JURL-MK1 (C-Kithigh/CD30−/CD40−) and JURL-MK2 (C

Home , CD135, CD30, CD33

JURL-MK1 (c-kithigh/CD30−/CD40−) and JURL-MK2 (c-kitlow/CD30+/CD40+) cell lines: ‘two-sided’ model for investigating leukemic megakaryocytopoiesis R Di Noto1, L Luciano2, C Lo Pardo3, F Ferrara4, F Frigeri5, O Mercuro1, ML Lombardi1, F Pane5, C Vacca3, C Manzo1, F Salvatore5, B Rotoli2 and L Del Vecchio3

1Divisione di Oncologia Sperimentale C, Istituto Nazionale dei Tumori; 2Divisione di Ematologia Clinica, Universita` Federico II; 3Servizio di Immunoematologia and 4Divisione di Ematologia, Ospedale A Cardarelli; 5CEINGE-Biotecnologie Avanzate, Dipartimento di Biochimica e Biotecnologie Mediche, Universita` Federico II, Naples, Italy

Two novel cell lines (JURL-MK1 and JURL-MK2) have been esis.3,4 We report here the establishment of two novel Ph+ MK- established from the peripheral blood of a patient in the blastic oriented cell lines, JURL-MK1 and JURL-MK2, obtained from phase of chronic myelogenous leukemia. The cells grow in a single cell suspension with doubling times of 48 h (JURL-MK1) a single patient affected by CML-BC. We show that JURL-MK1 and 72 h (JURL-MK2). Cytogenetic analysis has shown that and JURL-MK2 are characterized by divergent phenotypes, JURL-MK1 is hypodiploid whereas JURL-MK2 is near triploid karyotypes and responses to cell growth/differentiation and that both cell lines retain t(9;22). Moreover, JURL-MK1 and promoters. JURL-MK2 have a bcr/abl-fused gene with the same junction found in the patient’s fresh cells, and both cell lines express the b3/a2 type of hybrid bcr/abl mRNA. The morphology and immunophenotype of these cell lines are reminiscent of megak- Materials and methods aryoblasts. In both lines, a limited but consistent percentage of cells expresses gpIIbIIIa (CD41a), gpIIIa (CD61) and CD36, Case history with no expression of gpIb (CD42b), glycophorin A, hemoglobin and CD34. Both cell lines are clearly positive for CD33, CD43, In December 1983, a 63-year-old male patient was admitted CD45RO and CD63, while CD13, CD44, CD54, CD30 and CD40 to the Hematology Department for leukocytosis and spleen are specific features of JURL-MK2. Among cytokine receptors, + CD117/SCF-R is strongly displayed by a large fraction of JURL- enlargement. He was diagnosed as having Ph CML and was MK1 cells but is hardly detectable on about 20% JURL-MK2 treated with busulfan until July 1987, when he was switched cells. Both cell lines are clearly positive for CD25/IL2R␣, while to hydroxyurea. In November 1993, he complained of general a marked expression of CD116/GM-CSF-R and CDw123/IL3R␣ malaise and showed progressive unresponsive splenomegaly. is restricted to JURL-MK2. Induction of cell differentiation in Peripheral blood counts were: Hb 8.7 g/dl, platelets 22 000/␮l vitro has demonstrated that TPA is able to modulate the JURL- and WBC 140 000/␮l with 65% blast cells. Bone marrow MK1 phenotype, causing an increased expression of platelet- associated antigens. The JURL-MK2 phenotype is easily modu- examination revealed more than 90% blast cells with myeloid lated by both TPA and DMSO, which cause an increased appearance, which was confirmed by a cell surface marker + + + + + ± expression of CD41a and CD117 accompanied by a decreased study (CD33 , CD36 , CD45RO , CD43 , CD117 , CD25 , expression of CD30. Proliferation studies demonstrated that CD34−, HLA-DR−, CD11a−). Chromosome analysis confirmed JURL-MK1 cell growth is enhanced by stem cell factor, while the Ph+ karyotype; no additional abnormalities were found. JURL-MK2 proliferation is unaffected by this cytokine. JURL- A diagnosis of myeloid BC was made. The patient died of MK1 and JURL-MK2 are two novel cell lines with divergent biological features, representing a ‘two-sided’ model for complications (heart failure) a few days later. investigating new aspects of megakaryocytopoiesis. Keywords: megakaryoblastic cell lines; megakaryocytopoiesis; CD30; c-kit; CD117 Establishment of cell lines

Mononuclear cells from a peripheral blood sample drawn dur- Introduction ing the blastic phase (November 1993) were separated by Fic- oll–Hypaque (d = 1.077) density gradient centrifugation at Chronic myelogenous leukemia (CML) is characterized by 850 g for 20 min and seeded at 2 × 106/ml in 35-mm Petri progressive accumulation of differentiated blood cells derived dishes, containing 2 ml Dulbecco’s modified Eagle’s medium from a single pluripotent progenitor cell carrying the so-called (DMEM) with 4.5 g/l dextrose (ICN, Oxford, UK), sup- Philadelphia (Ph) chromosome. Virtually all CML patients plemented with 20% fetal bovine serum (FBS; Gibco, Life − enter the blastic crisis (BC) stage, in which the malignant cells Technologies, Paisley, UK), 2 mML-glutamine (Gibco), 10 5 M may show characteristics of various cell types, including those 2-mercaptoethanol and antibiotics. The cells were incubated 1 ° of the megakaryocytic lineage. at 37 C in a humidified 5% CO2 atmosphere and fresh DMEM Cell lines derived from patients with CML-BC may be useful was added at 7-day intervals, replacing half of the medium for investigating hematopoietic cell growth and differentiation, each time. Neither growth factors nor exogenous feeder layers providing information on the action of biological agents in were used. CML cells, as well as on the cascade of molecular events asso- In January 1994, the cells began to grow slowly, with large ciated with this myeloproliferative disorder.2 round cells proliferating on a substrate of fibroblast-like cells. Due to the scarceness of megakaryocytes (MK) in normal When the cells showed a doubling time of 7 days (May 1994), bone marrow, the establishment of MK cell lines can provide they were transferred to tissue culture flasks (Falcon; Becton models in order to perform studies on megakaryocytopoi- Dickinson, San Jose, CA, USA), and, within a few weeks (July 1994), a series of 12 subcultures were obtained. In this phase, the fibroblast-like cells disappeared. Four months later Correspondence: L Del Vecchio, Via Manzoni 24, 80123 Naples, Italy (November 1994), two subcultures began to grow vigorously, Received 20 January 1997; accepted 12 May 1997 and two cell lines (JURL-MK1 and JURL-MK2) were judged to JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1555 be established. The cells grew in single cell suspensions, and CSF-R), hIL-1R-M1 (CD121a, IL1R type 1), hIL-1R2-M22 the later passages showed doubling times of 48 h (JURL-MK1) (CD121b, IL1R type 2), 9F5 (CDw123, IL3R␣), S456C9 and 72 h (JURL-MK2). The cells grew at an optimal cell den- (CD124, IL4R), R34.34 (CD127, IL7R), 3B5 (CD132, IL2R ␶ sity of 1.5 × 106 (minimum: 8 × 105 cells/ml; maximum: chain), 4G8 (CD135, Flt3). As negative controls, IgG1, IgG2a 2.5 × 106 cells/ml). JURL-MK1 was split 1:2 every 48 h or 1:3 and IgG2b immunoglobulin, purchased from Immunotech every 72 h, while JURL-MK2 was split 1:2 every 72 h. and Ortho Diagnostic Systems, were utilized in direct and Aliquots of both JURL-MK1 and JURL-MK2 cells were cryo- indirect immunofluorescence experiments. As second step preserved in FBS/DMSO (6:1). Although we could maintain reagent for indirect immunofluorescence, a FITC-conjugated both JURL-MK1 and JURL-MK2 cells for 3 years in stable con- goat anti-mouse antiserum (Ortho Diagnostic Systems) was dition, the cryopreserved cells were routinely thawed and cul- employed. tured for further characterization of the cell lines. In all immunofluorescence experiments, non-specific bind- A test for mycoplasma contamination (Mycoplasma PCR ings due to Fc receptors were avoided by pre-incubating the ELISA; Boehringer Mannheim, Mannheim, Germany) was per- cells at 4°C with rabbit immunoglobulin. formed at regular intervals, but no contamination was found Intracellular hemoglobin was detected by intracytoplasmic in either cell culture. indirect immunofluorescence after cell permeabilization Cell lines were established at the Divisione di Oncologia (Becton Dickinson permeabilizing solution), using a rabbit Sperimentale C, Istituto Nazionale per lo Studio e la Cura dei polyclonal antiserum (Dako) and a FITC-conjugated swine Tumori, Fondazione G Pascale, Naples, Italy. anti-rabbit antiserum (Dako) as second step reagent. The findings obtained for individual cell lines were expressed as a percentage of positive cells with the back- Morphology, cytochemistry and cell marker analysis ground subtracted. The mean channel of fluorescence intensity of histograms characterized by unimodal distribution was Cytospin preparations of JURL-MK1 and JURL-MK2 cells were directly obtained from the cytometer and recorded. stained with May–Gru¨nwald–Giemsa and morphology was evaluated under a light microscope. Cytochemical stainings were performed for myeloperoxidase (MPO), alkaline phos- Karyotype and bcr rearrangement analysis phatase (AP), ␣-naphtyl acetate esterase (ANAE) and periodic acid-Schiff (PAS). Chromosome analysis was performed as follows: the cultured Cell surface antigens were tested by flow cytometry cells were treated with 75 mM hypotonic KCl solution for (Cytoron Absolute; Ortho Diagnostic Systems, Raritan, NJ, 30 min at 37°C and then fixed with methanol/acetate (3:1) sol- USA). Cell viability was routinely assessed by analysis of light ution. The chromosomes were banded by tripsin-Giemsa scatter properties. Living cells were also scored daily under a staining.5 phase contrast microscope. The rearrangements of M-bcr were tested in DNA samples The membrane expression of differentiation antigens was obtained from the patient’s bone marrow cells and from the investigated with a panel of monoclonal antibodies (MoAb) cell lines by Southern analysis, using a standard technique.6 directly coupled with fluorescein (FITC) or phycoerythrin (PE), Briefly, 15 ␮g of genomic DNA was digested with a three-fold including OKT11-FITC (CD2), Leu9-FITC (CD7), OKB19-FITC excess of BglII, HindIII, EcoRI or BamHI restriction enzymes (CD19), OKB2-FITC (CD24), LeuM9-PE (CD33), LeuM7-PE (New England Biolabs, Beverley, MA, USA), electrophoresed (CD13), OKM5-FITC (CD36), Leu15-PE (CD11b), LeuM5-PE on an 0.8% agarose gel, blotted on positively charged nylon (CD11c), 17F11-PE (CD117), anti-GPA-FITC (glycophorin-A), membranes (Hybond N-plus, Amersham, Buckingham, UK), anti-HLA-DR-PE (HLA-DR), IOM34-PE (CD34), HPCA-2-PE and hybridized to 50 ng of ␣32P labelled 0.7 kb HindIII- (CD34), anti-IIIa-FITC (CD61), IOP41a-FITC (CD41a), BamHI genomic probe encompassing exons 2 and 3 of the IOP42b-FITC (CD42b), IOP62-FITC (CD62P), IOP63-FITC M-bcr region of the bcr gene.7 (CD63), FMC8-FITC (CD9), Leu8-PE (CD62L), Leu44-FITC A method based on reverse-transcriptase polymerase chain (CD44), IOT16-FITC (CD11a), Leu19-PE (CD56), IOL54-FITC reaction (RT-PCR) was used to detect the presence of hybrid (CD54), 2H4-PE (CD45RA), UCHL1-PE (CD45RO), Ki-1-FITC bcr/abl transcripts in the patient’s blood and in the cell line (CD30), IOT14-PE (CD25), Leu23-FITC (CD69) and IL2R␤Ch- samples.8 FITC (CD122). Antibodies within the ‘OK’ series were obtained from Ortho Diagnostic Systems. ‘IO’ monoclonals, along with 17F11 and anti-GPA were obtained from Immuno- In situ hybridization (ISH) for Epstein–Barr virus (EBV) tech (Marseille, France). ‘Leu’ panel, HPCA-2 and anti-HLA- nuclear RNA DR were purchased from Becton Dickinson. Anti-IIIa, UCHL1 and Ki-1 were obtained from DAKO (Glostrup, Denmark). Detection of the two nuclear RNAs encoded by the EBV FMC8 was obtained from Seralab (Crawley Down, UK), (EBER) was performed with an ISH kit (Dako) used in accord- IL2R␤Ch from CLB (Amsterdam, The Netherlands) and 2H4 ance with the manufacturer’s instructions. Cytospins of the from Coulter Electronics (Hialeah, FL, USA). CD43 was stud- cell lines were fixed in 4% formalin, air-dried, rehydrated in ied by indirect immunofluorescence, using Leu22 (Becton pure water and treated with proteinase K diluted in tris-buff- Dickinson). CD40, CD126 (IL6-R␣) and CD130 (IL6R␤, ered saline (TBS). Following two washings in pure water and gp130) were also studied by indirect immunofluorescence immersion in 95% ethanol, the slides were air-dried and incu- using, respectively, the MoAbs G28-5, MT-18 and AM64, bated at 55°C for 30 min with a FITC-conjugated EBV peptide obtained at the V International Workshop on Leukocyte Dif- nucleic acid (PNA) probe complementary to the nuclear EBER. ferentiation Antigens. Other cytokine receptors were studied FITC-conjugated PNA probe directed against glyceraldehyde by indirect immunofluorescence using MoAbs obtained at the 3-phosphate dehydrogenase cellular RNA and FITC-conju- VI International Workshop on Leukocyte Differentiation Anti- gated random PNA probes were used as positive and negative gens. In particular, the MoAbs used were SCO4 (CD116, GM- controls, respectively. Following the hybridization step, the JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1556 slides were immersed in preheated stringent wash solution at washed twice with DMEM and cultured for 4 days at a density 55°C for 25 min with shaking. Then, the slides were incubated of 1 × 106/ml under the aforementioned culture conditions. with an anti-FITC AP-conjugated antibody for 30 min at room Differentiating agents were added to the cultures during logar- temperature. Following repeated washings in TBS and in pure ithmic growth. Control cultures without inducers were always water, the slides were incubated with two to three drops of established simultaneously. substrate (5-bromo-4-chloro-3-indolylphosphate, BCIP) com- Blast cell immunophenotyping was performed using the bined with nitroblue tetrazolium (NBT) and an inhibitor of described panel of MoAbs, after 4 days of liquid culture with endogenous AP (levamisole), immersed in tap water for 5 min and without inducers. The analytical gate included a percent- and mounted in glycergel. age of cells higher than 90%. The absolute number of viable cells before and after incubation was calculated cytometrically, and only experiments revealing comparable values Molecular typing of HLA class I and class II were included in the analysis.

