ReportsAmyotrophic Lateral Sclerosis

The Src/c-Abl pathway is a potential therapeutic target in amyotrophic lateral sclerosis

See allHide authors and affiliations

Science Translational Medicine  24 May 2017:
Vol. 9, Issue 391, eaaf3962
DOI: 10.1126/scitranslmed.aaf3962

A stepping stone to ALS drug discovery

ALS is a heterogeneous motor neuron disease for which there is no treatment and for which a common therapeutic target has yet to be identified. In a new study, Imamura et al. developed a drug screen using motor neurons generated from ALS patient induced pluripotent stem cells (iPSCs). They screened existing drugs and showed that inhibitors of Src/c-Abl kinases promoted autophagy and rescued ALS motor neurons from degeneration. One of the drugs was effective for promoting survival of motor neurons derived from ALS patients with different genetic mutations. The Src/c-Abl pathway may be a potential therapeutic target for developing new drugs to treat ALS.

Abstract

Amyotrophic lateral sclerosis (ALS), a fatal disease causing progressive loss of motor neurons, still has no effective treatment. We developed a phenotypic screen to repurpose existing drugs using ALS motor neuron survival as readout. Motor neurons were generated from induced pluripotent stem cells (iPSCs) derived from an ALS patient with a mutation in superoxide dismutase 1 (SOD1). Results of the screen showed that more than half of the hits targeted the Src/c-Abl signaling pathway. Src/c-Abl inhibitors increased survival of ALS iPSC-derived motor neurons in vitro. Knockdown of Src or c-Abl with small interfering RNAs (siRNAs) also rescued ALS motor neuron degeneration. One of the hits, bosutinib, boosted autophagy, reduced the amount of misfolded mutant SOD1 protein, and attenuated altered expression of mitochondrial genes. Bosutinib also increased survival in vitro of ALS iPSC-derived motor neurons from patients with sporadic ALS or other forms of familial ALS caused by mutations in TAR DNA binding protein (TDP-43) or repeat expansions in C9orf72. Furthermore, bosutinib treatment modestly extended survival of a mouse model of ALS with an SOD1 mutation, suggesting that Src/c-Abl may be a potentially useful target for developing new drugs to treat ALS.

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease that causes progressive loss of motor neurons (1, 2). The disease progression is fast, and there is no effective treatment. Most cases are classified as sporadic ALS, whereas about 10% are familial. About 25% of the familial ALS (FALS) cases are associated with mutations in the gene encoding Cu/Zn superoxide dismutase (SOD1) (3). Although mutant SOD1 transgenic mice recapitulate ALS phenotypes (4) and have been used for preclinical studies of ALS drug development, only a limited number of compounds have been tested. Thus, we developed a phenotypic screening assay for testing a number of compounds with a readout of ALS motor neuron survival. In previous studies of ALS, many kinds of causative genes have been discovered and multiple hypotheses about molecular pathogenesis have been proposed. However, it is clear that motor neuron death is an undisputed common phenotype among the heterogeneous forms of ALS (1, 2). Many drug screening platforms for ALS have been developed based on induced pluripotent stem cell (iPSC) technology (511). Here, we introduced transcription factors using the piggyBac vector system (12) to generate iPSCs from an ALS patient carrying an SOD1 mutation and then derived motor neurons from them. Using this phenotypic assay, we screened existing drugs (13) and identified Src/c-Abl inhibitors that ameliorated ALS motor neuron degeneration.

Src and c-Abl are ubiquitous nonreceptor tyrosine kinases that were identified as the mammalian homologs of the oncogene products of Rous sarcoma virus and Abelson murine leukemia virus, respectively. Activation of Src, which is associated with cell proliferation, angiogenesis, apoptosis, and invasion, has been observed in tumors and is a target for cancer therapy (14). The Bcr-Abl fusion protein, an oncogenic form of the c-Abl fusion kinase, is known to cause chronic myelogenous leukemia (CML) and Philadelphia chromosome–positive adult acute lymphoblastic leukemia, and c-Abl inhibitors have been developed as anti-CML drugs (15). Src/c-Abl is associated with various cellular functions (16, 17), and several studies have shown the involvement of Src family proteins and c-Abl in neurodegenerative diseases (1824).

RESULTS

To screen compounds, we used survival of ALS patient iPSC-derived motor neurons as readout. This required large-scale generation of motor neurons from patient iPSCs. We developed a motor neuron differentiation method by transducing three transcription factors, LIM homeobox protein 3 (Lhx3), neurogenin 2 (Ngn2), and ISL LIM homeobox 1 (Isl1), into iPSCs derived from healthy individuals and from ALS patients carrying SOD1 mutations. These transcription factors, delivered using adenoviral vectors, were reported previously to induce formation of mature spinal motor neurons from neural precursor cells (25). A polycistronic vector containing Lhx3, Ngn2, and Isl1 under control of the tetracycline operator was introduced into healthy control– and ALS patient–derived iPSCs (fig. S1, A to D, and tables S1 and S2) using the piggyBac vector, and the vector-introduced clones were established as stable iPSC clones after neomycin selection. After doxycycline treatment, motor neurons were generated from control iPSCs within 7 days (Fig. 1A). We applied this differentiation method to other healthy control– and ALS patient–derived iPSCs to generate motor neurons. These iPSC-derived motor neurons expressed motor neuron markers, including HB9, ChAT, and SMI-32 (Fig. 1, B and C). They also formed neuromuscular junctions with the human myoblast cell line Hu5/E18 and showed electrophysiological activity (Fig. 1, D to F, and fig. S2, A to C). To establish an ALS motor neuron phenotypic screening assay, we generated iPSCs from one ALS patient carrying the L144FVX mutation in the SOD1 gene (ALS1) (fig. S1, A to D). We corrected the mutation in established iPSCs using CRISPR-Cas9 to generate isogenic controls: corrected ALS1 (Fig. 1, G to K, and fig. S2, D to G) and corrected ALS1 (second clone) (fig. S2, D to F, H, and I). We differentiated iPSCs into motor neurons using a trio of transcription factors and measured the number of motor neurons expressing the motor neuron marker HB9. There were 62.3 ± 2.3% HB9+ motor neurons in the control, 60.3 ± 2.8% in ALS1, and 63.4 ± 1.3% in corrected ALS1 (Fig. 1G and fig. S2G). In the iPSC-derived motor neurons, we observed accumulation of misfolded mutant SOD1 protein (Fig. 1, H and I), which plays a pathological role in mutant SOD1–associated ALS (26, 27). Furthermore, we found that there was a decrease in ALS motor neuron survival, indicating increased vulnerability to cell death of ALS motor neurons compared to control motor neurons and mutation-corrected isogenic ALS motor neurons (Fig. 1, J and K).

Fig. 1. Generation of motor neurons from iPSCs derived from ALS patients and healthy controls.

Protocol for motor neuron generation from ALS patient iPSCs was as follows: (A) polycistronic vector containing Lhx3, Ngn2, and Isl1 under the control of the tetracycline operator was introduced into iPSCs. Motor neurons were generated from iPSCs within 7 days. Representative images of motor neurons generated from a control iPSC clone (control 1) are shown. Scale bars, 100 μm. (B) Motor neurons generated from iPSCs expressed spinal motor neuron markers, including HB9, ChAT, and SMI-32. Representative images are shown for control iPSC-derived motor neurons (control 1). Scale bars, 10 μm. (C) Real-time polymerase chain reaction analysis showed increased mRNA expression for HB9 and ChAT on day 7 after differentiation from iPSCs (each group represents mean ± SEM, n = 3; Student's t test, *P < 0.05). Data were obtained from control iPSC-derived motor neurons (control 1). (D) Cocultures of iPSC-derived motor neurons with the human myoblast cell line Hu5/E18. Immunoreactivity of a presynaptic protein synapsin colocalized with acetylcholine receptors labeled with α-bungarotoxin (αBTX), demonstrating formation of neuromuscular junctions. A representative image is shown for control iPSC-derived motor neurons (control 1). Scale bar, 10 μm. (E) Action potentials from current-clamp recordings of control iPSC-derived motor neurons (control 1), showing that these motor neurons had electrophysiological activity. (F) Functional neurotransmitter receptors expressed by control iPSC-derived motor neurons (control 1) were evaluated by electrophysiological analysis. Addition of 500 μM glutamate, 500 μM kainate, or 500 μM γ-aminobutyric acid (GABA) induced inward currents during voltage-clamp recordings, showing responses to excitatory and inhibitory neurotransmitters. (G) The percentage of HB9+ motor neurons on day 7 after differentiation was similar for motor neurons derived from healthy control (control 1), isogenic control (corrected ALS1), and ALS patient with the SOD1 mutation (ALS1). Each group represents mean ± SEM, n = 3. n.s., not significant. (H) Misfolded SOD1 protein accumulated in ALS patient iPSC-derived motor neurons with the SOD1 mutation. Scale bars, 10 μm. (I) Accumulation of misfolded SOD1 protein in ALS patient iPSC-derived motor neurons was demonstrated using an immunoprecipitation assay. IP, immunoprecipitates; IB, immunoblotting. (J) Survival of motor neurons on days 7 and 14 after differentiation is shown. Survival of ALS motor neurons carrying the L144FVX mutation (ALS1) was decreased compared to the isogenic control with the mutation corrected (corrected ALS1) and control motor neurons derived from healthy individuals (control 1). Scale bars, 100 μm. The ratio of surviving motor neurons on day 14 to surviving motor neurons on day 7 is shown in (K). Each group represents mean ± SEM, n = 6; *P < 0.05, one-way ANOVA.

