ARTHRITIS & RHEUMATISMVol. 63, No. 6, June 2011, pp 1562–1572DOI 10.1002/art.30328© 2011, American College of Rheumatology

Complement Regulatory Protein Crry/p65 CostimulationExpands Natural Treg Cells With Enhanced Suppressive

Properties in Proteoglycan-Induced Arthritis

Gloria Ojeda,1 Eliana Pini,1 Cesar Eguiluz,1 Marıa Montes-Casado,1 Femke Broere,2

Willem van Eden,2 Jose M. Rojo,3 and Pilar Portoles4

Objective. To investigate the costimulatory role ofCrry/p65 (Crry), a membrane complement regulatoryprotein, on the expansion and function of natural Tregcells and their ability to ameliorate proteoglycan-induced arthritis (PGIA), an animal model of inflam-matory arthritis in which the role of natural Treg cellsis not well established.

Methods. CD4�CD25� natural Treg cells fromBALB/c mice were activated in vitro and costimulated byCrry. The expanded cells were phenotypically charac-terized, and their suppressive effect on T cell prolifera-tion was assayed in vitro. The potential prophylactic andtherapeutic effects of this population versus those ofnatural Treg cells in PGIA were studied. The clinicalscore, histology, the antigen-specific isotype antibodypattern, in vitro T cell responses, and the presence ofTreg cells in the paws were studied.

Results. Crry costimulation enhanced the in vitroexpansion of natural Treg cells while maintaining theirphenotypic and suppressive properties. Crry-expanded

Treg cells had stronger suppressive properties in vivoand a longer ameliorating effect in the PGIA model thandid natural Treg cells. Crry-expanded Treg cells sup-pressed T cell– and B cell–dependent responses inPGIA, changing the pathogenic antibody isotype patternand decreasing antigen-dependent secretion of cyto-kines, including interferon-�, interleukin-12 (IL-12),and IL-17. Increased FoxP3 expression was detected inthe paws of mice transferred with Crry-expanded Tregcells.

Conclusion. Crry-mediated costimulation facili-tates in vitro expansion of natural Treg cells whilemaintaining their suppressive properties in vitro and invivo in the PGIA model. These results highlight thepotential of the complement regulatory protein Crry tocostimulate and expand natural Treg cells capable ofsuppressing disease in an animal model of chronicinflammatory arthritis.

Thymus-derived CD4�CD25�FoxP3� naturalTreg cells are important for self-tolerance and preven-tion of autoimmunity (1,2). However, the therapeuticapplication of Treg cells poses important problems,including the absence of specific, exclusive surface mark-ers allowing their isolation by standard methods; the lownumber of Treg cells in the lymphoid organs or blood ofhealthy individuals; the lack of definitive data concern-ing the growth factors and stimuli specifically controllingexpansion of Treg cells; and their own anergic nature,proliferating with difficulty in response to antigenicstimuli in vitro. In vitro, Treg cell–mediated suppressionrequires a signal through the T cell receptor, but onceactivated, suppression is nonspecific. Attempts havebeen made to expand natural Treg cells in vitro, usingdifferent stimuli and costimuli, without abrogating theirsuppressive activity (3–5).

Supported by the Fondo de Investigaciones Sanitarias (grantsPI070620, PI070484, and PI10/00648), the Ministerio de Educacion yCiencia (grant SAF-2004-06852), and the Fundacion Genoma Espana.Drs. Broere and van Eden’s work was supported by the Dutch ArthritisAssociation and the European Commission Seventh Framework Pro-gramme (TOLERAGE project grant 202156).

1Gloria Ojeda, PhD, Eliana Pini, PhD, Cesar Eguiluz, DVM,MSc, Marıa Montes-Casado, MSc: Instituto de Salud Carlos III,Madrid, Spain; 2Femke Broere, PhD, Willem van Eden, MD, PhD:Utrecht University, Utrecht, The Netherlands; 3Jose M. Rojo, PhD:Consejo Superior de Investigaciones Cientıficas, Madrid, Spain; 4PilarPortoles, PhD: Instituto de Salud Carlos III and Consejo Superior deInvestigaciones Cientıficas, Madrid, Spain.

Drs. Ojeda and Pini contributed equally to this work.Address correspondence to Pilar Portoles, PhD, Unidad de

Inmunologıa Celular, Centro Nacional de Microbiologıa, Instituto deSalud Carlos III, Ctra. Majadahonda-Pozuelo km. 2, Majadahonda,28220 Madrid, Spain. E-mail: pportols@isciii.es.

Submitted for publication June 2, 2010; accepted in revisedform February 24, 2011.

1562

Proteoglycan-induced arthritis (PGIA) is an ani-mal model of a systemic and progressive autoimmunedisease showing similarities to rheumatoid arthritis(RA) in genetics, clinical appearance, histopathology,immune regulation, inflammatory cell migration, andthe greater susceptibility of aging female mice (6,7).Several studies in patients with RA showed T cell and Bcell responses against human PG, suggesting that PGmay be a target of disease-associated T cell responses inpatients with RA (8–10). PGIA is induced in BALB/cmice by immunization with human cartilage PG, andthe development of disease is based on cross-reactivitybetween the immunizing human and mouse (self)cartilage PGs (11,12). The pathogenesis involves bothT cell and B cell responses, suggesting a breakdown ofperipheral tolerance and accumulation of autoreactiveT cells. Although there is growing evidence on theimplication of different CD4� T cell subpopulations inthe pathogenesis and control of PGIA (11,13–16), stud-ies of the role of natural Treg cells in this model arescarce (17,18). Because one major problem of usingnatural Treg cells in therapy is the anergic state of thecells, which precludes their in vitro expansion, we ex-plored the use of Crry as a costimulus to facilitate theirproliferation.

