C/EBPA methylation is common in T-ALL but not in M0 AML

Louis Terriou, Raouf Ben Abdelali, Christophe Roumier, Ludovic Lhermitte, John de Vos, Pascale Cornillet, Olivier Nibourel, Kheira Beldjord, Hervé Dombret, Guy Leverger, Vahid Asnafi, Claude Preudhomme and Elizabeth Macintyre

To the editor:

The CCAAT/enhancer binding protein α (C/EBPA) is a stage-specific transcription factor that controls proliferation and differentiation toward a myeloid and against a T-lymphoid fate.1 Nonconditional targeted disruption of C/EBPA leads to a selective early block to granulocyte maturation,2 similar to that observed in acute myeloid leukemia (AML), without affecting other hematopoietic lineages. The function of C/EBPA can be inhibited by several mechanisms in AML, including mutation,3 (post)transcriptional modulation,4 posttranslational inhibition,5,6 and, more recently, epigenetic modification of C/EBPA expression.7 Recently, Wouters et al,8 using a bisulfite genomic sequencing method, defined a novel specific subgroup of AML with C/EBPA silencing by methylation (C/EBPAmeth) and T-lymphoid features (expression of cytoplasmic [c] CD3 and CD7 transcripts, T-cell receptor γ [TCRγ] rearrangement, and NOTCH1 mutation [NOTCH1m]).

We initially searched for C/EBPAmeth by methylation-specific polymerase chain reaction (MSP), as described by Chim et al,7 in 5 AML with down-regulated C/EBPA by transcriptional profiling, and confirmed that the 3 cases (French-American-British [FAB]–M0/M1) with C/EBPAmeth demonstrated cCD3 (n = 1), CD7 (n = 3), and/or TCRγ (n = 1) rearrangement. The specificity of the MSP was investigated by DNA sequencing, which showed the expected patterns of bisulfite-induced changes, in keeping with the cases described by Wouters et al.8 To further investigate C/EBPAmeth in early T and myeloid acute leukemia, we searched for C/EBPAmeth and mutation and NOTCH1m in 99 T-ALL. Diagnosis of T-ALL was based on European Group for Immunophenotyping of Leukemia (EGIL) criteria and expression of cCD3 and CD7. TCR-based classification of T-ALL9 was performed centrally at Necker-Enfants Malades Hospital (Paris, France), using at least cCD3, CD3, TdT, CD2, CD5, CD7, CD1a, CD4, CD8, TCRαβ, TCRγδ, βF1, CD34, CD117, CD13, CD33, MPO, CD10, CD19, and cCD79a antibodies. No C/EBPA mutations were found, but 37/99 (37%) demonstrated C/EBPAmeth and 47/90 (52%) NOTCH1m. The stage of maturation arrest and immunophenotype of T-ALLs as a function of their C/EBPAmeth status is detailed in Tables 1 and 2, respectively. C/EBPAmeth was relatively rare in the most immature T-ALL (IM0) with no evidence of TCR rearrangement (12%) and in mature sTCRαβ+ cases (11%) compared with all other categories (34/73; 47%). C/EBPAmeth T-ALL expressed CD117 more frequently (P = .022) and CD2 (P = .001) less frequently but did not differ significantly from non-C/EBPAmeth T-ALL with respect to expression of CD13, CD33, CD34, CD56, CD123, CD1a, CD4, CD8, or CD10. Biphenotypic acute leukemia (AL), defined by an EGIL score greater than 2, was seen in 1 C/EBPAmeth and 1 non-C/EBPAmeth cases. NOTCH1m were seen in 63% of C/EBPAmeth T-ALLs, compared with 47% of non-C/EBPAmeth cases (P = .14).

Table 1

C/EBPAmeth status of 99 T-ALLs as a function of their stage of maturation arrest

Table 2

C/EBPAmeth status of 99 T-ALLs as a function of their immunophenotype

In contrast to T-ALL, C/EBPAmeth was found in only 1 of 49 immature cCD3-negative AML (FAB M010); this case was also CD7-positive. None of the 49 AML M0 expressed NOTCH1m.

These data demonstrate that C/EBPAmeth is common in T-ALL but not in M0 AML. It cosegregates with cCD3 expression and the onset of TCR rearrangements and is common at all stages of maturation arrest other than IM0 and TCRαβ T-ALL. As such it is a marker of neither immaturity nor “myeloid lineage/biphenotypic acute leukemia” and is virtually specific to T-ALL with the context of acute T/myeloid leukemia. Distinction of AML from early T-ALL is essentially based on detection of cCD3, which can be difficult to standardize in a multicenter setting. Detection of C/EBPAmeth could potentially be used as an alternative, although methylation-specific PCR may not be easier to standardize. Since C/EBPAmeth was not seen in most T-ALL, other mechanisms of silencing must be operational. These data also raise issues regarding AML versus ALL protocol assignment and the use of demethylating agents in T-ALL. They also suggest that the potentially misleading “biphenotypic leukemia” label be reassessed.


This work was supported by the Laurette Fugain Association, the Fondation de France (Comité Leucémie), and the North-West Canceropole (Onco-Hematology Axis).

Contribution: C.P. and E.M. conceptualized and designed the research, analyzed data, and wrote the paper; L.T. and R.B.A. performed the research and wrote the paper; L.L., J.d.V., O.N., K.B., V.A., and C.R. analyzed data; and P.C., H.D., and G.L. provided data and samples.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Elizabeth Macintyre, Laboratoire d'Hématologie, Tour Pasteur, Hôpital Necker, 149-161, rue de Sèvres, 75743 Paris cedex 15, France; e-mail: elizabeth.macintyre{at}