CONTEXT:A substantial number of patients with acute lymphoblastic leukemia (ALL) die from progressive disease after relapse that is frequently associated with clonal evolution. Clinical exome sequencing (CES) is a useful tool for mutation investigation, particularly in leukemic subclones during relapse.

OBJECTIVE:The objective of this pilot study was to determine the specific genetic abnormalities associated with primary tumor comparing with normal hematopoietic tissue and to define clonal evolution patterns at relapse.

DESIGN:CES (TruSightTM One Sequencing Panel) was performed on samples obtained from the patient with B-cell Ph-negative ALL at diagnosis, remission and relapse.

SETTING:DNA was extracted from bone marrow samples according to modified salting out method. Libraries were pooled together and underwent two rounds of hybridization and capture, and additional quantification and size assessment.

Between 21 and 26 million reads for each sample were obtained with a MiSeq (Illumina, USA), with a coverage of at least 20x for 97% of all sequences. More than 99% of reads passed quality and length control and were used in the analysis.

De-multiplexing of indexed reads and the generation of FASTQ files were performed by the MiSeq system on-instrument software. The sequencing data, composed of 150 bp paired-end reads, were processed for MiSeq Reporter (Illumina), which is endowed with Burrows-Wheeler Aligner (BWA) for alignment on a reference sequence of hg19 and Genome Analysis Toolkit (GATK) for variant calling. Then, the Variant Call format (VCF) file was processed for analyzing on VariantStudio (Illumina) and wANNOVAR (Yang H. et al, 2015), which contain annotation databases, such as the ClinVar database, OMIM, the dbSNP database), the Polymorphism Phenotyping v2 database (Polyphen-2), and the Sorting Intolerant from Tolerant database (SIFT) and some others.

PATIENTS OR OTHER PARTICIPANTS:A 27-year-old woman, who was admitted to the National Research Center for Hematology (Moscow, Russia) on March, 2017, presented with a three-month history of paleness and ossalgia. Common B-ALL (BII) ALL with complex karyotype and extramedullary lesions in ovaries, kidneys, intra-abdominal and peripheral lymph nodes has been diagnosed.

Patient achieved complete remission (CR) after first induction therapy according to ALL-2016 protocol ( NCT03462095). After 16 months of CR during maintenance therapy (July, 2019) was diagnosed a relapse, which was refractory to multiple chemotherapy approaches and the patient died from the disease (January, 2020).

RESULTS:Primary dataset analysis included more than 9000 gene variations for each of the three samples. Only clonal variations (at diagnosis) that are pathogenic either by the type of mutations (insertions (frameshift) and splice site mutations) or by prediction databases (PolyPhen-2, SIFT) were selected. As a result, about 70 variations were identified for each of the samples. Table 1 shows all clonal somatic mutations by which tumor tissue differs from normal hematopoietic tissue.

At the time of diagnosis, a clone with monosomy of chromosome 3 was identified, that disappeared at CR and reappeared as a minor clone at relapse. Taking in consideration that the loss of one allele was caused by monosomy, we searched for pathogenic variations in genes localized on the chromosome 3 in order to identify genes inactivated according to Knudson's "Two-Hit" Theory of cancer causation (Knudson A., 1971).

Out of 500 genes variations on chromosome 3, mutations in 5 genes according to prediction databases (PolyPhen-2, SIFT) were defined as pathogenic. All mutations were germinal. Moreover, four out of five mutations were homozygous. Table 2 shows the dynamics of pathogenic variations in genes localized on the chromosome 3 at different stages of the disease. The missense mutation ofCOL7A1gene (c.2791C> T), which was found in a small percentage (8%), was localized on the disappeared copy of chromosome 3, and its detection in this case was possible due to the presence of normal cells.

CONCLUSIONS:There is an evident change in tumor clones at relapse: a major clone diagnosed inde novoB-cell ALL patient became minor at relapse. The data obtained from this analysis suggest that at the time of diagnosis, several clones are present in the tumor, and that most of the diagnosed pathogenic variations are not associated with oncogenesis and are germinal.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.