Resistance to Anti-VEGF Therapy Mediated by Autocrine IL6/STAT3 Signaling and Overcome by IL6 Blockade Alexandra Eichten, Jia Su, Alexander P

Home , IL36G, Interleukin 10, Interleukin 6

Published OnlineFirst February 26, 2016; DOI: 10.1158/0008-5472.CAN-15-1443 Cancer Therapeutics, Targets, and Chemical Biology Research

Resistance to Anti-VEGF Therapy Mediated by Autocrine IL6/STAT3 Signaling and Overcome by IL6 Blockade Alexandra Eichten, Jia Su, Alexander P. Adler, Li Zhang, Ella Ioffe, Asma A. Parveen, George D. Yancopoulos, John Rudge, Israel Lowy, Hsin Chieh Lin, Douglas MacDonald, Christopher Daly, Xunbao Duan, and Gavin Thurston

Abstract

Anti-VEGF therapies benefit several cancer types, but drug ilarly, inhibition of IL6R enhanced the antitumor effects of resistance that limits therapeutic response can emerge. We gen- aflibercept in DU145 prostate tumor cells that displays high erated cell lines from anti-VEGF–resistant tumor xenografts to endogenous IL6R activity. In addition, post hoc stratification of investigate the mechanisms by which resistance develops. Of all data obtained from a clinical trial investigating aflibercept efficacy tumor cells tested, only A431 (A431-V) epidermoid carcinoma in ovarian cancer showed poorer survival in patients with high cells developed partial resistance to the VEGF inhibitor afliber- levels of circulating IL6. These results suggest that the activation of cept. Compared with the parental tumors, A431-V tumors secret- the IL6/STAT3 pathway in tumor cells may provide a survival ed greater amounts of IL6 and exhibited higher levels of phospho- advantage during anti-VEGF treatment, suggesting its utility as a STAT3. Notably, combined blockade of IL6 receptor (IL6R) and source of response biomarkers and as a therapeutic target to VEGF resulted in enhanced activity against A431-V tumors. Sim- heighten efficacious results. Cancer Res; 76(8); 1–13. 2016 AACR.

Introduction the environment of reduced tumor vessels (12–14). These resistance mechanisms cause insensitivity of tumor vasculature to anti- VEGF plays a key role in physiologic and pathologic angiogen- VEGF treatments and/or decreased dependence of tumors on esis, including tumor angiogenesis. Blockade of the VEGF path- angiogenesis. Combination treatments targeting VEGF and other way is effective at inhibiting angiogenesis in many tumors (1–3). angiogenic pathways such as angiopoietin-2 (Ang2) or Delta-like Several VEGF pathway inhibitors, including the monoclonal anti- 4 (Dll4) are being evaluated to increase the efficacy of angiogenic VEGF antibody bevacizumab, the soluble receptor aflibercept blockade. However, the underlying mechanisms of tumor cells (VEGF Trap, known as ziv-aflibercept in the United States), and becoming less dependent on angiogenesis in response to anti- the mAb to VEGF receptor 2 (ramucirumab) delay tumor growth VEGF therapies also need to be investigated. in preclinical tumor models (4–6) and extend the survival of To begin addressing this, we aimed to identify molecules cancer patients (7–11). involved in mediating anti-VEGF resistance using xenograft tumor Despite showing broad activity in both preclinical and clinical models with acquired resistance to VEGF blockade. Aflibercept is a settings, not all tumors respond to anti-VEGF therapies, and those recombinant fusion protein that potently binds all isoforms of that do may eventually become resistant. Clinical and preclinical human and murine VEGF-A, VEGF-B and placental growth factor studies suggest that resistance of tumors to anti-VEGF therapies (PlGF; ref. 4). A431 human epidermoid cell tumor xenografts are can occur via several mechanisms: (i) changes in the tumor sensitive to aflibercept, but can acquire resistance when treated for microenvironment resulting in upregulation of various proangio- longer time periods (several weeks or more). We established a genic factors, which lead to vessels that are less sensitive to VEGF stable variant cell line (A431-V), which is partially resistant to blockade and/or (ii) changes in the characteristics of cancer cells, aflibercept and allowed the direct comparison of resistant tumors such as mutations and epigenetic changes, which provide the cells to their sensitive parental A431 (A431-P) counterparts. Studies with a survival advantage and/or increased invasive potential in using these two cell lines revealed that A431-V tumors have increased levels of proinflammatory mediators including IL6, as Regeneron Pharmaceuticals, Tarrytown, New York. well as activation of the STAT3 signaling pathway. The activation of the IL6/STAT3 pathway led us to investigate the effects of an Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). anti-IL6R antibody (sarilumab), which blocks IL6 receptor signaling on human cells. Importantly, sarilumab does not bind or A. Eichten, J. Su, A.P. Adler, and L. Zhang contributed equally to this article. block murine IL6R. Treatment of A431-V tumor cells with sar- Corresponding Author: Gavin Thurston, Regeneron Pharmaceuticals, 777 Old ilumab reduced the phosphorylation of STAT3, and enhanced the Saw Mill River Road, Tarrytown, NY 10591. Phone: 914-847-7575; Fax: 914-847- antitumor activity of aflibercept against A431-V tumors in vivo. 7544; E-mail: [email protected] Similarly, IL6R inhibition enhanced the antitumor activity of doi: 10.1158/0008-5472.CAN-15-1443 aflibercept in another tumor model with high endogenous IL6R 2016 American Association for Cancer Research. activity, namely Du145 prostate tumors. In addition, IL6

www.aacrjournals.org OF1

Eichten et al.

