Article

"Bright field microscopy as an alternative to whole cell fluorescence in automated analysis of macrophage images" [PDF]

by Jyrki Selinummi, Pekka Ruusuvuori, Irina Podolsky, Adrian Ozinsky, Elizabeth Gold, Olli Yli-Harja, Alan Aderem, and Ilya Shmulevich. PLoS ONE, 4(10): e7497.

See the Downloads page for all the data used in the preparation of the paper.

Motivation and basics of the method

Flowchart of the image analysis procedure:

The whole cell fluorescence in the second step can be replaced by contrast enhanced bright field images.

(click to enlarge)

In the article, we present methods for improving contrast in bright field microscopy, removing the need for whole cell fluorescent staining in automated image analysis. We show that the enhanced bright field images can be directly applied in automated whole cell detection using the CellProfiler software, originally designed for fluorescence microscopy. Markers for each cell, such as fluorescent nuclei, are still required for single cell analysis.

The contrast enhanced images are constructed in the following way:
  1. Take a z-stack of 3 or more bright field images of different focus levels.
  2. For each x - y pixel location, measure how much the pixels' intensities vary in the z-dimension (from slice to slice).
Since the intensities vary only a little in background areas, the background appears dark in the resulting images. On the other hand, in areas occupied by cells there is more variation in the z-direction, making those pixels more bright. See the examples below.


Update Dec 17, 2009:
  • Note that you can easily test the projection by opening any z-stack of bright field images (e.g. the example stacks) in ImageJ, and selecting from the ImageJ menu: image - stacks - z-project - projection type: standard deviation
  • Based on very latest data, it seems the projections are somewhat sensitive to the slice selection after all. Commonly, stacks include slices that are too far away from optimal focus. When calculating the projection, try to select slices close to the best focus, and good results might also be obtained by selecting slices where cell borders are not very well visible. For the test data in this article, this phenomenon was not observed but any slices could be used for the projection as stated in the text.

Examples

Contrast enhancement and whole cell detection with bright field 3-D stack projections

Left: Original bright field image.
Center: Contrast enhanced projection using bright field z-stacks (no fluorescence).
Right: Automated cell segmentation result. Fluorescence nuclei used as markers for each cell, and whole cell areas detected using the contrast enhanced bright field images.

(click to enlarge)

Contrast enhancement example #2


(a) Low contrast bright field image.
(b) Fluorescence staining for whole cell detection.
(c) Standard deviation projection of stack of bright field images. No fluorescence required.
(d) Inverse of the projection for another visualization of the projection result in (c).

In addition to increased contrast, the projection (c and d) also suppresses background nonuniformities visible in (a).


Abstract


Background:

Fluorescence microscopy is the standard tool for detection and analysis of cellular phenomena. This technique, however, has a number of drawbacks such as the limited number of available fluorescent channels in microscopes, overlapping excitation and emission spectra of the stains, and phototoxicity.

Methodology:

We here present and validate a method to automatically detect cell population outlines directly from bright field images. By imaging samples with several focus levels forming a bright field z-stack, and by measuring the intensity variations of this stack over the z-dimension, we construct a new two dimensional projection image of increased contrast. With additional information for locations of each cell, such as stained nuclei, this bright field projection image can be used instead of whole cell fluorescence to locate borders of individual cells, separating touching cells, and enabling single cell analysis. Using the popular CellProfiler freeware cell image analysis software mainly targeted for fluorescence microscopy, we validate our method by automatically segmenting low contrast and rather complex shaped murine macrophage cells.

Significance:

The proposed approach frees up a fluorescence channel, which can be used for subcellular studies. It also facilitates cell shape measurement in experiments where whole cell fluorescent staining is either not available, or is dependent on a particular experimental condition. We show that whole cell area detection results using our projected bright field images match closely to the standard approach where cell areas are localized using fluorescence, and conclude that the high contrast bright field projection image can directly replace one fluorescent channel in whole cell quantification. Matlab code for calculating the projections can be downloaded from the supplementary site: http://sites.google.com/site/brightfieldorstaining