The exocyst subunit EXOC2 regulates the toxicity of expanded GGGGCC repeats in C9ORF72-ALS/FTD

SUMMARY

GGGGCC (G₄C₂) repeat expansion in C9ORF72 is the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). How this genetic mutation leads to neurodegeneration remains largely unknown. Using CRISPR-Cas9 technology, we deleted EXOC2, which encodes an essential exocyst subunit, in induced pluripotent stem cells (iPSCs) derived from C9ORF72-ALS/FTD patients. These cells are viable owing to the presence of truncated EXOC2, suggesting that exocyst function is partially maintained. Several disease-relevant cellular phenotypes in C9ORF72 iPSC-derived motor neurons are rescued due to, surprisingly, the decreased levels of dipeptide repeat (DPR) proteins and expanded G₄C₂ repeats-containing RNA. The treatment of fully differentiated C9ORF72 neurons with EXOC2 antisense oligonucleotides also decreases expanded G₄C₂ repeats-containing RNA and partially rescued disease phenotypes. These results indicate that EXOC2 directly or indirectly regulates the level of G₄C₂ repeats-containing RNA, making it a potential therapeutic target in C9ORF72-ALS/FTD.

In brief

Halim et al. deleted the gene EXOC2 from patient stem cells and then differentiated them into motor neurons. They found that several amyotrophic lateral sclerosis-related phenotypes were rescued in patient neurons when EXOC2 was deleted or knocked down by a drug. This study identifies EXOC2 as a potential therapeutic target.

Graphical Abstract

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are devastating and incurable neurodegenerative diseases that share genetic, molecular, and pathological mechanisms.^1–3 In particular, the most common genetic cause of both diseases is GGGGCC (G₄C₂) repeat expansion in the first intron of C9ORF72.^4,5 A key pathological hallmark of C9ORF72-ALS/FTD is the presence of five dipeptide repeat (DPR) proteins translated from expanded sense and antisense repeat RNAs: poly(GA), poly(GP), poly(GR), poly(PR), and poly(PA).^6–8 Poly(GR) expression patterns correlate with neurodegeneration in C9ORF72 patients’ brains,^9–11 and ectopic expression of poly(GR) confers strong toxicity in diverse in vitro and in vivo models.^12–19 Poly(GR) interacts with several cellular components such as ribosomes, stress granules, spliceosomes, and nucleolar proteins and impairs various cellular functions, including translation, nucleocytoplasmic transport, stress granule dynamics, splicing, mitochondrial function, DNA damage, and stress response.^{15,17,20–27}

Despite this recent progress, our understanding of altered pathways in C9ORF72-ALS/FTD is incomplete, and we do not know which pathway(s) can be targeted therapeutically in patient neurons. In an unbiased genetic screen in poly(GR)-expressing Drosophila, we identified Sec5 as a genetic suppressor of poly(GR) toxicity in non-neuronal cells.²¹ Sec5/EXOC2 is a subunit of the octameric exocyst complex. Exocysts participate in protein trafficking and spatiotemporal regulation of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex assembly for exocytosis, which is essential for neuronal function and other cellular processes during development and in adulthood.^28–30

In this study, we investigated the role of EXOC2 in the pathogenesis of C9ORF72-ALS/FTD in motor neurons (MNs) derived from induced pluripotent stem cells (iPSC-MNs). Using CRISPR-Cas9 technology and antisense oligonucleotides (ASOs), we found that the loss of EXOC2 activity suppresses disease-like phenotypes such as neuronal cell death and neurite degeneration, and, surprisingly, did so by reducing the levels of poly(GR) and expanded G₄C₂ repeats-containing RNA. Our findings indicate that EXOC2 directly or indirectly regulates the level of G₄C₂ repeats-containing RNA and may serve as a potential therapeutic target in C9ORF72-ALS/FTD.

RESULTS

Sec5 is a genetic modifier of poly(GR) toxicity in Drosophila neurons

We previously identified sec5 as a genetic suppressor of poly(GR) toxicity in non-neuronal Drosophila cells.²¹ To examine whether partial knockdown of sec5 suppresses poly(GR) toxicity in neuronal cells, we expressed poly(GR) in Drosophila photoreceptor neurons at the adult stage only. In this experiment, 80 copies of GR (GR₈₀) with a FLAG tag at the N terminus were expressed from the (GGXCGX)₈₀ sequence with an AUG start codon in photoreceptor neurons under the control of GMR-Gal4 in the presence of the temperature-sensitive yeast transcription suppressor Gal80 (Gal80^ts).^21,31 At 1–3 days of age, adult male and female flies were shifted from 18°C to 25°C, a temperature at which Gal80 does not inhibit Gal4-mediated poly(GR) transcription anymore.^21,31 Control flies had a normal eye phenotype characterized by well-organized ommatidia and lack of necrosis (Figure S1A). (GR)₈₀ expression led to an eye degeneration phenotype characterized by disarrangements in ommatidia and necrosis (Figure S1A). Partial loss of sec5 in two different RNAi lines suppressed poly(GR) toxicity in the eye (Figures S1A and S1B). The efficiency of sec5 RNAi knockdown was confirmed by qRT-PCR (Figure S1C).

Generation and characterization of EXOC2 deletion in C9ORF72 patient iPSCs

The human ortholog of Drosophila Sec5, EXOC2, encodes an essential subunit of the exocyst complex that plays a key role in multiple cellular processes.^28–30 To investigate whether and how EXOC2 regulates poly(GR) toxicity in human neurons, we first did western blot analyses of 3-week-old and 3-month-old iPSC-MNs and found no age-dependent change in EXOC2 levels between C9ORF72 patient neurons and their isogenic controls lacking expanded G₄C₂ repeats²¹ (Figures 1A and 1B).

Figure 1. Generation of a deletion in EXOC2 by the CRISPR-Cas9 technology in iPSC lines derived from two C9ORF72 patients.