overexpression rendered sensitive A431 cells resistant to afliber- manufacturer's instructions. Cytokines/chemokines assessed cept. Finally, IL6R blockade decreased the number of A431 were EGF, Eotaxin, FGF-2, Flt-3 ligand, fractalkine, G-CSF, GM- tumors that escaped long-term aflibercept treatment. To assess CSF, GRO, IFNa2, IFNg, IL1a, IL1b, IL1ra, IL2, sIL2ra, IL3, IL4, potential clinical relevance, we determined whether circulating IL5, IL6, IL8, IL9, IL10, IL12 (p40), IL12(p70), IL13, IL15, IL17, IP- levels of IL6 were associated with poorer outcome in a phase II 10, MCP-1, MCP-3, MDC, MIP-1a, MIP-1b, PDGF-AA, PDGF-AB/ clinical trial of ovarian cancer patients treated with aflibercept. A BB, RANTES, sCD40L, TGFa, TNFa, TNFb, and VEGFA. Samples correlation between high levels of IL6 and poorer tumor response were run using a Flexmap3D instrument (Luminex), and data to anti-VEGF therapy was observed. Taken together, these data were analyzed with Masterplex software. suggest that resistance to VEGF blockade can be mediated at least Phospho-STAT3 (Tyr705) levels in A431-P and A431-V cell in part by increased IL6/STAT3 signaling in tumor cells, and that lysates were assessed using Bio-Plex phosphoprotein detection kit blockade of IL6 signaling on tumor cells can overcome this (Bio-Rad) following manufacturer's instructions. The "relative resistance. fluorescent units" readout was used to compare the phosphorylation levels. Materials and Methods Tumor cells and reagents Flow cytometry Tumor cells including A431 human epidermoid cell skin car- A431-P and -V cells were treated with Enzyme Free Cell cinoma and Du145 human prostate carcinoma were obtained Dissociation Solution Hank's Based Solution (Millipore, cata- from the ATCC and grown according to ATCC guidelines. All cell log # S-004-C), blocked in 2% FBS/PBS and human truStain lines were authenticated between 2012 and 2015 using the STR FcX (BioLegend, catalog # 422302), incubated with anti-human Profile Testing by ATCC. A431 parental (P) and variant (V) cells IL6Ra-APC (catalog # 352805, BioLegend) for 1 hour at room fl were obtained by in vivo passaging, as described in Fig. 1. temperature and assessed by ow cytometry. Aflibercept (also known as VEGF Trap or ziv-aflibercept in the United States) is a recombinant fusion protein that potently binds In vivo tumor studies all isoforms of human and murine VEGF-A, VEGF-B, and PlGF. Tumor studies were performed in accordance with Regener- Sarilumab (REGN88) is a fully human IgG1 mAb that binds on's Institutional Animal Care and Use Committee guidelines. human IL6R and prevents binding of IL6. Cetuximab is a chimeric Atotalof1 106 A431-P or A431-V cells, or 5 106 Du145 mouse-human (30:70) IgG1 mAb that competitively inhibits the cells were implanted subcutaneously into CB.17 SCID mice binding of EGF to its receptor EGFR. (Taconic). Tumor volume was assessed with calipers using the L W2/2 formula. 100–200 mm3 tumor-bearing mice were Cell culture and in vitro cell growth randomized into groups and treated subcutaneously with fl A431 and Du145 cells were cultured in 10% FBS, DMEM, and 25 mg/kg hFc (control protein), a ibercept (VEGF Trap), sar- fl MEM, respectively. Preparation of conditioned media (CM) was ilumab, combination of a ibercept and sarilumab (25 mg/kg done by the following method: A431-P and A431-V cells were each), or cetuximab 2 times per week. Tumor growth curves: grown in serum-free media for 16 hours followed by media average mean SD. collection. A431-P and A431-V cells grown on a 6-well plate at a density of 1 106 cells/well or on 8-well chamber slides at a Generation of IL6-overexpressing A431 tumor cells density of 30,000 cells/well were serum starved for 16 hours Lentiviral pLOC vector encoding human IL6 and TurboGFP followed by exposure to CM for 1 hour. Cells were either lysed (nuc) reporter genes were purchased (Thermo Scientific) and to detect p-STAT3 status by Western blot analysis or fixed for packaged as per manufacturer's instructions. A431 carcinoma immunocytochemistry. Cell lysate preparation for the RTK array: cells were virally transduced with pLOC-IL6 lentivirus particles A431-V and A431-P cells were lysed in modified RIPA buffer and isolated by two rounds of flow cytometry cell sorting. (1% NP-40, 0.1% deoxycholate, 50 mmol/L Tris.HCl pH 7.4, 150 mmol/L NaCl, EDTA 1mmol/L, Na3VO4 1 mmol/L, NaF IHC and immunocytochemistry 5 mmol/L, glycerophosphate 5 mmol/L and Roche complete Cryosections. Optimal cutting temperature–embedded tissue was protease inhibitor cocktail tablets). cut into 5-mm sections, air dried, and fixed in acetone (20 C) for 10 minutes, and using Avidin-Biotin Blocking Kit (Vector Labo- 5 In vitro cell growth. A total of 1 10 A431-P or A431-V cells were ratories), blocked in 2.5 % normal goat serum/1 % BSA/PBS, Ki- seeded and cell numbers were assessed after a 3-day growth 67 antibody (BD; clone B56), biotinylated antibody (Vector ¼ t t 0 period. Doubling time (DT): DT (h) ( 0)log2/(logN Laboratories), and 3,3 -diaminobenzidine (DAB, Sigma). Tissue t t logN0), where and 0 are the times at which the cells were sections were counterstained with methyl green. N N t t counted, and and 0 are the cell numbers at times and 0, In Situ Cell Death Detection Kit (Sigma) was used according to respectively. the manufacturer's recommendations and apoptotic index was determined using HALO software (Indica Labs). RTK signaling antibody array PathScan RTK Signaling Array Kit (Cell Signaling Technology) FFPE tissue sections. Tumors were fixed in 10% neutral-buffered was used following manufacturer's instructions. formalin, dehydrated through graded ethanol and xylenes, embedded in paraffin, cut into 8-mm sections, and deparaffinized. Luminex analysis For p-STAT3 IHC, antigen retrieval was done, blocking was For A431-P and A431-V CM analysis, Milliplex Human Cyto- performed with 4% BSA, and incubation with p-STAT3 antibody kine/Chemokine 42 Plex Kit (Millipore) was used according to the (Cell Signaling Technology, catalog # 9145), HRP-conjugated

OF2 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

Reimplant (1x) A

Tumor growth stasis for 6–7 Re- 14-day Cell line ...... weeks, then implant treatment isolation 3 regrowth 100–150 mm A431 variant tumor-bearing mice Implant (A431-V)

50–100 mm3 A431 A431 tumor- Cells Continuous Re- 14-day Cell line bearing mice ...... tumor growth implant treatment isolation 100–150 mm3 A431 parental tumor-bearing mice (A431-P)

B C

800 1,500 A431-P - ctrl ) ) 3 Ctrl 3 A431-P - aflibercept 600 A431-V - ctrl A431-V ctrl 1,000 **** Aflibercept A431-V - aflibercept vs. P ctrl 400

500 A431-V afl 200 **** vs. V ctrl Tumor volume (mm Tumor volume (mm A431-P afl **** 0 vs. P ctrl 0 20 40 60 0 0 5 10 15 Days after treatment start Days post treatment start

D E

Ctrl Aflibercept Ctrl Aflibercept

20 ***

10 NS

% TUNEL-positivenuclei 0 A431-P A431-P c c F lib F lib h f h P V af - -P a - V 1 1 1 3 31- 43 4 43 A A A A4

1,500

1,000

500 Total nuclei analyzed nuclei Total

0 b A431-V A431-V li ib hFc fl P af a -P -V hFc -V 1 1 43 A431- A43 A A431

Figure 1. Generation and characterization of aflibercept-resistant A431 variants. A, schematic of A431-P and -V tumor cell line generation. B, SCID mice bearing A431 tumors (50–100 mm3) were treated with aflibercept or human Fc control protein for up to 7 weeks. C, growth kinetics and response to aflibercept treatment of tumors grown from in vivo passaged A431-V and A431-P cell lines. The average tumor volume SD is plotted over the course of treatment. , P < 0.0001 two- way ANOVA, Bonferroni post hoc test. D, representative images of cell proliferation assessed by Ki-67 IHC in A431-P and A431-V tumors treated with human Fc control protein or aflibercept for 14 days. Scale bar, 50 mm. E, representative images of cell apoptosis assessed by TUNEL staining in A431-P and A431-V tumors treated with human Fc control protein or aflibercept (aflib) for 24 hours. Scale bar, 10 mm. Quantitative analysis of apoptotic index based on TUNEL-positive (green) and -negative nuclei in A431-P and A431-V tumor tissue. , P < 0.001 one-way ANOVA with Bonferroni post hoc test. secondary antibody (Vector Laboratories, catalog # PI1000), and Triton X-100, 5% goat serum, 2.5% BSA in PBS, Ki-67 antibody 3,30-diaminobenzidine (DAB, Sigma) was done. For Ki-67 IHC, (BD, catalog # 556003), biotinylated antibody (Vector Labora- antigen retrieval was done and blocking was performed with 0.1% tories, catalog # BA-2001), ABC-ELITE (Vector Laboratories; ABC

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF3

Eichten et al.

VectaStain Elite), and 3,30-diaminobenzidine (DAB, Sigma). Tis- aflibercept from aflibercept-sensitive tumors. For this aim, we sue sections were counterstained with hematoxylin. selected four aflibercept-sensitive tumor lines, namely A498, Colo205, 786-0, and A431, and treated tumor-bearing mice with Immunocytochemistry. Cells were fixed in 4% PFA for 20 minutes, high-dose aflibercept (25 mg/kg, twice weekly), for up to 7 weeks. blocked for 45 minutes in 10% donkey serum/0.3% Triton X-100, For three of the four tumor lines, we observed complete growth incubated with rabbit anti-human phopho-STAT3 (Y705) anti- stasis (Colo205, A498) or gradual incremental tumor growth body (R&D Systems, catalog # AF4607), anti-rabbit NL557-con- (786-0), but no outgrowth of resistant tumors based on our jugated antibody (R&D Systems, catalog # NL004), and DAPI. criteria (Supplementary Fig. S1). However, A431 human epider- Slides were mounted in ProLong Gold. moid cell carcinoma tumors exposed to long-term high-dose aflibercept showed complete and prolonged tumor growth inhi- Immunoblotting analysis and ELISA assay bition for approximately 6 weeks followed by conspicuous out- Tumor cells or tissues were harvested in lysis buffer [50 mmol/L growth of some of the tumors (Fig. 1B and Supplementary Fig. S1 Hepes pH 8.0, 10% Glycerol, 1% Triton X-100, 150 mmol/L for individual tumor growth). One of the outgrowing A431 tumors was harvested, and frag- NaCl, 1 mmol/L EDTA, 1.5 mmol/L MgCl2, 100 mmol/L NaF, 10 ments of viable tumor tissue were reimplanted subcutaneously mmol/L NaP2O7, freshly supplemented with protease inhibitor into other SCID mice (the procedure is outlined in Fig. 1A). In tablet (Roche catalog # 1836153001) and 1mmol/L Na3VO4]. Tissues were homogenized on ice and tissue or cell lysates were some cases, reimplanted A431 tumors that were treated with run on a SDS-PAGE gel and Western blot analysis was conducted aflibercept continued to grow. This reimplantation and selection with the following antibodies overnight at 4C: Cell Signaling procedure was repeated once more, and we again observed tumor Technology: p-STAT3 (catalog # 9145), STAT3 (catalog # 9139), growth in the presence of aflibercept. In parallel, we performed a p44/42 MAPK (ERK1/2; catalog # 4348), phospho-p44/42 MAPK similar reimplantation procedure with A431 tumors treated with (ERK1/2; catalog # 4370), tubulin (catalog # 2125), Santa human control protein (Fc fragment only, Fig. 1A). After in vivo Cruz Biotechnology: NGAL (catalog # sc-50350), S100A7 (catalog passaging and selection, the aflibercept and control-treated # sc52948), SAA (catalog # sc-52214), IL36G (catalog # sc- tumors were harvested, minced, and transferred to tissue culture 80056), and b-actin (catalog # sc-69879), fibronectin (BD Trans- dishes for propagation in vitro. The cell lines derived from control duction, catalog # 610077). HRP-conjugated anti-rabbit and aflibercept-treated tumors were called A431-P (parental) and (Cell Signaling Technology, catalog # 7074), anti-mouse (Pierce, A431-V (variant), respectively (Fig. 1A). catalog # 31437), and anti-rat (Santa Cruz Biotechnology, catalog To confirm that the isolated A431-V cells retained resistance to # sc-2006) antibodies were applied for 1 hour at room temper- aflibercept, tumors from A431-V and A431-P cell lines were ature. Blots developed with ECL. Densitometry analysis was studied for growth kinetics and response to aflibercept. We performed using Carestream Software (Molecular Imaging). observed that control treated A431-V tumors grew slightly faster IL6 levels in conditioned medium of cultured cells, tumor than control-treated A431-P tumors (Fig. 1C). In addition, A431- lysates, and patient plasma were measured using Quantikine V tumors continued to grow during treatment with aflibercept, High Sensitivity IL-6 ELISA kit according to manufacturers' albeit at a slower rate (Fig. 1C), suggesting that A431-V tumors are instructions (R&D Systems, catalog # HS600B). VEGF levels of inherently resistant to VEGF blockade. These differences in tumor tumor lysates were measured using mouse and human VEGF growth were mirrored when assessing cell proliferation in tissue Quantikine ELISA Kit (R&D Systems, catalog # MMV00, DVE00) sections by detecting the proliferation marker Ki-67 (Fig. 1D). according to manufacturers' instructions. Control-treated A431-V tumors had higher numbers of proliferating cells in the tumor center than control treated A431-P tumors, Patient samples consistent with the faster growth rate of A431-V tumors (Fig. 1C fl Collection and testing of plasma for exploratory biomarker and D). A ibercept treatment dramatically reduced the number of analysis was specified in the amended clinical trial protocol for proliferating cells in A431-P tumors (Fig. 1D), with changes being fl ARD6122/AVE0005, a phase II, multicenter, randomized, dou- apparent as early as 24 hours after a ibercept administration fl ble-blind, parallel-arm, two-stage study of aflibercept in patients (Supplementary Fig. S2A). In contrast, a ibercept treatment with platinum-resistant and topotecan- and/or liposomal doxo- had a much smaller effect on cell proliferation in A431-V tumors fl rubicin-resistant advanced ovarian cancer. Informed consent for (Fig. 1D). Similarly, TUNEL staining revealed that a ibercept sample testing was obtained following protocol approval by local treatment increased apoptotic cells in A431-P tumors, but not Institutional Review Boards. in A431-V tumors (Fig. 1E). Interestingly, the growth rates of A431-P and A431-V cells in vitro did not differ significantly (Fig. 2A). In addition, A431-V Statistical analysis tumors had increased vessel density compared with A431-P Tumor sizes at the different time points from control and tumors, although vessel density after VEGF blockade was similar treatment groups were compared using two-way ANOVA and in A431-P and A431-V tumors (Supplementary Fig. S2B and S2C). Bonnferoni's multiple comparison tests (Prism software version 5). A431-P and A431-V tumor lysates did not show differences in murine (20.5 2 and 20.8 5 pg/mg, respectively) or human Results VEGF (13 1.9 and 12.1 4 ng/mg, respectively), suggesting that Generation and characterization of an aflibercept-resistant VEGF levels were not responsible for the difference in vessel A431 tumor cell line density. Human xenograft tumors grown in mice show different degrees We next wanted to determine whether the resistance of A431-V of response to VEGF blockade. To study resistance mechanisms, tumors was specific to anti-VEGF agents, or whether it was a more we aimed to derive tumor cell lines with acquired resistance to generalized resistance to targeted therapy. A431 cells express high

OF4 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

A B 7 6 ) 5 A431-P 5 A431-V 4 Day 0 P-ERK1/2 3 Day 3 2 ERK1/2

Cells per well (10 1 0 A431-P A431-V C D

1,200 1,200 ) A431-P - ctrl A431-P - ctrl 3

) 1,000 3 A431-P - aflib 1,000 A431-P - aflib A431-P - cetux A431-P - cetux 800 800 A431-V - ctrl A431-V - ctrl 600 A431-V - aflib 600 A431-V - aflib ** 400 A431-V - cetux 400 A431-V - cetux 200 Tumor volume(mm **** 200 ****

Tumor volume (mm **** 0

0 trl ib x rl ib x c fl tu ct u a e - afl 0 5 10 15 - c - cet P - - **** vs. A431-P - ctrl 1- P 1-V -V - 3 1- P 3 -V Days post treatment start 4 3 1- 4 A 3 A 431 31 A4 4 A **, **** vs. A431-V - ctrl A A4

