2,3-Dioxygenase Indoleamine the Immunomodulatory Enzyme Amino Acid Deprivation Links BLIMP-1 To

Amino Acid Deprivation Links BLIMP-1 to the Immunomodulatory Enzyme Indoleamine 2,3-Dioxygenase

This information is current as Nicholas A. Barnes, Sophie J. Stephenson, Reuben M. Tooze of September 29, 2021. and Gina M. Doody J Immunol 2009; 183:5768-5777; Prepublished online 14 October 2009; doi: 10.4049/jimmunol.0803480

http://www.jimmunol.org/content/183/9/5768 Downloaded from

References This article cites 58 articles, 20 of which you can access for free at: http://www.jimmunol.org/content/183/9/5768.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 29, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2009 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Amino Acid Deprivation Links BLIMP-1 to the Immunomodulatory Enzyme Indoleamine 2,3-Dioxygenase1

Nicholas A. Barnes, Sophie J. Stephenson, Reuben M. Tooze, and Gina M. Doody2

Catabolism of tryptophan by IDO1 plays an important role in the control of immune responses. Activation of the eukaryotic initiation factor 2␣ (eIF2␣) kinase general control nonderepressible-2 (GCN2) following tryptophan depletion is a major pathway mediating this effect. However, immunomodulatory target genes of GCN2 activation are poorly deﬁned. The transcriptional repressor B lymphocyte-induced maturation protein-1 (BLIMP-1) is a target of the eIF2␣ kinase1, protein kinase-like ER kinase (PERK) during the unfolded protein response of the endoplasmic reticulum. Thus, BLIMP-1 might also be a mediator of the GCN2 stress response pathway activated by IDO1 and tryptophan depletion. Indeed, in human monocytes BLIMP-1 mRNA and protein are up-regulated in response to both a pharmacological activator of GCN2 and tryptophan-depletion generated by IDO1-trans-

fected cells. This suggests a functional role for BLIMP-1 in the immunomodulatory effects of the IDO1-GCN2 axis. BLIMP-1 has Downloaded from been shown to repress IFN-␥-regulated promoters. As IDO1 is itself highly responsive to IFN-␥, we hypothesized that BLIMP-1 functions in a feedback loop to regulate IDO1 expression. We found that BLIMP-1 binds to IFN-responsive sites in the IDO1 promoter and represses IFN-dependent IDO1 activation. We propose that BLIMP-1 acts in a negative feedback loop to successfully balance the outcome of tolerance vs inﬂammation. The Journal of Immunology, 2009, 183: 5768–5777.

he metabolic enzyme IDO1 contributes to the balance be- that accumulate during cellular starvation (14, 15). The sole iden- http://www.jimmunol.org/ tween tolerance vs inflammation in a number of experi- tified target of GCN2 is the translation initiation factor, eukaryotic T mental models. Expression of IDO1 in APCs, such as initiation factor 2␣ (eIF2␣) (16). Phosphorylation of eIF2␣ medi- macrophages and dendritic cells, can suppress T cell responses as ates inhibition of global protein synthesis, but paradoxically in- observed during mammalian pregnancy, inflammatory conditions, creases the translation of activating transcription factor 4 (13, 17). autoimmunity, and tumor resistance (1–6). IDO1 is itself modu- Activating transcription factor 4, along with other selected tran- lated in the presence of inflammatory stimuli, exhibiting strong scriptional regulators, then activates the genetic program required induction in response to both IFN and TLR signaling (7, 8). This to achieve cellular homeostasis (18). enzyme catabolizes the essential amino acid tryptophan, thus de- The importance of GCN2 during the T cell response to trypto- creasing concentrations in the local microenvironment as well as phan depletion has been established. T cells from GCN2-deficient by guest on September 29, 2021 generating biologically active downstream metabolites (9, 10). mice are refractory to IDO1-mediated suppression of proliferation While either of these pathways may operate to regulate T cells, the in vitro and induction of anergy in vivo (19). There is also accu- existing evidence suggests that both mechanisms are required to mulating evidence that GCN2 activation plays a role in the IDO1- achieve the optimal effect (11). expressing APCs themselves. For example, a recent report dem- The ability to sense and respond to decreasing concentrations of onstrated that IDO1 and GCN2 regulate autocrine IFN-␣-driven amino acids is an important cellular survival mechanism, and the amplification of IDO1 expression in plasmacytoid dendritic cells pathways that mediate this effect are well conserved. Studies in (20). GCN2 belongs to a family of kinases that phosphorylate both yeast and mice revealed that the protein kinase general con- eIF2␣ in response to diverse forms of cellular stress (21). Heme- trol nonderepressible-2 (GCN2)3 plays a central role in the re- regulated inhibitor (HRI) is activated by oxidative stress and he- sponse to amino acid deprivation (12, 13). This kinase undergoes moglobin levels during erythrocyte maturation, protein kinase R a change in activation status upon binding to uncharged tRNAs by dsRNA, and the endoplasmic reticulum (ER) resident homolog of protein kinase R, protein kinase-like ER kinase (PERK), is activated during the unfolded protein response (22–24). We have Section of Experimental Haematology, Leeds Institute of Molecular Medicine, St previously shown that pharmacological activators of PERK up- James’s University Hospital, Leeds, United Kingdom regulate B lymphocyte-induced maturation protein-1 (BLIMP-1) Received for publication October 15, 2008. Accepted for publication August 31, in cells of the myeloid lineage (25). Furthermore, BLIMP-1 was 2009. originally identified as a postinduction transcriptional repressor of The costs of publication of this article were defrayed in part by the payment of page IFN-␤ in response to intracellular dsRNA accompanying viral in- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. fection (26, 27). These data suggest that BLIMP-1 may be a com- 1 This work was supported by a New Investigator Grant (BB/E000746/1) from the mon target of stress-induced signaling pathways. In turn, the in- U.K. Biotechnology and Biological Sciences Research Council (to G.M.D.). duction of BLIMP-1 may shape the outcome of the integrated 2 Address correspondence and reprint requests to Dr. Gina M. Doody, Section of stress response by altering gene expression. Experimental Haematology, Wellcome Trust Brenner Building, Leeds Institute of BLIMP-1 is a Kru¨ppel-type zinc-finger protein that exhibits Molecular Medicine, St. James’s University Hospital, University of Leeds, Leeds LS9 7TF, United Kingdom. E-mail address: [email protected] preferential binding to sequences with a high degree of overlap 3 Abbreviations used in this paper: GCN2, general control nonderepressible-2; eIF2␣, eukaryotic initiation factor 2␣; HRI, heme-regulated inhibitor; ER, endoplasmic re- homologous protein; XBP-1, X-box binding protein 1; ISRE, IFN-stimulated reticulum; PERK, protein kinase-like ER kinase; BLIMP-1, B lymphocyte-induced sponse element; ChIP, chromatin immunoprecipitation. maturation protein-1; IRF-E, IFN regulatory factor-element; IRF, IFN regulatory factor; FL, full length; siRNA, small interfering RNA; wt, wild type; CHOP, C/EBP Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.0803480 The Journal of Immunology 5769 with the IFN regulatory factor-element (IRF-E) found in promoters ates were also analyzed using an in-line RF 2000 fluorescence detector responsive to IFN-␥ (28, 29). Work from our laboratory has shown with 280 and 340 nm excitation and emission wavelengths. Fluorescence that BLIMP-1 is indeed capable of repressing genes regulated by detector gain settings were chosen to give more sensitive detection of tryp- ␥ tophan in more dilute samples than was possible using UV absorbance. IFN- (30, 31). In these examples, BLIMP-1 mediates repression Concentrations of tryptophan in media samples were determined from a through direct competition with IFN regulatory factors (IRFs) at calibration curve based on the peak area (calculated using Dionex Chrome- the IRF-E sites contained within their promoters. However, leon software) for absorbance at 278 nm or the fluorescence emission at BLIMP-1 can also exert its repressive effect through the recruit- 340 nm depending on the concentration. ment of histone-modifying enzymes (27, 32–35). Thus, BLIMP-1 RT-PCR has the potential to either transiently repress its targets or to alternatively ensure stable epigenetic silencing. RNA was isolated by the TRIzol method (Invitrogen) and subjected to DNase ␥ I treatment (DNA-free; Ambion). cDNA was generated using oligo(dT) and The IDO1 promoter is highly responsive to IFN- and is rapidly SuperScript II reverse transcriptase (Invitrogen). Quantitative real-time PCR induced in cell lineages in which BLIMP-1 is also expressed. A was performed using SYBR Green MasterMix (Applied Biosystems) on an structurally similar enzyme encoded by IDO2 can also be induced ABI Prism 7500 system (PerkinElmer) and evaluated with 7500 System Soft- by IFN-␥, but the regulatory elements conferring this response ware. Specificity of PCR was monitored with melting curve analysis, and amplification of a single product of the expected size was verified on agarose have not been characterized (36–39). In this study we tested the gels in initial experiments. Relative expression was normalized to GAPDH. hypothesis that BLIMP-1 operates as a postinduction repressor For quantitative RT-PCR the following primers were used: BLIMP-1 forward during IDO1-mediated immunoregulation, much like its originally (5Ј-ACCAAGGAATCTGCTTTTCAAGTATG-3Ј), BLIMP-1 reverse (5Ј-CA described function limiting IFN-␤ production during viral infec- TCACTCCAATAACCTCTTCACTGT-3Ј); C/EBP homologous protein (CHOP)/GADD153 forward (5Ј-TGGAAGCCTGGTATGAGGAC-3Ј), tion. We propose that BLIMP-1 regulates IDO1 levels in cell types Downloaded from CHOP/GADD153 reverse (5Ј-TGTGACCTCTGCTGGTTCTG-3Ј); DNAJB9 that express these two proteins, thereby maintaining the expression forward (5Ј-AGTCGGAGGGTGCAGGATATT-3Ј), DNAJB9 reverse (5Ј-CC of IDO1 at a level that successfully balances the outcome of tol- GATTTTGGCACACCTAA-3Ј); IDO1 forward (5Ј-GCCCTTCAAGTGTT erance vs inflammation. TCACCAA-3Ј), IDO1 reverse (5Ј-GCCTTTCCAGCCAGACAAAT AT-3Ј); GAPDH forward (5Ј-AACAGCGACACCCACTCCTC-3Ј), GAPDH reverse (5Ј-CATACCAGGAAATGAGCTTGACAA-3Ј). Materials and Methods X-box binding protein 1 (XBP-1) mRNA splicing was monitored using

