A combinatorial approach increases SMN level in SMA model mice

Abstract

Spinal muscular atrophy (SMA) is a neurodegenerative disease caused by reduced expression of the survival motor neuron (SMN) protein. Current disease-modifying therapies increase SMN levels and dramatically improve survival and motor function of SMA patients. Nevertheless, current treatments are not cures and autopsy data suggest that SMN induction is variable. Our group and others have shown that combinatorial approaches that target different modalities can improve outcomes in rodent models of SMA. Here we explore if slowing SMN protein degradation and correcting SMN splicing defects could synergistically increase SMN production and improve the SMA phenotype in model mice. We show that co-administering ML372, which inhibits SMN ubiquitination, with an SMN-modifying antisense oligonucleotide (ASO) increases SMN production in SMA cells and model mice. In addition, we observed improved spinal cord, neuromuscular junction and muscle pathology when ML372 and the ASO were administered in combination. Importantly, the combinatorial approach resulted in increased motor function and extended survival of SMA mice. Our results demonstrate that a combination of treatment modalities synergistically increases SMN levels and improves pathophysiology of SMA model mice over individual treatment.

Introduction

Spinal muscular atrophy (SMA) is a neurodegenerative disease characterized by the loss of motor neurons and atrophy of affected skeletal muscles (1). SMA is caused by a homozygous deletion or mutations of the survival motor neuron (SMN1) gene leading to functional insufficiency of the SMN protein (2–4). Complete loss of SMN protein is lethal; however, the SMN2 gene provides an additional source of SMN (5). The SMN2 gene largely produces a truncated splice isoform of SMN, referred to as SMN△7, because of the exclusion of exon 7 (6,7). SMN△7 is less stable than full length SMN (SMN-FL) and does not fully compensate for the loss of SMN1 (6,8–10). The severity of SMA in patients directly correlates with the amount of functional SMN protein that is present and is proportional to the number of copies of SMN2 that patient carries (11,12).

Substantial clinical advances have been made over the last few years culminating in the approval of three treatments for SMA (13–16). A gene therapy, onasemnogene abeparvovec-xioi, utilizes an AAV9 vector to introduce a working copy of the SMN1 gene. Nusinersen is an antisense oligonucleotide (ASO) that acts as a splicing modifier to promote the incorporation of exon 7 leading to more full-length SMN from the SMN2 gene. Risdiplam is a small molecule splicing modifier that, similar to nusinersen, acts to redirect exon 7 splicing to promote more full-length protein. These strategies have been milestones in the treatment of SMA patients as well as in the rare disease community, drastically increasing lifespan of severely affected children and improving motor and respiratory function. However, significant deficits remain for patients and additional treatment options still need to be developed to address this complex disorder.

We have previously shown that the small molecule ML372 selectively inhibits ubiquitination and slows the degradation of SMN (17). ML372 inhibits mindbomb-1 (Mib1), one of the E3 ligases responsible for SMN ubiquitination. ML372 improves functional outcomes in SMA model mice including motor function, endplate innervation, righting time and lifespan.

Although the approval of three SMN-specific therapeutics has profoundly changed the SMA landscape, there is a clear need for additional compounds or combinatorial strategies to address the breadth of the patient population and the complex pathology presented in many SMA cases that do not respond well to current therapies. Emerging evidence suggest that SMN induction is variable in patients. Here we investigated if ML372 acts synergistically with an ASO splicing modifier to improve the motor phenotype and lifespan of SMA model mice over either treatment alone. Our findings demonstrate that combination therapies may provide significant clinical improvement not reached by individual treatment modalities.

Results

ML372 and ASO increases SMN levels in cultured cells

Mechanisms to boost SMN expression are currently under investigation to complement current SMA therapies. We and others have shown in cell culture and mice that SMN is degraded by the ubiquitin proteasome system and targeting this pathway improves the SMA phenotype (8,17–20). We identified and provided evidence that the Mib1 inhibitor, ML372 (NCATS-SM4980), stabilizes the SMN protein by inhibiting its ubiquitination and subsequent degradation (17). Two of the three available SMA treatments increase SMN levels by promoting the inclusion of exon 7 of the SMN2 gene product. We first sought to determine if co-treating SMA patient cells with an ASO that targets SMN splicing (21,22) and ML372 that stabilizes the SMN protein could synergistically increase SMN protein levels. SMA patient-derived 3813 fibroblasts were co-administered 1 μM ML372 and 100 nM E1 ASO for 48 h and SMN protein levels assessed by western blot analysis. ML372 and ASO E1 increased SMN levels by 2-fold and 1.5-fold over vehicle treatment (Fig. 1A). Importantly, co-treatment increased SMN levels over individual treatments alone, supporting our hypothesis that a combinatorial approach targeting mechanistically discrete molecular pathways could synergize to increase SMN levels.

