Tumor-Produced Interleukin-8 Attracts Human Myeloid-Derived Suppressor Cells and Elicits Extrusion of Neutrophil Extracellular T

Published OnlineFirst March 8, 2016; DOI: 10.1158/1078-0432.CCR-15-2463

Biology of Human Tumors Clinical Cancer Research Tumor-Produced Interleukin-8 Attracts Human Myeloid-Derived Suppressor Cells and Elicits Extrusion of Neutrophil Extracellular Traps (NETs) Carlos Alfaro1,2,3, Alvaro Teijeira1,2,3, Carmen Onate~ 1,2,3, Guiomar Perez 1,2,3, Miguel F. Sanmamed2,3, Maria Pilar Andueza2,3, Diego Alignani4, Sara Labiano1, Arantza Azpilikueta1, Alfonso Rodriguez-Paulete1, Saray Garasa1, Juan P. Fusco2,3, Angela Aznar1, Susana Inoges 2,3, Maria De Pizzol5, Marcello Allegretti5, Jose Medina-Echeverz1, Pedro Berraondo1, Jose L. Perez-Gracia2,3, and Ignacio Melero1,2,3

Abstract

Purpose: Myeloid-derived suppressor cells (MDSC) are con- microscopy, fluorimetry, and time-lapse fluorescence confocal sidered an important T-cell immunosuppressive component in microscopy on short-term MDSC cultures. cancer-bearing hosts. The factors that attract these cells to the Results: IL8 chemoattracts both granulocytic (GrMDSC) and tumor microenvironment are poorly understood. IL8 (CXCL8) is monocytic (MoMDSC) human MDSC. Monocytic but not gran- a potent chemotactic factor for neutrophils and monocytes. ulocytic MDSC exerted a suppressor activity on the proliferation Experimental Design: MDSC were characterized and sorted by of autologous T cells isolated from the circulation of cancer multicolor flow cytometry on ficoll-gradient isolated blood leu- patients. IL8 did not modify the T-cell suppressor activity of cokytes from healthy volunteers (n ¼ 10) and advanced cancer human MDSC. However, IL8 induced the formation of NETs in patients (n ¼ 28). In chemotaxis assays, sorted granulocytic and the GrMDSC subset. monocytic MDSC were tested in response to recombinant IL8, IL8 Conclusion: IL8 derived from tumors contributes to the che- derived from cancer cell lines, and patient sera. Neutrophil motactic recruitment of MDSC and to their functional control. extracellular traps (NETs) formation was assessed by confocal Clin Cancer Res; 1–13. 2016 AACR.

Introduction genic activities by means of providing growth factors (12, 13), upregulating angiogenesis (2, 4, 7, 14), and suppressing and IL8 (CXCL8) is a chemokine originally discovered for its biasing antitumor immune responses (13, 15). In mice, a subset ability to attract neutrophils to inflammatory foci (1). In cancer of myeloid cells suppressing T-cell proliferation and activation its role in promoting angiogenesis is well characterized. IL8 accumulates in the spleen and tumor microenvironment of does not only attract neutrophilsbutalsomonocytesto tumor-bearing hosts (16). This population-termed myeloid- inflamed tissues (2). Tumor cells constitutively produce IL8 derived suppressor cells (MDSC) is characterized by surface (2–9) and hence plasma concentrations of IL8 reflect tumor coexpression of CD11b and Ly6 (16). Its functions cause sup- burden (2) and determine prognosis (2–9). IL8 is functional on pression of T-cell immunity by a variety of metabolic actions (17), CXCR1 and CXCR2 where it induces chemotactic responses. which ultimately repress the proliferation of T lymphocytes. There is no IL8 analogue in the mouse genome (10, 11), but IL8 Therefore, reductions in the numbers of such myeloid cells is functional on murine receptors (11). enhance therapeutic immune responses (18, 19) and have yielded Little is known about the function of IL8 to attract leukocytes in a therapeutic benefit in several tumor models. Two subsets of cancer. However, myeloid leukocytes are involved in protumoro- MDSC have been described based on selective expression of Ly6C þ and Ly6G (16, 20). Morphologically, Ly6C MDSC resemble þ 1Division of Gene Therapy and Hepatology, Centre for Applied Medical monocytes, whereas Ly6G MDSC share morphological features Research (CIMA), Pamplona, Spain. 2Department of Oncology, Uni- with polymorphonuclear (PMN) leukocytes. versity Clinic of Navarra, Pamplona, Spain. 3Department of Immunol- In the peripheral blood of human cancer patients, several ogy, University Clinic of Navarra, Pamplona, Spain. 4Cytometry Plat- fi form, Centre for Applied Medical Research (CIMA), Pamplona, Spain. investigators have de ned myeloid subsets that now also receive 5Dompé Farmaceutici SpA, L'Aquila, Italy. the name of MDSCs (21–25) even if their suppressive activity on Note: Supplementary data for this article are available at Clinical Cancer the activation and proliferation of T lymphocytes is not well fl Research Online (http://clincancerres.aacrjournals.org/). studied. Using ow cytometry multiple staining on peripheral blood leukocytes recovered from ficoll gradients permits the J.L. Perez-Gracia and I. Melero share equal credit for senior authorship. identification of human monocytic MDSC that are defined as þ þ þ Corresponding Author: Ignacio Melero, Avda.Pio XII, 55, Pamplona 31008, CD33 CD11b HLA-DR /low CD14 , whereas granulocytic Spain. Phone: 349-4819-4700; Fax: 349-4819-4717; E-mail: [email protected] þ þ þ MDSC are defined as CD33 CD11b HLA-DR /low CD15 (25). doi: 10.1158/1078-0432.CCR-15-2463 The numbers of both populations rise in advanced cancer 2016 American Association for Cancer Research. patients and several articles have reported a correlation

www.aacrjournals.org OF1

Alfaro et al.

