Leukemia (2002) 16, 1135–1142  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Direct signal transduction via functional -␣␤ receptors in CD34+ hematopoietic stem cells J Giron-Michel3, D Weill1, G Bailly1, S Legras2, PC Nardeux1, B Azzarone3, MG Tovey1 and P Eid1

1Laboratoire d’Oncologie Virale, UPR 9045, CNRS, Villejuif, France; 2Laboratoire d’He´matologie, Hoˆpital Paul Brousse, Villejuif, France; 3Inserm U506, Hoˆpital Paul Brousse, Villejuif, France

+ Affinity purified, freshly isolated CD34 progenitors were host defense against virus infection and neoplastic disease.4–6 shown to express low levels of type I interferon (IFN) receptors ␣ ± ± Recombinant IFN- has found wide application in clinical (740 60 binding sites/cell, Kd 0.7 0.04 nM) determined by Scatchard’s analysis using a radiolabelled, neutralizing, mono- medicine for the treatment of chronic viral hepatitis, solid clonal antibody directed against the IFNAR1 chain of the tumors, and a variety of hematological malignancies including human type I IFN receptor. Treatment of freshly isolated (day hairy cell leukemia and chronic myelogenous leukemia + 0), highly purified (Ͼ95% pure) CD34 cells with recombinant (CML).7–9 The mechanism(s) of the anti-tumor effects of IFN- ␣ IFN- resulted in rapid tyrosine phosphorylation and activation ␣ remain poorly understood, although it is well established of STAT1, Tyk2 and JAK1 as shown by Western immunoblot- ␣ ting. Similarly, IFN treatment was shown by confocal that IFN- exerts both direct anti-proliferative and pro-apop- microscopy to result in rapid nuclear localization of the tran- totic effects, and indirect immune-mediated effects on tumor scription factors IRF1 and STAT2, demonstrating the presence cells 10.10,11 In the case of hematological malignancies such + of functional IFN receptors on freshly isolated (day 0) CD34 as CML, IFN-␣ has been reported to inhibit the proliferation cells. The number of specific type I IFN receptor binding sites of both normal and leukemic progenitor cells.12–18 It remains expressed on hematopoietic progenitor cells increased to ␣ some 1440 ± 40 per cell after 11 days of cultivation of CD34+ unclear, however, whether IFN- exerts a direct effect on the cells in vitro suggesting that receptor expression increases proliferation of hematopoietic stem cells, and their immediate with cell differentiation. IFN-mediated signal transduction and progeny committed progenitor cells, or an indirect effect on the inhibitory effect of IFN-␣ on 7 or 14 days CFU-GM and BFU- stromal cells in the bone marrow, and the synthesis of growth E colony formation was abrogated in the presence of the ␣ factors and other required for stem cell prolifer- anti-IFNAR1 mAb, indicating that IFN- acts directly on the pro- 19–22 liferation of human hematopoietic progenitor cells via receptor ation. activated signal transduction without excluding the induction exert their characteristic biological effects by of other cytokines or growth factors by residual accessory binding to high-affinity receptors on the cell surface resulting cells. in the rapid tyrosine phosphorylation and activation of signal Leukemia (2002) 16, 1135–1142. DOI: 10.1038/sj/leu/2402492 transducers and activators of transcription (STATs) which regu- Keywords: interferon receptors; CD34+ cells; anti-IFNAR1 late the transcription of interferon-responsive genes. Thus, binding of IFN-␣ to its cell surface receptor activates the receptor-associated Janus tyrosine kinases Tyk2 and JAK123,24 Introduction leading to phosphorylation of the two transmembrane poly- + peptides, IFNAR1 and IFNAR2, which constitute the type I IFN Lymphohematopoietic CD34 stem cells and progenitors are receptor. Both chains of the receptor are members of the cyto- able to proliferate and differentiate into multiple lymphoid lin- kine receptor super family and are both required for high- eages, and understanding the mechanisms which control this process remains one of the central questions of modern day affinity ligand binding and the establishment of biological biology. Biochemical studies on primitive primary hemopo- activity. Tyk2 and JAK1 then phosphorylate STAT1 and ietic cells have been hindered by the difficulty in obtaining STAT2, leading to the formation of a heterodimer which is sufficient quantities of highly purified CD34+ cells. Further- translocated to the nucleus where it associates with a p48 more, analysis of mRNAexpression levels using techniques DNAbinding protein, a member of the IRF family of transcrip- 25–27 such as the polymerase chain reaction (PCR) does not provide tion factors (IRF9), leading to the activation of the tran- information on protein functionality. Thus, we have adopted scription of those genes which contain an IFN-sensitive an alternative approach based on the use of direct binding response element (ISRE) within their promoter region. It is the studies (Scatchard analysis), which provides information on activation of a specific set of IFN-responsive genes within a the level of translation of a receptor product and the level of given cell type which gives rise to the establishment of the its expression at the cell surface, coupled with an analysis of characteristic biological activities of the interferons. receptor signaling leading to the transcriptional activation of We show herein, using Scatchard analysis and a highly spe- effector genes. cific radiolabeled monoclonal antibody directed against the The proliferation and differentiation of hematopoietic stem IFNAR1 chain of the type I IFN receptor, which inhibits bind- cells is regulated by a complex network of cytokines and ing to both receptor chains and which neutralizes the biologi- growth factors,1–3 including the type I interferons (IFN-␣, IFN- cal activity of all the type I interferons,28,29 that freshly isolated ␤, IFN-␻, and IFN-␶) which are multifunctional cytokines that highly purified CD34+ cells (Ͼ95%) from human umbilical play an important role in the innate immune response and in cord blood, possess low levels of functional IFN receptors, the number of which increase during differentiation of myeloid progenitors in culture. These results suggest that IFN-␣ can Correspondence: P Eid, Laboratoire d’Oncologie Virale, CNRS-UPR exert direct effects on the proliferation of early myeloid 9045, 7, rue Guy Moquet-94801 Villejuif, France; Fax: 33-1 49 58 progenitors. 34 44 Received 10 September 2001; accepted 29 January 2002 Type I interferon receptors in CD34+ cells J Giron-Michel et al 1136 Materials and methods against purified human leukocyte IFN-␣ has been described previously.31 The preparation used in this study had a neu- Isolation of human CD34+ stem cells by affinity tralizing titer of 2 × 105 against 8 IU of recombinant IFN-␣2a. chromatography or by magnetic cell sorting Normal sheep serum was used as a control. (MiniMacs)

