Watanabe, Oue, and Miyamoto equally contributed to this work.
Identification of a quantitative trait locus for spontaneous osteoarthritis in STR/ort mice†
Article first published online: 15 JUN 2011
Copyright © 2011 Orthopaedic Research Society
Journal of Orthopaedic Research
Volume 30, Issue 1, pages 15–20, January 2012
How to Cite
Watanabe, K., Oue, Y., Miyamoto, Y., Matsuura, M., Mizuno, Y. and Ikegawa, S. (2012), Identification of a quantitative trait locus for spontaneous osteoarthritis in STR/ort mice. J. Orthop. Res., 30: 15–20. doi: 10.1002/jor.21483
- Issue published online: 16 NOV 2011
- Article first published online: 15 JUN 2011
- Manuscript Accepted: 26 MAY 2011
- Manuscript Received: 17 DEC 2010
- mouse model
Osteoarthritis (OA) is the most common joint disorder in humans. Most of the animal models of OA were developed by surgical destabilization of joints or through transgenic approaches, and information from naturally occurring models of OA is very limited. The mouse strain STR/ort is recognized as a spontaneous model of OA. This mouse is unique in that it develops late onset cartilage degeneration of the tibio-femoral joint, similar to human OA. The purpose of this study was to identify quantitative trait loci (QTL) for the OA phenotype in STR/ort. Whereas the trait had been reported to be recessive, a significant population of the F1 generation exhibited OA phenotype. Thus, backcrossed (BC) mice generated by crossing F1 male to C57BL/6N female mice were used for genetic analysis. Degeneration of articular cartilage in BC mice was evaluated by scanning electron microscopy. Linkage analysis was carried out using microsatellite markers covering the entire genome. Cartilage degeneration in STR/ort mice was a polygenic trait. A QTL for the OA phenotype was mapped to a region 20 centimorgans proximal to the centromere of chromosome 4 (LOD = 3.37, p = 0.0065). A QTL associated with the onset of cartilage degeneration in C57BL/6N mice was also identified on chromosome 5 (LOD = 3.04, p = 0.0147). These results suggest that multiple loci are involved in the OA phenotype in mice. © 2011 Orthopaedic Research Society Published by Wiley Periodicals, Inc. J Orthop Res 30:15–20, 2012
Osteoarthritis (OA) is a common disease of the elderly. It is a relatively slowly progressing disorder without a known etiology and pathogenesis. The pathophysiology of the OA joints is characterized by fissures, erosion, and loss of cartilage, as well as subchondral sclerosis and osteophyte formation. Although a number of genes have been implicated in OA by association studies in human populations, there are only a few genes that have been associated with OA in different ethnic groups,1, 2 and there are few genes that have definitive functional data to support a causative link to OA.1, 3–5 Thus, the etiology of OA remains unclear.
Animal models help elucidate the etiology and pathophysiology of OA. Models of surgically induced cartilage degeneration have been developed in rabbits, dogs, and guinea pigs, and there are other models as well.6 Recent studies using mouse reverse genetics have suggested that several genes involved in the regulation and organization of the extracellular matrix play critical roles in the pathogenesis of OA.7, 8 Transgenic mice that express a mutated form of α1(IX) collagen in chondrocytes develop an OA-like phenotype with mild chondrodysplasia.9 Ectopic expression of a dominant negative mutant of type II TGF-β receptor also results in an OA-like phenotype in mice.10 Mice deficient in Adamts5, an aggrecanase gene, are resistant to cartilage degradation in a surgically induced model of OA.11, 12 However, data from models of naturally occurring OA in mice is very limited.
The STR/ort mouse has an inherited OA phenotype in the knees.13STR/ort mice develop normally until 6 months of age when they begin to develop focal lesions in their cartilage. The late onset of the disorder suggests that it is not due to a developmental defect. Although STR/ort mice do not exhibit synovitis and have a low incidence of osteophyte formation, they possess a number of characteristics similar to human OA, such as a high degree of degeneration in the medial side of the articular cartilage, subchondral sclerosis, and an increase in glycosaminoglycan content prior to the onset of degeneration.13, 14 In the original characterization of STR/ort mice by Walton (1975), it was suggested that the OA-like phenotype developed as a recessive trait based on progeny from CBA mice.13 Recently, Jaeger et al.15 have reported an OA associated locus in STR/ort on chromosome 8. However, to date, there have been few report on detailed genetic studies of STR/ort mice.
