SEARCH

SEARCH BY CITATION

Keywords:

  • OSTEOBLAST;
  • OSTEOCLAST;
  • MINERALIZATION;
  • COLLAGEN;
  • BONE MINERAL DENSITY

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

By using a genome-wide N-ethyl-N-nitrosourea (ENU)-induced dominant mutagenesis screen in mice, a founder with low bone mineral density (BMD) was identified. Mapping and sequencing revealed a T to C transition in a splice donor of the collagen alpha1 type I (Col1a1) gene, resulting in the skipping of exon 9 and a predicted 18-amino acid deletion within the N-terminal region of the triple helical domain of Col1a1. Col1a1Jrt/+ mice were smaller in size, had lower BMD associated with decreased bone volume/tissue volume (BV/TV) and reduced trabecular number, and furthermore exhibited mechanically weak, brittle, fracture-prone bones, a hallmark of osteogenesis imperfecta (OI). Several markers of osteoblast differentiation were upregulated in mutant bone, and histomorphometry showed that the proportion of trabecular bone surfaces covered by activated osteoblasts (Ob.S/BS and N.Ob/BS) was elevated, but bone surfaces undergoing resorption (Oc.S/BS and N.Oc/BS) were not. The number of bone marrow stromal osteoprogenitors (CFU-ALP) was unaffected, but mineralization was decreased in cultures from young Col1a1Jrt/+ versus +/+ mice. Total collagen and type I collagen content of matrices deposited by Col1a1Jrt/+ dermal fibroblasts in culture was ∼40% and 30%, respectively, that of +/+ cells, suggesting that mutant collagen chains exerted a dominant negative effect on type I collagen biosynthesis. Mutant collagen fibrils were also markedly smaller in diameter than +/+ fibrils in bone, tendon, and extracellular matrices deposited by dermal fibroblasts in vitro. Col1a1Jrt/+ mice also exhibited traits associated with Ehlers-Danlos syndrome (EDS): Their skin had reduced tensile properties, tail tendon appeared more frayed, and a third of the young adult mice had noticeable curvature of the spine. Col1a1Jrt/+ is the first reported model of combined OI/EDS and will be useful for exploring aspects of OI and EDS pathophysiology and treatment. © 2014 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Osteogenesis imperfecta (OI), often referred to as brittle bone disease, affects approximately 1 in 20,000 births and is caused by mutations affecting any of the following processes in collagen type I biosynthesis: post-translational modification, folding, trafficking, secretion, or fibril formation. In particular, dominant mutations in the COL1A1 or COL1A2 genes have been associated with several classes of the disease ranging from mild type I, moderate and severe types IV and III to lethal type II OI.[1] Recessive mutations in genes encoding cartilage associated protein (CRTAP),[2] leprecan (LEPRE1),[3] peptidylprolyl isomerase B (PPIB),[4] FK506 binding protein 10 (FKBP10),[5, 6] and heat shock protein 47 (HSP47)[7] cause additional OI types. The connective tissue disorder Ehlers-Danlos syndrome (EDS) occurs at a higher frequency, about 1 in 5000 births,[8] and has a much broader etiology and pathology. In addition to COL1A1 and COL1A2 mutations, which account for a small fraction of EDS cases, mutations in genes for procollagen I N-proteinase (ADAMTS2), type III collagen (COL3A1),[9, 10] pro-alpha1(V) and pro-alpha2(V) chains of type V collagen (COL5A1, COL5A2), prolyl hydroxylase (PLOD1), and tenascin X (TNXB)[11-26] have also been implicated in the disease.

A subset of cases in which OI and EDS overlap has been described and is known as arthrochalasia (formerly known as EDS VII). Individuals with this type of EDS have hypermobility of the joints, hyperextensibility of the skin, as well as the OI clinical features of osteopenia. EDS VIIA and VIIB are caused by specific mutations in the COL1A1 and COL1A2 genes, respectively, the same genes affected in dominant forms of OI. In arthrochalasic patients, mutations in exon 6 of COL1A1 or COL1A2 have been uncovered. Because this exon encodes the recognition site of procollagen N-proteinase, its loss leads to retention of the N-propeptide and thereby abrogation of collagen fibril self-assembly. Not surprisingly, inactivating mutation(s) in ADAMTS2, the gene encoding the procollagen I N-proteinase, cause dermatosparaxis-type EDS in both humans and several animal models and is characterized by extremely fragile skin.[27, 28] Whereas the EDS features of patients with arthrochalasia are more prominent than their bone symptoms, a distinct group of COL1A1 and COL1A2 mutant probands called combined OI/EDS have been more recently described; these patients were identified primarily because of their bone phenotype but also have some skin and joint traits of EDS as well as early and rapidly progressive scoliosis.[29] Exhibiting severe osteopenia, multiple fractures, short stature, and blue schlera, these patients fall into the moderate to severe types IV and III OI classifications. In this group of patients, mutations are not found in exon 6 of the collagen genes, as in arthrochalasia, but rather in exons just downstream that encode the first 90 amino acids of the triple helical domain. To date, although more than 800 independent mutations have been identified in human OI[30, 31] and more than 200 in vascular type of EDS alone,[10] there exist only a few mouse models for OI[32, 33] and even fewer for EDS.[26, 34-38]

During an ENU-mutagenesis screen in mice, we identified a dominant mutation in the collagen alpha1 type I (Col1a1) gene that resulted in mice manifesting dual features of type IV OI and EDS. This new mouse model for combined OI/EDS will be a valuable tool for studying the pathophysiology of both OI and EDS and testing potential treatments for these connective tissue diseases.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Mice and ENU mutagenesis

C57BL/6J (B6), C3H/HeJ, and FVB males were purchased from Jackson Laboratory (Bar Harbor, ME, USA) at 6 to 8 weeks of age. B6 males received three intraperitoneal injections of ENU, 1 week apart, at a dose of 85 mg/kg as described previously.[39] ENU mutagenized mice were bred to C3H/HeJ females, and the first generation offspring were subjected to physiological screens, including those for bone traits. Affected female founder number 161-7-21, which was identified with low bone mineral density (BMD), was backcrossed to C3H/HeJ for three further generations. The second generation mice were tested for heritability and used for genetic mapping. Females carrying the mutation (Col1a1Jrt/+) from the fourth C3H/HeJ generation were then bred to FVB males. Mice from the second and third generation backcrosses onto FVB were used for the majority of the experiments reported here; some experiments were repeated with 10th-generation backcrosses with similar outcomes. All experimental procedures received approval from the local Animal Care Committees and were conducted in accordance with the guidelines of the Canadian Council on Animal Care.

