Macrophages From Cancer Patients: Analysis of TRAIL, TRAIL Receptors, and Colon Tumor Cell Apoptosis

Macrophages From Cancer Patients: Analysis of TRAIL,TRAIL Receptors, and Colon Tumor Cell Apoptosis

Jean-Philippe Herbeuval, Claude Lambert, Odile Sabido, Michèle Cottier,Pierre Fournel, Michel Dy, Christian Genin

Background: Tumor-infiltrating macrophages secrete cyto-kines, including Fas ligand, tumor necrosis factor-� (TNF-�),and TNF-related apoptosis-inducing ligand (TRAIL).TRAIL induces apoptosis in tumor cells but not in normalcells; however, regulation of TRAIL and its receptors incancer patients is relatively uncharacterized. We investi-gated whether macrophages from cancer patients produceTRAIL and whether apoptosis in cultured colon adenocar-cinoma cells involves TRAIL and its receptors. Methods:Macrophages isolated from pleural effusions of nine cancerpatients and five control patients with congestive heart fail-ure (whose effusions contained no tumor cells) were cul-tured. Levels of TRAIL, TNF-�, interferon �, and Fas ligandin conditioned medium were measured by enzyme-linkedimmunosorbent assays. Apoptosis of human colon adenocar-cinoma cell lines, including Colo 205, was determined by theAnnexin V method and terminal deoxynucleotidyltransfer-ase-mediated deoxyuridine 5�-triphosphate nick-end label-ing (TUNEL). Cell-surface TRAIL receptors were measuredby flow cytometry. Results: Conditioned culture mediumfrom macrophages isolated from pleural effusions contain-ing 1%–5% tumor cells (CM-A) contained TRAIL at 980–1300 pg/mL, whereas that from macrophages from pleuraleffusions containing more than 50% tumor cells or contain-ing no tumor cells (CM-B) contained TRAIL at 0–50 pg/mL.When cultured with medium containing 50% CM-A, 40%(95% confidence interval [CI] = 30% to 50%) of Colo 205cells underwent apoptosis; when cultured with 50% CM-B,8% (95% CI = 3% to 13%) underwent apoptosis. When Colo205 cells were cultured with 50% CM-A, cell-surface expres-sion of TRAIL death receptors DR5 and DR4 increased 13-fold and sixfold, respectively, compared with that of un-treated Colo 205 cells. Recombinant TRAIL induced 90%(95% CI = 85% to 95%) of Colo 205 cells to undergo apo-ptosis and acted synergistically with TNF-� to induce apo-ptosis. Conclusion: Macrophages from cancer patients ap-pear to be activated by tumor cells to produce TRAIL and toincrease the expression of TRAIL death receptors DR4 andDR5 on tumor cells. [J Natl Cancer Inst 2003;95:611–21]

Tumor infiltration by macrophages directly influences cancercell growth in vitro and in vivo (1–3). Tumor-infiltrating mac-rophages appear to be able to distinguish between malignant andnormal cells and can attack neoplastic cells in patients throughcontact-dependent and cytotoxin-mediated mechanisms (4); thepresence of tumor-infiltrating macrophages is associated withgood prognosis for patients with colorectal adenocarcinoma (5).However, the story is complicated because such macrophagescan secrete a variety of cytokines and thus can promote or inhibittumor growth, depending on which cytokines are secreted (6).When cytokines of the tumor necrosis factor-� (TNF-�) family

are secreted, as is frequently observed in macrophages that havebeen exposed to human lung cancer cells (7) or colorectal cancercells (8–10), apoptosis is induced.

TNF-� and Fas ligand (FasL), members of the TNF-� family,induce apoptosis by binding to their corresponding cell-surfacereceptors. The receptor for TNF-� is TNF-R1, and the receptorfor FasL is Fas. These receptors contain intracytoplasmic do-mains that, after ligand binding, induce apoptosis in tumor cells(11). TNF-related apoptosis-inducing ligand (TRAIL), anothermember of the TNF-� family (12), does not induce apoptosis innormal cells (13) but does induce apoptosis in several humantumor cell lines (14). Several melanoma cell lines (e.g., WM9,WM35, WM98–1, WM164, WM793, WM1205-Ln, WM1791-C,and WM3211) are resistant to apoptosis induced by FasL orTNF-� but are sensitive to TRAIL-mediated apoptosis (15,16).The following five TRAIL receptors have been described:TRAIL-R1 or death receptor 4 (DR4), TRAIL-R2 or DR5,TRAIL-R3 or decoy receptor 1 (DcR1), TRAIL-R4 or DcR2,and osteoprotegerin (OPG) (17). DR4 and DR5 contain cyto-plasmic death domains and induce apoptosis after ligand binding(18,19); DcR1, DcR2, and OPG lack such death domains and donot induce apoptosis (20,21). TRAIL is secreted by immunecells, including lymphocytes, natural killer cells (22), mono-cytes, dendritic cells (23), and macrophages (24). The expres-sion of decoy receptors for TRAIL in normal tissues, but not inmany tumor cell lines, may account for the resistance of normaltissues and for the broad sensitivity of tumor cell lines toTRAIL-induced apoptosis (25). However, the molecular mecha-nisms involved in the production and regulation of TRAIL andits receptors in vivo remain to be elucidated.

An inverse association between the number of macrophagesin pleural effusions from cancer patients and the extent of ma-lignant disease has been reported (26–28). Our objective in thisstudy was to investigate the cytotoxic effects of macrophages onhuman tumor cells. We sampled pleural effusions from ninepatients with cancer and from five control patients with conges-tive heart failure to obtain large numbers of macrophages thathad been in contact with large or small numbers of tumor cellsin pleural effusions from the cancer patients or no tumor cells inpleural effusions from patients with congestive heart failure. Wethen used human Colo 205, Colo 320, and Caco-2 colon adeno-

Affiliations of authors: J.-P. Herbeuval, C. Lambert, C. Genin (Groupe Im-munité des Muqueuses et Agents Pathogènes [GIMAP]), O. Sabido (Departmentof Flow Cytometry), M. Cottier (Department of Cytology), P. Fournel (Depart-ment of Pneumology), Jean Monnet University, Saint-Etienne, France; M. Dy,Unite Mixte de Recherche, Hopital Necker, Paris, France.

Correspondence to: Christian Genin, M.D., Ph.D., GIMAP, Faculté de Méde-cine Jacques Lisfranc, 15 rue A. Paré, 42023 Saint-Etienne Cédex 2, France(e-mail: [email protected]).

See “Notes” following “References.”Journal of the National Cancer Institute, Vol. 95, No. 8, © Oxford University

Press 2003, all rights reserved.

Journal of the National Cancer Institute, Vol. 95, No. 8, April 16, 2003 ARTICLES 611

by guest on Novem

ber 16, 2015http://jnci.oxfordjournals.org/

Dow

nloaded from

http://jnci.oxfordjournals.org/

carcinoma cells as target cell lines to develop an in vitro modelto test whether macrophages that had been activated by tumorcells in pleural effusions produced TRAIL, whether macro-phages induced increased expression of TRAIL death receptorson target tumor cells, and whether TRAIL induced target colonadenocarcinoma cell lines to undergo apoptosis.

MATERIALS AND METHODS

Materials

Human recombinant TNF-� was purchased from PreproTech,Inc. (Rocky Hill, NJ). Human recombinant TRAIL, interferon �(IFN �), and FasL were purchased from R&D Systems (Abing-don, U.K.). We used recombinant DR5 (R&D Systems) at 1, 5,10, and 20 �g/mL to selectively block TRAIL-mediated apo-ptosis. Membrane TRAIL was detected with a mouse anti-humanmonoclonal antibody from Diaclone Research (Besancon, France).Goat anti-human antibodies against TRAIL receptors DR4,DR5, DcR1, and DcR2 (all polyclonal antibodies) were pur-chased from R&D Systems. The tetrapeptide caspase inhibitorbenzyloxycarbonyl-Val-Ala-Asp(OMe)-CH2-fluoromethyl ke-tone (zVAD-fmk) was purchased from France Biochem(Meudon, France). Stock solutions of the inhibitor were pre-pared in dimethyl sulfoxide (Life Technologies, Gaithersburg,MD) and stored at 4 °C.

