A Critical Role for GRP78/BiP in the Tumor ... · and maintenance of the tumor (12, 13)....

Microenvironment and Immunology

A Critical Role for GRP78/BiP in the TumorMicroenvironment for Neovascularization during TumorGrowth and Metastasis

Dezheng Dong1, Christopher Stapleton2,5, Biquan Luo1, Shigang Xiong3, Wei Ye4, Yi Zhang1,Niyati Jhaveri2, Genyuan Zhu1, Risheng Ye1, Zhi Liu2, Kevin W. Bruhn6, Noah Craft6,Susan Groshen4, Florence M. Hofman2, and Amy S. Lee1

AbstractGlucose-regulated protein 78 (GRP78)/BiP is amultifunctional proteinwhich plays amajor role in endoplasmic

reticulum (ER) protein processing, protein quality control, maintaining ER homeostasis, and controlling cellsignaling and viability. Previously, using a transgene-induced mammary tumor model, we showed that Grp78heterozygosity impeded cancer growth through suppression of tumor cell proliferation and promotion ofapoptosis and the Grp78þ/� mice exhibited dramatic reduction (70%) in the microvessel density (MVD) ofthe endogenous mammary tumors, while having no effect on the MVD of normal organs. This observationsuggests that GRP78 may critically regulate the function of the host vasculature within the tumor microenviron-ment. In this article, we interrogated the role of GRP78 in the tumormicroenvironment. Inmouse tumormodels inwhich wild-type (WT), syngeneic mammary tumor cells were injected into the host, we showed that Grp78þ/�

mice suppressed tumor growth and angiogenesis during the early phase but not during the late phase of tumorgrowth. Growth of metastatic lesions of WT, syngeneic melanoma cells in the Grp78þ/� mice was potentlysuppressed. We created conditional heterozygous knockout of GRP78 in the host endothelial cells and showedsevere reduction of tumor angiogenesis and metastatic growth, with minimal effect on normal tissue MVD.Furthermore, knockdown of GRP78 expression in immortalized human endothelial cells showed that GRP78 is acriticalmediator of angiogenesis by regulating cell proliferation, survival, andmigration. Our findings suggest thatconcomitant use of current chemotherapeutic agents and novel therapies against GRP78 may offer a powerfuldual approach to arrest cancer initiation, progression, and metastasis. Cancer Res; 71(8); 2848–57. �2011 AACR.

Introduction

Glucose-regulated protein 78 (GRP78), also referred to asBiP or HSPA5, is a member of the HSP70 protein family with

established function as an endoplasmic reticulum (ER) cha-perone protein and a regulator of the ER stress signalingpathway (1). As a multifunctional protein, GRP78 plays criticalroles in physiologic and pathologic stress (2). GRP78 isinduced in a wide variety of tumors through intrinsic factorssuch as altered glucose metabolism in cancer cells, com-pounded by extrinsic factors such as glucose starvation andhypoxia in the microenvironment of poorly perfused solidtumors (3). The induction of GRP78 leads to an increase ofGRP78 in the ER compartment and promotion of GRP78 to thecell surface, where it assumes a new function of coreceptor forcell-surface signaling (4–7). Although a large number of stu-dies have established that GRP78 is required for tumor cellproliferation, survival, and therapeutic resistance (3, 8, 9), littleis known about the role of GRP78 in the tumor microenvir-onment, which is critical for support of tumor growth andmetastasis.

Previously, we established a heterozygous knockdownmouse model of GRP78 as homozygous deletion of Grp78leads to early embryonic lethality (10). The heterozygousGrp78 (Grp78þ/�) mice, despite reduced expression ofGRP78 to about half the wild-type (WT) level in all the tissues,are phenotypically normal with respect to organ developmentand antibody production (11). However, through breeding of

Authors' Affiliations: Departments of 1Biochemistry and Molecular Biol-ogy, 2Pathology, 3Medicine, 4Preventive Medicine, USC Norris Compre-hensive Cancer Center, University of Southern California Keck School ofMedicine, Los Angeles, California; 5Harvard-MIT, Division of HealthSciences and Technology, Harvard Medical School, Boston, Massachu-setts; and 6Divisions of Dermatology and Infectious Diseases, Los AngelesBiomedical Research Institute at Harbor-UCLA Medical Center, Torrance,California

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

C. Stapleton, B. Luo contributed equally to this work.

Corresponding Author: Amy S. Lee, Department of Biochemistry andMolecular Biology, USC Norris Comprehensive Cancer Center, Universityof Southern California, 1441 Eastlake Ave., Los Angeles, CA 90089. Phone:323-865-0507; Fax: 323-865-0094; E-mail: [email protected] orFlorence M. Hofman, Department of Pathology, University of SouthernCalifornia, 2011Zonal Ave., LosAngeles,CA90089. Phone: 323-442-1153;Fax: 323-442-3049; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-10-3151

�2011 American Association for Cancer Research.

CancerResearch

Cancer Res; 71(8) April 15, 20112848

Research. on January 24, 2021. © 2011 American Association for Cancercancerres.aacrjournals.org Downloaded from

