[CANCER RESEARCH 52, 4297-4305, August 15, 1992]

Interleukin-8 Is Produced in Neoplastic and Infectious Diseases of the HumanCentral Nervous System1

Erwin Van Meir,2 Miroslav Ceska, Fritz Effenberger, Alfred Walz, Eric Grouzmann, Isabelle Desbaillets,

Karl Frei, Adriano Fontana, and Nicolas de Tribolet

Neurosurgery Service, University Hospital (CHUV), 1011 Lausanne, Switzerland [E. V. M., E. G., I. D., N. d. T.J; Theodor Kocher Institut, UniversitätBern, 3000Bern, Switzerland [A. W.]; Sandoz Forschungsinstitut, Vienna, Austria [M. C., F. E.]; Section of Clinical Immunology, University Hospital, 8044 Zurich, Switzerland[K. F., A. F.]

ABSTRACT

The presence of ¡nterleukin-8 (IL-8), a leukocyte chemotactic factor,

was examined in primary and metastatic central nervous system tumorsand in nonneoplastic acute meningoencephalitides.

In vitro: (a) 11 of 12 glioblastoma cell lines constitutively expressedIL-8 mRNA; (b) 5 of 6 of these cell lines secreted IL-8 protein asdetected by enzyme-linked immunosorbent assay and a glucosaminidaserelease bioassay; and (e) II.-I/i or tumor necrosis factor was able toaugment both IL-8 mRNA steady state levels and protein secretion of allcell lines tested except IN-319.

IL-8 was also found in vivo, (a) IL-8 poly A mRNA was detected in

2 of 2 low grade astrocytomas, 1 of 2 anaplastic astrocytomas, and 6 of6 glioblastomas. (b) IL-8 protein was present in the cyst fluid of 1 of 4

low grade astrocytomas, 1 anaplastic astrocytoma, 2 of 2 glioblastomas,1 oligodendroglioma grade III, and one central nervous system cervicalcarcinoma metastasis, (c) The cerebrospinal fluid of 3 of 4 metastaticlymphomas, 2 of 16 glioblastomas, 1 of 2 low grade astrocytomas, butnone of 3 anaplastic astrocytomas and none of 9 meningiomas containedIL-8. The presence of IL-8 was not restricted to central nervous system

tumors as 2 of 2 bacterial meningitis and 5 of 5 acute viral meningitispatients contained considerable IL-8 levels in the cerebrospinal fluid.(d) Immunohistochemical analysis showed IL-8 immunoreactivity in

perivascular tumor cells in 11 of 15 glioblastoma sections.These data suggest that IL-8 secretion could be a key factor involved

in the determination of the lymphoid infiltrates observed in brain tumorsand the development of cerebrospinal fluid pleocytosis in meningoencephalitides.

INTRODUCTION

A feature common to certain neoplastic and nonneoplastichuman CNS3 diseases is the presence of a cellular inflammatory

response. Primary CNS tumors are infiltrated by mononuclearlymphoid populations (1-3), while nonneoplastic acute bacterial meningoencephalitides are characterized primarily by poly-morphonuclear lymphoid infiltrates (4). Neither the mediatorsnor the mechanism for the development of these specific cellinfiltrates have, as yet, been identified. However, cell specificityof the inflammatory infiltrates may be established by specificchemoattractants defining the particular classes of leukocytes.For example, N-formylated peptides such as yV-formyl-Met-Leu-Phe or polypeptides such as complement-derived C5a or

Received 12/11/91; accepted 5/29/92.The costs of publication of this article were defrayed in part by the payment of

page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by Grants 3.595.087 (N. d. T.) and 31.28402.90(A. F.) of the Swiss National Science Foundation.

2 To whom requests for reprints should be addressed, at Ludwig Institute forCancer Research, 9500 Oilman Drive. San Diego, CA 92093-0660.

3 The abbreviations used are: CNS, central nervous system; CSF, cerebrospinalfluid; cDNA, complementary DNA; ELISA, enzyme-linked immunosorbent assay;GFAP, glial fibrillary acidic protein; rhu, recombinant human; IL, interleukin;mAb, monoclonal antibody; PGE2, prostaglandin £2;TGF, transforming growthfactor; PBS, phosphate-buffered saline; MCP-1, monocyte chemoattractant pro-tein-1; TNF, tumor necrosis factor-«;NAP, neutrophil-activating protein.

leukotriene B4 are well known as neutrophil chemoattractantsthat have little effect on blood lymphocytes (5, 6). However,very little is known about lymphocyte-specific attractants (7,8).

We have chosen, therefore, to examine and compare glioblastoma, a CNS tumor, with nonneoplastic acute viral or bacterialmeningoencephalitides for the presence of a candidate chemotactic factor with potential to govern leukocyte accumulations.

Chemotactic factors may originate either in response to acellular immune reaction to a tumor and/or from tumor cellsthemselves. Chemotactic factors are released by sensitized lymphocytes when stimulated by specific antigens (9, 10). Mono-nuclear infiltrates of human glioblastomas consist predominantly of T-lymphocytes and macrophages as well asB-lymphocytes and natural killer cells (1-3). In addition, hu

man glioblastoma cell lines have been shown to release variouscytokines in vitro such as an interferon-/3-like activity (11), IL-1(12, 13), granulocyte CSF, and granulocyte-macrophage CSF( 14, 15), TGF-/32 (16), TNF ( 13, 17), MCP-1 ( 18-21 ), and IL-6(22). Although some of these cytokines such as IL-1 andMCP-1 have been reported to have chemotactic activities for

neutrophils (23) or monocytes (18, 19), no lymphocyte chemotactic factor has been identified as the mediator of the important T-lymphocyte infiltrates observed in glioblastoma. Recently, IL-8 has been shown to be a chemoattractant forT-lymphocytes as well as neutrophils and basophils but not

monocytes (8, 24, 25). Therefore, we decided to evaluate therole of this inflammatory cytokine as a candidate factor toexplain the specific T-lymphocyte infiltrates observed in CNStumors such as glioblastoma as compared to those observed innonneoplastic meningoencephalitides.

Inflammatory stimuli such as IL-1 and TNF have been shownto induce the synthesis of IL-8 mRNA in human endothelialcells (26, 27), T-lymphocytes (28), alveolar macrophages (29),dermal fibroblasts (30, 31), keratinocytes (32), retinal pigmentepithelial cells (33), synovial cells (34), neutrophils (35), as wellas 2 hepatoma cell lines (36), lung giant cell carcinoma lineLU65C (37), pulmonary epithelial cell line A549 (38), bladdercarcinoma cell line 5637 (39), and astrocytoma cell lineU373MG (20).

Here we demonstrate the inducible IL-8 mRNA expressionand the secretion of IL-8 protein by 13 glioblastoma cell lines.IL-8 mRNA was also present in frozen astrocytoma and glioblastoma tumors but not in meningioma. IL-8 was detected invivo by ELISA in cyst fluid and CSF of patients with primaryand metastatic CNS tumors as well as in the CSF of patientswith inflammatory diseases such as viral or bacterial meningitis.Immunohistochemical analysis of glioblastoma suggests thatIL-8 is produced by the tumor cells themselves. These datasuggest that in vivo IL-8 production might be an important

factor in determining the leukocyte infiltrates found in someCNS tumors or in meningoencephalitis.

4297

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

MATERIALS AND METHODS

Preparation of Supernatants from Glioma Cell Lines. The l S permanent glioblastoma cell lines used in this study and their cultureconditions have been described previously (22). For production of su-pernatants, 5 x IO5 cells of each glioblastoma cell line were plated in6-well plates (Costar). After 24 h, the subconfluent cultures werewashed 3 times and replaced with 2 ml serum-free RPMI 1640 (Se-romed). In some experiments, rhu-IL-10 (0.1 to 100 units/ml; Gen-zyme, Cambridge, MA), rhu-TNF (0.1 to 100 units/ml; Glaxo), rhu-IL-4 (IO3 units/ml; Genzyme), rhu-TGF-/32 (0.1 to 10 ng/ml; R&DSystems, Minneapolis, MN), and PGE2 ( 1O^8to 10~6 M; Sigma Chem

ical Co., St. Louis, MO) containing bovine serum albumin (Sigma) as acarrier protein were added at this step. The Supernatants were collected24 h later, centrifuged at 1000 x g, and stored at -20°C until being

tested. Supernatants were also prepared from a culture of pure astro-cytes from a human fetal brain (W323-HF 9.12:10.1 ) and from a mixedculture containing 60% astrocytes and 40% microglial cells (W360-HAM 10.31).

Harvesting Cerebrospinal Fluid and Glioma Cyst Fluid. CSF samples tested for IL-8 activity were obtained from 16 patients with suprat-entorial glioblastoma, 2 patients with benign astrocytoma, 3 patientswith anaplastic astrocytoma, 4 patients with brain metastatic lympho-mas, 9 patients with meningioma, 4 patients with nonhistologicallyconfirmed CNS tumors, 2 patients with bacterial meningitis, 5 patientswith viral meningitis, 3 patients with multiple sclerosis, and 18 patientswith unrelated diseases such as herniated lumbar disc syndrome ortension headache. In the latter patients, inflammatory or tumoral diseases of the CNS were excluded. CSF was collected by a lumbar puncture before operation at myelography. Glioma cyst fluids were obtainedperoperatively from 2 patients with glioblastoma, 4 patients with benign astrocytoma, 1 patient with anaplastic astrocytoma, 1 patient witholigodendroglioma grade III, and 1 patient with brain metastasis of acervical carcinoma. The CSF and cyst fluids were centrifuged, filteredthrough Millex GV 0.22 MHI(Millipore), and stored at -70°Cuntil the

assay.Quantification of IL-8 by ELISA. Mouse anti-NAP-l/lL-8 mAb

were used for coating microtiter plate wells (for 16 h at 4°C).After 4

washes (PBS, pH 7.5 containing 0.05% Tween-20), rhu-NAP-l/IL-8(kindly provided by Sandoz, Basel, Switzerland) at concentrations ranging from 0.02 to 10 ng/ml, or CSF (diluted at 1:2, 1:4, 1:10, or 1:100)was added to precoated plates and incubated for 2 h at 37°C.After 4

washes, goat anti-NAP-l/IL-8 alkaline phosphatase conjugate wasadded, and the plates were incubated for an additional 2 h at 37°C.After

the incubation with /7-nitrophenyl phosphate, the enzymatic reactionwas terminated by the addition of 2 N NaOH. Optical reading wasperformed at 405 nm. The sensitivity of the assay is 3 pg/ml in assaybuffer.

Bioassay for IL-8. The IL-8-like activity was measured by the releaseof A'-acetyl-/5-glucosaminidase from azurophil granules, released from

cytochalasin B-treated human neutrophils (40). In brief, samples werediluted 3- to 30-fold with PBS containing 2.5 mg/ml bovine serumalbumin (final volume 150 M!)and incubated for 15 min at 37°Cwith100 M!of cytochalasin B-pretreated (5 Mg/ml, 5 min at 37°C)humanneutrophils (IO7 cells/ml). After centrifugation, 50 M' of the cell-free

supernatant were transferred to a new microtiter plate and incubatedwith 50 n\ of 10 mm 4-methylumbelliferyl-2-acetamido-2-deoxy-/3-D-glucopyranoside (Sigma) in 0.1 Msodium citrate buffer for l h at 37°C.

The reaction was then stopped by addition of 100 M' of 0.4 M glycinebuffer. All samples were tested in the presence and absence of neutrophils to control for background levels of enzymes. To obtain absoluteIL-8 concentrations, corrected fluorescence values were fitted on a standard curve determined in the same assay with 72-amino acid rhu-IL-8.

RNA Isolation and Hybridization. Poly A+ selected mRJ^JAwas iso

lated from cell lines or solid tumors and used for Northern blottingexperiments with random labeled probes as described previously (22).Solid tumors from 8 low grade astrocytomas (patients AII-460, 466,471, 472, 483, 501, 528, and 552), 2 anaplastic astrocytomas (patientsAIII-485 and 496), 8 glioblastomas (patients G-403, 435, 467, 468,

473, 475, 515, and 523), and 5 meningiomas (patients M-391, 421,433,448, and 523) were used. The 500-base pair EcoRI fragment of theIL-8 gene coding region of plasmid pUC19hIL-8 was used as a probefor IL-8 (41), the 450-base pair EcoRl/Pstl fragment of the IL-la genecoding region as probe for IL-1«,and the ft/I 1100-base pair fragmentof plasmid pAL41 as probe for ß-actin(22).

Immunohistochemical Staining. Biopsy samples obtained from various CNS tumors and 3 normal adult brains (autopsy or lobectomy)were stored at -70°C. Serial sections of 7- to 8-Mm thickness were

processed for immunohistochemistry as described previously (22). Theantibodies used were an anti-IL-8 hybridoma supernatant 46E5 (42), a1:800 dilution of anti-GFAP mAb N358 (Amersham), a 1:20 dilutionof pan-T-lymphocyte anti-Leu-4 (CD3) mAb (Beckton Dickinson), 2antibodies commonly used to detect macrophages [a 1:20 dilution ofanti-Leu-M3 (CD14) mAb (Beckton Dickinson) and a 1:20 dilution ofanti-Leu-Ml (CD15) mAb (Beckton Dickinson)], and supernatant ofP3X63Ag8, a 7l-producing myeloma cell line as a negative control.The following 15 glioblastomas were analyzed: 403, 467,473, 475, 479,515, 523, 546, 549, 621, 628, 637, 658, 669, and 696.

Radiolabeling of IL-8. Four Mghuman recombinant IL-8 in 50 n\ of0.1 M phosphate buffer, pH 7.5, were added to 400 nC\ of [125I]Na

(Amersham IMS 30) for 15 min at room temperature in a tube containing 2.5 Mg lodogen (Pierce) (43). The reaction was stopped byadding 50 M!of a 120mM KI, 5 m\i ascorbic acid solution in0.9%NaCl.The mixture was fractionated using a 50-cm Sephadex G15 column andPBS containing 1% bovine serum albumin as eluent. The yield of labeling was evaluated to be 50% and the specific activity 50 pCi/pg.Biological activity of the labeled and cold IL-8 was evaluated in anelastase release assay (44). The biological activity of the iodinated IL-8was estimated to be 15% of the unlabeled IL-8.

Radioreceptor Assay of IL-8 on Glioblastoma Cell Lines. Bindingexperiments were performed in parallel l h at 37°Cor for 2 h at 4°Cto

avoid internalization (45) using RPMI containing 10 mg/ml bovineserum albumin and 10 mMA'-(2-hydroxyethyl)piperazine-A''-(2-ethane-sulfonic acid) as binding buffer and 90,000 cpm of I25l-labeled IL-8(1.18 ng/tube). Nonspecific binding was obtained with 1000-fold excessof unlabeled IL-8. Human blood granulocytes were used as positivecontrol. The 13 following glioblastoma cell lines were analyzed: LN-229, LN-319, LN-340, LN-382, LN-427, LN-428, LN-308, LN-Z308,LN-443, LN-444, LN-464, U-118, and U-251.

Quantification of IL-4 by ELISA. IL-4 was measured with the commercial Intertest-4 ELISA kit (Genzyme) according to the manufacturer's instructions. The Supernatants of 9 glioblastoma cell lines pre-

treated with rhu-IL-1 (0.1 to 100 units/ml) or untreated were tested at1:1 dilution. The detection limit at this dilution was 0.18 ng/ml. Thefollowing cell lines were tested: LN-18, LN-215, LN-229, U-251, LN-Z308, LN-319, LN-443, LN-444, and LN-464.

RESULTS

Twelve permanent glioblastoma cell lines were tested byNorthern blotting for the presence of IL-8 mRN A using an IL-8cDNA probe (Fig. 1). Variable amounts of constitutive IL-8mRNA were detected in 11 of 12 cell lines tested, in one skinfibroblast culture taken from the scalp (Fbl 445), and in thehistiocytic lymphoma line U937 (data not shown). No mRNAwas detected in cell line LN-319, whereas strong constitutiveexpression was obtained with cell lines LN-215, U-251, LN-308, and LN-443. All cell lines tested had low to high IL-8mRNA levels after induction with IL-1/3 or TNF (Fig. I, A andB). The enhancement ranged between 2 and 200 times as estimated by scanning the autoradiograph and standardizing to the/3-actin mRNA levels. IL-1/3 could also induce a weak expression of the IL-8 mRNA in cell line LN-319. The induction orenhancement of the IL-8 mRNA by IL-1/3 or TNF was maximalat 4 h and decreased slowly between 24 and 48 h (Fig. l C). Thestimulation was already detected with 0.1 unit/ml IL-1/3 (datanot shown) and 1 unit/ml TNF. The optimal concentrations

4298

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GO CO

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

LO iO co O Q CO CO ^J•¿�— CO •¿�— O f— ' '' ^ "I " -!''''

28S-

- 1.8Kb

- ßactin

Beo 1—¿�

T3

TNF ,

59 QO ' i

28S-

18S- •¿�H" -1.8Kb

28S-

28S-

1.8Kb 18S- -1.8Kb

•¿�•••* -ß actin

ßactin

Fig. l. IL-8 mRNA detection in glioblastoma cell lines. Northern blot analysis was performed on 10 Mgof total RNA using an IL-8 cDNA probe or a /3-actin cDNAprobe as control. I. glioblastoma cells were cultured for 24 h in medium alone (- columns) or in medium plus S units/ml IL-1/3 (+ columns). B, as in A but stimulationwith medium plus TNF (100 units/ml). C, as in A but the cells were cultured for 4, 24, or 48 h with IL-1/3 (5 units/ml). Five Mgof total fibroblast RNA were used asa positive control (Fbl 445). D, as in A but the cells were stimulated by IL-1/3 (1 unit/ml) or TNF (0.1 to 100 units/ml). The blots were autoradiographed 48 h (A), 11h (B), 1 days (O, and 24 h (D). Kb, Kilobase.

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

were 1 unit/ml IL-ißand 10 units/ml TNF for cell line LN-229(Fig. \D). Interferon-7 had no effect on IL-8 mRNA expressionby LN-229 cells at 100 units/ml (data not shown).

To test whether the synthesis of IL-8 was also detected at theprotein level and whether the increased amount of IL-8 mRNAobtained in response to IL-Ißor TNF resulted in a direct increase in secreted protein, we tested the supernatants of 6 selected cell lines expressing various levels of IL-8 mRNA forIL-8 by ELISA (Fig. 2). All of them except LN-319 showedvariable constitutive IL-8 levels by ELISA ranging from 3 to190 ng/ml. Both IL-1/3 (Fig. 2A) and TNF (data not shown)enhanced IL-8 production in a dose-dependent way. Althoughdoubling times of the different cell lines vary between 36 and 48h, a semiquantitative comparison is possible since equal cellnumbers (5 x 10s) were used and subsequent supernatants wereharvested at 48 h. The IL-8 levels increased up to more than850 ng/ml for cell line LN-215 with a 24-h stimulation with 100units/ml IL-1/3 for 24 h. The enhancement of IL-8 production

A. ELISA1000

U)

l(0

o.U)C

800 -

600 -

400 -

o>C

.H

control":i%101U/mlIL-1H10 U/mlIL-10100 U/mlIL-1\1!""J„1l;J|J

200 - r

B.

LN-18 LN-215 LN-229 LN-Z308 LN-308 LN-319 LN-428

cell lines

Glucosaminidase release assay100

80 -

60 -

40 -

20 -

l\pa

^\lrJp1JP1

control01U/mlIL-1010 U/mlIL-10100 U/mlIL-1il

1•Frj^spai|\\-vrílLN-18 LN-215 LN-229 LN-Z308 LN-308 LN-319 LN-428

cell lines

28S

18S -

18S

18S

- IL-8

- IL-1a

- ß-actin

Fig. 2. IL-8 detection in glioblastoma cell line supernatants. Cell lines werecultured for 24 h in medium alone (control) or in medium plus IL-1/3 (1 to 100units/ml). The supernatants were tested in parallel for IL-8 by ELISA (A) and theglucosaminidase release assay (B). LN-Z308 is a subculture of line LN-308 at 15passages that has been cultured independently in Zürichsince 1982.

4300

Fig. 3. IL-8 mRNA detection in ex vivo tumors. Northern blot analysis wasused to test IL-8 and IL-1«mRNA expression on variable amounts (0.1 to 5 Mg)poly A+ selected RNA extracted from 2 low grade astrocytoma (AH), 2 anaplastic

astrocytoma (AIII), and 6 glioblastoma (G) depending on the available tissue. Theblot was autoradiographed for 2 weeks for IL-8 and 3 weeks for IL-1«.ThemRNA degradation level and proportional mRNA amount loaded were assessedusing a ß-actincDNA probe. Kb. kilobase.

was proportional to the amount of IL-1/3 added and correlatedwith the mRNA levels detected by Northern blots. No IL-8 wasdetected in the supernatants of LN-319 cells treated with IL-lßat 1 to 100 units/ml, whereas a small amount (8 ng/ml) wasobtained with a TNF treatment at 100 units/ml (data notshown). IL-la could also enhance IL-8 secretion as tested oncell line LN-Z308, but at 10 times higher concentrations thanIL-1/3; whereas 0.1 to 10 ¿ig/mlLPS and 100 units/ml IL-6 had

no effect (data not shown).To further evaluate whether the IL-8 detected by ELISA is

biologically active, the same supernatants were tested for theirability to induce the secretion of glucosaminidase by neutro-phils (see "Materials and Methods"). All of the supernatants

tested were bioactive, and their IL-8-like activity was inducibleby IL-1/3 and TNF except for LN-319 (Fig. 2B). Although thegeneral profile of the data was identical to those obtained byELISA, the amount of IL-8-like activity detected by bioassaywas on average 10 times lower (compare Fig. 2, A and B). Thisdifference was not due to interfering activities in the supernatants as similar results were obtained with diluted samples (1:5and 1:10). To test whether proteases present in the supernatantscould be responsible for this result, we produced new supernatants in the presence of protease inhibitors (leupeptin 4 x 10~4

M and aprotinin 3 units/ml), however, no difference was noticed. Furthermore, 0.1 to 10 ng/ml IL-4, an IL-8 gene expression regulatory factor (46), 0.1 to 10 ng/ml TGF-/32, and 10~8to 10~6 M PGE2 showed no inhibitory effect (data not shown).

Supernatants from a purified astrocyte culture from humanfetal brain and a mixed culture of 60% astrocytes and 40%microglial cells contained 38 and 370 pg/ml IL-8, respectively.

The relevance of the in vitro data to the in vivo situation wasfirst addressed on a Northern blot with isolated poly A"1"mRNA

from ex vivo frozen tumors (Fig. 3). Two of 2 low grade astro-cytomas, 1 of 2 anaplastic astrocytomas, and 6 of 6 glioblasto-mas were faintly to strongly positive on the original autoradio-graph. No signal was obtained with mRNA of a 17-week fetalbrain and 5 meningiomas (data not shown).

