Quantification of immunofluorescence colocalization

Tissue sections were stained for immunofluorescence as described above and >20 images per testis were imaged with a ×40 1.2NA objective on a ZeissAX10 epifluorescence microscope, all at a single z-section. The percent overlap between the TCF21^lin population and several marker proteins was done using accustom written ImageJ macro (available upon request). Briefly, nuclear regions of interest (ROIs) were created from DAPI staining by blurring with a Gaussian filter, making the image binary, separating overlapping nuclei with a watershed function, then saving the outline of each binary nucleus to ImageJ’s ROI manager. The signal from TCF21^lin and the immunostaining were made binary using ImageJ’s automatic thresholding function and the overlap of the binary stain and the nuclear ROI was measured using Image J’s Measurement function. Cells positive for each stain and double-positive cells were sorted and identified in Microsoft Excel. The quantification from each testis was the sum of all quantified images taken from a single testis. The macro was optimized by contrasting its results to manual quantification from at least five images per immunostaining.