of July 2, 2015.This information is current as

MacrophagesChemoattractant Protein-1 Production by Hypoxia Selectively Inhibits Monocyte

Annamaria Rapisarda and Luigi VaresioZenghui Mi, Giovanni Melillo, Stefano Massazza, Maria Carla Bosco, Maura Puppo, Sandra Pastorino,

http://www.jimmunol.org/content/172/3/1681doi: 10.4049/jimmunol.172.3.1681

2004; 172:1681-1690; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/172/3/1681.full#ref-list-1

, 35 of which you can access for free at: cites 65 articlesThis article

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2004 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on July 2, 2015

http://ww

w.jim

munol.org/

Dow

nloaded from

Hypoxia Selectively Inhibits Monocyte ChemoattractantProtein-1 Production by Macrophages1

Maria Carla Bosco,2* Maura Puppo,* Sandra Pastorino,† Zenghui Mi,‡ Giovanni Melillo, §

Stefano Massazza,* Annamaria Rapisarda,§ and Luigi Varesio*

Hypoxia, a local decrease in oxygen tension occurring in inflammatory and tumor lesions, modulates gene expression in macro-phages. Because macrophages are important chemokine producers, we investigated the regulatory effects of hypoxia on macro-phage-derived chemokines. We demonstrated that hypoxia inhibits the production of the macrophage and T lymphocyte chemo-tactic and activating factor, monocyte chemoattractant protein-1 (MCP-1). Exposure of mouse macrophages to low oxygen tensionresulted in the down-regulation of constitutive MCP-1 mRNA expression and protein secretion. Hypoxia inhibitory effects wereselective for MCP-1 because the chemokines macrophage inflammatory protein-1� (MIP-1�), RANTES, IFN-�-inducible protein-10, and MIP-2 were not affected, and MIP-1� was induced. Hypoxia also inhibited, in a time-dependent fashion, MCP-1 up-regulation by IFN-� and LPS. Moreover, the inhibitory action of hypoxia was exerted on human monocytic cells. MCP-1 down-regulation was associated with inhibition of gene transcription and mRNA destabilization, suggesting a dual molecular mechanismof control. Finally, we found that the triptophan catabolite picolinic acid and the iron chelator desferrioxamine, which mimichypoxia in the induction of gene expression, differentially regulated the expression of MCP-1. This study characterizes a novelproperty of hypoxia as a selective inhibitor of MCP-1 production induced by different stimuli in macrophages and demonstratesthat down-regulation of gene expression by hypoxia can be controlled at both transcriptional and posttranscriptional levels.Inhibition of MCP-1 may represent a negative regulatory mechanism to control macrophage-mediated leukocyte recruitment inpathological tissues. The Journal of Immunology, 2004, 172: 1681–1690.

T he development of an inflammatory response requires thecoordinated recruitment and activation of different leuko-cyte populations at sites of injury and infection, driven by

the local secretion of appropriate chemotactic signals (1). A criticalrole in the regulation of leukocyte trafficking in vivo is played bythe chemokine superfamily, a group of low-m.w. secreted proteinsproduced by immune or nonimmune cells and endowed with che-motactic and activating properties for specific leukocyte subsets (2,3). The chemokine superfamily has been subdivided into four sub-families, which differ with respect to the number and arrangementof the conserved cysteine residues at the N terminus of the primaryamino acid sequence (CXC or �, CC or �, C or �, and CX3C or�) (2, 3). Monocyte chemoattractant protein-1 (MCP-1),3 the pro-

totype of the CC-chemokine subfamily, is endowed with chemo-tactic and activating properties for macrophages (M�), CD4�/CD8� T lymphocytes, NK cells, and basophils and is criticallyinvolved in the regulation of inflammatory processes and antitu-mor immune responses (4–6).

A primary source for chemokines is activated M� (1, 2, 4, 7),which are important effector and immunoregulatory cells activeagainst infections and tumors (8–10). The extent and magnitude ofa local M� response are regulated by a complex interplay betweenstimulatory and inhibitory signals of various natures, which in-clude stimuli derived from the immune system (9, 11, 12), metab-olites produced by surrounding cells (9, 13), microbial products (7,11), and tissue-specific signals such as changes in O2 tension andpH (14, 15). Modulation of chemokine production by M� wasobserved in response to pro- and anti-inflammatory stimuli, includ-ing various cytokines and growth factors, LPS, 12-O-tetradeca-noylphorbol-13-acetate, or the tryptophan catabolite picolinic acid(PA) (2, 4, 7, 16–18). However, the contribution of environmentalsignals to the control of M�-derived chemokines remains to beelucidated.

M� are ubiquitous cells that reside in the vast majority of tissuesunder normal physiological conditions but that markedly accumu-late in areas of inflammation and tumor growth, and evidence ex-ists that M� reactivity is modulated by stimuli that arise from thepathological microenvironment (10, 19). A common denominatorof many pathological processes is represented by low O2 tension(hypoxia). Hypoxia occurs in cardiovascular, hematological, andpulmonary disorders, inflammatory processes, and fibrosis (re-viewed in Ref. 20). Areas of low O2 concentration are present in

*Laboratory of Molecular Biology, G. Gaslini Institute, Genova, Italy; †Neuro-On-cology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD20892; and ‡Developmental Therapeutics Program and §Science Applications Inter-national Corporation, Tumor Hypoxia Laboratory, National Cancer Institute, Freder-ick, MD 21702

Received for publication June 9, 2003. Accepted for publication November 17, 2003.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by grants from the Italian Association for Cancer Re-search, the San Paolo Company, the Fondazione Italiana per la Lotta al Neuroblas-toma, and the Associazione Italiana per la Glicogenosi and has been funded in partwith federal funds from the National Cancer Institute (National Institutes of Health)under Contract NO1-CO-56000. The content of this publication does not necessarilyreflect the views or policies of the Department of Health and Human Services, nordoes mention of trade names, commercial products, or organization imply endorse-ment by the U.S. government.2 Address correspondence and reprint requests to Dr. Maria Carla Bosco, Laboratoriodi Biologia Molecolare, Istituto Giannina Gaslini, Padiglione 2, L.go GerolamoGaslini 5, 16147 Genova Quarto, Italy. E-mail address: mcbosco1@virgilio.it3 Abbreviations used in this paper: MCP-1, monocyte chemoattractant protein-1; M�,macrophage; PA, picolinic acid; VEGF, vascular endothelial growth factor; HIF-1,hypoxia-inducible factor-1; HRE, hypoxia-responsive element; DFX, desferrioxam-

ine; iNOS, inducible NO synthase; MIP, macrophage inflammatory protein; ActD,actinomycin D; RPA, RNase protection assay; AUH, AU RNA-binding protein/enoyl-coenzyme A hydratase; IP-10, IFN-�-inducible protein-10; Thio-M�, thiogly-collate-elicited peritoneal macrophage; ARE, AU-rich destabilizing element.

The Journal of Immunology

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

solid tumors and are known to contribute to tumor growth, met-astatization, and resistance to radio and chemotherapy (21). It isnow well recognized that hypoxia is an important environmentalstimulus capable of modulating the expression of specific genesinvolved in energy metabolism (glycolytic and mitochondrial en-zymes and glucose transporters), erytropoiesis (erytropoietin), an-giogenesis (vascular endothelial growth factor (VEGF) and plate-let-derived growth factor), vasomotor control (NO synthases), andiron metabolism (transferrin) (20, 22–24). Induction of gene ex-pression by hypoxia is mediated mainly by the hypoxia-induciblefactor-1 (HIF-1), which binds to and transactivates the hypoxia-responsive element (HRE) present in the promoter or enhancerelements of many hypoxia-responsive genes (reviewed in Refs. 20and 22). HIF-1 is a heterodimeric complex composed of the basic-helix-loop-helix periodic-aryl hydrocarbon receptor-single-mindedproteins, HIF-1� (or the recently identified homologues HIF-2�/3�), and HIF-1�, whose expression is tightly controlled by O2

concentrations (20, 22, 23, 25, 26). The � subunit, also known asthe aryl hydrocarbon receptor nuclear translocator, is constitutivelyexpressed in unstimulated cells, whereas the � molecule is thehypoxia-responsive subunit, which is posttranslationally stabilizedagainst ubiquitination and proteosomal degradation and accumu-lates under low O2 conditions (20, 22). Signals other than hypoxiacan activate HIF-1� and can induce gene expression through HREtransactivation in normal oxygen conditions (20, 22, 26, 27), in-cluding iron chelators such as PA (13) and desferrioxamine (DFX)(20, 28).

Hypoxic conditions have been shown to profoundly impact onM� proinflammatory and immunoregulatory responses (29, 30).Genes coding for various inflammatory cytokines, growth, and an-giogenic factors were demonstrated to be modulated in M� ex-posed to low O2 tension both in vitro and in vivo. Specifically,induction of TNF-�, VEGF, platelet-derived growth factor, fibro-blast growth factor ��, IL-6, and IL-1 production by human and/ormouse M� (29–34) and inhibition of GM-CSF expression in LPS-treated human M� were reported (34). Furthermore, we haveshown that hypoxia can act as a costimulus with IFN-� or LPS intriggering the transcriptional activation of the inducible isoform ofthe NO synthase (iNOS) gene in mouse M�, eliciting NO produc-tion (13,15).4

Recently, up-regulation of IL-8 and macrophage inflammatoryprotein-1� (MIP-1�) chemokines was reported in M� exposed tolow O2 concentrations (35, 36). Because the extent and impact ofhypoxia on chemokine production by M� are important questionsfor understanding leukocyte infiltration and activation in patholog-ical tissues, we were interested in further elucidating the regulatoryeffects of hypoxia on M�-derived chemokines. In this study, wecharacterize a novel property of hypoxia as a potent and selectiveinhibitor of both constitutive and IFN-�- or LPS-induced MCP-1production, we demonstrate that hypoxia inhibitory activity can beexerted on both mouse and human M�, and we establish the mo-lecular mechanisms accounting for this effect.

