of June 5, 2018.This information is current as

Osteoclast Development and FunctionActivated Invariant NKT Cells Regulate

Horwood, Irene A. G. Roberts and Anastasios KaradimitrisKotsianidis, Alan Boyde, Graham R. Williams, NikkiHowell, Ke Xu, Emmanouil Spanoudakis, Ioannis Ming Hu, J. H. Duncan Bassett, Lynett Danks, Peter G. T.

http://www.jimmunol.org/content/186/5/2910doi: 10.4049/jimmunol.1002353January 2011;

2011; 186:2910-2917; Prepublished online 28J Immunol 

Referenceshttp://www.jimmunol.org/content/186/5/2910.full#ref-list-1

, 20 of which you can access for free at: cites 50 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2011 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on June 5, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Activated Invariant NKT Cells Regulate OsteoclastDevelopment and Function

Ming Hu,* J. H. Duncan Bassett,†,1 Lynett Danks,‡,1 Peter G. T. Howell,x Ke Xu,*

Emmanouil Spanoudakis,{ Ioannis Kotsianidis,{ Alan Boyde,‖ Graham R. Williams,†

Nikki Horwood,‡ Irene A. G. Roberts,* and Anastasios Karadimitris*

Invariant NKT (iNKT) cells modulate innate and adaptive immune responses through activation of myeloid dendritic cells and

macrophages and via enhanced clonogenicity, differentiation, and egress of their shared myeloid progenitors. Because these same

progenitors give rise to osteoclasts (OCs), which also mediate the egress of hematopoietic progenitors and orchestrate bone remod-

eling, we hypothesized that iNKT cells would extend their myeloid cell regulatory role to the development and function of OCs. In

this study, we report that selective activation of iNKT cells by a-galactosylceramide causes myeloid cell egress, enhances OC

progenitor and precursor development, modifies the intramedullary kinetics of mature OCs, and enhances their resorptive

activity. OC progenitor activity is positively regulated by TNF-a and negatively regulated by IFN-g, but is IL-4 and IL-17

independent. These data demonstrate a novel role of iNKT cells that couples osteoclastogenesis with myeloid cell egress in

conditions of immune activation. The Journal of Immunology, 2011, 186: 2910–2917.

Invariant NKT (iNKT) cells are a subset of highly conservedimmunoregulatory T cells that possess a characteristic in-variant TCRVa14Ja18 chain and are activated by glycolipid

ligands presented by the nonpolymorphic MHC class I-like mol-ecule Cd1d (1, 2). Underpinning these characteristics, TCRJa182/2

and Cd1d2/2 mice lack development of iNKT cells (3, 4).iNKT cells regulate development and function of macrophages

and monocytes (5) as well as myeloid dendritic cells (mDC) (6, 7)and are required for effective myelopoiesis under physiologic con-ditions (8). Selective activation of iNKT cells by the glycolipidligand a-galactosylceramide (aGC) promotes myeloid progenitorclonogenicity, differentiation, and egress of myeloid cells from thebone marrow (BM) to the periphery (8, 9) as well as secondary acti-vation of many aspects of the innate and adaptive immune systems.

Osteoclasts (OCs), along with osteoblasts, play an essential rolein bone remodeling (10). In addition, OCs—when activated inresponse to various stimuli, including receptor activator of NFkBligand (RANKL)—facilitate egress of hematopoietic progenitorand mature cells from the BM to the periphery (11, 12). RANKL,together with M-CSF, is also essential for OC development fromCFU macrophages, the same myeloid progenitors that give rise tomonocytes-macrophages and mDCs (13–15). However, whetheriNKT cells promote the activation of OCs as well as macrophagesand mDCs, perhaps via effects on their common progenitor, is notknown. We therefore investigated the hypothesis that iNKT cells,under conditions of immune activation, are able to regulate thedevelopment and function of OCs and their progenitors.

Materials and MethodsAnimals

iNKT cell-deficient TCRJa182/2 (3), Cd1d12/2 (4), and TNFaRp552/2

(16) mice have been described and were bred with wild type (WT) mice inthe Medical Research Council Imperial College Animal Facility, Ham-mersmith Hospital, London. BALB/c IFN-g2/2 and IL-42/2 mice wereprovided by A. O’Garra (National Institute of Medical Research, London,U.K.). Approval for the use of mice was obtained under the authority ofa U.K. Home Office Project License.

Reagents

aGC (provided by Stephan Gadola, University of Southampton, South-ampton, U.K.) was dissolved in PBS at 200 mg/ml and stored at 4˚C.Mouse macrophage colony stimulating factor (mC-CSF) and mRANKLwere obtained from Peptrotech, and anti-mouse IL-17 blocking Ab andIgG control were obtained from R&D Systems (Abingdon, U.K.).

In vivo assays

For in vivo activation of iNKT cells, 2 mg aGC diluted in 200 ml PBS wasinjected i.p. into each mouse. Anti–IL-17 blocking Ab or IgG isotypecontrol (50 mg per mouse) were administered by tail vein injection 3 hprior to aGC injection.

Flow cytometry

BM cells were stained with anti–CD115-PE, anti–CD117-allophycocyanin,anti–CD3-FITC, anti–B220-FITC, and anti–Mac1-FITC (from BD Phar-mingen) using standard protocols. Data acquisition was performed on a

*Center for Hematology, Hammersmith Hospital, Imperial College London, LondonW12 0NN, United Kingdom; †Molecular Endocrinology Group, Department of Med-icine, Medical Research Council Clinical Sciences Centre, Imperial College London,Hammersmith Hospital, London W12 0NN, United Kingdom; ‡Kennedy Institute ofRheumatology, Faculty of Medicine, Imperial College London, London W6 8LH,United Kingdom; xDivision of Restorative Dental Sciences, Eastman Dental Institute,University College London, London WC1X 8LD, United Kingdom; {Departmentof Hematology, Democritus University of Thrace Medical School, Alexandroupolis68100, Greece; and ‖Centre for Oral Growth and Development, Institute of Dentistry,Bart’s and The London School of Medicine, Queen Mary University of London,London E1 2AT, United Kingdom

1J.H.D.B. and L.D. contributed equally to this work.

Received for publication July 13, 2010. Accepted for publication December 23, 2010.

This work was supported by Leukaemia and Lymphoma Research (to M.H., E.S., K.X.,I.K., I.A.G.R., and A.K.), the Medical Research Council (to J.H.D.B. and G.R.W.), andthe Arthritis Research Campaign (to L.D. and N.H.).

