Autoimmune Encephalomyelitis by NK In

In Vivo Regulation of Experimental Autoimmune Encephalomyelitis by NK Cells: Alteration of Primary Adaptive Responses

This information is current as Robin Winkler-Pickett, Howard A. Young, James M. of September 27, 2021. Cherry, John Diehl, John Wine, Timothy Back, William E. Bere, Anna T. Mason and John R. Ortaldo J Immunol 2008; 180:4495-4506; ; doi: 10.4049/jimmunol.180.7.4495

http://www.jimmunol.org/content/180/7/4495 Downloaded from

References This article cites 70 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/180/7/4495.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 27, 2021

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

In Vivo Regulation of Experimental Autoimmune Encephalomyelitis by NK Cells: Alteration of Primary Adaptive Responses1,2

Robin Winkler-Pickett,* Howard A. Young,3* James M. Cherry,† John Diehl,† John Wine,* Timothy Back,* William E. Bere,‡ Anna T. Mason,* and John R. Ortaldo*

Innate immune responses provide the host with its first line of defense against infections. Signals generated by subsets of lymphocytes, including NK cells, NKT cells, and APC during this early host response determine the nature of downstream adaptive immune responses. In the present study, we have examined the role of innate NK cells in an autoimmune model through the use of primary immunization with the myelin oligodendrocyte glycoprotein peptide to induce experimental autoimmune enceph- Downloaded from alomyelitis (EAE). Our studies have shown that in vivo depletion of NK cells can affect the adaptive immune responses, because NK cells were found to regulate the degree of clinical paralysis and to alter immune adaptive responses to the myelin oligodendrocyte glycoprotein peptide. The requirement for NK cells was reflected by changes in the T cell responses and diminished clinical disease seen in mice treated with anti-NK1.1, anti-asialo GM1, and selected Ly49 subtype-depleted mice. In addition to alteration in T cell responses, the maturational status of dendritic cells in lymph nodes was altered both quantitatively and qualitatively. Finally, examination of TCR V␤ usage of the brain lymphocytes from EAE mice indicated http://www.jimmunol.org/ a spectra-type change in receptor expression in NK- depleted mice as compared with non-NK-depleted EAE mice. These findings further establish a recently postulated link between NK cells and the generation of autoreactive T cells. The Journal of Immunology, 2008, 180: 4495–4506.

ncreasing evidence has emerged regarding the potential NK cells in the regulation of dendritic cell (DC)4 maturation in role for NK cells in regulating adaptive immune responses. both humans (4, 5) and mice (5, 6). I Early studies in the CMV infection models clearly dem- In many autoimmune diseases, lymphocytes have been shown to onstrated the important regulatory role of NK cells in the CD8- be the effector cells responsible for damage to the tissue target. T by guest on September 27, 2021 mediated T cell responses (1, 2). In addition to a role in con- cells and B cells, as part of the acquired arm of immunity, have trolling viral infections, NK cells producing IFN-␥ have long been known as key inducers of a variety of autoimmune dis- recently been shown to also regulate generation of T cell im- eases. T cells and B cells have been implicated in uveitis, hemo- munity (3) in a parasitic infection (Toxoplasma gondii). This lytic anemia, colitis, myasthenia gravis, lupus, and rheumatoid ar- innate-adaptive interface has been shown to involve a role for thritis among other diseases (7–12). NK cells, being part of the innate immune system, have been implicated in such autoimmune diseases as diabetes and insulitis, as well as rheumatoid arthritis (13–15). Experimental autoimmune encephalomyelitis (EAE) is a prototypic autoimmune disease induced in laboratory animals, *Laboratory of Experimental Immunology, Cancer and Inflammation Program, Na- bearing significant similarities to multiple sclerosis in clinical and tional Cancer Institute-Center for Cancer Research, †SAIC, Gene Expression Labo- ‡ histopathological aspects (16, 17). EAE is known to be mediated ratory, and Basic Research Laboratory, SAIC-Frederick, National Cancer Institute- ϩ Frederick, Frederick, MD 21702 by CD4 T cells that recognize peptides derived from encephali- Received for publication October 18, 2007. Accepted for publication January togenic proteins of the CNS. Cytokines, particularly TNF-␣, are 27, 2008. considered to be the mediators of the pathology that is observed in The costs of publication of this article were defrayed in part by the payment of page the CNS with conflicting analyses of their effects reported (18). In charges. This article must therefore be hereby marked advertisement in accordance the mouse, the disease is characterized by a paralysis proceeding with 18 U.S.C. Section 1734 solely to indicate this fact. from the hind limbs to the forelimbs. Paralysis initiates within 2 1 This work was supported in whole or in part with federal funds from the Na- tional Cancer Institute, National Institutes of Health, under Contract N01-CO- wk of injection of a myelin oligodendrocyte glycoprotein (MOG) 12400. This work was also supported in part by the Intramural Research Program peptide (19, 20). of the National Institutes of Health, National Cancer Institute, Center for Cancer NK cells are known to be a first line of defense in viral infec- Research. 2 tions. In addition, previous studies have suggested that both NK or The publisher or recipient acknowledges the right of the U.S. government to ϩ retain a nonexclusive, royalty-free license in and to any copyright covering this NK1.1 T (NKT) cells serve as regulatory cells in some T cell- article. The content of this publication does not necessarily reflect the views or mediated experimental autoimmune diseases, including murine policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. government. 3 Address correspondence and reprint requests to Dr. Howard A. Young, Labo- ratory of Experimental Immunology, Cancer and Inflammation Program, 4 Abbreviations used in this paper: DC, dendritic cell; EAE, experimental autoim- 1050 Boyles Street, Frederick, MD 21702. E-mail address: younghow@mail. mune encephalomyelitis; MOG, myelin oligodendrocyte glycoprotein; PT, pertussis nih.gov toxin; LN, lymph node; CT, threshold cycle. www.jimmunol.org 4496 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE

models of encephalomyelitis (EAE) (21, 22), colitis (23), and di- the procedures outlined in the “Guide for the Care and Use of Laboratory abetes (24). Studies by Ljunggren and colleagues (25, 26) have Animals (National Institutes of Health Publication no. 86-23, 1985). detailed the contribution of NK cells in primary B cell-mediated Antibodies autoimmunity. Autoantibodies produced by B cells are the primary cause of disease in a variety of autoimmune conditions, including The mAbs 4E5 (Ly49D), 3D10 (Ly49H), and 3A10 (NKG2D) were provided by Dr. W. Yokoyama (Washington University, St. Louis, MO). The hemolytic anemia, thyroiditis, stiff man syndrome, pemphigus vul- mAb 1F8 (Ly49C/I/H) was a gift from Dr. M. Bennett (University of Texas garis, and systemic lupus erythematosis. Shi et al. (27) demon- Southwestern Medical Center, Dallas, TX). Fluorochrome-labeled isotype- strated that MOG failed to induce EAE in IL-18Ϫ/Ϫ mice. EAE specific controls were purchased from BD Biosciences and BD Pharm- could be observed upon IL-18 administration but disease re- ingen along with fluorochrome-labeled Abs to Ly49D, NK1.1, CD4, CD3␧, and CD80. Fluorochrome-labeled Ab to CD86 was purchased quired the presence of IFN-␥-producing NK cells. These find- Ј from eBioscience. Rabbit F(ab )2 anti-rat IgG was used as a cross- ings established an important, unrecognized link between NK linking reagent (MP Biomedicals). Anti-asialo GM1 was purchased cells and autoimmunity in a primary in vivo model system. from Wako Pure Chemical Industries (Japan) and unlabeled anti-NK1.1 More recently, mice lacking CX3CR1 that developed MOG- and 5E6 (Ly49C/I) were prepared by Hazleton Laboratories (Falls ␧ induced EAE demonstrated a reduced recruitment of NK cells Church, VA). Unlabeled anti-CD3 was purchased from BD Bio- sciences and BD Pharmingen. Abs were used for in vivo depletions of to the CNS (28). cell subsets, flow cytometric, or functional studies. Intracellular IFN-␥ In a another report focused on the possible role of NK cells in detection was performed using Ab kits purchased from BD Biosciences the murine EAE model, Zhang et al. (21) noted that in vivo NK and BD Pharmingen. cell depletion resulted in exacerbation of clinical symptoms in Flow cytometry analysis (FCA) wild-type C57BL/6 (B6) mice. In addition, these investigators also Downloaded from found that depletion of NK cells resulted in an increased severity Cells were stained as previously described (29) and analyzed on an LSR flow cytometer (BD Biosciences) or a FACSort flow cytometer retrofitted of symptoms when disease was induced by passive transfer of a with a solid-state 635-nm laser (BD Biosciences). Data were analyzed us- MOG-specific T cell line. However, in this transfer system, it was ing either CellQuest software (BD Bisociences) or FCSExpress2 software difficult to directly examine the contribution of the endogenous (DeNovo Software). innate immune system. Analyses of results from such adoptive Cytokine measurement