Materials and MethodsCells and culture conditions

The mouse M� cell line ANA-1 was established by infecting fresh BM-derived cells from C57BL/6 mice with the J2 retrovirus (carrying the v-raf/v-myc oncogenes) and was shown to display the phenotypic and functionalfeatures and the morphology of well-differentiated �� (13, 18, 36).ANA-1 M� were cultured in DMEM (Euroclone; Celbio, Milano, Italy)

supplemented with 10% heat-inactivated FCS (HyClone Laboratories, Lo-gan, UT; Celbio), 2 mM L-glutamine, 100 U/ml penicillin, and 100 �g/mlstreptomycin (Celbio). The human THP-1 monocytic cell line was pur-chased from the American Type Culture Collection (Manassas, VA) andcultured in RPMI 1640 (Euroclone) supplemented as described for DMEM.Peritoneal M� were obtained from C57BL/6 mice injected i.p. with 1 mlof 3% thioglicollate broth (Sigma-Aldrich, Milano, Italy). After 4 days, theperitoneal exudate cells were collected by lavage of the peritoneal cavitywith 10 ml of sterile PBS (Euroclone). Cells were washed, resuspended,and plated in RPMI 1640. M� were isolated by adherence to tissue culturedishes, and their purity was �94% as assessed by morphology on Giemsa-stained cytocentrifuge slide preparations. Viability, determined by thetrypan blue dye exclusion test, was �99%. Cells were maintained at 37°Cin a humidified incubator containing 20% O2, 5% CO2, and 75% N2. Forexperimental purposes, cells were cultured in 10- or 15-cm Costar plates(Celbio) at 0.5 � 106 cells/ml and were stimulated for different time pointswith the indicated factors. Hypoxic conditions (i.e., 1% O2) were achievedby culturing the cells in a modular incubator chamber flushed with a gasmixture containing 1% O2, 5% CO2, and balanced N2 at 37°C in a humid-ified atmosphere. Reoxygenation was achieved by transferring the plates toa normoxic humidified incubator.

Reagents

Mouse IFN-� (specific activity � 107 IU/mg) and LPS (from Escherichiacoli serotype 011:B4) were purchased from Life Technologies (Milano,Italy). Recombinant human IFN-� (specific activity � 2.0 � 107 U/mg)was purchased from Roche Diagnostics (Mannheim, Germany). PA andDFX were from Sigma-Aldrich. During the course of experiments, severalbatches of PA were used, and all of them gave consistent and reproducibleresults. PA was dissolved in PBS, and the pH was adjusted to 7.4. Thestock solution was then passed through a 0.2-�m filter, aliquoted, andstored at �20°C. Actinomycin D (ActD; Calbiochem-Novabiochem, LaJolla, CA) was dissolved in ethanol at 1 mg/ml and used at a final con-centration of 5 �g/ml. The endotoxin content, as determined by a chro-mogenic Limulus amebocyte lysate test (QCL-1000; BioWhittaker, Walk-ersville, MD), was below the detection limit of 6 pg/ml in all of thereagents used.

Northern blot analysis

Total cellular RNA was purified from ANA-1 M� using the TRIzol reagent(Life Technologies) according to the manufacturer’s instructions (with theaddition of one extra chloroform extraction to improve the quality of re-covered RNA), resuspended in diethyl pyrocarbonate water, and quantifiedwith a spectrophotometer at 260 nm absorbance. A total of 20 �g of RNAfrom each sample was electrophoresed under denaturing conditions on a1.2% agarose gel containing 2.2 M formaldehyde, transferred onto Nytranmembranes (Schleicher & Schuell, Keene, NH), and cross-linked by UVirradiation. Filters were hybridized in Hybrisol I hybridization solution(Oncor, Gaithersburg, MD) at 42°C overnight with 2 � 106 cpm/ml of32P-labeled probe. Probes were labeled by random priming reaction usinga commercial kit (Life Technologies) and 5�-[�32P]dCTP (3000 Ci/mmol;Amersham, Milano, Italy). For MCP-1 and MIP-1� detection, the mouseJE/cDNA, the human MCP-1 full-coding sequence from the PUC19 vector,and the mouse MIP-1� full-length cDNA from the pBR322 vector, kindlyprovided by Dr. A. Sica (Istituto Mario Negri, Milano, Italy), were used.For iNOS detection, the cDNA probe specific for mouse M�-inducibleNOS was used (28). The pEMBL-8 vector containing the �-actin cDNAwas kindly provided by Prof. C. Garre’ (IBiG, Facolta’di Medicina eChirurgia, Universita’ di Genova, Italy). Membranes were then washedthree times at 42°C for 10 min in 2� SSC/0.1% SDS and twice at 65°C for15 min in 0.2� SSC/0.1% SDS before being autoradiographed usingKodak XAR-5 film (Eastman Kodak, Rochester, NY) and intensifyingscreens at �80°C. To obtain comparable band intensities, blots were ex-posed for different periods of time depending on the probe used for hy-bridization. When needed, densitometric analysis of the autoradiographswas performed using the VersaDoc Image Analyzer from Bio-Rad (Her-cules, CA), and quantitative assessment of the band intensities was con-ducted. The significance of mRNA expression differences was determinedby the Student t test (significant difference, p � 0.01).

RNase protection assay (RPA)

Total RNA extracted from ANA-1 M� with the TRIzol reagent was sub-jected to RPA analysis using the RiboQuant MultiProbe RNase ProtectionAssay System from BD PharMingen (Milano, Italy), as described (18).Briefly, a 32P-labeled antisense RNA probe set specifically for differentmouse ��-chemokines (mCK-5) was hybridized in excess to 10 �g of total

4 A. Rapisarda, L. S. Taylor, A. Brooks, L.Varesio, and G. Melillo. Synergistic ac-tivation of hypoxia inducible factor 1 (HIF-1) mediates induction of nitric oxidesynthase expression by lipopolysaccharide and hypoxia in macrophages. Submittedfor publication.

1682 MCP-1 INHIBITION BY HYPOXIA IN MACROPHAGES

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

RNA from each sample in solution, after which free probe and other single-stranded RNA were digested with RNases. The “RNase-protected” 32P-labeled probes, annealed to homologous sequences in the sample RNA,were resolved on denaturing PAGE and visualized by autoradiography(Kodak XAR-5 films).

Detection of cytokine release

Cell-free supernatants were harvested and assayed for mouse and humanMCP-1 content using specific ELISA kits from BioSource International(Milano, Italy) and R&D Systems (Milano, Italy) with a sensitivity of 9 and5 pg/ml, respectively. The ODs of the plates were determined using aDynatech MR 5000 plate reader set to 450 nm (Dynatech Laboratories,Chantilly, VA).

Nuclear run on

Run on experiments were performed as described (18). Briefly, nuclei wereisolated from 10 � 107 cells/sample by cell lysis in 6 ml of lysis buffer (10mM Tris-Cl (pH 7.4), 10 mM NaCl, 3 mM MgCl2, 150 mM sucrose, and0.4% Nonidet P-40 (Sigma-Aldrich)) for 5 min on ice. Nuclei were col-lected after centrifugation at 800 rpm for 5 min at 4°C, resuspended in 150�l of freezing buffer (50 mM Tris-Cl (pH 8.3), 40% glycerol, 5 mMMgCl2, and 0.1 mM EDTA), and stored at �80°C until they were used. Invitro RNA elongation was performed by adding 150 �l of 2� transcriptionbuffer (200 mM KCl, 20 mM Tris-Cl (pH 8), 10 mM MgCl2, 200 mMsucrose, 20% glycerol, 1 mM adenosine triphosphate lithium salt,guanosine triphosphate lithium salt, and cytidine triphosphate lithium salt(Boehringer Mannheim, Indianapolis, IN)) and 100 �Ci of 800 Ci/mmol[�32P]uridine triphosphate (NEN, Boston, MA) to 150 �l of nuclei sus-pension, and the mixture was incubated at 29°C for 30 min. Twenty mi-croliters of 100 mM CaCl2 and 20 U of RNase free DNase I (PromegaItalia, Milano, Italy) were added to the mixture, and the incubation wasallowed to continue for 10 min at 30°C with gentle mixing every 2 min.The nuclei were lysed with 1 ml of TRIzol and the RNA was isolatedaccording to the manufacturer’s procedure. Equal amounts of labeled elon-gated transcripts were added in 4 ml of Hybrisol I to Nytran membranes onwhich 1 �g of linearized MCP-1/JE and of �-actin cDNAs were immobi-lized using a dot blot apparatus (Bio-Rad). Hybridization was performed at42°C for 48 h, and filters were then washed as described for Northernanalysis and were autoradiographed. The autoradiographs were thenscanned using the VersaDoc Image Analyzer.

