Neuromuscular junctions (NMJs) innervated by motor neurons below spinal cord injury (SCI) have been reported to remain intact despite the interruption of supraspinal pathways and the resultant loss of activity. Here we report notably heterogeneous NMJ responses to SCI that include overt synapse disassembly. Complete transection of the thoracic spinal cord of adult rats evoked massive sprouting of nerve terminals in a subset of NMJs in ankle flexors, extensor digitorum longus, and tibialis anterior. Many of these synapses were extensively disassembled 2 weeks after spinal transection but by 2 months had reestablished synaptic organization despite continuous sprouting of their nerve terminals. In contrast, uniform and persistent loss of acetylcholine receptors (AChRs) was evident in another subset of NMJs in the same flexors, which apparently lacked terminal sprouting and largely maintained terminal arbors. Other synapses in the flexors, and almost all the synapses in the ankle extensors, medial gastrocnemius, and soleus, remained intact, with little pre- or postsynaptic alteration. Additional deafferentation of the transected animals did not alter the incidence or regional distribution of either type of the unstable synapses, whereas cycling exercise diminished their incidence. The muscle- and synapse-specific responses of NMJs therefore reflected differential sensitivity of the NMJs to inactivity rather than to differences in residual activity. These observations demonstrate the existence of multiple subpopulations of NMJs that differ distinctly in pre- and postsynaptic vulnerability to the loss of activity and highlight the anatomical instability of NMJs caudal to SCI, which may influence motor deficit and recovery after SCI.

Motor nerve terminals in adult mammalian slow-twitch muscles have lower levels of spontaneous and evoked neurotransmitter release than terminals in fast-twitch muscles. These reflect adaptive differences, allowing terminals in slow (postural) muscles to sustain release during the prolonged firing trains experienced in vivo. Here we ask whether these differences in terminal release properties in Sprague-Dawley rat extensor digitorum longus (EDL, fast) and soleus (slow) muscles reflect their early cytodifferentiation in the embryo or whether they might be adaptations to their distinct mature activity patterns, which emerge around two weeks postnatally. We find that the mature pattern of differences in release arise through co-ordinated increases in presynaptically dependent release properties (quantal content, spontaneous release frequency and evoked potential amplitude), beginning at three weeks, which are particularly substantial in EDL. In contrast, other synaptic properties are either consistently greater in the same muscle throughout development (evoked potential kinetics, muscle fibre diameter) or display no systematic muscle type-dependent differences (terminal area, input resistance, spontaneous release amplitude). Thus, the emergence of adaptive differences in terminal release properties correlates with the differentiation of locomotor activity patterns in postnatal rat hindlimb muscles.