Cyclic GMP−AMP Synthase Is the Cytosolic Sensor of Plasmodium falciparum Genomic DNA and Activates Type I IFN in Malaria

This information is current as Carolina Gallego-Marin, Jacob E. Schrum, Warrison A. of September 29, 2021. Andrade, Scott A. Shaffer, Lina F. Giraldo, Alvaro M. Lasso, Evelyn A. Kurt-Jones, Katherine A. Fitzgerald and Douglas T. Golenbock J Immunol published online 6 December 2017

http://www.jimmunol.org/content/early/2017/12/20/jimmun Downloaded from ol.1701048 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

*average by guest on September 29, 2021

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 © 2017 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published December 27, 2017, doi:10.4049/jimmunol.1701048 The Journal of Immunology

Cyclic GMP–AMP Synthase Is the Cytosolic Sensor of Plasmodium falciparum Genomic DNA and Activates Type I IFN in Malaria

Carolina Gallego-Marin,*,†,1 Jacob E. Schrum,*,1 Warrison A. Andrade,* Scott A. Shaffer,‡,x Lina F. Giraldo,† Alvaro M. Lasso,† Evelyn A. Kurt-Jones,* Katherine A. Fitzgerald,*,2 and Douglas T. Golenbock*,2

Innate immune receptors have a key role in the sensing of malaria and initiating immune responses. As a consequence of infection, systemic inflammation emerges and is directly related to signs and symptoms during acute disease. We have previously reported that plasmodial DNA is the primary driver of systemic inflammation in malaria, both within the phagolysosome and in the cytosol of effector cells. In this article, we demonstrate that Plasmodium falciparum genomic DNA delivered to the cytosol of human Downloaded from monocytes binds and activates cyclic GMP–AMP synthase (cGAS). Activated cGAS synthesizes 2939-cGAMP, which we subse- quently can detect using liquid chromatography–tandem mass spectrometry. 2939-cGAMP acts as a second messenger for STING activation and triggers TBK1/IRF3 activation, resulting in type I IFN production in human cells. This induction of type I IFN was independent of IFI16. Access of DNA to the cytosolic compartment is mediated by hemozoin, because incubation of purified malaria pigment with DNase abrogated IFN-b induction. Collectively, these observations implicate cGAS as an important cytosolic sensor of P. falciparum genomic DNA and reveal the role of the cGAS/STING pathway in the induction of type I IFN http://www.jimmunol.org/ in response to malaria parasites. The Journal of Immunology, 2018, 200: 000–000.

alaria remains a major cause of morbidity and mortality disease, the problems associated with malaria eradication remain worldwide. The World Health Organization has esti- significant. These include the increasing resistance of insect vec- M mated that there were ∼212 million cases of malaria tors to insecticides and the emerging resistance of Plasmodium globally in 2016 and ∼429,000 deaths, primarily (∼70%) occurring to the most efficacious antimalarial drugs (2). Current evidence in children under age 5 y (1). Despite many gains against the indicates that drug resistance to artemisinin derivatives, the last- generation treatment for asexual blood-stage infection, has de- by guest on September 29, 2021 veloped in Southeast Asia and Africa (3–5). Despite these *Program in Innate Immunity, Division of Infectious Diseases and Immunology, setbacks, efforts continue with the objective of achieving the Department of Medicine, University of Massachusetts Medical School, Worcester, global elimination of malaria. MA 01605; †Centro Internacional de Entrenamiento e Investigaciones Medicas, Cali 760001, Colombia; ‡Proteomics and Mass Spectrometry Facility, University Our understanding of the pathogenesis of malaria is still limited (6). of Massachusetts Medical School, Shrewsbury, MA 01545; and xDepartment of Therefore, a top priority in basic research is to dissect the mecha- Biochemistry and Molecular Pharmacology, University of Massachusetts Medical nisms involved in malaria disease development and provide new School, Worcester, MA 01605 approaches for therapeutic and prophylactic interventions. An im- 1 C.G.-M. and J.E.S. contributed equally to this work. portant component of the pathogenesis of malaria is the host innate 2 K.A.F. and D.T.G. contributed equally to this work. immune response to the parasite. The activation of innate immune ORCIDs: 0000-0001-7018-830X (C.G.-M.); 0000-0002-5264-3257 (S.A.S.); 0000-0002- cells and the associated systemic inflammationleadtotheinitial 3558-7249 (L.F.G.); 0000-0002-1230-6707 (A.M.L); 0000-0003-4669-2084 (E.A.K.-J.); 0000-0002-2447-2358 (D.T.G.). signs and symptoms of disease and can influence the development of Received for publication July 20, 2017. Accepted for publication November 6, 2017. severe disease (7). Inflammatory mediators during malaria infection are produced as a result of direct recognition of plasmodial pathogen- This work was supported by National Institute of Allergy and Infectious Dis- eases Grants R21AI124171 (to E.A.K.-J., D.T.G., and C.G.-M.) and associated molecular patterns (PAMPs) by innate immune receptors, R01AI079293 (to K.A.F. and D.T.G.). A.M.L. and L.F.G. were supported by including TLRs (7), Nod-like receptors (NLRs), and nucleic acid Departamento de Ciencia y Tecnologia – Colciencias Grants 0234-2014 and 0552-2015 (to C.G.-M.) and by National Institutes of Health/Fogarty International sensors (8). Concomitant with TLR activation, expression of sensor Center Training Grant D43TW006589. proteins, including NLRs, is augmented, and NLR inflammasomes Address correspondence and reprint requests to Dr. Douglas T. Golenbock, Univer- are assembled. Proinflammatory cytokines and mediators like TNF-a, sity of Massachusetts Medical School, LRB, Room 328, 364 Plantation Street, IL-12, caspase-1, and IL-1b are then released (9). Elevated expres- Worcester, MA 01605. E-mail address: [email protected] sion of IFN-stimulated genes in innate immune cells is also charac- Abbreviations used in this article: AT-rich ODN, adenine thymine–rich oligodeoxy- teristic during Plasmodium infection (8, 10). nucleotide; BMDM, bone marrow–derived macrophage; cGAMP, cyclic GMP–AMP; cGAS, cyclic GMP–AMP synthase; gDNA, genomic DNA; Hz, hemozoin; IFI16, The recognition of microbial DNA by the immune system IFN-g–inducible protein 16; iRBC, infected RBC; IRF, IFN regulatory factor; ISD, provides a general mechanism for the detection of pathogens immunostimulatory DNA; KO, knockout; LC-MS/MS, liquid chromatography– tandem mass spectrometry; MDM, monocyte-derived macrophage; NLR, Nod-like (11, 12). Delivery of foreign or self-DNA into the cytoplasm receptor; PAMP, pathogen-associated molecular pattern; p(dA:dT), poly(deoxyadenylic- (which is largely free of self-DNA) through microbial infection deoxythymidylic) acid; SeV, Sendai virus; STING, stimulator of IFN genes; TBK1, activates innate cytosolic nucleic acid sensors (13). Plasmodial tank-binding kinase 1; uRBC, uninfected RBC; WT, wild-type. DNA represents a major trigger of innate immunity during Copyright Ó 2017 by The American Association of Immunologists, Inc. 0022-1767/17/$35.00 infection (7, 8, 14). The plasmodial genome contains highly

