Fetal stromal-dependent paracrine and intracrine vascular endothelial growth factor-a/vascular endothelial growth factor receptor-1 signaling promotes proliferation and motility of

Fetal Stromal–Dependent Paracrine and Intracrine Vascular

Endothelial Growth Factor-A/Vascular Endothelial Growth

Factor Receptor-1 Signaling Promotes Proliferation and

Motility of Human Primary Myeloma Cells

Loı̈c Vincent,1David K. Jin,

1,2Matthias A. Karajannis,

1Koji Shido,

1Andrea T. Hooper,

1

William K. Rashbaum,3Bronislaw Pytowski,

4Yan Wu,

4Daniel J. Hicklin,

4

Zhenping Zhu,4Peter Bohlen,

4Ruben Niesvizky,

2and Shahin Rafii

1,2

Departments of 1Genetic Medicine, 2Medicine and Hematology-Medical Oncology, and 3Obstetrics and Gynecology, Weill Medical Collegeof Cornell University; and 4ImClone Systems, Inc., New York, New York

Abstract

Induction of neoangiogenesis plays an important role in thepathogenesis of multiple myeloma. However, the mechanismby which expression of vascular endothelial growth factor(VEGF)-A and its receptors modulate the interaction ofmultiple myeloma cells with stromal cells is not known. Here,we describe a novel in vitro coculture system using fetalbone stromal cells as a feeder layer, which facilitates thesurvival and growth of human primary multiple myelomacells. We show that stromal-dependent paracrine VEGF-Asignaling promotes proliferation of human primary multiplemyeloma cells. Primary multiple myeloma cells onlyexpressed functional VEGF receptor (VEGFR)-1, but notVEGFR-2 or VEGFR-3. VEGFR-1 expression was detected inthe cytoplasm and the nuclei of proliferating multiplemyeloma cells. Inhibition of VEGFR-1 abrogated multiplemyeloma cell proliferation and motility, suggesting that thefunctional interaction of VEGF-A with its cognate receptor isessential for the growth of primary multiple myeloma cells.Collectively, our results suggest that stromal-dependentparacrine and intracrine VEGF-A/VEGFR-1 signaling contrib-utes to human primary multiple myeloma cell growth andtherefore, VEGFR-1 blockade is a potential therapeuticstrategy for the treatment of multiple myeloma. (Cancer Res2005; 65(8): 3185-92)

Introduction

Multiple myeloma is a clonal B lymphocyte malignancy mainlycharacterized by the accumulation of terminally differentiatedantibody-producing cells in the bone marrow. Multiple myelomacells reside in close association with the stromal elements withinthe bone marrow. As the disease progresses, multiple myelomacells may proliferate in the extramedullary areas (1, 2). Currenttreatments for multiple myeloma offer only a median survival of3 years as investigators continue to search for novel therapeutictargets to combat the disease. One of the strategies is to targetthe multiple myeloma cells as well as the bone marrowmicroenvironment (3). It has been shown that moleculesdesigned to bind specifically to the tyrosine kinase domain of

vascular endothelial growth factor receptors (VEGFR) canefficiently block the growth of multiple myeloma cells (4, 5).VEGF-A and its receptor tyrosine kinases, namely VEGFR-1 (Flt-1)as well as VEGFR-2 (KDR, Flk-1), play a crucial role in neo-vascularization (6, 7).Functional VEGFRs are also expressed on subsets of leukemias,

resulting in autocrine loops that sustain leukemia migration andproliferation (8). VEGF-A is a potent activator of multiple myelomacell proliferation and migration (9, 10). Elevated plasma levels ofVEGF-A have been reported in patients with multiple myeloma (11)and correlated with increased angiogenesis in multiple myelomabone marrow (12–15). Moreover, binding of multiple myeloma cellsto bone marrow stromal cells markedly stimulates VEGF-Asecretion (16). This triggers interleukin-6 (IL-6) production frombone marrow stromal cells (17), thereby augmenting multiplemyeloma cell proliferation in a paracrine fashion. Thus, VEGF-Aplays an important role in both autocrine and paracrine control ofmultiple myeloma cell growth. Moreover, intracellular traffickingand nuclear localization of VEGFR-2 can promote leukemic cellproliferation (18). However, whether VEGF-A/VEGFR intracrinesignaling also supports the proliferation of primary multiplemyeloma cells is not known.Prior studies have shown that immortalized and primary

multiple myeloma cells express high-affinity VEGFR-1, but notVEGFR-2 (4, 5, 19). However, the precise functional role of VEGFRs,in particular VEGFR-1, in the regulation of proliferation andinteraction with the stromal cells is not known. In addition, studiesof multiple myeloma biology have been limited by the lack of acoculture model for cultivating human primary multiple myelomacells in vitro in order to study their behavior for sufficiently long-periods of time.Here we describe a novel in vitro coculture system using fetal

bone stromal cells as a feeder layer, which facilitates the growthand survival of human primary multiple myeloma cells. Thisprovides an opportunity to study the role of the VEGF-A/VEGFRaxis which accounts for the multiple myeloma cell growth.Remarkably, we show that VEGFR-1 is present in the cytoplasmand the nuclei of proliferating multiple myeloma cells. Blockade ofVEGFR-1 by a neutralizing monoclonal antibody (mAb) maintainsVEGFR-1 in a membrane-bound localization preventing its nucleartranslocation, and therefore blocks proliferation and migration ofprimary multiple myeloma cells. Collectively, our results suggestthat stromal-dependent paracrine and intracrine VEGF-A/VEGFR-1 signaling contributes to human primary multiple myeloma cellsurvival, proliferation, and migration.

