Committed Eosinophil Cell Line (YJ) Demonstrates Trilineage Potential Y Yamaguchi1, H Nishio1, T Kasahara2, SJ Ackerman3, H Koyanagi4 and T Suda1

Home , Eosinophil peroxidase, Megakaryocyte

Leukemia (1998) 12, 1430–1439  1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu Models of lineage switching in hematopoietic development: a new myeloid- committed eosinophil cell line (YJ) demonstrates trilineage potential Y Yamaguchi1, H Nishio1, T Kasahara2, SJ Ackerman3, H Koyanagi4 and T Suda1

1Department of Cell Differentiation, Institute of Molecular Embryology and Genetics, Kumamoto University School of Medicine, Kumamoto; 2Department of Biochemistry, Kyoritsu College of Pharmacy, Tokyo, Japan; 3Department of Biochemistry and Molecular Biology, University of Illinois, Chicago, IL, USA; and 4Department of Pediatrics, Jichi Medical School, Tochigi, Japan

A new human leukemia cell line with an eosinophilic pheno- Eosinophil differentiation from hematopoietic stem cells is type, designated YJ, was established from the peripheral blood regulated by a number of different cytokines, including IL-3, cells of a patient with chronic myelomonocytic leukemia 9–11 (CMMoL) with eosinophilia. When cultured in RPMI 1640 GM-CSF and IL-5. However, the molecular basis for the medium containing 10% fetal bovine serum, most YJ cells were commitment of progenitors to the eosinophil lineage remains myeloblastoid with a small number of the cells having eosino- unknown. One of the reasons why it has been difficult to ana- philic granules. Cell surface markers in the YJ cells were posi- lyze transcription factors that may be involved in the differen- tive for CD33 and were negative for CD34, CD16 and CD23. The tiation of myeloid progenitors to the eosinophil lineage has eosinophilic characteristics of YJ cells were confirmed by his- been the absence of eosinophilic cell lines suitable for elucid- tochemical staining with Fast-Green/Neutral-Red and by the expression of mRNAs for eosinophil-associated granule pro- ating the molecular basis of eosinophilic differentiation. teins, eosinophil cationic protein (ECP), eosinophil-derived We describe the establishment of a new eosinophilic leuke- neurotoxin (EDN), eosinophil peroxidase (EPO), and major mia cell line, YJ, capable of differentiating to the monocytic basic protein (MBP), and for the Charcot–Leyden crystal (CLC) lineage with PMA stimulation or to the neutrophil lineage with protein. The YJ cells could be induced towards monocytic dif- all-trans retinoic acid (RA). YJ cells may provide a potentially ferentiation by stimulation with phorbol 12-myristate 13-acetate useful model for analyzing the molecular basis for lineage (PMA). The monocytic characteristics of YJ cells treated with PMA were confirmed by morphological analysis with ␣-naph- switching among the eosinophil, macrophage and neutrophil thyl butyrate esterase staining, by CD14 expression, and by hematopoietic programs. In addition, this cell line may be increased expression of Egr-1 mRNA. Furthermore, YJ cells suitable to isolate the transcription factor(s) involved in could be differentiated towards the neutrophil lineage by stimu- eosinophil differentiation using the subtraction or the differen- lation with all-trans retinoic acid (RA). YJ cells treated in vitro tial display methods. with 2 ␮M RA differentiated into metamyelocytes and band neutrophils, and increased the number of nitroblue tetrazolium (NBT)-positive cells and increased gp91phox mRNA expression. Thus, the YJ cell line exhibited eosinophilic characteristics, but Materials and methods was able to differentiate to the monocytic or neutrophilic lineages in response to PMA or RA, respectively. The expression Case report of genes for transcription factors involved in myeloid differentiation was evaluated by Northern blot analysis. Increased HO, a 69-year-old Japanese male, was admitted to the Hema- expression of Egr-1 was observed with macrophage differentiation. In contrast, increased expressions of C/EBP␤ and MZF- tology Division of Jichi Medical School on 15 June 1994, for 1 mRNA occurred with neutrophilic differentiation. The YJ cell the treatment of intense breathing difficulty and high fever. line should be useful for elucidating the molecular mechanisms The history of his illness began in 1988 with the onset of ane- governing lineage switching from the eosinophil to monocytic mia. At that time, a diagnosis of myelodysplastic syndrome or neutrophil lineages. (MDS) with refractory anemia, was made according to Keywords: eosinophils; neutrophils; macrophages; cell line; tran- the classification of the French–American–British (FAB) scription factors nomenclature. The patient had been well except for eosinophilia and easy fatiguability due to anemia until June 1994, when