Real-time PCR

Total RNA from ANA-1 cells was obtained using an RNA Mini Kit (Qia-gen, Valencia, CA). One microgram of total RNA was used to performRT-PCR using an RT-PCR kit (PE Biosystems, Foster City, CA). Theconditions used for RT-PCR were as follows: 10 min at 25°C, 30 min at48°C, and 5 min at 95°C. To measure murine VEGF and murine AURNA-binding protein/enoyl-coenzyme A hydratase (AUH) expression, re-al-time PCR was performed using an ABI-Prism 7700 sequence detector(Applied Biosystems, Foster City, CA), as previously described (37). Typ-ically, 5 ng of reverse transcribed cDNA per sample was used to performreal-time PCR in triplicate. Real-time PCR cycles started with 2 min at50°C, 10 min at 95°C, and then 40 cycles of the following: 15 s at 95°C and1 min at 60°C. Primers and specific probes were obtained from AppliedBiosystems. The following primers and probes were used: VEGF forward,5�-GGCTGCACCCACGACAG-3�; reverse, 5�-CGCTGGTAGACGTCCATGAA-3� probe, 5�-FAM-GAGAGCAGAAGTCCCATGAAGTGATCAA-TAMRA-3�; AUH forward, 5�-CGCTACAAGGGAGAATAGGAGG-3�; reverse, 5�-GCTCTGAACCACTTCCAGCAC-3�. Detection of18S rRNA, used as an internal control, was performed using premixedreagents from Applied Biosystems. Detection of VEGF 18S rRNA wasperformed using TaqMan Universal PCR Master Mix (Applied Biosys-tems), whereas AUH detection was performed using Sybr Green PCR Mas-ter Mix (Applied Biosystems).

Relative quantitation values were expressed as follow: 2(Ctr � Ctt),where C is the value measured in each well, Ct is the mean of the replicatewells run for each sample, Ct is the difference between the mean Ctvalues of the samples in the target wells and those of the endogenouscontrol for the same wells (18S values), and Ctr � Ctt represents thedifference between Ct of the reference sample (medium) and Ct of thetested sample (treatment). Values are expressed as fold increases relative tothe reference sample (medium).

Preparation of nuclear extracts

Nuclear extracts were prepared by modification of a standard protocol, aspreviously described (13). Briefly, cells were washed twice with cold PBS

and pelletted by centrifugation at 1200 rpm for 5 min at 4°C. The cell pelletwas then washed once in a hypotonic buffer (10 mM Tris-HCl (pH 7.5), 1.5mM MgCl2, 10 mM KCl, 1 mM Pefabloc, 2 mM DTT, 2 mM sodiumvanadate, and 4 �g/ml leupeptin, aprotinin, and pepstatin) (Roche Diag-nostics), resuspended in the same buffer, and incubated on ice for 10 min.The cell suspension was subsequently homogenized with 18–20 strokes ina glass Dounce homogenizer. The nucler pellet was obtained after centrif-ugation at 1000 � g for 10 min at 4°C and was resuspended in a hypertonicbuffer (20 mM Tris-HCl (pH 7.5), 1.5 mM MgCl2, 0.42 M KCl, 20%glycerol, 1 mM Pefabloc, 2 mM DTT, 2 mM sodium vanadate, and 4�g/ml leupeptin, aprotinin, and pepstatin) to obtain nuclear extracts. Thenuclear suspension was rotated at 4°C for 30 min, and nuclear debris waspelletted by centrifugation at 15,000 � g for 30 min at 4°C. The superna-tant was used for Western blot analysis.

Western blot analysis

Fifty micrograms of nuclear protein extracted as described above was sep-arated on a 4–20% Tris-glycine gel (Invitrogen, Carlsbad, CA) and elec-troblotted on an Immobilon-P membrane (Invitrogen). Membranes wereblocked for 1 h at room temperature in blocking buffer containing 5%nonfat dry milk (Bio-Rad) in 1� TTBS (20 mM Tris, 150 mM NaCl (pH8.2), and 0.1% Tween 20) (Sigma-Aldrich). Membranes were then incu-bated with anti-HIF-1� mAb (Novus Biologicals, Littleton, CO) diluted1/500 in dilution buffer (1� TTBS/1% milk) overnight at 4°C. An anti-�-actin mAb (Sigma-Aldrich) diluted 1/10,000 in 1� TTBS/5% milk wasused as an internal control for loading. After washing three times in wash-ing buffer (1� TTBS), membranes were incubated for 30 min at roomtemperature with a peroxidase-conjugated goat anti-mouse Ab (Sigma-Al-drich) diluted 1/40,000 in dilution buffer. Membranes were then washedthree times in washing buffer, and chemiluminescence detection was per-formed using an ECL kit from Amersham Pharmacia Biotech (Piscataway,NJ), according to the manufacturer’s protocol.

ResultsHypoxia is a selective inhibitor of MCP-1 production by mouseM�

Initial experiments were performed to compare the expression pat-tern of several �- and �-chemokines by MultiProbe RPA in theANA-1 M� cell line cultured under hypoxic (1% O2) vs normoxic(20% O2) conditions for 18 h. Purified RNA was hybridized witha set of 32P-labeled mouse chemokine-specific antisense RNA

FIGURE 1. Identification by RPA analysis of MCP-1 as a hypoxia-in-hibited gene in M�. Total RNA was isolated from ANA-1 M� cultured for18 h under normoxic (20% O2) or hypoxic conditions (1% O2) and wassubjected to RPA analysis, as detailed in Materials and Methods. Quanti-tation of gene expression changes in hypoxic vs normoxic samples wasobtained by densitometric analysis of the band intensity. The mRNA levelsfor each indicated gene are presented as a bar graph and expressed as foldchanges of the hypoxic sample relative to the corresponding normoxicsample. Data shown are the mean SD of the results obtained in threeindependent experiments, with SD � 10% of the mean.

1683The Journal of Immunology

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

probes and was analyzed by PAGE. As depicted in Fig. 1, hypoxiacaused a strong down-regulation of the constitutive expression ofthe mRNA for the �-chemokine, MCP-1/JE, with the extent ofinhibition ranging from 2.3- to 2.8-fold in three independent de-terminations. In contrast, hypoxia increased the mRNA for another�-chemokine, MIP-1�, by �2.9-fold without affecting the expres-sion of the �-chemokines, MIP-1� and RANTES, or the �-che-mokines, IFN-�-inducible protein-10 (IP-10) and MIP-2, suggest-ing a selective inhibitory activity of hypoxia on MCP-1.

To confirm these results, MCP-1 mRNA was assessed by North-ern blot analysis. As shown in Fig. 2A, ANA-1 cells constitutivelyexpressed MCP-1 mRNA that was down-regulated by exposure to1% O2 for 18 h. Although slight fluctuations in the constitutivelevels were detected, we consistently observed inhibition ofMCP-1 mRNA expression under hypoxic conditions in five inde-pendent experiments (data not shown). Consistent with the RPAdata, MIP-1� mRNA was significantly enhanced by hypoxia, ex-cluding toxic effects (Fig. 2A). MCP-1 mRNA inhibition by hyp-oxia was paralleled by decreased protein release into the superna-tant (Fig. 2B). ANA-1 cells constitutively secreted MCP-1, whichaccumulated in the culture medium, ranging from 350 pg to 390pg/0.5 � 106 cells/ml in three different experiments after 24 h. Theamounts of secreted MCP-1 were decreased by �57% when cellswere exposed to 1% O2 for 24 h (Fig. 2B).

We conclude that hypoxia is a selective inhibitor of MCP-1production by M�.

Hypoxia inhibits MCP-1 induction by IFN-�

MCP-1 is an IFN-�-inducible chemokine (18, 38), and we wereinterested in determining whether hypoxia could inhibit IFN-�stimulatory activity.

ANA-1 M� were incubated for 18 h under normoxic or hypoxicconditions in the presence or absence of optimal concentrations ofIFN-� (18), and Northern blot analysis was performed to evaluateMCP-1 expression. IFN-� increased MCP-1 mRNA by �5-fold,and hypoxia abrogated such response (Fig. 3A). A similar patternof results was consistently observed in five independent experi-ments (data not shown). As a control, we show that hypoxia syn-

ergized with IFN-� in inducing iNOS mRNA expression (Fig. 3A),as reported previously (13), excluding the possibility that hypoxiainduced a deactivated cellular state. Secretion of MCP-1 proteinfollowed the same pattern of mRNA expression (Fig. 3B).

RPA analysis was then performed to establish whether hypoxiainhibitory effects could be exerted on other IFN-�-inducible che-mokines or were selective for MCP-1. As shown in Fig. 3C, inaddition to MCP-1, IFN-� up-regulated the expression of RAN-TES and IP-10, but this response was not affected by hypoxia.

Parallel experiments were conducted with the human monocyticcell line THP-1 to evaluate the response of human cells. THP-1cells did not express constitutive levels of MCP-1 mRNA but re-sponded to IFN-� stimulation with MCP-1 induction (Fig. 4A).Hypoxia effectively inhibited IFN-�-triggered MCP-1 mRNA ex-pression (Fig. 4A) and, accordingly, induction of MCP-1 proteinsecretion (Fig. 4B). Comparable results were observed in threeindependent experiments (data not shown).