Requests for reprints: Loı̈c Vincent and Shahin Rafii, Department of Medicine andHematology-Medical Oncology, Weill Medical College of Cornell University, 1300 YorkAvenue, New York, NY 10021. Phone: 212-746-2070; E-mail: [email protected] [email protected].

I2005 American Association for Cancer Research.

www.aacrjournals.org 3185 Cancer Res 2005; 65: (8). April 15, 2005

Research Article

Research. on October 26, 2015. © 2005 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Materials and Methods

Cytokine and antibodies. Recombinant VEGF-A was obtained from

R&D Systems (Minneapolis, MN). The expression profile of VEGFRs in

multiple myeloma cells was determined by immunohistochemical

staining and flow cytometry using 1 Ag/mL of the following mAbs:

anti-Flt-1 (clone FB5), anti-KDR (clone 1121) and anti-Flt-4 antibodies

(clone 3C5). Neutralizing mAb (10 Ag/mL) to VEGFRs were used to study

the role of each receptor in multiple myeloma cells: anti-Flt-1 (clone

6.12), anti-KDR (clone 1C11) and anti-Flt-4 antibodies (clone 3C5). All of

these mAbs were kindly provided by ImClone Systems, Inc (New

York, NY).

Bone marrow specimens and culture of multiple myeloma cells.

Heparinized bone marrow aspirates were freshly obtained from multiple

myeloma patients after appropriate informed consent approved by the

Institutional Review Board at Cornell University Medical College. The bone

marrow specimens were mixed to disaggregate cell clumps. Excess tissue

and clumps were removed by filtering the cell suspension through a 70-

Amol/L nylon tissue strainer (BD Falcon, Bedford, MA). The cell filtrate was

subjected to RBC lysis buffer (1:5, v/v, Roche, Indianapolis, IN). After

incubation at room temperature for 10 minutes, cells were pelleted by

centrifugation, and washed once with serum-free X-VIVO 20 medium (Bio-

Whittaker, Walkersville, CA). Multiple myeloma cells were then purified

using magnetic cell sorting CD138 Microbeads (clone B-B4; Miltenyi Biotec,

Bergisch Gladbach, Germany) following the manufacturers’ instructions.

Cells were positively selected over a magnetic column with 95% to 99%

purity.

Fetal bone stroma isolation and culture. The Institutional Review

Board of Cornell University Medical College approved the use of fetal

tissue. Femur from one 8- to 12-week-old human fetus from spontaneous

abortion was mechanically dissociated and the cell suspension was

filtered with a syringe through a 40-Am nylon mesh (Millipore, Billerica,

MA). Fetal bone stromal cells were cultured in M199 medium containing

10% fetal bovine serum, 100 units/mL penicillin and 100 Ag/mL

streptomycin.

Characterization of fetal bone stromal cells. Fetal bone stromal cells

were grown on glass slides and fixed with 10% paraformaldehyde. For

immunohistochemistry, fetal bone stromal cells were subjected to primaryantibodies raised against platelet endothelial cell adhesion molecule-1 (1:40,

Dako, Glostrup, Denmark), vimentin (1:50, Dako, Carpinteria, CA), and

a-smooth muscle actin (1:400, Sigma, St. Louis, MO). Detection of boundprimary antibody was done with the Dako EnVision horseradish peroxidase/

3,3V-diaminobenzidine staining kit (Dako). Fetal bone stromal cells were also

stained for VEGFR-1, VEGFR-2, and VEGFR-3. Fetal bone stromal cells were

incubated with the mAb or an unspecific, isotype-matched murine antibodyas a control, washed and were then incubated with secondary FITC-

conjugated antibodies. Cells were washed and analyzed using a Coulter

FC-500 Flow Cytometer.

Angiogenesis array of fetal bone stromal cells. Fetal bone stromalcells were grown to 50% confluence, washed with serum-free X-VIVO 20

medium, and then incubated with X-VIVO 20 medium. After 72 hours,

supernatant was collected and analyzed for soluble angiogenic factors withan angiogenesis antibody array (Panomics, Redwood City, CA) as per the

manufacturer’s instructions.

Multiple myeloma cell growth on fetal bone stromal cellmonolayer. Fetal bone stromal cells (5 � 104 cells per well) were seeded

in 12-well plates and let to adhere before a 24-hour starvation, and

multiple myeloma cells (3 � 104/mL) were then added to the stromal

feeder layer in serum-free medium. The physical interaction between fetalbone stromal and multiple myeloma cells was further shown using PKH2

(green) and PKH26 (red) fluorescent cell linkers, respectively, according

to the manufacturer’s instructions (Sigma). The growth of cocultured

multiple myeloma cells was compared with that without stromal feederlayer. Proliferation of the multiple myeloma cells was evaluated by first

collecting the nonattached multiple myeloma cells, and second by

detaching the residual attached multiple myeloma cells using a celldissociation solution (Sigma). The viability of the combined collected

fractions of multiple myeloma cells was determined by trypan blueexclusion after 7, 14, and 21 days in the coculture system. Some functional

studies to investigate the role of VEGFRs were designed in which

neutralizing mAb to VEGFRs were added every 2 days in the culture

medium. These experiments were done on four multiple myeloma samplesand each condition was done in triplicate.

EBV analysis. To rule out the possibility that multiple myeloma cell

growth is secondary to EBV transformation, bone marrow aspirate and

cultured multiple myeloma cells were examined for the presence of EBV by

in situ hybridization for EBER (EBV-encoded small RNAs, Dako Corpora-

tion, Via Real, CA).

