Gradual decorrelation of CA3 ensembles associated with contextual discrimination learning is impaired by Kv1.2 insufficiency

Abstract

The associative network of hippocampal CA3 is thought to contribute to rapid formation of contextual memory from one‐trial learning, but the network mechanisms underlying decorrelation of neuronal ensembles in CA3 is largely unknown. Kv1.2 expressions in rodent CA3 pyramidal cells (CA3‐PCs) are polarized to distal apical dendrites, and its downregulation specifically enhances dendritic responses to perforant pathway (PP) synaptic inputs. We found that haploinsufficiency of Kv1.2 (Kcna2+/−) in CA3‐PCs, but not Kv1.1 (Kcna1+/−), lowers the threshold for long‐term potentiation (LTP) at PP‐CA3 synapses, and that the Kcna2+/− mice are normal in discrimination of distinct contexts but impaired in discrimination of similar but slightly distinct contexts. We further examined the neuronal ensembles in CA3 and dentate gyrus (DG), which represent the two similar contexts using in situ hybridization of immediate early genes, Homer1a and Arc. The size and overlap of CA3 ensembles activated by the first visit to the similar contexts were not different between wild type and Kcna2+/− mice, but these ensemble parameters diverged over training days between genotypes, suggesting that abnormal plastic changes at PP‐CA3 synapses of Kcna2+/− mice is responsible for the impaired pattern separation. Unlike CA3, DG ensembles were not different between two genotype mice. The DG ensembles were already separated on the first day, and their overlap did not further evolve. Eventually, the Kcna2+/− mice exhibited larger CA3 ensemble size and overlap upon retrieval of two contexts, compared to wild type or Kcna1+/− mice. These results suggest that sparse LTP at PP‐CA3 synapse probably supervised by mossy fiber inputs is essential for gradual decorrelation of CA3 ensembles.

Measuring Pattern Separation in Hippocampus by in Situ Hybridization

“…However, DG-GCs express Arc-mRNA until 8 hr after the neuronal activity (Ramirez-Amaya, Angulo-Perkins, Chawla, Barnes, & Rosi, 2013). Because Arc mRNA migrates to the cytoplasm and accumulates, some GCs expressing H1a simultaneously express intranuclear and cytoplasmic Arc (intranuclear: white, cytosolic: cyan; Figure 2Bc and 2Bd), consistent with previous results (Eom et al, 2022). Thus, it is necessary to distinguish cells with intranuclear Arc from cells with cytoplasmic Arc.…”

“…The term "pattern separation" has been used interchangeably to describe computational processes, changes in cell ensemble activity, and even behavioral phenomena among researchers (Eom et al, 2022;GoodSmith et al, 2019;Nakashiba et al, 2012). When animals perform a task that requires them to cognitively separate similar contexts and show different responses, pattern separation at the ensemble level is a potential mechanism driving the different behavioral responses.…”

“…Cut off the olfactory lobe with microforceps. Afterward, cut the cranial nerve by sweeping the bottom of the skull with the tip of microspatula, and place the brain on the filter paper (Eom et al, 2019;Eom et al, 2022).…”

“…The frozen tissues may be stored up to 3 months at −70°C before further processing (Eom et al, 2022).…”

“…Neuronal IEGs, such as Arc and Homer1a (H1a), are transiently and rapidly expressed in hippocampal and neocortical neurons of rodents after the exploration of novel environments (Bottai et al, 2002;Eom et al, 2022;Maurer et al, 2017;Vazdarjanova, McNaughton, Barnes, Worley, & Guzowski, 2002). We used Arc and H1a as markers to label neural ensembles that are activated by the recent exploration of novel environments, based on our previous study (Eom et al, 2022). The riboprobes used to detect H1a or Arc mRNA each consist of a small piece of purified cDNA labeled with fluorescent dye that can be detected under a fluorescence or confocal microscope.…”