We targeted the MVNP gene to the OCL lineage in transgenic mice. These mice developed abnormal OCLs and bone lesions similar to those found in Paget's patients. These results show that persistent expression of MVNP in OCLs can induce pagetic-like bone lesions in vivo.
Introduction: Paget's disease (PD) of bone is the second most common bone disease. Both genetic and viral factors have been implicated in its pathogenesis, but their exact roles in vivo are unclear. We previously reported that transfection of normal human osteoclast (OCL) precursors with the measles virus nucleocapsid (MVNP) or measles virus (MV) infection of bone marrow cells from transgenic mice expressing a MV receptor results in formation of pagetic-like OCLs.
Materials and Methods: Based on these in vitro studies, we determined if the MVNP gene from either an Edmonston-related strain of MV or a MVNP gene sequence derived from a patient with PD (P-MVNP), when targeted to cells in the OCL lineage of transgenic mice with the TRACP promoter (TRACP/MVNP mice), induced changes in bone similar to those found in PD.
Results: Bone marrow culture studies and histomorphometric analysis of bones from these mice showed that their OCLs displayed many of the features of pagetic OCLs and that they developed bone lesions that were similar to those in patients with PD. Furthermore, IL-6 seemed to be required for the development of the pagetic phenotype in OCLs from TRACP/MVNP mice.
Conclusions: These results show that persistent expression of the MVNP gene in cells of the OCL lineage can induce pagetic-like bone lesions in vivo.
PAGET'S DISEASE (PD) of bone is the second most common bone disease, affecting 1-2 million patients in the United States. Although the etiology of PD is unknown, both genetic and nongenetic factors have been implicated. Studies of large families with PD have shown an autosomal dominant mode of inheritance, and recently, several loci have been linked to PD.(1-5) Mutations in the sequestosome-1 (SQSTM1) gene occur in ∼30% of patients with familial or 10% of patients with sporadic PD,(6) although the penetrance of PD in families with these mutations is variable. In addition, previous studies have suggested a viral etiology for PD. Electron microscopic studies first showed nuclear inclusions in pagetic osteoclasts (OCLs), which were similar to paramyxoviral nucleocapsids.(7) Immunohistochemical studies subsequently identified both respiratory syncytial virus and measles virus nucleocapsid proteins (MVNPs) in pagetic OCLs.(8) In situ hybridization studies also showed MVNP transcripts in cells from bone biopsy specimens from patients with PD,(9) and RT-PCR studies identified MVNP or canine distemper virus nucleocapsid transcripts in OCLs from patients with PD.(10,11) However, others have been unable to detect viral transcripts in pagetic OCLs.(12,13) Thus, the role of paramyxoviruses in the pathogenesis of PD is unclear.
We previously reported that transfection of normal human OCL precursors with the MVNP gene results in formation of OCLs that have many of the abnormal features of pagetic OCLs.(14) Both pagetic and MVNP-transfected normal OCL precursors form markedly increased numbers of OCLs in vitro, which contain many more nuclei per OCL and have an increased resorption capacity compared with normal OCLs. Furthermore, both pagetic and MVNP-transfected normal OCL precursors display marked hyper-responsivity to 1,25(OH)2D3, forming OCLs at concentrations that are one to two logs lower than required for normal OCL formation.(15) In addition, both pagetic and MVNP-transfected OCL precursors express high levels of TAFII-17, a member of the TF-IID transcription complex, which acts as a coactivator of vitamin D receptor-mediated gene transcription.(16) OCL from PD patients and OCL precursors transduced with the MVNP gene also secrete large amounts of IL-6.(14,17) Finally, when bone marrow cells from transgenic mice in which the CD46 MV receptor(18) is targeted to cells in the OCL lineage are infected in vitro with MV, they form OCLs that have the abnormal characteristics of pagetic OCLs.(19) However, it is unknown if MV can induce pagetic-like bone lesions in vivo that are similar to the abnormal bone present in PD.
In this study, we determined if persistent expression of the nucleocapsid gene from either an Edmonston variant of MV (E-MVNP) or the nucleocapsid sequence derived from a patient with PD (P-MVNP)(20) could induce changes in bone similar to those found in PD. The E-MVNP or P-MVNP gene was targeted to cells in the OCL lineage in transgenic mice using the TRACP promoter. These mice were analyzed at 4-16 months of age to determine if they developed bone abnormalities similar to those seen in PD.
