Hi again Andrej,

I agree that the isotype control will not solve your problems, but may still be useful when you are comparing different cell types (or stimulated vs non-stimulated cells) that may differ in their ability to bind antibodies. I would also use area, as this to some extent takt into account both density of antigen and cells size. You should also think about how you calculate the MFI. In general, median is likely the best as this is least likely to be influenced by outlayers or bi-modal binding to the cells.

In flow cytometry , there will in general be be a linear relationship between number of antibodies on a cell and the calculated fluorescence value for that cell, at least in the middle of your flow plots. Your cells are most likely to have the same amount of background autofluorescence regardless if there are any antibody binding or not, resulting in the same increase in MFI from a given number.

With these assumptions, I think that the delta MFI (i.e. the stained MFI - the unstained MFI) would be the most accurate to use rather than fold change. To be sure that your results actually represent true antigen binding I would use values for isotype MFI rather than unstained MFI as the background binding to antibody could change between stimulated vs non-stimulated cells. There are, however, many caveats (such as linearity), especially if you are dealing with small differences in MFI between your two conditions or between the autofluorescence and stained samples. If so, you should probably onside using a stronger fluorochrome (ideally PE).

If you have a clear increase in expression in one condition but not the other, you may consider using the proportion of cells that become positive after staining as a measurement - i.e. using an FMO sample to set a gate so that only 5% or 1% of the FMO cells are positive , and then, for the same cell condition, determine the proportion of cells that are positive after staining. Note - this may not actually be the number of cells that are in fact positive for expression, as most antigens display logarthmic expression in flow, and it is likely that the some of the weakest stained cells may overlap with the cells that show most autofluorescence.

If both your populations are clearly positive after staining, the proportion of positive cells will of course not help, since 100% of the cells are positive in both cases. On the other hand, then your MFI would be much higher than the background values for any of the populations (i.e. 10 fold higher than autofluorescence on any of the populations). Then, the positive MFI value will mostly represent antibody binding and you may not need to think about the background at all.

Best

Hello,

Do you use unstained cells as negative control or do you have a true isoform control? Does your treatment change the shape of your cells?

With a negative isoform control you may not have the same discrepancy between your samples.

Dear you should have a good control.

Where do you know that the autofluorescence of the treatment and control groups is different? Autofluorescence can be ruled out by blank control and unstained control.

Hi Andrej,

This is a common issue in flow cytometry, but the solution is not difficult. What Christine said is totally correct, I will just expand on her comments: I think you should use the same cells (ideally a mix of both untreated and treated cells) fr your staining controls, in that way you will have autofluorescence incorporated into the compensation calculations for flow. On the other hand, I would also have an isotype control mAb staining for your CD14, with all the rest of the mAbs and fluorophores exactly as in your sample to analyze. Lastly, I would also strongly recommend switching your PerCP for another fluorophore, as PerCP might not be ideal in the sense of overlapping signal with other fluorophores (depending on your mix of mAbs-fluorophores).

Viele Gruesse und ganz viel Glueck mit deinem Forschung!

Much better than isotype control is isoclonic control. And FMO, of course

What Christine said is totally correct, you should have a good control.

Dear Andrej,

This is a rather tricky question, and as some has pointed out you are also a bit unclear in what you are asking and what control was included.

In principle the answer is easy - the increase in fluorescence between a stained and non-stained cell population is due to increased binding of the antibody so it would be correct to use the increase in MFI as measurement of antibody binding to the cells for both populations.

However, there are technical concerns with this simplified view. To start with - what are your MFI values based on - height, width or area of the fluorescence? It is not absolutely clear which of these is best in your case, but if you are working with relatively large cells or cells that may differ in size I guess that the area measurement would best correspond to the total binding of antibody to your cells.

Secondly, the amount of antibody bound will be influenced by both molecules that bind antibodies specifically but are not antigens (i.e. Fc receptors on monocytes) or changes in the structure of the cell surface due to activation that may give unspecific binding. These may be controlled by blocking Fc receptors with specific antibodies and incubate the cells with serum before staining may help. However, as pointed out above, for this particular experiment it is also important to include both an FMO control (to ensure that no staining is observed in the absence of your specific antibody) and an isotype control as closely matched to the antibody used for detection (to control for binding to FcR and unspecific binding) should also be included.

Hi Andrej

autofluorescence is a fact we have to deal with

in your case :

1) I would use a channel were autofluo is low, if I remember well perCP is the worse for mono macro, I would have used far red or UV and efficient blocking to avoid FcR ect...

2) You can try to express the results as fold increase/decrease over the control (unstained or isotype control if you have a good one : MFI(CD14)/MFI(unstained)) for each treatment and no treatment

best

Christine Pich-Bavastro and Damian Maseda thank you for your answers. I use unstained control and Fc-Blocking Human IgG instead of isotype control as suggested in this publication: https://doi.org/10.1002/cyto.a.22995.

I could try using isotpye control but I have the feeling that this won't solve the problem I have which is high autofluerescence of my unstained controls. The compensation was actually done with beads not cells in order to get strong fluorescence signal. As far as I know that is the best way to do compensation.

Eric Espinosa thank you for your suggestion. Actually I only mentioned PerCP to give an example, but I observe the same phenomenon also in FITC, PE, APC, APC-Cy7 and DAPI channel. As for calculating the ratio of MFI(CD14)/MFI/(unstained): I have tried that and the results are the exact opposite of th "raw" MFI (e.g. untreated cells have higher CD14 expression when using war MFI whereas CD14 expression is lower in untreated cells when I do the calculation you suggest.) My problem is I don't know which of the two options displays the truth.

Mats Bemark thank you for your answer. I am aware of the problem of unspecific binding and Fc-Receptors which is why I use Fc-blocking agent. Apart from that I don't think that unspecific binding is the problem in my experiment but rather the vastly different autofluorescence (fluorescence of my unstained cells) between treated and untreated cells.

You state that the increase in fluorescence between a stained and non-stained cell population is due to increased binding of the antibody so it would be correct to use the increase in MFI . How would you calculate this increase? I do use area measurement. If something still remains unclear please ask.