Genetic mapping

Genomic DNA was extracted from tail tissue and subjected to genome scan using polymerase chain reaction (PCR) amplification of microsatellite markers as per standard procedures as described previously.[40]

Skeletal parameters

X-ray images were acquired from anesthetized mice on a Faxitron Specimen Radiography System Model MX-20 Digital (Faxitron X-ray Corporation, Wheeling, IL, USA). Dual energy X-ray absorptiometry (DXA; PIXImus, Lunar Corp., Madison, WI, USA) was used to measure bone mineral content (BMC), bone area, and areal BMD on the whole body (excluding the head). Scoring of animals for curvature of the spine was done from Faxitron images by a blinded reviewer.

The distal metaphysis of the left femurs and 4th lumbar vertebrae from 6 male Col1a1Jrt/+ and 6 +/+ littermates at 15 to 16 weeks of age were scanned with a Skyscan 1072 microCT (Bruker microCT, Kontich, Belgium) at the Centre for Bone and Periodontal Research (www.bone.mcgill.ca) as described.[41] Morphometric parameters were calculated with 3D Creator software supplied with the instrument. For dynamic histomorphometry, calcein-labeled femurs of 5-week-old mice were fixed in 4% paraformaldehyde, bone was infiltrated and embedded in methylmethacrylate using the osteo-bed (Polysciences, Inc., Warrington, PA, USA) at 35°C, and 5-µm coronal sections were analyzed.

Mechanical testing of bone and skin

Destructive three-point bending was performed on right femurs from 6 male Col1a1Jrt/+ and 5 +/+ littermates at 16 weeks of age using a screw-driven mechanical testing machine (Instron model 1011, Canton, MA, USA). Load was applied at a deformation rate of 1 mm/min.

Destructive tensile tests were conducted on ventral skin samples taken from 5 Col1a1Jrt/+ and 2 +/+ littermates at 20 weeks of age. Samples were mounted at two points 14 mm apart, preconditioned twice with a displacement of 3 mm, and then stretched to rupture at a deformation rate of 0.05 mm/s using a WF5 Load Cell (TestResources Inc., Shakopee, MN, USA). Load and displacement values were recorded using Testbuilder V2 software.

The maximum load (ultimate load) and maximum displacement (failure displacement) were measured from the load displacement curve, and the stiffness was determined from a linear regression of the initial region of the curve. The energy to failure was determined by measuring the area under the load deformation curve. Material properties were calculated from data normalized using femoral cross-sectional area (for bone) or sample cross-sectional area (for skin). The latter is the product of skin thickness, measured from hematoxylin and eosin-stained paraffin cross sections, and sample width (3.49 mm).

Isolation of bone marrow cells and CFU-O assay

Bone marrow cells were isolated from tibias and femora of 5- and 20-week-old age-matched sets (n = 3/genotype) of mice by centrifugation[42] and plated at a density of 1 × 106 live, nucleated cells/35-mm dish in α-MEM supplemented with 10% heat-inactivated fetal bovine serum (FBS) and an antibiotics cocktail (100 µg/mL penicillin G, 50 µg/mL gentamycin sulfate, and 0.3 µg/mL fungizone) and incubated at 37°C with 5% CO2. After 3 days, the medium was changed to differentiation medium (medium as above with 50 µg/mL ascorbic acid and 10 mM β-glycerophosphate). The medium was changed every second or third day for 19 days, cultures were fixed in 10% neutral-buffered formalin (NBF), stained for alkaline phosphatase activity and mineralization (Von Kossa),[43, 44] counted, then restained with methylene blue and total colony number quantified.

Quantitative RT-PCR

Total RNA was isolated from cortical bone and cell cultures from 5- and 20-week-old mice using TriReagent (Sigma-Aldrich, St. Louis, MO, USA) and reverse transcribed using Superscript II (Invitrogen, Carlsbad, CA, USA) and random hexamers. cDNA was combined with 0.5 µM each of the forward and reverse primers (Supplemental Table S1) and iQ SYBR Green Supermix (BioRad Laboratories, Hercules, CA, USA) in a final reaction volume of 20 µL and run in the MyIQ (BioRad Laboratories) Real-Time PCR system. Raw data were exported to PCR miner[45] for analysis and normalized using the internal control transcript for ribosomal protein L32.

Transmission electron microscopy (TEM)

Tail samples were fixed in paraformaldehyde/osmium tetroxide fixative at 4°C and decalcified in 10% EDTA/2.5% glutaraldehyde/3.2% paraformaldehyde in 0.1 M cacodylate buffer, washed in cacodylate buffer, then post-fixed in 1% osmium tetroxide for 2 hours. After dehydration in graded ethanol series and propylene oxide, samples were infiltrated with epoxy resin (Epon-Araldite) for 2 days, then embedded in resin at 60°C. Ultrathin sections of 70 to 80 nm were cut on a Reichert Ultracut E ultra-microtome, stained with uranyl acetate and lead citrate, and examined using a Hitachi H-7000 transmission electron microscope (Hitachi, Tokyo, Japan) at 75 kV; images were captured using an AMT XR60 CCD camera system (Advanced Microscopy Techniques Corp., Danvers, MA, USA).