clinics. Patients provided informed consent under a protocol Translational Relevance approved by the institutional ethics committee. The cell suspen- IL8 is a chemokine produced by cancer cells and whose sion was laid over 15 mL of Ficoll-Hypaque (GE Healthcare serum concentration correlates with poor prognosis. In this Europe GMBH) gradients in 50 mL tubes and centrifuged at study, we demonstrate that IL8 chemotactically attracts 2,200 rpm for 20 minutes at 20C. The leukocyte layer was myeloid-derived suppressor cells (MDSC) sorted from the resuspended in 5 mL ice-cold PBS. Subsequently, cells were peripheral blood of advanced cancer patients and that this incubated with PBS/human IgG (50 mg/mL; Beriglobina P; Behr- chemotactic activity can be pharmacologically disrupted in ing) for 10 minutes on ice, to block Fc receptors and then aliquots vivo. Surprisingly, we also found that IL8 activates granulo- of cells (106) were washed in ice-cold PBS and incubated 15 min- cytic MDSC to extrude DNA-forming neutrophil extracellu- utes at 4C with specific labeled mAb: Lin [CD3 PE (Biolegend), lar traps (NET). These IL8-mediated mechanisms are likely CD16 PE (Biolegend), CD19 PE (BD Biosciences), CD20 PE to be relevant in shaping a protumoral leukocyte microen- (Biolegend), CD56 PE (BD Biosciences)], HLA-DR APC (Biole- vironment in cancer. gend), CD11b PeCy5 (IoTest), CD33 FITC (Biolegend), CD14 BV421 (Biolegend), and CD15 BV510 (Biolegend). Immediately after, FACS sorting was performed using a BD FACSAria III with the gating strategy shown in Fig. 1A. For surface expression of CXCR1 and CXCR2 in human and mouse MDSC, the following between higher levels and worse prognosis (21, 26–28). We antibodies were used: Lin PE, CD11b PeCy5 (IoTest), CD33 have previously reported that IL8 attracts dendritic cells and BV421 (Biolegend), HLA-DR APC Cy7 (Biolegend), CD14 PeCy7 discovered that tumors producing IL8 were chemotactically (Biolegend), CD15 BV510 (Biolegend), human CXCR1 FITC retaining dendritic cells injected intratumorally, thereby avoid- (Biolegend), human CXCR2 AF647 (Biolegend), mouse CXCR1 ing their outmigration to draining lymph nodes (10, 29). In FITC (Biorbyt), and mouse CXCR2 APC (R&D Systems). For this study, we explored if IL8 could attract and functionally characterization the mouse MDSC, obtained from the spleen, the modulate MDSC. To this end, human MoMDSC and GrMDSC following antibodies were used: CD45.2 Pacific Blue (Biolegend), were identified and FACS-sorted from the peripheral blood of a CD11b APC (BD Biosciences), CD124 PE (BD Biosciences), Ly6G series of advanced cancer patients. It was found that IL8 readily BV510 (Biolegend), and Ly6C PerCP Cy5.5 (Biolegend). Samples attracts all these cell subsets, although it does not regulate their were analyzed using a FACSCanto flow cytometer (BD Bios- T-cell suppressive activity. However, IL8 induces the formation ciences) and data were analyzed with the FlowJo (FlowJo, TreeStar of neutrophil extracellular traps (NET) with projected DNA as Inc.) software. seen in human and mouse PMNs (30–32). Such NETs made of For neutrophil enrichment, 10 mL of peripheral blood DNA and DNA-binding proteins are known to have important drawn in heparin-containing tubes were mixed with 10 mL of bacteriostatic and bactericidal functions (32–35) and have cold PBS and 10 mL of 6% dextran/0.9% NaCl solution. After recently been discovered to play a role in autoimmunity being inverted 18 to 20 times to ensure adequate mixing, the (36) and in organizing metastatic emboli in cancer (37). mixture was pipetted into a 50 mL tube and the tube was placed Collectively, our experiments demonstrate that tumor-pro- upright at room temperature for 1 to 1.5 hours, or until duced IL8 chemoattracts and modulates human and mouse MDSC separation was completed. The yellowish supernatant was in a fashion that can be interfered with in vivo using the CXCR1/2 recovered into a 50 mL tube and spun at 1,200 rpm for 12 blocking agent reparixin, that are under clinical development in minutes at 4C with low brake. The supernatant was then oncology (38, 39). Given the protumorogenic immunosuppres- discarded and the pellet resuspended in 5 mL ACK buffer, sive function of these myeloid subsets, the IL8–MDSC axis offers 5 minutes at room temperature. After that incubation, the cells an attractive target for cancer immunotherapy strategies. were washed with 45 mL of cold PBS. The supernatant was discarded and the pellet was resuspended in 2.5 mL of PBS. The Materials and Methods cell suspension was laid over 3 mL of Ficoll-Hypaque (GE Healthcare Europe GMBH) gradients in a 15 mL tube and spun Cell lines and tumor grafting at 1500 rpm for 30 minutes at 4C. The pellet was resuspended HT29 was obtained from ATCC in 2011 and retrieved from the in2mLcoldPBSandkeptat4C. To ensure the correct verified master cell bank and CT26GM-CSF transfectants were a purification, the pellet was immunostained with CD15-PE. kind gift of Mario P. Colombo (Milano; ref.40). For tumor Neutrophil purity was >95% (CD15bright neutrophils). engraftment, 5 105 tumor cells were injected onto the flank of BALB/c mice (Harlan) or Rag2 / IL2Rg / mice bred in our Transwell migration assays pathogen-free animal facility. Animal experiments were con- In vitro MDSC and neutrophil migration was measured in ducted in accordance to Spanish laws and under approval by the Transwell chambers (5 mm; Corning Costar). Human and mouse Ethics committee of animal experimentation at the University of MDSC (105) or human neutrophils were seeded onto the top Navarra (study number 070-15). chamber and migration stimuli were placed in the bottom chamber. In this experiment, rhIL-8 (R&D Systems) was used at dif- FACS analyses and sorting of MDSC ferent concentrations. Duplicate experiments were performed Twenty milliliters of peripheral blood from a sequential series with or without IL8-neutralizing mAb or IgG as control (BD of advanced prostate cancer and NSCLC patients or from healthy Pharmigen) at 20 mg/mL that were added to the bottom chamber. volunteers were collected in 10 mL heparin-containing tubes and Transmigrated cells recovered from the bottom chamber were were mixed with 10 mL of ice-cold PBS. Blood samples from manually quantified using a Neubauer chamber. The chemotactic sequential patients were collected upon visit to our outpatient index was calculated as the number of migrated cells in each

OF2 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

experimental condition divided by number of migrated cells in that detects up to 0.15 EU/mL (Cambrex limulus amebocyte the negative control, which was serum-free RPMI medium. lysate QCL-1000).

