Leukemia (2000) 14, 576–585 2000 Macmillan Publishers Ltd All rights reserved 0887-6924/00 $15.00

www.nature.com/leu

Diphtheria toxin fused to human interleukin-3 is toxic to blasts from patients withmyeloid leukemiasAE Frankel1, JA McCubrey2, MS Miller1, S Delatte3, J Ramage1, M Kiser1, GL Kucera1, RL Alexander1, M Beran4, EP Tagge3,RJ Kreitman5 and DE Hogge6

1Department of Cancer Biology, Wake Forest University School of Medicine, Winston-Salem, NC; 2Department of Microbiology, EastCarolina University, Greenville, NC; 3Departments of Surgery, Medical University of South Carolina, Charleston, SC; 4Department ofMedicine, MD Anderson Cancer Center, Houston, TX; 5Laboratory of Molecular Biology, National Cancer Institute, NIH, Bethesda, MD;and 6Terry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada

Leukemic blasts from patients with acute phase chronicmyeloid leukemic and refractory acute myeloid leukemia arehighly resistant to a number of cytotoxic drugs. To overcomemulti-drug resistance, we engineered a diphtheria fusion pro-tein by fusing human interleukin-3 (IL3) to a truncated form ofdiphtheria toxin (DT) with a (G 4S)2 linker (L), expressed and pur-ified the recombinant protein, and tested the cytotoxicity of theDTLIL3 molecule on human leukemias and normal progenitors.The DTLIL3 construct was more cytotoxic to interleukin-3receptor (IL3R) bearing human myeloid leukemia cell lines thanreceptor-negative cell lines based on assays of cytotoxicityusing thymidine incorporation, growth in semi-solid mediumand induction of apoptosis. Exposure of mononuclear cells to680 pM DTLIL3 for 48 h in culture reduced the number of cellscapable of forming colonies in semi-solid medium (colony-for-ming units leukemia) >10-fold in 4/11 (36%) patients withmyeloid acute phase chronic myeloid leukemia (CML) and 3/9(33%) patients with acute myeloid leukemia (AML). Normalmyeloid progenitors (colony-forming unit granulocyte–macrophage) from five different donors treated and assayedunder identical conditions showed intermediate sensitivity withthree- to five-fold reductions in colonies. The sensitivity toDTLIL3 of leukemic progenitors from a number of acute phaseCML patients suggests that this agent could have therapeuticpotential for some patients with this disease. Leukemia (2000)14, 576–585.Keywords: diphtheria fusion toxin; interleukin-3; acute myeloid leu-kemia; chronic myeloid leukemia

Introduction

Both the natural history of acute phase CML and relapsed orrefractory AML is characterized by short survival (median ofaround 6 months) and resistance to cytotoxic chemo-therapy.1,2 Chemotherapy-resistant blasts are a frequent causeof treatment failure in these patients.3 In many cases, the blastsexhibit a multi-drug resistance phenotype. New reagents withunique mechanisms of action that can selectively modify theapoptotic threshold of malignant blasts are needed.

One such class of leukemia therapeutics is targeted toxinmolecules. These drugs consist of myeloid leukemia-directedligands covalently linked to protein synthesis inactivating pep-tide toxins. We previously selected human granulocyte–macrophage colony-stimulating factor (GM-CSF) as the ligandand fused the peptide to a truncated form of diphtheria toxin(DT) containing amino acids 1–388 of DT to produce a fusionprotein, DTGM.4 DTGM reduced leukemic progenitor colonyformation >1 log in 85% of AML patients, 71% of juvenilemyelomonocytic leukemia (JMML) patients and 60% of adult

Correspondence: AE Frankel, Hanes 4046, Wake Forest UniversitySchool of Medicine, Med Center Blvd, Winston-Salem, NC 27157,USA; Fax: 336 716 0255Received 18 November 1999; accepted 16 December 1999

chronic myelomonocytic leukemia (CMML) patients.5,6 Wehave scaled up production of DTGM under good manufactur-ing practice (GMP)7 and are currently performing a phase Idose-escalation clinical study for patients with relapsed orrefractory AML. However, because DTGM was ineffectiveagainst acute phase CML patient blasts and 15% of AMLpatient blasts, we sought an additional ligand for targeting.

Interleukin-3 (IL3) is a cytokine which supports the prolifer-ation and terminal differentiation of multipotential and com-mitted myeloid and lymphoid progenitors,8 but does not acton the most primitive hematopoietic stem cells.9 The IL3receptor (IL3R) has been demonstrated by 125I-labeled IL3binding to be present on human myeloid leukemic cells.10

Further, acute phase CML blasts proliferate in response toexogenous IL3.11,12 In fact, CML progenitors have beenreported to overexpress both IL3 and IL3R, yielding anautocrine loop stimulation of proliferation.13 This also occursin certain autocrine transformed mouse leukemia cell lines.14

The IL3R is composed of a and b subunits.15 The binding ofIL3 to its receptor causes rapid internalization of the ligand–receptor complex.16 Thus, based on the expression of itsreceptor on blasts from patients with acute phase CML andAML and its rapid receptor-mediated endocytosis on receptorcoupling, we chose to fuse IL3 to the catalytic and trans-location domains of diphtheria toxin (DT).

The goal of the present study was to synthesize an IL3-diph-theria toxin fusion molecule, purify and characterize the mol-ecule chemically, and evaluate the toxicity of the fusion pro-tein towards leukemia cell lines and progenitors from normalindividuals as well as patients with either acute phase CMLor AML.

Materials and methods

Plasmid construction

Primers and the polymerase chain reaction (PCR) were usedto introduce NdeI and HindIII restriction sites at the ends ofthe IL3 gene so that the target IL3 could then be directly lig-ated into a previously constructed expression vectorpRKDTGM encoding both diphtheria toxin (DT) and GM-CSF(GM). The GM was released from the expression vector byenzymatic digestion with NdeI plus HindIII and the newNdeI/HindIII IL3 fragment was ligated into the construct in itsplace. The IL3 DNA insert with NdeI and HindIII restrictionsites at the ends was synthesized by PCR from a pUC18 plas-mid containing mature human IL-3 cDNA (Catalog No.BBG14, Research Diagnostics, Minneapolis, MN, USA). AnIL3 DNA insert with a (Gly4Ser)2 linker was prepared. Ampli-fication was performed using the Perkin Elmer GeneAmp PCRReagent Kit (Perkin Elmer Cetus, Norwalk, CT, USA). The

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

577primer for the 59 end was 59-GCAGTCGACCATATGG-GCGGAGGCGGAAGTGGAGGAGGAGGCAGCGCTCC-CATGACCCAGACA-39. The primer for the 39 end was 59-GCAGTCGCAAAGC-TTCTAAAAGATCGCTAGCGACAA-39.The IL3 DNA was amplified according to the manufacturer’sinstructions in 10 mM Tris-HCl (pH 8.3)/2.5 mM MgCl2/50 mM

KCl reaction mixture containing 200 mM of each of the fournucleotide triphosphates, 1 mg/ml of each primer, and 3 unitsof TaqGold DNA polymerase (Perkin-Elmer Cetus). After aninitial denaturation step for 5 min at 94°C, the 40 ng ofpUC18-IL3 was amplified by 40 cycles of denaturation at94°C for 1 min, annealing at 50°C for 1 min, and extensionat 72°C for 1 min followed by an extended extension at 74°Cfor 7 min.