HLA class I and class II typing of the patient’s fresh cells and the derived cell lines JURL-MK1 and JURL-MK2 was performed using the PCR-sequence specific primers (SSP) method Proliferation assay described by Olerup et al.9,10 Commercial kits (Dynal, Oslo, Norway) for PCR-SSP low-resolution of HLA-A, -B, -C, DRB1 The proliferative response of the cell lines to various hemato- loci and PCR-SSP high-resolution of DRB1*01, DRB1*15/16, poietic effector molecules was examined by standard 3H-thy- DRB5, DQB1 loci were used according to the manufacturer’s midine incorporation and ␤-scintillation counting (Liquid instructions. Genomic DNA from the patient’s fresh cells and Scintillation Systems; Beckman Instruments, Fullerton, CA, from the derived cell lines was extracted as previously USA). The cells were washed extensively prior to the prolifer- described.11 Amplification was carried out using Taq-poly- ation experiments and the number of viable cells was quantit- merase (Perkin-Elmer Roche, Branchburg, NJ, USA). After ated by flow cytometry, according to their light scatter proper- PCR, electrophoresis on 2% agarose gel of amplified samples ties. The cells were seeded in triplicate in 100 ␮l medium in was carried out to detect both the presence of amplified con- flat-bottomed 96-well plates and incubated in the absence or trol bands and amplification of allele- or group-specific bands. presence of cytokines, at a concentration of 2 × 105 cells/ml. Since each primer pair gives rise to an amplified band which For the last 6 h of the 48 h incubation period, 1 ␮Ci methyl- can be identified by its molecular weight, this was tested by 3H-thymidine (Amersham-Buchler, Braunschweig, Germany) using DNA molecular weight marker VI (Boehringer was added to each well. The panel of recombinant human Mannheim). Alleles were assigned according to the cytokines was composed as follows (in parentheses, standard Nomenclature for Factors of the HLA System.12 abbreviation, supplier and concentrations at which the cytokines were applied): granulocyte colony-stimulating factor (G- CSF; Chugai-Rhone Poulenc, Vitry-sur-Seine, France, Ploidy studies 10 ng/ml); granulocyte–macrophage colony-stimulating factor (GM-CSF; Schering-Plough, Innishannon, Eire, 2.5 ng/ml), JURL-MK1 and JURL-MK2 cells were harvested, washed twice interleukin-2 (IL-2; Proleukin, EuroCetus, Amsterdam, The with Ca2+/Mg2+-free PBS and stained with a propidium iodide Netherlands, 50 U/ml), interleukin-3 (IL-3; Sandoz Pharma, (PI) ipotonic staining solution (PI 50 ␮g/ml in 0.1% sodium Basel, Switzerland, 100 U/ml), interleukin-6 (IL-6; Biosource citrate, plus RNAse, 1 mg/ml) (all chemicals were from Sigma International, Camarillo, CA, USA, 100 U/ml), interleukin-11 Chemical Co, St Louis, MO, USA). Briefly, one milliliter of PI (IL-11; Biosource International, 100 ng/ml) stem cell factor solution was added to 1 × 106 cells and incubated for 30 min (SCF/kit ligand; Biosource International, 50 ng/ml). SCF was at room temperature in the dark. The DNA content was ana- also used in combination with IL-3 and GM-CSF, at the con- lyzed by a FACScan flow cytometer (Becton Dickinson). Nor- centrations mentioned. These assays were performed three mal human leukocytes were used as the standard in all experi- times on different occasions, using the entire panel of factors. ments. Ploidy was defined by the DNA index (DI), which Moreover, dose–response curves (one cytokine per experi- represents the DNA content of G0/G1 cells in comparison to mental session) were determined with: G-CSF 0.1 ng/ml to the diploid leukocyte standard. The DI was calculated by 100 ng/ml; GM-CSF 0.25 ng/ml to 250 ng/ml; IL-2 0.5 U/ml dividing the mode channel value of the cell line G0/G1 peak to 500 U/ml; IL-3, IL-6 and IL-11 1 U/ml to 1000 U/ml; SCF by the mode channel value of the diploid leukocyte G0/G1 0.5 ng/ml to 500 ng/ml. peak.

