P15INK4B gene methylation and expression in normal ... - Nature

2 downloads 0 Views 123KB Size Report
P15INK4B gene methylation and expression in normal, myelodysplastic, and acute ..... While a diagnosis of myelodysplastic syndrome could be made for patient ...
Leukemia (2001) 15, 1589–1595  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu

P15INK4B gene methylation and expression in normal, myelodysplastic, and acute myelogenous leukemia cells and in the marrow cells of cured lymphoma patients HD Preisler, B Li, H Chen, L Fisher, J Nayini, A Raza, S Creech and P Venugopal Rush Cancer Institute, 1725 W. Harrison Street, Suite 809 Professional Building I, Chicago, Illinois 60612, USA

P15INK4B methylation and expression was studied in bone marrow cells obtained from normal individuals, from patients who had been cured of lymphoma, and from patients with either MDS or AML. The level of p15 methylation was very low in normal BM cells and in CD34+ and CD34− subpopulations (0–6.5%; med, = 2.5%). P15INK4B transcripts were present in each of these cell populations. In contrast, methylation was the usual situation in MDS and AML marrows. The presence of methylation of the p15INK4B gene did not always indicate an absence of expression nor was expression always present if methylation was absent. P15INK4B methylation was studied in the marrows of nine patients (one studied twice) who had been cured of lymphoma and in whom hemopoiesis was believed to be normal. Increased methylaton was present in all 10 marrows. These data indicate that p15INK4B methylation is likely to be a very early event in the development of the secondary hematologic disorders. Leukemia (2001) 15, 1589–1595. Keywords: secondary hematologic disorders; proliferation control; P15INK4B; gene methylation

Introduction Acute myelogenous leukemia cells exhibit at least three behavioral characteristics which distinguish these cells from normal cells: dysregulated proliferation with an apparent reduction in the controls that limit entry into S phase, greater proliferation potential, and impaired differentiation. Together, these characteristics provide the leukemia cells with a proliferation advantage over normal hemopoietic cells.1 The end result of this discordance in ‘fitness’ is the replacement of normal marrow cells by leukemia cells.1 Secondary AML is the end result of a progressive accumulation of genetic abnormalities. At the biological level, the accumulation of these abnormalities results in the sequential appearance of new abnormal clonal cell populations which have a proliferative advantage over their predecessors. At the clinical level this evolutionary process is represented by the sequential appearance of ‘benign’ MDS (RA/RARS), then ‘malignant’ MDS (RAEB/RAEBt), and finally ‘AML’.1 Patients cured of lymphoma by means of intensive cytotoxic therapy, especially those who have had an autologous stem cell transplant as part of their treatment, are at increased risk for developing a secondary hematologic disorder.2,3 This easily identifiable, seemingly normal patient population, together with the stepwise evolution of the secondary hematologic disorders, provides a unique opportunity for identifying the molecular abnormalities responsible for the development of the secondary hematologic disorders and for their evolution to AML. Recently, it has been reported that the transition of MDS to AML is accompanied by methylation of the p15INK4B tumor suppressor gene.4,5 Since the p15INK4B gene is involved in the

Correspondence: HD Preisler; Fax: 312 455 9635 Received 28 September 2000; accepted 15 May 2001

Table 1A P15 methylation and expression in normal bone marrow aspirate

Specimen No.

P15INK4B methylation 15INK4B methylation/total p15INK4B gene × 100%

P15INK4B expression P15INK4B/beta-actin

0 0 0 1.1% 2.5% 0 4.5% 6.5% N/A N/A 4.0 1.1 2.5 4.5 6.5

0.29 0.4 0.67 0.47 0.72 0.49 0.31 0.26 0.30 0.31 NA NA NA NA NA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

N/A, specimens were not available.

Table 1B Methylation and expression in normal CD34+ and CD34− subpopulations of bone marrow cells

Specimen No.