Tumor and Control Cell Collection

Pleural fluids containing tumor cells and leukocytes werecollected by drainage thoracocentesis from nine patients witheffusions related to various cancers (three with pleura mesothe-lioma, four with lung adenocarcinoma, one with breast adeno-carcinoma, and one with colon adenocarcinoma). All cancerpatients had stage IV metastatic disease. As a tumor-unrelatedcontrol, pleural fluids were also collected from five patientswith congestive heart failure, as previously described (29). Thisin situ control was considered appropriate because the macro-phages had not been in contact with tumor cells. Althoughin vitro adherence of macrophages to a culture dish might haveinduced the macrophages to produce TRAIL, we did not detectany TRAIL after a 48-hour culture of macrophages from controldonors, whereas we did detect the production of TRAIL bymacrophages from cancer patients after a 48-hour culture. If thepleural effusion was bloody, red cells (which represented lessthan 1% of total cells) were removed by Ficoll centrifugation (atroom temperature for 20 minutes at 800g) and, essentially, noneof the macrophages in the effusions came from the small amountof contaminating blood. After collection, pleural fluids wereplaced in sterile, screw-cap 2-L ethylene containers (PlastiquesGosselin, Hazebrouck, France) with 2000 IU of sodium heparin(Laboratoires LEO S.A., Saint Quentin en Yvelines, France) atroom temperature for 1–2 hours, for transportation to the labo-ratory. The thoracocentesis procedure was approved by the In-stitutional Review Board at the University Hospital Center,Saint-Etienne, France, and signed informed consent was ob-tained from all the cancer patients and patients with congestiveheart failure.

Cytologic Evaluation of Cells Present in the Pleural Fluid

After the pleural fluids were centrifuged at 400g at roomtemperature, the cells were transferred to slides (Merck Eurolab,Strasbourg, France) and spread thinly. At least four smears were

prepared from each sample. Air-dried smears were processedwith May–Grünwald–Giemsa staining, and wet-fixed smearswere processed with Papanicolaou staining.

The cytologic features that we used to distinguish malignantcells were their relatively large size, high nuclear-to-cytoplasmicvolume ratio, nuclear hyperchromasia with irregularities of thechromatin pattern, large nucleoli, and abnormal mitotic figures.Tumor cells were classified as adenocarcinoma, poorly differ-entiated carcinoma, oat-cell carcinoma, squamous cell carci-noma, and malignant mesothelioma. Macrophages showed anirregular nuclear shape and a voluminous cytoplasm with nu-merous vacuoles.

Preparation of Conditioned Medium

Pleural fluids were centrifuged at 400g for 10 minutes at 4 °C,and cell pellets were resuspended in RPMI-1640 medium con-taining 20 mM HEPES buffer (Life Technologies, Paisley, Scot-land), 5% type AB human serum (Western States Plasma–SeraCare, Inc., Oceanside, CA), 2 mM L-glutamine (Eurobio,Les Ullis, France), and antibiotics (penicillin G at 100 U/mL,streptomycin at 100 �g/mL, and amphotericin B at 0.25 �g/mL;Life Technologies). Cells were plated at 2 × 106 cells per mL,and macrophage-like cells were allowed to adhere to 75-cm2

tissue culture flasks (Falcon, Oxnard, CA) at 37 °C in an incu-bator in a humidified atmosphere of 5% CO2/95% air for 2hours. Nonadherent cells were removed by two washings withDulbecco’s phosphate-buffered saline (PBS) containing 0.02%EDTA (Eurobio) at room temperature. Macrophages accountedfor 95% of the remaining adherent cells, as determined by lightmicroscopy counting of May–Grünwald–Giemsa stained cells(30). After the 2-hour incubation, macrophages were removedfrom culture dishes by scraping and pipetting and plated in 75-cm2 culture dishes (Merck Eurolab) at 500 000 cells per milliliterof culture medium.

Adherent cells were cultured for 48 hours in the same me-dium at 37 °C in an incubator in a humidified atmosphere of 5%CO2/95% air. This time was chosen because a kinetic analysisindicated that all cytokines tested were maximally produced be-tween 24 and 48 hours of culture, and TRAIL was maximallyproduced by 48 hours. Conditioned culture medium was centri-fuged at 400g for 10 minutes at 4 °C, and the supernatant wasstored in aliquots at –80 °C until use. After thawing, supernatantswere filtered through 0.22-�m (pore size) filters. Culture super-natants from macrophages isolated from pleural effusions ofpatients with small numbers of tumor cells (1%–5%) were des-ignated as conditioned medium A (CM-A). Culture supernatantsfrom macrophages from pleural effusion of patients with largenumbers of tumor cells (>50%) or from patients with congestiveheart failure with no tumor cells were designated as CM-B. Eachof the five CM-A and nine CM-B samples were prepared andstudied individually. It should be noted that no patient pleuraleffusions contained tumor cells in the range of 6%–50%.

Tumor Cell Culture Conditions

The human colon cell lines Colo 205, Colo 320, and Caco-2were purchased from the European Collection of Cell Cultures(Salisbury, U.K.). Tumor cells were cultured under the sameconditions used for macrophages, except that RPMI-1640 me-dium was supplemented with 10% heat-inactivated fetal calfserum (Life Technologies). The appropriate cell density for plat-

612 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 8, April 16, 2003

by guest on Novem


Dow

nloaded from


ing was first determined under basal conditions to avoid contactinhibition of cell growth or depletion of nutrients from the me-dium (data not shown). Different numbers of Colo 205, Colo320, and Caco-2 cells were plated in 24-well plates (Falcon) andstained with trypan blue, and viable cells were counted with ahemocytometer. Exponential growth was observed at a platingdensity of 50 000 cells per well. Cells were cultured for anadditional 4 days with 50% conditioned medium (CM-A or CM-B). A concentration–response experiment, in which 0.1%–100%CM-A and CM-B were added to culture medium, determinedthat 50% CM-A or 50% CM-B was the optimal concentrationfor the production of TRAIL and the induction of apoptosis.

Enzyme-Linked Immunosorbent Assay (ELISA) forCytokines in Conditioned Medium

Levels of FasL, TNF-�, IFN �, and TRAIL were measured ineach conditioned medium and cell culture supernatant with com-mercial enzyme-linked immunosorbent assay (ELISA) kits, ac-cording to the manufacturer’s instructions: FasL and TNF-�,Quantikine from R&D Systems; trimeric TRAIL, Diaclone; andIFN �, Biosource International (Camarillo, CA). Positive con-trols were supplied in each kit. The optical density was measuredat 492 nm on a Titertek Multiskan reader (Flow Labs, Irvine,U.K.). Detection ranges of the various ELISA kits were as fol-lows: 15–1000 pg/mL for TNF-�; 12–500 pg/mL for IFN �;0.1–500 pg/mL for TRAIL; and 4–250 pg/mL for FasL. Thecytokine content of each of the 14 conditioned media was testedthree times. Levels of cytokines were expressed as picogramsper milliliter per 500 000 macrophages. Cytokines were not de-tected in the medium used to culture the macrophages.

