Abnormality of C-Kit Oncoprotein in Certain Patients with Chronic

Leukemia (2002) 16, 170–177  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Abnormality of c-kit oncoprotein in certain patients with chronic myelogenous leukemia – potential clinical signiﬁcance K Inokuchi, H Yamaguchi, M Tarusawa, M Futaki, H Hanawa, S Tanosaki and K Dan

Division of Hematology, Department of Internal Medicine, Nippon Medical School, Tokyo, Japan

Chronic myelogenous leukemia (CML) is characterized by the patients, there are two bcr/abl mRNAs for P210BCR/ABL, one Philadelphia (Ph) chromosome and bcr/abl gene rearrangement with and one without exon b3 (b3-a2 type and b2-a2 type).4 which occurs in pluripotent hematopoietic progenitor cells expressing the c-kit receptor tyrosine kinase (KIT). To elucidate In a smaller number of CML patients, there are two other types the biological properties of KIT in CML leukemogenesis, we of bcr/abl mRNAs based on the breakpoint positions of the performed analysis of alterations of the c-kit gene and func- bcr gene, ie m-bcr and ␮-bcr for P190BCR/ABL and P230BCR/ABL, tional analysis of altered KIT proteins. Gene alterations in the respectively.5,6 Extensive studies have been performed on the c-kit juxtamembrane domain of 80 CML cases were analyzed subtypes of the bcr/abl gene and their relation to the prognosis by reverse transcriptase and polymerase chain reaction-single and clinical features.7,8 The established findings regarding the strand conformation polymorphism (RT-PCR-SSCP). One case had an abnormality at codon 564 (AAT→AAG, Asn→Lys), and influence of the subtype of the bcr/abl gene (b3-a2 or b2-a2) six cases had the same base abnormality at codon 541 on the clinical characteristics have been similar for each sub- (ATG→CTG, Met→Leu) in the juxtamembrane domain. Because type, except for a higher platelet count in patients with the the change from Met to Leu at codon 541 was a conservative b3-a2 type.4,9 P190BCR/ABL in CML maybe associated with one which was also observed in the normal population and nor- monocytosis, and P230BCR/ABL maybe associated with the mal tissues of CML patients, it probably represents a polymor- chronic neutrophilic leukemia variant and marked throm- phic variation. Although samples of hair roots and leukemic 6,10 cells from the chronic phase of one CML patient showed no bocytosis. Other molecular factors which might control the abnormality, an abnormality at codon 541 (ATG→CTG, clinical features and hematological characteristics remain Met→Leu) was found only at blastic crisis (BC) of this case. In unclear. the case with the abnormality at codon 564, the mutation was Recently, the stem cell factor (SCF) c-kit signal system detected only in a sample of leukemic cells collected at BC. To (KIT/SCF) has been shown to playa crucial role in hematopo- examine the biological consequence and biological signifi- iesis.11 The c-kit receptor tyrosine kinase (KIT) is expressed on cance of these abnormalities, murine KITL540 and KITK563 expression vectors were introduced into interleukin-3 (IL-3)- progenitor stem cells as well as mast cells. SCF synergizes in dependent murine Ba/F3 cells to study their state of tyrosine vitro with other cytokines to increase the number and size of phosphorylation and their growth rate. Ba/F3 cells expressing colonies of hematopoietic progenitors.12 Thus, we speculated KITWT,KITL540 and KITK563 showed dose-dependent tyrosine that the KIT/SCF system possibly controls the hematological phosphorylation after treatment with increasing concentrations characteristics of CML. of recombinant mouse stem cell factor (rmSCF). The cells The present studywas designed to investigate the mutations expressing KITL540 and KITK563 were found to have greater tyrosine phosphorylation than cells expressing KITWT at 0.1 and 1.0 of the c-kit gene and the relationship between the in vitro ng/ml of rmSCF. The Ba/F3 cells expressing KITK563 proliferated function of mutant c-kit and the biological features of CML in response to 0.1 and 1.0 ng/ml of rmSCF as well as IL-3. The patients. Ba/F3 cells expressing KITL540 showed a relatively higher proliferative response to 0.1 ng/ml of rmSCF than the response of WT cells expressing KIT . These mutations and in vitro functional Materials and methods analyses raise the possibility that the KIT abnormalities influence the white blood cell counts (P Ͻ 0.05) and survival (P Ͻ 0.04) of CML patients. Patients Leukemia (2002) 16, 170–177. DOI: 10.1038/sj/leu/2402341 Keywords: c-kit; CML; Ba/F3; WBC We studied 116 bone marrow (BM) or peripheral blood (PB) samples obtained from 80 patients with CML in various clinical phases: 65 in chronic phase (CP), seven in accelerated Introduction phase (AP), and 44 in blastic crisis (BC). The diagnosis of CML was made on the basis of clinical features, hematological data Philadelphia (Ph) chromosome is the cytogenetic hallmark of and Ph chromosome. In 34 patients, both cytogenetic and chronic myelogenous leukemia (CML) and is found in up to molecular analyses were performed in more than two phases, 95% of CML patients.1 The demonstration of bcr/abl mRNA ie CP and BC (30 patients), CP and AP (two patients), and CP, is accepted as a reliable diagnostic marker for CML, and in AP and BC (two patients). Sixty-eight normal BM or PB some cases this evidence is more reliable than the Ph chromo- samples were obtained to studymutation and polymorphism some.2 The clinical signs and hematological findings probably of the c-kit gene. These samples were obtained with the depend partlyon the presence of P210 BCR/ABL, which plays a patients’ and normal volunteers’ informed consent. central role in the pathogenesis of the chronic phase of CML.3 According to the breakpoint site of the bcr gene in most CML Cells