Figure 2. Additional characterization of aflibercept-resistant A431 variants. A, in vitro growth kinetics of A431-P and A431-V cell lines as assessed by population doubling time (28.6 hours for A431-P and 27.9 hours for A431-V). Triplicates of each cell line were analyzed. B, total levels as well as phosphorylation levels ERK1/2 were assessed in A431-P and A431-V cells by Western blot analysis. C, tumor growth kinetics of A431-P and A431-V tumors (100–150 mm3) treated with human Fc control protein, aflibercept, or cetuximab for 14 days. The average tumor volume SD is plotted over the course of treatment. D, tumor growth changes of A431-P and A431-V tumors treated with control protein, aflibercept (aflib), or cetuximab (cetux) from start of treatment. , P < 0.01; , P < 0.0001 one-way ANOVA, Bonferroni post hoc test. levels of EGFR and are normally sensitive to EGFR antibodies. tumor cells was a marked increase in the phosphorylation of Variants of A431 that are resistant to EGFR blockers have been STAT3 at Tyr705 (Fig. 3A). The increase in phospho-STAT3 was described previously (15). EGFR activation was not altered confirmed by Luminex as well as Western blot analysis (Fig. 3B in A431-V cells, as no difference was observed in a phospho- and C, respectively). STAT3 can be phosphorylated in response to protein array (Fig. 3A). To determine EGFR activity, we assessed various secreted cytokines and chemokines, including IL6, IL10, phospho-ERK1/2 levels by Western blot analysis and observed and LIF (16). To determine whether the phosphorylation of that P-ERK1/2 in A431-V was somewhat lower compared with STAT3 in A431-V cells is mediated by secreted factors, we exposed A431-P (Fig. 2B), suggesting similar EGFR activity in both cell A431-P cells to CM from A431-V cells. We observed that phospho- lines. To determine whether A431-V tumors, which were selected STAT3 was increased, and was localized in the nucleus, in A431-P for resistance to aflibercept, maintained sensitivity to EGFR cells treated with CM from A431-V cells (Fig. 3D, third lane and E). blockade, mice bearing A431 variant or parental tumors were In contrast, phospho-STAT3 in A431-V cells was notably reduced treated with the EGFR antibody cetuximab. Both A431-V and -P after incubation with A431-P CM (Fig. 3D, fourth lane and E), tumors were similarly responsive to cetuximab, whereas, as suggesting that secreted factor(s) mediate the phosphorylation of expected the variant tumors were resistant to aflibercept (Fig. STAT3 in A431-V cells. 2C and D). Together, these findings indicate that the aflibercept To identify the factors secreted by A431-V cells that could resistance of A431-V tumors is a specific and stable trait, which mediate STAT3 activation, we compared the relative expression provides a model to study the underlying molecular mechanisms levels of secreted proteins in A431-P and A431-V cells using a of VEGF-resistant phenotype. quantitative proteomics technique called SILAC (stable isotope labeling with amino acids in cell culture; Supplementary Identification of pathways/factors mediating aflibercept Fig. S3A and S3B). We observed increased expression of 80 resistance in A431-V tumors proteins and decreased expression of 98 proteins in the secre- To explore the mechanisms of aflibercept resistance in A431-V tome of A431-V cells compared with A431-P cells (Supplemen- tumors, we first utilized a phospho-protein array to compare the tary Table S1). As many secreted signaling proteins, such as activities of 28 receptor tyrosine kinases and 11 downstream cytokines, are below the current detection limit of mass spec- signaling molecules in A431-V and A431-P cell lines (Fig. 3A). trometry, a multiplex Luminex assay was performed in parallel The most distinct difference between the resistant and sensitive to measure the concentrations of 42 cytokines in A431-P and

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF5

Eichten et al.

A B C A431-P Positive p-STAT3 (Tyr705) control 500

400 A431-P A431-V P-Stat3 A431-V 300

MFI Stat3 EGFR 200 (pan-Tyr) Stat3 (Tyr705) 100

0 A431-P A431-V

D E A431-P cells A431-V cells

A431-P A431-P A431-V CM A431-P A431-V

P-Stat3 A431-V Stat3 CM

F G

A431-P A431-V Protein fold Protein name change SAA IL8 135.8

CXCL-1 32.3 S100A7 GM-CSF 25.7 IL1α 24.5 FN1 IL6 23.8 IL36G

β-Actin

Figure 3. Constitutive STAT3 activation in A431-V cells. A, phospho-RTK signaling antibody array was used to examine phosphorylation status of 28 receptor tyrosine kinases and 11 downstream signaling molecules in A431-P and A431-V cell lysates. Each RTK or signaling node is spotted in duplicate. B, Tyr705 phosphorylation level of STAT3 in A431-P and A431-V cells assessed by Luminex. C, Tyr705 phosphorylation level of STAT3 in A431-P and A431-V cells assessed by Western blot analysis. D, A431-P and A431-V cells were exposed to CM from A431-P and A431-V cells as indicated by the red arrows and cell lysates were analyzed for phospho-STAT3 detection by Western blot analysis. E, A431-P and A431-V cells were exposed to CM from A431-P and A431-V cells as indicated and analyzed for phospho-STAT3 by IHC. Scale bar, 10 mm. F, top five cytokines with higher concentrations in A431-V conditioned media compared with A431-P CM, as measured by Luminex. G, increased expression of the IL6/STAT3-dependent protein SAA, S100A7, fibronectin, and IL36G in A431-V cell–conditioned media was confirmed by Western blot analysis.

A431-V CM. A total of 13 cytokines could be detected (Sup- expressing CXCL-1 and assessed the response to aflibercept in vivo. plementary Table S2), of which IL8, CXCL1, GM-CSF, IL1a,and CXCL-1 overexpression provided a slight growth advantage to IL6 were significantly increased in CM from A431-V cells A431 cells, but did not confer resistance to aflibercept (Supple- compared with that from A431-P cells (Fig. 3F). mentary Fig. S4A and S4B), suggesting that CXCL-1 is not a key Of these cytokines, CXCL-1 and IL6 are known to activate player in mediating aflibercept resistance in this model. STAT3 in various cell types (16–20). To assess the role of We next focused on IL6 and Ingenuity Pathway Analysis of the CXCL-1 in aflibercept resistance, we generated A431 cells over- 186 dysregulated proteins (SILAC plus Luminex) revealed the IL6

OF6 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

network as one of the most activated networks (in addition to IL8 agent caused a significant growth delay of Du145 tumors, suggest- and IL1a networks) in A431-V cells (Supplementary Fig. S5A and ing a role for autocrine IL6 signaling in Du145 tumor growth. S5B). In addition, increased protein expression of serum amyloid Aflibercept as a single agent had a more dramatic effect than A (SAA), S100 calcium-binding protein A7 (S100A7), fibronectin single-agent sarilumab, causing some tumor regression (Fig. 5B). (FN1), and IL36g (IL36G) was observed (Fig. 3G), all of which However, the combination of aflibercept plus sarilumab caused a have been reported to be downstream of the IL6/STAT3 pathway more rapid and pronounced tumor regression than aflibercept (21–24). Upregulation of SAA and NGAL mRNA was also alone (Fig. 5B), further supporting the possibility that IL6 signal- observed in A431-P cells exposed to A431-V CM (Supplementary ing limits the effectiveness of VEGF blockade in some tumors. Fig. S3C). On the basis of these data, we hypothesized that Consistent with active IL6/STAT3 signaling in Du145 tumors, secreted IL6 might be mediating phosphorylation of STAT3 in sarilumab decreased STAT3 phosphorylation as assessed by IHC A431-V cells. of tumor sections (Fig. 5C).

Blockade of IL6 signaling by anti-IL6R antibody partially Overexpression of IL6 promotes aflibercept resistance, whereas attenuates STAT3 activity in vitro and completely overcomes long-term blockade of both IL6 and VEGF decreases the aflibercept resistance in vivo emergence of aflibercept resistance To assess whether elevated STAT3 activation in A431-V cells To determine whether high IL6 expression is sufficient to was due to increased IL6 signaling, we determined whether A431- induce resistance to aflibercept, A431 cells were transduced with V cells and tumors have higher IL6 levels than their A431-P a lentiviral vector expressing human IL6 or an empty control counterparts. Indeed, IL6 protein levels in both A431-V cell CM vector. Ectopic IL6 expression resulted in accelerated tumor and A431-V tumor lysates were more than 3-fold higher than growth compared with control tumors (Supplementary those in A431-P counterparts (Fig. 4A). In contrast to IL6, levels of Fig. S6A). The high levels of ectopic IL6 expression also resulted IL6R were similar on A431-P and -V tumor cells (Fig. 4B). To in increased circulating IL6 protein level (100–200 ng/mL; data determine whether increased phosphorylation of STAT3 in A431- not shown), which was associated with body weight loss and early V tumors was due to the increased IL6, we used an IL6 receptor morbidity of tumor-bearing mice (Supplementary Fig. S6B). antibody (sarilumab) that blocks IL6 binding to IL6R (25). As Aflibercept treatment starting when tumors were approximately shown in Fig. 4C, STAT3 phosphorylation in cultured A431-V cells 100 mm3 had no effect on tumor growth in IL6-overexpressing was noticeably reduced by blocking IL6R with sarilumab, A431 tumors, while being very effective in A431 empty vector although not to quite the same level as seen in A431-P cells. control tumors (Fig. 6A and B). These data indicate that elevated To test the role of elevated IL6/STAT3 signaling in aflibercept IL6 expression in tumor cells is sufficient to induce resistance to resistance of A431-V tumors, we treated tumor-bearing mice with aflibercept in A431 xenograft tumors. sarilumab and/or aflibercept to assess the effects on A431-V and To investigate whether blockade of the IL6 pathway could A431-P tumor growth. Sarilumab binds and blocks human IL6R prevent the emergence of aflibercept-resistant A431 tumors, we but not murine IL6R, so its actions are specific to the A431 human treated A431 tumor bearing mice with control protein, aflibercept, tumor cells. Sarilumab had no effect on parental A431-P tumor sarilumab, or a combination of aflibercept plus sarilumab when growth and did not enhance the effects of aflibercept, indicating the tumors reached a volume of approximately 100 mm3. Single- that the growth of A431-P tumors is not dependent on agent sarilumab treatment did not delay A431 tumor growth, IL6 signaling, irrespective of whether VEGF activity is inhibited whereas both aflibercept single agent and aflibercept plus sarilu- (Fig. 4D, left). While single-agent sarilumab had no effect on the mab combination treatments resulted in prolonged growth stasis growth of A431-V tumors, the combination of sarilumab plus (Fig. 6C and D) for 6 weeks. As observed before, a subset of A431 aflibercept had a significantly greater effect than aflibercept alone, tumors (5/7 tumors; Fig. 6D top, Supplementary Fig. S1) began to completely inhibiting A431-V tumor growth (Fig. 4D, right). This grow significantly after 6 weeks of continuous aflibercept treat- finding indicates that IL6 signaling limits the effectiveness of ment. In contrast, no tumors showed steady growth during the aflibercept in A431-V tumors. The changes in growth kinetics of latter phases of continuous treatment with aflibercept plus sar- A431-V tumors upon aflibercept, sarilumab, or aflibercept plus ilumab (0/7 tumors; Fig. 6D, bottom). These results suggest that sarilumab were reflected in differences in proliferation, as assessed changes in the IL6 pathway are not only relevant after aflibercept by Ki-67 IHC (Fig. 4E). Consistent with a more prominent role for resistance occurred, but also play a role in the development of IL6/STAT3 signaling in A431-V tumors, Western blot analysis anti-VEGF resistance in A431 xenograft tumors. revealed that A431-V tumors have elevated levels of phospho- STAT3 compared to A431-P tumors, and that this STAT3 phos- IL6 serum levels may be predictive for outcome in ovarian phorylation can be blocked by sarilumab treatment (Fig. 4F). cancer patients treated with aflibercept These findings confirm that aflibercept resistance in A431-V cells is Taken together, our preclinical results show that activation of attributable to the increased STAT3 phosphorylation and IL6 the IL6/STAT3 pathway can provide evasive resistance to treat- signaling. ment with VEGF inhibitors such as aflibercept. To assess whether To extend our findings on the role of IL6/STAT3 signaling in this finding is clinically relevant, we examined the levels of IL6 in limiting aflibercept efficacy, we tested the effects of sarilumab as a ovarian cancer patients treated with single-agent aflibercept (26). single agent and in combination with aflibercept on the growth of Serum levels of IL6 were measured in a group of patients in an Du145 prostate tumor xenografts. Cultured Du145 cells exhibit international, double blind, phase II study of advanced ovarian constitutive STAT3 phosphorylation that can be inhibited by cancer, in which two different doses of aflibercept were used as blocking IL6R with sarilumab, indicating the presence of an monotherapy. Ninety-six patients were stratified into those with autocrine IL6/STAT3 signaling pathway (Fig. 5A) similar to that high and low circulating IL6 (cutoff 3.86 pg/mL). Although the in A431-V cells (Fig. 4C). As shown in Fig. 5B, sarilumab as a single study showed only a modest overall response rate to aflibercept,

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF7

Eichten et al.

A B C

1.0K

800 in CM 600 A431-P A431-V A431-P A431-V

400 hFc ++ + + Sarilumab + + + + 200

0 p-STAT3 100 101 102 103 104 105 106 FL4-A :: FL4-A t-STAT3 hIL6 (pg/mg) hIL6 (pg/mL) A431-P - unstained β-Actin A431-V - unstained p/t 2.62.51.01.05.23.81.31.0 A431-P - an-IL6R-APC STAT3 A431-V – an-IL6R-APC D

1,200 1,200 ) ) 3 3 900 900

600 600

300 300 vs. Fc vs. Fc vs. aﬂibercept vs. Fc Tumor volume (mm volume Tumor 0 vs. Fc (mm volume Tumor 0

E F A431-P A431-P A431-V A431-V Fc Sarilumab (Fc) (sarilumab) (Fc) (sarilumab) S S p-STAT3 T T t-STAT3

β-Acn

p-STAT3/ t-STAT3 (Densitomery) Aﬂibercept Combo 2.5 * S 2.0 S 1.5 T T 1.0 0.5 0.0 ) ) ) ) Fc b Fc b N ( a ( a -P m V m 1 ilu - ilu 3 r 31 r 4 sa 4 sa A ( A ( -P -V 1 1 3 3 4 4 A A

Figure 4. Effect of blocking cancer cell IL6 signaling pathway with sarilumab on the response of A431-P and A431-V tumors to aflibercept treatment. A, concentration of human IL6 was measured in CM from cultured A431-P or A431-V cells (left) and lysates from A431-P or A431-V tumor xenografts (right) by ELISA. , P < 0.01, unpaired t test. B, IL6R was detected on A431-P and A431-V cells by flow cytometry. C, A431-P and A431-V cells were treated with human Fc control or sarilumab in vitro and cell lysates were used for Western blot analysis detecting STAT3 and phospho-STAT3. Densitometry analysis was performed to calculate the ratio of phospho-STAT3 to total STAT3 (p/t STAT3). D, A431-P or A431-V tumor-bearing mice were treated with human Fc, sarilumab, aflibercept, or the combination of sarilumab plus aflibercept for 14 days. The average tumor volume SD is plotted over the course of treatment. , P < 0.05; , P < 0.01; , P < 0.0001, two-way ANOVA with Bonferroni post hoc test. E, representative images of cell proliferation assessed by Ki-67 IHC in A431-V tumors treated with control Fc, sarilumab, aflibercept, or the combination of sarilumab plus aflibercept for 14 days. Scale bar, 10 mm; S, stroma; T, tumor; N, necrosis. F, A431-P or A431-V tumor-bearing mice were treated with a single dose of human Fc or sarilumab and tumor lysates were prepared 18 hours after treatment. Phospho-STAT3 (p-STAT3) and total STAT3 (t-STAT3) levels were determined by Western blot analysis. Densitometry analysis was performed to calculate the ratio of phospho-STAT3 to total STAT3 (p/t STAT3). , P < 0.05, one-way ANOVA with Bonferroni post hoc test.

