(1997) 11, 1469–1477  1997 Stockton Press All rights reserved 0887-6924/97 $12.00

Two (AML-M5a) cell lines (MOLM-13 and MOLM-14) with interclonal phenotypic heterogeneity showing MLL-AF9 fusion resulting from an occult insertion, ins(11;9)(q23;p22p23) Y Matsuo1, RAF MacLeod2, CC Uphoff2, HG Drexler2, C Nishizaki1, Y Katayama3, G Kimura3, N Fujii3, E Omoto3, M Harada3 and K Orita1

1Fujisaki Cell Center, Hayashibara Biochemical Labs, Inc., Okayama; 3Department of Medicine II, Okayama University Medical School, Okayama, Japan and 2DSMZ-German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany

We describe two new human leukemia cell lines, MOLM-13 and and MOLM-14, established from this patient at relapse, in MOLM-14, established from the peripheral of a patient which RT-PCR and molecular cytogenetic analysis revealed at relapse of acute monocytic leukemia, FAB M5a, which had evolved from (MDS). Both cell lines MLL-AF9 fusion resulting from a micro-insertion which had express -specific esterase (MSE) and MLL-AF9 fusion been overlooked at diagnosis. mRNA. Gene fusion is associated with a minute chromosomal Continuous cell lines provide permanent sources of insertion, ins(11;9)(q23;p22p23). MOLM-13 and MOLM-14 are materials for researchers in several areas,5 but may be of parti- the first cell lines with, and represent the third reported case cular interest for studies on the molecular changes involving of, MLL gene rearrangement arising via chromosomal insertion. recurrent breakpoints. Leukemia–lymphoma cell lines have Both cell lines carry trisomy 8 which was also present during been instrumental in the molecular study of many break- the MDS phase, as well as the most frequent trisomies associa- 6 ted with t(9;11), ie, +6, +13, +19 variously present in different points. subclones. Despite having these features in common, differ- ences in antigen expression were noted between the two cell ؉ ؊ ؊ ؉ lines: that of MOLM-13 being CD34 , CD13 , CD14 , CD15 , Case report ,CD33؉; whereas MOLM-14 was CD4؉, CD13؉, CD14؉, CD15؉ ؉ CD33 . Differentiation to -like morphology could 4 be induced in both cell lines after stimulation with INF-␥ alone, The clinical details have been reported in detail elsewhere. or in combination with TNF-␣, which treatment also induced or Briefly, a 20-year-old male patient was readmitted to Oka- upregulated, expression of certain -associated yama University Hospital with fever and swelling of the lymph antigens, including CD13, CD14, CD15, CD64, CD65 and CD87. nodes in May 1995. In February 1995 he had received a trans- Together, these data confirm that both cell lines are likely to fusion of red blood cells and antibiotic treatment combined be novel in vitro models for studying monocytic differentiation with colony-stimulating factor (G-CSF) for hypo- and leukemogenesis. Keywords: cell lines; AML-M5a; interclonal phenotypic hetero- cellular MDS (refractory anemia with excess of blasts: RAEB) geneity; insertion; MLL-AF9 after which trilineage hematopoietic recovery was attained without detectable blast infiltration. At diagnosis of MDS, interface cytogenetic and RT-PCR analyses, respectively, Introduction showed trisomy 8 and absence of AF9-MLL rearrangement in the bone marrow (BM). On readmission with florid leukemia, 3 A subtle, reciprocal translocation exchanging the terminal his peripheral count was 13200/mm , with 56% leukemic blasts. His hemoglobin concentration was short and long arm segments of 9 and 11, 3 respectively t(9;11)(p21-22;q23), is associated with acute 15.0 g/dl, and his count was 22 000/mm . The bone (AML)-M5,1 particularly the M5a subtype.2 marrow aspiration showed 92% leukemic blasts, and the mor- phological diagnosis was made as AML-M5a. Cytogenetic Ascertainment may be difficult in suboptimal preparations + and, despite being regarded as the ‘standard’ cytogenetic analysis was interpreted as: 47, XY, 8, t(9;11)(p22;q23). change in acute monoblastic leukemia, its overall incidence Immunophenotyping analysis of fresh leukemia blasts and pattern of associations remain uncertain.2 A recent study revealed no significant expression of CD antigens associated comparing AML-M1 and -M5 patients, analyzed simul- with the myelo–monocytic lineage. CD34 was 30% positive taneously by reverse transcriptase-polymerase chain reaction and non-lineage-associated HLA-DR was found to be positive (RT-PCR), Southern blotting and fluorescence in situ hybridiz- at 70%; whereas those indicative of lymphoid lineage includ- ation (FISH) with an MLL-specific yeast artificial chromosome ing CD3, CD4, CD8, CD10, CD19 were absent or weakly (YAC) probe, suggests the overall incidence of MLL rearrange- expressed. Myeloperoxidase activity was weakly detected on ment in