Zoledronic Acid Down-Regulates Adhesion Moleculesof Bone Marrow Stromal Cells in Multiple MyelomaA Possible Mechanism for its Antitumor Effect

Alessandro Corso, M.D.1

Eleonora Ferretti, M.D.1

Monia Lunghi, M.D.1

Patrizia Zappasodi, M.D.1

Silvia Mangiacavalli, M.D.1

Mara De Amici, M.D.2

Chiara Rusconi, M.D.1

Marzia Varettoni, M.D.1

Mario Lazzarino, M.D.1

1 Division of Hematology, IRCCS Policlinico SanMatteo, University of Pavia, Pavia, Italy.

2 Department of Pediatrics, IRCCS Policlinico SanMatteo, University of Pavia, Pavia, Italy.

Address for reprints: Alessandro Corso, M.D., Di-vision of Hematology, IRCCS Policlinico San Mat-teo, Viale Golgi 19, 27100 Pavia, Italy; Fax: (011)39-0382-502250; E-mail: a.corso@smatteo.pv.it

Received November 3, 2004; revision receivedDecember 22, 2004; accepted February 10, 2005.

BACKGROUND. Myeloma plasma cells interact with the bone marrow microenvi-

ronment which, in turn, supports their growth and protects them from apoptosis.

In vitro studies have demonstrated the antitumor potential of zoledronic acid

(ZOL) on myeloma cell lines, but few data are available on its effects on bone

marrow stromal cells (BMSCs). The aim of the current study was to evaluate the

antiproliferative and apoptotic effect of ZOL on BMSCs, as well as its effect on the

expression of adhesion molecules.

METHODS. BMSCs, obtained from bone marrow mononucleated cells of 8 patients

with multiple myeloma, were treated with increasing concentrations of ZOL for 3

days. Cytotoxic effect was analyzed by 3-(4-5-dimethylthiazol-2-yl)-2,5 diphe-

nyltetrazolium bromide; thiazolyl blue (MTT) assay whereas the induction of

apoptosis was evaluated by flow cytometric detection of fluorescein isothiocya-

nate-labeled annexin V, terminal deoxynucleotidyl transferase-mediated dUTP

nick-end labeling assay, and nuclear changes. Moreover, expression of CD106,

CD56, CD50, CD49d, CD44, and CD40 was analyzed by flow cytometry. Data were

evaluated by the Friedman test.

RESULTS. After 3 days of exposure at concentrations of 10�4 to 10�5 M, ZOL

induced a decrease in proliferation (P � 0.0001) and an increase in apoptosis (P

� 0.002). Analysis of culture supernatants showed that myeloma BMSCs expressed

interleukin (IL)-6, negligible levels of tumor necrosis factor-alpha, and no IL-1�. In

vitro exposure to the lowest concentrations of ZOL decreased IL-6 production by

BMSCs. Among the adhesion molecules, CD106, CD54, CD49d, and CD40, which

were strongly expressed at baseline, showed a statistically significant reduction

compared with controls after exposure to ZOL.

CONCLUSIONS. ZOL interfered with myeloma BMSCs by reducing proliferation,

increasing apoptosis, and modifying the pattern of expression of adhesion mole-

cules, especially those involved in plasma cell binding. These effects on BMSCs

might explain the antitumor activity of ZOL. Cancer 2005;104:118 –25.

© 2005 American Cancer Society.

KEYWORDS: bone marrow stromal cells, adhesion molecules, zoledronic acid,antitumor effect, apoptosis.

Multiple myeloma (MM) is characterized by infiltration andgrowth of malignant plasma cells in the bone marrow (BM)

microenvironment, with consequent osteoclast activation and exten-sive osteolysis. Tumor cells localize within the BM through the inter-action of adhesion receptors with their ligands on BM stromal cells(BMSCs) and extracellular matrix (ECM) proteins.1 The adhesion ofmyeloma cells to the BM microenvironment is mediated by severalplasma cell membrane receptors. These play an important role ingrowth regulation and migration of myeloma cells. In particular, the

118

© 2005 American Cancer SocietyDOI 10.1002/cncr.21104Published online 13 May 2005 in Wiley InterScience (www.interscience.wiley.com).