Apoptosis.Multiple Myeloma cells cocultured on fetal bone stromal cells

or alone in serum-free X-VIVO 20 medium were collected and analyzed forthe presence of apoptotic cells using the ApoAlert Annexin V-FITC

propidium iodide (PI) Apoptosis Kit (Becton Dickinson, Palo Alto, CA),

following the manufacturer’s instructions. Flow cytometry analysis wasperformed using a Coulter FC-500 Cytometer. Results were shown as the

percentage of live cells (annexin V- PI-), early apoptotic cells (annexin V+

PI-), late apoptotic cells (annexin V+ PI+), and dead cells (annexin V- PI+).

Each set of experiments was done in triplicate.Immunofluorescence microscopy. After coculturing with fetal

bone stromal cells for 21 days, multiple myeloma cells were spun

onto glass microscope slides and VEGFR-1, VEGFR-2, and VEGFR-3

expression was detected by immunofluorescent technique. Multiplemyeloma cells were fixed in 3.7% paraformaldehyde and washed in

PBS. After permeabilization with methanol (90%), multiple myeloma

cells were incubated with the primary antibodies, washed and thenincubated with secondary FITC-conjugated antibodies (1:1,000, Vector

Laboratories, Burlingame, CA). After washing, samples were

mounted in Vectashield containing 4V,6-diamidino-2-phenylindole

(DAPI) and analyzed by fluorescence microscopy at 40� magni-fication (Olympus, NJ).

Flow cytometry. The multiple myeloma cells were collected following

the 21-day coculture with fetal bone stromal cells and stained for

VEGFR-1, VEGFR-2 and VEGFR-3. The multiple myeloma cells were fixed

in 4% paraformaldehyde and permeabilized in methanol (90%). The

multiple myeloma cells were incubated with the specific neutralizing

mAb or an unspecific, isotype-matched murine antibody as a control,

washed, and then incubated with secondary phycoerythrin (PE)- or FITC-

conjugated antibodies. Cells were washed and analyzed using a Coulter

FC-500 Cytometer. VEGFR expression was analyzed in four primary

multiple myeloma samples.

Transwell migration assay. Primary multiple myeloma cell migrationwas assayed using a Boyden chamber system with 8 Am pore size

inserts (Costar, NY). Multiple myeloma cells were starved in X-VIVO

20 for 6 hours prior incubation with neutralizing mAb to VEGFR-1,

VEGFR-2, or VEGFR-3 for 4 hours. Multiple myeloma cells (1.5 � 105

per insert) were then placed into the upper chamber of the transwell

system in X-VIVO 20 supplemented with 0.2 mg/mL bovine serum

albumin. The lower chamber was filled with X-VIVO 20 (2 mg/mL

bovine serum albumin) in the presence or absence of 50 ng/mL ofVEGF-A. A fetal bone stromal cell monolayer cultured in serum-free

X-VIVO 20 for 2 days was also used as a chemoattractant. After a 6-hour

incubation, the number of viable migrated cells harvested in the lower

chamber was assessed by trypan blue exclusion. A migration index,defined as the number of viable migrated cells in the sample divided by

the number of viable migrated cells in the control (no cytokine or fetal

bone stromal cells), was calculated to compare motility of cells relativeto control. The transwell migration assay was repeated thrice in

triplicate.

ELISA. Following a 72 hour-incubation of multiple myeloma cells and

fetal bone stromal cells cultured in serum-free condition (X-VIVO 20),VEGF-A and placental growth factor levels were measured in the

supernatants using commercially available ELISA kits (R&D Systems).

Results are shown as picograms per milliliter per 1 � 106 cells of VEGF-A or

placental growth factor in culture supernatants, and each sample was doneindependently in triplicate.

Cancer Research

Cancer Res 2005; 65: (8). April 15, 2005 3186 www.aacrjournals.org



Statistical analysis. Statistical significance was determined using anunpaired Student’s t test. The minimal level of significance was P < 0.05.

Results

Characterization of fetal bone stromal cells. The difficulty ofcultivating multiple myeloma cells in vitro has limited thestudies of multiple myeloma cell biology. In this study, we useda fetal bone stromal cell feeder layer to support the long-termpropagation of primary human multiple myeloma cells. Usingimmunohistochemistry, we showed that fetal bone stromal cellsexpressed the mesenchymal antigen vimentin (Fig. 1A) and thea-smooth muscle actin (Fig. 1B), but were negative for theexpression of endothelial cell marker platelet endothelial celladhesion molecule-1 and VEGF receptors (data not shown). Thefully confluent cells secrete VEGF-A and IL-6 (Fig. 1C). As shownin Fig. 1D , multiple myeloma cells (yellow arrows) werecocultured on a semiconfluent fetal bone stromal cell monolayer(black arrow). Using fluorescent red and green dye, respectively,to label the multiple myeloma cells growing on fetal bonestromal cells, the multiple myeloma cells were detected toclosely interact with the fetal bone stromal cells (Fig. 1D). Aftera 21-day coculture, clusters of multiple myeloma cells could

easily be identified as cellular clusters attached to stromal cells(Fig. 1E).Fetal bone stromal cells support multiple myeloma growth.