AML-M5 may be as high as 60%.3 It was apparent in the fresh leukemic blasts. Despite receiving chemotherapy, he that study that approximately half the cases with MLL succumbed to rapidly progressive leukemia on 15 August rearrangement went undetected by cytogenetic methods, 1995. including a cryptic t(6;11)(q27;q23) resulting from a cyto- genetically invisible insertion juxtaposing AF6 and MLL. A case of AML-M5a with de novo MLL-AF9 fusion evolving Materials and methods from MDS with trisomy 8 present at all phases has recently been described.4 We describe a pair of cell lines, MOLM-13 Cell culture During relapse, after chemotherapy, a heparinized peripheral Correspondence: Y Matsuo, Fujisaki Cell Center, Hayashibara blood specimen, obtained with informed consent, was pro- Biochemical Labs, Inc., 675-1 Fujisaki, Okayama 702, Japan vided by Okayama University Hospital, Second Department Received 3 April 1997; accepted 19 May 1997 of Medicine. Mononuclear cells were isolated by Ficoll– New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1470 Hypaque density centrifugation. The interface containing anti-human immunoglobulins (Igs) specific for ␬ and ␭ mononuclear cells was harvested and washed twice with fresh chains and for ␣, ␦, ␥ and ␮ heavy chains (Cappel Labora- RPMI 1640 (Nissui Pharmaceutical, Tokyo, Japan) culture tories, West Chester, PA, USA), rabbit anti-terminal deoxynu- medium. Mononuclear cells were suspended in RPMI 1640 cleotidyl transferase (TdT) (P-L Biochemicals, Milwaukee, WI, medium supplemented with 10% fetal calf serum (FCS: USA), monoclonal antibodies (mAbs) CD3 (NU-T3), CD4 GIBCO, Grand Island, NY, USA) and antibiotics, 100 U/ml (NU-Th/i), CD8 (NU-Ts/c), CD10 (NU-N1), CD20 (NU-B2), penicillin and 50 ␮g/ml streptomycin without any other CD13 (MCS-2), NU-Ia for HLA-class II (Nichirei, Tokyo, growth factors, and were incubated in two plastic culture Japan), CD19 (B4), CD33 (MY9), CD34 (MY10) (Coulter ° flasks at 37 C in a humidified 5% CO2 atmosphere. Each cul- , Hialeah, FL, USA). All mAbs were used in an ture was then fed once a week by replacing one half to one indirect immunofluorescence (IF) test with FITC-conjugated third of the volume of the culture contents with fresh nutrient goat anti-mouse IgG or IgM (Cappel Laboratories). In addition medium during the subsequent 2 weeks. In the third week, a to the immunophenotyping of the fresh leukemia cells, the slow yet sustained proliferation of cultured cells was noted following mAbs were used to type the established cell lines: in both flasks which were then designated as MOLM-13 and CD11a (SPV-L7; purchased from Nichirei); CD9 (BA-2; gift MOLM-14. from Dr TW LeBien, University of Minnesota, Minneapolis, The cell lines were found to be free of mycoplasma infec- MN, USA); CD14 (MY4; purchased from Coulter tion using standard broth-agar cultivation and DAPI staining. Immunology); CD15 (LeuM1; purchased from Becton Dickin- son, Mountain View, CA, USA); CD120A and CD120B (MR1- 4 and MR2-1, respectively; a gift from Dr WA Buurman, Uni- Morphological studies versity of Limburg, Maastricht, The Netherlands); CD115 Cytospin smears of the MOLM-13 and MOLM-14 cells, (D171; obtained through Dr T Kishimoto, Osaka University, stained with Wright–Giemsa, were used for cytochemical Japan via the 6th International Workshop on Human Leuko- myeloperoxidase (POX) staining. cyte Differentiation Antigens (HLDA); and all others were obtained through Dr S Shaw (National Institute of Health, Bethesda, MD, USA) (via the 5th International Workshop on Immunophenotyping HLDA). Immunofluorescent-positive cells were determined by fluorescent microscopy (Nikon, Tokyo, Japan). Immunophenotyping of fresh leukemia blasts from the peri- pheral blood was performed using the following antibodies (Table 1): fluorescein isothiocyanate (FITC)-conjugated goat Isoelectric focusing and isoenzyme staining Table 1 Marker profiles of MOLM-13 and MOLM-14 Enzyme extraction, separation by isoelectric focusing (IEF) and (% immunofluorescence positive visualization of esterase isoenzymes have been described in cells) detail elsewhere.7 In short, enzymes were extracted by repeated cycles of freezing and thawing and were solubilized Fresh leukemia MOLM-13 MOLM-14 by addition of Triton X-100 (Serva, Heidelberg, Germany). Ali- cells quots of enzyme-containing supernatants (referring to equal numbers of cells per sample) were separated on horizontal TdT Ͻ100polyacrylamide thin-layer gels composed of 4.8% CD4 NU-TH/I Ͻ 10 100 acrylamide/bisacrylamide and ampholytes of pH range 2–11 CD9 BA-2 nt 5 80 (Servalyt; Serva) using an LKB Multiphor System (Pharmacia, CD11a SPV-L7 NT 100 100 Ͻ Freiburg, Germany). The bands of the isoenzymes were vis- CD13 MCS-2 5580 ␣ CD14 MY4 NT 5 50 ualized directly on the gel by staining with -naphthyl acetate CD15 LeuM1 NT 50 80 (Sigma, Deisenhofen, Germany) as substrate and Fast Blue RR CD32 IV.3 NT 100 100 (Serva) as coupling diazo dye. Addition of 40 mM NaF inhibits CD33 MY9 Ͻ5 100 100 specifically the monocyte-specific esterase (MSE) bands, but CD34 MY10 30.6 0 0 does not affect the non-monocytic isoenzymes.8 CD64 22.2 NT 0 100 CD65 VIM-8 NT 5 100 CD68 Y-1/82a NT 0 0 CD87 3B10 NT 10 100 Cytogenetic preparation and analysis CD91 A2MR NT 0 0 CD92 VIM-15 NT 10 80 Standard cytogenetic harvesting, preparation and staining pro- CD93 VIM-D2 NT 100 100 cedures were used throughout. Briefly, mitotic cells were har- CD115 7-7A3-17 NT 0 10 ␮ CD116 hGMCSF NT 100 100 vested by colcemid arrest (0.004 M for 1.5 h) and, after hypo- CD119 GIR-208 NT 100 100 tonic treatment with 0.075 M KCl and 0.9% NaCitrate (1:1 for CD120A MR1-4 NT 0 0 5 min), fixed in chilled, methanol:acetic acid (3:1). Trypsin G- CD120R MR2-1 NT 0 10 banding and fluorescence in situ hybridization (FISH) were CD155 D171 NT 0 100 performed on slides after ageing for 10 days. For G-banding, HLA-DR NU-Ia1 70.0 0 0 slides were incubated briefly (10–15 s) in trypsin solution9 and stained for 15 min in 5% Giemsa in Sorensen’s buffer For the fresh leukemia cells, the values shown are the percent (pH 6.8). Analysis was performed using a Zeiss Axioplan (Carl immunofluorescence positive cells as analyzed by immunofluo- rescence microscopy. NT, not tested. For MOLM-13 and MOLM-14, Zeiss, Oberkochen, Germany) microscope configured to an the values shown are the means of percent immunofluorescence- image-analysis system (Cytoscan 3; Applied Imaging, War- positive cells as analyzed by immunofluorescence microscopy in rington, UK) equipped with a high-resolution laser printer multiple tests. (Lasertechnics, Tuscon, AZ, USA). Consensus karyotypes were New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1471 obtained by full analysis of 25 or more metaphases. FISH was Cytokine treatment performed using the following probes: MLL (Appligene Oncor, Heidelberg, Germany) chromosome libraries indirectly lab- The cells were cultured in RPMI 1640 with several stimuli. eled with FITC covering chromosomes 8, 11, 14 and 19 Prior to starting the experiments, the cultured cells were har- (Cambio, Cambridge, UK) and chromosome 9 directly labeled vested and washed twice with RPMI 1640 medium, and resus- with Spectrum Orange (Gibco-BRL, Paisley, UK). Hybridiz- pended in fresh medium; then the cell count was adjusted to ation and detection followed manufacturer’s protocols except 5 × 105 cells/ml. INF-␥ (1000 IU/ml), TNF-␣ (3 Japanese refer- that the stringency of the post-hybridization washes after ence units; JRU/ml) (Hayashibara Biochemical Labs, Okay- hybridizing with chromosome painting probes was reduced ama, Japan), macrophage colony-stimulating factor (M-CSF) by washing in 1 × SSC, intended to intensify signal strength, (50 U/ml), granulocyte–macrophage colony-stimulating factor albeit at the expense of ‘background noise’. Slides were coun- (GM-CSF) (20 ng/ml) (Genzyme, Cambridge, MA, USA) were terstained with 4,6-diamidino-2-phenylindole (DAPI) (Sigma) added to the culture at the appropriate conditions. The cells ° and mounted in a commercial antifade solution (Vectashield; were incubated at 37 C in an humidified 5% CO2 atmosphere Camon, Wiesbaden, Germany) and examined microscopi- for 3 days. After incubation, differentiation was judged by cally by epifluorescence. Photographic images of fluores- cellular morphology and CD antigen expression. Briefly, the cently illuminated chromosome preparations were recorded stimulated cells were harvested after the incubation period, on Kodak Ektachrome 400 daylight reversal film. and washed twice with PBS (phosphate-buffered saline). The Fc receptors on the cell surfaces were blocked with 1 mg/ml ␥ Reverse Transcriptase-Polymerase Chain Reaction (RT- -globulin (Sigma, St Louis, MO, USA) for 15 min at room PCR) temperature, then the cells were washed once with PBS. Using the mAbs listed in Table 2, an indirect IF test was performed Total RNA was isolated from cell pellets using standard guani- by the method described in the immunophenotyping section. dinium thiocyanate and ultracentrifugation methods. Five Immunofluorescent-positive cells were determined by flow micrograms of the total RNA were used for the synthesis of cytometry (EPICS-Profile; Coulter Electronics, Hialeah, FL, first strand cDNA applying a reverse transcriptase preamplifi- USA). cation (SuperScript; Gibco BRL, Eggenstein, Germany). The reverse transcription was carried out with 50 ng of random hexamers in a final volume of 20 ␮l RT buffer (containing Results

