Rel-deficient T cells exhibit defects in production of interleukin 3 and granulocyte-macrophage colony-stimulating factor

Abstract

The c-rel protooncogene encodes a subunit of the NF-kappa B-like family of transcription factors. Mice lacking Rel are defective in mitogenic activation of B and T lymphocytes and display impaired humoral immunity. In an attempt to identify changes in gene expression that accompany the T-cell stimulation defects associated with the loss of Rel, we have examined the expression of cell surface activation markers and cytokine production in mitogen-stimulated Rel-/- T cells. The expression of cell surface markers including the interleukin 2 receptor alpha (IL-2R alpha) chain (CD25), CD69 and L-selectin (CD62) is normal in mitogen-activated Rel-/- T cells, but cytokine production is impaired. In Rel-/- splenic T cell cultures stimulated with phorbol 12-myristate 13-acetate and ionomycin, the levels of IL-3, IL-5, granulocyte- macrophage colony-stimulating factor (GM-CSF), tumor necrosis factor alpha (TNF-alpha), and gamma interferon (IFN-gamma) were only 2- to 3-fold lower compared with normal T cells. In contrast, anti-CD3 and anti-CD28 stimulated Rel-/- T cells, which fail to proliferate, make little or no detectable cytokines. Exogenous IL-2, which restitutes the proliferative response of the anti-CD3- and anti-CD28-treated Rel-/- T cells, restores production of IL-5, TNF-alpha, and IFN-gamma, but not IL-3 and GM-CSF expression to approximately normal levels. In contrast to mitogen-activated Rel-/- T cells, lipopolysaccharide-stimulated Rel-/- macrophages produce higher than normal levels of GM-CSF. These findings establish that Rel can function as an activator or repressor of gene expression and is required by T lymphocytes for production of IL-3 and GM-CSF.

Full text

Proc.
Natl.
Acad.
Sci.
USA
Vol.
93,
pp.
3405-3409,
April
1996
Immunology
Rel-deficient
T
cells
exhibit
defects
in
production
of
interleukin
3
and
granulocyte-macrophage
colony-stimulating
factor
(transcription
factors)
STEVE
GERONDAKIS*,
ANDREAS
STRASSER,
DONALD
METCALF,
GEORGE
GRIGORIADIS,
JEAN-PIERRE
Y.
SCHEERLINCK,
AND
RAELENE
J.
GRUMONT
The
Walter
and
Eliza
Hall
Institute
of
Medical
Research,
P.O.
Box,
The
Royal
Melbourne
Hospital,
Parkville,
Victoria
3050,
Australia
Communicated
by
G.
J.
V.
Nossal,
The
Walter
and
Eliza
Hall
Institute
of
Medical
Research,
Victoria,
Australia,
December
27,
1995
(received
for
review
November
2,
1995)
ABSTRACT
The
c-rel
protooncogene
encodes
a
subunit
of
the
NF-KB-like
family
of
transcription
factors.
Mice
lacking
Rel
are
defective
in
mitogenic
activation
of
B
and
T
lympho-
cytes
and
display
impaired
humoral
immunity.
In
an
attempt
to
identify
changes
in
gene
expression
that
accompany
the
T-cell
stimulation
defects
associated
with
the
loss
of
Rel,
we
have
examined
the
expression
of
cell
surface
activation
mark-
ers
and
cytokine
production
in
mitogen-stimulated
Rel-/-
T
cells.
The
expression
of
cell
surface
markers
including
the
interleukin
2
receptor
a
(IL-2Ra)
chain
(CD25),
CD69
and
L-selectin
(CD62)
is
normal
in
mitogen-activated
Rel-/-
T
cells,
but
cytokine
production
is
impaired.
In
Rel-/-
splenic
T
cell
cultures
stimulated
with
phorbol
12-myristate
13-
acetate
and
ionomycin,
the
levels
of
IL-3,
IL-5,
granulocyte-
macrophage
colony-stimulating
factor
(GM-CSF),
tumor
ne-
crosis
factor
a
(TNF-a),
and
y
interferon
(IFN-y)
were
only
2-
to
3-fold
lower
compared
with
normal
T
cells.
In
contrast,
anti-CD3
and
anti-CD28
stimulated
Rel-/-
T
cells,
which
fail
to
proliferate,
make
little
or
no
detectable
cytokines.
Exoge-
nous
IL-2,
which
restitutes
the
proliferative
response
of
the
anti-CD3-
and
anti-CD28-treated
Rel-/-
T
cells,
restores
production
of
IL-5,
TNF-a,
and
IFN-y,
but not
IL-3
and
GM-CSF
expression
to
approximately
normal
levels.
In
con-
trast
to
mitogen-activated
Rel-/-
T
cells,
lipopolysaccharide-
stimulated
Rel-/-
macrophages
produce
higher
than
normal
levels
of
GM-CSF.
These
findings
establish
that
Rel
can
function
as
an
activator
or
repressor
of
gene
expression
and
is
required
by
T
lymphocytes
for
production
of
IL-3
and
GM-CSF.
Rel/NF-KB
transcription
factors
are
homo-
and
heterodimeric
proteins
composed
of
subunits
encoded
by
a
small
multigene
family
related
to
the
c-rel
protooncogene
(1,
2).
Rel/NF-KB
factors
regulate
transcription
by
binding
to
decameric
se-
quences
(KB
motifs)
in
the
promoters
and
enhancers
of
many
viral
and
cellular
genes,
in
particular
those
that
encode
pro-
teins
involved
in
immune,
acute
phase
and
inflammatory
responses
(3-5).