(A and B) Representative western blot images of EXOC2 in 3-week-old (A) and 3-month-old (B) isogenic control and C9ORF72 iPSC-MNs and densitometric quantification of EXOC2 protein level. Each data point represents one differentiation (n = 3 independent differentiations). Values are mean ± SEM. ns, not significant (one-way ANOVA with Tukey’s multiple comparisons test).

(C) Schematic representation of the deletion in EXOC2 near exon 13 in C9ORF72 iPSC lines.

(D) qRT-PCR analysis of EXOC2 mRNA levels in 3-week-old MNs differentiated from parental and EXOC2 deletion iPSC lines (EXOC2-del iPSC-MNs).

(E) Representative western blot analysis of 3-week-old parental and EXOC2-del iPSC-MNs. Both samples in parental C9-26 or Del-1 came from the same iPSC line. Arrowhead indicates a truncated form of EXOC2, which is present but weaker in Del-1 lanes.

(F) Representative immunostaining images for the dendrite-specific marker MAP2 in 3-week-old parental and EXOC2-del iPSC-MNs.

(G) Quantification of the numbers of primary dendrites from C9ORF72 and EXOC2-del iPSC-MNs as illustrated in (F). Twenty randomly chosen fields were quantified for analysis.

(H) Representative immunofluorescence image of surface NMDAR 2B staining in 3-week-old parental and EXOC2-del C9ORF72 iPSC-MNs.

(I) Mean fluorescence values of NMDAR 2B from C9ORF72 and EXOC2-del iPSC-MNs as illustrated in (H). Ten randomly chosen fields were quantified for analysis. Neurons were transduced with lentivirus expressing GFP. Each data point represents one differentiation in (D) and one neuron in (G) (n = 3–4 independent differentiations). Values are mean ± SEM. **p < 0.01; ****p < 0.0001 by one-way ANOVA with Dunnett’s multiple comparisons (patient 1) and two-tailed Student’s t test (patient 2). Scale bar, 20 μm.

See also Figures S1 and S2.

Next, we used CRISPR-Cas9 technology^32,33 to mutate EXOC2 in iPSC lines from two C9ORF72-ALS/FTD patients that caused a 97-nt deletion in exon 13 and intron 13 (Figure 1C). The deletion was confirmed by PCR analysis of genomic DNA (Figure S2A). Sanger sequencing analysis further confirmed the 97-nt deletion present in all cells that leads to a frameshift and a premature stop codon (Figure S2B). Indeed, qRT-PCR analysis showed a near-complete loss of EXOC2 mRNA in iPSC-MNs with EXOC2 homozygous deletion (EXOC2-del) (Figure 1D). Western blot analysis confirmed the complete loss of full-length EXOC2 protein (Figure 1E)–a surprising result because EXOC2 is required for cell viability.^34,35 However, a truncated EXOC2 fragment was detected at a low level in EXOC2-del iPSC-MNs differentiated from all three deletion iPSC lines, although the signal of the truncated band showed variability among the three lines (Figure 1E). The size of the truncated fragment corresponded to the fragment expected from the premature stop codon after CRISPR editing in EXOC2-del lines. Evidently, the EXOC2 fragment retains enough function to maintain cell viability. qRT-PCR analysis showed a 50% reduction in EXOC2 mRNA in iPSC-MNs carrying the EXOC2 heterozygous deletion (EXOC2-het del) (Figure S2C). However, the EXOC2 protein level in EXOC2-het del iPSC-MNs did not differ from that in the parental line (Figures S2D and S2E), indicating that a 50% reduction in EXOC2 mRNA is not sufficient to decrease its protein level, probably because it exists in the exocyst complex. Therefore, only EXOC2-del lines (homozygous deletion) were used for the rest of the study. Karyotyping analysis confirmed the absence of chromosomal abnormalities in these CRISPR-edited iPSC lines (Figure S2F).

Next, we determined whether EXOC2-del lines show partial loss of EXOC2 function. Indeed, 3-week-old EXOC2-del iPSC-MNs had fewer primary dendrites than parental neurons (Figures 1F and 1G), consistent with a previous report.³⁴ Since exocyst helps localize N-methyl-d-aspartate receptors (NMDARs) to the plasma membrane,³⁶ we analyzed the surface localization of NMDARs. In neurons transduced with GFP-expressing lentivirus, the NMDAR 2B signal showed a significant reduction in surface localization in EXOC2-del neurons (Figures 1H and 1I), but the total level of NMDAR 2B appeared to be the same as that in parental neurons (Figures S2G and S2H). Thus, a change in total NMDAR 2B expression level was not responsible for the decrease in NMDARs on neuronal cell membranes. These results suggest that EXOC2-del lines exhibit a partial loss of EXOC2 function.

EXOC2 deletion rescues disease-relevant cellular phenotypes in C9ORF72 iPSC-derived patient motor neurons

Expression of cleaved caspase 3 (CC3) level, an apoptosis induction marker, is greatly increased in 3- to 4-month-old C9ORF72 iPSC-MNs.²¹ To investigate whether EXOC2 deletion rescues apoptotic cell death, we examined CC3 levels in 3-month-old isogenic control, C9ORF72, and EXOC2-del iPSC-MNs. The CC3 level was variably but significantly higher in C9ORF72 iPSC-MNs than in isogenic controls (Figures 2A and 2B). In C9ORF72 neurons from patient 1, EXOC2 deletion decreased the CC3 level; in EXOC2-del C9ORF72 neurons from patient 2, the level was even lower than in the isogenic controls (Figures 2A and 2B).

Figure 2. EXOC2 deletion rescues C9ORF72-ALS/FTD-related disease phenotypes.

(A and B) Representative images (A) and densitometric quantification (B) for western blot analysis of CC3 in 3-month-old isogenic control, C9ORF72, and EXOC2-del iPSC-MNs. The upper half of the blot was probed for EXOC2 and actin, and the lower half of the blot was probed for CC3.

