Dietary NMN supplementation enhances motor and NMJ function in ALS

Abstract

Amyotrophic lateral sclerosis (ALS) is an adult-onset neurodegenerative disease that causes the degeneration of motor neurons in the motor cortex and spinal cord. Patients with ALS experience muscle weakness and atrophy in the limbs which eventually leads to paralysis and death. NAD⁺ is critical for energy metabolism, such as glycolysis and oxidative phosphorylation, but is also involved in non-metabolic cellular reactions. In the current study, we determined whether the supplementation of nicotinamide mononucleotide (NMN), an NAD⁺ precursor, in the diet had beneficial impacts on disease progression using a SOD1^G93A mouse model of ALS. We found that the ALS mice fed with an NMN-supplemented diet (ALS^+NMN mice) had modestly extended lifespan and exhibited delayed motor dysfunction. Using electrophysiology, we studied the effect of NMN on synaptic transmission at neuromuscular junctions (NMJs) in symptomatic of ALS mice (18 weeks old). ALS^+NMN mice had larger end-plate potential (EPP) amplitudes and maintained better responses than ALS mice, and also had restored EPP facilitation. While quantal content was not affected by NMN, miniature EPP (mEPP) amplitude and frequency were elevated in ALS^+NMN mice. NMN supplementation in diet also improved NMJ morphology, innervation, mitochondrial structure, and reduced reactive astrogliosis in the ventral horn of the lumbar spinal cord. Overall, our results indicate that dietary consumption of NMN can slow motor impairment, enhance NMJ function and improve healthspan of ALS mice.

1. Introduction

Amyotrophic lateral sclerosis (ALS) is an adult-onset fatal neurodegenerative disease characterized by the death of upper and lower motor neurons (MNs). ALS patients initially experience muscle weakness and atrophy in the limbs which progresses into paralysis and typically survive 2–5 years after symptoms appear (Zufiria et al., 2016). Numerous genes have been linked to ALS, including SOD1, TARDBP, and C9ORF72 (Mejzini et al., 2019). Currently the only approved treatments for ALS have displayed limited efficacy (Agrawal and Biswas, 2015; Sturmey and Malaspina, 2022; Gwathmey et al., 2023). As such, research into effective interventions that slow the progression of ALS is greatly needed. Common ALS symptoms include body weight loss, locomotor deficits, muscle weakness and atrophy, and paralysis. Symptoms arise from deficits in function of the central (spinal cord/brain) and peripheral (neuromuscular junctions, NMJ) nervous system. During ALS, NMJs undergo extensive structural and functional changes that contribute to muscle dysfunction and motor impairments and MN death, exhibiting denervation, broken motor endplates, mitochondrial dysfunction, and altered neurotransmission (Alhindi et al., 2022), making investigation of NMJs critical to understanding ALS development and progression.

Nicotinamide adenine dinucleotide (NAD⁺) is an abundant metabolite in mammalian cells and is involved in hundreds of cellular reactions and pathways, including DNA repair, oxidative stress, and energy metabolism (Verdin, 2015). NAD⁺ is synthesized using one of three different pathways: the de novo, Preiss-Handler, and NAD⁺ salvage pathways. In mammalian cells, the majority of NAD⁺ is generated using the salvage pathway, where nicotinamide (NAM) is converted into nicotinamide mononucleotide (NMN), by the rate-limiting enzyme nicotinamide phosphoribosyltransferase (NAMPT), and NMN is subsequently converted into NAD⁺ by nicotinamide mononucleotide acetyltransferases (NMNAT1–3) (Revollo et al., 2004; Canto et al., 2015; Chini et al., 2021). Nicotinamide riboside (NR) can also be converted to NMN and bypass the rate-limiting step in the salvage pathway. Altered NAD⁺ homeostasis has been linked to aging and age-related neurodegenerative diseases, including Alzheimer’s disease, Parkinson’s disease, and ALS (Lautrup et al., 2019; Covarrubias et al., 2021; Lundt and Ding, 2021). Many studies indicate that administering NAD⁺ precursor metabolites, typically NAM, NMN, or NR, can increase NAD⁺ levels and correct certain declines (Mills et al., 2016; Ryu et al., 2016; Wang et al., 2017; Hou et al., 2018; Lundt et al., 2020; Li et al., 2023).

Abnormal and defective bioenergetic homeostasis has been observed in ALS and NAD is critical for maintaining bioenergetic homeostasis (Dupuis et al., 2004; Browne et al., 2006). There is also growing evidence showing that NAD homeostasis is impaired in ALS, with the salvage pathway specifically affected. Metabolomic profiling of the ALS mouse motor cortex revealed alterations to the NAD⁺/NADH ratio and NAM levels, additionally, NAD⁺ contents are decreased in the brain and spinal cords of ALS mice (Roderer et al., 2018; Obrador et al., 2021; Gautam et al., 2022). Disrupted NAM metabolism has also been observed in both ALS mice and human patients. NAM levels are reduced in ALS patient blood serum and cerebrospinal fluid. ALS mice administered NAM subcutaneously exhibited better motor performance than ALS mice not provided with NAM (Blacher et al., 2019). In the salvage pathway, expression of NAMPT and NMNAT2, the NMNAT isoform predominant in neurons, are altered in human ALS spinal cords (Wang et al., 2017; Roderer et al., 2018; Harlan et al., 2020).

In our previous studies we demonstrated that deleting NAMPT from projection neurons caused locomotor deficits, neurodegeneration, and NMJ dysfunction, a phenotype reminiscent of ALS mice; however, administering NMN can enhance motor behavior, NMJ structure and function, and survival of these NAMPT deficient mice (Wang et al., 2017; Lundt et al., 2020). P7C3, a potential NAMPT activator, increased MN survival and improved walking gait in ALS mice (Tesla et al., 2012). NMN and NR can elevate NAD⁺ levels, improve motor activity, increase lifespan and counteract the toxicity of SOD1^G93A ALS astrocytes on MNs (Harlan et al., 2016; Harlan et al., 2020; Obrador et al., 2021). NMN treatment increased axonal and neurite outgrowth and improved mitochondrial structure in corticospinal MNs from ALS mice (Gautam et al., 2022).

However, in vivo investigation of NMN supplementation on NMJ function in ALS has not been reported. Here, using the SOD1^G93A ALS mouse model, which is the most widely studied and has a well-established disease phenotype, we observed NAD⁺ decline at the pre-symptomatic stage. We therefore hypothesized that NMN administration would have beneficial effects on ALS. To test this hypothesis, we fed ALS mice with NMN-supplemented diet, and studied its effect on survival, motor behaviors, NMJ function and structure, MN survival and mitochondrial morphology. Our study may provide a therapeutic strategy for ALS.

2. Materials and methods

2.1. Mice and NMN-supplemented diet

Mice were maintained on a 12 h light:12 h dark cycle (lights on 7 am-7 pm) in our AAALAC-accredited animal facility at the University of Missouri. All experimental procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals and experiments were approved by the University of Missouri Animal Care Quality Assurance Committee. Adult male and female wild-type C57BL/6 J (WT) and hemizygous transgenic SOD1^G93A ALS mice (Strain 002726 B6SJL-Tg(SOD1*G93A), Jackson Lab, ME) were used for this study. At 8 weeks old, ALS mice were assigned to a control diet (i.e., standard mouse chow pellets) or an NMN-supplemented diet (hereafter, ALS and ALS^+NMN mice). The NMN-supplemented diet contained 2.7 g NMN/kg mouse chow. Ad libitum feeding from this diet translates into 400 mg/kg/day in adult mice. NMN was obtained from dietary NMN supplement tablets (Gene Formulas, Davie, FL) and stored at 4 °C. Pellets of the NMN-supplemented diet consisted of standard mouse chow, NMN powder, xanthan gum (Bob’s Red Mill, Milwaukie, OR), and water. Standard mouse chow pellets were ground into a fine powder. Chow powder, NMN powder, xanthan gum, and water were mixed until well combined. Individual pellets were made using a granule machine and placed in dark room until completely dry. Pellets were made weekly, kept in an air-tight bag and stored at 4 °C until needed. For survival curve, a mouse was considered to have reached end-stage when it was unable to right itself within 20 s after being placed on its side (Hatzipetros et al., 2015).

