Vol. 3, 483-490, Marc/i 1997 Clinical Cancer Research 483
Decreased Antigen Presentation by Dendritic Cells in Patients with
Breast Cancer’
Dmitry I. Gabrilovich,2 Jadranco Corak,
Ilja F. Ciernik, Denise Kavanaugh, and
David P. Carbone
The Vanderbilt Cancer Center, Vanderbilt University School of
Medicine, Nashville, Tennessee 37232-6838 ID. I. G., D. P. Cl:
Hamon Center for Therapeutic Oncology Research, University of
Texas Southwestern Medical Center, Dallas, Texas [J. C., D. K.l: andDepartement fuer Innere Medizin, Universitaetsspital Zurich
Raemistrasse 100, 8091 Zurich, Switzerland [I. F. C.l
ABSTRACT
We evaluated T-cell responses to mitogens and to de-fined antigens in breast cancer patients. Significant defectsin responses to tetanus toxoid and influenza virus were
observed in patients with advanced-stage breast cancer. To
define whether these defects were associated with a defect inantigen presentation [dendritic cells (DCs)1 or effector func-
tion (T cells), these cells were studied separately Purified
DCs from 32 patients with breast cancer demonstrated a
significantly decreased ability to stimulate control allogeneicT cells, but stimulation of patient T cells with either controlallogeneic DCs or immobilized anti-CD3 antibody resulted
in normal T-cell responses, even in patients with stage IV
tumors. These data suggest that reduced DC function couldhe one of the major causes of the observed defect in cellularimmunity in patients with advanced breast cancer. We then
tested whether stem cells from these patients could give riseto functional DCs after in vitro growth with granulocyte/
macrophage colony-stimulating factor and interleukin 4.
Normal levels of control allogeneic and tetanus toxoid-
dependent T-cell proliferation were observed when DCs
obtained from precursors were used as stimulators. Those
cells also induced substantially higher levels of influenzavirus-specific CTL responses than mature DCs from theperipheral blood of these patients, although responses didnot quite reach control values. Thus, defective T-cell func-tion in patients with advanced breast cancer can be over-
come by stimulation with DCs generated from precursors,suggesting that these cells may better serve as autologous
antigen carriers for cancer immunotherapy than mature
peripheral blood DCs.
Received 9/12/96; revised 1 1/27/96: accepted I 2/I 3/96.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
I This work was supported by Grant RO1-CA61242 from the National
Cancer Institute. Part of this work was presented at the Thirty-Second
Annual Meeting of the American Society of Clinical Oncology, May
1996.2 To whom requests for reprints should be addressed, at The VanderbiltCancer Center, 648 Medical Research Building II, Nashville. TN 37232-
6838. Phone: (615) 936-3320: Fax: (615) 936-1790.
INTRODUCTION
Defective function ofthe host immune system is thought to
be a major mechanism by which tumors escape from immune
control. Numerous studies have described defective function of
T lymphocytes, including tumor-infiltrating cells, in tumor-
bearing hosts (reviewed in Ref. 1). The precise mechanism of
this dysfunction is not known, but it has been reported that it can
be reversed if sufficient stimulation with antigen-presenting
cells and support with cytokines (such as 1L3-2) is provided
(2-4). Because host professional APCs are one of the most
important elements in the induction of specific antitumor im-
mune responses (5, 6), APC could be partly responsible for the
observed defective immune responses in tumor-bearing hosts.
This group of cells includes dendritic cells, macrophage/mono-
cytes, and B cells. DCs are thought to be the most potent APCs,
and only these cells are capable of inducing primary immune
responses and, therefore, are actively being studied for use as
vehicles for delivery of tumor-specific antigens in both animal
experimental systems and in patients with cancer (7-9). In our
previous study, we showed that the nonresponsiveness ofT cells
from tumor-bearing mice was reversed after stimulation with
antigen-pulsed control DCs, even without additional cytokine
support (4). It is not known, however, whether DCs obtained
from patients with cancer have sufficient functional activity for
use as autologous antigen carriers. In particular, the ability of
these cells to present soluble antigens and induce cytotoxic
immune responses in patients with cancer is not clear. The
clarification of this question is important not only for under-
standing the mechanisms of tumor progression but also for
the development of effective cancer immmunotherapeutic
strategies.
Previously, we have shown that defects in T-cell function
in tumor-bearing mice could be reversed by stimulation with
DCs generated from stem cells from the bone marrow of tumor-
bearing mice (10). Here we investigated the function of DCs
from patients with breast cancer and found that these cells had
a considerably lower than normal ability to stimulate allogeneic
T cells as well as to present soluble antigen to autologous T
cells. To test the hypothesis that sufficient antigen presentation
may reverse observed defects in T-cell function, we generated
DCs from peripheral blood precursors and showed that those
cells were able to elicit an almost normal level of immune
response.
3 The abbreviations used are: IL, interleukin: APC, antigen-presentingcell; DC, dendritic cell: CCM, complete culture medium; GM-CSF,
granulocyte/macrophage colony-stimulating factor; NK. natural killer:
FLU. influenza virus: Ti’, tetanus toxoid: MLR. mixed leukocyte reac-tion.
