Activated Eosinophils Exert Antitumorigenic Activities in ...Research Article Activated Eosinophils Exert Antitumorigenic Activities in Colorectal Cancer Hadar Reichman1, Michal Itan1,

Research Article

Activated Eosinophils Exert AntitumorigenicActivities in Colorectal CancerHadar Reichman1, Michal Itan1, Perri Rozenberg1, Tal Yarmolovski1, Eli Brazowski2,Chen Varol2, Nathan Gluck2, Shiran Shapira3, Nadir Arber3, Udi Qimron1,Danielle Karo-Atar1, James J. Lee4,†, and Ariel Munitz1

Abstract

Immunotherapies targeting T lymphocytes are revolution-izing cancer therapy but only benefit a subset of patients,especially in colorectal cancer. Thus, additional insight intothe tumor microenvironment (TME) is required. Eosinophilsare bone marrow–derived cells that have been largely studiedin the context of allergic diseases and parasite infections.Although tumor-associated eosinophilia has been describedin various solid tumors including colorectal cancer, knowledgeis still missing regarding eosinophil activities and even thebasic question of whether the TME promotes eosinophilrecruitment without additional manipulation (e.g., immuno-therapy) is unclear. Herein, we report that eosinophilsare recruited into developing tumors during induction ofinflammation-induced colorectal cancer and in mice with the

Apcmin/þ genotype, which develop spontaneous intestinal ade-nomas. Using adoptive transfer and cytokine neutralizationexperiments, we demonstrate that the TME supported pro-longed eosinophil survival independent of IL5, an eosinophilsurvival cytokine. Tumor-infiltrating eosinophils consisted ofdegranulating eosinophils and were essential for tumor rejec-tion independently of CD8þ T cells. Transcriptome and prote-omic analysis revealed an IFNg-linked signature for intratu-moral eosinophils that was different from that of macro-phages. Our data establish antitumorigenic roles for eosino-phils in colorectal cancer. These findings may facilitate thedevelopment of pharmacologic treatments that could unleashantitumor responses by eosinophils, especially in colorectalcancer patients displaying eosinophilia.

IntroductionThe tumor microenvironment (TME) has been recognized as a

critical factor in tumor biology (1). The importance of theimmune system in the TME is nicely exemplified by the crucialfunctions that have been uncovered for cytotoxic T cells in tumorcell elimination (2), which have been utilized for therapy usingimmune-checkpoint blockade (3). However, the TME often pro-motes T-cell suppression, which complicates targeting of T cells,especially in solid tumors, and tumors may lack specific antigensor resist infiltration by cytotoxic cells. These limitations highlightthe need to identify additional cellular targets that possess anti-tumorigenic activities. In pursuing potentially antitumorigeniccellular targets, we focused on eosinophils.

Eosinophils are bone marrow–derived cells that have beenlargely studied in the context of allergic inflammation and par-

asitic infections. However, eosinophils infiltrate multiple tumorswhere they can display either pro- or antitumorigenic functions,leading to controversy about their role (4, 5). Nonetheless, mostof the experimental data assessing eosinophil function in cancerhave been gathered using tumor cell lines that secrete eosinophil-promoting cytokines (e.g., IL5 and CCL11; refs. 6, 7), geneticallymodified tumors that polarize a type 2 cytokine environment (8),responses to immunotherapy (e.g., IL4, GM-CSF, IL33, and TSLP)or in the absence of T regulatory cells (9–12). Thus, translation ofsuch data into insights about human disease is difficult.

Here, we studied eosinophils in cancer with experimentalmodels that fulfilled three conditions: (i) The models addressedtumor types with clinically reported data on increased tumor-infiltrating eosinophilia; (ii) tumor development occurred in ananatomically relevant tissue for eosinophils; and (iii) tumorprogression occurred gradually without exogenous tumor cellinjections, a process that allowed eosinophils to adapt to theirchanging microenvironment. Using these criteria led us to inves-tigate eosinophils in colorectal cancer. Several clinical studiesreported increased tissue eosinophilia in colorectal cancerpatients and that eosinophilia was associated with favorableprognosis (13–15). The largest homeostatic niche for eosinophilsis the gastrointestinal (GI) tract; and eosinophil accumulation inthe GI tract is a feature of numerous inflammatory GI diseases, inwhich eosinophils have key functions (16, 17). Finally, severalchronic experimental models exist for colorectal cancer, whichmimic the human disease, at least in part (18).

We show herein that the TME in human and experimentalcolorectal cancer is characterized by eosinophil recruitment, pro-longed survival, and degranulation. Eosinophils possessed anti-tumorigenic activities that were independent of CD8þ T cells.

1Department of Clinical Microbiology and Immunology, the Sackler School ofMedicine, Tel Aviv University, Ramat Aviv, Israel. 2Research Center for DigestiveTract and Disorders and Liver Diseases, Tel Aviv Sourasky Medical Center,Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. 3IntegratedCancer Prevention Center, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel.4Division of Pulmonary Medicine, Department of Biochemistry and MolecularBiology, Mayo Clinic Arizona, Scottsdale, Arizona.

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

†Deceased

Corresponding Author: Ariel Munitz, Tel Aviv University, Tel Aviv 69978, Israel.Phone: 972524388101; E-mail: [email protected]

doi: 10.1158/2326-6066.CIR-18-0494

�2019 American Association for Cancer Research.

CancerImmunologyResearch

Cancer Immunol Res; 7(3) March 2019388

on May 10, 2021. © 2019 American Association for Cancer Research. cancerimmunolres.aacrjournals.org Downloaded from

Published OnlineFirst January 21, 2019; DOI: 10.1158/2326-6066.CIR-18-0494

http://crossmark.crossref.org/dialog/?doi=10.1158/2326-6066.CIR-18-0494&domain=pdf&date_stamp=2019-2-8

http://cancerimmunolres.aacrjournals.org/

Unbiased RNA-seq and proteomics revealed that intratumoraleosinophils were characterized by an IFNg signaling signature.Consistent with this signature and the in vivo role for eosinophilsin colorectal cancer, activation of eosinophils with IFNg in vitropotentiated eosinophil-mediated killing of colorectal cancer cells.Collectively, our data established antitumorigenic functions foreosinophils in colorectal cancer and describe the phenotypiclandscape of intratumoral eosinophils. Thus, pharmacologicapproaches that target eosinophils have the potential to unleashantitumor activities.

Materials and MethodsHuman tissue arrays

Human tissue arrays as well as data on patients' characteristicswere obtained from US Biomax, Inc.. The following arrays wereused: CO601, CO602, CO703, CO702B, CO952, CO953,CO992A, TO54B, and TO55, and a total of 351 human sampleswere analyzed.

MiceWild-type (WT) C57BL/6 mice were originally obtained from

Harlan Laboratories and grown in-house. CD3-IL5 transgenicmice (NJ.1638, Il5Tg) mice were kindly provided by Dr. JamieLee (Mayo Clinic, Scottsdale, AZ). DdblGATA mice were kindlyprovided by Dr. August Avery (Cornell University, Ithaca, NY).ApcMin/þmice were obtained from The Jackson laboratory. ApcMin/

þ/DdblGATA mice were generated by mating male ApcMin/þ micewith female DdblGATA. All experiments were reviewed andapproved by the Animal Care Committee of Tel Aviv University(Number M-14-061, M-15-001) and were performed in accor-dance with its regulations and guidelines regarding the care anduse of animals for experimental procedures. All of the experimentswere conducted in the specific pathogen-free facilities of Tel AvivUniversity. In all experiments, age-, weight-, and sex-matchedmice were used.

Colitis-associated cancerMice were injected intraperitoneally with 12.5 mg/kg azoxy-

methane (AOM, Sigma). At 5 days after injection,mice were given2.25% dextran sodium sulfate (DSS) in drinking water for 5 days,followed by 16 days of normal drinking water. DSS treatment wasrepeated for 2 additional cycles of 2.25% and 2% DSS, and micewere euthanized at 10 to 13 weeks after AOM injection.

