Bone Malformations in Interleukin-18 Transgenic Mice

YUSUKE KAWASE,1 TOMOAKI HOSHINO,2 KOICHI YOKOTA,1 AKEMI KUZUHARA, 1

MASANORI NAKAMURA, 1 YU MAEDA, 1 EIJI NISHIWAKI,1 MICHIHISA ZENMYO,3 KOJI HIRAOKA,3

HISAMICHI AIZAWA, 2 and KOHICHIRO YOSHINO1

ABSTRACT

The in vivo effects of IL-18 on bone metabolism were investigated by histopathology in IL-18 transgenic mice.Deformed cortical bone and decreased turnover rate of lumbar trabecular bone are consistent with increasedexpression of IFN-� and IL-18 in the bone marrow.

Interleukin (IL)-18 has been demonstrated to inhibit osteoclastogenesis in an in vitro co-culture system. Weinvestigated the effects of IL-18 overexpression on bone metabolism by comparing bone characteristics in maleIL-18 transgenic (TG) mice, which secrete mature murine IL-18 from their B- and T-cells, and their wildtypelittermates (WT). Histopathological analysis revealed that the cortical bone of the femur was thinner and moredeformed in IL-18 TG mice. Bone histomorphometry showed that the cortical bone area of the mid-diaphysisof the femur and the trabecular bone volume of the lumbar vertebrae were significantly reduced in IL-18 TGmice. IL-18 TG mice also exhibited significantly fewer osteoclasts and a reduced bone formation rate in thetrabecular bones of their lumbar vertebrae. Real-time reverse transcriptase-polymerase chain reactionamplification of bone marrow cell mRNA revealed that interferon (IFN)-� mRNA expression was significantlyincreased, whereas IL-4 mRNA expression was significantly reduced, in IL-18 TG mice. However, theexpression ratio of receptor activator of NF�B ligand and osteoprotegerin mRNA was not significantly altered.Thus, deformed cortical bone and a decreased turnover rate of lumbar trabecular bone are characteristic ofIL-18 TG mice, and these features might be associated with the increased expression of IFN-� and IL-18 inthe bone marrow. (J Bone Miner Res 2003;18:975–983)

Key words: interleukin-18, osteoclast, bone histomorphometry, osteodystrophy, transgenic mouse

INTRODUCTION

BONE REMODELING INVOLVES the highly regulated cou-pling of bone resorption and formation. Osteoclasts are

the principal cells involved in bone resorption cells,(1)

whereas osteoblasts are responsible for the synthesis anddeposition of the bone extracellular matrix.(2) The functionsof osteoclasts and osteoblasts are intimately linked; bone isconstantly being destroyed or resorbed by osteoclasts thenreplaced by osteoblasts, and the bone remodeling process istightly regulated by local and endocrine factors.

Interleukin (IL)-18 was originally discovered as an inter-feron (IFN)-�–inducing factor in aProprionibacteriumacnes–induced shock model.(3) IL-18 shares some biologi-

cal similarities with IL-12, although the two are not struc-turally related.(4) It has been reported that IL-18 is expressedin a wide range of cells, including Kupffer cells, macro-phages, T-cells, B-cells, keratinocytes, intestinal epithelialcells, dendritic cells,(5) and chondrocytes.(6) The IL-18 re-ceptor is composed of at least two chains: a signalingcomponent (IL-18R�)(7) and a ligand-binding subunit (IL-18R�).(8) The mRNAs for both of these components areexpressed in the lung, spleen, colon, and leukocytes.(8)

IL-18 is also present in certain bone marrow cells; forexample, osteoblastic stromal cells have been shown to expressIL-18 mRNA.(9) IL-18 was demonstrated to inhibit osteoclas-togenesis through granulocyte/macrophage colony-stimulatingfactor (GM-CSF), but not IFN-�, production in a co-culturesystem containing mouse osteoblastic cells and bone marrow/spleen cells.(10) Furthermore, GM-CSF produced by T-cellsThe authors have no conflict of interest.

1R&D Laboratories, Nippon Organon K.K., Osaka, Japan.2Department of Internal Medicine 1, Kurume University School of Medicine, Kurume, Japan.3Department of Orthopedics, Kurume University School of Medicine, Kurume, Japan.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 18, Number 6, 2003© 2003 American Society for Bone and Mineral Research

975

was reported to contribute to the inhibition of osteoclast for-mation.(11) Subsequent studies demonstrated that mRNA ofIL-18 receptor complex is expressed in a mouse stromal cellline, ST2, a mouse osteoblast cell line, MC3T3-E1, and mousecalvarial osteoblasts, and IL-18 might inhibit osteoclastogen-esis by upregulating osteoprotegerin (OPG; also known asosteoclastogenesis inhibitory factor [OCIF]) in vitro.(12) Inaddition, it was also reported that both � and � chains of IL-18receptor are expressed in the osteoclast.(13) These findings,however, are limited to in vitro experiments, and it remainsunclear whether IL-18 influences bone mass or bone remod-eling in vivo.

Recently, we established IL-18 transgenic (TG) mice, inwhich mouse mature IL-18 was overexpressed in their B-and T-cells(14) under the control of the Ig-promoter. TheseTG mice exhibit high serum levels of IgE, IgG1, IL-4, andIFN-�. To investigate the effects of overexpression of ma-ture IL-18 in the bone marrow on the resulting bone, weanalyzed bone histopathology and histomorphometry inthese TG mice.