Address correspondence and reprint requests to Anastasios Karadimitris, Centre forHaematology, Imperial College London, Hammersmith Hospital, Imperial CollegeHealthcare NHS Trust, Du Cane Road, London W12 0NN, United Kingdom. E-mailaddress: a.karadimitris@imperial.ac.uk

Abbreviations used in this article: BM, bone marrow; BSE-SEM, backscattered elec-tron scanning electron microscopy; BV, bone volume; CTX, C-terminal telopeptidesof collagen type II; aGC, a-galactosylceramide; iNKT, invariant NKT; mDC, mye-loid dendritic cell; mMCS-F, mouse macrophage colony stimulating factor; OC,osteoclast; OCP, osteoclast progenitor; qBSE, quantitative BSE; RANKL, receptoractivator of NFkB ligand; TRAP, tartrate-resistant acid phosphatase; TV, tissue vol-ume; WT, wild type.

Copyright� 2011 by TheAmericanAssociation of Immunologists, Inc. 0022-1767/11/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1002353

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

two-laser FACSCalibur (Becton Dickinson), and analysis was performedwith FlowJo software.

Histology and histomorphometry

Mouse lower limbs were fixed for 24 h in 10% neutral buffered formalin anddecalcified in 10% EDTA, pH 7.4, for 14 d. Decalcified bones were em-bedded in paraffin, and 3-mm longitudinal midline sections were cut ontoSuperfrost plus slides (VWR), deparaffinized in xylene, and rehydrated. Ateach 80-mm level, serial sections were stained with H&E and Massontrichrome, and OC numbers were determined after staining sections fortartrate-resistant acid phosphatase (TRAP) activity. Staining using a SigmaTRAP kit (386A-1KT) was performed according to manufacturer’s in-structions, except that the incubation time was 4 h. Sections were photo-graphed using an Olympus BX51 microscope and DP71 digital camera.OC numbers and total bone surface were determined using OsteomeasureXP v1.2 software of a 1-mm2 area of the secondary spongiosa of the pro-ximal tibia, 500-mm below the growth plate.

Analysis of growth plate dimensions

Mouse lower limbs were fixed for 24 h in 10% neutral buffered formalin anddecalcified in 10% EDTA (pH 7.4) for 28 d. Decalcified bones were em-bedded in paraffin and 3-mm longitudinal midline sections. Sections werestained with van Gieson and alcian blue. Reserve zone, prehypertrophiczone, hypertrophic zone, and total growth plate heights were determined atfour separate positions across the growth plates using digital images ofproximal tibia sections (17). Results from two different levels of sectioningwere compared to ensure consistency of the data.

Faxitron point projection microradiographic analysis

Upper limbs and tail vertebraewere stored in 70% ethanol at 4˚C. Soft tissuewas removed and 10-mm–resolution scale images were obtained usinga Faxitron MX20 point projection x-ray source and digital imaging systemoperating at 26 kV and 35 magnification (Qados, Cross Technologies,Sandhurst, Berkshire, U.K.). Magnifications were calibrated by imaginga digital micrometer, and long bone lengths were determined using ImageJ1.41 software (http://rsb.info.nih.gov/ij/). Cortical bone thickness and di-ameter were determined in at least 10 locations in the middistal humerusdiaphysis (Fig. 6). The relative mineral content of calcified tissues wasdetermined using ImageJ software by comparison with three standardsincluded in each image frame (18). Increasing gradations of mineralizationdensity were represented in 16 equal intervals by a pseudocolor scheme forpresentation of digital images. The distributions of mineralization densitieswere analyzed by the Kolmogorov-Smirnov test.

Backscattered electron scanning electron microscopy

Analysis of bone microarchitecture was determined by backscatteredelectron scanning electronmicroscopy (BSE-SEM). Freshly dissected lowerlimbs were fixed and stored in 70% ethanol. Bones were opened longitu-dinally and cleaned by maceration with an alkaline bacterial pronase.Samples were coated with carbon and imaged using backscattered electrons,generally at a 20-kV beam potential. The images provide the best and mostdetailed view of bone surfaces, from which surface activity states (forming,resting, resorbing, resorbed) can be investigated. Cortical and trabecularbone surfaces were analyzed, and the fraction of bone surface with evidenceof osteoclastic resorption pits was determined (17).

The distribution of skeletal micromineralization densities was examinedby BSE-SEM and quantified by digital image analysis at the subcubicmicron-resolution scale (quantitative BSE [qBSE]). Fixed femurs wereembedded in poly-methyl-methacrylate. Block faces were cut through themidline of specimens, which were then polished, coated with carbon, andanalyzed by BSE-SEM (17). The mineralization densities of trabecular andcortical bone were determined by comparison with halogenated dimetha-crylate standards. Increasing gradations of micromineralization densitywere represented in eight equal intervals by a pseudocolor scheme forpresentation of digital images. Femur and caudal vertebra bone volume(BV) as a proportion of tissue volume (BV/TV) was quantified in qBSEimages using ImageJ (17).

In vitro OC assays

BM cells flushed from the femora and tibia of mice were cultured ina-MEM containing 10% FBS supplemented with mM-CSF (5 ng/ml)overnight. Nonadherent cells washed once with PBS were seeded at 0.5 3106 per well in a 48-well plate in the presence of 10 ng/ml mM-CSF and 50ng/ml mRANKL. Medium with fresh cytokines was changed every 2–3 d until mature OCs developed (usually on day 4).

ELISAs

ELISA kits for mRANKL, IFN-g, and IL-4 were obtained from R&DSystems Europe (Abingdon, U.K.). The Ratlaps C-terminal telopeptides ofcollagen type II (CTX) ELISA kit was obtained from Nordic BioscienceDiagnostics (Herlev, Denmark). ELISA assays were performed accordingto the manufacturer’s instructions.

Quantitative RT-PCR

BM cells from aGC-injected mice were flushed from femurs and tibias atdifferent time points after aGC injection and lysed in Trizol solution. TotalRNA was extracted according to the manufacturer’s instructions (Invitrogen,Paisley, U.K.). One microgram of total RNA was reverse-transcribed intocDNAs using the Superscript III cDNA synthesis kit (Invitrogen). TaqManreal-time quantitative PCR was performed using TNF-a–specific primer-probe (Assay ID: Mm00443258_m1; Applied Biosystems, Warrington, U.K.).

Statistics

Normally distributed data were analyzed by Student t test; p values ,0.05were considered significant. Relative and cumulative frequency histogramsof micromineralization densities from WT and Cd1d2/2 mice were ana-lyzed using the Kolmogorov-Smirnov test.

