MDS And Clonal Leukemic Differentiation

Methylation status of the p15INK4B gene in hematopoietic progenitors and peripheral blood cells in myelodysplastic syndromes

Article metrics


We previously reported that the hypermethylation of the p15INK4B gene promoter was frequently observed in myelodysplastic syndromes (MDS), and that it may be associated with disease progression. An unanswered question is whether p15INK4B gene methylation is restricted to undifferentiated blastic cells, or whether differentiated cells such as granulocytes or erythrocytes of MDS origin also harbor this epigenetic alteration. In this study, we analyzed the methylation status of the p15INK4B gene in MDS by the methylation-specific PCR (MSP) method, which is more sensitive than Southern blotting. The bone marrow mononuclear cells (BM-MNCs) of 23 MDS patients were analyzed, and six of them showed p15INK4B methylation. Progenitor assay with methylcellulose medium was also performed in all patients. In two of the six patients with p15INK4B-methylated BM-MNCs, erythroid and/or non-erythroid colonies formed were subjected to molecular analysis. Colonies with and without p15INK4B methylation were detected in both patients. Furthermore, X-chromosome inactivation (XCI) pattern of each colony was simultaneously determined by MSP-based human androgen receptor gene analysis (HUMARA-MSP), and all p15INK4B-methylated colonies showed the same XCI pattern, which was dominant among the colonies, while p15INK4B-unmethylated colonies showed both patterns of XCI, in each of the two patients. We then examined the methylation status of the p15INK4B gene of granulocyte (PB-PMN) fractions from 10 patients with available peripheral blood cells. In all four patients with p15INK4B-methylated BM-MNCs, their PB-PMNs showed p15INK4B methylation. These results suggest that p15INK4Bmethylation in hematopoietic cells in MDS patients is restricted to the MDS clone but not necessarily to blast cells.


Myelodysplastic syndrome (MDS) is a hematopoietic stem cell disorder characterized by peripheral blood cytopenia, morphological abnormalities of bone marrow cells and frequent progression to acute leukemia. Although several cytogenetic abnormalities, such as deletion of chromosomes 5 and 7, have been known to associate with this heterogeneous disorder, little is known about what genetic change(s) are responsible for this disease. Several genetic changes, including expression of chimerical proteins due to chromosome translocations, have been found in MDS,123 mostly in advanced cases. These abnormalities are also found in some de novo acute myelogenous leukemia (AML), suggesting that such genetic changes may be late events in MDS. In recent years, epigenetic changes, which do not alter primary structures of the genes but affect regulation of gene expression, have been recognized as playing certain roles in hematological and non-hematological malignancies.456 Hypermethylation of the regulatory sequences of tumor suppressor genes and consequent silencing of the gene could theoretically be equivalent to deletion or loss-of-function type of genetic mutations for tumorigenesis. In a previous study utilizing Southern blotting with a methylation-sensitive restriction enzyme, we reported that hypermethylation of the regulatory region of the p15INK4B tumor suppressor gene is frequently observed in bone marrow cells of MDS patients, especially, although not exclusively, in the advanced stage of the disease.7

Recent technical advances in studies of DNA methylation are the development of the bisulfite modification technique of DNA by Frommer et al,8 and that of the methylation-specific PCR (MSP) method, based on the former, by Herman et al.9 Cytosines are susceptible, while 5-methyl-cytosines are resistant to deamination and conversion to uracils by sodium bisulfite. Thus, the differences in methylation status in cytosines can be converted to those in primary structures of the DNA. Methylated-sequence-specific (M) primers and unmethylated-sequence-specific (U) primers can specifically amplify methylated and unmethylated alleles of the gene, respectively. An important advantage of this method in comparison to Southern blotting with methylation-sensitive restriction enzymes is its high sensitivity in detecting methylated alleles against the background of largely unmethylated genome.10 Quesnel et al11 examined bone marrow cells of 53 MDS patients by this method and found that most of them whose bone marrow contained more than 10% of blasts showed methylation of the p15INK4B gene, while none with less than 10% of same had p15INK4B methylation. These results suggest that it is the blasts that harbor hypermethylated allele of the p15INK4B gene. In our previous study by Southern blotting, which is much less sensitive in detecting methylation than the MSP method, however, the bone marrow cells of an RA patient with few blasts in the bone marrow showed significant methylation of this gene.7 Since not only the blasts but also the majority of the mature blood cells are considered to be of clonal origin in MDS,1213 it is of interest whether only the blasts have p15INK4B methylation, or whether cells that have differentiated into mature cells also harbor this epigenetic alteration. Another question is whether methylation of the p15INK4B gene in hematologic cells in MDS patients is restricted to the MDS clone. This issue, which is generally taken for granted, could be of importance because unlike mutations, cytosine methylation is considered to be a possibly reversible phenomenon.

