ARTHRITIS & RHEUMATISMVol. 63, No. 7, July 2011, pp 1878–1887DOI 10.1002/art.30336© 2011, American College of Rheumatology

Critical Proinflammatory Role ofThymic Stromal Lymphopoietin and its Receptor in

Experimental Autoimmune Arthritis

S. A. Y. Hartgring,1 C. R. Willis,2 C. E. Dean, Jr.,2 F. Broere,3 W. van Eden,3 J. W. J. Bijlsma,1

F. P. J. G. Lafeber,1 and J. A. G. van Roon1

Objective. The interleukin-7 (IL-7)–related cyto-kine thymic stromal lymphopoietin (TSLP) is a potentactivator of myeloid dendritic cells, enhancing Th2-mediated hypersensitivity, and it has been implicated inthe pathogenesis of atopic diseases. Although intra-articular concentrations of TSLP have been shown to beincreased in patients with rheumatoid arthritis (RA),the functional capacities of TSLP in arthritis are poorlystudied. The purpose of this study was to investigate theeffects of TSLP administration and TSLP receptordeficiency on immune activation, arthritis severity, andtissue destruction in T cell–driven arthritis models ofRA.

Methods. Immunopathology was studied in ar-thritic mice that were given multiple injections of mu-rine recombinant TSLP and in mice that were deficientin the TSLP receptor (TSLPR�/�). Arthritis severityand incidence were determined by visual examination ofthe paws. Joint destruction was determined by assessingradiographs and the immunohistochemistry of anklejoints. Total cellularity and numbers of T cell subsetswere assessed. Proinflammatory mediators were mea-sured by multianalyte profiling of serum or paw proteinextracts.

Results. Administration of TSLP significantly ex-acerbated the severity of collagen-induced arthritis and

the joint damage that was associated with increasedT cell activation. Furthermore, TSLPR�/� mice had lesssevere arthritis than did wild-type mice. TSLPR�/�

mice had diminished concentrations of local proinflam-matory and catabolic mediators, including IL-17, IL-1�,IL-6, basic fibroblast growth factor, and matrix metal-loproteinase 9, while levels of the regulatory cytokinesIL-10 and IL-13 were increased.

Conclusion. TSLP and its receptor enhance Th17-driven arthritis and tissue destruction in experimentalarthritis. The increased expression of TSLP as well asthe increased number of TSLPR-expressing cells in thejoints of patients with RA suggest that TSLP and itsreceptor constitute novel therapeutic targets in RA.

Rheumatoid arthritis (RA) is characterized bypersistent inflammation of the joints, which results inprogressive destruction of the joint tissues, particularlycartilage and bone (1). Several studies have demon-strated an important role of interleukin-17 (IL-17)–producing (Th17) and interferon-� (IFN�)–producing(Th1) CD4 T cells, B cells, and macrophages in theinflamed joints of RA patients. The numbers of thesecells in synovial fluid and synovial tissue are increased inpatients with RA and correlate with clinical symptoms(2–9). Moreover, therapies targeting these cells or theirproducts, such as CTLA-4Ig, anti-CD20, anti–tumornecrosis factor (anti-TNF), and anti–IL-6 receptor, areeffective in RA (7–10). Yet, there is a medical need inRA for new therapies targeting other inflammatorymolecules, since many RA patients either do not re-spond or respond only partially to these therapies. Werecently identified IL-7 as an arthritogenic cytokine thatis persistently expressed in RA patients who were resis-tant to TNF-blocking therapy (11–14).

Thymic stromal lymphopoietin (TSLP) is related

1S. A. Y. Hartgring, PhD, J. W. J. Bijlsma, MD, PhD,F. P. J. G. Lafeber, PhD, J. A. G. van Roon, PhD: University MedicalCenter Utrecht, Utrecht, The Netherlands; 2C. R. Willis, BSc, C. E.Dean, Jr., PhD, DVM: Amgen, Seattle, Washington; 3F. Broere, PhD,W. van Eden, MD, PhD: Utrecht University, Utrecht, The Nether-lands.

Ms Willis and Dr. Dean own stock or stock options in Amgen.Address correspondence to J. A. G. van Roon, PhD, Depart-

ment of Rheumatology and Clinical Immunology, University MedicalCenter Utrecht, PO Box 85500, F02.127, 3508 GA, Utrecht, TheNetherlands. E-mail: J.vanRoon@umcutrecht.nl.

Submitted for publication August 10, 2010; accepted inrevised form March 1, 2011.

1878

to IL-7 and is produced by epithelial cells, fibroblasts,mast cells, and keratinocytes (15). Epithelial cells pro-duce TSLP in response to Toll-like receptor 3 (TLR-3)signaling, cytokines such as IL-1� and TNF�, and pro-teases that interact with protease-activated receptor 2(16–18). Signaling of cells by TSLP requires IL-7 recep-tor � (19) and a distinctive receptor subunit, the TSLPreceptor (TSLPR), which is expressed by myeloid den-dritic cells (MDCs), monocytes, preactivated T cells,natural killer cells, and mast cells (20–23). TSLPR-deficient (TSLPR�/�) mice display normal T cell andB cell development. Yet, TSLP stimulates lymphocytedevelopment and expansion in young and lymphopenicmice (24,25). In humans, TSLP potently stimulatesMDCs, with up-regulated expression of HLA–DR,CD40, CD80, CD86, OX40L, and CD83 and productionof chemokines, including thymus and activation–regulated chemokine and macrophage-derived chemo-kine (26,27). TSLP-activated MDCs in their turn stimu-late naive autologous CD4 T cells to proliferate andexpand. TSLP-activated MDC-mediated expansion ofnaive CD4 T cells induces central memory T cells, whichretain the capacity to differentiate into Th1 or Th2 cells(26).

TSLP has been studied in human atopic disordersbecause of its effects on Th2 cytokine–secreting T cells(15). In atopic dermatitis, the cytokine is overexpressedin acute and chronic skin lesions (15), and in allergicasthma, TSLP is increased in the airways and correlateswith Th2 activity and disease severity (28). Overexpres-sion of TSLP in the mouse lung induces the spontaneousprogressive airway inflammation that is characteristic ofhuman asthma (29), whereas expression in the skin caninduce atopic dermatitis–like lesions. Interestingly, ex-pression of TSLP at one anatomical site can trigger Th2cytokine–driven inflammation at another (30).

However, TSLP may also contribute to Th1cytokine– and Th17 cytokine–dependent inflammation.First, TNF� and TLR ligands induce the up-regulationof TSLP in synovial fibroblasts derived from patientswith RA (31). Second, intraarticular TSLP concentra-tions are increased in a subset of patients with RA (31)and TSLP-activated MDCs have the ability to stimulateautologous T cells to secrete TNF�, IFN�, and IL-17(15,26,32), indicative of Th1 and Th17 activation, whichare critically involved in the pathogenesis of RA (4,33).Finally, in the T cell– and B cell–independent collagenantibody–induced arthritis model of TSLP blockade wasshown to inhibit arthritis severity (31).

Although the above-mentioned findings indicatea likely role of TSLP in arthritis, the functional proper-

ties of TSLP in T cell–driven autoimmune arthriticconditions have not previously been studied. Further-more, the mechanisms by which TSLP might aggravateexperimental arthritis are unknown. We thereforeinvestigated the contribution of TSLP and its receptor toTh1- and Th17-dependent experimental autoimmunearthritis.

MATERIALS AND METHODS

Induction, treatment, and assessment of collagen-induced arthritis (CIA) and proteoglycan-induced arthritis(PGIA). CIA and PGIA were produced as described previously(34–36). Mice were randomly divided into 3 groups of 15animals each. Recombinant mouse TSLP (10 �g; R&D Sys-tems) or phosphate buffered saline (PBS) was injected intra-peritoneally starting at the time the collagen booster was given(day 21) and was subsequently given every other day until day31 (total of 6 injections). Female BALB/c wild-type (WT) miceand TSLPR-deficient (TSLPR�/�) BALB/c mice (24 weeksold; Charles River Laboratories) were used for PGIA.TSLPR�/� mice were generated as previously described (25)and were backcrossed to a 100% BALB/c background atCharles River Laboratories. Both WT mice and TSLPR�/�

mice were randomly divided into 4 groups of 16 or 17 mice. Allexperiments were performed in accordance with the guidelinesof the animal ethical committee.

Assessment of radiologic and histologic joint damage.The ankles of each mouse were isolated and fixed in formalinfor 24 hours. Subsequently, lateral–medial radiographs of allankles were taken and were scored. After radiography, anklejoints were decalcified in 10% EDTA buffer and embedded inparaffin. Tissue sections (5 �m) were stained with hematoxylinand eosin. Sections were scored for severity of cell infiltrates(inflammation), subchondral bone erosion, periosteal exosto-ses (osteophytes), and articular cartilage erosion. Each histo-logic parameter was graded according to the following index:grade 0 � no abnormal findings, grade 1 � minimal, grade 2 �mild, grade 3 � moderate, and grade 4 � marked/severe.Staining for cathepsin K and for tartrate-resistant acid phos-phatase was done on deparaffinized tissue sections. Histologicparameters were graded according to the following index:grade 0 � no abnormal findings, grade 1 � minimal, grade 2 �mild, grade 3 � moderate, and grade 4 � marked/severe. Bothankles were graded, and the mean grade for each ankle wascalculated for each treatment group.

Spleen and thymus cell preparation and flow cytom-etry. After termination of the experiment, the spleen andthymus were collected, weighed, and different cell types wereassessed by fluorescence-activated cell sorting (FACS). For theidentification of different cell types by flow cytometry, thefollowing monoclonal antibodies were used: allophycocyanin(APC)/Cy7–labeled anti-CD4, Pacific Blue–labeled anti-CD8,PerCP/Cy5.5-labeled anti-CD19, and APC-labeled anti-CD62L (all from Becton Dickinson), as well as fluoresceinisothiocyanate–labeled anti-CD44 (eBioscience). A total of 105

events were collected using a FACS LSRII flow cytometer(Becton Dickinson), and the data were analyzed by gating on

TSLP AND ITS RECEPTOR PROMOTE ARTHRITIS 1879

viable lymphocytes, based on forward and side scatter patterns,and by specific staining, using FlowJo software (Tree Star).

