Quantitative X-ray microradiography for high-throughput phenotyping of osteoarthritis in mice

Summary Objective To investigate and validate digital X-ray microradiography as a novel, high-throughput and cost-effective screening approach to identify abnormal joint phenotypes in mice. Method Digital X-ray microradiography was used to quantify the subchondral bone mineral content (BMC) in the medial tibial plateau. Accuracy and reproducibility of the method were determined in 22 samples from C57BL/6(B6Brd;B6Dnk;B6N-Tyrc-Brd) wild-type mice. The method was then validated in wild-type mice that had undergone surgical destabilisation of medial meniscus (DMM) and in a genetically modified mouse strain with an established increase in trabecular bone mass. Results The measurement of subchondral BMC by digital X-ray microradiography had a coefficient of variation of 3.6%. Digital X-ray microradiography was able to demonstrate significantly increased subchondral BMC in the medial tibial plateau of male mice 4 and 8 weeks after DMM surgery and in female mice 8 weeks after surgery. Furthermore, digital X-ray microradiography also detected the increase in subchondral BMC in a genetically modified mouse strain with high trabecular bone mass. Conclusion Quantitation of subchondral BMC by digital X-ray microradiography is a rapid, sensitive and cost-effective method to identify abnormal joint phenotypes in mice of both genders at several ages.


Introduction
Osteoarthritis is the commonest joint disorder; it causes inexorable joint dysfunction, pain and disability as no drugs are available to prevent or delay disease progression. Patients are asymptomatic in the early stages of osteoarthritis and develop problems only after significant cartilage erosion has occurred. The aetiology of osteoarthritis is complex and multi-factorial with major genetic and environmental contributions 1 . Its heritability is estimated to be between 40 and 65% but currently identified disease susceptibility loci account for only a fraction of this heritability 2 . Thus, there is an urgent need to advance understanding of the pathogenesis of osteoarthritis and define new molecular pathways that facilitate the development of novel treatments.
Animal models are essential for in vivo study of disease mechanisms, drug targets and treatment responses 3 . A limited number of mouse models of spontaneous cartilage disorders have been described, including the STR/ort and STR-1N mouse. These models reinforce the importance of genetics in disease susceptibility but their use is limited by variable disease penetrance 4 . Surgically induced models of osteoarthritis are also used to investigate disease pathogenesis and response to treatment. However, this approach requires significant expertise and experience, and requires large numbers of mice at considerable expense. The best characterised and most reliable surgical model requires destabilisation of the medial meniscus (DMM), resulting in a high incidence of osteoarthritis in male mice 8 weeks following surgery 5 . Nevertheless, there is now a clear need to develop robust and rapid screening approaches to identify joint abnormalities in mouse models of spontaneous osteoarthritis.
The International Knockout Mouse Consortium and Medical Research Council N-ethyl-N-nitrosourea mutagenesis programmes aim to disrupt each of the >20,000 protein-coding genes 6 . These initiatives provide a unique opportunity to identify novel genetic determinants and in vivo models of osteoarthritis and define the cellular and molecular basis of disease. Nevertheless, to capitalise on these resources and identify mice with outlier joint phenotypes, new high-throughput, sensitive, specific and cost-effective joint phenotyping methods are needed.
Established osteoarthritis is characterised by articular cartilage destruction and subchondral bone sclerosis. Articular cartilage degradation progresses from surface fibrillation to formation of fissures and erosions, and finally cartilage loss with exposure of underlying bone. These changes are associated with increased chondrocyte hypertrophy, cartilage calcification and osteophyte formation 1 . Subchondral bone distributes load-bearing compressive forces and comprises the cortical plate underlying the articular cartilage and the adjacent trabecular bone. Early osteoarthritis is characterised by thinning of the cortical plate due to accelerated remodelling, whereas subchondral sclerosis is the hallmark of established disease 7 .
The gold standard histological assessment of articular cartilage type, architecture and integrity in accordance with Osteoarthritis Research Society International (OARSI) recommendations 4 is laborious, expensive and unsuitable for high-throughput screening. Micro-computed tomography (CT) analysis of osteoarthritis in mice is also time consuming; suitable high-resolution imaging equipment is expensive and not widely available and the vast quantity of data generated requires an extensive archive 8 .
We previously established that quantitative digital X-ray microradiography is a sensitive and specific method for determining bone mineral content (BMC) 9 and also demonstrated its suitability for high-throughput phenotyping 10 . Thus, we generated reference data for BMC in C57BL/6(B6Brd;B6Dnk;B6N-Tyr c-Brd ) wild-type mice and subsequently screened 100 knockout strains from the IKMC to identify nine outlier strains with either high or low BMC 10 . Since subchondral bone sclerosis is a consistent and established feature of both human osteoarthritis and surgically induced mouse models of osteoarthritis 11,12 , analysis of subchondral BMC represents a suitable target for high-throughput joint phenotype screening approaches.
We hypothesised, therefore, that digital X-ray microradiography can be used to quantify subchondral BMC and identify osteoarthritis phenotypes in genetically modified mice. We, therefore, studied knee joints from mice that had undergone DMM surgery 5,13 and genetically modified mice with increased trabecular bone mass 14,15 .

