Global Urinary Volatolomics with (GC×)GC-TOF-MS

Abstract

Urinary volatolomics offers a noninvasive approach for disease detection and monitoring. Herein we present an improved methodology for global volatolomic profiling. Wide coverage was achieved by utilizing a multiphase sorbent for volatile organic compound (VOC) extraction. A single, midpolar column gas chromatography (GC) assay yielded substantially higher numbers of monitored VOCs compared to our previously reported single-sorbent method. Multidimensional GC (GC×GC) enhanced further biomarker discovery while data analysis was simplified by using a tile-based approach. At the same time, the required urine volume was reduced 5-fold from 2 to 0.4 mL. The applicability of the methodology was demonstrated in a pancreatic ductal adenocarcinoma cohort where previous findings were confirmed while a series of additional VOCs with diagnostic potential were discovered.

Figure 1

Extracted VOCs and their contamination levels for comparison of sorptive extraction techniques, where the multiphase sorbent approach in headspace analysis clearly outperforms all the rest in terms of total peak area of summed extracted VOCs: URN, urine; HS, headspace; IM, immersive; MP, multiphase; PDMS, polydimethylsiloxane.

Figure 2

GC×GC- (a) and GC- (b) TOF-MS chromatograms of recollected pooled quality control (QC) urine samples. The recollection of multiple samples to a single thermal desorption (TD) tube facilitated metabolite annotation.

Figure 3

(a) Score plot of the OPLS-DA model generated on the preprocessed GC-TOF-MS data set, indicating cancer-control separation. (b) Permutation testing – 999 repeats: The Y-axis represents the R²Y and Q² values of each model. The X-axis shows the correlation coefficient between the “real” Y and the permuted Y. R²Y, 0.61; Q², 0.32.

Figure 4

Volcano plot of the tile-based Fisher ratio analysis generated on the GC×GC-TOF-MS data set, indicating cancer-control separation.