Microglia Activation-Induced Cell Death of Brain Gene-1 As An

Role of Antiproliferative B Cell Translocation Gene-1 as an Apoptotic Sensitizer in Activation-Induced Cell Death of Brain Microglia This information is current as of October 1, 2021. Heasuk Lee, Sanghoon Cha, Myung-Shik Lee, Gyeong Jae Cho, Wan Sung Choi and Kyoungho Suk J Immunol 2003; 171:5802-5811; ; doi: 10.4049/jimmunol.171.11.5802

http://www.jimmunol.org/content/171/11/5802 Downloaded from

References This article cites 52 articles, 14 of which you can access for free at: http://www.jimmunol.org/content/171/11/5802.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 October 1, 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 © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Role of Antiproliferative B Cell Translocation Gene-1 as an Apoptotic Sensitizer in Activation-Induced Cell Death of Brain Microglia1

Heasuk Lee,* Sanghoon Cha,† Myung-Shik Lee,‡ Gyeong Jae Cho,* Wan Sung Choi,* and Kyoungho Suk2*

Mouse brain microglial cells undergo apoptosis on exposure to inﬂammatory stimuli, which is considered as an autoregulatory mechanism to control their own activation. Here, we present evidence that an antiproliferative B cell translocation gene 1 (BTG1) constitutes a novel apoptotic pathway of LPS/IFN-␥-activated microglia. The expression of BTG1 was synergistically enhanced by LPS and IFN-␥ in BV-2 mouse microglial cells as well as in primary microglia cultures. Levels of BTG1 expression inversely correlated with a proliferative capacity of the microglial cells. Tetracycline-based conditional expression of BTG1 not only suppressed microglial proliferation but also increased the sensitivity of microglial cells to NO-induced apoptosis, suggesting a novel Downloaded from mechanism of cooperation between LPS and IFN-␥ in the induction of microglial apoptosis. An increase in BTG1 expression, however, did not affect microglial production of NO, TNF-␣, or IL-1␤, indicating that the antiproliferative BTG1 is important in the activation-induced apoptosis of microglia, but not in the activation itself. The synergistic action of LPS and IFN-␥ in the microglial BTG1 induction and apoptosis was dependent on the Janus kinase/STAT1 pathway, but not IFN-regulatory factor-1, as demonstrated by a pharmacological inhibitor of Janus kinase (AG490), STAT1 dominant negative mutant, and IFN-regulatory factor-1-deﬁcient mice. Taken together, antiproliferative BTG1 may participate in the activation-induced cell death of microglia http://www.jimmunol.org/ by lowering the threshold for apoptosis; BTG1 increases the sensitivity of microglia to apoptogenic action of autocrine cytotoxic mediator, NO. Our results point out an important link between the proliferative state of microglia and their sensitivity to apoptogenic agents. The Journal of Immunology, 2003, 171: 5802–5811.

icroglia are a type of neuroglia that support, nurture, tive diseases (4). From this point of view, one can speculate that and protect the neurons, maintaining homeostasis of the autoregulatory mechanisms that control the microglial activa- M the fluid that bathes neurons. Although the ontogeny tion may exist in vivo, and the failure of these autoregulatory of microglial cells has long been debated, recent works using mechanisms may be responsible for the deleterious effects of mi- by guest on October 1, 2021 mAbs specific for microglial cells indicated that these cells are croglial activation. Accordingly, the elucidation of molecular closely related to monocytes and macrophages (1). Microglia func- mechanisms underlying the autoregulation of microglial activation tion as macrophages in the CNS; they migrate to areas of injured may enhance our understanding of pathogenesis of neurodegen- nervous tissue, and they engulf and destroy microbes and cellular erative diseases. Recently, activated macrophages have been debris (2). They also secrete inflammatory cytokines and toxic shown to undergo apoptosis (5Ð7). It has been suggested that the mediators, which may amplify the inflammatory responses (3, 4). apoptosis of activated macrophages is one mechanism whereby an Activation of microglial cells may be intended to protect neurons organism may regulate immune and inflammatory responses in- at first. However, activation of microglial cells and inflammatory volving macrophages (7). We and others have demonstrated that a products derived from them have been also implicated in the neu- similar regulatory mechanism exists for microglial cells (8Ð10) ronal destruction commonly observed in various neurodegenera- and astrocytes (11) as well. We have previously shown that microglial cells and astrocytes undergo apoptosis upon inflammatory activation in a manner similar to that of activation-induced cell 3 *Department of Anatomy and Neurobiology and Research Institute of Natural Sci- death (AICD) of lymphocytes, and NO acts as an autocrine cy- ence, Gyeongsang National University College of Medicine, Institute of Health Sci- totoxic mediator in this process (9, 11). Inflammatory stimuli † ences, Jinju, Korea; Division of Food Science and Biotechnology, College of Agri- played a dual role in the microglial apoptosis; they not only in- culture and Life Sciences, Kangwon National University, Chunchon, Korea; and ‡Department of Medicine, Samsung Medical Center, Sungkyunkwan University duced NO production through IFN-regulatory factor-1 (IRF-1) but School of Medicine, Seoul, Korea also initiated the NO-independent apoptotic pathway via Received for publication May 12, 2003. Accepted for publication September caspase-11 induction followed by caspase-3 activation (12). How- 17, 2003. ever, a role of the inflammatory stimuli in the microglial AICD The costs of publication of this article were defrayed in part by the payment of page other than the induction of NO production and caspase-11 expres- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. sion has been also suggested, because pretreatment of microglia ␥ 1 This work was supported by Korea Research Foundation Grant KRF-2002-015- with LPS/IFN- markedly increased the sensitivity of microglia to ES0018 and by a grant from the Brain Korea 21 Project in 2003 (E021). M.-S.L. was NO-induced apoptosis (12). Therefore, it is of great interest to supported by a National Research Laboratory grant from the Korea Institute of Sci- ence and Technology Evaluation and Planning (2000-N-NL-01-C-232). 2 Address correspondence and reprint requests to Dr. Kyoungho Suk, Department of 3 Abbreviations used in this paper: AICD, activation-induced cell death; IRF-1, in- Anatomy and Neurobiology, Gyeongsang National University College of Medicine, 92 terferon-regulatory factor-1; JAK, Janus kinase; BTG1, B cell translocation gene 1; Chilam-dong, Jinju, Kyungnam 660-751, Korea. E-mail address: [email protected]. NMMA, N-monomethyl-L-arginine; SNAP, S-nitroso-N-acetylpenicillamine; SNP, ac.kr sodium nitroprusside; AG490, ␣-cyano(3,4-dihydroxy)-N-benzylcinnamide.

