Journal list menu
Interferon regulatory factor-1 is prerequisite to the constitutive expression and IFN-γ-induced upregulation of B7-H1 (CD274)
Abstract
Majority of cancer cells upregulate co-inhibitory molecule B7-H1 which confers resistance to anti-tumor immunity, allowing cancers to escape from host immune surveillance. We addressed the molecular mechanism underlying the regulation of cancer-associated B7-H1 expression in response to interferon-γ (IFN-γ). Using promoter constructs in luciferase assay, the region between 202 and 320 bp from the translational start site is responsible for B7-H1 expression. Electrophoretic mobility shift assay, site-directed mutagenesis and knockdown experiment using siRNA revealed that interferon regulatory factor-1 (IRF-1) is primarily responsible for the constitutive B7-H1 expression as well as for the IFN-γ-mediated B7-H1 upregulation in a human lung cancer cell line A549. Additionally, AG490, a Janus activated kinase/signal transducer and activator of transcription inhibitor, greatly abolished the responsiveness of A549 cells to IFN-γ by reducing the IRF-1 transcription. Our findings support a critical role of IRF-1 in the regulation of constitutive and IFN-γ-induced expression of B7-H1 in cancer cells.
1 Introduction
B7-H1 (CD274, PD-L1), a member of B7 family, is a co-signaling molecule that could be stimulatory or inhibitory for T cell immunity through engagement of different receptors expressed on T cells [1]. Small fraction of antigen-presenting myeloid dendritic cells and Kupffer cells in liver express a low level of endogenous B7-H1 on the surface. However, since the B7-H1 expression in peripheral tissues is tightly regulated in vivo, it is hard to detect B7-H1 expression in a variety of normal tissues through immunohistochemical analysis [2, 3]. Unlike normal cells of epithelial origin, human cancer cells, including lung cancer, ovarian cancer, and melanoma, express constitutively B7-H1 on the surface, and its expression is greatly upregulated by interferon-γ (IFN-γ) [4].
IFN-γ is a pro-inflammatory cytokine that is produced mainly by activated type 1 CD4+ helper T cells (Th1), CD8+ cytotoxic T lymphocytes (CTLs), macrophages, and natural killer (NK) cells, and regulates anti-viral and anti-tumor immunity [5]. Collectively, in cancer microenvironment, cancer-specific or non-specific immune cells that are recruited into cancer sites induce inflammation, a cellular process involving the release of a wide variety of inflammatory mediators including IFN-γ [6], subsequently ensuing induction of various cellular proteins such as B7-H1 in cancer cells.
Overwhelming data indicate that cancer-associated B7-H1 in murine cancer model facilitates apoptosis of cancer-reactive CTLs or forms a molecular shield, resulting in suppression of antitumor immunity [4, 7]. In addition to the blockade of cancer-associated B7-H1 using antagonistic B7-H1 antibody, downregulation of B7-H1 expression by genetic or pharmacological methods may offer alternative promising approaches for cancer treatment. To manipulate B7-H1 expression genetically or pharmacologically, it is necessary to understand how its expression is regulated within cancer cells [8, 9]. In a recent publication, Loke et al. [10] reported that B7-H1 expression was enhanced on inflammatory macrophages by classical activation by LPS and IFN-γ in a STAT1/3-dependent manner. However, it still remains unclear what molecular mechanisms are involved in the constitutive and IFN-γ-induced B7-H1 expression in cancer cells.
In this study, we demonstrated that endogenously expressed interferon regulatory factor-1 (IRF-1), a transcription factor, is primarily responsible for the constitutive B7-H1 expression as well as for the early induction of B7-H1 within 45 min after IFN-γ stimulation. Additionally, de novo synthesized IRF-1 through Janus activated kinase (JAK)/signal transducer and activator of transcription (STAT) pathway is involved in the late induction of B7-H1 expression in response to IFN-γ.