Published OnlineFirst April 5, 2011; DOI: 10.1158/0008-5472.CAN-10-3151

http://cancerres.aacrjournals.org/

the WT and Grp78þ/� mice with the transgenic mice expres-sing the middle T oncogene driven by the murine mammarytumor viral promoter, we discovered that Grp78 heterozyg-osity prolonged the latency period and significantly impededcancer growth by suppressing tumor cell proliferation andpromoting tumor cell apoptosis (11). Strikingly, the micro-vessel density (MVD) of the endogenous tumors in theGrp78þ/� mice was reduced by 70%, while the vasculatureof normal organs remained unaffected. This observationprovided the initial evidence that GRP78 may selectivelyregulate the function and survival of endothelial cells asso-ciated with the tumor.The tumor microenvironment contains a plethora of cells

that support tumor growth and progression. Tumor vascula-ture is essential for tumor growth and metastasis, as itsupplies the nutrients and oxygen crucial for the growthand maintenance of the tumor (12, 13). Endothelial cells aretherefore the requisite members of the tumor microenviron-ment. Previous studies on human glioma-derived endothelialcells have shown that these tumor-associated endothelial cellspossess unique functional characteristics, including an acti-vated phenotype (14). Tumors require an activated endothelialpopulation to provide a constant newblood vessel formation atthe growing border of the tumor, which allows for continuedexpansion and progression. Our recent investigations have alsoindicated that GRP78 expression is constitutively elevatedwithin the tumor vasculature, emphasizing the constant stateof activation of these tumor-associated endothelial cells (15).In this report, we tested the role of GRP78 in host cells

supporting tumor growth by using syngeneic cancer models.Our studies revealed that neovascularization during earlytumor growth and tumor metastasis was potently suppressedby host Grp78 heterozygosity. We further created an endothe-lial cell–specific Grp78 heterozygous knockout mouse model(Grp78F/þ;Tie2-Cre) and showed severe reduction of tumorangiogenesis and growth of metastatic lesions, with minimaleffect on normal tissue MVD. Furthermore, knockdown ofGRP78 in immortalized human endothelial cells results insuppression of endothelial cell proliferation, promotion ofapoptosis, and decreased migration. These findings revealGRP78 is a regulator of endothelial cell–specific angiogenicfunctions and provides a mechanistic explanation for therequirement of GRP78 in the microenvironment during tumorgrowth and metastasis.

Materials and Methods

Cell culturesThe E0771 mouse breast cancer cell line was generously

provided by Dr. Enrico Mihich (Roswell Park Cancer Institute)and was maintained in RPMI 1640, supplemented with 10%FBS, 10 mmol/L HEPES, and 1% penicillin antibiotics asdescribed (16). The immortalized human microvascularendothelial cells (HMEC) were kindly provided by Dr. BettyWu-Hsieh (Taiwan National University). HMECs were grownin Medium 200 (Invitrogen) supplemented with Low SerumGrowth Supplement (Invitrogen). All cells were maintained at37�C and 5% CO2.

Generation of the B16-Fluc-A1 melanoma cell lineB16-F0 cells (American Type Culture Collection) were

transduced with a lentiviral construct containing the fireflyluciferase gene driven by an internal CMV promoter. Clonalcell lines were selected by limiting dilutions in 96-well platesand expanded and screened for high levels of luciferaseexpression, using a Promega luciferase detection kit, perthe manufacturer's instructions. The cells were maintainedin high-glucose Dulbecco's modified Eagle's medium contain-ing 4.5 mg/mL glucose, supplemented with 10% FBS, 2 mmol/L glutamine, and 1% penicillin antibiotics.

Generation of genetically altered Grp78 mouse modelThe Grp78þ/� mice were generated as described (10) and

were backcrossed into the C57BL/6 genetic background for 8generations. To create endothelial cell–specific Grp78 het-erozygous knockout mice, mice carrying the Grp78 floxedallele (in C57BL6 and 129/Sv background; ref. 17) werecrossed with Tie2-Cre transgenic mice (Tek-Cre in C57BL6background; the Jackson Laboratory; ref. 18). Genotyping forthe WT, floxed, and KO alleles were carried out by PCR usinggenomic DNA extracted from mouse tail biopsies asdescribed (17). Genotyping was also carried out using geno-mic DNA extracted from enriched primary brain endothelialcells as previously described (19), with modifications(20). The Tie2-Cre transgene was identified with forwardprimer: 50-AAGAACCTGATGGACATGTTCAGGGA-30 andreverse primer: 50-ACGAACCTGGTCGAAATCAGTGCG-TTC-30. Three-month-old mice were used for the tumormodel studies. All animal protocols were conducted withthe approval of the USC University Animal Care and UseCommittee.

Generation of tumor modelsThe generation and monitoring of endogenous mammary

tumors driven by the MMTV-PyVT transgene in Grp78þ/þ orGrp78þ/� mice have been described (11). In the syngeneictumor models, E0771 cells (4� 106/mice) were resuspended in200 mL PBS and injected into the mammary fat pad of 4-week-old female mice; B16-Fluc-A1 melanoma cells (amount asindicated) were resuspended in 200 mL PBS and injectedthrough the lateral tail vein. Lung metastasis was monitoredby luminescence imaging (Xenogen) weekly. Following sacri-fice, the lungs were removed and the number of surfacemetastases was counted.

Immunofluorescent and immunohistochemicalstaining

Immunostaining was done on paraffin-embedded tumorsections as previously described (11). The primary antibodiesused were the following: mouse anti-PCNA (1:100), goat anti-VEGF (1:100) from Santa Cruz Biotechnology; and rabbit anti-CD31 (1:50) from Thermo Fisher Scientific. The processedsections were visualized using a fluorescence microscope.

Microvessel density measurementMouse tumor tissues were processed for MVD analysis as

previously described (11). The tissues were stained with rat

Role of GRP78 in the Tumor Microenvironment

www.aacrjournals.org Cancer Res; 71(8) April 15, 2011 2849




anti-mouse CD31 antibody (BD Pharmingen) and quantifiedusing the imaging processing program ImageJ (NIH).

siRNA transfectionHMECs were seeded at a density of 6� 104 per well in 6-well

plates and transfected with siRNA using Lipofectamine 2000Transfection Reagent (Invitrogen) per the manufacturer'sinstructions. The siRNA against Grp78 (siGrp78) is 50-ggagcg-cauugauacuagatt-30 as described (21). The control siRNA(siCtrl) is 50-aaggagacguauagcaacggu-30, composed of a 21-base pair scrambled sequence without significant homology toany known gene sequences from mouse, rat, or human. Theexperiments were repeated 3 times.

