- Email: [email protected]
Human T cells modulate myeloid-derived suppressor cells through a TNF-α-mediated mechanism
Accepted Manuscript Title: Human T cells modulate myeloid-derived suppressor cells through a TNF-␣-mediated mechanism Authors: Markus Bauswein, Anurag Singh, Anjali Ralhan, Davide Neri, Katharina Fuchs, Kelly Daryll Blanz, Iris Sch¨afer, Andreas Hector, Rupert Handgretinger, Dominik Hartl, Nikolaus Rieber PII: DOI: Reference:
S0165-2478(18)30136-6 https://doi.org/10.1016/j.imlet.2018.07.010 IMLET 6229
To appear in:
Immunology Letters
Received date: Revised date: Accepted date:
10-3-2018 23-7-2018 31-7-2018
Please cite this article as: Bauswein M, Singh A, Ralhan A, Neri D, Fuchs K, Blanz KD, Sch¨afer I, Hector A, Handgretinger R, Hartl D, Rieber N, Human T cells modulate myeloid-derived suppressor cells through a TNF-␣-mediated mechanism, Immunology Letters (2018), https://doi.org/10.1016/j.imlet.2018.07.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Bauswein et al.
Human T cells modulate myeloid-derived suppressor cells through a TNF-α-mediated mechanism
Markus Bauswein1*, Anurag Singh1*, Anjali Ralhan1, Davide Neri1, Katharina Fuchs1, Kelly
Nikolaus Rieber1,3
IP T
Daryll Blanz1, Iris Schäfer1, Andreas Hector1, Rupert Handgretinger1, Dominik Hartl1,2 and
Department of Pediatrics I, University of Tuebingen, Tuebingen, Germany
2
Roche Pharma Research & Early Development (pRED), Immunology, Inflammation and
SC R
1
U
Infectious Diseases (I3) Discovery and Translational Area, Roche Innovation Center Basel,
Department of Pediatrics, Kinderklinik Muenchen Schwabing, Klinikum Schwabing, StKM
A
3
N
Switzerland
M
GmbH und Klinikum rechts der Isar, Technical University of Munich, Munich, Germany
Correspondence:
CC E
PT
[email protected]
ED
*contributed equally
Highlights:
A
Unstimulated CD4+, but not CD8+ T cells, induce polymorphonuclear myeloid-derived suppressor cells (PMN-MDSC). The T-cell mediated induction of PMN-MDSC is dependent on direct cell-cell contact and requires transmembrane TNF-α signaling. Stimulated human CD3+ T cells delay the apoptosis of PMN-MDSC Human T cells modulate MDSC generation and survival
Abstract 1
Bauswein et al.
Manuscript body text word count: 3253
Abstract
IP T
Myeloid-derived suppressor cells (MDSC) represent an innate immune cell subset capable of suppressing T-cell responses in cancer and chronic inflammation. While the effect of MDSC
SC R
on T cells has been defined thoroughly, the reciprocal impact of T cells on MDSC homeostasis
remains poorly understood. Therefore, we comprehensively analyzed the effect of different T-
U
cell subsets on the generation and survival of human MDSC. Using an in vitro MDSC
N
generation assay, we demonstrate that unstimulated CD4+, but not CD8+ T cells, induce
A
polymorphonuclear MDSC (PMN-MDSC) from CD33+ myeloid cells. This effect was
M
dependent on direct cell-cell contact and required TNF-α signaling. Soluble TNF-α was dispensable for PMN-MDSC generation, suggesting that transmembrane TNF-α is involved in
ED
that trans-cellular process. Stimulated human CD3+ T cells delayed the apoptosis of PMNMDSC, which was independent of TNF-α signaling or direct cell-cell contact, but was
PT
recapitulated by IL-2. Taken together, our study shows that human T cells modulate MDSC
CC E
generation and survival through two distinct mechanisms and thereby fine-tune the homeostasis
A
of human MDSC in a regulated manner.
1. Introduction Myeloid-derived suppressor cells (MDSC) represent a novel innate immune cell subset that develop
under
disease
conditions
like
tumor,
infective
and
proinflammatory 2
Bauswein et al.
microenvironments. [1, 2] These myeloid cells are characterized by their capacity to potently suppress T-cell responses.[1] MDSC include two major subsets based on their phenotypical and morphological features: polymorphonuclear (PMN-) and monocytic (M-) MDSC. These subsets show unique, yet partially overlapping functional and biochemical characteristics.[1, 3-
CD66b+CD33+CD14-CD15+ and M-MDSC as CD33+CD14+HLA-DRlow.[4]
IP T
6] Phenotypically, human PMN-MDSC have most consistently been determined as
While the effect of MDSC on their functional target cells has been defined thoroughly
SC R
in mice and human cell systems, conversely, the reciprocal influence of immune cells on the
homeostasis and function of MDSC has not been extensively studied. Particularly, how T cells,
U
as the main functional targets of MDSC, reciprocally affect MDSC remains still incompletely
N
understood.[7] Several previous studies provided evidence on a dynamic and reciprocal
A
interaction between MDSC and T cells i) An influence of Th1 cells on the accumulation of
M
preferentially PMN-MDSC in an IFN-γ dependent manner was demonstrated in a mouse model of autoimmune hepatitis.[8] ii) Another study demonstrated that γδ Th17 cells drive the
ED
accumulation of PMN-MDSC in human colorectal cancer.[9] iii) In addition, Nagaraj et al. described in mice that antigen-specific CD4+ T cells promote the transition from antigen-
PT
specific MDSC, i.e. MDSC that present cognate antigens via MHC class I or II and suppress
CC E
antigen-specific T cells in a cell-contact dependent manner, to antigen-unspecific MDSC by crosslinking MHC class II on MDSC.[10] As a consequence, they proposed that MDSC are
A
part of a negative feedback loop which physiologically regulates T cell responses.[10] Based on these previous findings, we aimed to further dissect the influence of T cells on
human MDSC. Here we demonstrate for the first time that human CD4+ T cells in an in vitro MDSC generation system induce PMN-MDSC in a TNF-dependent manner. In addition, activated human T cells increased the survival of co-cultured PMN-MDSC. These mechanisms
3
Bauswein et al.
of T cell – MDSC interaction may serve as a future therapeutic target in malignant and nonmalignant disease conditions.
IP T
2. Materials and methods 2.1. Human blood samples
SC R
The study was conducted at the University Children’s Hospital Tuebingen (Germany) and all
study methods were approved by the local ethics committee. Buffy coats of healthy donors were
N
healthy volunteers (lab staff) after informed consent.
U
obtained from the blood donor center Tuebingen. Heparinized whole blood was taken from
M
A
2.2. Isolation of cells
PBMC were prepared from buffy coats / heparinized whole blood (see 2.1.) by Ficoll density
ED
gradient sedimentation (Biocoll Separating Solution; Biochrom) and washed twice in RPMI 1640 medium (Biochrom). Myeloid cells were isolated from the PBMC fraction by labelling
PT
all cells with anti-CD33 magnetic beads (Miltenyi Biotec) followed by two sequential magnetic bead separation steps according to the manufacturer´s protocol. For the isolation of different T-
CC E
cell subsets, all cells from the PBMC fraction were labelled either with anti-CD3 magnetic beads, or with anti-CD4 magnetic beads, or with anti-CD8 magnetic beads (all Miltenyi Biotec)
A
followed by two sequential magnetic bead separation steps according to the manufacturer´s protocol. For cell survival experiments, PMN-MDSC were defined as CD66b+ cells in the Ficoll low density fraction of freshly isolated heparinized whole blood from healthy volunteers. PMNMDSC were isolated by staining all cells in the Ficoll low density fraction (“PBMC fraction”) with FITC labeled anti-CD66b antibodies (BD), followed by two sequential anti-FITC magnetic bead separation steps (Miltenyi Biotec). 4
Bauswein et al.
2.3. In vitro generation and characterization of human MDSC Human MDSC were generated from myeloid PBMC, similarly to a recently published protocol.[11, 12] Isolated human CD33+ (myeloid) PBMC were cultured at a cell ratio of 500,000 cells/ml in RPMI 1640 medium supplemented with 10% FCS, 2mM L-glutamine, 100 IU/ml penicillin and 100 mg/ml streptomycin (all Biochrom) for 6d. 48-, 24- and 12-well flat
IP T
bottom plates (Corning) as well as 25cm2 culture flasks (BD) were used. For direct co-cultures, T cells at a concentration of 500,000 cells/ml were added to the myeloid PBMC in a cell ratio
SC R
of 5:1 (1,000,000 T cells; 200,000 myeloid PBMC). For transwell assays (ThinCerts for 24well plate; pore size 0.4 µm; greiner bio-one) T cells at a concentration of 5,000,000 cells/ml
U
within the inserts were cultured together with myeloid PBMC in a cell ratio of 5:1. The
N
following supplements were added as indicated: GM-CSF (10 ng/ml; genzyme Bayer
A
HealthCare), TNF-α (concentrations from 0.5 to 10 µg/ml; biomol), etanercept (10 µg/ml / 20
M
µg/ml; Wyeth Pharma), adalimumab (32 µg/ml; AbbVie Ltd) and infliximab (120 µg/ml; Janssen Biologics). Cells were incubated in a humidified atmosphere at 37˚C and 5% CO2.
ED
Medium and supplements were refreshed on day 4. On day 6 cells were harvested using Detachin (Genlantis) and stained with the following antibodies according to the manufacturer´s
PT
protocol: anti-CD14-FITC (BD), anti-CD33-PE (Miltenyi Biotec), anti-HLA-DR-PerCP (BD),
CC E
or anti-CD33-APC (BD). Rabbit serum (invitrogen) was added to block unspecific antibody binding.
