Pre-designed rat Bet1L siRNA was obtained from GE Healthcare Dharmacon (L-091291-02-0005, Chicago, IL). Scrambled sequence siRNA (Dharmacon, D-001810-10-05) was also purchased as a control. siRNA was mixed with Oligofectamine^TM transfection reagent (12252-011, Thermo Fisher Scientific, Waltham, MA) and RNAase (Qiagen, Hilden, Germany) in a buffer (diluted from 5 × siRNA buffer, NC1338268, Thermo Fisher Scientific) for a final concentration of 5 μM siRNA. Oligofectamine is a fast-acting, non-toxic transfection reagent that has been used to deliver siRNA into nuclear or cytoplasmic targets both in vitro and in vivo (Nakajima et al., 2012; Seven et al., 2018).

All the animal procedures in the present study were approved by the University of Wisconsin School of Veterinary Medicine Animal Care and Use Committee (IACUC, Protocol ID V005430) and carried out in accordance with the guidelines for our IACUC and NIH standards of animal care. SOD1^G93A transgenic and age-matched wild-type rats were used in this study. SOD1^G93A transgenic male founders, originally obtained from Taconic (Hudson, NY) (Howland et al., 2002), were crossed with wild-type female Sprague-Dawley rats to maintain colonies. This rat colony was maintained as previously described (Suzuki et al., 2007b; Hayes-Punzo et al., 2012; Van Dyke et al., 2016; Lynch et al., 2021). All rats were housed, bred, and sacrificed in accordance with UW-Madison and NIH standards of animal care. Genotyping of SOD1^G93A-positive rats was determined from ear punches using real-time PCR, which was performed by Transnetyx (Cordova, TN). Based on our latest analysis, the recent colonies of our ALS rats showed disease onset on average at day 179.5 ± 8 days (Lynch et al., 2021). The ALS rats used in this study were expected to be at the pre-symptomatic stage because all experiments were concluded before 96 days of age (Suzuki et al., 2007b; Hayes-Punzo et al., 2012; Lynch et al., 2021).

In the first study, we used healthy wild-type rats and evaluated the efficacy of Bet1L siRNA for gene silencing. The treatment groups were: (1) Bet1L siRNA, (2) scrambled siRNA, (3) vehicle with Oligofectamine, and (4) no injection (sham control). Animals were anesthetized with isoflurane, and then treatment solutions (30 μL per muscle) were injected unilaterally into tibialis anterior (TA) muscles of WT rats. The injections were repeated for 3 days. The subsequent injections were performed nearly at the same site as the original location of the muscle, slowly releasing siRNA or vehicle. On the 4th day, the animals were sacrificed, and their hindlimb muscles (tibialis anterior muscle, TA) were collected. The non-injected TA muscles from the unilaterally Bet1L siRNA animals were used in analysis as the no-injection treatment group.

In the subsequent experiments, pre-symptomatic SOD1^G93A rats received intramuscular injections at 75 days of age (Figure 1). Four treatment groups were prepared (Bet1L siRNA, scrambled siRNA, vehicle, and no injection). Similar treatment groups of age-matched WT rats were also prepared. Treatment solutions (30 μL per muscle) were unilaterally injected into the hind limb muscles (TA, gastrocnemius, and quadriceps femoris muscles). The injections were repeated 3 days in a row. After 1 week (i.e., 82 days of age) or 3 weeks (96 days of age) after the first injection, TA muscles and spinal cords were collected from each rat). The non-injected TA muscles of the unilaterally Bet1L siRNA animals were used in analysis as the no-injection treatment group.

Muscle tissues (~200–250 mg) were homogenized using a Silent Crusher M tissue homogenizer (Heidolph, Wood Dale, IL) in 1 mL of ice-cold homogenization buffer (20 mM Tris-Cl pH 7.8, 137 mM NaCl, 2.7 mM KCl, 1 mM MgCl₂, 1 mM EDTA, 1 mM DTT, 1% Triton X-100, 10% glycerol; all from Sigma Aldrich) supplemented with Halt Protease Inhibitor Single-use Cocktail (100 × , 78430, Thermo Scientific, Rockford, IL). Protein concentrations were determined using the Pierce BCA Protein Assay Kit (23227, Thermo Scientific).

