Quantitative detection of DNMT3A R882H mutation in acute myeloid leukemia

Abstract

Background: DNMT3A mutations represent one of the most frequent gene alterations detectable in acute myeloid leukemia (AML) with normal karyotype. Although various recurrent somatic mutations of DNMT3A have been described, the most common mutation is located at R882 in the methyltransferase domain of the gene. Because of their prognostic significance and high stability during disease evolution, DNMT3A mutations might represent highly informative biomarkers for prognosis and outcome of disease.

Methods: We describe an allele-specific PCR with a Blocking reagent for the quantitative detection of DNMT3A R882H mutation providing the possibility to analyze the quantitative amount of mutation during the course of disease. Next, we analyzed 62 follow-up samples from 6 AML patients after therapy and allogeneic stem cell transplantation (alloSCT).

Results: We developed an ASB-PCR assay for quantitative analysis of R882H DNMT3A mutation. After optimization of blocker concentration, a R882H-positive plasmid was constructed to enhance the accuracy of the sensitivity of quantitative detection. The assay displayed a high efficiency and sensitivity up to 10(-3). The reproducibility of assay analyzed using follow-up samples showed the standard deviation less than 3.1 %. This assay displayed a complete concordance with sequencing and endonuclease restriction analysis. We have found persistence of DNMT3A R882H mutations in complete remission (CR) after standard cytoreduction therapy that could be indicating presence of DNMT3A mutation in early pre-leukemic stem cells that resist chemotherapy. The loss of correlation between NPM1 and DNMT3A in CR could be associated with evolution of pre-leukemic and leukemic clones. In patients with CR with complete donor chimerism after alloSCT, we have found no DNMT3A R882H. In relapsed patients, all samples showed an increasing of both NPM1 and DNMT3A mutated alleles. This suggests at least in part the presence of NPM1 and DNMT3A mutations in the same cell clone.

Conclusion: We developed a rapid and reliable method for quantitative detection of DNMT3A R882H mutations in AML patients. Quantitative detection of DNMT3A R882H mutations at different time points of AML disease enables screening of follow-up samples. This could provide additional information about the role of DNMT3A mutations in development and progression of AML.

Fig. 1

Assay design of ASB-PCR. Primers and probe are located in exon 23 of DNMT3A. The allele-specific primer contains the mutational spot at its 3′-end whereas the wt spot is incorporated in the middle of the blocker (red box). Fluorescence detection was performed with a TaqMan probe which was designed near to the forward primer (13 bp distance)

Fig. 2

Performance of ASB-PCR. ai Qualitative analysis of the specificity of ASB-PCR. Enhancement of blocker concentration to 80 μM showed no significant change in the amplification properties. aii Quantitative analysis of Ct-changes induced by addition of 40 μM blocker. b Quantitative analysis displayed a C_t difference of 11–14 cycles between R882H mutated DNA and wt DNA. NTC, non-template control; wt, wild type

Fig. 3

Analysis of PCR sensitivity using serial dilutions of DNA (a) and (b) dilution of mutated DNA with wt DNA

Fig. 4

Sequence of R882H plasmid. a Heterozygote R882H mutation in a patient sample. The mutational spot is indicated by the red box. Both, guanine (wt) and adenine (mutation) were detected. b Homozygote R882H mutation in the generated plasmid. The mutational spot is indicated by the red box. The sequence displays only the mutated adenine base and no wt guanine base

Fig. 5

Absolute quantification of DNMT3A R882H mutations. Values of ASB-PCR are listed in the corresponding table. a Analysis of copy number dilutions of the R882H homozygote plasmid. Patient samples with approximately 50 % of R882H mutated allele are indicated by dotted lines; the 10° plasmid dilutions are pink line. b The standard curve (copy numbers over cycle). Patient samples with 50 % R882H are shown as a blue diamond and a green triangle. Wt samples are shown as red quadrats. R² = correlation coefficient. B = Intercept with the ordinate. M = Slope of the standard curve. c Absolute quantification was performed by percentage of mutated allele in a sample. Plasmid concentrations were adjusted to 50 % and 25 % of mutated R882H allele according to the measurement of heterozygote patient samples (dotted lines). d Standard curve (%R882H over cycle). The dotted red line shows the limit of 100 % R882H mutated allele in a sample. Patient samples are shown as a blue diamond and a green triangle. wt samples are shown as red quadrats. R² = correlation coefficienct. B = Intercept with the ordinate. M = Slope of the standard curve

Fig. 6

Concordance of ASB-PCR. a Representative endonuclease restriction analysis of follow-up samples of five patients (A-E). Wild type samples show two bands at 190 bp and 114 bp. Positive samples display three bands at 289 bp, 190 bp, 114 bp due to the loss of a restriction site of Fnu4HI caused by the mutation. Hyperladder II (Bioline) was used as marker. b Results of ASB-PCR analysis