Transcription factor genetics and biology in predisposition to bone marrow failure and hematological malignancy

Abstract

Transcription factors (TFs) play a critical role as key mediators of a multitude of developmental pathways, with highly regulated and tightly organized networks crucial for determining both the timing and pattern of tissue development. TFs can act as master regulators of both primitive and definitive hematopoiesis, tightly controlling the behavior of hematopoietic stem and progenitor cells (HSPCs). These networks control the functional regulation of HSPCs including self-renewal, proliferation, and differentiation dynamics, which are essential to normal hematopoiesis. Defining the key players and dynamics of these hematopoietic transcriptional networks is essential to understanding both normal hematopoiesis and how genetic aberrations in TFs and their networks can predispose to hematopoietic disease including bone marrow failure (BMF) and hematological malignancy (HM). Despite their multifaceted and complex involvement in hematological development, advances in genetic screening along with elegant multi-omics and model system studies are shedding light on how hematopoietic TFs interact and network to achieve normal cell fates and their role in disease etiology. This review focuses on TFs which predispose to BMF and HM, identifies potential novel candidate predisposing TF genes, and examines putative biological mechanisms leading to these phenotypes. A better understanding of the genetics and molecular biology of hematopoietic TFs, as well as identifying novel genes and genetic variants predisposing to BMF and HM, will accelerate the development of preventative strategies, improve clinical management and counseling, and help define targeted treatments for these diseases.

Keywords: bone marrow failure (BMF); germline; hematological malignancies (HM); pathogenic variant; transcription factor.

Figure 1

Impact of germline pathogenic variants in GATA2 on hematopoietic network and leukemic transformation. Schematic of the role of GATA2 interactions in transcriptional networks in normal hematopoiesis including interacting partners, target genes and upstream regulators. GATA2, a common target of aberration in leukemias, leads to inappropriate transcriptional programs resulting in TF dysregulation, impaired differentiation, and subsequent expansion of immature cell populations to initiate malignancy. While based on published data, the figure is somewhat hypothetical as each hematopoietic TF gene regulatory region is bound by a different combination, clustering, and arrangement of these TFs. Pathogenic germline variants (lightning bolt) may lead to disrupted DNA binding and/or interactions with other TFs (dashed arrows) resulting in a range of impacts on downstream target genes that may include disruptions to additive, synergistic and/or inhibitory transcriptional events. This image was created using BioRender.com.

Figure 2

Transcription factor expression in hematopoietic cells in human hematopoiesis Vs AML. Heatmap correlating the expression levels of TFs in hematopoietic cells in human hematopoiesis with AML. The mRNA expression levels of microarray data (log2) in the ‘Normal human hematopoiesis (DMAP)’ and ‘normal hematopoiesis with AML’ datasets from BloodSpot were used, where the probe with the overall highest intensity was selected and each quadruplicate/triplicate/duplicate averaged. TFs included those selected in Table 1 (56). common myeloid progenitor, CMP; dendritic cell, DC; Erythroid, Eryth; granulocyte/monocyte progenitor, GMP;CFU; colony forming unit, granulocyte, Gran; megakaryocyte/erythroid progenitor, MEP; mature natural killer, NK; natural killer T cell, NKT; memory. mem.

Figure 3

Pathogenicity predictors for TFs implicated in predisposition to BMF and HM. Comparison of pathogenicity predictor score for known predisposition TFs and potential TF genes isolated from Table 1. (A) The Probability of being loss-of-function (LOF) intolerant (pLI) score of each gene was collated using the gnomAD database (version 2.1.1). The pLI score reflects the tolerance of a given gene to the LOF based on the number of protein-truncating variants referenced in control databases weighted by the size of the gene and the sequencing coverage. The pLI score ranges from 0-1, where higher the score, the higher the intolerance of the gene. (B) Percentage of gene LOF variants in COSMIC database. The total number of somatic LOF variants within a particular gene were calculated by totaling positive mutation data for the selected gene. Variants called ‘LOF’ included nonsense substitutions, frameshift insertions and frameshift deletions. Percentage calculated using LOF variants over the total number of unique samples of each gene. (C) Percentage of point mutations observed in hematopoietic and lymphoid tissues. The distribution of mutations across the primary hematopoietic and lymphoid tissues curated by COSMIC were collated. The percentages were calculated by totaling the number of point mutations of each gene, over the total samples tested.

Figure 4

Gene Ontology of TFs implicated in hematological development and disease. Using ShinyGO0.77 analysis software (31), the known BMF/HM predisposition list in Table 1 and the potential list of BMF/HM predisposition genes in Table 4 (including ERG, GATA3 and SPI1) were analyzed for enrichment of GO terms in biological process and Molecular functions using all known human TF genes as background (i.e., normalization). All protein coding genes were used for background in GO cellular component analysis as there was no enrichment when using TFs as background.