MATERIALS AND METHODS

Mice

IL-18 TG mice, in which B- and T-cells express matureIL-18 under the control of an IgVH-promoter, were gener-ated as described previously.(14) The IL-18 TG mice weremated with C57BL/6N mice, purchased from Charles RiverJapan (Yokohama, Japan), in Nippon Organon, K.K. Themale hemizygous TG mice and their male wildtype litter-mates (WT) were used for this study at 8–10 weeks of age.They were allowed free access to tap water and a standarddiet (CE2, containing 1.08% calcium and 0.97% phospho-rus; Japan Clea, Osaka, Japan) and were kept in an air-conditioned room with a temperature of 24 � 2°C, a hu-midity of 55 � 5%, and a 12-h light/dark cycle. The micewere maintained and treated in accordance with the proce-dures outlined in the Guide for the Care and Use of Labo-ratory Animals.

Reagents

Calcein was purchased from Dojindo Laboratories (Ku-mamoto, Japan) and Isogen from Nippon Gene (Tokyo, Ja-pan). A phycoerythrin (PE)-conjugated anti-mouse CD45/B220 monoclonal antibody (mAb) (clone: RA3–6B2), an anti-mouse CD16/CD32 mAb, fluorescein-isothiocyanate (FITC)-conjugated rat IgG2a (clone: R35–95), and PE-conjugated ratIgG2a (clone: DAK-G09) were purchased from PharMingen(San Diego, CA, USA), and a FITC-conjugated, affinity-purified goat anti-mouse IgM �-heavy chain-specific antibody(Ab) (clone: LO-MM-9) was obtained from Zymed (San Fran-cisco, CA, USA). RPMI 1640 was purchased from GIBCOBRL (Gaithersburg, MD, USA). A deoxypyridinoline (Dpyr)ELISA kit, Pyridilinks-D, was obtained from Quidel (SantaClara, CA, USA), and the creatinine test kit (Creatinine TestWako) was purchased from Wako Pure Chemical Industry(Osaka, Japan). Villanueva bone stain powder was purchasedfrom Maruto (Tokyo, Japan).

Tissue collection and analysis

The TG and WT mice were injected intraperitoneallywith calcein (8 mg/kg body weight) for bone labeling 10and 4 days before death. Also, urine was collected for 18 h,and blood was taken for serum preparation before death. Allmice were killed by ether anesthesia. The tibias, femora, andhumeri from both sides were excised, together with thethoracic and lumbar vertebrae. For flow cytometric analysis,both ends of the humeri were cut, and the bone marrow cellswere isolated in RPMI 1640 containing 10% fetal calfserum (FCS). For the real-time reverse transcriptase-polymerase chain reaction (RT-PCR) analysis, both ends ofthe femora and tibia were cut, and the bone marrow cellswere extruded into 10% FCS RPMI 1640. After the centrif-ugation, Isogen (Nippon Gene) was added to the bonemarrow cells, and total RNA was isolated according tomanufacturer’s instructions.

Histomorphometry

The right femora and the lumbar vertebrae were stainedwith Villanueva bone stain for 5 days, after which they wereembedded in methyl methacrylate, and 8-�m-thick sagittalsections were prepared. Histomorphometry was performedusing image-analyzing software (Macscope; Mitani, Tokyo,Japan). The area of the secondary spongiosa, which ispresent in the region lying at least 0.1 mm from the corticalbone and 0.3 mm from the growth ends, was measured. Thetrabecular bone volume (BV), tissue volume (TV), bonesurface (BS), trabecular number (Tb.N), trabecular thick-ness (Tb.Th), single-labeled surfaces (sLS), double-labeledsurfaces (dLS), and interlabel distances (IrLth) were mea-sured. The mineralizing surface (MS/BS, %) was calculatedas (sLS/2 � dLS)/BS. The mineral apposition rate (MAR,�m/day) was calculated as the mean distance between thefirst and second label divided by the labeling interval (6days). The bone formation rate (BFR/BS, �m2/�m2/day)was calculated as the product of MAR, MS, and BS asdescribed by Parfitt et al.(15) To estimate the number ofosteoclasts, the femora and lumbar vertebrae were fixedwith 4% lysine periodate paraformaldehyde at 4°C for 6 h.Subsequently, they were decalcified with 10% EDTA-glycerol at �5°C for 10 days and embedded in paraffin.Three-micrometer-thick sagittal sections were made andstained with hematoxylin-eosin (HE) for TRACP. The os-teoclast number (N.Oc), BV, TV, and BS were measured.Multinuclear TRACP� cells with resorption lacunae, situ-ated at the surface of the trabecular bone, were identified asosteoclasts. Finally, cortical parameters were measured inthe femora by bisecting them transversely at the midpoint ofthe shaft and determining the internal (Ma.Dm) and external(B.Dm) diameters, and the cortical bone area (Ct.Ar) andthickness (Ct.Wi). Each parameter was analyzed using aspecimen from each mouse, and two pathologists indepen-dently scored these sections in a blinded manner.

Bone mineral density and ash weight

The bone mineral densities (BMDs) of the right femorawere measured by DXA on a QDR-1000 scanner (HologicInc., Waltham, MA, USA) with a line spacing of 0.254 mm

976 KAWASE ET AL.

and a point resolution of 0.127 mm. The femora wereburned in an oven at 700°C for 16 h, and the ash weight wasdetermined.

