Mutation Burden in Hematopoietic Stem Cells Is Not Increased in Congenital Neutropenia

Transformation to a myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) is a major clinical concern in patients with severe congenital neutropenia (SCN) and Shwachman-Diamond syndrome (SDS), with a cumulative risk of MDS/AML of more than 20%. The molecular mechanisms contributing to transformation to MDS/AML in these disorders are poorly understood, limiting the development of new therapies or strategies for risk stratification or early detection. Hematopoietic stem cells (HSCs) accumulate mutations with age (Welch et al, Cell, 2012). With each cell division, there is a finite chance of acquiring mutations in an HSC genome. In human cells, the mutation rate per cell division is highly variable. Thus, the number of mutations in HSCs is dependent on the cumulative number of HSC cell divisions and the mutation rate per cell division. We hypothesize that maladaptive compensatory changes induced by chronic severe neutropenia, including increased replicative stress, lead to a higher rate of mutation accumulation in HSCs, which contributes to the high rate of transformation to MDS/AML. To test this hypothesis, we measured the mutation burden in individual hematopoietic stem/progenitor cells (HSPCs) from patients with congenital neutropenia. Moreover, since increased G-CSF signaling has been implicated in leukemic transformation, we tested whether prolonged G-CSF therapy increases the HSPC mutation burden in mice.

We analyzed blood or bone marrow samples from four cohorts: SCN (n=13), SDS (n=4), cyclic neutropenia (CN) (n=5), and age-matched healthy donors (n=9); patients with MDS, AML, or identified clonal hematopoietic abnormalities were excluded. Single CD34+ lineage- CD38- cells were sorted from blood or bone marrow samples and cultured for 3-4 weeks on irradiated stromal feeder cells. The exomes of a minimum of 4 expanded HSPC clones were sequenced per patient; unsorted hematopoietic cells from the same patient served as the "normal control". A total of 124 exomes were sequenced using the NimbleGen SeqCap EZ Exome Kit v3.0, which captures 64 Mb of target DNA. Somatic variants with a variant allele frequency (VAF) greater than 25% were considered clonal somatic mutations (average VAF was 48%). The number of mutations per HSPC clone did not differ significantly between the cohorts. The average number of mutations per exome was 4.5 ± 0.6 for SCN, 1.6 ± 0.3 for SDS, 6.4 ± 1.4 for CN, and 3.4 ± 0.5 for the healthy controls. Moreover, linear regression analyses showed that the number of mutations per HPSC exome per year was similar in all cohorts. To investigate the effect of long term G-CSF treatment on mutation accumulation in HSPCs, we treated wild type or Csf3rd715X mice (which have increased G-CSF signaling) with G-CSF for 6 months. Mutation burden in expanded HSPC clones was measured by exome sequencing, essentially as described above. The average number of mutations per HSPC clone was similar in mice treated with G-CSF versus control. In wild type mice, 1.2 ± 0.4 mutations per HSPC exome were detected in saline-treated mice versus 1.8 ± 0.2 in G-CSF treated mice (n = 6 mice in each cohort, P=NS). In Csf3rd715X mice, there were 1.0 ± 0.2 mutations per HSPC exome in saline-treated mice versus 1.4 ± 0.2 in G-CSF treated mice (n = 5 mice in each cohort, P=NS). We conclude that 6 month exposure to G-CSF in mice does not increase mutation burden in HSPCs.

Collectively, these data suggest that an increased mutation rate in HSCs is not responsible for the increased risk of transformation to MDS/AML in SCN or SDS. Studies are underway to determine whether the incidence of clonal hematopoiesis is higher in congenital neutropenia.