Aberrant hypermethylation of tumor suppressor genes plays an important role in the development of acute myeloid leukemia (AML). In order to investigate the expression pattern of DNA methyltransferases (DNMTs) in AML patients, we established a tissue microarray (TMA) on bone marrow biopsies of 127 patients diagnosed with AML, who received intensive treatment in our department between 1999 and 2008 and with available bone marrow histology specimens. Immunohistochemistry was performed on this TMA with antibodies against DNMT1, DNMT3A, DNMT3B, methylated cytosine (mC) and hydroxymethylated cytosine (hmC).
127 patients having received intensive treatment in our department and with available clinical follow-up data were analyzed. Median follow-up after diagnosis was 56 (range 0–133) months. Median age at diagnosis of these patients cohort was 63 (range 21 – 85) years, 52 patients (41%) were female, 75(59%) were male. 88 patients (70%) had de novo AML, 38 (30%) had secondary AML (sAML) evolving from antecedent myelodysplastic syndrome or after previous cytotoxic treatment. For one patient the information was not available. Cytogenetic and molecular genetic risk assessment was performed according to the classification published by the European Leukemia Network (ELN). Of the 120 patients (95%) with available results, 20 patients (16.7%) were low risk, 49 (40.8%) intermediate-I, 20 (16.7%) intermediate-II and 31 (25.8%) high risk.
Of 126 patients with evaluable DNMT1 expression, only 9 patients (7%) had detectable expression in ≥5% positive blasts. DNMT1 expression did not associate with any outcome parameter. DNMT3A expression analysis could not be established due to unspecific staining. By contrast, DNMT3B expression was positive (≥5%) in 80 of 116 evaluable patients (69%). Patients with absent or low (positivity in <5% of blasts) DNMT3B expression had a higher pretreatment white blood cell count (WBC) (p= 0.046), a higher peripheral blood blast percentage before treatment (p= 0.02) and a more favorable cytogenetic and molecular risk profile (p=0.002) compared to patients with high DNMT3B expression. Patients with absent or low DNMT3B expression showed a slightly, but not significantly lower positivity for mC than patients with high DNMT3B expression (mean 54 versus 50% positive blasts, p=0.56) and a trend for lower hmC positivity (mean 47 versus 56%, p=0.17). Complete remission rates were not different between low/absent or high DNMT3B expression (69 versus 62%, p=0.27). DNMT3B low/absent expression was associated with a significantly improved overall (median 42 versus 10 months, p=0.014), event-free (median 13 versus 5 months, p=0.018) and relapse-free survival (64 versus 7 months, p=0.025) compared to patients with DNMT3B positivity in ≥5% bone marrow blasts. In multivariate Cox regression analyses with the inclusion of other risk factors associated with prognosis in our cohort (age, sex, ELN risk group, de novo versus secondary AML, WBC), absent or low DNMT3B expression remained an independent favorable risc factor for overall (proportional hazard risk (HR) 0.47, 95% confidence interval (CI) 0.26 – 0.85), event-free (HR 0.61, 95% CI 0.38 – 0.97) and relapse-free survival (HR 0.48, 95% CI 0.25 – 0.92).
Taken together, DNMT3B protein expression is an independent prognostic factor in patients with AML in our cohort. If validated, immunohistochemical detection of DNMT3B protein expression might become a valuable tool for prognostication in patients with AML.
No relevant conflicts of interest to declare.
This icon denotes a clinically relevant abstract
Asterisk with author names denotes non-ASH members.