A comprehensive in silico approach to measure spatio-temporal changes of bone tissue in mouse models of osteoporosis and osteoarthritis.

Lead Research Organisation: University of Sheffield
Department Name: Oncology and Metabolism


Musculoskeletal pathologies such as Osteoporosis (OP) and Osteoarthritis (OA) are major clinical problems that impair the quality of life of millions of people and cost over £5 billion to the NHS very year. In OP, patients have less dense bones, which are more likely to fracture compared to those of healthy patients. Patients who suffered of osteoporotic fractures lose mobility and independence and have increased mortality risk. OA affects the joints, which become stiffer and more painful with consequent reduced mobility for the patient. As there are no pharmacological treatments for OA, the only way to treat this disease is to perform an invasive surgery that replaces the degenerated joint with an implant. While there are effective treatments for OP, they are very expensive and they do not work for all the patients. Therefore, there is a need for development of pharmaceutical treatments for these musculoskeletal diseases.
Each new treatment needs to be tested in two animal species, one of which is usually a mouse, before testing the intervention in clinical trials. Most new drug treatments fail in phase I clinical trials even though they were found promising during the animal testing. This can be due to the differences in physiology between the animals and the humans, the fact that treatments are tested on animal disease models, which may not completely replicate the condition of patients and/or to the fact that in mice we usually perform simplified measurements. For example, in order to evaluate the ability of a new drug to reduce the risk of fracture in patients with OP, in mice the standard methodology suggests to measure the bone properties (density and morphology) in two small portions of a bone (e.g. tibia), while what we are really interested in increasing of bone strength. Moreover standard cross-sectional studies perform measurements on different animals assigned to treated or non-treated groups, increasing the inter-subject variability, which reduces measurement accuracy, which could hide the effect of interventions.
With standard approaches the bone strength can be measured only with with mechanical testing, which is invasive and can not be performed in vivo on the same mouse over time, leading to a large number of animal used in research and increasing measurement variability. This can be improved by using computational models. This project will focus on the enhancement of the assessment of bone and joint properties in preclinical studies, by combining longitudinal high-resolution imaging of the same mice, advanced image processing and computational modelling to non-invasively measure bone strength from the in vivo images. This improvement will also lead to a substantial reduction in the usage of mice in musculoskeletal research. We will create and validate computational models for the prediction of bone strength at each time point in mice scanned with in vivo microCT that allows for high-resolution scans of the mice tibia. We will create a service that measures automatically the bone properties in different portions of the tibia and that can be used worldwide by other researchers. Finally we will adapt our protocols to study in details bone changes in the mouse knee to evaluate the effect of OA.
This work will lead to an improvement of preclinical research for studying musculoskeletal diseases and for testing new interventions. Moreover, we will share our novel methodologies with the research community by creating a service with a web-interface that will allow to uptake these methodologies and substantially reduce the number of mice used in bone and joint research.

Technical Summary

This project aims to improve the current preclinical assessment of interventions for bone diseases (i.e. osteoporosis, OP, and osteoarthritis, OA), by developing more accurate longitudinal measurements in mice and evaluating endpoints, which are more relevant to predict the outcomes of the treatment in patients, and therefore facilitate the clinical translation.
We will develop and validate in vivo microCT based computational models for prediction of the mouse tibia strength, which is more related to risk of fracture than standard bone density measurements.
We will also develop a semi-automated web-service for allowing other researchers worldwide to perform non-invasive estimations of bone strength and spatio-temporal measurements of bone properties, which are more accurate than those performed with standard cross-sectional experimental designs, and lead therefore to a 63% reduction of the number of mice to be used. In this project we will convert the manual procedure we can use only in our laboratory into a semi-automatic approach, share the developed scripts with expert bone modellers, and offer a web-service to all researchers worldwide, who are willing to use this approach but do not have the expertise.
Finally we will extend the previously developed protocols for accurate measurements of early changes of knee subchondral bone in OA in vivo. We will evaluate the longitudinal changes induced by destabilization of the medial meniscus, the most common OA mouse model, and we will validate the outcomes of the longitudinal approach against ex vivo measurements of cartilage and bone from a standard cross-sectional experiment. We are confident that this will extend the reduction of required number of mice for projects, which aim to understand OA pathogenesis, to create new less invasive OA animal models and to develop new interventions for this disease.


10 25 50