Dual SMN inducing therapies can rescue survival and motor unit function in symptomatic Δ7SMA mice

Abstract

Spinal muscular atrophy (SMA) is an autosomal recessive disease characterized by survival motor neuron (SMN) protein deficiency which results in motor neuron loss and muscle atrophy. SMA is caused by a mutation or deletion of the survival motor neuron 1 (SMN1) gene and retention of the nearly identical SMN2 gene. SMN2 contains a C to T change in exon 7 that results in exon 7 exclusion from 90% of transcripts. SMN protein lacking exon 7 is unstable and rapidly degraded. The remaining full-length transcripts from SMN2 are insufficient for normal motor neuron function leading to the development of SMA. Three different therapeutic approaches that increase full-length SMN (FL-SMN) protein production are approved for treatment of SMA patients. Studies in both animal models and humans have demonstrated increasing SMN levels prior to onset of symptoms provides the greatest therapeutic benefit. Treatment of SMA, after some motor neuron loss has occurred, is also effective but to a lesser degree. The SMNΔ7 mouse model is a well characterized model of severe or type 1 SMA, dying at 14 days of age. Here we treated three groups of Δ7SMA mice starting before, roughly during, and after symptom onset to determine if combining two mechanistically distinct SMN inducing therapies could improve the therapeutic outcome both before and after motor neuron loss. We found, compared with individual therapies, that morpholino antisense oligonucleotide (ASO) directed against ISS-N1 combined with the small molecule compound RG7800 significantly increased FL-SMN transcript and protein production resulting in improved survival and weight of Δ7SMA mice. Moreover, when give late symptomatically, motor unit function was completely rescued with no loss in function at 100 days of age in the dual treatment group. We have therefore shown that this dual therapeutic approach successfully increases SMN protein and rescues motor function in symptomatic Δ7SMA mice.

Introduction

Spinal muscular atrophy (SMA) is an autosomal recessive motor neuron disease caused by survival motor neuron (SMN) protein deficiency (Lefebvre et al. 1995; Lefebvre et al. 1997; Coovert et al. 1997). In Caucasian populations SMA has an incidence of about 1/10,000 live births (Pearn 1973; Pearn 1978; Sugarman et al. 2012; Verhaart et al. 2017). As SMN protein is ubiquitous and essential, complete loss results in lethality (Schrank et al. 1997; Blatnik III 2020). SMA patients have a homozygous mutation and/or deletion in all copies of their survival motor neuron 1 (SMN1) gene and must rely on the nearly identical paralogous SMN2 gene as their sole source of SMN protein (Lefebvre et al. 1995). SMN1 and SMN2 differ by a single nucleotide change that greatly reduces exon 7 inclusion in SMN2 transcripts (Gennarelli et al. 1995; Monani et al. 1999; Lorson et al. 1999; Cartegni and Krainer 2002; Kashima and Manley 2003). The loss of exon 7 results in a truncated, unstable SMN protein that is incapable of oligomerization and rapidly degrades (Lorson et al. 1998; Pellizzoni, Charroux, and Dreyfuss 1999; Burnett et al. 2009). While SMN2 does make some FL-SMN protein, the amounts are insufficient for normal motor unit function (Lefebvre et al. 1995; Coovert et al. 1997; Lefebvre et al. 1997).

Severity of clinical presentation of SMA ranges from early-onset, severe Type 0 to later-onset, mild Type 4. The severity of SMA is inversely correlated to SMN2 copy number where more functional copies develop milder phenotypes (McAndrew et al. 1997; Calucho et al. 2018; Feldkotter et al. 2002; Burghes 1997; Mailman et al. 2002). Type 2, 3 and 4 patients typically have more than two SMN2 copies (Arnold and Burghes 2013). Additionally, variants in SMN intron 6 and exon 7 can modulate the amount of FL-SMN generated from a given SMN2, modifying the copy number association (Prior et al. 2009; Bernal et al. 2010; Wu et al. 2017; Ruhno et al. 2019).

Several studies in SMA animal models have previously shown a marked effect on survival, weight and motor unit function mediated by increased FL-SMN protein production. Antisense oligonucleotides that block ISS-N1, a negative regulator of exon 7, enhanced exon 7 inclusion, resulting in significantly improved survival and neuromuscular junction innervation in SMA mice (Hua et al. 2011; Passini et al. 2011; Porensky et al. 2012; Arnold et al. 2014). The single administration of a self-complementary adeno-associated virus subtype 9 with a SMN expression cassette (scAAV9-SMN) showed a remarkable increase in survival of the Δ7SMA mouse from 14 days to over 250 days with complete correction of motor unit function (Foust et al. 2010; Valori et al. 2010; Dominguez et al. 2011). Additionally, delivery of the same scAAV9-SMN to a pig model of SMA showed that early intervention was most effective in preventing muscle weakness and protecting motor neurons (Duque et al. 2015). RG7800, an orally active pre-clinical compound, increases SMN2 exon 7 inclusion by binding two sites in SMN2 transcripts (Sivaramakrishnan et al. 2017). RG7800 binds the weak 5’ splice site to stabilize a bulged adenine in the last nucleotide position of exon 7, while cooperatively binding with an internal splice regulator of exon 7 (ESE2) to promote exon 7 incorporation (Naryshkin et al. 2014; Ratni et al. 2016; Campagne et al. 2019). The daily administration of RG7800 increased FL-SMN levels and survival in SMA mice beyond 65 days (Ratni et al. 2016). Collectively these studies have clearly shown that early, pre-symptomatic SMN restoration is most effective in rescuing the SMA phenotype (Le et al. 2011; Lutz et al. 2011; Foust et al. 2010; Robbins et al. 2014).

There are currently three therapies approved for treatment of SMA patients: an antisense oligonucleotide nusinersen (Spinraza™), gene replacement therapy by onasemnogene abeparvovec (Zolgensma™), and a small molecule compound risdiplam (Evrysdi™)(Finkel et al. 2017; Mendell et al. 2017; Baranello et al. 2021; Mercuri et al. 2018). These therapies result in improved achievement of motor milestones and extended survival in SMA patients. As observed preclinically, all treatment options have shown that early induction of SMN yields the greatest therapeutic benefit as compared to patients treated after symptom onset (De Vivo et al. 2019; Strauss et al. 2020). While newborn screening is implemented in several US states and various other countries, the majority of SMA patients are diagnosed clinically after the onset of motor dysfunction. Consequently, these patients have missed the most optimal pre-symptomatic window for therapeutic intervention. Therefore, additional treatment options for symptomatic patients with motor neuron loss are necessary.

