Expression of enzymes and receptors of the leukotriene pathway in human neuroblastoma promotes tumor survival and provides a target for therapy

The FASEB Journal • Research Communication

Expression of enzymes and receptors of the leukotrienepathway in human neuroblastoma promotes tumorsurvival and provides a target for therapy

Baldur Sveinbjornsson,*,§,1,2 Agnes Rasmuson,*,1 Ninib Baryawno,* Min Wan,†

Ingvild Pettersen,§ Frida Ponthan,*,� Abiel Orrego,‡ Jesper Z. Haeggstrom,†

John I. Johnsen,* and Per Kogner**Childhood Cancer Research Unit, Department of Woman and Child Health, †Department ofMedical Biochemistry and Biophysics, and ‡Department of Oncology and Pathology, KarolinskaInstitutet, Stockholm, Sweden; §Department of Cell Biology and Histology, Faculty of Medicine,University of Tromso, Tromso, Norway; and �Northern Institute for Cancer Research, NewcastleUniversity, Newcastle upon Tyne, UK

ABSTRACT The metabolism of arachidonic acid bythe cyclooxygenase (COX) or lipoxygenase (LO) path-ways generates eicosanoids that have been implicated inthe pathogenesis of a variety of human diseases, includ-ing cancer. In this study, we examined the expressionand significance of components within the 5-LO path-way in human neuroblastoma, an embryonal tumor ofthe sympathetic nervous system. High expression of5-LO, 5-LO-activating protein (FLAP), leukotriene A4hydrolase, leukotriene C4 synthase, and leukotrienereceptors was detected in a majority of primary neuro-blastoma tumors and all cell lines investigated. Expres-sion of 5-LO and FLAP was evident in tumor cells butnot in nonmalignant adrenal medulla where neuroblas-tomas typically arise. Moreover, neuroblastoma cellsproduce leukotrienes, and stimulation of neuroblas-toma cells with leukotrienes increased neuroblastomacell viability. Inhibitors of 5-LO (AA-861), FLAP (MK-886), or the leukotriene receptor antagonist monte-lukast inhibited neuroblastoma cell growth by inductionof G1-cell cycle arrest and apoptosis. Similarly, specific5-LO and leukotriene receptor silencing by small inter-fering RNA decreased neuroblastoma cell growth.These findings provide new insights into the pathobiol-ogy of neuroblastoma, and the use of leukotrienepathway inhibitors as a novel adjuvant therapy forchildren with neuroblastoma warrants further consider-ation.—Sveinbjornsson, B., Rasmuson, A., Baryawno,N., Wan, M., Ingvild Pettersen, I., Frida Ponthan, F.,Orrego, A., Haeggstrom, J. Z., Johnsen, J. I., Kogner, P.Expression of enzymes and receptors of the leukotri-ene pathway in human neuroblastoma promotes tumorsurvival and provides a target for therapy. FASEB J. 22,3525–3536 (2008)

Key Words: eicosanoids � 5-lipoxygenase � apoptosis

Recent literature has extended the concept thatinflammatory mediators are critical components of

tumor progression. It is well established that manycancers arise from sites of infection and chronic inflam-mation. The tumor environment, which is orchestratedby the interplay between inflammatory cells, tumorcells, and other tumor-associated host cells, is an indis-pensable participant in the tumorigenic process. Exces-sively and chronically produced inflammatory media-tors are thought to promote tumor progression andsurvival. Tumor cells themselves may produce signalingmolecules such as arachidonic acid metabolites, inflam-matory cytokines, and their receptors for survival, inva-sion, and metastasis (1).

Neuroblastoma, the most common extracranial solidtumor in children, is characterized by a heterogeneousclinical behavior. Some tumors may undergo spontane-ous regression or differentiation, whereas the majorityof metastatic neuroblastomas have poor prognosis de-spite intensive therapy (2).

In comparison with nonmalignant nervous tissue,neuroblastoma cells contain increased levels of arachi-donic acid, the main substrate for eicosanoid biosyn-thesis catalyzed by cyclooxygenases (COXs) and lipoxy-genases (LOs; refs.3–5). The COX pathway generatesprostaglandins and thromboxane, whereas the LOpathway generates leukotrienes, lipoxins, and hy-droxyecosatetraenoic acid (Fig. 1). The role of COX-2and prostaglandins in cancer has been widely studiedand implicated in resistance to apoptosis as well asinduction of metastasis and angiogenesis (3).

In contrast to prostaglandins, leukotrienes are mostlyproduced by inflammatory cells, even though leukotri-ene production in cells of nonhematopoetic origin alsohas been reported (6–8). Leukotrienes are produced

1 These authors contributed equally to this work.2 Correspondence: Childhood Cancer Research Unit,

Astrid Lindgren Children’s Hospital, Q6:05, KarolinskaHospital, S-171 76 Stockholm, Sweden, E-mail: [email protected]

doi: 10.1096/fj.07-103457

35250892-6638/08/0022-3525 © FASEB

in a sequence of events that include phospholipaseA2-mediated release of arachidonic acid and transloca-tion of 5-LO to the nuclear envelope, where it associ-ates with 5-LO activating protein (FLAP). Subsequently,5-LO oxidizes arachidonic acid to 5-HPETE, which maybe reduced to 5-hydroxyeicosatetranoic acid (5-HETE).Alternatively, 5-HPETE is dehydrated into the unstableepoxide leukotriene (LT) A4. This intermediate canthen be converted into LTB4 by LTA4 hydrolase(LTA4H) or conjugated with glutathione to form LTC4in a reaction catalyzed by a specific LTC4 synthase(LTC4S). After extracellular transfer by multidrug-resis-tance-associated protein-1 (MRP1), LTC4 may be con-verted to LTD4 and further into LTE4 (5, 6). LTC4,LTD4, and LTE4 are collectively known as cysteinylleukotrienes (CysLT; ref. 6; Fig. 1).

LTB4 is a potent chemoattractant that exerts its effectthrough two G-protein- coupled receptors, designatedBLT1 and BLT2. BLT1 is a high affinity receptor forLTB4 and is expressed mainly on leukocytes, whereas

BLT2 is a low affinity receptor with high expression onleukocytes, liver, and spleen and weak expression inmost other tissues (9). Likewise, cysteinyl leukotrienessignal through two G-protein-coupled receptors,CysLT1 and CysLT2, which differ in ligand specificitiesand distribution (10). Cysteinyl leukotrienes have beenshown to have a prominent role in inflammatory dis-eases including asthma and inflammatory bowel dis-eases (6).

Aberrant expression of 5-LO has been detected invarious adult cancers (11–16). Moreover, the chemo-therapeutic and anticancer effects of 5-LO enzymepathway inhibitors have been demonstrated in experi-mental animal models (14, 17–19). Yet, little is knownabout whether the 5-LO enzyme pathway and leukotri-enes play a role in progression of tumors of neuralorigin.

The aim of this study was to assess the expression ofthe leukotriene enzyme pathway in childhood neuro-blastoma and in particular the effect of leukotrienes onneuroblastoma cell growth and survival.

MATERIALS AND METHODS

Human tissue samples

Tissue samples from tumors and nonmalignant adrenals wereobtained during surgery of neuroblastoma, snap-frozen inliquid nitrogen, and transferred to storage at �80°C forfuture analysis. Twenty-seven neuroblastoma samples derivedfrom children of different ages and all clinical stages, includ-ing different biological subsets (MYCN amplification, 7 of 27;1 p deletion, 9 of 27; Table 1) were analyzed. Three child-hood ganglioneuromas and 3 samples of nonmalignant adre-nals from children aged 12–25 months were also included.Ethical approval was obtained by the Karolinska UniversityHospital Research Ethics Committee (approval 03-308).

Immunohistochemistry

Formalin-fixed and paraffin-embedded tissue sections weredeparaffinized in xylene and graded alcohols, hydrated, andwashed in PBS. After antigen retrieval in sodium citrate buffer(pH 6) in a microwave oven, the endogenous peroxidase wasblocked by 0.3% H2O2 for 15 min. Sections were incubatedovernight at 4°C with a monoclonal mouse anti-5-LO anti-body (Research Diagnostics, Concord, MA, USA). As a sec-ondary antibody, the anti-mouse-horseradish peroxidase(HRP) SuperPicTure Polymer detection kit was used (ZymedLaboratories, San Francisco, CA, USA). For identification ofLTA4H and leukotriene receptors, sections were incubatedovernight at 4°C with rabbit polyclonal antibody againstLTA4H, CysLT1, CysLT2, and BLT1, respectively (CaymanChemicals, Ann Arbor, MI, USA). Immunodetection of LTC4synthase (LTC4S) and FLAP was performed by using rabbitpolyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz,CA, USA). Sections were subsequently washed and processedwith anti-rabbit-HRP SuperPicTure Polymer detection kit(Zymed). A matched isotype control was used as a control fornonspecific background staining.

