Expression of Proinflammatory and Proangiogenic Cytokines inPatients with Head and Neck Cancer1

Zhong Chen, Pramit S. Malhotra,Giovana R. Thomas, Frank G. Ondrey,Dianne C. Duffey, Conrad W. Smith,Ileana Enamorado, Ning T. Yeh, Glenn S. Kroog,Susan Rudy, Linda McCullagh, Shaker Mousa,Matha Quezado, Laurie L. Herscher, andCarter Van Waes2

Tumor Biology Section, Head and Neck Surgery Branch, NationalInstitute on Deafness and Other Communication Disorders, NIH,Bethesda, Maryland 20892-1419 [Z. C., P. S. M., G. R. T., F. G. O.,D. C. D., C. W. S., I. E., N. T. Y., G. S. K., S. R., C. V. W.]; WarrenGrant Magnussen Clinical Center, NIH, Bethesda, Maryland 20892-1419 [L. M.]; DuPont-Merck Research and Development,Wilmington, Delaware 19880-0400 [S. M.]; Surgical PathologySection, Pathology Branch, National Cancer Institute, NIH, Bethesda,Maryland 20892-1419 [M. Q.]; and Radiation Oncology Branch,National Cancer Institute, NIH, Bethesda, Maryland 20892-1419[L. L. H.]

ABSTRACTAltered immune, inflammatory, and angiogenesis re-

sponses are observed in patients with head and neck squamouscell carcinoma (HNSCC), and many of these responses havebeen linked with aggressive malignant behavior and a decreasein prognosis. In this study, we examined the hypothesis thatHNSCC cells produce cytokines that regulate immune, inflam-matory, and angiogenesis responses. We identified importantregulatory cytokines in supernatants of well-defined andfreshly cultured HNSCC cell lines by ELISA and determinedwhether these cytokines are detected in tumor cell lines andtissue specimens by immunohistochemistry. The serum con-centration of the cytokines and cytokine-dependent acute phaseinflammatory responses (i.e., fibrinogen, C-reactive protein,and erythrocyte sedimentation rate) from patients withHNSCC was determined, and the potential relationship ofserum cytokine levels to tumor volume was analyzed. Cyto-kines interleukin (IL)-1 a, IL-6, IL-8, granulocyte-macrophagecolony-stimulating factor (GM-CSF), vascular endothelialgrowth factor (VEGF), and basic fibroblast growth factor weredetected in similar concentration ranges in the supernatants of

a panel of established University of Michigan squamous cellcarcinoma (UM-SCC) cell lines and supernatants of freshlyisolated primary HNSCC cultures. Evidence for the expressionof IL-1 a, IL-6, IL-8, granulocyte-macrophage colony-stimulat-ing factor, and VEGF in HNSCC cells within tumor specimensin situ was obtained by immunohistochemistry. In a prospec-tive comparison of the cytokine level and cytokine-inducibleacute-phase proteins in serum, we report that cytokines IL-6,IL-8, and VEGF were detected at higher concentrations in theserum of patients with HNSCC compared with patients withlaryngeal papilloma or age-matched control subjects (atP <0.05). The serum concentrations of IL-8 and VEGF were foundto be weakly correlated with large primary tumor volume (R2

5 0.2 and 0.4, respectively). Elevated IL-1- and IL-6-inducibleacute-phase responses were also detected in cancer patients butnot in patients with papilloma or control subjects (at P < 0.05).We therefore conclude that cytokines important in proinflam-matory and proangiogenic responses are detectable in cell lines,tissue specimens, and serum from patients with HNSCC. Thesecytokines may increase the pathogenicity of HNSCC and proveuseful as biomarkers or targets for therapy.

INTRODUCTIONAlterations in host immunity, inflammation, angiogenesis,

and metabolism have been noted to be prominent clinical fea-tures in patients with HNSCC3 (1–3). These tumor-inducedT-lymphocyte, granulocyte, and neoangiogenesis responses inthe local tumor microenvironment (1–3) have been associatedwith increased growth and metastasis and decreased survival(4–6). Pathological changes in systemic responses have alsobeen observed, including induction of antibody and other acute-phase inflammatory protein responses (7, 8); a decrease indelayed-type hypersensitivity (9, 10), T-cell and natural killercell-mediated immunity (11, 12); and changes in catabolism thatoften lead to marked weight loss and hypercalcemia (13, 14).Although the origin of the signals and mechanisms underlyingthese responses are not well understood, the local and systemicnature of these responses suggests the hypothesis that cytokineswith proinflammatory, proangiogenic, and immunoregulatoryactivity are produced by squamous cell carcinomas and couldcontribute to the pathogenesis of HNSCC.

We reported previously that murine SCC cell lines derivedafter tumor progression exhibit increased malignant behavior inassociation with expression of a repertoire of proinflammatoryand proangiogenic cytokines (15–17). We found that humanReceived 11/12/98; revised 3/9/99; accepted 3/11/99.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisementin accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 This work was supported by National Institute on Deafness and OtherCommunications Disorders (NIDCD) Intramural Research Project Z01-DC-00016 and in part by NIDCD-DuPont Merck Cooperative Researchand Development Agreement DC 95-01.2 To whom requests for reprints should be addressed, at NIH, Building10, Room 5D55, MSC-1419, Bethesda, MD 20892-1419.

3 The abbreviations used are: HNSCC, head and neck squamous cellcarcinoma; IL, interleukin; VEGF, vascular endothelial growth factor;GM-CSF, granulocyte-macrophage colony-stimulating factor; TNF, tu-mor necrosis factor; TGF, transforming growth factor; FGF, fibroblastgrowth factor.

1369Vol. 5, 1369–1379, June 1999 Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

HNSCC cell lines express a similar repertoire of cytokines,including IL-1a, IL-6, IL-8, and GM-CSF, and that IL-1ainduces the expression of other members of this cytokine rep-ertoire by HNSCC cells (18). One or more of these cytokineshave also been detected in tumor homogenates, primary cul-tures, and established HNSCC lines from patients by others(19–22). The extent to which these and other cytokines impor-tant in the regulation of immune, inflammatory, and angiogen-esis responses can be detected in the local tumor environmentand systemically in patients remains to be determined.

In the present study, we surveyed the expression of 14 cyto-kines important in the regulation of immune, inflammatory, andangiogenesis responses in well-defined and freshly culturedHNSCC lines and sought to determine whether these cytokines areexpressed and can be detected in the local tumor environment andsystemicallyin vivo. Proinflammatory and proangiogenic cytokineswere most commonly detected, including IL-1a, IL-6, IL-8, GM-CSF, VEGF, and basic FGF. Other important proinflammatorycytokines, such as IL-1b, TNF-a, and TGF-b, and immunoregu-latory cytokines, such as IL-2, IL-12, IFN-g, IL-4, and IL-10 werenot detected in the HNSCC supernatants. We reported here thatcytokines IL-1a, IL-6, IL-8, GM-CSF, and VEGF were detected inHNSCCin situ by immunohistochemistry. Furthermore, elevationin the levels of IL-6, IL-8, and VEGF, as well as cytokine-inducibleacute phase responses, were found in the serum from patients withHNSCC. We therefore conclude that expression of cytokines im-portant in proinflammatory and proangiogenic responses were de-tected in the tumor environment and systemically in patients withHNSCC.