The
five
known
mammalian
Rel-related
pro-
teins,
NF-KB1
(p50,
p105),
NF-KB2
(p52,
plOO),
RelA
(p65),
RelB,
and
Rel
share
a
conserved
300-amino
acid
N
terminus
[Rel
homology
domain
(RHD)]
that
encompasses
sequences
required
for
DNA
binding,
protein
dimerization,
and
nuclear
localization
(1-5).
The
C
termini
of
Rel-related
proteins
are
divergent,
with
those
of
Rel,
RelA,
and
RelB
containing
transcriptional
transactivation
domains
(1,
4).
In
most
cells
prior
to
stimulation,
a
large
proportion
of
Rel/NF-KB
is
retained
in
the
cytoplasm
in
an
inactive
form
through
associ-
ation
with
regulatory
IKB
proteins
(1-5).
A
wide
range
of
stimuli
promote
Rel/NF-KB
nuclear
localization
by
a
mecha-
The
publication
costs
of
this
article
were
defrayed
in
part
by
page
charge
payment.
This
article
must
therefore
be
hereby
marked
"advertisement'"
in
accordance
with
18
U.S.C.
§1734
solely
to
indicate
this
fact.
nism
that
involves
phosphorylation
and
subsequent
degrada-
tion
of
the
IKB
proteins
(6).
Expression
of
the
Rel
subunit
is
largely
confined
to
hemo-
poietic
cells
(7,
8).
In
the
B-lymphocyte
lineage,
increased
Rel
expression
results
in
a
qualitative
change
in
the
subunit
composition
of
Rel/NF-KB
proteins
during
the
transition
from
a
pre-B
to
a
B
cell
(9-11).
A
rapid
increase
in
nuclear
levels
of
Rel
(9,
12,
13)
and
increased
c-rel
transcription (14)
during
mitogenic
activation
of
B
and
T
cells
implicates
Rel
in
lym-
phocyte
proliferation.
Such
a
function
is
also
supported
indi-
rectly
by
the
ability
of
the
avian
viral
oncogene
v-rel
to
transform
both
B
and
T
cells
(15).
While
genes
regulated
by
Rel/NF-KB
that
are
critical
for
B-cell
proliferation
are
yet
to
be
identified,
Rel
is
implicated
in
the
transcription
of
a
number
of
genes
that
are
thought
to
be
crucial
for
T-cell
activation
(5).
Rel
up-regulates
the
transcription
of
human
interleukin
2
receptor
a
(IL-2Ra)
chain
promoter
reporters
(16)
and
Rel/
NF-KB
is
involved
in
IL-2
transcription
(12,
17-19).
Rel/
NF-KB
has
also
been
implicated
in
transcriptional
regulation
of
other
cytokine
genes
expressed
in
activated
T
cells
(5,
18).
The
promoters
of
the
genes
encoding
IL-3,
IL-5,
IL-6,
tumor
necrosis
factor
a
(TNF-a),
granulocyte-macrophage
colony-
stimulating
factor
(GM-CSF)
and
y
interferon
(IFN-3y)
con-
tain
KB
sites
or
the
KB-related
CK1/CD28
response
element
(CD28RE;
refs.
20
and
21).
The
CD28RE,
which
is
involved
in
IL-2
(20)
and
GM-CSF
(22)
transcription,
has
recently
been
shown
to
bind
Rel/NF-KB
complexes
(21).
To
better
understand
the
physiological
role
of
Rel,
we
have
used
gene
targeting
to
generate
mutant
mice
with
an
inacti-
vated
c-rel
gene
(19).
In
these
mice,
the
development
of
cells
from
all
hemopoeitic
lineages
appears
to
be
normal,
but
humoral
immunity
is
impaired.
Mature
B
and
T
cells
are
largely
unresponsive
to
a
number
of
mitogenic
stimuli
and
IL-2
expression
is
reduced
in
activated
T
cells
(19).
In
an
attempt
to
identify
further
changes
in
T-cell
gene
expression
associated
with
the
loss
of
Rel,
we
examined
mitogen-activated
Rel-/-
T
cells
for
impaired
regulation
of
cell
surface
activation
markers
and
cytokines.
No
differences
were
detected
in
expression
of
a
range
of
cell
surface
markers.
However,
production
of
IL-3
and
GM-CSF,
but
not
IL-5,
TNF-a,
or
IFN--y,
was
diminished.
In
contrast
to
T
cells,
GM-CSF
production
by
lipolysaccharide-
stimulated
Rel-/-
resident
peritoneal
macrophages
was
higher
than
that
of
normal
cells.
These
findings
directly
implicate
Rel
in
controlling
expression
of
a
subset
of
cytokine
genes
and
indicate
that
Rel
is
important
for
positive
and
negative
regu-
lation
of
GM-CSF
expression
in
different
cell
types.
MATERIALS
AND
METHODS
Lymphocyte
and
Macrophage
Activation
in
Tissue
Culture.
Splenic
T
cells
of
-90%
purity
isolated
as
described
(19)
from
Abbreviations:
IL,
interleukin;
TNF-a,
tumor
necrosis
factor
a;
GM-
CSF,
granulocyte-macrophage
colony-stimulating
factor;
IFN-y,
in-
terferon
y;
CD28RE,
CD28
response
element;
PMA,
phorbol
12-
myristate
13-acetate.
*To
whom
reprint
requests
should
be
addressed.
3405

3406
Immunology:
Gerondakis
et
al.
4-
to
6-week-old
normal
and
Rel-deficient
mice
were
stimu-
lated
for
up
to
72
hr
in
RPMI
1640
medium
supplemented
with
10%
fetal
bovine
serum
and
50
,uM
2-mercaptoethanol
at
a
concentration
of
5
x
105
cells
per
ml
in
the
presence
or
absence
of
mouse
recombinant
IL-2
(200
units/ml;
Cetus)
with
con-
canavalin
A
(Pharmacia)
at
2
,tg/ml,
phorbol
12-myristate
13-acetate
(PMA;
Sigma)
plus
ionomycin
(Sigma)
at
2
ng/ml
and
1
jug/ml,
respectively,
or
on
tissue
culture
plates
coated
with
monoclonal
antibodies
specific
for
mouse
CD3
(clone
145-2C11;
ref.
23)
and
mouse
CD28
(clone
37.51;
ref.
24)
at
antibody
concentrations
of
10
,tg/ml
in
the
coating
solution.
Resident
peritoneal
macrophages
isolated
from
6-
to
8-week-
old
mice
were
first
purified
by
adherence
to
plastic
dishes
for
3
hr
and
then
cultured
at
5
x
105
cells
per
ml
in
Dulbecco's
Modified
Eagles
medium
containing
10%
fetal
calf
serum
and
lipopolysaccharide
at
0.1
or
1
jig/ml
for
24-48
hr.
Immunofluorescence
Staining
and
Flow
Cytometry.
The
expression
of
the
IL-2Ra
chain
(CD25),
CD69,
and
L-selectin
(CD62)
were
analyzed
on
T
cells
before
and
after
24,
48,
and
72
hr
of
stimulation.
Cells
were
stained
as
described
(25)
with
anti-Thyl.2-FITC
(clone
30H12;
Becton
Dickinson)
and
with
biotinylated
monoclonal
antibodies
specific
for
IL-2Ra
chain
(clone
PC61),
CD69
(clone
H1.2F3),
or
CD62/L-selectin
(clone
Mel-14)
or
with
a
biotinylated
isotype-matched
control
monoclonal
antibody.
Binding
of
biotinylated
antibodies
was
revealed
by
staining
with
phycoerythrin/streptavidin
(Caltag)
as
the
secondary
reagent.
Cytokine
Assays.
The
levels
of
IL-3,
IL-5,
TNF-a,
and
GM-CSF
in
T-cell
culture
supernatants
were
determined
by
ELISA
(Pharmingen,
San
Diego,
CA)
according
to
the
man-
ufacturer's
instructions.
IL-3
levels
were
also
determined
by
bioassay
using
IL-3-responsive
32D
cells
(26),
and
the
level
of
IFN-,y
was"
detected
by
using
the
IFN-3y
sensitive
cell
line
WEHI279
as
described
(27).
GM-CSF
levels
produced
by
stimulated
peritoneal
macrophages
were
determined
by
bio-
assay
as
described
(28).
All
cytokine
assays
were
standardized
by
including
a
titration
of
the
appropriate
purified
recombi-
nant
cytokine
of
known
activity.
RESULTS
The
Expression
Pattern
of
T
Cell
Surface
Markers
Is
Normal
in
Mitogen-Activated
Rel-Deficient
Cells.
Although
expression
of
a
wide
range
of
cell
surface
markers
was
shown
to
be
normal
in
naive
Rel-deficient
T
cells
(19),
a
lack
of
response
to
certain
modes
of
activation
prompted
a
study
of
surface
marker
expression
during
mitogenic
stimulation.
Changes
in
expression
of
a
number
of
surface
markers
nor-
mally
occurs
with
the
onset
of
T-cell
activation.
CD69
(29),
CD25/IL-2Ra
chain
(30),
CD44/Pgpl
(30),
and
the
Fas/
Apo-1
(CD95)
receptor
(31,
32)
are
all
up-regulated
during
T-cell
stimulation,
while
L-selectin/CD62
(33)
and
the
CD45
isoform
CD45RB
(33)
are
down-regulated.
Flow
cytometry
was
used
to
monitor
expression
of
these
cell
surface
markers
in
normal
and
in
Rel-/-
T
cells
stimulated
over
a
72-hr
time
course
with
PMA
and
ionomycin
or
with
anti-CD3
and
anti-
CD28.
A
representative
sample
of
results
from
these
experi-
ments
(Fig.
1)
clearly
establishes
that
72
hr
poststimulation
the
expression
of
CD25,
CD69,
and
MEL14
in
response
to
both
sets
of
stimuli
is
normal.
An
examination
of
these
and
other
cell
surface
markers
after
24
and
48
hr
of
stimulation
con-
firmed
that
the
kinetics
of
expression
on
Rel-/-
T
cells
was
also
normal
(results
not
shown).
These
results
therefore
doc-
ument
that
Rel
is
not
essential
for
expression
of
these
proteins.
The
Expression
of
IL-3
and
GM-CSF
Is
Reduced
in
Mito-
gen-Stimulated
Rel-Deficient
T
Cells.
Activated
T
cells
secrete
a
number
of
cytokines
that
are
critical
in
modulation
of
immune
responses
(34).