Results Induction of cell differentiation

Induction of differentiation was studied utilizing the following Morphology and cytochemical characteristics agents: 12-O-tetradecanoyl-phorbol-13-acetate (TPA; Sigma − Chemical Co, 10 7 M), all-trans retinoic acid (ATRA; Sigma Phase contrast microscopy revealed in both lines cell size het- Chemical Co, 0.5 ␮M), dimethyl sulfoxide (DMSO; Sigma erogeneity, with the presence of giant cells (3% in both cell Chemical Co, 1.2%, v/v), cytosine-arabinoside (Ara-C; lines) which were twice the size of the majority of cells. May– − Upjohn, Kalamazoo, MI, USA, 5 × 10 9M) and hemin (Sigma Gru¨nwald–Giemsa-stained cytospin preparations showed that Chemical Co; 0.1 mM). TPA was dissolved in 100% ethanol, JURL-MK1 cells were round, with large nuclei containing one ATRA in 100% DMSO, Ara-C and hemin in DMEM with 20% to several prominent nucleoli. The cytoplasm was basophilic FBS. The inducers were further diluted to a ﬁnal concentration with many vacuoles. Almost all cells showed cytoplasmic pro- in culture medium. Before each experiment, the cells were trusions. JURL-MK1 also occasionally showed cells charac- JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1557 terized by a larger size and lobulated nuclei (Figure 1A). JURL- Table 1 Surface marker analysis of the two novel MK cell lines MK2 morphology was similar, although a few differences could be noted. In particular, JURL-MK2 exhibited a cyto- JURL-MK1 JURL-MK2 plasm with frequent azurophilic granules, and cytoplasmic protrusions were observed in a limited fraction of cells. A Lymphoid large proportion of JURL-MK2 cells had the appearance of CD2 1.1 1.3 CD7 1.2 1.1 large cells with multi-lobulated nuclei (Figure 1B). As far as CD19 1.0 5.8 cytochemical reactions are concerned, both JURL-MK1 and CD24 2.9 7.5 JURL-MK2 cells were strongly positive for PAS and weakly positive for ANAE. MPO and AP activity was not observed Myeloid (data not shown). CD13 10.6 59.1 CD33 90.9 98.9

Erythroid Immunophenotype GPA 1.2 0.9 Hemoglobin 1.8 3.7 The surface marker analysis of JURL-MK1 and JURL-MK2 cells is reported in Table 1. Both cell lines were clearly positive for Stem Cell CD33, CD63, CD43 and CD45RO. In addition, in both lines HLA-DR 48.1 96.7 CD34 2.3 2.0 a limited but consistent percentage of cells expressed gpIIbIIIa (CD41a), gpIIIa (CD61) and CD36, with no expression of gpIb Platelet (CD42b), glycophorin A (GPA), hemoglobin and CD34 (tested CD9 2.8 3.7 by two different MoAbs, IOM34 and HPCA-2). Searching for CD36 30.0 55.0 differences between the two lines, we found that CD69 was CD41a 40.1 28.9 relatively more often expressed on JURL-MK1 cells, whereas CD42b 1.0 2.5 CD61 31.8 25.6 CD13, CD44, CD54, CD30 and CD40 were speciﬁc features CD63 77.0 98.0 of JURL-MK2 cells. Furthermore, HLA-DR was expressed by most JURL-MK2 cells and by a minority of JURL-MK1 cells. Adhesion Among cytokine receptors, CD117 (SCF-R) was strongly dis- CD11a 1.8 4.5 played by a large proportion of JURL-MK1 cells but hardly CD11b 1.3 6.4 CD11c 0.7 1.1 CD43 99.0 99.1 CD44 8.9 33.0 Cd45RA 1.0 5.4 CD45R0 84.4 99.0 CD54 7.7 98.8 CD62L 1.5 3.4 CD62P 1.3 2.5

Activation CD56 2.9 13.4 CD69 28.8 11.1

NGF-R superfamily CD30 1.1 95.8 CD40 0.2 78.8

Cytokine receptors CD25 90.8 94.7 CD116 24.3 47.6 CD117 95.6 18.9 CD121a 1.1 2.5 CD121b 4.2 5.0 CD122 1.1 1.5 CDw123 0.9 57.6 CD124 1.8 5.4 CD126 1.1 9.2 CD127 1.5 2.9 CD130 1.7 1.5 CD132 1.3 2.7 CD135 2.8 2.9

Data are expressed as a percentage of positive cells. Relevant ﬁg- ures (Ͼ15%) are in bold. During the last year, the phenotypic characterization was repeated monthly, with reproducible results. Data are referred to one representative characterization.

