Whole Exome Sequencing Identifies Somatic Mutation Patterns Of Myeloid Cells In Patients With Myeloproliferative Neoplasms Treated With Allogeneic Bone Marrow Transplantation

Allogeneic bone marrow transplantation (ABMT) is so far the only curative treatment option for patients with myelofibrosis (MF), although it carries risk of treatment related mortality and relapse. The effect of somatic mutations present in the clone causing the disease on the outcome of the ABMT is unknown. The aim of the study was to investigate the genetic basis of clonal evolution of MF patients undergoing ABMT and identify candidate prognostic markers of relapse. We performed high-resolution genetic analysis on DNA samples from five patients with MF who received ABMT. The analysis of acquired chromosomal deletions, gains and losses of heterozygosity was performed using SNP 6.0 microarrays. Somatic mutations in the genomes of the patients were identified using whole exome sequencing (WES) of tumor (granulocytes) and matched control tissue (T-lymphocytes). The 51-bp paired-end next generation sequencing was performed using HiSeq2000 system. The somatic origin of mutations was validated by Sanger sequencing. In the three relapsed cases we analyzed both the samples prior to ABMT and at relapse. For the two non-relapse cases only the baseline sample was analyzed, as the chimerism level at last follow up (3 and 5 years after ABMT) was 100% of the donor. We selected one relapse and one non-relapse case who did not carry mutations in JAK2 or MPL, as well as one relapse and one non-relapse case who had close to 100% JAK2-V617F mutational burden.

Patient 1 was a 49 year-old diagnosed with MPL-W515A positive post essential thrombocythemia (ET) MF. Following ABMT the patient achieved complete clinical and molecular remission. Three years later the patient developed myelodysplastic syndrome with re-appearance of the MPL-W515A positive clone and within three months refractory anemia with excess blasts II phenotype. The clonal evolution analysis revealed that the clone causing the relapse was the same one present at the initial post-ET MF diagnosis, which acquired additional somatic mutations and chromosomal aberrations. The initial clone causing post-ET MF carried 12 somatic mutations including MPL, DNMT3A, U2AF1, ASXL1, SIRT2, RAD50 and other genes, while no chromosomal aberrations were detectable. The relapsed clone acquired additional 9 somatic mutations in genes TP53, MYO18B, CDYL and others, as well as deletion (del) 7p, del7q, chromothripsis of chromosomes 11 and 16, and del22q.

Patient 2 is a 35 year-old diagnosed with post-ET MF. Mutations in JAK2 or MPL were not detected. Prior to ABMT a small clone with a deletion on chromosome 12q targeting SOCS2 gene among 3 other, could be detected by SNP array. The relapsed sample showed a full clone with 2 deletions targeting 2.5Mb region on chromosome 12 containing the SOCS2 gene, as well as a single gene ARID1B on chromosome 6. ARID1B is a member of SWI/SNF chromatin remodeling complex and a putative tumor suppressor. WES identified only 3 somatic mutations in this patient (MTUS2, SP3 and PCDH12). MTUS2 and SP3 mutations arose in the relapsed clone.

Patient 3 is a 57 year-old diagnosed with post polycythemia vera MF. The patient carried JAK2-V617F mutation coupled with a 9p UPD, therefore having a 99.2% mutational burden. Within 4 months after BMT the JAK2-V617F clone was detectable with 32% mutational burden, reaching 98% one month later. Apart from 9p UPD no other chromosomal aberration was detectable before BMT or at relapse. Preliminary WES data revealed 7 somatic mutations at relapse, including APBB1 and DOCK4.

Patient 4 was diagnosed with JAK2/MPL negative post-ET MF, while patient 5 had primary MF. In patient 4 we identified 10 somatic mutations and no chromosomal aberrations, while patient 5 carried the 9p UPD and a deletion of 20q, 11 somatic mutations and a germline variant in TET2 gene.

It seems that the sheer number of somatic mutations does not predict ABMT outcome. As the validation of WES hits is still ongoing, we expect to see other gene mutations that may serve as positive or negative predictors of the ABMT outcome in patients with MF. They will reflect the genomic adaptation that facilitates the escape from the graft-versus-myelofibrosis effect of the donor cells and allows the clonal dominance leading in some cases to disease progression. Further analysis of possible presence of subclones carrying relapse associated mutations before ABMT will have important implications for estimating success of ABMT in MF patients.