Prof. Hideyuki Okano, MD, PhD, Department of Physiology, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. Email: email@example.com
Amyotrophic lateral sclerosis (ALS) results in the selective loss of motor neurons within several years of disease onset; however, the molecular mechanisms behind ALS pathogenesis remain largely unknown. Among the genes responsible for familial ALS, mutations in TAR DNA-binding protein 43 (TDP-43) have been identified. The present study evaluated the cytotoxicity of TDP-43 fragments and ALS-associated mutants against human embryonic stem cell-derived motor neurons.
Magnetofection was used to investigate the mortality rates of motor neurons after forced expression of TDP-43 C-terminal fragments and ALS-associated TDP-43 missense mutants.
Neither wild-type TDP-43 nor the 35-kDa fragment induced cell death. However, the 26-kDa fragment, which forms insoluble aggregates in the motor neurons of ALS patients, showed significant cytotoxicity, similar to mutant forms of TDP-43, including p.A315T and p.A382T. Hence, a motor neuron-specific neurotoxic function might be attributed to both the 26-kDa fragment and the TDP-43 mutants. Thus, 26kDa C-terminal fragment was considered to play an important role in ALS pathology, but alterations in adenosine deaminase acting on RNA 2 (ADAR2) expression, an enzyme accountable for GluA2 (AMPA receptor subunit 2) mRNA editing insufficiency in sporadic ALS, did not lead to alterations in the expression of the 26-kDa fragment.
This is the first report showing the influence of specific ALS-associated factors on the mortality of motor neurons through the manipulation of gene expression. The motor neuron disease model described herein might prove useful for the identification of novel pathogenic factors and neuroprotective agents for the treatment of ALS.
Amyotrophic lateral sclerosis (ALS) was first described in the 1860s by JM Charcot, and is characterized by the rapid and selective loss of upper and lower motor neurons. Beginning with superoxide dismutase 1 (SOD1), genome analyses have allowed the identification of genes and gene mutations responsible for the pathogenesis of ALS. Mutations in TAR DNA-binding protein-43 (TDP-43),[1, 2] fused in sarcoma/translocated in liposarcoma (FUS/TLS),[3, 4] optineurin and hexanucleotide repeat expansion (GGGGCC) within C9ORF72[6, 7] are also linked to sporadic ALS.
Pathologically, TDP-43 is a major component of the tau-negative and ubiquitin-positive inclusions that typify ALS, and are related to intranuclear RNA metabolism. TDP-43 acts as an important processing regulator for many kinds of RNA, and is involved in microRNA biogenesis and self-splicing events; thus, TDP-43 has received much attention in the research for ALS pathomechanisms. Previous studies regarding familial ALS cases identified numerous missense mutations in the TDP-43 gene (e.g. p.A315T and p.A382T); however, the molecular mechanisms underlying ALS pathogenesis are still largely unknown.
TDP-43 is associated with two truncated C-terminal fragments: a 35-kDa fragment and 18-26-kDa fragments, which appear as a smear on polyacrylamide gel electrophoresis. The 35-kDa and 26-kDa fragments of TDP-43 are thought to be produced by: (i) translation downstream from the normal start codons of full-length TDP-43; and (ii) caspase-dependent cleavage in the cytoplasm. Unlike full-length TDP-43 confined to the nucleus, the 35-kDa C-terminal fragment lacks a nuclear localization signal (NLS), and is redistributed to the cytoplasm besides the nucleus. The 26-kDa also lacks a NLS, and is diffusely distributed to both the nucleus and cytosol throughout the entire cell.
Insoluble accumulation in the cytoplasm of the motor neurons in ALS cases consists of 18-26-kDa phosphorylated C-terminal fragments[11, 12] The formation of these aggregates is the most distinctive pathological change specifically observed in the spinal cords of ALS patients, but it is unknown whether the C-terminal fragment plays a role in survival or death of motor neurons in ALS pathomechanism. Therefore, the primary aim of the present study was to explore the cytotoxicity of the 35-kDa and 26-kDa fragments, as well as the p.A315T and p.A382T mutants, against human embryonic stem (ES) cell-derived motor neurons.
