Background: Approximately one third of chronic myeloid leukemia (CML) patients treated with tyrosine kinase inhibitors (TKI) fail to achieve optimal response due to resistance. It is known that BCR-ABL1 kinase domain (KD) mutations are associated with TKI resistance. However, only half on non-responsive patients carry BCR-ABL1 KD mutations. Other resistance mechanisms independent of BCR-ABL1 KD mutations, such as mutations in myeloid-related genes, may participate in the lack of response. Moreover, resistance to TKI therapy confers a much higher risk of progression to advanced disease stages of CML, such as accelerated phase (AP) and blast crisis (BC), and consequently, a worse prognosis. Aims: The main objective of this study was to explore the impact of mutations in myeloid-related genes in TKI resistance and progression to AP/BC by targeted deep-sequencing (TDS). Methods: The mutation profile study was done by TDS using a 32 myeloid-related gene panel (Myeloid Solution Capture Kit, Sophia Genetics, Switzerland). The platform Sophia DDM (Sophia Genetics) was used for data analysis. Regarding variant filtering, sequencing and mapping errors were eliminated and intronic and synonymous variants were filtered out. Variants present with a minor allele frequency (MAF) >1%, according to population databases (ExAc, 1000 genomes), were considered polymorphic changes without clinic relevance. COSMIC, VARSOME and Franklin databases as well as in silico functional predictors (SIFT, PolyPhen2 and MutationTaster) were used for variant interpretation. Variants were also filtered according to the variant allele frequency (VAF): all variants with VAF ≥5% were reported, as well as hot-spot variants with VAF 2-5% and at least 25 reads. Only variants described as pathogenic or potentially pathogenic were reported. Finally, BCR-ABL1 mutations were analyzed by Sanger in resistant patients. Results: TDS was performed in a total of 78 samples from 67 patients. Across patients with available sample at diagnosis, 25% (16/64) of patients harbored at least one mutation. Overall, 11 (17.2%) patients had 1 mutation, 3 (4.6%) had 2 mutations, 1 (1.6%) patient had 3, and 1 (1.6%) patient had 4 mutations (Figure 1A). Most frequently mutated genes at diagnosis were ASXL1 (14.1%), DNMT3A (6.3%), JAK2 (3.1%) and TET2 (3.1%), followed by ETV6, EZH2, IDH2, SETBP1, SRSF2, TP53 and WT1, all of them at a frequency of 1.6% (Figure 1B). Of the 67 CML patients included, 15 (22.4%) had resistance to first-line TKI and 5 of them (33%) progressed to AP/BC. From the 15 resistant patients, TDS was performed in 12 cases at diagnosis, in 5 at time of resistance, and at progression in 4/5 patients. At diagnosis, 6 (50%) resistant patients harbored 1 mutation in ASXL1 (3/12), DNMT3A (2/12) or TET2 (1/12) and 1 patient harbored 2 mutations (ASXL1 and WT1). Only 9 of 52 (17.3%) non-resistant patients harbored ≥1 mutation at diagnosis; most frequently mutated gene was ASXL1 (5/9). At time of resistance, 5 patients had BCR-ABL1 mutations by Sanger technique and 3 patients harbored cancer-related gene mutations by TDS (ASXL1 in 2 patients). At progression, all patients harbored at least 1 mutation by TDS, except one patient in which only 2 mutations in BCR-ABL1 KD were detected. Image:Summary/Conclusion:ASXL1 mutation was the most frequent mutation in CP at diagnosis and ASXL1 and ABL1 were the most common variants at resistance and progression. Acquisition of mutations was detected during CML progression. Acknowledgments: this work was supported by a grant from Incyte (EU-ES-H-18021)

Myelofibrosis (MF) occurs as part of the natural history of polycythemia vera (PV) and essential thrombocythemia (ET), and remarkably shortens survival. Although JAK2 V617F and CALR allele burden are the main transformation risk factors, inflammation plays a critical role by driving clonal expansion toward end-stage disease. NF-κB is a key mediator of inflammation-induced carcinogenesis. Here, we explored the involvement of miR-146a, a brake in NF-κB signaling, in MPN susceptibility and progression. rs2910164 and rs2431697, that affect miR-146a expression, were analyzed in 967 MPN (320 PV/333 ET/314 MF) patients and 600 controls. We found that rs2431697 TT genotype was associated with MF, particularly with post-PV/ET MF (HR = 1.5; p < 0.05). Among 232 PV/ET patients (follow-up time=8.5 years), 18 (7.8%) progressed to MF, being MF-free-survival shorter for rs2431697 TT than CC + CT patients ( p = 0.01). Multivariate analysis identified TT genotype as independent predictor of MF progression. In addition, TT (vs. CC + CT) patients showed increased plasma inflammatory cytokines. Finally, miR-146a −/− mice showed significantly higher Stat3 activity with aging, parallel to the development of the MF-like phenotype. In conclusion, we demonstrated that rs2431697 TT genotype is an early predictor of MF progression independent of the JAK2 V617F allele burden. Low levels of miR-146a contribute to the MF phenotype by increasing Stat3 signaling.