2.2. NAD assay

Red blood cell, plasma, and cortical NAD⁺ and NADH concentrations from 4, 6 (pre-symptomatic), 11–12 (symptom on-set) and 19–20 week-old (end-stage) ALS and age-matched WT mice were assessed using commercially available kits (Cat. No. E2ND-100, Bioassay systems, Hayward, CA) as we previously reported (Wang et al., 2017).

2.3. Behavioral assessment

Behavior testing was performed between 8:30 am-12:30 pm. All the animals were taken out from the animal facility in their home cages to the behavior test room and habituated for at least 30 min starting at 8:00 am. Mice were given an interval of at least 30 min for recovery in their home cages between different behavior tests. All instruments were sanitized between tests to avoid olfactory effects. Behavior testing began at 7 weeks of age for all mice, considered to be the pre-symptomatic period for SOD1^G93A mice (Rocha et al., 2013; Olivan et al., 2015) and occurred weekly during the testing period.

2.4. Rotarod test

Accelerated rotarod is a commonly used test to assess motor performance in rodents (Knippenberg et al., 2010). An ENV-557 M accelerating rotarod (Med Associates, VT) was used for testing. Mice were pre-trained to learn to walk on the accelerating rod prior to testing. Mice were placed on the rod and allowed to move freely. The 4 to 40 rpm acceleration setting was used. The time when the mouse fell off the rod was recorded, with a maximum time of 180 s. If a mouse dropped before 180 s, a second trial was given with at least 10 min between trials.

2.5. Hanging wire test

The hanging wire test is used to assess upper limb strength (Klein et al., 2012). Mice were pre-trained prior to testing. Mice utilized forelimbs to hang their body weight on a wire stretched between two posts separated by 50 cm. The wire was 60 cm high, with padding placed underneath to cushion mice should they drop. The time until the mouse dropped from the wire was recorded. A time of zero was assigned if the mouse dropped off immediately with a maximum time of 120 s. Two trials were performed for each mouse at each test period, with at least 10 min between trials.

2.6. Four limb grip force

Four limb grip force testing is used to measure peak muscle force (Dong et al., 2020). Grip strength was measured by a M4–2 digital force gauge (Mark-10 Corporation, Copiague, NY) attached with a rectangular grid. The apparatus was used to measure the combined strength of the forelimbs and hindlimbs. The mouse was held near the base of the tail and gently lowered to allow the forelimbs and hindlimbs to grab the grid. The mouse was gently pulled back by the tail and the mouse’s body stayed parallel to the grid. Peak force was recorded from the reading when the mouse released the grid. Mice were tested over five trials with at least 30 s between trials. The largest three values of the five were chosen, and the average value of them was used for analysis.

2.7. Open field test

The open field test is used to assess voluntary movement (Jambeau et al., 2022). After at least 30 min habituation in the testing area, the mouse was placed in the center of the open field recording chamber, which consists of a clear, open Plexiglass box (46.6 cm × 38.5 cm × 25.6 cm) with an overhead camera to record movements of mice for 10 min. The total distance travelled, total time immobile, and maximum speed during the 10 min testing period were analyzed by ANY- maze software (Stoelting, IL). Mean speed was calculated by dividing the total distance travelled by the total time mobile. The sensitivity for the detection of immobile was set as 65% of the body area. If the mouse remained motionless for 5 s it was considered immobile.

2.8. Walking gait analysis

Walking gait was assessed using a modified version of a previously established protocol (Wertman et al., 2019). Briefly, the forepaws and hindpaws of the mice were painted with two contrasting colors (orange and purple, respectively), using a non-toxic water-based paint. The mice were then gently placed on one end of textured watercolor paper strip that was approximately 43 cm long and 12 cm wide. The sides of the paper were bordered with blocks to ensure the mice walked in a straight line. The strip led into a darkened enclosure to encourage the mice to walk down the paper. The paws of the mice were then cleaned to remove any remaining paint. Parameters measured, for both forelimbs and hindlimbs, were stride length (the center of the paw pad to the center of the following paw pad) and stride width (the distance between the left and right forelimbs or hindlimbs. For both forelimbs and hindlimbs, the initial stride on the paper and the initial stride after the mouse had stopped, was not recorded.

2.9. End-Plate Potential (EPP) recording

Electrophysiological recordings of isolated semitendinosus muscles were conducted using methods previously described (Polo-Parada et al., 2005; Wang et al., 2017). The recording chamber, covered with Sylgard, was perfused with regular Tyrode’s solution (140 mM NaCl, 5.6 mM KCl,1 mM MgCl₂, 2 mM CaCl₂, 1.8 mM Na₂HPO₄, 10 mM NaHCO₃, 5.5 mM glucose) and continuously received 95% O₂/5% CO₂. The isolated muscle-nerve preparation was carefully extended and pinned flat on the bottom of the recording chamber. Nerves were sucked into a tight polyethylene electrode and stimulated with a Grass S44 stimulator (pulse duration of 0.5 ms, Grass Medical Instruments, Quincy, MA) through a PSIU6 photoelectric stimulus isolation unit (Grass Instrument Co, Quincy, MA). The minimum voltage and current threshold to induce muscle contraction was used for recording. Muscles lacking contractility were discarded. A muscle specific voltage-gated Na⁺ channel blocker μ-conotoxin GIIIB (1 μM, Alomone Labs, Jerusalem, Israel) was added to the chamber to prevent muscle contractions during recording. Sharp glass electrodes (40–60 MΩ) filled with 3 mM KCl solution were inserted into individual muscle fibers near motor endplates. EPPs were then recorded with a BA-1S bridge amplifier (NPI Electronic, Tamm, Germany) during stimulations and acquired with a Digidata 1440 A digitizer (Molecular Devices, San Jose, CA) using Axoscope 11.2 software (Molecular Devices, San Jose, CA). Only muscle fibers with resting membrane potentials between −60 and −80 mV were used for recording. Muscle fibers with resting membrane potential fluctuation of more than ±5 mV were discarded during the recording. Evoked EPPs were recorded during voltage pulse stimulations of 10, 20, 50, 100, and 200 Hz. The stimulation period for each frequency lasted approximately 2 s, with at least 5 s between stimulation periods. Paired-pulse facilitation, starting with 100 ms interval between pulses, was recorded following the frequency stimulation. Each interval was repeated 3 times, with at least 3 s between recordings. At the conclusion of both the frequency stimulation and the paired-pulse stimulation, the recording was continued without stimulation to acquire miniature EPPs (mEPPs) for assessing spontaneous vesicle release and quantal content.