(C) Representative images for immunostaining for TUJ1 showing neurite degeneration in 3-week-old isogenic control, C9ORF72, and EXOC2-del iPSC-MNs after neurotrophic factor withdrawal. Arrowheads indicate beading or fragmentation. Scale bar, 20 μm.

(D) Quantification of images as shown in (C) for MN cultures obtained from the two C9ORF72 patients. Degeneration index was measured by the ratio of fragmented neurites over total area. Twenty randomly chosen fields were quantified for analysis.

(E) Representative images for western blot analysis of p62 level in 2-month-old isogenic control, C9ORF72, and EXOC2-del iPSC-MNs.

(F) Densitometric quantification of p62 protein level in 2-month-old cultures from isogenic control, C9ORF72, and EXOC2-del iPSC-MNs as shown in (E). Each data point in (B) and (F) represents one differentiation (n = 3–6 independent differentiations); each data point in (D) represents one field. Values are mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by one-way ANOVA with Dunnett’s multiple comparisons test.

See also Figure S2.

Cortical neurons derived from iPSCs of C9ORF72 patients have more neurite degeneration than isogenic control neurons after 2–3 weeks of neurotrophic factor withdrawal.³¹ To examine whether EXOC2 deletion rescues neurite degeneration, we deprived iPSC-MNs of different genotypes of neurotrophic factor for 3 weeks. Neurite degeneration was significantly greater in C9ORF72 iPSC-MNs than in isogenic controls and was also rescued in EXOC2-del iPSC-MNs in both patient lines (Figures 2C and 2D).

Dysregulation of autophagy is another feature of C9ORF72-ALS/FTD. In C9ORF72 patient iPSC-derived neurons, levels of the selective autophagy receptor protein p62 are increased, and DPRs colocalize with p62 inclusions.^37,38 We found that the p62 level was significantly higher in 2-month-old C9ORF72 iPSC-MNs than in isogenic controls; this phenotype was rescued in EXOC2-del iPSC-MNs differentiated from both patient lines (Figures 2E and 2F).

EXOC2 deletion reduces transcripts containing expanded G₄C₂ repeats in C9ORF72 iPSC-MNs

To examine whether decreased DPR levels were responsible for the rescue of disease-relevant cellular phenotypes such as neuronal cell death and p62 accumulation, we measured poly(GR) levels with a Meso Scale Discovery immunoassay in a blinded manner.^21,31,39 The poly(GR) level was much lower in EXOC2-del iPSC-MNs than in C9ORF72 iPSC-MNs differentiated from both patient lines (Figure 3A).

Figure 3. EXOC2 deletion reduces poly(GR) and C9ORF72 mRNA levels.

(A) Poly(GR) level in 2-month-old isogenic control, parental, and EXOC2-del C9ORF72 iPSC-MNs.

(B) Schematic representation of the different C9ORF72 transcript variants and location of the primer sets used to quantify each transcript.

(C and D) qRT-PCR analysis of relative C9ORF72-V1 (C) and C9ORF72-V3 (D) mRNA levels in parental and EXOC2-del C9ORF72 iPSC-MNs.

(E) Schematic representation of the location of the (A/G) SNP in C9ORF72.

(F) Pyrosequencing quantification of the G allele in C9ORF72 and EXOC2-del iPSC-MNs from one patient. Each data point represents one differentiation (n = 3–6 independent differentiations). Values are mean ± SEM. **p < 0.01; ***p < 0.001; ****p < 0.0001 by one-way ANOVA with Dunnett’s multiple comparisons test (A and F and patient 1 in C and D) and two-tailed Student’s t test (patient 2 in B and C).

See also Figure S3.

To further explore the mechanism by which EXOC2 deletion reduces the poly(GR) level, because G₄C₂ repeat-containing RNA is the template for poly(GR) expression, we analyzed the levels of two variants, V1 and V3, which are splicing products of the expanded G₄C₂ repeat-containing C9ORF72 pre-mRNA. We used primers specific to V1 and V3 mRNAs³¹ (Figure 3B) and found that their levels in C9ORF72 neurons were lower than that in isogenic control neurons (Figure S3A), consistent with our earlier result comparing C9ORF72 neurons with non-isogenic control neurons.³⁷ We also found, surprisingly, that both V1 and V3 mRNA levels were significantly lower in EXOC2-del iPSC-MNs than C9ORF72 iPSC-MNs differentiated from both patient lines (Figures 3C and 3D). To confirm this finding, we differentiated C9ORF72 iPSC lines and EXOC2-del C9ORF72 iPSC lines again and measured EXOC2 mRNA (Figure S3B), V1 (Figure S3C), and V3 mRNA (Figure S3D) levels in a blinded manner and obtained the same result. We also quantified in a blinded manner the number of repeats-containing RNA foci (Figures S3E and S3F), the level of poly(GR) (Figure S3G), and the level of poly(GA) (Figure S3H) that is also translated from expanded G₄C₂ repeat RNA, and found they were all greatly reduced in C9ORF72 MNs with EXOC2 deletion. Next, we analyzed the allele-specific expression of C9ORF72-V3, taking advantage of the rs10757668 SNP (A/G) within exon 2.^4,31,40 While the wild-type allele is associated with nucleotide A, the mutant allele containing expanded G₄C₂ repeats is associated with nucleotide G (Figure 3E). The SNP G is present in patient 1 but not patient 2 neurons and preferentially expressed more than the A allele as shown by pyrosequencing and allele-specific quantification (Figure 3F). The relative level of the G allele was lower in EXOC2-del C9ORF72 iPSC-MNs than in the parental C9ORF72 iPSC-MNs of patient 1 (Figure 3F). To confirm this result, we used ASOs (see below) to knock down EXOC2 expression level in iPSC-derived MNs of another patient with the SNP G and confirmed the preferential downregulation of expanded G₄C₂ repeats-containing RNA (Figure S3I), albeit to a lesser extent. These findings suggest that EXOC2, either directly or indirectly, preferentially regulates the level of C9ORF72 RNAs containing expanded G₄C₂ repeats.