Chemicals

AA-861, SC 22716, baicalein, caffeic acid, esculetin, and3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoleum bromide

Figure 1. Overview of pathways involved in biosynthesis ofleukotrienes. Arachidonic acid is released from cellular phos-pholipids by phospholipase A2 (PLA2) and converted by 5-LOand FLAP to LTA4. LTA4 is transformed by either LTA4H orLTC4S into two classes of compounds, LTB4 and cysteinylleukotrienes (LTC4, LTD4, and LTE4). Cysteinyl leukotrienesbind two distinct receptors, CysLT1 and CysLT2. LTB4 acts atspecific receptors that are designated BLT1 and BLT2 (GGT:�-glutamyl transpeptidase, DP: dipeptidase).

3526 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal

(MTT) were purchased from Sigma-Aldrich (Solna, Sweden).MK-886, Rev-5901, NDGA, and LY171883 were purchasedfrom BioMol (Plymouth Meeting, PA, USA). MK-571,LY255283, U-75302, and 5(S)-HETE were purchased fromCayman Chemicals. Montelukast sodium was a generous giftfrom Merck MSD (Rahway, NJ, USA). All reagents weredissolved in DMSO and further diluted in culture medium(final DMSO concentration always �0.5%). Arachidonic acidwas from Nu-Check Prep (Elysian, MN, USA). The pan-caspase inhibitor carbobenzoyloxy-Val-Ala-Asp (Ome) flu-oromethyl ketone (zVADfmk; R&D Systems, Abingdon, UK)was used at a 10 �M final concentration in culture medium.

Cell lines

The human neuroblastoma cell lines SH-SY5Y, SK-N-BE(2),SK-N-SH, SK-N-AS, SK-N-FI, SK-N-DZ, and IMR-32 were cul-tured as described previously (20). The human myelocyticcell line U937 was grown in RPMI 1640 using the samesupplement as above.

Reverse transcriptase-polymerase chain reaction(RT-PCR) analysis

RT-PCR analysis was performed as described previously (21).Modifications of PCR running conditions and nucleotidesequences of PCR primers used are shown in Table 2. Thehuman myelocytic cell line U937 was used as a positivecontrol.

Western blot analysis

Isolation of proteins was performed as described previously(20). Equal quantities were separated by SDS-PAGE, trans-ferred to nylon membranes (Millipore, Sundbyberg, Swe-den), and probed with antibodies against 5-LO (ResearchDiagnostics), FLAP (FL-161), and LTC4S (H-100) (SantaCruz Biotechnology), LTA4H, CysLT1, CysLT2 (CaymanChemicals), cleaved caspase-3, cleaved caspase-9 (Cell Sig-naling, Beverly, MA, USA), and �-actin (Sigma-Aldrich).Anti-mouse IgG or anti-rabbit IgG (Pharmacia Biosciences,Uppsala, Sweden) conjugated with HRP served as second-

TABLE 1. 5-LO, FLAP, LTA4H, LTC4S, BLT1, CysLT1, and CysLT2 expression in neuroblastoma, ganglioneuroma, andnonmalignant adrenals

Sample DiaAge

(mo) GenderINSSstage

MYCNampl.

1pdel

DNAploidy

Highriska Outcome

Survival(mo) 5-LO FLAP LTA4H LTC4S BLT1 CysLT1 CysLT2

1 NB 21 M 1 No No 4n No NED 157� � � � � � � �2 NB 123 F 1 No No 3n No NED 129� � � � � � � �3 NB 7 F 1 Yes Yes 2n No DOD 8 � � � � � � �4 NB 13 M 1 No No No NED 39� � � � � � � �5 NB 18 F 1 No No No NED 151� � � � � � � �6 NB 31 M 2B No No 3n No NED 164� � � � � � � �7 NB 33 F 2A No No 3n No NED 177� � � � � � � �8 NB 8 F 2 No No 3n No NED 66� � � � � � � �9 NB 110 M 2 No No 2n No NED 56� � � � � � � �

10 NB 5 F 2 No No 3n No AWD 56� � � � � � � �11 NB 103 F 2B No No 2n No NED 171� � � � � � � �12 NB 6 M 3 No ND 3n No NED 172� � � � � � � �13 NB 12 F 3 No No 5n No NED 155� � � � � � � �14 NB 0 M 3 No No 3n No DOC 0 – � � � � � �15 NB 79 M 3 Yes Yes 3n Yes NED 94� � � � � � � �16 NB 43 F 3 No No 3n No NED 63� � � � � � � �17 NB 136 M 4 Yes Yes 2n Yes DOD 10 � � � � � � �18 NB 39 F 4 Yes Yes 2n Yes DOD 9 � � � � � � �19 NB 35 F 4 No Yes Yes NED 55� � � � � � � �20 NB 28 M 4 Yes Yes 3n Yes NED 107� � � � � � � �21a NB 8 M 4M

b No No 3n No NED 57� � � � � � � �21b NB 8 M 4M No No 4n/5n No NED 57� � � � � � � �22 NB 41 F 4 No Yes 4n Yes NED 91� � � � � � � �23 NB 22 M 4 Yes Yes Yes DOD 44 � � � � � � �24 NB 50 F 4 Yes Yes Yes DOD 18 � � � � � � �25 NB 0 M 4S No No 3n No NED 150� � � � � � � �26 NB 10 M 4S No No 3n No NED 73� � � � � � � �27 NB 0 M 4S No ND 4n No NED 172� � � � � � � �28 GN 145 M NED 106� � � � � � � �29 GN 30 F AWD 90� � � � � � � �30 GN 59 F NED 73� � � � � � � �31 GN 137 M NED 120� � � � � � � �32 ADR 19 F –c – � � � � �33 ADR 25 F – – � � � � �34 ADR 12 F – – � � � � �

ADR, nonmalignant adrenal gland; AWD, alive with disease; dia, diagnosis; DOC, dead of surgical complications; DOD, dead of disease; GN,ganglioneuroma; INSS, International Neuroblastoma Staging System; NB, neuroblastoma, ND, not determined; NED, no evidence of disease.aPatient fulfilling clinico-biological criteria to obtain high-risk therapy. bMultifocal primary. cNegative for 5-LO and FLAP expression inmedullary cells but positive for cortical cells.

3527LEUKOTRIENE ENZYME PATHWAY IN NEUROBLASTOMA

ary antibodies. Pierce Super Signal (Pierce, Rockford, IL,USA) was used for detection.

Analysis of BLT1 by flow cytometry

Quantitation of the BLT1 receptor on neuroblastoma celllines was performed with a phycoerythrin (PE) -conjugatedmonoclonal anti-BLT1 antibody according to the manufactur-er’s instructions (R&D Systems). PE-conjugated mouse IgG1isotype control (R&D Systems) was used as a control fornonspecific background staining. After being labeled, cellswere washed twice in PBS supplemented with 0.5% bovineserum albumin and analyzed on the FL2 channel on aFACSCalibur flow cytometer, using Cell Quest Software (Becton-Dickinson, San Jose, CA, USA).

Analysis of leukotriene biosynthesis in primaryneuroblastomas and cell cultures by enzyme immunoassay

Five fresh primary tumor samples (Table 3) were suspendedin Ca2�- and Mg2�-free PBS together with a cocktail ofprotease inhibitors (Roche Diagnostics GmbH, Mannheim,Germany), homogenized, and sonicated 3 times for 10 s onice. Homogenates were incubated with 2 mM ATP, 2 mMCa2�, and 80 �M arachidonic acid for 10 min at 37°C andsubsequently quenched with an equal volume of methanol.The same procedure was used for analysis of leukotriene

production in neuroblastoma cell cultures. After acidificationto pH 3–4 (for LTB4 assay) or pH 5–6 (for cysteinyl leuko-triene assay), the samples were purified by solid-phase extrac-tion (Supelclean LC-18, Supelco, Sigma-Aldrich) and elutedin methanol. Samples were dried under nitrogen and resus-pended in enzyme immunoassay buffer. The concentration ofLTB4 and cysteinyl leukotrienes was determined with enzymeimmunoassay according to manufacturer�s protocol (CaymanChemicals).

Cell viability of neuroblastoma cells on leukotrienestimulation in vitro

The neuroblastoma cell lines SK-N-AS (1�104 cells/well) andSK-N-BE(2) (3�103cells/well) were seeded in 96-well cultureplates. Cells were cultivated for 24 h and then changed toserum-free RPMI for 24 h followed by replacement with thesame medium containing different concentrations of eitherLTB4 or LTD4. Cells were incubated further for 48, 72, or96 h with medium changes every second day. Cell viability wasassessed by the MTT assay.