MATERIALS AND METHODSHuman Subjects. All tissue for derivation of cell lines

and immunohistochemical analysis and serum for cytokine andacute-phase reactant studies were obtained with informed con-sent under Institutional Review Board-approved protocols at theUniversity of Michigan (UM-SCC cell lines) or the NIH (97-DC-0044; primary NIH-SCC cell cultures, tissue, and serum).Patients with HNSCC treated at the University of Michigan orNIH received standard treatment with surgery without or withradiation or investigational therapy with concurrent paclitaxel

and radiation therapy under NIH Institutional Review Board-approved protocol 95-C-0162. Patients with papilloma under-went standard microlaryngoscopic excision with CO2 laser.

Culture of Established UM-SCC Squamous Cell Carci-noma Lines for Analysis of Cytokine Secretion. To identifycytokines produced by human HNSCC, we screened sevenwell-characterized cell lines from the UM-SCC series for pro-duction of a panel of 14 important regulatory cytokines withproinflammatory, proangiogenic, and immunoregulatory activ-ity by ELISA. The panel of seven established SCC cell linesfrom the University of Michigan series (UM-SCC) were ob-tained from Dr. T. E. Carey at the University of Michigan (AnnArbor, MI; Ref. 23). The UM-SCC cell lines were derived fromfive patients with SCC of the upper aerodigestive tract (Table 1).The seven UM-SCC cell lines selected were obtained frompatients with stages I–IV tumors, distributed among oral, pha-ryngeal, and laryngeal sites, and who died within 2 years oftherapy (Table 1). Cell lines established from single isolates ofa patient specimen are designated by a numeric designation, andwhere isolates from two time points or anatomical sites wereobtained, the designation includes an alphabetical suffix (i.e.,Aor B). Karyotype and epithelial marker analyses of the panel ofUM-SCC cell lines used have been published previously (24,25). The expression of the SCC tumor-associated A9 antigen bythe cell lines was confirmed in our laboratory prior to use (26).Cell lines used were tested and found to be free ofMycoplasmacontamination by two independent assays, Gibco MycoTect kit(Life Technologies, Inc., Gaithersburg, MD) andMycoplasmaDetection kit (Boehringer Mannheim, Manneim, Germany).Cell lines from relatively early (UM-SCC9, P 13; UM-SCC-11A, P 26 and B, P 48) or later passages (UM-SCC-38, P 49;UM-SCC-1, P 59; UM-SCC-22A and B, P 51, P 60) that wereavailable were included. Studies were performed using frozenstocks within five passages of receipt. The cell lines weremaintained in Eagle’s minimal essential media supplementedwith 10% fetal bovine serum and penicillin/streptomycin. Toprepare supernatants from UMSCC cell lines for ELISA, 12 mlof fresh medium were added to established UM-SCC tumor celllines when 60–80% confluent in 75-cm2 tissue culture flasksand collected after 48 h. Supernatants were centrifuged at

Table 1 Tumor treatment and response characteristics of patients from which UM-SCC cell lines were developed

Patient Agea Gender Stage TNMb Primary sitecSpecimen

sitedPrior

therapye Statusf Survivalg Cell line

1 72 M I T1N0M0 FOM Local recur R DWOD 15 UMSCC-12 72 F II T2N0M0 Tonsil/BOT Local recur R DOD 15 UMSCC-93 65 M IV T2N2aM0 Hypopharynx Pri bx N DOD 14 UMSCC-11A

Pri resect C UMSCC-11B4 59 F III T2N1M0 Hypopharynx Pri bx N DOD 10 UMSCC-22A

LN met N UMSCC-22B5 60 M IV T2N2aM0 Tonsil/BOT Pri N DOD 11 UMSCC-38a Patient age in years at diagnosis.b Tumor Node Metastasis system of staging.c FOM, floor of mouth; BOT, base of tongue.d Origin of tissue used to establish cultures; recur, recurrence, Pri, primary tumor site; bx, biopsy; resect, surgical resection specimen; LN, lymph

node; met, metastasis.e Therapy given before the specimen used for culture was obtained: N, none; R, radiation; C, chemotherapy.f DWOD, died without disease; DOD, died with disease.g Survival in months from diagnosis to last follow up.

1370Cytokines in Head and Neck Squamous Cell Carcinoma

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

15003 g to remove particles, and aliquots were stored frozen at280° until use in ELISA. Cell number in the culture flasks wasdetermined to standardize the quantity of cytokine secreted per106 tumor cells.

Primary Culture of Tumors From SCC Patients forAnalysis of Cytokine Secretion. To establish whether thesecytokines could be detected in supernatants of freshly isolatedHNSCC, we established nine primary HNSCC tumor cultures anddetermined the concentration of these cytokines in culture super-natants as part of a prospective study of immunoregulatory factorexpression in patients with HNSCC. Nine primary SCC cell lineswere successfully established from 7 of 11 untreated patients withadvanced stage III and IV HNSCC, distributed among oral, pha-ryngeal, and laryngeal sites (Table 2). The cell lines were givennumeric designation in order of receipt. Two tumor specimensavailable from patient 4 who underwent both a biopsy and surgerywere designated as 4A and 4B; and two tumor specimens frompatient 8 were from the primary tumor site (A) and metastaticlymph nodes (B). SCC tissue was prepared for cell culture after themethod of Krauseet al. (23). Briefly, the tumor specimens werewashed three times with complete EMEM medium, which contains10% FCS, 13 Antibiotic-Antimycotic solution (100 IU/ml peni-cillin, 100 mg/ml streptomycin, and 0.25mg/ml fungizone; LifeTechnologies, Inc.), 50mg/ml gentamicin (Cellgro, Herndon, VA),and 2 mM glutamine. The tissues were cut into small fragments ofabout 1-mm in diameter and covered with 2 drops of medium. Thenext day, 1 ml of medium was added into each flask, and themedium were changed every 3–4 days. After 2–3 weeks, culturescontaining colonies with epithelial morphology of 1–2-cm diameterwere selected and incubated with 2 ml of fresh medium for 48 h.Supernatants were then collected, centrifuged to remove cell debris,aliquoted, and stored at280°C until use in ELISA.

Immunohistochemical Staining of Cytokines in SCCTissue and Cultured Cells. Immunohistochemical staining ofhuman cytokines in SCC tissue sections was performed accord-ing to the method of Andersson and Andersson (27). UM-SCC-11B cells, which secrete cytokines IL-1a, IL-6, IL-8, and GM-CSF at elevated levels (18), were used as a positive control.

Cultured UM-SCC-11B cells (1–23 104) were plated on 8-wellchamber slides (Lab-Tek, Naperville, IL) for 2–3 days and werestained as positive controls (18). Cultured cells and frozen tissuesections (9–10mm) from SCC specimens preserved in OCTwere fixed and permeabilized in freshly made cold 4%paraformaldehyde (15 min) and 0.1% saponin (Sigma ChemicalCo., St. Louis, MO) in HBSS (Mediatech, Herndon, VA) con-taining 0.2% sodium azide (Sigma) and 1% H2O2 (Fisher Sci-entific, Fair Lawn, NJ) for blockade of endogenous peroxidase.The remaining procedures were carried out at room temperature.Ten % serum (Vector Lab, Inc., Burlingame, CA) was used toblock the nonspecific binding sites for 20 min and removedwithout washing. The samples were incubated with the follow-ing primary antibodies: 10mg/ml of mouse anti-human IL-1a(IgG1; Pharmingen, San Diego, CA), 5mg/ml of rat anti-humanIL-6 (IgG2a; Pharmingen), 10mg/ml of mouse anti-human IL-8antibody (IgG1; Genzyme, Cambridge, MA), 5mg/ml of mouseanti-human GM-CSF (IgG2a; PharMingen), 0.2mg/ml of mouseanti-human VEGF (IgG1; Santa Cruz Biotechnology, SantaCruz, CA), 5 mg/ml of mouse anti-human pan cytokeratin(IgG1; Novocastra Lab, Newcastle upon Tyne, United King-dom), or isotype controls, mouse IgG (Vector Lab) or rat IgG2a(Pharmingen), diluted in 0.1% saponin in balanced salt solutionwith 10% serum overnight at 4°C. The samples were blockedwith 5% serum for 20 min and incubated with the secondarybiotinylated antibody for 30 min (1:200 in 0.1% saponin inHBSS containing 5% serum; Vector Lab), followed by 30-minincubation with biotin/avidin horseradish peroxidase conjugates(Vectastain Elite ABC kit; Vector Lab) and 5–8 min withchromogen diaminobenzidine tetrahydrochloride (Vector Lab)according to the manufacturer’s specifications.

Serum for Cytokine and Acute-Phase Reactant Studies.Serum from 11 patients with HNSCC, 12 patients with squa-mous papilloma of the upper aerodigestive tract, and 12 unaf-fected age- and gender-matched control subjects were obtainedfor a prospective comparison of cytokine concentration andcytokine-dependent acute-phase responses (Table 3). Amongthe three subject groups, there were no significant differences in

Table 2 Tumor, treatment, and response characteristics of patients with head and neck squamous cell carcinoma from which primary SCC celllines, serum, and tissue were obtained

Patient Age Gender Stage TNM Primary siteSpecimen

siteaPrior

therapy Statusb Monthsc Cell line

1 64 F IV T4N1M0 Lateral tongue Pri bx N NED 19 NIH-SCC-12 24 M IV T4N0M0 Nasopharynx Pri bx N NED 18 NIH-SCC-23 77 M IV T3N2M0 Lateral tongue Pri bx N AWD 17 NIH-SCC-34 63 M IV T4N0M0 Retromolar Pri bx N AWD 16 NIH-SCC-4A

Trigone Pri resect NIH-SCC-4B5 53 F IV T4N1M0 Supraglottis Pri bx N NED 16 (2)6 55 M IV T3N3M1 Tongue base Pri bx N DOD 3 (2)7 29 M IV T4N3M0 Nasopharynx Pri bx N AWD 13 NIH-SCC-78 42 M IV T4N1M0 Maxillary sinus Pri bx N NED 12 NIH-SCC-8A

LN met NIH-SCC-8B9 72 F IV T4N2cM0 Tonsil Pri bx N DOD 9 NIH-SCC-9

10 46 M III T2N1M0 Tonsil Pri bx N NED 13 (2)11 46 M IV T4N3M0 Pyriform sinus Pri bx N NED 6 (2)a Pri, primary tumor site; bx, biopsy; resect, surgical resection specimen; LN, lymph node; met, metastasis.b NED, no evident disease; AWD, alive with disease; DOD, died with disease.c Months of follow-up after therapy.

1371Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

terms of age (mean age: HNSCC patients, 52 years; papillomapatients, 44 years; normal subjects, 55 years; Student’st test,P . 0.18), gender (Fisher’s exact test,P . 0.6), or alcoholconsumption (Fisher’s exact test,P . 0.2). There was a differ-ence in smoking history between patients with SCC (8 of 11)and normal subjects (3 of 12; Fisher’s exact test,P 5 0.04), andthis difference was consistent with differences in the prevalenceof smoking between patients with HNSCC and the generalpopulation. No subjects had a history of prior malignancy,immunodeficiency, autoimmune disorders, hepatitis, or HIVinfection. Blood was drawn from control subjects and patientsprior to treatment. Sera were collected by centrifuging wholeblood at 3000 rpm for 15 min at 10°C by a Sorvall RT6000Dcentrifuge (DuPont, Wilmington, DE), and the aliquots werestored at280°C until used in ELISA assay.

ELISA for Quantitation of Cytokine Concentration inSupernatants from Cultured Human SCC and Serum fromPatients and Normal Subjects. ELISA kits for specific cy-tokines were used (R&D systems, Minneapolis, MN, and En-dogen, Cambridge, MA) according to the manufacturer’s pro-tocol. Each sample was tested in duplicate in each of two ormore replicate experiments. After development of the colori-metric reaction, the absorbance at 450 nm was quantitated by aneight channel spectrophotometer (Biotek Systems, Winooski,VT), and the absorbance readings were converted to pg/mlbased upon standard curves obtained with recombinant cytokinein each assay. If the absorbance readings exceeded the linearrange of the standard curves, ELISA assay was repeated afterserial dilution of the supernatants. Each serum sample wastested in at least two independent ELISA experiments, and thedata were calculated from mean of two to four tests for eachsample.

Determination of Tumor Volume. Tumor volumeswere calculated from computerized tomographic images usingPowerTPS, the National Cancer Institute Radiation OncologyTreatment Planning System. For each patient, the tumor andinvolved lymph nodes were manually selected and contoured oneach axial computed tomography image. Tumor areas were firstdetermined in pixels by the intersection of the contoured regionwith the image matrix and then converted to slice volume bymultiplying the number of selected pixels by the pixel area andthe slice thickness. Total tumor volumes were then taken as theintegral of the slice volumes divided by the total number ofslices in the image set.

Measurement of Acute-Phase Responses.Fibrinogen,C-reactive protein, and sedimentation rate acute-phase re-sponses were measured by standardized clinical laboratory

methods in the Clinical Pathology Department of the WarrenGrant Magnussen Clinical Center at NIH using serum obtainedprior to treatment.

Histopathological Study. The cases were retrieved fromthe surgical pathology files of the National Cancer Institute.H&E-stained slides from in-house biopsies or from referringinstitutions (Kaiser Permanente, Kensington, MD) were re-viewed in all of the cases. The pathological diagnosis, degree ofdifferentiation, and infiltration of tumor by inflammatory cellswere graded independently by Dr. Martha Quezado of the Sur-gical Pathology Section, Pathology Branch, National CancerInstitute. Slides from patients 1 and 11 were not available for thereview at the time of this study, and the data were collected fromthe surgical pathology reports.

Statistical Analysis. Differences in gender, smoking,and alcohol consumption among SCC patients, papilloma pa-tients, and normal subjects were tested by Fisher’s exact testusing statistical software Prism 2.01. All of the other data werecalculated, and statistical significance was tested by independentStudent’st test using software SigmaPlot Scientific Graph Sys-tem (Jandel Scientific, San Rafael, CA).

RESULTSSecretion of Cytokines IL-1a, IL-6, IL-8, GM-CSF,

VEGF, and Basic FGF by Established and Primary HNSCCCell Lines. The characteristics of the patients and UM-SCCcell lines studied are shown in Table 1. Cytokines IL-1a, IL-6,IL-8, GM-CSF, VEGF, and basic FGF were most commonlydetected in culture supernatants from UM-SCC cell lines, assummarized in Table 4. Other important proinflammatory cyto-kines, such as IL-1b, TNF-a, and TGF-b, and immunoregula-tory cytokines IL-2, IL-12, IFN-g, IL-4, and IL-10 were notdetected by ELISA in the UM-SCC supernatants. Among the sixcytokines detected in the UM-SCC cell culture supernatants,IL-6, IL-8, and VEGF were detected in the highest concentra-tions.