In Δ7SMA mice, symptomatic treatment on postnatal day 4 (P4) using an ASO directed against ISS-N1 was much less effective in rescuing motor unit function and survival than pre-symptomatic intervention on the day of birth (Arnold et al. 2016). It has also been shown that following peripheral nerve injury, mice with low SMN levels demonstrate limited motor neuron repair, whereas those with high levels of SMN demonstrated enhanced motor neuron repair (Kariya et al. 2014);unpublished observation Arnold et al.). Therefore, we asked if increasing SMN production following symptom onset can enhance symptomatic restoration and recover motor unit function in SMA mice.

As the sites of interaction for the two SMN inducing therapies lie close to one another, it is important to know if they will interact in an additive or competitive manner, thus decreasing overall effectiveness. Δ7SMA mice were treated with morpholino ASO directed against ISS-N1 and the drug compound RG7800 at varying symptomatic stages of the disease. We sought to compare the effects of dual therapy versus single treatment in pre-symptomatic (P2), early symptomatic (P4) and late symptomatic (P6) Δ7SMA mice. We wished to determine if dual therapy would have an additive effect on SMN levels to thereby improving motor function and increasing survival in Δ7SMA mice.

Materials and Methods

Breeding

The use of all animals in this study were performed in accordance with our Ohio State University Laboratory Animal Resource approved protocol (2008A0089) following all Institutional Animal Care and Use Committee (IACUC) guidelines. The SMNΔ7 mouse model of SMA (Jackson Laboratory No: 005025), was used for all experiments (Le et al. 2005). Pairs of Δ7 heterozygotes (Smn^+/−; SMN2^+/+; SMNΔ7^+/+) were crossed to generate Δ7SMA mice (Smn^−/−; SMN2^+/+; SMNΔ7^+/+). The Δ7SMA mouse is phenotypically normal at birth but develops progressive weakness at 4 days of age and survives an average of 14 days without therapeutic intervention.

Genotyping

On day of birth (P0), each pup was marked with a unique tattoo pattern for identification and a tail snip was collected. Tails were dissociated in 180 μl of 50mM NaOH at 95°C for 20 minutes and neutralized with 20 μl of 1M Tris-HCl pH 8.0. Genomic DNA was added to a multiplex PCR reaction for the mouse Smn wild-type allele and the mouse Smn knock-out allele as previously described using FP 5`GATGATTCTGACATTTGGGATG, RP1 5` ACCGTTCTTTAGAGCATGCTATG and RP2 5` AACAAACGGCGGATTGAC (Foust et al. 2010).

Treatment Preparation and Assignment

Each litter was culled to 5 pups and any SMA, heterozygous or wild-type pup weighing ≤ 1.0 g at time of birth was excluded from the study (3 of 472 mice genotyped were excluded). The mean weight of Δ7 pups on the day of birth was 1.4 ± 0.2 g. Cross fostering was used to balance animals per treatment cage, and at least one untreated Δ7 heterozygote (Smn^+/−; SMN2^+/+; SMNΔ7^+/+) control was present in each treated litter. Each litter was subjected to the same exclusion criterion (Treat-NMD SOP Code: SMA_M.2.2.003). At P15, litters where all pups weighed less than half the normal weight of Δ7 heterozygotes (10.4 ± 0.3 g) were removed. These litters were considered failure to thrive due to lack of nutrition from mother. Overall, 3 litters were excluded which amounted to 9 of 118 treated SMA mice.

Mice were randomly assigned to one of three treatment groups: RG7800 only, morpholino ASO directed against ISS-N1 only, or RG7800 and ASO together. The same treatment was administered to all Δ7SMA mice in a cage to avoid cross contamination between treatment types. Δ7SMA mice began treatment at 2, 4 or 6 days of age (Fig. 1). RG7800 was administered daily at 3 mg/kg in 100% dimethyl sulfoxide (DMSO) via intraperitoneal (IP) injection from P2, P4 or P6 until P30. From P30 to P80, animals were dosed daily via oral gavage at 10 mg/kg in 0.5% hydroxypropyl methylcellulose (HPMC) and 0.1% Tween-80 (Naryshkin et al. 2014). The ASO was resuspended in sterile 0.9% sodium chloride and a total dose of 81 μg was delivered via a single intracerebroventricular (ICV) injection at P2, P4 or P6 as previously described (Porensky et al. 2012). At P100 all surviving animals were sacrificed for tissue collection.

Fig. 1.

Experimental design and dosing scheme. P0 indicates date of birth and P100 marks the end of the study. Δ7SMA mice were treated pre-symptomatically (P2), at early symptom onset (P4) or late symptomatically (P6). Tissue was collected after 8 days of dosing for mRNA and protein analysis. Electrophysiology measurements were recorded at P30 and P80 after SMN restoration (Arnold et al. 2014; Arnold et al. 2016). Untreated Δ7SMA mice die at 14 days of age.

Weight, survival, and phenotypic analysis

Treatment animals were monitored daily for weight, morbidity, and mortality (Fig. 1). Control and treatment mice were weighed daily from day of birth until 100 days of age. All 5 mice in the litter were weaned at 28 days of age and group housed. Treated Δ7SMA mice were housed together for the duration of the study regardless of sex with control heterozygous mice of only one sex. The development of hindlimb weakness and necrosis was noted. A minimum of 12 mice were studied in each group.

SMN cDNA quantification by droplet digital PCR

A separate group of mice began treatment at 2, 4, or 6 days of age and were sacrificed at 9, 11, and 13 days of age respectively, following 8 days of treatment (Fig. 1). Spinal cord tissue was isolated from treated mice, flash frozen in liquid nitrogen, and stored at −80 °C. In groups receiving RG7800, tissues were collected one hour after the final dose was administered (Zhao et al. 2016). RNA was isolated from spinal cord in all treatment groups using TRIzol reagent (Invitrogen). The RNA was purified with the Qiagen RNeasy Mini Kit and cDNA was synthesized using pdN6 primers (Sigma), RNaseOUT (Invitrogen), and AMV-RT enzyme (NEB) in accordance with the manufacturer’s protocol. mRNA expression was measured in a multiplex droplet digital PCR (ddPCR) reaction (QX200 BioRad System- QuantaSoft Software) as previously described (Porensky et al. 2012). FL-SMN was detected with SMN-FL-FP 5’-CAAAAAGAAGGAAGGTGCTCA, SMN-FL-RP 5’-TCCAGATCTGTCTGATCGTTTC and SMN-FL-FAM probe 5’ FAM-TTAAGGAGAAATGCTGGCATAGAGCAGCAC and normalized to expression of the housekeeping Cyclophilin (Cyclo-FP 5’-GTCAACCCCACCGTGTTCTT, Cyclo-RP 5’-TTGGAACTTTGTCTGCAAACA and Cyclo probe 5’ HEX-CTTGGGCCGCGTCT). Statistical significance was determined with three biological replicates per group and ratios are expressed as Relative Normalized Gene Expression (FL-SMN/Cyclophilin).

SMN protein expression quantification by ELISA

Mice began treatment at 2, 4, or 6 days of age and were sacrificed at 9, 11, or 13 days of age respectively, after 8 days of treatment (Fig. 1). A different cohort of mice were used to ensure sufficient tissue for measurement of RNA and protein. Spinal cord tissue was isolated from treated mice, flash frozen in liquid nitrogen, and stored at −80 °C. In groups receiving RG7800, tissues were collected one hour after the final dose was administered (Zhao et al. 2016). SMN protein was isolated from spinal cord tissue and measured at PharmOptima (Portage, MI) using an enzyme-linked immunosorbent assay (ELISA) as previously described (McGovern et al. 2015; Iyer et al. 2018). An electrochemiluminescence (ECL) immunoassay based on Meso Scale Discovery technology was used. The 2B1 mouse monoclonal antibody was used for capture and a rabbit polyclonal anti-SMN antibody (Protein Tech, Cat. No. 11708-1-AP) labeled with a SULFO-TAG™ was used for detection (Liu and Dreyfuss 1996). A standard curve was used to determine SMN levels using recombinant SMN protein (Enzo Life Sciences, Cat. No. ADI-NBP-201-050) for calibration. The dynamic range of the assay is 10 pg/ml to 20,000 pg/ml. The Meso Scale 6000 sector imager was used to read assay plates. Four biological replicates were analyzed per group and the amount of SMN protein was normalized to the total protein in each sample. The investigator was blinded to treatment groups when analysis was performed.

Electrophysiological analysis of motor unit function

Compound muscle action potential (CMAP) and motor unit number estimation (MUNE) measurements were performed at 30 and 80 days of age. These measurements were assessed from the right hind leg following stimulation of the sciatic nerve as previously described (Arnold et al. 2014; Arnold et al. 2015; Arnold et al. 2016). Additionally, repetitive nerve stimulation (RNS) was performed at a frequency of 40 Hz using surface recording electrodes to record repetitive CMAP responses at 30 days of age (Arnold et al. 2015). RNS decrement was calculated with the following equation: Percent of amplitude decrement = ((Amplitude of tenth response - Amplitude first response)/Amplitude of first response) multiplied by 100%. The investigator was blinded to treatment groups when analysis was performed. A minimum of 8 mice were studied in each group.

Immunofluorescent staining of NMJs

Mice began treatment at 2 or 6 days and were sacrificed at 17 days of age. The splenius capitis muscle was isolated from treated mice and fixed for 24 hours in 4% paraformaldehyde. Whole mount splenius capitis muscle was blocked with 4% Tween 20 (Fisher) and 4% goat serum (Sigma) in PBS for 1 hour. Whole mount splenius capitis was stained with mouse anti-neurofilament 160 kDa monoclonal (1:500, Millipore MAB5252) and rabbit anti-synaptophysin (1:100, Millipore AB9272) antibodies in 4% Tween 20 and 0.4% goat serum, in PBS overnight, followed by a 2-hour secondary incubation with Alexa Fluor-488 (1:1000, Invitrogen) and α–bungarotoxin 594 (1:1000, Invitrogen) and mounted in Flouromount-G (Southern Biotech).

Confocal microscopy

All images were captured using an Andor Revolution WD spinning disk confocal microscope (Andor Technology Ltd, Belfast, UK) equipped with a Yokogawa CSU-W1 confocal scanning unit (Yokogawa Electric Corporation, Tokyo, Japan) and a Nikon TiE inverted epifluorescence microscope (Nikon Instruments, Melville, NY). Excitation of fluorophores was achieved using a 50 mW 488 nm DPSS laser and a 50 mW 561 nm DPSS laser. A 405/488/561/640 multi-band pass dichroic was used in combination with 525/50 (GFP) and 600/50 (RFP) emission filters. Images were acquired using a 20X/0.75NA CFI Plan Apo objective and an Andor Neo 16-bit sCMOS camera. Subsequential image processing was performed with Adobe Photoshop CC.

Statistical analysis

Quantitative data are expressed as means and the standard error of the mean (SEM). Kaplan-Meier survival curves were generated using GraphPad Prism 8 and statistical significance was determined with the log-rank test. For comparison between singly treated groups and dual treated groups, results were analyzed using an unpaired t-test or one-way analysis of variance (ANOVA) with multiple comparisons. RNS results were analyzed using two-way ANOVA with multiple comparisons. Correction for multiple testing was done by the original false discovery rate (FDR) method of Benjamini and Hochberg (Benjamini and Hochberg 1995). Weight curve analysis was determined with the Compare Growth Curve function of the R-Package (Statmod) (Elso et al. 2004; Baldwin et al. 2007). For all experiments, p<0.05 was considered significant.

Results

In this study, we compared the effect of morpholino ASO and drug compound RG7800 on SMN expression and phenotype in Δ7SMA mice when combined and with each treatment alone. We and others have shown that early SMN restoration is most effective in rescuing both survival of the Δ7SMA mouse and motor unit function (Govoni et al. 2018; Arnold et al. 2016; Arnold and Burghes 2013). Here we administered single and dual therapies at three different time points: pre-symptomatically at P2, early symptomatically (P4) and late symptomatically at P6. RG7800 was administered continuously after P2, P4 or P6 and the ASO was administered once at P2, P4 or P6. The timeline and dosing paradigm is described in Fig 1.