2 Materials and methods
2.1 Cell lines and reagents
A549 human lung carcinoma cell line, colorectal carcinoma (Colo201 and WiDr), and HeLa cells were purchased from American Type Culture Collection (Manassas, VA). SNU739 hepatocellular carcinoma cell line was purchased from Korean Cell Line Bank (Seoul, Korea). Cells were grown in RPMI 1640 medium (GIBCO, NY) supplemented with 10% heat inactivated fetal bovine serum (FBS). IFN-γ and AG490 were purchased from R&D Systems (Minneapolis, MN) and Calbiochem (Ra Jolla, CA), respectively. PD98059, SB203580, SP600125, and LY294002 were purchased from Biomol (Plymoth, PA, USA). Rapamycin, PP1, Rottlerin, Staurosporine, and Genistein were purchased from Sigma–Aldrich (St. Louis, MO, USA). pGL3 basic vector, pRL CMV vector and dual luciferase reporter assay kit were purchased from Promega (Madison, WI). Restriction enzymes and modified enzyme were purchased from Promega and TaKaRa (Seta, Japan). Antibodies specific for IRF-1 and IRF-2 were purchased from Cell Signaling Technology (Beverly, MA). [γ-32P]-ATP and poly(dI-dC)2 were obtained from Amersham Biosciences (Piscataway, NJ). PE-conjugated streptavidin was obtained from Jackson ImmunoResearch (West Grove, PA). Anti-human B7-H1 (5H1) was generated in our laboratory, and was described previously [11, 12].
2.2 Flow cytometric analysis
Cells were stained with biotin-conjugated anti-human B7-H1 (5H1), detected with PE-conjugated streptavidin, and analyzed using flow cytometry and CellQuestPro software (BD science, Franklin Lakes, NJ). Cells were washed between the reactions in FACS buffer (1% FCS and 0.1% NaN3 in PBS). B7-H1 expression on the surface was represented as mean fluorescence intensity (MFI).
2.3 Western blot
Whole cell extracts were prepared in the lysis buffer as described previously [13]. Protein extracts were subject to SDS–PAGE (10% gel), transferred to nitrocellulose membranes (Schleicher & Schuell, NH), and probed with antibodies. The respective bands were detected with HRP-conjugated anti-rabbit IgG (Santa Cruz Biotechnology Inc., CA). The amount of chemiluminescence was measured by LAS-3000 SYSTEM (Fuji Photo Film, Japan).
2.4 Preparation of promoter constructs
Human genomic DNA was isolated from human embryonic kidney-293 (HEK-293) cell line using a DNeasy Tissue Kit (Qiagen, GmhH, Germany). The 3-kb of 5′-flanking promoter region was amplified from genomic DNA by PCR using upstream primer, 5′-TGA CTC GAG ACA CAT ATA GGA TGT GAG-3′; downstream primer, 5′-TCT CTC GAG CCC AAA GAA AGG GTG TAG-3′; The amplified 3036 bp fragment was cloned into TA-vector (pTA-3036) (Promega, WI). The luciferase constructs were generated by digesting pTA-3036 DNA with appropriate restriction enzymes mapped on promoter region and by subcloning into pGL3 basic (Promega). Restriction enzymes used for generation of promoter constructs were as follows; BamHI/XhoI (pH1-2405), SspI/BglII (pH1-1101), DraI/BglII (pH1-633), HindIII/HindIII (pH1-202), and PvuII/HindIII (pH1-88). Exceptionally, pH1-457 and pH1-320 plasmids were created by PCR cloning using specific forward primers (pH1-457: 5′-ACG CCG AGC TCA TAA AGG TTA AGG-3′, 457 bp; pH1-320: 5′-ATC TTC GGA GCT CTT CCC GGT G-3′, 320 bp), and the downstream primer used for PCR of 3036 bp fragment. The B7-H1 promoter sequence was analyzed for transcription factor binding sites within the promoter region using a sequence motif search program of GenomeNet (http://motif.genome.jp).
2.5 Transient transfection and luciferase assay
Luciferase assay was performed using Dual Luciferase Reporter Assay System (Promega) according to the manufacturer's instruction. Luciferase assays were carried out using Wallac 1420 multilabeling counter (Perkin–Elmer, CT). All firefly luciferase values were normalized to Renilla luciferase in order to compare the transfection efficiencies. Results were represented as means ± S.D. of a representative experiment performed in triplicate.
