A compact light-sheet microscope for the study of the mammalian central nervous system

Abstract

Investigation of the transient processes integral to neuronal function demands rapid and high-resolution imaging techniques over a large field of view, which cannot be achieved with conventional scanning microscopes. Here we describe a compact light sheet fluorescence microscope, featuring a 45° inverted geometry and an integrated photolysis laser, that is optimized for applications in neuroscience, in particular fast imaging of sub-neuronal structures in mammalian brain slices. We demonstrate the utility of this design for three-dimensional morphological reconstruction, activation of a single synapse with localized photolysis, and fast imaging of neuronal Ca(2+) signalling across a large field of view. The developed system opens up a host of novel applications for the neuroscience community.

Figure 1. Example light sheet images of a large, living pyramidal neuron.

These images were acquired from the CA3 region of an organotypic rat hippocampal slice, using the prototype light sheet fluorescence microscope described. The neuron was briefly patched and filled with two inorganic dyes, AF594 (a) and AF488 (b), then transferred to the light sheet microscope where it was imaged with illumination of corresponding wavelength under the 40× objective. Images a and b are composites of maximum intensity projections Z-stacks acquired with 1.4 μm intervals. Inset images 1 and 2 are single plane images, selected from the stacks shown in a, of dendrites from the basal (1) or apical (2) arbour, with spines clearly visible on both. These images are presented without any post-acquisition processing other than the application of a lookup table and adjusting brightness and contrast.

Figure 2. Demonstration of glutamate uncaging using 405 nm laser photolysis, with resulting Ca²⁺ transients recorded on the LSFM.

(a) Maximum intensity projection of an AF594-filled pyramidal neuron from the CA3 region of an organotypic rat hippocampal slice, acquired using the prototype LSFM with 40× objective. The cell was also filled with Ca²⁺-sensitive dye OGB-1. The slice was bathed in 1 mM MNI-glutamate during the experiment. (b) Region of the basal dendritic arbour (indicated by the white box in panel a) where the photolysis experiment was performed. The white circle indicates the location of the 405 nm photolysis spot. White squares indicate the regions of interest (ROIs) chosen for analysis of uncaging-evoked Ca²⁺ transients: a dendritic spine (S) and three locations on the dendrite (D1, D2 and D3). (c) Ca²⁺ traces from 10 photolysis trials are shown in grey, with the averaged trace overlaid in red. An OGB-1 image using 488 nm illumination was acquired on the LSFM every 5 ms (200 Hz). The Ca²⁺ trace for each ROI was calculated as the mean intensity for each time point, with background intensity subtracted, and displayed as % change in fluorescence divided by baseline fluorescence (%ΔF/F), where baseline is defined as the mean intensity for the 90 ms preceding the photolysis flash. The timing of the photolysis flash (duration 2 ms) is shown by the blue arrowhead. (d) Ca²⁺ trace for a single photolysis trial of interest selected from those shown in c. In this example a secondary Ca²⁺ transient is clearly seen in the dendrite about 200 ms after the initial response in the spine. This is probably a Ca²⁺-induced Ca²⁺ release (CICR) from endoplasmic reticulum in the dendrite.

Figure 3. Example LSFM recordings of Ca²⁺ events in axon and dendritic spines.

(a) Action potential-evoked Ca²⁺ influx detected by LSFM imaging at various points along a neuronal axon. The CA1 region of an organotypic rat hippocampal slice was virally transfected with Ca²⁺-sensitive dye GCaMP6s. Left-hand panel shows an example single-plane image of an axon, acquired on the prototype LSFM using 488 nm illumination. The slice was perfused with low-Mg²⁺ ACSF to induce action potentials, and images were acquired at 30 Hz. Right-hand panel shows the Ca²⁺ signal at various ROIs (indicated by numbered circles in left-hand panel). The regularly spaced transients reflect Ca²⁺ influx evoked by action potentials in the axon, and probably mediated by activation of presynaptic VGCCs. (b) Ca²⁺ influx at dendritic spines produced by spontaneous neurotransmitter release events. Left-hand panel shows an example single-plane image of a dendrite belonging to a GCaMP6s-transfected neuron from the CA1 region of an organotypic rat hippocampal slice, acquired on the prototype LSFM. The slice was perfused with ACSF containing 1 μM TTX to abolish action potentials during image acquisition at 30 Hz. The right-hand panel shows the Ca²⁺ signal at various spines (indicated by numbered circles in left-hand panel). Ca²⁺ signal was calculated as % change in fluorescence divided by baseline fluorescence (%ΔF/F).

Figure 4. 3D drawing of the inverted light sheet design, based on the openSPIM project.

Laser (L) provides illumination. Beam expander (BE) expands the laser beam. Adjustable slit (AS) adjusts the width of the beam and cylindrical lens (CL) focus the beam to a light sheet which is delivered to the perfusion sample chamber (SC) by relay lenses (RL) and illumination objective (O1). The two objectives (O1 and O2) are mounted on a customized holder. Excited fluorescent signal is collected by a detection objective (O2) and tube lens (TL), then projected onto a sCMOS camera (CAM). Stage (ST) and automatic linear actuator (LA) scans the sample chamber. Adjustable Fibre Collimator (FC) brings a diffraction-limited photolysis spot onto the imaging plane. An extra illumination laser is not shown on the graph. The blue beam shows the path of the 488 nm imaging laser, and the purple beam shows the path of the 405 nm photolysis laser.