Figure 1

ML372 and ASO increases SMN protein levels. (A) SMA patient fibroblast 3813 cells were treated with combination ML372 and ASO for 48 h. Western blot was used to assess levels of SMN protein in cell protein lysates. Densitometry analysis is shown as the mean ± standard error of mean (SEM) (n = 6). A one-way ANOVA was performed to compare the effects of treatment conditions on SMN levels [F(3,24) = 26.20, P < 0.0001] followed by post hoc Tukey’s test for multiple comparisons (vehicle versus ASO: P < 0.0001; vehicle versus ASO + ML372: P < 0.0001; ASO versus ML372: P = 0.0431; ASO versus ASO + ML372: P = 0.0285; ML372 versus ASO + ML372: P < 0.0001). (B) SMN△7 SMA mice were treated daily with combination ML372 and ASO starting on PND1 and sacrificed on PND13. Spinal cord tissue was collected from mice treated with combination ML372 and ASO. SMN levels were measured using western blot and densitometry analysis is shown as the mean ± SEM (n = 5). A one-way ANOVA was performed to compare the effect of treatment conditions on SMN levels [F(3,16) = 56.02, P < 0.0001] followed by post hoc Tukey’s test for multiple comparisons (Vehicle versus ML372: P < 0.05; Vehicle versus ASO: P < 0.0001; Vehicle versus ASO + ML372: P < 0.0001; ML372 versus ASO: P < 0.05; ML372 versus ASO + ML372: P < 0.0001; ASO versus ASO + ML372: P < 0.0001). (C) Tibialis anterior muscles were collected from mice treated with ML372, ASO alone or in combination. SMN levels were measured using western blot and densitometry analysis is shown as the mean ± SEM (n = 5). A one-way ANOVA was performed to compare the effect of treatment conditions on SMN levels [F(3,16) = 54.86, P < 0.0001] followed by post hoc Tukey’s test for multiple comparisons (vehicle versus ML372: P < 0.05; vehicle versus ASO: P < 0.0001; vehicle versus ASO + ML372: P < 0.0001; ML372 versus ASO: P < 0.05; ML372 versus ASO + ML372: P < 0.0001; ASO versus ASO + ML372: P < 0.0001).

Co-administering ML372 and SMN ASO increases SMN levels in mice

To assess the efficacy of the combination treatment in vivo we used the SMN△7 (Smn^−/−, SMN2^+/+, SMN△7^+/+) mouse model. Affected SMN△7 mice display significantly reduced SMN expression in all tissues, loss of motor function, reduced body weight and a lifespan of 14 days (9). We treated affected mice and unaffected littermates by a single intracerebroventricular (ICV) injection 2 nmol MOE^1v1.11 of ASO and intraperitoneal (IP) injections of 30 mg/kgML372 twice a day beginning at postnatal day 1 (PND1). The ASO dose was chosen because it was previously shown to not fully rescue the SMA phenotype and allowed us to determine if ML372 could boost SMN levels when splicing correction is suboptimal (23). Mice were sacrificed at PND13 and spinal cord and muscle tissues were collected. SMN levels were higher in spinal cord and muscle extracts from mice administered the ASO or ML372. Strikingly, we found that co-treated mice had significantly higher SMN levels compared with the single treated groups (Fig. 1B and C). Together these data provide strong evidence that SMN protein expression can be further increased by targeting complementary pathways.

Co-administering ML372 and SMN ASO improves neuropathology of SMA mice

Motor neuron loss is a key feature of SMA disease pathology and interventions that slow motor neuron degeneration and/or preserve the health of surviving motor neurons have been effective in improving the SMA phenotype in mice. We treated SMN△7 mice with a single ICV injection of ASO, daily IP delivery of ML372 or a combination of both treatment and lumbar spinal cord sections (L3–L5) were obtained from mice sacrificed at PND12. We analyzed the diameter and number of neurons >25 μm in the ventral horn of the lumbar spinal cord. The average diameter of ventral horn neurons was 43 μm in unaffected littermates, 30 μm in SMA vehicle-treated, 37 μm in SMA ML372-treated, 38 μm in SMA ASO-treated and 43 μm in ASO + ML372-treated mice. The average number of ventral horn neurons per section in unaffected mice was 31, compared with 21 in SMA vehicle-treated, 23 in SMA ML372-treated mice, 26 in SMA ASO-treated mice and 30 in SMA ASO + ML372-treated mice (Fig. 2A–C). Together these data indicate that administering ML372 in combination with a suboptimal dose of ASO almost completely rescued motor neuron size and number.

