 |
 |
 |
 |
> Related to the discussion on fluorescences beyond saturation, could anyone
> give me a vague idea of the typical concentration range over which
> directly labelled abs exhibit saturation? Or is this extremely variable
> depending on Ab type, cells, etc?
1. First, what does 'saturation' really mean?
In our hands, the one-site binding equation describes quite well the
binding of monoclonal antibodies to cell surface receptors. This model
predicts that 80% saturation occurs when the antibody concentration is 4
times the concentration which gives 50% saturation, that is, 4Ch (Ch
representing Concentration-giving-Half-saturation). 90% saturation would
require 9Ch, and 95%, 19Ch. Thus, to increase the intensity from 80% to
95% (a puny 19% increase), you have to use nearly 500% (5 times) more
antibody. 99% (a raging 4% more intensity) would require 99Ch, or 25 times
more antibody than 80%!
Therefore, when people say in their published methods that 'all antibodies
were used at saturation', I'm amazed at how much money they have to spend
on antibodies :-). Seriously, I construe that to mean that the antibodies
were used to within experimental error of saturation; within 5% of
saturation, shall we say -- i.e. at least 95% saturation.
2. Am I ever going to answer Kris' question? What concentrations of
antibodies 'saturate'?
In our hands, the Ch for several commercial monoclonal IgG antibodies (15
min incubation) is 0.1-0.2 micrograms/milliliter. It follows that 95%
saturation will occur at about 2-4 ug specific IgG/ml. Antibodies we have
prepared ourselves, undoubtedly less pure and likely partially inactive,
have Ch's rarely less than 1 ug/ml. These data are mostly for the first
antibody in indirect staining, but I don't see any reason why directly
conjugated antibodies should differ. For most antibodies and most
cell types, these values are fairly constant in our hands.
3. Is saturation really necessary? (Hardly ever!)
For subpopulation analyses, the fluorescence intensity (FI) is not usually
as important as the percentage positive cells. For other questions,
such as determining the relative densities of a receptor on different
cell lines, the FI does matter. There, we want the FI to be proportional
to the number of receptor epitopes on the cell surface.
If antibody binding is limited by diffusion (the rate at which antibody
molecules bump into the cell surface), then the FI will be directly
proportional to the epitope density. This will be true, I assert, even if
the antibody is employed at only 1% of saturation (but this will work only
if it still has a signal well above the blank). In the case of mouse Thy-1
(about one million receptors per T lymphocyte), the saturation FI signal to
noise ratio for indirect staining is several thousand, so one can actually
work at 1% of saturation and still have signal/noise around 10 (and get
a lot less aggregation!).
Only when binding is amount-limited might FI not be proportional to epitope
density. This means that the higher density cell lines are binding a
significant portion of the total antibody provided, hence they will achieve
a lower % saturation than the lower density cell lines.
One less-than-ideal experiment we did suggested that when antibody was
applied for 15 min in a fixed volume of 100 microliters per tube, binding
was diffusion-limited for up to 10 to the 7th (10e7) cells per tube, for
receptors of moderate density (about 10e5/cell). Binding became
amount-limited for high-density receptors (10e6/cell) at and above 10e6
cells/tube(*). Since we routinely use 3 x 10e5 cells per tube for flow
analyses, binding should always be diffusion-limited. Therefore, I
tentatively conclude that we need not employ antibodies near saturation
when signal/noise permits.
4. What is the equation? And what about polyclonal antibodies?
If I is the observed fluorescence intensity (corrected for blank
fluorescence), Imax is the true saturation plateau intensity, C is the
concentration of antibody employed, and Ch is the concentration of antibody
which half-saturates the cell surface,
I/Imax = C/(C + Ch), therefore:
C %Sat'n
---------------
0.1Ch 9%
0.5Ch 33%
1Ch 50% (by definition)
2Ch 67%
4Ch 80%
9Ch 90%
19Ch 95%
99Ch 99%
You can use any of several scientific plotting programs (e.g. Prism, Fig P)
to fit this equation to your titration data, thereby objectively determining
Ch and Imax.
Most commercial FITC or PE conjugates of polyclonal goat, rabbit, etc.
antibodies do not fit this equation. This is because as the concentration
increases, antibodies present at lower concentrations in the mixture
begin to contribute significant FI. So, as you keep increasing the
concentration, FI keeps increasing -- there is no true plateau! Since
you can't really 'saturate' these antibodies, it is especially good
news that you don't really need to!
5. Did I have a question (or several)?
Yes, and thank you for reading this far along. I've never seen a detailed,
authoritative discussion of these issues (one-site binding, diffusion- vs.
amount-limited binding, 'saturation'), in the context of flow cytometry, in
print. Please let me know where if you have. Also, if you know of any
data on the question of diffusion vs. amount-limited binding, please
share. And of course, if you think anything I've said is wrong (it's
happened before, as recently as ... today), please explain why.
*In our experiment, we determined the Ch by a titration at each cell
density. The density at which Ch began to increase was the density
at which amount-limited binding began to occur.
Sorry, Kris -- I've been wanting to have a discussion about these issues
for a long time.
/*- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Eric Martz, Professor of Immunology emartz@microbio.umass.edu
Dept Microbiology Voice: 413-545-2325 FAX: 413-545-1578
Morrill IVN 203, Box 35720, Univ Massachusetts, Amherst MA 01003-5720
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -*/
|
|
|
|