Blood and urinary biochemistry

Plasma levels of calcium and phosphorus were measuredby colorimetry using an autoanalyzer (type 7070; Hitachi,Tokyo, Japan). Serum levels of IL-18 were measured by themouse IL-18 ELISA kit (MBL, Nagoya, Japan). UrinaryDpyr levels were measured using an ELISA kit, and thecreatinine level in each urinary sample was measured usingCreatinine Test Wako, as described in the Reagents subsec-tion above. The urinary excretion of each marker was ex-pressed as a ratio relative to that of creatinine (nM/mMcreatinine).

Quantitative real-time RT-PCR

cDNAs were synthesized from total RNA (2 �g) usingthe Super Script Preamplification System for First-StrandcDNA Synthesis kit (Invitrogen, Carlsbad, CA, USA). Thesynthesized cDNA (2 �l), in a total reaction volume of 50�l comprised of TaqMan 1000 reaction PCR core reagents(Applied Biosystems Japan, Tokyo, Japan), 300 nM prim-ers, and 200 nM probes, was amplified using Gene-Amp5700 (Applied Biosystems). The mixture was incu-bated for 2 minutes at 50°C and denatured for 10 minutes at95°C and subjected to 45 two-step amplification cycles eachcomprised of an annealing/extension at 60°C for 1 minutefollowed by denaturation at 95°C for 15 s. Sequence-specific amplification was detected as an increase in thefluorescent signal generated by 6-carboxy-fluorescein dur-ing the amplification cycle. The primers and probes weredesigned using Primer Express software (Applied Biosys-tems) and synthesized. The primer and probe sequenceswere as follows: mIL-18 transgene forward primer: 5�-TGGGTACTGCTGCTCTGGGT-3�; mIL-18 transgenereverse primer: 5�-ATTCCGTATTACTGCGGTTGTA-CAG-3�; mIL-18 transgene probe: 5�-CCACTGGT-GACAACTTTGGCCGACTTC-3�; pro-mIL-18 forwardprimer: 5�-TCAGGACAAAGAAAGCCGCC-3�; pro-mIL-18reverse primer: 5�-TCTGACATGGCAGCCATTGT-3�;pro-mIL-18 probe: 5�-ACCTTCCAAATCACTTCCTCTT-GGCCC-3�; mIFN-� forward primer: 5�-AGCTCATC-CGAGTGGTCCAC-3�; mIFN-� reverse primer: 5�-AG-CAGCGACTCCTTTTCCG-3�; mIFN-� probe: 5�-TGTT-GCCGGAATCCAGCCTCAGG-3�; mIL-4 forward primer:5�-GGCATTTTGAACGAGGTCACA-3�; mIL-4 reverseprimer: 5�-AGGACGTTTGGCACATCCA-3�; mIL-4probe: 5�-CTCCGTGCATGGCGTCCCTTCT-3�; mGM-CSF forward primer: 5�- GGCGCCTTGAACATGACAG-3�; mGM-CSF reverse primer: 5�- TTTCACAGTCCGTT-TCCGG-3�; mGM-CSF probe: 5�- CAGCTACTACCAGA-CATACTGCCCCCCA-3�; mouse RANK (mRANK) forwardprimer: 5�-TGTGGTCTGCAGCTCTTCCA-3�; mRANKreverse primer: 5�-ATGAGACTGGGCAGGTAAGCC-3�;mRANK probe: 5�-CACTGAGGAGACCACCCAAG-GAGGC-3�; mRANK ligand (mRANKL) forward primer: 5�-GGATGTGGCCCAGCGAG-3�; mRANKL reverse primer:5�-TGCTGGCAGCATTGATGG-3�; mRANKL probe: 5�-CAAGCCTGAGGCCCAGCCATTTG-3�; mOPG forward

primer: 5�-GCTGCAGAGACGCACCTAGC-3�; mOPG re-verse primer: 5�-TTCATTGTGGTCCTCGGGA-3�; mOPGprobe: 5�-CTGACCCAGCGGCTGCCTCC-3�. The sense andantisense primers and probe for mGAPDH were purchasedfrom Applied Biosystems Japan. Amplification of themGAPDH was performed on all samples tested to control forvariations in their RNA contents, and all transcript values werenormalized with reference to mGAPDH mRNA levels. Ano-template control was included in each amplification reac-tion to control for contaminating templates. For a valid sampleanalysis, the fluorescence intensity of the no-template controlwas required to be zero.

Flow cytometric analysis

Bone marrow cells were prepared by flushing the bonemarrow out of the right and left humeri with RPMI 1640supplemented with 10% FCS, using a syringe with a 27-gauge needle. The cells were centrifuged and resuspendedin 1 ml of ACK lysing buffer (0.15 M NH4Cl, 10 mMKHCO3, and 0.1 mM EDTA�2Na, pH 7.4) to lyse the redblood cells. The cell suspension was washed with RPMI1640 supplemented with 10% FCS and resuspended in 1 mlCa2�- and Mg2�-free PBS containing 1% bovine serumalbumin (BSA). The bone marrow cells (1 � 106) wereincubated for 30 minutes on ice with anti-mouse CD16/32mAb to block nonspecific binding, washed, and resus-pended in PBS containing 1% BSA. They were then incu-bated for 20 minutes on ice with FITC-conjugated affinity-purified goat anti-mouse IgM �-heavy chain-specific mAband PE-anti-mouse CD45/B220 mAb, washed twice, andresuspended with 1 ml 4% paraformaldehyde. The negativecontrols were stained with isotype matched FITC-rat IgG2aand PE-rat IgG2a. The cells were analyzed using an EPICSflow cytometer (Beckman Coulter, Fullerton, CA, USA).

Statistical analysis

All results were expressed as means � SE. Differencesbetween the two groups were analyzed using the unpairedtwo-tailed Student’s t-test, and those with p values of lessthan 0.05 were considered to be significant.

RESULTS

IL-18 expression in IL-18 TG mice

Expression of mRNA for mature IL-18 transgene relativeto GAPDH mRNA was significantly higher in the IL-18 TGmice than that in the WT mice (Fig. 1). In addition, serumlevels of IL-18 in the IL-18 TG mice were significantlyhigher than WT mice as previously reported(14) (Fig. 1).

Cortical and trabecular bone volume of the femur

The IL-18 TG mice were normal in appearance, and theirbody weights were no different from those of the WT mice(Table 1). The skeletal phenotypes of the IL-18 TG micewere analyzed in their femora and lumbar vertebrae, andcortical bone parameters were assessed at the dissectionpoint in the mid-diaphysis of the femur. There was nosignificant difference in the B.Dm of the femur (data notshown). However, the Ct.Ar and the Ct.Wi of the corticalbone in the IL-18 TG mice were significantly reduced

977BONE MALFORMATIONS IN IL-18 TRANSGENIC MICE

compared with WT mice (Fig. 2A). The BV/TV of thetrabecular bone of the distal metaphysis of the femurshowed a tendency toward a reduction in the IL-18 TG micecompared with the WT mice (Fig. 2B), as did the BMD andash weight of the femur (data not shown).

Histopathology of the femur

Sagittal and cross-sections of the mid-diaphysis of thefemur were compared between the IL-18 TG and WT mice.A deranged cortical bone structure, as well as decreasedcortical bone thickness, was observed in the IL-18 TG mice(Fig. 3). However, the osteocytes were normal in appear-ance, and no apoptotic cell death was observed (data notshown). Thus, disturbance of cortical bone remodelingseems to occur in IL-18 TG mice. Chondrodysplasia wasnot observed in the IL-18 TG mice (data not shown).

Histomorphometry of the trabecular bone of thelumbar vertebrae

To assess trabecular bone remodeling, parameters asso-ciated with bone formation and resorption were analyzed inthe IL-18 TG and WT mice. When compared with WTmice, the reduced BV/TV in the lumbar vertebrae of theIL-18 TG mice was associated with a decrease in the Tb.Nand Tb.Th of trabecular structures. The MS/BS and theMAR also showed a tendency toward a decrease in theIL-18 TG mice compared with the WT mice. As a result, theBFR/BS was significantly reduced in the IL-18 TG mice.

The N.Oc/BS was also significantly lower in the IL-18 TGmice (Figs. 4 and 5). These observations indicate that theturnover rate of trabecular bone in the lumbar vertebrae islower in IL-18 TG than in WT mice. The BV/TV of thethoracic vertebrae showed only a tendency toward a reduc-tion in the TG mice (data not shown). Thus, the degree ofreduction in the trabecular bone volume in IL-18 TG miceseems to depend on the site.

Blood and urinary biochemical parameters

There were no significant differences in serum calciumand phosphorus levels between the IL-18 TG and WT mice.Urinary Dpyr levels were decreased by 13% in the IL-18TG mice compared with the WT mice, but it was notstatistically significant (Table 1).

Real time RT-PCR in bone marrow cells

It is well known that cytokines such as RANK, RANKL,OPG, GM-CSF, and IFN-� influence the differentiation ofosteoclasts. The expression of mRNA for these cytokineswas therefore assessed in bone marrow cells isolated fromthe left femur and tibia of the IL-18 TG and WT mice usingreal time RT-PCR (Fig. 6). The expression of IFN-� mRNAwas significantly increased, whereas IL-4 mRNA expres-sion was significantly decreased, in the IL-18 TG mice. Theexpression of GM-CSF mRNA was increased by 42% in theIL-18 TG mice. The expression of mRNA for RANK andRANKL, two osteoclast differentiation factors, and forOPG, an osteoclastgenesis inhibitory factor, were all de-creased by approximately 35% in the IL-18 TG mice. How-ever, the mRNA expression ratio of RANK/RANKL andOPG was not altered in the IL-18 TG mice.

Flow cytometric analysis

There was no significant difference in the total number ofbone marrow cells obtained from the humerus between theIL-18 TG and WT mice (data not shown). B220 is a cellsurface marker expressed on cells of B-lymphoid lineage atall stages of differentiation. The number of B220� and/orIgM� cells was significantly decreased in the bone marrowcells isolated from the IL-18 TG mice compared with WTmice (Fig. 7).

DISCUSSION

The IL-18 TG mice established in our laboratory areviable and have a normal appearance(14); however, theyshow high serum levels of IgE, IgG1, IFN-�, and IL-4, aswell as IL-18, and have a high potential for producing bothTh1 and Th2 cytokines, including IFN-� and IL-13, in theirspleen cells. In addition, they show an increased number ofCD8�CD44high T-lymphocytes and macrophages and a de-creased number of splenic B-lymphocytes. Thus, the IL-18TG mouse represents an important murine model for ana-lyzing the effects of IL-18 overexpression. The bone mar-row of these mice shows high IL-18 expression because thepE�IgH vector, which encodes a human E� enhancer and amouse IgVH promoter, was used to generate mature IL-18.To elucidate the effects of IL-18 overexpression on bone,we used young (8–10 weeks old) male IL-18 TG mice to

FIG. 1. Expression of mRNA for IL-18 in bone marrow cavity andserum IL-18 levels in the IL-18 TG and WT mice. The expression ofthe transgene-derived mature IL-18 mRNA was determined using aquantitative real-time RT-PCR, and referenced to that of GAPDHmRNA. The blood was taken just before death, and the serum wasprepared. Data represent the mean � SE for six animals. ND, notdetectable.