Ba/F3,13 a murine IL-3-dependent pro-B lymphoid cell line, Correspondence: K Inokuchi, Division of Hematology, Department of Internal Medicine, Nippon Medical School, 1-1-5 Sendagi, Bunkyo- was cultured in RPMI 1640 medium supplemented with 10% ku, Tokyo 113, Japan; Fax: 81-3-5814–6934. fetal calf serum (FCS) and 10% WEHI-3 cell conditioned Received 19 December 2000; accepted 3 October 2001 medium. Abnormality of KIT in CML K Inokuchi et al 171 Extraction of RNA and DNA a 10% polyacrylamide gel containing 90 mM Tris-borate, pH 8.3, 4 mM EDTA, and 10% glycerol. Electrophoresis was per- The total RNA of BM or PB leukocytes was extracted with an formed at 40 W for 3 h at 18° with cooling using a water RNAzol kit (Biotex Laboratories, Houston, TX, USA), which jacket. The gel was dried on a ﬁlter paper and exposed to was based on a technique described previously.14 The total X-rayﬁlm at −80° for 1–24 h with an intensifying screen. cellular DNA was extracted from BM and PB cells byprotease K digestion, phenol–chloroform extraction and ethanol precipitation. Sequence analysis of the c-kit gene

After the RT-PCR products were separated on 2% agarose gels RT-PCR of c-kit mRNA and stained with ethidium bromide, the amplified fragment was excised from the gel, electroeluted, purified with phenol RT-PCR was performed as described elsewhere.15 The sense and precipitated with ethanol. The fragments were subcloned primers were: kit560–1, 5Ј-CTGTTCACTCCTTTGCTGAT-3Ј into the EcoRV site of the pGEM-5Zf(+/−) cloning vector.16 The (residues 1582–1601); kit560–2, 5Ј-TTCGTAATCGT transfected cells were plated on to Luria-Beriani (LB)-ampicil- AGCTGGCAT-3Ј (residues 1605–1624); kit816–1, 5Ј- lin agar plates containing 5-bromo-4-chloro-3-indolyl-␤-D- ATCATGGAGGATGACGAGTTG-3Ј (residues 2287–2306); galactoside (X-Gal), isotransferred to fresh LB-ampicillin agar and kit816–2, 5Ј-CTAGACTTAGAAGACTTGCT-3Ј (residues plates containing X-Gal and isopropylthio-␤-D-galactoside, 2310–2329). The antisense primers were: kit560–3, 5Ј- and cultured overnight for secondaryselection. White colon- CATGTGATTACCAAGGTAA-3Ј (residues 1955–1974); ies were transferred into 150 ml of LB medium containing 50 kit560–4, 5Ј-GCTCCAAGTAGATTCACAAT-3Ј (residues mg/ml ampicillin and cultured at 37°C for 4 h. The cultures 1978–1997); kit816–3, 5Ј-ATTTCAGCAGGTGCGTGTTC-3Ј were sedimented bycentrifugation, resuspended in 20 ml of (residues 2698–2717); and kit816–4, 5Ј-TTTTTA water and heated at 98°C for 10 min. After centrifugation, the GGGGATCTGCATCC-3Ј (residues 2742–2761). Complemen- supernatants were amplified byPCR using the T7 or SP6 taryDNA was synthesizedfrom 500 ng of total cellular RNA primer. Three to five clones of the three independent PCR pro- extracted from cells using 100 ng of primer kit560–4 for analy- ducts were sequenced using a Model 377 ABI sequencer with sis of the sequence in the juxtamembrane domain or kit816– dye terminators (Perkin Elmer, Warrington, UK). All sequences 4 for analysis of the phosphotransferase domain. Briefly, RT were confirmed in both orientations. A mutation was defined reaction mixture contained 32 U of avian myeloblastosis virus as when three or more clones showed the same abnormality (AMV) RT (Takara Biochemicals, Shiga, Japan) in 25 ␮lofa of the base sequence. solution containing 200 ␮mol/l each of all four dNTPs, 80 U of RNase inhibitor, 50 mmol/l Tris-HCl (pH 8.3), 75 mmol/l

KCl, 10 mmol/l dithiothreitol and 3 mmol/l MgCl2. The reac- Site-directed mutagenesis and transfection tion was allowed to proceed for 60 min at 37°C, and the reaction mixture was used as the PCR substrate. A 35-cycle PCR To examine for functional abnormalityof c-kit mutations in reaction was performed in a DNA Thermal Cycler with slight activation of c-kit tyrosine kinase activity, site-directed modification of our original protocol.4 Briefly, 25 ␮l of the RT mutagenesis was performed using murine c-kit cDNA as reaction solution was mixed with a mixture containing 250 described byFuritsu et al.17 The genes encoding murine wild- ␮mol/l of each of all four dNTPs, 100 ng of 5Ј-primer ST1, 10 type c-kit and c-kit c-DNA with a mutation at codon 814 mmol/l Tris-HCl (pH 8.3) , 50 mmol/l KCl, 3 U of Taq DNA (GAC→GTC) in the XbaI site of the expression vector, pEF- polymerase (Takara Biochemicals) and 100 ng of primers BOS, were used as wild and mutant controls.18 These two c- kit560–1 and kit560–4 or primers kit816–1 and kit816–4. The kit expression vectors were kindlyprovided byProf Y Kana- reaction conditions and cycle number were the same as those kura (Osaka University, Osaka, Japan). To generate genes con- described above. A second PCR was performed for direct taining L-540, two oligonucleotide sets of 5Ј-GCAT- sequencing. Primers kit560–2 and kit560–3 or primers kit816– TATTGTGCTGATTCTGACCTACAAAT-3Ј and 3Ј-CAGAA 2 and kit816–3 were used to amplifythe first PCR products, TCAGCACAATAATGCACATCATGCC-5Ј were synthesized, whereas primers SC3 and ASC3 were used for the 3Ј-fragment. annealed to the template and extended. To generate genes containing K-563, two oligonucleotide sets of 5Ј-GAGGAGA TAAGTGGAAACAATTATGTTTAC-3Ј and 3Ј-ATTGTTTCC Single-stranded conformation polymorphism (SSCP) ACTTATCTCCTCAACAACCTTCC A-5Ј were synthesized, gel analysis annealed to the template and extended, according to the instructions accompanying the ExSite PCR-Based Site- Detection of c-kit mutations was performed bySSCP gel Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA). Fif- analysis. The 5Ј-ends of primers (100 pmol) were labeled with teen temperature-cycles of PCR were performed to extend and ␥-32P-ATP (3000 Ci/mmol) and polynucleotide (5 U, Boehr- incorporate mutation primers byPfu Turbo DNA polymerase, inger-Mannheim; Mannheim, Germany) in 10 ␮lof50mM resulting in formation of mutated circular strands. Methylated ° Tris-HCl, pH 8.3, 10 mM MgCl2 and 5 mM DTT at 37 C for parental DNA template was digested with 10 U of DpI. The 30 min. The PCR mixture contained 10 pmol of each of the mutated double-stranded plasmid was transformed into XL1- labeled primers, 2 nmol each of the four deoxynucleotides, Blue E. coli cells to expand the plasmid after repairing the 0.1 ␮g of sample cDNA or DNA, and 0.25 U of Taq poly- nicks with T4 DNA ligase. The c-kit cDNA and expression merase in 10 ␮l of the buffer provided in the GeneAmp kit. vectors were sequenced to confirm the mutations and absence The PCR products were mixed with 10 volumes of a loading of artifact abnormalities. buffer containing 95% formamide, 20 mM EDTA, 0.05% The expression vectors containing wild-type and mutated c- bromphenol blue and 0.05% xylene cyanol, denatured at kit cDNA (10 ␮g) were transfected byelectroporation into 5 94°C for 5 min, quenched on ice and applied (1 ␮l/lane) to × 107 Ba/F3 cells with pSV2Neo (0.2 ␮g), which encodes and

Leukemia Abnormality of KIT in CML K Inokuchi et al 172 expresses the neomycin-resistant protein. The Ba/F3 cells were cultured in RPMI 1640 containing 10% FCS 600 ␮g/ml G418 (Calbiochem-Novabiochem, San Diego, CA, USA) and 10% WEHI-3 cell-conditioned medium. After selection with G418 and limiting dilution of the cells, expression of KIT was conﬁrmed by ﬂow cytometry.

Flow cytometry

To detect cell surface expression of KIT, cells were incubated with ﬂuorescein isothiocyanate (FITC)-conjugated anti-KIT (2B8 clone; PharMingen, San Diego, CA, USA) MoAb at 4°C for 30 min and analyzed on a FACScan (Becton Dickinson, Los Angeles, CA, USA).15

Immune complexkinase assay Figure 1 RT-PCR-SSCP analysis and sequencing of the cloned RT- The procedures of immunoprecipitation, gel electrophoresis PCR products of the c-kit gene of the CML cases detected abnormali- and immunoblotting were performed according to the ties. (a) and (b) show the RT-PCR-SSCP and sequence of the RT-PCR products of case 1. (c) and (d) show the RT-PCR-SSCP and sequence methods described previously.19 The cells were washed with ␮ of the RT-PCR products of case 2. (a) Lane 1: RT-PCR product of RNA PBS and lysed in 500 l of RIPA buffer (50 mM Tris-HCl, pH sample from chronic phase; lane 2: RT-PCR product of blastic crisis; 7.5, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.5% deoxychol- lane 3: RT-PCR product of hair root; lane 4: RT-PCR product of skin. ate, and 0.1% SDS with protease inhibitors (1 mM PMSF, 50 The arrow indicates an abnormal band. (b) Sequence of the cloned ␮g/ml of antipain, 5 ␮g/ml of aprotinin, and 2 ␮g/ml of RT-PCR product of blastic crisis sample. The arrow indicates a one- → leupeptin)). Cell debris was removed bycentrifugation. The base change (codon 564, AAT AAG). (c) Lane 1: RT-PCR product of RNA sample from chronic phase; lane 2: RT-PCR product of blastic supernatant was precleared byincubation with protein-G crisis; lane 3: RT-PCR product of hair root. The arrow indicates an Sepharose (Pharmacia, Uppsala, Sweden) on ice for 30 min. abnormal band. (d) Sequence of the cloned RT-PCR product of blastic The precleared lysates were incubated with anti-mouse c-kit crisis sample. The arrow indicates a one-base change (codon 541, MoAb, 2B8 (PharMingen) for 1 h on ice and with Protein- ATG→CTG). G Sepharose for 30 min on ice to collect antigen–antibody complexes. The immunoprecipitates were washed three times initiation of the culture, 0.5 ␮Ci 3H-thymidine (specific with NP-40 buffer (50 mM Tris, pH 8.0, 1.0% NP-40, 50 mM activity, 5 Ci/mmol; Amersham, Arlington Heights, IL, USA) NaCl) containing protease inhibitors. The proteins released was added to each well. Five hours later, the cells were har- from the immunoprecipitates byLaemli’s sample buffer were vested with a semi-automatic cell harvester (Pharmacia), and subsequentlyanalyzedbyelectrophoresis on 5–25% SDS- the incorporation of 3H-thymidine was measured with a liquid PAGE. The proteins were electrophoreticallytransferred from scintillation counter. the gel on to a nitrocellulose membrane. After the transfer, the filter was blocked byincubation in 1% BSA/Tween-PBS (1.37 M NaCl, 27 mM KCl, 81 mM Na2HPO4,15mM KH2PO4,1% Determination of the type of bcr/abl mRNA Tween-20, 1% BSA). Immunoblotting was performed with an antiphosphotyrosine MoAb (Py20; PharMingen) or anti-c-kit Reverse transcriptase-polymerase chain reaction analysis (RT- antibody, and detected with HRP-conjugated goat anti-mouse PCR) for determination of the type of bcr/abl mRNA was car- IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA). All ried out as described previously.4,16 Briefly, complementary immunoprecipitates were detected with HPR-conjugated goat DNA (cDNA) was prepared from 250 ng of total RNA using anti-mouse IgG and visualized byautoradiography. an antisense ABL cDNA primer. Primers bcr-1 and abl-1 were Semiquantification of the photographic signals was perfor- used for PCR for p210 mRNA, whereas primers e-1 and abl- med using an MCID image analysis system (Imaging Research, 1 were used for p190 mRNA. The PCR products were St Catherines, Ontario, Canada). The entire width of the lane electrophoresed through a 2.5% agarose gel, stained with was analyzed with appropriate background subtraction. All ethidium bromide, Southern-transferred and hybridized. bands in one photograph were analyzed together. The relative intensitywas defined as the ratio of the KIT phosphorylation signals to KITV814 signals at 15 min incubation. Table 1 Frequencyof c-kit abnormalityin CML patients and hematologicallyhealthyvolunteers