homing of myeloma cells into the BM is mediated byvery late activation antigen (VLA)-1, VLA-4 (CD49d)VLA-5, and lymphocyte function-associated antigen-1(LFA-1). CD44 (Homing-associated cell adhesion mol-ecule [HCAM]) allows the attachment of plasma cellsto laminin and/or fibronectin on the ECM.2,3 MM cellsincrease their adhesion to BMSCs by binding VLA-4and LFA-1 expressed on plasma cells to, respectively,vascular cell adhesion molecule (VCAM)-1 (CD106)and intercellular adhesion molecule (ICAM)-1 (CD54)expressed on BMSCs.4,5

Recent studies have shown the importance of themicroenvironment in the pathophysiology of bonedisease. Adhesive interactions of plasma cells play animportant role in up-regulating the production of cy-tokines and growth factors by BMSCs, which conse-quently enhance tumor growth, bone destruction, andtumor survival.6 Binding of MM cells to the BM stromamaintains the plasma cells in the BM, and stimulatesthe production of osteoclast-activating factors by ei-ther myeloma or stromal cells.7 Among these factors,interleukin (IL)-6 seems primarily involved in my-eloma osteolysis as well as in the growth and survivalof malignant plasma cells. It is secreted by tumorplasma cells triggered via CD40,8 but there is strongevidence of a paracrine secretion by the BM microen-vironment (BMSCs, osteoblasts, osteoclasts).1,7,9,10

The hampering of the adhesion of MM cells to BMSCsby drugs such as thalidomide or proteasome inhibi-tors abrogates protection against apoptosis and blocksthe increased secretion of the cytokines involved inthe growth, survival, and migration of clonal cells.

The major mechanism of bisphosphonates (BPs)in reducing bone resorption is the induction of oste-oclast apoptosis.11–13 However, in vitro studies havedemonstrated a cytostatic and proapoptotic effect ofBPs on myeloma and other human tumor cell lines(e.g., breast carcinoma and prostate carcinoma celllines).14 –19 In addition to the direct antitumor effectson myeloma cells, BPs may also affect the proliferationand survival of these cells by inhibiting the release ofgrowth factors from osteoclasts and BMSCs.

The goals of the current study were to confirm thedirect antiproliferative effect of zoledronic acid (ZOL)on plasma cells and stromal cells, and to evaluate themodulation of the adhesion molecules on BMSCs,particularly those involved in the interaction withplasma cells.

MATERIALS AND METHODSReagentsZOL [zometa:1-hydroxy-2-(1H-imidazole-1-yl)ethyli-dene] was kindly supplied as the hydrated disodiumsalt by Novartis Pharma AG (Basel, Switzerland). Stock

solutions were prepared in phosphate-buffered saline(PBS), pH 7.4, filter sterilized using a 0.2-�m filter, andstored at �20 °C until use.

Cell CulturesThe human myeloma cell line RPMI-8226 was pur-chased from DSMZ (Braunschweig, Germany) andcultured according to the manufacturer’s instructionin RPMI-1640 medium supplemented with 10% fetalbovine serum, 2 mM glutamine, 100 U/mL penicillin,and 100 mg/mL streptomycin. Primary plasma cellswere freshly isolated from three patients. Mononu-clear cells from BM aspirates were isolated using Fi-coll-Hypaque reagent. CD138-positive cells were en-riched by the magnetic cell sorting system (MiltenyiBiotec, Auburn, CA) according to the manufacturer’sinstructions. Only samples with a myeloma cell con-tent � 95% were cultured in RPMI-1640 supplementedwith 20% fetal calf serum, 100 U/mL penicillin, 100U/mL streptomycin, 1 mM sodium pyruvate, 2 mMglutamine, and 2 ng/mL IL-6 (Peprotech, London, En-gland).20

BM mononucleated cells from eight patients withMM were separated by Ficoll-Hypaque density gradi-ent centrifugation. Adherent cells were long-term cul-tured and expanded in MyeloCult H5100 (StemCellTechnologies, Vancouver, British Columbia, Canada),10�6 M hydrocortisone sodium succinate, 100 U/mLpenicillin, and 100 mg/mL streptomycin at 37 °C with5% CO2. Adherent cell monolayers were obtained usu-ally after 2–3 weeks.21

Proliferation and Cytotoxicity AssaysRPMI-8226 cells were seeded in 96-well plates, each 1containing 104 cells/0.1 mL of medium. Cells werecultured for 72 hours in medium alone (control) ormedium containing ZOL (10�5 M, 5 � 10�5 M, 10�4 M,5 � 10�4 M), recombinant human IL-6 (5 ng/mL), or acombination thereof. The effect of ZOL on RPMI cellproliferation was evaluated by 3H-thymidine assay.Cytotoxicity was evaluated by 3-(4-5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide; thiazolyl blue(MTT) assay as previously described.22 All assays wereperformed in triplicate and at least three times.