To assess the effect of the fetal bone stromal cell on themultiple myeloma cell proliferation in serum-free condition, themultiple myeloma cells were collected every 7 days until day 21after the coculture started, and the number of multiple myelomacells was compared with the cells grown in the absence of astromal layer. Without fetal bone stromal cells, the percentage ofdead multiple myeloma cells increased to 78.78 F 13.33% and97.77 F 2.10% after 7 and 14 days of culture, respectively (Fig.2A). After a 21-day culture, all the multiple myeloma cellsunderwent apoptosis. In sharp contrast, the number of viablemultiple myeloma cells grown in coculture with the fetal bonestroma increased to 337.77 F 29.63%, which represents a 3-foldexpansion at the end of experiment. In the four specimenstested, all multiple myeloma cells were EBER-negative excludingthe possibility that the cultured multiple myeloma cells wereEBV-transformed (data not shown).Fetal bone stromal cells support multiple myeloma survival.

At day 21 of coculture, the collected multiple myeloma cells weresubjected to Annexin V and PI staining to assess the effect of fetalbone stromal cells on multiple myeloma cell survival. The majority

Figure 1. The fetal bone stromal cells support multiple myeloma cell growth. A and B, immunohistochemical staining of fetal bone stromal cells. The cells werestained with vimentin (A) and a-smooth muscle actin (B ). Fetal bone stromal cells were positive for the mesenchymal antigen vimentin and a-smooth muscle actinconsistent with an immunophenotype of activated fibroblasts. Magnification (�400); C, soluble factors expressed by fetal bone stromal cells. Fetal bone stromal cellswere grown in serum-free medium for 72 hours and supernatant was analyzed for soluble factors with an antibody array. Duplicate black dots show the presenceof soluble factors; D, association of fetal bone stromal and multiple myeloma cells. Multiple myeloma cells (yellow arrow ) were seeded on a semiconfluent fetal bonestromal cell monolayer (black arrow ). The interaction between fetal bone stromal and multiple myeloma cells was highlighted using PKH2 green and PKH26 redfluorescent cell linkers, respectively. Magnification (�400); E, sustained growth of multiple myeloma cells on fetal bone stromal cell monolayer. Multiple myeloma cellswere cocultured with fetal bone stromal cells in serum-free condition. Robust multiple myeloma cell proliferation was observed after 21 days. Magnification (�400).

VEGFR-1 Promotes Multiple Myeloma Growth




of the primary multiple myeloma cells (98% of the total cells) werealive (Annexin V� PI�) when they were seeded in serum-freecondition (Fig. 2B). Without fetal bone stromal cells to support themultiple myeloma cell growth, 97.8% of cells underwent apoptosisafter 21 days. In contrast, the fetal bone stromal cells significantlysupported the survival of the multiple myeloma cells (74.1%Annexin V� PI�, P < 0.05), although 25.6% of cells were in earlyand late apoptosis. The data are representative of three indepen-dent experiments.Vascular endothelial growth factor receptor-1 is localized to

the nucleus of the proliferating multiple myeloma cells. Wefurther evaluated the presence of VEGFR-1, VEGFR-2, and VEGFR-3 in primary human multiple myeloma cells. Using flow cytometryon three multiple myeloma samples, we were unable to detectexpression of VEGFR-1, VEGFR-2, and VEGFR-3 on the surface ofthe multiple myeloma cells (data not shown). However, VEGFR-1expression was detected within the permeabilized multiplemyeloma cells from all the studied bone marrow patients (4 outof 10) and the DAPI staining showed that VEGFR-1 waspredominantly present in the nuclei of the multiple myeloma cells(Fig. 3A). In contrast, neither VEGFR-2 nor VEGFR-3 staining wasdetectable in the four patient samples. The intracellular expressionof VEGFR-1 was confirmed by flow cytometry on four permeabi-lized multiple myeloma cells (Fig. 3B). The absence of VEGFR-2and VEGFR-3 was also confirmed by immunohistochemicalstaining.The blockade of vascular endothelial growth factor

receptor-1 inhibits multiple myeloma growth. Based on thenuclear presence of VEGFR-1 and the absence of VEGFR-2 andVEGFR-3 expression in the multiple myeloma cells, we hypothe-

sized that the expression of VEGFR-1 may play a critical role inmultiple myeloma biology. The multiple myeloma cell growthassay, in which the multiple myeloma cells were cocultured with asemiconfluent fetal bone stromal cell monolayer in serum-freeconditions, was done in the presence or absence of variousneutralizing mAb to VEGFRs. The neutralizing mAb against eitherVEGFR-2 or VEGFR-3 did not interfere with the multiple myelomacell growth after 21 days in the coculture system as compared withthe isotype IgG control (266.67 F 16% and 280 F 30% versus253.34 F 32.84%, respectively; Fig. 4). In contrast, the neutralizingmAb to VEGFR-1 markedly suppressed the growth of the multiplemyeloma cells as compared with the number of multiple myelomacells initially placed in the coculture system (47.66F 16%, P < 0.001,n = 4), suggesting that the expression of VEGFR-1 plays a key role inthe survival and proliferation of human primary multiple myelomacells. In addition, VEGFR-1 expression was localized to thecytoplasm but not the nucleus, when multiple myeloma cells weretreated with neutralizing mAb against VEGFR-1 (4 out of 10 bonemarrow samples; Fig. 5). The immunofluorescent staining forVEGFR-2 and VEGFR-3 was negative (data not shown).The blockade of vascular endothelial growth factor

receptor-1 inhibits multiple myeloma motility. Having shownthat VEGFR-1 supports multiple myeloma cell growth, we nextinvestigated whether VEGFR-1 blockade could interfere withmultiple myeloma cell motility and invasive potential. Cellmotility was assayed by measuring the transmembrane migrationactivity of multiple myeloma cells (3 multiple myeloma samplesout of 10) seeded in transwell inserts. As shown in Fig. 6, VEGF-Ainduced an increase in multiple myeloma cell migration (2.46 F0.08-fold, P < 0.001, n = 3; Fig. 6). Neutralizing mAb against