transfer experiments could be further complicated by the fact that http://www.jimmunol.org/ NK cells are known to reject autologous as well as allogeneic cell Cytokines were measured using ELISA kits (R&D Systems). Cell stimu- transfers, potentially resulting in fewer effector cells that are re- lations were performed in a 24-well Costar (Corning) plate at a cell con- centrations of 1 ϫ 106 cells/ml. Cells were stimulated with MOG peptide sponsible for initiating the histopathological effects on the CNS. (1–10 ␮g/106 cells) or anti-CD3 (2–5 ␮g/106 cells; BD Biosciences and These transfer model systems may, in fact, be measuring the ef- BD Pharmingen) plus IL-2 (100 IU/ml; Hoffmann-LaRoche), or MOG pep- fectiveness of cell transfer rather than a role for NK cells in the tide (2 ␮g/106 cells). Unless otherwise stated, samples were collected after actual disease pathology. overnight or 72-h incubation (37°C, 5% CO2) and were measured in du- plicate against the standard curve of the assay and reported as pg/ml. In all The results of many of these studies led to the hypothesis that assays, the SD was Ͻ5 pg/ml. NK cells appear to be involved in changing the balance of immunity by initiating or regulating the intensity of autoimmune reac- RNase protection assay by guest on September 27, 2021 tions and/or modifying the effector cells that can accumulate in the The multiprobe RNase Protection Assay was performed using the mck-1 or target organ(s). These mechanisms might include: 1) production of mck-5 template set (BD Pharmingen). Total cellular RNA was extracted cytokines that alter DC or T cell activation and/or proliferation; 2) using TRIzol (Invitrogen Life Technologies) and 1–5 ␮g of total mRNA 33 ϫ 6 direct interactions with APC that could alter Ag presentation; was hybridized with a [ P]UTP-labeled RNA probe (1 10 cpm/sample) prepared according to the manufacturer’s directions using the BD Pharm- and 3) alteration of regulatory cells by direct or indirect mech- ingen RiboQuant In Vitro Transcription Kit. Following hybridization, the anisms. The primary EAE mouse model provides an in vivo samples were treated with RNase A and T1 according to the procedure system to examine the regulatory role of NK cells because they provided by BD Pharmingen. The RNase was inactivated and precipitated ␮ may provide either protection from or exacerbation of the clin- using a master mixture containing 200 l of Ambion RNase inactivation reagent, 50 ␮l of ethanol, 5 ␮g of yeast tRNA, and 1 ␮l of Ambion ical course of autoimmune disease(s). In addition, the ability to GycoBlue coprecipitate per RNA sample. The samples were mixed well, selectively deplete NK cell subsets in a primary model allows incubated at Ϫ70°C for 15 min and centrifuged at 20,800 ϫ g for 15 min for a more speciﬁc interpretation of the role of NK cells in the at room temperature. The pellets were suspended in 3 ␮l of BD Pharmin- initiation of autoimmune disease. In this study, we examined gen sample buffer and subjected to PAGE as recommended by the manu- whether NK cells could serve as a regulatory element in pri- facturer (BD Pharmingen). mary, peptide-induced EAE. Reagents

The MOG35–55 (MEVGWYRSPFSRVVHLYRNGK) was commercially synthesized to 95% purity by HPLC (Invitrogen Life Technologies and Materials and Methods Advanced ChemTech). CFA was purchased from MP Biomedicals. Heat- Cell isolation killed Mycobacterium tuberculosis H37Ra was purchased from VWR. Per- tussis toxin (PT) was purchased from List Biological Laboratories. IL-2 Cells were isolated from spleen, liver, lymph nodes (LN; axial, inguinal, manufactured by Hoffmann-LaRoche was received from the BRB Preclin- mesenteric, and cervical), and perfused brain of C57BL/6 mice to examine ical Repository at the National Cancer Institute (NCI-Frederick, cell phenotype and cell functions. Liver cells were obtained as previously Frederick, MD). described (29). LN cells were obtained by gentle dissociation of the organ through a plastic mesh screen in ice-cold medium containing 10% FCS. Media Brain lymphocytes were obtained by anesthetizing the animal and perfus- ing saline into the left ventricle of the heart using a peristaltic pump. The RPMI 1640 was purchased from Invitrogen Life Technologies and sup- right atrium is clipped to allow release of vascular volume. Following plemented with L-glutamine, penicillin-streptomycin, and 10% FCS. Saline complete vascular perfusion and euthanasia, brains were harvested and was purchased from Baxter Healthcare. lymphocytes were isolated following the methodologies used to isolate Immunization lymphocytes from the liver tissue. The use of animals for this study was in compliance with policies regarding the humane treatment and care of an- For induction of active EAE, mice were injected s.c. in one ﬂank with 200 ␮ ␮ ␮ imals and was approved by the National Cancer Institute-Frederick Animal l of an emulsion containing 100 l of CFA, 200 gofMOG35–55, and Care and Use Committee. Animal care was provided in accordance with 300 ␮g of pulverized heat-killed M. tuberculosis in saline solution. On the The Journal of Immunology 4497

same day, the mice were injected i.p. with 400 ng of PT in 200 ␮l of saline Results and this treatment was repeated 48 h later. A booster immunization with an In vivo depletion of NK cells identical emulsion was given 1 wk later in the opposite flank with no PT. Onset of paralysis was anticipated to occur 3–7 days after the booster We chose to examine whether NK cells might play an important immunization. role in the development of autoimmunity by using a primary EAE model. The C57BL/6 (B6) model is particularly useful, since the EAE score model for EAE is well established in this mouse strain (31). Ad- Following immunization, mice were monitored daily for clinical signs ditionally, several Abs are routinely used in this strain for in vivo of EAE. The clinical grade was a modification of that which was pre- NK cell depletion because Abs to various NK subsets are available viously reported: 0, no clinical signs; 1, complete loss of tail tonicity; and various gene knockout mice have been generated on the B6 2, flaccid tail and abnormal gait ataxia and/or paresis of hind limb; 3, background (32–34). Given the availability of an autoimmune hind limb paralysis; 4, hind limb paralysis with foreleg involvement; model and specific NK reagents, we theorized that an alteration in and 5, death (21). the in vivo NK cell content would alter the adaptive responses to

the MOG35–55 peptide responsible for initiation of disease in the In vivo NK cell depletion B6 mouse. The use of a primary immune model is in contrast to Mice were i.p. injected with 100–200 ␮g of anti-NK1.1 mAb (PK136), studies that used adoptive transfer models or T cell lines (35, 36) anti-asialo GM1 (dilution 1/20; 0.2 ml), or 100 ␮g of anti-Ly49 Abs 1–3 to examine the role of NK cells in the development of EAE. Ex- days before first immunization with the emulsion containing the MOG 35–55 amination of the LN as the primary T and NK cell interactive site peptide. for initiation of an immune response resulting in EAE required Downloaded from verification of the presence of a NK population within the LN. T cell proliferative responses Preliminary phenotyping (Fig. 1) of normal B6 LN indicated the ϩ Ϫ To analyze primary T cell responses, 1 ϫ 105 LN cells isolated from consistent presence of 0.5–1.0% NK1.1 CD3 cells. These cells mice immunized with MOG35–55 peptide were cultured in 96-well flat- could be depleted with anti-NK1.1 or asialo GM1 Ab with 90% ␮ bottom plates with MOG35–55 peptide at a concentration of 2 g/ml in suppression maintained at day 7 and recovery of the population in 0.2 ml of RPMI 1640 medium supplemented with 10% FCS. The plates o the LN by day 14. The spleen and liver demonstrated similar ki- were incubated for 72 h at 37 C in humidified air containing 5% CO2. http://www.jimmunol.org/ Incorporation of [3H]thymidine (1␮Ci/well) for the final 18 h of the netics of suppression (data not shown). In addition to pan-NK- incubation was measured on a Trilux 1450 Microbeta liquid scintilla- depleting Abs, specific subsets of Ly49D (Fig. 1G) and Ly49H tion counter following harvest on a TomTec Harvester 96 Mach 3 (Fig. 1H) NK cells could be deleted efficiently from mice in vivo, (PerkinElmer). relative to control mice (Fig. 1F). To examine the clinical impact of a change in NK cell content Spectratype analysis of TCR V␤ usage during the EAE immunization/induction schedule, NK depletions TCR V␤ usage was analyzed following previously published methods (30). were done either before initial immunization of MOG peptide or Briefly, mRNA from lymphocytes isolated from draining, nondraining, and just before secondary immunization (at day 6). Fig. 2 shows the cervical LN and perfused brain was extracted using TRIzol (Invitrogen clinical score (Fig. 2), survival (Fig. 2), and frequency of dis- by guest on September 27, 2021 ␤ Life Technologies) according to the manufacturer’s protocol. TCR V ease (Fig. 2) for mice depleted of NK cells before primary im- products were synthesized using the Qiagen One-Step RT-PCR kit and primers unique for the 5Ј end of V␤1, 6, 8.1, 8.2, 8.3, 14, and 15 as well munization. EAE control mice that received either nothing, as a common 3Ј primer (Qiagen Operon). Primer sequences have been mIgG2a (Cntl), or normal rabbit serum (NRS; data not shown) previously published (30). generated a rapid clinical disease, starting about day 15 that peaked at days 22–25, and was generally maintained for 30 Analysis of gene expression by quantitative PCR additional days. Mice depleted of NK cells with either NK1.1 or asialo GM1 Abs exhibited significantly ( p Ͻ 0.05) reduced Quantitative PCR analysis was performed using SYBR Green Chemistry (Qiagen) according to the manufacturer’s instructions in 10-␮l final vol- average clinical scores; however, the time to onset of clinical umes in 384-well microtiter plates. Specific primers for detection of TCR disease was consistently delayed by 2–5 days in the non-NK- V␤ usage of tissue lymphocytes in EAE-immunized mice were defined depleted mice. Clinical manifestations of EAE in the mice as described above and were purchased from Qiagen. The endogenous tended to fade slightly after 30–35 days, with variability from control primer MuGAPDH was purchased from Applied Biosystems experiment to experiment. Survival in both NK-depleted groups and the sequence is proprietary in nature. Thermocycling conditions using an Applied Biosystems 7900 SDS were as follows; 95°C for 15 was significantly higher than the EAE control group; however, min and 40 cycles of 95°C for 15 s and 60°C for 1 min. Accurate the frequency of the number of animals displaying disease was quantification of each mRNA was achieved using the normalization of not statistically different. Mice depleted of NK cells after pri- ⌬ the sample CT values to one reference. This value, referred to as the mary, but before secondary immunization, failed to demonstrate ⌬ Ϫ ⌬ Ϫ ϭ Ϫ CT sample value ( CT sample CT reference), is derived by Ϫ⌬ Ϫ Ͼ any consistent differences in parameters shown in Fig. 2, A–C taking the result of the expression: if 2( CT) 1 0, then the ϭ Ϫ⌬ Ϫ ϭϪ ⌬ (data not shown). Thus, depletion of NK cells before immuni- result 2( CT) 1 or else the result 1/2(- CT). This equation changes the range for down-regulation from 0 through 1 to -ϱ through zation with MOG peptide decreased both the frequency and 0 and up-regulation from 1 through ϱ to 0 through ϰ. The samples that extent of disease compared with the EAE control mice. Also, were assayed for expression were normalized to murine GAPDH and any contribution of NKT cells can be ruled out since similar amplifications were conducted under identical conditions for each gene of interest. The target mRNA expression was normalized to the GAPDH results were obtained with both anti-NK1.1 and asialo GM1 expression, and the relative expression was calculated back to the EAE (Fig. 2A), as the latter treatment (in normal mice) does not controls for each cell type. Raw data from each quantitative PCR run remove NKT cells. were exported into a comparative CT analysis workbook. CT represents the threshold cycle or the PCR cycle at which an increase in reporter In vivo depletion of NK receptor (NKR)-bearing subsets fluorescence above baseline signal can be detected. The comparative CT workbook allows for normalization with different endogenous controls Because total NK depletion altered the course of disease, we on a number of samples and genes. Each graph displays the analyzed results in a format (both numerically and graphically) showing their sought to determine whether Ly49 NKR-bearing subsets of NK expression relative to not only an endogenous control but also a refer- cells might demonstrate a similar alteration. Thus, Abs to Ly49C/I ence sample. (present on ϳ50% of NK cells), Ly49D (present on 25–30% of NK 4498 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE

FIGURE 1. In vivo depletion of draining LN NK cells. A–D, Mice were treated with control Ig (A and B) Downloaded from or anti-NK1.1 (C and D). Mice were evaluated for the NK cell content of their draining LN on day 7 (A and C) and day 14 (B and D). Subsets of Ly49-activating receptors were evaluated by examination of CD3Ϫ NK1.1ϩ cells (E; gate r2) and evalu- http://www.jimmunol.org/ ating the expression of Ly49D and/or Ly49H (F). In vivo depletion of these subsets is shown in LN 3 days after Ab administration of anti-Ly49D (G) or anti-Ly49H (H). by guest on September 27, 2021

cells), and Ly49H (present on 25–30% of NK cells) were used to suggested that more effective disease alteration was coincident deplete NK subsets. Preliminary phenotypic screening verified with the activating Ly49-bearing NK subsets. In addition, it is that the LN NK cells expressed normal patterns of activating unlikely that these Ab treatments might be altering other leu- and inhibitory NKRs (Fig. 1) as compared with spleen and liver kocyte subsets, since these activating NKRs are not expressed (data not shown). The preliminary screening confirmed that a on T cells or monocytes/macrophages (37, 38). selective loss of the Ly49D NKR was observed when mice were treated with Abs to Ly49D or Ly49H (3D10). Thus, our NKR Role of Ly49H in MOG-induced EAE depletion can effectively alter NKR- bearing subsets in the LN, Data in Fig. 2 suggest that Ly49D- and Ly49H-bearing subsets the site where immunity to MOG is being induced. Elimination might be critical for EAE development. To further evaluate this of activating NKRs Ly49D and Ly49H resulted in a more ef- result, we used a specific anti-Ly49H Ab (3D10) for depletion. ficient alteration of clinical score (Fig. 2D) that was comparable Due to the major overlap between Ly49D and Ly49H (coex- to elimination of the entire NK population or the inhibitory pressed by ϳ75–80% of NK cells), we examined how critical NKRs Ly49C/I and Ly49G2 (data not shown; but similar to this Ly49 NKR was in the development of EAE. Table I com- NK1.1). These activating NKR-depleting Abs not only dimin- pares clinical scores, percent survival, and percentage of mice ished the clinical score but altered the frequency of disease with EAE between B6 control mice and mice that lack the (Fig. 2F) and improved mouse survival (Fig. 2E). These data Ly49H NK cell subset. We found that the removal of the The Journal of Immunology 4499

FIGURE 2. In vivo EAE after NK cell or subset depletion. EAE was measured by clinical score (A and D), number of mice surviving (B and E), or disease incidence (C and F)in

C57BL/6 mice. A–C, The mice were Downloaded from treated with control IgG (Œ), depleted of NK cells with anti-NK1.1 (PK136; F), or anti-asialo GM1 (), whereas in D–F, mice were depleted of NK subsetswithanti-Ly49D(4E5;‚),anti- Ly49C/I/H (1F8; ࡗ), or anti-Ly49C/I (5E6; Ⅺ). Values represent mean and http://www.jimmunol.org/ SD of 10 mice/group and are representative of more than three experiments. Abs were given to mice i.p. 1–3 days before immunization. by guest on September 27, 2021

Ly49H-expressing cells with Ab resulted in reduced clinical Adaptive immune regulation by in vivo depletion of scores. This alteration of autoimmunity was also observed in NKR-bearing subsets the frequency of disease, where only 40–50% of mice depleted with Ly49H Abs developed disease and had a 90% survival To further evaluate the alterations in adaptive immunity in NK- outcome compared with 100% of the EAE control mice devel- depleted mice, draining LN cells from EAE or control mice were oping disease with significantly diminished survival. Thus, col- evaluated for a T cell proliferative response (Fig. 3a) and for pro- lectively, these data are highly indicative of the specificity of duction of either Th1/2 (IL-2, IL-4) or inflammatory cytokines NK cells for altering the clinical course of disease with Ly49 (IFN-␥, TNF-␣, IL-6, IL-10) 10 days after immunization (Table receptors playing a role. II). Cells were unstimulated or stimulated in vitro with 2 ␮g/ml

Table I. Role of Ly49H in NK-mediated modulation of EAE

Strain Mean Clinical Score at Day 24 (SE) % Survival No. Surviving % with EAE No. with EAE

Expt. 1 C57BL/6 2.4 (0.4) 93 13/14 93 13/14 C57BL/6 anti-Ly49H 1.5 (0.5)a 100 10/10 90 9/10 C57BL/6 anti-NK1.1 1.4 (0.3)a 100a 10/10a 90a 8/10a

a Values are signiﬁcantly different from those of C57BL/6 control. 4500 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE Downloaded from http://www.jimmunol.org/

FIGURE 3. Responses of draining LN to MOG peptide. A, Proliferation of draining LN cells from MOG-immunized mice (day 15 following first immunization) were cultured with 2–25 ␮g/ml MOG peptide ( ), anti-CD3 (2 ␮g/ml), and IL-2 (100 U/ml) (s) or nothing (control (Cntl); f). Normal mice, EAE control mice, and EAE mice treated with anti-NK1.1, anti-asialo GM1, anti-Ly49H, and anti-Ly49D were evaluated. Cells were stimulated for ϩ 48 h, then pulsed with [3H]TdR (1 ␮Ci/well) for 18 h. B, Frequency of cytokine production in CD4 T cells from draining LN cells. Cells used in B were by guest on September 27, 2021 evaluated for intracellular production of TNF-␣ and IFN-␥ after8hofstimulation with MOG peptide. Cells were stimulated with 2 ␮g/ml MOG peptide for 2 h, then brefeldin A was added for the remaining 6 h. Values are pools of three to five mice per treatment. C, Cytokine production in CNS leukocytes and peripheral LN. Panel shows the production of IL-2 after 18 h in either brain lymphocytes (Lymphs) or draining LN cells after stimulation with 2 ␮g/ml MOG peptide in naive C57BL/6 mice (f), EAE control mice ( ), and EAE NK1.1-depleted (Dep., dep., Depl.) mice ( ). D, IFN-␥ mRNA expression at6hincells isolated from draining, nondraining, or cervical lymph nodes. Values are pools of three to five mice per treatment. The autoradiograph is a representative example of the RNase protection results. Lymphs, Lymphocytes.

