Distinct Roles for Ifn Regulatory Factor (irf)-3 and Irf-7 in the Activation of Antitumor Properties of Human Macrophages

When properly activated, macrophages can be tumoricidal, thus making them attractive additions to standard cancer therapies. To this end, tolerance and activity of human autologous IFN- ;-activated macrophages, produced in large scale for clinical use (MAK cells), have been assessed in pilot trials in cancer patients. In the present study, we tested the hypothesis that activation of IFN regulatory factor (IRF)-3 and IRF-7, with subsequent type I IFN production, may be involved in the acquisition of new antitumor functions by macrophages. Adenoviral vectors were generated for the delivery of constitutively active forms of IRF-3 (Ad-IRF-3) or IRF-7 (Ad-IRF-7) into primary human macrophages. Cell death was observed in Ad-IRF-3-transduced macrophages, whereas Ad-IRF-7-transduced macrophages produced type I IFNs and displayed increased expression of genes encoding tumor necrosis factor (TNF)–related apoptosis-inducing ligand, interleukin (IL)-12, IL-15, and CD80, persisting for at least 96 hours. Expression of iNOS, TNF-a, FasL, IL-1, and IL-6 genes was unaltered by Ad-IRF-7 transduction. Interestingly, Ad-IRF-3 or Ad-IRF-7 transduction negatively regulated the transcription of protumorigenic genes encoding vascular endothelial growth factor and matrix metalloproteinase-2. Furthermore, Ad-IRF-7-transduced macrophages exerted a cytostatic activity on different cancer cell lines, including SK-BR-3, MCF-7, and COLO-205; the latter cells were shown previously to be insensitive to MAK cells. In conclusion , transduction of active forms of IRF-3 or IRF-7 differentially modulate the apoptotic and antitumor properties of primary macrophages, with active IRF-7 leading to the acquisition of novel antitumor effector functions.


Introduction
Macrophages are present in a resting state in many tissues, where they control clearance of apoptotic cells, healing, and angiogenesis.When properly activated, macrophages exert tumoricidal activities that are mediated through phagocytosis, antibodydependent cell cytotoxicity, and production of interleukin (IL)-1, tumor necrosis factor (TNF), TNF-related apoptosis-inducing ligand (TRAIL), FasL, nitric oxide (NO), and oxygen radicals (reactive oxygen species; for review see ref. 1).In particular, IFN-g rapidly primes macrophages and increases expression of Fcg receptor, inducible NO synthase (iNOS), TRAIL, cell adhesion molecules, and production of IL-1, resulting in enhanced tumor cell killing.Adoptive transfer of activated macrophages is an attractive complement to conventional cancer therapies.Phase I and II studies have shown the feasibility and safety of transfer of IFNg-activated macrophages (MAK cells) in several malignancies, including mesothelioma, ovarian, and bladder cancer (2)(3)(4).
The IFN family includes type I IFNs and IFN-g, which bind two distinct cell surface receptors, type I and type II IFN receptors, respectively ( for review, see ref. 5).Some immunomodulatory effects are shared by IFN-a/h and IFN-g, although type I IFNs exert stronger antiviral, antiproliferative, and antiangiogenic effects than IFN-g and are widely administered as adjuvant therapy in cancer and viral-related diseases.Both type I IFNs and IFN-g are produced during the early innate immune response (6).IFN-g is secreted by T lymphocytes and natural killer (NK) cells, whereas IFN-h and/or IFN-a are rapidly produced following the sensing of viral nucleic acid in the cytoplasm of cells or after engagement of Toll-like receptors (TLR) in immune cells (7).Viral entry or engagement of TLR3 or TLR4 induces the phosphorylation of IFN regulatory factor (IRF)-3 by TANK-binding kinase 1 and InB kinase q kinases (8-10), leading to its dimerization, nuclear translocation, and transcriptional activation through binding to IFN-stimulated response element (ISRE) sites.IRF-3 cooperates with nuclear factor-nB (NF-nB) and activating transcription factor-2/c-Jun to form a transcriptionally active enhanceosome complex on the IFNB promoter (11).IRF-3 is also critical for activation of the human IFNA1 or mouse ifna4 promoters (12).Secretion of the newly produced IFN-h and binding to adjacent cognate type I IFN receptors lead to the transcription of the IRF7 gene.Activation of IRF-7 is crucial for the full induction of type I IFN production (13), and irf7-deficient mice show a profound defect in the type I IFN response (14).
In the present study, we tested the hypothesis that expression of IRF-3 and/or IRF-7 with subsequent production of type I IFN may be involved in macrophage activation as well as in the acquisition of new antitumor effector functions.Previous investigations showed that gene delivery of wild-type (WT) IRF-3 or IRF-7 proteins can be sufficient for the expression of some IRF target genes; however, only marginal levels of type I IFN are detected in the absence of virus infection (13,15).On the other hand, the IRF-3 5D mutant, in which residues at positions 396, 398, 402, 404, and 405 were replaced by the phosphomimetic aspartate amino acid, induces strong activation of the IFNB promoter in the absence of virus induction (16,17).An internal deletion of IRF-7, which removed amino acids 247 to 467 ( the inhibitory and the nuclear export domains, activates IFNA gene transcription >1,000-fold compared with WT IRF-7 (18).Thus, in the present study, recombinant adenoviruses were generated that expressed constitutively active forms of IRF-3 and IRF-7.Expression of IRF-3 5D in adenovirus-transduced primary human macrophages induced rapid cell death, whereas expression of IRF-7 D247-467 in macrophages resulted in production of type I IFNs, expression of IRF or IFN-a target genes, and increased tumoricidal activities.

Materials and Methods
Construction of recombinant adenoviral vectors.Constitutively active forms of IRF-3 and IRF-7 were chosen for these studies based on our previous analysis (16,19).To avoid possible toxicity related to the longterm expression of IRF-3 5D or IRF-7 D247-467 in adenovirus-producing cells, IRF-3 and IRF-7 active forms were subcloned into the pAdPS-CMV5-CuO adenovirus transfer vector, in which the cymene bacterial operator (CuO) was inserted downstream of the TATA box of the optimized cytomegalovirus (CMV) 5 promoter (20), such that this promoter is negatively regulated by the binding of the 28-kDa CymR repressor (Fig. 1A).Transgene expression was minimal in 293 cells stably expressing CymR (used for adenovirus recombination and amplification), whereas CMV5-driven expression was maximal in any other cell type (21).The pAdPS-CMV5-CuO adenovirus transfer vector also displays a dicistronic cassette for coexpression of GFPq that was removed before the subcloning of cDNA encoding GFPq/IRF-3 5D or GFPq/IRF-7 D247-467 fusion proteins (Fig. 1A).The full-length parental pAdPS-CMV5-CuO vector was used to prepare the control adenovirus vector (Ad-GFP).The viruses were generated using the positive selection method by transfection of the transfer vectors encoding the adenovirus protease (PS) gene and infection with an adenovirus vector deleted in the PS gene as previously detailed (22,23).To minimize the effect of adenovirus proteins on the cell cycle and immune activation, recombinant adenovirus vectors were amplified from a parental adenovirus vector deleted for the E1, E3, and E4 (with exception of the orf6 region) transcription units into 293 CymR cells because 293 cells complement the E1 functions necessary for virus replication.Therefore, Ad-GFP, Ad-GFPq/ IRF-3 5D (Ad-IRF-3), and Ad-GFPq/IRF-7 D247-467 (Ad-IRF-7) vectors shared the same backbone.The three adenovirus vectors were quantified in parallel and purified by double cesium chloride centrifugation as described previously (24).
Macrophage and MAK preparation.Primary macrophages and MAK cells were obtained essentially as described previously (25).MAK cells were generated by activation of macrophages with IFN-g (166 IU/mL; Imukin, Boehringer Ingelheim, Vienna, Austria) during the last 18 hours of the 7-day culture period.Macrophages and MAK cells were identified by assessment of CD14, CD64, and HLA-DR antigen expression.Fresh macrophages and MAK cells were used right after elutriation.The remaining cells were frozen in 4% human serum albumin and 10% DMSO and aliquoted in vials each containing 20 Â 10 6 to 40 Â 10 6 cells.Cells were thawed 1 to 12 months later in 4% human serum albumin and washed with McCoy's 5A medium (Cambrex Bio Science, Walkerville, MD) complemented with 10% FCS.
Cell line, infection, luciferase, and antiviral assays.For adenovirus transduction, macrophages were resuspended in complete medium at 25 Â 10 6 cells/mL and infected with adenovirus vectors for 3.5 hours at 50 to 150 multiplicities of infection (MOI) in a volume of 1.5 mL.Cells were then washed twice in complete medium and plated at 10 6 per mL in six-well plates.A549 cells were infected with adenovirus vectors at 10 to 100 MOI in 0.5 mL medium for 1.5 hours, washed twice, and plated at 2 Â 10 5 per mL.Alternatively, A549 cells were plated at 2 Â 10 5 per mL in six-well plates or 12-well plates and adenovirus vectors were added to the medium 18 hours later.For reporter gene assays, A549 cells (2 Â 10 5 ) were transfected with 100 ng pRLTK reporter (Renilla luciferase for internal control) and 400 ng pGL-3 plasmid harboring the Firefly luciferase reporter gene expressed under the control of the RANTES promoter (containing point mutations in the NF-nB binding elements; ref. 26), the IFNB promoter, or the IFNA4 promoter (18) using the LipofectAMINE reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions.At 5 hours after transfection, cells were transduced or not with adenovirus vectors.Reporter gene activities were measured 36 hours later as described by the manufacturer (Promega, Madison, WI).To study the antiviral response induced by IRF-3 or IRF-7, A549 cells (2 Â 10 5 ) were transduced or not with adenovirus vectors at 50 MOI, infected 48 hours later with the vesicular stomatitis virus (VSV) heat-resistant (VSV HR) strain at 0.5 MOI, and harvested 24 hours after VSV infection.WT VSV HR (Indiana serotype) propagation, purification, and titration were described previously (27).The human breast cancer cell lines SK-BR-3 (ATCC HTB-30) and MCF-7 (ATCC HTB-22) and the human colon cancer cell line COLO-205 (ATCC CCL-222) were grown in McCoy's 5A medium supplemented with 10% FCS and antibiotics.Tumor cell sensitivity to growth-inhibitory factors was evaluated in proliferation assays with recombinant TNF-a (10 ng/mL; R&D Systems, Minneapolis, MN), TRAIL (5-50 ng/mL; R&D Systems), or IFN-a (1,000 or 10,000 IU/mL; PBL Biomedical Laboratories, Piscataway, NJ).
Real-time PCR.Quantitative PCR assays were done in triplicates using the SYBR Green I on a LightCycler apparatus (Roche Diagnostics, Indianapolis, IN).The PCR primer pairs specific for IFN-a1/13, IFN-a2, and IRF-7 cDNAs were described previously (28).Other primers were as follows: IFN-b, 5 ¶-TTGTGCTTCTCCACTACAGC-3 ¶ ( forward) and 5 ¶-CTGTAAGTC-TGTTAATGAAG-3 ¶ (reverse); TNFSF10F (TRAIL), 5 ¶-TCCTGGGAATCAT-CAAGGAG-3 ¶ ( forward) and 5 ¶-ACTAAAAAGGCCCCGAAAAA-3 ¶ (reverse); and b-actin, 5 ¶-CCTTCCTGGGCATGGAGTCCT-3 ¶ ( forward) and 5 ¶-AATCT-CATCTTGTTTTCTGCG-3 ¶ (reverse).All data are presented as a relative quantification with efficiency correction based on the relative expression of target genes versus b-actin as reference gene.Standard curves and PCR efficiencies were obtained using serial dilutions of pooled cDNA prepared from macrophages or dendritic cells infected with Sendai virus.To analyze vascular endothelial growth factor (VEGF) mRNA levels, Taqman Gene Expression Assays (Applied Biosystems, Foster City, CA) were used.cDNA/ reaction (25 ng) was mixed in duplicate with Taqman Universal PCR Master Mix (Applied Biosystems) and then amplified using predesigned primers and probes specific for VEGF (Hs00173626_m1) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Hs99999905_m1).GAPDH was used Figure 1.Adenovirus construction and effects of IRF-3 5D and IRF-7 D247-467 on gene transcription, antiviral response, and cell viability in A549 lung epithelial tumor cells.A, cDNA encoding constitutively active GFP-IRF-3 5D (S396D/S398D/ S402D/T404D/S405D) and GFP-IRF-7 D247-467 were subcloned into the pAdPS-CMV5-CuO adenovirus transfer vector.GFP fusion proteins were expressed under the control of the human CMV promoter (CMV5).The cymene bacterial operator (CuO) was inserted downstream of the TATA box of the CMV5 promoter, such that this promoter is negatively regulated by the binding of the 28-kDa CymR repressor.The pAdPS-CMV5-CuO vector also contained (a) the adenovirus PS gene expressed under the control of the adenovirus major late promoter (MLP ) and (b ) parts of the Ad5 genome, including the left inverted terminal repeats (ITR ) sequence and an Ad5 genomic region sequence (adenovirus; Ad) for homologous recombination with genomic DNA from an infectious Ad5 virus.Recombinant adenoviruses were generated into 293 cells stably expressing CymR using the positive selection method by transfection with a pAdPS-CMV5-CuO-derived transfer vector, containing the PS gene, and infection with an Ad5 DE1, DE3, and DE4 (orf6) vector deleted in the PS gene.PS is dispensable for adenovirus infection but mandatory for adenovirus replication.The transfer vector was linearized by a single enzymatic digestion before transfection and the homologous recombination targeted its adenovirus sequence.Viruses were efficiently amplified in 293 CymR + cells because the E1 function necessary for virus replication was complemented in 293 cells and the expression and possible toxicity of the transgene was repressed by CymR.Transgene expression was CMV5 driven in any other cell types transduced with subsequent recombinant adenovirus vectors.B, gene transcription.A549 cells were transfected with plasmids harboring luciferase reporter genes under the control of IFNB, IFNA4 , or RANTES (nB mutated) promoters.Five hours later, cells were transduced or not with Ad-GFP, Ad-IRF-3, or Ad-IRF-7 at 10 or 100 MOI.Luciferase activity was analyzed 36 hours after transduction.Luciferase activity (RLU1 ) was normalized by cotransfection with Renilla luciferase (RLU2 )-encoding plasmid.C, antiviral activity.A549 cells were transduced or not with Ad-GFP (GFP), Ad-IRF-3 (IRF-3), or Ad-IRF-7 (IRF-7) at 50 MOI for 48 hours and then infected (last four right lanes ) or not (first four left lanes ) with WT VSV-HR (Indiana strain).Whole-cell extracts were prepared 24 hours later, and samples were run on a 10% SDS-PAGE and subsequently subjected to immunoblot analysis of VSV-G, VSV-N, and VSV-M protein expression and h-actin, as a control.D, apoptosis.A549 cells were transduced at 50 MOI for 16, 24, 32, and 48 hours.All cells were harvested after trypsinization, washed with PBS, and incubated with allophycocyanin (APC )-conjugated Annexin V and analyzed by flow cytometry for GFP expression and Annexin V binding.
as an endogenous control for normalization and the nontransduced samples were used as calibrator controls.Data were then collected using the AB 7500 Real-time PCR System (Applied Biosystems) and analyzed by comparative C T method using the SDS version 1.3.1 Relative Quantification software.
ELISA.To evaluate the concentration of cell-associated TRAIL, macrophages were transduced or not with adenovirus vectors for 24 hours.Total intracellular TRAIL levels were measured in cell pellets using a commercial ELISA kit from R&D systems according to the manucfacturer's instructions.IFN-a levels in supernatants were measured using a commercial ELISA kit from PBL Biomedical Laboratories according to the manufacturer's instructions.
Proliferation assay.Cytostatic activity of macrophages was evaluated on SK-BR-3, MCF-7, and COLO-205 tumor cell lines.Tumor cells (10 4 per well) were seeded in 96-wells plates.Macrophages were added to obtain final E:T ratios of 0.1:1, 0.3:1, 1:1, and 3:1.After 48 hours of culture at 37jC, 1 ACi [ 3 H]thymidine was added in each well.After an additional 24 hours, plates were centrifuged, supernatant was discarded, and 100 AL of 0.1 N NaOH were added to each well.Plates were then harvested on filters with Wallac (Turku, Finland) cell harvester and filters were read in a Wallac h-counter.The percentage of inhibition of proliferation was calculated as follows: [(P À E) / P] Â 100, where P is the average cpm obtained from wells with target cells only (proliferation of tumor cells) and E is the cpm obtained from experimental wells.Mean percentage of inhibition of proliferation and SD were calculated from triplicate experimental wells.To test if type I IFNs were involved in macrophages activity, mouse anti-IFN-a/h receptor (IFN-a/hR) chain II antibody (5 Ag/mL; PBL Biomedical Laboratories) or control mouse IgG1 (Becton Dickinson) was added to the wells.Percentage of blocking was calculated as follows: [1 À (T / C)] Â 100, where T is the mean percentage of inhibition of proliferation in the presence of the blocking antibody and C is the mean percentage of inhibition of proliferation in the presence of IgG1 control antibody.