20 mM Tris-HCl of pH 8.4, 50 mM KCl, 1.5 mM MgCl2, 0.1 mg/ml bovine serum albumin). The mixture was incubated Establishment of cell lines at 70°C for 10 min; 200 U of Moloney murine leukemia virus reverse transcriptase and 1 ␮lof10mMdNTP mix were added Peripheral blood mononuclear cells obtained from the patient to the reaction and incubated at 42°C for 50 min. The reaction were cultured as described in Materials and methods. The was stopped by heating to 90°C for 5 min and then quickly cells were maintained without any feeder cells or growth fac- chilled on ice. After brief centrifugation, 2 U of RNase H were tors. The first subculture was made from each flask at 4 weeks added to the reaction mixture and incubated for 20 min at after the cultures were started. The cells proliferated consist- 37°C. Two microliters of the first strand cDNA (corresponding ently as free-floating single cells in suspension. When estab- to 0.5 ␮g of RNA) were diluted with PCR buffer (10 × 500 mM lished, the cell lines were designated MOLM-13 and MOLM-

KCl, 15 mM MgCl2, 100 mM Tris-HCl pH 8.8, 0.8% Nonidet 14, respectively. Both cell lines have since been maintained P40) containing 20 pmol of each upstream and downstream in RPMI 1640 supplemented with 10% FCS, and each has a primer, 10 nmol of dNTP mix and 1.25 U of Taq DNA poly- merase (MBI Fermentas, St Leon-Rot, Germany) for the PCR reaction. To detect the MLL-AF9 fusion product the primers Table 2 Modulation of CD antigens by cytokines used were sense F-MLL (5′-CTGAATCCAAACAGGCCAC- ′ ′ Constitutive IFN-␥ IFN-␥ + M-CSF GM-CSF CACTC-3 ) and antisense (R-AF9 (5 -TCACCA- TNF-␣ TTCTTTATTTGCTTATCAGA-3′). For the reciprocal product AF9-MLL, the primers sense F-AF9 (5′-TCTCATGTCAA- ′ ′ MOLM-13 GATGGGAAAGGT-3 ) and antisense R-MLL (5 - CD13 2.6 5.8 21.0 2.8 3.5 TGCAGGGTGCCGCTCAGTACAG-3′) were employed. The CD14 5.9 19.3 34.3 7.0 9.2 sense and antisense primers for each gene were also used for CD15 69.5 56.1 78.1 62.2 56.6 the detection of the normal alleles. To determine the quality CD64 1.1 8.1 32.5 0.9 0.9 of the RNA, the RT reaction and the PCR amplification, the CD65 65.6 71.4 85.3 60.4 62.2 CD87 95.1 93.5 98.5 95.5 96.2 following primers were used to amplify ␤-actin: sense F- ACTIN (5′-ATGGATGATGATATCGCCGCG-3′) and antisense ′ ′ MOLM-14 R-ACTIN (5 -CTAGAAGCATTTGCGGTGGAC-3 ). The PCR CD13 37.7 60.9 94.5 32.6 36.8 reactions were performed with a DNA thermal cycler (Perkin CD14 25.2 60.1 95.0 18.2 22.3 Elmer Cetus, Heidelberg, Germany) under the following con- CD15 43.4 43.8 88.3 52.5 39.6 ditions: denaturation for 7 min at 95°C; 3 min at 75°C and CD64 90.5 96.0 98.9 84.1 86.8 addition of the Taq polymerase to perform a hot start PCR; 2 CD65 69.6 89.5 98.1 69.8 83.2 CD87 82.6 94.1 99.1 65.8 82.1 min at 60°C and 10 min at 72°C for one cycle. The amplifi- cation was carried out in 35 cycles of 30 s at 94°C, 30 s at ° ° The values are representative of triplicate analyses. The percent- 58 C and 60 s at 72 C with 2 s of extension time for each ages of immunofluorescence positive cells as analyzed by flow cycle. Nine microliters of the reaction mix were electrophor- cytometry are shown. The cytokines were used at the following con- esed on an ethidium bromide-stained 1.4% agarose gel and centrations: IFN-␥ (1000 IU/ml), TNF-␣ (3 JRU/ml), M-CSF (50 U/ml) observed under UV-light. and GM-CSF (20 ng/ml). New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1472 Table 3 Synopsis of data on new monocytic leukemia cell lines MOLM-13 and MOLM-14

Parameter MOLM-13 MOLM-14

Clinical data Age/Sex 20-year-old male Diagnosis MDS (RAEB) – AML-M5a Treatment status at relapse, after chemotherapy Specimen peripheral blood Year of establishment 1995 Cell lines relationship sister cell lines from same specimen Availability from first author Authentication cytogenetic authentication Cell culture Growth medium both RPMI 1640 + 10% FCS (growth factor-independent) Growth pattern both single cells in suspension Continuous culture both Ͼ18 months Doubling time both 3–4 days Maximum cell density both at c. 2.6 × 106 cells/ml Optimal split ratio both at 1:1 every 2–3 days Cryopreservation both in 80% medium, 10% DMSO, 10% FCS Morphology/Cytochemistry both monoblastoid, mononucleated, round/lobulated Morphology nucleus, prominent nucleoli, single basophilic cytoplasm Cytochemistry MSE (+), POX (−) MSE (+), POX (+) Immunoprofile T/NK cell markera negative negative markera negative negative Erythroid–megakaryocytic markera negative negative Myelomonocytic marker CD4+ CD13− CD14− CD4+ CD13+ CD14+ CD15+ CD32+ CD33+ CD15+ CD32+ CD33+ CD64− CD65− CD68− CD64+ CD65+ CD68− CD87+ CD91− CD92+ CD87+ CD91− CD92+ CD93+ CD93+ Progenitor/activation marker CD34− HLA-DR− TdT− CD34− HLA-DR− TdT− Adhesion marker CD11a+ CD11a+ Cytokines Receptor expression GM-CSFR␣+ M-CSFR− GM-CSFR␣+ M-CSFR+ IFN␥R+ IFN␥R+ TNFR-A− TNFR-B− TNFR-A− TNFR-B+ Cytogenetics Classical 49,Ͻ2nϾ,XY,+6,+8,+13, 49,Ͻ2nϾ, XY, +6,+8, ins(11;9)(q23;p22p23), ins(11;9)(q23;p22p23), del(14)(q23.3;q31.3) del(14)(q13.2q31.3) Molecular fusion gene MLL-AF9 fusion gene MLL-AF9 Induced differentiation IFN-␥ both to macrophage-like cells; upregulation of myelomonocytic markers IFN-␥ + TNF-␣ idem, but enhanced M-CSF, GM-CSF in both no effect

aThe following markers were additionally tested (data not shown). T/NK cell marker: CD1, CD2, CD3, CD5, CD7, CD8, CD57; B cell markers, CD10, CD19, CD20, CD21, CD22, CD23, CD40; erythroid–megakaryocytic markers: CD41a, CD42b, CD61, CD62p.