identify such LPS/IFN-␥-inducible apoptotic mediator(s) to under- plemented with 10% FBS, 2 mM glutamine, and penicillin-streptomy- stand the mechanism of microglial AICD to the fullest extent. cin (Life Technologies, Gaithersburg, MD). B cell translocation gene 1 (BTG1) was first identified as a trans- PCR-based subtractive hybridization cloning location gene in a case of B cell chronic lymphocytic leukemia The PCR-Select cDNA subtraction kit (Clontech, Palo Alto, CA) was used (13). BTG1 is constitutively expressed in quiescent cells, whereas to perform PCR-based subtractive hybridization. In brief, excess driver its expression is down-regulated as the cells enter the growth cycle cDNA (from untreated BV-2 cells) and adaptor-ligated tester cDNA (from (14, 15). Experiments with the forced expression of the gene BV-2 cells treated for 4 h with 100 ng/ml LPS plus 10 U/ml IFN-␥) were showed that BTG1 negatively regulates cell proliferation (16). The hybridized to subtract common cDNAs. The remaining unhybridized antiproliferative properties of BTG1 have been observed in many cDNA that represented differentially expressed genes was selectively amplified by PCR using primers specific for the adaptor sequences. The prod- different cell types including PBMC (16), testis (17), and macro- ucts of subtractive hybridization were then cloned into a T/A cloning vec- phages (18). Especially, in macrophages, we and others have found tor using TOPO TA cloning kit (Invitrogen, Carlsbad, CA) to make a that the levels of BTG1 expression negatively correlate with cell subtracted cDNA library. More than one hundred clones from the sub- proliferation (18) and that the gene expression colocalizes with an tracted cDNA library were differentially screened using the PCR-Select differential screening kit (Clontech) according to the manufacturer’s in- apoptotic marker (19). Moreover, overexpression of BTG1 in structions. BTG1 was one of the several clones that represented differen- NIH3T3 cells inhibited the cell proliferation (16) and increased tially expressed mRNA. apoptotic frequency (19). These previous findings and the func- Conditional expression tional similarity between macrophages and brain microglia led us to speculate that BTG1 might play a role in both limiting cell Synthetic inducible expression system on the basis of the tetracycline re- proliferation and inducing apoptosis in microglia. pressor of Escherichia coli (Tet-regulated expression system; Life Tech- nologies) (23) was used to reversibly control the expression of BTG1 in the Downloaded from In this work, we have demonstrated that BTG1 exerts antipro- microglial cells. In brief, the EcoRI/NheI fragment of BTG1 cDNA (1528 liferative activity in microglia and that the antiproliferative BTG1 bp) containing the entire coding region was cloned into the EcoRI/SpeI site constitutes a novel apoptotic pathway of LPS/IFN-␥-activated mi- of pTet-Splice plasmid (Life Technologies) that contains the Tet promoter croglia. LPS/IFN-␥-induced BTG1 participated in microglial apo- (consisting of the regulatory sequences from the tetracycline resistance ptosis by increasing the sensitivity of microglia to NO, which acts operon of Tn10 upstream of a minimal human CMV promoter) (Fig. 2). NheI and SpeI share a compatible cohesive end. The resulting construct as an autocrine apoptotic mediator in AICD of microglia. BTG1 (pTet-splice/BTG1) was contransfected into BV-2 cells along with induction, however, did not influence the production of inflamma- pSV2neo (24) and pTet-tTAk (Life Technologies), which contains the gene http://www.jimmunol.org/ tory mediators from activated microglia. The enhancement of for the tetracycline trans activator under the control of Tet promoter. Stable ␮ BTG1 expression by LPS/IFN-␥ and the ensuing apoptosis-sensi- transfectants were selected in the presence of G418 (500 g/ml) and tetracycline (1 ␮g/ml) 2 days after transfection according to the manufactur- tizing effects in microglia were dependent on Janus kinase (JAK)/ er’s instructions. For the conditional expression of BTG1, selected stable STAT1, but not IRF-1. Taken collectively, our results indicate that transfectants were grown in the culture medium without tetracycline BTG1 may lower the threshold for microglial apoptosis by its an- overnight. tiproliferative activity, and this BTG1 action seems to be indepen- Assessment of cytotoxicity by MTT assay or trypan blue dent of the microglial activation process (at least with respect to exclusion assay the production of NO, TNF-␣, and IL-1␤). Cells (3 ϫ 104 cells in 200 ␮l/well) were seeded in 96-well plates and by guest on October 1, 2021 treated with LPS and IFN-␥ for the indicated time periods. After the treat- Materials and Methods ment, the medium was removed and MTT (0.5 mg/ml) was added, fol- Reagents lowed by incubation at 37¡Cfor2hinaCO2 incubator. After a brief centrifugation, supernatants were carefully removed, and DMSO was LPS, N-monomethyl-L-arginine (NMMA), S-nitroso-N-acetylpenicillamine added to the cells. After insoluble crystals were completely dissolved, (SNAP), sodium nitroprusside (SNP), and tetracycline were obtained from OD was measured using a Thermomax microplate reader (Molecular Sigma-Aldrich (St. Louis, MO). Recombinant mouse IFN-␥ was purchased 540 Devices, Sunnyvale, CA). For trypan blue exclusion assay, cell suspension from R&D Systems (Minneapolis, MN). ␣-Cyano(3,4-dihydroxy)-N-ben- was mixed with the same volume of 0.4% trypan blue solution (Life Tech- zylcinnamide (AG490) was from Calbiochem (La Jolla, CA). All other nologies). Afterward, the number of stained cells and the total number of chemicals were obtained from Sigma-Aldrich, unless stated otherwise. cells were counted using a hemocytometer (Marienfeld, Lauda-Koenig- Mice and cells shofen, Germany). ␣ ␤ Mice with a targeted mutation in the IRF-1 gene (homozygous mice and TNF- and IL-1 ELISA their heterozygous littermates) were kindly provided by Dr. Y. C. Sung at TNF-␣ and IL-1␤ secreted in microglial culture supernatants was measured Postech (Pohang, Korea) (20) and bred in a virus-free facility at the Sam- ␣ ␤ Ϫ/Ϫ ϩ/Ϫ ϩ/ϩ by specific ELISA using rat monoclonal anti-mouse TNF- or IL-1 Ab as sung Medical Center. The IRF-1 , IRF-1 , and IRF-1 colonies ␣ ␤ ϩ/Ϫ Ϫ/Ϫ ϩ/Ϫ capture Ab and goat biotinylated polyclonal anti-mouse TNF- or IL-1 were maintained by mating IRF-1 to either IRF-1 or IRF-1 mice. Ab as detection Ab (ELISA development reagents; R&D Systems). The The genotype of all IRF-1 mice was determined by PCR of tail DNA using biotinylated anti-TNF-␣ or -IL-1␤ Ab was detected by sequential incuba- the following three primers as previously described (20): IRF-1 forward tion with streptavidin-HRP conjugate and chromogenic substrates. primer, TTC CAG ATT CCA TGG AAG CAC GC; IRF-1 reverse primer, ATG GCA CAA CGG AAG TTT GCC; neor reverse primer, ATT CGC Nitrite quantification CAA TGA CAA GAC GCT GG. PCR with these three primers amplifies ϩ Ϫ After cells (3 ϫ 104 cells in 200 ␮l/well) were treated with activating a 900-bp sequence for IRF-1 allele and a 700-bp sequence for IRF-1 Ϫ allele from genomic DNA. Mouse primary microglial cells were prepared agents in 96-well plates, NO2 in culture supernatants was measured to as previously described with minor modifications (9, 21). In brief, fore- assess NO production in microglial cells. Fifty microliters of sample ali- ␮ brains of newborn wild-type or IRF-1-deficient mice were chopped and quots were mixed with 50 l of Griess reagent (1% sulfanilamide-0.1% dissociated by trypsinization and mechanical disruption. The cells were naphthylethylenediamine dihydrochloride-2% phosphoric acid) in 96-well plates and incubated at 25¡C for 10 min. The OD550 was measured on a seeded into poly-L-lysine-coated flasks. After in vitro culture for 10 days, Ϫ microglial cells were detached by rapid and gentle shaking of the cul- microplate reader. NaNO2 was used as the standard to calculate NO2 ture flasks and seeded into plastic surfaces. After an additional 1-h concentrations. incubation, nonadherent cells were removed by replacing culture me- DNA ploidy analysis and annexin V-binding assay dium. The purity of microglial cultures was Ͼ92% as determined by isolectin B4 staining (data not shown). BV-2 mouse microglial cell line For DNA ploidy analysis, cells were suspended in PBS, 5 mM EDTA and originally developed by Dr. V. Bocchini at the University of Perugia fixed by adding 100% ethanol dropwise. RNase A (40 ␮g/ml) was added (Perugia, Italy) (22) was generously provided by Dr. E. Choi at Korea to the resuspended cells, and then incubation was conducted at room tem- University (Seoul, Korea). The cell line was maintained in DMEM sup- perature for 30 min. Propidium iodide (50 ␮g/ml) was then added for flow 5804 BTG1 IN MICROGLIAL APOPTOSIS cytometric analysis (FACSVantage; BD Biosciences, San Jose, CA). The 48 h after the transfection, the cells were treated with NO donor. After percentage of cells in S phase was determined by a cell cycle analysis another 24 h, the cells were fixed with 0.5% glutaraldehyde for 10 min at program, ModFitLT (Verity Software House, Topsham, ME). Apoptosis of room temperature and stained with 5-bromo-4-chloro-3-indolyl ␤-D-galac- microglial cells was also evaluated by annexin V-binding assay. In brief, topyranoside (1 mg/ml) in 4 mM potassium ferricyanide, 4 mM potassium cells were stained with annexin V-FITC, a fluorescent annexin V conjugate ferrocyanide, 2 mM magnesium chloride at 37¡C for the detection of blue (Clontech), and then subjected to flow cytometric analysis. cells expressing lacZ. Alternatively, in some experiments, BV-2 cells were cotransfected with a dominant negative mutant of STAT1 cDNA (provided Detection of DNA ladder and DNA fragmentation assay by T. Hirano, Osaka University, Osaka, Japan) along with 0.2 ␮goflacZ gene. At 48 h after the transfection, the cells were pretreated with 100 Agarose gel electrophoresis of genomic DNA was conducted to detect a ng/ml LPS plus 10 U/ml IFN-␥ for 16 h before the treatment with NO DNA ladder. For isolation of genomic DNA, BV-2 cells were lysed by donor for another 24 h. The cells were then fixed and stained as described incubation in the extraction buffer overnight at 55¡C (10 mM Tris-HCl (pH above. At least 250 blue cells were counted for each experiment, and trans- 8.0), 0.1 M EDTA, 0.5% SDS, 100 mM NaCl), followed by phenol-chlo- fection efficiency was 18Ð31%. The percentage of apoptotic cells was roform extraction and ethanol precipitation. The final pellet was dissolved based on the morphology of blue cells coexpressing either BTG1 or dom- in distilled water containing 0.1 mg/ml RNase A. Isolated genomic DNA inant negative mutant of STAT1 and lacZ. Dark blue and condensed cells was electrophoresed on 1.5% agarose gel and stained with ethidium bro- were considered to be apoptotic. For the analysis of BTG1 expression, the mide to detect internucleosomal cleavage. For quantitative analysis of cells were cotransfected with a dominant negative STAT1 along with neo- DNA fragmentation, the Cell Death Detection ELISA system (Roche Ap- resistant gene construct (pSV2neo). Seven days later, G418 (500 ␮g/ml)- plied Science, Indianapolis, IN) was used. The assay is based on the quan- resistant cells were selected, pooled, and then subjected to analysis of LPS/ titative sandwich enzyme immunoassay principle using mouse mAbs di- IFN-␥-induced BTG1 expression. rected against DNA and histones. This allows the specific determination of mono- and oligonucleosomes in the cytoplasmic fraction of cell lysates. After antihistone Ab was fixed adsorptively on the wall of the microplate Statistical analysis wells, the cytoplasmic fractions of the samples were added. In the next Downloaded from step, anti-DNA Ab conjugated to peroxidase was added to detect released All data are means Ϯ SEM from three or more independent experiments. nucleosomes in the sample. After removal of unbound peroxidase conju- Statistical comparison between different treatments was done by either Stu- gate, the amount of peroxidase retained in the immunocomplex was de- dent’s t test or one-way ANOVA with Dunnett’s multiple comparison test termined photometrically. using the GraphPad Prism program (GraphPad Software, San Diego, CA). Differences of p Ͻ 0.05 were considered statistically significant. RNA analysis

