Severe osteopetrosis, defective interleukin-1 signalling

and lymph node organogenesis in TRAF6-deficient mice

Asuka Naito1, Sakura Azuma1, Sakae Tanaka2, Tsuyoshi Miyazaki2, Satoshi Takaki3,
Kiyoshi Takatsu3, Kazuki Nakao4, Kenji Nakamura4, Motoya Katsuki4,
Tadashi Yamamoto1 and Jun-ichiro Inoue1,*
1
Department of Oncology, The Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku,
Tokyo 108-8639, Japan
2
Department Orthopaedic Surgery, Faculty of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo 113-0033, Japan
3
Department of Immunology, 4Division of DNA Biology & Embryo Engineering, Center for Experimental Medicine,
The Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo 108-8639, Japan

Abstract
Background: TRAF6, a member of the tumour
necrosis factor receptor-associated factor family,
was first identified as a transducer of CD40 and
interleukin-1 receptor (IL-1R) signals based on the
interaction of TRAF6 with the cytoplasmic tail of
CD40 and with the IL-1R associated kinase in vitro.
However, the functions of TRAF6 in vivo remain
unidentified.
Results: We show that TRAF6¹/¹ mice exhibit severe
osteopetrosis and are defective in osteoclast forma-
tion. In vitro culture experiments revealed that
osteoclast precursor cells derived from TRAF6¹/¹
mice are unable to differentiate to functional
osteoclasts in response to osteoclast differentiation
factor (ODF). In bone marrow of TRAF6¹/¹ mice,
the number of sIgMþB220þ immature B cells is
markedly reduced while the ratio of proB to preB
cells is not affected. In contrast, development of
thymocytes is not affected. Furthermore, TRAF6¹/¹
mice are defective in lymph node organogenesis and
IL-1 signalling in thymocytes.
Conclusions: The results identify TRAF6 as an
essential component of ODF signalling pathway,
and also show that TRAF6 plays pivotal roles in
immune and inflammatory systems in vivo.

Mosialos et al. 1995; Regnier et al. 1995; Sato et al.

Introduction
Tumor necrosis factor receptor (TNFR)-associated
factor (TRAF) family of proteins are recently char-
acterized cytoplasmic adapter proteins that can interact
with the cytoplasmic tail of cell surface receptors
(Arch et al. 1998). Among the receptors that have been
demonstrated to recruit the TRAF proteins are
members of the TNFR superfamily and the inter-
leukin-1 receptor (IL-1R) family. The Epstein–Barr
virus protein LMP-1, which was shown to be involved
in virus-mediated transformation of B cells, also
associates with the TRAF proteins. Six distinct TRAF
molecules, TRAF1 to TRAF6, have been identified
(Hu et al. 1994; Rothe et al. 1994; Cheng et al. 1995;
Communicated by: Shinichi Aizawa
* Correspondence: E-mail: jinoue@ims.u-tokyo.ac.jp
1995; Cao et al. 1996; Ishida et al. 1996a; Ishida et al.
1996b; Nakano et al. 1996). All TRAFs share a
common stretch of amino acids at their carboxyl
terminus, called the TRAF domain that has been
divided into two subregions (Cheng et al. 1995). The
carboxyl-terminal TRAF-C region is highly conserved
among the members of the TRAF family and has shown
to mediate the binding of the TRAF proteins to their
associated receptors. The amino-terminal half of the
TRAF domain, TRAF-N, is predicted to form a
coiled-coil structure and has shown to mediate both
homo- and heterodimerization of the TRAF proteins.
With the exception of TRAF1, all TRAF’s contain an
amino-terminal RING finger domain and a stretch of
predicted zinc fingers. The deletion of the RING finger
domain leads to the generation of dominant-negative
TRAF mutants (Rothe et al. 1995), suggesting that the

q Blackwell Science Limited Genes to Cells (1999) 4, 353–362 353

 A Naito et al.
RING finger domain could directly interact with the
downstream molecules or with the molecules which are
required to transduce signals to the downstream
molecules. Among TRAF family of proteins, TRAF2,
TRAF5, and TRAF6 activate Jun N-terminal kinase
(JNK) and nuclear factor kB (NFkB)(Arch et al. 1998).
Recently, signal–dependent interaction of MAP Kinase
Kinase Kinase (MAP3K) and TRAF’s has been
demonstrated. TRAF6 interacts TGFb-activated
kinase 1 (TAK1) to activate NFkB inducing
kinase (NIK) and JNK (Ninomiya et al. 1999).
TRAF2 interacts with Apoptosis signal-regulating
kinase-1(ASK1) to activate JNK (Nishitoh et al.
1998). Thus, TRAF2, TRAF6, and probably
TRAF5 exert their function by activating downstream
MAP3K. In contrast, TRAF3 was never shown to
activate any kinases so far. However, TRAF3¹/¹ mice
revealed an essential role of TRAF3 in
antibody response to T cell-dependent antigens (Xu
et al. 1996).
354 Genes to Cells (1999) 4, 353–362
Among members of the TRAF family, TRAF6
contains the most divergent TRAF-C domain (about
35% sequence identity with other TRAF’s, whereas
other TRAF’s share about 60% sequence identity in
their TRAF-C domains (Cao et al. 1996; Ishida et al.
