Interferon-α Acts on the S/G2/M Phases to Induce Apoptosis in the G1 Phase of an IFNAR2-expressing Hepatocellular Carcinoma Cell Line*

Background: The mode of action of interferon-α has been unknown. Results: Its point of action in the cell cycle was analyzed by single cell tracking using time lapse confocal imaging. Conclusion: Interferon-α activates p63 in S/G2/M and induces apoptosis and cell cycle arrest in the subsequent G1. Significance: Tracking cell cycle progression is crucial for understanding the mechanisms of interferon-α. Interferon-α (IFN-α) is used clinically to treat hepatocellular carcinoma (HCC), although the detailed therapeutic mechanisms remain elusive. In particular, IFN-α has long been implicated in control of the cell cycle, but its actual point of action has not been clarified. Here, using time lapse imaging analyses of the human HCC cell line HuH7 carrying a fluorescence ubiquitination-based cell cycle indicator (Fucci), we found that IFN-α induced cell cycle arrest in the G0/G1 phases, leading to apoptosis through an IFN-α type-2 receptor (IFNAR2)-dependent signaling pathway. Detailed analyses by time lapse imaging and biochemical assays demonstrated that the IFN-α/IFNAR2 axis sensitizes cells to apoptosis in the S/G2/M phases in preparation for cell death in the G0/G1 phases. In summary, this study is the first to demonstrate the detailed mechanism of IFN-α as an anticancer drug, using Fucci-based time lapse imaging, which will be informative for treating HCC with IFN-α in clinical practice.

Hepatocellular carcinoma (HCC) 3 is one of the most common malignant diseases worldwide. HCC is commonly preceded by chronic inflammation such as that caused by hepatitis C virus (1). Interferons are secreted from hepatocytes as an immunological response to hepatitis C virus infection (2). Interferons not only trigger innate and adaptive immune responses but also regulate proliferation, differentiation, and apoptosis of different cell types, and they have therefore been used for clinical treatment of leukemia and other cancers (3).
Human interferons are classified into three major subtypes (types I-III). Among them, interferon-␣ (IFN-␣) binds specifically to the heteromeric type I interferon receptor complex comprising interferon-␣ receptor-1 and -2 (IFNAR1 and IFNAR2), which utilize JAK tyrosine kinase and STAT transcription factors to activate interferon response genes (4). Phosphorylated STAT1 and STAT2 form a multimeric complex with IFN-regulatory factor 9 (IRF9), known as IFN-stimulated gene factor 3 (ISGF3), which then translocates to the nucleus and binds to IFN-stimulated response elements to initiate gene transcription (5,6). It was previously reported that combinatorial treatment with IFN-␣ and 5-fluorouracil (5-FU), an anticancer agent, against advanced HCC results in an efficient clinical outcome (20.0% of patients showed a complete response, and 13.3% showed a partial response) (7). The therapeutic effects could be seen only in HCCs expressing high levels of IFNAR2, indicating the importance of the IFN-␣/IFNAR2 axis (7,8).
Aside from the accumulated evidence of its clinical utility, the actual therapeutic mechanisms remain obscure, although some previous reports have indicated the significance of cell cycle regulation in the anticancer action of the IFN-␣/IFNAR2 axis. IFN-␣ was shown to induce G 1 arrest (9) and increase susceptibility to cytotoxic anticancer drugs (10), such as 5-FU, mainly targeting the S phase. However, these effects were evaluated based only on flow cytometry and remain controversial (11)(12)(13). Observation of apoptosis induction by IFN-␣ in living cells and monitoring of cell death at distinct cell cycle stages over time are essential to clarify the mechanisms of the cellular responses induced by IFN-␣.
The fluorescence ubiquitination-based cell cycle indicator (Fucci) is a fluorescent probe that detects the various stages of the cell cycle in living cells (14). In this system, geminin, a nuclear protein enriched in the S/G 2 /M phases, and Cdt1, which is enriched in the G 0 /G 1 phase, are marked by green and red fluorescent proteins, respectively.
In this study, we exploited these advanced imaging technologies to analyze the therapeutic mechanism of IFN-␣ against IFN-␣-susceptible HCCs, particularly in association with the cell cycle.