Thioglycollate-elicited peritoneal M� (Thio-M�) fromC57BL/6 mice were tested to investigate the effects of hypoxia onfresh M�. MCP-1 mRNA was constitutively expressed in Thio-M�, and stimulation with IFN-� for 18 h resulted in increasedmRNA accumulation, although to a lesser extent than in ANA-1cells (Fig. 4, C and D). A similar pattern of results was consistently

FIGURE 2. Inhibition of constitutive MCP-1 mRNA expression andprotein secretion by hypoxia in ANA-1 M�. A, ANA-1 cells were exposedto either normoxia (Med) or hypoxia (Hy) for 18 h, and total RNA wasassayed for MCP-1 mRNA expression by Northern blot. Blots were se-quentially hybridized with probes specific for MCP-1 and MIP-1� andwere exposed for 48 and 12 h, respectively, to obtain comparable bandintensities. Ethidium bromide staining of the 28S and 18S rRNA is shownas a control of RNA loading. B, ANA-1 cells were cultured for 24 h beforesupernatants were harvested and the amounts of secreted MCP-1 weremeasured by ELISA. Results from one representative experiment of threeperformed, expressed as picograms per 0.5 � 106 cells/ml, are shown.

FIGURE 3. Hypoxia inhibits IFN-�-dependent up-regulation ofMCP-1. A, Total RNA was isolated from ANA-1 M� treated for 18 h with100 IU/ml murine IFN-� under normoxic or hypoxic conditions and wasanalyzed by Northern blot. The blot was sequentially hybridized with theMCP-1/JE and the iNOS cDNAs and was exposed for 36 and 24 h, re-spectively, to obtain comparable band intensities. Ethidium bromide stain-ing of the 28S and 18S rRNA was assessed to ensure that comparableamounts of RNA were loaded in each lane. B, Supernatants from ANA-1cells cultured for 24 h with the indicated stimuli were assayed for secretedMCP-1 by ELISA. Results from one representative experiment, expressedas picograms per 0.5 � 106 cells/ml, are shown. C, Total RNA was ex-tracted from ANA-1 cells stimulated for 18 h with medium or murineIFN-� (100 IU/ml) under normoxic or hypoxic conditions and was sub-jected to RPA analysis, as described in Materials and Methods. TheGAPDH probe was used as a control of RNA loading. Data shown are fromone representative experiment of three performed.

1684 MCP-1 INHIBITION BY HYPOXIA IN MACROPHAGES

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

observed in two independent experiments, with the induction rang-ing from 2.5- to 3-fold over control. Hypoxia suppressed MCP-1basal expression and decreased by 72% mRNA induction byIFN-� (Fig. 4, C and D).

To determine the kinetics of MCP-1 inhibition by hypoxia,ANA-1 M� were cultured under normoxic or hypoxic conditionsin the presence or absence of IFN-� and were tested at differenttimes. As shown in Fig. 5, MCP-1 mRNA up-regulation by IFN-�occurred within 3 h of stimulation and reached plateau levels at24 h. Strong inhibition of IFN-�-induced MCP-1 mRNA by hyp-oxia was observed as early as 6 h after the onset of the culture, andmaximal suppression occurred after 24 h (Fig. 5). Comparable re-sults were observed in three independent experiments (data notshown).

These results demonstrate that hypoxia inhibits MCP-1 induc-tion by IFN-� both in mouse M� and in human monocytic cells.

Hypoxia inhibitory activity is exerted on MCP-1 induction byLPS and reverted by reoxygenation

To establish whether hypoxia antagonized MCP-1 up-regulationby other M� activators, in addition to that elicited by IFN-�, westudied the response to LPS (Fig. 6). An eightfold increase ofMCP-1 mRNA expression over basal levels was observed after 3 hof treatment with LPS, reaching maximal levels after 12 h. Induc-tion by LPS was susceptible to a time-dependent inhibition byhypoxia, although to a lesser extent than the response to IFN-�(Fig. 6A). Accordingly, MCP-1 release by LPS-treated cells wasdecreased by �68% after 24 h of hypoxia (Fig. 6B).

Previous studies have shown that the exposure of hypoxic M�to a reoxygenation period enhanced the effects of hypoxia on cy-

tokine gene expression (32, 39). To evaluate the effects of cellreoxygenation on MCP-1 expression, ANA-1 M� were treatedwith medium alone or supplemented with IFN-� or LPS, and thenthey were exposed to 1% O2 for a period of 12 h followed byre-exposure to normal oxygen tension for additional 12 h (reoxy-genation), or they were cultured under normoxic or hypoxic con-ditions for the entire length of the experiment. As shown in Fig. 7,reoxygenation reverted the inhibition by hypoxia both in the pres-ence or absence of IFN-� or LPS, restoring MCP-1 mRNA ex-pression to levels comparable with those detected in normoxiccells.

Taken together, these findings indicate that the inhibitory activ-ity of hypoxia on MCP-1 is stimulus independent and reversible.

Hypoxia decreases MCP-1 transcription and mRNA stability

Nuclear run-on assays were then performed to determine whetherMCP-1 down-regulation by hypoxia involved changes in genetranscription. Nuclei were isolated from control ANA-1 (Fig. 8,Med) and cells stimulated for 4 and 8 h with hypoxia, IFN-�, orhypoxia plus IFN-�. Fig. 8A shows a representative experiment ofthree performed, which gave comparable results. MCP-1 gene wastranscriptionally active in medium-treated M� and susceptible to a1.5-fold augmentation in response to stimulation with IFN-�. In-cubation of Med- and IFN-�-treated cells under hypoxia for atleast 8 h reduced MCP-1 transcription by �40 and 60%, respec-tively, compared with cells cultured under normoxic conditions(Fig. 8A), indicating that gene transcription is one level of controlof MCP-1 mRNA expression by hypoxia.

The effects of hypoxia on mRNA stability were next tested byassessing MCP-1 mRNA t1/2. ANA-1 M� were incubated for 12 hwith medium alone or supplemented with IFN-� under normoxicor hypoxic conditions, and the rate of MCP-1 transcript decay wasdetermined by evaluating mRNA expression immediately or atvarious time points after addition of the transcriptional inhibitorActD. As depicted in Fig. 8B, MCP-1 mRNA was rapidly de-graded in control cells, with calculated t1/2 of 40 min. IFN-�-

FIGURE 4. Inhibitory effects of hypoxia on MCP-1 expression in hu-man monocytic cells and Thio-M�. THP-1 and Thio-M� were exposed tonormoxia or hypoxia in the presence or absence of 500 U/ml human and100 IU/ml murine IFN-�, respectively. A and C, Representative MCP-1Northern blots performed on total RNA isolated from THP-1 cells (A) andThio-M� (C) after 18 h of culture are shown. B, Supernatants from THP-1cells cultured for 24 h with the indicated stimuli were assayed for secretedMCP-1 by ELISA. Results from one representative experiment, expressedas picograms per 0.5 � 106 cells/ml, are shown. D, Densitometric analysisof the blot shown in C. Relative MCP-1 mRNA levels normalized to therespective amounts of ethidium bromide-stained 18S rRNA are presentedas a bar graph and are expressed in arbitrary units.

FIGURE 5. Kinetics of MCP-1 mRNA inhibition by hypoxia in IFN-�-treated ANA-1 M�. ANA-1 M� were cultured for different time pointsunder normoxic or hypoxic conditions in the presence or absence of 100IU/ml murine IFN-�, and MCP-1 mRNA expression was then assessed. A,Representative Northern blot of IFN-�-treated cells is shown. B, Densito-metric analysis of the blot shown in A. Relative MCP-1 mRNA levelsnormalized to the respective amounts of ethidium bromide-stained 18SrRNA are expressed in arbitrary units.

1685The Journal of Immunology

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

treated M� displayed a greater MCP-1 mRNA stability relative tomedium-treated cells, with the transcript t1/2 increased to 70 min.In contrast, exposure to hypoxia reduced MCP-1 mRNA stability,with a 50% decrease in transcript levels observed after �30 min inmedium- and 40 min in IFN-�-treated cells (Fig. 8B). Consistentand reproducible results were obtained in three independent ex-periments (data not shown), indicating that hypoxia acceleratedMCP-1 mRNA decay.

Messenger RNA degradation rate is positively or negatively reg-ulated by trans-acting mRNA-binding proteins in response to var-ious stimuli and in a cell type-specific manner (17, 29, 40). Havingestablished that hypoxia destabilizes MCP-1 mRNA, it was of in-terest to determine whether this effect was associated with the in-creased expression of any mRNA-destabilizing factor. We foundby real-time PCR that hypoxia induced in ANA-1 M� a 2.1-foldup-regulation of the levels of the mRNA coding for AUH, anRNA-binding protein implicated in the regulation of mRNA decay(41, 42), relative to normoxic conditions (Fig. 8C). Comparableextent of increase was detectable after 6 and 18 h of exposure tohypoxia (Fig. 8C) and was consistently observed in three indepen-dent determinations (data not shown). As a control, we show thathypoxia significantly up-regulated VEGF mRNA expression (Fig.8C), as reported previously in human monocytes (33), causing 4.1-and 2.8-fold increases of constitutive levels at 6 and 18 h,respectively.

We conclude that the inhibitory action of hypoxia on MCP-1expression is controlled by a combination of transcriptional andposttranscriptional mechanisms.

