Insertion Mutation in Tnfrsf11a Causes a Paget's Disease–Like Phenotype in Heterozygous Mice and Osteopetrosis in Homozygous Mice

Early onset familial Paget's disease of bone (EoPDB), familial expansile osteolysis, and expansile skeletal hyperphosphatasia are related disorders caused by insertion mutations in exon 1 of the TNFRSF11A gene, which encodes receptor activator of nuclear factor κB (RANK) protein. To understand the mechanisms underlying these disorders, we developed a mouse model carrying the 75dup27 mutation which causes EoPDB. Mice heterozygous for the mutation (Tnfrsf11a75dup27/−) developed a PDB‐like disorder with focal osteolytic lesions in the hind limbs with increasing age. Treatment of these mice with zoledronic acid completely prevented the development of lesions. Studies in vitro showed that RANK ligand (RANKL)‐induced osteoclast formation and signaling was impaired in bone marrow cells from Tnfrsf11a75dup27/− animals, but that osteoclast survival was increased independent of RANKL stimulation. Surprisingly, Tnfrsf11a75dup27/75dup27 homozygotes had osteopetrosis at birth, with complete absence of osteoclasts. Bone marrow cells from these mice failed to form osteoclasts in response to RANKL and macrophage colony‐stimulating factor (M‐CSF) stimulation. This intriguing study has shown that in heterozygous form, the 75dup27 mutation causes focal osteolytic lesions in vivo reminiscent of the human disorder and extends osteoclast survival independently of RANKL signaling. In homozygous form, however, the mutation causes osteopetrosis due to failure of osteoclast formation and insensitivity to RANKL stimulation. © 2021 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR)..


Introduction
E arly onset Paget's disease of bone (EoPDB) is a rare condition that has phenotypic overlap with both classical PDB and familial expansile osteolysis. (1)(2)(3) It is caused by a heterozygous 27-basepair (bp) duplication in the signal peptide of TNFRSF11A gene (75dup27), leading to insertion of nine amino acids in the receptor activator of nuclear factor κB (NF-κB) (RANK) signal peptide. It belongs to a group of related focal osteolytic disorders with distinct but overlapping features in which various insertion mutations occur in exon 1 of TNFRSF11A causing duplications of between five and nine amino acids in the RANK signal peptide (reviewed in Ralston and Taylor (4) ). Previous in vitro studies have shown that the causal mutations prevent cleavage of the RANK signal peptide, causing the mutant RANK molecules to accumulate in the Golgi apparatus, impairing its ability to translocate to the cell surface. (3,5,6) This, in turn, has been shown to lead to defective RANK ligand (RANKL) signaling, indicating that these are loss-of-function mutations that impair RANKL-induced osteoclastogenesis. (5,7) There remains a conundrum of exactly how loss of function mutations in TNFRSF11A can cause an osteolytic disorder in vivo. In order to investigate this, we generated a mouse model of the 75dup27 insertion, confirming that this is a loss-of function mutation for RANKL-induced osteoclastogenesis in homozygous form, but is a gain of function mutation in heterozygous due to nonclassical activation of NF-κB signaling.