The
finding
that
Rel-/-
T
cells
make
reduced
levels
of
IL-2
(19)
prompted
us
to
compare
the
cytokine
secretion
profile
of
normal
and
Rel-/-
T
cells
stimulated
with
different
mitogenic
agents
for
72
hr
(Table
1).
Cytokine
levels
were
consistently
lower
in
stimulated
Rel-/-
T
cell
cultures,
with
the
relative
difference
in
expression
of
each
cytokine
dependent
on
the
stimulus.
For
example,
while
cytokine
levels
in
PMA
plus
ionomycin-stimulated
Rel-/-
T
cells
were
equivalent
to
or
only
severalfold
lower
than
in
normal
cultures,
cytokine
levels
in
anti-CD3
plus
anti-CD28-
stimulated
Rel-/-
T
cells
were
reduced
12-
and
'80-fold
for
GM-CSF
and
IL-3,
respectively.
The
poor
proliferative
response
of
anti-CD3
plus
anti-
CD28-stimulated
Rel-/-
T
cells
compared
with
that
elicited
by
PMA
and
ionomycin
(19)
prompted
an
examination
of
cyto-
kine
production
by
mitogen-stimulated
cultures
supplemented
with
IL-2.
Addition
of
IL-2,
which
restores
Rel-/-
T
cell
proliferative
responses
to
that
of
normal
cells
(19),
should
distinguish
between
reduced
cytokine
production
due
directly
to
the
loss
of
Rel
or
as
a
secondary
effect
resulting
from
the
lack
of
proliferation.
These
experiments
(Table
1)
establish
that
defects
in
cytokine
production
by
Rel-/-
T
cells
fall
into
two
groups.
The
amount
of
IFN-y,
TNF-a,
and
IL-5
now
produced
by
stimulated
Rel-/-
T
cells
was
similar
to
that
of
normal
cells,
indicating
that
reduced
levels
of
these
cytokines
appear
to
be
due
to
the
impaired
proliferative
response.
In
contrast,
the
amount
of
IL-3
and
GM-CSF
produced
by
anti-CD3
plus
anti-CD28-stimulated
Rel-/-
cells
supple-
mented
with
IL-2
was
still
16-
and
11-fold
lower,
respectively,
than
that
made
by
normal
T
cells,
indicating
that
Rel
was
essential
for
normal
expression
of
these
cytokines.
Rel
Has
Opposite
Effects
on
Regulation
of
GM-CSF
Ex-
pression
in
T
Cells
and
Macrophages.
GM-CSF
can
be
produced
not
only
by
activated
T
cells
but
also
by
a
number
of
cell
types
including
fibroblasts,
macrophages,
and
endothelial
cells
in
response
to
stimulation
by
cytokines,
antigens,
or
inflammatory
agents
(35).
Since
Rel
expression
is
largely
confined
to
hemopoietic
cells
(7,
8),
GM-CSF
production
was
also
examined
in
macrophages
from
Rel-/-
mice.
In
contrast
to
the
reduced
GM-CSF
levels
secreted
by
mitogen-stimulated
Rel-/-
T
cells,
resident
Rel-/-
peritoneal
macrophages
treated
with
lipopolysaccharide
for
24
or
48
hr
produce
10-
to
50-fold
more
GM-CSF
than
normal
macrophages
(Fig.
2).
This
did
not
appear
to
be
a
lipopolysaccharide
concentration-
dependent
phenomenon,
since
0.1
or
1
jig/ml
of
lipopolysac-
charide
per
ml
both
resulted
in
elevated
production
of
GM-
CSF
from
Rel-/-
macrophages.
The
basis
for
the
slight
reduction
in
GM-CSF
levels
produced
by
both
the
normal
and
Rel-/-
cultures
after
48
hr
stimulation
compared
with
that
seen
after
24
hr
most
likely
reflects
GM-CSF
consumption
by
the
macrophages.
Consistent
with
the
levels
of
GM-CSF
determined
by
bioassay
and
ELISA,
Northern
blot
analysis
confirmed
that
GM-CSF
mRNA
levels
were
reduced
in
Rel-/-
T
cells
and
elevated
in
Rel-/-
macrophages
compared
with
their
normal
counterparts
(S.G.
and
G.G.,
unpublished
results).
Collectively,
these
results
establish
that
Rel
can
function
as
both
an
activator
and
repressor
of
induced
GM-
CSF
expression
in
different
activated
cell
types.
DISCUSSION
Changes
in
expression
of
many
genes
accompany
T-cell
acti-
vation
(36).
The
finding
that
T
cells
from
Rel-/-
mice
exhibit
defects
in
proliferation
and
reduced
IL-2
expression
(19)
prompted
a
detailed
study
of
gene
expression
during
the
mitogenic
activation
of
Rel-/-
T
cells.
The
expression
pattern
of
cell
surface
markers
including
CD25,
CD44,
CD45RB,
CD69,
and
CD95,
all
of
which
normally
change
during
T-cell
activation
(29-33),
were
unaffected
in
Rel-/-
T
cells.
The
observation
that
Rel
is
not
essential
for
basal
or
induced
IL-2Ra
chain
gene
expression
in
murine
T
cells
was
surprising.
Although
Rel
binds
to
a
KB
site
in
the
promoter
of
the
human
IL-2Ra
chain
gene
required
for
Rel-dependent
transactivation
Proc.
Natl.
Acad.
Sci.
USA
93
(1996)

Proc.
Natl.
Acad.
Sci.
USA
93
(1996)
3407
control
(day
0)
C
~
~
CD69
CD25
CD62
......
