Competitive, intrinsic and stochastic determinants of neuromuscular synapse elimination in transgenic YFP expressing mice

Competition at motor endplates strongly influences elimination of polyneuronal innervation in neonatal muscles but it is still not known whether intrinsic prope rties of motor neurones also determine ultimate motor unit size. We are using transgenic mice expressing fluorescent protein in motor neurones to measure and model the changes in motor unit size as neuromuscular synapses form or become eliminated. First we designed experiments to test whether motor unit size in adult lumbrical muscles becomes reduced even when all other competing motor units are removed, by partial denervation at birth (neonates anaesthetised by chilling). We found that motor unit size was not reduced when these mice reached adulthood, suggesting that competition is sufficient to account for synapse elimination during development. Next we examined whether the initial motor innervation pattern is random. Analysis of YFP -expressing motor units in lumbrical muscles of thy1.2 -YFPH mice indicated that all connections made by a motor neurone are restricted to a single lumbrical muscle with no divergence between muscles. Stochastic modeling suggested that a key indicator of randomness is initial converge of branches of the same axon on single motor endplates. To test this we examined adult muscles in the earliest stages of reinnervation (12-14 days) after nerve crush under halothane/N20 anaesthesia. We found several compelling instances of within-unit convergence on motor endplates. Thus, stochastic properties play an important role in establishing the divergent and convergent innervation pattern within muscles but competitive interactions at polyinnervated junctions are more important than intrinsic properties of neurones during synapse elimination.