1

TNFα but not IL-17 is critical in the pathogenesis of rheumatoid arthritis

spontaneously occurring in a unique FcγRIIB-deficient mouse model

Hideki Okazaki1,2, Qingshun Lin1, Keiko Nishikawa1, Naomi Ohtsuji1, Hiromichi

Tsurui1, Mareki Ohtsuji3, Hirofumi Amano4, Norihiro Tada5, Katsuko Sudo6, Hiroyuki

Nishimura3, Toshikazu Shirai1, and Sachiko Hirose1

1Department of Pathology, 4Department of Internal Medicine, 5Atopy Research Center,

Juntendo University School of Medicine, Tokyo 113-8421, Japan 2Health and Life Science, Musashigaoka Junior College, Saitama 355-0154, Japan 3Toin Human Science and Technology Center, Department of Biomedical Engineering,

Toin University of Yokohama, Yokohama 225-8502, Japan 6Animal Research Center, Tokyo Medical University, Tokyo 160-8402, Japan

Number of text pages: 33

Numbers of tables and figures: 2 tables and 4 figures (2 color [Figs.1 and 2] and 2

black-and-white [Figs.3 and 4])

Key words: FcγRIIB, monocytosis, osteoclastogenesis, RANKL, OPG

Address correspondence and reprint requests to Sachiko Hirose, MD, PhD, Department

of Pathology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku,

Tokyo 113-8421, Japan. Phone: +81-3-5802-1671, Fax: +81-3-3813-3164, E-mail

address: sacchi@juntendo.ac.jp

2

Abstract

Objective TNFα and IL-17 have been shown to be the major inflammatory cytokines

involved in the pathogenesis of rheumatoid arthritis (RA). Here, we examined the effect

of these cytokines on spontaneously occurring RA in our newly established

arthritis-prone FcγRIIB-deficient C57BL/6 (B6) mice, designated KO1, by introducing

genetic deficiency of TNFα and IL-17 into KO1 mice.

Methods KO1.TNFα-/- and KO1.IL-17-/- mice were established by crossing KO1 with

TNFα-deficient and IL-17-deficient B6 mice, respectively. The incidence and severity

of RA, cartilage and bone destruction, immunological abnormalities, and transcription

levels of receptor activator of NF-κB ligand (RANKL)/osteoprotegerin (OPG) and

inflammatory cytokines/chemokines in ankle joints were compared among KO1,

KO1.TNFα-/-, and KO1.IL-17-/- mice.

Results The development of RA was completely inhibited in KO1.TNFα-/- mice. In

contrast, KO1.IL-17-/- mice unexpectedly developed severe RA comparable to KO1.

Compared with those in KO1 and KO1.IL-17-/- mice, frequencies of peripheral

monocytes, known to be containing osteoclast precursors, were significantly decreased

in KO1.TNFα-/- mice. Intriguingly, while RANKL expression levels in ankle joints did

not differ among the three strains, OPG expression levels were drastically decreased in

arthritis-prone, but not arthritis-free, mice. The expression levels of inflammatory

cytokines/chemokines, such as MCP-1, IL-6, and TNFα, were up-regulated in

arthritis-prone mice.

Conclusion TNFα is indispensable while IL-17 is dispensable in the pathogenesis of

RA in KO1 mice. In this model, TNFα may contribute to the development of arthritis,

through mediating the increase in frequencies of osteoclast precursors in circulation and

their migration into the joints, and the decrease in OPG expression, leading to the

up-regulated osteoclastogenesis associated with severe cartilage and bone destruction.

3

Introduction

Rheumatoid arthritis (RA) is one of the most serious systemic autoimmune diseases,

characterized by marked synovial hyperplasia associated with inflammatory cell

infiltration in multiple synovial joints, followed by the progressive destruction of

cartilage and bone. Accumulating evidence shows that the generation and activation of

osteoclasts in inflamed joint tissues are essential for bone loss in RA. Osteoclasts are

multinucleated giant cells positive for tartrate-resistant acid phosphatase (TRAP) and

cathepsin K, and resorb bone matrix [1, 2]. These cells differentiate from osteoclast

precursors of monocyte/macrophage lineage cells in the bone marrow [3] and in

peripheral blood [4]. The process of osteoclastogenesis is controlled by the interaction

of receptor activator of NF-κB (RANK) expressed on osteoclast precursors with its

ligand RANKL expressed on synovial fibroblasts, osteoblasts, and Th17 cells [5-7].

RANKL-mediated osteoclastogenesis is counterbalanced by the physiologically

expressed decoy receptor, osteoprotegerin (OPG) [8], and also by several cytokines

induced by activated immune cells, such as IFNγ [9].

In the last decade, TNFα has been shown to play a key role in the pathogenesis of

RA. Clinical trails have shown that blockade of TNFα is successful in the treatment of

4

RA [10-12] and is able to modulate the RANKL/OPG system in favor of bone

formation [13]. However, the therapies are not always effective in all patients [14].

Experimentally, anti-TNFα therapy has been shown to be effective in type II

collagen-induced arthritis (CIA) mouse model [15, 16]. Furthermore, TNFα-deficient

mice showed a significantly reduced degree of CIA arthritis; however, severe disease

still developed in some individuals [17]. In the K/BxN serum transfer arthritis model, a

proportion of TNFα-deficient mice were reported to develop robust disease [18].

