Nucleotidase) Up-Regulation ′ CD73

Home , Nucleotidase

IFN-α Induced Adenosine Production on the Endothelium: A Mechanism Mediated by CD73 (Ecto-5′-Nucleotidase) Up-Regulation

This information is current as Jussi Niemelä, Tiina Henttinen, Gennady G. Yegutkin, Laura of September 29, 2021. Airas, Anna-Maija Kujari, Pertti Rajala and Sirpa Jalkanen J Immunol 2004; 172:1646-1653; ; doi: 10.4049/jimmunol.172.3.1646 http://www.jimmunol.org/content/172/3/1646 Downloaded from

References This article cites 41 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/172/3/1646.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 29, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

IFN-␣ Induced Adenosine Production on the Endothelium: (A Mechanism Mediated by CD73 (Ecto-5؅-Nucleotidase Up-Regulation1

Jussi Niemela¨,* Tiina Henttinen,* Gennady G. Yegutkin,* Laura Airas,* Anna-Maija Kujari,* Pertti Rajala,† and Sirpa Jalkanen2*

CD73 (ecto-5؅-nucleotidase; EC 3.1.3.5) participates in lymphocyte binding to endothelial cells and converts extracellular AMP into a potent anti-inflammatory substance adenosine. However, the regulation of expression and function of CD73 has remained largely unknown. In this study, we show that IFN-␣ produces a time- and dose-dependent long-term up-regulation of CD73 on endothelial cells, but not on lymphocytes both at protein and RNA levels. Moreover, CD73-mediated production of adenosine is increased after IFN-␣ treatment on endothelial cells, resulting in a decrease in the permeability of these cells. Subsequent to Downloaded from induction with PMA, FMLP, dibutyryl cAMP, thrombin, histamine, IL-1␤, TNF-␣, and LPS, no marked changes in the level of CD73 expression on endothelial cells are observed. We also show that CD73 is up-regulated in vivo on the vasculature after intravesical treatment of urinary bladder cancers with IFN-␣. In conclusion, distinct behavior of lymphocyte and endothelial CD73 subsequent to cytokine treatment further emphasizes the existence of cell type-specific mechanisms in the regulation of CD73 expression and function. Overall, these results suggest that IFN-␣ is a relevant in vivo regulator of CD73 in the endothelial- leukocyte microenvironment in infections/inflammations, and thus has a fundamental role in controlling the extent of inflamma- http://www.jimmunol.org/ tion via CD73-dependent adenosine production. The Journal of Immunology, 2004, 172: 1646–1653.

he interaction between lymphocytes and endothelial cells cially interesting in light of the fact that the cDNA and protein is a multistep process. To be able to penetrate the vessel structure are very similar on both cell types (7). T wall and to reach the target site, circulating cells use a Adenosine, a purine nucleoside product of CD73 enzyme activ- very finely regulated set of adhesion molecules. Enhanced adhe- ity, has a role in many physiological and pathological events. It sion to endothelium and subsequent transmigration of recirculating binds to specific receptors on the cell surface. To date, four dif- leukocytes through the endothelial lining of vessel wall into the ferent subtypes of G protein-coupled adenosine receptors, A1R, by guest on September 29, 2021 tissue are characteristic for inflammation. Moreover, the release of A2aR, A2bR, and A3R, have been cloned (9). Ecto-5Ј-nucleoti- pro- and anti-inflammatory cytokines in a high extent takes place dase activity is shown to be an important mediator of the anti- at sites of inflammation. Those cytokines are potent regulators of inflammatory effect of methotrexate and sulfasalazine in vitro and the expression of adhesion molecules (1Ð3). in vivo in the murine air pouch inflammation model through in- Ј Ecto-5 -nucleotidase (CD73) is a 70-kDa GPI-anchored cell sur- creasing extracellular adenosine levels (10). Adenosine, by bind- face molecule with ecto-enzymatic activity. It is abundantly ex- ing to A1 and A2 receptors, regulates pathological consequences pressed on the vascular endothelium and at a low level on certain of inflammation by controlling leukocyte binding to endothelium subpopulations of human lymphocytes. It is part of the purine sal- (11) and acts as an anti-inflammatory agent by binding to A2 and Ј vage pathway by degrading nucleoside-5 -monophosphates (AMP A3 receptors, through the inhibition of neutrophil degranulation and IMP) into nucleosides such as adenosine and inosine (4, 5). (12). Adenosine also decreases eosinophil migration through acti- CD73 has also been suggested to mediate homing of skin-infil- vation of A3 receptor (13). Adenosine, converted from neutrophil- trating lymphocytes in vivo (6). Triggering of CD73 on the surface derived AMP, leads to increased endothelial barrier function by of lymphocytes, but not on endothelial cells, results in the shed- endothelial A2bR activation. This promoting effect of AMP is ding of the CD73 and increased adhesion of lymphocytes to en- CD73 mediated and is followed by an increase in intracellular dothelium via LFA-1 clustering (7, 8). This phenomenon is espe- cAMP (14). Recently, a critical role for A2a receptor has been shown in decreasing systemic and tissue-specific inflammatory re- *MediCity Research Laboratory and Department of Medical Microbiology, Turku sponses in vivo (15). University, Turku and National Public Health Institute, Turku, Finland; and †Depart- ment of Surgery, Turku University Central Hospital, Turku, Finland To date, practically nothing is known about the regulation of Received for publication May 28, 2003. Accepted for publication November endothelial CD73 expression and function. However, in inflam- 20, 2003. mation there may be some inducers secreted that in vivo specifi- The costs of publication of this article were defrayed in part by the payment of page cally control endothelial CD73 expression. This hypothesis was charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. suggested by a previous finding, that CD73 is up-regulated in inflamed skin (6). 1 This work was supported by grants from the Finnish Academy, the Sigrid Juselius Foundation, the Finnish Cancer Union, the Ida Montin Foundation, the Paulo Foun- As adenosine, having an anti-inflammatory and cell-protective dation, the Research and Science Foundation of Farmos, the Finnish Cultural Foun- effect, plays an important role in controlling the extent and con- dation, and the Turku Graduate School of Biomedical Sciences. sequences of inflammation, this work was designed to identify the 2 Address correspondence and reprint requests to Dr. Sirpa Jalkanen, MediCity Re- search Laboratory, Turku University, Tykisto¬katu 6A, FIN-20520 Turku, Finland. factors responsible for the regulation of CD73 expression as well E-mail address: sirpa.jalkanen@utu.fi as ecto-5Ј-nucleotidase-mediated adenosine production.