Detection of inflammation-associated proteins in pawlysates. The front paws of each mouse were collected andfrozen in liquid nitrogen directly upon removal. A tissue lyser(Qiagen) was used to shred paws into homogenous lysates indigestion buffer (50 mM Tris HCl, pH 7.4 containing 0.1M

NaCl, 0.1% Triton X-100), and Mini EDTA-free proteaseinhibitor tablets (Roche) with the use of 5-mm stainless steelbeads (Qiagen). The protein content of the lysates was nor-malized by first assaying with a Pierce BCA total proteinquantitation kit and then adjusting each sample to a concen-tration of 1 mg/ml for assaying. The concentrations of specificproteins in paw lysates were assayed by Rules-Based Medi-

Figure 1. Thymic stromal lymphopoietin (TSLP)–induced exacerbation of the severity of collagen-induced arthritisassociated with increased joint destruction. Mice were injected with phosphate buffered saline (PBS; control) or withTSLP every other day, starting on day 21 and continuing until day 31, and arthritis severity was assessed. On day 33,inflammation and immunopathology were graded based on histologic features. A, TSLP treatment significantly increasedarthritis severity. B, Inflammation, bone and cartilage erosion, and periosteal exostosis were also increased inTSLP-treated mice. C, In association with increased bone erosions, TSLP showed significantly increased expression ofcathepsin K, indicative of osteoclast activity. TSLP did not significantly change the incidence of arthritis (data notshown). Values are the mean � SEM of 15 mice per group. � � P � 0.05 versus PBS-treated controls.

Figure 2. Thymic stromal lymphopoietin (TSLP)–induced expansion of splenic CD4 T cells. A, The total number of thymocytes wasnot affected by TSLP administration. The absolute numbers of CD4 and CD8 double-negative (DN), double-positive (DP), andsingle-positive thymocytes were also not changed by treatment with TSLP. B, TSLP treatment significantly increased the total numbersof splenic CD4 T cells and slightly increased the number of CD19 B cells and CD8 T cells. Values are the mean � SEM of 15 miceper group. For the cell subsets, each data point represents a single mouse; horizontal lines show the mean. � � P � 0.05 versusphosphate buffered saline (PBS)–treated controls.

1880 HARTGRING ET AL

cine, using multianalyte profiling for rodent antigens (version2.0).

RESULTS

TSLP exacerbation of CIA and immunopathol-ogy associated with T cell activation. To test the effectsof TSLP on CIA, mice immunized with type II collagenreceived multiple injections of TSLP or, as a control,PBS. TSLP treatment significantly increased the severityof the arthritis (Figure 1A), but had no effect on theincidence of the arthritis. Arthritis developed in 93% ofmice treated with PBS, as compared with 100% of thosetreated with TSLP (data not shown).

The effects on arthritis severity were consistentwith the histologic features in the ankle joints, whichrevealed that TSLP treatment significantly increasedperiosteal exostosis (P � 0.05). TSLP also tended toenhance the infiltration by inflammatory cells, the for-

mation of bone erosions, and the induction of cartilagedamage (Figure 1B). Moreover, TSLP significantly stim-ulated the formation of osteoclasts, as measured by theexpression of cathepsin K (Figure 1C).

The immunomodulatory effects of TSLP wereanalyzed by assessing the total cell numbers and thenumbers of T cell subsets in the thymus and spleen. Thenumbers of thymocytes or thymocyte subsets were notsignificantly affected by TSLP treatment (Figure 2A). Inthe arthritic mice, the numbers of splenocytes, CD4T cells (P � 0.05), CD8 T cells, and B cells wereincreased by TSLP treatment (Figure 2B). Furthermore,TSLP treatment induced significant expansion of centraland effector CD4 memory T cells, whereas naive CD4T cells were not significantly altered (Figure 3A).

Since in several animal models of atopic disor-ders TSLP was found to increase Th2 activity, we nextstudied whether TSLP administration in arthritic mice

Figure 3. Thymic stromal lymphopoietin (TSLP)–induced expansion of memory and effector cytokine-secreting Thelper cells. A, The number of naive CD4 T cells (CD44–CD62L�) was not significantly altered by TSLP treatment. Incontrast, TSLP administration significantly increased the number of central memory (CD44�CD62L�) and effectormemory (CD44�CD62L–) T cells. B, Intracellular cytokine stainings for T cells positive for interferon-� (IFN�),interleukin-17 (IL-17), and IL-4 were performed in 5 of the 15 mice from each group in A that had a representativedegree of clinical arthritis severity. TSLP treatment enhanced the numbers of cytokine-secreting Th cells, but onlysignificantly increasing the numbers of Th2 (IL-4)–secreting cells. Each data point represents a single mouse; horizontallines show the mean. � � P � 0.05; �� � P � 0.01 versus phosphate buffered saline (PBS)–treated controls.

TSLP AND ITS RECEPTOR PROMOTE ARTHRITIS 1881

was also associated with Th2 activity and in what wayTh1 and Th17 activity are regulated. The increase inmemory CD4 T cells by TSLP treatment was associatedwith increased numbers of CD4 T cells that producedIL-17 (from a mean � SEM of 1.1 � 0.2 to 3.1 � 1.1 �105 cells with PBS versus TSLP), IFN� (from 1.0 � 0.8to 1.4 � 0.2 � 105 cells with PBS versus TSLP), and IL-4(from 0.8 � 0.5 to 1.6 � 0.2 � 105 cells;P � 0.05 with PBS versus TSLP) (Figure 3B).

