Choice of specimen and collection and handling variables

Pre-analytical variables can alter the analysis of blood-derived samples. Publications and protocols are generally deficient in this regard. Besides preparing the reference specimens of serum and plasma for direct comparisons, we undertook special studies on choice of sample type, stability during storage, use of protease inhibitors, and criteria for clinical standardization. The Specimens Committee concluded (Rai et al. [23]) that plasma is preferable to serum, due to less degradation ex vivo (as shown specifically by Tammen et al. [24] and Misek et al. [25]). Nevertheless, there is a view that standardization of proteomics assays with serum may be desirable, since archived specimens are so frequently sera.

They concluded that platelet-depletion of plasma may be desirable to avoid platelet activation with release of proteins, especially if there is a 4C step in the preparation. BD explained that 4C was chosen for centrifugation and holding of the tubes prior to aliquoting to aid in stabilizing labile biomarkers. For investigators concerned about platelet contamination, options include filtration of the plasma through a 0.2 m low protein binding filter; double centrifugation of the specimen; and use of additives that minimize platelet activation, such as CTAD, a mixture of citrate, theophylline, adenosine, and dipyridamole. Samples should be aliquoted and stored frozen with minimization of thaw/re-freeze cycles, preferably in liquid nitrogen, though — 80°C seems to be very nearly as good. Protease inhibitors would be desirable, but present cocktails introduce complications due to peptide inhibitors that may interfere in the MS and small molecule inhibitors that form covalent bonds with proteins, shifting the isoform pattern. The Committee recommends diligent tracking of pre-analytical variables, and development and use of certified reference materials for quality control and quality assurance.

Haab et al. [22] extensively analyzed the concentrations of assayable proteins in the PPP specimen sets. They noted a systematic 15% lower value for many proteins in citrate-plasma, compared with other specimens; it turns out that this can be attributed to dilution and osmotic effect with the citrate solution, without any impairment in detection of proteins compared with the other specimens. However, David Warunek and Bruce Haywood of BD advised us that results with citrate-anticoagulated plasma can be quite sensitive to the blood:additive ratio and the subject's hematocrit. EDTA, meanwhile, is a much better chelator of calcium and more effective at platelet inactivation.

The sets of four specimens from a given donor pool yielded rather similar numbers of proteins when analyzed by the same lab and same techniques (see Tab. 2). Naturally, the agreement on identification of specific proteins was greater for higher abundance proteins. Since the laboratories exercised considerable discretion in deciding how many and which of the reference specimens to request and how many to actually analyze, as well as how intensively to analyze them, compar isons across the specimen results is of limited validity in this exploratory phase of the PPP. However, comparisons within several laboratories (1, 2,11, 12, 28, 29, 41, 43 in Tab. 2) show quite close values for numbers of proteins identified, with deficiencies for B1-serum and B3-serum in Lab 1, and B1-heparin and possibly B3-heparin in Lab 29). It is curious that several laboratories chose citrate-plasma if they analyzed only one plasma specimen (Tables 1 and 2). Lab 28 shows greater similarity within each of the three donor pools for three citrate-plasma versus serum comparisons, than for citrate-plasma or sera across the three pools. The values for total number of proteins within each pair were quite close, whereas the B1 specimens yielded significantly fewer identifications than the B2 and B3 pairs. For B2 serum and B2 citrate plasma, they reported 365 proteins in common, of 542 and 572 identified in each. Ion current estimation of concentrations put 275 of the 365 within 62-fold; 59 proteins had plasma/serum values >2X and 31 had P/S values <0.5X (Adkins et al. [12], this issue). Lab 34 is a special case, because different instruments were used for B1-heparin (LCQ) and B1-serum (LTQ), as noted above (Section 3.1).

Tab. 2 summarizes the protein IDs by lab and specimen. As noted above, the numbers of proteins identified in the consolidated database may be different from those in the individual papers in this special issue due to the integration procedure applied to the Core Dataset and the expanded analyses for these papers. The most analyzed specimen, B1-serum (Caucasian American) had 1749 IDs among the 3020. The three anticoagulated B1 specimens yielded a total of 1904unduplicated IDs, of which 1023 were in common with the proteins identified in the B1-serum. The total number of unique IDs in the four B1 specimens that meet the two or more peptides criterion in either plasma or serum is 2630. A similar analysis ofthe combined C1 (Chinese) pooled specimens in just two labs yielded 1693 proteins, of which 1416 were identified in the B1 pool. With the exception of Labs 26 and 28, no very extensive analyses of the B2/B3 African-American and Asian American specimens were submitted. Combining all datasets, including the lyophilized NIBSC citrate-plasma specimen, we reached the 3020 protein dataset.

Tammen et al. [24] focused on the "peptidome" with mass <15 kDa. Peptides may be fragments of higher Mr proteins, or hormones, growth factors, and cyto-kines with specific biological functions. Their findings are not included in the Core Dataset since they used differential peptide display, plotting m/z ratios against retention time, with RP-HPLC-MALDI-TOF-MS. They do use nESQ-qTOF-MS/MS or MALDI-Tof-Tof-MS to confirm some peptide identifications. They did not actually attempt to identify proteins from the peptides. However, they made observations highly relevant to specimen processing. A large number of peptides, including many abundant peptides, are present only in serum, presumably due to the multi-protease events of clotting (AP-FXIII), enzyme activities (kallikrein), or pep-tides derived from cellular components, especially platelets, or the clot itself (thy-mosin beta-4, zyxin). In fact, at least 40% of the peptides detected in serum were serum-specific. Clotting is unpredictable due to influences of temperature, time, and medications, which are hard to standardize. These observations with serum may be highly relevant to the interpretation of SELDI results. They reported altered elution behavior of peptides in the presence of heparin, due to the polyanion nature of polydisperse low Mr heparin. Heparin acts through activation of antithrombi-n III, while citrate and EDTA inhibit coagulation and other enzymatic processes by chelate formation with ion-dependent enzymes. They recommend platelet-depleted EDTA or citrate-plasma, which gave consistent and similar results. They do not recommend addition of protease inhibitors, especially aprotinin, which requires mg/mL concentrations that interfere with analysis.

Was this article helpful?

0 0

Post a comment