I'm just adding a little to Deb's comments, supportively.
I want to emphasise that there are several assumptions made about linear
proportionality between cell-associated fluorescences and beads
standards'-associated fluorescences. Each of them needs to be verified.
Thus, a bead which has a known (= measured by another means) number of FITC
molecules attached or embedded will generally not have the same
fluorescence intensity as a cell which has the same number of FITC
molecules attached to it via the binding of FITC-conjugated antibodies. In
intuitive terms this can be seen as being due to the local ionic
microenvironment in which the Fluorescein finds itself being different
in/on the bead from what it is in/on the antibody and the cellular epitopes
to which it binds.
A good starting point is the paper from the EMBO Journal, authored by Ron
Chatelier and a football team: Vol5, #6 (1986) 1181-1186.
To restate a conclusion from that paper, the use of a bead standard "rests
on 2 assumptions. Firstly all bound ligands (read FITC molecules) have
identical quantum yields. ... if this is not the case, the standard is not
a valid calibrator of fluorescence; this can be assessed from the linearity
of the plot referred to below. Secondly, when the first assumption is
verified, the quantum yield of the cell-bound ligands must be the same as
that of the fluorescein residues in the calbration standard. This can be
determined by a comaprison of the plots of total mean fluorescence vs the
binding function for each of the cases, the cell binding and the
calibration standards binding. When they differ, the beads may themselves
be calibrated by the isoparametric method indicated in this paper."
Additional uncertainties arise in the estimation of receptor numbers as
well. The main one is from the ignorance of the actual stoichimetry of
binding which occurs when bivalent antibodies are used to label cell
surfaces: the degree of bivalent attachment (ie both antibody sites bind
two separate CD or receptor sites) relative to univalent attachment (only
one arm of the antibody is bound to one CD or receptor site). A further
issue arises from the discrimination of competable vs non-specific
attachment of ligand.
I guess that at the biological level, we have other sources of variation
which swamp these technical considerations.
The bottom line for me is that any estimation of CD or receptor numbers by
currently employed methods in flow have a precision that is at least worse
than 50%. This is not meant to be inflammatory or judgmental.
Regards, Bob.
-----Original Message-----
From: Deborah Berglund [SMTP:umbbd@gemini.oscs.montana.edu]
Sent: Saturday, August 30, 1997 6:11 AM
To: Cytometry Mailing List
Cc: Cytometry Mailing List
Subject: Re: Cell Surface Receptor Number
You can sort of do this, but you need to be aware that you are measuring
the amount of fluorescence from the label, not necessarily the actual
receptor. Also you need to know how much fluorescence you are getting
from your label. We have come close to an answer by using an
avidin-biotin linkage, with a known number of fluorescein molecules on the
biotin and using calibration particles with known numbers of fluorescein
molecules. Our probe is not an antibody. I would not presume to say that
our numbers are actual, just relative.
Deb Berglund
Montana State University Microbiology
On Thu, 28 Aug 1997, Gilbert A. Radcliff wrote:
>
> Dear Group:
> Could someone briefly explain a method by which one can determine
> cell surface receptor number in a population of cells?
>
> I assume that the intensity of staining is proportional to the
> density of receptors.
> Thank you for your replies!
>
> Richard Liu, Ph.D.
>
> Associate Scientist
> BMT Group
> Moffitt Cancer Center
> Tampa, FL
> 33612
>