Tumor-Associated Monocytes/Macrophages Impair …...Research Article Tumor-Associated Monocytes/Macrophages Impair NK-Cell Function via TGFb1 in Human Gastric Cancer Liu-sheng Peng1,

Research Article

Tumor-Associated Monocytes/MacrophagesImpair NK-Cell Function via TGFb1 in HumanGastric CancerLiu-sheng Peng1, Jin-yu Zhang1, Yong-sheng Teng1, Yong-liang Zhao2, Ting-ting Wang1,Fang-yuan Mao1, Yi-pin Lv1, Ping Cheng1,Wen-hua Li1, Na Chen1, Mubing Duan3,Weisan Chen3, Gang Guo1, Quan-ming Zou1, and Yuan Zhuang1

Abstract

Natural killer (NK) cells are a major component of the hostantitumor immune response in human cancer. However, thenature, functional regulation, and clinical relevance of NKcells in gastric cancer remain largely unknown. In this study,we showed that the percentages of NK cells in tumors weresignificantly decreased, and low percentages of tumor-infil-trating NK cells were positively correlated with poor survivaland disease progression. Although the expression of activatingand inhibitory receptors on NK cells was shown to be notdifferent between tumor and nontumor tissues, NK cells intumors had impaired effector functions, characterized bydecreased IFNg , TNFa, and Ki-67 expression. We found that

tumor-infiltrating monocytes/macrophages were physicallyclose to NK cells, and their percentages negatively correlatedwith IFNgþ and TNFaþ NK-cell percentages. Ex vivo studyshowed that isolated tumor-associated monocytes/macro-phages could impair NK-cell expression of IFNg , TNFa, andKi-67. Blockade of TGFb1 attenuated such monocytes/macro-phages-mediated impairment of NK-cell function. Our datasuggest that human NK-cell function was impaired by tumor-associated monocytes/macrophages, and that restoring NK-cell function may be an important therapeutic strategy toprevent tumor immune escape in gastric cancer. Cancer Immunol Res;5(3); 248–56. �2017 AACR.

IntroductionGastric cancer is one of themost commonhumanmalignancies

and accounts for approximately 900,000 total cases and 700,000deaths globally per annum (1). The prognosis of gastric cancerpatients with early-stage disease is generally good, but the controlof advanced-stage gastric cancer remains challenging. It is gener-ally accepted that the development and progression of gastriccancer is influenced by tumor and host immune system cross-talk,with many studies supporting a central role for adaptive immu-nity in determining the clinical outcomes of gastric cancer patients(2–4). However, less is known about the role of innate immunity

and innate immune cell cross-regulation during gastric cancerprogression.

Natural killer (NK) cells are a group of innate lymphoid cellsthat confer antitumor as well as antiviral host protection. Thisprocess is mediated by cell surface recognition of altered orcancerous cells compared with normal host cells, through theirdifferential engagement of NK-inhibitory or -activating receptors(5). Once activated, NK cells secrete cytotoxic enzymes andcytokines to kill their target cells. However, many studies focusingon the role of NK cells in cancer have shown that the function ofNK cells is impaired within the tumor microenvironment (6–8).In gastric cancer, impairment of the cytotoxic function of circu-lating NK cells has also been reported to be associated withadvanced disease progression (9), suggesting that the progressionof gastric cancer may be linked to defects in antitumor NK-celleffector function. However, the phenotype, functional regulation,and clinical relevance of NK cells within the human gastric cancermicroenvironment remain largely unknown.

Monocytes/macrophages are another important group ofinnate immune cells that reside within the tumor milieu andmodulate local antitumor immunity (10). Monocytes/macro-phages can be reprogrammed by tumor-derived signals to expressinhibitory molecules and/or release soluble inhibitory factors,which in turn suppress cytotoxic T-cell activation and effectorfunction (11, 12). Monocytes/macrophages isolated from tumorscan also induce NK-cell dysfunction during hepatocellular carci-noma (13), suggesting a role for tumor-associated monocytes/macrophages in impairing antitumor NK-cell function duringtumor progression. However, the presence of interactionsbetween tumor-associatedmonocytes/macrophages andNKcells,

1National Engineering Research Center of Immunological Products, Departmentof Microbiology and Biochemical Pharmacy, College of Pharmacy, Third MilitaryMedical University, Chongqing, PR China. 2Department of General Surgery andCenter of Minimal Invasive Gastrointestinal Surgery, Southwest Hospital, ThirdMilitary Medical University, Chongqing, PR China. 3La Trobe Institute for Molec-ular Science, La Trobe University, Bundoora, Victoria, Australia.

L.-s. Peng and J.-y. Zhang contributed equally to this article.

Note: Supplementary data for this article are available at Cancer ImmunologyResearch Online (http://cancerimmunolres.aacrjournals.org/).

