Delivery of c-myb Antisense Oligodeoxynucleotides to Human Neuroblastoma Cells Via...

Delivery of c-myb AntisenseOligodeoxynucleotides toHuman Neuroblastoma CellsVia Disialoganglioside GD2-Targeted Immunoliposomes:Antitumor Effects

Gabriella Pagnan, Darrin D. Stuart,Fabio Pastorino, Lizzia Raffaghello,Paolo G. Montaldo, TheresaM. Allen, Bruno Calabretta,Mirco Ponzoni

Background: Advanced-stage neuro-blastoma resists conventional treat-ment; hence, novel therapeutic ap-proaches are required. We evaluatedthe use of c-myb antisense oligodeoxy-nucleotides (asODNs) delivered to cellsvia targeted immunoliposomes to in-hibit c-Myb protein expression andneuroblastoma cell proliferation invitro. Methods: PhosphorothioateasODNs and control sequences wereencapsulated in cationic lipid, and theresulting particles were coated withneutral lipids to produce coated cat-ionic liposomes (CCLs). Monoclonalantibodies directed against the disialo-ganglioside GD2 were covalentlycoupled to the CCLs. 3H-labeled lipo-somes were used to measure cellularbinding, and cellular uptake of asODNswas evaluated by dot-blot analysis.Growth inhibition was quantified bycounting trypan blue dye-stained cells.Expression of c-Myb protein was exam-ined by western blot analysis.Results:Our methods produced GD2-targetedliposomes that stably entrapped 80%–90% of added c-myb asODNs. Theseliposomes showed concentration-dependent binding to GD2-positiveneuroblastoma cells that could beblocked by soluble anti-GD2 monoclo-nal antibodies. GD2-targeted liposomesincreased the uptake of asODNs byneuroblastoma cells by a factor of four-fold to 10-fold over that obtained withfree asODNs. Neuroblastoma cell pro-liferation was inhibited to a greaterextent by GD2-targeted liposomes con-taining c-myb asODNs than bynontar-geted liposomes or free asODNs.GD2-targeted liposomes containingc-myb asODNs specifically reduced ex-pression of c-Myb protein by neuro-

blastoma cells. Enhanced liposomebinding and asODN uptake, as well asthe antiproliferative effect, were notevident in GD2-negative cells.Conclu-sions: Encapsulation of asODNs intoimmunoliposomes appears to enhancetheir toxicity toward targeted cellswhile shielding nontargeted cells fromantisense effects and may be efficaciousfor the delivery of drugs with broadtherapeutic applications to tumor cells.[J Natl Cancer Inst 2000;92:253–61]

Advanced-stage neuroblastoma is re-fractory to conventional treatments suchas radiation therapy and chemotherapy(1,2).Thus, novel therapeutic approachesare required. The identification of acti-vated oncogenes and inactivated tumorsuppressor genes as fundamental geneticdifferences between malignant cells andnormal cells has made it possible to con-sider such genes as targets for antitumortherapy. The c-Myb proto-oncogene is thebest characterized member of a family oftranscription factor genes. Its proteinproduct plays a fundamental role in theproliferation of normal and leukemic cells(3–5), and c-Myb gene expression hasbeen reported in several solid tumors ofdifferent embryonic origins(6,7), includ-ing neuroblastoma, where it is linked tocell proliferation and differentiation(8,9).

In recent years, antisense oligodeoxy-nucleotides (asODNs) have shown effi-cacy in the selective inhibition of geneexpression(10–13),but therapeutic appli-cations of asODNs are limited by theirlow physiologic stability, unfavorablepharmacokinetics, low cellular uptake,and lack of tissue specificity. Instabilityhas been largely overcome by employingbackbone-modified oligodeoxynucleo-tides, such as phosphorothioate oligode-oxynucleotides(14). These analogues aremore resistant to nucleases(10,14,15),buttheir animal pharmacokinetics and cyto-toxicity appear relatively non-sequencespecific. For example, the same morpho-logic and functional abnormalities(mainly involving the spleen, liver, andkidneys) were observed in animals treatedwith several oligodeoxynucleotides ofdifferent sequences, likely as a result oftheir accumulation in these organs(16–18).

The problem of low cellular uptake ofasODNs has also been difficult to over-come. Endogenous uptake pathways haveinsufficient capacity to deliver the quan-tities of asODNs required to suppress

gene expression(14,19,20). Formingcomplexes between asODNs and cationicliposomes(21,22)or polylysine(23) hasenhanced intracellular delivery, but suchcomplexes have disadvantages forin vivoapplications. Cationic lipid complexes,for example, are rapidly cleared by thereticuloendothelial system and can benonspecifically cytotoxic(24).Liposomessterically stabilized with polyethyleneglycol derivatives have circulation half-lives of approximately 12 hours and goodstability in the presence of plasma, andthey have recently been shown to facili-tate delivery of asODNsin vivo (25,26).Sterically stabilized immunoliposomes,which have cell surface-directed antibod-ies on their exteriors, have been recog-nized as efficient tools for delivery ofdrugs and diagnostic agents to specifickinds of cells(27).

Among the antigens found on malig-nant cells, the disialoganglioside GD2 isan attractive target for immunoliposomaltherapy of tumors of neuroectodermal ori-gin (28,29), since it is extensively ex-pressed in these tumors(30,31)but is lesscommon in nonmalignant tissue. We havepreviously used GD2-targeted, stericallystabilized immunoliposomes to deliverthe apoptosis-inducing drug fenretinide tomelanoma cellsin vitro (32). In thisstudy, we investigated whether GD2-targeted, sterically stabilized liposomescontaining c-myb asODNs selectively de-livered to neuroblastoma cells sufficientasODNs to reduce c-Myb protein expres-sion and to inhibit cell growth.

MATERIALS AND METHODS

Chemicals and Monoclonal Antibodies

Hydrogenated soy phosphatidylcholine (HSPC),cholesterol, 1,2-distearoylglycero-3-phosphatidyle-thanolamine-N-polyethylene glycol-2000 (DSPE-PEG), and 1,2-dioleoyl-3-trimethylammonium pro-

Affiliations of authors:G. Pagnan, F. Pastorino,L. Raffaghello, P. G. Montaldo, M. Ponzoni, Labo-ratory of Oncology, G. Gaslini Children’s Hospital,Genoa, Italy; D. D. Stuart, T. M. Allen, Departmentof Pharmacology, University of Alberta, Edmonton,Canada; B. Calabretta, Kimmel Cancer Center,Thomas Jefferson University, Philadelphia, PA.

Correspondence to:Mirco Ponzoni, Ph.D., Labo-ratory of Oncology, G. Gaslini Children’s Hospital,Largo G. Gaslini 5, 16148, Genoa, Italy (e-mail:mircoponzoni@ospedale-gaslini.ge.it).

See“Notes” following “References.”