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

As further proof for the production of IL-8 in vivo, the presence of IL-8 in CSF and cyst fluid of patients was measured(Fig. 4). Eighteen CSF from patients with unrelated diseasessuch as herniated lumbar disc syndrome or tension headachewere used as controls. CSF of patients with infectious CNSdiseases consisting of 2 bacterial meningitis and 5 viral meningitis were also tested. Two of 16 glioblastomas (569 and 20pg/ml), 1 of 2 benign astrocytomas (110 pg/ml), 0 of 3 anaplas-

tic astrocytomas, 3 of 4 metastatic lymphomas (138, 66 and 54pg/ml), 0 of 9 meningiomas, and 2 of 4 other nonhistologicallyconfirmed CNS tumors (271 and 108 pg/ml) were positive forIL-8 by ELISA. Two of 18 control CSF (50 and 30 pg/ml), 2 of2 bacterial meningitis (24,340 and 289 pg/ml), and 5 of 5 viralmeningitis (4411, 1857, 1184, 549 and 250 pg/ml) were positive for IL-8. The CSF of 3 patients with multiple sclerosis werenegative. The tumor cyst fluids of 2 glioblastomas (7787 and612 pg/ml), 1 of 4 astrocytomas grade I or II (72 pg/ml), 1astrocytoma grade III (860 pg/ml), 1 oligodendroglioma gradeIII (885 pg/ml), and 1 metastasis of a cervical carcinoma (1239pg/ml) were also positive for IL-8.

The presence of IL-8 poly A+ mRNA in 9 of 10 ex vivoastrocytoma and glioblastoma specimens analyzed and of IL-8protein in 4 of 7 cyst fluids of gliomas suggests that cells withinthe tumor tissue release IL-8 in vivo. Some candidate cells arethe tumor cells themselves, reactive astrocytes, endothelialcells, T- and B-lymphocytes, and macrophage or microglialcells. To address this question, immunohistochemical analyseson frozen tissue sections were performed. Adjacent serial sections were studied by utilizing anti-IL-8 mAb 46E5 (42), anti-

GFAP serum, 2 antibodies commonly used to detect macrophages [anti-Leu-Ml (CD 15) mAb and anti-Leu-M3 (CD 14)mAb], a pan-T-lymphocyte antibody [anti-Leu-4 (CD3) mAb],and supernatant of P3X63Ag8 as a negative control (Fig. 5). Inthe glioblastoma sections of 11 of 15 patients studied, numerous dispersed strongly positive regions were observed. Some ofthe positive cells had a foamy cytoplasm and were mostly localized around tumor vessels where macrophages and T-lym-phocytes were present (Fig. 5, A and E). Pseudopalisading areas

100000.

10000-

f 1000.

100.

15

À0n

•¿�a+AA

A0

SVA

•¿�À•

ooTAAAA

AAAAAAAAAAAAAAAAAM*D•AAA•*Tocontrolmultiple

sclerosisviral

meningitisbacterial

meningitisbenign

astrocytomaanaplastic

astrocyloma

glioblastomametástasismeningioma

oligodendroglioma III

neurinomaCNS

tumorXXX

X X XX X X X*X

X X X *

Tumorcyst fluid

TumorCSF

MeningitisCSF

ControlCSF

Fig. 4. IL-8 detection in CSF and cyst fluid of patients with neoplastic andinfectious diseases of the CNS. Nine tumor cyst fluids. 39 tumor CSF, 7 viral orbacterial meningitis, 18 control CSF from patients with noninflammatory diseases such as herniated lumbar disc syndrome or tension headache, and 3 controlpatients with multiple sclerosis.

surrounding necrosis were also positive (data not shown). Adjacent sections were positive for GFAP, suggesting that thetumor cells produce IL-8 in vivo (Fig. 5D). No reactivity withneoplastic vascular endothelial cells or T-lymphocytes was

found. Gliosis surrounding metastatic CNS lymphomas wasalso positive, therefore reactive astrocytes may also produceIL-8. The presence of macrophages was assessed with anti-CD 14 and anti-CD 15 mAbs. In some glioblastomas, 10 to 50%of the cells were positive with anti-CD 14 as well as the gliosissurrounding metastatic CNS lymphomas (data not shown). Theexpression pattern of the CD 14 epitope correlated with theregions expressing IL-8. In contrast, only a few cells surrounding vessels were positive with anti-CD 15 (Fig. 5, B and F).These results suggest that infiltrating macrophages may alsoparticipate in the IL-8 secretion or that the tumor cells andreactive astrocytes expressing IL-8 are a subpopulation alsoexpressing the CD 14 antigen. No IL-8 staining was detected on3 normal adult brain sections examined (data not shown).

To determine whether IL-8 secretion could have a directbiological effect on the tumor cells, we checked for the presenceof IL-8 receptors on 13 glioblastoma cell lines with 125I-labeledIL-8. Freshly isolated human blood granulocytes were used aspositive control. The binding experiments were performed at4°Cto avoid internalization and subsequent degradation of the

ligand-receptor complex (45). Granulocytes exhibited manybinding sites with 92.5% specific binding, whereas no specificbinding was obtained for the 13 glioblastoma cell lines tested(data not shown).

DISCUSSION

The results presented here indicate that in vivo IL-8 secretion

occurs in both primary and metastatic brain tumors (but not inleptomeningeal tumors such as meningiomas) and in nonneo-plastic meningoencephalitides. We suggest that IL-8 may be animportant candidate factor determining the leukocyte infiltrations linked to these CNS diseases.

Brain Tumor Data. For brain tumors, our study shows that11 of 12 glioblastoma cell lines expressed IL-8 mRNA withoutany stimulation. ELISA and bioassay studies showed detectableIL-8 in the supernatants of 5 of 6 of the same cell lines. IL-1/3and TNF were able to enhance IL-8 mRNA steady state levelsand IL-8 secretion in these cell lines. LN-319 was the only cellline that did not secrete IL-8 constitutively. Despite the induction of IL-8 mRNA by IL-1 ßand TNF, only TNF could induceIL-8 production in LN-319, suggesting possible posttranscrip-tional control mechanisms. These results are in agreement withprevious studies; inflammatory cytokines regulate IL-8 mRNAsteady state levels and IL-8 secretion in astrocytoma cell lineU373MG (20) and in other cell types (26, 28-34, 36-38).

Interestingly, the IL-8 levels measured by ELISA in the supernatants of all the glioblastoma cell lines tested were about 10times higher than the IL-8-like activity obtained with the bio-

assay even in the presence of protease inhibitors. This difference is likely the result of comparing the 77-amino acid form ofIL-8 produced by glioblastoma cell lines4 with the 72-aminoacid rhuIL-8 (our standard), which is approximatively 10-fold

more bioactive in vitro (31, 47, 48).

4 M. Tada, Y. Sawamura. S. Sakuma, H. Abe, K. Suzuki, Y. Yamakawa, E. VanMeir, and N. de Tribolet. Human astroglial cells produce the 77-amino acid formof interleukin-8 (IL-877), submitted for publication.

4301

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

F

;»'•¿�•&.*•» *.,•¿�*,.*•:;* t*, *

-** •¿�

Fig. 5. Identification of IL-8 producing cells on ex vivoglioblastoma tumor specimens. Glioblastoma sections from patient G-696 stained with anti-IL-8 mAh 46ES(A), ami I,cu-Ml (CD15) mAb (A), anti-CD3 mAb (Q, or anti-GFAP serum (D). Glioblastoma section from patient G-403 stained with anti-IL-8 mAb 46E5 (E) oranti-Leu-Ml (CD15) mAb (F). Magnifications: A to D. x 250; E and F, x 400.

IL-8 was detected by ELISA in tumor cyst fluids of 1 of 4 lowgrade astrocytomas, 1 anaplastic astrocytoma, and 2 of 2 glio-blastomas. The cyst fluids of one oligodendroglioma grade IIIand one CNS cervical carcinoma metastasis were also positive.In addition, IL-8 was also present in the CSF of 3 of 4 meta-static lymphomas, in 2 of 16 glioblastomas, in 1 of 2 low gradeastrocytomas, but not in 3 anaplastic astrocytomas and 9 men-ingiomas analyzed. Other cytokines such as IL-1, IL-6, TGF-/3,

but not TNF have been reported in the cyst fluid or CSF ofbrain tumors (15, 22).

The presence of IL-8 in vivo in gliomas was further supportedby the finding that 2 of 2 low grade astrocytomas, 1 of 2 ana-

plastic astrocytomas, and 6 of 6 glioblastoma ex vivo tumorspecimens contained IL-8 poly A+ mRNA.

