The human interleukin-3 (IL-3) receptor is constitutively ex- pressed on certain hematopoietic cells where it mediates proliferation and differentiation, or ...
Interferon-y Upregulates Interleukin-3 (IL-3) Receptor Expression in Human Endothelial Cells and Synergizes With IL-3 in Stimulating Major Histocompatibility Complex Class I1 Expression and Cytokine Production By Eija I. Korpelainen, Jennifer R. Gamble, William B. Smith, Mara Dottore, Mathew A. Vadas, and Angel F. Lopez The human interleukin-3 (IL-3) receptor is constitutively expressed on certain hematopoietic cells where it mediates proliferation and differentiation, or functional activation. We have recently found that human umbilical vein endothelial cells (HUVECs) also express IL-3 receptors and that theexpression is enhanced by stimulation with the monokinetumor necrosis factor a.In this report we show that the lymphokine interferon y (IFNy) is a powerful stimulator of the IL-3 receptor of HUVECs and that the combination of IL-3 and IFNyhas a synergistic effect on major histocompatibility complex (MHCI class II expression and on the production of the early-acting hematopoietic cytokines IL-6 and granulocyte colony-stimulating factor (G-CSF). IFNy caused a timeand dose-dependent up-regulation of mRNA for both theLY and j3 chnins of the IL-3 receptor, with maximal effects occuring 12 t o 24 hours after stimulation with IFNy at 100 U/ mL. Induction of mRNA correlated with protein expression on thecell surface, as judged by monoclonal antibody staining of bothreceptor chains and by the abilityof HUVEC t o specifically bind '251-labefed IL-3 ['a51-lL-3). Scatchard
analysis of HUVECs stimulated with IFNy at 100 U/mL for 24 hours showed -6.300 IL-3 receptors percell thatwere of a high affinity class (dissociation constant [kd] = 500 pmol/ L) only. The addition ofIL-3 t o IFNy-treated HUVECs strongly enhanced the expression of MHC class II antigen. Importantly, IFNy and IL-3 also exhibited a synergistic effect in the induction of the mRNA for G-CSF and IL-6. This was reflected in increased amounts of G-CSF and IL-6 protein in HUVEC supernatants. In contrast, IFNy and IL-3 did not stimulate granulocyte-macrophage colony-stimulating factor(GMCSF) or IL-8production in HUVECs. These results show that IFNy is a strong stimulator of IL-3 receptor expression in HUVECs and suggest that in vivoT-cell activation, causing the concomitant production of IFNy and IL-3, may lead to enhanced endothelial MHC class II expression and to the selective production ofearly-actinghematopoieticcytokines. Thus, IL-3 could influence immunity and hematopoiesis by acting not only on hematopoietic cells, but also on vascular endothelium. 0 1995 by The American Society of Hematology.
express the IL-3R9-" and that the expression is enhanced by tumor necrosis factor a (TNFa)'. IL-3 stimulation of TNFaactivated endothelial cells enhanced IL-8 production, E-selectin expression, and neutrophil transmigration.' This suggested that IL-3 can play a role in inflammation not only by stimulating effector functions of mature leukocytes, but also by regulatingtheirlocalization to sites of inflammation through its action on endothelium. In this report we examine the regulation of endothelial IL-3R expressionand function by thelymphokine IFNy. IFNy is produced, like IL-3," upon T-cell activation and it is important in the initial phaqe of the immune response as it induces the expression of major histocompatibility complex (MHC) class 11 antigens on a variety of cell types including endothelial cells." IFNy is also able toinfluence hematopoiesis, enhancing [L-3-dependentproliferation of CD34' cells.'" Endothelial cells arean important source of hematopoietic cytokines such as granulocyte colony-stimulating factor (GCSF), GM-CSF, and IL-6, which can enhance the inflammatory response in several ways. G-CSF and GM-CSF notonly enhance effector functions of granulocytes and monocytes, but also stimulate their maturation,".'" thereby replenishing the pool of circulating leukocytes. IL-6 regulates immune responses and the acute phase reaction and also stimulates hematopoietic progenitor growth in synergy with 1L-3.I7 In keeping with their role in inflammation, the production of these cytokines by endothelium is induced by inflammatory stimuli such as TNFa, IL-10, and lipopolysaccharide (LPS).".'" We show here that IFNy induces IL-3R expression in human umbilical vein endothelial cells (HUVECs) and synergizes with IL-3 in stimulatingtheproduction of G-CSF and IL-6, but not GM-CSF or 1L-8. These results suggest that IL-3 may influence hematopoiesis not only directly by acting on progenitor cells but also indirectly.