EXOC2-specific ASOs reduce levels of mRNA and pre-mRNA containing expanded G₄C₂ repeats in C9ORF72 iPSC-MNs

Our finding that EXOC2 deletion decreases the levels of C9ORF72 transcripts containing expanded G₄C₂ repeats was surprising because EXOC2 has no known role in transcription or repeat RNA stability. It may do so indirectly through yet-to-be-identified molecular pathways. Moreover, it is unknown whether knocking down EXOC2 in postmitotic mature C9ORF72 patient neurons rescues disease phenotypes without affecting neuronal integrity. To address these questions, we used two ASOs to knock down EXOC2 in iPSC-MNs of two C9ORF72 patients and two healthy controls (Figure S4A). Control ASO did not change the EXOC2 mRNA level, whereas EXOC2 ASOs reduced the EXOC2 mRNA level by 70%–80% (Figure S4B), and as a result, EXOC2 protein was decreased by 50%–60% (Figures S4C and S4D).

qRT-PCR analysis revealed a significant decrease in the C9ORF72-V1 mRNA level in ASO-treated motor neurons differentiated from iPSC lines of two C9ORF72 patients (Figure S4E). qRT-PCR analysis of the C9ORF72-V3 mRNA level also showed similar results (Figure S4F). Using primers specific to the V1/V3 pre-mRNA (Figure 3B) also revealed a significant decrease in neurons of both patients (Figure S4G), indicating that the level of expanded G₄C₂ repeats-containing RNA is greatly down-regulated in C9ORF72 neurons without EXOC2. However, in both patient and healthy control neurons, ASO treatment did not alter C9ORF72-V2 levels in transcripts encoding full-length C9ORF72 protein lacking the G₄C₂ repeat sequence (Figure S4H).

Treatment of C9ORF72 iPSCs-derived mature motor neurons with EXOC2 ASOs rescues disease phenotypes

Early loss of EXOC2 results in neurodevelopmental defects in animal models and patients.^34,35 In contrast, ALS and FTD are age-dependent neurodegenerative disorders that arise in mid- to late adulthood. Since the loss of EXOC2 also ameliorated disease phenotypes associated with C9ORF72-ALS/FTD, we investigated the potential therapeutic implications of EXOC2 ASOs in iPSC-MNs. Old neurons tend to form large clumps, making ASO uptake inefficient and rendering it impossible to evaluate the effect of EXOC2 ASOs on apoptosis in 3-month-old neurons. Therefore, we implemented an ASO treatment regimen combined with a stress condition to detect apoptosis at an earlier time point. CC3 levels increase 18–40 h after neurotrophic factor withdrawal in SOD1 patient iPSC-MNs.⁴¹ Since neuronal maturation takes about 2 weeks, we treated mature neurons with control or EXOC2 ASOs for 10 days, starting on day 16, withdrew neurotrophic factors for 40 h, and then examined apoptosis in neurons or phosphorylated neurofilament heavy chain (pNfH) level in the culture medium (Figure 4A).

Figure 4. Treatment of mature iPSC-MNs with EXOC2 ASOs rescues disease phenotypes.

(A) Schematic of ASO treatment and stress induction in mature iPSC-MNs.

(B) Representative western blot image showing EXOC2 knockdown efficiency in mature isogenic control or C9ORF72 motor neurons.

(C) Densitometric quantification of EXOC2 protein levels.

(D) G₄C₂ repeats-containing pre-mRNA levels in C9ORF72 patient neurons treated with EXOC2 ASOs.

(E and F) Representative western blot image (E) and densitometric quantification (F) of CC3 in mature isogenic control and C9ORF72 iPSC-MNs treated with control or EXOC2 ASOs.

(G) ELISA showing pNfH in culture media from mature isogenic control and C9ORF72 iPSC-MNs treated with control or EXOC2 ASOs to assess neurodegeneration. pNfH levels were normalized to total protein concentration.

(H and I) qRT-PCR analysis of relative EXOC2 mRNA and C9ORF72 V1/V3 pre-mRNA levels in 3-week-old C9ORF72 patient 4 (H) or patient 3 (I) iPSC-MNs left untreated or treated with 5 μM nontargeting control ASO or EXOC2 ASOs.

(J) pNfH levels (normalized to total protein concentration) in culture media from mature iPSC-MNs of C9ORF72 patients and healthy subjects treated with control or EXOC2 ASOs to assess neurodegeneration. Each data point represents one differentiation (n = 3–4 independent differentiations). Values are mean ± SEM. **p < 0.01; ***p < 0.001; ****p < 0.0001 by one-way ANOVA with Dunnett’s multiple comparisons test.

See also Figure S4.

ASO treatment reduced the EXOC2 protein level by 50%–60% in both mature C9ORF72 and isogenic control iPSC-MNs (Figures 4B and 4C), which led to a reduction in the levels of G₄C₂ repeats-containing pre-mRNA (Figure 4D), but it did not change the number of MAP2⁺ primary dendrites (Figures S4J and S4K) or mean neurite length per neuron (Figure S4L). Thus, later interventions in mature neurons result in the partial loss of function of EXOC2 without causing neurite outgrowth defects.

Next, we assessed the induction of apoptosis 40 h after neurotrophic factor withdrawal. The CC3 level was increased in C9ORF72 neurons compared with control neurons, and EXOC2 ASOs rescued this phenotype. As a control, EXOC2 ASO did not alter the CC3 levels in isogenic control iPSC-MNs (Figures 4E and 4F).