Figure 2. Fetal bone stromal cells support multiple myeloma cellgrowth and survival. A, multiple myeloma cell growth on fetal bonestromal cells. Multiple myeloma cells were seeded in serum-freecondition in the presence or absence of a semiconfluent fetal bonestromal monolayer for 21 days. Multiple myeloma cells werecollected every 7 days for assessment for viability by trypan blueexclusion. After 21 days in culture, all the multiple myelomacells without fetal bone stromal cells were dead, whereassustained growth of multiple myeloma cells was observed in thecoculture system (***P < 0.001 as compared with the numberof cells at day 0, n = 4); B , multiple myeloma cell survival onfetal bone stromal cells. The multiple myeloma cells werestained with Annexin V/PI at days 0 and 21 of the culture in thepresence (+FBS) or absence (�FBS) of the fetal bone stromalcells. Results are shown as the percentage of viable cells(Annexin V� PI�), early apoptotic cells (Annexin V+ PI�), lateapoptotic cells (Annexin V+ PI+), and dead cells (Annexin V� PI+).Without fetal bone stromal cells to support the multiple myelomacell growth, the multiple myeloma cells were prone to celldeath (PI+). In contrast, the fetal bone stromal cells significantlysupported the survival of the multiple myeloma cells (74.1%Annexin V� PI�; P < 0.05, n = 3). The data presentedare representative of three independent experiments.

Cancer Research




VEGFR-1 decreased the VEGF-A-induced transmigration ofmultiple myeloma cells (1.48 F 0.21-fold, P < 0.001, n = 3). Thepresence of fetal bone stromal cells conferred a strongerchemoattractive effect on the multiple myeloma cells (4.09 F0.60-fold, P < 0.001, n = 3) as compared with that by VEGF-A.Neutralizing mAb against VEGFR-1 also decreased the fetal bonestromal cell–induced multiple myeloma cell migration (2.49 F0.10-fold, P < 0.01, n = 3). However, neutralizing mAb againstVEGFR-1 did not completely block multiple myeloma celltransmigration, suggesting that other factors besides VEGF-Amay support the migration of the multiple myeloma cells.Blockade of VEGFR-2 and VEGFR-3 had no significant effect onmultiple myeloma cell motility.The fetal bone stromal cells secrete high levels of vascular

endothelial growth factor-A. In order to determine how

autocrine and/or paracrine VEGF-A and placental growth factorstimulation promote the growth and migration of multiplemyeloma cells, the protein levels of VEGF-A and placental growthfactor were determined in the serum-free medium obtained fromfetal bone stromal cells. Fetal bone stromal cells secreted highlevels of VEGF-A (2,348.2 F 254.0 pg/mL/1 � 106 cells, n = 3) ascompared with multiple myeloma cells (105.8 F 34.7 pg/mL/1 �106 cells, n = 3). This drastic difference in VEGF-A secretionsuggests that VEGF-A through paracrine stimulation supportsprimary multiple myeloma cell growth and migration. In contrast,the level of placental growth factor secreted by fetal bone stromalcells and multiple myeloma cells was negligible (data not shown),suggesting VEGF-A as the primary factor released that activatesVEGFR-1 signaling, thereby promoting multiple myeloma cellsurvival, proliferation, and migration.

Figure 3. VEGFR-1, but not VEGFR-2 andVEGFR-3, is expressed on multiple myeloma cells.A, the multiple myeloma cells were harvestedfollowing the 21-day coculture with fetal bonestromal cell monolayer and stained for VEGFRs.The multiple myeloma cells were permeabilizedand subjected to immunofluorescent staining usingmAb against VEGFRs (FITC) and DAPI to stainthe DNA (nucleus). VEGFR-1 expression wasdetected inside the multiple myeloma cells and theDAPI staining showed the nuclear expression ofVEGFR-1. Neither VEGFR-2 nor VEGFR-3 weredetectable. Results were obtained from fourdifferent multiple myeloma samples in independentexperiments. Magnification (�1,000); B, themultiple myeloma cells were collected following the21-day coculture with fetal bone stromal cellmonolayer and stained for VEGFRs. The multiplemyeloma cells were permeabilized and stainedusing mAb against VEGFR-1 (phycoerythrin,FL2), VEGFR-2 and VEGFR-3 (FITC, FL1).The intracellular expression of VEGFR-1 wasconfirmed using flow cytometry. VEGFRexpression was analyzed in four multiplemyeloma samples. The absence of VEGFR-2and VEGFR-3 was also confirmed.





Discussion

Accumulating evidence has shown that VEGFRs, includingVEGFR-1, VEGFR-2, and VEGFR-3 are expressed in subsets ofprimary human multiple myeloma cells. However, the functionalrole of the activation of VEGF/VEGFR signaling axis on primarymultiple myeloma survival, proliferation, and mobilization in thecontext of interaction with their microenvironment remainsunclear. The majority of the in vitro observation originates fromthe use of established multiple myeloma cell lines due to the lackof an adequate coculture model to maintain primary multiplemyeloma cells in vitro . Here, we described a coculture modelinvolving primary human multiple myeloma cells growing on afeeder layer of fetal bone stromal cells in serum-free culturemedium.Remarkably, adult stromal cells were not effective in supporting thegrowth of primary multiple myeloma cells in serum-free condition(data not shown). Using fetal bone stromal cells, we were able tosustain the growth of multiple myeloma cells for at least 3 weeks.This highlights a salient feature in multiple myeloma pathobiology:the interaction of multiple myeloma cells with their permissivemarrow stromal microenvironment is essential for the long-termproliferation and migration of multiple myeloma cells.

