Interleukin-6StimulatesDefectiveAngiogenesiscancerres.aacrjournals.org/content/canres/75/15/3098.full.pdf ·...

Molecular and Cellular Pathobiology

Interleukin-6StimulatesDefectiveAngiogenesisGanga Gopinathan1, Carla Milagre1, Oliver M.T. Pearce1, Louise E. Reynolds1,Kairbaan Hodivala-Dilke1, David A. Leinster1, Haihong Zhong2, Robert E. Hollingsworth2,Richard Thompson1, James R.Whiteford3, and Frances Balkwill1

Abstract

The cytokine IL6 has a number of tumor-promoting activ-ities in human and experimental cancers, but its potential asan angiogenic agent has not been fully investigated. Here, weshow that IL6 can directly induce vessel sprouting in the ex vivoaortic ring model, as well as endothelial cell proliferation andmigration, with similar potency to VEGF. However, IL6-stim-ulated aortic ring vessel sprouts had defective pericyte coveragecompared with VEGF-stimulated vessels. The mechanism ofIL6 action on pericytes involved stimulation of the Notch

ligand Jagged1 as well as angiopoietin2 (Ang2). When peri-toneal xenografts of ovarian cancer were treated with an anti-IL6 antibody, pericyte coverage of vessels was restored. Inaddition, in human ovarian cancer biopsies, there was anassociation between levels of IL6 mRNA, Jagged1, and Ang2.Our findings have implications for the use of cancer therapiesthat target VEGF or IL6 and for understanding abnormalangiogenesis in cancers, chronic inflammatory disease, andstroke. Cancer Res; 75(15); 3098–107. �2015 AACR.

IntroductionIL6 is a major tumor-promoting cytokine produced by both

malignant and host cells in the tumormicroenvironment (1). It isalso a downstream product of oncogenic mutations, for example,ras and TP53 (2, 3). Typically via its major downstream signaltransducer STAT3, IL6 has both local and systemic protumoractions in experimental and human cancers. In the tumor micro-environment, these include stimulation of malignant cell growthand survival (4), promotion of invasion and metastasis (5),modulation of tumor-promoting T-cell subtypes, involvementin autocrine tumor cell cytokine networks (6), and regulation ofthe myeloid cell infiltrate (7). Systemic effects of excess IL6production include induction of acute phase reactants andinvolvement in the elevated platelet count (paraneoplastic throm-bocytosis; ref. 8) that is a complication of several commonhumancancers.

To add to this catalogue of tumor-promoting actions, there arereports that IL6 stimulates angiogenesis in the tumor microenvi-ronment (9) with evidence that STAT3 signaling induces hypoxia-inducible factor-mediated VEGF-A transcription (10). IL6 is alsoreported tohavedirect effects on endothelial cell proliferation andmigration (9, 11, 12) and has been implicated in resistance toanti-VEGF antibody treatment in patients (13, 14). In preclinical

and clinical studies, we found that a therapeutic neutralizing anti-IL6 antibody reduced systemic VEGF levels in ovarian cancerpatients, and that in peritoneal ovarian cancer xenografts, bloodvessels were reduced, with a concomitant inhibition of the Notchligand Jagged 1 (7).

This led us to study further the actions of IL6 in normal andcancer angiogenesis. In this article, we present novel evidencethat IL6 directly stimulates angiogenesis, but in contrast withVEGF, IL6-stimulated vessels have defective pericyte coverage.We show that this may be due to differential regulation ofNotch ligands and Ang2 by these two mediators. Our findingshave implications for the use of cancer therapies that targetVEGF or IL6.

Materials and MethodsEthics statement

All animal experiments were approved by the local ethicsreview process of the Biological Services Unit, Queen Mary Uni-versity of London (London, United Kingdom) and conducted inaccordancewith theUKCCCRguidelines for thewelfare anduseofanimals in cancer research.

Aortic ring assayAngiogenic sprouts were induced from mouse or rat thoracic

aortas according to the method of Nicosia and Ottinetti (15).Aortas were dissected from cervically dislocated 8- to 12-week-oldmale C57BL/6mice (Charles River) or 180–200 gmaleWistar rats(Harlan Laboratories) and sliced into 0.5 mm sections andincubated overnight in serum-free OptiMEM (Invitrogen) at37�C. Aortic rings were embedded in type I collagen (1 mg/mL)in E4 media (Invitrogen). For mouse aortic rings, the wells weresupplemented with OptiMEM with 1% FBS and 30 ng/mL ofVEGF (R&D Systems), 50 ng/mL of human IL6 (R&D Systems), or30 ng/mL of mouse IL6 (R&D Systems) and incubated at 37�C,10% CO2. Rat aortic ring wells were treated with OptiMEM with1% FBS and 10 ng/mL VEGF, 10 ng/mL rat IL6, or 10 nmol/LVEGFRi (Cediranib, VEGFR2 inhibitor) and incubated at 37�C,

1Barts Cancer Institute, Queen Mary University of London, Charter-house Square, London, United Kingdom. 2MedImmune, One MedIm-mune Way, Gaithersburg, Maryland. 3William Harvey Institute, QueenMary University of London, Charterhouse Square, London, UnitedKingdom.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Frances Balkwill, Barts Cancer Institute, CharterhouseSquare, London EC1M 6BQ, United Kingdom. Phone: 44-207-882-3851; Fax: 44-207-882-3885; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-15-1227

�2015 American Association for Cancer Research.

CancerResearch

Cancer Res; 75(15) August 1, 20153098

on February 15, 2019. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 16, 2015; DOI: 10.1158/0008-5472.CAN-15-1227

http://cancerres.aacrjournals.org/

10%CO2. Angiogenic sprouts were counted after 7 days of cultureformouse aortic ring and after 4 days of culture for rat aortic rings.The length of sprouts was quantified using ImageJ software bydrawing radial lines from thebase of the aortic ring to the tip of thesprouting new vessel. Pericytes were quantified 250microns fromthe tip of the aortic ring vessel to avoid false-positive quantifica-tion of activated fibroblast, which are normally found at the stalkof the vessel. Animals were housed and treated in accordancewithUK Home Office Regulations.