Crry is a complement regulatory molecule ex-pressed on the surface of most cells in the mouse.Previously, we demonstrated that P3D2 anti-Crry anti-bodies induce intracellular signals that increase T lym-phocyte proliferation in vitro while modifying interleu-kin secretion to favor a Th2 phenotype (19,20). In thecurrent study, we used Crry costimulation to expandhighly pure natural Treg cell populations in vitro. Theresulting population maintained high expression ofFoxP3 and efficiently suppressed CD4�CD25� cellproliferation in vitro. In PGIA, Crry-costimulated natu-ral Treg cells decreased the level of pathogenic anti-PGantibodies or antigen-dependent secretion of inflamma-tory cytokines. Furthermore, Crry-costimulated Tregcells accumulated differently in lymphoid organs and atthe site of inflammation and could suppress ongoingPGIA. Our results highlight the use of in vitro costimu-lation to facilitate the therapeutic potential of naturalTreg cells.

MATERIALS AND METHODS

Mice and antigens. BALB/c mice were bred understandard conditions in the animal facility of the Instituto deSalud Carlos III, from stock purchased from Charles River.Retired female BALB/c breeder mice were obtained fromCharles River and used for the PGIA experiments. They were

kept at the animal facility of the Instituto de Salud Carlos IIIunder standard conditions in filter-topped cages.

Human PG (aggrecan) was purified from human artic-ular cartilage as described elsewhere (14). All experimentalprocedures were performed according to established institu-tional and national guidelines.

Cell separation. CD4�CD25� cells were isolatedfrom pooled spleen cells from female BALB/c mice ages 8–14weeks, using the CD4�CD25� Regulatory T Cell IsolationKit (Miltenyi Biotech) according to the manufacturer’s instruc-tions, yielding a population of �80% CD4�CD25� cells, asdetermined by flow cytometry. Before in vitro culture, cellswere further purified by flow cytometry (fluorescence-activated cell sorting) in a FACSAria sorter (BD Biosciences),yielding a population of �95% CD4�CD25� cells aftersorting.

Cell culture, activation, and in vitro expansion.CD4�CD25� or CD4�CD25� T cells (106/well) in 1 ml ofClick’s medium (21) supplemented with 10% heat-inactivatedfetal calf serum (FCS) and 1,000 units/ml of recombinanthuman interleukin-2 (rhIL-2; Inmunotools) were activated in24-well culture plates (Costar) previously coated with anti-CD3 plus costimulatory anti-Crry P3D2 (20 �g/ml) orcontrol antibodies, as previously described (19). Anti-CD3(YCD3-1 [22]) was used at a concentration of 20 �g/ml toactivate CD4�CD25� cells or at 5 �g/ml to activateCD4�CD25� cells. Cells were activated for 4 days, resus-pended, washed, and expanded at 2 � 105 cells/ml for a further3 days in culture medium supplemented with 1,000 units/ml ofrhIL-2.

Intracellular staining. The intracellular IL-4, IL-10,and interferon-� (IFN�) content was determined by flowcytometry with a FACSCalibur (BD Biosciences) as describedpreviously (23), using specific antibodies purchased fromeBioscience. FoxP3 was stained using the Mouse FoxP3 Kit(eBioscience), according to the manufacturer’s instructions.

In vitro Treg cell assay. Treg cell assays were per-formed in 0.2 ml of culture medium in U-bottomed 96-wellplates (Costar) as previously described (23). CD4�CD25�responder T cells (5 � 104) were cultured in the presence ofanti-CD3 (YCD3-1; 5 �g/ml) and 105 mitomycin C–treated,T cell–depleted spleen cells, plus different amounts of theputative Treg cells. Proliferation was assessed at 72 hours, aspreviously described (24).

Induction and assessment of arthritis, and experimen-tal groups. Arthritis was induced by intraperitoneal injectionof 0.4 mg of purified and deglycosylated human PG emulsifiedin 2 mg of the synthetic adjuvant dimethyldioctadecylammo-nium bromide (DDA; Sigma) in 0.2 ml phosphate bufferedsaline (PBS) on day 0 and day 21, as described elsewhere(14,25). On day 20, mice were injected intravenously with 2 �106 freshly obtained CD4�CD25� natural Treg cells or withCrry-costimulated and expanded CD4�CD25� cells (Crry-expanded Treg cells); as a control, spleen cells were injected.Alternatively, cells were injected into mice on day 35 (day 14after the second immunization), when arthritis symptoms werealready clear. Each group included 5–7 mice.

Paws were examined in a blinded manner 3 timesweekly, to determine the severity of arthritis. Arthritis severitywas scored on a 0–4-point visual analog scale as follows: 0 � noswelling or redness, 1 � mild swelling and redness in the

Crry/p65 COSTIMULATION OF NATURAL Treg CELLS 1563

wrist/ankle, 2 � severe swelling and redness in the wrist/ankle,3 � very severe swelling and redness in the wrist/ankle (withswelling and redness also developing in the toes), and 4 � verysevere swelling and redness in the wrist/ankle and in �2 toesand ankylosis.

The degree of joint swelling for each paw (scored from0 to 4) was assessed and expressed as the cumulative arthritisseverity score for 4 paws, with a maximum possible score of16 per mouse.

Histologic assessment. The mouse limbs were dis-sected, fixed in 10% buffered formalin, decalcified in 10%formic acid, and embedded in paraffin. Paraffin sections(5 �m) were stained with hematoxylin and eosin and examinedfor joint histopathology.

Ex vivo antigenic stimulation of cells. Single-cell sus-pension of spleen cells (2 � 105/well) or axillary, inguinal, andpopliteal lymph node cells (4 � 105/well) from human PG–immunized mice were cultured in triplicate in flat-bottomed96-well tissue culture plates (Costar), in the presence orabsence of human PG (10–50 �g/ml) in 0.2 ml of Click’smedium. After 72 hours, 100 �l of supernatant was obtainedfrom each well for cytokine analysis.