EXPERIMENTAL PROCEDURES
Cell Lines and Reagents-The human HCC cell line HuH7 was obtained from the Japan Cancer Research Resources Bank (Tokyo, Japan). These cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ml streptomycin at 37°C in a humidified incubator with 5% CO 2 in air. Cells were treated with either purified human IFN-␣ (kindly provided by Otsuka Pharmaceutical Co., Tokyo, Japan) at final concentrations of 10 -10,000 IU/ml or 5-FU (Wako Pure Chemical Industries, Osaka, Japan) at final concentrations of 10, 100, or 1000 M. To synchronize the cell cycle of HuH7 cells, the cells were treated with aphidicolin (Calbiochem) at a final concentration of 1.6 g/ml for 24 h.
Quantitative Real-time RT-PCR and Immunoblot Analyses-Quantitative real-time RT-PCR (qRT-PCR), preparation of cell lysates, and immunoblot analyses were performed according to a previous report (18). All oligonucleotides used for qRT-PCR analyses were designed to amplify cDNA across exon-intron junctions, as described in Table 1. All data were normalized against internal GAPDH controls. Relative expression levels of IFNAR1 and IFNAR2 of WT controls were set as 1, and mRNA levels of mock and IFNAR2-expressing HuH7 samples were accordingly plotted as a -fold change. A full-length human p63 cDNA sequence, obtained from the Kazusa DNA Research Institute (Chiba, Japan), was transfected into HuH7 cells using FuGENE6 (Promega, Tokyo, Japan), according to the manufacturer's protocol. Cell lysates ware collected 1 day after transfection to perform qRT-PCR analyses.
To analyze activation of STAT1, -2, and -3, we performed immunoblot analyses using the primary antibodies described in Table 2. Mock control and IFNAR2-expressing cells were cultured in medium containing 10% FBS for 2 days and then in

TABLE 2
List of antibodies used for immunoblotting analysis in this study FBS-free medium for 2 h. They were then incubated in medium with IFN-␣ (100 IU/ml) for 0 -80 min, and the cells were lysed with mammalian lysis buffer (Promega, Tokyo, Japan). Cell lysates (10 g of protein/lane) were loaded on 4 -20% Mini-PROTEAN TGX TM gels (Bio-Rad), separated by electrophoresis, and blotted onto a nitrocellulose membrane. The membrane was blocked in PVDF Blocking Reagent for Can Get Signal (Toyobo, Osaka, Japan) for 1 h at room temperature and incubated with the specific antibodies in Can Get Signal Immunoreaction Enhancer Solution 1 TM (Toyobo, Osaka, Japan) overnight at 4°C. After washing in 1ϫ TBS-T, the membranes were incubated with HRP-labeled polyclonal secondary anti-rabbit (NA934V; GE Healthcare) or anti-mouse-specific antibodies (NA931V; GE Healthcare). Chemiluminescence was detected with an ECL Prime Kit (PerkinElmer Life Sciences) using an LAS4000 imaging system (GE Healthcare).
To analyze IFN-␣/IFNAR2 signaling in G 0 /G 1 or S/G 2 /M phases, IFNAR2-expressing Fucci-labeled HuH7 cells were treated with or without IFN-␣ (1,000 IU/ml) and sorted by FACS 24 h after treatment. Immunoblotting was performed as described above. The primary antibodies are listed in Table 2.
To deplete p63 expression, we performed RNAi with MISSION siRNA (Hs_TP63_6771 for TP63-specific and MISSION siRNA Universal Negative Control 1 for control; Sigma). RNA duplexes were transfected into IFNAR2-expressing HuH7 cells with RNAi MAX (Invitrogen) according to the manufacturer's instructions. After 24 h of transfection, cells were treated with 1,000 IU/ml IFN for 0, 24, or 48 h. qRT-PCR and immunoblot analyses were performed as described above (18).
Time Lapse Imaging-Cells (1.5 ϫ 10 4 /cm 2 ) were grown overnight at 37°C before imaging on a glass bottom dish in phenol red-free Dulbecco's modified Eagle's medium containing 10% FBS. Time lapse imaging was performed every 30 or 60 min with a confocal A1 microscope system (Nikon, Tokyo, Japan) equipped with a humidified imaging chamber (Nikon) at 37°C, 5% CO 2 in air. Time lapse images were analyzed using Nikon NIS-Elements software (Nikon).