TABLE 1. BODY WEIGHT, SERUM LEVELS OF CALCIUM AND

PHOSPHORUS, AND URINARY LEVELS OF DEOXYPYRIDINOLINE

IN IL-18 TG AND WT MICE

Bodyweight (g)

Calcium(mg/dl)

Phosphorus(mg/dl)

Deoxypyridinoline(nM/mM creatinine)

WT 18.8 � 0.6 9.0 � 0.1 9.2 � 0.7 28.3 � 2.9TG 18.4 � 1.1 9.3 � 0.1 9.6 � 0.4 24.7 � 2.4

Data represent the mean � SE for six animals. There was no statisticallysignificant difference in any of these parameters between the two groups(Student’s t-test). Blood samples were taken just before death, and urinarysamples were collected for 18 h before death.

978 KAWASE ET AL.

avoid the influence of differences in sex hormones and foodintake on bone metabolism.

A characteristic feature of the bone in IL-18 TG mice wasa cortical bone structure that was thinner and more irregu-larly shaped than that in WT mice. Although we carried outstaining for apoptotic cells, no apoptosis of the bone cellswas founded in the cortical bone area (data not shown).IL-18 is known to be an angiogenic inhibitory factor,(16) andit has been demonstrated that abnormal angiogenesis mightlead to aberrant bone formation.(17) Thus, we examined thedistribution of blood vessels in the metaphyseal area of thefemur, but no aberrant angiogenesis was found in either theIL-18 or the WT mice (data not shown). Bone histomor-phometry of the trabecular bone revealed that a parameter ofbone resorption, such as the number of osteoclasts on thetrabecular bone, was decreased in the IL-18 TG mice com-pared with the WT mice. The BFR/BS, a parameter of boneformation was also decreased in the IL-18 TG mice. The

trabecular bone volume in the lumbar vertebrae and thecortical bone area of the mid-diaphysis of the femur weresignificantly reduced. The trabecular bone volume of thedistal metaphysis of the femur as well as the BMD and ashweight of the femur showed a tendency toward a reductionin the IL-18 TG mice compared with the WT mice (data notshown). These observations indicate that the turnover rate oftrabecular bone may be reduced, and the overall bone vol-ume seems to be lower in IL-18 TG mice compared withWT mice.

A quantitative real time RT-PCR analysis revealed thatexpression of IFN-� mRNA in the bone marrow cells washigher in the IL-18 TG mice than in the WT mice. IFN-�has been reported to suppress collagen synthesis in achondrocyte culture,(18) and TG mice that overexpressedIFN-� in their bone marrow and thymus showed severeosteochondrodysplasia, including shortening and widen-ing of the long bones, fractures, thickening of the growth

FIG. 2. Cortical and trabecular bone volumeof the femur in the IL-18 TG and WT mice. (A)The cortical bone area (Ct.Ar) and cortical bonethickness (Ct.Wi) were measured at the mid-diaphysis of the femur by histomorphometry.(B) The trabecular bone volumes per tissue vol-ume (BV/TV) of the femur were measured at thedistal metaphysis by histomorphometry. Datarepresent the mean � SE for (A) six or (B) fiveanimals. *p � 0.05, **p � 0.01, significantlydifferent from WT mice (Student’s t-test).

FIG. 3. Histopathology of the mid-diaphysisof the femur in IL-18 TG and WT mice. (A)Vertical section of the cortical bone at the mid-diaphysis of the femur of WT mice. (B) Verticalsection of the cortical bone at the mid-diaphysisof the femur of IL-18 TG mice. (C) Transversesection of the cortical bone at the mid-diaphysisof the femur of WT mice. (D) Transverse sectionof the cortical bone at the mid-diaphysis of thefemur of IL-18 TG mice. All specimens werestained with hematoxylin-eosin. Compared withWT mice, the cortical bone of the IL-18 TG micewas immature and distorted; however, thegrowth plate and metaphyseal trabeculae ap-peared normal. Each bar indicates 200 �m.

979BONE MALFORMATIONS IN IL-18 TRANSGENIC MICE

plates, cartilaginous masses in the bone marrow cavity,and thinning and immaturity of the cortical bones.(19)

Although chondrodysplasia was not found in our IL-18TG mouse, the thin and immature cortical bone wasobserved. Thus, an influence of increased IFN-� expres-sion in the bone marrow cavity on collagen synthesismight lead to the cortical bone aberration seen in IL-18TG mice. In addition to inhibiting collagen synthesis,

IFN-� is known to suppress osteoclastogenesis and boneresorption in in vitro co-culture systems.(20) Recently, ithas been demonstrated that IFN-� suppresses osteoclas-togenesis by interfering with the RANKL-RANK signal-ing pathway in a co-culture system.(21) Moreover, one ofaberrant features of the IFN-� TG mouse phenotype isthat fewer osteoclasts are present at the metaphysis thanseen in WT mice.(19) Thus, increased expression of IFN-�

FIG. 4. The trabecular bone volume, the number of osteoclasts on the trabecular bone, and bone formation parameters for trabecular bone inthe lumbar vertebrae in IL-18 TG and WT mice. Bone histomorphometry parameters were measured using image-analyzing software as describedin the Materials and Methods section. The number of TRACP� cells was counted using microscopy at a magnification of 400�. The bone waslabeled with calcein twice, with a 6-day interval between the labeling. Data represent the mean � SE for four animals. *p � 0.05, **p � 0.01,significantly different from WT mice (Student’s t-test). BV/TV, bone volume per tissue volume; Tb.N, trabecular number; Tb.Th, trabecularthickness; N.Oc, osteoclast number; MS/BS, mineralizing surface per bone surface; MAR, mineral apposition rate; BFR/BS, bone formation rateper bone surface.

FIG. 5. Histomorphometry of the trabecularbone of lumbar vertebrae in IL-18 TG and WTmice. (A) TRACP� cells on the trabecular bonein WT mice. (B) TRACP� cells on the trabecu-lar bone in IL-18 TG mice. The lumbar vertebraewere decalcificated using EDTA, cut into 3-�msections, and stained for TRACP activity. Ar-rowheads indicate TRACP� cells. (C) Mineral-ization of trabecular bone of WT mice. (D) Min-eralization of the trabecular bone of IL-18 TGmice. The bone was labeled with calcein twice,with a 6-day interval between the labeling. Thedistance between the bands was smaller in theIL-18 TG mice than in WT mice. Each barindicates 100 �m.

980 KAWASE ET AL.

in the bone marrow cavity might be associated with thereduced number of osteoclasts observed in the IL-18 TGmice. The flow cytometric analysis of the bone marrowcells revealed a decreased population of progenitorB-cells in the IL-18 TG mice, which have previouslybeen demonstrated to have a significantly reduced sple-netic B-lymphocytes.(14) This feature is shared by TGmice that overexpress IFN-�.(22) The overproduction ofIFN-� in the bone marrow in IL-18 TG mice may there-fore be associated with an alteration in the myeloid cellcomposition as well as with bone morphological aberra-tions.

The differentiation of osteoclasts is principally regu-lated by M-CSF, RANKL, and OPG.(1) Previous reportshave shown that IL-18 inhibits osteoclastogenesis byupregulating GM-CSF(10) or OPG(12) in co-culture sys-tems. In the present study, we analyzed the expression ofmRNAs for RANK, RANKL, OPG, and GM-CSF in thebone marrow of these mice using real-time RT-PCR. Analteration in osteoclast differentiation is often accompa-nied by a change in the expression ratio of RANKL andOPG.(23) However, the mRNA expression ratio ofRANKL and OPG was not altered in the IL-18 TG mice.Makiishi-Shimobayashi et al.(12) reported that IL-18 up-regulates OPG expression in stromal/osteoblastic cells inin vitro systems. These results suggest that IL-18 mayplay different roles in OPG expression in vivo and invivo. Thus, further analysis is required to elucidatewhether OPG is associated with the reduced osteoclastnumber in IL-18 TG mice. B-lymphoid lineage cells have

been reported to be a major source of endogenousRANKL in the bone marrow and can themselves serve asosteoclast progenitors.(24) In our IL-18 TG mice, theexpression of RANKL mRNA was decreased to almostthe same extent as the number of B-lymphoid lineagecells in the bone marrow (Fig. 7). It is likely that thereduced expression of RANKL mRNA resulted from thedecrease in the number of B-lymphoid lineage cells in thebone marrow of the IL-18 TG mice.

The real-time RT-PCR analysis of bone marrow cellsrevealed a tendency toward an increase in the expression ofGM-CSF mRNA in the IL-18 TG mice. IL-18 is known toenhance the production of GM-CSF as well as IFN-� inperipheral blood mononuclear cells .(25) GM-CSF producedby T-cells under the influence of IL-18 has been reported toinhibit osteoclast formation in an in vitro co-culture sys-tem.(10) Previous studies have shown that many other kindsof bone marrow cell, including osteoblasts/stromal cells,monocytes, and B-lymphocytes, are able to produce GM-CSF, and that the IL-18 receptor is expressed in stromalcells, monocytes,(9) and B-lymphocytes.(26) It is possiblethat upregulation of GM-CSF in bone marrow cells contrib-utes to the reduction in osteoclast numbers in IL-18 TGmice.

We have previously demonstrated that IL-18 TG miceshow increased serum IL-4 levels,(14) although in thepresent study, the expression of IL-4 mRNA in the bonemarrow was lower in the IL-18 TG mice than in WT mice.The expression of IL-4 is induced in T-cells, especiallydifferentiated Th2 cells, and NK1.1� T-cells, in response to

FIG. 6. Expression of mRNAs for cytokinesand other genes relevant to osteoclast formationin the bone marrow cells of IL-18 TG and WTmice. The expression of each mRNA was deter-mined using quantitative real-time RT-PCR andwas corrected with reference to expression ofGAPDH mRNA. Data represent the mean � SEfor five (WT mice) or six animals (TG mice).*p � 0.05, significantly different from WT mice(Student’s t-test).

