Deletion of C9ORF72 Results in Motor Neuron Degeneration and Stress Sensitivity in C. elegans

An expansion of the hexanucleotide GGGGCC repeat in the first intron of C9ORF72 gene was recently linked to amyotrophic lateral sclerosis. It is not known if the mutation results in a gain of function, a loss of function or if, perhaps both mechanisms are linked to pathogenesis. We generated a genetic model of ALS to explore the biological consequences of a null mutation of the Caenorhabditis elegans C9ORF72 orthologue, F18A1.6, also called alfa-1. alfa-1 mutants displayed age-dependent motility defects leading to paralysis and the specific degeneration of GABAergic motor neurons. alfa-1 mutants showed differential susceptibility to environmental stress where osmotic stress provoked neurodegeneration. Finally, we observed that the motor defects caused by loss of alfa-1 were additive with the toxicity caused by mutant TDP-43 proteins, but not by the mutant FUS proteins. These data suggest that a loss of alfa-1/C9ORF72 expression may contribute to motor neuron degeneration in a pathway associated with other known ALS genes.


Introduction
Amyotrophic lateral sclerosis (ALS) is one of the most common neurodegenerative disorders and it is characterized by progressive death of motor neurons in the brain and spinal cord. In 1993, the first ALS gene identified was superoxide dismutase 1 (SOD1) [1] and thanks to recent genetic advances there are now over twenty genes linked to ALS [2]. Genes recently shown to be mutated in ALS include the DNA/RNA binding proteins TAR DNA binding protein 43 (TDP-43) and Fused-in-sarcoma (FUS) [3][4][5][6], and C9ORF72, the latter being a major cause of familial and sporadic ALS [7,8].
GGGGCC repeat expansions are found in the first intron of C9ORF72 and the presence of such long non-coding repeats is suggestive of a toxic gain of function mechanism driving neurodegeneration, perhaps through RNA toxicity, or uncontrolled translation of the repeat into non-native protein species [2,9]. Very little is known about the biological role of C9ORF72 other than its sequence similarity to the GDP/GTP exchange factor "Differentially Expressed in Normal and Neoplasia" (DENN) [10,11]. To learn more about the biological role of C9ORF72 we turned to the model organism Caenorhabditis elegans and characterized the C9ORF72 orthologue F18A1. 6, also called alfa-1 for ALS/FTD associated gene homolog, in a number of behavioral assays. Although appearing morphologically normal we observed that alfa-1(ok3062) null mutants developed an age-dependent motor phenotype and neurodegeneration specific to GABAergic motor neurons. Furthermore, alfa-1(ok3062) mutants showed hypersensitivity to osmotic stress which further exacerbated motor neuron degeneration. Lastly, we observed that alfa-1(ok3062) showed differential genetic interactions with mutant TDP-43 and FUS proteins suggesting a complex interaction amongst some ALS genes.