Cytokine quantification by enzyme-linked immunosor-bent assay (ELISA). Cytokines were determined in culturesupernatants by ELISA, using the RayBiotech Mouse Cyto-kine Antibody Array I, according to the manufacturer’s in-structions. Slides were read in a GenePix scanner (AxonInstruments), and analysis of the results was performed withthe Q-Analyzer application (RayBiotech).

Quantification of PG-specific antibodies. PG-specificantibodies were measured by ELISA. Briefly, 96-well plates(Costar) were coated overnight with 2 �g/ml of human PG in0.1M phosphate buffer, pH 9.2, and blocked with PBS/0.5%gelatin. Serial dilutions of the sera in PBS/0.5% gelatin wereincubated for 1 hour. After washing, horseradish peroxidase(HRP)–coupled antibodies against mouse IgG1 or IgG2a(Southern Biotechnology) were added and eventually devel-oped with SigmaFast OPD (o-phenylenediamine) HRP sub-strate. The serum titer was calculated as the inverse of thedilution giving 50% of maximal optical density.

Labeling with 5,6-carboxyfluorescein succinimidyl es-ter (CFSE). Cells to be transferred (freshly obtainedCD4�CD25� natural Treg cells, Crry-expanded Treg cells,or control CD4�CD25� cells [7 � 106/ml]) were incubatedfor 10 minutes at 37°C with CFSE (Vybrant CFDA SE CellTracer kit; Molecular Probes), 3 �M in PBS/5% FCS. Cellswere washed once in PBS/5% FCS and incubated again for5 minutes at 37°C. The reaction was stopped by the addition ofcold apyrogenic PBS (Sigma). The cells were thoroughlywashed before being transferred intravenously (2 � 106 cells/mouse).

Analysis of messenger RNA (mRNA) expression byreal-time quantitative polymerase chain reaction (qPCR).Total mRNA was extracted from paw tissue with the RNeasyFFPE (formalin-fixed, paraffin-embedded) Kit (Qiagen).On-column DNase treatment (Qiagen) and transcriptioninto complementary DNA (cDNA) using the iScript cDNASynthesis Kit (Bio-Rad) were carried out according tothe manufacturers’ protocols. Real-time qPCR was per-formed in a Bio-Rad MyiQ iCycler, as previously described(14). For each sample, mRNA expression was normalized to

the detected Ct value of hypoxanthine guanine phosphoribo-syltransferase.

Statistical analysis. Data are presented as the mean �SEM, as indicated. Statistical analyses were carried out withthe nonparametric Mann-Whitney U test, using GraphPadPrism version 4 software. P values less than or equal to 0.05were considered significant.

Figure 1. Crry costimulation fosters expansion of CD4�CD25� cellswhile increasing FoxP3 expression. A, Proliferation of purifiedCD4�CD25� or CD4�CD25� cells stimulated by plastic-boundanti-CD3 in the presence of interleukin-2 (IL-2) plus costimulatory(anti-CD28, anti-Crry) or control antibodies. Bars show the mean �SEM. � � P � 0.05, costimulated cultures versus anti-CD3 activation.B and C, CD25 and FoxP3 expression in CD4�CD25� cells (B) andCD4�CD25� cells (C), before (�) and after expansion in vitro.Numbers represent the percentage of cells in each quadrant. Thefluorescence intensity (FI) of CD25 (B) or both CD25 and FoxP3 (C)is shown in parentheses.

1564 OJEDA ET AL

RESULTS

Effect of Crry costimulation on facilitating invitro expansion of CD4�CD25� Treg cells while main-taining their suppressive properties. We previouslydemonstrated that costimulation of mouse CD4� T cellswith the anti-Crry antibody P3D2 enhances proliferationinduced by anti-CD3 and promotes a Th2 cytokineprofile (19). To analyze whether Crry could also en-hance the proliferation of CD4�CD25� natural Tregcells, purified CD4�CD25� cells were activated byplate-bound anti-CD3 in the presence or absence ofdifferent costimuli plus high doses of rhIL-2. Underthese conditions, the growth of CD4�CD25� cells waslower than that of CD4�CD25� cells, which was ex-pected because of the anergic nature of the expandedcells (Figure 1A). Anti-CD3 alone did not significantlyincrease cell proliferation after 7 days; however, cellproliferation was increased by costimulation with Crryand CD28 (Figure 1A).

CD4�CD25� cells expanded in vitro were as-sayed for their suppressive ability, as compared with thatof freshly obtained natural Treg cells. The freshly ob-tained CD4�CD25� Treg cell population suppressedup to 90% of the proliferation of CD4�CD25� responder

cells activated by anti-CD3 antibody and antigen-presenting cells. CD4�CD25� Treg cells, but notCD4�CD25� cells, expanded in vitro during 7 daysshowed an increased suppressive ability, strongly inhib-iting the proliferation of responder T cells (additionalinformation is available from the corresponding author).No clear differences in suppression ability were ob-served when CD4�CD25� cells expanded under differ-ent costimuli were used (additional information is avail-able from the corresponding author).

The expression of FoxP3 as a Treg cell marker inthe different populations used as suppressor cells wasalso analyzed. CD4�CD25� natural Treg cells expressedhigh levels of FoxP3 on a per-cell basis, which wereincreased upon expansion with CD3 stimulus and espe-cially with Crry costimulation (Figure 1C). In contrast,most CD4�CD25� cells expanded for 7 days (�97%)did not express FoxP3, even though they becameCD25� upon activation (Figure 1B). Crry costimulationalso increased surface CD25 expression in expandedCD4�CD25� cells and CD25� cells, as shown inFigures 1B and C. This effect was also observed incultures of total CD4� cells at the level of CD25 mRNAand protein (Portoles P: unpublished observations).

Figure 2. Crry costimulation results in different modification of cytokine expression in CD4�CD25� cells and CD4�CD25� cells. Intracellularexpression of interleukin-4 (IL-4), IL-10, and interferon-� (IFN�) in CD4�CD25� cells (A) and CD4�CD25� cells (B), before (�) and afterexpansion, in the presence or absence of costimulation with Crry is shown. Numbers represent the percentage of cells in each quadrant. See Materialsand Methods for details.

