Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
Contents lists available at ScienceDirect
Best Practice & Research Clinical Endocrinology & Metabolism journal homepage: www.elsevier.com/locate/beem
7
Standardization of hormone determinations Ulf-Håkan Stenman, M.D., Ph.D., Professor emeritus Department of Clinical Chemistry, Biomedicum, Helsinki University, PB 63, FIN-00014 Helsinki, Finland
Keywords: standardization standards reference methods hormones immunoassay mass spectrometry
Standardization of hormone determinations is important because it simplifies interpretation of results and facilitates the use of common reference values for different assays. Progress in standardization has been achieved through the introduction of more homogeneous hormone standards for peptide and protein hormones. However, many automated methods for determinations of steroid hormones do not provide satisfactory result. Isotope dilution-mass spectrometry (ID-MS) has been used to establish reference methods for steroid hormone determinations and is now increasingly used for routine determinations of steroids and other low molecular weight compounds. Reference methods for protein hormones based on MS are being developed and these promise to improve standardization. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction The aim of standardization is to obtain identical and correct results with different methods. The expression “harmonization” is often used as an alternative to standardization. However, while different harmonized methods may provide similar results, they may all be biased. Therefore, the goal should be to achieve standardization, but if standards are not available, harmonization is the only option. WHOapproved International Standards (IS) for most clinically important hormones are available from the National Institute for Biological Standards and Control (NIBSC) in the UK. However, only a few standards are available for cytokines and growth factors [1]. Standardization is feasible if a standard and a reference method are available. The reference method should preferably be a definitive method, in practice mass spectrometry (MS), which presently can be used for relatively small molecules like steroid and thyroid hormones and also for some peptide hormones [2–4]. Most hormone determinations are performed by immunoassay but MS is increasingly used for routine determination of steroid hormones. Standardization of immunoassays is challenging and results from quality assessment schemes show that for many methods, between-method variation is not yet on an adequate level [5]. E-mail address: ulf-hakan.stenman@helsinki.fi. 1521-690X/$ – see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.beem.2013.10.007
824
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
Development of hormone immunoassays Routine determination of hormones became possible with the development of the first radioimmunoassay for insulin by Yalow and Berson in 1959 [6] and an assay of thyroxin by use of thyroxin binding globulin by Ekins in 1960 [7]. During the following decade, immunoassays for most peptide and protein hormones were developed using antisera produced in-house. Some of these antisera were widely distributed and when purified protein hormone preparations became available from NIH and WHO, fairly well standardized assays could be established. When commercial assays gradually became available, the need for standardization became obvious. Assay manufacturers used different antisera, calibrators and assay conditions and this caused substantial variation in the results obtained. Thus twenty years ago the results obtained by different assays for human chorionic gonadotropin could vary by more than 10-fold [8]. Presently, hCG assays give results that differ by less than 50% from the mean [9]. Initially, binding inhibition assays were used to measure both small and large hormones. In these, the hormone to be measured compete with radiolabeled hormone for binding to the antiserum. Binding inhibition assays are still used to measure small hormones. In these, assay sensitivity increases with decreasing concentrations of antiserum and increasing incubation time. Optimal detection limits requires assay times of hours or even days. However, presently, most hormone immunoassays are performed with automated analyzers, in which incubations times have been minimized, often to less than 10–15 min, in order to optimize throughput. Unfortunately this has lead to impaired quality of steroid hormone assays [10–12]. Thus, it has been decided that steroid hormone determinations have to be performed by mass spectrometry in studies to be published in the Journal of Clinical Endocrinology and Metabolism [13]. Hormones with an MW larger than 3 kD can determined by immunometric, also called sandwich assays [14]. In these, a capture antibody (or antiserum) attached to a solid phase, i.e., a tube wall or a particle, binds the hormone through one epitope and a detector antibody, which is labeled with a radioisotope, an enzyme, a fluorophore or a luminescent molecule, binds to another epitope. A polyclonal antiserum containing antibodies to many epitopes on the antigen and can be used both as capture and tracer antibody. This approach is seldom used in commercial assays, but combinations of monoclonal and polyclonal antibodies are used. The detection limit of sandwich assays improves with increasing antibody concentrations. Thanks to this and the use of sensitive detection methods, very low detection limits can be achieved in spite of short incubation times. Problems for standardization Hormone concentrations Standardization of hormone determinations is complicated by a number of problems. First, the concentrations of hormones in circulation are very low in comparison to those of other similar substances. In healthy subjects, the concentrations of most peptide and protein hormones are in the range 1–50 pmol/l. The concentrations of the most abundant serum protein, albumin, is about 0.7 mol/l, which is 10–100 millionfold that of typical protein hormones. The concentrations of steroid hormones are in the range 20–300 pmol/ l for estradiol to 50–500 nmol/l for cortisol. These should be compared to the concentration of cholesterol, 5 mmol/l, which is 10,000–100,000,000–fold higher than those of steroid hormones. It is inevitable that the presence of a huge excess of similar substances affects hormone determinations by causing nonspecific interference. Considering these facts, it is amazing how specific and sensitive immunoassays can be, but it explains some problems with steroid hormone assays with automatic analyzers [13,15]. Hormone heterogeneity Most peptide and protein hormones occur in different forms in circulation, e.g., 22 and 20 kD growth hormone, intact hCG and its free b subunit (hCGb) or intact paratahormone and hormonally inactive, truncated forms. In order to standardize assays for such hormones, it is necessary to define which forms should be measured and to use antibodies that detect relevant epitopes. Genetic polymorphism of protein hormones is fairly common and a variant of luteinizing hormone (LH), which differs with respect to two amino acids (Trp8Arg and Ile15Thr) is not detected by some antibodies [16].