MOG peptide or anti-CD3 plus IL-2 (as a positive control). Table 17-producer T cell, termed Th17 (39–41). This Th17 cell has a II shows that although each group of mice had intact proliferative defined cytokine profile; it has its own set of lineage-specific de- responses to CD3 and IL-2, mice depleted of NK cells with Ab velopmental genes and is the main proinflammatory CD4ϩ effector treatment with anti-NK1.1, anti-Ly49H, anti-Ly49D, or anti-asialo T cell involved in murine models of CIA and EAE. To evaluate the GM1 (Fig. 3a) demonstrated a profound deficit in proliferative potential role for NK cells to modulate IL-17 and/or IL-23 capacity in response to MOG peptide. Mice depleted of other NKR-bearing subsets (Abs to Ly49C/I or Ly49G2) also had diminished proliferative responses but exhibited more variable re- Table II. Cytokine release after in vitro stimulation sults (data not shown). Because the NK or NK subset depletion altered the proliferation Immunization First First and of LN T cells, we evaluated the cytokine production in these drain- Cytokinea Normal Control Immunization NK Depleted ing LN leukocytes. As shown in Table II, there were strong levels TNF-␣ 2 899 189 of MOG-induced TNF-␣ and IFN-␥ production; modest levels of IFN-␥ 1 2203 468 IL-6, IL-12 p70, IL-10, and IL-2 while IL-4 and IL-5 were unde- IL-6 1 229 82 tectable. MOG-induced TNF-␣ and IFN-␥ production was strongly IL-12p70 7 29 21 diminished with NK depletion. When the frequency of CD4ϩ cells IL-10 16 32 44 IL-2 0 100 23 expressing these cytokines (Fig. 3B) was evaluated, a similar pat- IL-4 Ͻ10 Ͻ10 Ͻ10 tern of reduction was observed. Other cytokines evaluated (IL-4, IL-5 Ͻ10 Ͻ10 Ͻ10 IL-13, and IL-10) were either not produced or were not altered by a Cytokine production from draining LN cells. LN cells from normal mice, EAE NK depletion. Thus, the MOG-induced production of cytokines by control mice, and EAE mice treated with anti-NK1.1 were cultured with 2 ␮g/ml CD4ϩ cells was dramatically reduced when NK cells were de- MOG peptide. Cells were stimulated for 24 h, then supernatants were collected and evaluated for cytokine release. EAE mice LN were harvested after the second im- pleted before immunization. Recent studies have demonstrated that munization at day 10. Supernatants were measured using CBA bead technology (BD ϩ the classical Th1/Th2 CD4 T cells have been joined by an IL- Biosciences). All values are pools of three to five mice per treatment. The Journal of Immunology 4501 Downloaded from FIGURE 5. Cytokine production from CNS leukocytes. CNS was perfused with saline and lymphocytes were isolated as described in Materials and Methods. Bar graphs show the CD4ϩ or CD8ϩ T cell cytokine production in normal mice (u), MOG-immunized mice (f), or MOG-immunized and NK1.1-depleted mice (o). Values represent the total number of CD4ϩ or CD8ϩ T cells in the CNS isolate that are producing either TNF-␣ (A) or IFN-␥ (B). Values are pools of three to five mice per treatment. http://www.jimmunol.org/

lacked any significant IL-2 production, NK depletion of EAE-induced animals resulted in a significant reduction of IL-2 production by draining LN cells but not those cells isolated from the FIGURE 4. Production of Th17 cytokines after NK1.1 depletions. A, brain. Evaluation of T cell subsets in brain lymphocytes revealed Cytokine production in draining LN for IFN-␥ (Ⅺ and f), IL-17 (E and F), and IL-23 (छ and ࡗ). Panel shows the production after 18 h in drain- no change in CD4 or CD8 frequency after NK1.1 depletion (20.2 ing LN cells after stimulation with 1 ␮g/ml MOG peptide in individual and 23.1% for CD4 and 19.3 and 18.8% for CD8). In addition, mice from EAE control mice (E, Ⅺ, and छ) and EAE NK1.1-depleted CNS lymphocytes were evaluated for the presence of T regulatory by guest on September 27, 2021 ϩ ϩ ϩ mice (F, f, and ࡗ). B, Intracellular staining of CD4 (f) and CD8 (s)T cells (CD3 CD4 CD25 ) and no differences in frequency were cells. Cytokine production was analyzed in draining LN cells 6 h after observed upon comparison of control or NK- depleted mice (data stimulation with 1 ␮g/ml MOG peptide (three pooled mice). Values ex- not shown). When peripheral nodes (draining, nondraining, and press the number of IFN-␥- or IL-17-positive cells from the EAE- immu- cervical) were examined from the same mice for IFN-␥ mRNA nized mice with or without NK depletion. (Fig. 3D), NK depletion resulted in decreased IFN-␥ mRNA levels. This is consistent with the lower levels of MOG-induced T cell activation that were seen as measured by IL-2 production (Fig. production, draining node lymphocytes were evaluated for their 3C). Measurement of intracellular IFN-␥ and TNF-␣ from the ability to produce IFN-␥, IL-17, and IL-23 in either control or brain lymphocytes of EAE-induced mice demonstrated that these NK-depleted mice 10 days after immunization. Fig. 4a demon- lymphocytes were quite capable of producing both cytokines (Fig. strates the result of one typical experiment examining individual 5). The frequency of T cells producing cytokines after MOG pep- mice, where the production of IFN-␥ was significantly diminished. tide immunization was not significantly different (data not shown), When IL-17 or IL-23 production was examined, no consistent in- but the total number of TNF-␣ (Fig. 5a)- and IFN-␥ (Fig. 5b)-pro- crease was observed in the levels of either cytokine, although the ducing cells was significantly lower after NK cell depletion. Thus, ratio of IFN-␥ and IL-17 is dramatically altered by NK depletion. these data suggested that although NK depletion can significantly alter Fig. 4b shows the number of CD4- or CD8-positive cells produc- the number and quality of responses in the LN, the MOG-reactive ing either IFN-␥ or IL-17 from LN. A similar pattern of IFN-␥ cells that reach the CNS are capable of making cytokines at a similar reduction was seen, whereas IL-17-producing T cell subsets were frequency as compared to the non-NK-depleted mice. Therefore, it not changed upon NK depletion. appears that the decrease in the number of lymphocytes reaching the CNS in NK-depleted mice is responsible for the overall diminution of CNS lymphocyte analysis in NK-depleted mice clinical disease that we observed. Since autoimmunity in the EAE model is known to be caused by an infiltration of T cells that induces a demyelination of neural Role of DC in EAE alteration tissue as well as inflammation of the CNS (31), we further eval- Because the results above suggested that the CNS cells were iden- uated the brain lymphocytes that were found in the CNS (after tical in the EAE and the NK-depleted EAE mice, only differing in saline perfusion) using this primary immunization model system. numbers, while the draining LN T cells from the depleted mice IL-2 production was examined in either brain or draining LN lym- were functionally deficient, we examined this site to determine phocytes after MOG stimulation in vitro (Fig. 3C). Ag-induced whether an innate interaction between NK and DC might be al- IL-2 production was observed in lymphocytes from both brain and tered in NK-depleted mice. When LN and brain of EAE-NK-de- LN lymphocytes of EAE-induced mice. Although naive mice pleted mice were examined for the presence of DC (expression of 4502 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE

Table III. Evaluation of DC and NK cells in EAE brain and LN

Organ Site Group Total Cells (ϫ106) %DCa No. of DCa (ϫ105) %NKa No. of NKa (ϫ104)

Brain EAE 4.8 26.1 (100) 13.0 (100) 1.5 (100) 3.70 (100) Brain NK deplb 2.2 24.6 (94) 5.4 (41)c 0.2 (13)c 0.18 (5)c Brain Control 0.1 9.4 (36) 0.1 (0.7) 1.1 (73) 0.00 (0) Draining LN EAE 14.0 13.9 (100) 20.0 (100) 2.1 (100) 23.00 (100) Draining LN NK depl 8.0 21.9 (158)c 18.0 (90) 0.2 (10)c 12.00 (52)c Draining LN Control 18.0 3.8 (27) 6.8 (34) 1.1 (52) 6.00 (26) Cervical LN EAE 8.2 15.0 (100) 12.0 (100) 1.5 (100) 7.40 (100) Cervical LN NK depl 3.0 11.1 (74) 3.3 (27) 0.2 (13)c 0.48 (7)c Cervical LN Control 1.6 2.3 (15) 1.6 (13) 0.4 (26) 0.63 (6)

a Percent or number has been normalized to EAE group (100%) within each speciﬁc tissue examined. b depl, Depleted. c Values are indicators of signiﬁcant differences between the EAE group and the NK-depleted group.