Detection of cell-surface GRP78 proteinThe biotinylation and detection of cell-surface GRP78 by

Western blotting in HMECs 48 hours posttransfection weredone as previously described (4). The experiments wererepeated 3 times.

MTT cell viability assaysHMECs were seeded in quadruplicate at a density of 3� 103

per well in 96-well plates 48 hours posttransfection and grownfor an additional 48 hours. TheMTT assay was done accordingto the manufacturer's instructions (Sigma-Aldrich) and aspreviously described (15).

BrdU cell proliferation assaysHMECs were seeded in quadruplicate at a density of 3 �

103 per well in 96-well plates 48 hours posttransfection andgrown for an additional 24, 48, or 72 hours. The bromodeox-yuridine (BrdU) cell proliferation assay (Roche DiagnosticsCorp.) was conducted according to the manufacturer'sinstructions.

Migration assaysHMECs were seeded in duplicate at a density of 4 � 104

per modified Transwell Boyden chamber (BD BioCoat) 48hours posttransfection. Each chamber housed a 6.5-mm-diameter polyethylene terephthalate filter with 8-mm pores.Migration was stimulated by adding complete growth mediato the lower compartment of the experimental apparatus.After 6 hours, the filters were fixed and stained with HarlecoHemacolor staining solution (EMD Chemicals). Cells thatmigrated to the underside of the filter were quantified underhigh-power magnification (400�). A 6-hour migration periodwas used to eliminate the effect of disparate rates of cellproliferation.

TUNEL assayHMECs were seeded at a density of 1.5 � 104 per well on

glass coverslips 48 hours posttransfection. These cells weremounted with ProLong Gold Anifade mounting medium with40,6-diamidino-2-phenylindole (DAPI). The apoptotic cellsdetected by the In Situ Cell Death Detection Kit, TMR red(Roche Applied Science) were visualized using a fluorescencemicroscope. A total of 1,000 cells were counted per treatmentcondition.

Statistical analysisFor the syngeneic E0771 mammary tumor model, a linear

model was used to compare tumor volume over time, withslope and quadratic and cubic terms for eachmouse treated asrandom. The likelihood ratio test for the Group � Timeinteraction was used to indicate whether the tumor growthpatterns were significantly different between the 2 genotypes.This analysis was based on the logarithm of tumor volumeþ 1.For the B16 melanoma tumor model, the log-rank test wasused to compare time to lung metastasis between the 2genotypes stratifying by experiment. The Pike estimates ofrelative HR were calculated using the observed and expectednumbers of events based on the log-rank test statistic.Kaplan–Meier plots were graphed for time-to-lung metastasis.Two-way ANOVA was carried out for comparison of pulmon-ary metastases and BrdU incorporation in HMECs, withgenotype and weeks as the 2 factors. Prior to ANOVA, com-paring the endothelial cells with and without GRP78 knock-down, logarithm was taken of the responses to render the datacompatible with the assumptions of normality and homosce-dasticity. Pair-wise comparisons among the groups were doneusing the least significance difference method if the overall Pvalue was less than 0.05. t tests were used when only 2 groupswere compared for a measurement.

Results

Grp78 heterozygosity in host environment suppressesearly tumor growth

In the MMTV-PyT transgene-driven mammary tumormodel, tumors were formed in both the Grp78þ/þ andGrp78þ/� mice. Previously, we reported the latency periodfor tumor formation was delayed and the size of the primarytumor at 15 weeks was reduced by about 60% in the Grp78þ/�

mice (11). On following the mouse cohorts for a longer period(after 18 weeks), secondary tumors became apparent. In thisendogenous tumor model, in which both the tumor and hostcells were heterozygous for Grp78, there were fewer totalnumber of tumors (P < 0.05) and the tumors were of smallersize (P < 0.01) in Grp78þ/�PyTmice than in Grp78þ/þPyTmice(Fig. 1A and B).

To examine the role of GRP78 in tumor growth in thehost environment, the growth of tumors derived fromsyngeneic WT murine E0771 breast cancer cells injectedsubcutaneously into either Grp78þ/þ (n ¼ 12) or Grp78þ/�

(n ¼ 15) mice was monitored. Thus, in these experiments,only the host environment was varied. The likelihood ratiotest based on random coefficient model comparing the 2curves showed a significant reduction of E0771 tumorgrowth in Grp78þ/� mice as compared with the WT mice(P < 0.001; Fig. 1B, left panel). Tumor size was examined atdifferent times during tumor development. The early phasereferred to the period between 10 and 15 days followingtumor implantation, and the late phase referred to theperiod between 20 and 25 days posttumor injection. Ourresults showed that Grp78 heterozygosity in the host pri-marily affected the early phase of tumor growth (Fig. 1B,right panel); as after 20 days, the tumor growth rate

Dong et al.

Cancer Res; 71(8) April 15, 2011 Cancer Research2850




between the 2 host genotypes was parallel to each other(Fig. 1A, left panel). The slower growth of the early tumorsin the Grp78þ/� mice associated with a 50% reduction (P <0.001) in tumor cell proliferation compared with theGrp78þ/þ mice, as assayed by proliferating cell nuclearantigen (PCNA)-positive staining cells, whereas duringthe late phase, tumor cells in both host genotypesshowed similar number of proliferating cells (Fig. 1C).Collectively, these data suggest that Grp78 heterozygosityin the host environment is sufficient to impede tumorgrowth and proliferation at the early phase of tumorgrowth; however, at the late stage, the WT tumorcells can overcome the host deficiency due to Grp78heterozygosity.