Total RNA was extracted from BV-2 cells or primary microglial cells by http://www.jimmunol.org/ a sequential addition of 4 M guanidinium thiocyanate, 2 M sodium acetate, Results and acid phenol-chloroform. Reverse transcription was conducted using Superscript (Life Technologies-BRL) and oligodeoxythymidylate primer. Induction of BTG1 expression and microglial apoptosis by LPS/ PCR amplification using primer sets specific for IRF-1 was conducted at IFN-␥ ␤ 60¡C annealing temperature for 30 cycles. PCR for BTG1 and -actin was Previously, we have observed that microglial cells activated by conducted at 55¡C annealing temperature for 30 cycles. Nucleotide se- ␥ quences of the primers were based on published cDNA sequences of mouse inflammatory stimuli such as LPS and IFN- undergo apoptosis as BTG1, ␤-actin, and IRF-1 (BTG1 forward, GAT TGG ACT GAG CAG a possible autoregulatory mechanism for their own activation TCA GGA; BTG-1 reverse, TCT TAA CCT GAT ACA GTC ATC; states (9), and IRF-1-mediated NO production and caspase-11 in- ␤-actin forward, ATC CGT AAA GAC CTC TAT GC; ␤-actin reverse, duction play a pivotal role in this process (12). Our previous work AAC GCA GCT CAG TAA CAG TC; IRF-1 forward, TCT GAG TGG ␥ by guest on October 1, 2021 CAT ATG CAG ATG GAC; IRF-1 reverse, GGT CAG AGA CCC AAA also suggested that LPS/IFN- pretreatment might lower the CTA TGG TCG). The lengths of PCR product for each primer set were 261 threshold for microglial apoptosis through the induction of medi- bp for BTG1, 287 bp for ␤-actin, and 426 bp for IRF-1, respectively. For ators or pathways other than IRF-1 and caspase-11 (12). Thus, we Northern blot analysis, isolated total RNA (20 ␮g/sample) was electropho- initiated the current study by searching for LPS/IFN-␥-induced resed in a 1.0% agarose gel containing 0.7% formaldehyde, 20 mM MOPS, genes in BV-2 mouse microglial cells by PCR-based subtractive 5 mM sodium acetate, and 1 mM EDTA. RNA was transferred to the Nytran membrane (Schleicher & Schuell, Keene, NH). After UV cross- hybridization cloning in an attempt to better understand the mech- linking, the membranes were hybridized with a BamHI fragment (511 bp) anisms by which inflammatory stimuli induce the microglial apo- of BTG1 cDNA probe (106 cpm/ml). The membranes were then washed at ptosis. One of the clones identified by subtractive hybridization 65¡C, dried, and exposed to x-ray films. was BTG1. BTG1 was first identified in 1991 by molecular cloning Western blot analysis of chromosomal breakpoint in B cells (13). A negative correlation between BTG1 expression and cell proliferation has been observed Cells were lysed in triple-detergent lysis buffer (50 mM Tris-HCl (pH 8.0), in various cell types (16Ð18). In particular, BTG1 expression was 150 mM NaCl, 0.02% sodium azide, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1 mM PMSF). Protein concentration in cell lysates associated with both apoptosis and inhibition of cell proliferation was determined using the Bio-Rad protein assay kit. An equal amount of in macrophages (19), which are closely related to microglia. Thus, protein for each sample was separated by 12% SDS-PAGE and transferred in the current work, we explored the possibility that BTG1 might to Hybond ECL nitrocellulose membranes (Amersham, Arlington Heights, negatively regulate proliferation of microglial cells as observed in IL). The membranes were blocked with 5% skim milk and sequentially incubated with primary Abs (rabbit polyclonal anti-mouse BTG1, Genemed macrophages and this antiproliferative action of BTG1 might be Synthesis, South San Francisco, CA; rabbit polyclonal anti-human/mouse related with microglial apoptosis. Before testing this possibility, phospho-STAT1 (Tyr701), Cell Signaling Technology, Beverly, MA; rabbit the enhancement of BTG1 expression by LPS/IFN-␥ was first con- polyclonal anti-mouse IRF-1, Santa Cruz Biotechnology, Santa Cruz, CA; firmed by Northern blot as well as Western blot analysis (Fig. 1, ␣ mouse monoclonal anti- -tubulin, Sigma-Aldrich) and HRP-conjugated AÐC). The expression of BTG1 at both mRNA and protein levels secondary Abs (anti-rabbit or -mouse IgG, Amersham), followed by ECL ␥ detection (Amersham). Mouse BTG1-specific Ab was generated against was markedly increased by LPS/IFN- treatment in BV-2 cells two synthetic peptide sequences of mouse BTG1 protein (102LTLWVD- (Fig. 1, A and B) as well as mouse primary microglia cultures (Fig. PYEVSYRIGEDG119 and 156GRTSPSKNYNMM167; Ref.25) by custom 1C). However, either LPS or IFN-␥ alone did not significantly Ab service of Genemed Synthesis. affect BTG1 expression. An increase in the level of BTG1 expres- Transient transfection sion by LPS/IFN-␥ treatment correlated with microglial cell death; cotreatment with LPS plus IFN-␥ significantly increased cell death BV-2 cells in six-well plates were cotransfected with 1 ␮g of mouse BTG1 cDNA under control of CMV promoter (pcDNA3-BTG1) together with 0.2 of both BV-2 and primary microglia cultures compared with the ␮goflacZ gene (pCH110; Pharmacia, Peapack, NJ) using lipofectAMINE single treatment of either LPS or IFN-␥ alone (Fig. 1D). The re- reagent (Life Technologies) according to the supplier’s instructions. At duction of cell viability was similarly observed by trypan blue The Journal of Immunology 5805

