Anticorresponding p15 promoter methylation and microsatellite instability in acute myeloblastic leukemia

To the editor:

We read with interest the paper on p15promoter methylation in adult and childhood acute leukemias.1 Wong et al used methylation-specific polymerase chain reaction (MSP) to analyze p15promoter methylation patterns in acute leukemia. Their study included an analysis of 38 adults and 4 children with acute myeloid leukemia (AML). In this group they found higher p15methylation frequencies in patients with the M3, M4, M2, and M7 French-American-British (FAB) subtypes than in those with the M1, M6, or M5 subtypes. They also noted an association betweenp15 methylation and chromosomal translocations, inversions, and deletions, suggesting an interplay of these abnormalities andp15 methylation, but the number of patients with these abnormalities and p15 methylation in their cohort of patients with AML was small (n=3).

We have retrospectively analyzed p15methylation in 73 patients with AML aged 16 to 96 years (median age 63 years). Our study also utilized MSP using the same primer sets as Wong et al. Data relating to FAB type are available on 55 of our cases (Table). Unlike Wong et al, we failed to find any association between p15methylation frequencies and FAB subtype. We believe that the discrepancies between our results and those of Wong et al reflect the small numbers of patients involved in both studies and suggest that no correlations should be drawn between AML subtype and p15methylation status until a larger cohort of patients has been analyzed.

We also have cytogenetic data available on 53 of our patients. These have been analyzed according to the 3 prognostic cytogenetic risk groups defined in the Medical Research Council AML 10 trial.2 According to these criteria, the inv(16)(q13q22), t(15;17)(q22;q12) and t(8;21)(q22;q22) abnormalities described by Wong et al are characterized as good-risk abnormalities. In our study, 0 of 3 patients with good-risk cytogenetics, 19 of 34 with intermediate-risk cytogenetics, and 8 of 16 with poor-risk cytogenetics had evidence ofp15 methylation. The figures were compared using a Pearson chi-square test, and no correlation was found between p15methylation status and cytogenetic risk group (P = .178).

We have also examined the relationship betweenp15 methylation and functional evidence of a mismatch repair (MMR) gene defect as detected by microsatellite instability (MSI) using a panel of 11 microsatellite markers to compare constitutional DNA derived from mouthwash samples with that of leukemic DNA from the same patients.3 An analysis of MSI was successful in 61 of the 73 patients whose p15 methylation status was known. Six patients were found to have MSI, and this was shown to be associated with abnormal MSH2 protein expression as detected by Western blotting. Of the 6 MSI-positive cases, none had evidence of p15methylation. In contrast, 30 of 55 MSI-negative cases hadp15 methylation. Although the number of patients with MSI is small, the inverse correlation between MSI and p15methylation is statistically significant using the Fisher exact test (P = .024). These results represent an interesting and novel finding. p15 promoter hypermethylation appears to be a common late event in the pathogenesis of AML, whereas MMR gene defects are rare. The finding that MMR gene defects, when present, are not associated with p15 promoter hypermethylation implies a unique disease pathogenesis in those patients with MMR gene defects. In such cases leukemia is likely to evolve through novel pathways involving MMR gene defects and subsequent mutations in genes regulating growth and apoptosis independently of methylation defects.