Cell culture 5Ј-CTGAAGAGGAGGCGGAAGC-3Ј (forward) and 5Ј-AATACCGCCA http://www.jimmunol.org/ GAATCCATGG-3Ј (reverse) primers in conventional PCR reactions using U937, H929, U266 and HeLa cell lines were maintained in RPMI 1640 AmpliTaq Gold (Applied Biosystems). growth medium supplemented with 10% heat-inactivated FCS. COS7 cells were maintained in DMEM medium supplemented with 10% FCS. Tryp- Luciferase vectors and assays tophanol, L-tryptophan, DTT, PMA, and recombinant human IFN-␥ and IFN-␣ were all purchased from Sigma-Aldrich and added directly to the The promoter sequence of the human IDO1 locus was amplified from growth media in the indicated final concentrations. Tryptophan-depleted genomic DNA with the following primers: IDO1 FL (forward) (5Ј-TATA media was generated by transfecting HeLa cells with an IDO1 expression AGATCTCAGAATCATCTAAAACG-3Ј), IDO1 short (forward) (5Ј-TATA vector and cultured for 40 h. Media was filtered before secondary cell AGATCTAGCCTTCATTTTCATTC-3Ј), IDO1 (reverse) (5Ј-AATTAAG culture and supplemented with 10ϫ concentration of L-tryptophan where CTTCTTATAATAGTTGCTGC-3Ј). shown. The use of peripheral blood from healthy donors was approved by The amplicons were then cloned into the vector pXPG (40) generating by guest on September 29, 2021 the Local Research Ethics Committee. Primary monocytes were isolated by promoter constructs spanning Ϫ1298 to ϩ52 for IDO1 FL and Ϫ272 to negative selection using a monocyte-specific magnetic bead protocol ϩ52 for IDO1 short, relative to the transcriptional start sites of reference (Miltenyi Biotec). Purity was monitored by flow cytometry using CD14 sequences. Mutations were introduced into the short IDO1 promoter using and CD11b monocyte-macrophage markers. Human monocytes were the Gene Tailor site-directed mutagenesis kit (Invitrogen) as shown in Fig. maintained in RPMI 1640 supplemented with 30% heat-inactivated AB 5 using the following primers, with the mutated residues underlined: IDO1 human serum, 100 ␮g/ml streptomycin, and 100 U/ml penicillin and ad- mutant 1 (forward) (5Ј-AAAATGAAACCATTCCAAAACTGGAACTA hered overnight for 16 h before stimulation. ATTTCTCAC-3Ј), IDO1 mutant 1 (reverse) (5Ј-TTTTGGAATGGTTTCA TTTTCTAGTTTCTT-3Ј), IDO1 mutant 2 (forward) (5Ј-ACAAACAAAG Expression vectors and transfections AAACTAGAAACTGAAACCATTC-3Ј), IDO1 mutant 2 (reverse) (5Ј-TT TCTAGTTTCTTTGTTTGTAATATGAAAACC-3Ј). Mutant 2 was derived Expression vectors encoding full-length (FL) BLIMP-1, IRF-1, and IRF-2 from the mutant 1 plasmid, and all constructs were sequence verified before have been described previously (30). An expression vector encoding IDO1 use. For luciferase assays, three replicate transfections were performed for was generated by amplifying cDNA using primers (5Ј- ATAGAATTCG each condition. Experiments were done using the Promega luciferase assay CCACCATGGCACACGCTATGGA-3Јand5Ј-ATATGTCGACTTAACC system and analyzed on a Berthold Lumat LB luminometer. Each experi- TTCCTTCAAAA-3Ј) and cloned into the modified pIRES2-EGFP vector ment was performed in duplicate with similar results and was confirmed incorporating an N-terminal Myc epitope tag as previously described (30). with separate plasmid preparations. The data are from a representative Transfection of HeLa and COS7 cells were performed with GeneJuice experiment displayed as fold increase in light units relative to unstimulated reagent according to the manufacturer’s instructions. In RNA interference cells cotransfected with the empty vector. experiments, U937 cells were transfected with 400 pmol of small interfering RNA (siRNA) by electroporation with settings of 280 V and 975 ␮F. Antibodies Forty hours posttransfection, cells were processed for protein and mRNA evaluation. The oligos used contained the following sequences: GCN2 Rabbit antisera to BLIMP-1 have been described previously (25, 31). Rab- siRNA sense sequence (5Ј-CAGCAGAAAUCAUGUACGAUU-3Ј); con- bit polyclonal Abs to IRF-1 and IRF-2 were from Santa Cruz Biotechnol- trol siRNA sense sequence (5Ј-CUACCUCUAGAACGGACGUUU-3Ј). ogy. Nonimmune rabbit IgG was from Upstate Biotechnology. Abs used for flow cytometry were monoclonal anti-human CD14 conjugated to PE HPLC analysis (Dako) and monoclonal anti-human CD11b conjugated to PE (BD Bio- sciences). Rabbit polyclonal Ab to GCN2 was from Cell Signaling Tech- Tryptophan (Sigma-Aldrich, catalog no. T-0254) was dissolved in water nology, and mouse mAb to ␤-actin, clone AC15, was from Sigma-Aldrich. and the concentration accurately determined by UV absorbance spectro- photometry. A dilution series was then prepared in acetic acid (1% v/v) EMSA from 50 to 0.1 ␮g/ml. Standards and media samples were analyzed by HPLC on a C18 column (Agilent Zorbax ODS, 4.6 ϫ 250 mm) using a Nuclear extracts were prepared from COS7 cells transfected with expres- Dionex HPLC. An elution gradient was used from 100% eluent A (meth- sion vectors for BLIMP-1, IRF-1, or IRF-2 as described previously (30). anol 2.5% v/v, isopropanol 2.5% v/v, trifluoroacetic acid 0.1% v/v) to For EMSA, the double-stranded probes used contained the extended IFN- 100% eluent B (methanol 47.5% v/v, isopropanol 47.5% v/v, trifluoroace- stimulated response element (ISRE) sequence in the proximal region of the tic acid 0.1% v/v) at 20°C with a 0.5 ml/min flow rate. The column eluate IDO1 promoter within the following sequences: IDO1 wild-type (wt) (for- was monitored by UV absorbance at multiple wavelengths (230, 260, and ward) (5Ј-ACAAACAAAGAAACTAGAAAATGAAACCATTCCAAA 278 nm) using a Dionex PDA100 detector to aid peak identification. Elu- AGTGGAAGTAATTTCTCACTGCCC-3Ј), IDO1 wt (reverse) (5Ј-GGGC 5770 FEEDBACK INHIBITION OF IDO1 BY BLIMP-1