Cell proliferation assay Abnormality CML patients with Healthy volunteers abnormality/total with abnormality/ CML patients Total hematologically 3 Proliferation of cells was quantiﬁed by H-thymidine incor- (positive rate) healthy volunteers poration. Brieﬂy, the cells were washed twice with IMDM (positive rate) medium, and triplicate aliquots of cells (5 × 104) suspended in 200 ␮l of Cosmedium-001 (Cosmo Bio Co., Tokyo, Japan) Mutation at codon 564 1/80 (1.3%) 0/68 (0%) were cultured in 96-well microtiter plates for 72 h at 37°Cin Mutation at codon 541 6/80 (7.5%) 1/68 (1.5%)a the presence of various concentrations of rmIL-3 or rmSCF, provided byKirin BreweryCompany,Ltd. At 72 h after aP Ͻ 0.05 compared with 80-patient CML group.

Leukemia Abnormality of KIT in CML K Inokuchi et al 173 VP VP VP → → → VP VP HU HU HU HU + → + + + → Yes IFN + 94 Yes IFN +

BC , 17q − X 61 No IFN +

abnormality in (months)

Bcr/abl Additional Survival Giant Treatment

l) (g/dl) type chromosomal duration splenomegaly

␮ / b

Plt 10

× ; HU, hydroxyurea; VP, combined therapy with vincristine and prednisolone.

b ␣ l) (

␮ 75700 7.5 8.9 Minor t(9;14)(q34;q32) 7 No VP a Lys Leu Leu Leu Leu Leu Leu CTG BC 11200 35.8 13.7 B3a2 CTGCTG CP, BC, skin, HRCTG CP, BC, skin, HR 276800 CP, BC, skin, HR 67000 32.8 178000 37.0 10.2 30.4 B3a2 13.1 15q 12.0 B3a2 ND B3a2 None 32 75 No IFN CTG CP, BC, skin, HR 52000 34.0 9.0 Minor None 11 Yes HU CTG BC, skin AAG BC 100600 2.60 13.0 B2a2 del(3)(q21q23) 41 Yes IFN → → → → → → → → → → → → → → Asn Met Met Met Met Met Met

Amino acid detection

Codon Nucleotide/ Sample for Summaryof CML patients with a c-kit abnormality Numbers were those at diagnosis of CP. RNA samples of BC and skin were able to be analyzed. Table 2 1 M/52 564 AAT Normal range ofa neutrophil alkaline phosphataseb (NAP): 150–350. CP, chronic phase; BC, blastic crisis; HR, hair root; ND, not done; IFN, interferon- 2 M/57 541 ATG 34 M/235 F/36 541 F/45 ATG 541 541 ATG ATG 6 F/44 541 ATG CaseNo. Sex/Age (yr) Abnormality of c-kit WBC 7 M/64 541 ATG (/

Leukemia Abnormality of KIT in CML K Inokuchi et al 174 Statistical analysis during leukemic progression of case 2. In case 1, the abnormalityat codon 564 was detected onlyin an RNA sample Statistical analysis was performed using the Statview (Brain from leukemic cells at BC (Figure 1). Power, Calabashes, CA, USA) software package for the Macin- tosh personal computer. Comparisons of groups were analyzed using Fisher’s exact test for 2 × 2 tables. Values of P Ͻ c-kit receptor tyrosine kinase 0.05 were considered signiﬁcant. In order to determine the functional role of c-kit abnormalities in ligand-independent activation of c-kit products, expression Results vectors containing normal or mutated murine c-kit genes were transfected into a murine IL-3-dependent cell line, Ba/F3 cells. c-kit mutations in CML patients We used site-directed mutagenesis to construct two mutant murine c-kitR expression vectors: c-kitK563 coding for substi- EightyCML cases were screened for mutations in the coding tution of Lys for Asn at codon 563, and c-kitL540 coding for region of the c-kit gene byRT-PCR-SSCP. Expression of the c- substitution of Leu for Met at codon 540, which correspond to kit gene was detected in all 80 CML patients byRT-PCR. Lys-564 and Leu-541 of abnormal human c-kitR, respectively. Seven cases showed aberrant bands on RT-PCR-SSCP gels. These murine c-kit-expression vectors were co-transfected