Materials and Methods Table I. Regulators used to induce CD73 expression Cells, Abs, and reagents Inducer Concentration Used Incubation Time HUVEC (3) were isolated using a method modified from Jaffe et al. (16), and were cultured on gelatin-coated cell culture flasks in complete medium, Rapidly acting as described earlier (7). Human PBL from healthy volunteers were isolated FMLP 10Ϫ9 and 10Ϫ7 M 5 min and 2 h using Ficoll-Hypaque (Histopague-1077; Pharmacia, Uppsala, Sweden). Dibutyryl cAMP 0.5 and 5 ␮g/ml 5 min and 2 h PBL and U266B1 cell line, a gift from J.-Y. Bonnefoy (Glaxo Institute for Histamine 1 and 10 ␮g/ml 5 min and 2 h Molecular Biology, Geneva, Switzerland), was cultured in RPMI 1640 PMA 1 and 10 ng/ml 5 min and 2 h medium containing 10% FCS, 4 mM L-glutamine, 100 U/ml penicillin, and Slowly acting 100 ␮g/ml streptomycin. Anti-CD73 mAb 4G4 (mouse IgG1), anti- IL-1␤ 10 and 100 U/ml 4, 20 h ICAM-1 mAb 5C3 (IgG1) (17), anti-CD31 mAb 2C8 (IgG1), and mAb IL-4 10 and 100 U/ml 4, 20 h 3G6 (mouse IgG1) against chicken T cells as a negative control Ab were 100 and 500 U/ml 48 h used. ␣,␤ Methyleneadenosine 5Ј-diphosphate (AMPCP)3 and AMP were TNF-␣ 10 and 200 U/ml 4, 20 h from Sigma-Aldrich (St. Louis, MO). 200 and 1000 U/ml 48 h Inductions and immunofluorescence stainings LPS 10 and 100 ng/ml 4, 20 h 100 and 500 ng/ml 48 h IL-1␤, IL-4, TNF-␣, and IFN-␥ (Genzyme, Cambridge, MA); IFN-␣ IFN-␣ 1, 5, 10, 50, 100, 200, 72 h (Wellferon; The Wellcome Foundation, London, U.K.); LPS from Esche- 500, and 2000 U/ml richia coli serotype O:55 (Difco Laboratories, Detroit, MI); and dibutyryl 1000 U/ml 4, 12, 20, 24, 48, cAMP, PMA, thrombin from human plasma, histamine, and FMLP from 60, 72, and 96 h Sigma-Aldrich were used for inductions, as indicated in Table I. For every IFN-␥ 50 and 200 U/ml 4 and 20 h Downloaded from time point, a control flask was incubated without inducers. 1000 U/ml 12, 20, 24, 48, 60, and 72 h Three different protocols were used for immunofluorescence stainings 1) To study the effect of a panel of different inducers on surface expression of CD73 immunofluorescence analyses were performed, as reported before using TaqMan Universal PCR Master Mix and ABI PRISM 7700 Sequence (17). In brief, HUVEC were treated with or without inducers and detached Detector (Applied Biosystems, Foster City, CA). The expression of the http://www.jimmunol.org/ ϫ 5 with 5 mM of EDTA-trypsin. A total of 5 10 cells for each staining was housekeeping gene GAPDH was used as a reference for normalization, and incubated with saturating concentrations of mAb 3G6 negative control (neg the relative increase of CD73 mRNA expression between control and IFN- co), 4G4 (anti-CD73), and 5C3 (anti-ICAM-1) as hybridoma supernatants ␣-treated cells was calculated. or purified Ab (final concentration, 10 ␮g/ml) for 20 min at 4¡C and washed twice. Then the cells were incubated for 20 min at 4¡C with 1/100 Ecto-5Ј-nucleotidase assay diluted FITC-conjugated sheep anti-mouse-IgG mAb (Sigma-Aldrich) containing 5% AB serum. Finally, the cells were washed twice and fixed with Ecto-5Ј-nucleotidase activity was assayed by TLC, as described previously 1% paraformaldehyde. All incubations and washes were performed with (18). Briefly, the standard enzyme assay contained in a final volume of 120 ␮ ϫ 4 ϫ 5 PBS containing 2% FCS and 1 mM of NaN3. Fluorescence was then de- l of RPMI 1640, 4Ð6 10 detached HUVEC (or 1 10 lymphoid tected using FACS (BD Biosciences, San Jose, CA). The difference be- cells), 5 mmol/L ␤-glycerophosphate, and the indicated concentrations of 3 tween control and treated cells was calculated from: fold difference ϭ AMP with tracer [2- H]AMP (sp. act., 18.6 Ci/mmol; Amersham, Little by guest on September 29, 2021 treated cells (MFI, anti-CD73 Ϫ MFI, neg co)/nontreated cells (MFI, anti- Chalfont, U.K.). Incubation times were chosen to ensure the linearity of the CD73 Ϫ MFI, neg co). reaction with time, so that the amount of the converted AMP did not exceed 2) For the detection of intracellular CD73, lymphocytes were perme- 7Ð10% of the initially introduced substrate. Aliquots of the mixture were abilized before immunofluorescence stainings by incubating them for 2 applied to Alugram SIL G/UV254 TLC sheets (Macherey-Nagel, Du¬ren, min in acetone at Ϫ20¡C. Then the cells were washed with RPMI 1640 Germany) and separated with isobutanol/isoamyl alcohol/2-ethoxyethanol/ 3 medium containing 5% FCS and stained and analyzed by FACS, as de- ammonia/H20 (9:6:18:9:15) as solvent. H-labeled AMP and its dephos- scribed above. phorylated nucleoside derivatives were visualized in UV light and quanti- 3) To study the distribution of CD73 on HUVEC, cells were seeded on fied using a Wallac-1409 ␤-spectrometer. 10-mm-diameter gelatin-coated coverslips and cultured in the presence or absence of IFN-␣ (100 U/ml) for 72 h. The cells were washed twice with Permeability assay PBS and then incubated for 30 min on ice with saturating concentration of To evaluate barrier function of confluent monolayers, HUVEC were seeded mAb 4G4 (anti-CD73). After rinsing three times with PBS, coverslips were (50,000 cells/insert) on Transwell insert polycarbonate filters (6.5 mm in incubated further with FITC-conjugated sheep anti-mouse IgG mAb di- diameter, 0.4 in ␮m pore size; Costar, Cambridge, MA). The filters were luted 1/100 for 20 min on ice, washed three times in PBS, and mounted on treated for 1Ð2 h with fibronectin and air dried before seeding endothelial glass slides in Vectashield mounting medium (Vector Laboratories, Bur- cells. Typically, monolayers were studied 4Ð5 days postseeding. HUVEC lingame, CA). The cells were examined on a Zeiss (Oberkochen, Germany) were either induced with IFN-␣ (100 U/ml) for 72 h before the studies of LSM 510 laser-scanning confocal microscope. monolayer permeability or grown in medium without IFN-␣. Transport CD73 RNA analysis across endothelial monolayers was assessed by measuring the flux of ␮ FITC-labeled dextran (500 g/ml, Mr 70,000). Endothelial monolayers For RNA isolation, 2 ϫ 105 HUVEC were seeded onto six-well plates. were pretreated with AMP (50 ␮M) for 15 min before the FITC-dextran Typically, RNA was isolated after 4Ð5 days postseeding. IFN-␣ (100 transport was initiated. To evaluate the role of CD73 enzymatic activity on U/ml) was added to wells 72, 24, or 12 h before RNA isolation. Total RNA the endothelial cell permeability, the flux of FITC-dextran was measured in was isolated using RNase Easy Kit (Qiagen, Valencia, CA), according to the presence or absence of a specific inhibitor of ecto-5Ј-nucleotidase, the manufacturer’s instructions. Before real-time RT-PCR measurement, AMPCP (100 ␮M). In certain experiments, AMPCP was added to the 1Ð2 ␮g of total RNA was treated with DNase I, and cDNA was made by upper and lower chambers 30 min before the transport was initiated by using Superscript II reverse transcriptase (Life Technologies, Gaithersburg, adding FITC-labeled dextran. The inserts were removed from the bottom MD), according to the manufacturer’s instructions. Finally, cDNA was chamber (Visiplate; PerkinElmer Life Sciences, Turku, Finland) at the time treated with RNase H (Life Technologies). CD73 primers 5Ј-CTG GGA points 10, 20, 30, 40, and 100 min, and FITC-labeled dextran was mea- GCT TAC GAT TTT GCA-3Ј and 5Ј-CCT CGC TGG TCT GCT CCA-3Ј sured directly from the bottom chambers in a fluorometer (TECAN Ultra and probe 5Ј-CCA ACG ACG TGC ACA GCC GG-3Ј were designed fluorescence reader; Tecan Group, Maennedorf, Switzerland) using 485 using Primer Express computer software (PE Biosystems, Foster City, and 535 nm as the excitation and emission wavelengths, respectively. CA). Primers and probe for GAPDH housekeeping gene were used as internal controls. The real-time RT-PCR measurements were performed Patient characteristics Twelve patients having superficial epithelial bladder cancer were evaluated 3 Abbreviations used in this paper: AMPCP, ␣,␤ methyleneadenosine 5Ј-diphosphate; for operation from 1 to 3 wk before the actual operation. In connection to MFI, mean fluorescence intensity; neg co, negative control. the evaluation visit, biopsies were taken from normal area of the bladder 1648 CD73-MEDIATED ADENOSINE UP-REGULATION and from the tumor. Fifty million units of IFN-␣2b (IntronA; Schering- HUVEC were IFN-␣ and IFN-␥, with doses of Ͼ200 U/ml after Plough, Kenilworth, NJ) were instilled to the bladder 1 day before the induction for 20Ð24 h. As IFN-␣ is rather widely used in the clin- operation. The cystectomy was performed and the patients underwent con- ical medicine, its effects were evaluated in more detail. ventional ureteroenterocutaneostomy, enterocystoplasty, or ureteroente- oumbilicostomy as the reconstructive operation. Three patients did not re- ␣ ceive IFN-␣ before operation. Two of them received 100 mg epirubicin (a IFN- increases endothelial CD73 expression in a time- and standard treatment in the Turku University Hospital; Pharmorubicin; Phar- dose-dependent manner macia) instilled to the bladder 1 day before the operation, and one did not Practically all nonactivated HUVEC bear CD73 on their surface receive anything. Their tumors were analyzed before (biopsy) and after the operation and used as controls. All patients were Caucasian males. Patient when analyzed by FACS (7). Therefore, to measure the increase in characteristics appear in Table II. The study protocol was approved by the expression of CD73 molecules on cell surface, we analyzed the Ethical Board of the Turku University Hospital, and informed consent was MFI of HUVEC, as described in Materials and Methods. To fur- obtained from each patient. ther study the kinetics of IFN-␣ up-regulation, confluent mono- ␣ Immunohistochemical stainings layers of HUVEC were incubated using different doses of IFN- for the indicated periods of time. CD73 expression was increased The bladder sample specimens were snap frozen in liquid nitrogen and cut Ϯ ϭ ␮ time dependently almost 2-fold (92.4 11.5%; n 9) after 72 h into 5- m sections. Subsequently, serial sections were stained with anti- ␣ CD73 mAb 4G4, anti-CD31 mAb 2C8, or 3G6 (negative control) as pri- with 1000 U/ml IFN- (Fig. 1a). Longer exposure of the HUVEC mary Abs, and peroxidase-conjugated rabbit anti-mouse IgG (DAKO A/S, to IFN-␣ did not cause further significant increase in CD73 ex- Glostrup, Denmark) was used as a second-stage Ab. The reaction was pression (data not shown). A similar pattern of CD73 up-regulation developed by adding 3,3Ј-diaminobenzidine tetrahydrochloride (Poly- was seen after induction with IFN-␥ (data not shown). Up-regu-