FL-SMN mRNA expression is increased in dual treated mice

Reverse transcriptase-PCR was performed on spinal cord tissue collected from Δ7SMA mice treated for 8 days. Droplet digital PCR was used to quantify the inclusion of exon 7 in SMN2 mRNA, and FL-SMN was measured relative to the housekeeping gene Cyclophilin (Fig. 2). The assay used detects FL-SMN mRNA but not mouse Smn or SMNΔ7 mRNA. There was a significant increase in FL-SMN transcript in both single and dual treated groups over untreated Δ7SMA mice. Notably, dual treatment of Δ7SMA mice with both the morpholino ASO and the RG7800 drug compound increased the amount of FL-SMN RNA produced over single treatments at all timepoints including pre-symptomatically at P2, at early symptom onset (P4) and late symptoms (P6). In dual treated mice the amount of FL-SMN transcript increased by 75% (21.5 ± 0.8 relative normalize units (RNU)) when administered at P2, 77% (21.9 ± 0.7 RNU) at P4 and 17% (21.0 ± 0.6 RNU) at P6 when compared to mice treated with RG7800 only at those timepoints (P2 12.3 ± 0.9 RNU, P4 12.4 ± 0.2 RNU and P6 17.9 ± 0.3 RNU) (Fig. 2A, B and C). A similar level of total FL-SMN was reached in all dual treated groups at P2, P4 and P6 (21.5 RNU, 21.9 RNU, 21.0 RNU respectively), demonstrating a threshold is reached in the number of FL-SMN transcripts produced. In fact, the amount of transcript was not significantly different than unaffected heterozygous controls (P2 p=0.562, P4 p=0.701, P6 p=0.455). Thus, the ASO and the drug compound RG7800 can be administered at the same time without competition to achieve normal levels of FL-SMN transcripts in Δ7SMA mice.

Fig. 2.

Quantification of FL-SMN cDNA in spinal cord tissue using ddPCR indicates a significant increase in FL-SMN transcripts in dual treated over singly treated Δ7SMA mice treated for 8 days (n=3 in all groups). SMN cDNA expression was measured relative to Cyclophilin expression. (A) In P2 treated groups, SMN expression is increased in dual (ASO + RG7800) treated mice over ASO treated mice and RG7800 treated mice respectively. (B) In P4 treated groups, SMN expression is increased in dual treated mice over RG7800 treated mice. (C) In P6 treated groups, SMN expression is increased in dual treated mice over RG7800 treated mice. Untreated Δ7SMA mice and Δ7 heterozygous mice served as controls. See Fig. 1 for tissue collection time points. Error Bars = SEM, **p<0.01, ***p<0.001. Statistical test: one-way ANOVA for P2, unpaired t-test for P4 and P6.

Additionally, we measured FL-SMN mRNA in the liver and gastrocnemius muscle in dual treated versus RG7800 animals at P2 and P6 (Suppl. Fig. 1). Both liver (Suppl. Fig. 1A) and muscle (Suppl. Fig. 1B) had increased FL-SMN transcript in P2 and P6 dual treated and RG7800 treated Δ7SMA mice, however the difference was not statistically significant between the dual treatment group and drug alone. This would indicate that the ASO, which is delivered by ICV, makes limited alterations to SMN expression in the periphery, which is consistent with our previous results showing limited effects in tissues outside of the nervous system following this treatment (Porensky et al. 2012).

SMN protein expression is increased in dual treated mice

ELISA was used to measure the amount of SMN protein produced in isolated spinal cord tissue following 8 days of treatment (Fig. 1). Once again, in Δ7SMA mice receiving dual treatment (ASO + RG7800) we observed a significant increase in SMN protein at all time points when compared to singly treated groups (Fig. 3). As previously shown, the greatest increase in SMN protein compared to untreated mice (72% increase) occurred after pre-symptomatic administration at P2 (21.5 ± 0.8 ng/mg dual treatment vs. 12.5 ± 2.4 ng/mg drug only) (Fig. 3A). The absolute amount of SMN protein increase measured at P4 (12.2 ± 0.5 ng/mg) and P6 (10.9 ± 0.6 ng/mg) in the dual treated groups was not as high a P2 (Fig. 3B, 3C). However, there was still a 45% increase in SMN protein at P4 and a 47% increase at P6 over the drug only groups. This nearly 50% increase in SMN is striking as it demonstrates that in Δ7SMA mice, SMN protein can be significantly increased, not only at the onset of symptoms, but more importantly at a post symptomatic timepoint.

Fig. 3.

Quantification of SMN protein in spinal cord tissue measured by ELISA indicates a significant increase in protein expression in dual treated over singly treated Δ7SMA mice treated for 8 days (n=4 mice for each group). (A) In P2 treated groups SMN protein is increased in dual (ASO + RG7800) treated mice over ASO treated mice and RG7800 treated mice respectively. (B) In P4 treated groups, SMN protein is increased in dual treated mice over RG7800 treated mice. (C) In P6 treated groups, SMN protein is increased in dual treated mice over RG7800 treated mice. Untreated Δ7SMA and Δ7 heterozygous mice served as controls. See Fig. 1 for tissue collection time points. Error bars = SEM, *p<0.05, **p<0.01, ***p<0.001. Statistical test: one-way ANOVA for P2, unpaired t-test for P4 and P6.

Survival of Δ7SMA mice is increased following dual treatment

We did not observe a difference in mean survival between drug only and dual treated groups when given pre-symptomatically (P2) (Fig. 4A) (89.5 ± 7.1 days for dual vs 96.7 ± 3.3 days for drug only, p=0.181) and at onset of symptoms (P4) (Fig. 4B) (94.9 ± 5.1 days for dual treatment vs. 96.9 ± 3.1 days for drug only, p=0.901). Yet remarkably at P6, there was an 80% increase in survival in the dual treated group (94.4 ± 5.6 days) vs. the drug only group (52.5 ± 9.1 days) (p<0.0001) (Fig. 4C). The nearly doubling of mean survival in the dual treated group confirms that the increase in SMN protein has a functional impact in Δ7SMA mice (Suppl. Fig. 2A). Although P100 marked the end of the study, all treatment animals continued to be observed to obtain maximum survival for the treatment groups (Fig. 4). Treatment groups receiving RG7800 (RG7800 only or dual treatment) survived on average 60 days after the final dose was administered at P80 at all treatment time points.

Fig. 4.

Δ7SMA mice treated pre- and post-symptomatically improve weight and survival in all treatment groups. (A, B, C) Dual treated Δ7SMA mice rescue survival as compared to singly treated Δ7SMA mice. (A) In P2 treated groups, ASO treated mice (n=16) had a maximum survival of 118 days and RG7800 treated mice (n=18) a maximum of 138 days where dual (ASO + RG7800) treated mice (n=15) survived a maximum of 132 days. (B) In P4 treated groups, ASO treated mice (n=12) had a maximum survival of 86 days and RG7800 treated mice (n=13) a maximum of 158 days where dual treated mice (n=16) survived a maximum of 133 days. (C) In P6 treated groups, RG7800 treated mice (n=13) had a maximum survival of 110 days where dual treated mice (n=15) survived a maximum of 152 days. For comparison, untreated Δ7SMA mice (n=12) survive an average of 14 days and Δ7 heterozygous mice (n=12) survive beyond 200 days. (D, E, F) Dual treated Δ7SMA mice display improved weight as compared to singly treated Δ7SMA mice. (D) P2 treated groups: ASO (n=16, p<0.0001), RG7800 (n=15, p=0.006), and dual (n=18) treated mice. (E) P4 treated groups: ASO (n=12, p<0.0001), RG7800 (n=13, p=0.041), and dual (n=16) treated mice. (F) P6 treated groups: RG7800 (n=13, p=0.023) and dual (n=15) treated mice. Untreated Δ7SMA mice (n=12) and Δ7 heterozygous (n=12) served as controls. Statistical test: Kaplan-Meier survival analysis with log-rank test and Compare Groups of Growth Curve function of the R-Package (Statmod).