Adoptive transfer experimentsSpleens were extracted from Il5Tg mice and crushed through a

70-mm strainer. Red blood cells were lysed, and the remainingwhite blood cells were subjected to lymphocyte depletion usingDynabeads conjugated to antibodies against Thy1.2 and B220(Thermo Fisher Scientific). The purity of eosinophils was deter-mined using flow cytometry and was consistently >95%. There-after, DdblGATA mice were injected intravenously with 50 � 106

eosinophils on day 4 of the DSS treatment or at the age of 6 weeksin the Apcmin/þ mouse model.

Orthotopic injectionsMice were anesthetized using Ketamine/Xylazine. Submucosal

injections of MC38 colorectal cancer cells were accomplishedusing flexible stainless steel, 8 inch long, 30 gauge, 45 degreebevel hypodermic needles custom made according to our spec-

ification (Cadence, Inc.). The needle was inserted through Luerlock (S€ollner, GmbH) screwed on the working channel of theendoscope to avoid air leakage. After the scope was inserted intothe mouse colon and the colon inflated, the needle was broughtthrough the working channel to the scope's front. The colorectalcancer cell implantation procedure required two persons: one tonavigate the colonoscopywhile the other performed the injection.The injection consisted of a very gentle submucosal penetrationwith the open side of the bevel heading up at aflat angle. A volumeof 50 mL colorectal cancer tumor cells in saline was then injectedinto the colonic submucosa. Mice were euthanized on day 21 andhistologic specimens were prepared.

IL5 neutralizationApcmin/þ mice were treated for a period of up to 5 months with

anti-IL5 (TRFK5, 50 mg/mouse, twice a week) or isotype controlantibodies. Thereafter, mice were euthanized, and eosinophillevels were determined in the peripheral blood and intestines.

Punch biopsiesColons were flushed with phosphate-buffered saline and

opened along a longitudinal axis; 3 mm2 punch biopsies wereincubated for 24 hours in RPMI supplemented with 10% fetal calfserum and antibiotics. Supernatants were collected and assessedfor cytokine expression by ELISA (19).

Coculture experimentsPrimary eosinophils were isolated from the peritoneal lavage of

Il5Tg mice and enriched by lymphocyte depletion through use ofeither a MACS cell separation system with antibodies againstCD90.2 and CD45R (Miltenyi Biotec) or Dynabeads conjugatedto antibodies against Thy1.2 and B220 (Thermo Fisher Scientific).Eosinophils were then seeded togetherwithCT26orMC38 cells at800,000 cells per well at varying ratios. In several experiments,eosinophils were cultured in the presence of colorectal cancercell–conditioned media, which was obtained from the cells afterthey reached confluency. Eosinophils were cultured with tumorcell–conditioned media in the presence of anti-IL5 (clone TRFK5,0.15 mg/mL) or isotype control.

Enzymatic digestion of gastrointestinal lamina propria cellsColonic tissue was excised and flushed with 1 mL of calcium-

andmagnesium-free HBSS (CMF-HBSS). The colonwas dissectedlongitudinally and shaken (250 RPM) in 5 mL CMF-HBSS con-taining 5% FCS, 2 mmol/L EDTA, and 1mmol/L DTT (Ditiother-itol) for 40minutes at 37�C in order to remove epithelial cells andintraepithelial lymphocytes. Then, the colonic tissue was vortexedand strained through 70-mm gray mesh. The remaining tissue orisolated tumors were washed in PBS and then incubated andshaken (250 RPM) with complete PBS (containing calcium andmagnesium) supplemented with 5% FCS, 1mg/mL collagenase A(Roche), and 0.1mg/mLDnase I (Sigma) for 40minutes at 37�C.The cell suspension was filtered using gauze (70-mm mesh) andsuspended in flow cytometry staining buffer (HBSS, 1% FCS).

Flow cytometrySingle-cell suspensions of mouse cells were stained using the

following antibodies: anti-CD45-APC, anti-CD45-APC-eFluor780, anti-CD11b-PerCP-Cy5.5, anti-Gr1-PE, anti-CD8a-PE (obtained from eBioscience), anti-CD3e-PE-Cy7 (obtainedfromBioLegend), anti-CCR3-FITC (obtained fromR&DSystems),

Eosinophils in Colorectal Cancer

www.aacrjournals.org Cancer Immunol Res; 7(3) March 2019 389




anti-Siglec-F-PE (BD Biosciences), and DAPI (Sigma). Eosino-phils were identified as CD45þ/CD11bþ/Siglec-Fþ/Ly6c�/Ly6g�/MHC-II� (colon and ileum); CD45þ/Siglec-Fþ/CCR3þ

(Blood and bone marrow). T cells and myeloid-derived suppres-sor cells were identified as CD45þ/CD3þ/CD8þ and CD45þ/Gr1þ/CD11bþ cells, respectively.

IHCTissues were fixed, embedded, sectioned, and stainedwith anti-

EPX or anti-major basic protein (MBP) (kindly provided by Dr.Jamie Lee, Mayo Clinic, Scottsdale, AZ), as described previously(19). For anti-CD31 and anti-ki67 staining, slides were trypsi-nized with 0.1% trypsin at 37�C for 5 minutes (BD Difco),incubated with 180 mL methyl alcohol and 3 mL of 30% hydro-gen peroxide, blocked with 2% rabbit serum/PBS/Triton for2 hours, and incubated overnight with rat anti-mouse CD31(PECAM-1) or rat anti-mouse Ki67 (both obtained from Abcam),at 4�C. Slides were then incubated with biotinylated rabbit anti-rat (Vector Laboratory; BA-4001) for 60 minutes, developed withABC complex andDAB substrate (Vector Laboratory, and counter-stained with 0.1% nuclear fast red in 5% aluminum sulfate for2 minutes or hematoxylin. Images were captured using either anOlympus AX70 fluorescent microscope (Center Valley) equippedwith a DP72 camera or a Leica DM1000 equipped with an ICC50camera (Leica). Image analysis and quantitation were performedwith ImageJ (NIH).

ImmunofluorescenceTissues were placed in OCT (Tissue-TEK) and snap-frozen over

dry ice. Tissue sections were cut, air dried, fixed, and blocked.Thereafter, tissues were stained with anti-MBP (kindly providedby Dr. Jamie Lee, Mayo Clinic, Scottsdale, AZ) and anti-cleavedcaspase-3 (Cell Signaling) followed by the following secondaryantibodies: donkey anti-rat AlexaFluor 488 (1:300, JacksonImmunoResearch) and goat anti-rabbit AlexaFluor 546 (1:500,Life Technologies). Slides were stained with DAPI (Sigma) andmounted using gel-mount (Sigma). Images were captured usingan LSM 800 confocal microscope (Zeiss).

RNA-seqRNA was extracted using TRIzol (Invitrogen) according to the

manufacturer's instructions. Samples were preparedwith CEL-seqand sequenced using Illumina HiSeq 2500. Sample preparation,sequencing, quality control, and differential expression analyseswere conducted by the "Technion Genome Center," Life Scienceand Engineering Interdisciplinary Research Center, Technion,Haifa, Israel.

ProteomicsTo identify differentially expressed proteins and phospho-pro-

teins, eosinophils and macrophages were subjected to proteinisolation using scioExtract Pro (Sciomics) using adapted proto-cols. All samples to be analyzed were normalized to have equalprotein yields according to the respective cell counts. Individualeosinophil cell sorting batches were pooled to yield at least229,000 to 235,000 cells in total. The individual samples wereextracted in a cascade fashion,meaning that the first samples weretreated with scioExtract and the whole solution was carried to thenext sample tube. Three individual sampleswere pooled to yield asample for analysis. For all other samples, individual sampleswere extracted according to standard Sciomics protocols with

scioExtract Pro. The complete resulting protein fraction waslabeled with scioDye 1 (Sciomics) and purified with low cellcount additives (Sciomics) during the purification step. All sam-ples were incubated for three hours on scioDiscover arrays con-taining 1,130 antibodies against 900 proteins (Sciomics) accord-ing to standard protocols for microarray analysis for proteinexpression and phosphorylation analysis. Subsequently, slideswerewashed anddried. Slideswere scannedusing a Powerscanner(Tecan) to obtain fluorescent values. The resulting raw data wereanalyzed using the linear models for microarray data (LIMMA)package of R-Bioconductor for differential protein expressionincluding normalization (Cyclic Lowess) and P value as well aslogFc calculation. The false discovery rate was controlled accord-ing to Benjamini and Hochberg (20).