The characteristics of the patients and primary SCC celllines studied are shown in Table 2. Fig. 1 shows photomicro-graphs of the typical epithelial histology and morphology ofrepresentative tumors and primary cultures from patient 3 (Fig.1, A–C) and patient 4 (Fig. 1,D–F). The squamous epithelialorigin of the patient tumors was confirmed by immunohisto-chemical staining with pan cytokeratin, as shown in Fig. 1B.Cultured cells exhibited typical squamous epithelial morphol-ogy (Fig. 1,C andF). Table 5 shows that the primary HNSCCcultures produced the six cytokines in a range of concentrations

Table 3 Demographics of patients and control subjects

SubjectsSample size

(n)Gendera

(M/F)Age (yr)b

Mean6 SE (range) Alcohol usec Tobacco used

Cancer 11 8/3 51.96 5.0 (24–77) 8 8Papilloma 12 8/5 43.66 3.6 (25–73) 8 7Normal 12 9/4 54.86 4.5 (29–77) 8 3

a There is no statistically significant difference in gender between groups (Fisher’s exact test,P . 0.6).b There is no statistically significant difference in ages between groups (t test,P . 0.18).c There is no statistically significant difference in consuming tobacco and alcohol between groups (Fisher’s exact test,P . 0.2).d A significant difference is observed when comparing smoking history between SCC patients and control subjects (Fisher’s exact test,P 5 0.04).

1372Cytokines in Head and Neck Squamous Cell Carcinoma

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

similar to those detected in the supernatant of established UM-SCC cell lines. Cytokines IL-6, IL-8, and VEGF were producedin highest concentration by both established and freshly isolatedHNSCC cultures.

Detection of IL-1a, IL-6, IL-8, GM-CSF, and VEGF inHNSCC in Situ. We examined whether the cytokines ex-pressed in established and freshly isolated HNSCC culturescould be detected in HNSCC cells within tumor tissuein situby immunohistochemical staining. Fig. 2,A–E, shows thatHNSCC exhibit immunostaining with cytokine-specific anti-bodies for IL-1a, IL-6, IL-8, GM-CSF, and VEGF. A weakerstaining was evidenced with IL-6, and the staining withIL-8-specific antibody exhibited a nonhomogeneous pattern(Fig. 2, B and C). The immunostaining of cytokines IL-1a,IL-6, IL-8, GM-CSF, and VEGF in Fig. 2,A-E was detectedwithin keratin-positive malignant epithelium (Fig. 2,F–J),

whereas significant staining with GM-CSF was also detectedin the stroma of specimens (Fig. 2D). The strongest VEGFreactivity (Fig. 2E) was noted to be localized in areas oftumor that exhibited less epithelial differentiation and weakerkeratin staining (Fig. 2J). The staining of tissue with cytokineantibodies was greater than that observed with isotype con-trols (Fig. 2,K–O). Furthermore, the binding of these cyto-kine-specific antibodies to HNSCC could be specificallyinhibited by competition with recombinant cytokine in adose-dependent manner.4 We conclude that cytokines IL-1a,IL-6, IL-8, GM-CSF, and VEGF can be detected in HNSCCin situ.

4 Z. Chenet al., manuscript in preparation.

Fig. 1 Histology of tumor specimens and morphology of primary cell cultures from patients with HNSCC. Light photomicrographs of H&E-stainedtumor specimens from patient 3 (AandB) and patient 4 (DandE), showing representative areas of moderately differentiated HNSCC (3400). Theepithelial origin of tumor from patient 3 was confirmed by pan cytokeratin staining (B,3400). Typical morphology of epithelial colonies in primarycell cultures from patient 3 (C) and patient 4 (F), taken under phase contrast microscopy (3320).

Table 4 Secretion of cytokines by UMSCC cell lines

Cell linesa

Cytokinesb,c

Proinflammatory Proangiogenic

IL-1a IL-6 IL-8 GM-CSF VEGF Basic FGF

UMSCC-1 111c 111 1111 1 1111 11UMSCC-9 111 11 1111 111 1111 1UMSCC-11A 111 1111 1111 111 1111 11UMSCC-11B 111 1111 11111 1111 1111 111UMSCC-22A 11 2 111 2 1111 2UMSCC-22B 11 2 111 2 1111 2UMSCC-38 2 11 1111 11 1111 2

a UM-SCC (University of Michigan) cell lines.b Picograms/106 cells/48 h.1, 1–10;11, 10–100;111, 100–1,000;1111, 1,000–10,000.c Proinflammatory cytokines, IL-1b, TNF-a, TGF-b, and immunoregulatory cytokines IL-2, IL-12,g-IFN, IL-4, and IL-10 were also screened

in UM-SCC cell lines. The cell lines did not produce significant detectable levels of the cytokines.

1373Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Elevated IL-6, IL-8, and VEGF Cytokine Levels inSerum from Patients with HNSCC. On the basis of thefindings that proinflammatory and proangiogenic cytokines maybe expressed by HNSCCin vitro and in the local tumor envi-ronmentin situ, we prospectively examined and compared thelevels of these cytokines in the serum of unaffected subjects andpatients with HNSCC and benign squamous papilloma of theupper aerodigestive tract using ELISA. Serum from 11 patientswith HNSCC, 12 patients with laryngeal squamous papilloma,and 12 unaffected age- and gender-matched control subjectswere obtained (Table 3). Among the three subject groups, therewere no significant differences in terms of age, gender, oralcohol consumption. The difference in smoking history be-tween SCC (8 of 11) and normal subjects (3 of 12) was con-sistent with the increased prevalence of smoking history inpatients with HNSCC when compared with the general popula-tion. Fig. 3 shows that cytokines IL-6, IL-8, and VEGF weredetected in higher concentration in the serum of patients withHNSCC when compared with either patients with squamouspapilloma or control subjects (ttest, P , 0.05). When wecompared the serum cytokine levels between patients with pap-illoma and control subjects, we did not find significant differ-ences in the serum concentration of the three cytokines (Fig. 3).Within the limited sample size studied, we were unable to detectdifferences between smoking and nonsmoking subjects (data notshown). Concentrations of cytokines IL-1a, GM-CSF, and basicFGF were detected at the limits of sensitivity of the assays, andno significant differences were detected in serum of patientswith HNSCC, squamous papilloma, or normal subjects (data notshown).

Relationship of Proinflammatory and ProangiogenicCytokine Expression to Tumor Volume, Differentiation, andInflammation. We examined whether the expression ofproinflammatory and proangiogenic cytokines in the patientswith HNSCC was associated with differences in tumor volume,differentiation, and inflammation. Table 6 displays the concen-tration of cytokines IL-6, IL-8, and VEGF in serum, along withthe stage, volume, and the presence of differentiation and in-flammation in the tumor for each patient. The relationship of the

serum cytokine concentration to tumor volume in patients withHNSCC was examined. Tumor volume was calculated from thecomputed tomography/magnetic resonance imaging scans ob-tained for treatment planning, and the relationship of serumcytokine concentration to both primary and metastatic tumorvolume in the lymph nodes was determined. A weak relation-ship between primary tumor volume and VEGF concentration(Fig. 4, upper panel,linear regressionR2 5 0.420) and IL-8concentration (Fig. 4,lower panel,linear regressionR2 5 0.214)was detected. There was no correlation between IL-6 concen-tration with tumor volume. No significant relationship wasfound between the cytokines with tumor volume in metastaticlymph nodes or total tumor volume consisting of the sum ofvolumes of the primary and lymph node tumor volumes (Table6). When we examined whether the differences in cytokinelevels observed in serum were associated with differences in theother pathological correlates, such as differentiation or presenceof leukocytes, no relationship was evident (Table 6). Leukocyteswere detected within the tumor specimens of all of the patientswith HNSCC.