2.6 Site-directed mutagenesis
Mutagenesis of IRF-1 in B7-H1 promoter region was generated using a Quick Change site-directed mutagenesis system (Stratagene, CA). The mutated primers used are listed in Table 1 . PCR was performed using two antiparallel primers with the required nucleotide substitutions and the PCR product was used to transform XL1-Blue supercompetent Escherichia coli cells (Stratagene). The sequences of the mutation constructs were confirmed by bi-directional DNA sequencing.
| Oligonucleotides | Sequence (5′ → 3′) | Location |
|---|---|---|
| IRF-1α | ||
| Wild-type | ACT GGA CTG ACA TGT TTC ACT TTC TGT TTC | −291 to −262 |
| Mutant | ACT GGA CTG ACA TGT g T a AC c TTC TGT TTC | |
| IRF-1β | ||
| Wild-type | CTA GAT ACC TAA ACT GAA AGC TTC CGC CGA | −222 to −199 |
| Mutant | CTA GAT ACC TAA g CT t A c AGC TTC CGC CGA |
2.7 EMSA analysis
Sense and antisense oligonucleotides encompassing the IRF-1 binding site in B7-H1 promoter were synthesized (Table 1). Electrophoretic mobility-shift assay (EMSA) probes were made by annealing the oligonucleotides and end-labeling the resulting double-stranded oligonucleotides using T4 polynucleotide kinase and [γ-32P]-ATP. EMSA analysis was performed using the Gel Shift Assay System (Promega) with A549 nuclear extracts, prepared as previously described [14, 15]. For supershift assays, protein extracts were preincubated with 2 μg of anti-IRF-1, anti-IRF-2, and mouse IgG antibody for 4 h at 4 °C prior to binding reaction, and then subject to EMSA assay. The resultant protein/DNA complexes were electrophoretically resolved on 5% polyacrylamide gels in 0.5 × TBE buffer for 3 h at 120 V, and analyzed by autoradiography. The relative band intensity was analyzed using Image Acquisition and Analysis Program (UVP, Cambridge, UK).
2.8 RT-PCR
Total RNAs were isolated using Trizol reagent (Invitrogen, CA), according to the manufacturer's instructions. cDNA was generated using ImProm-II Reverse Transcription System (Promega), which was then amplified by PCR using specific primers for B7-H1 (forward, 5′-GAC CTG AAG GTT CAG CAT AG-3′; reverse, 5′-GTA TCT TGG ATG CCA CAT TT-3′, 613 bp), IRF-1 (forward, 5′-GTA AGG AGG AGC CAG AAA TTG ACA GC-3′; reverse, 5′-CTA CGG TGC ACA GGG AAT GGC CTG-3′, 157 bp), and IFN-γR (forward, 5′-AGT ACC AGA TCA TGC CAC AGG TCC-3′; reverse, 5′-ATC GCT AAC TGG CAC TGA ATC TCG-3′, 465 bp). The various inhibitors were used at working concentration such as AG490 (50 μM), PD98059 (50 μM), SP600125 (25 μM), SB203580 (25 μM), LY294002 (25 μM) Rapamycin (6 nM), PP1 (2 μg/ml), Rottlerin (10 μM), Staurosporine (0.1 μM) and Genistein (10 μM).
2.9 Preparation of small interfering RNA and transfection
Sense and antisense strands of small interfering RNA (siRNA) oligonucleotides directed against IRF-1 mRNA were synthesized as a duplex format (sense, 5′-CCAAGAACCAGAGAAAAGAdTdT-3′; antisense, 5′-UCUUUUCUCUGGUUCUUGGdTdT-3′) [16, 17]. A549 cells were plated at a density of 6 × 105 cells onto 6-well plate. After 24 h, the cells were transfected with various concentrations of IRF-1 siRNA or control siRNA using Lipofectamine 2000 (Invitrogen, CA) according to the manufacturer's protocol.
