The expression of IRF-1 and -2 can be altered by a host of factors/agents and thus they may serve as possible targets for the treatment of cancer. They have been most well studied in their role as transcription factors active in type I (IFN-α/β) and type II (IFN-γ) signal transduction pathways. There are several reports indicating that a decrease in the ratio of IRF-1/IRF-2 expression is prevalent in most tumor tissue types. In an analysis to study the protein expression profiles of IRF-1 and -2, we were able to show that although most normal and tumor tissues expressed IRF-1 and -2 the ratio of IRF-1 to IRF-2 dropped dramatically in the corresponding tumor tissue samples (Figure 1). This information builds a case for the relevance of IRF-1 and 2 in cancer biology and treatment.

To determine the cytotoxicity of VBL and pIFN-α, the cells were incubated with varying concentrations of the drugs at 37°C for 72 h. Tetrazolium (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra-zolium bromide (MTT)) cell proliferation assay was used to estimate the concentration required to inhibit cell growth by 50% (IC₅₀) for each drug. The IC₅₀ values of VBL and pIFN-α were approximately 1 nM and 0.5 μg/ml, respectively, for M14 cells (Figure 2a and b). Cells were treated with constant concentration of pIFN-α (0.5 μg/ml) and the indicated concentration of VBL (0–10⁴ pM) for 72 h. A greater than additive cell growth inhibition by MTT assay (70%) was observed when the cells were treated with a combination of 1 nM (10³ pM) VBL and 0.5 μg/ml of pegylated IFN-α (Figure 2c).

The effect of VBL and pegylated IFN-α on cell survival was also studied by clonogenic assay. Combination of VBL (1 nM) and pIFN-α (0.5 μg/ml) was 1.75-fold more toxic to the M14 cells than treatment with VBL (1 nM) alone. Treatment with half the concentration of VBL (0.5 nM) and pIFN-α (0.25 μg/ml) was as toxic as treatment with 1 nM of VBL alone (Figure 2d). These studies suggest that combination of pIFN-α and VBL may enhance the induction of cell death compared with either treatment alone.

IRF-1 is a critical transcriptional regulator in the IFN signaling pathway.⁵ Therefore, we investigated the expression of IRF-1 in response to pIFN-α and VBL in M14 melanoma cells. A dose of 30 nM for VBL was selected for treatment to ensure complete cell death. Cells were treated at the indicated time points with VBL (30 nM) or pIFN-α (0.5 μg/ml) alone and in combination at the same (30 nM VBL+0.5 μg/ml pIFN-α) or at half the concentration of each agent (15 nM VBL+0.25 μg/ml pIFN-α). IRF-1 was induced by 3–6 h and then decreased by 24 h in response to pIFN-α treatment. Interestingly, treatment with 0.25 or 0.5 μg/ml pIFN-α induced the same level of IRF-1. VBL treatment alone did not cause any noticeable induction of IRF-1. In addition, IRF-2 did not show any significant change in M14 cells after treatment with either VBL or pIFN-α (Figure 3a).

Subsequently, we looked at the induction of IRF-1 downstream targets and cell death in M14 melanoma cells, as assessed by poly (ADP-ribose) polymerase (PARP) cleavage, in response to pIFN-α and VBL individually or combined against M14 melanoma cells. Transcriptional induction of p21 is dependent on both p53 and IRF-1.⁸ M14 cells are p53 defective and possibly evade cell death by downregulation of p21. We have observed induction of p21 in M14 melanoma cells by 6 h and subsequent downregulation by 24 h in response to pIFN-α but not VBL (Figure 3b). Except for a delay in the induction of p21 by 3 h, the pattern of p21 protein expression in response to pIFN-α was similar to the pattern of IRF-1 induction suggestive of its transcriptional regulation by IRF-1.

Bak is a proapoptotic member of the Bcl-2 family of proteins and induces cell death by undergoing activation and homo-oligomerization.^{22, 23} We observed upregulation of total Bak in M14 melanoma cells upon treatment with pIFN-α but not VBL. Further, immunoblotting for PARP revealed that PARP cleavage, an indicator of cell death, occurs only upon treatment with VBL. However, combined treatment of VBL and pIFN-α at half the normal concentration of each agent (VBL (15 nM) and pIFN-α (0.25 μg/ml)) caused the same level of PARP cleavage as treatment with 30 nM VBL alone or the combined treatment (VBL (30 nM) and pIFN-α (0.5 μg/ml)). Treatment with pIFN-α alone did not cause PARP cleavage even by 48 h (Figure 3b).

