Re: Relative Fluorescence Intensity (RFI). What is that? HELP!

Ernesto, Hi!
I had myself questions similar to yours not so long ago, so let me try to explain what I understand is involved in your question. I am sure other people will correct me if I get something wrong. What you want to do is OK and commonly used, but you have to take care to set up and/or verify important points in the whole experimental system. How rigorous you want to be is up to you, and that'll determine how good are your results. I'll try to address the stuff in points to "bear in mind":

1) The flow cytometer (FC) is a machine without an internal standard. That is why beads are used to calibrate and set up the machine, verify it is working as it should. And that is why you need an external standard to do quantitative or semiquantitative work (beads with known MESF, beads that capture Abs, etc). 
2) Biological samples are not beads. What you are using for controls are biological controls, not fluorescence controls. Depending on your source for cells, they might be more or less heterogeneous from experiment to experiment, and the auto-fluorescence you'll get from them can change due to your treatments or due to things you do not control for. The specific fluorescence you'll get from stained cells can change as well due to Ab and fluorochrome degradation with shelf time, working at non-saturating conditions (where technical preparation errors can lead to artificial changes in fluorescence readings), changing Ab batch or brand, or a change in the protocol.
3) Machine working properly: when doing those ratios you are assuming the machine works linearly in the range you are using for measuring, and I understand that people have found that is not always the case, specially with older machines and analogical log amplifiers. Moreover, the low end of the scale where you are measuring the unstained/isotype is the noisiest.
4) Isotype controls: although usually used my most (including me), they have their caveats so take care to know what they represent.
5) Effect of necrotic and apoptotic cells: In my "professional" opinion, if I may already have one, judging from published literature this is one of the most overlooked aspects when using flow cytometry to measure changes in marker expression. Necrotic and apoptotic cells can have different autofluorescence, bind Abs unspecifically to different extents and change the expression of the marker in question during their death process. Moreover, these processes can affect different populations to different extents further increasing the possibility for bias in certain models. Maybe you are already doing this but I must say that the accounting for apo and nec is necessary, (like a necrotic exclusion dye like PI) regardless of what is it that you are measuring. A 10% decrease in marker expression could be due to the downregulating effect of your chemical or due to the toxicity of the chemical.
 
After all this, what is the bottom line, can it be done? For me, it is more or less as follows: make sure the machine is working well, ask the person in charge what kind of QC he/she uses to make sure the machine is calibrated and maintained well. Some kind of linearity confirmation is needed, and if you can determine where is the linear range stay there. Titrate all the Abs used for (semi)quantitative measurements, verify if you have Fc receptors intervening in the Ab binding. Stain from a prepared stock and not from the Ab bottle directly to ensure equal (and correct) amounts of Ab added to all treatments. Clean the machine well before using it, and if you have PI and non-PI samples (or PI "similars" like 7AAD etc) try to cluster them together and wash a little bit before putting a non-PI sample after a PI one (this might be a little paranoid, but I've had bad experiences already, maybe some of them due to lack of experience!). Use a necrotic exclusion or some other way of accounting for death processes. Bear in mind all the time that isotype controls and unstained controls are not constant from experiment to experiment but rather living (fixed?) cells, with all the changes that come with them. Add isotype in the same amount as specific Ab. Finally, I use median fluorescence and not mean fluorescence as the principal measure.

I am doing a lot of my work using this approach and it certainly can work very well, within its confines of course. You ask to measure protein expression, so you have to decide how quantitative you want to get. Knowing the "high end" of the quantitation procedures puts things in perspective. There are more advanced techniques for increasing the "quantitativeness" of the measurements as well as their accuracy (the precision is up to the machine and to you). The next step up is using MESF (molecules of equivalent soluble fluorochrome) beads, they have many advantages such as confirmation of linearity, semi (or relative) quantitation (using linear regression analysis), allow you to see visually if the machine is OK every time you do your experiment (in rough terms at least), they are an external standard to which to compare successive experiments. Some people use a single-peak MESF bead which does some of what I mentioned, but in my opinion if you are already running beads for MESF calculations a multi-peak mixture is much better for almost the same price (of course single peak beads have their uses for other things) If you want to go higher, there are beads with known number of actual fluorochromes instead of MESF, and also beads with known Ab binding sites. 

This email is surely "a long answer to a short question", but I hope I've helped you!

Regards,

Uriel.

Uriel Trahtemberg
MD/PhD student
The Laboratory for Cellular and Molecular Immunology
The Hebrew University - Hadassah Medical Organization
Jerusalem - ISRAEL 

"Chance favors the prepared mind"
	    Pasteur

----- Original Message ----- 
From: E.Oviedo-Orta 
To: cyto-inbox
Sent: Thursday, September 08, 2005 10:31 PM
Subject: Relative Fluorescence Intensity (RFI). What is that? HELP!


Dear all,

We're trying to measure protein expression by FACS. We're staining for CD86 and CD54 to measure the expression of the corresponding proteins in cell in culture in response to a drug treatment.

We've found that we could calculate the ratio of expression by using something I have never heard of before: The Relative Fluorescence Intensity (sorry for the ignorance).

For this we've found the following formula: The mean fluorescence intensity (MFI) of chemical treated cells minus MFI of chemical-treated isotype control cells divided by MFI of vehicle control cells minus MFI of vehicle isotype control cells.

Is this an appropriate method to use for what they want?

Is the formula correct?

I would really appreciate your help.

Regards,

Ernesto.