Corresponding Authors: Yuan Zhuang, Third Military Medical University, No.30Gaotanyan stree, Chongqing 400038, PR China. Phone: 086-023-68752377;Fax: 086-023- 68752377; E-mail: [email protected]; or Quan-mingZou, [email protected]

doi: 10.1158/2326-6066.CIR-16-0152

�2017 American Association for Cancer Research.

CancerImmunologyResearch

Cancer Immunol Res; 5(3) March 2017248

on April 24, 2020. © 2017 American Association for Cancer Research. cancerimmunolres.aacrjournals.org Downloaded from

Published OnlineFirst February 1, 2017; DOI: 10.1158/2326-6066.CIR-16-0152

http://cancerimmunolres.aacrjournals.org/

as well as their precise mechanisms of communication in gastriccancer, remains unknown.

In the present study, we showed that the percentages of tumor-infiltrating NK cells were significantly decreased, and low percen-tages of these cells were positively correlated with tumor progres-sion and poor overall survival of gastric cancer patients. Mean-while, tumor-infiltrating NK cells displayed functionalimpairment, and such impaired functions were mediated bytumor-associated monocytes/macrophages-derived TGFb1 pro-duction. Restoring NK-cell function may be an important thera-peutic strategy to prevent tumor immune escape in gastric cancer.

Materials and MethodsPatients and tissue samples

Fresh peripheral blood, autologous nontumor, and tumorgastric tissues were obtained from gastric cancer patients duringsurgery at the Southwest Hospital of the Third Military MedicalUniversity. None of the patients had received chemotherapy orradiotherapy before sampling, and individuals with autoimmunedisease, infectious diseases, or multi-primary cancer were exclud-ed. The clinical stages of tumors were determined according to thetumor–node–metastasis (TNM) classification systemby the Inter-national Union against Cancer (Edition 7). The study wasapproved by the Ethics Committee of the Third Military MedicalUniversity, andpriorwritten informed consentwasobtained fromeach patient. The clinical characteristics of gastric cancer patientswere presented in Supplementary Table S1.

Cell isolationFresh nontumor and tumor tissues were used for the isola-

tion of tissue-infiltrating lymphocytes as previously described(14). In brief, paired nontumor and tumor tissues were cutinto small pieces, then suspended in RPMI 1640 containing 1mg/mL collagenase IV (Sigma-Aldrich) and 10 mg/mL DNase I(Roche), and gently mechanically dissociated using the MACSDissociator (Miltenyi Biotec). Dissociated cell suspensionswere further incubated for 1 hour at 37�C under continuousrotation and then filtered through 70 mm cell strainers to obtainsingle-cell suspensions. Cell suspensions were then used forflow cytometry analysis.

Flow cytometric analysisCell suspensions were stained with appropriate surface anti-

bodies, and then fixed, permeabilized for 20 minutes usingCytofix/Cytoperm (BD Biosciences). Permeabilized cells weresubsequently stained with antibodies against intracellular mole-cules. For intracellular staining of IFNg and TNFa, cells werestimulated for 4 hours with phorbol myristate acetate (PMA, 50ng/mL) and ionomycin (1 mg/mL) in the presence of Golgistop(BD Biosciences) before staining. The fluorochrome-labeled anti-bodies are listed in Supplementary Table S2.

ImmunohistochemistryParaffin-embedded samples were cut into 5-mm sections. After

being deparaffinized and hydrated, sections immersed in Tris-EDTA buffer (pH 9.0) were subjected to heat-induced antigenretrieval in a microwave oven and then treated with 3% hydrogenperoxide for endogenous peroxide quenching. Following incu-bation with mouse monoclonal anti-human CD57 (NK-1) anti-body (Zhongshan Golden Bridge Biotech.) or TGFb1 antibody

(Abcam) overnight at 4�C, slides were incubatedwith horseradishperoxidase–conjugated secondary anti-mouse antibody (Zhong-shan Golden Bridge Biotech.) and then developed with 3,3'-diaminobenzidine tetrahydrochloride. For colocalization analy-sis, after CD57 staining, slides were incubated with a mouse anti-human CD68 antibody (DAKO) and stained using EnVision G2System/AP Rabbit/Mouse (Permanent Red; DAKO). Slides wereexamined using a microscope (Nikon Eclipse 80i; Nikon).