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
484 Decreased Antigen Presentation by DC in Cancer
PATIENTS AND METHODS
Patients. Thirty-two patients aged 33-69 years with his-
tologically confirmed breast cancer were enrolled into the study.
All patients were newly diagnosed with no prior therapy. Stag-
ing was performed in accordance with American Joint Commit-
tee on Cancer criteria. Four patients had stage I tumors, 10 had
stage II, 13 had stage III, and S had stage IV. Fourteen healthy
volunteers were used as controls. Informed consent was ob-
tamed from all individuals.
Reagents. CCM included RPMI 1640 (Life Technolo-
gies, Inc., Gaithersburg, MD) supplemented with 10% FCS and
antibiotics. For the generation of DCs from progenitors, medium
was additionally supplemented with 5 X l0� 2-mercaptoeth-
anol. Ficoll-paque was purchased from Pharmacia Biotech, Inc.
(Uppsala, Sweden). GM-CSF and IL-4 were obtained from
R&D Systems (Minneapolis, MN). Influenza virus strain PR8
(A/Puerto Rico/8/34) was obtained from the American Type
Culture Collection (Rockville, MD) and was propagated in the
Laboratory of Virology at the University of Texas Southwestern
Medical Center (Dr. J. Luby). Anti-HLA DR. CD3, CD14, and
CDI9 antibodies were purchased from Sigma Chemical Co. (St.
Louis, MO). Anti-HLA ABC, B7-l, B7-2, and CD56 antibod-
ies were purchased from PharMingen (San Diego, CA). Anti-
CD34 antibody was obtained from Becton Dickinson (Mountain
View, CA). Purified anti-CD3 antibody (64.1 hybridorna cell
lines produce an IgG2a monoclonal antibody) was a generous
gift from The Cancer Immunobiology Center, University of
Texas-Southwestern (Dr. E. S. Vitetta). Goat antimouse IgO
antibody conjugated with magnetic beads was purchased from
Dynal (Lake Success, NY).
Cell Isolation. DCs and T cells were isolated from the
peripheral blood of patients or control donors as described (11)
with minor modifications. Briefly, mononuclear cells obtained
after centrifugation of peripheral blood over Ficoll-paque gra-
dient were incubated with neurominidase-treated sheep red cells
(R). Cells that adhered to the red cells (R�) and those that did
not (R �) were separated on Ficoll-paque gradient. R cells were
then incubated for 18 h in CCM at 37#{176}C,followed by incubation
of nonadherent cells for 60 mm on Petri dishes coated with
human -y-globuhin to remove remaining FcR� rnonocytes. Fi-
nally, cells were centrifuged over a metrizamide gradient (14.5
g of metrizamide; Neyegaard, Oslo, Norway, in 100 ml of
culture medium). Cells collected from the interface were 50-
70% DCs, as estimated morphologically or by flow cytornetry
after labeling a mixture of monoclonal antibodies (CD3, CD 14,
CDL9, CD56, and HLA-DR). Contaminants were B cells, NK
cells, and few T cells. R� cells were further processed to obtain
an enriched T-cell fraction by overnight incubation in CCM at
37#{176}C,followed by osmotic lysis of the red cells. Nonadherent
cells were >90% T cells, as estimated by flow cytometry. In
several experiments, the DC fraction was additionally purified
using a magnetic bead separation technique. Briefly, the DC
fraction isolated as described above was treated at 4#{176}Cwith a
cocktail of monoclonal antibodies: anti-CD3, anti-CD19, anti-
CD14, and anti-CD57/HNK. After a 30-mm incubation, cells
were washed and labeled for 30 mm with goat antimouse IgO
antibody conjugated to magnetic beads (Dynal), followed by
washing and magnetic separation. Unlabeled cells were more
than 95% DCs, as estimated by flow cytometry.
DCs were generated from progenitors in peripheral blood
as described previously (12). Briefly, peripheral blood mononu-
clear cells were incubated for 2 h in six-well plates, nonadherent
cells were removed, and adherent cells were incubated with
GM-CSF (5 ng/ml) and IL-4 (5 ng/mJ) for 6-7 days. After that
time, cells were collected, washed, and analyzed. The purity of
the DCs was more than 70% with >95% viability.
MLR. The ability of DCs to stimulate allogeneic T cells
was tested in MLRs. To minimize the effect of differences in
HLA between individuals on allogeneic T-cell proliferation,
DCs from each of 14 normal controls were tested independently
against allogeneic T cells obtained from another three control
individuals, and the pairs giving maximal responses were used.
Cells from the patients were similarly tested against the panel of
normals, and those giving maximal responses were used. For
each experiment, T cells from patients were tested against DCs
from three control volunteers. DCs and T cells were cultured at
different ratios for 5 days in 96-well plates. Eighteen h prior to
harvesting the cells, 1 p.Ci of [3H}thymidine was added to each
well. [3H]Thymidine uptake was counted in a liquid scintillation
counter (Beckman, Palo Alto, CA).
Tetanus Toxoid-specific T-Cell Proliferation. DCs
were cultured with autologous T cells in the presence of 0.5
p.g/ml tetanus toxoid (this concentration was selected after
preliminary testing). [3H]Thymidine uptake was measured 5
days later. The background level of T-cell proliferation stimu-
lated by autologous DCs was subtracted from all experimental
values.
