with Dr. John Freed, she isolated the peptides bound 
to class II major histocompatibilty complex (MHC) 
proteins on spleen and thymus. Although many were 
expressed in both tissues, some were present only in 
spleen. These peptides come from proteins ex- 
pected to be at much higher concentration in the 
spleen than in the thymus because their expression 
is limited to B cells. Tolerance to this type of pep- 
tide is presumably not governed by deletion in the 
thymus, but rather by events in the periphery. 
In the past Dr. Marrack's group and others demon- 
strated that mature T cells can become inactive if 
they contact peripheral antigens. This has been 
thought to be a result of exposure to massive doses 
of antigen that first stimulate the cells and then 
kill or inactivate them — a process called clonal 
exhaustion. 
Experiments suggest, however, that this is not the 
only mechanism for induction of tolerance by pe- 
ripheral antigens. In mice a retrovirus transmitted in 
milk, a mouse mammary tumor virus (MMTV), en- 
codes a superantigen that reacts with most T cells 
bearing V^l4. The mice are infected at birth, but 
subsequently the T cells disappear. Surpris- 
ingly, they disappear very slowly, with deletion 
complete when the mice are about 6 months old. 
There is no obvious indication that deletion is pre- 
ceded by activation. 
Although this is an interesting model for tolerance 
induction, it has its limitations. For example, little is 
known about the timing or location of expression of 
the viral superantigen. To deal with such problems, 
Dr. Marrack and her co-workers (partially supported 
by the National Institute of Allergy and Infectious 
Diseases, National Institutes of Health) have devel- 
oped a different model. Adult mice are challenged 
every other day with minute doses of a bacterial su- 
perantigen, staphylococcal enterotoxin A (SEA). 
Chronic exposure to small amounts of SEA causes 
rapid and complete disappearance of mature target 
T cells, in this case bearing V|S3- 
Overall, these experiments show that tolerance 
can result from chronic exposure of mature T cells 
to small amounts of antigen. The work is leading to 
the suggestion that T cells are so programmed that 
they usually die or become inactive when they en- 
counter antigen. It is only under unusual circum- 
stances, when antigens are expressed in conjunction 
with some infectious agent, that T cells go into ac- 
tive response. 
Positive selection remains an unsolved mystery. It 
involves contact between the receptor on develop- 
ing thymocytes and MHC expressed on thymus cor- 
tical epithelial cells. Tolerance to MHC is also in- 
duced during thymocyte development, and it is not 
clear how self tolerance and positive selection can 
be reconciled. As mentioned above, thymocytes 
probably die when their receptors engage self be- 
cause engagement induces high intracellular Ca^"*" 
concentrations. However, Dr. Marrack's group has 
recently shown that when receptors are first ex- 
pressed on thymocytes, they do not associate com- 
pletely with CD3, the protein normally coexpressed 
with the receptor. Consequently, receptor engage- 
ment does not increase Ca^^ levels inside the cell. 
The group therefore suggested that positive selec- 
tion may occur during this early stage of the thymo- 
cyte life history, at a time when receptor engage- 
ment cannot cause death. 
It is clear that the receptor genes themselves af- 
fect the repertoire of T cell receptors. It is difficult 
to study such effects in the absence of other phenom- 
ena such as positive and negative selection. Since 
MMTV-associated superantigens have such a major 
effect on the T cell repertoire, a mouse strain of 
known MHC type and lacking all endogenous 
MMTVs is being developed for study. This is a 
lengthy process. Preliminary examination of several 
animals that are heterozygous for single MMTV inte- 
grants suggests that T cell use of V|8s is very uneven. 
For example, V/39 is rarely used on T cells, and V/32 
is used frequently. Similar findings apply in hu- 
mans; for example, V/310 is used rarely and V/313-1 
is used often. Such differences are probably due to 
the frequency with which the various V/? genes are 
rearranged. 
By and large, the repertoire of receptors on CD4- 
bearing T cells is very similar on young and old ani- 
mals. Surprisingly, this is not true for CD8-bearing T 
cells. As mice reach old age — 18 months or more — 
spikes appear in the repertoires of their CD8^ cells. 
For example, as many as 75% of the CD8^ T cells in 
old mice may bear a single V/?. The spiking cells do 
not appear to be leukemic. Currently the stimulat- 
ing agent for these cells is being sought. Perhaps it is 
a tumor antigen. 
Dr. Marrack is also a member of the Depart- 
ment of Medicine at the National Jewish Center 
for Immunology and Respiratory Medicine, 
Denver, and Professor of Biochemistry, Biophys- 
ics and Genetics, of Microbiology and Immunol- 
ogy, and of Medicine at the University of Colorado 
Health Sciences Center, Denver. 
Articles 
Choi, Y., Kotzin, B., Lafferty, J., White, J., Pigeon, 
M., Kubo, R., Kappler, J., and Marracli, P. 1991 ■ 
A method for production of antibodies to human 
348 
