CBFA2, Frequently Rearranged in Leukemia, Is Not Responsible for A

CBFA2, frequently rearranged in leukemia, is not responsible for a familial leukemia syndrome RD Legare1,DLu1, M Gallagher1,CHo1, X Tan1, G Barker1, K Shimizu2, M Ohki2, N Lenny3, S Hiebert3 and DG Gilliland1,4

1Department of Medicine, Division of Hematology/Oncology, Harvard Medical School, Boston, MA, USA; 2Division of Radiobiology, National Cancer Center Research Institute, Tokyo, Japan; 3Department of Tumor Cell Biology, St Jude Children’s Research Hospital, Memphis, TN; and 4Howard Hughes Medical Institute, Brigham and Women’s Hospital, Harvard Institute of Medicine, Boston, MA, USA

We have identified a family with an autosomal dominant platelet The placement of the FPD critical region on chromosome disorder with a predisposition for developing myeloid malig- 21q22 is of particular interest as this locus has been previously nancies and have previously demonstrated linkage of this trait to chromosome 21q22.1-22.2. The nearest flanking markers, implicated in AML and acute lymphoblastic leukemia (ALL) D21S1265 and D21S167, define the familial platelet disorder through the (8;21), (3;21) and (12;21) chromosomal translo- (FPD) critical region at a genetic distance of approximately 15.2 cations.4–6 As well, there is a potential association between centimorgans and physical distance of approximately 6 mega- the FPD gene and the hematologic abnormalities seen in bases. This locus is of particular interest as it has previously Down syndrome (trisomy 21), including an approximate 15- been implicated in the pathogenesis of acute myelogenous leu- fold increased risk of developing AML.7,8 kemia (AML) and acute lymphoblastic leukemia (ALL) through the (8;21), (3;21) and (12;21) chromosomal translocations. In As there is no discernible, constitutional cytogenetic abnor- each of these cases, the CBFA2 gene is rearranged. As well, mality in this pedigree, a positional cloning strategy has been there is a potential association of this locus with the hemato- undertaken to identify the mutant gene responsible for this logic abnormalities seen in Down syndrome (trisomy 21). To phenotype. Several known candidate genes map to this locus, identify the mutant gene in this pedigree, a positional cloning including CBFA2,4 IFNAR1,9 IFNAR2,10 CRFB4,11 GART,12 strategy has been undertaken. Several candidate genes map to SON,13 KCNE1,14 SCL5A315 and ATP50 (Antonarakis, per- this locus including: CBFA2, IFNAR1, IFNAR2, CRFB4, GART, SON, KCNE1, SCL5A3 and ATP50. CBFA2, as well as IFNAR1 sonal communication). We have first undertaken mutation and CRFB4, were the focus of initial mutational analysis efforts. analysis of three of these candidate genes: CBFA2, IFNAR1 In this report, we exclude CBFA2 as a candidate by Northern and CRFB4 for involvement in FPD. and Southern blotting, RNase protection, single-strand confor- CBFA2 (PEBP2␣A; AML13,4,6,16–26 encodes a subunit of the mational polymorphism (SSCP), direct sequencing and gel- heterodimeric transcription factor polyomavirus enhancer- shift analysis. Exons of the IFNAR1 and CRFB4 genes were binding protein 2 (PEBP2), also known as core binding factor also analyzed by SSCP and demonstrated no evidence of (CBF) and is involved in both DNA binding and protein– mutation. SSCP analysis identified a new polymorphism in the 16,23,25 second exon of the CRFB4 gene and confirmed a previously protein interaction. CBFA2 is the fusion partner on chro- described polymorphism in the fourth exon of IFNAR1. Efforts mosome 21 involved in the t(8;21) translocation seen in the are currently underway to delimit further the FPD critical region M2 subtype of AML,4,17,18,22 as well as as the t(3;21) translo- and to analyze the other known candidate genes, as well as cation associated with myelodysplastic syndrome, chronic novel candidate genes, which map to this locus. myelogenous leukemia in blast crisis, and therapy-related Keywords: AML; familial leukemia; CBFA2 (AML1) gene; positional 5,21,24 cloning; candidate gene; platelet disorder AML, and in t(12;21) associated with acute lymphoid malignancy.6,27 Several myeloid leukemia-associated fusion partners have been identified