High Bone Mass in Mice Expressing a Mutant LRP5 Gene

Authors


Abstract

A unique mutation in LRP5 is associated with high bone mass in man. Transgenic mice expressing this LRP5 mutation have a similar phenotype with high bone mass and enhanced strength. These results underscore the importance of LRP5 in skeletal regulation and suggest targets for therapies for bone disease.

A mutation (G171V) in the low-density lipoprotein receptor related protein 5 (LRP5) has been associated with high bone mass (HBM) in two independent human kindreds. To validate the role of the mutation, several lines of transgenic mice were created expressing either the human LRP5 G171V substitution or the wildtype LRP5 gene in bone. Volumetric bone mineral density (vBMD) analysis by pQCT showed dramatic increases in both total vBMD (30-55%) and trabecular vBMD (103-250%) of the distal femoral metaphysis and increased cortical size of the femoral diaphysis in mutant G171V transgenics at 5, 9, 17, 26, and 52 weeks of age (p < 0.01 for all). In addition, high-resolution microcomputed tomography (microCT) analysis of the distal femorae and lumbar vertebrae revealed an increase (110-232%) in trabecular bone volume fraction caused by both increased trabecular number (41-74%) and increased trabecular thickness (34-46%; p < 0.01 for all) in the mutant G171V mice. The increased bone mass was associated with significant increases in vertebral compressive strength (80-140%) and the increased cortical size with significant increases in femoral bending strength (50-130%). There were no differences in osteoclast number at 17 weeks of age. However, compared with littermate controls, the mutant G171V transgenic mice showed an increase in actively mineralizing bone surface, enhanced alkaline phosphatase staining in osteoblasts, and a significant reduction in the number of TUNEL-positive osteoblasts and osteocytes. These results suggest that the increased bone mineral density in mutant G171V mice was caused by increased numbers of active osteoblasts, which could in part be because of their increased functional lifespan. While slight bone anabolic activity was observed from overexpression of the wildtype LRP5 gene, it is clear that the G171V mutation, rather than overexpression of the receptor itself, is primarily responsible for the dramatic HBM bone effects. Together, these findings establish the importance of this novel and unexpected role of a lipoprotein receptor in regulating bone mass and afford a new model to explore LRP5 and its recent association with Wnt signaling in bone biology.

INTRODUCTION

GENETIC CONTROL OF bone mineral density (BMD) has been investigated by identification of osteoporosis susceptibility genes or chromosomal regions in both humans and in a variety of animal models.(1–4) While many chromosomal loci have been implicated in the control of BMD, human chromosome 11, and specifically the region 11q12–13, seems to be particularly intriguing. Recently, candidate genes from this locus were identified by traditional linkage analysis in three rare monogenic bone disorders. Autosomal recessive osteopetrosis results from inactivating mutations in the osteoclast proton pump TCIRG1 (OC116).(5, 6) Interestingly, the low-density lipoprotein receptor-related protein 5, LRP5, gene has been implicated in disorders resulting in either gain or loss of BMD. Inactivating mutations of LRP5 cause the juvenile-onset osteoporosis pseudoglioma syndrome (OPPG),(7, 8) while an autosomal dominant high bone mass (HBM) trait results from a single point mutation in LRP5.(9, 10) Further implications of genetic regulation of BMD from chromosome 11q12–13 come from linkage analysis that has refined a 6.6-cM region responsible for autosomal dominant osteopetrosis type I (ADOI).(11) At this time, both LRP5 and TCIRG1 remain ADOI candidates, residing within the critical interval. Last, quantitative trait locus (QTL) analysis suggested that one or more of the above genes, or other genes in the region, may contribute to the normal variation in BMD seen in the general population.(12, 13)

Affected family members from the high bone mass (HBM) kindred show a unique skeletal phenotype with hip and spine BMD values that are approximately 40–50% greater than age and gender-matched normal values.(9) The phenotype is present in young adults as well as elderly family members of both genders into the eighth decade.(9, 10) The bones of affected individuals, while appearing very dense radiographically, have normal external shape and outer dimensions and seem to have achieved a balance in bone turnover at a density that is significantly greater than necessary for normal skeletal stresses. Mutation analyses have revealed a single nucleotide mutation in the LRP5 gene resulting in an amino acid substitution (G171V) that was only present in HBM-affected individuals.(10) Remarkably, this observation was recently confirmed by the subsequent identification of a second high bone mass kindred with an identical LRP5 mutation.(14) Besides the HBM substitution, other mutations have been described for LRP5. In individuals with OPPG, nine disease-causing mutations in exons encoding the LRP5 extracellular domain have been identified, with each predicted to result in either frameshift or nonsense mutations.(8) These findings are consistent with in vitro data showing that C-terminal deletions of LRP5 result in dominant negative isoforms.(8) Furthermore, this group identified three potential disease-associated missense mutations in regions encoding the LRP5 extracellular domain: R494Q, R570W, and V667M.(8) It remains to be elucidated how these missense mutations result in loss of LRP5 function. Lastly, mutation analyses have identified two additional missense mutations in LRP5 resulting in Q89R and V1330A.(15) Their effects on bone, however, were not evaluated in this study. Thus, a variety of mutations in LRP5 have been identified, and several seem to be implicated with bone-related traits. In this study, we have focused on the HBM substitution at LRP5 residue 171 to further investigate its role in regulating BMD.

