Patrolling Monocytes Control NK Cell Expression of Activating and Stimulatory Receptors to Curtail Lung Metastases

This information is current as Prakash Babu Narasimhan, Tobias Eggert, Yanfang Peipei of September 27, 2021. Zhu, Paola Marcovecchio, Melissa A. Meyer, Runpei Wu and Catherine C. Hedrick J Immunol published online 25 November 2019 http://www.jimmunol.org/content/early/2019/11/22/jimmun

ol.1900998 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2019/11/22/jimmunol.190099 Material 8.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 27, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published November 25, 2019, doi:10.4049/jimmunol.1900998 The Journal of Immunology

Patrolling Monocytes Control NK Cell Expression of Activating and Stimulatory Receptors to Curtail Lung Metastases

Prakash Babu Narasimhan, Tobias Eggert, Yanfang Peipei Zhu, Paola Marcovecchio, Melissa A. Meyer, Runpei Wu, and Catherine C. Hedrick

The role of nonclassical, patrolling monocytes in lung tumor metastasis and their functional relationships with other immune cells remain poorly deﬁned. Contributing to these gaps in knowledge is a lack of cellular speciﬁcity in commonly used approaches for depleting nonclassical monocytes. To circumvent these limitations and study the role of patrolling monocytes in melanoma metastasis to lungs, we generated C57BL/6J mice in which the Nr4a1 superenhancer E2 subdomain is ablated (E22/2 mice). E22/2 mice

lack nonclassical patrolling monocytes but preserve classical monocyte and macrophage numbers and functions. Interestingly, NK Downloaded from cell recruitment and activation were impaired, and metastatic burden was increased in E22/2mice. E22/2 mice displayed un- changed “educated” (CD11b+CD27+) and “terminally differentiated” (CD11b+CD272) NK cell frequencies. These perturbations were accompanied by reduced expression of stimulatory receptor Ly49D on educated NK cells and increased expression of inhibitory receptor NKG2A/CD94 on terminally differentiated NK cells. Thus, our work demonstrates that patrolling monocytes play a critical role in preventing lung tumor metastasis via NK cell recruitment and activation. The Journal of Immunology, 2020, 204: 000–000. http://www.jimmunol.org/ ver the past decade, numerous studies have implicated bone marrow chimeric mice. Ab-mediated depletion of NK cells monocytes in the regulation of NK cell activity (1–5). In reversed the protective effect of monocytes against lung metastasis, O cancer, monocytes both enhance (1, 2, 5) and inhibit (3, 4) highlighting the interplay between these two immune cell types. NK cell antitumor immunity. Classical, inflammatory monocytes Collaboration between patrolling monocytes and NK cells (CD14+CD162 in humans and Ly6C+CCR2+CX3CR1low in mice) against melanoma metastases has been further evidenced by have been linked to NK cell proliferation, cytotoxicity, and two recent studies (8, 9). First, Plebanek and colleagues show cytokine production. For example, TLR 4 activation upregu- that exosomes from poorly metastatic melanomas expand the lates expression of the NKG2D ligand MICA on CD14+ monocytes patrolling monocyte population in the bone marrow through and induces NK cell–dependent IFN-g production (5). These MICA- expression of Nr4a1. In agreement with our own findings, this by guest on September 27, 2021 expressing monocytes enhance NK cell, Ab-dependent cytotoxicity paper also found that depletion of NK cells partially reverses against Ab-labeled cancer cells (6). Few studies, in contrast, have the antimetastatic effect of patrolling monocytes (8). Second, explored the role of nonclassical, patrolling monocytes Kubo and colleagues (9) provide data indicating that the (CD16+CD14low/2 in humans and Ly6Clow/2CX3CR1high in presence of primary, s.c. melanomas inhibit formation of mice) on NK cell antitumor function. melanoma lung metastases by expanding patrolling monocyte- Our laboratory has demonstrated that patrolling monocytes re- mediated NK cells in the lung. duce metastatic burden in mouse models of both melanoma lung A major limitation of the aforementioned studies is that the metastasis and primary breast cancer with spontaneous metastasis experimental approaches used for depleting patrolling monocytes to the lung. We have also shown that this effect is partially de- are not cell specific and therefore reduce numbers of additional pendent on NK cells (7). Specifically, we found that, by producing cell types. Furthermore, genetic ablation of Nr4a1 in mice also chemokines CCL3, CCL4, and CCL5, patrolling monocytes affects macrophage function (10). We have overcome these ob- recruit NK cells to the lung. This involvement of monocytes in stacles by ablating an Nr4a1 superenhancer E2 subdomain in the prevention of lung tumor metastasis was confirmed in pa- mice, resulting in efficacious depletion of patrolling monocytes trolling monocyte-deficient, Nr4a1 whole-body knockout and while leaving classical monocyte and macrophage function unaffected (11). In this study, we describe the use of these animals (henceforth referred to as E22/2 mice) to confirm the role of Division of Inflammation Biology, La Jolla Institute for Immunology, La Jolla, CA patrolling monocytes in NK cell–dependent prevention of mela- 92037 noma metastasis to lungs. In E22/2 mice, both recruitment and ORCIDs: 0000-0001-7111-7124 (P.B.N.); 0000-0001-6908-0860 (Y.P.Z.); 0000- 0003-3050-4819 (M.A.M.). activation of lung NK cells following i.v. injection of melanoma Received for publication August 19, 2019. Accepted for publication October 15, cells, as measured by a diverse set of surface receptors, were 2019. impaired in the absence of patrolling monocytes. In contrast, 2/2 This work was supported by the National Institutes of Health (NCI R01 CA202987 maturation of NK cells in E2 mice was unaffected. Long- and NCI U01 CA224766, both to C.C.H.). term monitoring of these mice showed increased metastatic Address correspondence and reprint requests to Dr. Catherine C. Hedrick, La Jolla burden in lungs, as assessed by imaging of melanoma luciferase Institute for Immunology, 9420 Athena Circle, La Jolla, CA 92037. E-mail address: expression and tumor nodule count. In summary, our work in [email protected] this study shows that patrolling monocytes provide important The online version of this article contains supplemental material. signals for proper NK cell recruitment, activation, and thereby Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 antimetastatic function.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1900998 2 PATROLLING MONOCYTES CONTROL METASTASIS VIA NK CELLS

Materials and Methods tumors (7). In this study, we set out to more rigorously define Cells patrolling monocyte-mediated recruitment and activation of NK cells by analyzing the accumulation of these cells in lungs soon B16F10 melanoma and Yac-1 cells were obtained from American Type Culture Collection. B16F10-luciferase expressing cells (B16luc) were after metastatic seeding. To this end, we studied the frequency + + low/2 kindly provided by the Kronenberg laboratory (La Jolla Institute for Im- of total monocytes (CD11b CD115 ), Ly6C monocytes 2 2 munology). Cell lines were tested for being pathogen free and maintained (CD11b+CD115+Ly6Clow/ ), and NK cells (CD3 NK1.1+)byflow in DMEM containing 10% heat-inactivated FBS, 2 mmol/l l-glutamine, cytometry(allAbslistedinTableI)at4and8hafteri.v.in- 1 mmol/l sodium pyruvate, 50 U/ml penicillin, and 50 mg/ml streptomycin. jection of B16F10 melanoma cells in wild-type (C57BL/6J) mice. Mice A significant increase in NK cell frequency (from 9 to 15%) was observed at 8 h after B16F10 melanoma cell seeding and C57BL/6J mice were purchased from The Jackson Laboratory. Congenic low/2 E22/2 mice were previously described (11) and maintained in house. In all coincided with a significant 2-fold expansion of Ly6C experiments, age-matched 8- to 12-wk-old male mice were used. Three monocytes (Fig. 1A, 1B). To confirm activation of recruited hundred thousand B16F10 or B16luc melanoma cells were injected via tail NK cells after B16F10 injection, we analyzed NK cells for the vein into recipient mice. After indicated time points, mice were euthanized expression of activating surface receptors CD226 and NKG2D. by CO2 inhalation, and lungs were perfused with 10 ml of PBS and har- vested for subsequent analyses. For metastatic burden quantification, lu- We observed that the fraction of activated NK cells signifi- ciferase activity was measured 13 d after i.v. injection of melanoma cells cantly increased upon recruitment to metastatic lungs, with using an Ami HT Bioluminescence Imager (Spectral Instruments Imaging) the NKG2D+NK cell population almost tripling at 8 h after after i.v. injection of 1 mg D-Luciferin (Caliper Life Sciences) in 100 ml B16F10 seeding compared with baseline (Fig. 1C–E). In agree- PBS. Twenty-four hours later, mice were euthanized, and lungs were removed ment with our previous work, these data suggest that NK cell Downloaded from and photographed, and metastases were counted. All experiments followed recruitment and activation to lung metastases is facilitated by guidelines of the La Jolla Institute for Immunology Animal Care and Use 2 Committee, and approval for use of rodents was obtained from La Jolla In- Ly6Clow/ monocytes. stitute for Immunology according to criteria outlined in the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health. NK cell recruitment and activation early after metastatic seeding is impaired in the absence of patrolling monocytes Flow cytometry

