way ANOVA). Accordingly, CTCF protein in these cells displayed similar expression patterns, consistent with the change in clinical samples. In view of the fact that approximately 85% of pediatric ALL cases are B-ALL and nearly 80% cases are with no chromosomal and molecular genetic abnormalities [1], Nalm-6 is selected to further explore the potential oncogenic mechanism of CTCF in leukemogenesis.

The high expression of the zinc finger protein CTCF in leukemic cells prompted us to investigate whether CTCF could affect apoptosis or proliferation in lymphoblastic cells. To further determine the effect of altered CTCF activity on cell apoptosis and proliferation, two CTCF shRNA-expressing plasmids (sh-CTCF-1 and sh-CTCF-2) were constructed. The shRNA plasmid specific for firefly luciferase (sh-luc) was used as a control. The RNA interference efficiency of the shRNAs was evaluated by Western blot and parallel semi-quantitative analysis in Nalm-6 cells. Our data revealed that sh-CTCF-2 was more effective at knock-down than sh-CTCF-1 (Figure 3A); hence, sh-CTCF-2 was selected for the subsequent assays.

Cellular apoptosis was detected at 72 h after transfection of the shRNA plasmids. Because all the plasmids carried a GFP tag, we detected cell apoptosis in GFP-positive cells sorted by flow cytometry. As expected, CTCF knock-down by sh-CTCF-2 resulted in a 2-fold increase in early cell apoptosis and an approximately 15-fold increase in late cell apoptosis (Figure 3B and 3C). Consistent with these findings, CTCF knock-down by sh-CTCF-2 decreased the expression of inactive procaspase-3, and triggered the activation of cleaved-caspase-3, indicating that inhibition of CTCF leads to activation of the apoptotic pathway in leukemic cells (Figure 3D).

We next sought to determine whether CTCF knock-down has deleterious effects on cell viability. Leukemic cells treated with CTCF-specific shRNAs consistently showed at least a 40% decrease in cell viability (Figure 3E and 3F, p = 0.018, paired samples t-test). These data support the notion that CTCF activity is involved in both leukemic cell death and proliferation.

To determine whether CTCF over-expression could rescue leukemic cells from the induction of apoptosis, the over-expression plasmid pEGFP-N2-CTCF was constructed from an in-frame fusion of CTCF and an enhanced-GFP tag (Figure 4A). In this experiment, pEGFP-N2-CTCF and the empty plasmid pEGFP-N2 were transiently transfected into the leukemic cell line. Expression levels from the transfected plasmids (pEGFP-N2-CTCF and pEGFP-N2) were assessed by Western blot (Figure 4B). As shown in Figures 4C and 4D, ectopic over-expression of CTCF rescued approximately 50% of early apoptotic cells from apoptosis. Additionally, the effect of over-expressed CTCF on cell apoptosis was further determined by analyzing caspase-3 activities. However, over-expression of CTCF in transfected Nalm-6 cells resulted in a counter effect against caspase-3 activation, and no cleaved caspase-3 was detected (Figure 4B).

In the subsequent experiment, we assessed the effect of CTCF over-expression on leukemic cell proliferation. An approximately 30% increase in cell viability was observed (Figure 4E and 4F, p = 0.047, paired samples t-test). However, the changes in cell apoptosis and proliferation were not as prominent as those observed in the knock-down experiments, primarily due to increased basal levels of CTCF in leukemic cells. These results demonstrate that CTCF serves as both an anti-apoptotic and proliferative factor in leukemic cells.

The transcription factor NF-κB is involved in many key cellular processes and has emerged as a major regulator of programmed cell death (PCD) via apoptosis or necrosis [17]. Increasing evidence suggests that NF-κB plays an important role in tumorigenesis. NF-κB possesses important regulatory functions for both normal and malignant hematopoiesis and is constitutively activated in pediatric ALL samples [18]. Therefore, we investigated whether the NF-κB pathway is involved in the anti-apoptotic or proliferative effects of CTCF in leukemic cells. To test this hypothesis, the effect of NF-κB on CTCF expression was examined via inhibition of NF-κB activity with ammonium pyrrolidinedithiocarbamate (PDTC), a NF-κB specific inhibitor. The activation of the NF-κB pathway was evaluated by Western blot using a specific antibody against nuclear p65 (Figure 5A). After inhibition of NF-κB activity, CTCF mRNA levels decreased (Figure 5B, fold change 2.49, p = 0.000). CTCF protein levels were down-regulated accordingly, consistent with the change in mRNA levels (Figure 5C, fold change 1.89). These data indicate that CTCF is likely involved downstream of the NF-κB pathway in leukemic cells.

