McCulloch (The Ontario Cancer Institute, Toronto, Canada)31; and ... mg/mL by pressure-driven ultrafiltration on an Ultrafree PF-filter ... After washing, SA-.
The Receptor Tyrosine Kinase ~ 1 8 Is 5Expressed ~ ~ on ~ a ~Subset of B-Lymphoid Blasts Prom Patients With Acute Lymphoblastic Leukemia and Chronic Myelogenous Leukemia By Hans-Jorg Buhring, lrmi Sures, Bahija Jallal, F. Ulrich Weiss, Friedrich W. Busch, Wolf-Dieter Ludwig, Rupert Handgretinger, Hans-Dierck Waller, and Axel Ullrich The class I receptor tyrosine kinase (RTK) HER2 is an oncoprotein that is frequently involved in the pathogenesis of tumors ofepithelial origin.Here we report mRNA expression in peripheral blood and bone marrow cells from healthy donors in hematopoietic cell lines and leukemic blasts from patients with acute lymphoblasticleukemia (ALL), acute myeloblastic leukemia (AML), chronic lymphoblastic leukemia (CLL), and chronic myeloid leukemia (CML). However, cell surface expression of HER2 protein ( ~ 1 8 5 ’ ~was ~ ~ found ) exclusively on a subset of leukemic cells of the B-lymphoblastic lineage. ~ 1 8 5 ’ ~expression ~’ was found on blasts in 2 of 15 samples from infants, 9 of 19 samples from adult patients with C-ALL [CDlS+CDIO+),and 1 of 2 samples from patients with pro-B ALL (CDISfCD1O-), whereas none ofthe leukemic
cells from patients with AML (0130). T-ALL (0/7), CLL (0/5) (CDIS+CD5+),or CML in chronic and accelerated phase (O/ 51 or in blast crisis with myeloid differentiation (0/14) were positive for ~ 1 8 5 ” ~ ~However, ’. cells from 3 of 4 patients with CML in B-lymphoid blastcrisis (CDlS+CDIO’) expressed high levels of ~185’~~’, which was also found on the surface of the CML-derived B-cell lines BV-173 and Nalm-l. Our study shows ~ 1 8 5 ” ~ expression ~’ on malignant cells of hematopoietic origin for the first time. Aberrant expression of this oncogenic receptor tyrosinekinase in hematopoietic cell types may bean oncogenic event contributing to thedevelopment of a subset of B-lymphoblastic leukemias. 0 1995 by The American Society of Hematology.
T
dimerization of p1 8SH“R2 molecules and the resulting receptor activation is facilitated by overexpression of p 18SHER2. l‘’ Although no consensus view exists to date about themechanism of ligand-dependent activation of p18SHER2,in two recent reports it wassuggested that neu-activating factor (NAF)”or ascitessialoglycoprotein-2 (ASGP-2) directly bind to and activate p18SH”R2.’2 p18SHER2 expressed is on several cell lines of nonhematopoietic origin that include cellsderived from neuroblastomas and carcinomas of the breast, stomach, and lung.’”’” Overexpression is found at a rate of 20% to 40%in cell lines derived from carcinomas of the breast, lung, ovary, stomach, colon, andcervix.’”” Overexpressionof p18SHRR2is frequently caused by an amplification of the cellular gene and in many cases has been shown to correlate with poor prognosis.” 1 8SHER’ expressionandpathogenic Althoughnormalp overexpression is usually found in epithelial cell types, we report in the present study that low levels of HER2 mRNA are expressed in normalhematopoietic cells and leukemic blasts of different sources, whereas detectable p I 8SHER2surface expression is found exclusively on the blasts from a fraction of patients with B-lymphoblastic leukemia.