Muscle tissue homogenates (20 μg protein per lane) were run on a 12.5% polyacrylamide gel and transferred onto a PVDF membrane (Millipore Sigma, Burlington, MA). The membrane was immunoblotted with anti-Bet1L (1:1000, mouse monoclonal, BDB610960, BD Biosciences, Franklin Lakes, NJ) or anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) antibodies (1:1000, rabbit monoclonal, D16H11, Cell Signaling Technology Inc., Danvers, MA). Following secondary incubation with anti-mouse or rabbit peroxidase-tagged antibody (1:1000, Promega, Madison, WI) and incubation with enhanced chemiluminescence (Pierce ECL Western Blotting Substrate, Thermo Fisher Scientific), the membrane was scanned by the LI-COR Odyssey XF imaging system (Lincoln, NE). The protein band signals were densitometrically analyzed using ImageJ software.

TA muscles were transversely sectioned at 20-micron thickness using a cryostat. Muscle sections spanned ~920 mm from the injection site (where the injected reagents had been distributed). The muscle sections were placed on a glass slide and fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at room temperature. Sections were then blocked in 5% normal donkey serum (NDS, Jackson ImmunoResearch Labs, West Grove, PA) in PBS for 2 h at room temperature, followed by primary antibody incubation in 5% NDS and 0.3% Triton-X 100 (Sigma Aldrich) overnight at 4°C. Then the sections were washed five times for 5 min each with PBS before adding secondary antibodies diluted in PBS supplemented with 5% NDS. After the wash step was repeated, some sections were incubated in Hoechst 33258 (1:10,000 in PBS, Sigma Aldrich) for 15 min at room temperature, followed by additional washes. Fluoromount-G mounting medium (Thermo Fisher Scientific) was applied to the slide after the last wash, and a glass coverslip was placed on top. Bet1L expression at the NMJs was identified by co-labeling with α-bungarotoxin (BTX) conjugated to Alexa Fluor 647 or 594 (1:1,000, Thermo Fisher Scientific) and an anti-Bet1L antibody (1:1,000, rabbit polyclonal, PA5-58943, Thermo Fisher Scientific). To determine any correlations between Bet1L expression and NMJ innervation, the degree of NMJ integrity was determined by co-staining for Bet1L, motor endplate (BTX), and motor axon terminals labeled by two antibodies against synaptic proteins and neurofilaments (1:40 with SV2 and 2H3 clones, respectively, mouse monoclonal, Developmental Studies Hybridoma Bank, Iowa City, IA).

For some images to secure their entire NMJ orientation, a modified immunostaining protocol was applied to longitudinal muscle sections (Nishimune et al., 2008; Rogers et al., 2017). Briefly, TA muscles were fixed with 2% PFA-PBS for 30 min at room temperature, rinsed in PBS a few times, and submerged in 20% sucrose in PBS for over 48 h. Fixed muscles were frozen in Optimal cutting temperature compound (Fisher Scientific) and longitudinally cut at 20-micron thickness using a cryostat. After processing with similar staining steps with antibody incubations and washing as described above, a cover slide was mounted on the slide glass with Prolong Glass Antifade Mountant (Thermo Fisher Scientific).

Fluorescent images were taken using a Nikon Eclipse 80i fluorescent microscope (Nikon Instruments Inc., Tokyo, Japan) with an ORCA-Fusion BT digital camera (C15440, Hamamatsu Photonics, Shizuoka, Japan). For some representative images, the three-dimensional renderings of rat NMJ immunohistochemistry were created using the z-stack feature of a Leica TCS SP8 confocal microscope (Leica, St. Galen, Switzerland). The number of Bet1L-positive or negative NMJs was quantified for each animal as an average from four different muscle sections. The innervation rate analysis of NMJs has been described previously (Rogers et al., 2017; Lynch et al., 2021). Briefly, fully innervated NMJs were identified by perfect overlap with adult motor nerve terminals and acetylcholine receptor clusters, which were identified by the antibodies against motor axon terminals (synaptic protein SV2A and neurofilament NH-M, SV2 + NF) and acetylcholine receptor (AChR)-positive endplates in the muscle (Rogers et al., 2017). NMJs were assessed for areas of the acetylcholine receptor clusters that were not occupied by nerves, whether in part or in full, which were considered to be denervated NMJs (Rogers et al., 2017). The average number of total NMJs counted per animal was 60–80. Quantifications were performed from 4 to 7 rats of each experimental group, leading to ~350 NMJs counted per group. When analyzing the number of immunolabeled NMJs, the researchers performing the quantification were blinded to the treatment group, and results were averaged between them to avoid potential bias. The results were then confirmed by multiple lab staff.