The differences in arthritis severity and immuno-pathology induced by TSLP coincided with changes inthe concentrations of circulating cytokines. TSLP-induced T cell activation was associated with significantincreases (P � 0.05 for all comparisons) in serumconcentrations of soluble CD40L, indicative of T cellactivation, and cytokines associated with chemotaxis ofneutrophils and monocyte/macrophages (macrophage-derived chemokine and monocyte chemotactic protein 5[MCP-5]) (data not shown).

Reduced experimental arthritis and immunopa-thology by prevention of TSLPR-mediated immune ac-tivation. To further investigate the contribution of TSLPto inflammatory reactions in arthritis, we investigatedthe development and severity of chronic relapsing PGIAin TSLPR deficient BALB/c mice as compared to wild-type BALB/c mice. This strategy selectively interfereswith TSLP signaling and has no effect on IL-7–mediatedimmune activation.

TSLPR�/� mice showed less severe arthritis(mean � SEM inhibition 56 � 6% on day 24–36) ascompared to WT mice (Figure 4A). This was associatedwith less joint damage, as measured radiographically(mean � SEM radiologic score 1.4 � 0.2 in WT miceversus 0.3 � 0.1 in TSLPR�/� mice [75% reduction];P � 0.001) (Figure 4B). This finding was consistent witha significantly improved immunopathology profile,showing reductions in cell infiltrates and in cartilage aswell as bone erosions (Figures 4C and D). The decrease

Figure 4. Inhibition of chronic relapsing proteoglycan-induced arthritis (PGIA) in thymic stromal lymphopoietin receptor(TSLPR)–deficient mice. PGIA was induced in wild-type (WT) and TSLPR�/� BALB/c mice, and the arthritis severity wasgraded. A, Arthritis severity was significantly decreased in TSLPR�/� mice as compared to WT mice. B, Radiographs of anklejoints were taken on day 36, and radiologic joint damage in each ankle was graded on a scale of 0–3. Significantly lowerradiologic damage was found in TSLPR�/� mice. C, Photomicrographs of representative ankle joint sections from a WT mouseand a TSLPR�/� mouse show differences in the condition of the articular surface and in the amount and severity of infiltratingcells in the connective tissues and joint spaces (arrows). Original magnification � 100. D, Ankle joint sections were scored ona scale of 0–3 for each of 4 features: inflammation, cartilage erosion, bone erosion, and periosteal exostosis. Each feature wassignificantly decreased in TSLPR�/� mice. E, Osteoclast formation was reduced in TSLPR�/� mice, as indicated by the numberof cells staining positive for tartrate-resistant acid phosphatase (TRAP). F, Total numbers of thymocytes, splenocytes, CD4 Tcells, and CD19 B cells were not significantly different in TSLPR�/� mice as compared to WT mice. Values are mean � SEMof 16 mice per group. � � P � 0.05; �� � P � 0.005; ��� � P � 0.001 versus WT mice.

1882 HARTGRING ET AL

in bone erosion was associated with reduced numbers oftartrate-resistant acid phosphatase–expressing oste-oclasts (Figure 4E).

Thymocyte and splenocyte numbers were notsignificantly changed in TSLPR�/� mice as compared toWT mice. Specifically, the splenic CD4 T cell and CD19B cell counts were not significantly different in WT micefrom those in TSLPR�/� mice (Figure 4F).

Association of reduced immunopathology inTSLPR�/� mice with reduced Th17 activity and proin-flammatory mediators. We found that IFN� levels inpaw extracts prepared from TSLPR�/� mice were sig-nificantly increased (Figure 5) and were associated withincreased IL-12 concentrations (mean � SEM 96 � 9pg/ml versus 118 � 7 pg/ml). Reduced IL-4 concentra-tions were also detected in TSLPR�/� mice (mean �SEM 11.1 � 1.7 in WT mice versus 7.8 � 0.8 pg/ml inTSLPR�/� mice; values below the least detectable dose)(data not shown). TSLPR�/� mice showed reducedIL-17 concentrations, as well as reduced local concen-trations of CD40L (and CD40 [data not shown]). Con-versely, the regulatory cytokine IL-10 was significantlyincreased in TSLPR�/� mice, and IL-13 also showedincreased concentrations, although the difference wasnot statistically significant (Figure 5).

In addition, we tested whether the reduced im-munopathologic features were associated with a reduc-tion in the levels of cytokines that are prominentlyproduced by monocyte/macrophages and have beendemonstrated to play pivotal roles in arthritic conditions.Whereas the TNF� levels were not significantly reducedin TSLPR�/� mice as compared to WT mice, the IL-1�levels were. In addition, local levels of IL-6 (P � 0.059)as well as systemic levels of IL-6 (P � 0.005) werereduced in TSLPR�/� mice.