The finding of the LRP5 gene, a member of the low-density lipoprotein receptor (LDLR) family, as a modulator of bone metabolism was initially surprising. In addition to LRP5, both the related LRP6 and Drosophila arrow genes comprise a subclass of the LDLR family with previously unknown function.(16) However, a number of recent studies in mouse, fly, and Xenopus have now implicated LRP5/LRP6/arrow as key modulators of the Wnt/wingless signal pathway.(17–19) Wnt signaling has been shown to play key roles in both embryonic development and oncogenesis.(20, 21) Wnts are secreted proteins that are thought to interact with LRP5/LRP6 in a complex with Frizzled receptors.(18, 22, 23) Intracellular signals are propagated through the disheveled protein, which in turn, inhibits glycogen synthase kinase 3β from phosphorylating β-catenin. Phosphorylated β-catenin is rapidly degraded, whereas stabilized, de-phosphorylated β-catenin accumulates and translocates to the nucleus where it acts as a cofactor of the T cell factor (TCF) transcription activator complex. In addition to Wnts and Frizzleds, LRP5 has been shown to interact with two additional Wnt pathway components: axin(22) and dickkopf.(24–26) Both of these molecules are negative regulators of the Wnt pathway and suggest a complex mechanism of LRP5-mediated signaling. Moreover, LRP5 is expressed in a wide variety of tissues and cells including osteoblasts,(10, 23, 27–29) which raises the question how LRP5 exerts its bone-specific effects. In this regard, although LRP5 expression is widespread, LRP5 null mice display two distinct tissue-specific phenotypes; osteoblast defects resulting in low bone mass and abnormal endothelial cell apoptosis resulting in persistent eye vascularization.(23) To validate and investigate the G171V substitution as being the cause of the human HBM phenotype, we created two sets of transgenic mice: one that overexpressed the human LRP5 G171V mutation and a second set that overexpressed the human wildtype LRP5 gene in bone; both use the same rat type I collagen promoter.(30, 31) Our data show that wildtype LRP5 overexpression in bone has a modest anabolic effect on bone mass and microarchitecture, whereas overexpression of the G171V mutant has a dramatic anabolic effect in bone, suggesting a direct effect of LRP5 in osteoblast function.

MATERIALS AND METHODS

Generation of transgenic mice

The G to T (G171V) mutation was introduced into human LRP5 cDNA (GenBank AF077820) by polymerase chain reaction (PCR), confirmed by DNA sequencing, and linked downstream of the 3.6-kb rat type I collagen promoter.(30, 31) An SV40 intron and polyadenylation cassette (GenBank NC001669) was placed downstream of the cDNA. A second set of transgenic mouse lines was generated by replacing the mutant cDNA with the wildtype human LRP5 cDNA. For both sets of transgenic mice, DNA was microinjected into C57BL/6Tac mouse embryos according to standard procedures. Offspring were tail-biopsied at 10–14 days of age and genotyped using PCR with the following primer set: forward 5′-GAATGGCGCCCCCGACGAC-3′(LRP5 nucleotides 4692–4710); and, reverse 5′-GCTCCCATTCATCAGTTCCATAGG-3′ (SV40 nucleotides 4541–4564). A 524-bp transgene-specific junction fragment was detected. In each PCR reaction, a 303-bp murine c-fos product was also amplified as an internal control. The same genotyping primers were used to identify mutant G171V and wildtype LRP5 transgenic mice. Southern blotting of genomic DNA was used to check for transgene integration and integrity. All studies were performed on heterozygotes with statistical comparisons made using nontransgenic littermates. Quantitation of LRP5 expression was performed using highly sensitive real-time Taqman reverse transcriptase (RT)-PCR (ABI Prism 7700; ABI, Foster City, CA, USA) on RNA isolated from femur and tibia tissue, flushed of marrow contents. An attempt was made to preserve the periosteum by very careful dissection and removal of surrounding tissue.

Species-specific primers and probe sets were developed based on human and mouse LRP5 cDNA sequences. Reagents for human LRP5 were as follows: forward primer, 5′-CGTGATTGCCGACGATCTC-3′ (nucleotides 2605–2623); reverse primer, 5′-TTCCGGCCGCTAGTCTTGT-3′ (nucleotides 2703–2721); probe, 6-FAM-5′-CGCACCCGTTCGGTCTGACGCAGTAC-3′ (nucleotides 2635–2650). Reagents for mouse Lrp5 (GenBank AF064984) were as follows: forward primer, 5′-CTTTCCCCACGAGTATGTTGGT-3′ (nucleotides 4372–4393); reverse primer, 5′-AAGGGACCGTGCTGTGAGC-3′ (nucleotides 4396–4424); probe, 6-FAM-5′-AGCCCCTCATGTGCCTCTCAACTTCATAG-3′ (nucleotides 4434–4452). Specificity was confirmed by measurement of endogenous LRP5 in cultured human HOB-03-C5(32) and mouse MC-3T3-E1 cells. Expression of endogenous mouse type I collagen was also determined using the following reagents: forward primer, 5′-TGCTTAGTTTTGCCCGCC-3′; reverse primer, 5′-GTAATGATGAGTTAACCACGCCC-3′; probe, 6-FAM-5′-TTGCCTTGATCAGGTCCTGGAGCG-3′. RNA from mouse E12-E14 embryos was used as a control for type I collagen. Results from experimental samples were normalized relative to 18S ribosomal RNA levels. Heterozygous animals were grouped by sex and age and maintained on a 14–10 h light-dark cycle with ad libitum access to water and a standard rodent diet containing 0.9% calcium and 0.7% phosphorous. For ex vivo analysis, the mice were killed by exposure to carbon dioxide at specified ages and the femorae and tibias from both limbs, the fifth lumbar vertebrae (L5), and the calvariae were dissected free of soft tissue. Bone densities as well as histology were determined on the right femur, which was stored in 70% ethanol until analyzed. The left femur and L5 vertebrae were stored frozen until analyzed for biomechanical strength. All animal studies were performed with the approval of the Institutional Animal Care and Use Committee at Wyeth.

In vivo peripheral DXA

Areal BMD (g/cm2) of the whole body, spine, and femoral diaphysis was evaluated in vivo using peripheral DXA (pDXA; PIXImus; GE-Lunar Corp., Madison, WI, USA). Scans were begun at 3 weeks of age and repeated at 2- or 4-week intervals for 12 months. The head was excluded from total body measurements. The region of interest for the spine extended from the 6th lumbar vertebra cranially and included the thoracic vertebrae. The region of interest for the femur included only the femoral diaphysis and therefore represents cortical bone only.

Ex vivo peripheral quantitative computed tomography

Volumetric BMD (vBMD, mg/cm3) of the right femur was evaluated using peripheral quantitiative computed tomography (pQCT; XCT Research; Stratec Medizintechnik, Pforzheim, Germany). One 0.5-mm-thick pQCT slice obtained 2.5 mm proximal from the distal end of the femur was used to compute total and trabecular density for the distal femoral metaphysis. A second slice acquired 6 mm proximal from the end of the femur, which was within 0.5–1.5 mm of the diaphyseal mid-shaft, was used to assess cortical geometry and density. The tomographic slices had an in-plane pixel size of 0.07 mm. After acquisition, the images were displayed, and the region of interest including the entire femur for each scan was outlined. The soft tissue was automatically removed using an iterative algorithm, and the density of the remaining bone (total density) in the first slice was determined. The outer 55% of the bone was then peeled away in a concentric spiral, and the density of the remaining bone (trabecular density) of the first slice was reported in mg/cm3. In the second slice, the boundary between cortical and trabecular bone was determined using an iterative algorithm and the volumetric density of the cortical bone was determined.

Microcomputed tomography

High-resolution microcomputed tomography (microCT) was used to evaluate trabecular volume fraction and microarchitecture in the distal femur (μCT20; Scanco Medical AG, Basserdorf, Switzerland) and the fifth lumbar vertebrae (μCT40; Scanco Medical AG).(33) The femur was scanned at 35 kEv with a slice increment of 9 μm. CT images were reconstructed with an isotropic voxel size of 9 μm, and the gray-scale images were segmented using a constrained three-dimensional (3D) Gaussian filter (σ = 0.8, support = 1.0) to remove noise, and a fixed threshold (35% of maximal gray scale value) was used to extract the structure of mineralized tissue. Scanning was started approximately at the growth plate and extended proximally for 200 slices. Morphometric analysis was performed on 135 slices extending proximally beginning with the first slice in which the femoral condyles had fully merged. The entire fifth lumbar vertebrae was scanned at 55 kEv, with a slice increment of 12 μm, and CT images were reconstructed with an isotropic voxel size of 12 μm. The gray-scale images were segmented using a constrained 3D Gaussian filter (σ = 0.8, support = 1.0) to remove noise, and a fixed threshold (22% of maximal gray scale value) was used to extract the structure of mineralized tissue. The trabecular bone within the vertebral body (excluding regions near the endplates) was identified using manually drawn contouring algorithms on approximately 200 CT slices per vertebrae (∼2.5 mm of vertebral height). Morphometric parameters computed for both skeletal sites included the bone volume fraction (BV/TV, %), trabecular thickness (Tb.Th, μm), trabecular number (Tb.N, mm−1), trabecular separation (Tb.Sp, μm), and connectivity density (Conn.D, mm−3). Tb.Th, Tb.N, and Tb.Sp were computed using algorithms that do not rely on assumptions about the underlying trabecular structure.(33–36)

Immunohistochemistry

Calvariae were removed intact, soft tissues were gently dissected, and the bones were fixed in 70% ethanol. After fixation, calvariae were decalcified in TBD-2 decalcifying agent (Shandon, Pittsburgh, PA, USA) for 7–8 h and dehydrated in graded alcohol. Four to six 5-μm-thick representative, nonconsecutive coronal step paraffin sections were cut. Detection of LRP5 in tissue sections used a rabbit polyclonal antibody that was generated by Sigma/Genosys (St Louis, MO, USA) using the peptide RWKASKYYLDLNSDSDPY, which represents a common cytoplasmic sequence (aa 1556–1573) in wildtype LRP5 and LRP5 G171V. This human LRP5 peptide sequence used to immunize rabbits was 94.4% identical (17/18 residues) to the mouse sequence and potentially could be expected to cross-react with the endogenous mouse protein. The binding of the antibody to the epitope was visualized (5 μg/ml) using an avidin-linked peroxidase system (Vector Laboratories, Burlingame, CA, USA). Preimmune serum was used a negative control and was not immunoreactive. Avidin-peroxidase in the absence of primary antibody served as additional controls.

Histology

The activity of alkaline phosphatase (AP) was assessed with a histochemical stain using an azo-dye method (MBT-BCIP; Roche Diagnostics, Indianapolis, IN, USA) in 6-μm frozen sections of the mouse parietal bone after fixing in 4% paraformaldehyde for 15 minutes. The TUNEL (Roche Diagnostics) method was use to detect cell apoptosis in paraffin-embedded mouse parietal bone (5 μm thick) as previously described.(37) Cells undergoing apoptosis (TUNEL-positive) were identified using fluorescence microscopy. TRACP activity in osteoclasts was measured as previously described.(38)

Histomorphometry

Trabecular mineralized bone area was measured in the distal region of the mouse femur, immediately below the growth plate in Goldner's trichrome-stained 5-μm tissue sections. An analysis image window, measuring 1.67 mm2, was established for evaluation of trabecular bone area. Measurement of erosion surface (ES) and bone surface (BS) were taken as linear measurements from a region directly below the growth plate and expressed as a percentage (ES/BS, %). To determine trabecular mineral apposition rate (MAR) and bone mineralizing surface (MS/BS, %), each animal was injected intraperitoneally with 15 mg/kg calcein at 9 and 2 days before necropsy. MAR was calculated by measuring the distance between the resulting two calcein fronts in bone using fluorescence microscopy. Linear measurements of single-label surface (SLS), double-label surface (DLS), and BS were taken, and the equation DLS + (1/2 SLS)/BS × 100 was used to calculate percent MS/BS. Measurements were made on unstained 10-μm sections at 40× magnification using a 0.03-mm2 image window and covering an area of approximately 1.67 mm2. All measurements were made using the Bioquant Image Analysis System (R&M Biometrics, Inc., Nashville, TN, USA).