RNA editing at the Q/R site of AMPA receptor subunit 2 (GluA2) mRNA is reportedly insufficient in the spinal motor neurons of sporadic ALS patients. A recent study showed that TDP-43 knockdown did not affect the efficiency of GluA2 mRNA editing in vitro. However, the precise relationship between RNA editing and the expression pattern of the 26-kDa fragment of TDP-43 remains to be elucidated. Thus, the present study also investigated whether alterations in the expression level of adenosine deaminase acting on RNA 2 (ADAR2), the enzyme accountable for GluA2 mRNA editing, influenced the expression level of the 26-kDa fragment of TDP-43.
Human ES cell-derived motor neurons
Human motor neuron progenitors derived from human ES cell lines were obtained from California Stem Cell (catalog number FP-6046; Irvine, CA, USA) and maintained in MotorBlast Culture Media (California Stem Cell) during the cell mortality assay.
Each member of the TDP-43 vector series, which was subcloned into the pcDNA3.1/V5-His A expression vector, was provided by Dr Daisuke Ito, Keio University, Japan. ADAR2-myc cDNA was generated by polymerase chain reaction by using a cDNA library from HeLa cells and sense (5′-GGAATTCACCATGGATATAGAAGATGAAGAAAACATGAG-3′) and antisense (5′-GGATATCTCACAGGTCTTCTTCAGAGATCAGTTTCTGTTCGGGCGTGAGTGAGAACTGG-3′) primers. The antisense primer contained a myc epitope. ADAR2-myc cDNA was excised with EcoRI and EcoRV, and cloned into the pcDNA3 vector (Invitrogen, Carlsbad, CA, USA). Plasmids were verified by DNA sequencing.
The sequences of the shRNA used to knockdown ADAR2 are shown in Table 1. ADAR2-specific shRNA were designed using the online siRNA design program, siDirect (http://sidirect2.rnai.jp/), Invitrogen BLOCK-iT RNAi Designer (https://rnaidesigner.invitrogen.com/rnaiexpress/) and the sequences of MISSION shRNA Plasmid DNA (Sigma-Aldrich, St. Louis, MO, USA). The negative control sequence was previously described. Sense and antisense oligonucleotides consisting of shRNA sequences were inserted into pENTR4-H1 and recombined with a CS-RfA-EG self-inactivating lentivirus vector (provided by Dr Hiroyuki Miyoshi, Riken BioResource Center, Japan) using Gateway LR Clonase (Invitrogen) according to the manufacturer's instructions. Finally, the constructed CS-RfA-EG vector was transfected into HEK293T cells by lipofection without packaging.
Table 1. shRNA sequences to knockdown adenosine deaminase acting on RNA 2
Mouse monoclonal anti-βIII tubulin antibody (TUJ-1, clone TU-20) was purchased from Millipore (Milford, MA, USA). Mouse monoclonal antibody against non-phosphorylated neurofilament (clone SMI-32) was from Covance (Indiana-polis, IN, USA). Rabbit polyclonal anti-Islet-1 and anti-HB9 antibodies were from Abcam (Cambridge, MA, USA). Rabbit polyclonal anti-GFAP was from DAKO (Carpenteria, CA, USA). Goat polyclonal anti-ADAR2 (E-20) was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit polyclonal anti-ADAR1 and anti-TDP-43 antibodies were from ProteinTech (Chicago, IL, USA). Mouse monoclonal anti-β-actin (AC-15) was from Sigma-Aldrich, and mouse monoclonal anti-V5 (R960-25) was from Invitrogen.