Here, we have carried out a genome-wide scan of OA loci in the STR/ort mouse using microsatellite markers, and found a locus on a region of chromosome 4 in a dominant manner. We also identified a quantitative trait loci (QTL) associated with OA in C57BL/6N mice on chromosome 5.
MATERIALS AND METHODS
All mice were housed in plastic cages and kept in a regulated environment with a 12-h light/dark cycle. STR/ort mice were originally obtained from Imperial College London and then cleaned to a specific pathogen free (SPF) status in Japan. They were bred in SPF conditions and fed a sterilized CE-2 pellet diet (CLEA, Tokyo, Japan) ad libitum. Internally bred male STR/ort mice were crossed with female C57BL/6N mice (Charles River, Yokohama, Japan) to generate F1 offspring. Male F1 mice were crossed with female C57BL/6N to generate backcrossed (BC) offspring. All animal studies were performed with approval from the institutional ethical committees of the Teijin Institute for Biomedical Research and the National Center for Geriatrics and Gerontology.
Analysis and Evaluation of Cartilage Degeneration
To analyze the cartilage degeneration phenotype of F1 and BC mice, paraffin-embedded tissue preparations of the proximal tibial articular cartilage were prepared at 45–50 weeks of age. Serial coronal sections (25–30 sections) were prepared and stained with hematoxylin and eosin. Cartilage degeneration was detected as fissures or defects of articular cartilage. The degree of degeneration was defined as follows: (−), normal; (+), <10% defect; (++), 10–20% defect; (+++), 30–40% defect; (++++), >50% defect of the medial or lateral articular cartilage area. Both right and left cartilages were analyzed and the maximum score between the cartilages for each mouse was used as the score.
Global images of the tibial articular cartilage were used to improve the efficiency and accuracy of phenotyping. The epiphyseal part of the tibia was resected, fixed in 10% neutral buffered-formaldehyde, and analyzed by scanning electron microscopy (SEM; S-3200N, HITACHI, Tokyo, Japan) obtained under low vacuum conditions (20 kV, 10 Pa). Back-scattered electron analysis revealed a colorless area in the articular cartilage in the SEM images that was interpreted as hydroxyapatite (data not shown).
To determine the consistency of the phenotype data obtained by the two evaluation methods, serial sections were prepared through the articular cartilage in both knee joints of 100 BC mice after SEM images were obtained. We examined approximately 60 sections from both tibiae and compared them to the SEM images.
Unlabeled and fluorescence-labeled primers for microsatellite markers (mouse MapPairs) were obtained from Invitrogen (Carlsbad, CA). ABI dye-labeled MapPairs and ABI Prism Mouse Mapping Primers v.1.0, were purchased from Applied Biosystems (Foster City, CA). Genomic DNA was extracted from the mouse tail using a standard method.16 The PCR reaction for the amplification of microsatellite markers was carried out according to manufacturers' protocols. Reaction mixtures containing unlabeled primers were subjected to 2.5% agarose gels electrophoresis (Agarose2000; Invitrogen) and then DNA was visualized by ethidium bromide staining. Labeled products were analyzed using an ABI3700 analyzer (Applied Biosystems).
To map the STR/ort loci, we used microsatellite markers that discriminated between STR/ort and C57BL/6N alleles. We genotyped more than 400 commercially available markers for both alleles, because of the absence of information on the heterogeneity of the markers for STR/ort mice. For the genome-wide QTL analysis, 122 autosomal microsatellite markers were selected (Table 1). Their positions were based on the genetic map on the Whitehead Institute web site.17 The most proximal markers that we genotyped were within 12 centimorgan (cM) of the centromere; internal markers were no more than 24 cM gap, and the most distal markers were within 12 cM of the telomere. The average spacing of the markers was 9.5 cM, with the exception of markers on chromosomes 9 and 10, where the most proximal markers were 13 and 19 cM from the centromere, respectively, and chromosomes 6 and 12, where the most distal markers were 26 and 21 cM from the telomere, respectively.
|Chromosome||Marker||Genetic map position (cM)||Chromosome||Marker||Genetic map position (cM)||Chromosome||Marker||Genetic map position (cM)|
As a trait variable for QTL analysis, a cartilage degeneration score was assigned to each mouse, as described above. We performed the QTL analysis using Map Manager QTX b20,18 which generates a likelihood ratio statistic (LRS) for each marker. To determine significance levels for the LRSs, a test of 10,000 permutations of all marker genotypes was performed. Results were considered “significant” for LRSs > 11.8 (p < 0.05) and “highly significant” for LRSs > 19.3 (p < 0.001). For chromosomal regions showing significant deviation in genotype distribution, additional informative markers were selected to further define the locus by interval mapping of all the markers on the chromosome. The LRS was converted to a LOD score by dividing by 4.6.
