Selected Histone Deacetylase Inhibitors Reverse the Frataxin Transcriptional Defect in a Novel Friedreich's Ataxia Induced Pluripotent Stem Cell-Derived Neuronal Reporter System

Abstract

Friedreich's ataxia (FRDA) is a neurodegenerative disorder caused by the expansion of guanine-adenine-adenine repeats within the first intron of the frataxin (FXN) gene. The location and nature of the expansion have been proven to contribute to transcriptional repression of FXN by decreasing the rate of polymerase II (RNA polymerase II) progression and increasing the presence of histone modifications associated with a heterochromatin-like state. Targeting impaired FXN transcription appears as a feasible option for therapeutic intervention, while no cure currently exists. We created a novel reporter cell line containing an FXN-Nanoluciferase (FXN-NLuc) fusion in induced pluripotent stem cells (iPSCs) reprogrammed from the fibroblasts of patients with FRDA, thus allowing quantification of endogenous FXN expression. The use of iPSCs provides the opportunity to differentiate these cells into disease-relevant neural progenitor cells (NPCs). NPCs derived from the FXN-NLuc line responded to treatments with a known FXN inducer, RG109. Results were validated by quantitative PCR and Western blot in multiple FRDA NPC lines. We then screened a commercially available library of compounds consisting of molecules targeting various enzymes and pathways critical for silencing or activation of gene expression. Only selected histone deacetylase inhibitors were capable of partial reactivation of FXN expression. This endogenous, FRDA iPSC-derived reporter can be utilized for high-throughput campaigns performed in cells most relevant to disease pathology in search of FXN transcription activators.

Keywords: Friedreich’s ataxia (FRDA); Nanoluciferase; induced pluripotent stem cells; neural progenitor cells (NPCs); reporter cell line; screening.

FIGURE 1

Generation and characterization of the FXN-NLuc reporter. (A) Schematic overview of FXN-NLuc reporter cell line generation. Primary fibroblasts of patients with FRDA were derived from a forearm skin biopsy (I) and reprogrammed to iPSCs (II). An in-frame fusion between the FXN gene and NLuc gene was generated via CRISPR/Cas9 genome editing (III). FXN-NLuc iPSCs were differentiated into NPCs (IV), generating a ready-to-use reporter cell line for compound screening (V). Image was created with BioRender.com. (B) Schematic presentation of the CRISPR/Cas9-mediated editing strategy used to generate the FXN-NLuc fusion. Approximate locations of the gRNA as well as primers used for confirmation of the editing are presented. (C) Validation of 5′ and 3′ junctions by PCR in two knock-in iPSC clones (C1 and C2). The 5′ junction was amplified using 5′jF and 5′jR (expected product size 1,068 bp); the 3′junction was amplified using 3′jF and 3′jR (expected product of 1,100 bp). M, size marker. (D) Confirmation of heterozygous knock-in of the NLuc gene in FRDA iPSCs. PCR was performed using primers WT_F and WT_R. C1 represents an edited clone [both unedited (WT) allele of 388 bp and edited WT + NLuc allele of 913 bp are amplified]; FRDA Par represents the unedited FRDA parental line. M, size marker. (E) Agarose gel electrophoresis of PCR products confirming the presence of expanded GAA repeats in edited FRDA iPSCs (clone C1 is shown). Primers amplify the GAA repeat tract and additional flanking sequences of 498 bp; M, size marker. (F) Western blot analysis of FXN and FXN-NLuc fusion expression in parental, unedited FRDA iPSCs (FRDA Par) and edited cells (C1, FXN-NLuc), FXN-M + NLuc (∼32 kDa)–fusion protein. GAPDH serves as a loading control. (G) In-gel detection of the luminescent FXN-NLuc fusion protein. Increasing amounts of cell lysates prepared from parental (FRDA Par) and edited (FXN-NLuc) iPSCs were separated by SDS-PAGE and the luminescent signal was detected using the Nano-Glo^® In-Gel Detection System. (H) Differentiation of FXN-NLuc iPSCs into Nestin and Sox2 positive NPCs. Scale bar, 20 mm.

FIGURE 2

Validation of FXN-NLuc reporter NPCs with known FXN inducer RG109. (A) Increase of FXN-NLuc expression (luminescence) by HDACi RG109 treatment of NPCs for 24 h at 1, 5, and 10 μM. Results are shown relative to DMSO control. (B) qRT-PCR determination of FXN mRNA and FXN-NLuc mRNA expression upon 24-h treatment of NPCs with 10 μM HDACi RG109. Primers used for amplification are located in exons 3 and 4 of FXN. DMSO treatment served as a vehicle control. (C) Representative results of Western blot analyses of FXN-NLuc NPCs treated with 10 μM RG109 for 24 h. DMSO was used as a control. Endogenous mature FXN (FXN-M, ∼13 kDa) and FXN-NLuc fusion (FXN-M + NLuc, ∼32 kDa) are indicated. GAPDH served as a loading control. (D) Western blot quantitation. FXN-M and FXN-M + NLuc were quantified and plotted independently. Statistical calculations were performed using one-way ANOVA (non-parametric Kruskal–Wallis test) for data in (A) or Student’s t-test for data in (B,D); *p < 0.05. Unless otherwise indicated, results represent mean ± SD from three independent experiments.