To further elucidate the regulatory role of CTCF in the NF-κB pathway, we treated Nalm-6 cells with different concentrations (5 or 10 μg/ml) of lipopolysaccharide (LPS), a potent activator of the NF-κB pathway [19]. The nuclear translocation of NF-κB p65 was enhanced with increasing LPS concentrations, indicating that the NF-κB pathway was activated effectively by LPS (Figure 6A). As shown in Figure 6B, NF-κB pathway activation with 5 or 10 μg/ml LPS resulted in a 1.33-fold (p = 0.027, paired samples t-test) and 1.66-fold (p = 0.025, paired samples t-test) increase in CTCF mRNA, respectively. Correspondingly, the protein level increased in a dose-dependent manner by 1.53-fold and 1.72-fold, respectively (Figure 6C). These data demonstrate that CTCF is regulated by NF-κB factors and may play a role downstream of the NF-κB pathway.

As previously shown, CTCF knock-down significantly increased Annexin V staining of Nalm-6 cells and decreased cell viability. To further determine whether CTCF was involved in the apoptotic or proliferative pathway mediated by NF-κB, we over-expressed CTCF in combination with treatment with PDTC or DMSO. After a 20-h pre-treatment of Nalm-6 cells with PDTC or DMSO, the over-expressing plasmids were transiently transfected; cellular apoptosis and viability were detected 48 h after transfection. Interestingly, the results demonstrated that ectopically over-expressed CTCF partially rescued the Annexin V stained cells, particularly the late apoptotic cells induced by the NF-κB-inhibitor PDTC (Figure 7A and 7B). By contrast, CTCF over-expression did not rescue the PDTC-induced proliferative inhibition of leukemic cells (Figure 7C, p = 0.070, paired samples t-test). Caspase-3 assay and the expression levels of the transfected plasmids were also assessed by Western blot (Figure 7D). These data imply that CTCF plays an important role in the anti-apoptotic pathway mediated by NF-κB factors.

A total of 28 children (7 months to 15 years, median age of 5 years) diagnosed with ALL and treated in the Hematology Oncology Center of Beijing Children’s Hospital between December 2002 and April 2009 were enrolled in this study. Informed consent was obtained from the parents or legal guardians of the patients. A single sample was obtained from a child with immune thrombocytopenic purpura (ITP) as a negative control. The study design followed the Helsinki guidelines and was approved by the Beijing Children’s Hospital Ethics Committee prior to initiating the study.

All patients were diagnosed with ALL using a combination of morphology, immunology, cytogenetics, and molecular biology (MICM). The cytogenetic ALL subtypes were experimentally identified by G-banding karyotype and multiplex nested reverse transcription-polymerase chain reaction (RT-PCR). A total of 29 fusion genes were assessed by RT-PCR, including TEL-AML1, BCR-ABL, E2A-PBX1, MLL-AF4, and SIL-TAL1.

Paired bone marrow (BM) samples from 16 pediatric patients (n = 32) were collected at the time the patient was characterized as newly diagnosed (ND) or in complete remission (CR). From these samples, 10 (n = 20) were randomly selected for quantitative real-time PCR (qRT-PCR) analysis (Table 1). In addition, 8 unpaired BM samples (n = 8) from 4 ND and 4 CR patients were collected. Matched BM samples were also collected from 4 relapsed patients at the time of ND, CR, and relapse (RE) (n = 12). The clinical features of these patients are described in detail in Table 2.

