Somatic Mutational Landscape of Hereditary Hematopoietic Malignancies Associated with Germline Variants in RUNX1, GATA2 and DDX41

BACKGROUND: Germline variants in RUNX1, GATA2 and DDX41 may confer a predisposition to hereditary haematopoietic malignancies (HHMs) such as MDS and AML yet have distinct age ranges of malignancy diagnosis and a highly variable overall risk for leukemogenesis. The increased awareness and identification of carriers of these germline variants, particularly before development of malignancy, has changed the way in which individuals and families need to be managed in the clinic. Individuals need lifelong monitoring and may also need modification to treatments when malignancy does develop, compared to sporadic counterparts. Gaps in understanding pre-malignant states in HHM syndromes have hampered efforts to design effective clinical surveillance regimes, provide personalized pre-emptive treatments, and appropriate counselling to patients.

METHODS: We have addressed some of these knowledge gaps using the largest known international comparative cohort of germline RUNX1, GATA2, or DDX41 variant carriers, which includes carriers both without a malignancy diagnosis “unaffected-carriers” and those with a malignancy “affected-carriers”. Our cohort included 245 patient samples evenly distributed between unaffected-carrier (n=122) and affected-carrier (n=123) samples. We utilized a uniform bioinformatics approach, to identify somatic variants from next generation sequencing data provided by our network of investigators.

RESULTS: We identified clonal hematopoiesis (CH) in 34% and 38% of RUNX1 and GATA2 unaffected-carriers, respectively (Figure 1A top), with overlap in the ages of observation CH and malignancy in different individuals observed (Figure 1A bottom). In contrast CH was seen in only 4% of DDX41 unaffected-carriers, significantly less frequently than RUNX1 or GATA2 (p=0.002). RUNX1- and GATA2 variant carriers were at an increased risk of early onset CH (0-29 years) present in 19% of RUNX1 carriers and 25% of GATA2 variant carriers compared to 0.2% of the general population (PMID: 25426837; PMID: 25426838). RUNX1 unaffected-carriers had an increased prevalence of CH at all ages compared to population controls, with BCOR being the most frequently mutated gene in RUNX1 carriers, present across the entire age spectrum from as young as 16 years to 76 years (Figure 1B). In RUNX1-driven CH specifically we detected “high-risk” variants in TET2, PHF6 and most frequently in BCOR, all of which were also recurrently mutated in RUNX1-driven malignancies, suggesting CH is a direct precursor to malignancy. Consistent with a multi-step process towards leukaemia, we found both an increase in clone size (variant allele frequency, VAF) and the number of variants present per sample, as the age of carriers increased (Figure 1B).

Tumor profiling similarly highlighted differences in the pathways to myeloid malignancy development for the three syndromes. Leukemogenesis in both RUNX1 and DDX41 carriers was often driven by second-hit variants in RUNX1 and DDX41 respectively. Somatic RUNX1 variants were not seen in germline RUNX1 unaffected-carriers, only in tumors, therefore this likely represents a “late” leukemogenic event. This was not able to be inferred for somatic DDX41 variants in DDX41 related predisposition where mutations before malignancy were rarely seen. In contrast, GATA2 germline driven malignancy was not associated with somatic mutation of GATA2, instead this process favored mutations in ASXL1. Both germline RUNX1 and DDX41 cohorts presented with a sex bias for HM development, but in opposite directions (AML 1:2.5 (RUNX1) and HM 3:1 (DDX41), Male:Female). Disruption of RUNX1 regulation of estrogen signalling, and DDX41 mediated exaggeration of the known sex difference in innate immunity, are postulated as mechanisms worth investigation to explain these observed sex biases (PMID: 25479752; PMID:28223406).

CONCLUSIONS: This study lays the foundation for developing novel-preventative therapies and the implementation of gene-specific clinical monitoring requirements for carriers of germline variants in RUNX1, DDX41 and GATA2. Individuals with RUNX1 variants will require regular monitoring for somatic variants in several genes throughout their lifetime, while DDX41 carriers are likely to benefit from monitoring throughout adulthood for driver DDX41 variants using sensitive technology to detect low-frequency initiating events.