MATERIALS AND METHODS
Development of TRACP/E-MVNP and TRACP/P-MVNP transgenic mice
These studies were approved by the IACUCs at both the University of Pittsburgh School of Medicine and Virginia Commonwealth University. The E-MVNP cDNA, originally derived from a measles patient, was generously provided by Dr Chris Richardson of the University of Toronto. Sequence analysis of this cDNA showed that it was from a virus belonging to the Edmonston group of MV strains, which is the most widespread group of MV strains and the origin of the majority of MV vaccines.(21) This cDNA encodes a protein that differs from the Edmonston strain wildtype MVNP (GI: 1041617)(22) at five amino acid residues: 26 (G to E), 453 (E to G), 467 (L to P), 473 (L to P), and 525 (D to G). Sequence analysis of the P-MVNP gene showed that it also encodes a closely related Edmonston strain MVNP that differs from the wildtype Edmonston MVNP at seven residues: 26 (G to E), 435 (K to R), 453 (E to G), 467 (L to P), 473 (L to P), 494 (A to T), and 525 (D to E). The majority of the amino acid differences between either the E-MVNP or P-MVNP and the wildtype MVNP fall within the hypervariable carboxy terminus of this protein. When we initially reported the detection of MVNP transcripts from the bone marrow of several Paget's patients,(20) the sequence from patient 1 (P1) diverged from the consensus MVNP sequence beginning at amino acid 497. It was subsequently determined that this apparent divergence was caused by a DNA sequencing error that produced a frameshift in P1 relative to the consensus MVNP sequence and that the P1 sequence in fact matches wildtype MVNP with the exceptions noted above. To generate the TRACP/E-MVNP and TRACP/P-MVNP transgenes, the E-MVNP and P-MVNP cDNA was inserted into the unique EcoRI site of the pBSmTRACP5′ plasmid.(23,24) This resulted in the addition of a 25 amino acid C-terminal tag to the P-MVNP but not the E-MVNP construct. The transgenes were excised with XhoI, and transgenic mice were generated by standard methods(25) in a CB6F1 (C57Bl/6 × Balb/c) genetic background. Transgenic founders were identified by Southern blot analysis of tail DNA, and transgenic mice of subsequent generations were identified by PCR analysis. Two TRACP/E-MVNP and four TRACP/P-MVNP founder mice were generated and bred to establish multiple independent lines of mice. OCL formation assays were performed on marrow cultures from all transgenic lines, and the TRACP/E-MVNP and TRACP/P-MVNP line that had the highest level of MVNP expression and OCL formation were selected for further characterization and longitudinal studies. This was necessitated by the large numbers of mice that had to be maintained and the extensive histomorphometric analysis that was required. All data presented here were derived from one line for each transgene that expressed the highest levels of the transgene, although increased levels of OCL formation and hypersensitivity to 1,25(OH)2D3 were found in marrow cultures from the additional transgenic lines.
Immunohistochemical detection and Western blot analysis of MVNP expression in marrow cells from TRACP/MVNP and WT mice and PD patients
OCLs from nonadherent mouse bone marrow cells (2 × 105 cells/well) from TRACP/MVNP or WT mice cultured for 7 days with 10−8 M 1,25(OH)2D3 were tested for cross-reactivity with a monoclonal antibody against the MVNP protein (Gene Tex, San Antonio, TX, USA) or mouse IgG (60 ng/ml) using a Vectastatin-ABC-AP kit (Vector Laboratories, Burlingame, CA, USA) as previously described.(19) For Western blot analysis, nonadherent mouse bone marrow cells (1.2 × 106 cells) from TRACP/MVNP or WT mice were cultured with macrophage-colony stimulating factor (M-CSF; 10 ng/ml)/RANKL (25 ng/ml) in α-MEM-10% FBS for 48 h. Cell lysates were prepared and processed for Western blot analysis as previously described.(14) The MVNP monoclonal antibody or β-actin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used at 1:1000 dilution in Tris-buffered saline containing 1% BSA. Nonadherent marrow mononuclear cells (5 × 105 cells) from patients with PD or normals were cultured for 3 weeks with 10−8 M 1,25(OH)2D3, to induce OCL formation, and the cells were processed for Western blot analysis as described above.