1996a)), indicating that the TRAF-C domain of
TRAF6 could recognize and bind to the unique
amino acid sequences. Consistent with this idea,
TRAF6 is the only one TRAF so far identified
that has been shown to be involved in the
signals emanating from IL-1R by interacting with the
IL-1R associated kinase (IRAK) (Cao et al. 1996).
Furthermore, TRAF2, TRAF3 and TRAF5 have been
shown to bind to the same domain in the cytoplasmic
tail of CD40, whereas TRAF6 interacts with mem-
brane-proximal domain distinct from those for other
TRAFs (Tsukamoto et al. 1999). However, it is not
known whether TRAF6 plays pivotal roles in
immune and inflammatory systems in vivo, and
whether it has additional functions yet to be identified.
Figure 1 Targeting of the TRAF6 gene.
(A) Genomic configuration of the germ-
line TRAF6 locus and construction of the
targeting vector. TRAF6 exons are shown
as filled boxes. The TRAF6 probes used
for Southern blotting and expected frag-
ment sizes after digestion of wild-type and
mutant genomic DNA by SphI are indi-
cated. E, EcoRI; S, SphI; neor, neomycin-
resistance gene; DT-A, diphtheria toxin A
fragment gene. (B) Southern blot analysis
of offspring from the heterozygote inter-
crosses. Genomic DNA was extracted
from mouse tails, digested with SphI,
electrophoresed, and hybridized with the
probe. Wild-type (9.0 kbp) and mutant
(4.5 kbp) bands are indicated. (C) Absence
of the TRAF6 protein in TRAF6¹/¹
mice. TRAF6 was first immunoprecipi-
tated from total brain lysate using rabbit
anti-TRAF6 polyclonal antibody. TRAF6
in the immune complex was detected by
Western blot using the same antibody for
immunoprecipitation. Cell lysate prepared
from 293T cell transfected with murine
TRAF6 expression vector was used as a
positive control (Transfect) (D) Absence of
mesenteric lymph nodes in TRAF6¹/¹
mice. Macroscopic view of mesenteric
lymph nodes from 15-day-old TRAF6þ/
þ
littermate mouse is shown as a control.
q Blackwell Science Limited

In order to address these questions, we have generated
TRAF6-deficient mice.
Results
Targeted disruption of TRAF6
Our gene-targeting strategy and genotyping analysis are
shown in Fig. 1A,B. The targeting vector deleted exon
4 (the second coding exon), which encodes amino acids
100–202 of TRAF6. The deleted region encodes part
of the RING finger domain and three zinc fingers that
are essential for signal transduction (Rothe et al. 1995).
Moreover, in the event of exon 3–5 splicing of primary
transcripts derived from the targeted allele, the deletion
of exon 4 would alter the reading frame and introduce
stop codon that would terminate translation after
synthesis of 167 amino acids. Neither full length nor
truncated forms of TRAF6 could be detected in brain
of TRAF6¹/¹ mice, using anti-TRAF6 antibody that
can recognize the both forms (Fig. 1C).
Of the first 254 live pups examined, only 12% were
TRAF6¹/¹, suggesting that TRAF6, like TRAF2,
plays a role in embryogenesis (Yeh et al. 1997).
However, examination of the embryos from
TRAF6þ/¹ matings revealed a normal Mendelian
ratio (30%) of TRAF6¹/¹ mice at day 14.5 post-
conception. Viable TRAF6¹/¹ mice appeared normal
at birth but became smaller than their normal
littermates by day 6. The mutant animals became
more runted with time and died at the age of 17–19 days.
At the time of death, the body weight of mutant animals
ranged from 50% to 70% of normal controls. Most organs
of TRAF6¹/¹ mice revealed no gross developmental
abnormalities with the exception of splenomegaly and
thymic atrophy (data not shown). Interestingly,
TRAF6¹/¹ mice lacked mesenteric (Fig. 1D), mandi-
bular, inguinal, axillary and para-aortic lymph nodes
(data not shown), indicating a critical role of TRAF6 in
the formation of second lymphoid organs. TRAF6þ/¹
mice did not show any defects or abnormalities
described in this paper.
Severe osteopetrosis in TRAF6¹/¹ mice
Since the most obvious phenotype of TRAF6¹/¹ mice
is failure of tooth eruption (Fig. 2A, right), which is
characteristics of osteopetrosis, a disorder of bone
remodeling caused by impaired osteoclast formation
or function (Popoff & Marks 1995), we first focused on
the bone structure. Whole body anteroposterior radio-
graphs of 12-day-old TRAF6¹/¹ mice and their
q Blackwell Science Limited
TRAF6-deficient mice
normal littermates revealed that the mutant mice had
shortened long bones that were radio-opaque and had
limited bone marrow cavities filled with spongy bone
(Fig. 2A, left and middle). Histological analysis of bone
sections from TRAF6¹/¹ mice revealed abnormal
bone formation with thickened epiphyseal growth
plates and limited marrow cavities. Even the diaphyseal
medullary cavity contained thick bony trabeculae,
confirming the presence of severe osteopetrosis
(Fig. 2B). Staining of bone sections for tartrate-resistant
acid phosphatase (TRAP) that is highly expressed in
osteoclasts (Minkin 1982) showed abundant, strongly
TRAPþ multinuclear cells in the metaphyseal bone of
wild-type mice. However, in most sections from the
bones of TRAF6¹/¹ mice, only a few weakly TRAPþ
mononuclear cells were observed, indicating that
TRAF6 is required for TRAPþ osteoclast formation
(Fig. 2C).
