Towards Standards for Human Fecal Sample Preparation in Targeted and Untargeted LC-HRMS Studies

Abstract

Gut microbiota and their metabolic products are increasingly being recognized as important modulators of human health. The fecal metabolome provides a functional readout of the interactions between human metabolism and the gut microbiota in health and disease. Due to the high complexity of the fecal matrix, sample preparation often introduces technical variation, which must be minimized to accurately detect and quantify gut bacterial metabolites. Here, we tested six different representative extraction methods (single-phase and liquid–liquid extractions) and compared differences due to fecal amount, extraction solvent type and solvent pH. Our results indicate that a minimum fecal (wet) amount of 0.50 gram is needed to accurately represent the complex texture of feces. The MTBE method (MTBE/methanol/water, 3.6/2.8/3.5, v/v/v) outperformed the other extraction methods, reflected by the highest extraction efficiency for 11 different classes of compounds, the highest number of extracted features (97% of the total identified features in different extracts), repeatability (CV < 35%) and extraction recovery (≥70%). Importantly, optimization of the solvent volume of each step to the initial dried fecal material (µL/mg feces) offers a major step towards standardization, which enables confident assessment of the contributions of gut bacterial metabolites to human health.

High-Throughput UHPLC-MS to Screen Metabolites in Feces for Gut Metabolic Health

“…Methanol was chosen as the solvent of extraction of metabolites from feces due to its versatility for dissolving compounds with different polarities [ 1 , 6 ]. Methanol has been widely used as a solvent for the extraction of free and conjugated bile acids from human feces, demonstrating excellent method recovery, precision and accuracy for quantitative MS-analysis [ 7 , 23 , 25 ].…”

“…High standards of method reproducibility are especially important in clinical studies, as they usually involve a large number of biological samples from various matrices such as plasma, urine, tissue and feces [ 1 , 3 , 4 ]. Due to the heterogeneity of fecal samples, there is a need for standardized protocols of sample preparation for analysis that allows direct comparisons between cohorts [ 2 , 5 , 6 , 7 ].…”

A Dual UHPLC-HRMS-Based Fecal Metabolomics and Lipidomics Analysis and Automated Data Processing Pipeline for Comprehen-sive Gut Phenotyping

“…A fecal sample mass of 100 mg, and methanol vs. UPW ratio of 75:25 (v/v) generated the highest number of untargeted components (26 296) with 15 966 of those untargeted features displaying a CV<20% and 135/161 targeted features displaying a CV<20%. This extraction protocol is relatively similar to earlier reported procedures for fecal metabolomics 6,[64][65][66] , the main differences pertain to the lower starting material mass (100 vs. >= 200 mg) and the different methanol vs. UPW ratio (75:25 vs 80:20). The residual metabolome fraction was used to perform a lipidomic extraction.…”

Development of an Untargeted LC-MS Metabolomics Method with Postcolumn Infusion for Matrix Effect Monitoring in Plasma and Feces

“…Precision was evaluated at each concentration level from the calibration lines. Pooled EDTA plasma, pooled male EDTA plasma, pooled female EDTA plasma, and one individual EDTA plasma were used as four different plasma samples for recovery, accuracy, and matrix effect 45 with the adjustment that 1 mL of water per gram of sample was added at the start to improve homogenization. The homogenized and aliquoted samples (around 2 g per tube) were stored at −80 °C for more than 48 h before lyophilization.…”