We also show that VEGF-A released from fetal bone stromalcells interacts with VEGFR-1 expressed on multiple myelomacells in a paracrine fashion, conferring a sustained proliferativepotential in vitro . This survival advantage is not mediatedthrough viral oncogenesis (EBV), as all primary human multiplemyeloma cells obtained from bone marrow aspirates testednegative for EBV. Neutralizing mAb to VEGFR-1 abrogatedmultiple myeloma cell proliferation through induction ofapoptosis, suggesting that functional interaction of VEGF-A withits cognate receptor is essential for multiple myeloma cellgrowth. VEGF-A secreted by multiple myeloma cells as well asby fetal bone stromal cells acts as an autocrine and paracrinefactor to stimulate multiple myeloma cell growth, in conjunctionwith IL-6 produced by stromal cells. Since IL-6 also enhances theproduction and secretion of VEGF-A by multiple myeloma cells(16, 17), VEGFR-1 blockade seems to be an effective strategy toblock the growth of multiple myeloma cells, overcoming themitogenic effect conferred by IL-6-induced VEGF-A productionby stromal and multiple myeloma cells.The intracellular trafficking and nuclear localization of VEGFR-1

suggests that a VEGF-A/VEGFR-1 intracrine signaling may be

Figure 4. VEGFR-1 blockade prevents multiple myelomacell growth on fetal bone stromal cells. Multiple myelomacells were seeded on a semiconfluent fetal bone stromalcell monolayer in serum-free condition and incubated in thepresence or absence of neutralizing mAb raised againstVEGFR-1, VEGFR-2, or VEGFR-3. The multiple myelomacells were collected every 7 days and viability wasdetermined by trypan blue exclusion. After 21 days,VEGFR-1 blockade suppressed the growth of multiplemyeloma cells in the coculture system (***P < 0.001 ascompared with the number of cells at day 0; jjP < 0.01,jjjP < 0.001 as compared with isotype-treated controlcells; n = 4).

Figure 5. Inhibition of VEGFR-1 impairs nuclear localization of VEGFR-1. The multiple myeloma cells incubated with neutralizing mAb to VEGFR-1 (clone 6.12)were collected following the 21-day coculture with fetal bone stromal cell monolayer and stained for VEGFR-1 (FITC) and DAPI to stain the DNA (nucleus). Cells treatedwith neutralizing mAb showed a clear shift in VEGFR-1 localization, now seen in the cytoplasm and on the cell surface. Results were obtained from four differentprimary multiple myeloma samples. Magnification (�1,000).

Cancer Research




present during multiple myeloma cell proliferation. Increasingevidence has shown that tumor-derived VEGF-A supports expan-sion of tumor vasculature, resulting in tumor growth. Subsets ofacute leukemia cells express functional VEGFR-2 (8, 19, 20), whichsupports extramedullary growth independent of angiogenesis,through an autocrine stimulation by leukemia-derived VEGF-A.As a result, the growth of leukemia is regulated by both paracrineand autocrine loops involving VEGF-A/VEGFR signaling. AutocrineVEGF-A/VEGFR loops are not unique to malignant cells, sinceVEGF-A can also regulate hematopoietic stem cell survival by aninternal autocrine loop mechanism (21). More recently, functionalVEGFR-2 was found to be expressed not only on acute leukemia cellsurfaces, but also predominantly in the nuclei of activated cells (18).These studies support the notion that the pathophysiologic role ofVEGF-A/VEGFR may not be exclusively mediated by the conven-tional paradigm of membrane-bound ligand-receptor interaction. Inthe present study, VEGFR-1 is mainly present in the nuclei ofproliferating human primary multiple myeloma cells, and neutral-izing mAb to VEGFR-1 could suppress the growth of multiplemyeloma cells. In fact, VEGFR-1 must traffic between themembrane for VEGF-A binding and the nucleus of proliferatingmultiple myeloma cells, and VEGFR-1 blockade is effective once thereceptor is present on the cell membrane. This can explain howneutralizing mAb to VEGFR-1 blocks the receptor and suppressestrafficking, providing for the retention of VEGFR-1 in the multiplemyeloma membrane after treatment with mAbs. There are othermembrane tyrosine kinase receptors, which are known to beinternalized after ligand binding, such as the epidermal growthfactor receptor (22–24). VEGFR-2, another cognate tyrosine kinasereceptor of VEGF-A, can interact with surface protein caveolin onendothelial cells and may be internalized upon VEGF-A binding (25,26). However, the mechanisms by which the intracellular traffickingof VEGFR-1 and its subsequent intracrine activation of multiplemyeloma cells in the context of survival, proliferation, andmigration remain undefined. Once inside the nucleus, the potentialactivity of VEGFR-1 may include recruitment of other signalingpartners or functioning as a transcription factor itself. Theidentification of the downstream target genes regulated by theVEGFR-1-mediated paracrine and intracrine loops may providenovel therapeutic targets for multiple myeloma.

Flow cytometric data suggests that subsets of the multiplemyeloma cells have a more robust expression of VEGFR-1. Thehigher level of expression of VEGFR-1 may be dependent on thecell cycle of the proliferating cells or reflect subsets of multiplemyeloma cells at different stages of their maturation. Whethermultiple myeloma cells expressing higher levels of VEGFR-1 aremore responsive to VEGF-A or placental growth factor is notknown and is the subject of future studies.Migration of primary multiple myeloma cells is necessary for the

homing of tumor cells to the bone marrow niche, for expansion ofmalignant plasma cells within the bone marrow microenviron-ment, and for the egress into peripheral blood contributing todisease progression (27–30). Here, we show that neutralizing mAbto VEGFR-1 diminished multiple myeloma cell motility in in vitrotranswell experiments, suggesting that the functional interaction ofVEGF-A with its cognate receptor is essential for multiple myelomahoming and migration. The motility of multiple myeloma cellswas partially blocked by a neutralizing mAb against VEGFR-1,suggesting that the fetal bone stromal cells may secrete othercytokines and/or chemokines to promote multiple myeloma cellmigration independent of VEGF-A/VEGFR-1 signaling. A primeexample is the chemokine stromal cell–derived growth factor-1a,which has been implicated in multiple myeloma cell migrationafter binding to its receptor CXCR4 (31). Therefore, it is logical tospeculate that VEGF-A and stromal cell–derived growth factor-1acould act synergistically to promote multiple myeloma cellmigration.We also show that the adult marrow stromal cells are

significantly less efficient than the human fetal derived cells insupporting multiple myeloma cell growth. The precise mechanismfor this intriguing functional diversity between fetal and adultstromal feeder layers is not known and is the subject of futurestudies. However, it is conceivable that in contrast to the adultstromal cells, fetal stromal cells are epigenetically programmed toproduce proangiogenic factors, and as such, are permissive forsupporting the growth of the multiple myeloma cells.Taken together, the results shown here show the feasibility

of blocking paracrine and intracrine VEGF-A/VEGFR-1 loopson multiple myeloma cells as a means of inducing multi-ple myeloma cell apoptosis. Alone or in combination with

Figure 6. VEGFR-1 activation supports multiple myeloma cellmotility. Multiple myeloma cell motility was assayed using a Boydenchamber system. Multiple myeloma cells were pretreated for4 hours with neutralizing mAb against VEGFR-1, VEGFR-2,VEGFR-3, or an isotype IgG antibody as control. Then, multiplemyeloma cells were plated in the upper chamber of a transwell andthe lower chamber was filled with a serum-free X-VIVO 20 mediumin presence or absence of 50 ng/mL of VEGF-A. A fetal bonestromal cell monolayer that was cultivated in serum-free X-VIVO20 for 2 days was also used as a platform for chemoattractingmultiple myeloma cells. After a 6-hour incubation, cells that hadmigrated in the lower chamber were then harvested and counted.Results are expressed as the number of migrated cells in thedifferent conditions of incubation divided by the number of migratedcells in the control (no cytokine or fetal bone stromal cells). VEGF-Apresent in the lower chamber induced an increase in multiplemyeloma cell motility, which was blocked by neutralizing mAbagainst VEGFR-1 (*P < 0.05, **P < 0.01, ***P < 0.001 as comparedwith unstimulated; jjP < 0.01, jjj P < 0.001 as compared withstimulated cell migration; n = 3).





chemotherapeutic agents, VEGFR-1 blockade may have clinicalrelevance for its therapeutic activity against multiple myeloma.Ongoing studies are in progress to further characterize theregulation of VEGFR-1 trafficking in multiple myeloma cells aswell as profiling of downstream targeted genes in order todelineate distinct intracellular signaling pathways involved inVEGF-A/VEGFR-1–mediated multiple myeloma growth andmigration.

Acknowledgments

Received 10/5/2004; revised 12/11/2004; accepted 1/12/2005.Grant support: Loı̈c Vincent would like to acknowledge La Fondation Bettencourt-

Schueller for its support. Shahin Rafii was supported by a grant from the AmericanCancer Society and the Leukemia and Lymphoma Society and National Institute ofHealth RO1-HL075234.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

References1. Hallek M, Bergsagel PL, Anderson KC. Multiplemyeloma: increasing evidence for a multistep transfor-mation process. Blood 1998;91:3–21.

2. Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D,Anderson KC. Adhesion of human myeloma-derived celllines to bone marrow stromal cells stimulates interleu-kin-6 secretion. Blood 1993;82:3712–20.

3. Anderson KC. Multiple myeloma. Advances in diseasebiology: therapeutic implications. Semin Hematol 2001;38:6–10.

4. Lin B, Podar K, Gupta D, et al. The vascularendothelial growth factor receptor tyrosine kinaseinhibitor PTK787/ZK222584 inhibits growth and mi-gration of multiple myeloma cells in the bone marrowmicroenvironment. Cancer Res 2002;62:5019–26.

5. Podar K, Catley LP, Tai YT, et al. GW654652, thepan-inhibitor of VEGF receptors, blocks the growthand migration of multiple myeloma cells in thebone marrow microenvironment Blood 2004;103:3474–9.

6. Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z.Vascular endothelial growth factor (VEGF) and itsreceptors. FASEB J 1999;13:9–22.

7. Ferrara N. Vascular endothelial growth factor and theregulation of angiogenesis. Recent Prog Horm Res2000;55:15–35.

8. Dias S, Hattori K, Zhu Z, et al. Autocrine stimulationof VEGFR-2 activates human leukemic cell growth andmigration. J Clin Invest 2000;106:511–21.

9. Podar K, Tai YT, Davies FE, et al. Vascular endothelialgrowth factor (VEGF) triggers signaling cascadesmediating multiple myeloma cell growth and migration.Blood 2001;98:428–35.

10. Podar K, Tai YT, Lin BK, et al. Vascular endothelialgrowth factor (VEGF)-induced migration of multiplemyeloma cells is associated with h1-integrin and PI3-kinase-dependent PKCa activation. J Biol Chem 2002;277:7875–81.