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1701048 2 cGAS IS THE CYTOSOLIC SENSOR OF P. FALCIPARUM DNA stimulatory CpG motifs, which are thought to activate TLR9 when extracted with phenol/chloroform/isoamyl alcohol and centrifuged at carried into the phagolysosomal compartment by the malaria 10,000 3 g for 10 min. The parasite DNA was precipitated overnight with 2 pigment hemozoin (Hz) (15–17). However, CpG-rich motifs are NaOAc and absolute ethanol at 80˚C, washed with 70% ethanol, and suspended in nuclease- and endotoxin-free water. The DNA concentration relatively rare in Plasmodium falciparum. In contrast, AT-rich was estimated by measuring absorption at 260 nm, and P. falciparum DNA motifs are abundant and induce type I IFN via a pathway gDNA was stored at 220˚C. The purity of P. falciparum gDNA was that is independent of TLRs, DNA-dependent activator of IFN- confirmed by PCR using primers for Plasmodium 18S RNA, as described regulatory factors, RNA polymerase III, and IFN-g–inducible pro- previously (29), and human TLR7 genes. tein 16 (IFI16/p204) but dependent on stimulator of IFN genes Cell culture and stimulation (STING) (18, 19), tank-binding kinase 1 (TBK1) (20, 21), and IFN PBMCs from healthy donors were obtained, as described previously, using regulatory factor (IRF)3 and IRF7 (8, 22). + Ficoll gradient separation (10). Human primary CD14 monocytes were The enzyme cyclic GMP–AMP synthase (cGAS) has been iden- purified from PBMCs using the Pan Monocyte Isolation Kit (Miltenyi tified as a cytosolic DNA sensor whose activation results in the Biotec) and MACS, according to the manufacturer’s instructions. subsequent activation of STING (23, 24). Specifically, in the presence Monocyte-derived macrophages (MDMs) were obtained by adherence and + of cytosolic dsDNA, cGAS catalyzes the synthesis of 2939–cyclic differentiation of CD14 monocytes in RPMI 1640–10% FBS medium for 7 d. Human promonocyte THP-1 cells (wild-type [WT], cGAS+/+, GMP–AMP (cGAMP) from ATP and GTP. 2939-cGAMP then cGAS2/2, IFI16+/+, and IFI162/2) were generated as previously described functions as a second messenger that binds to and activates STING. (30). WT and knockout (KO) THP-1 cells were grown in RPMI 1640/ Activated STING leads to IRF3 phosphorylation, type I IFN pro- glutamine supplemented with 10% FBS. Primary bone marrow–derived 2/2 2/2 duction, and expression of IFN-stimulated genes (12, 25). Recently, macrophages (BMDMs) from WT, STING ,andcGAS C57BL/6 mice were generated, as described previously (31), and cultured in the rodent pathogen Plasmodium yoelii wasreportedtoactivatethe Downloaded from DMEM (Life Technologies-BRL) supplemented with 4 mM glutamine negative regulator SOCS1 in a cGAS-dependent manner; under these and 10% FBS. For stimulations, poly(deoxyadenylic-deoxythymidylic) conditions, TLR7 was found to drive IFN-a/b production (26). In this acid [p(dA:dT)], immunostimulatory DNA (ISD), P. falciparum gDNA, article, we report that cGAS has an important role as a sensor of P. natural Hz, and adenine thymine–rich oligodeoxynucleotides (AT-rich falciparum genomic DNA (gDNA). Our data suggest that 2939- ODNs) were transfected at the indicated concentrations using Lip- ofectamine 2000 (Invitrogen), according to the manufacturer’s in- cGAMP acts as a second messenger after the sensing of P. falciparum structions. After transfection, celldeathwasmonitoredbyTrypanblue gDNA and other malaria PAMPs to induce IFN via the cGAS/STING staining (Corning). Sendai virus (SeV; Cantrell strain, 20 U/ml) was http://www.jimmunol.org/ pathway. We suggest that cGAS detection of cytosolic malaria DNA used as a control where indicated. DNase digestion of P. falciparum is an important molecular feature of malaria pathogenesis. gDNA and Hz was performed using DNase I (QIAGEN), according to the recommended protocol. Primary cells and cell lines were stimulated as stated and collected after the described time points for RNA ex- Materials and Methods traction or cell lysate preparation. Ethics statement Quantitative real-time PCR and ELISA The protocol and consent forms for experiments with human samples Total RNA was extracted using TRIzol Reagent (Invitrogen) or the RNeasy were approved by the Institutional Research Boards from the University Mini Kit (QIAGEN), according to the manufacturers’ instructions. Five of Massachusetts Medical School (IRB H-14839) and the Centro Internacional hundred nanograms of total RNA were used for cDNA synthesis using the by guest on September 29, 2021 de Entrenamiento e Investigaciones Medicas (CIEIH-1249). All experiments iScript Select cDNA Synthesis Kit (Bio-Rad). Quantitative real-time PCR involving animals were performed in accordance with guidelines set forth by was performed using iQ SYBR Green Supermix (Bio-Rad). Levels of the American Association for Laboratory Animal Science and were ap- human IFN-b mRNA were normalized relative to levels of b-actin mRNA proved by the Institutional Animal Care and Use Committee (A-1332) at the and expressed as a fold induction compared with unstimulated controls. University of Massachusetts Medical School. ELISA for mouse IFN-b protein was performed, as described in detail Subjects previously (32). Participants were healthy males and females between 18 and 60 y of age Detection of IRF3 phosphorylation with no prior exposure to malaria or residence in malaria-endemic regions. Cell lysates were prepared in RIPA lysis buffer, as described previously (33), Individuals with any comorbidity at the time of enrollment, recent or subjected to SDS-PAGE, and visualized by Western blotting using Abs concurrent treatment with anti-inflammatory or immunosuppressive drugs, against phospho-IRF3 (S386; Abcam), total IRF3 (D614C; Cell Signaling or pregnancy were excluded. Sixty to one hundred milliliters of total blood Technology), and monoclonal anti–b-actin (Sigma). To accurately determine were collected from healthy donors by phlebotomy. levels of IRF-3 phosphorylation, the intensity signal of the bands was mea- Culture of parasites and natural Hz preparation sured by densitometry using Image Studio Lite software. P. falciparum parasites (3D7 strain) were cultured as described previously Preparation of cytosolic extracts for analysis of endogenous (14). Briefly, plate cultures were prepared with human erythrocytes at a 5% cGAMP hematocrit and ∼1% parasitemia in malaria culture medium (27). Plates were put in a candle jar to produce low oxygen and placed at 37˚C. P. falciparum THP-1 cells were transfected with P. falciparum gDNA, and cytosolic 7 culture stage and parasitemia