Results
Activity of transduced IRF-3 and IRF-7 in lung epithelial cells.To evaluate the effect of IRF-3 and IRF-7 on immune cell functions, recombinant adenoviruses were generated for the expression of constitutively active forms of human IRF-3 or IRF-7.Active mutants of IRF-3 and IRF-7 chosen for these studies were based on our previous analysis (16,18).cDNA encoding GFPq/IRF-3 5D or GFPq/IRF-7 D247-467 fusion proteins were subcloned into the pAdPS-CMV5-CuO adenovirus transfer vector (Fig. 1A).This vector harbors the PS gene, which is dispensable for adenovirus infection but required for replication.The vector also contains the cymene bacterial operator (CuO) inserted downstream of the TATA box of the optimized CMV5 promoter and is negatively regulated by the binding of the CymR repressor (20).Transgene expression is minimal in 293 cells stably expressing CymR (used for adenovirus recombination and amplification), whereas CMV5 driven in any other cell type (21).Recombinant adenoviruses were generated using the positive selection method by transfection with the transfer vectors and infection with an Ad5 DE1, DE3, and DE4 (orf6) virus deleted in the PS gene (Fig. 1A) as detailed previously (22,23).
The transcriptional properties of IRF-3 5D and IRF-7 D247-467 delivered by adenovirus vectors (termed Ad-IRF-3 and Ad-IRF-7) were initially assessed in A549 lung epithelial tumor cells transfected with plasmids harboring luciferase reporter genes under the control of IFNB, IFNA4, or RANTES promoters.Expression of IRF-3 5D induced transcriptional activation of the IFNB and RANTES promoters, but not of the IFNA4, as shown previously (19), whereas expression of IRF-7 D247-467 strongly activated all promoters (Fig. 1B).Next, induction of antiviral activity by IRF-3 5D and IRF-7 D247-467 was evaluated in adenovirus-transduced A549 cells that were infected subsequently with VSV.Immunoblot analysis of VSV proteins as a measure of virus replication revealed that expression of either IRF-3 5D or IRF-7 D247-467 abrogated VSV replication (Fig. 1C).The antiviral effect correlated with type I IFN production because supernatants from transduced cells inhibited VSV replication, an effect that was partially abrogated by the addition of neutralizing antibodies specific for IFN-a and IFN-h (data not shown).
Previous studies have suggested that expression of IRF-3 and activation of the antiviral response contributed to cell death (29,30).Transduction with Ad-IRF-3 or Ad-IRF-7 for 48 hours led to a >2-fold decrease in the number of live A549 cells compared with transduction with green fluorescent protein (GFP) only (data not shown).Flow cytometry analysis of Annexin V binding showed 53% and 20% apoptotic A549 cells at 32 hours after transduction with Ad-IRF-3 or Ad-IRF-7 at 50 MOI, respectively, whereas only 7% apoptosis was detected in noninfected or GFP-expressing cells (Fig. 1D).The number of GFP-expressing and Annexin V-binding cells decreased progressively, with the accumulation of dead cells (Fig. 1D).Cell death induced by Ad-IRF-3 or Ad-IRF-7 was not caused by IFN-a alone because A549 cells were resistant to IFN growth inhibition (data not shown).Altogether, these data suggest that transduction of active IRF-3 or IRF-7 stimulated the production of type I IFN, induced an antiviral response, and increased the apoptotic response in A549 cells.
Expression of active IRF-3 in primary macrophages induces cell death.Primary macrophages and IFN-g-activated macrophages (MAK) were elutriated from 7-day cultures of healthy donor peripheral blood mononuclear cells.Hematopoietic cells of various lineages express low cell surface levels of the common coxsackie/ adenovirus receptor and a 5 integrins and are often poorly infected with adenovirus vectors compared with many established tumor cell lines (24).However, IRF-3 and IRF-7 transgene expression was readily detected by immunoblot following transduction with Ad-IRF-3 or Ad-IRF-7 (Fig. 2A), indicating that macrophages and MAK cells were susceptible to infection.Next, macrophages and MAK cells from two donors were transduced with adenovirus vectors, and transduction efficiency and apoptosis were monitored by flow cytometry 48 hours later.No significant difference in cell death between nontransduced macrophages and MAK cells isolated from the same donor was observed, suggesting that the treatment of macrophages with IFN-g did not contribute to increased cell death (Fig. 2B; data not shown).Expression of Ad-IRF-3 induced apoptosis in macrophages or MAK cells (Fig. 2B), whereas cell death related to the expression of Ad-IRF-7 was significantly lower; MAK cells from one donor displayed 42% and 8% Annexin V + cells after transduction with Ad-IRF-3 or Ad-IRF-7, respectively, whereas the percentages of GFP + cells were 58% and 79%, respectively (Fig. 2B).Differences in the fate of IRF-3-or IRF-7expressing cells were also observed after transduction of three other preparations of macrophages or MAK cells (see below; data not shown) or two preparations of fresh monocyte-derived dendritic cells (data not shown).Cell death induced by Ad-IRF-3 was observed as early as 20 hours after transduction.These data show a differential response of primary macrophages to IRF expression, with rapid cell death induced by IRF-3 and only low level cell death in IRF-7-expressing cells.