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
825
Cross-reacting hormones and binding proteins Standardization of steroid hormone determinations is complicated by the occurrence of closely related forms that crossreact with the antisera used. Some antibodies to hCG also recognize LH and vice versa. Specific assays can be developed by careful selection of epitope reactivity of the antibodies used [17]. The difference in structure between many steroid hormones, e.g., estradiol and estrone or testosterone are actually very small and therefore, some cross-reactivity between related hormones is inevitable. Most steroid hormones occur in circulation in complex with steroid-binding proteins that compete with antibodies. In many early steroid hormone immunoassays, this problem was eliminated by extraction and separation of related hormones by chromatography before the immunoassay. This type if pre-purification is not compatible with automatic analyzers. In these, dissociation of steroids from binding proteins is achieved by using agents that block binding to endogenous binding proteins. The incubation time with antisera and labels is short, often less than 10–15 min. In this time, binding equilibrium is not achieved which earlier was considered important. Thus is it not surprising that the quality of most automated steroid hormone assays is unsatisfactory [10–13]. Highly specific mass spectrometric (MS) methods have been used for decades as reference methods for determination of steroid hormones and other small molecules [18]. Thanks to recent advances in MS technology, these methods are now increasingly used for routine determinations. While the cost of MS still is higher than that of immunoassays, the better validity of MS often justifies its use [12,19]. A misleading result of an immunoassay often causes high additional costs by leading to unnecessary imaging and other examinations [15]. A problem with steroid and thyroid hormone determinations is that the concentration of the free, or bioavailable, fraction of the hormone is physiologically and clinically most relevant. Immunoassays for direct measurement of the concentrations of free steroid hormones have not been found to be reliable [20], but assays of free thyroid hormones are generally considered more relevant than the total hormone concentrations. The concentrations of free hormones can be determined by using dialysis or ultrafiltration. When combined with MS, these methods can used to establish reference methods [21,22]. Many peptide and some protein hormones occur in complex with specific binding proteins in circulation. Insulin-like growth factors (IGFs) occur mainly in complex with IGF-binding proteins (IGFBPs). Thus IGF-1 is bound to IGFBP-3, from which it is dissociated at low pH before assay. Rebinding is prevented by addition of an excess of IGF-2 before assay of total IGF-1. IGF-2 is also bound to other IGFBPs [23]. About 10% of growth hormone (GH) in plasma occurs in complex with the growth hormone binding protein (GHBBP) but the concentrations of GHBBP varies. Some assays recognize this form while other ones do not. This causes method-dependent variation in the results [24]. Interestingly, immunoassays recognizing naturally occurring and recombinant GH differently can be used to detect administration of recombinant GH for doping purposes [25]. Many peptide hormones, e.g., insulin, occur free in circulation unless bound to antibodies induced by injection of hormone analogs for therapeutic use, or occur spontaneously as often is the case with prolactin [26]. Matrix effects The composition of the sample affects the reaction between antigen and antibody and thus the results of hormone determinations. Therefore, the calibrators should be prepared in the same matrix as the sample, usually serum or plasma. Ideally, the calibrators are prepared by spiking serum devoid of the analyte with standards of desired concentrations. Sometimes, analyte-free serum can be obtained from sick people, but this is seldom possible in practice. Protein and peptide hormones can be removed by affinity chromatography using antibodies, but removal is seldom complete and leakage of antibodies disturbing the assay is hard to avoid. Serum free of steroid hormones can be prepared by charcoal absorption, which however removes many low MW components, e.g., cross-reacting steroids, that affect steroid hormone determinations. Thus it is obvious that preparation of the matrix for the calibrators is a crucial step in the development of an immunoassay [27]. Manufacturers seldom provide information about how this has been done. A common matrix problem is caused by human antibodies against animal immunoglobulins that are present at such concentrations that they disturb the assay. Circulating antibodies against ovine and
826
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
bovine immunoglobulin are actually quite common, obviously as a result of the ingestion of animal milk and meat [28]. In most cases the interference can be prevented by use of animal immunoglobulin in the assay buffer. This problem is discussed in detail elsewhere in the accompanying issue [29]. While the composition of serum (and plasma) is fairly constant, that of urine varies considerably. Due to variation of urinary flow rate, the density of urine varies from about 1.002 to more than 1.030, i.e. about 15-fold. The content of proteins, steroids and other solutes varies accordingly. Therefore, the measured concentrations of analytes in urine is corrected to correspond to a density of 1.020. Thus the result is multiplied by 2 if density is 1.010 and with 4 when density is 1.005. The salt concentration of urine is reflected by the density, and in dilute urine, salt concentration is lower than in serum, while it is higher when density is high. Furthermore, urine pH varies and together with variation in salt concentration this may have a considerable effect on the antigen–antibody reaction. This may cause a considerable bias if the sample volume in relation to total incubation volume is large, i.e., more than 10–15%. Therefore, rather few immunoassays have been validated for analysis of urine samples. Analysis of urinary steroid concentrations is especially complicated because of the large excess of steroid conjugates in urine. Therefore, analysis of drugs and hormones used for doping is performed by rigorously controlled methods based on mass spectrometry. Standards and units The first hormone standard for insulin was introduced 1925 to facilitate calibration of the bioactivity of preparations for treatment of diabetes. The standard was assigned values in arbitrary international units (IU) based on biological activity. The present international standard for insulin is still calibrated in IU, 1 mg corresponding to 26 IU [30]. These standards have later been used for immunoassay standardization, but recombinant insulin calibrated in mass or mol/l is increasingly used to calibrate immunoassays. The first standard for hCG established in 1939 was a crude extract of urinary proteins from pregnant women. It was assigned a value in units based on its bioactivity, and each subsequent hCG standards has been calibrated against the previous one using biological methods measuring the potency of the standards. With improving purity of the hCG standards, the biological activity in relation to mass has increased. At the same time, the immunoreactivity in relation to bioactivity has also changed (see below) [31]. Arbitrary units have also been assigned to hormone subunits and fragments lacking bioactivity. Therefore, the first standards for the free subunits of hCG, i.e., hCGa and hCGb were assigned units based on the protein content of the preparations. Thus 1 IU of the 3rd IS for hCGa or hCGb is equivalent to 1 mg [31]. These units are quite different from those for hCG, in which 1 mg corresponds to 9.3 IU. These units are not used in practice and, when concentrations of hCGb are measured with assays measuring hCG and hCGb together, the results are expressed in IU based on the hCG standard [9]. It is obvious that this practice is scientifically unacceptable. The present 5th IS for hCG has been purified from urine even if recombinant hCG (rhCG) is available. As determined by bioassay, the potency of this preparation in relation to mass is higher (about 13 IU/ mg) than that of the previous 3rd and 4th IS (9.3 IU/mg). In order to maintain calibration of immunoassays, the 5th IU has been assigned different values for bioassay and immunoassays, respectively. Interestingly, the bioactivity of recombinant hCG (rhCG) and the 5th IS are very similar [32]. Interestingly, a commercial pharmaceutical preparation of rhCG is not calibrated in IU but in mg taking into account only the peptide part of the molecule [33]. Purification of hormone standards from biological materials is demanding. With the exception of the early standards for insulin, which were purified from bovine and porcine pancreas, peptide and protein hormone standards have been purified from human body fluids or organs, e.g., urine and pituitary. The introduction of recombinant techniques has facilitated reproducible production of homogeneous peptide and protein hormones in sufficient amounts. Thus, many recently approved peptide and protein hormone standards have been produced by recombinant techniques. While recombinant proteins are fairly homogeneous, circulating peptide and protein hormones are mostly heterogeneous. Therefore, recombinant proteins may not be ideal for standardization of all immunoassays. Glycoproteins pose a potential problem for the use of recombinant proteins as standards. Presently, most recombinant proteins are produced in CHO cells, which have a different glycosylation machinery
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
827
than human cells. Thus the glycans of recombinant proteins are different than in their human counterparts [33] and this is likely to cause differences in bioactivity [34]. This may require the use of different standards for bioassays and immunoassays. Glycans are known to affect bioactivity, but they are not very immunogenic and, with few exceptions, monoclonal antibodies used in immunoassays do not recognize glycans [17]. However, a unique antibody, B152, recognizes the core-2 O-glycan on Ser132 of the b subunit of hCG [35]. This so called hyperglyosylated hCG that is produced in early pregnancy and by trophoblastic tumors [36] is of potential diagnostic utility [37]. Most hormone standards approved by WHO have been assigned values in IU based on bioactivity [1]. While this is important for calibration of pharmaceutical preparations, it is important to recognize that immunoassays do not measure bioactivity. The concentrations of highly purified glycoprotein hormones can be determined in substance concentrations and this is the primary method used to express the concentrations of the 1st International Reference Reagents (1st IRR) for human chorionic gonadotropin and its subunits and the 5th IS for hCG. The use of substance concentrations is based on the fact that antibodies detect the molar concentrations of epitopes rather than bioactivity or mass. Thus, molar concentrations correctly reflect the relative concentrations of related substances with different MW, e.g. intact hCG with an MW of 37500 with that of hCGb with a MW of 23000. Introduction of recombinant proteins as standards will facilitate reproducible value assignment based on mass and substance concentrations (mol/L) as well as bioactivity. The International Federation of Clinical Chemistry recommends the use of SI units, i.e., substance concentrations whenever possible but practice varies. Many endocrinology journals require the use of molar concentrations not only for steroids but also for some peptide hormones like insulin and parathormone. Effect of assay design on standardization The key reagents in immunoassays are antiserum or antibodies, calibrator, and tracer. The characteristics and quality of antisera and antibodies vary. Polyclonal antisera are of limited availability, and when a manufacturer has to replace the antiserum, the characteristics of the assay change, which may affect the results. The introduction of monoclonal antibodies (MAbs) has eliminated one source of variation and, in MAb-based assays, thus lot-to-lot variation is now a smaller problem than before. Reference methods MS can be used as a reference method for hormones of relatively low molecular weight (MW) like thyroid and steroid hormones as well as some peptide hormones. Recently, the use of MS has become possible also for determination of relatively large peptide hormones like hepcidin [38] and insulin [3]. With the development of MS technology, determination of protein hormones is becoming possible even as routine methods [38]. However, with increasing MW, protein hormones tend to become heterogeneous and standardization of these is challenging irrespective of the assay method used. Heterogeneity of protein hormones may be caused by variable glycosylation, expression of splice variants and occurrence of precursors, fragments and complexes of hormones with their soluble receptors or other binding proteins. Often only some of these forms are biologically active, and in most cases, these are clinically relevant. However, in other cases, a fragment or hormone subunit may be of diagnostic value [39]. Thus different problems complicate standardization of different assays. Determination of protein hormones by mass spectrometry is usually based the use of immunoextraction of the hormone, usually with antibody-coated magnetic beads, tryptic digestion of the bound hormone and mass spectrometric determination of one or several fragments. The use of an internal standard labeled with a stable isotope is essential for quantitative MS methods. Preparation of isotopelabeled large peptides and proteins is demanding and expensive. Therefore, internal standards corresponding to tryptic peptides have been used for protein hormones like growth hormone and hCG [2,4]. However, this does not compensate for incomplete extraction or digestion. Thus, further development in the preparation of internal standards is needed before MS can be used as a reference method for proteins. Recombinant proteins can be labeled with stable isotopes and serve as internal standards. A method called Protein Standard Absolute Quantification (PSAQ) has been successfully
828
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
used to prepare standards for cardiac markers [40]. This and other similar approaches promise to facilitate development of MS-based reference methods. However, this approach may not solve all problems, e.g., those caused by heterogeneous hormones, free subunits and fragments. It has been speculated whether MS could replace immunoassays for determination of protein hormones like hCG. However, at this stage MS is not considered a realistic alternative to immunoassay for routine protein hormone determinations [41]. Traceability In vitro diagnostic (IVD) measurements of hormones should follow the EU Directive 98/79/EC, which requires traceability of values assigned to commercial calibrators to higher-order reference materials or methods [42]. Ideally, the highest order primary standard in the traceability chain is defined using a primary method, or if this is not practical a primary ratio method or method of highest achievable metrological order, such as isotope dilution mass spectrometry (ID-MS) [43]. The IVD directive is intended to ascertain that hormone assays give correct and identical results. An individual patient result should relate back to a national or international standard through an unbroken chain of comparisons. However, while most manufacturers report that their assays have been calibrated according to the Directive, the problem for steroid hormones is that the immunoassays lack sensitivity and specificity [13], while for peptide and protein assays the problems are related to lack of reference methods and commutability of the standards spiked in a matrix [15,24,44]. No reference methods have yet been approved for large peptide and protein hormone determinations but several candidate methods based on immunocapture and ID-MS have recently been described [2,3,45,46]. If a definitive method is not available, the best method available may be determined to be the reference method [8]. This requires consensus regarding the method to use and so far this approach has not been utilized. However, even without reference methods, introduction of more homogeneous and well-characterized standards has facilitated considerable improvement in the standardization of several hormone determinations. When reference methods are not available, assay manufacturers appear to calibrate their assays according to the market leader, which does facilitate harmonization and in the best case improved standardization. Summary Determination of hormone concentrations is presently mainly performed by immunoassay on automatic analyzers. The quality of these assays is variable and improvement is needed for especially steroid hormones. The results of these assays is in most cases questionable. Because steroid hormones now can be determined by ID-MS, the use of this approach for routine clinical applications should be seriously considered. Large peptide and protein hormones can be determined with adequate sensitivity and specificity on automatic analyzers but there are too large differences in results obtained by different methods. Recent developments in protein expression and protein determinations by ID-MS will facilitate development of reference methods for peptide and protein hormones. Although protein hormone heterogeneity still is a problem, these developments will facilitate improved standardization.
Practice points 1. Determination of steroid hormone concentrations is unsatisfactory with most automated immunoassays. These should be replaced by mass spectrometric methods. 2. Reference methods are being developed for peptide and protein hormones. 3. Use of substance concentrations for peptide and protein hormones should be introduced as an alternative to International Units based on bioactivity.
