Functional description of metaplasticity at excitatory synapses

Long-term potentiation (LTP) of excitatory inputs is considered a neural mechanism of learning. LTP is a positive feed-back process and may saturate its effects, resulting in the occlusion of further modulation of synaptic function and, consequently, of learning. Hence there is a need for the homeostatic adjustment of the threshold for LTP induction. One described mechanism is “metaplasticity”: for example, a priming period of high frequency postsynaptic action potential activity (pAP-prime) before LTP induction reduces LTP expression, as predicted by the Bienenstock Cooper Munro theory of 1982. However, later experiments designed at testing the effect of pAP-prime have given conflicting results, with LTP being inhibited, facilitated or both. The dependance itself of metaplasticity on the presence of action potentials has been questioned. Moreover, a detailed description of the functional conditions in which metaplasticity may extert its homeostatic control over LTP is missing. We are aiming to shed light on metaplasticity functionality by testing different pAP-prime frequencies (30, 60 and 90Hz,) and different time delays between pAP-prime and LTP induction (0.5, 2 and 10min) in whole-cell patch clamp recordings of EPSPs evoked in CA1 hippocampal neurons of acute rat brain slices after Schaffer collateral electrical stimulation. On average, LTP expression is suppressed only when induced 2 or 10min after a 60Hz pAP-prime (EPSP amplitude measured 30min after LTP induction, relative to baseline: 2min delay, 1.1±0.13, n=8; 10min delay, 1.1±0.14, n=9; LTP only: 1.7±0.08, n=8; p = 0.003 and 0.004, respectively), but not in the other conditions tested so far (60Hz, 0.5min delay: 1.4±021, n=6, p=0.3; 90Hz, 2 and 10min delays: 1.4±0.23, n=9, p=0.3; 1.5±0.21, n=3, p=0.4; respectively). However, in all primed conditions a relevant percentage of tested neurons show no LTP (range 33-87%, cutoff at 1.3 relative to baseline), a data to be confronted with LTP-only experiments (0%). Conclusions: i) postsynaptic firing activity does exert a homeostatic control of future LTP; ii) this effect appears to be modulated by the frequency of firing and by the time delay with the LTP induction; iii) CA1 hippocampal neurons might be a non-homogeneous group of neurons as far as it concerns their sensitivity to metaplasticity.