Cell subsets

P15 methylation (%)

P15 expression

1

CD34+ CD34−

4% 7%

0.24 0.19

2

CD34+ CD34−

0% 1%

0.41 0.53

3

CD34+ CD34−

5% 1%

0.27 0.34

4

CD34+ CD34−

0% 0%

0.14 0.12

negative regulation of proliferation, a loss of function of the gene would result in dysregulated cell proliferation.6 The studies reported here demonstrate that detection of p15INK4B methylation in bone marrow does not always indicate an absence of expression of the gene in the marrow nor does the absence of methylation necessarily mean that the gene is being expressed. Further, evidence is presented that methylation of the gene is first detectable in the marrows of hematologically ‘normal’ cured lymphoma patients and that methylation progressively increases with the appearance of RA/RARS and with evolution to RAEB/RAEBt and AML. These observations suggest that abnormal p15INK4B function can be a very early event in the development of the secondary hematologic disorders.

The p15INK4B gene in MDS and AML HD Preisler et al

1590

Materials and methods

Patients studied Bone marrow specimens were obtained from 15 normal volunteers (Table 1), 10 unselected individuals who had been cured of lymphomas by means of intensive cytotoxic therapy and who were believed to be hematologically normal (referred to here as ‘post-treatment patients’) (Table 2), 25 unselected patients with MDS, and 44 patients with AML. Informed consent was obtained from each patient prior to their participation in this study. The FAB system was used to diagnose and classify a patient’s hematologic illness.7 Poor prognosis AML was defined as newly diagnosed AML in a patient with one or more of the following: age over 70 years, a history of toxic exposure, and/or AML following a preleukemic state. Standard prognosis AML refers to newly diagnosed patients with none of the preceding characteristics.8 A discrimination between standard and poor prognosis could be made for 40 of the 44 AML patients.

Acquisition of specimens for study Bone marrow was aspirated from the posterior iliac crest (10 ml) directly into tubes containing small amounts of preTable 2

Histories of patients cured of lymphoma

Patient Number

Patients age

Diagnosisa

Last treatment Time prior to BM studyb

P15INKb methylation

1

70

Low grade NHL

Chemotherapy 5 years

51%

2

69

Intermed grade NHL

Chemotherapy 2 years

3

46

Intermed grade NHL

4

42

5a

32

BM characteristicsc

PB countsd

Normal

Normal

70.6%

Erythroid hypoplasia

Normal

Autotransplant 4 years

60.6%

Mild hypercellularity Mild erythroid hyperplasia

Normal

Low grade NHL

Autotransplant 4 years

44.1%

Hypocellular erythroid series and myeloid hyperplasia

Platelet counts 138 000/␮l

Intermed grade NHL

Autotransplant 3 years

68.4%

Hypocellular bone marrow

Normal

Autotransplant 3. years

19.6%

Hypercellular bone marrow Megaloblastic erythropoiesis Myeloid hypoplasia

Normal

5b

6

42

Intermed grade NHL

Chemotherapy 1 year

5.3%

Erythroid hyperplasia

WBC 3600/␮l Platelet count 44 000/␮l Hemoglobin 12.9 g

7

66

Intermed grade NHL

Chemotherapy 6 years

88.8%

Hypercellular marrow Erythroid hyperplasia

WBC 2500/␮l Platelet count 82 000/␮l Hemoglobin 11.4 g

8

37

Intermed grade NH3

Chemotherapy 5 years

11.7%

Hypercellular marrow

Normal

9

49

Intermed grade NHL

Autotransplant 3 years

22.4%

Normal

Normal

a

Type of NHL. Time prior to BM study. c BM evaluation at time P15 study. d WBC × 103, Hg g, plt count × 103 at time of p15 study. ND, not done. The % myeloblasts was ⬍5% in every marrow. b

Leukemia

servative-free heparin. Marrow was usually obtained from two separate sites. The anticoagulated marrow cells were subjected to density centrifugation (sp. g. 1077) and the light density cells removed for study. Some light density specimens were frozen and stored at −160°C prior to study. Studies of normal marrow cells involved both whole bone marrow aspirates, as well as CD34+ and CD34− subpopulations, the latter permitting study of specimens in which there was a high proportion of immature cells and thus could serve as a normal control for specimens in which there was an increased proportion of immature cells (RAEB/RAEBt, AML). CD34 separation was carried out using a Cell Isolation Kit purchased from Miltenyi Biotech (Auburn, CA, USA). This cell isolation kit uses an indirect magnetic labeling system which permits the isolation of hematopoietic progenitor cells by the positive selection of CD34 expressing cells. Mononuclear cells separated by density gradient centrifugation using Ficoll– Paque are indirectly magnetically labeled using the haptenconjugated primary monoclonal antibody and an anti-hapten antibody coupled to MAC’s microbeads. These magnetically labeled cells are enriched on positive selection columns in the magnetic field of the MAC’s columns. Briefly, 108 cells are mixed with 100 ␮l of human Ig (FcR blocking reagent) followed by 100 ␮l of monoclonal hapten-conjugated CD34 antibody (clone: QBEND/10 isotype: mouse IGG1) and incubated for 15 min at 6–12°C. The final labeling volume is