ResultsIn vivo activated iNKT cells promote OC progenitor expansion

Activation of iNKT cells with aGC in vivo leads to rapid egress ofmyeloid progenitors and mature forms into peripheral blood (9),representing an important aspect of the ability of activated iNKT cellsto recruit effectors of the cellular innate immune system. In agree-ment with previous data (9), we found that within 3 h of aGC ad-ministration, the frequency of Mac+Gr1+ myeloid cells decreased inBM and concomitantly increased in peripheral blood (Fig. 1A, 1B).OC progenitors (OCPs) are myeloid cells that have the capacity

to differentiate to granulocytes, monocytes, and monocyte-derivedmDCs (13). The same progenitors develop into OCs under thedrive of RANKL (13, 19). It was shown previously that BM CD32

B2202CD11b2c-fmshic-Kithi cells are highly enriched in OCPs asassessed by in vitro, M-CSF- and RANKL-dependent assays (13,19). We confirmed that, in the presence of M-CSF and RANKL,flow-sorted, highly purified CD32B2202CD11b2c-fmshic-Kithi

cells form a large number of TRAP+, mature OCs displayingpolymerized F-actin upon staining with phalloidin. By contrast,CD32B2202CD11b2c-fmsnegc-Kithi/neg cells formed no, or veryfew, mature OCs (Fig. 1C, 1D).Because the CD32B2202CD11b2c-fmshic-Kithi cells were shown

to be bona fide OCPs, we tested whether aGC-activated iNKT cellswould promote the development of BM OCPs and mature OCs. Wefound that 3 d post-treatment of WT mice with aGC, OCP cells wereincreased by 5.8-fold (p, 0.01) compared with PBS-treated controls(Fig. 2A, 2B). This finding was an iNKT cell- and CD1d-dependenteffect, because the increase in OCP frequency in response to aGCwas not observed in TCRJa182/2 or Cd1d2/2 mice (Fig. 2A, 2B).

In vivo activated iNKT cells lead to accelerated maturationand activation of OCs

To test whether increased OCP frequency induced by aGC led toenhanced OC maturation and activation, BM cells from WT micewere collected and plated in vitro in the presence of M-CSF andRANKL 72 h after aGC injection. Consistent with the in vivoresults, day 3 TRAP+ multinucleated and day 4 mature OCs de-rived from aGC-treated mice developed in vitro in higher num-bers, suggesting increased OC clonogenicity and accelerated dif-ferentiation. This effect was also iNKT cell-dependent because itwas not observed in TCRJa182/2 and Cd1d2/2 mice. (Fig. 2C,2D). During the same period, serum CTX levels, a specific markerof in vivo OC activation and bone resorption, were increased by30% and remained increased in aGC-treated WT mice over a

The Journal of Immunology 2911

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

period of 72 h (p , 0.01 at 72 h; Fig. 2E) while remaining un-changed between baseline and 72 h in TCRJa182/2 and Cd1d2/2

mice (Fig. 2F). Thus, aGC promotes in vivo osteoclastogene-sis under conditions of immune activation in an iNKT cell-dependent manner.

Intramedullary kinetics of in vivo activated mature OCs

To visualize mature OC development in the BM after aGC ad-ministration, TRAP staining in histologic sections of BM wasperformed. This staining showed that although the overall numberof OCs did not change during a 72-h period after injection of aGC(data not shown), within 16 h and maximally at 24 h (Fig. 3A, 3B)the numbers of OCs attached to the trabecular and periosteal bonesurfaces (OC.S/BSA in Fig. 3B, left) were decreased. Interestingly,their numbers increased by ∼5-fold in the intramedullary space at24 h and returned to baseline levels at 72 h (Fig. 3B, right). Giventhe increase in OCPs and their accelerated maturation and increasein CTX levels (Fig. 2E), these observations suggest that aGCprovokes accelerated turnover of overactive OCs as well as their

redistribution from the periosteal and trabecular surfaces into theBM cavity. Therefore, aGC-dependent immune activation co-ordinately couples differentiation and egress of myeloid cells withenhanced OC development, altered intramedullary localization, andincreased resorptive activity of mature OCs, indicating a central roleof iNKT cells in linking myelopoiesis with OC function.

IFN-g is a negative regulator of iNKT cell-induced OCdevelopment

RANKL is the most potent pro-osteoclastogenic cytokine. Highlevels of RANKL as result of T cell activation lead to OC activationand bone destruction in experimental and clinical arthritis (20).We measured serum levels of RANKL in response to aGC over a

FIGURE 1. In vivo activated iNKT cells promote mature myeloid cell

egress and OCP expansion. A, Myeloid cell egress in response to aGC.

Mac+Gr-1+ mature myeloid cells in BM and peripheral blood (PB) as

assessed by flow cytometry at baseline and 3 h after aGC injection in WT

mice. While a 2.5-fold increase in the frequency of myeloid cells in PB is

observed at 3 h, their frequency in BM decreases by 2-fold, suggesting

a rapid egress of myeloid cells in response to aGC. B, Cumulative data of

the frequency of myeloid cells (data shown as mean 6 SD; n = 3 mice at

each time-point). C, Flow cytometric identification of BM OCPs as c-fmshi

c-Kithi (right dot blot, G1) gated on CD32B2202CD11b2 cells (left dot

blot). D, In vitro clonogenicity and generation of mature OCs from highly

purified CD32B2202CD11b2c-fmshic-Kithi cells in response to M-CSF

and RANKL. The left panel shows mature OCs identified either by TRAP

staining or phalloidin staining to demonstrate actin polymerization. The

right panel shows the number of mature OCs 4 d after plating 6500 CD32

B2202CD11b2c-fmshic-Kithi or 6,500 and 65,000 cells as shown in G2 in

Fig. 2C. Data are shown as mean 6 SD of triplicate wells and represen-

tative of two independent experiments.

FIGURE 2. aGC-induced maturation and activation of OCs. A, Flow

cytometric analysis of BM OCPs 72 h after WT C57BL/6 mice received

a single aGC (2 mg) or PBS injection i.p. OCPs are identified as c-fmshic-

kithi gated on CD32B2202CD11b2 cells. B, Frequency of OCPs in

C57BL/6 WT, TCRJa182/2, and Cd1d2/2 mice 72 h after treatment with

PBS or aGC (n = 10 for WT; n = 4 for TCRJa182/2 and Cd1d2/2 mice).

Data are shown as mean 6 SD. **p , 0.01 by Student t test. C, In vitro

assay of OCs derived from BM of WT mice injected with either PBS or

aGC. BM cells were obtained from mice 72 h after injection and plated in

the presence of M-CSF and RANKL. On days 3 and 4 of the culture, cells

were stained with TRAP to determine multinucleated and mature OCs.

More and larger multinucleated and mature OCs were observed in the

aGC-treated group. D, Total numbers of day 3 and day 4 mature and

multinucleated OCs generated from in vitro differentiation of BM cells

obtained from WT, TCRJa182/2, and Cd1d2/2 mice treated with PBS or

aGC. One representative of three experiments is shown; data are shown as

mean 6 SD of triplicates assays. E, Serum CTX levels at baseline and

different time points after aGC injection into WT mice (n = 6 for 0 and 72

h; n = 2 for 16 and 24 h). Data are shown as mean 6 SD. **p , 0.01 by

Student t test. F, Serum CTX levels at baseline and at 72 h after aGC

injection into TCRJa182/2 and Cd1d2/2 mice (n = 3 for each time point).

