Femke Mathot

10 General discussion and future perspectives 171 seeding on a decellularized allograft are described. Environmental signals are absent in this in vitro setting and the actual production of trophic factors was not quantified, but both gene expression profiles were altered by the extracellular matrix of the nerve allografts. The most striking suggestion described in this study, is the different effective phase of undifferentiated and differentiated MSCs. It seemed that undifferentiated MSCs needed some time to interact with the ECM to enhance expression of the appropriate genes, while differentiated MSCs directly expressed enhanced levels of genes essential for nerve regeneration. In accordance to our findings, previous studies also demonstrated that the ECM significantly enhances the neurotrophic characteristics of differentiated MSCs in particular on the short term, but these studies lack long term evaluation points or did not make direct comparisons to undifferentiated MSCs. 20, 37, 38 By all means, the enhanced short-term gene expression of differentiated MSCs and the enhanced long- term gene expression of undifferentiated MSCs perhaps suggest that a cocktail of both undifferentiated and differentiated MSCs together provide a better cover for the trophic factors needed during the entire Wallerian degeneration and axon regeneration process and deserves to be studied in future research. IN VIVO EVALUATION STRATEGIES OF FUNCTIONAL OUTCOMES Many measurements to evaluate functional outcome are possible, we however have chosen for compound muscle action potentials (CMAP), isometric tetanic force (ITF), muscle mass and cross-sectional tibial muscle area. Behavioral studies like walking track analysis were deliberately not used to evaluate nerve regeneration since they are very sensitive to several flaws. This includes dependency on walking velocity, weight shifts as a compensatory mechanism, difficulty to acquire representative paw prints and dependency on sensory feedback. 39 Likewise, CMAP and ITF outcomes are affected by electrode placement, manipulation of the nerve which can result in neuropraxia and overstimulation of the nerve and muscle leading to muscle fatigue. 40 However, muscle function was determined to represent the most relevant outcome factor for clinical practice and CMAP, ITF and muscle mass were rated to most optimally represent muscle function. Despite their flaws, we estimated CMAP and ITF to be less vulnerable than behavioral studies and our analyses resulted in significant differences with acceptable standard errors. The rationale of the non-invasive ultrasound evaluation of cross-sectional tibial muscle area is logical, but it seems to be not sensitive enough to significantly demonstrate subtle cross-sectional muscle area differences within groups in the rat model. Due to its inability to detect differences between groups, it is questionable if the ultrasound technique should be used in future research in rat-models. However, ultrasound evaluation does not require the sacrifice of extra animals, it has great intra-rater and inter-rater reliabilities, it demonstrates regeneration curves and might reveal significant differences in bigger animal/muscle models. 41-43 For future studies in larger muscle models it is therefore still recommended to use this non-invasive muscle function evaluation strategy.