Data was analyzed using Clampfit 11.2 (Molecular Devices, San Jose, CA). Only the first second of stimulation period for each frequency was used for analysis. The amplitude (resting baseline to peak of response) of the first five and last five responses of the first second of stimulation were recorded. These responses were averaged together to determine the mean EPP amplitude at the start (first five responses) and end (last five responses) of one second of stimulation. The EPP amplitude ratio was determined as the ratio of the mean EPP amplitudes of the last five responses to the first five responses. To study paired-pulse facilitation, the EPPs of the first and second responses with different pulse intervals were recorded and the ratio of the second response amplitude to the first response amplitude was calculated. At least three facilitation ratios were calculated for each pulse interval and were averaged together to determine the mean facilitation at each pulse interval.

To measure spontaneous vesicle release, mEPPs were recorded. At least 3 different end-plate recordings were used to assess mEPP amplitude, with at least 200 mEPPs for each recording. The responses were fitted with a Gaussian distribution, with peak of the best-fit line used as the mean amplitude. Each end-plate recording had a unique mEPP amplitude with individual amplitudes averaged together to determine mean mEPP for each condition. To determine mEPP frequency, the number of mEPPs that occurred in a 10–15 s period following the final frequency or paired-pulse stimulation event were counted. The 10–15 s period began at least 5 s after the final stimulation event. Quantal content was calculated by dividing the mean EPP amplitude at 10 Hz by the mean mEPP amplitude of each individual end-plate recording.

2.10. Immunostaining

2.10.1. Neuromuscular junction (NMJ)

To analyze the structure of NMJs, fresh semitendinosus muscles were isolated from 18-week-old WT, ALS, and ALS^+NMN mice following cervical dislocation. Muscles were incubated with α-bungarotoxin (α-BTX) conjugated with Alex Flour-555 (1:1000, B34451, Invitrogen) for 25–30 min, and then washed 3 times for 10 min with 1× PBS. Muscles were then fixed with 4% paraformaldehyde (PFA) overnight at 4 °C and washed 3 times for 20 min with 1× PBS. Fixed muscles were pinned to the bottom of a Sylgard dish and incubated with rabbit-neurofilament-200 (1:300, N4142, Millipore Sigma) on shaker at 4 °C for 1 week. Muscles were washed with 1× PBS for at least 4 days on shaker at 4 °C, with PBS replaced each day. Muscles were then incubated with donkey anti-rabbit AlexaFluor-488 (1:400, A21206, Invitrogen) on shaker at 4 °C for 1 week followed by washing with 1× PBS for at least 4 days on shaker at 4 °C, with PBS replaced each day. Muscles were embedded in optimal cutting temperature (OCT) compound and stored at −20 °C prior to cutting. Using a cryostat (Leica CM1900), 30 μm longitudinal sections of muscle were cut and placed onto coated glass microscope slides. Slides were allowed to dry, carefully washed with 1× PBS to remove excess OCT around muscle slices and sealed with glass coverslips using Permount.

2.10.2. Brain and spinal cord

18-week-old WT, ALS, and ALS^+NMN mice were anesthetized with urethane (200 mg/ml) and perfused with cold 1× PBS. The brains and spinal cords were quickly dissected, placed in 4% PFA, and stored at 4 °C for 4 days. After 2 days, the lumbar portion of the spinal cord was dissected out and returned to 4% PFA and stored at 4 °C for another 2 days. Brains and spinal cords were then transferred to 30% sucrose solution for 7 days. Tissues were embedded in ice and sectioned using a cryostat (Leica CM1900). 30 μm coronal (brain) and transverse (lumbar) sections were cut. Slices were then covered with a cryoprotective solution and stored at −20 °C until needed.

For immunolabelling, slices were washed 3 times for 10 min to remove cryoprotective solution. Slices were then blocked for 1 h (10% Donkey serum and 0.3% Triton-X in 1× PBS) and washed again 3 times for 10 min. Slices were incubated with goat-choline acetyltransferase (ChAT, 1:300, AB144P, Abcam), rabbit-NeuN (1:300, #12943, Cell Signaling), rabbit-GFAP (1:300, G9269), or rabbit-Iba1 (1:300, 019–19,741, FUJIFILM) at 4 °C on shaker for overnight. Slices were then washed 3 times for 10 min and incubated with donkey anti-rabbit Alexa Fluor-488 (1:400, A21206, Invitrogen) or donkey anti-goat Alexa Fluor-568 (1:400, A11051, Invitrogen) at room temperature on shaker for 4 h. Slices were then washed 3 times for 10 min. Slices were then carefully transferred to glass microscope slides and covered with DAPI (P36931, Invitrogen) or Permount mounting media.

2.10.3. Transmission electron microscopy (TEM)

18-week-old WT, ALS, and ALS^+NMN mice were anesthetized with urethane (200 mg/ml) and perfused with cold 1× PBS. The spinal cord and semitendinosus muscles were quickly dissected out. Immediately, muscles were cleaned of excessive non-semitendinosus muscle, connective tissue, and fat. The proximal and distal end of the muscle were discarded with only the central portion of the muscle, which contains the NMJs, was retained. The muscle was then cut into small pieces and fixed in 100 mM sodium cacodylate buffer containing 2% PFA and 2% glutaraldehyde (pH 7.35). Samples were left at room temperature for at least one hour and moved to 4 °C for at least 23 h. Spinal cords were placed in 4% PFA and stored at 4 °C for 4 days. The vertebrae were carefully removed, and L1/2 of lumbar spinal cord was dissected out. The dorsal portion was discarded, and the ventral horns were carefully cut into smaller sections and placed in 100 mM sodium cacodylate buffer containing 2% PFA and 2% glutaraldehyde (pH 7.35). Samples were left at room temperature for at least one hour and moved to 4 °C for at least 23 h.

Fixed samples were washed with 100 mM sodium cacodylate buffer (pH 7.35) containing 130 mM sucrose. Secondary fixation was performed using 1% osmium tetroxide in a cacodylate buffer using a Pelco Biowave (Ted Pella, Inc. Redding, California) operated at 100 W for 1 min. Samples were incubated at 4 °C for 1 h, then rinsed with cacodylate buffer, followed with distilled water. En bloc staining was performed using 1% aqueous uranyl acetate and incubated at 4 °C overnight, then rinsed with distilled water. A graded dehydration series (30%, 50%, 70%, 90%, 100%, 100%) was performed using ethanol at 4 °C, transitioned to acetone. Dehydrated samples were then infiltrated with Epon resin for 24 h at room temperature and polymerized at 60 °C for 48–72 h. Samples were cut into 85 nm thick longitudinal (semitendinosus) or transverse (spinal cord) sections using an ultramicrotome (Ultracut UCT, Leica Microsystems, Germany) and a diamond knife (Diatome, Hatfield, PA).

2.10.4. Imaging and analysis

Fluorescent images were taken using a Nikon Eclipse FN1 fluorescent microscope with a 10×, 20× or 40× water immersion Olympus objectives (LUMPlanFI/IR, 40×/0.8) and were acquired using a Photometric Cool SNAP EZ CCD camera controlled by Metaview software. Fluorescent intensity, area, length, breadth, and shape factor measurements were performed by Metamorph software using the integrated morphometry analysis function. Metamorph or ImageJ (NIH) were used for cell counting. At least 100 NMJ images were acquired for each mouse and neurofilament-200 staining was used to assess innervation. Only the gray matter of each ventral horn was used for analysis of L2–6 lumbar spinal cord staining. Only the CA1 region of the hippocampus was used for analysis. Representative images in figures were acquired using a Leica DMI4000B (Mannheim, Germany) confocal microscope using Leica 20× (HC PL APO, 20×/0.60) and 63× oil immersion (HD PL APO, 63×/1.40) lens. For TEM, 2000×, 2500×, and 5000× images were acquired with a JEOL JEM 1400 transmission electron microscope (JEOL, Peabody, MA) at 80 kV (0.35 s exposure time) on a Gatan Ultrascan 1000 CCD (Gatan, Inc., Pleasanton, CA). Mitochondrial area, perimeter, circularity, and Feret’s diameter were assessed using ImageJ. Circularity is scored from 0 to 1, with a score of 1 reflecting a perfect circle and the score decreasing as the shape becomes elongated. Feret’s diameter is the longest distance between any two points along the outlined area and can be referred to as caliper diameter (Leduc-Gaudet et al., 2015).

