Weird staining failure

[Ofir]

Here is a bit of a mystery... I ran an experiment in which staining worked fine for some markers and failed completely for others.

Attached is an image of four successful stains (TCRb, CD8, CD11b, and Ly6C) and two failures (CD4 and Ly6G). Top row is gating for live cells.
In my book the following markers did not stain: CD3*, CD4, CD25, NK1.1, F4/80, Ly6G
* CD3 is really a poor antibody so I'm not worried about its performance in this assay.

CyTOF staining problem

Here is what I can state with certainty:

* The antibodies were added for sure.
* The same antibodies worked in both preceding and following experiments.
* Cells were counted and volume adjusted to be appropriate for staining (3x10^6 cells in 150ul).

Also, the researcher who brought the samples reports that these stains worked for him in fluorescent flow in the past.
I suspect the preparation protocol is the culprit, however the researcher says he used the exact same protocol in his fluorescent flow experiments.
Protocol in brief: Cells were prepared from immune follicles in the liver and treated with collagenase IV and DNase I for 30 min at 37C. The reaction was stopped with EDTA and then erythrocytes were lysed.

There is one paper (PMID 8168120) that reports loss of surface markers upon treatment with collagenase IV and DNase, however the protocol used here was shorter.
Could it be that there is something different about CyTOF that makes the staining more sensitive to this protocol?
Hands-on experience would be much appreciated, as well as other ideas on possible culprits.

[hluche]

Hi Ofir,

I doubt the collagenase 4 treatment is responsible for what you see.
Col4 or ColD are equivalent in terms of purity and are with very low trypsin activity.
The trypsin like activity in less fractionated collagenase batch is responsible for this loss of marker expression (molecules get clived and epitope is lost).
We do regular treatment with ColD and do not see an effect in terms of CD4 or CD8 level by conventional flow cytometry (we work in a standardised manner and can trace this up).
In any case in mouse, CD8/anti-CD8 couple is the most sensitive pair to trypsin. Another one is CD11c. I have not seen this effect on Ly6G.

To adress that question, you could test your lot of collagenase to make sure the fractionaction worked OK by the vendor.
Therefore do a comparison of digested using your experimental conditions vs undigested thymocyte preparation, then do a CD4/CD8 staining (in mouse).
Was fixation performed before or after antibody staining ? Because NK11, CD25, F4/80 are really sensitive to PFA fixation (at least by conventional flow).

Best,

Hervé

[mleipold]

HI Ofir,

Just a question for clarification:

You state "The same antibodies worked in both preceding and following experiments."

Do you mean that you had, say, CyTOF experiments A, B, and C, and these two antibodies (CD4 and Ly6G) failed only in B, but worked in A and C, using the same type of cells and the same protocol?

Or that the antibodies worked in experiments A and C, that were done on a different kind of cells (eg, PBMC rather than disaggregated liver as in B)?


Mike

[Ofir]

The reason I suspect the protocol is that indeed the antibodies worked in different protocols (before and after).
As for the cells, the usual lymphocyte suspects - T, B, DC, Macs, Neuts, etc.