Genomic Landscape of RUNX1-Familial Platelet Disorder with Myeloid Malignancies Reveals Rising Clonal Hematopoiesis

Familial platelet disorder with associated myeloid malignancies (FPDMM) is a rare autosomal dominant disease caused by germline RUNX1 mutations. FPDMM patients have defective megakaryocytic development, low platelet counts, prolonged bleeding times, and a life-long risk (20-50%) of developing hematological malignancies. FPDMM is a rare genetic disease in need of comprehensive clinical and genomic studies. In early 2019 we launched a longitudinal natural history study of patients with FPDMM at the NIH Clinical Center and by May 2021 we have enrolled 98 patients and 100 family controls from 55 unrelated families. Genomic data have been generated from 56 patients in 24 families, including whole exome sequencing (WES), RNA-seq, and single-nucleotide polymorphism (SNP) array. We have identified 21 different germline RUNX1 variants among these 24 families, which include lost-of-function mutations throughout the RUNX1 gene, but pathogenic/likely pathogenic missense mutations are mostly clustered in the runt-homology domain (RHD). As an important form of RUNX1 germline mutations, five splice site variants located between exon 4-5 and exon 5-6 were identified in 6 families, which led to the productions of novel transcript forms that are predicted to generate truncated RUNX1 proteins. Large deletions affecting the RUNX1 gene are also common, ranging from 50 Kb to 1.5Mb, which were detected in 8 of the 55 enrolled families. Besides RUNX1, copy number variation (CNV) analysis from both SNP array and WES showed limited CNV events in non-malignant FPDMM patients. In addition, fusion gene analysis did not detect any in-frame fusion gene in these patients, indicating a relatively stable chromosome status in FPDMM patients. Somatic mutation landscape shows that the overall mutation burden in non-malignant FPDMM patients is lower than AML or other cancer types. However, in 13 of the 44 non-malignant patients (30%), somatic mutations were detected in at least one of the reported clonal hematopoiesis of indeterminate potential (CHIP) genes, significantly higher than the general population (4.3%). Moreover, 85% of our patients who carried CHIP mutations are under 65 years of age; in the general population, only 10% of people above 65 years of age and 1% of people under 50 were reported to carry CHIP mutations. Among mutated genes related to clonal hematopoiesis, BCOR is the most frequently mutated gene (5/44) in our FPDMM cohort, which is not a common CHIP gene among the general population. Mutations in known CHIP genes including SF3B1, TET2, and DNMT3A were also found in more than one patient. In addition, sequencing of 5 patients who already developed myeloid malignancies detected somatic mutations in BCOR, TET2, NRAS, KRAS, CTCF, KMT2D, PHF6, and SUZ12. Besides reported CHIP genes or leukemia driver genes, 3 unrelated patients carried somatic mutations in the NFE2 gene, which is essential for regulating erythroid and megakaryocytic maturation and differentiation. Two of the NFE2 mutations are nonsense mutations, and the other is a missense mutation in the important functional domain. NFE2 somatic mutations may play important roles in developing malignancy because 2 of the 3 patients already developed myeloid malignancies. For multiple patients in our cohort, we have sequenced their DNA on multiple timepoints. We have observed patients with expanding clones carrying FKBP8, BCOR or FOXP1 mutations. We have also observed a patient with relatively stable clone(s) with somatic BCOR, DNMT3A, and RUNX1T1, who have been sampled over more than four years. We will follow these somatic mutations through sequencing longitudinally and correlate the findings with clinical observations to see if the dynamic changes of CHIP clones harboring the mutations give rise to MDS or leukemia. In summary, the genomic analysis of our new natural history study demonstrated diverse types of germline RUNX1 mutations and high frequency of somatic mutations related to clonal hematopoiesis in FPDMM patients. These findings indicate that monitoring the dynamic changes of these CHIP mutations prospectively will benefit patients' clinical management and help us understand possible mechanisms for the progression from FPDMM to myeloid malignancies. Disclosures No relevant conflicts of interest to declare.