Intriguingly, the TNFα deficiency did not affect disease development at all in arthritis

observed in an RA model of human T-cell leukemia type I (HTLV-1) transgenic mice

[19].

Recently, the cytokine IL-17 has attracted attention in relation to the pathogenesis

of RA. IL-17 is a proinflammatory cytokine, produced by a variety of immune cells,

such as Th17 cells, macrophages, NK cells, NKT cells and γδ-T cells [20]. Since the

discovery of IL-17, many studies have shown that, while IL-17 plays a role in host

defense in many infectious diseases [21], it promotes the inflammatory process in

several autoimmune diseases, such as RA, systemic lupus erythematosus (SLE),

inflammatory bowel disease and experimental autoimmune encephalomyelitis [22].

5

Among these, RA has been intensively studied because Th17 cells can express RANKL

and induce osteoclastogenesis [7]. Early studies showed that high levels of IL-17 are

detected in the synovial tissues from RA patients, and that the culture medium of RA

synovial tissues induces osteoclast generation in vitro [23].

To confirm the roles of IL-17 in RA pathogenesis, the genetic deficiency of IL-17

was introduced into several murine RA models. The development of arthritis observed

in RA models of IL-1 receptor antagonist-deficient mice [24] and ZAP-70-mutant SKG

mice [25] was completely inhibited, and the arthritis in HTLV-1 transgenic mice [19]

and CIA mice [26] was moderately inhibited. In contrast, in the case of human cartilage

proteoglycan-induced arthritis, both inflammation and bone erosion were not suppressed

in IL-17 deficiency [27].

Since RA is a multifactorial disease under the control of multiple susceptibility

genes and environmental factors [28] and since the same RA phenotype may appear

under different mechanisms due to different sets of susceptibility genes and

environmental factors, further studies are required to clarify the mechanism of the

effects of RA-related proinflammatory cytokines using several kinds of RA mouse

models. We recently found that a C57BL/6 (B6) mouse strain genetically deficient in

6

inhibitory IgG Fc receptor IIB (FcγRIIB), designated KO1, spontaneously develops

severe RA closely resembling human RA [29]. Our model is unique, because another

FcγRIIB-deficient B6 strain of mice has been reported to develop severe SLE but not RA

[30, 31]. The mechanism responsible for this disease phenotype difference remains

undetermined; however, it is likely that these two sublines have the shared and

disease-specific predisposition toward RA and SLE. Indeed, we found that the

introduction of autoimmune susceptible Yaa mutation into KO1 mice causes the disease

phenotype conversion from RA to SLE [32]. In the present study, we examined the

effects of IL-17 and TNFα on RA using our unique arthritis-prone mouse model by

introducing genetic deficiency of IL-17 and TNFα.

Materials and methods

Mice

Arthritis-prone KO1 is an FcγRIIB-deficient B6 congenic line [29], obtained by

backcrossing the originally constructed FcγRIIB-deficient mice on a hybrid (129 x B6)

background [33] into a B6 background for over 12 generations. KO1.IL-17-/- mice were

established by crossing KO1 with IL-17-/- mice on a B6 background [34], a kind gift

7

from Dr. Nakae, the University of Tokyo. KO1.TNFα-/- mice were established by

crossing KO1 with TNFα-/- mice on a B6 background [35], obtained from Dr. Iwakura,

Tokyo University of Science. Female mice were analyzed in the present study. All mice

were housed under identical specific pathogen-free conditions, and all experiments were

performed in accordance with our institutional guidelines.

Scoring of arthritis

Ankle joint swelling was examined by inspection and arbitrarily scored as follows: 0, no

swelling; 1, mild swelling; 2, moderate swelling; and 3, severe swelling. Scores for both

ankle joints were summed for each mouse, and mice with the score 2 or over were

considered positive for arthritis.

Histopathology and tissue immunofluorescence

For histological examination, joint tissues were decalcified in 10% EDTA in 0.1 M Tris

buffer (pH 7.4), and fixed in 4% paraformaldehyde and embedded in paraffin. Tissue

sections were stained with hematoxylin/eosin, and also stained for TRAP using

TRAP/ALP stain kit (Wako Pure Chemical Industries Ltd.). For immunofluorescence of

spleen, tissues were embedded in Tissue-Tek OCT compound and frozen in liquid

nitrogen. Frozen sections were three-color stained for 30 min at room temperature with

8

Alexa 488-labeled anti-CD4 and anti-CD8 mAbs, Alexa 647-labeled anti-B220 mAb

and Alexa 546-labeled PNA. Antibodies and PNA were purchased from BD Pharmingen

and Vector Laboratories Inc. (Burlingame, CA, USA), respectively, and the labeling of

these reagents was carried out in our laboratory. Color images were obtained using laser

scanning microscopy (LSM510META Ver. 3.2, Carl Zeiss Co., Ltd., Germany).

Serum levels of antibodies

Serum levels of IgG class anti-cyclic citrullinated peptide (CCP) antibodies and

rheumatoid factor (RF) were measured employing commercially available kits (Cosmic

Corporation and Shibayagi, respectively), and are expressed as optical density.

Serum levels of IgG antibodies against type II collagen (CII) and double-stranded

(ds) DNA were measured using an ELISA plate pre-coated with 10 µg/ml bovine CII

(Sigma-Aldrich) and 5 µg/ml bovine dsDNA (Sigma-Aldrich). CII-binding activities are

expressed as optical density. DNA-binding activities are expressed in units, as

previously described [29].