In the same samples, we also measured the release of pNfH into the culture medium by ELISA. pNfH is released as a result of axonal damage, and an elevated pNfH level in cerebrospinal fluid is considered a diagnostic and prognostic biomarker for ALS.^42,43 We found that the pNfH level in the medium was higher in control ASO-treated C9ORF72 iPSC-MNs than in isogenic controls. EXOC2 ASOs rescued this phenotype, but it did not change the pNfH level in the medium of isogenic control iPSC-MNs (Figure 4G).

Lastly, to further confirm these findings, we used ASOs to knock down EXOC2 mRNA expression by 70%–80% in MNs differentiated from an additional two C9ORF72 patients and found that the level of V1/V3 pre-mRNA was significantly decreased as well (Figures 4H and 4I). Correspondingly, the pNfH levels in the medium of MN cultures from these two C9ORF72 patients but not from two healthy subjects were decreased after EXOC2 ASO treatment (Figure 4J). Evidently, the treatment of mature neurons with EXOC2 ASOs can rescue disease phenotypes and reduce expanded G₄C₂ repeats-containing RNA levels without causing morphological defects in mature neurons.

DISCUSSION

In this study, we identified Sec5/EXOC2 as another genetic modifier of poly(GR) toxicity in fly photoreceptor neurons and C9ORF72-ALS/FTD iPSC-derived motor neurons. EXOC2 is an exocyst subunit that contains an IPT/TIG (immunoglobulin-like fold, plexins, transcription factors/transcription factor immunoglobulin) domain, which is found in some receptors and several transcription factors (e.g., NF-kB and NFAT).^44–47 CRISPR-Cas9-mediated deletion of EXOC2 in iPSCs or ASO-mediated partial knockdown of EXOC2 in mature C9ORF72 iPSC-MNs rescued disease-relevant cellular phenotypes such as apoptosis and neurite degeneration. Surprisingly, the loss of EXOC2 reduced toxic poly(GR) levels by directly or indirectly decreasing the level of RNAs containing expanded G₄C₂ repeats. Because of this unexpected result, we could not examine whether the loss of EXOC2 could also modulate poly(GR) toxicity itself in patient neurons in a way similar to that in flies. Nonetheless, our findings suggest that EXOC2 directly or indirectly affects the level of RNAs containing expanded G₄C₂ repeats through a to-be-identified pathway.

We generated CRISPR-Cas9-mediated homozygous deletion lines of EXOC2 in C9ORF72 iPSC-MNs that express a truncated N-terminal EXOC2 fragment at a low level. Complete loss of EXOC2 and other exocyst subunits causes cell death and is embryonic lethal.^34,35,48,49 Our findings suggest that a truncated fragment of EXOC2 retains enough function to maintain cell viability. After we made this surprising observation, homozygous EXOC2 deletion mutations were identified in patients with a neurodevelopmental disorder characterized by intellectual disability, developmental delay, and brain abnormalities.³⁵ As a remarkable coincidence, that study reported that two adult patients with a homozygous deletion mutation produce a truncated EXOC2 fragment similar in size to the one generated in our C9ORF72 iPSC lines. Thus, a low level of the N-terminal EXOC2 fragment is sufficient to retain most of the function of the protein.

Using genetic knockout and ASO knockdown approaches, we found that the loss of EXOC2 function rescues neurodegeneration in C9ORF72 patient neurons. Thus, EXOC2 is another genetic modifier of the toxicity associated with expanded G₄C₂ repeats. Many genetic modifiers whose partial knockdown improves the neuronal survival of C9ORF72-ALS/FTD patient neurons have been reported, including the ones identified in our own lab such as Ku80,²¹ AFF2/FMR2,³¹ p53,²¹ HSP90, and TOPOII.⁵⁰ It is not known whether variants in these genes influence the onset or progression of ALS/FTD. Our findings indicate that EXOC2 knockdown by ASOs in mature C9ORF72 patient neurons can rescue cellular phenotypes relevant to ALS/FTD without compromising neuronal morphology. Thus, EXOC2 is a promising therapeutic target that merits further investigation.

Limitations of the study

One limitation of the present study is that the molecular mechanism by which loss of EXOC2 decreases the level of expanded G₄C₂ repeats-containing RNA remains unknown. We were unable to obtain convincing evidence that EXOC2 somehow may directly regulate transcription of expanded G₄C₂ repeats in C9ORF72. Other potential mechanisms are also possible. For instance, EXOC2 is known to interact with upstream small GTPase RALB and then recruit TBK1 to activate downstream signaling and transcriptional events in innate immune responses.⁵¹ Moreover, as part of the exocyst complex, EXOC2 regulates many other cellular processes.²⁹ It will be informative to examine how EXOC2 may act directly or indirectly through a specific molecular pathway to modulate the transcription and/or stability of expanded G₄C₂ repeats-containing RNA.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Fen-Biao Gao (fen-biao.gao@umassmed.edu)

Materials availability

Requests for resources and additional information should be directed to and will be fulfilled by the lead contact. Reagents generated in this study will be made available on request, but we may require a completed Materials Transfer Agreement if there is potential for commercial application.

Data and code availability

• Data supporting all the findings in this study are presented within the article and the supplementary information.

• This paper does not report original code.

• Any additional information required to reanalyze the results reported in this study is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

iPSC culture

Two iPSC lines from C9ORF72 carriers, 26L6 (Patient 1) and 27L11 (Patient 2), their isogenic control iPSC lines, 26Z90 (Control 1) and 27M91 (Control 2), three C9 patient iPSC lines, 16L14 (Patient 3), 40L3 (Patient 4) and 29iALS (Patient 5, for pyrosequencing), as well as iPSC lines from healthy subjects, 2L20 (Control 3) and 37L20 (Control 4) have been used before.^21,37 iPSCs were cultured as described.^21,50 Briefly, iPSCs were maintained in mTeSR1 medium (mTeSR1 Basal Medium with mTeSR1 5X supplement, Stem Cell Technologies) in six-well plates coated with Matrigel (cat. no. 354230, Corning), diluted 1:100 in KnockOut DMEM/F-12 (cat. no. 12660012, Gibco). The mTeSR1medium was replaced daily. iPSCs were passaged every 4–6 days. Cells were washed with Dulbecco’s phosphate-buffered saline (DPBS), dissociated in Accutase (Cat no. SCR005, Millipore) diluted 1:2 in DPBS for ~1 min at room temperature (RT), washed with DPBS, and scraped in fresh mTeSR1 medium with a cell lifter. After reaching the desired colony size, cells were seeded n fresh mTeSR1 medium in Matrigel-coated six-well plates. Spontaneously differentiated colonies were manually removed, split, and differentiated into motor neuron.

Motor neuron differentiation

iPSCs were differentiated into spinal motor neurons as described.^21,52 At ~ 50% confluency, small colonies of iPSCs were seeded in mTeSR1 medium in Matrigel-coated six-well plates. Next day, the medium was replaced with neuroepithelial progenitor (NEP) induction medium consisting of neural medium KnockOut DMEM/F-12 medium (cat. no. 12660012, Gibco) and Neurobasal medium (cat. no 21103049, Gibco) (1:1), 0.5X N2 Supplement (cat. no. 17502-048, Gibco), 0.5X B27 Supplement (cat. no. 17504044, Gibco), 0.1 mM ascorbic acid (cat. no. A4403, Sigma), 1X Glutamax (cat. no. 35050061, Thermo Fisher Scientific), supplemented with 3 μM CHIR99021 (cat. no. 72054, Stem Cell Technologies), 2 μM DMH1 (cat. no. 4126, Tocris Bioscience), and 2 μM SB431542 (cat. no. 04-0010-10, Stemgent). NEP medium was changed every other day for 6 days. NEPs were dissociated with 1:2 diluted Accutase and split 1:3 into Matrigel-coated six-well plates. NEPs were cultured in motor neuron progenitor (MNP) induction medium, the neural medium supplemented with 1 μM CHIR99021, 2 μM DMH1, 2 μM SB431542, 0.1 μM retinoic acid (cat. no. R2625, Sigma), and 0.5 μM Purmorphamine (cat. no. 540220, Calbiochem). MNP medium was changed every other day for 6 days. MNPs were dissociated with 1:2 diluted Accutase and cultured in suspension in motor neuron differentiation medium, neural medium supplemented with 0.5 μM retinoic acid and 0.1 μM purmorphamine) in 60- mm low-attachment plates for 6 days. The media were changed every other day for 6 days. Lastly, neurospheres were dissociated into single cells with Accutase for 10 min at 37°C, passed through a 40 μm filter into motor neuron medium. The neural medium supplemented with 0.5 μM retinoic acid, 0.1 μM purmorphamine, 0.1 μM Compound E (cat. no. 73954, Stem Cell Technologies), 10 ng/mL BDNF (cat. no. 248-BDB, R&D Systems), 10 ng/mL GDNF (cat. no. PHC7041, Thermo Fisher Scientific), and 1 μg/mL laminin (cat. no. L2020, Sigma). Cells were seeded in Matrigel-coated 12-well plates at 6x10⁵ cells/well density in motor neuron medium. For immunofluorescence staining experiments, cells were seeded onto poly-d-lysine/laminin-coated coverslips (cat. no. 354087, Corning) in 24-well plates in motor neuron medium at 3-6x10⁴ cells/well. After two weeks, neurons were maintained in motor neuron medium without retinoic acid or purmorphamine for up to 3 months; the medium was half-changed once a week for 3 weeks, and twice a week thereafter.

Drosophila strains and genetics

Flies were maintained at 25°C on standard food. UAS-(GR)₈₀ line was generated previously in our laboratory.¹⁶ UAS-GFP, w¹¹¹⁸ (BL3605), GMR-Gal4 (BL9146), Tub-Gal80^ts (BL7019), UAS-Sec5 RNAi-1 (BL27526), UAS-Sec5 RNAi-2 (BL50556) lines were from the Bloomington Drosophila Stock Center.

