Plasma phospholipidomic profile differs between children with phenylketonuria and healthy children

Phenylketonuria (PKU) is a disease of the catabolism of phenylalanine (Phe), caused by an impaired function of the enzyme phenylalanine hydroxylase. Therapeutics is based on the restriction of Phe intake, which mostly requires a modification of the diet. Dietary restrictions can lead to imbalances in specific nutrients, including lipids. In the present study, the plasma phospholipidome of PKU and healthy children (CT) was analyzed by hydrophilic interaction liquid chromatography-tandem mass spectrometry and gas chromatography-mass spectrometry. Using this approach, 187 lipid species belonging to nine different phospholipid classes and three ceramides were identified. Principal component analysis of the lipid species data set showed a distinction between PKU and CT groups. Univariate analysis revealed that 146 species of phospholipids were significantly different between both groups. Lipid species showing significant variation included phosphatidylcholines, containing polyunsaturated fatty acids (PUFA), which were more abundant in PKU. The high level of PUFA-containing lipid species in children with PKU may be related to a diet supplemented with PUFA. This study was the first report comparing the plasma polar lipidome of PKU and healthy children, highlighting that the phospholipidome of PKU children is significantly altered compared to CT. However, further studies with larger cohorts are needed to clarify whether these changes are specific to phenylketonuric children.


Introduction
Phenylketonuria (PKU) is the most prevalent inborn error in amino acid metabolism 1-3 . PKU is characterized by an impaired activity of the enzyme phenylalanine hydroxylase (PAH), resulting in a complete or partial inability of the enzyme to convert L-phenylalanine (Phe) to L-tyrosine. Deficiency of the PAH enzyme results in increased levels of P he in the blood (hyperphenylalaninemia) [4][5][6] . The accumulation of Phe can lead to intellectual impairment, microcephaly, seizures, and motor deficits 2,7,8 . Early diagnosis through neonatal screening and rapid therapeutic implementation are essential to prevent these possible complications 9 . The therapeutic approach for PKU is based on a lifelong Pherestricted diet, with a low intake of natural proteins 1 .
In the low-Phe diet, protein-rich foods like eggs, fish, meat and dairy products, which are the main natural dietary sources of long-chain polyunsaturated fatty acids (PUFA), are excluded 10 . Thus, to provide the required amount of daily protein intake, PKU patients consume commercial formulas with Phe-free essential amino acids, as well containing vitamins, minerals, and other nutrients 11 . To ensure that the essential PUFA dietary requirements are met, amino acid formulas containing PUFA and PUFA supplements are used together with PKU diet 12 . Despite the use of amino acid formulas and PUFA supplements, as well as some special low-protein foods available to PKU patients, high dietary restrictions can lead to nutritional imbalances, notably in lipids.
Alterations in blood lipids, especially in serum and plasma lipoprotein components, have already been described in PKU patients on a Phe-restricted diet 3 . Statistically significant lower levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) were observed in PKU patients compared to healthy subjects 5,13-16 . However, no statistical difference in the levels of TC, LDL-C and HDL-C has been reported in other studies 14,17 , highlighting the existence of dissimilar results in the plasma lipid profile of PKU patients.
Moreover, previous studies have revealed changes in the fatty acid (FA) profile of plasma and red blood cell (RBC) samples from PKU patients 3 . The existing literature is ambiguous about the prevalence of long-chain PUFA 4 deficiency in patients with PKU. However, some studies have shown a significant decrease in important PUFA, such as docosahexaenoic acid (DHA, 22:6n-3), eicosapentaenoic acid (EPA, 20:5n-3) and arachidonic acid (AA,, in patients with PKU who maintained a Phe-restricted diet (with no information available on PUFA supplementation) [18][19][20][21][22] . These reduced levels have been associated with dietary restrictions 22 , as PKU patients do not consume foods of animal origin rich in PUFA 2,3,23 . The low dietary intake of long-chain PUFA is thought to be the main cause of the decrease in the levels of DHA, EPA and AA, as their endogenous synthesis is very low 24,25 .
Given the important role of long-chain PUFA for the normal cognitive and visual development, controlled trials were also performed to elucidate the effect of PUFA supplementation on the FA status of patients with PKU, using plasma or RBC samples 4,23,[26][27][28][29][30][31] . Although some conflicting results between studies were reported, in most cases an increase in the content of n-3 PUFA was observed 3 .
Since FA are mainly esterified in other classes of lipids, namely in phospholipids (PL) which contain about 50% of the total amount of plasma FA 32 , variations in the plasma FA profile of PKU patients may impact these classes.
Some of the studies that have shown variations in the FA profile of PKU individuals have been performed using the plasma PL fraction 4,20,21,23,29,33 . However, changes at the level of individual PL species in PKU patients have never been explored.
Since PL have an important role in the signalling and regulatory functions in health and disease 3 , further characterization of plasma PL in PKU patients is needed to identify the plasticity of the PL profile due to dietary restrictions.
The present study represents the first report comparing the plasma phospholipidome of children with PKU and healthy children (control group, CT), highlighting the differences in the lipid profile of children with PKU at the level of PL species. For that, the PL plasma extracts obtained from children with PKU and healthy children (CT) were characterized by hydrophilic interaction liquid chromatography-tandem mass spectrometry (HILIC-MS/MS) for the identification of polar lipid species. Complementarily, gas chromatography-mass spectrometry (GC-MS) analysis was carried out for the determination of the composition of FA. 5

