the c.1085a g genetic variant of csf1r gene regulates tumor immunity … · cancer therapy:...

Cancer Therapy: Clinical

The c.1085A>G Genetic Variant of CSF1R GeneRegulates Tumor Immunity by Altering theProliferation, Polarization, and Function ofMacrophagesYu-Min Yeh1,2, Shan-Ju Hsu3, Peng-Chan Lin1,2, Keng-Fu Hsu1,4, Pei-Ying Wu4,Wu-Chou Su2, Jang-Yang Chang2,5, and Meng-Ru Shen3,4

Abstract

Purpose: Targeting tumor-associated macrophages with colo-ny-stimulating factor 1 receptor (CSF-1R) inhibition reveals astrategy for cancer therapy. Here, we studied the impact of CSF1Rgermline genetic variant on CSF-1R signaling and the suscepti-bility to CSF-1R inhibitors.

Experimental designs: CSF1R germline genetic variants werestudied in 140 cancer patients. CSF-1R phosphorylation, endo-cytosis, and macrophage polarization were measured as theresponse to CSF-1 stimulation. Tumor-associated macrophagesin surgical specimens and sensitivity to CSF-1R inhibitors wereused to determine macrophage function.

Results:ACSF1R c.1085A>Ggenetic variant causing the changeof histidine to arginine in the domain of receptor dimerizationwas identified as a high allele frequency in Eastern Asian popu-lation. Cancer patients with this variant allele had less M2-liketumor-associated macrophages accompanied by low VEGFexpression in tumor tissues. Importantly, CSF1R genetic variant

was significantly associated with disease-free survival in colorec-tal, endometrial, and ovarian cancer. In terms of differentiation,macrophages with CSF1R c.1085A>G genetic variant displayed arefractory response toCSF-1 stimulation andmacrophage survivalwas sensitive to CSF-1R inhibitors with IC50 of 0.1 to 1 nmol/Lrange. On contrast, CSF-1 induced a prominent phosphorylationand rapid endocytosis of CSF-1R, leading to anM2-like dominantpolarization inmacrophageswithCSF1R c.1085 genotype A_A, inwhich CSF-1R inhibitors of PLX3397, BLZ945, and GW2580inhibited macrophage survival with IC50 of 10 to 100 nmol/Lrange.

Conclusions: The CSF1R c.1085A>G genetic variant regu-lates tumor immunity by altering the polarization and func-tion of macrophages. This genetic variant confers the sensitiv-ity to CSF-1R inhibitors, implying as a biomarker in targetingCSF-1R signaling for cancer treatment. Clin Cancer Res; 23(20);6021–30. �2017 AACR.

IntroductionColony stimulating factor 1 receptor (CSF-1R) is the receptor

for CSF-1, a cytokine that controls the production, differentia-tion, and function of macrophages (1). Ligand binding activatesthe CSF-1R kinase through a process of oligomerization andtransphosphorylation. Two different polarized subpopulations

of macrophages, classically activated (M1) or alternatively acti-vated (M2) macrophage, have been identified. Exposure ofmacrophages to granulocyte/macrophage colony-stimulatingfactor (GM-CSF) leads to an M1-like state that produces thepro-inflammatory cytokines, such as tumor necrosis factor(TNF), IL6, and IL12, whereas exposure to CSF-1 leads macro-phages to be maintained in an M2-like polarized state (2). TheM2-like macrophages produce cytokines that could promotetumor progression, angiogenesis, and matrix remodeling.Tumor-associated macrophages are often found to have M2-likephenotype that is associated with a poor outcome in differenttypes of cancers (3). Thus, targeting CSF-1R signaling by mono-clonal antibody or small molecular inhibitors to manipulatemacrophage function has been studied in the treatment ofcancer, such as glioma (4), breast cancer (5), and pancreaticcancer (6). Different responses to CSF-1R inhibition are observedin the early-phase clinical trials and several clinical studies ofimmunotherapy are undergoing to evaluate the combination ofCSF-1R inhibition and blockade of PD-1/PD-L1 (7).

Because CSF-1R signaling controls the differentiation and sur-vival of macrophages, a critical question that remains to beanswered is whether mutations or nonsynonymous variants ofCSF1R gene would change macrophage function. For example,mutations in CSF1R gene affecting the tyrosine kinase domainwere found to interfere with the autophosphorylation of CSF-1R

1Graduate Institute of Clinical Medicine, National Cheng Kung University Hos-pital, College of Medicine, National Cheng Kung University, Tainan City, Taiwan.2Department of Internal Medicine, National Cheng Kung University Hospital,College of Medicine, National Cheng Kung University, Tainan City, Taiwan.3Department of Pharmacology, National Cheng Kung University Hospital, Col-lege of Medicine, National Cheng Kung University, Tainan City, Taiwan. 4Depart-ment of Obstetrics and Gynecology, National Cheng Kung University Hospital,College of Medicine, National Cheng Kung University, Tainan City, Taiwan.5National Institute of Cancer Research, National Health Research Institutes,Miaoli County, Taiwan.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Meng-Ru Shen, National Cheng Kung University Hos-pital, 138 Sheng-Li Road, Tainan 704, Taiwan. Phone: 886-6-2353535; Fax: 886-6-2766185; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-17-1007

�2017 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 6021

on July 23, 2021. © 2017 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst July 19, 2017; DOI: 10.1158/1078-0432.CCR-17-1007

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-17-1007&domain=pdf&date_stamp=2017-9-29

http://clincancerres.aacrjournals.org/

and cause hereditary diffuse leukoencephalopathy (8). By thegenome-wide association study, SNPofCSF1R gene, rs10079250,was significantly associated with lung cancer in never-smokingfemales (9). This rs10079250 polymorphism of CSF1R waslocated in the proximity of the binding sites of monoclonalantibody RG7155 and also correlated with a trend toward areduced response to the anti-CSF-1R antibody, RG7155 (10).