Dermal fibroblast culture and collagen analysis

Cells were derived from skin samples of mutant and control mice as previously described[46, 47] and grown in Dulbecco's modified Eagle's medium (D-MEM) (Wisent Inc., St-Jean-Baptiste, Canada) supplemented with 20% FBS (Wisent Inc.) and 1% antibiotics/antimycotics solution (Wisent Inc.). Fibroblast cultures were reseeded into 6-well plates in above medium and at confluency supplemented with 0.25 mM magnesium salt of L-ascorbic acid 2-phosphate (Wako Pure Chemical Industries Ltd., Osaka, Japan) and 0.2 mM of β-aminopropionitrile (Sigma-Aldrich). They were maintained in this medium for 17 to 21 days, after which cells were scraped into 0.5 M acetic acid, sonicated, extracted for 2 to 4 hours at 4°C, then digested at 4°C overnight with or without pepsin (0.1 mg/mL) (Sigma-Aldrich). Cell extracts were dried, dissolved in phosphate-buffered saline (PBS) containing 2 M urea, and precipitated with 75% (v/v) ethanol at 4°C for at least 1 hour. Ethanol precipitates were collected by centrifugation at 4°C, dried, solubilized in sample buffer (10% glycerol, 0.025 M Tris pH 6.8, 1 M urea, and 1% SDS) and loaded onto NuPAGE 3% to 8% Tris-acetate gels (Invitrogen). Gels were stained in PageBlue protein staining solution (Fermentas, Burlington, Canada) and destained in water according to the manufacturer's protocol.

Histomorphometry

Mice were given intraperitoneal injections of 30 mg/kg aqueous calcein at a defined interval (1 day for young mice) before euthanization. The left femur was embedded in polymethylmethacrylate (MMA) and 5-µm sections cut on a Leica Polycut S microtome (Leica Microsystems, Richmond Hill, Canada). Fluorescence images were captured using the Bioquant Osteoimager (BIOQUANT Image Analysis Corporation, Nashville, TN, USA). Bright field images were captured using a Leitz Orthoplan microscope equipped with a Nikon Coolpix 5100 digital and analyzed using ImageJ.

Determination of C-telopeptide fragments of collagen type I a1 chains (CTX-1) in serum

Serum samples were collected from fasted Col1a1Jrt/+ and +/+ mice via the tail vein, and detection of CTX-1 fragments, a marker of bone resorption, was assayed using the Serum CrossLaps ELISA (RatLaps EIA No. AC-06F1, Immunodiagnostic Systems, Fountain Hills, AZ, USA).

Immunoblotting

Both matrix and secreted collagen were isolated from stromal cultures and culture medium. Total protein was quantified by BioRad DC Protein Assay (BioRad Laboratories), and equal amounts from each sample were separated by SDS-PAGE, electrophoretically transferred to PVDF membrane, and probed for the presence of the N-propeptide with an N-propeptide-specific antibody (LF-39) (a gift from Dr Larry Fisher).

Statistical analysis

Data are expressed as the mean ± standard deviation or standard error of the mean as indicated, followed by analysis with two-sample, unpaired t tests using GraphPad InStat 3. A p value < 0.05 was considered statistically significant.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Generation and general characteristics of the Col1a1Jrt/+ mouse line

During screening for ENU-induced dominant mutations affecting BMD, a low BMD line was identified. Many of the affected mice had difficulty walking, and Faxitron analysis confirmed a high incidence of bone fractures. Mapping and sequencing identified a T to C transition within the splice donor of exon 9 of Col1a1 (Fig. 1); sequencing of Col1a1 cDNA from homozygous mutant mice, which are not viable, revealed that the mutation resulted in the skipping of exon 9, predicting an 18-amino acid deletion in the main triple helical domain of Col1a1. No additional products generated by this mutation through the use of cryptic splice sites were seen. Col1a1Jrt/+ mice were viable, but because of a combination of poor sperm quality in the mutant males (which was strain dependent) and overall physical weakness, breeding pairs involving +/+ males and Col1a1Jrt/+ females were used to generate the mice used in the experiments reported. The mutation was inherited in a dominant manner and not lethal as Col1a1Jrt/+ mice were found at the expected Mendelian ratio.

image

Figure 1. A Col1a1 gene splice donor mutation leads to an altered sequence in the triple helical domain of Col1a1. (A) The sequencing trace with the mutation indicated. (B) Schematic of part of the unprocessed transcript with the dotted line indicating the aberrant splicing of the transcript. Part of the sequence of exon 9 is shown with the splice donor underlined. Mutation of the splice donor (underlined) results in the deletion of exon 9 and the 18 amino acids within it (boxed in the partial amino acid sequence shown). (C) Schematic diagram of Col1a1 showing the location of the mutation.

Download figure to PowerPoint

At birth, Col1a1Jrt/+ mice were grossly indistinguishable from their +/+ littermates, but by 3 weeks of age, both male and female Col1a1Jrt/+ mice were visibly smaller, leaner, and displayed thin, flaccid tails. Their smaller size, reflected in both reduced weight and percent body fat (Supplemental Fig. S1), persisted throughout life. The skin of mutant mice was more easily torn. By 8 weeks of age, some Col1a1Jrt/+ mice began to exhibit a wobbly and uneven gait that became more pronounced with age and their bones were soft and friable on cutting. Plasma biochemical data were normal for Col1a1Jrt/+ mice (Supplemental Table S2). Histological screening at necropsy indicated myocardial hypertrophy and degenerative joint disease in mice 20 to 25 weeks of age (data not shown). There was no malocclusion or overt evidence of dentinogenesis imperfecta, but although dentin is formed, it is abnormally mineralized (personal communication, Dr Marc McKee). Although decreased auditory function is another common clinical feature of some types of OI, clickbox hearing tests did not reveal any evidence for hearing loss in Col1a1Jrt/+ mice (Supplemental Material).