Human umbilical cord blood (CB) was collected from full- Colony-forming cell assay term newborns following informed consent of the mother at the Hospital St Vincent de Paul (Paris, France). Cord blood CD34+ cells were seeded in StemGem medium mononuclear cells (CB-MNC) were isolated by density (1.077 (hatzfeldȰvjf.cnrs.fr; CNRS, Villejuif, France). The medium g/ml) gradient centrifugation (2000 rpm, 20 min) using Ficoll– was composed of 0.9% methylcellulose supplemented with Hypaque (Seromed, Berlin, Germany). After centrifugation 30% fetal calf serum (Dominique Deutsher, Brumath, France), CB-MNC present at the interface were harvested and washed 10 mg/ml transferrin, 1 mg/ml detoxified serum albumin + twice with phosphate-buffered saline (PBS). CD34 cells were (Sigma Chemical, St Louis, MO, USA), 2 mM glutamine (Life + − separated using the Ceprate LC CD34 (CellPro, Bothell, Technologies, Cergy-Pontoise, France), 5 × 10 5 M 2-mercap- WA, USA) to yield CD34+ cells with a purity ranging from 65 toethanol (Merck, Darmstadt, Germany), 1.7 units/ml of to 85%. For magnetic cell sorting, the procedure was perfor- recombinant human IL3, 10 units/ml of recombinant human med according to the manufacturer’s instructions. Blocking IL-6, 3 ng/ml of recombinant human SCF (all obtained from reagent and the antibody QBEND/10 from the direct CD34+ R&D Systems), and 18.3 units/ml of recombinant human GM- progenitor cell isolation kit (MACS; Miltenyi Biotec, Paris, CSF (Genetic Institute, Cambridge, MA, USA), 2 units/ml of France) were added simultaneously to the cell suspension, recombinant human EPO (Boehringer, Mannheim, Germany), mixed well and incubated for 30 min at 4°C on a rotating Approximately 400 enriched CD34+ cells were re-suspended wheel. After incubation, cells were washed with PBE in 0.4 ml Iscove’s modified Dulbecco’s medium (IMDM) sup- (phosphate-buffered saline (PBS), 0.5% bovine serum albumin plemented with 10% FCS and mixed with 2 ml of StemGem (BSA), 2 mM EDTA). The cells were re-suspended in 1 ml PBE methylcellulose and 1 ml of this mixture was plated in 35- and applied to the pre-filled column which was placed in a mm Petri dishes. Cultures were incubated for 21 days in 5%

magnetic field. Cells were centrifuged (1500 r.p.m., 10 min, CO2 in air and 100% humidity. Each culture was plated in 4°C) and re-suspended with PB buffer (PBS, 0.5% BSA). Only triplicate. The number of colonies containing more than 50 CD34+ progenitor cells were specifically retained by the col- cells was evaluated on days 7 and 14 with an inverted micro- umn matrix and subsequently eluted by removing the column scope. CFU-GM and BFU-E colonies were determined after from the magnetic field and flushing with PB. Using this iso- 14 days of incubation. lation procedure, the purity of CD34+ cells was routinely greater than 95%. The purity of CD34+ cells was determined by incubating cells for 15 minutes at 4°C with PerCP-labeled Liquid suspension culture of day 11 progenitor cells anti-CD34+ antibody (HPCA-2, clone 8G12, Becton Dickin- son, San Jose, CA, USA). Background staining was determined Purified CD34+ cells were seeded at a concentration of 103/ml using an irrelevant control PE-IgG1 (Coultronics, Margency, in StemGem liquid medium and incubated at 37°C with 5%

France), or Per-IgG1 (Becton Dickinson) for MACS. Pheno- CO2 in air for 11 days. The composition of the StemGem typic analysis was performed using FACscan analyser liquid medium was the same as for semi-solid medium except (Becton Dickinson). that methylcellulose was omitted. Cells were cultured in six- well flat-bottomed tissue culture plates (Falcon, Becton Dickinson). Interferons and antibodies