Our previously published data on patrolling monocytes and NK cell http://www.jimmunol.org/ Lungs were smashed through a 70-mm strainer to disperse cells. RBCs were lysed in RBC Lysis Buffer according to the manufacturer’s protocol function relied on Nr4a1 whole-body knockout and bone marrow (BioLegend). Cells (1 3 106–4 3 106) were resuspended in 100 mlflow chimeric mice, both of which lack specificity for the depletion of staining buffer (1% BSA [w/v] and 0.1% [w/v] sodium azide in PBS). Fcg patrolling monocytes (7). By deleting the Nr4a1 superenhancer receptors were blocked for 15 min, and surface Ags on cells were stained E2 subdomain, we generated C57BL/6J mice, termed “E22/2 for 30 min at 4˚C with directly conjugated fluorescent Abs (flow cytometry mice,” with preserved macrophage function but lacking Ly6Clow/2 Abs listed in Table I). LIVE/DEAD Fixable Dead Cell Stain (Invitrogen) monocytes even after tumor challenge. Thus, the E22/2 mouse was used for analysis of viability, and forward- and side-scatter parameters 2 were used for exclusion of doublets from analysis. Calculations of per- strain is an ideal model for studying Ly6Clow/ monocyte func- centages of CD45+ immune cells were based on live cells as determined by tions in cancer (11). In our present study, we used E22/2 mice forward and side scatter and viability analysis. Cell fluorescence was to assess the impact of Ly6Clow/2 monocyte loss on NK cell– by guest on September 27, 2021 assessed with a LSRII (BD Biosciences) and analyzed with FlowJo software (version 10.5.3). Mean fluorescence intensity was quantified by dependent, antitumor immunity. First, we investigated fre- calculating the geometric mean. quencies of NK cells in different tissues/organs of tumor-free E22/2 mice and compared these frequencies with those from In vitro NK cell degranulation/cytotoxicity assay wild-type (B6) mice (Supplemental Fig. 1A–D). In contrast to The in vitro NK cell degranulation assay has been described elsewhere (12). spleens, in which the percentage of NK cells was reduced, NK 2/2 Briefly, splenocytes were isolated from untreated C57BL/6J or E2 mice cell frequencies were unaltered in bone marrow, blood, and im- 3 6 and 2 10 cells were cultured in 96-well round bottom plates in com- portantly lungs of E22/2 mice, indicating that NK cell migration plete RPMI 1640 at 37˚C with or without 15 3 104 Yac-1 target cells. to lung tissue was unaffected by the absence of patrolling mono- After 1 h, FITC-conjugated anti-CD107a Ab was added and the plates 2 2 were incubated for another hour before addition of GolgiStop. Following cytes in tumor-free mice. E2 / mice subjected to i.v. injection of incubation for 5 h, cells were washed in flow staining buffer and cell B16F10 melanoma cells, in contrast, had reduced frequencies of surface markers were stained. Intracellular IFN-g was labeled with both total lung NK cells and activated NK cells, determined by PE-conjugated anti–IFN-g Ab using the Fixation/Permeabilization Solution Kit (BD Biosciences). measuring CD226 and NKG2D expression, 24 h after injection in comparison with wild-type mice (Fig. 2A–D). Importantly, CD226 Statistical analyses and NKG2D expression levels did not differ between tumor-free 2/2 Data for all experiments were analyzed with Prism software (GraphPad). E2 and B6 mice (Supplemental Fig. 1E, 1F). Hence, NK cell The sample size for each animal study was guided by previous studies in our recruitment and activation are impaired in the absence of patrolling 2/2 laboratory that have used the same E2 transgenic mouse strain. Un- monocytes soon after metastatic seeding in the lung. paired Student t tests and one- or two-way ANOVA analyses were used for comparisons between experimental groups as indicated in figure legends. Accelerated metastasis growth in Ly6Clow/2 monocyte-deficient We used the Brown–Forsythe and Welch ANOVA tests for one-way mice with impaired NK cell recruitment and activation analyses and the Fisher least significant difference ANOVA tests for two-way analyses. No animals were excluded prior to data analysis. The To gauge the long-term impact of impaired NK cell recruitment , p values 0.05 were considered statistically significant. and activation on metastasis growth, we used luciferase-expressing B16 melanoma cells for lung metastasis induction in wild-type and Results E22/2 mice. Metastasis growth was measured by IVIS imaging NK cell recruitment and activation occurs early after and metastasis counts, 13 and 14 d after i.v. injection, respectively. metastatic seeding and coincides with patrolling Luciferase expression and thus tumor volume doubled in lungs of monocyte accumulation patrolling monocyte-deficient E22/2 mice (Fig. 3A, 3B). Similarly, We have previously shown that patrolling (Ly6Clow/2) monocytes the number of metastases in lungs of E22/2 mice grew significantly control lung metastasis by scavenging tumor material from the from ∼20 to 60 nodules per lung (Fig. 3C, 3D). Collectively, these lung vasculature and promoting NK cell recruitment to established results reveal a strong increase in metastatic burden in mice The Journal of Immunology 3