PCR analysis for TAFII-17 expression
MVNP and WT mice OCL precursors were cultured for 2 days with 1,25(OH)2D3 (10−8 M) and subjected to RT-PCR for TAFII-17 mRNA as previously described.(16) The mouse TAFII-17 sense primer was 5′-CAGATTATGAACCAGTTTGGCCCTTCA-3′ and was derived from the TAFII-17 gene sequence (GenBank accession no. AL590442). The TAFII-17 antisense primer was 5′-CCTGTGTTATTTCTTGGTTGTTTTCCG-3′. The actin sense and antisense primers were 5′-GGCCGTACCACTGGCATCGTGATG-3′ and 5′-CLTGGCCGTCAGGCAGCTCGTAGC-3′ and were derived from the actin gene sequence (GenBank accession no. NM000492).
In vitro analysis of OCL formation by bone marrow cells from TRACP/MVNP or WT mice
Nonadherent marrow cells from long bones of TRACP/E-MVNP, TRACP/P-MVNP, or WT mice at 4-16 months of age were cultured for OCL formation in the presence of varying concentrations of 1,25(OH)2D3 (Roche, Indianapolis, IN, USA) or with 10 ng/ml of murine M-CSF and 25 ng/ml RANKL (R&D Systems, Minneapolis, MN, USA) for 7 days as previously described.(26) The cells were stained for TRACP activity using a commercial kit (Sigma, St Louis, MO, USA), and the number of TRACP+ multinucleated cells that contained at least three nuclei, as well as the number of nuclei per multinucleated cell, were scored microscopically. IL-6 levels in conditioned media from these marrow cultures were determined using a commercial ELISA kit (R&D Systems). In selected experiments, marrow cultures were treated with vehicle or 1,25(OH)2D3 in the presence of a neutralizing antibody to murine IL-6 (50 ng/ml) or isotype specific mouse IgG (R&D Systems) to assess the effects of IL-6 on OCL formation.
Histologic analysis of TRACP/E-MVNP and TRACP/P-MVNP vertebral bones
The first to fourth lumbar vertebrae from TRACP/E-MVNP, TRACP/P-MVNP, and WT mice were fixed in 10% buffered formalin and completely decalcified in 10% EDTA at 4°C and embedded in paraffin. Five-micrometer longitudinal sections were cut and mounted on glass slides. Deparaffinized sections were stained for TRACP as described by Liu et al.(27) OCLs containing active TRACP were stained red. Another set of sections was stained with 0.1% toluidine blue.
Histomorphometry was performed on the region of cancellous bone between the cranial and caudal growth plates of the third lumbar vertebral body under bright field and polarized light at a magnification of ×200, using the OsteoMeasure 4.00C morphometric program (OsteoMeasure; OsteoMetrics, Atlanta, GA, USA). Osteoclast perimeter (Oc.Pm) was defined as the length of bone surface covered with TRACP+ mono- and multinuclear cells. Osteoblast perimeter (Ob.Pm), cancellous bone volume (BV/TV), trabecular width (Tb.Wi), trabecular number (Tb.N), and trabecular separation (Tb.Sp) were also quantified and calculated.
To examine bone formation parameters, animals were given calcein (10 mg) subcutaneously on days 7 and 1 before death. Bones from these animals were embedded, undecalcified, in methyl methacrylate and sections were examined under fluorescent light for quantification of mineralized perimeter, mineral apposition rate, and bone formation rate. All variables were expressed and calculated according to the recommendations of the ASBMR Nomenclature Committee.(28,29)
In vitro culture results were analyzed by a two-way ANOVA. Histomorphometric variables were analyzed by two-factor ANOVA using NCSS 2004 software (NCSS Statistical Software, Kaysville, UT, USA). Genotype and age were assigned as factors and the responses of the measured variables were tested. A p value of <0.05 was considered statistically significant.
Analysis of OCLs from TRACP/E-MVNP and TRACP/P-MVNP mice for expression of MVNP
Immunohistochemical analysis showed that OCL formed in marrow cultures from TRACP/E-MVNP and TRACP/P-MVNP mice expressed MVNP (Fig. 1A). Similar levels of staining were detected in both the TRACP/E-MVNP and TRACP/P-MVNP OCLs, and no staining was seen in OCLs from nontransgenic control (WT) or normal human marrow cultures. These results were confirmed by Western blot analysis of MVNP expression, which showed expression of MVNP in TRACP/MVNP and PD patient samples but not in WT mice or normal marrow (Figs. 1B and 1C). Importantly, the levels of MVNP expression in both lines of transgenic mice was roughly comparable with that seen in OCL formed from marrow cultures of patients with PD (Fig. 1C).
Characterization of OCLs formed in marrow cultures of TRACP/E-MVNP and TRACP/P-MVNP mice
Significantly more OCLs were formed in marrow cultures from TRACP/E-MVNP and TRACP/P-MVNP mice than from WT mice in response to 1,25(OH)2D3 (Fig. 2A). Furthermore, both TRACP/E-MVNP and TRACP/ P-MVNP marrow cultures formed OCLs at concentrations of 1,25(OH)2D3 that were significantly lower than those required for WT marrow cultures, with OCL formation occurring in MVNP cultures at 10−11 to 10−12 M 1,25(OH)2D3. In addition, the OCL precursors from TRACP/P-MVNP and E-MVNP, but not WT mice, expressed high levels of TAFII-17 mRNA (Fig. 1D). The number of nuclei per OCL was also significantly increased in marrow cultures from TRACP/E-MVNP and TRACP/P-MVNP mice compared with WT mice (Fig. 2B), and the OCLs that formed were larger than those formed in WT marrow cultures (Fig. 2C). In contrast, there was no significant difference in the sensitivity of OCL precursors from WT, TRACP/E-MVNP, and TRACP/P-MVNP mice to RANKL or expression of RANK mRNA in OCL precursors from these mice (data not shown).
IL-6 production and effects of anti-IL-6 in bone marrow cultures from TRACP/P-MVNP or WT mice
Low levels of IL-6 were detected in bone marrow-conditioned media from marrow cultures of TRACP/P-MVNP and WT mice treated with vehicle. In contrast, high levels of IL-6 were present in conditioned media of cultures of TRACP/P-MVNP, but not WT mice, treated with 1,25(OH)2D3 to induce OCL formation (Fig. 3A). In contrast, IL-11 levels were similar in conditioned media from WT and TRACP/P-MVNP or TRACP/E-MVNP marrow cultures (data not shown).
We determined if IL-6 also played a role in the formation of pagetic-like OCLs in TRACP/P-MVNP marrow cultures. Addition of an anti-IL-6 antibody (50 ng/ml) to marrow cultures of TRACP/P-MVNP mice treated with 1,25(OH)2D3 significantly decreased OCL numbers and the number of nuclei per OCL (Fig. 3B; Table 1). In contrast, anti-IL-6 had no effect on OCL formation or nuclear number per OCL in WT bone marrow cultures (Fig. 3B; Table 1).
Table Table 1.. Effects of Anti-IL-6 on Nuclear Number/OCLs in TRACP/MVNP Mice
Histology and histomorphometry of bones from TRACP/E-MVNP and TRACP/P-MVNP mice
There were no significant differences in the measured histomorphometric variables between bones from TRACP/E-MVNP and P-MVNP mice, and their histological features were qualitatively very similar. The data from these two groups were therefore pooled as one MVNP group and compared with WT animals.
MVNP OCLs were larger in size and had more nuclei per cell and the resorption cavities were deeper in MVNP bone than in WT bone. Furthermore, tunneling resorption was present in MVNP bone but was rarely seen in WT bone (Figs. 4A-4C). Plump, cuboidal osteoblasts were more common in the MVNP than in the WT bone (Figs. 4D and 4E). Dynamic histomorphometric variables from nonlesioned bone in the calcein-labeled animals are shown in (Fig. 5A). Mineralized perimeter, mineral apposition rate, and bone formation rate were all significantly higher in the MVNP mice than in WT controls.