Compressive strength of vertebrae

The entire spine was dissected from T10-L6-L7 and frozen at −20°C. The L5 vertebrae were isolated, and two coplanar cuts 2.03 mm apart were made perpendicular to the cephalo-caudal axis using a Buehler diamond blade wafer saw (Buehler Ltd., Lake Bluff, IL, USA). The sample was compressed at a constant displacement rate (1.0 mm/min) parallel to the cephalo-caudal axis by an Instron 5534 loading device (Instron Corp., Canton, MA, USA). Load was measured using a 1000N load cell, and maximum load was calculated from the load-deformation curve and expressed as Newtons.

Femoral three-point bending strength

The left femur was cleaned of soft tissue, and femoral length was measured using a digital caliper. Periosteal and endocortical circumferences as well as cortical thickness were measured 6 mm from the distal end using pQCT; this measurement site was within 0.5–1.5 mm of the femoral midshaft. Load was applied at a constant displacement rate of 0.5 mm/s in the anteroposterior direction midway between two fixed supports that were 5 mm apart, on which the bone was mounted. The load was measured by a 1000N load cell (Instron 5534) mounted on the actuator. Load-displacement curves were recorded, and the breaking strength (peak load) was determined. Stress was calculated according to the published three-point bending formula using moment of inertia and radius calculated by pQCT measurements at the site of load application.(39)

Statistical analysis

Results are presented as mean ± SD or mean ± SE. Randomized block ANOVA was used to evaluate longitudinal changes in BMD measured by pDXA. Unless otherwise indicated, two-way ANOVA was performed to test for the significance of genotype and time main effects and genotype versus time interaction. Transgenic and nontransgenic values were compared at each time point using contrasts of genotype versus time least-squares means. Differences were considered significant at p < 0.05.

RESULTS

Overexpression of G171V mutation and wildtype LRP5 in bone

Transgene mRNA levels were measured relative to endogenous LRP5 transcripts expressed in a human osteoblast cell line. Analysis of mRNA by real time PCR identified four transgenic lines that expressed the LRP5 G171V mutation and three that expressed the wildtype LRP5 gene in tibia (data not shown). Mutant G171V line 19 and wildtype LRP5 line 19 were the highest expressing lines for each transgenic group (Fig. 1). The level of expression between mutant G171V line 35 and wildtype LRP5 line 19 was similar, thus allowing a comparison to be made of any phenotypic effects between these two lines. To obtain a qualitative measure of protein overexpression, immunohistochemistry using an LRP5 antibody was performed on sections taken from calvaria and long bones (Fig. 2). Mutant G171V protein expression from line 19 was evident in pre-osteoblasts and osteoblastic cells lining the periosteum, as well as in osteocytes present in mineralized bone (Figs. 2A and 2B). Periosteal osteoblasts in the mutant transgenics appeared plump and cuboidal, indicative of cells actively secreting extracellular matrix. In contrast, periosteal cells of the nontransgenic littermates appeared as flat, lining cells. There was also intense positive staining by the LRP5 antibody in osteoblasts of the primary spongiosa of the distal femur but not in the chondrocytes of the hypertrophic zone of the growth plate (Figs. 2C and 2D). Confirmation of LRP5 protein overexpression in wildtype LRP5 transgenic line 19 is shown in Figs. 2E and 2F. Staining confirmed that LRP5 protein was evident in osteocytes as well as osteoblasts.

Figure FIG. 1..

Transgene expression in mutant G171V and wildtype LRP5 mouse lines. mRNA expression in 5-week-old mouse tibias (male and female combined). Data for mutant G171V lines 19 (n = 11) and 35 (n = 10) and wildtype LRP5 line 19 (n = 8) were expressed relative to LRP5 expression measured in human HOB-03-C5 cells used as a control. A 1-fold level of transgene expression corresponds to the same level of LRP5 measured in HOB-03-C5 cells. Expression in mutant G171V line 19 was significantly greater than the other two lines (*p < 0.01). There was no significant difference between mutant G171V line 35 and wildtype LRP5 line 19. Values represent mean ± SD.

Figure FIG. 2..

Immunohistochemical staining of LRP5 protein in bone tissue from 9-week-old male mutant G171V (line 19) and wildtype LRP5 (line 19) transgenic mice. (A and B) Calvaria and (C and D) distal femur from (B and D) mutant G171V transgenic and (A and C) nontransgenic littermate control. (E and F) Calvaria from female (F) wildtype LRP5 transgenic and (E) nontransgenic littermate control. Bold arrows represent osteoblasts lining the periosteum of the calvaria, whereas open arrows indicate primary spongiosa adjacent to the growth plate. Staining was evident in both osteoblasts and osteocytes and was of greater intensity in mutant and wildtype transgenic examples compared with nontransgenic control.

Longitudinal assessment of increases in BMD in mutant G171V transgenic mice

All four expressing mutant G171V lines showed a HBM phenotype, and data are presented from two lines to show that effects were not caused by random integration of the transgene. Mutant G171V lines 19 and 35 were studied in further detail based on initial in vivo pDXA bone density analysis that revealed increased total body, spine, and femoral diaphyseal BMD as early as 3–5 weeks of age, which persisted for at least 12 months (Figs. 3A and 3B). In mutant line 19, the average increase for both genders combined for total body, spine, and femoral shaft BMD over 12 months was 28 ± 3%, 42 ± 10%, and 37 ± 8%, respectively (p < 0.01). Body weight was not altered when transgenics were compared with control littermates for both sexes and at all time points (data not shown). No other tissue abnormalities were detected in any transgenic animals after whole body necropsy examination at 17, 26, and 52 weeks of age (data not shown).

