ALBERT

Description: There is a constant need for red blood cells for transfusion therapy in the treatment of anemias and acute injury. As all blood products for transfusion come from donors, there are concerns over shortages and safety. Furthermore, many patients with transfusion-dependent anemias risk alloiumminization. The in vitro production of red blood cells would address these problems, especially as they can be genetically engineered to prevent alloimmunization. Numerous erythroid culture systems now exist for the in vitro production of red blood cells. Hematopoietic stem and progenitor cells (HSPCs) obtained from umbilical cord or peripheral blood can be differentiated into erythrocytes, however, they are limited in expansion. While umbilical cord HSPCs have greater expandability than peripheral blood, the resulting erythrocytes contain fetal globins. Pluripotent stem cells can also be used as a starting source, however only a small percentage of the cells can be differentiated into erythroblasts which also suffer from low enucleation rates. Presently, the cost of in vitro production of a unit of red cells is greater than an order of magnitude higher than obtaining it from a donor largely due to the medium and cytokine costs (Timmins & Nielsen, Trends Biotechnol, 2009). A relatively new approach of immortalizing early erythroblasts allowing unlimited expansion as well as terminal maturation and enucleation shows great therapeutic promise (Kurita et al., PLoS One, 2013; Huang et al., Mol Ther, 2014; Trakarnsanga et al., Nat Commun, 2017). However, these immortalized erythroblasts are still reliant on two costly cytokines: stem cell factor (SCF) and erythropoietin (Epo). Mutations in exon 17 of the receptor tyrosine kinase gene KIT are frequently seen in acute myeloid leukemias, gastrointestinal stromal tumors, and mast cells leading to mastocytosis. These mutations cause the c-Kit protein to spontaneously activate and transduce signal in the absence of SCF (Kit-ligand). To generate an SCF-independent HUDEP-2 cell line (Kurita et al., PLoS One, 2013), we used CRISPR/Cas9 to introduce missense and frameshifting mutations within the vicinity of Asp816 in exon 17 of the KIT gene. The resulting monoclonal cell lines were selected for by removing SCF from the expansion medium and were subsequently named KIT-CAT (KIT with Constitutively Activating Transformation). To better understand what KIT mutations allowed or impaired terminal maturation, monoclonal cell lines were genotyped by Sanger sequencing. Three cell lines with unique genotypes were chosen for further analysis. All three KIT-CAT lines had a shorter doubling time compared to HUDEP-2 cells (16.7 vs 18.9 hrs, p=0.020) and were no longer dependent on SCF or Epo. However, two of the three KIT-CAT lines showed more robust proliferation with Epo in the expansion medium. The addition of SCF to the medium caused no increase in c-Kit activation by Western blotting for phosphorylation at Tyr703. Furthermore, the low molecular weight and immature form of c-Kit is also phosphorylated in KIT-CAT cells, but not HUDEP-2 cells, indicating c-Kit activation occurs before trafficking to the cell membrane where SCF would bind (Tabone-Eglinger et al., Clin Cancer Res, 2008). Key features of erythroblast maturation are the decrease in cell and nuclear size which can be measured using imaging flow cytometry (McGrath et al., Methods, 2017). While in expansion phase, all 3 cell lines were larger in cell and nuclear area compared to the parental HUDEP-2 line. By day 6 of maturation, all three cell lines had statistically significant decreases in cell and nuclear size indicating maturation. By day 13 of culture, Wright-Giemsa staining showed that the majority of the cells were orthochromatic erythroblasts or enucleate reticulocytes. Reducing cell culture costs is needed for in vitro manufacturing of red blood cells to be economically feasible. These results show that a c-Kit activating mutations in human erythroblasts removes the cost of SCF and reduces the cost of Epo while still allowing for terminal maturation and enucleation. Disclosures No relevant conflicts of interest to declare.