Because of poor breeding success of Col1a1Jrt/+ mice, mice of different ages were used for assessment of certain phenotypic traits, as described below. However, the ages used were appropriate for the overall phenotypic characterization, and care was taken to match ages for related tests. Observations on Col1a1Jrt/+ mice at various ages and over several years from both the founder mouse and through line expansion and maintenance indicated that the phenotypic traits described (bone and skin mechanics) were stable throughout the observable life span of the mice.

Skeletal characteristics

Col1a1Jrt/+ mice had lower BMD, decreased bone parameters, and mechanically weaker, more brittle fracture-prone bones

Col1a1Jrt/+ mice had lower BMD and BMC compared with age-matched +/+ littermates at all ages tested (Fig. 2A–D), consistent with the lower Tb. BV/TV found in both femoral and vertebral (lumbar vertebrae) bones (Fig. 2E). In the femoral metaphysis, this was associated with an increase in trabecular separation and decrease in trabecular number but not in trabecular thickness, whereas in the lumbar vertebrae the lower BV/TV reflected decreases in all three parameters. Furthermore, at both bone sites, the decreased structural quality of the Col1a1Jrt/+ bone was reflected in a higher structure model index (SMI) than that from +/+ bone (2.64 ± 0.20 versus 1.42 ± 0.15, p < 0.001 in the femoral trabeculae and 2.10 ± 0.088 versus 1.42 ± 0.15 in the vertebral trabeculae, p < 0.001), consistent with a more diseased rodlike morphology. Finally, although there was no significant difference in femoral cortical thickness, the cross-sectional area and the polar moment of inertia were significantly lower in Col1a1Jrt/+ mice versus +/+ littermates (Fig. 2F).

image

Figure 2. Bone characteristics of Col1a1Jrt/+. Dual energy X-ray absorptiometry (PIXImus) measurement of BMD (A) and BMC (B) of the whole body excluding the head for Col1a1Jrt/+ (black bars) mice compared with +/+ littermates (white bars). Two-dimensional cross sections of 3D reconstructions of µCT scans of the third and fourth lumbar vertebrae (C) and distal metaphyseal region of left femurs (D) from 16-week-old +/+ (left) and Col1a1Jrt/+ (right) mice. (E) Histomorphometric analysis of femoral metaphyseal and vertebral trabeculae from 16-week-old +/+ (white bars) and Col1a1Jrt/+ (black bars) mice. All values are mean ± SD of 5 animals for each category. ***p < 0.001 versus +/+ mice. (F) Histomorphometric parameters of cortical bone from 16-week-old male mice: cortical thickness (Ct.Th), femoral cross-sectional area (FCSA), and polar moment of inertia (IP). Shown are representative images for 16-week-old male Col1a1Jrt/+ and +/+ mice (n = 4 per genotype). *p < 0.05 and **p < 0.01.

Download figure to PowerPoint

Whole-mount alcian blue and alizarin red staining of the skeletons and Faxitron analysis at E18.5 and at birth (day 0) revealed no gross skeletal structural anomalies (Supplemental Fig. S2A) or detectable fractures sustained before or at birth (Supplemental Fig. S2B) in Col1a1Jrt/+ mice. By 8 weeks of age, however, multiple fractures were evident in X-rays. Quantification of fracture number and fracture site revealed age- and bone site-related variation in the percent of Col1a1Jrt/+ mice affected (Fig. 3A). For example, in young mice, the most common sites for fractures were the pelvis (Fig. 3B, C), olecranon process (Fig. 3D, E), and skull specifically in the zygomatic arch (Supplemental Fig. S3) with frequencies of 82%, 80%, and 68% of mice, respectively. The proportion of Col1a1Jrt/+ mice with fractured tarsals increased from 20% at 8 weeks to 58% at 20 weeks of age; the percent with fractured scapulae and arthritic knees also increased between 8 and 20 weeks of age (28% to 63% and 0% to 33% for scapulae and knees, respectively). Although too few mice were examined at 30 and 50 weeks of age for robust statistical analysis (n = 7 and n = 5, respectively), there was a noticeable number of older mice with fractures in the tibial and fibular bones and the number of mice affected in the various categories overall continued to increase with age (Fig. 3A).

image

Figure 3. Col1a1Jrt/+ mice have a high incidence of fractures. Location of fractures and their frequency in male Col1a1Jrt/+ mice (A). The percentage of Col1a1Jrt/+ mice scoring positive for fractures at each site was determined for 8- (n = 71), 20- (n = 24), 30- (n = 7), and 50-week-old (n = 5) mice. *Mice were scored positively in this category when the knee joint was arthritic (enlarged or bony) or deformed. (B–E) Faxitron images of 5-week-old +/+ mice (B) and (D) and Col1a1Jrt/+ mice (C) and (E) illustrate most common fractures. (C) Pelvic region of a Col1a1Jrt/+ mutant with a femoral fracture (arrow) and laterally flared ischium (arrowhead). (E) Left forelimb of a Col1a1Jrt/+ mutant shows a typical avulsion fracture of the olecranon process (arrowhead).

Download figure to PowerPoint

Consistent with the fracture data, Col1a1Jrt/+ mice showed a drastic reduction in mechanical and material properties in cortical bone, as assessed by 3-point bending of femurs (Table 1). Bones from Col1a1Jrt/+ mice were significantly weaker, less tough, less stiff, and more brittle than those from +/+ mice. Col1a1Jrt/+ femurs had a 54% decrease in ultimate load, 90% decrease in energy to fail, 34% decrease in stiffness, and decrease of 74% in failure displacement when compared with those from their sex- and age-matched +/+ littermates. When normalized to cross-sectional area, there was a 39% decrease in strength and 89% decrease in toughness and 77% decrease in failure strain but no significant change in the elastic modulus between Col1a1Jrt/+ and +/+ bones (Table 1).

