Optimization of CSF3R Mutation Detection in Severe Congenital Neutropenia and Cyclic Neutropenia Patients for Routine Diagnostics Using Next Generation Sequencing

Severe congenital neutropenia (CN) is a group of bone marrow failure syndromes characterized by absolute neutrophil counts below 0.5x10⁹/L, susceptibility to bacterial infections and frequently associated with maturation arrest at promyelocyte stage in the bone marrow (BM).

There is a high incidence of malignant transformation among CN patients with a cumulative rate of MDS/AML 22 % after 15 years of G-CSF treatment. The acquisition of G-CSFR truncating mutations is a risk factor for leukemic transformation in CN patients. Therefore, annual monitoring of CSF3Rmutations by means of next generation sequencing (NGS) is required for identification of CN patients with high risk of MDS/AML development.

Since CSF3R mutations usually occur at low frequency without additional clinical features, it is important to carefully select suitable clinical sample type and methods for mutation detection. Next, it remains to be evaluated which CN genetic subgroups should be considered for annual screening of CSF3Rmutations.

We performed CSF3R mutational screening in DNA and/or cDNA in 101 patients (ELANE, n = 42; HAX1, n = 16; G6PC3, n = 7; JAGN1, n = 2; WASP, n = 1; digenic ELANE, HAX1, n =1; digenic HAX1 and G6PC3, n =1; inherited mutations in CSF3R, n = 2; genetically unclassified CN, n = 9; cyclic neutropenia (CyN), n = 20) from the European Branch of the Severe Congenital Neutropenia Registry (SCNIR).

Using DNA deep sequencing we screened 63 of 81 CN-patients and 20 CyN patients. Using this method, we identified CSF3R mutations in 22.2% (14/63) of CN patients and 10% (2/20) of CyN patients. The frequency of CSF3R mutations in CN patients with known inherited mutations was 20% (11/55): 30 % (3/10) in CN-HAX1 patients, and 22.9 % (8/35) in CN-ELANE patients. Interestingly, 3/8 (37.5 %) patients harbouring CSF3R mutations were observed in genetically unclassified CN. We did not detect any acquired CSF3R mutations in the small groups of CN patients (n=10) harbouring inherited G6PC3, JAGN1, CSF3Ror digenic mutations.

In order to increase the sensitivity of mutation detection we performed cDNA deep sequencing of the critical region of G-CSFR. We sequenced 38 CN patients (ELANE, n = 15; HAX1, n = 11; JAGN1, n = 2; G6PC3, n = 2; WASP, n = 1; germline CSF3R, n = 1; genetically unclassified, n = 6). We found 13% (2/15) CN-ELANE, 27% (3/11) CN-HAX1 and 33% (2/6) genetically unclassified CN patients to be positive for acquired mutations in the critical region of G-CSFR. One CN patient with WASP mutation also acquired CSF3R mutation. Based on our sequencing data we would suggest CSF3Rmutation sequencing in all studied groups of patients regardless of mutations in ELANE and HAX1 genes.

Intriguingly, 3 out of 5 CN patients with CSF3R mutations detected by cDNA deep sequencing were negative based on results of previous DNA deep sequencing. All of them were found to acquire low frequency CSF3R mutant clones (ELANE pos. patient with 0.3% of p.Q739* clone; genetically unclassified CN patient with 2% of p.Q749*clone; HAX1pos. patient with 0.9% of p.Q749* clone) in cDNA deep sequencing.

In 2 patients (one CyN-ELANE and one CN-HAX1) with multiple acquired CSF3R mutations we compared mutant clone enrichment in different cell types (BM MNC; BM PMN; PB MNC and PB PMN) by means of cDNA and DNA deep sequencing. In the CyN-ELANE patient with 2 CSF3R mutant clones, the highest mutant allele frequency (MAF) was detected in the cDNA sample of PB PMN (11% of p.Q749* clone and 0.44% of p.Q739* clone), whereas in the PB MNC cDNA sample clone p.Q749* had only 2.5% MAF and clone P.Q739 was not detectable. Similar to that, in the CN-HAX1 patient the highest MAFs for all 3 CSF3R mutant clones were in PB PMN cDNA and the lowest in PB MNC DNA sample. Frequency of mutated CSF3Rclones in BM PMNs of both patients was comparable to PB PMN samples.

Taken together, sequencing of cDNA extracted from peripheral blood or bone marrow PMN samples may provide better results than from MNC in terms of frequency of CSF3R mutation detection in CN and CyN patients. Sequencing of cDNA extracted from BM or PB samples allows enrichment of G-CSFR expressing mutant cells, but due to intrinsic low fidelity of reverse transcriptase the threshold level for positive calls could not be improved significantly (current threshold for candidate calls is 0.2-0.5%).

We would suggest CSF3R mutation screening using deep-sequencing of cDNA from peripheral blood PMN in all patient groups (CN and CyN) for routine diagnostics.