HE HER2/neu/erbB-2 proto-oncogene encodes a transmembrane receptor, p18SHRR2, whichbelongs to the class I receptor tyrosine kinase (RTK) family’.’ and is structurally related to the epidermal growth factorreceptor (EGFR), theHER3/erbB-3encoded protein p180“‘,R3, andthe These receptors HER4/erbB-4 encoded protein p1 80HER4.Z-5 share a commonmolecular architecture and contain two cystein-rich regions within their extracellular domains and structurallyrelated enzymatic regionswithintheircytoplasmatic domains. p180HERR’ most probably lacks enzymatic activitybecause fouraminoacid residues that are highly conserved orinvariant in the catalytic regions of both serine/ threonineandtyrosine proteinkinases are altered in the p180HER3 catalytic domain.‘ Recently, it has been shown that ~ 1 8 5 is~ not ~ ~only ’ activatedby receptor homodimerization, but also by formation of heterodimer complexes with EGFR, pI8OHER3,and p180HER4.7-y Ligand-stimulated heterodimerization is thought to be induced by EGF to form EGF-R/ p18SHER2complexesor by neu differentiation factodheregulin (NDF/HRG) form to p18SHER2/p180HER3 and/or p1 8SHER2/p I 80HER4 complexes.‘” Ligand-independent homo-
From the Second Department of Internal Medicine, Universih) of Tiibingen, Tiibingen; the Department of Molecular Biology, MarPlanck-Institute for Biochemistry, Martinsried; the Department of Hematology, Robert Bosch Hospital, Stuttgart; the Department of Medical Oncology, Robert Rdssle Hospital, Berlin; and Childrens Hospital, University of Tiibingen, Tiibingen, Germany. Submitted November IS, 1994; accepted April 19, 1995. Supported by the Deutsche ForschungsRemeinschaft (SF% 120, project CS) and SUGEN, Inc. Address reprint requests to Hans-JOrg Biihring, PhD, Second Department of Medicine, FACS LaboratotyOtfried-Muller-Str 10, 72076 Tiibingen, Germany. The publication costs of this article were defrayed in part by page chargepayment.Thisartide must therefore he hereby marked “advertisement” in rrccorhnce with I8 U.S.C. section 1734 solely Io indicate this juct. 0 1995 by The Americcm Society of Hematology. 0006-4971/95/8605-0038$3.00/0
1916
MATERIALS AND METHODS
Immunizution and Hybridoma Production Several monoclonal antibodies (MoAbs) with specificity for the extracellular domain of HER2 were raised by immunization of Balb/ c mice by subsequent injections of NIH/3T3 cells transfected with the human protooncogene HER2. The ~ 1 8 5 ~ transfected ~ ” cell line was constructed by transfecting an expression vector containing the complete human HER2 cDNA” into mouse NIH-3T3 cells.24The detailed immunization and fusion protocol is described elsewhere.?’ MoAb were selected that specifically recognize the transfectant cell line.but not NIH-3T3 cells transfected only with the expression vector. The specificity was confirmed by indirect immunofluorescence and immunoprecipitation of ~ 1 8 5 ~ ‘ ~MoAb ’. 24D2 (IgGI) was selected for screening by flow cytometric analysis because stainingof cells with this reagent resulted in optimal signals. MoAb 13AI (IgG2a) and 13D1 (IgG2a) were selected for downregulation experiments because they showed growth inhibitory activity on
Blood, Vol 86, No 5 (September 1). 1995: pp 1916-1923
1917
HER2 PROTEIN IS EXPRESSED ON LEUKEMIC BLASTS
SKBR3 breast carcinoma cells (A.U., unpublished observation, May 1994).