FIG. 7. B-lymphoid lineage cells were de-creased in the bone marrow cells of IL-18 TGand WT mice. Bone marrow cells were isolatedfrom IL-18 TG and WT mice and stained withFITC-conjugated anti-mouse IgM �-heavychain-specific Abs and PE-conjugated anti-mouse CD45/B220 mAb. Representative stain-ing patterns were shown. Data represent themean � SE for six animals. **p � 0.01, signif-icantly different from WT mice (Student’st-test).

981BONE MALFORMATIONS IN IL-18 TRANSGENIC MICE

stimulation through the T-cell receptor.(27,28) The high se-rum levels of IL-4 seen in IL-18 TG mice might therefore bederived from T-cells in the lymphatic tissues rather thanbone marrow cells. Previous studies have reported that IL-4inhibits osteoclastogenesis from bone marrow macrophagesin vitro,(29) and IL-4 TG mice exhibit osteoporosis.(30) InIL-4 TG mice, the expression of IL-4 mRNA is increased inthe bone marrow compartment compared with their WTmice.(30) Thus, the decreased expression of IL-4 mRNA inthe bone marrow may not be directly associated with thedeceased osteoclasts in IL-18 TG mice. However, it ispossible that high serum levels of IL-4 result in the decreasein osteoclasts indirectly by inhibiting osteoclastogenesis oraltering the systemic production of hormones such as para-thyroid hormone(30) in IL-18 TG mice. Further analysis isneeded to clarify this issue.

IL-18 and IL-12 synergistically inhibit osteoclast forma-tion(31) and osteoclastic bone resorbing activity(13) in vitrothrough unknown molecules as well as IFN-�. Although weare unable to exclude the possibility that factors other thanIFN-� contribute to bone metabolism in IL-18 TG mice, ourpresent results show that the overexpression of IL-18 in thebone marrow cavity, and possibly the lymphatic tissues,leads to a decrease in the turnover rate of trabecular boneand results in an aberrant cortical bone structure (osteodys-trophia).

ACKNOWLEDGMENTS

We thank Dr Howard A Young (NCI-Frederick, Freder-ick, MD, USA) for helpful discussions regarding this manu-script. TH is supported by the Japan Chemical IndustryAssociation (Tokyo, Japan), Uehara Memorial (Tokyo, Ja-pan), Kanae (Osaka, Japan), Nagao Memorial (Tokyo, Ja-pan), Ishibashi (Tokyo, Japan), and Mitsui Medical SciencePromotion (Tokyo, Japan) Foundations, and a Grant-in-Aidfor Scientific Research on Priority Areas (C) ”MedicalGenome Science“ from the Ministry of Education, Science,Sports and Culture of Japan.

REFERENCES

1. Teitelbaum SL 2000 Bone resorption by osteoclasts. Science 289:1504–1508.

2. Ducy P, Schinke T, Karsenty G 2000 The osteoblast: A sophisti-cated fibroblast under central surveillance. Science 289:1501–1504.

3. Okamura H, Tsutsi H, Komatsu T, Yutsudo M, Hakura A, Tani-moto T, Torigoe K, Okura T, Nukada Y, Hattori K, Akita K,Namba M, Tanabe F, Konishi K, Fukuda S, Kurimoto M 1995Cloning of a new cytokine that induces IFN-gamma production byT cells. Nature 378:88–91.

4. Dinarello CA 1999 Interleukin-18. Methods 19:121–132.5. Nakanishi K, Yoshimoto T, Tsutsui H, Okamura H 2001

Interleukin-18 regulates both th1 and th2 responses. Annu RevImmunol 19:423–474.

6. Olee T, Hashimoto S, Quach J, Lotz M 1999 IL-18 is produced byarticular chondrocytes and induces proinflammatory and catabolicresponses. J Immunol 162:1096–1100.

7. Hoshino K, Tsutsui H, Kawai T, Takeda K, Nakanishi K, TakedaY, Akira S 1999 Cutting edge: Generation of IL-18 receptor-deficient mice. Evidence for IL-1 receptor-related protein as anessential IL-18 binding receptor. J Immunol 162:5041–5044.

8. Born TL, Thomassen E, Bird TA, Sims JE 1998 Cloning of a novelreceptor subunit AcPL, required for interleukin-18 signaling. J BiolChem 273:29445–29450.

9. Atkins GJ, Haynes DR, Geary SM, Loric M, Crotti TN, FindlayDM 2000 Coordinated cytokine expression by stromal and hema-topoietic cells during human osteoclast formation. Bone 26:653–661.

10. Udagawa N, Horwood NJ, Elliott J, Mackay A, Owens J, OkamuraH, Kurimoto M, Chambers TJ, Martin TJ, Gillespie MT 1997Interleukin-18 (interferon-gamma-inducing factor) is produced byosteoblasts and acts via granulocyte/macrophage colony-stimulating factor and not via interferon-gamma to inhibit oste-oclast formation. J Exp Med 185:1005–1012.

11. Horwood NJ, Udagawa N, Elliott J, Grail D, Okamura H,Kurimoto M, Dunn AR, Martin T, Gillespie MT 1998 Interleu-kin 18 inhibits osteoclast formation via T cell production ofgranulocyte macrophage colony-stimulating factor. J Clin In-vest 101:595–603.

12. Makiishi-Shimobayashi C, Tsujimura T, Iwasaki T, Yamada N,Sugihara A, Okamura H, Hayashi S, Terada N 2001 Interleukin-18up-regulates osteoprotegerin expression in stromal/osteoblasticcells. Biochem Biophys Res Commun 281:361–366.