The present study aims to determine: (1) whether the methylation of the p15INK4B gene seen in hematologic cells in MDS patients is truly restricted to the MDS clones; and (2) whether p15INK4B gene methylation is restricted to blasts, or cells differentiated in vitro or in vivo also harbor the same epigenetic change.

Materials and methods


We analyzed 23 patients with MDS, including 11 RA, two RARS, five RAEB, three RAEB-T, and two secondary acute myelogenous leukemia (sAML) evolved from MDS. One patient developed MDS after high-dose chemotherapy and subsequent autologous bone marrow transplantation for non-Hodgkin's lymphoma. Seven of these patients were previously analyzed for p15INK4B methylation by Southern blotting.7 Characteristics of the patients are summarized in Table 1.

Table 1  Patients characteristics, p15 gene methylation status and colony formation

Cell fractionation and DNA extraction

Aspiration and venipuncture obtained bone marrow (BM) and peripheral blood (PB), respectively, with informed consent of the patients. The mononuclear cells (MNCs) were separated from bone marrow aspirates by density gradient centrifugation on Ficoll–Paque (Pharmacia, Uppsala, Sweden). The PB-polymorphonuclear (PMN) fractions were separated from heparinized peripheral blood as described.14 Briefly, 10 to 20 ml of heparinized blood was sedimentated through Ficoll–Paque, and mononuclear cells (MNCs) were removed from the interface layer. PB-PMNs were collected from the pellet by removing lysed erythrocytes after incubation in 17 mM Tris-NH4Cl buffer. High molecular genomic DNA from each fraction of hematologic cells and cell lines (HL60, ML1) was purified by the standard procedure with SDS-Proteinase K treatment and phenol extraction.15 DNA was also extracted from individual colonies by the same method as well as from PB-PMNs obtained from 30 healthy volunteers as controls with their informed consent.

Hematopoietic progenitor assay

BM-MNCs from MDS were incubated in RPMI with 20% fetal bovine serum (FBS) for 2 to 3 h in order to remove adherent cells, followed by 12 days culture in 1 ml of MethoCult GF H4434 (StemCell Technologies, Vancouver, Canada) containing methylcellulose in Iscove's MDEM, fetal bovine serum, bovine serum albumin (BSA), 2-mercapto-ethanol, L-glutamine, recombinant human erythropoietin (EPO), granulocyte–macrophage colony-stimulating factor (GM-CSF), interleukin-3 (IL-3) and stem cell factor (SCF) in a 35-mm culture dish (Nunc, Naperville, IL, USA) at 37°C with fully humidified atmosphere of 5% CO2. Individual erythroid and non-erythroid colonies were picked up by hand with a micropipette, and DNA was extracted and resuspended in 20 μl of TE buffer.

Bisulfite modification of DNA

DNA extracted from fractionated bone marrow, peripheral blood cells, or colonies was chemically modified by sodium bisulfite to convert all unmethylated (but not methylated) cytosines to uracils according to the method described by Herman et al,9 using CpGenome DNA Modification kit (Oncor, Gaithersburg, MD, USA). The chemical modification was performed according to the manufacturer's instruction. We modified 500 ng of DNA obtained from BM-MNCs, PB-PMNs and cell lines and 4 μl of 20 μl DNA solution from each colony. Modified DNA from each colony was resuspended in 20 μl of TE buffer. As controls to monitor efficiency of modification, DNA extracted from HL60 (p15INK4B-unmethylated) and ML1 (p15INK4B-methylated)7 was processed similarly.

MSP for the p15INK4B gene

With the primer sets described by Herman et al,9 the MSP for the p15INK4B was performed. Each primer set recognizes the nucleotide sequences of methylated and unmethylated p15INK4B gene promoter modified by sodium bisulfite, respectively, and PCR product can be seen only when methylated (or unmethylated) allele is present by M (or U) primer set. Thus, if the DNA sample contains both methylated and unmethylated p15INK4B alleles, which is often the case with clinical samples, bands can be seen both on the M and the U lanes. The PCR reaction mixture, which was described by Quesnel et al,11 contained 6.5 mmol/l MgCl2, 10 mmol/l Tris HCl, pH 9,50 mmol/l KCI, 0.1% Triton X100, 0.2 mg/l bovine serum albumin, dNTPs (each at 0.5 mM/l), 20 pmol of each primer, 5% dimethyl sulfoxide (DMSO), 1 U of Taq polymerase (Takara, Kyoto, Japan), and bisulfite modified DNA. Fifty nanograms of modified DNA from BM-MNCs were amplified for 40 cycles, while 2 μl of those from each colony were amplified for 50 cycles. PCR products were directly loaded on to 3% NuSieve agarose gel (FMC Bioproducts, Rockland, MD, USA), electrophoresed, stained with ethidium bromide, and visualized under UV illumination. As positive and negative controls for modification and PCR, modified DNA extracted from HL60 and ML1 were analyzed. To confirm the sensitivity of the MSP method, HL60 and ML1 cells were mixed at various ratios. We amplified 50 ng of modified DNA extracted from mixed cells for 50 cycles.

