Introduction: Sickle cell disease (SCD) and β-thalassemia are inherited blood disorders caused by mutations in the β-globin gene (HBB). Elucidation of the multiple pathophysiologic mechanisms in SCD and β-thalassemia has resulted in an increasing efforts to identify new treatment modalities to ameliorate the consequences of the disease. However, no consistent in vitro system exists for studies of pharmacological therapies for the diseases. Human umbilical cord-derived erythroid progenitor cells (HUDEP2) are an immortalized CD34+ hematopoietic stem cell-derived erythroid precursor cell line that can differentiate into red blood cells. Here, we engineered sickle HUDEP2 and β-thalassemia HUDEP2 clonal lines through CRISPR/Cas9-mediated editing of the human HBB. We sought to establish if these engineered cell lines exhibit disease phenotypes, and if upon in vitro erythroid differentiation they produce fetal hemoglobin (HbF) in response to hydroxyurea, the only FDA-approved drug for HbF induction. Our goal is to create an in vitro system to test new HbF inducers for treating SCD or β-thalassemia.

Materials and Methods: We delivered Hi-Fidelity Streptococcus pyogenes (Sp) Cas9 protein and CRISPR guide RNA as a ribonucleoprotein complex in conjunction with a single-stranded DNA donor (ssODN) template to introduce the sickle or K17X (A<T) or codon 6 [-G] β-thalassemia mutation into the HBB locus of HUDEP2 cells. Edited HUDEP2 cells were single-cell sorted into multiple 96-well plates and expanded. The genotype of the clones was determined using a probe-based droplet digital PCR assay and confirmed through Sanger sequencing. Native polyacrylamide gel electrophoresis and high-performance liquid chromatography (HPLC) were used to confirm the hemoglobin phenotype. Normal parental cell line, sickle clone, and two individual β-thalassemia clones were used to test the pharmacological induction of HbF. We initiated drug treatment in the expansion phase with 30 µM hydroxyurea. Trypan Blue staining and CD71/CD233/CD235 staining determined the effect of the drugs on the viability, growth rate and erythroid development of HUDEP2 lines. After 10 days of drug treatment, differentiated HUDEP2 were analyzed for globin expression through RT-qPCR and HPLC, and HbF positive cells (F-cells) were quantified via flow cytometry. Cells were placed at 2% O2 for four hours, fixed in glutaraldehyde, stained, and viewed under magnification to assess sickling potential.

Results and Discussion: We generated multiple clones with biallelic sickle or β-thalassemia mutations. Sickle HUDEP2 clones almost exclusively expressed sickle hemoglobin with low level of HbF and hemoglobin A2 (HbA2), and β-thalassemia HUDEP2 clones produced no normal adult hemoglobin, 8-10% HbF, and 26-28% HbA2. On HPLC analysis, β-Thalassemia HUDEP2 clones had an unknown tall peak (39-45%) between HbF and HbA consistent with an α-globin homotetramer (α4). When subjected to hypoxic conditions for 4 hours, sickle HUDEP2 produced sickle cells. HUDEP2 parent cells did not sickle under hypoxic conditions. Hydroxyurea induced 3.8-fold, 1.8-fold, and 1.6-fold increases in γ-globin gene (HBG) expression; 2.9-fold, 1.4-fold, and 1.4-fold increases in the percentages of F-cells; 1.4-fold, 1.2-fold, and 1.6-fold increase in the percentages of HbF in sickle, K17X(A<T) and codon 6[-G] β-thalassemia HUDEP2 clones, respectively. No change was observed in CD71/CD235 positive HUDEP2 cells in the presence hydroxyurea. This finding demonstrated that hydroxyurea treatment induces HBG expression as well as HbF and F-cells in engineered sickle and β-thalassemia HUDEP2 clones. Future work will include screening other pharmacological compounds as well as studying the mechanism of HbF induction by using HUDEP2 clones.

Conclusions: Our engineered sickle and β-thalassemia HUDEP2 cell lines have properties similar to those of patient erythroid cells and respond to the known HbF inducer hydroxyurea. This in vitro model system may facilitate the drug-discovery process by enabling multimodal drug screening on a large scale with consistent and reproducible results.

Acknowledgments: This work was supported by the Cancer Prevention and Research Institute of Texas grants RR140081 and RP170721 (to G.B.) and the National Heart, Lung and Blood Institute of NIH (1K08DK110448 to V.S.)


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.