The issue is that I would like to sort out various cell types (e.g. tumor cells, endothelial cells, microglia) from brain tumor samples using FACS and have trouble finding an appropriate protocol on how to treat the tissue before. Is it necessary to get rid of neurons? If yes, how do I do that? Has anybody got suggestions for a suitable Percoll gradient that would separate all other cell types from neurons? Does myelin somehow interfere with FACS? Is it inevitable to get rid of myelin and if yes how do I get rid of it?
Can I stick to collagenase digestion, which I used previously or is it better to use a papain digestion protocol?
In case it helps answering the questions: I'm working with orthotopic, syngeneic mouse tumor models using GL261 cells
A lot of questions, I know, but I would already be happy if you could answer only one of them. Thank you very much for all suggestions!
I don't know anything about your tumor model. My experience was with isolating brain endothelial cells from non-cancerous mouse brain.
I don't think you need to get rid of the neurons (and they'll probably mostly die anyway) as long as you have good markers for your cells of interest during FACS. For endothelial cells, I used CD45-CD31+ (if you only look at CD31, many leukocytes will have intermediate levels of CD31).
The problem with the endothelial cells is that they will largely stick to each other even with collagenase digestion (those pesky tight junctions). So after you pass your sample through a cell strainer, there will be a low yield of individual endothelial cells. It could work pretty well for your tumor cells and microglia though.
The myelin is going to mess up your antibody staining. You can get rid of it by putting your homogenate over a 30% Percoll gradient and centrifuging at 1900 x g for 10 min (no brake). The fatty myelin will be at the interface and your cells will be at the bottom. You may have heard of Miltenyi's myelin removal beads. These were absolutely worthless in my hands; don't buy them.
I have also used Miltenyi's papain-based neural tissue dissociation kit (see my linked paper). I did manage to get a few unicellular endothelial cells, but again, the yield is not so good.
Do you really need to sort endothelial cells from the tumor? Can you get away with isolating the blood vessel fragments (i.e. not unicellular and may have a little bit of contamination with leukocytes)? I have done a whole lot of isolating brain microvessels (see linked paper) and you may wish to see if the protocol works on your tumors. You can also make the endothelial cells grow (onto a collagen-coated plate) from the brain microvessels, using puromycin to kill off the other cell types. It wasn't really in the paper, but I could get unicellular endothelial cells with good yield by digesting the isolated microvessels with Liberase. This was not good for viability though, but it could be OK if you just want to extract RNA afterwards.
http://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1004963
Thank you very much for your detailed answer!
Since I want to isolate RNA from the endothelial cells, a pure culture would be advantageous. Unfortunately puromycin selection is out of the question because we don't use puromycin resistent cells. I guess I will stick to sorting, since in tumors there is extensive neo-angiogenesis I would expect a higher yield of endothelial cells compared to your model.
Thanks again for your valuable input.
Hi Timo,
Actually, brain endothelial cells are naturally puromycin-resistant because of high P-glycoprotein expression. But it sounds like culturing for a week would change the mRNA profile of your cells anyway.
I think Liberase would work the best for you (It's mostly highly purified collagenase, with some dispase to break down the tight junctions). I used Liberase DH. If you google for liberase + brain tumor, you will find some citations.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology