Building a Clinical Genetic GATA2 Variant Curation System with Activity Metrics

Human GATA2 genetic variation causes GATA2 deficiency syndrome, which can lead to myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). While pathogenic GATA2 variants alter the important zinc fingers and inter-zinc finger spacer, cellular aberrations underlying the disease predisposition are not established (Bresnick et al., 2020). Furthermore, as many questions exist regarding how GATA2 controls genome function, ascribing the functional significance of variants can be challenging. Our analysis of GATA2 variants in the United States population through the All of Us Research Program database (All of Us Research Program Genomics Investigators, 2024) revealed rare zinc finger variants, e.g., C295S, R308W, T356N (1/490,000+ alleles), that had not been described. Common variants included A164T (91,211/490,000+) and P250A (962/490,000+) in the N-terminus. To advance genetic curation, we established mouse and human genetic rescue systems to generate functional signatures that discriminate GATA2 from pathogenic variants.

GATA2 regulates its expression through two enhancers (-77 kb and +9.5 kb) required for hematopoietic development and function (Johnson et al., 2012; 2015). We are identifying essential GATA2-regulated enhancers genome-wide and described the importance of the GATA2-activated Cebpe +6 kb enhancer (Katsumura et al., 2024). Deletion of this enhancer in progenitors downregulated C/EBPε, expression of a GATA2-regulated gene cohort, and, surprisingly, genes regulated by the pathogenic variant T354M, which exhibits loss-of-function and gain-of-function activities. This analysis identified C/EBPε as a cooperating transcription factor with GATA2 and T354M, and Cebpe regulation did not discriminate GATA2 from T354M activity.

Using multiomics with the genetic rescue system in murine Gata2 -77-enhancer-deleted myeloid progenitors (hi-77^-/-) with ~75% lower GATA2, we identified metrics that discriminate GATA2 from a pathogenic variant. We compared how GATA2 and a human germline GATA2 pathogenic variant (9 amino acid insertion between the zinc fingers, 9aa-Ins) function genome-wide (Jung et al., 2023). GATA2 deficiency in hi-77^-/- cells upregulated expression of innate immune genes, including Il6st and Il6ra encoding cytokine receptor subunits IL6ST and IL6RA. GATA2, but not 9aa-Ins, reversed elevated Il6st and Il6ra expression and IL-6/STAT3 signaling. ATAC-seq revealed accessible chromatin 24 kb and 30 kb upstream of Il6st in hi-77^-/- cells or hi-77^-/- with 9aa-ins, but not wild type cells or hi-77^-/- rescued with GATA2. To ask whether these sites are enhancers mediating IL-6 signaling, CRISPR-Cas9 gene editing was used to delete the sequences in hi-77^-/- cells. Il6st expression decreased 13-fold (P<0.0001) and IL-6/STAT3 signaling decreased 85% (P<0.0001). Thus, GATA2 deficiency commissions enhancers that elevate Il6st expression and IL-6 signaling. To determine whether GATA2 deficiency commissions enhancers essential for other GATA2-repressed cytokine/chemokine receptor genes, we analyzed accessible chromatin in hi-77^-/- cells at all cytokine/chemokine receptor genes and identified prospective enhancers that are GATA2-, but not 9aa-Ins-, regulated. These included the Csf1r +2.6 kb enhancer (Rojo et al., 2019). As enhancer decommissioning has utility to discriminate GATA2 from a pathogenic variant, additional clinical variants are being analyzed. We are also analyzing whether variants are defective in utilizing transcription factors and coregulators (C/EBPε and SMARCD2 for activation and RUNX1 and PU.1 for repression) that functionally interact with GATA2 at enhancers, and are extending murine analyses to human with a GATA2^+/- HUDEP2 erythroid progenitor rescue system.

In aggregate, diverse metrics related to genome regulation, with and without inflammation, as well as subcellular localization and protein stability, are being used to innovate a machine learning-based classifier strategy to ascribe whether any variant most closely resembles GATA2 or pathogenic variants. Using a minimal metric cohort that rigorously segregates GATA2 from clinical variants, we will streamline assays to ensure rapid, yet definitive, variant curation, compatible with clinical decision making. This curation strategy is being extended to decipher genetic variation in other genes that create an MDS and AML predisposition.