Here we surveyed the CSF1R germline genetic variants incolorectal, ovarian, and endometrial cancer and studied theimpact of the variants on macrophage function. The resultsshowed thatmacrophageswithCSF1R c.1085A>Ggenetic variant,a high incidence in Eastern Asian population, displayed a refrac-tory response to CSF-1 stimulation and susceptibility to CSF-1Rinhibitors.

Materials and MethodsDatabases for analyzing CSF1R genetic variants

Germline CSF1R genetic variants were analyzed from thedatabase of an ongoing prospective study and a pilot studyinvestigating the association of germline genetic variants andchemotherapy-induced peripheral neuropathy (CIPN; ref. 11).This clinical study was approved by the institutional review boardof National Cheng Kung University Hospital (NCKUH) andregistered in ClinicalTrial.gov. All patients provided writteninformed consent. This study was conducted in accordance withtheDeclaration ofHelsinki. In this CIPN study, patientswith stageIII colorectal cancer receiving the standard adjuvant chemother-apy with mFOLFOX6 and patients with ovarian and endometrialcancers receiving the postoperative chemotherapy with the regi-men of carboplatin and paclitaxel were enrolled in NCKUH inTaiwan. Among these patients, DNAwas isolated from peripheralblood and the whole genome was sequenced by next-generationsequencing on Illumina HiSeq 2500 to analyze the germlinegenetic variants. Till December 2016, whole genome data of140 subjects, including 90 cases of colorectal cancer, 28 cases ofovarian cancer, and 22 cases of endometrial cancer, were available

for analysis of CSF1R genetic variants. In addition, to study theassociation between CSF1R genetic variant and clinical outcome,germlineDNAwas extracted fromnormal tissue andwhole exomewas sequenced by using Ion Proton Sequencer (Thermal FisherScientific) in another 18 cases of advanced stage endometrialcancer. Whole genome sequencing data of 499 normal Taiwanesewere provided by Taiwan Biobank to compare the distributionand frequency of CSF1R variant between cancer patients andnormal population (12). Publicly available data from 1,000GenomesProjectwas used to compare theworldwidedistributionof CSF1R c.1085 genetic variant (13).

PCR and Sanger sequencingSanger sequencing was used to determine the CSF1R c.1085

genotypes of normal volunteers who provided peripheral bloodfor isolation of monocytes and confirm the CSF1R c.1085 geno-type of the 140 subjects from the CIPN study. Briefly, genomicDNA was extracted from buffy-coat of peripheral blood usingQIAGEN genomic DNA Purification Kit. DNA pellets were dis-solved in MQ water for PCR. Coding exons containing CSF1Rc.1085were PCR-amplifiedusing the forward and reverse primers:50-ACAGTGGTCAACGTAGGCGA-30 and 30-ATGAATGTCCATAT-GACGCTTACC-50. Reactions were amplified with the followingprotocol: 95�C denaturation for 5 minutes, followed by 40 cyclesof denaturation at 95�C for 30 seconds, annealing at 56�C for 30seconds, and extension at 72�C for 1 minute, followed by a 6-minute extension at 72�C. PCR products were sequenced todetermine the genotype of CSF1R c.1085.

Chemicals and antibodiesPLX3397, BLZ945, and GW2580 were purchased from Sell-

eckchem. Recombinant human CSF-1 was obtained from R&Dsystems. The antibodies used for immunofluorescent stainingwere: rabbit anti-TNF-a (Abcam, ab6671), mouse anti-VEGF(GeneTex, A7-E11-G2),mouse anti-CD68 (GeneTex,GTX41865),mouse anti-CD163 (GeneTex,GTX42365), rabbit anti-iNOS(CellSignaling Technology, 13120), and Hoechst 33258 (Sigma-Aldrich, 861405). Alexa Fluor 594- (A11012) and Alexa Fluro488-conjugated second antibody (A11001) were purchased fromThermo Fisher Scientific Inc. The antibodies used for analysis offlow cytometry were: FITC mouse anti-human CD14 (#555397),PE mouse anti-human CD80 (#557227), PE-Cy5 mouse anti-human CD206 (#551136), FITC mouse IgG2a k isotype control(#555573), PE mouse IgG1 k isotype control (#555749), andPE-Cy5 mouse IgG1 k isotype control (#555750) from BDPharmingen.

Immunofluorescence, confocal images, and tissue scanningTumor specimens of 52 colorectal, 20 ovarian, and 16 endo-

metrial cancer patients enrolled in the CIPN study were availablefor immunofluorescent staining. We used anti-iNOS, anti-CD163, and anti-CD68 to stain M1-like, M2-like, and totalmacrophages, respectively. Anti-TNFa and anti-VEGF were usedto stain M1- and M2-associated cytokines, respectively. AlexaFluor 594-conjugated second antibody was used for iNOS andTNFa; and Alexa Fluor 488-conjugated second antibody was usedfor CD68, CD163, and VEGF. Tissue was co-stained with Hoechst33258 todetect thenucleus. Thefluorophoreswere excites by laserat 405, 488, and 594 nm, respectively, and detected by a scanningconfocal microscope (FV-1000, Olympus). The numbers of M1-and M2-like macrophages were counted in five different fields

Translational Relevance

Colony-stimulating factor 1 receptor (CSF-1R) signalingregulates the survival, proliferation, and differentiation ofmacrophages which plays an important role in innate immu-nity, inflammatory disease, and cancer. Multiple approachesto target CSF-1R signaling for cancer treatment have beentested clinically. However, the impact of CSF1R genetic var-iants on CSF-1R signaling and predicting the susceptibility toCSF-1R inhibitor is not clear.We analyzed the germline geneticvariants of CSF1R gene and identified a nonsynonymousvariant c.1085A>G, a high allele frequency in Eastern Asian,which causes the change of amino acid from histidine toarginine in the domain of receptor dimerization.Macrophageswith CSF1R c.1085A>G genetic variant displayed a refractoryresponse to CSF-1 stimulation and susceptibility to CSF-1Rinhibitors. The different ethnic frequency of genetic varianttogether with differential sensitivity to CSF-1R inhibitorssuggests this genetic variant is a potential biomarker to predictthe response in targeting CSF-1R signaling for cancertreatment.