sciences, Warrington, PA) in PBS. All incubations were 20 min, with sat- Downloaded from urating mAb concentrations, followed by two washes with PBS. The num- lation of CD73 expression was also dose dependent, as in concen- ber of positive vessels/microscopic field (ϫ200) was counted, and intensity trations ranging from 10 to 1000 U/ml the highest increase in in- of the staining was semiquantitatively evaluated. A combined score from 0 tensity was observed at 1000 U/ml (Fig. 1b). Comparable increase to 3 was given to each sample. Score 0 was assigned to samples with no subsequent to IFN-␣ treatment took place in HUVEC grown in positive blood vessels, and score 3 to samples with staining equal to in- Transwell chambers (data not shown). Immunofluorescence stain- flamed tonsil. Scores 1 and 2 were adjusted to cover the staining patterns ␣ in between. All samples were read blindly by two independent readers, and ings followed by confocal laser microscopy revealed that IFN- the mean of their scores is presented for each sample. treatment does not induce any significant changes in the distribution http://www.jimmunol.org/ or polarization of CD73 on HUVEC surface. Instead, CD73 is more Statistical analysis intensely, but similarly distributed on the cell surface (Fig. 2). Data are presented as mean Ϯ SEM of individual experiments. Statistical comparisons were made using StudentЈs t test, and p values Ͻ0.05 were Induction of CD73 RNA expression by IFN-␣ taken as significant. Data from kinetic experiments were subjected to computer analyses using the Michaelis-Menten equation to determine the K Next, we wanted to determine whether increase in surface CD73 m expression is mediated by increase in CD73 RNA expression. and Vmax values (GraphPad Prism version 3.0, San Diego, CA). Quantitative real-time PCR revealed prominent induction of CD73 Results by guest on September 29, 2021 CD73 expression is up-regulated on endothelial cells with IFN-␣ This work was designed to find potent regulators of CD73 expression or CD73-based ecto-5Ј-nucleotidase enzyme activity after cells were exposed to a wide range of well-known inducers. First inductions were done for 4 min or 2 h for rapidly acting inducers (i.e., inducers known to release preformed molecules from intracellular storage granules) PMA, FMLP, dibutyryl cAMP, thrombin, and histamine and for 4 or 20Ð24 h for slowly acting inducers (i.e., inducers leading to de novo synthesis of new molecules) IL- 1␤, LPS, IL-4, TNF-␣, IFN-␣, and IFN-␥ (see Table I). The only inducers leading to a marked change in the CD73 expression on

