Short-term synaptic depression in the neonatal mouse spinal cord: effects of calcium and temperature

We have studied short-term synaptic depression of excitatory postsynaptic potentials (EPSPs) in lumbosacral motoneurons in the isolated, in vitro spinal cord of neonatal mice at 2-4 days postnatal age. We used 2-amino-5-phosphonovaleric acid (AP5; 100 microM) to suppress spontaneous and stimulus-evoked polysynaptic activity. Monosynaptic EPSPs were generated by trains of 10 pulses stimuli delivered to a dorsal root at eight frequencies between 0.125 and 16 Hz. The amplitudes of the second (R2), third (R3), and the average of R8, R9, and R10 (tail) EPSPs, normalized by the first EPSP (R1), defined the shapes of synaptic depression curves. Tail responses were increasingly depressed as stimulation frequency increased but R2 and R3 exhibited relative facilitation at frequencies >1 Hz. Control experiments indicated that the depression curves were not explained by presynaptic activation failure. Lowering external Ca(2+) concentration ([Ca(2+)](o)) from 2.0 to 0.8 mM without changing [Mg(2+)](o) reduced average R1 amplitudes and R2 depression with little change in tail depression. Conversely, increasing [Ca(2+)](o) to 4.0 mM increased average R1 amplitude and R2 depression but again did not change tail depression. Increasing the bath temperature from 24 to 32 degrees C produced little change in R1 amplitudes but markedly reduced the depression of all responses at most frequencies. We developed an empirical model, based on mechanisms described in more accessible synaptic systems, that assumes: transmitter is released from a constant fraction, f, of release-ready elements in two presynaptic compartments (N and S) that are subsequently renewed by independent processes with exponential time constants (tau(N) and tau(S)); an activation-dependent facilitation of transmitter release with constant increment and fast exponential decay; and a more slowly decaying, activation-dependent augmentation of the rate of renewal (tau(N)) of N. The model gave satisfactory fits to data from all [Ca(2+)](o) conditions and implied that f and the increments of the facilitation and augmentation processes were all changed in the same direction as [Ca(2+)](o), without changing the time constants. In contrast, model fits to the 32 degrees C data implied that the process time constants all decreased by 40-45% while the presumably Ca(2+)-related weighting factors were unchanged. The model also successfully matched the normalized amplitudes of EPSPs during trains with irregular intervals.