.-
......J............
100
101
102
103
104
100
101
102
103 104
100
101
102
103
104
100
101
102
103
104
_
C
CD69
CD25
CD62
,
-,.J
1
.....
.....-
**,
.....J.1
......J
-......J
...
"......-
100
101
102
103
104
100
101
102 103
104
100
101
102
103
104
100
101
102
103
104
anti-CD3
+
anti-CD28
(day
3)
C
CD69
CD25
CD62
100
101
102
103
104
100
101
102
103
104
100
101
102
103
104
100
101
102
103
104
C
CD69
CD25
CD62
-A0.01
1-
*
:
ARel-/-
10°
101
102
103
104
10°
101
102
103
104
10°
101
102
103
104
10°
101
102
103
104
PMA
+plus
ionomycin
(day
3)
C
CD69
CD25
CD62
100
101
102
103
104 100
101
102
103
104 100
101
102
103
104
100
101
102
103
104
C
CD69
CD25
CD62
10°
101
102
103
104
100
101
102 103
104
100
101
102
103
104
100
101
102
103
104
FIG.
1.
Expression
of
cell
surface
activation
markers
on
T
cells
from
Rel-/-
mice.
Spleen
cells
from
normal
and
Rel-/-
mice
were
stimulated
in
tissue
culture
either
with
immobilized
anti-CD3
plus
anti-CD28
or
with
PMA
plus
ionomycin.
Expression
of
IL-2Ra
chain
(CD25),
CD69,
and
L-selectin
(CD62)
was
analyzed
on
T
cells
(identified
as
being
Thl.2
positive)
before
and
after
72
hr
of
stimulation.
Data
are
representative
of
three
independent
experiments
and
are
presented
as
histograms
of
fluorescence
intensity
of
Thyl.2-positive
gated
T
cells.
of
reporter
constructs
(16)
and
v-rel
upregulates
avian
IL-2Ra
chain
expression
in
Rev-T-transformed
lymphoid
cells
(37),
a
comparison
of
mouse
and
human
IL-2Ra
chain
gene
promoter
sequences
(38)
may
explain
the
basis
of
this
unexpected
species-specific
difference.
The
KB
site
in
the
human
IL-2Ra
chain
promoter,
5'-GGGAATCTCC-3',
differs
from
the
se-
quence
in
the
mouse
gene,
5'-GGAATCCCCC-3',
the
latter
of
which
is
a
poor
consensus
binding
site
for
Rel/NF-KB
proteins
(1,
3,
39).
Collectively,
these
findings
suggest
that
Rel
is
not
important
in
murine
IL-2Ra
chain
gene
expression.
In
contrast
to
the
regulation
of
cell
surface
markers,
cyto-
kine
gene
expression
is
perturbed
in
Rel-/-
T
cells.
The
levels
of
IL-2
(19),
IL-3,
IL-5,
GM-CSF,
TNF-a,
and
IFN-y
were
low
or
undetectable
in
concanavalin
A
or
anti-CD3
plus
anti-
CD28-stimulated
Rel-/-
T-cell
cultures.
Rescuing
the
prolif-
eration
of
Rel-/-
T
cells
by
adding
IL-2
revealed
that
IL-5,
TNF-a,
and
IFN-y
production
was
restored
to
normal
levels,
but
expression
of
IL-3
and
GM-CSF
remained
much
lower
than
in
the
normal
T-cell
cultures.
These
findings
indicate
that
in
this
system
expression
of
some
but
not
all
cytokine
genes
appear
to
be
directly
controlled
in
part
by
Rel
and
that
cytokine
production
rather
than
cell
surface
marker
expression
may
be
a
more
accurate
gauge
of
T-cell
activation.
While
the
ability
of
exogenous
IL-2
to
restitute
normal
expression
of
Immunology:
Gerondakis
et
al.

3408
Immunology:
Gerondakis
et
al.
Table
1.
Cytokine
production
by
Rel-/-
T
cells
in
response
to
different
mitogenic
stimuli
IFN-y,
IL-3,
pg/ml
IL-5,
pg/ml
GM-CSF,
pg/ml
TNF-a,
pg/ml
ng/ml
Stimulus
+/+
-/-
+/+
-/-
+/+
-/-
+/
+
-/-
/
None
<30
<30 <30
<30
<30 <30
<30
<30
<1
<1
IL-2
<30 <30 <30
<30 <30
<30
<30
<30
<1
<1
Concanavalin
A
600
+
70
<30
110
+
30
<30
625
+
80
120
+
30
315
+
50
175
_
40
20
ND
Concanavalin
A
+
IL-2
1,350+
110
65
_
20
190
40
160
40
1025
110
170
20
480
70
350
+50
20
15
PMA
+
ionomycin
75,000
+
5500
23,500
+
2100
410
+
60
250
+
30
5600
+
240
2350
+
180
1670
+
180
910
+
110
80
80
PMA
+
ionomycin
+
IL-2
68,000
+
4000
29,000
+
1500
520
+
40
470
+
50
7150
±+
570
2650
+
190
2240
+
270
1820
+
210
80
80
Anti-CD3
+
anti-CD28
34,500
+
2500
450
+
70
340
+
30
<30
4800
+
230 350
+
60
2750
+
230
130
±
40
160
20
Anti-CD3
+
anti-CD28
+
IL-2
42,000
±
3100
2,600
+
160
480
+
40
540
±
60
7100
+
480
650
+
80
3060
+
240
3160
+
190
160
>300
Specific
cytokine
levels
produced
by
T
cells
stimulated
for
72
hr
were
determined
by
comparing
supernatant
activities
with
those
of
purified
recombinant
cytokine
standards
and
were
converted
to
pg/ml
or
ng/ml.