Table I. Detailed information on Abs used

Ag Fluorochrome Clone Company CD45 PE-eFluor 610 30-F11 ThermoFisher Scientific CD3 AF700 17A2 BioLegend CD19 AF700 6D5 BioLegend F4/80 AF700 BM8 BioLegend CD11b BV711 M1/70 BioLegend CD27 BV421 LG.3A10 BioLegend NK1.1 Allophycocyanin/Cy7 PK136 BioLegend NK1.1 PerCP-Cyanine5.5 PK136 ThermoFisher Scientific NKG2D Allophycocyanin CX5 BioLegend NKG2AB6 PerCP-eFluor 710 16a11 ThermoFisher Scientific CD94 PE 18d3 BioLegend CD122 BV650 TM-Beta1 BD Biosciences CD226 PE/Cy7 10E5 BioLegend LY49D PE 4E5 BioLegend CD115 BV421 AFS98 BioLegend Ly6C Allophycocyanin/Cy7 HK1.4 BioLegend IFN-g PE XMG1.2 BioLegend CD107a Allophycocyanin/Cy7 1D4B BioLegend CD27 PE/Cy7 LG.3A10 BioLegend Downloaded from LIVE/DEAD Fixable Aqua Dead Cell Stain Kit ThermoFisher Scientific lacking Ly6Clow/2 monocytes, underscoring the significance of secretion (13, 14). In contrast, coexpression of NKG2A and CD94 patrolling monocytes in NK cell–mediated antitumor immunity. potently impedes NK cell activation and strongly impairs NK

cell-mediated tumor cell killing (15, 16). Therefore, we analyzed the http://www.jimmunol.org/ Absence of patrolling monocytes causes decreased expression expression of Ly49D and NKG2A/CD94 on NK cells 14 d after of stimulatory receptors and increased expression of inhibitory metastatic seeding in E22/2 and wild-type controls. Within the total receptors on NK cells following lung metastasis NK cell pool, the fraction of Ly49D+ NK cells was reduced in Ly49 receptors on NK cells are critical regulators of NK cell cancer melanoma metastases-bearing E22/2 mice compared with wild- immunosurveillance and Ly49D is a potent stimulator of cytokine type tumor-bearing mice, albeit not signiﬁcantly (Fig. 4A, 4B). by guest on September 27, 2021

FIGURE 1. NK cell recruitment and activation occurs early after metastatic seeding and coincides with patrolling monocyte accumulation. (A) Gating strategy used to identify NK cells, monocytes and Ly6Clow/2 monocytes. Cells were previously gated on singlets (using forward scatter width versus forward scatter area and side scatter width versus side scatter area gates). (B) Quantiﬁcation of NK cells, monocytes, and Ly6Clow/2 monocytes by ﬂow cytometry in lungs after indicated time points following B16F10 i.v. injection. Data are represented as means 6 SEM (n $ 8 mice); p values calculated by one-way ANOVA. *p , 0.05, ***p , 0.001. (C) Representative contour plots showing expression of activating receptors CD226 and NKG2D on NK cells after indicated time points following B16F10 i.v. injection. (D and E) Cumulative data from analysis shown in (C). Data are represented as means 6 SD (n = 6 mice); p values calculated by one-way ANOVA. *p , 0.05, **p , 0.01, ***p , 0.001. 4 PATROLLING MONOCYTES CONTROL METASTASIS VIA NK CELLS

FIGURE 2. NK cell recruitment and activation early after metastatic seeding is impaired in the absence of patrolling monocytes. (A) Quantiﬁcation of NK cells by ﬂow cytometry in lungs 24 h after B16F10 i.v. injection into B6 and E22/2 mice or in wild-type B6 mice. Data are represented as means 6 SD (n $ 8 mice); p values calculated by one-way ANOVA. **p , 0.01. (B) Representative contour plots showing expression of activating receptors CD226 and NKG2D on NK cells in lungs 24 h after B16F10 i.v. injection into B6 and E22/2 mice or in wild-type B6 mice. (C and D) Cu- mulative data from analysis shown in (B). Data are represented as means 6 SD (n $ 8 mice); p values

calculated by one-way ANOVA. *p , 0.05. Downloaded from http://www.jimmunol.org/

Expression of NKG2A/CD94, however, was significantly up- similar levels of Ly49D in metastases-bearing mice, expression of regulated on NK cells in mice lacking patrolling monocytes the inhibitory receptor NKG2A/CD94 was significantly upregulated (Fig. 4C). In addition to the total NK cell pool, we sought to in E22/2 mice (Fig. 4F, 4G). Nevertheless, the fraction of educated determine the expression of stimulatory Ly49D and inhibi- NK cells expressing Ly49D significantly decreased by almost 50% tory NKG2A receptors on “terminally differentiated” NK cells. in metastasis-bearing E22/2 mice (Supplemental Fig. 2B). These During NK cell maturation, “uneducated” CD27+ NK cells become changes in stimulatory and inhibitory receptor expression on late