Using this cellular model, we set up compound screening with a readout of ALS motor neuron survival. ALS patient–derived iPSCs were differentiated into motor neurons for 7 days. Chemical compounds were added to the cultures for another 7 days, and the surviving motor neurons at day 14 were evaluated using immunostaining for βIII-tubulin (Fig. 2A). Nearly 100% of βIII-tubulin+ neurons also expressed the HB9 motor neuron marker (fig. S2, J and K). Assay performance was determined by calculating the Z′ factor [Z′ factor = 0.42 ± 0.30 (mean ± SD)]. For positive control assays, cells were treated with 50 μM kenpaullone, which has been identified as a candidate compound for ALS (28); the positive effect of kenpaullone on ALS motor neuron survival was confirmed. In negative control assays, the motor neurons were treated with the vehicle dimethyl sulfoxide (DMSO). We conducted high-throughput screening of 1416 compounds that included existing drugs for treating various diseases that are commercially available or undergoing clinical testing. The results of the compound screening are shown in Fig. 2B. Hit compounds were defined as more than 3 SD above motor neuron survival in the negative controls; 27 compounds were identified as hits (hit ratio, 1.7%) (table S3). Representative images of ALS motor neurons showing the neuroprotective effect of the hit compound bosutinib are shown in Fig. 2C. We confirmed the dose dependency of the neuroprotective effect for seven of the hits (Fig. 2D).

Fig. 2. Screening compounds using an ALS motor neuron survival assay.

(A) Overview of compound screen using an ALS motor neuron survival assay. (B) Compounds (1416) consisting of approved drugs and those in clinical testing were screened using ALS motor neurons with the SOD1 mutation (ALS1). Scatterplot shows hits identified by the ALS motor neuron survival index. Hit compounds were defined as more than 3 SD above motor neuron survival in the vehicle control (shown by the yellow line). (C) Representative images of ALS motor neurons on day 7 (top) and on day 14 after treatment with vehicle (middle) or bosutinib (bottom). Treatment with bosutinib increased survival of ALS motor neurons compared to vehicle. Scale bar, 100 μm. (D) Hits and their dose-dependent effects on ALS motor neuron survival. Each group represents mean ± SEM, n = 6; *P < 0.05, one-way ANOVA. (E) Shown are 14 of the 27 hits; these 14 hits were all associated with receptor tyrosine kinase (RTK) and Src/c-Abl signaling pathways. PKC, protein kinase C. (F) Knockdown of Src or c-Abl with siRNAs increased the survival rate of ALS motor neurons carrying the SOD1 mutation. Each group represents mean ± SEM, n = 6; *P < 0.05 one-way ANOVA. (G and H) Phosphorylation of Src/c-Abl was increased in ALS motor neurons with the SOD1 mutation; bosutinib inhibited this phosphorylation as shown by Western blot analysis. Each group represents mean ± SEM, n = 3; *P < 0.05, two-way ANOVA. MW, molecular weight. (I) Representative images showing immunocytochemical staining of p-Src and p-c-Abl in ALS motor neurons (ALS1) and the isogenic control (corrected ALS1). Scale bars, 10 μm. (J) The increase in phosphorylation of Src/c-Abl in ALS motor neurons was blocked by bosutinib treatment, as shown by ELISA. Each group represents mean ± SEM, n = 3; *P < 0.05, two-way ANOVA. bos, bosutinib.

Fourteen of the 27 hits targeted the Src/c-Abl signaling pathway (Fig. 2E). Thus, we focused on Src/c-Abl as a common target of these hits. We re-evaluated other Src/c-Abl inhibitors that had appeared as nonhit compounds and confirmed that they also had a neuroprotective effect, although with reduced efficacy compared to the 14 hits (fig. S3A). Furthermore, knockdown of Src or c-Abl by small interfering RNAs (siRNAs) in ALS patient iPSC-derived motor neurons promoted motor neuron survival (Fig. 2F). This beneficial effect of Src or c-Abl knockdown was canceled if siRNA-resistant forms of Src or c-Abl were expressed by the motor neurons (fig. S3, B and C). Among the Src/c-Abl inhibitors, we focused on hits that directly blocked Src/c-Abl, such as bosutinib and dasatinib. Bosutinib showed dose dependency for ALS motor neuron protection without the bell-shaped responses observed with dasatinib (Fig. 2D); the protective effect was exhibited at lower doses compared to other hits in vitro. From these results, we selected bosutinib for further investigation.