Different cell types such as T- and B-lymphocytes, mono-cytes, and macrophages, fibroblasts, microglial cells, astrocytes,and endothelial cells or the tumor cells themselves are potentialcandidates for the production of IL-8 in vivo. Some of thesecells are known to release IL-8 in vitro (26-29, 31). Immuno-cytochemical analysis of glioblastoma sections showed that in11 of 15 patients, between 30 to 80% of the cells were positivefor IL-8 staining in regions expressing GFAP, suggesting thatthe tumor cells themselves are able to produce IL-8 in vivo.Although endothelial cells have been shown to secrete the IL-8

4302

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

77-amino acid form (26, 27), neoplastia vascular endotheliumwas not stained; neither were T-lymphocytes and sections of

normal brain. The positively staining cells were usually focusedaround vessels in regions containing macrophages and T-lymphocytes as shown by staining with anti-CD 15 and anti-CD3mAbs. Therefore, we cannot exclude that in addition to tumorcells, macrophages may participate directly in the IL-8 secretion; or that inflammatory cytokines such as IL-1 or TNF,produced by macrophages and microglial cells, could indirectlyelicit IL-8 secretion by the tumor cells in vivo, as occurs in vitro.Reactive astrocytes present in the gliosis surrounding meta-static CNS lymphomas were also positive, and primary culturesof astrocytes secreted IL-8, suggesting that when activated, normal astrocytes are also able to secrete IL-8.

What might be the mechanism of IL-8 production by theneoplastic cells? Our in vitro data show that glioblastoma cellsare able to secrete IL-8 constitutively, therefore the first possibility is that in vivo they also produce IL-8 without induction.The strong IL-8 induction capacities of IL-1 and TNF in vitrosuggest, however, that these factors may also play an autocrineor paracrine role in vivo. Although it is not a general characteristic, some astrocytoma and glioblastoma cell lines are ableto release TNF and IL-1 in vitro (12, 13, 49). The same heterogeneity is observed in vivo: we have been unable to detect IL-levmRNA in the tumors positive for IL-8 mRNA, and neitherIL-lt* nor IL-1/3 was detected by radioimmunoassay in the CSFof our glioma patients (49). Nevertheless, IL-la and -ßwererecently found in the cyst fluid but not CSF of 3 of 7 glioblastoma patients (15); IL-1/3 mRNA was detected in 3 fresh astro-cytomas (50), and IL-la immunostaining was reported on glioblastoma sections (49). These studies suggest that tumor cells,infiltrating macrophages, microglial cells, or reactive astrocytesthat all secrete IL-1 in vitro (12, 20, 51) could in some casesproduce IL-1 in vivo and so induce IL-8 secretion.

Nonneoplastic Meningoencephalitides Data. All 7 CSF ofinflammatory diseases such as viral and bacterial meningitiswere positive for IL-8 (up to 24,340 pg/ml in one case). Thepresence of other inflammatory cytokines such as TNF (52-54), IL-1 (54, 55), and IL-6 (15, 54, 56) in the CSF of menin

gitis has already been reported. The cellular origin of the cytokines produced in vivo has not yet been identified. Theinitiation of meningitis occurs primarily in the arachnoid layerof the leptomeninges and can subsequently spread to the piamater and the brain. Therefore, the activated endothelial cellsof the blood vessels and scattered macrophages as well as thearachnoid cap cells and fibrocytes present in these layers are allcandidates for the initial cytokine production. These cytokinescould then be released either directly in the blood circulation orin the CSF through the subarachnoid space and initiate CSFpleocytosis (57-59). Intrathecal injection of IL-1 and TNF orlipopolysaccharide causes meningitis and leukocytosis in theCSF of rabbits (60), and IL-6 and IL-8 production is induced byIL-1 and TNF in macrophages and glial cells in vitro (22, 29).Therefore, part of the action of TNF and IL-1 in meningitispathogenesis might be mediated by their induction of IL-8 andIL-6 in the inflamed regions of the CNS.

Although it is tempting to parallel the IL-8 production inmeningitis and CNS neoplasia, the involvment of IL-1 andTNF is better established in the first case than in the second, inwhich the tumor cells may constitutively secrete IL-8 withoutIL-1 or TNF induction. The finding of IL-1 but not TNF in invivo samples of glioblastoma suggests, however, that at leastIL-1 may play a role in IL-8 secretion. It will, however, be

difficult to establish whether IL-8 production in vivo in bothdiseases may be a common feature having different origins (tumor cells producing constitutively IL-1 and IL-8 as a result oftransformation and nonneoplastic cells producing IL-1 andIL-8 in response to viral or bacterial infection) or a more general phenomenon linked to the common production of inflammatory cytokines such as IL-1 by macrophages, microglial cells,or astrocytes in response to the presence of bacterial li-popolysaccharides and viral or tumor antigen presentation.

What Could Be the Consequence of IL-8 Secretion for BothDiseases in Vivo? First, does IL-8 act as an autocrine growthfactor for glioblastoma cells? To test whether IL-8 could have adirect effect on glioblastoma cells through an IL-8 receptor-mediated mechanism, we decided to check for the presence ofIL-8 receptors by binding of 125I-labeled rhIL-8 on 13 glioblas

toma cell lines. No binding was observed, suggesting by thistechnique that IL-8 has no direct effect on glioblastoma cells.

Second, does IL-8 play an important role for the leukocytechemotaxis in CNS tumors or meningitis? These diseases showdifferences in the localization (tumor or CSF) and the leukocytecontent of their infiltrates. CNS brain tumors have a predominance of mononuclear infiltrates within the tumor consistingmainly of T-lymphocytes and macrophages, whereas almost noleukocytes were found in the CSF (1-3). The leukocyte infiltrates in the CSF of bacterial meningoencephalitis patients consist of more than 90% polymorphonuclear leukocytes, whereasviral meningoencephalitis patients' CSF contain mostly lym

phocytes (4). IL-8 has been shown to be chemotactic in vitro forhuman neutrophils predominantly, as well as for T-lymphocytes and basophils (8, 24, 25). In our study, IL-8 was found inthe CSF of all 7 meningitis patients tested at low to high concentrations (up to 24,340 pg/ml). IL-8 was also present in thetumor cyst fluid at low to moderate concentrations (70 to 7000pg/ml) in 6 of 9 CNS tumors, but only in 8 of 35 CSF. Tada eta/.5 have also found a neutrophil chemotactic activity in 2 of 2

regional fluid and 2 of 4 CSF of operated glioma patients without the presence of neutrophils in vivo. Although IL-8 is presentin these body fluids, we only rarely detected the presence ofgranulocytes by immunohistochemistry in the tumor sectionsanalyzed in this study (data not shown). T-lymphocyte infiltrates without the presence of neutrophils after rhIL-8 injectionhave been observed previously in the guinea pig lung and the earskin of rats (24, 61).

Several factors may explain these different leukocyte contents, (a) The neutrophil response could be transient due todown-regulating functions that may impinge its further accumulation. Possible inhibitory factors include IL-4 (46), PGE2,and TGF-/32 (15). However, we have shown by ELISA that IL-4is not secreted by 9 glioblastoma cell lines in vitro and that IL-4,PGE2, and TGF-/32 do not affect the IL-1 stimulated IL-8secretion and bioactivity in vitro (data not shown), (b) Theconcentration of IL-8 could be below a threshold level necessaryto elicit a response. Injection of rIL-8 into the ear skin of rats at1 ng/ml induced a selective T-lymphocyte infiltration, whereasneutrophil recruitment necessitated 100-fold higher doses (24).Therefore, the low IL-8 concentrations (100 pg/ml to 1 ng/ml)found in the cyst fluid or CSF of the patients studied couldpossibly elicit only a T-lymphocyte infiltrate response. This is,however, unlikely since similar concentrations were observed in

5 M. Tada. Y. Sawamura. S. Sakuma. K. Suzuki. H. Ohta, T. Aida, and H. Abe.Cellular and cytokine responses in the human central nervous system to intracranialadministration of tumor necrosis factor-a for the treatment of malignant gliomas,submitted for publication.

4303

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

the CSF of one of our patients with bacterial meningitis whereneutrophil pleocytosis was present, (c) Tumor-secreting factorsmay inhibit patients' neutrophil migration abilities. Some cell

ular immune responses are impaired in patients with primaryintracranial tumors (62). However, no granulocytopenia existsin brain tumor patients (62) and no impairement of their neutrophil immune responsiveness has been reported so far. Inaddition, granulocyte-macrophage CSF, initially characterizedas a neutrophil migration inhibitory factor (63), is produced byglioblastoma cells in vitro but not in vivo (15). (d) The neoplas-tic vascular endothelium of the tumors was not stained for IL-8,suggesting either a lack of inducing factors such as TNF or adifference in function as compared with normal endothelium(27). Endothelial cells play an important role in neutrophildiapedesis: TNF, IL-1/3, and lipopolysaccharide induce endo-thelial cell-derived IL-8 secretion that in turn generates achemotactic gradient and allows neutrophil-endothelium interaction by regulating adhesion molecules such as LECAM-1 and/32 integrins (27). (e) The primary IL-8 form produced may byitself be unable to elicit neutrophil migration in the humanCNS. An attractive hypothesis would be that the 77-amino acidform of IL-8 produced by glioblastoma cells4 and the IL-8

molecular forms produced in meningoencephalitides wouldneed an additional proteolytic cleavage to become chemoattrac-tant in vivo for neutrophils. The 72-amino acid form of IL-8 isabout 10-fold more chemotactic in vitro than the 77- and 79-amino acid forms of IL-8 (31,47). TNF that is produced in vivoin bacterial meningitis (where neutrophil infiltrates are predominant) but not in viral meningitis (52) or glioblastoma (15)could be indirectly involved in this mechanism. TNF couldinduce the opening of endothelial junctions leading to an influxof blood components containing proteases able to cleave immature IL-8. TNF also regulates the coagulant/anticoagulant activity at the endothelial cell-blood interface, resulting in production of a proteolytic activity (47). Two clinical studies showthat a neutrophil response can be obtained when TNF or IL-2is injected in the CNS: first, Tada et a/.5 have shown that TNF

injected intracerebrally in human glioma patients could elicitleukocyte bursts both in regional fluid and CSF. These eventscorrelated with the appearance of a neutrophil chemotactic activity that copurified with IL-8 on high performance liquidchromatography.5 Second, intraventricular injection of IL-2 in

patients with leptomeningeal carcinomatosis induced a transient rapid influx of neutrophilic leukocytes, followed by a prolonged presence of lymphocytes correlating with the appearance of the IL-8 inducers IL-1 and TNF (64). Clearly thepotential of IL-8 as a leukocyte chemoattractant in the humanCNS will only be established with IL-8 injections into the CNS.This will only be possible if correlations between good prognosis and IL-8 levels can be established. In the meantime, injection of IL-8 in the CSF or the brain of an animal model willallow one to address the above-discussed hypotheses.

Finally, the fact that IL-8 belongs to a superfamily of cytok-ines sharing 4 half-cystines (the IL-8 and MCP-1 families) (21)opens a new chapter in the understanding of the leukocytechemotaxis in CNS diseases. Macrophage inflammatory protein-1 and -2 have been shown to exert important inflammatoryresponses in the CNS in a rabbit model (60, 65). MCP-1 ischemoattractant for monocytes but not neutrophils and is alsoinduced by IL-1 and TNF in human astrocytomas and glioblas-tomas in vitro (18-21). Of further particular interest will beRANTES, which attracts both T-lymphocytes of the memorytype and (at 200-fold higher doses) monocytes but not neutro

phils (66). Therefore, the possible concomitant release of othermembers of these 2 families and their concentrations in theCNS will influence and determine the final makeup of leukocytes in the tumor tissue.

ACKNOWLEDGMENTS

We wish to thank Dr. J. Antel, Dr. E. Christophers, Dr. T. Gauthier,Dr. I. Lindley, D. Piani, Dr. K. Matsushima, Dr. K. Schwechheimer,Dr. M. Sticherling, Dr. M. Tada, Dr. K. Williams, and Sandoz for thegenerous gifts of cytokines, cDNA probes, cell culture supernatants,CSF, antibodies, helpful advice, and unpublished data. We would alsolike to thank A-C. Diserens, I. Guignard, and M-F. Hamou for excellent technical assistance, Dr. T. Gauthier for expert computer assistance, and Dr. J-P. Deruaz and Dr. R. Janzer for the neuropathologicaldiagnosis of the tumors. We are grateful to Dr. W. Cavenee, Dr. T.Mikkelsen, Dr. S. Rempel, and Dr. Y. Sawamura for critically readingthe manuscript.

REFERENCES

1. von Hanwehr, R. I., Hofman, F. M., Taylor, C. R., and Apuzzo, M. L.Mononuclear lymphoid populations infiltrating the microenvironment of primary CNS tumors. Characterization of cell subsets with monoclonal antibodies. J. Neurosurg., 60: 1138-1147, 1984.

2. Kuppner, M. C., Hamou, M. F., and de Tribolet, N. Immunohistological andfunctional analyses of lymphoid infiltrates in human glioblastomas. CancerRes., 48: 6926-6932, 1988.

3. Wood, G. W., and Morantz, R. A. Immunohistologic evaluation of the lym-phoreticular infiltrate of human central nervous system tumors. J. Nati.Cancer Inst.. 62: 485-491, 1979.

4. Oehminger, M. Cerebrospinal Fluid Cytology. Stuttgart: Thieme, 1976.5. Snyderman, R., and Goetzl, E. J. Molecular and cellular mechanisms of

leukocyte chemotaxis. Science (Washington DC), 213: 830-837, 1981.6. Wilkinson, P. C., and Haston, W. S. Chemotaxis: an overview. Methods

Enzymol., 762:3-16, 1988.7. Wilkinson, P. C, Haston, W. S., and Shields, J. M. Some determinants of the

locomotory behaviour of phagocytes and lymphocytes in vitro. Clin. Exp.Immunol., 50: 461-473, 1982.

8. Leonard, E. J., Skeel, A., Yoshimura, T., Noer, K., Kutvirt, S., and Van Epps,D. Leukocyte specificity and binding of human neutrophil attractant/activation protein-1. J. Immunol.. 144: 1323-1330. 1990.

9. Ward, P. A., Remold, H. G.. and David, J. R. The production by antigen-stimulated lymphocytes of a leukotactic factor distinct from migration inhibitory factor. Cell Immunol.. /: 162-174. 1970.

10. Snyderman, R., Altman, L. C., Hausman, M. S., and Mergenhagen, S. E.Human mononuclear leukocyte chemotaxis: a quantitative assay for humoraland cellular chemotactic factors. J. Immunol.. 108: 857-860. 1972.

11. Larsson, L, Landstrom, L. E., Larner. E., Lundgren, E., Miorner, H., andStrannegard, L. Interferon production in glia and glioma cell lines. Infect.Immun., 22: 786-789, 1978.

12. Lee, J. C., Simon, P. L., and Young, P. R. Constitutive and PMA-inducedinterleukin-1 production by the human astrocytoma cell line T24. Cell Immunol., IIS: 298-311, 1989.

13. Velasco, S., Tarlow, M., Olsen, K., Shay, J. W., McCracken, G. H., andNisen, P. D. Temperature-dependent modulation of lipopolysaccharide-in-duced interleukin-1/j and tumor necrosis factor alpha expression in culturedhuman astroglial cells by dexamethasone and indomethacin. J. Clin. Invest.,87: 1674-1680, 1991.

14. Tweardy, D. J., Molt, P. L., and Glazer, E. W. Monokine modulation ofhuman astroglial cell production of granulocyte colony-stimulating factorand granulocyte-macrophage colony-stimulating factor. I. Effects of IL-1alpha and IL-beta. J. Immunol., 144: 2233-2241, 1990.

15. Frei, K., Malipieiro, U., Piani, D., Van Meir, E., de Tribolet, N., and Fontana, A. GM-CSF production by glioblastoma cells: despite the presence ofinducing signals GM-CSF is not expressed in vivo. i. Immunol.. 1480:3140-3146, 1992.

16. Bodmer, S., Strommer, K., Frei, K., Siepi, C., de Tribolet, N., Heid, I., andFontana, A. Immunosuppression and transforming growth factor-beta inglioblastoma. Preferential production of transforming growth factor-beta 2.J. Immunol., 143: 3222-3229, 1989.

17. Bethea, J. R., Gillespie, G. Y., Chung, J. Y., and Benveniste, E. N. Tumornecrosis factor production and receptor expression by a human malignantglioma cell line. D54-MG. J. Neuroimmunol., 30: 1-13, 1990.

18. Matsushima, K., and Oppenheim, J. Interleukin-8 and MCAF: novel inflammatory cytokines inducible by IL-1 and TNF. Cytokine, /: 2-13, 1989.

19. Robinson, E. A., Yoshimura, T., Leonard, E. J., Tanaka, S., Griffin, P. R.,Shabanowitz. J., Hunt, D. F., and Appella, E. Complete amino acid sequenceof a human monocyte chemoattractant, a putative mediator of cellular immune reactions. Proc. Nati. Acad. Sci. USA, 86: 1850-1854, 1989.

20. Kasahara. T., Mukaida. N., Yamashita, K., Yagisawa, H., Akahoshi, T., and

4304

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

IL-8 IS PRODUCED IN HUMAN CENTRAL NERVOUS SYSTEM DISEASES

Matsushima, K. IL-1 and TNF-a induction of IL-8 and monoeyte chemot-actic and activating factor (MCAF) mRNA expression in a human astrocy- 44.toma cell line. Immunology, 74: 60-67, 1991.

21. Leonard, E. J., and Yoshimura, T. Human monoeyte chemoattractant pro-tein-I (MCP-1). Immunol. Today, //: 97-101, 1990. 45.

22. Van Meir, E., Sawamura, Y., Diserens. A. C.. Hamou, M. F.. and de Tribolet,N. Human glioblastoma cells release interleukin 6 in vivoand in vitro. CancerRes., 50:6683-6688, 1990.

23. Kampschmidt, R. F., and Pulliam, L. A. Effect of human monoeyte pyrogen 46.on plasma iron, plasma zinc, and blood neutrophils in rabbits and rats. Proc.Soc. Exp. Biol. Med., 158: 32-35, 1978.

24. Larsen, C. G., Anderson, A. O., Appella, E.. Oppenheim, J. J., and Mat- 47.sushima, K. The neutrophil-activating protein (NAP-1) is also chemotacticfor T lymphocytes. Science (Washington DC), 243: 1464-1466, 1989.

25. Bacon. K. B., and Camp, R. D. Interleukin (ILj-8-induced in vitro humanlymphocyte migration is inhibited by cholera and pertussis toxins and inhib- 48.itors of protein kinase C. Biochem. Biophys. Res. Commun., 169: 1099-1104. 1990.

26. Stricter, R. M., Kunkel, S. L., Showell, H. J., Remick, D. G., Phan, S. H.,Ward, P. A., and Marks, R. M. Endothelial celi gene expression of a neu- 49.trophil chemotactic factor by TNF-a/pAu, LPS, and IL-1 bêta. Science(Washington DC), 243: 1467-1469, 1989.

27. Huber. A. R., Kunkel. S. L.. Todd, R. F.. Ill, and Weiss, S. J. Regulation of 50.transendothelial neutrophil migration by endogenous interleukin-8. Science(Washington DC), 254: 99-102, 1991. 51.

28. Gregory, H., Young, J., Schroder, J. M., Mrowietz, U., and Christophers, E.Structure determination of a human lymphocyte derived neutrophil activating peptide (LYNAP). Biochem. Biophys. Res. Commun., /5/: 883-890, 52.1988.

29. Stricter, R. M.. Chensue, S. W., Basha, M. A., Standiford, T. J., Lynch, J. P.,Baggiolini, M., and Kunkel, S. L. Human alveolar macrophage gene expression of interleukin-8 by tumor necrosis factor-alpha, lipopolysaccharide, and 53.interleukin-1 beta. Am. J. Respir. Cell Mol. Biol., 2: 321-326, 1990.

30. Stricter, R. M., Phan, S. H., Showell, H. J.. Remick. D. G., Lynch, J. P.,Genord, M., Raiford, C., Eskandari, M.. Marks, R. M., and Kunkel. S. L.Monokine-induced neutrophil chemotactic factor gene expression in humanfibroblasts. J. Biol. Chem., 264: 10621-10626, 1989. 54.

31. Schroder, J. M., Sticherling, M., Henneicke, H. H., Preissner, W. C., andChristophers, E. IL-1 alpha or tumor necrosis factor-a/pAa stimulate releaseof three NAP-l/IL-8-related neutrophil chemotactic proteins in human dermal fibroblasts. J. Immunol.. 144: 2223-2232, 1990. 55.

32. Larsen, C. G., Anderson, A. O., Oppenheim, J. J., and Matsushima, K.Production of interleukin-8 by human dermal fibroblasts and keratinocytes inresponse to interleukin-1 or tumour necrosis factor. Immunology, 68:31-36,1989. 56.

33. Einer, V. M., Stricter, R. M., Einer. S. G., Baggiolini, M., Lindley, I., andKunkel, S. L. Neutrophil chemotactic factor (IL-8) gene expression by cy-tokine-treated retinal pigment epithelial cells. Am. J. Pathol., 136: 745-750,1990. 57.

34. Brennan, F. M., Zachariae, C. O., Chantry, D., Larsen, C. G., Turner, M.,Maini, R. N., Matsushima, K., and Feldmann, M. Detection of interleukin 8biological activity in synovial fluids from patients with rheumatoid arthritis 58.and production of interleukin 8 mRNA by isolated synovial cells. Eur. J.Immunol., 20: 2141-2144, 1990.

35. Bazzoni, F., Cassatella, M. A., Rossi, F., Ceska, M., Dewald, B., and Bag- 59.giolini, M. Phagocytosing neutrophils produce and release high amounts ofthe neutrophil-activating peptide 1/interleukin 8. J. Exp. Med., 173: 771-774, 1991. 60.

36. Thornton, A. J., Stricter, R. M., Lindley, I., Baggiolini, M.. and Kunkel, S.L. Cytokine-induced gene expression of a neutrophil chemotactic factor/IL-8in human hepatocytes. J. Immunol., 144: 2609-2613, 1990.

37. Hotta, K., Hayashi, K., Ishikawa, J., Tagawa, M., Hashimoto, K., Mizuno, 61.S., and Suzuki, K. Coding region structure of interleukin-8 gene of humanlung giant cell carcinoma LU65C cells that produce LUCT/interleukin-8:homogeneity in interleukin-8 genes. Immunol. Lett., 24: 165-169, 1990.

38. Standiford, T. J., Kunkel, S. L., Basha, M. A., Chensue. S. W., Lynch, J. P.,Toews, G. B., Westwick, J., and Strieter, R. M. Interleukin-8 gene expression 62.by a pulmonary epithelial cell line. A model for cytokine networks in the lung.J. Clin. Invest., 86: 1945-1953, 1990.

39. Kaashoek, J. G., Mout, R., Falkenburg. J. H., Willemze, R.. Fibbe, W. E., 63.and Landegent. J. E. Cytokine production by the bladder carcinoma cell line5637: rapid analysis of mRNA expression levels using a cDNA-PCR procedure. Lymphokine Cytokine Res., 10: 231-235, 1991.

40. Baggiolini, M., and Dewald, B. Cellular models for the detection and evalu- 64.ation of drugs that modulate human phagocyte activity. Experientia, 44:841-848. 1988.

41. Matsushima, K., Morishita, K., Yoshimura, T., Lavu, S., Kobayashi, Y., Lew,W., Appella, E., Kung, H. F., Leonard, E. J., and Oppenheim, J. J. Molecularcloning of a human monocyte-derived neutrophil chemotactic factor(MDNCF) and the induction of MDNCF mRNA by interleukin 1 and tumor 65.necrosis factor. J. Exp. Med., 167: 1883-1893, 1988.

42. Sticherling, M., Schroder, J. M., and Christophers, E. Production and characterization of monoclonal antibodies against the novel neutrophil activatingpeptide NAP/IL-8. J. Immunol., 143: 1628-1634, 1989. 66.

43. Fraker, P. J., and Speck, J. C., Jr. Protein and cell membrane iodinationswith a sparingly soluble chloroamide, l,3,4,6-tetrachloro-3a,6a-diphrenylg-

4305

lycoluril. Biochem. Biophys. Res. Commun., 80: 849-857, 1978.Dewald, B., and Baggiolini, M. Evaluation of PAF antagonists using humanneutrophils in a microtiter plate assay. Biochem. Pharmacol., 36:2505-2510,1987.Samanta, A. K., Oppenheim, J. J., and Matsushima, K. Interleukin 8 (monocyte-derived neutrophil chemotactic factor) dynamically regulates its ownreceptor expression on human neutrophils. J. Biol. Chem.. 265: 183-189,1990.Standiford, T. J., Strieter, R. M., Chensue, S. W., Westwick, J., Kasahara,K., and Kunkel, S. L. II.4 inhibits the expression of IL-8 from stimulatedhuman monocytes. J. Immunol., 145: 1435-1439, 1990.Hebert, C. A., Luscinskas, F. W., Kiely, J. M., Luis, E. A., Darbonne, W. C.,Bennett, G. L., Liu, C. C., Obin, M. S., Gimbrone, M. A., Jr., and Baker, J.B. Endothelial and leukocyte forms of IL-8. Conversion by thrombin andinteractions with neutrophils. J. Immunol., 145: 3033-3040, 1990.Nourshargh, S., Perkins, J. A., Showell, H. J., Matsushima, K., Williams, T.J., and Collins, P. D. A comparative study of the neutrophil stimulatoryactivity in vitro and pro-inflammatory properties in vivoof 72 amino acid and77 amino acid IL-8. J. Immunol., 148: 106-111, 1992.Gauthier, T., Hamou, M. F., Monod, L., (¡allay.P., Carrel, S., and deTribolet, N. Expression and release of interleukin-1-a/pAa and beta by humanglioblastoma cells in vitro and in vivo. Acta Neurochir., in press, 1992.Lichtor, T., Dohrmann, G. J., and Gurney, M. E. Cytokine gene expressionby human gliomas. Neurosurgery. 26: 788-793. 1990.Fontana, A., Kristensen, F., Dubs, R., and Weber, E. Production of prostag-landin E and an interleukin-1 like factor by cultured astrocytes and C6 gliomacells. J. Immunol., 129: 2413-2419, 1982.Leist, T. P., Frei, K., Kam-Hansen, S., Zinkernagel, R. M., and Fontana, A.Tumor necrosis factor alpha in cerebrospinal fluid during bacterial, but notviral, meningitis. Evaluation in murine model infections and in patients. J.Exp. Med., 167: 1743-1748, 1988.Mustafa, M. M., Ramilo, O., Saez-Llorens, X., Olsen, K. D., Magness, R. R.,and McCracken, G. H., Jr. Cerebrospinal fluid prostaglandins, interleukin 1beta, and tumor necrosis factor in bacterial meningitis. Clinical and laboratory correlations in placebo-treated and dexamethasone-treated patients. Am.J. Dis. Child, 144: 883-887, 1990.Waage, A., Halstensen, A., Shalaby, R., Brandtzaeg, P., Kierulf, P., andEspevik, T. Local production of tumor necrosis factor alpha, interleukin 1,and interleukin 6 in meningococcal meningitis. Relation to the inflammatoryresponse. J. Exp. Med., 170: 1859-1867, 1989.Ramilo, O., Mustafa, M. N.. Porter, J., Saez-Llorens, X., Mensola, J., Olsen,K. D., Luby, J. P., Beutler, B., and McCracken, G. H., Jr. Detection ofinterleukin-Ib but not tumor necrosis factor-a in cerebrospinal fluid of children with aseptic meningitis. Am. J. Dis. Child, 144: 349-352, 1990.Gallo. P., Frei, K., Rordorf, C., Lazdins, J., Tavolato, B., and Fontana, A.Human immunodeficiency virus type 1 (HIV-1) infection of the central nervous system: an evaluation of cytokines in cerebrospinal fluid. J. Neuroim-munol., 23: 109-116, 1989.Righi, M., Mori, L., De-Libero, G., Sironi, M., Biondi, A., Mantovani, A.,Donini, S. D., and Ricciardi-Castagnoli, P. Monokine production by micro-glial cell clones. Eur. J. Immunol., 19: 1443-1448, 1989.Selmaj, K. J., Farooq, M., Norton, W. T., Raine, C. S., and Brosnan, C. F.Proliferation of astrocytes in vitro in response to cytokines. A primary role fortumor necrosis factor. J. Immunol., 144: 129-135, 1990.Chung, I. Y., and Benveniste, E. N. Tumor necrosis factor-alpha productionby astrocytes. Induction by lipopolysaccharide, IFN-gammo, and IL-lbeta. J.Immunol., 144: 2999-3007, 1990.Ramilo, O., Saez-Llorens, X., Mensola, J., Jafari, H., Olsen, K. D., Hansen,E., Yoshinaga, M., Ohkawara, S., Nariuchi, H., and McCracken, G. H.Tumor necrosis factor-a/cachectin and interleukin Ib initiate meningea! inflammation. J. Exp. Med., 772:497-507. 1990.Burrows, L. J., Piper, P. J., Lindley, I. D., and Westwick, J. Intraperitonealinjection of human recombinant neutrophil-activating factor/interleukin 8(hrNAF/IL-8) produces a T cell and eosinophil infiltrate in the guinea piglung. Effect of PAF agonist WEB2086. Ann. NY Acad. Sci., «29:422-424,1989.Brooks, W. H., Netsky, M. G., Normansell, D. E., and Horwitz, D. A.Depressed cell-mediated immunity in patients with primary intracranial tumors. J. Exp. Med., 136: 1631-1647, 1972.Weisbarth, R. H.. Chan, G., Spotter, L., and Golde, D. W. Further purification of neutrophil migration inhibitory factor from T lymphocytes (NIF-T):evidence that NIF-T and leukocyte inhibitory factor (LIF) are immunologi-cally distinct. Clin. Immunol. Immunopathol., 32: 269-274, 1984.List, J., Moser, R. P., Steuer, M., Loudon, W. G., Blacklock, J. B., andGrimm, E. A. Cytokine responses to intraventricular injection of interleukin2 into patients with leptomeningeal carcinomatosis: rapid induction of tumornecrosis factor a, interleukin Ib, interleukin 6. 7-interferon, and solubleinterleukin 2 receptor (M, 55,000 protein). Cancer Res.. 52: 1123-1128,1992.Saukkonen, K., Sande, S., Cioffe, C., Wolpe, S., Sherry, B., Cerami, A., andTuomanen, E. The role of cytokines in the generation of inflammation andtissue damage in experimental gram-positive meningitis. J. Exp. Med., 777:439-448, 1990.Schall, T. J., Bacon, K., Toy, K. J., and Goeddel, D. V. Selective attractionof monocytes and T lymphocytes of the memory phenotype by cytokineRANTES. Nature (Lond.), 347: 669-671, 1990.

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1992;52:4297-4305. Cancer Res   Erwin Van Meir, Miroslav Ceska, Fritz Effenberger, et al.   of the Human Central Nervous SystemInterleukin-8 Is Produced in Neoplastic and Infectious Diseases

  Updated version

  http://cancerres.aacrjournals.org/content/52/16/4297

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/52/16/4297To request permission to re-use all or part of this article, use this link

on May 10, 2019. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from