Chronic neutrophilic leukemia (CNL) and atypical chronic myeloid leukemia (aCML) are rare myeloproliferative and myelodysplastic/myeloproliferative neoplasms. So far, the diagnosis of CNL and aCML has been based on cytomorphology and the absence of JAK2V617F and PDGFR rearrangements. Recently, mutations in CSF3R and SETBP1 were identified and associated with CNL and aCML, respectively. Chronic myelomonocytic leukemia (CMML) and aCML also share several characteristics and need to be discriminated especially by the absolute number of monocytes in the peripheral blood.


To determine the frequency of CSF3R mutations (CSF3Rmut) in CNL, aCML, and CMML and to investigate a mutation pattern, cytogenetics and clinical data in all three entities.

Patients and Methods

To first delineate patients with potential CNL, we investigated blood and bone marrow smears and depicted patients with a white blood cell count >25x109/L, neutrophils >80%, immature granulocytes <10%, <1% myeloblasts and hypercellular bone marrow (according to WHO 2008). BCR-ABL1 fusion transcript, JAK2 and MPL mutations were excluded in all cases by RT-PCR and melting curve analyses. Indication for PDGFR rearrangements was precluded by over-expression analyses of PDGFRA and PDGFRB by quantitative real-time PCR, resulting in a final cohort of 20 cases declared as CNL patients. Additional 60 aCML and 252 CMML patients were included. Cytogenetics was available in 330/332 cases. Mutations in CSF3R exons 14 and 17 (n=332), in ASXL1 exon 13 (n=321), and the mutational hot spots in SETBP1 (n=331) and SRSF2 (n=320) were analyzed by Sanger sequencing.


In the total cohort of 332 patients we detected CSF3R mutations in 11 cases (3.3%). 8/11 cases showed a p.Thr618Ile mutation in exon 14, four of them carried an additional nonsense/frame-shift mutation in exon 17. One additional patient was mutated in p.Thr615Ala and showed a nonsense mutation in exon 17. Two cases showed a mutation in exon 17 only, one a nonsense the other a frame-shift mutation, respectively. Analyzing the mutation frequencies within the different entities revealed a clustering of CSF3Rmut within CNL cases with 7 of 20 (35%) mutated cases in contrast to 2 of 60 (3.3%; p=0.001) aCML and 2 of 252 (0.8%; p<0.001) CMML cases. Cytogenetics in CSF3Rmut cases showed that 9/11 cases had a normal karyotype and only one aCML patient harbored a del(3q) and one CMML patient a complex karyotype. Mutations in the three other analyzed genes ASXL1, SETBP1 and SRSF2 were detected in the total cohort in 156/321 (49%), 34/331 (10%), and 149/330 (45%) patients, respectively. Analyses regarding concomitant mutations of CSF3R with ASXL1, SETBP1 or SRSF2 revealed no additional mutation in two cases. In 8 of 11 parallel analyzed CSF3Rmut patients an ASXL1mut was identified, SETBP1 as well as SRSF2 were mutated in 3 of the 11 cases. Notably, the 7 CSF3Rmut within the CNL group had no mutation in SETBP1. Analysis of mutational loads in CNL showed that 6/7 CSF3Rmut had a higher mutational load than the second mutated gene (range: 25-50% vs. 10-30%). In one case both mutated genes had equal mutational loads (40%). In contrast, in CMML and aCML 3/4 patients had lower mutational loads in CSF3Rmut than in the additional mutated genes (20-50% vs. 40-50%), while also one case showed equal mutational loads in the mutated genes (50%). Combining the mutational results of these four genes indicate a specific and individual molecular pattern for these three different entities. While ASXL1 is frequently mutated in all entities (CNL: 8/11 (73%); aCML: 38/59 (64%); CMML: 110/251 (44%)), SRSF2 shows the highest mutation frequency in CMML cases (121/251; 48%), followed by aCML (24/60; 40%) and CNL (4/19; 21%). In contrast, SETBP1 is often mutated in aCML (19/60; 32%) and rarely in CMML (13/252; 5%) and CNL (2/19; 10.5%) patients. In addition, CSF3R is much more associated with the CNL cases (35%) and less frequently found in aCML (2%) and CMML (1%).


1) CNL, aCML and CMML are related diseases and difficult to distinguish by cytomorphology alone and therefore require additional diagnostic criteria, i.e. molecular mutations. 2) ASXL1 is the most frequently mutated gene in these entities and thus can help to prove clonality. 3) SETBP1 much more closely relates to aCML and SRSF2 to CMML. 4) Mutations in the novel marker CSF3R are closely related to CNL and thus qualify as a new molecular marker for diagnosis of CNL.


Meggendorfer:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schrauder:MLL Munich Leukemia Laboratory: Employment. Konietschke:MLL Munich Leukemia Laboratory: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Author notes


Asterisk with author names denotes non-ASH members.