Upsala J Med Sci 99: 259-266, 1994 5.3 Preparation of the Nordic Protein Calibrator Ole Blaabjergl, Mogens Blom2, Per Hyltoft Petersenl, Adam Uldal13, Hanne G@. 1, Department of Clinical Chemistry, Odense Uniuersity Hospital, DK-5000 Odense C, Denmark. 2. Department of Clinical Chemistry, Hjerring Sygehus, DK-9800 H j ~ r r i n g , Denmark. 3. Department of Clinical Chemistry, KAS Herleu, DK-2730 Herleu, Denmark. 4. Medi-lab, Adelgade 5-7, DK-2304 Copenhagen, Denmark. The problems related t o preparation of secondary calibrators for determination of single human serum proteins are numerous. The proteins in the preparation must be genuine and stable for years. Moreover, the preparation must be clear for the use in measurements based on turbidimetric and nephelometric principles. It would, further, be an advantage if the calibrator was prepared from serum from a well defined population so it could be repeatable reproduced. Commercial available calibrators for specific determination of human proteins are delipidated by treatment by organic solvents or similar materials. This treatment usually denaturates the proteins to a variable extent due to irreversible reactions, and this is often made worse by lyophilizing of the preparation. Therefore, the aim of the project was to produce a clear and reproducible liquid calibrator with genuine proteins stable for years. Production of Calibrator Reproducibility: The reproducibility is obtained by collecting blood from more than 1,000 male blood donors . The last few mL of blood in the tube after bleeding is collected, left at room temperature for coagulation for one to five hours, centrifuged (2,50O*g for 15 min), and the serum aliquots pooled. Solid NaN, was added to a concentration of 0.1 % (15 mmol/L), and stored a t - 80 "C. The blood donors were all tested for HIV and hepatitis B antibodies, and after a few days any pool where one sample was contaminated was discharged. 18 - 955059 259 Delipidation: The calibrator was cleared by ultracentrifugation, the lipid fraction was removed and t h e volume restored by 0.9 % NaCI. This procedure is very gentle and keep t h e proteins genuine without denaturation. When the sufficient amount of serum is collected, t h e serum is thawed and recentrifuged (2,5OO*g for 15 min), carefully mixed, weighed, and then ultracentrifugated for 40 hours at 180,00O*g a t 4 "C. After ultracentrifugation the lipid layer is sucked off and after a new weighing t h e calculated amount of 0.9 % NaCl (154 mmol/L with 15 mmol/L NaN3) is added in order to restore t h e original weight and t h e lot is frozen and kept at - 80 "C. When all lots are ultracentrifuged they are thawed and thoroughly mixed, t h e calibrator is then aliquoted in small portions (2 mL, 5 mL, and 10 mL) in cryotubes, frozen and stored at -80 "C until distribution for determination of t h e concentrations of t h e single proteins. U LT R AC E NT R IF U G AT I ON I CHYLO+VLDL 1 SERUM (turbid) C A L IB R A T 0 R (clear) Fig. 5.3.1. Schematic illustration of preparation of calibrator by ultracentri- fugation. From Blaabjerg et al. (1) with permission. Pools: In order to investigate t h e different steps of the procedure and different storage conditions, several pools were prepared from 1981 to 1989 and kept under different conditions. The pools are 1981: Ultracentrifuged (U-81) kept at - 20 "C (U-81-20) and at - 80 "C (U-81-80] and Not ultracentrifuged (NU-81) at - 20 "C (NU-81-20) and a t - 80 "C (NU- 8 1-80). 1983: Ultracentrifuged (U-83) kept at - 80 "C (U-83-80) and 1984: Ultracentrifuged (U-84) kept a t - 80 "C (U-84-80). 1989: Ultracentrifuged (U-89) kept at - 80 "C (U-89-80>. Not ultracentrifuged (NU-83) kept at - 80 "C (NU-83-80). 260 Methods for Documentation Turbidity o f p o o l s : The turbidity of the described batches of pools prepared as the calibrator was measured photometrically. The extinctions of undiluted pools in 10.0 mm cuvettes were determined spectrophotometrically on a Beckman DU-8 spectrophotometer at 650 nm. The same instrument was used for all measurements. Genuine proteins: It is difficult to demonstrate that the proteins are genuine. The pools were compared to commercial available calibrators and a fresh serum pool by agarose electrophoresis followed by immunofiiation with antibodies from DAKO according to Carlstrom and Johansson (3) . Table 5.3.1 Measurements of Turbidity in the Pools Extinction of undiluted pool in 10.0 mm cuvettes at 650 nm Time of Time of measurements collection of pools Pool Nov81 Nov82 Mar83 Apr84 Oct84 Mar89 Sep 1981 NU-81-20 0.587 0.759 0.791 1.111 1.948 2.417 NU-81-80 0.587 0.768 0.738 0.688 0.915 1.024 U-81-20 0.060 0.063 0.061 0.095 0.162 0.554 U-81-80 0.060 0.068 0.063 0.069 0.073 0.079 Jan 1983 NU-83-80 U - 8 3 - 8 0 0.365 0.049 0.600 0.062 Mar 1984 U-84-80 0.060 0.062 Au-De 1989 U-89-80 0.044 Combined reproducibility and stability: The stability of the preparations during storage as well as the reproducibility of preparing the pools are measured at the same time by the described procedures. This means that the results reflect both aspects. The measurements were performed at two occasions, in 1984 and in 1989. 1984: Radial immunodiffusion was performed according to Mancini (5). Electro immunoassay was performed according to Laurel1 (4). Turbichmetry was performed according to Blom and Hjplrne (2) on a Gemsaec centrifugal analyzer. Nephelometry was performed on a Technicon AIP Nephelometer@ 26 1 1989: Measurements were performed on a COBAS FARA centrifugal analyser For all measurements antibodies from DAKO were used. from ROCHE according to DAKO's recommendations. Results Turbidity of pools: The turbidity of the pools was determined at several occasions and the extensions are shown in Table 5.3.1. It is seen from the table that ultracen- trifugation is necessary for obtaining stable negligible values. Further, the storage at - 80 "C keeps the extinction low with insignificant increases for up to eight years. It should be mentioned that NU-83-80 was stored at - 20 "C from April 1984, and that the freezer where NU-81-20 and U-81-20 were kept was defect in July 1984 with temperatures up to - 10 "C for one week, but these facts do not change the general tendencies. t t t G - C D E r n L 1 I I 1 G G 1 2 3 - c + A B F D E Fig. 5.3.2. Comparison of three calibrator preparations (1:1981, 2:1984, and 3:1987) with six commercial protein calibrators (AB,C,D,E, and F) and a fresh serum pool (Gj by agarose electrophoresis and immunofixation. Arrows indicate changes in electrophoretic mobility compared to the fresh pool. From Blaabjerg et al. (1) with permission. Genuine proteins: Changes in protein mobility can be disclosed by agarose electro- phoresis and immunofixation If the mobility of the proteins is changed, this demonstrates changes in electrical charge and is an indication of denaturation. Three calibrator pools (from 1981, 1984, and 1987) and six commercial protein calibrators 262 were compared with a fresh serum pool by electrophoresis and immunofixation in 1987 and the results from the four proteins, which are most sensitive t o denaturation are shown in Fig. 5.3.2. Prealbumin shows changes in electrophoretic mobility in all the commercial calibrators (indicated by arrows) whereas the three preparations (from 1981 to 1987) of the frozen liquid calibrator all show the same mobility as in the fresh serum pool. In three of the commercial calibrators cxl-Antitrypsin shows lower concentration, an extra fraction and a broader band, respectively. No changes in mobility or concentration is seen in any of the frozen liquid calibrators. Ceruloplasmin may be the best indicator of denaturation and four of the commercial preparations show changed mobility and decreased concentrations. Again the frozen liquid pools behave like the fresh serum pool. Table 5.3.2 Combined Stability of Proteins and Reproducibility of Pools The data are normalized according t o protein and method Rad. Imm. Diff Elec. Imm. Ass Turbidimetry Nephelometry Protein 81 83 84 Prealbumin 0.978 1.007 1.015 Albumin 0.999 0.992 1.010 Orosomucoid 0.989 0.995 1.016 a l - A n t i t q p i n 1.021 1.000 0.980 Haptoglobin 0.993 0.997 1.009 Ceruloplasmin - az-Macroglobulin - - Transferrin 0.992 1.007 1.001 0.998 1.006 0.996 1.007 0.978 1.015 IgA 1.012 0.991 0.998 IgG IgM 81 83 84 1.001 0.985 1.013 1.063 0.966 0.970 1.015 0.980 1.005 1.045 0.974 0.981 1.017 0.978 1.005 1.013 0.991 0.995 1.010 0.982 1.008 0.978 1.000 1.022 0.989 0.997 1.014 1.020 0.996 0.985 1.016 0.91 0.989 81 83 84 1.011 0.999 0.991 1.028 0.987 0.985 1.016 0.983 0.999 1.012 0.983 1.005 1.003 1.005 0.993 1.020 0.984 0.995 1.015 0.993 0.992 1.044 0.972 0.984 81 83 84 - - - 1.000 0.998 1.002 0.992 0.996 1.012 Changes in Haptoglobin may be more difficult t o interpret as the combinations of phenotypes may be different from calibrator to calibrator. Combined reproducibility a n d stability: Three preparations of calibrator (from 1981, 1983, and 1884) were compared by quantification of eleven proteins by four methods in 1984. The results for the normalized values are shown in table 5.3.2 according to protein and method. 263 The results are all close to 1.000 and scattered with a variation close to the analytical., Only for al-Antitrypsin (and perhaps IgM) the results are consistent for the three methods indicating about 2 % higher values in the 1981 preparation. This, however, does not indicate instability as the highest values are in the oldest preparation. A new investigation in 1989 including a new preparation (collected in 1987), however, indicated lower values for the acute phase reactants. In table 5.3.3 the largest differences found between the 1984 and 1987 preparations are shown. Table 5.3.3 Combined Stability of Proteins and Reproducibility of Pools The data are given as differences in percentage for 1987 minus 1984 Protein M e a n difference 95 % Confidence Interval Prealbumin Albumin Transferrin Orosomucoid al-Antitrypsin Hap toglobin - 0.91 % - 0.89 % t 0.57 % - 0.45 % + 1.80 % - 0.52 % - 4.8 % - 3.5 % - 2.1 % - 2.24 to +0.42 % - 1.89 t o to.11 % +0.12 to + 1.02 % - 1.05 to t0.15 % + 1.60 to +2.00 % - 2.32 to + 1.28 % - 5.2 to - 4 . 4 % - 4.1 to - 2.9 % -2.6 to - 1.6 % The results from table 5.3.3 clearly show lower values for the acute phase reactants in the 1987 preparation. This is not a question of instability as the oldest preparation has the highest values. However, it illustrates that minor fluctuations in mean values of t h e population occurs e.g due to infectious diseases, whereby, the biological uncertainty is more important than the statistical uncertainty, due to sample size. Discussion It has been discussed for a long time whether protein calibrators should be frozen liquid or lyophilized, and how the delipidation should be performed. The IFCC-group decided on chemical delipidation and freeze drying. The documentation given here, 264 however, demonstrates that ultracentrifugation and storage as frozen liquid pool is very gentle to t h e proteins (compared to older commercial calibrators) and t h a t t h e proteins are stable for several years. For al-Antitrypsin, however, t h e quantifications indicate some minor changes in t h e structure, as described in the following two sections and in chapter 6. Except from this problem (which is well known for other protein calibrators) t h e frozen liquid calibrator is for years a reliable calibrator with genuine proteins when stored a t - 80 "C. Acknowledgements We want to thank Elsebeth Parlev and Karin Jorgensen for skilful and painstalung work with preparation of t h e pools and measurements of protein concentrations. References 1. 2. 3. 4. 5. Blaabjerg 0, Blom M, Gry H, Hyltoft Petersen P, Uldall A. Appropriate Sera for Calibration and Control of Specific Protein Assays. Scund J Clin Lab Inuest 1993;53 suppl. 212:13-15. Blom M, H j ~ r n e N. Profile Analysis of Blood Proteins with a Centrifugal Analyzer. Clin Chem 197622657-62. Carlstrom A, Johansson B-G. Agarose Gel Electrophoresis-Immunofixation. Scund J ImmunoZ 1983;17 SUPPI. 10~23-32. Laurel1 C-B. Quantitative Estimation of Proteins by Electrophoresis in Agarose Containing Antibodies. Anal Biochem 1966;15:45-52. Mancini G, Carbonara AO, Heremans JF. Immunochemical Quantitation of Antigens by Single Radial Immunodiffusion. Immunochemistry 1965;2:235-45. 265