“educated” by gaining reactivity and cytotoxicity. Further, NK cells stage metastasis aligns with our observation of reduced NK cell by guest on September 27, 2021 downregulate CD27 and upregulate CD11b expression, resulting recruitment and activation during early metastatic seeding in the in their terminal differentiation to CD11b+CD272 NK cells (17). absence of patrolling monocytes (Fig. 1). This further establishes a Neither the frequency of educated CD11b+CD27+ or terminally dif- role for patrolling monocytes in NK cell activation (13), whereas ferentiated CD11b+CD272 NK cells among all NK cells changed in they appear to be obsolete for NK cell maturation. Further evi- wild-type versus E22/2 mice (Fig. 4D, 4E, Supplemental Fig. 2A), dence for the latter notion stems from the observation of similar suggesting that patrolling monocytes are not necessary for NK cell expression levels of CD122, a marker for NK cell maturation (18), maturation. Whereas terminally differentiated NK cells expressed in B6 and E22/2 mice (Supplemental Fig. 2C, 2D).

FIGURE 3. Accelerated metastasis growth in Ly6Clow/2 monocyte-deficient mice with impaired NK cell recruitment and activation. (A) In vivo luciferase detection in B6 control and E22/2 mice 13 d after i.v. injection of 5 3 105 B16F10 melanoma cells expressing luciferase. (B) Luciferase quantification of experiment shown in (A). Data are represented as means 6 SD (n = 6 mice); p value calculated by unpaired t test. (C) Macroscopic im- ages of lungs from B6 control and E22/2 mice 14 d after i.v. injection of melanoma cells. (D) Quantifi- cation of metastases by tumor nodule count. Data are represented as means 6 SD (n = 6 mice); p value calculated by unpaired t test. **p , 0.01. The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/

FIGURE 4. Absence of patrolling monocytes causes decreased expression of stimulatory receptors and increased expression of inhibitory receptors by guest on September 27, 2021 on NK cells following lung metastasis. (A) Quantiﬁcation of NK cells per lung in B6 and E22/2 mice, with and without tumors. Data are represented as means 6 SD (n = 5–6 mice); p values calculated by one-way ANOVA. ***p , 0.001, ****p , 0.0001. (B) Representative contour plots showing expression of stimulatory receptor Ly49D and inhibitory receptor NKG2A/CD94 on NK cells in lungs 14 d after B16F10 i.v. injection into B6 and E22/2 mice. (C and D) Cumulative data from analysis shown in (B). Data are represented as means 6 SD (n = 6 mice); p values calculated by two-way ANOVA. (E) Representative contour plots showing surface expression of CD11b and CD27 on NK cells in lungs 14 d after B16F10 i.v. injection into B6 and E22/2 mice. (F) Cumulative data from analysis shown in (E). Data are represented as means 6 SD (n = 6 mice). (G and H) Expression of stimulatory receptor Ly49D and inhibitory receptor NKG2D/CD94 on terminally differentiated CD11b+CD272 NK cells in lungs 14 d after B16F10 i.v. injection into B6 and E22/2 mice. Data are represented as mean 6 SD (n = 6 mice); p values calculated by two-way ANOVA. **p , 0.01, ***p , 0.001.

Impaired NK cell degranulation in the absence of in the prevention of lung metastasis progression. In wild-type mice, patrolling monocytes NK cell recruitment to lung metastases occurred as early as 8 h after i.v. injection of B16F10 melanoma cells and was impaired in Finally, to confirm impaired NK cell function in the absence of 2/2 Ly6Clow/2 monocytes, we performed an in vitro, NK cell degranula- E2 mice, as indicated by reduced expression of NK cell sur- tion assay against Yac-1 target cells with splenocytes from either face markers CD226 and NKG2D. Long-term follow-up on these E22/2 mice or wild-type mice (Fig. 5). CD107a lines the inner animals revealed decreased expression of Ly49D and increased expression of NKG2A/CD94 on NK cells, as well as an increase in surface of cytotoxic granules and becomes temporarily acces- 2/2 sible for extracellular Abs during degranulation. Hence, positive metastatic burden. Collectively, our results using E2 mice staining for CD107a indicates release of granule content and support an observation by multiple groups that patrolling mono- NK cell cytotoxicity. Confirming our in vivo results, splenocytes cytes are crucial for effective antitumor functions by NK cells. from E22/2 mice insufficiently supported NK cell degranulation Defining the relationship between nonclassical monocytes and against target cells (Fig. 5A, 5B). Finally, in comparison with other immune cells in mice has been challenging (11). Earlier splenocytes from wild-type mice, there was a 30% reduction in studies, including our own, have used different means to ablate degranulation in the absence of Ly6Clow/2 monocytes. The data patrolling monocytes in mice. These efforts consisted of depletion presented in this article show that patrolling monocytes are nec- experiments using anti-CD115 Abs, liposomal clodronate, or the essary for proper NK cell recruitment, activation, cytotoxicity, and use of Nr4a1 whole-body knockout, myeloid conditional knock- thus NK cell–mediated antitumor immunity. out, and bone marrow chimeric mice. However, common to all these approaches is an additional effect on the number and/or Discussion function of other myeloid cells, including macrophages. For ex- Using a recently developed mouse model, we show that patrolling ample, CD115 is expressed not only on monocytes but also on monocytes play a critical role in NK cell recruitment and activation macrophages, and unsurprisingly, Ab targeting of CD115 depletes 6 PATROLLING MONOCYTES CONTROL METASTASIS VIA NK CELLS