Figure 1 Microphotographs of JURL-MK1 (A) and JURL-MK2 (B) cells (May–Gru¨nwald–Giemsa staining, × 1000). JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1558

Figure 2 JURL-MK1 (a) and JURL-MK2 (b) Giemsa-banded karyotype, displaying 39 and 72 chromosomes, respectively. For details see Results section. JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1559 detectable on a minority of JURL-MK2 cells. Both cell lines no staining, indicating that the cells were not infected with were clearly positive for CD25 (IL2-R␣ chain), while CD122 the Epstein–Barr virus (data not shown). (IL2-R␤ chain) and CD132 (IL2-R␶ chain) were undetectable. CD116 (GM-CSF-R) was more frequently expressed by JURL- MK2 cells as compared to JURL-MK1, while CDw123 (IL3- Molecular typing R␣ chain) was exclusively expressed by JURL-MK2. The other cytokine receptors tested were unexpressed by both the cell HLA class I and class II genotype of the patient’s fresh cells lines (Table 1). and of the derived cell lines was: A*33,68; B*13,14; Cw* 06,08; DRB1*0102, 1501; DRB5*0101; DQB1*0501,0602. Cytogenetic analysis As a representative experiment, PCR-SSP high resolution of DRB1*15/16 is presented in Figure 3. Results showed positive JURL-MK1 has a hypodiploid karyotype (33–43 DNA amplification of internal control (429 bp) in all gel lanes, chromosomes). The predominant karyotype was whereas specific DNA amplification, found in gel lanes 1, 3, 39XY,−4,−5,−9, der(9)t(9;22)(q34q11),−11,−12,i(17),−18,−19, 5, 10, 12, identified the presence of the DRB1*1501. −21,der(22)t(9;22)(q34q11),+mar1(six metaphases) (Figure 2a). JURL-MK2 was near-triploid (66–80 chromosomes). One paradigmatic karyotype is presented in Figure 2b (72XXY, +1,+3,+3,+4,+5,+5,+6,+7,+8,+8, der(9)t(9;22)(q34q11), Cellular ploidy analysis +der(9) t(9;22) (q34q11),+10,+10,+12,+12,+13,+14,+15,+15, +16,i(17),+18,+21,+21, der(22)t(9;22)(q34q11),+22, +der(22) Flow cytometry analysis demonstrated that the G0/G1 peak t(9;22)(q34q11). Among the 24 cells analyzed, three meta- of JURL-MK1 was comparable to the diploid peak of normal = phases displayed a pattern identical to that described, whereas leukocytes (DI 1.03), while the G0/G1 peak of JURL-MK2 = 14 metaphases had similar findings but with additional ran- indicated hyperdiploidy (DI 1.9) (data not shown). dom gains. The karyotypes have been tested on different occasions, with identical results.

In situ hybridization (ISH) for Epstein–Barr virus (EBV) nuclear RNA

Immunocytochemistry testing with FITC-conjugated EBV PNA probe complementary to the two nuclear EBER RNAs revealed

Figure 3 PCR-SSP high resolution of DRB1*15/16. A total of 12 reactions were performed per DNA sample. Each reaction contained a primer pair for the amplification of a specific allele and another one Figure 4 Southern analysis of the M-bcr region. Lane 1: DNA for the amplification of the internal control (429 bp). (a) DNA from samples from the patient at diagnosis; lane 2: DNA samples from the patient’s fresh cells; (b) DNA from JURL-MK1; (c) DNA from JURL-MK1 cell line; lane 3: DNA samples from JURL-MK2 cell line. JURL-MK2. m: standard molecular weight markers. SM: molecular size markers. JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1560 Detection of BCR rearrangements and DMSO (Figure 5d): a clear-cut increase of CD41a was paralleled by a significant augmentation of CD117 expression By Southern analysis of the M-bcr region following BglII diges- and a dramatic decrease of CD30 display. Dual fluorescence tion, DNA samples extracted from fresh bone marrow cells experiments were performed by simultaneously analyzing and from both cell lines shared the same rearranged band of CD30 and CD117 expression. A clear-cut reciprocal modu- 9.1 kb (Figure 4). No rearranged band was detected after lation of CD30 and CD117 was observed in JURL-MK2 cells digestions with the restriction enzymes (HindIII (Figure 4) and treated with either TPA (Figure 6) or DMSO. Along with these BamHI (not shown). These data were consistent with a hybrid main results documented in Figures 5 and 6, TPA stimulation bcr/abl gene with the junction b3/a2. Indeed, RT-PCR analysis exerted on JURL-MK1 a number of additional effects, ie a revealed in all samples a p210 type of bcr/abl transcript with marked increase in the percent expression of CD44, CD54 the b3/a2 junction. b2a3 transcript was not detectable by RT- and CD69 (from 8.9 to 76.1%, from 7.7 to 47.2% and from PCR (data not shown). 28.8 to 95.4%, respectively). On the same cells, the only effect produced by DMSO was a decrease in HLA-DR+ cells, from 48.1 to 19.6% (data not shown). In JURL-MK2, TPA Induction of cell differentiation stimulation increased CD44 and CD69 expression (from 30.0 to 94.5% and from 11.1 to 81.8%, respectively), while reduc- Exposure of the two cell lines over 4 days to Ara-C, ATRA and ing CD54 and CD40 expression, from 98.8 to 74.4% and from hemin did not show significant effects on the cell surface 78.8 to 30.6%, respectively. Similar results were obtained marker pattern when compared to the immunophenotype of with DMSO. In addition, in the JURL-MK2 cell line, the mean non-induced cells. By converse, exposure of JURL-MK1 to channel of fluorescence intensity of HLA-DR positive cells TPA and of JURL-MK2 to TPA or DMSO produced relevant after exposure to DMSO was significantly lower as compared surface marker pattern modifications. Marked differences to control cultures (data not shown). were observed when analyzing the expression of CD117 and CD30, along with a marker of MK differentiation, CD41a Effect of growth factors on JURL-MK1 and JURL-MK2 (Figure 5). In particular, in JURL-MK1 we observed a signifi- proliferation cant TPA-driven increase of CD41a expression (Figure 5a), while DMSO had no effect (Figure 5b). By contrast, in JURL- The JURL-MK1 and JURL-MK2 cell lines grow in the absence MK2 we observed comparable results by both TPA (Figure 5c) of any added growth factor. However, the effects of selected

Figure 5 Surface marker analysis of the two lines before (T0) and after (96 h) treatment with TPA or DMSO. (a) JURL-MK1 treated with TPA; (b) JURL-MK1 treated with DMSO; (c) JURL-MK2 treated with TPA; (d) JURL-MK2 treated with DMSO. Data are expressed as a percentage of positive cells and are referred to one representative experiment. Similar results were obtained in three independent experiments. JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1561