for CBFA2.17,19,20,21,24,26 A com- mon feature for each of these translocations is expression of Introduction the DNA binding domain of CBFA2 and disruption of the transactivating domain by the respective fusion partners. It has Analysis of the molecular basis of human tumors has provided been shown that the DNA binding domain then has dominant evidence for the multistep pathogenesis1 of many of these negative activity in precluding normal transactivation of genes malignancies, including acute myelogenous leukemia (AML). by CBFA2.28 It is therefore possible that an inherited missense However, it has been difficult to identify early mutations in or frameshift mutation which disrupts the CBFA2 transactivat- AML, due in part to the rarity of pedigrees with an inherited ing domain could lead to abnormal hematopoiesis in the pedi- predisposition to develop leukemias. In 1985, Dowton et al2 gree. This is an attractive hypothesis, because FPD is an auto- reported on a large kindred with an autosomal dominant bone somal dominant disorder in which there is one normal and marrow disorder, consisting of qualitative and quantitative one mutant allele; the phenotype could therefore be a conse- defects in platelet function and a propensity to develop AML. quence of a dominant negative mutant. Moreover, evidence Of 29 affected family members, eight have developed AML is mounting that the alternatively spliced species which con- or myelodysplastic syndrome (MDS). We have recently estab- tain the runt domain without the transactivation domain, may lished linkage of this disorder to chromosome 21q22.1-22.2.3 act in a similar fashion to decrease transactivation of the full The nearest flanking markers, D21S1265 and D21S167 define length transcript (unpublished data). the familial platelet disorder (FPD) critical region at a genetic The interferon ␣/␤ receptor (IFNAR1) gene encodes one of distance of approximately 15.2 centimorgans and physical the components of the fully functional interferon ␣/␤ receptor distance of approximately 6 megabases. and has been assigned to 21q22.1.9 IFNAR1 is a member of the cytokine/growth hormone/prolactin/interferon receptor family, and is a reasonable candidate gene based on its known Correspondence: DG Gilliland, Howard Hughes Medical Institute, function. Interferon mediates a variety of effects in hematopo- Brigham and Women’s Hospital, Harvard Institute of Medicine, 4 Blackfan Circle, Room 421, Boston, MA 02115, USA; Fax: 617 525 ietic cells through interaction with its receptor, including 5530 induction of expression of interferon regulatory factor 1 (IRF- Received 7 April 1997; accepted 13 August 1997 1), which has been implicated in pathogenesis of myelodys- CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2112 plastic syndromes.29 CRFB4 is located approximately 35 kb Single-strand conformational polymorphism (SSCP) from IFNAR1 on chromosome 21.30 Analysis of the extracellu- analysis lar domain indicates that CRFB4 is a typical member of the class II cytokine receptor family, which includes IFNAR, PCR-SSCP analysis was performed using a modification of a ␥ human tissue factor gene, and interferon- receptor gene. previously reported method,37 and using PCR conditions as Although the function and pattern of expression of the CRFB4 previously described.38 gene is unknown, it appears to be a gene duplication of IFNAR1 and by analogy may also play a role in hematopoiesis. In this report we describe the analysis of CBFA2, IFNAR1 and CRFB4 and show that no mutation can be identified in RNA-based PCR these genes using standard mutational analysis. Screening of the other known candidate genes and characterization and First-strand cDNA was synthesized as previously described,6 analysis of novel candidate genes which map to this locus using 1 ␮g of RNA and Moloney murine leukemia virus is ongoing. reverse transcriptase (GIBCO-BRL, Gaithersburg, MD, USA) according to the manufacturer’s instructions. To amplify the CBFA2 coding region, CBFA2 sense and antisense primers Materials and methods were generated from the AML1b coding region (GenBank accession U19601). Generation of lymphoblastoid cell lines and isolation of DNA and RNA