FIGURE 5. Impaired NK cell degranulation in the absence of patrolling monocytes. (A) Representative contour plots of CD107a staining on NK cells in in vitro degranulation assay against Yac-1 target cells using splenocytes from B6 or E22/2 mice. (B) Cumulative data from analysis shown in (A). (C) Total number of IFNg+ NK cells among CD45+ cells from experimental setup described in (A). Data are represented as means 6 SD (n = 4 mice); p values calculated by two-way ANOVA. *p , 0.05, **p , 0.01, ****p , 0.0001. both monocytes and macrophages (19). Clodronate is also widely Disclosures Downloaded from used to ablate macrophage populations in various tissues and was The authors have no financial conflicts of interest. shown to reduce liver and lung metastasis in a genetic mouse model of pancreatic cancer (20). Ablating Nr4a1 in the myeloid compartment, in contrast, diminishes patrolling monocyte References 1. Soderquest, K., N. Powell, C. Luci, N. van Rooijen, A. Hidalgo, F. Geissmann, numbers and increases macrophage numbers polarized toward an T. Walzer, G. M. Lord, and A. Martıń-Fontecha. 2011. Monocytes control natural

M1 phenotype (10). Deletion of an Nr4a1 enhancer subdomain killer cell differentiation to effector phenotypes. Blood 117: 4511–4518. http://www.jimmunol.org/ in E22/2 mice overcomes these limitations and preserves Nr4a1 2. Welte, S., S. Kuttruff, I. Waldhauer, and A. Steinle. 2006. Mutual activation of natural killer cells and monocytes mediated by NKp80-AICL interaction. Nat. gene expression in macrophages, making these animals an ideal Immunol. 7: 1334–1342. model to specifically study patrolling monocytes (11). Thus, 3. Peng, L.-S., J.-Y. Zhang, Y.-S. Teng, Y.-L. Zhao, T.-T. Wang, F.-Y. Mao, Y.-P. Lv, E22/2 mice allowed us to dissect the involvement of patrolling P. Cheng, W.-H. Li, N. Chen, et al. 2017. Tumor-associated monocytes/ macrophages impair NK-cell function via TGFb1 in human gastric cancer. monocytes on NK cell antitumor function in melanoma metas- Cancer Immunol. Res. 5: 248–256. tasis without implicating other myeloid cells. 4. Mao, Y., D. Sarhan, A. Steven, B. Seliger, R. Kiessling, and A. Lundqvist. 2014. Based on the results presented in this study, future projects will Inhibition of tumor-derived prostaglandin-e2 blocks the induction of myeloid- derived suppressor cells and recovers natural killer cell activity. Clin. Cancer need to focus on the monocyte-derived signals that induce NK cell Res. 20: 4096–4106. responses. One candidate signal is IL-15, a stimulatory cytokine 5. Kloss, M., P. Decker, K. M. Baltz, T. Baessler, G. Jung, H.-G. Rammensee, by guest on September 27, 2021 A. Steinle, M. Krusch, and H. R. Salih. 2008. Interaction of monocytes with NK found on monocytes and dendritic cells that prompts NK cell and cells upon toll-like receptor-induced expression of the NKG2D ligand MICA. J. T cell proliferation and differentiation (21). Supporting this notion Immunol. 181: 6711–6719. is the finding that IL-15 expression by activated monocytes is 6. Campbell, A. R., M. C. Duggan, L. P. Suarez-Kelly, N. Bhave, K. S. Opheim, E. L. McMichael, P. Trikha, R. Parihar, E. Luedke, A. Lewis, et al. 2017. MICA- required for optimal IFN-g secretion by NK cells (22). Moreover, expressing monocytes enhance natural killer cell Fc receptor-mediated antitumor in a mouse model of melanoma lung metastasis formation, IL-15 functions. Cancer Immunol. Res. 5: 778–789. was found to have a protective effect against metastases (9). 