OF8 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

A B Fc Sarilumab Fc Sarilumab +IL6 +IL6 ) 3 p-STAT3

STAT3 Tumor volume (mm volume Tumor

C Fc Sarilumab

Figure 5. Blocking the cancer cell IL6 signaling pathway with sarilumab affects the response of Du145 tumors to aflibercept treatment. A, Du145 prostate cancer cells were cultured in complete medium and treated with 10 mg/mL of human Fc control protein or 10 mg/mL of sarilumab for 18 hours (lanes 1–4), and followed by a treatment of 10 ng/mL of hIL6 for 30 minutes (lanes 3 and 4). Cell lysates were harvested after treatment and Western blot analysis was performed to assess the levels of phospho-STAT3 and total-STAT3. B, Du145 tumor-bearing mice were treated with human Fc control protein, sarilumab, aflibercept, or the combination of sarilumab plus aflibercept (Combo). The average tumor volume SD is plotted over the course of treatment. , P < 0.05; , P < 0.01; , P < 0.0001 versus hFc-treated group, two-way ANOVA with Bonferroni post hoc test. C, representative images of phospho-STAT3 IHC on FFPE Du145 tumors treated with human Fc or sarilumab. Scale bar, 50 mm. we observed that patients with high IL6 serum levels (> median) the lack of species cross-activity of some factors such as IL6, thus show significantly poorer overall survival compared with those potentially limiting tumor–stromal interactions that could with low IL6 levels (< median; Fig. 7). These results suggest a induce STAT3 signaling in other tumor compartments (43). correlation between high levels of IL6 and poorer tumor response Nevertheless, based on these findings, we conclude that in to anti-VEGF therapy. A431-V tumors, aflibercept resistance is mediated to a large extent by increased IL6/STAT3 signaling in tumor cells, and that blockade of IL6R can overcome resistance to aflibercept. Discussion Various clinical studies have correlated high pretreatment To uncover mechanisms underlying resistance to anti-VEGF serum IL6 level with poor outcome in patients with different therapies, we used serial passaging and selection in vivo to types of cancer including head and neck squamous cell carcinoma generate A431-V that is partially resistant to aflibercept, and (28), stage II and III gastric carcinoma (29), prostate cancer (30), compared this line to the aflibercept-sensitive parental counter- metastatic renal cell carcinoma (31), metastatic breast cancer part A431-P. Our studies found that increased levels of proin- (32), and ovarian cancer (33). Elevated levels of IL6 in cancer flammatory mediators, such as IL6, were associated with STAT3 patients could be due to increased expression in normal immune pathway activation in A431-V tumors, and can contribute to cells as a consequence of chronic (systemic) inflammation, and/or aflibercept resistance. Blockade of the IL6/STAT3 pathway with from stromal or cancer cells within tumors. The resulting activa- an anti-IL6R antibody (sarilumab), which is specifictohuman tion of the STAT3 transcription factor in cancer cells is essential for IL6R, rendered A431-V tumors more sensitive to aflibercept. IL6-dependent oncogenic activities such as promoting cancer cell Similarly, IL6R inhibition enhanced the activity of aflibercept in proliferation and survival (34). STAT3 activation has been shown another tumor model with strong endogenous IL6/STAT3 activ- to support cancer cell survival in the conditions of growth factor ity, namely Du145 prostate tumors. In addition, IL6 overexpres- deprivation (35) and to promote chemoresistance in cancer cells sion rendered A431 tumors resistant to aflibercept, suggesting exposed to hypoxia (36), suggesting that STAT3 may be a poten- that elevated IL6 expression is sufficient to provide a survival tially important cancer cell survival factor in tumors treated with advantage to tumor cells during aflibercept treatment. In these antiangiogenic therapies. Our findings that patients with high IL6 experiments, activation of IL6R/STAT3 signaling in cancer cells is serum levels showed poorer survival than those with low IL6 mediated by human IL6 because mouse IL6 does not activate the levels in a phase II study of advanced ovarian cancer where human IL6R (27), and murine IL6R is not inhibited by sarilu- aflibercept was used as monotherapy (26), suggests that our mab, a human-specific IL6R antibody. Limitations of working preclinical findings could have clinical significance and warrant with human tumor cells in immunocompromised mice include further investigation.

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF9

Eichten et al.

A B ns ****

500 400 ) ) 3 3 400 A431 IL6 ctrl 300 A431 IL6 afl 300 A431 empty vector ctrl 200

A431 empty vector afl 100 200 0

100 Tumor growth (mm -100 Tumor volume (mm fl fl 0 a trl c r a 6 r o 0 2 4 6 8 to IL6 ctrl 1 IL c ct 3 31 4 ve Days post treatment start 4 A ty ty ve A p p em 1 3 A431 em A4 C D

) 500 3 Aflibercept * 400 * 300 5/7 tumors escaped 200 * 1,000 * A431 Fc (25 mg/kg) 100 * ) 3 A431 sarilumab (25 mg/kg) Tumor volume (mm 0 800 0204060 A431 aflibercept (25 mg/kg) Days post treatment start 600 A431 combo (25 mg/kg/25 mg/kg)

) 500

3 Combo 400 400 0/7 tumors escaped 300 # 200 200 Tumor volume (mm 0 100 * 0 20 40 60

Tumor volume (mm 0 Days post treatment start 0 20 40 60 Days post treatment start

Figure 6. Overexpression of IL6 or long-term blockade of the IL6 pathway affects the response of A431 tumor xenografts to aflibercept treatment. A, A431 tumor cells were engineered to overexpress IL6 and tumor growth kinetics and aflibercept response was compared with A431 empty vector tumors for 7 days. The average tumor volume SD is plotted over the course of treatment. B, tumor growth changes of IL6 or empty vector expressing tumors in response to human Fc control protein or aflibercept from start of treatment. , P < 0.0001 one-way ANOVA, Bonferroni post hoc test. C, SCID mice bearing A431 tumors (100–150 mm3) were treated with human Fc control protein, sarilumab, aflibercept, or a combination of aflibercept þ sarilumab for up to 7 weeks. D, individual tumor growth kinetics for the aflibercept and aflibercept þ sarilumab-treated groups (Combo). The tumors that exhibit a "late escape," defined as a tumor showing 40% or more increase in tumor volume between day 42 and day 49, are indicated by a red star next to the tumor growth curve. One combination-treated tumor, as indicated by a black # symbol, showed intermittent growth throughout the experiment, but did not show escape as defined above. One combination-treated tumor grew at the end of the experiment, but only showed an increase in tumor volume of 33% between day 42 and 49, indicated by a blue star next to the tumor growth curve

Resistance to anti-VEGF therapies can occur via several mechan- giogenic factors, including IL8 and CXCL1 in A431-V tumors, isms: (i) upregulation of various proangiogenic factors resulting which could contribute to the increased vessel density. A key role in vessels that are less sensitive to VEGF blockade and/or (ii) for CXCL1 in mediating aflibercept resistance was ruled out in our changes in the characteristics of cancer cells, which provide the model system. Given the STAT3 activation we observed in A431-V cells with a survival advantage and/or increased invasive potential tumors, we decided to focus on IL6, rather than IL8 in subsequent in the environment of reduced tumor vessels (12–14). Our studies. A large body of literature suggests that IL6 activates STAT3 approach to induce variants of A431 tumor cells that are resistant (37, 38) and, in addition, our studies implicate a key role for the to aflibercept allowed direct comparisons between parental and IL6/STAT3 pathway in resistance to anti-VEGF therapies. variant tumors. For example, screening of phospho-protein arrays Notably, our studies on the role of IL6 in resistance to anti- found that phospho-STAT3 was the major difference between VEGF therapies focus on autocrine IL6/STAT3 signaling in the parental and variant A431 tumor cells. We also used unbiased cancer cells. Autocrine IL6 signaling has been reported in various SILAC screens to test for overall differences in the secreted protein cancer types, including multiple myeloma (39), prostate cancer profile (secretome) of parental and variant cells. Detailed expres- (40), lung cancer (41), and breast cancer (42). In our studies on sion analysis revealed upregulation of various potentially proan- A431-V and Du145 xenograft tumors, activation of IL6R/STAT3

OF10 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

1.00 Low IL6 High IL6

0.75

0.50

0.25

Overall surviving fraction HR (95% CI): 0.243 (0.156–0.379) 0.00 0 100 200 300 400 500 600 Days from rand to death STRATA: trt_median - Above Censored trt_median - Above trt_median - Below Censored trt_median - Below

Biomarker subset population Even/total Kaplan–Meier estimate 95% CI P High IL6 (>3.86 pg/mL) 66/94 198.000 (158–239) <.0001 Low IL6 (>3.86 pg/mL) 30/93 593.000 (499–593)

Figure 7. Aflibercept treatment is more effective in increasing overall survival in ovarian cancer patients with low IL6 levels. In an international, double blind, phase II study of advanced ovarian cancer, aflibercept was used as a monotherapy. Patients received either a 2 or 4 mg/kg of aflibercept every 2 weeks and were monitored for overall response. The study showed a modest response rate of 5% in both arms, and the ovarian indication has not been further vigorously pursued, in part due to this data. When the data was stratified on the basis of low (< median) and high (> median) IL6 serum levels, it became apparent that patients with high IL6 levels show a worse response compared with those with low IL6 levels.