were assessed daily by Giemsa staining. Where extracts from ∼1.8 3 10 cells were prepared by hypotonic lysis. Briefly, indicated, infected RBCs (iRBCs) were purified from P. falciparum cultures at cells were incubated in hypotonic buffer (10 mM Tris-HCl [pH 7.4], ∼8% parasitemia, and trophozoite or schizont stages were recovered as 10 mM KCl, 1.5 mM MgCl2) for 30 min and then dounce homogenized described (14, 28). The iRBCs suspension was loaded onto LD columns for 100 strokes. Cells lysates were heated at 95˚C for 5 min and then (Miltenyi Biotec), placed in a magnetic cell separator, and eluted with centrifuged at 17,000 3 g for 10 min to remove denatured proteins. The endotoxin-free Dulbecco’s PBS. Natural Hz was extracted from the parasite heat-resistant supernatant was recovered and stored at 280˚C until tested for cultures, as described (16). Briefly, supernatants of P. fa lc ipa rum cultures 2939-cGAMP. were pelleted and loaded onto LD columns as described above. Hz was eluted with Dulbecco’s PBS, quantified, and frozen at 220˚C. Quantification of 2939-cGAMP by liquid chromatography– tandem mass spectrometry Isolation of P. falciparum gDNA Quantification of 2939-cGAMP was performed by liquid chromatography– P. falciparum culture (∼30% parasitemia) at the trophozoite stage was tandem mass spectrometry (LC-MS/MS), as described previously (30). harvested, and parasites from iRBCs were released by treatment with 0.2% Extraction of cGAMP from cell lysates was performed with acetonitrile/ saponin. The released parasites were pelleted at 3000 3 g for 20 min, methanol/water (2:2:1, v/v/v) buffer. 3939 cGAMP (500 pg; Biolog) was washed with ice-cold Dulbecco’s PBS, suspended in Dulbecco’s PBS added to each sample as an internal standard. Extracts of untreated and supplemented with proteinase K (25 mg/ml), and incubated at 56˚C for P. falciparum gDNA–stimulated THP-1 cells were spiked with 3939-cGAMP, 10 min. Then, the resultant P. falciparum gDNA–containing solution was and the levels of 2939-cGAMP (endogenous) and 3939-cGAMP internal The Journal of Immunology 3 standard were measured in parallel. As an additional control, extracts gDNA, Hz, and AT-rich ODNs) activate type I IFN in human of untreated (medium control) THP-1 cells with an added spike of monocytes and MDMs. synthetic 2939-cGAMP (Biolog) were also prepared and used as posi- tive controls. Type I IFN are induced after cytosolic delivery of malaria DNA Statistical analysis To understand the effect of cytosolic location of P. falciparum gDNA, Differences between groups were analyzed with the Student t test or the we determined the role of cytosolic delivery of DNA in the ability of Wilcoxon rank test. Analyses were performed with GraphPad Prism 7 Hz to induce IFN-b expression in WT THP-1 cells. Hz significantly software (GraphPad, San Diego, CA), and p values , 0.05 were considered induced IFN-b expression in a dose-dependent manner when trans- statistically significant. Alternatively, data were analyzed using an unpaired fected into cells. This activity was abolished when Hz was pretreated two-tailed Student t test with a 95% confidence interval, a nonparametric with DNase I (Fig. 2A). Similarly, purified P. falciparum gDNA in- Mann–Whitney U test, or a nonparametric ANOVA (Kruskal–Wallis). duced IFN-b when delivered to the cytosol of THP-1 cells in a dose- Results and time-dependent manner (Fig. 2B and 2C, respectively). Type I IFN are induced in human primary cells in response to Type I IFN induction by P. falciparum gDNA is dependent on P. falciparum gDNA cGAS Previously, our group demonstrated that P. falciparum iRBCs, but cGAS is a DNA sensor that catalyzes the synthesis of cGAMP to not uninfected RBCs (uRBCs), stimulated IFN-b expression in drive activation of STING, resulting in the induction of type I IFN human PBMCs. P. falciparum gDNA also stimulated IFN-b when and other inflammatory genes (23–25). To determine the in- transfected into PBMCs (8). We extended these studies to deter- volvement of cGAS as a sensor for Plasmodium DNA, IFN-b Downloaded from mine the direct effect of P. falciparum on type I IFN induction in expression was assessed in cGAS+/+ and cGAS2/2 THP-1 cells. + + purified human CD14 monocytes. CD14 cells were stimulated Transfection of cultured Hz or P. falciparum gDNA induced ex- with iRBCs or uRBCs or were transfected with P. falciparum pression of IFN-b mRNA in cGAS+/+, but not in cGAS-KO, THP-1 gDNA as a control. A significant induction of IFN-b mRNA was cells (Fig. 3A). Detection of phosphorylated IRF3, an indicator of observed with iRBCs. As expected, uRBCs did not induce IFN-b STING activation and an upstream readout for type I IFN induction, mRNA (Fig. 1A). The induction of IFN-b by iRBCs was similar to +/+ revealed phosphorylation of IRF3 in cGAS cells, but not in http://www.jimmunol.org/ that elicited by transfected P. falciparum gDNA (Fig. 1A). IFN-a cGAS2/2 cells, when transfected with Hz or P. falciparum gDNA was also induced at the mRNA level (data not shown). (Fig. 3B). Cytosolic delivery of AT-rich ODNs also induced IFN-b Previous results demonstrated that Hz presents Plasmodium in cGAS+/+ THP-1 cells but did not induce IFN-b expression in DNA to TLR9 (16) and induces substantial amounts of IFN-b cGAS2/2 cells (Fig. 3C). Finally, transfection of IFI16-deficient + (8). Like CD14 monocytes, human MDMs expressed IFN-b cells with P. fa lc ip ar um gDNA induced equivalent levels of IFN- mRNA in response to transfected P. falciparum gDNA (Fig. 1B). b mRNA as in IFI16+/+ cells (Fig. 3D), suggesting that IFI16 is not We also determined whether other malarial products would in- required for detection of P. falciparum gDNA. Altogether, these duce IFN-b in MDMs. Hz and P. falciparum gDNA strongly results indicate that cGAS is the cytosolic sensor of P. falciparum stimulated IFN-b when delivered into the cytosol of MDMs gDNA and is required for the induction of IFN-b by malaria Hz as by guest on September 29, 2021 (Fig. 1B). Likewise, other known type I IFN inducers [p(dA:dT), carrier of P. fa lc ip ar um gDNA. ISD, and SeV], as well as the AT-rich ODNs AT5 and AT5 3x containing stem-loops [whose design was based on sequences The cGAS–STING pathway is involved in the induction of the from the P. falciparum genome (8)], significantly promoted IFN- type I IFN in response to malaria DNA b expression in MDMs (Fig. 1B). Collectively, these data indi- To corroborate the involvement of the cGAS–STING pathway in cate that malaria parasites and plasmodial PAMPs (P. falciparum the production of IFN-b, we measured IFN-b protein secreted