PA and DFX differentially regulate MCP-1 production by M�

Several stimuli, including PA and DFX, share with hypoxia theability to induce gene expression under normal O2 tension (13, 20,26–28, 43). To determine whether hypoxia inhibitory effects onMCP-1 could be mimicked by hypoxia-like stimuli under nor-moxic conditions, MCP-1 mRNA levels and protein secretion wereassessed in ANA-1 M� (Fig. 9, A and B). We found that PA andDFX exerted opposite effects on MCP-1 (Fig. 9, A and B), despitetheir common ability to induce iNOS and MIP-1� (13, 18, 28)(Fig. 9A). DFX reduced both constitutive and IFN-�-inducedMCP-1 mRNA expression (Fig. 9A) and protein release (Fig. 9B),mimicking the effects of hypoxia. Conversely, PA not only did notinhibit, but it even up-regulated MCP-1 mRNA levels in medium-treated cells, slightly enhancing the induction by IFN-� (Fig. 9A).Messenger RNA up-regulation was paralleled by increased MCP-1secretion into the supernatant (Fig. 9B). A similar pattern of resultswas obtained in the human THP-1 cell line (data not shown). Geneinduction by hypoxia is associated with increased expression ofHIF-1� (20, 22, 23, 26). To determine whether PA and DFX dif-ferential effects on MCP-1 correlated with a different regulation ofHIF-1�, HIF-1� protein levels were assessed in parallel experi-ments (Fig. 9C). Western blot analysis revealed barely detectableHIF-1� basal expression, which was unaffected by treatment withIFN-� for 6 h (Fig. 9C). Stimulation with PA or DFX up-regulatedHIF-1� protein, although the effects of PA were less pronouncedcompared with those of DFX (Fig. 9C). Interestingly, PA- or DFX-treated ANA-1 M� expressed significantly higher levels ofHIF-1� in the presence of IFN-� (Fig. 9C). A similar pattern of

FIGURE 6. Inhibition of LPS-induced MCP-1 production by hypoxia.ANA-1 M� were cultured for different time points under normoxic orhypoxic conditions in the presence or absence of 10 ng/ml LPS beforeMCP-1 mRNA expression and protein secretion were analyzed. A, Densi-tometric analysis of MCP-1 mRNA levels assessed by Northern blot. Rel-ative amounts of MCP-1 mRNA normalized to the respective amounts ofethidium bromide-stained 18S rRNA are presented as a bar graph and areexpressed in arbitrary units. Data shown are the mean SD of the resultsobtained in two different hybridization experiments, with SD � 10% of themean. B, Supernatants from ANA-1 cells cultured for 24 h with the indi-cated stimuli were assayed for secreted MCP-1 by ELISA. Results fromone representative experiment of two performed, expressed as picogramsper 0.5 � 106 cells/ml, are shown.

FIGURE 7. Reoxygenation restores MCP-1 mRNA expression in hy-poxic M�. Total RNA from medium-, IFN-�-, or LPS-treated ANA-1 M�,cultured for 12 h in the presence of 1% O2 (Hy, �) followed by additional12 h in the presence of normal oxygen levels (Reox, �) or exposed tohypoxic (�) or normoxic (�) conditions for 24 h, was analyzed for MCP-1expression. A, Representative Northern blot is shown. Ethidium bromidestaining of the 28S and 18S rRNA was assessed as a control of RNAloading. B, Densitometric analysis of MCP-1 mRNA expression. MCP-1mRNA levels in hypoxic (Hy) or reoxygenated (Reox) M� are expressedas a percentage of the amounts detected in cells cultured under normoxicconditions (Normo, 100%). Data shown are the mean SD of the resultsobtained in three separate hybridization experiments, with SD � 10% ofthe mean.

1686 MCP-1 INHIBITION BY HYPOXIA IN MACROPHAGES

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

results was obtained after 12 h of culture (data not shown), dem-onstrating that HIF-1� expression can be stimulated under nor-moxic conditions by PA or DFX in M� and that IFN-� can actsynergistically with either stimulus in inducing this factor.

These findings indicate the existence of a differential regulationof MCP-1 by hypoxia-like stimuli, which seems dissociated fromtheir activity on HIF-1� expression.

DiscussionThe network of physiopathological stimuli regulating MCP-1 ex-pression in M� is not fully understood and it extends beyond theclassical microbial or immune system-derived signals. In thisstudy, we demonstrate that hypoxia is a potent and selective in-hibitor of MCP-1 production by both resting and activated M�.Moreover, we show that inhibition of MCP-1 expression by hyp-oxia is the result of decreased gene transcription and reducedmRNA stability.

The influence of altered O2 concentrations on MCP-1 expres-sion has been previously demonstrated in other cell types, both invitro and in vivo. Increased MCP-1 mRNA and protein expressionwas reported in human dermal fibroblasts exposed to hypoxia (44),and human melanoma cell lines were shown to up-regulate MCP-1mRNA and protein under anoxia/reoxygenation (45). Moreover,stimulatory effects of hypoxia on MCP-1 have been observed inischemic rat neurons and myocardial cells (46, 47). Conversely, weconsistently observed down-regulation of MCP-1 mRNA expres-sion and decreased protein release by hypoxia both in the mouseM� cell line ANA-1 and in THP-1 human monocytic cells, dem-onstrating that hypoxia is active on different types of mononuclearphagocytes and across species in inhibiting MCP-1 expression.Furthermore, hypoxia inhibitory effects were exerted on primarythioglycollate-induced peritoneal M�, supporting the general va-lidity of the observation. Inhibition of MCP-1 expression has beenreported previously in TNF-�-stimulated human ovarian carci-noma cell lines cultured under hypoxia/anoxia (48), showing thatcells other than M� are susceptible to the inhibitory effects of lowO2 tension on MCP-1. Therefore, it is conceivable that expressionof MCP-1 is differentially regulated by hypoxia depending on thelineage of the target cell, similarly to what has been reported forother hypoxia-regulated genes (34, 49). Interestingly, cell-type-specific modulation of MCP-1 expression has also been observedin response to other stimuli (16).

M� produce several chemokines in addition to MCP-1 (2, 4, 7,17, 18), and MCP-1 expression can be induced in M� by variousproinflammatory stimuli, the prototypes of which are IFN-� andLPS, as previously shown (4, 7, 16, 18, 38) and confirmed here.Hence, we considered the potential selectivity and stimulus spec-ificity of the response to hypoxia. Interestingly, we found that hyp-oxia inhibitory effects were selective for MCP-1, because the ex-pression of other �- and �-chemokines, such as MIP-1�,RANTES, IP-10, and MIP-2, was not affected, and MIP-1� wasinduced. Importantly, moreover, hypoxia not only down-regulatedMCP-1 constitutive expression, but it also counteracted MCP-1

vested at 1, 2, 4, and 6 h and was examined for MCP-1 expression byNorthern blot. Data are presented as the relative amounts of MCP-1 mRNAremaining after addition of ActD and after normalization of the bands tothe respective amounts of ethidium bromide-stained 18S rRNA. C, TotalRNA was harvested after 6 and 18 h, reverse transcribed, and tested forAUH and VEGF mRNA expression by real-time PCR, as detailed in Ma-terials and Methods. 18S rRNA was tested in parallel as an internal controlfor input RNA. Results are expressed as fold increase relative to mRNAlevels under normoxic (Med) conditions (equal to 1) and are the average ofthree independent experiments

FIGURE 8. Inhibitory activity of hypoxia on MCP-1 gene transcriptionand mRNA half-life. ANA-1 M� were treated with medium alone or con-taining murine IFN-� (100 IU/ml) under normoxic or hypoxic conditions.A, MCP-1 gene transcription was determined by nuclear run-on analysis onnuclei isolated after 4 and 8 h of culture, as detailed in Materials andMethods. Upper panel, Representative run on assay is shown. Lower panel,Quantitative analysis of MCP-1 gene transcription. Relative transcriptionactivities normalized to �-actin are presented as a bar graph and are ex-pressed in arbitrary units. Data shown are the mean SD of the resultsobtained in three independent experiments, with SD � 10% of the mean.B, ActD (5 �g/ml) was added after 12 h of culture. Total RNA was har-

1687The Journal of Immunology

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

induction by IFN-�, without affecting IFN-�-dependent up-regu-lation of RANTES and IP-10. MCP-1 stimulation by LPS wassimilarly susceptible to inhibition by hypoxia, although to a lesserextent than was the response to IFN-�. These findings suggest thathypoxia can selectively modulate chemokine expression in M�and exert a targeted inhibition of MCP-1 production induced bydifferent M�-activating stimuli. Our recently reported observationthat hypoxia stimulatory activity on MIP-1� in M� is inhibited byIFN-� (36) further emphasizes the tight regulatory network con-necting hypoxia, IFN-�, and chemokine expression in these cells.We have previously shown that hypoxia acts synergistically withIFN-� or LPS in triggering iNOS gene expression and enzymaticactivity in murine M� (13, 15).4 Thus, it appears that hypoxia can

selectively synergize with or antagonize the stimulatory effects ofproinflammatory mediators on M� gene expression. The concertedaction of hypoxia with immune-derived or microbial factors islikely to play an important role in coordinating M� activity inpathological states, in that these stimuli are concomitantly presentat sites of inflammation and/or tumor growth (10, 20, 21).