The p15INK4B gene in MDS and AML HD Preisler et al

500 ␮l per 108 cells. The cells are carefully washed with buffer, centrifuged and the cell pellet resuspended in 400 ␮l of buffer. 100 ␮l of colloidal super paramagnetic MACs MicroBeads conjugated to an anti-hapten antibody reagent are added followed by incubation for 15 min at 6–12°C. The cell pellet is washed and resuspended in 500 ␮l of buffer. The retained cells are then recovered from the column. While too few cells were available to permit assessment of the % of CD34+ cells in the CD34+-enriched cell subpopulations used in our studies, in the last five CD34+ CD34− separations performed for other studies the % CD34+ cells ranged from 67% to 90%. P15 methylation was detected by P15INK4B methylation:9 PCR essentially as described by Uchida et al5 with some modifications. DNA was obtained from 5 × 106 bone marrow cells and divided into two aliquots. One of the aliquots was digested with 8 U of the methylation sensitive enzyme Eco521. The DNA was amplified in both aliquots using primers that flank the restriction site in the non-coding region of exon 1 (F:tcccagaagcaatccaggcg and R:tcagcttcattaccctcccg). Thirty cycles of 94°C 1 min, 58°C 1 min, 72°C 1 min were performed. The density of the PCR products and the degree of methylation were quantitated as described below. A standard curve derived from HL60 (methylation negative) and ML-1 (methylation positive) cells were constructed to measure the extent of p15 methylation. The cells were mixed in varying proportions (100% vs zero, 75% vs 25%, 50% vs 50%, 10% vs 90%). The band density was measured using Eagle Eye II software. The degree of methylation of the p15INK4B methylation in each specimen studied. The results are expressed as % p15INK4B gene band density of enzyme treated DNA p15INK4B gene band density of DNA not-treated with enzyme × 100%. RNA obtained from 5 × 106 bone P15INK4B expression:4 marrow cells were isolated using the RNAsol method (TelTest, Friendswood, TX, USA). In brief, the cells were dissolved in GITC. One ml RNAsol was then added. Total RNA was precipitated by ethanol and rehydrated with nuclease free water. cDNA was synthesized using a reverse transcription system kit (cat A3500; Promega, Madison, WI, USA). P15INK4B primers (p15INK4B 2F upstream primer 5⬘ ccagaagcaatccaggcgcg and p15INK4B 2R downstream primer 5⬘ cgttggcagccttcatcg) and beta-actin primer (upstream primer: 5⬘ GGGTCGGAAGGATTCCTATG and downstream primer: 5⬘ TCTCAAACATGATCTGGGTC). The PCR reaction conditions were chosen so that both PCR reactions were in a log phase. The PCR conditions were as follows: 35 cycles of 94°C 50 s, 62°C 50 s, 72°C 1 min. The RNA of HL-60 cells was used as a positive control.

Measurement of p15INK4B expression The densities of the p15INK4B and ␤-actin bands were measured using Eagle-Eye II software. The density of the p15INK4B band was divided by the band density of the housekeeping gene ␤-actin giving a ‘normalized’ value for p15INK4B transcript levels. This strategy permitted comparisons of the relative p15INK4B transcript levels in different specimens. This strategy also permitted the ␤-actin reaction to serve as a control for the rtPCR p15INK4B reaction since failure to detect ␤-actin

would indicate that the reaction had failed. The normalized level of p15INK4B expression (density p15INK4B band ⫼ density of ␤-actin band) are presented as arbitrary units.