Inability of splenocytes from TRAF6¹/¹ mice
to differentiate to functional osteoclasts
To address whether failure of osteoclast formation is due
to defects in accessory cells (osteoblast/stromal cells) or
cells of the osteoclast lineage (Yoshida et al. 1990;
Soriano et al. 1991; Grigoriadis et al. 1994; Franzoso
et al. 1997; Iotsova et al. 1997), we have performed in
vitro culture experiments. Primary osteoblasts, derived
from calvariae of newborn normal ddY mice were
cocultured with splenocytes either from TRAF6¹/¹
mice or their normal littermates. Functional osteoclasts
were formed from wild-type precursors present in the
spleen as determined by TRAP positivity (Fig. 3A) and
the formation of resorption pits (Fig. 3C) in the
presence of either 1,25 dihydroxyvitamin D3 (Vitamin
D3) or Prostaglandin E2 (PGE2). However, mature
osteoclasts were not formed from splenocytes of
TRAF6¹/¹ mice. Although small number of TRAPþ
cells were observed, they were small, mononuclear and
nonfunctional (Fig. 3A and 3C). It has been shown that
Vitamin D3 and PGE2 act on osteoblasts to produce
osteoclast differentiation factor (ODF also known as
OPGL, RANKL or TRANCE) (Lacey et al. 1998;
Yasuda et al. 1998). Thus, functional osteoclasts were
induced when splenocytes were cultured in the
presence of Macrophage colony stimulating factor
(M-CSF) and ODF without osteoblasts. Recently,
TRAF6 has been shown to bind to the cytoplasmic
tail of RANK (also known as TRANCE receptor)
(Darnay et al. 1998; Galibert et al. 1998; Wong et al.
1998), a receptor for ODF. Thus, we tested whether
splenocytes from TRAF6¹/¹ mice could differentiate
Genes to Cells (1999) 4, 353–362 355
 A Naito et al.
to mature osteoclasts in the presence of ODF and
M-CSF without osteoblasts. Only a limited number of
mononuclear TRAPþ cells were observed in the
culture of splenocytes from TRAF6¹/¹ mice, while
significant number of multinuclear functional osteo-
clasts were formed from wild-type splenocytes
(Fig. 3B,D, top), indicating that either RANK signalling
or M-CSF signalling could be defective in the precursor
cells of TRAF6¹/¹ mice. When splenocytes were
cultured with M-CSF alone, almost equal number of
macrophage-like cells (identified by the staining for
nonspecific esterase (NSE)) were induced from spleno-
cytes from either TRAF6¹/¹ mice or their normal
littermates (Fig. 3D, bottom). Since osteoclasts and
356 Genes to Cells (1999) 4, 353–362
macrophages are derived from the same precursor cells
(Franzoso et al. 1997; Tondravi et al. 1997), these results
strongly suggest that both the formation of precursor
cells and the c-Fms signalling stimulated by M-CSF in
the precursor cells are not impaired in TRAF6¹/¹
mice. Therefore, the failure of functional osteoclast
formation is likely due to a defect in RANK signalling
in the precursor cells of TRAF6¹/¹ mice.
Reduction of the relative number of immature
B cells in bone marrow of TRAF6¹/¹ mice
We next analysed the development of haematopoietic
cells. Due to the limited bone marrow cavities, only
Figure 2 Severe osteopetrosis in
TRAF6¹/¹ mice. (A) Radio-opaque
long bone shafts and failure of incisor
teeth to erupt in TRAF6¹/¹ mice. Shown
are whole body anterior-posterior (left)
and skull lateral (right) radiographs of a
12-day-old TRAF6¹/¹ mouse (bottom)
and a wild-type littermate control (top).
Cross sectional views of the distal meta-
physis were obtained by microfocus X-ray
computed tomography (middle). (B) His-
tological changes in the bone. Longitudi-
nal sections cut through the tibia from a
9-day-old TRAF6¹/¹ and a wild-type
littermate mice are stained with Masson’s
trichrome stain (MT) to visualize trabe-
cular bone. Original magnifications, ∞40.
(C) Absence of TRAPþ osteoclasts (red
cells in þ/þ) in TRAF6¹/¹ mice. A
different pair of a 9-day-old TRAF6¹/¹
mouse and a wild-type littermate are used
for preparing the sections which are then
stained for TRAP. Original magnifica-
tions, ∞100. Epiphysis is in the upper
direction.
