AUTHORS
Jacob Kurek, B.s., Crystal Potter, B.a.2, Xiaoqin Zhu, MD, MS3, Corbin R Azucenas, BS4, Devyani Sharma, MS5, Jan Frayne, PhD6, Marie-Dominique Filippi, PhD7, Katie Seu, PhD8, Theodosia M Kalfa Jr, MD, PhD9, Andrew Volk, PhD10
AFFILIATIONS
1 Cincinnati Children's Hospital, Cincinnati, Ohio, United States, University of Cincinnati, Cincinnati, Ohio, United States
2 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
3 Cincinnati Children's Hospital, Cincinnati, Ohio, United States, University of Cincinnati, Cincinnati, Ohio, United States
4 Cincinnati Children's Hospital, Cincinnati, Ohio, United States, University of Cincinnati, Cincinnati, Ohio, United States
5 Cincinnati Children's Hospital, Cincinnati, Ohio, United States, University of Cincinnati, Cincinnati, Ohio, United States
6 University of Bristol, Bristol, United Kingdom
7 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
8 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
9 Cincinnati Children's Hospital, Cincinnati, Ohio, United States
10 Cincinnati Children's Hospital, Cincinnati, Ohio, United States, University of Cincinnati, Cincinnati, Ohio, United States
ABSTRACT
Erythropoiesis is the most proliferative tissue in the human body. Following commitment into terminal erythropoiesis, each erythroblast must simultaneously condense and replicate chromatin to prepare for hemoglobinization and enucleation. Loss of control over chromatin structure or disruption of replication could have catastrophic consequences for erythropoiesis. In fact, a diagnostic hallmark of Congenital Dyserythropoietic Anemia Type I (CDA-I) is ineffective erythropoiesis accompanied by disruption of heterochromatin propagation in erythroblasts, leading to binucleated erythroblasts and nuclear heterochromatin with a Swiss cheese-like appearance under transmission electron microscopy. CDA-I is genetically driven by heritable mutations in Codanin-1, a protein whose exact function remains poorly understood, but recent crystallography suggests altered interaction kinetics with the histone chaperone ASF1. Thus, Codanin-1 mutations may disrupt chromatin assembly during replication. However, the mechanistic contribution of replication-dependent chromatin assembly to erythropoiesis is poorly understood. Using genetically engineered mouse models in vivo and the human BEL-A cell line in vitro, we have demonstrated that replication-dependent chromatin assembly is necessary for terminal erythroid maturation.
Replication-dependent chromatin assembly is mediated through the Chromatin Assembly Factor 1 (CAF-1) complex, comprised of three subunits (CHAF1A, CHAF1B, and RBBP4). CAF-1 facilitates histone H3/H4 heterodimer deposition at the replication fork with ASF1. We hypothesized that disruption of the histone supply and chromatin assembly during replication drives CDA-I pathogenesis. We tested this by assessing genetic deletion of the CAF-1 subunit CHAF1B in BEL-A cells in vitro, and knockout of Chaf1b in vivo in mouse erythroid cells. In the Rosa26CreERT2 Chaf1bfl/fl model, we observed a rapid loss of terminal erythropoiesis within 48 to 72 hours following Chaf1b deletion. BFU-e/CFU-e progenitors were the most mature erythroid cells persisting without CHAF1B, though they were unable to progress to proerythroblasts. Rosa26CreERT2 has rapid and robust deletion kinetics, though not erythroid-specific due to its broad expression throughout the body. Therefore, we confirmed the contribution of CHAF1B to hematopoiesis using a hematopoietic-specific Chaf1b deletion in VavCreERT2 Chaf1bfl/fl primary mice. Finally, we confirmed the specificity to erythropoiesis using EpoRCreChaf1bfl/fl which resulted in embryonic lethality and no Chaf1bD/D pups born. Together, our work demonstrates that replication-dependent chromatin assembly is mechanistically necessary for terminal erythropoiesis in primary murine models.
To probe the mechanistic contribution of replication dependent chromatin assembly to erythroid maturation, we turned to the BEL-A cell line, a model of early erythroid precursors that can be differentiated from proerythroblast stage to enuclueation as a model of terminal erythropoiesis. Cells tolerated CHAF1B loss for up to seven days resulting in cell cycle arrest and G0/G1 cell cycle arrest. Prior to this, we observed widespread accumulation of multi-nuclear dysplastic cells due to centrosome duplication in >50% of dividing cells. We observed similar morphological dysplasia in primary BFU-e/CFU-e from Chaf1b deleted mice at 72 and 96 hours, but not at 48 hours when erythropoiesis halts, indicating a cell intrinsic defect is manifesting prior to morphological dysplasia. The chromatin assembly function of CHAF1B is necessary to prevent binucleation, as G1er cells expressing a CHAF1A dominant negative allele preventing CAF-1 association with the replication fork also become bi-nucleated. These phenocopy CDA-I where, in addition to the chromatin phenotype, bi-nucleated erythroblasts accompany terminal erythropoiesis failure. Of note, CHAF1B depletion does not precipitate multinucleation in other healthy or malignant hematopoietic models, suggesting maturing erythroid cells are particularly reliant on CAF-1 and replication-dependent chromatin assembly to maintain erythropoiesis.
These results demonstrate that terminal erythropoiesis is reliant on replication-dependent chromatin assembly. In the context of CDA-I, there is potential for the dysregulation of chromatin structure to be the driving factor towards dyserythropoiesis and strategies that restore chromatin organization fidelity could serve as promising therapeutic interventions.
DISCLOSURES
Jacob Kurek, B.s.: No relevant conflicts of interest to declare.