Proliferation assaysCell proliferation was assessed using the Click-iT EdU kit

(Invitrogen) according to the manufacturer's instructions.

Statistical analysisData were analyzed by analysis of variance followed by Tukey

post hoc test or Student t test using GraphPad Prism 5. Survivalcurves were analyzed using the Gehan–Breslow–Wilcoxon testand the log-rank (Mantel–Cox) test. Data are presented asmean�SEM, and values of P < 0.05 were considered statistically signif-icant; P values for differential expression of RNA-seq datawere corrected for multiple testing according to the Benjamini–Hochberg procedure (20).

ResultsAccumulation of eosinophils in human colorectal cancerinversely correlates with tumor stage

To define the role of eosinophils in colorectal cancer, we firstanalyzed intratumoral eosinophils using tissue arrays containingbiopsies from 275 patients and healthy individuals (see thepatient characteristics in Supplementary Table S1). Tissues werestained with anti-eosinophil peroxidase (EPX), a commonlyusedmethod determine human eosinophil numbers and degran-ulation in situ (21). Stained biopsies were assessed using com-puterized analyses and divided into four groups based onthe numbers of intratumoral eosinophils: <10 eosinophils/mm2;10–40 eosinophils/mm2; 40–100 eosinophils/mm2; and >100eosinophils/mm2 (Fig. 1A). Segregation of the biopsies accordingto the tumor stage revealed an inverse correlation between tumorstage and intratumoral eosinophils (Fig. 1B), which was not dueto alterations in eosinophil numbers in the adjacent healthy tissue(Fig. 1C). Most specimens, independent of tumor grade, con-tained degranulated eosinophils as observed by extracellulardeposition of eosinophil granule content (Fig. 1D, blue arrows)as well as intact cells (Fig. 1D, black arrows).

Next, we asked whether eosinophils were also present indraining lymph node metastases. To this end, eosinophils inbiopsies from 91 patients who displayed lymph node metastasiswere assessed (Fig. 1E). Prominent eosinophilia (i.e., more than10 eosinophils/mm2) was observed in �27% of the biopsies(Fig. 1F).

Eosinophils are recruited and activated in experimentalcolorectal cancer

To delineate the roles of eosinophils in colorectal cancer, weinvestigated whether eosinophils accumulate in the GI tract

Reichman et al.

Cancer Immunol Res; 7(3) March 2019 Cancer Immunology Research390




during experimental colorectal cancer, using two models thatrepresent distinct etiologies: chronic inflammation (i.e., colitis-associated cancer; CAC) and genetic predisposition (Apcmin/þ

mice, resembling familial adenomatous polyposis patients—whodo not always have an inflammatory phenotype; ref. 22).

Anti-eosinophil MBP stain revealed a gradual elevation incolonic eosinophils following induction of CAC (Fig. 2A). Earlystages of this model (weeks 3 and 7) showed elevated eosino-philia, and the colon was highly infiltrated with eosinophils byweek 11. Intratumoral eosinophils were readily detected (Fig. 2B)and extracellular MBP, amarker of eosinophil degranulation, wasobserved in the colonic lamina propria (Fig. 2C).

Subsequently, we aimed to determine whether eosinophilsaccumulate in the ileum of Apcmin/þ mice, which spontaneouslydevelop intestinal polyps. Our analyses revealed that eosinophilsinfiltrate the lamina propria (Fig. 2D) and the tumor (Fig. 2E).Flow-cytometric analysis of excised adenomas revealed that intra-tumoral eosinophils (defined as CD45þ/CD11bþ/MHC-II�/Siglec-Fþ/Ly6g�/Ly6c�/SSChi) constituted up to 13% of all intra-tumoral leukocytes (Fig. 2F) and �30% of all intratumoralmyeloid cells (i.e., CD11bþ cells). Degranulating eosinophils andextracellular MBP were observed in 5-month-old Apcmin/þ mice,which displaymultiple adenomas at that timepoint (Fig. 2G, bluearrows).

Figure 1.

Eosinophils accumulate in human colorectal cancer and are inversely correlated with the tumor stage. A, Human tissue arrays from patients with colorectalcancer (n¼ 275) were stained with anti-eosinophil EPX. B, Stained tissues were divided according to the presence of eosinophils per mm2, ranging from <10 to>100 eosinophils/mm2 and assessed for their tumor stage (I–IV). C, The number of eosinophils in nontumor, healthy areas was determined and assessed fortumor grade.D, A representative anti-EPX stained slide demonstrating degranulating and intact eosinophils (blue and black arrows, respectively) is shown.Eosinophil numbers were determined in colorectal cancer lymph nodemetastasis using anti-EPX staining (n¼ 91, E) and divided according to the presence ofeosinophils per mm2, ranging from <10 to >100 eosinophils/mm2 (F).






Reichman et al.





The accumulation of eosinophils in the GI tract of mice under-going CAC and in 5-month-old Apcmin/þmice was associated withfewer eosinophils in the peripheral blood (Fig. 2H and I) and, tolesser extent, in the bone marrow (Supplementary Fig. S1).Finally, syngeneic intracolon injection of MC38 colorectal cancercells resulted in the peritumoral accumulation of eosinophils(Supplementary Fig. S2).

Increased expression of CCL11/eotaxin-1 in colorectal cancerAssessment of the eosinophil-specific chemokines CCL11/

eotaxin-1 and CCL24/eotaxin-2 in GI punch biopsies revealeda specific increase in CCL11, but not CCL24 (Supplementary Fig.S3). Thus, eosinophil homing signals are present in the TME, andaccumulation of eosinophils in colorectal cancer is independentof mode of disease induction.

The TME in colorectal cancer promotes eosinophil recruitmentTo demonstrate that eosinophils were actively and specifically

recruited into the TME, purified splenic eosinophils from Il5Tg

mice were adoptively transferred into DdblGATA mice undergo-ing CAC (Fig. 2J, top) or into Apcmin/þ mice that lack eosinophils(Apcmin/þ/DdblGATA; Fig. 2J, bottom). The colons of DdblGATAmice, which received no eosinophils (No Eos), or na€�veDdblGATAmice that were injected once with eosinophils (þEos),displayed few eosinophils (defined as CD45þ/CD11bþ/Siglec-Fþ/Ly6c�/Ly6G�/MHC-II�/SSChi cells; Fig. 2K). In contrast, asingle injection of eosinophils into DdblGATA mice undergoingCAC resulted inpreferential and rapidhomingof eosinophils intothe colon (Fig. 2K and L). Similarly, a single injection of eosino-phils into Apcmin/þ/DdblGATA mice resulted in significantly moreeosinophils in comparisonwithDdblGATAmice (Fig. 2M andN).

Prolonged eosinophil survival in the TME is independent of IL5The observation that eosinophils were still present in the colon

of DdblGATA mice undergoing colitis-associated cancer and theileum of Apcmin/þ/DdblGATA mice, respectively, for up to threemonths (Fig. 2K–N), following a single injection (Fig. 2J), sug-gests that as well as recruiting eosinophils, the TME supportsprolonged eosinophil survival. Indeed, 14 days following adop-tively transferring eosinophils, we could not detect peripheralblood eosinophils nor observe any eosinophil proliferation in situ(Supplementary Fig. S4).

IL5 is an eosinophil survival factor that regulates eosinophildifferentiation, maturation, and survival (23). To determinewhether eosinophil survival in colorectal cancer is regulated byIL5, Apcmin/þ mice were treated with anti-IL5 neutralizing anti-bodies for 5 months, at which point peripheral blood andcolonic eosinophils were assessed. Neutralization of IL5 in

Apcmin/þ mice significantly decreased peripheral blood eosino-phils in comparison with isotype control–treated mice (Sup-plementary Fig. S5). Nonetheless, eosinophils were detected inthe colons of anti–IL5-treated Apcmin/þ mice, and no differencewas observed between the anti–IL5-treated and isotype con-trol–treated mice (Supplementary Fig. S5). Consistently, con-ditioned media of MC38 colorectal cancer cells increasedeosinophil survival in vitro (Fig. 2O). Neutralization of IL5had no effect on eosinophil survival in response to MC38-conditioned media (Fig. 2O).

