that express one rearranged T cell receptor (TCR) 
on all T cells. By crossing TCR transgenic mice with 
mice expressing the antigen, tolerance can be di- 
rectly studied. Dr. Flavell's laboratory described a 
system with simian virus 40 (SV40) T antigen (SV-T) 
as tissue-specific antigen and an TCR from a cyto- 
toxic T lymphocyte (CTL) that recognizes T antigen. 
When antigen was expressed postnatally, spontane- 
ous autoimmunity resulted in double (TCR plus an- 
tigen) transgenic mice. A working hypothesis has 
been that the inflammatory effects of SV-T (e.g., tis- 
sue hyperplasia, necrosis) initiate this autoimmu- 
nity, causing presentation of SV-T to nontolerant T 
cells. 
To dissect the role of inflammation in autoimmu- 
nity, Dr. Flavell's laboratory created a chronic tis- 
sue-specific inflammatory state by directing the syn- 
thesis of the inflammatory cytokines tumor necrosis 
factor-a (TNF-a) and TNF-13 to the islets of Langer- 
hans in transgenic mice. Production of either cyto- 
kine is sufficient to cause a massive lymphocytic in- 
flammatory infiltrate. Inflammation does not lead to 
autoimmunity, since these mice never develop dia- 
betes. An inflammatory response seems therefore 
not to be a sufficient condition for autoimmunity. (A 
grant from the National Institutes of Health pro- 
vided partial support for the project described 
above.) 
A likely explanation for the absence of tissue de- 
struction in these mice is the failure of the lympho- 
cytes to become activated. Two transgenic models 
have been developed where T cells can be activated 
locally. In the first, interleukin-2 (IL-2) was di- 
rected to the islets of Langerhans to test whether it 
could cause autoimmunity, e.g., by the activation of 
"anergic" T cells that were previously tolerant. 
IL-2, however, has several other known abilities to 
activate lymphocytes. Transgenic mice expressing 
IL-2 in the islets develop a strong infiltrate predomi- 
nantly of T cells and become diabetic at approxi- 
mately 4-5 months of age. Current eff'orts are di- 
rected toward determining whether this diabetes is 
autoimmune. 
A second test of the role of T cell activation in 
autoimmunity involves co-stimulatory molecules. 
The two-signal model of T cell activation states that 
naive T cells are activated only when stimulated 
through the antigen receptor and simultaneously 
through a second receptor that delivers a co- 
stimulatory signal. T cells receiving stimulation 
only through the antigen-specific receptor are be- 
lieved to be inactivated or anergized. A favored mol- 
ecule for the co-stimulatory receptor on T cells is 
CD28; its ligand is the molecule B7, found on pro- 
fessional antigen-presenting cells (APCs). If this hy- 
pothesis is correct, then tolerance would be me- 
diated by stimulation of T cells by antigen presented 
on tissue cells that lack B7, such as islets of Langer- 
hans. Conversely, expression of co-stimulatory li- 
gand B7 on the islets in conjunction with MHC 
should lead to T cell activation. In this way the islet 
would become a quasi APC. 
Transgenic mice expressing the human (h) B7 
have been generated, and also transgenic mice coex- 
pressing both hB7 and the MHC class II molecule 
I-E. Islets transgenic for hB7 possess APC function in 
vitro and are capable of activating naive CDS T cells 
and thus of providing co-stimulatory function. De- 
spite this, transgenic mice expressing hB7 alone do 
not become autoimmune. However, mice express- 
ing hB7 and I-E on the islets acquire a lymphocytic 
infiltrate. Current efforts will test whether this is an 
autoimmune response. If so, the two-signal hypoth- 
esis will be strongly supported. 
Protective Role of MHC 
in Autoimmune Diabetes 
The murine I-E molecule protects against diabetes 
in nonobese diabetic (NOD) mice. Some time ago 
Dr. Flavell's laboratory, in collaboration with Dr. 
Ralph Brinster (University of Pennsylvania) and Dr. 
Richard Palmiter (HHMI, University of Washing- 
ton), generated transgenic mice in which the ex- 
pression of the I E molecule was directed to specific 
tissues that are a subset of the I-E"^ tissues in normal 
I-E^ mice. These mice therefore permit the determi- 
nation of the cell type that mediates this protective 
effect. Accordingly, Dr. Flavell's group, in collabora- 
tion with Dr. Pia Reich and Dr. Charles Janeway 
(HHMI, Yale University) crossed these transgenic 
mice to the NOD background and showed that only 
transgenic mice expressing I E on peripheral APCs 
were protected from IDDM. These results suggest 
that the protective effect is mediated by antigen pre- 
sentation, either by the induction of tolerance or by 
the preferential activation of "protective" T cells. 
Positive selection of T cells by I-E does not play a 
major role, since mice lacking I E on the thymic cor- 
tex were still protected against diabetes. (A grant 
from the National Institutes of Health provided par- 
tial support for the project described above.) 
Regulation and Expression of the MHC 
Dr. Flavell's laboratory has isolated mutant HeLa 
cell lines defective in the transcriptional activation 
of MHC genes by interferon-7 (IFN-7). The first 
group of mutants were defective in the transcrip- 
tional activation of both IFN-7- and IFN-a- 
IMMUNOLOGY 327 
