DNA Methylation Patterns at Relapse in Adult Acute Lymphocytic Leukemia1

Vol. 8, 1897–1903, June 2002 Clinical Cancer Research 1897

Guillermo Garcia-Manero,2 Carlos Bueso-Ramos, Conclusions: In summary, DNA methylation patterns Jerry Daniel, Jason Williamson, are stable in a majority of patients with relapsed ALL, but a subset of patients acquire new methylation changes, in Hagop M. Kantarjian, and Jean-Pierre J. Issa particular affecting cell cycle regulatory genes. Departments of Leukemia [G. G-M., J. D., J. W., H. M. K., J-P. J. I.] and Hematopathology [C. B-R.], University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 INTRODUCTION DNA methylation refers to the addition of a methyl group to the cytosine in a cytosine-guanosine pair (CpG). CpG pairs ABSTRACT are underrepresented in eukaryotic genomes but can be found at Purpose: Aberrant DNA methylation of promoter- near-expected frequency in intergenic repetitive sequences such associated CpG islands is an epigenetic DNA modification as Alu repeats or in the proximity of gene promoters. Methyl- observed in acute leukemias that in certain cases has been ation of these promoter-associated areas has been associated associated with a poor prognosis and increased relapse rates. with gene silencing and is a physiological compensatory dosage To study the role of DNA methylation in relapse mecha- mechanism in X-linked and imprinted genes. By silencing func- nisms in acute lymphocytic leukemia (ALL), we have com- tionally relevant genes, aberrant DNA methylation has been pared the methylation status of five genes at the time of linked to diverse human pathologies (1), including cancer (2) initial presentation and at first relapse in 25 adult patients and aging (3). Aberrant methylation of multiple promoter-asso- with ALL. ciated CpG islands has been found in acute leukemias (4–6). Experimental Design: Genes studied included the estro- These studies have revealed that subgroups of patients with gen receptor (ER), multidrug resistance gene 1 (MDR1), p73, acute leukemia are characterized by a hypermethylator pheno- p15, and p16. DNA was extracted from paraffin-embedded type defined by the simultaneous methylation of multiple genes. bone marrow biopsies. DNA methylation was analyzed using This is similar to the CpG island methylator phenotype de- PCR of bisulfite-modified DNA. scribed in colon cancer (7) and suggests that aberrant methyla- Results: Results indicate that methylation at the time of tion is an important oncogenic mechanism in acute leukemias. relapse was stable in 92% of patients for p73, 88% for ER, Methylation of several genes has been associated with poor 80% for p16, 72% for MDR1, and 60% for p15. Only one prognosis in patients with acute leukemia. Hypermethylation of the calcitonin gene has been associated with poor prognosis and case had p16 methylation at initial presentation, whereas 6 3 had methylation at relapse. Three disease progression in ALL (8–10). This phenomenon has also (0.0001 ؍ patients (P cases had concomitant methylation of p15 and p16 at re- been observed for HIC1 (11) and WT-1 (12) in AML. In con- trast, hypermethylation of ER has been associated with a better lapse. The degree of MDR1 methylation inversely correlated prognosis in AML (13), although abnormal methylation of this with the presence of MDR1 expression as detected by im- gene is also found at the time of relapse (14). In bladder cancer, munohistochemistry. Eighteen patients (72%) had acquired hypomethylation of the MDR1 gene has been associated with no or one methylation change, whereas the rest (28%) had protein reexpression and a chemoresistant phenotype at relapse methylation changes in two or three genes. No clinical- (15). p15 and p16, two cell cycle regulatory genes, are consid- biological correlations were found between methylation of ered tumor suppressor genes. Diverse abnormalities of these two any particular gene or pattern. genes, including p15 methylation, have been shown to be acquired at the time of relapse in pediatric ALL (16). Methylation of p15 has been shown to confer poor prognosis and to predict the risk of relapse in patients with acute leukemias (17, 18). Received 11/15/01; accepted 3/14/02. Methylation of p16, a rare event in human leukemias (6, 19), has The costs of publication of this article were defrayed in part by the been associated with disease progression in adult T-cell leuke- payment of page charges. This article must therefore be hereby marked mia (20). Other lymphoid neoplasms, such as non-Hodgkin’s advertisement in accordance with 18 U.S.C. Section 1734 solely to lymphomas, also are characterized by abnormal methylation of indicate this fact. 1 This study was supported in part by a Translational Research Grant these two genes, a phenomenon also associated with relapse from the Leukemia and Lymphoma Society of America and Research Grant RPG-99-098-01 from the American Cancer Society (to J-P. J. I.). G. G-M. is the recipient of a Career Development Award from the American Society of Clinical Oncology and the Peggy Howard Memo- rial Fund. 3 The abbreviations used are: ALL, acute lymphocytic leukemia; AML, 2 To whom requests for reprints should be addressed, at Department of acute myeloid leukemia; MRD, minimal residual disease; ER, estrogen Leukemia, Box 428, University of Texas M. D. Anderson Cancer Cen- receptor; MDR1, multidrug resistance 1; CR, complete remission; MSP, ter, 1515 Holcombe Boulevard, Houston, TX 77030. Phone: (713) 745- methylation-specific PCR; COBRA, combined bisulfite restriction 3428; Fax: (713) 794-4297; E-mail: [email protected]. analysis.

Table 1 Genes studied Gene accession Restriction enzyme Annealing temperatures number Primers (coordinates) Restriction sites (no. of cycles) ER GGTTTTTGAGTTTTTTGTTTTG (300) BstUI 60 (3), 57 (4), 54 (5), 51 (25) X03635 AACTTACTACTATCCAAATACACCTC (505) 1 p15 GGAGTTTAAGGGGGTGGG (24747) BstUI 59 (37) AC000049 CCTAAATTACTTCTAAAAAAAAAC (24918) 2 MDR1 GTTATAGGAAGTTTGAGTTT (140829) TaqI 54 (3), 51 (4), 48 (5), 45 (26) AC002457 AAAAACTATCCCATAATAAC (141001) 1 p73 methylated ACCCCGAACATCGACGTCCG 60 (34) (Ref. 24) GGACGTAGCGAAATCGGGGTTC p73 unmethylated ATCACAACCCCAAACATCAACATCCA 60 (34) (Ref. 24) AGGGGATGTAGTGAAATTGGGGTTT p16 methylated TTATTAGAGGGTGGGGCGGATCGC 60 (34) (Ref. 19) GACCCCGAACCGCGACCGTAA p16 unmethylated TTATTAGAGGGTGGGGTGGATTGT 60 (34) (Ref. 19) CAACCCCAAACCACAACCATAA

(21). These findings suggest that aberrant methylation may have to UpG without modifying methylated sites. This allows their a role in leukemic relapse. Knowledge of the dynamics of differentiation by allele specific PCR (MSP), restriction diges- methylation patterns at the time of relapse may have important tion (COBRA) or sequencing. Bisulfite treatment of genomic implications, including the understanding of cellular clonal DNA was performed as described (23). DNA was extracted changes during disease progression, the development of assays using standard phenol-chloroform methods after samples had to detect MRD, and the development of targeted therapies using been deparaffinized using xylene and ethanol followed by di- hypomethylating agents for high-risk patients. gestion with proteinase K. After extraction, 2 ␮g of DNA were We have previously studied the methylation status of ER, used for bisulfite treatment. DNA was denatured in 0.2 N NaOH MDR1, p73, p15, p16, THBS1, THBS2, CD10, c-abl, and Myod at 37°C for 10 min and incubated with 3 M sodium bisulfite at in 80 patients with ALL (6). We have selected five of them, ER, 50°C for 16 h. DNA was then purified using the Wizard cleanup MDR1, p73, p15, and p16, to study the methylation patterns system (Promega) and desulfonated with 0.3 N NaOH at 25°C both at the time of initial presentation and at relapse in 25 for 5 min. DNA was then precipitated with ammonium acetate patients with relapsed ALL who had achieved complete remis- and ethanol, washed with 70% ethanol, dried, and resuspended sion after initial induction chemotherapy. Our results indicate in H2O. that the methylation patterns of patients with relapsed ALL are Primer Design. Primer sequences, coordinates, Gen- stable in a majority of patients, but that a subgroup of patients Bank accession numbers, number of expected restriction acquire new methylation changes, in particular of genes in- fragments, and PCR conditions are shown in Table 1. For p73 volved in cell cycle regulation. and p16, we used MSP as described elsewhere (19, 24). For p15, MDR1, and ER, we used COBRA (25). To minimize MATERIALS AND METHODS overestimation of methylated alleles when using this method, Samples. Paraffin-embedded bone marrow biopsies from the following points were followed: (a) primers were de- 25 adult patients with ALL obtained at the time of original signed to contain a minimum number of CpG dinucleotides in diagnosis and at first relapse have been analyzed in this study. their sequence to avoid biased amplification of methylated Patients were selected from a cohort of 80 patients studied alleles. If primers contained CpG sites, they were designed to previously (6). Selection was based on evidence of relapse and amplify methylated and unmethylated equally (with a mix- sample availability (both at initial presentation and relapse). All ture of C or T used for the sense strand or G or A for patients had been treated with the hyper-CVAD chemotherapy antisense primers); (b) primers were designed to contain a program (protocol DM92-048) at the University of Texas M. D. maximum number of thymidines converted from cytosines to Anderson Cancer Center (22). Patient characteristics and results avoid amplification of the nonconverted genomic sequence; of this program have been reported elsewhere (22). Consent for (c) amplification of genomic DNA not treated with bisulfite sample collection was obtained from all patients following was always carried out to monitor lack of nonspecific ampli- institutional guidelines. This study was reviewed and approved fication; (d) primers were designed to be within 300 bp of by the Institutional Review Board. known transcription start sites; and (e) for each set of prim- Bisulfite Modification of DNA. Methods for bisulfite ers, we tested: (i) multiple restriction enzymes to confirm the modification of DNA and subsequent PCR techniques used in methylation status and sequence of the fragment analyzed; this study are described in detail.4 Bisulfite induces deamination and (ii) performed mixing experiments (using methylated and of unmethylated cytosines, converting unmethylated CpG sites unmethylated templates mixed at a known ratios) to exclude any potential amplification bias. PCR and Restriction Digestion. PCR reactions were carried in 50-␮l reactions. In each reaction, 2 ␮l of bisulfite- 4 Internet address: http://www.mdanderson.org/leukemia/methylation. treated DNA were used, as well as 1.25 mM deoxynucleotide