FIGURE 2. Construction of tetracycline-regulated BTG1 expression system in microglia. A, BTG1 cDNA fragment containing the entire coding region was cloned into the EcoRI/SpeI site of pTet-Splice plasmid under the control of the Tet promoter to generate pTet-Splice/BTG1 (see Mate- rials and Methods). B, After BV-2 cells were cotransfected with pTet- Downloaded from Splice/BTG1, pTet-tTAK, and pSV2neo, stable transfectants were selected in the presence of G418 and tetracycline. In a selected clone, BV-2/tet3, BTG1 induction was evaluated by removal of tetracycline from the growth medium. Western blot detection of BTG1 protein and densitometric analysis indicated ϳ5 fold increase in BTG1 protein level in the cells grown in the absence of tetracycline (Ϫ tet) compared with the cells grown in the

presence of tetracycline (ϩ tet). BTG1 protein level was normalized to http://www.jimmunol.org/ FIGURE 1. BTG1 induction and microglial cell death by LPS/IFN-␥. ␣-tubulin. Results are representative of three independent experiments. After treatment of BV-2 cells with LPS (100 ng/ml) and IFN-␥ (10 U/ml) either alone or in combination, BTG1 expression at mRNA (4 h after treatment) or protein level (8 h after treatment) was assessed by Northern blot analysis (A) or Western blot analysis (B), respectively. Only the combina- detected in BV-2/tet3, and this was enhanced 5.2-fold by the re- tion of LPS and IFN-␥ strongly increased BTG1 expression. C, Enhance- moval of tetracycline (Fig. 2B). The proliferative capacity of BV- ment of BTG1 expression by LPS/IFN-␥ (4 h after treatment) was similarly 2/tet3 cells in the presence or absence of tetracycline was then observed in mouse primary microglia cultures by RT-PCR. Ethidium bro- compared to evaluate the potential antiproliferative activity of mide staining of rRNA of 28S and 18S or detection of ␣-tubulin or ␤-actin BTG1 in microglia. Enhancement of BTG1 expression by the re- by guest on October 1, 2021 was done to confirm the equal loading and integrity of the samples. The moval of tetracycline suppressed the proliferation of BV-2/tet3 amplified products were not observed in RT-PCR without reverse tran- cells (Fig. 3, A and B) and led to the reduction in the percentage of scriptase (data not shown). D, When BV-2 cells or primary microglia cul- S phase cells (Fig. 3C). Induction of BTG1 expression per se, tures were treated with LPS and IFN-␥ alone or in combination for either 24 h (BV-2 cells) or 72 h (primary microglia cultures), Ͼ50% cytotoxicity however, did not significantly affect cell viability, as demonstrated was observed only after LPS plus IFN-␥ treatment. Viability of untreated by trypan blue staining (data not shown). Similarly, treatment of cells was set to 100%. Results are mean Ϯ SEM of three independent parental BV-2 cells with LPS/IFN-␥ that increased BTG1 expres- -Statistically significant differences from untreated control sion also attenuated the proliferative capacity of the cells as evi ,ء .experiments (p Ͻ 0.05). denced by the reduction of cells in S phase (24.5% in untreated cells, 11.6% in LPS/IFN-␥-treated cells). These results again support the antiproliferative activity of BTG1 in microglia. Because LPS/ exclusion assays (data not shown). The apoptotic nature of micro- IFN-␥ treatment itself causes cytotoxicity by inducing NO production, glial cell death under this condition has been previously deter- the experiments were done in the presence of NMMA, an inhibitor of mined by appearance of DNA ladder, subdiploidy cells, and nu- NO synthase. The percentage of microglia in S phase of cell cycle was clear condensation and fragmentation (9). assessed by DNA ploidy analysis of viable cells.

Antiproliferative activity of BTG1 in microglia Role of BTG1 in microglial apoptosis On the basis of the known antiproliferative properties of BTG1 in Previously, BTG1 has been shown to inhibit cell proliferation and macrophages and other cell types, we next sought to determine induce apoptosis when overexpressed in NIH3T3 cells (19). More- whether BTG1 negatively regulates microglial proliferation. Be- over, current results indicated that BTG1 was induced by LPS/ cause the antiproliferative nature of BTG1 may impede the gener- IFN-␥ and that BTG1 inhibited microglial proliferation. On the ation of stable transfectants that constitutively overexpress the basis of these results, we next asked whether BTG1 plays any role gene, we used a tetracycline-based conditional expression system in the apoptosis of microglia following inﬂammatory activation. to control the expression of BTG1 in BV-2 cells. A synthetic in- First of all, we investigated a possible correlation between BTG1 ducible expression system on the basis of the tetracycline repressor expression and the propensity of microglial cells to undergo apo- of E. coli (23) was constructed to reversibly control the expression ptosis on exposure to NO, an apoptosis-inducing agent in micro- of BTG1 in the microglial cells (Fig. 2A). BV-2/tet3 cells, a de- glia. Augmentation of BTG1 expression by the removal of tetra- rivative of the BV-2 cell line with tetracycline-controlled expres- cycline from the cultivating medium increased the sensitivity of sion of BTG1, were generated. In BV-2/tet3, transfected BTG1 BV-2/tet3 cells to the toxicity of the NO donor, SNAP, at a low gene expression is turned on in the absence of tetracycline in the concentration (0.2 mM; Fig. 4). Similar results were obtained by cultivating medium. A low constitutive expression of BTG1 was another NO donor, SNP (data not shown). Previously, we have 5806 BTG1 IN MICROGLIAL APOPTOSIS

though SNAP has been shown to kill other types of cells by necrosis, an absolute majority of BV-2 cells underwent apoptosis after exposure to SNAP as determined by both trypan blue exclusion assays and Hoechst nuclear staining (data not shown). Al- though, in the particular field of Hoechst staining results shown in Fig. 4C, ϳ30% of microglia cells appeared to be apoptotic, an average percentage of apoptotic cells from 10 microscopic fields was ϳ50%. This was in agreement with the percentage of cell death based on trypan blue exclusion assays (data not shown) and percent apoptosis in Fig. 4D. Role of BTG1 in inflammatory activation of microglia Because LPS or IFN-␥ is one of the major activators of macrophages and microglial cells in the inflammatory responses (25), we next asked whether LPS/IFN-␥-induced BTG1 plays a role in the inflammatory activation processes of microglia. When the production of inflammatory mediators such as TNF-␣, IL-1␤, and NO was compared among BV-2/tet3 microglial cells with different levels of BTG1 expression (with or without tetracycline in culture medium), no significant difference was found (Fig. 5). The results Downloaded from indicate that BTG1 is involved in the apoptosis of activated mi- FIGURE 3. Antiproliferative activity of BTG1 in microglia. Prolifera- croglial cells, but not activation processes (at least in terms of NO tive capacity of BV-2/tet3 cells was compared before and after removal of and proinflammatory cytokine production). tetracycline. When BTG1 expression was induced in BV-2/tet3 by tetra- Involvement of JAK/STAT1 pathway in LPS/IFN-␥ action in cycline removal from the growth medium, the proliferative capacity of the cells was decreased as determined by MTT assay (A) or phase contrast microglia http://www.jimmunol.org/ microscopy of the cells at day 3 during the MTT assay (ϫ200; B). Results To delineate the signaling pathway of LPS/IFN-␥-induced BTG1 Statistically signif- expression and microglial apoptosis, the possibility of involvement ,ء .are mean Ϯ SEM of three independent experiments ϩ icant differences between the cells grown in the presence ( tet) and those of various signaling components was investigated with a focus on Ϫ Ͻ Ϫ grown in the absence ( tet) of tetracycline (p 0.05). BTG1 induction ( the JAK/STAT pathway and IRF-1. IFN-␥, but not LPS, induced tet) also led to the reduction in S phase cells from 26.9% to 15.4% as activation of STAT1 in BV-2 microglial cells (Fig. 6A, left). Phos- determined by flow cytometric analysis of DNA ploidy and cell cycle analysis program at day 3 (C). Results are representative of four independent phorylation of STAT1 started at 30 min and peaked at 1 h after ␥ experiments. x-axis, DNA content; y-axis, cell number. stimulation with IFN- . A similar pattern of STAT1 phosphory- lation was observed after LPS/IFN-␥ treatment (Fig. 6A, right). by guest on October 1, 2021 Inhibition of STAT1 pathway by the transfection of a dominant determined that SNAP or SNP (0.2 mM) alone does not induce a negative mutant of STAT1 partially abolished the apoptosis-sen- significant change in the viability of BV-2 microglial cells (12). sitizing effect of LPS/IFN-␥ treatment (Fig. 6B). Compared with Thus, an increase in BTG1 expression seems to render microglial the empty vector-transfected cells, transfection of the dominant cells sensitive to an otherwise nontoxic dose of NO donors. Cy- negative mutant of STAT1 attenuated the effect of LPS/IFN-␥; i.e., totoxicity was based on the number of apoptotic cells assessed by LPS/IFN-␥ pretreatment sensitized microglia to NO donor-in- flow cytometric analysis of annexin V-binding cells (Fig. 4A)or duced apoptosis in the empty vector-transfected cells, whereas in quantification of DNA fragmentation using specific Ab against ap- the BV-2 cells transfected with dominant negative mutant of optotic nucleosomes (Fig. 4B). An MTT assay was not conducted, STAT1 the sensitizing effect of LPS/IFN-␥ on NO-induced apo- because BTG1 expression by itself influences cellular proliferation, ptosis was significantly reduced. The percentage of apoptosis of thereby obscuring the assay results. A prominent oligonucleosomal transfected cells was evaluated by separately counting dark blue cleavage of genomic DNA or nuclear condensation was also de- condensed apoptotic cells and healthy blue cells coexpressing tected in BV-2/tet3 microglial cells with a high BTG1 expression lacZ. To further investigate the role of the JAK/STAT1 pathway in (without tetracycline in culture medium) by agarose gel electro- the LPS/IFN-␥-induced BTG1 expression and microglial apopto- phoresis of genomic DNA or Hoechst nuclear staining, respec- sis, a JAK inhibitor (AG490) was used. A specific inhibition of tively (Fig. 4C). To eliminate a possible artifact of tetracycline JAK by AG490 diminished LPS/IFN-␥-induced BTG1 expression treatment and to further demonstrate the role of BTG1 in micro- (Fig. 6C), supporting the theory that the JAK/STAT1 pathway me- glial apoptosis, parental BV-2 cells were transiently cotransfected diates LPS/IFN-␥-induced BTG1 expression and the subsequent with BTG1 cDNA along with lacZ; then the effect of BTG1 ex- apoptosis sensitization in microglia. Treatment of the cells with pression on microglial apoptosis was again analyzed. Morpholog- AG490 also diminished LPS/IFN-␥-induced NO production as ical analysis of lacZ-expressing blue cells supported the role of well as cell death (Fig. 7A). The AG490 inhibition of BTG1 ex- BTG1 in the sensitization of the microglial cells to NO-induced pression was, however, not due to its inhibitory effect on NO pro- apoptosis (Fig. 4D). The number of condensed and dark blue ap- duction, because a similar inhibition of NO production by NOS optotic cells was significantly greater in BTG1-transfected cells inhibitor (NMMA) did not affect BTG1 expression (Fig. 7). A de- compared with the empty vector-transfected cells when treated finitive evidence for the involvement of the STAT1 pathway in with 0.2 mM SNAP. Based on our previous findings that LPS/ LPS/IFN-␥-induced BTG1expression was obtained by transfection IFN-␥ pretreatment sensitizes microglial cells to NO-induced ap- of dominant negative mutant of STAT1. Compared with empty optosis (12), our current results suggest that BTG1 might be one of vector-transfected cells, dominant negative STAT1 transfectants the LPS/IFN-␥-induced genes that may participate in AICD of showed a marked decrease in BTG1 induction after LPS/IFN-␥ microglia, possibly by lowering the threshold for apoptosis. Al- treatment (Fig. 7B). Because IRF-1 has been previously shown to The Journal of Immunology 5807

