Refinement of IKZF1 Genomic Status in Pediatric Philadelphia Positive Acute Lymphoblastic Leukemia

Background: IKZF1 deletions are the most frequent secondary genetic alterations with prognostic impact in pediatric Philadelphia-positive B-cell precursor acute lymphoblastic leukemia (Ph+ BCP-ALL). Studies of IKZF1 status have mainly focused on deletions of the gene whereas very little is known about other genetic alterations in the IKZF1 locus.

Aim: Using an amplicon deep sequencing approach to investigate IKZF1 mutations occurrence in pediatric Ph+ BCP-ALL reported as IKZF1wildtype (WT) for deletions by MLPA analysis.

Methods: Six European centres participated in this study with a total of 98 patients.The 454 GS Junior system (Roche) was used to sequence the coding regions of IKZF1, Exon1 - Exon7 (Mullighan C. et al, 2009). An average depth of 250 reads (minimum of 102 reads, maximum of 853 reads) per amplicon was achieved. Quality control analysis excluded variants coverage of <50 reads in both forward and reverse directions. Variant analyses were performed using Amplicon Variant Analyzer (AVA) software (Roche). Variants at intronic regions or in homopolymeric stretches were excluded from further analysis, and a cut-off ≥10% mutated reads was applied to define mutations. Finally, variants described as SNPs or silent mutations were excluded.

Results: In total, 882 amplicons were screened obtaining 337 variants. Of these, 14 variants corresponding to mutations with predicted deleterious impact on the function of IKZF1 passed the filters of our analysis pipeline. Twelve out of 98 (12%) IKZF1 non-deleted patients were mutated and 2 of them carried double mutations. Both cases showed different mutant allele frequencies of the 2 mutations, suggesting the presence of leukemic sub-populations with different IKZF1 status. Mutations can be subdivided in 2 categories: 5 miss-sense point mutations localized at the DNA binding domain with a dominant negative effect and 7 frameshift mutations leading to haploinsufficiency and impairing also the dimerization activity of IKZF1. The 5 point mutations were located in exon4 and 4 out of 5 were localized in the coding region of zinc finger 2 (ZF2), essential for DNA-protein interaction. Three patients carried the same c.698A>G substitution, that caused the change of Asparagine 159 into Serine. The 7 frameshift mutations (2 deletions, 4 insertions and 1 InDel) were identified all along the IKZF1locus. These nucleotide variations caused a shift in the reading frame and a consequent formation of premature stop codons depleting the C-terminal dimerization domain.

Six out of 12 patients carrying IKZF1 mutations experienced an event (either relapse or death in CCR): 6 patients were treated in the pre-TKI (tyrosine kinase inhibitor) era and had 5 events while the remaining 6 patients received imatinib and showed a more favourable outcome with only one patient who died in CCR. This trend is in line with the findings recently reported by Van der Veer et al. (2014). In this study on IKZF1 deletions were identified in 126 of 191 patients, while 65 were defined IKZF1 WT on the basis of deletions analysis alone. Our study revealed that 6 of these 65 IKZF1 WT patients carried IKZF1 mutations (3 in the pre-TKI and 3 in the EsPhALL cohort). The outcome of mutated patients and deleted patients taken together (IKZF1 aberrant) is consistent with results reported in the Van der Veer study, both for the pre-TKI cohort [(4-year disease free survival (DFS) was 28.1% [SE 6.4] for IKZF1 aberrant vs 64.3% [SE 9.7] for IKZF1 WT, P=0.0036); Van der Veer study (30.0% [SE 6.8] for IKZF1 deleted vs 57.5% [SE 9.4] for IKZF1 WT, P=0.013)] and for the EsPhALL cohort [(4-year DSF was 55.7% [SE 6.8] for IKZF1 aberrant vs 59.8% [SE 11.6] for IKZF1 WT (P=0.348); Van der Veer study (4-year DFS was 53.6 [SE 7.0] for IKZF1 deleted vs 63.1 [SE 11.0] for IKZF1WT, P=0.1676].

Conclusion: This work revealed IKZF1 mutations in >10% of Ph+ BCP-ALL pediatric patients previously classified as IKZF1 WT on the basis of deletions analysis. Our results, together with data of IKZF1 deletions, highlight a 70-75% incidence of IKZF1 aberrations in pediatric Ph+ patients. Exon4 carried 9 of 14 aberrations: 7 point mutations and 2 deletions, suggesting a hotspot region for somatic mutations acquisition. The outcome of mutated patients is in line with the data previously published on the impact of IKZF1 deletions in pre and post TKI cohorts, suggesting a similar impact of mutations and deletions on clinical outcome of pediatric Ph+ patients.