IL8 ELISA Time lapse microscopy of MSDCs NET formation IL8 concentrations in HT29 carcinoma culture supernatants GrMDSCs were FACS sorted and stained for 15 minutes in 6 (10 cells/mL) and serum of patients were tested using a com- RPMI media at 37C with DRAQ5 (ThermoFisher) at a final mercially purchased ELISA Kit used according to the manufac- concentration of 5 mmol/L. Cells were washed twice with HBSS turer's instructions (BD Biosciences). and a total of 5 104 cells were plated in eight-well imaging chambers (IBIDI) in a volume of 100 mL. HT29 cells were Hydrodynamic IL8 gene transfer stained with PKH26 (Sigma) according to manufacturer's The human IL8 expression plasmid (pUC19/human IL8) was instructions, and added to cultures in additional 100 mL. Imme- kindly given to us by Joost Oppenheim (Addgene). On day 14 diately prior to imaging Sytox Green (Invitrogen) was added in following subcutaneous tumor cell inoculation (2 106 HT29 the media at a final concentration of 0.1 mmol/L and 10 ng/mL cells), Rag2 / IL2Rg / mice received an intravenous injection of IL8 (R&D) were added. Imaging was performed with a Zeiss (tail vein) of 10 mg of plasmid in a volume of 100 mL/kg using a Axiovert 200 microscope (Zeiss) equipped with a Perkin Elmer 27-G needle at a rate of 0.4 mL/seconds (41). Reparixin (20 mg) Spining disk unit, a Hammamatsu C9100-50 CCD camera, and was administered subcutaneously daily between days 12 and 18. a40 Plan Apochromat Objective (N.A. 0.95), in an incubation At day 18, mice were sacrificed to obtain liver and spleen cell chamber at 37 C and 5% CO2. Twenty to twenty-four micro- suspensions. Organs were disaggregated to assess absolute num- meters of confocal stacks were acquired with a Z slice spacing of þ þ bers of MoMDSC (CD11b Ly6G Ly6C ) and GrMDSC 2 mm. Stacks were taken each 60 seconds for a total of 5 hours. þ þ (CD11b Ly6G Ly6Clow) by referring to total cells assessed by Image acquisition, Z maximum projections, video generations, manual counting on a Neubauer chamber. and gamma adjustments were generated with UltraVIEW ERS software (Perkin Elmer). Videos were further edited and com- Proliferation assay pressed using FinalCutPro software and are shown at a speed of Sorted peripheral blood GrMDSC or MoMDSC (105) were 10 fps. cocultured at different MDSC to T-cell ratios with autologous 5 T cells (10 ) derived from the same subject. T cells were stained Statistical analysis with carboxifluorescein succinimidyl ester (CFSE). Cells were Results were presented as mean SD or mean SEM and plated on flat-bottomed wells precoated overnight with anti-CD3 experimental groups were compared by Mann–Whitney U test (2.5 mg/mL) and anti-CD28 (1.2 mg/mL) mAb in PBS. In some and Student t tests as indicated in figure legends. conditions, IL8 (1,000 pg/mL) was added to the culture. After 5 days, lymphocytes were harvested and analyzed by flowcytometryusingaFACSCaliburFlowCytometer(Becton Results Dickinson). Lymphocyte division was measured by CFSE Mouse MDSC are responsive to human IL8 dilution (10). Mice engrafted with CT26 tumors expressing GM-CSF are known to accumulate high numbers of MDSC (40). IL8 is NET formation known to cross-react to the CXCR1 mouse receptor (10, 11). In To quantify NETs, 5 104 neutrophils, GrMDSC, or this regard, the spleens of mice subcutaneously engrafted with MoMDSC in 100 mL of HBSS (Invitrogen) without CaCl2 CT-26 transfected to express GM-CSF (CT26-GM) contained /MgCl2 were seeded in a black 96-well plate (Cayman Chem- abundant MDSC. This effect actually leads to evident spleno- ical). The cells were treated with 1 mg/mL PMA (Sigma-Aldrich), megalyasshowninSupplementaryFig.S1.Suchsplenic 100 ng/mL LPS (Sigma-Aldrich), IL8 (R&D Systems) at different MDSC were gated and FACS sorted as indicated in Fig. 1A. concentrations or left untreated. An equivalent cell sample was Immunostaining with specific mAb shows that both GrMDSC also lysed with 0.1% Triton X-100 (Union Carbide Chemicals) and MoMDSC express CXCR1 and CXCR2 on their plasma or the sample was frozen as a control for the total DNA in each membrane up to weak but detectable levels (Fig. 1B). In line well. Cells were incubated at 37C for 2 to 4 hours. The with this finding, both the MoMDSC and the GrMDSC from membrane-impermeable DNA-binding dye, Sytox Green (Invi- the spleen of such mice migrated in response to human trogen) was added (0.1 mmol/L) at each experimental time- recombinant IL8 in a dose-dependent fashion, even though point to bind extracellular DNA, and captured fluorescence was IL8 does not exist in the mouse genome. Such murine MDSC quantified using a fluorescence spectrophotometer (Fisher Sci- also chemotactically responded to mouse Gro-a (CXCL1), entific). In some experiments, HT29 cells labeled with PKH26 tested as a natural isospecies agonist for CXCR1 and CXCR2 were added together with neutrophils or MDSC. To digest and (Fig. 1C). destroy extracellular NET-DNA in samples, cells were incubated with 200 mg/mL of RNase-free DNase I (Roche Life Sciences) for Reparixin can block the chemoattraction of IL8 on mouse 2hoursat37 C. Photographs of representative fields were then MDSC in vivo taken under confocal microscopy. As the IL8-mediated chemo-attraction could be advanta- geous to tumors, we explored the possibility of therapeutic LPS detection test inhibition of such a function. To do so, in immunodeficient LPS concentrations in IL8 batches purchased from R&D Sys- Rag2 / IL2Rg / mice-bearing HT29 tumor xenografts to tems were below the detection limits of a sensitive commercial kit raise their MDSC numbers, we performed hydrodynamic gene

www.aacrjournals.org Clin Cancer Res; 2016 OF3

Alfaro et al.

Figure 1. Mouse MDSC express CXCR1 and CXCR2 and chemotactically respond to IL8. A, MDSC gating strategy following immunoﬂuorescence and FACS analysis. B, surface immunostaining for CXCR1 and CXCR2 on MoMDSC and GrMDSC derived from the spleen of mice bearing subcutaneous CT26-GMCSF–derived tumors for 7 days (n ¼ 10), indicating the percentage of positive cells. The mean ﬂuorescence intensity SEM is also given for each group. C, dose–response chemotaxis to IL8 and mouse Gro-a of sorted GrMDSC and MoMDSC. Results are shown as the chemotactic index. Statistical comparisons were performed by Mann–Whitney U test.

transfer of IL8 to the liver. This procedure achieves transient transfer of IL8 augmented the absolute number of MDSC in gene expression of IL8 in the liver of mice without detectable the liver of mice in the course of experiments described in expression in other organs (42). Indeed, hydrodynamic gene Fig. 2A. As can be seen in Fig. 2B and C, the numbers of

OF4 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

Figure 2. Reparixin blocks chemotaxis of MDSC to IL8. A, description of the in vivo experiment involving HT29 tumor- bearing Rag2 / IL2Rg / mice which were treated with reparixin from day 12 to 18. On day þ14, mice received a hydrodynamic gene transfer procedure with an IL8-encoding expression cassette (HD IL8) or control empty plasmid. B, C, show the liver and spleen absolute numbers of MoMDSC or GrMDSC in the indicated conditions determined by ﬂow cytometry referred to total cell counts (Neubauer chamber). Student t tests were used for statistical comparisons. D, experiment performed in BALB/c mice bearing CT26-GMCSF–derived tumors for 10 days and which were treated with control vehicle or reparixin (20 mg per dose) during the last 4 days prior to sacriﬁce. Total numbers of GrMDSC and MoMDSC retrieved from cell suspensions from eight tumors per group. Two repetitions of the experiments were performed and Mann–Whitney U tests were used for statistical comparisons.