FIGURE 4. BTG1 sensitizes microglia to NO-induced apoptosis. The effect of conditional expression (AÐC) or transient transfection (D)ofBTG1 on NO- induced apoptosis was assessed by annexin V-binding assay (A), DNA fragmentation (B and C), or cellular morphology (C and D). Forced expression of BTG1 increased the sensitivity of microglia to NO-induced apoptosis. Treatment of BV-2/tet3 cells with SNAP of 0.2 mM for 24 h in the presence of tetracycline (ϩ tet) induced a modest or insigniﬁcant apoptosis. However, enhancement of BTG1 expression in BV-2/tet3 cells by tetracycline removal (Ϫ tet) strongly increased microglial apoptosis on treatment with the same concentration of SNAP (24 h treatment), as judged by an increase in annexin V binding (A), cytoplasmic release of fragmented nucleosomes (B), the appearance of ev- Downloaded from ident DNA fragmentation (C, top), or nuclear condensation and fragmentation as indicated by arrows (C, bottom). Transient transfection of parental BV-2 cells with BTG1 cDNA also increased the sensitivity of the microglial cells to SNAP (0.2 mM, 24 h treatment; D). The percentage of apoptosis is based on the ratio of the number of apoptotic blue cells to total number of blue http://www.jimmunol.org/ cells expressing cotransfected lacZ (pCH110). Results are mean Ϯ SEM or representative of three indepen- -Statistically signiﬁcant differ ,ء .dent experiments ences from untreated control or the empty vector (pcDNA3)-transfected cells (p Ͻ 0.05). by guest on October 1, 2021

be induced by the JAK/STAT1 pathway and to play an important neither Fas-Fas ligand interaction nor TNF-␣ is important in AICD role in cellular apoptosis (12, 26, 27), we next examined the role of microglial cells. Instead, NO produced by activated microglial of IRF-1 as a downstream mediator of LPS/IFN-␥ action in mi- cells themselves is the major cytotoxic mediator (9). However, the croglia. IRF-1 protein expression was induced by IFN-␥ as early as presence of NO-independent cytotoxic mechanism has been also 2 h after the treatment in BV-2 cells, and gene induction was also reported. Our previous work indicated that inflammatory stimuli achieved by treatment with LPS alone or LPS plus IFN-␥ (Fig. 8, play a dual role in AICD of microglial cells (12) and astrocytes ␥ A and B). However, LPS/IFN- -induced BTG1 expression was not (31). They induce not only the indirect apoptotic pathway via pro- affected by IRF-1 gene deficiency, because a similar enhancement duction of NO, but also the direct apoptotic pathway through ␥ of BTG1 expression by LPS/IFN- was observed in primary mi- caspase-11 induction. Whereas caspase-11 induction and its acti- croglial cultures from both wild-type and IRF-1-deficient mice vation were required for the NO-independent apoptotic pathway, (Fig. 8C). Taken collectively, our results suggest that LPS/IFN-␥ IRF-1 and NF-␬B were involved in NO-dependent apoptosis of action on BTG1 expression and their effect on microglial apoptosis microglial cells mainly by mediating NO synthesis (via inducible might be mediated through JAK/STAT1, but not IRF-1. Although NO synthase induction). Up-regulated caspase-11 is autoactivated IRF-1 is a critical mediator of IFN-␥ action in both apoptosis and cell cycle control (28), IRF-1 does not appear to be involved in in microglia and triggers an activation cascade of downstream LPS/IFN-␥-mediated BTG1 induction or the ensuing proliferation caspases, which ultimately leads to cellular apoptosis. Meanwhile, inhibition in microglia. NO donor-induced apoptosis appears to directly activate downstream executioner caspases in microglia. Discussion Inflammatory stimuli such as LPS and IFN-␥ are more than just AICD is an autoregulatory mechanism by which the immune sys- cellular activators to microglial cells. They not only activate the tem removes unwanted activated immune cells after making ap- cells to produce inflammatory mediator such as TNF-␣ and NO but propriate use of them. Although AICD has been first identified in also induce autoregulatory apoptosis. Inflammatory stimuli appear lymphocytes (29, 30), recent works indicated that both microglial to induce or activate a specific group of genes and signaling path- cells (9) and astrocytes (11) in the CNS might be under the control ways, some of which are commonly involved in both activation of a similar regulatory mechanism. In contrast to AICD of T lym- and apoptosis. Our previous work suggested that caspase-11 may phocytes where Fas-Fas ligand interaction plays a central role, be an example of such genes (12). Induction and activation of 5808 BTG1 IN MICROGLIAL APOPTOSIS