Materials and Methods
Generation of the Tnfrsf11a 75dup27/75dup27 mouse model Mice with the 75dup27 mutation were generated by gene targeting in embryonic stem cells according to classical techniques. The targeting construct was created by generating a 4.7-kilobase (kb) fragment by polymerase chain reaction (PCR) from genomic DNA encompassing exon 1 of the mouse Tnfrsf11a gene and surrounding sequences. The 27-bp insertion mutation was inserted in the first exon ( Supplementary Fig. S1) by site-directed mutagenesis (Stratagene California, La Jolla, CA, USA), followed by cloning a floxed neomycin cassette as a positive selection marker, a 2-kb 3 0 homology region, and diphtheria toxin A (DTA) cassette as a negative selection marker. Supplementary  Fig. S2A shows schematic representation of the targeting construct. The targeting construct was electroporated into 129/Ola embryonic stem (ES) cells. Following selection by culturing the ES cells in neomycin, three positive clones were identified, and fidelity of the targeting event was confirmed by southern blotting with probes at the 5 0 and 3 0 ends ( Supplementary  Fig. S2B). Targeted ES cells were injected into C57BL/6 blastocysts to generate chimeric animals. The neomycin cassette was excised by crossing male chimeras with a cytomegaloviruscyclic recombinase (CMV-Cre) expressing C57BL/6 female. Presence of the 27-bp insertion mutation was confirmed by DNA sequencing (Supplementary Fig. S2C). The resulting C57BL/6-129Ola mixed-background animals were backcrossed into C57BL/6 for 10 generations and complete backcross was confirmed using Illumina mouse medium-density linkage panel following the manufacturer's recommended protocol (Illumina Inc, San Diego, CA, USA). All mice were housed in a standard animal facility with free access to food (pelleted RM1; SDS Diets, Essex, UK) and water. All experiments on mice were performed according to institutional and UK regulations.

Osteoclast cultures
Bone marrow was flushed from long bones of 3-month-old to 4-month-old mice of each genotype and cultured in α-minimum essential medium (α-MEM; Sigma, Dorset, UK) supplemented with 10% fetal calf serum (HyClone Laboratories, Logan, UT, USA), penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA), glutamine (Sigma, Dorset, UK), and macrophage colony-stimulating factor (MCSF; Prospec Bio, East Brunswick, NJ, USA) (100 ng/ml) for 48 h. The culture medium containing nonadherent cells was removed and adherent cells, containing MCSF-dependent osteoclast precursors were removed by trypsinization and plated at a density of 1 × 10 4 cells/well in 96-well plates with 150 μl of culture medium and MCSF at a concentration of 25 ng/ml and different concentrations of RANKL (R&D systems, Abingdon, UK) until osteoclasts were formed (4 to 5 days). At the end of the culture period the medium was removed and the cells were fixed with 4% formaldehyde in phosphatebuffered saline (PBS) and stained for tartrate-resistant acid phosphatase (TRAcP). (8) Osteoclasts, defined as TRAcP+ cells with three or more nuclei, were counted using a 10× objective lens on a bright field microscope. Osteoclasts carrying more than five nuclei were counted separately.

Osteoclast survival
Osteoclasts were generated from bone marrow cells as described in osteoclast cultures above. After 4 to 5 days in culture, RANKL was withdrawn and cultures continued for periods of up to 42 h. The cultures were stopped at the indicated time points and cells were fixed in 4% formaldehyde in PBS. Staining for osteoclasts was performed as described in osteoclast cultures above.