Cytotoxicity assay and fluorescence-activatedcell-sorting analysis

Cells were incubated in 96-well culture plates with theindicated concentrations of drugs dissolved in OptiMEM

TABLE 2. Primers and conditions used in RT-PCR

Gene Primer sequence

PCR conditions

Size(bp)

cDNA(�l)

Annealing(°C)

Cycles(n)

5-LO 5�-TTCCTGGGCATGTACCAGAAGAGC-3� (exon14A) (sense) 8 65 40 3685�-GCAGTCCTGCTCTGTGTAGAATGGG-3� (exon 14B) (antisense)

FLAP 5�-GCTGCGTTTGCTGGACTGATGTA-3� (sense) 2 70 40 2245�-TAGAGGGGAGATGGTGGTGGAGAT-3� (antisense)

LTA4H 5�-ATGACCTATTGGTCCTGCCACCAT-3� (sense) 2 58 35 4365�-TCTGGTCCTCCAAGCAGTTGTTCA-3� (antisense)

LTC4S 5�-GAGTCCTGCTGCAAGCCTACTTC-3� (sense) 7 65 40 1195�-ACCTGGGCTCGGTAGAC-3� (antisense)

CYSLT1 5�-AGCCCCCACAAGACAATCAA-3� (sense) 2 58 40 3585�-AGGAGAGGGTCAAAGCAACAA-3� (antisense)

CYSLT2 5�-GCAACCATCCATCTCCGTATC-3� (sense) 2 58 40 3925�-CCAGGAAACGCACAACACTC-3� (antisense)

BLT1 5�-TAGTGCCCTGGAAAACGAAC-3� (sense) 8 61 40 2615�-CCTCAGCCAGGTTCACCAC-3� (antisense)

�-Actin 5�-TGACGGGGTCACCCACACTGTGCCCATCTA-3� (sense) 1 55 30 6255�-ACTCGTCATACTCCTGCTTGCTGATCCA-3� (antisense)

TABLE 3. Leukotriene production in primary neuroblastomas

DiaAge

(mo) GenderINSSstage

MYCNampl. 1p del

DNAploidy

Highriska Outcome

Survival(mo)

LTB4

(ng/g tissue)CysLT

(ng/g tissue)

NB 5 F 2 No No 3n No AWD 56� 1.01 26.72NB 6 M 3 No ND 3n No NED 172� 1.25 13.33NB 13 M 1 No No No NED 40� 5.75 16.89NB 8 M 2 No No 2n No DOD 11 1.72 16.80NB 12 M 1 No No No NED 33� 3.49 33.06PBS�AA 0.068 0.03

AA, arachidonic acid; AWD, alive with disease; dia, diagnosis; DOD, dead of disease; INSS, International Neuroblastoma Staging System;NB, neuroblastoma, ND, not determined; NED, no evidence of disease. aPatient fulfilling clinico-biological criteria to obtain high-risk therapy.

3528 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal

(Gibco-BRL, Sundbyberg, Sweden) supplemented with 2mM l-glutamine, 100 IU/ml penicillin G, and 100 �g/mlstreptomycin (Life Technologies, Carlsbad, CA, USA) for48 h. Cell viability was measured using the MTT assay.Mean values of optical density measurements were calcu-lated from six separate wells. To determine the effect ofexogenous addition of leukotrienes and 5(S)-HETE oncytotoxicity mediated by 5-LO inhibition, SK-N-BE(2) cells(1�104 cells/well) were cultivated with the 5-LO inhibitorAA-861 (6 �M) in the absence or presence of 125 nM LTB4,125 nM LTD4, or 250 nM 5(S)-HETE respectively, for 48 h.

The mitochondrial transmembrane potential was as-sessed in SH-SY5Y and SK-N-BE(2) after 24 h of incubationwith AA-861 (10 �M), MK-886 (7 �M), or montelukast (10�M), respectively, using tetramethylrhodamine ethyl ester(TMRE; Molecular Probes, OR, USA), as described previ-ously (20). DNA content was assessed by flow cytometryessentially as described previously (22). Briefly, SH-SY5Yand SK-N-BE(2) cells were treated with AA-861 (10 �M),MK-886 (7 �M), or montelukast (10 �M) for 48 h, har-vested, and stained with DAPI and subjected to cell cycleanalysis by using single-parameter DNA flow cytometry.Assessment of early apoptosis in NB cells treated as de-scribed above was done using annexin V-fluorescein iso-thiocyanate and monitored by flow cytometry as recom-mended by the manufacturers (Sigma-Aldrich).

5-LO and CysLT1 siRNA

SK-N-BE(2) cells were seeded in 6-well culture plates inRPMI medium at a 30 –50% confluence. Cells were trans-fected with target-specific 5-LO (sc-29596), CysLT1 (sc-43712), control (Fluorescein Conjugate)-A (sc-36869), orscrambled control (sc-37007) siRNA (Santa Cruz Biotech-nology), respectively, at a concentration of 33 nM usingLipofectamin 2000 (Invitrogen, Carlsbad, CA, USA) inOptiMEM. To evaluate cell viability and siRNA efficiency,Western blot analysis of protein extracts and trypan blueexclusion assay were performed 72 h after the initialtransfection. Transfection efficiency was assessed in SK-N-BE(2) cells transfected with control (Fluorescein Conju-gate)-A using flow cytometry.

For measurement of leukotriene production in transfectedcells, 107 SK-N-BE(2) cells were seeded in 15 cm Petri dishesand the cells were grown to 50% confluence before transfec-tion with 5-LO using Lipofectamine 2000 as described above.Measurement of leukotriene biosynthesis was performed 72 hafter the initial transfection, as described above.

Statistical analysis

Two-sided unpaired t-tests were performed to evaluate leuko-triene synthesis, proliferation, and cell viability, the data waslog transformed when indicated.

RESULTS

5-LO, FLAP, LTA4H, LTC4S, and leukotrienereceptors are expressed in neuroblastoma primarytumors and cell lines

We analyzed neuroblastomas from different biologi-cal subsets and all clinical stages for 5-LO expression

(Table 1). Twenty-six of 27 neuroblastoma samplesshowed specific expression of 5-LO protein in theperinuclear area and cytoplasm of the tumor cells(Fig. 2a). Nonmalignant adrenals from childrenshowed significant staining for 5-LO in the cortex,whereas the medulla, where neuroblastomas typicallyarise, showed no staining (Fig. 2a). Three ganglion-euromas were investigated and showed 5-LO immu-nopositivity in the tumor-derived ganglion cells butnot in the surrounding stroma (Fig. 2a). Leukotrienesynthesis is dependent not only on 5-LO but also onFLAP, which acts to present arachidonic acid re-leased from the cellular membranes to 5-LO, facili-tating leukotriene production (23). All human neu-roblastoma cell lines investigated showed both 5-LOand FLAP mRNA and protein expression as detectedby RT-PCR and Western blot (Fig. 2i, j). Further-more, all neuroblastoma tumors investigated wereshown to express FLAP in the perinuclear area oftumor cells (Fig. 2b). Expression of FLAP in theganglioneuroma and normal adrenal was consistentwith the expression of 5-LO (Fig. 2a, b).

LTA4H is widely distributed in most tissues andcells, and clearly it is not restricted to cells expressing5-LO (24, 25). Specific immunostaining for LTA4Hwas detected in the cytoplasm and nucleus of tumorcells and adjacent stromal cells. Ganglioneuromacells, as well as cortex and medullary regions of theadrenal, were also immunopositive (Fig. 2c). LTC4Sconjugates LTA4 with glutathione to form LTC4, theparent compound of CysLTs (26). Strong nuclearstaining for LTC4S was observed in a majority of cellswithin the tumor samples investigated (Fig. 2d).Ganglioneuroma cells were strongly immunopositive,whereas the cortical and medullary regions of thenonmalignant adrenals were weakly positive (Fig.2d). We further confirmed the expression of LTA4Hand LTC4S in neuroblastoma cell lines by RT-PCRand Western blotting (Fig. 2i, j).