Adherent BMSCs (6 � 104 cells/mL) were seededin 24-well plates with ZOL at concentrations of 10�5, 5� 10�5, and 10�4 M for 3 days. Cells were incubatedwith 0.5 mg/mL of tetrazolium compound for 3 hoursat 37 °C and 5% CO2 conditions. Finally, the MTT-containing medium was removed and the purpleformazan crystals were dissolved with dimethylsulfox-ide. The plates were gently stirred until the color re-action was uniform and 200 �L of solution from eachwell was transferred to 96-well plates.

Zoledronic Acid and Adhesion Molecules/Corso et al. 119

Dye absorbance of RPMI-8226 cells and BMSCswas spectrophotometrically measured at 550 nm us-ing a microplate reader.

Annexin V Flow Cytometric Analysis of ApoptosisTo asses the effect of ZOL on apoptosis, BMSCs andprimary plasma cells were analyzed by flow cytometricdetection of fluorescein isothiocyanate (FITC)-labeledannexin V. Briefly, adherent cells of primary culturewere plated at final concentrations of 3 � 105 cells.After 3 days of exposure to ZOL (10�5 M or 10�4 M),cells were detached with trypsin/ethylenediaminetet-raacetic acid solution. Then, cells were collected andwashed with PBS. In primary plasma cells, the pres-ence of apoptotic death was evaluated daily for 1– 4days of exposure to ZOL at a concentration of 10�5 Mor 10�4 M. Myeloma BMSCs and plasma cells (approx-imately 1 � 105 cells per sample) were stained withannexin V-FITC (Molecular Probes, Poortgebouw, TheNetherlands) for 15 minutes at room temperature inthe dark. To better detect BM plasma cells, 10 �L ofCD38-TC and CD138-PE antibodies (Becton Dickin-son, San Jose, CA) was tested in combination withannexin V-FITC to gate BM plasma cells on the basisof the expression of CD38 and CD138 and side scatter(SSC) laser light diffraction. For each sample, an au-tocontrol (sham buffer) was processed. Flow cytomet-ric analysis was performed with a FACSCalibur flowcytometer (Becton Dickinson) and 5000 –10,000 eventswere acquired for each sample. Data were collectedand analyzed with CellQuest 3.1 software (BectonDickinson).

In Situ Apoptosis DetectionDNA fragmentation was detected with the fluorescein-based TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling) assay (In Situ CellDetection Kit, Fluorescein, Roche, Basel, Switzerland).

Stromal cells were seeded at a density of 2 � 104

cells in chamber slides (Nunc, Roskilde, Denmark).Cells were cultured for 24 hours. Then, 5 � 10�5M ofZOL was added.19 After 3 days of exposure, the cellmonolayer was fixed with 4% paraformaldehyde andpermeabilized with 0.1% Triton X-100 solution in PBSfor 5 minutes. Finally, 50 �L TUNEL reaction mixture(terminal deoxynucleotidyl transferase and nucleotidemixture in reaction buffer) was added to each cham-ber. A negative and positive control was included ineach experiment. To evaluate the cytoskeleton, cellswere stained for 20 minutes with rhodamine-labeledphalloidin (Molecular Probes), which selectively bindsto F-actin. After further rinsing with PBS, 1 �g/mL ofHoechst 33342 (Molecular Probes) was added to de-tect the chromatin condensation. After 15 minutes,

the slides were washed, mounted in Gel Mount(Sigma, St. Louis, MO), and immediately evaluatedusing a universal microscope (Axioskop, Zeiss, Ger-many) with UV, FITC, and rhodamine illuminationequipped with an MC100 automatic photomicro-graphic camera. Apoptotic cells were defined on thebasis of characteristic changes in the nuclear mor-phology and for the their positive reaction to theTUNEL assay.