MSP for the HUMARA gene

We developed an MSP-based method for X-chromosome-linked clonality analysis, which is a modification of the HUMARA method described by Allen et al.16 The detail is described elsewhere.17 Briefly, the upstream methylated (M) and unmethylated (U) primers were set on the sequence of the human androgen receptor (HUMARA) gene, described by Allen et al to be differentially methylated on the active and inactive X-chromosomes, and the common downstream primer was located in the area without CpG dinucleotides. Between the upstream primer and the downstream primer lies the (CAG) repeat, the number of which is highly polymorphic. As the M and U primers for the p15INK4B gene amplify the methylated and unmethylated p15INK4B gene, respectively, the M and U primer sets (namely the upstream M primer and the downstream common primer, and the upstream U primer and the downstream common primer) for the HUMARA gene amplify the methylated (on the inactive X-chromosome) and unmethylated (on the active X-chromosome) alleles, respectively. When a single cell, or a cell population derived from a single cell such as a hematopoietic colony, from a female heterozygous for the HUMARA polymorphism, is examined by this method, the M primer set should produce one band while the U primer set should generate another band of a different size (Figure 1). If the cell population is polyclonal, both the M primer set and the U primer set produce two bands of similar intensity, respectively. A schematic representation of the band patterns predicted to be seen with monoclonal and polyclonal cell populations, in comparison with those by the conventional HUMARA assay, is shown in Figure 1. Since a hematopoietic colony is considered to be of single-cell origin, each colony picked up should produce either A or B pattern, if contamination with neighboring cells is negligible (Figure 1).

Figure 1

 A schematic representation of the band patterns by HUMARA-MSP. The methylated allele is shown as a bold line. The U and M primer sets for the HUMARA gene amplify the unmethylated (ie active) and methylated (ie inactive) allele, respectively. A cell which has methylated long (CAG) repeat allele of the HUMARA gene is named A-pattern, while that which has methylated short (CAG) repeat allele is named B-pattern.

In order to examine cells from single colonies and increase the sensitivity of the method, we employed a semi-nested, two-step strategy of amplification for HUMARA-MSP. Five microliters of bisulfite-modified DNA solution was added to 45 μl of buffer (100 mmol/l Tris-HCl, 500 mmol/l KCI, 15 mM MgCl2) containing the outer set of primers (for M and U amplification, respectively, 20 pmol/l each), dNTP and 1 U of Taq polymerase, and amplified for 35 cycles in a thermal cycler (Takara, Kyoto, Japan). One tenth of each reaction solution was added to the second reaction solution containing the inner, semi-nested primers (for M and U amplification, respectively, 20 pmol/l each) and amplified for further cycles. After the second reaction, each solution was separated by electrophoresis through 10% polyacrylamide gel and stained with ethidium bromide. On the basis of amplified band pattern, which reflects XCI pattern, each colony was determined to be either A- or B-pattern (Figure 1). In one patient whose numbers of CAG repeats of two alleles seemed to be very close (UPN 19), it was difficult to determine XCI patterns by the non-isotopic method we described above, because of the insufficient separation of two bands. In this case, we carried out second amplification for 35 cycles with the addition of α-32P dCTP at 1 μCi/reaction. PCR products were separated on 6% denaturing polyacrylamide gels containing 7-M urea that should allow detection of one base difference. Then the dried gel was exposed to phosphorimaging plates (Imaging plate BASIII; Fuji Photo Film, Kanagawa, Japan), followed by analysis with a laser image analyzer (FUJIX BAS 2000; Fuji Photo Film).