In vitro coculture of tissue-infiltratingmonocytes/macrophagesand NK cells

Cell suspensions derived from paired nontumor and tumortissues were stained by PE-Cy7 anti-human CD45 and PerCP-Cy5.5 anti-human CD14 antibodies, and CD45þCD14þ mono-cytes/macrophages were then purified by fluorescence-activatedcell sorting (FACSAria II; BD Biosciences). Peripheral blood NKcells were purified using the humanNK-cell enrichment Kit (StemCell Technologies, Inc.). The purity of monocytes/macrophagesand NK cells was >90%. Purified tissue-associated monocytes/macrophages were cocultured with NK cells at a 1:5 ratio in RPIM1640 medium supplemented with 10% FCS and 100 U/mL IL2(Peprotech) for 5 to 7 days. In some experiments, a neutralizingantibody against TGFb1 (10 mg/mL; Abcam) or an isotype controlwas added into the coculture system. The proliferation andcytokine expression levels of NK cells were assessed by flowcytometry. In some cases, purified CD45þCD14þ monocytes/macrophages were cultured for 48 hours in vitro, and the super-natants were then collected for TGFb1 quantification by ELISA(Dakewe).

Statistical analysisAll results are presented asmean� SEM, and statistical analysis

was performed with the Prism 5.0 Software. The correlationanalysis between groups was determined by the Spearman cor-relation test. Differences between groups were evaluated by thetwo-tailed Student t test. When variance was detected, Mann–Whitney U tests were used to analyze the difference betweengroups. Cumulative survival time was measured in months andcalculated by the Kaplan–Meiermethod, and the log-rank test wasapplied to compare between groups. Multivariate analysis ofprognostic factors for overall patient survivalwas performedusingthe Cox proportional hazards model. P < 0.05 was considered asstatistically significant.

ResultsTumor-infiltrating NK cells are decreased in human gastriccancer

To evaluate the potential roles of NK cells in human gastriccancer, we first analyzed their proportion in peripheral blood,nontumor, and tumor tissues (Fig. 1). CD3–CD56þ NK cells andCD3þ T cells were characterized from the lymphocyte gate asdefined by a forward scatter (FSC) and side scatter (SSC) dot-plot.In comparison with peripheral blood, the percentages of tissue-infiltrating CD3–CD56þ NK cells and CD3þ T cells were signif-icantly decreased. Within patient's disease tissues, CD3þ T-cellpercentages were significantly higher in tumors compared withnontumor tissues, whereasCD3–CD56þNK-cell percentagesweresignificantly lower in tumors compared with nontumor tissues.Similar observations were made when analyzing the absolutenumber of NK cells and T cells per million cells in each sample

NK-Cell Functional Impairment in Gastric Cancer

www.aacrjournals.org Cancer Immunol Res; 5(3) March 2017 249




(Supplementary Fig. S1). Altogether, these results showed that thepercentages and numbers of tumor-infiltrating NK cells weredecreased in human gastric cancer.

The function of tumor-infiltrating NK cells is impaired inhuman gastric cancer

Basedon thesefindings,we further characterized thephenotypeof NK cells in the peripheral blood, nontumor, and tumor tissues.The expression of activating and inhibitory receptors on NK cellswas first determined by flow cytometry. Compared with periph-eral blood, we found that NK cells in nontumor and tumor tissuesdisplayed significantly lower percentages and mean fluorescenceintensity (MFI) of both activating receptors such asCD16,NKp30,NKp46, NKG2D, DNAM-1, 2B4, and CD94, and inhibitoryreceptors such asCD158a/h andCD158b, but a higher percentageandMFI of the activating receptorNKp44.However, expression ofthese receptors between nontumor and tumor-infiltrating NKcells was not significantly different (Fig. 2; Supplementary Fig.S2), suggesting that NK cells do not acquire significant alterationsin activating or inhibitory receptor expression after localizationwithin the tumor microenvironment. Further analysis showedthat the percentages of CD56brightCD16– and CD56dimCD16þ

NK-cell subsets in tissueswere significantly lower than those in the

peripheral blood. However, tumor and non–tumor-infiltratingNK-cell subsets were not significantly different (SupplementaryFig. S3). Next, we investigated the functional status of tumor-infiltratingNK cells (Fig. 3).NK cells in tumors expressed less IFNgand TNFa than those in nontumor tissues. In addition, theexpression of Ki-67 (a marker of cell proliferation) in tumor-infiltrating NK cells was also significantly decreased. However,expression of cytotoxic molecules granzyme B and perforinbetween nontumor and tumor-infiltrating NK cells was not sig-nificantly different, suggesting a selective impairment of cytokineproduction and proliferation capacities in tumor-infiltrating NKcells. Taken together, these results demonstrated that the functionof tumor-infiltrating NK cells was impaired in human gastriccancer.

NK-cell functions are impaired by tumor-associatedmonocytes/macrophages

Because our previous findings showed that monocytes/macrophages from gastric cancer tumors could modulate T-celleffector function (4), we postulated that tumor-associatedmonocytes/macrophages were also involved in the functionalregulation of NK cells. We thus investigated the relationshipbetween monocytes/macrophages and NK cells within tumors.

Figure 1.

