MHC restriction questions

Ian A. York iayork at
Tue Oct 20 14:26:13 EST 1998

In article <v03130303b25292724ea4@[]>,
Klaus D. Elgert <kdelgert at> wrote:
>self-peptides in our MHC molecules.  Why haven't we developed
>mechanisms to display only foreign peptides? Also, if somatic mutation

There are some workers who think there may be mechanisms by which
presentation *is* biased toward foreign peptides.  I find the arguments
(which are mainly theoretical) highly unconvincing, but I won't absolutely
say the conclusion is wrong.

As to reasons why the system has evolved the way it has, I think the main
reason is that identification of foreign vs. self (whether peptides or any
other system) is intrinsically difficult, and therefore costly.  If we're
still thinking of peptides (note that once we start getting hypothetical
like this you can postulate other systems of detecting self/non-self, but
it doesn't change the fundamental problem) it would mean that all peptides
which could be derived from the 50-100,000-odd self proteins must be
somehow tagged as 'self", and tagged in a way--this is crucial--that a
parasite could NEVER evolve similar tags, because that would completely
overcome the discrimination.

With the present system, the difficulty of the task can be estimated from
the cost it imposes.  I don't remember the exact figure, but something
like 99% of T cells are deleted in the thymus.  Even ignoring other layers
of redundant protection, this is a huge cost.  But that cost is limited to
one particular subset of cells.  

In general the cost is going to remain at least roughly similar.  (This is
basically information theory here, eh?)  Using your students' approach,
that huge cost will be shifted from the T cells, to every single nucleated
cell of the body.  Obviously here blithely deleting is not an option;
instead the cost would have to be metabolic. 

I can't even think of a way in which foreign peptides *could* be RELIABLY
screened by individual cells, but evolution has a much better imagination
than I do.  But I do think that there's no way of reducing the cost, and
that shifting the burden of discrimination from a subset of cells to all
cells will greatly increase the cost.

In other words, it works the way it does because that's the best solution.

Ian York (iayork at <>
"Bring back the Vicki Robinson .sig virus!" - Cindy Kandolf

More information about the Immuno mailing list