q Blackwell Science Limited
 TRAF6-deficient mice

Figure 3 Inability of osteoclast precursors derived from TRAF6¹/¹ mice to differentiate to functional osteoclasts. (A) In vitro
coculture of splenocytes with normal osteoblasts. Spleen cells from either wild-type or TRAF6¹/¹ mice were cocultured with normal
osteoblasts in the absence or presence of either 1,25-dihydroxyvitamin D3 (10 nM) or Prostaglandin E2 (1 mM). After 6 days culture, cells
were stained for TRAP activity, and TRAPþ red cells containing more than three nuclei were counted. (B) Induction of TRAPþ cells
from splenocytes without osteoblasts. Spleen cells from either wild-type or TRAF6¹/¹ mice were cultured in the absence or presence of
soluble osteoclast differentiation factor (sODF 30 ng/mL) and recombinant human macrophage-colony stimulation factor (rhM-CSF
10 ng/mL). After 6 days culture, cells were stained for TRAP activity, and TRAPþ red cells containing more than three nuclei were
counted. (C) Formation of resorption pits by osteoclasts induced in vitro. Co-cultures of splenocytes with normal osteoblasts in the
presence or absence of 1,25-dihydroxyvitamin D3 (10 nM) were performed on dentine slices. After 8 days culture, cells were removed
and the area of resorption pits formed was quantified. (D) Staining of TRAPþ cells and nonspecific esterase (NSE)þ cells induced from
splenocytes. Cell culture was performed as described in (B). After 6 days culture, cells were stained for TRAP activity (top). Splenocytes
were cultured in the presence of M-CSF (10 ng/mL) alone. After 6 days culture, cells were stained for non specific esterase activity
(NSE, bottom). NSEþ cells were stained black. Scale bar ¼ 100 mm.

q Blackwell Science Limited Genes to Cells (1999) 4, 353–362 357

 A Naito et al.
Figure 4 Impaired development of B lymphocytes in TRAF6¹/
¹
mice. Flow-cytometric analysis of lymphocytes from 12 to
14-day-old TRAF6¹/¹ mice and their normal littermates. Bone
marrow cells were stained for B220 and sIgM expression (top), or
B220 and CD43 expression on gated sIgM¹ (middle). Spleno-
cytes were stained for B220 and sIgM expression (bottom).
1–2 × 106 bone marrow cells were recovered from two
hindlimbs, which is about one tenth of those prepared
from wild-type mice. Flow cytometric analyses of bone
marrow cells from 12 to 14-day-old TRAF6¹/¹ mice
consistently revealed that significant reduction in the
relative amount of sIgMþB220þ immature B cells
(Fig. 4, top). The ratio of proB (CD43þB220þsIgM¹)
to preB (CD43¹B220þsIgM¹) cells was similar in bone
marrow of TRAF6¹/¹ mice and their normal litter-
mates (Fig. 4, middle). Although TRAF6¹/¹ mice are
splenomegaly, total cellularity of spleen of TRAF6¹/¹
358 Genes to Cells (1999) 4, 353–362
mice was not twice as large as that of control mice.
Significant reduction of sIgMþB220þ B cells was
observed in spleen of TRAF6¹/¹ mice (Fig. 4,
bottom), which is likely a reflection of the dramatic
decrease in the absolute number of immature B cells in
bone marrow. Relative number of Gr-1þ, Mac-1þ, or
TER119þ cells in spleen of TRAF6¹/¹ mice are not
dramatically different from those of control littermates
(data not shown). Thus, large population of spleen cells
of TRAF6¹/¹ mice are not identified by those surface
markers tested.
FACS profile of CD4þCD8þ double positive cells
and single positive cells in thymus of TRAF6¹/¹ mice
was similar to that of control mice (data not shown).
The development of triple negative thymocytes, that do
not express CD3, CD4, CD8, determined by the
expression profiles of CD25 and CD44 (Godfrey et al.
1993) was similar between TRAF6¹/¹ mice and their
control littermates (data not shown). These results
indicate that thymocyte development is normal in the
absence of TRAF6.
Defective IL-1 signalling in thymocytes of
TRAF6¹/¹ mice
Since TRAF6 is believed to mediate signals emanating
from IL-1R (Cao et al. 1996), we next examined IL-1-
mediated T cell proliferation (Adachi et al. 1998). In
wild-type mice, thymocytes displayed IL-1 dependent
enhancement of proliferation in the presence of
concanavalin A (ConA). However, thymocytes from
TRAF6¹/¹ mice did not show the IL-1-dependent
enhancement of proliferation (Fig. 5), leading to the
conclusion that TRAF6 is a downstream signal
transducer of IL-1R. We have previously demonstrated
that TRAF6 is a transducer of CD40 signals (Ishida et al.
1996a). However, the significant difference in the
relative number of CD40þ cells in spleen of TRAF6¹/¹
mice (2.9 6 0.8%) and that of control littermates
(47.7 6 9.4%) did not allow us to address directly the
role of TRAF6 in CD40 signalling.
Discussion
Analysis of ODF¹/¹ mice showed that the interaction
of RANK with ODF is essential for osteoclast
formation (Kong et al. 1999). We show here severe
osteopetrosis in TRAF6¹/¹ mice due to the inability of
osteoclast precursor cells to differentiate to functional
osteoclasts in response to ODF. RANK has been shown
to associate with TRAFs 1, 2, 3, 5 and 6 in in vitro
experiments (Darnay et al. 1998; Galibert et al. 1998;
q Blackwell Science Limited

Figure 5 Impaired IL-1-mediated T cell proliferation in
TRAF6¹/¹ mice. Thymocytes (2 × 105 cells/mL) from wild-
type and TRAF6¹/¹ mice were incubated with indicated
concentration of human IL-1a in the presence or absence of
ConA (1 mg/mL) for 72 h. [3H]-thymidine was pulsed for the last
12 h, and [3H] uptake was measured. Representative from three
independent experiments.