Staining of aortic ringsThe rat andmouse aortic rings were, respectively, cultured for 1

and 2 weeks before the staining. The rings were washed with PBS,fixed in 4% formaldehyde for 20 minutes. The wells were thenwashed once in PBS and the rings were permeabilized with 0.5%Triton X-100 in PBS for 30minutes, before being washed twice inPBS.Of note, 100mLof BS-1 Lectin FITC (1mg/mL; Sigma, cat. no.L9381/L5264; 1:200), anti-actin, a-SMA Cy3 (Sigma, cat. no.C6198; 1:500), or anti-NG2 (Millipore, ab5320; 1:200) wasadded and incubated overnight at 4�C. For IL6Ra staining onaortic rings, 100 mL of the unconjugated (1:200) IL6Ra antibodywas left overnight at 4�C. The following day the ringswerewashedwith PBS and incubated with goat anti-rabbit Alexa 488 antibody(Life Technologies, A-11034) for 2 hours at room temperature.The plates were washed twice in PBS and the rings were removedfrom the 96-well plate, using a syringe needle, placed on amicroscope slide, andmountedwith ProlongGoldDAPI contain-ing medium (Invitrogen, cat. no. P36931). The slides were left todry and imaged using confocal microscopy (Zeiss LSM 510META).

Tissue cultureMouse lung endothelial cell (MLEC) was kindly given by

Professor Kairbaan Hodivala-Dilke, and was used for most of thein vitro studies. This cell linewas isolated and cultured as describedpreviously (16). Human umbilical vein endothelial cells(HUVEC; HPA Laboratories) were grown and cultured in endo-thelial growth medium (HPA Laboratories) and maintainedwithin 3 to 4 passages.

Ovarian cancer cell linesThe IGROV-1 linewas recently characterized as a hypermutated

line but unlikely to represent HGSC. The cell line was mycoplas-ma tested (InvivoGen) and always maintained within 4 to 5passages before new cells were recovered from frozen masterstocks. Cells were cultured in RPMI-1640 supplemented with10% FCS and 1% pen-strep. Cells were counted using a Vi-cellcell counter (Beckman Coulter) on days 3 and 7.

Staining of MLECA total of 2� 105MLEC cells were plated on a coverslip in a 12-

well plate. The coverslips were fixedwith 4% formaldehyde for 30minutes and washed with PBS (3 times for 5 minutes each).Following fixation, the cells were permeabilized with 0.1% Tritonfor 20 minutes and washed again in PBS. The cover slips wereincubated with (1:200) rabbit IgG (R&D), (1: 300) Endomucinantibody (Santa Cruz Sc-65495), IL6Ra antibody (1:200; SantaCruz C-20, Sc-661) overnight at 4�C. The following day, the coverslips were incubated with secondary (1:2000) goat anti-rabbitAlexa 488 (Life Technologies, A-11034) for 2 hours at roomtemperature and then mounted on slides with Prolong Gold

DAPI containingmedium (Invitrogen, P36931). The images weretaken using confocal microscopy (Zeiss LSM 510 META).

Western blottingA total of 2 � 105 MLEC or HUVECs were plated in a 6-well

plate with 2mLMLECmedium (10%FBS). The following day, thesupernatantwas removedand the cellswere treatedwithVEGF (30ng/mL), human IL6 (30ng/mL), ormouse IL6 (30ng/mL) in 2mLof serum-free MLEC medium for 6 hours or 24 hours. Cells werethen washed with PBS and harvested using RIPA Buffer (R0278,Sigma) with 1� proteinase inhibitors. Protein quantification wasperformed using the Bradford reagent (Sigma-Aldrich), accordingto the manufacturer's instructions. Cell extracts (25 mg) were runon a NuPAGE Novex 4% to 12% Bis-Tris Gels, 1.5 mm andtransferred to a nylon membrane. The membrane was blockedovernight (4�C in PBS with 0.1% Tween and 5% milk powder)and probed using the following antibodies: Jagged1 (1:1000,Abcam ab7771), DLL4 (1:1000, Abcam ab7280), Ang2 (1:500,Abcam ab8452), Hey1 (1:500, Abcam ab22614), Phospho-Stat3(Tyr705) (1:1000, Cell Signaling Technology 9145), Stat3(1:1000, Cell Signaling Technology 4904), p-ERK (1:1000, SantaCruz Biotechnology sc-7383), ERK (1:1000, Cell Signaling Tech-nology 9102),b-Actin (1:5000, SigmaA5316/). A rabbit ormousehorseradish peroxidase-conjugated secondary antibody (GEHealthcare) incubation allowed visualization using enhancedchemiluminescence (GE Healthcare). Protein concentrationequivalence was confirmed by anti–b-actin antibody.

Scratch wound migration assayConfluentmonolayers ofMLEC cells were scratchedwith a p20

pipette tip, cells were then washed twice with PBS before theaddition of serum-free growth medium with VEGF 100 ng/mL,hIL6 100 ng/mL, and mIL6 100 ng/mL. Wounds were monitoredby time lapse microscopy using an Olympus IX81 MicroscopeHamamatsu Orca ER digital camera. Images were acquired every30 minutes and subsequently analyzed using ImageJ software.

Cell proliferation assayOf note, 70,000MLEC cells were seeded in a 24-well plate. The

following day cells were treated with either 50 or 100 ng/mL ofVEGF, human IL6, ormouse IL6. The plate was then incubated for72 hours at 37�C and 5% CO2. After incubation, the cells weretrypsinized and counted using the Vi-cell counter (BeckmanCoutler). Each condition was repeated in triplicate.