Figure FIG. 3..

Longitudinal in vivo pDXA analysis showing increased BMD in mutant G171V transgenic mice. Transgenic mice (diamonds, solid line) and nontransgenic littermates (squares, dotted line). (A) BMD values for mutant line 19 transgenics (n = 20–57) and nontransgenic littermates (n = 16–62). (B) BMD values for mutant line 35 transgenics (n = 5–54) and nontransgenic littermates (n = 4–45). Values for males and females are combined; all values represent mean ± SD (*p < 0.01).

Enhanced bone parameters in mutant G171V transgenic mice

In mutant G171V line 19 male transgenics, vBMD, as measured by pQCT, was significantly higher (p < 0.01) in both the total femoral (30–55%) and the trabecular region (103–250%) of the distal femoral metaphysis from 5 through 52 weeks of age compared with nontransgenic littermates (Figs. 4A and 4B). In contrast, cortical vBMD of the diaphysis showed a different age-related pattern. Cortical vBMD in the transgenics was significantly higher than the nontransgenics at 5 and 9 weeks of age; comparable with the nontransgenics at 17 and 26 weeks of age; and significantly lower than the nontransgenics at 52 weeks of age (Fig. 5A). Cortical thickness of the femoral diaphysis was significantly increased (p < 0.05) in the transgenics at all ages caused primarily by greater increases in periosteal circumference (Figs. 5B and 5C). However, endocortical circumference was significantly increased in the transgenics at 9, 26, and 52 weeks of age (Fig. 5D; p < 0.05). No changes were noted in femoral length between the transgenics and nontransgenics at any of the time points. Female transgenics from mutant line 19 displayed similar increases in vBMD and geometric parameters throughout the 52-week sampling interval. In comparison, the increase in vBMD for mutant line 35 was not as great, but both total femoral vBMD and trabecular vBMD were increased at 9 weeks of age (25% and 37%, respectively; p < 0.01; data not shown).

Figure FIG. 4..

Ex vivo pQCT analysis showing increased BMD in distal femoral metaphysis from mutant G171V transgenic mice. (A) Total bone density in mutant male line 19 transgenics (solid circles) and nontransgenic littermates (open circles) measured at 5, 9, 17, 26, and 52 weeks of age. (B) Trabecular bone density in same animals as A. Number of mice in each group is shown in parentheses; asterisks indicate significance (p < 0.01; ANOVA); plotted values represent mean ± SE.

Figure FIG. 5..

Age-related changes in cortical (A) vBMD (mg/cm3), (B) thickness (mm), (C) periosteal circumference (mm), and (D) endocortical circumference (mm) as determined by ex vivo pQCT analysis of a site adjacent to the mid-diaphysis in mutant G171V male transgenic mice (solid circles) and nontransgenic littermates (open circles). Number of mice in each group is shown in parenthesis; asterisks indicate significance (p < 0.05; ANOVA); values are plotted as mean ± SE.

Enhancement of trabecular bone compartment in femur and vertebra of mutant G171V transgenic mice

High-resolution microCT analysis of trabecular bone from distal femorae of mutant G171V line 19 transgenics at 5–52 weeks of age confirmed the increase in bone density shown by pQCT analysis. While the absolute values for BV/TV were greater for males than females (either transgenic or nontransgenic), the effects of the transgene within gender were comparable with trabecular BV/TV at 5 weeks of age, being 127% and 140% higher in the mutant G171V male and female mice, respectively (p < 0.01; Tables 1 and 2). BV/TV values for the transgenics after 52 weeks of age were essentially similar to those at 5 weeks, unlike those in the nontransgenics that had fallen by more than 60%. This progressive age-related decline in BV/TV, Tb.N, Tb.Th, and Conn.D, and associated increase in Tb.Sp in both male and female nontransgenic controls further accentuated the differences seen with age for these structural elements in the mutant G171V mice. The effects of the transgene on femoral metaphyseal BV/TV were associated with increased Tb.N (41–53%) and increased Tb.Th (43–46%), resulting in greater Conn.D (121–344%) and decreased Tb.Sp in both males and females (p < 0.01 for all parameters). Smaller but significant increases in trabecular BV/TV, Tb.N, Tb.Th, and Conn.D were also seen in mutant G171V line 35 at 5 and 9 weeks (data not shown). MicroCT analysis of the L5 vertebrae from 17-week-old male and female mutant G171V mice from line 19 revealed a 2-fold higher vertebral trabecular BV/TV compared with their nontransgenic littermates (Table 3; Figs. 6A and 6B). This increased BV/TV was associated with increased Tb.Th (34–38%) and Tb.N (47–74%) and decreased Tb.Sp (40–48%). Connectivity density was increased in male mutant G171V mice compared with their nontransgenic littermates; however, it was decreased in female mutants compared with their nontransgenic littermates (Table 3). This decrease in Conn.D with increased BV/TV likely reflects a transition from rod- to platelike architecture, thereby resulting in decreased connectivity at the highest values of BV/TV.

Table Table 1.. Elevated Bone Parameters Determined by High Resolution MicroCT Analysis of Distal Femur From Mutant G171V Line 19 Male Transgenic Mice (HET) and Nontransgenic Littermates (NTG)
original image
Table Table 2.. Elevated Bone Parameters Determined by High Resolution MicroCT Analysis of Distal Femur From Mutant G171V Line 19 Female Transgenic Mice (HET) and Nontransgenic Littermates (NTG)
original image
Table Table 3.. Vertebral Trabecular Morphometry in 17-Week-Old Mutant G171V Line 19 Transgenic Mice (HET) and Nontransgenic Littermate Controls (NTG), Assessed by MicroCT Imaging (Mean ± SEM)
original image
Figure FIG. 6..