Description: Maturation of erythroid progenitors is associated with significant changes in gene expression in the context of a nucleus that dramatically decreases in size in preparation for enucleation, and is regulated by the coordinated action of transcriptional regulators and epigenetic modifiers. In eukaryotes, all DNA is bound by histone proteins into chromatin. Posttranslational modifications of the N-terminal "tails" of these proteins are key regulators of chromatin structure and gene expression. We hypothesized that terminal erythroid maturation is associated with changes in the abundance of specific histone posttranslational modifications. To address this hypothesis, we utilized mass spectrometry to perform an unbiased assessment of the abundance histone post translational modifications in maturing erythroblasts. We cultured peripheral blood CD34+ hematopoietic stem and progenitor cells (HSPCs) down the erythroid lineage using a semi-synchronous culture system (as outlined in Gautier et al. Cell Reports 2016), and sent cells for mass spectrometry on day 7 of erythroid maturation, when the cells are predominately basophilic erythroblasts, and on day 12 of erythroid maturation, when they are predominately poly- and ortho- chromatic erythroblasts. The maturation stage of the cells was confirmed by both cytospins and imaging flow cytometric analyses. Two independent replicates were performed and key results confirmed by western blotting. Terminal erythroid maturation was associated with a dramatic decline in the abundance of multiple histone marks associated with active transcription elongation, including Histone H3 lysine 36 di- and tri-methylation (H3K36me2, H3K36me3), and Histone H3 Lysine 79 di-methylation (H3K79me2). Surprisingly, this was not accompanied by an increase in the abundance of repressive heterochromatin marks (H3K27me3, H3K9me3, and H4K20me3) or a global decline in histone acetylation. Histone H4 lysine 16 acetylation (H4K16Ac), associated with RNA polymerase II pause release (Kapoor-Vazirani MCB 2011) significantly declined, but multiple acetylation marks including H3K36Ac and H3K23Ac increased in abundance. As expected, the abundance histone H4 lysine 20 mono-methylation (H4K20me1), which is implicated both in erythroblast chromatin condensation (Malik Cell Reports 2017) and the regulation of RNA Polymerase II pausing (Kapoor-Vazirani MCB 2011) also significantly increased. Consistent with these data, integration of RNA-seq and ChIP-seq data identified 3,058 genes whose expression decreased from basophilic erythroblast to orthochromatic erythroblasts, which lost enrichment for H3K36me3 (mark of active elongation) without accumulating H3K27me3 (heterochromatin mark). Based on these data, we hypothesized that RNA polymerase II pausing is a critical regulator of gene expression in maturing erythroblasts. RNA Polymerase II (Pol II) pausing is a highly regulated mechanism of transcriptional regulation, whereby transcription is initiated, but pauses 30-60bp downstream of the transcription start site. For paused Pol II to be released into active elongation, pTEFb must hyper-phosphorylate Serine 2 of the Pol II c-terminal domain (CTD). Importantly, pTEFb can be directed to specific loci through interaction with transcription factors, including GATA1 (Elagib Blood 2008; Bottardi NAR 2011). Hexim1 is a key regulator of Pol II pausing that sequesters pTEFb and inhibits its action. Consistent with a central role for Pol II pausing dynamics in the regulation of terminal erythroid maturation, Hexim1 is highly expressed in erythroid cells compared to most other cell types and its expression increases during terminal erythroid maturation. Conversely, the expression of CCNT1 and CKD9, the components of pTEFb, decline during terminal maturation, and the level of elongation competent (Ser2 and Ser2/Ser5 CTD phosphorylated) Pol II also decreases dramatically. To gain insights into the function of Pol II pausing in maturing erythroblasts, we induced Hexim1 expression in HUDEP2 cells (Kurita PLoS One 2013) using hexamethane bisacetamide (HMBA). HMBA treatment increased Hexim1 levels a dose dependent manner and was associated with gene expression and phenotypic changes suggestive of accelerated erythroid maturation. Together, these data suggest that RNA Pol II pausing dynamics are an important regulator of terminal erythroid maturation. Disclosures No relevant conflicts of interest to declare.