In vivo IGROV-1 xenograftsA total of 1� 107 luciferase-expressing IGROV-1 cells (IGROV-

1luc) were injected intraperitoneally (i.p.) into 20 g 6- to 8-week-old female BALB/c nu/numice (purchased fromCharles River, UKLtd.). After 24 hours, anti-IL6 antibody (MEDI5117) was pre-pared at a concentration of 20mg/kg in sterile, endotoxin-free PBSand mice were injected i.p with 200 mL of this solution twiceweekly for 4 weeks. Control mice were injected with 200 mL of 20mg/kg IgG control antibody. MEDI5117 is a human IgG1kmonoclonal antibody that binds to IL6 with sub-pM affinity andneutralizes it by preventing the binding to IL6Ra. MEDI5117 wasgenerated using phage display technology. It bears a triple muta-tion (referred to as YTE) in the Fc domain of the heavy chain thatextends its half-life in circulation. MEDI5117 is active in severalpreclinical cancer models, including NSCLC, prostate cancer,breast cancer, and ovarian cancer (H. Zhong and colleagues; MolCancer Ther, in revision). An IgG1 isotype control antibody was

Interleukin-6 Stimulates Defective Angiogenesis

www.aacrjournals.org Cancer Res; 75(15) August 1, 2015 3099




Gopinathan et al.

Cancer Res; 75(15) August 1, 2015 Cancer Research3100




generated by Medimmune and was used at the same concentra-tion as MEDI5117. To visualize the architecture of blood vessels,some of the animals were anesthetized after 4 weeks of treatmentand injected with FITC-conjugated Lycopersicon esculentum lectin(tomato lectin; 100 mL, 2 mg/mL; Vector Laboratories) via the tailvein 3 minutes before animals were perfused with 4% parafor-maldehyde under terminal anesthesia. Samples were processedand frozen sections were cut for immunocytochemistry analysisfor pericytemarkers. The frozen sections from the lectin stained invivo sections were permeabilized in 0.5% Triton X-100 for 30minutes and washed three times in PBS. The sides were incubatedovernight with 1:100 anti-a-SMA FITC (abcam, ab8211), 1:100

Mouse IgG2a (FITC) isotype control (ab81197), 1:50 or anti-NG2(Millipore, ab5320) at 4�C. Primary antibody was washed withPBS. Slides were counterstained with DAPI Prolong Gold (Invi-trogen) and images captured by confocal microscopy (Zeiss LSMS10 META).

Protein extraction from mouse tumorsOf note, 75 mg of tumor tissue was lysed with 1 mL of ice-cold

lysis buffer (150mmol/LNaCl 20mmol/L Tris, pH7.5, 1mmol/LEDTA 1 mmol/L EGTA 1% Triton X-100) with protease andphosphatase Inhibitors. Samples were then dissociated usinggentleMACS Dissociator. After dissociation, samples were

Figure 2.IL6 has direct effects on mouselung endothelial cells. A,immunocytochemistry staining forendomucin (red) and IL6Ra (green) onMLEC. B, 70,000 MLECs were platedand treated with either PBS (control)or indicated concentrations of VEGF orIL6 for 72 hours. The cells were thentrypsinized and counted using a cellcounter and themean of the triplicateswas calculated. Proliferation assayshows a significant increase inproliferation with VEGF (50 or100 ng/mL), human IL6 (100 ng/mL),and mouse IL6 (50 or 100 ng/mL). C,scratch assay using a time-lapsemicroscope was used to measure themigration of MLEC after treatmentwith VEGF and IL6. Significantincreases in cell migration wereobserved with VEGF or IL6-treatedMLEC after 16 hours. Statisticalanalysis carried out using Studentt test is shown as � , P � 0.05;�� , P� 0.01; �� , P� 0.001. D, Westernblot analysis of protein extracted fromMLEC treated with IL6 (30 ng/mL) orVEGF (30 ng/mL) for 24 hours.IL6-induced downstream pSTAT3 andVEGF-induced downstream pERKlevels, indicating both pathways areindependent of each other forsignaling in MLEC.

Figure 1.IL6 stimulates angiogenesis in the aortic ring assay. A, phase contrast images of aortic rings embedded in collagen I, with the indicated concentrations of VEGF orhuman IL6. Human and mouse IL6 experiments were carried out with aortas isolated from wild-type C57BL/6 mice (8–12 weeks) and rat IL6 experimentswere carried outwith aortas isolated fromWistar ratsweighing 180 to 200g. B, angiogenic sproutswere counted after aweek in culture formouse aortic rings or after4 days in culture for the rat aortic rings following treatment with VEGF, hIL6 or mIL6, or rIL6. Significant increases in microvessel sprouting were observed inVEGF and IL6–treated rings (n¼ 9 per group) compared with PBS-treated controls; however, no difference in number of sprouts was observed between VEGF andIL6 treated rings. The P value for medium versus VEGF (P < 0.058) and medium versus VEGF (0.054) just failed to reach significance. Statistical analysis carriedout using Student t test is shown as ��, P� 0.01; �� , P� 0.001. C, length of sprouts was measured using ImageJ analysis. Mean length of sprouts (n¼ 9 per group)shows no significant difference between the VEGF and IL6-treated sprouts. D, 10 nmol/L of VEGFRi inhibited vessel sprouting in VEGF (10 ng/mL)-treatedrings but not in rat IL6 (10 ng/mL)-treated rings. Statistical analysis carried out using Student t test is shown as �� , P � 0.01; �� , P � 0.001. E, aortic ring vesselstained for endothelial cells using BS1 lectin (green) and for IL6Ra (red). F, Western blot analysis of protein extracted from mouse aortas treated with VEGF(30 ng/mL) or mouse IL6 (30 ng/mL). Mouse IL6 activated downstream pSTAT3 and VEGF induced downstream pERK levels.






centrifuged at 1,500 rpm for 2 minutes. Samples are always kepton ice between procedures. Next, using a probe sonicator set at40% amplitude, tissues were sonicated for 5 to 15 seconds bursts.Sonicated samples were then rotated for 30 minutes at 4�Cfollowed by a centrifugation for 15 minutes at 13,200 rpm at4�C. The pellet was discarded and protein concentration mea-sured. Lysates were frozen at �80�C until loaded on gels.