Measurement of TRAIL mRNA by ReverseTranscription–Polymerase Chain Reaction (RT–PCR)

Macrophages isolated from cancer patient pleural effusionswith small numbers of tumor cells and from pleural effusionsfrom patients with congestive heart failure were tested directlyfor TRAIL mRNA, without being cultured. mRNA was ex-tracted from 500 000 cultured macrophages with TRIzol reagent(Life Technologies), according to the manufacturer’s instruc-tions. RT was performed with a SuperScript preamplificationsystem for first-strand cDNA synthesis (Life Technologies) in aPerkin-Elmer thermocycler. To quantify total cDNA synthe-sized, we added deoxycytosine 5�-[�-33P]triphosphate (1 �L at10 mCi/mL; DuPont, Boston, MA). Total RNA was incubated50 minutes at 42 °C in the presence of the SuperScript enzyme,and the reaction was inactivated by heating at 70 °C for 15minutes. RT efficiency was estimated by the radioactivity asso-ciated with synthesized cDNA, and the number of copies wasused to standardize subsequent PCRs.

The production of TRAIL mRNA was semiquantitativelymeasured with RT–PCR. The following oligonucleotide primersfor human TRAIL (874 base pairs [bp]) and �2-microglobulin(280 bp) mRNAs were synthesized by Life Technologies:TRAIL 5� primer � TATTCCATCTATCAAGGGGGA and3� primer � AAACCAAGTCTCGCTCTGT; �2-microglobulin5� primer � CCAGCAGAGAATGGAAAGTC and 3� primer� GATGCTGCTTACATGTCTCG. The housekeeping gene�2-microglobulin was also amplified in parallel as a referencefor the quantification of TRAIL transcripts. Each cycle in the30-cycle amplification was as follows: 1 minute at 95 °C fordenaturation, 1 minute at 55 °C for annealing, and 1 minute at

72 °C for primer extension. All PCR-amplified products wereseparated by electrophoresis on 1.5% agarose gels containingethidium bromide (0.5 �g/mL) and visualized by UV irradiation.

Detection of Membrane-Bound TRAIL on Macrophagesby Flow Cytometry

Macrophages were cultured 24 hours or 48 hours, and mem-brane TRAIL expression was determined by incubating the mac-rophages for 15 minutes at room temperature with fluoresceinisothiocyanate (FITC)-conjugated mouse anti-human TRAILmonoclonal antibody or with relevant control isotype-matchedantibodies (each at 5 �g/mL; Diaclone) in PBS containing 3%bovine serum albumin. After two washes in ice-cold PBS, thecells were analyzed by flow cytometry (BD, San Jose, CA).

Apoptosis Studies

Morphologic examination of tumor cells treated with condi-tioned medium or recombinant cytokines showed the loss ofmembrane asymmetry and condensation of cytoplasm andnucleus, which are characteristics of apoptotic cells. Cells weretested further for apoptosis by using the Annexin V test and theterminal deoxynucleotidyltransferase-mediated deoxyuridine5�-triphosphate (dUTP) nick-end labeling (TUNEL) assay.

In early apoptosis, phosphatidylserine is exposed to the outerleaflet of the plasma membrane. Annexin V preferentially bindsto the exposed phosphatidylserine and thus detects early eventsin apoptosis (25). We cultured Colo 205, Colo 320, and Caco-2cells with 50% CM-A or CM-B or various recombinant cyto-kines for 24 hours. As a control, cells were grown in freshRPMI-1640 medium supplemented with 10% fetal calf serum.To determine the combination of cytokines that induces the op-timal level of apoptosis, we tested several combinations ofTRAIL, IFN �, TNF-�, and FasL (each at 0.1, 10, 25, 50, and100 ng/mL). Cells were then centrifuged, washed with the bufferfrom the Annexin kit (Coger, Paris, France), and incubated for20 minutes with medium only (negative control) or with FITC-conjugated Annexin V (Annexin V-FITC) (Coger) in the dark atroom temperature. To eliminate necrotic cells and membranefragments that might be Annexin V-FITC-positive, we addedpropidium iodide and selected propidium iodide-negative cellsas follows. After two washes, propidium iodide (100 �g/mL)was added directly to the cell pellet, and tubes were placedon ice. Labeled cells were quantitated by flow cytometry withFACStar Plus cell sorter (BD). Apoptotic cells were visualizedon a cytogram showing the number of propidium iodide-nega-tive cells versus the number of Annexin V-FITC-positive cells.

To test for later stages of apoptosis in Colo 205, Colo 320,and Caco-2 cells, we used flow cytometry to measure DNAfragmentation detected with the TUNEL assay (31). TheTUNEL assay labels DNA strand breaks in apoptotic cells; thesebreaks are characteristic of later stages of apoptosis. Propidiumiodide was used as a probe for total DNA content. For theTUNEL assay, 106 Colo 205 cells were washed twice in PBS andsuspended in a buffer containing terminal deoxynucleotidyltrans-ferase and biotin-conjugated dUTP (biotin-dUTP) (ApopDETEK,Enzo Dako, Trappes, France), according to the manufacturer’sinstructions. The cells were incubated at 37 °C for 60 minutesand washed twice in PBS. Colo 205 cells that incorporated bio-tin-dUTP in the terminal deoxynucleotidyltransferase reactionwere suspended in 200 �L of staining solution containing Avi-din-FITC (6 �g/mL; Immunotech-Coulter, Marseille, France),


by guest on Novem


Dow

nloaded from


5% nonfat dry milk (Boehringer Mannheim, Meylan, France),RNase (20 �g/mL; Boehringer Mannheim), and 0.1% TritonX-100 (Sigma-Aldrich, L’Isle d’Abeau Chesnes, France) instandard saline citrate buffer (Sigma-Aldrich). Cells were incu-bated for 30 minutes at 37 °C in the dark, washed twice withPBS, resuspended in 200 �L of propidium iodide (100 �g/mL inPBS), and analyzed by flow cytometry. To verify the activity ofterminal deoxynucleotidyltransferase as a positive control, Colo205 cells were treated with DNase (1 �g/mL; Boehringer Mann-heim), which fragments the DNA, and then subjected to theTUNEL assay. Apoptotic cells were visualized on a cytogramfor a dual parameter analysis plotting propidium iodide intensityversus Avidin-FITC intensity. Information on at least 10 000events was acquired for each TUNEL assay.

To determine whether the caspase cascade pathway is in-volved in Colo 205 cell death, we used the caspase inhibitorzVAD-fmk to specifically block caspase-mediated apoptosis.Colo 205 cells were incubated with 50 �M zVAD-fmk for 4hours at 37 °C and then cultured for 24 hours in 50% CM-A orwith recombinant TRAIL or TNF-� (each at 50 ng/mL). Cellswere then analyzed for apoptosis by the Annexin V method, asdescribed above.

Recombinant DR5 binds TRAIL and prevents TRAIL frombinding to its cell-surface receptors, resulting in inhibitionof TRAIL-associated pathways. To confirm that TRAIL wasinvolved in Colo 205 cell death, cells were cultured for 24 hourswith 50% CM-A and recombinant DR5 at 1, 5, 10, and20 �g/mL or with recombinant TRAIL (at 50 ng/mL) and DR5at 1, 5, 10, and 20 �g/mL, and apoptosis was assessed by theAnnexin V method, as described above.

Detection of TRAIL Receptors on Colo 205 CellMembranes by Flow Cytometry

Colo 205 cells were cultured alone or with CM-A, and thesurface expression of TRAIL receptors DR4, DR5, DcR1, andDcR2 was examined by incubating Colo 205 cells with goatanti-human DR4, DR5, DcR1, DcR2, or relevant control iso-type-matched antibodies (each at 5 �g/mL; R&D Systems)in PBS containing 3% bovine serum albumin for 45 minuteson ice. After three washes in PBS, the cells were incubated for45 minutes on ice with an FITC-conjugated mouse anti-goatantibody diluted 1 : 200 in PBS containing 3% bovine serumalbumin (R&D Systems). Cells were washed two times in ice-cold PBS and analyzed by flow cytometry (BD).

Statistical Analysis

All experiments were repeated a minimum of three times. Allcomparisons of data were made with a two-sided Student’s t test(data meet the requirements for a Student’s t test).

To test the effects of combined treatment (TNF-� plusTRAIL or TNF-� plus IFN �) for synergistic effects, combina-tions of cytokines were evaluated, and dose–response curves forsingle cytokines and combinations of cytokines were compared.Characteristics of the effects of the combined treatment werethen analyzed by the isobole method (32) for combinations ofmolecules A and B: Ac/Ae + Bc/Be � D, where Ac and Bc are theconcentrations of cytokines used in combination to induce aneffect, and Ae and Be are the concentrations of cytokines able toproduce the same magnitude of effect if used individually. If thecombination index D is less than 1, the combination of cytokinesact synergistically; if D is greater than or equal to 1, the respec-tive effects were antagonistic or additive, respectively. Eachexperiment was performed in triplicate, and the statistical sig-nificance (P) of the combination indices compared with theadditive combination index (D � 1) was determined by a two-sided Student’s t test.

RESULTS

Production of TRAIL, TNF-�, and IFN � by MacrophagesFrom Cancer Patients and Patients With CongestiveHeart Failure

Macrophages were isolated from cancer patient pleural effu-sions with small numbers of tumor cells (1%–5% of total cells),cancer patient pleural effusions with large numbers of tumorcells (>50% of total cells), and control pleural effusions frompatients with congestive heart failure (control for no tumorcells). These macrophages were cultured for 2 days, and thelevels of TRAIL and TNF-� secreted into culture supernatantswere determined. High levels of TRAIL (980–1300 pg/mL)were detected in culture supernatants from macrophages ex-posed in vivo to pleural effusions with 1%–5% malignant cells(CM-A; Table 1). Low levels of TRAIL (0–50 pg/mL) weredetected in culture supernatants from macrophages exposedin vivo to pleural effusions with greater than 50% malignant cellsor from macrophages exposed in vivo to pleural effusions frompatients with congestive heart failure containing no tumor cells(CM-B; Table 1). In contrast, no statistically significant differ-ence was observed in the levels of TNF-� and IFN � in CM-A

Table 1. Quantification of cytokine production by macrophages obtained from pleural effusions of cancer patients and control subjects*

Type of macrophage-conditioned medium

% of malignantcells in pleural fluids

Production of cytokines by macrophages (pg/mL)

TRAIL TNF-� IFN � FasL

CM-A 1–5† 980–1300 150–210 20–34 UDCM-B >50‡ 0–50 90–180 18–35 UDCM-B 0§ 0 60–130 16–24 UD

*TRAIL � TNF-related apoptosis-inducing ligand; TNF-� � tumor necrosis factor �; INF � � interferon �; FasL � Fas ligand; CM-A � conditioned mediumA (culture supernatants from macrophages isolated from pleural effusions of patients with small numbers of tumor cells [1%–5%]); CM-B � conditioned mediumB (culture supernatants from macrophages from pleural effusion of patients with large numbers of tumor cells [>50%] or from patients with congestive heart failurewith no tumor cells); UD � undetectable.

†Pleural effusions from cancer patients (one with adenocarcinoma of the colon, one with adenocarcinoma of the breast, one with adenocarcinoma of the lung, andtwo with pleural mesothelioma). All patients had stage IV metastatic disease.

‡Pleural effusions from cancer patients (three with adenocarcinoma of the lung and one with pleural mesothelioma). All patients had stage IV metastatic disease.§Pleural effusions from five patients with congestive heart failure.


by guest on Novem


Dow

nloaded from


or CM-B. FasL was not detected in any macrophage-conditionedmedium.

TRAIL mRNA expression was observed in freshly isolatedmacrophages only from cancer patients whose pleural effusionscontained a low level of tumor cells (<5%). This level of mRNAexpression was directly related to the amount of TRAIL proteinproduced after 48 hours of culture by macrophages isolated fromthe same pleural effusion (Fig. 1, lane 1). We used flow cytom-etry to assess macrophage membrane-associated TRAIL andfound that TRAIL was bound to the membranes of macrophagesthat produced CM-A but not to the membranes of macrophagesthat produced CM-B from control patients with congestive heartfailure. Membrane-associated TRAIL was weakly detected onmacrophages that produced CM-A from three of the five cancerpatients (data not shown).

Analysis of Apoptosis Mediated by Cytokines and byConditioned Medium

Colo 205, Colo 320, and Caco-2 cells were cultured in me-dium containing 50% CM-A or CM-B. As shown in Fig. 2, A,after 4 days of culture with CM-A, the number of cells in un-treated cultures was greater than that in treated cultures: for Colo205 cells, 590 000 untreated cells versus 230 000 treated cells(inhibition of 62%, 95% confidence interval [CI] � 57% to67%); for Colo 320 cells, 550 000 untreated cells versus 330 000treated cells (inhibition of 41%, 95% CI � 36% to 46%); forCaco-2 cells, 590 000 untreated cells versus 350 000 treated cells(inhibition of 40%, 95% CI � 37% to 43%). The numbers ofcells in untreated cultures and in cultures treated with mediumcontaining 50% CM-B were similar: for Colo 205 cells, 590 000untreated cells versus 545 000 treated cells (8% of inhibition,95% CI � 3% to 12%); for Colo 320 cells, 550 000 untreatedcells versus 530 000 treated cells (inhibition of 3%, 95% CI �1% to 5%); for Caco-2 cells, 590 000 treated cells versus 572 000treated cells (inhibition of 4%, 95% CI � 2% to 6%).

To determine whether apoptosis was responsible for the de-creased number of Colo 205 cells cultured with CM-A, we usedAnnexin V and flow cytometry to assay for apoptosis in cellscultured with 50% CM-A or 50% CM-B or in control cellscultured in medium alone. After 24 hours of culture, 40% (95%

CI � 30% to 50%) of Colo 205 cells cultured with CM-A wereapoptotic (Fig. 2, B, 1), but essentially no Colo 205 cells cul-tured with CM-B (8%, 95% CI � 3% to 13%) or medium alone(3%, 95% CI � 1% to 5%) were apoptotic (Fig. 2, B, 2). After3 days of culture, 66% (95% CI � 60% to 72%) of Colo 205cells cultured with CM-A were in the final stage of apoptosis,as determined by the TUNEL method (Fig. 2, C, lower leftpanel), whereas 12% (95% CI � 7% to 17%) of cells culturedwith CM-B (Fig. 2, C, lower right panel) and 8% (95% CI � 6%to 10%) of control cells cultured with medium alone (Fig. 2,C, upper right panel) were in the final stage of apoptosis. Asa positive control, when cells were treated with DNase to frag-ment the DNA, as observed in the final stage of apoptosis, 98%(95% CI � 96% to 100%) of cells were apoptotic. Similarresults were obtained with Colo 320 and Caco-2 cells (data notshown).

As shown in Fig. 3, A, recombinant TRAIL and recombinantTNF-� induced apoptosis in Colo 205 cells in a dose-dependentmanner. Recombinant TRAIL (100 ng/mL) induced 90% (95%CI � 85% to 95%) of Colo 205 cells to undergo apoptosis.Recombinant TRAIL (100 ng/mL) induced apoptosis in 80%(95% CI � 69% to 91%) of Colo 320 cells and 75% (95% CI� 64% to 86%) of Caco-2 cells. Recombinant TRAIL inducedapoptosis more effectively and more rapidly than recombinantTNF-�. With recombinant TRAIL (50 ng/mL), apoptosis wasobserved after only 4 hours of culture, but with recombinantTNF-� (50 ng/mL), apoptosis was observed after 12 hours ofculture (data not shown). Colo 205 cells were moderately sen-sitive to recombinant IFN � but were resistant to recombinantFasL at all concentrations tested.

Because only high concentrations of cytokines used individu-ally induced apoptosis, we determined whether combinationsof recombinant TRAIL, TNF-�, and IFN � at lower concentra-tions could induce apoptosis. As shown in Fig. 3, B, the com-bination of recombinant TNF-� and IFN � (each at 10 ng/mL)or the combination of recombinant TNF-� and TRAIL (each at10 ng/mL) induced 53% (95% CI � 43% to 63%) and 57%(95% CI � 51% to 64%), respectively, of Colo 205 cells toundergo apoptosis, whereas recombinant TNF-�, INF�, orTRAIL (each at 10 or 20 ng/mL) used alone did not induceapoptosis. Recombinant TNF-� and TRAIL acted synergisti-cally because the combination index D was 0.45 (P<.001). Thecombination of recombinant TRAIL, TNF-�, and IFN � (eachat 10 ng/mL) induced 65% (95% CI � 50% to 80%) of cellsto undergo apoptosis. When we tested recombinant TRAIL at1.3 ng/mL and TNF-� at 200 pg/mL (which are the concentra-tions detected in CM-A), we did not observe any apoptosis ofColo 205 cells.

To determine whether the caspase cascade pathway is in-volved in Colo 205 cell death, we used the caspase inhibitorzVAD-fmk to specifically block caspase-mediated apoptosis. Asshown in Fig. 4, A, zVAD-fmk at 50 �M dramatically reducedthe number of apoptotic cells that were induced by CM-A (63%reduction, 95% CI � 53% to 73%), recombinant TRAIL (82%,95% CI � 76% to 90%), or TNF-� (73%, 95% CI � 60% to86%).

Recombinant DR5 at 5 �g/mL, which competitively binds toTRAIL and blocks the interaction between TRAIL and DR5 ontarget cells, inhibited apoptosis mediated by CM-A and recom-binant TRAIL (Fig. 4, B). Only 35% (95% CI � 28% to 42%)of Colo 205 cells underwent apoptosis when cultured with re-

Fig. 1. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) mRNAproduced by macrophages. Macrophages were cultured for 2 days, and mRNAwas extracted and amplified by reverse transcription–polymerase chain reaction.Lane 1 � mRNA extracted from macrophages isolated from a pleural effusionwith low numbers of tumor cells; lane 2 � mRNA extracted from macrophagesisolated from a pleural effusion with no tumor cells; lane 3 � no complementaryDNA (cDNA; blank); lane M � molecular weight standards. TRAIL mRNAis 874 base pairs (bp), and �2-macroglobulin mRNA (the loading control) is280 bp. Data shown are representative of three experiments, all with similarresults.


by guest on Novem


Dow

nloaded from


combinant TRAIL (50 ng/mL) and recombinant DR5, comparedwith 88% (95% CI � 77% to 99%) when cultured with TRAILalone. The inhibition was greater when cells were cultured withCM-A and DR5—only 15% (95% CI � 8% to 22%) of cellswere apoptotic—compared with 40% (95% CI � 30% to 50%)when cells were cultured with CM-A alone. Recombinant DR5could also inhibit CM-A-mediated apoptosis in Colo 320 cells(65% inhibition, 95% CI � 61% to 69%) and Caco-2 cells (71%

inhibition, 95% CI � 68% to 74%). Use of higher concentra-tions of recombinant DR5 (10 and 20 mg/mL) did not increasethe inhibition of CM-A-mediated apoptosis.

Effect of CM-A on the Expression of TRAIL Receptor inColo 205 Cells

Colo 205 cells were cultured alone (control) or with 50%CM-A for 24 hours, and then TRAIL receptors were measured.

Fig. 2. Conditioned medium and Colo 205 cell apoptosis. A) Cell count. Colo205, Colo 320, and Caco-2 cells were cultured at a density of 5 × 104 cells perwell in 24-well plates for 4 days in medium containing 50% macrophage-conditioned medium A (CM-A) or macrophage-conditioned medium B (CM-B).Viable cells were counted after trypan blue dye exclusion. Data are representa-tive of five experiments (mean ± 95% confidence intervals), all with similarresults. B) Annexin V test for apoptosis. Colo 205 cells were cultured for 24hours in culture medium containing 50% CM-A or 50% CM-B. Apoptotic cellswere then quantified by the Annexin V method and flow cytometry. The cyto-gram shows apoptotic cells visualized by fluorescein isothiocyanate (FITC)-conjugated Annexin V (Annexin V-FITC) staining plotted against the number ofcells (propidium iodide-negative). Panel 1: Annexin V-FITC intensity, a mea-sure of the number of apoptotic cells, of Colo 205 cells cultured with 50% CM-Aversus control Colo 205 cells cultured with medium alone. The shift of thefluorescence peak for CM-A-treated cells corresponded to 33% of cells beingapoptotic. Panel 2: Annexin V-FITC intensity of Colo 205 cells cultured with50% CM-B versus control cells cultured with medium alone. A shift of the

fluorescence peaks of CM-B-treated cells and control cells was not detected.Data shown are representative of five experiments, all with similar results. C)Terminal deoxynucleotidyltransferase-mediated deoxyuridine 5�-triphosphate(dUTP) nick-end labeling (TUNEL) assay for apoptosis. Colo 205 cells werecultured for 3 days in culture medium containing 50% CM-A. Apoptosis wasanalyzed by the TUNEL assay and flow cytometry. Apoptotic cells were visu-alized on a cytogram with dual parameter analysis, and propidium iodide inten-sity was plotted versus biotinylate-dUTP staining, reflected by Avidin-FITCintensity. Upper windows � DNase-treated cells with total DNA fragmenta-tion, as observed at the final stage of apoptosis; C+ � DNase-treated cells(positive control); Untreated � Colo 205 cells cultured without conditionedmedium; CM-A � Colo 205 cells cultured with 50% CM-A; CM-B � Colo 205cells cultured with 50% CM-B. We observed an increase in DNA fragmentationin CM-A-treated cells (66% of cells in apoptosis) compared with CM-B-treatedand untreated cells. Data are representative of three experiments, all with similarresults.


by guest on Novem


Dow

nloaded from


The levels of TRAIL receptors DR4 and DR5 on Colo 205 cellmembranes were approximately 13-fold and sixfold, respec-tively, higher in treated cells than in untreated control cells (Fig.5). As measured by the mean fluorescence intensity (MFI), lev-els of DR4 and DR5 were 3.22 (95% CI � 1 to 5.5) and 4 (95%CI � 1 to 7.3), respectively, for untreated cells and 43 (95% CI� 32 to 54) and 25 (95% CI � 20 to 30), respectively, forCM-A-treated cells. The expression of DcR1 and DcR2 wasunchanged (Fig. 5): 8 (95% CI � 5 to 11) and 3 (95% CI � 1to 5), respectively, for untreated cells and 9 (95% CI � 4 to 14)and 3 (95% CI � 1 to 5), respectively, for CM-A-treated cells.

We also investigated the combination of recombinant TNF-�and IFN � and the combination of recombinant TNF-� andTRAIL on Colo 205 cells. Recombinant IFN � alone moderatelyincreased the expression of DR4 by 3.5-fold, but increased DR5expression by only 1.33-fold (Table 2). The combination of

recombinant TNF-� and IFN � induced a 5.6-fold increase inDR4 expression and a 1.22-fold increase in DR5 expression.Recombinant TNF-�, TRAIL, and IFN �, alone or in combina-tion, did not alter the expression of DcR1 and DcR2 on Colo 205cells, i.e., the increase of DcR1 or DcR2 expression was alwaysless than 1.1.