Lymphoblastoid cell lines (LBCL) were established as pre- Cloning and DNA sequencing viously described.31 DNA was extracted from lymphoblastoid cell lines according to the manufacturer’s instructions (DNA cDNA ampliﬁed by PCR was puriﬁed, cloned, miniprepped STAT 60; Tel-Test, Friendwood, TX, USA). Total RNA was and sequenced as previously described.6 PCR products gener- extracted from lymphoblastoid cells by the acid guanidinium ated from exon 4 of IFNAR1 and exon 2 of CRFB4 were simi- thiocyanate/phenol/chloroform method,32 according to the larly cloned and sequenced. manufacturer’s instructions (RNA STAT 60; Tel-Test).

Electrophoretic mobility shift assays Pulsed-field gel electrophoresis and Southern hybridization Cell extracts were prepared from 1 × 107 lymphoblastoid cells Southern blotting was performed according to standard as previously described.20 Lysates were clarified by high speed methods33 using LBCL DNA or genomic DNA obtained from centrifugation and 20 ␮g of cell lysate was used in EMSA. peripheral blood as previously described.34 Restriction endo- DNA-binding site selection and EMSA were performed as pre- nuclease digestions were performed using 10 ng of genomic viously described.20 CBFA2 antiserum was raised against a 16 DNA and BamHI, BglII, EcoRI, PvuII or HindIII according to amino acid peptide RIPVDASTSRRFTPPS and AML2 anti- the manufacturer’s instruction (NE BioLabs, Beverly, MA, serum was raised against an 18 amino acid peptide USA). Electrophoresis was performed using a CHEF II pulsed- KAQAAVGPGGRARPEVRS.28 field apparatus (Bio-Rad, Hercules, CA, USA), and DNA was transferred to Hybond-N nylon membranes (Amersham, Arlington Heights, IL, USA) and probed with an AML1b cDNA probe containing CBFA2 nucleotides 82–956 (GenBank accession number U19601), which includes both the t(8;21) Figure 1 (a) CBFA2 transcripts. Four of six major splice variants and t(3;21) breakpoint. After washing, the blot was subject for the CBFA2 gene are represented. The shaded area corresponds to autoradiography. with the runt, or DNA-binding domain. The transactivation domain is represented by vertical lines which are 3Ј of the DNA-binding domain. Riboprobes BH, HS and BB spanning the full-length CBFA2 coding sequence are noted. The locations of the t(12;21), t(8;21) and Northern hybridization t(3;21) are shown at the top of the figure. (b) Ribonuclease protection of CBFA2. RNA obtained from an affected family member (AF) and Northern blotting was performed according to standard proto- control RNA obtained from HL60, K562, an unrelated individual col33 with modifications by Fourney and colleagues (‘control’) and JF (an unaffected family member) was subjected to (Molecular Genetics and Carcinogenesis Laboratory, Cross ribonuclease protection analysis using CBFA2 riboprobes BH, HS and ␥ Cancer Institute, Edmonton, Alberta, Canada). DNA was trans- BB that span the CBFA2 gene. RNA was also hybridized to a -actin probe (lower panel) to control for quantity of RNA. Probes prior to ferred to Hybond-N nylon membranes (Amersham) and digestion with ribonuclease are shown in the lane marked ‘probes’. probed with an AML1b cDNA probe containing CBFA2 Yeast tRNA (ytRNA) served as a control for nonspecific protection of nucleotides 82–956 (GenBank accession number U19601). the probe. Probe BH gives a fully protected 513-nt fragment as well After washing, the blot was subject to autoradiography. as a 474-nt fragment and a 417-nt fragment corresponding to splice variants which may not contain exon 2 and exons 1 and 2, respectively. Probe HS gives a fully protected 443 bp fragment Ribonuclease protection analysis generated from the transcript containing exon 7b and a fully protected 314-bp fragment corresponding to the transcript containing exon 7a. ␥ Probe BB yields various protected fragments corresponding to various RNA was hybridized to CBFA2 riboprobes and to a -actin 3Ј alternative splicing. No abnormal CBFA2 transcripts could be riboprobe as an internal control of RNA integrity. Ribonu- detected in the affected family member (AF) when compared to clease protection was performed as previously described.35,36 controls. CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2113 CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2114 Table 1 CBFA2 oligonucleotide primer sequences

Exon Primer Primer sequence Fragment size (nucleotides)

1 AML1a CTGAAACAGTGACCTGTCTTG 350 AML1bR CTGAAGAGCTTCCATCTGAT 2 AML2a GTGTAAGAAGAAATGAGCTA 167 AML2R TATTTAAAAATATAACTTGG 3 AML3b AGCTGCTTGCTGAAGATCCG 331 AML3R CCTGTCCTCCCACCACCCTCT 4 AML4 ACATCCCTGATGTCTGCATT 240 AML4R CCATGAAACGTGTTTCAAGC 5 AML5 CCACCAACCTCATTCTGTTT 226 AML5R AGACATGGTCCCTGAGTATA 6 AML6b GGCATATCTCTAGCGAGTCT 286 AML6bR TGGGAAGGTGTGTGCACATG AML6 CTGATCTCTTCCCTCCCTCC AML6R AGCCGGCCAGTGGCTCCATC 7a AML7a ATTCGCGGCTCTATAAAGAA 126 AML7aR ACTTTCAGGGTAGGAGTATT 7b AML7b AATCCCACCCCACTTTACAT 244 AML7bR CTCAGCTGCAAAGAATGTGT

aCBFA2 (AML1) oligonucleotide primer sequences used for SSCP are shown. Sequences are in the 5Ј to 3Ј orientation. Primer sequences were generated from intronic sequence except for AML1a, which lies in the 5ЈUT, and AML 7aR and 7bR, which lie in the 3ЈUT. Intronic primers do not lie over splice donor or acceptor sites. SSCP for exon 6 was generated using primers AML6/6R after nesting with AML6b/bR.

Results evidence of CBFA2 gene rearrangement by pulsed-ﬁeld gel electrophoresis (data not shown). Northern blot analysis Mutation analysis of the CBFA2 gene showed no abnormalities in the size or pattern of expression of CBFA2 transcripts (data not shown). Six splice variants are We ﬁrst analyzed the CBFA2 gene in affected and unaffected known to exist from work in our laboratory in cloning the TEL- individuals by Northern and Southern blotting. There was no CBFA2 fusion transcript,6 and from work in other labora-

Table 2 IFNAR1 oligonucleotide primer sequencesa

Exon Primer Primer sequence Fragment size (nucleotides)