dyspnea and high fever developed. The eosinophilia Introduction and splenomegaly had developed gradually. At the time of admission, hepatosplenomegaly was observed and no lym- A large number of continuous cell lines from patients with phadenopathy was found. Hematological findings showed a myeloblastic leukemia have been established and many cases hemoglobin of 6.3 g/dl, a leukocyte count of 35.2 × 109/l, and 1–3 of eosinophilic leukemia have been reported. To date, how- a platelet count of 47.3 × 109/l. The leukocyte differential ever, only one eosinophil leukemia cell line, EoL, from a count revealed 5.4% blasts, 1.2% promyelocytes, 12% myelo- 4 patient with eosinophilic leukemia has been established. cytes, 3.4% metamyelocytes, 52% mature myelocytes, 12% Other eosinophilic cell lines have been subcloned from the eosinophils, 4.2% monocytes and 7% lymphocytes. The bone 5,6 HL-60 promyelocytic leukemia line, including HL-60-C15 marrow was hypocellular and myelofibrotic, with cytogenetic + 7 and HL-60-3 C5. Recently, an acute myelogenous leukemia studies of bone marrow showing 47XY, 1qh+, +8, del(12)(q15 (AML) cell line capable of terminally differentiating to eosino- q23) (100%). After admission, a diagnosis of myeloprolifer- phils in response to IL-5, IL-3 and GM-CSF has been ative disorder with eosinophilia was made and the patient was 8 reported. treated with cytosine arabinoside and prednisolone, and there was a striking clinical improvement. Three months after chemotherapy, leukocytosis with increased blast cells and Correspondence: Y Yamaguchi, Department of Cell Differentiation, marked splenomegaly developed. Hematological findings in Institute of Molecular Embryology and Genetics, Kumamoto Univer- × 9 sity School of Medicine, Honjo 2-2-1, Kumamoto, 860-0811 Japan; peripheral blood were: 166 10 /l leukocytes with 7% blast Fax: 81 96 373 5332 cells, 8.6% eosinophils, 11% neutrophils, 2% myelocytes and Received 19 September 1997; accepted 30 April 1998 metamyelocytes, 60% monocytoid cells, 6.2 g/dl hemoglobin Models of lineage switching in hematopoietic development Y Yamaguchi et al 1431 and 60 × 109/l platelets. Bone marrow aspirate was normocel- Reagents and cytokines lular with 90% myeloid cells including 6.2% blast cells, 20% neutrophils, 10.4% eosinophils, and 50% monocytoid cells. The cytokines used were human rIL-1␣, human rTNF-␣ The serum and urinary lysozyme levels were increased. A final (kindly donated by Dainippon Pharmaceutical, Suita, Osaka, diagnosis of blastic crisis of chronic myelomonocytic leuke- Japan), rhIL-2 (specific activity 1 × 107 U/mg, provided by mia (CMMoL) with eosinophilia was made. Despite additional Shionogi Pharmaceutical, Osaka, Japan), rhIL-3 (specific chemotherapy, the patient died of internal bleeding. activity, 21 900 U/ml, provided by Genentics Institute, Cam- bridge, MA, USA), rhIL-4 (specific activity, 1.3 × 107 U/ml, provided by DNAX Research Institute, Palo Alto, CA, USA), human rIL-5 (donated by Suntory, Tokyo, Japan), rhIL-6 Cell culture and cloning (specific activity, 3.9 × 109 U/mg, provided by Ajinomoto, Kawasaki, Japan), rhIL-8 (a half-maximal neutrophil chemo- Heparinized peripheral blood was obtained following infor- tactic activity, 0.5–1.0 ng/ml, provided by Dr Matsushima, med consent from the patient. The polymorphonuclear (PMN) Cancer Research Institute, Kanazawa, Japan), rhGM-CSF × 9 leukocyte fraction, composed mostly of cells of eosinophilic (specific activity, 1 10 U/ng, provided by Sumitomo Pharm- lineage including immature cells, was prepared by a discon- aceutical, Osaka, Japan), rhRANTES (specific activity assessed tinuous Percoll centrifugation method. The PMN cells were as effective chemotaxis on human monocytes detected at 10 cultured in tissue culture flasks at 5 × 105 cells/ml in RPMI to 100 ng/ml, purchased from Biosource, CA, USA), human ␥ 1640 supplemented with 10% fetal bovine serum (Hyclone, rIFN- (Takeda Pharmaceutical, Osaka, Japan), and human Logan, UT, USA), 2 mML-glutamine, 50 IU/ml of penicillin recombinant adult T cell leukemia-derived factor/thioredoxin and 50 ␮g/ml streptomycin, and the supernatant (10% v/v) of (rADF/TRX) (Ajinomoto, Tokyo, Japan). C5a, formyl- X63Ag8.653 plasmacytoma cells transfected with BMGNeo methionyl-leucyl-phenylalanine (fMLP), lipopolysaccharide murine IL-5 cDNA as a source of IL-5, in a humidified atmos- (LPS), phorbol 12-myristate 13-acetate (PMA), and all-trans 12 retinoic acid (RA) were purchased from Sigma Chemical (St phere of 5% CO2. The cultures were fed once a week, but the cells cultured in the absence of IL-5 did not proliferate. Louis, MO, USA). All these stimulants were used at optimal Cultured cells were subcloned by the in vitro colony forma- concentrations. tion method as described previously.10

Measurement of IL-8

Cytochemical staining IL-8 was measured by a sensitive ELISA method established as described previously.31 The minimum detection level of IL- The cells were stained with Wright–Giemsa, Fast-Green/ 8 by this ELISA was approximately 20 pg/ml. Neutral-Red, and naphthol AS-D chloracetate/␣-naphthyl butyrate esterase. Fluorescence-activated cell sorting (FACS) analysis

YJ cells were collected and resuspended in PBS with bovine Northern blot analysis serum albumin (4%), adjusted to a final concentration of 5 × 105 cells/100 ␮l, and incubated for 30 min at 4°C with monoclonal antibodies against CD16 (PharMingen, San Total cellular RNA was isolated from HL-60-C15 cells, YJ Diego, CA, USA), CD23 (Nichirei, Tokyo, Japan), CD33 cells, and PMA- or RA-induced YJ cells using the single step + (Dako, Glostrup, Denmark), CD34 (Nichirei), and CD14 RNA isolation method,13 followed by poly(A) mRNA selec- (Dako) conjugated with fluorescein isothiocyanate or phyco- tion using oligotex-dT30 (Takara Shuzo, Kyoto, Japan). The + erythrin. poly(A) mRNA (4 ␮g per lane) was denatured in formamide- formaldehyde and then subjected to electrophoresis in 1% agarose-formaldehyde gels. The RNA was then transferred to Biodyne nylon membranes (Pall BioSupport, East Hills, NY, Assay of MBP concentration USA) and the blots probed sequentially with radiolabeled cDNAs encoding human eosinophil peroxidase (EPO),14 The concentration of MBP in the supernatants of YJ cell sus- 32 human major basic protein (MBP),15 human eosinophil cat- pensions was measured by ELISA as described previously. ionic protein (ECP),16 human eosinophil-derived neurotoxin The YJ cells were washed twice with HBSS/2ME and resus- × 4 (EDN),17 human Charcot–Leyden crystal (CLC) protein,18 neu- pended in the same medium at 5 10 cell/ml. Aliquots of trophil elastase,19 gp91phox,20 myeloperoxidase (MPO),21 cell suspensions of 0.2 ml were added to the Eppendorf tubes. ° CD14,22 GATA-123 and GATA-2,24 C/EBP␣,25 C/EBP␤,26 Egr- Cell suspensions were incubated for 2 h at 37 C in the pres- 1,27 MZF-1,28 PU.129 and 18S ribosomal RNA30 to confirm ence of various reagents. Two hours after incubation, the con- equivalent RNA loading. All hybridizations were performed centration of MBP in the supernatants of each sample was with random-primed cDNA probes at 42°C for 18 h in 50% assayed. formamide, 6X SSC, 0.2% Ficoll-polyvinylpyrrolidone (PVP), and 0.1% sodium dodecyl sulfate (SDS). Filters were washed twice in 2X SSC containing 0.2% SDS at 53°C for 30 min and Nitroblue Tetrazolium (NBT) Reduction Test twice in 0.2% SSC containing 0.2% SDS at 55°C for 30 min. Autoradigraphy was performed at −80°C with Kodak XAR-5 The NBT assay was performed on RA-stimulated YJ cells as film. described by Baehner and Nathan.33 Models of lineage switching in hematopoietic development Y Yamaguchi et al 1432 Statistical analysis teristics of YJ cells were readily apparent by Fast- Green/Neutral-Red staining (Figure 1d). Results were expressed as means ± s.d. Statistical significance was evaluated using Student’s t-test. Chromosome analysis Results The chromosomal analysis of YJ cells was very complex, exhi- Establishment of YJ cell line biting a 45, X, −Y, −2, add(3)(q27), del(3)(p13), add(4)(q34), −5, add(6)(q25), −9, del(10)(p12), add(11)(p15), add(13)(q32), For the first month of culture, cells did not increase in number. −14, −14, −15, −16, add(16)(q23), −17, del(18)(q22), +m1-m8. However, by 2 months of culture the proliferation rate of the This was significantly different from that of the patient’s bone cultured cells had increased considerably. After culturing the marrow cells at the time of admission. cell line for 5 months and confirming its proliferation in the absence of IL-5, the line was designated as YJ cells. Once the line was established, IL-5, GM-CSF, and IL-3 had no effect on Expression of mRNAs for eosinophil granule proteins the proliferation or differentiation of YJ cells (data not shown). in the YJ cells This YJ cell line has a doubling time of 24 h. Morphologic and cytochemical studies suggested that the YJ Morphologic and cytochemical characteristics of YJ cells possessed eosinophilic characteristics. To determine cells whether YJ cells expressed cellular markers of the eosinophil lineage, we performed Northern blot analysis to investigate Most YJ cells were myeloblastic with a small number (about the expression of mRNAs for the eosinophil-associated gran- 10–15%) of the cells having eosinophilic granules by Wright– ule proteins (ECP, EDN, EPO and MBP) and cytosolic CLC Giemsa staining (Figure 1a). As well, the eosinophilic charac- protein. As shown in lane 2 of Figure 2, the level of mRNA a b

c d

Figure 1 Morphological appearance of YJ cells treated with PMA or RA. Wright–Giemsa-stained cytospin preparations of YJ cells treated not at all (a), or with 10 nM PMA for 2 days (b), 2 ␮M RA for 4 days (c), or Fast Green/Neutral Red stains of uninduced YJ cells (d) (original magniﬁcation ×1000). Models of lineage switching in hematopoietic development Y Yamaguchi et al 1433 secretion from YJ cells stimulated with cytokines (IL-1␣, IL-5, TNF␣, IFN␤), chemokines (RANTES), or other activating agents (ADF/TRX, C5a, fMLP, LPS, PMA, A23187) are summa- rized in Table 1. The levels of secreted IL-8 were assayed 24 h after culture in the presence of various stimulants. Whereas the amount of IL-8 secreted from YJ cells cultured without stimulants was 0.32 ng/ml, in cells cultured with PMA or TNF- ␣ for 24 h, this value increased 100-fold and six-fold, respectively. No signiﬁcant increase in the IL-8 level was observed for YJ cells cultured with any of the other stimulants.