IHCParaffin-embedded sections of tumor sections collected from

IGROV-1 mouse xenograft were stained with antibodies forJagged1 (1:100, R&D Systems, AF1277), DLL4 (1:100, R&DSystems, AF1389), and Ang2 (1:50, Abcam, AB8452). Slides werecounterstained with hematoxylin. Negative controls were isotypematched. The conditions used for staining with individual anti-bodies were in accord with manufacturers' recommendations.

Gene expression analysisTable generated fromaheatmap established in a previous study

(Coward and colleagues; ref. 16).

Statistical analysisStatistical analyses were carried out using Prism GraphPad

software. Statistical significance was calculated using Student ttest and c2 test. Findings are presented as SEM.

ResultsIL6 stimulates angiogenesis in the aortic ring assay

As we had found that anti-human IL6 reduced tumor bloodvessel density in human tumor xenografts (7), we tested theactivity of human IL6 in the aortic ring assay. This is an ex vivomodel of angiogenesis that studies the effects of mediators onnormal vessel sprouting. Optimal vessel sprouting was observedin themouse aortic ring assay 7 to 10 days after treatment with 30ng/mLVEGFor 50ng/mLhIL6 (Fig. 1A).Mouse IL6 (30ng/mL) inmouse aortas and rat IL6 (10 ng/mL) in rat aortas also stimulatedvessel sprouting and there was no significant difference in thenumber of sprouts between the VEGF and IL6 treatments (Fig.1B). VEGF and IL6 treatments also gave similar results in terms oflength of vessel sprouts (Fig. 1C).

Other reports have described that IL6 has indirect angiogenicactivity via stimulation of VEGF production, and thus inhibitionof VEGF action would also block IL6 activity (17). We thereforetested the actionof theVEGF receptor inhibitor cediranib in the rataortic ring assay. Cediranib significantly inhibited the sproutingactivity of VEGF but had no significant activity on the actions ofIL6 in this model (Fig. 1D). This result suggested that in the aorticring model IL6 may stimulate vessel sprout formation withoutinducing VEGF. To demonstrate that IL6Ra is required for IL6-mediated angiogenesis, we added an anti-IL6Ra antibody to theIL6-simulated aortic ring cultures. This abolished the effect of IL6

Figure 3.Pericyte coverage is defective inIL6-stimulated vessels. Aortic ringsgrown in type I collagen were stainedfor endothelial cells using BS1 lectin(green) and antibodies to a-SMA (red)after treatment with VEGF (top) orhuman IL6 (A), mouse IL6 (B), or rat IL6(bottom; C). Confocal microscopytaken using a Zeiss LSM 510 laser-scanning microscope (�40) showsVEGF-treated vessels with goodpericyte coverage and IL6-treatedvessels with poor pericyte coverage.D, significant difference is observed inpericyte coverage between the VEGF-and IL6-treated vessels. Statisticalanalysis carried out using Student t testis shown as � , P � 0.05; �� , P � 0.01(n ¼ 5–9 aortas per condition).

Gopinathan et al.





(data not shown). We also found that the endothelial cells of theaortic ring vessels expressed IL6 receptor, staining positive for thegp80 IL6Ra; Fig. 1E). Furthermore, Western blot analyses oflysates from the aortic ring assay showed strong induction ofSTAT3phosphorylation bymouse IL6 but onlyweak induction byVEGF (Fig. 1F and Supplementary Fig. S1). In contrast, there wasstrong phosphorylation of ERK in VEGF-stimulated aortic ringlysates but IL6 had no effect (Fig. 1F and Supplementary Fig. S1).

As the aortic ring assay has a mixture of cells that couldcomplicate any analysis of signaling pathways, we used a simplerexperimental system in our next set of experiments, the MLECmouse endothelial cell line (16).

IL6 has direct effects on endothelial cellsMLECs stained with the endothelial cell marker endomucin

express IL6Ra (Fig. 2A), and both human and mouse IL6 at 100ng/mL stimulatedMLECproliferation (Fig. 2B). The IL6 effect wasnot as potent as that of VEGF but was still significant. VEGF andhuman andmouse IL6 were equally potent in the "scratch" assay,which measured MLEC migration (Fig. 2C). We then used West-ern blotting to study the VEGF and IL6 signaling pathways in thesecells. As expected, IL6 increased STAT3 phosphorylation and

VEGF increased ERK phosphorylation, but IL6 stimulation didnot affect phospho-ERK levels and VEGF had no effect on phos-pho-STAT3 (Fig. 2D and Supplementary Fig. S2) after 24 hours oftreatment. Similar results were observed in HUVEC after treat-ment with human IL6 (hIL6) or VEGF (Supplementary Fig. S3).Earlier time points (2 and 6 hours) were also studied with similarresults (data not shown). Thus, we concluded that in both theaortic ring assay and in cultures ofMLECcells, IL6has direct effectson endothelial cells.

Pericyte coverage is defective in IL6-stimulated vesselsOver the 7 to 10 days mouse aortic ring assay incubation

period, and the 4 days of the rat aortic ring assay, pericytes alsodevelop around the endothelial cells of the vessel sprouts whenthey are stimulated by VEGF. Using a-SMA as a pericyte marker,we treated mouse aortic rings with VEGF, mouse or human IL6,and rat aortic rings with rat IL6, and we noticed that pericytecoverage was diminished in the IL6-stimulated cultures. Overall,there were less pericytes attached to the tips of the IL6-stimulatedvessels compared with the VEGF cultures and many detachedpericytes were observed in the presence of IL6 (Fig. 3A–C). Thisdifference between number of pericytes associatedwith the sprout

Figure 4.Differential regulation of VEGF- andIL6-mediated Notch ligands and Ang2may explain defective pericytecoverage. A total of 2 � 105 MLEC cellswas plated and treated with control(PBS), VEGF (30 ng/mL), or mIL6 (30ng/mL) for 24 hours. A, Western blotanalysis of Jagged1 and DLL4expression on MLEC cells. B, theinduction of Jagged1 by IL6 and DLL4by VEGF was further confirmed in theprotein analysis of aortas treated withIL6 or VEGF. C, hey upregulation in theVEGF-treated MLEC confirmed thepositive VEGF-induced Notch-DLL4interaction in theMLEC.D,Western blotanalysis of Ang2 in the MLEC treatedwith VEGF or IL6 shows similarupregulation of Ang2 in the MLECtreated with IL6 compared with VEGF.E, Ang2 expression in protein isolatedfrom three aortas per group from wild-type C57BL/6 mice (8–12 weeks) andtreatedwith either VEGF (30 ng/mL) ormouse IL6 (30 ng/mL) for 24 hours. F,model formechanismof action of VEGFor IL6 on endothelial cells. VEGFbinding to VEGF receptor leads toupregulation of DLL4, resulting inangiogenesis and maturation ofvessels. However, IL6 forms a complexwith IL6 receptors to activate Jagged1and Ang2, resulting in angiogenesiswith weak or defective pericytecoverage.






Figure 5.Effects of anti-IL6 antibody on in vivo IGROV-1 vasculature. A total of 1 � 107 IGROV-1-luc cells were injected i.p. into 8 weeks old Balb/c nude female mice. After24 hours, the mice were treated with either control IgG antibody (20 mg/kg) or anti-IL6 antibody (20 mg/kg) twice a week for 4 weeks. Following this,three of the animals from each groupwere injected with TRITC-conjugated Bandeiraea via the tail vein, 3 minutes before being perfused with 4% paraformaldehyde.The lectin-stained frozen sections were stained for a-SMA (green; A) or NG2 (green; B). (Continued on the following page.)

Gopinathan et al.





tips in VEGF and IL6 cultures was significant (Fig. 3D). Similarresults were observed with another pericyte marker NG2 (Sup-plementary Fig. S4).

Differences in VEGF and IL6 signaling may explain defectivepericyte coverage

We hypothesized that differential regulation of Notch familymembers may be involved in the differences in pericyte coveragebetween IL6- and VEGF-treated cultures. High levels of the Notchligand DLL4 in endothelial cells are associated with vessel matu-rity and good pericyte coverage (18–20). Another Notch ligand,Jagged1, is associated with increased vessel sprouting and lessmature vessels (21, 22). We first investigated this hypothesis inMLEC cells. Using Western blotting, we found that 24 hours oftreatment with IL6 and VEGF differentially regulated the expres-sion of DLL4 and Jagged1. Although IL6 stimulatedmore Jagged1thandidVEGF, VEGFhad a greater effect onDLL4 thanon Jagged1(Fig. 4A and Supplementary Fig. S5A). We also observed thisdifferential regulation in IL6 and VEGF-treated aortic ring lysates(Fig. 4B and Supplementary Fig. S5B). VEGF stimulation of DLL4leads to increased activation of the HEY transcription factor (21).VEGF treatment of MLEC cells increased levels of HEY, whereasIL6 hadno effect (Fig. 4C and Supplementary Fig. S5A). As Ang2 isimplicated in the detachment of pericytes fromblood vessels (23–29), we next investigated whether IL6 also induced Ang2. Figure4D shows that both VEGF and IL6 induced Ang2 in both MLECand the aortic ring model, but IL6 had a stronger effect (Fig. 4Dand E and Supplementary Fig. S5A and S5B). An earlier time pointof 6 hours treatment in MLEC with mIL6 or VEGF gave similarresults (Supplementary Fig. S5C). This was also observed inHUVEC after treatment with hIL6 or VEGF (Supplementary Fig.S6). Figure 4F is a summary diagram showing mechanism ofaction of VEGF or IL6 on endothelial cells. Binding of VEGF toVEGF receptor leads to upregulation of DLL4, resulting in angio-genesis and maturation of vessels. However, IL6 binds to IL6receptors to activate Jagged1 and Ang2, resulting in angiogenesiswith defective pericyte coverage.

Relevance of these findings to malignant diseaseOur results so far would suggest that anti-IL6 treatment would

increase pericyte coverage of tumor blood vessels. We previouslyreported that when peritoneal xenografts of IGROV-1 ovariancancer cells, that constitutively produce IL6, were treated withanti-human IL6 antibodies, vessel density was reduced, as wasJagged1 mRNA and protein (7). We repeated these experimentswith the anti-human IL6 neutralizing antibody MEDI5117, thistime assessing pericyte coverage of the vessels. Using a-SMA (Fig.5A) andNG2 (Fig. 5B) as pericytemarkers, we found that anti-IL6treatment increased pericyte coverage. The sections were scoredblind for the level of pericyte coverage of the tumor blood vesselsand the effect of anti-IL6 was statistically significant (Fig. 5C;Supplementary Fig. S7A shows an example of the way pericytecoverage was scored).

In addition, IHC of the anti-IL6–treated tumors showeddecreased Jagged 1 and Ang2 staining and increased DLL4 protein(Fig. 5D). The IHC results were scored blind and the results areshown in Supplementary Fig. S7B. This was confirmed inWesternblot analyses of tumor lysates (Fig. 5E and SupplementaryFig. S7C).

Correlations from human ovarian cancer biopsiesFinally, we revisited a previously published analysis of mRNAs

that are associated with high IL6 expression (7) using publicallyavailable datasets of ovarian cancer. Ang2 and Jagged1 hadsignificant a coefficient of correlation with high IL6 expression(P¼ 0.015 and P¼ 0.001, respectively; Fig. 6A). Further in-depthanalysis of these associations was carried out on RNAseq datafrom 27 samples of omental metastases from patients with stageIII/IV high-grade serous ovarian cancer (HGSC). The samplesweredivided into three groups based on histologic analysis; unin-volved omental tissue, established tumor, and stroma with lowtumor burden postchemotherapy. Figure 6B shows typical his-tology of each group. Only within the established tumor groupswas a strong and significant correlation (Spearman r) seenbetween IL6, Jagged 1, and Ang2 mRNA (Fig. 6C). There was nosignificant correlation between IL6 and DLL4 mRNA in the samesamples (data not shown).

DiscussionA role for IL6 in pathogenic angiogenesis has been suggested in

diseases such as stroke, rheumatoid arthritis, and various cancers(1, 30, 31). In aprevious publication,we showed that treatment ofovarian cancer xenografts with an anti-IL6 antibody reduced thetumor vasculature with concomitant inhibition of the NOTCHligand Jagged1,which has been implicated in vessel sprouting (7).Collectively, the published literature suggested that IL6 coulddrive abnormal angiogenesis and that the anti-IL6 antibody had apotential as antiangiogenic agent. Thus, we investigated the directeffects of IL6 on normal angiogenesis using endothelial in vitroand ex vivo studies, and used the findings from those studies toexplore its importance in tumor angiogenesis using peritonealmodels of ovarian cancer and ovarian cancer biopsies.

We found that IL6 is as potent as VEGF in inducing vesselsprouting in the aortic ring assay and was also able to stimulateendothelial cellmigration andproliferation inMLEC cells. VEGFRinhibition studies in the aortic ring model and protein analysis ofdownstream VEGF and IL6 signaling indicate that the angiogeniceffects observed with IL6may not depend on VEGF in endothelialcells. The effects on IL6 on malignant cells may be differentbecause of the complex autocrine signaling networks generatedin cancer cells. Preliminary experiments on malignant cell linesshowed that, in contrast with endothelial cells, IL6 can stimulateVEGF signaling and vice versa. In addition, in the RNAseq experi-ments of Fig. 6, we found that there was a positive associationbetween IL6 and VEGF mRNA levels, but this would be expected

(Continued.) Treatment with anti-IL6 antibody was able to restore the pericyte coverage in IGROV-1 xenografts. Staining of a-SMA and NG2 staining shows pooror detached pericyte coverage in the IgG antibody group compared with good pericyte coverage in the anti-IL6–treated group. C, quantification of thestaining shows a significant difference between the IGROV-1 antibody control and anti-IL6–treated group (P < 0.001). The quantification carried out by twoindependent reviewers was analyzed using c2 test (n ¼ 6 per group from two independent experiments). D, IHC staining for Jagged1, DLL4, and Ang2 in theIGROV-1 xenograft following 4 weeks of treatment with anti-IL6 antibody. The images were taken from five randomly selected areas per tumor section(n ¼ 6 per group from two independent experiments) using a �40 microscope. E, Western blot analysis of DLL4 and Ang2 from protein extracted from in vivoIGROV-1 xenografts treated with anti-IL6 antibody.






when studying isolates from a complex multicellular tumormicroenvironment. Hence, we suspect that IL6may have differenteffects on malignant cells and endothelial cells.

We found that the angiogenesis stimulated by IL6 leads toformationof vasculaturewithdefective pericyte coverage, and thatanti-IL6 treatment of the peritoneal ovarian cancer xenograftsleads to restoration of pericytes on the blood vessels. Investigatingthe mechanism of action of IL6 and VEGF in inducing thisdifferent phenotype of vessel maturation has shown roles forNotch ligands and Ang2.

This is, to our knowledge, the first study showing that IL6 caninduce a type of vessel sprouting with abnormal pericyte coveragecompared with VEGF. These observations have clinical implica-tions for malignant and other diseases, especially as studies invarious cancers suggest that pericyte depletion leads to increasedmetastasis (23, 32, 33). The ability of anti-IL6 treatment toimprove pericyte coverage of vessels in xenograft models suggeststhat in the tumor microenvironment, defective pericyte coveragemay be due to the action of IL6 in those tumors with high levels ofthis cytokine.

The regulation of Ang2 by IL6 is another interesting findingas Ang2 expression has been shown to correlate with lymph

node metastasis in various cancers (34–36). Moreover, theangiopoietin inhibitor trebananib increased progression-freesurvival in patients with recurrent ovarian cancer (37).

As STAT3 signaling is implicated in the treatment failure ofvarious anti-angiogenic agents (13, 38–40), combinations ofIL6 and angiogenesis antagonists may be worthy of furtherstudy.

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

Authors' ContributionsConception and design: G. Gopinathan, F. Balkwill, C. Milagre, R. Thompson,J.R. WhitefordDevelopment of methodology: G. Gopinathan, C. Milagre, K. Hodivala-Dilke,D.A. Leinster, R. Thompson, J.R. WhitefordAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): G. Gopinathan, C. Milagre, H. Zhong, J.R. WhitefordAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis):G.Gopinathan, F. Balkwill, C.Milagre, O.M.T. PearceWriting, review, and/or revisionof themanuscript:G.Gopinathan, F. Balkwill,C. Milagre, O.M.T. Pearce, L.E. Reynolds, H. Zhong, R.E. Hollingsworth

Figure 6.Relevance of these findings tomalignant disease. Gene expression data from 285 ovarian cancer biopsies from the Australian Ovarian Cancer Study (AOCS) datasetalong with 245 samples obtained by merging two other publicly available datasets were ranked for expression of IL6 pathway genes. Then, 50 sampleswith the highest (high IL6) and 50 samples with the lowest (low IL6) levels of expression were selected and from that a list was generated of the differentiallyexpressed genes between the high and low IL6 samples. These differentially expressed genes were then associated with various pathways and processes.Significant associations were found between the high IL6 pathway gene expression and with various processes including angiogenesis. A, the high IL6 pathwayexpression correlated positively with Jagged1 and Ang2. RNAseq analysis of omental metastases from patients with stage III–IV high-grade serous ovariancancer. The 27 tumor samples were divided into three groups based on histologic analysis: uninvolved omental tissue, established tumors, and stroma with lowtumor burden post chemotherapy. B, significant positive correlation was observed between expression levels for IL6, Jagged 1, and Ang2 mRNA, and the best-fitcorrelation (Spearman r) was found in samples that had a high burden of tumor.