Determination of Interleukin-6, Interleukin-1�, andTumor Necrosis Factor-Alpha ConcentrationsCytokine concentrations in BMSC culture superna-tants were examined by enzyme-linked immunosor-bent assays with commercially available kits (R&DSystems, Minneapolis, MN) and performed as previ-ously described.23 Briefly, 6 � 104 BMSCs were seededin MyeloCult 5100 for 24 hours in the presence ofdifferent concentrations of ZOL (10�11 to 10�6 M).Then, the culture medium was completely replacedwith �-minimum essential medium (stem cell, supple-mented with 2% FBS (stem cell), 2 mM glutamine, 100U/mL penicillin, 100 U/mL streptomycin, and the dif-ferent concentrations of ZOL. After an incubation of48 hours, the supernatants were collected and frozenat � 80 °C until analysis.

Adhesion Molecule Expression on Bone Marrow StromalCellsBMSCs (3 � 105) were cultured in a T25 flask in me-dium alone or in the presence of ZOL (10�5 M or10�4M) for 3 days. After enzymatic digestion, BMSCswere suspended in medium culture and washed withPBS.

BMSC samples (approximately 1 � 105 cells persample) were incubated (10 minutes at room temper-ature, lyse no wash technique) in the presence of 10�L of each monoclonal antibody (MoAb), according tothe recommendations of the manufacturer. For theflow cytometric analysis of adhesion molecule expres-sion, the following MoAbs were tested: CD40-phyco-erythrin (PE), CD44-FITC (HCAM), CD49d-PE (VLA-4),CD50-PE (ICAM-3), CD54-PE (ICAM-1), and CD106-PE(VCAM-1) (Ancell, Bayport, MN). After adding 2 mL pertube of PBS, 4-parameter, 2-color flow cytometry anal-ysis was performed with a FACSCalibur flow cytome-ter (Becton Dickinson). Approximately 10,000 eventswere acquired for each sample. For each sample, asham buffer (negative control) and an isotypic controlreagent were stained to analyze flow cytometric data.Data were collected and analyzed with the CellQuest3.1 program (Becton Dickinson). Immunophenotypicdata were expressed as the percentage of cells positivefor � 1 antigen.

120 CANCER July 1, 2005 / Volume 104 / Number 1

Statistical AnalysisData were evaluated by the Friedman test and a non-parametric analysis of variance, in view of the smallsample size. Moreover, the Wilcoxon test was used tocompare untreated and treated cells. The STATISTICAcomputer program was used. (Stat Soft, Inc., Tulsa,OK). Significance was set at P � 0.05.

RESULTSZoledronic Acid Induces Cytotoxicity to RPMI-8226 CellsDue to the few primary plasma cells selected fromeach patient, it was not possible to evaluate simulta-neously the cytotoxic and apoptotic effect of ZOL, and,therefore, the cytotoxicity was studied only on RPMI-8226 cells. Proliferation and cytotoxicity were ana-lyzed using 3H-thymidine and the MTT assay. As ex-pected, exposure for 72 hours with 10�5, 5 � 10�5,10�4, and 5 � 10�4 M ZOL induced a significant pro-gressive inhibition of the proliferation (P � 0.02) andan increase of cytotoxicity of the MM cell line (P� 0.007) (Fig. 1).

To ensure that the decrease of MM cell growthwas not related to a reduction in the production ofIL-6, 5 ng/mL recombinat human IL-6 was added tothe cultures. Our results showed that the addition ofthis cytokine was unable to reverse the antiprolifera-tive effect of ZOL.

Zoledronic Acid Induces Apoptosis to Multiple MyelomaCells Isolated from PatientsTumor plasma cells were isolated from BM aspiratesof 3 patients with myeloma and cultured with 10�5 M

or 10�4 M of ZOL for 4 days. All 3 fresh MM cellsuspensions contained � 95% CD138-positive plasmacells as detected by flow cytometry analysis. The per-centage of apoptotic cells was evaluated daily by flowcytometry detection of fluorescein-labeled annexin V,which recognized inverted phosphatidylserine on theexterior part of the plasma membrane as an early-stage apoptotic marker. As shown in Figure 2, ZOLapparently induced dose-dependent and time-depen-dent apoptosis, but the statistical analysis showed thatthe apoptotic effect reached statistical significance (P� 0.05) only at the highest concentration (10�4 M).

Zoledronic Acid Decreases Cell Viability in Bone MarrowStromal Cells by Apoptotic Cell DeathThe effect of ZOL on BMSC viability was analyzed in10 samples of BMSCs derived from patients with MMusing the MTT cytotoxic assay, in the presence orabsence of the drug. After a 96-hour incubation, cellsurvival was determined by the ability of viable cells toreduce the MTT dye to formazan. ZOL induced a sig-nificant (P � 0.0002) concentration-dependent celldeath in adherent cells when exposed to 10�5, 5� 10�5, and 10�4 M of drug for 3 days (Fig. 3). Thecytotoxic effect becomes more evident after exposureto a concentration � 10�5 M.

To determine if the reduction of cell viability ob-served with ZOL treatment was due to apoptoticdeath, BMSCs were analyzed by cytometric detectionof annexin V. After 3 days of exposure, BMSCs showeda significant (P � 0.001) dose-dependent increase inthe percentage of apoptotic cells, which was greaterthan that observed with the plasma cells at baseline(Fig. 3). Apoptosis was also evaluated through thetypical nuclear changes, including chromatin conden-sation (Hoechst staining) and DNA fragmentation de-tected by the TUNEL assay. ZOL at the dose of 5� 10�5 M for 3 days caused these nuclear changes in

FIGURE 1. Cytoreductive effects of zoledronic acid (ZOL) on RPMI-8226 cells.

Tumor cells were cultured for 72 hours with different concentrations of ZOL and

then assayed by incorporation of 3H-thymidine to determine percentage of

proliferation or by 3-(4-5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bro-

mide; thiazolyl blue (MTT) assay to determine the percentage of cytotoxicity

with respect to untreated control cells. The LD50 value calculated for the MTT

assay was 62 �M. Results are the mean � the standard deviation of three

independent experiments. Filled diamonds: cytotoxicity; open squares: prolif-

eration.

FIGURE 2. Dose-dependent apoptotic effect of zoledronic acid on primary

plasma cells at concentrations of 10 �M and 100 �M for 4 days. Data are

expressed as the mean of 3 independent experiments (P � 0.04). Filled

diamonds: control; filled triangles: 10 �M; filled squares: 100 �M.

Zoledronic Acid and Adhesion Molecules/Corso et al. 121

BMSCs. Beyond the morphologic changes of BMSCs,the exposure to ZOL provoked the loss of cytoskeletalintegrity, characterized by cell rounding and loss ofstress fibers (Fig. 4).

Effect of Zoledronate on Cytokine Secretion of BoneMarrow Stromal CellsTo determinate whether small concentrations of ZOL(10�11 to 10�6 M) could decrease cytokine secretion,IL-6, IL-1�, and tumor necrosis factor-alpha (TNF-�)were investigated in the supernatants of 8 differentsamples of BMSCs. IL-6 secretion in media alone was210 � 148 (mean � standard deviation) pg/mL,whereas IL-1� and TNF-� were undetectable apartfrom 2 samples in which the TNF-� concentration was23 � 6 pg/mL. After exposure to 10�6 to 10�7 M ofZOL for 3 days, IL-6 secretion was significantly de-creased (P � 0.05) with respect to untreated controls.No significant effect was observed for concentrations� 10�8 M.

Effect of Zoledronic Acid on Molecular AdhesionExpression in Bone Marrow Stromal CellsThe BMSCs of eight patients with MM were analyzedto assess the capacity of ZOL to modulate their expres-sion of the adhesion molecules. After exposure to thehighest concentration of ZOL (10�4 M), a statisticallysignificant down-regulation of CD40, CD49d, CD54,and CD106 was observed with respect to the baselinecondition (P � 0.05). In contrast, no statistical differ-ences were observed after exposure to the lower con-centration of ZOL (10�5 M) neither between ZOL 10�5

M and ZOL 10�4 M concentrations. As well, no statis-

tically significant variations were observed regardingthe expression of CD44 and CD50 (Fig. 5).