We investigated expression and phosphorylation of Src/c-Abl in ALS patient iPSC-derived motor neurons. Phosphorylation of Src/c-Abl was increased in cultures of ALS motor neurons carrying the SOD1 mutation compared with isogenic controls (Fig. 2, G and H) and treatment with bosutinib decreased phosphorylation of Src/c-Abl, as detected by Western blot analysis (Fig. 2, G and H). Representative immunocytochemical images showing phosphorylated Src (p-Src)/phosphorylated c-Abl (p-c-Abl) in ALS motor neurons and isogenic controls are presented in Fig. 2I. Using enzyme-linked immunosorbent assay (ELISA), we confirmed that phosphorylation of Src/c-Abl was increased in cultures of ALS motor neurons carrying the SOD1 mutation compared to isogenic controls and again found that treatment with bosutinib decreased phosphorylation of Src/c-Abl (Fig. 2J). Next, we evaluated expression and phosphorylation of Src/c-Abl in other types of cells. ALS patient–derived iPSCs (fig. S3, G and H) and astrocytes derived from them (fig. S3, D to F) exhibited increased phosphorylation of Src without increased phosphorylation of c-Abl.

To analyze the mechanism of neuroprotection by bosutinib in ALS patient iPSC-derived motor neurons, we investigated expression of autophagy markers and degradation of misfolded mutant SOD1. We found that expression of the autophagy protein p62 was elevated in ALS motor neurons but decreased after bosutinib treatment (Fig. 3, A to C). A change in the ratio of the autophagic markers LC3-II and LC3-I was suggestive of altered autophagy in ALS motor neurons (Fig. 3, A and B). To confirm whether autophagy was aberrant in ALS motor neurons, we investigated the effect of the mammalian target of rapamycin (mTOR) inhibitor rapamycin. Rapamycin, known to promote autophagy, increased ALS motor neuron survival as did knockdown of mTOR with siRNAs (Fig. 3, D and E). To investigate whether the protective effects of bosutinib were associated with the autophagy pathway, we added the autophagy inhibitors LY294002 and chloroquine to ALS motor neurons after bosutinib treatment. The autophagy inhibitors partially blocked the protective effect of bosutinib (Fig. 3F). Thus, our data suggested that the protective effect of bosutinib might be associated with an increase in autophagy. Furthermore, we found that bosutinib treatment reduced the amount of misfolded SOD1 protein in ALS motor neurons, as shown by Western blotting (Fig. 3G) and ELISA (Fig. 3H), without decreasing SOD1 mRNA expression (Fig. 3I). ALS motor neuron cultures also showed a decrease in intracellular adenosine 5′-triphosphate (ATP) that was partly reversed by bosutinib (Fig. 3J). To further investigate ALS patient iPSC-derived motor neurons, we performed transcriptome analysis using single-cell RNA sequencing (RNA-seq) (tables S4 and S5). We conducted Gene Set Enrichment Analysis to reveal the biological significance of differentially expressed genes between ALS patient iPSC-derived motor neurons and healthy control iPSC-derived motor neurons. We found an increase in mRNA expression of genes associated with the tricarboxylic acid (TCA) cycle and respiratory electron transport chain in ALS motor neurons compared to controls (Fig. 3K). After bosutinib treatment, there was a decrease in mRNA expression associated with TCA cycle and respiratory electron transport chain genes in ALS motor neurons (fig. S4).

Fig. 3. Neuroprotection of ALS motor neurons by Src/c-Abl inhibitors.

(A and B) ALS patient iPSC-derived motor neurons with an SOD1 mutation showed an increase in the autophagy protein p62, which was reversed by bosutinib treatment. Bosutinib also reversed the change in the ratio of the autophagy markers LC3-II and LC3-I in ALS motor neurons. Each group represents mean ± SEM, n = 3; *P < 0.05, two-way ANOVA. (C) ELISA showed an increase in p62 in ALS motor neurons with the SOD1 mutation. Bosutinib treatment decreased the amount of p62 expression in ALS motor neurons. Each group represents mean ± SEM, n = 3; *P < 0.05, two-way ANOVA. (D) Rapamycin treatment increased survival of ALS motor neurons with the SOD1 mutation (ALS1). Each group represents mean ± SEM, n = 6; *P < 0.05, one-way ANOVA. (E) Knockdown of mTOR increased survival of ALS motor neurons with the SOD1 mutation (ALS1). Each group represents mean ± SEM, n = 6; *P < 0.05, Student's t test. (F) Inhibitors of autophagy LY294002 and chloroquine decreased the protective effects of bosutinib on ALS motor neuron survival (ALS1). Each group represents mean ± SEM, n = 6; *P < 0.05, two-way ANOVA. (G and H) Immunoprecipitation (G) and ELISA (H) showed that bosutinib treatment decreased the accumulation of misfolded SOD1 protein in ALS motor neurons (ALS1). (I) Bosutinib treatment did not alter expression of SOD1 mRNA in ALS motor neurons (ALS1). (J) Intracellular ATP was decreased in ALS motor neurons with the SOD1 mutation (ALS1). Bosutinib partly reversed the decrease in ATP. Each group represents mean ± SEM, n = 6; *P < 0.05, two-way ANOVA. (K) Gene Set Enrichment Analysis of single-cell RNA-seq showed up-regulation of genes associated with the TCA cycle and the respiratory electron transport chain. Control motor neurons contained 10 cells of control 1 and 11 cells of control 2. ALS motor neurons with the SOD1 mutation (shown as mSOD1) contained 23 cells of ALS1 and 21 cells of ALS3 (see table S1 for iPSC clones). bos, bosutinib.

Next, we evaluated the effects of Src/c-Abl inhibitors on other proteins that are associated with ALS when mutated (TDP-43 mutations or C9orf72 repeat expansions). Diagnosis of FALS was confirmed by genotype (fig. S1A), and sporadic ALS was confirmed by resequencing ALS patient fibroblasts (table S2). Motor neurons were generated from iPSCs derived from 11 patients with either familial or sporadic ALS (Fig. 4A). Treatment with bosutinib increased survival of ALS motor neurons derived from all patients with FALS and some patients with sporadic ALS (Fig. 4B). Treatment with bosutinib decreased accumulation of misfolded proteins in familial and sporadic ALS motor neurons (fig. S5, A to C).

Fig. 4. Bosutinib treatment of an ALS mouse model and ALS patient iPSC-derived motor neurons.

(A) Images show control and ALS patient iPSC-derived motor neurons on day 7 of differentiation before bosutinib treatment. iPSC clones were generated from FALS patients carrying SOD1 mutations (ALS2 and ALS3), TDP-43 mutations (ALS4, ALS5, and ALS6), and C9orf72 repeat expansions (ALS7, ALS8, and ALS9) and from patients with sporadic ALS (ALS10, ALS11, and ALS12), respectively (see table S1). Motor neurons derived from each iPSC clone were stained with βIII-tubulin. Scale bars, 100 μm. (B) Bosutinib treatment increased survival of iPSC-derived motor neurons generated from patients with sporadic ALS and from eight different patients with FALS carrying SOD1 mutations, TDP-43 mutations, or C9orf72 repeat expansions. Each group represents mean ± SEM, n = 6; *P < 0.05, one-way ANOVA. (C) Kaplan-Meier survival curves showed that bosutinib treatment slightly delayed disease onset in transgenic mice carrying the SOD1 mutation. Mean onset was 123.2 ± 9.1 days for bosutinib and 112.4 ± 14.4 days for vehicle; mean ± SD, log-rank test, P = 0.0021, n = 26 per group. (D) Kaplan-Meier survival curves showed that bosutinib modestly extended the survival of transgenic mice carrying the SOD1 mutation. Mean survival was 164.1 ± 9.4 days for bosutinib and 156.3 ± 8.5 days for vehicle; mean ± SD, log-rank test, P = 0.0019, n = 26 per group. (E) ELISA showed a decrease in misfolded SOD1 protein accumulation in the spinal cords of bosutinib-treated transgenic mice carrying the SOD1 mutation at 12 weeks of age. Each group represents mean ± SEM; nontransgenic control littermates, n = 3; transgenic mice treated with vehicle, n = 3; transgenic mice treated with bosutinib, n = 3; *P < 0.05, one-way ANOVA and post hoc test. (F) Representative images of cresyl violet–stained sections from the lumbar spinal cord ventral horns of transgenic mice carrying the SOD1 mutation at 4.5 months (late symptomatic stage) treated with either vehicle or bosutinib. Scale bars, 50 μm. (G) Quantification of the number of motor neurons on one side of the lumbar spinal cord of transgenic mice carrying the SOD1 mutation treated with either vehicle or bosutinib. Each group represents mean ± SEM; nontransgenic littermate control mice, n = 4; transgenic mice treated with vehicle, n = 5; transgenic mice treated with bosutinib, n = 5; *P < 0.05, one-way ANOVA. bos, bosutinib.

To analyze whether Src/c-Abl inhibitors were effective in vivo, we administered bosutinib to transgenic mice carrying the SOD1 mutation, an established model of mutant SOD1–asssociated ALS. To investigate the effect of Src/c-Abl inhibitors on mouse motor neuron degeneration in vivo, we administered bosutinib (5 mg/kg per day) by intraperitoneal injection starting at 8 weeks of age and continuing until 13 weeks of age. Bosutinib slightly delayed disease onset by 10.8 days (Fig. 4C) and extended survival of mutant SOD1 transgenic mice by 7.8 days compared to untreated mutant SOD1 transgenic mice (Fig. 4D). Src/c-Abl was inhibited (fig. S5D), and there was a decrease in misfolded SOD1 protein in the spinal cords of the bosutinib-treated mutant SOD1 transgenic mice compared to vehicle control (Fig. 4E). The number of motor neurons was higher in bosutinib-treated mutant SOD1 transgenic mice compared to vehicle treatment (Fig. 4, F and G). These results indicated that Src/c-Abl inhibition delayed motor neuron degeneration induced by misfolded mutant SOD1 in vivo.

