Figure 1.
Schematic diagrams of the β-YAC construct and HTS work flow.
Panel A) γ-luc β-luc β-YAC construct used to generate transgenic mice and derivative CID-dependent murine BMCs. This β-YAC was assembled as described in Materials and Methods and was derived from the Ppo-155 β-YAC [22]. The β-YAC is indicated as a line with the β-like globin genes or β-like globin promoter-luciferase fusions shown as boxes with the names of the genes above them. More detailed information regarding the components of the two luc fusions are indicated above and below the β-YAC illustration. Boxes at the left and right ends are modified pYAC4 vector [42] sequences. The LCR 5′HSs, 3′HS1 and YAC/yeast gene components are indicated above the line. LYS2, yeast lysine synthesis gene; ARS1, autonomous replicating sequence (yeast origin of replication); CEN1, yeast centromere; TRP1, yeast tryptophan synthesis gene; PGKneo, mammalian G418-resistance cassette. Engineered restriction enzyme sites utilized YAC structural determinations are shown below the line. Panel B) High-throughput screening work flow and secondary assays. The process flow of the high-throughput screen for identification of active compounds up-regulating fetal γ-globin gene expression is shown. The assay utilized immortalized multi-potential cells derived from the bone marrow of transgenic mice stably expressing a dual luciferase construct with firefly luciferase under control of the fetal Aγ-globin promoter and Renilla luciferase under the control of the adult β-globin promoter. The screening parameters were optimized in 384-well format and the cells were characterized for their ability to respond to at least 10 known inducers of fetal globin including hydroxyurea, sodium butyrate, valproic acid, and valeric acid. The assay was used to screen 120,035 compounds from the KU compound collection; 232 of which were found to up regulate firefly luciferase. The actives were clustered into 12 structural groups and fresh compounds were repurchased from various vendors. Three cell-based secondary assays were performed using the freshly available compounds: 1) up-regulation of firefly luciferase, 2) activity of Renilla luciferase, and 3) general cytotoxicity. The active compounds were also tested for inhibition of purified luciferase in an optimized biochemical assay. Profiling of the compounds revealed that of the 232 compounds tested, at least 124 compounds selectively up-regulated firefly luciferase but did not up-regulate Renilla luciferase. The 124 compounds which selectively up-regulated firefly luciferase activity also did not inhibit purified luciferase enzyme activity and were largely non-toxic to bone marrow progenitor cells.
Figure 2.
High-throughput screening for firefly luciferase inducers.
Panel A) Distribution of Z′ scores. An average Z′ of 0.65±0.067 was obtained across all assay plates indicating suitability of the assay for high-throughput screening of 121,035 compounds. Panel B) Sodium butyrate-induced firefly luciferase expression. The treatment of cells with the positive control sodium butyrate resulted in an average increase in firefly luciferase expression by 6.3±0.77-fold across all 300 plates tested. Panel C) Scattergram of fold-induction of all 121,085 compounds tested using CID-dependent dual-luc β-YAC BMCs. 564 active compounds were identified that induced firefly luciferase greater than 3 standard deviations above the plate median.
Table 1.
Primary screen.
Table 2.
Reordered compounds.
Figure 3.
Reconfirmation and secondary assays of γ-globin inducer compounds.
Panel A) Reconfirmation of γ-luc inducibility by 232 actives from primary HTS. Four secondary assays were employed including two reconfirmation assays for firefly luciferase induction and cytotoxicity, and two specificity assays for Renilla luciferase activity and luciferase enzyme modulators. All assays were a 10-point dose-response. A summary of firefly induction is shown in this figure. 211 of the 232 compounds had firefly induction of 2-fold or higher; a 90% reconfirmation rate. Panels B-D) Performance of seven γ-firefly inducers in the initial four secondary assays – detailed 10-point dose-response data. Comparison of compound activity from dose-response data is shown for firefly and Renilla luciferase activity (Panel B), purified firefly luciferase enzyme inhibition (Panel C), and cytotoxicity (Panel D). Assays were performed as described in Materials and Methods.
