Pathways , and IL-15 Β

Published August 14, 2015, doi:10.4049/jimmunol.1500325 The Journal of Immunology

Priming of Human Resting NK Cells by Autologous M1 Macrophages via the Engagement of IL-1b, IFN-b, and IL-15 Pathways

Irene Mattiola,*,†,‡,1 Matthieu Pesant,*,1 Paolo F. Tentorio,† Martina Molgora,*,‡ Emanuela Marcenaro,x Enrico Lugli,† Massimo Locati,*,‡,2 and Domenico Mavilio†,‡,2

The cross talk between NK cells and macrophages is emerging as a major line of defense against microbial infections and tumors. This study reveals a complex network of soluble mediators and cell-to-cell interactions allowing human classically activated (M1) macrophages, but not resting (M0) or alternatively activated (M2) macrophages, to prime resting autologous NK cells. In this article, we show that M1 increase NK cell cytotoxicity by IL-23 and IFN-b–dependent upregulation of NKG2D, IL-1b–dependent upregulation of NKp44, and trans-presentation of IL-15. Moreover, both IFN-b–dependent cis-presentation of IL-15 on NK cells and engagement of the 2B4-CD48 pathway are used by M1 to trigger NK cell production of IFN-g. The disclosure of these synergic cellular mechanisms regulating the M1–NK cell cross talk provides novel insights to better understand the role of innate immune responses in the physiopathology of tumor biology and microbial infections. The Journal of Immunology, 2015, 195: 000–000.

atural killer cells are important components of the in- the heterogeneous family of inhibitory NK cell receptors and nate immune system and play an active role in the clear- MHC class I molecules allow the NK cell–mediated killing via the N ance of tumor-transformed and viral-infected cells either engagement of a large group of activating NKRs (aNKRs) that bind lacking or down-modulating molecules of the MHC class I without specific ligands expressed on stressed (i.e., infected by pathogens, a prior sensitization (1). Impairments of the interactions between transplanted, inflamed) or tumor-transformed cells (2–4). Upon activation, NK cells also secrete a variety of proinflammatory cytokines and chemokines, which play important regulatory func- *Leukocyte Biology Unit, Humanitas Clinical and Research Center, I-20089 Roz- tions in the context of immune responses, hematopoiesis, and cel- † zano, Milan, Italy; Unit of Clinical and Experimental Immunology, Humanitas lular homeostasis (1, 5). Clinical and Research Center, I-20089 Rozzano, Milan, Italy; ‡Department of Med- ical Biotechnologies and Translational Medicine, University of Milan, I-20089 Roz- Macrophages are well-known for being important players in zano, Milan, Italy; and xDipartimento di Medicina Sperimentale and Centro di the physiopathology of microbial infections and cancer. In line with Eccellenza per le Ricerche Biomediche, Universita` degli Studi di Genova, I-16132 the Th1/Th2 T cell polarization, the stimulation of resting macro- Genoa, Italy phages (M0) with Th1 cytokines (i.e., IFN-g) and the TLR4 ligands 1I.M. and M.P. contributed equally to this work. 2 (i.e., LPS) induces the classical polarization toward M1 macro- M.L. and D.M. are joint senior authors on this work. phages (M1), which display strong microbicidal and tumoricidal ORCIDs: 0000-0003-3077-590X (M.L.); 0000-0001-6147-0952 (D.M.). activities and preferentially promote inflammatory responses. In Received for publication February 11, 2015. Accepted for publication July 19, 2015. contrast, the alternative polarization toward M2 macrophages (M2) This work was supported by Italian Ministry of Health (Bando Giovani Ricercatori) is induced by the Th2 cytokine IL-4 (6, 7). M2 are involved in the Grant GR-2008-1135082 (to D.M.), a Ministry of University and Research Fondo per gli Investimenti della Ricerca di Base “Futuro in Ricerca” project (to M.L.), Italian control of parasite infections and in tissue remodeling and fi- Association for Cancer Research Grants IG 9104 and 14687 (to D.M.) and IG 14685 brosis, and they are endowed with potent proangiogenic features (to M.L.), European Union Marie Curie International Reintegration Grant 322093 (to that favor tumor progression by virtue of their immunoregulatory E.L.) and Collaborative Project TIMER-HEALTH-F4-2011-281608 (to M.L.), the intramural research program of Humanitas Clinical and Research Center (to D.M.), functions (8, 9). The different biological properties of M1 and a European School of Molecular Medicine/Structured International Postdoctoral M2 are also mirrored by distinctive cytokine and chemokine rep- Marie Curie International Mobility Fellowship (to M.P.), and a “Mario e Valeria ertoires (10). Tumor-associated macrophages (TAMs), present in Rindi” Fellowship from the Italian Foundation for Cancer Research (to I.M.). solid tumors at high frequency, display both proangiogenic and I.M., M.P., M.L., and D.M. conceived and designed the experiments; I.M., M.P., P.F.T., M.M., and E.L. performed the experiments; I.M. and M.P. analyzed the data; immune-regulatory features resembling the M2 phenotype (11). It E.M. provided reagents; and I.M., M.P., M.L., and D.M. wrote the manuscript. is now widely accepted that TAMs sustain cancer growth by en- Address correspondence and reprint requests to Prof. Massimo Locati or Prof. Do- gaging vicious cellular cross talks with malignant cells and other menico Mavilio, Leukocyte Biology Unit, Department of Medical Biotechnologies immune effectors within the tumor mass, thus worsening clinical and Translational Medicine, University of Milan School of Medicine, Humanitas Clinical and Research Center, Via Alessandro Manzoni 113, I-20089 Rozzano, Mi- prognosis and favoring disease progression (12–14). lan, Italy (M.L.) or Clinical and Experimental Immunology Unit, Department of Over the past decade, several lines of evidence clearly demon- Medical Biotechnologies and Translational Medicine, University of Milan School strated that NK cells are able to engage bidirectional interactions of Medicine, Humanitas Clinical and Research Center, Via Alessandro Manzoni 113, I-20089 Rozzano, Milan, Italy (D.M.). E-mail addresses: massimo.locati@ with other members of innate immunity, such as autologous den- unimi.it (M.L.) or [email protected] (D.M.) dritic cells (DCs), macrophages, and neutrophils (2, 15–20). The The online version of this article contains supplemental material. final outcome of these synergic interactions mediated by both cell- Abbreviations used in this article: aNKR, activating NKR; DC, dendritic cell; ILC, to-cell contacts and soluble mediators is the coordination and innate lymphoid cell; IL-15Ra, IL-15R a-chain; rh, recombinant human; RT, room optimization of both innate and adaptive immune responses. In temperature; TAM, tumor-associated macrophage. particular, the NK cell–macrophage cross talk is highly relevant Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 in the context of host–pathogen interactions (15, 16, 19) and