Three-dimensional microCT reconstruction of representative (median) L5 vertebral sections from 17-week-old female mutant G171V and wildtype LRP5 transgenics. Compared with (A) nontransgenic littermate controls, (B) mutant G171V transgenics show a 109% increase in vertebral body BV/TV. Between (C) nontransgenic littermate controls and (D) wildtype LRP5 transgenics, a more modest, but significant, increase of 25% in BV/TV is detected.

Mutant G171V expression results in increased bone strength

The striking differences in bone density and architecture of the vertebrae and the increased size of the femoral diaphysis in the mutant mice translated into increased biomechanical strength. Vertebral compressive failure load was 57–124% greater in male (Fig. 7A) and 112–154% greater in female (data not shown) mutant G171V line 19 transgenics at all time points (p < 0.01 for all). Bending strength of the femoral diaphysis was increased 61–162% in male transgenics (Fig. 7B) and 66–102% in females (data not shown; p < 0.05 for all time points). It is interesting that the femoral bending strength differential between the transgenic and nontransgenics continues to expand through 52 weeks. This does not seem to be caused by either decreases in geometric parameters or strength of the nontransgenic femur, which between the ages of 6 and 12 months seem to increase or remain constant, respectively. Although the increased structural strength of the femoral diaphysis of the transgenics seems to be largely attributable to increased bone size, when normalized for size, the pattern of the failure stress of the femur was found to vary with age in the transgenic and nontransgenics (Fig. 7C). These results suggest that greater femoral bending strength in the transgenics may be caused, in part, by increased material strength at early ages (5 and 9 weeks of age) and again at 52 weeks of age. Cortical vBMD was significantly greater (5% and 7%, respectively; p < 0.05) in the transgenics at 5 and 9 weeks of age but not at later ages. Although cross-sectional area measurements of the vertebrae were not taken to allow stress determinations, the differential in vertebral compression between the transgenics and nontransgenics remains fairly constant over the first 12 months of age. It will be instructive to explore these possible contrasting patterns in older animals because genetic studies in various strains of mice have shown that vertebral strength was not correlated consistently with femoral strength.(40) Nevertheless, the findings clearly show that the functional consequences of the G171V mutation results in increased strength of the vertebral body and the femoral shaft at every time point studied from 5 to 52 weeks of age.

Figure FIG. 7..

Enhanced biomechanical strength in bones from G171V mutant transgenic mice. Mutant line 19 male transgenic (solid circles) and nontransgenic littermates (open circles) were compared at 5, 9, 17, 26, and 52 weeks of age. Number of mice in each group is shown in parentheses; asterisks indicate significance (p < 0.05; ANOVA); plotted values represent mean ± SE. (A) Compressive strength of L5 vertebra was significantly enhanced in the transgenics at all time points. (B) Femoral diaphyses from transgenic animals displayed significantly elevated three-point bending strength. (C) Femoral three-point bending stress suggests that, at 5, 9, and 52 weeks of age, the femurs of the transgenics have greater material strength than the nontransgenics.

Mutant G171V expression enhances bone formation

Histomorphometric evaluation of the distal femoral metaphysis from mutant G171V line 19 male transgenics at 26 weeks showed a 260% increase (p < 0.01) in bone mineral area consistent with the changes seen by volumetric density analysis (Figs. 8A, 8B, and 8E). Bone mineral area changes in both males and females at all time points were consistent with density measurements (data not shown). Fluorescent micrographs of calcein-labeled bone revealed a 33% increase in actively mineralizing bone surface in the transgenic animals (Figs. 8C and 8D) and a slight but not significant increase in mineral apposition rate. The increase in mineralizing surface (Fig. 8F) is consistent with the significant increases in osteoblast surface seen at 5 and 26 weeks.

Figure FIG. 8..

Static and dynamic histomorphometry in distal femoral metaphysis showing increased trabecular bone fraction and mineralizing bone surface in 26-week-old male G171V mutant mice from line 19. (A and B) Goldner's trichrome staining showed increased trabecular bone area fraction in (B) mutant compared with (A) nontransgenic littermate controls. (C and D) Calcein-labeling of trabeculae at low resolution showed increased mineralizing bone surface in (D) mutants vs. (C) controls. (E) Quantitation of bone area (BA/TA, %) taken from A and B. (F) Quantitation of bone mineralizing surface (MS/BS, %) taken from C and D. Values represent mean ± SE; asterisks indicate significance (p < 0.05; Student's t-test).

Histochemical staining for AP, an osteoblast differentiation and functional marker, was elevated in the mutant calvarium, confirming the active secretory status of the cells in the transgenics compared with the controls (Figs. 9A and 9B). Furthermore, TUNEL staining revealed a reduced number of positive osteoblasts and osteocyte cells in mutant transgenic mouse calvaria, suggesting a reduction in apoptosis (Figs. 9C and 9D). There was a 60–70% reduction in TUNEL-positive-staining osteoblasts seen in calvaria from males at 9 and 17 weeks of age (Fig. 9G). Comparable reductions in osteocyte apoptosis were also seen (data not shown).

Figure FIG. 9..

Histological evaluation of 17-week-old mutant male G171V and nontransgenic mouse calvariae from line 19. (A and B) Alkaline phosphatase staining showed greater intensity in mutant (B) periosteum compared with (A) nontransgenic littermate controls. A reduction in TUNEL-positive osteoblasts and osteocytes in the (D) mutant vs. the (C) control animals indicates decreased apoptosis, which is shown quantitatively for osteoblasts in G. (E and F) TRACP staining for osteoclasts suggests no differences in the number of osteoclasts between (F) mutant and (E) control animals, which was confirmed quantitatively (H). Similar levels of (I) percent erosion surface between the mutant and control animals suggests comparable functional activity of osteoclasts. Values represent mean ± SE; asterisks indicate significance (p < 0.05; Student's t-test).