MCP-1 expression levels were dependent not only on the O2

concentration, but also on the extent and duration of the hypoxicexposure. In fact, we observed a time-dependent inhibition ofMCP-1 induction by IFN-� or LPS in ANA-1 M�, which occurredalready within 6 h of exposure to hypoxia and was maximal after24 h. Furthermore, cell reoxygenation reverted hypoxia inhibitoryeffects within 24 h, restoring MCP-1 mRNA expression to levelscomparable with those detected in normoxic cells. Our results sug-gest that the pattern of MCP-1 expression in M� in vivo may varydynamically with the degree of local oxygenation, which is quiteheterogeneous and rapidly fluctuating within pathological lesions(50), thus resulting in shifting gradients of secreted MCP-1.

Transcriptional modulation is the primary mechanism of regu-lation of MCP-1 expression, which is dependent on multiple tran-scription factors and exhibits both cell type- and stimulus-specificpatterns (16, 38, 51, 52). Here, we show that MCP-1 gene wastranscriptionally active in the ANA-1 M� cell line and that hyp-oxia decreased its transcriptional activity. Moreover, the exposureof IFN-�-treated cells to hypoxia resulted in a strong reduction ofIFN-�-dependent MCP-1 transcriptional increase. These findingssuggest that inhibition of transcription represents at least one of themechanisms by which hypoxia down-regulates MCP-1 expressionin M�, thus raising the issue of the regulatory pathway(s) in-volved. It is reasonable to suggest that one or more hypoxia-in-ducible transcription repressor(s) might target specific elements inthe MCP-1 promoter, thereby decreasing gene transcription, sim-ilarly to what has been recently reported for other genes (49, 53).Studies are currently underway to determine whether any of theregulatory elements previously identified in the MCP-1 5�-regula-tory region (38, 51, 52) or novel cis-acting sequences are the targetof hypoxia inhibitory effects. Interestingly, we found by computersearch of the published MCP-1 promoter sequence (GenBank ac-cession number U12470) a region of homology with the HRE(hereafter referred to as MCP-HRE) present in the promoter ofmany hypoxia-inducible genes (20, 22, 23), comprizing both anHIF-1 binding site (sequence GGCGTGGT; �1639/�1632) and adownstream HIF-1 ancillary sequence (sequence CACGC; �1626/�1622) required for the interaction with additional transcriptionfactors (54). At present, the role of MCP-HRE in the regulation ofMCP-1 expression is not clear, because MCP-1 transcriptional in-duction by hypoxia in human melanoma cell lines and dermal fi-broblasts was shown to be dependent on the activation of NF-�Band Sp-1 binding motifs (44, 45). Experiments are underway in ourlaboratory to determine whether MCP-1-HRE is active in the con-trol of MCP-1 transcription by hypoxia in M�.

Signals other than hypoxia can activate the HIF-1/HRE pathwayunder normoxic conditions and can mimic hypoxia in the inductionof gene expression (20, 22, 26, 27). We observed increasedHIF-1� protein levels in ANA-1 M� after stimulation with thehypoxia-like stimuli PA and DFX. This effect can probably beascribed to the iron-chelating property of these agents, becausechelation of iron was shown to affect an important biochemicalstep in the HIF-1� ubiquitin-proteosomal degradation pathway(26). These findings are consistent with our early reports showingthat PA or DFX can induce HRE-binding activity and activate theiNOS-HRE in mouse macrophages (13, 28). Induction of HIF-1�did not correlate with the regulation of MCP-1 expression, which

FIGURE 9. Differential regulation of MCP-1 by PA and DFX. ANA-1M� were stimulated with 4 mM PA or 400 �M DFX, in the presence orabsence of 100 IU/ml murine IFN-�. A, Total RNA was isolated after 18 h andwas assayed for MCP-1 mRNA expression by Northern blot. The blot wassequentially hybridized with MCP-1/JE, MIP-1�, and iNOS cDNAs. To ob-tain comparable band intensities, a 24-h exposure was required after hybrid-ization with MCP-1/JE and iNOS, whereas 8 h of exposure were sufficient forblots hybridized with MIP-1�. B, Supernatants were assayed after 24 h oftreatment for secreted MCP-1 by ELISA. Results from one representative ex-periment are presented as a bar graph and are expressed as picograms per0.5 � 106 cells/ml. C, Nuclear extracts (50 �g) isolated after 6 h of treatmentwere resolved by SDS-PAGE, and HIF-1� protein levels were assessed byWestern blot analysis, as described in Materials and Methods. �-Actin ex-pression was evaluated in parallel to control for protein loading. Results shownare from one representative experiment of two performed.

1688 MCP-1 INHIBITION BY HYPOXIA IN MACROPHAGES

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

was inhibited by DFX and increased by PA. These results dem-onstrate for the first time that PA or DFX can differentially controlthe expression of selected genes, suggesting that unique molecularevents can be triggered by either stimulus in addition to HIF-1�expression. Recent evidence has been provided that a variety ofgrowth factors and cytokines can induce HIF-1� accumulation incertain cell types (26, 27). Interestingly, we identified a novel syn-ergism between IFN-� and PA or DFX in the induction of HIF-1�,which probably in part accounts for the ability of the stimuli tocooperate in the activation of HRE-dependent iNOS transcription(13, 28), thus providing the first evidence that IFN-� can affect theexpression of this NF. Whether IFN-� reduces HIF-1� degradationor increases its transcriptional activation or translation will be ad-dressed experimentally in future studies. The synergism betweenPA and IFN-� may be potentially relevant for the expression ofvarious HIF-1-inducible genes under inflammatory conditions inwhich elevated levels of tryptophan metabolites have been de-tected (9, 13, 18, 55). In contrast, the cooperation between DFXand IFN-� may have practical clinical implications, because DFXis used as a therapeutic agent for the treatment of several patho-logical states in which IFN-� is produced (18, 28).

Posttranscriptional regulation of mRNA stability is another levelof control of gene expression by hypoxia, as shown in the case ofa number of hypoxia-regulated genes including VEGF, tyrosinehydroxylase, glucose transporter-1, erytropoietin, and eNOS (29,56–58). We found that MCP-1 inhibition by hypoxia was alsoassociated with destabilization of the mRNA, which decayed at afaster rate in hypoxic than in normoxic M�. These results areconsistent with the finding that the MCP-1 mRNA contains in its3� untranslated region an AU-rich destabilizing element (ARE)(59) present in the 3� untranslated region of many labile mamma-lian transcripts coding for cytokines, inflammatory mediators, andoncoproteins and serving as a signal for rapid mRNA degradation(17, 40). The regulation of mRNA turnover by AREs involvestheir association with any of a growing number of stabilizing ordestabilizing trans-acting mRNA-binding proteins with AU spec-ificity (17, 40, 60). Although a few ARE-binding factors mediatingmRNA stabilization by hypoxia have been identified (29, 56, 57),no information is currently available on hypoxia-inducible desta-bilizing proteins. We demonstrate that hypoxia increases the ex-pression of the mRNA for AUH, an AU-binding factor homolo-gous to a 32-kDa protein suggested to be implicated in ARE-directed mRNA decay (41, 42, 60), providing the first evidencethat AUH is a hypoxia-inducible gene. The possibility that AUHmay contribute to MCP-1 mRNA degradation in M� exposed tolow O2 tension is intriguing and under investigation.

In conclusion, in this study we have characterized a novel prop-erty of hypoxia as a selective inhibitor of MCP-1 production byM� and we have demonstrated the existence of a cross-talk amongenvironmental signals (hypoxia), stimuli derived from the immunesystem (IFN-�), and microbial factors (LPS) in the regulation ofM�-derived MCP-1. Given the central role of MCP-1 in the re-cruitment of various leukocyte populations, the modulation of itsexpression is a critical set point for the control of the kinetics andcomposition of the cellular infiltrate under various inflammatoryconditions and at tumor sites (4–6, 16, 61). Low O2 levels havebeen described in many pathological conditions characterized byM� infiltration and inflammatory responses (10, 20, 21, 29, 62).Thus, MCP-1 inhibition by hypoxia in M� is likely to be of patho-physiologic relevance and to represent a negative regulatory mech-anism to control M�-mediated leukocyte recruitment in patholog-ical tissues. MCP-1 inhibition may be particularly important atsites of inflammation where prolonged or excessive MCP-1 secre-tion represents a detrimental factor for the control and extinction of

an inflammatory response, which could favor the onset of chronicinflammatory disorders (3, 4, 47, 63). In contrast, inhibition ofMCP-1 production at sites of tumor growth may potentially con-tribute to tumor progression. In fact, MCP-1 was reported to dif-ferently influence the rate of tumor growth depending on its ex-pression levels, with high levels leading to tumor rejection (4, 61,64) and low levels supporting tumor growth (65). These and pre-vious findings showing impaired M� migratory ability to MCP-1under conditions of low O2 tension (66) indicate that hypoxia playsan immunosuppressive role on MCP-1 chemokine/receptor net-work in M�, which may in part explain the functional impairmentof tumor-associated macrophages demonstrated both in tumor-bearing mice and in cancer patients (10, 19, 67). It is important toemphasize, however, that hypoxia can exert a tight regulatory con-trol on M� production of several proinflammatory cytokines, mod-ulating the response to immune-derived and microbial stimuli (13,15, 29, 30, 32, 34–36).4 Thus, the activation state of hypoxic M�will be ultimately determined by the balance between hypoxia in-hibitory and stimulatory activities, which will establish the tem-poral and quantitative relationship of the various cytokines in thetissue microenvironment. Ongoing microarray experiments in-tended to fully characterize the profile of gene expression in hy-poxic M� will hopefully lead to a better understanding of the roleof hypoxia in shaping the functional profile of these cells.