Mo-MDSC and Gr-MDSC were clearly increased in the IL8- Reparixin is a described pharmacological inhibitor of CXCR1 gene-transferred livers but not in the spleens, indicating che- and CXCR2 (39, 43). Injection of this compound into mice moattraction to the IL8-producing organ, as a result of hydro- around the time of IL8 gene transfer almost completely abro- dynamic gene transfer. gated the attraction of MDSC to the liver (Fig. 2B and C). We

www.aacrjournals.org Clin Cancer Res; 2016 OF5

Alfaro et al.

performed similar experiments in tumor-free mice, which cancer patients have been extensively reported, the functional þ þ showed increases in Ly6G CD11b cell in the liver following effects of the human MDSC counterparts on T lymphocytes are IL8 gene transfer (Supplementary Fig. S2), albeit in this case not well explored. such cells are likely normal granulocytes because they do not In a series of metastatic stage cancer patients suffering from express CD124 (IL4Ra), which is considered a feature of Gr- non–small cell lung cancer (n ¼ 11) and carcinoma of the MDSC (44). prostate (n ¼ 7), the relative number of these MDSC was studied In immunocompetent mice-bearing CT26-GM-derived sub- among peripheral blood leukocytes collected from centrifuged cutaneous tumors, we tested the effect of four daily repeated ficoll gradients. Multiple color flow cytometry permitted analyses doses of reparixin on the intratumoral content of MDSC as on gated HLA-DR /low, lineage negative cells, that were charac- assessed by FACS on tumor-derived cell suspensions. As can be terized by CD33 and CD11b-positive staining and which we seen in Fig. 2D, reparixin treatment decreases the numbers of considered MDSC. Among these cells, subsets were characterized GrMDSC, whereas it did not affect the numbers of MoMDSC in by mutually exclusive coexpression of CD14bright or CD15. These this setting. cells correspond to the so-called monocytic (MoMDSC) and Our results confirm that IL8 chemoattraction of MDSC can take granulocytic (GrMDSC) MDSC subtypes. Figure 3A shows the place in in vivo settings of tumor-bearing mice and show that strategy of FACS staining for sorting and Fig. 3B the relative pharmacological inhibition of CXCR1 and CXCR2 is capable of numbers of these cells in the peripheral blood of cancer patients neutralizing such a function. (n ¼ 18) and healthy volunteers (n ¼ 10). Relative numbers of total MDSC were much higher in this sequential series of heavy Human MDSC are attracted by IL8 burden cancer patients (P < 0.001). Similar high numbers of Peripheral blood from advanced cancer patients is frequently circulating MDSC were assessed by analyzing blood leukocytes characterized by the relative abundance of myeloid populations following osmotic lysis of erythrocytes without ficoll separation þ defined as lymphoid lineage negative, CD11b , HLA-DR /low, (Supplementary Fig. S3). þ and CD33 (25). These cells are considered equivalent to mouse We next sought to determine if the known receptors for IL8, MDSCs (16). Although increases in number among PBMCs in CXCR1 and CXCR2 were expressed on human MDSC. As can be

Figure 3. Advanced lung and prostate carcinoma patients have increased numbers of MDSC in peripheral blood. A, FACS gating strategy to deﬁne MoMDSC and GrMDSC from ﬁcoll layers derived from peripheral blood of advanced cancer patients used for sorting. Purity following sorting was above 99% upon reanalysis. B, percentage of total MDSC in peripheral blood leukocytes of advanced cancer patients (n ¼ 18)incomparisontohealthydonors(n ¼ 10). LC, lung cancer; PC, prostate cancer. P assessed by Student t-test.

