Hi!

Look for my colleague Denis Hudrisier on RG, and ask him the question, he's the one.

Cheers

A handbook attached

Simple answer:

T cell: CD3

B cell: B220

NK cell: CD335

Real answer:

All these cells have many subsets based on their maturation, activation, expansion stages. Many times negative markers are as important as positive markers. I strongly suggest that you to pinpoint the cell types you are interested. Is it regulatory T cells, gamma-delta T cells, memory B cells, pre-B cells, NK-T cells etc? Then start looking for specific markers.

Korcan is correct of course. CD16 should be a good NK marker in the mouse (+CD 56 maybe human only?) Bcell CD19, CD3 good pan T cell marker. Then you need to get more specific.

I would choose B220 over CD19 for B cells as CD19 expression not as stable as often believed. Can down regulate with B cell activation.

Thank you all for your help..... Although the detail story is little complicated but just to make it simple, we want to compare the immune cells population in liver of wild type and a gene knockout mice,....our microarray data shows significant changes in many of the CDs expression between the knockout and its littermate control........Thank you again...

B220 is expressed on some NK cells. Therefore CD19 is a more reliable marker for B cells.

But double/triple staining sorts this?

Just to clarify a bit. It is important to know your markers! You may know this already but just in case: B220 is an isoform of CD45 that is expressed on all murine B cells (but interestingly only a subset of human B cells). CD45 is a protein tyrosine phosphatase with multiple isoforms that regulates activation through a range of lymphocyte cell surface receptors including the antigen receptor on T and B cells and cytokine receptors. So it is present on lots of different cell types. The B220 epitope (I say epitope because of course it is only one epitope on this isoform of CD45) appears to be fairly B cell specific in mice. I only say this because with this background knowledge you can interpret your flow cytometry more accurately. The 'real' specificity of any antibody should always be questioned. Given the biology why are anti-B220 antibodies B cell specific? Interestingly only a subset of human B cells express the B220 epitope (CD27 neg I think?). By knowjng the biology you won't make mistakes. If it really matters to you to be 100% sure what cells you are looking at you almost always need to use more than one marker. My CD19 comment was only because (many years ago!) I was caught out by this. Cross-linking of MHC class II on murine B cells downregulates CD19. In this instance I was saved by double staining for B220 demonstrating CD19 neg B220 pos B cells with the CD19 upregulating again over 24h following class II cross linking. I don't know the specifics of your cell populations but commonly used markers (both to positively identify and also exclude other cells) in this sort of work are CD11c, Gr-1, CD204, CD16, CD49b and of course CD3, CD4, CD8, CD19 and B220. A final thought. Non-specific staining can be a real problem with some of these cell types as can autofluorescence if activated so be very careful with your controls in particular ensuring you use isotype matched antibodies.

Hope all goes well!

It is just personal experience. Not interesting enough in it's own right to publish. But data was discussed at length with Doug Fearon, who was working in Cambridge at the time, so I am 100% confident of findings. We looked in detail at whether this was a 'staining' artefact given that CD19 associates with class II and so potentially CD19 could be caught up in class II antibody complexes with inhibition of anti-CD19 staining without loss from cell surface. This possibility was addressed definitively. I will see if I can find some of the original data. I think this only matters if cells are being treated in vivo with molecules that might affect phenotype of B cells. In most instances will be irrelevant but our findings were a complete surprise so if you can double stain (and I can't imagine why not, as still lots of channels free) then why not be 100% sure.

So in response to Dominic's question please see attached PDF. This work was looking at effects of anti-class II antibody on APC function. It was essential to find antibodies that did not deplete target cells as this is a pretty uninteresting explanation for their actions in vivo. This work taught me to be wary of cell lineage markers. The attached histograms are for splenocytes stained with anti-CD19 2h after in vivo administration of 1mg of various anti-class II antibodies. All staining on ice with azide and HINRS so changes are not staining artefacts. Can't find all of the data but this was looked at extensively. There is no B cell depletion at this time point with these antibodies as shown by B220 and IgM staining (not shown). CD19 clearly down regulates as a result of treatment with some antibodies. Interestingly at 5 days post treatment B cells upregulate class II but CD19 stays low. I don't know what Dominic's experiments are but this just sounds a warning re stability of these markers that we all use so I would always recommend double staining if these considerations are crucial for your experiments.

CD45 (B220) is upregulated in activated in CD4+ murine T cells and therefore not a good B cell lineage marker. 

Hi, for immunophenotyping of mouse-blood it is useful to use an anti-CD45 for characterisation of all leukocytes, and then you might make the following combinations using always two different antibodies to characterise one cell lineage. This helps the identification even for sub-optimal stainings.

T cells: CD3 + CD5, B cells: CD19+B220, Granulocytes: CD11b+Ly6G, NK cells: NK1.1 (only certain mouse strains, among others: B6) + NKp46 (comment: don't use DX5 for blood staining!), monocytes: CD11b+ (non NK cells and non granulocytes) either Ly6C+ or Ly6C- . As you can see: with a 10 color-cytometer you can go for all these markers.