Effect of Anticoagulants and Cell Separation Media as Preanalytical Determinants on Zymographic Analysis of Plasma Matrix Metalloproteinases

Matrix metalloproteinases (MMPs) are calcium/zinc-dependent endoproteinases involved in physiologic and pathologic processes, modulating extracellular matrix degradation (1). MMP-2 (EC 3.4.24.24) and MMP-9 (EC 3.4.24.35) circulating in the peripheral blood (PB) of patients with neoplasia showed contrasting results, revealing the preanalytical issue of whether the method of PB sampling influences MMP concentrations (2) and their zymographic profiles (3). We therefore analyzed the effects of anticoagulants and cell separation media on PB gelatinolytic profiles. PB samples from 30 healthy volunteers were collected into Vacutainer™ Tubes with clot activator (SST), lithium heparin (LH), dipotassium EDTA (K2E), sodium fluoride/potassium oxalate (NaF/KOx), and buffered/acidic citrate [natrium citrate (9NC), acid-citrate-dextrose (ACD), and citrate-phosphate-dextrose-adenine (CPDA); Becton Dickinson]. After centrifugation at 500g for 15 min at 4 °C, the supernatants and buffy coats were collected and analyzed. Leukocyte subpopulations were obtained after Lympholyte® gradient (5.64% Nycograde™ Polysucrose 400, 9.65% sodium diatrizoate; Cedarlane), and their subset recovery was tested through cytometric analysis (4). Gelatinases from leukocytes and plasma samples were analyzed by gelatin zymography, with 150 μg of total protein loaded on the gel (3). Calibrators were prepared from capillary PB (5). MMP-2 and MMP-9 were measured by ELISA (5). Differences were compared using the Mann–Whitney U-test and the paired t-test; P values <0.05 were statistically significant. The present work was carried out in accordance with the ethics standards of the Helsinki Declaration of 1975, as revised in 1983. We found in plasma a 72-kDa constitutive gelatinase that was produced by nonproteolytic activation of MMP-2 with sodium dodecyl sulfate, and additional MMP-9 forms at 92, 130, and 225 kDa. Western blot analysis, Ca2+/Zn2+ dependence, and p-aminophenyl-mercuric acetate activation (data not shown) identified the plasma gelatinases as fibroblast-derived proMMP-2 and neutrophil-derived proMMP-9, circulating as latent activatable proenzymes. MMP-9 was significantly higher in serum than in citrate plasma [mean (SD), 59 (7) vs 8 (2) μg/L, respectively; P <0.01], but the active/latent isoform proportions were similar (Fig. 1A ). Platelet aggregation during clotting (6) may have caused these differences. Gelatin zymograms of plasma and serum MMPs from the same healthy individuals (median age, 38 years). Sample aliquots (containing 150 μg of total protein) were analyzed on 7.5% polyacrylamide gels containing 2 g/L gelatin 90 Bloom Type A from porcine skin. Lane Std in all panels, capillary who e-blood gelatinases; molecular masses (kDa) are indicated. (A), MMP profiles from plasma samples collected in the presence of different anticoagulants: lane 1, 9NC (32 g/L trisodium citrate, 38 mL/L buffered solution); lane 2, ACD (26.3 g/L trisodium citrate, 3.3 g/L citric acid, 31.9 g/L dextrose); lane 3, CPDA (26.3 g/L trisodium citrate, 3.3 g/L citric acid, 2.5 g/L sodium dihydrogen phosphate, 31.9 g/L dextrose, 0.3 g/L adenine); lane 4, K2E (1.8 g/L dipotassium EDTA); lane 5, LH (144 USP units); lane 6, NaF/KOx (2.5 g/L sodium fluoride, 2 g/L potassium oxalate); lane 7, SST (clot activator-silica gel). (B), MMP zymogram of plasma collected in the presence of 0.9, 1.8, 3.6, and 7.2 g/L dipotassium EDTA (lanes 1–4, respectively). (C), MMPs in plasma collected in the presence of 9NC (lanes 1, 3, and 5) and LH (lanes 2, 4, and 6). The same samples were run before (lanes 1 and 2) and after Lympholyte gradient treatment (lanes 3 and 4) and after treatment with 9.65% sodium diatrizoate alone (lanes 5 and 6). MMP-9 forms detectable in buffy coat isolated physiologically from 9NC (lane 7) and LH PB (lane 8) are also shown. MMP expression was lower in K2E plasma than in LH plasma [14 (3) vs 25 (4) μg/L; P <0.05; Fig. 1A , lane 4 vs lane 5]. The concentrations of MMP-9 forms decreased significantly with increasing amounts of K2E during PB collection, whereas MMP-2 was increased (P <0.01; Fig. 1B ). When we added anticoagulants to the zymography buffer (to mimic the conditions in Vacutainer Tubes), only K2E inhibited the gelatinolytic activities (data not shown). Although EDTA may alter MMP expression (7), the reasons for the contrasting K2E effects remain unknown. To minimize interindividual variability, we collected PB from the same individual into different buffers. We found mainly proMMP-2 in the buffered/acidic citrate plasma [202 (15) μg/L], whereas there were no statistically significant differences among the 9NC, ACD, and CPDA plasmas. We found additional proMMP-9 in the K2E, LH, and NaF/KOx plasmas (Fig. 1A , lanes 1–3 vs lanes 4–6). Our observations revealed that anticoagulants can act as preanalytical determinants of PB MMPs. LH and 9NC plasmas collected after Lympholyte gradient (Fig. 1C , lanes 3 and 4 vs lanes 1 and 2), as well as after 9.65% sodium diatrizoate alone (Fig. 1C , lanes 5 and 6 vs lanes 3 and 4), showed increased concentrations of all MMPs. Polysucrose 400 alone did not affect MMP concentrations (data not shown) or isoform profiles. Cytometric analysis revealed differences in MMP composition between leukocytes from LH PB vs leukocytes from 9NC PB (data not shown). Physiologic buffy coats from 9NC PB showed only MMP-9 forms, had lower gelatinase activity, and had a different zymographic profile with respect to LH PB (Fig. 1C , lane 7 vs lane 8). The MMP differences between LH vs 9NC plasma could be caused by differential release of MMPs from, e.g., platelets and leukocytes, with a changed MMP content/profile depending on the anticoagulant used (6)(8). Although previous reports suggested heparin as the anticoagulant of choice to study circulating MMPs (2)(3), to optimize the diagnostic validity of PB MMPs as cancer biomarkers (1), we recommend the use of buffered/acidic citrate (9NC, ACD, and CPDA), whereas LH, K2E, and NaF/KOx, which affect the MMP content and zymographic profiles of plasma and leukocytes, should be avoided.