PCR product was then subcloned in the pCR-TOPO vectorfollowing the instructions of the manufacturer (TOPO TACloning, Version D, InVitrogen, Carlsbad, CA, USA). PlasmidDNA was isolated from stationary culture using the WizardPlus SV Miniprep DNA Purification System (Promega, Madi-son, WI, USA) following the manufacturer’s instructions andsequenced using the M13 universal forward and reverse pri-mers to confirm correct sequence for the IL3 construct. Boththe IL3 vector and the pRKDTGM plasmid17 DNAs weredigested with NdeI and HindIII. The 0.4 kb IL3 DNA fragmentand 3.7 kb pRKDT DNA fragment were isolated from 0.7%low melting point GTG NuSieve agarose gels (FMC Corpor-ation, Rockland, ME, USA) using the QiaQuick PCR Purifi-cation kit (Qiagen, Valencia, CA, USA) following the manu-facturer’s instructions. The vector fragment wasdephosphorylated with shrimp alkaline phosphatase(Boehringer Mannheim, Indianapolis, IN, USA), heat inacti-vated and ligated to the IL3 insert using T4 DNA ligase poly-merase (Promega). A Qiagen Maxi-Prep of pRKDTLIL3(containing a (G4S)2 linker between DT388 and IL3) was pre-pared and the plasmid sequenced with an Applied BiosystemsPrism Dye Terminator Dideoxy cycle-sequencing kit and auto-mated sequencer (Applied Biosystems, Foster City, CA, USA).The plasmid encoded a methionine followed by the first 388amino acid residues of diphtheria toxin followed by the His-Met and a (G4S)2 linker followed by IL3.

Expression and purification of DTLIL3

Competent E. coli BLR (lDE3) (Novagen, Milkwaukee, WI,USA) were transformed with 1.5 ml of pRKDTLIL3 (containingthe (G4S)2 linker) as previously described.7 Colonies from 20100 mm LB ampicillin plates were pooled and added to 3 ofsuperbroth containing 0.8 mM MgSO4 and 0.25% dextroseand shaken in two 6 l Fernbach flasks at 37°C. When the cul-ture reached an OD650 of 0.6 (2.5 h), 1 mM IPTG (Sigma, StLouis, MO, USA) was added. Two hours later, the cells wereharvested by centrifugation, resuspended in 480 ml TES(100 mM Tris pH 8, 1 mM EDTA, 50 mM NaCl) containing125 mg lysozyme, homogenized with a tissuemizer (IKAWorks, Wilmington, NC, USA), incubated 45 min at roomtemperature, and pelleted. The inclusion bodies were washedfour times with TES + 2.5% Triton X-100, and four times withTES alone. The inclusion bodies were then dissolved in 12 mlof 7 M guanidine HCl/100 mM Tris pH 8/1 mM EDTA contain-ing 120 mg dithioerythritol, homogenized and incubated 4 hat room temperature, and pelleted. The supernatant (80 mgtotal protein) was added dropwise to a stirring 4°C 1200 ml0.5 M L-arginine/1 mM EDTA/0.1 M Tris pH 8/664 mg oxidizedglutathione solution and refolding continued at 4°C for 48 h.

Leukemia

Refolded protein was dialyzed at 4°C against 20 mM TrispH7.4/1 mM EDTA, filter-sterilized and the 1450 ml loaded onan AKTA FPLC 5 ml HiTrap Q column in Buffer A (20 mM TrispH 7.4, 1 mM EDTA). Protein was washed with 0.05 partsBuffer B (Buffer A + 1 M NaCl) and 0.95 parts Buffer A, andeluted with 0.13 parts Buffer B and 0.87 parts Buffer A. Theeluant was diluted five-fold with Buffer A and loaded andwashed on a 1 ml HiTrap Q column. Protein was eluted with0.5 parts Buffer B and 0.5 parts Buffer A and sent to a Super-loop. From the Superloop, the protein was loaded on Hi Prep16/60 Sephacryl S-200 column and eluted in PBS. The mono-mer peak fractions were pooled, filter sterilized and storedat −80°C.

Characterization of DTLIL3

Protein was quantitated by the Coomassie Plus assay as perrecommendations of the supplier (Pierce, Rockford, IL, USA).

Aliquots of DTLIL3 and prestained Rainbow molecularweight standards (Amersham, Chicago, IL, USA), human IL3(Propertech, Rocky Hill, NJ, USA) and DTGM were run ona reducing 10% sodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE) and either stained with Coomas-sie Blue R-250 (Sigma Chemical) or transferred to nitrocellu-lose, blocked with 10% non-fat dry milk, 0.1% sodium azide,0.1% bovine serum albumin (BSA) in phosphate-bufferedsaline (PBS), washed with PBS with 0.05% Tween 20 and PBS,reacted with a 1:200 rat anti-human IL3 antibody(PharMingen, San Diego, CA, USA) or 1:200 mouse anti-diph-theria fragment A, rewashed, incubated with horseradish per-oxidase-conjugated goat anti-rat Ig 1:5000 (Jackson Immuno-Research Lab, West Grove, PA, USA) or horseradish peroxi-dase-conjugated goat anti-mouse Ig 1:2000, washed again,and developed with ECL reagent (Amersham, ArlingtonHeights, IL, USA) and exposed to X-ray film as per manufac-turer’s instructions. Gels and blots were scanned on a HewlettPackard ScanJet 6100C (Hewlett Packard, Palo Alto, CA,USA).

To measure the enzymatic activity of DTLIL3, the ADPribosylation assay modified from Collier and Kandel18 byHwang and colleagues19 was employed.18,19 Varying amountsof transferrin-CRM10720 or DTLIL3 in 480 ml DTEB bufferconsisting of 0.2% BSA, 40 mM DTT, 1 mM EDTA, and 50 mM

Tris pH 8.0 were mixed with 10 ml 14C-NAD (NEN-DuPont,NEC743, 588 Ci/mmol, 10 mCi/ml) and 10 ml wheat germextract (Promega, L418A). Samples were incubated at 37°Cfor 30 min, and protein precipitated for 15 min with 4°C0.5 ml 12% TCA. Samples were centrifuged at 25 000 g in amicrofuge, washed with 1 ml 6% TCA, and dissolved in0.3 ml 0.1 M NaOH. Samples were then added to scintillationvials with 0.2 ml 1 N HCL. Three ml scintillation fluid wasadded (Ecolite, ICN), and samples were counted in a BeckmanLS1800 liquid scintillation counter (Beckman Instruments,Palo Alto, CA, USA). Molarity of diphtheria fragment A con-taining molecules vs acid precipitable c.p.m. was plotted.