FLU-specific CTL Responses. DCs were cultured in
serum-free medium with influenza virus strain PR8 (A/Puerto
Rico/8/34) for 2 h, washed, and cultured with autologous T cells
at a ratio of 1:5 in 24-well plates (Costar, Cambridge, MA).
Seven days later, the cells were harvested and distributed at
various dilutions to 96-well plates. CTL activity was measured
against FLU-infected and uninfected 5tCr-labeled autologous
phytohemagglutinin blasts in a standard 6-h assay. At the ter-
mination of the assay, � ‘Cr release into the supernatants was
measured using absorption cartridges (Skatron Instruments, Inc.,
Sterling, VA) and counted in a gamma counter. The percentage
of specific 51Cr release was calculated from the following
formula:
% specific 51Cr release =
Release by CTL - spontaneous release
Total release - spontaneous release X 1
Spontaneous release was equal to or less than 25% of the total
release.
RESULTS
Defective Responses to Mitogenic and Antigemc Stim-ulation in Patients with Breast Cancer. We first examined
antigen- and mitogen-driven T-cell proliferation in patients with
breast cancer. Because this required the presence of APCs, DCs
and T cells were isolated from the peripheral blood of breast
cancer patients and mixed at different concentrations with var-
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
A 140000
120000
100000
80000
060000
40000
20000
0
B 20000
18000
16000
14000
12000
� 100000
8000
6000
4000
2000
�- --
ConA
stimulation
++
C �60000
50000 �-40000
�a
30000
20000
10000
0 � - --Control(n=14)
D 70000
60000
50000
40000
30000
20000
1 0000
0
.�jj�_jI �[�1. ri�Stage Stage Stage Stage
1 (n=3) 2 (n=7) 3 (n=4) 4anti-CD3
stimulation
Control Stage Stage Stage Stage(n=14) 1(n=3) 2 (n=7) 3 (n=4) 4 (n=4)
Control Stage Stage Stage Stage(n=14) 1(n=3) 2 (n=7) 3 (n=4) 4 (n=4)
Fig. I Function of DCs and I cells in patients with breast cancer. A, T-cell responses are decreased in response to ConA but maintained for anti-CD3
stimulation. For ConA stimulation, DCs and T cells were isolated as described in “Patients and Methods” and cultured together at a ratio of 1:50 (2 X
lO� DCs and lO� I cells) in the presence of 10 p.g/ml ConA. Averages (bars, SE) of 3H uptake on day 3 are shown. EJ, statistically significant
differences from control (P < 0.05). For anti-CD3 stimulation, 96-well plates were coated with S pg/mI anti-CD3 antibody and washed; then lO� T
cells were added to each well. 3H uptake in triplicates was measured on day 3. �, control (ii = I I); �, stage 1-2 (,z = 5); E, stage 3-4 (n = 4).
B, U-stimulated T-cell proliferation is decreased in patients with breast cancer. DCs and T cells were isolated as described in “Patients and Methods”
and mixed at different ratios (from 1:20 to 1: 160). iT-driven T-cell proliferation was tested as described in “Patients and Methods.” Averages (bars,
SE) of [3H]thymidine at a 1 :40 DC:T cell ratio are shown. Similar patterns of the results were found for all DC:T cell ratios. �, statistically significant
differences from control (P < 0.05). C, DCs from breast cancer patients show decreased stimulation of control allogeneic T cells. Patients’ DCs and
donors’ T cells were isolated, and allogeneic MLR was performed as described in “Patients and Methods.” The means (bars, SE) of [3H]thymidineuptake at a 1 :40 DC:T cell ratio are shown. Similar patterns of the results were found for all DC:T cell ratios. �, statistically significant differences
from control (P < 0.05). D, DCs from control individuals effectively stimulate breast cancer patients’ T cells. Experimental protocol was as describedin “Patients and Methods.” Means (bars, SE) of [3H]thymidine uptake at a 1 :40 DC:T cell ratio are shown. Similar patterns of the results were found
for all DC:T cell ratios.
Clinical Cancer Research 485
ious stimulators. Interestingly, patients with advanced breast
cancer demonstrated a substantially reduced yield and purity of
DCs, as assessed by flow cytometry and light microscopy. The
proportion of DCs in the fractions isolated from patients with
stages I and II was the same as that from control individuals
(55-70%), but samples from patients with stages III and IV
tumors yielded a substantially lower DC purity than control
(25-35%). Because of these differences in the purity of DCs,
DC numbers were always carefully adjusted to equal concen-
trations before any assays were performed. Significant decreases
in the ability of T cells to respond to ConA stimulation were
seen in patients with progressively advanced stages of breast
cancer (Fig. hA). This was accompanied by a dramatic reduction
in the response to the soluble antigen ‘FT (Fig. 1B). Defects in
‘VT-dependent T-cell proliferation were seen in patients at ear-
her stages (stage II) than the defects in responses to mitogens
(stage III; Fig. 1B). The majority of patients with stage II breast
cancer demonstrated normal or only slightly decreased levels of
influenza virus-specific CTL responses. However, a significant
decrease in this response was observed in patients with more
advanced disease (Fig. 2).
