c-FLIP is involved in tumor progression of peripheral T-cell lymphoma and targeted by histone deacetylase inhibitors

Compared with reactive hyperplasia, long and short isoform of c-FLIP gene (c-FLIP _L and c-FLIP _S ) were overexpressed in patients with PTCLs and T-cell acute lymphoblastic leukemia (T-ALL) (P all <0.001, Figure 1A), in agreement with significant downregulation of extrinsic apoptosis-inducing signaling ligand TRAIL and its receptor DR5 (P all <0.001, Figure 1B). Therefore, c-FLIP was potentially an indicator of defective extrinsic apoptosis in PTCLs.

Considering that c-FLIP _L was the main isoform expressed in PTCLs and did not vary from histological subtypes (Additional file 1: Figure S1), the relation of c-FLIP _L with clinical and biological parameters was studied. The median expression of c-FLIP _L in PTCLs was 70.06. The patients with c-FLIP expression level over and equal to the median value were regarded as high c-FLIP expression, whereas those below the median value were included in the low c-FLIP expression. Clinically, high c-FLIP expression was significantly associated with elevated serum lactate dehydrogenase (LDH) level and International Prognostic Index (IPI) indicating intermediate-high and high-risk (P = 0.036 and P = 0.010, respectively, Table 1).

To better define the biological function of c-FLIP in PTCLs, Jurkat and H9 cells were transfected with specific c-FLIP small-interfering RNA (siRNA). The effect of c-FLIP siRNA on c-FLIP expression was confirmed by western blot (Figure 2A). Comparing with the control siRNA (Con siRNA), c-FLIP siRNA resulted in remarkable induction of tumor cell apoptosis (Figure 2A, P = 0.014 and P = 0.005, respectively), as well as increase of TRAIL and DR5 expression (Representative results shown in Figure 2B). Moreover, effects of treatment of both cells with chemotherapeutic agents such as doxorubicin, cyclophosphamide and cisplatin that are regularly applied to treat lymphoma, were analyzed by MTT assay and the concentration of the drug requires for 50% growth inhibition (IC50) was determined. The c-FLIP siRNA transfected cells were more sensitive to these agents than those transfected with the Con siRNA (P = 0.005, P = 0.015, P = 0.012 in Jurkat cells and P = 0.002, P = 0.049, P = 0.012 in H9 cells, respectively, Figure 2C), with corresponding increase of tumor cell apoptosis in c-FLIP siRNA group (P = 0.021, P = 0.014, P = 0.008 in Jurkat cells and P = 0.012, P = 0.005, P = 0.020 in H9 cells, respectively, Figure 2D). These data indicated that c-FLIP conferred tumor cell resistance to chemotherapy.

To identify the possible role of bio-therapeutic agents HDACIs on c-FLIP, Jurkat and H9 cells were cultured with valproic acid (VPA) and suberoylanilide hydroxamic acid (SAHA). As shown in Figure 3A, HDACIs exerted substantial growth inhibition in both cells, which displayed characteristic morphological changes of apoptosis, such as shrinking cytoplasm, condensed chromatin and nuclear fragmentation with intact cell membrane (Figure 3B). Meanwhile, the percentage of Annexin V+/PI+ cells was significantly increased in HDACIs-treated cells (VPA, P = 0.002 and P = 0.023, SAHA, P = 0.013 and P = 0.003, respectively, Figure 3C).

HDACIs induced cleavage of Caspase-8, as well as decrease of c-FLIP expression both at the protein level (Figure 3D) and at the transcriptional level (Additional file 2: Figure S2A, VPA, P = 0.004 and P = 0.008, SAHA, P < 0.001 and P = 0.008, respectively). Meanwhile, significantly increased protein (Figure 3E) and gene expression of TRAIL and DR5 (Additional file 2: Figure S2B, TRAIL, VPA, P = 0.004 and P = 0.002, SAHA, P = 0.002 and P = 0.005; DR5, VPA, P = 0.001 and P = 0.012, SAHA, P = 0.001 and P = 0.009, respectively) were also observed upon HDACIs treatment, indicating that HDACI-induced extrinsic apoptosis was associated with downregulation of c-FLIP. In parallel to c-FLIP overexpression, T-lymphoma cells were less sensitive to VPA and SAHA, with decreased percentage of Annexin V+/PI+ cells (VPA, P = 0.031, SAHA, P = 0.018, respectively) (Figure 3F).