The expression of chemokines that mediate theattraction of T and B lymphocytes, monocyte/macrophages, and granulocytes to sites of inflammationwas reduced in the paws of TSLPR�/� mice. Figure 5shows the levels of representative chemokines. Chemo-kines that were also significantly reduced in TSLPR�/�

mice include macrophage inflammatory protein 1�(MIP-1�), MIP-3�, MCP-1, MIP-1�, MIP-2, granulo-cyte chemotactic protein 2, and eotaxin (P � 0.05 foreach comparison) (data not shown). Factors that medi-ate tissue destruction were also significantly reduced inTSLPR�/� mice. Figure 5 shows the concentrations ofrepresentative mediators. Other mediators that weresignificantly reduced include oncostatin M (OSM), IL-11, leukemia inhibitory factor (LIF), and tissue inhibitor

Figure 5. Association of arthritis inhibition with reduced local expression of proinflammatory cytokines in thymic stromallymphopoietin receptor (TSLPR)–deficient mice. Proteoglycan-induced arthritis (PGIA) was initiated in wild-type (WT) or TSLPR�/�

BALB/c mice, and the levels of mediators of inflammation were measured in paw lysates, except for interleukin-6 (IL-6), which wasmeasured in serum. Cytokine concentrations in paw protein lysates were measured by multicytokine analysis. Levels of cytokinesassociated with T cell and monocyte/macrophage activation, chemotaxis, and tissue destruction were reduced in TSLPR�/� mice. Incontrast, levels of IL-10 were significantly increased; IL-13 levels were increased, although not statistically significantly. Values aremean � SEM of 16 mice per group. � � P � 0.05; �� � P � 0.01; ��� � P � 0.005 versus WT mice. IFN� � interferon-�; TNF� �tumor necrosis factor �; MCP-1 � monocyte chemotactic protein 1; MDC � macrophage-derived chemokine; KC � keratinocytechemoattractant; MIP-1� � macrophage inflammatory protein 1�; FGF-b � basic fibroblast growth factor; MMP-9 � matrixmetalloproteinase 9.

TSLP AND ITS RECEPTOR PROMOTE ARTHRITIS 1883

of metalloproteinases 1 (TIMP-1) (P � 0.05 for eachcomparison).

DISCUSSION

This study demonstrates that TSLP treatmentenhanced arthritis and joint destruction in mice withCIA, which was associated with increased T cell activa-tion and circulating levels of mediators of inflammation.Consistent with this, TSLPR deficiency strongly amelio-rated the severity and joint destruction of PGIA, whichwas associated with down-regulation of Th17 activityand inhibition of proinflammatory cytokines and cata-bolic mediators at sites of inflammation.

It was recently shown that blockade of TSLPreduced the severity of collagen antibody–induced ar-thritis. This arthritis model is induced by passive immu-nization, is independent of T cell and B cell activation,and is largely dependent on TNF� induction triggeredby immune complexes (31). Since activated T cells andB cells play crucial proinflammatory roles in the immu-nopathology of RA (6,8), we studied the effects of TSLPand TSLPR deficiency in experimental arthritis modelsthat are critically dependent on T cell and B cellactivation in addition to activation of cells from theinnate immune system (CIA and PGIA). Furthermore,since the mechanisms and mediators of inflammationthat facilitate immune activation in animal models ofRA were unknown, these were studied as well.

Consistent with previous data demonstratingTSLP-induced homeostatic expansion of CD4 T cells(26), we found that TSLP enhanced the numbers ofcirculating central and effector memory T cells, thenumbers of IL-17–, IFN�-, and IL-4-producing CD4 Tcells, as well as systemic levels of soluble CD40L (26).Since CIA critically depends on Th17 cells and to alesser extent on Th1 cells, the TSLP-induced increase innumbers of Th17 and Th1 cells could locally contributeto the increased arthritis and joint destruction observedin our study (37). The TSLP-induced increase in IL-4–producing cells may, to a certain extent, control arthritisand joint destruction in CIA as has previously beendemonstrated, although in the present study this Th2induction does not prevent TSLP from enhancing im-munopathology (38,39).

The present study also shows reduced local con-centrations of IL-4 (Th2 activity) and increased IFN�and IL-12 (Th1 activity) in TSLPR�/� mice. These dataare consistent with the observation that TSLPR�/� miceare characterized by reduced IL-4 and increased IL-12and IFN� production (40). More importantly, in this

study, we demonstrated that this Th1 skewing wasaccompanied by a strong reduction in IL-17 levels in theinflamed paws of mice. This suggests that the diseasesuppression observed in this study was mediated by thedown-regulation of Th17 activity, which is consistentwith recent data showing that the later stages of PGIAare critically dependent on IL-17 production (41). This isalso supported by the finding that TSLP administrationcould increase Th17 numbers (this study) and thatTSLP-activated MDCs, from both the peripheral bloodand the synovial fluid of RA patients, potently inducedIL-17 by autologous CD4 T cells from the same patients(42). A selective reduction in IL-17 production couldalso explain the robustly reduced immunopathology inTSLPR�/� mice, since PGIA has been shown to bestrongly Th17-dependent (41,43,44).

The IL-17 production by TSLP/TSLPR may in-volve the CD40L/CD40 costimulatory pathway, sinceup-regulation of CD40L leads to increased IL-17 pro-duction by CD4 T cells (45). Considering that theinterplay between IL-17 and IL-6 strongly stimulatesautoimmunity (46), the breakdown of TSLP-driven im-munity may also involve IL-17/IL-6–mediated pathwaysin autoimmune arthritis. The strong reduction of CD40Land IL-6 in TSLPR�/� mice indicate that these media-tors could be involved in TSLP-driven Th17-associatedimmunopathology. A reduction in IL-6 levels has beensuggested to restore regulatory T cell function (47),resulting in reduced T cell activation and T cell–drivenresponses. The significantly increased IL-10 concentra-tions in TSLPR�/� mice are consistent with this assump-tion.