Crry/p65 COSTIMULATION OF NATURAL Treg CELLS 1565

Crry-mediated differential cytokine induction inCD4�CD25� cells and CD4�CD25� cells. We evalu-ated Crry-induced differences in cytokine expressionbetween CD4�CD25� cells and CD4�CD25� cellsfrom BALB/c mice. Neither CD3-dependent activationof CD4�CD25� Treg cells nor costimulation with Crrysignificantly changed the expression of IL-4 or IL-10(Figure 2B). However, and in spite of their origin inBALB/c mice (which is a Th2 cell–prone strain [12,26]),natural CD4�CD25� Treg cells activated by CD3 hadincreased IFN� expression.

In contrast, CD3-dependent activation and ex-pansion of CD4�CD25� cells from BALB/c mice in-creased the expression of intracellular IL-4 and antiin-flammatory IL-10 or both (Figure 2A). Crrycostimulation augmented the number of cells expressingthese cytokines, particularly IL-4–positive and IL-4–positive/IL-10–positive cells. The percentage of IL-4–negative/IFN�-positive cells was not significantly modi-fied upon CD3 activation of CD4�CD25� cells in theseTh2 cell–prone mice (26,27), and Crry costimulation didnot increase IFN� expression, as previously describedfor CD4� cells in C3H mice (19).

These results show that Crry differentially mod-ified the cytokine profile of CD4�CD25� cells andCD4�CD25� cells; this was not linked to differences insurface Crry levels in either subpopulation (additionalinformation is available from the corresponding author).

Role of Crry-expanded CD4�CD25� Treg cellsin the suppression of PGIA. PGIA is an animal model ofchronic/relapsing RA, induced by PG immunization ofBALB/c mice. B cell and T cell responses against jointcartilage PG are implicated, but a role for naturalCD4�CD25� Treg cells is not yet well known (17,18).Because Crry costimulation facilitates the in vitro expan-sion of these cells and increases CD25 and FoxP3expression as well as their suppressive properties, weexamined the role of natural Treg cells and Crry-expanded Treg cells in the PGIA model.

BALB/c mice were immunized with human PGplus DDA adjuvant; 1 day before the second immuniza-tion, 2 groups of mice received intravenous injections of2 � 106 freshly obtained natural CD4�CD25� Tregcells or CD4�CD25� Treg cells that were Crry-costimulated in vitro and expanded for 7 days; thecontrol group received normal spleen cells. As shown inFigure 3A, the development of clinical symptoms ofPGIA was delayed by the transfer of freshly obtainednatural Treg cells, but later these mice reached diseaselevels similar to those of control mice. In contrast,administration of Crry-expanded Treg cells significantlydiminished the onset and severity of arthritis in the

treated mice, showing an effect that continued for morethan 35 days after the second immunization. Histologic

Figure 3. Transfer of Crry-costimulated in vitro–expandedCD4�CD25� Treg cells delays the onset and decreases the severity ofproteoglycan-induced arthritis. A, Arthritis scores in mice that receivednormal spleen cells (control), freshly obtained purified CD4�CD25�natural Treg cells, or Treg cells that were costimulated with Crry andexpanded for 7 days, 1 day before administration of the second im-munization with human proteoglycan (hPG) (arrow). Values are themean � SEM. � � P � 0.05, Treg cell treatments versus control. B,Hematoxylin and eosin–-stained tissue sections from the joints ofcontrol and treated mice, obtained 10 days and 30 days after adminis-tration of the second immunization. Arrows indicate areas of bone ero-sion, infiltrates and synovial hyperplasia, and synovial cavities contain-ing inflammatory cells. Exp � expanded. Original magnification � 4.

1566 OJEDA ET AL

analysis of joints from control and treated mice on day10 after the second immunization (Figure 3B) showedthat inflammatory infiltrates and tissue damage werealmost absent in Crry-expanded Treg cell–treated miceas well as in the natural Treg cell–treated mice; however,on day 30 after the second immunization, infiltrates andjoint damage clearly appeared in the latter group.

To analyze the putative therapeutic effect ofCrry-expanded Treg cells, we adoptively transferredcells into mice 14 days after the second human PGimmunization, when signs of clinical inflammation werealready detectable. Figure 4 shows that clinical symp-toms of arthritis were diminished in mice receivingCrry-expanded Treg cells, even when a strong inflam-matory response was being produced; the effect ofCrry-expanded Treg cells was clearly stronger than thatof freshly obtained natural Treg cells.

Crry-expanded Treg cell transfer suppressesboth T cell and B cell responses in PGIA. On day 10after the second immunization, half of the mice in eachgroup were killed and analyzed for PG-specific antibod-ies. Figure 5A shows that the levels of anti–human PGantibodies of the IgG1 and IgG2a isotypes were clearlydiminished in the sera of mice treated with Crry-expanded natural Treg cells administered 1 day beforethe second PG immunization. This indicates that this

Figure 5. Modification of the immune response in human proteogly-can (hPG)–immunized mice upon transfer of Crry-expanded naturalTreg cells. A, Diminished levels of anti-hPG IgG1 and IgG2a in theserum of mice treated with Crry-expanded (Crry-exp) natural Tregcells, as determined by enzyme-linked immunosorbent assay. Serawere obtained 10 days after the second hPG immunization. Sera fromnonimmunized multiparous female mice (NMS, multip) and hPG-hyperimmunized mice (Hyperimm) were used as control. Bars showthe mean � SEM results for at least 5 mice per group and arerepresentative of 3 experiments. � � P � 0.05 versus control. B,Cytokine levels in the supernatants of in vitro cultures of lymphoidcells obtained 10 days after the second hPG immunization. Cells werecultured in medium alone (open bars) or were stimulated with hPG(solid bars). Bars show the mean � SEM and are representative of 3experiments. IL-4 � interleukin-4; IFN� � interferon-�; MCP-1 �monocyte chemotactic protein 1.