1591

Statistical analyses SAS and Splus statistical analysis software were used for analysis of quantitative and semi-quantitative data. The extent of the p15INK4B methylation and expression were summarized with descriptive statistics such as the mean ± s.e., and medians. The Mann–Whitney U test was used to determine relationships between parameters, which were interval/ordinal in nature, and parameters, which were discrete. The Spearman rho correlation was used to determine relationships between parameters, which were interval/ordinal in nature, and other parameters, which were also interval/ordinal in nature. Fisher’s exact test was used to evaluate relationships between discrete variables. P values ⭐ 0.05 were considered statistically significant. Results

p15INK4B methylation and expression in normal marrow cells (Table 1) The level of methylation in normal unseparated bone marrow cells ranged from 0 to 6.5% with mean ± s.e. and median values of 2.6 ± 0.6% and 2.5%. Mean and median values for four matched sets of normal CD34+ and CD34− cells were indistinguishable from that of the marrow cell population as a whole being 2.3 ± 1.6% (med, 1%) for the latter. Given these data, a level of methylation of 10% or more was deemed to indicate a level which is greater than that present in normal cells. P15INK4B expression among 10 normal marrow specimens ranged from 0.26 to 0.72 with mean ± s.e. and median values of 0.43 ± 0.05 and 0.37, respectively. Table 1 provides these data; p15INK4B methylation is almost nonexistent in normal marrow cells while, as will be seen below, the level of expression of the p15INK4B gene in normal marrow exceeded that of all other specimens studied.

p15INK4B methylation in ‘post treatment’ marrows (Table 2, Figure 1) P15INK4B methylation was measured in 10 bone marrow specimens obtained from nine patients believed to have been cured of lymphoma by cytotoxic therapy, three of whom had received an autotransplant as part of their treatment. The median time from the last cytotoxic therapy to marrow donation was 2.5 years with a range of 10 months to 5 years. White blood cell counts ranged from 5.9 to 6.63 × 103/␮l (med, 6.3 × 103/␮l), the platelet counts from 16 × 103 to 287 × 103/␮l (med, 186 × 103), and Hb values ranged from 9.9 to 15.1 g (med, 12.5). Methylation of the p15INK4B gene was detected in all 10 marrow specimens ranging from 5.3% to 70.6%, with a median value of 28.8%. The level of methylation was greater than 50% in three specimens. The peripheral blood counts of four of seven of the latter patients were abnormal with low platelet counts in one, low WBC in one, elevated WBC in one and pancytopenia in one. As described in Table 1, the bone marrows of two patients in whom methylation was ⬎50% Leukemia

The p15INK4B gene in MDS and AML HD Preisler et al

1592

Table 3 Relationship among P15 methylation, expression, and MDS classification

Patient No.

Figure 1 Percentage methylation in normal, post treatment, MDS, and AML marrows. ——, median. X-axis is staggered to reveal overlapping points.

were abnormal with erythroid hyperplasia being present in one. Of the 10 marrows studied, morphologically normal marrows were present in only four. A variety of abnormalities were detected in the other six marrows including monoclonality in two, hypocellularity in three, erythroid hyperplasia in two, and trilineage dysplasia in one. Patient number 5 was studied on two occasions 6 months apart. The level of p15 methylation fell from 68.4% to 19.6%. The change in methylation was accompanied by a change in the characteristics of the marrow as described in Table 2. There was no intervening treatment.

FAB classification

p15 methylation

1 2 3 4 5 6 7 8 9 10 11

RA NA RAEB-t RA RAEB CMMoL RAEB-t RA RAEBt RAEB-t RAEB

12 13 14 15 16 17 18 19 20 21 22 23 24 25

CMMoL RARS RA RA NA RA RA CMMoL RA RA RAEB CMMoL RA RAEBt

98% 86.26% 67.6% 65.17% 60% 55.99 55.4% 49.63% 41.2% 34.62% 27% N/A 20.9% 10.73% 7.69% 0 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

p15 expression 0 0 0.04 0 0 0 0.164 0.318 0 0.449 0.13 0.139 0.131 0.114 0.216 0 0.54 0.438 0.311 0.311 0.139 0 0 0 0 0

RA, refractory anemia; RAEB, refractory anemia with excess of blasts; RAEBT, RAEB in transformation; CMMOL, chronic myelomonocytic leukemia; N/A, unknown classification or no specimen available.

p15INK4B methylation and expression in MDS marrow cells (Table 3, Figures 1 and 2) Bone marrow specimens obtained from 15 MDS patients were studied for p15 methylation and 25 studied for p15INK4B expression. The level of methylation ranged from 0 to 99%, with a median of 50%. The level of methylation was less than 50% in eight specimens with methylation being undetectable in one. In seven marrows the level of methylation was ⭓50% with transcript levels ranging from 0 to 0.54. Transcripts were not detectable in 12 marrows while significant expression (⭓0.1 units) was detected in six of 25 specimens (24%). The median level of methylation in six RA/RARS marrows was 39% with a range of 0 to 99%, being 0 in one specimen and 7.7% and 10.7% in two others. In the other three specimens, methylation values ranged from 49.6% to 98%. Expression was measured in 11 RA/RARS specimens and ranged from 0 to 0.438 with a median value of 0.114. P15INK4B transcripts were not detected in five of the 11 marrows studied. The level of p15INK4B methylation was assessed in six specimens obtained from patients with RAEB or RAEBt. Methylation was present in all six with a median of 48% and with a range from 27% to 67.6%. P15INK4B transcripts were measured in eight RAEB/RAEBt specimens with a median value of zero and a range of 0 to 0.45. Transcripts were not detected in five of the eight specimens. In summary, the level of methylation was less than 50% in eight MDS specimens with expression being detected in six. Leukemia

Figure 2 p15INK4B expression in normal, MDS, and AML marrows. ——, median. X-axis is staggered to reveal overlapping points.

In contrast, methylation was ⬎50% in seven specimens with expression detected in one. The level of p15INK4B methylation in the marrow cells of two CMMoL patients was 20.9% and 56%. Transcript levels were measured in four CMMoL marrow specimens and were found to be 0, 0, 0.131, and 0.182. Table 3 provides these data.

The p15INK4B gene in MDS and AML HD Preisler et al

Table 4 Relationship among P15 methylation, expression and AML classification

Specimen No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44

FAB classification

p15 methylation

p15 expression

M2 M2 M2 NA M2 M3 M2 M4 M1 M3 NA M2 M1 M4 M2 M1/M2 M1/M2/M3 M5 M4 NA M1 M4 M4 M4 M1 M2 M1 M4 M2 M4 NA NA M1 M1 NA NA M1 NA NA NA NA NA NA NA

99.4% 99.1 98.9% 98.4% 97.3% 94.9% 90.3% 86% 85.8% 83% 77.2% 77.1% 66.8% 65.3% 52.7% 50.8% 48.35 48.2% 44.7% 43.1% 40.4% 33.7% 29.5% 29.2% 28.7% 19.7% 13.3% 12.7% 9.3% 1.1% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% NA NA

0 0 0.149 0 0.024 0 203 0 0 0 0 0 0 0 0 0 0.233 0.572 0.525 0 0.076 0.062 0.096 0.387 0 0.013 0.546 0 0.006 0.237 0 0 0.436 0.012 0.49 0 0.187 1.116 0 0 0 0 0.136 0

NA, specimen not available for evaluation.

p15INK4B methylation and expression in AML bone marrow cells (Table 4, Figures 1 and 2) Forty-two AML bone marrow specimens were studied for methylation. Values ranged from 0 to 99% (med, 43%) with methylation not being detected in 12 specimens (20%) and being ⭓50% in 16. In 29 specimens the level of methylation was ⬎10%. P15INK4B expression was not detected in 24, was ⬎ 0 ⬍ 0.1 in seven, and was greater than 0.1 in 1.3. Thus significant expression was present in 14 of 44 AML specimens (25%). P15INK4B methylation was detected in 74% of marrow specimens obtained from patients with poor prognosis AML, while methylation was detected in 83% specimens obtained from standard prognosis AML (P = 0.7). P15INK4B expression was measured in 44 marrow specimens. Expression was detected in marrows obtained from 15 (56%) patients with poor prognosis AML with the same being the case for marrows obtained

