Background: Skeletal muscles play an important role in systemic glucose homeostasis and are purported to be the origin of the altered metabolic state observed in amyotrophic lateral sclerosis (ALS). Objective: The purpose of this study was to evaluate whole-body and muscle-specific glucose metabolism in the SOD1-G93A mouse model of ALS. Methods: We assessed glucose tolerance in early-, middle-, and late-stage SOD1-G93A and control mice using an intraperitoneal glucose tolerance test. We then measured the respiratory exchange ratio (CO2 production/O2 consumption) as a function of fasting and feeding using indirect calorimetry in a subset of male mice at these time points. Finally, muscles from all mice were harvested to evaluate basal and insulin-stimulated glucose transport in fast- and slow-twitch muscles. Results: No changes in systemic glucose clearance were observed in SOD1-G93A mice at any stage, nor were there changes in fasting insulin levels. Indirect calorimetry revealed an increase in the respiratory exchange ratio during the fed state at middle, but not at early or late stages of disease. Middle-stage SOD1-G93A mice exhibited decreased insulin-stimulated glucose uptake in fast-twitch, but not slow-twitch, skeletal muscle. Late-stage SOD1-G93A mice exhibited decreased insulin-stimulated glucose uptake in both fast- and slow-twitch muscle, as well as increased basal (non-insulin-stimulated) glucose uptake. Conclusions: These results suggest that alterations in muscle metabolism occur in a fiber-type-specific manner in ALS, but do not necessarily lead to whole-body metabolic changes in SOD1-G93A mice.
Amyotrophic lateral sclerosis (ALS) is a progressive degenerative disease characterized by muscle weakness, paralysis, and eventual death. ALS affects upper and lower motor neurons, with marked weakness/atrophy, fasciculations, and reflex abnormalities in the innervated muscles occurring secondary to motor neuron degeneration . Recent evidence calls for a wider perspective on this multisystem disease, however. Some of the earliest signs of disease are seen in skeletal muscles long before the occurrence of motor neuron loss, including chronic weakness, neuromuscular junction pathology, and changes in gene regulation [2,3,4,5,6,7,8]. Although a causative role in early disease processes is controversial [9,10], skeletal muscles appear to be an early target of the disease. Thus, closer investigation of muscle function as it relates to ALS symptom onset and progression is warranted.
One potential target for investigation of skeletal muscles in motor neuron disease is peripheral glucose regulation. Skeletal muscles are the peripheral tissue responsible for the majority of glucose disposal in the body and thus play an important role in systemic glucose homeostasis . Defects in skeletal muscle glucose uptake are associated with insulin resistance. It follows that a disease that affects skeletal muscles such as ALS does will also affect glucose regulation and metabolism. While some studies document normal glucose tolerance and insulin secretion in ALS patients [12,13,14,15,16,17], others report impairment [18,19,20,21,22,23,24,25]. Recent studies have also reported hypermetabolism in patients with ALS [26,27]. Due to these discrepancies, it is important to determine the metabolic status of skeletal muscles in ALS, at different stages of disease, in order to determine whether drugs that increase muscle metabolism are appropriate.
The most common animal models of ALS are mice that express mutations in the gene coding for superoxide dismutase (SOD1). These models emerged after discovery of SOD1 mutations that accounted for a majority of the known familial forms of ALS [28,29]. Although the mechanisms are unclear, these mutations are believed to result in a toxic ‘gain of function' that leads to the ALS phenotype. Using SOD1 transgenic mice, Dupuis et al.  reported altered metabolic activity in these models of ALS. Most of their findings involved SOD1-G86R mice. The purpose of our study was to evaluate whole-body and muscle-fiber-type-specific glucose metabolism in the more commonly studied SOD1-G93A mutant mouse . We assessed whole-body metabolism and glucose tolerance in early (weanling)-, middle (presymptomatic)-, and late-stage (symptomatic) SOD1-G93A mice and age-matched wild-type littermates using a variety of measures. Although we did not confirm symptom onset in these animals, we selected ages based on numerous prior studies in this consistent and well-characterized model, including our own previous study . Because denervation in ALS occurs in a muscle-fiber-type-specific pattern (beginning with fast-twitch type II fibers and progressing to slow-twitch type I fibers) [5,7,32] we compared slow-twitch (soleus) and fast-twitch (extensor digitorum longus, EDL) fiber types. We are aware of no studies that have examined muscle-fiber-type-specific glucose uptake in the context of whole-body metabolism in ALS.