To determine whether combined treatment with pIFN-α and VBL results in elevated activation of Bak, the conformationally active form was immunoprecipitated under native conditions from M14 cells treated for 36 h with VBL (30 nM) or pIFN-α (0.5 μg/ml) alone and in combination (30 nM VBL+0.5 pIFN-α μg/ml). For immunoprecipitation, we used rabbit anti-Bak (NT) antibody (Millipore, Temecula, CA, USA), which recognizes the conformationally active form of Bak. Cells treated with pegylated IFN-α had an increased level of inactive Bak (Supplementary Figure 1). Treatment with VBL caused inactive Bak to undergo a conformational change and increased levels of activated protein were detected. Combined treatment with pIFN-α and VBL induced a further accumulation of active Bak leading to an increase in the total amount of cell death (Figure 3c).

Using the technique of BH3 profiling developed by Letai and colleagues,^{24, 25} we observed that a buildup of inactive Bak in the mitochondria as a consequence of pIFN-α treatment made the M14 melanoma cells more sensitive to VBL-induced cell death (i.e., pIFN-α conditions or primes the cells to death signaling induced by other agents). The essence of this technique is to use isolated mitochondria, which contain inactive Bak, and induce cytochrome c (cyt c) release through the addition of tBid or Bid (BH3 peptide), which promotes Bak activation. Mitochondria isolated from M14 melanoma cells treated with 0.5 μg/ml pIFN-α, and subsequently incubated with an activator (BH3 peptide) of Bak, induced cyt c release in a dose-dependent manner (Figure 3d). In contrast, only low background levels of cyt c are released from the mitochondria of untreated cells under similar conditions (Figure 3d).

Using the MTT assay, EA.hy926 human endothelial cells were found to be relatively more sensitive to VBL than the M14 melanoma cells with an IC₅₀ between 500 pM and 1 nM (Figure 4a). The IC₅₀ for pIFN-α in these cells was ∼0.5 μg/ml (Figure 4b). To test the combined effect of the drugs, cells were treated with a constant concentration of pIFN-α (0.5 μg/ml) and indicated concentration of VBL (0–500 pM) or indicated concentrations of VBL alone for 72 h. Cell growth inhibition measured by MTT assay was 40% greater with the combined treatment of 500 pM VBL and 0.5 μg/ml of pIFN-α in comparison with treatment with 500 pM VBL alone (Figure 4c). In subsequent clonogenic studies, cells were treated with the indicated concentration of pIFN-α and VBL alone and in combination, and the medium was replaced with fresh medium without the drugs after 24 h. In contrast to the results from the MTT assay, the clonogenicity of EA.hy926 cells was found to be markedly reduced after pIFN-α treatment alone. Combination of VBL (500 pM) and pIFN-α (0.5 μg/ml) had a similar effect to that of pIFN-α alone (Figure 4d).

Induction of IRF-1 in EA.hy926 cells that have wild-type p53 was found to occur by 6 h in the presence of pIFN-α but not VBL (Figure 5a, top panel). Cyclin D1, like p21, is a downstream target of p53²⁶ and IRF-1.¹² We observed the induction of p21 and cyclin D1 by both VBL and IFN-α. Upregulation of p21 at 6 h after treatment was much greater when EA.hy926 endothelial cells were treated with VBL and pIFN-α. Half the normal concentration of the drugs in the combined treatment was sufficient to cause maximum induction of p21 (Figure 5a, upper middle panel). Upregulation of cyclin D1 at 6 h was followed by its downregulation by 24 h. However, the combined treatment caused a greater reduction in cyclin D1 at 24 h (Figure 5a, lower middle panel). Although the level of Bak expression tended to vary in the presence of pIFN-α alone or in combination with VBL, it was completely absent at 48 h after treatment with VBL alone (Figure 5b). This suggests that IFN-α treatment increases and maintains the level of Bak expression to some extent, a function likely mediated through IRF-1 activity and also indicated by our results from western blotting described in Supplementary Figure 1 and the BH3 profiling results shown in Figure 3.