from three (25%) standard prognosis disease. Methylation was present in the marrow cells of six of nine patients whose leukemia followed a marrow disorder and in six of 11 marrows where this was not the case (55%). P15INK4B transcripts were present in eight of 11 leukemias which had occurred subsequent to toxic exposure (73%) and in five of 10 marrows without known toxic exposure (50%). P15INK4B expression was three for the latter (30%) and two of 11 for the former (18%). Figures 1 and 2 provide a comparison of p15INK4B methylation and expression of normal marrow aspirates obtained from cured lymphoma patients and from patients with MDS or AML. Table 4 provides the numerical data for p15INK4B methylation and expression in AML marrows. Note that the level of methylation was ⬍50% among the 24 AML marrows with expression (⬎0.11) being detected in 11. In contrast, p15 transcripts were detected in two of 16 marrows in which the level of methylation was ⬎50%.

1593

Relationship of p15INK4B expression to the level of p15INK4B methylation Methylation and expression of the p15INK4B gene were measured in marrows obtained from 15 patients with MDS with the level of methylation exceeding 50% in seven (Table 2, Figure 3). Significant levels of P15INK4B transcripts were present in one of seven specimens (0.164) in which methylation exceeded 50% and in six of eight marrows in which methylation was less than 50% (P = 0.04). Similarly, p15INK4B methylation and expression were measured in 44 AML marrows with methylation exceeding 50% in 16. Significant levels of p15 transcripts (⬎0.1) were detected in only two of the latter specimens with levels being 0.024 in one. In contrast, significant levels of transcript were present in 11 of the 26 marrows in which the level of methylation was ⬍50% with no transcripts detected in 15 (P = 0.084). Considering the MDS and AML specimens together, significant levels of p15INK4B transcripts were present in three of 23 specimens in which the level of methylation exceeded 50% and in 17/34 in which the level of methylation was ⬍50% (P = 0.000). It should be noted that a methylation level of ⬍50% did not guarantee that significant expression of the p15INK4B was present. For example, considering the eight MDS marrows in which the level of methylation was ⬍50%, transcripts were not detected in two. With respect to AML marrows, transcripts were not detected in 15 of 26 marrows in which the level of methylation was ⬍50%. In fact, taking together the 12 MDS or AML marrows in which p15INK4B methylation was 0, the transcript level was ⬎0.1 in five, was 0.012 in one, and was not detectable in six. Discussion The p15INK4B gene plays an important role in the negative regulation of the proliferation of hematopoietic cells and in the prevention of malignant transformation.6 With respect to the former, p15 protein binds to and inactivates the cyclin DCDK4/5 complex whose activity is necessary for the progression of cells through the G1 phase of the cell cycle and into the S phase. Without phosphorylation of the retinoblastoma protein (RB) by this complex the E2F transcription factors remain in an inactive state bound to RB protein.10 Phosphorylation of the RB protein by the cyclin D-CDK4/5 complex Leukemia