Study Design Overview
Plasma samples were obtained from PKU children (n = 15) between 3 and 17 years old, followed at the Centro de Referência de Doenças Hereditárias do Metabolismo -Centro Hospitalar Universitário de Coimbra (Coimbra, Portugal).
With an average of 9 years of age, the studied group included 7 females and 8 males. Patients were on a lifelong Phe-restricted diet, prescribed with Phe-free essential amino acid formulas (13 patients Supplementary Table S1. For the controls (CT, n = 12), recruited among those submitted to blood collection for other causes (e.g., minor surgery procedures) and in good general health and normal diet, the age of the CT children varied between 7 and 16 years. With an average of 12 years of age, the CT group included 4 females and 8 males. The collection of the blood samples was taken in the fasted state. Blood (2 mL) was collected into lithium-heparin tubes and centrifuged (2000 ×g, 10 min) to recover the plasma fraction. The plasma samples, after collection, were stored at -80ºC until the lipid extraction.

Plasma Phospholipids Extraction
The PL were extracted from plasma samples by solid-phase extraction (SPE) with a Visiprep SPE vacuum Manifold (Supelco, Sigma-Aldrich, Bellefonte, PA, USA) 34,35 . Briefly, protein precipitation was performed by mixing 100 µL of plasma sample with 900 µL of acetonitrile (ACN) with 1% formic acid, followed by vortexing (30 s) and centrifugation (626 ×g, 5 min). The resulting supernatant was transferred to a SPE column (HybridSPE-Phospholipid 30 mg, SUPELCO, Sigma-Aldrich, Bellefonte, PA, USA), after conditioning step (1 mL of ACN). As most supernatant was eluted, the column was washed with 1 mL of ACN with 1% formic acid and 1 mL of ACN. The PL were then eluted with ACN with 5% aqueous ammonia (2 x 1 mL). The PL were recovered and dried under a nitrogen stream. The extracts were then dissolved in dichloromethane and filtered using a HAMILTON glass syringe and syringe filter (0.22 µm pore size, 4 mm diameter, Millex-GV Durapore® (PVDF) membrane, hydrophilic, Millipore Corporation, Billerica, MA, USA). The filtered samples were collected in amber vials, dried and stored at -80 ºC until further lipid analysis.

Phospholipids Quantification
After SPE extraction, the total amount of PL was quantified by phosphorus (P) measurement 36 , as previously done 34,37 . PL-enriched extracts were dissolved in 100 μL of dichloromethane, and a volume of 10 μL was transferred, to a glass tube previously washed with 5% nitric acid, in duplicate. After drying the solvent under a stream of nitrogen, 125 µL of 70% perchloric acid were added to each tube. Samples were incubated in a heat block (Stuart, Staffordshire, U.K.) for 1 h at 180 °C, followed by cooling to room temperature. After that, 825 μL of Milli-Q water, 125 μL of ammonium molybdate (20 mM prepared in Milli-Q water), and 125 μL of ascorbic acid (567 mM prepared in Milli-Q water) were added to each sample, being vortexed between each addition. Samples were then placed during 10 min at 100°C in a water bath. Afterwards, using a cold-water bath, the samples were cooled down. The absorbance was measured at 797 nm in a Multiskan GO1.00.38 Microplate Spectrophotometer (Thermo Scientific, Hudson, 7 NH, USA) controlled by SkanIT software, version 3.2 (Thermo Scientific). The P content of each extract was calculated from a calibration curve prepared by carrying out the same procedure (without the heating block step) with standards containing up to 2 μg of P obtainedfrom a sodium dihydrogen phosphate dihydrate solution (100 μg mL -1 of P). The total amount of PL was then estimated by multiplying the amount of P by 25 38 .