Materials and Methods
A breeding colony was established with 2 male B6SJL-Tg (SOD1-G93A)1Gur/J (SOD1-G93A) and 8 female noncarrier (control) mice obtained from the Jackson Laboratory. The offspring of these breeders served as subjects in this experiment. Tail clips were sent to Transnetyx, Inc., for genotyping. Mice were housed in groups in an animal care facility approved by the Association for Assessment and Accreditation of Laboratory Animal Care International with free access to food and water, and were maintained on a 12-hour-light/12-hour-dark cycle. Mice were evaluated at 3 different stages: early-stage mice were aged 30-40 days; middle-stage mice were aged 65-75 days, and late-stage mice were aged 105-115 days. All procedures were approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee and were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals.
Body weight was recorded and systemic glucose tolerance was determined using an intraperitoneal glucose tolerance test in female and male early-, middle-, and late-stage control and SOD1-G93A mice (n = 12 control and n = 12 SOD1-G93A mice at each stage). Mice were fasted overnight (∼12 h) prior to testing. Body weight was recorded and basal glucose values were obtained from tail blood using a glucometer (Accucheck Active). The glucose load (60% glucose in 0.9% saline; 2 g glucose/kg i.p.) was administered and glucose values were obtained 15, 30, 60, 90, and 120 min after injection.
Fasting insulin levels were evaluated in a subset of female and male early-, middle-, and late-stage control and SOD1-G93A mice using a mouse insulin ELISA kit (Alpco Diagnostics; n = 6 control and n = 6 SOD1-G93A mice at each stage). Plasma was prepared from blood obtained during cardiac puncture and insulin levels were determined spectrophotometrically, according to the manufacturer's directions. The homeostasis model assessment of insulin resistance (HOMA-IR ) was calculated in these mice using the following equation: [fasting blood glucose (mg/dl) × fasting plasma insulin (ng/ml)]/405.
Oxygen consumption, carbon dioxide production and body temperature were evaluated in a subset of age-matched male early-, middle-, and late-stage control and SOD1-G93A mice using an open-circuit indirect calorimetric system (Columbus Instruments; n = 4 control and n = 4 SOD1-G93A mice at each stage). Mice were placed in individual metabolic chambers for 24-hour test sessions. The light/dark cycle was maintained during testing and test sessions were initiated at the beginning of the dark portion of the cycle. Mice were fasted during the first 12 h of testing. Food pellets were inserted for the remaining 12 h. Baseline respiratory exchange ratio (VCO2/VO2), O2 consumption, CO2 production, and body temperature were analyzed during the final seven 4-min data points prior to feeding. For statistical analysis of the shift from basal to carbohydrate metabolism following feeding, the time series was truncated to include data from 28 min (7 data points, 4 min/point) before the onset of the rise in the function to the end of the data series. The mean of the seven data points was computed for each animal and subtracted from each individual's data. All the results from animals in each group were combined and fit to the following function,
The parameter a is a value where the function saturates, b is the time constant (i.e. the time it takes for the function to rise to 63% of its maximum value), and x0 is the shift of the function along the abscissa. The function was fit using the Marquardt-Levenberg algorithm supplied in Sigmaplot 11.0. Student's t test was used for between-group comparisons of the means and standard errors of the means for a and b. In addition to the above measures, measurements of food pellet weight and water volume were performed before and after the sessions to monitor food and water consumption during the test sessions. Body weights were also recorded before and after the sessions.
Skeletal muscle glucose transport was evaluated in the EDL (fast-twitch) and soleus (slow-twitch) muscles from the mice used for the insulin ELISA and HOMA-IR. Mice were fasted overnight (∼12 h) prior to tissue harvest. As previously done in our laboratory , mice were deeply anesthetized with sodium pentobarbital (Nembutal, 50 mg/kg) and muscles were dissected and allowed to recover in Krebs-Henseleit buffer for 1 h. Muscles were then incubated in the presence or absence of insulin (1 mU/ml) in flasks containing oxygenated Krebs-Henseleit buffer and 2[1, 2-3H]-deoxyglucose (1.5 µCi/ml) and [14C] mannitol (0.2 µCi/ml). Intracellular 2-deoxyglucose (3H DPM) and extracellular space (14C DPM) were quantified in a scintillation counter.
Body weights were analyzed separately for female and male mice. In the females, body weight increased as a function of age group (F2, 32 = 32.5249, p < 0.0001; fig. 1a). There were no differences in body weight between SOD1-G93A and controls at any stage, nor did the increases in body weight differ between the two groups. Like the females, there was a significant increase in body weight with age in the males (F2, 38 = 106.078, p < 0.0001). Unlike the females, SOD1-G93A male mice weighed less than their wild-type littermates at each stage, leading to a main effect for genotype (F1, 38 = 35.6902, p < 0.0001; fig. 1b). Body weight increases were similar between the SOD1-G93A and wild-type mice at these time points.
Indirect Calorimetric Metabolic Activity
SOD1-G93A mice and controls did not differ with regard to the respiratory exchange ratio during the prefed state at any disease stage. Upon feeding, middle-stage SOD1-G93A mice achieved a significantly higher ratio of CO2 produced/O2 consumed (t282 = 13.47, p < 0.0001), indicating greater carbohydrate metabolism in this group (fig. 2). In addition to their greater maximal ratio, the rate of the increase from 0.7 to 1.0 (indicating a feeding-related shift from fat to carbohydrate metabolism) was significantly greater in SOD1-G93A mice during the middle stage (t282 = 3.08, p < 0.05). Metabolic test sessions led to essentially no changes in body weight (-0.67 ± 0.2 g/session). No significant differences between SOD1-G93A mice and controls were revealed for O2 consumption, CO2 production, body temperature, food or water consumption, or differences in body weight, during either the fasted or fed state at any stage (data not shown).
Fasting Glucose, Glucose Clearance, Fasting Insulin and HOMA-IR
No significant differences between SOD1-G93A and wild-type male and female mice were revealed for fasting blood glucose in early, middle, or late stages (fig. 3a-c: time point 0). Likewise, glucose clearance did not differ between SOD1-G93A and controls at any disease stage (fig. 3a-c: time points 15-120 min). No significant differences in fasting insulin (fig. 3d) or in HOMA-IR (data not shown) were revealed at early, middle or late stages, or between male and female mice, suggesting that fasting serum insulin levels and insulin function in early-, middle-, and late-stage SOD1-G93A mice are similar to that of control mice.
Muscle Glucose Uptake
In the presence of insulin, 2-deoxyglucose uptake was increased in the EDL at the early (F1, 20 = 14.11, p = 0.001), middle (F1, 15 = 13.82, p < 0.01), and late stages (F1, 16 = 4.46, p = 0.05; fig. 4) of male and female mice. Post hoc comparisons (t tests) revealed that a significant increase in 2-deoxyglucose uptake occurred in both control and SOD1-G93A EDL muscles at the early stage, but only in control EDL muscles at the middle and late stages (all p < 0.05). In addition, basal 2-deoxyglucose uptake was significantly increased in late-stage SOD1-G93A EDL muscles (fig. 4e). In the soleus, 2-deoxyglucose uptake was increased in the presence of insulin at the early (F1, 16 = 21.32, p < 0.001), middle (F1, 15 = 19.55, p < 0.001), and late stages (F1, 17 = 9.25, p < 0.01). Post hoc comparisons (t tests) revealed that the significant increase in insulin-stimulated 2-deoxyglucose uptake occurred in both control and SOD1-G93A soleus muscles at the early and middle stages (all p < 0.05), but only in control soleus muscles at the late stage.
Overall, our results indicate that despite decreased insulin-stimulated glucose uptake in skeletal muscles in middle and late stages of disease, SOD1-G93A mice do not exhibit alterations in systemic glucose metabolism. Our results also show that changes in glucose uptake occur in a fiber-type-specific manner, with changes occurring first in fast-twitch skeletal muscles and then in slow-twitch muscles. SOD1-G93A mice also exhibited increased carbohydrate metabolism following feeding during the middle stage measured by indirect calorimetry. These findings indicate that SOD1-G93A mice do not exhibit systemic metabolic deficits, but that fiber-type-specific changes in neuromuscular function lead to altered glucose uptake in skeletal muscles.
Our findings regarding the effects of the SOD1-G93A mutation on body weight in mice are consistent with previous preclinical studies, which report weight changes at widely variable points during the course of the disease [8,31,35,36,37]. Decreased body weight in this model has been attributed to several factors, including muscle wasting , increased metabolism , and decreased food intake due to orolingual motor deficits [31,39,40]. We and others have reported evidence of orolingual motor deficits in this model, but our findings do not support decreased food intake as an explanation for weight loss in these animals. Our results also do not support increased metabolism as a contributing factor. Although we did not measure muscle mass, our findings regarding decreased glucose uptake in both muscle types at end stage are consistent with muscle wasting, however.
Systemic Glucose Metabolism
Muscle wasting is associated with the development of systemic insulin resistance . However, studies of glucose tolerance in ALS patients report conflicting results [12,13,14,15,16,17,18,19,20,21,22,23,24,25]. The choice of control groups may have accounted for some of the inconsistencies in these studies. One well-controlled study compared a healthy with a neuromuscular-disease control group, and reported significant insulin resistance in ALS patients that could not be explained by a decrease in functioning muscle mass . A more recent study confirmed that a subset of ALS patients exhibited impaired systemic glucose clearance that was unrelated to muscle atrophy .