1 1 AACATGTAACTGGTGGGATC 292 1R AACACCACTTTCTCCTGGTT 2 2N GGGATAGAATAACATTTAGA 270 2RN CTAAAAATTATGGTGTTGTT 3 3 CATTTATACATTTGCTCACT 176 3R GGTAATATTTACTAAAGAAT 4 4 GTTCCATAGTAATTGTTTTG 226 4R CTCTGACATCGATTAAACAG 5 5 AGAACCAACTTATATTTGTG 227 5R GCCAAAATTGACAATCTGTT 6 6 CTTCTTGCCAGTTATCTCAC 220 6R ACAAACTGGGCAAAAGGAAA 7 7N ATACTATTTATTTCTGTAAC 390 7RN AATTTGTAAAGAAATAATGT 8 8 AATGTAAATCTCTGGTGCAA 296 8R AAGCTTTATTTACCCAAATC 9 9 AAAGCACATATTCCCTGATT 238 9R AAGACAATAAAAGATTCTGA 10 10 TGCTAAATATCACGTATATC 217 10R GTTCATTTCCTAGCTGTTGA 11 11 AATTGATTTCTACTCTTTCC 298 11R AACCTTATACTTGACACAGT

aIFNAR1 oligonucleotide primer sequences used for SSCP and sequencing (exon 4) are shown. Sequences are in the 5Ј to 3Ј orientation. Primers are from intronic sequence ﬂanking IFNAR1 exons except primer I 11R which lies in the 3Ј untranslated region. Primers do not overlap exon splice donor or splice acceptor sites. CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2115 Table 3 CRFB4 oligonucleotide primer sequencesa

Exon Primer Primer sequence Fragment size (nucleotides)

1 C1 TCGTGTGCTTGGAGGAAGCCGCG 130 C1R CCCGGTTCCCAAGCGCGCCCCCTG 2 C2 GAGGAGAAAAGTTGTAAATGTTTT 264 C2R CTCCTCCGGTGCGTTCCTGCCAAT 3 C3 CTCCTGATTGACCTATCTTT 209 C3R GGGAAAACAAAATGGCTTAC 4 C4 TTTGCTTGTTCTTCTGCTTA 223 C4R AAGCCTACTACCTCTGAAAT 5 C5 CACTGCTTAGTCATGTTCTT 227 C5R TGTTTAGGCATGCTTAGCGG 6 C6 ACCTGTGACAAGAATGTAAC 229 C6R TTCAAATCCACATCTCACTC 7 C7 GAATTCTCTTCCACAGCACC 251 C7R AGATGGAGTACGTCACTGTG aCRFB4 oligonucleotide primer sequences used for SSCP and sequencing (exon 2) are shown. Sequences are in the 5Ј to 3Ј orientation. Primer sequences were generated from intronic sequence except for: C1, which lies in the 5Ј untranslated region; C7, which lies in exon 6; and C7R which lies in the 3Ј untranslated region. C7 was therefore ampliﬁed from cDNA. Primers C3R and C4 lie over the splice acceptor and splice donor sites, respectively. tories39 (Figure 1a). Of these, four splicing variants have A series of riboprobes which would identify all known potential for dominant negative activity because they lack the CBFA2 splice variants were used to scan the CBFA2 coding 3Ј transactivating domain, but retain the DNA binding region (Figure 1b). Ribonuclease protection was noted to domain.39,40 These splice variants, if present in excess of the exclude variation in the relative amounts of each of the splice full length transcript, could have an activity similar to the leu- variants reported for the CBFA2 gene. Probe BH, which spans kemia-associated fusions involving CBFA2. CBFA2 nucleotide (1–513 (GenBank accession No. U19601),

Figure 2 SSCP analysis of CBFA2. Exons 1–7 of CBFA2 were amplified from genomic DNA using flanking 32P-labeled primers (Table 1) and the PCR products visualized by autoradiography after separation on MDE gels. No difference in the migration of single-stranded PCR product is appreciated between an affected (AF) and unaffected (JF) family member. Two control DNA samples (C and C1) obtained from unrelated individuals show variation in migration of amplified product from exon 1 when compared to AF and JF. Direct sequencing analysis demonstrates a polymorphism in intron 1. CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2116 gave a fully protected 513-nt fragment as well as a 474-nt fragment and a 417-nt fragment corresponding to splice variants which may not contain exon 2 and exons 1 and 2, respectively. Probe HS, which spans CBFA2 nucleotide 513– 956 (GenBank accession U19601), gave a fully protected 443- bp fragment generated from the transcript containing exon 7b and a fully protected 314-bp fragment corresponding to the transcript containing exon 7a. Probe BB, which spans CBFA2 nucleotide 865–1431 (latter BamHI site present in sequence deposited by Y Groner, GenBank accession No. X79549), yields various protected fragments corresponding to various 3Ј alternative splicing.39 No abnormal CBFA2 transcripts could be detected in the affected family member (AF) when compared to an unaffected sibling (JF) or the HL60 and K562 cell lines. There was no significant difference in the ratios of transcripts by phosphorimaging analysis except for a relative decrease in the 417-bp fragment obtained with probe BH in the affected family member, when compared to the normal sibling and cell lines. The absence of abnormal protected CBFA2 fragments in patient AF also effectively excludes a cryptic translocation involving CBFA2, as reported in t(12;21) involving TEL, a new member of the ETS family of transcription factors.6