Co-administering ML372 and SMN ASO improves muscle pathology in SMA mice

Loss of ventral horn neurons result in denervation-dependent atrophy in affected muscle in SMN△7 mice (9,24). Near disease end-stage (PND12) muscle fibers have fewer fibers and smaller fiber diameter (25–27). We thus examined if individual treatments improved SMA muscle pathology and if the combination treatment further improved the treatment outcome. Consistent with previous studies we saw improved myofiber numbers in SMA mice treated with ASO and ML372 (Fig. 3A and B). We observed further improvement in myofiber number in mice co-treated with both ASO and ML372 compared with individual treatment, comparable with myofiber numbers of wild-type (WT) mice (Fig. 3A and B). The total cross-sectional area and perimeter of affected abdominal muscle fibers [transverse abdominal muscle (TVA) and rectus abdominis (RA)] were also improved by co-treatment (Fig. 3C and D), confirming an improvement in overall muscle size. We found no evidence of fibrosis or inflammatory infiltrate, and very rare fibers showed centralized nuclei, the hallmark feature of regenerating myofibers. These data suggest that the SMA muscle pathology was improved by slowing denervation-dependent atrophy rather than promoting muscle regeneration.

Figure 2

Co-administering ML372 and SMN ASO improves neuropathology in SMA mice. (A) Representative images of Nissl-stained lumbar spinal cord sections from unaffected mice, untreated SMA mice, ML372, ASO and co-treated SMA mice. (×20; scale bar: 100 μm; arrows indicate representative motor neurons). (B) Mean number of ventral horn neuron number was reduced in SMA mice compared with heterozygous littermates (P < 0.0001) and was not changed by ML372 treatment. ASO treatment alone increased the number of ventral horn neurons (P < 0.05). Co-treatment resulted in increased motor neuron number compared with ASO alone (P < 0.001). (C) Mean neuron diameter was reduced in SMA mice compared with heterozygous littermates (P < 0.0001) and was improved by ML372 treatment (P < 0.001) and ASO treatment alone (P < 0.001). Co-treatment increased motor neuron diameter compared with ASO alone (P < 0.01).

Figure 3

Co-administering ML372 and SMN ASO improves muscle pathology in SMA mice. (A) Representative images of hematoxylin and eosin-stained RA from unaffected mice, untreated SMA mice and ML372 and ASO co-treated SMA mice. (×20; scale bar: 100 μm) (B) Myofiber units were counted to determine the effect of each treatment condition. The analysis is shown as the mean ± SEM (n = 20) A one-way ANOVA was performed to compare the effect of treatment conditions on the number of myofiber units [F(4,20) = 41.60, P < 0.0001] followed by post hoc Tukey’s test for multiple comparisons (WT versus SMA untreated: P < 0.0001; WT versus ASO: P = 0.0059; WT versus ML372: P = 0.0003; SMA untreated versus ASO: P < 0.0001; SMA untreated versus ML372; P = 0.0006; SMA untreated versus ASO + ML372: P < 0.0001; ASO versus ASO + ML372: P = 0.0006; ML372 versus ASO + ML372: P < 0.0001). (C) Muscle fiber area was measured in images from each of the treatment groups. The spread of individual data points is plotted. A one-way ANOVA was performed to compare the effect of treatment conditions on muscle fiber area [F(4,4410) = 411.6, P < 0.0001] followed by Tukey’s test for multiple comparisons (WT versus SMA untreated: P < 0.0001; WT versus ASO: P < 0.0001; WT versus ML372: P < 0.0001; WT versus ASO + ML372: P < 0.0001; SMA untreated versus ASO: P = 0.0454; SMA untreated versus ASO + ML372: P < 0.0001; ASO versus ASO + ML372: P < 0.0001; ML372 versus ASO + ML372: P < 0.0001). (D) Muscle fiber perimeter was measured in images from each of the treatment groups. Individual data points are plotted. A one-way ANOVA was performed to compare the effect of treatment conditions on muscle fiber perimeter [F(4,4419) = 194.2, P < 0.0001] followed by Tukey’s test for multiple comparisons (WT versus SMA untreated: P < 0.0001; WT versus ASO: P < 0.0001, WT versus ML372: P < 0.0001; WT versus ASO + ML372: P < 0.0001; SMA untreated versus ASO + ML372: P < 0.0001; ASO versus ML372: P = 0.0068; ASO versus ASO + ML372: P < 0.0001; ML372 versus ASO + ML372: P < 0.0001).