2.10.5. Western blotting analysis

The procedure was described as in our previous studies (Wang et al., 2016; Wang et al., 2019; Zhang et al., 2020). Briefly, cortical tissues were collected from end-stage ALS and age-matched WT mice and were homogenized in lysis buffer mixed with a protease inhibitor (Pierce Biotechnology, Rockford, IL) and phosphatase inhibitor cocktails (Sigma). The homogenized tissue was centrifuged at 13,500 g at 4 °C for 20 min and supernatant was retained. BCA assay kit (Thermo Scientific) was used to measure protein concentration. Equivalent amounts of protein from each sample were diluted with lysis buffer and boiled for 5 min. Samples were then subjected to electrophoresis in 10% SDS-polyacrylamide gels at 100 mV for 110–120 min. Gels were then transferred to polyvinylidene fluoride membranes using mixed molecular weight function of Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked for 1 h in 5% BSA in Tris-buffered saline containing (0.1%) Tween-20 (TBST). Membranes were then incubated overnight at 4 °C with primary antibodies Ms-NAMPT (1:2500, ALX-804-922-C100, Enzo Life Sciences), Ms-NMNAT1 (1:1000,sc-271557, Santa Cruz), Ms-NMNAT2 (1:1000, sc-515206, Santa Cruz), Gt-NMNAT3 (1:1000, ab121030, Abcam), Gt-ChAT (1:1000, AB144P, Abcam), Ms-PSD95 (1:1000, sc-32290, Santa Cruz), Ms-Synaptophysin (1:2500, s5768, Sigma), Ms-GluR1 (1:1000, NBP2–22399, Novus), Rab-GluR2 (1:1000, NBP2–75510), Ms-VGlut1 (1:1000, NBP2–59329, Novus), or Ms-β-actin (1:5000, sc-47778, Santa Cruz). Membranes were exposed to Clarity^™ Western ECL Substrate (Bio-Rad) and imaged with ChemiDoc^™ XRS+ system (BioRad). Precision Plus Protein^™ Dual Colour Standards (Bio-Rad) was used as the marker to evaluate the molecular size of protein bands.

2.10.6. Statistical analysis

Data were expressed as mean ± standard error of the mean (SEM). Comparisons were made using One-way ANOVA with Tukey post hoc test or unpaired Student’s t-test. Survival data was analyzed using Kaplan-Meier survival curve followed by Log-Rank test. Levels of statistical significance were set as *p < 0.05, **p < 0.01, and ***p < 0.001. Statistical analyses were performed using OriginPro 2022 and Excel.

3. Results

3.1. ALS causes reduction of NAD levels at pre-symptomatic stage

We first determined whether NAD homeostasis was altered in SOD1^G93A ALS mice. At 4 weeks of age, there was no difference in NAD in the plasma or red blood cells (RBC), however, NAD⁺ and NADH levels in the blood (both plasma and RBCs) were already reduced in pre-symptomatic (6-week-old) ALS mice and the deficit of NAD⁺ and NADH became larger at the symptom onset stage (11–12 weeks old) and end-stage (19–20 weeks old), however, the NAD⁺/NADH ratios remained unchanged at different stages (Fig. 1A and Fig. 2A), suggesting a disease-stage dependent decline in NAD levels. This progressive decline was most evident in plasma, where NAD⁺ was reduced by 4% (4 weeks old, without statistical significance) 14% (6-weeks old), 19% (11–12 weeks old), and 24% (end-stage) in ALS mice.

Fig. 1. SOD1^G93A ALS mice have reduced NAD⁺ levels in the plasma, cortex, and spinal cord.

A) Normalized plasma NAD⁺ and NADH levels and NAD⁺/NADH ratio in 4-week, 6-week (pre-symptomatic), 11–12-week (symptom on-set) and 19–20-week (end-stage) old ALS mice (N = 3–6) and age-matched WT mice (N = 5–6). C) NAD⁺ and NADH concentrations and the NAD⁺/NADH ratio in the cortex and spinal cord of 4-week-old and end-stage ALS (N = 3–7) and WT (N = 5–6) mice. D-E) Western blot images (D) and summary data (E) of NAD pathway salvage proteins from lumbar spinal cord of end-stage of ALS mice (N = 7) and age-matched WT mice (N = 6). Data were normalized to β-actin of WT mice. *p < 0.05; **p < 0.01; and ***p < 0.001. Student’s t-test. Supported by Fig. 2.

Fig. 2. NAD decline in RBCs and reduced lumbar synaptic protein expression in SOD1^G93A mice.

Normalized red blood cell NAD⁺ and NADH levels and NAD⁺/NADH ratio in 4-week, 6-week (pre-symptomatic), 11–12-week (symptom on-set) and 19–20-week (end-stage) old ALS mice (N = 3–6) and age-matched WT mice (N = 5–6). B) Western blot images (B) and summary data (C) of synaptic proteins in lumbar spinal cords of end-stage of ALS mice (N = 7) and age-matched WT mice (N = 6). Data were normalized to β-actin of WT mice. **p < 0.01 and ***p < 0.001. Student’s t-test. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Similarly, cortex and spinal cord NAD⁺ and NADH levels were not different at 4-weeks old but were reduced in end-stage ALS mice, with NAD⁺/NADH ratios remaining unchanged as compared with WT mice (Fig. 1B–C). These results demonstrate that SOD1^G93A ALS mice have impaired NAD homeostasis in circulation and CNS, but the redox balance was not affected. Along with the decline of NAD⁺ levels, the expression levels of NAD⁺ salvage pathway proteins NAMPT and NMNAT3 (the mitochondria specific isoform), were significantly reduced in spinal cord at end-stage of ALS mice (Fig. 1D–E). The NAD⁺ levels can be affected by both its synthesis and consumption; thus, the mechanism of NAD reduction should be further studied. Widespread synapse loss occurs in the spinal cord during ALS, with glutamate transport and signaling potentially being a driving factor (Rothstein et al., 1992; Baczyk et al., 2020; Nishimura and Arias, 2021). Supporting this, expression levels of important synaptic proteins were decreased in the lumbar spinal cord at the end-stage of ALS mice (Fig. 2B–C).

3.2. NMN supplementation improves healthspan and reduces motor dysfunction of SOD1^G93A ALS mice

Given that NAD levels were reduced in ALS mice, we studied whether dietary supplementation of NAD⁺ precursor NMN had any beneficial effects on disease progression in SOD1^G93A ALS mouse model. SOD1^G93A mice were fed either a diet of standard mouse chow (ALS mice) or standard mouse chow supplemented with NMN (NMN diet, ALS^+NMN mice) beginning at 8 weeks of age (Fig. 3A).

Fig. 3. NMN delayed motor and gait impairments in ALS mice.