Gopinathan et al.




Administrative, technical, or material support (i.e., reporting or organizingdata, constructingdatabases):G.Gopinathan,R.E.Hollingsworth,R. ThompsonStudy supervision: F. Balkwill, C. Milagre, R.E. Hollingsworth, J.R. Whiteford

Grant SupportThis work was supported by Arthritis Research UK Grant 19207, Cancer

Research UK ProgrammeGrant C587/A16354, and European Research CouncilAdvanced Grant 322566.

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 May 11, 2015; accepted May 13, 2015; published OnlineFirst June16, 2015.

References1. Taniguchi K, Karin M. IL6 and related cytokines as the critical lynchpins

between inflammation and cancer. Semin Immunol 2014;26:54–74.2. Hong DS, Angelo LS, Kurzrock R. Interleukin-6 and its receptor in cancer:

implications for translational therapeutics. Cancer 2007;110:1911–28.3. Ancrile B, Lim KH, Counter CM. Oncogenic Ras-induced secretion of IL6 is

required for tumorigenesis. Genes Dev 2007;21:1714–9.4. Grivennikov S, Karin E, Terzic J, Mucida D, Yu GY, Vallabhapurapu S, et al.

IL6 and Stat3 are required for survival of intestinal epithelial cells anddevelopment of colitis-associated cancer. Cancer Cell 2009;15:103–13.

5. Chang Q, Bournazou E, Sansone P, Berishaj M, Gao SP, Daly L, et al. TheIL6/JAK/Stat3 feed-forward loop drives tumorigenesis and metastasis.Neoplasia 2013;15:848–62.

6. Kulbe H, Chakravarty P, Leinster DA, Charles KA, Kwong J, Thompson RG,et al. A dynamic inflammatory cytokine network in the human ovariancancer microenvironment. Cancer Res 2012;72:66–75.

7. Coward J, Kulbe H, Chakravarty P, Leader D, Vassileva V, Leinster DA, et al.Interleukin-6 as a therapeutic target in human ovarian cancer. Clin CancerRes 2011;17:6083–96.

8. Stone RL, Nick AM, McNeish IA, Balkwill F, Han HD, Bottsford-Miller J,et al. Paraneoplastic thrombocytosis in ovarian cancer. N Engl J Med2012;366:610–8.

9. NilssonMB, Langley RR, Fidler IJ. Interleukin-6, secreted by humanovariancarcinoma cells, is a potent proangiogenic cytokine. Cancer Res 2005;65:10794–800.

10. Xu Q, Briggs J, Park S, Niu G, Kortylewski M, Zhang S, et al. Targeting Stat3blocks both HIF-1 and VEGF expression induced by multiple oncogenicgrowth signaling pathways. Oncogene 2005;24:5552–60.

11. Yao JS, Zhai W, YoungWL, Yang GY. Interleukin-6 triggers human cerebralendothelial cells proliferation andmigration: the role for KDR andMMP-9.Biochem Biophys Res Commun 2006;342:1396–404.

12. Fan Y, Ye J, Shen F, Zhu Y, Yeghiazarians Y, Zhu W, et al. Interleukin-6stimulates circulating blood-derived endothelial progenitor cell angiogen-esis in vitro. J Cereb Blood Flow Metab 2008;28:90–8.

13. de Groot J, Liang J, Kong LY, Wei J, Piao Y, Fuller G, et al. Modulatingantiangiogenic resistance by inhibiting the signal transducer and activatorof transcription 3 pathway in glioblastoma. Oncotarget 2012;3:1036–48.

14. KwonKA, KimSH,Oh SY, Lee S,Han JY, KimKH, et al. Clinical significanceof preoperative serum vascular endothelial growth factor, interleukin-6,andC-reactive protein level in colorectal cancer. BMCCancer 2010;10:203.

15. Nicosia RF. The aortic ring model of angiogenesis: a quarter century ofsearch and discovery. J Cell Mol Med 2009;13:4113–36.

16. Reynolds LE, Hodivala-Dilke KM. Primary mouse endothelial cell culturefor assays of angiogenesis. Methods Mol Med 2006;120:503–9.

17. Cohen T, Nahari D, Cerem LW, Neufeld G, Levi BZ. Interleukin 6 inducesthe expression of vascular endothelial growth factor. J Biol Chem1996;271:736–41.

18. Patel NS, Dobbie MS, Rochester M, Steers G, Poulsom R, Le Monnier K,et al. Up-regulation of endothelial delta-like 4 expression correlates withvessel maturation in bladder cancer. Clin Cancer Res 2006;12:4836–44.

19. Scehnet JS, Jiang W, Kumar SR, Krasnoperov V, Trindade A, Benedito R,et al. Inhibition of Dll4-mediated signaling induces proliferation of imma-ture vessels and results in poor tissue perfusion. Blood 2007;109:4753–60.

20. Schadler KL, Zweidler-McKay PA, Guan H, Kleinerman ES. Delta-likeligand 4 plays a critical role in pericyte/vascular smooth muscle cellformation during vasculogenesis and tumor vessel expansion in Ewing'ssarcoma. Clin Cancer Res 2010;16:848–56.

21. Benedito R, Roca C, S€orensen I, Adams S, Gossler A, Fruttiger M, et al. Thenotch ligandsDll4 and Jagged1have opposing effects on angiogenesis. Cell2009;137:1124–35.

22. Kume T. Novel insights into the differential functions of Notch ligands invascular formation. J Angiogenesis Res 2009;1:8.