alfa-1(ok3062) mutants develop an age-dependent motor phenotype
To better understand the pathogenesis that could result from decreased expression of C9ORF72, we examined ok3062 a null allele of alfa-1, the C. elegans orthologue of C9ORF72. ALFA-1 shares 58% homology with C9ORF72 (Blast e-value 2x10 -15 ) ( Figure 1A). The alfa-1(ok3062) mutation is a deletion spanning portions of exons 3 and 4 resulting in no detectable alfa-1 RNA expression ( Figure 1B, C). alfa-1(ok3062) mutants were superficially normal and had total progeny and lifespan comparable to wild type worms ( Figure S1A, B and Table S2). However, when worms were grown on solid media we observed motility defects when the alfa-1(ok3062) mutants reached adulthood, and it ended as an age-dependent paralysis phenotype affecting on average 60% of worms by day 12 of adulthood compared to approximately 20% seen in wild type N2 worms ( Figure 2A and Table S1). The progressive paralysis phenotype may indicate impaired transmission at the neuromuscular junction similar to what we previously observed in our ALS models expressing TDP-43 and FUS proteins in C. elegans motor neurons [12].  In worms, body movement is coordinated by excitatory input from acetylcholine and inhibitory inputs from GABA [13]. Aldicarb is an acetylcholinesterase inhibitor used to indirectly detect dysfunctional transmission at the neuromuscular junction in C. elegans [14], and worms with impaired GABA processing are hypersensitive to aldicarb-induced paralysis [15]. alfa-1(ok3062) mutants were more sensitive to aldicarb induced paralysis compared to wild type worms ( Figure 2B). These data suggest that alfa-1(ok3062) mutants may have impaired inhibitory GABAergic signalling, perhaps recapitulating the neurotransmitter imbalance observed in ALS patients [16]. When worms are grown in liquid culture they display a swimming behaviour, a vigorous activity that actively engages the neuromuscular junction to maintain activity of the body wall muscles. The paralysis phenotype of alfa-1(ok3062) mutants was greatly accelerated when the worms were grown in liquid culture, where approximately 60% of the worms became paralyzed in 8 hours ( Figure 2C), compared to 12 days when grown on solid media.
In addition to neuronal dysfunction, our previous TDP-43 and FUS models also showed age-dependent degeneration of motor neurons [12]. Therefore, to assess for similar phenotypes we examined several neuronal populations in our alfa-1(ok3062) mutants. To do so, we crossed the alfa-1(ok3062) mutation into strains with integrated reporters expressing GFP in different neurons including the GABAergic neurons (unc-47p::GFP), the dopaminergic neurons (dat-1p::GFP) and cholinergic neurons (unc-17p::GFP). At day 9 of adulthood we observed neurodegeneration, in the form of gaps and breaks, only within GABAergic neurons ( Figure 2D, E). Using RNA interference (RNAi), we confirmed that worms treated with alfa-1(RNAi) had motility defects and motor neuron degeneration compared to controls ( Figure S1C, D, Table S1). Thus, our data demonstrate that decreased expression of alfa-1 causes age-dependent motor defects accompanied by the specific neurodegeneration of the GABAergic motor neurons.

ALFA-1 is required for resistance to osmotic stress
A number of genes linked to ALS have roles in the cellular stress response [17], and C. elegans is a convenient system to investigate ALS gene orthologues and stress signalling [18,19].