Wong et al. 1998). TRAF6 interacted with membrane-
proximal region of the RANK cytoplasmic tail that is
distinct from those binding other TRAFs. Further-
more, deletion of the TRAF6 binding site of RANK
almost completely prevented the RANK-dependent
activation of NFkB (Galibert et al. 1998). These results
suggest that NFkB activation by RANK signalling via
TRAF6 could be essential for the formation of
functional osteoclasts. The lack of functional osteoclast
formation in the p50/p52 double knockout mice
(Franzoso et al. 1997; Iotsova et al. 1997) is consistent
with this idea, since p50 or p52 is required to form
transcriptionally active heterodimers (Miyamoto &
Verma 1995). One possibility is that NFkB activation
leads to the expression of one or more antiapoptotic
genes, consistent with the report that ODF stimulation
enhances Bcl-xL expression (Anderson et al. 1997;
Wong et al. 1997). Since normal number of osteoclast
precursor cells are present in the spleen of TRAF6¹/¹
mice, ODF–RANK interaction could lead to the apop-
tosis of the precursor cells of TRAF6¹/¹ mice during the
course of differentiation. It has recently been shown that
signal-dependent interaction of TRAF6 with TAK1
results in the activation of both IkB kinase (IKK) and
JNK (Ninomiya et al. 1999). Thus, the role of JNK
activation in osteoclast formation can not be ruled out.
ODF¹/¹ mice is defective in lymph node organo-
genesis (Kong et al. 1999) like TRAF6¹/¹ mice shown
q Blackwell Science Limited
TRAF6-deficient mice
here, while IL-1R¹/¹ mice have lymph node (Labow
et al. 1997). These observations strongly suggest that the
RANK signalling triggered by ODF but not IL-1
signalling is required for some steps of lymph node
formation. Since ODF-and RANK-expressing cells are
present in lymph nodes (Kong et al. 1999), RANK
signal could support growth or protect apoptosis of cells
that constitute mature lymph nodes.
Reduction of immature B cells in bone marrow of
TRAF6¹/¹ mice may result from the frequent
apoptosis of immature B cells. Although CD40-
deficient mice do not exhibit the reduction of immature
B cells in bone marrow (Kawabe et al. 1994), possible
defect in CD40 signalling in TRAF6¹/¹ mice could be
involved in this apoptosis in combination with defect in
unknown receptor that also requires TRAF6 for
signalling. Alternatively, the observed defect in B cell
development could be due to the impaired bone
marrow environment as a consequence of osteopetrosis
as in the case of c-fos¹/¹ mice (Okada et al. 1994) and
op/op mice (Wiktor-Jedrzejckaz et al. 1982). In contrast
to the normal thymocyte development in TRAF6¹/¹
mice, the development of CD3/CD4/CD8 triple
negative CD44¹CD25þ precursors to CD44¹CD25¹
thymocytes was blocked in ODF¹/¹ mice (Kong et al.
1999). This difference suggests that RANK-mediated
JNK activation instead of NFkB activation could be
critical for the development of thymocytes, since the
mutant RANK that has deletion in TRAF6 binding site
does not activate NFkB but still activates JNK (Galibert
et al. 1998). The thymic atrophy observed in a large
number of TRAF6¹/¹ mice correlates with their health
status, suggesting that this may be an indirect effect of
the TRAF6 mutation.
ODF¹/¹ mice were shown to born at the expected
Mendelian frequency (Kong et al. 1999), while
TRAF6¹/¹ mice were only 12% of total live pups.
The role of TRAF6 in embryogenesis might indicate
the involvement of the Toll receptor family, which has
recently been shown to use TRAF6 as a signal
transducer (Muzio et al. 1998), in embryogenesis.
Runting is a common phenotype of TRAF-deficient
mice, which might be due to the progressive decrease of
blood glucose levels as shown in the case of TRAF3¹/¹
mice (Xu et al. 1996).
While this paper was under review, Lomago et al.
(1999) reported the osteopetrosis in TRAF6¹/¹ mice.
However, they found normal number of TRAPþ
osteoclasts which lacked contact with bone surfaces in
TRAF6¹/¹ mice. Although their targeted exon is
different from ours, this discrepancy remains to be
elucidated.
Genes to Cells (1999) 4, 353–362 359
 A Naito et al.
Although members of the TRAF family share
common structures and have shown to bind to over-
lapping subsets of receptor (Arch et al. 1998), gene
knockout mice uncovered the unique role of each
TRAF protein in vivo. Taken together, we conclude
that TRAF6 is an essential transducer of the signal
leading to the functional osteoclast formation, which
may emanate from RANK. We also conclude that
TRAF6 is required for the formation of lymph nodes
and IL-1 signalling.
Experimental procedures
Generation of TRAF6¹/¹ mice
Murine TRAF6 genomic DNA fragments were isolated from a
129/SVJ mouse genomic library by screening with a radio-
labelled DNA fragment corresponding to nucleotides 1–701 of
the murine TRAF6 complementary DNA (GENBANK accession
number D84655). A targeting vector was constructed to replace
the 2.5 kbp genomic fragment with the 3.0 kbp DNA fragment
containing pol II promoter and neomycin resistance gene in an
opposite direction. The replaced genomic fragment contained
the second coding exon corresponding to nucleotides 556–865
of the cDNA. Thus, the mRNA from mutant allele could encode
first 98 amino acids of TRAF6 which does not cover whole Ring
finger domain. The fragment containing neomycin resistance
gene was flanked by the 6.8 kbp 50 genomic fragment and 1.9 kbp
30 fragment. A diphtheria toxin A fragment gene driven by
MC-1 promoter was introduced at 30 end of the 1.9 kbp
fragment. The linearized targeting vector was electroporated into
E14.1 ES cells, which were selected in G418. Southern blot
analysis using genomic DNA prepared from ES-cell colonies
identified eight recombinant ES clones with a single targeted
allele. Four independently identified targeted ES clones were
injected into fertilized blastcysts from C57BL/6 female mice.