To characterize human ES cell-derived motor neurons, cells were fixed, blocked and exposed to primary antibodies (TUJ-1 [1:1000], anti-HB9 [1:100], anti-GFAP [1:1000], SMI-32 [1:1000] and anti-Islet-1 [1:200]) overnight at 4°C. For Figure S1, cells were fixed 7 days after magnetofection (described later) with 4% paraformaldehyde at room temperature for 10 min and then permeabilized in 0.2% Triton X-100 for 5 min. After blocking in TNB blocking buffer (Perkin Elmer, Norwalk, CT, USA) to reduce non-specific binding, cells were incubated with monoclonal anti-V5 (1:400; Invitrogen). Primary antibody application was followed by CY3 (Alexafluor 555 or 488)-conjugated secondary antibodies (Invitrogen).
Magnetofection procedures were used for vector transfection into human ES cell-derived motor neurons. The conditions, including the volume ratio of the vector to the magnetofection agent (NeuroMag; OZ Biosciences, Marseille, France), were optimized for mature motor neurons according to the manufacturer's instructions. Briefly, a mixture of TDP-43 vector (0.5 μg), green fluorescent protein (GFP) vector (0.25 μg) and NeuroMag (0.9 μL) in serum-free OptiMEM medium (50 μL; Invitrogen-Gibco, Carlsbad, CA, USA) was prepared and added to 150 μL of MotorBlast Culture Media. Cells were then incubated on the MagnetoPlate (OZ Biosciences) for 15 min. One hour after removing the cells from the MagnetoPlate, the medium was exchanged with fresh MotorBlast Culture Media, resulting in the sufficient expression of a vector driven by the cytomegalovirus promoter.
Cell mortality assay
Seven days after magnetofection, cultures that coexpressed each TDP-43 vector and the GFP vector were incubated with 0.2 μmol/L ethidium homodimer-1 (EthD-1) in phosphate-buffered saline (PBS) for 30 min at 37°C, fixed with 4% paraformaldehyde for 15 min on ice and then washed twice with PBS. The mortality rate was calculated by counting the number of EthD-1-positive cells among 30 randomly chosen GFP-positive cells.
HEK293T cell culture
HEK293T cells were maintained in Dulbecco's modified Eagle medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum as previously described.[18, 19] Lipofection was carried out using GeneJuice Transfection Reagent (Novagen, Madison, WI, USA) according to the manufacturer's instructions.
A total of 48 h after lipofection, HEK293T cells were lysed in cold lysis buffer (50 mmol/L Tris-HCl, pH 7.4, 150 mmol/L NaCl, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, 0.25% sodium dodecyl sulfate (SDS), 5 μmol/L ethylenediaminetetraacetic acid (EDTA), 0.05 mmol/L phenylmethylsulfonyl fluoride and Complete, EDTA-free Protease Inhibitor Cocktail Tablets [Roche, Indianapolis, IN, USA]). Cell lysates were briefly sonicated and then separated by reducing SDS polyacrylamide gel electrophoresis on a 4–20% Tris-glycine gradient gel (Invitrogen). Proteins were then transferred onto a polyvinylidene difluoride membrane (Millipore). The membrane was incubated with primary antibodies (anti-ADAR2, 1:1000-1:500; anti-ADAR1, 1:500; anti-b-actin, 1:10,000; anti-TDP-43, 1:1000; and anti-V5, 1:1000), followed by horseradish peroxidase-conjugated secondary antibodies. Proteins were visualized using electrochemiluminescence (ECL) or ECL Prime Detection reagent (GE Healthcare, Waukesha, WI, USA) and an ImageQuant LAS 4000 digital imaging system (GE Healthcare). Optical density was quantified as the mean gray value in pixels using Image J software (http://rsbweb.nih.gov/ij/) followed by subtraction of the background density.
The Fisher's exact probability test was used to compare the mortality rates between human ES cell-derived motor neurons transfected with full-length and other TDP-43 vectors. To compare the intensities on immunoblot, values calculated from at least three independent experiments were compared using Student's t-test. P < 0.05 was considered statistically significant. Each error bar represents the standard deviation of the mean.