NTEIUEUKIN-3 (IL-3) is a hematopoietic growth factor that stimulates proliferation and differentiation of hematopoietic progenitor cells into various blood cell types,' a property that has led to its clinical usein bone marrow (BM) reconstitution after cancer therapy.* IL-3 also stimulates the effector functions of monocytes, eosinophils, and basophils, therebyhaving the potential to regulate inflammation and allergy."6 IL-3 exerts its biologic activities through binding to specifichigh-affinity receptors on the cell surface. The human IL-3 receptor (R) consists of at least two chains, an cr subunit, which is specific for IL-3 and binds IL-3 with low affinity,' and a ,!? subunit, which does not detectably bind IL-3 by itself but confers high-affinity binding when expressed together with the a chain. The ,B chain is shared betweentherelated receptors for granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-S and is critical for function.8 We and others have recently found that, in addition to hematopoietic cells, human endothelial cells also
From the Division of Human Immunology. HanSon Centre .fbr Cancer Research, Adelaide, South Australia. Submitted Octoher 3, 1994; uccepted Februup 17, 1995. Supported by the National Health & Medico1 Research Council of Au.srralia and by the National Institutes of Health Grant No. CA 45822. E.I.K. i.s a recipient of a Postgraduate Scholarship for Research from the University of Adelaide. Address reprint requests io Angel F. Lopez, MD, PhD, Division of Human Immunology, institute of Medical and Veterinary Science, Box 14, Rundle Mall Post Ofice, Adelaide, South Australia 5000. The publicationcosts of this article were defrayed in part by page charge payment. This at-tick must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. section 1734 sole[v t o indicate this fact. 0 1995 by The American Society of Hematology. 0006-4971~5/~601-0041$3.00/0 176
Blood, Vol 86, No 1 (July 1). 1995: pp 176-182
IFN-y UPREGULATES ENDOTHELIAL IL-3R EXPRESSION
TF1 t P HWEC+IFNT(U~I)
HWEC + IFN-y(h) 0
M TF1 t
12 18 24 48 72'
Fig 1. Selective induction of IL-3R a and chain mRNA in HUVECs by IFNy. RNase protection assay of IL-3R, GM-CSFR, and IL-5R subunit mRNAs in HUVECs incubated with (A) varying concentrations of IFNy for 24 hours at 37°C and (B)IFNy at 100 UlmL for various times as indicated. The samples were also probed for GAPDH mRNA, which is used as an internal control. RNA from TF-1 cells was used as a positive control (TF11 and tRNA as a negative control (t). Undigested probes (P) and DNA size markers (M; "P-labeled Hpa Il-digested pUC19 DNA) are also shown. The band between the IL-3R a and p chain signals appears consistently under the conditions of the protection assay and is caused by incomplete digestion of vector sequence contained in the IL-3R a chain probe.
through the production of early-acting hematopoietic cytokines from endothelial cells. MATERIALS AND METHODS Cell culture and stimulation. HUVECs were isolated from umbilical cords and cultured as previously described" in M 199 medium containing Earle's Salts (Cytosystems, Sydney, Australia), 20% fetal calf serum (FCS; PA Biologicals, Sydney, Australia), endothelial cell growth supplement (Collaborative Research) at 25 pglmL, and heparin (Sigma) at 37°C in a 5% CO, atmosphere. HUVECs were used between passages 2 and 5. Cells were stimulated with IFNy (2 X IO' Ulrng; Genentech, San Francisco, CA), IL-3 (Genetics Institute, Cambridge, MA), or TNFa (6.27 X IO' Ulmg; Genentech). LPS was undetectable (< 100 pglmL) in these cytokine preparations, as judged by the Limulus amoebocyte assay. TF-I cells2' were maintained in RPMI 1640 medium containing 10% FCS and recombinant GM-CSF at 2 ng/mL. RNase protection assays. Total cellular RNA was isolated from HUVECs using guanidinium thiocyanate?