Percentages of NK cells in theperipheral blood, nontumor, andtumor tissues of 65 gastric cancerpatients. A, Representative flowcytometry analysis of CD3�CD56þ NKcells and CD3þ T cells in the peripheralblood, nontumor, and tumor tissuesfrom gastric cancer patients. Wholeperipheral blood–, nontumor-, andtumor tissue–derived cell suspensionswere stained with PE-Cy7–conjugatedantibody (Ab) to human CD56 andAPC-conjugated Ab to human CD3.CD3–CD56þ NK cells and CD3þ T-cellpercentages analyzed were from thelymphocyte gate as defined by a FSCand SSC dot-plot. B, Statisticalanalysis of CD3–CD56þ NK-cell andCD3þ T-cell percentages in theperipheral blood, nontumor, andtumor tissue of 65 gastric cancerpatients. Data analyzed by theStudent t test and shown asmean � SEM. �� , P < 0.001.

Peng et al.

Cancer Immunol Res; 5(3) March 2017 Cancer Immunology Research250




Double-immunohistochemical staining showed that mostCD57þ cells in tumors were located within close proximity toCD68þ macrophages (Fig. 4A). In addition, monocytes/macro-phages' infiltration negatively correlated with the percentagesof IFNgþ and TNFaþ NK cells within tumors (Fig. 4B), suggest-ing that tumor-associated monocytes/macrophages may beassociated with the functional impairment of tumor-infiltratingNK cells in gastric cancer. Phenotypic analysis revealed thatmonocytes/macrophages from both nontumor and tumor tis-sues highly expressed HLA-DR, but expression of HLA-DR onsuch cells that had infiltrated into tumor tissues was lower thanthat in nontumor tissues (Supplementary Fig. S4), suggestingthat tumor-associated monocytes/macrophages possessed aless classically activated phenotype. We then investigated theeffects of tumor-associated monocytes/macrophages on regu-lating NK-cell functions ex vivo. Monocytes/macrophages weresorted from nontumor or tumor tissues and then coculturedwith purified allogeneic peripheral blood NK cells. We foundthat the production of IFNg and TNFa in NK cells coculturedwith tumor-associated monocytes/macrophages was signifi-cantly lower than those cocultured with monocytes/macro-phages from nontumor tissues (Fig. 4C and D). Moreover,

when compared with non–tumor-associated counterparts,tumor-associated monocytes/macrophages also significantlysuppressed Ki-67 expression in NK cells (Fig. 4E and F).Altogether, these results indicated that tumor-associated mono-cytes/macrophages impaired NK-cell effector function by sup-pressing their proliferation and cytokine production capacities.

Monocytes/macrophages-derived TGFb1 impairs NK-cellfunctions

We next sought to elucidate the molecular basis of monocytes/macrophages-mediated NK-cell functional impairment in gastriccancer tumors. Studies have previously reported that inhibitorymolecules including CD48 and PD-L1/PD-L2 are involved in thisprocess (12, 13). However, CD48 or PD-L1/PD-L2 expressionbetween tumor and non–tumor-associated monocytes/macro-phages was not significantly different (Supplementary Fig. S5).As we previously found significantly increased TGFb1 concentra-tions in tumor tissues, we questioned whether tumor-associatedmonocytes/macrophages produced TGFb1, a known inhibitor ofNK-cell function. TGFb1 was detectable in the gastric cancertumor microenvironment containing CD68þ macrophages, asanalyzed from serial paraffin sections of primary human gastric

Figure 2.

Phenotypic features ofNKcells from theperipheral blood, nontumor, and tumor tissues of gastric cancer patients. Peripheral blood, nontumor, and tumor-derived cellsuspensions were stained with Abs to CD3, CD56, CD16, NKp30, NKp44, NKp46, NKG2D, DNAM-1, 2B4, CD94, NKG2A, CD158a/h, CD158b, and CD158e1. NKcells gated as CD3–CD56þ events and the expression of CD16, NKp30, NKp44, NKp46, NKG2D, DNAM-1, 2B4, CD94, NKG2A, CD158a/h, CD158b, and CD158e1 thenanalyzed. Symbols represent individual values from 7 to 14 gastric cancer patients analyzed individually. � , P < 0.05; �� , P < 0.01; and �� , P < 0.001.






cancer specimens (Fig. 5A). Confocal microscope analysis alsoshowed the existence of tumor-infiltrating cells which positivelystained for both CD68 and TGFb1 (Supplementary Fig. S6),indicating that a proportion of CD68þ macrophages expressedTGFb1. In addition, monocytes/macrophages isolated fromtumors produced higher TGFb1 compared with those from non-tumor tissues (Fig. 5B). Flow cytometric analysis showed thatTGFb1was absent on the surface of tumor-associatedmonocytes/macrophages (Supplementary Fig. S7), suggesting that gastriccancer–associated monocytes/macrophages may secrete TGFb1tomediateNK-cell functional impairment. To further confirm thishypothesis, a neutralizing antibody against TGFb1 was added tothe monocytes/macrophages and NK-cell coculture system.TGFb1 blockade subsequently attenuated tumor-associatedmonocytes/macrophages-mediated suppression of IFNg , TNFa,and Ki-67 expression in NK cells (Fig. 5C and D).