␮ triphosphate, 6.7 mM MgCl2, 5 l of PCR buffer, 1 nmol of Table 2 Patient characteristics primers and 1 unit of Taq polymerase. All PCR reactions were Study cohort (%) performed using a hot start at 95°C for 5 min. After amplifica- Median age 26 tion, except for p73 and p16, PCR products were digested with Male sex 15 (60) specific restriction enzymes that digest alleles that were meth- WBC count Ͼ30 ϫ 109/liter 14 (56) ylated prior to bisulfite treatment (25) and separated in nonde- Platelet count Ͼ100 ϫ 109/liter 7 (28) Ͻ naturing polyacrylamide gels. Gels were stained with ethidium Hemoglobin level 10 g/dl 7 (28) Lactate dehydrogenase level 24 (96) bromide. The proportion of methylated versus unmethylated Ͼ600 units/liter product (digested versus undigested) was quantitated by densi- Karyotype tometric analysis, determining the density of methylation. Den- Diploid 3 (12) sitometric analysis was performed using a Bio-Rad Geldoc 2000 Ph-positive 7 (28) Hyperdiploid 1 (4) digital analyzer equipped with the Quantity One version 4.0.3 Miscellaneous 8 (32) software. For both p73 and p16, we used MSP as described Insufficient metaphases 6 (24) (19, 24). Immunophenotype Immunocytochemistry. Immunohistochemical staining Mature B 2 (8) T 1 (4) for MDR was performed on formalin-fixed, paraffin-embedded CALLA 19 (76) tissue sections of bone marrow biopsy or aspiration clot sec- Null 1 (4) tions, an avidin-biotin-peroxidase complex method, and an First CR rate 25 (100) a automated immunostainer (Dako, Carpinteria, CA). All tissue Second CR rate 10 (43) Median duration 1 CR 52 wk sections underwent heat-induced epitope retrieval using cit- Median duration 2 CR 85 wk rate buffer (pH 6.0). After rinsing in PBS, an anti-MDR1 Overall survival 107 wk antibody (1:25 dilution of C-19 goat polyclonal IgG; Santa a Only 23 patients evaluable. Cruz Biotechnology, Santa Cruz, CA) was applied as primary antibody for1hatroom temperature. The detection system used was LSAB2 detection kit (Dako). Negative and positive controls were also run. RESULTS Statistical Analyses. Criteria for stable, increased, or Patient Population. We have analyzed the methylation decreased methylation were defined based on the technique status of ER, MDR1, p73, p15, and p16 from paraffin-embedded used. For MSP (used for p73 and p16), stable methylation was biopsies from 25 patients with ALL obtained at the time of considered if the samples at original presentation and relapse initial presentation and at relapse. All patients were enrolled in were either methylated or unmethylated at both endpoints, in- the hyper-CVAD program at the University of Texas M. D. creased methylation if the bone marrow at presentation showed Anderson Cancer Center (22). All patients had achieved CR no methylation but there was evidence of methylation at relapse, prior to relapse. The clinical-pathological features of these pa- and decreased methylation if only the initial presentation sample tients are summarized in Table 2. showed evidence of methylation. For COBRA (used in ER, p15, Methylation Status of Five Selected Genes. To study and MDR1), methylation was stable if the samples had a density methylation patterns at the time of initial presentation and of methylation of 0–2% at both endpoints or more than that at relapse in ALL, we have analyzed the methylation status of five both endpoints, increased methylation if the initial sample had a genes in 25 adult patients with adult ALL. Genes studied in- density of 0–2% and Ͼ5% at relapse, and decreased methyla- cluded ER, MDR1, p73, p15, and p16. Different criteria were tion if the sample had a density of methylation Ͼ5% at initial used for gene selection. ER, MDR1, p15, and p73 are known to presentation and Ͻ2% at relapse. Methylation patterns refer to be methylated in ALL (6, 14, 17, 24, 26). Methylation of ER and the number of methylated or unmethylated genes in a single p15 have also been shown to correlate with disease progression patient. For those genes for which COBRA was used to measure (17, 27). Expression of MDR1 has been correlated with a che- methylation, and thus a qualitative analysis could be performed, moresistant phenotype in bladder cancer (15). MDR1 expression we used a t test for dependent samples to compare the methy- has been associated with a poor prognosis in elderly patients lation status at initial presentation and at relapse. For those two with AML (28). p16 is infrequently methylated in ALL (6, 19), genes where MSP was used (p73 and p16) and only a qualitative but it has been found to be deleted at higher frequency in relapse analysis could be performed, we used the McNemar ␹2 test. pediatric ALL (16). Correlation coefficients were computed using Spearman rank For all these genes, except for p73, we have analyzed four correlation coefficients. Variables analyzed include individual to five normal bone marrow specimens, none of which showed gene methylation values and the patient characteristics age, significant (Ͼ2%) methylation (data not shown). p73 has been gender, WBC, hemoglobin, platelet counts, albumin, ␤2-micro- shown not to be methylated in normal marrow (24). Fig. 1 globulin, cytogenetics, immunophenotype, and systemic and shows representative examples of methylation for each gene. central nervous system risk at presentation (22). Definition of Table 3 summarizes the distribution of methylation densities for CR and disease-free survival have been described elsewhere each individual gene both at initial presentation and at relapse. (22). Estimated 5-year disease-free survival and overall survival Definitions for stable, increased, and decreased methylation are were based on the Kaplan-Meier method, and differences were given above. ER methylation remained stable in 22 cases (88%) tested using the log-rank test. of cases, increased in 2 (8%), and decreased in 1 (4%). MDR1