A) Experimental timeline. B) 18-week-old WT, ALS, and ALS^+NMN mice. C–D) Survival curve (C) and mean lifespan (D) for WT (N = 10), ALS (N = 10), and ALS + NMN (N = 10). Dashed lines represent median lifespan (ALS = 138 days and ALS^+NMN = 143 days). E-H) Weekly behavior assessment for WT (N = 15–26), ALS (N = 7–17), and ALS^+NMN (N = 8–18). E) Body weight. F) Accelerated rotarod time to fall. G) Hanging wire time to drop. H) Four limb grip force. I-L) Open field measurements over 10 min period from mice at 8 and 18 weeks old. I) Total distance. J) Immobile time. K) Mean speed. L) Maximum speed. M) Representative walking gait tracks for WT, ALS, and ALS^+NMN mice at 120 days. Walking gait was assessed at 60 (pre-symptomatic), 90 (early symptomatic), and 120 (late symptomatic) days. Forepaws were painted orange and hindpaws were painted purple. N) Mean forelimb stride length for WT (N = 15–16, n = 30–32), ALS (N = 7–8, n = 14–16), and ALS^+NMN (N = 8, n = 16). O) Mean hindlimb stride width for WT (N = 15–16), ALS (N = 7–8), and ALS^+NMN (N = 8). P) Mean forelimb stride width for WT (N = 15–16, n = 30–32), ALS (N = 7–8, n = 14–16), and ALS^+NMN (N = 8, n = 16). Q) Mean hindlimb stride width for WT (N = 15–16), ALS (N = 7–8), and ALS^+NMN (N = 8). For C, E-H, N-Q, *WT vs ALS; #WT vs ALS^+NMN; $ALS vs ALS^+NMN. For I-J, *p < 0.05; **p < 0.01; ***p < 0.001. Student’s t-test. Fig. 3A, created with BioRender. Supported by Fig. 4. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

After 10 weeks of either the standard or our NMN diet, ALS mice were not paralyzed but had noticeably impaired mobility (Fig. 3B). ALS^+NMN mice had a 5-day increase in median lifespan (138 days ALS mice vs 143 days ALS^+NMN mice; Kaplan Meier, χ² = 0.089) (Fig. 3C–D), suggesting that NMN supplementation has a modest effect on lifespan extension. However, the initial mortality event for ALS^+NMN mice was delayed by nearly 2 weeks (120 days ALS mice vs 131 days ALS^+NMN mice). Nevertheless, ALS and ALS^+NMN mice had similar body weight loss during the testing period (Fig. 3E). One of the most prominent aspects of ALS is the progressive decline in motor function, therefore, we performed various motor behavior tests, starting prior to symptom onset, to determine if NMN is beneficial for motor function. ALS^+NMN mice exhibited slower motor impairment in both rotarod and hanging wire tests, with a two-week delay in dysfunction (Fig. 3F–G) however, four-limb grip force was not different between ALS and ALS^+NMN mice until 18 weeks of age (Fig. 3H). Open field testing was used to measure voluntary movement. While there was no difference in activity prior to symptom appearance, ALS^+NMN mice were more active, having increased distance travelled and mean speed, reduced immobile time, but similar maximal speed compared with ALS mice (Fig. 3I–L).

In addition to the weekly behavior tests, we also assessed walking gait by measuring stride length and stride width (Fig. 3M). We analyzed mice at three age time points corresponding to disease progression in SOD1^G93A mice: 60 (pre-symptomatic), 90 (early symptomatic), and 120 (symptomatic) days old. Both ALS and ALS^+NMN mice had altered walking gaits relative to WT mice. However, ALS^+NMN mice had improved forelimb and hindlimb stride length compared to ALS mice (Fig. 3N–O) but appeared to have similar forelimb and hindlimb stride widths (Fig. 3P–Q). The gait differences between ALS and ALS^+NMN mice started to appear around 90 days and were evident by 120 days. To determine how gait is changed during ALS progression, we compared stride length and width among the mouse groups between 60 and 90, 90–120, and 60–120 days old (Table 1). From 60 to 90 days, both WT and ALS^+NMN mice had a significant increase stride length while ALS mice was unchanged. From 60 to 120 days, ALS mice experienced a significant decrease in forelimb and hindlimb stride length (forelimb, p = 6.16 × 10⁻⁵; hindlimb, p = 1.80 × 10⁻⁵), while the forelimb and hindlimb stride lengths for ALS^+NMN mice did not change (forelimb, p = 0.38; hindlimb, p = 0.55) (Table 1). Generally, NMN diet prevented gait impairments due to ALS.

Table 1.

Test statistics for walking gait changes from 60 to 120 days old.

		P-value
		Hindlimb			Forelimb
Stride	Age Range (days)	WT	ALS	ALS+NMN	WT	ALS	ALS+NMN
	60–90	0.0144	0.4416	0.0042	0.1403	0.5950	0.0050
Length	90–120	0.2550	1.1E-05	0.0232	0.0621	1.4E-05	0.0684
	60–120	0.0014	1.8E-05	0.5480	0.0051	6.2E-05	0.3750
	60–90	0.4841	0.2622	0.4935	0.6182	0.9386	0.6822
Width	90–120	0.1651	0.6589	0.8855	0.6876	0.0839	0.2880
	60–120	0.4654	0.4325	0.5079	0.8505	0.1228	0.4839

To determine if our NMN diet increases NAD availability in mice, we placed WT mice on the NMN diet for two weeks. The NAD levels of WT mice fed the NMN diet were not increased in the cortex or plasma compared with WT mice fed with control diet (Fig. 4). However, based on our method of NMN administration (self-administered dietary supplementation), these NAD levels may not demonstrate the effects of long-term NMN supplementation (Mills et al., 2016). Overall, these results suggest that NMN supplementation can slow the progression of motor impairment, improve healthspan, and is beneficial against ALS.

Fig. 4. NAD levels in WT mice fed control diet and NMN diet.

A-B) Cortex and plasma NAD⁺ and NADH levels and NAD⁺/NADH ratio. In this experiment, adult 8-week-old WT mice were fed either control diet (N = 5) or NMN diet (N = 5) for 2 weeks. Student’s t-test.

3.3. NMN improves synaptic transmission at NMJs

ALS is a MN disease that also affects skeletal muscle by disrupting neurotransmission at NMJs. Based on the improvements to motor function in ALS^+NMN mice, we investigated how the function of NMJs were impacted by our NMN diet. We recorded evoked EPPs and mEPPs from freshly isolated semitendinosus muscles from 18-week-old WT, ALS, and ALS^+NMN mice. We recorded EPP responses at different stimulation frequencies (10, 20, 50, 100 and 200 Hz) and analyzed the amplitudes for the first five and final five responses during the first second of stimulation (Fig. 5A–F). At all frequencies, EPPs recorded from WT mice were significantly larger than those from both ALS groups, however, EPPs from ALS^+NMN mice were larger than ALS mice. At 10 and 20 Hz, ALS^+NMN mice had larger EPPs for the first and last five responses (Fig. 5B–C). At 50 Hz, ALS^+NMN mice had significantly improved EPP amplitudes for the first but not the last five responses (Fig. 5D). At the highest stimulation frequencies (100 and 200 Hz), ALS^+NMN mice again had increased amplitudes for the first and last five responses (Fig. 5E–F). To assess whether synaptic transmission at NMJs was maintained during the stimulation, we calculated the ratio of the last five EPPs to the first five EPPs. Expectedly, WT mice had a higher ratio than both ALS and ALS^+NMN mice at most frequencies. At 200 Hz, the amplitude ratio for WT and ALS^+NMN mice was not significantly different, and both were significantly higher than ALS mice (Fig. 5G). This indicates ALS^+NMN mice have an improved ability to maintain EPPs during sustained high frequency stimulation, though the response amplitude is reduced considerably compared to healthy mice.

