Metabolic Profile and Pathological Alterations in the Muscle of Patients with Early-Stage Amyotrophic Lateral Sclerosis

Abstract

Diverse biomarkers and pathological alterations have been found in muscle of patients with Amyotrophic lateral sclerosis (ALS), but the relation between such alterations and dysfunction in energetic metabolism remains to be investigated. We established the metabolome of muscle and serum of ALS patients and correlated these findings with the clinical status and pathological alterations observed in the muscle. We obtained data from 20 controls and 17 ALS patients (disease duration: 9.4 ± 6.8 months). Multivariate metabolomics analysis identified a distinct serum metabolome for ALS compared to controls (p-CV-ANOVA < 0.035) and revealed an excellent discriminant profile for muscle metabolome (p-CV-ANOVA < 0.0012). Citramalate was discriminant for both muscle and serum. High lauroylcarnitine levels in muscle were associated with low Forced Vital Capacity. Transcriptomics analysis of key antioxidant enzymes showed an upregulation of SOD3 (p = 0.0017) and GLRX2(1) (p = 0.0022) in ALS muscle. Analysis of mitochondrial enzymatic activity in muscle revealed higher complex II/CS (p = 0.04) and lower LDH (p = 0.03) activity in ALS than in controls. Our study showed, for the first time, a global dysfunction in the muscle of early-stage ALS patients. Furthermore, we identified novel metabolites to be employed as biomarkers for diagnosis and prognosis of ALS patients.

Keywords: Amyotrophic lateral sclerosis; metabolomics; mitochondria dysfunction; muscle; transcriptomics.

Figure 1

(A–C) Metabolomics analysis of serum from ALS and controls. (A) Score scatter plot based on the OPLS-DA models from serum to explain the diagnosis, with R2X = 0.452, R2Y = 0.477, Q2 = 0.269 and p < 0.035 for the CV-ANOVA test (blue: controls; green: ALS). (B) Loading scatter plot presenting the top 15 metabolites identified by the OPLS-DA. The horizontal axis displays the X-loadings p and the Y-loadings q of the predictive component. The vertical axis displays the X-loadings p(o) and the Y-loadings s(o) for the Y-orthogonal component. X-variables situated in the vicinity of the dummy Y-variables have the highest discriminatory power between the classes. The eight metabolites that had VIP scores higher than 1 are written in blue. (C) Pathway analysis with the 15 VIP metabolites highlighted alterations in arginine biosynthesis (p = 0.006), alanine, aspartate and glutamate metabolism (p = 0.02), biosynthesis of unsaturated fatty acids (p = 0.04) and linoleic acid metabolism (p = 0.04). Each node represents a metabolite set with its color based on its p-value and its size based on the pathway impact. The complete list of metabolic pathways is described in Supplementary File S4. (D–F) Metabolomics analysis of muscle from ALS and controls. (A) Univariate volcano plot analysis revealed different metabolites in the muscle metabolome from ALS patients and control subjects. Metabolites identified on the left are decreased in ALS patients compared to controls, while metabolites on the right of the diagram are increased in ALS patients. (B) Score scatter plot based on the OPLS-DA models from muscle to explain the diagnosis, with R2X = 0.76, R2Y = 0.555, Q2 = 0.446 and p < 0.0012 for the CV-ANOVA test (blue: controls; green: ALS). (C) Loading scatter plot presenting the top 15 metabolites identified by the OPLS-DA. The five metabolites that had VIP score higher than 1 are written in blue. (D) Pathway analysis with the 15 VIP metabolites highlighted alterations in the metabolism of glycine, serine and threonine (p < 0.001); biosynthesis (p = 0.002) and degradation (p = 0.04) of valine, leucine and isoleucine; aminoacyl-tRNA biosynthesis (p = 0.007) and glyoxylate and dicarboxylate metabolism (p = 0.03). The complete list of metabolic pathways is described in Supplementary File S4.

Figure 2

Venn diagram with discriminant metabolites revealed by univariate and univariate analysis identified 12 metabolites specific for serum of ALS patients, 20 metabolites specific to muscle of ALS patients, and citramalate as the metabolite commonly altered in serum and muscle of ALS patients when compared to control subjects. Venn diagram build with Venny 2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny/index.html accessed on 15 April 2021).

Figure 3

Alterations in muscle mitochondria. (A) Antioxidant genetic profile of ALS samples compared to controls. Expression levels of 25 transcripts of key antioxidant genes in the muscle of ALS patients (upper panel). Transcripts are ranked in decreasing order of expression in controls (bottom panel). Results are expressed as relative quantification (RQ) compared with control data (mean ± SEM). The horizontal gray line (y = 1) represents the healthy control profile (n = 20), and the black line represents patients’ data. Significant difference with the healthy controls (ΔCt values), *** p < 0.001. (B,C) Mitochondrial enzymatic activity from muscle of ALS patients and control subjects. (B) Ratios of analyzed complexes revealed alterations in the ratio Complex II/Cytrate synthase (II/CS; p = 0.04) and (C) in LDH activity (p = 0.03). No differences were found in the activity of the other complexes or ratios. Control: n = 20; ALS: n = 17. Results are shown as mean ± standard deviation. Data was analyzed using the Mann-Whitney statistical test. (D) Enrichment analysis of metabolites significantly correlated with LDH activity in muscle of ALS patients. Analysis performed with MetaboAnalyst tool. (E) Ultrastructural alterations in muscle mitochondria from ALS patients compared to control subjects. Representative transmission electron microscopy (TEM) images revealed the presence of mitochondria aggregates in the subsarcolemnic space in the muscle of ALS patient but not in the control subject. Scale bar: 2 µm. The insert corresponds to the zone indicated by the white box. Scale bar: 200 nm.

Figure 4

Alterations in actors of the energetic metabolism found in early-stage ALS patients. Metabolomics analysis performed in muscle and serum of ALS patients identified different metabolomes characterized by an increase in citramalate in both matrices. Besides this, increased levels of lauroylcarnitine were identified in the muscle of ALS patients as a bad prognostic factor. High levels of citramalate and lauroylcarnitine are associated with mitochondrial impairment. In ALS muscle, we observed a discrete accumulation of mitochondria in the subsarcolemnic space, suggestive of mitochondrial dysfunction. Mitochondrial dysfunction is a well-known source of reactive oxygen species (ROS). Transcriptomics analysis of muscle showed upregulation in ALS samples of two genes, SOD3 and GLRX2, that participate in the cellular antioxidant response. Furthermore, high levels of glycine—also found in the muscle of ALS patients—are associated with an upregulation of GLRX2. Upregulation of GLRX2 was shown to increase the activity of mitochondrial complex II, demonstrated in our analysis of mitochondrial activity in ALS muscle. Finally, mitochondrial dysfunction was also demonstrated by a decreased activity of LDH in ALS, compared to control samples. Our study confirmed the imbalance in muscle energetic metabolism in early-stage ALS and highlights metabolomics alterations associated with known pathological mechanisms described in ALS. These metabolomics alterations should be included in a panel of biomarkers to improve diagnosis and prognosis of ALS patients. Figure designed by Lucie Clarysse (Com&Sci).