Table 1. Mechanical and Material Properties of Col1a1Jrt/+ and +/+ Femurs
Bone mechanical test+/+Col1a1Jrt/+
  1. Destructive three-point bending was performed on femurs of 6 Col1a1Jrt/+ and age-matched +/+ littermates. Values shown are mean ± SD. The p values were determined using the unpaired t test. Any p > 0.05 is considered not significant (n.s.).

Non-normalized data
Ultimate load (N)25.80 ± 3.4111.90 ± 2.93 (p = 0.001)
Energy to fail (mJ)7.81 ± 2.010.79 ± 0.13 (p = 0.001)
Stiffness (N/mm)174.18 ± 32.24114.17 ± 36.69 (p = 0.019)
Length of femur (mm)16.57 ± 0.4315.16 ± 0.46 (p = 0.001)
Failure displacement (mm)0.41 ± 0.080.10 ± 0.04 (p = 0.001)
Normalized data
Ultimate stress (MPa)128.24 ± 18.0578.05 ± 29.36 (p = 0.012)
Toughness (MPa)7.72 ± 1.580.88 ± 0.23 (p = 0.001)
Modulus (MPa)4340.80 ± 949.524028 ± 1835.30 (n.s.)
Failure strain0.083 ± 0.020.019 ± 0.004 (p = 0.001)

Col1a1Jrt/+ mice display high bone turnover and aberrant mineralization

Both osteoblast surface (Ob.S/BS) and osteoblast number (N.Ob/BS) were significantly increased in Col1a1Jrt/+ compared with +/+ littermates (Fig. 4A). Although osteoclast surface (Oc.S/BS) and number (N.Oc/BS) were not significantly affected (Fig. 4B), serum from both 5-week-old and 20-week-old Col1a1Jrt/+ mice had higher concentrations of CTX-1, a collagen fragment released by bone resorption, than that from age-matched +/+ littermates (96.8 ± 27.1 ng/mL compared with 43.9 ± 17.9 ng/mL, p < 0.0001, and 12.45 ± 3.2 ng/mL compared with 8.90 ± 1.3 ng/mL, p = 0.0257, respectively, unpaired t test). In vitro, osteoclast number and size in Col1a1Jrt/+ spleen-derived cultures was normal (Supplemental Fig. S4). Calcein labels in Col1a1Jrt/+ bones, where present, were generally diffuse (Fig. 4C), suggesting impaired and disorganized mineralization, and did not allow for mineral apposition rate measurements. Most osteoblast markers were expressed at comparable levels in bones of 5- and 20-week-old Col1a1Jrt/+ and +/+ mice; however, Col1a1 and Ocn mRNA expression levels were significantly increased in cortical bone of 5-week-old Col1a1Jrt/+ versus +/+ mice (Fig. 4D). On the other hand, Alp mRNA expression levels were significantly increased in 20-week-old Col1a1Jrt/+ versus +/+ mice, although Col1a1 and Ocn were unaffected (Fig. 4E).

image

Figure 4. Cellular and molecular changes in Col1a1Jrt/+ bone. Static histomorphometric parameters of osteoblasts (A) and osteoclasts (B) from 8-week-old mice. Calcein labels in trabecular bone of 5-week-old mice (C). Expression of osteoblast-associated markers in diaphyseal bone of 5- and 20-week-old +/+ (white bars) and Col1a1Jrt/+ (black bars) mice (D, E). *p < 0.05, **p < 0.01, ***p < 0.001.

Download figure to PowerPoint

To analyze cellular mechanisms underlying these changes in expression level in more detail, and given the importance of collagen matrix deposition for osteoblast differentiation and activity,[48, 49] we next compared Col1a1Jrt/+ and +/+ osteoprogenitor frequency and differentiation capacity in vitro. Although the total number of mesenchymal progenitors (colony-forming unit fibroblast [CFU-F]) (Fig. 5A) and osteoprogenitors (CFU-ALP) (Fig. 5B) was the same, the number of mineralized osteoblast colonies (CFU-O; bone nodules) (Fig. 5C) was significantly lower in stromal cells isolated from 5-week-old but not 20-week-old Col1a1Jrt/+ versus +/+ mice.

image

Figure 5. Effect of Col1a1Jrt/+ mutation on stromal cultures. Total colonies (CFU-F) (A), alkaline phosphatase-positive colonies (CFU-ALP) (B), and the number of mineralized bone nodules (CFU-O) (C) were determined from bone marrow stromal cells isolated from 5- and 20-week-old +/+ (white bars) and Col1a1Jrt/+ (black bars) mice and grown under osteogenic differentiation conditions. The error bars represent SD of the mean. ***p < 0.001.

Download figure to PowerPoint

Col1a1Jrt/+ mice have aberrant matrix composition and fibrillogenesis

The extracellular matrices deposited in long-term cultures of Col1a1Jrt/+ dermal fibroblasts appeared more fragile than those of +/+ cells because they were easily dislodged from the plate by physical disturbance. Consistent with this observation, total collagen content of the Col1a1Jrt/+ matrices was ∼40% that of +/+ (Fig. 6A). The concentrations of types V and III collagens were not altered (Fig. 6B), but type I collagen in the Col1a1Jrt/+ matrices was reduced to only ∼30% of that in +/+ cultures, such that the relative amounts of types V and III collagen to the total were increased, whereas that of type I collagen was decreased (Fig. 6C). A faster migrating band is seen below the alpha2(I) band (Supplemental Fig. S5). Given the approximately equal amounts of wild-type and Col1a1Jrt/+ mRNAs produced by the fibroblasts (not shown), it seems likely that the mutant collagen chains exerted a dominant negative effect on the biosynthesis of type I collagen with a substantial reduction in the amount of normal protein that was available to form the heterotypic type I collagen fibrils. Electron microscopy showed that collagen fibrils in Col1a1Jrt/+ bone, tendon, and dermal fibroblast cultures had consistently smaller diameters than those in +/+ tissues (Fig. 6D–I). Other than the size difference, mutant fibrils appeared normal as did the banding pattern. However, expanded endoplasmic reticulum was clearly evident in Col1a1Jrt/+ osteoblasts, suggesting that abnormal collagen chains accumulate intracellularly (Fig. 6J).