Background:Telomere length (TL) is shortened with every cell division. Critically short telomeres are associated with genetic instability. In chronic myeloid leukemia (CML), increased cellular turnover of clonal BCR‐ABL1‐positive hematopoietic stem and progenitor cells leads to significantly shortened telomeres in myeloid cells from peripheral blood (PB). Retrospective studies have related accelerated telomere shortening in CML to disease stage, clinical risk score at diagnosis as well as cytogenetic remission.So far, no study has considered the possible association between TL at diagnosis of CML and the achievement of deep molecular response (DMR) with imatinib (IM) treatment.Aims:To explore if TL at diagnosis of CML is predictive of DMR achievement with IM as first‐line treatment.Methods:Samples from 100 patients with de novo CML at diagnosis of three different Catalan hospitals between 2004 and 2016 were included. Three patients were considered as missing due to no valid PCR results. All patients provided informed consent. Patients’ median age was 49 yrs (range 24–80 yrs) and 59% were males. Thirty‐one (38%) patients had the e13a2 p210 BCR‐ABL1 isoform whereas 51 (62%) had the e14a2 isoform. According to Sokal and ELTS scores, 48(50%) and 52(60%) patients were low risk, 30(32%) and 24(28%) were intermediate risk while 17(18%) and 10(12%) were high risk, respectively.Patients were treated with first line IM 400 mg QD orally. For the molecular response (MR) analysis, only response to IM was considered. Response during the first year was assessed according to European Leukemia Net 2013 guidelines. DMR was defined as MR4.0 or MR4.5 The median time from IM start to MR4.0 achievement was 16.1 months (5.5, 100.7). Follow‐up was at least 24 months for all patients.For TL determination, DNA was extracted from PB or bone marrow cells. TL was analysed by monochrome multiplex quantitative polymerase chain reaction and the results were expressed as the standardized T/S ratio. Age adaptation of TL, indicated as ΔTL (delta‐TL), was performed using data from 107 healthy controls (age range 16–84 years). The ΔTL for every patient was calculated subtracting the subject's linear predicted TL (linear regression analysis of healthy controls) from the observed TL.Results:Healthy controls showed the expected decline in TL with increasing age (R2 = 0.1693), which was not detectable in CML (R2 = 0.0002), suggesting that telomeres are not age‐dependent in CML at diagnosis.Analysing the clinical data, no significant correlation between ΔTL and clinical prognostic scores (Sokal and ELTS, with p = 0.816 and p = 0.314, respectively) was observed, nor between ΔTL and p210 BCR‐ABL1 onco‐protein isoforms (e13a2 and e14a2, p = 0.981).Finally, focusing on the ΔTL representing the difference between TL expected by age and TL determined in the CML cells, patients with lower ΔTL at diagnosis were significant associated with achievement of MR4.0 and MR4.5 (HR [95%CI]: 0.24 [0.1, 0.74], p = 0.013 and 0.23 (0.1, 0.76), p = 0.016, respectively). In addition, when patients were stratified according to median value of ΔTL (0.40), those with ΔTL<0.40 significantly achieved MR4.0 earlier than patients with ΔTL>0.40 (median in months [95%CI]: 21 [8.2, 33.7] vs 46.2 [23.6, 68.8], p = 0.034) (Fig1).Summary/Conclusion:TL at diagnosis of CML is a predictive biomarker of DMR achievement with IM as first‐line treatment. Acknowledgements: FIS (AES 2016) PI16/01200 and FEHH.image