TAT fusion proteins containing tyrosine 42-deleted IkappaBalpha arrest osteoclastogenesis.

In most circumstances, NF-kappaB, which is essential for osteoclastogenesis, is activated following serine 32/36 phosphorylation of its cytosolic inhibitory protein, IkappaBalpha. In contrast to other cell types, IkappaBalpha, in bone marrow macrophages (BMMs), which are osteoclast precursors, is tyrosine-phosphorylated by c-Src kinase. To address the role of IkappaBalpha phosphorylation in osteoclastogenesis, we generated TAT fusion proteins containing wild-type IkappaBalpha (TAT-WT-IkappaB), IkappaBalpha lacking its NH(2)-terminal 45 amino acids (TAT-IkappaB(46-317)), and IkappaBalpha in which tyrosine residue 42, the c-Src target, is mutated into phenylalanine (TAT-IkappaB(Y42F)). TAT-IkappaB efficiently enters BMMs, and the NF-kappaB-inhibitory protein, once intracellular, is functional. While TAT-WT-IkappaB only slightly inhibits osteoclastogenesis, osteoclast recruitment is diminished >80% by TAT-IkappaB(46-317), an event mirrored by dentin resorption. The fact that TAT alone does not impact osteoclastogenesis, which also resumes following withdrawal of TAT-IkappaB(46-317), establishes that the mutant's anti-osteoclastogenic properties do not reflect toxicity. Affirming a functional role for IkappaB(Tyr(42)) in osteoclastogenesis, TAT-IkappaB(Y42F) is as efficient as TAT-IkappaB(46-317) in blocking osteoclast differentiation. Thus, dominant-negative IkappaBalpha constructs block osteoclastogenesis, and Tyr(42) is essential to the process, increasing the possibility that nonphosphorylatable forms of IkappaBalpha may be a means of preventing pathological bone loss.

and p52 subunits of NF-B leads to osteoporosis, due to failed osteoclastogenesis, establishes that activation of this transcription complex is essential for commitment of BMMs to the osteoclast phenotype (5).
The NF-B family is composed of five members that can homo-or heterodimerize following activation (6). The transcription factor resides in its inactive form in the cytoplasm avidly bound to the inhibitory protein IB␣. In a variety of cell types, extracellular signals lead to phosphorylation and degradation of IB␣, thereby releasing NF-B. The transcription factor then translocates to the nucleus, binds to DNA sequences, and activates target genes (7). While in most circumstances IB␣ is serine-phosphorylated, it can undergo phosphorylation on tyrosine 42 (8 -10). This event, in TNF-induced osteoclast precursors, also leads to NF-B release and nuclear translocation (10).
While NF-B is clearly essential for osteoclast formation, the means by which it stimulates the process is unclear. The goals of the present exercise were, therefore, 2-fold. In the first instance, we wished to determine whether tyrosine phosphorylation of IB␣ is essential for BMM differentiation into osteoclasts. Second, we asked if osteoclastogenesis, and attendant bone resorption, can be inhibited by dominant-negative IB␣, specifically that in which Tyr 42 is unavailable for phosphorylation.
The planned exercise required expression of putative dominant-negative forms of IB␣ in osteoclast precursors. This strategy was hampered, however, by the fact that osteoclasts, and their monocyte/macrophage progenitors, are presently impossible to efficiently transfect by traditional methods. To overcome this obstacle we took advantage of TAT, a peptide derived from the human immunodeficiency virus, type-1 transduction domain (11,12). Proteins fused to this 11-amino acid sequence, when misfolded, enter many cell types, refold, and regain function (13)(14)(15)(16)(17). Thus, our first undertaking was to determine whether a TAT-based approach would permit us to efficiently express IB␣ constructs in bona fide osteoclast precursors.
Our data show that TAT-fused IB␣ readily enters osteoclast precursors in a dose-dependent manner and elevates NF-B levels in the cytosol. More importantly, IB␣ lacking all phosphorylation sites, inhibits osteoclastogenesis in a nontoxic reversible manner and blocks bone resorption.

