Droplet Digital PCR for DNMT3A and IDH1/2 Mutations to Improve Early Diagnosis of Acute Myeloid Leukemia Relapse after Allogeneic HSCT

INTRODUCTION:Despite the considerable improvement documented over the last two decades in the outcome of allogeneic Hematopoietic Stem Cell Transplantation (allo-HSCT) for Acute Myeloid Leukemia (AML), primary disease relapse still represents the main cause of mortality in transplanted patients. Since most of the available therapies for post-transplantation relapse display very limited activity when enacted in overt hematologic recurrence, efforts are aimed to anticipate relapse detection and treatment to the Minimal Residual Disease (MRD) stage. Still, the genetic heterogeneity and extensive clonal evolution which are distinctive features of AML hinder the identification of reliable MRD markers. Recent studies demonstrated that mutations in the DNMT3A and IDH1/2 genes occur very early during the step-wise process of leukemogenesis, possibly representing disease founder mutations, shared by all disease subclones and maintained throughout the patient longitudinal history. Moreover, by being present both in full-fledged transformed cells and their progenitors, their tracking might provide a wider scope on the efficacy of allo-HSCT in eradicating preleukemic stem cells.

METHODS: We took advantage of ultra-sensitive droplet digital PCR (ddPCR) to test a total of 52 bone marrow samples collected longitudinally over time from 17 patients who received myeloablative allo-HSCT for AML. All patients carried at least one mutation amongst DNMT3A R882H, IDH1 R132C, IDH1 R132H, IDH2 R140Q and IDH2 R172K, documented at diagnosis by conventional Sanger sequencing. As controls, we tested bone marrow samples collected at diagnosis from 7 patients typing negative for the mutations, and peripheral blood samples from 8 healthy individuals. ddPCR assays were performed using the Bio-Rad QX100 system: each sample was tested in duplicates, employing 25 ng of genomic DNA in each reaction well and using as reference for each mutation-specific assay the respective wild-type allele. Samples with a mutant-to-wild-type ratio above 0.1% were considered positive. ddPCR results were compared to those obtained testing the same samples by quantitative PCR (qPCR) assessment of the WT1 gene transcript (considering as threshold for relapse prediction 250 copies of WT1/10⁴ copies of ABL) and by qPCR-based hematopoietic chimerism assessment (employing the AlleleSEQR Chimerism Assay and considering as threshold for relapse prediction a host-specific signal above 1%).

RESULTS:All the 17 samples collected at diagnosis and typing positive for the mutations of interest by conventional Sanger sequencing resulted positive also for the corresponding ddPCR assay. None of the samples from healthy individuals or from patients typing negative for the mutations resulted positive by ddPCR. All the samples tested at post-transplantation relapse remained positive for the mutations present at diagnosis, except for one case, originally carrying both DNMT3A and IDH2 mutations and typing negative for the latter at relapse. This observation might argue against the putative role of IDH mutations as leukemia-founder events, and suggests that, when present, DNMT3A could represent a more reliable MRD marker. In samples harvested in overt leukemia, the population carrying the mutant allele, quantified by ddPCR, consistently exceeded the morphological count of leukemic blasts. When post-transplantation remission samples were tested, 32/32 (100%) of those harvested from patients who remained long-term leukemia-free (median follow-up after allo-HSCT: 19 months) resulted negative for the mutations of interest, whereas 3/5 (60%) of those from patients who subsequently relapsed resulted positive. Of notice, only 1 of those 5 samples displayed WT1 transcript overexpression and host chimerism above the 1% threshold, whereas the remaining 4 resulted negative by both qPCR-based techniques.

CONCLUSIONS: Although the very small number of patients included in this preliminary analysis warrants for caution, ddPCR for DNMT3A and IDH1/2 mutations appears extremely promising, displaying optimal specificity and very high sensitivity in relapse prediction, and comparing favorably with our present and historical results obtained by qPCR-based post-transplantation monitoring techniques.