Understanding Heterogeneity of Fetal Hemoglobin Induction through Comparative Analysis of Stage-Matched F- and a-Cells

Reversing the developmental switch from fetal (HbF, α₂γ₂) to adult (HbA, α₂β₂) hemoglobin is an important therapeutic approach in sickle cell disease (SCD) and β-thalassemia. Elevated HbF levels due to genetic variation or through therapeutic induction by hydroxyurea (HU) attenuate the severity of both disorders. HbF in healthy individuals, SCD patients, and patients treated with HU is present in a heterocellular fashion in a subset of red blood cells known as F-cells. Despite over 50 years of observations of F-cells, it is not known why only some cells in a genetically identical population are able to express HbF or respond to pharmacological inducers. Adult F-cells can potentially represent a reversion to a fetal-like epigenetic and transcriptional program, or alternatively isolated transcriptional or posttranscriptional events at the γ-globin genes. Here we set out to understand the heterogeneity of HbF activation and gain insights into whether the mechanisms underlying the heterocellular response are similar or distinct in response to different HbF inducers.

To this end we developed techniques to purify differentiation stage-matched late erythroblast F-cells and non-F cells (A-cells) from the human HUDEP2 erythroid cell line and primary CD34 cell erythroid cultures using a reversible fixation protocol enabling extraction of high-quality RNA and protein. Purified F-cells from both sources were enriched for γ-globin transcripts by 200-500 fold by RT-PCR, validating the purification scheme. We profiled these cells by RNA-seq using a modified method that depletes globin mRNAs and ribosomal RNAs and is capable of detecting low abundance transcripts, as well as by mass spectrometry using size fractionation to increase the number of detected proteins.

In differentiated clonal HUDEP2 cells, differences between F-cells and A-cells were remarkably small, with only 62 differentially expressed transcripts and 20 differentially expressed proteins. Top differentially expressed transcripts were γ-globin and the non-coding β-globin locus transcripts BGLT3 and HBBP1. Interestingly, there were no significant changes in known HbF regulators BCL11A, LRF, and HRI at the RNA or protein level. Gene set enrichment analysis (GSEA) using a previously generated set of differentially expressed transcripts from adult and fetal-derived CD34 erythroid cultures showed enrichment of fetal transcripts in F-cells and adult transcripts in A-cells. We also carried out transcriptome analysis of sorted matched late erythroblast F-cells and A-cells from human CD34+ cell erythroid cultures at different time points. Similar to HUDEP2 cells, only small numbers of transcripts were differentially expressed (33 at 8 days, 17 at 11 days, and 261 at 14 days). BCL11A, LRF, and HRI were not differentially expressed at the earlier timepoints, and BCL11A and HRI were at most decreased by about 20% at the 14-day mark. GSEA analysis did not show fetal transcript enrichment in day 8. At days 11 and 14, there was some enrichment of fetal transcripts in F-cells but not to the degree of HUDEP2 cells. Finally, we analyzed sorted F- and A-cells from day 11 CD34+ erythroid cultures treated with hydroxyurea and pomalidomide. Again, differences between F- and A-cells were small with hydroxyurea treatment (53 transcripts) and more significant with pomalidomide treatment (400 transcripts).

We have successfully established an approach to analyze stage-matched γ-globin containing cells from a genetically identical starting population, with high degree of enrichment. Our preliminary data indicate that these cells are overall highly similar to non-γ-containing cells, but do show some enrichment of fetal-specific transcripts, more so in HUDEP2 cells. The differences between F- and A- cells are overall smaller than those observed by us and others in profiling of fetal and adult-derived erythroblasts. This suggests that F-cells are not formed by reversion to a fetal-like state but rather through specific changes at the β-globin locus. Importantly, we do not find differential levels of any known γ-globin regulators, suggesting an alternative mechanism for the heterocellular expression pattern. Studies are currently ongoing to carry out epigenetic profiling of F-cells.