Cells Normal peripheral blood (PB) cells, bone marrow (BM) cells, and leukemic samples were drawn after informed consent. The study was approved by our local ethics committee. Erythrocytes were analyzed by using 5 pL of unfractionated heparinized PB cells for each assay. Platelets were obtained by collecting the plasma after settling of erythrocytes for 30 minutes and centrifugation at 2,000 rpm. The purity of erythrocyte and platelet populations was confirmed by labeling erythrocytes with the erythrocyte-specific marker glycophorin A and labeling platelets with the platelet-specific marker CD41. Analysis in a FACScan flow cytometer (Becton Dickinson, Mountain View, CA) showed that the corresponding populations were more than 99% glycophorin A+ and CD41+, respectively. A mixture of granulocytes, monocytes, and lymphocytes was obtained after lysis of erythrocytes by Optilyse reagent (Immunotech, Marseille, France) according to the manufacturer’s instructions. The three fractions were analyzed separately by appropriate gate setting in a dual scatter plot on a FACScan flow cytometer. Before analysis, normal mononuclear BM cells and leukemic cells (from PB) were purified by collecting the interphase cells on a Ficoll-Hypaque gradient (Pharmacia, Freiburg, Germany).
PuriJcation of CD34+ BM Cells CD34’ bone marrow cells were purified on a Ceprate LC-34 affinity column (CellPro, Bothell, WA). Gravity flow separation of 8 X lo7mononuclear cells was performed according to the manufacturer’s instructions. Nine X IO5cells containing 81% of CD34’ cells were collected after separation. Positivity was estimated using a conjugate of the human progenitor cell antigen-specific antibody, HPCA-2-FITC (Becton Dickinson).
Cell Lines K562 and HEL (erythroleukemias); KG-l, KG-la, HL-60, and U937 (myeloid leukemias); HSB-2, CCRF-CEM, and Molt-4 (Tlymphoblastic leukemias); and Daudi, Reh, and Cess (B-lymphoblastoid leukemias) were obtained from the American Type Culture Collection (ATCC; Rockville, MD). The cell lines M07-e (megakaryoblastic leukemia); T ” , Nalm-l, BV-173, CML-TI, SPI-801, and EM-2 (derived from chronic myeloid leukemia [CML]); and U-266 (myeloma) were purchased from the German Collection of Microorganisms and Cell Cultures (DSM). The following cell lines were kind gifts: UT-7 (erythroleukemic) from Dr Komatsu (Jichi Medical School, Tochigi, Japan)26; MEG-01(megakqoblastic) from Dr Saito (Nagoia University, School of Medicine, Nagoia, Japan)”; TF-l (erythroleukemic) from Dr Kitamura (University of Tokyo, Faculty of Medicine, Tokyo, Japan)”; MOLM-1 (megakaryoblastic) from Dr Minowada (Fujisaki Cell Center, Okayama, Japan)29;DU.528 (myeloid) from Dr Kurtzberg (Duke University Medical Center, Durham, NC)30; OCYAML-4 (myeloid) from Dr McCulloch (The Ontario Cancer Institute, Toronto, Canada)31;and Km3 (pre-B-lymphoblastoid) from Dr Greaves (Leukaemia Research Fund Centre, London, UK).
Biotinylation of MoAb 2402 After purification on a T-Gel column (Pierce, distributed by Kastner, Rottenburg, Germany) MoAb 24D2 was dialyzed against 0.1 m o m sodium borate buffer, pH 9.3, and concentrated to about 1 mg/mL by pressure-driven ultrafiltration on an Ultrafree PF-filter (10,000 nominal molecular weight limit [NMWL]; Millipore, Eschborn, Germany). A freshly prepared solution of 1 mg/mL €-amino caproic acid N-hydroxy succinimide biotin (biotin-X-NHS) was added to the antibody at a molar ratio of 3 0 1 and incubated for 2.5 hours at room temperature. The reaction was stopped by thorough dialysis against Hanks’ balanced salt solution. The resulting antibody concentration was estimated by Coomassie protein assay reagent. The optimal concentration of the biotinylated antibody for flow cytometric analysis was evaluated by incubating BV-173 cells with serial dilutions of the conjugate and staining with streptavidin-phycoerythrin (SA-PE).