Associated with the finding of Th17 activation,our data also suggest that TSLP-induced expression ofchemokines that mediate the attraction of T and Blymphocytes (MIP-1�, MIP-1�, IFN�-inducible 10-kdprotein), monocyte/macrophages (MCPs 1, 3, and 5) andgranulocytes (keratinocyte chemoattractant, MIP-1�,eotaxin, macrophage-derived chemokine) to the site ofinflammation strongly contributes to arthritis, since thelevels of these chemokines were reduced in the paws ofTSLPR�/� mice. Finally, TSLP-induced immune activa-tion seems to strongly contribute to joint destruction,since TSLPR�/� mice displayed decreased cartilage andbone erosion, decreased numbers of tartrate-resistantacid phosphatase–positive osteoclasts, and reduced localexpression of factors that are involved in tissue destruc-tion (basic fibroblast growth factor, matrix metallopro-teinase 9, OSM, IL-11, LIF, and TIMP-1).

This study showed that TSLPR deficiency doesnot alter splenic T cell or B cell numbers. We also

1884 HARTGRING ET AL

showed that thymic T cells were not significantly af-fected by TSLP administration or TSLPR deficiency, incontrast to other investigators, who have shown anincrease in thymocytes following TSLP treatment (24).This discrepancy presumably reflects differences in theages of the mice included in the different studies. In thepresent study, mice were 2 months old (TSLP) and 7months old (TSLPR�/�). Older mice are more suscep-tible to the development of arthritis and are character-ized by involuted thymuses with reduced thymic output.In the study by Al-Shami et al (24), younger mice (2weeks of age) had a higher thymic output and weretherefore likely to be more sensitive to the effects ofTSLP.

In addition to the previously suggested role ofTSLP in immune complex–induced arthritis that directlyinvolves the activation of cells of the innate immunesystem (31), the present data from our CIA and PGIAautoimmune arthritis models suggest that TSLP is ableto promote self-antigen–driven T cell–dependent jointinflammation and immunopathology. Our data are con-sistent with findings demonstrating that TSLP activatesMDCs and causes proliferation and expansion of self-reactive CD4 T cells (26). In healthy individuals, thiscould contribute to T cell homeostasis; in disease states,this could boost initiating or chronic T cell–drivenimmune responses. This might be dependent on geneticpredisposition as well as the local microenvironment,causing Th2-driven immunity to prevail in atopic dis-eases and Th17- or Th1-driven autoimmunity in diseasessuch as RA. In this regard, previous studies have dem-onstrated that TSLP-primed MDCs augment the induc-tion of Th1 and Th17 activity (48). In support of its rolein promoting autoimmunity, TSLP-activated MDCsstrongly express messenger RNA for the autoimmuneregulator autoimmune regulator (AIRE) protein, a tran-scriptional regulator that controls the expression of abroad range of self antigens on antigen-presenting cells,contributing to the in vitro expansion of human CD4 Tcells (26). In autoimmune diseases such as RA, TSLP-induced AIRE up-regulation might contribute to thediversified autoreactive T cell response that has beendocumented in this disease (49).

In support of previous findings (31), we haverecently demonstrated strongly and significantly in-creased concentrations of TSLP in the synovial fluid ofRA patients as compared to patients with osteoarthritis(42). In addition, we demonstrated that the percentagesof TSLPR-expressing CD1c� MDCs among the mono-nuclear cell population in RA synovial fluid were in-

creased as compared to the percentages in paired sam-ples of peripheral blood (42).

The present study demonstrates that exogenousTSLP increases T cell activation, levels of proinflamma-tory mediators, autoimmune arthritis, and joint destruc-tion in T cell–driven experimental arthritis. To ourknowledge, this is also the first study to demonstrate thatthe prevention of TSLPR-induced immune activation isaccompanied by a reduction in Th17-associated autoim-munity that causes the prevention of arthritis and jointdestruction. Together, these data indicate that increasedexpression of, and immune activation by, TSLP signifi-cantly contributes to the immunopathology in RA. Thissuggests that targeting TSLP or its receptor representsnovel strategies for the treatment of RA, but presumablyalso for other autoimmune diseases.

ACKNOWLEDGMENTS

Anh Leith and Dina Alcorn (Amgen) are kindly ac-knowledged for their technical assistance. We thank Prof. ErikHack for critical reading and suggestions on the manuscript.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising itcritically for important intellectual content, and all authors approvedthe final version to be published. Dr. van Roon 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. Hartgring, Willis, Dean, Broere, vanEden, Bijlsma, Lafeber, van Roon.Acquisition of data. Hartgring, Willis, Dean.Analysis and interpretation of data. Hartgring, Willis, Dean, vanRoon.

REFERENCES

1. Feldmann M, Brennan FM, Maini RN. Rheumatoid arthritis. Cell1996;85:307–10.

2. Chabaud M, Durand JM, Buchs N, Fossiez F, Page G, Frappart L,et al. Human interleukin-17: a T cell–derived proinflammatorycytokine produced by the rheumatoid synovium. Arthritis Rheum1999;42:963–70.

3. Morita Y, Yamamura M, Kawashima M, Harada S, Tsuji K,Shibuya K, et al. Flow cytometric single-cell analysis of cytokineproduction by CD4� T cells in synovial tissue and peripheralblood from patients with rheumatoid arthritis. Arthritis Rheum1998;41:1669–76.

4. Yudoh K, Matsuno H, Nakazawa F, Yonezawa T, Kimura T.Reduced expression of the regulatory CD4� T cell subset isrelated to Th1/Th2 balance and disease severity in rheumatoidarthritis. Arthritis Rheum 2000;43:617–27.