TRACP staining of osteoclasts revealed no significant differences in cell number at 5 and 17 weeks of age (Figs. 9E, 9F, and 9H). Moreover, the osteoclast erosion surface (Fig. 9I) was not different between the transgenics and nontransgenics at 17 weeks, suggesting that the phenotype was not caused by an osteoclast deficiency. These results, plus the histological data (Fig. 8), suggest that the increased bone formation was a result of increased numbers of active osteoblasts, which could in part be caused by their increased functional lifespan.

Overexpression of wildtype LRP5 results in a modest increase in bone mass

All four lines that expressed the mutant G171V protein showed an increase in bone mass. To address whether overexpression of the wildtype LRP5 protein resulted in increases in BMD, we also examined three independent transgenic lines overexpressing the human wildtype LRP5 in bone. Two of the three lines (31 and 47) showed a relatively low level of LRP5 expression in the tibia, but we could not find any evidence for the HBM phenotype at any of the time points studied (data not shown). However, in wildtype LRP5 line 19 where the level of LRP5 expression was similar to mutant G171V line 35 (Fig. 1), a bone response was indeed observed that was indicative of an anabolic effect.

pDXA analysis of male and female transgenic mice from wildtype LRP5 line 19 showed significant increases in total (5–8%) and vertebral (7–12%) BMD at 9 weeks of age (p < 0.01 for all; data not shown). High-resolution microCT analysis of trabecular bone from distal femorae of wildtype LRP5 line 19 female transgenics at 9 weeks of age revealed significant increases in several key bone parameters. Specifically, trabecular BV/TV was 35% higher in the transgenics (p < 0.01), as was Tb.N (15%) and Conn.D (47%; Table 4). In addition, vertebral trabecular BV/TV was 20–30% higher in 17-week-old wildtype LRP5 transgenics than their littermate controls (p < 0.01; Table 5; Figs. 6C and 6D). Vertebral Tb.Th was similar in wildtype LRP5 transgenics and littermate controls, but Tb.N was 20% greater and Tb.Sp was 18–22% lower in the wildtype LRP5 transgenics (p < 0.01 for both). A direct comparison between the femoral bone phenotype of LRP5 transgenics and G171V transgenic lines 19 and 35 (all at 9 weeks of age) expressed in terms of percent change from their respective nontransgenic controls is shown in Table 6. Although all of the lines had values that were statistically significantly (p < 0.01) greater than those for the nontransgenics, the magnitudes of the phenotypic changes in the LRP5 transgenics were far less than those seen with the mutant G171V transgenic lines. Thus, these results imply that regulation of bone mass was more closely tied to expression of the G171V mutation than the level of the LRP5 receptor.

Table Table 4.. Elevated Bone Parameters Determined by High Resolution MicroCT Analysis of Distal Femur From 9-Week-Old Female Wildtype LRP5 Transgenic Mice (HET) and Nontransgenic Littermates (NTG)
original image
Table Table 5.. Vertebral Trabecular Morphometry in 17-Week-Old Wildtype LRP5 Transgenic Mice (HET) and Nontransgenic Littermate Controls (NTG), Assessed by MicroCT Imaging (Mean ± SEM)
original image
Table Table 6.. Wildtype LRP5 Transgenics Show a Clear but Lower Magnitude Bone Anabolic Phenotype Than Mutant G171V Transgenics
original image

DISCUSSION

Transgenic mice expressing the human LRP5 (G171V) mutation in bone showed a remarkable high bone mass phenotype, which bore a very close resemblance to the phenotype described in the original HBM kindred.(9, 10) Several lines of evidence point to increased numbers of active osteoblasts with increased functional lifespan as contributing to the increase in bone mass. These studies validate the role of LRP5 in bone biology and provide new insights into the mechanism of LRP5 action and its effects on bone formation.

Histomorphometric evaluation of a biopsy sample from the iliac crest of an affected member of the HBM kindred revealed that BV/TV, trabecular number, and trabecular thickness were all significantly increased (p < 0.01) over the normal ranges for these values (RR Recker, ML Johnson, KM Davies, SM Recker, RP Heaney, personal communication, 2000). The similarity of the bone density and architectural changes in the affected bone compartments observed in the transgenic mice, together with the increased strength of the bone, suggested that the skeletal effects of mutant G171V transgene expression in the mouse closely mimics the human HBM phenotype.

Expression of the G171V mutation seemed to affect all skeletal sites. The magnitude of the changes in areal BMD (pDXA) in the transgenics is similar to the 37–56% increases in areal BMD observed in the vertebrae and hip of the HBM affected individuals.(9) Despite the dramatic changes in BMD, transgene expression did not affect femoral length (data not shown), in agreement with the clinically observed phenotype, wherein the external shape and dimensions of bones from affected individuals appear normal.(9) An increase in cortical thickness was also noted as a common finding in both the transgenic and human phenotype.

Recently, inactivating mutations in the LRP5 gene were correlated with the disease osteoporosis pseudoglioma,(8) and targeted disruption of the LRP5 gene in mice was found to produce a low bone mass phenotype that was attributed to decreased osteoblast proliferation and function.(23) While osteoblast proliferation was not specifically evaluated in the mutant G171V transgenic mice, it was clear that a reduction in osteoblast and osteocyte apoptosis as a result of the transgene had a role in the increased proportion of bone surfaces involved in osteogenesis. This contrasts with the LRP5 null mice where changes in apoptosis did not seem to have a role in the osteopenia that was observed. Nevertheless, it is clear that genetic alterations in the LRP5 gene seem to result in either gain or loss of function mutations that could define a continuum of skeletal phenotypes, potentially representing qualitative as well as quantitative differences in the mechanism of action of LRP5.