c-FLIP is a target gene of NF-κB [14],[15]. Extrinsic apoptosis-related genes, main NF-κB members and NF-κB target genes were assessed using Human NF-κB Signaling Pathway Plus PCR Array, before and after VPA or SAHA treatment (Figure 4A). In according with increased expression of apoptotic genes, NF-κB members (P65 and P50) and NF-κB target gene c-FLIP expression were decreased when treated with HDACIs. As revealed by western blot, nuclear P65 and P50 expression were downregulated, along with decrease of phosphorylated IKKα/β (p-IKKα/β) and phosphorylated IκBα (p-IκBα) expression in HDACI-treated cells (Figure 4B).

NF-κB member P50 is one of the regulators of c-FLIP [16]. As detected by immunofluorescence assay, VPA and SAHA blocked TNF-α-induced nuclear translocation of P50 in Jurkat and H9 cells (Figure 4C). Therefore, HDACIs could inhibit both intrinsic and extrinsic activation of P50. Chromatin immunoprecipitation (CHIP) assay was subsequently performed to explore the interaction of P50 with c-FLIP promoter region. The results showed that P50 was able to bind with the c-FLIP promoter and the binding activity declined after HDACIs treatment, as measured by PCR (Figure 4D). As shown by luciferase reporter assay, p50 activated the transcriptional activity of the c-FLIP promoter (P = 0.022, Figure 4E). To enforce the role of p50, Jurkat cells were transfected with specific P50 siRNA. The results showed that molecular silencing of p50 reduced the c-FLIP level (Figure 4F) and T lymphoma cells were less sensitive to apoptosis (P = 0.021, Figure 4F). Together, HDACI-induced c-FLIP downregulation was related to inactivation of NF-κB members, particularly P50.

As Class I HDACIs, VPA inhibited HDAC1 and HDAC3, whereas SAHA inhibited HDAC1, HDAC2 and HDAC3 expression (Figure 5A). Tissue array (Figure 5B) showed that HDAC1 was more frequently observed in T-cell lymphomas than in B-cell lymphomas (83.3% vs 38.9%, P = 0.016) and correlated with c-FLIP expression (r = 0.795). This was also present in tumor samples of patients with PTCLs (P = 0.035, Figure 5C). No significant difference of other members of Class I HDACIs (HDAC2, HDAC3, and HDAC8) was observed between T-cell lymphomas and B-cell lymphomas (data not shown).

Considering that HDAC1 was the main HDAC member involved in T-cell lymphomas, further study was focused on HDAC1 regulation by VPA and SAHA. In Jurkat and H9 cells, HDACIs-mediated HDAC1 downregulation was accompanied by reduced enzymatic activity of HDAC1 (VPA, P = 0.037 and P = 0.029, SAHA, P = 0.047 and P = 0.032, respectively, Figure 5D). Moreover, Jurkat cells were transfected with specific HDAC1 siRNA. Molecular silencing of HDAC1 significantly diminished HDACIs-mediated inhibition of tumor cell growth, consistent with decrease of P50 and c-FLIP expression (Figure 5E and 5F).

The in vivo anti-tumor activity of HDACIs was further evaluated in a murine xenograft model of T-cell lymphoma. Subcutaneous inoculation of Jurkat cells into nude mice resulted in a tumor formation at the site of injection in all mice. The size of tumors formed in mice treated with oral VPA was significantly smaller than those of the untreated group since 12 days of treatment (Day 12, P = 0.049, Day 14, P = 0.012, respectively, Figure 6A).

To search for more evidence of tumor cell apoptosis, terminal deoxytransferase-catalyzed DNA nick-end labeling (TUNEL) assay was performed on mice tumor sections. Compared with the Control group, the number of the apoptotic tumor cells was significantly increased in the VPA group (P = 0.023, Figure 6B). In according with in vitro results, P50 (Figure 6C), HDAC1 and c-FLIP expression (Figure 6D) were decreased, while TRAIL and DR5 expression (Figure 6D) were increased upon VPA treatment.

