Sensitive Monitoring of Minimal Residual Disease Status in CEBPA-Mutated Acute Myeloid Leukemia Using Amplicon Deep-Sequencing

CEBPA (CCAAT/enhancer binding protein alpha) encodes a member of the basic region leucine zipper (bZIP) transcription factor family essential for myeloid differentiation. CEBPA mutations occur predominantly in AML with a normal karyotype and CEBPA mutated AML has been included as provisional entity in the WHO classification. Cases with biallelic mutations were reported as being associated with a favorable clinical outcome, thus patients are spared from allogeneic transplantation in first CR. Screening for CEBPA mutations in patients with AML is often performed applying a combination of fragment length analysis, DHPLC and subsequent direct sequencing using Sanger technique (conventional methods). Study Design: Next-generation amplicon deep-sequencing (454 Life Sciences, Branford, CT) is a more sensitive quantitative detection method than Sanger sequencing and thus was used to analyze 144 samples from 29 CEBPA mutated AML patients with a normal karyotype. For a longitudinal analysis starting at diagnosis and following the course of treatment bone marrow (n=134) or peripheral blood (n=10) samples were obtained between 5/2006 and 6/2011. The sequencing assay targeted the complete coding region of CEBPA, covered with 4 amplicons, and was performed using genomic DNA extracted from mononuclear cells. In median, 711 reads per amplicon were generated using the NGS assay, thereby allowing a sensitive quantitative assessment of the CEBPA mutational burden in order to monitoring minimal residual disease (MRD). In median, 4 time points per patient (range: 2–9) were included with a median time span of 9.5 months (range: 1–45 months). The median sampling interval was 2 months (range: 0.3–45 months). Results: First, we evaluated the concordance of mutation detection by comparing data from NGS and conventional methods using the samples at initial diagnosis. In all 29 AML patients NGS concordantly detected the mutations known from conventional methods, i.e. in total 26 frame-shifts, 15 in-frame alterations, 8 missense, and 2 nonsense mutations. Further, at initial diagnosis, deep-sequencing detected the mutations with a median burden of 44% sequencing reads (range 3%–88%) and thus already allowed a quantitative assessment of the mutational load. There was no difference observed for 6 patients with monoallelic vs. 21 cases with biallelic mutations (excluding 2 cases with homozygous alterations). We next investigated the distribution of clones and their underlying kinetics of clone size reduction during subsequent high-dose chemotherapy cycles. Overall, 26/29 cases were evaluable and the clone size was assessed by NGS at the second analysis point during course of disease–in median 63 days from time of diagnosis (range 10–215 days): (i) In 16/26 cases, deep-sequencing was not able anymore to detect the mutations as observed at diagnosis. 14 of these 16 negative cases stayed in complete molecular remission till the end of follow-up (median follow-up 6.5 months, range 1–34.2; 2/14 cases with allogeneic stem cell transplantation). (ii) Interestingly, in 4/26 cases residual disease with clones ranging from 8%–50% was indicative of non-response to treatment. In this subgroup 3/4 patients were characterized by resistant disease or early relapse (1 case excluded due to short follow-up). (iii) In the remainder group of 6/26 patients with mutations still detectable in a range of 0.12%–3.7%, complete molecular remission status was achieved at subsequent time points. However, in this group also 3 relapses were observed including 2 cases with allogeneic stem cell transplantation. Of note, in 3/6 cases from the latter group, NGS had outperformed conventional methods and was able to still detect residual clones enabling a superior monitoring of therapy response. In all cases with biallelic mutations both clones responded in parallel with similar kinetics. Moreover, 5 patients were investigated following relapse of AML or non-response to therapy. In all 5/5 analyses including 2 monoallelic and 3 biallelic alterations the same mutations as harbored at initial diagnosis remained detectable. Conclusion:CEBPA mutations provide increasing clinical utility for the detection of MRD. We here demonstrated that deep-sequencing is a suitable unbiased and robust method to accurately detect and quantify CEBPA aberrations enabling an individualized monitoring of disease status and treatment efficacy.

Kohlmann:MLL Munich Leukemia Laboratory: Employment; Roche Diagnostics: Honoraria. Grossmann:MLL Munich Leukemia Laboratory: Employment. Fasan:MLL Munich Leukemia Laboratory: Employment. Stopp:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Schindela:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.