Previews
129
Pallanck, L., Ordway, R.W., Ramaswami, M., Chi, W.Y., Krishnan,
K.S., and Ganetzky, B. (1995). J. Biol. Chem. 270, 1874218744.
Sollner, T.H. (2003). Mol. Membr. Biol. 20, 209220.
Stamler, J.S., Lamas, S., and Fang, F.C. (2001). Cell 106, 675683.
Turner, K.M., Burgoyne, R.D., and Morgan, A. (1999). Trends Neu-
rosci. 22, 459464.
TSH, The Bone Suppressing
Hormone
The skeleton is a dynamic organ whose structural in-
tegrity depends on constant remodeling, controlled by
many local and systemic factors. In this issue of Cell,
Abe et al. (2003) identify thyroid-stimulating hormone
(TSH) as an important regulator of this process.
Thyroid disease and osteoporosis are prevalent in the
elderly, and von Recklinghausen recognized over 100
years ago that hyperthyroidism can cause osteoporosis.
Increased circulating thyroid hormone induces a high
turnover state in bone, with increased activity by osteo-
blasts and osteoclasts. The resorptive activity of osteo-
clasts, however, predominates, eventuating in decreased
structural integrity of the skeleton (Greenspan and
Greenspan, 1999). Thus far, investigators have focused
on the direct effects of active thyroid hormone (triiodo-
thyronine, T3) on bone cells, via thyroid hormone recep-
tors (TR1 and TR), members of the nuclear hormone
receptor family that induce transcription in a ligand-
dependent manner (Britto et al., 1994). Osteoblasts ex-
press TRs and respond to T3 with increased proliferation
and expression of lineage-specific markers such as al-
kaline phosphatase, osteocalcin, and collagen. Interest-
ingly, although osteoclasts have TRs, their response to
T3 appears to be mediated entirely by osteoblasts. In
the presence of vitamin D3, T3 induces osteoblasts to
express of RANK ligand (RANKL), the key osteoclasto-
genic cytokine. Additionally, mice lacking the known
active isoforms of TRs have retarded bone growth and
maturation, but do not manifest increased bone mineral
density, as would be predicted if, in fact, T3 is an impor-
tant stimulus of bone resorption in vivo (Gothe et al.,
1999).
A previously unexplored consequence of hyperthy-
roidism is suppression of thyroid stimulating hormone
(TSH), which is produced by the anterior pituitary and
directly controls production and release of thyroid hor-
mones by thyroid follicles. The TSH receptor (TSHR) is
a seven transmembrane glycosylated G protein-coupled
protein expressed by many tissues beyond the thyroid
(Davies et al., 2002). Although several extrathyroidal
functions for TSH have been proposed, evidence sup-
porting this contention is scant. In this issue of Cell, Abe
et al. (2003) demonstrate a critical role for TSH in skeletal
remodeling that is independent of its effects on circulat-
ing thyroid hormone. Using mice in which the TSHR
gene is replaced by GFP, they find bone mineral density,
in face of normal levels of thyroid hormone, depends
on an intact response to TSH. Thus, TSHR null mice
have severe osteoporosis. Heterozygous TSHR/ mice
are euthyroid, with normal levels of T3, T4, and TSH,
but nevertheless have a significant decrease in bone
density. Additionally, supplementation of TSHR/ mice
with thyroid hormone normalizes body weight, but not
bone mass. Interestingly, focal osteosclerosis accom-
panies global osteoporosis in TSHR/ mice. These ar-
eas of dense bone have histological features indicating
rapid formation. Thus, TSH appears to affect both osteo-
blasts and osteoclasts.
Because a role for TSH in bone remodeling was not
predicted, expression of TSHRs in bone cells had not
previously been examined. The present authors, how-
ever, convincingly demonstrate expression of TSHRs on
the surface of both osteoblast and osteoclast precur-
sors, in vivo and in vitro. They document suppressive
effects of TSH on osteoclast and osteoblast differentia-
tion, and enhanced maturation of cells lacking TSHR.
Therefore, unlike thyroid hormone, which acts directly
only on osteoblasts, TSH suppresses the function of
both osteoblasts and osteoclasts in a cell autonomous
fashion. In mice unable to respond to TSH due to ab-
sence of the receptor, the result is a high turnover state
in which bone resorption outpaces bone formation (Fig-
ure 1).
Osteoblast differentiation requires expression of the
transcription factors Runx-2 and osterix, but is also influ-
enced by the Wnt pathway (especially via the coreceptor
LRP-5) and activation of Flk-1, a VEGF receptor (Karse-
nty and Wagner, 2002). TSH suppresses expression of
LRP-5 and Flk-1, but does not affect Runx-2 or osterix.
Osteoblasts influence osteoclast differentiation in vivo
via their expression of RANKL and M-CSF (Teitelbaum
and Ross, 2003). The enhanced osteoclastogenesis of
TSHR/ mice cannot be attributed to enhanced expres-
sion of either M-CSF or RANKL, indicating that this is
not the point at which osteoclast and osteoblast activity
becomes uncoupled. Instead, the mutant animals have
increased levels of TNF, a cytokine that synergizes
with RANKL in the osteoclastogenic process.
RANKL activates a number of signaling pathways in
osteoclasts, including NF-B and the MAPKs (JNK, ERK,
and p38). Osteoclast precursors lacking one or both
copies of the TSHR gene exhibit increased signaling
through JNK and NF-B, but not ERK or p38. Likewise,
coadministration of TSH with RANKL in wild-type osteo-
clast cultures inhibits NF-B and JNK. Thus, TSH arrests
osteoclast differentiation through defined pathways
downstream of RANKL.