MATERIALS AND METHODS
Reagents-Polyclonal anti-IB␣ and anti-NF-B antibodies were purchased from Santa Cruz (Santa Cruz, CA). 1,25-(OH) 2 D 3 was provided by Dr. Milan Uskokovic (Hoffman La-Roche, Nutley, NJ). The ECL kit was obtained from Pierce. Wild-type and tyrosine 42-mutated IB␣ cDNAs were a gift of Dr. Jean-Francois Peyron (Nice, France). IB 46 -317 super-repressor mutant was generated by NH 2 -terminal deletion of residues 1-45 using a standard polymerase chain reaction approach. All other chemicals were obtained from Sigma. * This work was supported in part by an Arthritis Investigator Award (to Y. A.-A.); a Monsanto/Searle grant (to Y. A.-A.); and by National Institutes of Health Grants DE13754 (to Y. A.-A.), AR47096 (to J. C. C.), and AR32788, AR45623, and DE05413 (to S. L. T.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture-BMMs were isolated as described previously (4). Briefly, whole bone marrow of 4 -6 week mice was collected and incubated in tissue culture plates, at 37°C in 5% CO 2 , in the presence of 10 ng/ml macrophage-colony-stimulating factor (4). After 24 h in culture, the nonadherent cells were collected and layered on a Ficoll-Hypaque gradient. Cells at the gradient interface were collected and plated in ␣-minimal essential medium, supplemented with 10% heat-inactivated fetal bovine serum, at 37°C in 5% CO 2 in the presence of 10 ng/ml macrophage-colony-stimulating factor, and plated according to each experimental condition.
Osteoclast Generation-Whole marrow cultures were plated in 48 multi-well plates at 2 ϫ 10 6 cell/ml/well in the presence of 10 nM 1,25-(OH) 2 D 3 . Cultures were supplemented with 10 nM 1,25-(OH) 2 D 3 and fresh media on day 4 of culture. Osteoclasts develop on days 7-8 of culture at a point where cells are fixed and TRAP-stained Immunostaining-BMMs were plated on multi-well coverslips in the absence or presence (1 h) of TAT-IB. Cells were then fixed and stained with anti-HA antibody and detected with fluorescent secondary antibody.
Immunoblotting-Total cell lysates were boiled in the presence of 2ϫ SDS-sample buffer (0.5 M Tris-HCl (pH 6.8), 10% (w/v) SDS, 10% glycerol, 0.05% (w/v) bromphenol blue, distilled water) for 5 min and subjected to electrophoresis on 8 -12% SDS-polyacrylamide gel electrophoresis (18). Proteins were transferred to nitrocellulose membranes using a semi-dry blotter (Bio-Rad) and incubated in blocking solution (10% skim milk prepared in PBS containing 0.05% Tween 20), to reduce nonspecific binding. Membranes were washed with PBS/Tween buffer and exposed to primary antibodies (1 h at room temperature), washed again four times, and incubated with the respective secondary horseradish peroxidase-conjugated antibodies (1 h at room temperature). Membranes were washed extensively (5 ϫ 15 min), and an ECL detection assay was performed following the manufacturer's directions.