FGFR1 Fusion Genes Combine with RUNX1 Mutations to Drive Leukemogenesis

8p11 myeloproliferative syndrome is a myeloid tumor characterized by FGFR1 fusion gene. Its progression is fast, prognosis is bad and pathogenesis is not clear. We screened 10 cases of FGFR1 rearrangement (1 new fusion gene TFG-FGFR1), and 8 of them performed second generations of target exon capture and resequencing. The results showed that RUNX1 gene mutation occurred in 6 cases, with a mutation rate of 75%. The pretest results showed that FGFR1 fusion gene could significantly activate the downstream signaling pathway of FGFR1, and then promote cell proliferation, while RUNX1 gene mutation could block cell differentiation. It is presumed that FGFR1 gene fusion and RUNX1 mutation could lead to leukemia formation. Based on our previous work, we will explore the synergistic pathogenicity, the differences of variable opposite genes and possible mechanisms of FGFR1 fusion gene and RUNX1 mutation in vitro based on cell biology and animal models in vivo, and explore the therapeutic effect of targeted drugs in such diseases. This topic will help clarify the synergistic pathogenic effect of FGFR1 fusion gene and RUNX1 mutation and its related mechanisms, and provide theoretical and experimental evidence for the diagnosis and treatment of this disease. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Rhesus Macaques As a Preclinical Model Organism for RUNX1-FPD

Germline loss-of-function (LOF) heterozygous mutations in the central hematopoietic transcription factor RUNX1 gene cause the marrow failure/malignancy predisposition syndrome Familial Platelet Disorder with associated Myeloid Malignancies (FPDMM, or FPD). Patients with FPD have defective megakaryocytic development, low platelet counts, and a very high (35-50%) life-long risk of hematological malignancies, particularly MDS/AML. Murine heterozygous gene knockout models do not recapitulate the human phenotype in terms of thrombocytopenia or myeloid leukemia progression. Although gene correction of the RUNX1 mutation in hematopoietic stem and progenitor cells (HSPCs) is being considered as a possible treatment approach, it is unknown whether mutation-corrected HSPCs will have the hoped for advantage over RUNX1 mutant HSCs in vivo, likely necessary to significantly lower leukemia risk. In order to study the relative function of wildtype and RUNX1-mutated HSPCs in vivo in a model with close hematopoietic similarity to humans, we generated a rhesus macaque FPD competitive repopulation model via CRISPR/Cas9 NHEJ editing of the RUNX1 gene versus the AAVS1 safe-harbor control locus. We transplanted mixtures of autologous HSPCs edited at the two loci: 75% RUNX1-edited/25% AAVS1- edited CD34+ HSPCs in animal 1 and 25% RUNX1-edited/75% AAVS1-edited CD34+ HSPCs in animal 2, following conditioning with total body irradiation. Both animals engrafted tri-lineage hematopoiesis promptly following transplantation. However, platelet numbers remained below the normal range long-term in animal 1 receiving a higher ration of RUNX1-edited HSPCS and below counts of macaques receiving HSPCs edited at other loci (Figure 1). Bone marrow morphology at 6 months was normal. To assess the HSPC function of RUNX1 mutant versus AAV1 control and unedited WT cells we tracked RUNX1 and AAVS1-mutated allele frequencies in blood cells over time via deep sequencing (Figure 2). In the infusion products (IP), allele fractions reflected the desired ratios. In both animals, AAVS1-edited cells dominated compared to RUNX1-edited cells. However, in animal 1, RUNX1-mutated cells expanded over time eventually exceeding the ratio in the IP, and in animal 2, levels of RUNX1 and AAVS1-mutated cells were equivalent long-term. Marrow analyzed at 6 months showed heterozygous RUNX1-mutated CFU at levels concordant with mutation frequencies in the blood, but no homozygous RUNX1 mutated CFU, suggesting homozygous LOF is not compatible with long-term HSPC function. In conclusion, we have created pre-clinical model for FPD via CRISPR/Cas editing of HSPCs in rhesus macaques. The lack of a competitive advantage for wildtype or control-locus edited HSPCs over RUNX1 heterozygous-mutated HSPCs long-term in our model suggests that gene correction approaches for FPD will be challenging, particularly to reverse the MDS/AML predisposition phenotype. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Immunonano-Lipocarrier-Mediated Liver Sinusoidal Endothelial Cell-Specific RUNX1 Inhibition Impedes Immune Cell Infiltration and Hepatic Inflammation in Murine Model of NASH

Background: Runt-related transcription factor (RUNX1) regulates inflammation in non-alcoholic steatohepatitis (NASH). Methods: We performed in vivo targeted silencing of the RUNX1 gene in liver sinusoidal endothelial cells (LSECs) by using vegfr3 antibody tagged immunonano-lipocarriers encapsulated RUNX1 siRNA (RUNX1 siRNA) in murine models of methionine choline deficient (MCD) diet-induced NASH. MCD mice given nanolipocarriers-encapsulated negative siRNA were vehicle, and mice with standard diet were controls. Results: Liver RUNX1 expression was increased in the LSECs of MCD mice in comparison to controls. RUNX1 protein expression was decreased by 40% in CD31-positive LSECs of RUNX1 siRNA mice in comparison to vehicle, resulting in the downregulation of adhesion molecules, ICAM1 expression, and VCAM1 expression in LSECs. There was a marked decrease in infiltrated T cells and myeloid cells along with reduced inflammatory cytokines in the liver of RUNX1 siRNA mice as compared to that observed in the vehicle. Conclusions: In vivo LSEC-specific silencing of RUNX1 using immunonano-lipocarriers encapsulated siRNA effectively reduces its expression of adhesion molecules, infiltrate on of immune cells in liver, and inflammation in NASH.