Free thiols were determined by the use of Ellman’s reagentand spectrophotometry as described.21

Two-dimensional PAGE with isoelectric focusing was car-ried out using pH 3.5–10, pH 4–6 and pH 5–7 ampholinesby Wallin’s modification of the O’Farrell method.22

HPLC of DTLIL3 on a TSK3000 column with PBS at1 ml/min was performed along with protein standards BSAand ovalbumin. Optical density at 280 nm was monitored.

Reverse phase HPLC using a C4 Vydak column with a 0–

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

578

Leukemia

70% acetonitrile in 0.05% trifluoroacetic acid mobile phasewas performed with DTLIL3. Absorbance at 215 nm wasmeasured.

DTLIL3 (100 mg) was ethanol precipitated and analyzed byEdmond degradation and HPLC. The sequence of the N-ter-minal 14 amino acids was determined.

Tryptic peptide mapping of DTLIL3 was performed by themethod of Matsudaira.23

Tandem mass spectroscopy was performed on a Quattro IItriple quadrupole mass spectrometer running MassLynxsoftware after reverse phase HPLC as described above.

Cell lines

The hematopoietic growth factor-dependent human leukemiacell lines TF1, and the TF1 derivatives TF1/Bcl2,TF1DMEK+LNL6, TF1DMEK+Bcl2 were obtained from Dr JMcCubrey and grown in RPMI 1640 medium with 10% fetalbovine serum (FBS), 1 mM sodium pyruvate (GIBCO-BRL,Grand Island, NY, USA), 0.1 mM non-essential amino acids(GIBCO), 2 mM L-glutamine (GIBCO), 50 U/ml penicillin-G(GIBCO), 50 mg/ml streptomycin sulfate (GIBCO), and50 ng/ml human GM-CSF (Immunex, Seattle, WA, USA). OCI-AML and AML193 cells were obtained from the AmericanType Culture Collection (Rockville, MD, USA). Both cell lineswere grown in Iscove’s modified Dulbecco’s medium with25 mM Hepes and 10% FBS supplemented with 50 ng/ml GM-CSF. M07e cells were obtained from Dr D Hogge and grownin Dulbecco’s MEM with 10% FBS, 10% 5637 conditionedmedium, 50 mM 2-mercaptoethanol, and 5 ng/ml human IL-3.HL60, K562, and CEM human leukemia cells, P815 murinemastocytoma, and TK6 human lymphoblastoid lines wereobtained from the American Type Culture Collection andmaintained on RPMI 1640 medium plus glutamine with 10%FBS and penicillin/streptomycin. The properties of the celllines and their sources are described in Table 1.

Cell line sensitivity to DTLIL3 molecule

For thymidine incorporation inhibition assays, 5 × 104 cellswere incubated in 100 ml RPMI 1640/15% FBS/penicillin/streptomycin/glutamine in Costar 96-well flat-bot-

Table 1 Properties of assayed human cell lines

Cell line Description Growth Ref.factor

dependence

M07e Acute myeloid leukemia + 45TF1 Erythroleukemia + 46TF1/Bcl2 TF1 with Bcl2 transgene + Present

studyHL60 Acute myeloid leukemia − 47CEM T lymphoblastic leukemia − 48AML193 Acute monocytic leukemia + 49K562 Chronic myeloid leukemia − 50P815 Murine mastocytoma − 51TK6 Human lymphoblastoid − 52OCI-AML Human myeloid leukemia + 53TF1/DMEK+LNL6 TF1 with DMEK transgene + 54TF1/DMEK+Bcl2 TF1 with DMEK/Bcl2 + 54

transgenes

tomed plates. Twelve different concentrations of DTLIL3 wereadded to each column in 50 ml media, and the cells main-tained at 37°C/5% CO2 for 48 h. Then, 1 mCi3H-thymidine(NEN DuPont, Boston, MA, USA) in 50 ml medium was addedto each well, and incubation continued for an additional 18 hat 37°C/5% CO2. Cells were then harvested by a Skatron CellHarvestor (Skatron Instruments, Lier, Norway) on to glass fibermats and 3H c.p.m. were counted in an LKB liquid scintil-lation counter gated for 3H. The calculated IC50s were theconcentrations of toxin which inhibited thymidine incorpor-ation by 50% compared to control wells.

To evaluate inhibition of colony formation as another meas-ure of cell kill, 2 × 104 AML cell lines were incubated withdifferent concentrations of DTLIL3 (0–400 nM) in 150 ml ofRPMI 1640 medium plus 15% FBS supplemented with50 ng/ml G-CSF (Amgen, Thousand Oaks, CA, USA) in 96-well flat-bottomed Costar plates. After 48 h, 50 and 100 mlsamples from each well were mixed with 1 ml methylcellulosemedium (Methocult H4434, StemCell Technologies, Van-couver, BC, Canada) containing FBS, stem cell factor, GM-CSF, IL-3 and erythropoietin and poured into 35-mm griddedpetri dishes (Nunc, Naperville, IL, USA). Dishes were placedin humidified chambers at 37°C/5% CO2 for 14 days, afterwhich colonies containing greater than 20 cells were counted.Both the concentrations of toxin reducing colony formationby 50% (IC50) and the maximal log cell kill compared withcontrols were calculated as previously described.4

For assays of apoptosis induction, aliquots of 1 × 106 cellswere incubated in 24-well Costar plates at 37°C/5% CO2 for48 h in media with different concentrations of DTLIL3. Cellswere then washed in PBS and resuspended in 100 ml stainingsolution (containing Annexin-V fluorescein and propidiumiodide in HEPES buffer, Annexin-V-FLOUS Staining Kit;Boerhinger-Mannheim). Following incubation at room tem-perature for 15 min, cells were analyzed by flow cytometry aspreviously described.24 Cells were also examined by phasecontrast microscopy to assess morphologic changes ofapoptosis (nuclear condensation and fragmentation and cyto-solic membrane blebbing). In selected experiments, apoptoticinduction of internucleosomal fragmentation of genomic DNAwas determined. 1 × 107 cells (4 × 105/ml) were grown in theindicated conditions for 1 day. The cells were pelleted by cen-trifugation, resuspended in 500 ml of lysis buffer (20 mM Tris-HCl, pH 7.4, 10 mM EDTA, 0.2% Triton X-100) and placedon ice for 10 min. The lysate was then centrifuged for 10 minat 16 000 g, and the pellet discarded. Twenty ml of 4 mg/mlprotease K (Sigma) was added to the supernatant and the mix-ture was incubated overnight at room temperature. The sol-ution was phenol extracted twice followed by chloroformextraction once. The aqueous phase was saved each time. Thenucleic acids were ethanol precipitated and subsequentlyresuspended in water. The concentration of the nucleic acidswas determined and 40 mg was treated with 10 ml of0.25 mg/ml RNase A (Sigma) for 5–10 min. The remainingDNA was electrophoresed in an agarose gel with 1 × TBE(0.09 Tris borate, 0.002 M EDTA) running buffer and vis-ualized with ethidium bromide.