doubling time of 3–4 days. Both cell lines were grown con- The cells had round or lobulated nuclei with fine chromatin tinuously for 18 months. Maximum cell density is at 2.6 × 106 and prominent nucleoli and slightly basophilic cytoplasm cell/ml. Optimal split ratio is 1:1 at 0.5 × 106 cell/ml with fresh (Figure 1a and b). MOLM-13 cells were negative, whereas medium for every 2–3 days. The cells can be cryopreserved about 5% of MOLM-14 cells were positive for POX staining in defined medium (80% medium, 10% FCS, 10% DMSO), (Figure 2). stored in liquid nitrogen, thawed again (with viabilities of more than 70%) and successfully reconstituted. The data presented here were obtained from cells that had been in Immunophenotypic marker profiles culture for 14 months since establishment. The immunophenotypic marker profiles of the two cell lines and the patient’s original leukemia cells are summarized in Morphological and cytochemical characteristics Table 1. The cell lines expressed myelomonocyte-associated antigens including CD13, CD14, CD15, CD32, CD33, CD64, In Wright–Giemsa-stained preparations, MOLM-13 and CD65, CD87, CD92 and CD93. In addition, CD116 for GM- MOLM-14 cells exhibited similar monoblastoid morphologies. CSF␣ receptor and CD119 for the IFN-␥ receptor were New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1473 CD87, CD92, CD93, CD115, CD116, CD119 and CD155. Quantitative or qualitative differences in the expression of antigens between the two cell lines were seen for CD4, CD9, CD13, CD14, CD64, CD65, CD87, CD92, CD115, CD120B and CD115. The patient’s fresh leukemia cells displayed expression of CD34 and HLA-DR at percentages of 30 and 70%, respectively; however, those antigens were not detect- able on either of the established cells lines.

Expression of monocyte-specific esterase isoenzyme

Both MOLM-13 and MOLM-14 strongly express a band in the isoelectric focusing analysis that is specific for (Figure 3).

Chromosome analysis

The consensus GTG karyotypes of subclones present MOLM- 13 (that of the stemline is shown in Figure 4a) were: 49,Ͼ2nϽ,XY, +6, +8,ins(11;9)(q23.3;p22p23), +13, del(14) (q23.3q31.3)[27]/idem 48, −6 [11]/idem 51,+8,+19 in [10]. An additional X chromosome was observed in a single cell of the stemline. The frequency of near-tetraploid cells was 4%. Subclones of MOLM-14 exhibited the closely related consensus karyotypes: 49,Ͻ2nϾ,XY,+6,+8, ins(11;9) (q23.3;p22p23) [17]/idem der(2)t(1;2)(q11;q31) [4]/idem 50, +14, −del(14), +19 [3]. The frequency of near-tetraploid cells also approximated 4%. The 14q− present in both major sub- clones of MOLM-14 differed from that in MOLM-13; the proximal breakpoint in the former being closer to the centro- mere (14q13.2) vs (14q23.3); while the distal breakpoints Figure 1 Wright–Giemsa-stained preparation of (a) MOLM-13 (14q31.3) appeared the same in both deletions. cells and (b) MOLM-14 cells cultured under constitutive conditions Comparison of rearranged with normal homologues of both (original magnification × 312). chromosome pairs 9 and 11 (Figure 4b) indicated the deletion strongly expressed. CD115 for the M-CSF receptor was not detectable for MOLM-13 and was only 10% for MOLM-14. However, differences in antigen expression which are highly associated with myelomonocytic lineages were noted between the two cell lines: MOLM-13 was positive for CD4, CD11a, CD15, CD32, CD33, CD87, CD92, CD93, CD116 and CD119, whereas MOLM-14 was positive for CD4, CD9, CD11a, CD13, CD14, CD15, CD32, CD33, CD64, CD65,

Figure 3 extracted from various cell lines were separated by isoelectric focusing on a polyacrylamide gel and stained with an esterase staining. The monocyte-specific esterase (MSE) band is indi- cated by an arrow. Note that both MOLM-13 and MOLM-14 strongly express the MSE band, whereas six other cell lines do not display this enzyme marker. Derivation of cell lines: OCI-AML-10 from AML-M2; Figure 2 Myeloperoxidase staining of MOLM-14 cells. Arrows OCI-AML-11 from AML-M2; UT-7 from AML-M7; and SAM-1 from indicate positive cells (original magnification × 312). CML in myeloid blast crisis. New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1474

Figure 4 Karyotype of MOLM-13. (a) Depicts cell of the predominant clone (stemline): 49,Ͻ2nϾ,XY,+6,+8,ins (11;9)(q23.3;p21.3p22/p22p23.2), +13, del(14)(q23.3q31.3). The karyotype of MOLM-14 was identical excepting a more extensive 14q− which was del(14)(q13.2q31.3). Trisomies for chromosomes 6, 8, 13 and 19 (the last present in sublines of both sister cell lines) are the four commonest numerical aberrations secondary to t(9;11). (b) Partial karyotype shows normal and rearranged chromosomes 9 and 11. Lightly staining (G- negative) material comprising band p22 deleted from der(9) has been inserted at band q23 in the der(11). This rearrangement was not apparent in suboptimal preparations.

of lightly-staining (G-negative) material comprising band Hybridization with the MLL genomic probe revealed two 9p22, possibly together with a little adjacent G-positive signals of equal intensity contraindicating rearrangement or material (ie, 9p21.3 or 9p23.1), and its insertion at 11q23.3. deletion of this locus on 11q (Figure 5b). No signals were Although the der(9) was quite apparent in most cells, in detectable on the short arm region of the der(9). Thus MOLM- suboptimal preparations the der(11) bearing the insertion was 13 exhibited a hybridization pattern with MLL indistinguish- usually indistinguishable from its normal homologue. able from that of normal cells, whereas t(9;11) is normally Simultaneous hybridization of contrastingly labeled paint- accompanied by a split signal when hybridized with this ing probes for chromosomes 9 and 11 (Figure 5a) confirmed probe. the insertion of a small amount of chromosome 9 material in Together, the cytogenetic findings indicated that MOLM-13 distal 11q without the reciprocal transfer of chromosome 11 and MOLM-14 both comprised a number of closely related material, unlike controls positive for typical t(9;11)(p22;q23), clones, with the putative progenitor karyotype, eg, Mono Mac 6 (data not shown). 49,Ͻ2nϾ,XY,+6,+8, ins(11;9)(q23;p22p23), +13. New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1475 whereas the normal AF9 mRNA shows only weak expression in the RT-PCR assay (Figure 6). The expression pattern is simi- lar to that of the Mono Mac 6 cell line which expresses the (9;11) fusion transcript but not the normal AF9 mRNA.

