The CEBPA gene codes for the C/EBPa transcription factor which regulates hematopoietic stem cell homeostasis and orientates towards myeloid differentiation and against a T lymphoid fate. CEBPA silencing by promoter hypermethylation (CEBPAmeth) has recently been identified (Wouters, Blood 2007) in a minor subset of AML, all of which demonstrated T lymphoid features (expression of CD7 and CD3 transcripts, TCRG rearrangement and frequent NOTCH1 mutations (NOTCH1m)). These AML may be closer to immature T-ALLs, with CEBPa silencing and NOTCH1m determining lineage orientation. Although expression of CD7 and cytoplasmic CD3 (cCD3) are considered as defining markers in T-ALL, CD7 is commonly expressed in AML and reproducible multicenter detection of low level cCD3, when performed, can be difficult. We have classified CD7+, cCD3+ T-ALLs on the basis of their TCR profiles: immature cases arrested prior to TCRB VDJ rearrangement include TCRgd lineage precursors and can be divided into those with no TCR rearrangement (IM0), TCRd only rearrangement (IMD) or TCRd and TCRg rearrangement (IMG). The CALM-AF10 fusion transcript is found in both AML and T-ALL, where it is restricted to the TCRgd lineage. We initially searched for in CEBPAmeth by methylation specific PCR in 5 AML with downregulated CEBPand confirmed that the 3 cases (FAB-M0/M1) with CEBPAmeth expressed CD3, CD7 and/or TCRg rearrangement. Only 1/49 AML M0, all of which were cCD3 negative, demonstrated CEBPAmeth.We then searched for CEBPAmeth and mutation and NOTCH1m in 18 IM T-ALL (4 aged <15y; 5 IM0, 5 IMD, 8 IMG) compared to 6 adults with TCRB VDJ+, TCRab lineage T-ALL. No CEBPA mutations were found but 11/24 (56%) demonstrated CEBPAm and 10/23 (43%) NOTCH1m. Both correlated with IMG T-ALL, when 6/8 demonstrated CEBPAmeth and NOTCH1m and a seventh only NOTCH1m, compared to only 4/10 CEBPAmeth and 1/10 NOTCH1m in IM0/D cases, with no overlap. In ab-lineage T-ALLs 2/6 demonstrated NOTCH1m, one of which was also CEBPAmeth, and notably expressed TLX1/HOX11. CEBPAmeth cases were more frequently CALM-AF10+ (5/10), CD2-, CD123/IL3R-, HOXA9 transcript+ and IgH DJ rearranged but did not differ from non-CEBPAmeth T-ALL in CD34, CD5, CD117, CD13/33, CD79a or CD56 expression. As such CEBPAmeth is rare in cCD3- M0 AML but common in cCD3+ immature T-ALL, independently of myeloid antigen expression. CEBPAmeth and NOTCH1m in IM T-ALL are preferentially associated with cases having undergone TCRG rearrangement but are not found in the majority of more immature cases, demonstrating that CEBPAmeth is not necessary for CD7 and cCD3 expression, within a T-ALL context. It also suggests that CALM-AF10 should be searched for in CEBPAmeth AML, raises issues regarding AML vs. ALL protocol assignement and the use of demethylating agents in IMG T-ALLs and suggests that the potentially misleading “biphenotypic leukaemia” label be reassessed.

Author notes

Disclosure: No relevant conflicts of interest to declare.

Sign in via your Institution