Neither DNA hypomethylation nor changes in the kinetics of erythroid differentiation explain 5-azacytidine's ability to induce human fetal hemoglobin

Abstract

5-azacytidine (5-Aza) is a potent inducer of fetal hemoglobin (HbF) in people with beta-thalassemia and sickle cell disease. Two models have been proposed to explain this activity. The first is based on the drug's ability to inhibit global DNA methylation, including the fetal globin genes, resulting in their activation. The second is based on 5-Aza's cytotoxicity and observations that HbF production is enhanced during marrow recovery. We tested these models using human primary cells in an in vitro erythroid differentiation system. We found that doses of 5-Aza that produce near maximal induction of gamma-globin mRNA and HbF do not alter cell growth, differentiation kinetics, or cell cycle, but do cause a localized demethylation of the gamma promoter. However, when we reduced gamma promoter methylation to levels equivalent to those seen with 5-Aza or to the lower levels seen in primary fetal erythroid cells using DNMT1 siRNA and shRNA, we observed no induction of gamma-globin mRNA or HbF. These results suggest that 5-Aza induction of HbF is not the result of global DNA demethylation or of changes in differentiation kinetics, but involves an alternative, previously unrecognized mechanism. Other results suggest that posttranscriptional regulation plays an important role in the 5-Aza response.

Figure 1

5-azacytidine increases γ-globin and decreases β-globin mRNA levels. (A) The in vitro differentiation system used in this research. (B) γ- and β-globin steady-state mRNA levels during in vitro differentiation. The first panel shows RNA levels on the same scale. The second panel shows mRNA levels normalized to the maximal expression for each mRNA to emphasize the different time courses of expression. (C) Effects of daily 300 nM 5-Aza on steady-state mRNA levels for the α-, β-, and γ-globin and the GATA-1 genes. CD34⁺ cells were from different normal donors were used for experiments in panels B and C. Each RT-PCR analysis was performed in triplicate. Error bars represent 1 SD. Microscopy was performed using an Olympus model CHT light microscope (Olympus Optical, Tokyo, Japan). Magnification of Figure 1A was 100× (10×/10×; numeric aperture of the 10× objective was 0.25). Images were acquired with a FujiFilm FinePix A345 digital camera (Fujifilm USA, Valhalla, NY). Images were processed using iPhoto version 6.0.6 (Apple Computer, Cupertino, CA).

Figure 2

Dose-response effects of 5-azacytidine during in vitro erythroid differentiation. Effects of different doses of 5-azacytidine on (A) γ-globin gene expression during differentiation, (B) β-globin gene expression during differentiation, (C) total γ- and β-globin gene expression, expressed as the area under the curve (AUC) for each dose, (D) γ/γ plus βAUC for each dose, (E) cell expansion, (F) differentiation kinetics as assessed by CD235a (glycophorin A) cell surface expression, (G) the proportion of cells in the G1 phase of the cell cycle, and (H) the proportion of cells in the S phase of the cell cycle. CD34⁺ cells were from different normal donors were used for experiments in panels A-F and G,H. Each RT-PCR analysis was performed in triplicate. Error bars represent 1 SD.

Figure 3

Dose-response effects of 5-azacytidine on hemoglobin production. (A) Ion exchange HPLC tracings of hemolysates from cells at the completion of the in vitro differentiation procedure. Fetal (F) and adult (A) hemoglobin peaks are indicated. (B) Summary of Hb HPLC data showing a reciprocal effect of 5-azacytidine on fetal and adult hemoglobin levels.

Figure 4

5-Azacytidine decreases methylation of the γ-globin promoter. (A) The sodium bisulfite modification method was used to determine the methylation status of the 6 CpGs between −162 and +50 of the γ-globin promoter during in vitro differentiation. Each bar represents the proportion of methylated CpGs at each site for different time points during differentiation. (B) Methylation of individual γ-globin CpGs on day +11 of differentiation with () and without (□) 300 nM 5-azacytidine treatment. For comparison, the methylation of the CpGs at day +5 without 5-azacytidine is shown (▨). (C) Methylation of individual γ-globin CpGs on day +14. (D) Methylation of individual γ-globin CpGs on day +18. (D) Comparison of total γ-globin promoter methylation during differentiation with and without 300 nM 5-azacytidine. (E) Comparison of total γ-globin promoter methylation during differentiation with and without 300 nM 5-azacytidine. (F) As a measure of genome-wide DNA methylation, the effect of 300 nM 5-azacytidine on LINE methylation was determined. CD34⁺ cells from a single normal donor were used for this experiment. Error bars represent 1 SD. P values were determined by t test.

Figure 5

siRNA knock-down of DNMT1 produces decreased γ-globin promoter methylation but no increase in γ-globin mRNA or fetal hemoglobin. siRNA for the DNMT1 gene was transiently transfected into cells on day +11 of in vitro erythroid differentiation. Controls for this experiment were cells from the same initial culture that were treated with a nonspecific siRNA (siCTRL) or underwent mock transfection without siRNA (CTRL). (A) The effect of siRNA treatment on DNMT1 mRNA levels during differentiation as determined by quantitative RT-PCR. (B) The effect of siRNA treatment on DNMT3a mRNA levels. (C) The effect of siRNA treatment on DNMT1 protein levels as determined by Western blotting. β-actin and DNMT3a serve as nonspecific controls. (D) The effect of siRNA treatment on global DNA methylation as determined by bisulfite conversion analysis of LINE elements. Note that for the day 15 control sample no error bars are shown. This is because all 5 sequences, which included a total of 42 individual CpGs, were 100% methylated so the standard deviation for this data point was 0. (E) The effect of siRNA treatment on γ-globin promoter DNA methylation. (F) The effect of siRNA treatment on γ-globin mRNA during differentiation. (G) The effect of siRNA treatment on hemoglobin production as assessed by HPLC analysis of cell lysates at the end of in vitro differentiation. CD34⁺ cells from a single normal donor were used for this experiment. Error bars represent 1 SD. P values were determined by t test.

Figure 6

shRNA knock-down of DNMT1 produces decreased γ-globin promoter methylation but no increase in γ-globin mRNA or fetal hemoglobin. Lentiviral vector LL3.7 containing DNMT1 shRNA and GFP sequences was used to transduce differentiating cells on day +5. On day +7 cells expressing GFP were sorted and allowed to differentiate into erythroid cells. Controls for this experiment were cells from the same initial culture transduced with vectors carrying a mismatched version of the DNMT1 shRNA, an empty vector, or cells that were untreated. Effects of DNMT1 shRNA on (A) DNMT1 steady-state mRNA levels during in vitro differentiation, (B) DNMT1 protein levels as determined by Western blot, (C) γ-globin promoter DNA methylation, (D) LINE element DNA methylation, (E) γ-globin steady-state mRNA, and (F) Hb levels as determined by HPLC at the end of differentiation. CD34⁺ cells from a single healthy donor were used for this experiment. Error bars represent 1 SD. P values were determined by t test.

Figure 7

5-Azacytidine increases nascent mRNA production from the γ, β, and α-globin genes. Quantitative RT-PCR was performed using primers that crossed intron-exon boundaries or were within an intron to determine levels of nascent or pre-mRNA as an indicator of gene activity. (A) Effect of 5-azacytidine on nascent γ-globin gene expression. (B) Effect of 5-azacytidine on nascent β-globin gene expression. (C) Effect of 5-azacytidine on nascent α-globin gene expression. (D) Effect of 5-azacytidine on nascent GAPDH gene expression. CD34⁺ cells from a single normal donor were used for this experiment. Each RT-PCR analysis was performed in triplicate.