Abstract: Murine. I-Ak epidermal antigen-presenting cells. (APCs) have been shown to be capable of presenting soluble tumor fragments. (TFs), as a source.
lnterferon-y inhibits tumor antigen antigen-presenting cells Stephan
Grabbe,
Sandra
MGH/Harvard
Cutaneous
Medical
Charlestown,
Abstract: cells (APCs)
Murine have
School,
been
Bruvers,
Biology
Research
Stefan Center,
Beissert, Department
I-Ak epidermal shown to be
antigen-presenting capable of presenting
J.
Key
Words:
Langerhans interferon-y
cells .
.
and
tumor
immunity
.
antigen
mice
INTRODUCTION For the whether
past few years, our laboratory has investigated antigen-presenting cells (APCs) are involved in immune responses against cutaneous neoplasms and how Langerhans cell (LC) presentation of tumor-associated antigens (TAAs) is regulated by cytokines and other factors derived from the local microenvironment. LCs are dendritic APCs that reside in the epidermis, make up -2-5% of the total cell population of the epidermis, and have been shown to be capable ofpresenting a variety ofantigens Previous studies have shown that both splenic APCs are able to recognize tumor-associated
can also inhibit oftumor-specific was also found epidermal LCs mal cell-lymphocyte of IFN-’y in the immune responses
to T cells [1]. and cutaneous antigens and to
DTH.
granulocyte-macrophage
13th Street, Received
are
usually
antigenic
and
rejected
General
Hospital,
Harvard
by
normal,
tumors non-UV-
this
cytokine.
interferon-gamma; lymphocyte
TAA, logy
Journal
requests:
of inhibiting by GM-CSF
the
Center, Charlestown, September
of Leukocyte
epidermal
colony-stimulating cells;
natural
antigens; Richard
EC,
hypersensitivity;
Langerhans nMu,
tumor-associated Research
delayed-type
LC, reaction;
Reprint
is capable presentation
upand
tumor antigen presentation for the elicitation DTH responses by epidermal APCs. IFN-’y to inhibit alloantigen-presenting capacity by in the primary and secondary mixed epiderresponse, indicating that the presence cutaneous microenvironment may modulate induced or elicited by epidermal APCs.
-CSF,
observation
(UV)-induced
by
Abbreviations:
GM
ultraviolet
capacity
Our data show that IFN-’y regulation of tumor antigen
initiate tumor immunity in naive mice as well as to elicit a tumor-specific delayed-type hypersensitivity (DTH) response in tumor-immune mice [2, 3]. The clinical observation that immunosuppressed patients develop greatly increased numbers of cutaneous malignancies [4-7], accompanied by the murine
D. Granstein
Massachusetts
irradiated hosts, but grow progressively in UV-irradiated animals [8, 9], suggests that the appearance and growth of malignancies within the cutaneous microenvironment may be subject to immunologic control. Therefore, we investigated whether cytokines present in the vicinity of emerging tumors and/or in the cutaneous microenvironment would be capable of down-regulating LC tumor antigen presentation. Several lines of evidence indicate that various cytokines may affect the ability of LCs to induce or to elicit immune responses. Down-regulatory potential in this regard has so far been demonstrated for transforming growth factor /3 (TGF-f3), which may inhibit alloantigen presentation by cultured but not by fresh LCs [10]. However, our recent studies demonstrated that granulocyte-macrophage colonystimulating factor (GM-CSF) greatly augments the ability of epidermal cells (ECs) to present TAA for induction of immunity in vivo and that tumor necrosis factor a (TNF-a) downregulates TAA presentation by GM-CSF-treated ECs [2, 3]. This is in agreement with earlier studies demonstrating augmentation of LC antigen-presenting function by GM-CSF I 11]. In other experimental systems, studies suggested that interferon-’y (IFN--y) may also