Table II. Patient characteristics of IFN-␣-treated patients

Interval between Age Biopsies and Patient (years) Cystectomy Histology (WHO grade)

1 59 1 wk UCa grade 2 eosinophilic cystitis 2 73 1 wk UC grade 3 3 64 1 wk UC grade 2 4 66 3 wk UC grade 3 5 70 1 wk SCCb grade 2 6 59 1 wk UC grade 3 with squamocellular FIGURE 1. Induction of CD73 surface expression on HUVEC by component IFN-␣ is both time and dose dependent. a, HUVEC were exposed to 1000 7 68 2 wk UC grade 3 U/ml IFN-␣ for indicated time periods. b, HUVEC were cultured with 8 65 3 wk UC grade 2 different concentrations of IFN-␣ for 72 h. Relative means of MFI Ϯ SEM 9 70 1 wk UC grade 1, invasive of three to nine experiments are shown. Control expression is the expres- ␣ a UC, urothelial cancer. sion of CD73 without IFN- at each time point. Background (the negative .p Ͻ 0.01 ,ءء ;p Ͻ 0.05 ,ء .b SCC, squamous cell cancer. control staining) is subtracted The Journal of Immunology 1649

Despite the structural similarity of endothelial and lymphoid CD73, IFN-␣ promotes different effects on these two cell types After finding out that CD73 expression on endothelial cells is IFN-␣ inducible, we wanted to determine whether CD73 on lymphocytes would also behave similarly in the same conditions. A total of 1000 U/ml IFN-␣ did not increase CD73 expression on PBL significantly (data not shown). Even with longer induction time up to 48 h, we observed only minor changes in CD73 expression on lymphocyte surface. To exclude the possibility that freshly isolated lymphocytes do not survive well in culture conditions and, therefore, fail to up-regulate CD73, we also treated CD73-expressing lymphoid cell line U266B1 with IFN-␣. They also were incapable of up-regulating their CD73 expression even after 48 or 72 h of induction. Instead, there was a decrease at 48-h time point after IFN-␣ treatment as compared with control cells (relative MFI 82.1 Ϯ 5.6% vs 100%; n ϭ 3). To elucidate whether there is also up-regulation in intracellular CD73 protein level, PBL and HUVEC were permeabilized with acetone before immunofluorescence staining after IFN-␣ induc- Downloaded from tion. No up-regulation of intracellular expression of CD73 could be observed in PBL. Similar results were obtained when analyses were done with FACS and fluorescence microscopy. In HUVEC, a slightly increased intracellular staining with anti-CD73 mAb was seen after IFN-␣ induction (data not shown). http://www.jimmunol.org/ CD73 is up-regulated in vessels of healthy and tumor areas of ␣ FIGURE 2. Induction of CD73 with IFN- leads to increased expres- bladder after treatment with IFN-␣ in vivo sion rather than changes in its distribution. HUVEC were either grown in medium or induced with IFN-␣ for 72 h, and the cell surface expression of To investigate whether IFN-␣ would also regulate the expression CD73 was detected with mAb 4G4 against CD73 and FITC-conjugated of CD73 in vivo, tissue specimens from superficial epithelial blad- anti-mouse IgG Ab. The characteristic distribution of CD73 on control (a der cancers were collected before and after IFN-␣2b treatment, and b) and on IFN-␣-treated (c and d) endothelial cells. On control stained, and analyzed (Table II). In two tumors, the malignant cells HUVEC CD73 is expressed on the cell surface in a punctate-like pattern. were CD73 positive, reflecting the fact that also some epithelial After IFN-␣ induction, expression of CD73 is more intense, but surface ␣

cells are CD73 positive. IFN- produced a clear up-regulation of by guest on September 29, 2021 distribution is similar as on control HUVEC. Images (a and c) are projec- CD73 in vascular endothelium both in normal and cancer tissue in tions generated from confocal serial sections of ﬂuorescently labeled cells, and images (b and d) are confocal XZ proﬁles (Z-scans). Bar ϭ 10 ␮m.

mRNA expression after 12- and 24-h exposure to IFN-␣ (100 U/ml) (2.0 Ϯ 0.5-fold increase and 2.1 Ϯ 0.2-fold increase, respectively, compared with untreated cells after normalization to GAPDH, p Ͻ 0.05). Treatment of HUVEC with IFN-␣ for 72 h resulted in maximal increase of CD73 mRNA (2.5 Ϯ 1.1-fold increase, p Ͻ 0.05) (Fig. 3).

FIGURE 3. IFN-␣ increases the relative CD73 mRNA expression in HUVEC. HUVEC were incubated with IFN-␣ (100 U/ml) for 12, 24, and 72 h or left untreated (0 h). Total RNA was isolated, and real-time PCR was used to determine CD73 mRNA levels. Data were calculated relative FIGURE 4. IFN-␣ induces CD73 expression in vivo. A summary of to internal control gene (GAPDH) and are expressed as fold increase over semiquantitative analysis of immunohistochemical stainings of urine blad- untreated cells (0 h) Ϯ SEM at each indicated time. The data present the der samples from healthy (a) and tumor (b) areas before and after IFN-␣ .p Ͻ 0.05. treatment is presented ,ء ;means of three experiments 1650 CD73-MEDIATED ADENOSINE UP-REGULATION vivo when compared with expression levels before and after treat- 5Ј-nucleotidase activity, we applied a radiochemical assay for di- ment in specimens of the control patients treated with epirubicin rect measurement of [3H]AMP conversion into [3H]adenosine. (Fig. 4). However, no CD73 up-regulation was detected among the Pretreatment of HUVEC monolayers with IFN-␣ (1000 U/ml for few normal lymphocytes present within the tumors. Similarly, the 48 h) caused significant increase in the rate of [3H]AMP hydrolysis expression level of tumor cells remained constant during IFN-␣2b (Fig. 6a), whereas no significant activation of the enzyme activity treatment in those tumors that were positive for CD73 (Fig. 5). was detected after PBL treatment with IFN-␣ (Fig. 6b). To further Three patients, who did not receive IFN-␣ and were used to control elucidate the mechanism of ecto-5Ј-nucleotidase activation, kinetic the possible up-regulation of CD73 caused by the biopsy and op- analysis of [3H]AMP hydrolysis by control and IFN-␣-treated eration itself, did not show any significant increase in their CD73 HUVEC was conducted, and these saturating curves can be seen in expression (one patient did not show any change and two patients Fig. 6c. Statistical analysis revealed that IFN-␣ significantly in- Ј had an increase of 0.5 in endothelial CD73 expression within the creased the maximum hydrolysis rate (Vmax)of5-nucleotidase as tumor tissue). Thus, the mean change of the control patients was compared with nontreated cells (525 Ϯ 30 vs 350 Ϯ 29 nmol/106 ϭ 0.3, and that of treated patients 1.3 ( p 0.02). cells/h) without any modification of the enzyme affinity (Km ϳ50Ð60 ␮mol/L). These data suggest that IFN-␣ increases the IFN-␣ increases ecto-5Ј-nucleotidase activity on endothelial number of enzymatically active 5Ј-nucleotidase molecules on the cells endothelial surface rather than induces conformational changes of To determine whether IFN-␣-induced increase of CD73 expres- the enzyme catalytic site. Interestingly, use of the same approach sion on HUVEC is accompanied by concomitant induction in ecto- with other nucleotide [3H]ATP did not reveal significant changes

of ATP-hydrolyzing activities after HUVEC treatment with IFN-␣ Downloaded from (data not shown), conﬁrming the speciﬁcity of ecto-5Ј-nucleotidase induction. To ensure that CD73 is not continuously secreted from lymphocytes into cell culture supernatant producing increased enzyme activity, we analyzed [3H]AMP conversion into [3H]adenosine in cell culture medium from IFN-␣-induced lym-

phocytes and nontreated control cells. No signiﬁcant change in http://www.jimmunol.org/ enzymatic activity of cell culture medium was found between control and IFN-␣ treatment (data not shown).

IFN-␣ increases HUVEC membrane function To study whether IFN-␣-up-regulated CD73 expression and CD73-mediated increase in adenosine production are able to regulate HUVEC membrane function, we measured the ﬂux of FITC- labeled dextran through conﬂuent endothelial monolayers growing by guest on September 29, 2021

FIGURE 5. IFN-␣ induces expression of endothelial CD73 also in vivo in patients suffering from bladder carcinoma. a, A specimen stained with anti-CD73 mAb before IFN-␣ treatment. The vessels are practically negative for CD73 (score 0). b, A serial section of the same sample stained with anti-CD31 mAb demonstrating the vessels. c and d, Serial sections FIGURE 6. IFN-␣ increases cell surface ecto-5Ј-nucleotidase activity. from a sample of the same patient taken after IFN-␣ treatment stained with HUVEC (a) and PBL (b) were pretreated for 48 h without (Ⅺ) or with 1000 anti-CD73 mAb (c) or with anti-CD31 mAb (d). CD31-positive vessels U/ml IFN-␣ (f). Ecto-5Ј-nucleotidase activity was assayed by using 300 express CD73 abundantly (score 3). eÐh, An example of a tumor (t) ex- ␮mol/L [3H]AMP and expressed on ordinate as nmoles of substrate de- p Ͻ 0.05 ,ء .(pressing CD73 before (e) and after (g) IFN-␣ treatment. Also in this case, phosphorylated by 106 cells per hour (mean Ϯ SEM; n ϭ 4Ð5 IFN-␣ increased CD73 expression on endothelial cells (e, score 1 and g, as compared with control cells. c, Rate of [3H]AMP hydrolysis by control score 2). f and h, Anti-CD31 stainings. a--d and g--h, The arrows point to (F) and IFN-␣-treated (E) HUVEC vs substrate concentration plot. Values individual vessels. e and f, The arrows point toward a neovascular network are expressed as mean Ϯ SEM for two independent experiments. The ki- of vessels. Insets in d and h show stainings with a negative control Ab 3G6. netic parameters (Vmax and Km) were calculated from the presented curves Scale bar 100 ␮minallfigures. and summarized in the text. The Journal of Immunology 1651 on permeable insert wells. At all time points examined, there was CD73 is up-regulated at sites of inflammation and especially in the a significant difference ( p Ͻ 0.05) in the permeabilities of HUVEC skin (6), but the mediators causing this effect have remained untreated with IFN-␣ (100 U/ml) for 3 days compared with untreated known. In this work, we screened a multitude of rapidly and long HUVEC, as indicated by decreased flux of FITC-dextran (Fig. 7a). activating mediators and found IFN-␣ to be a potent activator of Pretreatment of HUVEC monolayers with a specific CD73 enzyme CD73 expression. IFN-␣ produced specific time- and dose-depen- inhibitor, AMPCP, reversed the permeability decrease associated dent in vitro up-regulation of CD73 expression on endothelium, with IFN-␣ treatment as demonstrated by increased flux of FITC- but not on PBL, in concentrations measured in patients suffering dextran (Fig. 7b). from infections (21). Even more importantly, it up-regulated CD73 expression in tumor vessels of bladder carcinoma patients in vivo. Discussion We also demonstrated that up-regulated endothelial CD73 after In inflammatory conditions, many cytokines and chemokines are IFN-␣ induction is enzymatically functional, producing adenosine secreted that markedly modulate the expression and/or activation from 5Ј-AMP, leading to enhanced barrier function in endothelial status of multiple adhesion molecules (19, 20). Expression of cells. Moreover, we found a cell type-specific difference in regulation of endothelial and lymphocyte CD73. During normal cell maturation, CD73 expression and its enzymatic activity increase (22, 23). One of the known activators is the end product of the enzymatic activity of CD73 itself: adenosine increases CD73 expression and enzymatic activity in a paracrine fashion (24). The human CD73 promoter sequence contains pos- Downloaded from sible binding sites for many transcription factors such as Sp1, Sp1/ AP-2, cAMP response element, and Sp1/NF-AT (25), and among them is agtttcgtatcac sequence, which rather closely resembles IFN-␣-stimulated response elements in certain known genes. Al- though this sequence does not fully match the consensus site (agtt tcnntttcnc) of those known genes (26), it remained still possible http://www.jimmunol.org/ that this sequence in CD73 gene acts as an IFN-␣ response element. Therefore, we tested whether nuclear proteins isolated from HUVEC can bind to this sequence by using EMSA (data not shown). Based on these results, IFN-␣ seems not to use this element, but instead exerts its effect indirectly inducing other mediators, which then cause up-regulation of CD73. None of the rapidly acting factors tested in this work was able