DNA Sequencing and SSCP analysis of the full length CBFA2 cDNA

Since point mutations could also give rise to dominant inhibi- tory CBFA2 gene products, we sequenced the full length coding region of the AML1b transcript. Primers were chosen Figure 3 Electrophoretic mobility shift assay (EMSA) of CBFA2. from cDNA sequence spanning exons 1–8 and PCR products Lymphoblastoid cell (LBC) extracts obtained from an affected (AF) and cloned into Bluescript were sequenced in the sense and anti- unaffected (JF) family member were analyzed by EMSA with sense direction using the chain termination method. No point consensus CBFA2-binding site as probe showing similar binding. Both mutation, insertion or deletion was identified in sequence samples were similarly supershifted with anti-CBFA2 antibody and analysis of between 7 and 16 clones for each cDNA fragment. competed by excess exogenous CBFA2 peptide. Anti-AML2 antibody In addition to sequence analysis of cDNA, SSCP analysis of was used to demonstrate specificity of probes. Ab1, antibody directed against CBFA2; Ab2, antibody directed against AML2; pep1, CBFA2 was performed using intronic primers flanking each unlabeled, exogenous CBFA2 peptide; pep2, unlabeled exogenous exon, including splice donor and acceptor sites (41; Table 1). AML2 peptide. There was no difference in the migration of single-stranded PCR product between affected and unaffected siblings (Figure 2). flank splice donor and acceptor sites (Tables 2 and 3). There is no difference in migration of amplified products between Electrophoretic mobility shift analysis (EMSA) affected and unaffected individuals (Figures 4 and 5). SSCP of IFNAR exon 4 showed a difference in migration of control Although no abnormalities in splicing, expression or primary sample ‘C’ (Figure 4). Sequencing of amplified product dem- sequence could be detected, EMSA was performed to docu- onstrated a previously reported polymorphism in exon 4 at position 19 158 with a G substitution for C leading to a leu- ment expression of a functional CBFA2 protein. Electrophor- → etic mobility shift analysis with a consensus CBFA2-binding cine 168 valine substitution. SSCP of CRFB4 showed no differences between affected and unaffected individuals but site as a probe demonstrated similar binding of CBFA2 protein → from cell lysates generated from an affected family member identified a novel polymorphism in exon 2 with an A G and the unaffected sibling (Figure 3). CBFA2 band shifts were substitution at nucleotide 182 (GenBank accession → supershifted with anti-CBFA2 antibody and competed by No. Z17227) yielding a lysine 47 glutamate. excess exogenous CBFA2. The amount of CBFA2 DNA binding activity in this EMSA was small, consistent with the low level of expression in lymphoblastoid cell lines as demon- Discussion strated by our laboratory and others.42 AML2, which serves as a control (Figure 3), is highly expressed in lymphoblastoid There is evidence that myeloid leukemias, like other malig- cell lines. nancies, result from the acquisition of serial mutations. Evi- dence in support of the multistep pathogenesis of AML includes: (1) clonal remission from AML;43 (2) preleukemic SSCP analysis of IFNAR1 and CRFB4 syndromes, such as MDS, which progress to AML with associated acquisition of karyotype abnormalities;44 and (3) rare SSCP analysis was performed for the 11 exons of IFNAR1 and families with an inherited predisposition to develop AML.3,45 seven exons of CRFB4. Intronic primers were chosen which By establishing linkage to chromosome 21q22.1-22.2 of an CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2117

