Therapy-Related Myeloid Neoplasms (tMN) Following Treatment of Acute Myeloid Leukemia (AML): Exome Sequencing Reveals the Presence of an Ancestral Clone Refractory to Chemotherapy

The WHO recognizes tMN as a distinct disease entity. The majority of tMN arise after treatment of solid tumors with cytotoxic chemotherapy or radiation therapy. In many tMN patients, clonal hematopoiesis was demonstrated to be already present at diagnosis of the solid tumor and later gave rise to the tMN. Cases of tMN following successful treatment of AML are rare and poorly characterized. We therefore set out to study the genetic profile of these patients and evaluated if AML and subsequent tMN are two genetically distinct diseases or share a common ancestral clone.

We selected cases of AML sent to our laboratory for diagnostic work-up from 2005 to 2018 who subsequently developed a tMN. To exclude events of AML relapse from the cohort, known AML driver mutations such as in NPM1 or AML defining rearrangements present at diagnosis (Dx) of AML had to be absent at Dx of tMN. Based on sample availability, whole exome sequencing (WES) was carried out for 25 pts at the time points of Dx of AML and Dx of tMN as well as on matched remission samples that were available for 19 pts. Enrichment based library preparation was performed using the xGen Exome Research Panel and sequenced (2x151bp) on a NovaSeq instrument. Data was processed with BaseSpace using the BWA Enrichment app with BWA for Alignment (against hg19) and GATK for variant calling with default parameters. Data was subsequently loaded into BaseSpace Variant Interpreter to filter and prioritize variants of interest. Only passed protein changing variants were considered with an ExAC population frequency of less than 1% for further analysis.

The cohort included 13 females and 12 males, aged 28 to 77 years (median: 60 yrs). Median time to Dx of tMN following Dx of AML was 42 months (range: 8-96 months). Thirteen pts had developed tMDS and 12 pts tAML. While WES identified numerous recurrently mutated genes, we focused our further analyses on 28 genes associated with myeloid neoplasms and detected a total of 97 mutations (1-8 mutations per patient, median: 4). Following genes were mutated in >15% of pts: NPM1 (14/25 of pts), DNMT3A (9/25), TET2 (9/25), SRSF2 (7/25), CEBPA (6/25), RUNX1 (6/25), ASXL1, FLT3, IDH2 and TP53 (4/25 of pts, each). We did not identify a mutation pattern distinguishing between pts who developed a tAML or tMDS, however, pts with tAML had a significantly higher total number of mutations than those with tMDS (median of 5 vs 3 mutations, p= 0.022). When comparing matched initial AML and subsequent tMN samples for each patient, a total of 43 mutations were detected at Dx of AML only, while 25 mutations were exclusive to the tMN sample. Interestingly, 18/25 (72%) of pts harbored mutations present at both time points (Figure 1). Moreover, in 8/22 of pts with available material, mutations in DNMT3A (n=6), IDH1, SRSF2 and TET2 (n=1, each) persisted during remission. Of the 7 pts without shared mutations between time point of AML and tMN, 2 were initially positive for the PML-RARA rearrangement, and 1 patient showed a CBFB-MYH11 rearrangement at Dx of AML. Next, we compared gene mutation data with cytogenetic information available for 21/25 pts at both time points. Nine patients with a normal karyotype (NK) at Dx of AML developed chromosome aberrations at Dx of tMN. In this group, 8/9 pts harbored mutations in epigenetic modifiers (DNMT3A, TET2 and SRSF2) present both at AML and tMN. Similarly, 2/3 patients with NK AML and NK tMN shared mutations in IDH2, SRSF2 and ASXL1 at both time points. Eight pts showed either independent cytogenetic aberrations between AML and tMN or harbored chromosome aberrations at Dx of AML that were absent at Dx of tMN. Even in this group of cytogenetically independent clones, 3/8 pts harbored the same mutations at both time points.

Our study provides novel insights into the pathogenesis of tMN. In the majority of analyzed cases, we detected mutations present in the AML and tMN clones of the same patient. Considering the absence of known AML driver gene mutations such as in NPM1 in the tMN sample, we speculate the presence of a common ancestral clone with mutations mostly affecting epigenetic modifiers which have previously been linked to clonal hematopoiesis. Unlike the full-blown leukemia, this clone is refractory to chemotherapy and later gives rise to the tMN. The absence of shared mutations between AML and tMN in patients with PML-RARA or CBFB-MYH11 rearrangement is in line with thought that these entities are true de novo AML that do not evolve from a clonal hematopoiesis.

View largeDownload slide

View largeDownload slide

Close modal