Recombinant human IFN-␣2a with a specific activity of Radiolabeling of the monoclonal antibody approximately 2 × 108 IU/mg was obtained from Hoffman LaRoche (Meylan, France). Recombinant human IFN-␥ with a Affinity purified 64G12 mAb was radiolabeled using a modi- specific activity of approximately 107 IU/mg was obtained fied chloramine-T method as described previously 28.28,29 from R&D Systems (Mineapolis, MN, USA). Briefly, 30 ␮g of the antibody in 0.2 ml PBS was incubated The 64G12 mAb was prepared by immunizing mice with with 10 ␮l of chloramine-T (Sigma), 0.5 mg/ml diluted in PBS a recombinant protein corresponding to the extra-cellular and 1 mCi (10 ␮l) of 125I-Na. After 20 s, the reaction was domain of the IFNAR1 chain of the human IFN-␣/␤ receptor quenched by addition of 30 ␮l of 1 mg/ml filtered sodium expressed in mammalian cells as described previously.30 The metabisulfite diluted in PBS containing 10 mM tyrosine 64G12 mAb is a mouse IgG1 which inhibits both the binding (Sigma). Unreacted iodide and other low molecular weight and biological activity of all the human type I interferons reagents were removed by filtration on G-25 Sephadex PD- tested.28 An irrelevant mouse IgG1 was used as a negative 10 columns (Pharmacia Biotech, Saclay, France) equilibrated control. Anti-phospho-STAT1 and anti-phospho-Tyk2, were with PBS, and 0.5% BSA. For equilibrium saturation binding obtained from New England Biolabs (Beverly, MA, USA). Anti- studies, 0.5 or 1.0 × 106 cells were incubated in a final volume STAT1 (clone IF9A7), anti-STAT2 (S21220), were obtained of 0.5 ml with various concentrations of 125I-64G12 mAb. from Pharmingen-BD (San Diego, CA, USA), and BD Trans- After incubation for 2 h at 4°C, cells were washed extensively duction Laboratories (San Jose, CA, USA), respectively. Anti- and bound radioactivity was determined using a gamma phospho-JAK1 (44–422) and anti-IRF1 (C-20) were obtained counter (LKB 1282). Nonspecific binding was determined by from Biosource International (Camarillo, CA, USA), and Santa incubation of cells with the 125I-64G12 mAb in the presence Cruz Biotechnology (Santa Cruz, CA, USA), respectively. of a 200-fold excess of unlabeled 64G12 mAb, and subtracted Asheep polyclonal anti-human IFN- ␣ antibody raised from the total radioactivity. Results are expressed as the mean

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1137 of duplicate determinations. The number of receptors per cific radioactivity than 125I-labeled IFN-␣2, and binding of the cell and their dissociation constants were calculated by 125I-labeled 64G12 mAb has been shown previously to be a Scatchard analysis. sensitive means of detecting IFN-␣/␤ receptor expression on a number of different human cell lines including Daudi, K562, Ly28 and HL60.28 Thus, in the present study, specific binding Immunoblotting of phosphorylated proteins of the 125I-64G12 mAb was determined either alone or in the presence of a 200-fold excess of the unlabeled mAb, using The 64G12 mAb (20 ␮g/ml) was added or not added to pur- freshly isolated CD34+ cells (purity 85% ± 5%) from CB. ified CD34+ cells (0.5 million) in 1 ml DMEM culture medium Experiments were carried out at 4°C until binding equilibrium containing 15% FCS, followed by the addition of 1000 IU/ml was reached (routinely 2 h). IFN-␣2 for 15 min at 37°C. The cell pellet was lysed in 2% The results of two independent representative experiments SDS sample buffer, boiled for 5 min, centrifuged and proteins are shown in Figure 1a. CD34+ cells were found to express a were fractionated by sodium dodecyl sulfate-polyacrylamide single class of high affinity IFN-␣ receptors, with saturation gel (7.5%) electrophoresis (SDS-PAGE) under reducing con- binding occurring at 2 nM. Scatchard analysis revealed that ditions, transferred to polyvinylidene difluoride filters (PVDF, freshly isolated CD34+ cells expressed 740 ± 40 high affinity ± 2 = Boehringer Mannheim) and probed with the appropriate pri- receptors per cell with a Kd of 0.7 0.04 nM (R 0.989), mary-specific antibodies for 1 h at 20°C. Protein–antibody compared with approximately 4300 receptors per cell and a 2 = complexes were then revealed following incubation for 45 Kd of 0.8 nM (R 0.980) for Daudi cells determined under min with a 1/20000 dilution of a horse-radish peroxidase the same experimental conditions (data not shown). (HRP)-conjugated goat anti-mouse or anti-rabbit secondary AsCD34+ cells developed into more differentiated cells antibody (Jackson Immunoresearch, Baltimore, MD, USA). during incubation at 37°C in vitro, specific binding of 125I- PVDF blots were then developed using an enhanced chemi- labeled 64G12 mAb increased markedly and the level of luminescence kit (Amersham Pharmacia Biotech, Bucks, UK). receptor expression rose to some 1440 ± 40 receptors per cell ± 2 = after 11 days in culture, with a Kd of 2.1 0.2 nM (R 0.996) as shown in Figure 1b. These results suggest that specific bind- Confocal microscopy ing sites for type I IFN increase with the differentiation of pro- genitor cells. It was estimated from the number of IFN-␣/␤ Purified CD34+ cells were incubated for 30 min with 100 receptors on CB-MNC that specific binding of 125I-64G12 ␮g/ml of the anti-IFNAR1 mAb (64G12) and then activated mAb to accessory cells present in samples of purified CD34+ with 1000 IU/ml IFN-␣2 for 30 min in FCS-free medium. cells did not exceed 10% of total binding (data not shown). For confocal microscopy of STAT2 and IRF-1, activated cells were fixed and permeabilized with ORTHOpermeafix (Ortho Diagnostic Systems, Raritan, NJ, USA). Cells were then Effect of 64G12 mAb on the inhibitory action of IFN- washed once with PBS and incubated for 30 min with 20 ␣ on CD34+ cell proliferation ␮g/ml of a rabbit polyclonal anti-STAT2 (Transduction Labora- tories, Lexington, KY, USA) or anti-IRF-1 (Santa Cruz In agreement with the results of previous studies,13,14,21 culti- Biotechnology). vation of human CD34+ cells in methylcellulose in the pres- Staining was revealed using 10 ␮g/ml of Alexa FluorTM ence of 1000 IU/ml of IFN-␣ was found to result in 63% ± 488-labeled goat anti-rabbit IgG antibody (Molecular Probes, 5% inhibition of the number of 7 day colonies relative to cul- Montluc¸on, France). Nuclear coloration was performed with tures treated with medium alone (Figure 2). We have shown propidium iodide diluted 1/50, 2 ␮g/ml final concentration, previously that the 64G12 mAb inhibits both the anti-prolifer- (Sigma). Stained cells were washed with PBS, cytocentrifuged ative and antiviral activities of IFN-␣ in a variety of different in a cytospin 3 (Shandon, Pittsburg, PA, USA) and analyzed cell types.28 It was of interest therefore to determine the ability by laser scanning confocal microscopy using a Leica TCS of the 64G12 mAb to neutralize the effect of IFN-␣ on colony Confocal System (Wetzler, Germany). formation by CD34+ cells. Thus, in the presence of 1 ␮g/ml of 64G12 mAb, the inhibition of colony formation observed at 7 days in cultures treated with 1000 IU/ml of IFN-␣ was Results reduced to only 26% ± 5% (Figure 2a). The suppressive effect of the 64G12 mAb was still apparent IFN-␣/␤ receptor expression on freshly isolated CD34+ when the number of CFU-GM and BFU-E was determined cells and day 11 progenitor cells after 14 days in culture (Figure 2b). Thus, in the presence of 1000 IU/ml of IFN-␣ alone the number of CFU-GM and BFU- The results of preliminary experiments showed that the level E was inhibited by 60% ± 5% and 62% ± 5%, respectively, of IFN-␣/␤ receptor expression on freshly isolated CD34+ cells 14 days after treatment, relative to untreated control cultures from human umbilical cord blood was too low to allow quan- (Figure 1b). In contrast, in the presence of the 64G12 mAb, titative Scatchard analysis to be carried out on highly purified the inhibitory effect of 1000 IU/ml of IFN-␣ on CFU-GM and CD34+ cells due to the large number of cells required using BFU-E colony formation was reduced to 20% ± 5% and 28% radio-labeled IFN-␣2, the low number of CD34+ cells present ± 5% inhibition, respectively. in cord blood, and the limited volume of blood available. Treatment of CD34+ cells with 1000 IU/ml of IFN-␥ was This was overcome by the use of a 125I-labeled monoclonal also found to inhibit colony formation at 7 days relative to antibody specific for the human type I IFN receptor, prepared control cultures. Treatment of CD34+ cells with IFN-␥ in the by immunizing animals with a recombinant protein corre- presence of 64G12 mAb did not affect 7 day colony formation sponding to the extra-cellular domain of the IFNAR1 chain of relative to cultures treated with IFN-␥ alone, indicating that the human type I IFN receptor. The 125I-labeled anti-IFNAR1 the effect of the 64G12 mAb was specific for type I IFN monoclonal antibody, 64G12, had a significantly higher spe- (Figure 2a).