2912 ACTIVATED iNKT CELLS REGULATE OSTEOCLASTS

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

72-h period. Serum RANKL levels were the same in aGC-treatedWT and PBS-treated animals (Fig. 4A), suggesting that solubleRANKL is not a major mediator of the pro-osteoclastogenic effectof aGC-activated iNKT cells.We therefore tested the role of IFN-g and IL-4, the two ar-

chetypal Th1 and Th2 cytokines, in this process. First, we mea-sured serum levels, and in agreement with previous reports (7, 21)we found that a single injection of aGC leads to a rapid secretionof IL-4 and IFN-g in the first 3–6 h in WT but not TCRJa182/2

mice (Fig 4B, 4C). To investigate the relative contribution of IFN-g and IL-4, we injected aGC into WT, IL-42/2, and IFN-g2/2

BALB/c mice. We found that in each strain, OCP frequency wassignificantly higher in mice treated with aGC compared with PBScontrols (p , 0.01 for aGC versus PBS in all three strains).However, although the fold increment of OCP frequency in re-sponse to aGC was similar between WTand IL-42/2mice (IL-42/2,2.1-fold versus WT, 2.3-fold; p . 0.05; Fig. 4D, 4E), it wassignificantly higher in IFN-g2/2 mice (IFN-g2/2, 6.2-fold versusWT, 2.3-fold; p , 0.01; Fig. 4D, 4E), suggesting that IFN-g is animportant negative regulator of the pro-osteoclastogenic effect ofactivated iNKT cells.

TNF-a is a positive regulator of iNKT cell-induced OCdevelopment

In search of positive regulators of iNKT cell-mediated enhance-ment of OC development, we tested the roles of IL-17 and TNF-a.

IL-17, the prototypic Th17 cytokine, was shown recently to besecreted by activated iNKT cells (22) and to promote OC de-velopment (23, 24). Similarly, TNF-a is a strongly pro-osteo-clastogenic cytokine (25, 26).We found that OCP frequency in C57BL/6WTmice treated with

aGC plus anti-IL-17 Ab [previously shown to effectively neu-

tralize IL-17 in vivo (27)] was not different to aGC-plus-IgG

control-treated mice (Fig. 5A, 5B) indicating that IL-17 is not

a major regulator of OC development in this context. To test the

role of TNF-a, mice lacking the TNF-aR p55 subunit were

injected with aGC. The p55 but not the p75 subunit of the TNF-

aR was previously shown to mediate the pro-osteoclastogenic

effect of TNF-a (25, 26). We found that, although in response

to aGC, the frequency of OCPs in TNF-aR2/2 mice rose by 2.3-

fold. Compared with PBS controls, the increment was signifi-

cantly lower compared with C57BL/6 WT controls (TNF-aR2/2,

2.3-fold versus WT, 6-fold; p , 0.01; Fig. 5C, 5D), suggesting

that TNF-a is indeed a major mediator of the pro-osteoclastogenic

effect of aGC-activated iNKT cells. Consistent with these results,

aGC injection triggered an up to 10-fold increase of TNF-a

mRNA production in total BM cells (Fig. 5E).These findings suggest that aGC-induced osteoclastogenesis

is independent of IL-4 and IL-17, whereas IFN-g and TNF-a

are negative and positive regulators of this process, respec-

tively.

FIGURE 3. In vivo kinetics of OC development and

activity in response to aGC. A, TRAP staining of BM

sections 24 h after WT mice receiving PBS control or

aGC. Sections of trabecular bone near the growth plate

(original magnification 310), periosteum (original

magnification 320), and medullary cavity (original

magnification 340) are shown, in addition to insets

(original magnification, 340). TRAP+ OCs are indi-

cated by arrows. In the mouse receiving aGC, TRAP+

OCs appear to have vacated the trabecular and peri-

osteal surfaces and accumulated in the BM cavity. B,

Left graph shows OC surface per bone surface (OCs/

BS) at the indicated time points (n = 3 mice for each

time point). The right panel shows total TRAP+ OCs

within the BM cavity (TRAP+ in BM).

The Journal of Immunology 2913

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Skeletal phenotype of Cd1d2/2 mice

iNKT cells are known to be autoreactive—that is, they are acti-vated by endogenous ligands in steady state in the absence ofexogenous glycolipid Ags (28). It is possible that this backgroundiNKT activation might be important also for steady state OC de-velopment and activation; if so, then chronic lack of iNKT cellsmight result in perturbed skeletal phenotype and bone development. To address this prediction, we assessed skeletal development and adult bone structure in Cd1d2/2 mice using severalindependent specific and sensitive methods (Fig. 6).First, as shown by histologic analysis of the growth plates,

endochondral ossification is not affected in 11-wk-old femaleCd1d2/2 mice as compared with age- and gender-matched WTcontrols (Fig. 6A).Furthermore, Faxitron radiographic analysis showed that linear

growth, diaphyseal diameter, cortical bone thickness and adult bonemorphology in Cd1d2/2 mice are similar to WT controls (Fig.6B). These data are consistent with the in vitro OC developmentassays and serum CTX levels in Cd1d2/2 and TCR Ja182/2 mice,which were not different in WT mice (Fig. 2). However, furtheranalysis indicated that bone mineral content was increased in adultCd1d2/2 mice (Fig. 6B). Because cortical bone thickness wassimilar in WT and Cd1d2/2 mice, the increased bone mineral

content determined by Faxitron analysis suggested a phenotype ofincreased cortical bone mineralization.To investigate bone structure in detail, we first analyzed bone

microarchitecture by BSE-SEM. As shown in Fig. 6C, Cd1d2/2

mice had increased trabecular BV in long bones (23% increase inBV fraction [BV/TV]) compared with WT and vertebrae (19%increase in BV/TV).qBSE-SEM analysis demonstrated increased cortical bone min-

eralization density in femurs from Cd1d2/2 mice (Fig. 6E, 6F),which was accompanied by reduced osteoclastic resorption sur-faces on endosteal (35% lower in Cd1d2/2 mice) and trabecular(26% reduced) bone, although these reductions did not reachstatistical significance.These detailed studies indicate that, although Cd1d2/2 mice

have normal endochondral ossification, the increased bone min-eralization and BV result from low bone turnover with impairedosteoclastic bone resorption, which is consistent with a role foriNKT cells in OC development and activation even under basalconditions.