DISCUSSION

To investigate the possible effects that macrophages mayhave on colorectal cancer cells, we developed a human cellculture model that approximates physiologic conditions. Culturesupernatants from macrophages that were isolated from pleuraleffusions of patients with benign or malignant diseases wereused to stimulate Colo 205, Colo 320, and Caco-2 colon adeno-carcinoma cell lines. Our findings indicate that macrophagesconsistently produced similar amounts of TNF-� and IFN �,irrespective of the patient’s disease. However, macrophagesfrom some cancer patients produced higher levels of TRAILthan macrophages from other cancer patients or from noncancerpatients, suggesting that the former macrophages may have beenactivated by tumor cells in the pleural space. We also showedthat TRAIL was the most active of the three apoptosis-associated cytokines secreted by macrophages into conditionedmedium.

We found that, for macrophages isolated from pleural effu-sions, the level of TRAIL produced by macrophages was depen-dent on the number of tumor cells in the corresponding pleuraleffusion. High levels of TRAIL were produced by macrophagesthat had been in contact with small numbers of tumor cells(CM-A), and low levels of TRAIL were produced by macro-phages that had been in contact with large numbers of malignantcells or no malignant cells (CM-B). Thus, the number of tumorcells to which macrophages were exposed in the pleural effusionappeared to directly affect the production of TRAIL by macro-phages. Macrophages may need to be exposed to malignant cellsto produce TRAIL, but exposure to a large number of cancercells appears to reduce the production of TRAIL. A similareffect has been described for the production of TNF-�, and thiseffect is inhibited by exposure of the macrophages to immuno-suppressive molecules such as interleukin 10 that are secreted bytumor cells (33,34).

We investigated whether conditioned medium from macro-phages had antitumor activity against Colo 205, Colo 320, andCaco-2 colon tumor cells and found that CM-A containing highlevels of TRAIL induced apoptosis in tumor cells but that CM-Bcontaining low levels of TRAIL did not induce apoptosis, de-spite similar concentrations of TNF-� and IFN � in both CM-Aand CM-B. Thus, TRAIL appears to play a predominant role inthe induction of apoptosis in Colo 205 cells. Ligand-bound deathreceptors of the TNF receptor family, such as TRAIL receptorsDR4 and DR5, activate a signaling pathway that activates thecaspase cascade. This caspase cascade can be blocked by thecaspase inhibitor zVAD-fmk (35). In our model, apoptosis wasblocked by zVAD-fmk, suggesting that apoptosis in our modelinvolved the caspase cascade pathway. Conditioned medium-mediated apoptosis was complete, as demonstrated by DNAfragmentation detected by the TUNEL method. Finally, recom-binant DR5 inhibited 62.5% of CM-A-induced apoptosis, dem-onstrating that TRAIL participated in CM-A-mediated celldeath. However, because the inhibition was not complete, wecannot exclude the possibility that other factors produced by

Fig. 3. Recombinant tumor necrosis factor (TNF)-related apoptosis-inducingligand (TRAIL), TNF-�, interferon � (IFN �), Fas ligand (FasL), and Colo 205cell apoptosis. A) Dose–response experiments with individual cytokines. Colo205 cells were cultured for 24 hours with increasing amounts of recombinantTRAIL, TNF-�, IFN �, or FasL (from 0 to 100 ng/mL). Apoptosis was quan-tified by the Annexin V method and flow cytometry. Results are expressed as thepercentage of apoptotic Colo 205 cells (mean ± 95% confidence intervals [CIs]of five experiments). B) Combination cytokine treatment. Colo 205 cells werecultured for 24 hours alone (control) or with recombinant TRAIL, TNF-�, and/orIFN � at 10 ng/mL, separately or in combinations, as indicated. Apoptosis wasanalyzed by the Annexin V method and flow cytometry. Results are presented asthe mean ± 95% CIs of five experiments.


by guest on Novem


Dow

nloaded from


macrophages may play a role in conditioned medium-inducedapoptosis.

We showed that only TRAIL and TNF-� of the cytokinestested were able to induce Colo 205 cells to undergo apoptosis,with TRAIL being more efficient and exerting a more rapideffect than TNF-�. Colo 205 cells contain high levels of DR4and DR5 cell-surface TRAIL death receptors but contain TNF-R1, the only known TNF-� death receptor, which may explainthis observation. TNF-� can induce apoptosis in tumor cells andalso promote tumor invasion and metastasis (36). In breast can-cer, TNF-� produced by tumor-infiltrating macrophages in-creases angiogenesis (37). However, the severe toxicity ofTNF-� in normal tissues has ruled it out for cancer therapy(38,39). In contrast, TRAIL, which has antitumor activity bothin vivo and in vitro, does not appear to be cytotoxic to normalhuman tissues (13,40).

Griffith et al. (14,15) showed that a wide variety of humantumor cell lines are sensitive to TRAIL in vitro, and the antitu-mor activity of TRAIL has been demonstrated against humanthyroid and colon cancer cells transplanted into nude mice (41–43). Moreover, the tumor doubling times of established Colo205 tumors in mice receiving daily intravenous injections ofsoluble recombinant human TRAIL are statistically significantly

longer in a TRAIL dose-dependent manner (44). We also dem-onstrate that, although Colo 205 cells were sensitive to TRAIL-mediated apoptosis, they were resistant to FasL-mediated apo-ptosis, as observed with human melanoma cell lines (16).

Oligomeric TRAIL is more biologically active than mono-meric TRAIL. The fact that recombinant TRAIL forms oligo-mers less efficiently than natural TRAIL (information providedby R&D Systems) could account for our findings that only highlevels of recombinant TRAIL (50 ng/mL) induced apoptosis inour Colo 205 cell model and that the TRAIL concentration inconditioned medium was quite low (1 ng/mL). Consequently,we tested recombinant TRAIL in combination with recombinantTNF-� and/or IFN �. TRAIL and TNF-� acted synergistically toinduce Colo 205 cells to undergo apoptosis, perhaps becauseTRAIL death receptors and TNF-� death receptors use similarsignaling pathways to induce apoptosis. DR4, DR5, and TNF-R1activate the adapter proteins TRADD (TNF receptor-associateddeath domain) and FADD (Fas-associated death domain) (45).FADD in turn activates caspase 8, thereby initiating a series ofcaspase-dependent events that lead to cell death. Therefore, lowlevels of TNF-� and TRAIL may stimulate death receptorsenough to activate FADD protein, which in turn activatescaspase 8.

Fig. 4. Inhibition of Colo 205 cell apoptosis.A) Inhibition of apoptosis with the caspase in-hibitor zVAD-fmk [benzyloxycarbonyl Val-Ala-Asp(OMe)-CH2]. Colo 205 cells were incubatedwith 50 �M zVAD-fmk for 4 hours at 37 °C andthen cultured for 24 hours in the presence of50% conditioned medium and recombinantTRAIL or TNF-�. Cells were then analyzed byflow cytometry using the Annexin V method.Results shown are the means ± 95% confidenceintervals of four experiments. B) Inhibition ofcell apoptosis with recombinant DR5. Colo 205cells were cultured for 24 hours with recombi-nant TRAIL (panel 1) or with CM-A (panel 2)and recombinant DR5 (5 �g/mL) was added toall cultures. Apoptosis was quantified withAnnexin V and flow cytometry.


by guest on Novem


Dow

nloaded from


Because the concentration of TRAIL in our conditioned me-dium was too low to induce apoptosis, even in combination withTNF-� or IFN �, we investigated the expression of TRAILreceptors DR4 and DR5 on Colo 205 cell membranes. We found

that treatment with CM-A increased the expression of TRAILdeath receptors DR4 and DR5 but did not affect the expressionof decoy receptors. Consequently, we hypothesize that increas-ing the expression of TRAIL death receptors increases the sen-sitivity of Colo 205 cells to TRAIL, even to low concentrationsof TRAIL.