Defective Function of DCs but not T Cells in Breast
Cancer Patients. These data are consistent with defective
function of either T cells or APCs. To characterize the mecha-
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
40�
35-
� 30.
C
C 25
L)
a�
10.
60000
50000
40000
0
0
p
0
0
.
.
.
0
20000
10000
0
486 Decreased Antigen Presentation by DC in Cancer
5.
I #{149} I
Control Stage Stagedonors I-Il lll-IV
Fig. 2 Anti-FLU-specific CU responses are decreased in patients with
breast cancer. CTL assays were performed as described in “Patients and
Methods” at an E:T ratio of 40: 1 . Each point represents the cytotoxicityassay results from one individual. Bar, average.
nism of the observed effects, we purified DCs and T cells from
patients and controls and tested the ability of both patient and
control DCs to stimulate allogeneic T cells. As shown in Fig.
1C, a reduced activity of DCs (not statistically significant, P =
0.07) was observed in patients with stages I and II breast cancer.
A significant decrease in stimulation was observed in patients in
stages III and IV. In contrast, no differences were found in the
isolated T cells of these patients, for stimulation of patient T
cells with either control allogeneic DCs or immobilized anti-
CD3 antibody resulted in normal T-cell proliferation, even in
patients in stage IV (Fig. 1, A and D). These data indicate that
defects in the functional activity of DCs could be one of the
major reasons for the observed decreased cellular immune re-
sponses in patients with advanced breast cancer. Because the
purity of the DC preparation obtained from patients with ad-
vanced disease was lower than that from control individuals,
there is always the possibility that described effects could be due
to contamination of DC fractions by other cells. To evaluate
this, several experiments were performed using highly purified
DCs. In these experiments, DCs were additionally purified using
magnetic bead separation, and their ability to stimulate control
allogeneic T cells was compared with that of the partially
purified DCs described previously. Even the highly purified
DCs isolated from patients with stage III tumors showed the
same low level of activity as those from partially purified
fractions (Fig. 3).
In the next group of experiments, we tested the hypothesis
that changes in the level of expression of molecules on the
surface of DCs might contribute to their defective function. DCs
were isolated from four control donors and five patients with
stage III breast cancer. Cells were double labeled with a cocktail
of monoclonal antibodies against T, B, and NK cells and mono-
cytes and with antibodies against MHC class I, class II, and B-7
molecules. Only DC (“cocktail-negative” cells) were analyzed.
As shown in Fig. 4, no differences were observed in the level of
Ea_ 30000C-)
50 100 200 400
T cell:DC ratio
Fig. 3 Additional purification of DCs did not improve their specific
function. DCs were isolated from patients with stage III breast cancer asdescribed in “Patients and Methods”. One-half of the cells were addi-
tionally purified using monoclonal antibodies against T, B, and NK cells
and monocytes using magnetic beads as described in “Patients andMethods.” Control allogeneic T cells were stimulated with DCs fromanother donor, and equal numbers of partially and highly purified DCswere obtained from the patient. Allogeneic MLR was performed intriplicate as described in “Patients and Methods”. Means (bars, SD) areshown. #{149}, control; �, pure DCs; A, DC fraction.
expression of MHC class I molecules (HLA A, B, and C),
whereas a substantial decrease in the expression of MHC class
II (HLA DR and HLA DQ) was seen. A decline in the expres-
sion of B7-l and B7-2 was also notable, albeit less profound
than that of MHC class II molecules.
Function of DCs Generated from Progenitors. Theabove data demonstrate serious defects in the function of pe-
ripheral blood DCs in patients with advanced breast cancer. We,
therefore, investigated the mechanisms of this dysfunction. Pre-
viously, we have shown that soluble factors released by animal
and human tumor cells affected the maturation of DCs but did
not impair the function of mature DCs from peripheral blood
(4). We, therefore, hypothesized that DCs generated in vitro
from precursors from advanced cancer patients, without the
influence of any tumor-produced factors, could be functionally
intact. To test this hypothesis, peripheral blood from patients
with stage III and control donors was processed to obtain T cells
and DCs as described before. In parallel, DCs were generated
from precursors using OM-CSF and IL-4. Generation of DCs
from precursors requires at least 7 days of in vitro culture.
Mature DCs isolated from peripheral blood are optimally func-
tional on the day after harvest. To use autologous T cells and to
standardize experimental conditions, T cells from each patient
were cryopreserved in ahiquots immediately after isolation and
thawed before testing. DCs generated from precursors showed
high levels of expression of MHC class II and costimulatory
molecules consistent with phenotype of relatively mature DCs.
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
500
400
300
200
100
oS000.
0
p.
0.
I
� I �
0
EU
U.
I0
�. �g’
Isotype HLA-ABC HLA-DRcontrol
Surface molecules
Control donors
IiLA-L)Q b/-I b/-2
A
Patients with breast cancer
a.C-)
45000
40000
35000
30000
25000
20000
15000
10000
BAB
40 60 ‘ 40 80160 * 40
T cells: DC ratio
Fig. 5 DCs grown from the peripheral blood precursors of patientswith breast cancer are capable of normal levels of allogeneic stimula-
tion. A, DCs isolated from peripheral blood; B, DCs generated fromprecursors in peripheral blood. The experimental protocol was as de-scribed in “Patients and Methods.” [3H]Thymidine uptake by T cells on
day S is shown. #{149},donor 2; 0, donor 4; U, patient B; A, patient K; V.patient L.