Yeh et al.

Clin Cancer Res; 23(20) October 15, 2017 Clinical Cancer Research6022




with40�oil immersion lens andM1/M2 ratiowas calculated. TheiNOS expression level of macrophages in confocal imaging wasanalyzed by FV10-ASW 4.0 microscopy software. To investigatethe total macrophages and cytokine expression, whole tissue wasscanned with TissueGnostics GmbH FACS-like Tissue Cytometry(TissueFAXS Plus). Series of separate images per fluorescencechannel and field of view were acquired automatically andmerged. HistoQuest software was used to analyze the positivestaining area of TNFa and VEGF in total tumor area. In eachsample, the number of macrophages (CD68-positive) cells wasquantitated in 10 fields of tumor microenvironment. Each fieldwas 1.2 � 1.2 mm2.

Culture of human monocyte-derived macrophagesMonocytes were enriched from whole blood by negative selec-

tion using the Lymphoprep (STEMCELL) according to the man-ufacturer's instruction. To differentiate these monocytes intomonocyte-derived macrophages, 1 � 107 cells were plated in 6-cmdish andmaintained in RPMI plus 10%FBSwith recombinanthuman CSF-1 10 ng/mL for 6 days.

Analysis and cell sorting by flow cytometryFor detection of cell surface markers and cell sorting, 1 mg of

monoclonal FITCmouse anti-human CD14 (BD Biosciences), PEmouse anti-human CD80 (BD Biosciences), PE-Cy5 mouse anti-human CD206 antibodies (BD Biosciences), or the relevantisotypes were incubated with samples containing 2 � 105 cellsfor 15 minutes at 4�C. Following incubation, samples werewashed and resuspended in PBS. Flow cytometric analysis wasperformed by using a four color flow cytometer CytoFLEX (Beck-man Coulter). Forward and side scatter light measuring the sizeand granularity of the cells, respectively, was used to gate thepopulation of macrophages. CD80 and CD206 were used tocharacterize M1- and M2-like macrophages, respectively. Tenthousand events were recorded and the data were analyzed usingFlowJo software, version 10.1 (Tree Star, Inc.). To prevent thecontamination of nonmacrophage cells, CD14 and CD206 wereused to purify the mature macrophages by using a BeckmanCoulterMoFlow XDP cell sorting system (Beckman Coulter).Macrophages with or without purification were used for analysisof CSF-1R phosphorylation and endocytosis.

CSF-1R phosphorylationAfter the human monocytes were differentiated into macro-

phages by 6-day incubation with CSF-1, the CSF-1-containingmedium was replaced by medium without serum and CSF-1 toserum starve the macrophages for 18 hours. After serum starva-tion, the cells were pretreated with CSF-1R inhibitors for 2 hoursand then stimulated with CSF-1 (100 ng/mL) for 5 minutes. Celllysates were generated with RIPA buffer containing protease andphosphatase inhibitor cocktails. DC protein assay (Bio-Rad) wasused to determine the protein concentrations and lysates at 0.2mg/mL were used to quantify the phosphorylated CSF-1R by thePathScan Phospho-M-CSF-Receptor sandwich ELISA Kit, accord-ing to manufacturer's instruction.

CSF-1R endocytosisMonocytes, which were enriched from whole blood by using

the Lymphoprep (STEMCELL), were seeded on 20mmcover glassand induced differentiation intomacrophages by the stimulationof CSF-1 (10 ng/mL) for 6 days. Cells were blocked with CAS-

Block Histochemical reagent (Thermo Fisher Scientific Inc.) at37�C for 1 hour. After that, cells were stained with CSF-1R-GFPantibody for 1 hour andwashedbyfiltered PBS.Dynamics of CSF-1R-GFP was acquired with an APO 100�/1.65 NA oil immersionobjective at 37�C using a total internal reflection fluorescence(TIRF) microscopy (Olympus). After the region of interest wasselected, the 488-nm excitation laser was angled until reflectionwasobservedon the cover glass. The evanescent penetration depthof 80 nm was used and images were continuously acquired onceper 30 seconds for 45 minutes to monitor the endocytosis ofCSF-1R.

Cell viabilityMonocytes were enriched from whole blood by using the

Lymphoprep (STEMCELL) and 7 � 105 cells were seeded intothewells of 96-well cell culture plate. CSF-1 10 ng/mLwas used toinduce the differentiation of macrophages after seeding anddifferent concentrations of CSF-1R inhibitor were added at thesame day of CSF-1 stimulation. After incubation of CSF-1 andCSF-1R inhibitor for 8 days, CellTiter-Glo Luminescent CellViability Assay was used to determine the cell viability accordingto manufacturer's instruction.

Statistical analysisAll valueswere reported asmean� SEM. Fisher's exact test, Chi-

square test, and unpaired t test were used to compare the differ-ence between groups. Pearson correlation was used to determinethe correlation between the M1/M2 ratio and VEGF expression.Kaplan–Meier survival analysis and log-rank test were used toestimate the survival and compare the difference between groups.A P value <0.05 was considered statistically significant.

ResultsGlobal distribution of CSF1R c.1085A>G genetic variant

The germline genetic variants of CSF1R from 140 cancerpatients were analyzed, including 90 cases of colorectal cancer,28 cases of ovarian cancer, and 22 cases of endometrial cancer. Inthese 140 cancer patients, 496 genetic variants of CSF1R, includ-ing the single nucleotide variants (SNV) and small insertion/deletions (Indel), were identified in 3' untranslated region (UTR),5'UTR, introns, and exons. Among the 496 genetic variants, 14variants were located in exons and only 7 of 14 exonic variantswere nonsynonymous substitution leading to a change in aminoacid (Fig. 1A and B). The frequency of these nonsynonymousvariants was low (<2%), except for the genetic variant c.1085A>Gwith a high allelic frequency of 42.86%. However, there was nosignificant difference in this allele frequency between cancer types(Supplementary Fig. S1A). We also checked the frequency anddistribution of CSF1R genetic variants in Taiwan Biobank (12)that enrolled healthy Taiwanese, the same ethnic groupwith those140 cancer patients. The c.1085A>G variant was the only non-synonymous genetic variant of CSF1R that could be identified inthe exonic regionswith the allele frequencymore than 2%(Fig. 1Band C). There was no significant difference in the frequency ofCSF1R genetic variants between healthy Taiwan population andcancer patients. We further studied the worldwide distribution ofc.1085A>G genetic variant of CSF1R based on 1,000 GenomesProject. The high frequency of c.1085A>G variant (about 38%)was also observed in the population of East Asia (Fig. 1D andSupplementary Fig. S1B). In contrast, only 7% to 11% alternative