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1138

Figure 2 Effect of IFN-␣ on colony formation by CD34+ cells in the absence or presence of the anti-IFN-␣␤ receptor mAb 64G12. Four hundred purified CD34+ cells (purity 80 ± 5%) were incubated in the presence of 1␮g/ml of the 64G12 mAb for 1 h at 37°C then seeded in the presence of IFN-␣ (1000 IU/ml) or IFN-␥ (1000 IU/ml) and plated in methylcellulose medium. The data presented are representa- tive from a total of three independent experiments. Colonies were counted on day 7 (a) and on day 14 to determine the number of CFU- GM and BFU-E (b). The data are presented as the number of colonies formed per 103 cells and expressed as mean ± s.d. of triplicate plates.

Anti-IFNAR1 mAb (64G12) inhibits tyrosine phosphorylation of the STAT1, Tyk2 and JAK1 proteins involved in signal transduction following IFN- ␣ activation

Binding of IFN to its cell surface receptor results in rapid tyro- Figure 1 Analysis of IFN-␣␤ receptor expression on freshly iso- + sine phosphorylation and activation of the various compo- lated CD34 cells and 11 day progenitor cells. Interferon receptor expression was analyzed in freshly isolated CD34+ cells from human nents of the signal transduction pathway leading from the cell cord blood (a), or on progenitor cells after 11 days cultivation of surface to the nucleus. We have shown previously that the CD34+ cells in liquid medium (b), using the 125I-labeled anti-IFN-␣␤ anti-IFNAR1 monoclonal antibody, 64G12, neutralizes the receptor monoclonal antibody, 64G12, as described in the Materials biological activities mediated by IFN-␣, ␤ and ␻28–30 as well and methods. The data presented are representative of two inde- + as the formation of the ISGF3 complex (IFN-stimulated gene pendent experiments for both CD34 and progenitor cells. The purity 32 + ± factor 3) in nuclear extracts from human cell lines. In order of CD34 cells was 85% ( 5%). As shown in the insert of panel a, ␣ CD34+cells were estimated to possess 740 ± 60 binding sites per cell to elucidate the events which follow the binding of IFN- to with a K = 0.7 ± 0.04 nM (R2 = 0.989) by extrapolation of the linear the cell surface and which culminate in inhibition of the pro- d + portion of the Scatchard plot, and 1440 ± 40 binding sites per cell, liferation of CD34 cells, we analyzed the tyrosine phos- = ± 2 = Kd 2.1 0.2 nM (R 0.996) for 11 day progenitor cells as shown phorylation and activation of the various components of the in the insert of panel (b). IFN signal transduction pathway following IFN treatment in