Profile of type I IFN, TRAIL, and IRF-7 mRNA induced by active IRF-3 and IRF-7 in macrophages.Next, the profile and expression of type I IFN mRNA were examined by quantitative RT-PCR after Ad-IRF transduction of primary macrophages.Ad-IRF-3 induced low levels of IFN-b (Fig. 3A) and IFN-a2 mRNA (Fig. 3B) but no detectable IFN-a1 mRNA.In contrast, increased IFN-b, IFN-a1, and IFN-a2 mRNA levels were observed following transduction with Ad-IRF-7.Levels of IFN-a1 and IFN-a2 transcripts were enhanced >50-fold when macrophages were infected with Ad-IRF-7 at 150 MOI versus 50 MOI.Quantification of secreted IFN-a levels by ELISA showed that macrophages from this donor produced 24.6 F 3.0 ng/10 6 and 1.6 F 0.0 ng/10 6 cells 48 hours after transduction with Ad-IRF-7 at 150 MOI versus 50 MOI.IFN-a levels were below detectability in supernatants from Ad-IRF-3-infected cells or control cells.The quantitative difference in expression of IFN-a transcripts and protein in Ad-IRF-7-transduced macrophages at 150 MOI versus 50 MOI did not correlate with the number of GFP + cells (64% and 55% of total macrophages, respectively), suggesting that higher MOI may induce more transcripts per cell.Data obtained from multiple donors showed that IRF-7 selectively enhanced the transcription of IFNB and IFNA genes in primary macrophages (data not shown).
Next, expression levels of the IFN-a/h target genes IRF7 and TNFSF10F (TRAIL) were examined.Transduction with Ad-IRF vectors induced transcription of IRF7 and TRAIL genes, although Ad-IRF-7 was more potent than Ad-IRF-3 in up-regulating TRAIL mRNA levels (Figs.3B and 4A).Cellular TRAIL accumulation was detected by ELISA in Ad-IRF-7-transduced or recombinant IFN (rIFN)-a2b-treated macrophages (Fig. 3C), indicating that TRAIL release may contribute to macrophage effector functions.Regulation of IRF7 and TRAIL gene expression by active IRF-7 was subsequent to type I IFN induction because a significant decrease in IRF-7 or TRAIL mRNA levels was quantified by RT-PCR in the presence of neutralizing antibodies specific for the IFN-a/hR chain II (Fig. 3D).Thus, active IRF-7 controls the expression of endogenous IRF7 and TRAIL genes through the production of type I IFN in macrophages.
Up-regulation of TRAIL, IL-15, and CD80 by Ad-IRF in macrophages.Direct tumor cell killing by macrophages can be exerted through the production of TNF-a, IL-1, NO (regulated by iNOS), FasL, or TRAIL ( for review, see ref. 1).In addition, tumor antigen presentation to T cells by macrophages can be optimized by expression of costimulatory molecules, such as CD80, and the secretion of IL-15 or IL-12 (p35), T helper 1 (Th1)-polarizing ILs (31,32).Expression of some of these immunomodulatory genes was analyzed by RT-PCR in Ad-IRF-transduced macrophages or MAK cells from three different donors (Fig. 4A; Supplementary Fig. S1A).Up-regulation of TRAIL, IL-15, IL-12 p35, ISG-56, and CD80 mRNA levels was detected in both Ad-IRF-3-and Ad-IRF-7transduced cells (Fig. 4A), whereas TNF-a, FasL, iNOS, IL-6, IL-1a, and IL-1b mRNA levels were unaffected by IRF expression (Fig. 4A; Supplementary Fig. S1A).MAK cells showed an up-regulation of iNOS mRNA levels; however levels of other ISGs were comparable with macrophages (Fig. 4A).This expression pattern is likely due to the transient effect of IFN-g activation on MAK cells because macrophages activated for 18 hours but not 48 hours with IFN-g display increased levels of TRAIL or ISG-56 mRNA (Fig. 4B).In contrast, up-regulation of ISG-56, TRAIL, and IL-15 mRNA was observed in Ad-IRF-7-expressed macrophages at 48 hours (Fig. 4B) and 4 days (Supplementary Fig. S1B).
Constitutively active IRF-7 increases cytostatic properties of macrophages.Given that tumoricidal factors were stimulated in Ad-IRF-transduced cells, the antitumor activity of macrophages against different human tumor cell lines was evaluated.SK-BR-3 (breast cancer cells), MCF-7 (breast cancer cells), and COLO-205 (colon cancer cells) were tested for their sensitivity to IFN-a2b, TRAIL, and TNF-a, with a cell proliferation assay measured by [ 3 H]thymidine incorporation.As shown in Fig. 5A, all tumor cells displayed sensitivity to IFN-a2b (moderate for COLO-205) and TNF-a, and MCF-7 and COLO-205 cells were sensitive to growth inhibition by TRAIL.The effect of MAK on these two breast cancer cell lines could be inhibited by anti-TNF-a-neutralizing antibodies, suggesting a role for TNF-a in MAK cells activity (ref.25; data not shown).In contrast, COLO-205 cells were insensitive to inhibition by MAK cells (data not shown).
In light of these observations, the effects of IRF-3 and IRF-7 on the in vitro inhibitory activity of macrophages were tested against SK-BR-3 and COLO-205 cells in cocultures of macrophages and tumor cells at different E:T ratios.With one donor, 55% and 64% of total macrophages were GFP + 48 hours after transduction with Ad-IRF-7 at 50 or 150 MOI, respectively (Fig. 5B).Cell viability was 87% and 70.5%, respectively, comparable with viability of nontransduced cells or GFP-tranduced cells (89% and 85%, respectively).In contrast, cell viability was significantly decreased in IRF-3 5D-expressing macrophages, 59% at 24 hours and 26% at 48 hours after transduction.Control or GFP-expressing macrophages were unable to inhibit the growth of SK-BR-3 cells (Fig. 5B), whereas Ad-IRF-3-transduced macrophages displayed weak inhibition of SK-BR-3 cell growth and had either no effect or a stimulatory effect on COLO-205 cell proliferation (>40% increase in proliferation at ratio 0.3:1).In contrast, Ad-IRF-7-transduced macrophages inhibited the proliferation of both SK-BR-3 and COLO-205 cells.Significant growth inhibition (>35%) was observed at ratios as low as 0.3:1 (Fig. 5B).Supernatants obtained from macrophages 48 hours after transduction with Ad-IRF-7, but not with Ad-GFP, inhibited the growth of SK-BR-3 and COLO-205 cells as evaluated by a tetrazolium compound-based viability assay (data not shown).Furthermore, the antitumor effect of Ad-IRF-7-transduced macrophages was partially inhibited by addition of anti-IFN antibody (Fig. 5C), indicating that IFN-a/h secretion and other mechanisms may be involved in tumor inhibition by macrophages.These results were confirmed with macrophages from another donor (Fig. 5D).Despite a low adenovirus transduction efficiency, inhibition of SK-BR-3, MCF-7, and COLO-205 tumor cell growth was observed in cocultures with Ad-IRF-7-transduced macrophages (Fig. 5D).Importantly, pretreatment with rIFNa-2b for 24 hours did not increase their tumoricidal effect on any of the tumor cells (Fig. 5D).Thus, the constitutively active form of IRF-7 is able to increase the antitumor effector functions of macrophages.

Discussion
In the present study, we show that the antitumor properties of primary macrophages are stimulated by ectopic expression of a constitutively active form of IRF-7, delivered using a regulated adenovirus vector system.Expression of active IRF-3 5D by primary human macrophages resulted in rapid cell death, whereas expression of IRF-7 D247-467 was not as proapoptotic and induced long-lasting up-regulation of type I IFN genes.In addition, macrophages transduced with IRF-7 D247-467 displayed enhanced tumor inhibitory effects that were mediated in part by induction and secretion of type I IFNs.
Transduction of Ad-IRF-3 induced rapid apoptosis in macrophages (Fig. 2B), whereas Ad-IRF-7 induced low level apoptosis, observed mainly when macrophage preparations were highly transduced.Similar responses were also observed in primary monocyte-derived dendritic cells (data not shown).Cell death induced by IRF-3 or IRF-7 could be mediated by IFN-a/h because IFN-h is known to up-regulate p53 expression and thereby the apoptotic cellular response (33).However, the cell death observed in the present study is not solely associated with IFN-a/h expression because Ad-IRF-7 was a more potent inducer of IFNA and IFNB transcription compared with Ad-IRF-3 (Fig. 3A and B).IRF-3 and IRF-7 harbor distinct DNA binding recognition sequences and other transcriptional properties, such as the recruitment of transcriptional coactivators, and may therefore .Gene expression profiles in Ad-IRF-transduced primary human macrophages.A, primary macrophages (Mac) or MAK cells (donor CP116) were transduced or not with Ad-GFP (GFP), Ad-IRF-3 (IRF-3), or Ad-IRF-7 (IRF-7) at 50 MOI, washed, and plated.At 24 hours after transduction, total RNA was prepared from adherent live cells and processed for RT-PCR analysis of mRNA levels of iNOS, TNF-a, TRAIL, CD80, IL-12 p35, IL-15, and IL-1b using ISG-56 as an ISG control and b-actin, as a control.Cycles (20, 25, 30, and 35) were carried out for each amplification and PCR products were analyzed by electrophoresis on 2% agarose gels.Ethidium bromide fluorescence of PCR products before reaching saturation levels.Negative controls included PCRs done with RNA samples not treated (À) with reverse transcriptase.B, persistence of expression of ISGs.Macrophages (donor 353) were transduced or not with adenovirus vectors at 100 MOI, washed, and plated or treated with rIFNa-2b or IFN-g at 166 units/mL.At 18 hours, cells were washed twice and fresh medium was added.Total RNA was prepared from adherent live cells at 18 and 48 hours and processed for RT-PCR analysis of b-actin, TRAIL, and ISG-56 levels as described in (A ).C, macrophages or MAK cells (donor CP116) were transduced or not with adenovirus vectors at 50 MOI, washed, and plated.At 24 hours, total RNA was prepared from adherent live cells and processed for RT-PCR analysis of MMP-2, IL-8, and b-actin mRNA levels as described in (A).D, macrophages or MAK cells (donor SA6088) were transduced or not with adenovirus vectors at 50 MOI, washed, and plated.At 24 hours, total RNA was prepared from adherent live cells and processed for quantitative real-time PCR analysis of VEGF mRNA levels.Data on gene expression are representative of experiments done with cells from three different donors, with the exception of (B ).

Cancer Research
Cancer Res 2006; 66: (21).November 1, 2006 differentially regulate expression of apoptotic genes ( for review, see ref. 8).Studies are under way to investigate IRF-3 or IRF-7 target genes in the apoptotic response in primary macrophages or dendritic cells.
Gene expression profile analysis indicated that stable expression of IRF-7 WT into the BJAB B cell line does not induce IFNB or IFNA transcription but up-regulates the expression of IRF-8 mRNA.Virus infection of IRF-7 WT/BJAB cells leads to the increase of many genes, including immune genes, such as IFN-b, IFN-a, and CD80encoding genes (15,16,26,34).It was reported that IRF-3 can bind directly to ISRE elements of the TRAIL gene promoter, whereas no direct evidence for IRF-7 binding was observed (35).TRAIL transcription is also induced in many cell types by treatment with IFN-a or IFN-g, probably through the binding of ISGF3 (36) or IRF-1 (37), respectively.We show here that the classic IFNdependent target genes encoding IRF-7, TRAIL, IL-12 p35, and CD80 are stimulated by both Ad-IRF-3 and Ad-IRF-7.Neutralization of the type I IFN receptor inhibited IRF-7-induced TRAIL mRNA up-regulation in macrophages (Fig. 3D).Expression levels of IFN-b or IFN-a mRNA and protein were, however, not strictly correlated with ISG expression in adenovirus-transduced macrophages.In general, cells display low levels of type I and type II IFN receptors (100-1,000 molecules on the cell surface), but these receptors are very efficient for signal transduction.This could account for the similar transcriptional induction of ISG, such as IRF7 or TRAIL, in macrophages transduced with Ad-IRF-7 at different MOI (Fig. 3B).