Finally, we investigated postmortem spinal cord tissue from ALS patients. Immunoreactivity of p-Src was slightly increased in the remaining motor neurons in the spinal cords of four ALS patients (fig. S6A and table S6), although the trend toward increased Src phosphorylation was not significant (fig. S6B and table S7).

DISCUSSION

We developed a phenotypic screen with a readout of motor neuron survival using motor neurons generated from iPSCs derived from a patient with FALS carrying a mutation in the SOD1 gene. Using this assay, we showed that Src/c-Abl inhibitors and knockdown of Src/c-Abl with siRNAs rescued degeneration of ALS motor neurons. Further analysis revealed that these compounds promoted autophagy, reduced the amount of misfolded SOD1 protein, and restored energy homeostasis in ALS motor neurons. Furthermore, treatment with the Src/c-Abl inhibitor bosutinib rescued iPSC-derived motor neurons from patients with other types of ALS caused by mutations in TDP-43 or repeat expansions in C9orf72. Finally, we found that the Src/c-Abl inhibitor bosutinib prolonged the survival of an ALS mouse model by 7.8 days.

Mutations in the SOD1 gene cause conformational changes, leading to misfolding and aggregation. Misfolded aggregated SOD1 is localized in pathological lesions in the spinal cords of animal models and human ALS patients (29). We showed that Src/c-Abl inhibitors reduced the amount of misfolded SOD1 in ALS motor neurons. We observed that Src/c-Abl inhibitors promoted autophagy, supporting previous observations (30, 31). Misfolded SOD1 has been shown to induce endoplasmic reticulum stress, mitochondrial dysfunction (7), and altered membrane properties (32). We observed reduced ATP in ALS motor neurons, prompting us to speculate that misfolded mutant SOD1 induced endoplasmic reticulum stress, altered membrane excitability (33, 34), or mitochondrial dysfunction. Increased expression of TCA cycle and respiratory electron transport chain genes that was reversed by Src/c-Abl inhibitors suggested that ALS motor neurons may be trying to compensate for reduced ATP. A computational model of ALS motor neuron degeneration showed an ATP shortage, implicating mitochondrial involvement (35). We speculated that Src/c-Abl inhibitors could restore ATP levels by decreasing the amount of misfolded proteins through an increase in autophagy. Binding of ATP to Src/c-Abl is required for their activation; by blocking this binding, Src/c-Abl inhibitors (36) may have increased available ATP. It remains unclear why phosphorylation of Src/c-Abl was increased in ALS motor neurons, as shown both here and in a previous report of postmortem ALS spinal cord tissue (24). We speculate that an RTK-mediated mechanism might be associated with Src/c-Abl activation in ALS motor neurons. The signal activation mechanism, potentially promoted by oligomerized proteins in ALS motor neurons, may be similar to the signaling pathways activated by the oncoproteins BCR-ABL and EML4-ALK (37). As oncoproteins, BCR and EML4 form dimers with ABL or ALK, respectively, resulting in activation of cell proliferation and tumor formation. In contrast, misfolded proteins associated with RTKs may form oligomers, leading to activation of Src/c-Abl and the induction of neurodegeneration in nondividing cells such as motor neurons. c-Abl activation is known to result in neurodegeneration of adult mouse neurons (20) and apoptosis in a rat cell line (38).

ALS is a genetically heterogeneous disease (2). Although this heterogeneity may complicate the identification of potential therapeutics, the analysis of iPSC-derived motor neurons from patients with different types of ALS should help to resolve this problem. Our data showed that Src/c-Abl inhibition was effective not only in helping to protect iPSC-derived motor neurons carrying SOD1 mutations, but also ALS motor neurons carrying TDP-43 mutations and C9orf72 repeat expansion. It has been reported that mutant TDP-43 forms oligomers that exhibit reduced DNA binding capability and neurotoxicity (39). C9orf72 repeat expansions lead to the formation of toxic RNA foci and accumulation of dipeptide repeat proteins, resulting in cytotoxicity (10). Sporadic ALS is characterized by accumulation of inclusions containing TDP-43 in motor neurons (40). We observed a decrease in misfolded TDP-43 after treatment with Src/c-Abl inhibitors and speculated that common pathways for neuronal death through apoptosis may be suppressed by Src/c-Abl inhibitors. Both studies of mutant SOD1 transgenic mice treated with dasatinib and TDP-43 transgenic mice treated with bosutinib have shown attenuation of ALS phenotypes (24, 41).

In our compound screen, the Z′ factor was below 0.5, suggesting that the assay might not meet standards. This low score may stem from the fact that the screening takes 7 days to observe neuronal cell death. Although we confirmed the effects of both Src and c-Abl in neuronal degeneration through knockdown experiments using siRNAs, many kinase inhibitors are not specific for just one kinase but inhibit multiple kinases (42). The possibility that the efficacy of Src/c-Abl inhibitors was associated with common off-target effects cannot be ruled out. Furthermore, Src and c-Abl may interact with each other in the pathway of motor neuron death. Further study is needed to identify more specific targets among the Src family including c-Src, Lck, and Lyn (43), and c-Abl.