A
2.4. T cell suppression assay MDSC were generated from myeloid cells of the PBMC fraction as described above and isolated from cell cultures by magnetic bead cell sorting for CD33. If indicated we depleted CD3+ cells first (anti-CD3 magnetic beads) before sorting for CD33. Responder-PBMC were obtained either from buffy coats or from healthy volunteers´ heparinized full blood and stained with CFSE (life technologies) according to the manufacturer´s protocol. CFSE-labelled PBMC 5
Bauswein et al.
were stimulated with 100 U/ml IL-2 (R&D Systems) and 1 µg/ml OKT3 (Janssen-Cilag). Both MDSC and CFSE-labelled PBMC were added to RPMI 1640 medium supplemented with 10% human serum, 2mM L-glutamine, 100 IU/ml penicillin and 100 mg/ml streptomycin. In a 96well round bottom plate (greiner bio-one), either 10,000 / 30,000 MDSC or, as a control supplemented medium only, were added to 60,000 PBMC per well. Cells were incubated in a
IP T
humidified atmosphere at 37˚C and 5% CO2. On day 5 cells were harvested and stained with
anti-CD8a-APC, anti-CD4-PE antibodies (BioLegend) and propidium iodide (BD). PI positive
SC R
cells were excluded in flow cytometry. CFSE signals of CD4+ and CD8+ PBMC were analyzed. 2.5. Annexin V assay
U
Isolated CD66b+ PBMC (PMN-MDSC) were cultured in 48- and 24-well flat bottom plates
N
(Corning) at a cell ratio of 500,000 cells/ml in RPMI 1640 medium (Biochrom) supplemented
A
with 10% human serum (several healthy donors), 2mM L-glutamine, 100 IU/ml penicillin and
M
100 mg/ml streptomycin. For direct co-cultures, T cells at a concentration of 500,000 cells/ml were added to the PMN-MDSC in a cell ratio of 5:1. For transwell assays (ThinCerts for 24-
ED
well plate; pore size 0.4 µm; greiner bio-one) T cells at a concentration of 5,000,000 cells/ml
PT
within the inserts were cultured together with PMN-MDSC in a cell ratio of 5:1. The following supplements were added as indicated: IL-2 (R&D Systems), OKT3 (Janssen-Cilag), etanercept
CC E
(20 µg/ml; Wyeth Pharma). After incubation in a humidified atmosphere at 37˚C and 5% CO2 for one day, cells were collected and stained with FITC anti-CD66b antibodies (BD), Annexin
A
V-APC (BD) and propidium iodide (BD). 2.6. Measurement and analysis Flow cytometry was performed on a FACS Calibur (BD), data was analyzed with CellQuest analysis software (BD) and FlowJo 10 (FlowJo). 2.7. Statistics 6
Bauswein et al.
Graphics and statistical analysis were performed in Graph Pad Prism 6. For CFSE proliferation assays a Kruskal-Wallis test was performed as group analysis, followed by multiple one-tailed Mann-Whitney tests for groups of interest based on an a priori hypothesis. For MDSC surface characterization and survival assays Kruskal-Wallis tests were performed as group analyses, followed by Dunn´s multiple comparisons tests as post-tests. For comparisons of only two
IP T
groups, Mann-Whitney tests were performed. A significant difference was assumed for p ≤ 0.05
SC R
and is indicated by an asterisk.
N
U
3. Results
A
3.1. T cells induce PMN-MDSC
M
To define the impact of T cells for MDSC generation, we built on an established in vitro MDSC generation system[11, 12] to induce MDSC from myeloid cells out of the human PBMC
ED
fraction. For this purpose, PBMCs were obtained from buffy coats by Ficoll density gradient
PT
sedimentation. Subsequently, myeloid cells (CD33+) and CD3+ T cells from the PBMC fraction were isolated using magnetic bead cell sorting. The isolated myeloid cells from the PBMC
CC E
fraction were cultured for 6 days either in medium only or under addition of GM-CSF, a cytokine well-known for its potential to generate MDSC, or in co-culture with CD3+ T cells. In order to assess the suppressive potential of the myeloid cells in culture, we performed T cell
A
suppression assays. Myeloid cells were isolated from the cell cultures after 6 days, again using magnetic bead isolation for CD33. The effect of the isolated myeloid cells on the proliferation of polyclonal stimulated allogeneic T cells was analyzed by flow cytometry. Myeloid cells that had been cultured for 6 days efficiently suppressed polyclonal T cell proliferation (Fig. 1 and S1). 7
Bauswein et al.
After co-culture with T cells the surface marker profile of the generated suppressive myeloid cells was CD33+CD11b+CD16+CD14-, consistent with polymorphonuclear MDSC (PMN-MDSC) (Fig. 2B). When compared to 6 day culture in medium only we recognized a striking downregulation of CD14 when MDSC were generated in the presence of T cells (Fig. 2B). Cytospins showed a substantial number of cells with segmented nuclei (Supplemental
IP T
Figure S2). CD14- MDSC have been classified as PMN-MDSC in contrast to CD14+ monocytic MDSC (M-MDSC).[4] When MDSC were generated in the presence of T cells 91% of myeloid
SC R
cells exhibited the CD14- PMN-MDSC phenotype, whereas only 11% or 39% did, when
myeloid cells were cultured in medium only and under addition of GM-CSF, respectively (Fig.
N
3.2. CD4+, but not CD8+ T cells induce PMN-MDSC
U
2C).
A
To dissect which human T-cell subsets were responsible for the PMN-MDSC generation in
M
our experimental system, we compared CD4+ and CD8+ T cells side-by-side in the respective MDSC generation assay. For the isolation of different T-cell subsets, PBMC were labelled
ED
either with anti-CD4 or with anti-CD8 magnetic beads, followed by two sequential magnetic
PT
bead separation steps. For co-cultures, CD4+ or CD8+ T cells at a concentration of 500,000 cells/ml were added to the myeloid PBMC in a cell ratio of 5:1 (1,000,000 T cells; 200,000
CC E
myeloid PBMC). These studies showed that CD4+ T cells led to a proportion of 84% PMNMDSC (mean, n=7), whereas CD8+ T cells did not significantly induce PMN-MDSC (mean
A
8% PMN-MDSC, n=6) (Fig. 2D) in these assay systems. 3.3. T-cell mediated induction of PMN-MDSC requires direct cell-cell contact and transmembrane TNF-α Next, we addressed the question if the T-cell mediated induction of PMN-MDSC requires direct cell-cell contact, which is required for a variety of MDSC-mediated T cell suppression effector responses. For this purpose we used a transwell system and found that T-cell mediated induction 8
Bauswein et al.
of PMN-MDSC required direct cell-cell contact (Fig. 2E). In search for possible cellular pathways involved in T-cell / MDSC homeostasis we speculated on a role for the transmembrane T-cell cytokine TNF-α, which has recently been reported in T cell-mediated MDSC activation in mice.[13] Adding the TNF-α inhibitor etanercept [20 µg/ml] to co-cultures with CD4+ T cells substantially decreased the percentage of PMN-MDSC to 12% (mean, n=7)
IP T
(Fig. 3A). This effect was not due to cellular toxicity of etanercept on PMN-MDSC in coculture with T cells (Fig. 3B). Additionally, we evaluated two other TNF-α inhibitors,
SC R
infliximab [120 µg/ml] and adalimumab [32 µg/ml]. Both inhibitors confirmed our previous
findings (Fig. 3A). In accordance with the role of membrane-bound TNF-α, we did not find an effect of soluble TNF-α in broad concentration ranges [0.5 µg/ml - 10 µg/ml] on the induction
U
of PMN-MDSC (Fig. 3C). When viewed in combination, these studies demonstrated that T-cell
N
mediated induction of PMN-MDSC requires direct cell-cell contact and transmembrane TNF-
M
A
α.
3.4. CD3+ T cells delay the apoptosis of PMN-MDSC
ED
To assess whether human T cells delay the apoptosis of PMN-MDSC and thereby contribute to
PT
their net accumulation, we isolated PMN-MDSC from the PBMC fraction by magnetic cell sorting against CD66b. Subsequently, we cultivated PMN-MDSC in medium only or in co-
CC E
culture with autologous CD3+ T cells [cell ratio PMN-MDSC / T cells 1:5]. On day one after the isolation we studied the proportion of living, early apoptotic and late apoptotic / necrotic PMN-MDSC by flow cytometry after staining with Annexin V and PI (Fig. 4A). These studies
A
showed that co-culture with CD3 T cells nearly doubled the survival rate of PMN-MDSC from 12% to 22% (Fig. 4B). In a further step, we activated the CD3+ T cells with OKT3 [1 µg/ml]. This led to a further increase in survival of co-cultured PMN-MDSC, whereas OKT3 alone had no direct effect on PMN-MDSC (Fig. 4C). Comparing OKT3 stimulated CD4+ and CD8+ Tcell subpopulations there was no significant difference in increasing MDSC survival (Fig. 4C). 9
Bauswein et al.
Transwell assays further revealed that the T-cell mediated increase of PMN-MDSC survival did not require direct cell-cell contact (Fig. 4D). Furthermore, etanercept [20 µg/ml] did not block T-cell mediated MDSC survival (Fig. 4D). Therefore, these studies showed that the Tcell mediated induction and survival of PMN-MDSC employs different mechanisms. Interestingly, the T-cell cytokine IL-2 alone without direct T cell help increased the proportion
IP T
of living PMN-MDSC in a dose-dependent manner up to 26% living PMN-MDSC when added in a concentration of 10,000 U/ml to PMN-MDSC (Fig. 4E). In summary, these studies
SC R
demonstrated that human CD3+ T cells delay the apoptosis of PMN-MDSC.
U
4. Discussion
N
While effects of MDSC on T cells have been established thoroughly, the reciprocal impact of
A
T cells on MDSC is incompletely understood.[7, 14, 15] Therefore, we analyzed the effect of
M
different human T-cell subsets on the generation and survival of human MDSC. We
ED
demonstrate for the first time that human T cells induce human PMN-MDSC from myeloid cells. In addition, human T cells delayed the apoptosis of co-cultured PMN-MDSC. Taken
PT
together, our study shows two different mechanisms, how T cells reciprocally influence the
CC E
homeostasis of human MDSC.