to up-regulate CD73. This is interesting in light that expression of by guest on September 29, 2021 lymphocyte CD73 can be increased by triggering of CD38 in 20 min. In this case, the additional CD73 is translocated to the cell surface from a pre-existing intracellular pool (27). Based on the fact that a relatively long time was needed for IFN-␣ to up-regulate CD73 in vitro, we may argue that its up-regulation is not an early event at sites of inflammation in vivo either. It may further indicate that up-regulation of CD73 and increased production of adenosine are the body’s own defense mechanisms to decrease and limit inflammation. This is well in line with the findings that adenosine prevents cell damage during heart and central nervous ischemia (9, 28, 29). After hypoxia, ecto-5Ј-nucleotidase activity increases due to phenomenon known as preconditioning (30Ð32). This results in release of large amounts of adenosine, leading to increased resistance of cells to infarction, for example, in cardiac hypoxia. The prerequisite for adenosine formation is availability of AMP. AMP is an important intermediate metabolite of ecto-enzymatic ATP metabolism that can be either irreversibly hydrolyzed into adenosine by ecto-5Ј-nucleotidase or regenerated into ATP via ecto-nucleotidase kinase reactions. Concentration of AMP is not a FIGURE 7. Effect of IFN-␣ on permeability of HUVEC monolayers. a, limiting factor, because CD73 functions as a master switch, deter- HUVEC were plated on porous polystyrene membrane (pore size of 0.4 mining the shift from the ATP-consuming/adenosine-producing ␮m) and grown to confluency. HUVEC were grown in medium or treated pathway to the ATP-generating pathway, and whenever more with 100 U/ml IFN-␣ for 72 h. Fifteen minutes after adding AMP, the AMP is needed for adenosine production less ATP is generated membrane function was analyzed by measuring the flux of 70-kDa FITC- (18, 33). Moreover, pathological conditions such as hypoxia and dextran through HUVEC monolayer to the lower chamber with a fluorom- tissue injury trigger release of high amounts of AMP (34, 35). eter. The FITC-dextran flux was measured up to 100 min. Values are IFNs are very potent immunomodulatory substances. They pro- .p Ͻ 0.05 as compared with IFN-␣-treated cells ,ء .means Ϯ SEM, n ϭ 3 b, Confluent monolayers were exposed to AMPCP (100 ␮M), a specific duce antiproliferative effects, induce antiviral resistance, and reg- ␣ ecto-5Ј-nucleotidase inhibitor, 30 min before addition of FITC-dextran. ulate immune responses (36). Levels of IFN- in healthy individ- p Ͻ 0.05 as compared with uals are usually low, but in inflammatory conditions IFN-␣ is ,ء .Data shown are mean values Ϯ SEM, n ϭ 3 IFN-␣-treated cells. secreted already at early stages of inflammation from APCs. The 1652 CD73-MEDIATED ADENOSINE UP-REGULATION lower concentrations used in our experiments that cause up-regu- 4. Thompson, L. F., J. M. Ruedi, A. Glass, G. Moldenhauer, P. Moller, M. G. Low, lation of CD73 are comparable to those found endogeneously in M. R. Klemens, M. Massaia, and A. H. Lucas. 1990. Production and character- ization of monoclonal antibodies to the glycosyl phosphatidylinositol-anchored patients suffering from inflammations (21, 37), and thus expected lymphocyte differentiation antigen ecto-5Ј-nucleotidase (CD73). Tissue Antigens to up-regulate CD73 also in vivo. Interestingly, IFN-␣ seems to 35:9. 5. Zimmermann, H. 1992. 5Ј-Nucleotidase: molecular structure and functional as- maintain the integrity of vascular wall, because it has been shown pects. Biochem. J. 285:345. to enhance endothelial barrier function of bovine retinal microvas- 6. Arvilommi, A.-M., M. Salmi, L. Airas, K. Kalimo, and S. Jalkanen. 1997. CD73 cular endothelial cells (38). The mechanisms were not elucidated mediates lymphocyte binding to vascular endothelium in inflamed human skin. Eur. J. Immunol. 27:248. in this earlier work by Gillies and Su, but our present results 7. Airas, L., J. Niemela¬, M. Salmi, T. Puurunen, D. J. Smith, and S. Jalkanen. 1997. strongly suggest that up-regulation of CD73 contributes to this Differential regulation and function of CD73, a glycosyl-phosphatidylinositol- phenomenon via increased adenosine production. linked 70-kD adhesion molecule, on lymphocytes and endothelial cells. J. Cell ␣ Biol. 136:421. IFN- has also been used to treat various cancers. In this work, 8. Airas, L., J. Niemela¬, and S. Jalkanen. 2000. CD73 engagement promotes lym- we found that bladder carcinoma patients treated with IFN-␣ up- phocyte binding to endothelial cells via a lymphocyte function-associated anti- regulate their CD73 expression specifically on endothelial cells. gen-1-dependent mechanism. J. Immunol. 165:5411. 9. Olah, M. E., and G. L. Stiles. 1995. Adenosine receptor subtypes: characteriza- This type of up-regulation was not detectable in normal lympho- tion and therapeutic regulation. Annu. Rev. Pharmacol. Toxicol. 35:581. cytes always present in variable numbers within the tumor tissues 10. Morabito, L., M. C. Montesinos, D. M. Schreibman, L. Balter, L. F. Thompson, or in malignant cells of two tumors positive for CD73. CD73 pos- R. Resta, G. Carlin, M. A. Huie, and B. N. Cronstein. 1998. Methotrexate and sulfasalazine promote adenosine release by a mechanism that requires ecto-5Ј- itivity is thought to provide survival advantage as a salvage path- nucleotidase-mediated conversion of adenine nucleotides. J. Clin. Invest. way for CD73-positive tumor cells when the patients are treated by 101:295. antimetabolites that block the de novo synthesis of purines (39, 11. Cronstein, B. N., R. I. Levin, M. Philips, R. Hirschhorn, S. B. Abramson, and Downloaded from G. Weissmann. 1992. Neutrophil adherence to endothelium is enhanced via aden- 40). In contrast, in the hypoxic conditions often present in tumors, osine A1 receptors and inhibited via adenosine A2 receptors. J. Immunol. increased adenosine production might give the normal cells, espe- 148:2201. cially endothelial cells and CD73-positive tumor-infiltrating lym- 12. Bouma, M. G., T. M. M. A. Jeunhomme, D. L. Boyle, M. A. Dentener, N. N. Voitenok, F. A. J. M. van den Wildenberg, and W. A. Buurman. 1997. phocytes, an advantage to survive better than CD73-negative can- Adenosine inhibits neutrophil degranulation in activated human whole blood: cer cells, which have a higher proliferation rate and thus are more involvement of adenosine A2 and A3 receptors. J. Immunol. 158:5400. 13. Knight, D., X. Zheng, C. Rocchini, M. Jacobson, T. Bai, and B. Walker. 1997. vulnerable to cytotoxic drugs. Trophic actions of adenosine on http://www.jimmunol.org/ Adenosine A3 receptor stimulation inhibits migration of human eosinophils. endothelial cells may be advantageous to cancer growth as such, J. Leukocyte Biol. 62:465. because angiogenesis is a prerequisite for tumor growth (41). 14. Lennon, P. F., C. T. Taylor, G. L. Stahl, and S. P. Colgan. 1998. Neutrophil- Ј Our present results of the distinct effects of IFN-␣ on endothelial derived 5 -adenosine monophosphate promotes endothelial barrier function via CD73-mediated conversion to adenosine and endothelial A2B receptor activation. and lymphocyte CD73 further demonstrate the difference between J. Exp. Med. 188:1433. the cell types to regulate the expression of CD73. This difference 15. Ohta, A., and M. Sitkovsky. 2001. Role of G-protein-coupled adenosine receptors ␣ in down-regulation of inflammation and protection from tissue damage. Nature is not explained by lack of IFN- receptors on lymphocytes, be- 414:916. cause both B and T cells express high affinity receptor for IFN-␣ 16. Jaffe, E. A., R. L. Nachman, C. G. Becker, and C. R. Minick. 1973. Culture of (42). In our earlier work, we discovered that triggering of CD73 human endothelial cells derived from umbilical veins: identification by morpho-

logic and immunologic criteria. J. Clin. Invest. 52:2745. by guest on September 29, 2021 with anti-CD73 mAb (that mimics the ligand binding) results in 17. Airas, L., M. Salmi, and S. Jalkanen. 1993. Lymphocyte-Vascular adhesion pro- shedding of lymphocyte, but not endothelial cell CD73, although tein-2 is a novel 70-kDa molecule involved in lymphocyte adhesion to vascular the cDNA sequences of CD73 in these cell types are practically endothelium. J. Immunol. 151:4228. 18. Yegutkin, G. G., T. Henttinen, and S. Jalkanen. 2001. Extracellular ATP forma- identical (7, 8). In this context, it is worth mentioning that the tion on vascular endothelial cells is mediated by ecto-nucleotide kinase activities amount of CD73 varies markedly between lymphocytes and en- via phosphotransfer reactions. FASEB J. 15:251. 19. Constantin, G., M. Majeed, C. Giagulli, L. Piccio, J. Y. Kim, E. C. Butcher, and dothelial cells. Only 10Ð15% of lymphocytes express CD73, and ␤ C. Laudanna. 2000. Chemokines trigger immediate 2 integrin affinity and mo- the expression level is low in comparison with, for example, bility changes: differential regulation and roles in lymphocyte arrest under flow. HUVEC, which all are positive for CD73. This type of cell-spe- Immunity 13:759. cific differences in amount and regulation of CD73 may be fun- 20. Butcher, E. G., and L. J. Picker. 1996. Lymphocyte homing and homeostasis. Science 272:60. damental for appropriate behavior of lymphocytes, the role of 21. Pirisi, M., C. Fabris, P. Toniutto, E. Falleti, S. Tisminetzky, M. Gerotto, which is to actively deaminate the existing adenosine and extrav- G. Soardo, D. Vitulli, M. Del Forno, F. Baralle, and E. Bartoli. 1997. Endogenous asate to lymphoid tissues or to sites of inflammation. In contrast, interferon-␣ concentration and outcome of interferon treatment in patients with chronic hepatitis C. Dig. Dis. Sci. 42:767. adenosine is necessary for endothelial cells to maintain their bar- 22. Edwards, N. L., E. W. Gelfand, L. Burk, H. M. Dosch, and I. H. Fox. 1979. rier function. Distribution of 5Ј-nucleotidase in human lymphoid tissues. Proc. Natl. Acad. Sci. In conclusion, we have demonstrated in this work how CD73 USA 76:3474. 23. Resta, R., S. W. Hooker, K. R. Hansen, A. B. Laurent, J. L. Park, can be up-regulated both in vitro and in vivo by IFN-␣. As aden- M. R. Blackburn, T. B. Knudsen, and L. F. Thompson. 1993. Murine ecto-5Ј- osine is highly anti-inflammatory in its nature, manipulation of its nucleotidase (CD73): cDNA cloning and tissue distribution. Gene 133:171. endogenous production via up-regulation of CD73 may be a po- 24. Narravula, S., P. F. Lennon, B. U. Mueller, and S. P. Colgan. 2000. Regulation of endothelial CD73 by adenosine: paracrine pathway for enhanced endothelial tential way to treat harmful inflammatory conditions such as reper- barrier function. J. Immunol. 165:5262. fusion injuries in connection to myocardial infarction and stroke. 25. Spychala, J., A. G. Zimmermann, and B. S. Mitchell. 1999. Tissue-specific regulation of the ecto-5Ј-nucleotidase promoter: role of the cAMP response element site in mediating repression by the upstream regulatory region. J. Biol. Chem. Acknowledgments 274:22705. We thank Maritta Pohjansalo, Laila Reunanen, Sari Ma¬ki, and Pirjo Heinila¬ 26. Darnell, J. E., Jr., I. M. Kerr, and G. R. Stark. 1994. Jak-STAT pathways and for excellent technical assistance; Anne Sovikoski-Georgieva for secre- transcriptional activation in response to IFNs and other extracellular signaling tarial help; and Erkki Nieminen for help with the figures. proteins. Science 264:1415. 27. Peola, S., P. Borrione, L. Matera, F. Malavasi, A. Pileri, and M. Massaia. 1996. Selective induction of CD73 expression in human lymphocytes by CD38 ligation: References a novel pathway linking signal transducers with ecto-enzyme activities. J. Im- 1. Butcher, E. C. 1991. Leukocyte-Endothelial cell recognition: three (or more) munol. 157:4354. steps to specificity and diversity. Cell 67:1033. 28. Heurteaux, C., I. Lauritzen, C. Widmann, and M. Lazdunski. 1995. Essential role ϩ 2. Springer, T. A. 1994. Traffic signals for lymphocyte recirculation and leukocyte of adenosine, adenosine A1 receptors, and ATP-sensitive K channels in cerebral emigration: the multistep paradigm. Cell 76:301. ischemic preconditioning. Proc. Natl. Acad. Sci. USA 92:4666. 3. Carlos, T. M., and J. M. Harlan. 1994. Leukocyte-Endothelial adhesion mole- 29. Linden, J. 2001. Molecular approach to adenosine receptors: receptor-mediated cules. Blood 84:2068. mechanisms of tissue protection. Annu. Rev. Pharmacol. Toxicol. 41:775. The Journal of Immunology 1653