Chimeric mice were mated with C57BL/6 female mice for
germinal transmission. Heterozygous mice were intercrossed to
obtain homozygotes. To confirm the absence of TRAF6 protein,
cell lysates from whole brain were analysed. Whole brain was
disrupted in TNE buffer (10 mM Tris-HCl [pH 7.5], 150 mM
NaCl, 1 mM EDTA, 1% NP-40) and centrifuged to remove
debris. From the resulting supernatant, TRAF6 was immuno-
precipitated by adding rabbit anti-TRAF6 polyclonal antibody
(gift of Dr Randy Noelle) and protein G Sepharose (Amersham
Pharmacia Biotech). TRAF6 in the immunocomplex was
detected by Western blot using the same rabbit anti-TRAF6
polyclonal antibody and horse radish peroxidase (HRP)-
conjugated anti-rabbit IgG (Amersham Pharmacia Biotech).
Radiography and histology
Groups of TRAF6þ/þ, TRAF6þ/¹ and TRAF6¹/¹ mice were
necropsied on day 12 after birth. Radiography of whole bodies
was performed on a Softex CMB-2 (Softex Co., Kanagawa,
360 Genes to Cells (1999) 4, 353–362
Japan). Cross-sectional views of the distal metaphysis were
obtained by microfocus X-ray computed tomography (MCT)
manufactured for trial (ELE SCAN, Nittentsu Elex, Sakai,
Japan). The tomographicing was performed at energy; 0.25 KeV,
current; 0.1 mA, slice thickness; 8.86 mm. Total body, spleen and
thymus were weighed. Spleens and thymi were fixed in 10%(V/V)
buffered formalin, dehydrated, infiltrated with paraffin, and
sectioned at 6–8 mm for haematoxylin and eosin (H&E) staining.
Tibias were fixed in 70%(V/V) ethanol, dehydrated in an ethanol
series, embedded in paraffin and sectioned for Masson’s
trichrome staining (Takano-Yamamoto & Rodan 1990). For
analysing osteoclast, tibias were fixed in cold 70%(V/V) ethanol
instead of formalin and sections were stained for tartrate-resistant
acid phosphatase (TRAP) activity (Yasuda et al. 1998).
Induction and activation of osteoclast in cell
culture systems
To test the ability of splenocytes to differentiate into TRAPþ
osteoclasts, two different in vitro culture systems were used.
Spleen cells (1 × 105 cells/well) from either wild-type or
TRAF6¹/¹ mice were cocultured in the absence or the presence
of either Prostaglandin E2 (1 mM) or 1, 25-dihydroxyvitamin D3
(10 nM) with primary osteoblast (104 cells/well) isolated from
calvariae of 3-day-old mice. Alternatively, spleen cells (1 × 105
cells/well) were cultured in the presence of soluble form of ODF
(30 ng/mL) and M-CSF (10 ng/mL). After 6 days culture, cells
were stained for TRAP activity (Yasuda et al. 1998) and red cells
containg more than three nuclei were counted. To check the
functional property of TRAPþ osteoclasts, cocultures of spleen
cells with primary osteoclasts described above were performed on
dentine slides. After 8 days culture, cells were removed and the
area of resorption pits formed were quantified using an image
analysis system (SYSTEM SUPPLY, Nagano, Japan). For
macrophage induction, splenocytes were cultured in the presence
of M-CSF (10 ng/mL) alone. After 6 days culture, cells were
stained for non-specific esterase (NSE) activity according to the
manufacturer’s recommendation.
Flow cytometric analysis of haematopoietic cells
Single-cell suspensions of bone marrows, spleens and thymi were
stained with FITC-, phycoerythrin- or biotin-conjugated
antibodies (Pharmingen) reactive to Thy-1, CD3, CD4, CD8,
CD19, CD25, CD40, CD43, CD44, B220, sIgM, Mac-1, Gr-1,
IL-7R. Stained cells were analysed with a FACSCalibar (Becton
Dickinson) using the CELLQUEST software.
Lymphocyte proliferation assay
Thymocytes (2 × 105) were cultured in 96-well plates for
72 h with the indicated amount of human IL-1a (30, 100,
300 U/mL) in the presence or absence of 1 mg/mL ConA. [3H]-
thymidine was pulsed for the last 12 h, and [3H] uptake was
measured.
q Blackwell Science Limited

Acknowledgements
We thank Y. Yamanashi, Y. Yoshida, H. Hayashi, A. Aiba,
H. Toki, H. Yamato, R. Noelle and T. Kato for helpful discussion
and various materials, and R. Wisdom for critical reading of the
manuscript. This work was supported by grants from the
Ministry of Education, Science, Sports and Culture of Japan
and by a grant for AIDS Research from the Japan Health Science
Foundation.