Sequencing of the RT-PCR fragments which showed aberrant into the Ba/F3 murine IL-3-dependent cell line with pSV2Neo, bands on the RT-PCR-SSCP gels revealed one of the seven which contains the neomycin resistance gene, byelectropor- cases had an abnormalityat codon 564 (AAT →AAG, ation. As a negative control, the pEF-BOS vector, without the Asn→Lys) in the juxtamembrane domain, while six cases had c-kit gene, was transfected into Ba/F3 cells. After selection in an abnormalityat codon 541 (ATG →CTG, Met→Leu) (Figure a G418-containing medium for 2 weeks, surface expression 1). These two base abnormalities were observed in cDNA of KIT on the transfected cells was examined with DX2, an clones generated from three independent PCR products. Sixty- anti-mouse c-kit MoAb. Flow cytometric analysis showed that, eight normal BM aspirate samples were obtained from although transfection of Ba/F3 cells with the pEF-BOS vector hematologicallyhealthyvolunteers after obtaining informed alone resulted in no expression of KIT, Ba/F3 cells transfected consent and used as normal controls. One of the 68 healthy with pEF-BOS-KITWT, pEF-BOS-KITV814, pEF-BOS-KITK563 or normal BM samples had the same migration pattern bySSCP pEF-BOS-KITL540 showed abundant surface expression of and the same alteration as that of our CML patients at codon KITWT, KITV814, KITK563 or KITL540 on their surface, respectively 541 in which ATG was changed to CTG, resulting in a change (Figure 2a). in the encoded amino acid from Met to Leu (Table 1). The same codon 541 abnormalitywas detected in RNA samples from skin and hair roots of cases 3, 4, 5 and 6 (Table 2). State of tyrosine phosphorylation of KITK563 or KITL540 Because the amino acid substitution was a conservative one which was also observed in the normal population and nor- To examine the state of KIT-tyrosyl phosphorylation in the mal tissues of the CML patients, it probablyrepresents a poly- transfected Ba/F3 cells, the cells were deprived of serum and morphic variation. Intriguingly, however, RNA samples from growth factors for 12 h and then stimulated with 0.1, 1 or 100 the hair roots and leukemic cells at CP of case 2 were demon- ng/ml of rmSCF for 15 min. KIT was then immunoprecipitated strated to have no abnormality, whereas this case showed the and assayed by immunoblotting with either antiphosphotyros- same abnormalityat codon 541 (ATG →CTG, Met→Leu) in ine MoAb or anti-c-kit MoAb. As shown in Figure 2b, when the leukemic cells at BC. Thus, we think the alteration of wild-type or mutated c-kit genes were transfected into Ba/F3, codon 541, in which ATG was changed to CTG, occurred the c-kit gene products were synthesized in the cells as 145-

Figure 2 Flow cytometric analysis and tyrosine phosphorylation of Ba/F3Vector, BaF3- KITWT, BaF3- KITV814, BaF3- KITK563 and BaF3- KITL540. (a) Flow cytometric analysis of the surface binding of a monoclonal anti-c-kit antibody to Ba/F3 cells transfected with pEF-BOS vector alone (Vector), pEF-BOS-KITWT, pEF-BOS-KITV814, pEF-BOS-KITK563 or pEF-BOS-KITL540. Cells were incubated with either FITC-conjugated negative- control antibody (----) or 2B8 (——), washed and analyzed on a FACScan. (b) Tyrosine phosphorylation of KIT in Ba/F3 cells expressing KITWT, KITV814, KITK563 and KITL540. KIT was immunoprecipitated with an anti-c-kit MoAb (2B3) from lysates of the indicated cells before and after stimulation with rmSCF (0, 0.1, 1.0 or 100 ng/ml). The immunoprecipitates were separated bySDS-PAGE and then immunoblotted with antiphosphotyrosine (anti-P-Tyr) MoAb (second to ﬁfth rows). The immunoprecipitates of Ba/F3 cells were cultured with 0 ng/ml of SCF, divided into two aliquots, separated bySDS-PAGE, and then immunoblotted with anti-c-kit (2B3) (ﬁrst row). All immunoprecipitates were detected with HPR-conjugated goat anti-mouse IgG and visualized byautoradiography.

Leukemia Abnormality of KIT in CML K Inokuchi et al 175 and 125-kDa proteins, respectively. Immunoblotting with an KITL540 and KITK563 modulation in IL-3-dependent antiphosphotyrosine MoAb showed that increased Ba/F3 growth phosphotyrosine was observed in KITV814 regardless of rmSCF stimulation. In contrast, KITWT, KITL540 and KITK563 were dose- To determine if KITL540 and KITK563 could modulate IL-3- or dependently phosphorylated on tyrosine after treatment with SCF-dependent growth, Ba/F3 cells expressing KITWT, KITW814, increasing concentrations of rmSCF. KITL540 and KITK563 were KITL540 or KITK563 were cultured in the presence of 0 to 100 found to be more phosphorylated on tyrosine than KITWT at ng/ml of rmIL-3 or 0 to 1000 ng/ml of rmSCF for 72 h, fol- 0.1 and 1.0 ng/ml of rmSCF (Figure 2). lowed bymeasurement of cell proliferation using 3H-thymid- Repeated semi-quantiﬁcation of immunoblots (Figure 3a) ine uptake assay(Figure 4). It was demonstrated that rmIL-3 and the time course for the induction using a narrower range induced dose-dependent proliferation whereas rmSCF of concentrations between 0 and 1.0 ng/ml of rmSCF (Figure induced no proliferation in the parental Ba/F3 or pEF-BOS- 3b) successfullyshowed dose-dependent activation of KIT L540 transfected Ba/F3 cells. In addition to the proliferative WT K563 and KITK563 as well as KITWT, although KITV540 and KITK563 response to rmIL-3, Ba/F3 cells expressing KIT , KIT and L540 showed relativelyhigher efﬁciencythan KIT WT between 0.1 KIT induced dose-dependent proliferation in response to and 1.0 ng/ml of rmSCF (Figure 3). rmSCF over the range of 0.1 to 1000 ng/ml, indicating functional expression of KITWT, KITK563 and KITL540. KITK563 induced a higher proliferative response to 0, 0.1 and 1 ng/ml of rmSCF than KITWT. The proliferative response to 0.1 ng/ml of rmSCF byBa/F3 cells expressing KIT L540 was relatively higher than the response bycells expressing KIT WT. In contrast, Ba/F3 cells expressing KITV814 proliferated even in the absence of both rmIL-3 and rmSCF.