OF6 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

seen in Fig. 4A, GrMDSC brightly expressed CXCR1 and CXCR2 PMN granulocytes to extrude their DNA that is decorated with whereas MoMDSC mainly expressed CXCR2. antimicrobial polypeptides to entrap and kill microorganisms The serum of advanced cancer patients is usually rich in IL8 (30). These structures are termed NETs (31). We studied if IL8 content, as this chemokine is commonlyproducedbymalig- can modify these functions. nant cells (2). We, therefore, asked if IL8 can chemotactically To our surprise, freshly isolated PMN from healthy volunteers attract human MDSC in classical transwell migration assays. extruded NETs upon exposure to endotoxin-free recombinant IL8. Figure 2C shows that FACS-sorted human MoMDSC and In fact, the degree of NET formation was comparable to that GrMDSC were attracted by recombinant IL8 in a dose-depen- elicited by incubation in the presence of LPS (Fig. 6A). We dent fashion and by IL8-containing culture supernatants of carefully excluded contamination of endotoxin in the rhIL8 HT29 colon carcinoma cells and by IL8-containing sera of commercial batches (see Materials and Methods). patients (Fig. 4B). Importantly, a neutralizing anti-IL8 mAb Next, we investigated if NETs could be induced among was able to reduce the migration as induced by patient sera human FACS-sorted MDSC from cancer patients. As can be and HT29-conditioned media (Fig. 4B). Figure 4C shows the seen in Fig. 6B and C, GrMDSC formed NETs in short-term concentration of IL8 quantitated by ELISA in the culture cultures in the presence of endotoxin-free IL8, whereas supernatants of HT29 and the sera from cancer patients used MoMDSCfailedtodoso. in these experiments. Furthermore, additional chemotaxis Figure 6D shows representative images of NETs formation transwell experiments were performed with GrMDSC and as induced by the autologous serum of a cancer patient MoMDSC from four advanced cancer patients assessing the diluted 1:5. As can be seen in the figure, the extent of NETs response to their own serum or to serum from a healthy formation induced by serum was similar to that achieved by donor. Supplementary Figure S4 shows that although both LPS. However, in the case of patient sera, the IL8 neutralizing subsetsofMDSCmigratedtowardtheserumofthecancer antibody failed to inhibit NETs formation indicating that the patient, this was not the case with serum of the healthy patients sera contains additional factors involved in NETs volunteers. induction, which are absent in sera from healthy volunteers (Fig. 6D). In this setting, a series of experiments indicated that IL8 induced GrMDSC netosis in a dose-dependent manner T-cell suppressor activity of peripheral blood human MDSC is (Supplementary Fig. S6A) and was abrogated by an anti-IL8 not modified by IL8 mAb that could not inhibit netosis induced by LPS (Supple- Reportedly, MDSC in mice have the ability to downregulate mentary Fig. S6B). T-cell activation and proliferation through a number of molec- Formation of NETs as induced by IL8 in human GrMDSC was ular mechanisms that are exerted by cell-to-cell contact (16). also documented by time-lapse fluorescence confocal microscopy Human MDSC have not been consistently tested for such an that observed GrMDSC whose DNA had been prelabeled with immunosuppressive activity on T-cell cultures (25). Hence, we DRAQ5 in medium that contained Sytox-green (a non –cell pear- set up co-cultures of sorted peripheral blood MoMDSC and meant DNA probe) to stain extracellular DNA. As can be seen in GrMDSC from patients with autologous T cells derived from Supplementary Video S1, exposure to IL8 (10 ng/mL) results in the same samples at different MDSC to T-cell ratios. CFSE- extrusion of DNA that formed double-stained NETs in a 2- to 3- labeled T lymphocytes in total PBMCs were stimulated with hour time frame. Such an effect was not seen if IL8 were not added plastic-bound anti-CD3 and anti-CD28 mAb in the same plate. to the cultures (data not shown). Proliferation of T lymphocytes was assessed by CFSE dilution In cocultures of stimulated Gr-MDSC and red-labeled HT29 assays, which showed a partial inhibition by MoMDSC at 1:1, cancer cells, we observed that elongated NETs seemed to entrap 1:5, and 1:20 ratios (Fig. 5A, left). By contrast, GrMDSC did tumor cells when that GrMDSC extruded NETs in response to not inhibit autologous T-cell proliferation in these in vitro IL8 or LPS (Supplementary Fig. S6C). Supplementary video 2 experiments (Fig. 5A, right). The extent of the proliferation (right) shows similar cocultures with time-lapse confocal inhibition by MoMDSC in a representative case is shown in the microscopy clearly indicating that the formed NETs upon CFSE dilution histograms. Importantly, IL8 added to the cul- response to IL8 were able to entrap tumoral HT29 cells (pre- tures did not modify the extent of T-cell suppression by labeled in red). This phenomenon of NET formation did not MoMDSC in six out of six cases. In conclusion, although IL8 take place if IL8 was not added to the cultures (Supplementary attracts these myeloid subsets, it does not affect to a measur- Video S2, left). able extent the capability of these MoMDSC to inhibit T-cell According to these results, it is conceivable that both proliferation on per cell basis. As experiments in Fig. 5 had GrMDSC and neutrophils of cancer patients under the path- been performed with total PBMC as the source of proliferating ophysiological conditions of exposure to IL8 may be induced lymphocytes, we performed two independent experiments to extrude TRAPs with potential involvement in disease-rele- with purified CD4 and CD8 T cells in co-culture with autol- vant mechanisms. ogous GrMDSC or Mo-MDSC. Similar results indicating suppressive activity on CD4 and CD8 T-cell proliferation by MoMDSC but not by GrMDSC were observed in both experi- Discussion ments (Supplementary Fig. S5 shows histograms from one of Decades ago it was discovered that the myeloid compartment the cases). of tumor-bearing mice contained subsets of cells capable of suppressing T-cell proliferation and activation (45), determin- IL8 induces formation of NETs in human MDSC ing a worse course of the transplanted malignancies (21). In the recent years it has been discovered that a key mech- The field of study of myeloid cells in experimental can- anism in the defense against bacteria and fungi is the ability of cer has extensively shown that both tumor-associated

www.aacrjournals.org Clin Cancer Res; 2016 OF7

Alfaro et al.

Figure 4. MDSC express CXCR1 and CXCR2 and chemotactically respond to IL8. A, surface immunostaining for CXCR1 and CXCR2 on human MDSC derived respectively from peripheral blood of advanced cancer patients (n ¼ 6). Representative FACS histograms are included as insets. Background immunostaining was controlled by ﬂuorescenceminus1.B,chemotacticresponseofGrMDSC and MoMDSC from patients or dextran-gradient puriﬁed neutrophils (PMN) from healthy donors in response to the indicated stimuli including tissue culture supernatant of colon cancer HT29 cells, IL8-containing serum from cancer patients, and recombinant IL8 used at different concentrations. When indicated, an IL8 neutralizing mAb was added to the lower transmigration chamber. C, the IL8 concentration of HT29 supernatants and patient sera used in the experiments. Results are representative of four independent repetitions. Statistical comparisons were performed by Mann–Whitney U test.

OF8 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

Figure 5. Human MoMDSC suppress T-cell proliferation in culture whereas GrMDSC do not. A, coculture experiments of CFSE-labeled PBLs and MoMDSC or GrMDSC from the same cancer patient and assessment of the level of T-cell proliferation in response to anti-CD3 and anti-CD28 mAb by CFSE dilution. Experiments represent three independent cases. B, representative histograms of a case of MoMDSC T-cell coculture in which rhIL8 failed to modify the level of suppression. This experiment was repeated three times with similar results. Mann–Whitney U tests were performed for statistical comparisons. macrophages and MDSC exert protumor activities that In humans, abundant data indicate that the accumulation include depressing the cellular antitumor immune response of MDSC (21) and M2-polarized macrophages (15) cor- using a variety of mechanisms (16). Indeed, therapies that relates with a worse clinical outcome of the disease in a variety selectively damage such myeloid cell subsets exert therapeu- of cancer types (25). MDSC have also been proposed to worsen tic effects (18, 46) and are actively pursued in the clinical the efﬁcacy of ipilimumab immunotherapy in melanoma, but setting (47). it is unclear whether this is predictive or prognostic (21, 25).

www.aacrjournals.org Clin Cancer Res; 2016 OF9

Alfaro et al.

Figure 6. IL8 induces NETs formation in human neutrophils and GrMDSC. A, PMN from healthy subjects or (B) GrMDSC and (C) MoMDSC from cancer patients were exposed in culture to LPS or rhIL8. Results show representative confocal microscopy photographs of NETs and ﬂuorimetric assessment of extracellular DNA ﬂuorescence staining per well. Statistical comparisons were performed with Mann–Whitney U test. D, representative NET formation in GrMDSC and lack of formation in Mo-MDSC following addition of LPS, IL8-containing patient serum or serum from a healthy volunteer. Of note anti-IL8 mAb failed to completely inhibit the patient-serum induction of NETs, suggesting the presence of other factors absent in healthy serum.