play a role in the induction of specific unresponsiveness to various antigens [12-14]. Furthermore, IFN-’y has been demonstrated to induce accessory molecules on the surface of macrophages leading to preferential antigen presentation to suppressor T cells [14]. Gene transcripts for IFN--y have been demonstrated in normal and inflamed murine skin [15]. Therefore, experiments were performed to investigate the effect of IFN-y on Langerhans cell tumor antigen presentation with regard to induction as well as elicitation of tumor immunity, gaining insight into the effect of IFN-’y on cutaneous tumor immune responses as well as on the regulation of Langerhans cell antigen-
presenting
that
Richard
of Dermatology,
by epidermal
Massachusetts
soluble tumor fragments (TFs), as a source of tumorassociated antigens (TAAs), for primary and secondary tumor immune responses. In this study we investigated whether incubation of epidermal APCs in interferon--y (IFN-’y) modulates their ability to present TAA and whether the effects of IFN--y on the presentation of tumor antigen correspond to its effects on alloantigen presentation in both primed and unprimed systems. Our results show that three weekly subcutaneous injections of naive mice with GM-CSF-cultured but not with fresh TAApulsed epidermal APCs induce protective tumor immunity in naive mice and that the immunostimulatory effect of GM-CSF in this system is abrogated by coculture of epidermal cells in IFN--y. Furthermore, epidermal APCs are able to present TAA to primed, tumor-immune mice, as assessed by the elicitation of tumor-specific delayedtype hypersensitivity after injection of TAA-pulsed epidermal APCs. IFN-’y was found to inhibit tumor antigen presentation by freshly prepared epidermal APCs in this system. The effects of IFN-’y on the presentation of tumor antigen correlated well with its effects on the primary and secondary mixed epidermal cell-lymphocyte reaction, indicating that IFN-’y differentially modulates the function of epidermal APCs with regard to induction versus elicitation of immunity. Leukoc. Biol. 55: 695-701; 1994.
p resentation
presentation
TF,
Massachusetts MA 02129. 29, 1993; accepted
mixed
rMu,
tumor
D. Granstein,
Biology
MELR,
murine;
fragments;
January
Volume
IFN--y,
epidermal
recombinant Ts,
MGH/Harvard General
cells;
factor;
cell murinc;
T
suppressor.
Cutaneous
Hospital-East, 26,
Bldg.
Bin149,
1994.
55, June
1994
695
MATERIALS
AND METHODS
washing complement
Mice
treatment Ca2/Mg2-free
Female (BALB/c C3H/HeJ (H-2’) old, were obtained bor,
x A/J) F1 LCAF1JJ and BALB/c (H-2 from the Jackson
(H2’a), mice, Laboratory
)
A/J (H-2’), 6 to 12 weeks (Bar Har-
ME).
Tumors The
S1509a
methylcholanthrene-induced
line, originally by Dr. Mark Philadelphia. and 5% inactivated
Grand
CO2
0.1
in fetal
RPMI-1640 calf serum
NY), mM
mM L-glutamine, HEPES buffer progressively demonstrated responses
spindle
cell
tumor
derived from A/J mice, was kindly provided I. Greene, University of Pennsylvania, It was maintained in tissue culture at 37#{176}C
Island,
tomycin,
supplemented (FCS; Gibco
100 U/ml
essential
and
1 mM (“complete
penicillin,
100 tg/ml
nonessential
amino and
sodium medium”).
in normal syngeneic to induce a in the host 116, 17].
with 10% heatLaboratories,
pyruvate, S1509a
recipients variety of
strepacids, 0.01
usually
2 M
grows
[1] and has been immunological
Reagents Cytokines used in this study include natural murine GMCSF (Genzyme, Cambridge, MA) and recombinant murine IFN-’y (rMuIFN--y, 10 U/mg, Genzyme). This rMuIFN-’y preparation contains
(4#{176}C) -> iT
No TF
400
z
0
against S1509a procedure. As with GM-CSF-in-
Grabbe
5 DAYS
Fig.
1.
IFN-y
munity. CSF
inhibits
CAF1 (A,
0),
ECs 100
U/mI
fragments
were
incubated
were
immunized 2 x the
ECs.
n
vs.