Mice treated pre- and post-symptomatically with dual treatment (ASO and RG7800) displayed a marked increase in weight throughout the duration of the study as compared to singly treated groups (Fig. 4D, E, F). Importantly, at 100 days of age all dual treated animals weighed more than drug only treatment groups (Table 1). In fact, the greatest increase in mean weight was observed in the late symptomatic P6 treatment group (Suppl. Fig. 2B). Dual treated mice weighed on average 25% more than drug only treated mice (17.3 ± 0.7g vs 13.8 ± 0.5g) at 100 days of age.

Table 1.

Mean weight of Δ7SMA mice treated pre-symptomatically, at early symptom onset, and late symptomatically. Dual treated mice at P2 (n=15) exhibited improved weight as compared to dual treated mice at P4 (n=16, p=0.058) and P6 (n=15, p<0.0001). RG7800 treated mice at P2 (n=18) exhibited improved weight as compared to RG7800 treated mice at P4 (n=13, p=0.039) and P6 (n=13, p<0.0001). ASO treated mice at P2 (n=16) weighed more than ASO treated mice at P4 (n=12, p<0.0001). Δ7 heterozygous (n=12) served as controls. Statistical test: Compare Groups of Growth Curve function of the R-Package (Statmod).

Mean weight (grams) of treated Δ7SMA mice
	P2			P4			P6
Age (Days)	ASO	RG7800	ASO+ RG7800	ASO	RG7800	ASO+ RG7800	ASO	RG7800	ASO+ RG7800	Δ7 Control
30	13.1 ± 0.6	12.7 ± 0.5	14.5 ± 0.5	8.2 ± 0.5	11.2 ± 0.7	13.5 ± 0.5	---	6.6 ± 0.6	9.4 ± 0.5	21.7 ± 0.7
60	17.1 ± 0.9	20.9 ± 0.7	22.2 ± 0.6	12.2 ± 0.6	16.6 ± 0.9	19.9 ± 0.7	---	11.5 ± 0.9	14.5 ± 0.7	26.1 ± 0.5
80	17.7 ± 0.5	20.8 ± 0.5	22.8 ± 0.6	10.7 ± 0.5	18.6 ± 0.9	21.3 ± 0.7	---	13.7 ± 0.8	16.5 ± 0.7	27.9 ± 0.6
100	17.0 ± 0.8	20.8 ± 0.6	24.9 ± 0.8	---	18.9 ± 1.1	21.7 ± 0.8	---	13.8 ± 0.5	17.3 ± 0.7	29.1 ± 0.6

Interestingly, we found that drug only treatment groups had an improved phenotype versus the ASO only treatment group. As previously described, necrosis of the tail, eyes and ears can develop in ASO or drug treated Δ7SMA mice at 45 to 65 days of age (Porensky et al. 2012; Zhao et al. 2016). Specifically, those treated with ASO present with a truncated tail at weaning age (P21) that necroses at 45 to 65 days of age. In this study mice dosed symptomatically (P4 and P6) with drug only displayed signs of ear necrosis on average at 60 days of age. Mice at P6 with drug only additionally presented with eye and tail necrosis on average at 50 days of age. Of note, all dual treated groups have better grooming, normal tail length, and did not develop necrosis of any kind during the study (Suppl. Fig. 3).

Electrophysiology of motor unit function is fully restored following dual treatment

CMAP and MUNE are well-established clinical techniques used to assess motor unit function in SMA patients (Swoboda et al. 2005). CMAP size provides a measure of the total output of motor units supplying a muscle or group of muscles, whereas MUNE estimates the number of motor units supplying a certain muscle (Arnold et al. 2014). We have shown that CMAP and MUNE, which are decreased in untreated Δ7SMA mice, recover by P30 following SMN restoration (Arnold et al. 2014; Arnold et al. 2016). As corrected motor unit function is a measure of therapeutic efficacy in SMA, we assessed motor function in dual treatment versus single treatment groups with CMAP and MUNE measurements at 30 and 80 days of age (Fig. 5 and 6).

Fig. 5.

CMAP and MUNE measures in dual treated versus singly treated Δ7SMA mice at 30 days of age. CMAP measures showed no significant difference in (A) P2 groups: ASO (n=8, p=0.524) and RG7800 treated mice (n=13, p=0.278) compared to dual (ASO + RG7800) treated mice (n=11), (B) P4 groups: ASO (n=10, p=0.041) and RG7800 treated mice (n=14, p=0.600) compared to dual treated mice (n=14), nor in (C) P6 groups: RG7800 treated mice (n=11, p=0.105) versus dual treated mice (n=14). MUNE measures showed no significant difference in (D) P2 groups: ASO (n=8, p=398) and RG7800 treated mice (n=13, p=0.028) compared to dual treated mice (n=11), nor in (F) P6 groups: RG7800 treated mice (n=11, p=0.449) versus dual treated mice (n=14). (E) MUNE measures in P4 ASO treated mice (n=10, p=0.005) were statistically significant from dual treated mice (n=13), where RG7800 treated mice at P4 (n=14, p=0.695) showed no difference from dual treated mice. The same Δ7 heterozygous mice (n=8) served as controls for all treatment time points. Untreated Δ7SMA mice survive an average of 14 days and could not be used as controls. **p<0.01. Statistical test: one-way ANOVA for P2 and P4, unpaired t-test for P6.

Fig. 6.

CMAP and MUNE measures in dual treated versus singly treated Δ7SMA mice at 80 days of age. CMAP measures showed no significant difference in (A) P2 groups: ASO (n=10, p=0.2854) and RG7800 treated mice (n=17, p=360) compared to dual (ASO + RG7800) treated mice (n=10), (B) P4 groups: ASO (n=4, p=0.929) and RG7800 treated mice (n=12, p=0.388) compared to dual treated mice (n=16) nor in (C) P6 groups: RG7800 treated mice (n=5, p=0.453) versus dual treated mice (n=14). MUNE measures showed no significant difference in (D) P2 groups: ASO (n=10, p=0.608) and RG7800 treated mice (n=17, p=0.482) compared to dual treated mice (n=10), (E) P4 groups: ASO (n=4, p=0.687) and RG7800 treated mice (n=12, p=0.551) compared to dual treated mice (n=16), nor in (F) P6 groups: RG7800 treated mice (n=4, p=0.119) versus dual treated mice (n=14). The same Δ7 heterozygous mice (n=6) served as controls for all treatment time points. Untreated Δ7SMA mice survive an average of 14 days and could not be used as controls. Statistical test: one-way ANOVA for P2 and P4, unpaired t-test for P6.