Leukotrienes exert their effect through G-proteincoupled receptors. Hence, we analyzed the expres-sion of the leukotriene receptors BLT1, CysLT1, andCysLT2 in neuroblastoma primary tumors and celllines. CysLT1 receptor staining was evident in allclinical tumor specimens and in addition was de-tected in the vasculature of the adjacent stroma (Fig.2e). Similarly, the immunopositivity for BLT1 wasabundant in the cell membrane of the tumor cellsand in the tumor vasculature (Fig. 2g). Adrenal glandshowed expression of CysLT1, CysLT2, and BLT1 inboth the cortical and medullary compartments(Fig. 2f). RT-PCR revealed transcripts for BLT1,CysLT1, and CysLT2 in all neuroblastoma cell linesinvestigated (Fig. 2e– g). At the protein level, theleukotriene receptors CysLT1, CysLT2, and BLT1

were expressed in all neuroblastoma cell lines, asshown by Western blot and flow cytometry, respec-tively (Fig. 2j, k).

3529LEUKOTRIENE ENZYME PATHWAY IN NEUROBLASTOMA

Endogenous production of leukotrienes byneuroblastoma cells increases cell viability

Having established the expression of leukotriene path-way enzymes and receptors in neuroblastomas, weinvestigated endogenous production of leukotrienes inneuroblastoma cells. The leukotriene content in pri-mary tumors and neuroblastoma cell lysates was mea-sured by enzyme immunoassay. Incubation of cell ho-mogenates with 80 �M arachidonic acid resulted in asignificant leukotriene production in both primarytumors and cells (Table 3; Fig. 3a).

Moreover, given the high expression of leukotrienereceptors, we analyzed the survival rate of neuroblas-toma cells in response to exogenous leukotrienes. BothLTB4 and LTD4 significantly induced a concentration-dependent increase in cell viability of SK-N-AS cells(P�0.001; Fig. 3b). Interestingly, both leukotrienesinduced cell proliferation of SK-N-BE(2) cells(P�0.001; Fig. 3b). Taken together, these results indi-

cate the presence of a leukotriene-driven autocrinesurvival loop in neuroblastoma cells.

The rationale for using SK-N-BE(2) cells in themajority of the experiments in our study was based onthe fact that this cell line is a multiresistant, MYCN-amplified, p53-mutated neuroblastoma cell line thatdiffers from the majority of other neuroblastoma celllines in that it is considerably more resistant to mostconventional cancer therapeutic drugs. The other twocell lines SH-SY5Y and SK-N-AS do not exhibit anyMYCN-amplification or p53 mutations. On the otherhand, the SK-N-AS cell line is multiresistant whereasSH-SY5Y is not.

Leukotriene enzyme pathway inhibitors haveprofound effect on neuroblastoma cell growth

We proceeded to investigate whether inhibition of theleukotriene pathway could affect neuroblastoma cellgrowth and survival. Hence, we treated six neuroblas-

Figure 2. 5-LO pathway enzymes are expressed in neuroblastomaprimary tumors and cell lines. a–h) Immunohistochemistry showingspecific 5-LO (a), FLAP (b), LTA4H (c), LTC4S (d), CysLT1 (e),CysLT2 (f), and BLT1 (g) expression in tumor cells (arrow) of aneuroblastoma (left panels), differentiated ganglion cells (arrow) ina benign ganglioneuroma (middle panels), and nonmalignant adre-nal tissue showing positive staining in the cortical area (open arrow).The medullary compartments (closed arrow) were negative for 5-LOand FLAP but positive for LTA4H, LTC4S, CysLT1, and CysLT2 (rightpanels). h) Routin staining with hematoxylin/eosin (�400). i) RT-

PCR analysis of 5-LO, FLAP, LTA4H, LTC4S, CysLT1, CysLT2, and BLT1 in neuroblastoma cells. The human myelocytic cell lineU937 was used as a positive control. RT-PCR of �-actin confirmed the integrity of cDNA in all samples. j) Western blot detectedprotein bands of 78, 18, 69, 18, 38, and 58 kDa corresponding to 5-LO, FLAP, LTA4H, LTC4S, CysLT1, and CysLT2 respectively,in all neuroblastoma cell lines investigated. Blots were stripped and probed with �-actin to ensure equal protein loading. k)Analysis of BLT1 expression in neuroblastoma cells by flow cytometry. Black lines represent cells stained with BLT1 antibody,gray lines represent cells stained with isotype-matched control antibody.

3530 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal

toma cell lines with increasing concentrations of differ-ent LO inhibitors, a FLAP inhibitor, a LTA4H inhibitor,or leukotriene receptor antagonists. Irrespective of thedrug, the treatment resulted in a dose-dependent inhi-bition of neuroblastoma cell growth. The drug concen-tration needed to inhibit 50% of cell viability, EC50, isshown for the different drugs and cell lines in Table 4.The most effective drugs were MK-886, an inhibitor ofFLAP; montelukast, a CysLT1 receptor antagonist; andAA-861, an inhibitor of 5-LO. Inhibition of leukotrienefunction with montelukast was as effective in reducingneuroblastoma cell growth as either inhibition of 5-LOor FLAP (Table 4). Furthermore, exogenous additionof LTD4 and LTB4 or 5(S)-HETE to SK-N-BE(2) cellstreated with AA-861 (6 �M) was shown to significantlyinhibit the AA-861-mediated cytotoxicity (P�0.001; Fig.3c). Neither MYCN-amplification nor a drug-resistantphenotype had any effect on the sensitivity to LOpathway inhibitors (Table 4).

Leukotriene pathway inhibitors induce cell cyclearrest and apoptosis of neuroblastoma cells

Depolarization of the mitochondrial trans-membranepotential and the subsequent release of proapoptotic

factors are required for the induction of the intrinsicapoptotic pathway that is affected in aggressive neuro-blastomas (27). Treatment of SH-SY5Y and SK-N-BE(2)cells with AA-861, MK-886, or montelukast induceddepolarization of the mitochondrial membrane poten-tial (Fig. 4a) activation of caspase-9 and caspase-3 [(Fig.4b; similar data for SK-N-BE(2) not shown] followed byan accumulation of neuroblastoma cells in the sub-G1

phase of the cell cycle (Fig. 4c). Moreover, the broadcaspase inhibitor zVAD-fmk almost completely blockedthe cytotoxic effects of MK886, AA861, and monte-lukast, suggesting that all three drugs induced caspase-dependent apoptosis of SH-SY-5Y and SK-N-BE(2) cells(data not shown). Treatment of SH-SY5Y cells withMK-886 resulted in a 97% accumulation of cells insub-G1, whereas treatment with AA-861 or montelukastresulted in a 68 and 43% accumulation of cells withsub-G1 DNA content, respectively. For the p53-mutated,MYCN-amplified, multidrug-resistant cell line SK-N-BE(2), 80% of the cells showed a sub-G1 contentwhen treated with montelukast, whereas MK-886 andAA-861 induced prominent arrest of cells in G1 (Fig.4c). However, no accumulation of cells with a sub-G1

DNA content was observed with these drugs (Fig. 4c).

Figure 3. Leukotrienes are produced by neuroblastoma cells and increased cellviability in vitro. a) Leukotriene production by neuroblastoma cells. Neuroblastomacell homogenates of SK-N-BE(2) (open bar), SK-N-AS (gray bar), and SH-SY5Y (blackbar) were incubated with 80 �M arachidonic acid as described in Materials andMethods, and the concentration of produced LTB4 and the cysteinyl leukotrienes(CysLT) was measured using EIA. Statistical analysis was performed using unpairedt test on log transformed data; P � 0.05. Values are means � se for two independentexperiments. b) Effects of leukotrienes on neuroblastoma cell viability. Neuroblas-toma cells were incubated with different concentrations of LTB4 (31.2, 62.5, or 125nM) or LTD4 (31.2, 62.5, 125, or 250 nM) for 48, 72, or 96 h, and cell viability wasassessed using the MTT assay. Statistical significance was determined by unpaired ttest on log transformed data; P � 0.001 at 72 and 96 h. Data are expressed as mean �se percentage of untreated control at 48h. c) Effect of 5-LO metabolites on

AA-861-mediated cytotoxicity. SK-N-BE(2) cells were incubated with AA-861 (6 �M) alone or together with both LTB4 and LTD4(125 nM) or 5(S)-HETE (250 nM); cells were cultivated for 48 h. Cell viability was measured by the MTT assay. Statistical analysiswas performed using unpaired 2-sided t test; P � 0.05. Graph shows mean � se percentage of untreated control; values arerepresentative of two independent experiments.

3531LEUKOTRIENE ENZYME PATHWAY IN NEUROBLASTOMA

We therefore performed annexin V-staining to monitorearly apoptotic stages. SK-N-BE(2) cells treated withMK-886 or AA-861 demonstrated positive annexin V-staining that corresponded to the activation of caspases(Fig. 4b and data not shown). These results demon-strate that inhibitors of the 5-LO enzyme pathway andthe CysLT1 receptor antagonist induce a combinationof cell cycle arrest and apoptosis of neuroblastomacells, acting via the intrinsic apoptotic pathway. Tofurther confirm the role of 5-LO and leukotrienes inneuroblastoma growth, SK-N-BE(2) cells were trans-fected with 5-LO, CysLT1 or scrambled siRNA. Asshown in Fig. 4d, transfection with either 5-LO orCysLT1 siRNA resulted in a significant decrease in cellgrowth compared with cells transfected with scrambledsiRNA (P�0.05). Specific suppression of 5-LO andCysLT1 protein by siRNA transfection was verified byWestern blot (Fig. 4d) together with a significant de-crease in leukotriene production (P�0.05; Fig 4e).Transfection efficiency was shown to be 30% asevaluated using flourescein-labeled scrambled siRNAand flow cytometry analysis (data not shown).

DISCUSSION

The metabolism of arachidonic acid by the COX or LOpathways generates eicosanoids that have been impli-cated in the pathogenesis of a variety of human dis-eases, including cancer (28). Eicosanoid-producing en-zymes are over-expressed in adult epithelial cancersand result in production of a spectrum of mediatorsthat may be involved in resistance to apoptosis, induc-tion of metastasis, and angiogenesis (28, 29).

We have previously reported that COX-2 is expressedin human neuroblastoma and that inhibitors of COXhave significant effects in the treatment of neuroblas-toma in vivo (20, 30, 31). Furthermore, simultaneousinhibition of COX and LO by diclofenac and a pan-LO

inhibitor, NDGA, in the presence of arachidonic acidinhibited cell viability more potently than inhibitingeither pathway alone (20). The current study wasdesigned to investigate the role of the 5-LO enzymepathway in neuroblastoma growth and survival. Weanalyzed neuroblastoma primary tumors from differentbiological subsets and all clinical stages and detectedhigh 5-LO expression in 26 out of 27 samples. 5-LOexpression was also detected in tumor cells of the moredifferentiated ganglioneuromas (Fig. 2a) but not in thenonmalignant adrenal medullas, where neuroblasto-mas typically arise (Fig. 2a).

The significance of the expression of 5-LO in cells ofneuronal origin is not well understood. It has beenshown that the expression of 5-LO mRNA and proteinwas higher in proliferating immature cortical neuronsthan in mature cultures (32). Furthermore, the pres-ence of 5-LO seems to be essential for maintainingnormal proliferation of immature neuroblasts, as treat-ment with 5-LO antisense or 5-LO inhibitors reducedcell number and proliferation (32, 33). LO activity hasalso been reported to be involved in cell cycle progres-sion of neuroblastoma cells (34). Together with ourresults these data indicate that 5-LO is particularlyexpressed in proliferating immature neuronal cells.

Aberrant expression of 5-LO has been detected invarious adult cancers (11–16). Furthermore, polymor-phisms in the promoter region of the 5-LO gene areassociated with a lower risk of colon cancer (35). Arecent study (36) suggests that 5-LO may be a markerfor early pancreatic intraepithelial neoplastic lesions. Inour study, we could not detect any differences in 5-LOexpression between different stages and biological sub-sets of neuroblastomas. Also, results from analysis ofleukotriene production in primary tumor samplescould not be related to different biological stages oftumors (Table 3).

Expression of FLAP was consistent with the expres-

TABLE 4. EC50 values of LO-pathway inhibitors on neuroblastoma cell viability in vitro

Drug Drug target

EC50 (�M)

SK-N-BE(2)a,b IMR-32a SH-SY5Y SK-N-ASb SK-N-SHb SK-N-FIb

AA-861 5-LO 10–12 10–12 8–10 12–14 12–14 8–10MK-886 FLAP 3–4 2–3 3–4 4–5 4–5 4–5NDGA pan LO 20–25 5–10 5–10 15–20 5–10 15–20Caffeic acid 5-LO/15-LO 200 200 200 200 200 200Esculetin 5-LO/12-LO 50–100 10–25 50–100 50–100 50–100 25–50Baicalein 12-LO 100–200 50–100 100–200 100–200 200 200REV-5901 CysLT1

c/5-LO 20–25 25–50 15–20 25–50 25–50 25–50Montelukast CysLT1

c /5-LO 4–8 4–10 2–8 6–8 2–8 4–10MK 571 CysLT1

c 30–34 20–34 26–30 40–50 34–40 40–50LY171883 CysLT1

c 100–125 125–200 100–125 100–125 125–200 200SC 22716 LTA4H 5–25 20–30 20–30 20–30 30–40 40–50U-75302 BLT1

c 30–40 20–30 20–30 30–40 40–50 40–50LY2552833 BLT2

c 50–60 30–40 30–40 50–60 50–60 70–80

Cells were treated with different LO-pathway inhibitors at a range of concentrations for 48 h. Effective concentration 50% (EC50) valueswere estimated from repeated experiments and lie within the range shown. aMYCN amplification. bMultidrug-resistant phenotype. cReceptorantagonist.

3532 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal

sion of 5-LO in neuroblastoma samples, ganglioneu-roma, and adrenal cortex (Fig. 2b), and both 5-LO andFLAP mRNA and protein were detected in all neuro-blastoma cell lines investigated (Fig. 2i, j). The data onFLAP expression in tumor cells are limited. FLAPexpression has been shown to be up-regulated inesophageal squamous cell carcinoma, and a recentstudy demonstrated an association between FLAP ex-pression and poor prognosis in breast cancer (37, 38).Although the mechanism of action for FLAP has notbeen fully elucidated, it has been suggested that its roleis to present arachidonic acid to 5-LO. FLAP was shownto bind arachidonic acid, and this binding could becompeted by compounds like MK-886 (39). FLAP alsostimulates the utilization of arachidonic acid by 5-LOand increases the efficiency by which 5-LO converts5-HPETE into LTA4 (40). The requirement of FLAP for5-LO function was demonstrated when human osteo-sarcoma cells transfected with 5-LO required a concom-

itant transfection with FLAP for ionophore-inducedleukotriene production (41).

LTA4 is an unstable epoxide intermediate fromwhich all other leukotrienes are synthesized along twomajor metabolic pathways, the first one involving enzy-matic hydrolysis by LTA4H, which leads to the produc-tion of LTB4. The second metabolic pathway involvesconjugation of LTA4 with gluthathione by LTC4S toform LTC4, the parent compound of cysteinyl leukotri-enes (Fig. 1; ref. 26). As shown by immunohistochem-istry, strong expression of both LTA4H and LTC4S wasobserved in neuroblastoma cells, in ganglioneuroma,and in the adrenal gland (Fig. 2c, d). Furthermore,both enzymes were expressed in all neuroblastoma celllines investigated, as shown by RT-PCR and Westernblot (Fig. 2i, j). The significance of 5-LO enzymepathway expression in the adrenal gland is unclear,although it has been previously reported and suggestedthat 5-LO metabolites may play a role in hormone

Figure 4. 5-LO enzyme pathway inhibitors induce cellcycle arrest and apoptosis of neuroblastoma cells. a)Effect of 5-LO enzyme pathway inhibitors and leuko-triene receptor antagonist on the mitochondrialtransmembrane potential in neuroblastoma cells. SK-N-BE(2) or SH-SY5Y cells were treated for 24 h withAA-861 (10 �M), MK-886 (7 �M), or montelukast (10�M), respectively, and the change in mitochondrialtransmembrane potential was measured by loss of

TMRE fluorescence signal assessed by flow cytometry. Dark lines represent treated cells; gray lines represent untreated controlcells. b) Analysis of proteins in the apoptotic pathways by Western blotting. Treatment of SH-SY5Y cells with AA-861, MK-886, ormontelukast activates caspase-9 and caspase-3. Arrows indicate specific cleavage products. The blot was stripped and probed with�-actin to ensure equal protein loading. c) Effects of 5-LO enzyme pathway inhibitors and leukotriene receptor antagonist onthe cell cycle distribution of neuroblastoma cells. Flow cytometry (DAPI) showing sub-G1 accumulation and cell cycledistribution of SK-N-BE(2) or SH-SY5Y cells exposed to AA-861 (10 �M), MK-886 (7 �M), or montelukast (10 �M) for 48 h. d)Left: siRNA-mediated suppression of 5-LO and CysLT1 expression reduced viability of SK-N-BE(2) cells (P�0.05; t test, bothtreatments). Data are expressed as mean � se for two independent experiments. Right: representative Western blots showing5-LO, CysLT1, and �-actin in SK-N-BE(2) cells transfected with either 5-LO, CysLT1, or scrambled siRNA for 72 h. e)Measurement of leukotriene production in neuroblastoma cells transfected with 5-LO siRNA. Cell homogenates were incubatedwith or without 80 �M of AA before LTB4, and CysLT production was measured using EIA. Results are expressed as mean �se of two independent experiments. Statistical analysis was performed using 2-sided unpaired t test. *P � 0.05; **P � 0.01.Western blots and flow cytometric histograms are representative of two independent experiments.

3533LEUKOTRIENE ENZYME PATHWAY IN NEUROBLASTOMA

production and secretion in adrenal cortical cells (42–45).

After having shown that 5-LO pathway enzymes werehighly expressed in neuroblastomas, we wanted toexamine whether the cells were able to produce leuko-trienes. Primary neuroblastomas were shown to pro-duce leukotrienes (Table 3). When neuroblastomacultures were incubated in the presence of arachidonicacid, significant levels of leukotrienes were produced inall three cell lines investigated, indicating that theleukotriene synthesis is a general feature of neuroblas-toma cells (Fig. 3a). Biosynthesis of leukotrienes mayalso involve transport of LTA4 from the cytoplasm ofthe cell into another cell containing LTA4H and LTC4Sby a process called transcellular biosynthesis (46). Inthe context of the tumor environment, LTA4 couldtheoretically be provided by infiltrating inflammatorycells and further converted into leukotrienes by neuro-blastoma cells. However, the fact that the leukotrienepathway enzymes are present in neuroblastoma cellsand leukotriene production can be detected in bothprimary tumors and cells in vitro clearly demonstratesendogenous leukotriene production.

Leukotrienes possess a wide variety of effects attrib-uted to tumor growth such as increased vascular per-meability, chemotaxis, and cell adhesion, as well as cellsurvival and proliferation (47–49). In addition, a rolefor leukotrienes in neoangiogenesis has been sug-gested, because they have been shown to increasemigration and viability of endothelial cells (6, 50). Theeffects of cysteinyl leukotrienes are mediated throughCysLT1 and CysLT2, whereas BLT1 and BLT2 are acti-vated by LTB4 (9, 10). A third receptor for LTB4 is thenuclear peroxisome proliferator-activated receptor �(PPAR-�), involved in termination of inflammatoryresponses (51). Increased expression of leukotrienereceptors in human tumors has previously been re-ported and was shown to negatively correlate withpatient survival (11, 16, 52). In this study, we demon-strate the presence of CysLT1, CysLT2, and BLT1 inneuroblastoma cell lines and tumors (Fig. 2e–g). BothLTB4 and LTD4 induced a survival and concentration-dependent proliferative response in neuroblastomacells (Fig. 3b). Our results demonstrate that these lipidmediators can influence the growth and survival ofneuroblastoma cells in an autocrine and/or paracrinemanner. Other studies (18, 53, 54) suggest that themechanisms underlying increased proliferation andcell survival response induced by leukotrienes involveactivation of extracellular related kinase 1/2 and phos-phoinositide 3-kinase and increased expression of anti-apoptotic proteins such as BCL-2.

In vitro inhibition of the 5-LO enzyme pathway in avariety of cancer cells results in growth inhibition andapoptosis (13, 17, 18, 55). To investigate whether theleukotriene signaling pathway represents a possibletherapeutic target in neuroblastomas, we assessed theeffect of different inhibitors of the 5-LO enzyme path-way and leukotriene receptor antagonists on a panel ofneuroblastoma cells in vitro. Out of 13 different drugs

investigated, the 5-LO inhibitor AA-861, the FLAPinhibitor MK-886, and the CysLT1 receptor antagonistmontelukast were the most effective in reducing neu-roblastoma cell growth. The drug concentration thatwas needed to inhibit 50% of cell viability (EC50)ranged from 2–14 �M for these drugs (Table 4). Thecytotoxic effect of AA-861 on neuroblastoma cells wasinhibited by addition of leukotrienes or 5(S)-HETE tothe cultures (Fig. 3c), indicating the importance of5-LO metabolites in neuroblastoma growth and sur-vival.

Cell death by apoptosis after inhibition of the 5-LOenzyme pathway in adult epithelial cancer cells hasbeen reported, and specific 5-LO enzyme pathwayinhibitors have been shown to be effective in experi-mental tumor models in vivo (14, 17, 18, 55). Weobserved that treatment of neuroblastoma cells withAA-861, MK-886, or montelukast resulted in a depolar-ization of the mitochondrial transmembrane potentialand subsequent induction of apoptosis (Fig. 4a, b). Thethree different drugs have different primary targets inthe leukotriene pathway, and their cytotoxic effect overtime is different. Therefore, a distinct caspase cleavagepattern was not unexpected. Hence, we monitoredinduction of early apoptosis by annexin V staining.Treatment of NB cells with MK-886 or AA-861 resultedin positive annexin V staining. These data together withthe measurements of the mitochondrial transmem-brane potential, activation of caspases, and cell cycleanalysis demonstrate that inhibitors of the 5-LO path-way induce apoptosis of NB cells (Fig. 4a–c and datanot shown). The fact that a transient siRNA transfec-tion was less potent in neuroblastoma cell growthinhibition compared with chemical inhibitors or leuko-triene receptor antagonist may be due to the lowtransfection efficiency of the siRNA constructs used inour study or the existence of a 5-LO enzyme pathway-independent mechanism of the drugs (Fig. 4d, e).

Montelukast had a significant cytotoxic effect onneuroblastoma cells in vitro (Table 4; Fig. 4). In addi-tion to its function as a CysLT1 receptor antagonist,montelukast may also have a direct inhibitory effect on5-LO at therapeutically relevant concentrations (56).Moreover, inhibitory effects of montelukast on nuclearfactor-k� activity (NF- B) and VEGF expression hasbeen reported (57, 58). Therefore, it cannot be ex-cluded that targets of montelukast other than CysLT1may be involved in the tumor cell growth inhibition. Ofparticular interest for a potential future application ofmontelukast for neuroblastoma therapy is the fact thatthe drug is available for oral clinical treatment ofpediatric asthma and the extraordinary tolerabilityshown in clinical use (59, 60).

In comparison with nonmalignant nervous tissue,neuroblastomas contain increased levels of arachidonicacid, the main substrate for eicosanoid biosynthesiscatalyzed by COX-2 and 5-LO (4). Interestingly, cystei-nyl leukotrienes and prostaglandins are transportedout of producing cells by MRP1 and MRP4, respectively.In addition to inhibiting prostaglandin synthesis, some

3534 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal

nonsteroidal anti-inflammatory drugs may inhibitMRP4 directly (61). Similarly, MRP1 can be inhibitedby leukotriene receptor antagonists (62). Expression ofboth MRP1 and MRP4 is correlated to MYCN expres-sion and malignancy in neuroblastoma and confersresistance to chemotherapy (63, 64). Hence, the leuko-triene pathway inhibitors may potentiate the effect ofchemotherapeutic agents.

Our results demonstrate for the first time that a fullyactive and functional leukotriene synthesis pathway ispresent in childhood neuroblastoma and that neuro-blastoma cells produce leukotrienes that can promotesurvival in an autocrine manner. Inhibition of theleukotriene pathway also induced apoptosis of neuro-blastoma cells in vitro. Our findings provide new insightinto the pathobiology of neuroblastoma. Pharmacolog-ical interventions that target the leukotriene signalingpathway may be an important adjuvant therapy forchildren with neuroblastoma, but further preclinical invivo studies are warranted.

We thank Lotta Elfman, Anja Pedersen, and Roy Lysaa fortechnical assistance. This work was supported by the SwedishResearch Council, the Swedish Children Cancer Foundation,the Swedish Cancer Society, the Stockholm Cancer Society(Sweden), the Wenner-Gren Foundation (Stockholm, Swe-den), EC FP6 (LSHM-CT-2004–005033), and the AakreFoundation (University of Tromso, Tromso, Norway).

REFERENCES

1. Coussens, L. M., and Werb, Z. (2002) Inflammation and cancer.Nature 420, 860–867

2. Brodeur, G. M. (2003) Neuroblastoma: biological insights into aclinical enigma. Nat. Rev. Cancer 3, 203–216

3. Gasparini, G., Longo, R., Sarmiento, R., and Morabito, A.(2003) Inhibitors of cyclo-oxygenase 2: a new class of anticanceragents? Lancet Oncol. 4, 605–615

4. Reynolds, L. M., Dalton, C. F., and Reynolds, G. P. (2001)Phospholipid fatty acids and neurotoxicity in human neuroblas-toma SH-SY5Y cells. Neurosci. Lett. 309, 193–196

5. Samuelsson, B. (1983) Leukotrienes: mediators of immediatehypersensitivity reactions and inflammation. Science 220, 568–575

6. Funk, C. D. (2001) Prostaglandins and leukotrienes: advances ineicosanoid biology. Science 294, 1871–1875

7. Janssen-Timmen, U., Vickers, P. J., Wittig, U., Lehmann, W. D.,Stark, H. J., Fusenig, N. E., Rosenbach, T., Radmark, O.,Samuelsson, B., and Habenicht, A. J. (1995) Expression of5-lipoxygenase in differentiating human skin keratinocytes. Proc.Natl. Acad. Sci. U. S. A. 92, 6966–6970

8. Luo, M., Lee, S., and Brock, T. G. (2003) Leukotriene synthesisby epithelial cells. Histol. Histopathol. 18, 587–595

9. Tager, A. M., and Luster, A. D. (2003) BLT1 and BLT2: theleukotriene B(4) receptors. Prostaglandins Leukot. Essent. FattyAcids 69, 123–134

10. Kanaoka, Y., and Boyce, J. A. (2004) Cysteinyl leukotrienes andtheir receptors: cellular distribution and function in immuneand inflammatory responses. J. Immunol. 173, 1503–1510

11. Hennig, R., Ding, X. Z., Tong, W. G., Schneider, M. B., Standop,J., Friess, H., Buchler, M. W., Pour, P. M., and Adrian, T. E.(2002) 5-Lipoxygenase and leukotriene B(4) receptor are ex-pressed in human pancreatic cancers but not in pancreaticducts in normal tissue. Am. J. Pathol. 161, 421–428

12. Gupta, S., Srivastava, M., Ahmad, N., Sakamoto, K., Bostwick,D. G., and Mukhtar, H. (2001) Lipoxygenase-5 is overexpressedin prostate adenocarcinoma. Cancer 91, 737–743

13. Hoque, A., Lippman, S. M., Wu, T. T., Xu, Y., Liang, Z. D.,Swisher, S., Zhang, H., Cao, L., Ajani, J. A., and Xu, X. C. (2005)Increased 5-lipoxygenase expression and induction of apoptosisby its inhibitors in esophageal cancer: a potential target forprevention. Carcinogenesis 26, 785–791

14. Avis, I. M., Jett, M., Boyle, T., Vos, M. D., Moody, T., Treston,A. M., Martinez, A., and Mulshine, J. L. (1996) Growth controlof lung cancer by interruption of 5-lipoxygenase-mediatedgrowth factor signaling. J. Clin. Invest. 97, 806–813

15. Boado, R. J., Pardridge, W. M., Vinters, H. V., and Black, K. L.(1992) Differential expression of arachidonate 5-lipoxygenasetranscripts in human brain tumors: evidence for the expressionof a multitranscript family. Proc. Natl. Acad. Sci. U. S. A. 89,9044–9048

16. Ohd, J. F., Nielsen, C. K., Campbell, J., Landberg, G., Lofberg,H., and Sjolander, A. (2003) Expression of the leukotriene D4receptor CysLT1, COX-2, and other cell survival factors incolorectal adenocarcinomas. Gastroenterology 124, 57–70

17. Hennig, R., Ding, X. Z., Tong, W. G., Witt, R. C., Jovanovic,B. D., and Adrian, T. E. (2004) Effect of LY293111 in combina-tion with gemcitabine in colonic cancer. Cancer Lett. 210, 41–46

18. Tong, W. G., Ding, X. Z., Witt, R. C., and Adrian, T. E. (2002)Lipoxygenase inhibitors attenuate growth of human pancreaticcancer xenografts and induce apoptosis through the mitochon-drial pathway. Mol. Cancer Ther. 1, 929–935

19. Gunning, W. T., Kramer, P. M., Steele, V. E., and Pereira, M. A.(2002) Chemoprevention by lipoxygenase and leukotrienepathway inhibitors of vinyl carbamate-induced lung tumors inmice. Cancer Res. 62, 4199–4201

20. Johnsen, J. I., Lindskog, M., Ponthan, F., Pettersen, I., Elfman,L., Orrego, A., Sveinbjornsson, B., and Kogner, P. (2004)Cyclooxygenase-2 is expressed in neuroblastoma, and nonste-roidal anti-inflammatory drugs induce apoptosis and inhibittumor growth in vivo. Cancer Res. 64, 7210–7215

21. Pettersen, I., Figenschau, Y., Olsen, E., Bakkelund, W., Smed-srod, B., and Sveinbjornsson, B. (2002) Tumor necrosis factor-related apoptosis-inducing ligand induces apoptosis in humanarticular chondrocytes in vitro. Biochem. Biophys. Res. Commun.296, 671–676

22. Ponthan, F., Johnsen, J. I., Klevenvall, L., Castro, J., and Kogner,P. (2003) The synthetic retinoid RO 13–6307 induces neuro-blastoma differentiation in vitro and inhibits neuroblastomatumour growth in vivo. Int. J. Cancer 104, 418–424

23. Miller, D. K., Gillard, J. W., Vickers, P. J., Sadowski, S., Leveille,C., Mancini, J. A., Charleson, P., Dixon, R. A., Ford-Hutchinson,A. W., Fortin, R., Gauthier, J. Y., Rodkey, J., Rosen, R., Rouzer,C., Sigal, I. S., Strader, C. D., and Evans, J. F. (1990) Identifica-tion and isolation of a membrane protein necessary for leuko-triene production. Nature 343, 278–281

24. Chen, X., Li, N., Wang, S., Wu, N., Hong, J., Jiao, X., Krasna,M. J., Beer, D. G., and Yang, C. S. (2003) Leukotriene A4hydrolase in rat and human esophageal adenocarcinomas andinhibitory effects of bestatin. J. Natl. Cancer Inst. 95, 1053–1061

25. Izumi, T., Shimizu, T., Seyama, Y., Ohishi, N., and Takaku, F.(1986) Tissue distribution of leukotriene A4 hydrolase activityin guinea pig. Biochem. Biophys. Res. Commun. 135, 139–145

26. Lam, B. K. (2003) Leukotriene C(4) synthase. ProstaglandinsLeukot. Essent. Fatty Acids 69, 111–116

27. Abel, F., Sjoberg, R. M., Nilsson, S., Kogner, P., and Martinsson,T. (2005) Imbalance of the mitochondrial pro- and anti-apo-ptotic mediators in neuroblastoma tumours with unfavourablebiology. Eur. J. Cancer 41, 635–646

28. Romano, M., and Claria, J. (2003) Cyclooxygenase-2 and 5-li-poxygenase converging functions on cell proliferation andtumor angiogenesis: implications for cancer therapy. FASEB J.17, 1986–1995

29. Nie, D., and Honn, K. V. (2004) Eicosanoid regulation ofangiogenesis in tumors. Semin. Thromb. Hemost. 30, 119–125

30. Johnsen, J. I., Lindskog, M., Ponthan, F., Pettersen, I.,Elfman, L., Orrego, A., Sveinbjornsson, B., and Kogner, P.(2005) NSAIDs in neuroblastoma therapy. Cancer Lett. 228,195–201

31. Ponthan F., Wickstrom, M., Gleissman, H., Fuskevåg, O. M.,Segerstrom, L., Sveinbjornsson, B., Redfern, C. P., Eksborg, S.,Kogner, P., and Johnsen, J. I. (2007) Celecoxib prevents neuro-blastoma tumor development and potentiates the effect of

3535LEUKOTRIENE ENZYME PATHWAY IN NEUROBLASTOMA

chemotherapeutic drugs in vitro and in vivo. Clin. Cancer Res. 13,1036–1044

32. Uz, T., Manev, R., and Manev, H. (2001) 5-Lipoxygenase isrequired for proliferation of immature cerebellar granule neu-rons in vitro. Eur. J. Pharmacol. 418, 15–22

33. Wada, K., Arita, M., Nakajima, A., Katayama, K., Kudo, C.,Kamisaki, Y., and Serhan, C. N. (2006) Leukotriene B4 andlipoxin A4 are regulatory signals for neural stem cell prolifera-tion and differentiation. FASEB J. 20, 1785–1792

34. Van Rossum, G. S., Bijvelt, J. J., van den Bosch, H., Verkleij, A. J.,and Boonstra, J. (2002) Cytosolic phospholipase A2 and lipoxy-genase are involved in cell cycle progression in neuroblastomacells. Cell. Mol. Life Sci. 59, 181–188

35. Goodman, J. E., Bowman, E. D., Chanock, S. J., Alberg, A. J., andHarris, C. C. (2004) Arachidonate lipoxygenase (ALOX) andcyclooxygenase (COX) polymorphisms and colon cancer risk.Carcinogenesis 25, 2467–2472

36. Hennig, R., Grippo, P., Ding, X. Z., Rao, S. M., Buchler, M. W.,Friess, H., Talamonti, M. S., Bell, R. H., and Adrian, T. E. (2005)5-Lipoxygenase, a marker for early pancreatic intraepithelialneoplastic lesions. Cancer Res. 65, 6011–6016

37. Zhi, H., Zhang, J., Hu, G., Lu, J., Wang, X., Zhou, C., Wu, M.,and Liu, Z. (2003) The deregulation of arachidonic acidmetabolism-related genes in human esophageal squamous cellcarcinoma. Int. J. Cancer 106, 327–333

38. Jiang, W. G., Douglas-Jones, A. G., and Mansel, R. E. (2006)Aberrant expression of 5-lipoxygenase-activating protein (5-LOXAP) has prognostic and survival significance in patients withbreast cancer. Prostaglandins Leukot. Essent. Fatty Acids 74, 125–134

39. Mancini, J. A., Abramovitz, M., Cox, M. E., Wong, E., Charleson,S., Perrier, H., Wang, Z., Prasit, P., and Vickers, P. J. (1993)5-lipoxygenase-activating protein is an arachidonate bindingprotein. FEBS Lett. 318, 277–281

40. Abramovitz, M., Wong, E., Cox, M. E., Richardson, C. D., Li, C.,and Vickers, P. J. (1993) 5-lipoxygenase-activating protein stim-ulates the utilization of arachidonic acid by 5-lipoxygenase. Eur.J. Biochem. 215, 105–111

41. Dixon, R. A., Diehl, R. E., Opas, E., Rands, E., Vickers, P. J.,Evans, J. F., Gillard, J. W., and Miller, D. K. (1990) Requirementof a 5-lipoxygenase-activating protein for leukotriene synthesis.Nature 343, 282–284

42. Hirai, A., Tahara, K., Tamura, Y., Saito, H., Terano, T., andYoshida, S. (1985) Involvement of 5-lipoxygenase metabolites inACTH-stimulated corticosteroidogenesis in rat adrenal glands.Prostaglandins 30, 749–767

43. Jones, D. B., Marante, D., Williams, B. C., and Edwards, C. R.(1987) Adrenal synthesis of corticosterone in response to ACTHin rats is influenced by leukotriene A4 and by lipoxygenaseintermediates. J. Endocrinol. 112, 253–258

44. Omura, M., Hirai, A., Tamura, Y., and Yoshida, S. (1990)Transformation of arachidonic acid by 5- and 15-lipoxygenasepathways in bovine adrenal fasciculata cells. Prostaglandins Leu-kot. Essent. Fatty Acids 40, 93–102

45. Schroder, O., Sjostrom, M., Qiu, H., Jakobsson, P. J., andHaeggstrom, J. Z. (2005) Microsomal glutathione S-transferases:selective up-regulation of leukotriene C4 synthase during li-popolysaccharide-induced pyresis. Cell Mol. Life Sci. 62, 87–94

46. Folco, G., and Murphy, R. C. (2006) Eicosanoid transcellularbiosynthesis: from cell-cell interactions to in vivo tissue re-sponses. Pharmacol. Rev. 58, 375–388

47. Dahlen, S. E., Bjork, J., Hedqvist, P., Arfors, K. E., Ham-marstrom, S., Lindgren, J. A., and Samuelsson, B. (1981)Leukotrienes promote plasma leakage and leukocyte adhesionin postcapillary venules: in vivo effects with relevance to theacute inflammatory response. Proc. Natl. Acad. Sci. U. S. A. 78,3887–3891

48. Palmblad, J., Malmsten, C. L., Uden, A. M., Radmark, O.,Engstedt, L., and Samuelsson, B. (1981) Leukotriene B4 is apotent and stereospecific stimulator of neutrophil chemotaxisand adherence. Blood 58, 658–661

49. Paruchuri, S., and Sjolander, A. (2003) Leukotriene D4 medi-ates survival and proliferation via separate but parallel pathwaysin the human intestinal epithelial cell line Int 407. J. Biol. Chem.278, 45577–45585

50. Bogatcheva, N. V., Sergeeva, M. G., Dudek, S. M., and Verin,A. D. (2005) Arachidonic acid cascade in endothelial pathobi-ology. Microvasc. Res. 69, 107–127

51. Devchand, P. R., Keller, H., Peters, J. M., Vazquez, M., Gonzalez,F. J., and Wahli, W. (1996) The PPARalpha-leukotriene B4pathway to inflammation control. Nature 384, 39–43

52. Matsuyama, M., Hayama, T., Funao, K., Kawahito, Y., Sano, H.,Takemoto, Y., Nakatani, T., and Yoshimura, R. (2007) Overex-pression of cysteinyl LT1 receptor in prostate cancer andCysLT1R antagonist inhibits prostate cancer cell growththrough apoptosis. Oncol. Rep. 18, 99–104

53. Paruchuri, S., Broom, O., Dib, K., and Sjolander, A. (2005) Thepro-inflammatory mediator leukotriene D4 induces phosphati-dylinositol 3-kinase and Rac-dependent migration of intestinalepithelial cells. J. Biol. Chem. 280, 13538–13544

54. Paruchuri, S., Hallberg, B., Juhas, M., Larsson, C., and Sjolan-der, A. (2002) Leukotriene D(4) activates MAPK through aRas-independent but PKCepsilon-dependent pathway in intesti-nal epithelial cells. J. Cell Sci. 115, 1883–1893

55. Ghosh, J., and Myers, C. E. (1998) Inhibition of arachidonate5-lipoxygenase triggers massive apoptosis in human prostatecancer cells. Proc. Natl. Acad. Sci. U. S. A. 95, 13182–13187

56. Ramires, R., Caiaffa, M. F., Tursi, A., Haeggstrom, J. Z., andMacchia, L. (2004) Novel inhibitory effect on 5-lipoxygenaseactivity by the anti-asthma drug montelukast. Biochem. Biophys.Res. Commun. 324, 815–821

57. Maeba, S., Ichiyama, T., Ueno, Y., Makata, H., Matsubara, T.,and Furukawa, S. (2005) Effect of montelukast on nuclear factorkappaB activation and proinflammatory molecules. Ann. AllergyAsthma Immunol. 94, 670–674

58. Lee, K. S., Kim, S. R., Park, H. S., Jin, G. Y., and Lee, Y. C. (2004)Cysteinyl leukotriene receptor antagonist regulates vascularpermeability by reducing vascular endothelial growth factorexpression. J. Allergy Clin. Immunol. 114, 1093–1099

59. Migoya, E., Kearns, G. L., Hartford, A., Zhao, J., van Adelsberg,J., Tozzi, C. A., Knorr, B., and Deutsch, P. (2004) Pharmacoki-netics of montelukast in asthmatic patients 6 to 24 months old.J. Clin. Pharmacol. 44, 487–494

60. Knorr, B., Maganti, L., Ramakrishnan, R., Tozzi, C. A., Migoya,E., and Kearns, G. (2006) Pharmacokinetics and safety ofmontelukast in children aged 3 to 6 months. J. Clin. Pharmacol.46, 620–627

61. Reid, G., Wielinga, P., Zelcer, N., van der Heijden, I., Kuil, A., deHaas, M., Wijnholds, J., and Borst, P. (2003) The humanmultidrug resistance protein MRP4 functions as a prostaglandinefflux transporter and is inhibited by nonsteroidal antiinflam-matory drugs. Proc. Natl. Acad. Sci. U. S. A. 100, 9244–9249

62. Gekeler, V., Ise, W., Sanders, K. H., Ulrich, W. R., and Beck, J.(1995) The leukotriene LTD4 receptor antagonist MK571 spe-cifically modulates MRP associated multidrug resistance. Bio-chem. Biophys. Res. Commun. 208, 345–352

63. Peaston, A. E., Gardaneh, M., Franco, A. V., Hocker, J. E.,Murphy, K. M., Farnsworth, M. L., Catchpoole, D. R., Haber, M.,Norris, M. D., Lock, R. B., and Marshall, G. M. (2001) MRP1gene expression level regulates the death and differentiationresponse of neuroblastoma cells. Br. J. Cancer 85, 1564–1571

64. Norris, M. D., Smith, J., Tanabe, K., Tobin, P., Flemming, C.,Scheffer, G. L., Wielinga, P., Cohn, S. L., London, W. B.,Marshall, G. M., Allen, J. D., and Haber, M. (2005) Expressionof multidrug transporter MRP4/ABCC4 is a marker of poorprognosis in neuroblastoma and confers resistance to irinotecanin vitro. Mol. Cancer Ther. 4, 547–553

Received for publication December 4, 2007.Accepted for publication May 15, 2008.

3536 Vol. 22 October 2008 SVEINBJORNSSON ET AL.The FASEB Journal