FIGURE 3

Screening of epigenetic compound library to identify novel inducers of FXN expression. (A) An overview of the screening process. All compounds of the APExBIO DiscoveryProbe™ Epigenetics Compound Library were tested at 10 μM concentration. Luminescence signal was plotted independently per each plate. The green line corresponds to plate average signal; dashed lines indicate average signal plus 1, 2, and 3 standard deviations of the mean. All data from two rounds of screening are included as Supplementary Figure 2. (B) Secondary validation of the efficacy of 16 compounds selected by initial screens. Luminescence detection was performed after 24 and 48 h of treatment. DMSO- and RG109-treated FXN-NLuc NPCs were used as controls. Five compounds were selected for further studies: CI994, Mocetinostat, Entinostat, UF 010, and Cerdulatinib. Results are mean ± SD from three technical replicates; * indicates p < 0.05 by one-way ANOVA. (C) Dose–response analyses for the selected compounds. Luminescence analyses were performed after 48 h of treatment with compounds at 0.5, 1, 2, 5, and 10 μM. EC50 is indicated for each compound. (D) Determination of cytotoxicity in FXN-NLuc NPCs using an LDH assay. Cytotoxicity is calculated relative to the spontaneous LDH release detected in DMSO-treated cells. Cells were treated with 10 μM of each compound for 48 h. Pirarubicin served as a positive control for cytotoxicity. Results are mean ± SD from three independent experiments; *** indicates p < 0.001 by one-way ANOVA.

FIGURE 4

Selected HDACi increase FXN transcript and protein in FXN-NLuc NPCs. (A) Results of qRT-PCR analysis of FXN mRNA expression upon 48-h treatment with 10 μM CI994, Mocetinostat, Entinostat, UF 010, and Cerdulatinib. Results represent mean ± SD from three independent experiments; one-way ANOVA, **p < 0.01. DMSO treatment served as a control. (B) FXN and FXN-NLuc protein levels determined after treatment with indicated compounds for 48 h at 10 μM. Signals corresponding to mature FXN (FXN-M), intermediate FXN isoform (FXN-I), and FXN-NLuc fusion (FXN-M + NLuc) are indicated. GAPDH and Ponceau S staining served as loading controls. (C) Quantification of FXN and FXN-NLuc protein expression relative to Ponceau S staining. Cumulative signal from FXN and FXN-NLuc is shown. Results represent mean ± SD from three independent experiments; *p < 0.05 and **p < 0.01 determined by one-way ANOVA.

FIGURE 5

Validation of efficacy of identified HDACi in additional FRDA NPC lines. (A) Results of qRT-PCR analysis of FXN mRNA expression upon 48-h treatment with 10 μM CI994, Mocetinostat, Entinostat, UF 010, and Cerdulatinib. Results represent mean ± SD from three independent experiments conducted using three different NPC lines; *p < 0.05 determined by one-way ANOVA. DMSO treatment served as a control. (B) FXN protein expression determined after treatment with indicated compounds for 48 h at 10 μM. Signals corresponding to mature FXN (FXN-M) and intermediate FXN (FXN-I) are indicated. GAPDH and Ponceau S staining served as loading controls. A representative Western blot for the FRDA NPC lines tested is shown. (C) Quantification of FXN-M levels relative to Ponceau S staining. Results represent mean ± SD from three independent experiments obtained using three different FRDA NPC lines; **p < 0.01 and ***p < 0.001 determined by one-way ANOVA.

FIGURE 6

Validation of the most potent identified HDACi in terminally differentiated FRDA neurons. (A) Results of qRT-PCR analysis of FXN mRNA expression upon 48-h treatment with 10 μM CI994, Mocetinostat, and Entinostat. Results represent mean ± SD from three independent experiments conducted using three different FRDA neuronal lines; *p < 0.05 determined by independent Student’s t-test. DMSO treatment served as a control. (B) FXN protein expression determined after treatment with indicated compounds for 48 h at 10 μM concentration. The signals corresponding to mature FXN-M and FXN-I are indicated. HPRT1 and Ponceau S staining served as loading controls. A representative Western blot is shown. (C) Quantification of FXN-M levels relative to Ponceau S staining. Results represent mean ± SD from two independent experiments obtained using two different FRDA neuronal lines; *p < 0.01 determined by independent Student’s t-test.