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1500325 2 PRIMING OF NK CELLS BY AUTOLOGOUS M1 MACROPHAGES during the course of viral infections (20). However, the mechanism(s) Flow cytometry and cell sorting adopted by activated human macrophages to regulate autologous For direct multicolor flow-cytometry analysis (FACSCanto II; BD Bio- resting NK cell functions are largely unknown. A proper activation sciences) or cell sorting (FACSAria II; BD Biosciences), NK cells, mono- of NK cells can be achieved with different proinflammatory cyto- cytes, and polarized macrophages were washed with HBSS (Lonza) kines, such as IL-2, IL-12, IL-15, IL-18, and type I IFNs (1, 21, 22). supplemented with 2% FBS and incubated for 20 min at 4˚C with the In particular, IL-15 has been shown to play an important role in the following directly conjugated Abs labeled with indicated fluorochromes: CD3/CD56-FITC/PC5 (Beckman Coulter), CD3-FITC (BD Pharmingen), context of NK development, homeostasis, and functions. After its CD3-Brilliant Violet 650 (Biolegend), CD14-FITC (BD Pharmingen), CD14- binding to the high-affinity IL-15R a-chain (IL-15Ra) expressed Brilliant Violet 570 (Biolegend), CD16-PE-Cy7 (BD Pharmingen), CD16- by DCs and monocytes/macrophages, IL-15 is trans-presented to PerCP-Cy5.5 (BD Pharmingen), CD19-allophycocyanin-Cy7 (BD Pharmingen), surrounding cells bearing the lower-affinity IL-15Rbg-chains (22, CD48-PE (BD Pharmingen), CD56-Brilliant Violet 421 (Biolegend), CD69- Pe-Cy7 (BD Pharmingen), CD80-PE (BD Pharmingen), CD117-PE-Cy7 23). The membrane-associated IL-15–IL-15Ra complex can be (Biolegend), CD127-PE-Cy5 (eBioscience), CD206-FITC (BD Pharmingen), also cleaved from cell surface and released in biologic fluids, such CD209-PE (BD Pharmingen), M-DC8-allophycocyanin (Miltenyi), NKp46- as plasma (24, 25). More recently, it emerged that IL-15 can also PE (Beckman Coulter), NKp44-PE (Beckman Coulter), NKp30-PE (Beckman be presented by IL-15Ra to IL-15Rbg-chains expressed on the Coulter), NKG2D-PE (Beckman Coulter), DNAM-1–PE (Beckman Coulter), same cell, thus allowing a mechanism of IL-15 cis-presentation that 2B4-PE (R&D Systems), IL-15Ra–PE (Biolegend), IL-1RII–PE (R&D Sys- tems), and their appropriate isotype controls. For IL-1RI indirect staining, triggers NK cell effector functions (26). Another potent proin- NK cells were incubated for 45 min at room temperature (RT) with purified flammatory cytokine produced by M1 is IL-1b, which has been goat anti-human IL-1RI (R&D Systems) or its polyclonal IgG control and reported to modulate homeostasis and activation of innate lym- for 30 min at RT with mouse anti-goat Alexa Fluor 647–labeled sec- phoid cells (ILCs) in secondary lymphoid tissues (27–29). ILCs ondary Ab (Invitrogen). For sorting experiments, thawed NK cells were stained for 15 min at RTwith LIVE/DEAD Aqua (Invitrogen), followed have been recently identified as an additional lymphocytic pop- by incubation for 20 min at RT with fluorochrome-conjugated Abs and ulation that could be divided in different subsets (ILC1, ILC2, then sorted. FACS-sorted NK cells showed a purity of $98%, and the ILC3) (30). Because NK cells differ from ILC1 by the expression frequencies of cellular contaminants were #0.4% for CD19pos B cells of eomesodermin during their differentiation (30), they are con- and CD3pos T/NKT cells, #0.8% for CD56neg/CD127pos ILCs, and #0.2% pos pos sidered as a distinct immune cell population. However, whether for CD14 monocytes and M-DC8 slan-DCs (Supplemental Fig. 1D, 1E). Data were analyzed with FACSDiva (BD) and FlowJo (Tree Star) software. NK cells are responsive to IL-1b is still being debated in both human and murine settings (26–28, 31, 32). Incubation of NK cells with macrophage-conditioned media or This study investigates on the cross talk between human resting autologous macrophages NK cells and autologous polarized macrophages, and discloses the NK cells were incubated for 24 h at 1.25–2.5 3 105 cells/ml with selective ability of M1 to prime NK cells through a complex net- 30% conditioned media taken from macrophage (M0/M1/M2) cultures, work of interactions that comprises mechanisms associated with 200 U/ml rhIL-2, 20 ng/ml rhIL-12, 10 ng/ml rhIL-15 (Peprotech), 1 ng/ml both the production of soluble mediators and cell-to-cell contacts. rhIL-1b (Miltenyi), 100 ng/ml rhIL-18, or 200 U/ml rhIFN-b (R&D Sys- tems). For direct NK cell–macrophage coculture, 1 3 105 NK cells were grown for 24 h at a 1:1 cell ratio with resting or polarized macrophages, Materials and Methods which were extensively washed to eliminate all soluble factors accumu- NK cell and macrophage isolation and culture lated during the polarization period before analysis. Human PBMCs were isolated from buffy coats of healthy donors obtained Fluorescence microscopy in accordance with clinical protocols approved by the Institutional Review Board of Desio Hospital, Milan, Italy. After centrifugation on a Ficoll den- NK cells treated with macrophage-conditioned media or cytokines were sity gradient (GE Healthcare Biosciences), NK cells and monocytes were attached to poly-L-lysin solution (Sigma)–coated cover glasses. NK cells isolated by negative magnetic cell purification (STEMCELLTechnologies) were fixed and nonspecific staining were prevented by incubation for (33, 34). The purity of enriched resting NK cells was $95%, and the fre- 1 h with PBS (Biosera) supplemented with 2% BSA (Sigma) and 5% quency of other contaminating cells was #3% for NKT/T cells, #2% donkey serum (Sigma). NK cells were then incubated for 1 h at RTwith 2.5 for CD56neg/CD127pos ILCs, #0.5% for CD19pos B cells, #0.2 for CD14pos mg/ml purified goat anti-human IL-1RI (R&D Systems), 5 mg/ml purified monocytes, and #0.2% for M-DC8pos slan-DCs (Supplemental Fig. 1A, goat anti-human IL-15Ra (R&D Systems), or their corresponding isotype 1B). Freshly purified and enriched NK cells were frozen in FBS supple- controls, then washed and incubated for 1 h at RT with a donkey anti- mented with 10% DMSO (Sigma) immediately after their isolation, thawed goat Alexa Fluor 488–labeled secondary Ab (Invitrogen). After repeated when autologous resting or polarized macrophages where ready, and cul- washes, NK cells were incubated 10 min at RT with DAPI (Invitrogen). tured in absence of any stimuli for 24 h at 3 3 106 cells/ml. The viability of CD14pos monocytes were plated directly on cover glasses, differentiated NK cells before the coculture was always $90%, as assessed by trypan blue into resting macrophages, and then polarized, as described earlier. They exclusion (data not shown). Resting NK cells did not undergo a signifi- were then fixed, incubated for 1 h with PBS supplemented with 2% BSA cant cellular activation after being thawed and cultured (Supplemental and 5% donkey serum, and stained for IL-15Ra or IL-1RI, following the Fig. 1C). To generate resting macrophage (M0), we cultured CD14pos same protocol used for NK cells. NK cells and macrophages were not monocytes (cell purity $ 95%) for 7 d with RPMI 1640 supplemented with permeabilized. Images were acquired in oil with an Olympus FV1000 10% FBS, 1% L-glutamine, 1% penicillin/streptomycin (Lonza) and 100 ng/ml (UPL SAPO 603 O NA: 1.35) and analyzed with FV1000 using the Imaris recombinant human (rh) M-CSF (R&D Systems). M0 macrophages were software. then incubated for 24 h with 100 ng/ml LPS purified from Escherichia coli 055:B5 (Sigma) and 20 ng/ml rhIFN-g (Peprotech) to induce M1 polarization Real-time PCR (evaluated by CD80 expression) or 20 ng/ml rhIL-4 (Peprotech) to induce M2 polarization (evaluated by CD206 and CD209 expression; Supplemental Cells were lysed with QIAzol Reagent (Qiagen), and total RNA was ex- Fig. 2A–C). The role of inflammasome was investigated treating resting tracted using miRNeasy mini kit (Qiagen). RNA was converted in cDNA macrophages for 1 h with 50 mg/ml glibenclamide (Labogen) or its vehicle using the High-Capacity cDNA Reverse Transcription Kit (Applied Bio- (DMSO) before stimulation with 100 ng/ml LPS and 20 ng/ml rhIFN-g. systems), and quantification of the following transcripts was performed following the recommended protocols for SYBR Green Master Mix Cell lines (Applied Biosystems) or TaqMan Fast Advanced Master Mix (Applied Bio- systems): IFN-g (forward: 59-CTCTTGGCTGTTACTGCCAGG-39;re- K562 (erythroleukemia), HEK-293T (embryonic kidney), JA3 (Jurkat leu- verse: 59-CTCCACACTCTTTTGGATGCT-39), IL-1RI (TaqMan assay kemia), Raji (Burkitt’s lymphoma), and 221 (B lymphoblastoid) cell lines Hs00991002_m1), IL-15 (TaqMan assay Hs01003716_m1), and IL-15Ra were cultured in DMEM supplemented with 10% FBS, 1% L-glutamine, 1% (TaqMan assay Hs00542604_m1). Results were normalized on s18 (for- pen/strept. The MOLT-4 (acute lymphoblastic leukemia) cell line was grown ward: 59-CCGCAGCTAGGAATAATGGAATA-39; reverse: 59-CGAAA- in RPMI 1640 medium supplemented with 10% FBS, 1% L-glutamine, 1% ACCAACAAAATAGAACCG-39; or TaqMan assay Hs99999901_s1) for pen/strept, and 1% HEPES (Lonza). NK cell analysis or GAPDH (TaqMan assay Hs99999905_m1) for mac- The Journal of Immunology 3 rophage analysis. Data were analyzed by SDS 2.4 software with a 7900HT stimulated with K562 cells (E:T ratio 2:1) for 4 h in both experimental Fast Real Time PCR System (Applied Biosystems) and Opticon Monitor 3 approaches (indirect or direct culture). Macrophage secretion of IL-1b, with a PTC-200 PCR (MJ Research). IL-12p70, IL-15, and IL-15/IL-15Ra weredosedbycommercialELISA kit (Duoset; R&D Systems). Data were analyzed by SoftMaxPro 5.3 Masking experiments software. Saturating concentrations of the following neutralizing Abs were added to Statistical analysis the NK cell culture: NKG2D (1 mg/ml; Biolegend), IL-1b (5 mg/ml; BD Pharmingen), IL-15 (0.5 mg/ml; R&D Systems), IL-18 (5 mg/ml; MBL Results were expressed as mean 6 SEM from multiple independent ex- International Corporation), IL-23p19 (0.8 mg/ml; R&D Systems), IFN-b periments. Two-tailed Student t test or one-sample two-tailed t tests for (0.2 mg/ml; R&D Systems), IL-1RI (2 mg/ml; R&D Systems), and IL-15Ra comparison against fixed values (% of max, set to 100%) were performed (5 mg/ml; R&D Systems). Anti-NKp30 (clone F252), anti-NKp44 (clone using Prism (GraphPad 4) and/or Excel (Microsoft) software: n.s., not sig- KS38), anti-NKp46 (clone KL247), anti-2B4 (clone C054), and anti- nificant; *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. DNAM-1 (clone DX11) mAbs were kindly provided by Prof. Alessandro Moretta (University of Genova, Italy). Macrophage-conditioned media were incubated for 45 min at RT with blocking Abs against cytokines be- Results fore NK cell stimulation. IL-1RI on NK cells was blocked at RT for 45 min Priming of resting NK cells requires M1 polarization before macrophage-conditioned media treatment and left until harvest of the cells. Anti–IL-15Ra was given at the beginning of the treatment To study the ability of either resting or polarized macrophages to with macrophage-conditioned media or NK cell–macrophage coculture prime autologous resting NK cells, we set up two distinct ex- and left until the conclusion of the experiment. Blocking Abs to activating perimental approaches. In the first one, the potential effects of receptors were incubated for 45 min at RT with macrophage-primed macrophage-derived soluble mediators was assessed by incu- NK cells before they were incubated with target cells and left until har- bating NK cells with 30% macrophage-conditioned media taken vest of the cells and on resting NK cells before they were cocultured with polarized macrophages and left for all the period of the coculture. Each from cultures of M0 (M0-primed NK), M1 (M1-primed NK), or neutralizing condition was compared with a matched isotype control. M2 (M2-primed NK) (34). Compared with controls (i.e., untreated resting NK cells), we found a significant increase in IFN-g secre- CD107a degranulation assay and cytokine secretion tion, CD107a degranulation, and CD69 expression selectively in We used CD107a degranulation assay as indicative of NK cell cytotox- M1- but not in M0- or M2-primed NK cells (Fig. 1A–C). In the icity, as previously demonstrated (35). NK cells treated with macrophage- second experimental approach, we analyzed the effects of cel- conditioned media were incubated with K562 at 2:1 E:T ratio or lular interactions by directly coculturing M0, M1, or M2 with au- HEK-293T at 5:1 E:T ratio for 4 h in the presence of PE-labeled CD107a Ab (BD Pharmingen) or its isotype control. K562, JA3, MOLT-4, Raji, or tologous resting NK cells. Macrophages were previously washed 221 (E:T ratio 2:1) were incubated for 4 h directly into NK cell–macro- to eliminate soluble factors accumulated during the polarization phage cocultures, together with PE-labeled CD107a Ab or its isotype period. Our data showed that M1, but not M0 and M2, induced control. Thereafter, cells were washed, stained with CD56 and CD3 Abs to a significant induction of IFN-g secretion, CD107a degranulation, allow gating on NK cells, and analyzed by flow cytometry. IFN-g secretion by NK cells was dosed by commercial ELISA kit (Duoset; R&D Systems). NK and CD69 surface expression compared with resting NK cells cells were washed extensively after treatment with autologous macrophage- (Fig. 1D–F). In both experimental settings, M1 increased NK cy- conditioned media to eliminate any cytokine carryover. NK cells were totoxicity against different target cells, as assessed by CD107a