The Impact of CSF1R Genetic Variant on CSF-1R Signaling

www.aacrjournals.org Clin Cancer Res; 23(20) October 15, 2017 6023




allele frequency was observed in Africa, America, South Asia, andEuropean population (Fig. 1D).

CSF1R variant is associated with clinical outcomeTo study the clinical relevance, the total macrophage content,

expression patterns of M1- and M2-like tumor-associated macro-phages, and associated cytokines were examined in the surgicalspecimens of colorectal, ovarian, and endometrial cancer (Fig.2A–E and Supplementary Fig. S2). CD68 was used to stain thetotalmacrophages, andM1andM2macrophagesweremonitoredusing cluster of inducible nitric oxide synthase (iNOS) andCD163 staining, respectively. As shown in Fig. 2A and B, thenumber of tumor-associatedmacrophage is significantly higher intumor tissues derived from the genetic background of CSF1Rc.1085 genotype A_A. Compared with that of reference group, thetumor M1/M2 ratio was significantly higher in the surgical speci-

mens of c.1085A>G genetic variant of CSF1R (Fig. 2C and D andSupplementary Fig. S2). In addition, the M1/M2 ratio of tumor-associated macrophages inversely correlated with VEGF expres-sion in tumor tissues (Fig. 2F, r2 ¼ 0.37, P < 0.0001). A higherM1/M2 ratio of tumor-associated macrophages accompanied bylower VEGF expression in tumor tissues was noted in the surgicalspecimens from cancer patients with CSF1R c.1085 A>G geneticvariant (Fig. 2C–E). More importantly, CSF1R genetic variant issignificantly associated with disease-free survival in stage IIIcolorectal cancer (Fig. 2G and Supplementary Table S1) andoverall survival in endometrial cancer (Supplementary Fig. S3Aand S3B). For patients with ovarian cancer, there was also a trendshowing poor outcome in the group of CSF1R c.1085 genotypeA_A than genotype A_G (Supplementary Fig. S3C and S3D, P ¼0.059). These results imply the impact ofCSF1R genetic variant onthe clinical outcome of cancer patients.

Figure 1.

The frequency and distribution of germline CSF1R genetic variants. A, Germline CSF1R genetic variants in 140 cancer patients. The whole genome data of 140 cancerpatients from a clinical study evaluating the association between genetic variants and chemotherapy-induced neurotoxicity were used to analyze thedistribution and allele frequency of germline CSF1R genetic variants in this study. The variants located at exons were shown here. Solid and open symbols indicatenonsynonymous and synonymous variants, respectively. Ig, immunoglobulin domain; Tyr kinase, tyrosine kinase domain. B, Comparison of the distributionof CSF1R genetic variants between cancer patients and normal population. The allele frequency of germline CSF1R variants was compared between 140 cancerpatients and 499 normal Taiwanese by Fisher's exact test. Bold indicates the nonsynonymous variants. NS, no significance. C, Germline CSF1R geneticvariants in Taiwan general populations. The database of Taiwan Biobank, which included 499 normal Taiwanese, was used to analyze the distribution of germlineCSF1R genetic variants in Taiwan general population. D, The alternative allele frequency in CSF1R c.1085 in different ethnic groups. The allele frequency of variantallele in CSF1R c.1085 was compared between this study, normal Taiwanese, and different ethnic groups enrolled in the 1,000 Genomes Project.

Yeh et al.





CSF1R c.1085A>G genetic variant regulates the dynamics ofreceptor endocytosis

We further explored the mechanism underlying the impact ofCSF1R c.1085 genetic variant on macrophage function. CSF-1stimulates CSF-1R dimerization which results in CSF-1R activa-tion, followed by kinase inactivation, dephosphorylation, andinternalization of the receptor–ligand complex (14). Because theCSF1R c.1085 genetic variant is located in the domain of dimer-ization, we studied whether CSF1R c.1085 genetic variant isinvolved in the regulation of CSF-1-induced internalization andendocytosis of CSF-1R. Themononuclear cells were isolated fromthe peripheral blood, seeded on 20 mm cover glass, and differ-

entiated into macrophages by incubation with 10 ng/mL CSF-1for 6 days. To study the endocytosis of CSF-1R, anti-CSF-1R-GFPantibody was used to stain CSF-1R and the dynamics of CSF-1Rendocytosis inmacrophage were real-timemonitored by the totalinternal reflection fluorescence microscopy (TIRFM) with anevanescent penetration depth of 80 nm (Fig. 3A). It took 35–45 minutes from the exposure to CSF-1, activation of CSF-1Rsignaling, clustering, internalization, endocytosis, to final degra-dation of CSF-1R (15). Accordingly, we monitored the CSF-1-induced changes of fluorescent signals for 40minutes to visualizethe dynamics of CSF-1R endocytosis. When macrophages withCSF1R genotype A_A were stimulated by CSF-1, the fluorescent

Figure 2.