Figure 4. Therapeutic effect of Crry-expanded Treg cells inproteoglycan-induced arthritis. Arthritis scores in mice that receivednormal spleen cells (control), freshly obtained purified CD4�CD25�natural Treg cells, or Treg cells that were costimulated with Crry andexpanded for 7 days, 14 days after administration of the secondimmunization with human proteoglycan (hPG) (arrow) are shown.Values are the mean � SEM. � � P � 0.05 versus control.

Crry/p65 COSTIMULATION OF NATURAL Treg CELLS 1567

population can suppress B cell–dependent responsesrelevant to the development of PGIA.

T cell responses were analyzed in the same miceby activation of lymph node or spleen cells with antigen(soluble PG) in vitro. Culture supernatants obtained onday 4 were screened for cytokine content. Figure 5Bshows that lower levels of IFN�, IL-6, IL-12, and IL-17,together with other inflammatory cytokines such asmonocyte chemotactic protein 1 and RANTES, wereobserved in culture supernatants from Crry-expandedTreg cell–treated mice as compared with controls.Low IL-4 levels were detected in lymph node cultures,but the IFN�/IL-4 balance pointed toward a Th1 cellresponse in control immunized mice and toward a Th2cell response in Crry-expanded Treg cell–treated mice.These differences may explain the lower degree ofinflammation observed in mice treated with Crry-expanded Treg cells.

At the time of the assay, 10 days after the secondimmunization, low levels of IL-10 were detected in bothcontrol and Crry-expanded Treg cell–treated mice.However, mice treated with freshly obtained naturalTreg cells produced higher levels of IL-10 and IL-4,which might balance the high levels of IL-17 produced,temporarily delaying the inflammatory response.

Different homing patterns of Crry-expandedTreg cells and natural Treg cells. We also evaluatedwhether Crry-expanded Treg cells might have distinctproperties of trafficking or homing to the inflammationsite that also determine their ameliorating effect.Freshly obtained natural or Crry-expanded Treg cellswere loaded with CFSE and injected intravenously intoBALB/c mice, and their presence in secondary lymphoidorgans was analyzed by flow cytometry. Five days aftercell transfer, similar levels of CFSE-positive naturalTreg cells, expanded Treg cells, and controlCD4�CD25� cells were detected in the spleens (Figure6A). In contrast, the level of CFSE-positive cells waslower in lymph nodes of mice treated with Crry-expanded Treg cells (Figure 6B), indicating that thesecells have different homing or trafficking propertiescompared with those of freshly obtained natural Tregcells. Analysis of the cell cycle did not show proliferationof the CFSE-loaded cells in vivo in any case (data notshown), further suggesting that the differences observedwere attributable to the trafficking or homing pattern ofCrry-expanded cells.

The in vivo efficacy of suppression may bedetermined by the ability of Treg cells to reach andremain at the site of inflammation, as described in

Figure 6. In vivo homing patterns of Crry-expanded Treg cells and freshly obtained natural Treg cells. Cells loaded with 5,6-carboxyfluoresceinsuccinimidyl ester (CFSE) were transferred intravenously into mice 1 day before challenge with proteoglycan. Mice were treated with CD4�CD25�cells (control; open bars), Crry-expanded Treg cells (shaded bars), or freshly obtained natural Treg cells (solid bars). A and B, Percentage ofCFSE-labeled spleen (A) and lymph node (B) cells obtained 5 days after treatment, as analyzed by flow cytometry. C, Expression of FoxP3 mRNA,as analyzed by quantitative polymerase chain reaction. Hypoxanthine guanine phosphoribosyltransferase (HPRT) was used as a housekeeping gene.Bars show the mean � SEM. � � P � 0.05.

1568 OJEDA ET AL

other experimental models (28,29). To analyze thispossibility, we performed qPCR to detect FoxP3 in thepaw tissue of the different groups of mice. Figure 6Cshows that higher levels of FoxP3 were detected in micethat were treated with Crry-expanded Treg cells ascompared with control mice; similar results were ob-served when CD4 was used as the control gene (resultsnot shown). Natural Treg cell–treated mice showedintermediate levels of FoxP3 in the paws. Thus, accumu-lation of Treg cells at the site of inflammation might alsocontribute to the differences observed between groupstreated with freshly obtained or in vitro–expanded Tregcells.

DISCUSSION

The importance of natural Treg cells in theregulation of immune homeostasis is well established inhumans and mice, yet their therapeutic application isprecluded by their low numbers in the organism andtheir activation characteristics (for review, see refs.30–32). In the current study, we took advantage of thecostimulatory properties of the complement regulatorymolecule Crry (19,20) to efficiently expandCD4�CD25�FoxP3� natural Treg cells in vitro. TheCrry-expanded cells maintained their suppressive prop-erties in vitro and were more potent on a per-cell basisthan were natural Treg cells at diminishing and delayingthe onset and degree of PGIA, a model for chronicand progressive arthritis that depends on T cell– andB cell–mediated responses (11,12). This is relevant,because some patients with RA exhibit antigen-specificT cell and B cell responses against cartilage matrixcomponents (15,33,34).

Treg cells need T cell receptor triggering andcostimulation to become fully active, but no singlepathway is known to specifically costimulate Treg cells.This precludes the administration of agents capable ofselectively targeting specific T cell subsets. AdoptiveTreg cell therapy is an attractive alternative, which takesadvantage of the immunosuppressive activity of Tregcells (for review, see refs. 31 and 35). However, thisapproach requires the isolation, purification, and in vitroexpansion of a scarce population lacking a specific,exclusive surface marker, with anergic properties andbadly known requirements for in vitro expansion. Weshowed that CD3 signals in the presence of added IL-2potentiate the suppressive capacity of natural Treg cellsin vitro (additional information is available from thecorresponding author), but that the Crry costimulusallows an enhanced proliferative response (Figure 1A),

thereby facilitating the use of expanded Treg cells foradoptive transfer in vivo.