Eosinophils prevent the development of colorectal cancerTo address the in vivo function of eosinophils in colorectal

cancer, CAC was induced in WT and DdblGATA mice. The fre-quency of colitis-associated cancer mortality in WT mice was13.3% (n ¼ 30 mice). In contrast, in the absence of eosinophils,themortality rose by 4.2-fold, and 56.2%of themice died (n¼ 32mice, P < 0.05); Fig. 3A). Eosinophil-deficient mice undergoingCAC had an increased tumor load (Fig. 3B), as determined by theincreased total tumor counts and size (Fig. 3C andD).Despite ourefforts to assess whether eosinophil reconstitution will decreasetumor load, and similar to previously published data (24),maximum eosinophil reconstitution reached only 50% in com-parison with the levels of eosinophils in WT mice undergoingCAC. Thus, as an alternative approach, we examined tumor loadin hypereosinophilic Il5Tg mice. Il5Tg mice undergoing CAC dis-played decreased tumor burden following the induction of CAC(Supplementary Fig. S6).

To further establish the antitumorigenic activity of eosinophilsin colorectal cancer, we aimed to determine their role in Apcmin/þ

mice. Apcmin/þ/DdblGATA mice displayed approximately a 3-foldincrease in cancer-associated mortality (Fig. 3E). Apcmin/þ micedisplayed a mortality frequency of 13.8% (n¼ 41mice), whereasApcmin/þ/DdblGATAmice displayed 40.9% (n¼ 44mice; Fig. 3E, P< 0.01). Increased mortality in Apcmin/þ/DdblGATA mice wasaccompanied by an increase in tumor load (Fig. 3F and G).

Antitumorigenic activities of eosinophils in colorectal cancerare independent of CD8þ T cells

Eosinophilsmay orchestrate tumor rejection in part by enhanc-ing the infiltration of CD8þ T cells (11). However, immunephenotyping of the cellular infiltrate that was present inthe colonic tissue in DdblGATA mice undergoing CAC and inApcmin/þ/DdblGATA mice revealed that increased tumor load inthe absence of eosinophils was not associatedwith any alterationsin CD8þ T cells or myeloid-derived suppressor cells (Supplemen-tary Fig. S7).We assessedwhether the antitumorigenic activities ofeosinophils in colorectal cancer were dependent on CD8þ T cells.

Figure 2.Eosinophils are recruited to the TME, which fosters their prolonged survival. The presence of eosinophils was determined using anti-eosinophil MBP staining incolonic (A–C) and ileal (D–E,G) sections, which were obtained at the indicated time points frommice undergoing colitis-associated cancer (A–C) or fromApcmin/þmice at the age of 5 months (D–E, G). Intratumoral MBPþ cells were identified following colitis-associated cancer (B) and in Apcmin/þ adenomas (E). Thepresence of intratumoral eosinophils was confirmed by flow cytometry of single-cell suspensions obtained from isolated adenomas (F). Eosinophils wereidentified as CD45þ/CD11bþ/Siglec-Fþ/MHC-II�/Ly6g�/Ly6c�/SSChi cells. The percentage of peripheral blood (H, I) eosinophils out of the total CD45þ cells wasdetermined throughout colitis-associated cancer (H) and in 3- and 5-month-old Apcmin/þmice (I). Data represent 4–5 mice per time point; �� , P < 0.01;�� , P < 0.001; ns, nonsignificant. Highly purified eosinophils from the peritoneal cavity of Il5Tgmice were adoptively transferred into na€�ve DdblGATAmice(DdblGATA), DdblGATAmice undergoing colitis-associated cancer or DdblGATA/Apcmin/þmice using a single time point injection protocol (J). Thereafter, themice were euthanized at the indicated time points and colonic (K–L) and ileal (M–N) eosinophils were assessed. Data in K and L represent n¼ 5 experimentsconducted with 3–5 mice per group. Data inM–N are from n¼ 3 with 3–5 mice per group; �� , P < 0.01. Eosinophil survival following overnight incubation in MC38colorectal cancer cell conditionedmedia (CM) was determined by flow cytometry (O). Data are from n¼ 3; ns, not significant; � , P < 0.05; �� , P < 0.01;�� , P < 0.001.






Therefore, Apcmin/þ and Apcmin/þ/DdblGATA mice were treatedwith antibodies to deplete CD8þ T cells (25). Depletion of CD8þ

T cells in Apcmin/þ mice (Supplementary Fig. S8) resulted in 16%tumor-associated death in comparison with isotype control-treated mice, who showed no mortality (Fig. 3H). Isotypecontrol–treated Apcmin/þ/DdblGATA mice displayed 50%tumor-associateddeath,whichwas further increased to66%upondepletion of CD8þ T cells (Fig. 3H). Tumor burden was increasedin anti–CD8-treatedApcmin/þ/DdblGATAmice in comparisonwithanti–CD8-treated Apcmin/þ mice (Fig. 3I). Collectively, these dataindicate that eosinophils display antitumorigenic activities incolorectal cancer independent of CD8þ T cells.

Transcriptome signatures of intratumoral eosinophils andmacrophages

To identify potential mechanisms that mediate the antitumori-genic activities of eosinophils in colorectal cancer, we definedtheir transcriptional signature. We compared the transcriptionalsignatures of intratumoral eosinophils with those of macro-phages, which have been better characterized. To this end, na€�vecolonic eosinophils and macrophages as well as intratumoraleosinophils and macrophages were sorted (purity of both cell

types was �95%) and subjected to RNA-seq following inductionof CAC. Principal component analysis (PCA) revealed four dis-tinct cellular populations: na€�ve colonic eosinophils, na€�ve colon-ic macrophages, intratumoral eosinophils, and intratumoralmacrophages (Fig. 4A). In comparison with na€�ve eosinophils,intratumoral eosinophils displayed multiple differentiallyexpressed transcripts. Of these, 587 transcripts were upregulatedand438downregulated [fold change>2;P value adjusted for falsediscovery rate (FDR) < 0.05; Fig. 4B; Supplementary Tables S2 andS3]. Unbiased STRING analysis, which identifies known andpredicted protein interactions (26), revealed that the transcrip-tome signature of intratumoral eosinophils was associated with aproinflammatory phenotype and was divided into three clusters(Fig. 4C; Supplementary Tables S4–S6). Cluster 1 included enrich-ment of transcripts that represent a response to interferons andregulation of TLR signaling (e.g., Ifit1, Irf1, Irf7, Irf9, Ifi47, Ifitm1,Nos2, and Stat1; FDR¼2.37�10�13). Cluster 2was enrichedwithtranscripts associated with chemokines and cell migration. Clus-ter 3 comprised transcripts associated with inflammatory andinnate immune responses (e.g., Ccl2, Cd14, Cd19, Il12b, Tlr2,Il17a, and Traf1). Consistent with our STRING analysis, visuali-zation and integrated discovery (DAVID) annotation analysis of

Figure 3.

Eosinophils mediate antitumorigenic activities in colorectal cancer independent of CD8þ T cells. Colitis-associated cancer-inducedmortality was determinedinWT and DdblGATA (A) n¼ 30–32mice, P < 0.05. Representative photos of colonic adenomas (B) and quantitative assessment of tumor load (C), sizeand number (D) in WT and DdblGATA undergoing colitis-associated cancer are shown. Data are from n¼ 4 experiments with 8–21 mice per group;�� , P < 0.001. Colorectal cancer–associated mortality was determined in Apcmin/þ and DdblGATA/Apcmin/þmice. E, n¼ 41–44mice; P < 0.01. Representativephotos of ileal adenomas (F) and quantitative assessment of tumor load are shown (G). Data are from n¼ 4 with experiments with 6–17 mice per group;� , P < 0.05. Colorectal cancer–associated mortality (H) and representative photos of ileal adenomas (I) in Apcmin/þ and DdblGATA/Apcmin/þmice followingdepletion of cytotoxic T cells using anti-CD8 antibodies (aCD8;H). n¼ 6–9mice.