5. 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.

6. Edwards JC, Szczepanski L, Szechinski J, Filipowicz-Sosnowska A,Emery P, Close DR, et al. Efficacy of B cell–targeted therapy with

TSLP AND ITS RECEPTOR PROMOTE ARTHRITIS 1885

rituximab in patients with rheumatoid arthritis. N Engl J Med2004;350:2572–81.

7. Mulherin D, FitzGerald O, Bresnihan B. Synovial tissue macro-phage populations and articular damage in rheumatoid arthritis.Arthritis Rheum 1996;39:115–24.

8. Kremer JM, Westhovens R, Leon M, Di Giorgio E, Alten R,Steinfeld S, et al. Treatment of rheumatoid arthritis by selectiveinhibition of T cell activation with fusion protein CTLA4Ig. N EnglJ Med 2003;349:1907–15.

9. Olsen NJ, Stein CM. New drugs for rheumatoid arthritis. N EnglJ Med 2004;350:2167–79.

10. Ding C, Cicuttini F, Li J, Jones G. Targeting IL-6 in the treatmentof inflammatory and autoimmune diseases. Expert Opin InvestigDrugs 2009;18:1457–66.

11. Van Roon JA, Glaudemans KA, Bijlsma JW, Lafeber FP. Inter-leukin 7 stimulates tumour necrosis factor � and Th1 cytokineproduction in joints of patients with rheumatoid arthritis. AnnRheum Dis 2003;62:113–9.

12. Van Roon JA, Hartgring SA, Wenting-van Wijk M, Jacobs KM,Tak PP, Bijlsma JW, et al. Persistence of interleukin 7 activity andlevels on tumour necrosis factor � blockade in patients withrheumatoid arthritis. Ann Rheum Dis 2007;66:664–9.

13. Van Roon JA, Verweij MC, Wenting-van Wijk M, Jacobs KM,Bijlsma JW, Lafeber FP. Increased intraarticular interleukin-7 inrheumatoid arthritis patients stimulates cell contact–dependentactivation of CD4� T cells and macrophages. Arthritis Rheum2005;52:1700–10.

14. Hartgring SA, van Roon JA, Wenting-van Wijk M, Jacobs KM,Jahangier ZN, Willis CR, et al. Elevated expression of interleu-kin-7 receptor in inflamed joints mediates interleukin-7–inducedimmune activation in rheumatoid arthritis. Arthritis Rheum 2009;60:2595–605.

15. Soumelis V, Reche PA, Kanzler H, Yuan W, Edward G, Homey B,et al. Human epithelial cells trigger dendritic cell mediated allergicinflammation by producing TSLP. Nat Immunol 2002;3:673–80.

16. Lee HC, Ziegler SF. Inducible expression of the proallergiccytokine thymic stromal lymphopoietin in airway epithelial cells iscontrolled by NF-�B. Proc Natl Acad Sci U S A 2007;104:914–9.

17. Kato A, Favoreto S Jr, Avila PC, Schleimer RP. TLR3- and Th2cytokine-dependent production of thymic stromal lymphopoietinin human airway epithelial cells. J Immunol 2007;179:1080–7.

18. Kouzaki H, O’Grady SM, Lawrence CB, Kita H. Proteases induceproduction of thymic stromal lymphopoietin by airway epithelialcells through protease-activated receptor-2. J Immunol 2009;183:1427–34.

19. Park LS, Martin U, Garka K, Gliniak B, Di Santo JP, Muller W,et al. Cloning of the murine thymic stromal lymphopoietin (TSLP)receptor: formation of a functional heteromeric complex requiresinterleukin 7 receptor. J Exp Med 2000;192:659–70.

20. Reche PA, Soumelis V, Gorman DM, Clifford T, Liu M, Travis M,et al. Human thymic stromal lymphopoietin preferentially stimu-lates myeloid cells. J Immunol 2001;167:336–43.

21. Allakhverdi Z, Comeau MR, Jessup HK, Yoon BR, Brewer A,Chartier S, et al. Thymic stromal lymphopoietin is released byhuman epithelial cells in response to microbes, trauma, or inflam-mation and potently activates mast cells. J Exp Med 2007;204:253–8.

22. Rochman I, Watanabe N, Arima K, Liu YJ, Leonard WJ. Cuttingedge: direct action of thymic stromal lymphopoietin on activatedhuman CD4� T cells. J Immunol 2007;178:6720–4.

23. Nagata Y, Kamijuku H, Taniguchi M, Ziegler S, Seino K. Differ-ential role of thymic stromal lymphopoietin in the induction ofairway hyperreactivity and Th2 immune response in antigen-induced asthma with respect to natural killer T cell function. IntArch Allergy Immunol 2007;144:305–14.

24. Al-Shami A, Spolski R, Kelly J, Fry T, Schwartzberg PL, Pandey

A, et al. A role for thymic stromal lymphopoietin in CD4� T celldevelopment. J Exp Med 2004;200:159–68.

25. Carpino N, Thierfelder WE, Chang MS, Saris C, Turner SJ,Ziegler SF, et al. Absence of an essential role for thymic stromallymphopoietin receptor in murine B cell development. Mol CellBiol 2004;24:2584–92.