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
829
Research agenda 1. Production of hormone standards with recombinant techniques. 2. Development of standards labeled with stable isotopes and establishment of reference methods for peptide and protein hormones based on MS. 3. Establishment of conversion factors for IU to substance concentrations (mol). Role of the funding source None to declare. Conflict of interest Ulf-Håkan Stenman has served as a consultant for PerkinElmer Wallac. Acknowledgments Ulf-Håkan Stenman served as the first chairman of the IFCC working group for standardization of hCG determinations. The input of the members of the working group is gratefully acknowledged; Peter Berger, Jean-Michel Bidart, Steven Birken, Rob Norman and Cathie Sturgeon. Antibody B152 for determination of hCG-h was kindly provided by Steven Birken. U-HS also chaired working groups established by ISOBM for mapping of the epitopes of hCG. Peter Berger, Elisabeth Paus, Steven Birken and Jean-Michel were key persons in this group. References [1] WHO international biological reference preparations. Endocrinological substances. http://www.who.int/bloodproducts/ catalogue/EndoFeb2013.pdf (assessed October 20, 2013). [2] Arsene CG, Henrion A, Diekmann N, et al. Quantification of growth hormone in serum by isotope dilution mass spectrometry. Anal Biochem 2010;401:228–35. [3] Miller WG, Thienpont LM, Van Uytfanghe K, et al. Toward standardization of insulin immunoassays. Clin Chem 2009;55: 1011–8. [4] Lund H, Snilsberg AH, Halvorsen TG, et al. Comparison of newly developed immuno-ms method with existing Delfia immunoassay for human chorionic gonadotropin determination in doping analysis. Bioanalysis 2013;5:623–30. *[5] Sturgeon C. Quality assessment of hormone determinations. Best Pract Res Clin Endocrinol Metab 2013;27. [6] Yalow RS, Berson SA. Assay of plasma insulin in human subjects by immunological methods. Nature 1959;184(Suppl. 21): 1648–9. [7] Ekins RP. The estimation of thyroxine in human plasma by an electrophoretic technique. Clinica Chim Acta Int J Clin Chem 1960;5:453–9. *[8] Stenman UH, Bidart JM, Birken S, et al. Standardization of protein immunoprocedures. Choriogonadotropin (CG). Scand J Clin Lab Invest Suppl 1993;216:42–78. [9] Sturgeon CM, Berger P, Bidart JM, et al. Differences in recognition of the 1st who international reference reagents for hCGrelated isoforms by diagnostic immunoassays for human chorionic gonadotropin. Clin Chem 2009;55:1484–91. *[10] Rosner W, Hankinson SE, Sluss PM, et al. Challenges to the measurement of estradiol: an endocrine society position statement. J Clin Endocrinol Metab 2013;98:1376–87. *[11] Taieb J, Mathian B, Millot F, et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography-mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003;49:1381–95. *[12] Keevil BG. Novel liquid chromatography tandem mass spectrometry [LC-MS/MS) methods for measuring steroids. Best Pract Res Clin Endocrinol Metab 2013;27:663–74. *[13] Handelsman DJ, Wartofsky L. Requirement for mass spectrometry sex steroid assays in the journal of clinical endocrinology and metabolism. J Clin Endocrinol Metab 2013;98:3971–3. [14] Miles LE, Hales CN. Immunoradiometric assay of human growth hormone. Lancet 1968;2:492–3. *[15] Stenman UH. Immunoassay standardization: is it possible, who is responsible, who is capable? Clin Chem 2001;47:815–20. [16] Pettersson K, Ding YQ, Huhtaniemi I. An immunologically anomalous luteinizing hormone variant in a healthy woman. J Clin Endocrinol Metab 1992;74:164–71. *[17] Berger P, Paus E, Hemken PM, et al. Candidate epitopes for measurement of hCG and related molecules: the second ISOBM TD-7 workshop. Tumour Biol 2013 (in press). [18] Bjorkhem I, Blomstrand R, Lantto O, et al. absolute methods in clinical chemistry: application of mass fragmentography to high-accuracy analyses. Clin Chem 1976;22:1789–801.
830
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
[19] Turpeinen U, Hamalainen E, Haanpaa M, et al. Determination of salivary testosterone and androstendione by liquid chromatography-tandem mass spectrometry. Clinica Chim Acta Int J Clin Chem 2012;413:594–9. *[20] Rosner W, Auchus RJ, Azziz R, et al. Position statement: utility, limitations, and pitfalls in measuring testosterone: an endocrine society position statement. J Clin Endocrinol Metab 2007;92:405–13. [21] Thienpont LM, Van Uytfanghe K, Poppe K, et al. Determination of free thyroid hormones. Best Pract Res Clin Endocrinol Metab 2013;27:689–700. [22] Faix JD. Principles and pitfalls of free hormone measurements. Best Pract Res Clin Endocrinol Metab 2013;27:631–45. [23] Hjorteberg R, Frystyk J. Determination of IGFs and their binding proteins. Best Pract Res Clin Endocrinol Metab 2013;27. *[24] Boulo S, Hanisch K, Bidlingmaier M, et al. Gaps in the traceability chain of human growth hormone measurements. Clin Chem 2013;59:1074–82. [25] Bidlingmaier M, Suhr J, Ernst A, et al. High-sensitivity chemiluminescence immunoassays for detection of growth hormone doping in sports. Clin Chem 2009;55:445–53. [26] Fahie-Wilson M, Smith TP. Determination of prolactin: the macroprolactin problem. Best Pract Res Clin Endocrinol Metab 2013;27:725–42. [27] Lantto O, Bjorkhem I, Blomstrand R, et al. Interlaboratory evaluation of four ria kits for determination of plasma cortisol, with special reference to accuracy: influence of matrix in calibration standards. Clin Chem 1980;26:1899–902. [28] Hunter WM, Budd PS. Circulating antibodies to ovine and bovine immunoglobulin in healthy subjects: a hazard for immunoassays. Lancet 1980;2:1136. [29] Bolstad N, Warren DJ, Nustad K. Heterophilic antibody interference in immunometric assays. Best Pract Res Clin Endocrinol Metab 2013;27:647–61. [30] Miles AA, Mussett MV, Perry WL. Third international standard for insulin. Bull World Health Organ 1952;7:445–59. [31] Stenman U. hCG standards. In: Cole LA, Butler S, editors. Human chorionic gonadotropin (hCG). Amsterdam: Elsevier; 2010. p. 264–71. [32] Stenman UH, Hotakainen K, Alfthan H. Gonadotropins in doping: pharmacological basis and detection of illicit use. Br J Pharmacol 2008;154:569–83. [33] Gervais A, Hammel YA, Pelloux S, et al. Glycosylation of human recombinant gonadotrophins: characterization and batchto-batch consistency. Glycobiology 2003;13:179–89. [34] Rose MP. Follicle stimulating hormone international standards and reference preparations for the calibration of immunoassays and bioassays. Clinica Chim Acta Int J Clin Chem 1998;273:103–17. [35] Birken S, Yershova O, Myers RV, et al. Analysis of human choriogonadotropin core 2 o-glycan isoforms. Mol Cell Endocrinol 2003;204:21–30. [36] Kovalevskaya G, Birken S, Kakuma T, et al. Differential expression of human chorionic gonadotropin (hCG) glycosylation isoforms in failing and continuing pregnancies: preliminary characterization of the hyperglycosylated hCG epitope. J Endocrinol 2002;172:497–506. [37] Keikkala E, Vuorela P, Laivuori H, et al. First trimester hyperglycosylated human chorionic gonadotrophin in serum – a marker of early-onset preeclampsia. Placenta 2013;34:1059–65. [38] Itkonen O, Parkkinen J, Stenman UH, et al. Preanalytical factors and reference intervals for serum hepcidin LC-MS/MS method. Clinica Chim Acta Int J Clin Chem 2012;413:696–701. [39] Lempiainen A, Stenman UH, Blomqvist C, et al. Free beta-subunit of human chorionic gonadotropin in serum is a diagnostically sensitive marker of seminomatous testicular cancer. Clin Chem 2008;54:1840–3. [40] Huillet C, Adrait A, Lebert D, et al. Accurate quantification of cardiovascular biomarkers in serum using protein standard absolute quantification (PSAQ) and selected reaction monitoring. Mol Cell Proteomics 2012;11:M111008235. [41] Gronowski AM. Clinical assays for human chorionic gonadotropin: what should we measure and how? Clin Chem 2009; 55:1900–4. [42] Directive 98/79/ec of the European parliament and of the council of 27 October 1998 on in vitro diagnostic medical devices. Off J Eur Communities 1998 Dec 7;L 331:1–37. [43] Pritchard C, Quaglia M, Mussell C, et al. traceable absolute protein quantification of somatropin that allows independent comparison of somatropin standards. Clin Chem 2009;55:1984–90. [44] Stenman UH. Pitfalls in hormone determinations. Best Pract Res Clin Endocrinol Metab 2013;27:629–30. [45] Fierens C, Stockl D, Thienpont LM, et al. Strategies for determination of insulin with tandem electrospray mass spectrometry: implications for other analyte proteins? Rapid Commun Mass Spectrom RCM 2001;15:1433–41. [46] Thomas A, Schanzer W, Delahaut P, et al. Immunoaffinity purification of peptide hormones prior to liquid chromatography-mass spectrometry in doping controls. Methods 2012;56:230–5.
Contents lists available at ScienceDirect
Best Practice & Research Clinical Endocrinology & Metabolism journal homepage: www.elsevier.com/locate/beem
7
Standardization of hormone determinations Ulf-Håkan Stenman, M.D., Ph.D., Professor emeritus Department of Clinical Chemistry, Biomedicum, Helsinki University, PB 63, FIN-00014 Helsinki, Finland
Keywords: standardization standards reference methods hormones immunoassay mass spectrometry
Standardization of hormone determinations is important because it simplifies interpretation of results and facilitates the use of common reference values for different assays. Progress in standardization has been achieved through the introduction of more homogeneous hormone standards for peptide and protein hormones. However, many automated methods for determinations of steroid hormones do not provide satisfactory result. Isotope dilution-mass spectrometry (ID-MS) has been used to establish reference methods for steroid hormone determinations and is now increasingly used for routine determinations of steroids and other low molecular weight compounds. Reference methods for protein hormones based on MS are being developed and these promise to improve standardization. Ó 2013 Elsevier Ltd. All rights reserved.
Introduction The aim of standardization is to obtain identical and correct results with different methods. The expression “harmonization” is often used as an alternative to standardization. However, while different harmonized methods may provide similar results, they may all be biased. Therefore, the goal should be to achieve standardization, but if standards are not available, harmonization is the only option. WHOapproved International Standards (IS) for most clinically important hormones are available from the National Institute for Biological Standards and Control (NIBSC) in the UK. However, only a few standards are available for cytokines and growth factors [1]. Standardization is feasible if a standard and a reference method are available. The reference method should preferably be a definitive method, in practice mass spectrometry (MS), which presently can be used for relatively small molecules like steroid and thyroid hormones and also for some peptide hormones [2–4]. Most hormone determinations are performed by immunoassay but MS is increasingly used for routine determination of steroid hormones. Standardization of immunoassays is challenging and results from quality assessment schemes show that for many methods, between-method variation is not yet on an adequate level [5]. E-mail address: ulf-hakan.stenman@helsinki.fi. 1521-690X/$ – see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.beem.2013.10.007
824
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
Development of hormone immunoassays Routine determination of hormones became possible with the development of the first radioimmunoassay for insulin by Yalow and Berson in 1959 [6] and an assay of thyroxin by use of thyroxin binding globulin by Ekins in 1960 [7]. During the following decade, immunoassays for most peptide and protein hormones were developed using antisera produced in-house. Some of these antisera were widely distributed and when purified protein hormone preparations became available from NIH and WHO, fairly well standardized assays could be established. When commercial assays gradually became available, the need for standardization became obvious. Assay manufacturers used different antisera, calibrators and assay conditions and this caused substantial variation in the results obtained. Thus twenty years ago the results obtained by different assays for human chorionic gonadotropin could vary by more than 10-fold [8]. Presently, hCG assays give results that differ by less than 50% from the mean [9]. Initially, binding inhibition assays were used to measure both small and large hormones. In these, the hormone to be measured compete with radiolabeled hormone for binding to the antiserum. Binding inhibition assays are still used to measure small hormones. In these, assay sensitivity increases with decreasing concentrations of antiserum and increasing incubation time. Optimal detection limits requires assay times of hours or even days. However, presently, most hormone immunoassays are performed with automated analyzers, in which incubations times have been minimized, often to less than 10–15 min, in order to optimize throughput. Unfortunately this has lead to impaired quality of steroid hormone assays [10–12]. Thus, it has been decided that steroid hormone determinations have to be performed by mass spectrometry in studies to be published in the Journal of Clinical Endocrinology and Metabolism [13]. Hormones with an MW larger than 3 kD can determined by immunometric, also called sandwich assays [14]. In these, a capture antibody (or antiserum) attached to a solid phase, i.e., a tube wall or a particle, binds the hormone through one epitope and a detector antibody, which is labeled with a radioisotope, an enzyme, a fluorophore or a luminescent molecule, binds to another epitope. A polyclonal antiserum containing antibodies to many epitopes on the antigen and can be used both as capture and tracer antibody. This approach is seldom used in commercial assays, but combinations of monoclonal and polyclonal antibodies are used. The detection limit of sandwich assays improves with increasing antibody concentrations. Thanks to this and the use of sensitive detection methods, very low detection limits can be achieved in spite of short incubation times. Problems for standardization Hormone concentrations Standardization of hormone determinations is complicated by a number of problems. First, the concentrations of hormones in circulation are very low in comparison to those of other similar substances. In healthy subjects, the concentrations of most peptide and protein hormones are in the range 1–50 pmol/l. The concentrations of the most abundant serum protein, albumin, is about 0.7 mol/l, which is 10–100 millionfold that of typical protein hormones. The concentrations of steroid hormones are in the range 20–300 pmol/ l for estradiol to 50–500 nmol/l for cortisol. These should be compared to the concentration of cholesterol, 5 mmol/l, which is 10,000–100,000,000–fold higher than those of steroid hormones. It is inevitable that the presence of a huge excess of similar substances affects hormone determinations by causing nonspecific interference. Considering these facts, it is amazing how specific and sensitive immunoassays can be, but it explains some problems with steroid hormone assays with automatic analyzers [13,15]. Hormone heterogeneity Most peptide and protein hormones occur in different forms in circulation, e.g., 22 and 20 kD growth hormone, intact hCG and its free b subunit (hCGb) or intact paratahormone and hormonally inactive, truncated forms. In order to standardize assays for such hormones, it is necessary to define which forms should be measured and to use antibodies that detect relevant epitopes. Genetic polymorphism of protein hormones is fairly common and a variant of luteinizing hormone (LH), which differs with respect to two amino acids (Trp8Arg and Ile15Thr) is not detected by some antibodies [16].