CD11c and MHC class II, I-Ab; Table III) and compared with the the ratio of bright (mature) to dim (immature) CD80ϩCD86ϩ DC EAE control mice on day 7 following immunization, there were during treatment (Fig. 6, B and C), we found that the NK depletion ϩ ϩ signiﬁcantly fewer DC present in the NK-depleted EAE brain and resulted in an increased frequency of CD80 CD86 (mature) DC Downloaded from cervical LN. Although the draining LN did not show signiﬁcant in the draining nodes, whereas immature DC are favored in the changes in the actual number of DC, the percentage of DC was cervical nodes. increased. The number of brain DC was diminished by 59% while the percentage was almost identical. In addition, the frequency and intensity of expression of CD80 and CD86 on DC (CD11cϩclass TCR usage in EAE after NK depletion ϩ

II ) was not altered by NK cell depletion (data not shown). As Previous studies using MOG35–55 in various H-2 mice have indi- http://www.jimmunol.org/ expected after NK1.1 depletion, a dramatic decrease was observed cated that TCR V␤1, 6, 8,14, and 15 (16, 43–45) are used in in the NK cell numbers in the LN that began to return by day 14 C57BL/6 mice (46, 47) with the major TCR ␤s being 1 and 8. (data not shown). It should be noted that immunization with MOG Since our studies have demonstrated an alteration in T cell re- peptide increased the number of draining LN NK cells compared sponses that is characterized by decreased clinical disease, loss of with naive mice, a finding recently reported in another adaptive Ag responsiveness, and reduced cytokine production, especially in model (42). These data suggest the potential for tissue-specific the peripheral lymphoid tissue, we examined the TCR V␤ usage in interactions between DC and NK cells. the NK-depleted EAE mice by quantitative real-time PCR (30, 47) We next examined the maturational status (based on expression primer sequences. The data from a representative experiment of CD80 and CD86) of draining and cervical LN DC from EAE (n ϭ 3) are shown in Fig. 7. TCR V␤ quantitative usage in by guest on September 27, 2021 and NK-depleted EAE mice evaluated at day 7 postimmunization EAE-immunized mice was compared with NK-depleted, immu- (Fig. 6). NK depletion modified the DC composition as found in nized mice. All TCR V␤s were normalized to a housekeeping draining LN (site of immunization); there was an increase in the gene and compared with the EAE controls. At the site of clinical percentage of CD80ϩCD86ϩ DC in EAE NK-depleted mice, sug- disease in the brain, NK depletion resulted in a decrease in TCR gesting a more mature DC population as compared with the EAE V␤8.1 usage, by 100–200%. Decreases were also observed in control mice. This change was not seen in the cervical LN and TCR V␤14 and 15. However, when either cervical or draining these DC effects were lost by day 14 of treatment. By comparing LN were examined, an overall reduction in the measured TCR

FIGURE 6. Effects of NK depletion on DC. DC were enumerated from LN at specified times after immunization. A, Results of flow cytometric analysis of DC (examining CD11cϩclass IIϩ cells) evaluating the expression of CD80 and CD86 from cervical (left dot plots) or draining LN (right dot plots) at the immunization site at days 7 and 14, with or without anti-NK1.1 treatment (NK depl.). B (draining) and C (cervical) nodes were evaluated for relative expression of CD80/86bright–dim DC in mice immunized with MOG peptide with (EAE NK Dep.; Œ) or without NK depletion (EAE; ‚). n, Values are pools of three to five mice per treatment. The Journal of Immunology 4503

from NKT cells (58–60), a novel regulatory lymphocyte population that produces a large amount of IL-4 after TCR ligation (61). To overcome the problems inherent in the previous studies, we

selected a model of primary EAE induced in B6 with MOG35–55 peptide (31). Our initial studies evaluated the effects of depleting NK cells with anti-NK1.1, anti-asialo GM1, and select anti-Ly49s to alter populations of NK cells in adaptive tissue and draining LN and evaluate the effects of these treatments on the clinical course of EAE in vivo. Our studies clearly demonstrated that NK depletion before immunization diminished the onset of EAE and extent of paralysis. The use of anti-asialo GM1 and select anti-Ly49s ruled out a role for NKT cells since these cells remained unaffected by these treatments. In addition, the depletion of NK cells after immunization failed to alter clinical disease in vivo, thus supporting a regulatory role for NK cells early in immune development but not in the effector phase. Our ﬁndings that elimination of NK cells expressing Ly49D and Ly49H, a highly overlapping but mi- nor subset of NK cells, had an effect similar to that of total NK

depletion implicated this subset of the NK cells in the clinical Downloaded from response. Furthermore, when our studies were translated into an in vitro evaluation of T cell responses to cognate MOG Ag, NK- depleted lymphoid tissues demonstrated a diminished T cell response as measured by proliferation and cytokine production (IFN-␥ and TNF-␣).

Recent studies evaluating IL-17 production by T cells has im- http://www.jimmunol.org/ plicated this cytokine as a major autoimmune regulator. Komiyama et al. (39) demonstrated that IL-17Ϫ/Ϫ mice had sup- pressed development of EAE. These animals exhibited delayed onset, reduced clinical scores, altered histological changes, and early recovery. The major producer of IL-17 was CD4ϩ T cells and their data suggested that IL-17 and IFN-␥ mutually cross- regulate expression of each cytokine. These observations indicate that IL-17 plays a crucial role in the development of EAE. Fur- thermore, Carballeda et al. (40) demonstrated that mice lacking by guest on September 27, 2021 IL-23 (p19Ϫ/Ϫ) did not develop EAE. In addition, disease resis- tance by IL-23 knockout mice was associated with loss of IL-17- producing CD4ϩ T lymphocytes. These studies demonstrated that the IL-17/23 inflammation and related molecules were critical ini- FIGURE 7. Analysis of TCR usage in EAE mice depleted of NK cells. tiators of EAE and perhaps other autoimmune diseases. Suryani Lymphocytes from perfused brains (A) or cervical (B) or draining (C) et al. (41) demonstrated that MOG peptide induced an IL- ␤ ϩ ϩ ϩ nodes were removed and analyzed for TCR V usage by real-time quan- 17 IFN-␥ population of CD4 CNS-infiltrating MOG -spe- ␤ 35–55 titative PCR using primers specific for the V s specified. Mice were im- cific T cells, which is the majority population relative to IL- munized with MOG without NK1.1 depletion and values were normalized ϩ Ϫ 17 IFN-␥ cells. These studies implied that the Th1 lineage is to this treatment (error bar). Values represent change in EAE NK-depleted mice relative to EAE autoimmune mice (with SE). The NK1.1 depletion more encephalitogenic than is suggested by adoptive transfer of Ϫ ␥ϩ values (f) are shown for selected V␤s. Th1 (IL-17 /IFN- ) cells since these cells may be terminally differentiated. In the present study, the removal of NK cells did not consistently alter the IL-17 or IL-23 levels induced by MOG stim- V␤s was seen across the analysis, consistent with the lack of ulation, although a dramatic decrease in the ratio of IL-17:IFN-␥ reactivity to in vitro MOG peptide seen in these same tissues. production was observed. These findings suggest that the target of NK effects is not within the IL-17/23 regulation of effector T cells. Discussion Evaluation of the CNS leukocytes, unlike the draining LN, in- Numerous recent reviews (48–50) have postulated the concept that dicated both a qualitative alteration (TCR V␤) and a quantitative NK cells can play an important role in the regulation of adaptive change, since a reduced frequency of MOG-reactive T cells that immunity. However, to date, there are limited definitive reports emigrate to the CNS and mediate autoimmune reaction was ob- that define where and how NK cells can modulate their complex served. Analysis of TCR V␤ usage after NK cell depletion re- interactions. Perhaps the most studied and well-defined reports are vealed a qualitative shift in the T cell population because T cells the studies in CMV infection and the development of CD8 effector containing V␤8, 14, and 15 were decreased. Because these T cells T cells (1, 2). Although encephalitogenic peptides (20) or TCR have been shown to be the major effector repertoire in C57BL/6 peptides (19, 51) in association with MHC molecules are recog- mice (31), loss of this population as a result of NK cell depletion nized as the receptor ligands for some regulatory T cells, less is would explain the altered clinical outcome observed in our study. known about how other regulatory cells are triggered. Finally, the evaluation of LN DC demonstrated an alteration in In fact, early studies (44, 52) used anti-asialo GM1 polyclonal mature DC levels (based on CD80 and CD86 expression), indi- sera for NK cell deletion, which can damage macrophages (53) or cating that NK cells may prevent optimal secondary T cell re- T cells (54). Other studies (45, 55–57) did not distinguish NK cells sponses and Ag presentation. Thus, the role of NK cells may be 4504 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE nonlytic since studies with perforin knockout mice (data not NKT cells, and monocytes/macrophages, were unchanged. The shown) demonstrated no regulatory role in primary EAE induction lack of NK cells was associated with increase clinical EAE mor- and disease. Collectively, these data support the recent hypothesis tality and hemorrhagic inflammatory lesions. These changes were (62) that in addition to being antitumor and antiviral effector cells, not observed in CD1d-deficient mice, effectively ruling out a role NK cells mediate a critical link between NK cells and adaptive T for NKT cells. These results strongly suggested a critical role for and B cell responses as indicated in previous reports in the CMV CX3CR1ϩ NK cells in CNS pathology. Although our studies did infection model system (1, 63) and in the multiple sclerosis model not examine subsets of NK cells, we did observe primarily changes (25, 28, 64). in T cell numbers with intact activation status in the CNS. How- The mechanisms by which NK cells regulate these adaptive re- ever, increased NK cell numbers in the CNS compared with con- sponses are not yet fully elucidated, but their site of regulation has trol mice were observed in our studies (Table III). Thus, our stud- been clearly shown to involve the draining LN after MOG peptide ies are consistent with the report by Huang et al. (28) that NK cells immunization, an immunoregulatory site not generally thought to do increase in the EAE CNS. involve NK cells until recently. The cytokines that are critical for Overall and in the context of the studies discussed above, our the maturation of EAE responses and that determine the differen- results are consistent with recent postulations (25, 47–49) that NK tiation of T cells into MOG- specific Th1 (IL-2, IFN-␥, TNF-␣)T cells serve as an important potential link between innate and adap- cells would be expected to have a regulatory impact on numerous tive immunity. However, this link represents a very complex in- other adaptive responses. Our findings are consistent with the pro- teraction and may be highly dependent upon the model system posed hypothesis that through cytokine production, NK cells can used. Thus, further studies are required to elucidate the role of NK regulate the ability of intracellular bacteria and viruses to induce cells in innate-adaptive interactions. Downloaded from and regulate Ag-specific T cells (65, 66). Our findings are in contrast to those reported by Zhang et al. Our data suggest an important LN interaction among NK cells, (21). In that study, in vivo NK cell depletion increased EAE clin- DC, and the development of Ag-specific T and perhaps B cells. ical pathology (Fig. 1 in Ref. 21) and increased MOG-induced NK cells have been proposed to promote Th1 responses in part by proliferation and IFN-␥ production from T cells (Fig. 5 in Ref. 21). early production of IFN (67) and regulation of DC (6). These re- In addition, NK depletion did not alter monophasic EAE (Fig. 4 in