Fig. 5. NMN improved synaptic transmission at NMJs in ALS.

A) Representative traces for 10 Hz, 20 Hz, 50 Hz, 100 Hz, and 200 Hz stimulation. B–F) Mean amplitude for first 5 and last 5 responses for 10 Hz [WT (N = 5, n = 42), ALS (N = 6, n = 50), and ALS^+NMN (N = 4, n = 26)], 20 Hz [WT (N = 5, n = 42), ALS (N = 6, n = 50), and ALS^+NMN (N = 4, n = 25)] 50 Hz [WT (N = 5, n = 41), ALS (N = 6, n = 50), and ALS^+NMN (N = 4, n = 21)], 100 Hz [WT (N = 5, n = 39), ALS (N = 6, n = 49), and ALS^+NMN (N = 4, n = 24)], and 200 Hz [WT (N = 5, n = 29), ALS (N = 6, n = 43), and ALS^+NMN (N = 4, n = 23)]. G) Ratio of the mean amplitude for the last 5 responses over the mean amplitude for the first 5 responses. H) Representative mEPP amplitude trace. I) Peak mEPP amplitude. Determined using best fitting of Gaussian function for WT (N = 5, n = 14), ALS (N = 6, n = 18), and ALS^+NMN (N = 4, n = 12) mice. J) Quantal content. Quantal content was determined dividing amplitude of first 5 responses to 10 Hz stimulation by the peak mEPP amplitude for WT (N = 5, n = 14), ALS (N = 6, n = 18), and ALS^+NMN (N = 4, n = 12) mice. K-L) Histogram and cumulative probability of mEPP amplitude. M) Representative mEPP frequency trace. Analyzed 10–15 s, with analysis starting 5–10 s after final stimulation event. N) Mean mEPP frequency (mEPP/s) for WT (N = 5, n = 14, ALS (N = 6, n = 16), and ALS^+NMN (N = 4, n = 10) mice. All mice used for recording were 18-weeks old. N, number of mice; n, number of traces analyzed. For B–F, *p < 0.05; **p < 0.01; ***p < 0.001. For G, *WT vs ALS; #WT vs ALS^+NMN; $ALS vs ALS^+NMN. One-way ANOVA with Tukey post hoc test (B-G) or Student’s t-test (I, L, N).

Spontaneous NMJ activity was determined by recording mEPPs (Fig. 5H). The amplitude of mEPPs was significantly decreased in ALS mice but restored in ALS^+NMN mice (Fig. 5I). There was also a distinct shift in the probability of small mEPPs (<0.5 mV), which were more likely in ALS mice. ALS^+NMN mice had an increased probability of moderate mEPPs (0.5–0.8 mV), but there did not appear to be any change in very large mEPPs (>1 mV) between WT, ALS, and ALS^+NMN mice (Fig. 5J–K). However, quantal content was not different between ALS and ALS^+NMN mice (Fig. 5L). ALS^+NMN mice also had improved mEPP frequency compared with ALS mice (Fig. 5M–N). These findings show that our NMN diet was able to correct deficits to spontaneous vesicle release present in ALS mice.

To assess short-term plasticity at the NMJ, we evaluated paired-pulse facilitation by applying progressively shorter intervals between pulses. At longer intervals (20–100 ms), the facilitation ratio was similar between the three conditions, though some small differences were found (Fig. 6A–E). As intervals became shorter (≤10 ms), alterations to the facilitation ratio appeared. At 8 ms, ALS^+NMN mice had a larger facilitation ratio than ALS mice, though ALS and WT mice were not different (Fig. 6F–G). At 6 ms, WT and ALS^+NMN mice both had higher facilitation ratios (Fig. 6H), and at 4 ms, ALS and ALS^+NMN mice had similarly reduced ratios (Fig. 6I) with a noticeable lack of facilitation as compared with WT mice. ALS^+NMN mice were able to maintain facilitation until a pulse interval of 6 ms while WT mice exhibited facilitation at all intervals (Fig. 6J). Notably, starting at 10 ms, ALS mice displayed larger variability in the facilitation ratio compared to WT and ALS^+NMN mice. ALS^+NMN mice only had similar response variance at 4 ms (Fig. 6I). These results indicate that ALS mice improved short-term synaptic plasticity albeit with no improvement at very short pulse intervals.

Fig. 6. NMN increased synaptic (short-term plasticity) facilitation in ALS.

A-I) Representative trace and facilitation ratio (amplitude of the second response over amplitude of the first response) for 100 ms [A, WT (N = 5, n = 38), ALS (N = 6, n = 46), and ALS^+NMN (N = 4, n = 25)], 80 ms [B, WT (N = 5, n = 38), ALS (N = 6, n = 48), and ALS^+NMN (N = 4, n = 25)], 60 ms [C, WT (N = 5, n = 38), ALS (N = 6, n = 48), and ALS^+NMN (N = 4, n = 25)], 40 ms [D, WT (N = 5, n = 38), ALS (N = 6, n = 48), and ALS^+NMN (N = 4, n = 25)],20 ms [E, WT (N = 5, n = 38), ALS (N = 6, n = 48), and ALS^+NMN (N = 4, n = 25)], 10 ms [F, WT (N = 5, n = 32), ALS (N = 6, n = 48), and ALS^+NMN (N = 4, n = 22)],8 ms [G, WT (N = 5, n = 26), ALS (N = 6, n = 45), and ALS^+NMN (N = 4, n = 19)], 6 ms [H, WT (N = 5, n = 20), ALS (N = 6, n = 45), and ALS^+NMN (N = 4, n = 16)], and 4 ms [I, WT (N = 5, n = 15), ALS (N = 6, n = 34), and ALS^+NMN (N = 4, n = 15)] pulse interval. J) Mean facilitation ratio as a function of time interval. Insert: Mean facilitation ratio from 10 ms to 4 ms interval. All mice used for recording were 18-weeks old. N, number of mice; n, number of traces analyzed. For A-I, *p < 0.05. For J, *WT vs ALS; #WT vs ALS^+NMN; $ALS vs ALS^+NMN. One-way ANOVA with Tukey post hoc test.

Overall, electrophysiological recording demonstrated that diet supplementation of NMN can enhance synaptic function at NMJs in ALS.

3.4. NMN supplementation reduces MNJ and intermyofibrillar (IMF) mitochondrial abnormalities

We observed similar body weight reduction of ALS and ALS^+NMN mice compared with WT mice (Fig. 3E), however, compared with WT mice, ALS^+NMN mice had less impairment of NMJ function (Fig. 5–6) and reduction of semitendinosus muscle mass than ALS mice (Fig. 7A–B). Because synaptic transmission at a NMJ is related to its structure, we studied the effect of NMN on NMJ morphology and innervation (Ryten et al., 2007; Wang et al., 2017; Lundt et al., 2020; Fralish et al., 2021; Zou and Pan, 2022).

Fig. 7. NMN diet reduced skeletal muscle and NMJ abnormalities.

