Inhibition of Rho-Associated Kinases ROCK1 and ROCK2 as a Therapeutic Strategy to Reactivate the Repressed FXN Gene in Friedreich Ataxia

Abstract

Friedreich ataxia (FA) is an autosomal recessive disease characterized by progressive damage to the nervous system and severe cardiac abnormalities. The disease is caused by a GAA•TTC triplet repeat expansion in the first intron of the FXN gene, resulting in epigenetic repression of FXN transcription and reduction in FXN (frataxin) protein which results in mitochondrial dysfunction. Factors and pathways that promote FXN repression represent potential therapeutic targets whose inhibition would restore FXN transcription and frataxin protein levels. Here, we performed a candidate-based RNAi screen to identify kinases, a highly druggable class of proteins, that when knocked down upregulate FXN expression. Using this approach, we identified Rho kinase ROCK1 as a critical factor required for FXN repression. ShRNA-mediated knockdown of ROCK1, or the related kinase ROCK2, increases FXN mRNA and frataxin protein levels in FA patient-derived induced pluripotent stem cells (iPSCs) and differentiated neurons and cardiomyocytes to levels observed in normal cells. We demonstrate that small molecule ROCK inhibitors, including the FDA-approved drug belumosudil and fasudil, reactivate FXN expression in cultured FA iPSCs, neurons, cardiomyocytes, and FA patient primary fibroblasts and ameliorate the characteristic mitochondrial defects in these cell types. Remarkably, treatment of transgenic FA mice of both sexes with belumosudil or fasudil upregulates FXN expression, ameliorates the mitochondrial defects in the brain and heart tissues, and improves motor coordination and muscle strength. Collectively, our study identifies ROCK kinases as critical repressors of FXN expression and provides preclinical evidence that FDA-approved ROCK inhibitors may be repurposed for treatment of FA.

Keywords: Friedreich ataxia; ROCK kinases; frataxin; mitochondrial function; neurodegeneration; reactivation of FXN.

Figure 1.

A candidate-based shRNA screen identifies protein kinases as potential drug targets for FXN reactivation. A, Schematic of the kinome-wide shRNA screening strategy (created with BioRender.com). TRC, The RNAi Consortium; NGS, next-generation sequencing. B, qRT-PCR analysis monitoring relative FXN mRNA levels in FA iPSCs expressing a nonsilencing (NS) shRNA or an shRNA targeting one of eight kinase candidates. The results were normalized to that obtained in normal iPSCs, which was set to 1. C, Immunoblot analysis showing frataxin protein levels in FA iPSCs expressing a NS shRNA or an shRNA targeting a kinase candidate. α-tubulin (TUBA) was monitored as a loading control. Data are represented as mean ± SD (n = 3 biological replicates). *p < 0.05, **p < 0.01.

Figure 2.

FXN-RFs mediate repression of FXN in FA neurons and cardiomyocytes. A, Top, Micrograph showing TUJ1-positive neurons (green) derived from FA iPSCs. TUJ1 is a pan-neuron marker. Nuclear DAPI staining is shown in blue. Bottom, Quantification of the percentage of TUJ1-positive cells in FA iPSCs and FA neurons. B, Quantification of the percentage of FA iPSCs or FA neurons that are positive for phosphorylated histone H3 (pH3), a marker for actively diving cells. C, D, qRT-PCR analysis monitoring relative FXN mRNA levels (C) and immunoblot analysis showing frataxin protein levels (D) in FA neurons expressing a NS or FXN-RF shRNA. E, Top, Micrographs showing cardiac troponin T (cTNT)- or sarcomeric actinin (SA)-positive cardiomyocytes (CMCs) derived from FA iPSCs. cTNT staining is shown in red, SA staining in green, and DAPI staining in blue. Bottom, Quantification of the percentage of cells staining positive for cTNT or SA in FA iPSCs or FA cardiomyocytes. F, G, qRT-PCR analysis monitoring relative FXN mRNA levels (F) and immunoblot analysis showing frataxin protein levels (G) in FA cardiomyocytes (CMCs) expressing a NS or FXN-RF shRNA. Data are represented as mean ± SD (n = 3 biological replicates for C and F). *p < 0.05, **p < 0.01.

