I read your message with interest and would like to receive Part B.
>From the viewpoint of a binding study, I think the generalised analysis we
did in Ron Chatelier's paper in the EMBO J still raises the question of
what happens when there are several different cell-types which occur within
a mixture such as whole blood, each of which has (different?) binding
specificity for the target antigen, as applies with all of the non-lineage
specific antibodies.
This is why the initial process needs to be done with "pure"(?) cell lines,
which express just a single class of specific binding sites; of course,
this needs to be demonstrated via proper binding curves (best done by the
isoparametric method I would say).
I don't accept your enthronement of the radio method as the gold standard,
because of its bulk reporting response and because of the influence of
dead-cell uptake.
The gold standard will be fluorescent ligands with defined univalent
antigen-specific binding, used initially via an isoparametric study and
calibrated against a series of bi-valent Mabs, eg those offered by Dako, BC
and BD.
I am intersted to hear more on QIFI though.
Bob
-----Original Message-----
From: Michel Canton [SMTP:mcanton@wcube.fr]
Sent: Wednesday, September 10, 1997 3:37 PM
To: Cytometry Mailing List
Subject: Re: Cell surface receptor
Dear group,
I fully agree with Bob Ashcroft's answer to this very interesting question
regarding the use of antibodies and the concern of monovalent/bivalent
attachment as well as the influence of affinity.
The most recognized and popular method of measuring cell-associated
receptor
numbers is radio-binding. This method may be considered as a gold standard
for any other method.
Flow cytometry (FCM) methods presents many advantages over radio-binding
(and in addition, we all like it so much!). This is Titus et al. ((1981),
PNAS, 78, p519) measured Fc receptors on human cells with FCM after
calibrating their assay using radio-labeled aggregated IgG. They
demonstrated a linear relationship between MFI (corrected from background)
and the number of cell bound IgG molecules as determined by radio-binding.
By simply applying the same principle with anti-CD5 MAbs, we designed the
QIFI technology (quantitative indirect fluorescence immuno assay) that
proved to be valid for any mouse IgG Mab (J.Immunol.Methods, 85: 65-74,
1985). This method is now being fully exploited in application kits.
Initially CEM sub-clones with known CD5 Mab binding capacity were used as
reference calibration standards for quantitative FCM (Leucocytes Typing II,
vol.2, 329-343). Later on non fluorescent latex beads covered with mouse
IgG
were calibrated towards our cellular standards. Such beads still serve as
pseudo-cells covered with known numbers of IgG Mab to relate MFI to Mab
binding, i.e., number of bound Mab molecules per cell.
Infering antigen density from Mab binding is a more critical approach as
underlined by Bob Ashcroft. We experienced various situations:
1- CD5 T101 Mab had proved in our hands to monovalently bind to CEM cells
under saturating conditions as checked by radio-binding using both IgG and
Fab fragments.
2- During the 2nd International Workshop we found that a majority of
different CD8 Mabs revealed about 130,000 sites per CD8+ lymphocytes
whereas
few others revealed only half this number , i.e., about 65,000 sites. In
fact each CD8 complex consists in one alpha-beta heterodimer that could
express an epitope present on both chains. Each CD8 complex binds either 2
Mab molecules or only one depending on the specificity of the Mab.
3- A recent study from Wagner et al. (Blood, 88:907, 1996) showed that even
when the target antigen is expressed at a very high density and
inter-molecule distance is shorter than the size of a F(ab')2 fragment,
divalent binding occurs even under conditions of entibody excess. This is
demonstrated for the GpIIa-IIIa complex on the platelet surface.
Although monovalent binding could be taken as a rule when saturating
concentration is used there may be exceptions that one needs to know to be
able to measure antigenic sites more appropriately (use of Fab fragment).
Under the monovalent binding assumption we published some numbers for major
CDs on human leucocytes which could (?) serve as a basis for further
studies
(Cytometry , CCC, 26: 137, 1996). Some of these numbers have been, at
least,
rather consistent over time (years), when using various cytometers and also
between different laboratories. There are for instance about 50,000 CD4
sites/CD4+T lymphocyte (Res. Immunol., 142: 291, 1991) or about 200,000
CD45
sites on normal PBL (Cytometry, 12 suppl. 5, 418A, 1991).
To summarize I just wanted to illustrate that besides fluorescent beads
there are other means (non fluorescent beads covered with mouse IgG or
whatever) that can also be useful and at least very flexible and that in
any
of these different quantitation strategies radio-binding must still remain
the referecence when designing such quantitative flow assays.
If you allow me I shall send you the part B of my answer in few days
focusing on the specific question of quantifying receptors and measuring
receptor occupancy.
Hope this long part A (too long ?) can already help!
Philippe Poncelet, PhD
Scientific Director
BioCytex
140, chemin de l'armee d'afrique
13010 Marseille
France
Tel +33 (0) 4 91 94 29 39
Fax: +33 (0) 4 91 47 24 71
E-mail: use my boss's E-mail: mcanton@wcube.fr