Tumor-infiltrating NK cells correlate with multiple clinicalparameters

Finally, we investigated the implications of having tumor-infiltrating NK cells with regards to human gastric cancerprogression (Fig. 6). We found that the percentages oftumor-infiltrating NK cells were significantly decreased atadvanced TNM stages, and low NK-cell percentages were pos-itively correlated with poor overall survival of gastric cancerpatients when using the medium value of all tumor-infiltratingNK-cell percentages as a comparison point. We also found thattumor-infiltrating NK-cell percentages independently predicted

patient survival, which was verified by multivariate analysesusing a Cox proportional hazard model (Supplementary TableS3). In addition, the percentages of tumor-infiltrating NK cellsnegatively correlated with other advanced-stage clinical para-meters including tumor size, tumor invasion, lymph node anddistant metastasis, and neural invasion status. Nevertheless,age, gender, H. pylori infection status, histologic type, andvascular invasion status were not significantly different. Aboveall, these results suggested that decreased tumor-infiltrating NK-cell percentages were associated with gastric cancer progressionand patients' poor overall survival.

DiscussionDeciphering host innate and adaptive immune cell roles within

the tumor milieu is critical for understanding the developmentand progression of human tumors (15, 16). Althoughmuch efforthas been focused on delineating the functions of adaptiveimmune cells in gastric cancer (17, 18), the roles of antitumorinnate immune cells remain lesswell understood. In this study,weshowed that the percentage of NK cells in tumors was significantlydecreased at advanced stages of gastric cancer, with a low per-centage of NK cells positively correlating with poor overall sur-vival of gastric cancer patients. In addition, tumor-infiltrating NKcells colocalized with monocytes/macrophages, and monocytes/macrophages isolated from tumors induced the functionalimpairment of NK cells, indicating a new pathway of tumorimmune escape during gastric cancer progression.

Figure 3.

Functional characteristics of NK cells in gastric cancer patients. A, Representative dot-plots of IFNg , TNFa, Ki-67, granzyme B, and perforin expression levels in NKcells from paired nontumor and tumor tissues from a gastric cancer patient. B, Statistical analysis on IFNgþ, TNFaþ, Ki-67þ, granzyme Bþ, and perforinþ

NK-cell percentages from nontumor and tumor tissues of gastric cancer patients. Symbols represent individual values from 11 to 12 gastric cancer patientsanalyzed individually. � , P < 0.05; �� , P < 0.01; NS, P > 0.05.

Peng et al.





In humans, NK-cell infiltration in tumors influences diseaseprogression and patient survival (19). Therefore, the analysis ofNK-cell infiltration in gastric cancer is a critical area of clinicalinvestigation. In the present study, we observed a significantnegative association between the percentage of NK cells andadvanced clinical features of gastric cancer, such as tumor size,neural invasion, tumor invasion, and lymph node and distantmetastasis status. Low percentages of tumor-infiltrating NKcells independently predicted a lower rate of patient overallsurvival, suggesting that tumor-infiltrating NK cells may be auseful clinical marker of gastric cancer progression. Regardlessof patient outcome, the percentage of NK cells was significantlydecreased in tumors compared with nontumor tissues. Thismay be a consequence of either a decreased capacity of NK cellsto migrate from the peripheral blood or enhanced apoptosis ofNK cells that infiltrate tumors. Colorectal cancer and renal cellcarcinoma have greatly impaired NK-cell infiltration, despite

high concentrations of locally available chemokines (20, 21).NK cells are also reported to be sensitive to H2O2-inducedapoptosis, with high concentrations of H2O2 found withintumor microenvironments, including that of gastric cancer(22). Therefore, it is possible for increased NK-cell apoptosisto additionally contribute to the decreased infiltration of NKcells in gastric cancer tumors.

Along with decreased percentages of NK cells, we observed thatthe function of tumor-infiltrating NK cells in gastric cancer wasimpaired, as indicated by their decreased expression of Ki-67,IFNg , and TNFa. However, the mechanism of NK-cell functionalimpairment in gastric cancer remains unclear. Tumor-associatedmonocytes/macrophages are known tomodulate the functions ofvarious immune cells, such as T cells and NK cells, through theproduction of anti-inflammatory or immunomodulatory factors(10, 23). Thus, we investigated the relationship between mono-cytes/macrophages and NK cells in gastric cancer tumors.

Figure 4.