References
Adachi, O., Kawai, T., Takeda, K., et al. (1998) Targeted
disruption of the MyD88 gene results in loss of IL-1- and
IL- 18-mediated function. Immunity 9, 143–150.
Anderson, D.M., Maraskovsky, E., Billingsley, W.L., et al. (1997)
A homologue of the TNF receptor and its ligand enhance
T-cell growth and dendritic-cell function. Nature 390,
175–179.
Arch., R.H., Gedrich, R.W. & Thompson, C.B. (1998) Tumor
necrosis factor receptor-associated factors (TRAFs)—a family
of adapter proteins that regulates life and death. Genes Dev. 12,
2821–2830.
Cao, Z., Xiong, J., Takeuchi, M., Kurama, T. & Goeddel, D.V.
(1996) TRAF6 is a signal transducer for interleukin-1. Nature
383, 443–446.
Cheng, G., Cleary, A.M., Ye, Z.S., Hong, D.I., Lederman, S. &
Baltimore, D. (1995) Involvement of CRAF1, a relative of
TRAF, in CD40 signaling. Science 267, 1494–1498.
Darnay, B.G., Haridas, V., Ni, J., Moore, P.A. & Aggarwal, B.B.
(1998) Characterization of the intracellular domain of receptor
activator of NF-kB (RANK). Interaction with tumor necrosis
factor receptor-associated factors and activation of NF-kB and
c-Jun N-terminal kinase. J. Biol. Chem. 273, 20551–20555.
Franzoso, G., Carlson, L., Xing, L., et al. (1997) Requirement for
NF-kB in osteoclast and B-cell development. Genes Dev. 11,
3482–3496.
Galibert, L., Tometsko, M.E., Anderson, D.M., Cosman, D. &
Dougall, W.C. (1998) The involvement of multiple tumor
necrosis factor receptor (TNFR)-associated factors in the
signaling mechanisms of receptor activator of NF-kB, a
member of the TNFR superfamily. J. Biol. Chem. 273,
34120–34127.
Godfrey, D.I., Kennedy, J., Suda, T. & Zlotnik, A. (1993) A
developmental pathway involving four phenotypically and
functionally distinct subsets of CD3-CD4-CD8-triple-nega-
tive adult mouse thymocytes defined by CD44 and CD25
expression. J. Immunol. 150, 4244–4252.
Grigoriadis, A.E., Wang, Z.Q., Cecchini, M.G., et al. (1994) c-
Fos: a key regulator of osteoclast-macrophage lineage
determination and bone remodeling. Science 266, 443–448.
Hu, H.M., O’Rourke, K., Boguski, M.S. & Dixit, V.M. (1994) A
novel RING finger protein interacts with the cytoplasmic
domain of CD40. J. Biol. Chem. 269, 30069–30072.
Iotsova, V., Caamano, J., Loy, J., Yang, Y., Lewin, A. & Bravo, R.
(1997) Osteopetrosis in mice lacking NF-kB1 and NF-kB2.
Nature Med. 3, 1285–1289.
Ishida, T., Mizushima, S., Azuma, S., et al. (1996a) Identification
of TRAF6, a novel tumor necrosis factor receptor-associated
factor protein that mediates signaling from an amino-terminal
q Blackwell Science Limited
TRAF6-deficient mice
domain of the CD40 cytoplasmic region. J. Biol. Chem. 271,
28745–28748.
Ishida, T.K., Tojo, T., Aoki, T., et al. (1996b) TRAF5, a novel
tumor necrosis factor receptor-associated factor family
protein, mediates CD40 signaling. Proc. Natl. Acad. Sci. USA
93, 9437–9442.
Kawabe, T., Naka, T., Yoshida, K., et al. (1994) The immune
responses in CD40-deficient mice: impaired immunoglobulin
class switching and germinal center formation. Immunity 1,
167–178.
Kong, Y.Y., Yoshida, H., Sarosi, I., et al. (1999) OPGL is a key
regulator of osteoclastogenesis, lymphocyte development and
lymph-node organogenesis. Nature 397, 315–323.
Labow, M., Shuster, D., Zetterstrom, M., et al. (1997) Absence of
IL-1 signaling and reduced inflammatory response in IL-1 type
I receptor-deficient mice. J. Immunol. 159, 2452–2461.
Lacey, D.L., Timms, E., Tan, H.L., et al. (1998) Osteoprotegerin
ligand is a cytokine that regulates osteoclast differentiation and
activation. Cell 93, 165–176.
Lomago, M.A., Yeh, W.-C., Sarosi, I., et al. (1999) TRAF6
deficiency results in osteopetrosis and defective interleukin-1,
CD40, and LPS signaling. Genes Dev. 13, 1015–1024.
Minkin, C. (1982) Bone acid phosphatase: tartrate-resistant acid
phosphatase as a marker of osteoclast function. Calcif. Tissue
Int. 34, 285–290.
Miyamoto, S. & Verma, I.M. (1995) Rel/NF-kB/IkB story. Adv.
Cancer Res. 66, 255–292.
Mosialos, G., Birkenbach, M., Yalamanchili, R., VanArsdale, T.,
Ware, C. & Kieff, E. (1995) The Epstein-Barr virus
transforming protein LMP1 engages signaling proteins for
the tumor necrosis factor receptor family. Cell 80, 389–399.
Muzio, M., Natoli, G., Saccani, S., Levrero, M. & Mantovani, A.