AGTGAGAAATTACTTCCACTTTTGGAATGGTTTCATTTTCTAGTT genes were analyzed in the same samples. CHOP represents a TCTTTGTTTGT-3Ј). The underlined bases represent sites in which common target of eIF2␣ kinases and was elevated to similar levels positions of the ISRE sites were altered as described in Fig. 5. as BLIMP-1 in response to both DTT and tryptophanol, but not to DNA probes [32P] labeled with T4 polynucleotide kinase were incubated with nuclear extract in the presence of poly(dI:dC) (Amersham Bio- PMA (13). In contrast, DNAJB9 is a specific target of the ER stress sciences) for 30 min at room temperature. Supershift was performed by the response pathway and accordingly is only up-regulated in response addition of antisera to the extract before mixing with radioactive probe, and to DTT (42). As an additional marker of the unfolded protein re- competition assays included the addition of unlabeled probe to the reaction sponse, we evaluated XBP-1 mRNA splicing, which was also con- mixture. fined to DTT-stimulated samples (Fig. 1C) (43, 44). These results Chromatin immunoprecipitation (ChIP) confirm that tryptophanol does not induce a stress response via the ER pathway. Chromatin prepared from myeloma cells or primary monocytes was precipitated with 2–4 ␮g of Ab to BLIMP-1, IRF-1, IRF-2, or control rabbit Ab and To demonstrate that the response observed in U937 cells was protein A-Sepharose. DNA was eluted and cross-links were reversed by over- not a property unique to this monocytic cell line, we evaluated night incubation at 65°C. Input DNA was prepared from an equal volume of BLIMP-1 levels in primary human monocytes purified from pe- chromatin following RNase treatment and resuspended in the same final vol- ripheral blood (Fig. 1B). Samples from five healthy individuals ume as the ChIP samples. A standard curve of input DNA was generated from were subjected to stimulation with tryptophanol. After 6 or 24 h of each chromatin sample. Target sequences were analyzed by real-time PCR as described above. The amount of precipitated material was calculated from the incubation the levels of BLIMP-1 and CHOP mRNA were as- standard curve and normalized relative to control immunoprecipitates. The sessed. In each case, the basal levels of CHOP were reduced by data are representative of a minimum of two independent chromatin prepara- prolonged incubation without stimulation. However, treatment tions, and they represent average and SD of at least two independent immu- with tryptophanol increased the relative amount of CHOP at each Downloaded from noprecipitations. The following primers were used: IDO1 Ϫ1449 (forward) (5Ј-ATCCGCCCTCCTCAGCTT-3Ј), IDO1 Ϫ1351 (reverse) (5Ј-ATTCCG time point. BLIMP-1 levels remained essentially unchanged in TTTATCCAGTCATCTCACT-3Ј), IDO1 Ϫ812 (forward) (5Ј-TAGAGGAA samples treated with media alone for 6 h, but showed a reproduc- CGGGCAACTTGG-3Ј), IDO1 Ϫ722 (reverse) (5Ј-GCATGCAAGTCTGT ible 2- to 3-fold increase in the tryptophanol-treated samples over GGTTCACT-3Ј), IDO1 Ϫ610 (forward) (5Ј-GCCAGACTGTTGCCTCAT the same time frame. At 24 h a similar increase was maintained in Ј Ϫ Ј CA-3 ), IDO1 512 (reverse) (5 -CATTTGCCCTTCTCACATCTCT tryptophanol-treated samples relative to samples treated with me- AA-3Ј); IDO1 Ϫ378 (forward) (5Ј-AGGTTGTGTTTCCGGGCTG-3Ј), IDO1 Ϫ288 (reverse) (5Ј-CATTTAGTTATTTCTCATTGGCTTCCTT-3Ј), IDO1 dia alone at the same time point, although absolute levels had http://www.jimmunol.org/ ϩ3 (forward) (5Ј-TTCTCACTGCCCCTGTGATAAA-3Ј), IDO1 ϩ93 returned to starting levels. (reverse) (5Ј-GTGCCCCTCAGTGTCTGAAGA-3Ј), IDO1 ϩ90 (forward) The above data show that BLIMP-1 mRNA is rapidly up-regu- Ј Ј ϩ Ј (5 -GCACCAGAGGAGCAGACTACAA-3 ), IDO1 183 (reverse) (5 - lated in response to both ER stress and an activator of GCN2. We CAAAGCCCACTTCTTCATCAATATG-3Ј), IDO1 ϩ565 (forward) (5Ј- CCATGAGTCAAACAGAAATCCCT-3Ј), IDO1 ϩ655 (reverse) (5Ј-CA had previously shown that BLIMP-1 protein expression is delayed ATCTCATAAACCCCATCAAATTAGA-3Ј), IDO1 ϩ912 (forward) during the DTT stress response, similar to stress response genes (5Ј-CCATTCTCATAGCCATCCCAA-3Ј), IDO1 ϩ1002 (reverse) such as BiP (GRP78) (25, 45). To determine whether induction of (5Ј-CCCAGGTGACAGAGCGAGAC-3Ј). BLIMP-1 in response to tryptophanol leads to a similar delay in Protein analysis translation, lysates were prepared from U937 cells treated for the indicated amount of time and evaluated for BLIMP-1 protein (Fig. by guest on September 29, 2021 Equivalent numbers of U937 cells were treated as indicated, then washed 1D). As anticipated, PMA stimulation of U937 results in BLIMP-1 with PBS and lysed in reducing Laemmli buffer. Samples were separated protein induction that is entirely concordant with the levels of on SDS-PAGE, transferred to nitrocellulose, and blocked in Tris-buffered saline containing 1% BSA and 0.1% Tween 20. Abs used to detect mRNA, whereas tryptophanol treatment gives rise to abundant BLIMP-1, GCN2, and ␤-actin were incubated with the membrane followed protein at a later time point. These results indicate that BLIMP-1 by secondary Abs coupled to HRP and visualized using an ECL (Pierce) is transiently expressed when aminoacyl-tRNA formation is protocol. prevented. Results The GCN2 activator tryptophanol induces BLIMP-1 IDO1-mediated tryptophan depletion induces BLIMP-1 BLIMP-1 is rapidly expressed following PMA stimulation of Pharmacological agents such as DTT and tryptophanol offer a con- U937 monocytic cells and is involved in the differentiation of this venient means for eliciting a robust stress response, but may not cell line (41). Furthermore, we have shown that U937 cells sub- fully recapitulate the same response to a physiological stimulus in jected to a range of stimuli that target ER stress response pathways vivo. To better approximate the effect of amino acid deficiency on induce BLIMP-1 at the level of transcription (25). BLIMP-1 induction, we generated tryptophan-deficient media by To assess whether BLIMP-1 is also responsive to stress path- transfecting HeLa cells with an expression vector encoding the ways initiated by nutritional deprivation, U937 cells were treated human IDO1 gene. HPLC analysis of the resulting conditioned with the amino acid alcohol tryptophanol. Amino acid alcohols act media confirmed the selective depletion of tryptophan in the pres- as competitive substrates for tRNA synthetases, thus inhibiting the ence of IDO1-expressing cells, reducing the detectable concentra- aminoacylation of tRNAs. GCN2 is subsequently activated by the tion from 5 ␮g/ml in standard media to 0.37 Ϯ 0.01 ␮g/ml (Fig. resulting increase in uncharged tRNA (19). For comparison, U937 2A). The observed reduction in tryptophan levels between empty cells were also treated with PMA or DTT, a potent inducer of the vector and IDO1-conditioned media was 83 Ϯ 8.5%, determined unfolded protein response. from three independent preparations. U937 cells were cultured for A kinetic analysis of BLIMP-1 induction confirmed previous 24 h in conditioned media derived from HeLa cells transfected results, demonstrating a gradual increase in response to PMA stim- with empty vector or an IDO1 expression vector (Fig. 2B). Tran- ulation that was maintained throughout the time course, whereas scripts for both BLIMP-1 and CHOP increased 4-fold in cells in- DTT stimulation generates an ϳ6-fold increase in transcript level cubated with the IDO1-conditioned media, while DNAJB9 levels that peaks between 4 and 6 h and then declines (Fig. 1A) (25). remained unaffected. The catabolism of tryptophan by IDO1 not Tryptophanol treatment of U937 also increased BLIMP-1 mRNA only lowers the available concentration of this essential amino levels in a pattern that was most similar to that observed with DTT. acid, but it also results in generation of biologically active metab- To control for the specificity of each stimulus, other stress-induced olites from the kynurenine pathway (9, 10). To determine whether The Journal of Immunology 5771 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 1. GCN2 activation by tryptophanol increases BLIMP-1 expression in U937 cells and primary monocytes. A, BLIMP-1, CHOP, and DNAJB9 mRNA levels were determined by quantitative RT-PCR relative to GAPDH in U937 cells cultured in the presence of 0.5 mM tryptophanol, 2 mM DTT, 10 ng/ml PMA or media alone for the indicated times. These data are derived from a minimum of three separate experiments and are displayed as the means Ϯ SD. B, The expression of BLIMP-1 and CHOP mRNA was assessed by quantitative RT-PCR in primary human monocytes treated with 1 mM tryptophanol or media alone for the indicated times. These data are derived from ﬁve healthy donors and are normalized to expression in a 0-h control from ,p Ͻ 0.05 by unpaired t test). C, XBP-1 splicing (U ,ء p Ͻ 0.01 and ,ءء) each individual. The mean value from the ﬁve samples is indicated on the plot unspliced; S, spliced) was assessed in the samples described in A by conventional RT-PCR. D, The kinetics of BLIMP-1 protein expression in U937 cells after stimulation with 0.5 mM tryptophanol or 10 ng/ml PMA was assessed by Western blotting.