2.10 Statistical analysis
Student's t test was performed wherever applicable. Mean ± S.D. is shown unless otherwise stated.
3 Results
3.1 B7-H1 is upregulated in A549 cells in response to IFN-γ
RT-PCR analyses showed detectable expressions of constitutive B7-H1 mRNA in most cancer cells tested, including Colo201, WiDr, HeLa, and A549 cells. Exceptionally, SNU739, a hepatoma cell line, expressed higher level of endogenous B7-H1 mRNA. Treatment of the various cancer cells with 100 U/ml IFN-γ for 24 h resulted in the induction of high levels of B7-H1 mRNA expression (Fig. 1 A). This finding was consistent with protein expression as evidenced by FACS analysis (Fig. 1B). We also observed the upregulated B7-H1 expression in other cancer cell lines such as lung cancer cells (A427 and Calu-3), hepatoma (HepG2 and SNU449), and colon cancer cells (LS513) (Fig. S1). We found that the constitutive expression of B7-H1 protein was more likely prominent than B7-H1 mRNA expression in cancer cells. The finding that the same treatment did not seem to affect the expression of the IFN-γ receptor (IFN-γ R) in any of the cancer cells tested suggest that B7-H1 upregulation in response to IFN-γ does not occur by modulation of its receptor. Given the similarity of responses in the different cancer cell lines, A549 was used in this study. To determine the concentration of IFN-γ revealing a maximum expression of B7-H1, we treated A549 cells with various concentrations of IFN-γ for 24 h. B7-H1 mRNA (Fig. 1C) and protein (Fig. 1D) was maximally increased at 100–200 U/ml IFN-γ. As shown in Fig. 1E and F, B7-H1 mRNA was slightly increased after 1 h of IFN-γ treatment and was further increased in a time-dependent manner thereafter. The maximal induction of B7-H1 mRNA and protein occurred at 9–12 and 24 h, respectively. From the experiment to demonstrate a stability of IFN-γ-induced B7-H1 after withdrawal of INF-γ at 24 h of culture, we found that B7-H1 expression was decreased time-dependently at a similar rate to that under IFN-γ stimulation (Fig. S2).
3.2 Cloning and characterization of the B7-H1 promoter
A 3.0-kb fragment of the B7-H1 promoter was amplified from human genomic DNA by PCR and was cloned into a luciferase reporter plasmid, pGL3 basic vector. The luciferase activity of pH1-2405 was induced 3.0 (±0.6)-fold higher following IFN-γ treatment compared to that of untreated control. Reporter constructs containing promoter regions covering from nt −2405 to nt −320 remained highly inducible in response to IFN-γ (Fig. 2 A). There was a dramatic loss of promoter activity in pH1-202 and pH1-88. While these constructs showed greatly reduced luciferase activity, they still displayed B7-H1 promoter activity compared to pGL vector (Fig. 2, inset). Specifically, luciferase activity of pH1-202 construct that contains NF-κB was likely induced by IFN-γ stimulation, but the relative luciferase activity is significantly lower than that of pH1-320. These results suggest that the promoter region between nt −320 and nt −202 is more likely responsible for the basal or IFN-γ-induced B7-H1 promoter activity in A549 cells. A sequence motif search of GenomeNet further identified transcription factor binding sites in that region, including two IRF-1 sites (IRF-1α, -1β) (Fig. 2B).
3.3 IRF-1 binds to the putative binding sites in B7-H1 promoter
To test the binding of IRF-1 to B7-H1 promoter region, we performed EMSA analyses. As shown in Fig. 3 A, IFN-γ increased the protein complex formation between IRF-1 and the DNA probes. The formation of detectable complex was competed out with 50-fold excess of unlabeled cold probe (Fig. 3A). This result provides evidence that IRF-1 is induced by IFN-γ and binds to a specific site in the B7-H1 promoter. Additionally, signal generated by the IRF-1–DNA complex was supershifted in the presence of anti-IRF-1, not anti-IRF-2, antibody, strongly indicating that the nuclear protein complex that binds to the putative IRF-1 binding sequence in B7-H1 promoter region contains IRF-1 protein. Furthermore, a time course analysis showed that maximum binding activity of IRF-1 to B7-H1 promoter appeared in two phases; the first peak of binding activity within 5 min and the second peak at 7 h after IFN-γ stimulation (Fig. 3B).