TDP-43-induced cellular toxicity
To investigate whether TDP-43 fragments and/or mutant forms have cytotoxic or cytoprotective effects on mature motor neurons, human ES cell-derived motor neurons were generated in vitro as reported previously. In addition to positive staining of HB-9 and Islet-1 (89.4 ± 4.0% and 88.8 ± 9.3%, respectively), immunocytochemistry showed that 88.3 ± 7.8% and 56.8 ± 20.7% of the cells expressed βIII-tubulin, a marker for neurons, and non-phosphorylated neurofilament protein, a marker for mature neurofilaments, respectively. Meanwhile, just 4.9 ± 3.2% of the cells expressed glial fibrillary acidic protein (GFAP), a marker for astrocytes (Fig. 1a–d). These results are indicative of the efficient differentiation of the human ES cells into mature motor neurons.
After the successful demonstration of neuronal maturation in vitro, cellular toxicity assays were carried out by transfecting each TDP-43 fragment or ALS-associated mutant into ES cell-derived motor neurons. Several gene delivery methods can be carried out to introduce foreign genes of interest into neurons, including lipofection and calcium phosphate-mediated transfection. Magnetofection, in contrast, is less common; however, this procedure affords comparatively low cytotoxicity and sufficiently high efficiency for the transfection of primary neurons. Thus, as in previous studies,[18, 21, 22] this novel transfection procedure, combined with the magnetofection reagent, NeuroMag (OZ Biosciences, Marseille, France), was chosen to enable the study of cellular toxicity in this study (Fig. 2a).
First, the coexpression of GFP- and TDP-43-containing vectors was confirmed after their transfection into ES cell-derived neurons. As a result, the cotransfected vectors were both efficiently expressed after magnetofection (Fig. S1). Cell death induced by the expression of the truncated C-terminal fragments or the TDP-43 mutants was definitively detected by EthD-1 staining (Fig. 2b). The mortality rate was calculated as the ratio of the number of EthD-1-positive cells to the number of GFP-positive cells cotransfected with each TDP-43 construct.
Cell lethality was not significantly increased after transfection of the 35-kDa C-terminal fragment into neurons relative to full-length TDP-43 (13.3 ± 0.0% vs 10.9 ± 4.2%, respectively, P = 0.693; Fig. 2c). By contrast, transfection of the 26-kDa fragment and the mutant TDP-43 forms, p.A315T and p.A382T, resulted in mortality rates of 30.0 ± 5.5%, 43.4 ± 4.7% and 49.2 ± 6.3%, respectively. These figures were significantly higher than that after transfection of full-length TDP-43 (P < 0.001; Fig. 2c). Thus, the 26-kDa C-terminal fragment, in addition to the mutant forms of TDP-43, caused apparent neurotoxicity in mature human motor neurons.
Effect of ADAR2 knockdown or overexpression on expression of the 25-26-kDa TDP-43 fragments
Recent reports suggest that the accumulation of insoluble 18-26-kDa TDP-43 fragments and/or insufficient RNA editing of GluA2 mRNA in the spinal cord might reflect the pathological pathway implicated in sporadic ALS.[14, 19] To test whether ADAR2 expression affected the expression pattern of the 26-kDa fragment, two ADAR2-specific short hairpin RNA (shRNA), shR-1 and shR-2, were designed (Table 1), and transfected into HEK293T cells (Fig. S2). The efficiency and specificity of the shRNA vectors were verified by immunoblot analysis of ADAR2 protein expression with cotransfection of ADAR2(-myc) (Fig. 3a). ADAR2 protein levels were markedly suppressed by shR-1 and shR-2, but no suppressive effect was observed on the p150 or p110 isoform of ADAR1, the other mammalian enzyme with RNA editing activity. The band intensity of the 25-26-kDa fragment relative to that of full-length TDP-43 was not altered by ADAR2 knockdown. The intensity ratios were 1.05 ± 0.28 for shR-1, 1.03 ± 0.44 for shR-2 and 1.00 ± 0.45 for shR-n/c (negative control) treatment (Fig. 3b). To confirm this finding, full-length TDP-43 tagged with a V5 epitope at the C-terminus was coexpressed with an ADAR2 vector or a control vector, again resulting in no difference between the intensity ratios (1.03 ± 0.24 for ADAR2 vs 1.00 ± 0.29 for control vector expression; Fig. 3c,d).