DiscussioniNKT cells have the unique property of modulating immuneresponses by recruiting and activating innate cellular components,including NK cells and mature myeloid cells such as monocytes-macrophages and granulocytes (2, 9, 29). Our work adds OCs,a specialized form of macrophage, to the cellular circuit that ismobilized and activated by iNKT cells to promote immediate andeffective innate immune responses.Our findings suggest that activated iNKT cells act on OC de-

velopment, maturation, and activation. Indeed, the frequency of

FIGURE 4. Role of IFN-g and IL-4 in aGC-enhanced OC development.

A, Serum RANKL in response to aGC. Serum RANKL was measured at

different time points by ELISA following injection of aGC or PBS into

WT C57BL/6 mice (n = 3 mice per time point). B and C, Serum levels of

IFN-g and IL-4 as determined by ELISA at different time points after aGC

injection in C57BL/6 WT and TCRJa182/2 mice. Values are mean 6 SD

from three mice at each time point. D and E, Frequency of OCP in IFN-g2/2

and IL-42/2 mice at 72 h after treatment with PBS or aGC. Because both

IFN-g2/2 and IL-42/2 mice were on the BALB/c background, BALB/c

WT mice were used as controls (n = 7–9 for each strain). **p , 0.01, ns

(not significant) refers to comparisons of the fold increment between

groups by Student t test.

FIGURE 5. Role of IL-17 and TNF-a in aGC-enhanced OC de-

velopment. A and B, Frequency of OCPs in C57BL/6 WT mice treated

with aGC and anti–mIL-17 blocking Ab. C57BL/6 WT mice were injected

i.v. with isotype IgG or anti–IL-17 blocking Ab 3 h prior to aGC injection.

Flow cytometric analysis of OCPs was performed 72 h later (n = 4). C and

D, Frequency of OCPs in TNF-aRp552/2 and C57BL/6 WT mice 72 h

after treatment with PBS or aGC (n = 4 for each strain). **p , 0.01, ns

(not significant) refers to comparisons of the fold increment between

groups by Student t test. E, TNF-a mRNA levels relative to GAPDH in total

BM cells at the indicated time points after injection of aGC as assessed by

Taqman real-time quantitative PCR (n = 3 mice per time point).

2914 ACTIVATED iNKT CELLS REGULATE OSTEOCLASTS

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

OCPs increases in the BM in response to aGC in an iNKT cell- andCd1d-dependent manner. In addition, the in vitro clonogenicassays reflect enhanced commitment and accelerated differentia-tion toward precursors and mature OCs. The modest but sustainedincrease in serum CTX levels provides an in vivo measurement ofincreased OC activity and bone resorption. Kollet et al. (11, 30)have shown that activated OCs, through their proteolytic activity,contribute significantly to myeloid cell egress and mobilization inresponse to growth factors such as G-CSF, stress such as bleeding,and inflammatory stimuli such as LPS. Previous work has dem-onstrated the ability of aGC-activated iNKT cells, through se-cretion of GM-CSF and IL-3, to mobilize myeloid cells to theperiphery (8, 9)—a finding also observed in our studies. Theseobservations strongly support the hypothesis that the promotion ofOC development and activation by iNKT cells contributes sig-

nificantly to the ability of these cells to mobilize myeloid innateeffectors during the course of an immune response.Direct visualization of the mature OC in the BM in response to

aGC revealed a complex, dynamic picture. First, in contrast to theincreased frequency of OCP, OC clonogenic and resoprtive ac-tivity, overall numbers of TRAP+ cells (i.e., mature osteoclasts)on histological sections were not increased. This apparent dis-crepancy would suggest that aGC induces higher turnover andhyperactivity of OCs that would be supported by their accelerateddevelopment and maturation in vitro (Fig. 2). It was also in-teresting to observe the redistribution of a large proportion ofmature OCs from the bone surfaces to the intramedullary area. Itremains to be determined whether this phenomenon is of func-tional significance and whether it is linked to the role of OCs inthe mobilization of progenitor and mature myeloid forms.

FIGURE 6. Skeletal phenotype of Cd1d2/2 mice. A, Proximal tibia sections from 11-wk-old mice stained with alcian blue (cartilage) and van Gieson

(osteoid; original magnification 3200). Graphs indicate growth plate and chondrocyte zone measurements (left) and proportions relative to total growth

plate height (right). Student t test: Cd1d2/2 versus WT not significant (n = 4). Scale bar, 200 mm. B, Faxitron images of humerus from 11-wk-old WT and

Cd1d2/2 mice. Gray scale images were pseudocolored according to a 16-color palette; low mineralization density is black, high density is white. Upper

graphs show ulna length as determined from grayscale images of whole upper limb and cortical thickness measured as indicated by white arrows. Student t

test: Cd1d2/2 versus WT not significant (n = 4). Lower relative and cumulative frequencies histograms of bone mineralization densities in whole humerus

from WT and Cd1d2/2 mice. Kolmogorov-Smirnov test; ***p, 0.001 Cd1d2/2 versus WT (n = 4). Scale bar, 200 mm. C, Bone microarchitecture in adult

WT and Cd1d2/2 mice. BSE-SEM views of whole femur and tibia (original magnification36) and distal femur (original magnification313) from WT and

Cd1d2/2 mice. Scale bar, 200 mm. D, BSE-SEM images of distal femur and caudal vertebrae from 11-wk-old WT and Cd1d2/2 mice. Graphs show

quantitative analysis of BVas a proportion of TV (BV/TV) in P77 WT and Cd1d2/2 mice. Student t test: *p , 0.05 Cd1d2/2 versus WT (n = 4). Scale bar,

200 mm. E–G, Quantitative BSE-SEM images of distal femur (E), mid diaphyseal cortical bone (F), and metaphyseal trabecular bone (G) from WT and

Cd1d2/2 mice. Grayscale images were pseudocolored according to an eight-color palette; low mineralization density is dark blue, and high density is gray.

Relative and cumulative frequency histograms of micromineralization densities from WT and Cd1d2/2 mice. Kolmogorov-Smirnov test; ***p , 0.001

Cd1d2/2 versus WT (n = 4). Scale bar, 200 mm. H and I, Trabecular (H) and endosteal (I) bone surfaces from 11-wk-old WT and Cd1d2/2 mice imaged by

BSE-SEM (original magnification350). White arrows indicate roughened OC resorption surfaces. Scale bar, 200 mm. Graphs show quantitative analysis of

OC resorption surfaces (percentage of total bone surface). Student t test: Cd1d2/2 versus WT not significant (n = 4). GP, growth plate; HZ, hypertrophic

zone; PZ, proliferative zone; RZ, reserve zone.