FIGURE 1. Priming of resting NK cells by macrophage-conditioned media or macrophages. Se- cretion of IFN-g (A–D), percentages of CD107apos (B–E) (gated on CD3neg CD56pos cells) in the presence of K562 target cell line, and expression of CD69 presented as mean fluorescence intensity (MFI) (C–F) (gated on CD3neg CD56pos cells) after incubation of resting NK cells with either M0-, M1-, or M2-conditioned media from autologous macrophages (M0-primed, M1-primed, M2-primed NK cells) (A–C, black bars) or with autologous M0, M1, and M2 (D–F, black bars) compared with negative controls (i.e., untreated resting NK cells in white bars). Mean 6 SEM; n = 11, 17, 7, 10, 15, and3for(A)–(F), respectively. *p , 0.05, **p , 0.01, ***p , 0.001. 4 PRIMING OF NK CELLS BY AUTOLOGOUS M1 MACROPHAGES degranulation assay (35, 36) (Supplemental Fig. 3A, 3C). Taken Moreover, our data showed a significant increase of IL-1R1 on together, these results demonstrate that proinflammatory M1 M1-primed NK cells as compared with M0-primed counterparts. prime resting NK cells by both secreting soluble mediators and Conversely, we did not detect any expression of the decoy IL- establishing cell-to-cell interactions. In contrast, neither M0 nor 1RII, neither on resting nor on macrophage-primed NK cells (42) M2 influence NK cell activation and effector functions under (Fig. 4). The key role of this cytokine in inducing NKp44 up- these experimental settings. regulation was also confirmed by incubating resting FACS-sorted and highly pure NK cells directly with rhIL-1b (Supplemental M1 macrophages induce NKp44 and NKG2D upregulation via Fig. 1D–F). NKp44 expression on FACS-sorted NK cells was also IL-1b and IFN-b induced by M1-conditioned media (Supplemental Fig. 1F), thus To identify the mechanisms involved in the M1-mediated priming ruling out the contribution of those cellular contaminants present of NK cells, we first evaluated the expression of a large panel of at very low frequencies after NK cell enrichment (i.e., monocytes aNKRs after activation of NK cells with either M1-conditioned or slan-DCs) and able to release IL-1b (43, 44). Finally, we also medium or autologous M1. Among all receptors tested on M1- demonstrated that the NKp44 upregulation on M1-primed NK primed NK cells, we detected the selective upregulation of NKp44 cells did not occur in the presence of a prior incubation of M1 with and NKG2D to levels comparable with those observed on IL-2– the inflammasome inhibitor glibenclamide (Fig. 3E), which is in- activated NK cells (37). The incubation of resting NK cells with deed capable of blocking IL-1b secretion by M1 (Supplemental either IFN-g or LPS did not affect the expression of these two Fig. 4B). aNKRs, thus ruling out a potential carryover artifact of the in vitro Although the masking of IL-18, IL-15, and IL-15Ra (Fig. 3B), culture. Finally, the stimulation of resting NK cells with the su- as well as the inflammasome pathway inhibition (Fig. 3F), had no pernatant of autologous M0 activated only with LPS induced an effect on NKG2D expression on M1-primed NK cells, a role was upregulation of both NKp44 and NKG2D similar to that of M1- evident for IFN-b (Fig. 3D). These results are also consistent with primed NK cells (Fig. 2, Supplemental 3D–O). Because these find- a recent report describing that IFN-b produced by activated murine ings indicate that the upregulation of these two aNKRs is associated macrophages regulates NKG2D expression on NK cells (45). The with the production of LPS-dependent cytokines, we then pro- surface levels of NKG2D on M1-primed NK cells also decreased ceeded to selectively block those M1-derived cytokines known to after the masking of IL-23p19 (Fig. 3B), which is known to share modulate NK cell functions (11, 15, 38–40). We first ruled out with IL-12 some functional properties (46, 47). IL-12p70, which was undetectable in the M1-conditioned media Finally, to assess the functional relevance of NKG2D and NKp44 (Supplemental Fig. 4A), as well as IL-18, IL-15, and IL-15Ra, upregulation on M1-primed NK cells, we evaluated NK cell cy- whose masking did not decrease the expression of NKp44 (Fig. tolytic activity against K562 (Fig. 3G) and HEK-293T cell lines 3A). We then analyzed the potential contribution given by IL- (Fig. 3H). We found that K562 are preferentially killed via 23p19 and IL-1b, two LPS-induced proinflammatory cytokines NKG2D, as also previously reported (48), and via NKp30, but not that have been reported to regulate NKp44 expression on ILCs through the engagement of the NKp44 pathway (Supplemental (41). The blocking of IL-1b in the M1-conditioned medium and/or Fig. 3B). Hence we chose K562 as a target cell line to evaluate IL-1RI on NK cells significantly reduced NKp44 upregulation, the NKG2D-mediated cell killing of M1-primed NK cells. Indeed, whereas an anti–IL-23p19 mAb did not decrease the surface ex- the masking of NKp44 did not decrease the degree of degranu- pression of this aNKR (Fig. 3A, 3C). Consistent with these latter lation of M1-primed NK cells against K562, whereas the blocking results and in agreement with previous reports (28, 29, 31, 41), we of NKG2D significantly reduced their percentage of CD107a ex- detected constitutive expression of IL-1RI on resting NK cells. pression to levels similar to those observed on M0-primed NK cells (Fig. 3G). Differently from K562 target cells, HEK 293T cell line expresses ligands for both NKp44 and NKG2D (38, 49, 50). Therefore, the degree of CD107apos M1-primed NK cells was significantly reduced in the presence of NKp44 masking and was re- stored to almost basal levels when NKp44 and NKG2D were blocked together (Fig. 3H). These results confirm that the upregulation of NKp44 and NKG2D on M1-primed NK cells is functionally relevant and contributes to increasing the cytolytic potential of M1-primed NK cells. IFN-b secreted by M1 macrophages induces IL-15 cis-presentation on NK cells and enhances IFN-g production NK cells represent a major source of IFN-g, an important antiviral and immune-regulatory cytokine that plays a key role in the induction of M1 polarization (11). In line with what has been recently reported (51), we first hypothesized that the engagement of the NKG2D pathway could contribute to the increased production of IFN-g by M1-primed NK cells (Fig. 1A). However, our results ruled out this possibility because the masking of NKG2D did not reduce the synthesis and the secretion of IFN-g in M1-primed FIGURE 2. Upregulation of NKp44 and NKG2D on M1-primed NK NK cells (Supplemental Fig. 4C, 4D). A second potential mecha- cells. Surface expression of NKp44 (A) and NKG2D (B)(gatedonCD3neg CD56pos cells) on M0-primed, M1-primed, M2-primed NK cells, and nism enhancing the production of IFN-g could be associated with on NK cells incubated with conditioned media of LPS-activated macro- the M1 secretion of proinflammatory molecules known to trigger phage (black bars) compared with untreated NK cells (white bars). Mean 6 NK cell effector functions (15, 22, 26, 39). Our results showed that SEM; n = 7 (for NKp44) and 13 (for NKG2D). *p , 0.05, **p , 0.01, ***p , blocking of both IL-15 and IL-15Ra significantly reduced IFN-g 0.001. n.s., not significant. secretion in M1-primed NK cells (Fig. 5A). Nevertheless, and in The Journal of Immunology 5

FIGURE 3. Mechanisms and functional relevance of NKp44 and NKG2D upregulation on M1-primed NK cells. (A and B) Expression of NKp44 (A) and NKG2D (B) (gated on CD3neg CD56pos cells) on M0- and M1-primed NK cells either in the absence (isotype controls, white bars) or in the presence of neutralizing mAbs against IL-18 (black bars), IL- 23p19 (gray bars), and IL-15 plus IL- 15Ra (hatched bars). Results are presented as fold change (FC) versus untreated resting NK cells calculated on mean ﬂuorescence intensity (MFI). Mean 6 SEM; n =4(A)and n =3(B). (C and D)Expressionof NKp44 (C) and NKG2D (D)(gated on CD3neg CD56pos cells) on M0- primed and M1-primed NK cells either in the absence (white bars) or in the presence of neutralizing mAbs against IL-1b [black bars in (C)], IL-1RI [gray bars in (C)], IL-1b and IL-1RI together [hatched bars in (C)], and IFN-b [black bars in (D)]. Re- sults were presented as FC versus untreated resting NK cells, calculated on MFI. Mean 6 SEM; n =6(for NKp44) and n = 4 (for NKG2D). (E and F) Expression of NKp44 (E) and NKG2D (F)(gatedonCD3neg CD56pos cells) on NK cells primed with the conditioned medium com- ing from M0 or M1 treated with vehicle (white bars) or glibenclamide (black bars). Results were presented as FC versus M0-primed NK cells (M0-primed NK), calculated on MFI. Mean 6 SEM; n =4.(G and H)Per- centages of CD107apos M0-primed and M1-primed NK cells (gated on CD3neg CD56pos cells) incubated with K562 (G) or HEK-293T (H) target cell lines either in the absence (white bars) or in the presence of neutralizing mAbs against NKp44 (gray bars), NKG2D (black bars) alone, or together (hatched bars). Results are presented as relative percentages (%) referred to the highest level of cell degranulation (M1-primed NK cells, set as 100%). Mean 6 SEM; n =5. *p , 0.05, **p , 0.01, ***p , 0.001. n.s., not signiﬁcant.

line with a previous study (15), we did not detect any level of IL- M1 polarization is associated with a significant production of 15 in M1-conditioned medium, neither in a free form nor con- IFN-b (52, 53). Indeed, M1-primed NK cells showed signifi- jugated to IL-15Ra (Supplemental Fig. 4E, 4F), thus excluding cantly higher levels of both IL-15 and IL-15Ra transcripts com- thepossibleroleofIL-15/IL-15Ra soluble complexes (24, 25). pared with those of either resting or M0-primed NK cells, albeit Because it has been recently reported that IFN-b produced by at lower levels compared with M1 macrophages (Fig. 5B, 5C). DCs induces IL-15/IL-15Ra cis-presentation on murine NK These results also parallel the significantly higher expression of cells (26), we also investigated whether such a mechanism takes IL-15Ra observed on M1-primed NK cells compared with their place in M1-primed human NK cells, taking into account that resting or M0-primed counterparts (Fig. 5D, 5E). Similar to what 6 PRIMING OF NK CELLS BY AUTOLOGOUS M1 MACROPHAGES

FIGURE 4. M1-primed NK cells and IL-1b pathway. (A and B) Expression of IL-1RI (A)andIL-1RII(B)(gatedonCD3neg CD56pos cells) on resting, M0-primed, and M1-primed NK cells (black bars) compared with their isotype controls (white bars). Data are presented as MFI. Mean 6 SEM; n =5(A)and n =2(B). (C) Transcript levels of IL-1RI in resting, M0-primed, M1-primed NK cells, and M1 macrophages. Results are presented as 22DCt. Mean 6 SEM; n =4.(D) Fluorescence microscopic images from one representative example of two different experiments from unrelated healthy donors showing the surface expression of IL-1RI (green) on untreated (i.e., resting), M0-primed, and M1-primed NK cells. The blue DAPI staining indicates cell nuclei. Original magniﬁcation 360. *p , 0.05.

has been observed with murine DCs (26), IL-15 cis-presentation utes to enhanced NK cell cytolytic potential. In contrast, this latter was IFN-b dependent; as in M1-primed NK cells, its masking re- mechanism does not play any role in increasing the production of duced both the expression of IL-15Ra (Fig. 5D) and the secretion IFN-g by NK cells cultured with autologous M1 (Fig. 6B). of IFN-g (Fig. 5A). Furthermore, rhIFN-b directly induced ex- Finally, we analyzed whether M1 polarization was associated pression of IL-15Ra selectively on M1-primed NK cells (Fig. 5E). with the induction of ligands for aNKRs, focusing our attention Taken together, these data demonstrate that IFN-b secreted by M1 on the 2B4-CD48 pathway, which is known to be involved in reg- triggers NK cell production of IL-15 and surface expression of ulating the interaction between NK cell and macrophages (15, 16), IL-15Ra, allowing a mechanism of cis-presentation that, in turn, and contributes to cytokine production by NK cells (16). Our data increases the secretion of IFN-g by NK cells. show that CD48 is highly induced in M1 but not M2 macrophages (Fig. 6C), and demonstrate that the engagement of 2B4 by CD48 is Mechanisms regulating cell-to-cell–dependent NK cell priming functionally relevant in the context of M1-NK cell cross talk because by autologous M1 macrophages the masking of 2B4 significantly reduced the levels of IFN-g pro- Because M1 polarization was confirmed to be associated with a duced by NK cells only when they were incubated with M1 (Fig. 6D). significant increase in IFN-b production, as well as with an induced expression of IL-15 and IL-15Ra (Supplemental Fig. 2) M1-primed NK cells reverse M2 macrophage polarization (39, 52, 53), we next investigated the role of these mediators in Our data demonstrate that M1 macrophages activate NK cells autologous resting NK cell priming (Fig. 1D–F). As expected given and induce IFN-g production. Because IFN-g plays a key role in the pathway normally involved in the NK cell lysis of K562 cell inducing M1 polarization (7), we asked whether M1-primed NK line (Supplemental Fig. 3B), the simultaneous masking of NKp30 cells could influence macrophage polarization by reverting an and NKG2D significantly reduced the amount of CD107a de- established M2 phenotype. To test this, we assessed expression granulation by NK cells incubated with either autologous M0 levels of a panel of M1- and M2-associated genes on M2 macro- or M1 against this target. Interestingly, the additional blocking of phages incubated with the supernatant of autologous M1-primed IL-15Ra,butnotofIFN-b, significantly further reduced the de- NK cells (M1-p NK). Our results showed that the supernatant of gree of CD107a degranulation only in NK cells cocultured with M1-primed NK cells induced a significant upregulation of M1 M1 and not with M0 (Fig. 6A), thus demonstrating that IL-15/IL- markers (CD80, IL-15Ra,IL-15,IL-1b)inM2,similartothat 15Ra trans-presentation by M1 to autologous NK cells contrib- detected in M1 (Fig. 7A–D). A significant reduction of M2 The Journal of Immunology 7

FIGURE 5. IL-15 cis-presentation on M1-primed NK cells. (A) Secretion of IFN-g by M0- and M1-primed NK cells in the presence of neutralizing mAbs against IFN-b (black bars) or IL-15 plus IL-15Ra (hatched bars) compared with isotype controls (white bars). Results are presented as relative percentage referred to the highest level of IFN-g secretion (M1-primed NK cells, set as 100%) (mean 6 SEM; n = 6). (B and C) Transcript levels of IL-15 (B) and IL-15Ra (C) in resting (NK), M0-primed, and M1-primed NK cells and M1 macrophages. Results are presented as 22DCt [mean 6 SEM; n = 4 for (B) and n = 5 for (C)]. (D) Surface levels of IL-15Ra (gated on CD3neg CD56pos cells) on untreated (NK), M0-primed, and M1-primed NK cells either in the absence (isotype control, white bars) or in the presence (black bars) of a neutralizing mAb against IFN-b (means 6 SEM; n = 7). (E) Fluorescence microscopic images from one representative example of two experiments performed with cells of unrelated healthy donors showing the surface expression of IL-15Ra (green) on untreated (i.e., resting), M0-primed, or M1-primed NK cells and NK cells treated for 3 h with 200 U/ml rIFN-b. Blue staining indicates cell nuclei (DAPI). Original magniﬁcation 360. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. markers in already polarized M2 was also observed, although Discussion the M2 phenotype was fully reverted for some markers (CD206, Beyond being effector cells able to eliminate virally infected or ALOX15) and only partially for others (CCL18, CCL22; Fig. 7E– tumor-transformed target cells, NK cells are also endowed with H). Taken together, these results reveal a number of different immune-regulatory functions. Indeed, their ability to interact with pathways sustaining the ability of M1 to promote NK cell anti- other innate immunity cells, such as DCs or neutrophils, is rele- tumoral properties (Fig. 8) and suggest that conversely soluble vant for the development of optimal immune responses against mediators released by M1-primed NK cells have the potential pathogens and tumors (54). The cross talk between NK cells and to re-educate the protumoral M2 phenotype. macrophageshasalsobeenrecentlyreportedtoplayakeyrole

FIGURE 6. Mechanisms that regulate the cell-to-cell–dependent priming of NK cells by autologous M1 macrophages. (A) Percentages of CD107apos (gated on CD3neg CD56pos cells) NK cells cocultured with autologous M0 and M1 either in the absence (isotype controls, white bars) or in the presence of neutralizing mAbs against NKG2D plus NKp30 (black bars), NKG2D plus NKp30 plus IFN-b (gray bars), NKG2D plus NKp30 plus IL-15Ra (hatched bars). Results are presented as relativepercentages referred to the highestlevels of CD107a degranulation (NK cell cocultured with M1,set as 100%)(mean 6 SEM;n = 3). (B) Secretion of IFN-g by NK cells cocultured with M0 or M1 autologous macrophages either in the absence (isotype controls, white bars) or in the presence of neutralizing mAbs against IFN-b (black bars) or IL-15 plus IL-15Ra (hatched bars). Results are presented as relative percentage referred to the highest level of IFN-g secretion (NK cells cocultured with M1, set as 100%; mean 6 SEM; n = 4). (C) Expression of CD48 on M0, M1, M2, or resting macrophages activated with LPS (M0+LPS) or IFN-g (M0+IFN-g; means 6 SEM;n = 4). (D) Secretion of IFN-g by NK cells cocultured with M0 or M1 autologous macrophages either in the absence (isotype control, white bars) or in the presence of a neutralizing mAb against 2B4 (black bars). Results are presented as relative percentages referred to the highest level of IFN-g secretion (NK cells cocultured with M1; set as 100%; mean 6 SEM; n = 4). *p , 0.05, ***p , 0.001, ****p , 0.0001. 8 PRIMING OF NK CELLS BY AUTOLOGOUS M1 MACROPHAGES

FIGURE 7. M1-primed NK cells are endowed to revert M2 polarization. Transcript levels of CD80 (A), IL-15Ra (B), IL-15 (C), IL-1b (D), CD206 (E), ALOX15 (F), CCL-18 (G), and CCL-22 (H) in M0, M1, and M2 macrophages treated or not with the supernatant of M1-primed NK cells. Results are presented as 22DCt [means 6 SEM; n = 6 for (A)–(D), n = 5 for (E)–(H)]. *p , 0.05, **p , 0.01, ***p , 0.001. in the modulation of innate immunity in the presence of inflam- showing that human NK cells express detectable levels of IL-1RI matory insults (15–17), but the underlying mechanisms regulating and that IL-1b has biological activities on human NK cells (28, this interplay are still poorly understood. This study demonstrates 29, 31). In particular, it has been also reported that IL-1b acts as a that only M1 primes human NK cells, whereas M0 and M2 do not synergistic cytokine with IL-12, IL-15, and IL-23 for IFN-g induce NK cell activation and effector functions. This M1-mediated production (28, 31, 32) and, to the best of our knowledge, this priming increasing NK cell cytolytic potential and IFN-g pro- study reports for the first time that IL-1b also increases human duction occurs through the engagement of four different pathways NK cell cytotoxicity via the induction of NKp44 expression. Indeed, working in synergy to ensure the establishment of effective innate our results demonstrate that: 1) M1-primed NK cells have a signifi- immune responses (Fig. 8). cantly higher expression of IL-1RI compared with their resting We disclosed two distinct mechanisms with impact on NK counterparts; 2) the incubation of NK cells in vitro with rhIL-1 cell cytotoxicity, evaluated as CD107a degranulation (35). The first upregulated NKp44; 3) the masking of IL-1b and IL-1RI de- relies on soluble mediators and is sustained by the M1 production creased NKp44 expression on M1-primed NK cells; and 4) the of IL-23p19, IFN-b,andIL-1b, which, in turn, induces the up- use of the inflammasome inhibitor glibenclamide inhibited M1 regulation of NKG2D and NKp44. The binding of these two production of IL-1b, thus preventing the M1-induced upregula- aNKRs to their putative ligands on target cells enhances NK cell tion of NKp44. Of note, this natural cytotoxicity receptor is cytotoxicity. The involvement of type I IFN in regulating the ex- present in humans but is not conserved in mice (57, 58). This is pression of NKG2 receptors has been previously reported (55), one of several significant differences in the repertoires of sur- and our results showing the role of M1-secreted IFN-b in in- face receptors and lineage markers occurring between these two creasing the surface levels of NKG2D extend to a human setting species and not only in respect to NK cell phenotype, but also a finding already reported for murine NK cells (45). Interestingly, in the context of macrophage polarization (39). These discrep- it has been reported that IL-12 is also able to regulate NKG2D ancies underline the importance of taking into account species expression on NK cells (46) and that STAT3 is involved in the specificity when investigating cellular interactions and ask for transcriptional regulation of this aNKR (56). Our data showing caution before transferring to human immunology results ob- the absence of IL-12 in M1-conditioned media and the IL-23– tained in murine experimental settings. mediated increase of NKG2D on M1-primed NK cells suggest The second mechanism enhancing the cytolytic activity of that the STAT3 pathway is likely associated with this phenomenon NK cells requires cell-to-cell interactions and is mediated by downstream of the IL-12b1/IL-23R complex. IL-15. It is well-known that IL-15 is a potent inducer of NK ef- We also provide in this article experimental evidence showing fector functions (21, 38) and is endowed with the ability to work that human NK cells express IL-1RI and are responsive to IL-1b, as a membrane-bound IL-15–IL-15Ra complex trans-presented advancing a matter of current debate. Indeed, although the rela- in a paracrine fashion to cells expressing IL-15Rb (22, 23). We tively low NK cell expression of IL-1RI, as compared with ILCs, previously demonstrated that upregulation of IL-15 and IL-15Ra has led to the general consensus that NK cells are negative for genes is a hallmark of M1 polarization (39), and consistent with this receptor (30), our findings are in line with previous reports this background, we disclose in this article that IL-15Ra expressed The Journal of Immunology 9

FIGURE 8. Macrophage–NK cell interactions. Diagram summarizes the cellular pathways that regulate the cross talk between human NK cells and M1 macrophages. First, polarization to M1 induces CD48 expression on macrophages, whose binding to 2B4 triggers NK cell production of IFN-g (cellular interaction, upper left). Second, M1 secretion of IFN-b induces NK cell production and surface expression of IL-15 and IL-15Ra, respectively. In turn, the IL-15–IL-15Ra membrane-bound complex is cis-presented in an autocrine loop to the same NK cell expressing the IL-15Rb–IL-15Rg receptor complex, thus enhancing NK cell production of IFN-g (soluble interaction, upper right). Third, M1 secretion of IL-1b and IFN-b or IL-23 induces the upregulation of NKp44 and NKG2D, respectively, that, in turn, increases NK cell cytotoxicity against target cells expressing their putative ligands (soluble interaction, lower right). Fourth, the IL-15–IL-15Ra membrane-bound complex expressed by M1 is trans-presented in a paracrine loop to surrounding NK cells expressing the IL-15Rb–IL-15Rg receptor complex and induces NK cell activation and cytotoxicity (cellular interaction, lower left). by M1 trans-presents IL-15 to surrounding NK cells, thus en- either tumor-transformed or infected by opportunistic pathogens, hancing their cytotoxicity properties. Different from IFN-b and in the production of antiviral cytokines/chemokines (including IL-1b, which selectively upregulate specific aNKRs, this IL-15 IFN-g), and in the interaction with autologous DCs explain, at trans-presentation pathway might favor NK cell cytotoxicity least in part, HIV-1 escape from innate immune responses of the by increasing intracellular levels of granzyme B (59). host (5). These pathologic NK cell effector functions are directly Our study also provides experimental evidence of two addi- associated with the expansion of a highly anergic CD56neg/CD16pos tional pathways involved in the M1-induced NK cell secretion of (CD56neg) NK cell subset, whose frequency is very low in healthy IFN-g. The first mechanism relies on the ability of M1-polarized individuals (35, 37). Moreover, the presence of these aberrant macrophages to synthetize IFN-b (45). We demonstrate that CD56neg NK cells is also paralleled by a preferential expansion of NK cells respond to IFN-b by increasing their surface expression M2 in AIDS patients (61). In contrast, macrophages are suscep- of IL-15Ra and their synthesis of IL-15. This allows autocrine tible to HIV-1 infection and represent an important reservoir for IL-15 cis-presentation on NK cells, which, in turn, enhances their the virus (61, 62). Hence it is conceivable to hypothesize that the production of IFN-g. These results are in line with a recent report physiopathology of NK cells and macrophages in HIV-1 infection investigating the DC–NK cell cross talk in the murine setting is associated with a dysfunctional M1–NK cell cross talk. This (26). The second cell contact–dependent mechanism enhancing study sheds new light on the mechanisms required for an effective IFN-g production by NK cells relies on the activation of the NK cell–M1 interplay, and therefore paves the groundwork to CD48-2B4 pathway (15, 16, 20). We show in this article that this better understanding and breaks this vicious cycle to overcome binding occurs only in the presence of M1 polarization. In fact, the defective host immune responses against HIV-1. although NK cells constitutively express high levels of 2B4, Both NK cells and macrophages are also present in tumor M0 and M2 are negative for the expression of its natural ligand stroma and are naturally endowed with antitumoral activities CD48. It has been reported that the engagement of the CD48-2B4 (1, 9). However, under the effect of tumor-derived signals, macro- pathway induces to the activation of FYNT, which, in turn, induces phages acquire a protumoral M2-like functional profile, which the NK cell production of IFN-g (60), thus further supporting our not only develops protumoral functions (63) but also, as shown in findings showing the importance of the CD48–2B4–IFN-g axis this study, compromises their priming effects on NK cells. Thus, in the context of NK cell–macrophage cross talk. acting on TAMs, the tumor microenvironment interferes with the The M1-induced NK cell production of IFN-g likely plays a role antitumoral potential of both cell types. Several approaches have in the context of rescuing M2 toward a proinflammatory M1 profile been proposed to target TAMs therapeutically, including blocking to boost and optimize innate immune responses against infections. their homing to tumors (64) or depleting them in situ (65). As Indeed, a defective or aberrant NK cell–M1 cross talk might favor recently demonstrated with CD40 agonists in pancreatic cancer viral or bacterial infections, as reported in HIV-1–infected patients. (66), TAM reprogramming to M1-like phenotypes represents a Indeed, the reported NK cell impairment in the clearance of autol- promising approach that not only detracts the tumor from its pro- ogous HIV-1–infected CD4pos T cells, in the killing of cell targets tumoral activities but also converts them into efficient antitumor 10 PRIMING OF NK CELLS BY AUTOLOGOUS M1 MACROPHAGES cell mediators (67). Our results indicate that the ability of M1 to 17. Michel, T., F. Hentges, and J. Zimmer. 2012. Consequences of the crosstalk between monocytes/macrophages and natural killer cells. Front. Immunol. 3: 403. engage a positive cross talk with NK cells improves NK cytolytic 18. Costantini, C., F. Calzetti, O. Perbellini, A. Micheletti, C. Scarponi, S. Lonardi, potential and increases IFN-g production, which would feedback M. Pelletier, K. Schakel, G. Pizzolo, F. Facchetti, et al. 2011. Human neutrophils on macrophages and potentiate their antitumoral activity (Fig. 8). interact with both 6-sulfo LacNAc+ DC and NK cells to amplify NK-derived IFNgamma: role of CD18, ICAM-1, and ICAM-3. Blood 117: 1677–1686. On this important matter, we also demonstrate in this article that 19. Lapaque, N., T. Walzer, S. Me´resse, E. Vivier, and J. Trowsdale. 2009. Inter- M1-primed NK cells are able to downmodulate the expression of actions between human NK cells and macrophages in response to Salmonella protumoral genes (M2 markers) and to increase the expression of infection. J. Immunol. 182: 4339–4348. 20. Romo, N., G. Magri, A. Muntasell, G. Heredia, D. Baıá, A. Angulo, M. Guma, antitumoral genes (M1 markers) in M2. Interestingly, the condi- and M. Lo´pez-Botet. 2011. Natural killer cell-mediated response to human tioned media from M1-primed NK cells induced in M2 the ex- cytomegalovirus-infected macrophages is modulated by their functional polari- pression of inflammatory genes (IL-15, IL-15Ra, IL-1b), which, zation. J. Leukoc. Biol. 90: 717–726. 21. Cooper, M. A., T. A. Fehniger, and M. A. Caligiuri. 2001. The biology of human in turn, sustain NK cell activation. This suggests that therapeutic natural killer-cell subsets. Trends Immunol. 22: 633–640. approaches able to re-educate TAM to M1-like macrophages may 22. Fehniger, T. A., and M. A. Caligiuri. 2001. Interleukin 15: biology and relevance to human disease. Blood 97: 14–32. also take advantage of this positive feed-forward loop with NK 23. Jakobisiak, M., J. Golab, and W. Lasek. 2011. Interleukin 15 as a promising cells, which should, in principle, result is setting in motion a pow- candidate for tumor immunotherapy. Cytokine Growth Factor Rev. 22: 99–108. erful antitumoral immune mechanism. 24. Steel, J. C., T. A. Waldmann, and J. C. Morris. 2012. Interleukin-15 biology and its therapeutic implications in cancer. Trends Pharmacol. Sci. 33: 35–41. In summary, this study describes a complex network of cellu- 25. Bergamaschi, C., J. Bear, M. Rosati, R. K. Beach, C. Alicea, R. Sowder, lar pathways that regulate, in the context of a human autologous E. Chertova, S. A. Rosenberg, B. K. Felber, and G. N. Pavlakis. 