BM samples were collected in ethylenediaminetetraacetic acid (EDTA) tubes. Mononuclear cells were isolated by Ficoll gradient centrifugation (MD Pacific, Tianjin, China, density: 1.077 g/ml) and cryo-preserved in a -80°C freezer for subsequent experiments. Total RNA from the BM samples and cell lines was extracted using Trizol reagent (Invitrogen, Paisley, UK) and the mirVana™ Protein and RNA Isolation System (Ambion, USA), respectively, according to the manufacturers’ instructions. cDNA was synthesized using random hexamers and Moloney murine leukemia virus reverse transcriptase (Promega, Madison, USA). For BM samples, qRT-PCR was performed with the GenomeLab GeXP Genetic Analysis System (Beckman Coulter, CEQ8000, USA) using the GenomeLab™ GeXP Start Kit (Beckman Coulter, USA). The GAPDH gene was used as an internal control. The primer sequences were as follows: CTCF, 5′-AGGTGACACTATAGAATACAGCAGGAGGGTCTGCTATC-3′ and 5′-GTACGACTCACTATAGGGAGTGTGGCTTTTCATGTGACG-3′; GAPDH, 5′-AGGTGACACTATAGAATAAATCCCATCACCATCTTCCA-3′ and 5′-GTACGACTCACTATAGGGATTCACACCCATGACGAACAT-3′. The qRT-PCR reaction was performed with a starting temperature of 95°C for 10 min, followed by 35 cycles of 30 s at 94°C, 30 s at 55°C, and 1 min at 72°C. Each assay was repeated three times to ensure reproducibility and reliability. For cell lines, real-time PCR was done as described previously [16]. The threshold cycle (Ct) values for both CTCF and GAPDH on each PCR array were used to calculate the fold-changes in mRNA expression. The relative expression level was normalized to the GAPDH by the method of 2^-∆∆Ct.

The full-length cDNA encoding human CTCF (727 amino acids, NP_006556.1) was cloned into the pEGFP-N2 vector (EcoRI/ApaI digestion), which carries an enhanced-GFP tag; the resulting construct was named pEGFP-N2-CTCF. For the RNA interference (RNAi) experiment, the U6 promoter-driven shRNA expression vector pNeoU6 + 1 and the shRNA plasmid specific for firefly luciferase (sh-luc) were prepared by our lab facility [30]. Both plasmids contain a GFP tag. The two target sites in the CTCF mRNA coding regions were sh-CTCF-1 (658–677, ATGTAGATGTGTCTGTCTAC) and sh-CTCF-2 (953–971, TACTCGTCCTCACAAGTGC). These targeted sequences were verified as unique sequences in the human genomic and transcriptional sequence database (NCBI). The plasmids were purified using a Plasmid Mini Kit (Omega, Bio-tek, USA) in accordance with the manufacturer’s instructions.

The anti-CTCF polyclonal antibody was generated by injecting the pET28a-CTCF antigen into a rabbit. The pET28a-CTCF antigen was constructed by inserting the N-terminal region of CTCF (amino acids 1–280) into the pET28a expression vector. The antiserum had good specificity and could be used to detect human CTCF by Western blot. The anti-GAPDH antibody was produced by the animal center of our institution [16].

Samples containing 20 μg of total protein were separated on 8% ~ 12% SDS-PAGE gels according to the different molecular weight and then transferred onto nitrocellulose membranes (Whatman, Germany) in transfer buffer (25 mM Tris-base, 40 mM glycine, and 20% methanol) using a Mini Trans-Blot Cell (BIO-RAD) at 400 mA for 2 h. The membranes were blocked by incubation in 5% nonfat milk in TBS-T (20 mM Tris, 137 mM NaCl, and 0.1% Tween 20) for 1 h at room temperature. Proteins were detected using specific rabbit polyclonal anti-CTCF (1:2,000) or rabbit polyclonal anti-GAPDH (1:5,000) antibodies. After washing with TBS-T, the membranes were incubated with goat anti-rabbit immunoglobulin G secondary antibodies (1:5,000, Pierce, USA) in TBS-T containing 5% nonfat milk for 45 min at room temperature. The proteins were visualized using an enhanced chemiluminescence kit (Amersham, USA).