23. Gerhardt H, Semb H. Pericytes: gatekeepers in tumour cell metastasis? JMol Med 2008;86:135–44.

24. QinD, Trenkwalder T, Lee S, ChilloO,Deindl E, Kupatt C, et al. Early vesseldestabilization mediated by Angiopoietin-2 and subsequent vessel mat-uration via Angiopoietin-1 induce functional neovasculature after ische-mia. PLoS ONE 2013;8:e61831.

25. Feng Y, vom Hagen F, Pfister F, Djokic S, Hoffmann S, Back W, et al.Impaired pericyte recruitment and abnormal retinal angiogenesis as aresult of angiopoietin-2 overexpression. Thromb Haemost 2007;97:99–108.

26. Hammes HP, Lin J, Wagner P, Feng Y, Vom Hagen F, Krzizok T, et al.Angiopoietin-2 causes pericyte dropout in the normal retina: evidence forinvolvement in diabetic retinopathy. Diabetes 2004;53:1104–10.

27. Keskin D, Kim J, Cooke VG, Wu CC, Sugimoto H, Gu C, et al. Targetingvascular pericytes in hypoxic tumors increases lung metastasis via angio-poietin-2. Cell Rep 2015;10:1066–81.

28. Ribatti D, Nico B, Crivellato E. The role of pericytes in angiogenesis. Int JDev Biol 2011;55:261–8.

29. Armulik A, Abramsson A, Betsholtz C. Endothelial/pericyte interactions.Circ Res 2005;97:512–23.

30. Gertz K, Kronenberg G, K€alin RE, Baldinger T, Werner C, Balkaya M, et al.Essential role of interleukin-6 in post-stroke angiogenesis. Brain 2012;135:1964–80.

31. Kayakabe K, Kuroiwa T, Sakurai N, Ikeuchi H, Kadiombo AT, Sakairi T,et al. Interleukin-6 promotes destabilized angiogenesis by modulatingangiopoietin expression in rheumatoid arthritis. Rheumatology 2012;51:1571–9.

32. Xian X, Ha�kansson J, Sta

�hlberg A, Lindblom P, Betsholtz C, Gerhardt H,

et al. Pericytes limit tumor cell metastasis. J Clin Invest 2006;116:642–51.33. CookeVG, LeBleuVS,KeskinD,KhanZ,O'Connell JT, TengY, et al. Pericyte

depletion results in hypoxia-associated epithelial-to-mesenchymal transi-tion and metastasis mediated by met signaling pathway. Cancer Cell2012;21:66–81.

34. Jo MJ, Lee JH, Nam BH, Kook MC, Ryu KW, Choi IJ, et al. Preoperativeserum angiopoietin-2 levels correlate with lymph node status in patientswith early gastric cancer. Ann Surg Oncol 2009;16:2052–7.

35. Sfiligoi C, de Luca A, Cascone I, Sorbello V, Fuso L, Ponzone R, et al.Angiopoietin-2 expression in breast cancer correlates with lymph nodeinvasion and short survival. Int J Cancer 2003;103:466–74.

36. Schulz P, Fischer C, Detjen KM, Rieke S, Hilfenhaus G, von Marschall Z,et al. Angiopoietin-2 drives lymphatic metastasis of pancreatic cancer.FASEB J 2011;25:3325–35.

37. Monk BJ, Poveda A, Vergote I, Raspagliesi F, Fujiwara K, Bae DS, et al. Anti-angiopoietin therapy with trebananib for recurrent ovarian cancer (TRI-NOVA-1): a randomised, multicentre, double-blind, placebo-controlledphase 3 trial. Lancet Oncol 2014;15:799–808.

38. Jain RK, Duda DG, Willett CG, Sahani DV, Zhu AX, Loeffler JS, et al.Biomarkers of response and resistance to antiangiogenic therapy. Nat RevClin Oncol 2009;6:327–38.

39. Willett CG,DudaDG, di Tomaso E, Boucher Y, AncukiewiczM, Sahani DV,et al. Efficacy, safety, and biomarkers of neoadjuvant bevacizumab, radi-ation therapy, and fluorouracil in rectal cancer: amultidisciplinary phase IIstudy. J Clin Oncol 2009;27:3020–6.

40. Zhu AX, Sahani DV, DudaDG, di Tomaso E, AncukiewiczM, CatalanoOA,et al. Efficacy, safety, andpotential biomarkers of sunitinibmonotherapy inadvancedhepatocellular carcinoma: a phase II study. J ClinOncol 2009;27:3027–35.






2015;75:3098-3107. Published OnlineFirst June 16, 2015.Cancer Res Ganga Gopinathan, Carla Milagre, Oliver M.T. Pearce, et al. Interleukin-6 Stimulates Defective Angiogenesis

Updated version

10.1158/0008-5472.CAN-15-1227doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2015/06/25/0008-5472.CAN-15-1227.DC2

http://cancerres.aacrjournals.org/content/suppl/2015/06/18/0008-5472.CAN-15-1227.DC1Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/75/15/3098.full#ref-list-1

This article cites 71 articles, 38 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/75/15/3098.full#related-urls

This article has been cited by 5 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://cancerres.aacrjournals.org/content/75/15/3098To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-15-1227

http://cancerres.aacrjournals.org/content/suppl/2015/06/18/0008-5472.CAN-15-1227.DC1 http://cancerres.aacrjournals.org/content/suppl/2015/06/25/0008-5472.CAN-15-1227.DC2

http://cancerres.aacrjournals.org/content/suppl/2015/06/18/0008-5472.CAN-15-1227.DC1 http://cancerres.aacrjournals.org/content/suppl/2015/06/25/0008-5472.CAN-15-1227.DC2

http://cancerres.aacrjournals.org/content/75/15/3098.full#ref-list-1

http://cancerres.aacrjournals.org/content/75/15/3098.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/75/15/3098


Interleukin-6StimulatesDefectiveAngiogenesiscancerres.aacrjournals.org/content/canres/75/15/3098.full.pdf ·...

Documents

Transcript of Interleukin-6StimulatesDefectiveAngiogenesiscancerres.aacrjournals.org/content/canres/75/15/3098.full.pdf ·...