FIGURE 1. Human CD14+ monocytes produce IFN in response to P. falciparum gDNA and infected RBCs. (A) CD14+ monocytes were isolated from PBMCs from human healthy donors by MACS and transfected with P. falciparum gDNA (10 mg/ml) using Lipofectamine 2000 or were incubated with P. falciparum iRBCs at a 20:1 RBC/monocyte ratio. RNA was prepared after 6 h of incubation or transfection, and IFN-b and b-actin levels were measured by quantitative real-time PCR. (B) Human MDMs from healthy donors were transfected with natural Hz (50 mM), AT5 and AT5 3x ODNs (3 mM), or P. falciparum gDNA (0.5 ml total at 1 mg/ml). IFN-b and b-actin levels were measured as in (A). ISD, p(dA:dT), and SeV were used as positive controls. Data are presented as mean 6 SD and are representative of three independent experiments. All samples were compared with cells incubated with medium alone or between specific groups, when indicated, and analyzed with the Mann–Whitney U test. *p , 0.05, **p , 0.001. Pf, P. falciparum. 4 cGAS IS THE CYTOSOLIC SENSOR OF P. FALCIPARUM DNA

FIGURE 2. Type I IFN are induced by P. falciparum gDNA. (A) The effect of P. falciparum Hz in the induction of IFN-b mRNA was assessed by treatment with DNase I and quantitative real-time PCR. THP-1 cells were transfected with Lipofectamine 2000 and the indicated amounts of Hz for 6 h, followed by measurement of IFN-b mRNA by quantitative real-time PCR. (B) Cells were transfected as described in (A) with the indicated concentrations (mg/ml) of P. falciparum gDNA for 6 h, followed by measurement of IFN-b mRNA by quantitative real-time PCR. (C) Time course of induction of IFN-b mRNA in THP-1 cells transfected with 50 mg/ml P. falciparum gDNA. THP-1 cells incubated with Lipofectamine 2000 or medium alone were used as controls. Data are presented as mean 6 SD and are representative of three independent experiments. *p , 0.05, **p , 0.001, paired Student t test. Pf, P. falciparum. Downloaded from by mouse BMDMs stimulated with P. falciparum gDNA. in the induction of type I IFN in response to the sensing of Cytosolic delivery of P. falciparum gDNA or AT-rich ODNs P. falciparum DNA. induced the secretion of IFN-b from WT BMDMs (Fig. 4). This induction of IFN-b protein secretion was not observed 2939-cGAMP is induced after sensing of P. falciparum gDNA in STING-KO or cGAS-KO macrophages (Fig. 4). These cGAS catalyzes the synthesis of 2939-cGAMP in the presence

data indicate the involvement of the cGAS–STING pathway of cytosolic DNA (24). To further characterize this activity in http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. P. falciparum gDNA induces IFN-b expression through the cGAS–STING pathway. (A) cGAS+/+ and cGAS2/2 THP-1 cells were transfected with 10, 50, or 100 mM of natural Hz for 6 h, followed by detection of IFN-b mRNA by quantitative real-time PCR. P. falciparum gDNA was used as control. Data were analyzed for the difference in IFN-b mRNA induction in WT cells versus KO cells by the paired Student t test. (B) Phosphorylation of IRF3 was detected by SDS-PAGE, followed by Western blot. Detection of total IRF3 and b-actin was used as loading control. (C) Transfection of AT-rich ODNs (AT5, AT5 3x, and dsAT5 4x, 3 mM each) results in a cGAS-dependent induction of type I IFN. SeV was used as control for cGAS-independent induction of IFN-b. (D)IFI16+/+ and IFI16-KO (clones 1 and 2) THP-1 cells were transfected with 10 mg/ml P. falciparum gDNA for 6 h and then IFN-b mRNA was measured by quantitative real-time PCR. Data were analyzed for the difference in IFN-b mRNA induction in WT cells versus KO cells by the Wilcoxon rank test. Data are presented as mean 6 SD and are representative of three independent experiments. *p , 0.05, ***p , 0.0001. Pf, P. falciparum. The Journal of Immunology 5

