Lidocaine injections targeting CA3 hippocampus impair long-term spatial memory and prevent learning-induced mossy fiber remodeling

Abstract

Learning a spatial location induces remodeling of the mossy fiber terminal field (MFTF) in the CA3 subfield of the dorsal hippocampus (Ramirez-Amaya et al. (2001) J Neurosci 21:7340-7348; Holahan et al. (2006) Hippocampus 16:560-570; Rekart et al. (2007a) Learn Mem 14:416-421). These fibers appear to grow from the stratum lucidum into distal stratum oriens. Is this axonal growth dependent on “repeated and persistent” neural activity in the CA3 region during training? To address this issue, we targeted local inactivation of the MFTF region in a post-training, consolidation paradigm. Male Wistar rats, bilaterally implanted with chronic indwelling cannulae aimed at the MFTF CA3 region, were trained on a hidden platform water maze task (10 trials per day for 5 days). Immediately after the 10th trial on each training day, rats were injected with lidocaine (4% w/v; 171 mM; n=7) or phosphate-buffered saline (PBS; n=7). Behavioral measures of latency, path length, and thigmotaxis were recorded, as was directional heading. A retention test (probe trial) was given 7 days after the last training day, and brains were subsequently processed for MFTF distribution (Timm's stain) and cannula location. Lidocaine treatment was found to block the learning-associated structural remodeling of the MFTF that was reported previously and observed in the PBS-injected controls. During training, the lidocaine group showed elevated latencies and a misdirected heading to locate the platform on the first trial of each training day. On the 7-day retention probe trial, the lidocaine-injected group showed poor retention indicated by the absence of a search bias in the area where the platform had been located during training. These data suggest that the reduction of neuronal activity in the CA3 region impairs long-term storage of spatial information. As this was associated with reduced MFTF structural remodeling, it provides initial anatomical and behavioral evidence for an activity-dependent, presynaptic growth model of memory.