Cells were defined as apoptotic if they showed morphological changes, such as cell shrinkage and fragmentation into membrane-bound apoptotic bodies. The fluorescent signal detected in apoptotic cells was used to identify the cell cycle phases at which apoptosis occurred (red, G 0 /G 1 ; green, S/G 2 /M). To analyze the relationship between the cell cycle and apoptosis, each individual cell was tracked, and cell cycle changes and cell fate were monitored for 72 h. The frequency of cell death was calculated by dividing the initial total cell number in the visual field at t ϭ 0 by the number of apoptotic cells for each cell cycle phase at the time of cell death (t ϭ 0 until t ϭ 72 h).
Flow Cytometry-To analyze the DNA content of Fuccitransfected HuH7 cells, we stained the cells with Hoechst 33342 (final concentration, 3.6 g/ml; Invitrogen). After incubation for 30 min, cells were harvested and analyzed using a FACSCanto II flow cytometer (BD Biosciences). Both mKO2 and mAG were excited by a 488-nm laser, and Hoechst 33342 dye was excited by a 325-nm laser. Fluorescence signals were collected at 530 nm (530/28 BP) for mAG, at 575 nm (575/26 BP) for mKO2, and at 400 nm (380 LP) for Hoechst 33342 dye (14).
Preparative FACS sorting was performed using a FACSAria (BD Biosciences) cell sorter equipped with a 488-nm laser using 530/30BP or 585/42BP filters, respectively. The data were analyzed using FlowJo software (Tree Star, Inc., Ashland, OR).

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Bromide
(MTT) Assay-The MTT assay was performed with the Cell Proliferation Kit 1 (Roche Applied Science) according to the manufacturer's protocol. In short, cells (1 ϫ 10 3 cells/96-well dish) were grown overnight at 37°C in a 96-well dish. Following treatment with or without IFN-␣ (30, 100, 300, 1,000, 3,000, and 10,000 IU/ml) for 72 h at 37°C, cells were labeled with MTT-labeling reagent (MTT final concentration, 0.5 mg/ml; Roche Applied Science) for 4 h at 37°C and subsequently solubilized with Solubilization Solution (Roche Applied Science) for 16 h at 37°C. The absorbance was measured in a microplate reader (PowerScanHT; DS Pharma Biomedical, Osaka, Japan) at a wavelength of 550 nm with a 650-nm reference. The assays were carried out in 12 replicates at each IFN-␣ concentration in three individual experiments, and the results were plotted as a percentage of the absorbance relative to untreated controls. The concentration of IFN-␣required to reduce the cell viability to 70% that of control cells (IC 70 ) was calculated from the spline curve generated using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA).
Statistical Analyses-Differences between the control and treated groups were assessed by an unpaired Student's t test or Mann-Whitney U test and considered to be significant at p Ͻ 0.05 (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.005). Values are given as means Ϯ S.E. Statistical analysis was performed using GraphPad Prism software (version 5.0; GraphPad Software).

IFN-␣ Reduces the Viability of IFNAR2-expressing HCC
Cells-To study the role of IFN-␣ in the induction of apoptosis of HCC cells, we transduced IFNAR2 into the human HCC cell    line HuH7 with a constitutive retroviral expression vector for IFNAR2 (pMXs-IFNAR2) because the endogenous expression level of IFNAR2 in HuH7 cells is quite low (19). To confirm expression in HuH7 cells, we performed qPCR and immunoblot analyses. We obtained a HuH7 cell line expressing a higher level of IFNAR2, 30-fold higher in mRNA (Fig. 1A) and 9-fold higher in protein (Fig. 1B) than the mock control. Expression of IFNAR1 was also moderately increased in both the mock control and IFNAR2-expressing HuH7 cells (Fig. 1, C and D), possibly due to intrinsic cell responses against viral infection (20). To determine activation of the JAK/STAT signaling pathway upon IFN-␣ treatment of these cells, we performed immunoblot analyses to detect phosphorylation of STATs (STAT1, -2, and -3) (Fig. 1E). All STATs were phosphorylated in both the mock control and IFNAR2-expressing HuH7 cells, although the phosphorylation level was increased in IFNAR2-expressing HuH7 cells, indicating the function of exogenously expressed IFNAR2.