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
825
Cross-reacting hormones and binding proteins Standardization of steroid hormone determinations is complicated by the occurrence of closely related forms that crossreact with the antisera used. Some antibodies to hCG also recognize LH and vice versa. Specific assays can be developed by careful selection of epitope reactivity of the antibodies used [17]. The difference in structure between many steroid hormones, e.g., estradiol and estrone or testosterone are actually very small and therefore, some cross-reactivity between related hormones is inevitable. Most steroid hormones occur in circulation in complex with steroid-binding proteins that compete with antibodies. In many early steroid hormone immunoassays, this problem was eliminated by extraction and separation of related hormones by chromatography before the immunoassay. This type if pre-purification is not compatible with automatic analyzers. In these, dissociation of steroids from binding proteins is achieved by using agents that block binding to endogenous binding proteins. The incubation time with antisera and labels is short, often less than 10–15 min. In this time, binding equilibrium is not achieved which earlier was considered important. Thus is it not surprising that the quality of most automated steroid hormone assays is unsatisfactory [10–13]. Highly specific mass spectrometric (MS) methods have been used for decades as reference methods for determination of steroid hormones and other small molecules [18]. Thanks to recent advances in MS technology, these methods are now increasingly used for routine determinations. While the cost of MS still is higher than that of immunoassays, the better validity of MS often justifies its use [12,19]. A misleading result of an immunoassay often causes high additional costs by leading to unnecessary imaging and other examinations [15]. A problem with steroid and thyroid hormone determinations is that the concentration of the free, or bioavailable, fraction of the hormone is physiologically and clinically most relevant. Immunoassays for direct measurement of the concentrations of free steroid hormones have not been found to be reliable [20], but assays of free thyroid hormones are generally considered more relevant than the total hormone concentrations. The concentrations of free hormones can be determined by using dialysis or ultrafiltration. When combined with MS, these methods can used to establish reference methods [21,22]. Many peptide and some protein hormones occur in complex with specific binding proteins in circulation. Insulin-like growth factors (IGFs) occur mainly in complex with IGF-binding proteins (IGFBPs). Thus IGF-1 is bound to IGFBP-3, from which it is dissociated at low pH before assay. Rebinding is prevented by addition of an excess of IGF-2 before assay of total IGF-1. IGF-2 is also bound to other IGFBPs [23]. About 10% of growth hormone (GH) in plasma occurs in complex with the growth hormone binding protein (GHBBP) but the concentrations of GHBBP varies. Some assays recognize this form while other ones do not. This causes method-dependent variation in the results [24]. Interestingly, immunoassays recognizing naturally occurring and recombinant GH differently can be used to detect administration of recombinant GH for doping purposes [25]. Many peptide hormones, e.g., insulin, occur free in circulation unless bound to antibodies induced by injection of hormone analogs for therapeutic use, or occur spontaneously as often is the case with prolactin [26]. Matrix effects The composition of the sample affects the reaction between antigen and antibody and thus the results of hormone determinations. Therefore, the calibrators should be prepared in the same matrix as the sample, usually serum or plasma. Ideally, the calibrators are prepared by spiking serum devoid of the analyte with standards of desired concentrations. Sometimes, analyte-free serum can be obtained from sick people, but this is seldom possible in practice. Protein and peptide hormones can be removed by affinity chromatography using antibodies, but removal is seldom complete and leakage of antibodies disturbing the assay is hard to avoid. Serum free of steroid hormones can be prepared by charcoal absorption, which however removes many low MW components, e.g., cross-reacting steroids, that affect steroid hormone determinations. Thus it is obvious that preparation of the matrix for the calibrators is a crucial step in the development of an immunoassay [27]. Manufacturers seldom provide information about how this has been done. A common matrix problem is caused by human antibodies against animal immunoglobulins that are present at such concentrations that they disturb the assay. Circulating antibodies against ovine and
826
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
bovine immunoglobulin are actually quite common, obviously as a result of the ingestion of animal milk and meat [28]. In most cases the interference can be prevented by use of animal immunoglobulin in the assay buffer. This problem is discussed in detail elsewhere in the accompanying issue [29]. While the composition of serum (and plasma) is fairly constant, that of urine varies considerably. Due to variation of urinary flow rate, the density of urine varies from about 1.002 to more than 1.030, i.e. about 15-fold. The content of proteins, steroids and other solutes varies accordingly. Therefore, the measured concentrations of analytes in urine is corrected to correspond to a density of 1.020. Thus the result is multiplied by 2 if density is 1.010 and with 4 when density is 1.005. The salt concentration of urine is reflected by the density, and in dilute urine, salt concentration is lower than in serum, while it is higher when density is high. Furthermore, urine pH varies and together with variation in salt concentration this may have a considerable effect on the antigen–antibody reaction. This may cause a considerable bias if the sample volume in relation to total incubation volume is large, i.e., more than 10–15%. Therefore, rather few immunoassays have been validated for analysis of urine samples. Analysis of urinary steroid concentrations is especially complicated because of the large excess of steroid conjugates in urine. Therefore, analysis of drugs and hormones used for doping is performed by rigorously controlled methods based on mass spectrometry. Standards and units The first hormone standard for insulin was introduced 1925 to facilitate calibration of the bioactivity of preparations for treatment of diabetes. The standard was assigned values in arbitrary international units (IU) based on biological activity. The present international standard for insulin is still calibrated in IU, 1 mg corresponding to 26 IU [30]. These standards have later been used for immunoassay standardization, but recombinant insulin calibrated in mass or mol/l is increasingly used to calibrate immunoassays. The first standard for hCG established in 1939 was a crude extract of urinary proteins from pregnant women. It was assigned a value in units based on its bioactivity, and each subsequent hCG standards has been calibrated against the previous one using biological methods measuring the potency of the standards. With improving purity of the hCG standards, the biological activity in relation to mass has increased. At the same time, the immunoreactivity in relation to bioactivity has also changed (see below) [31]. Arbitrary units have also been assigned to hormone subunits and fragments lacking bioactivity. Therefore, the first standards for the free subunits of hCG, i.e., hCGa and hCGb were assigned units based on the protein content of the preparations. Thus 1 IU of the 3rd IS for hCGa or hCGb is equivalent to 1 mg [31]. These units are quite different from those for hCG, in which 1 mg corresponds to 9.3 IU. These units are not used in practice and, when concentrations of hCGb are measured with assays measuring hCG and hCGb together, the results are expressed in IU based on the hCG standard [9]. It is obvious that this practice is scientifically unacceptable. The present 5th IS for hCG has been purified from urine even if recombinant hCG (rhCG) is available. As determined by bioassay, the potency of this preparation in relation to mass is higher (about 13 IU/ mg) than that of the previous 3rd and 4th IS (9.3 IU/mg). In order to maintain calibration of immunoassays, the 5th IU has been assigned different values for bioassay and immunoassays, respectively. Interestingly, the bioactivity of recombinant hCG (rhCG) and the 5th IS are very similar [32]. Interestingly, a commercial pharmaceutical preparation of rhCG is not calibrated in IU but in mg taking into account only the peptide part of the molecule [33]. Purification of hormone standards from biological materials is demanding. With the exception of the early standards for insulin, which were purified from bovine and porcine pancreas, peptide and protein hormone standards have been purified from human body fluids or organs, e.g., urine and pituitary. The introduction of recombinant techniques has facilitated reproducible production of homogeneous peptide and protein hormones in sufficient amounts. Thus, many recently approved peptide and protein hormone standards have been produced by recombinant techniques. While recombinant proteins are fairly homogeneous, circulating peptide and protein hormones are mostly heterogeneous. Therefore, recombinant proteins may not be ideal for standardization of all immunoassays. Glycoproteins pose a potential problem for the use of recombinant proteins as standards. Presently, most recombinant proteins are produced in CHO cells, which have a different glycosylation machinery
