Concurrent profiling of polar metabolites and lipids in human plasma using HILIC-FTMS

Abstract

Blood plasma is the most popularly used sample matrix for metabolite profiling studies, which aim to achieve global metabolite profiling and biomarker discovery. However, most of the current studies on plasma metabolite profiling focused on either the polar metabolites or lipids. In this study, a comprehensive analysis approach based on HILIC-FTMS was developed to concurrently examine polar metabolites and lipids. The HILIC-FTMS method was developed using mixed standards of polar metabolites and lipids, the separation efficiency of which is better in HILIC mode than in C5 and C18 reversed phase (RP) chromatography. This method exhibits good reproducibility in retention times (CVs < 3.43%) and high mass accuracy (<3.5 ppm). In addition, we found MeOH/ACN/Acetone (1:1:1, v/v/v) as extraction cocktail could achieve desirable gathering of demanded extracts from plasma samples. We further integrated the MeOH/ACN/Acetone extraction with the HILIC-FTMS method for metabolite profiling and smoking-related biomarker discovery in human plasma samples. Heavy smokers could be successfully distinguished from non smokers by univariate and multivariate statistical analysis of the profiling data, and 62 biomarkers for cigarette smoke were found. These results indicate that our concurrent analysis approach could be potentially used for clinical biomarker discovery, metabolite-based diagnosis, etc.

Figure 1

Total ion chromatograms of the standard mixture of 7 polar metabolites and 7 lipids on (a) Silica HILIC, (b) C5 and (c) C18. Ade: adenine; Guo: guanosine; Cr: creatinine; Arg: L-Arginine; His: L-histidine; Phe: L-phenylalanine; Suc: sucrose; LPC: LPC (17:0); PC: PC (17:0/17:0); PE: PE (17:0/17:0); PA: PA (17:0/17:0); PG: PG (17:0/17:0); Cer: Cer (d18:1/17:0(2S-OH)); DG: DG (15:0/15:0).

Figure 2

Base peak chromatograms (BPCs) for the four extracts: (a) methanol extract (ME), (b) MeOH/ACN/Acetone extract (MAAE), (c) polar fraction of CHCl₃/MeOH/H₂O extraction (CMHEp) and (d) non polar fraction of CHCl₃/MeOH/H₂O extraction (CMHEn), in positive ionization mode.

Figure 3. Comparison of the metabolite recovery of ME, MAAE and CMHE.

The peak areas of the 42 metabolites were normalized to the peak area of an ion (m/z = 427.39118, RT = 14.5 min) which could be detected in all the blanks and samples, and then multiplied by 10,000, followed by a base-10 log-transformation. Mean values from replicates were used for comparison.

Figure 4

(a) PLS-DA scores plots for comparison of the global metabolite profiles in (∆) heavy smokers and (+) non-smokers. 26.2% and 13.1% are the scores of the component 1 and 2, respectively, in the PLS-DA analysis. (b) Volcano plot of the 294 features with VIP values > 1 in PLS-DA analysis. Volcano plot is a combination of fold change and t-tests. The 62 features with a fold change >2 and p-value < 0.05 were marked in red.

Figure 5. The levels and correlations of the 62 significant features in smokers and non-smokers.

The log-transformed normalized intensities of the 62 features in the 9 samples were shown in the heat map. Obviously, 43 features were elevated in heavy smokers when compared to non-smokers while 19 features were reduced in the former. The correlations of the 62 features were shown by the Hierarchical clustering dendrogram. H1-H4: heavy smokers; N1-N5: non-smokers.