To confirm the responsiveness of these HuH7 cells to IFN-␣ treatment, we performed MTT assays (Fig. 1F). As expected, upon treatment with increasing concentrations of IFN-␣, the viability of IFNAR2-expressing cells was significantly lower than that of WT and mock controls (Fig. 1F), indicating specific growth inhibition, such as apoptosis or cell cycle arrest induction through an IFN-␣/IFNAR2 interaction. This result confirms that IFN-␣ negatively regulates cell viability in an IFNAR2-dependent manner and validates these IFNAR2-expressing HuH7 cells as a useful model with which to investigate the cellular responses of HCC to IFN-␣ therapy in vitro.
IFN-␣ Specifically Induces Apoptosis by an IFNAR2 Signaling Pathway at the G 0 /G 1 Phases of the Cell Cycle-To examine the association between IFN-␣ action and cell cycle status, we introduced Fucci into the IFNAR2-expressing HuH7 cells and control (mock) HuH7 cells, both of which we have established (Fig. 1). Fucci-introduced IFNAR2-expressing HuH7 cells and mock controls allowed for tracking of individual cells over time and cell cycle changes during cell division by time lapse imaging using confocal microscopy ( Fig. 2A). Furthermore, the DNA content (detected by a nuclear acid dye, Hoechst 33342) of Fucci-labeled HuH7 cells in the G 0 /G 1 and S/G 2 /M phases varies according to the fluorescently indicated stages of the cell cycle in flow cytometric analyses (Fig. 2B), indicating the proper functioning of the Fucci system.
Concerning putative differential roles for IFN-␣ and 5-FU in effective clinical combination therapies for HCC, we performed in vitro time lapse imaging of Fucci-labeled IFNAR2-expressing HuH7 cells treated with IFN-␣ or 5-FU. Cells were defined as apoptotic if they showed morphological changes indicative of cell shrinkage and fragmentation into membrane-bound apoptotic bodies (21). First, we demonstrated that treatment with 5-FU, a nucleic acid analog that prevents cell division, led to accumulation of green (S/G 2 /M) cells over time (Fig. 3A and supplemental Video 1) in a dose-dependent manner (Fig. 4, A  and B). In addition, the cell death events (Fig. 3A, arrowheads) occurred preferentially in the green S/G 2 /M phases in both the control and IFNAR2-expressing HuH7 cells with comparable efficiency (Fig. 3B). These results are consistent with the conventional idea that nucleic acid analogs, such as 5-FU, block the cell cycle at the S phase and exert cytotoxicity (22), and it was expected that this cytotoxic effect induced by 5-FU was not dependent on the expression of IFNAR2.
In contrast, IFN-␣ treatment resulted in an accumulation of red (G 0 /G 1 ) IFNAR2-expressing HuH7 cells, and the results were not obvious in mock controls (Figs. 3C and 4 (C and D) and supplemental Video 2). Moreover, a high frequency of cell death in red (G 0 /G 1 ) cells was also observed by time lapse imaging, albeit in only IFNAR2-expressing HuH7 cells (Fig. 3, C (arrowheads) and D). It should be noted that IFN-␣ treatment did not increase the rate of death of green (S/G 2 /M) cells but resulted in an increased proportion of red (G 0 /G 1 ) apoptotic cells at 100 or 1,000 IU/ml IFN-␣, suggesting that IFN-␣ causes G 0 /G 1 arrest.
These observations identified unique roles for IFN-␣ and 5-FU in inducing cell death at specific stages of the cell cycle. In addition, our data emphasize the importance of the IFN-␣/ IFNAR2 signaling pathway in regulating the efficiency of apoptosis induction at the G 0 /G 1 phases of the cell cycle.
