Methodology problem

[habu]

---

Members don't see this ad.

Hello - I've just got to let you know-


I'm doing a research on pathology and I'm now stuck with a evaluation of ihc expression patterns. I'm trying to figure out a method which would allow me to get relative expression of cd4 when comparing to cd8. I've got 20 cd4 stained slides and 20 cd8 stained slides. First 10 slides are case-group and second 10 slides are the control-group. So I want to know is there relative expression difference between case- and control-group(I know sounds quit stupid).

I practically made up method of my own by just explicitly evaluating each slide pair. When both slides were equally strong/weak, I gave them both value 3. When one was more expressed than another, I gave the stronger value 3 and the weaker value 2 to 1, depending on how weak the another was. For example if slide pair 2 has cd4-slide which is stronger and cd8- slide which is very weak, I give cd4-slide value 3 and cd8-slide value 1


Now I had values in two different tables. Then I just took CD4/CD8 ratio. Now I had data in one table and I could do some statistical analysis. I had five values
CD4/CD8 = 1/3 strong cd8 expression(dominant)
CD4/CD8 = 2/3 moderate cd8 expression
CD4/CD8 = 3/3 none difference
CD4/cD8 = 3/2 moderate cd4 expression
CD4/CD8 = 3/1 strong cd4 expression


Thank you for the (coming) help

 

[JHopRevisit]

Do you need the stain to remain in situ? That is, do you need spatial information?

If you don't, or if you could sacrifice some of your samples instead of using them for ihc labelling/imaging, just chop up your sample (ie run it through a cell strainer, or for some samples you might need to digest some connective tissue first, you can find protocols that are similar to trypsinizing), run it through a lysis buffer, stain it, and then run it on Flow Cytometry. That should solve the problem of having to guesstimate the expression level by eye.

There are techniques for evaluating labeling intensity on an image but they involve a little bit of image processing, how comfortable are you with that? Know any MATLAB? Worked with ImageJ before? I've also only used them for fluorescence stuff, it's a totally different thing for a simple stain.

Otherwise, yea, just eyeball it. Most imaging stuff in immunology is done that way, "here's an out-context image I took that I promise is representative, see how it has more blue stuff?" Usually there's other techniques that make your case and the pictures are just there to look pretty and drive the point home, so it's not that big of a deal.

Also, I'm not totally sure I'm interpreting your question correctly, and a few more details might be helpful. What's the sample, what type of staining, etc.

 

[reine1jb]

if you don't need a spatial resolution, I would recommend something a little more quantitative if you have the equipment for it. For instance, I use quantitative real-time PCR to measure the relative expression of a gene in one population versus another population..

 

[mercaptovizadeh]

Hello - I've just got to let you know-


I'm doing a research on pathology and I'm now stuck with a evaluation of ihc expression patterns. I'm trying to figure out a method which would allow me to get relative expression of cd4 when comparing to cd8. I've got 20 cd4 stained slides and 20 cd8 stained slides. First 10 slides are case-group and second 10 slides are the control-group. So I want to know is there relative expression difference between case- and control-group(I know sounds quit stupid).

I practically made up method of my own by just explicitly evaluating each slide pair. When both slides were equally strong/weak, I gave them both value 3. When one was more expressed than another, I gave the stronger value 3 and the weaker value 2 to 1, depending on how weak the another was. For example if slide pair 2 has cd4-slide which is stronger and cd8- slide which is very weak, I give cd4-slide value 3 and cd8-slide value 1


Now I had values in two different tables. Then I just took CD4/CD8 ratio. Now I had data in one table and I could do some statistical analysis. I had five values
CD4/CD8 = 1/3 strong cd8 expression(dominant)
CD4/CD8 = 2/3 moderate cd8 expression
CD4/CD8 = 3/3 none difference
CD4/cD8 = 3/2 moderate cd4 expression
CD4/CD8 = 3/1 strong cd4 expression


Thank you for the (coming) help

Do you know anything about the affinity of the antibodies for CD4 and CD8? Also, I would think a better approach would be lysing the tissue and then co-staining with both antibodies tagged with different flurochromes, and then do flow cytometry on it. That will make it a bit more quanititative.

 

[habu]

Thanks for the quick answers

First I don't need spatial information and the stain remains in situ. At the moment I can't do PCR or Flow Cytometry. I've only got single stained slides. I've not used ImageJ and I'm bit novise with MatLab. Also my tutor wants just some eyeball method to estimate cd4/cd8 ratio. About the affinity - I could get the values(I've the data from ihc-kit).

My study is about mucosal tissue(oral). I'm just looking for immunohistochemichal differences between two similar diseases. IHC-staining was conducted with LabVision machine and with Envision+ polymerase-based staining method.

Jeah I know this is crappy way to do quantitive analysis, but it's all I have at the moment.