Clinical features of CML patients with c-kit abnormality

Table 2 shows the relationships between the clinical data and the molecular characteristics of the seven CML cases with a c-kit abnormality. Three kinds of bcr/abl junctions, ie the b2- a2 and b3-a2 types of major bcr breakpoint, and the minor bcr/abl type of minor bcr breakpoint, are shown in Table 2. One of the seven CML cases with c-kit gene alterations had the b2-a2 type, while four had the b3-a2 type. Intriguingly, two CML patients with a c-kit abnormalityhad the rare, minor bcr/abl type. Case 6 had extramedullary onset, and case 7 had received 5-FU medication for several years for esophageal cancer. As shown in Table 2, the platelet counts of all seven cases at diagnosis of CP were within or below the normal range, regardless of the type of bcr/abl junction. The platelet counts of CML with 564Lys-, 541Leu-KIT were marginallylower (P Ͻ 0.07, Table 3) than the counts of CML with wild-type KIT (WT-KIT). The WBC count at diagnosis of six mutated cases was markedlyhigher than the other 73 cases of CML with KITWT (P Ͻ 0.05, Table 3). The WBC count was relativelylow in case 2 with a normal c-kit gene in CP. The mean survival duration of the seven CML patients with 564Lys- or 541Leu-KIT was 45.9 months, while that of the CML patients without any Figure 3 Semiquantification of tyrosine-phosphorylation signals of KIT abnormalitywas 69.6 months (Table 3). 564 Lys-KIT and KIT in Ba/F3 cells expressing KITWT, KITV814, KITK563 and KITL540. 541Leu-KIT maybe prognostic factors ( P Ͻ 0.04), as shown in (a) Semiquantification of tyrosine-phosphorylation signals of KIT in Table 3. Ba/F3 cells expressing KITWT, KITV814, KITK563 and KITL540 in Figure 2b. The representative data from three independent experiments are shown with error bars. The relative intensities of tyrosine-phosphorylation signals of KIT in Ba/F3 cells expressing KITWT, KITK563 and Discussion KITL540 compared with those of KITV814 are given as the mean ± s.d. for three independent experiments. (b) Time course of tyrosine-phos- The interstitial cells of Cajal (ICCs) and hematopoietic stem phorylation signals of KIT in Ba/F3 cells expressing KITWT, KITV814, cells express both KIT and CD34. Gastrointestinal stromal K563 L540 KIT and KIT . Incubation times were 0, 5, 10 and 15 min. SCF tumors (GISTs) mayoriginate from the ICCs. 20 CML mayorig- concentrations were six points between 0 and 1.0 ng/ml. Relative intensities of tyrosine-phosphorylation signals of KIT in Ba/F3 cells inate from hematopoietic stem cells that are double-positive WT K563 L540 V814 for KIT and CD34. High expression of c-kit has been demon- expressing KIT , KIT and KIT compared with those of KIT + are given as the mean ± s.d. for three independent experiments. Sig- strated on CD34 cells from chronic-phase CML patients.20 nificantlyhigher values were obtained for Ba/F3 cells expressing Here, we successfullydemonstrated that the juxtamembrane KITK563 or KITL540 compared with those expressing KITWT.*P Ͻ domain mutant occurs in some CML cells which are double- Ͻ 0.05; **P 0.01. positive for KIT and CD34 as well as in ICCs. The juxtamembrane domain mutant is found in GISTs.21 The juxtamembrane

Leukemia Abnormality of KIT in CML K Inokuchi et al 176

Figure 4 Proliferation of Ba/F3 cells in response to various concentrations of rmSCF (left panel) and rmIL3 (right panel). Quadruplicate aliquots of cells expressing KITWT, KITV814, KITK563 or KITL540 were cultured with each factor, and cell proliferation was measured by 3H-thymidine incorporation assay. The results are shown as the mean ± s.d. for three separate experiments. Wild, V814, L563 and L540 mean Ba/F3 cells expressing KITWT, KITV814, KITK563 and KITL540, respectively. Vector means Ba/F3 cells transfected with the pEF-BOS vector. Signiﬁcantly higher values were obtained for Ba/F3 cells expressing KITK563 or KITL540 compared with those expressing KITWT.*P Ͻ 0.05; **P Ͻ 0.01.

