CSF3R, SETBP1 and CALR mutations in chronic neutrophilic leukemia

Abstract

The WHO 2008 definition of chronic neutrophilic leukemia (CNL) is based on clinical and laboratory parameters but not on molecular abnormalities. Mutations in CSF3R, SETBP1 and CALR are reported in patients with chronic neutrophilic leukemia (CNL). However, because CNL is rare, there are few large studies of this issue. We sequenced these genes in 14 patients who met the WHO-criteria of CNL. 8 subjects had CSF3R^T618I, 6 SETBP1 mutations and 1 a CALR mutation. Our data suggest mutation analysis of CSF3R, SETBP1 and CALR should be included in the diagnostic criteria for CNL. These data may also have therapy implications.

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 CSF3R^T618I 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 SETBP1^I871T. 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 JAK2^V617F mutation (Table 1).

Table 1.

Clinical characteristics and laboratory variables

No.	Diagnosis	Gender/Age	Hb (g/L)	WBC (×10 9 /L)	ANC (×10 9 /L)	PLT (×10 9 /L)	Spleen (cm,LCM)	Karyotypes	CSF3R	SETBP1	CALR	JAK2 V617F	Treatment	Survival (month)
1	CNL	M/80	120	27.19	25.42	91	3	46,XY,t(1,7) (p32,q11)[10]	T618I	D874N	wt	wt	Hydroxyurea	52
2	CNL	M/64	123	86.83	ND	394	0	46,XY[5]	T618I	D868N	wt	wt	Hydroxyurea	5+
3	CNL	F/77	125	35.68	25.92	351	4	46,XY[20]	T618I	G870S	wt	wt	Hydroxyurea	22+
4	CNL	F/49	104	85.61	79.47	20	6	46,XX[20]	T618I	I871T	wt	wt	Hydroxyurea	13+
5	CNL	M/70	86	146.77	121.82	104	15	ND	T618I	I871T	wt	wt	Hydroxyurea	17
6	CNL	M/43	55	112.65	101.1	98	10	46,XY[5]	T618I	D868N	wt	wt	Hydroxyurea	4+
7	CNL	F/69	102	57.40	40.16	231	6	46,XX[12]	T618I	wt	c.1154-1155insTTGTC	wt	Hydroxyurea	10
8	CNL	M/45	119	32.76	28.73	286	6	46,XY[9]	T618I	wt	wt	wt	Hydroxyurea	32
9	MGUS-CNL	M/46	63	65.30	60.7	101	7	46,XY[2]	wt	wt	wt	wt	Hydroxyurea	6+
10	MGUS-CNL	F/52	121	26.53	19.94	170	3	46,XX[20]	wt	wt	wt	wt	Hydroxyurea	27+

All the parameters in Table 1 were measured at the initial diagnosis in our hospital.

MGUS-CNL: monoclonal gammopathy with uncertain significance associated CNL; Hb: hemoglobin; WBC: White Blood Cell Count; ANC: Absolute Neutrophil Count; PLT: Platelet Count; Spleen (cm): Spleen size under left costa. ND: not done.

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 CSF3R^T618I 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

	Gotlib J et al. (2013) [ 7 ]	Pardanani A et al. (2013) [ 3 ]	This series	Total
T618I only	1	5	1	7
T618I + SETBP1	4	4	6	14
Compound CSF3R mutationsa only	2	1	0	3
Compound CSF3R mutations + SETBP1 mutations	1	0	0	1
Others	1b	2c	1d	4

^a: compound CSF3R mutations mean nonsense or frameshift mutations that truncate the cytoplasmic tail (truncation mutations) combined with point mutations in the extracellular domain (membrane proximal mutation). In the 3 cases of compound CSF3R mutations Tyner et al. reported, two patients harbored T618I and one harbored T615A in the membrane proximal domain. In Tefferi’s study, the compound CSF3R mutation showed T618I + c.2341_2342insC.

^b: A case with JAK2 mutation only.

^c: A case with I598I and a case with M696T in CSF3R.

^d: A case with CSF3R T618I and CALR frameshift mutation.

Gotlib et al. reported JAK2^V617F 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 CSF3R^T618I 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 CSF3R^T618I mutation [2,11]. However, ruxolitinib was ineffective in a patient with CSF3R^T618I and SETBP1 mutations in whom fedratinib suppressed CFU-GM colony formation [12].