B-cell precursor acute lymphoblastic leukaemia (BCP-ALL) is a heterogeneous disease in which patient outcome is influenced by genetic lesions. Cytogenetic classification has improved survival through risk stratification for treatment. Translocations involving the Immunoglobulin Heavy Chain Locus (IGH) comprise 5% of BCP-ALL and lead to overexpression of juxtaposed genes, due to the powerful IGH enhancer elements. Multiple IGH partner genes have been described in BCP-ALL, including five members of the Ccaat Enhancer-Binding Protein (CEBP) transcription factor family: CEBPA, CEBPB, CEBPD, CEBPE and CEBPG, which comprise 12% of the IGH-translocation subgroup.
Patients with potential IGH-CEBP translocations were identified by mining cytogenetic data from UK ALL trials. The translocations were confirmed by fluorescence in situ hybridisation (FISH), followed by screening for commonly occurring BCP-ALL lesions using Multiplex Ligation-dependent Probe Amplification (MLPA) (SALSA P335 kit, MRC Holland), and detection of novel genetic lesions using SNP6.0 arrays. A total of 33 IGH-CEBP patients were identified; CEBPD (n=11, 34%); CEBPA (n=10, 30%); CEBPB (n=8, 24%); CEBPE (n=3, 9%); CEBPG (n=1, 3%). Cohort median age was 15 years (range, 2-65 years). Interestingly the majority of IGH-CEBPD patients were under the age of 10 years, while CEBPA and CEBPB patients were older, all over the age of 10 years (p=0.005). The median white blood cell (WBC) count of the cohort was low at 8.24x109/l (range, 0.9-430 109/l), the highest WBCs were observed in the CEBPB subgroup (p=0.04). Associated abnormalities were mixed: 5 Down syndrome (DS) patients were identified in the cohort all exclusive to the IGH-CEBPD subgroup, an observation which has been previously reported, BCR-ABL1 translocation (n=1), high hyperdiploidy (n=1) and hypodiploidy (n=1). Although numbers were small, some trends were observed: 27% (n=9) of patients were deceased: CEBPB (n=4, 44%), CEBPA (n=2, 22%), CEBPD (n=2, 22%) and CEBPE (n=1, 13%). Overall 15% (n=5) relapsed: CEBPD (n=3, 60%), CEBPA and CEBPB (each n=1, 20%). Minimum residual disease data (MRD) at day 28 were available for 8 (24%) patients, CEBPD (n=3, 38%), CEBPA (n=2, 25%), CEBPB, CEBPE and CEBPG (n=1, 13% each). Although numbers were low, all 3 MRD negative patients were IGH-CEBPD while the remaining patients, all MRD positive, comprised the remaining CEBP subgroups.
MLPA screening of 28 patients identified that 25% had whole gene deletions of CDKN2A/CDKN2B (n=7): the most common copy number abnormality (CNA) observed in IGH-CEBP. IKZF1 deletions of exons 4-7, a deletion resulting in the formation of the dominant negative isoform IK6, were present in 21% (n=6) patients: CEBPB (n=4, 67%) and CEBPD (n=2, 33%) (P=0.04). Four CNAs of the PAX5 gene were observed: individual gains of exon 5 and exon 7 (n=1 each, 50%); single exon gains of PAX5 have been predicted to lead to altered protein activity, whole gene deletion (n=1), deletion of exon 1 along with gain of exons 7 and 8 (n=1).
Using SNP6.0 arrays (n=13) a recurrent focal deletion of the ABL2 gene was discovered (n=4, 31%). These patients had a common region of deletion covering exon 2 of the gene. This deletion covers the Src Homology 2 domain (SH2), which allows for the recognition of phosphorylated tyrosine residues, vital for cell signalling. Patients spanned three subgroups; CEBPB (n=2, 50%), CEBPA (n=1, 25%) and CEBPD (n=1, 25%).
In this study, we show that in the IGH-CEBP subgroup, the CEBP partner gene influences disease phenotype. CEBPB patients showed more aggressive BCP-ALL characteristics, with the highest WBC, death rate (50%), older age and highest levels of associated abnormalities, with all patients showing one or more genetic lesions. The CEBPD subgroup comprises the youngest patients, a statistically significant finding both with and without inclusion of DS status. This subgroup also had the highest number of relapses in the cohort, and was the only CEBP subgroup to be MRD negative at day 28, albeit based on low patient numbers. A focal deletion within ABL2, a paralog of the ABL1 gene, is an interesting new abnormality in this subgroup, which may contribute to disease development.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.