© Oxford University Press

Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000 REPORTS 253

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

pane (DOTAP) were from Avanti Polar Lipids, Inc.(Alabaster, AL). A derivative of DSPE-PEG with amaleimide group at the distal terminus of the poly-ethylene glycol chain (DSPE-PEG-MAL) was syn-thesized by Shearwater Polymers (Huntsville, AL).[3H]Cholesterol hexadecylether ([3H]CHE) wasfrom Du Pont NEN (Boston, MA). All the otherreagents of biochemical and molecular biologygrade were obtained from Sigma Chemical Co. (St.Louis, MO).

Two different monoclonal antibodies (MAbs) di-rected against disialoganglioside GD2 (anti-GD2s)were employed. In the first set of experiments, weused 14.G2a, a murine MAb of immunoglobulin (Ig)G2a isotype subclass, specific for the GD2 antigen,provided by R. A. Reisfeld (The Scripps Institute,La Jolla, CA)(33). To reduce the possibility of ad-verse reactions to murine antibodies, we used, in asecond set of experiments, the chimeric human–mouse MAb ch14.18. This antibody was constructedby combining constant regions of human IgG mol-ecules with variable regions of a GD2-specific mu-rine MAb (29) and was provided by R. Hand-gretinger (University of Tubingen, Tubingen,Germany).

Oligodeoxynucleotides

Three 24-mer phosphorothioate oligodeoxy-nucleotides were prepared, purified, and dried undervacuum by Lynx Therapeutics, Inc. (Hayward, CA).Our asODN, which is referred to hereafter as myb-as, is complementary to codons 2–9 of human c-Myb messenger RNA (mRNA) and has the sequence58-TATGCTGTGCCCGGGTCTTCGGGC-38. It iscurrently being used in phase II clinical studies,where it is known as LR-3001(34) . A sense oligo-deoxynucleotide (myb-s) (58-GCCCGAAGACCC-GGGCACAGCATA-38) and a scrambled sequence(58-TCGCGGATGTCCGGGTCTGTCGCT-38) oli-godeoxynucleotide were used as controls.

Liposome Preparation andOligodeoxynucleotide Encapsulation

A 700-mg portion of a given oligodeoxynucleo-tide with a trace of125I-labeled oligodeoxynucleo-tide (to allow us to estimate the percentage of oli-godeoxynucleotide present at various stages of theprocedure) was dissolved in 0.25 mL distilled deion-ized water. Next, 0.51 mL methanol and 0.25 mLCHCl3 containing 2.17mmol DOTAP were added,and the mixture was gently vortex mixed to form amonophase. After a 30-minute incubation at roomtemperature, 0.25 mL distilled deionized water and0.25 mL CHCl3 were added. After mixing and cen-trifugation (800g for 7 minutes at room tempera-ture), the aqueous methanol layer was removed. Un-der these conditions, which were developed in pilotstudies using 10mg oligodeoxynucleotide and vari-ous amounts of DOTAP, 90%–95% of the oligode-oxynucleotide was recovered in the organic phase.The resulting oligodeoxynucleotide-to-lipid mole ra-tio is 1 : 24, and a positive-to-negative charge ratioof 1 : 1 is obtained.

Following the extraction, 4.3mmol cholesteroland appropriate amounts of HSPC, DSPE-PEG, orDSPE-PEG-MAL to give cholesterol-to-phospho-lipid mole ratios of 1 : 2 were added to the organicphase. Next, 0.2 mL distilled deionized water wasadded, and the mixture was vortex mixed and then

emulsified by sonication for 2 minutes. The organicphase was evaporated with the use of a rotary evapo-rator (Buchi, Flawil, Switzerland). Liposomes,which form as the organic phase evaporates, werefreeze-thawed in liquid nitrogen 10 times and thenreduced in size to approximately 100 nm by extru-sion six times through 200-nm and five timesthrough 100-nm polycarbonate membranes (Aves-tin, Inc., Ottawa, ON, Canada) in a Liposofast ex-truder (Avestin, Inc.), as previously described(32,35). Liposomes prepared in this manner arecalled coated cationic liposomes (CCLs)(26,27).

Coupling of Anti-GD 2 MAbs toLiposomes

A previously described method(36), slightlymodified by us(32), was used to covalently linkMAbs to the maleimide terminus of DSPE-PEG-MAL. To activate the anti-GD2s for reactivity to-ward the maleimide, we utilized 2-iminothiolane(Traut’s reagent) to convert exposed amino groupson the antibody into free sulfhydryl groups. A 20 : 1mole ratio of 2-iminothiolane to MAb and 1 hour ofincubation at room temperature with occasionalmixing gave optimal MAb activation. After separa-tion of thiolated MAb from iminothiolane, with theuse of Sephadex G-25 column chromatography, theMAb was slowly added to a 5-mL test tube contain-ing the liposomes (with encapsulated oligodeoxy-nucleotides) and a small magnetic stirring bar. Op-timal coupling was obtained with the use of aphospholipid-to-MAb mole ratio of 1500–2000 : 1.Oxygen was displaced by running a slow stream ofnitrogen over the reaction mixture. The tube wascapped and sealed with Teflon tape, and the reactionmixture was incubated overnight at room tempera-ture with continuous slow stirring. The resulting im-munoliposomes were separated from unreactedMAb by chromatography with the use of SepharoseCL-4B, sterilized by filtration through 0.2-mm porecellulose membranes (Millipore Corp., Bedford,MA), and stored at 4 °C. Coupling of MAbs to li-posomes was quantified by adding trace amounts of125I-labeled MAb to the coupling reaction with li-posomes, followed byg-counting (Cobra 5002;Canberra Packard, Meriden, CT).

Coated cationic liposomes with covalently at-tached anti-GD2 MAb and encapsulated myb-as arereferred to as GD2-targeted liposomes and abbrevi-ated aGD2-CCL-myb-as throughout this report.Those that lack the antibody are referred to as non-targeted liposomes, and the “aGD2” prefix is omit-ted from the abbreviation. Liposomes with encapsu-lated myb-s have “myb-s” as their suffix.

Cell Lines and Culture Conditions

To broadly cover the phenotypes exhibited byneuroblastoma cellsin vitro, we used four GD2-positive human neuroblastoma cell lines: GI-LI-N,ACN, HTLA-230, and LAN-5 (32,37).Four GD2-negative cell lines, adherent cell lines A431 (humanepidermoid carcinoma) and HeLa (human cervicalcarcinoma) and suspension cell lines Jurkat (humanT cell) and HL-60 (human leukemia), were used insome experiments. All cell lines were maintained inthe logarithmic phase of growth at 37 °C in 75-cm2

plastic culture flasks (Corning Inc., Corning, NY) ina 5% CO2–95% air humidified incubator. They weresplit, washed, counted, and replated in fresh com-

plete medium (RPMI-1640 medium [Biochrom,Berlin, Germany], supplemented with 10% heat-inactivated fetal calf serum [Biochrom], 50 IU/mLsodium penicillin G, 50mg/mL streptomycin sulfate,and 2 mM L-glutamine), as previously described(32,38).A 1 mM EDTA solution in Hanks’ balancedsalt solution (Flow Laboratories, Milan, Italy) wasused to release adherent cell lines from the surfaceof the culture flasks.

Binding and Uptake of Liposomes

Neuroblastoma cells or control GD2-negativecells were incubated in complete medium for 2hours at 37 °C with various concentrations of[3H]CHE-labeled liposomes with encapsulated myb-as. Cells were washed extensively, treated with tryp-sin, and lysed with 1 N NaOH prior to measurementof radioactivity. In competition experiments, a 50-fold excess of free MAb was added 30 minutes be-fore addition of the liposomes. In some experiments,the chimeric human–mouse variant ch14.18 anti-GD2 MAb or the nonspecific isotype-matchedantibody, code X 0943 (Dako Corp., Glostrup, Den-mark), was coupled to myb-as-entrapping lipo-somes, as described above.

Oligodeoxynucleotide Release

Release of oligodeoxynucleotides from liposomeswas measured by dialyzing the oligodeoxynucleo-tide-containing liposomes in 25% human plasmafrom healthy donors or in complete medium againsta large volume of the same solvent at 37 °C, usingdialysis tubing with a molecular mass cutoff of100 000 daltons. The dialysis bag was sampled atintervals, and the radioactivity was measured.

Uptake of asODNs by NeuroblastomaCells

For the uptake studies, 50mg/mL free or lipo-some-encapsulated myb-as was added to each wellof six-well culture plates (Corning, Bibby Sterilin,Ltd., Staffordshire, U.K.) containing 5 × 105 cells/well. After incubation at 37 °C for 2 hours, the me-dium was removed, and cells were washed twicewith ice-cold 1% bovine serum albumin (BSA) inphosphate-buffered saline (PBS), pH 7.4. Cells wereresuspended in Versene (Life Technologies, Inc.[GIBCO BRL], Gaithersburg, MD), washed againwith BSA in PBS, and centrifuged at 300g for 5minutes at 4 °C. The cell pellet was suspended in 0.5M NaCl in 0.2M acetic acid (pH 2.5) and held at4 °C for 10 minutes to elute surface-bound oligode-oxynucleotides. Samples were centrifuged at 300gfor 5 minutes at 4 °C. The pellet was resuspended ata volume of 100mL and then frozen at −20 °C. Fordot-blot analyses, 400mL 12.5 mMEDTA in 0.5 MNaOH was added to each frozen sample, and thesamples were boiled for 10 minutes. They were thenblotted onto a nylon membrane (Amersham LifeScience, Inc., Arlington Heights, IL) by use of adot-blot or a slot-blot apparatus (Bio-Rad Laborato-ries, Richmond, CA). The membrane was washedwith 0.4M NaOH and with 2× standard saline citrate(i.e., 300 mMNaCl and 30 mMsodium citrate),dried in an oven at 80 °C for 1 hour, and then hy-bridized with a32P-labeled probe complementary tomyb-as, as outlined in the ExpressHyb protocol(Clontech Laboratories, Inc., Palo Alto, CA). Time-

254 REPORTS Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

dependence studies were carried out in the sameway, except that the time of incubation was varied.Competition experiments were carried out by addi-tion of a 50-fold excess of free anti-GD2 MAb 30minutes before the addition of aGD2-CCL-myb-as.

Inhibition of Cell Growth

GI-LI-N or HeLa cells were plated in T-25 cultureflasks (Corning) and treated with free myb-as, non-targeted liposomes (CCL-myb-as), GD2-targeted li-posomes (aGD2-CCL-myb-as), or control oligode-oxynucleotide sequences (free or encapsulated intargeted or nontargeted liposomes) at an initial con-centration of 80mg/mL. After 18, 36, and 72 hours,additional oligodeoxynucleotides (same formula-tions as were added initially) were added at a con-centration of 40mg/mL.

In some experiments, 100mg/mL of the variousoligodeoxynucleotide formulations was added tocells at the beginning of the experiment and every 2days thereafter. Two hours after each addition, thecells were washed to remove unbound oligodeoxy-nucleotides and transferred to fresh complete me-dium.

All experiments were continued for 8 days; at2-day intervals, the cells were detached with EDTA,stained with trypan blue dye, and counted micro-scopically.

Western Blot Analysis of c-Myb inNeuroblastoma Cells

GI-LI-N cells (5 × 105) were treated with freemyb-as, CCL-myb-as, or aGD2-CCL-myb-as at aninitial concentration of 100mg/mL. After 18 and 36hours, additional myb-as (same concentration andformulations as were used initially) was added. Twohours after each addition, the cells were washed andtransferred to fresh complete myb-as-free medium.Control cells were untreated with any myb-as for-mulation, washed, and transferred in the same wayas the experimental cells. At 12-hour intervals fromthe beginning of the experiment, cells were har-vested and stored frozen at −80 °C for later analysis.Frozen cell samples were solubilized in lysis buffer(i.e., 0.01 M Tris–HCI [pH 7.5], 0.144M NaCl,0.5% Nonidet P-40, 0.5% sodium dodecyl sulfate[SDS], 0.1% aprotinin, 10mg/mL leupeptin, and 2mM phenylmethylsulfonyl fluoride) and sonicated.The protein content of the samples was quantifiedwith the use of the BCA protein assay (PierceChemical Co., Rockford, IL). Forty-microgramsamples of protein were subjected to 10% SDS–polyacrylamide gel electrophoresis, with the use ofprestained molecular weight markers (AmershamInternational, Little Chalfont, Buckinghamshire,U.K.) run in parallel to aid in the localization of the75-kd c-Myb or 64- to 67-kd c-Myc protein. (Thelatter protein was used to test for equivalence ofprotein gel loading and transfer to the nitrocellulosemembranes and the specificity of the myb-as.) Theresolved proteins were blotted onto nitrocellulosemembranes, and an anti-human c-Myb MAb (cloneUBI 05–175; Upstate Biotechnology, Lake Placid,NY) or an anti-c-Myc MAb (clone 9E10; CambridgeResearch Biochemicals, London, U.K.) was used tolocalize the c-Myb or c-Myc polypeptides on theblots. Peroxidase-conjugated goat anti-mouse anti-bodies (Bio-Rad Laboratories) were used as second-ary antibodies. Immune complexes (i.e., proteins of

interest) were visualized with the use of an enhancedchemiluminescence system (Pierce Chemical Co.).The relative amount of transferable c-Myb protein ina given sample was quantified by densitometry ofx-ray films and normalization by staining the blottednitrocellulose membranes with Ponceau red to esti-mate the amount of protein loaded.

Statistics

Results are expressed as means ± 95% confidenceintervals (CIs). All data derive from at least fourindependent experiments.