Fig. 2 Gene-specific variations in DNA methylation at relapse in ALL. The DNA methylation characteristics of five genes were studied both at initial presentation and at relapse in 25 patients with ALL. Two different techniques have been used in this study: COBRA for ER, MDR1, and p15, and MSP for p73 and p16. A patient was considered as having stable methylation if the samples at original presentation and relapse were either methylated or unmethylated at both endpoints (for MSP) or if the samples had a density of methylation of 0–2% at both endpoints or more than that at both endpoints (for COBRA), as having increased methylation if the bone marrow at presentation showed no methylation but there was evidence of methylation at relapse (for MSP) or if the initial sample had a density of 0–2% and Ͼ5% at relapse (for COBRA), or as having decreased methylation if only the initial presentation sample showed evidence of methylation (MSP) or if the sample had a density of methylation Ͼ5% at initial presentation and Ͻ2% at relapse.

Fig. 1 Changes in DNA methylation at relapse in ALL. For ER, The analysis of multiple genes at presentation and relapse MDR1, and p15, we have analyzed DNA methylation using COBRA (bisulfite PCR followed by restriction enzyme digestion). pre, initial allows the evaluation and comparison of methylation patterns at presentation samples; post, relapsed samples. Top arrow, unrestricted, relapse (Table 4). Eighteen patients (72%) had changes (in- unmethylated product; bottom arrow, restricted, methylated product. creased or decreased methylation) of 0 or 1 genes. Three pa- The name of the cell line used as control is indicated in the top tients had methylation changes of two genes (12%) and 4 right-most lane. For p73 and p16, DNA methylation has been analyzed using MSP. Two different PCR reactions were carried out, one with patients (16%) had changes in three genes, including 1 patient primers specific for the methylated sequence (p73-m and p16-m) and with complete demethylation of all three previously methylated one with primers for the unmethylated sequence (p73-um and p16-um). genes. Of importance, of the 5 cases with increased p16 meth- A sample was considered methylated if a band was visualized in the ylation at relapse, 3 had concomitant increase in p15 methyla- methylated reaction. tion and 1 was methylated at p15 at initial presentation. DNA Methylation Status Correlates with Protein Ex- pression. DNA methylation has been associated with tran- Table 3 Distribution of methylation densities at initial presentation scriptional inactivation and gene silencing. To demonstrate that and at relapse in ALL the changes in DNA methylation observed in this study corre- Density lated with the level of gene activation, we performed immunohistochemical analysis of MDR1 in cases with known methyl- 0–1.9% 2–14.9% Ն15% ation densities. Because of the sample source, we could not Prea Post Pre Post Pre Post study RNA levels. Three paired samples with known methyla- Gene n (%) n (%) n (%) n (%) n (%) n (%) tion levels were blindly evaluated for MDR1 expression by a ER 17 (68) 17 (68) 1 (4) 3 (12) 7 (28) 5 (20) hematopathologist. None of the samples studied with methyla- p15 17 (68) 13 (52) 0 4 (16) 8 (32) 8 (32) tion of MDR1 had detectable levels of the corresponding protein MDR1 17 (68) 16 (64) 1 (4) 0 7 (28) 9 (36) p16b 1 (4) 6 (24) by immunostaining, whereas all unmethylated samples had ev- p73b 14 (56) 14 (56) idence of protein expression, representing an excellent correla- a Pre, at initial presentation; Post, at relapse. tion between DNA methylation and MDR1 protein expression. b p16 and p73 were analyzed using MSP, and only qualitative Clinical and Pathological Correlates. Using Spearman assessment was made. Rank correlation analysis, we have explored possible correlations between methylation of a particular locus or methylation pattern with clinical and pathological characteristics and known methylation was stable in 18 cases (72%), increased in 4 (16%), prognostic factors in ALL. Methylation patterns analyzed in- and decreased in 3 (12%), and p15 remained stable in 15 (60%), cluded patients with 0–1 methylation changes versus more than increased in 7 (28%), and decreased in 3 (12%). The methyla- one gene and patients with concomitant p15 and p16 methylation status of p73 and p16 was analyzed using MSP; therefore, tion. Factors analyzed included age, gender, WBC, hemoglobin, no quantitative analysis could be performed. For p73, 23 cases platelet counts, albumin, ␤2-microglobulin, cytogenetics, im- remained stable (92%), whereas 1 increased and 1 decreased munophenotype, and systemic and central nervous system risk (4% each). p16 remained stable in 20 cases (80%) but increased at presentation. No significant association was found between in 5 (20%; P ϭ 0.0001; Fig. 2). No case of p16 demethylation methylation of any of the genes studied, any of the above was observed. patterns, and any of the studied clinical characteristics. Next, we

Table 4 Methylation patterns and individual patient clinical dataa

a Summary of individual patient characteristics and corresponding methylation patterns. Each row (#) represents a patient, and each column represents a characteristic. Age is shown in years. WBC, white blood cell count Ͼ30 ϫ 109/liter. Cyto (cytogenetics): 1, diploid; 2, t(9:22); 3, hyperdiploid; 5, miscellaneous; 7, insufficient metaphases. Pheno (immunophenotype): 1, T cell; 2, B cell; 3, null; 4, CALLA. Resp: (duration of first remission in weeks). 2nd CR: Y, yes; N, no. White boxes represent stable methylation at relapse; black boxes represent increased methylation, and gray boxes represent decreased methylation.