The induction of PMN-MDSC from CD33+ PBMC by T cells has not previously been
described in the literature. We would discriminate these CD33+CD11b+CD16+CD14- cells from
A
what has been designated as nonclassical CD14lowCD16+ monocytes[16] as these cells in our assays were completely CD14 negative, showed high granularity in the forward-side scatter and showed distinct T cell suppressive capacity. Furthermore, cytospins demonstrated cells with segmented nuclei, however not as distinctive as in the publication by Youn et al..[17] Some evidence, that T cells influence the fate of PMN-MDSC has already been documented. Cripps et al. demonstrated in a mouse model of autoimmune hepatitis that Th1 cells contribute to the 10
Bauswein et al.
accumulation of PMN-MDSC.[18] Wu et al. described the γδ-T17-cell mediated accumulation of PMN-MDSC in tumor tissue in patients with colorectal carcinoma.[9] This effect was due to increased cell migration, increased proliferation and prolonged PMN-MDSC survival.[9] To better define the T cell interaction leading to the induction of PMN-MDSC, we first investigated the effect of different T cell subpopulations. These experiments revealed that only CD4+, but
IP T
not CD8+ T cells, induce PMN-MDSC. Previous studies also described the selective action of certain T cell subpopulations, such as those of Th1 cells[18] or of γδ T17 cells[9] on MDSC.
SC R
TNF-α belongs to a superfamily of cytokines and receptors that mediate pleiotropic
effects in the immune system.[19] TNF-α exerts its effects both in a transmembrane form
U
(tmTNF-α) and as a soluble cytokine after it has been proteolytically released from the cell
N
membrane.[19] Several receptors, e.g. TNF receptor I (TNFRI) and TNF receptor II (TNFRII),
A
mediate the pleiotropic effects of the TNF superfamily cytokines.[19] In our studies soluble
M
TNF-α showed no significant effect on the induction of PMN-MDSC in a broad concentration range. The induction of PMN-MDSC by unstimulated CD4+ T cells, however, was dependent
ED
on direct cell-cell contacts and could be blocked by three different TNF-α inhibitors (etanercept, adalimumab and infliximab), indicating a role for tmTNF-α in this process. Effects of tmTNF-
PT
α are preferably mediated via TNFRII.[20] We, therefore, postulate that tmTNF-α on CD4+ T
CC E
cells, but not soluble TNF-α, mediates its action via TNFRII on myeloid cells. Further experiments, for example by selective blockade or selective knock-down (e.g. by microRNA) of TNF-α in T cells or TNFRII in myeloid cells, could contribute to a better understanding of
A
this process. A role for TNF-α in the T-cell mediated induction of PMN-MDSC has not been previously described. More generally, however, an impact of TNF-α on MDSC has repeatedly been demonstrated. Recently, Sade-Feldman et al. reported in a mouse model of chronic inflammation that TNF-α induces accumulation of MDSC in spleen and bone marrow by blocking the differentiation of immature myeloid cells into macrophages and dendritic cells. 11
Bauswein et al.
Furthermore, TNF-α increased the suppressive activity of MDSC.[21] Our study now broadens the spectrum of TNF-α effects on MDSC. For the experiments on PMN-MDSC apoptosis, PMN-MDSC were MACS-isolated using the granulocytic surface marker CD66b in the PBMC fraction of healthy donors.[11, 22] A possible caveat of this method is that it is not completely selective between activated
IP T
granulocytes and immunosuppressive PMN-MDSC, since both express the surface marker CD66b and the separation occurs only via cell density / buoyance. Although the method used
SC R
is the most accepted for human PMN-MDSC isolation, it cannot be ruled out that some activated
granulocytes can also be found in the low density PBMC fraction.[4] Furthermore, one general
U
limitation of our study is that positive selection of cells with magnetic beads can lead to the
N
activation of cells and may influence their function. This has to be taken into account when
A
interpreting the results of our in vitro study. The experiments focusing on MDSC apoptosis
M
demonstrated that human CD3+ T cells act anti-apoptotic on PMN-MDSC, an effect, which was enhanced after stimulation with OKT3 and mediated both by CD4+ T cells and CD8+ T cells.
ED
The fact that certain T cell subpopulations influence the survival of PMN-MDSC has already been described in the literature. Wu et al. reported that γδ-T17 cells via anti-apoptotic effects
PT
contribute to the accumulation of PMN-MDSC in tumor tissue of patients with colorectal
CC E
carcinoma.[9] However, also the opposite effect that activated T cells induce apoptosis of MDSC has been documented in the literature.[23] Zhao et al. reported that in the mouse model tmTNF-α by signaling through TNFRII delays apoptosis of MDSC.[24] In our experiments the
A
TNF-α antagonist etanercept did not inhibit the T-cell mediated anti-apoptotic effect on PMNMDSC. Thus, TNF-α does not appear to be the relevant cytokine in this context. In contrast, our present study shows that the T-cell cytokine IL-2 dose-dependently increases the survival rate of PMN-MDSC. IL-2 is known to play a crucial role in T-cell activation and T-cell homeostasis.[25] Interestingly, polymorphonuclear granulocytes also express an IL-2 12
Bauswein et al.
receptor.[26] However, this receptor has differences in its composition compared to the high affinity receptor expressed by T cells and has only an intermediate affinity for IL-2.[27, 28] IL2 shows various effects on polymorphonuclear granulocytes.[26] Among others, it increases the antifungal activity[27], prolongs their survival by inhibiting apoptosis[29], and increases their production of TNF-α[28] and IL-8[30]. By contrast, effects of IL-2 on PMN-MDSC have
IP T
not been described in the literature before. In vivo, the effect of IL-2 on PMN-MDSC survival
might actually be augmented by activated T cells and their GM-CSF production induced by
SC R
autocrine IL-2.[31]
Nagaraj et al. hypothesized that MDSC could be part of a physiological negative
U
feedback loop that serves to regulate T cells.[10] In our present study we confirmed in vitro that
N
T cells reciprocally influence human MDSC. Induction and prolongation of survival of their
A
own suppressor cells are considered to be negative feedback programs for the activity of T cells.
M
Our results further demonstrate differences depending on the T cell subpopulation and activation state.
ED
In summary, our study demonstrates that human T cells dynamically modulate MDSC
PT
generation and survival through two distinct mechanisms and thereby fine-tune the homeostasis of human MDSC in a regulated manner. Further research is required to evaluate these human
CC E
findings in vivo, which could lead to new therapeutic options targeting MDSC in inflammation
A
and cancer.
Author contributions M.B. performed experiments, analyzed the data and co-wrote the manuscript. A.S. performed revision experiments, supervised experiments, discussed the data and revised the manuscript. A.R. and D.N. discussed the data and helped with experiments, K.F., K.D.B., and I.S. helped 13
Bauswein et al.
with experiments. A.H. and R.H. provided guidance in the study and revised the manuscript. D.H. co-designed the study, discussed the data and revised the manuscript. N.R. co-designed the study, supervised the experiments, discussed the data and co-wrote the manuscript. Conflict of interest disclosure
IP T
All authors declare that no conflict of interest exists. Funding
SC R
This work was funded by the IZKF Promotionskolleg (University of Tübingen) to M.B. and the
M
A
N
U
DFG project RI 2511/2-1 to N.R.