Clinical Cancer Research 487
Fig. 4 Expression of MHC class Iand class II and costimulatory mole-cules on the surface of DCs isolated
from control individuals and patients
with breast cancer. DCs were directlyisolated from peripheral blood or gen-
crated from precursors in peripheralblood as described in “Patients andMethods”. Blood was obtained fromcontrol individuals and patients withstage III breast cancer. Cells were Ia-beled with a cocktail of phycoerythrin-labeled antibodies (CD3, CD19, CD56,and CD14) and FITC-labeled antibod-ies against MHC class I, MHC class II,
B7-l, and B7-2 molecules. Expres-sion of the FITC-labeled markers was
analyzed in phycoerythrin-negative
cells (DCs). The means of fluorescence
in each experiment are shown. 0, con-
trol individuals; #{149},patients. 0, cellsgenerated using OM-CSF and IL-4from precursors in peripheral bloodfrom control donors; U, DCs generatedfrom precursors in peripheral bloodfrom cancer patients.
Direct comparison of the levels of expression of surface mole-
cules between DCs generated from precursors from control
individuals and from the patients showed no differences (Fig. 4).
DCs generated from precursors in the peripheral blood of pa-
tients with breast cancer had a normal ability to stimulate
control allogeneic T cells, whereas DCs from peripheral blood
from the same patients had more than 2-fold lower stimulatory
capacity (Fig. 5). Normal levels of U-dependent T-cell stimu-
lation were observed when DCs obtained from precursors were
used as stimulators (Fig. 6). Those cells also induced a substan-
tially higher level of FLU-specific Cli responses than DCs
from peripheral blood. However, that response did not reach the
control figures (Fig. 7).
DISCUSSION
The effective function of APCs is an integral part of
immune responses in general, and antiturnor immune responses
in particular. Defects in antigen presentation would make the
development of specific immune responses almost impossible.
However, until now it was not clear whether the function of
APCs was normal or suppressed in tumor-bearing hosts, and if
suppressed, whether methods could be developed to restore
function. The few reports in this field are controversial and
incomplete. It has been shown, for example, that a subset of
I-At epidermal APCs may be capable of inducing tolerance to
tumor antigens and that activated rnacrophages might induce
structural abnormalities of the T-cell receptor/CD3 complex (13,
14). Zou et a!. (15) have reported norma] and even increased
function of APCs (mostly macrophages) in tumor-bearing mice
(15). However, defective function of macrophages in patients
with cancer has also been described in several reports ( 16, 17).
Dendritic cells are the most potent APCs; therefore, we
focused on investigation of their ability to present antigens and
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Control donors Patients
A B A
Donor 1 Donor 5 Patient W Patient P Patient T
Fig. 7 Improved stimulation of influenza virus-specific CTLs by breastcancer patient DCs grown from peripheral blood precursors. T:E ratio1 :40, experimental protocol as described in “Patients and Methods.”�,
peripheral blood DCs; 0, DCs from precursors. Bars. SE.
488 Decreased Antigen Presentation by DC in Cancer
40000�
35000
30000
25000
a. 20000C.)
15000
10000
5000
0. #{149}, , #{149}� I I I I I I I I I 1
40 80160 * 40 80160 #{149}40 80160 #{149}40 80160
T cells:DC ratio
Fig. 6 DCs grown from the peripheral blood precursors of patientswith breast cancer are capable of improved induction of U-dependentautologous T-cell proliferation. A, DCs isolated from peripheral blood.B, DCs isolated from precursors in peripheral blood. The experimentalprotocol was as described in “Patients and Methods.” #{149},donor 1 ; 0,donor 2; U, donor 3; A, patient D; V. patient N: I, patient R. Bars, SE.
induce CTL responses in patients with breast cancer. Previously,
we reported defective functioning of DCs in tumor-bearing mice
(4), and we speculated that the same defects may occur in
patients with cancer. Several other reports have suggested pos-
sible DC dysfunction in patients with cancer (16, 18). However,
these observations did not answer the main question of whether
DCs from patients with cancer are able to stimulate specific
immune responses.
In this study, we investigated the immune function of
patients with breast cancer and particularly the ability of DCs to
present antigens to T cells. Unlike animal experimental models,
it is much more difficult to analyze the function of APCs in
patients because antigen-specific T-cell lines and clones with
the same haplotype are not readily available. Therefore, we
chose two antigens (iT and influenza virus) with prevalent
immune responses in the general population. We studied a
sufficient number of control individuals and patients to reach
statistically significant conclusions. A clear inverse correlation
between the stage of breast cancer and the level of antitetanus
toxoid or FLU immune responses was found. In almost all
patients with stages III and IV breast cancer, no FLU-specific
CTL responses were observed. These patients’ cells also showed
a dramatic decrease in T-cell response to mitogeneic stimulation
with ConA. These data are consistent with defects in either T
cells or DCs because in these experimental systems, the pres-
ence of APCs is important to getting a maximal T-cell response.