Figure 6 Four-dimensional flow cytometry analysis of JURL-MK1 and JURL-MK2 cells prior to and following induction of differentiation with TPA (4-day exposure). Viable cells were identified in a dual parameter cytogram of orthogonal and 180° light scatter, then cells were analyzed in two color cytograms for their expression of a combination of CD30 (x axis) and CD117 (y axis). Vertical and horizontal cursors were set on the basis of the reaction with isotype-matched controls. (a) JURL-MK1 without TPA; (b) JURL-MK1 with TPA; (c) JURL-MK2 without TPA; (d) JURL-MK2 with TPA. In JURL-MK2, TPA induced a measurable up-modulation of CD117-c-kit, accompanied by a significant reduction of CD30 expression. DMSO produced similar results. cytokines on the proliferation of these cells were assessed by lished from the peripheral blood of a patient with Ph+ CML- measuring the 3H-thymidine incorporation into DNA. The BC, which are characterized by different cytogenetic, pheno- values were calculated in comparison to the incorporation typic and functional features. The possibility of generating level without growth factor (51 000 ± 5200 c.p.m.) and are more than one cell line from a single patient has been pre- averages of triplicates. SCF stimulated the 3H-thymidine incor- viously described, for example as regards the myelomonocytic poration (1.5-fold basal level) of JURL-MK1 cells, while it was cell line ME-1.13 There are two possible explanations for the ineffective on JURL-MK2 (Figure 7). No synergistic effect with generation of morphologic, cytochemical, phenotypic and IL-3 or GM-CSF was observed. G-CSF, GM-CSF, IL-2, IL-3, IL- cytogenetic changes in different sublines: the occurrence of 6 and IL-11 had no significant effect on 3H-thymidine incor- an intracellular stochastic event or, alternatively, the establish- poration in either JURL-MK1 or JURL-MK2. Three different ment of unforeseeable and imperceptible divergent conditions assays gave identical results. Dose–response curves, perfor- in split subculture microenvironments. med with all factors, confirmed that SCF stimulated JURL-MK1 In order to obtain unequivocal evidence to demonstrate that (but not JURL-MK2) in a concentration-dependent manner our cell lines are derived from the same ancestry, as well as with the maximal stimulation at 50 ng/ml (data not shown). to formally exclude cross-contamination and misidentifi- cation, we showed the HLA class I and class II molecular typing of the original cells in direct comparison with the Discussion resulting cell lines. As for cytogenetic differences, JURL-MK1 is hypodiploid, Cell lines of CML origin, representing the early differentiation whereas JURL-MK2 is near triploid. This finding, obtained by stages of various hemopoietic lineages are useful models to conventional cytogenetic analysis, was confirmed by the investigate biological events occurring in this type of leuke- results of ploidy analysis by flow cytometry, although the con- mia.2 Moreover, among the various CML cell lines, cells cordance between these two methods cannot be absolute. arrested at the MK differentiation stage can provide new In order to clarify the differentiation stage of the cell lines insights into biological studies on megakaryocytopoiesis, JURL-MK1 and JURL-MK2 we evaluated a large series of cell which are often hampered by difficulty in obtaining MK pro- markers, including platelet glycoproteins, the CD34 antigen genitors from normal human bone marrow.3,4 In the present and erythromyeloid specificities. A ‘pure’ MK cell line should study, we describe two novel MK-oriented cell lines, estab- display the following biological characteristics: expression of JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1562

Figure 7 Effect of growth factors on the proliferation of JURL-MK1 (a) and JURL-MK2 (b) cells. The values were calculated in comparison to the incorporation level without growth factor (51 000 ± 5200 c.p.m.) and are averages from triplicates. Three independent experiments gave identical results.