In other experiments, cells were treated in 24-well Costarplates with 50 nM DTLIL3 for 0, 6, 12, 24, 48 and 60 h per-iods. Cell samples were reacted with 6 mg/ml Hoeschst 33342(Aldrich, Milwaukee, WI, USA) and apoptotic cells quantifiedusing fluorescence microscopy. A total of at least 200 cellswere counted, and the data calculated as percent apoptoticcells.

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

579Acute phase CML, AML and normal marrow cells

With the approval of the respective Institutional ReviewBoards and after obtaining informed patient or parental con-sent, heparinized blood samples were obtained from 20patients with myeloid acute phase CML and 17 patients withAML. Heparinized marrow aspirates were also obtained fromfive allogeneic bone marrow donors. Blood and marrow cellswere diluted 1:1 with RPMI 1640 medium and layered over0.3 vol of Ficoll–Hypaque (1.070 g/ml; Pharmacia, Piscata-way, NJ, USA). After density gradient centrifugation at 1000 gfor 30 min, light density cells (,1.077 g/ml) were dilutedthree-fold with RPMI 1640 and centrifuged again at1300 r.p.m. for 10 min. Cells were then usually cryopreservedin 50% FBS with 10% dimethylsulfoxide. Upon thawing, cellswere suspended in RPMI 1640 medium with 15% FBS and2 mM L-glutamine, 50 U/ml penicillin G, 50 mg/ml strepto-mycin sulfate with 1 mg/ml DNAse I (Sigma Chemicals). TheDNAse reduces cell clumping with DNA released fromnecrotic cells.

Leukemic and normal progenitor cell sensitivity

To evaluate inhibition of colony formation, aliquots of 1 or2 × 105 acute phase CML, AML and normal marrow light den-sity cells were incubated with different concentrations ofDTLIL3 (0–400 nM) in 150 ml of RPMI 1640 medium plus 15%FBS supplemented with 50 ng/ml G-CSF (Amgen) in 96-wellflat-bottomed Costar plates at 37°C/5% CO2 in air. After 48 h,50 and 100 ml aliquots were plated in Methocult as describedabove for cell lines. Colonies were counted at 14–21 days.

Results

We designed an IL3R-directed fusion toxin for potential ther-apy of acute phase CML and AML. Our choice was based onthe presence of IL3R on leukemic progenitors10–13 and on thebiological potency for murine leukemic cells for similar diph-theria fusion molecules with murine IL3.25,26 Because of con-cern that steric hindrance from the diphtheria toxin mightreduce the binding of the IL3 moiety to its receptor as wasreported with murine IL3,26 we synthesized a linker form ofdiphtheria IL3 fusion protein. The construct consisted of thefirst 388 amino acids of diphtheria toxin including both thecatalytic and translocation domains fused via a His-Met-(Gly4

Ser)2-linker to human IL3. The bacterial expression plasmidwas successfully prepared and sequence confirmed and theconstruct is shown in Figure 1.

Preparation of recombinant diphtheria IL3 toxin

Three liter fermentations were done using BR(DE3) E. colitransformed with pRKDTLIL3 DNA. The inclusion body yieldfor DTLIL3 was 79 mg protein. The refolded protein (pI of 6)was dialyzed against a low salt pH 7.4 buffer and bound toan anion exchange matrix (HiTrap Q) and eluted with a highersalt buffer (0.13 M NaCl once and 0.5 M NaCl for the concen-trating chromatography step. Approximately 1.5 mg of recom-binant protein was recovered from the columns from the start-ing 79 mg protein for a yield of 2%. All the bioactive, properlyfolded protein was in this fraction. Unbound and lower saltelution material (0–0.05 M NaCl) was not biologically active.

Leukemia

Figure 1 Construction of pRKDTLIL3 bacterial expression plasmid.The BBG14 plasmid is a pUC18 plasmid with synthetic IL3 cDNAcloned between the HindIII and EcoRI sites in the polylinker. A PCRreaction was performed with the 59-oligonucleotide: 59-GCAGTCGACCATATGGGCGGAGGCGGAAGTGGAGGAGGAGGCAGCGCTCCCATGACCCAGACA-39 and the 39-oligonucleotide: 59-GCAGTCGCAAAGCTTCTAAAAGATCGCTAGCGACAA-39. The PCRproduct was subcloned in the pCR-TOPO vector, restricted with NdeIand HindIII and ligated to the Ndel–HindIII restricted, alkaline phos-phatase treated pRKDTGM 3.7 kb DNA fragment. Ligated DNA trans-formants were used to make plasmid preparations and these weresequenced and the pRKDTLIL3 plasmids isolated.

Similarly, higher salt elution protein was not biologicallyactive and consisted of DT fragments (30–40 kDa Mr) with apI of 5. A S200 column was then run in PBS at 0.5 ml/min.Monomeric DTLIL3 was recovered (0.45 mg/3) separate fromapproximately equal amounts of dimer and aggregate. Again,all the bioactivity was in the monomeric peak. Thus, the finalyield of .80% pure DTLIL3 was 0.6% of the startingdenatured protein. Endotoxin levels were minimized by exten-sive Triton X-100 washes of the inclusion bodies and by separ-ating on anion exchange and size exclusion chromato-graphic matrices.

Characterization of DTLIL3

Reducing SDS-PAGE showed that the recombinant moleculewas .80% pure. The recombinant protein had an apparentmolecular weight of 58 kDa (Figure 2a). Immunoblots withanti-DTA and anti-huIL3 revealed reactivity of a major proteinspecies of the anticipated molecular weight (Figure 2b and c).There are minor species of 50–55 kDa which react with anti-DTA but not anti-huIL3. They constitute between 10 and 20%of the product. ADP-ribosylating activity was found in DTLIL3.Compared to the enzymatic activity of CRM107 (a form ofdiphtheria toxin with the single point mutation S525F in thereceptor binding domain and a fully active enzymaticdomain), DTLIL3 was equally active. DTLIL3 lacked free thiolsbased on reaction with Ellman’s reagent. The control ricintoxin A chain had the expected two free thiols/molecule (oneinternal and one external) and the diphtheria fusion moleculehad <0.24 free thiols/molecule. Two-dimensional SDS-PAGE

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

580

Leukemia

Figure 2 10% SDS-PAGE of DTLIL3. (a) Commassie stained. (b) Anti-IL3 immunoblot. (c) Anti-DTA immunoblot. Molecular weight standardswere run on each gel and replaced by arrows at the left of each gel.

with isoelectric focusing revealed an approximately 58 kDaprotein with a pI of 6 and .80% purity by densitometry.HPLC of DTLIL3 on a TSK3000 column with PBS revealed asingle dominant peak with a retention time of 7.7 minwhereas BSA (67 kDa) and ovalbumin (43 kDa) gave retentiontimes of 7.5 and 8.3 min, respectively. This data was consist-ent with the predicted MW of DTLIL3 of 58 kDa. Reversephase HPLC was performed using a C4 Vydak column with a0–70% acetonitrile in 0.05% trifluoroacetic acid mobilephase. A single peak was observed at 49.4 min (49.5 min forDTGM). DTLIL3 (100 mg) was ethanol precipitated and ana-lyzed by Edmond degradation and HPLC (run No. 258, WakeForest Protein Analysis Core Lab). The N-terminal 14 aminoacids of DTLIL3 matched those of DT and DTGM. Trypticpeptide mapping of DTLIL3 showed a unique peptide pattern,different than DTGM. Tandem mass spectroscopy was perfor-med on a Quattro II triple quadrupole mass spectrometer run-ning MassLynx software and revealed a determined MW of58 320 daltons vs a calculated MW for DTLIL3 based on com-plete DNA sequencing of the plasmid and intact disulfidebonds of 58 321 daltons.