To characterize the contribution of each of these types of
cells to the observed decreased function, we studied allogeneic
MLR reactions using purified cell fractions. In this system,
control allogeneic T cells were stimulated with DCs from pa-
tients and vice versa. To minimize the effect of chance MHC
matching in the DC-T cell pairs, DCs from each patient were
tested against T cells from at least three different control mdi-
viduals, and the maximum responses were scored. These studies
showed that the ability of DCs from patients with advanced
breast cancer to stimulate control allogeneic T cells was signif-
icantly impaired. Initially, we used partially purified fractions of
DCs in all of our experiments. In control individuals and pa-
tients with stages I and II, it provided 50-70% pure DCs.
However, we found that the purity of identically prepared cells
from patients with advanced disease was significantly lower.
Because we always adjusted the cell concentration so that the
DC numbers were equal, fractions from patients with stages III
and IV contained more “non DC cells” than in controls. It was
possible that these contaminating cells produced factors or di-
rectly interfered with DC function, rather than the defect being
an intrinsic property of the DC. To evaluate this possibility, we
additionally purified DCs using a magnetic bead separation
technique, effectively eliminating all T, B, and NK cells and
rnonocytes from this fraction. The remaining cells had the same
low level of activity as those from the partially purified frac-
tions, indicating that the differences we observed were not due
to possible contamination of the final DC fractions with lym-
phocytes or monocytes. However, this approach does not rule
out interference from immature cells lacking the markers spe-
cific for T, B, and NK cells and monocytes. We have shown
previously that supernatants from breast cancer cells inhibited
the functional maturation of DCs in vitro and resulted in the
production of as yet incompletely characterized immature cells
of rnyelomonocytic lineage (19). These in vitro findings could
provide an explanation for the in vivo observations reported
here. Decreased expression of MHC class II and some costimu-
latory molecules on the surface of DCs isolated from the pe-
ripheral blood of cancer patients might be one of the possible
causes of the observed impaired ability of these cells to stimu-
late allogeneic MLR and antigen-specific T-cehl responses.
These molecules are known to be directly involved in T-cell
stimulation (20, 21). CD4� cells engaged through MHC class II
molecules play an important role in induction of the CD8� -
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 489
specific CTL responses. Depletion of CD4� cells dramatically
decreased CD8� cell-mediated cytotoxicity against tumor-spe-
cific antigens (4).
A normal level of proliferation was observed when T cells
from patients were stimulated with DCs from control individu-
als. To confirm the normal proliferative capacity of the T cells
from cancer patients, we tested their responses to immobilized
anti-CD3 antibody, which is able to stimulate T-cehl prolifera-
tion almost independently of the presence of APCs through
direct triggering of the CD3/T-cell receptor complex (22). Using
this approach, we did not find any differences in the level of
proliferative responses between control individuals and breast
cancer patients. These data indicate that peripheral blood T cells
from these patients are capable of normal responses to these
stimuli.
The molecular mechanism of this DC dysfunction is not
clear. Previously, we proposed that defects in DC maturation
mediated by soluble substances released by tumor cells could be
partly responsible ( 10). DCs generated from bone marrow pro-
genitors from tumor-bearing mice showed a normal level of
activity (presentation of the soluble antigen to CTL and stimu-
lation of allogeneic T cells). Recently, we found that vascular
endothelial growth factor released by tumor cells was one of the
factors responsible for inhibiting the functional maturation of
DCs, without affecting the function of the relatively mature cells
in peripheral blood ( 19). The data presented here suggest that
defective DC function consistent with the defects observed in
vitro can be observed in patients with breast cancer. They
further suggest that this defect may be caused by tumor-released
soluble factor(s) as seen in vitro.
To explore the latter hypothesis, we generated DCs from
peripheral blood precursors from these breast cancer patients
using OM-CSF and IL-4. This combination of cytokines pro-
motes the growth of DCs with high activity for taking up and
processing soluble antigens (12, 23). Several studies have dem-
onstrated that it is possible to generate DCs from progenitors in
patients with cancer using similar combinations of cytokines
(24, 25). The cells obtained after this procedure had the pheno-
type of DCs and were able to stimulate antigen-dependent T-cell
responses. However, in all of these cases, the patients were
pretreated with chemotherapy and granulocyte-CSF, which
could influence the function of the derived cells, and in addition,
the function of freshly isolated peripheral blood DCs from the
same patient was not investigated. Therefore, from these data it
is impossible to conclude whether those patients had any defects
in DC function and whether DCs generated from precursors had
any functional advantage over those obtained from peripheral
blood. To answer all of these questions, we compared the
function of DCs isolated from peripheral blood with cells gen-
crated from precursors from the same patients. We showed that
in control individuals, there were no significant differences
between the functional activity of DCs isolated from peripheral
blood and of those generated from precursors. U-specific T-
cell proliferation in patients with breast cancer was completely
restored when DCs generated from precursors were used as
APCs. Increased FLU-specific CU responses were also ob-
served, albeit not quite to normal levels. Why were DCs gen-
crated from precursors in peripheral blood functionally compe-
tent whereas those directly isolated from peripheral blood were
not? One of the possible explanations could be the fact that the
DCs we generated in the presence of OM-CSF and IL-4 were
derived not from CD34� stem cell precursors but from cells
which have the phenotypic features of monocytes (26-28).
