Impact of Preanalytical Variations in Blood-Derived Biospecimens on Omics Studies: Toward Precision Biobanking?

Introduction

Laboratory testing and clinical research that requires biospecimens can be performed immediately after sample collection, or using biobank samples stored for future research such as biomarker discovery and analysis of disease mechanisms. However, variations in the results of laboratory testing and clinical research may occur in the preanalytical phase, including test requisition, biospecimen preparation, storage, and transportation to the laboratory (Kellogg et al., 2015; Plebani et al., 2014). That is, research data and outcomes do vary across populations and persons, but this is not always due to experimental or true biological variation.

During the preparation phase of blood-derived biospecimens, the preanalytical variables include, for example, the type of blood collection tube, pre- and postcentrifugation conditions, long-term storage temperature, and duration. Preanalytical variations may occur differentially, depending on the type of biospecimen and the assay method (Betsou et al., 2016). Thus, it is important to collect, manage, and utilize biospecimens based on an in-depth understanding of the factors that cause preanalytical variations. Based on this understanding, the National Cancer Institute Early Detection Research Network (EDRN, https://edrn.nci.nih.gov) has prepared a standard operating procedure (SOP) for collection and management of blood-derived biospecimens for clinical research such as biomarker discovery depending on the sample type (Tuck et al., 2009). The International Society for Biological and Environmental Repositories (ISBER, www.isber.org) Biospecimen Science Working Group has developed a Standard PREanalytical Code (SPREC) to facilitate the preanalytical annotation of biospecimens (Lehmann et al., 2012).

It is interesting to note that “precision medicine” has been on the postgenomics research agenda for several decades. Similar focus on “precision biobanking” appears to be important and timely for provision of high quality biosamples and conduct of research in the age of Big Data whereby research noise is minimized, at least from the perspective of the biosamples accessed from biobanks.

The present review article and biobanking innovation analysis offer new insights with a discussion on preanalytical variables, for example, the type of blood collection tube, centrifugation conditions, long-term sample storage temperature, and duration, on the output of omics analyses of blood-derived biospecimens: whole blood, serum, plasma, buffy coat, and peripheral blood mononuclear cells (PBMCs).

Materials and Methods

We searched and synthesized the biomedical literature related to the preanalytical variations of blood-derived biospecimens using the PubMed database (www.ncbi.nlm.nih.gov/pubmed), using the keywords such as preanalytical, variation, serum, plasma, blood, next-generation sequencing, proteomics, or metabolomics (Tables 1–4). The word “significant” in the tables indicates statistical significance as noted by the authors of the respective publications reviewed. The titles of the columns in Tables indicate the following:

The word biospecimens and biosamples were used interchangeably in the article.

Results

Tables 1–4 present the impact of various preanalytical variables on the output of omics analyses. Preanalytical variations depending on the biospecimen type, the analyte type, the type of blood collection tube, pre- and postcentrifugation conditions, long-term storage temperature and time, and assay method have all contributed to variations in the outcomes concerned. Such variations that can occur in the curation of the biospecimens for biobanking ought to be recorded as part of the metadata.

Discussion

Population-based biobanks have been recognized as scientific (Björkesten et al., 2017), as well as political and economic, instruments (Birch et al., 2016). Biobanks play a formidable function in planning and implementation of data-intensive omics studies, be they genomics, transcriptomics, proteomics, or metabolomics. In addition, biobanks have vast gatekeeper roles in medical and integrative biology research (Mirsafian et al., 2016). The outcomes of such omics and biological research are often variable, but this variability is not always due to veritable biological variation but also due to artifacts that can emerge from preanalytical biospecimen curation. Only recently attention has begun to focus on biospecimen stability, curation methods, and preanalytical variables that can collectively impact the experimental data generated using biospecimens available in biobanks.

Table 1 provides information about the differences occurring in the omics analyses output by the type of blood collection tube. For example, cfDNA concentration (Lam et al., 2004) and miRNA integrity (Basso et al., 2017) in the plasma were stable in both citrate and EDTA tubes. Serum (from red-top and tiger-top tubes) and plasma (from citrate, EDTA, fluoride, or heparin tubes) protein levels were significantly changed by the type of blood collection tube used (Evans et al., 2001; Hebels et al., 2013; Hsieh et al., 2006). Serum metabolites were not significantly changed by the type of blood collection tube (red-top and tiger-top tubes) (Breier et al., 2014), but plasma metabolites were significantly affected (among citrate, EDTA, and heparin tubes) (Hebels et al., 2013; Pinto et al., 2014; Yin et al., 2013). The quality of genomic DNA and RNA isolated from the buffy coat exhibited no significant differences among the citrate, EDTA, and heparin tubes (Hebels et al., 2013).