pTAT Construct and Protein Coupling-Various IB␣ constructs were cloned into the pTAT-HA bacterial expression vector described previously by Nagahara et al. (13), which contains a six-histidine tag, for easy purification, an HA tag for detection followed by the TAT transduction domain, and finally the IB sequence. The resultant plasmid, pTAT-IB, was transformed into the DH5␣ strain of Escherichia coli. The transformants were screened initially by restriction enzyme mapping. A recombinant containing the correct restriction fragments was then sequenced on both strands. This plasmid was then used to express the TAT-coupled IB␣ in the BL21 (DE3) strain of E. coli. Following 4 -6 h of induction, the cells were sonicated in 8 M urea, and the TAT-coupled IB␣ was purified on a nickel-Sepharose column (Qiagen), then applied to an ionic exchange column (Mono Q) in 4 M urea. To shock misfold the protein, the ionic exchange column was switched in one step from 4 M urea to aqueous buffer (20 mM HEPES). IB␣ was eluted by stepping from 50 mM to 1 M NaCl followed by desalting on a PD-10 column (Amersham Pharmacia Biotech) into PBS or 20 mM HEPES (pH 7.2), 137 mM NaCl and frozen in 10% glycerol at Ϫ80°C. pTAT-coupled and misfolded proteins remain highly concentrated, and resistant to freeze-thaw denaturation, and readily enter the cells upon incubation. Once in hand, the coupled TAT proteins were added to cultured marrow cells without the aid of transfection agents.
Pit Assay-Osteoclast precursors were plated on whale dentin slices, in 24-well plates, and allowed to differentiate, under osteoclastic conditions, in the absence or presence of 100 nM TAT or TAT-IB 46 -317 . After 8 days slices were placed in 0.25 M NaOH and sonicated briefly to remove cells. Dentin slices were then rinsed in PBS and subjected to scanning electron microscopy to identify resorption pits.

TAT-IB Enters Osteoclast Progenitors and Retains NF-
B-To test TAT-IB transduction, we treated BMMs, for 1 h, with a TAT-IB fusion protein containing HA tag (HA-TAT-IB). As demonstrated by immunostaining, using anti-HA monoclonal antibody, HA-TAT-IB transduces the vast majority of BMMs in culture (Fig. 1). To determine whether the HA-TAT-IB enters BMMs in a concentration-dependent manner, we incubated osteoclast precursors with increasing amounts of the purified fusion protein. Lysates of 1-h-treated cells were immunoblotted with anti-HA monoclonal antibody. As shown in Fig. 2A, HA-TAT-IB enters BMMs, dose-dependently.
Because IB␣ normally binds to, and retains, NF-B in the cytoplasm, we asked if TAT-IB affects cellular localization of NF-B. Attesting to the TAT fusion protein's functionality, the transduced inhibitory protein, dose-dependently, retains NF-B in the cytosol as assessed by measuring cytosolic levels of p65 NF-B subunit from cell lysates (Fig. 2B).
TAT-IB Proteins Block Osteoclastogenesis-Given that NF-B deletion blunts osteoclastogenesis, we asked whether inhibition of NF-B activation would also block osteoclast development. We reasoned that introduction of IB␣ protein, the NF-B inhibitory protein, devoid of its functional phosphorylation sites, into osteoclast precursors might block osteoclast recruitment. To address the role of IB␣ phosphorylation in osteoclastogenesis, we generated TAT fusion proteins containing wild-type IB␣, (TAT-WT-IB) and IB␣ lacking its NH 2 -terminal 45 amino acids (TAT-IB 46 -317 ). Marrow, placed in osteoclastogenic conditions (4), was treated, on day 3 of culture, with the various TAT-IB constructs or with TAT alone. In some cultures the inhibitory proteins were withdrawn following 48 h of exposure. Tartrate-resistant acid phosphatase (TRAP) expression by these cultures indicates that TAT alone fails to impact osteoclast recruitment, indicating that the human immunodeficiency virus-derived peptide is nontoxic in these circumstances (Fig.  3A, row D). Osteoclast recruitment is, however, diminished Ͼ80% by TAT-IB 46 -317 (Fig. 3A, wells C1-C2, and B), suggesting that the functional phosphorylation sites are central to the process. The resumed osteoclast differentiation following withdrawal of the inhibitory construct establishes that the TAT fusion protein's anti-osteoclastogenic properties are reversible (Fig. 3A, rows 3-4). In contrast to TAT-IB 46 -317 , TAT-WT-IB only slightly inhibits osteoclastogenesis (Fig.  3A, wells B1-B2). As expected, inhibition of osteoclast formation by TAT-IB 46 -317 is mirrored by arrested dentin resorption (Fig. 4). The number of resorption lacunae counted in 1 cm 2 was 187 Ϯ 39 in PBS ϩ TAT compared with 20 Ϯ 14 in TAT-IB-treated conditions. Furthermore, lacunae in TATtreated conditions were well defined, wide and deep tracks compared with shallow focal pits in TAT-IB-treated conditions.
Tyrosine 42-mutated IB Inhibits NF-B Activation and Osteoclastogenesis-Having established that deletion of the functional phosphorylation sites to IB␣ blunts osteoclastogenesis, we asked if mutation of tyrosine 42, which is specifically phosphorylated in osteoclast precursors, mirrors this event. To this end we generated TAT-IB(Y42F) and examined its cellular transduction activity. First, HA-TAT-IB(Y42F) was added to osteoclast precursors, and the presence of the protein was examined with time by immunoblots. We find that the protein is efficiently transduced in osteoclast precursors within 1 h, and its intracellular levels are further elevated in the following 8 h (Fig. 5). Residual levels of TAT-IB(Y42F) remain detectable up to 2 days in culture. Furthermore, we find that dose-dependent transduction of this mutant protein (Fig. 6A) retains NF-B in the cytosol (Fig. 6B) in a manner similar to effects exerted by other IB mutants (Fig. 2). Second, we asked if TAT-IB(Y42F) inhibits TNF-induced NF-B nuclear translocation. To this end, cells were left untreated or treated with TNF (10 ng/ml) in the absence or presence of TAT-IB(Y42F) for different time points. The data depicted in Fig. 7A indicate that, while as expected, TNF reduces cytosolic levels of NF-B (lane 2), tyrosine-mutated TAT-IB prevents this reduction within 1 h, and levels of NF-B in the cytosol are sustained up to 16 h. Reduced levels of NF-B in days 2 and 3 (lanes 6 and 7) most likely reflect decay of the TNF signal and/or degradation of the TAT protein. The observations in Fig. 7A are further supported by our findings that TNF enhances nuclear levels of NF-B (Fig. 7B, lane 2) and that the various IB␣ proteins block it (lanes 3-5). Finally, we examined whether, similar to TAT-IB 46 -317 , addition of TAT-IB(Y42F) to osteoclastogenic cultures blocks osteoclastogenesis. Establishing a functional role for IB(Y42F) in osteoclastogenesis, we find that TAT-IB(Y42F) is as efficient as TAT-IB 46 -317 in blocking osteoclast differentiation (Figs. 8 and 9). DISCUSSION Osteoclast development and activation are requisite for normal bone resorption (1,19). On the other hand, in inflammatory conditions such as periodontitis, rheumatoid arthritis, and peri-prosthetic implant loosening, excessive osteoclast recruitment leads to pathological osteolysis (1,2). In fact, accelerated osteoclastogenesis is the cellular defect responsible for the most common of the metabolic bone disorders, namely post-menopausal osteoporosis. Thus, understanding the means by which monocyte/macrophage precursors differentiate into osteoclasts provides potential avenues for preventing these diseases.
Experimental and spontaneous osteoporosis has yielded seminal observations regarding the molecular regulation of osteoclast differentiation and function. Among the most important information derived from these animals is the pivotal role that NF-B plays in the osteoclastogenic process. Not only does the p50/p52 NF-BϪ/Ϫ mouse lack osteoclasts, but the phenotype is cured by marrow transplantation (5,20), establishing that the transcription complex exerts its osteoclastogenic effect on the osteoclast precursor, per se, and not an accessory cell such as those of stromal lineage. Thus, understanding regulatory events of NF-B activation in osteoclast progenitors may provide insights to manage pathologic osteoclastic activity.