AcknowledgmentsWe thank Dr. Howard Young for RPA analysis and Chantal Dabizzi forsecretarial assistance.

References1. Ben-Baruk, A., D. Michiel, and J. Oppenheim. 1995. Signals and receptor in-

volved in the recruitment of inflammatory cells. J. Biol. Chem. 270:11703.2. Baggiolini, M., B. Dewald, and B. Moser. 1994. Interleukin 8 and related che-

motactic cytokines: CXC and CC chemokines. Adv. Immunol. 55:97.3. Murphy, P. 1996. Chemokine receptors: structure, functions and role in microbial

pathogenesis. Cytokine Growth Factor Rev. 7:47.4. Matsushima, K., E. T. Baldwin, and N. Mukaida. 1992. Interleukin-8 and MCAF:

novel leukocyte recruitment and activating cytokines. Chem. Immunol. 51:236.5. Fuentes, N. E., S. K. Durham, M. R. Swerdel, A. C. Lewin, D. S. Barton,

J. R. Megill, R. Bravo, and S. A. Lira. 1995. Controlled recruitment of monocytesand macrophages to specific organs through transgenic expression of monocytechemoattractant protein-1. J. Immunol. 155:5769.

6. Carr, M. W., S. J. Roth, E. Luther, S. S. Rose, and T. A. Springer. 1994. Mono-cyte chemoattractant protein 1 acts as a T-lymphocyte chemoattractant. Proc.Natl. Acad. Sci. USA 91:3652.

7. Kopydlowski, K. M., C. A. Salkowski, M. J. Cody, N. Van Rooijen, J. Major,T. A. Hamilton, and S. N. Vogel. 1999. Regulation of macrophage chemokineexpression by lipopolysaccharide in vitro and in vivo. J. Immunol. 163:1537.

8. Adams, D. O., and T. A. Hamilton. 1992. Macrophages as destructive cells inhost defense. In Inflammation: Basic Principles and Clinical Correlates.J. I. Gallin, I. M. Goldstein, and R. Snyderman, eds. Raven, New York, p. 637.

9. Varesio, L., I. Espinoza-Delgado, G. Gusella, G. Cox, G. Melillo, T. Musso, andM. C. Bosco. 1995. Role of cytokines in the activation of monocytes. In HumanCytokines: Their Role in Disease and Therapy. B. B. Aggarwal and R. K. Puri,eds. Blackwell Scientific Publications, Cambridge, MA, p. 55.

10. Sica, A., A. Saccani, and A. Mantovani. 2002. Tumor-associated macrophages: amolecular perspective. Int. Immunopharmacol. 2:1045.

11. Paulnock, D. M., K. P. Demick, and S. P. Coller. 2000. Analysis of interferon-�-dependent and -independent pathways of macrophage activation. J. LeukocyteBiol. 67:677.

12. Hamilton, T. A., Y. Ohmori, J. M. Tebo, and R. Kishore. 1999. Regulation ofmacrophage gene expression by pro- and anti-inflammatory cytokines. Pathobi-ology 67:241.

13. Melillo, G., T. Musso, A. Sica, L. S. Taylor, G. Cox, and L. Varesio. 1995. Ahypoxia responsive element mediates a novel pathway of activation of the in-ducible nitric oxide synthase promoter. J. Exp. Med. 182:1683.

14. Shi, M. M., J. J. Godleski, and J. D. Paulauskis. 1996. Regulation of macrophageinflammatory protein-1� mRNA by oxidative stress. J. Biol. Chem. 271:5878.

15. Melillo, G., L. S. Taylor, A. Brooks, G. Cox, and L. Varesio. 1996. Regulationof inducible nitric oxide synthase expression in IFN-�-treated murine macro-phages cultured under hypoxic conditions. J. Immunol. 157:2638.

16. Colotta, F., A. Borre, J. M. Wang, M. Tattanelli, F. Maddalena, N. Polentarutti,G. Peri, and A. Mantovani. 1992. Expression of a monocyte chemotactic cytokineby human mononuclear phagocytes. J. Immunol. 148:765.

17. Bosco, M. C., G. Gusella, I. Espinoza-Delgado, D. Longo, and L. Varesio. 1994.Interferon-� upregulates interleukin-8 gene expression in human monocytic cellsby a posttranscriptional mechanism. Blood 83:537.

1689The Journal of Immunology

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from

18. Bosco, M. C., A. Rapisarda, S. Massazza, G. Melillo, H. Young, and L. Varesio.2000. The tryptophan catabolite picolinic acid selectively induces the chemokinesmacrophage-inflammatory protein-1� and -1� in macrophages. J. Immunol.164:3283.

19. Elgert, K. D., D. G. Alleva, and D. W. Mullins. 1998. Tumor-induced immunedysfunction: the macrophage connection. J. Leukocyte Biol. 64:275.

20. Semenza, G. L. 2001. Hypoxia-inducible factor 1: oxygen homeostasis and dis-ease pathophysiology. Trends Mol. Med. 7:345.

21. Harris, A. L. 2002. Hypoxia: a key regulatory factor in tumour growth. Nat. Rev.Cancer 2:38.

22. Wenger, R. H. 2002. Cellular adaptation to hypoxia: O2-sensing protein hydroxy-lases, hypoxia-inducible transcription factors, and O2-regulated gene expression.FASEB J. 16:1151.

23. Ratcliffe, P. J., J. O’Rourke, P. H. Maxwell, and C. W. Pugh. 1998. Oxygensensing, hypoxia-inducible factor-1 and the regulation of mammalian gene ex-pression. J. Exp. Biol. 201:1153.

24. Ebert, B. L., J. M. Gleadle, J. F. O’Rourke, S. M. Bartlett, J. Poulton, andP. J. Ratcliffe. 1996. Isoenzyme-specific regulation of genes involved in energymetabolism by hypoxia: similarities with the regulation of erythropoietin. Bio-chem. J. 313:809.

25. Talks, K. L., H. Turley, K. C. Gatter, P. H. Maxwell, C. W. Pugh, andP. J. Ratcliffe. 2000. The expression and distribution of the hypoxia-induciblefactors HIF-1� and HIF-2� in normal human tissues, cancers, and tumor-asso-ciated macrophages. Am. J. Pathol. 157:411.

26. Huang, L. E., and H. F. Bunn. 2003. Hypoxia-inducible factor and its biomedicalrelevance. J. Biol. Chem. 278:19575.

27. Richard, D. E., E. Berra, and J. Pouyssegur. 2000. Nonhypoxic pathway mediatesthe induction of hypoxia-inducible factor 1� in vascular smooth muscle cells.J. Biol. Chem. 275:26765.

28. Melillo, G., L. S. Taylor, A. Brooks, T. Musso, G. Cox, and L. Varesio. 1996.Functional requirement of the hypoxia responsive element in the activation of theinducible nitric oxide synthase promoter by the iron chelator desferrioxamine.J. Biol. Chem. 272:12236.

29. Lewis, J. S., J. A. Lee, J. C. E. Underwood, A. L. Harris, and C. E. Lewis. 1999.Macrophage responses to hypoxia: relevance to disease mechanisms. J. Leuko-cyte Biol. 66:889.

30. Cramer, T., Y. Yamanishi, B. E. Clausen, I. Forster, R. Pawlinski, N. Mackman,V. H. Haase, R. Jaenisch, M. Corr, V. Nizet, et al. 2003. HIF-1� is essential formyeloid cell-mediated inflammation. Cell 112:645.

31. Kuwabara, K., S. Ogawa, M. Matsumoto, S. Koga, M. Clauss, D. J. Pinsky,P. Lyn, J. Leavy, L. Witte, J. Joseph-Silverstein, et al. 1995. Hypoxia-mediatedinduction of acidic/basic fibroblast growth factor and platelet-derived growthfactor in mononuclear phagocytes stimulates growth of hypoxic endothelial cells.Proc. Natl. Acad. Sci. USA 92:4606.

32. Koga, S., S. Ogawa, K. Kuwabara, J. Brett, J. Leavy, J. Ryan, Y. Koga,J. Plocinski, W. Benjamin, and D. K. Burns. 1992. Synthesis and release ofinterleukin-1 by reoxygenated human mononuclear phagocytes. J. Clin. Invest.90:1007.

33. Melillo, G., E. A. Sausville, K. Cloud, T. Lahusen, L. Varesio, andA. M. Senderovicz. 1999. Flavopiridol, a protein kinase inhibitor, down-regulateshypoxic induction of vascular endothelial growth factor expression in humanmonocytes. Cancer Res. 59:5433.

34. Guida, E., and A. Stewart. 1998. Influence of hypoxia and glucose deprivation ontumor necrosis factor-� and granulocyte-macrophage colony-stimulating factorexpression in human cultured monocytes. Cell. Physiol. Biochem. 8:75.