7. Hanna, R. N., C. Cekic, D. Sag, R. Tacke, G. D. Thomas, H. Nowyhed, E. Herrley, N. Rasquinha, S. McArdle, R. Wu, et al. 2015. Patrolling monocytes Conclusive evidence that IL-15 is produced by patrolling mono- control tumor metastasis to the lung. Science 350: 985–990. cytes is missing in this study. In contrast, we have previously 8. Plebanek, M. P., N. L. Angeloni, E. Vinokour, J. Li, A. Henkin, D. Martinez-Marin, shown that patrolling monocytes produce chemokines that may be S. Filleur, R. Bhowmick, J. Henkin, S. D. Miller, et al. 2017. Pre-metastatic cancer exosomes induce immune surveillance by patrolling monocytes at the metastatic involved in NK cell recruitment to lung metastases (7). Quanti- niche. Nat. Commun. 8: 1319. tative PCR analysis of sorted classical and patrolling monocytes 9. Kubo, H., S. Mensurado, N. Gonçalves-Sousa, K. Serre, and B. Silva-Santos. revealed upregulation of CCL3, CCL4, and CCL5 in patrolling 2017. Primary tumors limit metastasis formation through induction of IL15- mediated cross-talk between patrolling monocytes and NK cells. Cancer monocytes compared with classical monocytes. All three of Immunol. Res. 5: 812–820. these chemokines can induce human NK cell migration and 10. Hanna, R. N., I. Shaked, H. G. Hubbeling, J. A. Punt, R. Wu, E. Herrley, C. Zaugg, H. Pei, F. Geissmann, K. Ley, and C. C. Hedrick. 2012. NR4A1 enhance cytolysis in vitro (23). Furthermore, CCL3 and CCL5 (Nur77) deletion polarizes macrophages toward an inflammatory phenotype and have been shown to play critical roles in NK cell antitumor increases atherosclerosis. Circ. Res. 110: 416–427. immunity in vivo (24, 25). Future studies need to dissect the 11. Thomas, G. D., R. N. Hanna, N. T. Vasudevan, A. A. Hamers, C. E. Romanoski, S. McArdle, K. D. Ross, A. Blatchley, D. Yoakum, contribution of each of these chemokines to patrolling monocyte- B. A. Hamilton, et al. 2016. Deleting an Nr4a1 super-enhancer subdomain mediated NK cell recruitment to lung metastases. NK cells have ablates Ly6Clow monocytes while preserving macrophage gene function. recently gained attention as possible targets for cancer im- Immunity 45: 975–987. 12. Speak, A. O., D. Te Vruchte, L. C. Davis, A. J. Morgan, D. A. Smith, munotherapy (26). Therefore, an in-depth understanding of the N. M. Yanjanin, L. Simmons, R. Hartung, H. Runz, E. Mengel, et al. 2014. mechanisms by which patrolling monocytes control NK cell Altered distribution and function of natural killer cells in murine and human Niemann-Pick disease type C1. Blood 123: 51–60. function will be crucial to leverage these processes for anti- 13. Tu, M. M., A. B. Mahmoud, A. Wight, A. Mottashed, S. Be´langer, M. M. Rahim, cancer immunotherapy. E. Abou-Samra, and A. P. Makrigiannis. 2014. Ly49 family receptors are required for cancer immunosurveillance mediated by natural killer cells. Cancer Res. 74: 3684–3694. Acknowledgments 14. Ortaldo, J. R., and H. A. Young. 2003. Expression of IFN-g upon triggering of We thank Deborah Yoakum for assistance with our mouse colony, Cheryl activating Ly49D NK receptors in vitro and in vivo: costimulation with IL-12 or Kim and the LJI Flow Cytometry Core, and Dr. Daniel Araujo for critical IL-18 overrides inhibitory receptors. [Published erratum appears in 2003 J. Immunol. 170: 5799.] J. Immunol. 170: 1763–1769. reading of the manuscript. The visual abstract was created using BioRender 15. Andre´, P., C. Denis, C. Soulas, C. Bourbon-Caillet, J. Lopez, T. Arnoux, (https://biorender.com/). M. Ble´ry, C. Bonnafous, L. Gauthier, A. Morel, et al. 2018. Anti-NKG2A mAb is The Journal of Immunology 7

a checkpoint inhibitor that promotes anti-tumor immunity by unleashing both T 21. Waldmann, T. A. 2006. The biology of interleukin-2 and interleukin-15: impli- and NK cells. Cell 175: 1731–1743.e13. cations for cancer therapy and vaccine design. Nat. Rev. Immunol. 6: 595–601. 16. Kamiya, T., S. V. Seow, D. Wong, M. Robinson, and D. Campana. 2019. 22. Carson, W. E., M. E. Ross, R. A. Baiocchi, M. J. Marien, N. Boiani, Blocking expression of inhibitory receptor NKG2A overcomes tumor resistance K. Grabstein, and M. A. Caligiuri. 1995. Endogenous production of interleukin to NK cells. J. Clin. Invest. 129: 2094–2106. 15 by activated human monocytes is critical for optimal production of interferon- 17. Boudreau, J. E., and K. C. Hsu. 2018. Natural killer cell education and the gamma by natural killer cells in vitro. J. Clin. Invest. 96: 2578–2582. response to infection and cancer therapy: stay tuned. Trends Immunol. 39: 23. Taub, D. D., T. J. Sayers, C. R. Carter, and J. R. Ortaldo. 1995. Alpha and beta 222–239. chemokines induce NK cell migration and enhance NK-mediated cytolysis. J. 18. Yu, J., A. G. Freud, and M. A. Caligiuri. 2013. Location and cellular stages of Immunol. 155: 3877–3888. natural killer cell development. Trends Immunol. 34: 573–582. 24. Allen, F., I. D. Bobanga, P. Rauhe, D. Barkauskas, N. Teich, C. Tong, J. Myers, 19. MacDonald, K. P., J. S. Palmer, S. Cronau, E. Seppanen, S. Olver, N. C. Raffelt, and A. Y. Huang. 2017. CCL3 augments tumor rejection and enhances CD8+ R. Kuns, A. R. Pettit, A. Clouston, B. Wainwright, et al. 2010. An antibody T cell infiltration through NK and CD103+ dendritic cell recruitment via IFNg. against the colony-stimulating factor 1 receptor depletes the resident subset of Oncoimmunology 7: e1393598. monocytes and tissue- and tumor-associated macrophages but does not inhibit 25. Mgrditchian, T., T. Arakelian, J. Paggetti, M. Z. Noman, E. Viry, E. Moussay, inflammation. Blood 116: 3955–3963. K. Van Moer, S. Kreis, C. Guerin, S. Buart, et al. 2017. Targeting autophagy 20. Griesmann, H., C. Drexel, N. Milosevic, B. Sipos, J. Rosendahl, inhibits melanoma growth by enhancing NK cells infiltration in a CCL5- T. M. Gress, and P. Michl. 2017. Pharmacological macrophage inhibition dependent manner. Proc. Natl. Acad. Sci. USA 114: E9271–E9279. decreases metastasis formation in a genetic model of pancreatic cancer. 26. Morvan, M. G., and L. L. Lanier. 2016. NK cells and cancer: you can teach Gut 66: 1278–1285. innate cells new tricks. Nat. Rev. Cancer 16: 7–19. Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021 Supplemental Figure 1 A) B) C)

4 20 20 cells (%) cells (%) + cells (%) +

+ 3 15 15

2 10 10

1 5 5

0 0 0 BM NK cells / CD45

B6 E2 B6 E2 Lung NK cells / CD45 B6 E2 Blood NK cells / CD45

D) E) F)

*** 5 8 4 cells (%) + 4 6 3 NK cells / NK cells / +

3 + cells (%) 4 cells (%) 2 + 2 + 2 1 CD45 1 CD45 Lung CD226 0 0 Lung NKG2D 0 B6 E2 B6 E2 B6 E2 Spleen NK cells / CD45

Supplemental Figure 1. Absence of patrolling monocytes does not alter frequencies of NK cells and activating receptor expression on NK cells in tumor-free lungs. Cumulative data from ﬂow cytometric analysis of NK cell frequencies in bone marrow (A), blood (B), lung (C) and spleen (D) of tumor-free B6 and E2-/- mice. Data are represented as means ± SD (n= 5-8 mice); p value calculated by unpaired t-test; ***p < 0.001. (E) Cumulative data from ﬂow cytometric analysis of lung NK cells expressing CD226 or NKG2D (F) in tumor-free B6 and E2-/- mice. Data are represented as means ± SD (n = 3-4 mice). Supplemental Figure 2 A) B)

40 10 * B6 NK cells (%) NK cells (%)

8 E2

30 + 6

20 CD27 +

NK cells / 4 + 10 2 CD27 / CD11b + 0 + 0 tumor-free B16 tumor-free B16 Ly49D CD11b C) D)

80 50 B6 70 E2 40 NK cells (%) -

60 NK cells / +

CD27 30 +

NK cells / (%) 50 + CD122

40 CD11b 20

CD122 tumor-free B16 tumor-free B16

Supplemental Figure 2. Absence of patrolling monocytes does not alter NK cell maturation in tumor-free lungs or lungs with metastases. (A) Cumulative data from ﬂow cytometric analysis of terminally differentiated CD11b+CD27- NK cells within the entire NK cell pool or of the frequency of terminally differentiated NK cells expressing Ly49D (B). Analysis performed on leukocytes isolated from lungs 14 days after B16F10 IV injection into B6 and E2-/- mice. Data are represented as means ± SD (n = 6 mice); p value calculated by two-way ANOVA. *p < 0.05. (C,D) Cumulative data from ﬂow cytometric analysis of NK cell frequencies expressing CD122 in the entire NK cell pool (C) or within terminally differentiated CD11b+CD27- NK cells (D) in lungs 14 days after B16F10 IV injection into B6 and E2-/- mice. Data are represented as means ± SD (n = 6 mice).