11. Di Raimondo F, Azzaro MP, Palumbo G, et al.Angiogenic factors in multiple myeloma: higher levels

in bone marrow than in peripheral blood. Haemato-logica 2000;85:800–5.

12. Munshi NC, Wilson C. Increased bone marrowmicrovessel density in newly diagnosed multiplemyeloma carries a poor prognosis. Semin Oncol 2001;28:565–9.

13. Rajkumar SV, Leong T, Roche PC, et al. Prognosticvalue of bone marrow angiogenesis in multiplemyeloma. Clin Cancer Res 2000;6:3111–6.

14. Ribatti D, Vacca A, Nico B, et al. Bone marrowangiogenesis and mast cell density increase simulta-neously with progression of human multiple myeloma.Br J Cancer 1999;79:451–5.

15. Vacca A, Ribatti D, Presta M, et al. Bone marrowneovascularization, plasma cell angiogenic potential,and matrix metalloproteinase-2 secretion parallelprogression of human multiple myeloma. Blood 1999;93:3064–73.

16. Gupta D, Treon SP, Shima Y, et al. Adherence ofmultiple myeloma cells to bone marrow stromal cellsupregulates vascular endothelial growth factor secre-tion: therapeutic applications. Leukemia (Baltimore)2001;15:1950–61.

17. Dankbar B, Padro T, Leo R, et al. Vascular endothelialgrowth factor and interleukin-6 in paracrine tumor-stromal cell interactions in multiple myeloma. Blood2000;95:2630–6.

18. Santo S, Dias S. Internal and external autocrineVEGF/KDR loops regulate survival of subsets of acuteleukemia through distinct signaling pathways. Blood2004;103:3883–9.

19. Bellamy WT, Richter L, Frutiger Y, Grogan TM.Expression of vascular endothelial growth factor and itsreceptors in hematopoietic malignancies. Cancer Res1999;59:728–33.

20. Gerber HP, Malik AK, Solar GP, et al. VEGFregulates haematopoietic stem cell survival by aninternal autocrine loop mechanism. Nature 2002;417:954–8.

21. Gabrilove JL. Angiogenic growth factors: autocrineand paracrine regulation of survival in hematologicmalignancies. Oncologist 2001;6 Suppl 5:4–7.

22. Wang Y, Pennock S, Chen X, Wang Z. Internalizationof inactive EGF receptor into endosomes and thesubsequent activation of endosome-associated EGFreceptors. Sci STKE 2002;161:PL17.

23. Sorkin A. Internalization of the epidermal growthfactor receptor: role in signaling. Biochem Soc Trans2001;29:480–4.

24. Wells A, Marti U. Signaling shortcuts: cell-surfacereceptors in the nucleus? Nat Rev Mol Cell Biol2002;3:697–702.

25. Feng Y, Venema VJ, Venema RC, Tsai N, Caldwell RB.VEGF induces nuclear translocation of Flk-1/KDR,endothelial nitric oxide synthase, and caveolin-1 invascular endothelial cells. Biochem Biophys Res Com-mun 1999;256:192–7.

26. Labrecque L, Royal I, Surprenant DS, Patterson C,Gingras D, Beliveau R. Regulation of vascular endothe-lial growth factor receptor-2 activity by caveolin-1 andplasma membrane cholesterol. Mol Biol Cell 2003;14:334–47.

27. Hideshima T, Richardson P, Anderson KC. Noveltherapeutic approaches for multiple myeloma. Immu-nol Rev 2003;194:164–76.

28. Masellis-Smith A, Belch AR, Mant MJ, TurleyEA, Pilarski LM. Hyaluronan-dependent motility ofB cells and leukemic plasma cells in blood, butnot of bone marrow plasma cells, in multiplemyeloma: alternate use of receptor for hyaluronan-mediated motility (RHAMM) and CD44. Blood 1996;87:1891–9.

29. Borset M, Hjorth-Hansen H, Seidel C, Sundan A,Waage A. Hepatocyte growth factor and itsreceptor c-met in multiple myeloma. Blood 1996;88:3998–4004.

30. Borset M, Hjertner O, Yaccoby S, Epstein J, Sander-son RD. Syndecan-1 is targeted to the uropods ofpolarized myeloma cells where it promotes adhesionand sequesters heparin-binding proteins. Blood 2000;96:2528–36.

31. Gazitt Y, Akay C. Mobilization of myeloma cellsinvolves SDF-1/CXCR4 signaling and downregulation ofVLA-4. Stem Cells 2004;22:65–73.

Cancer Research




2005;65:3185-3192. Cancer Res Loïc Vincent, David K. Jin, Matthias A. Karajannis, et al. Motility of Human Primary Myeloma CellsFactor Receptor-1 Signaling Promotes Proliferation andEndothelial Growth Factor-A/Vascular Endothelial Growth

Dependent Paracrine and Intracrine Vascular−Fetal Stromal

Updated version

http://cancerres.aacrjournals.org/content/65/8/3185

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/65/8/3185.full.html#ref-list-1

This article cites 30 articles, 17 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/65/8/3185.full.html#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

Subscriptions

Reprints and

[email protected] at

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

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications


http://cancerres.aacrjournals.org/content/65/8/3185

http://cancerres.aacrjournals.org/content/65/8/3185.full.html#ref-list-1

http://cancerres.aacrjournals.org/content/65/8/3185.full.html#related-urls

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

mailto:[email protected]

mailto:[email protected]


Documents

Fetal stromal-dependent paracrine and intracrine vascular endothelial growth factor-a/vascular endothelial growth factor receptor-1 signaling promotes proliferation and motility of