OF10 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

Recent deconvoluted transcriptomic evidence also indicates performed in vitro, GrMDSC failed to suppress T-cell cultures but that the presence of neutrophils (indistinguishable from this does not preclude that such an activity could be exerted in vivo GrMDSC in this large series of data) in the tumor microenvi- once conditioned by the tumor microenvironment. ronment is a major factor correlating with worse overall sur- In a search of activities modulated by IL8 in MDSC, we vival prognosis in human cancer (48). discovered that IL8 can induce the formation of NETs by neu- In this study we sought to investigate if IL8, as abundantly trophils. Moreover, similar experiments performed on sorted produced by human cancer cells across malignant diseases (1), GrMDSC from cancer patients indicated that this cell subset also could be a chemotactic factor attracting MDSC. In humans suf- protruded its DNA in a comparable fashion. We ruled out artifacts fering from advanced metastatic cancer it is well known that potentially related to microbial contaminants and, importantly, distinct subsets of myeloid cells accumulate in peripheral blood observed the same effect with IL8-containing serum samples from (25). This is reminiscent of the mouse MDSC counterparts and cancer patients. Of note, healthy donor serum did not induce these cells can also be classified as granulocytic and monocytic NETs whereas cancer patient serum did. Interestingly, the IL8 based on morphology and surface markers (25). Our data con- neutralizing antibody did not completely block this action; thus, clusively demonstrate that IL8 can attract these circulating cells suggesting the presence of other NET-inducing factors in the expressing CXCR1 and CXCR2 on their surface, and it is tempting serum of the patients. to speculate that the chemotactic gradient of IL8 can guide such IL8-induced formation of NETs might have profound implica- MDSC into the tumor microenvironment. In our hands, neutro- tions for microbial defense (31). In the context of our studies, phils from healthy volunteers transmigrate in response to IL8 NETs may be involved in the propensity of cancer patients to chemotaxis better than MDSC. This could be due to differences in develop thromboses (49) and in favoring the adhesive mechan- the levels of receptor expression but also to a more fit cell motility isms of metastasis (16, 37). Indeed, it has recently been described machinery. It is also clear that other chemotactic factors could be that neutrophil-extruded NETs favor the nesting of circulating present in the conditioned culture media and the serum of the tumor cells in the liver sinusoids of mice (37, 50). In our ex vivo patients resulting in incomplete inhibitions by the IL8 neutral- experiments, we observed that MDSC extruded NETs in response izing antibody. to IL8 and thereby could entrap cancer cells ex vivo by adhesive We were able to model that scenario by means of genetically mechanisms in coculture (Supplementary Fig. S6C). Supplemen- enforcing the production of IL8 in mouse liver and attracting tary Video S2 shows similar entrapping events in cocultures with mouse MDSC to this organ. Importantly, this activity can be time-lapse confocal microscopy. It is early to speculate about the therapeutically interfered with using the CXCR1/2 antagonist functional and pathological in vivo consequences of these findings. reparixin (39, 43), thus hinting at a potential cancer immuno- Indeed, nothing is known in cancer patients of what the con- therapy approach. This was also seen in experiments showing the sequences of MDSC attraction and activation to form NETs could effect of reparixin at reducing the numbers of GrMDSC in be. However, our results further argue in favor of the convenience CT26GM tumors implanted in immunocompetent mice. In this of pharmacologically inhibiting IL8 as previously carried out in case, other CXCR1/2 agonist chemokines ought to be involved xenografted mouse models and currently in clinical trials with because mouse tumors cannot produce IL8, which is absent from neutralizing antibodies (51) and small drug inhibitors (52). the murine genome. As a whole, our study unambiguously shows that IL8 is a Apart from its effects on MDSC, IL8 has been found to attract powerful chemotactic stimulus for peripheral patient-derived and retain dendritic cells in the tumor (29) in a disorienting MDSC and that, to our surprise, IL8 directly induces the formation manner that eventually decreases antitumor immunity (10). of NETs out of GrMDSC. Therefore, IL8 could be a multifaceted inflammatory mediator used by the tumor to simultaneously promote angiogenesis Disclosure of Potential Conflicts of Interest – – (1, 3 9), co-opt immunosuppressive populations (1, 3 9), and I. Melero is a consultant/advisory board member for Boehringer Ingelheim, disrupt the routes of migration of professional antigen-present- Bristol-Myers Squibb, Incyte, Novartis, and Roche-Genentech. No potential ing dendritic cells (29). conflicts of interest were disclosed by the other authors. In our series of cancer patients, only MoMDSC from peripheral blood showed a T-cell suppressive activity that was not Authors' Contributions modulated by IL8. In our hands, peripheral blood GrMDSC did Conception and design: C. Alfaro, P. Berraondo, J.L. Perez-Gracia, I. Melero not exert any suppressive effect on T lymphocytes, at least in Development of methodology: C. Alfaro, M.F. Sanmamed, J. Medina-Echeverz, cocultures. We do not have them a clear mechanistic explana- J.L. Perez-Gracia tion to this is in spite of the fact that both subsets express Acquisition of data (provided animals, acquired and managed patients, similar levels of arginase and reactive oxygen species (data not provided facilities, etc.): C. Alfaro, A. Teijeira, C. Onate,~ G. Perez, shown). Other mechanisms remain to be explored to under- M.F. Sanmamed, M.P. Andueza, D. Alignani, S. Labiano, A. Azpilikueta, stand suppression by human primary MDSC and differences A. Rodriguez-Paulete, J.P. Fusco, A. Aznar, J.L. Perez-Gracia Analysis and interpretation of data (e.g., statistical analysis, biostatistics, between subsets. These may encompass mechanisms acting computational analysis): C. Alfaro, A. Teijeira, C. Onate,~ G. Perez, M.P. both in cell-to-cell contacts and at longer distance. It is impor- Andueza, S. Labiano, A. Azpilikueta, A. Rodriguez-Paulete, P. Berraondo, tant to highlight that IL8 and the other chemokines acting via J.L. Perez-Gracia, I. Melero CXCR1/2 are unlikely to be the only chemotactic factors used Writing, review, and/or revision of the manuscript: C. Alfaro, A. Teijeira, by tumors to attract and retain MDSC. M.F. Sanmamed, M.P. Andueza, Rodriguez-Paulete, A. Aznar, S. Inoges, We were able to observe T-cell suppression in culture as medi- P. Berraondo, J.L. Perez-Gracia, I. Melero Administrative, technical, or material support (i.e., reporting or organizing ated by MoMDSC. However, on per-cell basis IL8 did not modify data, constructing databases): C. Alfaro, C. Onate,~ G. Perez, M.P. Andueza, suppression albeit by simply attracting higher MDSC numbers, IL8 S. Garasa, M. De Pizzol, M. Allegretti should result in stronger local T-cell suppression. In experiments Study supervision: M.P. Andueza, J.L. Perez-Gracia, I. Melero

www.aacrjournals.org Clin Cancer Res; 2016 OF11

Alfaro et al.