Cytotoxicity of DTLIL3 to leukemic cell lines

Toxicity of DTLIL3 for leukemic cell lines was assessed bymeasuring inhibition of cell proliferation using 3H-thymidineincorporation. The IC50 was calculated as the concentrationof fusion protein inhibiting thymidine incorporation by 50%compared to control. Representative cell cytotoxicity curveswith 24 h incubations in DTLIL3 are shown in Figure 3. Apanel of 12 cell lines was incubated for 48 h with the fusiontoxin and the results are described in Table 2. Receptor posi-tive cell lines were sensitive to drug with IC50s of 1–28 pM

whereas proliferation of the receptor negative cell lines wasnot impaired at concentrations exceeding 1400 pM. The lowreceptor density HL60 cell line showed intermediate sensi-tivity (IC50 = 119 pM). Co-incubation with human IL3(1500 ng/ml) blocked toxicity to TF1/Bcl2 cells by 120-fold

Figure 3 DTLIL3 cell cytotoxicity. Cell were incubated for 24 hwith fusion protein and then pulsed for 18 h with 3H-thymidine, andthen harvested on glass fiber mats and counted. (j) TF1/Bcl2; (G)CEM.

(Figure 4). The IC50 for TF1/Bcl2 without blocking IL3 was1 pM and with IL3 was 162 pM.

Inhibition of colony formation by DTLIL3 for the TF1/Bcl2IL3 receptor positive cell line and the CEM IL3 receptor nega-tive cell line was measured. The results are shown in Figure5. The IC50 for TF1/Bcl2 cells was 14 pM and at 1000 pM over3 logs cell kill was observed. This is the maximal log cell killdetectable with this assay. In contrast, no inhibition of colonyformation by CEM cells was seen even when incubated with20 000 pM DTLIL3.

AML193 cells were incubated with 2 ng/ml IL3 and with orwithout 624 pMDTLIL3 for 1, 2, 3 or 4 days and apoptosismeasured by morphology under phase contrast microscopyand flow cytometry after staining with annexin V-FITC andpropidium iodide. Apoptosis measured by flow cytometry in

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

581Table 2 Sensitivity of leukemia cell lines to diphtheria IL3 fusiontoxins

Cell line DTLIL3a IC50 (pM)

M07e 1TF1/DMEK+Bcl2 2TF1/Bcl2 4TF1/DMEK+LNL6 14TF1 17OCI-AML 23AML193 28HL60 119K562 .1400P815 .1400TK6 .1400CEM .1400

aAssay performed at least two different times with duplicatesamples at each of 12 fusion toxin concentrations. Results arethe mean.

Figure 4 DTLIL3 cell cytotoxicity for TF1/Bcl2 cells with 24 hincubation with (j) and without (G) excess blocking human IL3.

the absence of fusion toxin was 10% for each day, whileapoptosis in the presence of toxin rose to 30% on day 1, 31%on day 2, 40% on day 3 and 60% on day 4. Morphologicchanges and nuclear DNA fragmentation associated withapoptosis were clearly demonstrated after 48 h in the presenceof fusion protein (Figure 6). The experiment was performedtwice.

TF1 cells incubated with 50 ng/ml human GM-CSF and5 nM DTLIL3 for 0, 6, 12, 24, 48 and 60 h and evaluated forapoptotic nuclear bodies by Hoeschst staining showed 11%,10%, 10%, 12%, 31%, 47%, and 68% apoptosis, respect-ively. Control TF1 cells showed 10% apoptosis over the sametime period.

Leukemic progenitor sensitivity

Colony growth was observed under our semi-solid cultureconditions for 11/20 myeloid acute phase CML samples and9/17 AML samples. For 4/11 myeloid acute phase CML patient

Leukemia

Figure 5 DTLIL3 inhibition of colony formation in semi-solidmedium. Cells were exposed to fusion protein at different concen-trations for 48 h and then aliquoted in 0.3% agarose. Colonies with.20 cells counted on day 7. (j) CEM; (d) TF1/Bcl2.

progenitor cell samples, there was >1 log cell kill at 680 pM

with IC50s of 7–27 pM (Figure 7 and Table 3), and 3/9 AMLpatient progenitor cell samples had >1 log cell kill at 680 pM

with IC50s of 3–68 pM. Because colony growth on manysamples was limited to 10–200 colonies/105 cells, the maxi-mal log cell kill may be an underestimate. This assay detectedonly the reduction in colony growth seen in 35 mm dishes sothat the range varied from 1 to 3 logs.

Normal progenitor sensitivity

Five normal marrow light density mononuclear cell prep-arations were incubated with DTLIL3 for 48 h and thenassayed for CFU-GM. The IC50 was 81–675 pM with 0.5 logcell kill at 680 pM (Figure 7 and Table 3).

Discussion

Drug-resistant myeloid blasts frequently show overexpressionof membrane-dependent drug transporters (P-glycoproteinand lung resistance protein)27,28 and drug metabolizingenzymes (glutathione S-transferase p).29 Fusion toxins bypassthese drug resistance pathways due to their distinct molecularshape and mechanism of action. We have previously demon-strated the cytotoxic potency of the DTGM fusion toxin bothfor multi-drug resistant cell lines and fresh patient leukemicprogenitors.4,5 DTLIL3 may be particularly useful in acutephase CML where the majority of patients present with diseaserefractory to cytotoxic drugs that target cell proliferation orDNA.

The yield of DTLIL3 was low and reflects the lack of optim-ization of refolding. Nevertheless, the purity of final product,.80%, suggests the purification method is good. The con-taminant appears to be proteolytically cleaved DTLIL3 withsmaller 45 kDa Mr material that is immunoreactive with anti-DTA but not anti-IL3. Use of newer expression vectors anduse of affinity tags may yield improved yields and purity. Infact, preliminary results with a pET21 construct have giventhree-fold higher yields and .90% purity after Ni2+ affinitychromatography (unpublished results). We have generated

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

582

Leukemia

Figure 6 DTLIL3 apoptosis induction in AML193 cells. Cells were incubated for 48 h with or without 2 ng/ml IL3 and with or without 621 pM

DTLIL3. Panels a and b are phase contrast micrographs (100×; arrows indicate apoptotic bodies in panel b). Panel c is an agarose gel electro-phoresis with ethidium bromide staining of chromosomal DNA. (a) IL3 alone. (b) IL3 + DTLIL3. (c) Lane 1, media alone; lane 2, IL3 alone; lanes3 and 4, IL3 + DTLIL3. n = 2.