Western blot
The cells were washed in 1 ml of ice-cold PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer. Protein quantification was performed using the Pierce BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA). Protein lysates were mixed with the appropriate volume of 5× sample loading protein buffer (10% sodium dodecylsulfate [SDS], 500mM dithiothreitol [DTT], 50% glycerol, 250mM Tris-hydrochloride [Tris-HCL], and 0.5% bromophenol blue dye, pH 6.8) and denatured at 95 C on a dry bath for 3 to 5 min, then 60 to 70 μg was loaded into precast acrylamide gels (Bio-Rad Laboratories, Hercules, CA, USA). Band separation was performed by electrophoresis at 200 V for 35 to 45 min and proteins were transferred to a nylon membrane at 60 mA for 2 to 3 h. Cellular oncogene Fos (c-Fos), extracellular signal-regulated kinase (ERK) 1/2 (Thr202/Tyr204 p44/p42 mitogen-activated protein kinase [MAPK]), Ser32 inhibitor of NF-κB [IκB]-α, Thr180/Tyr182 p38 MAPK, as well as their phosphorylated counterparts were detected using rabbit polyclonal antibodies (Supplementary Table S1). β-Actin or hypoxanthine phosphoribosyltransferase (HPRT) was used for normalization and detected using rabbit polyclonal antibody (Supplementary  Table S1). Horseradish peroxidase-conjugated anti-rabbit secondary antibodies were used to visualize results using a Supersignal chemiluminescence kit (Thermo Fisher Scientific, Waltham, MA, USA) in a Syngene Genegnome bioimaging system (Thermo Fisher Scientific, Waltham, MA, USA). Intensities of bands were quantified using Image Studio Lite Ver 3.1 software (LI-COR Biosciences, Lincoln, NE, USA). NF-κB luciferase reporter assay MCSF-dependent macrophages were generated from mice as described above in osteoclast cultures and seeded at a density of 1 × 10 6 cells and cultured in complete α-MEM medium for 24 h. Subsequently, the cells were transduced with a lentiviral NF-κB luciferase reporter (NF-κB Cignal lenti luciferase reporter assay; SABiosciences, Frederick, MD, USA) at a multiplicity of infection (MOI) of 3 in 3 ml serum free α-MEM media for 5 h. An additional 3 ml of standard α-MEM was then added to each flask and the cultures were incubated for 2 days followed by selection for another 48 h using Puromycin (5 μg/ml). The cells were then harvested, plated at 1 × 10 4 cells/well in 96-well plates and cultured in complete α-MEM for 24 h. Luciferase activity was measured with a SteadyGlo-luciferase reporter assay system (Promega UK Ltd, Southampton, UK) at the indicated time points following RANKL stimulation (100 ng/ml) until osteoclasts were formed, as well as 24 and 48 h after RANKL withdrawal using a Bio-Tek Synergy HT plate reader (BioTek, Winooski, VT, USA).
Bone micro-computed tomography analysis Micro-computed tomography (μCT) analysis was performed using a Skyscan 1172 system (Bruker, Kontich, Belgium). The hind limbs of animals were dissected free of soft tissue, fixed in 4% formaldehyde in PBS, and tibial metaphysis and diaphysis were scanned ex vivo at a resolution of 5 μm (60 kV, 167 μA, using a 0.5-mm aluminum filter, and a rotation step size of 0.3 degrees). Image reconstruction was performed using the Skyscan NRecon package and trabecular and cortical bone parameters were measured using Skyscan CTAn software as described. (9) Briefly, trabecular bone parameters were measured in a stack of 200 slices immediately distal from the growth plate and, for cortical bone parameters, in 100 slices at the mid-shaft of the femur.

Semiquantitative analysis of osteolytic lesions
The presence and numbers of osteolytic lesions were assessed on μCT images of the hind limbs. Bone lesions were analyzed using a semiquantitative three-point scoring system to grade the lesions based on the size and number of lesions observed. Each animal was given a total lesion score based on the sum of scores generated from the left and right hind limbs.

Bone histology and histomorphometry
Histological analysis was performed as described. (10) After dissection, hind limbs were fixed for 24 h in 4% buffered formalin and stored in 70% ethanol. The samples were embedded in methyl methacrylate and sections were cut using a tungsten steel knife. Bone sections were stained for TRAcP to visualize osteoclasts and counterstained with Aniline blue. Cartilage matrix was visualized using Safranin O staining and Fast green as a counterstain. To achieve imaging of entire sections, image fields were stitched together using Microsoft Ice software (Microsoft Corp., Redmond, WA, USA). Histomorphometry was performed using custom software based on Image J (NIH, Bethesda, MD, USA; https://imagej.nih.gov/ij/) (11) and followed standardized nomenclature and recommendations. (12) Bisphosphonate therapy Heterozygous Tnfrsf11a 75dup27/− female mice were treated with zoledronic acid at a dose of 85 μg/kg (n = 8) or vehicle (PBS; n = 8) between 4 months and 12 months of age by subcutaneous injection every 2 months. The experiment was terminated by culling the mice at 15 months of age; the limbs were recovered and analyzed for the presence of bone lesions by μCT and histological analysis.

Statistical analysis
Differences between groups were assessed by unpaired Student's t test using SPSS for Windows version 19 (IBM Corp., Armonk, NY, USA). A p value <.05 was considered statistically significant. Data are presented as box and whiskers plots showing all data points, the interquartile range, the minimum and maximum values.