RESULTS

Characterization of GD2-TargetedImmunoliposomes

In experiments for which data are notshown, we noted the following: 1) Lipo-somes prepared by our procedure are typi-cally 70–120 nm in diameter, as observedby electron microscopy(35); 2) an aver-age of 65% (95% CI4 50%–80%) of theanti-GD2 antibody was associated with li-posomes (presumably covalently attachedthrough thioether bonds), with an anti-body density of 60–80mg MAb/mmolphospholipid; 3) the efficiency of asODNtrapping in liposomes was estimated to be80%–90%; 4) during 2–4 hours of dialy-sis at 37 °C in 25% human plasma, 10%–20% of asODN was released, and theseasODNs were probably associated withthe liposomal exterior; 5) continued dialy-sis for up to 1 week produced no morefree asODN; and 6) our dialyzed lipo-somes retained their ability to bind toGD2-expressing cells over these sametime periods.

Binding of GD2-TargetedImmunoliposomes to GD2-Positiveand GD2-Negative Cells

We studied the concentration depen-dence of aGD2-CCL-myb-as-binding toGD2-positive (GI-LI-N and ACN) andGD2-negative (HeLa and HL-60) cells bymeasuring uptake of phospholipid fromliposomes. Phospholipid uptake by GD2-positive neuroblastoma cell lines in-creased with phospholipid (i.e., liposome)concentration, showing subsaturation at400 nmol phospholipid/mL (Fig. 1). Simi-lar results were obtained with the use ofboth mouse 14.G2a or its chimeric hu-man–mouse variant ch14.18 anti-GD2

MAbs (data not shown). The same lipo-some preparation without anti-GD2 MAb(CCL-myb-as) showed very low phos-pholipid uptake with no evidence of satu-ration. Preincubation of cells for 30 min-utes with soluble anti-GD2 MAb, but not

with an unrelated MAb, almost com-pletely blocked phospholipid uptake bycells, suggesting that aGD2-CCL-myb-asbinding occurred through specific anti-gen–antibody recognition. Moreover,coupling to liposomes of nonspecific iso-type-matched antibody did not increasephospholipid uptake by neuroblastomacells (data not shown). Both GD2-negative cell lines showed very low cellphospholipid uptake in all cases examined(Fig. 1).

Uptake of Free and EncapsulatedOligodeoxynucleotides byNeuroblastoma Cells

Four GD2-positive neuroblastoma celllines and four GD2-negative unrelated tu-mor cell lines were incubated with free orencapsulated myb-as, and uptake of theoligodeoxynucleotide was measured bydot-blot analysis. Since there were no ob-vious differences in the degree of uptakeof liposome-encapsulated myb-as by thefour GD2-positive cell lines, only the dataobtained with two GD2-positive cell lines(GI-LI-N and ACN) and one GD2-negative cell line (HeLa) are shown (Fig.2). Fig. 2, A, is an example of the dot-blotanalysis of myb-as uptake. The amount ofmyb-as taken up by GD2-positive cellstreated with aGD2-CCL-myb-as (dots 7and 8) was substantially greater than thattaken up by cells treated with free myb-as(dots 3 and 4) or CCL-myb-as (dots 5 and6). By using densitometry, we estimatedthat the uptake of myb-as into GI-LI-Nand ACN cells was fourfold to fivefoldhigher when they were incubated withaGD2-CCL-myb-as than when they wereincubated with an equal amount of freemyb-as or CCL-myb-as (Fig. 2, B).Longer incubation led to an approxi-mately 10-fold greater uptake of myb-asfrom aGD2-CCL-myb-as, relative to up-take of free myb-as. After 24 hours’ in-cubation, uptake from aGD2-CCL-myb-as relative to uptake from CCL-myb-asdecreased to only twofold (Fig. 2, C). In acompetition experiment, the uptake ofaGD2-CCL-myb-as was inhibited by a50-fold excess of free anti-GD2 (Fig. 2, B).

Effect of Liposome-Encapsulatedmyb-as on Neuroblastoma CellProliferation

To determine the effects of myb-as onneuroblastoma cell proliferation, wetreated GI-LI-N cells with free or lipo-some-encapsulated myb-as, using a pro-

Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000 REPORTS 255

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

tocol previously shown to inhibit neuro-blastoma cell growth by asODNs(9). Asexpected, free myb-as inhibited GI-LI-Ncell proliferation up to 70%. Treatment ofcells with CCL-myb-as or aGD2-CCL-myb-as induced a more rapid and strongerinhibition of neuroblastoma cell growth(Fig. 3, A). Conversely, in GD2-negativeHeLa cells, myb-as, in all formulations,affected cell proliferation to a smaller ex-tent (Fig. 3, B). To demonstrate the se-quence specificity of these effects, wealso delivered myb-s into GI-LI-N andHeLa cells. With the use of formulationsof the same types as those used with myb-as, myb-s showed little or no growth in-hibition (Fig. 3, A and B). Scrambled-sequence oligodeoxynucleotides alsowere ineffective (data not shown).

To assess the efficiency of GD2-targeted liposomes in delivering myb-asto target cells, we used a protocol thatallows high levels of antibody-mediated

binding while minimizing nonspecific ad-sorption of liposomes to cells(32,37).Inthese experiments, tumor cells were ex-posed to oligodeoxynucleotides for 2hours at the beginning of the experimentand every 2 days thereafter (for 8 days)and then washed and transferred to oligo-deoxynucleotide-free complete medium.Under these conditions, free myb-as,CCL-myb-as, free myb-s, CCL-myb-s,and aGD2-CCL-myb-s had no substantialcytotoxic effect on GI-LI-N cells (Fig. 3,C). In contrast, aGD2-CCL-myb-as in-duced a marked decrease in neuroblas-toma cell proliferation over the same8-day period. Under the same conditions,none of our oligodeoxynucleotide formu-lations inhibited proliferation of the GD2-negative cell lines HeLa (Fig. 3, D) orHL-60, a cell line that has been shown tobe strictly dependent on c-Myb proteinfor cell proliferation (5,39) (data notshown). Hence, antiproliferative activity

under these conditions appears to requireexpression of the GD2 antigen on the cellsurface to allow binding followed by in-ternalization of the aGD2-CCL-myb-as.

To investigate whether the enhancedcytotoxicity of aGD2-CCL-myb-as de-pends on specific recognition of cells byantibodies, competition studies were per-formed. Addition of a 50-fold excess offree anti-GD2 MAb to a culture of GI-LI-N cells that were growing in the pres-ence of aGD2-CCL-myb-as caused a sub-stantial reduction in the antiproliferativeeffect of aGD2-CCL-myb-as, while addi-tion of an unrelated MAb had little effect(data not shown).