Altered systemic metabolism and glucose homeostasis are recapitulated in SOD1-G86R mice . Interestingly, however, these animals exhibit increased whole-body metabolism and augmented systemic glucose clearance rather than impairment. Despite similar neuromuscular phenotypes in SOD1-G86R and SOD1-G93A mice , we found no significant differences in systemic glucose clearance in early-, middle-, or late-stage SOD1-G93A mice when they were compared to age-matched controls. In addition (and also unlike the SOD1-G86R mice ), we measured no differences in fasting glucose or fasting insulin levels, or in the HOMA-IR assessment of insulin resistance, at any stage examined. Although our finding of increased carbohydrate metabolism in 75-day-old SOD1-G93A mice measured using indirect calorimetry was consistent with the findings of Dupuis et al.  in similarly aged SOD1-G93A mice, these authors reported increased metabolism at subsequent ages in SOD1-G86R mice. It is unclear what accounts for the metabolic differences between the G93A and G86R mutants because direct comparisons between them are few. Disease onset and duration are similar for these two models, but other factors have been reported to differ [30,35]. While these differences merit further study, our results indicate that, unlike SOD1-G86R mice, systemic glucose metabolism in SOD1-G93A mice is maintained at levels similar to controls throughout the course of disease.
Muscle Denervation, Oxidative Stress, and Glucose Uptake
Muscle denervation drastically increases the production of reactive oxygen species (ROS) . Through the activation of serine/threonine kinases, oxidative stress negatively regulates the insulin signaling cascade . In addition, oxidative stress alters the transcription of the glucose transporters responsible for basal and insulin-stimulated glucose uptake (GLUT1 and GLUT4, respectively), resulting in increased basal glucose uptake, decreased insulin-stimulated glucose uptake, and the exacerbation of oxidative stress in the cell . In accordance with this, experimental muscle denervation decreases GLUT4 mRNA and impairs GLUT4 translocation, thereby reducing insulin signaling and altering glucose uptake [44,45].
ALS is characterized by oxidative stress, and the accumulation of ROS is believed to play a central role in disease pathogenesis . Markers of oxidative stress and the activity of antioxidant enzymes are increased in SOD1-G93A muscles [41,46,47]. Importantly, a recent study shows a strong ROS-triggered upregulation of Rad (Ras-related associated with diabetes) protein in skeletal muscle from sporadic ALS patients as well as asymptomatic SOD1-G93A and SOD1-G86R mice . This protein may negatively regulate glucose uptake in muscle and fat cells . Muscle denervation in ALS occurs in a functional-fiber-type-specific manner [5,7,32]. Synapses on type IIb fibers (fast twitch) undergo denervation first, followed by those on type IIa fibers (fast twitch). Synapses on type I fibers (slow twitch) are resistant until end stage. Evidence of oxidative stress in SOD1-mutant muscle is apparent before signs of denervation .
Accordingly, we found that SOD1-G93A mice exhibit alterations in skeletal muscle glucose uptake. By middle stage, these animals exhibited decreased insulin-stimulated glucose uptake in fast-twitch muscle. By late stage, insulin-stimulated glucose uptake was decreased in both fast- and slow-twitch muscles. This mimics the pattern of fiber-type-specific denervation described above [5,7,32]. In addition, an increase in basal, non-insulin-stimulated glucose uptake was apparent in late-stage fast-twitch muscles. Overall, our results suggest that while systemic glucose regulation is maintained in SOD1-G93A mice, disturbances occur in skeletal muscle glucose metabolism. Further study is required to determine the importance of these fiber types and their differential vulnerabilities in ALS-related metabolic disease.
Varying effects of different SOD1 mutations are documented in the literature [50,51]. Some of these differences extend to humans with familial, SOD1-related disease, underscoring the importance of the role of mutation type in disease pathogenesis [41,52]. ALS-related metabolic disturbances are still not fully understood, and evidence exists for differential effects of SOD1 mutations in altered glucose metabolism in ALS. Our studies of systemic and peripheral glucose metabolism in SOD1-G93A mice found that these animals developed fast-twitch muscle insulin resistance by 75 days of age. By end stage, both fast- and slow-twitch muscles were affected. In contrast, 60- to 80-day-old SOD1-G86R mice exhibit augmented systemic glucose metabolism, increased glucose uptake in skeletal muscle and white adipose tissue, and a hypermetabolic state . Due to the equivocal state of the clinical literature regarding metabolism and ALS, any judgment regarding which rodent model best replicates metabolic function in the human disease is premature. These discrepancies highlight the need for further investigation into relationships between motor neuron disease, and whole-body and muscle metabolism.
This work was supported by NIH grants AG023549 and AG026491 to J.A.S., and by the Kansas Intellectual and Developmental Disabilities Research Center (HD02528).