We did not see substantial cell death even at 1 μM IFN-α treatment in the 72 h MTT assay (Figure 4b) but observed a significant reduction in cell survival in the clonogenic studies (Figure 4d). These perplexing results led us to inspect the doubling time of the cells after treatment with pIFN-α. Interestingly, although the control cells doubled in 0.9167 days (about 24 h), the pIFN-α-treated cells failed to increase in total number (Figure 6a). This loss of the proliferative capacity of EA.hy926 cells was subsequently observed to be at least partially caused by pIFN-α-induced senescence. We observed that upon exposure to pIFN-α, cells became enlarged, flattened, irregularly shaped and significantly more cells expressed SA β-gal+, suggesting a senescent phenotype (Figure 6b, middle and lower panel) as compared with untreated cell cultures (Figure 6b, upper panel).

To observe the effect of pIFN-α on the morphology of EA.hy926 endothelial cells over time, 1 × 10⁵ cells were treated for 1, 2 and 6 days with 0.25 and 0.5 μg/ml of pIFN-α and observed at × 200 magnification. Cells treated with IFN-α for 1 and 2 days were rinsed and grown in fresh medium until the sixth day. There were significantly more cells exhibiting a senescent phenotype in cultures treated continuously with 0.5 μg/ml of pIFN-α compared with cells treated with 0.25 μg/ml of pIFN-α (Supplementary Figure 2a). The cells were subsequently harvested and replated with fresh medium into new culture vessels for 2 weeks and again observed by phase-contrast microscopy. In order to prevent the cells from being over confluent, only one-fifth of the total cells in the untreated plate were replated. The senescent phenotype continued to persist in these freshly plated cells previously exposed to pIFN-α independent of total length of the initial exposure to pIFN-α (Supplementary Figure 2b).

Treatment with the microtubule inhibitor, VBL, causes the cells to arrest at G2/M. Cells tend to get rounded, start floating and eventually undergo cell death. In EA.hy926 cells, a similar morphological change was observed by 48 h (Supplementary Figure 3a) upon treatment with 30 nM VBL or with a combination of 15 nM VBL and 0.25 μg/ml pIFN-α. A quantification of adherent versus nonadherent cells after treatment with pIFN-α and VBL is shown in Supplementary Figure 3b and represents data from three replicates per condition. As unsynchronized cells were used for plating, we did observe a percentage (11%) of nonadherent cells upon IFN-α treatment as compared with almost 50% of nonadherent cells after VBL or combined treatment with IFN-α and VBL.

M14 is a human malignant melanoma cell strain, originally denoted as UCLA-SO-M14, later called M14, which expresses most melanoma-associated markers.^{40, 41} EA.hy926 is an immortalized human endothelial-like cell line derived from the fusion of primary human endothelial vein endothelial cells with the lung carcinoma cell line A549.⁴² The cell lines were maintained in monolayer culture at 37°C and 5% CO₂. The M14 human melanoma cell line was maintained in RPMI (Cellgro, Manassas, VA, USA), supplemented with 10% fetal bovine serum (Atlas, Fort Collins, CO, USA), 50 units/ml penicillin and 50 μg/ml streptomycin (Hyclone, Logan, UT, USA). The EA.hy926 endothelial cell line was cultured in DMEM/F-12 (Cellgro) with 10% bovine calf serum (Hyclone), 2 mM -glutamine, 100 units/ml penicillin, 100 mg/ml streptomycin and 1 : 50 HAT (Cellgro).

Pegylated IFN-α2b (PEGINTRON) referred as pIFN-α was purchased from Schering-Plough (Kenilworth, NJ, USA). VBL and thiazolyl blue tetrazolium bromide (MTT) were purchased from Sigma-Aldrich (St Louis, MO, USA). IRF-1, IRF-2 and cyclin D1 (DCS-6) antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Bak (NT) antibody for immunoprecipitation was from Millipore and Bak (Ab-1) antibody for immunoblotting was from Oncogene (Boston, MA, USA). Glyceraldehyde-3-phosphate dehydrogenase antibody was from Cell Signaling Technology (Danvers, MA, USA). Cytochrome c and p21 antibodies were from BD Pharmingen (San Jose, CA, USA). Crystal violet (Fisher Scientific, Fairlawn, NJ, USA) and protein A/G PLUS-agarose beads were from Santa Cruz Biotechnology.

