SERIAL E-M AND SIMULATION STUDY OF PRESYNAPTIC INHIBITION ALONG A GROUP IA COLLATERAL IN THE SPINAL-CORD

1. A muscle spindle primary afferent (group Ia) was physiologically identified and labeled intracellularly with the use of horseradish peroxidase (HRP) in the cat lumbar spinal cord. Serial-section electron microscopy (EM) was used to examine and reconstruct an entire axon collateral and its branches within Clarke's column. In the present study the existence and location of presynaptic contacts on Ia afferent boutons along these collateral branches were determined from examination of the serial-section electron-micrographs. 2. Of 36 Ia boutons examined in serial sections along the branches of the same collateral, 3 presynaptic contacts were found. Two of these contacts were made with Ia boutons in a complex nodal region consisting of two unmyelinated side branches exhibiting a total of six Ia boutons. The other presynaptic contact was made with a Ia bouton in a nodal region consisting of two Ia boutons connected by a thin unmyelinated bridge. 3. Computer simulations, based directly on the serial-section reconstructions, were used to investigate the possible effects of these presynaptic contacts on membrane potential and on a propagating action potential along the Ia collateral. The effect of a presynaptic contact was modeled by a sustained gamma-aminobutyric acid-A (GABA(A))-activated chloride conductance. 4. The simulation results indicated that the effect of a presynaptic contact on membrane potential and action-potential amplitude is likely to extend beyond the contacted bouton to other boutons occurring along the short unmyelinated branches arising from the same node. However, despite a large reduction of the action-potential amplitude within such a nodal region, the action potential was relatively unaffected at prior and subsequent nodes along the myelinated collateral. 5. Simulations of action-potential generation, based on different densities of voltage-dependent sodium and potassium conductances and the resting membrane leak conductance, indicated that a very large sustained chloride conductance compared with known synaptically mediated GABA, conductances was required to significantly reduce the action-potential amplitude, in agreement with a recent theoretical study on presynaptic inhibition by B. Graham and S. J. Redman. The possibility that presynaptic inhibition operates via a depolarization-induced inactivation of presynaptic calcium channels and the role of GABAB receptor activation is discussed.