signaling in cancer cells is solely mediated by IL6 derived from the tance and thus might be a valid therapeutic target when combined human cancer cells, but not the murine stromal cells, because with anti-VEGF therapies in cancer patients. Although the data mouse IL6 does not activate the human IL6R (27). Although presented in this article do not provide details on how IL6/STAT3 human IL6 activates mouse IL-6R on stromal cells, this interaction activation in the tumor cells can confer resistance to aflibercept is not inhibited by sarilumab, which is a human-specific IL6R treatment, we hypothesize that a double hit of targeting the antibody. Therefore, the effects of sarilumab reported here are due vasculature, thus diminishing oxygen and nutrient supply for the to the interruption of autocrine IL6/STAT3 signaling in cancer tumor cells, and blocking IL6, thereby shutting down the IL6/ cells. Although other studies have reported that increased activa- STAT3–regulated survival and proliferation pathway in the tumor tion of tumor or stromal STAT3 can lead to enhanced STAT3 cells and limiting IL6-mediated inflammation in the tumor signaling in tumor endothelial cells (43), such a mechanism is microenvironment, can overcome anti-VEGF resistance in tumors apparently not at play in the A431-V model. Instead, our data with measurable IL6/STAT3 signaling. Further studies and clinical highlight a mechanism of resistance to anti-VEGF therapy, in trials are necessary to validate or refute the role of IL6 in anti-VEGF which the cancer cell characteristics change and provide a survival resistance in different types of cancer. advantage in the environment of reduced tumor vessels. Although IL6 has previously been implicated as a potential fl biomarker for targeted anti-VEGF therapy, the results have been Disclosure of Potential Con icts of Interest inconsistent. When treated with bevacizumab plus a chemother- E. Ioffe is an employee and stock holder at Regeneron Pharmaceuticals. I. Lowy is a vice president and has ownership interest (including patents) in apeutic regimen, a better outcome was reported for metastatic Regeneron. H.C. Lin is a senior director, molecular profiling, and reports colorectal cancer (mCRC) patients with a signature that included receiving a commercial research grant and other commercial research sup- lower than average baseline IL6 serum levels (44), whereas in port from Regeneron Pharmaceuticals Inc. No potential conflicts of interest another study of mCRC, a worse outcome was associated with a were disclosed by the other authors. signature that included lower than average IL6 serum levels (45). High baseline IL6 levels were also reported to predict shorter Authors' Contributions progression-free survival (PFS) and overall survival (OS) in Conception and design: A. Eichten, J. Su, A. Adler, L. Zhang, E. Ioffe, patients with advanced hepatocellular carcinomas (46). In line G.D. Yancopoulos, D. MacDonald, G. Thurston with these later studies, we observed in an ovarian cancer trial that Development of methodology: A. Adler, E. Ioffe, A.A. Parveen, X. Duan, aflibercept-treated patients with high IL6 had poorer survival than G. Thurston patients with low IL6 serum levels. However, caution should be Acquisition of data (provided animals, acquired and managed patients, taken when using serum IL6 levels as a biomarker, as our studies provided facilities, etc.): J. Su, A. Adler, X. Duan suggest that IL6R signaling on tumor cells is the relevant feature Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Eichten, J. Su, A. Adler, L. Zhang, I. Lowy, C. Lin, that confers resistance to anti-VEGF therapy. C. Daly, G. Thurston IL6 could be a potential biomarker for anti-VEGF therapies, but Writing, review, and/or revision of the manuscript: A. Eichten, J. Su, A. Adler, our data now suggest that IL6 can also mediate anti-VEGF resis- L. Zhang, I. Lowy, C. Lin, D. MacDonald, X. Duan, G. Thurston

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF11

Eichten et al.

Administrative, technical, or material support (i.e., reporting or organizing critical comments and suggestions, and Alshad Lalani and Bo Gao for assistance data, constructing databases): J.S. Rudge with processing clinical samples and data analysis. Study supervision: A. Eichten, L. Zhang, D. MacDonald, G. Thurston The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Acknowledgments The authors thank Bo Luan, Thomas Nevins, Benjamin Strober, Bradley Received May 28, 2015; revised December 16, 2015; accepted December 29, Hagan, and Jennifer Espert for technical assistance, Cristina Abrahams for 2015; published OnlineFirst February 26, 2016.