All CD4�CD25�FoxP3� natural Treg cells inthe mouse express surface Crry but not other comple-ment regulatory proteins such as decay-acceleratingfactor, CR1, or CD59, confirming the important role ofthis molecule in autologous complement attack andhomeostasis (36). Crry is a functional counterpart of thehuman complement regulator CD46, which is also areceptor for important pathogens including measlesvirus and Streptococcus pyogenes. CD46 ligation caninduce an IL-10–secreting Treg cell phenotype (Tr1)(37) or alter the pattern of granzyme expression innatural or Tr1 cells (38,39), suggesting a potential rolefor complement regulatory proteins in immune regula-tion. Here, we report for the first time that Crryfacilitates the expansion of isolated natural Treg cells invitro in the presence of CD3 signals and IL-2. Underthese conditions, expression of the transcription factorFoxP3 and surface CD25 is increased in Crry-costimulated Treg cell cultures for at least 10 days(Figures 1B and C, and results not shown). Furthermore,the expanded Treg cell population keeps its character-istic anergic properties in vitro when restimulated withCD3 (data not shown) and does not proliferate in vivo,as demonstrated by our results using CFSE-loaded cells,which (in our hands) makes them a safe population foradministration in vivo. CD3 activation and Crry costimu-lation of the CD4�CD25� cell population also inducedCD25 expression, yet under the conditions used theresulting population did not express FoxP3 or havesuppressor properties in vitro.

The effect of Crry costimulation in vitro couldbe observed not only in isolated CD4�CD25� cellsbut also in the total population of CD4� cells. In thatcase, the CD4�CD25�FoxP3� subpopulation prolifer-ated actively, decreasing the percentage of CD25�FoxP3� cells. However, these cells had a clear suppres-sive capacity in Treg cell assays in vitro. The naturalligand of Crry capable of costimulating T cells is notknown yet, but one could speculate that natural ligandsof Crry (i.e., C3b), in the presence of T cell receptor/CD3 stimuli, could fine-tune immune responses bymodulating/balancing the proliferation of CD25� cellsand the regulatory function of CD25� Treg cells, asrecently described for human CD46 (40). This high-lights the role of complement regulators in the controlof immune responses.

Crry costimulation modifies the cytokine secre-tion pattern of CD4� T cells (19), favoring IL-4 versusIFN�. We now report that Crry costimulation also

Crry/p65 COSTIMULATION OF NATURAL Treg CELLS 1569

increases IL-10 secretion when CD4�CD25� cells be-come activated (Figure 2) and in total CD4� cellcultures (results not shown). However, in spite of theenhanced expression of IL-10, the expanded isolatedCD4�CD25� cells did not show in vitro suppressorproperties (additional information is available from thecorresponding author). We also observed that in vitroactivation of natural Treg cells slightly increased theirexpression of IFN� and did not significantly modify theirexpression of IL-10. Although this could partially ac-count for the increased suppressive capability observedin Crry-expanded Treg cells, blocking of IL-10 did notsignificantly modify the suppression of natural orcostimuli-expanded Treg cells in vitro (data not shown).Thus, our data suggest that Crry-expanded Treg cells usea mechanism different from that recently reported (14),showing that inducible, IL-10–expressing Treg cells de-velop in vivo upon oral or intranasal administration ofspecific antigen, abolishing arthritis.

The participation of natural Treg cells in PGIAhas been studied in 2 different manners, with differentresults. Bardos et al (17) used a model of pathogeniccells transferred to SCID mice to show that cotransfer ofnatural CD4�CD25� Treg cells did not protect themice or delay illness onset. In contrast, Roord et al (18)reported that depletion of CD25� cells worsened PGIA,an effect that reverted following the transfer of normalbone marrow cells, in a 2-phase manner determined bynatural Treg cells. In our study, the transfer of naturalCD4�CD25� Treg cells to immunized mice delayed theonset of PGIA and ameliorated the clinical symptoms(Figures 3 and 4), but the effect was limited, and thearthritis score eventually reached the levels of untreatedmice. However, Treg cells activated and costimulatedthrough Crry were far more effective in delaying theonset of PGIA (Figure 3) and ameliorating inflamma-tion in PGIA (Figure 4), thus showing therapeuticpotential.

The differences between freshly obtained and invitro Crry–expanded Treg cells may be attributable todifferent regulatory mechanisms, because activation ofnatural Treg cells induced changes in cytokine expres-sion and the levels of CD25, FoxP3, CD103, latency-activating peptide, and neuropilin, as well as granzymecontent. Previous data provided by other investigatorsalso showed an increased suppressive function in invitro–activated natural Treg cells (5).

The ability of Treg cells to protect against auto-immune damage in a particular organ requires homingof the Treg cells to that organ (41). We observedincreased FoxP3 content in the paws of mice injected

with Crry-expanded Treg cells (Figure 6C), which mightcontribute to their therapeutic effect. In our system,improved clinical symptoms correlated with changes inT cell– and B cell–dependent responses on day 10 afterthe second immunization (Figure 5). Thus, mice admin-istered Crry-expanded Treg cells showed decreased se-rum anti-PG IgG2a levels and diminished in vitro anti-gen responses, yielding lower levels of secreted Th1 andproinflammatory cytokines. Mice treated with freshlyobtained natural Treg cells did not show lower IgG2aanti-PG antibody levels; although a balance toward aTh2 response was still present in these mice on day 10after the second immunization, expression of some ofthe inflammatory cytokines reached an intermediatelevel, which is indicative of evolution toward a stronginflammatory response.

Inflammation in arthritis models is probably theresult of several proinflammatory and antiinflammatorycytokines, the balance of which might determine thefinal outcome, as observed in other models of inflam-matory disease (42). There is growing evidence thatIL-17–secreting CD4� T cells (Th17 cells) play a keyrole in autoimmune diseases such as RA and multiplesclerosis (43). Although IL-17–knockout mice are sensi-tive to PGIA (44), recent data showed that both IL-17and IFN� have the potential to induce PGIA, yet it is thestrength of the IFN� response that regulates the contri-bution of each of these Th effector cytokines (16).