3.4 IRF-1 is prerequisite to constitutive and IFN-γ-induced B7-H1 transcription
In order to examine the effect of IRF-1 on B7-H1 transcription, we mutagenized the predicted transcription factor binding sites, IRF-1α and IRF-1β, in the pH1-320 promoter construct. Luciferase assay using those mutants revealed that mutation at the IRF-1α and -1β binding sites resulted in a great reduction of constitutive- as well as IFN-γ-induced B7-H1 promoter activity by ∼80–90% (Fig. 4 A), implying the importance of these cis-acting elements and their involvement in B7-H1 promoter activation. This notion was further supported by the finding that IRF-1 siRNA significantly decreased the IFN-γ-mediated IRF-1 expression and subsequently suppressed IFN-γ-induced B7-H1 mRNA and protein expression by about 80% and 60%, respectively (Fig. 4B). Additionally, IRF-1 siRNA greatly diminished B7-H1 promoter activity in response to IFN-γ in luciferase assay. Interestingly, IRF-1 siRNA also reduced the basal expression of IRF-1, and subsequently declined the constitutive expression of B7-H1 mRNA and protein (Fig. 4B), and downregulated basal B7-H1 promoter activity in luciferase assay (Fig. 4C). In line with this observation, we found that IRF-1 siRNA largely reduced the constitutive B7-H1 expression in SNU739 that expresses high level of basal B7-H1 and greatly diminished the IFN-γ-induced B7-H1 expression in SNU739 and HeLa cells (Fig. 4D). Collectively, the data suggest that IRF-1 plays a key role in constitutive B7-H1 expression and in IFN-γ-induced B7-H1 transcription in A549 cells. This notion was the case with primary normal cells including human dermal fibroblasts and adipose stem cells (Fig. S2).
3.5 JAK/STAT pathway is responsible for regulation of IFN-γ-induced B7-H1 expression
It has been reported that IFN-γ activates various intracellular signaling proteins such as JAK/STAT, PI3K, MTOR/p70-S6K, ERK, JNK, p38, and protein kinases C (PKCs) [18]. We sought to examine whether inhibition of these signaling proteins affects IFN-γ-mediated B7-H1 expression in A549 cells. We found that treatment with AG490 (an inhibitor of JAK/STAT) greatly reduced the transcription of B7-H1 in the presence or absence of IFN-γ compared to other inhibitors for signaling proteins (Fig. 5 A and B). Interestingly, there was a detectable upregulation of B7-H1 transcription in A549 cells treated with rottlerin and genistein (Fig. 5B). Taken together, JAK/STAT pathway may be predominantly responsible for the IFN-γ-mediated B7-H1 transcription. This postulate was further supported by the finding that pretreatment with AG490 strongly reduced both basal expression and IFN-γ-induced upregulation of B7-H1 mRNA and protein in dose-dependent mode (Fig. 6 A). This was the case with IRF-1 transcription in the presence of AG490 (Fig. 6A). In consistent with this finding, basal and IFN-γ-induced B7-H1 promoter activity was substantially diminished in the presence of AG490 (Fig. 6B), an observation suggesting that JAK/STAT pathway may be a key signaling pathway involved in basal expression and IFN-γ-induced upregulation of B7-H1 in A549 cancer cells.