NK-cell functions are impairedby tumor-associatedmonocytes/macrophages. A,Immunohistochemical staining ofCD68þ macrophages (red) and CD57þ

cells (brown) on the same section oftumor tissue from gastric cancerpatients. B, CD45þCD14þ monocytes/macrophages percentagesnegatively correlated with IFNgþ

(black dot) and TNFaþ NK cells (reddot) expression in tumors from 12gastric cancer patients. C–F,Monocytes/macrophages purified fromnontumor and tumor tissues ofgastric cancer patientswere coculturedwith NK cells from healthy allogenicindividuals for 5 to 7 days, and NK-cellexpression of IFNg , TNFa, and Ki-67then determined by flow cytometry(n ¼ 3). � , P < 0.05.






Immunohistochemical staining showed that most gastric cancer–infiltrating CD57þ cells localized in close proximity to CD68þ

macrophages, and we found that the majority of gastric cancer–infiltrating CD57þ cells belonged to NK cells by flow cytometry(data not shown).Moreover, CD57 has been reported as a specificmarker for NK cells including gastric cancer (13, 24). Therefore, inaddition to modulating T-cell function, monocytes/macrophagesmay also interact with NK cells within gastric cancer tumors. Thepercentages of tumor-associated monocytes/macrophages nega-tively correlated with the percentages of IFNgþ and TNFaþ NKcells, and ex vivo coculture experiments confirmed that tumor-associated monocytes/macrophages suppressed Ki-67, IFNg , andTNFa expression inNK cells. Thus, our results demonstrated a rolefor tumor-associated monocytes/macrophages in NK-cell func-tional impairment.

TGFb is a pleiotropic cytokine capable of suppressingantitumor immune responses (25). Although the concentra-tion of TGFb1 is significantly increased in gastric cancertumors (26), the cellular source of TGFb1 remained unclear.Here, we found that tumor-associated monocytes/macro-phages produced high amounts of TGFb1, with TGFb1 block-ade restoring NK-cell expression of IFNg , TNFa, and Ki-67 exvivo. Previous studies have demonstrated that human regula-tory T cells (Treg) express membrane-bound TGFb, capable ofdirectly inhibiting NK-cell effector function (27). However,our data showed that TGFb1 was not localized on the surfaceof gastric cancer–associated monocytes/macrophages, butsecreted extracellularly. Tumor-infiltrating Tregs in gastriccancer tumors have been reported to express littleTGFb1,potentially ruling them out as a predominant source of local

Figure 5.

TGFb1-dependent regulation of NK-cell function by tumor-associated monocytes/macrophages. A, Representative immunohistochemical staining of TGFb1 andCD68 from serial paraffin tumor tissue sections (�20 magnification). B, Monocytes/macrophages sorted from nontumor and tumor tissues of gastric cancerpatients were cultured for 48 hours and analyzed for TGFb1 production (n ¼ 3). C–D, Purified monocytes/macrophages from nontumor and tumor tissues werecultured with NK cells from allogenic healthy individuals for 5 to 7 days with or without anti-TGFb1 (10 mg/mL) or an isotype control antibody, and NK-cellexpression of IFNg , TNFa, and Ki-67 then detected by flow cytometry (n ¼ 3). � , P < 0.05.

Peng et al.





TGFb1 in gastric cancer (28). Our results suggested that gastriccancer–associated monocyte/macrophage-secretion of TGFb1induced the functional impairment of NK cells within gastriccancer tumors.

Tumor-associated monocytes/macrophages have also beenreported to induce NK-cell dysfunction through high expressionof CD48 and PD-L1/PD-L2 molecules (13, 29). However, expres-sion of CD48 or PD-L1/PD-L2 on monocytes/macrophagesbetween tumors and nontumor tissues was not significantlydifferent in gastric cancer patients. Therefore, tumor-associatedmonocytes/macrophages may inhibit NK-cell function throughdifferent signaling pathways in different tumor microenviron-

ments, and further work is needed to elucidate the precise micro-environment parameters and mechanisms involved in thisprocess.

In conclusion, our study has highlighted an important role forNK cells in human gastric cancer and identified a newmechanismof gastric cancer–associated monocytes/macrophages-dependentNK-cell functional impairment. Critically, it highlights anotherpathway whereby maladaptive host innate immune cells canactively inhibit local antitumoral effector immune cells. Overall,restoring the function of NK cells that infiltrate tumors could be auseful therapeutic strategy for preventing gastric cancer tumorimmune escape.

Figure 6.

The percentage of tumor-infiltrating NK cells correlated with multiple clinical parameters of gastric cancer. Percentages of tumor-infiltrating CD3–CD56þ NK cellswere analyzed for putative correlations withmultiple clinical parameters. For the cumulative survival curves, patientswere separated into two groups by themedianvalue of tumor-infiltrating NK cells percentages, and Kaplan–Meier plots used to calculate cumulative survival differences. � , P < 0.05; �� , P < 0.01; NS, P > 0.05. Eachdot represents 1 patient. Ab, antibody; Diff, differentiated; Undiff, undifferentiated.






Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: L.-s. Peng, J.-y. Zhang, W. Chen, Q.-m. Zou, Y. ZhuangDevelopment of methodology: L.-s. Peng, J.-y. Zhang, Y.-s. Teng, Y.-l. Zhao,N. Chen, Y. ZhuangAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): L.-s. Peng, J.-y. Zhang, Y.-s. Teng, T.-t. Wang, F.-y. Mao,Y.-p. Lv, P. Cheng, W.-h. LiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): L.-s. Peng, J.-y. Zhang,W.Chen,Q.-m. Zou, Y. ZhuangWriting,review,and/orrevisionofthemanuscript:L.-s. Peng,M.Duan,W.Chen,Q.-m. Zou, Y. Zhuang

Administrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): L.-s. Peng, Y.-l. Zhao, G. GuoStudy supervision: L.-s. Peng, J.-y. Zhang, Q.-m. Zou, Y. Zhuang

Grant SupportThis work was supported by the National Key Research and Development

Program of China (2016YFC1302200) and grants of the National NaturalScience Foundation of China (NSFC, No. 81502457 and No. 81402355).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 4, 2016; revisedDecember 12, 2016; accepted January 10, 2017;published OnlineFirst February 1, 2017.

References1. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer

statistics, 2012. CA Cancer J Clin 2015;65:87–108.2. Lee HE, Chae SW, Lee YJ, Kim MA, Lee HS, Lee BL, et al. Prognostic

implications of type and density of tumour-infiltrating lymphocytes ingastric cancer. Br J Cancer 2008;99:1704–11.

3. Apetoh L, Smyth MJ, Drake CG, Abastado JP, Apte RN, Ayyoub M, et al.Consensus nomenclature for CD8þ T cell phenotypes in cancer. Oncoim-munology 2015;4:e998538.

4. Zhuang Y, Peng LS, Zhao YL, Shi Y, Mao XH, ChenW, et al. CD8(þ) T cellsthat produce interleukin-17 regulate myeloid-derived suppressor cells andare associated with survival time of patients with gastric cancer. Gastro-enterology 2012;143:951–62.e8.

5. Mandal A, Viswanathan C. Natural killer cells: In health and disease.Hematol Oncol Stem Cell Ther 2015;8:47–55.

6. Langers I, Renoux VM, Thiry M, Delvenne P, Jacobs N. Natural killer cells:Role in local tumor growth and metastasis. Biologics 2012;6:73–82.

7. Mamessier E, Sylvain A, Thibult ML, Houvenaeghel G, Jacquemier J,Castellano R, et al. Human breast cancer cells enhance self tolerance bypromoting evasion from NK cell antitumor immunity. J Clin Invest 2011;121:3609–22.

8. Platonova S, Cherfils-Vicini J, Damotte D, Crozet L, Vieillard V, Validire P,et al. Profound coordinated alterations of intratumoral NK cell phenotypeand function in lung carcinoma. Cancer Res 2011;71:5412–22.

9. Peng YP, Zhu Y, Zhang JJ, Xu ZK, Qian ZY, Dai CC, et al. Comprehensiveanalysis of the percentage of surface receptors and cytotoxic granulespositive natural killer cells in patients with pancreatic cancer, gastric cancer,and colorectal cancer. J Transl Med 2013;11:262.

10. Chanmee T, Ontong P, Konno K, Itano N. Tumor-associated macrophagesas major players in the tumor microenvironment. Cancers (Basel) 2014;6:1670–90.

11. Noy R, Pollard JW. Tumor-associated macrophages: From mechanisms totherapy. Immunity 2014;41:49–61.

12. ZhaoQ, Xiao X,Wu Y,Wei Y, Zhu LY, Zhou J, et al. Interleukin-17-educatedmonocytes suppress cytotoxic T-cell function through B7-H1 in hepato-cellular carcinoma patients. Eur J Immunol 2011;41:2314–22.

13. Wu Y, Kuang DM, Pan WD, Wan YL, Lao XM, Wang D, et al. Monocyte/macrophage-elicited natural killer cell dysfunction in hepatocellular car-cinoma is mediated by CD48/2B4 interactions. Hepatology 2013;57:1107–16.

14. Zhuang Y, Peng LS, Zhao YL, Shi Y, Mao XH, Guo G, et al. Increasedintratumoral IL-22-producing CD4(þ) T cells and Th22 cells correlate withgastric cancer progression and predict poor patient survival. Cancer Immu-nol Immunother 2012;61:1965–75.

15. Galon J, Angell HK, Bedognetti D,Marincola FM. The continuum of cancerimmunosurveillance: Prognostic, predictive, and mechanistic signatures.Immunity 2013;39:11–26.

16. Schreiber RD, Old LJ, Smyth MJ. Cancer immunoediting: integratingimmunity's roles in cancer suppression and promotion. Science 2011;331:1565–70.