Increase in IL-1- and IL-6-dependent Acute Phase Re-sponses in Patients with HNSCC. We examined whetherthere was evidence for induction of acute phase proteins in thepatients with upper aerodigestive neoplasms. Fig. 5 shows thatpatients with HNSCC exhibit an increase in the mean fibrino-gen, C-reactive protein, and the erythrocyte sedimentation ratewhen compared with patients with papilloma or unaffectedsubjects (ttest,P , 0.05). There were no significant differencesin the values of fibrinogen, C-reactive protein, and the erythro-cyte sedimentation rate observed between papilloma patientsand normal subjects or within the groups between smoking andnonsmoking subjects (data not shown). The increase in acute-phase proteins detected in the patients with HNSCC showed nocorrelation with tumor volume and was not associated withconcurrent infection, fever, or elevation in WBC count (data notshown). There were no consistent changes in the concentrationof several other acute-phase components tested, includinga-1anti-trypsin,a-2 macroglobulin, complements factors C3/C4 orfactor B, haptoglobin, or albumin (data not shown,P . 0.05).

Table 5 Secretion of cytokines by primary head and neck SCC lines

Cell line

Cytokinesa

Proinflammatory Proangiogenic

IL-1a IL-6 IL-8 GM-CSF VEGF Basic FGF

NIH-SCC-1 11 111 1111 111 NDb NDNIH-SCC-2 11 111 1111 111 ND NDNIH-SCC-3 111 111 1111 111 ND NDNIH-SCC-4Ac 11 111 1111 111 ND NDNIH-SCC-4B 11 111 1111 111 ND NDNIH-SCC-7 2 111 1111 111 1111 2NIH-SCC-8Ad 11 111 1111 1111 1111 1NIH-SCC-8B 2 111 1111 11 111 2NIH-SCC-9 11 111 1111 1111 111 11

a The level of cytokine production in the supernatants was shown as picograms/ml/48 h.1, 1–10;11, 10–100;111, 100–1,000;1111,1,000–10,000.

b ND, not done.c 4A and 4B primary cultures were isolated from the biopsy (A) and surgery (B) specimens.d 8A and 8B primary cultures were isolated from the primary tumor site (A) and lymph node metastasis (B).

1374Cytokines in Head and Neck Squamous Cell Carcinoma

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

DISCUSSIONIn this study, we show that established and primary SCC

lines from patients with head and neck cancer can produceproinflammatory cytokines IL-1aIL-6, IL-8, and GM-CSF andproangiogenic cytokines VEGF and basic FGF. In an extendedsurvey of other important proinflammatory and immunoregula-tory cytokines, we did not detect secretion of IL-1b, IL-2, IL-4,IL-10, IL-12, TNF-a, or TGF-b in any of the UM-SCC celllines. Several of the proinflammatory cytokines detected in thisstudy, including IL-1a, IL-6, IL-8, and GM-CSF, have beendetected in tumor homogenates, primary cultures, establishedHNSCC lines, and in serum from patients by ourselves andothers (18–22). In addition, we show that cytokines IL-1a, IL-6,IL-8, GM-CSF, and VEGF could be detectedin situ in HNSCCwithin the local tumor environment. The detection of thesecytokines by immunohistochemical methods is consistent withsimilar findings for IL-1a, and IL-6 by others (22), and resultswe obtained demonstrating localization of IL-1a, IL-6, andGM-CSF mRNA in HNSCCin situ (28). We conclude that the

six cytokines selectively produced by HNSCC possess proin-flammatory, proangiogenic, and immunoregulatory activitiesthat are consistent with pathological alterations observed inpatients with these neoplasms. These cytokines may play animportant role in promoting tumorigenesis.

We compared the concentrations of these cytokines in theserum of patients with HNSCC, benign squamous papilloma,and unaffected age- and gender-matched controls in a prospec-tive study to explore whether elevated levels of these cytokinesmay be detected in serum of HNSCC patientsin vivo. Weobtained evidence that cytokines IL-6, IL-8, and VEGF, as wellas proinflammatory cytokine-induced, acute-phase proteins, aredetected at higher concentrations in the serum of patients withHNSCC but not in patients with benign squamous papilloma orunaffected control subjects. Within the repertoire of proinflam-matory and proangiogenic cytokines detected in the superna-tants of established and primary HNSCC cell lines, the mostabundant cytokines detected were IL-6, IL-8, and VEGF, whichis consistent with the elevation of these cytokines detected in the

Fig. 2 Detection of IL-1a, IL-6, IL-8,GM-CSF, and VEGF in HNSCC by im-munohistochemistryin situ. IL-1a (A),IL-6 (B), IL-8 (C), GM-CSF (D), andVEGF (E) were detected in primary tumorspecimens from HNSCC by immunohisto-chemical staining, as described in “Mate-rials and Methods.” Positive staining oftumor with pan cytokeratin antibody inadjacent sections is demonstrated (F–J)but is not observed with matched isotypecontrol antibodies (K–O). Specimens fromA, F, K andC, H, M were counterstainedwith hematoxylin to demonstrate nuclearstructures. The magnification of all lightphotomicrographs was3400, except forE, J, O,which were taken at3200.

1375Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

serum of patients with HNSCC (Fig. 3). Although IL-1a wasnot detected at significant levels in serum, we did detect IL-1ain situ and observed an elevation in the concentration of certainIL-1a- and IL-6-dependent acute-phase responses (Fig. 5).Within the limited sample studied, we did not detect increasedconcentrations of GM-CSF and basic FGF in serum, althoughthese two cytokines were detected in cell culture supernatants(Tables 4 and 5) orin situ (Fig. 2). We have not excluded thepossibility that these differences in the ability to detect cyto-kines in serum could be due to kinetic (half-life), metabolism, orbinding protein modulation by HNSCC.

The evidence that HNSCCs express proinflammatory andproangiogenic cytokinesin vitro and in situ suggests thatHNSCCs are an important source for the elevated serum levelsof proinflammatory and proangiogenic cytokines. Presently, weobtained evidence that increased serum levels of IL-8 andVEGF are correlated with tumor volume (Fig. 4). Conversely,preliminary results of an analysis of the effect of surgery orchemo- and radiation therapy upon cytokine levels indicate thatserum cytokine levels decrease in posttreatment patients.4 How-ever, HNSCC cells may not be the only source of the elevatedserum cytokine levels in patients. In the immunohistochemicalanalysis of cytokines IL-6, IL-8, and GM-CSF, we also detectedareas of staining of cytokine IL-6 in fibroblasts and IL-8 andGM-CSF in tumor-infiltrating leukocytes within tumor stroma,which was negative for cytokeratin staining. This evidenceindicates that the serum levels of cytokines may also depend inpart upon individual host inflammatory responses within thetumor, which may explain some of the variation in serumcytokine concentration observed among individual patients withHNSCC. However, we did not observe a well-defined relation-ship between serum cytokine levels and the degree of tumordifferentiation, presence of inflammatory infiltrating cells, orlymph node metastasis in this pilot study (Table 6). Develop-ment of more sensitive and quantitative measurement of cyto-kines in tissue could complement serum cytokine analysis. Theconcentration of cytokines detected in serum in this study be-tween or within groups did not appear to be associated withgender, age, or alcohol or tobacco use. When we compared theserum cytokine levels between smokers and nonsmokers in thethree groups, no significant difference was found, although thiswas limited by the number of patient samples in this study.