Cell Cycle Synchronization Revealed That IFN-␣/IFNAR2 Signaling Acts on S/G 2 /M but Induces Apoptosis at the G 0 /G 1 Phases in HuH7 Cells-To more precisely monitor the time course of the cell cycle-dependent IFN-␣ effect on IFNAR2expressing HuH7 cells, we synchronized the cell cycle of HuH7 cells during G 0 /G 1 to the early S boundary using aphidicolin, a reversible inhibitor of DNA polymerases (23). We were able to synchronize Fucci-transfected IFNAR2-expressing HuH7 cells to the early S phases (green) by the addition of aphidicolin for 24 h or to the G 0 /G 1 phases (red) by removal of aphidicolin for 12 h (Fig. 5A). The relative populations of the cell cycle phases of the synchronized cells were evaluated by FACS, which indi- cated that 77.7 and 86.3% of cells were enriched to early S (green) or G 0 /G 1 (red) by these methods, respectively (Fig. 5B).
Using this synchronization technique, we examined the effect of IFN-␣ exposure on the cell cycle of HuH7 cells. Most of the early S-synchronized cells (ϳ57.1%) underwent apoptosis when they entered the next G 0 /G 1 phase (Fig. 5, C (arrowheads) and D, and supplemental Video 3), suggesting that IFN-␣-induced cell death occurs during mainly the G 0 /G 1 phase. In contrast, a majority of the G 0 /G 1 -synchronized cells (ϳ29.4%) died in the re-entered G 0 /G 1 phase after going through the next S/G 2 /M phases (Fig. 5, E (arrowheads) and F, and supplemental Video 4). Concordantly, tracking of respective cell fates showed that cell deaths occurred faster in S phase-synchronized cells (40.6 Ϯ 2.1 h) than in G 0 /G 1 -synchronized cells (52.4 Ϯ 1.2 h) (Fig. 5G). We also checked the cell cycle status using FACS and found that IFN-␣-induced accumulation of cells in the G 0 /G 1 phase, which can be assumed to be a preapoptotic population, was detected earlier in the early S-synchronized cells (ϳ24 h) (Fig. 6A), compared with the case with the G 0 /G 1 -synchronized cells (ϳ72 h) (Fig. 6B). These results indicate that the S/G 2 /M phase is critical for receiving the "cell cycle arrest and death" signals from IFN-␣, although the cells died mainly during the following G 0 /G 1 phase.

Biochemical Analyses of IFN-␣/IFNAR2-induced Signaling
Cascades in HuH7 Cells-Finally, we assessed the molecular mechanisms underlying cell cycle-dependent IFN-␣ action. The interaction of IFN-␣ and IFNAR2 has been reported to activate the JAK/STAT signaling cascade, which is considered to be involved in G 1 arrest, leading to apoptosis (5) (Fig. 7A). To confirm the activation of this pathway, we performed immunoblotting for signaling components of the proapoptotic JAK/ STAT pathways on IFNAR2-expressing Fucci-labeled HuH7 cells, which were used in the confocal time lapse imaging. Cells were treated or not treated with IFN-␣ for 24 h and separately collected for identification of G 0 /G 1 or S/G 2 /M populations by cell sorting.
Expression of p53, which has been reported to play a central role in cell cycle arrest and apoptotic induction (6), was unaltered in HuH7 cells treated by IFN-␣, irrespective of the cell cycle status (Fig. 7, B (lane 1) and C). This was probably because p53 in HuH7 cells carries a point mutation (Y220C) that renders it non-susceptible to IFN-␣ (24). Instead, we monitored the expression of p63, a molecule homologous to p53 (25,26) and thus considered to play roles comparable with those of p53. We detected marked elevation of p63, especially during the S/G 2 /M phases (Fig. 7, B (lane 2) and C), suggesting that IFN- ␣-induced apoptotic signals were activated preferentially in the S/G 2 /M phase. In the case of p63, there are two distinct promoters that result in two different types of proteins with opposing functions (i.e. p53-like proteins containing the transactivation (TA) domain (TAp63) and inhibitory proteins lacking the TA domain) (⌬Np63) (27). In our HuH7 cells, only TAp63 was confirmed to be expressed both at the mRNA (Fig. 7D) and protein levels (data not shown). In contrast, the increase in p21, which has been shown to be responsible for G 1 arrest (28), and the activation of caspase-3 and -7 could be preferentially detected in the G 0 /G 1 phase (Fig. 7, B (lanes 3-6) and C), both of which have been shown to be inducible by p63 (29). These results were in agreement with the time lapse imaging results and suggest that IFN-␣ initiates its action during the S/G 2 /M phase and that cell cycle arrest and apoptosis occur in the subsequent G 0 /G 1 phase. We detected significant up-regulation of STAT1/2 expression and phosphorylation upon stimulation with IFN-␣ in both the G 0 /G 1 and S/G 2 /M phases (Fig. 8, lanes  1-4). In contrast, STAT3 was phosphorylated upon stimulation with IFN-␣ in both the G 0 /G 1 and S/G 2 /M phases, and total expression of STAT3 was not affected by IFN-␣ treatment. IFN-␣-activated JAKs are also known to regulate the activation of Vav or other guanine nucleotide exchange factors, resulting in the regulation of p38 mitogen-activated protein kinase (MAPK) (5). Both the expression and phosphorylation of p38 were not changed by these treatments (Fig. 8, lanes 7 and 8), suggesting that these pathways may not be dependent on the cell cycle. To examine whether p63 can induce cell cycle arrest and apoptosis in HuH7 cells, we examined the mRNA expres-sion of p21 and p53 up-regulated modulator of apoptosis (PUMA) in p63 (TAp63)-transfected HuH7 cells. Upon overexpression of TAp63, both p21 and PUMA were significantly up-regulated ϳ4.0and 3.7-fold, respectively (Fig. 9), indicating the potency of TAp63 for inducing proapoptotic signaling in HuH7 cells.