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1139 the presence or absence of the anti-IFNAR1 monoclonal anti- nucleus following IFN treatment of CD34+ cells as shown in body 64G12. It has been shown previously that IFNAR1, Figure 4 (lower panel). Furthermore, the nuclear localization IFNAR2, Tyk2 and JAK1 are all maximally tyrosine phos- of STAT2 is again inhibited by treatment of CD34+ cells with phorylated within 10 min, and STAT1 and STAT2 within 60 the 64G12 mAb which neutralizes the biological activity of min, of contact of cells with IFN-␣.23–27 The tyrosine phos- type I IFN (Figure 4, lower panel). phorylation of STAT1, Tyk2 and JAK1, was therefore analyzed 15 min after treatment of freshly isolated highly purified CD34+ cells (Ͼ95% purity) with IFN-␣ in the presence or not Discussion of the 64G12 mAb. As shown in Figure 3a, tyrosine phos- phorylation is clearly stimulated in extracts of purified CD34+ Interferon-␣ is used extensively for the treatment of myelopro- cells treated with IFN-␣ compared to extracts from excipient liferative disorders, hairy cell leukemia (B cell leukemia) and treated control cells (Figure 3a). Furthermore, tyrosine phos- chronic myeloid leukemia.22,33–35 Although numerous studies phorylation of STAT1, Tyk2 and JAK1, is effectively abrogated have shown that IFN-␣ inhibits the proliferation of leukemic in extracts of purified CD34+ cells treated with IFN-␣ in the progenitor cells,15–19 relatively little is known, however, about presence of the 64G12 mAb (Figure 3a). As shown in Figure the effect of IFN-␣ on the proliferation of hematopoietic pro- 3b the quantities of protein applied were comparable for all genitor cells from normal individuals,36 and in particular the cell extracts. Similar results were also obtained using whether the proliferation of such cells is influenced directly extracts of highly purified CD34+ cells isolated from bone by IFN, or indirectly by cytokines and growth factors induced marrow (data not shown). by the action of IFN on bone marrow stromal cells. Interferons exert their characteristic biological effects by binding to high-affinity receptors on the cell surface resulting Nuclear localization of IFN-␣ activated STAT2 and in rapid tyrosine phosphorylation and activation of STAT pro- IRF1 teins which regulate the transcription of IFN responsive genes resulting in the characteristic biological activities of the inter- IFN signal transduction culminates in nuclear translocation of ferons. Thus, IFN sensitivity is determined by receptor an activated transcription complex containing a number of expression, and the biological activity of IFN-␣ is proportional transcription factors including IRF1 and STAT2. To assess the to the number of IFN-␣ receptors expressed on the cell sur- effects of IFN-␣ on the nuclear translocation of activated tran- face.37–39 It was of interest, therefore to determine whether scription factors in CD34+ cells, cells were treated with 1000 freshly isolated CD34+ cells from human cord blood express IU/ml of IFN-␣ in the presence or absence of the 64G12 mAb, detectable levels of IFN receptors, and if so whether such and analyzed by confocal microscopy as described in receptors are functional. Thus, we have shown for the first Materials and methods. time that freshly isolated CD34+ cells from human cord blood The nuclear localization of IRF1 following treatment of express approximately 740 ± 60 IFN-␣/␤ receptors per cell, CD34+ cells with IFN-␣ is shown in Figure 4. The upper panel compared with 4300 IFN-␣/␤ receptors per cell detected on shows that the nuclear localization of IRF1 in IFN-␣-treated the human B cell line Daudi using either radiolabeled IFN, or CD34+ cells is inhibited in the presence of the 64G12 mAb. the same 125I-64G12 binding assay used in this study (P Eid, Similarly, STAT2, which is specific to the transduction of the unpublished results). The number of IFN-␣/␤ receptors type I IFNs, and which forms a complex with STAT1 and IRF9 detected on freshly isolated CD34+ cells is significantly greater (p48 DNAbinding protein) is clearly translocated to the than the 36 ± 1 GM-CSF receptors per cell,40 and the absence

Figure 3 Inhibition of tyrosine phosphorylation of the components of the signaling pathway in CD34+ cells following IFN-␣ activation in the presence of the 64G12 mAb. Highly purified (Ͻ95% purity) freshly isolated (day 0) CD34+ cells (5 × 105 per lane) were treated (+) or not (−) with anti-IFNAR1 mAb 64G12 (100 ␮g/ml) followed by the addition of IFN-␣2 (1000 IU/ml) for 15 min at 37°C. The cells were lysed under reducing conditions and protein extracts were separated by SDS/7.5% PAGE, transferred to PVDF membranes and probed with anti-phospho- STAT1, anti-phospho-Tyk2 and anti-phospho-JAK1 antibodies (panel a). After stripping, the filters were re-probed individually with the anti- STAT1, anti-Tyk2 and anti-JAK1 antibodies in order to demonstrate that similar amounts of proteins were present in all lanes (panel b). Immuno- reactive bands were visualized by ECL Western blotting system (Amersham, UK).