YJ cells differentiate to the monocytic lineage in response to PMA

Because it was observed that YJ cells produced a considerable amount of IL-8 following stimulation with PMA, we analyzed whether there was a morphological change in YJ cells after treatment with PMA. The YJ cells became more adherent and aggregative after treatment with PMA. Wright–Giemsa cytospin preparations demonstrated that PMA induced YJ cell differentiation towards the macrophagic lineage. PMA-induced YJ cells showed a significant decrease in the nuclear-cytoplasmic ratio, and an increase in cytoplasmic vesicles (Figure 1b). The ␣-naphthyl butyrate esterase reaction was positive in the PMA-induced macrophage-differentiated YJ cells (data not shown). To investigate further the monocytic differentiation of YJ cells induced by PMA, expression of Egr-1 mRNA by North- Figure 2 Northern blot analysis for eosinophil-associated proteins ern blot analysis and expression of CD14 surface antigen was (CLC, ECP, EDN, EPO, and MBP), neutrophil elastase, gp91phox, MPO, analyzed by FACS. In the same cells, expression of CD14 was CD14, and 18S rRNA mRNA expression in HL-60-C15 (lane 1), unin- significantly increased (Figure 4), with increasing duration of duced YJ cells (lane 2), YJ cells induced with 10 nM PMA for 2 days exposure to PMA (data not shown). (lane 3), and YJ cells induced with 2 ␮M RA for 4 days (lane 4). expression in the YJ cells for the eosinophil granule cationic YJ cells differentiate to the neutrophilic lineage in proteins, ECP, EDN, EPO and MBP, as well as CLC protein, response to RA were similar to those expressed in an eosinophilic subline of HL-60, HL-60-C15 cells. These findings demonstrate the It is well known that the human promyelocytic leukemia cell eosinophil characteristics of the YJ cell line. line, HL-60, differentiates into neutrophilic cells following treatment with RA. We tested whether YJ cells have the capacity to differentiate towards the neutrophilic lineage in Flow cytometric analysis of surface antigens expressed response to RA, even though this cell line was not derived on YJ cells from a patient with acute promyelocytic leukemia. Surpris- ingly, RA-induced YJ cells differentiated morphologically into To characterize surface antigen expression on YJ cells, FACS mature neutrophils (Figure 1c). YJ cells were cultured with analysis was performed. As shown in Figure 3, YJ cells were various concentrations of RA for 4 days and their ability to mainly CD33-positive and contained a small number of reduce NBT was measured. As shown in Figure 5, 2 ␮M RA CD14-positive cells. No expression of CD16 (Fc␥RIII), CD23 induced about 85% of YJ cells to become NBT-positive. Also, (FC⑀RII), or CD34 (stem cell antigen) antigens was observed as shown in Figure 2, the RA-induced YJ cells showed in YJ cells. It has been reported that B cells and myeloid cells expression of mRNA for gp91phox and CD14 and decreased can be generated from a single progenitor cell.34 Accordingly, expression for MPO. Importantly, mRNA expression for the we analyzed YJ cells for the expression of CD19 and CD20. eosinophil-associated CLC and MBP genes was not observed, However, YJ cells were negative for both of these B cell mark- in YJ cells induced with RA towards the neutrophil lineage, ers (data not shown). These results indicate that YJ cells do suggesting a loss of eosinophilic characteristics. not express characteristics of more mature myeloid cells or B lymphocytes. Induction of genes for transcription factors in the differentiation of eosinophil-committed cells to IL-8 production from YJ cells in response to various macrophages or neutrophils stimulants Northern blot analysis showed that the expression of mRNA To determine whether YJ cells respond to various cytokines for the eosinophil-associated proteins CLC, ECP, EDN, EPO, or other activating stimuli, we measured the amount of IL-8, and MBP was either decreased (ECP, EDN, EPO) or no longer which is mainly produced by myeloid cells including eosino- evident (CLC, MBP) in YJ cells cultured with PMA for 2 days phils,35,36 using a sensitive ELISA method.31 Results for IL-8 or with RA for 4 days. These findings support a loss of eosino- Models of lineage switching in hematopoietic development Y Yamaguchi et al 1434

Figure 3 Flow cytometric analysis of surface antigens expressed on YJ cells. The expression of CD16, CD23, CD33, CD34 and CD14 on YJ cells is shown. Uninduced YJ cells were clearly positive for CD33 and marginally positive for CD14.

philic characteristics by YJ cells in the process of lineage switching to the macrophage or neutrophil programs. Table 1 IL-8 production in YJ cells treated with various reagents Hematopoietic differentiation is regulated by the expression of transcription factors that activate lineage-speciﬁc genes. It Stimulants IL-8 ng/ml was therefore of interest to analyze the expression of transcription factors by YJ cells in their differentiation from the eosino- None 0.32 ± 0.09 phil to the macrophage or neutrophil lineages. Among tran- IL-1␣ (20 ng/ml) 0.49 ± 0.02 scription factors known to be involved in myeloid IL-5 (1 ␮g/ml) 0.27 ± 0.04 differentiation, the expression of GATA-1, GATA-2, C/EBP␣, TNF-␣ 20 ng/ml) 1.84 ± 0.20 C/EBP␤, Egr-1, MZF-1 and PU.1 transcripts were analyzed ␥ 3 ± IFN- (10 U/ml) 0.48 0.03 (Figure 6). The expression of GATA-1, which is involved in RANTES (100 ng/ml) 0.37 ± 0.05 ADF (200 ng/ml) 0.39 ± 0.02 the differentiation of erythroid/megakaryocyte, eosinophil, − C5a (5 × 10 7 M) 0.48 ± 0.03 and mast cell lineages, did not change with macrophage or − fMLP (5 × 10 7 M) 0.44 ± 0.10 neutrophil differentiation induced by PMA or RA, respectively. LPS (20 ␮g/ml) 0.43 ± 0.08 Expression of GATA-2 also remained unchanged with macro- ± PMA (20 nM) 33.7 3.6 phage or neutrophil differentiation. No GATA-3 transcript was A23187 (100 ng/ml) 1.2 ± 0.3 observed in all the cells, HL-60-C15, YJ, YJ induced with

× 5 PMA, and YJ induced with RA (data not shown). In contrast, Cells (5 10 cells/ml in 1% FCS-RPMI 1640) were incubated with ␣ the indicated stimulants for 24 h. IL-8 content in the culture super- expression of C/EBP decreased slightly, with both macro- natant was determined by ELISA. phage and neutrophil differentiation, whereas expression of C/EBP␤ signiﬁcantly increased. These results suggest that Models of lineage switching in hematopoietic development Y Yamaguchi et al 1435 Discussion