studied the impact of gene-specific methylation and methylation frequently methylated in acute leukemias and has been shown to patterns on second CR rate, duration of first and second remis- track with disease activity in ALL (14). Although ER methyl- sion, and overall survival. No correlation with second CR rate, ation has been shown to confer good prognosis to patients with duration of first and second remission, or overall survival was AML (13), we were unable to demonstrate such observation in observed. a large series of ALL patients (6). ER methylation remains stable at relapse in close to 90% of the cases studied. Unfortu- DISCUSSION nately, its role in disease prognostication and in MRD evalua- In the present study, we have analyzed the methylation tion is limited because ER methylation is observed in nonneo- patterns at the time of initial presentation and relapse in 25 adult plastic tissues (27). p73 is a p53 homologue that has been shown patients with ALL. Results indicate that the methylation status to be selectively methylated in lymphoid malignancies but not in of the five genes analyzed remained stable at relapse in 60–92% normal tissues (24, 26). No prior information exists regarding its of the cases, depending on the gene studied. As a consequence, methylation frequency in relapsed ALL. In our series, this gene methylation patterns are not altered at the time of relapse in a remains stable in Ͼ90% of cases, suggesting that evaluation of majority of patients (72%), although a significant minority p73 methylation could have a role in studies of MRD in ALL. (28%) acquires changes at relapse. These changes most fre- Expression of MDR1 has been shown to correlate with a poorer quently involved MDR1, p15, and p16. It is important to note prognosis in elderly patients with AML (28). Methylation of that p16 methylation is significantly increased at relapse, and MDR1 has been shown to correlate with the level of expression that this is usually accompanied by concomitant p15 methyla- and resistance in leukemia cell lines and in chronic lymphocytic tion. No associations were observed between methylation at leukemia (29). In bladder cancer, methylation of MDR1 has initial presentation and any of the clinical-biological character- been shown to inversely correlate with MDR1 overexpression, istics analyzed or with CR rates, disease-free survival, or overall an effect promoted by exposure to chemotherapy. In our series, survival. MDR1 methylation remained stable in 72% of cases, the rest of This is the first study to simultaneously analyze the meth- cases affected both by demethylation and hypermethylation of ylation status of multiple genes both at initial presentation and the gene. These changes in methylation density correlated with relapse in an homogeneously treated cohort of patients. ER is the level of MDR1 protein expression. Several studies have

