Toxoplasma gondii - The Journal of Immunology

The Journal of Immunology

Toxoplasma gondii Dysregulates IFN-␥-Inducible Gene Expression in Human Fibroblasts: Insights from a Genome-Wide Transcriptional Profiling1 Seon-Kyeong Kim, Ashley E. Fouts, and John C. Boothroyd2 Toxoplasma gondii is an obligate intracellular parasite that persists for the life of a mammalian host. The parasite’s ability to block the potent IFN-␥ response may be one of the key mechanisms that allow Toxoplasma to persist. Using a genome-wide microarray analysis, we show here a complete dysregulation of IFN-␥-inducible gene expression in human fibroblasts infected with Toxoplasma. Notably, 46 of the 127 IFN-␥-responsive genes were induced and 19 were suppressed in infected cells before they were exposed to IFN-␥, indicating that other stimuli produced during infection may also regulate these genes. Following IFN-␥ treatment, none of the 127 IFN-␥-responsive genes could be significantly induced in infected cells. Immunofluorescence assays showed at single-cell levels that infected cells, regardless of which Toxoplasma strain was used, could not be activated by IFN-␥ to up-regulate the expression of IFN regulatory factor 1, a transcription factor that is under the direct control of STAT1, whereas uninfected cells in the same culture expressed IFN regulatory factor 1 normally in response to IFN-␥. STAT1 trafficked to the nucleus normally and indistinguishably in all uninfected and infected cells treated with IFN-␥, indicating that the inhibitory effects of Toxoplasma infection likely occur via blocking STAT1 transcriptional activity in the nucleus. In contrast, a closely related apicomplexan, Neospora caninum, was unable to inhibit IFN-␥-induced gene expression. A differential ability to interfere with the IFN-␥ response may, in part, account for the differences in the pathogenesis seen among Toxoplasma and Neospora parasite strains. The Journal of Immunology, 2007, 178: 5154 –5165.

T

oxoplasma gondii is an obligate intracellular parasite that induces a strong IFN-␥-driven cell-mediated immune response in its mammalian hosts. This response is critical for the resolution of acute infection (1) and control of a chronic, latent infection in the CNS (2). Various cell types are activated by IFN-␥ to acquire potent toxoplasmacidal mechanisms, including inducible NO synthase and IFN-␥-induced GTPases expression (3– 6). In addition, IFN-␥ plays a role in driving the conversion of tachyzoites (the acute phase form) to bradyzoites (the chronic phase form) (7) and suppresses the reactivation to tachyzoites (8). IFN-␥ is produced at high levels in acutely infected mice (100 – 300 ng/ml serum at 7 days postinfection) (9) and remains at 1– 6 ng/ml serum even at 3 wk postinfection (10). Clearly, IFN-␥ is a potent immune effector and, yet, Toxoplasma is not cleared but persists in immunocompetent hosts. For optimal control of Toxoplasma replication in vitro, host cells should be activated with IFN-␥ before infection (3– 6); if the cells are infected first, subsequent exposure to IFN-␥ is unable to control infection and the parasite grows normally (11–13). Toxoplasma disseminates rapidly following oral infection of a mammalian host and reaches

Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305 Received for publication September 14, 2006. Accepted for publication February 7, 2007. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by National Institutes of Health Grants AI41014 and AI21423. 2 Address correspondence and reprint requests to Dr. John C. Boothroyd, Department of Microbiology and Immunology, Stanford University School of Medicine, 299 Campus Drive, Stanford, CA 94305-5124. E-mail address: [email protected]

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org

distant tissues such as the lungs, muscles, and brain within a few days (14). It is plausible that the parasite reaches these tissues before the arrival of IFN-␥-secreting NK and T cells and, given time to establish an infection and co-opt/subvert host cell processes, Toxoplasma can effectively inhibit the IFN-␥ signaling pathway and enhance its intracellular survival and persistence. Indeed, recent studies have shown that several IFN-␥-responsive genes could not be induced by IFN-␥ in Toxoplasma-infected cells (11–13). IFN-␥ exerts its effects via the transcriptional activation of numerous genes involved in antimicrobial activity, Ag processing and presentation, lymphocyte trafficking, cell growth, and apoptosis (15). IFN-␥ binds a ubiquitously expressed cell surface IFN-␥ receptor, which is associated with JAK1 and JAK2 that phosphorylate STAT1 at Tyr701 (16). Phosphorylated STAT1 forms homodimers that then translocate to the nucleus and initiate the transcription of IFN-␥ target genes. Additional phosphorylation at Ser727 is thought to be necessary for a maximal transcriptional activity (17, 18). Toxoplasma appears to use various mechanisms to interfere with the STAT1-dependent IFN-␥ signaling pathway: in heavily infected murine macrophages, a type I Toxoplasma strain (BK) was reported to induce a high-level SOCS1 mRNA expression and degradation of STAT1 protein (13). In contrast, a type II strain (NTE) inhibited IFN-␥-inducible gene expression without altering STAT1 protein stability (11, 12). As discussed in Ref. 11, STAT1 trafficking was found to be intact in infected NIH3T3 cells overexpressing STAT1-GFP fusion protein. In contrast, the same group reported in (12) that nuclear trafficking of endogenous STAT1 appeared to be partially inhibited by infection in heavily infected macrophages. These previous studies examined the expression of only several IFN-␥-responsive genes mostly by performing RT-PCR using a population of heavily infected macrophages. The extent to which Toxoplasma blocks IFN␥-inducible gene expression, the effects of infection in cell types other than macrophages (e.g., in nonphagocytic cells in which the

The Journal of Immunology parasite persists at low numbers), and the exact mechanisms of parasite interference remain unknown. In this study, we examined how Toxoplasma-infected human fibroblasts respond to IFN-␥. Fibroblasts in the smooth muscle may be important cell types for infection and persistence (19, 20). Fibroblasts are now known to play an important role in shaping the course of an immune response in tissues by regulating the switch from acute inflammation to adaptive immunity and tissue repair (21). Nonphagocytic, nonprofessional APCs, such as fibroblasts, critically depend on IFN-␥ to develop antimicrobial properties and an ability to interact with T cells effectively (15). It would be to the parasite’s advantage to infect these cells and block the expression of IFN-␥ target genes to persist. We compared whether and by what mechanisms three major clonal types (I, II, and III) of Toxoplasma (22) and its closely related apicomplexan species Neospora caninum (23, 24) inhibit the IFN-␥-inducible gene expression, because the host range, virulence, and ability to persist vary greatly among these parasites (24 –26), although IL-12 and IFN-␥ are key to controlling infection in all cases (1, 27, 28). We show in this study that Toxoplasma, but not Neospora, successfully blocks the effects of IFN-␥. Microarray analysis reveals that by some means (probably an early proinflammatory cytokine response) Toxoplasma infection causes a differential expression of ⬃50% of the 127 IFN-␥-inducible genes in human foreskin fibroblasts (HFFs)3 before infected cells are exposed to IFN-␥, but that the other half shows no such change. Addition of IFN-␥ to infected cells had no significant effect on any of the 127 IFN-␥-inducible genes, indicating a highly efficient means by which Toxoplasma defends itself against this potent immune mediator.

Materials and Methods Parasites and cell culture The following T. gondii strains representing the three major clonal types were used in this study: RH (type I) (10) and CL14 (type III) (provided by Dr. J. Saeij, Stanford University, Stanford, CA) expressing GFP under the GRA1 promoter, and Pru (type II) expressing GFP under the GRA2 promoter. The minimal GRA2 promoter region (29) was PCR-cloned from the Pru strain genomic DNA using the following primers: 5⬘-GACCAAGCT TCCCGTCGCACGGTGATACTG-3⬘ and 5⬘-CGGCATGCATTGTGAG GCGATATGTGGAGAA-3⬘. The 317-bp PCR product was digested with HindIII and NsiI and replaced the GRA1 promoter in the pGRA1-GFP/ pDHFR-HPT plasmid as described in Ref. 10. Transformation of the Pru strain and selection for hypoxanthine-xanthine-guanine phosphoribosyltransferase activity was performed as described previously (10). Toxoplasma strains and N. caninum (NC-1) were maintained by passage in HFF monolayers cultured in DMEM supplemented with 10% FCS, penicillin (100 U/ml), streptomycin (100 ␮g/ml), and L-glutamine (2 mM) (Invitrogen Life Technologies) in a humidified, 5% CO2 incubator.

IFN-␥ treatment of infected HFFs Parasites were released from intracellular vacuoles by syringe lysis as described previously (10) and washed three times in complete medium before inoculation of HFFs grown to confluency in flasks or on coverslips. Cells were infected or left uninfected for 18 h at a nominal multiplicity of infection (MOI) of 3 for microarray and Western blot experiments and at a MOI of 0.5 for immunofluorescence microscopy (IFM). For IFN-␥ treatment, the cells were given fresh medium containing recombinant human IFN-␥ at 100 U/ml (⬃2 ng/ml) (R&D Systems) and incubated further for the indicated amount of time.

Human cDNA microarray analysis Type II Toxoplasma strain, Pru, was used to infect confluent monolayers of HFFs in 175-cm2 flasks. At 18 h postinfection, culture medium was removed from all uninfected or infected flasks and fresh medium containing 3 Abbreviations used in this paper: HFFs, human foreskin fibroblasts; MOI, multiplicity of infection; IFM, immunofluorescence microscopy; SAM, statistical analysis of microarrays; IRF, IFN regulatory factor; SOCS1, suppressor of cytokine signaling 1; FDR, false discovery rate; PIAS, protein inhibitor of activated STAT.

5155 or lacking recombinant human IFN-␥ (100 U/ml) was added. After 2, 4, and 8 h of IFN-␥ treatment, total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies) and poly(A)⫹ mRNA was isolated using the Oligotex kit (Qiagen). As a reference, mRNA was isolated from commercially available total RNA pooled from 10 human cell lines (Stratagene). cDNA preparation and hybridization were essentially conducted as described previously (30). Briefly, reference and sample cDNAs were directly labeled with Cy3-dUTP and Cy5-dUTPs (Amersham), respectively, through random nonamer priming using Klenow enzyme. Following quantitation and purification, reference (250 ng) and sample (300 ng) cDNA were simultaneously hybridized to human cDNA microarrays spotted on Corning Ultragap slides (Stanford Functional Genomics Facility, Stanford University, Stanford, CA; 40,996 spots representing 23,228 unique putative genes). Duplicate cultures were analyzed for each time point. Microarrays were scanned using Axon Genepix 4000A and gridded using Genepix 5.1 (Molecular Devices). Data were entered into Stanford Microarray Database and two-dimensional spatial local estimation was used to normalize the spots. Data were filtered to remove poor quality spots (regression correlation ⬍0.6 and Cy3 channel intensity ⬍3-fold background) and retrieved as Log2 of Cy5/Cy3 normalized ratio (median). Twoclass SAM (statistical analysis of microarrays) (31) was performed using MeV 3.0 (www.tigr.org) to identify the genes showing significantly different expression levels in uninfected cells before and after IFN-␥ treatment. Biological network analyses were performed using the software available from the Ingenuity Systems (www.ingenuity.com) with the IFN-␥responsive genes showing differential expression levels in uninfected and infected cells before IFN-␥ treatment.

Cytokine Ab arrays For qualitative determination of the kinds of cytokines and chemokines secreted by infected HFFs, cytokine Ab arrays were performed with culture supernatant. Confluent HFF monolayers were left uninfected or infected with Pru strain parasites at a nominal MOI of 3 as described in human cDNA microarray analysis. Culture supernatant (containing 10% FCS) was harvested after 18 h and used in human cytokine Ab arrays according to the manufacturer’s instructions (RayBiotech). Briefly, supernatant was incubated with an array membrane on which 42 different human cytokine Abs were bound. The membrane was then incubated with a mixture of biotinylated anti-cytokine Abs followed by HRP-conjugated streptavidin. Detection was by chemiluminescence reagents provided by the manufacturer.

Immunofluorescence microscopy HFFs grown on coverslips were infected with indicated parasite strains for 18 h at a MOI of 0.5 and treated with IFN-␥ (100 U/ml) for the indicated amount of time. Cells were fixed in 3.7% formaldehyde (20 min at room temperature), blocked in PBS containing 3% FCS (30 min at room temperature), and permeabilized in 0.2% Triton X-100 (1 h at room temperature or overnight at 4°C). Cells were incubated with the following mouse mAbs in PBS containing 3% FCS (1 h at room temperature): anti-IFN regulatory factor (IRF)-1 (BD Biosciences; clone 20); anti-total STAT1 (BD Biosciences; clone 42; recognizes a C-terminal epitope shared by STAT1␣ and STAT1␤); anti-pTyr701 STAT1 (BD Biosciences; clone 4a); ¨ 39). Detection Ab and anti-HLA-DR/DP/DQ (BD Biosciences; clone TU was goat anti-mouse IgG coupled with Alexa Fluor 594 (Molecular Probe). Coverslips were mounted on a glass slide for fluorescence microscopy as described previously (10).

Western blots Western blots were performed as previously described (10) with total cellular proteins prepared from HFFs grown in 25-cm2 flasks (⬃106 cells per flask) infected or not with the indicated parasite strains at a MOI of 3. At 18 h postinfection, ⬎70% of the cells were infected with 1–3 vacuoles containing 2– 8 parasites. Cells were washed three times in ice-cold PBS and scraped into ice-cold lysis buffer (400 ␮l per flask). Lysates were boiled in the presence of 0.1 M DTT, and 20 ␮l was loaded per lane in a 12% denaturing SDS-PAGE (8% gel for JAK1 and JAK2 immunoblots). Proteins were transferred to nitrocellulose membrane and probed with the following Abs: mouse monoclonal IgG against total STAT1; pTyr701STAT1 and IRF1 that were also used in IFM; rabbit polyclonal antipSer727-STAT1, anti-pTyr1022/1023-JAK1, and anti-pTyr1007/1008-JAK2 IgG (Cell Signaling Technology); rabbit polyclonal anti-suppressor of cytokine signaling (SOCS)1 IgG (Santa Cruz Biotechnology; clone H-93); and mouse monoclonal anti-GAPDH IgG (Calbiochem; clone 6C5). Membranes were incubated overnight at 4°C with primary Abs diluted in 2% nonfat dry milk in Tris-buffered saline (140 mM NaCl, 2.7 mM KCl, 25

5156

INHIBITION OF IFN-␥-INDUCIBLE GENE EXPRESSION BY Toxoplasma mM Tris base (pH 7.4)). HRP-conjugated anti-mouse and anti-rabbit IgG were used as secondary reagents (Bio-Rad). Detection was by SuperSignal chemiluminescence reagents (Pierce Biotechnology). The manufacturer of anti-pTyr1022/1023-JAK1 and anti-pTyr1007/1008-JAK2 IgG noted a possible mutual cross-reactivity.

Results A genome-wide microarray analysis reveals a complete dysregulation of IFN-␥-inducible gene expression in Toxoplasma-infected human fibroblasts To understand the extent of Toxoplasma inhibition of IFN-␥-inducible gene expression in human fibroblasts, we performed mRNA expression profiling using human cDNA microarrays having 40,996 spots representing 23,228 unique putative genes. Because the effects of IFN-␥ are mediated largely by the transcriptional activation of target genes, the microarray experiments were performed with a population of HFFs in which ⬎90% of the cells were infected to detect parasite inhibition of IFN-␥-inducible gene expression. At 18 h postinfection with a type II strain (Pru) at a MOI of 3, infected cells had 1–3 vacuoles containing 2– 8 parasites per vacuole. At this time, cells were treated with IFN-␥ (100 U/ml) for 0, 2, 4, and 8 h (named P0, P2, P4, and P8, respectively). Uninfected HFFs treated with IFN-␥ for equal amounts of time were used as controls (named N0, N2, N4, and N8, respectively). Duplicate cultures were analyzed for each condition. Among the 17,004 good quality spots represented in at least 9 of the total 16 arrays, 150 spots representing 127 unique putative genes were identified as being significantly up-regulated in uninfected cells following IFN-␥ treatment (2-class SAM of N0 vs N2, N4, and N8 at 10% false discovery rate (FDR)) (Fig. 1 and Supplemental Table I) (the on-line version of this article contains supplemental material). No genes were identified as being significantly suppressed by IFN-␥ in uninfected cells. Remarkably, none of the 127 IFN-␥-responsive genes were found to be significantly induced by IFN-␥ in infected HFFs by 2-class SAM of P0 vs P2, P0 vs P4, or P0 vs P8 at 10% FDR (Fig. 1). Notably, 65 of 127 IFN-␥-responsive genes showed differential expression in Pru-infected (P0) vs uninfected cells (N0) before IFN-␥ treatment (Table I) (see Supplemental Table I for infectioninduced fold change values for all IFN-␥-responsive genes): with the P0:N0 ratio of 2 as a cutoff value for a significant difference in expression levels, 46 genes showed increased expression in infected cells (“P0 ⬎ N0” group) and 19 genes showed decreased expression in infected cells (P0 ⬍ N0 group), whereas the remaining 64 genes were expressed at similar levels (P0 ⫽ N0 group). Although type I IFNs and IFN-␥ share numerous common target genes (32), type I IFNs is an unlikely cause for the P0 ⬎ N0 and P0 ⬍ N0 groups because Toxoplasma-infected HFFs produce little, if any, IFN-␣ and -␤ (M. Lodoen, unpublished data). Additionally, infected HFFs may be in an unresponsive state to IFN-␣ and -␤, because the expression of IRF9 and STAT2 that form a

FIGURE 1. IFN-␥-inducible gene expression in human fibroblasts uninfected or infected with Toxoplasma. Human cDNA microarrays were performed with HFFs uninfected (N) or infected with type II Pru strain for 18 h (P) and then treated with IFN-␥ (100 U/ml) for the given amount of time. For example, N2 indicates an uninfected culture 2 h after IFN-␥ treatment. Data from duplicate arrays per time point were retrieved as log2(sample/reference) and are shown as a color scheme. Shown are data for all 150 spots (representing 127 unique putative genes) identified as being significantly induced by IFN-␥ in uninfected HFFs (2-class SAM of N0 vs N2, N4, and N8 at 10% FDR). Gene symbols and GenBank identifications are shown. Gray rectangles indicate no data. See Supplemental Table I for gene names and fold change values.


5157

Table I. Basal expression levels of IFN-␥-inducible genes in uninfected and Toxoplasma-infected HFFs before IFN-␥ treatmenta Gene Symbol

GenBank Identification

Gene Symbol


Gene Symbol


C18orf54 C1GALT1 C21orf91 C6orf192 C8orf1 CBR3 CCL2 CCL2 CCNB2 DDX46 DIP2C FAS

R91258 AA487750 AA281932 AA991624 H57105 AA281744 N73031 AA400378 R67222 H25041 AI352345 T77816 AA425102 AI932735 H89698 AA873341 AA293570

P0 ⬎ N0 group (46 genes) FLJ13273 N32085 FLJ34922 AA465168 FLJ34922 H80525 FLJ39370 N38839 FST AA701860 HPS3 AA918982 ICAM1 N92072 IRF7 AA477347 KLF6 AA156946 LOC25391 AA400157 LOC90355 R52538 MSX1 R33154 MT1G AI745626 MYCBP AA424831 NFKBIZ H70961 NT5E R60343 NUDCD1 N32587

PBEF1 PHLDA1 PIK3AP1 PMAIP1 PRKAA1 RIPK2 SCML1 SLC15A2 SLCO1B1 SOD2 SOD2 STAMBP1 VAV2 VRK2 ZNF9

AA489629 AA258396 R62339 AA458838 R33152 AA913804 AI537061 AA425352 H62893 T60269 W78148 AI017607 AA682337 AA282291 AA625995

GBP2 GBP2 IFIT1 IFIT1 IFIT1 IFIT2 IFIT2

H11453 AA971543 W72748 W01896 AI953299 AA074989 AA489640 AA143609 N63988

P0 ⬍ N0 group (19 genes) IFITM1 AA419251 IGF1 AA456321 ISGF3G AA291389 MX1 AA405416 MX2 AA286908 OAS1 AA146772 OAS2 R72243 RARRES3 W47350 RNF36 AA133281

SAMD9L SAMD9L SAMD9L SOCS1 TRAFD1 TRIB2 UBE2L6 UBE2L6

AA490264 AA988857 AA996042 AA280136 N21170 AA053865 AW071596 AA292031

AFF3 APOL6 ATXN7 BCL6 BCL6 C12orf11 CCDC75 CFHL3 CMAH CSF1 CSTF3 DMN EPSTI1 ETV1 FLJ20160 G1P2 G1P3 G1P3

T55558 AA149783 AA433993 T72621 AA402719 AI275489 AA973917 AI004671 R68682 R63241 R99749 AA521434 AA479106 AA781508 T74567 N29639 AA878257 AA465143 AA877815 AI286247 T60063 AA259115 AA120862 AA432030 AA075725

P0 ⫽ N0 group (64 genes) GBP1 AA486849 GBP3 R78509 GBP4 AI268082 HERC6 AA487462 HLA-E R94660 HSPB8 H57493 HSPB8 AA010110 IFI16 AA287732 IFI35 AA827287 IFIH1 AA911194 INDO AA478279 ING1 N47308 IRF2 AA416883 IRF7 AA877255 JUN AA293362 JUN W96134 KCTD3 AA448160 KIAA1033 R53810 KLF4 AW008766 LAP3 R69306 MX1 AA456886 NCOA3 AI302669 PAPD5 AA699802 PAPD5 AA029273 PARP14 T64956

PELI1 PLSCR1 PLSCR1 PSMB8 RALA SAMHD1 SFRS1 SLMAP SP110 STAT1 TAP1 TAP1 TMPRSS13 TRIM25 TRIM25 TRIM25 USP42 WARS WARS XPR1 XRN1 ZKSCAN1 ZNF313 ZNF313 ZNFX1

W86504 AA058514 N25945 AI983836 H97948 AA421603 AA455164 N49107 AA504832 AA486367 AA487429 AI346384 AA290867 N73575 AA464251 AA281936 R71889 AA857888 AA664040 AA453474 AA504116 W84769 R38967 AA504825 AA099652

a Expression levels were compared in HFFs uninfected (N0) and infected with Pru strain for 18 h (P0) that were not treated with IFN-␥. Two-fold increase or decrease in expression levels was used as cutoff value to classify genes into the P0 ⬎ N0, P0 ⬍ N0, or P0 ⫽ N0 group. Data for 150 spots representing 127 unique genes are shown. Two spots for MX1 and IRF7 are represented in two different groups, causing the sum of all genes in three groups to be 129.

trimeric complex with STAT1 in response to IFN-␣ and -␤ (16) appear to be suppressed in infected cells (P0:N0 ratio ⫽ 0.24 and 0.12, respectively; see Supplemental Table II for information on all 3,245 genes differentially expressed in N0 and P0). Although IFN-␥ signaling is mainly transduced via the JAK/ STAT1 pathway, other cytokines, such as IL-1␤ (33, 34) and TNF-␣ (35), are also known to activate some IFN-responsive genes (16). A biological network analysis of the IFN-␥-inducible genes in the P0 ⬎ N0 group identified one major network that showed an association of the genes in our data set with the TNF-␣

and IL-1␤ pathways that lead to the activation of downstream transcription factor NF-␬B (Fig. 2A) (36, 37). Indeed, many of the IFN-␥-inducible genes in the P0 ⬎ N0 group have been shown to be NF-␬B target genes (e.g., CCL2, FAS, FST, ICAM1, IRF7, MSX1, NFKBIZ, and SOD2) (http://people.bu.edu/gilmore/nf-kb/ target/index.html). In addition, IL-1␤ is one of the most highly induced genes by infection (140-fold increase) and likely expressed as a functional protein, because its known target gene PTGS2/COX2 is also up-regulated 174-fold during infection (also see Fig. 3 below). TNF-␣ was induced 1.7-fold in infection and its

5158

INHIBITION OF IFN-␥-INDUCIBLE GENE EXPRESSION BY Toxoplasma

FIGURE 2. Biological networks relevant to the IFN-␥-inducible genes differentially expressed in Toxoplasma-infected cells before IFN-␥ exposure. Three groups of IFN-␥-responsive genes (Table I) were uploaded onto the Ingenuity Pathway Analysis (www.ingenuity.com) to identify molecular networks in which these genes are known to function. The analyses show direct (e.g., binding, modification; indicated as thin lines) and indirect (e.g., regulation of gene expression; indicated as dotted lines) relationships between genes that are known in the literature. The analyses found only one major network for each of the following three groups of IFN-␥-responsive genes. The software chose a certain number of IFN-␥-inducible genes in each of the three groups as focus genes to build molecular networks (indicated by underlining and bolding); genes shown in plain face are interacting genes that the software identified but that did not emerge as IFN-␥-inducible genes in our microarray analysis. Gene names are positioned in the figures based on their known subcellular localization. Such information is not available for the four genes shown in the “Unknown” box in B and C. A, Analysis for the IFN-␥-inducible genes in the P0 ⬎ N0 group (i.e., expression levels increased ⱖ2-fold during infection). The software used 13 of 46 genes from this group as focus genes in creating networks. B, Analysis for the IFN-␥-inducible genes in the P0 ⬍ N0 group (i.e., expression levels decreased ⱖ2-fold during infection). Twelve of 19 genes from this group were used as focus genes. C, Analysis of the IFN-␥-inducible genes in the P0 ⫽ N0 group (i.e., infection-induced changes in expression levels were ⬍2-fold). Eighteen of 64 genes from this group were used as focus genes.

target genes, TNFAIP3 and TNFAIP6 (TNF-␣-induced proteins), were up-regulated up to 30-fold. FAS, encoding a member of the TNF receptor superfamily known to activate NF-␬B, is induced 3.66-fold. Both TNF-␣ and IL-1␤ are known to be potent inducers of NF-␬B-dependent CCL2 production (38), and CCL2 was found to be induced 25-fold during infection in our microarray experiments. The expression of NF-␬B1 itself was found to be induced 5-fold during infection. Taken together, the above data suggest that early proinflammatory cytokines produced by type II Toxoplasmainfected HFFs may also regulate a large number of IFN-␥-responsive genes via the recruitment of NF-␬B and independently of STAT1, and that these genes cannot be further induced upon ex-

FIGURE 3. Human cytokine arrays with Toxoplasma-infected HFF culture supernatant. Culture supernatant from uninfected HFFs (A) and HFFs infected with Pru parasites for 18 h at MOI of 3 (B) were analyzed in human cytokine arrays as described in Materials and Methods. Both uninfected and infected cells were cultured in medium containing 10% FCS. The table identifies the position of a given cytokine analyzed in the arrays. Cytokines abundantly produced by infected cells are indicated in bold. Asterisks indicate the cytokines also produced by uninfected cells at high levels.

posure to IFN-␥. This result is consistent with a recent report that NF-␬B p50/p65 dimers were basally bound to the promoters of some IFN-responsive genes and that the stimulation of these genes by IFN was higher in NF-␬B-deficient mouse fibroblasts than in wild-type cells (39). Nineteen IFN-␥-responsive genes in the P0 ⬍ N0 group (Table I) are also associated with the TNF-␣ pathway (Fig. 2B). Among these genes, IFIT1, IFIT2, IFITM1, MX1, MX2, and OAS1 have been shown to be up-regulated in reovirus-infected HeLa cells in a NF-␬B-dependent manner (40). Thus, it seems that the IFN-␥responsive genes in the P0 ⬍ N0 group may also be regulated by NF-␬B in Toxoplasma-infected HFFs, although mechanisms by


5159

FIGURE 4. IFN-␥-inducible IRF1 expression in Toxoplasma- and Neospora-infected HFFs. A, Top panels are IFM images showing IRF1 expression (antiIRF1 in red) in uninfected cells after IFN-␥ treatment (100 U/ml) for the indicated amount of time. Same exposure time was used for all IRF1 images. Corresponding Hoechst nuclear staining is shown in middle panels. Bottom panels for 0- and 0.5-h time points are longer exposure images for the same cells shown at top and demonstrate the presence of IRF1 in the host cell nucleus. B, IFN-␥-induced IRF1 expression in HFFs infected with GFP-expressing Toxoplasma strains RH, Pru, and CL14 representing types I, II, and III, respectively. HFFs were infected for 18 h and then treated with IFN-␥ for 2 or 6 h. Top panels in each time point show parasite GFP (green) and anti-IRF1 (red); lower panels in each vertical pair are images after merging Hoechst nuclear staining (blue). C, IFN-␥-induced IRF-1 expression (red) in HFFs infected with N. caninum for 18 h. The NC-1 strain used lacks GFP expression and infected cells can be identified in the phase images. D, Western blots showing basal expression levels of STAT1 (91 kDa), IRF1 (48 kDa), and SOCS1 (24 kDa), in HFFs uninfected or infected with indicated parasite strains for 18 h at a MOI of 3. Approximately 70 – 80% of the cells were infected with one to three vacuoles per cell containing two to eight parasites per vacuole at the time of lysate preparation. GAPDH (36 kDa) was used as a loading control. E, Western blots showing SOCS1 expression levels in HFFs uninfected or infected with RH or Pru strains as in D following IFN-␥ stimulation for 2 and 4 h. Data shown are representative of three experiments performed.

which these genes are suppressed in Toxoplasma infection are unclear. It is possible that Toxoplasma is able to directly suppress a specific subset of IFN-␥-responsive genes. The IFN-␥-responsive genes in the P0 ⫽ N0 group (Table I) were largely associated with the IFN-␥/STAT1 signaling pathway (Fig. 2C). This group of genes may be regulated mainly by IFN-␥ in a STAT1-dependent manner, and not by other cytokines such as IL-1␤ and TNF-␣. Included in this group are STAT1 itself and the genes that function in the MHC Ag processing and presentation pathway (i.e., PSMP8, LAP3, TAP1, and HLA-E). We then performed a human cytokine array analysis to examine the kind of cytokines and chemokines produced by infected HFFs that may regulate the expression of a subset of IFN-␥-responsive genes in the absence of IFN-␥. Interestingly, we found that uninfected HFFs secreted several cytokines, chemokines, and growth factors into the culture medium, including high levels of GRO (⫽ CXC), MCP-1 (⫽ CCL2), IL-6, and IL-8 (Fig. 3A). This finding was not simply due to the presence of 10% FCS in the culture medium, because uninfected cells cultured for 24 h without FCS still produced the same cytokines but at somewhat reduced amounts (data not shown). Infection with Pru strain (type II)

increased the repertoire of secreted factors, and infected cells produced IL-1␤, GRO-␣ (⫽ CXCL1), MCP-2 (⫽ CCL8), MCP-3 (⫽ CCL7), and RANTES (⫽ CCL5) in addition to MCP-1, IL-6, and IL-8. These cytokines and chemokines produced by infected HFFs are known to be target genes of NF-␬B (http://people.bu.edu/gilmore/nf-kb/target/index.html). Thus, data indicate the activation of NF-␬B in infected HFFs, which may regulate the expression of a subset of IFN-␥-responsive genes. In summary, our microarray experiments demonstrated that none of the IFN-␥-responsive genes could be significantly up-regulated in Toxoplasma-infected HFFs upon IFN-␥ treatment, indicating an efficient means by which the parasite blocks the effects of IFN-␥. Moreover, we have shown that numerous IFN-␥-responsive genes can also be regulated by other stimuli produced during infection. All three clonal types of Toxoplasma, but not N. caninum, block the IFN-␥-inducible expression of the transcription factor IRF1 As mentioned above, various cytokines and chemokines produced at high levels in Toxoplasma-infected cultures may modulate the

5160


expression of numerous IFN-␥-inducible genes. Such immune mediators may have played a role in the inability to detect IFN-␥inducible gene expression in macrophages heavily infected with Toxoplasma (11–13), given that macrophages are potent producers of proinflammatory cytokines and chemokines, and the milieu produced by infected macrophages may be far more complex than that produced by HFFs. To minimize the effects of other cytokines produced during infection and examine specific interactions between Toxoplasma and the STAT1-dependent IFN-␥ signaling pathway, we performed IFM using HFFs infected at a low MOI of 0.5 with parasites that had been thoroughly washed. STAT1 transcription factor is the main signal transducer of the IFN-␥ response and is required for resistance to Toxoplasma (41). To see whether Toxoplasma specifically interferes with STAT1 in HFFs, we examined by IFM the IFN-␥-inducible expression of IRF1, a transcription factor that is under the direct control of STAT1 and cannot be induced by IFN-␥ in STAT1⫺/⫺ cells (42). IRF1 protein turns over rapidly (half life ⬃30 min) and is constitutively expressed at low levels to drive the basal expression of MHC class I and other genes involved in Ag presentation, antimicrobial activity, and apoptosis (43). IRF1⫺/⫺ mice are extremely susceptible to Toxoplasma infection (44), consistent with a crucial role for this downstream mediator of the IFN-␥ response. HFFs constitutively expressed low levels of IRF1 that were predominantly localized in the nucleus (Fig. 4A; 0 h). As expected, IFN-␥ treatment rapidly induced IRF1 expression in HFFs and resulted in a strong nuclear staining with anti-IRF1 after 2 h of IFN-␥ treatment (Fig. 4A). Uninfected cells within Toxoplasmainfected cultures also showed normal levels of IRF1 induction after stimulation with IFN-␥ (Fig. 4B). However, all cells infected with Toxoplasma, regardless of which strain was used, failed to up-regulate IRF1 expression within 2 h of IFN-␥ treatment (Fig. 4B). IFN-␥ stimulation for 6 h (Fig. 4B) and 24 h (data not shown) did not result in IRF1 induction in Toxoplasma-infected cells. Even increasing the IFN-␥ concentration up to 500 U/ml did not induce IRF1 in Toxoplasma-infected cells (data not shown). Often, we observed a lack of IFN-␥-induced IRF-1 expression in cells containing a vacuole of only one or two parasites (Fig. 4B; Pru at 2 h and RH at 6 h), suggesting that impairment of the STAT1dependent IFN-␥ signaling pathway is not dependent on parasite replication. In contrast, IFN-␥ induced normal levels of IRF1 expression in cells infected with N. caninum (Fig. 4C). Previously, it was reported that infection with type I Toxoplasma caused STAT1 degradation and induced high levels of SOCS1 mRNA in murine macrophages (MOI of 10 for 4 h) (13). In our microarray experiments, STAT1 and IRF1 mRNA levels were similar in uninfected and infected HFFs (MOI of 3 for 18 h), and infection did not change basal STAT1 and IRF1 protein levels (Fig. 4D). Thus, under the conditions used here, the effects of Toxoplasma infection do not seem to be through degradation of STAT1 and IRF1 mRNA or their protein products. Also, in our microarray experiments, mRNA levels of SOCS1 and PIAS1, negative regulators of STAT1-dependent IFN-␥ signaling pathway (45, 46), were not induced by infection. Consistent with this result, neither RH (type I) nor Pru (type II) altered SOCS1 protein levels in infected cells (Fig. 4D). In uninfected HFFs, SOCS1 protein levels increased ⬃1.5-fold after 2 and 4 h of IFN-␥ treatment (Fig. 4E), consistent with a previous report that SOCS1 mRNA levels significantly increase in hepatic cells during 3– 6 h of IFN-␥ treatment (47). However, IFN-␥-induced increase in SOCS1 protein levels was not observed with RH- or Pru-infected HFFs (Fig. 4E), as expected from our microarray data showing that none of the IFN-␥-responsive genes could be induced in infected cells.

Thus, parasite induction of inhibitory regulators such as SOCS1 and protein inhibitor of activated STAT (PIAS) 1 is unlikely to be responsible for the lack of IRF1 induction in Toxoplasma-infected HFFs. It is more likely that the block in IRF1 expression is due to a defect in STAT1 activation and/or its transcriptional activity in the nucleus. IFN-␥-induced STAT1 nuclear translocation is normal in Toxoplasma-infected cells but does not result in autoinduction of STAT1 or MHC class II expression IFN-␥ binding to its cell surface receptor triggers the phosphorylation of cytoplasmic STAT1 at Tyr701, a prerequisite for STAT1 dimerization and nuclear translocation (48). We examined whether STAT1 activation and nuclear translocation is intact in Toxoplasma-infected HFFs in response to IFN-␥. In uninfected HFFs not treated with IFN-␥, IFM for total STAT1 (i.e., irrespective of phosphorylation state) showed low levels of STAT1 broadly distributed in the cytoplasm and nucleus of all cells (Fig. 5A; 0 h). This finding is consistent with the fact that STAT1 is found in the nuclei of unstimulated cells (49, 50) and functions as a constitutive transcription factor for some genes without needing ligand-mediated tyrosine phosphorylation and homodimer formation (51). At 0.5 h of IFN-␥ treatment, cytoplasmic STAT1 intensity decreased and nuclear STAT1 intensity increased sharply in all HFFs (Fig. 5A), indicating mobilization of cytoplasmic STAT1 to the nucleus. At 2 h of IFN-␥ treatment, some STAT1 reappeared in the cytoplasm, likely indicating redistribution of nuclear STAT1. Cytoplasmic STAT1 staining increased noticeably at 6 h and was very bright at 12 h of IFN-␥ treatment (Fig. 5A), indicating new STAT1 protein synthesis. In Toxoplasma-infected cultures, all cells, infected or not and regardless of infecting strains, showed total STAT1 expression levels and localization patterns similar to that in uninfected cultures at 0, 0.5, and 2 h of IFN-␥ treatment (Fig. 5B for Pru) (data not shown for RH and CL14). Thus, early events of STAT1 activation in the cytoplasm and nuclear trafficking proceeded normally in infected cells. At 6 h of IFN-␥ treatment, uninfected cells within the infected culture began to accumulate STAT1 in the cytoplasm, but Pru-infected cells tend to lack such increase in cytoplasmic STAT1 despite the persistent presence of STAT1 in the nucleus (Fig. 5B). At 12 h of IFN-␥ treatment, regardless of which Toxoplasma strain was used, 20 – 40% of infected cells lacked cytoplasmic STAT1 staining despite positive nuclear staining (Fig. 5B, 12 h (a)), and the remaining 60 – 80% of infected cells lacked both cytoplasmic and nuclear STAT1 staining (Fig. 5B, 12 h (b)). The STAT1 levels at 6 and 12 h of IFN-␥ treatment may be the balance of normal turnover of STAT1 molecules that were already present in the cell before IFN-␥ treatment and the lack of new protein synthesis due to parasite inhibition of IFN-␥ signaling. Although the half-life of STAT1 is known to be 16 –24 h in unstimulated cells (52, 53), it may change depending on culture conditions. The effects of Toxoplasma infection and IFN-␥ stimulation on the STAT1 half-life in HFFs are not known. Western blot results were consistent with the IFM data: total STAT1 protein levels were similar in uninfected and Pru-infected cultures (MOI of 3 for 18 h) (Fig. 5C). Stimulation with IFN-␥ for 24 h substantially increased STAT1 protein levels in uninfected cultures. Infected cultures also showed an increase in STAT1 levels due to the presence of uninfected cells in the cultures but this increase was much less than that in uninfected cultures, indicating an inhibition in STAT1 expression due to infection. Similar results were obtained with RH-infected cultures (data not shown).


5161

FIGURE 5. STAT1 nuclear trafficking and autoinduction in Toxoplasma- and Neospora-infected HFFs following exposure to IFN-␥. A and B, IFM images of anti-total STAT1 staining using a mAb that recognizes STAT1 irrespective of its phosphorylation state. Uninfected HFFs (A) and HFFs infected with a type II Toxoplasma Pru for 18 h (MOI of 0.5) (B) are compared. Times indicate the duration of IFN-␥ treatment. Phase images in B allow for the delineation of infected cells. White triangles indicate parasite vacuoles. C, Western blots with uninfected HFFs or HFFs infected with Pru strain for 18 h (MOI of 3) and then treated with IFN-␥ for 24 h or left untreated. D, IFM for total STAT1 in HFFs infected with N. caninum for 18 h (MOI of 0.5). Triangles indicate parasite vacuoles.

Thus, as with IRF1, all three types of Toxoplasma had inhibitory effects on the IFN-␥-inducible STAT1 expression, and the inhibition was entirely confined to infected cells. In contrast, cells infected with N. caninum up-regulated STAT1 expression normally when stimulated with IFN-␥: at 12 h of IFN-␥ treatment, infected cells showed equally high levels of STAT1 expression both in the cytoplasm and nucleus to those seen in uninfected cells (Fig. 5D). One of the critical effects of IFN-␥ on nonprofessional APCs is the up-regulation of MHC class II and costimulatory molecules to enable the cells to interact with CD4⫹ T cells. Nonprofessional APCs, such as HFFs, do not normally express MHC class II molecules (Fig. 6A). After 24 h of IFN-␥ treatment, ⬃50% of the HFFs in the uninfected culture could be induced to express MHC class II molecules (Fig. 6B). All cells in Pru-infected cultures lacked MHC class II expression before stimulation with IFN-␥ (Fig. 6C). When exposed to IFN-␥, ⬃50% of the uninfected cells in the infected culture expressed MHC class II but none of the infected cells did (Fig. 6D). Cells infected with Toxoplasma types I and III strains were also unable to express MHC class II upon exposure to IFN-␥, whereas N. caninum-infected cells did (data not shown).

IL-1, IL-6, MCP-3, RANTES, and other cytokines by HFFs that activate the PI3K, protein kinase C, and MAPK pathways (16, 54). In Western blots, the extent of IFN-␥-induced phosphorylation of STAT1 at Tyr701 appeared to be reduced by infection. Before IFN-␥ treatment, infected cultures showed low levels of pTyr701STAT1 that were undetectable in uninfected cultures (Fig. 7A). IFN-␥ caused a substantial increase in pTyr701-STAT1 in both

Toxoplasma modulates the Tyr701 phosphorylation state of STAT1 in the nucleus of some infected host cells Immunofluorescence assays with anti-total STAT1 Ab suggested that early events of STAT1 activation in the cytoplasm and subsequent nuclear trafficking were intact in Toxoplasma-infected cells but did not lead to the downstream gene expression. Indeed, Western blot analysis showed that the extent and kinetics of tyrosine phosphorylation of JAK1/2 in Pru-infected cultures was similar to that in uninfected cultures following IFN-␥ treatment (Fig. 7A). IFN-␥-induced phosphorylation of STAT1 at Ser727 also appeared unchanged by infection (Fig. 7A). Under the culture conditions used, both uninfected and Pru-infected cultures showed low but detectable levels of pSer727-STAT1 before IFN-␥ treatment (Fig. 7A; 0 min). This effect may be caused by the production of

FIGURE 6. Lack of MHC class II expression by Toxoplasma-infected HFFs after IFN-␥ treatment. HFF monolayers uninfected (A and B) or infected with Pru (type II) strain for 18 h (MOI of 0.5) (C and D) were treated with IFN-␥ (100 U/ml) for 24 h (B and D). Parasite GFP is shown in green, staining with mouse monoclonal anti-HLA-DR/DP/DQ Ab in red, and Hoechst nuclear staining in blue.

5162


FIGURE 7. STAT1 phosphorylation at the Tyr701 residue in Toxoplasma- and Neospora-infected cells. A, Western blots performed with HFFs uninfected or infected with Pru strain for 18 h at a MOI of 3 before IFN-␥ treatment for the indicated amount of time. Abs used are rabbit polyclonal Abs to pTyr1007/1008-JAK2 (may cross-react with pTyr1022/1023-JAK1) and pSer727STAT1, and mouse mAb to pTyr701-STAT1. B–F, HFFs were uninfected or infected with indicated parasite strains for 18 h at a MOI of 0.5. Cells were subsequently treated, or not, with IFN-␥ (100 U/ml) for 45 min before IFM. For Toxoplasma-infected cultures, parasite GFP (green) and anti-pTyr701-STAT1 (red) are shown in top panels at each time point; Hoechst nuclear staining (blue) is merged onto the corresponding bottom panels. For N. caninum that lacks GFP, anti-pTyr701STAT1 (red) are shown in top panels; infected cells can be identified in corresponding phase images shown in bottom panels.

infected and uninfected cultures, but the increase was less for infected cultures, indicating that infection may interfere with the tyrosine phosphorylation state of STAT1. We then performed IFM to examine STAT1 phosphorylation at Tyr701 at single-cell levels. Before IFN-␥ exposure, IFM with antipTyr701-STAT1 revealed a predominantly cytoplasmic or perinuclear staining in uninfected cells (Fig. 7B; 0 min). Toxoplasmainfected cells, regardless of which strain was used, showed similar cytoplasmic or perinuclear localization of pTyr701-STAT1 before IFN-␥ exposure (Fig. 7, C–E; 0 min). What causes this basal level of STAT1 phosphorylation at Tyr701 and why this does not induce nuclear translocation of STAT1 are not clear. Nonetheless, data indicate that Toxoplasma does not appear to cause dephosphorylation of STAT1 at Tyr701 in the cytoplasm. Forty-five minutes after IFN-␥ treatment, anti-pTyr701-STAT1 staining was strongly nuclear in uninfected cultures (Fig. 7B) and in uninfected cells within the Toxoplasma-infected cultures (Fig. 7, C–E). In contrast, ⬃60% of type I RH-infected cells lacked such nuclear staining (Fig. 7C). With type II Pru and type III CL14, the effect was less dramatic with only 20 –30% of infected cells lacking the nuclear staining and the remainder showing nuclear staining similar to that seen in uninfected cells (Fig. 7, D and E). Because Tyr701 phosphorylation is required for STAT1 nuclear translocation, and because total STAT1 staining is nuclear in all infected cells soon after exposure to IFN-␥ (Fig. 5B), our data suggest that Toxoplasma does not interfere with the pTyr701STAT1 formation in the cytoplasm but may cause dephosphorylation of pTyr701-STAT1 in some cells, not in the cytoplasm, but after it reaches the nucleus. Indeed, infected cells lacking nuclear pTyr701-STAT1 staining still showed cytoplasmic staining at 45 min of IFN-␥ treatment (Fig. 7C). In contrast to Toxoplasma, and as expected from the normal IFN-␥-induced expression of IRF1 (Fig. 4C) and STAT1 (Fig.

5D), N. caninum did not alter the phosphorylation state of STAT1: anti-pTyr701-STAT1 staining was predominantly nuclear in all infected cells after 45 min of IFN-␥ treatment and indistinguishable from that seen in uninfected cells (Fig. 7F).

Discussion The ability of Toxoplasma to modify host cell processes and render the cell unresponsive to IFN-␥ may be one of the key mechanisms that allow the parasite to persist in the face of a robust IFN-␥-mediated cellular immunity elicited in immunecompetent hosts. Using HFFs as a model, we investigated how Toxoplasma infection of nonphagocytic cells in tissues affects the cell’s responsiveness to IFN-␥, which is critical for developing antimicrobial properties and maturing into a fully functional APC. We compared three major clonal types of Toxoplasma and its closely related species, N. caninum, to examine whether their ability or inability to inhibit the IFN-␥ response would correlate with the differences in host range, virulence, and persistence. Over 200 genes are known to be regulated by IFN-␥ in various cell types, and functional significance is yet to be attributed to many of them (15, 42). Our microarray experiments with human fibroblasts identified 127 unique genes that can be induced by IFN-␥, and no genes were found to be suppressed by IFN-␥ (in mouse macrophages, ⬃10% of the IFN-␥-responsive genes appear to be suppressed by IFN-␥ (42)). Remarkably, we have found that none of the 127 IFN-␥-responsive genes could be up-regulated in Toxoplasma (type II Pru)-infected HFFs following IFN-␥ treatment. This unresponsiveness can be accounted for by two different mechanisms: first, infection causes the activation of a large number of IFN-␥-responsive genes, presumably through the activation of NF-␬B, and these genes cannot be further induced by subsequent exposure to IFN-␥. Secondly, Toxoplasma seems to specifically

The Journal of Immunology target and disrupt the STAT1 transcriptional activity in the host cell nucleus, thus blocking the expression of STAT1-depedent, IFN-␥-responsive genes. We found that ⬃50% of the 127 IFN-␥-responsive genes were already induced or suppressed during infection before IFN-␥ treatment, and, according to biological network analyses, these genes were associated with the NF-␬B pathway regulated by proinflammatory cytokines, such as TNF-␣ and IL-1␤. Although several reports have suggested that TNF-␣ and IL-1␤ could activate IFNresponsive genes, only a limited number of genes have been examined in their studies (33–35). Our microarray experiments suggest that a remarkable number of IFN-␥-responsive genes may be transcriptionally regulated by these and other cytokines produced during infection. It has been shown that the promoter regions of some IFN-inducible genes already occupied by NF-␬B are not readily accessible to STAT1 dimers (39). This may be one of the reasons why the genes belonging to the P0 ⬎ N0 and P0 ⬍ N0 groups could not be further up-regulated by IFN-␥ in Toxoplasma-infected HFFs. The other half of the 127 IFN-␥-responsive genes showed similar expression levels in infected and uninfected cells (P0 ⫽ N0 group) and was closely associated with the IFN-␥/STAT1 network. This result suggests that promoter regions of these genes may remain free of other transcription factors during infection and may be regulated primarily by STAT1. Investigation of the promoter regions of the genes in the P0 ⬎ N0, P0 ⬍ N0, and P0 ⫽ N0 groups may lead to new information as to the mechanisms by which a given set of genes is regulated by several distinct pathways and why a second signal cannot override the existing conditions of the promoter regions to further activate the target genes. Moreover, the genes that are regulated by IFN-␥ but not by other cytokines (i.e., the P0 ⫽ N0 group) may hold an answer to why the immune system has evolved to produce IFN-␥ in addition to many other cytokines and type I IFNs, which apparently accomplish many redundant functions. In this regard, it is noteworthy that STAT1 itself and such genes as PSMP8, LAP3, TAP1, and HLA-E that function in the MHC Ag processing and presentation pathway belong to the P0 ⫽ N0 group, indicating a uniquely important role of IFN-␥ in facilitating the interaction between APCs and T cells (15). We have demonstrated by IFM that all three types of Toxoplasma, but not N. caninum, block the IFN-␥-inducible expression of IRF1 and STAT1. The inhibition was confined to infected cells, and neighboring, uninfected cells expressed IRF1 and STAT1 normally in response to IFN-␥. This strongly argues against the possibility that soluble factors secreted by infected cells into the culture medium exert an inhibitory paracrine effect. For IRF1 and STAT1, as well as for other IFN-␥-responsive genes of the P0 ⫽ N0 group, the recruitment of NF-␬B to their promoter regions alone cannot explain the lack of their expression in Toxoplasmainfected cells following IFN-␥ treatment; consistent with our microarray analysis results with a type II Pru strain suggesting the activation of NF-␬B, Toxoplasma type II strains have been shown to cause a persistent presence of NF-␬B in the host cell nucleus but types I and III strains do not (J. Saeij, unpublished results). N. caninum also recruits NF-␬B to the host cell nucleus (G. Alvarez, unpublished results). Because IRF1 expression cannot be induced by IFN-␥ in STAT1⫺/⫺ cells (42) and is blocked by all three types of Toxoplasma but not by N. caninum, our data indicate that all Toxoplasma strains have evolved a common ability to directly target and disturb STAT1 transcriptional activity, which has probably been key to their intracellular survival in a wide variety of hosts. Unlike a previous report (13), we have observed that neither type I nor type II Toxoplasma causes the degradation of STAT1 or

5163 increased expression of SOCS1 in HFFs. Thirty minutes after IFN-␥ treatment, STAT1 was found in the nucleus in all infected and uninfected cells (according to IFM with anti-total STAT1 Ab), indicating that early events of STAT1 activation (phosphorylation at Tyr701 by JAK1/2 in the cytoplasm and subsequent nuclear translocation of STAT1 dimers) are intact in infected HFFs. However, we found that not all STAT1 in the nucleus was phosphorylated at the Tyr701 residue, suggesting that infection appeared to cause dephosphorylation of STAT1 in some cells after it reaches the nucleus. This could negatively regulate STAT1 transcriptional activity because tyrosine phosphorylation is essential for the tight binding of STAT1 dimers to DNA. However, such dephosphorylation occurred at varying degrees, depending on parasite strains used, and cannot fully account for the lack of IFN-␥-inducible IRF1 and STAT1 expression seen in every infected cell. In contrast to the parasite modulation of STAT1 phosphorylation at Tyr701 that was clearly detectable in Western blots, pSer727STAT1 formation did not appear to be altered by infection. However, it is possible that the parasite causes dephosphorylation of STAT1 at the Ser727 residue in the nucleus, which is thought to be necessary for a maximal transcriptional activity of STAT1. We have been unable to address this possibility conclusively because anti-pSer727-STAT1 Abs acquired from two independent commercial sources were contaminated with reactivities toward Toxoplasma in addition to host cell nuclei in IFN-␥-treated cells (data not shown). Because we know that some secreted Toxoplasma proteins are targeted to the host cell nucleus (see below), we cannot rule out that the nuclear staining we observed may be due to reactivities to parasite Ags. Another possible mechanism of STAT1 inhibition is an autocrine effect within Toxoplasma-infected cells. It has been shown that TGF-␤1 production restricted to type I Toxoplasma (RH)infected murine microglial cells could inhibit the induction of inducible NO synthase by IFN-␥ (55). Recently, it has been reported that TGF-␤1 may block the IFN-␥-inducible NO production in dendritic cells and macrophages by inhibiting tyrosine phosphorylation of STAT1, presumably, as a result of a physical association between TFG-␤1 receptor and IFN-␥ receptor (56). This autocrine effect of TFG-␤1 is an unlikely explanation for the impaired STAT1 activity we observed in Toxoplasma-infected HFFs because STAT1 nuclear translocation that requires phosphorylation at Tyr701 in the cytoplasm is intact in all infected cells. Based on the fact that STAT1 activation and nuclear translocation occurs normally in infected cells despite the lack of its downstream effects on gene expression, we propose that secreted Toxoplasma proteins present in the host cell nucleus (i.e., Tyr-phosphatase or PIAS1-like inhibitory factors) may interfere with the STAT1 transcriptional activity. Many secreted Toxoplasma molecules are now known to be targeted to specific destinations within a host cell (57). Notably, among the secretome are rhoptry proteins that include at least one protein phosphatase (PP2C-hn) (58) and one protein kinase (Rop16) that are targeted to the host cell nucleus (59). For Rop16, it has been shown that the presence of this highly polymorphic kinase is associated with strain-specific activation of STAT3 and STAT6 in infected cells, such that at 18 h postinfection, cells infected with type I or III strain, but not type II, contained phosphorylated STAT3 and STAT6 in the host cell nuclei (59). The fact that infection itself does not cause STAT1 activation regardless of which strain was used (see Fig. 5 above) and that the IFN␥-inducible gene expression can be blocked by all three Toxoplasma strains indicates that a mechanism shared by all three strains and independent of STAT3/6 regulation is responsible