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
827
than human cells. Thus the glycans of recombinant proteins are different than in their human counterparts [33] and this is likely to cause differences in bioactivity [34]. This may require the use of different standards for bioassays and immunoassays. Glycans are known to affect bioactivity, but they are not very immunogenic and, with few exceptions, monoclonal antibodies used in immunoassays do not recognize glycans [17]. However, a unique antibody, B152, recognizes the core-2 O-glycan on Ser132 of the b subunit of hCG [35]. This so called hyperglyosylated hCG that is produced in early pregnancy and by trophoblastic tumors [36] is of potential diagnostic utility [37]. Most hormone standards approved by WHO have been assigned values in IU based on bioactivity [1]. While this is important for calibration of pharmaceutical preparations, it is important to recognize that immunoassays do not measure bioactivity. The concentrations of highly purified glycoprotein hormones can be determined in substance concentrations and this is the primary method used to express the concentrations of the 1st International Reference Reagents (1st IRR) for human chorionic gonadotropin and its subunits and the 5th IS for hCG. The use of substance concentrations is based on the fact that antibodies detect the molar concentrations of epitopes rather than bioactivity or mass. Thus, molar concentrations correctly reflect the relative concentrations of related substances with different MW, e.g. intact hCG with an MW of 37500 with that of hCGb with a MW of 23000. Introduction of recombinant proteins as standards will facilitate reproducible value assignment based on mass and substance concentrations (mol/L) as well as bioactivity. The International Federation of Clinical Chemistry recommends the use of SI units, i.e., substance concentrations whenever possible but practice varies. Many endocrinology journals require the use of molar concentrations not only for steroids but also for some peptide hormones like insulin and parathormone. Effect of assay design on standardization The key reagents in immunoassays are antiserum or antibodies, calibrator, and tracer. The characteristics and quality of antisera and antibodies vary. Polyclonal antisera are of limited availability, and when a manufacturer has to replace the antiserum, the characteristics of the assay change, which may affect the results. The introduction of monoclonal antibodies (MAbs) has eliminated one source of variation and, in MAb-based assays, thus lot-to-lot variation is now a smaller problem than before. Reference methods MS can be used as a reference method for hormones of relatively low molecular weight (MW) like thyroid and steroid hormones as well as some peptide hormones. Recently, the use of MS has become possible also for determination of relatively large peptide hormones like hepcidin [38] and insulin [3]. With the development of MS technology, determination of protein hormones is becoming possible even as routine methods [38]. However, with increasing MW, protein hormones tend to become heterogeneous and standardization of these is challenging irrespective of the assay method used. Heterogeneity of protein hormones may be caused by variable glycosylation, expression of splice variants and occurrence of precursors, fragments and complexes of hormones with their soluble receptors or other binding proteins. Often only some of these forms are biologically active, and in most cases, these are clinically relevant. However, in other cases, a fragment or hormone subunit may be of diagnostic value [39]. Thus different problems complicate standardization of different assays. Determination of protein hormones by mass spectrometry is usually based the use of immunoextraction of the hormone, usually with antibody-coated magnetic beads, tryptic digestion of the bound hormone and mass spectrometric determination of one or several fragments. The use of an internal standard labeled with a stable isotope is essential for quantitative MS methods. Preparation of isotopelabeled large peptides and proteins is demanding and expensive. Therefore, internal standards corresponding to tryptic peptides have been used for protein hormones like growth hormone and hCG [2,4]. However, this does not compensate for incomplete extraction or digestion. Thus, further development in the preparation of internal standards is needed before MS can be used as a reference method for proteins. Recombinant proteins can be labeled with stable isotopes and serve as internal standards. A method called Protein Standard Absolute Quantification (PSAQ) has been successfully
828
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
used to prepare standards for cardiac markers [40]. This and other similar approaches promise to facilitate development of MS-based reference methods. However, this approach may not solve all problems, e.g., those caused by heterogeneous hormones, free subunits and fragments. It has been speculated whether MS could replace immunoassays for determination of protein hormones like hCG. However, at this stage MS is not considered a realistic alternative to immunoassay for routine protein hormone determinations [41]. Traceability In vitro diagnostic (IVD) measurements of hormones should follow the EU Directive 98/79/EC, which requires traceability of values assigned to commercial calibrators to higher-order reference materials or methods [42]. Ideally, the highest order primary standard in the traceability chain is defined using a primary method, or if this is not practical a primary ratio method or method of highest achievable metrological order, such as isotope dilution mass spectrometry (ID-MS) [43]. The IVD directive is intended to ascertain that hormone assays give correct and identical results. An individual patient result should relate back to a national or international standard through an unbroken chain of comparisons. However, while most manufacturers report that their assays have been calibrated according to the Directive, the problem for steroid hormones is that the immunoassays lack sensitivity and specificity [13], while for peptide and protein assays the problems are related to lack of reference methods and commutability of the standards spiked in a matrix [15,24,44]. No reference methods have yet been approved for large peptide and protein hormone determinations but several candidate methods based on immunocapture and ID-MS have recently been described [2,3,45,46]. If a definitive method is not available, the best method available may be determined to be the reference method [8]. This requires consensus regarding the method to use and so far this approach has not been utilized. However, even without reference methods, introduction of more homogeneous and well-characterized standards has facilitated considerable improvement in the standardization of several hormone determinations. When reference methods are not available, assay manufacturers appear to calibrate their assays according to the market leader, which does facilitate harmonization and in the best case improved standardization. Summary Determination of hormone concentrations is presently mainly performed by immunoassay on automatic analyzers. The quality of these assays is variable and improvement is needed for especially steroid hormones. The results of these assays is in most cases questionable. Because steroid hormones now can be determined by ID-MS, the use of this approach for routine clinical applications should be seriously considered. Large peptide and protein hormones can be determined with adequate sensitivity and specificity on automatic analyzers but there are too large differences in results obtained by different methods. Recent developments in protein expression and protein determinations by ID-MS will facilitate development of reference methods for peptide and protein hormones. Although protein hormone heterogeneity still is a problem, these developments will facilitate improved standardization.