image

Figure 6. Effect of Col1a1Jrt/+ mutation on matrix composition and fibril formation. Total collagen content (A), amount (B), and relative ratios (C) of collagen types I, III, and V (labeled I, III, and V) in matrix formed by dermal fibroblast cultures from +/+ (white bars, n = 11) and Col1a1Jrt/+ (black bars, n = 8) mice. *p < 0.05, **p < 0.01. Transmission electron micrographs of collagen fibers in bone (D, G), tendon (E, H), and dermal fibroblast cultures (F, I) of +/+ (top row) and Col1a1Jrt/+ (bottom row) samples. Black bar represents 500 nm (D, G) and 100 nm (F, I). Mean diameters of Col1a1Jrt/+ versus +/+ bone and dermal fibrils shown are 8.3 ± 0.73 versus 11.8 ± 1.2 nm; p < 0.0001 and 26.7 ± 2.2 versus 29.6 ± 3.8 nm; p = 0.0053, respectively (unpaired t test; errors represent the SD of n > 15 fibril measurements). Osteoblasts from Col1a1Jrt/+ show distended endoplasmic reticulum (*) compared with +/+ samples (J). Western blot of collagen isolated from bone marrow stromal cultures matrix (lanes 2–6, left panel) and culture medium (lanes 7–10, left panel, all of right panel) (K). Shown are the relative mobilities of the major bands of the protein standard (lane 1). The left panel was probed with antibody against mouse collagen type I, whereas the right panel is probed with antibody specific for the N-propeptide of collagen type I (LF-39). Pro, pN, pC, and α indicate the procollagen, the uncleaved pN-collagen species, the uncleaved pC-collagen species, and collagen bands, respectively. Between or among littermates, the ratio of incompletely to fully processed α-chain was greater in the Col1a1Jrt/+ than in the +/+ sample (lanes 2–3, 8.4 versus 4.7; lanes 4–6, 24 versus 7.2; lanes 7–8, 24 versus 1.7; lanes 9–10, 4.0 versus 0.26).

Download figure to PowerPoint

Because mutations in the first 90 residues of the triple helical domain have been shown, in human dermal cultures, to affect N-propeptide processing,[29] we looked for the presence of the uncleaved pN-collagen species in stromal culture matrix. In immunoblots of matrix protein probed with antibody against mouse collagen type I, Col1a1Jrt/+ matrix had proportionately greater amounts of incompletely proteolytically processed procollagen species, specifically procollagen Proα1(I) molecule and pNα1(I), relative to processed collagen than did +/+ matrix (Fig. 6K). The stronger signals of the bands corresponding to underprocessed collagen likely reflects the greater immunogenicity of the N- and C-propeptide domains.

Col1a1Jrt/+ mice have features of Ehlers-Danlos syndrome

The location of the mutation within the stretch of amino acids shown to cause a combined OI/EDS phenotype in humans led us to compare Col1a1Jrt/+ against +/+ mice for EDS features, ie, fragile skin, tendons, and kyphosis. Although there was no significant difference in thickness of the skin (Table 2), tensile tests suggested that Col1a1Jrt/+ skin had a lower failure strain than +/+ skin, was less extensible with a lower failure displacement, and required less energy to failure. Similarly, tendon teased out from the tail of Col1a1Jrt/+ mice was more frayed than that isolated from +/+ tails (Fig. 7A). These findings were consistent with the ease with which both skin and tendon were torn during normal mouse dissection. Finally, Col1a1Jrt/+ mice as young as 8 weeks of age exhibited noticeable curvature of the spine in X-ray images (Fig. 7B); 48% (n = 61) of male and 28% (n = 16) of female Col1a1Jrt/+ mice but only 1 (n = 50) +/+ mice had obvious signs of spinal curvature.

Table 2. Properties of Col1a1Jrt/+ and +/+ Skin
Parameter+/+Col1a1Jrt/+
  1. Destructive tensile tests were conducted on skin samples taken from 5 Col1a1Jrt/+ and 2 +/+ littermates (males aged 20 weeks). Shown are mechanical and material properties that were affected in Col1a1Jrt/+ skin. The latter data were normalized to the cross-sectional area of the samples where the thickness of the samples were determined using measurements taken from paraffin sections. Values shown are mean ± SD. The p values were determined using the unpaired t test. Any p > 0.05 is considered not significant (n.s.).

Thickness (mm)0.39 ± 0.0460.32 ± 0.077 (n.s.)
Failure displacement (mm)11.9 ± 3.27.1 ± 1.8 (p = 0.046)
Energy to fail (mJ)27.3 ± 7.613.9 ± 4.84 (p = 0.03)
Failure strain93.7 ± 9.053.7 ± 13.1 (p = 0.012)
image

Figure 7. Features of Ehlers-Danlos syndrome in Col1a1Jrt/+ mice. Tail tendon isolated from Col1a1Jrt/+ and +/+ mice (A). Whole-body faxitron images of Col1a1Jrt/+ and +/+ mice at 8 weeks of age (B).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

We report here a new ENU-induced mouse mutant, Col1a1Jrt/+, with dominantly inherited low BMD and other clinical features of type IV OI, including weaker and more brittle fracture-prone bones as well as EDS features of fragile skin and tendons, kyphosis, and early development of osteoarthritis (latter data not shown). Although not the focus of the current study, it should be noted that the physical weakness of the mutant mice and defective dentin may affect feeding and thus the severity of the overall phenotype, issues to be explored in detail in future. A higher osteoblast number in vivo despite similar osteoprogenitor numbers coupled with increased serum levels of a resorption marker CTX-1 indicated that the Col1a1Jrt/+ mutation causes high-turnover osteopenia. These combined features make Col1a1Jrt/+ a useful model for further exploring aspects of OI-EDS pathophysiology and treatment.