tryptophan depletion was the predominant activity leading to en- To conﬁrm that the increased levels of BLIMP-1 mRNA obtained hanced gene expression, the conditioned media were supplemented in U937 cells exposed to tryptophan-depleted media were mirrored in with a 10-fold excess (10ϫ) of the normal concentration of L- BLIMP-1 protein levels, we performed Western blot analysis on the tryptophan. Inclusion of excess tryptophan completely abolished same samples. Incubating U937 cells with media from HeLa cells induction of BLIMP-1 and CHOP in the presence of IDO1-condi- transfected with empty vector had no effect on BLIMP-1 levels (Fig. tioned media, indicating that the absence of free tryptophan is es- 2C). In contrast, cells grown in media from IDO1-transfected HeLa sential for this response, as opposed to the generation of kynure- cells increased BLIMP-1 protein, while inclusion of supplemental nine metabolites. tryptophan restored the protein to basal levels. 5772 FEEDBACK INHIBITION OF IDO1 BY BLIMP-1

FIGURE 2. IDO1-mediated tryptophan depletion induces BLIMP-1 in U937 cells and primary monocytes. A, Levels of free tryptophan (Trp) in conditioned medium (CM) obtained from HeLa cells transfected with an empty vector or IDO1 expression construct were determined by HPLC analysis. The peak position of tryptophan was veriﬁed by inclusion of excess tryptophan (data not shown). B, BLIMP-1, CHOP, and DNAJB9 mRNA expression was determined by quantitative RT- PCR in U937 cells cultured for 24 h in CM. These data are derived from a minimum of three separate experiments and are displayed as the means Ϯ SD. C, BLIMP-1 protein induction in U937 cells exposed to the CM was determined by Western blotting. D, The ex- Downloaded from pression of BLIMP-1 and CHOP mRNA was assessed by quantitative RT-PCR in primary human monocytes treated with CM for 6 h. These data are derived from four healthy donors and are normalized to expression in cells exposed to vector CM from each individual. The mean value from the four samples is indicated on the

/p Ͻ 0.05 by unpaired t test). http://www.jimmunol.org ,ء) plot

To further substantiate these observations, we subjected primary human monocytes to culture with media derived from transfected by guest on September 29, 2021 HeLa cells. Consistent with the results obtained from U937 cells, primary monocytes exposed to IDO1-conditioned media showed a reproducible induction of BLIMP-1, as well as an expected increase in CHOP, relative to cells cultured with media from empty vector transfectants (Fig. 2D). Supplementation of IDO1-conditioned media with excess tryptophan once again prevented the response. Thus, a tryptophan-depleted microenvironment generated by expression of IDO1 is sufﬁcient to induce BLIMP-1.

Induction of BLIMP-1 mRNA is dependent on GCN2 The induction of BLIMP-1 by tryptophanol and tryptophan depletion implicates GCN2 as a key mediator of the response. To directly address this we therefore examined the effect of depleting GCN2 expression on BLIMP-1 induction. Transfection of U937 cells with GCN2-specific but not control siRNA resulted in knockdown of GCN2 protein expression (Fig. 3A). To determine whether induction of BLIMP-1 is downstream of GCN2 in the amino acid stress response, the mRNA levels of BLIMP-1 and CHOP were then assessed in these cells after treatment with IDO1- conditioned media (Fig. 3B). While cells treated with control siRNA induced the genes encoding BLIMP-1 and CHOP when subjected to tryptophan depletion, in cells expressing low levels of GCN2 protein, the induction of both BLIMP-1 and the stress re- FIGURE 3. siRNA-mediated knockdown of GCN2 inhibits BLIMP-1 in- sponse gene CHOP was abolished. Thus, the induction of duction. A, siRNA knockdown of GCN2. U937 cells were transfected with BLIMP-1 mRNA by tryptophan-deficient media is dependent on a control or GCN2-specific siRNA and evaluated for protein expression by Western blotting at 40 h. B, Assessment of BLIMP-1 and CHOP transcript functional GCN2 signaling pathway. levels by quantitative RT-PCR relative to GAPDH from the U937 samples in BLIMP-1 binds to regulatory regions in the IDO1 promoter A treated with IDO1-conditioned media for 24 h. These data are derived from p Ͻ 0.05 ,ء) two separate knockdowns and are displayed as the means Ϯ SD p Ͻ 0.05 ,ءء Expression of IDO1 in multiple cell types is modulated in response by unpaired t test, BLIMP-1 in control vs GCN2 knockdown, and to local inflammatory signals, most notably IFN-␥ and IFN-␣ by unpaired t test, CHOP in control vs GCN2 knockdown). The Journal of Immunology 5773 Downloaded from

FIGURE 4. BLIMP-1 binds to the endogenous IDO1 promoter. ChIP analysis of (A) BLIMP-1, (B) IRF-1, and (C) IRF-2 binding in myeloma cells http://www.jimmunol.org/ lines. Chromatin prepared from U266 or H929 myeloma cells was immunoprecipitated with control rabbit IgG, anti-BLIMP-1, anti-IRF-1, or anti-IRF-2. D, ChIP analysis of BLIMP-1 binding in primary human monocytes. Chromatin prepared from monocytes purified from two donors stimulated for 6 h with tryptophanol was immunoprecipitated with control rabbit IgG or anti-BLIMP-1. In each instance the amount of promoter sequence present at the indicated positions was then quantified by real-time PCR. The ratio of amplified material relative to control IgG is displayed as fold enrichment. These data are derived from a minimum of two separate immunoprecipitations from each sample and are displayed as the means Ϯ SD.