Reichman et al.





gene ontology (GO) pathways revealed enrichment of inflam-matory pathways in intratumoral eosinophils including defenseand innate immune responses to various stimuli involving bac-teria, stress, lipids, and external stimuli such as cytokines, specif-ically IFNg (Fig. 4D). Intratumoral eosinophils also displayedincreased mRNA expression of the antiapoptotic molecule Bcl2l1and decreased mRNA expression of the proapoptotic moleculeCasp3 (Supplementary Tables S2 and S3).

A similar analysis of the macrophage transcriptome revealedthat in comparison with na€�vemacrophages, intratumoralmacro-phages displayed 1,392 transcripts that were differentiallyexpressed. Of these, 654 were upregulated and 738 were down-regulated (fold change > 2; P value adjusted for FDR < 0.05;

Supplementary Tables S7 and S8). Intratumoral eosinophils andmacrophages displayed distinct phenotypes, because only 270(26% of the eosinophils and 19% of the macrophage transcriptsignatures, respectively) were shared between eosinophils andmacrophages (Fig. 4E; Tables S7–S9). Analysis of these 270 sharedtranscripts in intratumoral eosinophils andmacrophages revealeddiscrete expression patterns between the two cells (Fig. 4F andG).For example, expression of eosinophil-associated ribonuclease 2(Ear2) was increased in eosinophils and decreased in macro-phages, whereas vascular endothelial growth factor (Vegf)decreased in eosinophils but increased in macrophages(Fig. 4G). Transcripts that are associated with innate immuneactivation such as formyl peptide receptor 1 (Fpr1) andCd14were

Figure 4.

Intratumoral eosinophils and macrophages display distinct transcriptome signatures. Na€�ve colonic and intratumoral eosinophils and macrophageswere sorted from mice undergoing colitis-associated cancer and subjected to RNA-seq (n ¼ 2 for na€�ve groups and 6 for intratumoral groups). PCAof the different experimental groups is shown (A). The transcriptome signature of na€�ve (N1–2) colonic eosinophils and intratumoral (C1–6)eosinophils is shown (B; fold change > 2, P value adjusted for FDR < 0.05). Significantly upregulated transcripts were subjected to STRING analysis(C). Furthermore, the transcriptome signature of intratumoral eosinophils was subjected to bioinformatics analysis using the database for annotation,visualization, and integrated discovery (DAVID) and annotation of gene ontology (GO) pathways (D). The transcriptome signature of intratumoraleosinophils was compared with that of macrophages using a Venn diagram (E) and heat plot analyses (F). G, A comparison of the expression ofselected transcripts from intratumoral eosinophils and macrophages is shown. The transcriptome signature of intratumoral macrophages wassubjected to bioinformatics analysis using the DAVID and the annotation of GO pathways (H).






increased to a greater extent in eosinophils, whereas transcriptsthat were associated with tissue repair, such as matrix metallo-proteinase 7 (Mmp7) and arginase 2 (Arg2), increased to a greaterextent in macrophages. The resistin-like molecule a (Retnla),a hallmark of alternatively activated macrophages, whichresemble tumor-associated macrophages (27), was decreasedin eosinophils and macrophages, although it was downregu-lated to a greater extent in macrophages (Fig. 4H; a completetranscript list of genes that are differentially regulated in macro-phages and eosinophils can be found in Supplementary TableS8). Bioinformatics analysis, using the DAVID annotation ofGO pathways, revealed enrichment of pathways that are asso-ciated with developmental and tissue repair processes in intra-tumoral macrophages (Fig. 4H).

IFNg as a key activator of eosinophils in colorectal cancerNext, we characterized the proteomic profile of intratumoral

eosinophils following CAC by means of antibody array (28).PCA analysis and subsequent hierarchical clustering segregatedthe samples according to cell type and treatment (Fig. 5A).Hierarchal clustering revealed that all four cell groups segre-gated distinctly (Fig. 5A, bottom tree cluster). However, na€�vecolonic eosinophils were distinctly separated from na€�vemacrophages as well as intratumoral macrophages and eosi-nophils (Fig. 5A, bottom tree cluster). Intratumoral and na€�ve

eosinophils differentially expressed 155 proteins. Of theseproteins, 49 were abundant in na€�ve eosinophils and 106 wereabundant in intratumoral eosinophils (Fig. 5B and see list inSupplementary Table S10). STRING analysis of the differen-tially regulated proteins in intratumoral eosinophils revealedthat they were grouped into three unique clusters (Fig. 5C).Cluster 1 comprised proteins associated with cell survival (e.g.,BAX, BCL2, caspase-3); cluster 2 included the enrichment ofvarious cytokines and cell-surface receptors (e.g., IL4, IFNg ,IL12b, CD44, and CD79); cluster 3 comprised growth factorsand enzymes that can regulate the integrity of the extracellularmatrix (e.g., MMP7, TIMP-1, and FGF7). Comparison of theproteomic signature, which was retrieved for eosinophils,revealed that only 40 of 155 proteins (25% of the eosinophilproteomic signature) were shared between intratumoral macro-phages and eosinophils (Fig. 5D; Supplementary Tables S1–S13). Although these 40 proteins were the majority of theproteins detected in macrophages, further analysis revealedthat the expression pattern (i.e., increased vs. decreased expres-sion) of 90% of them (36 proteins) was higher in intratumoraleosinophils in comparison with macrophages (Fig. 5E). Con-sistent with our transcriptome data, unbiased bioinformaticsanalysis, with DAVID annotation of GO pathways, revealedthat the pathway most enriched in intratumoral eosinophilswas associated with IFNg signaling (Fig. 5F).

Figure 5.

Proteomic analysis of tumor-associated eosinophils in colorectal cancer. Protein lysates of na€�ve colonic and intratumoral eosinophils and macrophages (n¼ 3for na€�ve groups and 2 for intratumoral groups) were subjected to scioDiscover antibody microarrays. PCA of the different experimental groups is shown (A). Avolcano plot depicting the distribution of proteins, which were differentially expressed in na€�ve and intratumoral eosinophils, is shown (B). Differentiallyexpressed proteins were subjected to STRING analysis, and the identified clusters were circled (C). Venn plot analysis (D) of differentially regulated proteins inintratumoral eosinophils and macrophages (D) as well as representative expression (E) of selected proteins from these cells is shown. The proteomic pathwaysignature of intratumoral eosinphils, as analyzed by the database for annotation, visualization, and integrated discovery (DAVID) and the annotation of GOpathways is shown (F).

Reichman et al.





IFNg potentiates eosinophil-mediated colorectal cancer cellkilling

The increased abundance of IFNg-regulated pathways in intra-tumoral eosinophils suggested a role for IFNg in their antitumori-genic activities.

Coculture of eosinophils with MC38 or CT26 cells resulted inincreased colorectal cancer cell death without eosinophil activa-tion (Fig. 6A). Nonetheless, IFNg enhanced the ability of eosi-nophils to kill colorectal cancer cells in vitro (Fig. 6B). Increasedcytotoxic activities of eosinophils in response to IFNg were not

Figure 6.

IFNg potentiates eosinophil-mediated killing of colorectal cancer (CRC) cells. Eosinophils were cocultured with CT26 and MC38 colorectal cancer cells at theindicated ratios. Thereafter, the survival of colorectal cancer cells was determined by flow cytometry (A). Eosinophils were activated with IFNg (B) or additionalstimuli (C) and cocultured with MC38 (B, C), CT26 colorectal cancer cells (B) as well as with PyMT breast cancer cells (D) at a 20:1 ratio and the percentage ofdead tumor cells was determined (B,D). Data inA–D represent n¼ 5 experiments; ns, nonsignificant; � , P < 0.05; �� , P < 0.01; �� , P < 0.01. Human peripheralblood eosinophils were activated with IFNg and cocultured with SW-480/HCT-116 colorectal cancer cells (E). Thereafter, tumor cell viability was determined byflow cytometry. Data represent n¼ 3; � , P < 0.05; �� , P < 0.01. Tumors were excised from Apcmin/þmice andmice undergoing CAC and stained with anti-eosinophil MBP, anti-cleaved caspase-3, and DAPI (F–G). Representative photomicrographs of MBPþ cells residing in the proximity of cleaved caspase-3þ cellsare shown (F–G). Colonic and ileal samples obtained fromWT and DdblGATA undergoing colitis-associated cancer as well as from ApcMin/þ and ApcMin/þ/DdblGATAmice, respectively, were stained with anti-cleaved caspase-3 (H). Quantitative analysis of cleaved caspase-3 staining in the GI tract (I–J) are shown.Data are from n¼ 6–10 mice; � , P < 0.05.






due to increased eosinophil death in vitro, because nonstimulatedand stimulated eosinophils displayed�12% cell death. Increasedcytotoxicity of eosinophils toward colorectal cancer cells wasspecifically enhanced by IFNg because stimulation of eosinophilswith additional stimuli such as E. coli, poly I:C, TNFa, andf-met-leu-phe (FMLP) did not increase eosinophil cytotoxicity(Fig. 6C). Eosinophil-mediated cytotoxicity is likely not gen-eralized to all tumor cells, because eosinophils did not inducecytotoxicity in vitro toward PyMT breast cancer cells (Fig. 6D).

Similar to ourfindingswithmouse eosinophils, humanperiph-eral blood eosinophils cocultured with SW480 colorectal cancercells induced cell death, an effect potentiated by activation ofeosinophils with IFNg (Fig. 6E).

Antitumorigenic activities of eosinophils are associated withcytotoxicity

Dual immunofluorescence staining of MBP and cleavedcaspase-3 in WT mice undergoing CAC demonstrated intratu-moral eosinophils in the vicinity of apoptotic (i.e., activecaspase-3þ) tumor cells (Fig. 6F and G). The majority ofeosinophils (>99%) were negative for active caspase-3 staining,confirming our observation regarding prolonged eosinophilsurvival in the TME (Fig. 2).

Moreover, tissues from WT and DdblGATA mice undergoingCAC as well as tissues from Apcmin/þ and DdblGATA/Apcmin/þ

mice were obtained and stained with anticleaved caspase-3. Thenumber of cleaved caspase-3þ cells was decreased in the absenceof eosinophils both in colitis-associated cancer and in Apcmin/þ

mice (Fig. 6H–J). No differences, however, were observed in thenumbers of Ki67þ epithelial cells (Supplementary Fig. S9) and/orCD31þ blood vessels (Supplementary Fig. S9).

DiscussionOur perceptions regarding the roles of eosinophils in health

and disease have changed because functions for these cells havebeen identified in settings that are beyond classic type-2 immunity(29, 30). In this study, we dissected the roles of eosinophils incolorectal cancer. We demonstrate that elevated eosinophilia inhuman colorectal cancer was associatedwith an improved diseasestage, suggesting beneficial roles for eosinophils in colorectalcancer. Experimentally, eosinophils were recruited to the TME,which supported prolonged eosinophil survival andCD8þT-cell–independent antitumorigenic activities. Transcriptome and prote-omic analysis of intratumoral eosinophils revealed an activatedeosinophil phenotype, which was associated with IFNg signaling.These data provide insight into the transduction mechanismsunleashing antitumor activities from eosinophils and identifythese cells as targets for future immunotherapy especially incolorectal cancer.

Our analyses revealed that tumor eosinophilia in colorectalcancer was inversely correlated with tumor grade, in line withstudies that have associated tumor eosinophilia with improvedoverall and/or colorectal cancer–specific survival (13, 14, 31–34).Furthermore, we demonstrated that 27% of colorectal cancerpatients with lymph node metastasis displayed lymph nodeeosinophilia. Although previous reports in colorectal cancer didnot directly assess eosinophils in metastatic sites such as thelymph node, eosinophilia in the primary site was inverselycorrelated with local recurrence and distant metastases, andtumors with more tumor eosinophilia displayed less metastasis

(35). Future studies will be required to assess the role of eosino-phils in lymph nodes and distant metastatic sites during tumorprogression in colorectal cancer.

We established that eosinophils display prolonged survival inthe TME independently of IL5 and confirmed this with adoptivetransfer experiments and the sparse number of active caspase-3þ

eosinophils in the GI tract. Consistent with previous reports inmurinemodels of allergic GI inflammation and IBD (36–38), theeosinophil-specific chemokine CCL11/eotaxin-1 (but notCCL24/eotaxin-2) was increased during the progression of colo-rectal cancer. Increased CCL11/eotaxin-1 was associated withaccumulation of eosinophils in human colorectal cancer (39),although CCL24/eotaxin-2 may also be involved (40). The find-ing that the TME supports eosinophil accumulation and survivalsuggests that eosinophil degranulation in vivo and eosinophil-mediated tumor elimination is not necessarily a consequence ofeosinophil cell death.

Eosinophils can eliminate tumor cells by direct and indirectmechanisms (41). Under appropriate settings, eosinophils arehighly suited for eliminating tumors (42–46). It is unclear,however, what signaling mechanisms induce eosinophils to dis-play antitumorigenic functions. Our unbiased empiricalapproach, subjecting isolated primary intratumoral eosinophilsto RNA-seq and proteomics, revealed that tumor-infiltratingeosinophils in colorectal cancer displayed an IFNg-associatedsignature with multiple innate immune–signaling components,such as pattern recognition receptors and IFNg-dependent genes(e.g., Stat1, Ifi202b, Fpr1, Fpr2, Rtp4, Nos2, Slfn4, Ifit1bl2, ifitm1,and ifit3b). Unbiased proteomic and subsequent bioinformaticsanalysis substantiated the IFNg-associated signature by revealingthat the top-enriched pathway (with a P value of 7.83–10�4) wasIFNg-dependent signaling. A reported microarray analysis ofeosinophils differentiated ex vivo with recombinant IL18, whichinduces IFNg production, suggests a distinct gene-expressionsignature characteristic of "inflammatory eosinophils" (47).These inflammatory eosinophils display increased expression ofseveral transcripts that we observed in intratumoral eosinophils,including Cd274, Saa3, Ly6a, Ifit3, Oasl2, Spp1, and Rtp4. More-over, Il18bp, anegative regulator of IL18,which is induced by IFNg(48), is also increased in intratumoral eosinophils. The IFNgsignature, which we identified by means of RNA-seq and prote-omics, was functionally validated by activation of mouse andhuman eosinophils with IFNg , and consequently, the cytotoxicitytoward colorectal cancer cells increased. The finding that IFNg-activated eosinophils displayed tumoricidal activities suggeststhat, in colorectal cancer, eosinophils display an activated phe-notype that resembles that of classically activated macrophages(also termed M1 cells), which also have the ability to kill tumorcells (49). Indeed, IFNg-activated eosinophils are capable ofreleasing reactive oxygen species, mitochondrial DNA, and nitricoxide, which are capable of killing tumor cells (17, 50). Com-paring the transcriptome and proteome signature of tumor-infiltrating eosinophils to that of macrophages revealed thatalthough these two cell types were exposed to similar stimuliin the TME, their responses differ. Eosinophils were polarizedinto an inflammatory state, with increased proinflammatorycytokines, chemokines, and signaling pathways. In contrast,macrophages displayed a phenotype that was associated withtissue repair and organ development, with increased expression ofgrowth factors and matrix metalloproteinases. Future researchis needed to determine whether, while retaining their

Reichman et al.





antitumorigenic activities, eosinophil-derived IL4/IL13 contri-butes to the suppressive function of macrophages in the TME.