Comparison between Col1a1Jrt and other OI mouse models

Two mouse models of type IV OI, BrtlIV[50] and G610C (Amish),[51] have been reported. Both mimic corresponding human mutations involving the substitution of a glycine residue with a cysteine in the helical domain. The position of the Col1a1Jrt/+ 18-amino acid deletion is the furthest N-terminal of this group of mutations and, notably and uniquely, does not disrupt the Gly-X-Y motif.

Although all three of these OI mouse models exhibit fragile bone and low bone mass, the cellular basis of the phenotypes, including the contributions of osteoblast and osteoclast anomalies, appear different. Whereas osteoblast activity is unaffected in 2-month BrtlIV bone,[52] the N.Ob/BS and Ob.S/BS is increased in Col1a1Jrt/+ mice compared with +/+ at the same age, although mineralization appeared to be impaired. Bone resorption in both BrtlIV mice[52] and Col1a1Jrt/+ was increased but for different reasons. Osteoclast number was reported to be increased in BrtlIV mice[52] and in bone marrow-derived osteoclast cultures from an OI type III model, the oim/oim mouse,[53] yet neither N.Oc/BS nor Oc.S/BS was affected in Col1a1Jrt/+ mice. Furthermore, osteoclast number and size in Col1a1Jrt/+ spleen-derived cultures was normal (Supplemental Fig. S5), indicating a lack of intrinsic abnormalities in Col1a1Jrt/+ osteoclast precursor number or differentiation. Differences in the way the associated collagen abnormalities of these various mutations affect the bone cell populations, which may be attributable to genetic background, methodological differences, or other possibilities, deserve additional study.

Col1aJrt/+ mice exhibit age-related changes in bone and osteoblast phenotypes

Fracture rates in OI patients have been clinically observed to fall dramatically after puberty.[54] In several mouse models of OI, a similar “postpubertal adaptation” in bone such as thickening of the periosteum, improvement in bone geometry, and mechanical properties has been observed.[55-58] Col1a1Jrt/+ mice also exhibit pubertal and postpubertal bone changes, manifested as changes in mineralized colonies in culture and changes in osteoblast gene expression profiles with age. Indeed, the increased expression of the mineralization-promoting ecto-enzyme Alp[59] and decreased expression of the mineralization-inhibiting Opn[60] likely contribute to the rescue of mineralization seen in cultures from older mice. Further molecular and biochemical understanding of the switch and the nature of the age-related rescue of mineralization are needed, but taken together, our results suggest a mechanism for postpubertal adaptation involving cell autonomous age-dependent changes in osteoblast maturation, osteoblast-associated marker expression, and mineralization in Col1a1Jrt/+ mice.

Col1a1Jrt/+ mice share many similarities with human combined OI/EDS

A subset of human mutations has emerged based on a uniquely strong OI phenotype combined with some typical EDS features that distinguish it from existing classes of EDS.[29] EDS patients with deletions or substitutions of Gly residues in the first 90 amino acids of the helical domain have the EDS features of joint hypermobility and progressive scoliosis together with a moderate to severe form of OI. The Col1a1Jrt/+ mutation causes a deletion of 18 amino acids (Δ37–54) that lie within a comparable region of the mouse collagen protein corresponding to the human OI/EDS genotype. Not only do Col1a1Jrt/+ mice have an obvious OI bone phenotype, but they also have the fragile skin, tendon, and early kyphosis phenotypes displayed clinically in OI/EDS. Although joint hypermobility was not directly assessed, tendon tissues from Col1a1Jrt/+ mice, which determine joint mobility, had a visibly frayed appearance and tails detached readily from the body during mouse manipulations, suggesting that they had a low breaking strength. Although not studied here, it is worth noting that Col1a1Jrt/+ mice also showed signs of early onset osteoarthritis (data not shown).

As yet, no exact corresponding human mutation to that of the Col1a1Jrt/+ mice has been reported; however, two exon 9 splice donor mutations exist in the Human Type I Collagen Mutation Database (http://www.le.ac.uk/ge/collagen/)[10, 30] with associated clinical phenotypes varying from type I OI to type III/IV.

Collagen from OI/EDS probands have been shown to have smaller fibril diameters than that from normal patients.[29] Our TEM results on fibrils from bone, tendon, and dermal fibroblast cultures of Col1a1Jrt/+ mice are consistent with this combined OI/EDS fibril phenotype. To note, in the human study, this indirect evidence was combined with matrix deposition and differential scanning calorimetry to show that mutant collagen chains with mutations within the N-terminal region of the helical domain are secreted and that the N-propeptide remains associated with the collagen protein interfering with the processing of the N-propeptide and by extension fibril assembly.[29]

Although the incorporation of the mutant collagen molecule into fibrils may itself be the size-limiting step of fibril formation,[29] the role of other molecules in regulating fibril size may also be a contributing factor. Type I collagen fibrils are heterotypic and also contain types III and V collagen, both of which are implicated in the regulation of fibril diameter size, which itself can change as tissues develop or in response to mechanical stress.[61] In this regard, it is worth noting that at least one (albeit not all of those studied) OI-EDS proband (G88E) exhibited an increase in type III collagen content in fibroblast matrix (2% ± 3% compared with control 13% ± 2%),[62] as we report here for Col1a1Jrt/+ fibroblast matrix. Thus, given the reduction not only in total collagen deposition but also in the proportion of types I collagen relative to type III and V collagens in matrices of dermal fibroblasts from Col1a1Jrt/+ mice, it is perhaps not surprising that fibril diameter size is decreased in dermal fibroblast cultures, as well as in tendon and bone matrix.

