Dynamic Changes in the Clonal Structure of MDS and AML in Response to Epigenetic Therapy

Hematopoietic cells from patients with myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML) contain gene mutations that are variably distributed between the founding clone and daughter subclone(s). Traditional response criteria in MDS and AML are based on bone marrow morphology and may not accurately reflect antitumor activity and clinical benefit in patients treated with hypomethylating agents. We used digital sequencing of serial bone marrow samples to monitor tumor burden and to characterize the changes in the clonal structure of MDS and AML that occur during treatment with epigenetic therapy. We hypothesized that digital sequencing may provide an alternative measure of antitumor activity and identify the persistence or emergence of resistant clones during treatment which mediate disease relapse.

We conducted a phase I/II study in older adults (age ≥ 60) with advanced MDS (IPSS ≥ 1.5) or AML. Subjects received a combination of decitabine 20 mg/m² on d1-5 with the histone deacetylase inhibitor, panobinostat 10-40 mg po 3x/week every 28 days for up to 12 cycles. Serial bone marrow samples were collected for digital sequencing at baseline, after every 2 cycles of treatment and at the time of relapse.

A total of 52 patients, 14 with MDS and 38 with AML were enrolled in this study. For AML patients, 10% achieved a complete remission (CR+CRi) with an additional 18% of patients achieving a morphologic leukemia-free state (mLFS) using IWG response criteria. For patients with MDS, 14% achieved a CR and 21% achieved a marrow CR. We identified 9 MDS and 16 AML patients that had banked, paired bone marrow and skin (as a source of normal DNA) samples and a somatic mutation in at least 1 of 54 recurrently mutated MDS/ AML genes. DNA was enriched for 285 genes commonly mutated in MDS and AML (n=24 patients) or whole exome probes spiked-in with the 285 genes (enhanced exome sequencing; EES) (n=7 patients), and sequenced on a HiSeq2000 instrument with 2x101bp reads. We detected an average of 4.9 SNVs and indels per patient (range 1-15) when only the 285 gene panel was used, compared to 27.4 mutations per patient (range 9-43) using EES. Ten genes were mutated in at least 3 pre-study samples. The presence of a TP53 mutation (N=8) was associated with a trend towards achieving a response (p=0.09).

We then analyzed variant allele frequencies (VAF) of mutations in serial samples. We observed five distinct patterns that were associated with different clinical responses, including i) AML patients achieving a CR+CRi (n=2): mutation VAFs were undetectable by cycle 2 using standard sequencing, ii) AML with mLFS (n=2): mutation VAFs remained detectable but decreased to <10%, iii) MDS with CR/cCR+mCR (n=3): mutation VAFs decreased to <10% and were intermittently below the level of detection, iv) MDS with stable disease (n=2): mutation VAFs decreased but some remained >10%, and v) AML with treatment failure (n=5): mutation VAFs were essentially unchanged and remained >30%.

We observed responding patients can have persistent measurable clonal hematopoiesis for at least one year without disease progression. Sequencing also revealed selective AML subclone clearance in a patient with treatment failure, nominating a set of mutations that may mark super-responder clones. We observed that the blast percentage decreases prior to mutation VAFs in some patients, suggesting that the differentiation of blasts could falsely underestimate tumor burden. Finally, sequencing revealed that tumor burden can be measured even in patients achieving a CR. Using an ultra-sensitive barcode sequencing approach, we sequenced 1 MDS and 1 AML patient achieving a clinical and molecular CR (based on standard sequencing). We detected extremely rare TP53 mutations months to years prior to disease relapse (VAFs = 0.23% in MDS and 0.05% in AML during a CR - equivalent to a sensitivity of 1 in 2000 heterozygous mutant cells). While patients can live with persistent clonal hematopoiesis in a CR or stable disease, ultimately we find evidence that expansion of a rare subclone drives relapse or progression from MDS to secondary AML. Digital sequencing provides an alternative measure of disease response which may augment traditional clinical response criteria and should be explored in future clinical trials.