5164


for the disruption of the IFN-␥ signaling pathway. The results with Rop16 suggest that the as-yet-unidentified parasite molecules might be targeted to the host cell nucleus to specifically interfere with STAT1 activity. In striking contrast to Toxoplasma, its closely related apicomplexan cousin N. caninum lacked the ability to block the IFN-␥inducible expression of IRF1, STAT1, and MHC class II. Although Toxoplasma is ubiquitous in various mammalian species, N. caninum predominantly infects dogs, cattle, and sheep (23, 24). In mice, both Toxoplasma and N. caninum induce a strong IFN-␥ response that is critical for the resolution of acute infection. However, N. caninum requires a 100 –10,000-fold higher inoculum than Toxoplasma strains to establish a productive infection in mice by i.p. injection (10, 60, 61), and, even then, brain cysts are almost nonexistent during a persistent infection (62) (G. Alvarez, unpublished results). The inability to block the host’s IFN-␥ response may cause N. caninum to be readily cleared during infection and unable to persist, unlike Toxoplasma that rapidly expands and disseminates, and establishes a life-long persistent infection in many tissues. Our present study provides a convenient screening tool for identifying Toxoplasma molecules that disturb the STAT1-dependent IFN-␥ signaling pathway. The lack of IFN-␥-inducible IRF1 expression seen in infected fibroblasts was reproducible in HeLa cells with all three Toxoplasma strains (data not shown). Genetic and biochemical analyses of mutant parasites that fail to block the activation of the IRF1 promoter in reporter cells will reveal exact mechanisms by which Toxoplasma dysregulates the expression of IFN-␥-responsive genes and aid in assessing their roles in the pathogenesis and persistence of the parasite in vivo.

Acknowledgments We are grateful to the Stanford University Functional Genomics Facility for advice on microarray experiments and to Drs. M. Lodoen, J. Saeij, and G. Alvarez for communicating unpublished results. We thank Drs. M. Lodoen and J. Saeij for helpful comments on the manuscript.

Disclosures The authors have no financial conflict of interest.

References 1. Suzuki, Y., M. A. Orellana, R. D. Schreiber, and J. S. Remington. 1988. Interferon-␥: the major mediator of resistance against Toxoplasma gondii. Science 240: 516 –518. 2. Suzuki, Y., F. K. Conley, and J. S. Remington. 1989. Importance of endogenous IFN-␥ for prevention of toxoplasmic encephalitis in mice. J. Immunol. 143: 2045–2050. 3. Aline, F., D. Bout, and I. Dimier-Poisson. 2002. Dendritic cells as effector cells: ␥ interferon activation of murine dendritic cells triggers oxygen-dependent inhibition of Toxoplasma gondii replication. Infect. Immun. 70: 2368 –2374. 4. Ceravolo, I. P., A. C. Chaves, C. A. Bonjardim, D. Sibley, A. J. Romanha, and R. T. Gazzinelli. 1999. Replication of Toxoplasma gondii, but not Trypanosoma cruzi, is regulated in human fibroblasts activated with ␥ interferon: requirement of a functional JAK/STAT pathway. Infect. Immun. 67: 2233–2240. 5. Halonen, S. K., G. A. Taylor, and L. M. Weiss. 2001. ␥ Interferon-induced inhibition of Toxoplasma gondii in astrocytes is mediated by IGTP. Infect. Immun. 69: 5573–5576. 6. Pfefferkorn, E. R. 1984. Interferon ␥ blocks the growth of Toxoplasma gondii in human fibroblasts by inducing the host cells to degrade tryptophan. Proc. Natl. Acad. Sci. USA 81: 908 –912. 7. Bohne, W., J. Heesemann, and U. Gross. 1993. Induction of bradyzoite-specific Toxoplasma gondii antigens in ␥ interferon-treated mouse macrophages. Infect. Immun. 61: 1141–1145. 8. Jones, T. C., K. A. Bienz, and P. Erb. 1986. In vitro cultivation of Toxoplasma gondii cysts in astrocytes in the presence of ␥ interferon. Infect. Immun. 51: 147–156. 9. Robben, P. M., D. G. Mordue, S. M. Truscott, K. Takeda, S. Akira, and L. D. Sibley. 2004. Production of IL-12 by macrophages infected with Toxoplasma gondii depends on the parasite genotype. J. Immunol. 172: 3686 –3694. 10. Kim, S. K., and J. C. Boothroyd. 2005. Stage-specific expression of surface antigens by Toxoplasma gondii as a mechanism to facilitate parasite persistence. J. Immunol. 174: 8038 – 8048.

11. Lang, C., M. Algner, N. Beinert, U. Gross, and C. G. Luder. 2006. Diverse mechanisms employed by Toxoplasma gondii to inhibit IFN-␥-induced major histocompatibility complex class II gene expression. Microbes. Infect. 8: 1994 –2005. 12. Luder, C. G., W. Walter, B. Beuerle, M. J. Maeurer, and U. Gross. 2001. Toxoplasma gondii down-regulates MHC class II gene expression and antigen presentation by murine macrophages via interference with nuclear translocation of STAT1␣. Eur. J. Immunol. 31: 1475–1484. 13. Zimmermann, S., P. J. Murray, K. Heeg, and A. H. Dalpke. 2006. Induction of suppressor of cytokine signaling-1 by Toxoplasma gondii contributes to immune evasion in macrophages by blocking IFN-␥ signaling. J. Immunol. 176: 1840 –1847. 14. Dubey, J. P. 1997. Bradyzoite-induced murine toxoplasmosis: stage conversion, pathogenesis, and tissue cyst formation in mice fed bradyzoites of different strains of Toxoplasma gondii. J. Eukaryotic Microbiol. 44: 592– 602. 15. Boehm, U., T. Klamp, M. Groot, and J. C. Howard. 1997. Cellular responses to interferon-␥. Annu. Rev. Immunol. 15: 749 –795. 16. Platanias, L. C. 2005. Mechanisms of type-I- and type-II-interferon-mediated signalling. Nat. Rev. Immunol. 5: 375–386. 17. Varinou, L., K. Ramsauer, M. Karaghiosoff, T. Kolbe, K. Pfeffer, M. Muller, and T. Decker. 2003. Phosphorylation of the Stat1 transactivation domain is required for full-fledged IFN-␥-dependent innate immunity. Immunity 19: 793– 802. 18. Wen, Z., Z. Zhong, and J. E. Darnell, Jr. 1995. Maximal activation of transcription by Stat1 and Stat3 requires both tyrosine and serine phosphorylation. Cell 82: 241–250. 19. Dubey, J. P. 1997. Distribution of tissue cysts in organs of rats fed Toxoplasma gondii oocysts. J. Parasitol. 83: 755–757. 20. Dubey, J. P. 1997. Tissue cyst tropism in Toxoplasma gondii: a comparison of tissue cyst formation in organs of cats, and rodents fed oocysts. Parasitology 115(Pt. 1): 15–20. 21. Buckley, C. D., D. Pilling, J. M. Lord, A. N. Akbar, D. Scheel-Toellner, and M. Salmon. 2001. Fibroblasts regulate the switch from acute resolving to chronic persistent inflammation. Trends Immunol. 22: 199 –204. 22. Boyle, J. P., B. Rajasekar, J. P. Saeij, J. W. Ajioka, M. Berriman, I. Paulsen, D. S. Roos, L. D. Sibley, M. W. White, and J. C. Boothroyd. 2006. Just one cross appears capable of dramatically altering the population biology of a eukaryotic pathogen like Toxoplasma gondii. Proc. Natl. Acad. Sci. USA 103: 10514 –10519. 23. Speer, C. A., J. P. Dubey, M. M. McAllister, and J. A. Blixt. 1999. Comparative ultrastructure of tachyzoites, bradyzoites, and tissue cysts of Neospora caninum and Toxoplasma gondii. Int. J. Parasitol. 29: 1509 –1519. 24. Dubey, J. P., and D. S. Lindsay. 1993. Neosporosis. Parasitol. Today 9: 452– 454. 25. Saeij, J. P., J. P. Boyle, and J. C. Boothroyd. 2005. Differences among the three major strains of Toxoplasma gondii and their specific interactions with the infected host. Trends Parasitol. 21: 476 – 481. 26. Boothroyd, J. C., and M. E. Grigg. 2002. Population biology of Toxoplasma gondii and its relevance to human infection: do different strains cause different disease? Curr. Opin. Microbiol. 5: 438 – 442. 27. Baszler, T. V., M. T. Long, T. F. McElwain, and B. A. Mathison. 1999. Interferon-␥ and interleukin-12 mediate protection to acute Neospora caninum infection in BALB/c mice. Int. J. Parasitol. 29: 1635–1646. 28. Walker, W., C. W. Roberts, D. J. Ferguson, H. Jebbari, and J. Alexander. 1997. Innate immunity to Toxoplasma gondii is influenced by gender and is associated with differences in interleukin-12 and ␥ interferon production. Infect. Immun. 65: 1119 –1121. 29. Mercier, C., S. Lefebvre-Van Hende, G. E. Garber, L. Lecordier, A. Capron, and M. F. Cesbron-Delauw. 1996. Common cis-acting elements critical for the expression of several genes of Toxoplasma gondii. Mol. Microbiol. 21: 421– 428. 30. Eisen, M. B., and P. O. Brown. 1999. DNA arrays for analysis of gene expression. Methods Enzymol. 303: 179 –205. 31. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. USA 98: 5116 –5121. 32. Baechler, E. C., F. M. Batliwalla, G. Karypis, P. M. Gaffney, W. A. Ortmann, K. J. Espe, K. B. Shark, W. J. Grande, K. M. Hughes, V. Kapur, et al. 2003. Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc. Natl. Acad. Sci. USA 100: 2610 –2615. 33. Rivieccio, M. A., G. R. John, X. Song, H. S. Suh, Y. Zhao, S. C. Lee, and C. F. Brosnan. 2005. The cytokine IL-1␤ activates IFN response factor 3 in human fetal astrocytes in culture. J. Immunol. 174: 3719 –3726. 34. Zhu, H., and C. Liu. 2003. Interleukin-1 inhibits hepatitis C virus subgenomic RNA replication by activation of extracellular regulated kinase pathway. J. Virol. 77: 5493–5498. 35. Chernajovsky, Y., and T. R. Reid. 1990. Regulation of the human interferoninducible 6 –16 promoter in tumor necrosis factor-sensitive and resistant mouse cells: role of cAMP as a mediator of signal transduction. J. Interferon Res. 10: 627– 636. 36. O’Neill, L. A., and C. Greene. 1998. Signal transduction pathways activated by the IL-1 receptor family: ancient signaling machinery in mammals, insects, and plants. J. Leukocyte Biol. 63: 650 – 657. 37. Tian, B., D. E. Nowak, M. Jamaluddin, S. Wang, and A. R. Brasier. 2005. Identification of direct genomic targets downstream of the nuclear factor-␬B transcription factor mediating tumor necrosis factor signaling. J. Biol. Chem. 280: 17435–17448. 38. Lockwood, C. J., P. Matta, G. Krikun, L. A. Koopman, R. Masch, P. Toti, F. Arcuri, S. T. Huang, E. F. Funai, and F. Schatz. 2006. Regulation of monocyte


39.

40.

41.

42.

43.

44.

45.

46. 47.

48. 49.

50.

chemoattractant protein-1 expression by tumor necrosis factor-␣ and interleukin-1␤ in first trimester human decidual cells: implications for preeclampsia. Am. J. Pathol. 168: 445– 452. Wei, L., M. R. Sandbulte, P. G. Thomas, R. J. Webby, R. Homayouni, and L. M. Pfeffer. 2006. NF␬B negatively regulates interferon-induced gene expression and anti-influenza activity. J. Biol. Chem. 281: 11678 –11684. O’Donnell, S. M., G. H. Holm, J. M. Pierce, B. Tian, M. J. Watson, R. S. Chari, D. W. Ballard, A. R. Brasier, and T. S. Dermody. 2006. Identification of an NF-␬B-dependent gene network in cells infected by mammalian reovirus. J. Virol. 80: 1077–1086. Collazo, C. M., G. S. Yap, S. Hieny, P. Caspar, C. G. Feng, G. A. Taylor, and A. Sher. 2002. The function of ␥ interferon-inducible GTP-binding protein IGTP in host resistance to Toxoplasma gondii is Stat1 dependent and requires expression in both hematopoietic and nonhematopoietic cellular compartments. Infect. Immun. 70: 6933– 6939. Gil, M. P., E. Bohn, A. K. O’Guin, C. V. Ramana, B. Levine, G. R. Stark, H. W. Virgin, and R. D. Schreiber. 2001. Biologic consequences of Stat1-independent IFN signaling. Proc. Natl. Acad. Sci. USA 98: 6680 – 6685. Taniguchi, T., K. Ogasawara, A. Takaoka, and N. Tanaka. 2001. IRF family of transcription factors as regulators of host defense. Annu. Rev. Immunol. 19: 623– 655. Silva, N. M., C. V. Rodrigues, M. M. Santoro, L. F. Reis, J. I. Alvarez-Leite, and R. T. Gazzinelli. 2002. Expression of indoleamine 2,3-dioxygenase, tryptophan degradation, and kynurenine formation during in vivo infection with Toxoplasma gondii: induction by endogenous ␥ interferon and requirement of interferon regulatory factor 1. Infect. Immun. 70: 859 – 868. Alexander, W. S., R. Starr, J. E. Fenner, C. L. Scott, E. Handman, N. S. Sprigg, J. E. Corbin, A. L. Cornish, R. Darwiche, C. M. Owczarek, et al. 1999. SOCS1 is a critical inhibitor of interferon ␥ signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell 98: 597– 608. Shuai, K., and B. Liu. 2005. Regulation of gene-activation pathways by PIAS proteins in the immune system. Nat. Rev. Immunol. 5: 593– 605. Radaeva, S., B. Jaruga, W. H. Kim, T. Heller, T. J. Liang, and B. Gao. 2004. Interferon-␥ inhibits interferon-␣ signalling in hepatic cells: evidence for the involvement of STAT1 induction and hyperexpression of STAT1 in chronic hepatitis C. Biochem. J. 379: 199 –208. Reich, N. C., and L. Liu. 2006. Tracking STAT nuclear traffic. Nat. Rev. Immunol. 6: 602– 612. Schindler, C., K. Shuai, V. R. Prezioso, and J. E. Darnell, Jr. 1992. Interferondependent tyrosine phosphorylation of a latent cytoplasmic transcription factor. Science 257: 809 – 813. Vinkermeier, U. 2004. Getting the message across STAT: design principles of a molecular signaling circuit. J. Cell Biol. 167: 197–201.

5165 51. Ramana, C. V., M. Chatterjee-Kishore, H. Nguyen, and G. R. Stark. 2000. Complex roles of Stat1 in regulating gene expression. Oncogene 19: 2619 –2627. 52. Andrejeva, J., D. F. Young, S. Goodbourn, and R. E. Randall. 2002. Degradation of STAT1 and STAT2 by the V proteins of simian virus 5 and human parainfluenza virus type 2, respectively: consequences for virus replication in the presence of ␣/␤ and ␥ interferons. J. Virol. 76: 2159 –2167. 53. Siewert, E., W. Muller-Esterl, R. Starr, P. C. Heinrich, and F. Schaper. 1999. Different protein turnover of interleukin-6-type cytokine signalling components. Eur. J. Biochem. 265: 251–257. 54. Xuan, Y. T., Y. Guo, Y. Zhu, O. L. Wang, G. Rokosh, R. O. Messing, and R. Bolli. 2005. Role of the protein kinase C-⑀-Raf-1-MEK-1/2-p44/42 MAPK signaling cascade in the activation of signal transducers and activators of transcription 1 and 3 and induction of cyclooxygenase-2 after ischemic preconditioning. Circulation 112: 1971–1978. 55. Rozenfeld, C., R. Martinez, S. Seabra, C. Sant’anna, J. G. Goncalves, M. Bozza, V. Moura-Neto, and W. De Souza. 2005. Toxoplasma gondii prevents neuron degeneration by interferon-␥-activated microglia in a mechanism involving inhibition of inducible nitric oxide synthase and transforming growth factor-␤1 production by infected microglia. Am. J. Pathol. 167: 1021–1031. 56. Takaki, H., Y. Minoda, K. Koga, G. Takaesu, A. Yoshimura, and T. Kobayashi. 2006. TGF-␤1 suppresses IFN-␥-induced NO production in macrophages by suppressing STAT1 activation and accelerating iNOS protein degradation. Genes Cells 11: 871– 882. 57. Dunn, J. D., B. Butcher, E. Denkers, and J. C. Boothroyd. 2006. Alterations in host-cell biology due to Toxoplasma gondii. In Toxoplasma gondii: The Model Apicomplexan-Perspectives and Methods. K. Kim and L. Weiss, eds. Elsevier, New York, pp. 317–340. 58. Gilbert, L. A., S. Ravindran, J. M. Turetzky, J. C. Boothroyd, and P. J. Bradley. 2007. Toxoplasma gondii targets a protein phosphatase 2C to the nucleus of infected host cells. Eukaryot. Cell 6: 73– 83. 59. Saeij, J. P. J., S. Coller, J. P. Boyle, M. E. Jerome, M. W. White, and J. C. Boothroyd. 2007. Toxoplasma co-opts host gene expression by injection of a polymorphic kinase homologue. Nature 445: 324 –327. 60. Long, M. T., T. V. Baszler, and B. A. Mathison. 1998. Comparison of intracerebral parasite load, lesion development, and systemic cytokines in mouse strains infected with Neospora caninum. J. Parasitol. 84: 316 –320. 61. Saeij, J. P., J. P. Boyle, M. E. Grigg, G. Arrizabalaga, and J. C. Boothroyd. 2005. Bioluminescence imaging of Toxoplasma gondii infection in living mice reveals dramatic differences between strains. Infect. Immun. 73: 695–702. 62. Rettigner, C., T. Leclipteux, F. De Meerschman, C. Focant, and B. Losson. 2004. Survival, immune responses and tissue cyst production in outbred (Swiss white) and inbred (CBA/Ca) strains of mice experimentally infected with Neospora caninum tachyzoites. Vet. Res. 35: 225–232.