Suitability of Dried Blood Spots for Accelerating Veterinary Biobank Collections and Identifying Metabolomics Biomarkers With Minimal Resources

Abstract

Biomarker discovery using biobank samples collected from veterinary clinics would deliver insights into the diverse population of pets and accelerate diagnostic development. The acquisition, preparation, processing, and storage of biofluid samples in sufficient volumes and at a quality suitable for later analysis with most suitable discovery methods remain challenging. Metabolomics analysis is a valuable approach to detect health/disease phenotypes. Pre-processing changes during preparation of plasma/serum samples may induce variability that may be overcome using dried blood spots (DBSs). We report a proof of principle study by metabolite fingerprinting applying UHPLC-MS of plasma and DBSs acquired from healthy adult dogs and cats (age range 1-9 years), representing each of 4 dog breeds (Labrador retriever, Beagle, Petit Basset Griffon Vendeen, and Norfolk terrier) and the British domestic shorthair cat (n = 10 per group). Blood samples (20 and 40 μL) for DBSs were loaded onto filter paper, air-dried at room temperature (3 h), and sealed and stored (4°C for ~72 h) prior to storage at -80°C. Plasma from the same blood draw (250 μL) was prepared and stored at -80°C within 1 h of sampling. Metabolite fingerprinting of the DBSs and plasma produced similar numbers of metabolite features that had similar abilities to discriminate between biological classes and correctly assign blinded samples. These provide evidence that DBSs, sampled in a manner amenable to application in in-clinic/in-field processing, are a suitable sample for biomarker discovery using UHPLC-MS metabolomics. Further, given appropriate owner consent, the volumes tested (20-40 μL) make the acquisition of remnant blood from blood samples drawn for other reasons available for biobanking and other research activities. Together, this makes possible large-scale biobanking of veterinary samples, gaining sufficient material sooner and enabling quicker identification of biomarkers of interest.

Keywords: biobank; biomarker; cat; dog; dried blood spots (DBSs); metabolomics.

Figure 1

Principal component analysis (PCA) scores plots (PC1 vs. PC2) for data collected for three sample types applying the HILIC positive ion mode assay. The sample classes are pooled QC: yellow; domestic short hair cat: green, Norfolk terrier: light blue, Labrador retriever: dark blue, Beagle: red, and Petit Basset Griffon Vendeen: pink. The PC1 and PC2 variances were 31.1 and 8.2% (dried blood spot, DBS, 20 μL), 30 and 7.9% (DBS 40 μL), 25.5 and 7.5% (plasma).

Figure 2

PLS-DAscores plots for dog breed data collected for three sample types applying the HILIC positive ion mode assay. The sample classes were Norfolk terrier: yellow, Labrador retriever: turquoise, Beagle: dark blue, and Petit Basset Griffon Vendeen: purple. The two blinded DBS samples, blind sample 1: orange and blind sample 2: green, were overlaid on the relevant DBS plots. The 95% confidence ellipses were plotted for each breed class. Components 1 and 2 variances were 34 and 12% (DBS 20 μL), 33 and 13% (DBS 40 μL), and 32 and 10% (plasma).

Figure 3

Heatmap visualization of a small data set showing breed discriminatory ability collected using the lipid positive assay for (A) 20 μL DBS samples, (B) 40 μL DBS samples, and (C) 40 μL plasma samples. Each breed average is colored differently (Beagle, red; Labrador retriever, green; Norfolk Terrier, dark blue; Petit Basset Griffon Vendeen, light blue). Each heatmap defines a small number of metabolites, and the breed with darkest red indicates the highest relative concentrations, darkest blue indicates the lowest concentrations, and the spectrum of blue and red in between represents intermediate concentrations. AcCa, acyl carnitine; Cer, ceramide; DG, diacylglyceride; FA, fatty acid; LysoPC, lysoglycerophosphocholine; LysoPG, lysoglycerophosphoglycerol; PA, phosphatidic acid; PC, glycerophosphocholine; PE, glycerophosphoethanolamine; PG, glycerophosphoglycerol; PI, glycerophosphoinositol; PS, glycerophosphoserine; TG, triacylglyceride.