Hernández-Alonso Pablo, Becerra-Tomás Nerea, Papandreou Christopher, Bulló Mònica, Guasch-Ferré Marta, Toledo Estefanía, Ruiz-Canela Miguel, Clish Clary B, Corella Dolores, Dennis Courtney, Deik Amy, Wang Dong D, Razquin Cristina, Drouin-Chartier Jean-Philippe, Estruch Ramon, Ros Emilio, Fitó Montserrat, Arós Fernando, Fiol Miquel, Serra-Majem Lluís, Liang Liming, Martínez-González Miguel A, Hu Frank B, Salas-Salvadó Jordi
Universitat Rovira i Virgili, Departament de Bioquímica i Biotecnologia, Unitat de Nutrició Humana, Hospital Universitari San Joan de Reus, Reus, 43201, Spain.
Institut d'Investigació Pere Virgili (IISPV), Reus, 43003, Spain.
Mol Nutr Food Res. 2020 Jun;64(12):e2000178. doi: 10.1002/mnfr.202000178. Epub 2020 May 25.
The plasma metabolomics profiles of protein intake have been rarely investigated. The aim is to identify the distinct plasma metabolomics profiles associated with overall intakes of protein as well as with intakes from animal and plant protein sources.
A cross-sectional analysis using data from 1833 participants at high risk of cardiovascular disease is conducted. Associations between 385 identified metabolites and the intake of total, animal protein (AP), and plant protein (PP), and plant-to-animal ratio (PR) are assessed using elastic net continuous regression analyses. A double 10-cross-validation (CV) procedure is used and Pearson correlations coefficients between multi-metabolite weighted models and reported protein intake in each pair of training-validation datasets are calculated. A wide set of metabolites is consistently associated with each protein source evaluated. These metabolites mainly consisted of amino acids and their derivatives, acylcarnitines, different organic acids, and lipid species. Few metabolites overlapped among protein sources (i.e., C14:0 SM, C20:4 carnitine, GABA, and allantoin) but none of them toward the same direction. Regarding AP and PP approaches, C20:4 carnitine and dimethylglycine are positively associated with PP but negatively associated with AP. However, allantoin, C14:0 SM, C38:7 PE plasmalogen, GABA, metronidazole, and trigonelline (N-methylnicotinate) behave contrarily. Ten-CV Pearson correlation coefficients between self-reported protein intake and plasma metabolomics profiles range from 0.21 for PR to 0.32 for total protein.
Different sets of metabolites are associated with total, animal, and plant protein intake. Further studies are needed to assess the contribution of these metabolites in protein biomarkers' discovery and prediction of cardiometabolic alterations.
蛋白质摄入量的血浆代谢组学特征鲜少被研究。本研究旨在识别与蛋白质总摄入量以及来自动物和植物蛋白源的摄入量相关的独特血浆代谢组学特征。
对1833名心血管疾病高危参与者的数据进行横断面分析。使用弹性网连续回归分析评估385种已识别代谢物与总蛋白、动物蛋白（AP）、植物蛋白（PP）摄入量以及植物与动物蛋白比例（PR）之间的关联。采用双10折交叉验证（CV）程序，并计算每对训练-验证数据集中多代谢物加权模型与报告的蛋白质摄入量之间的Pearson相关系数。大量代谢物与所评估的每种蛋白质来源始终相关。这些代谢物主要包括氨基酸及其衍生物、酰基肉碱、不同的有机酸和脂质种类。蛋白质来源之间很少有代谢物重叠（即C14:0 SM、C20:4肉碱、γ-氨基丁酸和尿囊素），但它们的关联方向均不同。关于AP和PP方法，C20:4肉碱和二甲基甘氨酸与PP呈正相关，但与AP呈负相关。然而，尿囊素、C14:0 SM、C38:7 PE缩醛磷脂、γ-氨基丁酸、甲硝唑和胡芦巴碱（N-甲基烟酸）的情况则相反。自我报告的蛋白质摄入量与血浆代谢组学特征之间的十折交叉验证Pearson相关系数范围从PR的0.21到总蛋白的0.32。
不同组的代谢物与总蛋白、动物蛋白和植物蛋白摄入量相关。需要进一步研究来评估这些代谢物在蛋白质生物标志物发现和心脏代谢改变预测中的作用。