Together with a recent report showing that TDP-43 knockdown did not consistently influence RNA editing in vitro, there seems to be no direct relationship between ADAR2 expression and the expression pattern of the 25-26-kDa TDP-43 fragments.
The neuroprotective versus neurotoxic actions of the 35-kDa and (18-)26-kDa C-terminal fragments of TDP-43 have not been elucidated in mature human motor neurons. The present study used human ES cell-derived motor neurons as an in vitro model for ALS and transfected each TDP-43 fragment or well-known ALS-associated mutants into the neurons to test its neurotoxicity.
Transfection of full-length TDP-43 or the 35-kDa fragment showed no significant cytotoxicity against motor neurons compared with mock transfection in our system. Because the 35-kDa fragment contributes to the formation of stress granules, this fragment might protect neurons under the disease condition induced by ALS. However, the 26-kDa fragment of TDP-43, which forms insoluble aggregates in ALS-afflicted motor neurons, induced significant cell death as efficiently as mutant forms of TDP-43, p.A315T and p.A382T. These findings suggest that the 26-kDa fragment and the ALS-associated mutants would exert their toxic actions through a gain-of-function mechanism.
Immunoblotting of the insoluble fraction isolated from the ALS spinal cord showed hyperphosphorylation of the (23-)26-kDa TDP-43 fragment.[11, 12] Furthermore, insoluble TDP-43 fractions isolated from ALS (Q343R, M337V)-induced pluripotent stem (iPS) cell-derived motor neuron-containing neural populations also showed increased levels of the 26-kDa fragment. Hence, the 26-kDa fragment of TDP-43 might have pro-apoptotic effects in ALS motor neurons.
A recent report showed that suppression of TDP-43 expression does not necessarily induce changes in editing efficiency at the Q/R site of GluA2 mRNA. This implies that TDP-43 pathology does not take place upstream of reduced ADAR2 activity in the ALS pathomechanism. Additionally, knockdown of ADAR2 did not have an effect on the production of the 25-26-kDa C-terminal fragments of TDP-43 in the present study. Another recent report shows that TDP-43 mislocalization in the cytoplasm of murine motor neurons, accompanied by defects in ADAR2 localization and GluA2 mRNA editing, occurs with increased age. Further investigation will be required to clarify the connection between ADAR2 activity and pathological TDP-43 aggregation.
An in-depth characterization of ES cell-derived motor neurons, and the expression levels and intracellular distribution of endogenous TDP-43 will also be required to further understand the role of TDP-43 in ALS etiology. For example, the possibility that full-length TDP-43, or its 35-kDa C-terminal fragment, safeguard neurons against the loss of vital cellular functions cannot be disregarded. Furthermore, the lack of epigenetic considerations in the present study (including factors that might alter RNA metabolism) complicates our gain-of-function interpretation of the ALS pathomechanism discussed earlier.
Phenotypic changes and alterations in neuronal process formation, TDP-43 distribution and RNA expression profiles have all been shown in motor neurons generated from iPS cells derived from human ALS patients (ALS-iPS cells). Cell-specific vulnerability was also observed in motor neurons derived from human ALS-iPS cells. However, to the best of our knowledge, no previous reports have evaluated cell mortality induced by exogenous expression of a TDP-43 C-terminal fragment in human motor neurons. Better characterization of disease progression in ALS-iPS cell-derived motor neurons and evaluation of cell death induced by ALS-associated factors in control iPS cell-derived motor neurons will be needed to clarify the ALS pathomechanism. Despite the availability of mutant SOD1 transgenic mice and ADAR2 conditional knockout mice, there is currently no animal model system that completely recapitulates the progression of ALS in humans. Thus, human stem cell-derived motor neuron disease models, such as that described herein, are likely to prove increasingly useful for the identification of unknown pathogenic factors and novel neuroprotective agents for the treatment of ALS.
