Functional Classification of Germline RUNX1 Variants Identified in Patients Suspected of Familial Platelet Disorder with Associated Myeloid Malignancies

RUNX1 is a master regulator of hematopoiesis. A highly conserved runt-homology domain (RHD) located near the N-terminus is responsible for DNA binding and interaction with its heterodimeric partner CBFβ, whereas the C-terminus contains the transactivation domain (TAD). Somatic variants and chromosomal abnormalities affecting RUNX1 are commonly observed in hematological malignancies, such as myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). Heterozygous germline RUNX1 variants are responsible for familial platelet disorder with associated myeloid malignancies (FPDMM), an autosomal dominant disorder with incomplete penetrance and a broad spectrum of clinical phenotypes, including thrombocytopenia, functional platelet defects, and an increased risk (35-45%) of developing MDS and AML. In 2019, our group launched a longitudinal natural history study of patients with germline RUNX1 variants to increase our understanding of FPDMM, which will hopefully lead to better clinical management and potentially new therapies. The majority of RUNX1 variants are private with limited segregation data available, making variant interpretation challenging. A significant number of cases have variants of unknown significance (VUS) in RUNX1. Therefore, it is important to determine if these VUS are pathogenic (or benign) in order to improve the diagnosis and clinical management for these individuals.

To functionally characterize each variant, we used well-established in vitro functional assays to assess the damaging effect of each RUNX1 VUS, including transactivation reporter assays, electrophoretic mobility shift assays, co-immunoprecipitation, immunofluorescence, and western blot with cell fractionation. All assays include RUNX1 wildtype and known benign and pathogenic controls, as well as co-expression with its binding partner CBFβ. The VUS Ser94Arg located in RHD of RUNX1 has showed impaired CSF1R transactivation (< 20% of wildtype), reduced subcellular co-localization of RUNX1 with CBFβ, and reduced heterodimerization with CBFβ. These results together with population frequency data, in silico prediction, phenotype and segregation data support the reclassification of this variant into the likely pathogenic category. On the other hand, the VUS Asp123His shows no damaging effect in vitro, and could be reclassified to likely benign. Using these assays, we have confirmed or reclassified several RUNX1 variants using the criteria set by the ACGM Myeloid Malignancy VCEP (Luo et al, Blood Adv. 2019; 3:2962). Together, these results highlight the importance of using in vitro functional assays for RUNX1 variant classification.

Moreover, we sought to determine the potential of using platelet transcriptomics for RUNX1 variant classification and to gain a deeper understanding of FPDMM pathogenesis. Clariom™ D microarrays identified 218 differentially expressed genes (DEG; absolute fold change ≥ 2; q-value < 0.05) between FPDMM patients and healthy controls, with 90 genes (41%) being upregulated and 128 genes (59%) being downregulated in patients. Hierarchical clustering showed that the selected FPDMM patients with pathogenic and likely pathogenic RUNX1 variants clustered together and separately from healthy controls, with well-established pathogenic variants, such as Arg201Ter and Arg201Gln, at the opposite end of the dendrogram, and likely pathogenic variants such as Tyr403Cysfs153 closer to healthy controls. Our results show that FPDMM patients exhibit reduced expression of important genes related to thrombocytopenia and platelet dysfunction. Furthermore, several top upregulated genes identified in FPDMM patient's platelets are associated with poorer outcomes in AML patients, according to the TCGA dataset.

Overall, our study shows the importance of performing functional assays to aid RUNX1 variant classification and to facilitate the study of FPDMM pathogenesis. We are currently analyzing the platelet proteome data of FPDMM patients by LC-MS/MS, which will be reported at the annual meeting.