Table Table 2.. Cancellous Bone in 12-Month MVNP and WT Mice
Markedly abnormal bone structure was seen in at least two of the four vertebrae examined in a subset of 4 of the 14 MVNP mice (29%) at 12 months of age. Two of these animals were TRACP/E-MVNP mice and two were TRACP/P-MVNP mice. These lesions were histologically similar to those seen in PD and were characterized by focally thickened and irregular trabeculae composed mainly of woven bone (Fig. 6). Cancellous bone volume, trabecular number, trabecular width, OCL perimeter, and osteoblast perimeter were all significantly increased in these four animals compared with age-matched WT controls, whereas trabecular separation was significantly reduced (Fig. 5B). None of these histological features was seen in WT controls.
To determine whether the dramatic changes seen in the vertebrae from these four animals were localized to individual vertebrae, we measured the histomorphometric variables in adjacent vertebrae that, qualitatively, did not appear to be as severely affected. These data are shown in Table 2. Whereas bone microarchitecture and turnover variables were not as abnormal in the adjacent vertebrae, they were still significantly different from those in WT animals. Data from the animals studied at 4, 8, and 12 months of age, including the lesioned bone from the four 12-month-old mice in Fig. 6, are given in Table 3. OCL perimeter was increased by 20-58% in MVNP mice compared with those from WT, and osteoblast perimeter was increased by 26-61%. The magnitude of the differences between MVNP and WT in OCL and osteoblast perimeters increased with age.
Table Table 3.. Cancellous Bone Structure in MVNP and WT Mice
In this study, we determined the capacity of two different MVNP genes, one from an Edmonston group of MV originally isolated from a measles patient and one derived from a patient with PD, to induce a Paget's-like phenotype in transgenic mice. Expression of the two nucleocapsid genes was directed to cells in the OCL lineage using the TRACP promoter, which is highly expressed in OCLs and OCL precursors and has been used previously to target expression of multiple genes to cells in the OCL lineage.(30) TRACP is also expressed in chondrocytes and occasional osteoblasts in bone(31) but at very low levels compared with OCLs.
OCL precursors from TRACP/E-MVNP and TRACP/ P-MVNP mice were found to be similar to each other and express almost all the features of pagetic OCL precursors. These include increased levels of OCL formation and a marked hyper-responsivity to 1,25(OH)2D3. In addition, the OCLs that form are larger and contain many more nuclei per OCL. The only phenotypic difference that distinguished OCL precursors from TRACP/E-MVNP or TRACP/P-MVNP mice from OCL precursors from PD patients is that TRACP/MVNP OCL precursors are not hyper-responsive to RANKL.(32,33) These data suggest that additional factors, possibly genetic factors linked to PD, may be responsible for the hyper-responsivity of pagetic OCLs to RANKL.
Importantly, bones from the TRACP/E-MVNP or TRACP/P-MVNP mice displayed many of the histologic and histomorphometric features of bone lesions from patients with PD. These include an increase in mineralized perimeter, mineral apposition rate, bone formation rate, an increase in OCL and osteoblast perimeters, increases in the number and size of OCLs with more nuclei/OCL, deeper resorption cavities and tunneling resorption, and abundant large cuboidal osteoblasts. Furthermore, the bone that was formed was abnormal and was woven in character, similar to that seen in pagetic lesions. These changes were particularly evident in 30% of the animals at 12 months of age. In addition to the marked increase in turnover, the lesions in these animals displayed a dramatic increase in bone volume and trabecular thickness. The abnormally thickened and coarse trabeculae were remarkably similar to those seen in PD, and the fact that these dramatic lesions were only observed in the oldest animals is also consistent with the slow development of pagetic lesions. Furthermore, not all of the bones in these animals were as severely affected, consistent with a variable rate of expression of the phenotype in different bones, although adjacent bones showed increased OCL and osteoblast activity (Table 2). These data suggest that the rate of development of the pagetic-like lesions differed in the individual vertebral from these mice. In contrast to these findings, high turnover states are generally associated with reduced, rather than increased, bone volume in both humans and experimental animals. For example, transgenic mouse models of both primary and secondary hyperparathyroidism display increased resorption and formation, but bone volume is reduced.(34,35) Also, unlike the histological changes accompanying chronic PTH excess,(36) the MVNP mice did not show any peritrabecular or marrow fibrosis. Increased bone volume does accompany high bone turnover states, but only when the stimulus is anabolic (e.g., intermittent, exogenous administration of PTH or PGE2).(37,38) Woven bone formation is also seen under such circumstances but requires very high doses.(37,38) Furthermore, such anabolic effects occur throughout the skeleton, rather than being restricted to individual bones, as seen in the TRACP/MVNP mice. Furthermore, other transgenic mouse models in which the TRACP promoter has been used to target genes to the OCL lineage do not develop pagetic-like bone lesions similar to those found in TRACP/MVNP mice.(30) When taken together, these observations strongly suggest that the bone lesions observed in the MVNP mice are the result of osteoclastic expression of the nucleocapsid genes rather than being caused by a generalized high turnover state.