Characterization of Fatty Acids Profile by Gas Chromatography Coupled to Mass Spectrometry (GC-MS)
FA analysis was performed by GC-MS after transmethylation of PL-enriched extracts 39 , as routinely used in authors' laboratory 35,37 . Briefly, an amount of lipid extract equivalent to 15 μg of total PL was dissolved in a volume of 1 mL of internal standard methyl nonadecanoate (Sigma, St. Louis, MO, USA) prepared in n-hexane (3.63 µM). FA were converted to fatty acid methyl esters (FAME) by adding 200 μL of methanolic KOH solution (2 M) and intense vortex-mixing for 1-2 min. Thereafter, 2 mL of NaCl solution (171 mM) were added, being centrifuged at 626 ×g for 5 min. A volume of 600 µL of the organic phase was then collected and dried under a nitrogen stream. The resulting FAME derivatives were dissolved in 100 μL of n-hexane and 2 μL were then used for GC-MS analysis (Agilent Technologies 8860 GC System, Santa Clara, CA, USA). The GC equipment was connected to an Agilent 5977B Mass Selective Detector operating with an electron impact mode at 70 eV and scanning the range m/z 50-550 in a 1 s cycle in a full scan mode acquisition. The oven temperature was programmed from an initial temperature of 58 °C for 2 min, a linear increase to 160 °C at 25 °C min -1 , followed by a linear increase at 2 °C min -1 to 210 °C, then at 20 °C min -1 to 225 °C, standing at 225 °C for 15 min. The injector and detector temperatures were 220 and 230°C, respectively. Helium was used as the carrier gas at a flow rate of 1.4 ml min -1 . The data acquisition software used was GCMS5977B/Enhanced MassHunter. The acquired data were analysed using the software Agilent MassHunter Qualitative Analysis 10.0. FA identification was 8 achieved by MS spectrum comparison with the chemical database NIST library and "The Lipid Web" 40 , and take into account the retention times and MS spectra of FAME standards (Supelco 37 Component FAME Mix, Sigma-Aldrich, Darmstadt, Germany). FA quantification was conducted using calibration curves obtained from FAME standards under the same instrumental conditions. . Initially, 5% of mobile phase A was held isocratically for 2 min, followed by successive linear increases to 70% of A within 11 min, and then to 90% of A in 7 min. The solvent (90% of A) was held isocratically during 30 min, returning to the initial conditions in 5 min, followed by a 5 min re-equilibration period prior next injection. In a glass vial with a micro-insert, a volume of 5 µL of each PLenriched extract, previously resuspended in dichloromethane (1 µg µL -1 PL), was

Statistical Analysis
Multivariate and univariate statistical analyses were performed using R were corrected for multiple testing using Benjamin-Hochberg method (with R built-in function) for the false discovery rate (FDR, q-values) 47 . Heatmaps were created from autoscaled data using the R package pheatmap 48 , and using "Euclidean" as clustering distance, and "ward.D" as the clustering method. All graphics and boxplots were created using the R packages ggplot2 49 , plyr 50 , dplyr 51 , tidyr 52 and ggrepel 53 .

Phospholipids quantification
The total amount of PL recovered after SPE (expressed in µg PL per 100 µL of plasma; mean ± standard deviation) was 41.27 ± 7.30 and 53. 78 ± 21.73 for PKU and CT samples, respectively. No statistically significant differences were found by univariate analysis between the two groups (q value > 0.05).

Fatty Acid Composition analysis by GC-MS
For an overview of the composition of esterified FA, GC-MS analysis was performed. A total of 9 FA were identified in the PKU and CT groups ( showed that the two groups were not differentiated, with the model capturing 84.7% of the total variance in the data set (Dim 1: 75.4%; and Dim 2: 9.3%) ( Figure 1). Univariate analysis (Shapiro-Wilk test followed by Welch t-test or Mann-Whitney test) was used to assess the existence of significant differences in the FA composition by comparing the two groups (CT vs PKU). Of the 9 FA identified, 6 showed significant differences between groups, with higher content in children with PKU when compared to the CT group ( Figure 2, Supplementary Table S2).