Table 3 Summaryof total CML patients of the two groups The 541Leu-KIT mutation maybe one of polymorphism, since it is found even in normal healthypeople. In our CML CML with CML with patients, however, the frequencyof the 541 Leu-KIT polymor- c-kit mutation WT c-kit phism was relativelymore common compared with in normal (n = 7) (n = 73) individuals (Table 1). This higher frequencyin CML was partly due to a newlyoccurring mutation at BC, such as in case 2. Mean age ± s.d. (yr) 45.9 ± 13.6 44.6 ± 16.5 L540 Male 4 40 The Ba/F3-cell proliferation-inducing activityof KIT was WT L540 Female 3 33 the same as that of KIT . Tyrosine kinase activation of KIT WBC (×104/␮l) 12.5 ± 8.7 7.0 ± 5.8 was slightlyhigher than that of KIT WT in medium containing (n = 6)c (P Ͻ 0.05)b 0.1 ng/ml of SCF. The tyrosine kinase activation and Ba/F3- Platelet (×104/␮l) 29.07 ± 10.1 69.8 ± 64.4 cell proliferation activities of KITK563 were relativelyhigher Ͻ b (P 0.07) than those of wild-type KIT in medium containing between Overall survival (months) 45.9 ± 32.5 69.6 ± 22.4 (P Ͻ 0.04)a 0.1 and 1.0 ng/ml of SCF. Based on these in vitro data on tyrosine kinase activation and proliferation activities, KITL540 K563 WT a and KIT do not have exactlythe same function as KIT . Log rank P value. WT L540 K563 bStudent’s t-test. Unlike KIT , KIT and KIT maycause verylow levels cCase 2 having WT-KIT in CP was omitted. of spontaneous tyrosine kinase activation and cell proliferation even without SCF, as shown in Figures 3 and 4. These ﬁndings mayindicate that 541 Leu-KIT does not fall within the domain mutant is also found in mast cell leukemia.22 The tyro- categoryof polymorphism.This is because the KIT abnor- sine kinase domain mutant occurs onlyin human and murine malityof these cases was more common (8.8%) than in the mast cell leukemia, but the mechanism of tumorigenesis dif- normal healthycases (1.5%; P Ͻ 0.05). fers between the juxtamembrane domain mutant and the tyro- We speculate that these tinydifferences in in vitro function sine kinase domain mutant. The former, the juxtamembrane may inﬂuence the clinical phenotype of CML. Recently, an domain mutant, is constitutivelydimerized and activated internal tandem duplication (ITD) of the juxtamembrane (JM) without binding SCF, whereas the latter, the tyrosine kinase domain-coding sequence of the FLT3 gene in acute myeloid domain mutant, is constitutivelyactivated without forming leukemia patients was found to lead to leukocytosis and dimers.23 In GISTs, mutations are located within an 11-amino shorter survival.25 The KIT and FLT3 proteins are members of acid stretch (Lys-550 to Val-560).21 The tyrosine kinase and the class III receptor tyrosine kinase family.26 The ITD mutant proliferation of Ba/F3 cells expressing these mutated KIT pro- of FLT3 is constitutivelydimerized and autophosphorylated teins were constitutivelyactivated without SCF. 21 Another on tyrosine residues.27 The KITL540 and KITK563 abnormalities report shows two major types of deletion mutations involving of JM maybe more easilydimerized and autophosphorylated codons 550–565 and codons 566–580, stretches which were than KITWT between 0 and 1.0 ng/ml SCF. This phenomenon especiallycommon in GISTs. 22 In murine mastocytoma cells, mayresult in a tinychange in the structure of the KIT molecule the mutation is deletion of a 7-amino acid stretch (Thr-573 to which causes CML cells to undergo greater proliferation and His-579).24 Point mutations/polymorphism at codon 564 or culminates in shorter survival of CML patients. 541 is within or near these stretches involved in GIST Overall, the present analysis found that these c-kit mutations and deletion mutations of murine mastocytoma, mutations do not greatlyaffect the pathogenesis of CML, although these mutations resulted in these mutant KITs not although the presence of these mutations resulted in clinical being constitutivelyactivated. These in vitro biological differ- heterogeneityof the blood and a poorer prognosis for CML ences maybe partlydue to differences in the mutation type, patients bearing a mutation vs CML patients with normal c-kit. such as deletions of manynucleotides or one nucleotide The poor prognosis is probablydue to the higher proliferation- mutation. inducing activities of 564Lys-, 541Leu-KIT than WT-KIT between