METHOD DETAILS

Gene editing in iPSC lines

CRISPR/Cas9 technology was used to create a deletion in EXOC2 as described^32,33 with some modifications. Two sgRNAs targeting EXOC2 were designed with CHOPCHOP in CRISPR/Cas9 knockout mode (https://chopchop.cbu.uib.no/, see Table S1). sgRNAs were cloned into pCAG-eCas9-GFP-U6-gRNA plasmid (Key Resources Table) and nucleofected into iPSCs. Cells with high GFP expression were FACS sorted and seeded at very low density (50,000 cells/plate) in Matrigel-coated 100-mm dishes. Single colonies were picked and seeded in Matrigel-coated 24-well plates and passaged for genotyping and maintenance. DNA was extracted with the DNeasy Blood & Tissue kit (cat. no. 69506, Qiagen) following the manufacturer’s instructions. Colonies were screened by PCR (see primers in Table S1) and agarose gel electrophoresis. The deletions were confirmed in selected colonies by Sanger sequencing.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-EXOC2 [EPR9420]	Abcam	Cat#ab140620
Mouse anti-β actin	Sigma	Cat#A2228
Mouse anti-MAP2	Sigma	Cat#M9942
Mouse anti-βIII Tubulin	Promega	Cat#G7121
Rabbit anti-NMDAR 2B	Alomone Labs	Cat#AGC-003
Rabbit anti-Cleaved Caspase 3 (Asp175)	Cell Signaling Technology	Cat#9661
Rabbit anti-SQSTM1/p62	Abcam	Cat#91526
Mouse anti-SQSTM1/p62	Sigma	Cat#WH0008878M1
Mouse anti-α Tubulin	Sigma	Cat#T6199
Rabbit anti-(GR)8	Covance	Custom
Anti-(GR)8 GOLD SULFO	MSD	Cat#R31AA
Critical commercial assays
DNeasy Blood&Tissue Kit	Qiagen	69506
RNeasy Mini Kit	Qiagen	74106
TaqMan Reverse Transcription Kit	Thermo Fisher Scientific	N8080234
SYBR Select Master Mix	Thermo Fisher Scientific	4472918
PNFH ELISA Kit	EnCor Biotechnology	ELISA-pNfH-V2
Experimental models: Cell lines
Human: EXOC2−/− C9ORF72 iPSC line 1 (EXOC2 del-1)	This paper	N/A
Human: EXOC2−/− C9ORF72 iPSC line 2 (EXOC2 del-2)	This paper	N/A
Human: EXOC2−/− C9ORF72 iPSC line 3 (EXOC2 del-3)	This paper	N/A
Human: Isogenic control iPSC line 26Z90 (Control 1)	Lopez-Gonzalez et al.21	N/A
Human: Isogenic control iPSC line 27M91 (Control 2)	Lopez-Gonzalez et al.21	N/A
Human: Control iPSC line 2L20 (Control 3)	Almeida et al.37	N/A
Human: Control iPSC line 37L20 (Control 4)	Zhang et al.55	N/A
Human: C9ORF72 carrier iPSC line 26L6 (Patient 1)	Almeida et al.37	N/A
Human: C9ORF72 carrier iPSC line 27L11 (Patient 2)	Almeida et al.37	N/A
Human: C9ORF72 carrier iPSC line 16L14 (Patient 3)	Lopez-Gonzalez et al.17	N/A
Human: C9ORF72 carrier iPSC line 40L3 (Patient 4)	Freibaum et al.56	N/A
Human: C9ORF72 carrier iPSC line 29ALS	Sareen et al.40	N/A
Experimental models: Organisms/strains
D. melanogaster: UAS-(GR)80	Yang et al.16	N/A
D. melanogaster: UAS-GFP	Yang et al.16	N/A
D. melanogaster: w[1118]	Bloomington Drosophila Stock Center	BDSC:3605
D. melanogaster: w[1118]; P{GMR-GAL4.w[-]}2/CyO	Bloomington Drosophila Stock Center	BDSC:9146
D. melanogaster: w[*]; P{w[+mC] = tubP-GAL80[ts]}20;	Bloomington Drosophila	BDSC:7019
TM2/TM6B, Tb[1]	Stock Center
D. melanogaster: y[1] v[1]; P{y[+t7.7]	Bloomington Drosophila	BDSC:27526
v[+t1.8] = TRiP.JF02676}attP2	Stock Center
D. melanogaster: y[1] v[1]; P{y[+t7.7]	Bloomington Drosophila	BDSC:50556
v[+t1.8] = TRiP.GLC01676}attP2	Stock Center
Oligonucleotides
EXOC2 sgRNAs, see Table S1	This paper	N/A
EXOC2 PCR primers, see Table S1	This paper	N/A
qRT-PCR primers, see Table S1	Yuva-Aydemir et al.31	N/A
C9ORF72-V3 pyrosequencing primers, See Table S1	Yuva-Aydemir et al.31	N/A
Recombinant DNA
pCAG-eCas9-GFP-U6-gRNA	Deposited by Jizhong Zou	Addgene plasmid # 79145
pCMV delta R8.2	Deposited by Didier Trono	Addgene plasmid # 12263
pCMV-VSV-G	Stewart et al.57	Addgene plasmid # 8454
pLenti CMV GFP Puro (658-5)	Campeau et al.58	Addgene plasmid # 17448
Software and algorithms
ImageJ	NIH	https://imagej.nih.gov/ij/index.html
GraphPad	Prism	https://www.graphpad.com/features
Zen	Zeiss	N/A
ImageStudio	Li-Cor	https://www.licor.com/bio/image-studio-lite/

RNA extraction and real-time quantitative PCR

Total RNA from iPSC-MNs was extracted with the RNeasy Mini kit (cat. no. 74106, Qiagen). RNA (0.25–1 μg) was reverse transcribed to cDNA with random hexamers and the TaqMan Reverse Transcription kit (cat. no. N8080234, Thermo Fisher Scientific). Real-time quantitative PCR was done with SYBR Select Master Mix (cat. no. 4472918, Thermo Fisher Scientific) and an Applied Biosystems Quant Studio 3 System. Ct values for each sample were normalized to cyclophilin. Relative mRNA levels of each gene were calculated with the 2^−ΔΔCt method. (See Table S1 for the qRT-PCR primers).

Pyrosequencing for C9ORF72 allele specific expression

Total RNA from C9ORF72-26 and EXOC2 del iPSC-MNs was used to generate cDNA with OneStep RT-PCR Master Mix (cat. no. 978801, Qiagen) and previously described V3-specific primers.³¹ Pyrosequencing for allele-specific quantification and data analysis were done by EpigenDx Inc. (Hopkinton, MA) in a blinded manner, using a previously described sequencing primer³¹ (Table S1).

Western blotting

iPSC-MNs were washed with PBS and lysed in RIPA buffer (cat. no. 89901, Thermo Fisher Scientific) containing EDTA-free Halt Protease and Phosphatase Inhibitor Cocktail (100X) (cat. no. 78441, Thermo Fisher Scientific). Lysates were incubated on ice for 15 min, sonicated on ice at a 50% pulse rate for 10 s and centrifuged at 16,000 x g for 20 min at 4°C. The supernatant was collected, and protein concentration was determined with the Pierce BCA protein assay (cat. no. 23227, Thermo Fisher Scientific). Protein (20 μg) was mixed with 2X Laemmli Sample Buffer (cat. no. 1610737, Bio-Rad) containing β-mercaptoethanol (cat. no. M3148, Sigma) and boiled at 95°C for 10 min. Protein (60 μg) was loaded for detection of Cleaved caspase 3 and truncated fragments of EXOC2. Samples were run on SDS-PAGE gel, transferred to PVDF membranes, incubated with the primary antibodies listed in Key Resources Table at 4°C overnight. After washing three times with TBS-T, the membranes were incubated with secondary antibodies IRDye 680 goat anti-rabbit (cat. no. 926–68071, LI-COR Biosciences) and/or IRDye 800CW donkey anti-mouse (cat. no. 926–32212, LI-COR Biosciences) for 1 h at room temperature (RT). Odyssey Blocking Buffer (cat. no. 927–40010, LI-COR Biosciences) was used for blocking and antibody dilutions. The images were acquired with the Odyssey Imaging System (LI-COR Biosciences).