revealed that the induction of type I IFN is driven by a pathway that did not involve TLR9, DNA-dependent activator of IFN- regulatory factors, RNA polymerase III, IFI16/p204 (8), or DDX41 and IFI203, as reported previously (26). Recently, cGAS has been described as a DNA sensor for detecting malaria DNA in the context of P. yoelii infection. The role of cGAS in P. yoelii was somewhat complex, because cGAS activation appeared to down- regulate IFN-a/b production rather than enhance it through in- duction of the negative regulator SOCS1 (26). In this article, we demonstrate that P. falciparum gDNA and AT-rich ODNs induce type I IFN production through cGAS and the synthesis of 2939- cGAMP, which drives activation of the STING pathway and provides a mechanism for IFN-a/b in malaria pathogenesis. The sensing of pathogen-derived DNA is a central strategy used by the innate immune system to initiate immune responses following microbial invasion (11). cGAS acts as cytosolic DNA sensor that activates innate immune responses through production of the second messenger 2939-cGAMP, which, in turn, activates the adaptor STING FIGURE 4. Type I IFN are produced through the activation of the and leads to type I IFN production (12). The importance of this Downloaded from cGAS–STING pathway in response to P. falciparum gDNA. BMDMs from 2 2 2 2 cytosolic DNA sensor and type I IFN release is now being appre- WT, STING / , and cGAS / mice were transfected with P. falciparum gDNA (1 mg/ml), AT5 and AT5 3x ODNs (3 mM), or 2939 cGAMP (10 nM). ciated in the context of various infectious diseases, beyond their Supernatants were recovered after 18 h of stimulation, and levels of mouse traditional roles in antiviral immunity. A growing list of pathogens as IFN-b were measured by ELISA. Synthetic 2939-cGAMP, ISD, p(dA:dT), diverse as Neisseria gonorrhoeae (30), CMV (35), Mycobacterium and SeV were used as controls. Data are presented as mean 6 SD and are tuberculosis (36), HIV (37), Streptococcus agalactiae (38), Listeria , , representative of three independent experiments. *p 0.05, **p 0.001, monocytogenes (39), and Chlamydia trachomatis (40) all induce http://www.jimmunol.org/ ***p , 0.0001, unpaired t test corrected for multiple comparisons using the type I IFN through the cGAS–STING pathway. The data from this Holm–Sidak method. Pf, P. falciparum. studyalsoshowthatthisistrueforP. falciparum. Exogenous DNA that gains access to the cytosol is a particularly potent and clear danger signal for the innate immune system (12). response to Plasmodium DNA, we transfected THP-1 cells with We have previously demonstrated that natural Hz activates TLR9 P. falciparum gDNA and measured 2939-cGAMP production. We as a result of the delivery of plasmodial DNA to the endosomal began by heat-treating total extracts of transfected cells, because compartment (16). However, a TLR9-independent response also cGAMP is heat stable. We confirmed the induction of 2939- occurs when purified Hz acts as vehicle to deliver Plasmodium cGAMP in cells transfected with P. falciparum gDNA by LS-MS/ DNA into the cytosol, leading to IFN-b production (8). Experi- by guest on September 29, 2021 MS, consistent with the cGAS-dependent induction of IFN-b mentally, we addressed this hypothesis by transfecting P. falciparum shown earlier in response to P. falciparum gDNA (Fig. 3). 2939- gDNA and Hz with Lipofectamine 2000. In contrast to our experi- cGAMP was produced in P. falciparum gDNA–transfected THP-1 mental in vitro conditions, Hz crystals appear to destabilize the cells (Fig. 5A2) but not in control THP-1 cells (Fig. 5A1; 2939- phagolysosome during infection, which allows the delivery of phag- cGAMP peak is represented by a single asterisk [*]). As a positive osomal contents, including DNA, to the cytosol (9). This highlights a control, the lysate from untreated THP-1 cells was spiked with mechanism by which Hz-associated cargo, such as plasmodial DNA, synthetic 2939-cGAMP (Fig. 5A3). All lysates also included an might access the cytosol and suggests that P. falciparum gDNA can internal standard of 3939-cGAMP (500 pg per sample; represented drive IFN-b production upon access to the cytosolic compartment. by the double asterisk [**]). Tandem mass spectrometry of the The large number of AT motifs in the P. falciparum genome 2939-cGAMP peak revealed several fragmented ions with the ex- (present .6000 times) raised the question of their impact in the pected m/z values for product ions of 2939-cGAMP, and they were cGAS/STING-dependent type I IFN response. Our previous observed in similar ratios in DNA-transfected samples and spiked studies showed the immune-stimulatory effect of this unique motif controls (Fig. 5B), confirming the induction of 2939-cGAMP by containing stem-loop secondary structures in human and mouse transfected P. falciparum gDNA. Altogether, our data indicate that cells (8). cGAS has been described as a sensor of dsDNA, in a the cGAS–P. falciparum DNA complex activates 2939-cGAMP nucleotide length–dependent sequence-independent manner (41). synthesis and, consequently, the induction of IFN-b. The data also This DNA sensor has also been linked to the recognition of stem- suggest that cGAS is an important sensor for P. falciparum gDNA. loop DNA structures formed from the HIV-1 genome reverse- transcribed into ssDNA triggering type I IFN production (42). Discussion Collectively, our results suggest that cGAS sensing of AT-rich Much progress has been made in our understanding of how ODNs is involved in the type I IFN response elicited during phagocytes sense Plasmodium and their associated host receptors P. falciparum infection. that elicit inflammation (7). Three major P. falciparum PAMPs Data presented in this study identify the basis of type I IFN have been described: GPI anchors (34), Hz crystals (9, 16), and production mediated by the sensing of cytosolic plasmodial DNA ISD (8, 16). Still, the receptor or family of receptors for Plas- by cGAS and subsequent activation of the cGAS–STING pathway. modium DNA-driven IFN responses has been elusive. In previous This conclusion is supported by a recent study showing that articles, we described “unknown” cytosolic DNA sensor(s) that Tmem173gt mice (coding for a null allele of STING) were com- coupled to STING, TBK1, and the IRF3–IRF7 signaling pathway. pletely resistant and Mb21d12/2 mice (coding for cGAS) showed This receptor or family of receptors acted as a sensor for Plas- partial protection in a lethal model of P. yoelii infection (26). This modium DNA, iRBCs, and oligonucleotides containing the AT- study highlights cGAS as sensor for detection of P. yoelii gDNA rich motif, leading to type I IFN production (7, 8). We also by identifying a decrease in IFN-b mRNA induction after cGAS 6 cGAS IS THE CYTOSOLIC SENSOR OF P. FALCIPARUM DNA Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 5. 2939-cGAMP is induced by P. falciparum gDNA. Lysates from THP-1 cells that were left untreated (medium) or transfected and incubated for 6 h with 50 mg/ml P. falciparum gDNA were heated at 95˚C for 5 min to denature proteins. The heat-resistant supernatants were recovered after centrifugation at 17,000 3 g and analyzed by LC-MS/MS. As a positive control, isolate from untreated THP-1 cells was spiked with 100 nM 2939-cGAMP. (A1) Untreated THP-1 cells (medium), (A2)THP-1cellstransfectedwithP. falciparum gDNA, and (A3) THP-1 cells left untreated and spiked with synthetic 2939-cGAMP. Reconstructed ion chromatograms of cGAMP fragment ion at m/z 312.049 following LC-MS/MS fragmentation of cGAMP protonated ion. Peaks marked with a single asterisk (*) are from endogenous or positive control 2939-cGAMP; double asterisks (**) indicate 3939-cGAMP spiked at 500 pg to each isolate as internal standard. (B)Tandem mass spectra of the peak observed at 6.7 min [from (A2)and(A3)]. Peaks are 2939-cGAMP fragment ions and are observed in similar ratios. Pf, P. falciparum. knockdown (cGAS small interfering RNA in RAW264.7 cells) central to the progression of cerebral malaria, given that mice defi- and an increase in IFN-b mRNA induction in HEK293T cells after cient in these factors survived far longer than WT mice (8). In transfection with cGAS. The current study adds to our knowledge contrast, resistance of mice in the lethal model of P. yoelii has been of plasmodial infections because we have focused on the major linked to early robust IFN-a/b production by plasmacytoid dendritic human pathogenic species of Plasmodium (i.e., P. falciparum). cells, showing that STING and cGAS–deficient mice were more Importantly, by using LC-MS/MS, we provide direct evidence that resistant compared with WT mice (26). A recent report identifies an 2939-cGAMP is produced in response to P. falciparum gDNA. immunological regulatory effect of type I IFN in patients during The consequences of type I IFN induction during malarial in- blood-stage P. falciparum infection. Type I IFN were found to fection are currently under investigation in numerous laboratories. suppress IL-6, but not TNF-a, production by blood monocytes, Several reports have revealed the ability of Plasmodium spp. to in- promote IL-10–producing CD4+ T cells, and generate regulatory Tr1 duce type I IFN and a type I IFN gene signature (8, 10, 15, 43–45). cells (45). Altogether, these data indicate the relevant and complex Previous studies have shown that exogenous rIFN-a can even involvement of type I IFN in the pathogenesis of malaria. inhibit experimental cerebral malaria and reduce parasite burden The activation of innate immune cells and consequent systemic in mice infected with Plasmodium berghei ANKA (43). However, inflammation lead to the initial signs and symptoms of malaria and other studies demonstrate that signaling via the type I IFN re- can also influence development of the more severe forms of the ceptor impairs dendritic cell function and T cell responses that disease. Our data corroborate the role of Plasmodium DNA as a control parasitemia in mouse models of malaria (46). Our studies in potent PAMP during malaria infection and identify cGAS as an the P. b erg hei ANKA model have demonstrated that production of important cytosolic sensor that activates the second messenger type I IFN in a manner dependent on TBK1, IRF3, and IRF7 is 2939-cGAMP. We also reveal the role of the cGAS–STING pathway The Journal of Immunology 7 in the consequent type I IFN production. Our results suggest that 21.Ishii,K.J.,T.Kawagoe,S.Koyama,K.Matsui,H.Kumar,T.Kawai,S.Uematsu, O. Takeuchi, F. Takeshita, C. Coban, and S. Akira. 2008. TANK-binding kinase-1 any attempts at immunomodulatory therapy for patients with severe delineates innate and adaptive immune responses to DNA vaccines. Nature 451: 725–729. malaria will need to take into account the role of DNA-induced type 22. Gun, S. Y., C. Claser, K. S. Tan, and L. Re´nia. 2014. Interferons and interferon I IFN and its impact on malaria pathogenesis. regulatory factors in malaria. Mediators Inflamm. 2014: 243713. 23. Sun, L., J. Wu, F. Du, X. Chen, and Z. J. Chen. 2013. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 339: 786–791. Acknowledgments 24. Wu, J., L. Sun, X. Chen, F. Du, H. Shi, C. Chen, and Z. J. Chen. 2013. Cyclic GMP-AMP is an endogenous second messenger in innate immune signaling by We thank Elizabeth J. Thatcher for assistance with sample preparation cytosolic DNA. Science 339: 826–830. and data analysis and Maria Adelaida Gomez for critically reviewing the 25. Zhang, X., H. Shi, J. Wu, X. Zhang, L. Sun, C. Chen, and Z. J. Chen. 2013. manuscript. Cyclic GMP-AMP containing mixed phosphodiester linkages is an endogenous high-affinity ligand for STING. Mol. Cell 51: 226–235. 26. Yu,X.,B.Cai,M.Wang,P.Tan,X.Ding,J.Wu,J.Li,Q.Li,P.Liu,C.Xing,etal.2016. Disclosures Cross-regulation of two type I interferon signaling pathways in plasmacytoid dendritic The authors have no financial conflicts of interest. cells controls anti-malaria immunity and host mortality. Immunity 45: 1093–1107. 27. Trager, W., and J. B. Jensen. 1976. Human malaria parasites in continuous culture. Science 193: 673–675. 28. Baratin, M., S. Roetynck, C. Le´polard, C. Falk, S. Sawadogo, S. Uematsu, References S. Akira, B. Ryffel, J. G. Tiraby, L. Alexopoulou, et al. 2005. Natural killer cell 1. World Health Organization. 2016. World Malaria Report 2016. World Health and macrophage cooperation in MyD88-dependent innate responses to Plasmo- Organization Press, Geneva, Switzerland. dium falciparum. Proc. Natl. Acad. Sci. USA 102: 14747–14752. 2. Cotter, C., H. J. Sturrock, M. S. Hsiang, J. Liu, A. A. Phillips, J. Hwang, 29. Singh, B., A. Bobogare, J. Cox-Singh, G. Snounou, M. S. Abdullah, and C. S. Gueye, N. Fullman, R. D. Gosling, and R. G. Feachem. 2013. The changing H. A. Rahman. 1999. A genus- and species-specific nested polymerase chain reaction epidemiology of malaria elimination: new strategies for new challenges. Lancet malaria detection assay for epidemiologic studies. Am.J.Trop.Med.Hyg.60: 687–692. Downloaded from 382: 900–911. 30. Andrade, W. A., S. Agarwal, S. Mo, S. A. Shaffer, J. P. Dillard, T. Schmidt, 3. Lu, F., R. Culleton, M. Zhang, A. Ramaprasad, L. von Seidlein, H. Zhou, G. Zhu, V. Hornung, K. A. Fitzgerald, E. A. Kurt-Jones, and D. T. Golenbock. 2016. J. Tang, Y. Liu, W. Wang, et al. 2017. Emergence of indigenous artemisinin- Type I interferon induction by Neisseria gonorrhoeae: dual requirement of cyclic resistant Plasmodium falciparum in Africa. N. Engl. J. Med. 376: 991–993. GMP-AMP synthase and Toll-like receptor 4. Cell Reports 15: 2438–2448. 4. Ashley, E. A., M. Dhorda, R. M. Fairhurst, C. Amaratunga, P. Lim, S. Suon, 31. Charrel-Dennis, M., E. Latz, K. A. Halmen, P. Trieu-Cuot, K. A. Fitzgerald, S. Sreng, J. M. Anderson, S. Mao, B. Sam, et al; Tracking Resistance to Arte- D. L. Kasper, and D. T. Golenbock. 2008. TLR-independent type I interferon misinin Collaboration (TRAC). 2014. Spread of artemisinin resistance in Plas- induction in response to an extracellular bacterial pathogen via intracellular modium falciparum malaria. N. Engl. J. Med. 371: 411–423. recognition of its DNA. Cell Host Microbe 4: 543–554.

