Delineating specific brain regions and counting nuclei with FIJI / ImageJ (Part 1)

[@gvozdemape6636]

Hi KRU neuro, my name is martin im from Argentina, nowdays im working on hipocampus of rat brians and im wondering if maybe you can help me with the count of nuclei and apoptosis on images colored with HE. Sorry for the disturb. and thank you very much for your videos, each one was and are very helpful to me.

[@kruneuro]

Hello Martin! I think you should be able to separate the colored H&E image into its red/green/blue components. I played around with an H&E image I found on the web. However, rather than using the split color function, I used the Image > Color > Color Deconvolution function. If this function isn't listed, you'll need the newest version of FIJI as it should be there. Within those individual color "channels", you'll notice the H stain (mostly blue) be distinct from the E stain (mostly red). The green channel will be mostly blank. The key here is that your marker for apoptosis (TUNEL?) must be a substantially different color from the rest of the tissue coloring. Is the apoptosis marker something more distinct, like brown or black? If so, it should be most visible/distinct in the green channel. Then, you should be able to threshold the green channel to highlight particles of apoptosis markers. If you are trying to do percentage of cells in apoptosis, then you would just do the same thresholding and counts for the blue (H stain) image to get total cells. You'd then divide the green counts from the blue counts. Hopefully I read you right and gave you helpful pointers. -Kevin

[@gvozdemape6636]

​@@kruneuro Ohh, thank you very much for your answer, it´s really helpful for me. I´m going to try the color deconvolution later today cause the RGB separation show huge differences between similar images. Relateted to apoptosis, we didn´t have the chance to use any marker because we couldn´t afford it. However, i´m using digital markers to distinguish apoptotics cells from the normal cells but i have to work image by image and cell by cell. In that way, I think that i can try to use that markers as anyother marker. so i´ll be following your advices soon. Yes, it was an excellent explanation. once again I want to thank you so much for your answer and advices Kevin. Martin

[@kruneuro]

@@gvozdemape6636 Ah, I understand budgetary constraints. I saw that the Color Deconvolution option has multiple settings; H&E + DAB might be another one to try if apoptosing cells have any coloration difference leaning towards brown. Otherwise, if it's a cell or subcellular shaping difference (perhaps nuclear fragmentation visible in the H stain), there's probably a way to have different particle count settings that I'm less aware of. Whatever the case, good luck!