Novel Markers of the Metabolic Impact of Exogenous Retinoic Acid with A Focus on Acylcarnitines and Amino Acids

Abstract

Treatment with all-trans retinoic acid (ATRA), the carboxylic form of vitamin A, lowers body weight in rodents by promoting oxidative metabolism in multiple tissues including white and brown adipose tissues. We aimed to identify novel markers of the metabolic impact of ATRA through targeted blood metabolomics analyses, with a focus on acylcarnitines and amino acids. Blood was obtained from mice treated with a high ATRA dose (50 mg/kg body weight/day, subcutaneous injection) or placebo (controls) during the 4 days preceding collection. LC-MS/MS analyses with a focus on acylcarnitines and amino acids were conducted on plasma and PBMC. Main results showed that, relative to controls, ATRA-treated mice had in plasma: increased levels of carnitine, acetylcarnitine, and longer acylcarnitine species; decreased levels of citrulline, and increased global arginine bioavailability ratio for nitric oxide synthesis; increased levels of creatine, taurine and docosahexaenoic acid; and a decreased n-6/n-3 polyunsaturated fatty acids ratio. While some of these features likely reflect the stimulation of lipid mobilization and oxidation promoted by ATRA treatment systemically, other may also play a causal role underlying ATRA actions. The results connect ATRA to specific nutrition-modulated biochemical pathways, and suggest novel mechanisms of action of vitamin A-derived retinoic acid on metabolic health.

Keywords: acylcarnitines; amino acids; retinoic acid; targeted metabolomics; vitamin A.

Figure 1

Relative plasma levels of carnitine and the indicated acylcarnitines species in NMRI mice treated with all-trans retinoic acid (ATRA; 50 mg/kg body weight, subcutaneous injection) or the vehicle (olive oil) during 4 days. Metabolite peak area (as arbitrary units, AU) was normalized by the volume of plasma used in LC-MS/MS. Data are expressed as percentage with respect to the normalized value in the control group, which was set to 100%, and are the mean ± SEM of 11 animals per group. * Indicates significant difference between the ATRA and the control group after Benjamini–Hochberg correction (observed p values in two-tailed Student’s t test are indicated when p < 0.1).

Figure 2

Relative plasma levels of amino acids in NMRI mice treated with all-trans retinoic acid (all-trans retinoic acid (ATRA; 50 mg/kg body weight, subcutaneous injection) or the vehicle (olive oil) during 4 days. Metabolite peak area (as arbitrary units, AU) was normalized by the volume of plasma used in LC-MS/MS. Data are expressed as percentage with respect to the normalized value in the control group, which was set to 100%, and are the mean ± SEM of 11 animals per group. * Indicates significant difference between the ATRA and the control group after Benjamini–Hochberg correction (observed p values in two-tailed Student’s t test are indicated when p < 0.1).

Figure 3

Relative plasma levels of the indicated lipids in NMRI mice treated with all-trans retinoic acid (ATRA; 50 mg/kg body weight, subcutaneous injection) or the vehicle (olive oil) during 4 days. Metabolite peak area (as arbitrary units, AU) was normalized by the volume of (plasma used in LC-MS/MS. Data are expressed as percentage with respect to the normalized value in the control group, which was set to 100%, and are the mean ± SEM of 11 animals per group. * Indicates significant difference between the ATRA and the control group after Benjamini–Hochberg correction (observed p values in two-tailed Student’s t test are indicated when p < 0.1). DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid.

Figure 4

Relative plasma levels of the indicated compounds of intermediary metabolism in NMRI mice treated with all-trans retinoic acid (ATRA; 50 mg/kg body weight, subcutaneous injection) or the vehicle (olive oil) during 4 days. Metabolite peak area (as arbitrary units, AU) was normalized by the volume of plasma used in LC-MS/MS. Data are expressed as percentage with respect to the normalized value in the control group, which was set to 100%, and are the mean ± SEM of 11 animals per group. * Indicates significant difference between the ATRA and the control group after Benjamini–Hochberg correction (observed p values in two-tailed Student’s t test are indicated when p < 0.1).

Figure 5

Score biplot representation of the first three principal components (PCs) in Principal Component Analysis. Data (of 11 animals per group) were spread by treatment to assess possible relationships between 117 variables (shown in Supplementary Table S1).

Figure 6

Relative plasma levels of carnitine and acylcarnitines (A), amino acids (B), the indicated fatty acids (C), the indicated metabolites/cofactors of intermediary metabolism (D), and the indicated mRNAs (E) in PBMC of mice treated with all-trans retinoic acid (ATRA; 50 mg/kg body weight, subcutaneous injection) or the vehicle (olive oil) during 4 days. Data are the means ± SEM of 3–5 mice per group and are expressed relative to the mean value of the control (vehicle-treated) group, which was set to 100. In A–D, metabolite peak area (as arbitrary units, AU) was normalized per cell, and observed p values in Mann–Whitney U test are indicated when p < 0.1 (significances disappeared after Benjamini–Hochberg correction). In E, mRNA levels were normalized to the mRNA levels of the reference gene Actb, * indicates a significant difference between the ATRA and the control group (p < 0.05), and p values are indicated when 0.05 < p < 0.1 (Mann–Whitney U test).

Figure 7

Effects of short-term, high-dose all-trans retinoic acid (ATRA) treatment in mice revealed in this work and suggested connection to oxidative metabolism in adipose tissues. Direct or indirect ATRA effects on metabolite levels and gene expression are indicated by arrows. Question marks are used to distinguish observed trends from statistically significant effects.