Animals
Studies were conducted in accordance with the Animals (Scientific Procedures) Act 1986 and recommendations of the Weatherall report, and were approved by the Imperial College Ethical Review Committee. All wild-type mice used for DMM surgery were of the same C57BL/6(B6Brd;B6Dnk;B6N-Tyr c-Brd ) genetic background 10 . Additional studies were performed using TRa 0/0 and control mice originally derived in a mixed C57BL/6 and 129Sv genetic background 14,15 , and subsequently backcrossed onto C57BL/ 6N for more than 20 generations.

DMM surgery
Transection of the right medial meniscotibial ligament was performed at 10 weeks of age and sham surgery was performed on the left knee 5 . Males were sacrificed 4 weeks (n ¼ 6 per group) or 8 weeks (n ¼ 12 per group) post-surgery. Females were sacrificed 8 weeks post-surgery (n ¼ 6 per group).

Sample preparation
Lower limb samples were fixed in 10% neutral buffered formalin overnight and stored in 70% ethanol. Soft tissue was removed but integrity of the knee capsule and patella was preserved [ Fig. 1(A)].

Histological analysis
Knees were decalcified in 10% Ethylenediaminetetraacetic acid (EDTA) for 10 days and embedded in paraffin. 5 mm coronal sections were cut at 80 mm intervals at 16 levels (10 sections per level) through the entire thickness of the joint, stained with safranin O and fast green, according to OARSI recommendations 4 . Qualitative analysis was performed to confirm the expected presence of cartilage degradation that co-localised with subchondral bone thickening in the medial tibial plateau following DMM surgery 11 .

X-ray microradiography
Knee joints were held in a flexed position of 105 using synthetic rubber adhesive within a 37 mm diameter plastic ring. This resulted in consistent location of the apex of the knee joint 12 mm above the mount with the tibial growth plate in a vertical plane. The patella was used as a landmark to ensure correct orientation of the joint [ Fig. 1 Anterioreposterior projection X-ray images were recorded at 10 mm pixel resolution using a Faxitron MX-20 variable kV point projection X-ray source and digital image system (Qados, Cross Technologies plc, Berkshire, UK) operating at 26 kV for 15 s and 5Â magnification. Images were calibrated by X-raying a digital micrometer. The relative medial tibial plateau subchondral BMC was determined by comparison with standards included in each image (1 mm diameter steel, aluminium and polyester wires) [ Fig. 1(B)]. The 2368 Â 2340 16 bit DICOM images were converted to 8 bit TIFF using ImageJ (http://rsb.info.nih.gov/ij/) and the grey scale histogram was stretched from the polyester (grey level 0) to steel (grey level 255) standards. Increasing gradations of mineralization density were represented in 16 equal intervals by a pseudocolour scheme 9,10 .

Image analysis
A 1000 mm Â 250 mm area of the medial tibial plateau was

Statistics
As grey level frequency distributions were not normally distributed [ Fig. 1(D)], the median value was determined for each sample. To define the variation in median BMC within the population of 22 female, 16-week-old wild-type mice [ Fig. 1(D)] the mean and standard deviation of the median BMC values were determined. The KolmogoroveSmirnov test was used to compare grey level cumulative frequency distributions between DMM and sham-operated mice, and between TRa 0/0 and control mice. P values for the D statistic were D ! 6.01, P < 0.05; D ! 7.20, P < 0.01; and D ! 8.62, P < 0.001 9 .

Results
To determine the accuracy and reproducibility of digital X-ray microradiography, we analysed knee joints from 22 female wildtype C57BL/6(B6Brd;B6Dnk;B6N-Tyr c-Brd ) mice. The 2.0SD reference interval for relative subchondral BMC was 178e206 with a median grey level of 192 and coefficient of variation (CV) of 3.6% [ Fig. 1(D)].
To validate this novel imaging approach, DMM was selected as an established surgical model of osteoarthritis 5,13 and histology was performed according to OARSI recommendations 4 and analysed qualitatively. X-ray microradiography of joints from male mice 8 weeks following DMM surgery confirmed displacement of the medial meniscus and demonstrated increased medial tibial plateau subchondral BMC [ Fig. 1(E), (F)]. Importantly these changes colocalised with articular cartilage lesions and histological evidence of increased subchondral bone [ Fig. 1(G)]. Quantitative ROI analysis demonstrated a significant increase in medial tibial plateau subchondral BMC 8 weeks following DMM surgery compared to the contralateral sham-operated knee [ Fig. 1(H)].
To investigate sensitivity of the method, knees were analysed from males 4 weeks following DMM surgery and from females 8 weeks post-surgery [ Fig. 2(A), (B)]. Medial tibial plateau subchondral BMC was increased 4 weeks post-surgery in males (P < 0.001) and also in females 8 weeks post-surgery (P < 0.001) compared to the sham-operated knee.
To determine whether the method could be applied for rapid detection of altered subchondral BMC in genetically modified mice, we analysed knees from control and TRa 0/0 mice 15 , which we previously demonstrated exhibit increased trabecular bone mass 14 [ Fig. 2(C)]. Quantitative X-ray microradiography identified the marked increase in medial tibial plateau subchondral BMC in TRa 0/ 0 mice compared to control mice (median grey level 210 vs 192; TRa 0/0 vs control; P < 0.001).