Table 2 provides an analysis of the impact of delayed separation after blood collection on the output of omics analyses. The cfDNA concentrations in plasma were stable when the centrifugation of whole blood was delayed at 4°C for 24 h or at room temperature (RT) for 6 h (Board et al., 2008; Jung et al., 2003; Lam et al., 2004; Norton et al., 2013). The output of DNA methylation analysis by microarray using genomic DNA from buffy coat was not affected by delayed separation at RT for 8 h (Hebels et al., 2013). Serum and plasma proteins, analyzed through an aptamer-based proteomics array, were not generally altered by delayed separation at RT for 2 h (Ostroff et al., 2010). Metabolic profiles in serum by delayed separation at 4°C for 24 h (Breier et al., 2014; Dunn et al., 2008) and in plasma by delayed separation at RT for 4 h (Breier et al., 2014; Hebels et al., 2013) were not significantly changed when analyzed by mass spectrometry.

Table 3 provides an overview of the impact of delayed freezing after biospecimen preparation on the output of omics analyses. The stability of genomic DNA was maintained in whole blood stored at 4°C for 24 h before freezing (Halsall et al., 2008; Permenter et al., 2015). Serum (Guo et al., 2013; Hsieh et al., 2006; Evans et al., 2001; Ostroff et al., 2010) and plasma proteins (Evans et al., 2001; Guo et al., 2013; Hsieh et al., 2006; Ostroff et al., 2010) were generally stable under delayed freezing conditions at 4°C for 24 h or at RT for 2 h. Serum metabolites were not significantly changed by delayed freezing at 4°C for 24 h when analyzed through nuclear magnetic resonance (NMR) spectroscopy (Barton et al., 2008). Plasma metabolites were significantly changed by delayed freezing at 4°C for 16 h or at RT for 2.5 h when analyzed by mass spectrometry or NMR spectroscopy (Kamlage et al., 2014; Pinto et al., 2014).

Table 4 provides information about the impact of long-term storage of biospecimens on the output of omics analyses. The quality and quantity of RNA were stable when whole blood was stored in Tempus tubes at −80°C for 6 years (Duale et al., 2014). The levels of miRNA were stable when plasma was stored at −80°C for 4 years (Balzano et al., 2015). Proteomic approaches using mass spectrometry showed minimal changes in serum stored at −80°C for 3 months (Hsieh et al., 2006), but no significant changes in plasma stored at −70°C for 4 years (Mitchell et al., 2005). Plasma metabolites were analyzed in a stable manner when plasma samples stored at −80°C for 20–30 months were used for NMR spectroscopy analysis (Pinto et al., 2014).

With the arrival of Big Data, omics-based biomarker discovery research is being actively conducted (Brooks et al., 2017; Matthews et al., 2016). It is important that biospecimens are secured with an in-depth understanding concerning the preanalytical variables because omics studies require large numbers of biospecimens with high quality (Anton et al., 2015). Many researchers have reported the preanalytical variables affecting biospecimen quality (El Messaoudi et al., 2013; Hubel et al., 2011; Lee et al., 2016a, 2016b). In this review, we summarized topline information about variations on the output of omics analyses that occur during the processing stages of blood-derived biospecimens, through an analysis of the data available in literature. Furthermore, we describe the sample handling conditions required to maintain the stability of blood-derived biospecimens for omics studies. Preanalytical variations depend on the biospecimen type, the analyte type, and the assay method, as well as the processing conditions of biospecimens.

Considering these points, researchers or biobank workers would be well served to establish SOPs for the collection and management of biospecimens suitable for research purposes and should also collect the preanalytical annotations of the biospecimens (such as SPREC). To be clearer, we define the collection and management of biospecimens as described above as “precision biobanking.” Precision biobanking will enable the generation of accurate omics data, thus contributing to the realization of precision medical care. As with the field of precision medicine, the future of precision biobanking rests, in part, on taking into account preanalytical variations so as to minimize their influence on the outcome of omics data, analyses, and sensemaking, particularly in clinical omics applications. We make a call for further research on a greater range of preanalytical variables and under different biospecimen curation conditions and settings in the future. We also underscore the need for using broad, critical, and independent social sciences and humanities (SSH) research so as to understand the multiple possible future trajectories of precision biobanking, something that is increasingly relevant in the current age of Big Data (López and Lunau, 2012; Petersen, 2013; Williams, 2006). To the extent that precision medicine rests on precision biobanking, both of these research knowledge domains should advance commensurately and in ways informed by independent SSH scholarship.