TUNEL fluorescence quantification

cellprofiler

#1

Hi,

I’m working with a mammalian cell culture system in which i’ve induced cell death and am trying to sort out the mechanism. I’ve done TUNEL staining to look at endpoint apoptosis. The kit is from Roche and stains TUNEL positive nuclei green (FITC). I’ve also stained nuclei with DAPI. The images load as greyscale images in cellprofiler.

My goal is to quantify the difference in TUNEL fluorescence activity between samples - the image acquisition was via a fluorescence microscope. A post-doc in my lab turned me onto this program after I had little sucess with image-pro plus.

I have a basic understanding of this program, but using a sample (human cells from the website) pipeline in which they first isolate the nuclei by way of the DAPI image in the green image, and then apply the objectintensity module - to my images yields inconsistent and confusing results. In images with little intensity, i get higher numerical values for the intensity vs. images in which I have strong green signal. I suspect that the background, which is “black”, but is likely influencing the quantification and producing the inconsistencies i’m seeing. I’m trying to build a module to isolate the green nuclei (it is crucial that i ignore any cytoplasmic green signal) and then eliminate the “background” before quantifying it and i’ve had little success in doing so. Has anyone had any success with trying to build such a module, or is this application not suited for cellprofiler?

Thanks for all the help!

Shashank

p.s. i can upload the images if necessary - just ask


#2

Please do post an example image (including the various channels) for your experiment. And posting the pipeline you’ve developed so far may help us determine what is going on.
Anne


#3

Hi,

No problem, here is my pipeline:

loadimages
identifyprimautomatic (to id my nuclei)
identifysecondary (to id my cytoplasm)
measureobjectintensity (where i derive my nuclear TUNEL stain intensity quantification from)
measureobjectintensity (to quantify my cytoplasmic signal - there should be none)
saveimages
exporttoexcel

I’ve attached the images, however, I took them w/ ImageProPlus as .tif files and they wouldn’t attach (something about improper file format) so I took a snashot of the image, converted it to jpeg and then attached it.

Let me know if there’s anything else you need. Thanks.[attachment=0]3g.tif Snapshot.jpg[/attachment][attachment=1]3b.tif Snapshot.jpg[/attachment]


#4




#5

Hi Shashank,

Thanks for posting the images. But rather than describing the modules in words, could you upload the pipeline file itself to the forum? Then we should be able to fully diagnose the issue.
-Mark


#6

No problem

Also, the mask image module is to create an outline of nuclie with the DAPI image, and then apply it to the TUNEL (green) image and only quantitate the signal within the mask (i.e. the nuclear signal)

Thanks
TUNELPIPE3.mat (837 Bytes)


#7

Hi Shashank,

It is very likely that a high background fluorescence level could be responsible for the differences you are seeing. One way to tell is to open each image in which you see this difference in CellProfiler by double-clicking on it in the box in the lower left. Then you can go to CellProfilerImageTools > ShowOrHidePixelData and use the cursor to see what the relative values are between the background and the nuclei.

If you want to get rid of the background so you can normalize everything, you can use the CorrectIllumination_Calculate and _Apply modules. The Help for the module can get you started, and we have a paper which may also be of use to you.

Uploading examples of the images where you have this discrepancy would help sort things out. Since you converted these to JPG, which changes the data, I unfortunately can’t say for certain.

Another point is that the MaskImage module is actually redundant to the MeasureObjectIntensity module. It seems you are attempting to mask the green channel with the Nuclei objects, then measure the intensity within this new TUNEL image. MeasureObjectIntensity will report the intensity values from a given image using the spatial boundaries of a given object. By applying MeasureObjectIntensity with the Nuclei as the object and the green FITC as the image, you can do the same thing in one module.

Hope this helps!
-Mark