Effect of myb-as on c-Myb Expressionin Neuroblastoma Cells

The expression of c-Myb protein inGI-LI-N cells treated with free myb-as,CCL-myb-as, or aGD2-CCL-myb-as wasanalyzed by western blot analysis (Fig. 4,A). Cells treated with free myb-as orCCL-myb-as (by addition of 100mg/mLinhibitor at 0, 18, and 36 hours followed,2 hours later, by transfer to fresh inhibi-tor-free complete medium) showed levelsof protein expression similar to those inthe control (untreated) cells. In contrast,GI-LI-N cells treated with the aGD2-CCL-myb-as showed a reduction in c-Myb protein levels of about 70% in com-parison with control cells. Free orliposome-entrapped sense or scrambledanalogues of myb-as did not affect c-Mybprotein expression (data not shown). Thetime dependence of this effect is shown inFig. 4, B. To examine the specificity ofprotein repression by myb-as, the effect ofaGD2-CCL-myb-as on the expression ofc-Myc protein was determined. As ex-pected, there was no repression of c-Mycexpression by aGD2-CCL-myb-as (Fig. 4, C).

DISCUSSION

We have shown that an asODNcomplementary to eight contiguouscodons of c-Myb mRNA inhibited thegrowth of neuroblastoma cellsin vitroand that its inhibitory effect was greatestwhen it was delivered to the cells in GD2-targeted, sterically stabilized liposomes.Our immunoliposomal asODN prepara-tion also inhibits expression of c-Mybprotein by these cells. Our results demon-strate the feasibility of the selective deliveryof myb-as to GD2-positive neuroblastoma

Fig. 1. Concentration dependence of liposome binding to disialoganglioside GD2-positive (GI-LI-N andACN) and GD2-negative (HL-60 and HeLa) tumor cells. Cells were incubated for 2 hours with the indicatedconcentrations of3H-labeled liposomes with encapsulated c-myb antisense oligodeoxynucleotides (myb-as)and with (open circles) or without (open triangles) coupled anti-GD2 monoclonal antibodies (14.G2a MAb).In competition experiments, cells were preincubated for 30 minutes with a 50-fold excess of either 14.G2aMAb (closed circles) or an isotype-matched nonspecific antibody (code X 0943;open squares) beforeaddition of3H-labeled liposomes with coupled anti-GD2 MAb and encapsulated myb-as. Cells were washed,treated with trypsin, and lysed with 1 N NaOH prior to measurement of radioactivity.Error bars in the toptwo panelsare 95% confidence intervals. Only the upper portions of the error bars are shown for the uppercurve, and only the lower portions are shown for the lower curves.

256 REPORTS Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

cell lines by targeted immunoliposomes.The myb-as oligodeoxynucleotide, LR-3001, used in this study is being tested inan ongoing phase II clinical trial for thetreatment of chronic myelogenous leuke-mia (34).

Neuroblastoma is a pediatric neuroec-toderm-derived tumor that often presentsat an advanced stage and is characterizedby a very poor prognosis. Despite devel-opment of very aggressive multimodaltherapeutic approaches, the long-term sur-vival of patients with neuroblastomas hasnot improved substantially over the lasttwo decades(1,2). Additional therapeutic

strategies are continuously being exploredfor their clinical potential as alternative oradjuvant treatment.

asODNs are being investigated as pos-sible therapeutic agents that take directadvantage of molecular complementarity(10,17,18)to inhibit gene expression in aspecific manner. The antisense approachuses a short oligodeoxynucleotide de-signed to bind to a target mRNA tran-script through Watson-Crick base-pairingto form an asODN–RNA heteroduplexthat inhibits gene expression.

While this conceptually simple,straightforward, and rational approach to

novel drug design is attractive for thetreatment of cancers and other human dis-eases(10,40),using it therapeutically toreduce or to eliminate expression of geneproducts important for the growth of neu-roblastoma cells is much more complexfor several reasons. These include theneed to deliver asODNs selectively tospecific areas of the body in order tomaximize their action and to minimizetheir side effects(41), instability, and im-permeability. A previously developedtherapeutic strategy that addresses manyof these potential problems uses lipo-somes as carriers for asODNs(42).

Fig. 2. Cellular uptake of c-myb antisense oligodeoxynucleotides (myb-as).Panel A: dot-blot evaluation of intracellular myb-as. For the standards lane(STD), various amounts of free myb-as (200, 100, 50, 25, 12.5, 6.25, 3.12, 1.55ng in dots 1–8,respectively) were added to frozen untreated cell samples.Dotsin the other lanes show myb-as released from the indicated types of cellsincubated at 37 °C for 2 hours with the following additions: none (dots 1 and 2),free myb-as (dots 3 and 4), myb-as encapsulated in liposomes without anti-disialoganglioside GD2 monoclonal antibody (MAb) (dots 5 and 6), and myb-asencapsulated in liposomes with anti-GD2 MAb (dots 7 and 8). After addition offree myb-as (STD lane) or incubation (other lanes), 400mL 12.5 mMEDTA in0.5 M NaOH was added to the mixtures, and the samples were boiled for 10minutes. They were then blotted onto a nylon membrane by use of a dot-blotapparatus. The membrane was washed with 0.4M NaOH and with 2× standardsaline citrate (i.e., 300 mM NaCl and 30 mMsodium citrate), dried in an ovenat 80 °C for 1 hour, and then hybridized with a32P-labeled probe complementary

to myb-as.Panel B: densitometric analysis, after hybridization and autoradiog-raphy, of released intracellular myb-as. The uptake of myb-as mediated byliposomes was expressed as percentage of the uptake obtained with free myb-asin these independent experiments taken as 100%. CCL-myb-as4 cells incu-bated with myb-as encapsulated in liposomes without anti-GD2 MAb. aGD2-CCL-myb-as4 cells incubated with myb-as encapsulated in liposomes withanti-GD2 MAb. aGD2-CCL-myb-as + aGD24 cells incubated with myb-asencapsulated in liposomes with anti-GD2 MAb following preincubation with50-fold excess free anti-GD2 MAb for 30 minutes.Error bars are 95% confi-dence intervals.Panel C: time-dependent intracellular uptake into GI-LI-N cellsof myb-as from free myb-as, CCL-myb-as, or aGD2-CCL-myb-as (abbreviationsas in panel B). Scanning densitometry was used to quantify myb-as, using datafrom the STD lane in panel A to generate a standard curve.Error bars are 95%confidence intervals. ODN4 oligodeoxynucleotide.

Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000 REPORTS 257

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

Here, we have described a system inwhich asODNs are encapsulated withinsmall (about 100 nm) liposomes that havehigh trapping efficiencies (up to 90% ofadded oligodeoxynucleotide is stably as-sociated with lipids) and bear covalentlyattached MAbs specific for the GD2 anti-gen on their exterior. The rationale forthis approach is the extensive expressionof GD2 on neuroblastoma cells in com-parison to its limited presence on normaltissues, such as the peripheral nervoussystem and the cerebellum(28–31).