Flow cytometric analysis

Spleen cells were four-color stained with the following mAbs: anti-B220, anti-CD69,

anti-CD138, anti-CD4, anti-ICOS, and ant-PD1, and with peanut agglutinin (PNA). For

9

peripheral monocyte staining, peripheral leukocytes were four-color stained with the

following mAbs: anti-CD11b, anti-Gr1, anti-FcγRIV (9E9), anti-CCR2, and

anti-CX3CR1. Stained cells were analyzed using a FACSAria cytometer and FlowJo

software (Tree Star Inc.).

Quantitative real-time PCR (qRT-PCR) analysis

Total RNA was isolated from ankle joints using QIAGEN RNeasy Lipid Tissue Minikit

(Cat. No. 74804). Briefly, ~25 mg of ankle joint tissue was added in 500 µl of QIAzol

lysis reagent in a 2 ml tube containing 5-mm-diameter zirconia beads (Hirasawa YTZ-5)

and homogenized on TissueLyser (Qiagen) for 1 min at 30 Hz. Total RNA was

extracted from homogenized materials using Minikit according to the manufacturer’s

instructions, and the single-stranded cDNA was synthesized using an oligo(dT)-primer

with Superscript II First-Strand Synthesis kit (Invitrogen, Carlsbad, CA). The cDNA

product was used for qRT-PCR. The data were normalized to β-actin reference. The

primer pairs used and the length of PCR products are shown in Table 1. The quantity

was normalized using the formula of the 2-∆∆CT method. Values of B6 mice were

designated as 1, and values of KO1, KO1.IL-17-/- and KO1.TNFα-/- mice are evaluated

as fold change compared with the values in B6 mice.

10

Statistics

Statistical analysis was carried out using Mann-Whitney’s U test for antibody levels and

qRT-PCR analysis, and Student’s t-test for flow cytometric analysis. A value of P < 0.05

was considered as statistically significant.

Results

Comparison of disease incidence and severity among KO1, KO1.IL-17-/-, and

KO1.TNFα-/- mice

Fig. 1a compares the cumulative incidence and severity of arthritis in ankle joints

among the three strains. KO1 mice spontaneously developed severe arthritis with

swelling and limited mobility of ankle joints symmetrically after 5 months of age, and

severity of arthritis increased with aging. The arthritis observed in KO1 mice was

almost completely inhibited in KO1.TNFα-/- mice and there was no evidence of arthritis

even by 12 months of age. In contrast, KO1.IL-17-/- mice unexpectedly developed

severe arthritis with comparable incidence and severity to that observed in KO1 mice.

Fig. 1b shows representative macroscopic findings of hind paws of KO1, KO1.IL-17-/-,

and KO1.TNFα-/- mice at 10 months of age. Histological examination revealed severe

11

synovitis with destruction of cartilage and bone due to remarkable pannus formation in

KO1 and KO1.IL-17-/-, but not KO1.TNFα-/-, mice (Fig. 1c upper panel). TRAP staining

showed an increase in the number of TRAP-positive osteoclasts at the erosion site of

cartilage and bone in KO1 and KO1.IL-17-/- mice (Fig. 1c lower panel).

Activation status of lymphocytes and serum autoantibody levels

Fig. 2a shows immunohistochemical findings of spleen sections. Age-associated

germinal center (GC) formation was found in KO1 and KO1.IL-17-/- mice. In contrast,

spontaneous GC formation was not observed in KO1.TNFα-/- mice, a finding consistent

with a previous report [35]. Table 2 compares the spleen weight and the frequencies of

lymphocyte subsets among the three strains. KO1 mice developed splenomegaly, and

the spleen weight was significantly suppressed in KO1.IL-17-/- and KO1.TNFα-/- mice.

As for B-cell subsets, while there was no difference in the frequencies of total B cells

among the three strains, the frequencies of CD69+ activated B cells and PNA+ GC B

cells per total B cells were significantly lower in KO1.TNFα-/- mice than those in KO1

mice. These differences were not observed between KO1 and KO1.IL-17-/- mice. There

was no difference in the frequencies of CD138+ plasma cells, total T cells, CD69+

activated CD4+ T cells and ICOS+PD1+-phenotype CD4+ T cells among the three

12

strains.

Fig. 2b compares serum levels of IgG class RF and IgG class autoantibodies

against CCP, CII and dsDNA at 8 months of age. Levels of all autoantibodies in KO1,

KO1.IL-17-/- and KO1.TNFα-/- mice were significantly higher than those in normal B6

mice. There was no significant difference in these autoantibody levels among KO1,

KO1.IL-17-/- and KO1.TNFα-/- mice, indicating that there was no association between

the levels of these autoantibodies and the development of RA in our mouse models.

Frequencies of peripheral monocytes

To clarify the mechanism involved in the disease suppression in KO1.TNFα-/- mice, we

then examined the frequency of peripheral CD11b+ monocytes, since peripheral

monocytes contain osteoclast precursors [4]. As shown in Fig. 3a, flow cytometric

analysis revealed that the frequencies of CD11b+ monocytes of especially Gr1- subsets

were increased in KO1 and KO1.IL-17-/- mice at 8~9 months of age; however, this was

not observed in KO1.TNFα-/- mice. Fig. 3b showed the significant difference in the

frequencies of CD11b+ monocytes between arthritis-prone and arthritis-free mice. Fig.