To gain further insight about the role of ALS genes and stress, we subjected alfa-1(ok3062) mutants to several, distinct environmental insults. A major regulator of the cellular, and organism-wide stress response signalling in C. elegans is the Insulin-IGF pathway. DAF-2 is the sole Insulin/IGF-like receptor in C. elegans and hypomorphic mutations in daf-2 result in stress resistant and long-lived phenotypes compared to wild type animals [18,20]. In our environmental stress assays, wild type N2 worms are typically stress-sensitive and show progressive lethality while daf-2(e1370) animals are highly resistant to stress-induced lethality. Thus, we asked where alfa-1(ok3062) mutants functioned along this stress sensitivity axis. Wild type N2 worms and alfa-1(ok3062) mutants were equally sensitive to thermal stress, while daf-2(e1370) and alfa-1(ok3062);daf-2(e1370) mutants were both highly resistant ( Figure 3A). We used the natural compound juglone to test for oxidative stress associated lethality and observed that N2 and alfa-1(ok3062) mutants were comparably sensitive, while daf-2(e1370) and alfa-1(ok3062);daf-2(e1370) mutants were equally resistant to oxidative stress ( Figure 3B). Finally, we examined osmotic stress using sodium chloride and observed that alfa-1(ok3062) mutants were more sensitive to osmotic stress associated lethality compared to N2 worms ( Figure 3C). Results were also confirmed by alfa-1(RNAi) ( Figure S2A). In the absence of alfa-1(ok3062), daf-2(e1370) mutants were slightly less resistant to osmotic stress at a concentration of 400 mM NaCl ( Figure 3C). When increasing the concentration to 500 mM NaCl, a significant difference was observed when comparing alfa-1(ok3062);daf-2(e1370) to daf-2(e1370), where a loss of alfa-1 impairs the resistance of daf-2(e1370) worms ( Figure 3D). At 600 mM NaCl, both strains die after 60 minutes ( Figure 3D). alfa-1(ok3062) had no effect on dauer formation or the long-lived phenotypes of daf-2(e1370) mutants ( Figure  S2C, D and Table S2). These data suggest that ALFA-1 has a specific role in protecting worms against osmotic stress, perhaps involving the insulin-IGF pathway.
It has been hypothesized that in addition to causative mutations, secondary genetic or environmental factors may contribute to motor neuron degeneration in ALS [21]. Thus, we investigated whether an impaired response to osmotic stress in alfa-1(ok3062) worms would impact the degeneration of motor neurons. Using the unc-47p::GFP reporter to visualize the GABAergic motor neurons, we subjected unc-47p::GFP or unc-47p::GFP;alfa-1(ok3062) worms to acute thermal stress that induced comparable levels of neurodegeneration ( Figure  3E). However, we observed that acute osmotic stress resulted in a higher rate of motor neurodegeneration in unc-47p::GFP;alfa-1(ok3062) animals compared to unc-47p::GFP transgenic controls ( Figure 3F). The same experiment was carried using alfa-1(RNAi) and similar results were obtained ( Figure S2B). These data suggest that the motor neurons of alfa-1(ok3062) animals are specifically sensitive to osmotic stress and that this type of environmental stress may be relevant to the function of C9ORF72.
to either alfa-1(ok3062) or FUS S57∆ alone ( Figure 4B and Table  S1). These data suggest that the genetic interactions between alfa-1 and TDP-43 or FUS are not equivalent, and that perhaps alfa-1 and FUS S57∆ function in the same pathway, while TDP-43 may use parallel or independent pathways resulting in motor neuron dysfunction.