Co-administering ML372 and SMN ASO improves motor endplate pathology in SMA mice

One of the pathological features of SMA is compromised motor endplates in vulnerable muscle (24,28). In affected mice we observe distinct neuromuscular junction (NMJ) denervation in TVA and RA muscle (Fig. 4A). To assess the effect of our treatment we mapped the presynaptic motor neuron endings onto the postsynaptic endplate using neurofilament (NF) and synaptophysin. This method allows the innervation of the junction to be visualized and the degree of innervation to be assessed. Mice treated with ML372 or ASO increased the percent of fully innervated endplates from partially innervated (Fig. 4B and C) and reduced the percent of fully denervated endplates in the RA muscles (Fig. 4E) and trending toward a reduction in the TVA muscle (Fig. 4D). ASO and ML372 co-treated mice showed a more dramatic shift from partially to fully innervated endplates compared with untreated affected mice (Fig. 4B and C) with a reduction in the percent of fully denervated endplates to levels observed in unaffected littermates (Fig. 4D and E).

Figure 4

Co-administering ML372 and SMN ASO improves motor endplate pathology in SMA mice. (A) Representative images of NMJs isolated from abdominal muscles in treated SMA mice. NMJs are visualized in ASO and ML372 treated mice using NF + synaptophysin (green) for presynaptic staining and a-bungarotoxin (red) for postsynaptic staining. The percentage of total endplates was assessed by level of innervation defined as fully denervated, partially innervated or fully innervated. (B) A two-way ANOVA was performed to assess the effect of treatment condition on the percent of neurons either fully innervated, partially innervated or fully denervated in the TVA [F(8,135) = 2.891, P = 0.0053] followed by a post hoc Tukey’s test for multiple comparisons for fully innervated NMJs (WT versus SMA untreated: P = 0.0318), partially innervated NMJs (no significant differences found) and fully denervated NMJs (no significant differences found). (C) A two-way ANOVA was performed to assess the effect of treatment condition on the percent of neurons either fully innervated, partially innervated or fully denervated in the RA [F(8,132) = 13.54, P < 0.0001] followed by a post hoc Tukey’s test for multiple comparisons for fully innervated NMJs (WT versus SMA untreated: P < 0.0001; WT versus ASO: P < 0.0001; WT versus ML372: P = 0.0285; SMA untreated versus ML372: P = 0.0070; SMA untreated versus ASO + ML372: P = 0.0021; ASO versus ML372: P = 0.0037; ASO versus ASO + ML372: P = 0.0011), partially innervated NMJs (WT versus ASO: P = 0.0001; SMA untreated versus ASO: P = 0.0135; ASO vs ML372: P = 0.0349) and fully denervated NMJs (WT versus SMA untreated: P < 0.0001; SMA untreated versus ASO: P = 0.0307; SMA untreated versus ML372: P = 0.0014; SMA untreated versus ASO + ML372: P < 0.0001). (D) The percent of fully denervated NMJs in the TVA and (E) the RA are shown. Results are shown as mean ± SEM. Mice treated with combination ASO and ML372 have fewer NMJs that are fully denervated. A one-way ANOVA was performed to assess the effect of treatment conditions on the percent of fully denervated NMJs in the TVA [F(4,45) = 2.913, P = 0.0317] followed by post hoc Tukey’s test for multiple comparisons (SMA untreated versus ASO + ML372: P = 0.0176). A one-way ANOVA was performed to assess the effect of treatment conditions on the percent of fully denervated NMJs in the RA [F(4,44) = 13.79, P < 0.0001] followed by post hoc Tukey’s test for multiple comparisons (WT versus SMA untreated: P < 0.0001; SMA untreated versus ASO: P = 0.0023; SMA untreated versus ML372: P < 0.0001; SMA untreated versus ASO + ML372: P < 0.0001).