Figure 3.

Confirmation that both ROCK1 and ROCK2 are FXN-RFs. A, qRT-PCR analysis monitoring relative ROCK1 and ROCK2 mRNA levels in FA iPSCs expressing a NS shRNA or one of several shRNAs targeting ROCK1 or ROCK2. The results were normalized to that obtained with a NS shRNA, which was set to 1. The ROCK1-1 and ROCK1-2 shRNAs correspond to those used in Figure 1B,C and Extended Data Figure S1. The results show that the ROCK1-3 and ROCK2-1 shRNAs were selective for knockdown of ROCK1 and ROCK2, respectively. B–D, qRT-PCR analysis monitoring relative FXN mRNA levels in FA iPSCs (B), neurons (C), and cardiomyocytes (D) expressing a NS, ROCK1-3, or ROCK2-1 shRNA. The results were normalized to that obtained in normal cells, which was set to 1. Data are represented as mean ± SD (n = 3 biological replicates). *p < 0.05, **p < 0.01.

Figure 4.

ROCK inhibitors reactivate FXN and ameliorate the mitochondrial defects in FA neurons and cardiomyocytes. A–C, qRT-PCR analysis monitoring relative FXN mRNA levels (left) and immunoblot analysis showing frataxin levels (right) in FA iPSCs (A), neurons (B), and cardiomyocytes (C) treated with DMSO (control), fasudil, belumosudil or ripasudil. D, Activation curves showing FXN mRNA levels following treatment of FA neurons and cardiomyocytes with increasing concentrations of fasudil, belumosudil or ripasudil. E, Mitochondrial ROS levels, as assessed by MitoSOX Red staining followed by flow cytometry, in FA iPSCs, neurons, and cardiomyocytes treated with DMSO, fasudil, belumosudil, or ripasudil. F, Oxygen consumption rate of FA iPSCs, neurons, and cardiomyocytes treated with DMSO, fasudil, belumosudil, or ripasudil. Data are represented as mean ± SD (n = 3 biological replicates). *p < 0.05, **p < 0.01.

Figure 5.

ROCK inhibitors reactivate FXN and ameliorate mitochondrial defects in cultured FA patient primary fibroblasts. A, qRT-PCR analysis monitoring relative FXN mRNA levels in fibroblast cell lines derived from six FA patients. The results were normalized to that obtained with DMSO-treated fibroblasts from an unaffected individual, which was set to 1. B, Relative aconitase activity levels in unaffected or FA patient fibroblasts treated with DMSO, fasudil, or belumosudil. Aconitase activity was normalized to endogenous citrate synthase activity, which was set to 1 (data not shown). Data are represented as mean ± SD. *p < 0.05, **p < 0.01.

Figure 6.

ROCK inhibitors reactivate FXN, ameliorate mitochondrial defects, and rescue disease phenotypes in a mouse model of FA. A, qRT-PCR analysis monitoring relative FXN mRNA levels in the brain (left) or heart (right) tissue of YG8sR mice (n = 4 per group) treated with vehicle (PBS) or fasudil or belumosudil (50, 100, or 200 mg/kg/d) for 2 weeks. The results were normalized to that obtained with vehicle, which was set to 1. FXN expression in vehicle-treated Y47R mice is shown as a normal control. B, Immunoblot analysis showing frataxin levels in brain (left) or heart (right) tissue in YG8sR mice treated with vehicle or fasudil or belumosudil at 100 mg/kg/d for 2 weeks. C, Relative aconitase activity levels in the brain (left) or heart (right) tissue of YG8sR mice (n = 4 per group) treated with vehicle or fasudil or belumosudil (50, 100 or 200 mg/kg/d) for 2 weeks or, as a normal control, vehicle-treated Y47R mice. Aconitase activity was normalized to endogenous citrate synthase activity, which was set to 1 (data not shown). D, Inverted screen test showing grip time for YG8sR mice treated with vehicle (n = 7) or 100 mg/kg/d fasudil (n = 7) or belumosudil (n = 5), for 2, 3, or 4 weeks or, as a normal control, vehicle-treated Y47R mice. E, Rotarod test showing latency time for the mice described in D.