|Tubulin Inhibition: Evidence from Silent Neurons||Memory|
The central theme of this book is that the NR2B-containing NMDARs can be inhibited by tubulin, thus suppressing NMDA spikes and dendritic plateau potentials which are required for the generation of action potentials in spiny neurons. Long-term memories are likely stored in these neurons, with dendritic branches as the memory units (Govindarajan et al., 2011). Memory extinction occurs when the dendritic branches encoding the memory are unable to produce sufficient NMDA spikes and dendritic plateau potentials due to the inhibition of NR2B-containing NMDARs by tubulin. CRMP2, which can bind both tubulin and NR2B, is also necessary for strengthening the binding between NR2B and tubulin (Chapter 7).
Chapter 10 and Chapter 11 have explained how tubulin may inhibit NR2B-containing NMDARs in the dendritic spines and shaft. Chapter 12 presents compelling evidence that BDNF is capable of increasing NR2B-containing NMDARs, thereby promoting memory extinction. In an adult brain, the memories of numerous events have become extinct. If these extinction memories result from the suppression of neuronal firing by NMDAR inhibition, a majority of spiny neurons should be suppressed. This is indeed the case.
Prevalence of Silent Neurons
Silent neurons refer to the neurons that rarely fire. In the hippocampus, nearly two-thirds of all pyramidal neurons are behaviorally silent (Thompson and Best, 1989) and over 90% of granule cells opt for early "retirement" (Alme et al., 2010). The prevalence of silent neurons has also been observed in the neocortex where a minority of neurons are responsible for the majority of spikes (Barth and Poulet, 2012). In the medial prefrontal cortex, most neurons showed only sporadic activity (Blaeser et al., 2017).
Regarding sparse firing, the dentate gyrus (DG) of the hippocampus is probably the most unique. A singe dentate granule cell receives 3600 - 5600 inputs from the entorhinal cortex (Amaral et al., 2007). Yet the vast majority of DG neurons are silent during exploration (Diamantaki et al., 2016). Moreover, DG is located in a pathway that connects hippocampal regions with high propensity for generating seizures. This has led to the hypothesis that DG may serve as a control point for seizures in the hippocampus and that a breakdown of the dentate gate causes seizures (Lothman et al., 1992). The dentate gate hypothesis is supported by recent studies (Krook-Magnuson et al., 2015).
Dendritic Branches and Silent Neurons
In the dentate gyrus, silent neurons can be distinguished from active neurons simply by their morphology: silent cells have less dendritic branches than active cells (Figure 13-1). CRMP2 is known to play pivotal roles in dendritic branching. Binding of CRMP2 to tubulin has been shown to increase the number of dendritic branching points (Niisato et al., 2013). In another study, the hippocampal pyramidal neurons from CRMP2 knockout mice exhibited a significant decrease in both total dendritic length and dendritic arborization complexity (Zhang et al., 2016). Therefore, the correlation between silent neurons and reduced dendritic branches suggests that the CRMP2 activity could decrease in silent neurons, consistent with the hypothesis that silent neurons might result from the inhibition of NMDARs by tubulin and CRMP2. Only after the tubulin/CRMP2 complex dissociates from NMDARs, can the silent neurons be activated. The dissociated tubulin/CRMP2 complexes may further promote dendritic branching. Hence, active neurons are linked to the NMDARs with frequently dissociated tubulin/CRMP2 complexes.
Activation of Silent Neurons
Figure 13-2 shows the firing characteristics of silent neurons. These data were obtained from patients who were implanted with depth electrodes for clinical reasons. The implanted microelectrode allows for single-cell recordings. In the experiments, patients were presented with pictures according to their preferences and background. The microelectrode was implanted into the patient's medial temporal lobe which includes the hippocampus, amygdala, and entorhinal, parahippocampal and perirhinal cortices. In each trial, the recording lasts for about 1500 milliseconds. The recorded action potentials (spikes) are represented as dots in this figure. We see that the neurons in Cluster 3 respond selectively to the picture of Russian President Putin while Cluster 5 reacts preferentially to the picture of Taj Mahal in India.
The above result suggests that the memory of an object or event is encoded in a particular set of neurons. These neurons typically remain silent until they are stimulated by the encoded object or event. If the binding of tubulin to NMDARs causes neuronal silence, then dissociation of tubulin from NMDARs should be able to reactivate silent neurons. This could be the underlying mechanism of memory retrieval. Further details are discussed in later chapters.
Author: Frank Lee