Description: Setd8 is the sole methyltransferase capable of mono-methylating histone H4, lysine 20. Setd8 mRNA is expressed ~10-fold higher in erythroid cells than any other cell type (biogps.org) and Setd8 protein levels increase in concert with GATA1 levels during erythroid differentiation of CD34+ HSPCs, suggesting Setd8 may have a role regulating the erythroid transcriptome. Consistent with this hypothesis, erythroid deletion of Setd8 is embryonic lethal by embryonic day 11.5 (E11.5) due to profound anemia and global transcriptomic analyses of sorted populations of E10.5 Sed8 null and control erythroblasts demonstrated a profound defect in transcriptional repression, with 340/345 differentially expressed genes (DEG) expressed at higher levels in the Setd8 null cells than controls (Malik Cell Reports 2017). Primitive erythroblasts mature and enucleate in a semi-synchronous manner in circulation. To better understand the function of Setd8 in regulating the erythroid transcriptome, we extended our transcriptomic analyses by performing RNA-seq in sorted E9.5 Sed8 null (EpoRCre+; Setd8 Δ/Δ) and control (EpoRCre+; Setd8 Δ/+) erythroblasts. The Setd8 null cells failed to repress 20/137 (15%) of the genes that are down regulated in control cells from E9.5 to E10.5. Although relatively few genes were impacted, those genes were enriched for the pathway "Oxidative Stress" (adjusted p-value 0.009) suggesting that Setd8 may regulate specific functions during terminal erythroid maturation. We next compared the DEG in Setd8 null erythroblasts to transcriptomic changes that occur as a cell transcends the hematopoietic hierarchy, gaining lineage specificity while suppressing the multi-lineage transcriptome (GSE14833). A large fraction, 105/345 (~30%), of genes up-regulated in Setd8 null erythroblasts, are also up-regulated in multipotent progenitors compared to proerythroblasts. In contrast, only 16/345 (5%) were also up-regulated in granulocyte-monocyte progenitors suggesting that Setd8 does not repress other lineage restricted signatures. Together, these results suggest that Setd8 regulates repression of the multi-lineage transcriptome during erythroid differentiation from multipotent progenitor cells. To gain insights into how Setd8 regulates the erythroid transcriptome, we performed ATAC-seq (Buenrostro Nature Methods 2013) on sorted populations of erythroblasts from E10.5 Sed8 null and control embryos. Cell number for the Setd8 null samples was limited due to anemia, with ~1000 cells used for each replicate. Setd8 and control replicates were aggregated and accessible regions were identified using MACS2. Regions more accessible in Setd8 null cells were identified by computing a log2 ratio between Setd8 null and control samples using deepTools bamCompare. In addition, we utilized ChIPmentation (Schmidl Nature Methods 2015) to assay H3K27me3 occupancy across the genome of WT E10.5 erythroblasts to identify regions of heterochromatin in maturing erythroblasts. Two replicates were performed using 2.5-5x105 cells per assay, and peak called was done using MACS2. A total of 157 genes were identified that had more accessible chromatin in Setd8 null cells and contained an enrichment for H3K27me3 in WT cells suggesting that these genes should be repressed during normal erythropoiesis. Among these were several DEG that were up-regulated in the Setd8 null cells including Hhex, Cd63, and Gata2. Genomic data integration also identified several additional transcriptional regulators that are active in earlier hematopoietic progenitors but typically silenced during erythroid differentiation including Notch1 and Cebpa. Pathway analysis of the 157 genes identified several stemness-related pathways including "Transcriptional regulation of pluripotent stem cells" and "OCT4, SOX2, NANOG repress genes related to differentiation" (adjusted p-values 0.005 and 0.008, respectively). The chromatin regions that were more accessible in the Setd8 null cells were enriched for the DNA binding motifs of the transcription factor ERG (p-value 1-257), SCL (p-value 1e-193), and NRF1 (p-value 1e-101). Taken together, these data suggest that Setd8 works in concert with erythroid transcription factors to repress the transcriptional network in stem and progenitor cells and establish appropriate patterns of gene expression during erythroid differentiation. Disclosures No relevant conflicts of interest to declare.