ED
References Literature
A
CC E
PT
[1] D.I. Gabrilovich, S. Nagaraj, Myeloid-derived suppressor cells as regulators of the immune system, Nat Rev Immunol 9(3) (2009) 162-174. [2] I. Marigo, L. Dolcetti, P. Serafini, P. Zanovello, V. Bronte, Tumor-induced tolerance and immune suppression by myeloid derived suppressor cells, Immunol Rev 222 (2008) 162-79. [3] C.A. Dumitru, K. Moses, S. Trellakis, S. Lang, S. Brandau, Neutrophils and granulocytic myeloidderived suppressor cells: immunophenotyping, cell biology and clinical relevance in human oncology, Cancer immunology, immunotherapy : CII 61(8) (2012) 1155-67. [4] V. Bronte, S. Brandau, S.H. Chen, M.P. Colombo, A.B. Frey, T.F. Greten, S. Mandruzzato, P.J. Murray, A. Ochoa, S. Ostrand-Rosenberg, P.C. Rodriguez, A. Sica, V. Umansky, R.H. Vonderheide, D.I. Gabrilovich, Recommendations for myeloid-derived suppressor cell nomenclature and characterization standards, Nat Commun 7 (2016) 12150. [5] F.J.N. Lelis, J. Jaufmann, A. Singh, K. Fromm, A.C. Teschner, S. Poschel, I. Schafer, S. Beer-Hammer, N. Rieber, D. Hartl, Myeloid-derived suppressor cells modulate B-cell responses, Immunol Lett 188 (2017) 108-115. [6] A. Florcken, A. Takvorian, A. Singh, A. Gerhardt, B.N. Ostendorf, B. Dorken, A. Pezzutto, J. Westermann, Myeloid-derived suppressor cells in human peripheral blood: Optimized quantification in healthy donors and patients with metastatic renal cell carcinoma, Immunol Lett 168(2) (2015) 2607. [7] S. Nagaraj, J.I. Youn, D.I. Gabrilovich, Reciprocal relationship between myeloid-derived suppressor cells and T cells, J Immunol 191(1) (2013) 17-23. 14
Bauswein et al.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[8] J.G. Cripps, J. Wang, A. Maria, I. Blumenthal, J.D. Gorham, Type 1 T helper cells induce the accumulation of myeloid-derived suppressor cells in the inflamed Tgfb1 knockout mouse liver, Hepatology 52(4) 1350-9. [9] P. Wu, D. Wu, C. Ni, J. Ye, W. Chen, G. Hu, Z. Wang, C. Wang, Z. Zhang, W. Xia, Z. Chen, K. Wang, T. Zhang, J. Xu, Y. Han, T. Zhang, X. Wu, J. Wang, W. Gong, S. Zheng, F. Qiu, J. Yan, J. Huang, gammadeltaT17 cells promote the accumulation and expansion of myeloid-derived suppressor cells in human colorectal cancer, Immunity 40(5) (2014) 785-800. [10] S. Nagaraj, A. Nelson, J.I. Youn, P. Cheng, D. Quiceno, D.I. Gabrilovich, Antigen-specific CD4(+) T cells regulate function of myeloid-derived suppressor cells in cancer via retrograde MHC class II signaling, Cancer Res 72(4) (2012) 928-38. [11] N. Rieber, A. Brand, A. Hector, U. Graepler-Mainka, M. Ost, I. Schafer, I. Wecker, D. Neri, A. Wirth, L. Mays, S. Zundel, J. Fuchs, R. Handgretinger, M. Stern, M. Hogardt, G. Doring, J. Riethmuller, M. Kormann, D. Hartl, Flagellin induces myeloid-derived suppressor cells: implications for Pseudomonas aeruginosa infection in cystic fibrosis lung disease, J Immunol 190(3) (2013) 1276-84. [12] M.G. Lechner, D.J. Liebertz, A.L. Epstein, Characterization of cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells, J Immunol 185(4) (2010) 2273-84. [13] X. Hu, B. Li, X. Li, X. Zhao, L. Wan, G. Lin, M. Yu, J. Wang, X. Jiang, W. Feng, Z. Qin, B. Yin, Z. Li, Transmembrane TNF-alpha promotes suppressive activities of myeloid-derived suppressor cells via TNFR2, J Immunol 192(3) (2014) 1320-31. [14] D.I. Gabrilovich, Myeloid-Derived Suppressor Cells, Cancer Immunol Res 5(1) (2017) 3-8. [15] P.L. Raber, P. Thevenot, R. Sierra, D. Wyczechowska, D. Halle, M.E. Ramirez, A.C. Ochoa, M. Fletcher, C. Velasco, A. Wilk, K. Reiss, P.C. Rodriguez, Subpopulations of myeloid-derived suppressor cells impair T cell responses through independent nitric oxide-related pathways, Int J Cancer 134(12) (2014) 2853-64. [16] L. Ziegler-Heitbrock, P. Ancuta, S. Crowe, M. Dalod, V. Grau, D.N. Hart, P.J. Leenen, Y.J. Liu, G. MacPherson, G.J. Randolph, J. Scherberich, J. Schmitz, K. Shortman, S. Sozzani, H. Strobl, M. Zembala, J.M. Austyn, M.B. Lutz, Nomenclature of monocytes and dendritic cells in blood, Blood 116(16) (2010) e74-80. [17] J.I. Youn, M. Collazo, I.N. Shalova, S.K. Biswas, D.I. Gabrilovich, Characterization of the nature of granulocytic myeloid-derived suppressor cells in tumor-bearing mice, J Leukoc Biol 91(1) (2012) 16781. [18] J.G. Cripps, J. Wang, A. Maria, I. Blumenthal, J.D. Gorham, Type 1 T helper cells induce the accumulation of myeloid-derived suppressor cells in the inflamed Tgfb1 knockout mouse liver, Hepatology 52(4) (2010) 1350-9. [19] T. Hehlgans, K. Pfeffer, The intriguing biology of the tumour necrosis factor/tumour necrosis factor receptor superfamily: players, rules and the games, Immunology 115(1) (2005) 1-20. [20] G.D. Kalliolias, L.B. Ivashkiv, TNF biology, pathogenic mechanisms and emerging therapeutic strategies, Nat Rev Rheumatol 12(1) (2016) 49-62. [21] M. Sade-Feldman, J. Kanterman, E. Ish-Shalom, M. Elnekave, E. Horwitz, M. Baniyash, Tumor necrosis factor-alpha blocks differentiation and enhances suppressive activity of immature myeloid cells during chronic inflammation, Immunity 38(3) (2013) 541-54. [22] S. Brandau, S. Trellakis, K. Bruderek, D. Schmaltz, G. Steller, M. Elian, H. Suttmann, M. Schenck, J. Welling, P. Zabel, S. Lang, Myeloid-derived suppressor cells in the peripheral blood of cancer patients contain a subset of immature neutrophils with impaired migratory properties, J Leukoc Biol 89(2) (2011) 311-7. [23] P. Sinha, O. Chornoguz, V.K. Clements, K.A. Artemenko, R.A. Zubarev, S. Ostrand-Rosenberg, Myeloid-derived suppressor cells express the death receptor Fas and apoptose in response to T cellexpressed FasL, Blood 117(20) (2011) 5381-90. [24] X. Zhao, L. Rong, X. Zhao, X. Li, X. Liu, J. Deng, H. Wu, X. Xu, U. Erben, P. Wu, U. Syrbe, J. Sieper, Z. Qin, TNF signaling drives myeloid-derived suppressor cell accumulation, J Clin Invest 122(11) (2012) 4094-104.
15
Bauswein et al.
SC R
IP T
[25] S.L. Gaffen, K.D. Liu, Overview of interleukin-2 function, production and clinical applications, Cytokine 28(3) (2004) 109-23. [26] D. Girard, J. Gosselin, D. Heitz, R. Paquin, A.D. Beaulieu, Effects of interleukin-2 on gene expression in human neutrophils, Blood 86(3) (1995) 1170-6. [27] J.Y. Djeu, J.H. Liu, S. Wei, H. Rui, C.A. Pearson, W.J. Leonard, D.K. Blanchard, Function associated with IL-2 receptor-beta on human neutrophils. Mechanism of activation of antifungal activity against Candida albicans by IL-2, J Immunol 150(3) (1993) 960-70. [28] S. Wei, D.K. Blanchard, J.H. Liu, W.J. Leonard, J.Y. Djeu, Activation of tumor necrosis factor-alpha production from human neutrophils by IL-2 via IL-2-R beta, J Immunol 150(5) (1993) 1979-87. [29] F. Pericle, J.H. Liu, J.I. Diaz, D.K. Blanchard, S. Wei, G. Forni, J.Y. Djeu, Interleukin-2 prevention of apoptosis in human neutrophils, Eur J Immunol 24(2) (1994) 440-4. [30] S. Wei, J.H. Liu, D.K. Blanchard, J.Y. Djeu, Induction of IL-8 gene expression in human polymorphonuclear neutrophils by recombinant IL-2, J Immunol 152(7) (1994) 3630-6. [31] F.J. Hartmann, M. Khademi, J. Aram, S. Ammann, I. Kockum, C. Constantinescu, B. Gran, F. Piehl, T. Olsson, L. Codarri, B. Becher, Multiple sclerosis-associated IL2RA polymorphism controls GM-CSF production in human TH cells, Nat Commun 5 (2014) 5056.
N
U
Figure legends
A
Figure 1. Differently in-vitro generated MDSC suppress the proliferation of allogenic
M
polyclonal T cells.
MDSC were generated from CD33+ PBMC of buffy coats by culturing for 6 days with GM-
ED
CSF or in presence of autologous T cells (cell ratio PBMC / T cells 1:5). MDSC were isolated from cultures by MACS for CD33 and were added to OKT3 + IL-2 stimulated allogenic
PT
CFSE-labelled PBMC (cell ratio MDSC / stimulated PBMC 1:2 / 1:6). The CFSE signal of
CC E
CD4+ and CD8+ T cells was analyzed. (A) Overview of different proliferation assays (B) Gating strategy of CD4+ and CD8+ T cells (C) Percentages of proliferating CD4+ and CD8+ T cells + / - MDSC. Depicted are means + SEM. n = 4 (MDSC-GM) – 7 (no MDSC);; * p ≤
A
0.05.
Figure 2. T cells induce PMN-MDSC. MDSC were generated in-vitro from CD33+ PBMC in absence or presence of unstimulated autologous T cells as described above. The phenotype of differently generated MDSC was analyzed by flow cytometry. CD14‾ MDSC were regarded as PMN-MDSC. (A) Gating 16
Bauswein et al.
strategy. (B) Representative surface marker profile of CD33+ myeloid cells directly after MACS isolation (day 0), after 6 day culture in medium only, and after co-culture with CD3+ T cells (C) CD3+ T cells induce PMN-MDSC. (D) CD4+ T cells, but not CD8+ T cells induce PMN-MDSC. (E) T cells in transwell systems (TW) do not induce PMN-MDSC. Depicted are
Figure 3. Induction of PMN-MDSC requires transmembrane TNF.
IP T
means + SEM n = 3 (co-culture CD3+ + GM-CSF) – 18 (medium). * p ≤ 0.05.
SC R
(A) The TNF-antagonists etanercept, adalimumab and infliximab inhibit the induction of
PMN-MDSC by CD3+ T cells / CD4+ PBMC. n = 1 (CD3+ + etanercept [10 µg/ml]) – 18 (medium). (B) Etanercept is not toxic on isolated PMN-MDSC in co-culture with T cells. n =
U
5 (co-culture CD3+ + etanercept) – 21 (control). (C) Soluble TNF-α does not induce PMN-
M
A
Figure 4. T cells delay apoptosis of PMN-MDSC.
N
MDSC. Depicted are means +/- SEM. n = 1 (TNF-α [10 µg/ml]) – 18 (medium); * p ≤ 0.05.
PMN-MDSC were isolated using MACS for CD66b and were co-cultured with autologous T-
ED
cell subsets (cell ratio PMN-MDSC / T cells 1:5) for 1 day where indicated. PMN-MDSC cultured in medium only served as a control. After cultivation, cells were subsequently stained
PT
with anti-CD66b antibodies, Annexin V and PI. (A) Flow cytometry gating strategy. (B-E)
CC E
Frequencies of live PMN-MDSC under different cultivation conditions: (B) Co-culture with CD3+ T cells (n = 9 (co-culture CD3+ + OKT3) – 24 (control)), (C) co-culture with CD4+ / CD8+ T cells (n = 3 (co-cultures CD4/8 ± OKT3) – 24 (control)), (D) co-culture with CD3+ T
A
cells in transwell system (TW) and additiont of etanercept (n = 5 (co-culture CD3+ + etanercept) – 24 (control)), (E) dose-dependent effect of IL-2 (n = 1 (IL-2 [50 U/ml]) – 24 (control)). (F) Frequencies of live, early and late apoptotic PMN-MDSC under different cultivation conditions. Depicted are means +/- SEM. n = 1-24; * p ≤ 0.05.