* A IO-pg sample of total RNA was analyzed by RNase protection assay as previously described:' except that only RNase A was used (20 pglmL) and the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe used as an internal control was synthesized at a lower specific activity using [a-'2P] uridine triphosphate at 4 Cilmmol. The probes for the different receptor subunits were designed to protect fragments of different sizes to allow simultaneous analysis. For probe synthesis, cDNAs were subcloned into pGEMl (IL-5Ra. GM-CSFRa, common p chain, G-CSF, IL-6). pGEM2 (IL-3Ra), pGEM4Z (GMCSF),or Bluescript KS (GAPDH). The probes protect the following fragments ofthemRNA,with the enzymes usedto linearize the transcription template shown in parentheses: I , I 19-1,280 (Psr I) for IL-3Ra chain, 937-1.191 (Pvu 11) for IL-5Ra chain, 606-731 (Rsa I) for GM-CSFRa chain, 833-1.032 (Bal I) for the common p chain, 1,233-1.508 (Mlu I) for G-CSF, 910-1.128 (Rsa I) for IL-6, 549756 (NCOI) for GM-CSF and 707-810 (Sty I) for GAPDH. Gels were quantified by Phosphorhager analysis (Molecular Dynamics, Menlo Park, CA), with GAPDH being used as an internal control. Binding of '"I-/L-3. Binding experiments were performed essentially as described.24In brief, Escherichia coli-derived rhIL-3 was radioiodonated to high specific activity (range, 1.1 to 1.3 X IO6 cpndpmol) using iodine monochloride and the radio-labeled cytokine separated from iodide ions by chromatography on Sephadex G-25
(Pharmacia) in phosphate-buffered saline containing 0.02% Tween 20. Before use, radiolabeled protein was purified from non-proteinassociated radioactivity by cation-exchange chromatography using carboxymethyl-Sepharose CL-6B. The Scatchard transformation*' was derived from a saturation binding curve using 2 X lo6 HUVECs per sample and six different concentrations of"'I-IL-3, ranging from 80 pmol/L to 1.4 nmoll L. HUVEC were detached with EDTA (40 mmol/L) and chondroitin sulphate (100 pg/mL) before incubation with "'I-IL-3 for 3 hours at 22°C in0.15 mL of binding medium (RPMI 1640, pH 7.4, supplemented with 20 mmol/L HEPES and 0.5% bovine serum albumin) containing 0.1% sodium azide to prevent receptor internalization. Cell suspensions were then overlayed onto 0.2mLofFCSand centrifuged for 30 seconds at maximum speed in a Beckman Microfuge 12. The visible cell pellet was recovered by cutting the tube, and the radioactivity determined in a Packard Auto Gamma 5650. Specific binding was determined by subtracting the cpm from parallel samples containing a 100-fold excess ofunlabeled IL-3. All samples were analyzed in duplicate. Binding data were analyzed by EBDA and LIGAND software (Elsevier-Biosoft, Cambridge, UK). Specific "'1-IL-3 binding to HUVEC during an IFNy-time-course was measured in a similar fashion, except that I2'I-IL-3 was used at 1 nmol/L. Flow cytometry. The surface expression of the IL-3R a and p chains was examined using the specific monoclonal antibodies (MoAbs) 7.G3 and 4.F3, respectively.*"HUVEC were stimulated as described in the text and then incubated with the anti-IL-3R a chain MoAb 7.G3 or the anti-8 chain MoAb 4.F3 for 30 minutes at 37°C. After three washes, HWECs were incubated further withFITCconjugated rabbit antimouse IgG (Silenus) at 4°C washed three times with RPM1 1640 medium containing 2.5% FCS, detached with TrypsinEDTA (0.05%/0.02%; Cytosystems, Castle Hill, New South Wales, Australia) and fixed in 0.4% formaldehyde. A minimum of IO4 HUVECs were then analyzed using the EPICS-Profile I1 (Coulter Electronics, Hialeah, FL) flow cytometer. The relative mean fluorescence intensity (MFI) values, obtained directly by the instrument's linear conversion of log units, were used for quantitative analysis. An irrelevant MoAb of matching isotype was used as a negative control to determine background fluorescence. All samples were analyzed in duplicate. MHC class I1 expression was detected and quantitated in a similar fashion byflow cytometry using the anti-MHC class I1 MoAb FMC52.
in HUVEC by IFNy. To examine whether the enhanced 1L3R tnRNA levels were accompaniedby an increase in protein - 1FN-y expression, the induction of IL-3R subunitso n the surface of - - - TNF-u, IFNr IFNy-stimulated HUVEC was monitored using the specific ;-, MoAbs 7.G3 and 4.F3, directed against the IL-3R a and 0 chain, respectively.26 IFNy increased the surface expression of both the IL-3R a and ,O chains as judgedby How cytometry (Fig 2 ) . In five separate experiments, the IL-3R 01 chain expressionincreased by 3.5- t- 0.8-fold(mean i SD) and kg fluonrscence kg fluorescence the 0 chain expression by 1.6- 2 0.4-fold (mean t- SD) upon 24 hours IFNy-stimulation. Fig 2. Flow cytometry analysis of IL-3R a and p chain induction Because we have previously shown that TNFa upregulates in HUVECs. Cellswere stimulated with medium alone, IFNy 1100 U/ mL), TNFa 1100 UlmL), or both