3c shows the expression levels of stimulatory IgG Fc receptor, FcγRIV, and chemokine

receptors, CCR2 and CX3CR1, in the CD11b+ gated cell population, indicating that

13

Gr1+ monocytes were FcγRIV-CCR2+CX3CR1dull, while Gr1- monocytes were

FcγRIV+CCR2-CX3CR1high. These two phenotypically different subsets correspond to

the previously reported monocyte subsets; the former is the “inflammatory” subset that

has the capacity to migrate to the inflamed tissues because of the high expression of

CCR2 (MCP-1 receptor), and the latter is the “resident” subset that is recruited to the

non-inflamed tissues [4, 36].

qRT-PCR analysis of RANKL/OPG and cytokine/chemokine mRNA expression

levels in ankle joints

To evaluate the mRNA expression levels, qRT-PCR analysis was performed using

mRNA extracted from ankle joint tissues of each of the KO1, KO1.IL-17-/- and

KO1.TNFα-/- mice, as well as normal B6 mice as a relative control. We first examined

the relationship between RANKL/OPG expression levels and the development of RA.

The expression levels in KO1, KO1.IL-17-/- and KO1.TNFα-/- mice were evaluated as

fold change compared with the level in B6 mice tentatively designated as 1. As shown

in Fig. 4a, while the RANKL expression levels in the three strains were similar to that in

normal B6 mice and there was no strain difference in RANKL expression levels, OPG

expression level was significantly suppressed in KO1 and KO1.IL-17-/- mice with

14

severe arthritis, but not in arthritis-free KO1.TNFα-/- mice.

As MCP-1 and RANTES are known to be the major chemokines involved in the

pathogenesis of RA [37], we then compared the expression levels of MCP-1 and

RANTES in ankle joints. As shown in Fig. 4b, the MCP-1 expression level was

markedly up-regulated in arthritis-prone KO1 and KO1.IL-17-/- mice, compared with

the finding in arthritis-free KO1.TNFα-/- mice. In contrast, there was no difference in

RANTES expression levels among the three strains of mice.

Fig. 4C compares the expression levels of IL-6, IL-10, IFNγ and TNFα among the

three strains of mice. IL-6 expression level was up-regulated in arthritis-prone KO1 and

KO1.IL-17-/- mice, but the level in KO1.TNFα-/- mice was similar to that in normal B6

mice. IL-10 expression levels in the three strains were higher than that in B6 mice;

however, there was no difference in expression levels among the three strains.

Intriguingly, IFNγ expression level was up-regulated in arthritis-free KO1.TNFα-/ mice.

The levels of TNFα were up-regulated in arthritis-prone KO1 and KO1.IL-17-/- mice.

As for IL-17 expression, there was no detectable level in ankle joint tissues from all the

mouse strains examined.

15

Discussion

Disease features of RA were completely inhibited in KO1 mice with TNFα deficiency,

indicating a pivotal role of TNFα in the pathogenesis of RA in KO1 mice. In contrast, RA

features were seldom suppressed in IL-17-deficient KO1 mice. On the basis of studies

using several kinds of the arthritis mouse model, it has been shown that IL-17 plays a

central role in the development of RA through activating synoviocytes, fibroblasts,

endothelial cells and macrophages to induce TNFα, which subsequently induces

inflammation and osteoclast differentiation in multiple synovial joints [14, 22]. However,

the present study showed that severe RA developed in KO1 mice in the absence of IL-17,

showing that IL-17 is dispensable in the pathogenesis of RA in this model.

qRT-PCR analysis revealed that there was no detectable level of IL-17 mRNA

expression in arthritic joints of arthritis-prone KO1 and non-arthritis KO1.TNFα-/- mice.

This is consistent with our earlier findings in flow cytometric analysis of

cytokine-producing CD4+ T cells in inflamed synovial tissues from KO1 mice, revealing

that nearly 30% of the cells were positive for TNFα, while IL-17-producing cells were

seldom observed [29]. The current studies clearly indicated that several different

mechanisms underlie the processes of joint inflammation and osteoclastogenesis in

16

clinically similar RA individuals. Recent phase II clinical studies of RA patients under

anti-IL-17 antibody treatments showed variable results [38], suggesting the presence of

different disease mechanisms underlying in individual patients.

Recently, Pisitkun et al. [39] reported that the IL-17 deficiency protected SLE-prone

FcγRIIB-deficient B6 mice from kidney damage, even though this did not prevent the

spontaneously developing GC formation and autoantibody production. The protection

from kidney disease was suggested to be due to the decrease in inflammatory cell

infiltration in the kidney. Consistent finding in our model is no influence on the

spontaneous formation of GCs and the production of autoantibodies; however, the IL-17

deficiency did not protect KO1 mice from RA with respect to its incidence, severity, and

histopathology of the joint disease. Taken collectively, it is clear that the level of IL-17

production shows differential effects between SLE and RA seen in the two different

sublines of FcγRIIB-deficient B6 mice. As these two sublines of FcγRIIB-deficient B6

mice appear to carry similar genetic background, different environmental factors may

underlie the difference in their disease phenotypes. Intriguingly, Ivanov et al. [40]

reported that B6 mice obtained from different commercial venders may have different

commensal intestinal bacteria, and that this difference affects the immune system, most

17

strikingly, IL-17 production, which may possibly affect the specificity and severity of

autoimmune diseases.