17. Feichtenbeiner A, Haas M, B€uttner M, Grabenbauer GG, Fietkau R, DistelLV. Critical role of spatial interaction between CD8þ and Foxp3þ cells inhuman gastric cancer: The distancematters. Cancer Immunol Immunother2014;63:111–9.

18. Liu K, Yang K, Wu B, Chen H, Chen X, Chen X, et al. Tumor-infiltratingimmune cells are associated with prognosis of gastric cancer. Medicine(Baltimore) 2015;94:e1631.

19. Desbois M, Rusakiewicz S, Locher C, Zitvogel L, Chaput N. Natural killercells in non-hematopoietic malignancies. Front Immunol 2012;3:395.

20. Halama N, BraunM, Kahlert C, Spille A, Quack C, Rahbari N, et al. Naturalkiller cells are scarce in colorectal carcinoma tissue despite high levels ofchemokines and cytokines. Clin Cancer Res 2011;17:678–89.

21. Sconocchia G, Spagnoli GC, Del Principe D, Ferrone S, Anselmi M,Wongsena W, et al. Defective infiltration of natural killer cells in MICA/B-positive renal cell carcinoma involves beta(2)-integrin-mediated inter-action. Neoplasia 2009;11:662–71.

22. Izawa S, Kono K,Mimura K, Kawaguchi Y, WatanabeM,Maruyama T, et al.H2O2 production within tumor microenvironment inversely correlatedwith infiltration of CD56(dim) NK cells in gastric and esophageal cancer:Possible mechanisms of NK cell dysfunction. Cancer Immunol Immun-other 2011;60:1801–10.

23. SolinasG,GermanoG,Mantovani A, Allavena P. Tumor-associatedmacro-phages (TAM) as major players of the cancer-related inflammation. JLeukoc Biol 2009;86:1065–73.

24. Li T, Zhang Q, Jiang Y, Yu J, Hu Y, Mou T, et al. Gastric cancer cells inhibitnatural killer cell proliferation and induce apoptosis via prostaglandin E2.Oncoimmunology 2015;5:e1069936.

25. AchyutBR, YangL. Transforming growth factor-b in the gastrointestinal andhepatic tumor microenvironment. Gastroenterology 2011;141:1167–78.

26. Peng LS, Zhuang Y, Shi Y, Zhao YL, Wang TT, Chen N, et al. Increasedtumor-infiltrating CD8(þ)Foxp3(þ) T lymphocytes are associated withtumor progression in human gastric cancer. Cancer Immunol Immunother2012;61:2183–92.

27. Ghiringhelli F, M�enard C, Terme M, Flament C, Taieb J, Chaput N, et al.CD4þCD25þ regulatory T cells inhibit natural killer cell functions in atransforming growth factor-beta-dependent manner. J Exp Med 2005;202:1075–85.

28. Kindlund B, Sj€oling Å, Yakkala C, Adamsson J, Janzon A, Hansson LE,et al. CD4þ regulatory T cells in gastric cancer mucosa are proliferatingand express high levels of IL-10 but little TGFb. Gastric Cancer 2016;20:116–25.

29. Noy R, Pollard JW. Tumor-associated macrophages: From mechanisms totherapy. Immunity 2014;41:49–61.


Peng et al.




2017;5:248-256. Published OnlineFirst February 1, 2017.Cancer Immunol Res Liu-sheng Peng, Jin-yu Zhang, Yong-sheng Teng, et al.

1 in Human Gastric CancerβFunction via TGFTumor-Associated Monocytes/Macrophages Impair NK-Cell

Updated version

10.1158/2326-6066.CIR-16-0152doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerimmunolres.aacrjournals.org/content/suppl/2017/02/01/2326-6066.CIR-16-0152.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerimmunolres.aacrjournals.org/content/5/3/248.full#ref-list-1

This article cites 29 articles, 4 of which you can access for free at:

Citing articles

http://cancerimmunolres.aacrjournals.org/content/5/3/248.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerimmunolres.aacrjournals.org/content/5/3/248To request permission to re-use all or part of this article, use this link



http://cancerimmunolres.aacrjournals.org/lookup/doi/10.1158/2326-6066.CIR-16-0152

http://cancerimmunolres.aacrjournals.org/content/suppl/2017/02/01/2326-6066.CIR-16-0152.DC1

http://cancerimmunolres.aacrjournals.org/content/5/3/248.full#ref-list-1

http://cancerimmunolres.aacrjournals.org/content/5/3/248.full#related-urls

http://cancerimmunolres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerimmunolres.aacrjournals.org/content/5/3/248


Documents

Tumor-Associated Monocytes/Macrophages Impair …...Research Article Tumor-Associated Monocytes/Macrophages Impair NK-Cell Function via TGFb1 in Human Gastric Cancer Liu-sheng Peng1,