Figure 7 DTLIL3 inhibition of colony formation in semi-solidmedium. Cells exposed to fusion protein at different concentrationsfor 48 h and then aliquoted in Methocult H4434. Mean values withnumber of patients given in parentheses. (d) CFU-GM (5); (j) acutephase CML (4).

adequate quantities (2.5 mg) of DTLIL3 for further in vitro andin vivo testing.

Diphtheria fusion protein directed to the mouse IL3R hasbeen developed by two laboratories.25,26 In both cases, selec-tive toxicity to receptor positive murine leukemic blasts wasseen. Less damage to normal murine progenitors wasobserved, and adoptive transfer experiments indicated the

absence of IL3 receptors on at least a subset of repopulatingprogenitor cells.25 Our study with a diphtheria fusion proteintargeting the human IL3R has given similar results so far inthe ex vivo targeting of human IL3R-bearing cells.

DTLIL3 cytotoxicity was mediated by IL3R binding basedboth on blocking with human IL3 and selective toxicity toIL3R-positive leukemic cell lines. Toxicity was measured byinhibition of cell proliferation, dye exclusion, colony forma-tion and by the induction of apoptosis. The delay in apoptosisto 48–60 h corresponds with the time required for toxininternalization, translocation and EF2 inactivation.30

Mutation of the p53 gene and loss of its function is fre-quently observed in acute phase CML.31 Mutant p53 cellshave poor apoptosis induction by conventional cytotoxicdrugs.32 In our study, p53-deficient cell lines (including TF1)33

were readily killed by DTLIL3 suggesting the fusion toxinbypasses this mechanism of drug resistance.

Toxicity of DTLIL3 to fresh leukemic progenitors as well ascell lines is encouraging, since some targeted toxins havebeen much less effective on patient samples than celllines.34,35 However, a number of our patient samples wererelatively resistant to DTLIL3 suggesting that expression of theIL-3R on leukemic progenitors is variable. It is interesting tohypothesize that the sensitivity of some patient samples to thetoxin may reflect the presence of an IL3-IL3R autocrine loopin these cells as has been previously described.13

The reduced but measurable sensitivity of CFU-GM toDTLIL3 suggests the resistance of CFU-GM to DTGM is not ageneral resistance to diphtheria toxin-mediated cell damage.Specific attributes such as altered GM-CSF receptor physi-ology on normal vs malignant cells may account for theDTGM selectivity.36 Since earlier normal myeloid progenitors(LTC-IC) are less likely to require IL3 and express IL3 recep-

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

583Table 3 Sensitivity of normal and leukemic progenitors to DTLIL3fusion protein

Patients Assay Colonies IC50 (pM) Log cell kill(No./105 at 680 pM

cells)

Normals1 CFU-GM 20 675 0.52 CFU-GM 89 81 0.63 CFU-GM 39 81 0.54 CFU-GM 22 81 0.65 CFU-GM 107 675 0.5

Myeloidacute CML6 CFU-L 630 7 1.57 CFU-L 1138 9 18 CFU-L 350 12 19 CFU-L 67 27 1

10 CFU-L 23 2160 0.511 CFU-L 32 756 0.412 CFU-L 805 405 0.413 CFU-L 945 1080 014 CFU-L 1890 .6750 015 CFU-L 840 .6750 016 CFU-L 490 6750 0

AML17 CFU-L 40 3 1.518 CFU-L 36 3 1.019 CFU-L 66 68 1.020 CFU-L 23 270 0.521 CFU-L 30 270 0.422 CFU-L 86 81 0.423 CFU-L 58 810 0.324 CFU-L 133 34 0.725 CFU-L 15 41 0.6

tors, they may show greater resistance to DTLIL3.9 We arecurrently evaluating this hypothesis.

An in vitro therapeutic index reported by Ghielmini andothers, can be calculated by comparison of the IC50 of DTLIL3on sensitive leukemic CFC with that of normal CFU-GM.37–39

In these studies, we observed an in vitro index of 4–12, wherevalues greater than 1 would suggest selective toxicity towardsmalignant progenitors. In comparison, when conventionalchemotherapeutic agents including doxorubicin, melphalanand cytarabine have been tested in similar assays,37–39 theindices obtained were ,5. It will be interesting to observewhether or not the relative toxicity of DTLIL3 against normaland leukemic targets will be similar when more primitive pro-genitors that can be detected in long-term stromal co-culturesor immunudeficient mice are studied.

DTLIL3 may also target non-hematopoietic tissues pos-sessing high affinity IL3R.40–42 In vivo toxicology studies inanimals with cross-reactive receptors will be necessary toaddress potential non-hematopoietic toxicities.

Synergy with cytotoxic chemotherapy is likely to beobserved with DTLIL3 as has been documented withDTGM.24,43 Preliminary results suggest similar DTLIL3 synergywith cytarabine as was seen with DTGM (Frankel, unpub-lished observations). Applications of such biologicals withconventional chemotherapy should permit reductions in doseand toxicities of each agent and improved efficacy for thecombination.

DTLIL3 may be a clinically useful reagent. The related diph-theria fusion protein, DAB389 IL2, has completed a phase II

Leukemia

trial with 73 refractory cutaneous T cell lymphoma patients.44

At 9 and 18 mg/kg/day given by intravenous infusion over15 min for 5 days for two to eight courses, 10% of patientshad complete remissions and 20% of patients had partialremissions lasting 3–12+ months. The AML fusion toxin,DTGM, is currently in a phase I dose escalation study. Four-teen relapsed AML patients have been treated at 1, 2, 3, or4 mg/kg/day for 5 days by 15 min intravenous infusion withside-effects limited to rare fever, transaminasemia and hypo-tension ameliorated by steroid prophylaxis. Thus, there is agrowing clinical experience in leukemia and lymphomapatients with related molecules. DTLIL3 has the advantage ofminimal IL3R expression on monocytes/macrophages. In theabsence of effective strategies for the treatment of acute phaseCML and patients with refractory AML, the experiments in thisstudy suggest the need for further preclinical development ofDTLIL3 as a possible new therapeutic agent for this group ofpoor prognosis patients. DTLIL3 may be used clinically eitheras a purging agent for autologous stem cell or bone marrowtransplants or for systemic treatment of chemotherapy-resistant AML or blast crisis CML.

Acknowledgements

This work was supported in part by Leukemia Society ofAmerica Grant No. 6114–98 (to AF), NIH Grants No.R01CA76178 (to AF), R01CA54116 (to ET), R01CA51025 (toJM) and grants from the National Cancer Institutes of Canada(NCIC) to DH with funds from the Terry Fox Run.

References

1 Kantarjian HM, Keating MJ, Talpaz M. Chronic myelogenous leu-kemia in blast crisis. Am J Med 1987; 83: 445–454.

2 Keating MJ, Kantarjian H, Smith TL, Estey E, Walters R, AnderssonB, Beran M, McCredie KB, Freireich EJ. Response to salvage ther-apy and survival after relapse in acute myelgenous leukemia. JClin Oncol 1989; 7: 1071–1080.

3 List AF. Role of multidrug resistance and its pharmacologicalmodulation in acute myeloid leukemia. Leukemia 1996; 10:937–942.