analyzed the role of p15 and p16 abnormalities in relapse ALL. explored for remission maintenance in patients with known In a study of pediatric ALL (16), p16 was shown to be deleted methylation patterns. at relapse in 5 of 18 samples (30%). The same study also In summary, our results indicate that methylation patterns evaluated p15 deletion and methylation at relapse; 3 of 18 (16%) at relapse in ALL are stable in a majority of patients and confirm cases acquired p15 deletion at relapse. All 3 cases had also the importance of p15 and p16 in relapsed leukemia. acquired concomitant p16 deletion. Four of these cases were evaluated for p15 methylation both at initial presentation and at relapse; in 2 cases, methylation appeared at relapse; 1 case ACKNOWLEDGMENTS remained unmethylated; and 1 case was demethylated. We thank Sherry Pierce for data management and analysis. There are several limitations to our study. The first one is related to the number of patients and genes studied. Obviously, by increasing the number of patients, we could have observed REFERENCES relevant clinical-biological associations not found presently. 1. Robertson, K. D., and Wolffe, A. P. DNA methylation in health and Also, by studying more genes, we could have found other disease. Nat. Rev. Genet., 1: 11–19, 2000. important genes aberrantly methylated at relapse. The second 2. Herman, J. G. Hypermethylation of tumor suppressor genes in cancer. Semin. Cancer Biol., 9: 359–367, 1999. limitation is related to the techniques used to study methylation. 3. Toyota, M., and Issa, J. P. CpG island methylator phenotypes in In this study, we used archival paraffin-embedded biopsies, the aging and cancer. Semin. Cancer Biol., 9: 349–357, 1999. only relapse material available to us. The source and small 4. Melki, J. R., Vincent, P. C., and Clark, S. J. Concurrent DNA quantity of DNA obtained precluded the use Southern blot as hypermethylation of multiple genes in acute myeloid leukemia. Cancer done in prior studies. Bisulfite sequencing of PCR fragments, Res., 59: 3730–3740, 1999. the current gold standard for methylation analyses, would have 5. Toyota, M., Kopecky, K. J., Toyota, M. O., Jair, K. W., Willman, required close to 2500 sequencing reactions to conduct a com- C. L., and Issa, J. P. Methylation profiling in acute myeloid leukemia. prehensive analysis of the cohort studied here. To circumvent Blood, 97: 2823–2829, 2001. these issues, we have used two different methods, depending on 6. Garcia-Manero, G., Daniel, J., Pierce, S., Kornblau, S., Kantarjian, the genes studied. Both techniques, COBRA and MSP, exploit H., and Issa, J-P. CpG island methylation profile in acute lymphocytic leukemia (ALL). Blood, 96: 102a, 2000. the modification of DNA by bisulfite. By specifically deami- 7. Toyota, M., Ahuja, N., Ohe-Toyota, M., Herman, J. G., Baylin, S. B., nating nonmethylated cytosines, it allows the design of specific and Issa, J. P. CpG island methylator phenotype in colorectal cancer. primers for unmethylated and methylated sequences (MSP) or Proc. Natl. Acad. Sci. USA, 96: 8681–8686, 1999. the use of specific restriction enzymes (COBRA). Because we 8. Leegwater, P. A., Lambooy, L. H., De Abreu, R. A., Bokkerink, J. P., are comparing the methylation patterns at both endpoints, the and van den Heuvel, L. P. DNA methylation patterns in the calcitonin results obtained should be representative of the methylation gene region at first diagnosis and at relapse of acute lymphoblastic changes occurring in the promoter areas of the genes analyzed. leukemia (ALL). Leukemia (Baltimore), 11: 971–978, 1997. It is important to emphasize that in the present study, we 9. Thomas, X., Teillon, M. H., Belhabri, A., Rimokh, R., Fiere, D., Magaud, J. P., and Archimbaud, E. Hypermethylation of calcitonin gene have not attempted to evaluate the role of p15 and p16 deletions in adult acute leukemia at diagnosis and during complete remission. in relapse ALL, nor have we attempted to clarify whether the Hematol. Cell Ther., 41: 19–26, 1999. methylation changes are acquired in the initial dominant malig- 10. Roman, J., Castillejo, J. A., Jimenez, A., Bornstein, R., Gonzalez, nant cell or whether they represent a shift toward a selected M. G., Del Carmen Rodriguez, M., Barrios, M., Maldonado, J., and cellular clone at relapse. Torres, A. Hypermethylation of the calcitonin gene in acute lympho- To our knowledge, this is the most extensive study of DNA blastic leukaemia is associated with unfavourable clinical outcome. methylation at the time of relapse in ALL. It is important to note Br. J. Haematol., 113: 329–338, 2001. that all patients studied here were treated homogeneously with 11. Issa, J. P., Zehnbauer, B. A., Kaufmann, S. H., Biel, M. A., and Baylin, S. B. HIC1 hypermethylation is a late event in hematopoietic the same chemotherapy program. Although these results need to neoplasms. Cancer Res., 57: 1678–1681, 1997. be confirmed in prospective larger scale studies, this work has 12. Plass, C., Yu, F., Yu, L., Strout, M. P., El-Rifai, W., Elonen, E., important implications in the understanding of relapse mecha- Knuutila, S., Marcucci, G., Young, D. C., Held, W. A., Bloomfield, nisms in leukemia. The results suggest that relapse in ALL is C. D., and Caligiuri, M. A. Restriction landmark genome scanning for dominated by the original cellular malignant clone in a majority aberrant methylation in primary refractory and relapsed acute myeloid of patients, but a subgroup of patients acquires methylation leukemia: involvement of the WIT-1 gene. Oncogene, 18: 3159–3165, 1999. changes at relapse. The frequency of concomitant methylation 13. Li, Q., Kopecky, K. J., Mohan, A., Willman, C. L., Appelbaum, of p16 and p15, together with the high rate of deletions reported F. R., Weick, J. K., and Issa, J. P. Estrogen receptor methylation is by other groups (30), confirms the important role of these two associated with improved survival in adult acute myeloid leukemia. genes in relapsed leukemia. The fact that we have not found any Clin. Cancer Res., 5: 1077–1084, 1999. defining clinical or biological characteristic of this subgroup of 14. Issa, J. P., Zehnbauer, B. A., Civin, C. I., Collector, M. I., Sharkis, patients is probably a result of the relatively few patients stud- S. J., Davidson, N. E., Kaufmann, S. H., and Baylin, S. B. The estrogen ied. These results have implications in the development of receptor CpG island is methylated in most hematopoietic neoplasms. methylation techniques to study MRD in ALL. Genes such as Cancer Res., 56: 973–977, 1996. p73 and p16 could be used to monitor disease progression in 15. Tada, Y., Wada, M., Kuroiwa, K., Kinugawa, N., Harada, T., Nagayama, J., Nakagawa, M., Naito, S., and Kuwano, M. MDR1 gene patients with ALL. The information obtained could be used to overexpression and altered degree of methylation at the promoter region treat patients still in morphological remission with hypomethy- in bladder cancer during chemotherapeutic treatment. Clin. Cancer Res., lating agents. The use of 5-aza-2-deoxycytidine could also be 6: 4618–4627, 2000.