Response to cytokines

When the MOLM-13 cells were cultured with IFN-␥, or IFN- ␥ and TNF-␣, differentiation to macrophage-like cells was induced as judged by cellular morphology (Figure 7a) and by the induction and upregulation of CD antigen expression (Table 2). IFN-␥ upregulated myelomonocyte-associated CD antigen expression of CD13, CD14 and CD64, and the combi- nation of IFN-␥ and TNF-␣ further enhanced this effect. How- ever, M-CSF (50 U/ml) and GM-CSF (20 ng/ml) did not pro- duce any changes in the expression of those antigens. The cellular morphology did not change upon stimulation with M- CSF or GM-CSF. Changes in the expression of CD antigens by IFN-␥ or the combination of IFN-␥ and TNF-␣ on MOLM-14 were also determined. The myelomonocyte-associated CD antigens tested, including CD13, CD14, CD15, CD64, CD65 and CD87, were all upregulated. However, with M-CSF and GM-CSF, as shown in MOLM-13, no significant changes were seen. The cellular differentiation of MOLM-14 stimulated with IFN-␥ or the combination of IFN-␥ and TNF-␣ showed a more pronounced differentiation towards macrophage-like cells than that of MOLM-13, as judged by cellular morphology. In particular, MOLM-14 cells stimulated with the combination of IFN-␥ and TNF-␣ were larger in diameter, and had irregularly shaped nuclei with a rough chromatin network and vacuo- lated cytoplasm. Degenerated small round cells presenting as apoptotic bodies were also seen (Figure 7b).

Figure 5 FISH analysis of ins(11;9) in MOLM-13. (a) Depicts chro- mosome painting with probes for chromosomes 9 (Spectrum Orange) and 11 (FITC). The arrow indicates insertion of material from chromo- some 9 into distal 11q. To intensify signal-strength of insertion, post- hybridization washes were performed at reduced stringency resulting in increased background labeling of centromeric DNA. (b) Depicts hybridization with an MLL probe labeled with digoxigenin, detected using an anti-digoxin mAb labeled with FITC. Only two signals were present (arrows) mimicking result obtained with normal cells. Absence of signal-splitting thus distinguishes ins(11;9) from t(9;11), although both rearrangements affect MLL-AF9 fusion. FISH preparations in both (a) and (b) were counterstained with DAPI and viewed through a three-color bandpass filter.