Sixty-one patients diagnosed as PTCLs were enrolled in this study, including PTCL-not otherwise specified (33 cases), angioimmunoblastic T-cell lymphoma (16 cases), and anaplastic large-cell lymphoma (12 cases). Histologic diagnoses were established according to the World Health Organization classification [32]. Induction chemotherapy consisted of 6 to 8 cycles of CHOP or CHOP-like regimen. The clinical and biological features of the patients were listed in Table 1. Twenty cases with T-ALL and thirty cases with reactive hyperplasia were referred as controls. The study was approved by the Shanghai Rui Jin Hospital Review Board with informed consent obtained from all patients in accordance with the Declaration of Helsinki.

T lymphoma cell lines (Jurkat and H9) and HEK-293 T cells were obtained from American Type Culture Collection (Manassas, VA, USA). Cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated fetal bovine serum in humidified atmosphere of 95% air and 5% CO₂ at 37°C. VPA was from Sigma-Aldrich (St Louis, MO, USA). SAHA was from Merck & Co (Darmstadt, Germany). Antibodies against Caspase-8, p-IKKα/β, p-IκBα and HDAC3 were purchased from Cell Signaling (Beverly, MA, USA). Antibodies against c-FLIP, P65, P50, IKKα/β, IκBα, HDAC1, HDAC2, HDAC8 and Lamin B were from Abcam (Cambridge, UK). Anti-β-actin antibody was from Sigma-Aldrich. Horseradish peroxidase-conjugated goat anti-mouse IgG and goat anti-rabbit IgG were from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Cell proliferation was assessed by MTT assay and the absorbance was measured at 490 nm by spectrophotometry. Cell morphology was evaluated by Wright’s staining under light microscopy.

Cell apoptosis was assessed using Annexin V-FITC Apoptosis Kits (Becton Dickinson, Franklin Lakes, NJ, USA) according to the manufacturer’s instructions. Expression levels of TRAIL and DR5 were quantified using antibodies against TRAIL (Cell signaling) and DR5 (Abnova, Walnut, CA, USA) as the primary antibodies, and DyLight 405 labeled anti-rabbit antibody (KPL, Washington DC, USA) as the secondary antibody. The median fluorescent intensity (MFI) was measured by flow cytometry.

Total RNA was extracted from frozen tissue of PTCLs and reactive hyperplasia, as well as bone marrow blasts of T-ALL, using TRIzol reagent (Invitrogen). cDNA was synthesized using PrimeScript RT reagent Kits with gDNA Eraser (TaKaRa, CA, USA) following the manufacturer’s instructions. Real-time polymerase chain reaction (PCR) was performed using SYBR Premix Ex Taq™ II (TaKaRa) on ABI Prism 7500 (Applied Biosystems, Bedford, MA, USA). The relative gene expression levels were calculated using the SDS2.4 software. The primer sequences were as follows: c-FLIP _L forward, 5′-ATTGCATTGGCAATGAGACAGAGC-3′; reverse, 5′-TCGGTGCTCGGGCATACAGG-3′, c-FLIP _S forward, 5′-ACCCTCACCTTGTTTCGGACTAT-3′; reverse, 5′-TGAGGACACATCAGATTTATCCAAA-3′, TRAIL forward, 5′-TCAGCACTTCAGGATGATGG-3′; reverse, 5′-CACCAGCTGTTTGGTTCTCA-3′, DR5 forward, 5′-TGACGGGGAAGAGGAACTGA-3′; reverse, 5′-GGCTTTGACCATTTGGATTTGA-3′, GAPDH forward, 5′-GAAGGTGAAGGTCGGAGTC-3′; reverse, 5′-GAAGATGGTGATGGGATTTC-3′.

Real-time PCR of NF-κB signaling pathway was performed using the RT² profiler PCR Array-Human NF-κB signaling pathway (QIAGEN Sciences, Frederick, MD, USA). GAPDH was used as the endogenous control and Jurkat cells for calibration. A relative quantification was calculated using the ^ΔΔCT method.

A human lymphoma tissue array was purchased from US Biomax, Inc. (Rockville, MD, USA) containing 12 T-cell lymphomas and 18 B-cell lymphomas. Levels of protein expression were graded according to staining intensity (SI) and distribution using the immunoreactive score (IRS). IRS = SI × PP (percentage of positive cells). SI was defined as 0 = negative; 1 = weak; 2 = moderate; and 3 = strong. PP was scored as 1, <25%; 2, 25-50%; 3, 50-75%; and 4, 75-100% positive cells. IRS > 4 was determined as positive.