Southern hybridization

Southern hybridization with a methylation-sensitive restriction enzyme for the p15INK4B gene was performed as previously described.7 Briefly, 5 μg of DNA was digested with 20 U each of HindIII and Eco52I, a methylation-sensitive restriction enzyme, and separated through 0.7% agarose gel, transferred to a nylon membrane, hybridized to the α-32P dCTP-labeled p15INK4B cDNA probe and analyzed. For quantitative evaluation, we measured the intensity of the bands and then calculated methylation intensity as previously reported.7


Methylation status of the p15INK4B gene in bone marrow mononuclear cells

We performed methylation analysis of the p15INK4B gene by the MSP on the BM-MNC samples from the 23 MDS patients. Amplification bands with the unmethylated sequence-specific primer pair (p15-U primers) were detected in all samples, which probably reflects the existence of normal cells in all patients. Specific bands amplified with the methylated sequence-specific primer pair (p15-M primers) were observed in six (26%) of the patients (Table 1, Figure 2a). Hypermethylation of the p15INK4B gene was found mostly in the high-risk subtypes (RAEB, RAEB-T and sAML; 5/10,50%), and only one of the 13 low-risk patients (RA and RARS; 8%) showed p15INK4B gene methylation. This result confirmed previous reports.711 DNA obtained from PB-PMNs from 30 healthy volunteers and HL60 cell line showed amplification with only U primers, while DNA from MLI cell line showed M but not U bands (Figure 2a). In the cell mixture experiment, methylation could be detected in 1% but not in 0.1% of ML1 cell line sample (Figure 2b).

Figure 2

 p15-MSP on BM-MNCs and cell lines. (a) Methylation status analyses by MSP on cell lines and BM-MNC obtained in two patients (UPN 2 and 22). p15-U and -M primer sets produce specific products, respectively. HL60 (p15-unmethylated) showed U band only and ML1 (p15-methylated) showed M band only. UPN 2, RA patient, has unmethylated p15 gene. In contrast, UPN 22 with sAML showed p15 methylation. (b) HL60 and ML1 cells were mixed at various ratios.

Methylation status of the p15INK4B and HUMARA genes in hematopoietic colonies

After 12-day culture in cytokine-supplied methylcellulose medium, BM-MNCs from 13 out of 23 MDS patients formed erythroid and/or non-erythroid colonies. The distinction between erythroid and non-erythroid colonies was made based on morphology determined under an inverted microscope.18 As shown in Table 1, two of the six patients with methylated BM-MNCs formed colonies, and of the 17 patients without the p15INK4B gene methylation, 11 showed colony formation.

Among the 23 MDS patients analyzed, only two showed both p15INK4B methylation and colony formation. One was UPN 1, with RA, whose BM-MNCs formed non-erythroid colonies only, and the other was UPN 19 with RAEB-T and both erythroid and non-erythroid colony formation. Both patients showed normal karyotype on cytogenetic analysis of the bone marrow cells. The individual colonies from these two patients were picked up by hand and subjected to MSP analyses. After DNA extraction and bisulfite modification, PCR with p15-M and P15-U primer sets was performed, respectively, to examine if any of these colonies had p15INK4B gene methylation. In both patients, some colonies showed amplified bands with the p15-M primer set, indicating they actually had methylated the p15INK4B gene. Simultaneous analysis by HUMARA-MSP was done for both patients to determine the XCI pattern of each colony. Both patients were heterozygous for HUMARA polymorphism, but one of them (UPN 19) showed a very small difference in the sizes of the two alleles, and we needed to employ an isotopic method to detect the difference. Representative results are shown in Figure 3, and the results in the two patients are summarized in Table 2. For UPN 19, whose HUMARA alleles had a very small size difference, the representative band pattern would be as below: due to the construction of the primers, the expected product size for the U primer set is 5 bp longer than from the M primer set. Presumably UPN 19 has one (CAG) repeat (ie 3 bp) difference between the two alleles. If so, when her long repeat allele is methylated (A-pattern), the band size produced from the U primer set would be 2 bp longer than from the M primer set, while if her short repeat allele is methylated (B-pattern), the size difference between the products from the U and M primer sets would be 8 bp. The actual band patterns are compatible with these presumed results (Figure 3b).

Figure 3

 Representative patterns of p15-MSP and HUMARA-MSP on colonies from UPN 1 and UPN 19. (a) p15 methylation status and X-chromosome inactivation (XCI) pattern on some colonies obtained from UPN 1. p15-unmethylated colonies (colony 1 and 2) showed both A - and B-XCI pattern, whereas p15-methylated colonies such as colony 3 showed only A-pattern. (b) p15 methylation status and X-chromosome inactivation (XCI) pattern on some colonies obtained from UPN 19. The expecting product size from the U primer set is 5 bp longer than that from the M primer set. Since the patient has only one (CAG) repeat difference for HUMARA polymorphism, presumably, when her long (CAG) repeat allele is methylated (A-pattern), the size of the U band would be 2 bp longer than that of the M band, while when her short (CAG) repeat allele is methylated (B-pattern), the U band size would be 8 bp longer than the M band size.