Discussion
Many questions remain to be answered about the role of C9ORF72 in the pathogenesis of ALS. It is still not clear whether the GGGGCC repeat expansion results in a loss of function, a gain of function or both, or if the size of the repeat has differential effects on these mechanisms of action. Recent reports have observed decreased expression of C9ORF72 when the GGGGCC repeat reaches pathogenic length [7,22,23]. Since no clear mechanisms have been demonstrated for C9ORF72 toxicity, in vivo models are important tools to investigate normal biological functions that may lead to insights about the disease state. C9ORF72 protein sequence is highly similar to ALFA-1 protein sequence. It was hypothesized by two different groups that C9ORF72 share common feature with DENN proteins [10,11]. Interestingly, Zhang et al. have also shown that the amino acids most conserved between C9ORF72 and other DENN proteins are also conserved between C9ORF72 and ALFA-1. Therefore, we hypothesize that depletion of ALFA-1 represents the depletion of C9ORF72 and its impact as a DENN protein. DENN proteins are involved in the regulation of Rab GTPases including Rab35. Rab35 has roles in endocytosis, exosome fusion, synaptic vesicle function and regulation of the actin cytoskeleton [24]. It remains to be seen if C9ORF72 functions in any of these cellular processes and whether or not hexanucleotide mutations impact normal C9ORF72 function perhaps leading to neurodegeneration.
We investigated the biological consequences of deleting the ALFA-1 from C. elegans as a putative model for decreased expression of C9ORF72 in ALS. alfa-1(ok3062) mutant worms displayed motility defects that progressed into age-dependent paralysis accompanied by the specific neurodegeneration of GABAergic motor neurons. A locomotion deficit caused by a decreased expression of C9ORF72 was recently reported in zebrafish [23] corroborating our results that decreased expression of this protein can cause a motor phenotype. Further characterization of ALFA-1 remains to be done, as it will be important to determine that ALFA-1 is expressed in the nervous system, which has been reported for mouse and fish models [8,23]. However, the rapid onset of motor phenotypes in alfa-1(ok3062) mutants in liquid culture sets the stage for chemical screens using neuroprotective molecules. Our nascent drug testing experiments suggest that alfa-1(ok3062) toxicity may be distinct from TDP-43, since protective molecules identified previously [19,25] do not suppress motor alfa-1(ok3062) phenotypes (data not shown).
Of interest is the recurring theme involving ALS genes in the cellular stress response [17]. We have previously shown that TDP-1, the orthologue of TDP-43 in C. elegans, is also involved in response to osmotic stress [18]. Also, in cellular models FUS was shown to robustly react to osmotic stress and increase resistance to this stress [26]. We observed that alfa-1 mutants were specifically sensitive to osmotic stress and that exposure to this stress enhanced motor neuron degeneration. Cells maintain extensive quality control mechanisms to preserve protein homeostasis against environmental or intrinsic challenges [27]. Osmotic stress can lead to cellular shrinkage, macromolecular crowding and increased protein aggregation with perhaps irreversible degenerative outcomes [28]. Thus it is easy to appreciate the importance of maintaining osmotic balance over the life of a neuron especially since many proteins linked to ALS have a propensity to misfold and aggregate. Here we showed that osmotic stress enhanced neurodegeneration in alfa-1(ok3062) mutants, but the impact of osmotic stress on motor neuron health awaits further investigation, which may in time open a new avenue for potential therapeutic strategies.
Recent progress in genetics is developing a more complete picture of the ALS spectrum, but with this information comes the need to better understand the interactions of ALS genes under normal, pathogenic and aging conditions [2]. Work from genetically expedient model organisms has investigated the genetic interactions of several ALS genes [18,[29][30][31]. However, the picture is not yet complete and nothing is currently known about genetic interactions of C9ORF72 with other ALS genes. We observed that deletion of alfa-1 enhanced motor defects by mutant TDP-43, but not by mutant FUS. One interpretation of these data is that since the motor neuron degeneration caused by the loss of alfa-1 is additive to the toxicity of mutant TDP-43 proteins, these mechanisms function in parallel or separate pathways. Oppositely, a mechanism may be shared by alfa-1 and mutant FUS, as the level of paralysis observed was comparable to each condition alone or in combination. Additional tools and experiments are required to better understand the basis of these genetic interactions. Our findings are intriguing in light of observations that C9ORF72 ALS cases show TDP-43 pathology in the absence of FUS pathology [2]. If C9ORF72 functions together in the same pathway as FUS, one might expect to observe FUS pathology along with C9ORF72 mutations, but this has not been observed. Furthermore, the involvement of TDP-43 may not be specific since TDP-43 pathology is observed in a number of neurological disorders. Finally our genetic approach uncovered the effects of alfa-1 loss of function on mutant TDP-43 or mutant FUS neuronal toxicity, and while informative, is not representative of hexanucleotide mutations in C9ORF72 and their effects on the aggregation of wild type TDP-43 or FUS proteins.
The molecular pathogenic mechanisms behind the genetic mutations of many ALS genes are not fully understood. For many of these genes it is not known whether mutations lead to a gain of function, a loss of function, or both. An informative example comes from studies of TDP-43 where several in vivo models suggested that both occur simultaneously [18,30,32,33]. A similar situation may exist for C9ORF72 where loss of expression leads to motor phenotypes, in conjunction with recent findings demonstrating that the expression of GGGGCC RNA is toxic in a Drosophila model [34], and that repeats can be inappropriately translated into different peptides with additional potential cytotoxic effects [35,36]. Thus, further characterization of both mechanisms will unravel the toxicity caused by the presence of the GGGGCC repeat in the first intron of C9ORF72.

