"Lumbar spine finite element (FE) modelling can bring important information about load distributions in tissues involved in low back pain. However, current verifications of spine models cannot validate predictions at single tissue level (Noailly, 2007). Mechanistic models of spine tissues, respecting the link between composition, structure, and biophysical properties, can ensure accurate local predictions, but their use in complete segment models is still poorly developed. Thus, this project aims to mechanistically describe and validate the behaviour of a L3-L5 lumbar spine bi-segment FE model, by focussing on intervertebral disc collagen fibres and proteoglycans. Further mechano-biological developments of the model will be also initiated. In a first step, the lumbar intervertebral disc proteoglycan matrix will be modelled as an osmo-poroviscoelastic (OPVE) material. For the annulus fibrosus, proteoglycan matrix will be coupled to a previously developed fibre-reinforced composite model. The OPVE theory will be validated with a pre-existing lumbar disc FE geometry associated to a set of model-specific static and dynamic in vitro measurements. In a second step, the OPVE theory will be used to validate both the tissue and segment mechanics predicted by a previously developed L3-L5 lumbar spine bi-segment model. At tissue level, verifications will be done by correlating predicted intervertebral disc and vertebra load distributions with load-related tissue structures (Smit, 1997), compositions (Brickley-Parson, 1984), and cell phenotypes (Bruehlmann, 2002). Finally, a proteoglycan mechano-regulation theory will be developed, based on the influence of hydrostatic stresses and principal strains on the fixed charge densities described by the OPVE model. The new theory will be assessed thanks to the large amount of reported data about disc experimental mechano-biology. Beyond this project, the evolutive model should be further coupled to a transport model."