It remains to be determined why the 26k-Da fragment had a toxic effect on motor neurons. In this regard, the 35-kDa fragment retains the RNA recognition motifs (RRM) 1 and 2 found in full-length TDP-43, but lacks the nuclear localization signal (NLS). By contrast, the 26-kDa fragment lacks all of RRM1 and part of RRM2, in addition to the NLS. These structural defects suggest that the 26-kDa fragment could interact with functional RNA differently from full-length TDP-43 or the 35-kDa fragment. For instance, alterations in the binding affinity of RNA (or RNA processing proteins) to RRM1, RRM2 or the C-terminal domain might be involved in the pathogenesis of ALS. The 26-kDa fragment has also been detected in the nucleus, suggesting that it might have an aberrant role in nuclear RNA metabolism.
As such, screening RNA bound to ALS-associated RNA-binding proteins, such as TDP-43, provides an important direction for future studies of ALS pathology and progression. These studies should optimally be carried out in an environment that closely reflects the endogenous environment of ALS motor neurons. Therefore, it would be important to evaluate the extent to which ES cell- and iPS cell-derived motor neurons actually recapitulate the molecular and cellular characteristics of bonafide motor neurons. Future studies should focus on RNA-binding proteins associated with ALS (e.g. TDP-43, FUS/TLS and other ALS-associated RNA processing factors), and the intramolecular pathogenic changes in the target RNA should be assessed by taking advantage of this stem cell-derived motor neurons model.
Although the present study identified the role of the 26-kDa fragment in ALS pathology, the mechanisms associated with motor neuron degeneration, including RNA metabolism, require further investigation at the molecular level. To this point, innovative screening methods, such as HITS-CLIP and RNA sequencing procedures using Next Generation Sequencers, can now target all kinds of RNA, including non-coding and small RNA. We anticipate that a deeper understanding of the molecular pathogenesis of ALS will lead to the identification of novel therapeutic targets and agents for its treatment.
Conflict of interest
HK has financial interests in California Stem Cell, a company with interests related to the study. HK serves as the Chairman of the Scientific Advisory Board at California Stem Cell. HK also owns equity interests in California Stem Cell, which is in addition to his salary from the University of California at Irvine. AP is an employee of California Stem Cell, and as such, has financial interest in California Stem Cell. Other authors claim no conflicts of interests related to this study.
We thank Dr Satoshi Suyama, Dr Satoshi Kawase and Dr Ken-ichiro Kuwako (Keio University, School of Medicine) for technical assistance, and Dr Daisuke Ito (Keio University, School of Medicine) for providing the TDP-43 plasmid series. This study was supported by a Grant-in-Aid for Young Scientists (B) from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan to YN; a Keio University Grant-in-Aid for Encouragement of Young Medical Scientists to YN; an Inochi-no-Iro ALS Research Grant to YN; the Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST Program) to HO; and a Grant-in-Aid for the Global COE (Center of Excellence) program from the MEXT to Keio University.
All research dealing with human ES cells and motor neuron progenitor cells were carried out at California Stem Cell. All research using human ES cell-derived motor neurons in Japan was carried out after the definite differentiation into the nervous system. In Keio University, all the experiments were carried out according to the Guidelines on the Utilization of Human Embryonic Stem Cells of the MEXT, using human ES cell-derived motor neurons that are classified as “differentiated cells” in article 2 of this guideline. All studies were carried out in accordance with the Declaration of Helsinki.