factors together for 24 hours and the IL-3R in HUVEC,' we investigated whether IFNy could stainedasdescribed in Materials and Methods. Shown is the further increase this induction. We found that the combinasmoothed data set. The isotype control is represented by the grey tion of IFNy and TNFa had a synergistic effect on surface histogram. expression of both the IL-3R 01 and p chain on HUVEC (Fig 2), increasing the L 3 R a chain expression by 14. l - 0.8fold(mean t SD; n = 4) and the p chain expression by Cytokine measurements. For 1L-6, G-CSF,andGM-CSFmea8.0- 5 4.2-fold (mean 5 SD; n = 4), as judged by How treated with either surements,HUVECs ( 5 X IO' perwell)were cytometry. This effect was also observed at the mRNA level culture medium alone, TNFa for 24 hours, IFNy for 48 hours, IL(data not shown). 3 for 24 hours, or IFNy for 48 hours with1L-3 added for the last 24 The induction of both the IL-3R a and p chain mRNA hours. Immunoreactive IL-6 was measured in HUVEC supernatants and protein by IFNy correlated with the ability of HUVEC using an enzyme-linked immunosorbent assay (ELISA; Quantikine; to specifically bind IL-3. Whereas the numberof IL-3 recepR&DSystems,Minneapolis,MN). G-CSF measurementswere torsonunstimulated cells variedbetween endothelial cell kindly performed by Dr J. Layton(LudwigInstituteforCancer Research, Melbourne, Australia) by anE L B A method as previously lines (ranging from undetectable to 2,000 receptors per cell), described." GM-CSF was detected using a sandwich ELISA method. IFNy consistently enhanced IL-3binding. Scatchard analysis In brief, mouse-antihuman GM-CSF MoAb 4.D4 was used to coat of a saturation curve using '"I-IL-3 and HUVEC stimulated round-bottomed96-wellplatesbeforeloadingsamples and stanwith IFNy for 24 hours at 100 U/mL showed ~ 6 , 3 0 01Ldards. An avidin-biotin peroxidase system was then used to detect 3R per cell,which were of ahigh affinity class (kd, S00 thesecondstagerabbit-antihumanGM-CSFpolyclonalantibody. pmol/L) only (Fig 3A). We monitored the induction of ILFor IL-8 measurements, HUVEC (10' per well) were treated with 3 binding to IFNy-stimulated HUVECs more closely over TNFa for 24 hours, IFNy for 24 hours, IL-3 for 6 hours, or IFNy a 48-hour period using 1 nmol/L '2sI-IL-3. '"I-IL-3 binding for 24 hours with IL-3 added for the last 6 hours. The medium was to HUVECreached a plateau at20 hours after IFNy-stimulathen changed and supernatants were collected after 4 hours for ILtion (100 U/mL), and remained elevated thereafter (Fig 3B). 8 measurements. IL-8 was quantified with an ELISA as described.2x IL-3 potentiates the IFNy-induced MHC class I1 expresRESULTS sion i n HUVECs. Given that IFN y induces the expression of MHC classI1 antigens o n endothelial cells," we examined Selective induction of IL-3R a and p chain mRNA expreswhether IL-3 could regulate this expression. iL-3 (30 ng/ sion in HUVECs by IFNy. HUVECs were simultaneously mL) alonedid not induceMHCclass I1 on HUVECsas examined for the presence of mRNA for the IL-3R a and p judged by flow cytometry, but it strongly enhanced the exchains as well as for therelated GM-CSFR and IL-SR a pression induced by IFNy compared with the levels elicited chains by using RNase protection probes of different sizes. It was important to study the expression of the GM-CSFR and IL-SR a chains, because they share the p subunit with the IL-3R. Unstimulated HUVECs expressed mRNA for the A 0.3 IL-3R a chain and the common p chain, butnot for the related GM-CSF or IL-S receptor a chains (Fig l). Treatment of HUVECs with increasing concentrations of IFNy for 24 0.2 U hours led to increases in mRNA for both the a and p chains U of the IL-3R, whereas the mRNA for the GM-CSFR and ILrna 0.1 SR a chains remained undetected. As a positive control for thereceptor mRNAs, we used RNA from TF-l cells that 0 l express all three receptor^.*^^'^ The maximal effect of IFNy M 4 0 O M ) l W 0 8 1 6 2 4 3 2 4 0 4 8 bound (PM) time (h) was achieved at a concentration of 100 to 1,000 U/mL (Fig 1 A). The inductionof both the IL-3R a and p chain mRNAs Fig 3. CharacterizationofIFNy-inducedIL-3Rexpression in HUby IFNy in HUVECs was transient,withlevelsbeinginVECs. (A) Scatchard analysis of '251-lL-3 binding to HUVECs treated creased at 2 hours, reaching a maximum at 12 to 24 hours with IFNy at100 U/mL for 24 hours. (B)Timetourse of specific '''land decreasing slowly thereafter (Fig 1B). IL-3 l1 nmollL) binding to HUVECs stimulated with IFNy at 100 U / mL. Induction of IL-3R protein expression and IL-3 binding