Therefore, it is possible that tumor-derived factors specifically
affect the ability of CD34� cells to different into DCs while not
altering the production of these monocytoid cells or their ability
to be diverted to a DC phenotype in vitro. A second explanation
could relate to the fact that DCs isolated from blood are almost
terminally matured cells and do not undergo cell division,
whereas cells we generate in vitro in the presence of OM-CSF
and IL-4 actively divide (at least during the first 3-4 days in
culture). It is possible then that rather than a selective effect on
a different lineage of cells, several cycles of cell division in the
absence of the tumor-derived factors are required to produce
fully functional DCs, regardless of the lineage of the precursor.
These questions require additional clarification.
Several major conclusions can be drawn from these find-
ings: (a) cellular immune deficits in patients with advanced
cancer can at least partially be explained by a decreased number
and potency of DC; (b) defective DC function in patients with
advanced stage breast cancer can be at least partially overcome
when these cells are generated in vitro from precursors; (c)
T-cell hyporesponsiveness to soluble antigens can be overcome
by stimulation with functionally potent DCs generated from
precursors. These results not only suggest that the antigen pres-
entation function of DCs may be selectively targeted by cancer
cells as a means of immune escape but also that DCs generated
from precursors ex vivo may have a significant advantage over
DCs obtained directly from peripheral blood as vehicles for
antigen delivery in the immunotherapy of cancer.
REFERENCES
1 . Ioannides, C., and Whiteside, T. T cell recognition of human tumors:implication for molecular immunotherapy of cancer. Clin. Immunol.
Immunopathol., 66: 91-106, 1993.
2. Cozzolino, F., Torcia, M., Carossino, A. M., Oiordani, R., Selli, C.,
Talini, 0., Reali, E., Novelli, A., Pistole, V., and Ferrarini, M. Charac-
terization of cells from invaded nodes in patients with solid tumors.Lymphokine requirement for tumor-specific lymphoproliferative re-
sponse. J. Exp. Med., 166: 303-318, 1987.
3. Belldegrun, A., Kasid, A., Uppenkamp, M., Topalian, S. L., and
Rosenberg, S. A. Human tumor infiltrating lymphocytes: analysis oflymphokine mRNA expression and relevance to cancer immunotherapy.
J. Immunol., 42: 4520-4526, 1989.
4. Gabrilovich, D., Ciernik, F., and Carbone, D. P. Dendritic cells in
anti-tumor immune responses. I. Defective antigen presentation in tu-mor-bearing hosts. Cell. Inimunol., 170: 101-1 10, 1996.
5. Huang, A. Y. C., Golumbek, P., Ahmadzadeh, M., Jaffee, E., Pardoll,
D., and Levitsky, H. Role of bone marrow-derived cells in presentingMHC class I-restricted tumor antigens. Science (Washington DC), 264:961-965, 1994.
6. Rock, K. L., Rothstein, L., Gamble, S., and Fleischacker, C. Char-
acterization of antigen-presenting cells that present exogenous antigensin association with class I MHC molecules. J. Immunol., 150: 438-446,1993.
7. Zitvogel, L., Mayordomo, J. I., Tjandrawan, T., DeLco, A. B.,
Clarke, M. R., Lotze, M. T., and Storkus, W. J. Therapy of murinetumors with tumor peptide-pulsed dendritic cells: dependence on Tcells, B7 costimulation, and T hepler cell 1-associated cytokines. J. Exp.Med., 183: 87-97, 1996.
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
490 Decreased Antigen Presentation by DC in Cancer
8. Mayordomo, J. I., Zorina, T., Storkus, W. J., Celluzzi, C., Falo, L. D.,
Melief, C. J., Ildstad, T., Kast, W. M., DeLco, A. B., and Lotze, M. T.Bone-marrow derived dendritic cells pulsed with synthetic tumour pep-tides elicit protective and therapeutic antitumour immunity. Nat. Med.,
I: 1297-1302, 1995.
9. Hsu, F., Benike, C., Fagnoni, F., Liles, T. M., Czerwinsky, D., Taidi,B., Engleman, E. G., and Levy, R. Vaccination of patients with B-celllymphoma using autologous antigen-pulsed dendritic cells. Nat. Med.,2: 52-58, 1996.
10. Gabrilovich, D. I., Nadaf, S., Corak, J., Berzofsky, J. A., andCarbone, D. P. Dendritic cells in anti-tumor immune responses. II.Dendritic cells grown from bone marrow precursors, but not mature DC
from tumor-bearing mice are effective antigen carriers in the therapy ofestablished tumors. Cell. Immunol., 170: 111-120, 1996.
I I . Bhardwaj, N., Bender, A., Gonzalez, N., Kim Bui, L., Garrett,M. C., and Steinman, R. M. Influenza virus-infected dendritic cellsstimulate strong proliferative and cytolytic responses from human
CD8+ T cells. J. Clin. Invest., 94: 797-807, 1994.