We performed motor neuron analysis similarly as described in previous papers (Suzuki et al., 2007a, 2008, 2010; Krakora et al., 2013; Austin et al., 2022). Briefly, small tissue segments of rat spinal cord (T13-L6) were embedded in paraffin, sectioned by a Leica RM2255 microtome, and processed for cresyl violet staining. Previous studies demonstrated that the T13-L6 segments of the spinal cord innervate the siRNA-injected hindlimb muscles (Suzuki et al., 2007a; Mohan et al., 2014, 2015). Bright-field microscopy images were captured from the ventral horn of the spinal cord using a Keyence BZ-X710 microscope (Osaka, Japan). A standardized circular area of interest was placed over the ventral horn of the spinal cord image, and a scale bar was added in Google Draw (San Francisco, CA). Nissl-stained motor neurons residing in the area of interest were characterized by a distinct nucleolus and a darkly stained cytoplasm. The diameter (an average of the minimal and maximal diameter) and average number of motor neurons were analyzed by NIH ImageJ software. As in the NMJ quantification above, the analysis was blindly performed by multiple lab staff.

For the ALS rats that received intramuscular siRNA injection for 3 weeks, changes in hindlimb motor function were assessed weekly in a small enclosure using the Basso, Beattie, and Bresnahan (BBB) locomotor rating scale (Basso et al., 1995). The BBB locomotor rating scale was obtained as previously described (Suzuki et al., 2007b; Hayes-Punzo et al., 2012). Each rat was allowed to walk around in an open cage while we observed hindlimb movements for ~3–5 min. Each hind limb score was based on the 21-point scoring scale from no movement (0) to normal locomotion (21). Pre-symptomatic was defined as ranging from 21 to 18, symptom onset from 17 to 14, and symptomatic from 13 to 0 (Suzuki et al., 2007b; Hayes-Punzo et al., 2012). Scoring criteria include paw rotation, toe clearance, weight support, the frequency of each abnormality to locomotion, and the amount of movement occurring from each joint (Basso et al., 1995). To avoid potential bias, the BBB scaling was performed in a blinded manner with regard to the treatment group. Additionally, video clips were recorded from each rat and used to confirm the results by three different lab staff.

GraphPad Prism 9.4.0 (GraphPad Software, Inc., La Jolla, CA) was used to perform statistical analyses. Quantitative data were graphed as means ± standard error of the mean (SEM) from at least two independent experiments. One-way analysis of variance (ANOVA) was used to compare data in the same animal group. A statistically significant one-way ANOVA was followed up with Tukey's post-hoc test for multiple comparisons. A two-way ANOVA was calculated to compare multiple groups, followed up with Tukey's multiple comparisons test. Differences were considered significant when P < 0.05.

In order to explore the effects of Bet1L siRNA-based knockdown, we first confirmed siRNA efficacy for gene silencing using healthy wild-type rats. Bet1L siRNA was prepared at a final concentration of 5 mM and injected into the tibialis anterior (TA) muscles of healthy adult WT rats. Three control groups were prepared: scrambled siRNA, vehicle only (Oligofectamine), and no-injection sham. After the injections were repeated for 3 days, the animals were sacrificed on the 4th day, and the injected TA muscles were collected. We performed Western blotting using muscle homogenates (Figure 2A). The densitometric analysis of proteins revealed that Bet1L siRNA injection significantly reduced its protein level up to ~18% (P < 0.05, 18.3 ± 2.8%) when compared to the level in the other three controls (80.7 ± 6.5% in scrambled siRNA, 89.7 ± 4.1% in vehicle only, 100.0 ± 7.3% in no injection; n = 6 in each group) (Figure 2B).

Next, we performed immunohistochemistry to identify siRNA-based inhibition of Bet1L levels at the NMJs in the siRNA-injected muscle (Figures 2C, D). The overlap of acetylcholine receptor (AChR)-positive motor endplates (labeled by fluorochrome-conjugated BTX) and Bet1L proteins indicated the presence of Bet1L at the NMJs in the control rats. In contrast, the lack thereof depicted the reduction or absence of Bet1L in the Bet1L siRNA-injected muscle (Figure 2C). Corresponding to our recent observations (Lynch et al., 2021), confocal microscopy using longitudinal muscle sections showed the specific localization of Bet1L protein at the NMJ of a no-injection control rat (Supplementary Movie 1). However, Bet1L expression was reduced at the NMJ by intramuscular Bet1L siRNA injection (Supplementary Movie 2). After quantitative comparison between four experimental groups (n = 4 each), we found that Bet1L siRNA-injected muscles had a lower percentage of NMJs positive with Bet1L when compared to three control groups (Figure 2D). Specifically, the difference was statistically significant between Bet1L siRNA-injected and the non-injection control (P < 0.01) (Figure 2D).