17
A ED
PT
CC E
IP T
SC R
U
N
A
M
Bauswein et al.
18
A ED
PT
CC E
IP T
SC R
U
N
A
M
Bauswein et al.
19
S0165-2478(18)30136-6 https://doi.org/10.1016/j.imlet.2018.07.010 IMLET 6229
To appear in:
Immunology Letters
Received date: Revised date: Accepted date:
10-3-2018 23-7-2018 31-7-2018
Please cite this article as: Bauswein M, Singh A, Ralhan A, Neri D, Fuchs K, Blanz KD, Sch¨afer I, Hector A, Handgretinger R, Hartl D, Rieber N, Human T cells modulate myeloid-derived suppressor cells through a TNF-␣-mediated mechanism, Immunology Letters (2018), https://doi.org/10.1016/j.imlet.2018.07.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Bauswein et al.
Human T cells modulate myeloid-derived suppressor cells through a TNF-α-mediated mechanism
Markus Bauswein1*, Anurag Singh1*, Anjali Ralhan1, Davide Neri1, Katharina Fuchs1, Kelly
Nikolaus Rieber1,3
IP T
Daryll Blanz1, Iris Schäfer1, Andreas Hector1, Rupert Handgretinger1, Dominik Hartl1,2 and
Department of Pediatrics I, University of Tuebingen, Tuebingen, Germany
2
Roche Pharma Research & Early Development (pRED), Immunology, Inflammation and
SC R
1
U
Infectious Diseases (I3) Discovery and Translational Area, Roche Innovation Center Basel,
Department of Pediatrics, Kinderklinik Muenchen Schwabing, Klinikum Schwabing, StKM
A
3
N
Switzerland
M
GmbH und Klinikum rechts der Isar, Technical University of Munich, Munich, Germany
Correspondence:
CC E
PT
[email protected]
ED
*contributed equally
Highlights:
A
Unstimulated CD4+, but not CD8+ T cells, induce polymorphonuclear myeloid-derived suppressor cells (PMN-MDSC). The T-cell mediated induction of PMN-MDSC is dependent on direct cell-cell contact and requires transmembrane TNF-α signaling. Stimulated human CD3+ T cells delay the apoptosis of PMN-MDSC Human T cells modulate MDSC generation and survival
Abstract 1
Bauswein et al.
Manuscript body text word count: 3253
Abstract
IP T
Myeloid-derived suppressor cells (MDSC) represent an innate immune cell subset capable of suppressing T-cell responses in cancer and chronic inflammation. While the effect of MDSC
SC R
on T cells has been defined thoroughly, the reciprocal impact of T cells on MDSC homeostasis
remains poorly understood. Therefore, we comprehensively analyzed the effect of different T-
U
cell subsets on the generation and survival of human MDSC. Using an in vitro MDSC
N
generation assay, we demonstrate that unstimulated CD4+, but not CD8+ T cells, induce
A
polymorphonuclear MDSC (PMN-MDSC) from CD33+ myeloid cells. This effect was
M
dependent on direct cell-cell contact and required TNF-α signaling. Soluble TNF-α was dispensable for PMN-MDSC generation, suggesting that transmembrane TNF-α is involved in
ED
that trans-cellular process. Stimulated human CD3+ T cells delayed the apoptosis of PMNMDSC, which was independent of TNF-α signaling or direct cell-cell contact, but was
PT
recapitulated by IL-2. Taken together, our study shows that human T cells modulate MDSC
CC E
generation and survival through two distinct mechanisms and thereby fine-tune the homeostasis
A
of human MDSC in a regulated manner.
1. Introduction Myeloid-derived suppressor cells (MDSC) represent a novel innate immune cell subset that develop
under
disease
conditions
like
tumor,
infective
and
proinflammatory 2
Bauswein et al.
microenvironments. [1, 2] These myeloid cells are characterized by their capacity to potently suppress T-cell responses.[1] MDSC include two major subsets based on their phenotypical and morphological features: polymorphonuclear (PMN-) and monocytic (M-) MDSC. These subsets show unique, yet partially overlapping functional and biochemical characteristics.[1, 3-
CD66b+CD33+CD14-CD15+ and M-MDSC as CD33+CD14+HLA-DRlow.[4]
IP T
6] Phenotypically, human PMN-MDSC have most consistently been determined as
While the effect of MDSC on their functional target cells has been defined thoroughly
SC R
in mice and human cell systems, conversely, the reciprocal influence of immune cells on the
homeostasis and function of MDSC has not been extensively studied. Particularly, how T cells,
U
as the main functional targets of MDSC, reciprocally affect MDSC remains still incompletely
N
understood.[7] Several previous studies provided evidence on a dynamic and reciprocal
A
interaction between MDSC and T cells i) An influence of Th1 cells on the accumulation of
M
preferentially PMN-MDSC in an IFN-γ dependent manner was demonstrated in a mouse model of autoimmune hepatitis.[8] ii) Another study demonstrated that γδ Th17 cells drive the
ED
accumulation of PMN-MDSC in human colorectal cancer.[9] iii) In addition, Nagaraj et al. described in mice that antigen-specific CD4+ T cells promote the transition from antigen-
PT
specific MDSC, i.e. MDSC that present cognate antigens via MHC class I or II and suppress
CC E
antigen-specific T cells in a cell-contact dependent manner, to antigen-unspecific MDSC by crosslinking MHC class II on MDSC.[10] As a consequence, they proposed that MDSC are
A
part of a negative feedback loop which physiologically regulates T cell responses.[10] Based on these previous findings, we aimed to further dissect the influence of T cells on
human MDSC. Here we demonstrate for the first time that human CD4+ T cells in an in vitro MDSC generation system induce PMN-MDSC in a TNF-dependent manner. In addition, activated human T cells increased the survival of co-cultured PMN-MDSC. These mechanisms
3
Bauswein et al.
of T cell – MDSC interaction may serve as a future therapeutic target in malignant and nonmalignant disease conditions.
IP T
2. Materials and methods 2.1. Human blood samples
SC R
The study was conducted at the University Children’s Hospital Tuebingen (Germany) and all
study methods were approved by the local ethics committee. Buffy coats of healthy donors were
N
healthy volunteers (lab staff) after informed consent.
U
obtained from the blood donor center Tuebingen. Heparinized whole blood was taken from
M
A
2.2. Isolation of cells
PBMC were prepared from buffy coats / heparinized whole blood (see 2.1.) by Ficoll density
ED
gradient sedimentation (Biocoll Separating Solution; Biochrom) and washed twice in RPMI 1640 medium (Biochrom). Myeloid cells were isolated from the PBMC fraction by labelling
PT
all cells with anti-CD33 magnetic beads (Miltenyi Biotec) followed by two sequential magnetic bead separation steps according to the manufacturer´s protocol. For the isolation of different T-
CC E
cell subsets, all cells from the PBMC fraction were labelled either with anti-CD3 magnetic beads, or with anti-CD4 magnetic beads, or with anti-CD8 magnetic beads (all Miltenyi Biotec)
A
followed by two sequential magnetic bead separation steps according to the manufacturer´s protocol. For cell survival experiments, PMN-MDSC were defined as CD66b+ cells in the Ficoll low density fraction of freshly isolated heparinized whole blood from healthy volunteers. PMNMDSC were isolated by staining all cells in the Ficoll low density fraction (“PBMC fraction”) with FITC labeled anti-CD66b antibodies (BD), followed by two sequential anti-FITC magnetic bead separation steps (Miltenyi Biotec). 4
Bauswein et al.
2.3. In vitro generation and characterization of human MDSC Human MDSC were generated from myeloid PBMC, similarly to a recently published protocol.[11, 12] Isolated human CD33+ (myeloid) PBMC were cultured at a cell ratio of 500,000 cells/ml in RPMI 1640 medium supplemented with 10% FCS, 2mM L-glutamine, 100 IU/ml penicillin and 100 mg/ml streptomycin (all Biochrom) for 6d. 48-, 24- and 12-well flat
IP T
bottom plates (Corning) as well as 25cm2 culture flasks (BD) were used. For direct co-cultures, T cells at a concentration of 500,000 cells/ml were added to the myeloid PBMC in a cell ratio
SC R
of 5:1 (1,000,000 T cells; 200,000 myeloid PBMC). For transwell assays (ThinCerts for 24well plate; pore size 0.4 µm; greiner bio-one) T cells at a concentration of 5,000,000 cells/ml
U
within the inserts were cultured together with myeloid PBMC in a cell ratio of 5:1. The
N
following supplements were added as indicated: GM-CSF (10 ng/ml; genzyme Bayer
A
HealthCare), TNF-α (concentrations from 0.5 to 10 µg/ml; biomol), etanercept (10 µg/ml / 20
M
µg/ml; Wyeth Pharma), adalimumab (32 µg/ml; AbbVie Ltd) and infliximab (120 µg/ml; Janssen Biologics). Cells were incubated in a humidified atmosphere at 37˚C and 5% CO2.
ED
Medium and supplements were refreshed on day 4. On day 6 cells were harvested using Detachin (Genlantis) and stained with the following antibodies according to the manufacturer´s
PT
protocol: anti-CD14-FITC (BD), anti-CD33-PE (Miltenyi Biotec), anti-HLA-DR-PerCP (BD),
CC E
or anti-CD33-APC (BD). Rabbit serum (invitrogen) was added to block unspecific antibody binding.