Leukemia Abnormality of KIT in CML K Inokuchi et al 177 0 and 1.0 ng/ml SCF. The serum SCF concentration in our Sudo T, Kina H, Nakauchi H, Nishikawa S-I. Expression and func- CML patients was between 0.5 and 1.5 ng/ml (preliminary tion of c-kit in hematopoietic progenitor cells. J Exp Med 1991; data). Thus, CML blasts exposed to SCF concentrations 174: 63–71. 12 Bernstein ID, Andrew RG, Zsebo KM. Recombinant human stem between 0.5 and 1.5 ng/ml possiblyundergo greater phos- cell factor enhances formation of colonies byCD34 + and phorylation and proliferation. CD34+lin− cells, and the generation of colony-forming cell pro- Although our findings do not fullyexplain the mechanisms genyfrom CD34 +lin− cells cultured with interleukin-3, granulocyte bywhich 564 Lys-, 541Leu-KIT lead to clinical heterogeneityof colony-stimulating factor, or granulocyte–macrophage colony-sti- CML, a small functional difference from KITWT is thought to be mulating factor. Blood 1991; 77: 2316–2321. one cause of differences in the prognosis and the phenotype of 13 Palacios R, Steinmetz M. IL3-dependent mouse clone that express B-220 surface antigen, contain Ig genes in germ-line configration, leukocytosis in CML. and generate B lymphocytes in vivo. Cell 1985; 41: 727–734. 14 Tanosaki S, Inokuchi K, Shimada T, Dan K. Relation between microsatellite instabilityand N-ras mutation and duration of dis- ease free survival in patients with acute leukemia. Cancer 1998; Acknowledgements 83: 475–481. 15 Abo J, Inokuchi K, Dan K, Nomura T. p53 and N-ras mutations We thank Dr Y Kanakura for providing full-length murine c- in two new leukemia cell lines established from a patient with kit expression vectors and Kirin BreweryCompany,Ltd for multilineage CD7-positive acute leukemia. Blood 1993; 82: providing rmIL-3 and rmSCF. This work was supported by 2829–2836. 16 Inokuchi K, Fukaki M, Hanawa H, Tanosaki S, Yamaguchi H, Grants-in-Aid from the Ministryof Education, Science and Iwakiri R, Nomura T, Dan K. Heterogenous expression of bcr-abl Culture of Japan and the Takahashi Foundation. fusion mRNA in a patient with Philadelphia-chromosome-positive acute lymphoblastic leukemia. Br J Haematol 1997; 97: 837–840. 17 Furitsu T, Tsujimura T, Yono T, Ikeda H, Kitayama H, Koshimizu U, Sugahara H, Butterfield JH, Ashman LK, Kanayama Y, Matsuz- References awa Y, Kitamura Y, Kanakura Y. Identification of mutations in the coding sequence of the proto-oncogene c-kit in a human mast cell 1 Groffen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR, leukemia cell line causing ligand-dependent activation of c-kit. J Grosveld G. Philadelphia chromosomal breakpoints are clustered Clin Invest 1993; 92: 1736–1744. within a limited region, bcr, on chromosome 22. Cell 1984; 36: 18 Mizushima S, Nagata S. pEF-BOS, a powerful mammalian 93–99. expression vector. Nucleic Acids Res 1990; 18: 5322. 2 Kantarjian HM, Smith TL, McCredie KB, Keating MJ, Walters RS, 19 Inokuchi K, Miyake K, Takahashi H, Dan K, Nomura T. DCC pro- Talpaz M, Hester JP, Bligham G, Gehan E, Freireich EJ. Chronic tein expression in hematopoietic cell populations and its relation myelogenous leukemia: a multivariate analysis of the associations to leukemogenesis. J Clin Invest 1996; 97: 852–857. of patient characteristics and therapywith survival. Blood 1985; 20 Martin-Henao GA, Quiroga R, Sureda A, Garcia S. CD7 66: 1326–1335. expression on CD34+ cells from chronic myeloid leukemia in 3 Shtivelman E, Lifshiftz B, Gale RP, Canaani E. Fused transcripts of chronic phase. Am J Hematol 1999; 61: 178–186. abl and bcr genes in chronic myelogenous leukemia. Nature 1985; 21 Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, Ishiguro 315: 550–554. S, Kawano K, Hanada M, Kurata A, Takeda M, Tunio GM, Matsuz- 4 Inokuchi K, Inoue T, Tojo A, Futaki M, Miyake K, Yamada T, Tan- awa Y, Kanakura Y, Shinomura Y, Kitamura Y. Gain-of-function abe Y, Ohki I, Dan K, Ozawa K, Asano S, Nomura T. A possible mutations of c-kit in human gastrointestinal stromal tumors. correlation between the type of bcr-abl hybrid messenger RNA Science 1998; 279: 577–580. and platelet count in Philadelphia-positive chronic myelogenous 22 Tsujimura T, Furitsu T, Morimoto M, Isozaki K, Nomura S, Matsuz- leukemia. Blood 1991; 78: 3125–3127. awa Y, Kitamura Y, Kanakura Y. Ligand-independent activation of 5 Yamaguchi H, Inokuchi K, Shinohara T, Dan K. Extramedullary c-kit receptor tyrosine kinase in a murine mastocytoma cell line P- presentation of chronic myelogenous leukemia with p190 815 generated bya point mutation. Blood 1994; 83: 2619–2626. BCR/ABL transcripts. Cancer Genet Cytogenet 1998; 102: 74–77. 23 Sakurai S, Fukasawa T, Chong JM, Tanaka A, Fukayama M. C-kit 6 Yamagata T, Mitani K, Kanda Y, Yazaki Y, Hirai H. Elevated plate- gene abnormalities in gastrointestinal stromal tumors (tumors of let count features the variant type of BCR/ABL junction in chronic interstitial cells of Cajal). Jpn J Cancer Res 1999; 90: 1321–1328. myelogenous leukemia. Br J Haematol 1996; 94: 370–372. 24 Tsujimura T, Morimoto M, Hashimoto K, Moriyama Y, Kitayama 7 Mills KI, Birnie GD. Does the breakpoint within the major break- H, Matsuzawa Y, Kitamura Y, Kanakura Y. Constitutive activation point cluster region (M-bcr) influence the duration of the chronic of c-kit in FMA3 murine mastocytoma cells caused by deletion of phase in chronic myeloid leukemia? An analytical comparison of seven amino acids at the juxtamembrane domain. Blood 1996; current literature. Blood 1991; 78: 1155–1161. 87: 273–283. 8 Shtalrid M, Talpaz M, Kurzrock R, Kantarjian H, Trujillo J, Gut- 25 Kiyoi H, Naoe T, Nakano Y, Yokota S, Minami S, Miyawaki S, terman J, Galina Y, Mark B. Analysis of breakpoints within the bcr Asou N, Kuriyama K, Jinnai I, Shimazaki C, Akiyama H, Saito K, gene and their correlation with clinical course of Philadelphia- Oh H, Moyoji T, Omoto E, Saito H, Ohno R, Ueda R. Prognostic positive chronic myelogenous leukemia. Blood 1988; 72: 485– implication of FLT3 and N-RAS gene mutations in acute myeloid 490. leukemia. Blood 1999; 93: 3074–3080. 9 Faderl S, Talpaz M, Estrov Z, O’Brien S, Kurzrock R, Kantarjian 26 Matthews W, Jordan CT, Wiegand GW, Pardoll D, Lemischka IR. HM. The biologyof chronic myeloidleukemia. N Engl J Med A receptor tyrosine kinase specific to hematopoietic stem and pro- 1999; 341: 164–172. genitor cell-enriched populations. Cell 1991; 65: 1143–1152. 10 Melo JV, Myint H, Galton DAG, Goldman JM. P190 BCR-ABL 27 Hayakawa F, Towatari M, Kiyoi H, Tanimoto M, Kitamura T, Saito chronic myeloid leukemia: the missing link with chronic myelo- H, Naoe T. Tandem-duplicated Flt3 constitutivelyactivates STAT5 monocytic leukemia? Leukemia 1994; 8: 208–211. and MAP kinase and introduces autonomous cell growth in IL-3- 11 Ogawa M, Matsuzaki Y, Nishikawa S, Hayashi S-I, Kunisada T, dependent cell lines. Oncogene 2000; 19: 624–631.

Leukemia