16. Maloney, K. W., McGavran, L., Odom, L. F., and Hunger, S. P. 23. Clark, S. J., Harrison, J., Paul, C. L., and Frommer, M. High Acquisition of p16(INK4A) and p15(INK4B) gene abnormalities be- sensitivity mapping of methylated cytosines. Nucleic Acids Res., 22: tween initial diagnosis and relapse in children with acute lymphoblastic 2990–2997, 1994. leukemia. Blood, 93: 2380–2385, 1999. 24. Corn, P. G., Kuerbitz, S. J., van Noesel, M. M., Esteller, M., 17. Wong, I. H., Ng, M. H., Huang, D. P., and Lee, J. C. Aberrant p15 Compitello, N., Baylin, S. B., and Herman, J. G. Transcriptional silenc- promoter methylation in adult and childhood acute leukemias of nearly ing of the p73 gene in acute lymphoblastic leukemia and Burkitt’s Ј all morphologic subtypes: potential prognostic implications. Blood, 95: lymphoma is associated with 5 CpG island methylation. Cancer Res., 1942–1949, 2000. 59: 3352–3356, 1999. 25. Xiong, Z., and Laird, P. W. COBRA: a sensitive and quantitative 18. Tien, H. F., Tang, J. H., Tsay, W., Liu, M. C., Lee, F. Y., Wang, DNA methylation assay. Nucleic Acids Res., 25: 2532–2534, 1997. C. H., Chen, Y. C., and Shen, M. C. Methylation of the p15(INK4B) gene in myelodysplastic syndrome: it can be detected early at diagnosis 26. Kawano, S., Miller, C. W., Gombart, A. F., Bartram, C. R., Matsuo, or during disease progression and is highly associated with leukaemic Y., Asou, H., Sakashita, A., Said, J., Tatsumi, E., and Koeffler, H. P. Loss of p73 gene expression in leukemias/lymphomas due to hyper- transformation. Br. J. Haematol., 112: 148–154, 2001. methylation. Blood, 94: 1113–1120, 1999. 19. Herman, J. G., Civin, C. I., Issa, J. P., Collector, M. I., Sharkis, S. J., 27. Issa, J. P., Baylin, S. B., and Herman, J. G. DNA methylation and Baylin, S. B. Distinct patterns of inactivation of p15INK4B and changes in hematologic malignancies: biologic and clinical implica- p16INK4A characterize the major types of hematological malignancies. tions. Leukemia., 11 (Suppl. 1): S7–S11, 1997. Cancer Res., 57: 837–841, 1997. 28. Leith, C. P., Kopecky, K. J., Godwin, J., McConnell, T., Slovak, 20. Nosaka, K., Maeda, M., Tamiya, S., Sakai, T., Mitsuya, H., and M. L., Chen, I. M., Head, D. R., Appelbaum, F. R., and Willman, C. L. Matsuoka, M. Increasing methylation of the CDKN2A gene is associated Acute myeloid leukemia in the elderly: assessment of multidrug resist- with the progression of adult T-cell leukemia. Cancer Res., 60: 1043– ance (MDR1) and cytogenetics distinguishes biologic subgroups with 1048, 2000. remarkably distinct responses to standard chemotherapy. A Southwest 21. Martinez-Delgado, B., Richart, A., Garcia, M. J., Robledo, M., Oncology Group study. Blood, 89: 3323–3329, 1997. Osorio, A., Cebrian, A., Rivas, C., and Benitez, J. Hypermethylation of 29. Kantharidis, P., El-Osta, A., deSilva, M., Wall, D. M., Hu, X. F., P16ink4a and P15ink4b genes as a marker of disease in the follow-up of Slater, A., Nadalin, G., Parkin, J. D., and Zalcberg, J. R. Altered non-Hodgkin’s lymphomas. Br. J. Haematol., 109: 97–103, 2000. methylation of the human MDR1 promoter is associated with acquired 22. Kantarjian, H. M., O’Brien, S., Smith, T. L., Cortes, J., Giles, F. J., multidrug resistance. Clin. Cancer Res., 3: 2025–2032, 1997. Beran, M., Pierce, S., Huh, Y., Andreeff, M., Koller, C., Ha, C. S., 30. Maloney, K. W., McGavran, L., Odom, L. F., and Hunger, S. P. Keating, M. J., Murphy, S., and Freireich, E. J. Results of treatment with Different patterns of homozygous p16INK4A and p15INK4B deletions hyper-CVAD, a dose-intensive regimen, in adult acute lymphocytic in childhood acute lymphoblastic leukemias containing distinct E2A leukemia. J. Clin. Oncol., 18: 547–561, 2000. translocations. Leukemia (Baltimore), 12: 1417–1421, 1998.

Guillermo Garcia-Manero, Carlos Bueso-Ramos, Jerry Daniel, et al.

Clin Cancer Res 2002;8:1897-1903.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/8/6/1897

Cited articles This article cites 30 articles, 16 of which you can access for free at: http://clincancerres.aacrjournals.org/content/8/6/1897.full#ref-list-1

Citing articles This article has been cited by 12 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/8/6/1897.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/8/6/1897. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.