ENDOTHELIAL IFN-y UPREGULATES
cytokine mRNA expression in response to IL-3 in both resting and IFNy-stimulated HUVEC. We found that IL-3 (30 ng/mL) did not increase the IL-6 mRNA in untreated HUVEC, whereas IFNy alone had only a modest effect. However, IL-3 strongly enhanced the IL-6 mRNA in HUVECs pretreated with lFNy (Fig 5A) to levels similar to those observed with TNFa, which was used as a positive control. A similar pattern of induction was observed for the G-CSF mRNA; IL-3 alone didnot induce the G-CSF mRNAbut had a synergistic effect on G-CSF expression with IFNy (Fig 5B). In contrast, GM-CSF mRNA was not induced by IL-3, either alone or in combination with IFNy (Fig 5B). On the other hand, TNFa strongly induced both G-CSF and GM-CSF mRNA (Fig 5B). To examine whether increased cytokine mRNA levels were reflected in increased protein production we examined HUVEC supernatants for the presence of IL-6, G-CSF, and GM-CSF. In keeping with the mRNA results, IL-3 did not induce IL-6 secretion in unstimulated HUVECs but caused a 3.2- 2 1.0-fold (mean +- SD; n = 3) increase in IL-6 secretion in HUVEC pretreated with IFNy for 24 hours (Fig 6). Similarly, whereas IL-3 did not increase G-CSF secretion in unstimulated HUVEC, it did cause a 3.0- t- 1 .2-fold (mean +- SD; n = 3) increase in G-CSF production in HUVECs that had been treated with IFNy, comparcd with thc levels clicited by IFNy alone (Fig 6). In contrast with IL-6 and GCSF production, GM-CSF production was not induced by IL-3 and IFNy, either alone or in combination. Consistent with the GM-CSF mRNA levels, immunoreactive GM-CSF could be detected only in supernatants from TNFa-stimulated cells (Fig 6). We also measuredproduction of the chemokine IL-8, which we havepreviouslyshown to be enhanced by IL-3 in TNFa-activated HUVEC.9 Interestingly, IL-3 did not induce IL-8 production in IFNy-activated HUVEC under conditions where the IL-3R expression was maximal (Fig 6).
24 48 time (h)
Fig 4. IL-3 synergizes with IFNy in upregulating MHC class II antigen expression on HUVECs. Time-course of indirect immunofluorescence quantified byflow cytometry, ofHUVECs stimulated with IFNy (100 U/mL), of IL-3 (30 ng/mLI, or of both factors together. The Yaxis shows the relative MFI above background, as described in Materials and Methods. Data are from one experiment representative of three and are expressed as the mean 2 SD determined in duplicate samples. (W), IFNy; (01,IFNy IL-3; (A),IL-3.
by lFNy alone. The enhancing effect of IL-3 was first noted after 48 hours of stimulation and lasted for at least 72 hours (Fig 4). At 48 hours, the MHC class I1 expression was increased 3.1- -t 0.7-fold (mean % SD; n = 4) by IL-3, comparedwith the levels induced by IFNy alone. In contrast, IL-3 did not affect MHC class I expression, either alone or in combination with IFNy (data not shown). IL-3 enhances G-CSF and IL-6, hut not GM-CSF or IL8 production in IFNy-treated HUVECs. To further investigate the functional role of the IL-3R in IFNy-stimulated cells, we measured the production of cytokines involved in hematopoiesis and inflammation. We initially measured
In this study we report the upregulation of the IL-3R in human endothelial cells by IFNy and the functional effects
Fig 5. Induction of IL-6 and G-CSF, but not GMCSF mRNA by IL-3 in HUVECs.RNase protection assay probed for IL-6 mRNA (A) or G-CSF and GMCSF mRNA (B) using RNA from HUVECs incubated with medium alone, IFNy (l00 U/mL) for 48 hours, IL-3 (30ng/mL) for 24 hours, IFNy for 48 hours with 11-3 added for the last 24 hours, or with TNFa (100 UlmL) for 24 hours. The samples were also probed for GAPDH, which is used as an internal control. tRNA was used as a negative control It)and lane P represents undigested probes. (A) Phosphor image of the RNase protection gel.
HUVEC t P I
KORPELAINEN ET AL
Fig 6. IL-3 enhances IL-6 and G-CSF, but not GM-CSF or IL-8 production in IFhly-treated HUVECs. Cells were treated and cytokines measured as described in Materials and Methods. Dataare from one experiment representative of three and are expressed as the mean ? SEM determined in duplicate wells. (H),G-SCF; (D),IL-6; (W), IL8; (U),GM-CSF.