To avoid the complication of comparisons among different transgenic lines, introducing the HBM mutation into the endogenous Lrp5 locus of the mouse through homologous recombination could be contemplated to overcome the limitations of the transgenic approach. Such a model would directly assess the role of the LRP5 G171V mutation when expressed under the control of the endogenous gene promoter. Nevertheless, the HBM phenotype was evident in four independent transgenic mouse lines expressing the G171V mutation, thereby ruling out any integration-related effects of the transgene. Moreover, despite the similarity in expression between mutant G171V line 35 and wildtype LRP5 line 19, the magnitude of the phenotype in the wildtype line was substantially smaller than the G171V mutant line. Thus, while we cannot rule out the possibility that even greater levels of wildtype LRP5 expression would lead to a greater phenotype, the evidence suggests that the G171V mutation is largely responsible for the increase in bone mass. Although further characterization of the potential phenotypic changes induced by overexpression of the wildtype LRP5 gene will be needed, it is clear that the G171V mutation maximizes the anabolic potential of the LRP5 pathway. Together, these data reveal a novel function for LRP5 in regulating normal bone biology.

During the course of phenotype evaluation, we found that some of the skeletal parameters differed modestly, but significantly, between the nontransgenic littermate controls from the mutant G171V and the wildtype LRP5 lines (Tables 1, 2, 3, 4, 5, and 6). Because these animals were generated on the same inbred background (C57BL6/Tac), it is unlikely that these differences result from any genetic contributions. Furthermore, we were very careful in ensuring that animals were correctly genotyped, and the small SEMs associated with all of our endpoints tend to support this fact. In looking for potential environmental effects, it should be noted that the mutant G171V transgenics were raised and housed at a different geographical location from the wildtype LRP5 transgenic animals. As a result, the mice were fed different diets that were nearly identical in calcium (0.95% vs. 0.93%) and phosphate (0.67% vs. 0.74%) content but differed slightly in both total protein (23.4% vs. 13.1%) and total fat (10% vs. 13.4%) content (all values listed for mutant G171V vs. wildtype LRP5). Consequently, environmental factors may have contributed to the subtle differences observed in some of the skeletal parameters. These small differences between the controls in the different lines emphasize the importance of using littermate controls as opposed to a single C57BL6/Tac wildtype reference. Thus, because all of the transgenic mice were compared with their corresponding nontransgenic littermates, maintained under identical conditions, the comparisons made in this study are valid.

A second HBM family was recently identified with the identical G171V mutation that had been previously reported.(10, 14) Analysis of biochemical markers of bone turnover in this study revealed a significant elevation in the bone formation marker, osteocalcin, and no significant change in the bone resorption marker, NTX.(14) This would be consistent with our findings in the mouse of greater involvement of bone surfaces with active osteoblasts in the absence of significant changes in osteoclast number and function. Despite the apparent lifelong positive balance in osteogenesis in the HBM kindreds and the transgenic mice, there is only relatively modest and clearly not pathological changes in bone shape. Further studies using the mutant G171V transgenic mouse will hopefully further illuminate the dynamics involved in this unique phenotype.

As a result of recent studies showing that LRP5 and its closely related LRP family member, LRP6, are Wnt co-receptors capable of interacting with several key components of the Wnt pathway,(17–19, 22, 24) attention regarding the signaling mechanisms involved in bone regulation by LRP5 have focused on this pathway. Wnt proteins have been previously shown to specify critical events during skeletogenesis. For example, Wnt control of limb patterning has been demonstrated in both the developing mouse and chick.(41–44) In addition, Wnt14 has been implicated in initiating synovial joint formation in the chick limb bud.(45) Wnt signaling in the adult, however, is poorly understood, with the possible exception of the study of some human cancers. Primary calvarial osteoblasts from LRP5 null mice were shown to have lower Wnt responsiveness (Lef1-dependent transcription) than wildtype cells, which could be reversed by LRP5 co-transfection.(23) In studies with the G171V construct, Boyden et al.(14) found that transfected mouse fibroblast cells responded normally to Wnt activation but unlike cells transfected with the wildtype receptor, signaling was not blocked by the Wnt signaling inhibitor DKK-1. Using a series of deletion constructs, Mao et al.(24) have shown that the third and fourth β-propeller domains of LRP6 are necessary for DKK1 binding. Because the HBM G171V mutation resides in the first β-propeller of LRP5,(10) its ability to modulate DKK1 activity may not occur by simply interfering with DKK1-LRP5 binding. Rather, the first two β-propellers of LRP6 are reported to be critical for interactions with Wnt and Frizzled.(24) As a result, the LRP5 G171V mutant may affect these interactions thereby repressing the inhibitory effects of DKK1. On the other hand, no interaction was recently found between the Drosophila homologs of LRP5, Wnt, and Frizzled, thus raising the possibility that a more complicated mechanism exists.(46) Clearly much remains to be learned regarding the unique bone selectivity of the G171V mutation given the wide tissue distribution of LRP5 expression and the many gene responses regulated by the canonical Wnt pathway. The data presented here together with the recent identification of human LRP5 gain and loss-of-function kindreds(8, 10, 14) support a normal role for LRP5 and the Wnt signaling pathway in maintaining BMD throughout adult life. Thus, modulation of LRP5 and the pathways through which it signals such as the Wnt pathway suggest an exciting new route for development of novel osteogenic agents.

Acknowledgements

We thank Michael Cain, Christopher Childs, and Lisa Neuhold for plasmid constructs; John Kulik for DNA microinjections; Terri Haire, James Murray, Amy Espindle, and Glen Pedneault for animal breeding and phenotyping; Sera Mallette and Michael Rhodin for pDXA analysis; Vanessa Dell, Melissa Tarby, Paula Green, and Melissa Russ for ex vivo bone analysis; Beth Goad and Barbara Sheppard for pathology examination; Charles Richard, C Richard Lyttle, Christopher Miller, Kristina M Allen, Tim Keith, and Randall Little for critical discussions and reading of the manuscript; and Robert Recker and Mark Johnson for discussions on the HBM kindred.

All authors are employees of Wyeth Research.

Ancillary