Amino acid sequence alignment
The ALFA-1 (F18A1.6) isoform 1 protein sequence (WormBase) was used as the subject sequence and the C9ORF72 protein sequence as the query sequence (accession number NM_001256054.1) and were aligned using BLASTP (http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins). Sequence alignment and visualization were done using ClustalW [38] and BoxShade (http://www.ch.embnet.org/ software/BOX_form.html). Conserved amino acids are marked in grey and identical amino acids in black.

RT-PCR
RNA was extracted using Trizol. After worm lysis and homogenization, chloroform was added and tubes were centrifuged. RNA was precipitated from the aqueous phase using isopropanol, pellets were washed with 75% ethanol and resuspended in water. RNA was reverse transcribed with the QuantiTect kit (Qiagen) preceded by gDNA wipeout. 1 µl of cDNA was used for act-3 and F18A1.6 amplification using the following primers; F18A1.6 forward 5' AATGAGCGGAACATCAAGC 3', F18A1.6 reverse 5' TTCGGATATGTCAGGCTGAAG 3', act-3 forward 5'GTTGCCGCTCTTGTTGTAGAC 3', act-3 reverse 5' GGAGAGGACAGCTTGGATGG3' Paralysis assay 30 adult, day-one worms were transferred to NGM with FUDR plates and scored daily for movement. Worms were counted as paralysed if they failed to move after being prodded in the nose. Experiments were conducted at 20°C and done in triplicates. Survival curves were produced and compared using the Log-rank (Mantel-Cox) test. 30 adult, day-one worms were transferred to NGM with 1mM aldicarb plates. Worms were scored every 30 minutes for two hours and counted paralysed if they failed to move after being prodded on the nose. Experiments were conducted at 20°C and done in triplicates. Survival curves were produced and compared using the Log-rank (Mantel-Cox) test using GraphPad Prism software.

Liquid culture
Synchronized populations of worms were obtained by hypochlorite extraction. 20-30 young adults were distributed in 96-well plate containing OP 50 and incubated for eight hours at 25°C. Worms were counted paralysed if they failed to move after gently tapping the side of the plate. The mean and SEM were calculated and two-tailed t-tests were used for statistical analysis. 30 adult, day-one worms were transferred to NGM-FUDR plates and counted every two days. Worms were counted as dead if they did not respond to tactile stimulus. Survival curves were produced and compared using the Log-rank (Mantel-Cox) test.

Stress assays
Worms were grown at 20°C on normal NGM plates until day one of adulthood. 30 adult, day one worms were then transferred to NGM plates + 240 µM juglone (oxidative stress), or NGM + 400 mM NaCl, or NGM + 500mM NaCl, or NGM + 600 mM NaCl (osmotic stress). Tests were carried at 20°C for oxidative and osmotic stresses and at 37°C for thermal stress. Worms were counted every two hours for up to 14 hours. For all experiments, worms were counted as dead if they did not respond to tactile stimulus. Survival curves were produced and compared using the Log-rank (Mantel-Cox) test.

Neurodegeneration assay
To score gaps or breaks, synchronized animals were selected at day one, five and nine of adulthood for in vivo visualization. For neurodegeneration count during stress tests, adult day one worms were transferred to NGM + 400 mM NaCl at 20°C (osmotic stress) or normal NGM and put at 37°C (thermal stress) for six hours. To confirm those results with RNAi, rrf-3(pk1426) worms submitted to alfa-1 or EV RNAi up to day 1 of adulthood. Worms were then transferred on 400 mM NaCl for six hours. For visualization, animals were immobilized in M9 with 5 mM of levamisole and mounted on slides with 2% agarose pads. Neurons were visualized with a Leica 6000 microscope and a Leica DFC 480 camera. For all experiments, a minimum of 100 worms was scored over at least 3 trials for all conditions. The mean and SEM were calculated and two-tailed t-tests were used for statistical analysis. Figure S1. alfa-1(ok3062) worms had normal (A) progeny and (B) lifespan compared to N2 worms. (C) rrf-3(pk1426) worms submitted to alfa-1 RNAi display motility defects causing paralysis at day 12 of adulthood compared to rrf-3(pk1426) worms submitted to empty vector (EV). alfa-1 RNAi showed an increase neurodegeneration of GABAergic motor neurons (unc47p::GFP) after 6 hours under osmotic stress compared to rrf-3(pk1426) worms submitted EV in the same conditions (P<0.0001). The alfa-1(ok3062) mutation had no effect on (C) dauer formation or (D) the extended lifespan of daf-2(e1370) mutants. (TIF)