Nalm-6 is a pre-B ALL cell line with no fusion gene, while Reh is a pre-B ALL cell line with the TEL-AML1 fusion gene. Jurkat is a T-lineage ALL cell line. Cells were cultured in a modified HyQ RPMI-1640 medium (Hyclone, USA) supplemented with 10% fetal bovine serum (FBS, PAA, USA) in a 5% CO₂ humidified atmosphere at 37°C. In the NF-κB-inhibited drug experiments, the cells were treated with ammonium pyrrolidinedithiocarbamate (PDTC, 100 μM/ml) [31] or dimethyl sulfoxide (DMSO) for 20 h. In the NF-κB-activated drug experiments, the cells were treated with various concentrations of lipopolysaccharides (LPS, 5 or 10 μg/ml) [19] for 12 h. The cells were harvested and washed twice with PBS. The cells were incubated on ice for 30 min in 1× cell lysis buffer [20 mM Tris, 50 mM NaCl, 2 mM Na₃VO₄, 10 mM NaF, 1 mM EDTA, 0.1% Triton X-100, and Proteinase Inhibitor Cocktail (Roche)] and then sonicated. Following centrifugation at 4°C for 30 min, the supernatants were frozen at -80°C or used immediately.

Nuclear proteins (including NF-κB p65) were isolated and analyzed by Western blot. The cells were washed twice with ice-cold PBS and suspended in NE buffer A [10 mM Hepes-NaOH (pH 7.9), 1.5 mM MgCl₂, 10 mM KCl, proteinase inhibitor, 1 mM DTT, and 1 mM PMSF]. Intact nuclei were released from the cells by several washes with NE buffer B [NE Buffer A supplemented with 0.3% NP-40]. Nuclear membranes were damaged by adding NE buffer C [12.5% glycerol, 1mMTris-HCl (pH 6.5), 0.1 mM EDTA], followed by three cycles of sonication. A mouse monoclonal anti-p65 antibody (1:2,000, Santa Cruz, USA) was used to analyze the translocation of NF-κB to nuclei by standard Western blot analysis as described above. A rabbit polyclonal anti-histone H3 CT pan antibody (1:5,000, Upstate, USA) and a mouse monoclonal anti-α-tubulin antibody (1:10,000, Sigma, USA) were used as loading controls for nuclear and cytoplasmic proteins, respectively.

In the knock-down and over-expression experiments, the shRNA and over-expressing plasmids were transiently transfected into Nalm-6 cells (2 × 10⁶ seeding density) using the Amaxa Cell Line Nucleofector Kit T and the Nucleofector Device (Lonza, Swiss) according to the manufacturer’s instructions. The cells were incubated for 72 h in 2 ml of antibiotic-free media containing 10% FBS and harvested for apoptosis analysis and Western blot. A total of 1 × 10⁴ cells per well were seeded in a 96-well plate after transfection, with triplicate seedings per clone. Viable cells were counted using the Cell Counting Kit-8 (CCK-8, Dojindo, Japan) assay for 5 days according to the manufacturer’s instructions. All values were normalized to the non-treated cells (sh-luc plasmid and pEFG-N2 vector). For the drug treatment experiments, PDTC or DMSO was added to the Nalm-6 cells to inhibit NF-κB activation 20 h prior to the transfection of pEGFP-N2-CTCF or pEGFP-N2. The cells were harvested for apoptosis analysis and Western blot 48 h after transfection. Cell viability was assessed by CCK-8 analysis as described above.

For the apoptotic assays, GFP-positive cells were sorted and collected by flow cytometry (BD, FACSAria II, USA) to measure the silencing efficiency. The percentages of Annexin V-APC/PI stained (BD, USA) positive and negative cells were analyzed with FlowJo software. Besides, caspase-3 activity in cells was further determined by Western blot using specific mouse monoclonal antibody against caspase-3 (Beyotime Institute of Biotechnology, Nanjing, China), which contains specificities for detecting both procaspase-3 (1:500) and cleaved caspase-3 (1:250).

Western blots were subjected to semi-quantitative analysis using Gel-Pro Analyzer 4.0 software. The relative expression level of CTCF was normalized to the integrated optical density (IOD) of CTCF compared with GAPDH (loading control).