Transcriptional Regulation of RUNX1: An Informatics Analysis

The RUNX1/AML1 gene encodes a developmental transcription factor that is an important regulator of haematopoiesis in vertebrates. Genetic disruptions to the RUNX1 gene are frequently associated with acute myeloid leukaemia. Gene regulatory elements (REs), such as enhancers located in non-coding DNA, are likely to be important for Runx1 transcription. Non-coding elements that modulate Runx1 expression have been investigated over several decades, but how and when these REs function remains poorly understood. Here we used bioinformatic methods and functional data to characterise the regulatory landscape of vertebrate Runx1. We identified REs that are conserved between human and mouse, many of which produce enhancer RNAs in diverse tissues. Genome-wide association studies detected single nucleotide polymorphisms in REs, some of which correlate with gene expression quantitative trait loci in tissues in which the RE is active. Our analyses also suggest that REs can be variant in haematological malignancies. In summary, our analysis identifies features of the RUNX1 regulatory landscape that are likely to be important for the regulation of this gene in normal and malignant haematopoiesis.

Towards prevention of childhood ALL by early-life immune training

B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is the most common form of childhood cancer. Chemotherapy is associated with life-long health sequelae and fails in approximately 20% of cases. Thus, prevention of leukemia would be preferable to treatment. Childhood leukemia frequently starts before birth, during fetal hematopoiesis. A first genetic hit (e.g. the ETV6-RUNX1 gene fusion) leads to the expansion of pre-leukemic B-cell clones, which are detectable in healthy newborn cord blood (up to 5%). These pre-leukemic clones give rise to clinically overt leukemia in only about 0.2% of carriers. Experimental evidence suggests that a major driver of conversion from the pre-leukemic to the leukemic state is exposure to immune challenges. Novel insights have shed light on immune host responses and how they shape the complex interplay between (A) inherited or acquired genetic predispositions, (B) exposure to infection, and (C) abnormal cytokine release from immunologically untrained cells. Here, we integrate the recently emerging concept of "trained immunity" into existing models of childhood BCP-ALL and suggest future avenues towards leukemia prevention.

Characterization of the Platelet Phenotype Caused by a Germline RUNX1 Variant in a CRISPR/Cas9-Generated Murine Model

RUNX1-related disorder (RUNX1-RD) is caused by germline variants affecting the RUNX1 gene. This rare, heterogeneous disorder has no specific clinical or laboratory phenotype, making genetic diagnosis necessary. Although international recommendations have been established to classify the pathogenicity of variants, identifying the causative alteration remains a challenge in RUNX1-RD. Murine models may be useful not only for definitively settling the controversy about the pathogenicity of certain RUNX1 variants, but also for elucidating the mechanisms of molecular pathogenesis. Therefore, we developed a knock-in murine model, using the CRISPR/Cas9 system, carrying the RUNX1 p.Leu43Ser variant (mimicking human p.Leu56Ser) to study its pathogenic potential and mechanisms of platelet dysfunction. A total number of 75 mice were generated; 25 per genotype (RUNX1WT/WT, RUNX1WT/L43S, and RUNX1L43S/L43S). Platelet phenotype was assessed by flow cytometry and confocal microscopy. On average, RUNX1L43S/L43S and RUNX1WT/L43S mice had a significantly longer tail-bleeding time than RUNX1WT/WT mice, indicating the variant's involvement in hemostasis. However, only homozygous mice displayed mild thrombocytopenia. RUNX1L43S/L43S and RUNX1WT/L43S displayed impaired agonist-induced spreading and α-granule release, with no differences in δ-granule secretion. Levels of integrin αIIbβ3 activation, fibrinogen binding, and aggregation were significantly lower in platelets from RUNX1L43S/L43S and RUNX1WT/L43S using phorbol 12-myristate 13-acetate (PMA), adenosine diphosphate (ADP), and high thrombin doses. Lower levels of PKC phosphorylation in RUNX1L43S/L43S and RUNX1WT/L43S suggested that the PKC-signaling pathway was impaired. Overall, we demonstrated the deleterious effect of the RUNX1 p.Leu56Ser variant in mice via the impairment of integrin αIIbβ3 activation, aggregation, α-granule secretion, and platelet spreading, mimicking the phenotype associated with RUNX1 variants in the clinical setting.