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1140

Figure 4 Effect of IFN-␣2 on nuclear translocation of STAT2 and IRF1 in CD34+ cells. Freshly isolated (day 0), and highly purified (Ͼ95%) CD34+ cells were fixed, permeabilized, and stained with anti-STAT2, or anti-IRF-1 (green spots) and propidium iodide (for nucleal staining, red spots). Yellow spots indicate nuclear localization. (a) Resting CD34+ cells; (b) CD34+ cells 30 min after IFN-␣2 activation; (c) CD34+ cells 30 min after treatment with neutralizing anti-IFNARI antibodies followed by 30 min treatment with IFN-␣2.

of detectable levels of the CXC receptor for the when CD34+ cells were treated with IFN-␣ in the presence of IFN-␥-inducible protein IP 10 and IFN-␥-inducible monokine a neutralizing antibody specific for the human type I IFN Mig,41 reported previously for freshly isolated CD34+ cells. receptor. We have also shown using confocal microscopy that Primitive progenitor cells can be induced to proliferate and phosphorylation of STAT proteins in freshly isolated (day 0) differentiate in vitro by the supply of different combinations of CD34+ cells is followed by nuclear translocation of a complex cytokines.42 Thus, progenitor cells derived from CD34+ cells of transcription factors containing STAT2 and IRF1, preceding cultivated for 11 days in liquid suspension culture were found transcriptional activation of IFN responsive genes and the to express some 1440 ± 40 IFN-␣/␤ receptors per cell. establishment of the biological effects of IFN-␣. These results Although this is only a two-fold increase in receptor contrast with those obtained for GM-CSF by Wheadon et al,40 expression relative to the levels detected on freshly isolated who reported that freshly isolated (day 0) CD34+ cells which CD34+ cells. FACS analysis revealed that only 2.2% of the express very low numbers of GM-CSF receptors, are hypores- progenitor cells were found to express the CD34+ marker after ponsive to GM-CSF as determined by activation of down- 11 days in liquid suspension culture (data not shown). The stream signaling. Thus, JAK 2 activation was detected in one number of IFN-␣/␤ receptors seemed to increase during the of four experiments only and the level of STAT 5 activation differentiation of CD34+ cells. Although the biological signifi- obtained was relatively low, and there was no apparent acti- cance of these data can hardly be ascertained, these results vation of MAP kinase by GM-CSF. Indeed, CD34+ did not are in keeping with previous reports that IFN-␣/␤ receptor exhibit maximal downstream signaling from the GM-CSF expression increases during the course of and receptor until some 3 days in culture. megakaryocyte differentiation in vitro.43,44 Similarly, in vitro The detection of significant numbers of functional IFN-␣/␤ maturation of CD34+ progenitors is accompanied by an receptors on freshly isolated (day 0) and highly purified increase in the number of GM-CSF receptors per cell while (Ͼ95% purity) CD34+ cells, and the abrogation of the inhibi- 40 ␣ the Kd remains unchanged. The CXC tory effect of IFN- on the formation of 7 or 14 day CFU- for the IFN-␥-inducible proteins IP 10 and Mig also becomes GM and BFU-E colonies when CD34+ were cultivated in the detectable after cultivation of CD34+ cells in vitro in the pres- presence of an anti-IFN-␣/␤ receptor monoclonal antibody, is ence of GM-CSF.41 The number of IFN-␣/␤ receptors detected in favor of a direct action of IFN-␣ on the proliferation of on 11 day CD34+ progenitors is somewhat lower, however, CD34+ cells via receptor-mediated signal transduction without than the 1910 to 2070 IFN-␥ receptors per cell reported excluding an indirect effect on the induction of other cyto- previously for 6 day erythroid colony-forming cells.45 kines or growth factors by residual accessory cells. In common with numerous other cytokines binding of IFN- ␣ to its cell surface receptor is followed by rapid tyrosine phosphorylation and activation of specific STAT proteins by Acknowledgements Janus kinases.26,46 Activated STAT proteins then dimerize, and translocate to the nucleus leading to activated transcription of This work was supported by grants from Corixa Inc, and the a specific set of IFN-responsive genes.6,25 Thus, treatment of Association Nouvelles Recherches Biome´dicales. freshly isolated and highly purified CD34+ cells with IFN-␣ was shown to result in rapid tyrosine phosphorylation of References STAT1, Tyk2 and JAK1 indicating that freshly isolated (day 0) + CD34 cells possess functional type I IFN receptors. The speci- 1 Kurata H, Arai T, Yokota T, Arai K. Differential expression of gra- ficity of this activation was demonstrated by its abrogation nulocyte– colony-stimulating factor and IL-3 receptor