Acknowledgments 2020 programs (AICR and PROCROP), SUDOE-IMMUNONET, Fundacion ~ The authors thank Eneko Elizalde for animal facility care and Dr. Carlos Ortiz de la Asociacion Espanola Contra el Cancer (AECC), Fundacion Mutua ~ de Solorzano for his help with in vivo confocal imaging. Dr. Mario P. Colombo Madrilena, Fundacion BBVA and Fundacion Caja Navarra. C. Alfaro receives (Molecular Immunology Unit, Fondazione IRCCS Istituto Nazionale dei aSaraBorrellcontractfromISCIII.S. Labiano is recipient of predoctoral Tumori, Milano, Italy) is acknowledged for kindly supplying CT26-GM-CSF scholarship from Ministerio de Economía. transfectants. Reparixin was kindly provided by Dompé Farmaceutici SpA. 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 Grant Support in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. This work was ﬁnancially supported by grants from MICINN (SAF2011- 22831 and SAF2014-52361-R). I. Melero was also funded by the Departa- mento de Salud del Gobierno de Navarra, Redes tematicas de investigacion Received October 11, 2015; revised January 30, 2016; accepted February 3, cooperativa RETICC, European Commission VII Framework and Horizon 2016; published OnlineFirst March 8, 2016.

References 1. Baggiolini M, Walz A, Kunkel SL. Neutrophil-activating peptide-1/inter- derived suppressor cell frequency for prediction of clinical outcomes. leukin 8, a novel cytokine that activates neutrophils. J Clin Invest Cancer Immunol Res 2014;2:812–21. 1989;84:1045–9. 22. Kotsakis A, Harasymczuk M, Schilling B, Georgoulias V, Argiris A, White- 2. Sanmamed MF, Carranza-Rua O, Alfaro C, Onate C, Martin-Algarra S, Perez G, side TL. Myeloid-derived suppressor cell measurements in fresh and et al. Serum interleukin-8 reflects tumor burden and treatment response across cryopreserved blood samples. J Immunol Methods 2012;381:14–22. malignancies of multiple tissue origins. Clin Cancer Res 2014;20:5697–707. 23. Nagaraj S, Gabrilovich DI. Myeloid-derived suppressor cells in human 3. Infanger DW, Cho Y, Lopez BS, Mohanan S, Liu SC, Gursel D, et al. cancer. Cancer J 2010;16:348–53. Glioblastoma stem cells are regulated by interleukin-8 signaling in a 24. Pak AS, Wright MA, Matthews JP, Collins SL, Petruzzelli GJ, Young MR. tumoral perivascular niche. Cancer Res 2013;73:7079–89. Mechanisms of immune suppression in patients with head and neck 4. Kotyza J. Interleukin-8 (CXCL8) in tumor associated non-vascular extra- cancer: presence of CD34(þ) cells which suppress immune functions cellular fluids: its diagnostic and prognostic values. A review. Int J Biol within cancers that secrete granulocyte-macrophage colony-stimulating Markers 2012;27:169–78. factor. Clin Cancer Res 1995;1:95–103. 5. Singh JK, Simoes BM, Howell SJ, Farnie G, Clarke RB. Recent advances 25. Poschke I, Kiessling R. On the armament and appearances of human reveal IL-8 signaling as a potential key to targeting breast cancer stem cells. myeloid-derived suppressor cells. Clin Immunol 2012;144:250–68. Breast Cancer Res 2013;15:210. 26. Kalathil S, Lugade AA, Miller A, Iyer R, Thanavala Y. Higher frequencies of 6. Walz A, Peveri P, Aschauer H, Baggiolini M. Purification and amino acid GARP(þ)CTLA-4(þ)Foxp3(þ) T regulatory cells and myeloid-derived sequencing of NAF, a novel neutrophil-activating factor produced by suppressor cells in hepatocellular carcinoma patients are associated with monocytes. Biochem Biophys Res Commun 1987;149:755–61. impaired T-cell functionality. Cancer Res 2013;73:2435–44. 7. Waugh DJ, Wilson C. The interleukin-8 pathway in cancer. Clin Cancer Res 27. Pico de Coana Y, Masucci G, Hansson J, Kiessling R. Myeloid-derived 2008;14:6735–41. suppressor cells and their role in CTLA-4 blockade therapy. Cancer Immu- 8. Xie K. Interleukin-8 and human cancer biology. Cytokine Growth Factor nol Immunother 2014;63:977–83. Rev 2001;12:375–91. 28. Weide B, Martens A, Zelba H, Stutz C, Derhovanessian E, Di Giacomo AM, 9. Yuan A, Chen JJ, Yao PL, Yang PC. The role of interleukin-8 in cancer cells et al. Myeloid-derived suppressor cells predict survival of patients with and microenvironment interaction. Front Biosci 2005;10:853–65. advanced melanoma: comparison with regulatory T cells and NY-ESO-1- or 10. Alfaro C, Suarez N, Martinez-Forero I, Palazon A, Rouzaut A, Solano S, et al. melan-A-specific T cells. Clin Cancer Res 2014;20:1601–9. Carcinoma-derived interleukin-8 disorients dendritic cell migration with- 29. Feijoo E, Alfaro C, Mazzolini G, Serra P, Penuelas I, Arina A, et al. Dendritic out impairing T-cell stimulation. PLoS One 2011;6:e17922. cells delivered inside human carcinomas are sequestered by interleukin-8. 11. Fan X, Patera AC, Pong-Kennedy A, Deno G, Gonsiorek W, Manfra DJ, et al. Int J Cancer 2005;116:275–81. Murine CXCR1 is a functional receptor for GCP-2/CXCL6 and interleukin- 30. BrinkmannV,ReichardU,GoosmannC,FaulerB,UhlemannY,Weiss 8/CXCL8. J Biol Chem 2007;282:11658–66. DS, et al. Neutrophil extracellular traps kill bacteria. Science 2004;303: 12. Gordon S, Martinez FO. Alternative activation of macrophages: mecha- 1532–5. nism and functions. Immunity 2010;32:593–604. 31. Brinkmann V, Zychlinsky A. Neutrophil extracellular traps: is immunity the 13. Noy R, Pollard JW. Tumor-associated macrophages: from mechanisms to second function of chromatin? J Cell Biol 2012;198:773–83. therapy. Immunity 2014;41:49–61. 32. Nauseef WM, Borregaard N. Neutrophils at work. Nat Immunol 14. Ferrara N.Role of myeloid cells in vascular endothelial growth factor- 2014;15:602–11. independent tumor angiogenesis. Curr Opin Hematol 2010;17:219–24. 33. Clark SR, Ma AC, Tavener SA, McDonald B, Goodarzi Z, Kelly MM, et al. 15. Mantovani A, Germano G, Marchesi F, Locatelli M, Biswas SK. Cancer- Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in promoting tumor-associated macrophages: new vistas and open questions. septic blood. Nat Med 2007;13:463–9. Eur J Immunol 2011;41:2522–5. 34. Pilsczek FH, Salina D, Poon KK, Fahey C, Yipp BG, Sibley CD, et al. A novel 16. Gabrilovich DI, Ostrand-Rosenberg S, Bronte V. Coordinated regulation mechanism of rapid nuclear neutrophil extracellular trap formation in of myeloid cells by tumours. Nat Rev Immunol 2012;12:253–68. response to Staphylococcus aureus. J Immunol 2010;185:7413–25. 17. Arina A, Bronte V. Myeloid-derived suppressor cell impact on endogen- 35. Urban CF, Rembert D, Schmid M, Abu-Abed U, Goosmann C, Nacken W, ous and adoptively transferred T cells. Curr Opin Immunol 2015;33:120–5. et al. Neutrophil extracellular traps contain calprotectin, a cytosolic protein 18. Berraondo P, Nouze C, Preville X, Ladant D, Leclerc C. Eradication of large complex involved in host defense against Candida albicans. PLoS Pathog tumors in mice by a tritherapy targeting the innate, adaptive, and regulatory 2009;5:e1000639. components of the immune system. Cancer Res 2007;67:8847–55. 36. Sangaletti S, Tripodo C, Chiodoni C, Guarnotta C, Cappetti B, Casalini P, 19. Ugel S, Delpozzo F, Desantis G, Papalini F, Simonato F, Sonda N, et al. et al. Neutrophil extracellular traps mediate transfer of cytoplasmic neu- Therapeutic targeting of myeloid-derived suppressor cells. Curr Opin trophil antigens to myeloid dendritic cells toward ANCA induction and Pharmacol 2009;9:470–81. associated autoimmunity. Blood 2012;120:3007–18. 20. Condamine T, Ramachandran I, Youn JI, Gabrilovich DI. Regulation of 37. Cools-Lartigue J, Spicer J, McDonald B, Gowing S, Chow S, Giannias B, et al. tumor metastasis by myeloid-derived suppressor cells. Annu Rev Med Neutrophil extracellular traps sequester circulating tumor cells and pro- 2015;66:97–110. mote metastasis. J Clin Invest 2013:pii:67484. 21. Kitano S, Postow MA, Ziegler CG, Kuk D, Panageas KS, Cortez C, et al. 38. Bizzarri C, Beccari AR, Bertini R, Cavicchia MR, Giorgini S, Allegretti M. Computational algorithm-driven evaluation of monocytic myeloid- ELRþ CXC chemokines and their receptors (CXC chemokine receptor 1