The p15INK4B gene in MDS and AML HD Preisler et al

1594

Leukemia

results in the release of these transcription factors with initiation of progression from G0 or from early G1 to and through the restriction point (R1) in G1 phase of the cell cycle and into S phase.11 The importance of the anti-oncogene nature of the activities of the p15 and other similar genes is illustrated by the fact that under the usual conditions mutant ras genes do not transform cells in the absence of the cooperation of another proto-oncogene.12–14 However, if the p15 class of gene or another tumor suppressor gene is inactivated a single mutant ras gene suffices to cause malignant transformation.15 It has been reported that methylation of the p15 gene, and hence its inactivation, is ‘acquired during disease progression’.4,5 The paper by Quesnel and colleagues4 reported that methylation of the gene was only detected once the percent of blast cells in the marrow exceeded 10%, that is only after an aggressive phase of MDS had appeared (RAEB or RAEBt). The authors of the second paper, Uchida et al,5 assessed the activity of the p15 gene by measuring the level of methylation of the gene, assuming that a level of methylation of 20% or greater indicated that the gene was not being transcribed. Our data are very different from those described above. With respect to the report of Quesnel et al, not only did we detect methylation in the marrows of patients with early MDS (RA/RARS), but we also detected significant levels of p15 methylation in the marrows of patients who are at high risk of developing MDS, but in whom MDS has not as yet developed (these individuals will be discussed in detail below). With respect to the studies of Uchida and colleagues,5 in the studies described here p15 transcripts were not detectable in one of three RA or RARS marrows in which the level of methylation was ⬍20%. Further, while six of six marrows obtained from patients with RAEB/RAEBt demonstrated levels of methylation exceeding 20%, p15 transcripts were nevertheless detected in two of the six marrows. With respect to the situation in AML, p15 methylation levels exceeding 20% were detected in 27 marrows with p15 transcripts being present in five. Conversely, methylation levels of less than 20% were present in 14 AML marrows and yet p15 transcripts were not detected in six. Hence, there must be mechanisms independent of methylation which regulate p15 expression. These data clearly show the perils of making assumptions about gene expression, when expression itself is not being measured. In our experience, a level of methylation of 50% is highly likely to indicate that p15 transcripts will not be detected. However, expression can occasionally be found at these levels of methylation. Perhaps, the most unexpected data were those obtained during the study of patients who had been cured of a lymphoma by means of treatment regimens which included cytotoxic chemotherapy. This group of individuals have long been known to be high risk of developing a secondary hematologic disorder, especially those in whom an autotransplant was performed during their treatment.2,3 The latter was the case for seven of the patients whom we studied. When last seen by their physicians the patients had been found to be well with normal peripheral blood counts. When studied by us, the blood counts of two of these individuals were found to be abnormal (patients 5 and 6, Table 2). Methylation of the p15 gene was found to be present in every bone marrow studied with the level ranging from 5.3% to 70.6%. Methylation levels exceeding 50% were present in the marrows of three patients. It is of special interest that the bone marrow was found to be completely normal in only three patients. Among the seven other patients, monoclonality was found to be present in two,

hypercellularity in one, hypocellularity in three, hyperplasia in one or more cell lineages in two, and trilineage dysplasia in one. Taken together, the data discussed in the preceding paragraph indicate that the bone marrow was abnormal in eight of 10 individuals who had been cured of lymphoma in the past. While a diagnosis of myelodysplastic syndrome could be made for patient 6, this was not the case for the other seven individuals. Nevertheless, the marrows of all seven of these individuals were abnormal with changes that are consistent with a high likelihood that the seven individuals are at extremely high risk of developing a clinically apparent secondary hematologic disorder. The observations that the marrow abnormalities became less in one of the patients over time (5a, b, Table 2) suggests that in some individuals the detection of abnormalities in a single marrow examination should not be taken to be an absolute indication that a secondary hematologic disorder is inevitable. The detection of significant levels of methylation in the marrows of cured lymphoma patients, together with the observations that mutations of the N-ras and c-fms genes are especially common in these individuals,16,17 suggests a possible mechanism underlying the development of monoclonal hemopoiesis. If a mutation is acquired by a cell in which the normal p15 restraint on abnormal cell proliferation is no longer present (because of the methylation of the p15 gene), then it is not unlikely that monoclonal hemopoiesis will appear. Further, it is likely that as cell numbers increase beyond normal levels there will be an increase in the production of several negative regulators of proliferation in the marrow, including TGF␤.18 Since TGF␤ suppression of proliferation is mediated via the p15 pathway,19 the absence of functional p15 genes in the newly emerging cells will protect these cells against the growth inhibitory effects of TGF␤ while the normal cells in the marrow remain sensitive to the proliferation inhibitory effects of this cytokine. This difference in TGF␤ effect on normal cells and on cells in which the p15 genes have been inactivated increases the proliferative advantage which the newly emerging cell clones enjoy over their normal competitors. As illustrated by patient 5a, b in Table 2, the level of p15INK4b methylation fell from 68.4% to 19.6% over a period of 6 months without intervening treatment. Additionally we found that the level of p15INK4b expression increased threefold in the AML marrow cells in vivo in a patient who was treated with 5-azacytidine for 3 days. These observations demonstrate that it is therapeutically possible to increase expression of the p15INK4b gene and perhaps to prevent or reverse the progression of the secondary hematologic disorders. Further, it will be of interest to determine if an increase in p15 expression will induce the apoptosis of cells carrying mutant oncogenes.

Acknowledgements This work was supported by the National Cancer Institute grant 1-PO-1CA75606-04.