We have compared the cellular uptakesof free asODN, asODN encapsulated inGD2-targeted liposomes, and asODN en-capsulated in nontargeted liposomes. En-capsulation of asODN in GD2-targeted li-posomes was found to promote a fourfoldto 10-fold increase in cellular uptake of

asODNs when compared with the free oli-godeoxynucleotide. The increase in up-take relative to CCL-myb-as approaches10-fold at 6 hours’ incubation time, butthe relative size of the increase diminisheswith time, to only twofold at 24 hours.Uptake was cell type specific: Of the celltypes we examined, only those that ex-press GD2 antigens took up more asODNfrom receptor-targeted liposomes thanfrom nontargeted liposomes. That the up-take of asODN by neuroblastoma cellscould be competitively inhibited by ex-cess free anti-GD2 MAb is consistent withthe hypothesis that intracellular deliveryof asODN from GD2-targeted liposomesis mediated by GD2 molecules in theplasma membrane. Similar findings havebeen reported for the folate receptor(13,43).

Although we did not perform fraction-ation assays to assess the subcellular lo-calization of immunoliposome-deliveredasODN, we can conclude that it was notbound to the plasma membrane, sinceacetic acid was used to elute surface-bound asODN in our studies of asODNuptake. Because aGD2-CCL-myb-as wasable to inhibit cell growth and to suppressc-Myb expression, we conclude that myb-as was delivered intracellularly in func-tionally active form. Since it has beenshown that 14.G2a MAb was internalizedby melanoma cells, even when conjugatedto ricin A-chain (44), these findings sug-gest that, in our experimental model sys-tem, liposomes bind to the cell surfaceand are then internalized via a receptor-dependent endocytic pathway. This con-clusion agrees with previous reports

Fig. 3. Growth inhibition of neuro-blastoma cells by c-myb antisenseoligodeoxynucleotides (myb-as) en-capsulated in anti-disialogangliosideGD2 liposomes.Panels AandB rep-resent GI-LI-N and HeLa cells, re-spectively, treated with free senseoligodeoxynucleotide (free-myb-s),free-myb-as, sense oligodeoxynucleo-tide encapsulated in liposomes with-out anti-GD2 monoclonal antibody(CCL-myb-s), antisense oligodeoxy-nucleotides encapsulated in lipo-somes without anti-GD2 monoclonalantibody (CCL-myb-as), sense oligo-deoxynucleotide encapsulated in li-posomes with anti-GD2 monoclonalantibody (aGD2-CCL-myb-s), ormyb-as encapsulated in liposomeswith anti-GD2 monoclonal antibody(aGD2-CCL-myb-as). The initial oli-godeoxynucleotide concentrationwas 80 mg/mL. Oligodeoxynucleo-tides were added at 40mg/mL after18, 36, and 72 hours. At the begin-ning of each experiment and at 2-dayintervals thereafter, the cells weredetached with EDTA, stained withtrypan blue, and counted microscopi-cally. Panels CandD represent thesame cell lines treated by adding thesame oligodeoxynucleotide prepara-tions to the growing cells at a con-centration of 100mg/mL at the be-ginning of the experiment and every2nd day thereafter. Two hours aftereach of these additions, the cellswere washed to remove unbound oli-godeoxynucleotides and transferredto fresh complete medium. Cellswere detached, stained, and countedas in the experiments shown in pan-els A and B. Data are expressed aspercentage of control, anderrorbars are 95% confidence intervals.

258 REPORTS Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

(13,45) that showed that endocytosis ofimmunoliposomes leads to nondestructiverelease of much of the internalized mate-rial into the cytosol, and it confirms theimportance of internalizing receptors(27,36).

One of the most attractive features ofantisense-based inhibitors is the potentialfor great specificity of their inhibitory ef-fects(40).Demonstrating such specificityis usually considered to be the most im-portant criterion for concluding that a trueantisense mechanism explains the bio-logic effects of a particular oligodeoxy-nucleotide. For this reason, we have ex-amined the specificity of inhibition ofc-Myb expression and tumor growth byLR-3001, an oligodeoxynucleotidecomplementary to codons 2–9 of humanc-Myb mRNA, and have demonstratedthat the effects of this oligodeoxynucleo-tide are specific, as measured by a numberof parameters. Scrambled and sense ana-logues of LR-3001 had no effect on c-Myb protein expression or on neuroblas-toma growth in cell culture. LR-3001inhibits c-Myb protein expression but

does not affect the expression of c-Myc, arelated transcription factor with a similarhalf-life (46). These findings suggest thatthe inhibitory effects of LR-3001 on c-Myb expression and its biologic conse-quences occur through an antisensemechanism.

Our results illustrate that encapsulationof asODN in GD2-targeted liposomes canprotect nontargeted cells from a poten-tially toxic molecule while simulta-neously enhancing the molecule’s toxicitytoward a targeted cell population. Onecannot, however, expect that even a care-fully targeted liposome will end up onlyat the intended tumor site; a portion of thedose could accumulate at other bodilylocations with potentially toxic conse-quences. In futurein vivo studies, the tox-icity of aGD2-CCL-myb-as at nontar-geted sites deserves attention.

The great advantages of liposome-encapsulated cytotoxic agents over freemolecules have been unquestionablydemonstrated in several animal models ofcancer(47–50),mainly against small pri-mary and micrometastatic solid tumors

(36,51).Thus, the use of myb-as, encap-sulated into GD2-targeted liposomes, mayhold promise for the treatment of neuro-blastoma in the minimal residual diseasesetting. Moreover, the great selectivity ofaGD2-CCL-myb-as for GD2-positivecells suggests exploring using it to purgecontaminating neuroblastoma cells frombone marrow before reinfusion into pa-tients undergoing autologous bone mar-row transplantation(52),possibly in com-bination with CD34+ cell selection.

Finally, the fact that many melanomaspecimens have altered expression of c-Myb, coupled with the previously demon-strated inhibition of melanoma growth byc-myb asODNs, suggests that there maybe a potential role for c-myb antisensetherapy in the treatment of this tumor(53). Thus, immunotargeting of lipo-somes to neuroectoderm-derived tumorcells may provide an effective strategy fortumor site-selective delivery of drugswith broad therapeutic applications.

REFERENCES

(1) Pritchard J. “Megatherapy” for advanced neu-roblastoma—rationale and role. Eur J Cancer1995;31A:134–6.

(2) Niethammer D, Handgretinger R. Clinicalstrategies for the treatment of neuroblastoma.Eur J Cancer 1995;31A:568–71.

(3) Bister K, Nunn M, Moscovici G, Perbal B,Baluda MA, Duesberg PH. Acute leukaemiaviruses E26 and avian myeloblastosis virushave related transformation-specific RNA se-quences but different genetic structures, geneproducts, and oncogenic properties. Proc NatlAcad Sci U S A 1982;72:3677–81.

(4) Gewirtz AM, Calabretta B. A c-myb antisenseoligodeoxynucleotide inhibits normal humanhematopoiesisin vitro. Science 1988;242:1303–6.