OF12 Clin Cancer Res; 2016 Clinical Cancer Research

Effects of IL8 on MDSC

and CXC chemokine receptor 2) as new therapeutic targets. Pharmacol 46. Di Mitri D, Toso A, Chen JJ, Sarti M, Pinton S, Jost TR, et al. Tumour- Ther 2006;112:139–49. infiltrating Gr-1þ myeloid cells antagonize senescence in cancer. Nature 39. Citro A, Cantarelli E, Maffi P, Nano R, Melzi R, Mercalli A, et al. CXCR1/2 2014;515:134–7. inhibition enhances pancreatic islet survival after transplantation. J Clin 47. Lesokhin AM, Hohl TM, Kitano S, Cortez C, Hirschhorn-Cymerman Invest 2012;122:3647–51. D, Avogadri F, et al. Monocytic CCR2(þ) myeloid-derived suppres- 40. Bronte V, Serafini P, De Santo C, Marigo I, Tosello V, Mazzoni A, et al. IL-4- sor cells promote immune escape by limiting activated CD8 T-cell induced arginase 1 suppresses alloreactive T cells in tumor-bearing mice. infiltration into the tumor microenvironment. Cancer Res 2012;72: J Immunol 2003;170:270–8. 876–86. 41. Ochoa MC, Fioravanti J, Duitman EH, Medina-Echeverz J, Palazon A, Arina 48. Gentles AJ, Newman AM, Liu CL, Bratman SV, Feng W, Kim D, et al. The A, et al. Liver gene transfer of interkeukin-15 constructs that become part of prognostic landscape of genes and infiltrating immune cells across human circulating high density lipoproteins for immunotherapy. PLoS One cancers. Nat Med 2015;21:938–45. 2012;7:e52370. 49. Yalavarthi S, Gould TJ, Rao AN, Mazza LF, Morris AE, Nunez-Alvarez C, 42. Ochoa MC, Fioravanti J, Rodriguez I, Hervas-Stubbs S, Azpilikueta A, et al. Antiphospholipid antibodies promote the release of neutrophil Mazzolini G, et al. Antitumor immunotherapeutic and toxic properties extracellular traps: a new mechanism of thrombosis in the antiphospho- of an HDL-conjugated chimeric IL-15 fusion protein. Cancer Res 2013; lipid syndrome. Arthritis Rheumatol 2015;67:2990–3003. 73:139–49. 50. Tanaka K, Koike Y, Shimura T, Okigami M, Ide S, Toiyama Y, et al. In vivo 43. Citro A, Valle A, Cantarelli E, Mercalli A, Pellegrini S, Liberati D, et al. characterization of neutrophil extracellular traps in various organs of a CXCR1/2 inhibition blocks and reverses type 1 diabetes in mice. Diabetes murine sepsis model. PLoS One 2014;9:e111888. 2015;64:1329–40. 51. Highfill SL, Cui Y, Giles AJ, Smith JP, Zhang H, Morse E, et al. Disruption of 44. Gallina G, Dolcetti L, Serafini P, De Santo C, Marigo I, Colombo MP, et al. CXCR2-mediated MDSC tumor trafficking enhances anti-PD1 efficacy. Sci Tumors induce a subset of inflammatory monocytes with immunosup- Transl Med 2014;6:237ra67. pressive activity on CD8þ T cells. J Clin Invest . 2006;116:2777–90. 52. Boppana NB, Devarajan A, Gopal K, Barathan M, Bakar SA, Shankar EM, 45. Serafini P, De Santo C, Marigo I, Cingarlini S, Dolcetti L, Gallina G, et al. et al. Blockade of CXCR2 signalling: a potential therapeutic target for Derangement of immune responses by myeloid suppressor cells. Cancer preventing neutrophil-mediated inflammatory diseases. Exp Biol Med Immunol Immunother 2004;53:64–72. 2014;239:509–18.

www.aacrjournals.org Clin Cancer Res; 2016 OF13

Tumor-Produced Interleukin-8 Attracts Human Myeloid-Derived Suppressor Cells and Elicits Extrusion of Neutrophil Extracellular Traps (NETs)

Carlos Alfaro, Alvaro Teijeira, Carmen Oñate, et al.

Clin Cancer Res Published OnlineFirst March 8, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-15-2463

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2016/03/08/1078-0432.CCR-15-2463.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2016/05/19/1078-0432.CCR-15-2463. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.