Investigation of neurotransmitter diffusion in three-dimensional reconstructions of hippocampal neuropil

Introduction

In the manuscript titled "Extracellular sheets and tunnels modulate glutamate diffusion in hippocampal neuropil", published in the Journal of Comparative Neurology, the authors present a reconstruction of the 3D geometry of 180 cubic microns of rat CA1 hippocampal neuropil from serial electron microscopy and corrected for tissue shrinkage. The reconstruction revealed an interconnected network of tunnels, formed at the junction of three or more cellular processes, spanned by sheets between pairs of cell surfaces. Tunnels tended to occur around synapses and axons and sheets were enriched around astrocytes. Simulations suggested that the rate of diffusion of neurotransmitter was slower in sheets than in tunnels.

A three dimensional reconstruction of a 6µ by 6µ by 5µ volume of hippocampal neuropil based on high-resolution electron microscope images reveals the extracellular space as a patchwork of flat sheets separated by a network of tubular tunnels of larger extracellular width. In this rendering of the reconstructed cell membranes, small spheres were placed at each vertex of the surface mesh and shaded yellow if the extracellular width measured at that vertex is larger than 15 nanometers and otherwise shaded cyan. The extracellular space forms sheets when surrounded by a single pair of cells and pores at the junction of three or more cells.

Animation:

Downloads

The following bzipped tar files contain the EM images, traces, 3D reconstructions, and software described in the manuscript. For more details about how the data was collected and the reconstructions were created, please refer to the manuscript

To extract the contents of tar file file.tar.bz2 , issue the command
tar xvfj file.tar.bz2 which will create a folder named file in the current directory.

EM Images

em_image_stack.tar.bz2 contains the EM images (.jpg) and the manually-traced contours of the cell processes in the images (Volumejosef.).

Reconstructions

reconstructions.tar.bz2 contains the surface meshes of the 3D reconstructions illustrated in Figure 2B-H in the manuscript. The mesh file is assumed to be in the Hughes Hoppe mesh file format which first psts the location of all vertices followed by a pst of faces that reference the vertices. Vertices are specified by a vertex index i and three coordinate positions; x, y, z: [Vertex i x y z]. Faces are specified by a face index j and three vertex indices; v1, v2, v3: [Face j v1 v2 v3]. All indices are positive integers larger than zero. For more information, see Chapter III, Section 2 of Kinney (2009)

Meshmorph

meshmorph contains the source code for a program used to manipulate reconstructed surfaces. This custom program written in c++ accepts a reconstructed surface as input. The program models the surface as a system of interconnected springs and and manipulates the location of the surface on the nanometer scale to reduce the potential energy of the springs. In this way, the program was used to control the morphlogy of the extracellular space. For documentation of the program see Chapter II, Section 1.F of Kinney (2009) You can checkout a copy with git clone git://github.com/jpkinney/meshmorph.git
Static files: meshmorph.tar.bz2

Meshalyzer

meshalyzer contains the source code for a mesh analysis program. This custom program written in c++ accepts a reconstructed surface as input, analyzes the geometry of the surface, and prints a summary report. For documentation of the program see Chapter 3 of Kinney (2009) You can checkout a copy with git clone git://github.com/jpkinney/meshalyzer.git
Static files: meshalyzer.tar.bz2

Electron micrograph of 50 nm thin section (1 of 100 in series) of rat hippocampal area CA1 with manual tracing of dendrites (yellow), axons (green), and astrocyte (blue) at a resolution of 2.3 nm per pixel.

One cubic micron from the reconstruction shown in Fig. 2F was rendered with the ECS as solid and intracellular space invisible. Coloring the ECS according to extracellular width reveals the interconnectedness of the tunnel network (blue).

Authors

Justin P. Kinney,1*
Josef Spacek,2
Thomas M. Bartol,1,3
Chandrajit L. Bajaj,4
Kristen M. Harris,5
Terrence J. Sejnowski 1,3,6

1Howard Hughes Medical Institute, Salk Institute for Biological Studies, La Jolla, CA 92037, USA
2Charles University Prague, Faculty of Medicine, Hradec Kralove, Czech Republic
3Center for Theoretical Biological Physics, University of California at San Diego, La Jolla, CA 92093-0374, USA
4Computational Visualization Center, Department of Computer Sciences, University of Texas, Austin, TX 78712, USA
5Center for Learning and Memory, Department of Neurobiology, University of Texas, Austin, TX 78712-0805, USA
6Division of Biological Sciences, University of California at San Diego, La Jolla, CA 92093, USA

*Now a CRC Postdoc at McGovern Institute, MIT.