Although steady progress of effective chemotherapy in childhood acute lymphoblastic leukemia (ALL) carried with exceeding 80% of individuals now cured, the majority of adult patients with ALL are not cured by chemotherapy, and allogeneic hematopoietic stem cell transplantation (allo-HSCT) is the only curative option. However, relapse remains the most leading cause of death after allo-HSCT. Adverse genetic alterations are generally accepted to be responsible for treatment failure and relapse. Several structural chromosomal alterations including rearrangement of the myeloid-lymphoid or mixed-lineage leukemia gene (MLL) and Philadelphia chromosome (Ph), have been mostly found in relapsed ALL. However, many Ph-negative (Ph-) ALL patients with normal karyotype , lacking known risk factors, also experienced relapse. The underlying pathologic determinants leading to relapse and prognostic markers in these cases remain poorly understood. More importantly, allo-HSCT is a distinct treatment option from tradtional chemotherapy and has 2 important forms to eliminate and select on malignant cells. The malignant cells that go on to causing relapse must initially survive ablation of chemotherapy before allo-HSCT and conditioning regimen in allo-HSCT. Then, after allo-HSCT, they must survive the effect of graft-versus- leukemia (GVL) reaction. Following this rationale, we hypothesized that there may be pivotal genetic causes confer leukemic cells a fitness advantage to undergo huge selective pressures and expand after allo-HSCT. To elucidate the genomic basis underlying relapse after allo-HSCT to aid to discover novel predictive biomarkers and identify therapeutic targets, we carried out the first whole-exome sequencing analysis in longitudinal matched samples from diagnosis to relapse after allo-HSCT in adult patients with the most common subtype of ALL, Ph- B-cell ALL (B-ALL).


Whole-exome sequencing was conducted for 9 genomic DNA samples from 3 relapsed cases with Ph- B-ALL (discovery cohort) at 3 specific time points including: diagnosis, complete remission (CR) after induction chemotherapy before allo-HSCT, relapse after allo-HSCT to discover candidate relapse-associated mutated genes. We identified putative somatic mutations by comparing each tumor ( diagnostic samples or relapsed samples) to normal (CR samples) from the same patient. To confirm candidate somatic gene mutations, screen relapse-associated gene mutations and define the frequency of somatic mutations identified by whole-exome sequencing analysis, we further carried out target genes whole coding regions sequencing in an ALL extended validation cohort including 58 adult Ph- B-ALL cases, where 27 patients experienced relapse at a median time of 6.5 (range 2-33) months after allo-HSCT and 31 patients did not relapse after allo-HSCT at a median follow-up for 34 (range 12–56) months.


(1) We discovered novel associations of recurrently mutated genes (CREBBP, KRAS, PTPN21) with the pathogenesis of adult Ph- B-ALL relapse after allo-HSCT, which were mutated in at least two relapsed cases, but were not mutated in non- relapsed patients. (2) The generation of high-depth whole-exome sequencing data in longitudinal matched samples from diagnosis to relapse after allo-HSCT in initial 3 patients allowed us to directly assessed the evolution of somatic mutations. Our data suggested that in the progression of leukemia relapse after allo-HSCT, the relapse clone had a clear relationship to the diagnosis clone, either arising from a subclone already exsiting in the diagnostic tumor, or originating from a common preleukemic progenitor with the diagnosis clone. In the latter pattern, the relapse clone acquires new genetic alterations while retaining some but not all of the alterations found in the diagnostic tumor. In contrast, in some cases, leukemia recurrences afer allo-HSCT may be composed of second malignancies with completely distinct sets of mutations from the primary tumor.


Our study is the first to explore genetic basis of adult Ph- B-ALL from diagnosis to relapse after allo-HSCT over time, which will provide novel genetic biomarkers on risk “index” to improve individualized treatment intensification and intervention strategies, and potential therapeutic targets for Ph--ALL relapse after allo-HSCT.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract