Microscopy and Image Analysis

PI:
Christian Waeber
waeber@helix.mgh.harvard.edu
617- 726-0768

Technician:
Sarah Tyndall
styndall@partners.org
617- 726-6939

The Microscopy and Image Analysis core offers tools to both acquire and analyze images of various nature. An essential function of the core is to offer training, consultation and assistance during image acquisition and analysis.

For image acquisition, users have a choice between a Zeiss LSM Pascal confocal microscope (operated by the Confocal Microscope Core) or a Nikon E800 microscope with epifluorescence illumination. The Nikon microscope has a XY motorized stage and focus drive with encoder unit attached to a Roper Coolsnap cf mono cooled integrating camera, connected to a workstation running the central image analysis application of this core, Imaging Research’s MCID Elite.
Images of objects such as whole brains, microtiters plates or films can also be acquired with the Roper Coolsnap camera with a macro lens mounted onto a copy stand. The Elite system can analyze images of any kind or format, acquired using the Zeiss confocal microscope or any other imaging device (including video recordings for time lapse studies) and brought to the workstation via the network, on a CD-Rom or a Zip disk.

2.1. Quantitative Regional Autoradiography: This mode of analysis is used for regional receptor binding and functional mapping (e.g. blood flow). Absolute cerebral blood flow can be determined using the  iodoantipyrine tissue equilibration autoradiography technique, while brain activity is assessed using 2-deoxyglucose. The Image Analysis Core provides tools to quantify the results of these experiments.

Image of a human brain section showing the regional distribution of 5-HT2A receptors. From Waeber and Palacios, 1994.

2.2. Gel/blot Analysis: The Elite software includes linear and nonlinear calibration in two dimensions, corrections for gel artifacts (e.g. “smiling” gel), automated and manual lane and peak detection, and flexible baseline subtraction, as well as calibration for known standards (e.g. radioactive, calibrated optical density tablet). The amount of sample in the band of interest can then be measured, taking into account its density AND size.

2.3. 3D Reconstruction: The Elite software includes a powerful, flexible, and easy-to-use 3D reconstruction system, fully integrated with the MCID™ image analyzer. The software contains image editing, alignment, rendering and display functions that are well-suited to most bioscience 3D applications.

2.4. Grain Counting: This generic counting mode is applicable to cells in sections or cultures, grains, and other discrete objects. Valid targets are discriminated from background using intensity, color and spatial criteria (size, shape). Image transformation and combination functions assist in discriminating difficult targets (e.g. particulates overlapped with cells). Basic morphometric measurements (area, perimeter, length, etc.) can easily be performed on the selected objects. Note that quantitation of immunolabelled cells in brain sections can be done using profile-based sampling methods (methodologically similar to grain counting) or stereological methods.

2.4. Fluorescence Imaging: The Elite system includes extensive image fusion functions, with flexible adjustment of color, intensity, and transparency of each discrete fluorochrome, combination of images of different fluorochromes, as well as digital confocal deconvolution using nearest-neighbor or no-neighbor algorithms.

2.5. Stereology: The Elite system includes commonly used procedures and parameters for design-based stereological measures. An automated optical fractionator for unbiased cell counting is also included and can be used to assess the number of labeled cells in a given brain region, the length of blood vessels segments positive for a given merker or the volume of heterogeneous, non confluent cell death regions surrounding hemorrhage, to give a few examples

MCID’s stereological measures are grouped in 3 categories: 2D frames (Ns:number of 2D targets, As: area sampled, Na: 2D numerical density, Vv: volume ratio), 3D probes (Aa: area ratio, Sv: mean surface density, Vm: mean volume) and 3D disectors (N: number of 3D targets, Nv: numerical density of 3D targets, Vs volume sampled).

In addition, MCID implements optical fractionator methods, most commonly applied to cell counting (the system uses the Tiled Field Mapping module and the XY motorized stage and focus drive with encoder unit for this function). To use fractionators, the user indicates the number of sections cut through the region of interest, the actual section thickness after fixing/staining, and the sampling frequency. A montage of each section is created at low power, and an area of interest is outlined on this montage. A higher power is selected, and the system uses random systematic sampling to move the microscope stage to locations within the region of interest. At each location, interactive sampling is performed. To sample, the user focuses up and down and marks targets within the counting frame. An on-screen indicator displays the current Z position and guard volume locations. The computer keeps track of the counts and calculates an estimate of the number of targets.