After confirming the ability of Bet1L siRNA to reduce Bet1L expression in rat skeletal muscle, we next determined whether distal morphological changes could occur by silencing Bet1L expression. WT rats at 75 days of age were unilaterally injected with their respective treatment (Bet1L siRNA, scrambled siRNA, vehicle only, or sham) into their TA muscles (n = 4). The injection was repeated daily for 3 days. At 4 days after the initial injection, the TA muscles were collected and sectioned to determine both NMJ innervation and Bet1L expression at the NMJs by immunohistochemistry (Figures 3A, B). A quantitative analysis of NMJ innervation showed that intramuscular Bet1L siRNA injection significantly decreased the percentage of innervated NMJs when compared to all three controls (Figure 3B). Corresponding to these innervation results, the percentage of denervated NMJs was increased in Bet1L siRNA-injected rats (Figure 3B).

Next, we accounted for the potential impact of the ALS genetic background on the morphological outcome of silencing Bet1L expression. Similarly to the last study using WT rats, presymptomatic ALS rats (75 days old) were unilaterally injected with their respective treatment (Bet1L siRNA, scrambled siRNA, vehicle only, or sham; n = 4 each group) into their TA muscles, and we analyzed NMJ innervation and Bet1L expression by immunohistochemistry (Figure 3C). The knockdown of Bet1L in ALS rat muscles significantly reduced NMJ integrity (23.9 ± 14.2%), decreasing the percentage of innervated NMJs (Figure 3C).

We further analyzed the correlation between NMJ integrity and Bet1L expression following Bet1L siRNA-based gene silencing (Figures 3D, E). In both WT and ALS rats with Bet1L siRNA treatment, the majority of innervated NMJs had Bet1L present, whereas the majority of denervated NMJs did not express Bet1L. Similar correlations between Bet1L expression and NMJ integrity were observed in all three control groups.

Next, we prepared another cohort of pre-symptomatic ALS and age-matched WT rats and unilaterally injected Bet1L siRNA in their hindlimb muscles (TA, gastrocnemius, and quadriceps femoris muscles). The animals received the injection once per day for the first 3 days in a week, and these weekly injections were repeated for up to 3 weeks. We then observed the animals closely for any changes to motor neuron function and viability from prolonged Bet1L knockdown.

At the end of the experiment, in the third week after the first siRNA injection, TA muscles were harvested, sectioned, and co-stained to analyze any changes in NMJ innervation. In WT rats (Figure 4A), intramuscular siRNA injection for 3 weeks significantly decreased the percentage of innervated NMJs (P < 0.05, 74.4 ± 4.7%) when compared to two control groups, vehicle (93.8 ± 0.9%) and no injection (96.0 ± 1.9%, n = 4 each group). There was a similar trend in the difference between Bet1L and scrambled siRNA-treated rats, but the difference did not reach significance. Bet1L siRNA-injected ALS rats had significantly lower levels of innervated NMJs (P < 0.05, 47.4 ± 7.3%) when compared to all three controls (76.0 ± 7.9% scrambled, 77.9 ± 4.3% vehicle, and 72.9 ± 5.8% no injection, n = 7 each group; Figure 4B).

Together with the results in Figure 3, we calculated differences in the effect of Bet1L gene silencing in NMJ innervation using two-way ANOVA in terms of the genotype (WT vs. ALS), duration (Week 1 vs. Week 3), and their interaction in Bet1L siRNA-treated animals (Figure 4C). While there was no statistical difference in the duration, a significant difference was observed in the genotype (P < 0.001). The interaction between genotype and duration was not significantly different. Additional post-hoc comparisons revealed that ALS rats showed a lower level of innervated NMJs at both Week 1 and Week 3 (P < 0.05, Figure 4C), indicating that more severe effects in NMJ integrity were induced by siRNA-based Bet1L in ALS model rats. In contrast, a similar statistical analysis did not identify any statistical differences in the three control groups.

We next hypothesized that the distal denervation at the NMJs, which were induced by intramuscular Bet1L siRNA injection, would lead to motor neuron degeneration. To answer this, we prepared spinal cord sections from the animals used for NMJ analyses above and analyzed cresyl violet (Nissl)-stained motor neurons within the region (T13-L6) that innervated the siRNA-injected hindlimb muscles (Suzuki et al., 2007a; Mohan et al., 2014, 2015) (Figure 5A).

