course of chronic myelogenous leukemia - Europe PMC

Proc. Natl. Acad. Sci. USA

Vol. 91, pp. 10722-10726, October 1994 Medical Sciences

Progressive de novo DNA methylation at the bcr-abl locus in the course of chronic myelogenous leukemia (do wove promoter

a /tumor Inp

)

MICHAL ZION*, DINA BEN-YEHUDAt, AYELET AvRAHAM*, OFER COHEN*, MEIR WETZLER*, DANIELLE MELLOUL*, AND YINON BEN-NERIAH*§ *Lautenberg Center for General and Tumor Immunology, The Hebrew University-Hadassah Medical School, P.O. Box 12272, Jerusalem 91120, Israel;

tDepartment of Hematology, Hadassah Medical Center, Jerusalem 91120, Israel; and tDepartment of Medical Oncology, M.D. Anderson Cancer

Center, 1515 Holcombe Boulevard, Houston, TX 77030

Communicated by Laszlo Lorand, June 6, 1994 (received for review January 24, 1994)

ABSTRACT De novo methylation of CpG Islands is a rare event in mammalian cells. It has been observed In the course of developmental processes, such as X chromosome inactivation and genomic imprinting. The methylation of DNA, an Important factor in the epigenetic control of gene expression, may also be involved In tumorigenesis. After the t(9;22) chromosomal translocation and generation of the Philadelphia chromosome, the Initiating event In chronic myelogenousleemia (CML), most of the abl coding sequence is fused to the 5' region of the bcr gene. Expression of the hybrid bcr-abl gene is, therefore, regulated by the bcr promoter. In most cases of CML, one of the two abl promoters (Pa) is nested within the bcr'-bl transcrptional unit and should be able to tr be the type Ia 6-kb normal abi mRNA from the Philadelphia chromosome. However, we have found that the 6-kb tapt is present only in CML cell lines coning a normal abl allele and that the apparent inactivation of the nested Pa promoter is associated with allele-specfic methylation. Furthermore, we have noticed that the Pa promoter is contaied within a CpG island and undergoes progressive de nova methylation in the course of the disease. This ls atts to by the fact that DNA samples from CML patients that are methylation-free at the time of diagnosis invariably become methylated in advanced CML. Since tumor progression in CML caunot always be inferred from the clinical presentation, a nt of de novo CpG methylation may prove to be of critical value in management of the disease. It could herald blastic transformation at a stage when bone marrow mnsplantation, the only potentially curative therapeutic procedure in CML, is still effective.

on an individual basis and there is no evidence that conventional antileukemic therapy at the chronic stage alters the risk of transformation, survival is determined mainly by the intrinsic biology of the disease (6). In the past decade, there has been definitive evidence that many patients in the chronic phase of the disease can be cured by bone marrow transplantation (7). The question is not whether to perform a marrow transplant but the optimal timing. The related mortality within the first 2 years of bone marrow transplantation is =25% and the morbidity rate, including severe graft versus host disease, is even higher (8). Therefore, if one could accurately define a cohort of patients likely to remain in the chronic phase for several years, it would be reasonable to delay the procedure as long as possible. Such a definition cannot be based on clinical presentation or on the results of any currently used laboratory test. It is, however, conceivable that the progression of the disease is the consequence of genetic alterations that must accumulate prior to changes in hematopoietic stem cell behavior. A great deal of effort has been invested in identifying these genetic alterations and, indeed, certain mutations and chromosomal aberrations are found to be associated with the blastic transformation (9-11). Nevertheless, there is thus far no distinct molecular finding commonly related to tumor progression in CML that precedes and can, therefore, predict the blastic transformation. In searching for such an event, we noticed that part of the bcr-abl gene undergoes de novo methylation in the course of CML. Samples of DNA that are methylation-free at the time of diagnosis invariably become methylated before the blast crisis. It is thus possible that this molecular finding could serve as a useful marker for CML progression and predict the onset of blast crisis.