Discrepancy between the levels of autoantibody production and GC formation in the

spleen was observed in KO1.TNFα-/- mice. KO1.TNFα-/- mice had no GC formation, in

keeping with previous reports [35, 41]. Despite the lack of GC formation,

KO1.TNFα-/- mice showed comparable serum levels of autoantibodies to those found in

KO1 and KO1.IL-17-/- mice. These findings suggest that the production of autoantibodies

could be performed by extrafollicular B cells, as reported elsewhere [42]. TNFα

stimulates the vascular endothelium to express adhesion molecules and to accelerate the

inflammatory cell migration from circulation into the inflammatory site [19]. Thus,

KO1.TNFα-/- mice did not develop arthritis because of the lack of inflammatory cell

infiltration, irrespective of the deposition of ICs in the joint tissue.

FcγRIIB molecules are expressed not only on B cells but also on

monocytes/macrophages, and negatively regulate their activation mediated by

stimulatory FcγRs and cytokine/chemokine production [43, 44]. Analysis of mice

deficient in stimulatory FcγRs demonstrated that monocyte frequencies are increased as a

consequence of IC-triggered activation through stimulatory FcγRs [45]. Lack of FcγRIIB

18

expression may strengthen IC-triggered activation signals from stimulatory FcγRs and

accelerate monocytosis. In the present study, age-associated monocytosis was evident in

KO1 and KO1.TIL-17-/-, but not KO1.TNFα-/-, mice. It has been shown that monocytosis

is one of the characteristic features in SLE-prone mice and well associates with

autoantibody production and lupus nephritis [46]. Our present study revealed that

monocytosis is also associated with the development of RA, suggesting that peripheral

monocytes might be a potential therapeutic target for RA. Monocytosis did not develop in

KO1.TNFα-/- mice, irrespective of the lack of FcγRIIB expression and of comparable

serum levels of autoantibodies as those found in KO1 and KO1.IL-17-/- mice. Yao et al.

[47] reported that the administration of TNFα induces an increase in circulating CD11b+

monocytes. This event is reversible, and anti-TNF therapy reduces monocytosis, owing to

inhibition of monocyte mobilization from the bone marrow. Thus, it appears that, in

KO1.TNFα-/- mice, monocytosis does not develop even in the presence of circulating ICs.

It has been shown that monocytes in peripheral blood contain osteoclast precursors

[4]. Peripheral monocytes are composed of heterogeneous populations and could be

subdivided into two phenotypically and functionally distinct subsets in mice, namely a

CD11b+CX3CR1lowCCR2+Gr1+ “inflammatory” subset that is actively recruited to

19

inflamed tissues and a CD11b+CXCR3highCCR2-Gr1- “resident” subset that is recruited to

non-inflamed tissues [4, 36]. Intriguingly, this latter subset is selectively expanded in

lupus-prone mice and expresses the stimulatory FcγRIV, suggesting that this subset may

be involved in IgG-mediated kidney damage [45]. Furthermore, there is a report that the

“resident” subset gives rise to specialized cell types, including osteoclasts [47]. Thus, the

originally classified Gr1- “resident” subset may contain functionally different cell types,

one responsible for lupus nephritis and the other osteoclastogenesis in joints. However,

further studies are required for the role of the CCR2+Gr1+ “inflammatory” subset in

osteoclastogenesis, since MCP-1, the CCR2 ligand, is shown to play a role in promoting

the fusion of CCR2+ osteoclast precursors, resulting in an acceleration of multinucleated

osteoclast generation [37].

An additional important finding in the present study is that the development of

arthritis is associated with OPG, but not RANKL, expression levels. There was no

difference in RANKL expression levels among KO1, KO1.IL-17-/- and

KO1.TNFα-/- mice, while OPG expression levels in arthritis joints in KO1 and

KO1.IL-17-/- mice were markedly decreased, but those in KO1.TNFα-/- mice were at

normal levels comparable to those in B6 mice. These findings suggest that the balance

20

between RANKL and OPG determines the degree of osteoclastogenesis, in keeping with

the findings reported in studies of RA patients under anti-TNF therapy [13].

OPG is produced by a variety of cells including bone marrow stromal cells, B cells,

dendritic cells and vascular endothelial cells [48, 49]. Marked decrease in OPG

expression in arthritic joints is suggested to be due to the exhaustion of OPG production

over extended periods of chronic stimulation by TNFα [49]. In addition, although the

precise mechanism for up-regulated expression of IFNγ in joint tissues in

KO1.TNFα-/- mice remains unknown, this may also contribute to the inhibitory effect of

osteoclastogenesis, since IFNγ inhibits RANK signaling by accelerating the degradation

of TRAF6 through activation of the ubiquitin/proteasome system [9]. Thus, TNFα may

contribute to the development of arthritis in KO1 mice, through mediating the increase in

frequencies of osteoclast precursors in circulation and their migration into the joints, and

the decrease in OPG expression, leading to the up-regulation of osteoclastogenesis.

Taken collectively, our RA mouse model is useful for obtaining a more thorough

understanding of the complex disease process of RA, and for obtaining precise

therapeutic targets for tailor-made therapy of such a complex disease.

21

Acknowledgments We thank Dr. A. Sato-Hayashizaki and Ms. K. Kojo for excellent

support in establishing IL-17-deficient and TNFα-deficient KO1 mice. We also thank

Drs. T. Ikegami and T. Koga for excellent technical support in qRT-PCR and TRAP

staining, respectively. This work was supported in part by a Grant-in-Aid (24590491)

from the Ministry of Education, Science, Technology, Sports and Culture of Japan, a

grant for Research on Intractable Diseases from the Ministry of Health, Labour and

Welfare of Japan, and a grant from the Strategic Japanese-Swiss Cooperative Program.

The authors have no conflicting financial interests.

Conflict of interest None.

References

1. Connor JR, Dodds RA, James IE, Gowen M. Human osteoclast and giant cell

differentiation: the apparent switch from NSE to TRARP activity coincides with the

in situ expression of osteopontin mRNA. J Histochem Cytochem.

22

1995;43:1193-201.

2. Dodds RA, Connor JR, Drake FH, Gowen M. Expression of cathepsin K messanger

RNA in giant cells and their precursors in human osteoarthritic synovial tisseus.

Arthritis Rheum. 1999;42:1588-93.

3. Boyle W, Simonet WS, Lacey DL. Osteoclast differentiation and activation. Nature.

2003;423:337-42.

4. Gordon S, Taylor PR. Monocyte and macrophage heterogeneity. Nat Rev Immunol.

2005;5:953-64.

5. Takayanagi H. Osteoimmunology and the effect of the immune system on bone. Nat

Rev Rheumatol. 2009;5:667-76.

6. Walsh MC, Kim N, Kadono Y, Rho J, Lee SY, et al. Osteoimmunology: interplay

between the immune system and bone metabolism. Annu Rev Immunol.

2006;24:33-63.

7. Sato K, Suematsu A, Okamoto K, Yamaguchi A, Morishita Y, Kadono Y, et al. Th17

functions as an osteoclastogenic helper T cell subset that linkes T cell activation

and bone destruction. J Exp Med. 2006;203: 2673-82.

8. Simonet WS, Lacey DL, Dunstan CR, Kelley M, Chang MS, Luthy R, et al.

23

Osteoprotegerin: a novel secreted protein involved in the regulation of bone density.

Cell. 1997;89:309-19.

9. Takayanagi H, Ogasawara K, Hida S, Chiba T, Murata S, Sato K, et al,

T-cell-mediated regulation of osteoclastogenesis by signaling cross-talk between

RANKL and IFN-γ, Nature. 2000;408:600-5.

10. Maini RN, Elliott MJ, Brennan FM, Feldmann M. Beneficial effects of tumor

necrosis factor-alpha (TNF-α) blockade in rheumatoid arthritis (RA). Clin Exp

Immunol. 1995;101:207-12.

11. Taylor PC. Anti-tumor necrosis factor therapies. Curr Opin Rheumatol.

2001;13:154-9.

12. Zwerina J, Redlich K, Schett G, Smolen JS. Pathogenesis of rheumatoid arthritis:

targeting cytokines. Ann N Y Acad Sci. 2005;1051:715-29.

13. Catrina AI, af Klint E, Ernestam S, Catrina S-B, Makrygiannakis D, Botusan IR, et

al. Anti-tumor necrosis factor therapy increases synovial osteoprotegerin expression

in rheumatoid arthritis. Arthritis Rheum. 2006;54:76-81.

14. Voll RE, Kalden JR. Do we need new threatment that goes beyoun tumor necrosis

factor blockers for rheumatoid arthritis? Ann N Y Acad Sci. 2005;1051:799-810.

24

15. Thorbecke GJ, Shah R, Leu CH, Kuruvilla AP, Hardison AM, Palladino MA.

Involvement of endogenous tumor necrosis factor α and transforming growth factor

β during induction of collagen type II arthritis in mice. Proc Natl Acad Sci USA.

1992;89:7375–79.

16. Piguet PF, Grau GE, Vesin C, Loetscher H, Gentz R, Lesslauer W. Evolution of

collagen arthritis in mice is arrested by treatment with antitumour necrosis factor

(TNF) antibody or a recombinant soluble TNF receptor. Immunology.

1992;77:510–14.

17. Campbell IK, O’Donnell K, Lawlor KE, Wicks IP. Severe inflammatory arthritis

and lymphadenopathy in the absence of TNF. J Clin Invest. 2001;107:1519-27.

18. Ji H, Pettit A, Ohmura K, Oriz-Lopez A, Duchatelle V, Degott C, et al. Critical roles

for interleukin 1 and tumor necrosis factor a in antibody-induced arthritis. J Exp

Med. 2002;196:77-85.

19. Iwakura Y, Nakae S. Saijo S, Ishigame H. The roles of IL-17A in onflammatory

immune responses and host defense against pathogens. Immunol Rev.

2008;226:57-79.

20. Korn T, Bettelli E, Oukka M, Kuchroo VK. IL-17 and Th17 cells. Ann Rev Immnol.

25

2009;27:485-517.

21. O’Quinn D, Palmer M, Lee Y, Weaver C. Emergence of the Th17 pathway and its

role in host defense. Adv Immunol. 2008;99:115–63.

22. Onishi RM, Gaffen SL. Interleukin-17 and its target genes: mechanisms of

interleukin-17 function in disease. Immunology. 2010;129:311-21.

23. Kotake S, Udagawa N, Takahashi N, Matsuzaki K, Itoh K, Ishiyama S, et al. IL-17

in synovial fluids from patients with rheumatoid arthritis is a potent stimulator of

osteoclastgenesis. J Clin Invest. 1999;103:1345-52.