(7, 46, 47). The recruitment of STAT1 and IRF-1 to the IDO1 The region of the human IDO1 promoter where we detected promoter mediates the potent response to IFN-␥, and animals de- BLIMP-1 binding contains a characterized ISRE motif that confers by guest on September 29, 2021 ficient in either IRF-1 or IFN-␥ fail to produce IDO1 during in- IFN responsiveness (46). In addition to this known site, a more fection (48, 49). We have previously shown that the transcriptional degenerate sequence that is similar to an ISRE motif is present repressor BLIMP-1 is an effective competitor for binding to sites immediately downstream (Fig. 5B). Examination of these se- that are alternatively occupied by IRF-1 or IRF-2 (30, 31). To quences suggested that, although divergent from the optimum determine whether BLIMP-1 is capable of binding to the IDO1 BLIMP-1 consensus, the tandem sites may be sufficient to confer promoter in vivo, we performed ChIP from myeloma cell lines, BLIMP-1 binding (29). Using a probe that incorporated both po- which express constitutively high levels of BLIMP-1 protein. An tential ISREs, we evaluated the relative binding of BLIMP-1, earlier characterization of the IDO1 promoter identified two ISREs IRF-1, and IRF-2 by EMSA (Fig. 5A). A single dominant band was Ϫ Ϫ located 1050 and 30 bp upstream of the transcriptional start observed for all three proteins consistent with monomeric interac- site that contain potential IRF binding sites (50). Evaluation of tions. The specificity of these complexes was confirmed by Ab BLIMP-1 binding to a region spanning both sites with a series of supershift. A higher molecular mass species, possibly representing primer pairs revealed a preferential 6- to 10-fold association with two molecules, was observed with BLIMP-1-containing nuclear the proximal ISRE in both cell lines (Fig. 4A). These myeloma cell extract. lines also constitutively express both IRF-1 and IRF-2, which can We had previously found that BLIMP-1 exhibits a strong pref- compete with BLIMP-1 for binding site occupancy. ChIP analysis erence for a G rather than C at key positions of overlapping of IRF-1 and IRF-2 binding to the IDO1 promoter region demon- BLIMP-1/IRF binding sites (30, 31). The presence ofaCatoneof strated a similar pattern of occupancy to that of BLIMP-1 (Fig. 4, B and C). There was no detectable enrichment at the distal ISRE the equivalent positions in the upstream ISRE suggested that the with any of these transcription factors. downstream motif could be more relevant to BLIMP-1 binding. To To demonstrate that BLIMP-1 inducibly expressed in the my- assess the contribution of this site we performed a competitor as- eloid lineage can also bind to the IDO1 promoter, we performed say using excess unlabelled probe in which the indicated G resi- ChIP from primary human monocytes that had been cultured in the dues were mutated to Cs (mutant 1, Fig. 5B). As expected, the presence of tryptophanol for 6 h (Fig. 4D). BLIMP-1 demonstrated introduction of these mutations had no effect on the binding of an ϳ2-fold enrichment using primer pairs closest to the predicted either IRF-1 or IRF-2. However, BLIMP-1 binding was not abol- binding site. Assessment of binding using additional primers that ished but only slightly reduced. This suggested that binding to the amplify within the coding region of the IDO1 gene showed no upstream ISRE was indeed important. We therefore assayed a enrichment. Thus, the ability of BLIMP-1 to bind the IDO1 pro- probe containing a further unfavorable base substitution in this moter is maintained, whether it is constitutively or inducibly ex- upstream site (mutant 2, Fig. 5B). While IRF-1 and IRF-2 binding pressed, and is observed in a cell type in which IDO1 normally were modestly affected by the alteration, this change was sufficient functions. to block competition for BLIMP-1 binding. These results indicate 5774 FEEDBACK INHIBITION OF IDO1 BY BLIMP-1

lation of HeLa cells transfected with this construct produced a 300-fold induction of the reporter gene, reﬂecting co-operation between the multiple cis-regulatory elements. Cotransfection of BLIMP-1 had a profound effect on the IFN-␥ induction, reducing the stimulation by Ͼ10-fold. Although IFN-␥ is generally the most potent inducer of IDO1 expression, the type I IFN IFN-␣ has also been shown to regulate this gene (46, 47). Examination of the response of the luciferase constructs to IFN-␣ stimulation showed an 8-fold induction of the FL reporter and a weaker activation of the short construct (Fig. 6B). The inclusion of the BLIMP-1 expression vector substantially reduced the luciferase activity driven by either promoter. To further determine the importance of the BLIMP-1 binding, we performed site-directed mutagenesis of the short IDO1 promoter containing the proximal ISRE and overlapping BLIMP-1 binding site. Equivalent mutations were introduced into the critical residues of the extended BLIMP-1 site as deﬁned by EMSA (Fig. 5). The response to IFN-␥ was retained after mutation of

either one or both halves of the extended BLIMP-1 binding site, Downloaded from although efﬁciency of induction was somewhat reduced (Fig. 6C). As predicted by EMSA, cotransfection of BLIMP-1 also inhibited mutant 1 where only the ﬁrst half of the extended site retains preferred residues for BLIMP-1 binding, although the retained inhibition is not as effective as that seen on the wt

promoter. Mutation of both halves of the extended ISRE/ http://www.jimmunol.org/ BLIMP-1 site in mutant 2 was required to eliminate BLIMP-1- dependent repression. Taken together, these results indicate that BLIMP-1 binds to and regulates the IDO1 promoter and provide further evidence to suggest that some promoters containing tandem BLIMP-1 sites allow greater divergence from the optimal consensus sequence.

BLIMP-1 prevents IFN-␥-dependent induction of endogenous by guest on September 29, 2021 IDO1 To ensure that the observed inhibition of the IDO1 promoter construct by BLIMP-1 could be reproduced in the context of the FIGURE 5. BLIMP-1 binds to an extended ISRE site. A, EMSA eval- native promoter, we evaluated regulation of the endogenous uation of binding to the proximal IDO1 ISRE site. Nuclear extracts from gene. HeLa cells were transfected with bicistronic BLIMP-1- IRF-1-, IRF-2-, or BLIMP-1-transfected COS cells were evaluated for EGFP or control EGFP vector and EGFPhigh-expressing cells binding to wild-type ISRE sequence alone, in the presence of specific Abs, were sorted (Fig. 7). The transfected populations were treated or in the presence of 5- or 50-fold excess unlabelled wt or mutated com- with IFN-␥, and the relative expression levels of IDO1 mRNA petitor probe. B, The core sequence of each probe and mutated versions were measured after 6 or 24 h. Cells transfected with the empty F used in A. Bands corresponding to BLIMP-1 (arrowhead), IRF-1 ( ), and vector substantially increased IDO1 levels after6hoftreatment E IRF-2 ( ). Two arrowheads indicate high-molecular mass species of with IFN-␥ (100% relative induction), and levels remained high BLIMP-1. at the 24 h time point. In contrast, cells that expressed BLIMP-1 generated Ͻ20% of the amount of IDO1 expressed by control cells at 6 h. At 24 h, the levels of IDO1 remained substantially that greater divergence from the optimal consensus is tolerated by suppressed, although to a somewhat lesser extent. Thus, con- BLIMP-1 at this tandem binding site. sistent with the demonstration of direct promoter regulation detailed above, BLIMP-1 represses the IFN-␥ response of the en- BLIMP-1 represses IDO1 promoter induction dogenous IDO1 gene. To address whether BLIMP-1 regulates the IDO1 promoter as predicted by promoter binding, we generated luciferase reporter constructs. In the first instance, we evaluated a construct that con- Discussion tained the basal promoter and the single proximal ISRE (short The original identification of BLIMP-1 as a repressor of IFN-␤ IDO1). Stimulation of transfected HeLa cells with IFN-␥ for6h during viral infection provides a groundwork for establishing how resulted in a 16-fold increase in luciferase activity relative to cells this transcription factor may be functioning outside the B cell lin- treated with media alone (Fig. 6A). Cotransfection of a vector en- eage during inflammatory and immune responses (26). In contrast coding BLIMP-1 inhibited the IFN-␥ response by Ͼ4-fold. To to the requirement for sustained BLIMP-1 expression in the gen- determine whether BLIMP-1 could operate in the presence of ad- eration and maintenance of plasma cells, exposure of other cell ditional upstream regulatory elements, we generated a construct lineages to acute stimuli may lead to a transient increase in that included the distal ISRE as well as an IFN-responsive IFN-␥ BLIMP-1 protein to repress genes likely to impact on the cell’s activation site (GAS) element (FL IDO1) (48, 50). IFN-␥ stimu- responsive behavior (51). Our previous data identified BLIMP-1 as The Journal of Immunology 5775