We first determined the diameter of motor neuron cell bodies (the average of the minimal and maximal diameter) in the ventral horn of the spinal cord (Figure 5B). In WT rats, the diameter of motor neuron cell bodies was commonly lower in Bet1L siRNA-injected rats (116.0 ± 6.1 mm, n = 63) when compared to scrambled siRNA (119.8 ± 6.5 mm, n = 67), vehicle only (127.1 ± 5.7 mm, n = 61), and no injection (125.2 ± 6.5 mm, n = 58). However, this difference did not reach significance (Figure 5B left). In contrast, the cell diameter of motor neurons in Bet1L siRNA-injected ALS rats (73.3 ± 1.9 mm, n = 140, P < 0.05) was significantly lower than all controls: 91.6 ± 1.8 mm in scrambled siRNA (n = 91), 94.9 ± 2.6 mm in vehicle only (n = 85), and 81.4 ± 1.9 mm in no injection (n = 126) (Figure 5B right).

Next, the number of motor neurons was determined using the spinal cord sections (Figure 5C). We counted the neurons with large-sized cell bodies in the ventral horn, which represent alpha motor neurons. In WT rats (Figure 5C left), there was no significant difference in the number of motor neurons between the four treatment groups: 10.7 ± 0.8 in Bet1L-siRNA, 12.1 ± 0.8 in scrambled siRNA, 12.1 ± 0.8 in vehicle-only, and 12.1 ± 0.8 in non-injection (n = 4 each). In ALS rats, the number of alpha motor neurons was lower in the Bet1L siRNA-injected rats (8.0 ± 0.5; n = 7) when compared to scrambled siRNA-treated rats (13.1 ± 0.9, n = 7; P < 0.05) (Figure 5C right). The two other controls also showed a similar trend (vehicle only 12.1 ± 1.4; no injection 10.9 ± 1.5; n = 7 each), although the difference was not statistically significant.

Furthermore, we analyzed differences in the effect of Bet1L gene silencing by comparing the values between genotypes (WT vs. ALS). Although no statistical differences were identified in the motor neuron diameter, ALS rats showed a lower number of motor neurons than WT rats when Bet1L-siRNA was treated (P < 0.05, Figure 5C).

Over the experimental period, individual limb function had been analyzed using the Basso-Beattie-Bresnahan (BBB) rating test (Basso et al., 1995) (Figure 6). A BBB score of 21 indicated consistent plantar stepping and coordinated gait, consistent toe clearance, predominant paw position parallel throughout stance, consistent trunk stability, and a tail consistently up, whereas a BBB score of 0 indicated no observable hindlimb movement (complete paralysis) (Basso et al., 1995).

In WT rats (Figure 6A), there was no obvious difference in the locomotor function between treatment groups after 1 week of unilateral injection. The BBB score was decreased by intramuscular Bet1L siRNA injection in the second (14.8 ± 1.1) and third (17.5 ± 1.0) weeks of the injection, while all rats in the other three groups were identified as having a BBB score of 21. A two-way ANOVA revealed that there was a statistical difference in the treatment (F_{3, 48} = 33.16, P < 0.0001), week (F_{3, 48} = 7.478, P = 0.0003), and their interaction (F_{9, 48} = 6.534, P < 0.0001). Based on multiple comparisons using the post-hoc analysis, Bet1L siRNA-treated rats showed significantly lower BBB scores after 2 weeks when compared to the other control groups (P < 0.01, Figure 6A).

In ALS rats, Bet1L siRNA-based knockdown showed a similar trend. Specifically in the third week, Bet1L siRNA-treated rats significantly decreased BBB scores (15.8 ± 0.6, P < 0.01) when compared to three control groups (scrambled siRNA, 17.4 ± 0.4; vehicle, 18.0 ± 0.3; and no injection control, 21.0 ± 0.0) (Figure 6B). Similarly, as seen in WT rats, a two-way ANOVA identified a statistical difference in the treatment (F_{3, 96} = 72.49, P < 0.0001), week (F_{3, 96} = 42.66, P < 0.0001), and their interaction (F_{9, 96} = 6.45, P < 0.0001). Additional post hoc analysis found that Bet1L siRNA-treated rats significantly reduced their BBB scores after 1 week of the treatment (P < 0.01, Figure 6B). Although all three controls in WT rats fully recovered their motor function by the 3rd week (Figure 6A), ALS rats treated with either scrambled siRNA or the vehicle showed lower BBB scores when compared to the no-injection control (Figure 6B).

In summary, our results indicated that Bet1L siRNA-based knockdown of Bet1L influenced NMJ innervation and motor function in both WT and ALS rats. Bet1L gene silencing commonly showed more severe effects in ALS rats, indicating that the disease progression was accelerated by Bet1L siRNA injection, and the disease stage shifted from pre-symptomatic to symptomatic.