Tyrosine kinase fusion genes in pediatric BCR-ABL1-like acute lymphoblastic leukemia

Abstract

Approximately 15% of pediatric B cell precursor acute lymphoblastic leukemia (BCP-ALL) is characterized by gene expression similar to that of BCR-ABL1-positive disease and unfavorable prognosis. This BCR-ABL1-like subtype shows a high frequency of B-cell development gene aberrations and tyrosine kinase-activating lesions. To evaluate the clinical significance of tyrosine kinase gene fusions in children with BCP-ALL, we studied the frequency of recently identified tyrosine kinase fusions, associated genetic features, and prognosis in a representative Dutch/German cohort. We identified 14 tyrosine kinase fusions among 77 BCR-ABL1-like cases (18%) and none among 76 non-BCR-ABL1-like B-other cases. Novel exon fusions were identified for RCSD1-ABL2 and TERF2-JAK2. JAK2 mutation was mutually exclusive with tyrosine kinase fusions and only occurred in cases with high CRLF2 expression. The non/late response rate and levels of minimal residual disease in the fusion-positive BCR-ABL1-like group were higher than in the non-BCR-ABL1-like B-others (p<0.01), and also higher, albeit not statistically significant, compared with the fusion-negative BCR-ABL1-like group. The 8-year cumulative incidence of relapse in the fusion-positive BCR-ABL1-like group (35%) was comparable with that in the fusion-negative BCR-ABL1-like group (35%), and worse than in the non-BCR-ABL1-like B-other group (17%, p=0.07). IKZF1 deletions, predominantly other than the dominant-negative isoform and full deletion, co-occurred with tyrosine kinase fusions. This study shows that tyrosine kinase fusion-positive cases are a high-risk subtype of BCP-ALL, which warrants further studies with specific kinase inhibitors to improve outcome.

Keywords: BCR-ABL1-like; minimal residual disease; pediatric B cell precursor acute lymphoblastic leukemia; tyrosine kinase fusion.

Figure 1. Distribution of tyrosine kinase fusions and CRLF2 high expression cases

Pie diagrams showing the percentages of tyrosine kinase (TK) fusion cases and CRLF2 high expression among A. 77 BCR-ABL1-like B-other cases, and B. 76 non-BCR-ABL1-like B-other cases. Within the CRLF2 high expression cases, a sub-distribution of cases with JAK2 mutation and/or PAR1 deletion is shown. In the non-BCR-ABL1-like B-other cases, 2 PAR1-deleted cases were not tested for JAK2 mutations.

Figure 2. BCR-ABL1-like tyrosine kinase fusion cases

A. Clinical characteristics and follow-up for the tyrosine kinase fusion cases. Barplot representing years from diagnosis to event or censoring. Treatment protocol and arm, prednison window response (Pred), morphological response after induction therapy (Resp), minimal residual disease (MRD), and deletion status of IKZF1 are shown. MRD monitoring by PCR was performed for research purposes in ALL-9 and COALL03 and for MRD-guided risk stratification in ALL-10. WT indicates no IKZF1 deletion. For definition of Pred, Resp, and MRD see footnotes Table 3. Response ‘unkn’ indicates patients with low cellularity bone marrow which could not be evaluated. * Indicates patients who received bone marrow transplant, for which HR-treated cases were eligible. B. Cumulative incidence of relapse and non-response curves for tyrosine kinase fusion positive BCR-ABL1-like cases, fusion-negative BCR-ABL1-like cases, and non-BCR-ABL1-like B-other cases. Relapse and non-response were considered as events, death as competing event. Cumulative incidence probabilities were estimated using a competing risk model, equality was tested with the Gray test.