The Journal of Immunology 2915

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

aGC-mediated activation of iNKT cells is known to lead toa secondary and nonspecific activation of T cells as well as acti-vation of different components of the innate immune system, suchas macrophages and NK cells. This process is not dissimilar to theeffects of LPS, which are also associated with increased secretionof IFN-g and TNF-a, the latter produced by IFN-g–activatedT cells and macrophages (31).Our results suggest that in response to aGC and activation of

iNKT cells, these two cytokines are major determinants of OCdevelopment. Previous work showed that the net effect of IFN-gon OC development and function is determined by the balance ofits direct inhibitory effect on hematopoiesis and OCPs (32, 33)and its ability to promote osteoclastogenesis through the inductionof RANKL expression on activated T cells (20, 34, 35). Ourfinding that IFN-g is a negative regulator of aGC-promoted os-teoclastogenesis is consistent with previous observations showingthat IFN-g inhibits osteoclastogenesis when nonspecific T cellactivation occurs in the course of an innate immune response,such as in response to LPS (35). In that context, IFN-g inhibitsOC development through ubiquitination and proteasome degrada-tion of TRAF6, an adapter protein required for RANK-dependentNF-kB and Jnk signaling (35).As discussed above, and similar to LPS stimulation, activation

of iNKT cells by aGC results in rapid secretion of IFN-g byiNKT cells and in nonspecific activation and IFN-g secretion by Tand NK cells (2, 29, 36), hence the enhanced frequency of OCP inresponse to aGC in IFN-g2/2 mice.Activation of T cells and macrophages by iNKT cell-derived

IFN-g also leads to secretion of TNF-a (37), a strongly pro-osteoclastogenic cytokine. In an inflammatory context, TNF-aenhances osteoclastogenesis in a RANKL-dependent manner ei-ther directly, by promoting commitment of progenitors to OClineage and differentiation, or indirectly by stimulating secretionof RANKL and M-CSF by osteoblasts (25, 26). We addressed therole of TNF-a using mice that lack expression of p55R—that is,the TNF-aR that in contrast to p75R was shown previously tomediate the pro-osteoclastogenic effect of TNF-a (38). Indeed, theenhancing effect of aGC on OCP was significantly dampened inTNFap55R2/2 mice, suggesting that TNF-a is a major mediatorof the effects of aGC and iNKT cells in osteoclastogenesis.Given that overall in response to aGC osteoclastogenesis is

enhanced, we presume that the positive effect of TNF-a in this

process prevails over the suppressive effect of IFN-g, such that thepro-osteoclastogenic effect of iNKT cell activation is coupled withmyeloid cell egress.We found that aGC-mediated promotion of OC development is

not regulated by IL-4, the prototypical Th2 cytokine that nega-tively regulates osteoclastogenesis by inhibiting maturation of OCprecursors (39–41), or by IL-17, a recently described strongly pro-osteoclastogenic cytokine (23). Similarly, serum RANKL levelswere not increased in response to aGC, suggesting that solubleRANKL is not a major mediator of OC development in thisprocess. However, it is possible that membrane-bound RANKL(e.g., on activated T cells, osteoblasts, or other stroma cells) couldcontribute along with TNF-a in the enhanced development andactivation of OCs in response to iNKT cell activation.The ability of activated iNKT cells to modulate OC function

raised the prospect of these cells being required for OC devel-opment and activation under basal conditions. Indeed, iNKTcells are known to be autoreactive and in an activated state invivo, even in the absence of exogenous antigenic challenge[reviewed in Ref. 28]. Cord blood iNKT cells, which bear markersof Ag experience and activation, are a characteristic example ofthis endogenous activation of iNKT cells (42, 43). It is possiblethat the baseline activation of iNKT cells might have an effect onOC function and turnover and might ultimately result in impairedbone resorption and increased bone mass over time. Accordingly,detailed analysis of bone structure and mineralization in 11-wk-old Cd1d2/2 mice revealed a mild phenotype characterized byincreased BV/TV and increased bone mineralization density, ac-companied by evidence of reduced osteoclastic bone resorption.We propose that the effects of iNKT cells on OC function cover

a spectrum that ranges from baseline, homeostatic, chronic, low-level OC activation to the immediate and robust OC activationin response to exogenous stimuli, such as the one provided by aGC.Exogenous activation of iNKT cells results in secondary activation

of T cells, which are likely to be a major source of the cytokinesmodulating OC function. Indeed, T cells are a major source of pro-osteoclastogenic cytokines and indispensable for OC activation andbone loss in the context of immune activation in inflammatory andautoimmune disorders (13, 44–46). To our knowledge, our dataprovide evidence for the first time of the involvement of an im-munoregulatory T cell subset (i.e., iNKT cells) in this process.Based on this and previous studies (8, 9), we propose a model

integrating the role of iNKT cells in myeloid cell recruitmentduring an immune response (Fig. 7). iNKT cells promote activa-tion of mature myeloid cells (i.e., macrophages, mDCs, OCs) andenhance the clonogenic activity and differentiation of their com-mon progenitors and egress of their mature forms. Activation ofOCs by iNKT cells further potentiates this process. Our findingsalso suggest that therapeutic manipulation of iNKT cells to controlchronic inflammatory conditions, such as experimental (47) andrheumatoid arthritis (40, 48–50), might be useful for controllingthe associated cytokine-induced OC activation and bone de-struction (20).

DisclosuresThe authors have no financial conflicts of interest.

References1. Brigl, M., and M. B. Brenner. 2004. CD1: antigen presentation and T cell

function. Annu. Rev. Immunol. 22: 817–890.2. Kronenberg, M. 2005. Toward an understanding of NKT cell biology: progress

and paradoxes. Annu. Rev. Immunol. 23: 877–900.3. Cui, J., T. Shin, T. Kawano, H. Sato, E. Kondo, I. Toura, Y. Kaneko, H. Koseki,

M. Kanno, and M. Taniguchi. 1997. Requirement for Valpha14 NKT cells in IL-12-mediated rejection of tumors. Science 278: 1623–1626.

FIGURE 7. Proposed model of the role of exogenously activated

iNKT cells in the regulation of OCPs and mature cells. iNKT cells promote

growth and differentiation of myeloid progenitors (CFU-GM, CFU-M) as

well as activation of their mature progeny (i.e., macrophages, mDCs,

granulocytes). OCs develop from the same CFU-M progenitor as macro-

phages and mDCs. Activation of mature OCs leads to bone resorption,

contributes to maintenance of the hematopoietic stem cell niche, and

promotes progenitor mobilization. iNKT cells can potentiate this process

by promoting OCP growth and differentiation and mature OC activation.

Although IFN-g is a negative regulator of this process, the strong pro-

osteoclastogenic effect of TNF-a prevails.

2916 ACTIVATED iNKT CELLS REGULATE OSTEOCLASTS

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

4. Smiley, S. T., M. H. Kaplan, and M. J. Grusby. 1997. Immunoglobulin E pro-duction in the absence of interleukin-4-secreting CD1-dependent cells. Science275: 977–979.