5. Bhattarai, A., A. S. Ali, S. P. Kachur, A. Martensson,˚ A. K. Abbas, R. Khatib, 32. Roberts, Z. J., N. Goutagny, P. Y. Perera, H. Kato, H. Kumar, T. Kawai, S. Akira, http://www.jimmunol.org/ A. W. Al-Mafazy, M. Ramsan, G. Rotllant, J. F. Gerstenmaier, et al. 2007. R. Savan, D. van Echo, K. A. Fitzgerald, et al. 2007. The chemotherapeutic agent Impact of artemisinin-based combination therapy and insecticide-treated nets on DMXAA potently and specifically activates the TBK1–IRF-3 signaling axis. malaria burden in Zanzibar. PLoS Med. 4: e309. J. Exp. Med. 204: 1559–1569. 6. Lozano, R., M. Naghavi, K. Foreman, S. Lim, K. Shibuya, V. Aboyans, 33. Seth, R. B., L. Sun, C. K. Ea, and Z. J. Chen. 2005. Identification and charac- J. Abraham, T. Adair, R. Aggarwal, S. Y. Ahn, et al. 2012. Global and regional terization of MAVS, a mitochondrial antiviral signaling protein that activates mortality from 235 causes of death for 20 age groups in 1990 and 2010: a NF-kappaB and IRF 3. Cell 122: 669–682. systematic analysis for the Global Burden of disease study 2010. [Published 34. Zhu, J., G. Krishnegowda, and D. C. Gowda. 2005. Induction of proinflammatory erratum appears in 2013 Lancet. 381: 628.] Lancet 380: 2095–2128. responses in macrophages by the glycosylphosphatidylinositols of Plasmodium 7. Gazzinelli, R. T., P. Kalantari, K. A. Fitzgerald, and D. T. Golenbock. 2014. falciparum: the requirement of extracellular signal-regulated kinase, p38, c-Jun Innate sensing of malaria parasites. Nat. Rev. Immunol. 14: 744–757. N-terminal kinase and NF-kappaB pathways for the expression of proinflammatory 8. Sharma, S., R. B. DeOliveira, P. Kalantari, P. Parroche, N. Goutagny, Z. Jiang, cytokines and nitric oxide. J. Biol. Chem. 280: 8617–8627.