Identification of the phospholipid profile by HILIC-MS/MS
The plasma phospholipid profile of the two experimental groups (PKU and CT) was characterized using high-resolution HILIC-MS/MS. In total, we identified

Univariate analysis (Shapiro-Wilk test followed by Welch t-test or Mann-
Whitney test) was also performed to evaluate lipid species with significant variation between the two groups (CT vs PKU). The results revealed that 146 of 190 lipid species were statistically different between groups (q-value < 0.05) (Supplementary Table S4). Boxplots of the 16 main species with the lowest qvalues (q-value < 0.001) are shown in Figure 4. These 16 species that exhibited major variation included 7 PC, 5 SM, 3 PS, and 1 LPC. All of these PC species

Discussion
Previous studies have revealed alterations in lipoprotein levels of serum/plasma and the FA profile of plasma and red blood cells of PKU individuals 3,5,16,19,26,29,33,54,55  In accordance to the multivariate analysis of the FA dataset by PCA (Figure 1), the PKU and CT groups were not separated.
Regarding to the univariate analysis of the FA dataset, the PKU group showed a significant increase in the level of PUFA DHA and EPA, compared to the CT group ( Figure 2). Consistent with these results, in seven studies 4,23,[26][27][28][29]58 , higher levels of DHA in the total serum/plasma PL fraction were observed in PKU individuals supplemented with PUFA. Only two of these seven studies showed an increase in the level of EPA in plasma 23,28 , as observed in this study.
It is known that in healthy individuals only 2-5% and 5-10% of α-linolenic acid are converted to DHA and EPA, respectively 59 . Consequently, the DHA and EPA status is determined by food intake 60  the increased generation of anti-inflammatory cytokines and the generation of special pro-resolving lipid mediators, such as resolvins, protectins and maresins 63 . Therefore, it is possible to speculate that DHA supplementation would result in an increased capability to produce special pro-resolving lipid mediators and a decrease in the inflammatory state. Such an anti-inflammatory action may be relevant, considering that a pro-inflammatory state has already been described in patients with PKU 24,65 . Furthermore, the higher levels of n-3 PUFA are associated with a reduction in the risk of developing cardiovascular diseases and atherosclerosis 62,63,66,67 . Despite the health benefits associated with consuming PUFA, particularly n-3 PUFA, it should be noted that the question about the dosage of PUFA needed to achieve optimal effects has remained unanswered.
The PCA of the lipid species showed a separation between the PKU and CT groups ( Beyond the common diacyl PC, also an up-regulation of some ether PC as well as ether PE has been observed in PKU children, both representing 15.06% of all statistically significant lipid species (Supporting Information, Table S4). The physiological functions of these unique alkyl-acyl PLs are not fully understood.
However, it has been reported that ether-linked PL, especially plasmalogens, have antioxidant properties because of their ability to scavenge oxygen radicals, as they are preferentially oxidized [69][70][71] , and seem to be transformed into inoffensive products that can be reutilized by the organism 71,72 . The increase of 20 some ether-like PC and PE may be beneficial for PKU children because they can decrease oxidative stress and lipid peroxidation 3 .
Regarding the increase of PI species, namely PI (36:4) and PI (38:4), observed in PKU children, it is known that PI species influence the structure and function of plasma lipoproteins, but their role in plasma is unclear and requires further investigation 73 . Alterations in PI content are probably associated with their signalling roles, as they are important precursors of signalling molecules, such as PIPs that regulate metabolic processes 74 . Therefore, we speculate that the observed increase in PI species could be associated with a decrease in its phosphorylation to phosphoinositides, by phosphatidylinositol kinases (e.g. phosphatidylinositol 3-kinase), due to the reduction in oxidative stress 75 .
PS is the major class of anionic phospholipids in eukaryotic membranes 76 . The

Conclusions
The present study reports, for the first time, the application of mass

Supporting Information
( fasting, from PKU patients and healthy controls (mass error < 5 ppm). (Table S4) The 146          . The confirmation as PE species was also achieved by observing the product ion at m/z 140.0 (formula: C2H7NO4P; exact mass: 140.0113), corresponding to phosphoethanolamine polar head. For LPE, the same fragmentation was observed, with the exception that only one product ion corresponding to a fatty acid was detected.