Ooh Ooh Ooh! They Finally Did It!
Over the last few years there has been increasing interest in behavioral pattern separation. This has primarily focused on spatial pattern separation in rats and mice or pattern separation of visual objects as evaluated in humans. What has been overlooked is a vast literature of temporal pattern separation extending into the early 1990s. The manuscript I will cover in this post not only addresses this type of temporal pattern separation, but also used spatial and temporal pattern separation abilities as outcome measures for a mutant mouse to demonstrate the neurophysiological deficits that are dissociated across different parts of the hippocampus map onto functional deficits at the behavioral level.
The manuscript discussed in this post is “GluN2A−/− Mice Lack Bidirectional Synaptic Plasticity in the Dentate Gyrus and Perform Poorly on Spatial Pattern Separation Tasks” by Timal Kannangara in Brian Christie’s lab at the University of Utah. The article can be found in the journal Cerebral Cortex.
Disclaimer: In this manuscript they authors used behavioral tasks developed and popularized by my wife and myself over the last 8 years. As such, I am excited to see my behavioral work being adopted by others in this way.
What is Being Studied and Why?
Kannangara and colleagues set out to determine if the GluN2A subunit of the NDMA receptor is specifically important for behavior More info Here. They did this by knocking out the subunit and evaluating the neurophysiological characteristics of the cells in two subregions of the hippocampus: the dentate gyrus and CA1. They also used behavioral tests to specifically determine if there were functional deficits that map onto dentate gyrus or CA1 specific disruption.
An Innovative Approach
Kannangara and colleagues demonstrated that there was reduced long term potentiation (LTP) as well as reduced long term depression (LTD) in the dentate gyrus of mice lacking the GluN2A subunit of the NMDA receptor. Using these same mice, Kannangara and colleagues showed that LTP was only slightly in CA1. They also demonstrated that there was only a minuscule reduction to LTD in CA1. They demonstrated this was not due to presynaptic mechanisms and showed normal AMPA receptor activity in these mice. Overall in these mice, there was an 80% reduction in the NMDAR:AMPAR ratio in the dentate gyrus, whereas only a 50% reduction was observed in CA1. Critically, neither the dentate gyrus or CA1 cells had an altered appearance or odd morphology as viewed by a Golgi stain and Scholl analysis. These data suggest the GluN2A subunit may be differentially expressed in the dentate gyrus and CA1, and thus knocking out function may affect these regions differently.
To get a readout of this differential involvement, the authors used two behavioral tasks shown over the years to effectively dissociate dentate gyrus and CA1 function: a spatial pattern separation task called the Metric Task and a temporal pattern separation task called a Temporal Ordering Task.
Briefly, the Metric Task requires the mice to explore 2 objects in a box separated by a given distance for 15 min. after a 5 min break, the objects are moved closer together and a normal mouse will explore the objects like crazy if they notice the objects were moved. A lack of exploration is taken as evidence that the mice did not notice the change-and thus impaired spatial processing.
For the Temporal Ordering Task mice are allowed to explore two copies of an object for 5 min. After a 5 min break the mouse then explored two copies of second object, and then a third. After a final 5 min break, the mouse is given a change to explore a copy of the first and third objects they encountered. Typical mice will prefer to explore the first over the third object if given a choice. A lack of preference is taken as evidence that the mice do not remember the order the objects were presented-and thus impaired temporal processing.
Based on the physiology, it could be hypothesized there would be a deficit for the Metric task but not the temporal ordering task. This was precisely what the authors observed. There were absolutely no deficits for the temporal processing task, all the mice performed optimally. However, for the metric task, there was a profound deficit. Intriguingly, although not mentioned by the authors, this deficit almost copied the deficits seen after ablating the dentate gyrus (CA1 lesions do not have a profound effect on performance of this task). What I mean is that their GluN2A knockout mice acted as if they did not have a functioning dentate gyrus, not just a slightly inefficient dentate gyrus.
Also important was the fact that impaired NMDA function in the dentate gyrus did not appear to negatively impact performance on behavioral tasks shown to be unaffected by dentate gyrus lesions. These data are important because they suggest that behavioral tasks developed to test hippocampus subregional function are useful as functional outcomes in studies such as this, not just for studies using mouse models of genetic disorders, as I have written about before Here and Here.
Take Home Message
Kannangara and colleagues demonstrated that synaptic plasticity mechanisms in the dentate gyrus appear to be a common mechanism with spatial pattern separation function of the dentate gyrus. However, they found no similar pattern between plasticity in CA1 and temporal pattern separation. In fact, they found the GluN2A subunit of the NMDA receptor appeared necessary for dentate gyrus LTP and LTD, but was not involved in CA1 LTP or LTD. Importantly, the authors demonstrated that in the dentate gyrus, knocking out the GluN2A subunit of the NMDA receptor has a profound effect on plasticity, but did not alter the ultrastructure of the granule cells as quantified from Golgi stains.
The combination of altered synaptic plasticity and a normal cell structure does suggest that the disruptions to synaptic plasticity (impaired LTP and impaired LTD in the dentate gyrus) were sufficient to disrupt spatial pattern separation (although the authors make clear their data can draw no causal link). As predicted, since CA1 seemed to have relatively normal plasticity, CA1 function as measured by temporal processing tasks was normal in GluN2A knockout mice.
TL;DR The GluN2A subunit of the NMDA receptor appears to be important for dentate gyrus function, but less so for CA1 function. This can be observed both in neurophysiology as well as behavioral studies. In fact, the behavior and the neurophysiology provided converging evidence that the GluN2A subunit of the NMDA receptor is more involved in spatial than temporal processing.
Behavioral analysis of hippocampus function rocks!