13. Yamada N, Niwa S, Tsujimura T, Iwasaki T, Sugihara A, Futani H,Hayashi S, Okamura H, Akedo H, Terada N 2002 Interleukin-18and interleukin-12 synergistically inhibit osteoclastic bone-resorbing activity. Bone 30:901–908.

14. Hoshino T, Kawase Y, Okamoto M, Yokota K, Yoshino K,Yamamura K, Miyazaki J, Young HA, Oizumi K 2001 Cuttingedge: IL-18-transgenic mice. In vivo evidence of a broad role forIL-18 in modulating immune function. J Immunol 166:7014–7018.

15. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H,Meunier PJ, Ott SM, Recker RR 1987 Bone histomorphometry:Standardization of nomenclature, symbols, and units. Report of theASBMR Histomorphometry Nomenclature Committee. J BoneMiner Res 2:595–610.

16. Coughlin CM, Salhany KE, Wysocka M, Aruga E, Kurzawa H,Chang AE, Hunter CA, Fox JC, Trinchieri G, Lee WM 1998Interleukin-12 and interleukin-18 synergistically induce murinetumor regression which involves inhibition of angiogenesis. J ClinInvest 101:1441–1452.

17. Gerber HP, Ferrara N 2000 Angiogenesis and bone growth. TrendsCardiovasc Med 10:223–228.

18. Goldring MB 1987 Control of collagen synthesis in human chon-drocyte cultures by immune interferon and interleukin-1. J Rheu-matol 14:64–66.

19. Nii A, Reynolds DA, Young HA, Ward JM 1997 Osteochondro-dysplasia occurring in transgenic mice expressing interferon-gamma. Vet Pathol 34:431–441.

20. Takahashi N, Mundy GR, Roodman GD 1986 Recombinant hu-man interferon-gamma inhibits formation of human osteoclast-likecells. J Immunol 137:3544–3549.

21. Takayanagi H, Ogasawara K, Hida S, Chiba T, Murata S, Sato K,Takaoka A, Yokochi T, Oda H, Tanaka K, Nakamura K, TaniguchiT 2000 T-cell-mediated regulation of osteoclastogenesis by signal-ling cross- talk between RANKL and IFN-gamma. Nature 408:600–605.

22. Young HA, Klinman DM, Reynolds DA, Grzegorzewski KJ, NiiA, Ward JM, Winkler-Pickett RT, Ortaldo JR, Kenny JJ,Komschlies KL 1997 Bone marrow and thymus expression ofinterferon-gamma results in severe B-cell lineage reduction. T-celllineage alterations, and hematopoietic progenitor deficiencies.Blood 89:583–595.

23. Khosla S 2001 Minireview: The OPG/RANKL/RANK system.Endocrinology 142:5050–5055.

24. Manabe N, Kawaguchi H, Chikuda H, Miyaura C, Inada M, NagaiR, Nabeshima Y, Nakamura K, Sinclair AM, Scheuermann RH,Kuro-o M 2001 Connection between B lymphocyte and osteoclastdifferentiation pathways. J Immunol 167:2625–2631.

25. Ushio S, Namba M, Okura T, Hattori K, Nukada Y, Akita K,Tanabe F, Konishi K, Micallef M, Fujii M, Torigoe K, TanimotoT, Fukuda S, Ikeda M, Okamura H, Kurimoto M 1996 Cloning ofthe cDNA for human IFN-gamma-inducing factor, expression inEscherichia coli, and studies on the biologic activities of theprotein. J Immunol 156:4274–4279.

26. Nakamura S, Otani T, Okura R, Ijiri Y, Motoda R, Kurimoto M,Orita K 2000 Expression and responsiveness of human

982 KAWASE ET AL.

interleukin-18 receptor (IL-18R) on hematopoietic cell lines. Leu-kemia 14:1052–1059.

27. Paul WE 1991 Interleukin-4: A prototypic immunoregulatory lym-phokine. Blood 77:1859–1870.

28. Wei S, Wang MW, Teitelbaum SL, Ross FP 2002 Interleukin-4reversibly inhibits osteoclastogenesis via inhibition of NF-kappa Band mitogen-activated protein kinase signaling. J Biol Chem 277:6622–6630.

29. Kasono K, Sato K, Sato Y, Tsushima T, Shizume K, Demura H1993 Inhibitory effect of interleukin-4 on osteoclast-like cellformation in mouse bone marrow culture. Bone Miner 21:179 –188.

30. Lewis DB, Liggitt HD, Effmann EL, Motley ST, Teitelbaum SL,Jepsen KJ, Goldstein SA, Bonadio J, Carpenter J, Perlmutter RM1993 Osteoporosis induced in mice by overproduction of interleu-kin 4. Proc Natl Acad Sci USA 90:11618–11622.

31. Horwood NJ, Elliott J, Martin TJ, Gillespie MT 2001 IL-12 aloneand in synergy with IL-18 inhibits osteoclast formation in vitro.J Immunol 166:4915–4921.

Address reprint requests to:Tomoaki Hoshino, MD, PhD

Department of Internal Medicine 1Kurume University School of Medicine

67 Asahi-machiKurume 830-0011, Japan

E-mail: hoshino@med.kurume-u.ac.jp

Received in original form August 8, 2002; in revised form October22, 2002; accepted December 13, 2002.

983BONE MALFORMATIONS IN IL-18 TRANSGENIC MICE