Here, we generated motor neurons from ALS patient–derived iPSCs through forced expression of transcription factors. Although ALS motor neurons recapitulated vulnerability to cell death, forced gene expression may contribute to cell death as a stress condition. In ALS pathogenesis, noncell autonomous mechanisms have been well explored (2), and cells other than motor neurons are thought to contribute to motor neuron death. By focusing on and generating motor neurons alone, our screening method could guarantee robustness and scalability, but the contributions of cell types other than motor neurons have not been considered. In this regard, it is essential to support the results in vitro by studying the effects of pathogenesis in multiple cell types in vivo. Furthermore, we found that motor neuron death was not observed in motor neurons derived from one sporadic ALS patient iPSC clone, suggesting that the heterogeneity of ALS should be considered in future studies.

Given that our in vitro screen analyzed motor neuron survival, we administered a hit from that screen, bosutinib, to a mouse model of ALS at a time point when motor neurons were beginning to die but before the onset of motor symptoms. The proper dose of bosutinib for mouse experiments could be determined based on the dose used clinically in humans; this bosutinib dose has been used previously in mice (41). We found that bosutinib slightly delayed the onset of disease and modestly extended survival of mutant SOD1 transgenic mice. These in vivo results confirmed our in vitro data but also suggested that this drug treatment is not yet ready for clinical translation. It will be important to examine other doses of bosutinib and other Src/c-Abl inhibitors, such as those that are able to cross the blood-brain barrier. Although the mutant SOD1 transgenic mouse model is useful for evaluating new therapeutics, studies may not always predict human responses in clinical trials (44). It will be important to combine studies of ALS patient iPSC-derived motor neurons and ALS transgenic mouse models in future studies.

MATERIALS AND METHODS

Study design

The objective of our study was to identify a candidate compound or a target for ALS treatment. High-throughput compound screening was performed using ALS patient iPSC-derived motor neurons, and hits were selected based on increased motor neuron survival. One of the hits, the Src/c-Abl inhibitor, bosutinib, was used to treat a mouse model of ALS. Generation and use of human iPSCs was approved by the Ethics Committees of the respective departments at Kyoto University, University of Edinburgh, and Kumamoto University. All methods were performed in accordance with approved guidelines. Informed consent was obtained from all subjects. All mice analyzed in this study were cared for, and procedures were performed in accordance with the Kyoto University Animal Institutional Guidelines, and all experiments were approved by the Center for iPS Cell Research and Application (CiRA) Animal Experiment Committee. Human postmortem samples with written informed consent were obtained from the Department of Medicine and Graduate Schools of Medicine, Kyoto University; Jichi Medical University; and Kansai Medical University.

Generation of iPSCs

iPSCs were generated from skin fibroblasts, peripheral blood mononuclear cells, or immortalized B lymphocytes using retrovirus (Sox2, Klf4, Oct3/4, and c-Myc), sendai virus (Sox2, Klf4, Oct3/4, and c-Myc), or episomal vectors [Sox2, Klf4, Oct3/4, L-Myc, Lin28, and p53-shRNA (short hairpin RNA)], as reported previously (4547), and were cultured on an SNL feeder layer with human iPSC medium (primate embryonic stem cell medium; ReproCELL) supplemented with basic fibroblast growth factor (4 ng/ml) (Wako Chemicals) and penicillin-streptomycin.

Generation of motor neurons

iPSCs, in which polycistronic vectors containing Lhx3, Ngn2, and Isl1 (LNI cassette) were introduced, were dissociated to single cells using Accutase (Innovative Cell Technologies), were plated onto Matrigel-coated dishes or coverslips with neuronal medium [Dulbecco’s modified Eagle’s medium/F12 (Thermo Fisher Scientific), apotransferrin (100 μg/ml) (Sigma-Aldrich), insulin (5 μg/ml) (Sigma-Aldrich), 30 nM selenite (Sigma-Aldrich), 20 nM progesterone (Sigma-Aldrich), and 100 nM putrescine (Sigma-Aldrich)], containing 1 μM retinoic acid (RA) (Sigma-Aldrich), 1 μM smoothened agonist (SAG), brain-derived neurotrophic factor (BDNF) (10 ng/ml) (R&D Systems), glial cell line–derived neurotrophic factor (GDNF) (10 ng/ml) (R&D Systems), and neurotrophin-3 (NT-3) (10 ng/ml) (R&D Systems) with doxycycline (1 μg/ml) (TAKARA), and were cultured for 7 days.