4 Frankel AE, Hall PD, Burbage C, Vesely J, Willingham MC, BhallaK, Kreitman RJ. Modulation of the apoptotic response of humanmyeloid leukemia cells to a diphtheria toxin granulocyte–macro-phage colony-stimulating factor fusion protein. Blood 1997; 90:3654–3661.

5 Hogge DE, Willman CL, Kreitman RJ, Berger M, Hall PD, KopeckyKJ, McLain C, Tagge EP, Eaves CJ, Frankel AE. Malignant progeni-tors from patients with acute myelogenous leukemia are sensitiveto a diphtheria toxin–granulocyte–macrophage colony-stimulatingfactor fusion protein. Blood 1998; 92: 589–595.

6 Frankel AE, Lilly M, Kreitman RJ, Hogge D, Beran M, FreedmanMH, Emanuel PD, McLain C, Hall PD, Tagge EP, Berger M, EavesC. Diphtheria toxin fused to granulocyte–macrophage colony-sti-mulating factor is toxic to blasts from patients with juvenile myelo-monocytic leukemia and chronic myelomonocytic leukemia.Blood 1998; 92: 4279–4286.

7 Frankel AE, Ramage J, Latimer A, Feely T, Delatte S, Hall PD,Tagge EP, Kreitman RJ, Willingham MC. High-level expression andpurification of the recombinant diphtheria fusion toxin DTGM forphase I clinical trials. Protein Expression Purification 1999; 16:190–201.

8 Suda T, Suda J, Ogawa M, Ihle JN. Permissive role of interleukin3 (IL-3) in proliferation and differentiation of multipotential hema-topoietic progenitors in culture. J Cell Physiol 1985; 124: 182–190.

9 Ogata H, Taniguchi S, Inaba M, Sugawara M, Ohta Y, Inaba K,

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

584

Leukemia

Mori J, Ikehara S. Separation of hematopoietic stem cells into twopopulations and their characterization. Blood 1992; 80: 91–95.

10 Park LS, Waldron PE, Friend D, Sassenfeld HM, Price V, AndersonD, Cosman D, Andrews RG, Bernstein ID, Urdal DL. Interleukin-3,GM-CSF, and G-CSF receptor expression on cell lines and primaryleukemia cells: receptor heterogeneity and relationship to growthfactor responsiveness. Blood 1989; 74: 56–65.

11 Miyauchi J, Kelleher CA, Wong GG, Yang YY, Clark SC, MinkinS, Minden MD, McCulloch EA. The effects of combinations of therecombinant growth factors GM-CSF, G-CSF, IL-3, and CSF-1 onleukemic blast cells in suspension culture. Leukemia 1988; 2:382–387.

12 Ailles LE, Gerhard B, Hogge DE. Detection and characterization ofprimitive malignant and normal progenitors in patients with acutemyelogenous leukemia using long-term coculture with supportivefeeder alyers and cytokines. Blood 1997; 90: 2555–2564.

13 Jian X, Lopez A, Holyoake T, Eaves A, Eaves C. Autocrine pro-duction and action of IL-3 and G-CSF in chronic myeloid leuke-mia. Proc Natl Acad Sci USA 1999; 96: 12804–12809.

14 Steelman LS, Algate PA, Blalock WL, Wang XY, Prevost KD, HoylePE, McCubrey JA. Oncogenic effects of overexpression of theinterleukin-3 receptor on hematopoietic cells. Leukemia 1996; 10:528–542.

15 Hara T, Miyajima A. Two distinct functional high affinity receptorsfor mouse interleukin-3 (IL-3). EMBO J 1992; 11: 1875–1884.

16 Nicola NA, Peterson L, Hilton DJ, Metcalf D. Cellular processingof murine colony-stimulating factor (Multi-CSF, GM-CSF, G-CSF)receptors by normal hematopoietic cells and cell lines. GrowthFactors 1988; 1: 41–49.

17 Kreitman RJ, Pastan I. Recombinant toxins containing human gra-nulocyte–macrophage colony-stimulating factor and eitherPseudomonas exotoxin or diphtheria toxin kill gastrointestinalcancer and leukemia cells. Blood 1997; 90: 252–259.

18 Collier RJ, Kandel J. Structure and activity of diphtheria toxin. JBiol Chem 1971; 246: 1496–1503.

19 Hwang J, FitzGerald DJ, Adhya S, Pastan I. Functional domainsof Pseudomonas exotoxin identified by deletion analysis of thegene expressed in E. coli. Cell 1987; 48: 129–136.

20 Greenfield L, Johnson VG, Youle RJ. Mutations in diphtheria toxinseparate binding from entry and amplify immunotoxin selectivity.Science 1987; 238: 536–539.

21 Ellman GL. Tissue sulfhydryl groups. Arch Biochem Biophys 1959;82: 70–75.

22 Wallin R, Culp EN, Coleman DB, Goodman SR. A structuralmodel of human erythrocyte band 2.1: alignment of chemical andfunctional domains. Proc Natl Acad Sci USA 1984; 81: 4095–4099.

23 Matsudaira P. A Practical Guide to Protein and Peptide Purifi-cation for Microsequencing. Academic: San Diego, CA, 1993.

24 Kim CN, Bhalla K, Kreitman RJ, Willingham MC, Hall P, TaggeEP, Jia T, Frankel AE. Diphtheria toxin fused to granulocyte–macrophage colony-stimulating factor and Ara-C exert synergistictoxicity against human AML HL-60 cells. Leukemia Res 1999; 23:527–538.

25 Chan CH, Blazar BR, Greenfield L, Kreitman RJ, Vallera DA. Reac-tivity of murine cytokine fusion toxin, diphtheria toxin390 murineinterleukin-3 (DT390-mIL-3), with bone marrow progenitor cells.Blood 1996; 88: 1445–1456.

26 Liger D, vanderSpek JC, Gaillard C, Cansier C, Murphy JR,Leboulch P, Gillet D. Characterization and receptor specific tox-icity of two diphtheria toxin-related interleukin-3 fusion proteins,DAB389-mIL3 and DAB389-(Gly4Ser)2-mIL-3. FEBS Lett 1997; 406:157–161.

27 Gsur A, Zochbauer S, Gotzl M, Kyrle PA, Lechner K, Pirker R.MDR1 RNA expression as a prognostic factor in acute myeloidleukemia: an update. Leuk Lymphoma 1993; 12: 91–94.

28 List AF, Spier C, Grogan T, Johnson C, Roe D, Greer J, Wolff S,Broxterman H, Scheffer G, Scheper R, Dalton W. Overexpressionof the major vault transporter protein lung-resistance protein pre-dicts treatment outcome in acute myeloid leukemia. Blood 1996;87: 2464–2469.

29 Russo D, Marie JP, Zhou DC, Faussat AM, Melli C, Damiani D,Michelutti A, Michieli M, Fanin R, Baccarani M. Evaluation of theclinical relevance of the anionic glutathione-S-transferase (GST pi)

and multidrug resistance (mdr-1) gene coexpression in leukemiasand lymphomas. Leuk Lymphoma 1994; 15: 453–468.