To examine whether the apoptosis induced by IFN-␣ was dependent on p63, we performed knockdown experiments of p63. Expression of p63 was confirmed to be reduced in p63knockdown cells (Fig. 10A), and p63-knockdown cells showed less activation of caspase-3 and -7 (Fig. 10B). This result clearly suggests that p63 was critically involved in apoptosis induced by the IFN-␣/IFNAR2 signaling pathway.

DISCUSSION
IFN-␣ has been used in the clinic to treat HCC, indicative of its therapeutic potential. However, its actual modes of action are less clear, precluding further development of this line of novel therapies (30). Flow cytometric analysis is commonly used for studying the cell cycle, although we did not observe cell death events with this conventional methodology. This study is the first to use confocal time lapse cell cycle imaging analyses with Fucci to visualize the specific relationship between cell cycle and cell death. The results demonstrated that IFN-␣ exerts its action on IFNAR2-expressing tumor cells by first sensitizing cells in the S/G 2 /M phases prior to inducing apoptosis in the G 1 phase. Combination therapy with IFN-␣ and 5-FU, commonly used clinically (30), can now be identified as an ideal regimen, because 5-FU is known to increase the S/G 2 /M population (see Fig. 4), which makes the tumor cells more susceptible to IFN-␣. The present evidence provides clinicians with the rationale for use of this combination therapy and enables them to further improve the regimen based on its mode of action.
However, this study raises new questions. Among them is how the action of IFN-␣ is dependent on the cell cycle. Biochemical analyses indicated that "cell cycle-independent" activation of STAT1/2 signaling with up-regulation of p63, a proapoptotic signaling molecule, was observed predominantly in the S/G 2 /M phase. A promoter region of p63 reportedly contains putative binding motifs for E2F1, an E2F family transcription factor responsible for cell cycle regulation (31), although E2F1 may not bind to this motif by itself (32). Nevertheless, we can hypothesize that E2F1 or other cell cycle-dependent factors induce p63 expression if accompanied by STAT1 and -2 activation stimulated by IFN-␣. On the other hand, it is also possible that the event of mitosis may be required for the induction of cell death by IFN-␣. Further investigation is necessary to elucidate the whole picture of this intricate apoptotic signaling reg- FIGURE 8. Biochemical analyses of JAK/STAT and MAPK signaling. The expression of several signaling molecules activated by IFN-␣, including STAT1, phospho-STAT1 (pSTAT1), STAT2, phospho-STAT2 (pSTAT2), STAT3, phospho-STAT3 (pSTAT3), p38, phospho-p38 (pp38), and of ␤-actin (control) was analyzed by immunoblotting after culturing in the absence or presence of IFN-␣ (1,000 IU/ml) for 24 h. IFNAR2-expressing Fucci-introduced HuH7 cells were sorted by FACS for the G 0 /G 1 (mKO single positive) or S/G 2 /M (mAG single positive) samples. ulation associated with the cell cycle and controlled by therapeutic use of IFN-␣.