Large-scale screen using ALS motor neurons

iPSCs with the LNI cassette were dissociated to single cells using Accutase and plated onto Matrigel-coated 96-well plates (BD Biosciences) with neuronal medium containing 1 μM RA, 1 μM SAG, BDNF (10 ng/ml), GDNF (10 ng/ml), and NT-3 (10 ng/ml) with doxycycline (1 μg/ml). The libraries used in the compound screens were MicroSource US Drugs (MicroSource Discovery Systems), MicroSource International Drugs (MicroSource Discovery Systems), and kinase inhibitors from EMD and Selleck Chemicals. We selected existing drugs and clinical trial testing drugs from these libraries using Integrity (Thomson Reuters) and NextBio database (NextBio) and used them for throughput screening. Compounds (1416) (final concentration, 10 μM) were added on day 7, and cells were fixed and stained on day 14. DMSO was used as negative control, and kenpaullone, which was identified as an effective compound for ALS treatment, was used as positive control. The number of surviving motor neurons stained with βIII-tubulin was quantified by IN Cell Analyzer 6000 (GE Healthcare).

Statistical analysis

Results were analyzed using one-way or two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test to determine statistical significances of the data. Analysis of disease onset and survival period was performed by log-rank test. Differences were considered significant at P < 0.05. Analyses were performed using SPSS software (IBM).

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/9/391/eaaf3962/DC1

Materials and methods

Fig. S1. Generation of iPSCs from control and ALS patients.

Fig. S2. Characterization of motor neurons and genetic correction of mutant SOD1 iPSCs.

Fig. S3. Investigation of the effects of Src/c-Abl inhibitors.

Fig. S4. mRNA expression changes after bosutinib treatment by single-cell analysis.

Fig. S5. Decrease in misfolded proteins after bosutinib treatment.

Fig. S6. Analysis of postmortem ALS spinal cord tissue.

Table S1. List of iPSC clones.

Table S2. Sequence variations in exon regions for sporadic ALS.

Table S3. List of hit compounds.

Table S4. Genes highly expressed in mutant SOD1 ALS motor neurons identified by single-cell RNA-seq.

Table S5. Genes highly expressed in control motor neurons identified by single-cell RNA-seq.

Table S6. List of postmortem spinal cord tissue for immunohistochemistry.

Table S7. List of postmortem spinal cord tissue for ELISA.

Table S8. Primer list for editing of SOD1 gene.

Table S9. Primer list for quantitative PCR.

References (4856)

REFERENCES AND NOTES

  1. Acknowledgments: We thank all of our co-workers and collaborators including T. Enami, R. Shibukawa, M. Funayama, M. Kawada, K. Goto, H. Houlden, E. Preza, and C. Okada for their technical support. We acknowledge P. Karagiannis for critical reading of the paper and N. Endo and R. Taniguchi for their administrative support. Funding: This work was funded in part by a grant from the iPS Cell Research Fund (S.Y.); the Program for Intractable Diseases Research utilizing Disease-specific iPS cells from the Japan Agency for Medical Research and Development (AMED) (H. Inoue); the Research Center Network for Realization of Regenerative Medicine from AMED (A.H., S.Y., and H. Inoue); Research Project for Practical Applications of Regenerative Medicine from AMED (A.O., T.E., and H. Inoue); Parkinson’s UK Senior Fellowship (F-0902) (T. Kunath); grant-in-aid for scientific research from the Japan Society for the Promotion of Science (15H04270; H. Ito, 15H05581; A.H.); and the Daiichi Sankyo Foundation of Life Science (H. Inoue). Author contributions: H. Inoue conceived the project; K.I. and H. Inoue designed the experiments; K.I., K.T., T.Y., T. Kondo, and S. Kitaoka performed cell culture, molecular experiments, and compound screen; A.W. performed single-cell analysis; K.I. and A. Tanaka performed animal experiments; S. Kaneko, T.A., and H. Ito performed human-sample analysis; N.O., M.H., and H.A. performed resequencing; K.I., A.W., T.Y., D.W., and H. Inoue analyzed the data; K.W., A.H., A.O., T. Kunath., S.W., T.E., T.F., H.N., K.H., H. Ichijo, J-.P.J., and S. Kaneko contributed reagents, materials, and analysis tools; Y.I., M.M., H.T., A. Tamaoka, H.F., K.M., K.O., and R.K. recruited patients; D.W., R.T., and S.Y. provided critical reading and scientific discussions; K.I. and H. Inoue wrote the paper. Competing interests: S.Y. is an unpaid scientific advisor to iPS Academia Japan. Kyoto University has filed patents related to this manuscript: PCT application PCT/JP2014/058142, entitled “Pluripotent stem cells for neuronal differentiation” with K.I. and H. Inoue as coinventors; and PCT application PCT/JP2016/050883, entitled “Agent for preventing and/or treating Amyotrophic lateral sclerosis” with K.I. and H. Inoue as coinventors. All other authors declare that they have no competing interests.
View Abstract

Navigate This Article