Immunohistochemistry was performed on 5 μm-paraffin sections with an indirect immunoperoxidase method using the primary antibody against c-FLIP (1:200), TRAIL (1:800), DR5 (1:200), HDAC1 (1:100), HDAC2 (1:100), HDAC3 (1:100) and HDAC8 (1:100). Immunofluorescence assay was performed on methanol-fixed cells using P50 as the primary antibody and diaminotriazinylaminofluorescein-labeled donkey anti-rabbit-IgG antibody (Abcam) as the second antibody.

Jurkat cells were transfected with c-FLIP, P50 and HDAC1 siGENOME SMARTpool or Non-Targetingpool (Dharmacon) as the negative control using DharmaFECT2 transfection reagent (Dharmacon) following the manufacturer’s instruction. As for overexpression assay, Jurkat cells were transfected with c-FLIP-overexpressing vector or a control vector, electroporated at 250 V 25 ms in 4-mm cuvettes using a BTX ECM 830 and replated in fresh medium for further experiments.

Cells were suspended in 400 μL lysis buffer (10 mM HEPES, 10 mM KCl, 1.5 mM MgCl₂, 0.5 mM DTT, pH 7.9) with 0.2% Nonidet P-40 and protease inhibitor cocktail for 1 min on ice. After centrifuged for 1 min at 2500 × g, the supernatants were collected as cytoplasmic protein extracts. The pellets were washed with lysis buffer without Nonidet P-40, then re-suspended in 150 μL extraction buffer (20 mM HEPES, pH 7.9, 420 mM NaCl, 0.5 mM DTT, 0.2 mM EDTA and 25% glycerol), and incubated for 20 min on ice. After centrifuged at 12000 × g for 10 min, the supernatants were collected as nuclear protein extracts.

Cells were collected and lysed in 200 μL lysis buffer (Sigma Aldrich). Protein lysates (20 μg) were electrophoresed on 10% sodium dodecyl sulfate polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked with 5% non-fat dried milk and incubated overnight at 4°C with appropriate primary antibody, followed by horseradish peroxidase-linked secondary antibody. The immunocomplexes were visualized using chemiluminescence phototope-horseradish peroxidase Kits. Lamin B and β-actin were used to ensure equivalent loading of nuclear and whole cell protein, respectively.

CHIP assay was performed using EZ-ChIP Kits (Millipore, Billerica, MA, USA) to identify the interaction between DNA and protein following the manufacturer’s instructions. Antibody against RNA Polymerase II was referred as the positive control and non-specific Ig and PCR with primers amplifying the distal region of the c-FLIP promoter as the negative control (forward, 5′-CCCGGGTTCAAGCAATTCTC-3′; reverse, 5′-GGATCACGAGGTCAGGAGTT-3′).

HEK-293T cells were transfected with luciferase reporter and P50 overexpressing vector, using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Protein was collected 24 h after transfection, using the Passive Lysis Buffer (30 μL per well) provided as part of the Dual-Luciferase Reporter Assay System kit (Promega). Firefly and Renilla luciferase activities were examined by the Dual-Luciferase Reporter Assay System and detected by a Centro XS3 LB960 Luminometer (Berthold).

Enzymatic activity of HDAC1 was quantified by enzyme-linked immunosorbent assay using HDAC colorimetric Kit (BioVision, Milpitas, CA, USA) according to the manufacturer’s instructions.

In situ cell apoptosis was determined by detection of fragmented DNA, using DeadEnd Colorimetric Terminal deoxytransferase-catalyzed DNA nick-end labeling System (Promega Corporation, Madison, WI, USA), on 5 μm-paraffin sections according to the manufacturer’s instructions.

To test the in vivo efficiency of VPA, nude mice (5-6-week-old, obtained from Shanghai Laboratory Animal Center, Shanghai, China) were injected with 4 × 10⁷ Jurkat cells into the right flank. Treatments started after tumor became about 0.5 cm × 0.5 cm in surface (Day 0). The control group received saline, the treatment group received oral VPA for 14 days (0.4%w/v in the drinking water daily).

Difference of c-FLIP expression among groups were calculated using Mann–Whitney U test. The association between c-FLIP expression and clinical parameters was analyzed by Chisquare test. In vitro experimental results were expressed as mean±s.d. of data obtained from three separate experiments and determined by t-test to compare variance. P<0.05 was considered to be significant. Statistical analyses were performed on SPSS13.0 software.