DISCUSSIONThe myeloma microenvironment is characterized bythe presence of clonal plasma cells, stromal cells (e.g.,osteoclasts, osteoblasts, fibroblasts, and endothelialcells), ECM protein, and several growth factors. It hasbeen shown that BMSCs support MM cell growth andsurvival, protect them from apoptosis, enhance cyto-kine production, and induce development and activa-tion of osteoclasts. All these events are primed by theinteraction between MM and stromal cells through theexpression of cell adhesion molecules. Several studieshave demonstrated that the abrogation of this contactinhibits tumor cell growth and suppresses osteoclasticactivation.24

BPs are a class of nonhydrolyzable analogs of py-rophosphate which have shown, besides the strongantiosteoclastic activity, a direct antitumor effect onseveral cancer cells, including myeloma cells.

In the current study, we investigated the antitu-mor effect of ZOL with particular attention to themodulation of adhesion molecules expressed by BM-SCs. As previously demonstrated, we observed anti-proliferative and apoptotic effects on myeloma celllines, primary plasma cells, stromal cells, and a reduc-tion in IL-6 production.

These data are in accordance with the majority ofin vitro studies on myeloma cell lines. In our study,however, in contrast to that of Derenne et al.,23 whoshowed antiproliferative and proapoptotic effects ofZOL on BMSCs only at the highest concentration, weobserved that at all concentrations (10�5 and 10�4 M)had a dose-dependent pattern.

Apoptosis induced by ZOL was associated withDNA fragmentation and morphologic changes in BM-SCs (Fig. 4). These peculiar features have been shownto be related to the BP down-regulation of bcl-2,which, in turn, causes a release of cytochrome C frommitochondria and caspase-3 activation.25 Other au-thors have also hypothesized that nitrogen-containing(N)-BPs mediate this effect through the inhibition ofthe mevalonate pathway,26 –28 which purports theblock of the synthesis of isoprenoid compounds, in-cluding Ras, Rho, and Rac, involved in important cel-lular pathways.

To evaluate if the apoptosis of the BMSCs inducedby ZOL was correlated to an inhibition of cell adhe-sion, we studied the expression of the following adhe-sion molecules on BMSCs: CD40, CD44 (HCAM),CD49d (VLA-4), CD50 (ICAM-3), CD54 (ICAM-1), andCD106 (VCAM-1).

We found that ZOL down-regulates the expression

FIGURE 3. Apoptotic effect versus cell viability in bone marrow stromal cells

(BMSCs). BMSCs were exposed to increasing concentrations of zoledronic acid

for 3 days and then analyzed by annexin V assay to determine the percentage

of apoptotic cells and by 3-(4-5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium

bromide; thiazolyl blue (MTT) assay to determine the loss of cell viability. The

two effects were significant at all the concentrations tested. Data are expressed

as the mean of eight independent experiments. Filled diamonds: annexin V;

filled squares: MTT assay.

122 CANCER July 1, 2005 / Volume 104 / Number 1

FIGURE 4. Morphologic changes in bone marrow stromal cells induced

by zoledronic acid (ZOL) after 3 days of treatment. Analysis was performed

by fluorescent microscopy. Nuclei were stained with Hoechst 33342 and

cytoskeleton with rhodamine-labeled phalloidin. The micrographs were

obtained after double exposure using an MC100 automatic photomicro-

graphic camera. (A) Untreated cells. (B, D, E) Cells were treated with 50

�M of ZOL. In contrast to normal cells, the nuclei of apoptotic cells (arrow)

have highly condensed chromatin and show a typical change in their

morphology (B–D). (E) After the terminal deoxynucleotidyl transferase-

mediated dUTP nick-end labeling assay, only the nuclei of apoptotic cells

were labeled with fluorescein (green). Original magnification � 100 (A, B);

� 400 (C–E).

Zoledronic Acid and Adhesion Molecules/Corso et al. 123

of some of these molecules, in particular those in-volved in the cell-to-cell contact with plasma cells. Tothe best of our knowledge, no other studies have fo-cused on this effect of ZOL. We observed that theexposure to the highest concentration of ZOL reducedsignificantly the expression of CD40, CD49d, CD54,and CD106. The strong modulation of integrin CD49d(P � 0.005) may indicate a possible effect of ZOL oncell anchorage. Conversely, the down-regulation ofCD54 (ICAM-1) and CD106 (VCAM-1), which mediatethe cell-to-cell-contact with plasma cells, can affecttumor cell growth, survival, and local secretion of IL-6.Actually, the evaluation of IL-6 levels in the superna-tant of BMSC long-term culture after exposure to ZOL,even at very low concentrations, produced a signifi-cant down-regulation of IL-6 production. This obser-vation could be consistent with the central role of IL-6even in the modulation of adhesion molecules. Thishypothesis is also supported by recent evidence sup-porting the reduced expression of VCAM-1 in an IL-6 – deficient stroma with respect to wild-type stroma.29