Immunojluorescence Staining Single-color staining. For blocking of unspecific binding to Fc receptors, cells were incubated with 10% human AB serum (Behring, Marburg, Germany) for 10 minutes on ice. In the next step, cells were incubated with the designated antibody (2 pg/mL) for 15 minutes and stained after washing with the F(ab‘), fragment of a PE-conjugated sheep-antimouse serum (Sigma, Mlinchen, Germany). Background staining was performed with an IgGl control antibody (Southern Biotechnology, distributed by Dunn, Asbach, Germany). BM cells that were selected with Ceprate LC-34 columns (CD34’ cells) were labeled with MoAb 24D2 (IgG1) and stained with an isotypematched goat-antimouse antibody conjugated with PE (Southern Biotechnology). Dual-color staining. Two-color staining of leukemic blasts and BM cells was performed by labeling the cells with biotinylated MoAbs 24D2 (24D2-b) and HPCA-2-FITC (CD34), CALLA-FITC (CDlO), or Leul2-FITC (CD19), respectively. After washing, SAPE was added to stain MoAb 24D2-b. To indicate background staining, cells were labeled with a mixture of control IgGl-FITC and control IgGl -biotin plus SA-PE.
Flow Cytometric Analysis Flow cytometric analysis was performed on a FACScan flow cytometer equipped with an argon ion laser. The fluorescence of the cells was excited at 488 nm with an argon laser. The FITC signals were detected through a 530-nm band pass filter and the PE signals were detected through a 570-nm band pass filter. Ten thousand events were acquired and analyzed on an Amatic 486 PC using inhouse programs. For two-color analysis, 25,000 cells were acquired after proper setting of the compensation with Calibrite beads (Becton Dickinson). Dot-plot displays were analyzed using the Lysys I1 program (Becton Dickinson).
MoAb-Mediated Downregulation of p m H E R 2
Antibody-mediated downregulation was estimated using a slight modification of a previously described protocol.3zIn brief, cells were incubated in phosphate-buffered saline (PBS) for 2 hours at 37°C Antibodies either with a control antibody or 5 pg/mL of p185HERZ-reactive 13A1 or 13D1. After blocking and washing with ice-cold The following MoAbs were used 24D2, 13A1, 13D1 ( ~ 1 8 5 ~ ~MoAb ) produced in our own laboratories; CALLA (CDlO), Leu-12 (CD19), PBS containing 0.1% NaN3, cells were labeled either with an IgGl and HPCA-2-FITC (CD34) from Becton Dickinson; and IOM41control antibody or with 13A1 or 13D1, respectively, and stained FITC (CD41) and glycophorin A-FITC from Immunotech. for fluorescence analysis as described above.
1918
NIH-3T3-control NIH-3T3-HER2
BUHRING ET AL
with the expression vector. Figure 1 shows that MoAb 24D2 recognizes exclusively the transfectant, but not the control cell line. Similar reactivity profiles were obtained with the other p185H"R'-reactive MoAb (data not shown). The isotypes of the MoAb obtained from cloned hybridoma cells were determined by flow cytometric analysis using PE-conjugated isotype-specific secondary stepantimouse sera. MoAb 24D2 (IgG1 ) wasselected for further experiments because immunofluorescence staining with this reagent reFig 1. Immunofluorescence stainingof HEW-transfected (left)and sulted in optimal signals(data not shown).MoAbs 13A1 control NIH-3T3 cells (right) with the p185HEw-specificMoAb 2402. The cells were labeled with either MoAb 24D2 oran isotype-matched and 13Dl (both IgG2a) were used for downregulation expercontrol antibody (IgG1) and stained with a PE-conjugate of a goatiments because they showed a growth inhibitory effect on antimouse