Our data indicate that the antitumorigenic activities of eosi-nophils in colorectal cancer are independent of CD8þ T cells andare associated with tumor cell death. In our models, the anti-tumorigenic function of eosinophils wasmore potent than that ofCD8þ T cells. This conclusion is important because clinical datashow that current T-cell–based immunotherapies have limitedsuccess for most colorectal cancer patients (42). The finding thatthe antitumorigenic activities of eosinophils are independent ofCD8þ T cells is in contrast to a report showing that eosinophilscoordinate antitumor immunity via recruitment of CD8þ T cells(11). In that system, eosinophils recruit CD8þ T cells and rendermacrophages antitumorigenic via secretion of CXCL9 and pro-motion of an environment characterized by elevated IFNg andTNFa (11).We also identified increased expression of IFNg ,Cxcl9,and additional IFNg-associated genes in intratumoral eosino-phils. In addition, experimental models of GI infection also showincreased activity of eosinophils in response to IFNg . Whereas ininfectious disease settings, IFNg-dependent expression of PDL-1on eosinophils restricted Th1-induced immune responses (43), acollective view of these data highlight IFNg as a key activator ofeosinophils, especially in the GI tract.

The association between the antitumorigenic activities of eosi-nophils in response to IFNg is consistent with previous reportsregarding the role of IFNg in colorectal cancer. Given the longevityof experimental models for colorectal cancer (3–5 months) andthe accumulation of multiple cells that may differentially expressIFNg , future experiments are required to identify the cellularsource of the IFNg and the in vivo role of IFNg signaling in theantitumorigenic activities of eosinophils. Nonetheless, the find-ing that eosinophils can mediate antitumor activities indepen-dent of CD8þ T cells may have therapeutic implications forcombinatorial therapies targeting these cells.

Oneof the conclusions thatwe candraw fromour studies is thatthe role of eosinophils in the TME is largely tissue- and context-dependent. Although this concept is accepted for other immunecells such as neutrophils and macrophages (44, 45), to date,eosinophils have been examined from a dichotomic point ofview, leading to the overall notion that the roles of eosinophils incancer are controversial (41, 46). On the basis of our data, wesuggest an alternative explanation, whereby the differential rolesof tumor immunology involving eosinophils (i.e., pro- vs. anti-tumorigenic activities) are dependent upon the TME. Undersettings in which eosinophils are exposed to innate immune

stimuli and in the presence of IFNg , they will be polarized todisplay antitumorigenic activities. In contrast, settings lackingIFNg or innate immune activation may polarize eosinophils toproduce tumor-promoting factors.

In summary, we provided evidence that, in colorectal cancer,eosinophils have antitumorigenic activity in vivo and their func-tions can be distinguished from cytotoxic T cells and intratumoralmacrophages. These data enhance our understanding of themolecular pathways regulating tumor eosinophilia. Our findingshave implications for cancer therapy, in particular for patientswith colorectal cancer.

Disclosure of Potential Conflicts of InterestA. Munitz is a consultant/advisory board member for GSK and Augmanity

Nano Ltd. No potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: H. Reichman, M. Itan, N. Arber, A. Munitz, J.J. LeeDevelopment of methodology:H. Reichman, M. Itan, E. Brazowski, A. MunitzAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): E. Brazowski, C. Varol, N. Gluck, N. ArberAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): H. Reichman, M. Itan, E. Brazowski, C. Varol,A. MunitzWriting, review, and/or revision of the manuscript: H. Reichman, C. Varol,N. Arber, U. Qimron, D. Karo-Atar, J.J. Lee, A. MunitzAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): P. Rozenberg, T. Yarmolovski, S. ShapiraStudy supervision: A. Munitz

AcknowledgmentsA. Munitz is supported by the US-Israel Bi-national Science Foundation

(grant nos. 2009222 and 2011244), the Israel Science Foundation(grant no. 886/15), a project grant from the Israel Cancer Research Foun-dation, the Israel Cancer Association (grant no. 20150002), the IsraelMinistry of Health (grant no. 3-10117), and the Boaz and Varda DotanCenter Grant for Hemato-oncology Research. H. Reichman was funded inpart by the Constantiner Institute for Molecular Genetics and performed thiswork in partial fulfillment of the requirements for a PhD degree at the SacklerFaculty of Medicine, Tel Aviv University, Israel. The authors wish to thankProfessor Marc Rothenberg for critically reviewing this manuscript.

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 23, 2018; revised October 14, 2018; accepted December 17,2018; published first January 21, 2019.

References1. Hanahan D, Coussens LM. Accessories to the crime: functions of cells

recruited to the tumor microenvironment. Cancer Cell 2012;21:309–22.2. Zhang N, Bevan MJ. CD8(þ) T cells: foot soldiers of the immune system.

Immunity 2011;35:161–8.3. Pardoll DM. The blockade of immune checkpoints in cancer immuno-

therapy. Nat Rev Cancer 2012;12:252–64.4. Rothenberg ME, Hogan SP. The eosinophil. Annu Rev Immunol 2006;

24:147–74.5. Reichman H, Karo-Atar D, Munitz A. Emerging roles for eosinophils in the

tumor microenvironment. Trends Cancer 2016;2:664–75.6. Simson L, Ellyard JI, Dent LA, Matthaei KI, Rothenberg ME, Foster PS,

et al. Regulation of carcinogenesis by IL-5 and CCL11: a potential rolefor eosinophils in tumor immune surveillance. J Immunol 2007;178:4222–9.

7. Kataoka S, Konishi Y,Nishio Y, Fujikawa-Adachi K, TominagaA. Antitumoractivity of eosinophils activated by IL-5 and eotaxin against hepatocellularcarcinoma. DNA Cell Biol 2004;23:549–60.

8. Mattes J, Hulett M, Xie W, Hogan S, Rothenberg ME, Foster P, et al.Immunotherapy of cytotoxic T cell-resistant tumors by T helper 2 cells:an eotaxin and STAT6-dependent process. J Exp Med 2003;197:387–93.

9. Lucarini V, Ziccheddu G, Macchia I, La Sorsa V, Peschiaroli F, Buccione C,et al. IL-33 restricts tumor growth and inhibits pulmonary metastasis inmelanoma-bearing mice through eosinophils. Oncoimmunology 2017;6:e1317420.

10. Zhang B,Wei CY, Chang KK, Yu JJ, ZhouWJ, YangHL, et al. TSLP promotesangiogenesis of human umbilical vein endothelial cells by strengtheningthe crosstalk between cervical cancer cells and eosinophils. Oncol Lett2017;14:7483–8.






11. Carretero R, Sektioglu IM, Garbi N, Salgado OC, Beckhove P,H€ammerling GJ. Eosinophils orchestrate cancer rejection by normaliz-ing tumor vessels and enhancing infiltration of CD8(þ) T cells. NatImmunol 2015;16:609–17.

12. Tepper RI, Coffman RL, Leder P. An eosinophil-dependent mechanism forthe antitumor effect of interleukin-4. Science 1992;257:548–51.

13. Harbaum L, Pollheimer MJ, Kornprat P, Lindtner RA, Bokemeyer C,Langner C. Peritumoral eosinophils predict recurrence in colorectal cancer.Mod Pathol 2015;28:403–13.

14. PrizmentAE, Vierkant RA, Smyrk TC, Tillmans LS, Lee JJ, SriramaraoP, et al.Tumor eosinophil infiltration and improved survival of colorectal cancerpatients: Iowa Women's Health Study. Mod Pathol 2016;29:516–27.

15. Yalcin AD, Kargi A, Gumuslu S. Blood eosinophil and platelet levels,proteomics patterns of trail and CXCL8 correlated with survival in bev-acizumab treated metastatic colon cancers. Clin Lab 2014;60:339–40.

16. Hogan SP, Rothenberg ME. Eosinophil function in eosinophil-associatedgastrointestinal disorders. Curr Allergy Asthma Rep 2006;6:65–71.

17. Yousefi S, Gold JA, Andina N, Lee JJ, Kelly AM, Kozlowski E, et al. Catapult-like release of mitochondrial DNA by eosinophils contributes to antibac-terial defense. Nat Med 2008;14:949–53.

18. RosenbergDW,Giardina C, Tanaka T.Mousemodels for the study of coloncarcinogenesis. Carcinogenesis 2009;30:183–96.

19. Moshkovits I, Reichman H, Karo-Atar D, Rozenberg P, Zigmond E, Haber-man Y, et al. A key requirement for CD300f in innate immune responses ofeosinophils in colitis. Mucosal Immunol 2017;10:172–83.

20. Benjamini Y, Hochberg Y. On the adaptive control of the false discoveryrate in multiple testing with independent statistics. J Educ Behav Statist2000;25:60–83.