ported effects of NK cells influencing the Ag-specific autoimmune Ref. 21). In contrast to these results, our studies using anti-NK1.1, http://www.jimmunol.org/ response brings together a number of these regulatory circuits. anti-asialo GM1, and anti-Ly49 reagents all demonstrated a de- However, such observations do not exclude the possibility that crease in clinical pathology and a decreased frequency of disease other cell types or other signals from NK cells might be involved (current study, Fig. 1) as well as decreased MOG-induced prolif- in NK-DC-T cell interactions. For example, we have already eration and IFN-␥ production (current study, Fig. 3). In contrast to shown that cytokines (IL-12 and IL-18) induced by other cell types our study which only evaluated primary EAE generation, Zhang can significantly costimulate NK cells in an in vivo environment et al. used adoptive transfer models of either primary T cells or T and alter the inhibitory circuit that is mediated by NKRs (68). cell lines (Figs. 7 and 8 in Ref. 21) that mediated an EAE auto- Thus, further experimentation is required to more fully define all of immune response. A possible explanation for the differences be- the signals that regulate NK-DC interactions. tween our report and that of Zhang et al. is that in transfer studies by guest on September 27, 2021 The study of NK cell function in vivo has been challenging and it is well known that NK cells rapidly eliminate transplanted cells has resulted in contradictory results as reviewed by Shi (25). These in the circulation. Thus, some of the conclusions reached by Zhang previous studies have used anti-NK1.1 or anti-asialo GM1 (not and coworkers could have been due to the measured alterations in present on NKT cells) to selectively examine NK cells in a variety transfer efficiency. The reasons for the discrepancies in the in vivo of autoimmune models and our findings are not inconsistent with depletions are not apparent; however, there are several notable the data reported in those model systems. differences between our report and that of Zhang and coworkers. Shi (27) demonstrated in IL-18-deficient mice that immuniza- First, the clinical EAE observed in the control mice in the studies Ͻ tion with MOG35–55 failed to induced Th1 and autoantibody re- by Zhang et al., where NK depletion was performed, was 1, sponses and subsequent clinical EAE. This lack of autoimmunity which represents a very low level of disease. Our studies generally could be restored by IL-18 but was dependent on the presence of had control EAE that was 2–3 in magnitude (equivalent to 4–5 NK cells. In addition, these studies suggested a role for NK-in- based on the scale used by Zhang et al.). Second, the anti-NK1.1 duced IFN-␥ for CNS pathology. Our studies are quite consistent was used at a much higher dose by Zhang et al. (500 ␮g/mouse), with this report for a central role of NK cells in the development whereas our studies used a split dose of 100 ␮g/mouse. These split of Th1 cells and MOG-induced EAE. doses deplete NK cells more efficiently in liver and LN. Both doses In another study, Shi (26) reported that NK cells can determine removed detectable NK cells as reported but the large excess of the outcome of B cell-mediated autoimmunity. Using this myas- anti-NK1.1 might have other consequences not observed with ad- thenia gravis model, mice depleted of NK cells or NK cell-defi- ministration of the lower dose. Third, Zhang et al. used footpad cient IL-18Ϫ/Ϫ mice failed to develop Th1 responses and Abs to injection, followed by a booster immunization after 1 wk with PT the acetylcholine receptor. These results established an important being given i.v. Our studies used s.c. immunization followed by a link between NK cells and autoreactive T and B cells. Our studies 48-h boost with the PT given i.p. Fourth, our PT source was dif- are quite consistent with this report for a central role of NK cells ferent from that of Zhang et al., a variable that might change the in the development of this B cell-mediated autoimmunity. clinical score in this murine EAE model (our unpublished data). Other reports of mice lacking IL-15 or T-bet demonstrated both Fifth, our in vitro studies emphasized the LN (draining, nondrain- defective Th1 responses and autoimmunity. However, these defi- ing, and cervical) and brain lymphocyte responses to MOG pep- ciencies target multiple immune effectors and are a bit more dif- tide, whereas studies by Zhang et al. evaluated the spleen. Sixth, ficult to apply directly to NK cells (reviewed in Ref. 64). our studies evaluated the brain immunology, DC maturation status, Recent studies by Huang (28) demonstrated a link between and TCR V␤ repertoire, issues that were not examined by Zhang CX3CR1 expression and brain pathology in EAE, showing that et al. Seventh, both studies found that NK depletion after immu- NK cells were markedly reduced in the CNS of CX3CR1-deficient nization failed to have any effect on clinical disease, supporting a mice. However, the recruitment of other leukocytes, e.g., T cells, role for NK cells early in the generation of an adaptive T cell The Journal of Immunology 4505 response. Thus, the differences in the timing of immunization and 8. Playfair, J. H., and S. Marshall-Clarke. 1973. Induction of red cell autoantibodies PT routes as well as the cells evaluated may play an important role in normal mice. Nat. New Biol. 243: 213–214. 9. Radford-Smith, G. 1997. Ulcerative colitis: an immunological disease? Baillieres in the discrepancies in our results with those previously reported Clin. Gastroenterol. 11: 35–52. by Zhang et al. 10. Mountz, J. D., W. C. Gause, and R. Jonsson. 1991. Murine models for systemic Studies by Bakker et al. (69) demonstrated an important role for lupus erythematosus and Sjo¨gren’s syndrome. Curr. Opin. Rheumatol. 3: 738–756. DAP12 in autoimmune disease and concluded that an impaired Ag 11. Samuels, J., Y. S. Ng, C. Coupillaud, D. Paget, and E. Meffre. 2005. Impaired priming was the underlying mechanism responsible for the re- early B cell tolerance in patients with rheumatoid arthritis. J. Exp. Med. 201: sistance of DAP12-deficient mice to the development of EAE. 1659–1667. 12. Huang, S. W., J. W. Rose, and R. F. Mayer. 1977. Assessment of cellular and Since DAP12 is present in both APC and NK cells, these con- humoral immunity of myasthenics. J. Neurol. Neurosurg. Psychiatry 40: clusions are not in conflict with the data that we have presented 1053–1059. here because our studies clearly demonstrated that primary T cell 13. Flodstrom, M., A. Maday, D. Balakrishna, M. M. Cleary, A. Yoshimura, and N. Sarvetnick. 2002. Target cell defense prevents the development of diabetes responses are minimized and that DC are altered when NK cells after viral infection. Nat. Immunol. 3: 373–382. are depleted. In addition, the lack of DAP12 in NK cells, and as 14. Poirot, L., C. Benoist, and D. Mathis. 2004. Natural killer cells distinguish in- shown by the role of activating Ly49 cells, would significantly nocuous and destructive forms of pancreatic islet autoimmunity. Proc. Natl. Acad. Sci. USA 101: 8102–8107. alter their ability to secrete cytokines, since many cytokines are 15. Dalbeth, N., R. Gundle, R. J. Davies, Y. C. Lee, A. J. McMichael, and expressed upon triggering activating receptors (67, 68, 70). Fur- M. F. Callan. 2004. CD56bright NK cells are enriched at inflammatory sites and thermore, our data directly implicate Ly49D and Ly49H NKR sub- can engage with monocytes in a reciprocal program of activation. J. Immunol. 173: 6418–6426. sets, one of which is a potent source of NK cytokine production 16. Alvord, E. C., Jr. 1984. Experimental Allergic Encephalomyelitis: A Useful (71–73). Consistent with regulation of early Ag presentation by Model for Multiple Sclerosis. Liss, New York. Downloaded from DC, our data also implicate an early involvement of NK cells since 17. Tabira, T., and J. Kira. 1992. Strain and species differences of encephalitogenic determinants of myelin basic protein and proteolipid apoprotein. In Myelin: Bi- their depletion after the primary immunization (after day 7) did not ology and Chemistry. R. E. Martenson, ed. CRC, Boca Raton, FL, pp. 783–800. modulate EAE clinically or the T cell responses. 18. Merrill, J. E., and E. N. Benveniste. 1996. Cytokines in inflammatory brain le- In this current study, we demonstrate that NK cells, a major arm sions: helpful and harmful. Trends Neurosci. 19: 331–338. 19. Kumar, V., and E. E. Sercarz. 1993. The involvement of T cell receptor peptide- of innate immunity, participate in the development of primary specific regulatory CD4ϩ T cells in recovery from antigen-induced autoimmune