CMAP responses in dual treated groups were recovered at 30 days (Fig. 5A, B, C) (P2 33.9 ± 2.2 mV, P4 37.6 ± 2.1 mV, P6 37.0 ± 2.3 mV). Single treatment groups also had increased CMAP amplitudes to levels not statistically different from dual treatment; ASO (P2 31.7 ± 3.5 mV, P4 30.3 ± 3.3 mV) or RG7800 (P2 37.1 ± 1.3 mV, P4 35.9 ± 1.8 mV, P6 42.5 ± 2.1 mV). Additionally, MUNE showed an increase of motor unit numbers in dual treated groups (Fig. 5D, E, F) (P2 291.9 ± 19.6, P4 278.2 ± 20.2, P6 244.3 ± 19.1). Mice treated with RG7800 only also showed improved MUNE measures at all treatment time points (P2 362.0 ± 22.1, P4 265.9 ± 18.9, P6 266.4 ± 21.4). MUNE measures in P4 ASO treated mice (175.9 ± 31.8) were reduced when compared to RG7800 only treated (265.9 ± 18.9) and dual treated mice (278.2 ± 20.2) (Fig. 5E). Collectively, these results demonstrate that motor unit function in dual treated groups is rescued at P30.

Longitudinal electrophysiological responses revealed no change in motor unit function correction at 80 days of age (Fig. 6). Dual treated groups still demonstrated increased CMAP (P2 32.8 ± 4.2 mV, P4 33.6 ± 2.3 mV, P6 39.1 ± 2.4 mV) (Fig. 6A, B, C) and MUNE (P2 247.6 ± 17.7, P4 232.7 ± 19.2, P6 219.3 ± 15.6) (Fig. 6D, E, F) measurements at all treatment time points. Similar to what we observed at P30, we again did not observe a difference in CMAP and MUNE between dual treatment versus single treatment groups at all time points. RG7800 only treated mice still demonstrated full recovery of CMAP (P2 36.3 ± 2.0 mV, P4 36.6 ± 2.9 mV, P6 42.3 ± 2.4 mV) and MUNE (P2 273.4 ± 21.8, P4 254.3 ± 31.7, P6 274.6 ± 31.5).

Additionally, treatment mice were analyzed for electrophysiological dysfunction of the neuromuscular junction (NMJ). This dysfunction, known as decrement, has been observed in SMA patients (Wadman et al. 2012). To assess for defects and failure of NMJ transmission, CMAP amplitude decrement was recorded and determined present if amplitude decreased following 10 repetitive stimuli. Dual treated and RG7800 treated Δ7SMA mice both exhibited features of NMJ transmission defects at 30 days of age (Fig. 7). CMAP decrement recorded was similar in RG7800 and dual treated groups at all timepoints, demonstrating neither treatment timing nor treatment were able to effectively eliminate NMJ transmission defects in treated Δ7SMA mice.

Fig. 7.

Treated Δ7SMA mice exhibited NMJ defects at 30 days of age following repetitive nerve stimulation at a frequency of 40 Hz. Treatment with either RG7800 or ASO + RG7800 had no effect on improving NMJ defects (p=0.999). Treatment timing also had no effect on improving CMAP decrement (p=0.970). The observed effect of treatment on CMAP decrement is consistent across treatment time points (p=0.477). Statistical test: 2-way ANOVA.

Innervation of the splenius capitis muscle is increased in dual treated mice

To assess whether dual treatment had an effect on NMJ innervation, we examined NMJ innervation of the splenius capitis muscle—a relatively severely denervated muscle in the SMNΔ7 mouse model (Ling et al. 2012). By P14, the splenius capitus was shown to be 60% innervated, however induction of SMN at P1 rescues this phenotype (Feng et al. 2016b). The splenius capitis was immunostained for nerves, presynaptic terminals, and acetylcholine receptors (AChR) with anti-neurofilament antibody, anti-synaptophysin antibody and α-bungarotoxin, respectively (Fig. 8). Δ7SMA mice treated at P2 with drug only (Fig. 8A) and the P2 dual treatment groups (Fig. 8B) have similar denervation in the splenius capitis, in contrast to the fully innervated heterozygous mice (Fig. 8E). Late treatment at P6 with drug only (Fig. 8C) displays more NMJ denervation when compared to dual treatment at P6 (Fig. 8D). Thus, dual treatment improves innervation of the splenius capitis muscle in the late symptomatic group.

Fig. 8.

Confocal micrographs showing the morphology of NMJs in the splenius capitus muscles of Δ7SMA mice treated pre- and late symptomatically. Δ7SMA mice treated at (A) P2 with drug, (B) P2 with dual treatment, (C) P6 with drug or (D) P6 with dual treatment were compared to (E) a Δ7 heterozygous control animal. NMJs were labeled with anti-neurofilament (green), anti-synaptophysin (green) antibodies and alpha-bungarotoxin (red). Dual treatment improves NMJ innervation of the splenius capitis muscle at P2 and P6 Scale bar = 500 μm in each micrograph.