Co-administering ML372 and ASO co-treatment improves motor function and survival of SMA mice

Given the improvement in histopathology of the ventral horn neurons, skeletal muscles and NMJ, we next sought to determine if co-treatment improved motor function of SMN△7 mice. Mice were placed on their backs and the time for them to right themselves was recorded. Trials were stopped if mice were unable to right themselves after 30 s. We performed righting time assessment on mice treated with ML372 and ASO, alone or in combination starting at PND6. The co-treated animals showed a significant improvement over ML372 or ASO treated animals, closely approaching the righting times of unaffected WT mice by PND16 (Fig. 5A).

Weight gain strongly correlates with improved health and survival in SMA model mice. We next investigated if co-treating SMA mice with ML372 and an ASO would improve weight and survival of SMA mice. Consistent with previous findings we show that SMA mice gain weight more slowly compared with healthy littermates (Fig. 5B). Individual treatment with ASO and ML372 improved weight gain of SMA mice up until near disease end stage. Importantly, we found significant improvement in weights of mice co-treated with ASO and ML372 compared with individual treatment, with weights indistinguishable from WT littermates by PND70. The improved weight gain tracked well with extended lifespans. ML372-treated mice showed no improvement in survival, which was not surprising given we used almost half the effective dosage we reported previously. Notably, co-treated mice lived for 80 days on average compared with 30 days with the suboptimal dose of ASO, demonstrating that an SMN protein stabilizer can synergize with splice correcting treatments to further improve the lifespan of SMA mice (Fig. 5C).

Discussion

The approval of three SMA therapies has dramatically improved the survival, motor function and quality of life of SMA patients. Although these life changing developments are ground breaking, the current treatments are not cures and many treated patients remain severely debilitated (16). One potential strategy to improve treatment outcomes is to identify therapies with relevant synergistic interactions that may exert greater effects while limiting side effects and providing an opportunity to reduce the effective dosing. We have previously shown that inhibiting the proteasome and increasing SMN gene expression through complementary pathways that augment SMN expression improves lifespan and motor function of SMA model mice (19,29). Those proof-of-principle studies utilized compounds that were not SMN selective and thus could not be advanced for clinical consideration. Here we showed that a combination of ML372, which more selectively targets SMN ubiquitination, and an SMN splicing ASO can synergistically increase SMN protein levels and improve the disease phenotype in SMA mice.

Stabilizing SMN protein complements current SMA therapies

We recently completed a genome-wide RNAi screen to identify modifiers of SMN protein expression (30). Findings from that study strongly implicated the RNA processing and protein degradation pathways as major SMN modifying genes. Our group and others have investigated if SMN protein can be stabilized by targeting the degradation machineries. In cultured cells we have shown that SMN is degraded primarily through the ubiquitin proteasome system (8). In neurons there has been compelling evidence that SMN is degraded by lysosomes suggesting that the regulation of SMN turnover is likely tissue dependent (31). Two main questions have emerged from these studies: (i) what are the enzymes responsible for catalyzing SMN degradation? and (ii) can these enzymes be targeted to increase SMN in vivo? So far, several enzymes have been implicated in SMN turnover including the E3 ligases mib1, Neurl2, SCF^Slmb and UHCL1, the deubiquitinase Usp9x and glycogen synthase kinase 3 (GSK-3) (20,30,32–35). In follow-up studies, mib1 and GSK-3 inhibitors increased SMN levels and improved the SMA phenotype in SMA mouse models (17,34). Given that current approved strategies increase SMN gene expression, modalities that stabilize the SMN protein would likely increase the steady state levels of SMN with potential implications of treatment efficacy.

Where is SMN needed and how much?

How much SMN is required in each tissue remains controversial. Findings from rodent models suggest that most tissues require some level of SMN since tissue-specific SMN knockout results in malformations or malfunction. Restoring SMN in the central nervous system is necessary to significantly modify the disease course in rodent models. Nusinersen, which is delivered through intrathecal administration, has shown impressive efficacy even though SMN is primarily boosted in the Central Nervous System (CNS) and limited by the amount of SMN2 gene expression. The SMN gene therapy strategy, whereas delivered once systemically, could see reduced SMN expression in dividing cells since the SMN gene remains ectopically expressed and not incorporated in the genome. Risdiplam, which is delivered systematically to promote exon 7 inclusion, could in theory provide both CNS and peripheral enhancement of SMN but, like Nusinersen, is limited by SMN2 gene expression. Although it is difficult to functionally assess the relative expression of full-length SMN protein produced by the SMN2 splice modifiers compared with the SMN1gene in humans, whole-tissue SMN protein levels from SMA patient autopsies demonstrate dramatic variability in SMN induction (36).