Table 2  p15 gene methylation status and X-chromosome inactivation pattern of colonies

We picked up 41 non-erythroid colonies from the progenitor assay dishes for the RA patient (UPN 1), and 26 erythroid and 27 non-erythroid colonies from the RAEB-T patient (UPN 19). For p15-MSP in colonies from UPN 1, 21 of 41 non-erythroid colonies showed U bands only (p15-U colonies), eight showed M bands only (p15-M colonies), six showed both U and M bands, and six did not show either M or U bands (Table 2A). On simultaneous HUMARA-MSP analysis, 29 colonies were the A-pattern of XCI, five showed B-pattern and no bands were detectable in seven colonies. The failure to detect any PCR products either by p15-or HUMARA-MSP may well be due to the small amount of DNA utilized, as we and others experienced in similar experiments.181920 In p15-U colonies, the numbers of colonies showing A-pattern and B-pattern were 13/21 and 4/21, respectively, while all p15-M colonies were the A-pattern of XCI except for one colony with no detectable HUMARA-MSP bands.

In both erythroid and non-erythroid colonies from UPN 19, there were four types of colonies (U band only, M band only, both U and M bands, and no detectable band colonies) as shown in Table 2B. Among seven p15-U colonies in 27 non-erythroid colonies, five showed the A-pattern of XCI, one showed B-pattern. On the other hand, in 14 colonies with M band, 12 were the A-pattern of XCI and none was B-pattern. Among erythroid colonies, 15 colonies showed U band only, and 11 of those were A-pattern, four were B-pattern. For the six p15-M colonies, five showed the A-pattern and none the B-pattern of XCI.

Methylation status of the p15INK4B gene in peripheral blood cells

Peripheral blood samples were available for 10 of the 23 patients, and PB-PMN fractions were analyzed for p15INK4B gene methylation by the MSP. PB-PMNs are the fraction of the peripheral blood cells, in which polymorphonuclear granulocytes can be separated with 97% purity, when assessed with samples without blasts.14 As shown in Figure 4a, in all four patients whose BM-MNCs had p15INK4B methylation, the PB-PMNs were positive for p15INK4B methylation, while the PB-PMNs of all six patients with p15INK4B-unmethylated BM-MNCs did not show p15INK4B methylation. In the three patients with PB-PMNs positive for p15INK4B methylation, blasts were present in the peripheral blood (15–29%, Table 3). The observations and the fact that the MSP is a sensitive method for detection may indicate that the p15INK4B methylation in PB-PMNs is the result of blast cells in the PB-PMN fractions. However, one RA patient (UPN 1) with no blast in her PB showed p15INK4B methylation in PB-PMN. Moreover, the PB-PMNs of one of these patients (UPN 22) obtained earlier when her disease was in RA stage (blasts were not present in this peripheral blood) also showed p15INK4B methylation (Table 3). To assess the methylation status of the p15INK4B gene in PB-PMNs more quantitatively, we performed Southern blot hybridization for one patient (UPN 19). Methylation intensity7 of PB-PMNs obtained from this patient was 21% (Figure 4). She had 15% of blasts in her PB (Table 3), and the percentage of blasts in PB-PMN fraction is expected to be much lower than that in the unfractionated PB cells. Thus, the Southern blot result confirms that not only the blasts, but also at least some of the granulocytes, have methylated the p15INK4B gene.

Figure 4

 p15-MSP and Southern blotting on PB-PMNs. Methylation status on BM-MNC and PB-PMN in some patients. (a) PB-PMNs obtained from patients (UPN 1, 19, 20, 22) were positive for p15 methylation. (b) Southern blotting also showed p15-methylation of PB-PMN obtained from UPN 19 with RAEB-T. Lane 1, DNA from HL60; lane 2, DNA from ML1; lanes 3 and 4, DNA from UPN 19; lanes 5 and 6, control DNA (normal spleen B cell) digested by HindIII or HindIII and Eco52I.

Table 3  p15 gene methylation status in BM-MNCs and PB-PMNs


The importance of inactivation of tumor suppressor genes by methylation of their promoters has recently been acknowledged in hematologic and non-hematologic malignancies.456 The development of the MSP method9 is a recent landmark in the study of DNA methylation. This method has much higher sensitivity than Southern hybridization with methylation-sensitive restriction enzymes, and can detect specific methylation changes in, theoretically, any sequences with CpG dinucleotides. We earlier confirmed the usefulness of p15-MSP in a case of MDS/AML evolved from paroxysmal nocturnal hemoglobinuria (PNH).21 In the present study five out of 10 patients with more than 5% blast cells in the bone marrow (RAEB/RAEB-T/sAML) showed p15INK4B gene methylation, whereas only one in the 13 patients with less than 5% of blasts (RA/RARS) were positive for methylation. The results for the seven patients we had previously analyzed for p15INK4B methylation by Southern hybridization are compatible with those obtained by the MSP analysis in the present study. These results confirm previous observations,7891011 and it is clear that the more advanced the stage of disease is, the more frequent the methylation of the p15INK4B gene.

Quesnel et al11 examined bone marrow cells of 53 MDS patients by the MSP method and found p15INK4B gene methylation exclusively in samples from the patients with more than 10% of blast cells in their bone marrow. These results appear to indicate that there is no p15INK4B gene methylation when there are no blasts. It has been established, however, that MDS is a stem cell disorder, and not only immature blastic cells but also most differentiated cells in granulocytic, erythroid, megakaryocytic, and to a certain extent, B lymphocytic cells, can be derived from the MDS clone.1213 Thus, even if less than 10% of cells of the BM were blasts, some, if not most, variably differentiated cells in the bone marrow of the patients examined by Quesnel et al may well belong to the MDS clone. These considerations led us to examine whether blasts are the only cells that can be p15INK4B-methylated in MDS, or alternatively, whether cells with the capacity for in vitro and in vivo differentiation may also have hypermethylation of this gene.

Of the 23 patients, 13 (RA 9/12, RARS 1/1, RAEB 2/5, RAEB-T 1/3, sAML 0/2) showed erythroid and/or non-erythroid colony formation. Although statistical analysis could not be performed due to the small number of patients in our study, patients in advanced-stage (RAEB/RAEB-T/sAML) clearly tended to form fewer colonies than those in less-advanced stage (RA/RARS), a finding compatible with previous reports.22 When compared for the presence or absence of p15INK4B methylation, two of six p15INK4B-methylated patients showed colony formation, while BM-MNCs from 11 of 17 p15INK4B-unmethylated patients formed colonies. It is not clear whether the presence of p15INK4B methylation has any selection bias on colony formation, because, as mentioned above, p15INK4B methylation was found more frequently in patients in advanced stage. Among the six patients whose BM-MNCs showed p15INK4B methylation, only two formed colonies. To determine whether p15INK4B methylation was also present in some of the colonies of these two patients, these individual colonies were subjected to molecular analyses.

The result that some non-erythroid colonies from two patients (UPN 1 and 19) showed M bands for p15-MSP clearly demonstrates the methylation of the p15INK4B gene in hematopoietic progenitor cells, which have the capacity to differentiate into myeloid lineage in vitro. Also, in all four patients with BM-MNCs positive for p15INK4B methylation and whose peripheral blood samples were available, the granulocytic fractions (PB-PMNs) showed p15INK4B gene methylation by the MSP method. This result directly demonstrates that at least some cells undergoing terminal differentiation into myeloid lineage in vivo were p15INK4B-methylated. As shown in Table 2, six non-erythroid colonies derived from UPN 1 and three non-erythroid and one erythroid colony from UPN 19 demonstrated both U and M bands for p15-MSP. This fact deserves some consideration. The contamination of neighboring cells is almost inevitable in such experiments,181920 especially because in this study we performed colony assays with BM-MNCs and not with CD34-positive cells, which would effectively reduce background cells. Another possibility is that those cells actually have both methylated and unmethylated p15INK4B alleles. However, this seems less likely, because in our experience of p15-MSP with leukemia cell lines, which are considered to be homogeneous, most of them showed either U or M bands. Another explanation for this fact is that partial loss of p15INK4B gene methylation may have occurred during in vitro differentiation, however, there is no evidence, as far as we know, that epigenetic alteration occurs during a short-time culture without specific reagents. Since MSP is highly sensitive and can detect methylation from 1% of contamination cells, the more likely explanation for colonies with both M and U bands for p15-MSP seems to be the contamination of neighboring cells.

In UPN 19 with RAEB-T, not only non-erythroid but also erythroid colonies were formed. The result that six colonies showed M band only for p15-MSP among 26 erythroid colonies demonstrates that, at least in this case, hematopoietic cells capable of differentiating into erythroid lineage harbor p15INK4B methylation. We need to be cautious, however, to conclude from the results with these two patients, that p15INK4B-methylated progenitor cells in MDS always have the capacity to differentiate in vivo. Of the six patients with p15INK4B-methylated BM-MNCs, only cells from these two formed colonies, and they may not necessarily represent all the MDS patients with p15INK4B methylation.