Description: Setd8 is the sole histone methyltransferase capable of mono-methylating histone H4, lysine 20. Setd8 is expressed at basal levels in most cell types and is important for many basic cellular functions, including cell cycle progression, transcriptional regulation, and mitotic chromatin condensation. Setd8 is expressed ~10-fold higher in erythroblasts than any other cell type and during erythroid maturation of human CD34+ HSPC, Setd8 protein levels increase in parallel with Gata1 levels, suggesting that Setd8 may have an erythroid-specific function(s). Consistent with this hypothesis, erythroid-specific deletion of Setd8 was embryonic lethal, resulting in profound anemia. Setd8-null erythroblasts had cell cycle abnormalities, failure of transcriptional repression, and defective terminal erythroid maturation. (Malik et al., Cell Reports, 2017). These studies provided important insights into the function of Setd8 in erythroid cells, but were not able to clearly delineate the "housekeeping" functions of Setd8 from its specific functions in erythropoiesis. To identify the erythroid-specific functions of Setd8, we sought to identify and disrupt the enhancer that drives high level Setd8 expression in erythroid cells. Using publically available ChIP-seq data sets, we identified a putative enhancer located in intron 1 of the SETD8 gene that was occupied by Gata1, Tal1, and H3K4me1 in human erythroblasts derived from culture of CD43+ HSPCs. This putative enhancer was able to drive luciferase expression in a reporter gene assay, and deletion of the Gata1:Tal1 site at the center of this region was sufficient to abrogate reporter gene activity. Based on these data, we hypothesized that this was the enhancer that drives high level expression of Setd8 in erythroid cells. To test this hypothesis, we used CRISPR/Cas9 genome editing to delete this region in HUDEP-2 cells. Briefly, Cas9 and guide RNA ribonucleoprotein complexes targeting the enhancer were delivered into the cells using electroporation (Gundry et al., Cell Reports, 2015). PCR and sequencing were used to confirm genome editing in monoclonal cell lines. Homozygous deletion of the enhancer (Δ/Δ) reduced SETD8 expression to 27.8% of WT (+/+) controls by RT-qPCR (n=3 for each genotype; p=0.0018). Decreased Setd8 protein levels and H4K20 mono-methylation was confirmed by Western blot. Further supporting an important function of Setd8 in erythropoiesis, deletion of the enhancer and exon 7 in CD34+ HSPCs resulted in a decreased efficiency of erythroid colony formation to 49.6% of control (n=5, p=0.0359). To gain insights into Setd8 gene regulation in erythroid cells, we performed RNA-seq, comparing the Δ/Δ and +/+ enhancer lines. In total, there were 603 genes differentially expressed (p1.5), including SETD8, FAS, and CDKN1A (p21Cip1). Pathway analyses identified numerous genes associated with apoptosis and cell death to be up-regulated. Intriguingly, multiple genes in important for stress erythropoiesis were differentially expressed in the Setd8 Δ/Δ and +/+ enhancer lines and were also differentially expressed in Setd8-null murine erythroblasts (Malik et al., Cell Reports, 2017). Most notably, both the Δ/Δ enhancer lines and the Setd8-null erythroblasts had significantly higher levels of Fas death receptor transcript than control cells. Down-regulation of Fas is essential for stress erythropoiesis (Liu et al., Blood, 2006). We therefore hypothesized that Setd8 is important for the stress erythropoiesis response. To test this hypothesis, we subjected EpoR-Cre+/-;Setd8fl/+ (Setd8Δ/+) and EpoRCre+/-;Setd8+/+ (Setd8+/+) mice to anemic stress by retro-orbital bleeding. Setd8Δ/+ and Setd8+/+ mice had similar hematocrit after anemic stress (26.6 vs 29.4%; p=0.216), but the Setd8Δ/+ had an impaired ability to mount a stress response, with a lower MCV (43.0 vs 45.1 fL, p=0.003) and reticulocyte count (8.05 vs 2.14%, p=0.031) Consistent with the transcriptomic data, Setd8Δ/+ mice had higher levels of Fas transcript in splenic erythroblasts than Setd8+/+ controls. Together, these data suggest that high level Setd8 expression is important for normal erythroid maturation and gene expression, and for regulating the stress erythropoiesis response. Disclosures No relevant conflicts of interest to declare.