24. Nakae S, Saijo S, Horai R, Sudo K, Mori S, Iwakura Y. IL-17 production from

activated T cells is required for the spontaneous development of arthritis in mice

deficient in IL-1 receptor antagonist. Proc Natl Acad Sci USA. 2003;100:5986-90.

25. Hirota K, Hashimoto M, Yoshitomi H, Tanaka S, Nomura T, Yamaguchi T, et al. T

cell self-reactivity forms a cytokine milieu for spontaneous development of IL-17+

Th cells that cause autoimmune arthritis. J Exp Med. 2007;204:41-7.

26. Nakae S, Nambu A, Sodo K, Iwakura Y. Suppression of immune induction of

collagen-induced arthritis in IL-17-deficient mice. J Immunol. 2003;171:6173-7.

26

27. Doodes PD, Cao Y, Hamel KM, Wang Y, Farkas B, Iwakura Y, et al. Development

of proteoglycan-induced arthritis is independent of IL-17. J Immunol.

2008;181:329-37.

28. Delgado-Vega A, Sánchez E, Löfgren S, Castillejo-López C, Alarcón-Riquelme ME.

Recent findings on genetics of systemic autoimmune diseases. Curr Opin Immunol.

2010;22:698–705.

29. Sato-Hayashizaki A, Ohtsuji M, Lin Q, Hou R, Ohtsuji N, Nishikawa K, et al.

Presumptive role of 129 strain-derived Sle16 locus for rheumatoid arthritis in a new

mouse model with FcγRIIB-deficient C57BL/6 genetic background. Arthritis

Rheum. 2011;63:2930-8.

30. Bolland S, Ravetch JV. Spontaneous autoimmune disease in FcγRIIB-deficient mice

results from strain-specific epistasis. Immunity. 2000;13:277-85.

31. Nimmerjahn F, Ravetch JV. Antibody-mediated modulation of immune responses.

Immunol Rev. 2010;236:265-75.

32. Kawano S, Lin Q, Amano H, Kaneko T. Nishikawa K, Tsurui H, et al. Phenotype

conversion from rheumatoid arthritis to systemic lupus erythematosus by

27

introduction of Yaa mutation into FcγRIIB-deficient G57BL/6 mice. Eur J Immunol.

2013;43:770-8.

33. Takai T, Ono M, Hikida M, Ohmori H, Ravetch JV. Augmented humoral and

anaphylactic responses in FcγRII-deficient mice. Nature. 1996;379:346-9.

34. Nakae S, Komiyama Y, Nambu A, Sudo K, Iwase M, Homma I, et al.

Antigen-specific T cell sensitization is impaired in IL-17-deficient mice, causing

suppression of allergic cellular and humoral responses. Immunity. 2002;17:375-87.

35. Taniguchi T, Takada M, Ikeda A, Momotani E. Sekigawa K. Failure of germinal

center formation and impairment of response to endotoxin in tumor necrosis factor

α-deficient mice. Lab Invest. 1997;77:647-58.

36. Geissmann F, Jung S, Littman DR. Blood monocytes consist of two principal

subsets with distinct migratory properties. Immunity. 2003;19:71-82.

37. Kim MS, Day CJ, Morrison NA. MCP-1 is induced by receptor activator of nuclear

factor-κB ligand, promotes human osteoclast fusion, and rescues granulocyte

macrophage colony-stimulating factor suppression of osteoclast formation. J Biol

Chem. 2005;280:16163-9.

38. van den Berg WB, Mclnnes IB. Th17 cells and IL-17 A-Focus on immunopathology

28

and immunotherapeutics. Semin Arthritis Rheum. 2013;43:158-70.

39. Pisitkun P, Ha H-L, Wang H, Claudio E, Tivy CC, Zhou H, et al. Interleukin-17

cytokines are critical in development of fetal lupus glomerulonephritis. Immunity.

2012;37:1104-15.

40. Ivanov II, Atarashi K, Manel N, Brodie EL, Shima T, Karaoz U, et al. Induction of

intestinal TH17 cells by segmented filamentous bacteria. Cell. 2009;139:485-98.

41. Marino MW, Dunn A, Grail D, Inglese M, Noguchi J, Richards E, et al.

Characterization of tumor necrosis factor-deficient mice. Proc Natl Aca Sci USA.

1997;97:8093-8.

42. Sweet RA, Christensen SR, Harris ML, Shupe J, Sutherland JL, Shlomchik MJ. A

new site-directed transgenic rheumatoid factor mouse model demonstrates

extrafollicular class switch and plasmablast formation. Autoimmunity.

2010;43:607-18.

43. Ravetch JF and Kinet P-J. Fc receptors. Annu Rev Immunol. 1991;9:457-92.

44. Takai T. Roles of Fc receptors in autoimmunity. Nat Rev Immunol. 2002;2:580-92.

45. Santiago-Raber M-L, Amano H, Amano E, Baudino L, Otani M, Lin Q, et al.

FcγR-dependent expansion of a hyperactive monocyte subset in lupus-prone mice.

29

Arthritis Rheum. 2009;60:2408-17.

46. Kikuchi S, Santiago-Rober M-L, Amano H, Amano E, Fossati-Jimack L, Moll T, et

al. Contribution of NZB autoimmunity 2 to Y-linked autoimmune

acceleration-induced monocytosis in association with murine systemic lupus. J

Immunol. 2006;176:3240-7.