References 1. Ferrara N, Kerbel RS. Angiogenesis as a therapeutic target. Nature 2005; 17. Song L, Turkson J, Karras JG, Jove R, Haura EB. Activation of Stat3 by 438:967–74. receptor tyrosine kinases and cytokines regulates survival in human non- 2. Papadopoulos N, Martin J, Ruan Q, Rafique A, Rosconi MP, Shi E, et al. small cell carcinoma cells. Oncogene 2003;22:4150–65. Binding and neutralization of vascular endothelial growth factor (VEGF) 18. Naugler WE, Sakurai T, Kim S, Maeda S, Kim K, Elsharkawy AM, et al. and related ligands by VEGF Trap, ranibizumab and bevacizumab. Gender disparity in liver cancer due to sex differences in MyD88-dependent Angiogenesis 2012;15:171–85. IL-6 production. Science 2007;317:121–4. 3. Folkman J. Fundamental concepts of the angiogenic process. Curr Mol Med 19. Grivennikov S, Karin E, Terzic J, Mucida D, Yu GY, Vallabhapurapu S, et al. 2003;3:643–51. IL-6 and Stat3 are required for survival of intestinal epithelial cells and 4. Holash J, Davis S, Papadopoulos N, Croll SD, Ho L, Russell M, et al. VEGF- development of colitis-associated cancer. Cancer Cell 2009;15:103–13. Trap: a VEGF blocker with potent antitumor effects. Proc Natl Acad Sci U S A 20. Wei ZW, Xia GK, Wu Y, Chen W, Xiang Z, Schwarz RE, et al. CXCL1 2002;99:11393–8. promotes tumor growth through VEGF pathway activation and is 5. Rudge JS, Thurston G, Davis S, Papadopoulos N, Gale N, Wiegand SJ, et al. associated with inferior survival in gastric cancer. Cancer Lett 2015;359: VEGF trap as a novel antiangiogenic treatment currently in clinical trials for 335–43. cancer and eye diseases, and VelociGene- based discovery of the next 21. Hagihara K, Nishikawa T, Isobe T, Song J, Sugamata Y, Yoshizaki K. IL-6 generation of angiogenesis targets. Cold Spring Harb Symp Quant Biol plays a critical role in the synergistic induction of human serum amyloid A 2005;70:411–8. (SAA) gene when stimulated with proinflammatory cytokines as analyzed 6. Rudge JS, Holash J, Hylton D, Russell M, Jiang S, Leidich R, et al. VEGF Trap with an SAA isoform real-time quantitative RT-PCR assay system. Biochem complex formation measures production rates of VEGF, providing a Biophys Res Commun 2004;314:363–9. biomarker for predicting efficacious angiogenic blockade. Proc Natl Acad 22. Hagihara K, Nishikawa T, Sugamata Y, Song J, Isobe T, Taga T, et al. Essential Sci U S A 2007;104:18363–70. role of STAT3 in cytokine-driven NF-kappaB-mediated serum amyloid A 7. Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, Hainsworth J, Heim gene expression. Genes Cells 2005;10:1051–63. W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for 23. West NR, Watson PH. S100A7 (psoriasin) is induced by the proinflam- metastatic colorectal cancer. N Engl J Med 2004;350:2335–42. matory cytokines oncostatin-M and interleukin-6 in human breast cancer. 8. Giantonio BJ, Catalano PJ, Meropol NJ, O'Dwyer PJ, Mitchell EP, Alberts Oncogene 2010;29:2083–92. SR, et al. Bevacizumab in combination with oxaliplatin, fluorouracil, and 24. Sullivan NJ, Sasser AK, Axel AE, Vesuna F, Raman V, Ramirez N, et al. leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: Interleukin-6 induces an epithelial-mesenchymal transition phenotype in results from the Eastern Cooperative Oncology Group Study E3200. J Clin human breast cancer cells. Oncogene 2009;28:2940–7. Oncol 2007;25:1539–44. 25. Huizinga TW, Fleischmann RM, Jasson M, Radin AR, van Adelsberg J, Fiore 9. Van Cutsem E, Tabernero J, Lakomy R, Prenen H, Prausova J, Macarulla T, S, et al. Sarilumab, a fully human monoclonal antibody against IL-6Ralpha et al. Addition of aflibercept to fluorouracil, leucovorin, and irinotecan in patients with rheumatoid arthritis and an inadequate response to improves survival in a phase III randomized trial in patients with metastatic methotrexate: efficacy and safety results from the randomised SARIL-RA- colorectal cancer previously treated with an oxaliplatin-based regimen. MOBILITY Part A trial. Ann Rheum Dis 2014;73:1626–34. J Clin Oncol 2012;30:3499–506. 26. Tew WP, Colombo N, Ray-Coquard I, Del Campo JM, Oza A, Pereira D, 10. Wilke H, Muro K, Van Cutsem E, Oh SC, Bodoky G, Shimada Y, et al. et al. Intravenous aflibercept in patients with platinum-resistant, advanced Ramucirumab plus paclitaxel versus placebo plus paclitaxel in patients ovarian cancer: results of a randomized, double-blind, phase 2, parallel- with previously treated advanced gastric or gastro-oesophageal junction arm study. Cancer 2014;120:335–43. adenocarcinoma (RAINBOW): a double-blind, randomised phase 3 trial. 27. Chiu CP, Moulds C, Coffman RL, Rennick D, Lee F. Multiple biological Lancet Oncol 2014;15:1224–35. activities are expressed by a mouse interleukin 6 cDNA clone isolated from 11. Garon EB, Ciuleanu TE, Arrieta O, Prabhash K, Syrigos KN, Goksel T, bone marrow stromal cells. Proc Natl Acad Sci U S A 1988;85:7099–103. et al. Ramucirumab plus docetaxel versus placebo plus docetaxel 28. Duffy SA, Taylor JM, Terrell JE, Islam M, Li Y, Fowler KE, et al. Interleukin-6 for second-line treatment of stage IV non-small-cell lung cancer after predicts recurrence and survival among head and neck cancer patients. disease progression on platinum-based therapy (REVEL): a multi- Cancer 2008;113:750–7. centre, double-blind, randomised phase 3 trial. Lancet 2014;384: 29. Liao WC, Lin JT, Wu CY, Huang SP, Lin MT, Wu AS, et al. Serum interleukin- 665–73. 6 level but not genotype predicts survival after resection in stages II and III 12. Bergers G, Hanahan D. Modes of resistance to anti-angiogenic therapy. gastric carcinoma. Clin Cancer Res 2008;14:428–34. Nat Rev Cancer 2008;8:592–603. 30. Michalaki V, Syrigos K, Charles P, Waxman J. Serum levels of IL-6 and TNF- 13. Loges S, Schmidt T, Carmeliet P. Mechanisms of resistance to anti-angio- alpha correlate with clinicopathological features and patient survival in genic therapy and development of third-generation anti-angiogenic drug patients with prostate cancer. Br J Cancer 2004;90:2312–6. candidates. Genes Cancer 2010;1:12–25. 31. Negrier S, Perol D, Menetrier-Caux C, Escudier B, Pallardy M, Ravaud A, 14. Smith NR, Baker D, Farren M, Pommier A, Swann R, Wang X, et al. Tumor et al. Interleukin-6, interleukin-10, and vascular endothelial growth factor stromal architecture can define the intrinsic tumor response to VEGF- in metastatic renal cell carcinoma: prognostic value of interleukin-6–from targeted therapy. Clin Cancer Res 2013;19:6943–56. the Groupe Francais d'Immunotherapie. J Clin Oncol 2004;22:2371–8. 15. Viloria-Petit A, Crombet T, Jothy S, Hicklin D, Bohlen P, Schlaeppi JM, et al. 32. Zhang GJ, Adachi I. Serum interleukin-6 levels correlate to tumor progres- Acquired resistance to the antitumor effect of epidermal growth factor sion and prognosis in metastatic breast carcinoma. Anticancer Res receptor-blocking antibodies in vivo: a role for altered tumor angiogenesis. 1999;19:1427–32. Cancer Res 2001;61:5090–101. 33. Scambia G, Testa U, Benedetti Panici P, Foti E, Martucci R, Gadducci A, et al. 16. Yu H, Pardoll D, Jove R. STATs in cancer inflammation and immunity: a Prognostic significance of interleukin 6 serum levels in patients with leading role for STAT3. Nat Rev Cancer 2009;9:798–809. ovarian cancer. Br J Cancer 1995;71:354–6.

OF12 Cancer Res; 76(8) April 15, 2016 Cancer Research

IL6/STAT3 Signaling Mediates Anti-VEGF Resistance

34. Bollrath J, Phesse TJ, von Burstin VA, Putoczki T, Bennecke M, Bateman 41. Gao SP, Mark KG, Leslie K, Pao W, Motoi N, Gerald WL, et al. Mutations in T, et al. gp130-mediated Stat3 activation in enterocytes regulates the EGFR kinase domain mediate STAT3 activation via IL-6 production in cell survival and cell-cycle progression during colitis-associated tumor- human lung adenocarcinomas. J Clin Invest 2007;117:3846–56. igenesis. Cancer Cell 2009;15:91–102. 42. Hartman ZC, Poage GM, den Hollander P, Tsimelzon A, Hill J, Panupinthu 35. Yin W, Cheepala S, Roberts JN, Syson-Chan K, DiGiovanni J, Clifford JL. N, et al. Growth of triple-negative breast cancer cells relies upon coordinate Active Stat3 is required for survival of human squamous cell carcinoma autocrine expression of the proinflammatory cytokines IL-6 and IL-8. cells in serum-free conditions. Mol Cancer 2006;5:15. Cancer Res 2013;73:3470–80. 36. Selvendiran K, Bratasz A, Kuppusamy ML, Tazi MF, Rivera BK, Kuppusamy 43. Kujawski M, Kortylewski M, Lee H, Herrmann A, Kay H, Yu H. Stat3 P. Hypoxia induces chemoresistance in ovarian cancer cells by activation of mediates myeloid cell-dependent tumor angiogenesis in mice. J Clin Invest signal transducer and activator of transcription 3. Int J Cancer 2009;125: 2008;118:3367–77. 2198–204. 44. Abajo A, Boni V, Lopez I, Gonzalez-Huarriz M, Bitarte N, Rodriguez J, et al. 37. Hong DS, Angelo LS, Kurzrock R. Interleukin-6 and its receptor in Identification of predictive circulating biomarkers of bevacizumab-con- cancer: implications for translational therapeutics. Cancer 2007;110: taining regimen efficacy in pre-treated metastatic colorectal cancer patients. 1911–28. Br J Cancer 2012;107:287–90. 38. Sansone P, Bromberg J. Targeting the interleukin-6/Jak/stat pathway in 45. Liu Y, Starr MD, Bulusu A, Pang H, Wong NS, Honeycutt W, et al. human malignancies. J Clin Oncol 2012;30:1005–14. Correlation of angiogenic biomarker signatures with clinical outcomes in 39. Kawano M, Hirano T, Matsuda T, Taga T, Horii Y, Iwato K, et al. Autocrine metastatic colorectal cancer patients receiving capecitabine, oxaliplatin, generation and requirement of BSF-2/IL-6 for human multiple myelomas. and bevacizumab. Cancer Med 2013;2:234–42. Nature 1988;332:83–5. 46. Boige V, Malka D,Bourredjem A,DromainC,Baey C,Jacques N,et al. Efficacy, 40. Giri D, Ozen M, Ittmann M. Interleukin-6 is an autocrine growth factor in safety, and biomarkers of single-agent bevacizumab therapy in patients with human prostate cancer. Am J Pathol 2001;159:2159–65. advanced hepatocellular carcinoma. Oncologist 2012;17:1063–72.

www.aacrjournals.org Cancer Res; 76(8) April 15, 2016 OF13

Resistance to Anti-VEGF Therapy Mediated by Autocrine IL6/STAT3 Signaling and Overcome by IL6 Blockade

Alexandra Eichten, Jia Su, Alexander P. Adler, et al.

Cancer Res Published OnlineFirst February 26, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-1443

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/02/26/0008-5472.CAN-15-1443.DC1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2016/04/05/0008-5472.CAN-15-1443. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.