Grp78 host heterozygosity suppresses tumorangiogenesis during early tumor growth

Initiation of tumor growth requires neovascularization. Toexamine the effects of Grp78 host heterozygosity on tumorMVD, E0771 breast tumors harvested from Grp78þ/þ orGrp78þ/� mice were snap frozen and stained with anti-CD31 antibody, which is specific for endothelial cells. Tumorsfrom both early- and late-phase periods were analyzed. Theearly tumors of the Grp78þ/� mice showed 70% reduction inMVD compared with Grp78þ/þ mice (P < 0.01), whereas thedifferences in the MVD between the 2 host genotypes were notsignificant in the late-phase tumors (P ¼ 0.68; Fig. 2A). Theseresults were confirmed with fluorescence staining of CD31 inparaffin-embedded tumor sections (Fig. 2B). In contrast,

Figure 1. Grp78 host heterozygosity suppresses early-phase tumor growth. A, comparison of the size and total number of endogenous tumors(primary and secondary) in female Grp78þ/þ;PyT mice, and Grp78þ/�;PyT mice (n ¼ 7 for each genotype) at 18 weeks old (size, *, P < 0.05; number,**, P < 0.01). B, left, cohorts of femaleGrp78þ/þmice (n¼ 12) andGrp78þ/�mice (n¼ 15) were monitored for the time of appearance and size of the implantedWT E0771 tumor in each mouse. Dots, observed mean tumor volumes; lines, model-predicted tumor volumes. The likelihood ratio test based on randomcoefficient model comparing the 2 curves showed the significant reduction of E0771 tumor growth in heterozygous mice compared with the WT mice(P < 0.001). E and L denote early (10–15 days) and late (20–25 days) phase following tumor implantation. Right, enlarged view of the initial tumor growth patternfor the 2 host genotypes. C, left, paraffin-embedded E0771 tumor sections were subjected to immunofluorescent staining for PCNA (green), and the nucleiwere counterstained with DAPI (blue). The merged images are representative of 6 animals examined. Bar, 25 mm. Right, quantitation of the percentage ofPCNA-positive staining cells. The difference in early-phase tumor between the 2 genotypes is significant (***, P < 0.001).






staining for the VEGF was similar between the 2 host geno-types in both early- and late-phase tumors (Fig. 2C). Thissuggests that VEGF, primarily produced and released by theWT tumor cells (22), was not obviously affected by Grp78 hostheterozygosity.

Grp78 host heterozygosity suppresses growth ofmetastatic lesions

Formation of new tumor vasculature is critical for theestablishment and growth of metastatic colonies. To testwhether the inhibition of early-phase tumor angiogenesisby Grp78 host heterozygosity negatively impacts tumormetas-tasis, we utilized the syngeneic B16 melanoma tumor model. Anewly derived melanoma clonal cell line (B16-Fluc-A1) wasgenerated through transduction with a lentivirus expressingluciferase, followed by serial dilution and screening. Luciferaseexpression was approximately 5 times stronger in the B16-Fluc-A1 cell line than in the original B16-Fluc cell line that alsoexpressed the green fluorescent protein, with comparable invitro and in vivo growth characteristics (23). Following tail veininjection of the WT tumor cells (5� 104/mice; n¼ 10 for eachhost genotype), the progress of metastasis was first deter-mined by whole-body luminescence imaging and representa-tive images are shown in Figure 3A. Visible large tumorcolonies were detected in the lungs of syngeneic Grp78þ/þ

mice beginning at 3 weeks andmore were observed at 4 weeks,whereas Grp78þ/� mice showed no visible tumor at 3 weeksand only 1 mice developed a small visible tumor at 4 weeks(Fig. 3A). Following sacrifice at 2, 3, and 4 weeks, the lungs ofthe Grp78þ/þ mice showed considerably larger tumor andgreater number of tumors than did Grp78þ/� mice (Fig. 3B).The number of tumors per lung for the 2 genotypes issummarized in Figure 3C. The result of stratified log-ranktest showed that the risk of developing lung metastasis forGrp78þ/� mice was 47% (95% CI: 25–89) of that for Grp78þ/þ

mice (P < 0.001; Fig. 3D). Thus, consistent with the notion thatGRP78 is required for tumor neovascularization, Grp78 het-erozygosity in the host environment suppressed the number ofvisible metastatic lesions and slowed time to progression ofmetastatic tumors.

Endothelial cell–specific Grp78 heterozygosity inhibitsangiogenesis and proliferation in metastases

Endothelial cells, the principal cell component of the angio-genesis process, are a critical cellular component of the tumormicroenvironment. Therefore, to understand the role of GRP78in endothelial cells, we created a new mouse model of hetero-zygous knockout of Grp78 specifically in endothelial cells bybreeding the Grp78 floxed/floxed mice (17) with the Tie2-Cre transgenic mice (18). The genotypes of the mouse strains

Figure 2. Grp78 hostheterozygosity inhibits tumorangiogenesis. A, left:representative CD31 stainingpattern of snap-frozen E0771tumor sections during early (E) andlate (L) tumor growth phase. Thesections were lightlycounterstained with hematoxylin.Reddish purple indicates positivestaining. Right, quantitation of theMVD in the tumor sections (n ¼ 6from each group). One unit isdefined as 1 mm2 for 1 � 200magnification field. The differencein MDV in early-phase tumors issignificant (**, P < 0.01). Thedifference in MDV from late-phasetumors is not significant (P¼ 0.68).Columns, mean MVD; bars, 95%CI. B and C, paraffin-embeddedE0771 tumor sections from early-and late-phase tumors subjectedto immunofluorescent staining forCD31 (B) or VEGF (C), denoted ingreen. Blue, nuclei werecounterstained with DAPI.Representative merged stainingpatterns are shown from eachhost genotype. Bar, 50 mm.