CSF1R c.1085A>G genetic variant wasassociated with the polarized tumor-associated macrophages, relatedcytokine, and clinical outcome. A andB, Representative images andquantitative analysis of tumor-associated macrophages in 52 tumortissues from colorectal cancer. Scalebar, 100 mm. C, Representativeconfocal images co-stained with iNOS(for M1 macrophages), CD163 (for M2macrophages) andHoechst 33258 (fornucleus) in tumor specimens ofcolorectal cancer were shown in topand those co-stained with TNFa (M1-related cytokine), VEGF (M2-relatedcytokine), and Hoechst 33258 wereshown in bottom (n ¼ 52). Scale bar,40 mm. D, The number of M1- and M2-like macrophages quantified in fivedifferent fieldswith 40� oil immersionlens was used to calculate the M1/M2ratio. TheM1 toM2 ratiowas comparedbetween the group of CSF1R c.1085A_A and CSF1R c.1085 A_G. E, Thepercentage of positive staining areafor VEGF relative to total tumor areawas measured and analyzed byTissueGnostics GmbH FACS-likeTissue Cytometry (TissueFAXS Plus).The percentage of positive stainingarea to total tumor areawas comparedbetween the group of CSF1R c.1085A_AandCSF1R c.1085A_G. Solid lines,mean � SEM. Parentheses indicatecase number in each group.�� , P < 0.001 by unpaired t test.F, Correlation between the positivestaining area of VEGF and M1/M2 ratioin tumor specimens was determinedby Pearson correlation coefficient.G, The Kaplan–Meier curves ofdisease-free survival in colorectalcancer patients with CSF1R c.1085genotype A_A (n ¼ 21) and A_G(n ¼ 31) were shown and comparedby log-rank test.






signals first became brighter in the first 5 minutes, suggesting theclustering of CSF-1R. After that, the intensity of fluorescent signalsremarkably decreasedby70% in the following 5minutes and thenrecovered gradually in the next 5 minutes which indicated therapid endocytosis of CSF-1R induced by CSF-1 (Fig. 3B and C;Movie 1). By striking contrast, whenmacrophages with theCSF1Rgenotype A_Gwere stimulated by CSF-1, the fluorescent intensityinitially surged and gradually decreased by 10% to 20% in thewhole period (Fig. 3B and D; Movie 2). It seemed that CSF-1induced CSF-1R clustering which was followed by slower inter-nalized process in the genotype of CSF1R c.1085 A_G. Theseresults indicate that CSF1R c.1085 genetic variant, located in thedomain of dimerization, is involved in the regulation of CSF-1-induced endocytosis of CSF-1R.

CSF1R genetic variant determines macrophage polarizationThe biological effects of CSF-1 are mediated by the auto-

phosphorylation of CSF-1R, which triggers the downstream sig-naling and determines macrophage differentiation. BecauseCSF1R c.1085A>G genetic variant regulates the dynamics of

CSF-1R endocytosis, it is likely that CSF-1 differentially activatesCSF-1R phosphorylation and affects the subsequent macrophagefunction. Accordingly, we studied the effects of the CSF1R geneticvariant on the CSF-1-induced phosphorylation of CSF-1R and thedifferentiation of M2-like macrophages in vitro. After peripheralbloodmononuclear cellswere differentiated intomacrophages byincubated with CSF-1 (10 ng/mL) for 6 days, macrophages wereserum starved for 18 hours, stimulated by CSF-1 for 5 minutes,and then harvested for collection of cell lysates. Phospho-MCSF-Receptor sandwich ELISA kit was used to determine the phos-phorylation of CSF-1R. As shown in Figure 4A, there was asignificant difference in CSF-1-induced CSF-1R phosphorylationbetween the reference (CSF1R c.1085A_A) and variant groups(CSF1R c.1085A_G). In macrophages from reference group(CSF1R c.1085 A_A), CSF-1 induced 2.2-fold increase in CSF-1R phosphorylation, in which 10 nmol/L PLX3397, a CSF-1Rinhibitor, blocked 70% of CSF-1-induced phosphorylation. Incontrast, in the variant groups, CSF-1 induced 1.4-fold increase inCSF-1R phosphorylation that was abolished by 10 nmol/LPLX3397. We further studied whether CSF1R genetic variant

Figure 3.

CSF1R c.1085A>G genetic variant regulates the endocytosis of CSF-1R. A, Schematic showing TIRFM monitored the endocytosis of CSF-1R by selective excitationof fluorophores at plasma membrane and submembrane region near the interface of the cell and cover glass. Cells were stained with anti-CSF-1R-GFPantibody and endocytosis of CSF-1R was monitored by imaging the changes of fluorescent signals with the evanescent penetration depth of 80 nm continuously.B, Representative TIRFM images showing the changes of fluorescent signals after stimulation by CSF-1. Macrophages differentiated from peripheral bloodmononuclear cells were seeded on 20 mm cover glass and stained by anti-CSF-1R-GFP antibody. CSF-1 20 ng/mL was used to stimulate macrophages and thechanges of fluorescent signals were real-time monitored by TIRFM once per 30 second for 40 minutes. C and D, Changes of CSF-1R-GFP fluorescent intensity inmacrophages with CSF1R c.1085 A_A and CSF1R c.1085 A_G, respectively. Y axis, normalized fluorescent intensity in TIRFM. ROI, region of interest.

Yeh et al.





determined macrophage polarization (Fig. 4B–E). Cells differen-tiated from peripheral blood mononuclear cells were harvestedafter incubation with CSF-1 for 6 days. Forward scatter and sidescatter were used to gate the population of macrophages (Fig. 4B)and polarized M2-like macrophages were characterized by theexpression of CD206 (Fig. 4D). Compared to the variant group,the percentage of CSF-1-induced macrophage differentiation wassignificantly higher in the reference group (Fig. 4C, P < 0.05).When CD206 was used to characterize the M2-like macrophages,the percentage ofM2-likemacrophages after incubationwithCSF-1 was also significantly higher in the reference group than in thevariant group (Fig. 4E, P < 0.05). Accompanied with the higherpercentage of M2-like macrophages, macrophages with CSF1Rc.1085 genotype A_A had lower iNOS expression than those withCSF1R c1.085 genotype A_G (Supplementary Fig. S4, P < 0.05).