26. Watanabe N, Hanabuchi S, Soumelis V, Yuan W, Ho S, de WaalMalefyt R, et al. Human thymic stromal lymphopoietin promotesdendritic cell–mediated CD4� T cell homeostatic expansion. NatImmunol 2004;5:426–34.

27. Liu YJ. TSLP in epithelial cell and dendritic cell cross talk. AdvImmunol 2009;101:1–25.

28. Ying S, O’Connor B, Ratoff J, Meng Q, Mallett K, Cousins D, etal. Thymic stromal lymphopoietin expression is increased in asth-matic airways and correlates with expression of Th2-attractingchemokines and disease severity. J Immunol 2005;174:8183–90.

29. Zhou B, Comeau MR, De Smedt T, Liggitt HD, Dahl ME, LewisDB, et al. Thymic stromal lymphopoietin as a key initiator ofallergic airway inflammation in mice. Nat Immunol 2005;6:1047–53.

30. Demehri S, Morimoto M, Holtzman MJ, Kopan R. Skin-derivedTSLP triggers progression from epidermal-barrier defects toasthma. PLoS Biol 2009;7:e1000067.

31. Koyama K, Ozawa T, Hatsushika K, Ando T, Takano S, Wako M,et al. A possible role for TSLP in inflammatory arthritis. BiochemBiophys Res Commun 2007;357:99–104.

32. Watanabe N, Hanabuchi S, Marloie-Provost MA, Antonenko S,Liu YJ, Soumelis V. Human TSLP promotes CD40 ligand-inducedIL-12 production by myeloid dendritic cells but maintains theirTh2 priming potential. Blood 2005;105:4749–51.

33. Miossec P, Korn T, Kuchroo VK. Interleukin-17 and type 17helper T cells. N Engl J Med 2009;361:888–98.

34. 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.

35. 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.

36. Hartgring SA, Willis CR, Alcorn D, Nelson LJ, Bijlsma JW,Lafeber FP, et al. Blockade of the IL-7 receptor inhibits collagen-induced arthritis and is associated with reduction of T cell activityand proinflammatory mediators. Arthritis Rheum 2010;62:2716–25.

37. 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.

38. Van Roon JA, Bijlsma JW, Lafeber FP. Diversity of regulatory Tcells to control arthritis. Best Pract Res Clin Rheumatol 2006;20:897–913.

39. Lubberts E, Joosten LA, Chabaud M, van den Bersselaar L,Oppers B, Coenen-de Roo CJ, et al. IL-4 gene therapy for collagenarthritis suppresses synovial IL-17 and osteoprotegerin ligand andprevents bone erosion. J Clin Invest 2000;105:1697–710.

40. Taylor BC, Zaph C, Troy AE, Du Y, Guild KJ, Comeau MR, et al.TSLP regulates intestinal immunity and inflammation in mousemodels of helminth infection and colitis. J Exp Med 2009;206:655–67.

41. Boldizsar F, Tarjanyi O, Nemeth P, Mikecz K, Glant TT. Th1/Th17polarization 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.

42. Moret FM, Hack CE, van der Wurff-Jacobs KM, Lafeber FP, vanRoon JA. Increased intraarticular levels of IL-7-related TSLPactivate myeloid dendritic cells to stimulate Th1 and Th17 activity

1886 HARTGRING ET AL

in patients with rheumatoid arthritis [abstract]. Ann Rheum Dis2010;69 Suppl 3:323.

43. Lubberts E, Koenders MI, Oppers-Walgreen B, van den Bersse-laar L, Coenen-de Roo CJ, Joosten LA, et al. Treatment with aneutralizing anti-murine interleukin-17 antibody after the onset ofcollagen-induced arthritis reduces joint inflammation, cartilagedestruction, and bone erosion. Arthritis Rheum 2004;50:650–9.

44. Lubberts E. IL-17/Th17 targeting: on the road to prevent chronicdestructive arthritis? Cytokine 2008;41:84–91.

45. Iezzi G, Sonderegger I, Ampenberger F, Schmitz N, Marsland BJ,Kopf M. CD40–CD40L cross-talk integrates strong antigenicsignals and microbial stimuli to induce development of IL-17-producing CD4� T cells. Proc Natl Acad Sci U S A 2009;106:876–81.

46. Ogura H, Murakami M, Okuyama Y, Tsuruoka M, Kitabayashi C,

Kanamoto M, et al. Interleukin-17 promotes autoimmunity bytriggering a positive-feedback loop via interleukin-6 induction.Immunity 2008;29:628–36.

47. Chen X, Das R, Komorowski R, Beres A, Hessner MJ, Mihara M,et al. Blockade of interleukin-6 signaling augments regulatory Tcell reconstitution and attenuates the severity of graft-versus-hostdisease. Blood 2009;114:891–900.

48. Tanaka J, Watanabe N, Kido M, Saga K, Akamatsu T, Nishio A,et al. Human TSLP and TLR3 ligands promote differentiation ofTh17 cells with a central memory phenotype under Th2-polarizingconditions. Clin Exp Allergy 2009;39:89–100.

49. Van Eden W, van der Zee R, Taams LS, Prakken AB, van RoonJ, Wauben MH. Heat-shock protein T cell epitopes trigger aspreading regulatory control in a diversified arthritogenic T cellresponse. Immunol Rev 1998;164:169–74.

TSLP AND ITS RECEPTOR PROMOTE ARTHRITIS 1887