In most cells, phosphorylation of IB␣ at Ser 32 and Ser 36 prompts its dissociation from NF-B, marking the inhibitory protein for ubiquitination and subsequent degradation. NF-B, in turn, translocates to the nucleus to function as a transcriptional complex. On the other hand, NF-B activation, in a restricted population, is attendant upon tyrosine IB␣ phosphorylation, a process mediated by Src family kinases. For example, deletion of c-Src in osteoclast precursors or p56 lck in T-lymphocytes inhibits NF-B activation and interleukin-6 secretion (8,10). Moreover, we and others (9,10,21) find that IB(Tyr 42 ) is a functional phosphorylation site. In this regard, our previous studies unveiled that the tyrosine kinase c-Src, also essential for osteoclastogenesis, mediates tyrosine phosphorylation of IB␣ on tyrosine residue 42, in osteoclast progenitors. This phosphorylation is followed by activation of the NF-B transcriptional complex.
Extending these observations to a possible therapeutic approach, we reasoned that nonphosphorylatable IB␣ might serve as a dominant-negative inhibitor of osteoclastogenesis. Specifically, we speculated that mutating IB␣ amino-terminal phosphorylation sites might preserve the IB⅐NF-B complex in osteoclast precursors. However, significant effect requires introduction of the mutated proteins in the vast majority of primary osteoclast progenitors, an unattainable goal by traditional techniques. To resolve this contingent, we utilized the TAT-mediated protein transduction method, developed and successfully used by our group to introduce a large number of proteins into a wide spectrum of cells, including primary osteoclast precursors (17,22).
Our data show that dominant-negative forms of IB␣, delivered into osteoclast precursors, as TAT fusion proteins, prevent the cells' commitment to the osteoclast phenotype. Importantly, delivery of the mutated TAT-IB proteins is a rapid event, which is sustained for at least 24 h. Moreover, transduced IB␣ mutant proteins in osteoclast precursors, dose-dependently, retain NF-B in the cytoplasm under basal and TNF-stimulating conditions. Thus, TAT-IB mutants appear to act as dominant-negative proteins by displacing endogenous IB␣; the later remains susceptible for phosphorylation and processing.
The biological significance of TAT-IB delivery into osteoclast precursors is manifested by its potent anti-osteoclastogenic effect. In this regard, we document that while TAT-WT-IB, is only slightly inhibitory, serine/serine/tyrosine-deleted and tyrosine-mutated IB␣ proteins are highly effective in blocking osteoclast development. These effects correlate well with their observed effect on NF-B inactivation. While yet to be proven, our observations suggest that the raised levels of WT-IB, delivered as a TAT construct, are subject to phosphorylation and removal and, thus, are insufficient to block prolonged activation of NF-B in osteoclast precursors, and hence, osteoclastogenesis continues. Additionally, the lack of impact of TAT peptide or TAT-WT-IB on osteoclast recruitment together with reversibility of the effect underscores the nontoxic nature of these fusion proteins.
The capacity of these constructs to effectively transduce osteoclasts reinforces the potential of TAT fusion proteins as therapeutic moieties. As regard to the present constructs, preventing phosphorylation of IB␣ at its NH 2 -terminal domain, in general, and tyrosine phosphorylation on residue 42, in particular, substantially diminishes osteoclastogenesis. Taken with the restricted cell specificity of IB␣ tyrosine phosphorylation, these observations raise the possibility that nonphosphorylatable TAT-IB may be a means of averting pathological bone loss.
FIG. 9. Nonphosphorylatable IB proteins block osteoclastogenesis. Marrow was placed in osteoclastogenic conditions. On day 5, TAT peptide, alone, or TAT-WT-IB, TAT-IB 46 -317 , or TAT-IB(Y42F) (100 nM) were added for an additional 3 days. Cultures were stained for TRAP activity and the number of osteoclasts/well counted. Data are presented as the average number of multi-nucleated (3ϩ nuclei), TRAPpositive cells in triplicate wells from two independent experiments. * represents p Ͻ 0.001 relative to TAT.