In this paper, we report the establishment of a new human eosinophilic cell line (YJ) derived from the peripheral blood of a patient with chronic myelomonocytic leukemia (CMMoL) with eosinophilia. For the first 3 months of culture, the polymorphonuclear (PMN) fraction of cells from the peripheral blood required IL-5 for maintenance of proliferation, after which they became factor-independent. This suggests that only the cells capable of proliferating in response to IL-5 developed the ability to grow autonomously. The proliferation of the YJ cell line may therefore have been due to an autocrine mechanism involving IL-5. Human eosinophilic cell lines established to date include the following: HL-60-C15 cells,5,6 HL-60-3+C5 cells,7 EoL cells,4 and AML 14 cells.8 The HL-60-C15 and HL-60-3+C5 eosinophil sublines were derived from the acute promyelocytic leukemia cell line HL-60, and the EoL cell line was derived from an eosinophilic leukemia. The AML14 cell line was derived from an acute myelocytic leukemia, M2.8 The YJ line reported here was derived from CMMoL, and the YJ cell line is the only eosinophilic cell line to be derived from a myelodysplastic syndrome (MDS) patient among the eosino- Figure 4 Expression of CD14 in YJ cells induced with 10 nM PMA for 2 days. Cells were incubated with biotinylated CD14 and FITC- philic cell lines established to date. Table 2 shows the conjugated streptavidin. (A) unstained YJ cells, (B) YJ cells cultured chromosomal abnormalities reported in eosinophilic cell lines without PMA, (C) YJ cells cultured with 10 nM PMA for 2 days. established to date. One chromosomal abnormality in all cell lines involves chromosome 9. In human neoplasia, 5.6% of the cytogenetic aberrations described involve chromosome 9,39 most of which are the Ph1 chromosome. Sreekantaiah et al40 have reported five patients with acute nonlymphocytic C/EBP␣ plays an important role in immature myeloid cells and leukemia (ANLL) with chromosomal aberrations involving that C/EBP␤ may regulate the cellular functions of more bands 9q21-q22. The diagnoses were ANLL FAB type M4 mature myeloid cells, macrophages and neutrophils. The (myelomonocytic leukemia), M4 with eosinophilia (M4eo), expression of Egr-1, which apparently plays a role in the M4, M5 (monocytic leukemia), and M2 (myelocytic development of hematopoietic cells along the macrophage leukemia), respectively. The report by Sreekantaiah et al40 lineage, was induced in YJ cells with PMA stimulation and suggests that abnormalities in chromosome 9 may be associa- macrophage differentiation. However, the Egr-1 transcript was ted with the differentiation of myelocytic or monocytic lin- also increased with RA-induced neutrophil differentiation. The eages. Recently, it has been shown that homozygous deletion level of MZF-1 mRNA, which is an apparently necessary fac- of the cyclin-dependent-4 inhibitor gene (CD4I; p16), which is tor for granulocytic differentiation,37 was also increased some- mapped to chromosome 9p21, is one of the tumor-suppressor what in YJ cells differentiated with RA towards the neutrophil genes, that may result in immortalization of human leukemia lineage. PU.1, the product of the Spi-1 oncogene and a mem- cells.41 Cytogenetic study of our patient’s BM cells did not ber of the ets family, is essential for myeloid cell develop- show deletion of chromosome 9, but YJ cells did. This may ment.38 Expression of PU.1 mRNA was increased slightly after indicate that only those cells with deletion of p16 were cap- RA treatment, but remained unchanged in response to PMA. able of becoming immortalized or that the YJ cell deletion of These results suggest that the Egr-1 and C/EBP␤ transcription p16 developed in long-term culture of the leukemic cells. factors accumulate during differentiation of YJ cells to both We have previously reported the presence of precursor cells macrophage and neutrophil lineages. in bone marrow which are able to respond to both G-CSF and IL-5 and form trilineage (neutrophils, macrophages, and eosinophils) colonies.42 In addition, it was reported that large Activity of various stimuli on the release of eosinophil eosinophil colonies derived from single cells contain a very granule MBP from YJ cells small number of neutrophils or macrophages.43 For the present studies, we speculate that YJ cells are induced towards To determine whether YJ cells respond to various cytokines, the monocyte/macrophage or neutrophil lineages from blast chemokines, or other activating stimuli in a manner equival- cells capable of differentiating the eosinophil lineage in ent to mature human eosinophils, we examined the effect of response to PMA or RA stimulation (Figure 8). From the evi- ADF/TRX, IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-8, GM-CSF, and dence described above, the developmental stage of YJ cells RANTES on MBP secretion by YJ cells. As shown in Figure might be slightly advanced still toward the eosinophil lineage 7, rADF, IL-3, IL-5, GM-CSF, RANTES, and IL-4 significantly at a stage which can nevertheless differentiate to all three increased MBP release from YJ cells compared to a control in myeloid lineages (eosinophils, macrophages and neutrophils) which YJ cells were cultured in the absence of cytokines. depending on the inducing stimulus (Figure 8). Paul et al44 Among the cytokines examined, IL-4 (40 U/ml) augmented the reported that a cloned subline of the AML14 cell line release of MBP most potently. These results suggest that the (AML14.3D10), which was capable of spontaneously differen- YJ cell line may be useful for analyzing the mechanisms of tiating to eosinophilic myelocytes in the absence of cytokine eosinophil activation, signal transduction, or degranulation. stimulation, could change in the differentiation program to the Models of lineage switching in hematopoietic development Y Yamaguchi et al 1436

Figure 5 NBT reduction test in YJ cells treated with various concentrations of RA. YJ cells were cultured without or with 20, 200 and 2000 nM RA for 4 days. (a) Only neutrophilic cells with a large black deposit were classiﬁed as NBT positive. (b) Quantitation of the NBT reduction test of cells shown in (a).

Figure 7 Effect of various agonists on MBP release from YJ cells. YJ cells (5 × 104 cells/ml) were incubated with medium alone, ADF (10 ␮g/ml), IL-1␣ (20 U/ml), IL-2 (20 U/ml), IL-3 (50 ng/ml), IL-4 (40 U/ml), IL-5 (50 ng/ml), IL-6 (50 ng/ml), IL-8 (50 ng/ml), GM-CSF (10 ng/ml), or RANTES (10 nM) for 2 h at 37°C. After incubation, the amount of MBP in the culture supernatants was measured by ELISA. The mean ± s.d. for triplicate determinations is shown. *P Ͻ 0.05

neutrophil lineage after exposure to RA. They suggest that the lineage switch occurs at least in part by a change in the differentiation program of cells that already show advanced evi- Figure 6 Northern blot analysis for transcription factors associated dence of eosinophilic differentiation. Accordingly, it can not with myeloid differentiation. HL-60-C15 cells (lane 1), untreated YJ be excluded that the differentiation of YJ cells to cells (lane 2), YJ cells differentiated towards macrophage (lane 3) or monocyte/macrophage or neutrophil lineages do not develop neutrophil (lane 4) lineages. Lane 1, Eosinophil-committed cell line, from the blast-like progenitors in YJ cells, but from the promy- HL-60-C15 cells; lane 2, uninduced YJ cells; lane 3, YJ cells treated elocytic stage of eosinophils in YJ cells. with 20 nM PMA for 2 days; lane 4, YJ cells treated with 2 ␮M RA for 4 days. Each lane contained 4 ␮g of poly(A)+ mRNA. The blot was It has been reported that GATA-1 mRNA is expressed in not probed sequentially with each cDNA probe and an 18S rRNA probe only erythroid and megakaryocyte but also eosinophil and to control for equivalent loading. mast cell lineages. We have previously demonstrated Models of lineage switching in hematopoietic development Y Yamaguchi et al 1437 Table 2 Chromosomal abnormalities in eosinophilic cell lines