4 Discussion
IRF-1, a member of IRF family of 10 mammalian transcription factors (IRF-1-10), reveals a great functional diversity in the regulation of cellular response in host defense [19]. IRF-1 mRNA is expressed in a variety of cell types, and its expression is dramatically upregulated upon viral infection or IFN stimulation [20]. In this study, we demonstrated that IRF-1 is a critical transcription factor for both constitutive and IFN-γ-induced B7-H1 expression. This result was evidenced by the observations from a site-directed mutagenesis study, in which mutations of the IRF-1 binding site in B7-H1 promoter greatly diminished the basal transcription of B7-H1 in A549 cells that were unstimulated with IFN-γ. Additionally, those mutations appeared to completely abolish the responsiveness of A549 cells to IFN-γ. In support of this finding, knockdown of IRF-1 using siRNA greatly decreased both the constitutive and IFN-γ-induced transcription and translation of B7-H1 in A549 cells. The data provide evidence that IRF-1 serves as an essential transcription factor involved not only in the constitutive but also in IFN-γ-induced B7-H1 expression.
B7-H1 transcription and translation was induced as early as 1 h post-IFN-γ stimulation. This early induction of B7-H1 expression may be largely mediated by the basally expressed IRF-1, as demonstrated by the finding that cytoplasmic IRF-1 was significantly mobilized into a nucleus and to bind B7-H1 promoter region as early as 5 min after IFN-γ stimulation, a time period during which IFN-γ stimulation is probably not sufficient to de novo synthesize IRF-1 through JAK/STAT pathway [21, 22]. In addition, the late IFN-γ-mediated B7-H1 upregulation at 12–24 h after IFN-γ stimulation may occur through de novo synthesized IRF-1, which shows a maximal binding activity to B7-H1 promoter at 7 h following IFN-γ stimulation. Based on the result of IRF-1 and B7-H1 transcription in the presence of AG490, JAK/STAT can be demonstrated as key signaling molecules responsible for basal and IFN-γ-induced transcription of IRF-1 and B7-H1. From the data on PKC inhibitor-induced upregulation of B7-H1 in the absence of IFN-γ, however, we cannot exclude the possibility that other signaling pathway such as PKC might be involved in basal expression of B7-H1.
From the promoter activity assay using various promoter constructs, we found that promoter region from nt −457 and nt −2405 was implicated in an inhibition of promoter activity in response to IFN-γ compared to pH1-457 and pH1-320 which showed maximal B7-H1 promoter activity. This observation suggests that promoter region upstream to nt −457 may contain binding site(s) for some inhibitory transcription factor(s). It remains to be illustrated in the future study. NF-κB is a well-known transcription factor involved in many immunological events, such as inflammatory cytokine production, TCR-mediated T cell activation, and regulation of co-signaling molecules [12, 23, 24]. Interestingly, although B7-H1 promoter contains an NF-κB binding site at nt −182, NF-κB is unlikely to play a key role in the transcription of B7-H1 in A549 lung cancer cells, as evidenced by the luciferase assay in which promoter construct containing NF-κB binding site greatly abolished the responsiveness of A549 cells to IFN-γ. Recently, a publication suggested that NF-κB is essential for IFN-γ-induced B7-H1 upregulation in normal dermal fibroblast, and PI3K is involved in this upregulation [25]. We suppose that this discrepancy in terms of transcription factor and signal pathway involving B7-H1 transcription may result from the different cell types, cancer and normal cells, in which IFN-γ may trigger activation of different signaling proteins, whose activation may differently involve B7-H1 expression.
It is conceivable that tumor cells that are exposed to IFN-γ in the cancer microenvironment, which can exert harsh conditions for cancer cells, can escape immune surveillance by suppressing cancer-specific CTLs through upregulation of a co-inhibitory molecule B7-H1 on the surface. Therefore, one should be cautious to the use of IFN-γ as a therapeutic agent for cancer treatment. Thus understanding the molecular mechanism of regulation of B7-H1 expression in cancers may shed a light on a development of alternative strategies for cancer treatment including cytokines or pharmaceutical compounds downregulating the cancer-associated B7-H1 expression.
Acknowledgments
This study was supported by National R&D Program Grant (M10417060006-04N1706-00610) of The Ministry of Science and Technology and 2,000 Inje University Grant (I.-H.C), and by National Institutes of Health Grant CA97085 and CA106861 (L.C.).
Appendix A A
Supplementary data associated with this article can be found, in the online version, at 10.1016/j.febslet.2005.12.093.
