Metformin treatment has no beneficial effect in a dose-response survival study in the SOD1(G93A) mouse model of ALS and is harmful in female mice

Abstract

Background: Amyotrophic Lateral Sclerosis (ALS) is a devastating neurological disorder characterized by selective degeneration of upper and lower motor neurons. The primary triggers for motor neuron degeneration are unknown but inflammation, oxidative stress and mitochondrial defects have been identified as potential contributing factors. Metformin is an anti-type II diabetes drug that has anti-inflammatory and anti-oxidant properties, can bring about mitochondrial biogenesis and has been shown to attenuate pathology in mouse models of Huntington's disease and multiple sclerosis. We therefore hypothesized that it might increase survival in the SOD1(G93A) murine model of ALS.

Methodology/principal findings: Treatment of male and female SOD1(G93A) mice (n = ≥6 per sex) with 2 mg/ml metformin in the drinking water from 35 days, resulted in a significant increase in motor unit survival, as measured by in vivo electrophysiology at 100 days, in male EDL muscles (24+/-2 vs. 14+/-2 motor units, p<0.005) and female TA muscles (21+/-1 vs. 15+/-2 motor units, P = 0.0134). We therefore continued to test the effect of 0.5, 2 and 5 mg/ml metformin in the drinking water from 35 days on disease onset and progression (identified by twice weekly determination of weight and neurological score) as well as survival in male and female SOD1(G93A) mice (n = ≥14 per sex). Results for all groups were compared using Kaplan-Meier time to event analyses. In this survival study, metformin was unable to reduce pathology at any dose and had an unexpected dose-dependent negative effect on the onset of neurological symptoms (P = 0.0236) and on disease progression (P = 0.0362) in female mice.

Conclusions/significance: This study suggests that metformin is a poor candidate for clinical trial in ALS patients and that the possibility of harmful effects of metformin in female ALS patients with type II diabetes should be investigated.

Figure 1. Effect of metformin treatment on motor unit number in male and female SOD1^G93A mice.

Representative traces for in vivo motor unit assessment protocol (A) untreated SOD1^G93A male Extensor Digitorum Longus (EDL) muscle and (B) metformin treated SOD1^G93A EDL muscle at 100 days of age. (C) Graph showing mean number of motor units surviving in the EDL muscles of control (SOD1) and metformin treated (SOD1 tx.) SOD1^G93A mice at 100 days of age (SOD1^G93A male vs. metformin treated SOD1^G93A male, P = 0.005 (**)), (SOD1^G93A female vs. metformin-treated SOD1^G93A female, P = 0.056) (D) graph showing mean number of motor units surviving in the Tibialis Anterior (TA) muscles of control (SOD1) and metformin treated (SOD1 tx.) SOD1^G93A mice at 100 days of age (SOD1^G93A female vs. metformin-treated SOD1^G93A female, P = 0.0134 (*)). Error bars in (C) and (D) represent standard error of the mean (SEM). n = 7 SOD1 males, n = 6 SOD1 females, n = 9 metformin treated SOD1 males and n = 15 metformin treated SOD1 females.

Figure 2. Effect of metformin on body weight, disease onset and disease progression in male SOD1^G93A mice.

(A) Mean group body weight plotted for male SOD1^G93A mice over time from 35 days of age until the date at which the last death within each individual group occurred. As mice within groups died at different ages, the final weights of all mice were carried forward when calculating group mean values, until the date at which the last mouse within each group died. This prevented decomposition of the mean at later stages. Error bars represent standard error of the mean (SEM). (B) Kaplan-Meier survival plot for time to peak weight (an indicator of disease onset) in male SOD1^G93A mice. (C) Kaplan-Meier survival plot for time from peak weight to end stage (an indicator of disease progression) in male SOD1^G93A mice. Graphs (A), (B) and (C) present data for male SOD1^G93A mice in all four experimental groups: Control = green, 0.5 mg/ml metformin = yellow, 2 mg/ml metformin = blue, 5 mg/ml metformin = red). Results for statistical analyses performed on data presented in (B) and (C) are given in Table 1.

Figure 3. Effect of metformin on neurological score, neurological disease onset and survival in male SOD1^G93A mice.

(A) Mean group combined neurological score plotted for male SOD1^G93A mice over time from 41 days of age until the date at which the last death within each group occurred. Combined neurological score was calculated for each mouse at each time point through summation of the neurological scores recorded for its left and right hindlimbs. As mice within groups died at different ages, the final combined neurological scores of all mice were carried forward when calculating group mean values, until the date at which the last mouse within each group died. This prevented decomposition of the mean at later stages. Error bars represent standard error of the mean (SEM). (B) Kaplan-Meier survival plot for the time taken for male SOD1^G93A mice to reach a neurological score of 2 in both hindlimbs (an indicator of the onset of symptomatic neurological disease). (C) Kaplan-Meier survival plot for the time taken for male SOD1^G93A mice to reach the humane end stage of attaining a neurological score of 4 (representative of survival). Graphs (A), (B) and (C) present data for male SOD1^G93A mice in all four experimental groups: Control = green, 0.5 mg/ml metformin = yellow, 2 mg/ml metformin = blue, 5 mg/ml metformin = red). Results for statistical analyses performed on data presented in (B) and (C) are given in Table 1.

Figure 4. Effect of metformin on body weight, disease onset and disease progression in female SOD1^G93A mice.

(A) Mean group body weight plotted for female SOD1^G93A mice over time from 35 days of age until the date at which the last death within each individual group occurred. As mice within groups died at different ages, the final weights of all mice were carried forward when calculating group mean values, until the date at which the last mouse within each group died. This prevented decomposition of the mean at later stages. Error bars represent the standard error of the mean (SEM). (B) Kaplan-Meier survival plot for time to peak weight (an indicator of disease onset) in female SOD1^G93A mice. (C) Kaplan-Meier survival plot for time from peak weight to end stage (an indicator of disease progression) in female SOD1^G93A mice. Graphs (A), (B) and (C) present data for female SOD1^G93A mice in all four experimental groups: Control = green, 0.5 mg/ml metformin = yellow, 2 mg/ml metformin = blue, 5 mg/ml metformin = red). Results for statistical analyses performed on data presented in (B) and (C) are given in Table 2.

Figure 5. Effect of metformin on neurological score, neurological disease onset and survival in female SOD1^G93A mice.

(A) Mean group combined neurological score plotted for female SOD1^G93A mice over time from 41 days of age until the date at which the last death within each group occurred. Combined neurological score was calculated for each mouse at each time point through summation of the neurological scores recorded for its left and right hindlimbs. As mice within groups died at different ages, the final combined neurological scores of all mice were carried forward when calculating group mean values, until the date at which the last mouse within each group died. This prevented decomposition of the mean at later stages. Error bars represent standard error of the mean (SEM). (B) Kaplan-Meier survival plot for the time taken for female SOD1^G93A mice to reach a neurological score of 2 in both hindlimbs (an indicator of the onset of symptomatic neurological disease). (C) Kaplan-Meier survival plot for the time taken for female SOD1^G93A mice to reach the humane end stage of attaining a neurological score of 4 (representative of survival). Graphs (A), (B) and (C) present data for female SOD1^G93A mice in all four experimental groups: Control = green, 0.5 mg/ml metformin = yellow, 2 mg/ml metformin = blue, 5 mg/ml metformin = red). Results for statistical analyses performed on data presented in (B) and (C) are given in Table 2.