Expression of iNOS was not up-regulated by Ad-IRF-7 in primary macrophages.Other reports showed that iNOS is induced by IFN-g priming of macrophages by an IRF-1-dependent mechanism (38); however, the role of NO in the antitumor effects of human macrophages is not clearly established.As expected, expression of IRF-3 or IRF-7 did not change expression levels of IL-1 or TNF-a mRNA (Fig. 4A), and anti-TNF-a-neutralizing antibodies had no effect on the growth of tumor cells in coculture with macrophages (data not shown).Our previous work showed that IFN-g-activated macrophages produced in large scale for clinical application exhibited increased antitumor activity in vitro partly due to TNF-a production (25,39).Ongoing studies will compare antitumor properties of macrophages or MAK cells after Ad-IRF-7 transduction.
Macrophages and tumor cells are known to release protumorigenic factors, which constitute a major drawback for macrophage adoptive transfer in cancer therapy.Recombinant IFN-a inhibits expression of IL-8, VEGF, MMP-2, and MMP-9 by tumor cells in vitro and in vivo (40,41).Molecular mechanisms underlying these effects are not well characterized, although IFN signaling Figure 5. Antitumor properties of Ad-IRF-7-transduced primary human macrophages.A, SK-BR-3, MCF-7, and COLO-205 tumor cell sensitivity to TNF-a, TRAIL, and IFN-a.Tumor cells (10 4 ) were seeded in 96-well plates in the presence or absence of recombinant TNF-a (10 ng/mL), TRAIL (50 ng/mL), or rIFN-a2b (10,000 IU/mL).Twenty-four hours later, 1 ACi [ 3 H]thymidine was added to each well.Plates were harvested 24 hours later and radioactivity was counted to determine the percentage of inhibition of tumor cell proliferation.B, tumor inhibition by macrophages.Fresh macrophages from donor HD594 were transduced or not (gray squares ) with Ad-GFP (black squares ), Ad-IRF-3 (gray circles), and Ad-IRF-7 (black circles ) at 50 (IRF7-50 ) or 150 (IRF7-150 ) MOI.Alternatively, nontransduced macrophages were treated with 1,000 IU/mL rIFN-a2b (black pentagon ) for 24 hours in (D ).Macrophages transduced at 50 MOI were harvested 24 hours later and seeded in 96-well plates together with tumor cells (10 4 ) at E:T ratios of 0.1 or 0.3:1 to 3:1.[ 3 H]thymidine (1 ACi) was added to each well 48 hours later.Plates were harvested 24 hours later and radioactivity was counted to determine the percentage of inhibition of tumor cell proliferation.Some macrophages were harvested 48 hours after adenovirus transduction and fixed for flow cytometry analysis of GFP expression.C, inhibition of type I IFN receptor.Neutralizing anti-type IFN chain II receptor antibody (5 Ag/mL; plain line ) or control mouse IgG1 (dotted line ) was added to cocultures of Ad-IRF-7-transduced macrophages (50 MOI) and tumor cells.D, tumor inhibition by macrophages.Fresh macrophages from another donor (HD598) was transduced and analyzed as described in (B).Data were analyzed using two-way ANOVA and Bonferroni post-tests to compare treatments against the nontransduced control (B and D ) or IgG-treated control (C ).*, P < 0.05 (5%); **, P < 0.01 (1%); ***, P < 0.001 (0.1%).

IRF-7
Increases Antitumor Properties of Macrophages may create competition for transcriptional modulators that bind preferentially to ISG promoters and decrease the expression of other genes (42).Interestingly, we observed that the transcription of proangiogenic and metastatic genes (such as VEGF) was downregulated by expression of IRF-7 active mutants in macrophages (Fig. 4C and D), suggesting that IRF-7 may increase macrophage antitumor properties while reducing their protumorigenic effects.
Previous studies have explored the role of type I IFNs and IRF-7 in antitumor immunity and macrophage activation (43).Okada et al. reported that intratumoral delivery of IFN-a-transduced mouse dendritic cells leads to the induction of tumor-specific CTLs with effector function mediated by perforin, FasL, and TRAIL.Other studies showed that tumor cell rejection by the immune system was impaired in IFN-a/h-unresponsive mice (44).Immunomodulating activities of IFN-a/h favor the differentiation of monocytes into dendritic cells (45)(46)(47), the activation of dendritic cells (48) and NK cells (49), the cross-priming of CD8 + T cells, and Th1-biased responses by T and B cells (50,51).On the other hand, overexpression of IRF-7 WT was shown to support the macrophage differentiation of the U937 monocytic cell line, a process that could not be reproduced by treatment with rIFN-a (52).No clear data were obtained on the effect of IRF-7 on primary monocytes; however, their differentiation into dendritic cells is observed after exposure to granulocyte macrophage colonystimulating factor and IFN-a, which is likely to up-regulate IRF-7 mRNA levels (53,54).
In vivo, primary macrophages transduced with Ad-IRF-7 may mediate their antitumor effects by four distinct mechanisms (Fig. 6

Figure 3 .