J. Chan, D. C. Bartholomeu, F. Lauw, J. P. Hall, et al. 2011. Innate immune 35. Lio, C. W., B. McDonald, M. Takahashi, R. Dhanwani, N. Sharma, J. Huang, by guest on September 29, 2021 recognition of an AT-rich stem-loop DNA motif in the Plasmodium falciparum E. Pham, C. A. Benedict, and S. Sharma. 2016. cGAS-STING signaling regulates genome. Immunity 35: 194–207. initial innate control of cytomegalovirus infection. J. Virol. 90: 7789–7797. 9. Kalantari, P., R. B. DeOliveira, J. Chan, Y. Corbett, V. Rathinam, A. Stutz, 36. Wiens, K. E., and J. D. Ernst. 2016. The mechanism for type I interferon induction by E. Latz, R. T. Gazzinelli, D. T. Golenbock, and K. A. Fitzgerald. 2014. Dual Mycobacterium tuberculosis is bacterial strain-dependent. PLoS Pathog. 12: e1005809. engagement of the NLRP3 and AIM2 inflammasomes by plasmodium-derived 37. Gao, D., J. Wu, Y. T. Wu, F. Du, C. Aroh, N. Yan, L. Sun, and Z. J. Chen. 2013. hemozoin and DNA during malaria. Cell Reports 6: 196–210. Cyclic GMP-AMP synthase is an innate immune sensor of HIV and other retro- 10. Hirako, I. C., C. Gallego-Marin, M. A. Ataide, W. A. Andrade, H. Gravina, viruses. Science 341: 903–906. B. C. Rocha, R. B. de Oliveira, D. B. Pereira, J. Vinetz, B. Diamond, et al. 2015. 38. Andrade, W. A., A. Firon, T. Schmidt, V. Hornung, K. A. Fitzgerald, E. A. Kurt- DNA-containing immunocomplexes promote inflammasome assembly and re- Jones, P. Trieu-Cuot, D. T. Golenbock, and P. A. Kaminski. 2016. Group B lease of pyrogenic cytokines by CD14+ CD16+ CD64high CD32low inflam- Streptococcus degrades cyclic-di-AMP to modulate STING-dependent type I matory monocytes from malaria patients. MBio 6: e01605–e01615. interferon production. Cell Host Microbe 20: 49–59. 11. Barbalat, R., S. E. Ewald, M. L. Mouchess, and G. M. Barton. 2011. Nucleic acid 39. Hansen, K., T. Prabakaran, A. Laustsen, S. E. Jørgensen, S. H. Rahbæk, recognition by the innate immune system. Annu. Rev. Immunol. 29: 185–214. S. B. Jensen, R. Nielsen, J. H. Leber, T. Decker, K. A. Horan, et al. 2014. Listeria 12. Chen, Q., L. Sun, and Z. J. Chen. 2016. Regulation and function of the cGAS- monocytogenes induces IFNb expression through an IFI16-, cGAS- and STING- STING pathway of cytosolic DNA sensing. Nat. Immunol. 17: 1142–1149. dependent pathway. EMBO J. 33: 1654–1666. 13. Barber, G. N. 2014. STING-dependent cytosolic DNA sensing pathways. Trends 40. Zhang, Y., L. Yeruva, A. Marinov, D. Prantner, P. B. Wyrick, V. Lupashin, and Immunol. 35: 88–93. U. M. Nagarajan. 2014. The DNA sensor, cyclic GMP-AMP synthase, is essential 14. Wu, X., N. M. Gowda, S. Kumar, and D. C. Gowda. 2010. Protein-DNA complex for induction of IFN-b during Chlamydia trachomatis infection. J. Immunol. 193: is the exclusive malaria parasite component that activates dendritic cells and 2394–2404. triggers innate immune responses. J. Immunol. 184: 4338–4348. 41. Stetson, D. B., and R. Medzhitov. 2006. Recognition of cytosolic DNA activates 15. Pichyangkul, S., K. Yongvanitchit, U. Kum-arb, H. Hemmi, S. Akira, an IRF3-dependent innate immune response. Immunity 24: 93–103. A. M. Krieg, D. G. Heppner, V. A. Stewart, H. Hasegawa, S. Looareesuwan, 42. Herzner, A. M., C. A. Hagmann, M. Goldeck, S. Wolter, K. Kubler,€ S. Wittmann, et al. 2004. Malaria blood stage parasites activate human plasmacytoid dendritic T. Gramberg, L. Andreeva, K. P. Hopfner, C. Mertens, et al. 2015. Sequence- cells and murine dendritic cells through a Toll-like receptor 9-dependent path- specific activation of the DNA sensor cGAS by Y-form DNA structures as found way. J. Immunol. 172: 4926–4933. in primary HIV-1 cDNA. Nat. Immunol. 16: 1025–1033. 16. Parroche,P.,F.N.Lauw,N.Goutagny,E.Latz,B.G.Monks,A.Visintin,K.A.Halmen, 43. Viga´rio, A. M., E. Belnoue, A. C. Gruner,€ M. Mauduit, M. Kayibanda, M. Lamphier, M. Olivier, D. C. Bartholomeu, et al. 2007. Malaria hemozoin is im- J. C. Deschemin, M. Marussig, G. Snounou, D. Mazier, I. Gresser, and L. Re´nia. munologically inert but radically enhances innate responses by presenting malaria DNA 2007. Recombinant human IFN-alpha inhibits cerebral malaria and reduces to Toll-like receptor 9. Proc.Natl.Acad.Sci.USA104: 1919–1924. parasite burden in mice. J. Immunol. 178: 6416–6425. 17. Gowda, N. M., X. Wu, and D. C. Gowda. 2011. The nucleosome (histone-DNA 44. Rocha, B. C., P. E. Marques, F. M. S. Leoratti, C. Junqueira, D. B. Pereira, complex) is the TLR9-specific immunostimulatory component of Plasmodium L. R. D. V. Antonelli, G. B. Menezes, D. T. Golenbock, and R. T. Gazzinelli. 2015. falciparum that activates DCs. PLoS One 6: e20398. Type I interferon transcriptional signature in neutrophils and low-density granulocytes 18. Ishikawa, H., Z. Ma, and G. N. Barber. 2009. STING regulates intracellular DNA- are associated with tissue damage in malaria. Cell Reports 13: 2829–2841. mediated, type I interferon-dependent innate immunity. Nature 461: 788–792. 45. Montes de Oca, M., R. Kumar, F. L. Rivera, F. H. Amante, M. Sheel, 19. Zhong, B., Y. Yang, S. Li, Y. Y. Wang, Y. Li, F. Diao, C. Lei, X. He, L. Zhang, R. J. Faleiro, P. T. Bunn, S. E. Best, L. Beattie, S. S. Ng, et al. 2016. Type I P. Tien, and H. B. Shu. 2008. The adaptor protein MITA links virus-sensing interferons regulate immune responses in humans with blood-stage Plasmodium receptors to IRF3 transcription factor activation. Immunity 29: 538–550. falciparum infection. Cell Reports 17: 399–412. 20. Fitzgerald, K. A., D. C. Rowe, B. J. Barnes, D. R. Caffrey, A. Visintin, E. Latz, 46. Haque, A., S. E. Best, A. Ammerdorffer, L. Desbarrieres, M. M. de Oca, B. Monks, P. M. Pitha, and D. T. Golenbock. 2003. LPS-TLR4 signaling to IRF- F. H. Amante, F. de Labastida Rivera, P. Hertzog, G. M. Boyle, G. R. Hill, and 3/7 and NF-kappaB involves the toll adapters TRAM and TRIF. [Published erratum C. R. Engwerda. 2011. Type I interferons suppress CD4+ T-cell-dependent parasite appears in 2003 J. Exp. Med. 198: 1451.] J. Exp. Med. 198: 1043–1055. control during blood-stage Plasmodium infection. Eur. J. Immunol. 41: 2688–2698.