Figure 5 SSCP analysis of CRFB4. Exons 1–7 of CRFB4 were amplified from genomic DNA using flanking 32P-labeled primers (Table 3) and the PCR products visualized by autoradiography after separation on MDE gels. No difference in the migration of single- stranded PCR product is appreciated between an affected (AF) and unaffected (JF) family member. Control DNA sample ‘C2’ obtained from an unrelated individual shows a difference in migration of amplified product from exon 2 when compared to AF, JF and control Figure 4 SSCP analysis of IFNAR1. Exons 1–11 of IFNAR1 were 32 samples ‘C’ and ‘C1’, also obtained from unrelated individuals, amplified from genomic DNA using flanking P-labeled primers consistent with a polymorphism. (Table 2) and the PCR products visualized by autoradiography after separation on MDE gels. No difference in the migration of single- stranded PCR product is appreciated between an affected (AF) and unaffected (JF) family member. Control DNA sample ‘C’ obtained from are being identified as the transcript map of chromosome 21 an unrelated individual shows a difference in migration of amplified progresses. To identify the gene responsible for the phenotype product from exon 4 when compared to AF, JF and control samples in this pedigree, a positional cloning strategy has been under- ‘C1’ and ‘C2’, also obtained from unrelated individuals, consistent taken, with initial efforts focused on known candidate genes. with a previously reported polymorphism. Through conventional mutational analysis, we have demonstrated that no mutation in three candidate genes CBFA2, IFNAR1 and CRFB4 can be identified in our pedigree. autosomal dominant disorder of platelet function and number, The data presented are convincing that CBFA2 is not and a striking propensity to develop AML, we have a unique mutated in affected family members. No gross structural opportunity to identify and characterize a gene which may be rearrangement was detected on Southern analysis and no involved in the very earliest events in the multistep pathogen- abnormal patterns of expression or altered transcript size were esis of hematologic malignancy. The physical distance of detected on Northern analysis between affected and unaffec- approximately 6 mB which spans the familial platelet disorder ted individuals. Ribonuclease protection assays show no dif- critical region is bound proximally by marker D21S1265 and ferences in splicing of CBFA2 between affected and unaffec- distally by D21S167. Several known candidate genes map to ted individuals and exclude a cryptic translocation affecting this area including CBFA2,4 IFNAR1,9 IFNAR2,10 CRFB4,11 CBFA2.6 CBFA2 has no point mutations as assessed by SSCP GART,12 SON,13 KCNE1,14 SCL5A315 and ATP50 analysis of genomic DNA followed by sequencing of the full (Antonarakis, personal communication). As well, novel genes length cDNA. Finally, gel shift analysis documented CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2118 expression of functional CBFA2 from affected and unaffected contribute to interferon-alpha/beta responsiveness. Somatic Cell individuals, with no difference in protein binding to CBFA2- Mol Genet 1995; 21: 139–145. specific sequence, supershifting with anti-CBFA2 antibody or 11 Lutfalla G, Gardiner K, Uze G. A new member of the cytokine receptor gene family maps on chromosome 21 at less than 35 kb competition with exogenous CBFA2. We also found no evi- from IFNAR. Genomics 1993; 16: 366–373. dence of mutation in the IFNAR1 or CRFB4 genes, when stud- 12 Henikoff S, Eghtedarzadeh M. Conserved arrangement of nested ied by SSCP analysis. As the sensitivity of SSCP ranges from genes at the Drosophila Gart locus. Genetics 1987; 117: 711–725. 70 to 98%36,46–48 and is dependent on nucleotide sequence, 13 Khan I, Fisher R, Johnson K, Bailey M, Siciliano M, Kessling A, fragment size and electrophoretic conditions, it is possible that Farrer M, Carritt B, Kamalati T, Buluwela L. The SON gene enco- a mutation in IFNAR1 or CRFB4 could have been missed des a conserved DNA binding protein mapping to human chromo- using this technique. Efforts are currently underway to analyze some 21. Ann Hum Genet 1994; 58: 25–34. 14 Chevillard C, Attali B, Lesage F, Fontes M, Garhanin K, Lazdunski the other known candidate genes as well as to characterize M, Mattei M. Localization of a potassium channel gene (KCNE1) and analyze novel genes which map to this locus. to 21q22.1-22.2 by in situ hybridization and somatic cell hybridiz- Identification and characterization of the gene responsible ation. Genomics 1993; 15: 243–245. for causing this disorder may provide further insight into the 15 Berry G, Mallee J, Kwon H, Rim J, Mulla W, Muenke M, Spinner molecular basis of platelet production and function, as well N. The human osmoregulatory Na+/myo-inositol contransporter as leukemogenesis. The mutant gene responsible for the trait gene (SLC5A3): molecular cloning and localization to chromo- in this pedigree may also be involved in other familial leuke- some 21. Genomics 1995; 25: 507–513. 