3.2.1. Plasma and Serum

Föcker et al. studied the serum lipidome by targeted metabolomics. In a first approach, they reported increased lipid concentrations in AN patients during acute starvation and after weight recovery compared to healthy controls. Glycerophospholipids, including phosphatidylcholines (PC), lysophosphatidylcholines (LPC) and sphingomyelins (SM) were significantly increased in patients (e.g., LPC(14:0), LPC(17:0), PC(32:2), PC(32:3), SM(16:0), SM(18:1). In addition, they also observed increased concentrations of some carnitines in serum of AN patients at both time points (e.g., carnitine, acetyl-carnitine) [48]. These results are consistent with previous studies that described the lipid profile of plasma in AN by non-metabolomics approaches [60]. However, Miyata et al. reported that by untargeted metabolomics lower concentrations of some acylcarnitines (AC), such as palmitoylcarnitine, butyrylcarnitine, O-acetylcarnitine, and octanoylcarnitine are observed [51]. Additionally, the changes found by Föcker et al. after renourishment therapy were higher than in the acute state, similar to that found with amino acids, meaning that metabolism is highly susceptible to maintained starvation but even more to the subsequent weight recovery [48]. Nonetheless, in a subsequent study, Föcker et al. found fewer differences between controls and the acute starvation state. The most significant changes were between the starvation state and short-term weight recovery. After complete renourishment therapy, the metabolome was restored, reaching values close to those of the healthy controls. Hence, metabolism seems to adapt to long starvation and renourishment processes, reaching stable metabolic states. The discordances between these studies were justified by the methodological differences and the small sample size in both cases. The most relevant findings in this second study are some compounds that are proposed as potential biomarkers of different states in disease and treatment of AN. These compounds showed significant associations with their respective states and homogeneous time-course behavior in the tested samples. For the starvation state, PC(34:4) and PC(38:3) are significantly decreased and are restored after therapy. In short-term weight recovery, LPC(16:1) and LPC(20:3) are increased, while PC(38:6) and pimelylcarnitine are decreased, suggesting that they could serve as possible markers of the metabolic state during renourishment therapies in AN [47].

Shih et al. used an **untargeted** metabolomics approach by using GC-MS for the determination of polyunsaturated fatty acids (PUFAs) and a **targeted** metabolomics analysis (HPLC-MS/MS) for oxylipins measurement in plasma. Oxylipins are derived from PUFAs by enzymatic (cyclooxygenases, lipoxygenases, and cytochrome P450) or non-enzymatic oxidations, and they are the most relevant mediators of PUFAs functions in the human body. The concentrations of the free fatty acids n-3 (alpha-linolenate-ALA, stearidonate-SDA, eicosapentaenoate-EPA, and docosahexaenoate-DHA) and n-6 (gamma-linolenate-GLA, dihomo-gamma-linolenate-DGLA, arachidonate-ARA, and osbond acid-OBA) were re-

ported to be increased in the plasma of individuals with AN compared to controls. They analyzed the two major ratios between n-3 and n-6 PUFAs (LA (linolenate): ALA and ARA: EPA), which are significantly decreased in AN compared to controls. Moreover, those ratios were inversely correlated with anxiety in individuals with anorexia nervosa, and ARA: EPA was significantly correlated with BMI in patients as well. They also reported significant differences in individual oxylipins and oxylipins ratios. The eicosanoids significantly altered in AN included DHA and ARA metabolites, which belong to the CYP450 pathway. They also studied sEH (soluble epoxide hydrolase) activity which is an enzyme involved in the inactivation of epoxy-fatty acids from CYP catabolism of PUFAs. They finally suggested a greater in vivo activity, concentration, or efficiency of sEH in AN patients when compared to controls. The higher activity of this enzyme, involved in the CYP oxylipin pathway, has been related to increased inflammation and psychiatric disorders such as depression or anxiety, which are comorbidities of AN [61]. As a general overview, individuals with AN showed altered lipidome profiles that were correlated with increased neuroinflammation, anxiety disorders, and lower BMI [52,53].

Nguyen et al. also studied the plasma lipidic profile in acute and recovered individuals with anorexia nervosa compared to healthy controls at two different time points: fasting and postprandial. They examined 26 compounds, including saturated and unsaturated FA by GC-MS. Out of these 26 FA, AN patients presented significant increases in four species under fasting conditions and in only one of them after food intake. Similar to what Shih et al. described, laurate, EPA, and DPA (docosapentaenoate) were increased under fasting while ALA was increased at both timepoints in AN patients [52,56].

Bulant et al. analyzed the steroid profile of serum samples from 33 hospitalized women with AN. The aim was to determine the steroid variations after renourishment therapy. By GC-MS in selected ion monitoring (SIM) mode, they found significantly decreased concentrations of 7β-hydroxy-metabolites of C19Δ5steroids (7β-hydroxydehydroepiandrosterone and 5-androstene-3β,7β,17β-triol) which have been related to immunostimulation and anti-inflammatory properties. They also observed increased concentrations of the steroids at the beginning of the steroidogenic pathway, pregnenolone sulfate, and 20α-dihydropregnenolone sulfate after treatment [50].