Dong et al.





were confirmed by PCR analysis (Supplementary Fig. S1A). TheCre-recombinase was driven by the receptor tyrosine kinaseTek promoter/enhancer, which is active predominantly inendothelial cells starting at embryonic E9.5 (24). The knockoutof the Grp78 floxed allele in primary endothelial cells isolatedfrom the Grp78F/þ;Tie2-Cre mice was confirmed by PCRanalysis (Supplementary Fig. S1B). Mice with the Grp78F/þ;Tie2-Cre genotype, which is phenotypically normal served asthe experimental group, with sibling Grp78F/þ mice as nega-tive controls. A high dose of B16-Luc-A1 melanoma cells wereinjected via tail vein into Grp78F/þ;Tie2-Cre and Grp78F/þmice (2� 105/mice; n¼ 21 for each genotype) to generatemore

metastatic lesions for comparison. Lungs were removed at 2, 3,and 4 week intervals (n ¼ 7 at each time point). The macro-scopic inspection showed clear reduction in metastasis in theGrp78F/þ;Tie2-Cremice and representative images are shownin Figure 4A. The visible pulmonary surface metastases werefewer in the Grp78F/þ;Tie2-Cre mice and the differencebecame greater as time increased (P < 0.001) for the interactionbetween group and time. The mean difference (and associated95%CI) in number of colonies per lung between the 2 groups ofmice was 60 (32–88), 167 (139–195), and 355 (327–383) at 2weeks, 3 weeks, and 4 weeks (all with P < 0.001), respectively(Fig. 4B). Thus,Grp78 heterozygosity in endothelial cells causessignificant delay and decrease in metastatic growth. In thesesame groups of mice, CD31 staining of cryostat sections ofnormal organs such as brain, liver, and heart showed minimaldifference in the blood vessel density (Fig. 4C and D). Incontrast, immunofluorescence staining of paraffin-embeddedlung sections of these mice with the CD31 antibody revealedfewer vessels at the border between the tumor and the adjacentnormal tissue in the Grp78F/þ;Tie2-Cre tumor sections than inGrp78F/þ controls (Fig. 5A), correlating with reduced PCNAstaining in the Grp78F/þ;Tie2-Cre tumor sections (Fig. 5B).Thus, endothelial cell–specific heterozygous knockdown ofGRP78 is sufficient to impede growth of metastatic lesionsand vascularization at the growing edge of the tumor withminimal effect on normal organs.

Knockdown of GRP78 suppresses proliferation,migration, and promotes apoptosis in immortalizedendothelial cells

To identify the functional consequence resulting fromdecreased GRP78 levels in endothelial cells mimicking thosein the tumor microenvironment, in vitro studies were con-ducted with HMECs. The immortalized HMECs are SV40-transformed endothelial cells (25), which constitutively expresshigh levels of GRP78, as observed in tumor-associated endothe-lial cells (14, 15). The HMECs were transiently transfected withsiRNA specifically targeting Grp78 (siGrp78) or control siRNA(siCtrl). Western blot analysis of whole-cell lysates indicatedthat a greater than 85% knockdown of GRP78 expression wasachieved (Fig. 6A). In the HMECs, biotinylation of cell-surfaceprotein followed by avidin agarose pull down andWestern blotrevealed that approximately 1% of total intracellular GRP78was present on the cell surface, and following siGrp78 treat-ment, cell-surface GRP78 was below detection limits (Fig. 6A).Investigation of the overall viability of cells using the MTTassay showed a 55% (95% CI: 49–61; P < 0.001) decrease in thenumber of viable HMECs at 48 hours following GRP78 knock-down (Fig. 6B). Analysis of cell proliferation via incorporationof BrdU showed that knockdown of GRP78 reduced cell pro-liferation and the reduction became greater as time increased(P < 0.001 for the interaction between cell type and time).Knockdown of GRP78 reduced cell proliferation by 24% (95%CI: 4–40; P ¼ 0.024), 61% (95% CI: 50–69; P< 0.001), and 79%(95% CI: 73–83; P < 0.001) when compared with siCtrl-trans-fected HMECs at 24, 48, and 72 hours, respectively (Fig. 6B). Inaddition to proliferation, within 6 hours of the knockdown ofGRP78, HMECs showed reduced overall cell migration by 31%

Figure 3. Grp78 host heterozygosity inhibits pulmonary melanomametastasis. A, representative luminescence images of Grp78þ/þ andGrp78þ/� mice injected with B16-Fluc-A1 melanoma cells (5 � 104/mice)at the 1- to 4-week following tail vein injection (n ¼ 10 for each hostgenotype). B, representative macroscopic photographs of lungs removedat 2, 3, and 4 weeks after tumor cell injection. Examples of metastaticlesions are indicated by blue arrows. C, lung surface metastases for eachhost genotype were quantitated and SD was indicated. D, the stratifiedlog-rank test was used to compare time to lung metastasis between the 2groups. The risk of developing lungmetastasis forGrp78þ/�mice was 47%(95% CI: 25–89) of that for Grp78þ/þ mice (P < 0.001).






(95% CI: 16–43; P¼ 0.001) cells per hpf (Fig. 6B). Partial loss ofGRP78 led to cell death; however, it required at least 48 hours.Terminal deoxynucleotidyl transferase–mediated dUTP nickend labeling (TUNEL) staining revealed that apoptosis insiGrp78-transfected cells at 48 hours posttransfection wassignificantly increased, from 2.5% in siCtrl-transfected cellsto 23% (P < 0.001; Fig. 6C). Thus, reduced GRP78 expression in

immortalized endothelial cells suppresses cell proliferation,promotes apoptosis, and decreases migration.

Discussion

In an array of tissue culture and tumor models, GRP78 hasbeen established to be important for tumor progression and

Figure 4. Endothelial cell–specificGrp78 heterozygosity suppressespulmonary melanoma metastasis.A, representative macroscopicphotographs of lungs fromGrp78F/þ andGrp78F/þ;Tie2-Cremice at 2, 3, and 4 weeks followingtail vein injection of B16-Fluc-A1melanoma cells (2 � 105/mice;n ¼ 21 for each genotype).B, quantitation of averagepulmonary surface metastases at2, 3, and 4 weeks. Significantdifferences between the 2 hostgenotypes are detected for all timepoints (***, P < 0.001). The meansand associated 95% CIs areindicated. C, representative CD31staining pattern of sections offrozen brain tissue from Grp78F/þand Grp78F/þ;Tie2-Cre mice.Reddish purple indicates positivestaining. D, summary of MVD ofbrain, liver, and heart fromGrp78F/þ andGrp78F/þ;Tie2-Cremice. The standard error is shown.