16 Bae S, Yamaguchi-Iwai Y, Ogawa E, Maruyama M, Inuzuka M, mia syndromes, leukemia and hematologic abnormalities Kagoshima H, Shigesada K, Satake M, Ito Y. Isolation of PEBP2␣B associated with Down Syndrome, and in the majority of leuke- cDNA representing the mouse homolog of human acute myeloid mias in which there is no apparent inherited predisposition. leukemia gene, CBFA2. Oncogene 1993; 8: 809–814. 17 Downing J, Head D, Brint-Curcio A, Hulshof M, Motroni T, Carroll A, Drabkin H, Willman C, Theil K, Erickson P. An AML1/ETO fusion transcript is consistently detected by RNA-based polymerase chain reaction in acute myelogenous leukemia containing Acknowledgements the (8;21)(q22;q22) translocation. Blood 1993; 81: 2860–2865. 18 Erickson P, Gao J, Chang K-S, Look T, Whisenant E, Raimondi S, We thank Y Groner, H Miyoshi and N Sacchi for providing Lasher R, Trujillo J, Rowley J. Drabkin H. Identification of break- CBFA2 sequence data used to generate primers and H Miyoshi points in t(8;21) acute myelogenous leukemia and isolation of a for providing a cDNA probe. This study was supported in part fusion transcript, AML1/ETO, with similarity to Drosophila seg- mentation gene, runt. Blood 1992; 80; 1825–1831. by American Cancer Society Grant DHP-48. DGG is the 19 Kozu T, Miyoshi H, Shimizu K, Maseki N, Kaneko Y, Asou H, Stephen Birnbaum Scholar of the Leukemia Society of Kamada N, Ohnki M. Junctions of the AML1/MTG8(ETO) fusion America and an Assistant Investigator of the Howard Hughes are constant in t(8;21) acute myeloid leukemia detected by reverse Medical Institute. transcription polymerase chain reaction. Blood 1993; 82: 1270– 1276. 20 Meyers S, Downing J, Hiebert S. Identification of AML-1 and the (8;21) translocation protein (AML-1/ETO) as sequence-specific DNA-binding proteins: the runt homology domain is required for References DNA binding and protein–protein interactions. Mol Cell Biol 1993; 13: 6336–6345. 1 Weinberg RA. Oncogenes, antioncogenes, and the molecular 21 Mitani K, Ogawa S, Tanaka T, Miyoshi H, Kurokawa M, Kiroyuki basis of multistep carcinogenesis. Cancer Res 1989; 49: 3713– M, Yazaki Y, Ohki M, Hirai H. Generation of the AML1EVI-1 3721. fusion gene in the t(3;21)(q26;q22) causes blastic crisis in chronic 2 Dowton S, Beardsley D, Jamison D, Blattner S, Li F. Studies of a myelocytic leukemia. EMBO J 1994; 13: 504–510. familial platelet disorder. Blood 1985; 65: 557–563. 22 Miyoshi H, Kozu T, Shimizu K, Enomoti K, Maseki N, Kaneko Y, 3 Ho C, Otterud B, Legare RD, Varvil T, Saxena R, DeHart D, Kohler Kamada N, Ohki M. The t(8;21) translocation in acute myeloid S, Aster J, Dowton SB, Li FP, Li F, Leppert M, Gilliland DG. Link- leukemia results in production of an AML1-MTG8 fusion tran- age of a familial platelet disorder with a propensity to develop script. EMBO J 1993; 12: 2715–2721. acute myeloid leukemia to human chromosome 21q21-22. Blood 23 Nuchprayoon I, Meyers S, Scott L, Suzow J, Hiebert S, Friedman 1996; 87: 5218–5224. A. PEBP2/CBF, the murine homolog of the human myeloid AML1 4 Miyoshi H, Shimizu K, Kozu T, Maseki N, Kaneko Y, Ohki M. and PEBP2B/CBFB proto-oncoproteins, regulates the murine mye- t(8;21) breakpoints on chromosome 21 in acute myeloid leukemia loperoxidase and neutrophil elastase genes in immature myeloid are clustered within a limited region of a single gene, AML1. Proc cells. Mol Cel Biol 1994; 14: 5558–5568. Natl Acad Sci USA 1991; 88: 10431–10434. 24 Nucifora G, Larson R, Rowley J. Persistence of 8:21 translocation 5 Nucifora G, Begy C, Erickson P, Drabkin H, Rowley J. The 3;21 in patients with acute myelogenous leukemia type M2 in long- translocation in myelodysplasia results in a fusion transcript term remission. Blood 1993; 82: 712–715. between the AML1 gene and the gene for EAP, a highly conserved 25 Ogawa E, Maruyama M, Kagoshima H, Inuzuka M, Lu J, Satake M, protein associated with the Epstein–Barr virus small RNA EBER 1. Proc Natl Acad Sci USA 1993; 90: 7784–7788. Shigesada K, Ito Y. PEBP2/PEA2 represents a family of transcription 6 Golub T, Barker G, Bohlander S, Hiebert S, Ward D, Bray-Ward fctors homologous to the products of the Drosophila runt gene P, Morgan E, Raimonke S, Rowley J, Gilliland D. Fusion of the and the human CBFA2 gene. Proc Natl Acad Sci USA 1993; 90: TEL gene on 12p13 to the AML1 gene on 21q22 in acute lym- 6959–6863. phoblastic leukemia. Proc Natl Acad Sci USA 1995; 92: 4917– 26 Nucifora G, Rowley J. AML1 and the 8;21 and 3;21 translocations 4921. in acute and chronic myeloid leukemia. Blood 1995; 86: 1–14. 