Ethical considerations
Ethical approval for the experiments was obtained from the Animal Welfare and the Ethical Review Body of the University of Edinburgh and the experiments were conducted according to the UK Animals Act of 1986 (Scientific Procedures).

Results
Phenotype of homozygous Tnfrsf11a 75dup27/75dup27 mice Mice that were homozygous for the 27-bp insertion of RANK (Tnfrsf11a 75dup27/75dup27 ) were viable but were smaller than their wild-type (WT) littermates. By the time of weaning, Tnfrsf11a 75dup27/75dup27 mice lacked incisors, and they were fed a soft food diet. They also had reduced survival; 31% (5/16) of Tnfrsf11a 75dup27/75dup27 mice died by 4 weeks of age compared to 0% (0/16) of WT littermates. All Tnfrsf11a 75dup27/75dup27 mice became progressively unwell with age due to the severe osteopetrosis phenotype and they had to be culled by the age of 4 months. None of the Tnfrsf11a 75dup27/75dup27 mice had incisors by the time of culling. Radiographic analysis of these mice showed metaphyseal widening and sclerosis of the tibia and femur characteristic of osteopetrosis (Fig. 1). The osteopetrosis was confirmed by μCT, which also showed metaphyseal widening and thinning of the cortex ( Fig. 2A-F). This phenotype was fully penetrant and affected all 16 mice analyzed. Detailed μCT analysis showed that Tnfrsf11a 75dup27/75dup27 mice had significantly higher trabecular bone volume/tissue volume (BV/TV), trabecular number (Tb.N), and trabecular thickness (Tb.Th) compared to WT, whereas the trabecular separation (Tb.Sp) was significantly lower than WT (Fig. 2G). The trabecular area and peripheral circumference of tibias were also significantly higher in homozygous mice. Histological analysis of Safranin-O-stained and Fast green-stained sections showed expansion of the growth plate and the bone marrow cavity was filled with mineralized cartilage. No osteoclasts were detected upon TRAcP staining of bone sections from homozygous animals (Fig. 2C,F). In keeping with these observations, bone marrow-derived macrophages from Tnfrsf11a 75dup27/75dup27 mice failed to form osteoclasts in vitro in response to RANKL and M-CSF (Fig. 2H).
Phenotype of heterozygous Tnfrsf11a 75dup27/− mice Heterozygous Tnfrsf11a 75dup27/− mice were viable and fertile, had normal survival, and had a body size similar to WT littermates. Analysis by x-ray showed no major differences between heterozygous and WT but some animals (3/18) aged between 12 and 15 months showed osteolytic lesions in the tibia (Fig. 1). μCT analysis of female 4-month-old mice showed normal trabecular  Table S2) with no evidence of osteolytic bone lesions. Consistent with the μCT results, histomorphometry of 4-month-old animals also showed no differences in bone density between WT and Tnfrsf11a 75dup27/− animals (Supplementary Table S3). There was a trend for reduced osteoclast surface/bone surface (Oc.S/BS) and osteoclast number/bone surface (Oc.N/BS) in heterozygous animals compared to WT but this was not statistically significant. Analysis of 4-month-old male mice revealed similar results to those from female mice (data not shown).
We went on to analyze bone density and architecture by μCT of 12-month-old and 15-month-old mice. We found that BV/TV, Tb.Th, and Tb.N were significantly higher in 12-monthold Tnfrsf11a 75dup27/− mice compared to WT (Fig. 3A,B). Additionally, Tb.Sp was significantly lower in Tnfrsf11a 75dup27/− mice compared to WT. There was a trend for larger cortical thickness (Cort.Th) in Tnfrsf11a 75dup27/− mice, but this was not statistically significant (Fig. 3B). We observed larger peripheral circumference in Tnfrsf11a 75dup27/− mice compared to WT, indicating bone enlargement (Fig. 3B). However, analysis of the skull bone showed no significant differences in morphology, width, length, or height between heterozygous (Het) and WT mice (Fig. 3C,D).
We then assessed the mice for the presence of bone lesions by analysis of μCT images of the hind limbs. Results from 12-monthold mice showed lytic lesions at the proximal tibia and distal femur close to the growth plate and at both lower femoral condyles in Tnfrsf11a 75dup27/− mice (Fig. 4A). These lesions were similar to those reported in other mouse models of Paget's disease. (10,13,14) Most 12-month-old heterozygous animals (7/8) showed lesions similar to those shown in Fig. 4A, but by 15 months all (10/10) Tnfrsf11a 75dup27/− mice had developed lesions. However, some animals showed lesions in one limb only (2/8 by 12 months and 2/10 by 15 months), reminiscent of the focal nature of PDB. Analysis of bone lesion severity score showed a significantly higher score in heterozygous animals compared to WT (Fig. 4B). The severity score was found to increase with age because 15-month-old animals had significantly higher score than 12-month-old animals (p = .005).
A small number of Het animals (3/18) developed larger focal lesions in the tibial shaft of one limb as shown in Fig. 4C and E. Histological analysis of these lesions revealed disorganized bone  Fig. 4D,F). Similarly, numbers of osteoblasts (OB) were higher in the lesion area (147.1 ± 8.8 OB/mm 2 ) compared to WT (21.6 ± 3.9; p = 5 × 10 −9 ).
Osteoclasts cultured from bone marrow cells of heterozygous Tnfrsf11a 75dup27/− mice were fewer in number and had a smaller number of nuclei than those formed from WT littermates (Fig. 5A,B). They also appear to have a denser TRAcP staining compared to the WT osteoclasts. The percentage of larger osteoclasts (more than five nuclei) was decreased in cultures from Tnfrsf11a 75dup27/− mice compared to WT at both 100 and 200 ng/ml (Fig. 5C). Analysis of osteoclast survival following withdrawal of RANKL showed that the percentage of surviving osteoclasts was significantly higher in the Tnfrsf11a 75dup27/− cultures compared to WT at all time points analyzed (Fig. 5D).