As data continue to emerge regarding the tissue and temporal requirement of SMN from patients currently undergoing treatment, it will soon become clear where and when additional interventions may be needed. Identifying multiple treatment modalities that enhance SMN induction or stabilize the protein could help broaden the scope of current therapeutic options.

Conclusion

The current goal of SMA therapy is to maintain the gains that have been achieved through the approved SMN boosting therapies. Although these therapies have dramatically improved the disease outcome there remains room to (i) improve the clinical outcomes by targeting pathways that cannot be reversed by SMN restoration or (ii) boost SMN protein expression over current therapies. We show here that a compound that slows SMN degradation can synergistically enhance SMN levels to further improve the SMA phenotype in mice. This pathway offers an additional therapeutic modality to potentially reduce the number of non-responders and broaden the therapeutic range of current treatments.

Materials and Methods

Transfection

Plasmid transfection of SMN-GFP (1 μg) was performed using Lipofectamine 3000 (Invitrogen, Cat. #100 022 052) and P3000 (Invitrogen, Cat. #100 022 058) for 48 h according to the manufacturer protocols. The SMN-GFP was generously donated by Greg Matera. SMN siRNA knock-down was done using 40 pmol of pre-validated siRNA (IDT, Cat. #37 206 943) and RNAi Max (Life Technologies, Cat. #56 532) per manufacturer protocol for 48 h. Opti-MEM (Gibco, Cat. #11 058-021) diluent was used in both protocols. 3813 patient fibroblast cells were transfected with 100 ng/μl E1 ASO using Endo-Porter (GeneTools) and harvested 48 hours post-transfection. Transfection was validated by western blot.

Protein expression analysis

Cell and tissues were lysed in 1% NP-40 (Cat. #28324, Thermofisher), 50 mM Tris–HCl pH 8 (Cat. #15568-025, Thermofisher), 150 mM NaCl and protease inhibitor cocktail (Cat. #04693-159 001, Roche, Indianapolis, IN) on ice for 10 min. The lysates were centrifuged at 4°C for 10 min and the supernatants collected. Protein concentration was determined using the detergent compatible (DC) assay (Cat. #5000116, Bio-Rad, Laboratories, Hercules, CA) according to the manufacturer’s protocol. Protein lysates were resolved by SDS–PAGE (4–12%) (Cat. #XP04122BOX, Thermofisher) and transferred to polyvinylidene difluoride (PVDF) membranes (Cat. #1704272, Bio-Rad). The membranes were blocked in 5% milk (Cat. #1706404, Bio-Rad) and probed with mouse anti-SMN (1, 1000, Cat. #610647, clone 8, BD Biosciences, San Diego, CA), mouse anti-β-actin-peroxidase (Cat. #A3854, Millipore Sigma) and goat anti-mouse IgG horseradish peroxidase (HRP) conjugate (Cat. #BML-A204-0100, Enzo Life Sciences, Farmingdale, NY).

SMA mice

The original breeding pairs for the SMA mice (Smn^+/–SMN2^+/+SMN△7^+/+) were purchased from Jackson Laboratories. The breeding colony was maintained by interbreeding Smn^+/–SMN2^+/+SMN△7^+/+ mice, and offspring were genotyped using polymerase chain reaction (PCR) assays on tail DNA as previously described (17).

ASO and ML372 co-treatment

Animals were treated with 2 nmol MO^E1v1.11 via a single ICV injection on PND1 into the superficial vein on PND2. On day PND11 animals were anesthetized using isoflurane and sacrificed via decapitation. Skeletal muscles were harvested and flash frozen in liquid nitrogen. Tissues were stored at −80°C until processed.

Figure 5

Co-administering ML372 and ASO co-treatment improves motor function and survival of SMA mice. A blinded analysis was performed in litters that were randomized into vehicle- and compound-treated groups. (A) Motor function was assessed using a time to right test in SMA mice. Results shown as mean ± SEM for each day in a treatment group. (B) Body weight of SMA mice in each treatment group was measured daily and is represented as the mean ± SEM. (C) A Kaplan–Meier curve shows the total survival of SMA mice treated with vehicle, ML372, ASO or a combination of ASO and ML37 (log-rank test, P < 0.0001).