Figure 6 RNA derived from the cell line MOLM-13 was used in RT-PCR an RT-PCR reaction with MLL− and AF9− and actin-specific primers as indicated above the lanes. Aliquots of the amplification products were separated by agarose gel electrophoresis and visualized by ethid- The RT-PCR reactions show that MOLM-13 expresses the ium bromide staining. Lane 1: water control; lane 2: actin control; MLL-AF9 fusion mRNA, but not the reciprocal fusion mRNA. lane 3: MLL-AF9 fusion product; lane 4: normal AF9 product (503 bp); The normal MLL mRNA is expressed in normal quantities, lane 5: size marker (pBR-322, cut with Hinf1). New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1476 like cells accompanied by an upregulation of the intensity and percentage of several myelomonocyte-associated antigens including CD14, CD15, CD64, CD65 and CD87. The expression of receptors for IFN-␥ and TNF-␣, both type A and type B, was analyzed using specific mAbs. The receptor for IFN-␥ was positive in both cell lines, however, receptor type A for TNF-␣ was negative for both cell lines; receptor type B was detected in only 10% of MOLM-14 cells. Differentiation induced by the combination of IFN-␥ and TNF-␣ can be explained by undetectable expression due to the limited num- ber of receptor molecules. The apoptotic bodies which were seen after IFN-␥ and TNF-␣ treatment might be caused by the direct effect of IFN-␥ or TNF-␣. Lee and Esteban12 showed that transfection and expression of the IFN-induced double- stranded RNA-activated kinase, a serine/threonine kin- ase, in a human tumor cell line resulted in rapid apoptotic cell death.12,13 This mechanism could be involved partially in the apoptotic cell death caused by IFN-␥ and TNF-␣ in the present study. We have shown that both sister cell lines, MOLM-13 and MOLM-14, are closely related karyotypically and are likely to represent mixtures of clones derived from a common progeni- tor carrying ins(11;9) as well as trisomies for chromosomes 6, 8 and 13. The presence of MLL-AF9 chimeric mRNA together with the common involvement of 9p22/23 and 11q23 break- points in the insertion shows that the ins(11;9) rearrangement mimics t(9;11), both structurally and functionally. Interest- ingly, all trisomies observed in the various subclones of both MOLM-13 and MOLM-14 ie, +6, +8, +13, +19, together with +X, seen in one (MOLM-13) cell, comprise all the commonest numerical abnormalities associated with t(9;11) in AML.14 The consistency of this association, together with their presence in both sister cell lines, indicates an in vivo origin for all numeri- Figure 7 Wright–Giemsa-stained preparation of (a) MOLM-13 cal changes. cells and (b) MOLM-14 cells cultured for 3 days with a combination The MLL gene has been found to be involved in several of IFN-␥ (1000 IU/ml) and TNF-␣ (3 JRU/ml) (original magnification × 312). translocations affecting 11q23 including t(4;11), t(6;11), t(11;19) as well as t(9;11) reported to occur in 20–30% of childhood and adult acute monocytic .15,16 AF9, Discussion located at chromosome 9p22, has been cloned and demon- strated to be a fusion partner of the MLL gene located at chro- Two acute monocytic leukemia cell lines carrying the mosome band 11q23. The MLL/AF-9 fused gene produces a ins(11;9)(q23;p22p23) chromosomal abnormality, MOLM-13 chimeric fusion protein and appears to play a crucial role in being CD4+, CD13−, CD14−, CD15+, CD33+, and MOLM-14 leukemogenesis.17,18 The cytogenetic interpretational differ- being CD4+, CD13+, CD14+, CD15+, CD33+, were established ences in the present case imply significant under-ascertain- from the peripheral blood of a patient with AML-M5a. In ment of MLL rearrangement via small deletions of which only addition to the above antigens, expression of others including two instances have been described.3,19 However, as the CD9, CD64, CD65, CD87, CD92 and CD155 was also found present report documents, improved ascertainment of MLL to be different between the two cell lines. The differences seen rearrangement due to minute insertions may not be guaran- in the immunoprofiles of the two cell lines, however, suggest teed by molecular cytogenetic detection. interclonal phenotypic heterogeneity despite the fact that they The donor patient is reportedly the first case of MLL-AF9 were derived from the same specimen, which is supported by arising de novo in AML evolving from MDS4 implying distinct the fact that they share the common cytogenetic abnormality, primary roles for +8 and MLL-AF9 in the respective ontogenies t(11;9)(q23;p22p23). Most cells of both cell lines had a blastic of MDS and AML phases. The presence of both rearrange- appearance, but a small percentage of the cells showed ments in the cell lines, together with additional trisomies for macrophage-like features with myeloperoxidase positivity. chromosomes 6, 13, and 19, suggests a direct line of leukemo- Both cell lines were found to be positive for MSE, clear-cut genic descent marked by cumulative chromosome change at evidence that they were both of monocytic origin.7,8,10 The successive stages from MDS (+8) through onset of acute leuke- cause of this heterogeneity remains unknown, thus the follow- mia (evolution of MLL-AF9) and onward to relapsing AML ing possibilities are proposed: (1) the two cell lines may rep- (additional trisomies). resent distinct stages of differentiation; or (2) the cell lines may Reports of monocytic leukemia cell lines are still limited. be of oligo- or biclonal origin. U-937,20 THP-1,21 DOP-M1,22 Mono Mac 6,23–25 IMS-M1,26 The cytokines IFN-␥ and TNF-␣ can activate monocytes and P31/FUJIOKA,27 YK-M228 and KP-MO-TS29 are some of the appear to be promising therapeutic agents against tumor cell lines described so far. Among these cell lines, THP-1, cells.11 Stimulation of MOLM-13 and MOLM-14 with IFN-␥ Mono Mac 6, and IMS-M1 are known to have the t(9;11) and TNF-␣ resulted in cellular differentiation to macrophage- chromosomal abnormality.6 It is of note that the immunopro- New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1477 files of these cell lines are quite heterogeneous. In the present 11 Watanabe N, Niitsu Y, Yamauchi N, Umeno H, Sone H, Neda H, Urushizaki I. Antitumor synergism between recombinant human study, despite the fact that MOLM-13 and MOLM-14 orig- ␥ inated from the same blood specimen (from a patient with tumor necrosis factor and recombinant human interferon- . J Biol Response Mod 1988; 7: 24–31. AML-M5a) and were established at the same time, they 12 Lee SB, Esteban M. The interferon-induced double-stranded RNA- showed distinct antigen marker profiles. Although both cell activated protein kinase induces apoptosis. Virology 1994; 199: lines carried the same chromosomal abnormality, they 491–496. revealed quantitative and/or qualitative differences in several 13 Meurs EF, Galabru J, Barber GN, Katze MG, Hovanessian AG. myelomonocyte-associated antigens including CD13, CD14, Tumor suppressor function of the interferon-induced double- CD64, CD65, CD87, CD92 and CD115. However, they were stranded RNA-activated protein kinase. Proc Natl Acad Sci USA equally positive for CD4, CD11a, CD15, CD33, CD93 and 1993; 90: 232–236. 