Chronic myelogenous leukemia (CML) is a fatal clonal stemcell disorder that accounts for -25% of all leukemia cases. It results from a reciprocal translocation, t(9;22), cytogenetically detectable by the presence of the Philadelphia chromosome (Ph'). At the molecular level, the protooncogene abl located on chromosome 9 is fused to the bcr gene on chromosome 22, resulting in the formation of the bcr-abl hybrid gene (1). The latter, which is under the transcriptional control of the bcr promoter, gives rise to an 8-kb mRNA (2) that is translated into a 210-kDa fusion protein (3), implicated in the pathogenesis of the chronic leukemia (4). Clinically, CML is characterized by a triphasic course: chronic and accelerated stages and blast crisis, exemplifying its malignant evolution. In the late phases of the disease, a myeloid or lymphoid progenitor, a descendent of the originally affected stem cell, loses its capacity for terminal differentiation (5). Sequential transformation from the chronic to the accelerated phase and blastic crisis appears to occur at random (5). Since the onset ofthe disease cannot be predicted

MATERIALS AND METHODS Cell Lines and Clinical Samples. BV173 (12), KBM5 (13), EM2 (14), and K562 (15) are Ph'-positive CML cell lines; ALL1 (16) is a Ph'-positive acute lymphoblastic leukemia cell line. Disease stage was determined by the criteria of the Bone Marrow Transplant Registry for Classifying the Phases of CML (17). PCR Analysis. A 400-ng quantity of genomic DNA extracted from cell lines or bone marrow was digested with the indicated restriction enzymes for 6 h in 20-pI aliquots. To ensure complete cleavage, an additional 5 units of the enzyme was added and digestion was allowed to proceed for an additional 16 h. PCR amplification of the reaction products was performed as follows: 10 A1 containing 200 ng of the cleaved genomic DNA was added to 50 t1 of PCR mixture

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: Ph', Philadelphia chromosome; CML, chronic myelogenous leukemia.

§To whom reprint requests and correspondence should be addressed.

10722

Medical Sciences: Zion et al. containing all four dNTPs (each at 500 FM), 67 mM Tris*HCl (pH 8.8), 3 mM MgCl2, 16.6 mM (NH4)2SO4, 10 mM 2-mercaptoethanol, 10o (vol/vol) dimethyl sulfoxide, and 10 pmol of Ia3 primer (5'-GGAAGCTTGATAACAGCTGGAGGAC3'). The reaction mixture was preheated for 10 min at 940C and then 10 pmol of CG20 primer (5'-CGGGGGGGCGCGCGGGCCGA-3') and 2.5 units of Taq DNA polymerase (Boehringer Mannheim) were added at 85TC. This was followed by 28 cycles of 1 min at 94°(, 1 min at 59°C, and 1 min at 72TC, terminated by 10 min at 720C in a Perkin-Elmer/ Cetus thermocycler. A 10-jd aliquot of the amplified product was separated by electrophoresis through 4% NuSieve agarose/0.5% agarose gel, transferred to a nylon membrane (Hybond-N, Amersham), and prehybridized for 2 h at 42°C. The hybridization solution contained 20% (vol/vol) formamide, 5x Denhardt's solution, 6x SSC, 0.1% SDS, 0.05% PP1, and tRNA (25 pg/ml). The Ia5 internal oligomer (5'GGTCTAGACAAAATGTTGGAGATCTG-3') was 32P-endlabeled and used as a probe. After a 2-h hybridization at 42°C, the membrane was washed at room temperature for 10 min with 6x SSC and autoradiographed.