We next compared the results of p15-MSP and HUMARA-MSP in colonies from two patients to see if there was any correlation. The results from the colonies that showed either only U or only M band for P15-MSP and either A- or B-pattern for HUMARA-MSP showed a clear tendency (shaded boxed region in Table 2). In UPN 1, of the 17 p15-U colonies, 13 were A-pattern and four were B-pattern for the HUMARA-MSP, while all seven p15-M colonies were the A-pattern of XCI determined by the HUMARA-MSP. Among non-erythroid and erythroid colonies of UPN 19, p15-U colonies showed either A- or B-pattern for HUMARA-MSP, while all the p15-M colonies showed the A-pattern of XCI. Since A-pattern colonies are clearly more prevalent that B-pattern colonies in both patients, it can be postulated that the MDS clone is of the dominant A-pattern.21 These results are compatible with the notion that p15INK4B methylation in MDS is not a random event but is restricted, as is generally assumed but not yet confirmed, to the MDS clone. One important, and disturbing, result is that there was one colony that showed both M and U bands for p15-MSP, and B-pattern for HUMARA-MSP in UPN 1. This may indicate that certain hematopoietic precursors, which have a separate origin from other putative MDS clone, harbor one methylated and one unmethylated p15INK4B allele. This may be the case. However, another, seemingly more probable explanation is that this result came from contamination of two different types of colonies. If one colony with the B-pattern of XCI with unmethylated p15INK4B gene was contaminated with relatively small number of p15INK4B-methylated, A-pattern cells, the results of p15-MSP and HUMARA-MSP could be as was seen with the colony in question. That is because HUMARA-MSP is a method for comparing the ratio between methylated and unmethylated alleles, while p15-MSP is the one for detecting the presence of methylated and/or unmethylated alleles. Thus, the latter could be more sensitive than the former for detecting the presence of relatively small numbers of the target sequence. For UPN 19, no such colony was found among the 53 colonies that we examined.

These results suggest that methylation of the p15INK4B gene in hematologic cells in MDS is restricted to the MDS clone, but not necessarily to the cells that had lost the capacity to differentiate into myeloid and/or erythroid lineage. These observations raise the question whether p15INK4B methylation is an early or a late event in MDS leukemogenesis. If p15INK4B methylation were restricted to blasts and not shared by differentiated cells, it could have provided good evidence for substantiating the idea that it is a late event and p15INK4B-methylated cells exist as a subclone of the ‘early’ MDS clone. Our results, however, do not contradict this hypothesis, because an advanced subclone could well retain some capacity to differentiate into mature cells. The fact that among the p15-U colonies from both UPN 1 and UPN 19, A-pattern colonies were more prevalent than B-pattern colonies may suggest that some of the p15-U colonies may also be of MDS origin. Simultaneous examination of p15INK4B methylation and a more specific marker of MDS clones, such as chromosome abnormalities, may provide a definite answer.


  1. 1

    Russel M, List A, Greenberg P, Woodward S, Glismann B, Parganas E, Ihle J, Taetle R . Expression of EVI1 in myelodysplastic syndromes and other hematologic malignancies without 3q26 translocations Blood 1994 84: 1243–1248

  2. 2

    Nucifora G, Birn DJ, Espinosa R, Erickson P, LeBeau MM, Roulston D, McKeithan TW, Drabkin H, Rowley JD . Involvement of the AML1 gene in the t(3;21) in therapy-related leukemia and in chronic myeloid leukemia in blast crisis Blood 1993 81: 2728–2734

  3. 3

    Nucifora G, Begy CR, Kobayashi H, Roulston D, Claxton D, Pedersen-Bjergaard J, Parganas E, Ihle JN, Rowley JD . Consistant intergenic splicing and production of multiple transcripts between AML1 at 21q22 and unrelated genes at 3q26 in (3;21)(q26;q22) translocations Proc Natl Acad Sci USA 1994 91: 4004–4008

  4. 4

    Laird PW, Jaenisch R . The role of DNA methylation in cancer genetic and epigenetics Ann Rev Genet 1996 30: 441–464

  5. 5

    Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP . Alterations in DNA methylation: a fundamental aspect of neoplasia Adv Cancer Res 1998 72: 141–196

  6. 6

    Issa JP, Baylin SB, Herman JG . DNA methylation changes in hematologic malignancies: biologic and clinical implications Leukemia 1997 11: (Suppl. 1) S7–11

  7. 7

    Uchida T, Kinoshita T, Nagai H, Nakahara Y, Saito H, Hotta T, Murate T . Hypermethylation of the p15INK4B gene in myelodysplastic syndromes Blood 1997 90: 1403–1409