Discussion
Osteoarthritis is a major cause of pain and disability, its pathogenesis is poorly defined, and no drugs are available that prevent the onset or progression of disease. Advancement of the understanding of disease pathogenesis has been limited by a lack of suitable mouse models, but the generation of knockout mice for all protein-coding genes 6 now provides a timely opportunity to identify new in vivo models of osteoarthritis. These models will be essential to investigate the cellular and molecular basis of osteoarthritis and identify novel signalling pathways for future therapeutic targeting.
Nevertheless, identification of knockout strains with outlier joint phenotypes requires development of novel, robust, highthroughput and cost-effective screening approaches since current methods are too labour intensive, time consuming and costly.
We have previously shown that X-ray microradiography is an ideal high-throughput bone phenotyping method for determination of BMC in knockout mice 9,10 . We, thus, investigated this approach for the quantitation of subchondral BMC, which is characteristically increased in established osteoarthritis 7 . Induction of osteoarthritis by transection of the right medial meniscotibial ligament 5,13 was used to validate the use of X-ray microradiography for identification of osteoarthritis phenotypes in mice. Establishment of a normal reference range for medial tibial plateau subchondral BMC revealed a CV of only 3.6%, which is similar to the CV of 1.7% previously determined for whole femur BMC 10 , thus demonstrating excellent sensitivity of the method.
Digital X-ray microradiography clearly demonstrated increased subchondral BMC following DMM surgery, and these changes colocalised with the characteristic histological abnormalities in mice following surgically induced osteoarthritis 5,13 . Group sizes of only six animals were sufficient to identify increased subchondral BMC 4 weeks after DMM surgery in males and 8 weeks postsurgery in females (in which provocation of osteoarthritis is less penetrant 5 ). Comparing groups of six samples also allowed the difference in subchondral BMC in TRa 0/0 mice compared to genetically identical controls to be identified. Importantly, these group sizes are commensurate with the number of samples available from knockout lines generated for the International Mouse Phenotyping Consortium 6 .
These data indicate that X-ray microradiography is a sensitive and robust technique for quantitation of subchondral BMC, but a number of technical points need to be considered. Firstly, all soft tissue should be removed whilst carefully maintaining the knee joint capsule intact. Secondly, joints should be mounted at a constant angle of flexion such that the tibial growth plate is in a vertical plane with the patella central. Thirdly, X-ray intensity and exposure should be optimised for each imaging system to ensure the full dynamic range of the detector is used. Finally, standards should be included within every image for internal calibration 9,10 .
In conclusion, these studies demonstrate that quantitation of subchondral BMC by digital X-ray microradiography is a rapid new method to identify abnormal joint phenotypes in mice. It should be used as an initial high-throughput screening tool to identify mutant mice for further investigation. Such detailed studies, however, are costly and labour intensive, requiring high-resolution micro-CT, surgical provocation and gold standard OARSI histological scoring. Thus, the role of this novel method is in the cost-effective and robust identification of mouse strains with joint abnormalities that merit detailed analysis 4,11 .
Overall, we report the novel application of digital X-ray microradiography that, in combination with complementary methods of joint analysis, will help to identify new genetic determinants and animal models for osteoarthritis.

Author contributions
GRW (graham.williams@imperial.ac.uk) and JHDB (d.bassett@ imperial.ac.uk) take responsibility for the integrity of the work as a whole from its inception to the finished article. JAW acquired, assembled, analysed and interpreted data, and drafted the article. SAM acquired and analysed data, and helped draft the article. SG acquired and analysed data, and helped draft the article. AG analysed and interpreted data, and helped draft the article. JGL interpreted data and helped draft the article. GRW conceived and designed the studies, provided statistical advice, interpreted data, and drafted and critically revised the article for important intellectual content. JHDB conceived and designed the studies, provided statistical advice, analysed and interpreted data, prepared final figures and critically revised the article. All authors approved the final article. JAW, GRW and JHDB obtained funding from Arthritis Research UK, the Wellcome Trust and Imperial College London.