We previously hypothesized that the sequence variants of the MVNP transcripts isolated from Paget's patients might contribute to the unique pathogenic role of MVNP in PD.(20) However, in this study, we introduced two MVNP variants into transgenic mice, one originally derived from a MV patient and one derived from a Paget's patient, and the resulting phenotype is indistinguishable. The majority of the amino acid substitutions between either E-MVNP or P-MVNP and the Edmonston WT MVNP, as well as the differences between E-MVNP and P-MVNP, fall within the C-terminal region of the protein, which is known to be hypervariable.(21) Thus, it now seems unlikely that the unique amino acid substitutions seen in P-MVNP specifically contribute to its pathogenic role in PD.
The MVNP gene can have effects on other human cells in addition to OCLs. During initial infection by MV, transient profound immune suppression occurs, followed by development of long-term immunity to MV.(39) The mechanism for this immunosuppression involves binding of the nucleocapsid protein to the Fc-γ receptors on dendritic cells, resulting in suppression of IL-12 expression and increased IL-6 expression.(40) These data suggest that MVNP can have profound effects on cellular function in cells of the monocyte macrophage lineage, and it is the same precursor cell in this lineage that gives rise to both OCL and dendritic cells. Thus, it is reasonable that the OCL precursors in the monocyte macrophage lineage could be affected by MV.
The mechanism(s) underlying the capacity of the MVNP gene to induce pagetic-like OCLs in vitro and in vivo are still being defined. Previous reports have shown that chronic infection of human glial cells with MV markedly increases IL-6 production with little or no increase in IL-1β or TNF-α expression.(41) Induction of IL-6 expression in OCL seems to play a role in the abnormal OCLs formed in TRACP/MVNP mice. lL-6 levels were increased in conditioned media from marrow cultures of TRACP/P-MVNP mice induced to form OCLs, and an anti-IL-6 antibody decreased both OCL formation and nuclei/OCL in TRACP/ P-MVNP marrow cultures. In contrast, IL-6 levels were not increased in WT marrow cultures treated with 1,25(OH)2D3, and anti-IL-6 had no effect on OCL formation in WT cultures. We previously showed that OCLs formed in cultures of marrow from PD patients also produce high levels of IL-6.(17) Taken together, these data support an important role for IL-6 in the abnormal OCL formation in PD and suggest that MVNP may be responsible for the increased IL-6 expression in pagetic OCLs. These results further suggest that other factors yet to be defined are involved in the abnormal OCL formation in TRACP/MVNP mice because anti-IL-6 did not reduce OCL formation to control levels. However, this factor does not seem to be IL-11, because IL-11 levels were not increased by MVNP expression in OCL precursors.
Although the experiments reported here do not prove that MV can cause PD, they clearly show that expression of MVNP in cells of the OCL lineage can result in bone lesions and abnormalities in OCL precursors that are very similar to those found in patients with PD. In addition, just as in patients with a genetic predisposition to PD (e.g., an inherited SQSTM1 mutation), the development of PD is variable, not all of the mice in our cohorts developed PD-like lesions. Furthermore, like patients with PD, the lesions that did arise were focal despite the forced expression of MVNP in the majority of OCLs in the transgenic mice.
Thus, persistent expression of MVNP in cells of the OCL lineage can induce pagetic-like lesions in vivo. These results suggest that persistent expression of MVNP in OCLs is an important component in the complex etiology of PD.
This work was supported by PO1-AR049363, VCU Massey Cancer Center Support Grant P30-CA16059, and USAMRMC Award DAMD17-03-1-0763. We thank the GCRC at the University of Pittsburgh for assistance with obtaining the marrow samples.