A-B) Representative photograph of semitendinosus muscle and mean mass for 18-week-old WT (N = 6, n = 11), ALS (N = 10, n = 14), and ALS^+NMN (N = 6, n = 9) mice. C–D) Representative innervated, denervated and non-innervated NMJs stained with BTX and neurofilaments in semitendinosus muscles from WT, ALS, and ALS^+NMN mice. E-I) Morphological measures of motor endplate area (E), perimeter (F), breadth (G), length (H), and shape factor (I), for WT (N = 3, n = 375), ALS (N = 3, n = 387), and ALS^+NMN (N = 3, n = 394) mice. J) NMJ innervation ratio for WT (N = 3), ALS (N = 3), and ALS^+NMN (N = 3) mice. K-L) IMF mitochondria TEM images at 5000× (K) and 10,000× (L). M) Mitochondrial density for WT (N = 2, n = 21), ALS (N = 2, n = 21), and ALS^+NMN (N = 2, n = 21) mice. N-Q) Morphological measures for mitochondria area (N), perimeter (O), Feret’s diameter (P), and circularity (Q) for WT (N = 2, n = 558), ALS (N = 2, n = 530), and ALS^+NMN (N = 2, n = 547) mice. All mice used for analysis were 18-weeks old. N, number of mice, n, number of muscles (B), NMJs (E-J), images (M), or mitochondria (N-Q). *p < 0.05, **p < 0.01, ***p < 0.001. One-way ANOVA with Tukey post hoc test.

At NMJs, acetylcholine receptors were labeled with α-BTX and axons were stained with neurofilament (Fig. 7C–D). Both motor endplate area and perimeter were reduced in ALS and ALS^+NMN mice, but the area reduction was smaller for ALS^+NMN mice (Fig. 7E–F). Similarly, the length and breadth of the motor endplate were improved in ALS^+NMN mice (Fig. 7G–H). The endplates were more circular in both ALS groups (Fig. 7I). At a functional level, NMN significantly increased the innervation ratio of NMJs, counteracting the loss caused by ALS (Fig. 7I–J).

Growing evidence suggest mitochondria defects in ALS human patient and mouse models (Krasnianski et al., 2005; Crugnola et al., 2010; Luo et al., 2013; Smith et al., 2019), and our previous study also showed deletion of NAMPT affects mitochondrial function (Lundt et al., 2021) (our cell report); therefore, using TEM we investigated whether the NMN diet influences skeletal muscle IMF mitochondria, which are important for muscle contraction (Fig. 7K–L). IMF mitochondrial density was reduced in both ALS conditions with no improvement by NMN (Fig. 7M). However, mitochondrial area and perimeter were significantly larger in ALS mice than those in ALS^+NMN mice (Fig. 9N–O). NMN also significantly increased Feret’s diameter of mitochondria and restored mitochondria circularity in ALS mice (Fig. 7P–Q). These results indicate that our NMN diet can alleviate some of the changes to IMF mitochondria that occurs in ALS.

Fig. 9. NMN reduced astrocyte activation in lumbar spinal cord of ALS mice.

A-B) Representative Iba1 image (A) and mean intensity (B) for WT (N = 2, n = 6), ALS (N = 2, n = 6), and ALS^+NMN (N = 2, n = 6) mice. C–D) Representative GFAP image (C) and mean intensity (D) for WT (N = 2, n = 6), ALS (N = 2, n = 6), and ALS^+NMN (N = 2, n = 6) mice. N, number of mice; n, average intensity in ventral horns from one section. All mice used for analysis were 18-weeks old. *p < 0.05 and ***p < 0.001. One-way ANOVA with Tukey post-hoc test. Figure supported by Fig. 10.

3.5. NMN supplementation ameliorates ALS pathology

The loss of lower MNs from the ventral horn is a hallmark of ALS. Here we examined whether NMN can ameliorate ALS pathology in spinal cord and cortex. To determine if our NMN diet improves spinal cord MN survival, we did immunostaining of cross-sectioned lumbar spinal cords from 18-week-old mice with antibodies against choline acetyltransferase (ChAT), a marker of MNs, and NeuN, a marker for neurons (Fig. 8A). There was no difference in ChAT+ MN number between the ALS groups, with both having a 50% reduction compared with WT mice (Fig. 8B). Morphologically, the area and perimeter of the ChAT+ MNs were improved in ALS^+NMN mice (Fig. 8C–D). Similarly, NMN significantly restored the length, breadth and shape in ALS^+NMN mice (Fig. 8E–G). These analyses suggest that while ALS^+NMN mice have similar MN number to ALS mice, they have larger size of surviving MNs with morphologies more similar to those in WT mice.

Fig. 8. NMN improved spinal cord MN and mitochondrial morphology in ALS^+NMN mice.

A) Representative spinal cord image for WT, ALS, and ALS^+NMN mice. B) ChAT+ MN density for WT (N = 2, n = 21), ALS (N = 2, n = 21), and ALS^+NMN (N = 2, n = 24) mice. C-G) Morphological measures of MN area (C), perimeter (D), length (E), breadth (F), and shape factor (G), for WT (N = 2, n = 316), ALS (N = 2, n = 235), and ALS^+NMN (N = 2, n = 226). H–I) TEM image of ventral horn mitochondria at 5000× (H) and 10,000× (I). J) Mitochondrial density for WT (N = 2, n = 60), ALS (N = 2, n = 70), and ALS^+NMN (N = 2, n = 44) mice. K–N) Morphological analyses for mitochondria area (D), perimeter (E), Feret’s diameter (F), and circularity (H) for WT (N = 2, n = 475), ALS (N = 2, n = 334), and ALS^+NMN (N = 2, n = 384) mice. All mice used for analysis were 18-weeks old. N, number of mice; n, number of sections (B), MNs (C-G), images (J), or mitochondria (K–N). *p < 0.05, **p < 0.01, ***p < 0.001. One-Way ANOVA with Tukey post hoc test.

To further investigate any changes to spinal cord MNs, we assessed the mitochondrial structure using TEM (Fig. 8H–I). Mitochondrial density was significantly reduced in ALS mice but restored in ALS^+NMN mice (Fig. 8J). ALS mice had an increase in mitochondria area, but this was modestly corrected by the NMN diet (Fig. 8K). Mitochondrial perimeter and Feret’s diameter were returned to normal length in ALS^+NMN mice (Fig. 8L–M), but there was no difference in mitochondrial shape between ALS and ALS^+NMN groups (Fig. 8N). Overall, these results indicate that spinal cord mitochondrial density and structure are restored to normal in ALS^+NMN mice.

Though primarily associated with MNs, ALS is a non-cell autonomous disease with glial cells, especially astrocytes, contributing to neuroinflammation and MN toxicity (Van Harten et al., 2021). Thus, we determined whether our NMN diet had any effect on glial cell activation in lumbar spinal cords by immunostaining of Iba1, a marker of microglia, and GFAP, a marker of astrocytes (Fig. 9A–B). We found that both Iba1 and GFAP signals were increased in the ALS groups (Fig. 9C–D); however, the ALS^+NMN mice had a significant reduction in GFAP signal compared to ALS mice (Fig. 9D). This suggests that NMN can reduce astrocyte activation in the lumbar spinal cord.

Finally, we examined whether our NMN diet had any benefits to the brain. In the motor cortex, there was no difference in NeuN intensity between the three groups but ALS^+NMN mice had an increase in NeuN+ cell number (Fig. 10A–C). ALS and ALS^+NMN mice had similar GFAP intensities that were significantly lower compared to WT mice and no difference in GFAP+ cell density was found among 3 groups (Fig. 10D–F). There was no difference in Iba1 intensity, though ALS^+NMN mice had a significant reduction in Iba + cells compared to ALS mice (Fig. 10G–I). In the hippocampus, GFAP intensity and the density of GFAP+ astrocytes were significantly higher in both ALS and ALS^+NMN mice than in WT mice, but there was no difference between ALS and ALS^+NMN mice (Fig. 10J–L). There was no difference in Iba1 intensity and both ALS conditions had significantly more Iba + cells (Fig. 10M–O). Glial cells in the brains of ALS mice do not appear to be strongly impacted by our NMN diet.