Cells (2000 per well) were seeded in 96-well plates and on the following day VBL and/or pIFN-α was added. After 72 h, MTT (50 mg per well) was added for 4 h followed by dimethylsulfoxide solubilization of the cells, and an absorbance reading was done at 540 nm.

Cells were seeded at 200 cells per well in six-well tissue culture plates and on the next day treated with the indicated concentration of drugs for 24 h, and then colony formation was assessed. Cells were allowed to grow for 2 weeks after rinsing and replenishing the wells with fresh media without the drugs at 24 h. Colonies were stained with crystal violet and counted to calculate the surviving fraction.

Customized western blots (ProSci Inc., Poway, CA, USA) of 15 normal human and tumor tissues each were probed for IRF-1 (1 : 1000). The blots were stripped and reprobed for IRF-2 (1 : 1000). The level of IRF-1 and -2 protein expression was estimated by analyzing the relative band intensities using the Image J software (NIH, Bethesda, MD, USA).

Whole-cell extracts were prepared by suspending cells in 0.25 ml of lysis buffer (25 mM HEPES, pH 7.5, 0.5% sodium deoxycholate, 5 mM EDTA, 5 mM dithiothreitol, 20 mM glycerophosphate, 1 mM Na₃VO₄, 50 mM NaF, 1% Triton X-100, 20 μg/ml aprotinin, 50 μg/ml leupeptin, 10 μM pepstatin, 1 μM okadaic acid and 1 mM phenylmethylsulfonyl fluoride). Cells were incubated at 4°C on a nutator for 1 h, insoluble material was removed by centrifugation (15 min at 12 000 × g) and protein concentration was determined using the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA, USA).

Mitochondria (25 μg per reaction) isolated from M14 cells without and after treatment with 0.5 μg/ml pIFN-α for 48 h were incubated with the indicated concentrations of BH3 peptide for 30 min at room temperature. Supernatant was collected and cyt c release was observed by immunoblotting. DMSO was used as the vehicle (2% final concentration in all samples).²⁴

To evaluate Bak activation status, the cells were lysed in 0.5 ml of 10 mM HEPES (pH 7.4), 150 mM NaCl, 1% CHAPS, plus protease inhibitors (as above) and incubated at 4°C for 45 min on nutator. The cell extract was centrifuged at 12 000 × g for 15 min, and 500 μg of cell lysate was precleared using protein A/G-PLUS agarose beads for 30 min followed by incubation with 4 μg of Bak (NT) polyclonal antibody (Millipore) in a total of 500 μl of lysis buffer for 1 h. The lysates were then incubated with 25 μl of protein A/G-agarose beads (Santa Cruz Biotechnology) overnight at 4°C. The beads were pelleted by centrifugation and washed three times with 0.2 ml of lysis buffer. The beads were resuspended in 50 μl of 2 × SDS loading buffer and electrophoresed on 12.5% SDS-PAGE precast acrylamide gels (Bio-rad). Following electrophoresis, immunoblotting was performed using mouse anti-Bak (Ab-1) monoclonal antibody (Oncogene) at 1 : 1000 dilution.²³

The SA-β-Gal assay was performed as previously described in the manufacturer's instructions (Senescence β-galactosidase staining kit, Cell Signaling Technology). Briefly, cells were washed in phosphate-buffered saline and fixed in 2% formaldehyde–0.2% glutaraldehyde. Then the cells were washed and incubated at 37°C overnight with fresh senescence-associated β-Gal stain solution (1 mg of 5-bromo-4-chloro-3-indolyl-β--galactopyranoside (X-Gal per ml), 40 mM citric acid–sodium phosphate (pH 6.0), 150 mM NaCl, 2 mM MgCl₂, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide). For all visualization and photography, a Nikon Eclipse Te2000-U microscope (Kanagawa, Japan) was used.