Figure 5. Reduction of tumorangiogenesis and proliferation inmice with endothelial cell–specificGrp78 heterozygous knockoutmice. Paraffin-embeddedpulmonary B16-Fluc-A1melanoma sections from theGrp78F/þ and Grp78F/þ;Tie2-Cre mice were subjected toimmunofluorescent staining forCD31 (A) and PCNA (B) denoted ingreen, with nuclei counterstainedwith DAPI (blue). Representativehematoxylin and fluorescencestaining patterns are shown foreach host genotype. Bar, 50 mm.

Dong et al.





chemotherapeutic drug resistance, correlating with clinicalstudies that patients with high GRP78 expression levels intheir tumors generally have a poorer prognosis and experienceearlier relapse (3, 9, 11, 21, 26, 27). These discoveries suggestthat GRP78 can be a potent therapeutic target for anticancertherapy. Previous studies on GRP78 function in cancer pro-gression have largely focused on tumor cells and how GRP78exerts its protective antiapoptotic effect to allow them toescape immunosurveillance and stress associated with tumorprogression (3, 9). Here, we examined the role of GRP78 in thetumor microenvironment. In 2 syngeneic tumor models, WTtumors in the Grp78 heterozygous host environment weresignificantly impeded in tumor angiogenesis and proliferation.Through the creation of a new mouse model with endothelialcell–specific heterozygous knockout of Grp78, and coupledwith in vitro studies on GRP78 knockdown in immortalizedendothelial cells, we further showed that GRP78 is an impor-

tant mediator of endothelial cell proliferation, survival, andmigration and provided a newmechanistic explanation for therequirement of GRP78 in the microenvironment during tumorgrowth and metastasis.

How might GRP78 regulate function and promote survivalof endothelial cells supporting tumor growth? Expression ofGRP78, with well-established chaperone function (28), isinduced in hypoxic endothelial cells (29). Induction ofGRP78 leads to increase in both intracellular form ofGRP78 and cell surface of GRP78 (4). Upregulationof GRP78 in the ER can protect endothelial cells against ERstress, maintain ER calcium homeostasis, and suppress onsetof stress-induced apoptosis through suppression of caspase 7and CHOP induction, which mediate cytotoxicity in cellsundergoing ER (17, 30). On the other hand, cell-surfaceGRP78 is reported to colocalize with proangiogenic growthfactor receptors, promoting growth signaling, proliferation,

Figure 6. GRP78 regulates function and viability of endothelial cells. A, Western blot analysis of total and cell-surface GRP78 expression level inHMECs transfected with control siRNA (siCtrl) or siRNA against Grp78 (siGrp78), with b-actin serving as a loading control for the total lysate samplesand a control free of cytoplasmic contamination for the cell-surface protein samples. The amount of lysate applied onto the gel represented 20% ofinput. B, left, analysis of cell viability at 48 hours posttransfection by the MTT assay showed that knockdown of GRP78 significantly reduced cell survival(***, P < 0.001). Middle, BrdU incorporation in HMECs showed knockdown of GRP78 significantly decreased cell proliferation at 24, 48, and 72 hours(**, P¼ 0.024 at 24-hour time point and ***, P� 0.001 for 48- and 72-hour time points). Data are expressed as themean OD of the siGRP78 group relative to themean OD of the siCtrl group at each time point. The 95% CIs are indicated. Data are based on 2 independent experiments. Right, HMEC migrationwas assessed by the modified transwell Boyden chamber assay. After a 6-hour migration period, the number of migrated HMECs was significantly lower in thesiGrp78-transfected group than in the siCtrl-transfected group (***, P < 0.001). Data represent the mean number of cells per hpf that migrated to the undersideof the porous filter and are representative of 2 independent experiments. The 95% CIs are indicated. C, left, detection of apoptotic cells (in redindicated by arrows) by TUNEL assay in the transfected HMECs. Bar, 50 mm. Right, quantitation of apoptosis cells in the transfected groups withsignificant differences detected (***, P < 0.001).






and cell migration (5, 6, 8). GRP78 cell-surface expression isinduced by VEGF in human umbilical vein endothelial cellsand is required for VEGF-induced proliferation and angio-genic signaling (31). Therefore, the observed decrease inHMEC proliferation and migration, 2 critical angiogenic func-tions, correlates well with the consequences of GRP78 knock-down in these cells. Interestingly, our studies indicate that theeffects of Grp78 host heterozygosity are most evident in aminimal growth factor environment at the early stage oftumor growth. A possible explanation is that as the tumorgrows bigger, sufficient levels of proangiogenic growth factorsare produced and/or secreted by the WT tumor cells tostimulate tumor endothelial cell growth, hence can compen-sate for the effects of GRP78 loss in the host microenviron-ment.