  8. 8

    Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW, Molloy PL, Paul CL . A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in indicidual DNA strands Proc Natl Acad Sci USA 1992 89: 1827–1831

  9. 9

    Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB . Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands Proc Natl Acad Sci USA 1996 93: 9821–9826

  10. 10

    Herman JG, Civin CI, Issa JP, Collector MI, Sharkis SJ, Baylin SB . Distinct patterns of inactivation of p15INK4B and p16INK4A characterize the major types of hematological malignancies Cancer Res 1997 57: 837–841

  11. 11

    Quesnel B, Guillerm G, Vereecque R, Wattel E, Preudhomme C, Bauters F, Vanrumbeke M, Fenaux P . Methylation of the p15INK4B gene in myelodysplastic syndrome is frequent and acquired during disease progression Blood 1998 91: 2985–2990

  12. 12

    Janssen JW, Buschle M, Layton M, Drexler HG, Lyons J, van den Berghe H, Heimpel H, Kubanek B, Kleihauer E, Muffi GJ, Bartram CR . Clonal analysis of myelodysplastic syndromes: evidence of multipotent stem cell origin Blood 1989 73: 248–254

  13. 13

    Tsukamoto N, Morita K, Maehara T, Okamoto K, Karasawa M, Omine M, Naruse T . Clonality in myelodysplastic syndromes: demonstration of pluripotent stem cell origin using X-linked restriction fragment length polymorphism Br J Haematol 1993 83: 589–594

  14. 14

    Ohashi H, Hotta T, Ichikawa A, Kinoshita T, Taguchi R, Kiguchi T, Ikezawa H, Saito H . Peripheral blood cells are predominatly chimeric of affected and normal cells in patients with paroxymal noctural hemoglobinuria (PNH): simultaneous investigation on clonality and expression of glycophosphatidylinositol-anchored proteins Blood 1994 83: 853–859

  15. 15

    Maniatis T, Fritsch EF, Sambrook J . Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory, Cold Spring Harbor 1982 pp 9–14

  16. 16

    Allen RC, Zoghbi HY, Moseley AB, Rosenblatt HM, Belmont JW . Methylation of Hpa II and Hha I sites near the polymorphic CAG repeat in the human androgen-receptor gene correlates with X chromosome inactivation Am J Hum Genet 1992 51: 1229–1239

  17. 17

    Uchida T, Ohashi H, Aoki E, Nakahara Y, Hotta H, Murate T, Saito H, Kinoshita T . Clonality analysis by methylation-specific PCR for the human androgen-receptor gene (HUMARA-MSP) Leukemia (in press)

  18. 18

    Suzuki H, Asano H, Ohashi H, Kinoshita T, Murate T, Saito H, Hotta T . Clonality analysis of refractory anemia with ring sideroblasts: simultaneous study of clonality and cytochemistry of bone marrow progenitors Leukemia 1999 13: 130–134

  19. 19

    Gilliland DG, Blanchard KL, Levy J, Perrin S, Bunn HF . Clonality in myeloproliferative disorders: analysis by means of the polymerase chain reaction Proc Natl Acad Sci USA 1991 88: 6846–6852

  20. 20

    Asano H, Ohashi H, Ichihara M, Kinoshita T, Murate T, Kobayashi M, Saito H, Hotta T . Evidence for nonclonal hematopoietic progenitor cell populations in bone marrow of patients with myelodysplastic syndromes (MDS) Blood 1994 84: 588–594

  21. 21

    Uchida T, Ohashi H, Kinoshita T, Saito H, Taguchi R, Hotta T, Murate T . Hypermethylation of p15INK4B gene in a patient with acute myelogenous leukemia (AML) evoked from paroxymal noctural hemoglobinuria (PNH) Blood 1998 92: 2891–2893

  22. 22

    Oscier DG, Worsley A, Darlow S, Figes A, Williams JD, Hamblin TJ . Correlation of bone marrow colony growth in the myelodysplastic syndromes with the FAB classification and the Bournemouth score Leukemia Res 1989 13: 833–839

Download references


We thank Dr Suzuki and Dr Nakahara for suggestions on the experimental procedures, and Ms Hatano and Ms Kondo for technical assistance. This study was supported in part by the Ministry of Health and Welfare's Research Grant for Specific Disease, Japan.

Author information

Correspondence to H Ohashi.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Aoki, E., Uchida, T., Ohashi, H. et al. Methylation status of the p15INK4B gene in hematopoietic progenitors and peripheral blood cells in myelodysplastic syndromes. Leukemia 14, 586–593 (2000) doi:10.1038/sj.leu.2401719

Download citation


  • myelodysplastic syndromes
  • DNA methylation
  • p15INK4B gene

Further reading