Ephraim,
It strikes me that your arguments really stand or fall on the
question of whether or not B cells tolerise naive T cells. As for
activated B cells (LPS activated), both Anne O'Garra's and Kim Bottomley's
labs have data showing that B cells CAN prime naive (L-selectin hi, small,
dense) T cells.
For resting B cells, the situation is less clear, but I would like
to suggest to you that (a) it is unlikely that a resting B cell would meet
an idiotype specific naive T cell and (b) even if they did meet, the
interaction does not necessarily result in the tolerance of the T cell,
but the T cell may "ignore" the B cell.
Now, what are my reservations. Firstly, this brings up a question
first posed, I think, by Peter Medawar, which is what is the probability
of a chance encounter between an antigen specific T cell and an
appropriate B cell (in his case "appropriate" meant specific for the same
antigen, but here we can take it to mean "presenting a crossreactive
idiotope"). In the absence of antigen we can't invoke the original
arguments about lymphoid organs, as the resting cells will recirculate
through these. So how many T cells will each B cell encounter? If
we make some back-of-the-envelope assumptions (life span of B cells
is 6 weeks; average time to encounter a T cell and be scanned for
the presence of antigen is 1 minute - and I believe this is a GROSS
underestimate) then the average B cell will encounter 6x10^4 T cells,
somewhere around 0.001% of the total T cell number. So it is by no means
certain that the appropriate B and T cells would ever get together.
Secondly, peripheral tolerance does not delete T cells, nor is it
permanent - tolerised T cells do recover responsiveness over time.
Since B cells (in the absence of activation) have a half life of
about 3 weeks, the proportion of the repertoire that is tolerised
will change over time. Now if we assume that we need two anchor
positions with a nonamer to make an MHC-binding epitope, the potential
repertoire of epitopes is of the order of 10^17. The total number of
B cells at any one time is of the order of 2x10^9, and for T cells
is about 10^10. Ok, ok - I know that many of the peptides will
be indistinguishable to T cells, but with only 4 TCR contact
residues the potential repertoire is still 1000-fold greater than
the total number of B cells. So B cells are likely at best to
tolerise no more than 0.1% of the repertoire at any one time, and
that only transiently.
Thirdly, how efficiently does a resting B cell present self-Ig peptides?
The level of class II expressed by resting B cells is relatively low.
In the absence of antigen, my guess is that the recycling rate of the cell
surface Ig is also relatively low and that the density of Ig-derived
peptides will also be low. Since T cell recognition does not appear
to be an all-or-nothing phenomenon, but dependent both on ligand
density and receptor affinity, I would suggest that most resting B cells
are unable to signal T cells at all, and are ignored by the immune system.
We (well, someone) can do an experiment to test this out - or it may well
have been done already. Antonio Lanzavecchia and Colin Watts isolated
human T cells specific for murine Ig peptides (though in this case I
think it was C region peptides - nevertheless the principle still
applies). I don't recall what the restriction element was, but should
a transgenic exist with that restriction element, one could ask whether
resting transgenic B cells could be recognised. It's not a perfect
experiment, but it would be interesting to see what happens.
Alternatively, since Diane Mathis and Christophe Benoist have
identified regulatory elements that control expression of
class II in B cells, by crossing an A knockout with an Ea transgenic
(on b or s background) one could in principle make a mouse that
was A-E+ except for its B cells, which would be A-E-. I can't
remember whether they have made such a mouse, but how would you
expect its T cells to behave towards, say, resting wild type (E+)
B cells vs LPS or Ag stimulated?
Paul
--
Paul J Travers phone : +44-(0)71-631-6862 (office)
ICRF Structural Biology Unit " " " 6868 (lab)
Birkbeck College fax : +44-(0)71-631-6803
Malet Street
London WC1E 7HX email : p_travers at icrf.icnet.uk
England or : paul at histo.cryst.bbk.ac.uk