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1141 subunits on human CD34+ cells and leukemic cell lines. J Allergy chronic myeloid leukemia primitive progenitor cells to bone mar- Clin Immunol 1995; 96: 1083–1099. row stromal cells. Blood 1991; 78: 499–505. 2 McKinstry WJ, Li CL, Rasko JE, Nicola NA, Johnson GR, Metcalf 20 Aman MJ, Keller U, Derigs G, Mohamadzadeh M, Huber C, Pes- D. receptor expression on hematopoietic stem and pro- chel C. Regulation of cytokine expression by interferon-alpha in genitor cells. Blood 1997; 89: 65–71. human bone marrow stromal cells: inhibition of hematopoietic 3 Majika M, Janowska-wieczorek A, Ratajczak J, Ehrenman K, growth factors and induction of -1 receptor antagonist. Pietrzkowski M, Kowalska MA, Gewirtz AM, Emerson SG, Blood 1994; 84: 4142–4150. Ratajczak MZ. Numerous growth factors, cytokines, and chemo- 21 Means RT Jr, Krantz SB. Inhibition of human erythroid colony- kines are secreted by human CD34+ cells, myeloblasts, erythro- forming units by interferons alpha and beta: differing mechanisms blasts and megakaryoblasts and regulate normal hematopoiesis in despite shared receptor. Exp Hematol 1996; 24: 204–208. an autocrine/paracrine manner. Blood 2001; 97: 3075–3085. 22 Peschel C, Aulitzky WE, Huber C. Influence of interferon-alpha 4 De Maeyer E, De maeyer-Guignard J. Interferons and Other Regu- on cytokine expression by the bone marrow microenvironment- latory Cytokines. John Wiley: New York, 1988, pp 69–85. impact on treatment of myeloproliferative disorders. Leuk Lym- 5 Pestka S, Langer JA, Zoon KC, Samuel CE. Interferons and their phoma 1996; 22 (Suppl. 1): 129–134. actions. Annu Rev Biochem 1987; 56: 727–777. 23 Velazquez L, Fellous M, Stark GR, Pellegrini S. Aprotein tyrosine 6 Stark GR, Kerr IM, Williams BRG, Silverman RH, Schreiber RD. kinase in the interferon alpha/beta signaling pathway. Cell 1992; How cells respond to interferons. Annu Rev Biochem 1998; 67: 70: 313–322. 227–264. 24 Uze G, Lutfalla G, Mogensen KE. Alpha and beta interferons and 7 Allan NC, Richards SM, Shepherd PC. UK Medical Research their receptor and their friends and relations. J Interfer Cytok Res Council randomised, multicentre trial of interferon- alpha for 1995; 15: 3–26. chronic myeloid leukaemia: improved survival irrespective of 25 Darnell JE Jr, Kerr IM, Stark GR. Jak-STAT pathways and transcrip- cytogenetic response. The UK Medical Research Council’s Work- tional activation in response to IFNs and other extracellular sig- ing Parties for Therapeutic Trials in Adult Leukaemia. Lancet 1995; naling proteins. Science 1994; 264: 1415–1421. 345: 1392–1397. 26 Ihle JN, Kerr IM. Jaks and Stats in signaling by the cytokine recep- 8 Messora C, Bensi L, Vecchi A, Longo R, Giacobbi F, Temperani tor superfamily. Trends Genet 1995; 11: 69–74. P, Bevini M, Emilia G, Sacchi S. Cytogenetic conversion in a case 27 Horvath CM, Stark GR, Kerr IM, Darnell JE. Interactions between of polycythaemia vera treated with interferon-alpha. Br J Haematol STAT and non-STAT proteins in the interferon-stimulated gene fac- 1994; 86: 402–404. tor 3 transcription complex. Mol Cell Biol 1996; 16: 6957–6964. 9 Kantarjian HM, Talpaz M. Long-term follow-up results of alpha 28 Benoit P, Maguire D, Plavec I, Kocher H, Tovey M, Meyer F. A interferon therapy in chronic myelogenous leukemia at MD Ander- monoclonal antibody to recombinant human IFN-alpha receptor son Cancer Center. Leuk Lymphoma 1993; 11: 169–174. inhibits biologic activity of several species of human IFN-alpha, 10 Gresser I, Maury C, Kaido T, Bandu MT, Tovey MG, Maunoury IFN-beta, and IFN-omega. Detection of heterogeneity of the cellu- MT, Fantuzzi L, Gessani S, Greco G, Belardelli F. The essential lar type I IFN receptor. J Immunol 1993; 150: 707–716. role of endogenous IFN alpha/beta in the anti-metastatic action of 29 Eid P, Langer JA, Bailly G, Lejealle R, Guymarho J, Tovey MG. sensitized T lymphocytes in mice injected with Friend erythroleu- Localization of a receptor nonapeptide with a possible role in the kemia cells. Int J Cancer 1995; 63: 726–731. binding of the type I interferons. Eur Cytok Netw 2000; 11: 11 Tarumi T, Sawada K, Sato N, Kobayashi S, Takano H, Yasukouchi 560–573. T, Takashashi T, Sekiguchi S, Koike T. Interferon-alpha-induced 30 Eid P, Tovey MG. Characterization of a domain of a human type apoptosis in human erythroid progenitors. Exp Hematol 1995; 23: I interferon receptor protein involved in ligand binding. J Interfer 1310–1318. Res 1995; 15: 205–211. 12 Klimpel GR, Fleischmann WR Jr, Klimpel KD. Gamma interferon 31 Mogensen KE, Pyhala L, Cantell K. Raising antibodies to human (IFN gamma) and IFN alpha/beta suppress murine myeloid colony leukocyte interferon. Acta Pathol Microbiol Scand [B] 1975; 83: formation (CFU-C)N: magnitude of suppression is dependent upon 443–450. level of colony-stimulating factor (CSF). J Immunol 1982; 129: 32 Benizri E, Gugenheim J, Lasfar A, Eid P, Blanchard B, Lallemand 76–80. C, Tovey MG. Prolonged allograft survival in cynomolgus monk- 13 Broxmeyer HE, Lu L, Platzer E, Feit C, Juliano L, Rubin BY. Com- eys treated with a monoclonal antibody to the human type I inter- parative analysis of the influences of human gamma, alpha and feron receptor and low doses of cyclosporine. J Interfer Cytok Res beta interferons on human multipotential (CFU-GEMM), erythroid 1998; 18: 273–284. (BFU-E) and granulocyte–macrophage (CFU-GM) progenitor cells. 33 Talpaz M, Kantarjian H, Kurzrock R, Gutterman JU. Interferon J Immunol 1983; 131: 1300–1305. alpha in the therapy of CML. Br J Haematol 1991; 79 (Suppl. 1): 14 Mazur EM, Richtsmeier WJ, South K. Alpha-interferon: differential 38–41. suppression of colony growth from human erythroid, myeloid, and 34 Mahon FX, Montastruc M, Faberes C, Reiffers J. Predicting com- megakaryocytic hematopoietic progenitor cells. J Interfer Res plete cytogenetic response in chronic myelogenous leukemia 1986; 6: 199–206. patients treated with recombinant interferon alpha. Blood 1994; 15 Carlo-Stella C, Cazzola M, Gasner A, Barosi G, Dezza L, Meloni 84: 3592–3594. F, Pedrazzoli P, Hoelzer D, Ascari E. Effects of recombinant alpha 35 Hehlmann R, Heimpel H. Current aspects of drug therapy in Phila- and gamma interferons on the in vitro growth of circulating hema- delphia-positive CML: correlation of tumor burden with survival. topoietic progenitor cells (CFU-GEMM, CFU-Mk, BFU-E, and Leuk Lymphoma 1996; 22 (Suppl. 1): 161–167. CFU-GM) from patients with myelofibrosis with myeloid meta- 36 Thiele J, Wickenhauser C, Neuwirth C, Schulze HJ, Flucke U, plasia. Blood 1987; 70: 1014–1019. Kvasnicka HM, Borchmann P, Krech R, Fischer R. Effect of IFN- 16 Ganser A, Carlo-Stella C, Greher J, Volkers B, Hoelzer D. Effect alpha on normal human hematopoiesis: an immunohistochemical of recombinant interferons alpha and gamma on human bone and morphometric study on trephine biopsy specimens. J Interfer marrow-derived megakaryocytic progenitor cells. Blood 1987; 70: Cytok Res 1998; 18: 247–253. 1173–1179. 37 Fukuda R, Ishimura N, Kushiyama Y, Moriyama N, Ishihara S, 17 Geissler D, Gastl G, Aulitzky W, Tilg H, Gaggl S, Konwalinka G, Nagasawa S, Miyake T, Niigaki M, Satoh S, Sakai S, Akagi S, Wat- Huber C. Recombinant interferon-alpha-2C in chronic myelogen- anabe M, Fukumoto S. Effectiveness of interferon-alpha therapy in ous leukaemia: relationship of sensitivity of committed haemato- chronic hepatitis C is associated with the amount of interferon- poietic precursor cells in vitro (BFU-E, CFU-GM, CFU-Meg) and alpha receptor mRNAin the liver. J Hepatol 1997; 26: 455–461. clinical response. Leuk Res 1990; 14: 629–636. 38 Yatsuhashi H, Yamasaki K, Aritomi T, Delcarmen PM, Inoue O, 18 Mahon FX, Chahine H, Barbot C, Pigeonnier V, Jazwiec B, Reiffers Koga M, Yano M. Quantitative analysis of interferon alpha/beta J, Ripoche J. All-trans retinoic acid potentiates the inhibitory receptor mRNAin the liver of patients with chronic hepatitis C: effects of interferon alpha on chronic myeloid leukemia progeni- correlation with serum hepatitis C virus RNAlevels and response tors in vitro. Leukemia 1997; 11: 667–673. to treatment with interferon. J Gastroenterol Hepatol 1997; 12: 19 Dowding C, Guo AP, Osterholz J, Siczkowski M, Goldman J, Gor- 460–467. don M. Interferon-alpha overrides the deficient adhesion of 39 Mizukoshi E, Kaneko S, Yanagi M, Ohno H, Kaji K, Terasaki S,