Figure 3. Profile of type I IFNs and expression of IRF7 and TRAIL in Ad-IRF-transduced primary human macrophages.Macrophages (donor HD594) were transduced or not with Ad-GFP (GFP), Ad-IRF-3 (IRF-3), or Ad-IRF-7 (IRF-7) at 50 MOI (GFP-50, IRF3-50 , and IRF7-50 ) or 150 MOI (IRF7-150 ) or treated with rIFNa-2b at 1,000 IU/mL (B and C ) or 100 IU/mL (D).A, expression levels of IFN-b mRNA.Total RNA was prepared from fresh macrophages 24 hours later and cDNA samples were subjected to PCR analysis with primers specific for IFN-b or b-actin, as a control.Ethidium bromide fluorescence of PCR products before reaching saturation levels.Negative controls included PCRs done with RNA samples not treated (À) with reverse transcriptase (RT).B, expression levels of IFN-a1/13, IFN-a2, IRF-7, and TRAIL mRNA.Total RNA was prepared from fresh macrophages 24 hours and/or 48 hours later and subjected to real-time RT-PCR analysis.Data are presented as the relative expression of target genes versus b-actin as reference gene.C, concentration of cell-associated TRAIL.Cell pellets were collected 24 hours later and processed for ELISA analysis of TRAIL.Data were compared with the GFP-transduced control by a Student's two-tailed impaired t test.*, P < 0.05 (5%).D, inhibition of IRF7 and TRAIL transcription by neutralization of the type I IFN receptor signaling.Adenovirus-transduced or rIFNa-2b-treated macrophages were plated in the presence of 20 Ag/mL of a neutralizing rabbit antibody specific for type I IFN receptor chain II or rabbit IgG as a control.At 24 hours, total RNA was prepared from adherent cells and processed for real-time RT-PCR analysis of TRAIL and IRF-7 mRNA levels as in (B).

Figure 4
Figure 4. Gene expression profiles in Ad-IRF-transduced primary human macrophages.A, primary macrophages (Mac) or MAK cells (donor CP116) were transduced or not with Ad-GFP (GFP), Ad-IRF-3 (IRF-3), or Ad-IRF-7 (IRF-7) at 50 MOI, washed, and plated.At 24 hours after transduction, total RNA was prepared from adherent live cells and processed for RT-PCR analysis of mRNA levels of iNOS, TNF-a, TRAIL, CD80, IL-12 p35, IL-15, and IL-1b using ISG-56 as an ISG control and b-actin, as a control.Cycles(20, 25, 30, and 35)  were carried out for each amplification and PCR products were analyzed by electrophoresis on 2% agarose gels.Ethidium bromide fluorescence of PCR products before reaching saturation levels.Negative controls included PCRs done with RNA samples not treated (À) with reverse transcriptase.B, persistence of expression of ISGs.Macrophages (donor 353) were transduced or not with adenovirus vectors at 100 MOI, washed, and plated or treated with rIFNa-2b or IFN-g at 166 units/mL.At 18 hours, cells were washed twice and fresh medium was added.Total RNA was prepared from adherent live cells at 18 and 48 hours and processed for RT-PCR analysis of b-actin, TRAIL, and ISG-56 levels as described in (A ).C, macrophages or MAK cells (donor CP116) were transduced or not with adenovirus vectors at 50 MOI, washed, and plated.At 24 hours, total RNA was prepared from adherent live cells and processed for RT-PCR analysis of MMP-2, IL-8, and b-actin mRNA levels as described in (A).D, macrophages or MAK cells (donor SA6088) were transduced or not with adenovirus vectors at 50 MOI, washed, and plated.At 24 hours, total RNA was prepared from adherent live cells and processed for quantitative real-time PCR analysis of VEGF mRNA levels.Data on gene expression are representative of experiments done with cells from three different donors, with the exception of (B ).
): (a) directly, via secretion of type I IFN ( for type I IFNsensitive tumors); (b) after activation of macrophages by either IRF-7 or type I IFN, enhancing their effector functions; (c) indirectly, via recruitment and polarization of other immune cells by type I IFN or other macrophage-derived factors, including chemo-kines; and (d) via down-regulation of expression of genes known to promote metastasis and angiogenesis.Our RT-PCR and ELISA data suggest that type I IFN and TRAIL are major tumoricidal effector molecules up-regulated in Ad-IRF-7-transduced macrophages.Supporting this hypothesis, inhibition of tumor cell proliferation by IRF-7-transduced macrophages or supernatants was partially blocked by anti-IFN-a/hR antibody (Fig.5C).Immunorelated genes up-regulated by IRF-7 in macrophages included IL-12 p35, CD80 and IL-15, a cytokine-stimulating NK cells and memory T cells (Fig.4A).Importantly, modulation of gene expression was relatively long lasting (up to 4 days), which may constitute an advantage compared with simple IFN-a pretreatment of macrophages.The present results show that macrophages expressing constitutively active IRF-7 are attractive candidates for in vivo cancer therapy via adoptive transfer.In vivo delivery of IRF-7-expressing macrophages to the tumor site may result in the release of type I IFNs in the local tumor environment and in the enhancement of macrophage effector functions.These immune modulatory effects would sustain direct tumor cell killing by macrophages as well as favor the induction of Th1-biased antitumor immune responses.

Figure 6 .
Figure 6.Proposed mechanisms of the antitumor effects of Ad-IRF-transduced macrophages.The antitumor properties of Ad-IRF-transduced macrophages may be mediated by four mechanisms. 1, secretion of type I IFN, for IFN-sensitive tumors; 2, activation of macrophages by IRF-7 or type I IFN, leading to the up-regulation of effector molecules, such as TRAIL; 3, modulation of the tumor microenvironment and polarization of other immunes cells as T and NK cells induced by type I IFN or other macrophage-derived factors (cytokines as IL-12 and IL-15; chemokines as macrophage inflammatory protein-1b, RANTES, and IP-10; and costimulatory molecules as CD80); 4, down-regulation of production of proangiogenic and metastatic factors (MMP-2 and VEGF).Red and green boxes, enclose factors up-regulated and down-regulated, respectively, by macrophages transduced with Ad-IRF.Arrows, induction; bars, inhibition.