Moreover, other authors have demonstrated that thecontact of mouse myeloma plasma cells with BMSCsvia �4�1 integrin/VCAM-1 interactions enhances theproduction of osteoclastogenesis factors and that dis-ruption of this cell-to-cell contact reduces their pro-duction24 and can suppress the development of MMand associated osteoclastic osteolysis.30

In conclusion, ZOL shows in vitro antitumor ac-tivity on myeloma BMSCs and plasma cells, inducingapoptosis and reducing cell proliferation. It interferesalso with the BMSC down-regulation of cytokine se-cretion and modifies the expression of adhesion mol-ecules, in particular those involved in cell-to-cell-con-tact. The modulation of the adhesion molecules could

represent the main pathogenetic mechanism of theantitumor activity of ZOL.

REFERENCES1. Vidriales MB, Anderson KC. Adhesion of multiple myeloma

cells to the bone marrow microenvironment: implicationsfor future therapeutic strategies. Mol Med Today. 1996;2:425– 431.

2. Uchiyama H, Barut BA, Mohrbacher AF, Chauhan D, Ander-son KC. Adhesion of human myeloma-derived cell lines tobone marrow stromal cells stimulates interleukin-6 secre-tion. Blood. 1993;82:3712–3720.

3. Ahsmann EJ, Lokhorst HM, Dekker AW, Bloem AC. Lympho-cyte function-associated antigen-1 expression on plasmacells correlates with tumor growth in multiple myeloma.Blood. 1992;79:2068 –2075.

4. Kawano MM, Huang N, Tanaka H, et al. Homotypic cellaggregations of human myeloma cells with ICAM-1 andLFA-1 molecules. Br J Haematol. 1991;79:583–588.

5. Sanz-Rodriguez F, Ruiz-Velasco N, Pascual-Salcedo D,Teixido J. Characterization of VLA-4-dependent myelomacell adhesion to fibronectin and VCAM-1. Br J Haematol.1999;107:825– 834.

6. Roodman GD. Role of stromal-derived cytokines and growthfactors in bone metastasis. Cancer. 2003;97(3 Suppl):733–738.

7. Chauhan D, Uchiyama H, Akbarali Y, et al. Multiple my-eloma cell adhesion-induced interleukin-6 expression inbone marrow stromal cells involves activation of NF-kappaB. Blood. 1996;87:1104 –1112.

8. Urashima M, Chauhan D, Uchiyama H, Freeman GJ, Ander-son KC. CD40 ligand triggered interleukin-6 secretion inmultiple myeloma. Blood. 1995;85:1903–1912.

9. Bataille R, Chappard D, Marcelli C, et al. Recruitment of newosteoblasts and osteoclasts is the earliest critical event in thepathogenesis of human multiple myeloma. J Clin Invest.1991;88:62– 66.

10. Caligaris-Cappio F, Bergui L, Gregoretti MG, et al. Role ofbone marrow stromal cells in the growth of human multiplemyeloma. Blood. 1991;77:2688 –2693.

11. Hughes DE, Wright KR, Uy HL, et al. Bisphosphonates pro-mote apoptosis in murine osteoclasts in vitro and in vivo.J Bone Miner Res. 1995;10:1478 –1487.

12. Selander KS, Monkkonen J, Karhukorpi EK, Harkonen P,Hannuniemi R, Vaananen HK. Characteristics of clodr-onate-induced apoptosis in osteoclasts and macrophages.Mol Pharmacol. 1996;50:1127–1138.

13. Hiroi-Furuya E, Kameda T, Hiura K, et al. Etidronate (EHDP)inhibits osteoclastic-bone resorption, promotes apoptosisand disrupts actin rings in isolate-mature osteoclasts. CalcifTissue Int. 1999;64:219 –223.

