To the Editor

The WHO defines chronic neutrophilic leukemia (CNL) as a myeloproliferative neoplasm (MPN) with sustained elevated neutrophils and <10% immature cells [1]. Recently, recurrent somatic mutations in the membrane proximal domain of CSF3R were reported in patients with CNL [2],[3]. CSF3R was mutated in 100% [3], SETBP1 33% [3] and CALR in 12.5% of WHO-defined cases of CNL [4]. We analyzed mutations in CSF3R, SETBP1 and CALR in 14 subjects who met the WHO-criteria.

Findings

CSF3R exon 14–17 [3], SETBP1 exon 4 [3] and CALR exon 9 [5] were amplified by PCR and sequenced. 8 subjects who met the WHO 2008 CNL criteria had a CSF3RT618I mutation, 6 SETBP1 mutations (2 D868N, 2 I871T, 1 G870S and 1 D874N) and the last had a CALR mutation (c.1154-1155insTTGTC). All mutations were heterozygous except 1 case of SETBP1I871T. 6 other subjects, 2 with monoclonal gammopathy of unknown significance (MGUS)-associated CNL and 4 with reactive neutrophilic leukocytosis had no mutation of these genes. No subject had a JAK2V617F mutation (Table 1).

Table 1 Clinical characteristics and laboratory variables
Full size table

The consistent association between CSF3RT618I and CNL in our study is similar to data of Maxson et al. [2] and Pardanani et al. [3] (Table 2). Tefferi et al.[6] suggested including CSF3RT618I or other membrane proximal CSF3R mutations as a criteria for diagnosis of CNL. We also confirmed the high incidence SETBP1 mutations in patients with CNL. The mutations we detected focused on a hotspot area from D868 to D874 (Table 1). Although these mutations also occur in other hematologic neoplasms such as atypical chronic myeloid leukemia aCML and chronic myelomonocytic leukemia (CMML), analysis of SETBP1 mutations could help distinguish CNL from reactive conditions such as infection, inflammatory conditions and non-haematologic neoplasms.

Table 2 CSF3R and SETBP1 mutations in CNL
Full size table

Gotlib et al. reported JAK2V617F mutation in a subject of CNL [7]. Lasho et al. reported a CALR mis-sense mutation in a subject with CNL [4]. We found concurrent CSF3RT618I and CALR frame-shift mutations in 1 subject. The 5 bp insertion into CALR exon 9 is reported in BCR/ABL1- and JAK2-negative MPNs and results in a 1+ base-pair frame-shift with an altered C-terminus.

There is controversy whether co-existence of MGUS and CNL is one or two diseases. The 2 MGUS subjects in our study had no mutation in CSF3R, SETBP1, JAK2 or CALR. In another study, none of 6 cases of MGUS-associated CNL had CSF3R mutations [3]. Also, survival of patients with MGUS-associated CNL is significantly longer survival than those with CNL only. These data support the notion patients with MGUS and CNL are 2 diseases [8].

There may be therapy implications of our findings. CSF3R truncation mutations may be sensitive to SRC kinase-inhibitors such as dasatinib whereas CSF3R membrane proximal mutations may be sensitive to JAK kinase-inhibitors such as ruxolitinib [9],[10]. Ruxolitinib was reportedly effective in a mouse model of CNL and a patient with CNL and a CSF3RT618I mutation [2],[11]. However, ruxolitinib was ineffective in a patient with CSF3RT618I and SETBP1 mutations in whom fedratinib suppressed CFU-GM colony formation [12].

Authors’ contributions

XZJ designed the study and drafted the article. CYJ collected the data, analyzed the molecular aberrations, and drafted the article. RPG drafted the typescript. LB, JQ, XZF, QTJ, ZPH, and ZY reviewed the clinical and pathology data. All authors read and approved the final typescript.