Measuring Mitochondrial Shape with ImageJ

Abstract

Mitochondria are shaped by opposing fission (division) and fusion events. Mounting evidence indicates that mitochondrial shape influences numerous aspects of mitochondrial function, including ATP production, Ca²⁺ buffering, and quality control. Despite the recognized importance of mitochondrial dynamics, the literature is rife with subjective, categorical estimates of mitochondrial morphology, preventing reliable comparison of results between groups. This chapter describes stringent, but easily implemented methods for quantification of mitochondrial shape changes using the open-source software package ImageJ. While we provide examples for analysis of epifluorescence images of cultured primary neurons, these methods are easily generalized to other cell types and imaging techniques.

Appendix—Morphometry Macro

var ch = 0; // channel to be analyzed for RGB images/** Measure mitochondrial morphology in the current selection* Ctrl+Shift+O closes current and opens next image*/macro “Morphometry [F7]” {title = getTitle();morphometry(title, false); // not batch mode}/** Batch-apply a set of “named” ROIs to analyze images with that file name*/macro “Batch Morphometry [F8]” {dir = getDirectory(“Select an image directory”);while (roiManager(“Count”) == 0)waitForUser(“Please open named ROIs into ROI manager”);prevName = imgName = “”;n = roiManager(“Count”);for (i = 0; i < n; ++i) { // loop through the ROI Manager tableprevName = imgName;imgName = call(“ij.plugin.frame.RoiManager.getName”, i);if (isOpen(imgName)) { // named image is openselectWindow(imgName);} else { // done with current image, close and open nextif (isOpen(prevName)) {selectWindow(prevName);close();}open(dir + imgName);}roiManager(“Select”, i);morphometry(imgName, true); // batch mode}}function morphometry(title, batchMode) {while (ch < 1 || ch > 3) { /* RGB channel not yet selected, initialize; reinstall macro to change channel */ch = getNumber(“Analyze RGB channel(1-3):”, 1);run(“Set Measurements...”, “decimal=5 area perimeter fit”);print(“image\t n\t area2\t area-weighted ff\t form factor\t aspect ratio\t length”); /* header for results table */}if (bitDepth == 24) // RGB imagerun(“Make Composite”);if (isOpen(“Binary”)) {selectWindow(“Binary”);close();} // close previous working imageif (isOpen(“Skeleton”)) {selectWindow(“Skeleton”);close();} // close previous working imageselectWindow(title);if (selectionType() == -1) // no selectionrun(“Select All”);if (!batchMode) {roiManager(“Add”); // save selection to ROI Manager for batch processinglast = roiManager(“Count”) - 1;roiManager(“Select”, last);roiManager(“Rename”, title);/* roiManager(“Save”, File.directory + “named_ROIs.zip”); */ /* un-comment to save ROIs automatically */}// copy selection to new window and clear outsidesetSlice(ch); // ignored if grayscalerun(“Duplicate...”, “title=Binary”);run(“Make Inverse”);if (selectionType != -1) { // outside of ROI is selectedrun(“Duplicate...”, “ “); // make a mask of the backgroundrun(“Convert to Mask”);run(“Create Selection”);run(“Make Inverse”);roiManager(“Add”);close();n = roiManager(“Count”);roiManager(“Select”, n - 1);getRawStatistics(_area, backG); // mean is backgroundsetColor(backG);run(“Restore Selection”); // fill outside of selection with backgroundfill();run(“Gaussian Blur...”, “radius=64”); // smooth abrupt background transitionroiManager(“Delete”); /* delete masking selection (ROI manager has cell selections) */}run(“Select None”);// subtract background and thresholdrun(“Subtract Background...”, “rolling=50”); /* non-destructive filter even if already applied */run(“Make Binary”);// also try other threshold methods included with Fiji, e.g.: run(“Auto Threshold”, “method=Li white”);// create Results table of metrics, one line/particlerun(“Analyze Particles...”, “size=9-Infinity circularity=0.00-1.00 show=Masks pixel clear”);awff = ff = ar = sum_a = a2 = len = 0;for (i = 0; i < nResults; i++) { // for every particle in tablea = getResult(“Area”, i);p = getResult(“Perim.”, i);ar += getResult(“Major”, i) / getResult(“Minor”, i); /* aspect ratio = length / width */sum_a += a;a2 += a * a; // area2 = a2 / (sum_a * sum_a)awff += b = (p * p) / (4 * 3.14159265358979); // awff = ff * (a / sum_area)ff += b / a; // ff = p^2 / (4 * pi * a)}nParticles = nResults;// skeletonize to get lengthselectWindow(“Mask of Binary”); /* created by Analyze Particles .., excludes noise (< 9 pixels) */rename(“Skeleton”);run(“Skeletonize”);run(“Analyze Particles...”, “size=0-Infinity show=Nothing pixel clear”);for (i = 0; i < nResults; i++)len += getResult(“Area”, i);// average and outputa2 /= sum_a * sum_a;awff /= sum_a;ff /= nParticles;ar /= nParticles;len /= nResults;print(title + “\t “ + nParticles + “\t “ + a2 + “\t “ + awff + “\t “ + ff + “\t “ + ar + “\t “ + len);selectWindow(title);