antiserum and analyzed on a FACScan flow cytometer. SKBR3 cells (A.U., unpublished observation, May 1994). Background staining is indicated by the dotted lines. To confirm the specificities of the MoAb, the molecular mass of therecognizedantigenwas estimated by sodium dodecyl sulfate-polyacrylamidegelelectrophoresis (SDSPolymerase Chain Reaction (PCR) Analysisof HER2 PAGE) of immunoprecipitatedlysates fromSKBr3 cells. mRNA Expression All theinvestigated MoAbs precipitatedaprotein with a molecular mass of about 185 kD, which corresponds to the RNA was isolated from frozen cell pellets (1 X l Oh to 5 X 1Ox expected molecularweight of the HER2 protein (data not cells) as described by Puissant and Houdebine.33 Two micrograms each of total RNA was treated with RNase-free DNase and subjected shown). to reverse transcription and PCR amplification (55°C annealing temperature, 30 cycles) essentiallyas described by Fuqua et al.34 Intron- Expression ofplHSHLn2on Established Cell Lines and flanking sense and antisense primers were designed complementary Blasts From Leukemia Patients to HER2 nucleotides at position 935-953 and 1,245-1.263, respecThe leukemic cell lines used for analysis of p185H"R2extively, according to Coussens et aIz3 in Fig 1. PCRproductswere pression are summarizedin Table 1. Only the B-lymphoblasvisualized on 8% polyacrylamide gels by ethidium bromide staining, tic cell lines BV-173 and Nalm-l express detectable numbers eluted, cloned, and identified by DNA sequence analysis." of ~ 1 8 5 on ~ their ~ ~ surfaces. ' Figure 2 shows that ~ 1 8 5 ~ " ~ ' Tyrosine Phosphorylation of p18jHER2 in Leukemic Cell expression is higher on BV- I73 (Fig 2A) than on Nalm- 1 cells (Fig 2B). Lines The blasts of 18 infant and86 adultpatients suffering Tyrosine phosphorylation of p18SHERZ in leukemic cell lines was from chronic or acute leukemia wereanalyzed on a flow estimated using comparable amountsof cells ( IO-cm dish) that were cytometer for their cell surface expression of ~ 1 8 5 " ~ As ~'. either grownin RPM1 1640 plus 10% fetal bovine serum (CCRFshown in Table 2, the cells from 2 of 15 infants and 9 of CEM, BV-173, and Nalm-l) supplementedwith 10 ng/mLof interleukin-3 (M07e), or incubated for additional 10 minutes with 100 19 adult patients withcommon acutelymphoblastic leukemia &mLpervanadate.Cells were disruptedin 0.5 mL lysis buffer " ~their ' surface. In addition, (C-ALL) expressed ~ 1 8 5 ~ on containing 50 mmol/L HEPES, pH 7.5, 150 rnmoVL NaCl, 5 mmol/ the cells of one of two pro-B ALL (CD19TD10-) and three L EDTA,10% (voUv01) glycerol, 1% (voUvol) Triton X-100, 1 of four CML in blast crisis (CML-BC) samples with a CmmoVL phenylmethylsulfonyl fluoride, 1 mmol/L sodium orthovanFigure ALL phenotype (CD19'CDlO') expressed p1 85HER2. adate, and 10 m g h L aprotinin.Immunoprecipitation of ~ 1 8 5 " ~ ~ ' 2C shows the distribution of ~ 1 8 5 " on ~ ~the ' cells of three was performed using a polyclonal antibody directed against the last one CML-BCsample representative C-ALLsamplesand 17 C-terminalaminoacids,andtheresultingimmunoprecipitates with B-lymphoblastic blast crisis (Fig 2C, bottom right). No were separated on a 7.5% polyacrylamide gel and transferred onto expression of ~ 1 8 5 was ~ ~ observed ~ ' on theblasts from nitrocellulose. After incubation with a monoclonal antiphosphotyropatients with T-ALL (0/3 samples from infants and 0/7 from sine antibody (5E2) immune complexes were detected witha horseradish-peroxidaseconjugatedgoat-antimouseantiserum(BioRad, adultpatients), acute myeloblasticleukemia (AML; 0/30), Miinchen, Germany), and an enhancedchemiluminescence(ECL) 0/5), CML in chronic chronic lymphoblastic leukemia (CLL; detection system (Amersham, Dreieich, Germany). After stripping or accelerated phase (0/5),or CML-BC with myeloid differaccording to the manufacturer'sprotocolthemembrane was reentiation (0/14). probed with a polyclonal anti-p185HER2 specific antibody followed Toestimate whether the percentage of p1UHER2' cells by ECL detection. roughly reflects the percentage of leukemic blasts within the samples, peripheral blood C-ALL samples containing cells RESULTS with highlevels of p 185HER2 expression and morphologically MoAbs 13A1, 1301, and 2 4 0 2 Recognize p18jHER2 well-distinguishable blasts were selected.To achieve optimal Fusion of SP2/0 myeloma cells with the spleen cells from discrimination between the leukemic blasts and normal cells we double-labeled thecells of foursamples with MoAb Balb/c mice immunized with NIH-3T3 cells transfected with the HER2p r o t o - ~ n c o g e n e ~resulted " ~ ~ in 24 hybridoma cells 24D2 (24D2-biotin plus SA-PE) and with MoAbs against CD34 (CD34-FITC), CD19 (CD19-FITC), or CD10 (CD10secreting MoAbs thatspecifically recognizetheHER2 FITC), respectively. Figure 3 shows that about 77% of the transfected cell line, but not NIH-3T3 cells transfected only
1919
HER2 PROTEIN IS EXPRESSED ON LEUKEMIC BLASTS
Table 1. Expression of ~ 1 8 5 on " ~Hematopoietic ~ Cell Lines Cell Line
Erythroblastic/megakaryoblastic cell lines UT-7 TF-1 M07e MEG-01 HEL K562 SPI-801 MOLM-1 Myeloid cell lines €M2 KG-l KG-la HL60 DU.528 U937 OCI/AML-4 B-cell lines T MM BV-l73 Nalm-l Daudi Reh U266 Km3 Cess T-cell lines CML-TI HSB-2 CCRF-CEM Molt-4
Characteristics of Cell Line
~ 1 8 5 ' ~ ~ ~
Erythroblastidmegakaryoblastic leukemia leukemia Erythroleukemic Megakaryoblastic leukemia Megakaryoblastic (CML) leukemia Ph+ Erythroblastic/megakaryoblastic leukemia Erythroblastic leukemia Ph+ (CML) Erythroblastic leukemia (K562 derived) Ph' (CML) Megakaryoblastic leukemia Ph' (CML)
-
Myeloblastic leukemia leukemia Myeloblastic Immature subline leukemia Myeloblastic Myeloblastic leukemia phenotype) (mixed leukemia Myeloblastic
-
leukemia -
B-lymphoblastic leukemia (CD10-)
+ +(CD109 -
-
-
of KG-l
Myeloblastic
-
-
Ph+ (CML)
leukemia B-lymphoblastic Ph' leukemia B-lymphoblastic leukemia B-lymphoblastic Pre B-lymphobl. leukemia Myeloma Pre B-lymphobl. leukemia leukemia B-lymphobl.
Ph- (CML) (CD10')
(CML) Ph' (CML)
(CD10+) (CDlO')
leukemia T-lymphoblastic Ph' (CML) T-lymphoblastic leukemia T-lymphoblastic leukemia T-lymphoblastic leukemia
Cells were labeled with the ~ 1 8 5 ~ ~ ~ ' - r e a c t i v e M 24D2, o A b stained with a PE-conjugated antimouse antiserum, and analyzed on a FACScan flow cytometer. Abbreviation: Ph, Philadelphia chromosome.
cells in sample one and about 82% of the cells in sample two were positive for p1UHER2 (samples three and four are not shown). The fraction of leukemic blasts determined by morphologic criteria was 81% in sample one and 89% in sample two. Evaluation of four C-ALL samples resulted in the following ratios between percentage of ~ 1 8 5 ~ cells ~'" and percentage of leukemic blasts: 77%/81%, 82%/89%, 72%/77%, and 69%/73%, respectively. These data suggest that a low percentage of blasts exists that do not express p1MHER2. As shown in Fig 3, both samples contain a small fraction of cells that express CDlO and CD34, but do not coexpress plMHERZ. Because expression of these antigens is not detected at significant levels on normal PB lymphocytes (
