 |
 |
 |
 |
1) Establish the absolute quantity of monoclonal antibody present. This is
often provided as mg/ml of antibody protein by the manufacturer, or you can
measure it yourself. The antibody should be pure.
2) Knowing what isotype of antibody you have you can calculate the molar
amount of antibody, the absolute number of antibody molecules present per
ml of solution at a given dilution.
3) You can then do staining of cells with antibody at a limiting dilution,
i.e. at a dilution at which all the antibody is "used" up by binding to the
cells. Preferably you should do a number of dilutions, all of which are at
dilutions in which all the antibody is used up. Of course technically there
is never a condition in which there is absolutely NO free antibody left, no
matter how long one incubates or how strong the affinity of the antibody is
since the affinity constant is always less than infinity, however with
sufficiently long incubations or sufficiently strong affinity antibody the
residual free antibody can be negligibly small amount.
Free MoAb + Free Cells -----------> MoAb=Cells
K
With a K that is large enough the ratio of MoAb=Cells to
[FreeMoAb][FreeCells] can be very high, and by increasing the number of
free cells it is also possible to "use up" all of the antibody.
3) Now you know that all the antibody originally free in your solution is
on your cells. Assume that this antibody is evenly distributed over all
cells, and that each cell has an equivalent number of ligand molecules (or
that the distribution is essentially normal with a representative mean).
You do know exactly how many cells you had in your assay tube. Now just
divide the moles of antibody (which you are assuming are on the cells) by
the number of cells, and you have the number of moles of antibody bound per
cell. If the antibody and ligand are interacting in a Mole :: Mole way then
you have the number of moles of ligand per cell. VOILA!!!
The problems with this method have to do with some of the assumptions,
which need to be proven true. Lack of bivalent or multivalent binding, very
wide F/P distributions, poor antibody affinity, nonspecific binding etc.
can all interfere with the calculations, but this is true with ANY
quantitative antibody method.
It would also be desirable to get a nomogram for cells which have varying
numbers of ligand molecules per cell so that the fluorescence that
corresponds to differing amounts of antibody can be assessed beyond the
range provided by amounts of antibody that are insufficient to saturate a
normal cell. For HLA antigens, for example (which are codominantly
expressed), allele specific antibodies may detect different amounts of
ligand in a homozygote than a heterozygote, and therefore the correlation
of fluorescence intensity to molar amount of antibody bound (allele
specific monoclonal) can be determined over a two-fold range beyond the
heterozygote.
The other theoretical way of handling the problem would be with some kind
of reference beads with ligand on them, but I am not sure that the
fluorescence characteristics would necessarily be the same as with real
cells.
Edgar.
Edgar L. Milford, M.D.
Brigham and Women's Hospital
Tissue Typing Laboratory
75 Francis Street, Boston, MA, 02115
Tel: 617 - 732-5872
Fax: 617 - 566-6176
|
|
|
|