It will be important to determine whether elevated levels ofproinflammatory and proangiogenic cytokines in serum of pa-tients with HNSCC at the time of diagnosis can allow monitor-ing of patients for therapeutic response and recurrence and helpwith defining prognosis. The present small pilot study offersencouragement, because we have observed that four of fivepatients who have developed progressive disease (patients 4, 6,7, and 9) showed significantly elevated serum levels of one ormore of IL-6, IL-8, or VEGF. This is in accordance with reportsthat elevated serum levels of proinflammatory and proangio-genic cytokines are correlated with advanced stage, metastaticdisease or large tumor burden in other types of cancer. In theseother cancers, increased serum IL-6 levels correlated well withthe disease status and prognosis, including patients with epithe-lial ovarian cancer (29, 30), metastatic renal cell carcinoma (31),colorectal cancer (32), esophageal squamous cell carcinoma(33), and cervical cancer (34). Elevated serum IL-8 was found inpatients with colorectal cancer (32), hepatocellular carcinoma(35), metastatic melanoma (36), and endometrial cancer (37).Increased serum VEGF level has been found in patients withinvasive breast carcinoma (38) and is associated with the pooroutcome in small-cell lung cancer (39). Our laboratory is con-tinuing to monitor serum cytokine levels in the patients enrolledin this clinical study and will investigate the relevance of serumcytokine levels as an indicator for prognosis, effectiveness oftherapy, and disease recurrence.

The detection of elevated expression of proinflammatoryand proangiogenic cytokines(IL-1a, IL-6, IL-8, GM-CSF, and

Fig. 3 Concentration of proinflammatory and proangiogenic cytokinesIL-6, IL-8, and VEGF in serum from patients with HNSCC, papilloma,and normal subjects by ELISA. The levels of cytokines, IL-6, IL-8, andVEGF in serum from age- and gender-matched patients with HNSCC,papilloma, and normal subjects were measured by ELISA. The meanconcentration of cytokine detected in replicate samples in serum ofpatients with cancer (F), papilloma (M), and normal subjects (‚) isshown.Solid bars,mean concentration (6SEM) of samples for eachgroup.p, a significant difference byt test between affected and normalsubjects atP , 0.05. The sample size and individual serum cytokinevalues of patients with HNSCC are shown in Table 6.

1376Cytokines in Head and Neck Squamous Cell Carcinoma

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

VEGF) in cultured squamous cell carcinoma linesin vitro and intumor specimens from cancer patients suggests that such cyto-kine expression may play a role in the increased pathogenicityof HNSCC by providing a growth advantage. We have observed

that rIL-1a, rIL-6, and rIL-8 can promote growth of UM-SCClines in vitro,5 which is consistent with our detection of proin-flammatory and proangiogenic cytokines in human SCC celllines and in patient serum. Moreover, we have evidence fromoral and cutaneous SCC murine models, that murine SCCsconstitutively produce a similar repertoire of proinflammatoryand proangiogenic cytokines, consisting of IL-1a, IL-6, KC, anIL-8/gro homologue, and GM-CSF (16, 17).6 Such high cyto-kine-producing lines exhibit a selective advantage in the rate oftumor growth and metastasisin vivo (15, 16).6

IL-1a, IL-6, IL-8, or VEGF may have effects on other non-neoplastic cells that could enhance survival and growth of SCCcells through a paracrine mechanism. The diverse local and sys-temic inflammatory responses observed in patients with SCC (1–14) are consistent with the broad proinflammatory and proangio-genic activity that have been reported for the cytokines detected inHNSCC. Thus, IL-1a, IL-6, and GM-CSF have been reported toplay important roles in the initiation of local inflammation, activa-tion of CD41 and B-lymphocyte responses, as well as systemicacute-phase responses. For example, chronic elevation of IL-6 canpromote immune unresponsiveness and induction of wasting, ca-chexia, and hypercalcemia, all of which are observed in patientswith SCC who have a poor prognosis (32, 33, 40). IL-8 plays animportant role in the stimulation of angiogenesis, proliferation, andchemotaxis of granulocytes and macrophages, which are prominentconstituents in the stroma of SCCs (2, 4, 5 32, 35, 36). VEGF isthought to be a major angiogenic growth factor in animal andhuman cancer (41–43) that provides sufficient vascularization fortumor growth expansion (42–44). Thus, there is considerable evi-dence that IL-1a, IL-6, IL-8, GM-CSF, and VEGF cytokinespromote tumorigenesisin vivo.

The constitutive expression of cytokines in SCC cells couldresult from activation of pathways involving IL-1a, other con-

5 S. H. Hong, F. G. Ondrey, I. M. Avis, Z. Chen, P. F. Cavanaugh, Jr.,C. Van Waes, and J. L. Mulshuhe. Effect of cyclooxygenase on theregulation of inflammatory cytokines and growth of human oropharyn-geal squamous cell carcinomas, manuscript in preparation.6 G. Thomas, unpublished observations.

Table 6 Serum cytokine levels, tumor load, and pathological characteristics of patients with head and neck squamous cell carcinoma

Patient

Cytokinesa

TNM

Tumor size (ml) Pathology

IL-6 IL-8 VEGF Primary Metastasis Differentiationb Inflammatoryc cells

1 2.7 29.5 779.7 T4N1M0 16.8 1.9 NA 112 28.5 24.4 900.6 T4N0M0 95.4 0.0 0 1113 7.5 27.6 430.1 T3N2M0 NA NA 11 14 7.2 130.9 1084.4 T4N0M0 92.4 0.0 11 15 53.5 40.9 699.1 T4N1M0 23.2 0.0 11 16 68.1 194.5 380.9 T3N3M1 NA NA 1 1117 129.6 0.0 146.8 T4N3M0 22.2 82.6 0 118 2.0 0.0 362.9 T4N1M0 39.7 0.0 11 1119 4.8 3.3 1073.3 T4N2cM0 44.3 43.7 11 111

10 0.5 23.6 374.2 T2N1M0 14.8 1.6 1 1111 1.9 63.6 234.9 T4N3M0 24.3 123.3 1 NA

a Cytokines were detected in patients’ serum and presented as pg/ml.b Differentiation was classified as well differentiated,111; moderately differentiated,11; poorly differentiated,1; and undifferentiated, 0.c The inflammatory infiltrate cells were evaluated as: intense,111; moderate,11; mild, 1; or absent. NA, not applicable.

Fig. 4 Correlation between serum VEGF and IL-8 levels and tumorload in patients with HNSCC. Linear regression analysis of cytokineserum levels (pg/ml) of VEGF (upper panel) and IL-8 (lower panel) asa function of primary tumor volume (cm3) in patients with HNSCC.Cytokine concentration was plotted against tumor volume and wasdetermined from computed tomography scan.Rs shown in each figurewere determined by linear regression analysis.

1377Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

stitutively expressed cytokines, signal transduction molecules,or activation of these genes by transcription factor(s). IL-1a-inducible acute-phase responses were elevated in the SCC pa-tients (Fig. 5). Furthermore, we have shown that IL-1a is astrong inducer of IL-8 and GM-CSF in HNSCC lines (18). Othercytokines, such as IL-1b, TNF, and platelet-derived growthfactor, have also been reported to modulate IL-6, IL-8, andVEGF gene expression (45). We did not detect IL-1b or TNFsecretion in any of the cell lines tested, but we did detectplatelet-derived growth factors in the culture supernatants fromUM-SCC.7 Alternatively, transcription factors nuclear factor-

kB, nuclear factor IL-6, and AP-1, which induce the genesencoding the proinflammatory and proangiogenic cytokines,could be activated in HNSCC (46, 47).

ACKNOWLEDGMENTSWe thank Drs. James Mulshine and Carol Langford for critical

reading the manuscript and providing helpful suggestions. We alsothank Dr. Gang Dong for assistance in preparing the photographs andJoanne Corum for assistance in manuscript preparation.