(1998) The human toll signaling pathway: divergence of
nuclear factor kB and JNK/SAPK activation upstream of
tumor necrosis factor receptor-associated factor 6 (TRAF6).
J. Exp. Med. 187, 2097–2101.
Nakano, H., Oshima, H., Chung, W., et al. (1996) TRAF5, an
activator of NF-kB and putative signal transducer for the
lymphotoxin-b receptor. J. Biol. Chem. 271, 14661–14664.
Ninomiya, T.J., Kishimoto, K., Hiyama, A., Inoue, J., Cao, Z. &
Matsumoto, K. (1999) The kinase TAK1 can activate the
NIK-IkB as well as the MAP kinase cascade in the IL-1
signaling pathway. Nature 398, 252–256.
Nishitoh, H., Saitoh, M., Mochida, Y., et al. (1998) ASK1 is
essential for JNK/SAPK activation by TRAF2. Mol. Cell 2,
389–395.
Okada, S., Wang, Z.-Q., Grigodiadis, A.E., Wagner, E.F. & von
Rüden, T. (1994) Mice lacking c-fos have normal hema-
topoietic stem cells but exhibit altered B-cell differentiation
due to an impaired bone marrow environment. Mol. Cell. Biol.
14, 382–390.
Popoff, S.N. & Marks, S.C. Jr (1995) The heterogeneity of the
osteopetroses reflects the diversity of cellular influences during
skeletal development. Bone 17, 437–445.
Regnier, C.H., Tomasetto, C., Moog-Lutz, C., et al. (1995)
Presence of a new conserved domain in CART1, a novel
member of the tumor necrosis factor receptor-associated
protein family, which is expressed in breast carcinoma. J. Biol.
Chem. 270, 25715–25721.
Rothe, M., Sarma, V., Dixit, V.M. & Goeddel, D.V. (1995)
TRAF2-mediated activation of NF-kB by TNF receptor 2
and CD40. Science 269, 1424–1427.
Genes to Cells (1999) 4, 353–362 361
 A Naito et al.
Rothe, M., Wong, S.C., Henzel, W.J. & Goeddel, D.V. (1994) A
novel family of putative signal transducers associated with the
cytoplasmic domain of the 75 kDa tumor necrosis factor
receptor. Cell 78, 681–692.
Sato, T., Irie, S. & Reed, J.C. (1995) A novel member of the
TRAF family of putative signal transducing proteins binds to
the cytosolic domain of CD40. FEBS Lett. 358, 113–118.
Soriano, P., Montgomery, C., Geske, R. & Bradley, A. (1991)
Targeted disruption of the c-src proto-oncogene leads to
osteopetrosis in mice. Cell 64, 693–702.
Takano-Yamamoto, T. & Rodan, G.A. (1990) Direct effects of
17b-estradiol on trabecular bone in ovariectomized rats. Proc.
Natl. Acad. Sci. USA 87, 2172–2176.
Tondravi, M.M., McKercher, S.R., Anderson, K., et al. (1997)
Osteopetrosis in mice lacking haematopoietic transcription
factor PU.1. Nature 386, 81–84.
Tsukamoto, N., Kobayashi, N., Azuma, S., Yamamoto, T. &
Inoue, J. (1999) Two differently regulated nuclear factor kB
activation pathways triggered by the cytoplasmic tail of CD40.
Proc. Natl. Acad. Sci. USA 96, 1234–1239.
Wiktor-Jedrzejckaz, W., Ahmed, A., Szczylik, C. & Skelly, R.R.
(1982) Hematological characterization of congenital osteope-
trosis in op/op mouse: Possible mechanism for abnormal
macrophage differentiation. J. Exp. Med. 156, 1516–1527.
Wong, B.R., Josien, R., Lee, S.Y., et al. (1997) TRANCE
(tumor necrosis factor [TNF]-related activation-induced
362 Genes to Cells (1999) 4, 353–362
cytokine), a new TNF family member predominantly
expressed in T cells, is a dendritic cell-specific survival
factor. J. Exp. Med. 186, 2075–2080.
Wong, B.R., Josien, R., Lee, S.Y., Vologodskaia, M., Steinman,
R.M. & Choi, Y. (1998) The TRAF family of signal
transducers mediates NF-kB activation by the TRANCE
receptor. J. Biol. Chem. 273, 28355–28359.
Xu, Y., Cheng, G. & Baltimore, D. (1996) Targeted disruption of
TRAF3 leads to postnatal lethality and defective T-dependent
immune responses. Immunity 5, 407–415.
Yasuda, H., Shima, N., Nakagawa, N., et al. (1998) Osteoclast
differentiation factor is a ligand for osteoprotegerin/osteoclas-
togenesis-inhibitory factor and is identical to TRANCE/
RANKL. Proc. Natl. Acad. Sci. USA 95, 3597–3602.
Yeh, W.C., Shahinian, A., Speiser, D., et al. (1997) Early lethality,
functional NF-kB activation, and increased sensitivity to
TNF-induced cell death in TRAF2-deficient mice. Immunity
7, 715–725.
Yoshida, H., Hayashi, S., Kunisada, T., et al. (1990) The murine
mutation osteopetrosis is in the coding region of the
macrophage colony stimulating factor gene. Nature 345,
442–444.
Received: 13 April 1999
Accepted: 15 May 1999
q Blackwell Science Limited