B, C,
=
IFN-y
mean all
NS;
2 x
S1509a tumor
l0
cells
of not
of
P
to
in
100
U/mI
IFN-
a 2-h
pulse
by
Negative
control
and
last
E,
P
C
with
immunity
Earlier studies had shown that I-At epidermal APCs are able to present TAA derived from the cutaneous spindle cell tumor S1509a after short-term incubation in GM-CSF [2]. Thus, TAA-pulsed epidermal APCs can be used to immunize naive syngeneic mice against S1509a and to induce protective tumor immunity against this tumor in vivo. To investigate whether IFN-’y can modulate tumor antigen presentation by epidermal LCs and also to determine whether IFN--y affects cutaneous APC function in an unprimed system in general, we compared the effect of IFN-’y on the ability of untreated and GM-CSF-incubated ECs to
CSF
IFN-y
-J
weekly
the ability
+
tumor tumor
RESULTS
S1509a
GM-CSF
6OO
tumor
IFN-y inhibits
TF
200
three times. The of the maximal
diameter in three perpendicular directions, measured with a Vernier caliper. This method has previously been confirmed to correlate well with the tumor weight [2]. To avoid unnecessary distress to the experimental animals, mice were sacrificed after the tumor volume exceeded 1000 mm3. To evaluate statistical differences between the mean tumor volume in the various experimental groups, a two-way ANOVA based on the slopes of the regression curve in each group was performed. DTH results were analyzed using Swdent’s t-test.
induce
->
IFN-’y->TF
-D-----
evaluation
at least product
GM-CSF
vs.
treated D,
NS;
E
.028.
=
cubated and TAA-pulsed ECs led to induction of profound protective tumor immunity against S1509a. ECs incubated in either IFN-y or medium alone and pulsed with TAAs mediated no significant tumor immunity. Mice that were immunized with ECs which had been incubated in GM-CSF plus 100 U/ml IFN-’y showed no enhanced tumor immunity compared to mice that received ECs which had not been incubated in any cytokines, indicating that IFN-’y reversed or prevented the immunostimulatory effect of GM-CSF in this system. inhibited
Thus, the
coincubation ability of
of ECs epidermal
in GM-CSF plus APCs to acquire
IFN-y potent
antigen-presenting function for the induction of de novo tumor immunity. To ensure that the effect of IFN-y observed was not actually due to a contaminant in the IFN--y preparation, an additional experiment was performed in the same manner except that in one group IFN-’y was neutralized with monoclonal anti-IFN--y. Neutralization was performed by incubation of IFN-’y with anti-IFN-’y for 1 h at 37#{176}Cprior to treatment of ECs. As shown by the data in Figure 2, neutralization of IFN-’y led to complete loss of inhibition of the induction of tumor immunity. In a control group ECs were treated with anti-IFN-’y alone; these cells were fully capable of inducing immunity.
et at.
IFN-y
inhibits
EC
tumor
antigen
presentation
697
IFN-’y inhibits alloantigen-presenting epidermal antigen-presenting cells
1000
-U-
T1
GM-CSF
->
A
GM-CSF
+
IFN-’y
-0--
GM-CSF
+
(IFN-’1’
GM-CSF
+
anti-IFN.’y
CM-CSF
->
No TF
-0--
To investigate whether in the S1509a system modulation of primary
iT
->
anti-IFN’y)
+
iT
->
TF
>
epidermal
EC
APCs
in
purpose,
medium IFN-y,
general,
we
capacity
ECs
plus 50 or medium
of
the effects of IFN-y that were representative of immune responses
antigen-presenting
this
capacity
were
studied
the
in another
modulation
assay
preincubated
U/mI GM-CSF, plus 100 U/mI
we observed its effect on induced by
in
of
system.
medium
medium IFN-’y plus
For alone,
plus 50 U/mi
100 U/ml GM-CSF
for 18 h and used as stimulators in the mixed allogeneic epidermal cell-lymphocyte reaction. A trend toward increased allostimulation was observed with GM-CSF-treated ECs compared to medium alone treatment with many, but not all, experiments demonstrating this. As shown in Table 1, coincubation ofECs in GM-CSF and IFN-7 inhibited the allostimulatory capacity of epidermal antigen-presenting cells compared with ECs that were incubated in GM-CSF alone in four of four experiments. IFN--y significantly inhibited
the
allostimulatory
GM-CSF 0
Fig.