Table 3.
EC50, IC50 and therapeutic index for seven cherry-picked compounds.
Figure 4.
γ-globin and β-globin mRNA levels in compound-treated CID-dependent wild-type β-YAC mouse bone marrow cells.
qRT-PCR was performed as described in Materials and Methods. Fold change in mRNA level is shown on the y-axis; sample names are shown on the x-axis. Ctrl, untreated cells; DMSO, cells treated with DMSO only; NaB, 2 mM sodium butyrate; #7, 50 µM; #42, 15 µM, #87; 100 µM; #125, 30 µM; #157, 100 µM; #208-25, 25 µM. Gray bars, γ-globin mRNA expression; black bars, β-globin mRNA expression. Data shown are the results of two-three separate experiments, with each sample duplicated within an experiment. P≤0.01 for NaB and compound # 42; P≤0.05 for compounds #42 and #208.
Figure 5.
Induction of HbF by lead compounds in CID-dependent mouse BMCs.
Panel A) FACS analysis of HbF levels in compound-treated CID-dependent wild-type β-YAC mouse bone marrow cells. The protocol was carried out as described in Materials and Methods using anti-human HbF FITC-conjugated antibody. Samples are labeled as described in the legend to Figure 4, but compound concentration follows the compound number. Representative data from one experiment is shown here, but the experiment was replicated two to three times for each sample with similar results (summarized in Panels B and C). Panels B-C) Summary of FACS analysis of HbF levels in compound-treated CID-dependent wild-type β-YAC mouse bone marrow cells. Panel B) fold change in γ-globin (FITC)+ cells. Panel C) percent viable cells. Bars show mean and standard error of the mean in control and compound-treated cells for each panel. Panel D) HbF induction measured by ELISA in compound-treated CID-dependent wild-type β-YAC mouse bone marrow cells. The assay was carried out as described in Materials and Methods. Fold induction of HbF is shown on the y-axis; cell treatment and concentration, where applicable, are shown on the x-axis. Data represent the mean and standard error of the mean from four experiments with duplicate samples within each experiment.
Figure 6.
γ-globin expression is induced in compound-treated human primary erythroid progenitors.
Human erythroid progenitors were generated from adult CD34+ cells in liquid culture as described in Materials and Method). Cells were analyzed for morphology, globin gene expression and HbF induction after treatment with the lead compounds. Panel A) Erythroid cells were harvested on the days indicated for cell morphology determination by Geimsa staining. The percentage of different erythroid progenitors at each stage is shown as a function of days in culture. At least 500 cells were counted by light microscopy from duplicate slides. Representative cell morphology is shown in the images at 40× magnification. Panel B) qRT-PCR was performed as described in Materials and Methods. The levels of γ- and β-globin expression were normalized to GAPH; expression of each gene is shown as a fraction of the total globin (γ+β). Note the γ-to β-globin switch around day 10. Panel C) Fold change in γ-globin/β-globin mRNA ratio after normalization to GAPDH is shown on the y-axis. UN, untreated cells; DMSO, cells treated with DMSO only; NaB, 2 mM sodium butyrate; lead compounds: #7, 50 µM; #42, 15 µM, #87; 100 µM; #208, 25 µM. Data shown are the mean and standard error of the mean of three independent samples.
Figure 7.
The lead compounds induce HbF expression in human primary erythroid cells.
FACS analysis was performed with erythroid progenitors treated for 48 hours with each of the four lead chemical compounds. The protocol was carried out as described in Materials and Methods using anti-human HbF FITC-conjugated antibody. Panel A) Representative FACS tracings are shown; experiments were replicated three times for each compound with similar results. Panel B) Based on the FACS data the % HbF-positive cells (FITC-A) was calculated for the different treatment conditions. Shown in the graph is the fold change in HbF-positive cells. Data are shown as the mean and standard error of the mean in control and compound-treated cells for each sample. *; P<0.01.