MOG-induced T cell autoimmune disease. The requirement of NK disease. J. Exp. Med. 178: 909–916. http://www.jimmunol.org/ cells was demonstrated in vivo by the reduced ability of NK or 20. Chen, Y., V. K. Kuchroo, J. Inobe, D. A. Hafler, and H. L. Weiner. 1994. Reg- ulatory T cell clones induced by oral tolerance: suppression of autoimmune en- NKR subset-depleted mice to exhibit potent clinical disease. When cephalomyelitis. Science 265: 1237–1240. studies were translated in vitro from mice depleted of their NK 21. Zhang, B., T. Yamamura, T. Kondo, M. Fujiwara, and T. Tabira. 1997. Regula- cells or a NKR subset, T lymphocytes demonstrated both a qual- tion of experimental autoimmune encephalomyelitis by natural killer (NK) cells. J. Exp. Med. 186: 1677–1687. itative and quantitative failure to exhibit a T cell anti-MOG re- 22. Matsumoto, Y., K. Kohyama, Y. Aikawa, T. Shin, Y. Kawazoe, Y. Suzuki, and sponse as compared with EAE non-NK-depleted mice. NK cells N. Tanuma. 1998. Role of natural killer cells and TCR ␥␦ T cells in acute au- were shown to regulate in the LN the immune responses that ini- toimmune encephalomyelitis. Eur. J. Immunol. 28: 1681–1688. 23. Fort, M. M., M. W. Leach, and D. M. Rennick. 1998. A role for NK cells as tiate the autoimmune disease, since the absence of the appropriate regulators of CD4ϩ T cells in a transfer model of colitis. J. Immunol. 161: NK cells resulted in decreased production of inflammatory cyto- 3256–3261. by guest on September 27, 2021 kines (TNF-␣ and IL-6) and a decrease in the growth-promoting 24. Lehuen, A., O. Lantz, L. Beaudoin, V. Laloux, C. Carnaud, A. Bendelac, J. F. Bach, and R. C. Monteiro. 1998. Overexpression of natural killer T cells cytokine IL-2. protects V␣14-J␣281 transgenic nonobese diabetic mice against diabetes. J. Exp. In summary, the current data indicate that NK cells function in Med. 188: 1831–1839. autoimmune pathology and that the alteration of the immune status 25. Shi, F., H. G. Ljunggren, and N. Sarvetnick. 2001. Innate immunity and autoimmunity: from self-protection to self-destruction. Trends Immunol. 22: 97–101. of NK cells that either activate or inhibit their function would 26. Shi, F. D., H. B. Wang, H. Li, S. Hong, M. Taniguchi, H. Link, L. Van Kaer, and promote or interfere with normal innate homeostasis. Just as NK H. G. Ljunggren. 2000. Natural killer cells determine the outcome of B cell- cells have been proposed to be a first line of defense against tumor mediated autoimmunity. Nat. Immunol. 1: 245–251. 27. Shi, F. D., K. Takeda, S. Akira, N. Sarvetnick, and H. G. Ljunggren. 2000. IL-18 cell progression, this study indicates that NK cells can contribute directs autoreactive T cells and promotes autodestruction in the central nervous to the quality and quantity of adaptive T and B cell responses. system via induction of IFN-␥ by NK cells. J. Immunol. 165: 3099–3104. 28. Huang, D., F. D. Shi, S. Jung, G. C. Pien, J. Wang, T. P. Salazar-Mather, T. T. He, J. T. Weaver, H. G. Ljunggren, C. A. Biron, et al. 2006. The neuronal Disclosures chemokine CX3CL1/fractalkine selectively recruits NK cells that modify exper- The authors have no financial conflict of interest. imental autoimmune encephalomyelitis within the central nervous system. FASEB J. 20: 896–905. 29. Ortaldo, J. R., R. Winkler-Pickett, and G. Wiegand. 2000. Activating Ly-49D NK References receptors: expression and function in relation to ontogeny and Ly-49 inhibitor 1. Biron, C. A., K. B. Nguyen, and G. C. Pien. 2002. Innate immune responses to receptors. J. Leukocyte Biol. 68: 748–756. LCMV infections: natural killer cells and cytokines. Curr. Top. Microbiol. Im- 30. Gussoni, E., M. A. Panzara, and L. Steinman. 2003. Evaluating human T cell munol. 263: 7–27. receptor gene expression by PCR. In Current Protocols in Immunology. 2. Tay, C. H., E. Szomolanyi-Tsuda, and R. M. Welsh. 1998. Control of infections J. Coligan, Ed. Wiley, New York, pp. 10.26.1–10.26.14. by NK cells. Curr. Top. Microbiol. Immunol. 230: 193–220. 31. Mendel, I., N. Kerlero de Rosbo, and A. Ben Nun. 1995. A myelin oligodendro- 3. Goldszmid, R. S., A. Bafica, D. Jankovic, C. G. Feng, P. Caspar, cyte glycoprotein peptide induces typical chronic experimental autoimmune en- b ␤ R. Winkler-Pickett, G. Trinchieri, and A. Sher. 2007. TAP-1 indirectly regulates cephalomyelitis in H-2 mice: fine specificity and T cell receptor V expression CD4ϩ T cell priming in Toxoplasma gondii infection by controlling NK cell of encephalitogenic T cells. Eur. J. Immunol. 25: 1951–1959. IFN-␥ production. J. Exp. Med. 204: 2591–2602. 32. Koller, B. H., P. Marrack, J. W. Kappler, and O. Smithies. 1990. Normal devel- ␤ ϩ 4. Moretta, A., E. Marcenaro, S. Parolini, G. Ferlazzo, and L. Moretta. 2008. NK opment of mice deficient in 2m, MHC class I proteins, and CD8 T cells. cells at the interface between innate and adaptive immunity. Cell Death Differ. Science 248: 1227–1230. 15: 226–233. 33. Shinkai, Y., G. Rathbun, K. P. Lam, E. M. Oltz, V. Stewart, M. Mendelsohn, 5. Walzer, T., M. Dalod, E. Vivier, and L. Zitvogel. 2005. Natural killer cell-den- J. Charron, M. Datta, F. Young, A. M. Stall, and F. W. Alt. 1992. RAG-2- dritic cell crosstalk in the initiation of immune responses. Expert Opin. Biol. deficient mice lack mature lymphocytes owing to inability to initiate V(D)J re- Ther. 5(Suppl. 1): S49–S59. arrangement. Cell 68: 855–867. 6. Makarenkova, V., A. K. Chakrabarti, J. A. Liberatore, P. Popovic, G. Lu, 34. Dalton, D. K., S. Pitts-Meek, S. Keshav, I. S. Figari, A. Bradley, and S. Watkins, and N. L. Vujanovic. 2005. Dendritic cells and natural killer cells T. A. Stewart. 1993. Multiple defects of immune cell function in mice with interact via multiple TNF family molecules. J. Leukocyte Biol. 77: 408–413. disrupted interferon-␥ genes. Science 259: 1739–1742. 7. Mor, F., F. Quintana, A. Mimran, and I. R. Cohen. 2003. Autoimmune enceph- 35. Fritz, R. B., and M. L. Zhao. 2001. Regulation of experimental autoimmune alomyelitis and uveitis induced by T cell immunity to self ␤-synuclein. J. Im- encephalomyelitis in the C57BL/6J mouse by NK1.1ϩ, DX5ϩ, ␣␤ϩ T cells. munol. 170: 628–634. J. Immunol. 166: 4209–4215. 4506 A ROLE FOR NK CELLS IN THE DEVELOPMENT OF EAE

36. Xu, W., G. Fazekas, H. Hara, and T. Tabira. 2005. Mechanism of natural killer 53. Yohe, H. C., S. Ye, B. B. Reinhold, and V. N. Reinhold. 1997. Structural char- (NK) cell regulatory role in experimental autoimmune encephalomyelitis. J. Neu- acterization of the disialogangliosides of murine peritoneal macrophages. Glyco- roimmunol. 163: 24–30. biology 7: 1215–1227. 37. Ortaldo, J. R., L. H. Mason, T. A. Gregorio, J. Stoll, and R. T. Winkler-Pickett. 54. Kosaka, A., U. Lee, D. Wakita, N. Matsubara, Y. Togashi, S. Nishimura, ϩ 1997. The Ly-49 family: regulation of cytokine production in murine NK cells. H. Kitamura, and T. Nishimura. 2006. Interleukin-12-responding asialoGM1 ϩ J. Leukocyte Biol. 62: 381–388. CD8 central memory-type T cells as precursor cells for interferon-␥-producing 38. Ortaldo, J. R., R. Winkler-Pickett, J. Willette-Brown, R. L. Wange, killer T cells. Cancer Sci. 97: 1236–1241. S. K. Anderson, G. J. Palumbo, L. H. Mason, and D. W. McVicar. 1999. Struc- 55. Takeda, K., and G. Dennert. 1993. The development of autoimmunity in ture/function relationship of activating Ly-49D and inhibitory Ly-49G2 NK re- C57BL/6 lpr mice correlates with the disappearance of natural killer type 1-pos- ceptors. J. Immunol. 163: 5269–5277. itive cells: evidence for their suppressive action on bone marrow stem cell pro- 39. Komiyama, Y., S. Nakae, T. Matsuki, A. Nambu, H. Ishigame, S. Kakuta, liferation, B cell immunoglobulin secretion, and autoimmune symptoms. J. Exp. K. Sudo, and Y. Iwakura. 2006. IL-17 plays an important role in the development Med. 177: 155–164. of experimental autoimmune encephalomyelitis. J. Immunol. 177: 566–573. 56. Orange, J. S., B. Wang, C. Terhorst, and C. A. Biron. 1995. Requirement for ␥ 40. Furuzawa-Carballeda, J., M. I. Vargas-Rojas, and A. R. Cabral. 2007. Autoim- natural killer cell-produced interferon in defense against murine cytomegalo- mune inflammation from the Th17 perspective. Autoimmun. Rev. 6: 169–175. virus infection and enhancement of this defense pathway by interleukin 12 ad- 41. Suryani, S., and I. Sutton. 2007. An interferon-␥-producing Th1 subset is the ministration. J. Exp. Med. 182: 1045–1056. major source of IL-17 in experimental autoimmune encephalitis. J. Neuroimmu- 57. Wilder, J. A., C. Y. Koh, and D. Yuan. 1996. The role of NK cells during in vivo nol. 183: 96–103. antigen-specific antibody responses. J. Immunol. 156: 146–152. 42. Martin-Fontecha, A., L. L. Thomsen, S. Brett, C. Gerard, M. Lipp, 58. Bix, M., and R. M. Locksley. 1995. Natural T cells: cells that co-express NKRP-1 and TCR. J. Immunol. 155: 1020–1022. A. Lanzavecchia, and F. Sallusto. 2004. Induced recruitment of NK cells to ϩ ␥ 59. Bendelac, A. 1995. Mouse NK1 T cells. Curr. Opin. Immunol. 7: 367–374. lymph nodes provides IFN- for TH1 priming. Nat. Immunol. 5: 1260–1265. 43. Mendel, I., N. Kerlero de Rosbo, and A. Ben-Nun. 1995. A myelin oligodendro- 60. Taniguchi, M., Y. Makino, J. Cui, K. Masuda, T. Kawano, H. Sato, E. Kondo, ␣14ϩ cyte glycoprotein peptide induces typical chronic experimental autoimmune en- and H. Koseki. 1996. V NK T cells: a novel lymphoid cell lineage with regulatory function. J. Allergy Clin. Immunol. 98: S263–S269. cephalomyelitis in H-2b mice: fine specificity and T cell receptor V ␤ expression 61. Yoshimoto, T., and W. E. Paul. 1994. CD4pos, NK1.1pos T cells promptly pro-

of encephalitogenic T cells. Eur. J. Immunol. 25: 1951–1959. Downloaded from duce interleukin 4 in response to in vivo challenge with anti-CD3. J. Exp. Med. 44. Mendel, I., N. Kerlero de Rosbo, and A. Ben-Nun. 1996. Delineation of the 179: 1285–1295. minimal encephalitogenic epitope within the immunodominant region of myelin 62. Farag, S. S., J. B. VanDeusen, T. A. Fehniger, and M. A. Caligiuri. 2003. Biology oligodendrocyte glycoprotein: diverse V ␤ gene usage by T cells recognizing the and clinical impact of human natural killer cells. Int. J. Hematol. 78: 7–17. core epitope encephalitogenic for T cell receptor V ␤ and T cell receptor V ␤ 63. Tay, C. H., L. Y. Yu, V. Kumar, L. Mason, J. R. Ortaldo, and R. M. Welsh. 1999. H-2b mice. Eur. J. Immunol. 26: 2470–2479. The role of LY49 NK cell subsets in the regulation of murine cytomegalovirus 45. Zhong, M. C., L. Cohen, A. Meshorer, N. Kerlero de Rosbo, and A. Ben-Nun. infections. J. Immunol. 162: 718–726. 2000. T cells speciﬁc for soluble recombinant oligodendrocyte-speciﬁc protein 64. Shi, F. D., and K. L. Van. 2006. Reciprocal regulation between natural killer cells induce severe clinical experimental autoimmune encephalomyelitis in H-2(b) and

and autoreactive T cells. Nat. Rev. Immunol. 6: 751–760. http://www.jimmunol.org/ H-2(s) mice. J. Neuroimmunol. 105: 39–45. 65. Romagnani, S. 1992. Induction of TH1 and TH2 responses: a key role for the 46. Mendel, I., A. Katz, N. Kozak, A. Ben-Nun, and M. Revel. 1998. Interleukin-6 “natural” immune response? Immunol. Today 13: 379–381. functions in autoimmune encephalomyelitis: a study in gene-targeted mice. Eur. 66. Fearon, D. T., and R. M. Locksley. 1996. The instructive role of innate immunity J. Immunol. 28: 1727–1737. in the acquired immune response. Science 272: 50–53. 47. Mendel, I., N. Kerlero de Rosbo, and A. Ben-Nun. 1996. Delineation of the 67. Trinchieri, G. 1989. Biology of natural killer cells. Adv. Immunol. 47: 187–376. minimal encephalitogenic epitope within the immunodominant region of myelin 68. Ortaldo, J. R., and H. A. Young. 2003. Expression of IFN-␥ upon triggering of ␤ oligodendrocyte glycoprotein: diverse V gene usage by T cells recognizing the activating Ly49D NK receptors in vitro and in vivo: costimulation with IL-12 or ␤ ␤ core epitope encephalitogenic for T cell receptor V b and T cell receptor V and J. Immunol. b IL-18 overrides inhibitory receptors. 170: 1763–1769. H-2 mice. Eur. J. Immunol. 26: 2470–2479. 69. Bakker, A. B., R. M. Hoek, A. Cerwenka, B. Blom, L. Lucian, T. McNeil, 48. Raulet, D. H. 2004. Interplay of natural killer cells and their receptors with the R. Murray, L. H. Phillips, J. D. Sedgwick, and L. L. Lanier. 2000. DAP12- adaptive immune response. Nat. Immunol. 5: 996–1002. deﬁcient mice fail to develop autoimmunity due to impaired antigen priming.

49. Yokoyama, W. M. 2005. Specific and non-specific natural killer cell responses to Immunity 13: 345–353. by guest on September 27, 2021 viral infection. Adv. Exp. Med. Biol. 560: 57–61. 70. Ortaldo, J. R., and R. B. Herberman. 1984. Heterogeneity of natural killer cells. 50. Zingoni, A., T. Sornasse, B. G. Cocks, Y. Tanaka, A. Santoni, and L. L. Lanier. Annu. Rev. Immunol. 2: 359–394. 2005. NK cell regulation of T cell-mediated responses. Mol. Immunol. 42: 71. Smith, K. A., J. Wu, A. B. H. Bakker, J. H. Phillips, and L. L. Lanier. 1998. 451–454. Ly49D and Ly49H associate with mouse DAP12 and form activating receptors. 51. Kozovska, M. F., T. Yamamura, and T. Tabira. 1996. T-T cellular interaction J. Immunol. 161: 7–10. Ϫ Ϫ between CD4 CD8 regulatory T cells and T cell clones presenting TCR pep- 72. Dokun, A. O., S. Kim, H. R. Smith, H. S. Kang, D. T. Chu, and tide: its implication for TCR vaccination against experimental autoimmune en- W. M. Yokoyama. 2001. Specific and nonspecific NK cell activation during virus cephalomyelitis. J. Immunol. 157: 1781–1790. infection. Nat. Immunol. 2: 951–956. 52. Gilbertson, S. M., P. D. Shah, and D. A. Rowley. 1986. NK cells suppress the 73. Ortaldo, J. R., E. W. Bere, D. Hodge, and H. A. Young. 2001. Activating Ly-49 generation of Lyt-2ϩ cytolytic T cells by suppressing or eliminating dendritic NK receptors: central role in cytokine and chemokine production. J. Immunol. cells. J. Immunol. 136: 3567–3571. 166: 4994–4999.