Modulating the endoplasmic reticulum stress response attenuates neurodegeneration in a Caenorhabditiselegans model of spinal muscular atrophy

Abstract

Spinal muscular atrophy (SMA) is a devastating autosomal recessive neuromuscular disease resulting in muscle atrophy and neurodegeneration, and is the leading genetic cause of infant death. SMA arises when there are homozygous deletion mutations in the human SMN1 gene, leading to a decrease in corresponding SMN1 protein. Although SMN1 is expressed across multiple tissue types, much of the previous research into SMA focused on the neuronal aspect of the disease, overlooking many of the potential non-neuronal aspects of the disease. Therefore, we sought to address this gap in knowledge by modeling SMA in the nematode Caenorhabditis elegans We mutated a previously uncharacterized allele, which resulted in the onset of mild SMA-like phenotypes, allowing us to monitor the onset of phenotypes at different stages. We observed that these mutant animals recapitulated many key features of the human disease, and most importantly, we observed that muscle dysfunction preceded neurodegeneration. Furthermore, we tested the therapeutic efficacy of targeting endoplasmic reticulum (ER) stress in non-neuronal cells and found it to be more effective than targeting ER stress in neuronal cells. We also found that the most potent therapeutic potential came from a combination of ER- and neuromuscular junction-targeted drugs. Together, our results suggest an important non-neuronal component of SMA pathology and highlight new considerations for therapeutic intervention.

Keywords: Caenorhabditis elegans; ER stress; Genetics; Muscle pathology; Spinal muscular atrophy.

Fig. 1.

A point mutation in smn-1 recapitulates key neuronal features of SMA. (A) Smn-1(gk) mutant animals display a reduction in lifespan (P<0.0001). (B) Smn-1(gk) mutants exhibit an increase in age-dependent paralysis compared to wild-type N2 animals (P<0.0001). (C) smn-1(gk)/⁺ heterozygous animals display lower paralysis levels compared to smn-1(gk)/ smn-1(gk) homozygous mutants, but higher paralysis than N2 animals (Mantel–Cox test, P<0.0001 and P<0.001, respectively). (D) The smn-1(gk) mutation does not significantly affect smn-1 mRNA expression levels. (E) Representative images of GABAergic neurodegeneration observed in day 9 adult wild-type and smn-1(gk) mutant animals. Black arrowheads indicate neuronal gaps indicative of neurodegeneration. (F) Smn-1(gk) mutant animals display GABAergic motor neuron degeneration at day 9 (P<0.001, n=100 animals per condition), but do not show any degeneration of cholinergic motor neurons (n=100 animals per condition). (G) Non-neuronal RNAi treatment of smn-1(gk) animals with smn-1 RNAi exacerbates paralysis (P<0.0001). (H) Non-neuronal smn-1 RNAi treatment does not increase GABAergic motor neuron degeneration in smn-1(gk) mutants (n=100 animals per condition). (I) Swimming defects in smn-1(gk) animals are observed at day 5 of adulthood (P<0.0001). (J) Smn-1(gk) mutant animals do not display degeneration of GABAergic motor neurons at days 3 or 5 (not significant, n=100 animals per condition). (K) At day 1 of adulthood, smn-1(gk) mutants display hypersensitivity to aldicarb compared to N2 animals, but not to the same extent as unc-47(e307) animals (P<0.0001 and P<0.0001, respectively). (L) Smn-1(syb1923) animals show slightly higher levels of paralysis than smn-1(gk) animals (P<0.01). (M) Swimming defects in smn-1(syb1923) are nearly identical to those in smn-1(gk) mutants. (N) Smn-1(syb1923) animals display GABAergic neuron degeneration (P<0.01, n=100 animals per condition). Data are mean±s.d. Statistical significance was determined using a Mantel–Cox test (A-C,G,K,L), two-tailed unpaired Student's t-test (D,F,H,J), or two-way ANOVA (I,M).

Fig. 2.

smn-1(gk) mutants show signs of early muscle defects. (A) Day 1 adult smn-1(gk) mutant animals are hypersensitive to the nAChR agonist levamisole (P<0.0001). (B) Wild-type N2 animals become paralyzed when they are fed RNAi against smn-1 (P<0.0001). (C) Transgenic worms sensitive to RNAi only in their body-wall muscle cells became paralyzed when they were fed RNAi against smn-1 (P<0.0001). (D) RNAi knockdown of smn-1 in intestinal cells did not result in a paralysis phenotype. (E) Representative images of transgenic animals expressing GFP::MYO-3 in their body-wall muscle cells. Levels of morphological distortion were characterized as ‘low’, ‘medium’, or ‘high’ depending on the extent of GFP::MYO-3 disorganization in the cells. (F,G) At days 3 and 5 of adulthood, greater numbers of smn-1(gk) mutants displayed ‘high’ levels of muscle dysfunction (P<0.0001, n=100 animals per condition). (H) Smn-1(syb1923) CRISPR mutants displayed increased muscle dysfunction compared to wild-type animals (P<0.0001). (I) Representative images of transgenic animals expressing TOM20::mRFP in body-wall muscle cells. Mitochondrial organization was quantified as either ‘linear’, ‘intermediate’, or ‘fragmented’. Arrowheads indicate muscle defects. (J,K) At days 1 and 5, smn-1(gk) mutants had higher levels of mitochondria impairment compared to wild-type animals (P<0.0001, n=100 animals per condition). Data are mean±s.d. Statistical significance was determined using a Mantel–Cox test (A-D) or two-way ANOVA (F-H,J,K).

Fig. 3.

Targeting ER stress chemically and genetically is protective in smn-1 mutants, and is more effective when targeted to non-neuronal cells. (A) Treatment of smn-1(gk) mutant animals with either 50 µM guanabenz or salubrinal attenuates the paralysis phenotype (P<0.0001 and P<0.01, respectively). (B) Pharmacological inhibition of ER stress with either 50 µM guanabenz or salubrinal prevents neurodegeneration in GABAergic motor neurons in smn-1(gk) mutants (P<0.05 and P<0.01, respectively; n=100 animals per condition). (C) Salubrinal and guanabenz treatment partially restores muscle morphology in smn-1(gk) mutants (P<0.0001 and P<0.0001, respectively). (D) Treating smn-1(syb1923) animals with either 50 µM guanabenz or salubrinal ameliorates their paralysis phenotype (P<0.0001). (E) Salubrinal and guanabenz treatments restore neurodegeneration at day 9 in smn-1(syb1923) animals (P<0.001; n=100 animals per condition). (F) Salubrinal and guanabenz treatments partially restore muscle dysfunction of smn-1(syb1923) mutants (P<0.01 and P<0.05, respectively). (G) Non-neuronal RNAi against ire-1, atf-6 and pek-1 restores paralysis in smn-1(gk) animals compared to EV treatment (P<0.0001). (H) Treatment of smn-1(gk) animals with non-neuronal ire-1, atf-6 and pek-1 RNAi also protects against neurodegeneration (P<0.0001, P<0.001 and P<0.01, respectively; n=100 animals per condition). (I) Treating smn-1(gk) mutants with non-neuronal ire-1, atf-6 and pek-1 RNAi helps restore muscle defects (P<0.0001, n=100 animals per condition). (J) Non-neuronal ire-1, atf-6 and pek-1 RNAi treatment from day 5 of adulthood is still capable of partially attenuating paralysis phenotypes in smn-1(gk) mutants (P<0.01, P<0.001 and P<0.01, respectively). (K) Neuronal ire-1, atf-6 and pek-1 RNAi treatment had no effect on smn-1(gk) paralysis compared to EV treatment. (L) Neuronal knockdown of ire-1, atf-6 pek-1 in smn-1(gk) animals does not significantly rescue neurodegeneration (n=100 animals per condition). Levels of morphological distortion were characterized as ‘low', ‘medium' or ‘high’ depending on the extent of GFP::MYO-3 disorganization in the cells. Data are mean±s.d. Statistical significance was determined using a Mantel–Cox test (A,D,G,J,K), a two-tailed unpaired Student's t-test (B,E,H,L) and two-way ANOVA (C,F,I).

Fig. 4.

Simultaneously targeting ER stress and NMJ stabilization has an additive effect on smn-1(gk) mutants. (A) Treatment of smn-1(gk) mutant animals with 40 µM pimozide significantly improved the swimming behavior of the animals (P<0.0001), whereas 40 µM riluzole treatment significantly reduced it (P<0.0001). (B) Treatment with 40 µM pimozide or riluzole was able to significantly reduce paralysis in smn-1(gk) animals (P<0.001 and P<0.0001, respectively). (C) GABAergic motor neuron degeneration was significantly reduced in smn-1(gk) mutant animals upon pimozide or riluzole treatment (P<0.01 and P<0.001, respectively). (D) Muscle dysfunction was partially restored in smn-1(gk) animals upon treatment with pimozide (P<0.0001), but riluzole had no effect. (E) Treatment of smn-1(gk) animals simultaneously with 20 µM pimozide and 25 µM salubrinal effectively restored paralysis (P<0.0001). (F) Simultaneous treatment with both salubrinal and pimozide greatly restored muscle defects in smn-1(gk) mutants (P<0.0001). (G) Individually, treatment of smn-1(gk) mutants with salubrinal and pimozide from day 5 had no effect on paralysis. (H) Treatment of smn-1(gk) with both pimozide and salubrinal from day 5 resulted in a slight decrease of paralysis levels (P<0.01). Levels of morphological distortion were characterized as ‘low', ‘medium' or ‘high’ depending on the extent of GFP::MYO-3 disorganization in the cells. Data are mean±s.d. Statistical significance was determined using two-way ANOVA (A,D,F), a Mantel–Cox test (B,E,G) and a two-tailed unpaired Student's t-test (C).