We also purified mature macrophages by sorting to repeat theexperiments of CSF-1R phosphorylation and endocytosis. Theprotocol of sorting mature macrophages was shown in Supple-mentary Fig. S5A. In brief, after macrophages were harvested,CD14 and CD206 were used to sort mature macrophages for

functional study (Supplementary Fig. S5B and S5C). Consistentwith the results shown in Fig. 3 and 4A, CSF-1 induced a rapidreceptor endocytosis and remarkable phosphorylation in maturemacrophages with the genotype of CSF1R c.1085 A_A.

CSF1R c.1085A>G genetic variant confers the sensitivity ofmacrophage survival to CSF-1R inhibitors

Small molecule inhibitors of CSF-1R offer an attractive strategyfor reducing macrophage numbers associated with cancer as wellas autoimmune and inflammatory disease. However, in the earlyphase clinical trials, the results of CSF-1R inhibitors were contro-versial, either positive or detrimental for cancer treatment. Wehere tested the hypothesis that the genetic alteration within theCSF1R locus could determine macrophage susceptibility to CSF-1R inhibitors. Three CSF-1R kinase inhibitors, includingPLX3397, BLZ945, and GW2580, were used to study their effectson macrophage proliferation. In cellular assays of macrophageswith CSF1R c.1085A>G genetic variant, these three CSF-1R inhi-bitors dose-dependently inhibitedmacrophage survival with IC50

in the 0.1 to 1 nmol/L range (Fig. 5A and B). By contrast, for

Figure 4.

CSF1R c.1085A>G genetic variant determines the phosphorylation of CSF-1R and the differentiation and polarization of macrophages. A, Quantitative analyses ofCSF-1R phosphorylation induced by CSF-1 in macrophages with (open column) and without (solid column) CSF1R c.1085A>G genetic variant. Macrophagesdifferentiated from peripheral blood mononuclear cells were serum starved for 18 hours followed by stimulation with CSF-1 100 ng/mL for 5 minutes with orwithout pretreatment with CSF-1R inhibitor, PLX3397. The phosphorylation of CSF-1R was measured by phospho-MCSF-receptor sandwich ELISA Kit. Y axis,normalized CSF-1R phosphorylation. Each value represents mean � SEM from at least seven different samples in each group. B, Macrophages were differentiatedfrom peripheral blood mononuclear cells by incubation with CSF-1 10 ng/mL for 6 days. Cells were harvested for flow cytometry analysis. Forward scatter(FSC) and side scatter (SSC) were used to gate the population of macrophages. C, The percentage of macrophage differentiation induced by CSF-1 was comparedbetween reference (CSF1R c.1085 A_A) and variant group (CSF1R c.1085 A_G). D, The expression of CD206 was used to characterize the M2-like macrophages.E, The percentage of CD206 (þ) cells was compared between the reference and variant group. Data were presented as means � SEM. �� , P < 0.01 byunpaired t test.






macrophages with CSF1R genotype A_A, PLX3397, BLZ945, andGW2580 inhibited themacrophage survival with IC50 in the 10 to100 nmol/L range. These results suggest that CSF1R geneticvariant confers sensitivity to the treatment of CSF1R inhibitors.

DiscussionHere we identified the genetic alteration within CSF1R locus

mediates macrophage function. This study highlights an impor-tant finding that macrophages with CSF1R c.1085A>G geneticvariant display a refractory response to CSF-1 stimulation andsusceptibility to CSF-1R inhibitors. This conclusion is supportedby the following evidences. (i) In the reference group (CSF1Rgenotype A_A), CSF-1 induced a remarkable CSF-1R phosphor-ylation and a fast process of receptor endocytosis. However,macrophages with CSF1R c.1085A>G genetic variant showedrefractory to CSF-1 stimulation, in terms of receptor phosphor-ylation and endocytosis. (ii) Significant CSF-1-mediated macro-phage differentiation and M2 polarization were observed in thereference group. In contrast,macrophageswithCSF1R c.1085A>Ggenetic variant demonstrated a poor response to CSF-1 stimula-tion while determined by the percentage of macrophage differ-entiation and expression ofM2marker. (iii) The genetic alterationwithin CSF1R locus confers the sensitivity to CSF-1R inhibitors.Compared to macrophages with CSF1R c.1085A>G genetic var-iant, the IC50 of CSF-1R inhibitors to inhibit macrophage survivalwas 10- to 100-fold higher in the reference group.

In terms of molecular and structural aspects, we propose twopossible mechanisms to explain the great impact of c.1085A_Gallele on macrophage function. First, the CSF1R c.1085 A>Ggenetic variant results in a change of amino acid from histidineto arginine within the immunoglobulin-like domain 4 which isessential for the formation of receptor dimerization. Althougharginine andhistidine are both basic amino acids, histidine has animidazole side chain that would lose one proton at a pH above itspK of 6.0 leading to the change from positive charge to a neutralstatus (16). The different side chain of amino acid and the changeof positive charge in dimerization domain of CSF-1R caused bythe substitutionof histidine by argininemight affect the process ofCSF-1-induced dimerization, subsequent phosphorylation andendocytosis of CSF-1R. Second, the substitution of amino acidcould change the protein stability and alter the level of proteinexpression. Zhao and colleagues reported arginine-to-histidine

mutations of voltage-gated potassium channel also had an impacton the stability of the protein and reduced the protein expressionon the cell surface (17). The substitution of histidine by argininecaused CSF1R c.1085 A>G genetic variant might also have aneffect on CSF-1R expression. Different levels of CSF-1R expressionmay explain the different responses to CSF-1 stimulationobserved inmacrophageswithdifferentCSF1R genotypes. Furtherstudy is needed to confirm these hypotheses.

