Abstract
Introduction: Clonal evolution has been described in various hematologic malignancies. Persistence of one founding clone is the basis for relapse of AML, and diversity of accompanying mutations greatly varies in AML [Ding et al. Nature 2012]. On the contrary, a targeted sequencing approach revealed that in myeloproliferative neoplasia, few additional mutations arise during progression [Lundberg et al. Blood 2014]. The dynamics of clonal evolution has been linked to response to Ruxolitinib therapy [Newberry et al. Blood 2017].
Patients and methods: Here, we describe the clonal evolution of 30 pts (median age at alloSCT 60, range 36-75 years) with myelofibrosis (16 primary, 6 post-ET, 7 post-PV MF) who either progressed directly (n=14) or after a period of complete donor hematopoiesis for 417 days (median, range 7-2240 days) after allogeneic stem cell transplantation. Amplicon-based next generation sequencing of a custom panel consisting of 17 different genes (178 amplicons in 2 primer pools, targeting regions in TET2, IKAROS, JAK2, MPL, SH2B3, CBL, IDH1, IDH2, ASXL1, EZH2, SRSF2, SF3B1, DNMT3A, RUNX1, FLT3, NRAS, KRAS) was performed for all pts using a Personalized Genome Machine (PGM, ThermoFisher/ LifeTechnologies). Additional whole exome sequencing was performed for 6 out of these 30 pts, both before alloSCT and at relapse of the disease after alloSCT, using a Proton (ThermoFisher/ LifeTechnologies). Mutations in CALR were determined using digital PCR.
Results: Pts carried JAK2 V617F (n=23), CALR L367fs* (n=3), CALR K368fs*, CALR K385fs* (each n=1), or were triple-negative (n=2). The mutations in the driver genes JAK2, CALR, and MPL that were present before alloSCT were also present at relapse. One pt carrying a JAK2 V617F before alloSCT had developed an additional MPL L494V missense mutation at relapse. With targeted amplicon sequencing, a median of 3 nonsynonymous mutations (range 0-13) was found in a median of 2 (0-8) genes before alloSCT. Before alloSCT, 19 pts were mutated in TET2 (I1762V/ L1721W/ P363L/ L34F/ V218M/ H1778R in 13/ 7/ 3/ 2/ 2/ 2 pts, 9 further mutations occurring only once), trailed by ASXL1 (7 pts) and SRSF2 (6 pts). DNMT3A and IDH2 were mutated in 3 pts, each, with the latter exclusively altered at R140. KRAS (G12V, G12H), RUNX1 (D198Y, A424S), SF3B1 (K666N, V695L) were mutated in 2 pts each. Mutations in CBL (C381S), IKZF1 (Q7R), EZH2 (Y733C), SH2B3 (S186I), and IDH1 (R132H) occurred in one pt each. At relapse, using the identical amplicon based sequencing strategy, a median of 3 (0-13) nonsynonymous mutations was found in a median of 2.5 (0-6) genes. Counting gains and losses of nonsynonymous mutations at relapse, a median of 1 (0-9) changes was observed. Identities of gains and losses were more variable when compared to the mutations before alloSCT. Most changes occurred in TET2, followed by ASXL1, SRSF2, RUNX1, and DNMT3A . Taking only pts who had achieved full donor chimerism (n=16) into account, short molecular remission and number of mutational changes correlated (R²=0.013, figure). This correlation was stronger than the correlation between number of mutations before alloSCT and time in molecular remission after alloSCT (R²=0.0001). Notably, of the 6 pts in whom additional whole exome sequencing was performed at first presentation and upon progression, none showed clonal evolution beyond genes depicted in the targeted amplicon sequencing panel.
Conclusion: Clonal evolution of myelofibrosis pts who relapse after alloSCT is modest. A net loss of mutations can be found at relapse. In comparison to the situation in MDS and AML, the gain of novel mutations at relapse seems to be moderate in myelofibrosis. Future research has to identify mutational spectra with specific prognostic and predictive value.
No relevant conflicts of interest to declare.