The participation of complement activation inTh17 cell differentiation and expansion in chronicautoimmune arthritis was recently described (45).Those results are compatible with our findings showingthat signals delivered in vitro through the membranecomplement regulatory protein Crry costimulatenatural Treg cells and enable them to diminish inflam-mation in PGIA, controlling levels of proinflammatorycytokines including IFN�, IL-6, and IL-17. Taken to-gether with results on human CD46 (37–40), our find-ings highlight the implication of regulators of comple-ment activation as an important link between innate andadaptive immunity in tolerance and in autoimmuneillnesses.

ACKNOWLEDGMENTS

The technical assistance of A. Nadal, M. Blanco, E.Martorell, and the Flow Cytometry Unit of Centro Nacional deMicrobiologıa is acknowledged.

AUTHOR CONTRIBUTIONSAll authors were involved in drafting the article or revising it

critically for important intellectual content, and all authors approved

1570 OJEDA ET AL

the final version to be published. Dr. Portoles had full access to all ofthe data in the study and takes responsibility for the integrity of thedata and the accuracy of the data analysis.Study conception and design. Ojeda, Pini, van Eden, Rojo, Portoles.Acquisition of data. Ojeda, Pini, Eguiluz, Montes-Casado, Broere, vanEden, Rojo, Portoles.Analysis and interpretation of data. Ojeda, Pini, Eguiluz, Montes-Casado, van Eden, Broere, Rojo, Portoles.

REFERENCES

1. Fontenot JD, Rasmussen JP, Williams LM, Dooley JL, Farr AG,Rudensky AY. Regulatory T cell lineage specification by theforkhead transcription factor Foxp3. Immunity 2005;22:329–41.

2. Sakaguchi S. Naturally arising CD4� regulatory T cells for immu-nologic self-tolerance and negative control of immune responses.Annu Rev Immunol 2004;22:531–62.

3. Karim M, Feng G, Wood KJ, Bushell AR. CD25�CD4� regula-tory T cells generated by exposure to a model protein antigenprevent allograft rejection: antigen-specific reactivation in vivo iscritical for bystander regulation. Blood 2005;105:4871–7.

4. Shevach EM. Mechanisms of Foxp3� T regulatory cell-mediatedsuppression. Immunity 2009;30:636–45.

5. Thornton AM, Shevach EM. Suppressor effector function ofCD4� CD25� immunoregulatory T cells is antigen nonspecific.J Immunol 2000;164:183–90.

6. Glant TT, Finnegan A, Mikecz K. Proteoglycan-induced arthritis:immune regulation, cellular mechanisms, and genetics. Crit RevImmunol 2003;23:199–250.

7. Tarjanyi O, Boldizsar F, Nemeth P, Mikecz K, Glant TT. Age-related changes in arthritis susceptibility and severity in a murinemodel of rheumatoid arthritis. Immun Ageing 2009;6:8.

8. Guerassimov A, Zhang Y, Cartman A, Rosenberg LC, Esdaile J,Fitzcharles MA, et al. Immune responses to cartilage link proteinand the G1 domain of proteoglycan aggrecan in patients withosteoarthritis. Arthritis Rheum 1999;42:527–33.

9. Li NL, Zhang DQ, Zhou KY, Cartman A, Leroux JY, Poole AR,et al. Isolation and characteristics of autoreactive T cells specific toaggrecan G1 domain from rheumatoid arthritis patients. Cell Res2000;10:39–49.

10. Zou J, Zhang Y, Thiel A, Rudwaleit M, Shi SL, Radbruch A, et al.Predominant cellular immune response to the cartilage autoanti-genic G1 aggrecan in ankylosing spondylitis and rheumatoidarthritis. Rheumatology (Oxford) 2003;42:846–55.

11. Berlo SE, Guichelaar T, ten Brink CB, van Kooten PJ, Hauet-Broeren F, Ludanyi K, et al. Increased arthritis susceptibility incartilage proteoglycan–specific T cell receptor–transgenic mice.Arthritis Rheum 2006;54:2423–33.

12. Finnegan A, Mikecz K, Tao P, Glant TT. Proteoglycan (aggrecan)-induced arthritis in BALB/c mice is a Th1-type disease regulatedby Th2 cytokines. J Immunol 1999;163:5383–90.

13. Boldizsar F, Tarjanyi O, Nemeth P, Mikecz K, Glant TT. Th1/Th17 polarization and acquisition of an arthritogenic phenotype inarthritis-susceptible BALB/c, but not in MHC-matched, arthritis-resistant DBA/2 mice. Int Immunol 2009;21:511–22.

14. Broere F, Wieten L, Klein Koerkamp EI, van Roon JA, Guiche-laar T, Lafeber FP, et al. Oral or nasal antigen induces regulatoryT cells that suppress arthritis and proliferation of arthritogenicT cells in joint draining lymph nodes. J Immunol 2008;181:899–906.

15. De Jong H, Berlo SE, Hombrink P, Otten HG, van Eden W,Lafeber FP, et al. Cartilage proteoglycan aggrecan epitopesinduce proinflammatory autoreactive T cell responses in

rheumatoid arthritis and osteoarthritis. Ann Rheum Dis 2010;69:255–62.

16. Doodes PD, Cao Y, Hamel KM, Wang Y, Rodeghero RL, MikeczK, et al. IFN-� regulates the requirement for IL-17 in proteoglycan-induced arthritis. J Immunol 2010;184:1552–9.

17. Bardos T, Czipri M, Vermes C, Finnegan A, Mikecz K, Zhang J.CD4�CD25� immunoregulatory T cells may not be involved incontrolling autoimmune arthritis. Arthritis Res Ther 2003;5:R106–13.