14. Fromigue O, Lagneaux L, Body JJ. Bisphosphonates inducebreast cancer cell death in vitro. J Bone Miner Res. 2000;15:2211–2221.

15. Jagdev SP, Coleman RE, Shipman CM, Rostami HA,Croucher PI. The bisphosphonate, zoledronic acid, inducesapoptosis of breast cancer cells: evidence for synergy withpaclitaxel. Br J Cancer. 2001;84:1126 –1134.

16. Lee MV, Fong EM, Singer FR, Guenette RS. Bisphosphonatetreatment inhibits the growth of prostate cancer cells. Can-cer Res. 2001;61:2602–2608.

FIGURE 5. Effect of zoledronic acid (ZOL) on adhesion molecule expression

of bone marrow stromal cells after exposure for 3 days to different concen-

trations of ZOL. Data are the means of eight independent experiments � the

standard deviation. The expression of CD40, CD49d, CD54, and CD106 was

significantly reduced after treatment with the highest dosage of ZOL. Bars with

heavy stippling: control; bars with light stippling: 10 �M; bars with stripes: 100

�M.

124 CANCER July 1, 2005 / Volume 104 / Number 1

17. Tassone P, Forciniti S, Galea E, et al. Growth inhibition andsynergistic induction of apoptosis by zoledronate and dexa-methasone in human myeloma cell lines. Leukemia. 2000;14:841– 844.

18. Shipman CM, Rogers MJ, Apperley JF, Russell RG, CroucherPI. Bisphosphonates induce apoptosis in human myelomacell lines: a novel anti-tumour activity. Br J Haematol. 1997;98:665– 672.

19. Senaratne SG, Pirianov G, Mansi JL, Arnett TR, Colston KW.Bisphosphonates induce apoptosis in human breast cancercell lines. Br J Cancer. 2000;82:1459 –1468.

20. Chatterjee M, Honemann D, Lentzsch S, et al. In the pres-ence of bone marrow stromal cells human multiple my-eloma cells become independent of the IL-6/gp130/STAT3pathway. Blood. 2002;100:3311–3318.

21. Gartner S, Kaplan HS. Long-term culture of human bonemarrow cells. Proc Natl Acad Sci USA. 1980;77:4756 – 4759.

22. Twentyman PR, Luscombe M. A study of some variables ina tetrazolium dye (MTT) based assay for cell growth andchemosensitivity. Br J Cancer. 1987;56:279 –285.

23. Derenne S, Amiot M, Barille S, et al. Zoledronate is a potentinhibitor of myeloma cell growth and secretion of IL-6 andMMP-1 by the tumoral environment. J Bone Miner Res.1999;14:2048 –2056.

24. Michigami T, Shimizu N, Williams PJ, et al. Cell-cell contactbetween marrow stromal cells and myeloma cells viaVCAM-1 and alpha(4)beta(1)-integrin enhances productionof osteoclast-stimulating activity. Blood. 2000;96:1953–1960.

25. Cohen GM. Caspases: the executioners of apoptosis. Bio-chem J. 1997;326(Pt 1):1–16.

26. van Beek E, Pieterman E, Cohen L, Lowik C, Papapoulos S.Farnesyl pyrophosphate synthase is the molecular target ofnitrogen-containing bisphosphonates. Biochem Biophys ResCommun. 1999;264:108 –111.

27. Bergstrom JD, Bostedor RG, Masarachia PJ, Reszka AA, Ro-dan G. Alendronate is a specific, nanomolar inhibitor offarnesyl diphosphate synthase. Arch Biochem Biophys. 2000;373:231–241.

28. Coxon FP, Helfrich MH, Van’t Hof R, et al. Protein gera-nylgeranylation is required for osteoclast formation, func-tion, and survival: inhibition by bisphosphonates and GGTI-298. J Bone Miner Res. 2000;15:1467–1476.

29. Rodriguez Mdel C, Bernad A, Aracil M. Interleukin-6 deficiencyaffects bone marrow stromal precursors, resulting in defectivehematopoietic support. Blood. 2004;103:3349–3354.

30. Mori Y, Shimizu N, Dallas M, et al. Anti-alpha4 integrin anti-body suppresses the development of multiple myeloma andassociated osteoclastic osteolysis. Blood. 2004;104:2149–2154.

Zoledronic Acid and Adhesion Molecules/Corso et al. 125