Our clinical study showed that CSF1R c.1085 genetic variantwas associated with clinical outcomes. M2-like macrophage hasbeen reported to be associated with a poor outcome in differenttypes of cancers (3). Here we found that tumor M1/M2 ratio wassignificantly higher in the surgical specimens of CSF1Rc.1085A>G, compared with that of reference group. In addition,M1/M2 ratio of tumor-associated macrophages inversely corre-lated with VEGF expression in tumor tissues. It seems that CSF1Rc.1085A>G genetic variant regulated tumor immunity throughdifferential recruitment of tumor-associated macrophages andrelated cytokines. Importantly, when analyzing the effect ofCSF1R genetic variant on clinical outcome in patients with stageIII colorectal cancer, the reference group demonstrated a poorerdisease-free survival than patients with CSF1R c.1085 geneticvariant. Endometrial cancer patients wereCSF1R c.1085 genotypeA_Aalso had apoor overall survival than thosewithCSF1R c.1085A_G. For patients with ovarian cancer, there was also a trendshowing a shorter survival in the group ofCSF1R c.1085 genotypeA_A than genotype A_G. These results imply that CSF1R geneticvariant has an impact on tumor immunity associatedwith clinicaloutcomes.

CSF-1 induced a strong CSF-1R signaling in macrophages withCSF1R c.1085 genotype A_A, leading to more macrophage dif-ferentiation and M2 polarization. Importantly, we observedhigher concentration of CSF-1R inhibitors was required to inhibitCSF-1-induced receptor phosphorylation in macrophages withCSF1R c.1085 genotype A_A. Because CSF-1R singling is impor-tant for survival, proliferation, and differentiation of macro-phages (18), it is not surprised that higher concentration ofCSF-1R inhibitors was required to inhibit the survival of macro-phages with CSF1R genotype A_A. We observed the IC50 of CSF-1R inhibitors was 10- to 100-fold lower in macrophages withCSF1R c.1085 A_G than A_A. These results imply that CSF-1Rinhibitors should be used with caution and highlight the impor-tance of testing CSF1R genotypes before prescribing CSF-1R

Figure 5.

CSF1R c.1085A>G genetic variantconfers the sensitivity of macrophagesurvival to CSF-1R inhibitors.Macrophages differentiated fromperipheral blood mononuclear cellswere incubated with variousconcentrations of three CSF-1Rinhibitors, including PLX3397 (A),BLZ945 (B), and GW2580 (C) for 8days. Percentage of cell survival wasdetermined by CellTiter-GloLuminescent Cell Viability Assay.

Yeh et al.





inhibitors clinically. In patients with CSF1R genotype A_G, tar-geting CSF-1R signaling by CSF-1R inhibitors may cause excessdeaths of macrophages and the beneficial M1 macrophages mayalso be depleted. The excess death of macrophages would jeop-ardize innate immunity which would put the host in the risk ofinfections. And, the depletion of beneficial M1macrophages mayreduce the efficacy of CSF-1R inhibitors in patients with CSF1Rc.1085 genotype A_G.

The allele frequency ofCSF1R c.1085A>G varies in population.This genetic variant has a high incidence in Eastern Asian pop-ulationwith the allele frequency of 42.86%.However, only 7% to11%of this genetic variant was observed in Africa, America, southAsia, and European. In the era of precisionmedicine, using geneticmarkers to predict the benefit of treatment or avoid potentialtoxicities is getting more andmore important. A good example toshow the importance of varied frequency of genetic variantsamong different ethnic groups is the efficacy of EGFR inhibitionin phase III trial of lung cancer, the INTEREST study (19). TheAsian lung adenocarcinomahad the high epidermal growth factorreceptor (EGFR) mutation frequency (�50%) and the low EGFRmutation frequency occurred in west population, around 10%.The 20% of Asian population in INTEREST study led to failure ofdemonstrating better efficacy of EGFR inhibitor than traditionalchemotherapy and withdrawal of gefitinib from the USmarket. Arecent study tried to show the different effects of CSF-1R mono-clonal antibody on macrophage survival (10). Pradel and collea-gues (10) investigated the impact of SNP rs10079250 on thedepletion of macrophages by RG7155, a humanized anti-CSF-1Rmonoclonal antibody blocking the receptor dimerization. Theresults showed a trend toward a reduced response to RG7155 indonors carrying the variant allele. The SNP rs10079250 is thegenetic variant, CSF1R c.1085A>G investigated in our study. Thedifferent frequency ofCSF1R c.1085 genetic variant among ethnic

groups associated with the different response to CSF-1R inhibi-tion raises an important issue thatCSF1R genetic variantmight bea potential predictive biomarker in targeting CSF-1R signalingclinically.

Taken together, this is the first study to demonstrate CSF1Rgenetic variant regulates the CSF-1R signaling and sensitivity toCSF-1R inhibitors (summarized in Table 1). Our work showedCSF-1 induced remarkable phosphorylation and endocytosis ofCSF-1R in macrophages with CSF1R c.1085 genotype A_A. Theprominent activation and dynamics of CSF-1Rwere accompaniedby the significant CSF-1-mediated macrophage differentiationand M2 polarization. M2 predominant distribution of tumor-associated macrophages was also observed in clinical specimenswith CSF1R genotype A_A and colorectal cancer patients withCSF1R genotype A_A had a poorer disease-free survival. Differentsusceptibility to CSF-1R inhibitors in macrophages with differentCSF1R genotypes suggests this genetic variant is worth beingtested as a predictive marker when targeting CSF-1R signaling bysmall molecular inhibitors, especially in the population in whichthe prevalence of CSF1R c.1085A>G genetic variant is high.

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

Authors' ContributionsConception and design: Y.-M. Yeh, S.-J. Hsu, P.-C. Lin, M.-R. ShenDevelopment of methodology: Y.-M. Yeh, S.-J. HsuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): Y.-M. Yeh, S.-J. Hsu, P.-C. Lin, K.-F. Hsu, P.-Y. Wu,W.-C. SuAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): Y.-M. Yeh, S.-J. Hsu, P.-C. Lin, K.-F. Hsu, P.-Y. Wu,J.-Y. ChangWriting, review, and/or revision of the manuscript: Y.-M. Yeh, M.-R. ShenAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): Y.-M. Yeh, S.-J. Hsu, M.-R. ShenStudy supervision: M.-R. Shen

AcknowledgmentsThis work was supplied by a grant from the Ministry of Science and

Technology, Ministry of Health and Welfare (MOHW106-TDU-B-211-113003), and National Cheng Kung University Hospital.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received April 6, 2017; revised June 27, 2017; accepted July 14, 2017;published OnlineFirst July 19, 2017.