14 Johannsson B, Mertens F, Mitelman F. Secondary chromosomal CD116. HLA-DR was negative in both cell lines. CD14 is abnormalities in acute leukemias. Leukemia 1994; 8: 953–962. regarded as an immunological marker of relatively mature + 15 Brodeur GM, Williams DL, Kalwinsky DK, Williams KJ, Dahl GV. monocytes.30 While MOLM-14 was constitutively CD14 , Cytogenetic features of acute nonlymphoblastic leukemia in 73 MOLM-13 did not express this antigen under steady-state cul- children and adolescents. Cancer Genet Cytogenet 1983; 8: 93– ture conditions; however, MOLM-13 also became CD14+ after 105. treatment with IFN-␥. 16 Kaneko Y, Rowley JD, Maurer HS, Variokojis D, Moohr JW. Chro- To our knowledge, this report is the first to document two mosome pattern in childhood acute nonlymphocytic leukemia (ANLL). Blood 1982; 60: 389–399. cell lines with interclonal phenotypic heterogeneity of this 17 Rubnitz JE, Behm FG, Downing JR. 11q23 rearrangements in acute type. MOLM-13 and MOLM-14 are unique leukemia cell lines leukemia. Leukemia 1996; 10: 74–82. derived from AML-M5a which can be useful in studying the 18 Corral J, Forster A, Thompson S, Lampert F, Kaneko Y, Slater R, pathogenesis of AML-M5a, and the novel cytogenetic abnor- Kroes WG, van der Schoot CE, Ludwig WD, Karpas A, Pocock C, mality, ins(11;9)(q23;p22p23). The presence of unpre- Cotter F, Rabbitts TH. Acute leukemias of different lineages have cedentedly high degrees of subclonal karyotypic diversity of similar MLL gene fusions encoding related chimeric proteins apparent in vivo origin in both cell lines may be attributable resulting from chromosomal translocation. Proc Natl Acad Sci USA 1993; 90: 8538–8542. to their relative ‘youth’ compared to most currently available 19 Kobayashi H, Espinosa R III, Thirman MJ, Gill HJ, Fernald AA, leukemia cell lines which appear to have undergone signifi- Diaz MO, Le Beau MM, Rowley JD. Heterogeneity of breakpoints cant karyotypic evolution in vitro. of 11q23 rearrangements in hematologic malignancies identified A synopsis of the data on MOLM-13 and MOLM-14 is with fluorescence-in-situ-hybridization. Blood 1993; 82: 547– shown in Table 3. 551. 20 Sundstrom C, Nilsson K, Establishment and characterization of a human histiocytic lymphoma cell line (U-937). Int J Cancer 1976; Acknowledgements 17: 565–577. 21 Tsuchiya S, Yamabe M, Yamaguchi Y, Kobayashi Y, Konno T, We thank Dr J Minowada for critical discussion and Ms A Tada K. Establishment and characterization of a human acute Okochi for her excellent technical assistance. monocytic leukemia cell line (THP-1). Int J Cancer 1980; 26: 171–176. 22 Sugita K, Sakakibara HJ, Eguchi M, Hayashi Y, Furukawa T. Estab- References lishment of a monocytoid cell line (DOP-M1) from an infant with acute monocytic leukemia. Int J Hematol 1992; 55: 45–51. 1 Hagemeijer A, Haehlen K, Sizoo W, Abels J. Translocation 23 MacLeod RAF, Voges M, Drexler HG. MONO MAC 6: a mature (9;11)(p21;q23) in three cases of acute monoblastic leukemia. monoblastic leukemia cell line with t(9;11)(p21;q23). Blood 1993; Cancer Genet Cytogenet 1982; 5: 95–105. 82: 3221–3222. 2 Mitelman F. Catalogue of Chromosome Aberrations in Cancer, 5th 24 Quentmeier H, Kolsdorf K, Zaborski M, Drexler HG. IL-4R ␣ and edn. Wiley-Liss: New York, 1994. ␥ chain expression in LPS- and IL-4-stimulated MONO-MAC-6 3 Poirel H, Rack K, Delabesse E, Radford-Weiss I, Troussard X, cells. Leukemia 1995; 9: 336–340. Debert C, Lebef D, Bastard C, Picard F, Veil-Buzyn A, Flandrin G, 25 Quentmeier H, Kolsdorf K, Zaborski M, Drexler H. IL-4 inhibits Bernard O, Macintyre E. Incidence and characterization of MLL the LPS-induced expression of CD14 and monocyte-specific ester- gene (11q23) rearrangements in M1 and ase mRNA in MONO-MAC-6 cells. Leukemia 1994; 8: 1551– M5. Blood 1996; 87: 2496–2505. 1556. 4 Katayama Y, Takimoto H, Fujii N, Kimura G, Omoto E, Harada M. 26 Iida S, Seto M, Yamamoto K, Komatsu H, Tojo A, Asano S, Kam- Clonal evolution to acute myeloblastic leukemia with MLL gene ada N, Ariyoshi T, Takahashi T, Ueda R. MLLT3 gene on 9p22 rearrangement from hypocellular myelodysplastic syndrome with involved in t(9;11) leukemia encodes a serine/proline rich protein trisomy 8. Leukemia 1997; 11: 1380–1382. homologous to MLLT1 on 19p13. Oncogene 1993; 8: 3085–3092. 5 Drexler HG, Gignac SM, Minowada J. Hematopoietic cell lines. 27 Hirose R, Minato K, Tobinai K, Shimoyama M, Watanabe S, Abe In: Hay RJ, Park JG, Gazdar A (eds). Atlas of Human Tumor Cell T. A novel monocytoid cultured cell line, P31/FUJIOKA, derived Lines. Academic Press: San Diego, 1994, pp. 213–250. from acute monoblastic leukemia. Gann 1982; 73: 735–741. 6 Drexler HG, MacLeod RAF, Borkhardt A, Janssen JW. Recurrent chromosomal translocations and fusion genes in leukemia– 28 Ohno H, Fukuhara S, Doi S, Yamasowa M, Arita Y, Sasada M, lymphoma cell lines. Leukemia 1995; 9: 480–500. Ohno Y, Akasaka K, Uchino H. Establishment of a novel acute 7 Gignac SM, Drexler HG. Monocyte-specific esterase isoenzyme monoblastic leukemia cell line (YK-M2) having a near triploid kar- demonstrated by isoelectric focusing. Electrophoresis 1990; 11: yotype. Cancer Res 1986; 46: 6400–6405. 819–824. 29 Ikushima S, Yoshihara T, Fujiwara F, Todo S, Misawa S, Nakagawa 8 Uphoff CC, Hu ZB, Gignac SM, Ma W, Rainey FA, Kreutz M, H, Yumura K, Ha-Kawa K, Morioka Y, Imashuku S. Establishment Ludwig WD, Drexler HG. Characterization of the monocyte- of a monocytoid cell line (KP-MO-TS) from a patient with acute specific esterase (MSE) gene. Leukemia 1994; 8: 1510–1526. monocytic leukemia and a mediastinal mass. Acta Hematol Jap 9 Seabright M. A rapid banding technique for chromosomes. Lancet 1990; 53: 678–687. 1971; 2: 971–972. 30 Abrink M, Gobl AE, Huang R, Nilsson K, Hellman L. Human cell 10 Uphoff CC, Gignac SM, Metge K, Zschunke F, Radzun HJ, Drexler lines U-937, THP-1 and Mono Mac 6 represent relatively imma- HG. Expression of the monocyte-specific esterase gene in ture cells of the monocyte–macrophage cell lineage. Leukemia leukemia–lymphoma cell lines. Leukemia 1993; 7: 58–62. 1994; 8: 1579–1584. New AML-M5a cell lines with ins(11;9)(q23;p22p23) Y Matsuo et al 1478