the MK specific antigen CD41; expression of c-kit/CD117; MK evidence for the MK nature of JURL-MK1 and JURL-MK2 differentiation induced by TPA.4 The absence of the cell lines. CD42b/gpIb molecule does not preclude the attribution of MK JURL-MK1 and JURL-MK2 do not express CD34. It is not origin to JURL-MK1 and JURL-MK2, since even in other pre- clear whether this phenotype reflects a relatively late differen- viously described MK cell lines its expression is variable.14 tiation stage, as suggested by the existence of CD34−/CD61+ Moreover, when typing acute leukemia cells, CD41a+/CD42b− MK precursors16 or, alternatively, if it indicates a malignancy- cells are considered as megakaryoblasts,15 while the simul- related surface membrane abnormality. The lack of CD34 taneous expression of CD41a and CD42b has been referred positivity cannot be due to a defective reactivity in the to as characteristic of (pro)megakaryocytic leukemia. At the reagents used, since the finding was confirmed using two dif- same time, the constitutive expression of CD61/gpIIIa along ferent MoAbs. Of note, other published MK lines are charac- with CD36/thrombospondin receptor15 provides additional terized by low or even absent CD34 expression.4 The pheno- JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1563 typic pattern CD41a+CD42b−CD34− has been reported on MK cular, while JURL-MK1 phenotypic changes are only pro- cell lines, in the context of a wide spectrum of different anti- moted using a phorbol ester, JURL-MK2 shows a clear-cut genic combinations. A data bank search has shown that 28/29 response to both TPA and DMSO, as testified by the signifi- MK cell lines were CD41a+, 15/25 CD42b+ and 18/24 CD34+. cant modulation of several surface markers. In addition, TPA The JURL-MK1 and JURL-MK2 cell lines exhibit high levels and DMSO induced a marked up-regulation of CD117/c-kit of the CD45 isoform R0. Recently, we have demonstrated a expression in JURL-MK2. This finding is in keeping with pre- preferential expression of CD45R0 on CML-BC cells as com- vious data, indicating that the treatment of MK cell lines with pared to de novo acute myeloblatic leukemia (AML) blast TPA and bryostatin 1 triggers a significant increase in c-kit cells, which very frequently express CD45RA.17 Since mRNA levels.22 Our data demonstrate that c-kit expression CD45R0 has been associated with adhesive cell proper- can be augmented by various differentiation stimuli, not ties,18,19 JURL-MK1 and JURL-MK2 may be a tool for further necessarily confined to protein kinase C activators. Differen- analyzing the potential involvement of CD45R0 in MK tiation experiments performed on JURL-MK2 indicated a adhesion processes. reciprocal modulation of CD30 and CD117 expression, TPA- In this report we demonstrate that the IL-2 receptor ␣-chain and DMSO-driven CD117 up-modulation being consistently (CD25) is strongly expressed by both JURL-MK1 and JURL- paralleled by decreased CD30 expression. MK2, while ␤ and ␶ chains are cytometrically undetectable. To date, about 20 megakaryoblastic cell lines, including the It has been demonstrated that the lack of detection by flow present two, have been published, generally derived from cytometry of the II-2 receptor ␤ chain (p75, CD122) is caused AML-M7 or CML-BC patients. Some of these cell lines (MEG- by the low number of molecules per cell and that this chain 01, DAMI, CMK, MKPL-1, M-07e, MOLM-1, UT-7) represent expression is a common feature of myeloid leukemia cells.20 nowadays a standard tool for studying MK biology.3,4 In our In non-lymphoid malignancies the expression of CD25 is opinion, our novel cell lines can be employed in studying new related to the co-expression of markers of different lineages as aspects of megakaryocytopoiesis, such as the expression and well as to the presence of the Philadelphia chromosome.21 role of NGF-Rs and c-kit, and for further elucidating the bio- Another relevant finding in this work is the divergent logical meaning of the apparently reciprocal regulation of expression of CD117/c-kit, CD116/GM-CSF-R and these molecules. CDw123/IL3R␣ in these two cell lines. The c-kit product, reported on most of the megakaryoblastic cell lines estab- 22,23 lished so far, was more expressed on JURL-MK1, as com- Availability of the cell lines pared to JURL-MK2. In accordance with surface phenotype, the growth of the JURL-MK1 cell line could be significantly JURL-MK1 and JURL-MK2 cell lines will be made available to enhanced by exposure to SCF. By contrast, although CD116 outside investigators upon request to the first (RDN) or senior and CDw123 were an almost exclusive feature of JURL-MK2, author (LDV). GM-CSF and IL-3 were unable to stimulate JURL-MK2 cell growth. Moreover, GM-CSF and IL-3 were unable to synergize with SCF on both JURL-MK1 and JURL-MK2 cell lines. These data are divergent from those obtained with another well- Acknowledgements described MK cell line, ELF-153,3,4 and could enable additional studies concerning the effects of SCF on R Di Noto was a recipient of an AIRC fellowship throughout megakaryocytopoiesis. the period of establishment of the two cell lines (1993–1995). Further studies on the effects of other hematopoietic cyto- Ploidy was studied by Dr Stefano Pepe and Dr Angela Ruggi- kines, in particular the recently cloned human recombinant ero, Chair of Oncology, Universita` Federico II, Naples. thrombopoietin (TPO) are required. A recent review reported that while none of 30 growth factor independent erythro-MK cell lines respond to TPO with increased proliferation, this References growth factor strongly augments the growth of cytokine- dependent cell lines (HU-3, M0-7e, TF-1), which can be 1 Kantarjian HM, Deisseroth A, Kurzrock R, Estrov Z, Talpaz M. rendered TPO-dependent and used as bioassays.24 Chronic myelogenous leukemia: a concise update. Blood 1993; The vast majority of JURL-MK2 cells express CD30, which 82: 691–703. is a member of the nerve growth factor-receptor (NGF-R) 2 Drexler HG. Leukemia cell lines: in vitro models for the study of superfamily25,26 and is typically expressed by cells of certain chronic myeloid leukemia. Leukemia Res 1994; 18: 919–927. 