30 Falnes PO, Olsnes S. Modulatiion of the intracellular stability andtoxicity of diphtheria toxin through degradation by the N-end rulepathway. EMBO J 1998; 17: 615–625.

31 Mashal R, Shtalrid M, Talpaz M, Kantarjian H, Smith L, Beran M,Cork A, Trujillo J, Gutterman J, Deisseroth A. Rearrangement andexpression of p53 in the chronic phase and blast crisis of chronicmyelogenous leukemia. Blood 1990; 75: 180–189.

32 McCurrach ME, Connor TM, Knudson CM, Korsmeyer SJ, LoweSW. Bax-deficiency promotes drug resistance and oncogenictransformation by attenuating p53-dependent apoptosis. Proc NatlAcad Sci USA 1997; 94: 2345–2349.

33 Orashima M, Teoh G, Chauhan D, Ogata A, Shirahama S, ChiharuKK, Matsuzaki M, Matsushima H, Akiyama M, Yuza Y, MaekawaK, Anderson K. MDM2 protein overexpression inhibits apoptosisof TF-1 granulocyte–macrophage colony-stimulating factor-dependent acute myeloblastic leukemia cells. Blood 1998; 92:959–967.

34 Hertler A, Schlossman D, Borowitz M, Blythman H, Casellas P,Frankel AE. An anti-CD5 immunotoxin for chronic lymphocyticleukemia: enhancement of cytotoxicity with human serum albu-min-monensin. Int J Cancer 1989; 43: 215–219.

35 Kreitman RJ, Chaudhary VK, Waldmann T, Willingham MC, Fitz-Gerald DJ, Pastan I. The recombinant immunotoxin anti-Tac (Fv)-Pseudomonas exotoxin 40 is cytotoxic toward peripheral bloodmalignant cells from patients with adult T-cell leukemia. Proc NatlAcad Sci USA 1990; 87: 8291–8295.

36 Cannistra SA, Koenigsmann M, DiCarlo J, Groshek P, Griffin JD.Differentiation associated expression of two functionally distinctclasses of granulocyte–macrophage colony-stimulating factorreceptor by human myeloid cells. J Biol Chem 1990; 265:12656–12563.

37 Ghielmini M, Colli E, Bosshard G, Pennella G, Geroni C, Torri V,D’Incalci M, Cavalli F, Sessa C. Hematotoxicity on human bonemarrow- and umbilical cord blood-derived progenitor cells andin vitro therapeutic index of methoxymorpholinyldoxorubicin andits metabolites. Cancer Chemother Pharmacol 1998; 42: 235–240.

38 Ghielmini M, Bosshard G, Capolongo L, Geroni MC, Pesenti E,Torri V, D’Incalci M, Cavalli F, Sessa C. Estimate of the haemato-logical toxicity of minor groove alkylators using tests on humancord blood cells. Br J Cancer 1997; 75: 878–883.

39 Leglise MC, de Tailly PD, Vignot JL, Le Bot MA, Le Roux AM,Riche C. A cellular model for drug interactions on hematopoiesis.Cell Biol Toxicol 1996; 12: 39–53.

40 Morikawa Y, Tohya K, Hara T, Kitamura T, Miyajima A. Expressionof IL3 receptor in testis. Biochem Biophys Res Commun 1996;226: 107–112.

41 Korpelainen EI, Gamble JR, Vadas MA, Lopez AF. IL3 receptorexpression, regulation and function in cells of the vasculature.Immunol Cell Biol 1996; 74: 1–17.

42 Tabira T, Chui DH, Fan JP, Shirabe T, Konishi Y. Interleukin-3 andinterleukin-3 receptors in the brain. Ann NY Acad Sci 1998; 840:107–116.

43 Frankel AE, Hall PD, McLain C, Safa AR, Tagge EP, Kreitman RJ.Cell-specific modulation of drug resistance in acute myeloid leu-kemic blasts by targeted toxins. Bioconj Chem 1998; 9: 490–496.

44 Kreitman RJ. IBC’s International Conference on novel therapeuticproteins for oncology (drug discovery and development). ExpOpin Invest Drugs 1998; 7: 1–8.

45 Hendrie PC, Miyazawa K, Yang YC, Langefeld CD, Broxmeyer HE.Mast cell growth factor (c-kit ligand) enhances cytokine stimu-lation of proliferation of the human factor-dependent cell line,M07e. Exp Hematol 1991; 19: 1031–1037.

46 Kitamura T, Tange T, Terasawa T, Chiba S, Kuwaki T, MiyagawaK, Piao YF, Miyazono K, Urabe A, Takaku F. Establishment andcharacterization of a unique human line that proliferates depen-dently on GM-CSF, IL-3 or erythropoietin. J Cell Physiol 1989;140: 323–334.

47 Dalton WT, Ahearn M, McCredie KB, Freireich EJ, Stass SA, Tru-jillo JM. HL-60 cell line was derived from a patient with FAB-M2and not FAB-M3. Blood 1988; 71: 242–247.

48 Uzman BG, Foley GE, Farber S, Lazarus H. Morphologic vari-ations in human leukemic lymphoblasts (CCRF-CEM) after long

DTLIL3 is toxic to blasts from patients with myeloid leukemiasAE Frankel et al

585term culture and exposure to chemotherapeutic agents. Cancer1966; 19: 1725–1742.

49 Lange B, Valtieri M, Santoli D, Caracciolo D, Mavilio F, Gemperl-ein I, Griffin C, Emanuel B, Finan J, Nowell P, Rovera G. Growthfactor requirements of childhood acute leukemia: establishmentof GM-CSF-dependent cell lines. Blood 1987; 70: 192–199.

50 Klein E, Ben-Bassat H, Neumann H, Ralph P, Zeuthen J, PolliackA, Vanky F. Properties of the K562 cell line, derived from a patientwith chronic myeloid leukemia. Int J Cancer 1976; 18: 421–431.

51 Ralph P, Nakoinz I. Phagocytosis and cytolysis by a macrophagetumour and its cloned cell line. Nature 1975; 257: 393–394.

52 Liber HL, Thilly WG. Mutation assay at the thymidine kinase locusin diploid human lymphoblasts. Mutat Res 1982; 94: 467–485.

Leukemia

53 Isogai C, Murate T, Tamiya-Koizumi K, Yoshida S, Ito T, Nagai H,Kinoshita T, Kagami Y, Hotta T, Hamaguchi M, Saito H. Analysisof bax protein in sphringosine-induced apoptosis in the humanleukemic cell line TF1 and its bcl-2 transfectants. Exp Hematol1998; 26: 1118–1125.

54 McCubrey JA, Steelman LS, Hoyle PE, Blalock WL, Weinstein-Oppenheimer C, Franklin RA, Cherwinski H, Bosch E, McMahonM. Differential abilities of activated Raf oncoproteins to abrogatecytokine dependency, prevent apoptosis and induce autocrinegrowth factor synthesis in human hematopoietic cells. Leukemia1998; 12: 1903–1929.