2. Neutralization
IFN--y
from
munity.
A/J
(A,
inhibiting ECs
100 0),
for of TAA.
IFN-’y
or
U/mi
50
to TFs.
differentially
with
All
and
mice
mice.
were
The
with
differently
100
s.c. n
graph treated
=
gg/ml
were
the
ECs.
A
A),
50
2 x
all groups mean vs.
B,
tumor E:
a total
106 live except
(TFs) GM-CSF
with
for
2 x
10
in A
mice vs.
experiment but in a preliminary
not
exposed
(inhibition
was
was
to
also
not statistically experiment, ad-
C,
of IFN-y
on secondary
In analogy to primary immune
the
immune
investigation of responses, we also
responses
the effect tested the
in vitro
of IFN--y on effect of IFN-
of these immuni-
I week
B, which
volume
P < 0.001;
cells
(D,
as a source but not
of three
S1509a group
Effects
GM-CSF anti-IFN-y
anti-IFN-y
immunized
the fourth Interestingly,
ECs
im-
GM-CSF
U/ml
neutralizing
in
in
of
experiments
of anti-TNF-a antibody during the preincubation appeared to restore partially the inhibitory effect of in this system (data not shown), suggesting that IFNrny-induced TNF-a secretion may play a role in this regard.
prevents tumor
50 U/ml
neutralizing
intervals
with
shows
S1509a
with tumor fragments El) were incubated
mice
5 for
(B,
sg/ml
observed significant).
capacity
of four
dition period IFN--y
antisera
induce
ofeither
IFN-y
0.7
at weekly
challenged
to
presence 0.7
with
of A/J
neutralizing
with
20
CHALLENGE
ECs
U/mI
a 2-h pulse cells (E,
EC
specific fresh
in the
preincubated
15
TUMOR
of
incubated
Groups
immunization.
four
ability
GM-CSF
treated
zations. last
of IFN--y
18 h, followed by Negative control
exposed
the
AFIER
GM-CSF
U/mi
10
DAYS
the
were
#{149}), 50 U/mI
and
(C, L)
5
in three
after
contained
IF
immunized D:
NS.
No IF
Effects of IFN-’y on elicitation of antitumor DTH responses by TAA-pulsed epidermal APCs In addition
to investigating
the effect
of IFN-y
IFNy
on the modu-
lation of primary immune responses by epidermal antigenpresenting cells, we also performed experiments to study the effect of IFN-’y on secondary immune responses in primed systems. For this purpose, mice were immunized against S1509a as described in Materials and Methods and the elicitation of a DTH response against S1509a after injection of TAA-pulsed ECs was measured. As seen in Figure 3, injection of freshly prepared TF-pulsed ECs into hind footpads of tumor immune mice leads to elicitation of a significant footpad swelling response against S1509a-TAA presented by epidermal antigen-presenting cells. Earlier studies demonstrated the specificity and genetic restriction of the DTH response [2, 3]. Preincubation ofECs for 3 h in IFN--y before or after exposure to TAAs suppressed the elicitation of a DTH response in this system. Other cytokines such as TGF/3 or TNF-a did not inhibit DTH responses against S1509a TAAs in this system, indicating that the observed effect was specific statistically
698
(data
not shown). However, significant, is of relatively
Journal
of Leukocyte
Biology
the effect, although modest magnitude.
Volume
->NoIF
55, June
1994
IFN-y
->
IF
lEN
->
IF
-
10
0
Mean Fig.
3.
Incubation
DTH
specific by
s.c.
their
enriched
Control
with
TF x 10
footpad
1 x
(B,
ECs C).
of these thickness
D, E: P