7 Ganick DJ. Hematological abnormalities in Down syndrome. Ped- 27 Romana S, Le Coniat M, Berger R. t(12;21): a new recurrent trans- iatr Hematol Oncol Forum 1994; 1: 1–8. location in acute lymphoblastic leukemia. Genes Chromosom 8 Zipursky A, Poon A, Doyle J. Leukemia in Down syndrome: a Cancer 1994; 9: 186–191. review. Pediatr Hematol Oncol 1992; 9: 139–149. 28 Meyers S, Lenny N, Sun W, Hiebert S. AML2 is a potential target 9 Lutfalla G, Gardiner K, Proudhan D, Vielh E, Uze G. The structure for transcriptional regulation by the t(8;21) and t(12;21) fusion pro- of the human interferon alpha/beta receptor gene. J Biol Chem teins in acute leukemia. Oncogene 1996; 13: 303–312. 1992; 267: 2802–2809. 29 Willman C, Sever C, Pallavicini M, Harada H, Tanaka N, Slovak 10 Raz R, Cheung K, Levy D. Three distinct loci on chromosome 21 M, Yamamoto H, Harada K, Meeker T, List A. Deletion of IRF-1, CBFA2 is not responsible for a familial leukemia syndrome RD Legare et al 2119 mapping to chromosome 5q31.1, in human leukemia and preleu- 39 Levanon D, Bernstein Y, Negreanu V, Ghozi M, Bar-Am I, Aloya kemic myelodysplasia. Science 1993; 259: 968–971. R, Goldenberg D, Lotem J, Groner Y. A large variety of alterna- 30 Cheng S, Lutfalla G, Uze G, Chumakou I, Gardiner K. GART, tively spliced and differentially expressed mRNAs are encoded by SON, IFNAR, and CRF2-4 genes cluster on human chromosome the human acute myeloid leukemia gene AML1. DNA Cell Biol 21 and mouse chromosome 16. Mammalian Genome 1993; 4: 1996; 15: 175–185. 338–342. 40 Meyers S, Lenny N, Hiebert S. The t(8;21) fusion protein interferes 31 Neitzel H. A routine method for the establishment of permanent with AML-1B-dependent transcriptional activation. Mol Cell Biol growing lymphoblastoid cell lines. Hum Genet 1986; 73: 320– 1995; 15: 1974–1982. 326. 41 Miyoshi H, Miki O, Kimiko O, Mitani K, Hirai H, Imai T, Yoko- 32 Chromczynski P, Sacchi N. Single-step method of RNA isolation yama K, Soeda E, Ohki M. Alternative splicing and genomic struc- by acid guanidinium thiocyanate phenol-chloroform extraction. ture of the AML1 gene involved in acute myelogenous leukemia. Ann Biochem 1987; 162: 156–159. Nucleic Acids Res 1995; 23: 2762–2769. 33 Sambrook J, Fritsch E, Maniatis T. Molecular Cloning: A Labora- 42 Levanon D, Negreanu V, Bernstein Y, Bar-Am I, Avivi L, Groner tory Manual, 2nd edn. Cold Spring Harbor Laboratory Press: Cold Y. AML1, AML2, and AML3, the human members of the runt Spring Harbor, NY, 1989. domain gene-family; cDNA structure, expression, and chromo- 34 Bell G, Karam J, Rutter W. Polymorphic DNA region adjacent to somal localization. Genomics 1994; 23: 425–432. the 5Ј end of the human insulin gene. Proc Natl Acad Sci USA 43 Busque L, Gilliland DG. Clonal evolution in acute myeloid leuke- 1981; 78: 5759–5763. mia. Blood 1993; 82: 337–342. 35 Barker GF, Beemon K. Nonsense codons in the Rous sarcoma 44 Legare R, Gilliland D. Myelodysplastic syndrome. Curr Opin virus gag gene decrease the stability of unspliced RNA. Mol Cell Hematol 1995; 2: 283–292. Biol 1991; 11: 60–68. 45 Horwitz M. The genetics of familial leukemia. Leukemia 1997; 36 Golub T, Barker G, Lovett M, Gilliland D. Fusion of PDGF recep- 11: 1347–1359. tor B to a novel ets-like gene, tel, in chronic myelomonocytic leu- 46 Ravnik-Glavac M, Glavac D, Dean M. Sensitivity of single-strand kemia with t(5;12) chromosomal translocation. Cell 1994; 77: conformation polymorphism and heteroduplex method for 307–316. mutation detection in the cystic fibrosis gene. Hum Mol Genet 37 Orita M, Iwahana H, Kanazawa H, Hayaski K, Sekiya T. Detection 1994; 3: 801–807. of polymorphisms of human DNA by gel electrophoresis and 47 Kneppers ALJ, Deutz-Terlouw PP, den Dunnen JT, van Ommen single-strand conformation polymorphisms. Proc Natl Acad Sci GJB, Bakker E. Point mutation screening for 16 exons of the dystro- USA 1989; 86: 2766–2770. phin gene by multiplex single-strand conformation analysis. Hum 38 Stegmaier K, Takeuchi S, Golub TR, Bohlander SK, Bartram CR, Mutat 1995; 5: 235–242. Koeffler, PH, Gilliland DG. Mutational analysis of the candidate 48 Sarkar G, Yoon H-S, Sommer S. Screening for mutations by RNA tumor suppressor gene Tel and Kip1 in childhood acute lympho- single-strand conformation polymorphism (rSSCP): comparison blastic leukemia. Cancer Res 1996; 56: 1413–1417. with DNA-SSCP. Nucleic Acids Res 1992; 20: 871–874.