Characterization of the Cartilage Degeneration
Cartilage degeneration in STR/ort and C57BL/6N mice was evaluated by SEM. The SEM data corroborated the observed structural changes in tibial articular cartilage (Fig. 1). The consistency between the two methods was confirmed using some BC mice (data not shown). By SEM imaging, cartilage degeneration (+ to ++++) was observed in all five STR/ort mice (50–55 weeks of age), whereas, subtle degeneration (+) was detected in one of seven C57BL/6N mice (60–65 weeks of age).
Inheritance of the OA Phenotype in STR/ort Mice
To obtain a mating strain for mapping, male STR/ort mice were crossed to inbred mouse strains, C57BL/6N, Balb/c, C3H/HeN, and ICR. C57BL/6N was selected as the mating strain because it was the best propagating partner of the STR/ort mouse in terms of litter size, and exhibited the lowest incidence of cartilage degeneration among the strains tested. Since the penetrance of the OA phenotype of STR/ort mice is higher in males than females,13, 19 male mice were used for the analysis. Forty-eight percent (32/67) male mice in F1 generation (STR/ort male X C57BL/6N female) exhibited OA phenotype, whereas, the incidence of the phenotype in parental STR/ort mice at 10 months of age was more than 90%. The BC generation, which was subjected to whole genome scanning, was generated by mating randomly selected F1 males to C57BL/6N females. Thus, the X chromosome in BC males, as well as F1 males, originated from a C57BL/6N mouse. The prevalence of the OA phenotype in BC males at 47–52 weeks of age was 46% (92/199), and the means of the degeneration scores expressed as number of “+” marks of C57BL/6N, STR/ort, and BC mice were 0.14, 3.0, and 0.93, respectively (Fig. 2).
Analysis of 199 male BC mice and 122 genome-wide microsatellite markers using the Map Manager QTX program detected a significant QTL on chromosome 4 and on chromosome 5 (D4Mit97 and D5Mit223), respectively (Table 2). The genotypes of 20 and 11 informative markers on chromosomes 4 and 5, respectively, were then analyzed. Significant loci were located at 8.5 cM (LRS = 15.5, LOD score = 3.37) on chromosome 4 and at 77.6 cM (D5Mit223) on chromosome 5 (Fig. 3). Six additional markers on chromosome 4 (D4Mit103, D4Mit235, D4Mit263, D4Mit193, D4Mit228, and D4Mit236) had LRSs, which were higher than the threshold of genome-wide significance.
|Chromosome||Position (cM)||Marker||LRS||LOD score||p-Value*||Risk allele for cartilage degeneration|
In the current study, we used SEM images to assess the OA phenotypes of BC mice. The evaluation of cartilage degeneration by SEM has been reported.20–22 SEM enables visualization of the whole articular surface of the knee joint with more comprehensive coverage for assessing external degeneration and better detection of fine structure than radiographic assessments. The technique can omit the specimen preparation steps, such as slicing and staining, and improve the throughput, suggesting that the observation by SEM is suitable and takes advantage for phenotyping in this genetic association study.
It has been reported earlier that the OA phenotype in the parental strain, STR/1N, is polygenic trait.23 In the agreement with the finding, the OA phenotype did not conform to a Mendelian pattern of inheritance, but appeared to be polygenic when crossed with C57BL/6N mice. The trait is recessive when mated with CBA mice.13 Spontaneous OA-like cartilage degeneration in aged C57BL/6J mice have been previously reported.24, 25 Although the incidence may be affected by breeding conditions, established laboratory strains, including C57BL/6 and Balb/c, frequently develop OA after the ages of 2. However, in C57BL/6N mice, which was the strain used for the backcross in the current study, only 14% of the mice exhibited degeneration of the cartilage at 13 months of age, and degeneration was less severe than in BC and parental STR/ort (Fig. 2). In humans, the estimated heritability of knee OA is relatively low as compared to other joints.26 Thus, the incidence of OA at F1 and BC generations was likely influenced by penetrance of OA related genes and the genetic background of STR/ort and C57BL/6N mice.