of IL-3 on thesecells. No effect was observed with IL-3 alone. However, L 3 stimulation of IFNy-activatedHUVEC-enhanced MHC class 11 expression, and the production of the early acting hematopoietic cytokines IL-6 and G-CSF, but not production of GM-CSF orIL-8. These results suggest that IL-3 mayinfluence immunityand hematopoiesis,not only by acting on hematopoietic cells, but also on vascular endothelium, where it acts as an amplification factor. The upregulation of the IL-3R in HUVEC by IFNy was detected using a combination of approaches. IFNyincreased both the L 3 R a and p chain mRNAs (Fig l ) , a finding which was reflected in increased surface expression of both proteinsas judged by MoAbstaining (Fig 2). In keeping with these results, we could also detect anincreasc i n specific lz'I-1L-3 binding to IFNy-stimulated HUVEC (Fig 3). This coordinate up-regulation of both IL-3R chains is important from the functional point of view because both chains are required for IL-3 signaling."] The induction of the TL-3R was selective in that mRNAs for the related GM-CSFR and IL-SRcu chains were not detected (Fig 1). We and others have previously shown that the L 3 R is expressed in HUVEC" I ' and can be up-regulated by TNFa inatime-anddose-dependentmanner."Interestingly,the expression of both L 3 R a and p chainscanbe further increased by combining IFNy and TNFa (Fig 2), the synergistic effect suggesting that they operate through different mechanisms. The kinetics of IL-3R induction by IFNy differed from that mediated by TNFa in being more prolonged. This maintenance of IL-3R expression over time together with increased MHC class I1 expression by IL-3 and IFNy (Fig 4), suggestamechanism by whichthese two factors can contribute to inflatnmation and delayed-typehypersensitivity. The concomitant production of IFNy and IL-3 upon T-cell activation in vivo may lead to enhanced MHC class I1 expression and, thus, antigen presentation by endothelium, thereby amplifying the T-helper response.
TFNy has been previously shown to selectively upregulate the chain, but not the GM-CSFRa chain in human monocytes," a finding which is similar to our results with HUVEC, suggesting that the expression of these genes is under similar regulatory control in endothelial and hematopoietic cells. In contrast, the findings that IFNy upregulates the IL3R LY chain in HUVEC (Fig 2), but does not affect IL-3R a chain expression in CD34' progenitor cells," indicatesa different pattern of regulation of the IL-3R a chain in hematopoietic cells and endothelium. Stimulation of IFNy-activated HUVEC with IL-3 led to an increased production of G-CSF and 1L-6. They are early acting hematopoietic growth factors that,together with 1L-3, form part of the optimal cytokine combination for hematopoiesis in vitro." TL-6 also has an importantrole in inflammation, being themain inducer of theacute-phase response.17 Both IL-6 and G-CSF can be foundin the circulation underinflammatory conditions and,interestingly,increased serum levels of IL-6 and acute-phase proteins have also been detected during clinical administration of IL-3.14 The increased secretion of TL-6 and G-CSF into the circulation by vascular endothelium in response to IL-3 and IFNy could initiate a cascade of events to enhance hematopoiesis inthe BM during inflammation. It wouldalsorepresenta second, indirect mechanism by which both IL-3 and IFNy can regulate hematopoiesis. In contrast with TNFa and other prototypic inducers of endothelial cytokine production, IL-3 and lFNy had the ability to selectively induce G-CSF and L 6 without increasing GM-CSF or the neutrophil chemoattractant 1L-8. Although we have measured a limited number of cytokines, the pattern of cytokine expression with IL-3 and IFNy and with TNFa suggests a bias towards different functions. Whereas TNFa induces cytokines involved predominantly in inflammation, IFNy and IL-3 may direct cytokine production more toward hematopoiesis. TL-3 acts in this system essentially as an amplitication factor: lFNy induces G-CSF and IL-6 i n endothelial cells, and this is enhanced by IL-3. By contrast, genes that are not induced by IFNy, such as IL-8 and GM-CSF, are not affected by IL-3. However, when IL-8 is induced by TNFa, L 3 amplifies this response.' The mechanism underlying the synergistic effect of 1L-3 and IFNy on G-CSF and IL-6 production in HUVEC is not known at present. IFNy nay render the cells more responsive to IL-3 by up-rcgulating the L 3 R , or the synergism may be caused by some post-receptor mechanism along the signalingpathway.Interestingly, IL-3 hasrecently been shown to activate JAK2 kinase,'< which has also been implicated in IFNy signaling, and to induce DNA binding complexes that recognize GAS (IFNy activation site) elements." Both increased transcription and mRNA stability have been shown to play a role in the induction of TL-6 and G-CSF by inflammatory ~ t i m u l i ' ~whether ~ ~ ~ ;theirinduction by IL-3 and IFNyin HUVEC is similarly mediated is currently being investigated. In this report, we show that IFNy is a strong stimulator of IL-3 receptor exprcssion in endothelial cells and suggest that in vivo T-cell activation, causing the concomitant production of IFNy and IL-3, may lead to enhanced MHC class