21. ProtheroeC,Woodruff SA, de Petris G,Mukkada V,Ochkur SI, JanarthananS, et al. A novel histologic scoring system to evaluatemucosal biopsies frompatients with eosinophilic esophagitis. Clin Gastroenterol Hepatol2009;7:749–55e711.

22. McCart AE, Vickaryous NK, Silver A. Apc mice: models, modifiers andmutants. Pathol Res Pract 2008;204:479–90.

23. Yamaguchi Y, Hayashi Y, Sugama Y, Miura Y, Kasahara T, Kitamura S, et al.Highly purifiedmurine interleukin 5 (IL-5) stimulates eosinophil functionand prolongs in vitro survival. IL-5 as an eosinophil chemotactic factor. JExp Med 1988;167:1737–42.

24. Wen T, Besse JA, Mingler MK, Fulkerson PC, Rothenberg ME. Eosinophiladoptive transfer system to directly evaluate pulmonary eosinophil traf-ficking in vivo. Proc Natl Acad Sci USA 2013;110:6067–72.

25. Carmi Y, Spitzer MH, Linde IL, Burt BM, Prestwood TR, Perlman N, et al.Allogeneic IgG combined with dendritic cell stimuli induce antitumourT-cell immunity. Nature 2015;521:99–104.

26. SzklarczykD,Morris JH,CookH,KuhnM,Wyder S, SimonovicM, et al. TheSTRING database in 2017: quality-controlled protein–protein associationnetworks, made broadly accessible. Nucleic Acids Res 2017;45:D362–8.

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

28. Sill M, Schr€oder C, Shen Y, Marzoq A, Komel R, Hoheisel JD, et al. Proteinprofiling gastric cancer and neighboring control tissues using high-contentantibody microarrays. Microarrays 2016;5:pii: E19.

29. Rosenberg HF, Dyer KD, Foster PS. Eosinophils: changing perspectives inhealth and disease. Nat Rev Immunol 2013;13:9–22.

30. Lee JJ, Jacobsen EA, McGarry MP, Schleimer RP, Lee NA. Eosinophilsin health and disease: the LIAR hypothesis. Clin Exp Allergy 2010;40:563–75.

31. Nielsen HJ, Hansen U, Christensen IJ, Reimert CM, Br€unner N, MoesgaardF. Independent prognostic value of eosinophil andmast cell infiltration incolorectal cancer tissue. J Pathol 1999;189:487–95.

32. Prizment AE, Anderson KE, Visvanathan K, Folsom AR. Inverse associationof eosinophil countwith colorectal cancer incidence: atherosclerosis risk incommunities study. Cancer Epidemiol Biomarkers Prev 2011;20:1861–4.

33. Saraiva AL, Carneiro F. New insights into the role of tissue eosinophils inthe progression of colorectal cancer: a literature review. Acta Med Port2018;31:329–37.

34. Fernandez-Acenero MJ, Galindo-Gallego M, Sanz J, Aljama A. Prognosticinfluence of tumor-associated eosinophilic infiltrate in colorectal carcino-ma. Cancer 2000;88:1544–8.

35. Nagtegaal ID, Marijnen CA, Kranenbarg EK, Mulder-Stapel A, Hermans J,van de Velde CJ, et al. Local and distant recurrences in rectal cancer patientsare predicted by the nonspecific immune response; specific immuneresponse has only a systemic effect–a histopathological and immunohis-tochemical study. BMC Cancer 2001;1:7.

36. Hogan SP,MishraA, Brandt EB, Foster PS, RothenbergME.A critical role foreotaxin in experimental oral antigen-induced eosinophilic gastrointestinalallergy. Proc Natl Acad Sci USA 2000;97:6681–6.

37. Ahrens R, Waddell A, Seidu L, Blanchard C, Carey R, Forbes E, et al.Intestinal macrophage/epithelial cell-derived CCL11/eotaxin-1 mediateseosinophil recruitment and function in pediatric ulcerative colitis. JImmunol 2008;181:7390–9.

38. Hogan SP, Mishra A, Brandt EB, Royalty MP, Pope SM, Zimmermann N,et al. A pathological function for eotaxin and eosinophils in eosinophilicgastrointestinal inflammation. Nat Immunol 2001;2:353–60.

39. WagsaterD, Lofgren S,Hugander A,DienusO,Dimberg J. Analysis of singlenucleotide polymorphism in the promoter and protein expression of thechemokine eotaxin-1 in colorectal cancer patients. World J Surg Oncol2007;5:84.

40. Cho H, Lim SJ, Won KY, Bae GE, Kim GY, Min JW, et al. Eosinophils incolorectal neoplasms associated with expression of CCL11 and CCL24. JPathol Transl Med 2016;50:45–51.

41. Gatault S, Legrand F, Delbeke M, Loiseau S, Capron M. Involvement ofeosinophils in the anti-tumor response. Cancer Immunol Immunother2012;61:1527–34.

42. Emambux S, Tachon G, Junca A, Tougeron D. Results and challenges ofimmune checkpoint inhibitors in colorectal cancer. Expert Opin Biol Ther2018;18:561–73.

43. Arnold IC, Artola-Bor�anM, Tall�on de Lara P, Kyburz A, Taube C,OttemannK, et al. Eosinophils suppress Th1 responses and restrict bacterially inducedgastrointestinal inflammation. J Exp Med 2018;215:2055–72.

44. Sionov RV, Fridlender ZG, Granot Z. The multifaceted roles neutrophilsplay in the tumor microenvironment. Cancer Microenviron 2015;8:125–58.

45. Qian BZ, Pollard JW. Macrophage diversity enhances tumor progressionand metastasis. Cell 2010;141:39–51.

46. Davis BP, Rothenberg ME. Eosinophils and cancer. Cancer Immunol Res2014;2:1–8.

47 Venkateshaiah SU, Mishra A, Manohar M, Verma AK, Rajavelu P, NiranjanR, et al. A critical role of IL-18 in transformation and maturation of naiveeosinophils to pathogenic eosinophils. J Allergy Clin Immunol 2018;142:301–5.

48. Carbotti G, Barisione G, Orengo AM, Brizzolara A, Airoldi I, Bagnoli M,et al. The IL-18 antagonist IL-18-binding protein is produced in the humanovarian cancer microenvironment. Clin Cancer Res 2013;19:4611–20.

49. Allavena P, Sica A, Garlanda C, Mantovani A. The yin-yang of tumor-associated macrophages in neoplastic progression and immune surveil-lance. Immunol Rev 2008;222:155–61.

50. Phipps S, LamCE,Mahalingam S, NewhouseM, Ramirez R, Rosenberg HF,et al. Eosinophils contribute to innate antiviral immunity and promoteclearance of respiratory syncytial virus. Blood 2007;110:1578–86.


Reichman et al.




2019;7:388-400. Published OnlineFirst January 21, 2019.Cancer Immunol Res Hadar Reichman, Michal Itan, Perri Rozenberg, et al. Colorectal CancerActivated Eosinophils Exert Antitumorigenic Activities in

Updated version

10.1158/2326-6066.CIR-18-0494doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerimmunolres.aacrjournals.org/content/suppl/2021/03/13/2326-6066.CIR-18-0494.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerimmunolres.aacrjournals.org/content/7/3/388.full#ref-list-1

This article cites 50 articles, 12 of which you can access for free at:

Citing articles

http://cancerimmunolres.aacrjournals.org/content/7/3/388.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/7/3/388To 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-18-0494

http://cancerimmunolres.aacrjournals.org/content/suppl/2021/03/13/2326-6066.CIR-18-0494.DC1

http://cancerimmunolres.aacrjournals.org/content/7/3/388.full#ref-list-1

http://cancerimmunolres.aacrjournals.org/content/7/3/388.full#related-urls

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

mailto:[email protected]

http://cancerimmunolres.aacrjournals.org/content/7/3/388


Activated Eosinophils Exert Antitumorigenic Activities in ...Research Article Activated Eosinophils Exert Antitumorigenic Activities in Colorectal Cancer Hadar Reichman1, Michal Itan1,

Documents