FIGURE 6. BLIMP-1 represses IDO1 promoter activity. Luciferase reporter assay of promoter activation by IFN-␥ (A) or IFN-␣ (B) HeLa cells transfected with luciferase reporter constructs containing FL or short promoter regions of IDO1 with either 50 ng of empty expression vector or 50 ng of expression vector encoding BLIMP-1 were evaluated for luciferase activity after treatment with media alone or containing 200 IU/ml IFN-␥ or 1000 IU/ml IFN-␣ for6h.C, BLIMP-1 requires speciﬁc residues for inhibition of IDO1 promoter Downloaded from activity. Luciferase reporter constructs encoding the truncated IDO1 promoter were mutated as described in Fig. 5 and cotransfected into HeLa cells with either 50 ng of empty expression vector or 50 ng of expression vector encoding BLIMP-1 and evaluated for luciferase activity after treatment with media alone or containing ␥

200 IU/ml IFN- for 6 h. These data are derived from http://www.jimmunol.org/ three transfections displayed as the means Ϯ SD and are representative of three separate experiments. by guest on September 29, 2021

a gene that is induced downstream of the kinase PERK after ex- ing cellular stress. We now show that an additional pathway lead- posure of cells to agents that initiate a stress response of the en- ing to eIF2␣ phosphorylation, triggered by low levels of an essen- doplasmic reticulum (25). Since PERK is one of four kinases that tial amino acid, can stimulate BLIMP-1 expression through converges on eIF2␣, and another of these, protein kinase R, is activation of the stress kinase GCN2. activated during viral infection, we hypothesized that BLIMP-1 The identification of BLIMP-1 as a molecular target of IDO1- may represent a transcription factor that is selectively induced dur- mediated GCN2 activation not only reinforces the concept that BLIMP-1 should be considered as a transcriptional regulator of the integrated stress response, but also highlights its recurring role in the feedback attenuation of cellular responses. For example, a recent report documenting the function of BLIMP-1 in the T cell lineage demonstrated that upon activation of mature T cells, IL-2 drives the late induction of BLIMP-1 and then is itself extin- guished as BLIMP-1 protein accumulates (52). These kinetics contrast with the acute induction observed in response to cell stress, indicating that the timing of BLIMP-1 expression is very much cell-type and stimulus specific. This may in turn translate into differential effects on target gene regulation. Our results show that the activation of GCN2, in the absence of other stimulatory inputs, results in the rapid induction of BLIMP-1 in monocytes. In another setting, the activation of GCN2 in CD19ϩ dendritic cells stimulated through B7 ligation leads to an amplification of IDO1 (20). FIGURE 7. BLIMP-1 prevents IDO1 induction by IFN-␥. Transcript In this instance the dynamics of IDO1 expression may be far more levels of IDO1 were assessed by quantitative RT-PCR in sorted EGFPhigh HeLa cells transfected with bicistronic BLIMP-1 IRES2 EGFP or control complex, requiring the coordinated activity of a series of different IRES2 EGFP vector after treatment with 200 IU/ml IFN-␥ for 6 h. The data transcription factors to achieve the desired outcome. are derived from two separate transfections and are displayed as the The IDO1 gene is evolutionarily ancient and is likely to have means Ϯ SD. evolved from an enzyme used by unicellular organisms competing 5776 FEEDBACK INHIBITION OF IDO1 BY BLIMP-1 for limiting nutrients, to one used by an intermediate immune sys- required for effective Ag presentation. Our previous work demon- tem checking the growth of pathogens, to one used by a sophisti- strating the repression of components of both MHC class I and cated higher vertebrate such as placental mammals where IDO1 is class II Ag presentation pathways by BLIMP-1 is consistent with involved in regulatory networks integrating both innate and adap- this suggestion (30, 31). tive immune systems (10). As part of the evolutionary process, the Furthermore, when one considers the wider immune context in IDO1 gene has acquired the ability to respond to signals specific to which monocytic cells are expressing IDO1 in response to IFN-␥, the adaptive immune system. Most notably, the IDO1 promoter is tryptophan is depleted not just from the IDO1-expressing cell, but highly responsive to IFN-␥ and requires the action of the tran- also from the surrounding tissue microenvironment. As a result, scription factor IRF-1 (48, 49). The closely related IDO2 gene is neighboring cells will also be subjected to inadequate levels of free also responsive to IFN-␥ in a number of cell lines, but not in a tryptophan. In particular, T cells are exposed to this microenvi- murine model of malaria infection in which IDO1 expression de- ronment as they are recruited to participate in the immune re- pends on IFN-␥ (36–39). The ability of IRF-1, and potentially sponse. T cell proliferation is known to be susceptible to the effects BLIMP-1, to modulate IDO2 will require a more thorough inves- of enzymes expressed by APCs that catabolize and lower concen- tigation of the regulatory elements governing expression of this trations of amino acids, such as tryptophan and arginine, in mixed locus. In agreement with the previous descriptions of BLIMP-1 as leukocyte reactions (19, 56). With respect to IDO1, this is in part an antagonistic regulator of IRF-1-bound promoters, the data pre- due to an amino acid stress response since proliferation can be sented herein identify the IDO1 promoter as belonging to this restored through the addition of excess tryptophan (5, 19). More- group as well (29–31). Indeed, IDO1 may represent a target that over, the starvation stress response in T cells is GCN2 dependent, exemplifies how the regulatory module of BLIMP-1 and IRF tran- since GCN2-knockout T cells are both resistant to the effects of Downloaded from scription factors was reused as the immune system developed (31). IDO1 and fail to differentiate into regulatory T cells (11, 19). At In addition to its ability to compete with the IRF family of present, the downstream components linking GCN2 to the atten- transcription factors, BLIMP-1 can also recruit epigenetic modifi- uation of T cell responses remain uncharacterized. However, the ers to repress target loci. Among the growing number of BLIMP-1 fate of effector and regulatory T cells is controlled and rerouted epigenetic partners are histone deacetylases 1 and 2 (32, 35). through the expression of BLIMP-1 (57, 58). Thus, in cells of both