I1 expression and production of early acting hematopoietic cytokines by endothelium. In addition to acting directly on hematopoietic progenitors, IL-3 and 1FN-y may hence also influence hematopoiesis indirectly by enhancing cytokine secretion from the vascular endothelium. However, to draw firm conclusions this hypothesis will need to be tested further in vivo. It would also be of interest to determine whether endothelial cells isolated from BM3' respond to IL-3 similarly to HUVEC, allowing IL-3 to regulate cytokine production locally in the hematopoietic microenvironment. ACKNOWLEDGMENT We thank G. Goodall and C. Lagnado for expert help and advice, L. Noack for technical assistance; the staff of the Delivery Ward of the Queen Victoria Hospital for collecting umbilical cords; J. Layton of the Ludwig Institute for Cancer Research (Melbourne, Australia) for the G-CSF protein measurements; Genentech for providing IFNy and TNFa; and Genetics Institute for providing IL-3. REFERENCES 1. Metcalf D: The Hemopoietic Colony-Stimulating Factors. Amsterdam, The Netherlands, Elsevier, 1984 2. Orazi A, Cattoretti G, Schiro R, Siena S, Bregni M, Di Nicola M, Gianni AM: Recombinant human interleukin-3 and recombinant granulocyte-macrophage colony-stimulating factor administered in vivo after high-dose cyclophosphamide cancer chemotherapy: Effect on hematopoiesis and microenvironment inhumanbone marrow. Blood 79:2610, 1992 3. Elliott JI, Vadas MA, Cleland LC, Gamble JR, Lopez AF: IL3 and granulocyte-macrophage colony-stimulating factor stimulate distinct phases of adhesion inhuman monocytes. J Immunol 145:167, 1990 4. Lopez AF, Dyson P, To L-B, Elliott MJ, Milton S, Russel J, Juttner C, Yang Y-C, Clark SC, Vadas MA: Recombinant human interleukin-3 stimulation of hematopoiesis in humans: Loss of responsiveness with differentiation in the neutrophilic myeloid series. Blood 72:1797, 1988 5. Haak-Frendscho M,AraiN,AraiK-I, Baeza ML,Finn A, Kaplan AP: Human recombinant granulocyte-macrophage colonystimulating factor and interleukin 3 cause basophil histamine release. J Clin Invest 82:17, 1988 6. Lopez AF, Eglinton JM, Lyons AB, Tapley PM, To LB, Park LS, Clark SC, Vadas MA: Human interleukin-3 inhibits the binding of granulocyte-macrophage colony-stimulating factor and interleukin-5 to basophils and strongly enhances their functional activation. J Cell Physiol 145:69, 1990 7. Kitamura T, Sato N, Arai K-I, Miyajima A: Expression cloning of the human IL-3 receptor cDNA reveals a shared p subunit for the human IL-3 and GM-CSF receptors. Cell 66:1165, 1991 8. Lopez AF, Elliott MJ, Woodcock J, Vadas MA: GM-CSF, IL3 and IL-5: cross-competition on human haemopoietic cells. Immunology Today 13:495, 1992 9. Korpelainen EI, Gamble JR, Smith WB, Goodall GJ, Qiyu S , Woodcock JM, Dottore M, Vadas M, Lopez AF: The receptor for interleukin 3 is selectively induced in human endothelial cells by tumor necrosis factorcy and potentiates interleukin 8 secretion and neutrophil transmigration. Proc Natl Acad Sci USA 90: 11 137, 1993 10. Colotta F, Bussolino F, Polentarutti N, Guglielmetti A, Sironi M, Bocchietto E, De Rossi M, Mantovani A: Differential expression of the common p and specific a chains of the receptors for GMCSF, IL-3 and IL-5 in endothelial cells. Exp Cell Res 206:311, 1993 11. BrizziMF, Garbarino G, Rossi PR, Pagliardi GL, Arduino C, Avanzi GC, Pegoraro L: Interleukin 3 stimulates proliferation
