Skip to main content

The neural circuit mechanisms of anti-NMDA receptor encephalitis: revealing the role of NMDA receptors in spatial memory in mice

Final Report Summary - ISOLM (The neural circuit mechanisms of anti-NMDA receptor encephalitis: revealing the role of NMDA receptors in spatial memory in mice.)

Summary description of the project objectives:
The main goal of our project was to investigate the NMDA receptor disfunction in the hippocampus during memory formation and retrieval by using human antibodies from patients with anti-NMDA-receptors (anti-NMDAr) encephalitis, a disease discovered by Josep Dalmau in 2007 (co-supervisor of this project). The project the following objectives:

1) Generate a good mouse behavioral model for long-term spatial memory impairments: a) Including the adjustment of the anti-bodies dosage to minimize effect in other behavioral aspects than the spatial navigation task. b) Compare the effects of local hippocampus injection with intra-ventricular injection.

2) Quantify behaviorally the impact of anti-NMDAr infusion in the formation of new memories and the retrieval of existing ones. Characterize the reversibility of the effects after discontinuing the anti-NMDAr infusion. Collect neural data simultaneously to the behavioral experiments.

3) Characterize the neural basis of memory deficits by measuring the collective activity of hippocampal place-cells across the behavioral experiment (pre-, during and post-injection). Quantify changes in the place-fields activity, which reflect memory stability at the single cell level. Compute decoding accuracy and signal to noise variability across days, which underlie memory impairments at the population level

Description of the work performed since the beginning of the project:
We brought the calcium imaging technique utilizing miniature microscopes in freely moving animals to the host country (Spain) and recorded hundreds of neurons in a novel preparation. This implies combining Calcium imaging of hippocampal circuits and the study of the with the animal model of chronic injection of antibodies through intra-ventricles injection of antibodies utilizing osmotic-pumps previously developed by Planagummá et al. (2015).

To combine calcium-imaging with this animal model we re-designed the way the infusion of anti-bodies is carried out and the way the mini-micrioscope is implanted (Figure 1). This implied a new built cannula scull-base and mounting a novel base-plate piece for the mini-microscope.

We developed a novel behavioral memory task to quantify the learning and retrieval of spatial memories while performing imaging of hippocampal populations or electrophysiology (Figure 2). This new paradigm, allows running many animals per day (up to 12) and with a high throughput number of trials per session (~100trials). The entire task is automatically controlled by computer allowing to run several animals in parallel provided there are multiples behavioral boxes (we used two until now). All variables of the experiment are automatically stored in real time, building a large dataset of dozens of Terabytes size, for detailed off-line analysis. Tuning the parameters of this new task (e.g. training staged, session duration, retrieval design, etc) took training several batches of animals in slightly different versions of the task until arriving to the current version in which animals can learn in about 1-2 weeks.
We developed a novel behavioral set-up to control the memory task, and a python code with a user-friendly interphase (GUI) to control the various effectors and effectors (sound and water valves) and receptors (licks in water ports, video tracking). This required the use of an Arduino board and designing electrical circuits to detect licks in each of the 8 ports. Moreover, because the task required close loop tracking of the position of the animal, it involved programing in Python a detection algorithm which was fast and robust.

We trained a Ph.D. student and an animal technician to perform same-day data analysis and run the behavioral experiments, respectively.

Description of the main results achieved so far
We successfully built an animal model preparation where animals can carry the mini-microscope and the osmotic pumps for our study. This model could be used for different types of research utilizing pharmacology combined with freely moving calcium imaging (to be published).
We developed a new memory paradigm to run in mice, which is completely computer-controlled that tackles the limitations of previous attempts of generating high throughput data sets (Dupret et al 2010, Roux et al 2017) doubling the number of trials per session and in a completely human un-supervised manner. For this we designed and built the equipment and software that allow us to record neural activity utilizing freely moving calcium imaging and drug infusion in a spatial navigation task with a “learning” and “memory” components.
We successfully brought the technique to Europe as one of the first labs in the continent performing this king of methods for high number of “same neurons” of neurons recordings over days (see Ziv et al 2013).
We are finalizing our third batch of animals that will allow us to have results for a publication in the next months. We have found that animals can readily learn the task in 1-2 weeks. Once learned, they learn to recognize the rewarded port in a few trials (5-10) during which they sometimes start exploring the ports that were rewarded the day before. Once they are able to obtain reward from the new target (rewarded) port, one trial is enough to create a solid memory so that the performance in the subsequent trials is above 80% (chance would be 12,5% given that there are 8 ports). In retrieval sessions 3 hours later we find that results are heterogeneous: some animals remember well the location of the target port based on the distal cues, whereas others use the local cues (e.g. odors). Some animals in contrast do not seem to remember the rewarded port three hours later. We are currently quantifying these memory traces and assessing the impact of NMDA antagonist (CPP) as well as a NMDAr antibodies.

The expected final results and their potential impact and use (including the socio-economic impact and the wider societal implications of the project so far)
We would like to establish our behavioral experiment as a new paradigm for spatial navigation in rodents and to become a reference for other researchers as a general method for studying spatial memory and retrieval in rodents as currently occupies the “Morris Water Maze” experiments with no electrophysiology registration.
To confirm in a big data set that our behavioral task is susceptible in a controlled fashion to the pharmacological manipulations of hippocampus as well as the inmuno-ablation of the NDMA receptors.
To set a new reference model for clinical studies with animal research to study hippocampus and prefrontal cortex malfunctions.
To describe which type of memory component (formation and/or retrieval) is impaired by the effect of the NMDAr antibodies in the patients. This could shed light on how to treat patients after the acute phase of the disease to reduce the long recovery period of 1-2 years. This will improve quality of life of this patient’s population and their families, as well as reduce the cost of public health treatment on these patients.