While displacement of IRF-1 by BLIMP-1 at the IDO1 promoter the monocytic and T cell lineages, the stress-mediated up-regula- http://www.jimmunol.org/ may be sufficient to down-regulate expression, histone modifica- tion of BLIMP-1 may contribute more widely to the immunoregu- tion, and deacetylation in particular, could contribute to maintain- latory effects of the IDO1-GCN2 axis. ing appropriate IDO1 levels. Notably, histone deacetylase inhibi- tors were recently shown to induce IDO1 in dendritic cells, Acknowledgments resulting in suppression of inflammation in a model of graft-vs- We thank Liz Straszynski for cell sorting and Iain Manfield for assistance host disease (53). with the HPLC experiments. HPLC analysis was performed in the Well- Although the IDO1 promoter contains two potential sites for come Trust-funded Centre for Biomolecular Interactions, Astbury Centre BLIMP-1/IRF counterregulation, we observed preferential binding for Structural Molecular Biology, Faculty of Biological Sciences, Univer- to the proximal ISRE site in myeloma cell lines. This does not sity of Leeds. by guest on September 29, 2021 exclude a role for the distal site in the regulation of IDO1, but suggests that it is either not operative in these cells or is potentially Disclosures occupied by other factors. Moreover, the distal ISRE sequence The authors have no financial conflicts of interest. contains C residues at both critical positions for BLIMP-1 binding, and it is unlikely to bind BLIMP-1 with high affinity. Given these References findings, we focused our analysis on the proximal ISRE. A detailed 1. Munn, D. H., M. Zhou, J. T. Attwood, I. Bondarev, S. J. Conway, B. Marshall, C. Brown, and A. L. Mellor. 1998. Prevention of allogeneic fetal rejection by inspection of the sequence demonstrated two potential BLIMP-1 tryptophan catabolism. Science 281: 1191–1193. binding sites adjacent to each other, immediately upstream of the 2. Gurtner, G. J., R. D. Newberry, S. R. Schloemann, K. G. McDonald, and IDO1 transcription start site. This element may therefore provide W. F. Stenson. 2003. Inhibition of indoleamine 2,3-dioxygenase augments trini- trobenzene sulfonic acid colitis in mice. Gastroenterology 125: 1762–1773. a platform for co-operative or multimeric binding, leading to en- 3. Seo, S. K., J. H. Choi, Y. H. Kim, W. J. Kang, H. Y. Park, J. H. Suh, B. K. Choi, hanced transcriptional repression. Alternatively, the extended se- D. S. Vinay, and B. S. Kwon. 2004. 4-1BB-mediated immunotherapy of rheu- quence may be engaging additional zinc finger domains of matoid arthritis. Nat. Med. 10: 1088–1094. 4. Grohmann, U., F. Fallarino, R. Bianchi, C. Orabona, C. Vacca, M. C. Fioretti, and BLIMP-1 other than those that have been previously described. P. Puccetti. 2003. A defect in tryptophan catabolism impairs tolerance in nono- The available data suggest that only the first two of five zinc finger bese diabetic mice. J. Exp. Med. 198: 153–160. domains of BLIMP-1 are responsible for contacting the positive 5. Munn, D. H., M. D. Sharma, D. Hou, B. Baban, J. R. Lee, S. J. Antonia, J. L. Messina, P. Chandler, P. A. Koni, and A. L. Mellor. 2004. Expression of regulatory domain I (PRDI) site within the IFN-␤ promoter (54). indoleamine 2,3-dioxygenase by plasmacytoid dendritic cells in tumor-draining Whether this holds true for all BLIMP-1 targets has not been rig- lymph nodes. J. Clin. Invest. 114: 280–290. 6. Uyttenhove, C., L. Pilotte, I. Theate, V. Stroobant, D. Colau, N. Parmentier, orously tested. However, the presence of a high-molecular mass T. Boon, and B. J. Van den Eynde. 2003. Evidence for a tumoral immune resis- complex in the EMSA in Fig. 5A and the graded inhibition of tance mechanism based on tryptophan degradation by indoleamine 2,3-dioxyge- BLIMP-1-mediated repression observed when mutating one or nase. Nat. Med. 9: 1269–1274. 7. Taylor, M. W., and G. S. Feng. 1991. Relationship between interferon-␥, in- both potential BLIMP-1 binding sites in the extended proximal doleamine 2,3-dioxygenase, and tryptophan catabolism. FASEB J. 5: 2516–2522. ISRE site in Fig. 6C suggest that co-operative binding is more 8. Fallarino, F., and P. Puccetti. 2006. Toll-like receptor 9-mediated induction of the likely. immunosuppressive pathway of tryptophan catabolism. Eur. J. Immunol. 36: 8–11. The finding that BLIMP-1 has the potential to modulate IDO1 9. Grohmann, U., F. Fallarino, and P. Puccetti. 2003. Tolerance, DCs and trypto- expression opens up the possibility that BLIMP-1 plays a more phan: much ado about IDO. Trends Immunol. 24: 242–248. 10. Mellor, A. L., and D. H. Munn. 2004. IDO expression by dendritic cells: toler- expansive role during immune suppression. There are a number of ance and tryptophan catabolism. Nat. Rev. Immunol. 4: 762–774. lines of evidence documenting cell-autonomous effects of IDO1 on 11. Fallarino, F., U. Grohmann, S. You, B. C. McGrath, D. R. Cavener, C. Vacca, gene expression in APCs, including the alteration of metallopro- C. Orabona, R. Bianchi, M. L. Belladonna, C. Volpi, et al. 2006. The combined effects of tryptophan starvation and tryptophan catabolites down-regulate T cell tease expression and prostaglandin synthesis (55). The transient receptor ␨-chain and induce a regulatory phenotype in naive T cells. J. Immunol. expression of BLIMP-1 in these cells could repress other genes 176: 6752–6761. The Journal of Immunology 5777

12. Hinnebusch, A. G. 2005. Translational regulation of GCN4 and the general amino 37. Ball, H. J., A. Sanchez-Perez, S. Weiser, C. J. Austin, F. Astelbauer, J. Miu, acid control of yeast. Annu. Rev. Microbiol. 59: 407–450. J. A. McQuillan, R. Stocker, L. S. Jermiin, and N. H. Hunt. 2007. Characteriza- 13. Harding, H. P., I. Novoa, Y. Zhang, H. Zeng, R. Wek, M. Schapira, and D. Ron. tion of an indoleamine 2,3-dioxygenase-like protein found in humans and mice. 2000. Regulated translation initiation controls stress-induced gene expression in Gene 396: 203–213. mammalian cells. Mol. Cell 6: 1099–1108. 38. Lob, S., A. Konigsrainer, D. Zieker, B. L. Brucher, H. G. Rammensee, G. Opelz, 14. Wek, S. A., S. Zhu, and R. C. Wek. 1995. The histidyl-tRNA synthetase-related and P. Terness. 2009. IDO1 and IDO2 are expressed in human tumors: levo- but sequence in the eIF-2␣ protein kinase GCN2 interacts with tRNA and is required not dextro-1-methyl tryptophan inhibits tryptophan catabolism. Cancer Immunol. for activation in response to starvation for different amino acids. Mol. Cell. Biol. Immunother. 58: 153–157. 15: 4497–4506. 39. Witkiewicz, A. K., C. L. Costantino, R. Metz, A. J. Muller, G. C. Prendergast, 15. Dong, J., H. Qiu, M. Garcia-Barrio, J. Anderson, and A. G. Hinnebusch. 2000. C. J. Yeo, and J. R. Brody. 2009. Genotyping and expression analysis of IDO2 Uncharged tRNA activates GCN2 by displacing the protein kinase moiety from in human pancreatic cancer: a novel, active target. J. Am. Coll. Surg. 208: 781– a bipartite tRNA-binding domain. Mol. Cell 6: 269–279. 787; discussion 787–789. 16. Dever, T. E., L. Feng, R. C. Wek, A. M. Cigan, T. F. Donahue, and 40. Bert, A. G., J. Burrows, C. S. Osborne, and P. N. Cockerill. 2000. Generation of ␣ A. G. Hinnebusch. 1992. Phosphorylation of initiation factor 2 by protein kinase an improved luciferase reporter gene plasmid that employs a novel mechanism GCN2 mediates gene-speciﬁc translational control of GCN4 in yeast. Cell 68: for high-copy replication. Plasmid 44: 173–182. 585–596. 41. Chang, D. H., C. Angelin-Duclos, and K. Calame. 2000. BLIMP-1: trigger for 17. Scheuner, D., B. Song, E. McEwen, C. Liu, R. Laybutt, P. Gillespie, T. Saunders, differentiation of myeloid lineage. Nat. Immunol. 1: 169–176. S. Bonner-Weir, and R. J. Kaufman. 2001. Translational control is required for 42. Lee, A. H., N. N. Iwakoshi, and L. H. Glimcher. 2003. XBP-1 regulates a subset the unfolded protein response and in vivo glucose homeostasis. Mol. Cell 7: of endoplasmic reticulum resident chaperone genes in the unfolded protein re- 1165–1176. sponse. Mol. Cell. Biol. 23: 7448–7459. 18. Ron, D., and P. Walter. 2007. Signal integration in the endoplasmic reticulum unfolded protein response. Nat. Rev. Mol. Cell Biol. 8: 519–529. 43. Calfon, M., H. Zeng, F. Urano, J. H. Till, S. R. Hubbard, H. P. Harding, 19. Munn, D. H., M. D. Sharma, B. Baban, H. P. Harding, Y. Zhang, D. Ron, and S. G. Clark, and D. Ron. 2002. IRE1 couples endoplasmic reticulum load to Nature A. L. Mellor. 2005. GCN2 kinase in T cells mediates proliferative arrest and secretory capacity by processing the XBP-1 mRNA. 415: 92–96. anergy induction in response to indoleamine 2,3-dioxygenase. Immunity 22: 44. Yoshida, H., T. Matsui, A. Yamamoto, T. Okada, and K. Mori. 2001. XBP1 633–642. mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to Downloaded from 20. Manlapat, A. K., D. J. Kahler, P. R. Chandler, D. H. Munn, and A. L. Mellor. produce a highly active transcription factor. Cell 107: 881–891. 2007. Cell-autonomous control of interferon type I expression by indoleamine 45. Novoa, I., Y. Zhang, H. Zeng, R. Jungreis, H. P. Harding, and D. Ron. 2003. 2,3-dioxygenase in regulatory CD19ϩ dendritic cells. Eur. J. Immunol. 37: Stress-induced gene expression requires programmed recovery from translational 1064–1071. repression. EMBO J. 22: 1180–1187. 21. Sood, R., A. C. Porter, D. A. Olsen, D. R. Cavener, and R. C. Wek. 2000. A 46. Dai, W., and S. L. Gupta. 1990. Regulation of indoleamine 2,3-dioxygenase gene mammalian homologue of GCN2 protein kinase important for translational con- expression in human ﬁbroblasts by interferon-␥: upstream control region discrim- trol by phosphorylation of eukaryotic initiation factor-2␣. Genetics 154: inates between interferon-␥ and interferon-␣. J. Biol. Chem. 265: 19871–19877.

787–801. 47. Hassanain, H. H., S. Y. Chon, and S. L. Gupta. 1993. Differential regulation of http://www.jimmunol.org/ 22. Chen, J. J., and I. M. London. 1995. Regulation of protein synthesis by heme- human indoleamine 2,3-dioxygenase gene expression by interferons-␥ and -␣: regulated eIF-2␣ kinase. Trends Biochem. Sci. 20: 105–108. analysis of the regulatory region of the gene and identiﬁcation of an interferon- 23. Williams, B. R. 2001. Signal integration via PKR. Sci. STKE 2001: RE2. ␥-inducible DNA-binding factor. J. Biol. Chem. 268: 5077–5084. 24. Harding, H. P., Y. Zhang, and D. Ron. 1999. Protein translation and folding are 48. Chon, S. Y., H. H. Hassanain, and S. L. Gupta. 1996. Cooperative role of inter- coupled by an endoplasmic-reticulum-resident kinase. Nature 397: 271–274. feron regulatory factor 1 and p91 (STAT1) response elements in interferon-␥- 25. Doody, G. M., S. Stephenson, and R. M. Tooze. 2006. BLIMP-1 is a target of inducible expression of human indoleamine 2,3-dioxygenase gene. J. Biol. Chem. cellular stress and downstream of the unfolded protein response. Eur. J. Immunol. 271: 17247–17252. 36: 1572–1582. 49. Silva, N. M., C. V. Rodrigues, M. M. Santoro, L. F. Reis, J. I. Alvarez-Leite, and 26. Keller, A. D., and T. Maniatis. 1991. Identiﬁcation and characterization of a R. T. Gazzinelli. 2002. Expression of indoleamine 2,3-dioxygenase, tryptophan novel repressor of ␤-interferon gene expression. Genes Dev. 5: 868–879. degradation, and kynurenine formation during in vivo infection with Toxoplasma 27. Gyory, I., J. Wu, G. Fejer, E. Seto, and K. L. Wright. 2004. PRDI-BF1 recruits gondii: induction by endogenous ␥ interferon and requirement of interferon reg-

the histone H3 methyltransferase G9a in transcriptional silencing. Nat. Immunol. ulatory factor 1. Infect. Immun. 70: 859–868. by guest on September 29, 2021 5: 299–308. 50. Chon, S. Y., H. H. Hassanain, R. Pine, and S. L. Gupta. 1995. Involvement of two 28. Keller, A. D., and T. Maniatis. 1991. Selection of sequences recognized by a regulatory elements in interferon-␥-regulated expression of human indoleamine DNA binding protein using a preparative Southwestern blot. Nucleic Acids Res. 2,3-dioxygenase gene. J. Interferon Cytokine Res. 15: 517–526. 19: 4675–4680. 51. Calame, K. L., K. I. Lin, and C. Tunyaplin. 2003. Regulatory mechanisms that 29. Kuo, T. C., and K. L. Calame. 2004. B lymphocyte-induced maturation protein determine the development and function of plasma cells. Annu. Rev. Immunol. 21: (Blimp)-1, IFN regulatory factor (IRF)-1, and IRF-2 can bind to the same regu- 205–230. latory sites. J. Immunol. 173: 5556–5563. ␥ 52. Gong, D., and T. R. Malek. 2007. Cytokine-dependent Blimp-1 expression in 30. Tooze, R. M., S. Stephenson, and G. M. Doody. 2006. Repression of IFN- activated T cells inhibits IL-2 production. J. Immunol. 178: 242–252. induction of class II transactivator: a role for PRDM1/Blimp-1 in regulation of 53. Reddy, P., Y. Sun, T. Toubai, R. Duran-Struuck, S. G. Clouthier, E. Weisiger, cytokine signaling. J. Immunol. 177: 4584–4593. Y. Maeda, I. Tawara, O. Krijanovski, E. Gatza, et al. 2008. Histone deacetylase 31. Doody, G. M., S. Stephenson, C. McManamy, and R. M. Tooze. 2007. PRDM1/ inhibition modulates indoleamine 2,3-dioxygenase-dependent DC functions and BLIMP-1 modulates IFN-␥-dependent control of the MHC class I antigen-pro- regulates experimental graft-versus-host disease in mice. J. Clin. Invest. 118: cessing and peptide-loading pathway. J. Immunol. 179: 7614–7623. 2562–2573. 32. Ren, B., K. J. Chee, T. H. Kim, and T. Maniatis. 1999. PRDI-BF1/Blimp-1 repression is mediated by corepressors of the Groucho family of proteins. Genes 54. Keller, A. D., and T. Maniatis. 1992. Only two of the five zinc fingers of the Dev. 13: 125–137. eukaryotic transcriptional repressor PRDI-BF1 are required for sequence-specific 33. Yu, J., C. Angelin-Duclos, J. Greenwood, J. Liao, and K. Calame. 2000. Tran- DNA binding. Mol. Cell. Biol. 12: 1940–1949. scriptional repression by blimp-1 (PRDI-BF1) involves recruitment of histone 55. Marshall, B., D. B. Keskin, and A. L. Mellor. 2001. Regulation of prostaglandin deacetylase. Mol. Cell. Biol. 20: 2592–2603. synthesis and cell adhesion by a tryptophan catabolizing enzyme. BMC Biochem. 34. Ancelin, K., U. C. Lange, P. Hajkova, R. Schneider, A. J. Bannister, 2: 5. T. Kouzarides, and M. A. Surani. 2006. Blimp1 associates with Prmt5 and directs 56. Matlack, R., K. Yeh, L. Rosini, D. Gonzalez, J. Taylor, D. Silberman, histone arginine methylation in mouse germ cells. Nat. Cell Biol. 8: 623–630. A. Pennello, and J. Riggs. 2006. Peritoneal macrophages suppress T-cell activa- 35. Su, S. T., H. Y. Ying, Y. K. Chiu, F. R. Lin, M. Y. Chen, and K. I. Lin. 2009. tion by amino acid catabolism. Immunology 117: 386–395. Involvement of histone demethylase LSD1 in Blimp-1-mediated gene repression 57. Martins, G. A., L. Cimmino, M. Shapiro-Shelef, M. Szabolcs, A. Herron, during plasma cell differentiation. Mol. Cell. Biol. 29: 1421–1431. E. Magnusdottir, and K. Calame. 2006. Transcriptional repressor Blimp-1 regu- 36. Metz, R., J. B. Duhadaway, U. Kamasani, L. Laury-Kleintop, A. J. Muller, and lates T cell homeostasis and function. Nat. Immunol. 7: 457–465. G. C. Prendergast. 2007. Novel tryptophan catabolic enzyme IDO2 is the pre- 58. Kallies, A., E. D. Hawkins, G. T. Belz, D. Metcalf, M. Hommel, L. M. Corcoran, ferred biochemical target of the antitumor indoleamine 2,3-dioxygenase inhibi- P. D. Hodgkin, and S. L. Nutt. 2006. Transcriptional repressor Blimp-1 is essen- tory compound D-1-methyl-tryptophan. Cancer Res. 67: 7082–7087. tial for T cell homeostasis and self-tolerance. Nat. Immunol. 7: 466–474.