RESULTS Expression of the Normal abl 6-kb mRNA Is Absent in CML Cell Lines Lacking an Intact Chromosome 9. The protooncogene abl is expressed in all normal human cells. The gene has two first exons, Ia and Tb, which are differentially transcribed from their own promoters. The proximal promoter, Pa, and a distal one, Pb, are 175 kb apart and direct the synthesis of two mRNA species of 6 and 7 kb, respectively (18). In -90% of Ph' translocations, the proximal promoter (Pa) is nested within the bcr-abl transcriptional unit (19). The location of Pa, with respect to the bcr promoter, is often analogous to that in the native gene, where it is nested within the Pb transcriptional unit (19, 20). In normal cells, both abl promoters are active in gene expression; the activity of the Pa downstream promoter appears to be unaffected by the upstream Pb promoter (18, 21, 22). To date, however, the transcriptional activity of Pa within the bcr-abl gene has not been investigated. Certain CML cell lines established from CML patients in blast crisis carry multiple Ph' chromosomes but have no intact chromosome 9. In our experiments, the only mRNA species found to hybridize with the abl probe was the 8-kb bcr-abl transcript common to all CML cells (Fig. 1A). In K562, a CML cell line with multiple Ph' chromosomes in addition to an intact chromosome 9 (15), the normal abl 6-kb and 7-kb mRNAs and the 8-kb bcr-abl transcript were observed (Fig. 1A). The 6-kb band is also abundant in ALL1, a Ph'-positive acute lymphoblastic leukemia cell line that contains a normal copy of chromosome 9. The 7-kb band of ALL1 is largely the product of the bcr-abl hybrid gene (data not shown). An RNase protection assay localized the Pa transcriptional start point to nt -72, corresponding-to the ATG codon. A specific 150-bp protected fragment (Fig. 1B) was detected using three probes of different sizes (data not shown). The 240-bp protected band (Fig. 1B) was detected only with one probe and is, therefore, nonspecific. The RNase protection assay confirmed the observation that the BV173, EM2, and KBM5 cell lines did not express the 6-kb mRNA initiated from the main Pa transcriptional start point, whereas it was expressed in the K562 and ALL1 Ph'-positive cell lines (Fig. 1B). For RNA quality control, samples were hybridized with an abl lb exon RNA probe, the expression of which is not affected by Ph' translocation (data not shown). Hence, it appears that the translocation event compromises Pa transcription from the Ph' chromosome but spares the transcriptional activity of abl in the unaffected chromosome.

Proc. Nadl. Acad. Sci. USA 91 (1994) , K.5 D)audi I IeLa

8kb 7kb 6kb

_* U, soL

-

Ph B

1\ 1-3

I

10723

\1 rW\5I .! iL

*memaw4m

lo

O 2 ix4 0 0 x4 / 0) u 2 HeLa Daudi Raji H160 K!562 BY KBMS EM12 ALLI rRF tRNA Probe

527 -o 404

-1

309 - I

242 238 -

NA

Ih

IWi

2 7--

201-1. -~ 190 I g o -I 180 160 -

147

j "040 MO

Uim

123 -

,1i0

FIG. 1. Expression of the normal abl 6-kb mRNA. (A) Northern blot analysis of CML and control cell lines. Poly(A)+ RNA from Daudi, HeLa, K562, EM2, KBM5, and ALL1 cell lines was electrophoresed through a formaldehyde/agarose gel, transferred to a nitrocellulose filter, and hybridized with a 32P-labeled cDNA frgment, P18, corresponding to the abI exons (2). The three abl-related mRNAs are indicated in the figure. The 7-kb ALLI band corresponds to the acute lymphoblastic leukemia-specific bcr-abl fusion transcript (16). Note the additional minor bcr-ablof 7.7 kb in KBM5 (13). (B) RNase protection analysis. A 0.64-kb Nco I-BamHI genomic abI Ia template including exon Ia and its 5' flankin region was used for the in vitro synthesis of a 32P-labeled antisense RNA probe. The probe was annealed to 3-5 pg of poly(A)+ RNA derived from K562, BV173, KBM5, and EM2, CML cell lines; ALL1, a Ph'-positive ALL cell line; and Daudi, Raji, HL60, TRF, and HeLa cells, used as control cell lines. Yeast tRNA (5 pg) was used as a negative control. The Ia-protected fragment of 150 bp is indicated by an arrowhead. The number of Ph' chromosomes in each cell line is indicated at the top. A 32P-labeled pBR322 Msp I digest was used as a size marker.

The Pa Promoter Is Nested Witin a CpG Island That Is

Methylated in CML Cell Lhies. The loss of Pa transcriptional activity in the various CML cell lines led us to study the Pa promoter region. The genomic DNA upstream of the Pa exon is composed of 85% CG (Fig. 2B). In vertebrate genomes, the dinucleotide sequence CpG is relatively rare. It is often clustered in a discrete region of %1 kb, mostly surrounding promoters, a sequence known as a CpG island (23). While the

10724

Proc. Nad. Acad. Sci. USA 91

Medical Sciences: Zion et al.

5tJ JO=3

3'

.i~~~~~~Ij r b ia;ln8 i. _._ .!. _._

.

T~~~~~~~~~ .* _ _

rid D

(1994)

/ PiZ

oSma I ,'

Sac'h0

Sac

i-

II=I

Bna,mt I II

BglII

7I IIPHUI

L

_ Island_ _CpG

_

bp

L.

B

khhL

PhII

I

IN

B

Hpan Hpafl CG20 MspI MSpI SacU 5'CGGGGGGGCGCGCGGGCCGAGCCGGGCCrGAGCCGGGCCCGCGGACCGAGCrGG HpalI UaS BgI MspI GAGAGGCrCCGGCCCCCGACGTCGTGGCGCGGGAAAATGTrGGAGATCrGCCTGA

------.

_._, j. I_

,.

iT

',i

KBM

II

Ix

x4

HS

-LN

I

IfS

H

| N HS I HS

1

I-N HS

11

1

IN ITS

EI 23

23.0

-

9.4

-

6.6

-

i.

".., t'. 1,

4

';

4.3

AGCrGGTGGGCrGCAAATCCAAGAAGGGG

GcGTCCTCCAGCrGITATCrGG

3'

2 -3 ..p

2.3

-

2.0-

FIG. 2. Characterization of the Pa promoter. (A) Schematic organization of the abl gene and restriction map of the CpG island containing the Ia exon and its 5' flanking region. Boxes, exons; arrows, transcriptional start sites. (B) Sequence of the Ia exon and its 5' flanking region. The arrows denote oligonucleotide length and directionality. The starting ATG is marked with an arrow and the restriction sites are overlined.

observed-to-expected ratio of CpG throughout most of the genome is is

~

A

Accelerated Phaqe r

Aqpp W? W ,4

\\1~ t

i__

o~o

B1

_ c

c

40

OfW

"

FIG. 4. Progressive Pa methylation in CML patients. (A) Methylation state of the Pa promoter in the accelerated phase. Genomic DNA was extracted from the bone marrow of three CML patients in the accelerated phase. (B) Methylation state of the Pa promoter in chronic and blast crisis phases. Genomic DNA was extracted from cells of five patients (rows A-E) in the chronic phase and in blast crisis.

10726

Medical Sciences: Zion et al.

tion of the abl gene would result in 50% loss of the relevant mRNA. The situation is more complex, however, since in most cases the translocation leaves one transcriptional abl unit (Pa) within the Ph' chromosome and only the Pb unit is compromised. Surprisingly enough, we have found that in CML cell lines lacking the normal abl alleles and retaining only the Ph' alleles, there is no evidence of either Pb or Pa abl transcription. By assuming the transcription defect is specific to the Ph' allele, it is not known whether the lack of Pa transcription is an immediate or a late consequence of the translocation. Immediate inactivation of the Pa transcription unit after translocation would result in an abrupt decline in abl transcription and could, therefore, be correlated with the onset of the disease. Later blockage of abl transcription might have an impact on disease progression. Overexpression of abl retards cell cycle progression (32, 33), and interference with its cellular function through a transdominant abl mutant promotes oncogene transformation (33). Therefore, abl may act as a growth suppressor gene or an antioncogene. It is conceivable that even a modest decline in normal abl expression could affect the ratio of oncogene (bcr-abl) to antioncogene (abi) expression and further promote tumorigenesis. Indeed, the ratio ofthe transforming N-ras mutant to the N-ras protooncogene has been found to affect tumorigenesis in neuroblastoma cell lines (34). A recent observation concerning the origin of the t(9;22) translocation hints that genomic imprinting may underlie the pathogenesis of CML (35). The translocation always occurs between a maternal bcr allele and a paternal abl allele, but it is not known whether the choice of partner alleles is made prior to the translocation or if it stems from the selective advantage enjoyed by certain random combinations. Genomic imprinting could play a role in selecting the partner alleles for translocation and would account for the compromised abl Pa transcription observed in CML cells. To our knowledge, there is no evidence for the direct involvement of either bcr or abl in genomic imprinting. Human Ph'-negative cells appear to express both bcr alleles (36) and gene targeting experiments in mice have failed to reveal genomic imprinting in abl +/- heterozygous mice (37, 38). Therefore, the choice of translocation partners may be due to the imprinting of other genes on chromosomes 9 and 22. On the other hand, masked genomic imprinting of abl may become apparent only after chromosomal translocation. Also, progressive Pa methylation may be a consequence of the translocationreactivated imprinting, similar to the delayed methylation following X chromosome-linked HPRT gene inactivation (39). Methylation has an important role in the establishment or maintenance of imprinting (26). Indeed, methyltransferase mouse mutants do not show a normal pattern of imprinting (40). The function of progressive Pa methylation may thus be adjunctive to an imprinting program that works to inactivate normal abl transcription at the Ph' chromosome. Thus, the observed methylation, apart from being a marker of tumor progression in CML, could also contribute to the process of blastic transformation. Chromosomal translocations contribute to tumorigenic processes in many ways, among which deregulation of the partner gene expression is a frequent event in certain translocations (41). It is conceivable that progressive methylation near translocation sites is associated with the progression of other tumors as well and not restricted to CML. We are grateful to Dr. Eli Canaani for his advice, CML cell lines, and the abi Ia genomic clone, to Drs. Howard Cedar and Aharon

Razin for their constructive comments, and to Iris Nefesh for her excellent technical assistance. We thank Dr. Lynn Wang and Offer Gerlitz for critical reading of the manuscript. The work was sup-

Proc. Nad. Acad. Sci. USA 91

(1994)

ported by the Israel Cancer Research fund and the Research Associates of the Lautenberg Center. 1. Heisterkamp, N., Stephenson, J. R. & Groffen, J. (1983) Nature (London) 306, 239-242. 2. Shtivelman, E., Lifshitz, B., Gale, R. P. & Canaani, E. (1985) Nature (London) 315, 550-554. 3. Ben-Neriah, Y., Daley, G. Q., Mes-Mason, A. M., Witte, 0. N. & Baltimore, D. (1986) Science 233, 212-214. 4. Daley, G. Q. & Ben-Neriah, Y. (1991) Adv. Cancer Res. 57, 151-184. 5. Athens, J. W. (1993) in Wintrobe's Clinical Hematology, eds. Lee, R. G., Bithell, T. C., Foerster, J., Athens, J. W. & Lukens, J. N. (Lea & Febiger, Philadelphia), 9th Ed., Vol. 2, pp. 1969-1998. 6. Kantarjian, H. M., Smith, T. L., McCredie, K. B., Keating, M. J., Walters, R. S., Talpaz, M., Hester, J. P., Bligham, G., Gehan, E. & Freireich, E. J. (1985) Blood s, 1326-1335. 7. Allan, H. C. (1989) Blood Rev. 3, 45-50. 8. Segal, G. B., Simon, W. & Lichtman, M. A. (1986) Blood 68, 1055-1064. 9. Rowley, J. D. (1990) Semin. Hematol. 27, 122-136. 10. Ahuja, H., Bar-Eli, M., Advani, S. H., Benchimol, S. & Cline, M. J. (1989) Proc. Natl. Acad. Sci. USA 86, 6783-6787. 11. Liu, E., Hielde, B. & Bishop, J. M. (1988) Proc. Natd. Acad. Sci. USA 85, 1952-1956. 12. Westbrook, C. A., Rubin, C. M., Carrino, J. J., Le beau, M. M., Bernards, A. & Rowley, J. D. (1988) Blood 71, 697-702. 13. Romero, P., Beran, M., Shtalrid, M., Andersson, B., Talpaz, M. & Blick, M. (1989) Oncogene 4, 93-98. 14. Keating, A., Martin, P. J., Bernstein, I. D., Papayanopoulous, T., Raskind, W. & Singer, J. W. (1983) in Proceedings of the UCLA Symposia on Molecular and Cellular Biology, eds. Gold, D. & Marks, P. (Liss, New York), pp. 513-520. 15. Collins, S. J. & Groudine, M. T. (1983)Proc. Natl. Acad. Sci. USA 8, 4813-4817. 16. Erikson, J., Griffin, C. A., Ar-Rushdi, A., Valtieri, M., Hoxie, J., Finan, J., Emanuel, B. S., Rovera, G., Nowell, P. C. & Croce, C. M. (1986) Proc. Natd. Acad. Sci. USA-83, 1807-1811. 17. Hughes, T. P. & Goldman, J. M. (1990) in Hematology: Basic Principle and Practice, eds. Hoffman, R., Benz, E. J., Shattill, S. J., Furie, B. & Cohen, H. J. (Churchill Livingstone, New York), p. 858. 18. Shtivelman, E., Lifshitz, B., Gale, R. P., Roe, B. A. & Canaani, E.