Leukemia Type I interferon receptors in CD34+ cells J Giron-Michel et al 1142 Shimoda A, Matsushita E, Kobayashi K. Expression of interferon IL-6, IL-3, interferon-gamma, and . Blood 1993; 81: alpha/beta receptor in the liver of chronic hepatitis C patients. J 2579–2584. Med Virol 1998; 56: 217–223. 43 Martyre MC, Wietzerbin J. Characterization of specific functional 40 Wheadon H, Roberts PJ, Watts MJ, Linch DC. Changes in signal receptors for HuIFN-alpha on a human megakaryocytic cell line transduction downstream from the granulocyte–macrophage col- (Dami): expression related to differentiation. Br J Haematol 1994; ony-stimulating factor receptor during differentiation of primary 86: 244–252. hemopoietic cells. Exp Hematol 1999; 27: 1077–1086. 44 Fantuzzi L, Eid P, Malorni W, Rainaldi G, Gauzzi MC, Pellegrini 41 Jinquan T, Quan S, Jacobi HH, Jing C, Millner A, Jensen B, Madsen S, Belardelli F, Gessani S. Post-translational up-regulation of the HO, Ryder LP, Svejgaard A, Malling HJ, Skov PS, Poulsen LK. CXC cell surface-associated alpha component of the human type I inter- chemokine receptor 3 expression on CD34(+) hematopoietic pro- feron receptor during differentiation of peripheral blood mono- genitors from human cord blood induced by granulocyte–macro- cytes: Role in the biological response to type I interferon. Eur J phage colony-stimulating factor: and adhesion Immunol 1997; 27: 1075–1081. induced by its ligands, -inducible protein 10 and 45 Taniguchi S, Dai CH, Krantz SB. Specific binding of interferon- monokine induced by interferon gamma. Blood 2000; 96: gamma to high affinity receptors on human erythroid colony-for- 1230–1238. ming cells. Exp Hematol 1997; 25: 193–198. 42 Brugger W, Mocklin W, Heimfeld S, Berenson RJ, Mertelsmann 46 Taniguchi T. Cytokine signaling through nonreceptor protein tyro- R, Kanz L. Ex vivo expansion of enriched peripheral blood CD34+ sine kinases. Science 1995; 268: 251–255. progenitor cells by , interleukin-1 beta (IL-1 beta),

Leukemia