Multiomic, mass spectrometry-based analysis of dried blood: toward deep phenotyping of sepsis

Rationale: The quantification of circulating proteins and metabolites can enable deep phenotyping of diseases, including sepsis. However, population-based studies have been largely confined to plasma, which suffers from pre-analytical variability and ignores the importance of cellular changes relevant to disease pathobiology. We have been exploring the applicability of dried blood, collected using volumetric absorptive microsampling, for multomic analysis. We have previously found that protein abundance changes in dried blood from COVID-19 ICU patients versus normal controls largely replicates many of those observed in matched cell-free plasma. We are extending this approach to the analysis of protein post-translational modifications and metabolites, using analytical figures of merit, as well as differential abundance measurements based on inter-individual variability and in vitro exposures, for method development and validation.

Methods: Lithium-heparin blood was collected from n=4 normal human donors. Blood was treated with sham or 1 microgram/mL E. coli lipopolysaccharide and incubated at 37 degC for up to 48 hours. Blood was exposed to a phosphatase inhibitor cocktail at the 6 h timepoint. Twenty microliters of blood was collected on a Neoteryx Mitra tips and dried before storing at -80 degC. For proteomic analysis, Mitra tips were denatured in 5% deoxycholate, reduced and alkylated and digested with trypsin for 2 h. Approximately 1 mg of digests were enriched for glycopeptide and phosphopeptides using Ti-IMAC or TiO. Metabolites were extracted using MeOH/ammonium acetate. Proteome and PTM analyses utilized microflow and nanoflow liquid chromatography, respectively, coupled to high-resolution accurate mass (HR/AM) mass spectrometry. Metabolite analysis utilize ZipChip capillary electrophoresis coupled to HR/AM MS.

Results: We developed facile methods for processing of Mitra tips for proteome, PTM and metabolites that can be accomplished in one day in a 96-well format. Approximately 1000 proteins, 400 N-linked glycopeptides, 1000 phosphopeptides and 200 metabolites were identified and quantified in dried blood from two Mitra tips per subject timepoint in <2.5 h of total analysis time. As expected, hierarchical clustering of the blood proteome revealed high inter-subject variability. LPS treatment had modest effects on protein/PTM abundance, whereas 138/940 quantified phosphopeptides were significantly increased with phosphatase inhibitor, establishing a subset of the blood phosphoproteome that is highly dynamic and may be useful for studying PTM stability during long-term storage.

Conclusions: Mass spectrometry-based analysis of dried blood can be accomplished in a relatively short amount of time and cover a potentially large biological space, setting up the analysis of large biobanks for deep phenotyping.

Funding: GM146142