References 1 Preisler HD. Fitness landscapes and the myeloid leukemias. Leukemia Res 1999; 23: 167–176. 2 Darrington DL, Vose JM, Anderson JR, Bierman PJ, Bishop MR,

The p15INK4B gene in MDS and AML HD Preisler et al

3 4

5 6 7

8

9 10

Chan WC, Morris ME, Reed EC, Sander WG, Tarantole SR. Incidence and characterization of secondary myelodysplastic syndrome and acute myelogenous leukemia following high-dose chemoradiotherapy and autologous stem-cell transplantation for lymphoid malignancies. J Clin Oncol 1994; 12: 2527–2534. Stone RM. Myelodysplastic syndrome after autologous transplantation for lymphoma: the price of progress? Blood 1994; 83: 3437–3440. Quesnel B, Guillerm G, Vereecque R, Watte E, Preudhome C, Bauters F, Vanrumbeke M, Fenaux P. Methylation of the p15 INK4b gene in myelodysplastic syndromes is frequent and acquired during disease progression. Blood 1998; 91: 2985–2990. Uchida T, Kinoshita T, Nagai H, Nakahar Y, Saito H, Hott AT, Murate T. Hypermethylation of p14 (INK4B) gene in MDS. Blood 1997; 90: 1403–1409. Herman JB, Jen J, Merlo A, Baylin SB. Hypermethylation-associated inactivation indicates a tumor suppressor role for p15 INK4B1. Cancer Res 1996; 56: 722–727. Bennett JM, Begg CB. Eastern Cooperative Oncology Group study of the cytochemistry of adult acute myeloid leukemia by correlation of subtypes with response and survival. Cancer Res 1981; 41: 4833–4837. Preisler HD, Larson RA, Raza A, Browman G, Goldberg J, Vogler R, Day R, Gottlieb A, Vardiman JW, Bennett J, Kukla C, Grunwald H. The treatment of patients with newly diagnosed poor prognosis acute myelogenous leukemia: response to treatment and treatment failure. Br J Haematol 1991; 79: 390–397. Herman JG, Graff J, 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. Sellers WR, Kaelin WG. Role of the retinoblastoma protein in the pathogenesis of human cancer. J Clin Oncol 1997; 15: 3301– 3312.

11 Horton LE, Qian Y, Templeton DJ. G1 cyclins control the retinoblastoma gene product growth regulation activity via upstream mechanisms. Cell Growth Diff 1995; 6: 395–407. 12 Parada LF, Land H, Weinberg RA. Cooperation between genes inducing p53 tumor antigen and ras in cellular transformation. Nature 1984; 312: 649–651. 13 Land H, Parada LF, Weinberg RA. Tumorigenic conversion of primary embryo fibroblasts requires at least two cooperating oncogenes. Nature 1983; 304: 596–602. 14 Preisler HD. A hypothesis regarding the development of acute myeloid leukemia from preleukemic disorders: the role of protooncogenes. Cancer Genet Cytogenet 1988; 321: 133–142. 15 Tanaka N, Ishihar M, Kitagawa M, Harada H, Kimura T, Matsuyama T, Lamphier MS, Aizawa S, Mak TW, Tanguchi T. Cellular commitment to oncogene-induced transformation or apoptosis is dependent on the transcription factor IRF-1. Cell 1994; 77: 829–839. 16 Carter G, Ridge S, Padua RA. Genetic lesions in preleukemia. Crit Rev Oncol 1992; 3: 339–364. 17 Preisler HD, Li B, Yang J, Horvath E, Raza A, Loew J, Gregory S, Hsu W-T, Showel J, Adler S, Bi S. The premyelodysplastic state and secondary hematologic disorders. In: Hiddemann W, Buchner T, Wormann B, Ritter J, Creutzig U, Keating M, Plunkett W (eds). Acute Leukemia VII Symposium published in Acute Leukemias VII, Experimental Approaches and Novel Therapies. Springer Verlag: Berlin, 1998; pp 102–107. 18 Blobe GC, Schiemann WP, Lodish HF. Role of transforming growth factor ␤ in human disease. N Eng J Med 2000; 342: 1350–1358. 19 Hannon G, Beach D. P15ink4b is a potential effector of TGF-␤induced cell cycle. Nature 1994; 371: 257–261.

1595

Leukemia