Given the critical role of GRP78 both in tumor cell survivaland in endothelial cells supporting tumor growth, it repre-sents a dual target for anticancer therapy. GRP78 is thereceptor for Kringle 5, which induces apoptosis of both cancercells and endothelial cells (29). Sensitization of Kringle 5–induced apoptosis of brain microvessel endothelial cells byradiation requires GRP78 (32). Proteomic studies suggest thatGRP78 reduces the efficacy of topoisomerase inhibitors toinduce endothelial cell apoptosis (33). Furthermore, sometumor cells are capable of secreting GRP78, which confersbortezomib resistance to endothelial cells (34). We show herethat reducing GRP78 affects several critical aspects of angio-genesis: endothelial cell number and endothelial cell migra-tion. Migration is especially important because blood vesselgrowth involves the migration of endothelial progenitor cellsto the tumors and the expansion of neighboring blood vessels(35). Furthermore, other antiangiogenic drugs [i.e., bevacizu-mab (Avastin)] alter the VEGF levels in the patient, resulting ina change in the delicate balance of angiogenic growth factors.Such an imbalance can cause severe blood vessel regrowthand resistance to drugs (36). It is recently reported thatangiogenesis inhibitors targeting the VEGF and other proan-

giogenic growth factor pathways, while showing antitumoreffects, concomitantly elicit tumor adaptation leading toaccelerated metastasis (37, 38). Here, we showed that Grp78host heterozygosity does not seem to affect VEGF staining inthe tumor sections, suggesting that VEGF production in thetumor tissue is not altered. Furthermore, the decrease intumor metastatic growth in both Grp78þ/� and Grp78F/þ;Tie2-Cre mice persisted throughout the entire course of theexperiments, which was terminated at 4 weeks when animalsrequired euthanasia. One explanation is that altering thecytokine balance by decreasing VEGF activation is likely toinduce a response that is different from decreasing GRP78expression, as GRP78 regulates a much wider repertoire ofcellular function (2, 7). Thus, concomitant use of currentchemotherapeutic agents and novel therapies againstGRP78 may offer a powerful new approach to arrest cancerinitiation, progression, and metastasis.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Dr. Enrico Mihich of Roswell Park Cancer Institute for providingthe E0771 cells, Sam Li, Grant Dagliyan, and the USC Norris Cancer Center SmallAnimal Imaging Core Facility for assistance and advice in mouse imaging.

Grant Support

The work was supported by NIH grant CA027607 (A.S. Lee), NIH grantsCA127565, CA127135, andMelanoma Research Foundation grant (N. Craft), NIHgrant CA014089 (S. Groshen), and Sarnoff Cardiovascular Research Foundation(C. Stapleton).

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 25, 2010; revised February 9, 2011; accepted February 17,2011; published OnlineFirst April 5, 2011.

References1. Ni M, Lee AS. ER chaperones in mammalian development and human

diseases. FEBS Lett 2007;581:3641–51.2. Pfaffenbach KT, Lee AS. The critical role of GRP78 in physiologic and

pathologicstress.CurrOpinCellBiol2010Oct21. [Epubaheadofprint].3. Lee AS. GRP78 induction in cancer: therapeutic and prognostic

implications. Cancer Res 2007;67:3496–9.4. Zhang Y, Liu R, Ni M, Gill P, Lee AS. Cell surface relocalization of the

endoplasmic reticulum chaperone and unfolded protein responseregulator GRP78/BiP. J Biol Chem 2010;285:15065–75.

5. Shani G, Fischer WH, Justice NJ, Kelber JA, Vale W, Gray PC. GRP78and Cripto form a complex at the cell surface and collaborate to inhibittransforming growth factor beta signaling and enhance cell growth.Mol Cell Biol 2008;28:666–77.

6. Gonzalez-Gronow M, Selim MA, Papalas J, Pizzo SV. GRP78: amultifunctional receptor on the cell surface. Antioxid Redox Signal2009;11:2299–306.

7. Ni M, Zhang Y, Lee AS. Beyond the endoplasmic reticulum: atypicalGRP78 in cell viability, signaling and therapeutic targeting. Biochem J2011;434:181–8.

8. Wang M, Wey S, Zhang Y, Ye R, Lee AS. Role of the unfolded proteinresponse regulator GRP78/BiP in development, cancer, and neuro-logical disorders. Antioxid Redox Signal 2009;11:2307–16.

9. Li J, Lee AS. Stress induction of GRP78/BiP and its role in cancer. CurrMol Med 2006;6:45–54.

10. Luo S, Mao C, Lee B, Lee AS. GRP78/BiP is required for cell pro-liferation and protecting the inner cell mass from apoptosis duringearly mouse embryonic development. Mol Cell Biol 2006;26:5688–97.

11. Dong D, Ni M, Li J, Xiong S, Ye W, Virrey JJ, et al. Critical role of thestress chaperone GRP78/BiP in tumor proliferation, survival andtumor angiogenesis in transgene-induced mammary tumor develop-ment. Cancer Res 2008;68:498–505.

12. Bergers G, Benjamin LE. Tumorigenesis and the angiogenic switch.Nat Rev Cancer 2003;3:401–10.

13. Folkman J. The role of angiogenesis in tumor growth. Semin CancerBiol 1992;3:65–71.

14. Charalambous C, Chen TC, Hofman FM. Characteristics of tumor-associated endothelial cells derived from glioblastoma multiforme.Neurosurg Focus 2006;20:E22.

15. Virrey JJ, Dong D, Stiles C, Patterson JB, Pen L, Ni M, et al. Stresschaperone GRP78/BiP confers chemoresistance to tumor-associatedendothelial cells. Mol Cancer Res 2008;6:1268–75.

16. Ewens A, Luo L, Berleth E, Alderfer J, Wollman R, Hafeez BB, et al.Doxorubicin plus interleukin-2 chemoimmunotherapy against breastcancer in mice. Cancer Res 2006;66:5419–26.

Dong et al.





17. Wang M, Ye R, Barron E, Baumeister P, Mao C, Luo S, et al. Essentialrole of the unfolded protein response regulator GRP78/BiP in protec-tion from neuronal apoptosis. Cell Death Differ 2010;17:488–98.

18. Koni PA, Joshi SK, Temann UA, Olson D, Burkly L, Flavell RA.Conditional vascular cell adhesion molecule 1 deletion in mice:impaired lymphocyte migration to bone marrow. J Exp Med 2001;193:741–54.

19. Wu Z, Hofman FM, Zlokovic BV. A simple method for isolation andcharacterization of mouse brain microvascular endothelial cells. JNeurosci Methods 2003;130:53–63.