A
2.4. T cell suppression assay MDSC were generated from myeloid cells of the PBMC fraction as described above and isolated from cell cultures by magnetic bead cell sorting for CD33. If indicated we depleted CD3+ cells first (anti-CD3 magnetic beads) before sorting for CD33. Responder-PBMC were obtained either from buffy coats or from healthy volunteers´ heparinized full blood and stained with CFSE (life technologies) according to the manufacturer´s protocol. CFSE-labelled PBMC 5
Bauswein et al.
were stimulated with 100 U/ml IL-2 (R&D Systems) and 1 µg/ml OKT3 (Janssen-Cilag). Both MDSC and CFSE-labelled PBMC were added to RPMI 1640 medium supplemented with 10% human serum, 2mM L-glutamine, 100 IU/ml penicillin and 100 mg/ml streptomycin. In a 96well round bottom plate (greiner bio-one), either 10,000 / 30,000 MDSC or, as a control supplemented medium only, were added to 60,000 PBMC per well. Cells were incubated in a
IP T
humidified atmosphere at 37˚C and 5% CO2. On day 5 cells were harvested and stained with
anti-CD8a-APC, anti-CD4-PE antibodies (BioLegend) and propidium iodide (BD). PI positive
SC R
cells were excluded in flow cytometry. CFSE signals of CD4+ and CD8+ PBMC were analyzed. 2.5. Annexin V assay
U
Isolated CD66b+ PBMC (PMN-MDSC) were cultured in 48- and 24-well flat bottom plates
N
(Corning) at a cell ratio of 500,000 cells/ml in RPMI 1640 medium (Biochrom) supplemented
A
with 10% human serum (several healthy donors), 2mM L-glutamine, 100 IU/ml penicillin and
M
100 mg/ml streptomycin. For direct co-cultures, T cells at a concentration of 500,000 cells/ml were added to the PMN-MDSC in a cell ratio of 5:1. For transwell assays (ThinCerts for 24-
ED
well plate; pore size 0.4 µm; greiner bio-one) T cells at a concentration of 5,000,000 cells/ml
PT
within the inserts were cultured together with PMN-MDSC in a cell ratio of 5:1. The following supplements were added as indicated: IL-2 (R&D Systems), OKT3 (Janssen-Cilag), etanercept
CC E
(20 µg/ml; Wyeth Pharma). After incubation in a humidified atmosphere at 37˚C and 5% CO2 for one day, cells were collected and stained with FITC anti-CD66b antibodies (BD), Annexin
A
V-APC (BD) and propidium iodide (BD). 2.6. Measurement and analysis Flow cytometry was performed on a FACS Calibur (BD), data was analyzed with CellQuest analysis software (BD) and FlowJo 10 (FlowJo). 2.7. Statistics 6
Bauswein et al.
Graphics and statistical analysis were performed in Graph Pad Prism 6. For CFSE proliferation assays a Kruskal-Wallis test was performed as group analysis, followed by multiple one-tailed Mann-Whitney tests for groups of interest based on an a priori hypothesis. For MDSC surface characterization and survival assays Kruskal-Wallis tests were performed as group analyses, followed by Dunn´s multiple comparisons tests as post-tests. For comparisons of only two
IP T
groups, Mann-Whitney tests were performed. A significant difference was assumed for p ≤ 0.05
SC R
and is indicated by an asterisk.
N
U
3. Results
A
3.1. T cells induce PMN-MDSC
M
To define the impact of T cells for MDSC generation, we built on an established in vitro MDSC generation system[11, 12] to induce MDSC from myeloid cells out of the human PBMC
ED
fraction. For this purpose, PBMCs were obtained from buffy coats by Ficoll density gradient
PT
sedimentation. Subsequently, myeloid cells (CD33+) and CD3+ T cells from the PBMC fraction were isolated using magnetic bead cell sorting. The isolated myeloid cells from the PBMC
CC E
fraction were cultured for 6 days either in medium only or under addition of GM-CSF, a cytokine well-known for its potential to generate MDSC, or in co-culture with CD3+ T cells. In order to assess the suppressive potential of the myeloid cells in culture, we performed T cell
A
suppression assays. Myeloid cells were isolated from the cell cultures after 6 days, again using magnetic bead isolation for CD33. The effect of the isolated myeloid cells on the proliferation of polyclonal stimulated allogeneic T cells was analyzed by flow cytometry. Myeloid cells that had been cultured for 6 days efficiently suppressed polyclonal T cell proliferation (Fig. 1 and S1). 7
Bauswein et al.
After co-culture with T cells the surface marker profile of the generated suppressive myeloid cells was CD33+CD11b+CD16+CD14-, consistent with polymorphonuclear MDSC (PMN-MDSC) (Fig. 2B). When compared to 6 day culture in medium only we recognized a striking downregulation of CD14 when MDSC were generated in the presence of T cells (Fig. 2B). Cytospins showed a substantial number of cells with segmented nuclei (Supplemental
IP T
Figure S2). CD14- MDSC have been classified as PMN-MDSC in contrast to CD14+ monocytic MDSC (M-MDSC).[4] When MDSC were generated in the presence of T cells 91% of myeloid
SC R
cells exhibited the CD14- PMN-MDSC phenotype, whereas only 11% or 39% did, when
myeloid cells were cultured in medium only and under addition of GM-CSF, respectively (Fig.
N
3.2. CD4+, but not CD8+ T cells induce PMN-MDSC
U
2C).
A
To dissect which human T-cell subsets were responsible for the PMN-MDSC generation in
M
our experimental system, we compared CD4+ and CD8+ T cells side-by-side in the respective MDSC generation assay. For the isolation of different T-cell subsets, PBMC were labelled
ED
either with anti-CD4 or with anti-CD8 magnetic beads, followed by two sequential magnetic
PT
bead separation steps. For co-cultures, CD4+ or CD8+ T cells at a concentration of 500,000 cells/ml were added to the myeloid PBMC in a cell ratio of 5:1 (1,000,000 T cells; 200,000
CC E
myeloid PBMC). These studies showed that CD4+ T cells led to a proportion of 84% PMNMDSC (mean, n=7), whereas CD8+ T cells did not significantly induce PMN-MDSC (mean
A
8% PMN-MDSC, n=6) (Fig. 2D) in these assay systems. 3.3. T-cell mediated induction of PMN-MDSC requires direct cell-cell contact and transmembrane TNF-α Next, we addressed the question if the T-cell mediated induction of PMN-MDSC requires direct cell-cell contact, which is required for a variety of MDSC-mediated T cell suppression effector responses. For this purpose we used a transwell system and found that T-cell mediated induction 8
Bauswein et al.
of PMN-MDSC required direct cell-cell contact (Fig. 2E). In search for possible cellular pathways involved in T-cell / MDSC homeostasis we speculated on a role for the transmembrane T-cell cytokine TNF-α, which has recently been reported in T cell-mediated MDSC activation in mice.[13] Adding the TNF-α inhibitor etanercept [20 µg/ml] to co-cultures with CD4+ T cells substantially decreased the percentage of PMN-MDSC to 12% (mean, n=7)
IP T
(Fig. 3A). This effect was not due to cellular toxicity of etanercept on PMN-MDSC in coculture with T cells (Fig. 3B). Additionally, we evaluated two other TNF-α inhibitors,
SC R
infliximab [120 µg/ml] and adalimumab [32 µg/ml]. Both inhibitors confirmed our previous
findings (Fig. 3A). In accordance with the role of membrane-bound TNF-α, we did not find an effect of soluble TNF-α in broad concentration ranges [0.5 µg/ml - 10 µg/ml] on the induction
U
of PMN-MDSC (Fig. 3C). When viewed in combination, these studies demonstrated that T-cell
N
mediated induction of PMN-MDSC requires direct cell-cell contact and transmembrane TNF-
M
A
α.
3.4. CD3+ T cells delay the apoptosis of PMN-MDSC
ED
To assess whether human T cells delay the apoptosis of PMN-MDSC and thereby contribute to
PT
their net accumulation, we isolated PMN-MDSC from the PBMC fraction by magnetic cell sorting against CD66b. Subsequently, we cultivated PMN-MDSC in medium only or in co-
CC E
culture with autologous CD3+ T cells [cell ratio PMN-MDSC / T cells 1:5]. On day one after the isolation we studied the proportion of living, early apoptotic and late apoptotic / necrotic PMN-MDSC by flow cytometry after staining with Annexin V and PI (Fig. 4A). These studies
A
showed that co-culture with CD3 T cells nearly doubled the survival rate of PMN-MDSC from 12% to 22% (Fig. 4B). In a further step, we activated the CD3+ T cells with OKT3 [1 µg/ml]. This led to a further increase in survival of co-cultured PMN-MDSC, whereas OKT3 alone had no direct effect on PMN-MDSC (Fig. 4C). Comparing OKT3 stimulated CD4+ and CD8+ Tcell subpopulations there was no significant difference in increasing MDSC survival (Fig. 4C). 9
Bauswein et al.
Transwell assays further revealed that the T-cell mediated increase of PMN-MDSC survival did not require direct cell-cell contact (Fig. 4D). Furthermore, etanercept [20 µg/ml] did not block T-cell mediated MDSC survival (Fig. 4D). Therefore, these studies showed that the Tcell mediated induction and survival of PMN-MDSC employs different mechanisms. Interestingly, the T-cell cytokine IL-2 alone without direct T cell help increased the proportion
IP T
of living PMN-MDSC in a dose-dependent manner up to 26% living PMN-MDSC when added in a concentration of 10,000 U/ml to PMN-MDSC (Fig. 4E). In summary, these studies
SC R
demonstrated that human CD3+ T cells delay the apoptosis of PMN-MDSC.
U
4. Discussion
N
While effects of MDSC on T cells have been established thoroughly, the reciprocal impact of
A
T cells on MDSC is incompletely understood.[7, 14, 15] Therefore, we analyzed the effect of
M
different human T-cell subsets on the generation and survival of human MDSC. We
ED
demonstrate for the first time that human T cells induce human PMN-MDSC from myeloid cells. In addition, human T cells delayed the apoptosis of co-cultured PMN-MDSC. Taken
PT
together, our study shows two different mechanisms, how T cells reciprocally influence the
CC E
homeostasis of human MDSC.