2012. Circu- setting, cellular and soluble pathways operated by macrophages to lating IL-15 exists as heterodimeric complex with soluble IL-15Ra in human and mouse serum. Blood 120: e1–e8. efficiently prime resting NK cells. In particular, we disclose in 26. Zanoni, I., R. Spreafico, C. Bodio, M. Di Gioia, C. Cigni, A. Broggi, T. Gorletta, this article cytokine (IL-15, IL-1b, IFN-b) and membrane receptor M. Caccia, G. Chirico, L. Sironi, et al. 2013. IL-15 cis presentation is required pathways (CD48-2B4) working in synergy to ensure appropriate for optimal NK cell activation in lipopolysaccharide-mediated inflammatory conditions. Cell Reports 4: 1235–1249. support from M1-polarized macrophages to NK cell–mediated 27. Cella, M., K. Otero, and M. Colonna. 2010. Expansion of human NK-22 cells immune responses, highly relevant with respect to tumor biology with IL-7, IL-2, and IL-1beta reveals intrinsic functional plasticity. Proc. Natl. Acad. Sci. USA 107: 10961–10966. and microbial infections. Our findings also open new insights with 28. Hughes, T., B. Becknell, A. G. Freud, S. McClory, E. Briercheck, J. Yu, C. Mao, regard to the potential ability of M1-activated NK cells in reverting C. Giovenzana, G. Nuovo, L. Wei, et al. 2010. Interleukin-1beta selectively M2 polarization and boosting a preferential proinflammatory mac- expands and sustains interleukin-22+ immature human natural killer cells in secondary lymphoid tissue. Immunity 32: 803–814. rophage response. 29. Glatzer, T., M. Killig, J. Meisig, I. Ommert, M. Luetke-Eversloh, M. Babic, D. Paclik, N. Bluthgen,€ R. Seidl, C. Seifarth, et al. 2013. RORgt⁺ innate lym- Disclosures phoid cells acquire a proinflammatory program upon engagement of the activating receptor NKp44. Immunity 38: 1223–1235. The authors have no financial conflicts of interest. 30. Spits, H., D. Artis, M. Colonna, A. Diefenbach, J. P. Di Santo, G. Eberl, S. Koyasu, R. M. Locksley, A. N. McKenzie, R. E. Mebius, et al. 2013. Innate lymphoid cells–a proposal for uniform nomenclature. Nat. Rev. Immunol. 13: References 145–149. 1. Vivier, E., E. Tomasello, M. Baratin, T. Walzer, and S. Ugolini. 2008. Functions 31. Cooper, M. A., T. A. Fehniger, A. Ponnappan, V. Mehta, M. D. Wewers, and of natural killer cells. Nat. Immunol. 9: 503–510. M. A. Caligiuri. 2001. Interleukin-1beta costimulates interferon-gamma pro- 2. Moretta, A. 2002. Natural killer cells and dendritic cells: rendezvous in abused duction by human natural killer cells. Eur. J. Immunol. 31: 792–801. tissues. Nat. Rev. Immunol. 2: 957–964. 32. van de Wetering, D., R. A. de Paus, J. T. van Dissel, and E. van de Vosse. 2009. 3. Moretta, A., C. Bottino, M. Vitale, D. Pende, R. Biassoni, M. C. Mingari, and Salmonella induced IL-23 and IL-1beta allow for IL-12 production by mono- L. Moretta. 1996. Receptors for HLA class-I molecules in human natural killer cytes and Mphi1 through induction of IFN-gamma in CD56 NK/NK-like T cells. cells. Annu. Rev. Immunol. 14: 619–648. PLoS One 4: e8396. 4. Lanier, L. L. 2005. NK cell recognition. Annu. Rev. Immunol. 23: 225–274. 33. Gupta, N., J. Arthos, P. Khazanie, T. D. Steenbeke, N. M. Censoplano, 5. Brunetta, E., K. L. Hudspeth, and D. Mavilio. 2010. Pathologic natural killer cell E. A. Chung, C. C. Cruz, M. A. Chaikin, M. Daucher, S. Kottilil, et al. 2005. subset redistribution in HIV-1 infection: new insights in pathophysiology and Targeted lysis of HIV-infected cells by natural killer cells armed and triggered by clinical outcomes. J. Leukoc. Biol. 88: 1119–1130. a recombinant immunoglobulin fusion protein: implications for immunotherapy. 6. Mills, C. D., K. Kincaid, J. M. Alt, M. J. Heilman, and A. M. Hill. 2000. M-1/M-2 Virology 332: 491–497. macrophages and the Th1/Th2 paradigm. J. Immunol. 164: 6166–6173. 34. Solinas, G., S. Schiarea, M. Liguori, M. Fabbri, S. Pesce, L. Zammataro, 7. Murray, P. J., J. E. Allen, S. K. Biswas, E. A. Fisher, D. W. Gilroy, S. Goerdt, F. Pasqualini, M. Nebuloni, C. Chiabrando, A. Mantovani, and P. Allavena. 2010. S. Gordon, J. A. Hamilton, L. B. Ivashkiv, T. Lawrence, et al. 2014. Macrophage Tumor-conditioned macrophages secrete migration-stimulating factor: a new activation and polarization: nomenclature and experimental guidelines. [Pub- marker for M2-polarization, influencing tumor cell motility. J. Immunol. 185: lished erratum appears in 2014 Immunity. 41: 339–340.] Immunity 41: 14–20. 642–652. 8. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of mac- 35. Brunetta, E., M. Fogli, S. Varchetta, L. Bozzo, K. L. Hudspeth, E. Marcenaro, rophage activation. Nat. Rev. Immunol. 8: 958–969. A. Moretta, and D. Mavilio. 2009. The decreased expression of Siglec-7 rep- 9. Murray, P. J., and T. A. Wynn. 2011. Protective and pathogenic functions of resents an early marker of dysfunctional natural killer-cell subsets associated macrophage subsets. Nat. Rev. Immunol. 11: 723–737. with high levels of HIV-1 viremia. Blood 114: 3822–3830. 10. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. 36. Alter, G., J. M. Malenfant, and M. Altfed. 2004. CD107a as a functional marker for The chemokine system in diverse forms of macrophage activation and polari- the identification of natural killer cell activity. J. Immunol. Methods 294: 15–22. zation. Trends Immunol. 25: 677–686. 37. Mavilio, D., J. Benjamin, M. Daucher, G. Lombardo, S. Kottilil, M. A. Planta, 11. Biswas, S. K., and A. Mantovani. 2010. Macrophage plasticity and interaction E. Marcenaro, C. Bottino, L. Moretta, A. Moretta, and A. S. Fauci. 2003. Natural with lymphocyte subsets: cancer as a paradigm. Nat. Immunol. 11: 889–896. killer cells in HIV-1 infection: dichotomous effects of viremia on inhibitory and 12. Steidl, C., T. Lee, S. P. Shah, P. Farinha, G. Han, T. Nayar, A. Delaney, activating receptors and their functional correlates. Proc. Natl. Acad. Sci. USA S. J. Jones, J. Iqbal, D. D. Weisenburger, et al. 2010. Tumor-associated mac- 100: 15011–15016. rophages and survival in classic Hodgkin’s lymphoma. N. Engl. J. Med. 362: 38. Varchetta, S., B. Oliviero, D. Mavilio, and M. U. Mondelli. 2013. Different 875–885. combinations of cytokines and activating receptor stimuli are required for human 13. Mantovani, A., P. Allavena, A. Sica, and F. Balkwill. 2008. Cancer-related in- natural killer cell functional diversity. Cytokine 62: 58–63. flammation. Nature 454: 436–444. 39. Martinez, F. O., S. Gordon, M. Locati, and A. Mantovani. 2006. Transcriptional 14. Noy, R., and J. W. Pollard. 2014. Tumor-associated macrophages: from mech- profiling of the human monocyte-to-macrophage differentiation and polarization: anisms to therapy. [Published erratum appears in 2014 Immunity. 41: 866.] Im- new molecules and patterns of gene expression. J. Immunol. 177: 7303–7311. munity 41: 49–61. 40. Biswas, S. K., P. Allavena, and A. Mantovani. 2013. Tumor-associated macro- 15. Bellora, F., R. Castriconi, A. Dondero, G. Reggiardo, L. Moretta, A. Mantovani, phages: functional diversity, clinical significance, and open questions. Semin. A. Moretta, and C. Bottino. 2010. The interaction of human natural killer cells Immunopathol. 35: 585–600. with either unpolarized or polarized macrophages results in different functional 41. Bernink, J. H., C. P. Peters, M. Munneke, A. A. te Velde, S. L. Meijer, K. Weijer, outcomes. Proc. Natl. Acad. Sci. USA 107: 21659–21664. H. S. Hreggvidsdottir, S. E. Heinsbroek, N. Legrand, C. J. Buskens, et al. 2013. 16. Nedvetzki, S., S. Sowinski, R. A. Eagle, J. Harris, F. Ve´ly, D. Pende, Human type 1 innate lymphoid cells accumulate in inflamed mucosal tissues. J. Trowsdale, E. Vivier, S. Gordon, and D. M. Davis. 2007. Reciprocal regulation Nat. Immunol. 14: 221–229. of human natural killer cells and macrophages associated with distinct immune 42. Garlanda, C., F. Riva, E. Bonavita, and A. Mantovani. 2013. Negative regulatory synapses. Blood 109: 3776–3785. receptors of the IL-1 family. Semin. Immunol. 25: 408–415. The Journal of Immunology 11