Effect of Tnfrsf11a 75dup27/− mutation on NF-κB signaling in vitro
In order to investigate the signaling mechanisms underlying the in vitro results for the Tnfrsf11a 75dup27/− model, we analyzed ERK, IκB, and p38 MAPK, c-Fos, and NF-κB pathways in MCSFdependent macrophages in response to RANKL. We found no difference in ERK phosphorylation at baseline, but following RANKL stimulation there was an increased phosphorylation of ERK in osteoclasts cultured from Tnfrsf11a 75dup27/− mice compared to WT, and by 30 min the levels of phosphorylated ERK (pERK) in Tnfrsf11a 75dup27/− mice were significantly higher than WT (Fig. 6A,B). However, IκB phosphorylation was significantly reduced in Tnfrsf11a 75dup27/− osteoclasts after 10 to 15 min of stimulation with RANKL compared with WT. Levels of phosphorylated p38 were greatly reduced in osteoclasts cultured from Tnfrsf11a 75dup27/− mice compared to WT, such that the mutant osteoclasts showed attenuated responses to RANKL. Analysis of c-Fos showed a significant increase in its level in mutant osteoclast precursor cells compared to WT, but after RANKL stimulation c-Fos levels were comparable to WT (Fig. 6C). To assess NF-κB activity during osteoclast differentiation, we used a NF-κB luciferase reporter assay and found decreased NF-κB activity in the Tnfrsf11a 75dup27/− at all stages during osteoclast differentiation compared with WT (Fig. 6D). However, when RANKL was withdrawn, NF-κB activity was higher in Tnfrsf11a 75dup27/− cells, compared with WT (Fig. 6E).
We then assessed the effect of Tnfrsf11a 75dup27/− mutation on serum levels of RANKL and biochemical markers of bone turnover in 4-month-old mice. We found no difference in serum levels of P1NP or CTX-I between Tnfrsf11a 75dup27/− and WT (Fig.  7A,B), but serum levels of RANKL were significantly lower in the mutant mice compared to WT (Fig. 7C).
Zoledronic acid prevents development of osteolytic lesions in heterozygous Tnfrsf11a 75dup27/− mice Administration of zoledronic acid between 4 and 12 months of age completely prevented the development of osteolytic lesions in heterozygous Tnfrsf11a 75dup27/− mice, whereas multiple lesions developed in the group receiving vehicle to a similar extent as in  pp38. (C) Western blot images of cFos in primary osteoclast cultures from Het compared to WT mice after RANKL stimulation. The box plot to the right represents band intensity analysis of cFos. All Western blot results are shown as a percentage of unstimulated WT (0 min) and were from two to four independent biological replicates. (D) Luciferase assay for NF-κB activation during osteoclast differentiation from bone marrow-derived cells upon stimulation with RANKL in WT compared to Het. (E) RANKL withdrawal; luciferase assay for NF-κB activation in osteoclasts and up to 48 h after RANKL withdrawal. Results are representative of three independent biological replicates. Abbreviations: cFos, Cellular oncogen Fos; Het, heterozygous; NF-κB, nuclear factor κB; pERK, phosphorylated extracellular signal-regulated kinase; pIκB, phosphorylated inhibitor of nuclear factor κB; pp38, phosphorylated p38; RANKL, receptor activator of nuclear factor κB ligand; WT, wild-type. untreated heterozygous mice (Fig. 8A,B). No obvious adverse effects were detected in mice that received zoledronic acid or PBS.