Cell line Cytogenetic analysis

HL-60-C15 −5, −9, −10, −16, −16,−17, +?18, der(5)t(5:17)(q13;q12), +der(9)t(9;?)(q34.3;?), +der(10)t(10;13)(p11.2; q12.1), +der(16)t(16;?)(q22;?), +der(16)t(16;?)(123;?), del(9)(q13) EoL-1 48XY, +6, +8, 9q- EoL-2 49XY, +8, 9q-, +12, +13 EoL-3 49XY, +4, +8, 9q-, +13 AML 14 −5, +8, −9, +13, −14, −18, −21, +der(3), del(5q) AML14.3D 10 65-84XY, +X, del(5)(q23;q34), +6, +6, −7, add(8)(Q24.3), del(9)(q22), add(13)(p13), add(14)(p13), add(14)(p13), +16, add17(p13), +19, −20, −21, add(22)(q13), +12–14 mar YJ X, Y, −2, −5, −9, −14, −14, −15, −16, −17, add(3)(q27), del(3)(p13), add(4)(q34), add(6)(q25), del 10(p12), add(11)(p15), add13(q32), add16(q23), del(18)(q22)

In addition, it has been shown that eosinophil MBP itself stimulates the production of IL-8 by human eosinophils.48 Our results indicate, however, that the YJ eosinophil cell line is a poor producer of IL-8, but did turn out to be a good producer when differentiated to the monocyte/macrophage lineage by PMA treatment. To date, the transcription factors shown to be involved in monocyte/macrophage differentiation include Egr-1 and AP- 1.49,50 In YJ cells, macrophage differentiation induced by PMA was accompanied by an up-regulation of Egr-1 mRNA level (Figure 6) and protein at an early phase (data not shown). Mol- linedo et al50 have reported that expression of AP-1 is not required for the induction of HL-60 cells towards granulo- cytes, whereas induction of monocytic differentiation was cor- related with an increase in AP-1 activity. Although there is some evidence that AP-1 and Egr-1 play a role in monocyte differentiation, the mechanism by which transcription factors such as AP-1 and Egr-1 act during monocytic differentiation remains unknown. In contrast, the paper has also been reported that Egr-1 is not required for macrophage differentiation or activation.51 Accordingly, the role of Egr-1 on the differentiation along the macrophage lineage is controversial. In addition, the effect of Egr-1 on the differentiation or acti- Figure 8 The YJ cell line is composed of blastoid progenitor cells vation of myeloid lineage remains to be elucidated. In these which undergo a constitutive eosinophilic differentiation. These blas- regards, the YJ cell line may be useful for elucidating the mol- toid progenitors can differentiate towards the macrophage or neutro- ecular mechanisms that govern lineage switching of myeloid phil lineages in response to PMA or RA, respectively. progenitors amongst the eosinophil, monocyte/macrophage and neutrophil hematopoietic programs. increased expression of mRNA for the GATA-binding proteins GATA-1, GATA-2, and GATA-3 during eosinophil differentiation of myeloid leukemic cell lines and constitutive 45 expression in mature eosinophils and basophils. Further- Acknowledgements more, we have shown a major positive regulatory role for GATA-1 and negative regulatory role for GATA-2 in the tran- 46 scription of eosinophil major basic protein (MBP) gene. In This work was supported by Grants-in-Aid from the Ministry addition, as shown in Figure 6, expression of GATA-1 mRNA of Education, Science, and Culture of Japan (YY). SJA is sup- was very little changed in uninduced YJ cells, YJ cells induced ported by Grants AI 33043 and AI 25230 from the National with PMA, and YJ cells induced with RA. These results indi- Institute of Health, USA. We thank Mr Masayuki Sonoyama cate that GATA-1 may play an important role not only in the for performing the chromosomal analysis of this cell line, Dr differentiation of erythroid/megakaryocyte but also in the Motohiro Hashiyama for the flow cytometric analysis. We also functions of myeloid cells including eosinophils, macrophages thank Dr Ronald G Crystal, Dr Michio Nakamura, Dr Shun- and neutrophils. suke Yamamoto, Dr Masayuki Yamamoto, Dr Kleanthis G We have provided evidence that the YJ cell line is able to Xanthopoulos, Dr Shizuo Akira, Dr Takashi Yokota, Dr Robert produce high levels of IL-8 in response to PMA and TNF-␣, Hromas, and Dr Françoise Moreau-Gachelin for providing the but not other agents, including IL-1␣, IL-5 and C5a. Recently, neutrophil elastase, gp91phox, CD14, GATA-1 and -2, C/EBP␣, it has been reported that human mature eosinophils are able ␤ to produce IL-8 in response to C5a or calcium ionophore.35,47 C/EBP , Egr-1, MZF-1, and PU.1 cDNAs, respectively. Models of lineage switching in hematopoietic development Y Yamaguchi et al 1438 References Activity and tissue-specific expression of the transcription factor NF-E1 multigene family. Gene Dev 1990; 4: 1650–1662. 1 Wimmer RS, Raney RB, Naiman JL. Hypereosinophilia with acute 24 Dorfman DM, Wilson DB, Bruns GA, Orkin SH. Human transcrip- lymphocytic and acute myelomonocytic leukemia in childhood. J tion factor GATA-2. Evidence for regulation of preproendothelin- Pediatr 1978; 92: 244–246. 