Means
±
SD
for
IL-3,
IL-5,
GM-CSF,
and
TNF-a
were
determined
from
four
to
six
separate
experiments.
IFN-7
levels
are
means
of
two
experiments.
<,
Cytokine
activities
lower
than
the
limit
of
sensitivity
for
detection
by
either
cytokine
ELISA
or
bioassay;
ND,
not
determined.
certain
T-cell
cytokines
is
consistent
with
cytokine
production
being
intimately
linked
to
proliferation,
it
remains
a
formal
possibility
that
the
IL-2-mediated
rescue
of
IL-5,
IFN-y,
and
TNF-a
expression
in
anti-CD3
plus
anti-CD28-treated
Rel-
deficient
T
cells
is
associated
with
an
aspect
of
T-cell
activation
that
is
independent
of
the
IL-2-dependent
proliferative
response.
The
transcriptional
control
regions
of
the
IL-2,
IL-3,
and
GM-CSF
genes
share
a
number
of
regulatory
sequences
including
the
CK1
or
CD28RE,
which
is
distinct
from
but
related
to
the
KB
motif
and
has
been
shown
to
bind
complexes
containing
NF-KB1,
RelA,
and
Rel
(21).
A
compilation
of
CD28REs
within
cytokine
gene
promoters
is
shown
in
Fig.
3.
The
GM-CSF
motif
is
a
perfect
match
with
the
Rel
homodimer
consensus
binding
site
5'-NGGNNA/TTTCC-3'
(39),
while
the
sequences
from
the
IL-2, IL-3,
and
G-CSF
genes
differ
by
only
a
single
nucleotide
from
the
consensus
motif.
Although
STIMULUS
NONE
LPS
0.1
gg
1
pg
30000
10000
E
uL
1000
C)
0
CD
100
30
NONE
LPS
0.1
pg
1
lg
FIG.
2.
Level
of
GM-CSF
produced
by
lipopolysaccharide
(LPS)-
activated
Rel-/-
peritoneal
macrophages
is
elevated
compared
with
normal
cells.
Resident
peritoneal
macrophages
from
normal
(open
bars)
or
Rel-/-
(solid
bars)
mice
were
stimulated
with
lipopolysac-
charide
(0.1
or
1
j,ig/ml)
for
24
and
48
hr.
Each
histogram
represents
the
mean
±
SE
from
four
or
six
experiments.
the
IFN-y
CD28RE
differs
by
only
2
nucleotides
from
the
consensus
sequence,
these
changes
reside
in
the
3'
half-site,
which
is
the
most
conserved
region
of
the
Rel
binding
element
(39).
Therefore,
the
ability
of
Rel
to
preferentially
bind
to
particular
CD28REs
may
account
for
the
relative
defects
in
expression
of
the
different
cytokine
genes
in
the
Rel-/-
T
cells.
Despite
evidence
that
phorbol
ester
and
calcium
ionophore
treatment
mimics
activation
mediated
through
cross-linking
of
the
T-cell
receptor
and
CD28
costimulatory
receptor
(18),
the
relative
levels
of
IL-3
and
GM-CSF
produced
by
normal
and
Rel-/-
T
cells
in
response
to
these
two
stimuli
differ.
The
ability
of
PMA
and
ionomycin
to
induce
IL-2
(19),
IL-3,
and
GM-CSF,
albeit
several
fold
less
than
in
control
cells
in
the
absence
of
Rel,
may
indicate
that
in
T
cells,
Rel
is
either
upstream
of
the
point
at
which
these
agents
engage
a
common
intracellular
pathway
required
for
the
transcription
of
these
genes
or
that
these
agents
operate
through
a
distinct
pathway.
Alternatively,
PMA
and
ionomycin,
but
not
anti-CD3
plus
anti-CD28
costimulation
may
activate
other
Rel
family
proteins
or
different
transcription
factors
that
substitute
in
part
for
Rel.
A
comparative
analysis
of
GM-CSF
production
by
mitogen-
activated
Rel-/-
T
cells
and
macrophages
indicates
that
Rel
can
function
both
as
an
activator
or
repressor
of
gene
expres-
sion.
While
this
dual
activity
is
well
established
for
the
Drosophila
Rel
homologue
Dorsal
(40,
41),
the
results
pre-
sented
here
are
the
first
clear
demonstration
that
a
mammalian
GENE
IL-2
GM-CSF
IL-3
G-CSF
INF-y
CD28RE
SEQUENCE
AaGAAATTCC
AGGAGATTCC
TGGAGgTTCC
Ca
GAG
AT
T
CC
AGGAAAcTC
t
LOCATION
-163
to
-154
-109
to
-98
-116
to
-105
-193
to
-182
-171
to
-160
Consensus
Rel
binding
motif
5'-N
G
G
N N
A/T
T
T
C
C-3'
FIG.
3.
Compilation
of
CD28
response
elements
in
the
promoters
of
murine
cytokine
genes.
These
sequences
are
compared
with
the
consensus
Rel
binding
site
(39),
and
nonconserved
bases
are
shown
in
lowercase.
Nucleotide
numbering
assigned
to
these
elements
in
the
5'
flanking
sequence
of
the
different
mouse
cytokine
genes
is
according
to
the
start
sites
of
transcription.
24h
48h
I~~~~~~~~~~~~~
L--A-
L--L--M
A
7
Proc.
Natl.
Acad.
Sci.
USA
93
(1996)

Proc.
Natl.
Acad.
Sci.
USA
93
(1996)
3409
Rel
protein
possesses
these
properties.
One
plausible
model
that
could
explain
how
Rel
is
an
activator
of
GM-CSF
expres-
sion
in
T
cells
but
a
repressor
of
the
same
gene
in
macrophages
is
the
modulation
of
Rel
activity
through
interaction
with
different
transcription
factors
in
the
two
cell
types.
This
model
is
consistent
with
the
ability
of
Rel/NF-KB
proteins
to
asso-
ciate
with
a
number
of
transcription
factors
including
ATF-2
(42),
SP1
(43),
and
the
C/EBP
proteins
(44).
The
finding
described
here
could
result
from
Rel
associating
with
different
Rel-like
subunits
or
with
unrelated
proteins.
For
example,
different
high
mobility
group
(HMG)
proteins
can
determine
whether
Rel/NF-KB
factors
function
as
activators
or
repres-
sors
of
transcription.
Binding
of
HMG
I
(Y)
at
sites
adjacent
to
the
KB
element
in
the
IFN-f3
gene
augments
both
NF-KB
binding
and
transcriptional
activity
(45),
while
the
DSP1
HMG
protein
binds
to
a
negative
regulatory
element
flanking
the
IFN-f3
KB
site
and
converts
NF-KB
from
a
transcriptional
activator
to
a
repressor
(46).
The
identification
of
genes
whose
expression
is
induced
or
repressed
by
Rel
should
provide
a
basis
for
elucidating
the
mechanism(s)
governing
the
tran-
scription
regulatory
properties
of
Rel/NF-KB
factors.
We
thank
Dr.
James
Allison
for
anti-CD28-producing
hybridoma
cells;
Dr.
Jan
Allison
for
monoclonal
antibodies
to
CD44,
CD45RB,
and
CD62;
and
Dr.
Doug
Hilton
and
the
Cetus
Corporation
for
purified
murine
GM-CSF
and
IL-2,
respectively.
This
work
was
supported
by
the
National
Health
and
Medical
Research
Council
(Australia),
Australian
Cooperative
Research
Centre
Grant
91007,
and
a
Human
Frontiers
Science
Program
Grant,
Principal
Investigator,
Dr.
Diane
Mathis.
A.S.
was
supported
by
fellowships
from
the
Leukemia
Society
of
America
and
the
Swiss
National
Science
Foundation.
1.
Grilli,
M.,
Chiu,
J.-S.
&
Lenardo,
M.
J.
(1993)
Int.
Rev.
Cytol.
143,
1-62.
2.
Siebenlist,
U.,
Franzoso,
G.
&
Brown,
K.
(1994)
Annu.
Rev.
Cell
Biol.
10,
405-455.
3.
Baeuerle,
P.
A.
(1991)
Biochim.
Biophys.
Acta
1072,
63-80.
4.
Nolan,
G.
&
Baltimore,
D.
(1992)
Curr.
Opin.
Genet.
Dev.
2,
211-220.
5.
Baeuerle,
P.
A.
&
Henkel,
T.
(1994)
Annu.
Rev.
Immunol.
12,
141-179.
6.
Finco,
T.
S.
&
Baldwin,
A.
S.
(1995)
Immunity
3,
263-272.
7.
Brownell,
E.,
Mathieson,
B.,
Young,
H.
A.,
Keller,
J.,
Ihle,
J.
N.
&
Rice,
N.
R.
(1987)
Mol.
Cell.
Biol.
7,
1304-1309.
8.
Grumont,
R.
J.
&
Gerondakis,
S.
(1990)
Oncogene
Res.
2,
245-
254.
9.
Grumont,
R.
J.
&
Gerondakis,
S.
(1994)
Cell
Growth
Differ.
5,
1321-1331.
10.
Liou,
H.-C.,
Sha,
W.
C.,
Scott,
M.
L.
&
Baltimore,
D.
(1994)
Mol.
Cell.
Biol.
14,
5249-5359.
11.
Miyamoto,
S.,
Schmitt,
M.
J.
&
Verma,
I.
M.
(1994)
Proc.
Natl.
Acad.
Sci.
USA
91,
5056-5060.
12.
Bryan,
R.
G.,
Li,
Y.,
Lai,
J.-H.,
Van,
M.,
Rice,
N.,
Rich,
R.
R.
&
Tan,
T.-H.
(1994)
Mol.
Cell.
Biol.
14,
7933-7942.
13.
Venkataraman,
L.,
Burakoff,
S.
J.
&
Sen,
R.
(1995)
J.
Exp.
Med.
181,
1091-1099.
14.
Grumont,
R.
J.
&
Gerondakis,
S.
(1990)
Cell
Growth
Differ.
1,
345-350.
15.
Gilmore,
T.
(1991)
Trends
Genet.
7,
318-322.
16.
Tan,
T.-H.,
Huang,
G.
P.,
Sica,
A.,
Ghosh,
P.,
Young,
H.
A.,
Longo,
D.
L.
&
Rice,
N.
R.
(1992)
Mol.
Cell.
Biol.
12,4067-4075.
17.