Somewhat surprising was the identification of a QTL on chromosome 5 as a significant region for cartilage degeneration was more potent in C57BL/6N mice than in STR/ort mice. To our knowledge, this is the first description of a locus in C57BL/6 mice that is significantly related to cartilage degeneration. The allele may be an accelerator or predictor of lesions, which in the STR/ort mouse contributes less to the phenotype. The C57BL/6N used in this study exhibited a lower incidence of OA, which suggests that the expression of the cartilage degeneration phenotype is influenced by breeding conditions or other genetic elements of the STR/ort mice. It is also possible that there is a suppressor locus of the QTL in C57BL/6N mice, and that the contribution of the QTL might be more apparent for OA phenotype as described in previous reports.24, 25
The region of mouse chromosome 4 that contained the QTL for cartilage degeneration corresponds to human chromosomal regions 6q, 8q, and 9p. These regions have not been previously reported to be susceptibility loci of OA. The chromosomal region in mice contains several genes, including Mmp16 (MT3-MMP); Gdf6, a member of TGF-β superfamily; Map3K7 (TAK1), a mediator kinase of TGF-β signaling; and Tgfbr1 (ALK5), type I receptor for TGF-β superfamily. Recently, Miyamoto et al.4 described a functional SNP in the 5′-untranslated region of GDF5, also a member of TGF-β superfamily, that is associated with human OA. Gdf6 knockout mice exhibit developmental defects in joints, ligaments, and cartilage formation that are similar, yet distinct from those observed in Gdf5 knockout mice.27 Although Gdf6 expression was decreased in the articular cartilage of STR/ort comparing to those of age-matched C57BL/6 mice analyzed by quantitative PCR (data not shown), it is not clear whether the insufficiency of Gdf6 expression may account for the dominant trait. Further study will be required for identification of the causative alteration in the region.
Jaeger et al.15 previously identified a QTL associated with articular cartilage degeneration in STR/ort mice. Although C57BL/6 was also used as the mating partner, similar to our study, the authors detected an OA-related QTL on chromosome 8. In current study eight markers were used for mapping chromosome 8, including D8Mit258, D8Mit190, and D8Mit100, which cover the previously reported QTL (21.3 cM; Table 1). However, we did not detect a significant association for these markers. Our approach used the BC animals to seek the QTL for the dominant trait, suggesting that the previously detected QTL on chromosome 8 may be a recessive trait among the polygenetic factors in STR/ort and thus was not detected in BC generation. The QTL on chromosome 4, for which seven markers exhibited significance, was not detected by Jaeger et al., whose study adopted only a maker (D4Mit101) in the corresponding region.
The results of our phenotypic characterization were similar to that of Jaeger et al., which reported that there were differences in the area of cartilage degeneration between parental STR/ort and F2 mice, that there was more severe damage at the medial plateau of the articular cartilage in STR/ort mice,14 and that degeneration was pronounced in the lateral area of the joint in F2 mice.15 In the previous work, the authors used F2 animals for phenotypic analysis and the median OA scores of STR/ort and F2 mice, which were of mixed genotype at each locus (STR/STR, STR/C57, or C57/C57), were almost identical (∼1.5). In current study, while alleles were either heterozygous (STR/C57) or homozygous (C57/C57) in BC mice, lateral degeneration was observed in some joints of BC mice. Assuming that cartilage degeneration is a complex polygenic trait in STR/ort, protective allele(s) against degeneration may also be present in both strains. However, these putative protective alleles might not be actualized by the power of detection (e.g., the population of mice and number of markers) in the studies by Jaeger et al. and us.
Notably, medial degeneration, as seen in the parental phenotype, was more persistent in the affected cartilage of mice with the STR/C57 genotype at the QTL on chromosome 4, and the genotype (STR/C57 vs. C57/C57) at the QTL (D4Mit97) was significantly associated with the regional phenotype (medial vs. lateral, p = 0.0013 by χ2-test). In contrast, the QTL at chromosome 5 was not associated with the regional phenotype. We propose that the QTL on chromosome 4 is associated with the inheritance of the cartilage degeneration phenotype in the parental STR/ort strain, and that the QTL of chromosome 5 has a protective effect that is independent of regional phenotype. Additional studies including fine mapping with more detailed phenotyping, will be required to fully characterize these loci.
We thank Ms. Naomi Matsui and Ms. Hideko Takagi (Teijin Pharma) for the excellent assistance to conduct the animal breeding and phenotyping. This work was supported by the Research Grant (17A-2) for Longevity Sciences from the Ministry of Health, Labor, and Welfare and the Research Funding for Longevity Sciences (21A-16) from National Center for Geriatrics and Gerontology (NCGG), Japan.