5. Nieuwenhuis, E. E. S., T. Matsumoto, M. Exley, R. A. Schleipman, J. Glickman,D. T. Bailey, N. Corazza, S. P. Colgan, A. B. Onderdonk, and R. S. Blumberg.2002. CD1d-dependent macrophage-mediated clearance of Pseudomonas aeru-ginosa from lung. Nat. Med. 8: 588–593.

6. Fujii, S., K. Shimizu, C. Smith, L. Bonifaz, and R. M. Steinman. 2003. Acti-vation of natural killer T cells by alpha-galactosylceramide rapidly induces thefull maturation of dendritic cells in vivo and thereby acts as an adjuvant forcombined CD4 and CD8 T cell immunity to a coadministered protein. J. Exp.Med. 198: 267–279.

7. Hermans, I. F., J. D. Silk, U. Gileadi, M. Salio, B. Mathew, G. Ritter, R. Schmidt,A. L. Harris, L. Old, and V. Cerundolo. 2003. NKT cells enhance CD4+ and CD8+T cell responses to soluble antigen in vivo through direct interaction with den-dritic cells. J. Immunol. 171: 5140–5147.

8. Kotsianidis, I., J. D. Silk, E. Spanoudakis, S. Patterson, A. Almeida,R. R. Schmidt, C. Tsatalas, G. Bourikas, V. Cerundolo, I. A. Roberts, andA. Karadimitris. 2006. Regulation of hematopoiesis in vitro and in vivo by in-variant NKT cells. Blood 107: 3138–3144.

9. Leite-de-Moraes, M. C., M. Lisbonne, A. Arnould, F. Machavoine, A. Herbelin,M. Dy, and E. Schneider. 2002. Ligand-activated natural killer T lymphocytespromptly produce IL-3 and GM-CSF in vivo: relevance to peripheral myeloidrecruitment. Eur. J. Immunol. 32: 1897–1904.

10. Novack, D. V., and S. L. Teitelbaum. 2008. The osteoclast: friend or foe? Annu.Rev. Pathol. 3: 457–484.

11. Kollet, O., A. Dar, S. Shivtiel, A. Kalinkovich, K. Lapid, Y. Sztainberg,M. Tesio, R. M. Samstein, P. Goichberg, A. Spiegel, et al. 2006. Osteoclastsdegrade endosteal components and promote mobilization of hematopoieticprogenitor cells. Nat. Med. 12: 657–664.

12. Lymperi, S., N. Horwood, S. Marley, M. Y. Gordon, A. P. Cope, and F. Dazzi.2008. Strontium can increase some osteoblasts without increasing hematopoieticstem cells. Blood 111: 1173–1181.

13. Lorenzo, J., M. Horowitz, and Y. Choi. 2008. Osteoimmunology: interactions ofthe bone and immune system. Endocr. Rev. 29: 403–440.

14. Fogg, D. K., C. Sibon, C. Miled, S. Jung, P. Aucouturier, D. R. Littman,A. Cumano, and F. Geissmann. 2006. A clonogenic bone marrow progenitorspecific for macrophages and dendritic cells. Science 311: 83–87.

15. Miyamoto, T., O. Ohneda, F. Arai, K. Iwamoto, S. Okada, K. Takagi,D. M. Anderson, and T. Suda. 2001. Bifurcation of osteoclasts and dendritic cellsfrom common progenitors. Blood 98: 2544–2554.

16. Rothe, J., W. Lesslauer, H. Lotscher, Y. Lang, P. Koebel, F. Kontgen, A. Althage,R. Zinkernagel, M. Steinmetz, and H. Bluethmann. 1993. Mice lacking the tu-mour necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highlysusceptible to infection by Listeria monocytogenes. Nature 364: 798–802.

17. Bassett, J. H. D., P. J. O’Shea, S. Sriskantharajah, B. Rabier, A. Boyde, P. G.T. Howell, R. E. Weiss, J. P. Roux, L. Malaval, P. Clement-Lacroix, et al. 2007.Thyroid hormone excess rather than thyrotropin deficiency induces osteoporosisin hyperthyroidism. Mol. Endocrinol. 21: 1095–1107.

18. Bassett, J. H., A. Boyde, P. G. Howell, R. H. Bassett, T. M. Galliford,M. Archanco, H. Evans, M. A. Lawson, P. Croucher, D. L. St Germain, et al.2010. Optimal bone strength and mineralization requires the type 2 iodothyr-onine deiodinase in osteoblasts. Proc. Natl. Acad. Sci. USA 107: 7604–7609.

19. Jacquin, C., D. E. Gran, S. K. Lee, J. A. Lorenzo, and H. L. Aguila. 2006.Identification of multiple osteoclast precursor populations in murine bone mar-row. J. Bone Miner. Res. 21: 67–77.

20. Horwood, N. 2008. Lymphocyte-derived cytokines in inflammatory arthritis.Autoimmunity 41: 230–238.

21. Silk, J. D., I. F. Hermans, U. Gileadi, T. W. Chong, D. Shepherd, M. Salio,B. Mathew, R. R. Schmidt, S. J. Lunt, K. J. Williams, et al. 2004. Utilizing theadjuvant properties of CD1d-dependent NK T cells in T cell-mediated immu-notherapy. J. Clin. Invest. 114: 1800–1811.

22. Lee, K. A., M. H. Kang, Y. S. Lee, Y. J. Kim, D. H. Kim, H. J. Ko, andC. Y. Kang. 2008. A distinct subset of natural killer T cells produces IL-17,contributing to airway infiltration of neutrophils but not to airway hyperreac-tivity. Cell. Immunol. 251: 50–55.

23. Sato, K., A. Suematsu, K. Okamoto, A. Yamaguchi, Y. Morishita, Y. Kadono,S. Tanaka, T. Kodama, S. Akira, Y. Iwakura, et al. 2006. Th17 functions as anosteoclastogenic helper T cell subset that links T cell activation and bone de-struction. J. Exp. Med. 203: 2673–2682.

24. Richardson, P. G., B. Barlogie, J. Berenson, S. Singhal, S. Jagannath, D. Irwin,S. V. Rajkumar, G. Srkalovic, M. Alsina, R. Alexanian, et al. 2003. A phase 2study of bortezomib in relapsed, refractory myeloma. N. Engl. J. Med. 348:2609–2617.

25. Lam, J., S. Takeshita, J. E. Barker, O. Kanagawa, F. P. Ross, andS. L. Teitelbaum. 2000. TNF-alpha induces osteoclastogenesis by direct stimu-lation of macrophages exposed to permissive levels of RANK ligand. J. Clin.Invest. 106: 1481–1488.