Histological analysis of motor neurons and trunk muscle fibers

Mice were transcardially perfused with 4% paraformaldehyde. Lumbar spinal cords (L1–L5) and trunk muscles were dissected and postfixed overnight. Lumbar spinal cords were dissected and hind limb tissues were decalcified, embedded in paraffin and cross-sectioned at the midpoint of the muscle. Embedded tissues were serially cryosectioned at 10 μm thickness with every 10th section from the spinal cord tissue being collected for immunohistochemistry. All analyses were done blinded. Fifteen sections per mouse were mounted on slides and stained with choline acetyltransferase primary antibody (1, 100, catalog AB144P, Millipore Sigma), donkey anti-goat Alexa Fluor-594 secondary antibody (1, 250, Jackson ImmunoResearch) and NeuroTrace Green Fluorescent Nissl (1, 100, catalog N21480, Thermo Fisher Scientific). Digital images were captured using a Zeiss Axiovert 100 M microscope and analyzed with NIH ImageJ software for number and diameter of ventral horn cells, total myofiber number (original magnification, ×20), myofiber diameter (original magnification, ×20) and cross-sectional area. Myofiber diameter was determined by measuring the largest diameter of at least 300 myofibers across all sections per animal.

NMJ pathology

NMJ morphology was assessed using previously published and validated methods (37–41) Briefly, NMJ pathology was examined in two muscles with high vulnerability to denervation. The TVA and RA muscles were dissected from P18 mice and prepared for whole-mount NMJ staining. Muscles were stained using specific antibodies, including anti-NF-H (1, 2000, catalog AB5539, Chemicon, EMD Millipore), anti-synaptophysin (1, 200, catalog YE269, Life Technologies). Acetylcholine receptors were labeled with Alexa Fluor 594-conjugated α-bungarotoxin (1, 200, Life Technologies). NMJ analysis was performed on at least four randomly selected fields of view scattered throughout the innervation region of the entire muscle without any bias for specific regions (×20 objective; Leica DM5500 B, Leica Microsystems Inc.). These images consisted of z-stacks that encompassed the entire thickness of the muscle as to include all the NMJs in that area. The images were analyzed using freely available Fiji Software (NIH) in a double-blinded manner to reduce user bias. NMJ classification was performed as follows: Endplates missing overlapping nerve terminal staining were considered completely denervated, endplates with partial overlap were considered partially denervated and endplates with complete overlap were considered fully innervated.

Righting reflex

Time to right was determined as previously described (42). Briefly, righting time was defined as the average of two trials of the time required for a pup, placed on its back, to turn over and stabilize on all four paws (maximum 30 s).

Weight and survival

Treated mice were weighed daily. Mice that lost 30% of their body weight and were unable to right themselves were euthanized. Survival was displayed using Kaplan–Meier and analyzed using a log-rank test for significance.

Statistics

All statistical analyses were done using GraphPad Prism v7.05 (GraphPad Software, Inc.). Outliers were determined using Grubbs’ test where α = 0.05. Normal distribution was assessed using the Shapiro–Wilk normality test and homogeneity of variances was assessed using either an F test or a Brown–Forsythe test. Data were analyzed using either an unpaired t-test (for comparisons with one independent variable and two groups), one-way analysis of variance (ANOVA) (for comparisons with one independent variable and more than two groups) or two-way ANOVA (for comparisons with two independent variables and more than two groups). Post-hoc analyses were performed using Tukey’s multiple comparisons test. Survival curves were analyzed using a log-rank (Mantel–Cox) test. A P-value <0.05 was considered statistically significant.

Study approval

Animals were housed and treated in accordance with the Animal Care and Use Committee guidelines of the Uniformed Services University of the Health Sciences and the University of Missouri. The Animal Care and Use Committee of the Uniformed Services University and the University of Missouri approved these studies.

Conflict of Interest statement. S.A.D., E.V., E.M.B., J.J.M. and B.G.B. declare no conflicts of interest. C.L.L. is the co-founder and current chief scientific officer of Shift Pharmaceuticals.

Funding

CureSMA; National Institute of Neurological Disorders and Stroke (R01NS091575 to B.G.B.); Missouri Spinal Cord Injury and Disease Research Program to C.L.L.).

Author notes

Co-First authors