Discussion

The clinical severity of SMA patients roughly correlates to variations in SMN2 copy number. Patients with Type 1 SMA typically have 2 copies of SMN2, account for about a third of cases, and present with a severe phenotype by 6 months of age (Zerres and Rudnik-Schoneborn 1995; Calucho et al. 2018). The milder disease phenotypes of Type 2 (typically 3 SMN2 copies) and Type 3 patients (3-4 SMN2 copies) account for 20% and 30% of cases respectively, with disease onset occurring later than 6 months of age (Arnold, Kassar, and Kissel 2015). Although there is a range of severity and disease onset among SMA patients, muscle weakness and atrophy are the predominant clinical features in all disease types. These symptoms are caused by progressive and irreversible motor neuron loss due to reduced amounts of FL-SMN (Crawford and Pardo 1996; Arnold and Burghes 2013; Kuru et al. 2009; Stam et al. 2019). Hence, age of onset and SMA type correlate to SMN2 copy number and SMN protein levels (Coovert et al. 1997; Lefebvre et al. 1997; McAndrew et al. 1997; Calucho et al. 2018; Burghes 1997). Mild forms of SMA with a predominant motor neuron phenotype have proven difficult to replicate in transgenic mice (Burghes et al. 2017). A pharmacologically induced mouse model displays a disease phenotype similar to milder forms of SMA following suboptimal administration of a SMN2 splicing modifier to Δ7SMA mice beginning on the day of birth (Feng et al. 2016a). At 32 days of age, a switch from the suboptimal dose to a high dose of drug results in increased survival, weight, and muscle size. However, there was no correction of motor unit function and denervation was still present. Here in our study, rather than extending survival with a suboptimal dose to mimic disease variability, we employed a different approach by treating Δ7SMA mice at P4 and P6 when weakness and motor neuron loss are present (Monani et al. 2000; Baumer et al. 2009; Arnold et al. 2014). We have previously shown recovered electrophysiological measures in Δ7SMA mice with post-symptomatic (P4 and P6) restoration of SMN with an ASO directed at ISS-N1, however fewer motor units are rescued the later treatment is administered (Arnold et al. 2016). The motor neuron phenotype in SMA patients is more pronounced than in the mouse models, as evidenced by the dramatic loss of motor neurons in Type 1 autopsy patients (Crawford and Pardo 1996; Harding et al. 2015). Significant denervation and loss of motor neurons also occurs in milder forms of SMA in patients (Kuru et al. 2009; Arnold and Burghes 2013; Swoboda et al. 2005; Stam et al. 2019). Therefore, our preference here was to maintain the severe motor neuron phenotype in the Δ7SMA mouse when testing therapeutics.

In the last several years, three effective SMN inducing therapies have been approved for treatment in SMA patients: antisense oligonucleotide nusinersen (Spinraza™), gene replacement therapy onasemnogene abeparvovec (Zolgensma™), and a small molecule compound risdiplam (Evrysdi™) (Finkel et al. 2017; Baranello et al. 2021; Mendell et al. 2017; Mercuri et al. 2018). Delivery of these treatments result in remarkable rescue of motor function and increased survival, particularly when administered pre-symptomatically (De Vivo et al. 2019; Strauss et al. 2020). Currently, newborn screening for SMA is not universally accessible across all populations and most patients will receive a clinical diagnosis after symptom onset. It is imperative that newborn screening efforts continue to advance and be implemented in all US states and countries. Yet even with implementation of newborn screening, the 5% of SMA cases caused by a small mutation in the SMN1 gene instead of the typical deletion will be missed by current diagnostic methods (Burghes and McGovern 2017; Burghes and Beattie 2009; Dangouloff et al. 2020). While treatment administered to symptomatic SMA patients can stabilize and prevent further motor neuron loss, there is a critical need for improved therapies in this population (Darras et al. 2019).

The opportunity to combine treatments for SMA could improve therapeutic outcomes in symptomatic patients. This could be achieved by administering an SMN independent therapy with one of the approved SMN inducing therapies. SMN independent therapies include enhancing muscle strength through either myostatin inhibition or Troponin C activators, which should result in greater muscle function with limited motor neuron input (Rose et al. 2009; Russell et al. 2012; Feng et al. 2016a). Combining myostatin inhibition and SMN restoration has been performed in SMA mice. Long et al. used the pharmacological model of mild SMA to test the addition of SRK-015P, an antibody that blocks myostatin action (Long et al. 2019). The mouse was first treated with a low dose of the SMN inducing compound SMN-C1 for 24 days and then switched to a high dose of SMN-C1 with the addition of SRK-015P. An increase in muscle size and force was observed. In addition, Zhou et al. delivered a myostatin propeptide via AAV in combination with a morpholino ASO against ISS-N1 in mice with similar results (Zhou et al. 2020). In both cases the weight of muscle was increased but survival of the mouse was not markedly altered. Finally, some have suggested the administration of a neuroprotectant with a SMN inducing therapy, however this may not improve the outcome as SMN itself protects the neurons that are present (Dessaud et al. 2014). Additionally, low SMN levels have been shown to affect the repair of motor neuron connectivity with muscle following nerve injury (Kariya et al. 2014). Conversely, we have found that increased SMN levels, expressed from a high copy number SMN2 transgene, resulted in an improved rate of repair following nerve crush injury (Arnold et al unpublished results).

An alternative method for improving therapeutic outcomes in symptomatic patients is to further increase SMN levels by combining two SMN inducing therapies. Here we report, for the first time in mice, the effect of administering the two SMN induction therapies currently approved for treatment in children and adults with SMA. In this study we found a marked increase in survival of Δ7SMA mice treated at P6 with both ASO and RG7800 drug compound. In fact, dual treated mice at P6 exhibited survival similar to those treated with either single treatment starting at P2. Moreover, dual treatment at P6 revealed an 80% increase in survival over mice treated with drug only at P6. Notably, mice treated at P4 with RG7800 alone had improved survival and motor unit function when compared to mice treated with ASO alone at P4. This may be due to differences in biodistribution of the ASO in the spinal cord and the periphery. We additionally attempted dual treatment at P8, but survival was markedly reduced to less than 30 days. Thus, unlike myostatin combined approaches, administration of two SMN inducing therapies resulted in a marked extension of survival. We also observed an increase in total body weight with dual treatment. Finally, combining treatments clearly increased SMN protein levels. It has been previously reported that neonatal functional tests like righting reflex are extremely variable, especially in Δ7SMA mice with a more pronounced phenotype (Butchbach, Edwards, and Burghes 2007). Therefore, due to difficulty in reproducibility and the number of mice required for significance, we did not perform righting reflex in this study. Furthermore, open field tests, grip strength, and rotarod, which can only be performed in older mice, also have reproducibility issues requiring large cohorts of mice to obtain significance. As such, we chose to assess motor neuron physiology as the readout for correction of the motor unit and measurement of survival and weight for improved phenotype of the mice.

Even though the RG7800 site of action that resolves the poor U1 site and the negative splice regulator ISS-N1 lie adjacent each other, there was no observed competition or toxicity when administered together (Singh et al. 2017; Sivaramakrishnan et al. 2017; Campagne et al. 2019). Of note, preclinical studies with continuous dosing of RG7800 displayed retinal toxicity in cynomolgus monkeys, but this has not been reported in mice (Ratni et al. 2018). While no overt signs of toxicity such as early death were observed in our study, we did not measure any specific metabolic pathway in dual treated mice to assess this. However, lack of toxicity in dual treated mice is supported by the increase in SMN levels and improvement of the SMA phenotype at all time points.