1 gene expression in endothelial cells. J Biol Chem 1992; 267: 2 Marione G, Rossi G, Verzura P. Eosinophilic blast crisis in a case 1279–1285. of chronic myeloid leukemia. Br J Hematol 1983; 55: 251–256. 25 Antonson P, Xanthopoulos KG. Molecular cloning, sequence, and 3 Benvenisti DS, Ultmann JE. Eosinophilic leukemia. Report of five expression patterns of the human gene encoding CCAAT/enhancer cases and review of literature. Ann Intern Med 1969; 71: 731–745. binding protein ␣ (C/EBP␣). Biochem Biophys Res Commun 1995; 4 Saito H, Bourinbaiar A, Ginsburg M, Minato K, Ceresi E, Yamada 215: 106–113. K, Machover D, Breard J, Mathe G. Establishment and characteriz- 26 Akira S, Issiki H, Sugita T, Tanabe O, Kinoshita S, Nishio Y, Naka- ation of a new human eosinophilic leukemia cell line. Blood 1985; jima T, Hirano T, Kishimoto T. A nuclear factor for IL-6 expression 66: 1233–1240. (NF-IL6) is a member of a C/EBP family. EMBO J 1990; 9: 5 Fishkoff SA. Graded increase in probability of eosinophilic differ- 1897–1906. entiation of HL-60 promyelocytic leukemic cells induced by cul- 27 Suggs SV, Katzowitz JL, Tsai-Morris C, Sukhatme VP. cDNA ture under alkaline conditions. Leukemia Res 1988; 12: 679–686. sequence of the human cellular early growth response gene Egr- 6 Fishkoff SA, Pollak A, Gleich GJ, Testa JR, Misawa S, Reber TJ. 1. Nucleic Acids Res 1990; 18: 483. Eosinophilic differentiation of the human promyelocytic leukemia 28 Hromas R, Collins SJ, Hickstein D, Raskind W, Deaven LL, O’Hara cell line, HL-60. J Exp Med 1984; 160: 179–196. P, Hagen FS, Kaushansky K. A retinoic acid-responsive human 7 Tomonaga M, Gasson JC, Quan SG, Golde DW. Establishment of zinc finger gene, MZF-1, preferentially expressed in myeloid cells. eosinophilic sublines from human promyelocytic leukemia (HL- J Biol Chem 1991; 266: 14183–14187. 60) cells: demonstration of multipotentiality and single-lineage 29 Ray D, Culine S, Tavitian A, Moreau-Gachelin F. The human commitment of HL-60 stem cells. Blood 1986; 67: 1433–1441. homologue of the putative proto-oncogene Spi-1: characterization 8 Paul CC, Tolbert M, Mahrer S, Singh A, Grace MJ, Baumann MA. and expression in tumors. Oncogene 1990; 5: 663–668. Cooperative effects of interleukin-3 (IL-3), IL-5, and granulocyte– 30 Radzioch D, Clayton M, Varesio L. Interferon-␣,-␤, and -␥ aug- macrophage colony-stimulating factor: a new myeloid cell line ment the levels of rRNA precursors in peritoneal macrophage but inducible to eosinophils. Blood 1993; 81: 1193–1199. not in macrophage cell lines and fibroblasts. J Immunol 1987; 139: 9 Silberstein DS, David JR. The regulation of human eosinophil func- 805–812. tion by cytokines. Immunol Today 1987; 8: 380–385. 31 Ko Y, Mukaida N, Panyutich A, Voitenok NN, Matsushima K, 10 Yamaguchi Y, Suda T, Suda J, Eguchi M, Miura Y, Harada N, Tom- Kawai T, Kasahara T. A sensitive enzyme-linked immunosorbent inaga A, Takatsu K. Purified interleukin 5 supports the terminal assay for human interleukin-8. J Immunol Meth 1992; 149: differentiation and proliferation of murine eosinophilic precursors. 227–235. J Exp Med 1988; 67: 43–56. 32 Koyanagi H, Wakasugi N, Yoshimatsu K, Takashima Y, Yodoi J, 11 Clutterbuck EJ, Hirst EM, Sanderson CJ. Human interleukin-5 (IL- Momoi M, Suda T, Kasahara T, Yamaguchi Y. Role of ADF/TRX 5) regulates the production of eosinophils in human bone marrow and its inhibitor on the release of major basic protein from human cultures: comparison and interaction with IL-1, IL-3, IL-6, and GM-CSF. Blood 1989; 73: 1504–1512. eosinophils. Biochem Biophys Res Commun 1995; 213: 1140– 12 Karasuyama H, Melchers F. Establishment of mouse cell lines 1147. which constitutively secrete large quantities of interleukin 2, 3, 4, 33 Baehner RL, Nathan DG. Quantitative nitroblue tetrazolium test or 5, using modified cDNA expression vectors. Eur J Immunol in chronic granulomatous disease. New Engl J Med 1968; 278: 1988; 18: 97–104. 971–976. 13 Chomzynski P, Sacchi N. Single-step RNA isolation by acid guani- 34 Ohara A, Suda T, Tokuyama N, Suda J, Nakayama K, Miura Y, dinium thiocyanate-phenol-chloroform extraction. Anal Biochem Nishikawa S, Nakauchi H. Generation of B lymphocytes from a 1987; 162: 156–159. single hemopoietic progenitor cell in vitro. Int Immunol 1991; 3: 14 Sakamaki K, Tomonaga M, Tsukui K, Nagata S. Molecular cloning 703–709. and characterization of a chromosomal gene for human eosino- 35 Braun RK, Franchini M, Erard F, Rihs S, Devries IJ, Blaser K, Hansel phil peroxidase. J Biol Chem 1989; 264: 16828–16836. TT, Walker C. Human peripheral blood eosinophils produce and 15 Barker RL, Gleich GJ, Pease LR. Acid precursor revealed in human release interleukin-8 on stimulation with calcium ionophore. Eur eosinophil granule major basic protein cDNA. J Exp Med 1988; J Immunol 1993; 23: 956–960. 