Regulation of the TRAIL receptors could be important foroptimizing the therapeutic use of TRAIL. TNF-� and TRAIL actsynergistically to increase DR5 expression on epithelial tumorcells and to induce these cells to undergo apoptosis (46). Inepithelial tumor cells, the binding of TNF-� to its receptor TNF-R1increases the expression of DR5 and thereby enhances the abilityof TRAIL to induce apoptosis. In our study, we observed that,in Colo 205 cells, the combination of TNF-� and IFN � or thecombination of TNF-� and TRAIL moderately increased the ex-pression of DR4 but not of DR5. These findings raise the pos-sibility that other macrophage-secreted molecules mediate theexpression of DR5.

In summary, we have demonstrated for the first time, to thebest of our knowledge, that human macrophages harvested froma metastatic site can, presumably, be activated by tumor cellslocated at that site to produce large amounts of TRAIL. Weshowed that TRAIL could induce Colo 205 tumor cells to un-

Table 2. Effect of recombinant TRAIL, TNF-�, and IFN � molecules onexpression of death and decoy receptors on Colo 205 cell membrane, as

determined by flow cytometry*

Treatment of Colo 205 cells

Mean fluorescence intensity (arbitrary units)

DR4 DR5 DcR1 DcR2

None 3 4.5 8 2CM-A 43 24 9 2.6TRAIL 5 5 9 3TNF-� 5 5 8 3IFN � 11 6 7.5 2.5TNF-� + IFN � 17 5.5 8 2.5TRAIL + TNF-� 5 5 8 3TRAIL + IFN � 11 6 8 3

*TRAIL � TNF-related apoptosis-inducing ligand; TNF-� � tumor necrosisfactor �; IFN � � inteferon �; DR � death receptor; DcR � decoy receptor;CM-A � conditioned medium A (culture supernatants from macrophagesisolated from pleural effusions of patients with small numbers of tumor cells[1%–5%]). Data presented are representative of three experiments.

Fig. 5. Macrophage-conditioned medium A (CM-A) and membrane-associatedtumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) receptorexpression on Colo 205 cells. Colo 205 cells were cultured for 24 hours alone orwith 50% CM-A. Cells were then incubated with goat anti-human death receptor4 (DR4), death receptor 5 (DR5), decoy receptor 1 (DcR1), decoy receptor 2(DcR2), or isotype control antibodies (each at 5 �g/mL). Expression of TRAILreceptors was analyzed by flow cytometry. Colo 205 cells with isotypic mono-clonal antibodies (left column), untreated cells with anti-receptor monoclonal

antibodies (middle column), and CM-A-treated cells with anti-receptor mono-clonal antibodies (right column) are shown. The mean fluorescence intensity(MFI), which indicates the relative amount of fluorescent antibody bound to thecell examined, is shown below each panel. Expression of DR4 and DR5 in-creased by approximately 13-fold and sixfold, respectively, on cells culturedwith CM-A compared with untreated control cells. Expression of DcR1 andDcR2 was unchanged. Data are representative of three experiments.


by guest on Novem


Dow

nloaded from


dergo apoptosis and that TRAIL and TNF-� act synergisticallyto induce apoptosis in these cells. The synergistic effect ofTRAIL and TNF-� may have implications for subsequent clini-cal applications, i.e., TRAIL might act as a coactivator of tumorcell death that would permit the use of subtoxic doses of anotheranticancer drug. We also showed that macrophages producemolecules that increased the expression of death receptorsDR4 and DR5 on tumor cells, which may make such cells sus-ceptible to TRAIL-induced apoptosis at lower concentrations ofTRAIL.

REFERENCES

(1) Abe K, Harada M, Tamada K, Ito O, Li T, Nomoto K. Early-appearingtumor-infiltrating natural killer cells play an important role in the nitricoxide production of tumor-associated macrophages through their interferonproduction. Cancer Immunol Immunother 1998;45:225–33.

(2) Koshita Y, Lu Y, Fujii S, Neda H, Matsuyama T, Satoh Y, et al. Efficacyof TNF-alpha gene-transduced tumor cells in treatment of established invivo tumor. Int J Cancer 1995;63:130–5.

(3) Andreesen R, Hennemann B, Krause SW. Adoptive immunotherapy ofcancer using monocyte-derived macrophages: rationale, current status, andperspectives. J Leukoc Biol 1998;64:419–26.

(4) Strassmann G, Springer TA, Somers SD, Adams DO. Mechanisms of tu-mor cell capture by activated macrophages: evidence for involvement oflymphocyte function-associated (LFA)-1 antigen. J Immunol 1986;136:4328–33.

(5) Toomey D, Harmey J, Condron C, Kay E, Bouchier-Hayes D. Phenotypingof immune cell infiltrates in breast and colorectal tumours. Immunol Invest1999;28:29–41.

(6) Joseph IB, Isaacs JT. Macrophage role in the anti-prostate cancer responseto one class of antiangiogenic agents. J Natl Cancer Inst 1998;90:1648–53.

(7) Ishihara Y, Iijima H, Matsunaga K. Contribution of cytokines on the sup-pression of lung metastasis. Biotherapy 1998;11:267–75.

(8) Bonta IL, Ben-Efraim S. Involvement of inflammatory mediators in mac-rophage antitumor activity. J Leukoc Biol 1993;54:613–26.

(9) Eifuku R, Yoshimatsu T, Yoshino I, Takenoyama M, Imahayashi S, So T,et al. Heterogeneous response patterns of alveolar macrophages from pa-tients with lung cancer by stimulation with interferon-gamma. Jpn J ClinOncol 2000;30:295–300.

(10) Mace KF, Ehrke MJ, Hori K, Maccubbin DL, Mihich E. Role of tumornecrosis factor in macrophage activation and tumoricidal activity. CancerRes 1988;48:5427–32.

(11) Gruss HJ, Duyster J, Herrmann F. Structural and biological features of theTNF receptor and TNF ligand superfamilies: interactive signals in thepathobiology of Hodgkin’s disease. Ann Oncol 1996;7:19–26.

(12) Wiley SR, Schooley K, Smolak PJ, Din WS, Huang CP, Nicholl JK, et al.Identification and characterization of a new member of the TNF family thatinduces apoptosis. Immunity 1995;3:673–82.

(13) Gura T. How TRAIL kills cancer cells, but not normal cells. Science1997;277:768.

(14) Griffith TS, Chin WA, Jackson GC, Lynch DH, Kubin MZ. Intracellularregulation of TRAIL-induced apoptosis in human melanoma cells. J Im-munol 1998;161:2833–40.

(15) Griffith TS, Lynch DH. TRAIL: a molecule with multiple receptors andcontrol mechanisms. Curr Opin Immunol 1998;10:559–63.

(16) Thomas WD, Hersey P. TNF-related apoptosis-inducing ligand (TRAIL)induces apoptosis in Fas ligand-resistant melanoma cells and mediates CD4T cell killing of target cells. J Immunol 1998;161:2195–200.

(17) Golstein P. Cell death: TRAIL and its receptors. Curr Biol 1997;7:R750–3.(18) Pan G, O’Rourke K, Chinnaiyan AM, Gentz R, Ebner R, Ni J, et al. The

receptor for the cytotoxic ligand TRAIL. Science 1997;276:111–3.(19) Schneider P, Bodmer JL, Thome M, Hofmann K, Holler N, Tschopp J.

Characterization of two receptors for TRAIL. FEBS Lett 1997;416:329–34.(20) Degli-Esposti MA, Smolak PJ, Walczak H, Waugh J, Huang CP, DuBose

RF, et al. Cloning and characterization of TRAIL-R3, a novel member ofthe emerging TRAIL receptor family. J Exp Med 1997;186:1165–70.

(21) Degli-Esposti MA, Dougall WC, Smolak PJ, Waugh JY, Smith CA, Good-win RG. The novel receptor TRAIL-R4 induces NF-kappaB and protectsagainst TRAIL-mediated apoptosis, yet retains an incomplete death do-main. Immunity 1997;7:813–20.

(22) Smyth MJ, Cretney E, Takeda K, Wiltrout RH, Sedger LM, Kayagaki N,et al. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)contributes to interferon gamma-dependent natural killer cell protectionfrom tumor metastasis. J Exp Med 2001;193:661–70.

(23) Vidalain PO, Azocar O, Lamouille B, Astier A, Rabourdin-Combe C,Servet-Delprat C. Measles virus induces functional TRAIL production byhuman dendritic cells. J Virol 2000;74:556–9.

(24) Strebel A, Bachmann F, Wernli M, Erb P. Tumor necrosis factor-related,apoptosis-inducing ligand supports growth of mouse mastocytoma tumorsby killing tumor-infiltrating macrophages. Int J Cancer 2002;100:627–34.

(25) Sheridan JP, Marsters SA, Pitti RM, Gurney A, Skubatch M, Baldwin D,et al. Control of TRAIL-induced apoptosis by a family of signaling anddecoy receptors. Science 1997;277:818–21.

(26) Weissflog D, Kroegel C, Luttmann W, Grahmann PR, Hasse J. Leukocyteinfiltration and secretion of cytokines in pleural drainage fluid after tho-racic surgery: impaired cytokine response in malignancy and postoperativecomplications. Chest 1999;115:1604–10.

(27) Weissler JC, Lipscomb MF, Lem VM, Toews GB. Tumor killing by humanalveolar macrophages and blood monocytes. Decreased cytotoxicity of hu-man alveolar macrophages. Am Rev Respir Dis 1986;134:532–7.

(28) Yanagawa H, Takeuchi E, Suzuki Y, Ohmoto Y, Bando H, Sone S. Pres-ence and potent immunosuppressive role of interleukin-10 in malignantpleural effusion due to lung cancer. Cancer Lett 1999;136:27–32.

(29) Gjomarkaj M, Pace E, Melis M, Spatafora M, Toews GB. Mononuclearcells in exudative malignant pleural effusions. Characterization of pleuralphagocytic cells. Chest 1994;106:1042–9.

(30) Lombard Y, Giaimis J, Makaya-Kumba M, Fonteneau P, Poindron P.A new method for studying the binding and ingestion of zymosan particlesby macrophages. J Immunol Methods 1994;174:155–65.

(31) van de Loosdrecht AA, Ossenkoppele GJ, Beelen RH, Broekhoven MG,Drager AM, Langenhuijsen MM. Apoptosis in tumor necrosis factor-alpha-dependent, monocyte-mediated leukemic cell death: a functional, morpho-logic, and flow-cytometric analysis. Exp Hematol 1993;21:1628–39.

(32) Berenbaum MC. Criteria for analyzing interactions between biologicallyactive agents. Adv Cancer Res 1981;35:269–335.

(33) Ito S, Ansari P, Sakatsume M, Dickensheets H, Vazquez N, Donnelly RP,et al. Interleukin-10 inhibits expression of both interferon alpha- and in-terferon gamma-induced genes by suppressing tyrosine phosphorylation ofSTAT1. Blood 1999;93:1456–63.

(34) Erwig LP, Kluth DC, Walsh GM, Rees AJ. Initial cytokine exposure de-termines function of macrophages and renders them unresponsive to othercytokines. J Immunol 1998;161:1983–8.

(35) Luschen S, Ussat S, Scherer G, Kabelitz D, Adam-Klages S. Sensitizationto death receptor cytotoxicity by inhibition of fas-associated death domainprotein (FADD)/caspase signaling. Requirement of cell cycle progression.J Biol Chem 2000;275:24670–8.

(36) Miles DW, Happerfield LC, Naylor MS, Bobrow LG, Rubens RD, BalkwillFR. Expression of tumour necrosis factor (TNF alpha) and its receptors inbenign and malignant breast tissue. Int J Cancer 1994;56:777–82.

(37) Leek RD, Landers R, Fox SB, Ng F, Harris AL, Lewis CE. Association oftumour necrosis factor alpha and its receptors with thymidine phosphory-lase expression in invasive breast carcinoma. Br J Cancer 1998;77:2246–51.

(38) Bonavida B, Ng CP, Jazirehi A, Schiller G, Mizutani Y. Selectivity ofTRAIL-mediated apoptosis of cancer cells and synergy with drugs: the trailto non-toxic cancer therapeutics. Int J Oncol 1999;15:793–802.

(39) Ashkenazi A, Pai RC, Fong S, Leung S, Lawrence DA, Marsters SA, et al.Safety and antitumor activity of recombinant soluble Apo2 ligand. J ClinInvest 1999;104:155–62.

(40) French LE, Tschopp J. The TRAIL to selective tumor death. Nat Med1999;5:146–7.

(41) Ahmad M, Shi Y. TRAIL-induced apoptosis of thyroid cancer cells: po-tential for therapeutic intervention. Oncogene 2000;19:3363–71.

(42) Gliniak B, Le T. Tumor necrosis factor-related apoptosis-inducing ligand’santitumor activity in vivo is enhanced by the chemotherapeutic agent CPT-11. Cancer Res 1999;59:6153–8.


by guest on Novem


Dow

nloaded from


(43) Walczak H, Miller RE, Ariail K, Gliniak B, Griffith TS, Kubin M, et al.Tumoricidal activity of tumor necrosis factor-related apoptosis-inducingligand in vivo. Nat Med 1999;5:157–63.

(44) Kelley SK, Harris LA, Xie D, Deforge L, Totpal K, Bussiere J, et al. Pre-clinical studies to predict the disposition of Apo2L/tumor necrosis factor-related apoptosis-inducing ligand in humans: characterization of in vivoefficacy, pharmacokinetics, and safety. J Pharmacol Exp Ther 2001;299:31–8.

(45) Sprick MR, Weigand MA, Rieser E, Rauch CT, Juo P, Blenis J, et al.FADD/MORT1 and caspase-8 are recruited to TRAIL receptors 1 and 2and are essential for apoptosis mediated by TRAIL receptor 2. Immunity2000;12:599–609.

(46) Sheikh MS, Burns TF, Huang Y, Wu GS, Amundson S, Brooks KS, et al.p53-dependent and -independent regulation of the death receptor KILLER/DR5 gene expression in response to genotoxic stress and tumor necrosisfactor alpha. Cancer Res 1998;58:1593–8.

NOTES

Supported by a grant from the Ligue Nationale Contre le Cancer.We thank C. Ruesch (Hospital of Saint-Chamond, France), P. Sagnol (Hos-

pital of Firminy, France), M. Perol (University Hospital Center of Lyon, France),and J. M. Vergnon (University Hospital Center of Saint-Etienne, France) forproviding pleural effusions. We thank Dr. Shearer (National Cancer Institute,National Institutes of Health, Bethesda, MD), Pr. M. Crépin (University of ParisXIII, France), and Dr. O. Garraud and Pr. D. Pearson (University of Saint-Etienne, France) for helpful discussion and review of the manuscript. We alsothank H. Hamzeh, D. Laurent, and F. Duplat for technical help and C. Moulin(University of Saint-Etienne) for secretarial assistance.

Manuscript received June 18, 2002; revised January 30, 2003; acceptedFebruary 14, 2003.


by guest on Novem


Dow

nloaded from


Documents

Macrophages From Cancer Patients: Analysis of TRAIL, TRAIL Receptors, and Colon Tumor Cell Apoptosis