Reduction of collagen in secreted matrix may be caused by the retention of mutant protein and/or misfolded collagen type I in the ER

The amount of type I collagen in the matrix of dermal fibroblast cultures from Col1a1Jrt/+ samples was only 30% of that found in the +/+ matrix, which is consistent with the expected frequency (25%) of type I collagen containing only wild-type α1(I) chains. This, of course, assumes that both mutant and wild-type chains are produced equally and that type I collagen containing mutant chains are retained or degraded within the cell. Immunoblot analysis of matrix from cultured stromal osteoblasts showed a greater presence of procollagen and partially processed intermediates in Col1a1Jrt/+ samples. Although the pN-collagen band was not seen in matrix proteins isolated from dermal fibroblast cultures, this might have been expected because fibroblast cultures from human OI/EDS subjects with mutations closer to the C-terminal half of the N-anchor domain (Gly34, Gly76, and Gly88) did not differ in the ratio of the pN to alpha1(I) band compared with control samples.[29] Interestingly, what appears to be a higher mobility band in the Col1a1Jrt/+ dermal matrix may be caused by the exposure of a novel proteolytic site within the mutant collagen type I similar though not necessarily identical to a truncated species of alpha2(I) found in some human OI/EDS.[29]

Mutations in the helical domain of the procollagen α1 chain result in misfolded proteins targeted for degradation, a process termed procollagen suicide[63, 64] or, more generally, ER-specific unfolded protein response,[65] a response documented in at least some mouse models of OI such as Aga2.[66] Two consequences of this are an accumulation of both misfolded protein and native/nonmutant procollagen chains during triple helix formation in the ER and an overall reduction of collagen available for extracellular fibrillogenesis. The distended ER observed in Col1a1Jrt/+ tissues analyzed by TEM and the marked reduction in both type I and total collagen in the matrices formed by Col1a1Jrt/+ dermal fibroblasts in culture are consistent with this type of process occurring. Unlike the Aga2 mouse model in which a mutation in the C-terminal end of α1(I) results in increased osteoblast apoptosis, we did not find any evidence of this or the unfolded protein response in the Col1a1Jrt/+ model (eg, BiP, GADD, Hsp47 expression are unaffected, data not shown). This is consistent with the fact that Col1a1Jrt/+ collagen retains the Gly-X-Y motif and should be able to fold properly. Thus, the distended ER observed in the Col1a1Jrt/+ dermal fibroblasts is probably a sign of a slowed folding process.

Experiments to determine the biochemical mechanism underlying the Col1a1Jrt/+ phenotype are beyond the scope of the present study because of their challenging and involved nature. Future work on this mutant, however, will include matrix deposition assays to detect whether the mutant collagen chain is retained or secreted, and if the latter, to what extent and its impact on matrix structure and function. Such studies will help to elucidate to what degree the mouse model parallels presently known cases of human OI/EDS. In summary, the Col1a1Jrt/+ model highlights again the differences in how genotype affects phenotype in collagen-associated diseases. Although many of the overt physical and skeletal characteristics of the Col1a1Jrt/+ mouse are consistent with previously published mouse models of OI, in particular those modeling type IV OI, there exist several interesting biochemical, cellular, and molecular differences in particular features of EDS. Thus, the Col1a1Jrt/+ line adds to the available murine models in which to assess not only the etiology but also genotype-phenotype-specific therapeutic options for OI-EDS.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

We thank other members of the Centre for Modeling Human Disease (http://www.cmhd.ca) for their enthusiasm, technical expertise, and help, including Celeste Owen, Lois Kelsey, Igor Vukobradovic, Geoffrey Wood, Susan Newbigging, Dean Percy, and Edward Weiss. We thank Usha Bhargava, Lee Wei, Catherine Chan, and Alex Chebotarev from the labs of Jane Aubin, Janet Henderson, William Cole, and Craig Simmons, respectively, as well as Feryal Sarraf from the Faculty of Dentistry, and Battista Calvieri and Steven Doyle of the Faculty of Medicine Microscopy Imaging Lab for technical help and discussions. We also thank Harvey Goldberg for helpful discussions during revisions.

This work was supported by funding from the Canadian Institutes of Health Research (CIHR; MOP 69198 to JEA; a Group grant to JR, JEA, and SLA; a CIHR Distinguished Scientist Award to JR; and CIHR/Osteoporosis Canada Fellowship to RAZ); Genome Canada and the Ontario Genomic Institute; and the Canadian Arthritis Network.

Authors' roles: Study design: JEA, FC, AF, WC, MG, WV, LO, LA, and JR. Data collection: FC, RG, SI, LM, ER, TZ, RAZ, and WV. Data analysis: FC, JEA, RG, LM, RAZ, WC, MG, LO, and WV. Data interpretation: FC, JEA, LM, RAZ, WC, MG, LO, and WV. Drafting manuscript: FC and JEA. Revising manuscript content: RG, SI, LM, ER, TZ, RAZ, AF, WC, MG, LO, LA, and JR. FC and JEA take responsibility for the integrity of the data analysis.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmr2177-sm-0001-SupFig-S1.tif146KSupplementary Figure S1.
jbmr2177-sm-0002-SupFig-S2.tif2687KSupplementary Figure S2.
jbmr2177-sm-0003-SupFig-S3.tif715KSupplementary Figure S3.
jbmr2177-sm-0004-SupFig-S4.tif51KSupplementary Figure S4.
jbmr2177-sm-0005-SupFig-S5.tif997KSupplementary Figure S5.
jbmr2177-sm-0006-SupInfo-S1.doc45KSupporting Information S1.
jbmr2177-sm-0007-SupInfo-S2.doc20KSupporting Information S2.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.