The induction of PMN-MDSC from CD33+ PBMC by T cells has not previously been
described in the literature. We would discriminate these CD33+CD11b+CD16+CD14- cells from
A
what has been designated as nonclassical CD14lowCD16+ monocytes[16] as these cells in our assays were completely CD14 negative, showed high granularity in the forward-side scatter and showed distinct T cell suppressive capacity. Furthermore, cytospins demonstrated cells with segmented nuclei, however not as distinctive as in the publication by Youn et al..[17] Some evidence, that T cells influence the fate of PMN-MDSC has already been documented. Cripps et al. demonstrated in a mouse model of autoimmune hepatitis that Th1 cells contribute to the 10
Bauswein et al.
accumulation of PMN-MDSC.[18] Wu et al. described the γδ-T17-cell mediated accumulation of PMN-MDSC in tumor tissue in patients with colorectal carcinoma.[9] This effect was due to increased cell migration, increased proliferation and prolonged PMN-MDSC survival.[9] To better define the T cell interaction leading to the induction of PMN-MDSC, we first investigated the effect of different T cell subpopulations. These experiments revealed that only CD4+, but
IP T
not CD8+ T cells, induce PMN-MDSC. Previous studies also described the selective action of certain T cell subpopulations, such as those of Th1 cells[18] or of γδ T17 cells[9] on MDSC.
SC R
TNF-α belongs to a superfamily of cytokines and receptors that mediate pleiotropic
effects in the immune system.[19] TNF-α exerts its effects both in a transmembrane form
U
(tmTNF-α) and as a soluble cytokine after it has been proteolytically released from the cell
N
membrane.[19] Several receptors, e.g. TNF receptor I (TNFRI) and TNF receptor II (TNFRII),
A
mediate the pleiotropic effects of the TNF superfamily cytokines.[19] In our studies soluble
M
TNF-α showed no significant effect on the induction of PMN-MDSC in a broad concentration range. The induction of PMN-MDSC by unstimulated CD4+ T cells, however, was dependent
ED
on direct cell-cell contacts and could be blocked by three different TNF-α inhibitors (etanercept, adalimumab and infliximab), indicating a role for tmTNF-α in this process. Effects of tmTNF-
PT
α are preferably mediated via TNFRII.[20] We, therefore, postulate that tmTNF-α on CD4+ T
CC E
cells, but not soluble TNF-α, mediates its action via TNFRII on myeloid cells. Further experiments, for example by selective blockade or selective knock-down (e.g. by microRNA) of TNF-α in T cells or TNFRII in myeloid cells, could contribute to a better understanding of
A
this process. A role for TNF-α in the T-cell mediated induction of PMN-MDSC has not been previously described. More generally, however, an impact of TNF-α on MDSC has repeatedly been demonstrated. Recently, Sade-Feldman et al. reported in a mouse model of chronic inflammation that TNF-α induces accumulation of MDSC in spleen and bone marrow by blocking the differentiation of immature myeloid cells into macrophages and dendritic cells. 11
Bauswein et al.
Furthermore, TNF-α increased the suppressive activity of MDSC.[21] Our study now broadens the spectrum of TNF-α effects on MDSC. For the experiments on PMN-MDSC apoptosis, PMN-MDSC were MACS-isolated using the granulocytic surface marker CD66b in the PBMC fraction of healthy donors.[11, 22] A possible caveat of this method is that it is not completely selective between activated
IP T
granulocytes and immunosuppressive PMN-MDSC, since both express the surface marker CD66b and the separation occurs only via cell density / buoyance. Although the method used
SC R
is the most accepted for human PMN-MDSC isolation, it cannot be ruled out that some activated
granulocytes can also be found in the low density PBMC fraction.[4] Furthermore, one general
U
limitation of our study is that positive selection of cells with magnetic beads can lead to the
N
activation of cells and may influence their function. This has to be taken into account when
A
interpreting the results of our in vitro study. The experiments focusing on MDSC apoptosis
M
demonstrated that human CD3+ T cells act anti-apoptotic on PMN-MDSC, an effect, which was enhanced after stimulation with OKT3 and mediated both by CD4+ T cells and CD8+ T cells.
ED
The fact that certain T cell subpopulations influence the survival of PMN-MDSC has already been described in the literature. Wu et al. reported that γδ-T17 cells via anti-apoptotic effects
PT
contribute to the accumulation of PMN-MDSC in tumor tissue of patients with colorectal
CC E
carcinoma.[9] However, also the opposite effect that activated T cells induce apoptosis of MDSC has been documented in the literature.[23] Zhao et al. reported that in the mouse model tmTNF-α by signaling through TNFRII delays apoptosis of MDSC.[24] In our experiments the
A
TNF-α antagonist etanercept did not inhibit the T-cell mediated anti-apoptotic effect on PMNMDSC. Thus, TNF-α does not appear to be the relevant cytokine in this context. In contrast, our present study shows that the T-cell cytokine IL-2 dose-dependently increases the survival rate of PMN-MDSC. IL-2 is known to play a crucial role in T-cell activation and T-cell homeostasis.[25] Interestingly, polymorphonuclear granulocytes also express an IL-2 12
Bauswein et al.
receptor.[26] However, this receptor has differences in its composition compared to the high affinity receptor expressed by T cells and has only an intermediate affinity for IL-2.[27, 28] IL2 shows various effects on polymorphonuclear granulocytes.[26] Among others, it increases the antifungal activity[27], prolongs their survival by inhibiting apoptosis[29], and increases their production of TNF-α[28] and IL-8[30]. By contrast, effects of IL-2 on PMN-MDSC have
IP T
not been described in the literature before. In vivo, the effect of IL-2 on PMN-MDSC survival
might actually be augmented by activated T cells and their GM-CSF production induced by
SC R
autocrine IL-2.[31]
Nagaraj et al. hypothesized that MDSC could be part of a physiological negative
U
feedback loop that serves to regulate T cells.[10] In our present study we confirmed in vitro that
N
T cells reciprocally influence human MDSC. Induction and prolongation of survival of their
A
own suppressor cells are considered to be negative feedback programs for the activity of T cells.
M
Our results further demonstrate differences depending on the T cell subpopulation and activation state.
ED
In summary, our study demonstrates that human T cells dynamically modulate MDSC
PT
generation and survival through two distinct mechanisms and thereby fine-tune the homeostasis of human MDSC in a regulated manner. Further research is required to evaluate these human
CC E
findings in vivo, which could lead to new therapeutic options targeting MDSC in inflammation
A
and cancer.
Author contributions M.B. performed experiments, analyzed the data and co-wrote the manuscript. A.S. performed revision experiments, supervised experiments, discussed the data and revised the manuscript. A.R. and D.N. discussed the data and helped with experiments, K.F., K.D.B., and I.S. helped 13
Bauswein et al.
with experiments. A.H. and R.H. provided guidance in the study and revised the manuscript. D.H. co-designed the study, discussed the data and revised the manuscript. N.R. co-designed the study, supervised the experiments, discussed the data and co-wrote the manuscript. Conflict of interest disclosure
IP T
All authors declare that no conflict of interest exists. Funding
SC R
This work was funded by the IZKF Promotionskolleg (University of Tübingen) to M.B. and the
M
A
N
U
DFG project RI 2511/2-1 to N.R.
ED
References Literature
A
CC E
PT
[1] D.I. Gabrilovich, S. Nagaraj, Myeloid-derived suppressor cells as regulators of the immune system, Nat Rev Immunol 9(3) (2009) 162-174. [2] I. Marigo, L. Dolcetti, P. Serafini, P. Zanovello, V. Bronte, Tumor-induced tolerance and immune suppression by myeloid derived suppressor cells, Immunol Rev 222 (2008) 162-79. [3] C.A. Dumitru, K. Moses, S. Trellakis, S. Lang, S. Brandau, Neutrophils and granulocytic myeloidderived suppressor cells: immunophenotyping, cell biology and clinical relevance in human oncology, Cancer immunology, immunotherapy : CII 61(8) (2012) 1155-67. [4] V. Bronte, S. Brandau, S.H. Chen, M.P. Colombo, A.B. Frey, T.F. Greten, S. Mandruzzato, P.J. Murray, A. Ochoa, S. Ostrand-Rosenberg, P.C. Rodriguez, A. Sica, V. Umansky, R.H. Vonderheide, D.I. Gabrilovich, Recommendations for myeloid-derived suppressor cell nomenclature and characterization standards, Nat Commun 7 (2016) 12150. [5] F.J.N. Lelis, J. Jaufmann, A. Singh, K. Fromm, A.C. Teschner, S. Poschel, I. Schafer, S. Beer-Hammer, N. Rieber, D. Hartl, Myeloid-derived suppressor cells modulate B-cell responses, Immunol Lett 188 (2017) 108-115. [6] A. Florcken, A. Takvorian, A. Singh, A. Gerhardt, B.N. Ostendorf, B. Dorken, A. Pezzutto, J. Westermann, Myeloid-derived suppressor cells in human peripheral blood: Optimized quantification in healthy donors and patients with metastatic renal cell carcinoma, Immunol Lett 168(2) (2015) 2607. [7] S. Nagaraj, J.I. Youn, D.I. Gabrilovich, Reciprocal relationship between myeloid-derived suppressor cells and T cells, J Immunol 191(1) (2013) 17-23. 14
Bauswein et al.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[8] J.G. Cripps, J. Wang, A. Maria, I. Blumenthal, J.D. Gorham, Type 1 T helper cells induce the accumulation of myeloid-derived suppressor cells in the inflamed Tgfb1 knockout mouse liver, Hepatology 52(4) 1350-9. [9] P. Wu, D. Wu, C. Ni, J. Ye, W. Chen, G. Hu, Z. Wang, C. Wang, Z. Zhang, W. Xia, Z. Chen, K. Wang, T. Zhang, J. Xu, Y. Han, T. Zhang, X. Wu, J. Wang, W. Gong, S. Zheng, F. Qiu, J. Yan, J. Huang, gammadeltaT17 cells promote the accumulation and expansion of myeloid-derived suppressor cells in human colorectal cancer, Immunity 40(5) (2014) 785-800. [10] S. Nagaraj, A. Nelson, J.I. Youn, P. Cheng, D. Quiceno, D.I. Gabrilovich, Antigen-specific CD4(+) T cells regulate function of myeloid-derived suppressor cells in cancer via retrograde MHC class II signaling, Cancer Res 72(4) (2012) 928-38. [11] N. Rieber, A. Brand, A. Hector, U. Graepler-Mainka, M. Ost, I. Schafer, I. Wecker, D. Neri, A. Wirth, L. Mays, S. Zundel, J. Fuchs, R. Handgretinger, M. Stern, M. Hogardt, G. Doring, J. Riethmuller, M. Kormann, D. Hartl, Flagellin induces myeloid-derived suppressor cells: implications for Pseudomonas aeruginosa infection in cystic fibrosis lung disease, J Immunol 190(3) (2013) 1276-84. [12] M.G. Lechner, D.J. Liebertz, A.L. Epstein, Characterization of cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells, J Immunol 185(4) (2010) 2273-84. [13] X. Hu, B. Li, X. Li, X. Zhao, L. Wan, G. Lin, M. Yu, J. Wang, X. Jiang, W. Feng, Z. Qin, B. Yin, Z. Li, Transmembrane TNF-alpha promotes suppressive activities of myeloid-derived suppressor cells via TNFR2, J Immunol 192(3) (2014) 1320-31. [14] D.I. Gabrilovich, Myeloid-Derived Suppressor Cells, Cancer Immunol Res 5(1) (2017) 3-8. [15] P.L. Raber, P. Thevenot, R. Sierra, D. Wyczechowska, D. Halle, M.E. Ramirez, A.C. Ochoa, M. Fletcher, C. Velasco, A. Wilk, K. Reiss, P.C. Rodriguez, Subpopulations of myeloid-derived suppressor cells impair T cell responses through independent nitric oxide-related pathways, Int J Cancer 134(12) (2014) 2853-64. [16] L. Ziegler-Heitbrock, P. Ancuta, S. Crowe, M. Dalod, V. Grau, D.N. Hart, P.J. Leenen, Y.J. Liu, G. MacPherson, G.J. Randolph, J. Scherberich, J. Schmitz, K. Shortman, S. Sozzani, H. Strobl, M. Zembala, J.M. Austyn, M.B. Lutz, Nomenclature of monocytes and dendritic cells in blood, Blood 116(16) (2010) e74-80. [17] J.I. Youn, M. Collazo, I.N. Shalova, S.K. Biswas, D.I. Gabrilovich, Characterization of the nature of granulocytic myeloid-derived suppressor cells in tumor-bearing mice, J Leukoc Biol 91(1) (2012) 16781. [18] J.G. Cripps, J. Wang, A. Maria, I. Blumenthal, J.D. Gorham, Type 1 T helper cells induce the accumulation of myeloid-derived suppressor cells in the inflamed Tgfb1 knockout mouse liver, Hepatology 52(4) (2010) 1350-9. [19] T. Hehlgans, K. Pfeffer, The intriguing biology of the tumour necrosis factor/tumour necrosis factor receptor superfamily: players, rules and the games, Immunology 115(1) (2005) 1-20. [20] G.D. Kalliolias, L.B. Ivashkiv, TNF biology, pathogenic mechanisms and emerging therapeutic strategies, Nat Rev Rheumatol 12(1) (2016) 49-62. [21] M. Sade-Feldman, J. Kanterman, E. Ish-Shalom, M. Elnekave, E. Horwitz, M. Baniyash, Tumor necrosis factor-alpha blocks differentiation and enhances suppressive activity of immature myeloid cells during chronic inflammation, Immunity 38(3) (2013) 541-54. [22] S. Brandau, S. Trellakis, K. Bruderek, D. Schmaltz, G. Steller, M. Elian, H. Suttmann, M. Schenck, J. Welling, P. Zabel, S. Lang, Myeloid-derived suppressor cells in the peripheral blood of cancer patients contain a subset of immature neutrophils with impaired migratory properties, J Leukoc Biol 89(2) (2011) 311-7. [23] P. Sinha, O. Chornoguz, V.K. Clements, K.A. Artemenko, R.A. Zubarev, S. Ostrand-Rosenberg, Myeloid-derived suppressor cells express the death receptor Fas and apoptose in response to T cellexpressed FasL, Blood 117(20) (2011) 5381-90. [24] X. Zhao, L. Rong, X. Zhao, X. Li, X. Liu, J. Deng, H. Wu, X. Xu, U. Erben, P. Wu, U. Syrbe, J. Sieper, Z. Qin, TNF signaling drives myeloid-derived suppressor cell accumulation, J Clin Invest 122(11) (2012) 4094-104.
15
Bauswein et al.
SC R
IP T
[25] S.L. Gaffen, K.D. Liu, Overview of interleukin-2 function, production and clinical applications, Cytokine 28(3) (2004) 109-23. [26] D. Girard, J. Gosselin, D. Heitz, R. Paquin, A.D. Beaulieu, Effects of interleukin-2 on gene expression in human neutrophils, Blood 86(3) (1995) 1170-6. [27] J.Y. Djeu, J.H. Liu, S. Wei, H. Rui, C.A. Pearson, W.J. Leonard, D.K. Blanchard, Function associated with IL-2 receptor-beta on human neutrophils. Mechanism of activation of antifungal activity against Candida albicans by IL-2, J Immunol 150(3) (1993) 960-70. [28] S. Wei, D.K. Blanchard, J.H. Liu, W.J. Leonard, J.Y. Djeu, Activation of tumor necrosis factor-alpha production from human neutrophils by IL-2 via IL-2-R beta, J Immunol 150(5) (1993) 1979-87. [29] F. Pericle, J.H. Liu, J.I. Diaz, D.K. Blanchard, S. Wei, G. Forni, J.Y. Djeu, Interleukin-2 prevention of apoptosis in human neutrophils, Eur J Immunol 24(2) (1994) 440-4. [30] S. Wei, J.H. Liu, D.K. Blanchard, J.Y. Djeu, Induction of IL-8 gene expression in human polymorphonuclear neutrophils by recombinant IL-2, J Immunol 152(7) (1994) 3630-6. [31] F.J. Hartmann, M. Khademi, J. Aram, S. Ammann, I. Kockum, C. Constantinescu, B. Gran, F. Piehl, T. Olsson, L. Codarri, B. Becher, Multiple sclerosis-associated IL2RA polymorphism controls GM-CSF production in human TH cells, Nat Commun 5 (2014) 5056.
N
U
Figure legends
A
Figure 1. Differently in-vitro generated MDSC suppress the proliferation of allogenic
M
polyclonal T cells.
MDSC were generated from CD33+ PBMC of buffy coats by culturing for 6 days with GM-
ED
CSF or in presence of autologous T cells (cell ratio PBMC / T cells 1:5). MDSC were isolated from cultures by MACS for CD33 and were added to OKT3 + IL-2 stimulated allogenic
PT
CFSE-labelled PBMC (cell ratio MDSC / stimulated PBMC 1:2 / 1:6). The CFSE signal of
CC E
CD4+ and CD8+ T cells was analyzed. (A) Overview of different proliferation assays (B) Gating strategy of CD4+ and CD8+ T cells (C) Percentages of proliferating CD4+ and CD8+ T cells + / - MDSC. Depicted are means + SEM. n = 4 (MDSC-GM) – 7 (no MDSC);; * p ≤
A
0.05.
Figure 2. T cells induce PMN-MDSC. MDSC were generated in-vitro from CD33+ PBMC in absence or presence of unstimulated autologous T cells as described above. The phenotype of differently generated MDSC was analyzed by flow cytometry. CD14‾ MDSC were regarded as PMN-MDSC. (A) Gating 16
Bauswein et al.
strategy. (B) Representative surface marker profile of CD33+ myeloid cells directly after MACS isolation (day 0), after 6 day culture in medium only, and after co-culture with CD3+ T cells (C) CD3+ T cells induce PMN-MDSC. (D) CD4+ T cells, but not CD8+ T cells induce PMN-MDSC. (E) T cells in transwell systems (TW) do not induce PMN-MDSC. Depicted are
Figure 3. Induction of PMN-MDSC requires transmembrane TNF.
IP T
means + SEM n = 3 (co-culture CD3+ + GM-CSF) – 18 (medium). * p ≤ 0.05.
SC R
(A) The TNF-antagonists etanercept, adalimumab and infliximab inhibit the induction of
PMN-MDSC by CD3+ T cells / CD4+ PBMC. n = 1 (CD3+ + etanercept [10 µg/ml]) – 18 (medium). (B) Etanercept is not toxic on isolated PMN-MDSC in co-culture with T cells. n =
U
5 (co-culture CD3+ + etanercept) – 21 (control). (C) Soluble TNF-α does not induce PMN-
M
A
Figure 4. T cells delay apoptosis of PMN-MDSC.
N
MDSC. Depicted are means +/- SEM. n = 1 (TNF-α [10 µg/ml]) – 18 (medium); * p ≤ 0.05.
PMN-MDSC were isolated using MACS for CD66b and were co-cultured with autologous T-
ED
cell subsets (cell ratio PMN-MDSC / T cells 1:5) for 1 day where indicated. PMN-MDSC cultured in medium only served as a control. After cultivation, cells were subsequently stained
PT
with anti-CD66b antibodies, Annexin V and PI. (A) Flow cytometry gating strategy. (B-E)
CC E
Frequencies of live PMN-MDSC under different cultivation conditions: (B) Co-culture with CD3+ T cells (n = 9 (co-culture CD3+ + OKT3) – 24 (control)), (C) co-culture with CD4+ / CD8+ T cells (n = 3 (co-cultures CD4/8 ± OKT3) – 24 (control)), (D) co-culture with CD3+ T
A
cells in transwell system (TW) and additiont of etanercept (n = 5 (co-culture CD3+ + etanercept) – 24 (control)), (E) dose-dependent effect of IL-2 (n = 1 (IL-2 [50 U/ml]) – 24 (control)). (F) Frequencies of live, early and late apoptotic PMN-MDSC under different cultivation conditions. Depicted are means +/- SEM. n = 1-24; * p ≤ 0.05.
17
A ED
PT
CC E
IP T
SC R
U
N
A
M
Bauswein et al.
18
A ED
PT
CC E
IP T
SC R
U
N
A
M
Bauswein et al.
19