26. Teitelbaum, S. L. 2006. Osteoclasts; culprits in inflammatory osteolysis. ArthritisRes. Ther. 8: 201.

27. Irmler, I. M., M. Gajda, and R. Brauer. 2007. Exacerbation of antigen-inducedarthritis in IFN-gamma-deficient mice as a result of unrestricted IL-17 response.J. Immunol. 179: 6228–6236.

28. Gapin, L. 2010. iNKT cell autoreactivity: what is ‘self’ and how is it recognized?Nat. Rev. Immunol. 10: 272–277.

29. Bendelac, A., P. B. Savage, and L. Teyton. 2007. The biology of NKT cells.Annu. Rev. Immunol. 25: 297–336.

30. Kollet, O., A. Dar, and T. Lapidot. 2007. The multiple roles of osteoclasts in hostdefense: bone remodeling and hematopoietic stem cell mobilization. Annu. Rev.Immunol. 25: 51–69.

31. Ozmen, L., M. Pericin, J. Hakimi, R. A. Chizzonite, M. Wysocka, G. Trinchieri,M. Gately, and G. Garotta. 1994. Interleukin 12, interferon gamma, and tumornecrosis factor alpha are the key cytokines of the generalized Shwartzman re-action. J. Exp. Med. 180: 907–915.

32. Gowen, M., and G. R. Mundy. 1986. Actions of recombinant interleukin 1, in-terleukin 2, and interferon-gamma on bone resorption in vitro. J. Immunol. 136:2478–2482.

33. Takahashi, N., G. R. Mundy, and G. D. Roodman. 1986. Recombinant humaninterferon-gamma inhibits formation of human osteoclast-like cells. J. Immunol.137: 3544–3549.

34. Gao, Y., F. Grassi, M. R. Ryan, M. Terauchi, K. Page, X. Yang,M. N. Weitzmann, and R. Pacifici. 2007. IFN-gamma stimulates osteoclastformation and bone loss in vivo via antigen-driven T cell activation. J. Clin.Invest. 117: 122–132.

35. Takayanagi, H., K. Ogasawara, S. Hida, T. Chiba, S. Murata, K. Sato,A. Takaoka, T. Yokochi, H. Oda, K. Tanaka, et al. 2000. T-cell-mediated regu-lation of osteoclastogenesis by signalling cross-talk between RANKL and IFN-gamma. [gamma] Nature 408: 600–605.

36. Carnaud, C., D. Lee, O. Donnars, S. H. Park, A. Beavis, Y. Koezuka, andA. Bendelac. 1999. Cutting edge: Cross-talk between cells of the innate immunesystem: NKT cells rapidly activate NK cells. J. Immunol. 163: 4647–4650.

37. Biburger, M., and G. Tiegs. 2005. Alpha-galactosylceramide-induced liver injuryin mice is mediated by TNF-alpha but independent of Kupffer cells. J. Immunol.175: 1540–1550.

38. Abu-Amer, Y., J. Erdmann, L. Alexopoulou, G. Kollias, F. P. Ross, andS. L. Teitelbaum. 2000. Tumor necrosis factor receptors types 1 and 2 differ-entially regulate osteoclastogenesis. J. Biol. Chem. 275: 27307–27310.

39. Moreno, J. L., M. Kaczmarek, A. D. Keegan, and M. Tondravi. 2003. IL-4suppresses osteoclast development and mature osteoclast function bya STAT6-dependent mechanism: irreversible inhibition of the differentiationprogram activated by RANKL. Blood 102: 1078–1086.

40. Guillot, P. V., H. T. Cook, C. D. Pusey, N. M. Fisk, S. Harten, J. Moss, I. Shore,and G. Bou-Gharios. 2008. Transplantation of human fetal mesenchymal stemcells improves glomerulopathy in a collagen type I alpha 2-deficient mouse. J.Pathol. 214: 627–636.

41. Okada, Y., I. Morimoto, K. Ura, Y. Nakano, Y. Tanaka, S. Nishida, T. Nakamura,and S. Eto. 1998. Short-term treatment of recombinant murine interleukin-4rapidly inhibits bone formation in normal and ovariectomized mice. Bone 22:361–365.

42. van Der Vliet, H. J. J., N. Nishi, T. D. de Gruijl, B. M. von Blomberg, A. J.M. van den Eertwegh, H. M. Pinedo, G. Giaccone, and R. J. Scheper. 2000.Human natural killer T cells acquire a memory-activated phenotype before birth.Blood 95: 2440–2442.

43. D’Andrea, A., D. Goux, C. De Lalla, Y. Koezuka, D. Montagna, A. Moretta,P. Dellabona, G. Casorati, and S. Abrignani. 2000. Neonatal invariant Valpha24+NKT lymphocytes are activated memory cells. Eur. J. Immunol. 30: 1544–1550.

44. Walsh, M. C., N. Kim, Y. Kadono, J. Rho, S. Y. Lee, J. Lorenzo, and Y. Choi.2006. Osteoimmunology: interplay between the immune system and bone me-tabolism. Annu. Rev. Immunol. 24: 33–63.

45. Nakashima, T., and H. Takayanagi. 2008. The dynamic interplay betweenosteoclasts and the immune system. Arch. Biochem. Biophys. 473: 166–171.

46. Takayanagi, H. 2005. Mechanistic insight into osteoclast differentiation inosteoimmunology. J. Mol. Med. 83: 170–179.

47. Teige, A., R. Bockermann, M. Hasan, K. E. Olofsson, Y. Liu, and S. Issazadeh-Navikas. 2010. CD1d-dependent NKT cells play a protective role in acute andchronic arthritis models by ameliorating antigen-specific Th1 responses. J.Immunol. 185: 345–356.

48. Linsen, L., M. Thewissen, K. Baeten, V. Somers, P. Geusens, J. Raus, andP. Stinissen. 2005. Peripheral blood but not synovial fluid natural killer T cellsare biased towards a Th1-like phenotype in rheumatoid arthritis. Arthritis Res.Ther. 7: R493–R502.

49. van Amerongen, M. J., G. Bou-Gharios, E. Popa, J. van Ark, A. H. Petersen,G. M. van Dam, M. J. van Luyn, and M. C. Harmsen. 2008. Bone marrow-derived myofibroblasts contribute functionally to scar formation after myocardialinfarction. J. Pathol. 214: 377–386.

50. van der Vliet, H. J. J., B. M. von Blomberg, N. Nishi, M. Reijm, A. E. Voskuyl,A. A. van Bodegraven, C. H. Polman, T. Rustemeyer, P. Lips, A. J. M. van denEertwegh, et al. 2001. Circulating V(alpha24+) Vbeta11+ NKT cell numbers aredecreased in a wide variety of diseases that are characterized by autoreactivetissue damage. Clin. Immunol. 100: 144–148.

The Journal of Immunology 2917

by guest on June 5, 2018http://w

ww

.jimm

unol.org/D

ownloaded from