We would predict that combining AAV9-SMN and RG7800 or ASO would also increase SMN to provide a therapeutic benefit. However, this cannot be tested in late symptomatic Δ7SMA mice as there is a rapid switch of AAV9-targeting of motor neurons to astrocytes at P4 and P6, a feature that does not occur in primates or pigs (Foust et al. 2009; Meyer et al. 2015; Bevan et al. 2011). Moreover, at this time gene therapy is only approved for treatment in children under the age of two.

To evaluate if dual treatment rescued motor unit function, we employed the clinically relevant electrophysiological measurements of CMAP and MUNE. Previously untreated Δ7SMA mice were reported to have reduced CMAP and MUNE outputs beginning at P6 (Arnold et al. 2014). Furthermore, Δ7SMA mice after symptom onset (P4 and P6) with ASO directed against ISS-N1 displayed reduced motor unit function at P21 and P30. Here we again confirmed that P4 ASO only treated mice have reduced MUNE measurements at P30. Remarkably, when ASO was combined with the drug at P4, and even at P6, motor unit function was completely rescued. Consequently, we extended the analysis to P80 and did not detect abnormalities in motor unit function as measured by CMAP and MUNE. It should be noted that full recovery of CMAP and MUNE also occurred with the single treatment in the drug only groups at P4 and P6. In SMA, it is suggested that the loss of NMJ function precedes the death of the motor neuron. In our study, rescue of motor unit function (MUNE) at P6 during late symptomatic treatment suggests that the remaining motor units are still functional. However, it is unclear if there is a similar window of rescue in SMA patients where improvement in NMJ function can result in the maintenance of motor unit survival. Due to the limitations of the mouse model, we are unable to determine this without administering dual treatment within a clinical setting.

NMJ transmission defects are observed in patients and animal models of SMA (Kariya et al. 2008; Wadman et al. 2012; Pera et al. 2019). CMAP amplitude decrement is a measure of NMJ transmission used in the clinic and is determined through repetitive nerve stimulation. P4 treated ASO Δ7SMA mice demonstrated decrement when compared to Δ7 heterozygous controls. In this study dual treatment did not eliminate decrement at any treatment time point, indicating that the NMJ transmission defect is still present. The exact reason for decrement in SMA is not known, but it has been suggested that high levels of SMN are required early for full development of the neuromuscular junction (Kariya et al. 2014). The presence of this defect has also been observed after nusinersen treatment in SMA patients (Arnold et al. 2021). Although NMJ transmission defects are still present, we found that dual treatment improved NMJ innervation of the splenius capitis, a vulnerable muscle known to be denervated in Δ7SMA mice at P14 (Ling et al. 2012). This finding is not surprising given CMAP and MUNE are rescued in the P2 drug and P2 dual treatment groups. MUNE estimates the number of motor units that contribute to CMAP therefore, the restoration of these measures indicates recovery of the motor unit in the muscles they specifically innervate. The presence of decrement indicates persistent NMJ transmission dysfunction and abnormality of vesicle recycling, suggesting that the motor neuron or the synaptic space is altered. We did not include rotorod, grip strength, or righting reflex behavioral tests because CMAP and MUNE directly assess motor unit function and are not complicated by gross stimulation from upper motor and sensory neurons. CMAP, MUNE, and decrement measures allowed us to determine that the rescued NMJs are functional but that they aren’t entirely normal, which may not have been observed with behavioral tests.

While the loss of motor neurons remains the hallmark of SMA phenotype, the overall severity of motor neuron dysfunction appears to be less pronounced in the mouse than in humans. In addition, SMA mice develop bradycardia and dilated cardiomyopathy with decreased contractility, which can be the cause of sudden death among treatment groups (Bevan et al. 2010; Heier et al. 2010; Wijngaarde et al. 2017). While mouse models of SMA indicate involvement of tissues outside of the central nervous system such as the spleen, kidney, and heart, these same peripheral defects are only rarely found in SMA patients (Bevan et al. 2010; Heier et al. 2010; Shababi et al. 2010; Somers et al. 2016; Szunyogova et al. 2016; Deguise et al. 2017; Thomson et al. 2017; Nery et al. 2019; Allardyce et al. 2020). Some autopsy reports of SMA patients have indicated other organ involvement, but without appropriate controls it is impossible to determine if the effect is primary or secondary to SMA. For example, smaller spleens with abnormal pathology have been reported in mice (Thomson et al. 2017). Yet autopsy reports indicated either no change in spleen size or the presence of red pulp that could be due to infection. Moreover, the presence of an accessory spleen is unlikely to be related to SMA as this occurs in ten percent of the normal population (Moore 1992; Thomson et al. 2017). As reviewed by Iascone et al, most reported changes, such as bradycardia and spleen abnormalities, are not consistently present in SMA patients (Iascone, Henderson, and Lee 2015). The possibility of combined treatment influencing peripheral organs that are present in mouse models of SMA but not represented in the same manner in SMA patients could influence the survival of SMA. As such it is important to perform clinical trials in SMA patients to confirm the additive effect of these SMN inducing therapies.

In summary, combining an ASO, which modifies SMN2 splicing, with the drug compound RG7800, which also modifies SMN2 splicing, resulted in increased SMN levels and extended survival in symptomatically treated Δ7SMA mice. Moreover, dual treated mice displayed increased weight and full rescue of motor unit function at 100 days of age. These positive outcomes indicate there is no toxicity or competition when two SMN splice modifying therapies are co-administered. As such clinical trials of combined therapies should be considered in symptomatic SMA patients.

Highlights:

• Dual treatment with SMN2 splice modifying therapies, an ASO directed against ISS-N1 and the RG7800 drug compound, significantly increased FL-SMN transcript and protein levels in SMA mice.

• Post-symptomatic (P6) administration of dual therapies increased SMN levels, extended survival and recovered weight gain comparable to that obtained by either treatment when administered pre-symptomatically (P2).

• Dual treatment completely rescued CMAP and MUNE deficits in SMA mice.

• The promising additive effect of dual therapy observed in SMA mice should be confirmed in clinical trials with SMA patients.

Abbreviations

SMA

spinal muscular atrophy

SMN

survival motor neuron

ASO

antisense oligonucleotide

AAV9

adeno-associated virus subtype 9

ddPCR

droplet digital polymerase chain reaction

CMAP

compound muscle action potential

MUNE

motor unit number estimation

RNS

repetitive nerve stimulation

NMJ

neuromuscular junction

ANOVA

analysis of variance

P

postnatal

SEM

standard error of the mean

ICV

intracerebroventricular

IP

intraperitoneal