and triggers endothelial-leukocyte adhesion molecule 1 gene activation of human endothelial cells. J Clin Invest 91:2887, 1993 12. Niemeyer CM, Sieff CA, Mathey-Prevot B, Wimperis JZ, Bierer B, Clark SC, Nathan DC: Expression of human interleukin3 (multi-CSF) is restricted to human lymphocytes and T-cell tumor lines. Blood 73:945, 1989 13. Pober JS, Gimbrone MA, Cotran RS, Reiss CS, Burakoff SJ, Fiers W, Ault KA: Ia expression by vascular endothelium is inducible by activated T cells and humany interferon. J Exp Med 157:1339, 1983 14. Caux C, Moreau I, Saeland S, Banchereau J: Interferon-y enhances factor-dependent myeloid proliferation of human CD34' hemopoietic progenitor cells. Blood 79:2628, 1992 15. Demetri GD, Griffin JD: Granulocyte colony-stimulating factor and its receptor. Blood 78:2791, 1991 16. Gasson JC: Molecular physiology of granulocyte-macrophage colony-stimulating factor. Blood 77:1131, 1991 17. Kishimoto T, Akira S, Taga T: Interleukin-6 and its receptor: A paradigm for cytokines. Science 258:593, 1992 18. Seelentag WK, Mermod J-J. Montesano R, Vassalli P: Additive effects of interleukin l and tumour necrosis factor-cy on the accumulation of the three granulocyte and macrophage colony-stimulating factor mRNAs in human endothelial cells. EMBO J 6:2261, 1987 19. Jirik FR, Podor TJ, Hirano T, Kishimoto T, Loskutoff DJ, Carson DA, Lotz M: Bacterial lipopolysaccharide and inflammatory mediators augment IL-6 secretion by human endothelial cells. J Immunol 142:144, 1989 20. Gamble JR, Elliott MJ, Jaipargas E, Lopez AF, Vadas MA: Regulation of human monocyte adherence by granulocyte-macrophage colony-stimulating factor. Roc Natl Acad Sci USA 86:7169, 1989 21. Kitamura T, Tange T, Terasawa T, Chiba S , Kuwaki T, Miyagawa K, Pia0 Y-F, Miyazono K, Urabe A, Takaku F: Establishment and characterization of a unique human cell line that proliferates dependently on GM-CSF, IL-3 or erythropoietin. J Cell Physiol 140:323, 1989 22. Chomczynski P, Sacchi N: Single-step method of RNA isolation by acid guanidiniun thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156, 1987 23. Goodall GJ, Wiebauer K, Filipowicz W: Analysis of premRNA processing in transfected plant protoplasts. Methods Enzymol 181:148, 1990 24. Lopez AF, Eglinton JM, Gillis D, Park LS, Vadas MA: Reciprocal inhibition of binding between interleukin 3 and granulocytemacrophage colony-stimulating factor tohuman eosinophils. Proc Natl Acad Sci USA 86:7022, 1989 25. Scatchard G: The attraction of proteins for small molecules and ions. Ann N Y Acad Sci 5 1:660, 1949 26. Woodcock JM, Zacharakis B, Plaetinck G, Bagley CJ, Qiyu S , Hercus TR, Tavernier J, Lopez AF: Three residues in the common chain of the human GM-CSF, IL-3, and IL-5 receptors are essential for GM-CSF and IL-5 but notIL-3 high affinity binding and interact with Glu2' of GM-CSF. EMBO J 13:5176, 1994 27. Zoellner H, Filonzi EL, Stanton HR, Layton JE, Hamilton JA: Human arterial smooth muscle cells synthesize granulocyte colonystimulating factor in response to interleukin-la and tumor necrosis factor-cy. Blood 80:2805, 1992 28. Van Zee KJ, De Forge LE, Fischer E, MaranoMA, Kenney JS, Remick D C , Lowry SF, Moldawer LL: IL-8 in septic shock, endotoxemia, and after IL-I administration. J Immunol 146:3478, 1991 29. Kitamura T, Takaku F, Miyajima A: IL-l up-regulates the expression of cytokine receptors on a factor-dependent human hemopoietic cell line TF-I. Int Immunol 3:571, 1991 30. Kitamura T, Miyajima A: Functional reconstitution of the human interleukin-3 receptor. Blood 80:84,1992
31.HallekM,LepistoEM,Slattery KE, Griffin JD,Ernst TJ: Interferon-y increases the expression of the gene encoding the p subunit of the granulocyte-macrophage colony-stimulating factor receptor. Blood 80: 1736, 1992 32. Sato N. Caux C, Kitamura T, Watanabe Y, Arai K-I, Bancherau J, Miyajima A: Expression and factor-dependent modulation of the interleukin-3 receptor subunits on human hemopoietic cells. Blood 82:752, 1993 33. Haylock DN, To LB. Dowse TL, Juttner CA, Simmons PJ: Ex vivo expansion and maturation of peripheral blood CD34' cells into the myeloid lineage. Blood 80:1405, 1992 34. Lindemann A, Ganser A, Hoelzer D, Mertelsmann R, Herrmann F: In vivo administration of recombinant human interleukin3 elicits an acute phase response involving endogenous synthesis of interleukin-6. Eur Cytokine Netw 2:173, 1991
KORPELAINEN ET AL
35. Silvennoinen 0, Witthuhn BA, Quelle F W , Cleveland JL, Yi T, Ihle JN: Structure of the murine Jak2 protein-tyrosine kinase and its role in interleukin 3 signal transduction. Proc Natl Acad Sci USA 90:8429,1993 36.Lamb P, Kessler LV, Suto C, Levy DE,SeidelHM.Stein RB,Rosen J: Rapidactivationofproteinsthatinteract with the interferon y activation site in response to multiple cytokines. Blood 832063. 1994 37. Elias JA, Lentz V: 1L-l and tumor necrosis factor synergistically stimulate fibroblast IL-6 production and stabilize IL-6 messenger RNA. J Immunol 145:161, 1990 38. Rafii S. Shapiro F, Rimarachin J, Nachman RL. FerrisB, Weksler B, Moore MAS, Asch AS: Isolation and characterization of human bone marrow microvascular endothelial cells: Hematopoietic progenitor adhesion. Blood 84:10, 1994