Hoyos,
B.,
Ballard,
D.
W.,
Bohnlein,
E.,
Siekvitz,
M.
&
Greene,
W.
C.
(1989)
Science
244,
457-460.
18.
Crabtree,
G.
R.
&
Clipstone,
N.
A.
(1994)
Annu.
Rev.
Biochem.
63,
1045-1083.
19.
Kontgen,
F.,
Grumont,
R.
J.,
Strasser,
A.,
Metcalf,
D.,
Li,
R.,
Tarlinton,
D.
&
Gerondakis,
S.
(1995)
Genes
Dev.
9,
1965-1977.
20.
Fraser,
J.
D.,
Irving,
B.
A.,
Crabtree,
G.
R.
&
Weiss,
A.
(1991)
Science
251,
313-316.
21.
Ghosh,
P.,
Tan,
T.-H.,
Rice,
N.,
Sica,
A.
&
Young,
H.
A.
(1993)
Proc.
Natl.
Acad.
Sci.
USA
90,
1696-1700.
22.
Shannon,
M.
F.,
Gamble,
J.
R.
&
Vadas,
M.
A.
(1988)
Proc.
Natl.
Acad.
Sci.
USA
85,
674-678.
23.
Leo,
O.,
Foo,
M.,
Sachs,
D.,
Samelson,
L.
&
Bluestone,
J.
(1987)
Proc.
Natl.
Acad.
Sci.
USA
84,
1374-1378.
24.
Harding,
F.
A.,
McArthur,
J.
G.,
Gross,
J.
A.,
Raulet,
D.
H.
&
Allison,
J.
P.
(1992)
Nature
(London)
356,
607-609.
25.
Strasser,
A.,
Harris,
A.
W.
&
Cory,
S.
(1991)
Cell
67,
889-899.
26.
Metcalf,
D.
(1985)
Blood
65,
357-362.
27.
Kelso,
A.
(1990)
J.
Immunol.
145,
2167-2176.
28.
Metcalf,
D.
&
Rasko,
J.
E.
J.
(1993)
Leukemia
7,
878-886.
29.
Croft,
M.
(1994)
Curr.
Opin.
Immunol.
6,
431-437.
30.
Sagerstrom,
C.
G.,
Kerr,
E.
M.,
Allison,
J.
P.
&
Davis,
M. M.
(1993)
Proc.
Natl.
Acad.
Sci.
USA
90,
8987-8991.
31.
Trauth,
B.
C.,
Klas,
C.,
Peters,
A.
M.
K.,
Matzku,
S.
M.,
Moller,
P.,
Falk,
W.,
Debatin,
K.-M.
&
Krammer,
P.
H.
(1989)
Science
244,
301-305.
32.
Itoh,
N.,
Yonehara,
S.,
Ishii,
A.,
Yonehara,
M.,
Mizushima,
S.-I.,
Sameshima,
M.,
Hase,
A.,
Seto,
Y.
&
Nagata,
S.
(1991)
Cell
66,
233-243.
33.
Kosaka,
H.,
Surh,
C.
D.
&
Sprent,
J.
(1992)
J.
Exp.
Med.,
1291-1302.
34.
Thompson,
C.
B.,
Lindsten,
T.,
Ledbetter,
J.
A.,
Kunkel,
S.
L.,
Young,
H.,
Emerson,
S.
G.,
Leiden,
J.
M.
&
June,
C.
H.
(1989)
Proc.
Natl.
Acad.
Sci.
USA
86,
1333-1337.
35.
Rasko,
J.
&
Gough,
N.
M.
(1994)
The
Cytokines
Handbook
(Academic,
New
York),
pp.
343-369.
36.
Crabtree,
G.
R.
(1989)
Science
243,
355-361.
37.
Hrdlickova,
R.,
Nehyba,
J.
&
Humphries,
E.
H.
(1994)J.
Virol.
68,
308-319.
38.
Suzuki,
N.,
Matsunami,
N.,
Kanamori,
H.,
Ishida,
N.,
Shinizu,
A.,
Yaoita,
Y.,
Nikaido,
T.
&
Honjo,
T.
(1988)
J.
Biol.
Chem.
262,
5079-5086.
39.
Kunsch,
C.,
Ruben,
S.
M.
&
Rosen,
C.
A.
(1992)
Mol.
Cell.
Biol.
12,
4412-4421.
40.
Anderson,
K.
V.
(1987)
Trends
Genet.
3,
91-97.
41.
Levine,
M.
(1988)
Cell
52,
785-786.
42.
Kaszubska,
W.,
van
Huijsduijnen,
R.
H.,
Ghersa,
P.
&
DeRaemy-
Schenk,
A.
M.
(1993)
Mol.
Cell.
Biol.
13,
7180-7190.
43.
Perkins,
N.
D.,
Edwards,
N.
L.,
Duckett,
C.
S.
&
Agranoff,
A.
B.
(1993)
EMBO
J.
12,
3551-3558.
44.
Stein,
B.
&
Baldwin,
A.
S.
(1993)
Mol.
Cell.
Biol.
13,
7191-7198.
45.
Thanos,
D.
&
Maniatis,
T.
(1992)
Cell
71,
777-789.
46.
Lehming,
N.,
Thanos,
D.,
Brickman,
J.
M.,
Ma,
J.,
Maniatis,
T.
&
Ptashne,
M.
(1994)
Nature
(London)
371,
175-179.
Immunology:
Gerondakis
et
al.