REFERENCES1. Wolf, F., Hudson, J., Peterson, K., Miller, H., and McClatchey, K.Lymphocyte subpopulations infiltrating squamous carcinomas of thehead and neck: correlations with extent of tumor and prognosis. Oto-laryngol. Head Neck Surg.,95: 142–152, 1986.2. Young, M. R. I., Wright, M. A., Lozano, Y., Prechel, M. M.,Benefield, J., Leonetti, J. P., Collins, S. L., and Petruzelli, G. J. In-creased recurrence and metastasis in patients whose primary head andneck squamous cell carcinomas secreted granulocyte-macrophage col-ony stimulating factor and contained CD341 natural suppressor cells.Int. J. Cancer,74: 69–74, 1997.3. Gleich, L. L., Biddinger, P. W., Duperier, F. D., and Gluckman, J. L.Tumor angiogenesis as a prognostic indicator in T2–T4 oral cavitysquamous cell carcinoma: a clinical-pathologic correlation. Head Neck,19: 276–280, 1997.4. Pak, A. S., Wright, M. A., Matthews, J. P., Collins, S. L., Petruzelli,G. J., and Young, M. R. I. Mechanisms of immune suppression in patientswith head and neck cancer: presence of CD341 cells which suppressimmune functions within cancers that secrete granulocyte-macrophage col-ony-stimulating factor. Clin. Cancer Res.,1: 95–103, 1995.5. Smith, D. R., Polverini, P. J., Kunkel, S. L., Orringer, M. B., Whyte,R. I., Burdick, M. D., Wilke, C. A., and Strieter, R. M. Inhibition ofinterleukin 8 attenuates angiogenesis in bronchogenic carcinoma. J.Exp. Med.,179: 1409–1415, 1994.6. Norton, J. A., Peacock, J. L., and Morrison, S. D. Cancer cachexia.Crit. Rev. Oncol. Hematol.,7: 289–327, 1987.7. Vlock, D. R., Arnold, B., Humpierres, J., Schwartz, D. R., Baker,S. R., Krause, C. J., Swanson, N., and Carey, T. E. Serial studies ofautologous antibody reactivity to squamous cell carcinoma of the headand neck. Cancer Immunol. Immunother.,34: 329–336, 1992.8. Veltri, R., Rodman, S., Maxim, P., Baseler, M., and Sprinkle, P.Immune complexes, serum proteins, cell-mediated immunity, and im-mune regulation in patients with squamous cell carcinoma of the headand neck. Cancer (Phila.),57: 2295–2308, 1986.9. Mandel, M., and Kiehn, C. The prognostic significance of delayedcutaneous reactivity in head and neck cancer patients. J. Plast. Reconstr.Surg.,53: 72–76, 1974.10. Burdick, J., Wells, S., and Herberman, R. Immunologic evaluationof patients with cancer by delayed hypersensitivity reactions. Surg.Gynecol. Obstet.,141: 779–794, 1975.11. Das, S., and Khanna, N.In vivo andin vitro observation of cellularimmune parameters in squamous cell carcinoma of the oral cavity andits correlation with tumor load and prognosis. Cancer Invest.,4: 207–216, 1986.12. Mickel, R., Kessler, D., Taylor, J., and Lichtenstein, A. Natural killercell cytotoxicity in the peripheral blood, cervical lymph nodes, and tumor ofhead and neck cancer patients. Cancer Res.,48: 5017–5022, 1998.13. Yoneda, T., Aufdemorte, T., Nishimura, R., Nishikawa, N., Sakuda,M., Alsina, M., Chavez, J., and Mundy, G. Occurrence of hypercalcemiaand leukocytosis with cachexia in a human squamous cell carcinoma of themaxilla in athymic nude mice. A novel experimental model of threeconcomitant paraneoplastic syndromes. J. Clin. Oncol.,9: 468–477, 1991.14. Saito, K., Kuratomi, Y., and Yamamoto, K., Saito, T., Kuzuya, T.,Yoshida, S., Moriyama, S. I., and Takahashi, A. Primary squamous cellcarcinoma of the thyroid associated with marked leukocytosis andhypercalcemia. Cancer (Phila.),48: 2080–2083, 1981.7 Unpublished observations.

Fig. 5 Determination of acute-phase responses in serum from patientswith head and neck cancer, papilloma, and normal subjects. The meanconcentration of proteins or erythrocyte sedimentation rate were de-tected in replicate samples in patients with cancer (F), papilloma (M),and normal subjects (‚) is shown.Solid bars,mean concentration ofsamples for each group (6SEM). p, statistical difference byt testbetween affected and normal subjects atP , 0.05.

1378Cytokines in Head and Neck Squamous Cell Carcinoma

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

15. Chen, Z., Smith, C. W., Kiel, D., and Van Waes, C. Metastaticvariants derived followingin vivo tumor progression of anin vitrotransformed squamous cell carcinoma line acquire a differential growthadvantage requiring tumor-host interaction. Clin. Exp. Metastasis,15:527–537, 1997.16. Dong, G., Loukinova, E., Smith, C. W., Chen, Z., and Van Waes, C.Genes differentially expressed with malignant transformation and met-astatic tumor progress on of murine squamous cell carcinoma. J. CellBiochem. Suppl.,28/29: 90–100, 1997.17. Smith, C. W., Chen, Z., Dong, G., Loukinova, E., and Van Waes, C.The host environment promotes the development of primary and met-astatic squamous cell carcinomas that constitutively express proinflam-matory cytokines IL-1a, IL-6, GM-CSF and KC. Clin. Exp. Metastasis,16: 655–664, 1998.18. Chen, Z., Colon, I., Ortiz, N., Callister, M., Dong, G., Pegram,M. Y., Arosarena, O., Strome, S., Nicholson, J. C., and Van Waes, C.Effects of interleukin-1a, interleukin-1 receptor antagonist, and neutral-izing antibody on proinflammatory cytokine expression by human squa-mous cell carcinoma lines. Cancer Res.,58: 3668–3676, 1998.19. Mann, E., Spiro, J., Chen, L., and Kreutzer, D. Cytokine expressionby head and neck squamous cell carcinoma. Am. J. Surg.,164:567–573,1992.20. Cohen, R., Contrino, J., Spiro, J., Mann, E., Chen, L., and Kreutzer,D. Interleukin-8 expression by head and neck squamous cell carcinoma.Arch. Otolaryngol. Head Neck Surg.,121: 202–209, 1995.21. Yamamura, M., Modlin, R., Ohmen, J., and Moy, R. Local expres-sion of antiinflammatory cytokines in cancer. J. Clin. Invest.,91:1005–1010, 1993.22. Woods, K. V., El-Nagger, A., Clayman, G. L., and Grimm, E. A.Variable expression of cytokines in human head and neck squamous cellcarcinoma cell lines and consistent expression in surgical specimens.Cancer Res.,58: 3132–3141, 1998.23. Krause, C., Carey, T., Ott, R., Hurbis, C., McClatchey, K., andRegezi, J. Human squamous cell carcinoma. Establishment and charac-terization of new permanent cell lines. Arch. Otolaryngol.,107: 703–710, 1981.24. Bradford, C. R., Kimmel, K. A., Van Dyke, D. L., Worsham, M. J.,Tilley, B. J., Burk, D., del Rosario, F., Lutz, S., Tooley, R., Hayashida,D. J. S., and Carey, T. E. 11p deletions and breakpoints in squamous cellcarcinoma: association with altered reactivity with the UM-E7 antibody.Genes Chromosomes Cancer,3: 272–282, 1991.25. Van Dyke, D. L., Worsham, M. J., Benninger, M. S., Krause, C. J.,Baker, S. R., Wolf, G. T., Drumheller, T., Tilley, B. C., and Carey, T. E.Genes Chromosomes Cancer,9: 192–206, 1994.26. Van Waes, C., Mi Surh, D., Cheng, Z., Kirby, M., Rhim, J. S.,Brager, R., Sessions, R. B., Poore, J., Wolf, G. T., and Carey, T. E.Increase in suprabasilar integrin adhesion molecule expression in humanepidermal neoplasms accompanies increased proliferation occurringwith immortalization and tumor progression. Cancer Res.,55: 5434–5444, 1995.27. Ågren, K., Andersson, V., Litton, M., Funa, K., Nordlander, B., andAndersson J. Production of immunoregulatory cytokines is located tothe extrafollicular area of human tonsils. Acta Otolaryngol. (Stockh),116: 477–485, 1996.28. Van Waes, C., Chen, Z., Callister, M., Colon, I., Ortiz, N., Smith,C. W., Thomas, G. R., and Dong, G. Cytokines in the immune patho-genesis and therapy of head and neck cancer.In: J. E. Veldman, D.Passali, and D. J. Lim (eds.), Immuno- and Molecular Biology inORL–State of the Art, in press. The Hague, the Netherlands: KuglerPublications, 1999.29. Berek, J. S., Chung, C., Kaldi, K., Watson, J. M., Knox, R. M., andMartinez-Maza, O. Serum interleukin-6 levels correlate with diseasestatus in patients with epithelial ovarian cancer. Am. J. Obstet. Gynecol.,164: 1038–1042, 1991.30. Watson, J. M., Sensintaffar, J. L., Berek, J. S., and Martinez-Maza, O.Constitutive production of interleukin 6 by ovarian cancer cell lines and byprimary ovarian tumor cultures. Cancer Res.,50: 6959–6965, 1990.