27 3 Hassan HT, Freund M. Review of megakaryoblastic cell lines. lymphoid neoplasms; JURL-MK1 cells lack this antigen. The Characteristic biological features of human megakaryoblastic leu- occasional expression of CD30 in myeloid malignancies has kaemia cell lines. Leukemia Res 1995; 19: 589–594. recently been reported, particularly in myelodysplasia-AML.28 4 Hassan HT, Drexler HG. Interleukins and colony-stimulating fac- It has also been shown that LAMA-84, an MK-oriented cell tors in human myeloid leukemia cell lines. Leuk Lymphoma 1995; line, produces CD30 transcripts.29 The results of published 20: 1–15. observations indicate that the binding of the CD30 ligand to 5 Seabright M. A rapid banding technique for human chromosomes. Lancet 1971; II: 971–972. its receptor confers the capacity for signal transduction to dif- 6 Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Labora- 29,30 ferent target cells. Interestingly, JURL-MK2 (and not JURL- tory Manual, 2nd edn. Cold Spring Harbor Laboratory Press: New MK1) cells express CD40, another member of the NGF-R York, 1989. superfamily. In this respect, it will be of interest to investigate 7 Negrini M, Tallarico A, Pazzi I, Castagnosli A, Cuneo A, Castoldi whether CD30 and/or CD40 ligands are able to trigger a bio- GL. A new chromosomal breakpoint in Ph positive, bcr negative logical response in the JURL-MK2 cell line. chronic myelogenous leukemia. Cancer Genet Cytogenet 1992; 61: 11–13. Treatment of the two novel megakaryoblastic cell lines here 8 Pane F, Frigeri F, Camera A, Sindona M, Brighel F, Martinelli V, reported with differentiation inducers has indicated that JURL- Luciano L, Selleri C, Del Vecchio L, Rotoli B, Salvatore F. Com- MK2 is more inducible as compared to JURL-MK1. In parti- plete phenotypic and genotypic lineage switch in a Philadelphia JURL-MK1 and JURL-MK2 megakaryoblastic cell lines R Di Noto et al 1564 chromosome-positive acute lymphoblastic leukemia. Leukemia stromal cell elements via recognition of stromal heparan sulfate. 1996; 10: 741–749. Blood 1994; 84: 739–752. 9 Olerup O, Zetterquist H. HLA-DR typing by PCR amplification 20 Pizzolo G, Rigo A, Zanotti R, Vinante F, Vincenzi C, Cassatella with sequence-specific primers (PCR-SSP) in 2 hours: an alterna- M, Carra G, Castaman G, Chilosi M, Semenzato G, Zambello R, tive to serological DR typing in clinical practice including donor– Trentin L, Libonati M, Perona G. ␣ (p55) and ␤ (p75) chains of recipient matching in cadaveric transplantation. Tiss Antigens the interleukin-2 receptor are expressed by AML blasts. Leukemia 1992; 39: 225–235. 1993; 7: 418–425. 10 Olerup O, Aldener A, Fogdell A. HLA-DQB1 and DQA1 typing 21 Nakase K, Kita K, Otsuji A, Anazawa H, Hoshino K, Sekine T, by PCR amplification with sequence specific primers (PCR-SSP) in Shirakawa S, Tanaka I, Nasu K, Tsutani H, Dohy H, Tsudo M. 2 hours. Tiss Antigens 1993; 41: 119–134. Diagnostic and clinical importance of interleukin-2 receptor alpha 11 Lombardi ML, Mercuro O, Tecame G, Fusco C, Ruocco V, Salerno chain expression on non-T cell acute leukaemia cells. Br J Haema- A, Pirozzi G, Manzo C. Molecular analysis of HLA DRB1 and tol 1992; 80: 317–326. DQB1 in Italian patients with Pemphigus vulgaris. Tiss Antigens 22 Hu ZB, Ma W, Uphoff CC, Quentmeier H, Drexler HG. c-kit 1996; 47: 228–230. expression in human megakaryoblastic leukemia cell lines. Blood 12 Bodmer JG, Marsh SGE, Albert ED, Bodmer WF, Bontrop RE, Char- 1994; 8: 2133–2144. ron D, Dupont B, Erlich HA, Mach B, Mayr WR, Parham P, Sasa- 23 Morita S, Tsuchiya S, Fujie H, Itano M, Ohashi Y, Minegishi M, zuki T, Schreuder GMT, Strominger JL, Svejgaard A, Terasaki PI. Imaizumi M, Endo M, Takano N, Konno T. Cell surface c-kit Nomenclature for factors of the HLA system, 1995. Tiss Antigens receptors in human leukemia cell lines and pediatric leukemia: 1995; 46: 1–18. selective preservation of c-kit expression on megakaryoblastic cell 13 Yanagisawa K, Sato M, Horiuchi T, Hasegawa H, Fujita S. Differ- lines during adaptation to in vitro culture. Leukemia 1996; 10: 102–105. ence in response to colony-stimulating factors and involvement of 24 Drexler HG, Quentmeier H. Thrombopoietin: expression of its protein kinase C signal transduction system in three subclones receptor MPL and proliferative effects on leukemic cells. Leukemia from the ME-1 cell line and two sublines. Blood 1994; 84: 84–93. 1996; 10: 1405–1421. 14 Takeuchi S, Sugito S, Uemura Y, Miyagi T, Kubonishi I, Taguchi H, 25 Durkop H, Latza U, Hummel M, Eitelbach F, Seed B, Stein H. Enzan H, Ohtsuki Y, Miyoshi I. Acute megakaryoblastic leukemia: Molecular cloning and expression of a new member of the nerve establishment of a new cell line (MKPL-1) in vitro and in vivo. growth factor receptor family that is characteristic for Hodgkin’s Leukemia 1992; 6: 588–594. disease. Cell 1992; 68: 421–427. 15 Imamura N, Ota H, Abe K, Kuramoto A. Expression of the throm- 26 Goodwin RG, Armitage RJ. CD30 antigen, a marker for Hodgkin’s bospondin receptor (CD36) on the cell surface in megakaryoblas- lymphoma, is a receptor whose ligand defines an emerging family tic and promegakaryocytic leukemias: increment of the receptor of cytokines with homology to TNF. Cell 1993; 73: 1349–1360. by megakaryocyte differentiation in vitro. Am J Hematol 1994; 45: 27 Smith CA, Gruss HJ, Davis T, Anderson D, Farrah T, Baker E, Suth- 181–184. erland GR, Brannan C, Copeland NG, Jenkins NA, Grabstein KH, 16 Debili N, Issaad C, Masse´ JM, Guichard J, Katz A, Breton-Gorius Gliniak B, McAlister IB, Fanslow W, Alderson M, Falk B, Gimpel J, Vainchenker W. Expression of CD34 and platelet glycoprotein S, Gillis S, Din WS, Goodwin RG, Armitage RJ. CD30 antigen, a during human megakaryocytic differentiation. Blood 1992; 80: marker for Hodgkin’s lymphoma, is a receptor whose ligand 3022–3035. defines an emerging family of cytokines with homology to TNF. 17 Schiavone EM, Lo Pardo C, Di Noto R, Manzo C, Ferrara F, Vacca Cell 1993; 73: 1349–1360. C, Del Vecchio L. Expression of the leukocyte common antigen 28 Masuya M, Kita K, Shimizu N, Ohishi K, Katayama N, Sekine T, (LCA, CD45) isoforms RA and R0 in acute haematological malig- Otsuji A, Miwa H, Shirakawa S. Biologic characteristics of acute nancies: possible relevance in the definition of new overlap points leukemia after myelodysplastic syndrome. Blood 1993; 81: between normal and leukaemic haemopoiesis. Br J Haematol 3388–3394. 1995; 91: 899–906. 29 Gruss HG, DaSilva N, Zhen-Bo H, Uphoff CC, Goodwin RG, 18 Stamenkovic I, Sgroi D, Aruffo A, Sy MS, Anderson T. The B lym- Drexler HG. Expression and regulation of CD30 ligand and CD30 phocyte adhesion molecule CD22 interacts with leukocyte com- in human leukemia–lymphoma cell lines. Leukemia 1994; 8: mon antigen CD45R0 on T cells and alpha 2-6 sialyltransferase, 2083–2094. CD75, on B cells. Cell 1991; 66: 1133–1144. 30 Gruss HJ, Boiani N, Williams DE, Armitage RJ, Smith CA, Good- 19 Coombe DR, Watt SM, Parish CR. Mac-1 (CD11b/CD18) and win RG. Pleiotropic effects of the CD30 ligand on CD30-express- CD45 mediate the adhesion of hematopoietic progenitor cells to ing cells and lymphoma cell lines. Blood 1994; 83: 2045–2056.