18. Roord ST, de Jager W, Boon L, Wulffraat N, Martens A, PrakkenB, et al. Autologous bone marrow transplantation in autoimmunearthritis restores immune homeostasis through CD4�CD25�

FoxP3� regulatory T cells. Blood 2008;111:5233–41.19. Fernandez-Centeno E, de Ojeda G, Rojo JM, Portoles P. Crry/

p65, a membrane complement regulatory protein, has costimula-tory properties on mouse T cells. J Immunol 2000;164:4533–42.

20. Jimenez-Perianez A, Ojeda G, Criado G, Sanchez A, Pini E,Madrenas J, et al. Complement regulatory protein Crry/p65-mediated signaling in T lymphocytes: role of its cytoplasmicdomain and partitioning into lipid rafts. J Leukoc Biol 2005;78:1386–96.

21. Peck AB, Bach FH. A miniaturized mouse mixed leukocyte culturein serum-free and mouse serum supplemented media. J ImmunolMethods 1973;3:147–63.

22. Portoles P, Rojo J, Golby A, Bonneville M, Gromkowski S,Greenbaum L, et al. Monoclonal antibodies to murine CD3�define distinct epitopes, one of which may interact with CD4during T cell activation. J Immunol 1989;142:4169–75.

23. Rojo JM, Pini E, Ojeda G, Bello R, Dong C, Flavell RA, et al.CD4�ICOS� T lymphocytes inhibit T cell activation ‘in vitro’ andattenuate autoimmune encephalitis ‘in vivo’. Int Immunol 2008;20:577–89.

24. Mosmann T. Rapid colorimetric assay for cellular growth andsurvival: application to proliferation and cytotoxicity assays. J Im-munol Methods 1983;65:55–63.

25. Hanyecz A, Berlo SE, Szanto S, Broeren CP, Mikecz K, Glant TT.Achievement of a synergistic adjuvant effect on arthritis inductionby activation of innate immunity and forcing the immune responsetoward the Th1 phenotype. Arthritis Rheum 2004;50:1665–76.

26. Hsieh CS, Macatonia SE, O’Garra A, Murphy KM. T cell geneticbackground determines default T helper phenotype developmentin vitro. J Exp Med 1995;181:713–21.

27. Hondowicz BD, Park AY, Elloso MM, Scott P. Maintenance ofIL-12-responsive CD4� T cells during a Th2 response in Leish-mania major-infected mice. Eur J Immunol 2000;30:2007–14.

28. Leavy O. Regulatory T cells: taking the right route. Nat RevImmunol 2009;9:311.

29. Zhang N, Schroppel B, Lal G, Jakubzick C, Mao X, Chen D, et al.Regulatory T cells sequentially migrate from inflamed tissues todraining lymph nodes to suppress the alloimmune response.Immunity 2009;30:458–69.

30. Josefowicz SZ, Rudensky A. Control of regulatory T cell lineagecommitment and maintenance. Immunity 2009;30:616–25.

31. June CH, Blazar BR. Clinical application of expandedCD4�CD25� cells. Semin Immunol 2006;18:78–88.

32. Pini E, Ojeda G, Portoles P. The renaissance of T regulatory cells:looking for markers in a haystack. Inmunologıa 2007;26:100–7.

33. Burmester GR, Stuhlmuller B, Keyszer G, Kinne RW. Mononu-clear phagocytes and rheumatoid synovitis: mastermind or work-horse in arthritis? [review]. Arthritis Rheum 1997;40:5–18.

34. Goronzy JJ, Weyand CM. Rheumatoid arthritis. Immunol Rev2005;204:55–73.

35. Riley JL, June CH, Blazar BR. Human T regulatory cell therapy:take a billion or so and call me in the morning. Immunity2009;30:656–65.

36. Li Q, Nacion K, Bu H, Lin F. Mouse CD4�CD25� T regulatory

Crry/p65 COSTIMULATION OF NATURAL Treg CELLS 1571

cells are protected from autologous complement mediated injuryby Crry and CD59. Biochem Biophys Res Commun 2009;382:223–6.

37. Kemper C, Chan AC, Green JM, Brett KA, Murphy KM, Atkin-son JP. Activation of human CD4� cells with CD3 and CD46induces a T-regulatory cell 1 phenotype. Nature 2003;421:388–92.

38. Grossman WJ, Verbsky JW, Barchet W, Colonna M, AtkinsonJP, Ley TJ. Human T regulatory cells can use the perforin path-way to cause autologous target cell death. Immunity 2004;21:589–601.

39. Grossman WJ, Verbsky JW, Tollefsen BL, Kemper C, AtkinsonJP, Ley TJ. Differential expression of granzymes A and B inhuman cytotoxic lymphocyte subsets and T regulatory cells. Blood2004;104:2840–8.

40. Cardone J, Le Friec G, Vantourout P, Roberts A, Fuchs A,

Jackson I, et al. Complement regulator CD46 temporally regulatescytokine production by conventional and unconventional T cells.Nat Immunol 2010;11:862–71.

41. Feuerer M, Hill JA, Mathis D, Benoist C. FoxP3� regulatoryT cells: differentiation, specification, subphenotypes. Nat Immu-nol 2009;10:689–95.

42. Bettelli E, Oukka M, Kuchroo VK. TH-17 cells in the circle ofimmunity and autoimmunity. Nat Immunol 2007;8:345–50.

43. Korn T, Bettelli E, Oukka M, Kuchroo VK. IL-17 and Th17 Cells.Annu Rev Immunol 2009;27:485–517.

44. Doodes PD, Cao Y, Hamel KM, Wang Y, Farkas B, Iwakura Y, etal. Development of proteoglycan-induced arthritis is independentof IL-17. J Immunol 2008;181:329–37.

45. Hashimoto M, Hirota K, Yoshitomi H, Maeda S, Teradaira S,Akizuki S, et al. Complement drives Th17 cell differentiation andtriggers autoimmune arthritis. J Exp Med 2010;207:1135–43.

1572 OJEDA ET AL