Perhaps the most intriguing issue posed by the dis-
covery of this new regulator of skeletal remodeling is
the mechanism by which TSH exerts its effects. The
TSHR is a member of the seven transmembrane G pro-
tein-coupled receptor family that also includes the calci-
tonin and parathyroid hormone (PTH) receptors, both
regulators of bone turnover. Although the calcitonin re-
ceptor is expressed on both osteoclasts and osteo-
blasts, its primary function is inhibition of bone resorp-
tion, decreasing osteoclast motility and secretion, by
osteoclasts, of acid and proteases (Samura et al., 2000).
The PTH receptor is highly expressed only on osteo-
blasts, which PTH impacts in a biphasic manner (Locklin
et al., 2003). Specifically, intermittent administration of
Cell
130
the hormone enhances bone formation, and continuous
exposure increases RANKL expression, thereby favoring
osteoclast differentiation and bone resorption. The TSH,
calcitonin, and PTH bone-active seven transmembrane
receptors each signal through cAMP, phospholipase C,
and protein kinase A. However, none has been pre-
viously shown to affect the signaling pathways found
to be suppressed by TSH in the current study. Determin-
ing the molecular basis for differences in signaling
downstream of these related receptors, and for differ-
ences in signaling by the same receptor in different
cell types is important because this receptor family has
already been targeted for therapeutic intervention in os-
teoporosis.
In recent years, genetic targeting of a number of pro-
teins, including c-src, leptin, and TSHR, have yielded
unexpected bone phenotypes. In retrospect, the discov-
ery of TSH as an important player in bone homeostasis
should perhaps not be so surprising, given its relation-
ship to other critical mediators of bone cell function.
Nevertheless, the connection would not likely have been
made without the use of gene-ablated mice. The contin-
ued generation of mutant mice, and the examination of
their bones, will provide many additional regulators of
bone homeostasis in the years to come.
Deborah Veis Novack
Department of Medicine
Division of Bone and Mineral Diseases
Washington University School of Medicine
Saint Louis, Missouri 63110
Selected Reading
Abe, E., Marians, R.C., Wu, X.-B., Iqbal, J., Ando, T., Yanan, L., Blair,
H.C., Davies, T.F., and Zaidi, M. (2003). Cell 115, this issue, 151162.
Britto, J.M., Fenton, A.J., Holloway, W.R., and Nicholson, G.C.
(1994). Endocrinology 134, 169176.
Davies, T.F., Marians, R., and Latif, R. (2002). J. Clin. Invest. 110,
161164.
Gothe, S., Wang, Z., Ng, L., Kindblomm, J.M., Barrosm, A.C., Ohls-
Figure 1. Bone Turnover Is Enhanced in the
Absence of TSHR
In normal cells, TSH inhibits the differentia-
tion of both osteoblast and osteoclast precur-
sors. In TSHR/ mice, this inhibition is re-
moved. Osteoblast precursors upregulate
LRP-5 and FLK-1, leading to increased num-
bers of differentiated osteoblasts (OB) and
increased new bone formation. Production of
TNF is also increased, and this cytokine
enhances the differentiation of osteoclast
precursors. In the absence of the TSHR,
osteoclast precursors show enhanced RANKL-
mediated differentiation, with increased phos-
phorylation of JNK (JNK-p) and activation of
NF-B. The net result is a marked increase
in mature osteoclasts (OC) with concomitant
bone resorption that outpaces new bone
formation. Thus, TSHR/ mice are osteopo-
rotic.
sonm, C., Vennstrom, B., and Forrest, D. (1999). Genes Dev. 13,
13291341.
Greenspan, S.L., and Greenspan, F.S. (1999). Ann. Intern. Med.
130, 750758.
Karsenty, G., and Wagner, E.F. (2002). Dev. Cell 2, 389406.
Locklin, R.M., Khosla, S., Turner, R.T., and Riggs, B.L. (2003). J.
Cell. Biochem. 89, 180190.
Samura, A., Wada, S., Suda, S., Iitaka, M., and Katayama, S. (2000).
Endocrinology 141, 37743782.
Teitelbaum, S.L., and Ross, F.P. (2003). Nat. Rev. 4, 638649.
The Double Life of Ribosomal
Proteins
Many integral proteins of the ribosome also carry out
extra-ribosomal functions as independent polypep-
tides, raising questions as to their evolutionary deriva-
tion. In this issue of Cell, Mazumder et al. report a
surprising new twist in the dual life of these molecules:
as part of a cellular response to interferon, a large-
subunit protein dramatically exits the ribosome to bind
and inhibit the translation of a specific mRNA.
Although it is widely held that the earliest protein-syn-
thesizing machines were composed exclusively of RNA,
proteins presently make up one-third to one-half of the
total ribosome mass and efficient translation cannot
take place without them. In bacteria, however, the ab-
sence of individual r-proteins does not necessarily deal
the translation apparatus a fatal blow; in many cases,
protein-deficient ribosomes still function, albeit with var-
ious degrees of impairment. The origins and evolution-
ary history of these polypeptides remain an enigma.
Numerous ribosomal proteins are bifunctional, that is,
they are not only integral components of the ribosome
but carry out other tasks in the cell often unrelated to