43. Scotton, C. J., F. O. Martinez, M. J. Smelt, M. Sironi, M. Locati, A. Mantovani, 55. Zhang, C., J. Zhang, R. Sun, J. Feng, H. Wei, and Z. Tian. 2005. Opposing effect and S. Sozzani. 2005. Transcriptional profiling reveals complex regulation of the of IFNgamma and IFNalpha on expression of NKG2 receptors: negative regu- monocyte IL-1 beta system by IL-13. J. Immunol. 174: 834–845. lation of IFNgamma on NK cells. Int. Immunopharmacol. 5: 1057–1067. 44. Wehner, R., A. Bitterlich, N. Meyer, A. Kloss, K. Schakel, M. Bachmann, and 56. Goriely, S., M. F. Neurath, and M. Goldman. 2008. How microorganisms tip the M. Schmitz. 2013. Impact of chemotherapeutic agents on the immunostimula- balance between interleukin-12 family members. Nat. Rev. Immunol. 8: 81–86. tory properties of human 6-sulfo LacNAc+ (slan) dendritic cells. Int. J. Cancer 57. Hayakawa, Y., N. D. Huntington, S. L. Nutt, and M. J. Smyth. 2006. Functional 132: 1351–1359. subsets of mouse natural killer cells. Immunol. Rev. 214: 47–55. 45. Zhou, Z., C. Zhang, J. Zhang, and Z. Tian. 2012. Macrophages help NK cells to 58. Moretta, A., C. Bottino, M. Vitale, D. Pende, C. Cantoni, M. C. Mingari, attack tumor cells by stimulatory NKG2D ligand but protect themselves from R. Biassoni, and L. Moretta. 2001. Activating receptors and coreceptors involved NK killing by inhibitory ligand Qa-1. PLoS One 7: e36928. in human natural killer cell-mediated cytolysis. Annu. Rev. Immunol. 19: 197– 46. Zhang, C., J. Zhang, J. Niu, Z. Zhou, J. Zhang, and Z. Tian. 2008. Interleukin-12 223. improves cytotoxicity of natural killer cells via upregulated expression of 59. Mortier, E., R. Advincula, L. Kim, S. Chmura, J. Barrera, B. Reizis, NKG2D. Hum. Immunol. 69: 490–500. B. A. Malynn, and A. Ma. 2009. Macrophage- and dendritic-cell-derived 47. Zhu, S., P. V. Phatarpekar, C. J. Denman, V. V. Senyukov, S. S. Somanchi, interleukin-15 receptor alpha supports homeostasis of distinct CD8+ T cell H. T. Nguyen-Jackson, E. M. Mace, A. F. Freeman, S. S. Watowich, J. S. Orange, subsets. Immunity 31: 811–822. et al. 2014. Transcription of the activating receptor NKG2D in natural killer cells 60. Veillette, A. 2006. Immune regulation by SLAM family receptors and SAP- is regulated by STAT3 tyrosine phosphorylation. Blood 124: 403–411. related adaptors. Nat. Rev. Immunol. 6: 56–66. 48. Bae, D. S., Y. K. Hwang, and J. K. Lee. 2012. Importance of NKG2D-NKG2D 61. Herbein, G., and A. Varin. 2010. The macrophage in HIV-1 infection: from ligands interaction for cytolytic activity of natural killer cell. Cell. Immunol. 276: activation to deactivation? Retrovirology 7: 33. 122–127. 62. Varchetta, S., P. Lusso, K. Hudspeth, J. Mikulak, D. Mele, S. Paolucci, 49. Baychelier, F., A. Sennepin, M. Ermonval, K. Dorgham, P. Debre´, and R. Cimbro, M. Malnati, A. Riva, R. Maserati, et al. 2013. Sialic acid-binding Ig- V. Vieillard. 2013. Identification of a cellular ligand for the natural cytotoxicity like lectin-7 interacts with HIV-1 gp120 and facilitates infection of CD4pos receptor NKp44. Blood 122: 2935–2942. T cells and macrophages. Retrovirology 10: 154. 50. Byrd, A., S. C. Hoffmann, M. Jarahian, F. Momburg, and C. Watzl. 2007. Ex- 63. Hanahan, D., and L. M. Coussens. 2012. Accessories to the crime: functions of pression analysis of the ligands for the Natural Killer cell receptors NKp30 and cells recruited to the tumor microenvironment. Cancer Cell 21: 309–322. NKp44. PLoS One 2: e1339. 64. Qian, B. Z., J. Li, H. Zhang, T. Kitamura, J. Zhang, L. R. Campion, E. A. Kaiser, 51. Horng, T., J. S. Bezbradica, and R. Medzhitov. 2007. NKG2D signaling is coupled L. A. Snyder, and J. W. Pollard. 2011. CCL2 recruits inflammatory monocytes to to the interleukin 15 receptor signaling pathway. Nat. Immunol. 8: 1345–1352. facilitate breast-tumour metastasis. Nature 475: 222–225. 52. Tamassia, N., V. Le Moigne, F. Calzetti, M. Donini, S. Gasperini, T. Ear, 65. Ries,C.H.,M.A.Cannarile,S.Hoves,J.Benz,K.Wartha,V.Runza,F.Rey-Giraud, A. Cloutier, F. O. Martinez, M. Fabbri, M. Locati, et al. 2007. The MyD88- L. P. Pradel, F. Feuerhake, I. Klaman, et al. 2014. Targeting tumor-associated independent pathway is not mobilized in human neutrophils stimulated via macrophages with anti-CSF-1R antibody reveals a strategy for cancer therapy. TLR4. J. Immunol. 178: 7344–7356. Cancer Cell 25: 846–859. 53. Yang, H. T., Y. Wang, X. Zhao, E. Demissie, S. Papoutsopoulou, A. Mambole, 66. Beatty, G. L., E. G. Chiorean, M. P. Fishman, B. Saboury, U. R. Teitelbaum, A. O’Garra, M. F. Tomczak, S. E. Erdman, J. G. Fox, et al. 2011. NF-kB1 W. Sun, R. D. Huhn, W. Song, D. Li, L. L. Sharp, et al. 2011. CD40 agonists inhibits TLR-induced IFN-b production in macrophages through TPL-2- alter tumor stroma and show efficacy against pancreatic carcinoma in mice and dependent ERK activation. J. Immunol. 186: 1989–1996. humans. Science 331: 1612–1616. 54. Hudspeth, K., E. Pontarini, P. Tentorio, M. Cimino, M. Donadon, G. Torzilli, 67. Hagemann, T., T. Lawrence, I. McNeish, K. A. Charles, H. Kulbe, E. Lugli, S. Della Bella, M. E. Gershwin, and D. Mavilio. 2013. The role of R. G. Thompson, S. C. Robinson, and F. R. Balkwill. 2008. “Re-educating” natural killer cells in autoimmune liver disease: a comprehensive review. J. tumor-associated macrophages by targeting NF-kappaB. J. Exp. Med. 205: Autoimmun. 46: 55–65. 1261–1268.