168: 1493–1498. 36 Nakajima H, Gleich GJ, Kita H. Constitutive productions of IL-4 16 Rosenberg HF, Ackerman SJ, Tenen DG. Human eosinophil cat- and IL-10 and stimulated production of IL-8 by normal peripheral ionic protein. Molecular cloning of a cytotoxin and hel- blood eosinophils. J Immunol 1996; 156: 4859–4866. minthotoxin with ribonuclease activity. J Exp Med 1989; 170: 37 Bavisotto L, Kausansky K, Lin N, Hromas R. Anti-sense 163–176. oligonucleotides from the stage-specific myeloid zinc finger gene 17 Rosenberg HF, Tenen DG, Ackerman SJ. Molecular cloning of the MZF-1 inhibit granulopoiesis in vitro. J Exp Med 1991; 174: human eosinophil-derived neurotoxin: a member of the ribonu- 1097–1101. clease gene family. Proc Natl Acad Sci USA 1989; 86: 4460–4464. 38 Chen H, Zhang P, Voso MT, Hohaus S, Gonzalez DA, Glass CK, 18 Ackerman SJ, Corrette SE, Rosenberg HF, Bennett JC, Mastriani Zhang D-E, Tenen DG. Neutrophils and monocytes express high DM, Nichoson-Weller A, Weller PF, Chin DJ, Tenen DG. Molecu- levels of PU.1 (Spi-1) but not Spi-B. Blood 1995; 85: 2918–2928. lar cloning and characterization of human eosinophil Charcot– 39 Mitelman F. Catalog of chromosome Aberrations in Cancer, third Leyden crystal protein: similarity to IgE-binding protein and the S- edn. Alan R Liss: New York, 1988. type animal lectin superfamily. J Immunol 1993; 150: 456–468. 40 Sreekantaiah C, Baer MR, Preisler HD, Sandberg AA. Involvement 19 Takahashi H, Nukiwa T, Basset P, Crystal RG. Myelomonocytic of bands 9q21-q22 in five cases of acute nonlymphocytic leuke- cell lineage expression of the neutrophil elastase gene. J Biol mia. Cancer Genet Cytogenet 1989; 39: 55–64. Chem 1988; 263: 2543–2547. 41 Ogawa S, Hirano N, Sato N, Takahashi T, Hangaishi A, Tanaka 20 Bjorgvinsdottir H, Zhen L, Dinauer MC. Cloning of murine K, Kurokawa M, Tanaka T, Mitani K, Yazaki Y, Hirai H. Homo- gp91phox cDNA and functional expression in a human X-linked zygous loss of the cyclin-dependent kinase 4-inhibitor (p16) gene chronic granulomatous disease cell line. Blood 1996; 87: 2005– in human leukemia. Blood 1994; 84: 2431–2435. 2010. 42 Yamaguchi Y, Suda T, Suda J, Eguchi M, Miura Y, Mita S, Hitoshi 21 Morishita K, Kubota N, Asano S, Kaziro Y, Nagata S. Molecular Y, Tominaga A, Takatsu K. Analysis of eosinophil-predominant cloning and characterization of cDNA for human myeloperoxi- colonies formed by human hemopoietic precursor cells in the dase. J Biol Chem 1987; 262: 3844–3851. presence of purified interleukin-5. Acta Hematol 1990; 53: 22 Nasu N, Yoshida S, Akizuki S, Higuchi Y, Setoguchi M, Yamamoto 688–698. S. Molecular and physiological properties of murine CD14. Int 43 Suda T, Suda J, Ogawa M, Ihle JN. Permissive role of interleukin Immunol 1991; 3: 205–213. 3 (IL-3) in proliferation and differentiation of multipotential hemo- 23 Yamamoto M, Ko LJ, Leonard MW, Beug H, Orkin SH, Engel JD. poietic progenitors in culture. J Cell Physiol 1985; 124: 182–190. Models of lineage switching in hematopoietic development Y Yamaguchi et al 1439 44 Paul CC, Mahrer S, Tolbert M, Elbert BL, Wong I, Ackerman SJ, 48 Kita H, Abu-Ghazaleh RI, Sur S, Gleich GJ. Eosinophil major basic Baumann MA. Changing the differentiation program of protein induces degranulation and IL-8 production by human hematopoietic cells: retinoic acid-induced shift of eosinophil- eosinophils. J Immunol 1995; 154: 4749–4758. committed cells to neutrophils. Blood 1995; 86: 3737–3744. 49 Nguyen HQ, Hoffman-Liebermann B, Liebermann DA. The zinc 45 Zon LI, Yamaguchi Y, Yee K, Albee EA, Kimura A, Bennett JC, finger transcription factor egr-1 is essential for and restricts differ- Orkin SH, Ackerman SJ. Expression of mRNA for the GATA-bind- entiation along the macrophage lineage. Cell 1993; 72: 197–209. ing proteins in human eosinophils: potential role in gene transcrip- 50 Mollinedo F, Gajate C, Tgores A, Flores I, Naranjo JR. Differences tion. Blood 1993; 81: 3234–3241. in expression of transcription factor AP-1 in human promyelocytic 46 Yamaguchi Y, Ackerman SJ, Minegishi N, Yamamoto M, Suda T. HL-60 cells during differentiation towards macrophages vs gra- Mechanisms of transcription in eosinophils: GATA-1 but not nulocytes. Biochem J 1993; 294: 137–144. GATA-2 transactivates the promoter of the eosinophil granule 51 Lee SL, Wang Y, Milbrandt J. Unimpaired macrophage differen- major basic protein gene. Blood 1998; 91: 3447–3458. tiation and activation in mice lacking the zinc finger transcription 47 Miyamasu M, Hirai H, Takahashi M, Iida M, Yamaguchi M, Kosh- factor NGFI-A (EGR1). Mol Cell Biol 1996; 16: 4566–4572. ino T, Takaishi T, Morita Y, Ohta K, Kasahara T, Ito K. Chemotac- tic agonists induce cytokine generation in eosinophils. J Immunol 1995; 154: 1339–1349.