47. Yao Z, Li P, Zhang Q, Schwarz EM, Keng P, Arbini A, et al. Tumor necrosis

factor-α increases circulating osteoclast precursor numbers by promoting their

proliferation and differentiation in the bone marrow through up-regulation of c-Fms

expression. J Biol Chem. 2006;281:11846-55.

48. Yun TJ, Chaudhary PM, Shu GL, Frazer JK, Ewings MK, Schwartz SM, et al.

OPG/FDCR-1. A TNF receptor family member, is expressed in lymphoid cells and

is up-regulated by ligating CD40. J Immunol. 1998;161:6113-21.

49. Zannettio ACW, Holding CA, Diamond P, Atkins GJ, Kostakis P, Farrugia A, et al.

Osteoprotegerin (OPG) is localized to the Weibel-Palade bodies of human vascular

endothelial cells and is physically associated with von Willebrand factor. J Cell

Physiol. 2005;204:714-23.

30

Figure legends

Fig. 1 Comparisons of arthritic changes. a The cumulative incidence of arthritis and

arthritis score (mean and SE) in KO1 (n=33), KO1.IL-17-/- (n=21) and KO1.TNFα-/-

(n=19) mice. b Representative macroscopic findings of hind paws in KO1, KO1.IL-17-/-

31

and KO1.TNFα-/- mice at 10 months of age. KO1 and KO1.IL-17-/-, but not

KO1.TNFα-/-, mice developed marked swelling and stiffness of the ankle joints. c

Representative histopathological changes in finger joints in KO1, KO1.IL-17-/- and

KO1.TNFα-/- mice at 10 months of age. KO1 and KO1.IL-17-/- mice show marked

synovitis with inflammatory cell infiltration and the destruction of bone due to pannus

formation (Pn) and TRAP-positive osteclast generation (arrowhead). There are no such

changes in KO1.TNFα-/- mice. Representative results obtained from six female mice in

each strain. Hematoxylin/eosin and TRAP staining. Bars = 50 µm.

Fig. 2 Comparisons of spontaneous germinal center formation and serum levels of

autoantibodies among KO1, KO1.IL-17-/- and KO1.TNFα-/- mice. a Frozen spleen

sections of 10-month-old mice were triple stained with a mixture of anti-CD4 and

anti-CD8 mAbs, anti-B220 mAb and PNA to examine the extent of germinal center

formation. Representative results obtained from six female mice in each strain. b Serum

levels of IgG class RF, anti-CCP, -CII and -dsDNA antibodies were compared at 8

months of age. The horizontal bar represents the mean level. Statistical significance is

shown (* P<0.05, ** P<0.01, ***P<0.001).

32

Fig. 3 Comparisons of peripheral monocyte frequencies and cell surface phenotypes of

monocytes among KO1, KO1.IL-17-/- and KO1.TNFα-/- mice at 8~9 months of age. a

Peripheral mononuclear cells were double stained with anti-CD11b and -Gr1 mAbs, and

frequencies of Gr1+CD11b+ and Gr1-CD11b+ monocytes are shown. b CD11b+

monocyte frequencies are compared (** P<0.01). c Peripheral mononuclear cells were

stained with anti-CD11b, -Gr1, -FcγRIV, -CCR2 and -CX3CR1 mAbs, and cell surface

phenotypes were examined using CD11b+-gated cells. Results were obtained from at

least four mice in each group.

Fig. 4 Comparisons of mRNA expression levels of a RANKL and OPG, b MCP-1 and

RANTES, and c IL-16. IL-10, IFNγ and TNFα in ankle joints by quantitative real-time

PCR analysis among KO1, KO1.IL-17-/- and KO1.TNFα-/- mice at 8~9 months of age.

The value of normal B6 mice was designated as 1, and values of each of the KO1,

KO1.IL-17-/- and KO1.TNFα-/- mice were evaluated as fold change compared with the

value in B6 mice. Data are shown as mean + SEM of four mice for each strain and

representative of three experiments performed. Statistical significance was shown

33

(**P<0.01, *** P<0.001).

Table 2 Spleen weight and splenic lymphocyte subpopulations in KO1, KO1.IL-17-/-

and KO1.TNFα-/- micea

KO1 KO1.IL-17-/- KO1.TNFα-/-

Spleen weight (gm) 0.28 ± 0.03 0.20 ± 0.02b 0.17 ± 0.02b

Spleen cell populations (%)

B220+B cells/total cells 60.2 ± 2.8 47.7 ± 3.3 57.4 ± 2.7

CD69+B220+B cells/total B cells 7.4 ± 0.5 8.0 ± 0.9 5.0 ± 0.8c

PNA+B220+B cells/total B cells 4.5 ± 0.9 3.9 ± 0.5 2.0 ± 0.4c

CD138+plasma cells/total cells 1.0 ± 0.2 1.6 ± 0.1 1.2 ± 0.3

CD4+T cells/total cells 17.3 ± 2.2 17.3 ± 2.8 13.3 ± 2.0

CD69+CD4+T cells/total T cells 27.7 ± 1.4 23.0 ± 2.6 27.1 ± 6.8

CD4+ICOS+PD1+T cells/total T cell 3.2 ± 1.3 2.2 ± 0.6 2.8 ± 1.0

aValues are the mean ± SEM of at least 6 female mice aged 9-10 months.bDifferences were statistically significant versus KO1 (P < 0.01).cDifferences were statistically significant versus KO1 (P < 0.05).