Discussion
EoPDB is a rare condition caused by a heterozygote duplication of nine amino acids in the signal peptide of the RANK protein due to a 27-bp duplication in exon 1 of TNFRSF11A. It shares some of the clinical features of classical PDB, but has an earlier age of onset, and is characterized by early-onset deafness, premature tooth loss, and prominent involvement of the mandible and maxilla. (1) Early functional studies of the TNFRSF11A 27-bp duplication showed that when constructs containing the mutant allele were overexpressed in HEK293 cells, there was an evidence of increased constitutive NF-κB activation in promoter-reporter assays when compared with WT. (3) These observations were confirmed by Crockett et al. (5) but when the 75dup27 allele and the related 84dup18 and 84dup15 alleles were stably expressed as a single copy in HEK 293 cells, there was no evidence of increased NF-κB activation. Furthermore, Crockett et al. (5) demonstrated that cells stably expressing these three mutant alleles failed to respond to RANKL stimulation and that the mutant proteins failed to translocate to the cell surface, but instead accumulated in organized smooth endoplasmic reticulum (OSER)-like structures. The overall conclusion from Crockett et al.'s experiments were that the insertion mutations were loss-of-function variants that result in a failure to respond to RANKL in vitro.
The experiments described here confirm that the 75dup27 mutation is a loss-of function variant which fails to respond to RANKL. Reflecting this fact, mice homozygous for the 75dup27 variant had features of osteopetrosis in vivo that include metaphyseal widening and sclerosis of the tibia and femur and total absence of osteoclasts in histological sections. Furthermore, the bone marrow cavity was filled with mineralized cartilage, which is confirmed by significantly higher BV/TV, Tb.N, Tb.Th, and trabecular area. Consistent with this, bone marrow cells from these mice failed to differentiate into osteoclasts in response to MCSF and RANKL stimulation, a phenotype that is similar to that of mice with targeted inactivation of RANK. (15) In contrast, Tnfrsf11a 75dup27/− mice appeared normal at a young adult age (4 months), but analysis by histomorphometry showed a trend for lower OC.N/BS compared to WT; nonetheless, trabecular bone indices analyzed by μCT showed no significant differences between heterozygous and WT mice. Additionally, the serum levels of bone turnover markers P1NP and CTX-I were comparable to those observed in age-matched WT mice; however, serum levels of RANKL were significantly reduced in heterozygous mice.
When heterozygous mice were analyzed at an older age (12 months), they exhibited higher BV/TV, Tb.N, and Tb.Th, and there was an evidence of bone enlargement as evident from larger peripheral circumference in hind limbs of heterozygous mice compared to WT. However, the skull bones from these mice were normal with no significant differences to WT.
We also found that heterozygous mice developed focal osteolytic lesions in the lower limbs with increasing age, consistent with the phenotype observed in humans with this mutation. At the age of 12 months, most animals developed osteolytic lesions at the proximal tibia and distal femur close to the growth plate and the lesion severity was found to increase with age.