Fig. 10. Glial cell activation in motor cortex and hippocampus of ALS^+NMN mice.

A-C) Representative NeuN images (A) and mean NeuN intensity (B) and NeuN+ density (C) in motor cortex. D–F) Representative motor cortex GFAP images (D), mean GFAP intensity (E) and density of GFAP+ astrocytes (F). G-I) Representative motor cortex Iba1 images (G), mean Iba1 intensity (G) and density of Iba1+ microglia (H). J-L) Representative CA1 hippocampal GFAP images (J), mean GFAP intensity (K) and density of GFAP+ astrocytes (L). M-O) Representative CA1 hippocampal Iba1 images (M), mean Iba1 intensity (N) and density of Iba1+ microglia (O). For staining, WT (N = 2, n = 5–12), ALS (N = 2, n = 6–11), and ALS^+NMN (N = 2, n = 10). All mice used for analysis were 18-weeks old. N, number of mice; n, average from 1 slice. **p < 0.01. One-way ANOVA with Tukey post-hoc test.

4. Discussion

In this study, we investigated the effect of dietary NMN on the SOD1^G93A ALS mouse model. Our findings demonstrated that NMN delayed motor and gait impairments that occur as ALS progresses and modestly extended lifespan of ALS mice. Importantly, NMJs, one of the earliest affected sites in ALS, benefited greatly from our NMN diet. ALS^+NMN mice had improved neurotransmission, motor endplate morphology, and innervation at NMJs. Skeletal muscle atrophy and mitochondrial morphology were partially restored in ALS by NMNs. MN survival was not affected, though MNs were larger in ALS^+NMN mice. Glial cell activation was reduced in ventral horn in ALS^+NMN mice compared with ALS mice. Thus, NMN supplementation to ALS mouse diet can exert various beneficial effects against ALS, indicating NMN supplementation can significantly improve healthspan of ALS mice.

Prior studies reported a loss of blood NAD at end-stage in ALS mice (Roderer et al., 2018; Obrador et al., 2021). NAD loss could potentially be a critical early event that precedes physical symptoms, with worsening NAD⁺ deficits driving ALS disease during on-set and progression. Impairments to NAD homeostasis are due to either decreased biosynthesis or increased consumption. In ALS patient spinal cords, NAMPT and NMNAT2 expression are significantly altered (Wang et al., 2017; Harlan et al., 2020). It is also possible that increased acetylation of NAMPT might reduce NAMPT activity, as reported in aged mice (Yoshino et al., 2011; Yang et al., 2023). Increased NAD consumption by PARP1 or CD38 could also cause NAD⁺ reduction in ALS as inhibition or knockout of PARP1 or CD38 can nearly restore NAD⁺ levels (Camacho-Pereira et al., 2016; Zha et al., 2018). Interestingly, mutations to SARM1, an NADase controlling axonal degradation, that result in increased activity are more common in ALS human patients (Gilley et al., 2021; Bloom et al., 2022). The mechanism by which NAD homeostasis is disrupted in ALS warrants further study.

How impactful NAD⁺ precursors are to survival appears variable in different studies. NAM and NR have each demonstrated modest to significant increases on survival in ALS mice (Blacher et al., 2019; Harlan et al., 2020; Zhou et al., 2020; Obrador et al., 2021). A role for NAD⁺ availability and healthspan is evident, both against normal aging and disease. NAD⁺ declines naturally with age but augmenting the NAD salvage pathway can improve motor and metabolic activity in aged mice (Mills et al., 2016; Yoshida et al., 2019). Administration of a single dose of NMN by oral gavage to mice can increase NAD⁺ levels (Wang et al., 2017). However, NMN is quickly converted into NAD⁺ making it challenging to assess the direct impact of NMN on NAD⁺. As such, supplementing mice with NMN in diet or drinking water, where the mice receive a small amount of NMN over time, rather than administering a single NMN dose may not produce a significant increase to NAD⁺ levels in most tissues but rather a constant slight elevation in circulating NAD⁺ (de Picciotto et al., 2016; Mills et al., 2016; Ru et al., 2022). The lack of an observable increase in NAD⁺ levels from the NMN diet could potentially explain the modest effect on lifespan and minor impacts on spinal cord motor neurons. Increased NAD⁺ availability improves motor performance against neurodegenerative diseases (Hathorn et al., 2011; Wang et al., 2017; Schondorf et al., 2018; Lundt et al., 2020). Motor dysfunction can be an indicator of reduced healthspan (Fang et al., 2016; Mitchell et al., 2018). Other NAD⁺ precursor metabolites, NR and NAM, were also shown to slow motor decline (Blacher et al., 2019; Zhou et al., 2020; Obrador et al., 2021). Moreover, ALS mice administered P7C3, a purported NAMPT activating molecule, not only exhibited improved rotarod performance but also had preservation of walking gait (Tesla et al., 2012). How improving NAD homeostasis affects other healthspan-related loss during ALS requires more study. Additionally, while NMN is a precursor of NAD⁺, no observed effect on NAD⁺ levels could suggest that the benefits from NMN could be a mechanism different from NAD⁺ homeostasis.

Electrophysiological recording of EPP is a gold standard approach to assess NMJ function (Arbour et al., 2015; Orr et al., 2020), however, how NMJ function is affected by NAD⁺ precursors in ALS has not been reported. We found ALS mice had reduced evoked and spontaneous EPP activity indicating both pre- and post-synaptic alterations, which was expected based on previous findings, (Rocha et al., 2013; Tremblay et al., 2017; Chand et al., 2018) and that NMN significantly improved NMJ activity. Potential post-synaptic changes could be alterations to nAChR density, motor end-plate disorganization (i.e., endplate fragmentation, loss of innervation, etc), and/or muscle fiber diameter (Rocha et al., 2013). Muscle fiber type and diameter should also be investigated. There is strong evidence indicating problems with the release and cycling of synaptic vesicle in the NMJ during ALS (Cappello et al., 2012; Chand et al., 2018; Butti et al., 2021). NMN can improve the SV cycle in NMJs in mice undergoing significant neurodegeneration (Lundt et al., 2020). More investigation into the SV cycle at NMJs of ALS mice is needed. Overall, our study increases the known benefits of NAD⁺ precursors on ALS pathology because previous studies on ALS mice did not investigate NMJs (Harlan et al., 2020; Zhou et al., 2020; Obrador et al., 2021).

The effects of NMN alone on the spinal cord may be limited. As such, pairing NMN with other interventions, such as the inhibitors of NAD⁺-consumers, may elicit stronger benefits. It is also important to note that we only investigated the benefits of dietary NMN in the SOD1^G93A ALS mouse model, whether NMN has similar effects on other ALS mouse models, such as TDP-43 or C9orf72, should be studied (Alhindi et al., 2022). SOD1 mutations comprise a small percentage of total ALS cases, as such, whether NMN has a similar impact on the majority of ALS cases is unknown (Mejzini et al., 2019). Generally, more investigation is needed to understand the best manner to incorporate NMN into an ALS treatment regimen. In summary, our findings demonstrate that NMN supplementation in diet can modestly extend lifespan and significantly enhance healthspan by improving motor performance and NMJ function in ALS mice. The current study provides new insights into the role of NMN in the development of ALS.

Data availability

Data will be made available on request.