30. Kitakaze, M., M. Hori, T. Morioka, T. Minamino, S. Takashima, Y. Okazaki, derived paracrine factor that elicits chloride secretion from T84 intestinal epi- K. Node, K. Komamura, K. Iwakura, T. Itoh, et al. 1995. ␣1-Adrenoceptor ac- thelial cell monolayers. J. Clin. Invest. 91:2320. tivation increases ecto-5Ј-nucleotidase activity and adenosine release in rat car- 36. Schindler, C. W. 2002. Series introduction: JAK-STAT signaling in human dis- diomyocytes by activating protein kinase C. Circulation 91:2226. ease. J. Clin. Invest. 109:1133. 31. Kitakaze, M., K. Node, T. Minamino, K. Komamura, H. Funaya, Y. Shinozaki, 37. Ronnblom, L., and G. V. Alm. 2001. A pivotal role for the natural interferon M. Chujo, H. Mori, M. Inoue, M. Hori, and T. Kamada. 1996. Role of activation ␣-producing cells (plasmacytoid dendritic cells) in the pathogenesis of lupus. of protein kinase C in the infarct size-limiting effect of ischemic preconditioning J. Exp. Med. 194:F59. Ј through activation of ecto-5 -nucleotidase. Circulation 93:781. 38. Gillies, M. C., and T. Su. 1995. Interferon-␣2b enhances barrier function of 32. Minamino, T., M. Kitakaze, T. Morioka, K. Node, K. Komamura, H. Takeda, bovine retinal microvascular endothelium in vitro. Microvasc. Res. 49:277. M. Inoue, M. Hori, and T. Kamada. 1996. Cardioprotection due to precondition- Ј 39. Ujha«zy, P., M. Klobusicka«, O. Babusõ«kova«, P. Strausbauch, E. Mihich, and ing correlates with increased ecto-5 -nucleotidase activity. Am. J. Physiol. Ј 270:H238. M. J. Ehrke. 1994. Ecto-5 -nucleotidase (CD73) in multidrug-resistant cell lines 33. Henttinen, T., S. Jalkanen and G. G. Yegutkin. 2003. Adherent leukocytes pre- generated by doxorubicin. Int. J. Cancer 59:83. Ј vent adenosine formation and impair endothelial barrier function by ecto-5Ј-nu- 40. Pieters, R., and A. J. P. Veerman. 1988. The role of 5 nucleotidase in therapy- cleotidase/CD73-dependent mechanism. J. Biol. Chem. 278:24888. resistance of childhood leukemia. Med. Hypotheses 27:77. 34. Bhatnagar, A. 2003. Surviving hypoxia: the importance of rafts, anchors, and 41. Burnstock, G. 2002. Purinergic signaling and vascular cell proliferation and ﬂuidity. Circ. Res. 92:821. death. Arterioscler. Thromb. Vasc. Biol. 22:364. 35. Madara, J. L., T. W. Patapoff, B. Gillece-Castro, S. P. Colgan, C. A. Parkos, 42. Paul, W. E. 1993. Fundamental Immunology, Chapter 21. Raven, New York, p. C. Delp, and R. J. Mrsny. 1993. 5Ј-Adenosine monophosphate is the neutrophil- 815. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021