35. Hirani, N., F. Antonicelli, R. M. Strieter, M. S. Wiesener, C. Haslett, andS. C. Donnelly. 2001. The regulation of interleukin-8 by hypoxia in human mac-rophages: a potential role in the pathogenesis of the acute respiratory distresssyndrome (ARDS). Mol. Med. 7:685.

36. Carta, L., S. Pastorino, G. Melillo, M. C. Bosco, S. Massazza, and L. Varesio.2001. Engineering of macrophages to produce IFN� in response to hypoxia.J. Immunol. 166:5374.

37. Rapisarda, A., B. Uranchimeg, D. A. Scudiero, M. Selby, E. A. Sausville,R. H. Shoemaker, and G. Melillo. 2002. Identification of small molecule inhib-itors of hypoxia-inducibile factor 1 transcriptional activation pathway. CancerRes. 62:4316.

38. Zhou, Z.-H. L., P. Chaturvedi, Y. L. Han, S. Aras, Y. S. Li, P. Kolattukudy,D. Ping, J. Boss, and R. Ransohoff. 1998. IFN-� induction of the human mono-cyte chemoattractant protein (hMCP)-1 gene in astrocytoma cells: functional in-teraction between an IFN-�-activated site and a GC-rich element. J. Immunol.160:3908.

39. Maeda, B. Y., M. Matsumoto, O. Hori, K. Kuwabara, S. Ogawa, S. D. Yan,T. Ohtsuki, T. Kinoshita, T. Kamada, and D. M. Stern. 1994. Hypoxia/reoxy-genation-mediated induction of astrocyte interleukin-6: a paracrine mechanismpotentially enhancing neuron survival. J. Exp. Med. 180:2297.

40. Chen, C. Y., and A. B. Shyu. 1995. AU-rich elements: characterization and im-portance in mRNA degradation. Trends Biochem. Sci. 20:465.

41. Nakagawa, J., H. Waldner, S. Meyer-Monard, J. Hofsteenge, P. Jeno, andC. Moroni. 1995. AUH, a gene encoding an AU-specific RNA binding proteinwith intrinsic enoyl-CoA hydratase activity. Proc. Natl. Acad. Sci. USA 92:2051.

42. Vakalopoulou, E., J. Schaack, and T. Shenk. 1991. A 32-kilodalton protein bindsto AU-rich domains in the 3� untranslated regions of rapidly degraded mRNAs.Mol. Cell. Biol. 11:3355.

43. Rapella, A., A. Negrioli, G. Melillo, S. Pastorino, L. Varesio, and M. C. Bosco.2002. Flavopiridol inhibits vascular endothelial growth factor production inducedby hypoxia or picolinic acid in human neuroblastoma. Int. J. Cancer 99:658.

44. Galindo, M., B. Santiago, J. Alcami, M. Rivero, J. Martin-Serrando, andJ. L. Pablos. 2001. Hypoxia induces expression of the chemokines monocytechemoattractant protein-1 (MCP-1) and IL-8 in human dermal fibroblasts. Clin.Exp. Immunol. 123:36.

45. Kunz, M., G. Bloss, R. Gillitzer, G. Gross, M. Goebeler, U. R. Rapp, andS. Ludwig. 2002. Hypoxia/reoxygenation induction of monocyte chemoattractantprotein-1 in melanoma cells: involvement of nuclear factor-�B, stimulatory pro-tein-1 transcription factor and mitogen-activated protein kinase pathways. Bio-chem. J. 366:299.

46. Ivacko, J., J. Szaflarski, C. Malinak, C. Flory, J. S. Warren, and F. S. Silverstein.1997. Hypoxic-ischemic injury induces monocyte chemoattractant protein-1 ex-pression in neonatal rat brain. J. Cereb. Blood Flow Metab. 17:759.

47. Ono, K., A. Matsumori, Y. Furukawa, H. Igata, T. Shioi, K. Matsushima, andS. Sasayama. 1999. Prevention of myocardial reperfusion injury in rats by anantibody against monocyte chemotactic and activating factor/monocyte chemoat-tractant protein-1. Lab. Invest. 79:195.

48. Negus, R. P. M., L. Turner, F. Burke, and F. R. Balkwill. 1998. Hypoxia down-regulated MCP-1 expression: implications for macrophage distribution in tumors.J. Leukocyte Biol. 63:758.

49. Kitamuro, T., K. Takahashi, K. Ogawa, R. Udono-Fujimori, K. Takeda,K. Furujama, M. Nakayama, J. Sun, H. Fujita, W. Hida, et al. 2003. Bach 1functions as a hypoxia-inducible repressor for the heme oxygenase-1 gene inhuman cells. J. Biol. Chem. 278:9125.

50. Evans, S. M., S. M. Hahn, D. P. Magarelli, and C. J. Koch. 2001. Hypoxicheterogeneity in human tumors: EF5 binding, vasculature, necrosis, and prolif-eration. Am. J. Clin. Oncol. 24:467.

51. Freter, R. P., J. A. Alberta, G. Y. Hwang, A. L. Wrentmore, and C. D. Stiles.1996. Platelet-derived growth factor induction of the immediate-early geneMCP-1 is mediated by NF-�B and a 90-KDa phosphoprotein coactivator. J. Biol.Chem. 271:17417.

52. Ueda, A., Y. Ishigatsubo, T. Okubo, and T. Yoshimura. 1997. Transcriptionalregulation of the human monocyte chemoattractant protein-1 gene. J. Biol. Chem.272:31092.

53. Michelotti, G. A., M. J. Bauman, M. P. Smith, and D. A. Schwinn. 2003. Cloningand characterization of the rat �1a-adrenergic receptor gene promoter. J. Biol.Chem. 278:8693.

54. Kimura, H., A. Weisz, T. Ogura, Y. Hitomi, Y. Kurashima, K. Hashimoto, F.D’Acquisto, M. Makuuchi, and H. Esumi. 2001. Identification of hypoxia-induc-ible factor-1 ancillary sequence and its function in vascular endothelial growthfactor gene induction by hypoxia and nitric oxide. J. Biol. Chem. 276:2292.

55. Rapisarda, A., S. Pastorino, S. Massazza, L. Varesio, and M. C. Bosco. 2002.Antagonistic effect of picolinic acid and interferon-� on macrophage inflamma-tory protein-1�/� production. Cell. Immunol. 220:70.

56. Levy, N. S., S. Chung, H. Furneaux, and A. P. Levy. 1998. Hypoxic stabilizationof vascular endothelial growth factor mRNA by the RNA-binding protein HuR.J. Biol. Chem. 273:6417.

57. McGary, E. C., I. J. Rondon, and B. S. Beckman. 1997. Post-transcriptionalregulation of erythropoietin mRNA stability by erythropoietin mRNA-bindingprotein. J. Biol. Chem. 272:8628.

58. McQuillan, L. P., G. K. Leung, P. A. Marsden, S. K. Kostyk, andS. Kourembanas. 1994. Hypoxia inhibits expression of eNOS via transcriptionaland posttranscriptional mechanisms. Am. J. Physiol. 267:1921.

59. Koerner, T. J., T. Hamilton, M. Introna, C. Tannenbaum, R. Bast, and D. Adams.1987. The early competence genes JE and KC are differentially regulated inmurine peritoneal macrophages in respone to lipopolysaccharide. Biochem. Bio-phys. Res. Commun. 149:969.

60. Myer, V. E., X. C. Fan, and J. A. Steitz. 1997. Identification of HuR as a proteinimplicated in AUUUA-mediated mRNA decay. EMBO J. 16:2130.

61. Zhang, L., A. Khayat, H. Cheng, and D. T. Graves. 1997. The pattern of mono-cyte recruitment in tumors is modulated by MCP-1 expression and influences therate of tumor growth. Lab. Invest. 76:579.

62. Semenza, G. L. 2002. HIF-1 and tumor progression: pathophysiology and ther-apeutics. Trends Mol. Med. 8:S62.

63. Koch, A. E., S. L. Kunkel, L. A. Harlow, B. Johnson, H. L. Evanoff,G. K. Haines, M. D. Burdick, R. M. Pope, and R. M. Strieter. 1992. Enhancedproduction of monocyte chemoattractant protein-1 in rheumatoid arthritis. J. Clin.Invest. 90:772.

64. Manome, Y., P. Y. Wen, A. Hershowitz, T. Tanaka, B. D. Rollins, D. W. Kufe,and H. A. Fine. 1995. Monocyte chemoattractant protein-1 (MCP-1) gene trans-duction: an effective tumor vaccine strategy for non-intracranial tumors. CancerImmunol. Immunother. 41:227.

65. Nesbit, A., H. Schaider, T. H. Miller, and M. Herlyn. 2001. Low-level monocytechemoattractant protein-1 stimulation of monocytes leads to tumor formation innon-tumorigenic melanoma cells. J. Immunol. 166:6483.

66. Turner, L., C. Scotton, R. Negus, and F. Balkwill. 1999. Hypoxia inhibits mac-rophage migration. Eur. J. Immunol. 29:2280.

67. Bingle, L., N. J. Brown, and C. E. Lewis. 2002. The role of tumor-associatedmacrophages in tumor progression: implication for new anticancer therapies.J. Pathol. 196:254.

1690 MCP-1 INHIBITION BY HYPOXIA IN MACROPHAGES

by guest on July 2, 2015http://w

ww

.jimm

unol.org/D

ownloaded from