FIGURE 6. Essential role of JAK/STAT1 signaling in LPS/IFN-␥-me-

FIGURE 5. BTG1 is not involved in the inflammatory activation of mi- Downloaded from croglia. Role of BTG1 in the inflammatory activation of microglia was diated apoptosis sensitization and BTG1 induction in microglia. A, STAT1 investigated by comparing the production of inflammatory mediators such was activated by IFN-␥ (10 U/ml) or combination of LPS (100 ng/ml) and as TNF-␣, IL-1␤, and NO from BV-2/tet3 cells cultured in the presence (ϩ IFN-␥, but not by LPS alone, as measured by Western blot detection of 701 tet) or absence (Ϫ tet) of tetracycline by specific ELISA or Griess reaction. phosphorylated STAT1 protein (at Tyr ) at the indicated time points in An increase in BTG1 expression by removal of tetracycline from growth parental BV-2 cells. B, Inhibition of STAT1 signaling by transient trans- medium did not significantly affect LPS-induced production of TNF-␣, fection of the dominant negative mutant of STAT1 cDNA (DN STAT1) ␥ IL-1␤, or NO (100 ng/ml LPS, 24 h treatment). Results are mean Ϯ SEM of suppressed LPS/IFN- -induced sensitization of microglia to apoptogenic http://www.jimmunol.org/ three independent experiments. No statistically significant difference was action of SNAP. LPS/IFN-␥ pretreatment rendered BV-2 microglial cells found between ϩ tet and Ϫ tet for any of the inflammatory mediators tested. sensitive to otherwise nontoxic dose of SNAP (0.2 mM, 24 h treatment). However, STAT1 inhibition partially abolished the apoptosis-sensitizing effect of LPS/IFN-␥. LPS/IFN-␥ pretreatment (16 h) was done in the presence of a NOS inhibitor, NMMA (0.5 mM) to eliminate confounding ef- caspase-11 are involved in the production of proinflammatory cy- fects of LPS/IFN-␥-mediated inducible NO synthase induction and result- tokines such as IL-1␤ and IL-18 by activating the processing en- ing NO production. The percentage of apoptosis was determined by zyme (caspase-1) required for the conversion of procytokines to differential counting of healthy blue cells and apoptotic blue cells coex- their mature form (32, 33). Caspase-11 induction also causes mi- pressing lacZ (pCH110). Empty vector indicates pcDNA3. Results are by guest on October 1, 2021 ء Ϯ croglial apoptosis. Therefore, inflammatory stimuli seem to acti- mean SEM of three independent experiments. , Statistically significant differences (p Ͻ 0.05). C, Treatment of BV-2 cells with JAK inhibitor vate microglial cells to produce various inflammatory mediators AG490 (10 ␮g/ml), attenuated LPS/IFN-␥-induced BTG1 protein expres- and concomitantly activate a built-in autoregulatory apoptosis sion (100 ng/ml LPS, 10 U/ml IFN-␥, 8 h treatment). The numbers indicate mechanism. The role of inflammatory stimuli in microglial AICD a fold induction of BTG1 expression normalized to ␣-tubulin as deter- was not limited to the NO production or caspase-11 induction. Our mined by densitometric analysis of Western blot results. previous findings that pretreatment of microglia with inflammatory stimuli enhanced their sensitivity to exogenous NO independently of caspase-11 suggested the presence of yet other mechanisms whereby the inflammatory stimuli affect microglial AICD process (12). microenvironment could be an ideal apoptogenic signal to remove Here, we present evidence that BTG1 constitutes a novel apo- these overactivated microglia with a high BTG1 expression. Thus, ␥ ptotic pathway during AICD of microglia. Because LPS/IFN- has overactivation of microglia may mark themselves with a high been previously shown to increase the sensitivity of microglia to BTG1 expression for the apoptotic elimination by NO. NO-induced apoptosis (12), we hypothesized that LPS/IFN-␥ may We do not know how antiproliferative activity of BTG1 medi- lower the threshold for microglial apoptosis through the expression ates or enhances apoptogenic action of NO. In many cases, block- of inducible mediators. We now report that BTG1 is one of such ade of cell cycle by anticancer drugs leads to cellular apoptosis. It mediators induced by LPS/IFN-␥. This is supported by our results is well documented that cell cycle control and apoptosis are closely that 1) LPS/IFN-␥ synergistically augmented BTG1 expression in linked intracellular events (34, 35). Apoptosis could be viewed as microglia, 2) BTG1 inhibited microglial proliferation, and 3) conditional expression of BTG1 increased the sensitivity of microglial a safeguard mechanism that allows the elimination of cells the cells to NO-induced apoptosis. Thus, the LPS/IFN-␥-mediated in- physiological cell cycle of which has been blocked. In this respect, crease in BTG1 expression may prevent microglia from entering BTG1-mediated inhibition of cellular proliferation may trigger this cell cycle, and this antiproliferative action of BTG1 may render the type of safeguard mechanism, thereby rendering microglial cells cells more sensitive to apoptotic signals. Our current results sup- extremely sensitive to incoming apoptogenic signals. BTG1 has port the model of microglial apoptosis as a self-regulatory mech- been shown to interact with transcription factors such as CAF1 anism that has been previously proposed (8, 9). Activation of mi- (36, 37) and Hoxb9 (38) that control the expression of prolifera- croglia may lead to their proliferation under certain conditions. tion-regulatory genes. BTG1 also forms a complex with PRMT1 However, overactivating signals that are strong enough to enhance that methylates various intracellular proteins, thereby modulating BTG1 expression, e.g., LPS plus IFN-␥, may retard microglial pro- its methyltransferase activity (39). BTG1 may exert its antiprolif- liferation, making them ready to be eliminated by apoptosis. NO erative activity through the interaction with these proteins. The that has already been produced by activated microglia in the same apoptosis-sensitizing effect of BTG1 demonstrated in the current The Journal of Immunology 5809

FIGURE 7. Attenuation of LPS/IFN-␥-induced microglial cell death and NO production by AG490. A, Treatment of BV-2 cells with AG490 (10 ␮g/ml) attenuated LPS/IFN-␥-induced microglial NO production (left) and cell death (right) as assessed by MTT assays and Griess reaction at 24 h. NMMA (0.5 mM) was used for comparison. Results are ,ء .mean Ϯ SEM of three independent experiments Statistically significant differences from LPS/IFN-␥ treatment alone (p Ͻ 0.05). B, Although LPS/IFN- ␥-induced BTG1 message induction was not influ- enced by NMMA (left), it was significantly reduced by the introduction of the dominant negative mutant STAT1 (DN STAT1) (right). RT-PCR was per- formed to evaluate BTG1 or ␤-actin messages after the stimulation for 4 h. Downloaded from