20. Perriere N, Demeuse P, Garcia E, Regina A, Debray M, Andreux JP,et al. Puromycin-based purification of rat brain capillary endothelialcell cultures. Effect on the expression of blood-brain barrier-specificproperties. J Neurochem 2005;93:279–89.

21. Fu Y, Wey S, Wang M, Ye R, Liao CP, Roy-Burman P, et al. Pten nullprostate tumorigenesis and AKT activation are blocked by targetedknockout of ER chaperone GRP78/BiP in prostate epithelium. ProcNatl Acad Sci U S A 2008;105:19444–9.

22. Kim KJ, Li B, Winer J, Armanini M, Gillett N, Phillips HS, et al. Inhibitionof vascular endothelial growth factor-induced angiogenesis sup-presses tumour growth in vivo. Nature 1993;362:841–4.

23. Craft N, Bruhn KW, Nguyen BD, Prins R, Liau LM, Collisson EA, et al.Bioluminescent imaging of melanoma in live mice. J Invest Dermatol2005;125:159–65.

24. Hamada K, Sasaki T, Koni PA, Natsui M, Kishimoto H, Sasaki J, et al.The PTEN/PI3K pathway governs normal vascular development andtumor angiogenesis. Genes Dev 2005;19:2054–65.

25. Ades EW, Candal FJ, Swerlick RA, George VG, Summers S, BosseDC, et al. HMEC-1: establishment of an immortalized human micro-vascular endothelial cell line. J Invest Dermatol 1992;99:683–90.

26. Pyrko P, Schonthal AH, Hofman FM, Chen TC, Lee AS. The unfoldedprotein response regulator GRP78/BiP as a novel target for increasingchemosensitivity in malignant gliomas. Cancer Res 2007;67:9809–16.

27. Baumeister P, Dong D, Fu Y, Lee AS. Transcriptional induction ofGRP78/BiP by histone deacetylase inhibitors and resistance to his-tone deacetylase inhibitor-induced apoptosis. Mol Cancer Ther2009;8:1086–94.

28. Hendershot LM. The ER function BiP is a master regulator of ERfunction. Mt Sinai J Med 2004;71:289–97.

29. Davidson DJ, Haskell C, Majest S, Kherzai A, Egan DA, Walter KA,et al. Kringle 5 of human plasminogen induces apoptosis of endothe-lial and tumor cells through surface-expressed glucose-regulatedprotein 78. Cancer Res 2005;65:4663–72.

30. Reddy RK, Mao C, Baumeister P, Austin RC, Kaufman RJ, Lee AS.Endoplasmic reticulum chaperone protein GRP78 protects cells fromapoptosis induced by topoisomerase inhibitors: role of ATP bindingsite in suppression of caspase-7 activation. J Biol Chem 2003;278:20915–24.

31. Katanasaka Y, Ishii T, Asai T, Naitou H, Maeda N, Koizumi F, et al.Cancer antineovascular therapy with liposome drug delivery systemstargeted to BiP/GRP78. Int J Cancer 2010;127:2685–98.

32. McFarland BC, Stewart J Jr, Hamza A, Nordal R, Davidson DJ, HenkinJ, et al. Plasminogen Kringle 5 induces apoptosis of brain microvesselendothelial cells: sensitization by radiation and requirement forGRP78 and LRP1. Cancer Res 2009;69:5537–45.

33. Bruneel A, Labas V, Mailloux A, Sharma S, Royer N, Vinh J, et al.Proteomics of human umbilical vein endothelial cells applied to etopo-side-induced apoptosis. Proteomics 2005;5:3876–84.

34. Kern J, Untergasser G, Zenzmaier C, Sarg B, Gastl G, Gunsilius E,et al. GRP-78 secreted by tumor cells blocks the antiangiogenicactivity of bortezomib. Blood 2009;114:3960–7.

35. Patenaude A, Parker J, Karsan A. Involvement of endothelial progeni-tor cells in tumor vascularization. Microvasc Res 2010;79:217–23.

36. Xu L, Duda DG, di Tomaso E, AncukiewiczM, Chung DC, Lauwers GY,et al. Direct evidence that bevacizumab, an anti-VEGF antibody, up-regulates SDF1alpha, CXCR4, CXCL6, and neuropilin 1 in tumors frompatients with rectal cancer. Cancer Res 2009;69:7905–10.

37. Paez-Ribes M, Allen E, Hudock J, Takeda T, Okuyama H, Vinals F,et al. Antiangiogenic therapy elicits malignant progression of tumorsto increased local invasion and distant metastasis. Cancer Cell2009;15:220–31.

38. Ebos JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG,Kerbel RS. Accelerated metastasis after short-term treatment with apotent inhibitor of tumor angiogenesis. Cancer Cell 2009;15:232–9.






2011;71:2848-2857. Published OnlineFirst April 5, 2011.Cancer Res Dezheng Dong, Christopher Stapleton, Biquan Luo, et al. for Neovascularization during Tumor Growth and MetastasisA Critical Role for GRP78/BiP in the Tumor Microenvironment

Updated version

10.1158/0008-5472.CAN-10-3151doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2011/04/04/0008-5472.CAN-10-3151.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/71/8/2848.full#ref-list-1

This article cites 37 articles, 18 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/71/8/2848.full#related-urls

This article has been cited by 14 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/71/8/2848To 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-10-3151

http://cancerres.aacrjournals.org/content/suppl/2011/04/04/0008-5472.CAN-10-3151.DC1

http://cancerres.aacrjournals.org/content/71/8/2848.full#ref-list-1

http://cancerres.aacrjournals.org/content/71/8/2848.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/71/8/2848


A Critical Role for GRP78/BiP in the Tumor ... · and maintenance of the tumor (12, 13)....

Documents

Transcript of A Critical Role for GRP78/BiP in the Tumor ... · and maintenance of the tumor (12, 13)....