(5) Calabretta B, Sims RB, Valtieri M, CaraccioloD, Szczylik C, Venturelli D, et al. Normal andleukemic hematopoietic cells manifest differ-ential sensitivity to inhibitory effects ofc-myb antisense oligodeoxynucleotides: anin vitro study relevant to bone marrow purg-ing. Proc Natl Acad Sci U S A 1991;88:2351–5.

(6) Alitalo K, Winquist R, Linn CC, de La ChapelleA, Schwab M, Bishop JM. Aberrant expressionof an amplified c-myb oncogene in two celllines from a colon carcinoma. Proc Natl AcadSci U S A 1984;81:4534–8.

(7) Griffin CA, Baylin SB. Expression of the c-myb oncogene in human small cell lung carci-noma. Cancer Res 1985;45:272–5.

(8) Thiele CJ, McKeon C, Triche TJ, Ross RA,Reynolds CP, Israel MA. Differential protoon-cogene expression characterizes histopatholog-ically indistinguishable tumors of the periph-eral nervous system. J Clin Invest 1987;80:804–11.

(9) Raschella G, Negroni A, Skorski T, Pucci S,Nieborowska-Skorska M, Romeo A, et al. In-

Fig. 4.Selective inhibition of c-Myb protein expression in GI-LI-N cells. Except in control experiments, freec-myb antisense oligodeoxynucleotide (free-myb-as), myb-as encapsulated in liposomes without anti-disialoganglioside GD2 monoclonal antibodies (CCL-myb-as), or myb-as encapsulated in liposomes withanti-GD2 monoclonal antibody (aGD2-CCL-myb-as) was added to GI-LI-N cells at a concentration of 100mg/mL at the beginning of the experiment and 18 and 36 hours later. Two hours after each addition, the cellswere washed and transferred to fresh myb-as-free complete medium. The cells were harvested at 48 hours(panel A) or at the indicated time points (panels BandC) and frozen for later analysis. Analysis for proteinexpression (c-Myb for panels A and B and c-Myc for panel C) was by immunoblotting. (See“Materials andMethods” section for more details.) Kd4 kilodalton.

Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000 REPORTS 259

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

hibition of proliferation by c-myb antisenseRNA and oligodeoxynucleotides in trans-formed neuroectodermal cell lines. Cancer Res1992;52:4221–6.

(10) Ho PT, Parkinson DR. Antisense oligonucleo-tides as therapeutics for malignant diseases.Semin Oncol 1997;24:187–202.

(11) Monia BP, Johnston JF, Geiger T, Muller M,Fabbro D. Antitumor activity of a phosphoro-thioate antisense oligodeoxynucleotide tar-geted against C-raf kinase. Nat Med 1996;2:668–75.

(12) Neurath MF, Pettersson S, Meyer zum Buchen-felde KH, Strober W. Local administration ofantisense phosphorothioate oligonucleotides tothe p65 subunit of NF-kB abrogates estab-lished experimental colitis in mice. Nat Med1996;2:998–1004.

(13) Wang S, Lee RJ, Cauchon G, Gorenstein DG,Low PS. Delivery of antisense oligodeoxyribo-nucleotides against the human epidermalgrowth factor receptor into cultured KB cellswith liposomes conjugated to folate via poly-ethylene glycol. Proc Natl Acad Sci U S A1995;92:3318–22.

(14) Crooke RM. Cellular uptake, distribution andmetabolism of phopsphorothioate, phosphodi-ester, and methylphosphonate oligonucleo-tides, chapt 24. In: Agrawal S, Crooke ST,editors. Antisense research and applications.New York (NY): Springer-Verlag; 1998. p.427–49.

(15) Agrawal S, Goodchild J, Civeira MP, ThorntonAH, Sarin PS, Zamecnik PC. Oligodeoxy-nucleoside phosphoramidates and phosphoro-thioates as inhibitors of human immunodefi-ciency virus. Proc Natl Acad Sci U S A 1988;85:7079–83.

(16) Srinivasan SK, Iversen Pl. Review ofin vivopharmacokinetics and toxicology of phospho-rothioate oligonucleotides. J Clin Lab Anal1995;9:129–37.

(17) Henry SP, Templin MV, Gillett N, Rojko J,Levin AA. Correlation of toxicity pharmacoki-netic properties of a phosphorothioate oligo-nucleotide designed to inhibit ICAM-1. Toxi-col Pathol 1999;27:95–100.

(18) Stein C, Cheng YC. Antisense oligonucleo-tides as therapeutic agents—Is the bullet reallymagical?. Science 1993;261:1004–12.

(19) Corrias MV, Guarnaccia F, Ponzoni M. Bio-availability of antisense oligonucleotides inneuroblastoma cells: comparison of efficacyamong different types of molecules. J Neuro-Oncol 1997;31:171–80.

(20) Beltinger C, Saragovi HU, Smith RM, LeSau-teur L, Shah N, De Dionisio L, et al. Binding,uptake, and intracellular trafficking of phos-phorothioate-modified oligodeoxynucleotides.J Clin Invest 1995;95:1814–23.

(21) Bennet CF, Chiang MY, Chan H, ShoemakerJE, Mirabelli CK. Cationic lipids enhance cel-lular uptake activity of phosphorothioate anti-sense oligonucleotides. Mol Pharmacol 1992;41:1023–33.

(22) Litzinger DC, Brown JM, Wala I, KaufmanSA, Van GY, Farrell CL, et al. Fate of cationicliposomes and their complex with oligonucleo-tide in vivo.Biochim Biophys Acta 1996;1281:139–49.

(23) Trubetskoy VS, Torchilin VP, Kennel S,Huang L. Cationic liposomes enhance targeteddelivery and expression of exogenous DNAmediated by N-terminal modified poly(L-lysine)-antibody conjugate in mouse lung en-dothelial cells. Biochim Biophys Acta 1992;1131:311–3.

(24) Thierry A, Dritschilo A. Intracellular availabil-ity of unmodified, phosphorothioated and lipo-somally encapsulated oligodeoxynucleotidesfor antisense activity. Nucleic Acids Res 1992;20:5691–8.

(25) Stuart DD, Allen TM. Pharmacokineticsof pegylated liposomal asODNs. Proc IntSymp Control Rel Bioact Mater 1998;25:366–7.

(26) Stuart DD, Kao GY, Allen TM. A novel, long-circulating and functional liposomal formula-tion of antisense oligodeoxynucleotides tar-geted against MDR1. Cancer Gene Ther. Inpress 1999.

(27) Allen TM, Moase EH. Therapeutic opportuni-ties for targeted liposomal drug delivery. AdvDrug Del Rev 1996;21:117–33.

(28) Frost JD, Hank JA, Reaman GH, Frierdich S,Seeger RC, Gan J, et al. A phase I/IB trial ofmurine monoclonal anti-GD2 antibody 14.G2aplus interleukin-2 in children with refractoryneuroblastoma: a report of the Children’s Can-cer Group. Cancer 1997;80:317–33.

(29) Uttenreuther-Fischer MM, Huang CS, Yu AL.Pharmacokinetics of human-mouse chimericanti-GD2 mAb ch14.18 in a phase I trial inneuroblastoma patients. Cancer Immunol Im-munother 1995;41:331–8.

(30) Mujoo K, Cheresh DA, Yang HM, ReisfeldRA. Disialoganglioside GD2 on human neuro-blastoma cells: target antigen for monoclonalantibody-mediated cytolysis and suppressionof tumor growth. Cancer Res 1987;47:1098–1104.

(31) Schulz G, Cheresh DA, Vark NM, Yu A,Staffileno LK, Reisfeld RA. Detection of gan-glioside GD2 in tumor tissues and sera of neu-roblastoma patients. Cancer Res 1984;44:5914–20.

(32) Pagnan G, Montaldo PG, Pastorino F, Raffa-ghello L, Kirchmeier M, Allen TM, et al. GD2-mediated melanoma cell targeting and cytotox-icity of liposome-entrapped fenretinide. Int JCancer 1999;81:268–74.

(33) Honsik CJ, Jung G, Reisfeld RA. Lymphokine-activated killer cells targeted by monoclonalantibody to the disialoganglioside GD2 andGD3 specifically lyse human tumor cells ofneuroectodermal origin. Proc Natl Acad SciU S A 1986;83:7983–7.

(34) Luger SM. Phase II study of autologous bonemarrow transplant using marrow treated withc-myb antisense oligodeoxynucleotide (LR-3001) in chronic myelogenous leukemia inchronic or accelerated phase. PQD ClinicalTrial Abs.; Protocol IDs: UPCC-3492, NCI-H94-0532. Bethesda (MD): CancerNet™; lastmodified 1998.

(35) Pagnan G, Caridi G, Montaldo PG, Bado M,Chiesa V, Allen TM, et al. Apoptosis of humanneuroblastoma cells induced by liposome-encapsulated fenretinide. J Liposome Res1998;8:401–23.

(36) Kirpotin D, Park JW, Hong K, Zalipsky S, LiWL, Carter P, et al. Sterically stabilized anti-HER2 immunoliposomes: design and targetingto human breast cancer cellsin vitro. Biochem-istry 1997;36:66–75.

(37) Montaldo PG, Pagnan G, Raffaghello R, Pas-torino F, Allen TM, Kirchmeier M, et al. AntiGD2-immunoliposome-mediated targeting of[125I]metaiodobenzylguanidine to neuroblas-toma and melanoma cellsin vitro. J LiposomeRes 1999;9:367–85.

(38) Ponzoni M, Bocca P, Chiesa V, Decensi A,Pistoia V, Raffaghello L, et al. Differential ef-fects of N-(4-hydroxyphenil)retinamide andretinoic acid on neuroblastoma cells: apoptosisversus differentiation. Cancer Res 1995;55:853–61.

(39) Tondelli L, Ricca A, Laus M, Lelli M, Citro G.Highly efficient cellular uptake of c-myb anti-sense oligonucleotides through specifically de-signed polymeric nanospheres. Nucleic AcidsRes 1998;26:5425–31.

(40) Crooke ST. Therapeutic applications of oligo-nucleotides. Annu Rev Pharmacol Toxicol1992;32:329–76.

(41) Davis SS. Biomedical applications of nano-technology—implications for drug targetingand gene therapy. Trends Biotechnol 1997;15:217–24.

(42) Zelphati O, Szoka FC Jr. Liposomes as a car-rier for intracellular delivery of antisense oli-gonucleotides: a real or magic bullet? J Con-trolled Release 1996;41:99–119.

(43) Li S, Huang L. Targeted delivery of antisenseoligodeoxynucleotides formulated in a novellipidic vector. J Liposome Res 1998;8:239–50.

(44) Wargalla UC, Reisfeld R. Rate of internaliza-tion of an immunotoxin correlates with cyto-toxic activity against human tumor cells. ProcNatl Acad Sci U S A 1989;86:5146–50.

(45) Lee RJ, Low PS. Folate-mediated tumor celltargeting of liposome-entrapped doxorubicininvitro. Biochim Biophys Acta 1995;1233:134–44.

(46) Ramsay G, Evan GI, Bishop JM. The pro-tein encoded by the human proto-oncogene c-myc. Proc Natl Acad Sci U S A 1984;81:7742–6.

(47) Allen TM, Brandeis E, Hansen CB, Kao GY,Zalipsky S. A new strategy for attachmentof antibodies to sterically stabilized lipo-somes resulting in efficient targeting to cancercells. Biochim Biophys Acta 1995;1237:99–108.

(48) Huwyler J, Wu D, Pardridge WM. Brain drugdelivery of small molecules using immunoli-posomes. Proc Natl Acad Sci U S A 1996;93:14164–9.

(49) Spragg DD, Alford DR, Greferath R, LarsenCE, Lee KD, Gurtner GC, et al. Immunotar-geting of liposomes to activated vascular en-dothelial cells: a strategy for site-selective de-livery in the cardiovascular system. Proc NatlAcad Sci U S A 1997;65:491–7.

(50) Lopes De Menezes DE, Pilarski LM, AllenTM. In vitro and in vivo targeting of immuno-liposomal doxorubicin to human B-cell lym-phoma. Cancer Res 1998;58:3320–30.

(51) Allen TM, Ahmad I, Lopes De Menezes DE,Moase EH. Immunoliposome-mediated target-

260 REPORTS Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

ing of anti-cancer drugsin vivo. Biochem SocTrans 1995;23:1073–9.

(52) Leung W, Chen AR, Klann RC, Moss TJ,Davis JM, Noga SJ, et al. Frequent detection oftumor cells in hematopoietic grafts in neuro-blastoma and Ewing’s sarcoma. Bone MarrowTransplant 1998;22:971–9.

(53) Walker MJ, Silliman E, Dayton MA, Lang JC.The expression of c-myb in human metastaticmelanoma cell lines and specimens. AnticancerRes 1998;18:1129–35.

NOTES

Supported by Associazione Italiana Ricerca sulCancro, by Associazione Italiana Neuroblastoma,and by Ricerca Corrente Ministeriale Gaslini.

We thank R. A. Reisfeld and R. Handgretinger forfurnishing us with 14.G2a and ch14.18 antibodies;G. Zon for providing us with LR-3001; C. Brignoleand P. Bocca for expert technical assistance; E.Campochiaro for editing; and R. A. Reisfeld, L. Va-resio, V. Corrias, and V. Pistoia for helpful discus-sions.

Manuscript received June 22, 1999; revised No-vember 18, 1999; accepted December 1, 1999.

Journal of the National Cancer Institute, Vol. 92, No. 3, February 2, 2000 REPORTS 261

by guest on June 22, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from