To identify the molecular mechanisms involved in the maintenance of the osteoclast activity, we evaluated a number of molecular targets from MAPK and NF-κB pathways that were associated with osteoclastogenesis and osteoclast function. We found that cells cultured from mice heterozygous for the 75dup27 mutation showed a decreased IκB and p38 phosphorylation, compared to WT controls. A decrease in the expression of the active form of IκB (pIκB) has been associated with reduced osteoclastogenesis (reviewed in Boyce et al. (16) ). In the same line, p38, a member of the MAPK family, plays a role in the differentiation of bone marrow cells into osteoclasts. (17) Therefore, the reduction in osteoclastogenesis found in the Tnfrsf11a 75dup27/− model could be caused by a decrease in IκB and p38 activation.  We also detected an increase of ERK phosphorylation in the heterozygotes compared with WT. ERK is involved in a number of processes in osteoclasts, including survival, proliferation, cell polarity, and differentiation, among others. (18) Increasing ERK phosphorylation could be associated to the increased osteoclast survival found in our model, because the increase in pERK was sustained for longer time and became significant at 30 min after RANKL activation. Additionally, basal c-Fos expression was higher in osteoclast precursors of heterozygous animals, which may also partially explain the increased survival of osteoclasts.
Analysis of NF-κB activity during osteoclast differentiation in the Tnfrsf11a 75dup27/− mice showed that RANKL-induced NF-κB activation was reduced in heterozygous animals compared to WT, consistent with our Western blot results and with previous in vitro studies. (5) However, after withdrawal of RANKL, the NF-κB expression is sustained and increased compared with WT controls. NF-κB is important for osteoclast function and survival (reviewed in Novack (19) ), and its continuous activation to trigger the RANK pathway in absence of the ligand could be associated with increased osteoclast survival, ultimately leading to the bone lesions reminiscent of Paget's disease that were found in the Tnfrsf11a 75dup27/− mice in vivo.
Previous in vitro studies have reported a continuous activation of the RANK pathway caused by mutations in the signal peptide of RANK leading to a hyperactivity of the osteoclasts, but the mechanism involved was unclear. (3) It has been demonstrated that the 75dup27 and other mutations in the signal peptide of RANK impair cleavage of the signal peptide and prevent the RANK protein from migrating to the plasma membrane. (20,21) Instead the mutant protein has been shown to accumulate in OSER-like structures in endoplasmic reticulum. (22) The mechanism by which the abnormal protein stimulates osteoclast survival is unknown, but a possibility would be by triggering the unfolded protein response (UPR), which is known to cause activation of NF-κB signaling. (23) Furthermore, accumulation of unfolded proteins and/or a decreased capacity to degrade unfolded proteins with age (24) could contribute to increased penetrance of the Paget's disease-like phenotype with age that we observed in Tnfrsf11a 75dup27/− mice.
In addition to clarifying the mechanisms of action by which these mutations cause osteolytic lesions, the present experiment has shown that when administered early in life, zoledronic acid can prevent the development of lesions at older age. Although bisphosphonates have not been shown to reverse bone deformity, improve mobility, or prevent deafness and tooth loss in patients with EoPDB, (2) it should be noted that the treatment was administered at a late stage in the disease process. Our data suggest that if prophylactic therapy was given at an early stage, it may be possible to prevent or inhibit the progression of bone disease in this severe and disabling disorder.