Is my boss wrong? - head hurts... (Nov/16/2007 )

Hello, I have another problem and I am already being desperate.
My boss wants to examine levels of a certain protein by quantitative rt- PCR in lymphoid cancers. Problem is, patient samples have different percentage of cancer cells (by flow cytometry), and the protein is not restricted to cancer cells only, it is also expressed in normal cells. I have a gut feeling that in this case, quantitative rt-pcr is a wrong method, because of additional variability stemming from different percentage of cancer cells, no matter how many controls we take. Nevertheless I can’t convince my boss about this. He just wants me to find more appropiate internal controls and produce reults.
My thought is, could the problem might be solved by sorting cells of some sort? I’d do it if I had antibodies. Or my boss is right and nobody cares about the percentage? Or I am right, and then, what to do? I surfed the net for papers, read up and I am desperate. He’s telling me I am just being stubborn and not motivated enough. Every time I think we got to an agreement, and later he doesn't seem to remember what conclusion we've got to.

I believe the percentage of cancer cells do matter, it obviously has too. Else how will you know if the concentration of protein X has any link to the cancer cells?

But the experiment has to be done, you want to see if the level of expression of protein X varies with the level/stage of progress/agressiveness of said lymphoid cancer.

The question become can you isolate the cancer cells? What techniques do you have available to you? Can the flow cytometer sort your cells for you?
If it can't perhaps you can devide your samples into groups....
Group A - 0 - 5% cancer cells
Group B - 5 - 10 % cancer cells
and so on.

Run the rt-PCR and see if the protein levels correspond to the group.

Unfortunately it can't. I can't really group it, too, because in qrtpcr small differences in the beginnning produce big differences at the end. I did some semi-q pre-runs, and the samples are really variable, and they don't simply correlate with the percentage, so there must be something about it.
I'd sort it on a magnetic column, if right antibodies were provided. If anyone knows another (easier? faster? better?) way to do it, I'd be very grateful.

Don't know if this will be of any help to you, but it seems to me that you would have to find some way to make a pure population of cancer cells and a pure population of normal cells, then do the quantitative RT-PCR to compare them. I don't know much about how this would be done, but I'm thinking along the lines of culturing a population from individual cells. I seem to recall some cancer researchers at my university who made cancer cell lines from isolated tumors, so maybe you should check the literature if such a thing has been done before.

I'll certainly think about it. Lymphoid cancers aren't easy to culture, they often die outside their environment. Of course, it depends on the type of cancer. So I'd be talking with the clinicist who takes the samples...
I did some talking at the lab today, maybe I'll get the solution (it seems the boss agrees now that we have to somehow sort it and I really hope he was listening, not just nodding).

It seems immunostain would serve the purpose best, if possible.

qRT-PCR for PROTEIN levels?? RNA levels do not always reflect protein levels. An immunostain is by far the better way to go.

I know there is a type of laser-assisted sample selection apparatus that you can practically burn off non-canerous cells under microscope on a tissue section. If you can get a hand on one of those, that will be the perfect solution.

h2so4hurts: I know. Oh man, you know what I mean.

genehunter: Yes, and we do have one in the institute. I've been investigating if it's an appropiate method. There are some technical problems (the method is heaps better for DNA than RNA, there's high risk of sample degrading), but I think they can be solved. You're quite right, though, dropping the whole RNA stuff and going immunostain is the most sensible thing to do. It's just the boss much prefers the former. I can understand that, if you want to measure another gene expression you just buy another cheap set of primers.
Thanks all for the extensive help, I'm going to look around and see what I am going to do

Just for a basic assessment, you could control for the percentage of cancer cells just by calculation. I woudn't trust it that much, but your boss is your boss and what's the sense of discussing if he is so ignorant?
So, let's say you have sample A with 20percent and sample B with 40 percent cancer cells. If the starting amount of your mRNA was the same per cell, you should end up with a CT +1 in sample B. You could standardize your measurement to the amount of cancer cells...