References1. Wynn TA, Chawla A, Pollard JW. Macrophage biology in development,

homeostasis and disease. Nature 2013;496:445–55.2. Van Overmeire E, Stijlemans B, Heymann F, Keirsse J, Morias Y, Elkrim Y,

et al. M-CSF and GM-CSF receptor signaling differentially regulate mono-cyte maturation and macrophage polarization in the tumor microenvi-ronment. Cancer Res 2016;76:35–42.

3. Mantovani A, Sozzani S, Locati M, Allavena P, Sica A. Macrophage polar-ization: tumor-associated macrophages as a paradigm for polarized M2mononuclear phagocytes. Trends Immunol 2002;23:549–55.

4. Pyonteck SM, Akkari L, Schuhmacher AJ, Bowman RL, Sevenich L, QuailDF, et al. CSF-1R inhibition alters macrophage polarization and blocksglioma progression. Nat Med 2013;19:1264–72.

5. DeNardoDG, BrennanDJ, Rexhepaj E, Ruffell B, Shiao SL,Madden SF, et al.Leukocyte complexity predicts breast cancer survival and functionallyregulates response to chemotherapy. Cancer Discov 2011;1:54–67.

6. Zhu Y, Knolhoff BL, Meyer MA, Nywening TM, West BL, Luo J, et al. CSF1/CSF1Rblockade reprograms tumor-infiltratingmacrophages and improvesresponse to T-cell checkpoint immunotherapy in pancreatic cancermodels.Cancer Res 2014;74:5057–69.

7. https://clinicaltrials.gov/ct2/results?term¼csf1r&Search¼Search.8. Rademakers R, Baker M, Nicholson AM, Rutherford NJ, Finch N, Soto-

Ortolaza A, et al. Mutations in the colony stimulating factor 1 receptor(CSF1R) gene cause hereditary diffuse leukoencephalopathy with spher-oids. Nat Genet 2011;44:200–5.

Table 1. The genetic alteration within CSF1R locus mediates macrophagefunction

Genetic variant CSF1R c.1085 A_A CSF1R c.1085 A_G

Global distribution �90% in African,American, European,and South Asian

�40% in East Asian

Response to CSF-1CSF-1R phosphorylation Strong WeakCSF-1R endocytosis Rapid SlowM2 polarization Prominent Less prominent

Sensitivity to CSF-1R inhibitorsIC50 10–100 nmol/L 0.1–1 nmol/L

Potential clinical impact Poor outcome Favorable outcome





https://clinicaltrials.gov/ct2/results?term=csf1r&Search=Search





9. Kang HG, Lee SY, Jeon HS, Choi YY, Kim S, Lee WK, et al. A functionalpolymorphism in CSF1R gene is a novel susceptibility marker for lungcancer among never-smoking females. J Thorac Oncol 2014;9:1647–55.

10. Pradel LP, Ooi CH, Romagnoli S, Cannarile MA, Sade H, Ruttinger D, et al.Macrophage susceptibility to emactuzumab (RG7155) treatment. MolCancer Ther 2016;15:3077–86.

11. https://clinicaltrials.gov/ct2/show/NCT02481336?term¼meng-ruþshen&rank¼1.

12. Fan CT, Lin JC, Lee CH. Taiwan Biobank: a project aiming to aid Taiwan'stransition into a biomedical island. Pharmacogenomics 2008;9:235–46.

13. Genomes Project C, Auton A, Brooks LD, Durbin RM, Garrison EP, KangHM,et al. A global reference for human genetic variation. Nature 2015;526:68–74.

14. Li W, Stanley ER. Role of dimerization and modification of the CSF-1receptor in its activation and internalization during the CSF-1 response.EMBO J 1991;10:277–88.

15. Lou J, Low-Nam ST, Kerkvliet JG, Hoppe AD. Delivery of CSF-1R to thelumen ofmacropinosomes promotes its destruction inmacrophages. J CellSci 2014;127:5228–39.

16. Tu Z, Young A, Murphy C, Liang JF. The pH sensitivity of histidine-containing lytic peptides. J Pept Sci 2009;15:790–5.

17. Zhao J, Zhu J, Thornhill WB. Spinocerebellar ataxia-13 Kv3.3 potas-sium channels: arginine-to-histidine mutations affect both function-al and protein expression on the cell surface. Biochem J 2013;454:259–65.

18. Pixley FJ, Stanley ER. CSF-1 regulation of the wandering macrophage:complexity in action. Trends Cell Biol 2004;14:628–38.

19. Kim ES, Hirsh V, Mok T, Socinski MA, Gervais R, Wu YL, et al.Gefitinib versus docetaxel in previously treated non-small-cell lungcancer (INTEREST): a randomised phase III trial. Lancet 2008;372:1809–18.


Yeh et al.



https://clinicaltrials.gov/ct2/show/NCT02481336?term=meng-ru+shen&rank=1






2017;23:6021-6030. Published OnlineFirst July 19, 2017.Clin Cancer Res Yu-Min Yeh, Shan-Ju Hsu, Peng-Chan Lin, et al. Macrophages

ofImmunity by Altering the Proliferation, Polarization, and Function Gene Regulates TumorCSF1RThe c.1085A>G Genetic Variant of

Updated version

10.1158/1078-0432.CCR-17-1007doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2017/07/19/1078-0432.CCR-17-1007.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/23/20/6021.full#ref-list-1

This article cites 17 articles, 6 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/23/20/6021.full#related-urls

This article has been cited by 1 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 at

Permissions

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

.http://clincancerres.aacrjournals.org/content/23/20/6021To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-17-1007

http://clincancerres.aacrjournals.org/content/suppl/2017/07/19/1078-0432.CCR-17-1007.DC1

http://clincancerres.aacrjournals.org/content/23/20/6021.full#ref-list-1

http://clincancerres.aacrjournals.org/content/23/20/6021.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/23/20/6021


the c.1085a g genetic variant of csf1r gene regulates tumor immunity … · cancer therapy:...

Documents