study may not be solely due to its antiproliferative activity. Pre- because the cell cycle-inhibitory drugs alone may cause cell death liminary results in our laboratory indicated that cell cycle-inhibi- (depending on the concentrations used) as opposed to BTG1 ex- http://www.jimmunol.org/ tory drugs did sensitize microglia to NO-induced apoptosis; how- pression, which does not affect the cell viability by itself. Further ever, their sensitizing activity was much less effective than the works are required to resolve this issue and to see whether BTG1 forced expression of BTG1. The results suggest that antiprolifera- induction can sensitize microglia to other apoptosis-inducing tive BTG1 may have an additional function, which is uniquely agents besides NO. involved in the NO-induced apoptosis. However, a firm conclusion Role of BTG1 in cell cycle control has been previously docu- on the role of BTG1 cannot be drawn with these experiments, mented (14). Studies with PHA-activated PBL and serum stimulation of growth-arrested NIH3T3 cells showed that BTG1 expression negatively correlates with the percentage of cells entering S phase (16). Involvement of BTG1 in cellular apoptosis has been by guest on October 1, 2021 also suggested by a previous report, in which the forced expression of BTG1 in NIH3T3 cells increased the frequency of apoptosis and BTG1 expression colocalized not only with cells positive for mark- ers of apoptosis but also with macrophage-rich regions in Wa- tanabe heritable hyperlipidemic rabbits (19). In addition, we have previously demonstrated a negative correlation between BTG1 expression and proliferation of mouse peritoneal macrophages as well as the RAW264.7 macrophage-like cell line (18). Based on functional similarities between macrophages and microglia, antiproliferative action of BTG1 currently observed in microglia is not greatly unexpected. Furthermore, our results on the role of BTG1 in microglial apoptosis are in agreement with the previous studies that demonstrated the apoptosis-related role of BTG1 in NIH3T3 cells (19). In this work, we focused on the effects of LPS and IFN-␥ as representative inflammatory stimuli. Other known microglial acti- FIGURE 8. No role of IRF-1 in LPS/IFN-␥-induced BTG1 induction. vators include thrombin (40), chromogranin A (10), gangliosides Expression of IRF-1 protein was induced by both IFN-␥ (10 U/ml) and (41), and amyloid ␤ peptide (42). These agents induce the inflam- LPS (100 ng/ml) in BV-2 cells as determined by Western blot analysis (A matory activation of microglia. Many of these inflammatory stim- and B). IRF-1 or ␣-tubulin protein was detected at the indicated time points uli initiate common intracellular signaling pathways such as after IFN-␥ treatment (A) or 4 h after treatment with the indicated stimuli NF-␬B activation, p38 MAPK, and/or ERK activation. In particular, (B). A similar level of BTG1 message induction by LPS/IFN-␥ was ob- gangliosides uniquely evoke activation of JAK/STAT signaling path- served in primary microglial cells from wild-type mice (IRF-1ϩ/ϩ) and Ϫ/Ϫ ways in microglia (43). Considering the similarity between ganglio- IRF-1 deficient mice (IRF-1 ; C). The absence of an IRF-1 gene did not sides and IFN-␥ in the initiation of the JAK/STAT pathway, the two affect LPS/IFN-␥ induction of BTG1 messages in microglia as assessed by stimuli may share BTG1 as a common downstream mediator to in- RT-PCR (4 h treatment with the stimuli). The absence of IRF-1 transcripts in microglia from IRF-1-deficient mice was confirmed by RT-PCR (D). duce microglial apoptosis. Moreover, because LPS is a canonical ac- Mouse ␤-actin was used as an internal control in RT-PCR. The amplified tivator of NF-␬B (44), there might be other unexpected overlaps in product was not observed in RT-PCR without reverse transcriptase (data signal transduction pathways among these various microglial activa- not shown). Results are representative of three independent experiments. tors. All these activators may potentially induce microglial apoptosis, 5810 BTG1 IN MICROGLIAL APOPTOSIS and it will be interesting to examine whether they are capable of 6. Albina, J. E., S. Cui, R. B. Mateo, and J. S. Reichner. 1993. Nitric oxide-mediated inducing BTG1 expression either alone or in various combinations in apoptosis in murine peritoneal macrophages. J. Immunol. 150:5080. 7. Adler, B., H. Adler, T. W. Jungi, and E. Peterhans. 1995. Interferon-␣ primes the process of AICD of microglia. The mechanism of LPS/IFN-␥- macrophages for lipopolysaccharide-induced apoptosis. Biochem. Biophys. Res. induced AICD of microglial cells proposed in the current studies may Commun. 215:921. well be applicable to other inflammatory stimuli. Either LPS or IFN-␥ 8. Liu, B., K. Wang, H. M. Gao, B. Mandavilli, J. Y. Wang, and J. S. Hong. 2001. Molecular consequences of activated microglia in the brain: overactivation in- alone was not sufficient for the strong induction of BTG1 expression; duces apoptosis. J. Neurochem. 77:182. a marked induction was achieved only by cotreatment. Moreover, in 9. Lee, P., J. Lee, S. Kim, H. Yagita, M. S. Lee, S. Y. Kim, H. Kim, and K. Suk. ␥ 2001. NO as an autocrine mediator in the apoptosis of activated microglial cells: contrast to IFN- , LPS failed to activate STAT1 signaling. Thus, the correlation between activation and apoptosis of microglial cells. Brain Res. JAK/STAT1 signaling pathway may be necessary, but not sufficient, 892:380. for LPS/IFN-␥-induced BTG1 expression. It is speculated that IFN-␥ 10. Kingham, P. J., M. L. Cuzner, and J. M. Pocock. 1999. Apoptotic pathways mobilized in microglia and neurones as a consequence of chromogranin A-in- uses the JAK/STAT1 signaling pathway, whereas LPS may trigger duced microglial activation. J. Neurochem. 73:538. other signaling events such as NF-␬B or mitogen-activated protein 11. Suk, K., J. Lee, J. Hur, Y. S. Kim, M. S. Lee, S. H. Cha, S. Y. Kim, and H. Kim. kinases. Cooperation between these signaling events seems to be re- 2001. Activation-induced cell death of rat astrocytes. Brain Res. 900:342. 12. Lee, J., J. Hur, P. Lee, J. Y. Kim, N. Cho, M. S. Lee, S. Y. Kim, H. Kim, and quired for the efficient BTG1 induction and the subsequent apoptosis K. Suk. 2001. Dual role of inflammatory stimuli in activation-induced cell death sensitization in microglia. We and others (26, 27, 45Ð48) have pre- of mouse microglial cells: initiation of two separate apoptotic pathways via in- viously reported that STAT1/IRF-1 pathway plays a central role in duction of interferon regulatory factor-1 and caspase-11. J. Biol. Chem. 276:32956. cellular apoptosis induced by inflammatory cytokines. IRF-1 is a 13. Rimokh, R., J. P. Rouault, K. Wahbi, M. Gadoux, M. Lafage, E. Archimbaud, member of a family of IFN-␥-inducible transcription factors. IRF-1, C. Charrin, O. Gentilhomme, D. Germain, J. Samarut, et al. 1991. A chromosome as one of the genes the transcription of which is up-regulated by 12 coding region is juxtaposed to the MYC protooncogene locus in a t(8;12)(q24;

q22) translocation in a case of B-cell chronic lymphocytic leukemia. Genes Chro- Downloaded from STAT1, is known to mediate apoptosis as well as cell cycle arrest in mosomes Cancer 3:24. IFN-␥ responses (49). Although caspase induction has been suggested 14. Matsuda, S., J. Rouault, J. Magaud, and C. Berthet. 2001. In search of a function as a possible downstream event following IRF-1 induction in IFN-␥- for the TIS21/PC3/BTG1/TOB family. FEBS Lett. 497:67. 15. Tirone, F. 2001. The gene PC3(TIS21/BTG2), prototype member of the PC3/ induced apoptosis (47), studies using IRF-1-deficient mice have BTG/TOB family: regulator in control of cell growth, differentiation, and DNA shown that p53 and p21 are associated with IRF-1 action in IFN-␥- repair? J. Cell. Physiol. 187:155. induced cell cycle arrest (50). IRF-1 was also an important player in 16. Rouault, J. P., R. Rimokh, C. Tessa, G. Paranhos, M. Ffrench, L. Duret, M. Garoccio, D. Germain, J. Samarut, and J. P. Magaud. 1992. BTG1, a member http://www.jimmunol.org/ microglial apoptosis by mediating NO production (12, 51). However, of a new family of antiproliferative genes. EMBO J. 11:1663. the absence of IRF-1 gene did not affect the inducibility of caspase-11 17. Raburn, D. J., K. G. Hamil, J. K. Tsuruta, D. A. O’Brien, and S. H. Hall. 1995. Stage-specific expression of B cell translocation gene 1 in rat testis. Endocrinol- (12) or BTG1 (Fig. 8). Although we have not tested the induction or ogy 136:5769. activation of other caspases, apoptotic or cell cycle mediators in IRF- 18. Suk, K., D. G. Sipes, and K. L. Erickson. 1997. Enhancement of B-cell translo- 1-deficient microglial cells, a role of IRF-1 in microglial apoptosis cation gene-1 expression by prostaglandin E2 in macrophages and the relationship to proliferation. Immunology 91:121. seems to be related to NO production rather than caspase induction or 19. Corjay, M. H., M. A. Kearney, D. A. Munzer, S. M. Diamond, and cell cycle control. J. K. Stoltenborg. 1998. Antiproliferative gene BTG1 is highly expressed in ap- There is now growing evidence that toxic mediators produced optotic cells in macrophage-rich areas of advanced lesions in Watanabe heritable hyperlipidemic rabbit and human. Lab. Invest. 78:847. by activated microglial cells might be involved in the pathogenesis 20. Matsuyama, T., T. Kimura, M. Kitagawa, K. Pfeffer, T. Kawakami, N. Watanabe, by guest on October 1, 2021 of various neurodegenerative diseases such as Parkinson’s disease, T. M. Kundig, R. Amakawa, K. Kishihara, A. Wakeham, et al. 1993. Targeted Alzheimer’s disease, and HIV-associated dementia (3, 4, 52). disruption of IRF-1 or IRF-2 results in abnormal type I IFN gene induction and aberrant lymphocyte development. Cell 75:83. Thus, in CNS, the production of toxic inflammatory mediators by 21. Aloisi, F., R. De Simone, S. Columba-Cabezas, and G. Levi. 1999. Opposite ␥ activated microglial cells must be strictly regulated to avoid harm- effects of interferon- and prostaglandin E2 on tumor necrosis factor and inter- ful effects. Because the apoptosis of activated microglia is believed leukin-10 production in microglia: a regulatory loop controlling microglia pro- and anti-inflammatory activities. J. Neurosci. Res. 56:571. to be one of the autoregulatory mechanisms for microglial activa- 22. Blasi, E., R. Barluzzi, V. Bocchini, R. Mazzolla, and F. Bistoni. 1990. Immor- tion, the elucidation of microglial AICD mechanism and the better talization of murine microglial cells by a v-raf/v-myc carrying retrovirus. J. Neu- understanding of associated apoptosis pathways will lead to a roimmunol. 27:229. 23. Gossen, M., and H. Bujard. 1992. Tight control of gene expression in mammalian novel way of controlling microglial activation. This will in turn cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89:5547. allow us an opportunity for the therapeutic interventions against 24. Hoglund, M., T. Siden, and D. Rohme. 1992. Different pathways for chromo- the neurodegenerative diseases, in which overactivation of micro- somal integration of transfected circular pSVneo plasmids in normal and estab- lished rodent cells. Gene 116:215. glia plays a pathogenic role. 25. Adams, D. O., and T. A. Hamilton. 1984. The cell biology of macrophage acti- In conclusion, we identified BTG1 as an important constituent in vation. Annu. Rev. Immunol. 2:283. microglial AICD process. In the activation-induced apoptosis of 26. Suk, K., S. Kim, Y. H. Kim, K. A. Kim, I. Chang, H. Yagita, M. Shong, and M. S. Lee. 2001. IFN-␥/TNF-␣ synergism as the final effector in autoimmune microglial cells, inflammatory stimuli (LPS/IFN-␥) play a multiple diabetes: a key role for STAT1/IFN regulatory factor-1 pathway in pancreatic ␤ role. They induce not only direct apoptotic mediators (NO, cell death. J. Immunol. 166:4481. 27. Suk, K., I. Chang, Y. H. Kim, S. Kim, J. Y. Kim, H. Kim, and M. S. Lee. 2001. caspase-11) but also BTG1 expression to sensitize the target cells Interferon ␥ (IFN-␥) and tumor necrosis factor ␣ synergism in ME-180 cervical to an apoptotic mediator, NO. This BTG1 action appears to be cancer cell apoptosis and necrosis. IFN-␥ inhibits cytoprotective NF-␬B through related to its antiproliferative activity. STAT1/IRF-1 pathways. J. Biol. Chem. 276:13153. 28. Tanaka, N., and T. Taniguchi. 2000. The interferon regulatory factors and onco- genesis. Semin. Cancer Biol. 10:73. 29. Donjerkovic, D., and D. W. Scott. 2000. Activation-induced cell death in B lym- References phocytes. Cell Res. 10:179. 1. Stoll, G., and S. Jander. 1999. The role of microglia and macrophages in the 30. Crispe, I. N. 1999. Death and destruction of activated T lymphocytes. Immunol. pathophysiology of the CNS. Prog. Neurobiol. 58:233. Res. 19:143. 2. Gehrmann, J., Y. Matsumoto, and G. W. Kreutzberg. 1995. Microglia: intrinsic 31. Suk, K., S. Y. Kim, and H. Kim. 2002. Essential role of caspase-11 in activation- immuneffector cell of the brain. Brain Res. Brain Res. Rev. 20:269. induced cell death of rat astrocytes. J. Neurochem. 80:230. 3. Minghetti, L., and G. Levi. 1998. Microglia as effector cells in brain damage and 32. Sanz, J. M., and F. Di Virgilio. 2000. Kinetics and mechanism of ATP-dependent repair: focus on prostanoids and nitric oxide. Prog. Neurobiol. 54:99. IL-1␤ release from microglial cells. J. Immunol. 164:4893. 4. Gonzalez-Scarano, F., and G. Baltuch. 1999. Microglia as mediators of inflam- 33. Conti, B., L. C. Park, N. Y. Calingasan, Y. Kim, H. Kim, Y. Bae, G. E. Gibson, matory and degenerative diseases. Annu. Rev. Neurosci. 22:219. and T. H. Joh. 1999. Cultures of astrocytes and microglia express interleukin 18. 5. von Knethen, A., A. Lotero, and B. Brune. 1998. Etoposide and cisplatin induced Brain Res. Mol. Brain Res. 67:46. apoptosis in activated RAW 264.7 macrophages is attenuated by cAMP-induced 34. Blagosklonny, M. V. 1999. A node between proliferation, apoptosis, and growth gene expression. Oncogene 17:387. arrest. Bioessays 21:704. The Journal of Immunology 5811

