A Novel High-Content Immunofluorescence Assay as a Tool to Identify at the Single Cell Level γ-Globin Inducing Compounds

Abstract

The identification of drugs capable of reactivating γ-globin to ameliorate β-thalassemia and Sickle Cell anemia is still a challenge, as available γ-globin inducers still have limited clinical indications. High-throughput screenings (HTS) aimed to identify new potentially therapeutic drugs require suitable first-step-screening methods combining the possibility to detect variation in the γ/β globin ratio with the robustness of a cell line. We took advantage of a K562 cell line variant expressing β-globin (β-K562) to set up a new multiplexed high-content immunofluorescence assay for the quantification of γ- and β-globin content at single-cell level. The assay was validated by using the known globin inducers hemin, hydroxyurea and butyric acid and further tested in a pilot screening that confirmed HDACs as targets for γ-globin induction (as proved by siRNA-mediated HDAC3 knockdown and by treatment with HDACs inhibitors entinostat and dacinostat) and identified Heme-oxygenases as novel candidate targets for γ-globin induction. Indeed, Heme-oxygenase2 siRNA knockdown as well as its inhibition by Tin protoporphyrin-IX (TinPPIX) greatly increased γ-globin expression. This result is particularly interesting as several metalloporphyrins have already been developed for clinical uses and could be tested (alone or in combination with other drugs) to improve pharmacological γ-globin reactivation for the treatment of β-hemoglobinopathies.

Fig 1. Analysis of γ/β globin levels by immunofluorescence and automated image capture.

A) Image acquisition and analysis for β-K562 and K562. Merged signals of DNA (Hoechst-33342), β-globin and γ-globin are read in channel 1 (Ch1), channel 2 (Ch2) and channel 3 (Ch3), respectively (see also S1D Fig). Bar = 50μm. The intensity value of signals is automatically assigned by the instrument and converted into a corresponding intensity of color. The relative scatter plots show the distribution of double γ^-β^- negative, single γ⁺β^- positive, single γ^-β⁺ positive and double γ⁺β⁺ positive cells (x axis: FITC-β-globin; y axis: PE-γ-globin). Numbers within plots refer to the averaged percentage of cells within each population from three independent experiments (n = 3). The relative st.errors are shown in panel C: *p<0,05; ** p<0,01; ***p<0,001. B) Quantitative fluorescence imaging of single cells: cells numbered from 1 to 6 in panel A are taken as an example of γ^-β^- double negative (1 and 2), single γ⁺β^- positive (5), single γ^-β⁺ positive (4) and γ⁺β⁺ double positive (3 and 6). C) Statistical analysis (n = 3): γ^-β^- cells; red: γ⁺β^- cells; yellow: γ⁺β⁺ cells; green: γ^-β⁺ cells. D) RTqPCR on α, ε, γ- and β-globins. Histograms show the relative levels of expression normalized on glyceraldehyde-3-phosphate dehydrogenase (GAPDH). n≥3, statistical analysis: *p<0,05; **p<0,01; ***p<0,001.

Fig 2. High-content analysis of compound-induced changes in globins accumulation.

β-K562 cells were treated with 800μM hydroxyurea and 900μM butyric acid (n = 3, a representative experiment is shown here) and the same cells were analyzed in parallel by immunofluorescence and by RTqPCR 4 days after the addition of the drugs. A) Immunofluorescence images (Bar = 50μm) and relative scatter plots. Data from three independent experiments are presented and statistically analyzed (B) as in Fig 1. C) RTqPCR on α-, γ- and β- globins. Histograms show the relative levels of expression relative to GAPDH. D) Confocal analysis of β-K562 cells untreated or treated with HU as in panel A and subjected to a quadruple staining with Hoechst (blue), anti β- (green), anti γ-globin (red) and anti-CD235a (white). Magnification: 20x. Right panel: 40x magnification of individual cells γ⁺CD235a⁺ or β⁺CD235a⁺ double positive and γ⁺β⁺CD235a⁺ triple positive, respectively.

Fig 3. High-content γ/β globin analysis as readout of siRNA screening in β-K562 confirms HDAC as targets for γ-globin activation.

A) Cells were transfected with a non-targeting oligo (siNTO) as negative control and with a siRNA directed to HDAC3. Two siRNAs were tested, with two technical replicates. C) β-K562 treated with two different HDAC inhibitors: entinostat and dacinostat (see also S3 Fig). A and C) Immunofluorescence images (Bar = 50μm) and relative scatter plots. Data from three independent experiments are presented and statistically analyzed (B and D) as in Fig 1.

Fig 4. HMOX2 siRNA-mediated knockdown and hemin or Tin-PPIX treatment have similar effects on β-K562 hemoglobinization levels.

A) Cells were transfected with a non-targeting oligo (siNTO) as negative control and with a siRNA directed to HMOX2. Two siRNAs (see also S4A Fig) were tested, with two technical replicates. C) Cells were treated with 50μM of either hemin or Tin-PPIX. A, C) Immunofluorescence images (Bar = 50μm) and relative scatter plots. Data from three independent experiments are presented and statistically analyzed (B and D) as in Fig 1. E) RTqPCR on α-, γ- and β-globins from cells treated with hemin or Tin-PPIX. Histograms show levels of globins expression relative to GAPDH (n = 3).