31. Blay, J. Y., Rossi, J. F., Wijdenes, J., Menetrier-Caux, C., Scheman,S., Negrier, S., Philip, T., and Favrot, M. Role of interleukin-6 in theparaneoplastic inflammatory syndrome associated with renal-cell carci-noma. Int. J. Cancer,72: 424–430, 1997.

32. Ueda, T., Shimada, E., and Urakawa, T. Serum levels of cytokinesin patients with colorectal cancer: possible involvement of interleukin-6and interleukin-8 in hematogenous metastasis. J. Gastroenterol.,29:423–429, 1994.

33. Oka, M., Yamamoto, K., Takahashi, M., Hakozaki, M., Abe, T.,Iizka, N., Hazama, S., Hirazawa, K., Hayashi, H., Tangoku, A., Hirose,K., Ishihara, T., and Suzuki, T. Relationship between serum levels ofinterleukin 6, various disease parameters, and malnutrition in patientswith esophageal squamous cell carcinoma. Cancer Res.,56: 2776–2780,1996.

34. Chopra, V., Dinh, T. V., and Hannigan, E. V. Circulating serumlevels of cytokines and angiogenic factors in patients with cervicalcancer. Cancer Invest.,16: 152–159, 1998.

35. Sakamoto, K., Masuda, T., Mita, S., Ishiko, T., Nakashima, Y.,Arakawa, H., Egami, H., Harada, S., Matsushima, K., and Ogawa, M.Interleukin-8 is constitutively and commonly produced by various hu-man carcinoma cell lines. Int. J. Clin. Lab. Res.,22: 216–219, 1992.

36. Scheibenbogen, C., Mohler, T., Haefele, J., Hunstein, W., andKeilholz, U. Serum interleukin-8 (IL-8) is elevated in patients withmetastatic melanoma and correlates with tumour load. Melanoma Res.,5: 179–181, 1995.

37. Chopra, V., Dinh, T. V., and Hannigan, E. V. Serum levels ofinterleukins, growth factors and angiogenin in patients with endometrialcancer. J. Cancer Res. Clin. Oncol.,123: 167–172, 1997.

38. Obermair, A., Kucera, E., Mayerhofer, K., Speiser, P., Seifert, M.,Czerwenka, K., Kaider, A., Leodolter, S., Kainz, C., and Zeillinger, R.Vascular endothelial growth factor (VEGF) in human breast cancer:correlation with disease-free survival. Int. J. Cancer,74: 455–458, 1997.

39. Salven, P., Ruotsalainen, T., Mattson, K., and Joensuu, H. Highpre-treatment serum level of vascular endothelial growth factor (VEGF)is associated with poor outcome in small-cell lung cancer. Int. J. Cancer,79: 144–146, 1998.

40. Lissoni, P., Rovelli, F., Tisi, E., Brivio, F., Ardizzoia, A., Barni, S.,Tancini, G., Saudelli, M., Cesana, E., and Vigano, M. G. Relationshipbetween macrophage and T helper-2 lymphocyte functions in humanneoplasms: neopterin, interleukin-10 and interleukin-6 blood levels inearly or advanced solid tumors. J. Biol. Regul. Homeostatic Agents,9:146–149, 1995.

41. Dvorak, H. F., Sioussat, T. M., Brown, L. F., Berse, B., Nagy, J. A.,Sotrel, A., Manseau, E. J., Van De Water, L., and Senger, D. R.Distribution of vascular permeability factor (vascular endothelial growthfactor) in tumors: concentration in tumor blood vessels. J. Exp. Med.,174: 1275–1278, 1991.

42. Kolch, W., Martiny-Baron, G., Kieser, A., and Marme, D. Regula-tion of the expression of the VEGF/VPS and its receptors: role in tumorangiogenesis. Breast Cancer Res. Treat.,36: 139–155, 1995.

43. Weidner, N., and Folkman, J. Tumoral vascularity as a prognosticfactor in cancer. Important Adv. Oncol.,1996: 167–190, 1996.

44. Folkman, J. Seminars in Medicine of the Beth Israel Hospital,Boston. Clinical applications of research on angiogenesis. N. Engl.J. Med.,333: 1757–1763, 1995.

45. Paul, W. Fundamental Immunology, Ed. 3. pp 801–836. New York:Raven Press, Ltd., 1993.

46. Duffey, D., Ondrey, F. G., Chen, Z., Dong, G., and Van Waes, C.Activation of immediate-early transcription factor NF-kB in humanhead neck squamous carcinoma cell lines that are resistant to TNFa.Proc. Am. Assoc. Cancer Res.,39: 451, 1998.

47. Ondrey, F. G., Dong, G., and Van Waes, C. Constitutive expressionof AP-1 and NF IL-6 in squamous carcinoma of the head and neck. Proc.Am. Assoc. Cancer Res.,39: 453, 1998.

1379Clinical Cancer Research

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

1999;5:1369-1379. Clin Cancer Res   Zhong Chen, Pramit S. Malhotra, Giovana R. Thomas, et al.   in Patients with Head and Neck CancerExpression of Proinflammatory and Proangiogenic Cytokines

  Updated version

  http://clincancerres.aacrjournals.org/content/5/6/1369

Access the most recent version of this article at:

   

   

  Cited articles

  http://clincancerres.aacrjournals.org/content/5/6/1369.full#ref-list-1

This article cites 42 articles, 10 of which you can access for free at:

  Citing articles

  http://clincancerres.aacrjournals.org/content/5/6/1369.full#related-urls

This article has been cited by 61 HighWire-hosted articles. Access the articles at:

   

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

  Subscriptions

Reprints and

  .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://clincancerres.aacrjournals.org/content/5/6/1369To request permission to re-use all or part of this article, use this link

Research. on August 10, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from