Dot blot

iPSC-MN lysates were prepared as described above. Duplicate samples (20 μg protein) were loaded onto the Bio-Dot Microfiltration System dot blot module (cat. no. 1703938, Bio-Rad) transferred to a nitrocellulose membrane in vacuum mode following manufacturer’s instructions and stained with Ponceau S for detection of total protein. Samples were incubated with primary and secondary antibodies as described for western blotting.

Immunofluorescence staining

iPSC-MNs were fixed in 4% paraformaldehyde for 10 min at RT and permeabilized with 0.2% Triton X-100 for 5 min (except for surface NMDAR staining experiments), blocked with 3% BSA for 30 min at RT, and incubated with primary antibodies (Key Resources Table) for 1 h at RT. After three washes with PBS, cells were incubated with donkey anti-mouse IgG Alexa Fluor 488 (cat. no. A-21202, Thermo Fisher Scientific) and/or Donkey anti-Rabbit IgG Alexa Fluor 568 (cat. no. A-10042, Thermo Fisher Scientific; 1:500) for 1 h at RT. Coverslips were mounted with Prolong Glass Antifade Mountant with NucBlue Stain (cat. no. P36981, Thermo Fisher Scientific). Images were acquired with Zeiss LSM 800 microscope equipped with an AxioCam 506 camera.

Karyotyping analysis

Two 25cm flasks (50% confluent) per iPSC line were shipped to KaryoLogic Inc. (Durham, NC). 10% FBS was added to the media to reduce detachment.

Lentivirus production

Lentiviral eGFP expression vector was purchased from Addgene (Plasmid#17448) and GFP-expressing lentivirus was generated in HEK cells as described.⁵⁴

ASO treatment in iPSC-MNs

Control ASO (nontargeting control) and two EXOC2 ASOs were designed by and purchased from IDT Custom Design Service. The ASOs were 20 nt long and contained the following modifications: phosphorothioate linkage in backbone, 5-10-5 gapmer structure with 2′-O-methoxy-ethyl bases (2′-MOE). ASOs (5 or 10 μM) were directly added to culture medium for treatment of iPSC-MNs.

MSD assay

Poly(GR) from iPSC-MNs was measured with an MSD immunoassay as described.^31,39 iPSC-MN lysates were prepared in RIPA buffer as described in Western blotting Section. Samples were adjusted to an amount of 50 μg/well with Tris-buffered saline (TBS) and tested in duplicate wells on 96-well plate (cat. no. L45XA, MSD) precoated with a custom-made polyclonal rabbit anti-(GR)₈ antibody (1 μg/mL; Covance). Anti-(GR)₈ antibody tagged with GOLD SULFO (0.5 μg/mL; cat. no. R31AA, Meso SCale Discovery) was used as detection antibody. Signals were detected using QuickPlex SQ120 Instrument (Meso Scale Discovery). For background correction, values from samples were subtracted those of corresponding isogenic control iPSC-MNs. Similarly, poly(GA) was quantified as reported before.³⁹

Detection of phosphorylated neurofilament heavy chain (pNFH)

Culture medium from iPSC-MNs was collected 40 h after withdrawal of neurotrophic factor. Samples were briefly centrifuged to remove debris and supplemented with protease and phosphatase inhibitor. The pNFH signal was measured in medium diluted 1:2.5 with the pNFH ELISA kit (cat. no. ELISA-pNF-H-V2, EnCor Biotechnology) following the manufacturer’s instructions.

QUANTIFICATION AND STATISTICAL ANALYSIS

Neurite degeneration index

iPSC-MNs stained with Tuj1 antibody were imaged with a 20x objective as described above. The ratio of degenerated neurites to total area (degeneration index) was calculated as described.^31,53 Briefly, images were binarized and the particle analyzer of ImageJ was used to quantify 20–10,000 pixels or degenerated neurites and 20–infinity pixels for total area.

Neurite quantification

iPSC-MNs stained for MAP2 were imaged with a 20x objective as described above. Neurites were traced with the NeuronJ plugin in ImageJ to quantify the number of primary dendrites (defined as those originating from the soma). Average neurite length per neuron was also quantified using NeuronJ, limiting measurement to primary dendrites from soma to the first branchpoint.

Quantification of Drosophila external eye phenotype

The external eye phenotype of control and (GR)₈₀-expressing flies was characterized as described.⁵⁰ Briefly, flies were raised at 23°C for 14 days. Adult flies were immobilized and examined under a dissecting microscope. The phenotype was categorized as absent, weak, or medium based on ommatidia structure, pigmentation, and level of necrosis.

Statistical analysis

Values are reported as mean ± S.E.M. GraphPad Prism software (version 9.5.1) was used for statistical analysis. Unpaired two-tailed t-tests were used to compare two groups. One-way ANOVA with Dunnett’s or Tukey’s multiple comparisons tests were used to compare three or more groups.

Highlights.

• The N-terminal fragment of EXOC2 is sufficient to maintain cell viability

• Deletion of EXOC2 in C9ORF72-ALS/FTD motor neurons rescues disease phenotypes

• EXOC2 ASOs reduce axon degeneration and apoptosis in C9ORF72 motor neurons

• Loss of EXOC2 decreases the levels of DPR proteins and expanded G₄C₂ repeats RNA