35. King, K. L., and J. A. Cidlowski. 1998. Cell cycle regulation and apoptosis. Annu. 43. Kim, O. S., E. J. Park, E. H. Joe, and I. Jou. 2002. JAK-STAT signaling mediates Rev. Physiol. 60:601. gangliosides-induced inflammatory responses in brain microglial cells. J. Biol. 36. Prevot, D., A. P. Morel, T. Voeltzel, M. C. Rostan, R. Rimokh, J. P. Magaud, and Chem. 277:40594. L. Corbo. 2001. Relationships of the antiproliferative proteins BTG1 and BTG2 44. Raetz, C. R., and C. Whitfield. 2002. Lipopolysaccharide endotoxins. Annu. Rev. with CAF1, the human homolog of a component of the yeast CCR4 transcrip- Biochem. 71:635. tional complex: involvement in estrogen receptor ␣ signaling pathway. J. Biol. 45. Tanaka, N., M. Ishihara, M. Kitagawa, H. Harada, T. Kimura, T. Matsuyama, Chem. 276:9640. M. S. Lamphier, S. Aizawa, T. W. Mak, and T. Taniguchi. 1994. Cellular com- mitment to oncogene-induced transformation or apoptosis is dependent on the 37. Rouault, J. P., D. Prevot, C. Berthet, A. M. Birot, M. Billaud, J. P. Magaud, and transcriptional factor IRF-1. Cell 77:829. L. Corbo. 1998. Interaction of BTG1 and p53-regulated BTG2 gene products 46. Tamura, T., M. Ishihara, M. S. Lamphier, N. Tanaka, I. Oishi, S. Aizawa, with mCaf1, the murine homolog of a component of the yeast CCR4 transcrip- T. Matsuyama, T. W. Mak, S. Taki, and T. Taniguchi. 1995. An IRF-1-dependent tional regulatory complex. J. Biol. Chem. 273:22563. pathway of DNA damage-induced apoptosis in mitogen-activated T lymphocytes. 38. Prevot, D., T. Voeltzel, A. M. Birot, A. P. Morel, M. C. Rostan, J. P. Magaud, and Nature 376:596. L. Corbo. 2000. The leukemia-associated protein Btg1 and the p53-regulated 47. Kano, A., T. Haruyama, T. Akaike, and Y. Watanabe. 1999. IRF-1 is an essential protein Btg2 interact with the homeoprotein Hoxb9 and enhance its transcrip- mediator in IFN-␥-induced cell cycle arrest and apoptosis in primary cultured tional activation. J. Biol. Chem. 275:147. hepatocytes. Biochem. Biophys. Res. Commun. 257:672. 39. Lin, W. J., J. D. Gary, M. C. Yang, S. Clarke, and H. R. Herschman. 1996. The 48. Suk, K., Y. Kim, I. Chang, J. Y. Kim, Y. Choi, K. Y. Lee, and M. Lee. 2001. mammalian immediate-early TIS21 protein and the leukemia-associated BTG1 IFN␣ sensitizes ME-180 human cervical cancer cells to TNF␣-induced apoptosis protein interact with a protein-arginine N-methyltransferase. J. Biol. Chem. by inhibiting cytoprotective NF-␬B activation. FEBS Lett. 495:66. 271:15034. 49. Harada, H., T. Taniguchi, and N. Tanaka. 1998. The role of interferon regulatory factors in the interferon system and cell growth control. Biochimie 80:641. 40. Ryu, J., H. Pyo, I. Jou, and E. Joe. 2000. Thrombin induces NO release from 50. Tura, B. J., K. E. Bunyan, and D. J. Harrison. 2001. The effect of IFN␥ on the cultured rat microglia via protein kinase C, mitogen-activated protein kinase, and hepatocyte: cell cycle and apoptosis. Int. J. Exp. Pathol. 82:317. NF-␬B. J. Biol. Chem. 275:29955. 51. Kamijo, R., H. Harada, T. Matsuyama, M. Bosland, J. Gerecitano, D. Shapiro, 41. Pyo, H., E. Joe, S. Jung, S. H. Lee, and I. Jou. 1999. Gangliosides activate J. Le, S. I. Koh, T. Kimura, S. J. Green, et al. 1994. Requirement for transcription cultured rat brain microglia. J. Biol. Chem. 274:34584. factor IRF-1 in NO synthase induction in macrophages. Science 263:1612. 42. Meda, L., M. A. Cassatella, G. I. Szendrei, L. Otvos, Jr., P. Baron, M. Villalba, 52. McGeer, P. L., and E. G. McGeer. 1995. The inflammatory response system of Downloaded from D. Ferrari, and F. Rossi. 1995. Activation of microglial cells by ␤-amyloid pro- brain: implications for therapy of Alzheimer and other neurodegenerative dis- tein and interferon-␥. Nature 374:647. eases. Brain Res. Brain Res. Rev. 21:195. http://www.jimmunol.org/ by guest on October 1, 2021