12. Romani, N., Gruner, S., Brang, D., Kampgen, E., Lenz, A.,
Trockenbacher, B., Konwalinka, 0., Fritsch, P., Steinman, R. M., andSchuler, G. Proliferating dendritic cell progenitors in human blood.
J. Exp. Med., 180: 83-93, 1994.
I 3. Aoe, T., Okamoto, Y., and Saito, T. Activated macrophages induce
structural abnormalities of the T cell receptor-CD3 complex. J. Exp.
Med., /81: 1881-1886, 1995.
14. Tan, K. C., Hosoi, J., Grabbe, S., Asahina, A., and Granstein, R. D.
Epidermal cell presentation of tumor-associated antigens for induction
of tolerance. J. Immunol., 153: 760-767, 1994.
15. Zou, J-P., Shimizu, J., Ikegame, K., Yamamoto, N., Ono, S.,
Fujiwara, H., and Hamaoka, T. Tumor-bearing mice exhibit a progres-
sive increase in tumor antigen-presenting cell function and a reciprocal
decrease in tumor antigen responsive CD4+ T cell activity. J. Immunol.,148: 648-655, 1992.
16. Tas, M., Simons, P., Balm, F., and Drexhage, H. Depressed mono-cyte polarization and clustering of dendritic cells in patients with head
and neck cancer: in vitro restoration of this immunosuppressionby thymic hormones. Cancer Immunol. Immunother., 36: 108-1 14,
1993.
17. Erroi, A., Sironi, M., Chiaffarino, F., Chen, Z. 0., Mengozzi, M.,
and Mantovani, A. ILl and 1L6 released by tumor-associated macro-phages from human ovarian carcinoma. lnt. J. Cancer, 4/: 65-68, 1989.
18. Toriyama, K., Wen, D. R., Paul, E., and Cochra, A. J. Variations inthe distribution, frequency, and phenotype of Langerhans cells duringthe evolution of malignant melanoma of the skin. J. Invest. Dermatol.,100: 269S-273S, 1993.
19. Gabrilovich, D. I., Chen, H. L., Girgis, K. R., Cunningham, H. T.,Meny, G. M., Nadaf, S., Kavanaugh, D., and Carbone, D. P. Productionof vascular endothelial growth factor by human tumors inhibits thefunctional maturation ofdendritic cells. Nat. Med., 2: 1096-1 103, 1996.
20. Liu, Y., and Linsley, P. S. Costimulation of T cell growth. Curr.
Opin. Immunol., 4: 265-270, 1992.
21. Wade, W. F., Davoust, J., Salamero, J., Andre, P., Watts, T. H., and
Cambier, J. C. Structural compartmentalisation of MHC class II signal-ling function. Immunol. Today, 14: 539-546, 1993.
22. Geppert, T. D., and Lipsky, P. E. Activation of T lymphocytes by
immobilized monoclonal antibodies to CD3. Regulatory influences of
monoclonal antibodies to additional T cell surface determinants. J. Clin.
Invest., 81: 1497-1505, 1988.
23. Sallusto, F., and Lanzavecchia, A. Efficient presentation of soluble
antigen by cultured human dendritic cells is maintained by granulocyte/
macrophage colony-stimulating factor plus interleukin 4 and downregu-lated by tumor necrosis factor. J. Exp. Med., 179: 1109-11 18, 1994.
24. Bernhard, H., Disis, M. L., Heimfeld, S., Hand, S., Gralow, J. R.,
and Cheever, M. A. Generation of immunostimulatory dendritic cellsfrom human CD34+ hematopoietic progenitor cells of the bone marrow
and peripheral blood. Cancer Res., 55: 1099-1 104, 1995.
25. Fisch, P., Koller, 0., Garbe, A., Herbst, B., Wider, D., Kohler, H.,
Schaefer, H. E., Mertelsmann, R., Brugger, W., and Kanz, L. Generationof antigen-presenting cells for soluble protein antigens ex vivo fromperipheral blood CD34+ hematopoietic progenitor cells in cancer pa-tients. Eur. J. Immunol., 26: 595-600, 1996.
26. Thomas, R. L., Davis, L. S., and Lipsky, P. E. Isolation and
characterization of human peripheral blood dendritic cells. J. Immunol.,
150: 821, 1993.
27. O’Doherty, U., Peng, M., Gezelter, S., Swiggard, W. J., Betjes, M.,
Bahardawaj, N., and Steinman, R. M. Human blood contains twosubsets of dendritic cells, one immunologically mature and the other
immature. Immunology, 82: 487-493, 1994.
28. Pickl, W. F., Majdic, 0., Kohl, P., Stockl, J., Riedl, E., Scheinecker,
C., Bello-Fernandez, C., and Knapp, W. Molecular and functionalcharacteristics of dendritic cells generated from highly purified CD 14+peripheral blood monocytes. J. Immunol., 157: 3850-3859, 1996.
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1997;3:483-490. Clin Cancer Res D I Gabrilovich, J Corak, I F Ciernik, et al. with breast cancer.Decreased antigen presentation by dendritic cells in patients
Updated version
http://clincancerres.aacrjournals.org/content/3/3/483
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/3/3/483To request permission to re-use all or part of this article, use this link
Research. on October 7, 2020. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from