(1986) Cell 47, 277-284.

19. Jiang, X., Trujillo, J. M., Dao, D. & Liang, J. C. (1989) Cancer Genet. Cytogenet. 42, 287-294. 20. Bernards, A., Rubin, C. M., Westbrook, C. A., Paskind, M. & Baltimore, D. (1987) Mol. Cell. Biol. 7, 3231-3236. 21. Ben-Neriah, Y., Bernards, A., Paskind, M., Daley, G. Q. & Baltimore, D. (1986) Cell 44, 577-586. 22. Renshow, M. W., Capozza, M. A. & Wang, J. Y. J. (1988) Mol. Cell. Biol. 8, 4547-4551. 23. Bird, A. P. (1993) Curr. Biol. 3, 275-277. 24. Stoger, R., Kubicka, P., Liu, C. G., Kafri, T., Razin, A., Cedar, H. & Barlow, D. P. (1993) Cell 73, 61-71. 25. Lindsay, S. & Bird, A. P. (1987) Nature (London) 327, 336-338. 26. Razin, A. & Cedar, H. (1994) Cell 77, 473-476. 27. Antequera, F., Boyes, J. & Bird, A. (1990) Cell 62, 503-514. 28. Sakai, T., Toguchida, J., Ohtani, N., Yandell, D. W., Rapaport, J. M. & Dryja, T. P. (1991) Am. J. Hum. Genet. 48, 880-8. 29. Nelkin, B. D., Przepiorka, D., Burke, P. J., Thomas, E. D. & Baylin, S. B. (1991) Blood 77, 2431-2434. 30. Mills, K. I., Sproul, A. M. & Burnett, A. K. (1993) Leukemia 7, 707-711. 31. Feinberg, A. P. & Vogelstein, B. (1983) Nature (London) 301, 89-92. 32. Jackson, P. & Baltimore, D. (1989) EMBO J. 8, 449-456. 33. Sawyers, C. L., McLaughlin, J., Goga, A., Havlik, M. & Witte, 0. (1994) Cell 77, 121-131. 34. Wylie, J. A. & Green, A. R. (1986) in Oncogenes and Growth Control, eds. Kohn, P. & Graf, T. (Springer, Berlin), pp. 340-345. 35. Haas, 0. A., Argyriou-Tirita, A. & Lion, T. (1992) Nature (London) 359, 414-416. 36. Rigins, G. J., Zhang, F. & Warren, S. T. (1994) Nat. Genet. 6,226. 37. Tybulewicz, V. L. J., Crawford, P. K. J., Bronson, R. T. & Mudligan, R. C. (1991) Cell 65, 1153-1163. 38. Schwartzberg, P. L., Stall, A. M., Hardin, J. D., Bowdish, K. S., Hamaran, T., Boast, S., Harbison, M. L., Robertson, E. J. & Goff, S. P. (1991) Cell 65, 1165-1175. 39. Lock, L. F., Takagi, N. & Martin, G. R. (1987) Cell 48, 39-46. 40. Beard, L. E. & Jaenisch, R. (1993) Nature (London) 36, 362-365. 41. Taub, R., Moulding, C., Battey, J., Murphy, W., Vasicek, T., Lenoir, G. M. & Leder, P. (1984) Cell 36, 339-348.