Practice points 1. Determination of steroid hormone concentrations is unsatisfactory with most automated immunoassays. These should be replaced by mass spectrometric methods. 2. Reference methods are being developed for peptide and protein hormones. 3. Use of substance concentrations for peptide and protein hormones should be introduced as an alternative to International Units based on bioactivity.
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
829
Research agenda 1. Production of hormone standards with recombinant techniques. 2. Development of standards labeled with stable isotopes and establishment of reference methods for peptide and protein hormones based on MS. 3. Establishment of conversion factors for IU to substance concentrations (mol). Role of the funding source None to declare. Conflict of interest Ulf-Håkan Stenman has served as a consultant for PerkinElmer Wallac. Acknowledgments Ulf-Håkan Stenman served as the first chairman of the IFCC working group for standardization of hCG determinations. The input of the members of the working group is gratefully acknowledged; Peter Berger, Jean-Michel Bidart, Steven Birken, Rob Norman and Cathie Sturgeon. Antibody B152 for determination of hCG-h was kindly provided by Steven Birken. U-HS also chaired working groups established by ISOBM for mapping of the epitopes of hCG. Peter Berger, Elisabeth Paus, Steven Birken and Jean-Michel were key persons in this group. References [1] WHO international biological reference preparations. Endocrinological substances. http://www.who.int/bloodproducts/ catalogue/EndoFeb2013.pdf (assessed October 20, 2013). [2] Arsene CG, Henrion A, Diekmann N, et al. Quantification of growth hormone in serum by isotope dilution mass spectrometry. Anal Biochem 2010;401:228–35. [3] Miller WG, Thienpont LM, Van Uytfanghe K, et al. Toward standardization of insulin immunoassays. Clin Chem 2009;55: 1011–8. [4] Lund H, Snilsberg AH, Halvorsen TG, et al. Comparison of newly developed immuno-ms method with existing Delfia immunoassay for human chorionic gonadotropin determination in doping analysis. Bioanalysis 2013;5:623–30. *[5] Sturgeon C. Quality assessment of hormone determinations. Best Pract Res Clin Endocrinol Metab 2013;27. [6] Yalow RS, Berson SA. Assay of plasma insulin in human subjects by immunological methods. Nature 1959;184(Suppl. 21): 1648–9. [7] Ekins RP. The estimation of thyroxine in human plasma by an electrophoretic technique. Clinica Chim Acta Int J Clin Chem 1960;5:453–9. *[8] Stenman UH, Bidart JM, Birken S, et al. Standardization of protein immunoprocedures. Choriogonadotropin (CG). Scand J Clin Lab Invest Suppl 1993;216:42–78. [9] Sturgeon CM, Berger P, Bidart JM, et al. Differences in recognition of the 1st who international reference reagents for hCGrelated isoforms by diagnostic immunoassays for human chorionic gonadotropin. Clin Chem 2009;55:1484–91. *[10] Rosner W, Hankinson SE, Sluss PM, et al. Challenges to the measurement of estradiol: an endocrine society position statement. J Clin Endocrinol Metab 2013;98:1376–87. *[11] Taieb J, Mathian B, Millot F, et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography-mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003;49:1381–95. *[12] Keevil BG. Novel liquid chromatography tandem mass spectrometry [LC-MS/MS) methods for measuring steroids. Best Pract Res Clin Endocrinol Metab 2013;27:663–74. *[13] Handelsman DJ, Wartofsky L. Requirement for mass spectrometry sex steroid assays in the journal of clinical endocrinology and metabolism. J Clin Endocrinol Metab 2013;98:3971–3. [14] Miles LE, Hales CN. Immunoradiometric assay of human growth hormone. Lancet 1968;2:492–3. *[15] Stenman UH. Immunoassay standardization: is it possible, who is responsible, who is capable? Clin Chem 2001;47:815–20. [16] Pettersson K, Ding YQ, Huhtaniemi I. An immunologically anomalous luteinizing hormone variant in a healthy woman. J Clin Endocrinol Metab 1992;74:164–71. *[17] Berger P, Paus E, Hemken PM, et al. Candidate epitopes for measurement of hCG and related molecules: the second ISOBM TD-7 workshop. Tumour Biol 2013 (in press). [18] Bjorkhem I, Blomstrand R, Lantto O, et al. absolute methods in clinical chemistry: application of mass fragmentography to high-accuracy analyses. Clin Chem 1976;22:1789–801.
830
U.-H. Stenman / Best Practice & Research Clinical Endocrinology & Metabolism 27 (2013) 823–830
[19] Turpeinen U, Hamalainen E, Haanpaa M, et al. Determination of salivary testosterone and androstendione by liquid chromatography-tandem mass spectrometry. Clinica Chim Acta Int J Clin Chem 2012;413:594–9. *[20] Rosner W, Auchus RJ, Azziz R, et al. Position statement: utility, limitations, and pitfalls in measuring testosterone: an endocrine society position statement. J Clin Endocrinol Metab 2007;92:405–13. [21] Thienpont LM, Van Uytfanghe K, Poppe K, et al. Determination of free thyroid hormones. Best Pract Res Clin Endocrinol Metab 2013;27:689–700. [22] Faix JD. Principles and pitfalls of free hormone measurements. Best Pract Res Clin Endocrinol Metab 2013;27:631–45. [23] Hjorteberg R, Frystyk J. Determination of IGFs and their binding proteins. Best Pract Res Clin Endocrinol Metab 2013;27. *[24] Boulo S, Hanisch K, Bidlingmaier M, et al. Gaps in the traceability chain of human growth hormone measurements. Clin Chem 2013;59:1074–82. [25] Bidlingmaier M, Suhr J, Ernst A, et al. High-sensitivity chemiluminescence immunoassays for detection of growth hormone doping in sports. Clin Chem 2009;55:445–53. [26] Fahie-Wilson M, Smith TP. Determination of prolactin: the macroprolactin problem. Best Pract Res Clin Endocrinol Metab 2013;27:725–42. [27] Lantto O, Bjorkhem I, Blomstrand R, et al. Interlaboratory evaluation of four ria kits for determination of plasma cortisol, with special reference to accuracy: influence of matrix in calibration standards. Clin Chem 1980;26:1899–902. [28] Hunter WM, Budd PS. Circulating antibodies to ovine and bovine immunoglobulin in healthy subjects: a hazard for immunoassays. Lancet 1980;2:1136. [29] Bolstad N, Warren DJ, Nustad K. Heterophilic antibody interference in immunometric assays. Best Pract Res Clin Endocrinol Metab 2013;27:647–61. [30] Miles AA, Mussett MV, Perry WL. Third international standard for insulin. Bull World Health Organ 1952;7:445–59. [31] Stenman U. hCG standards. In: Cole LA, Butler S, editors. Human chorionic gonadotropin (hCG). Amsterdam: Elsevier; 2010. p. 264–71. [32] Stenman UH, Hotakainen K, Alfthan H. Gonadotropins in doping: pharmacological basis and detection of illicit use. Br J Pharmacol 2008;154:569–83. [33] Gervais A, Hammel YA, Pelloux S, et al. Glycosylation of human recombinant gonadotrophins: characterization and batchto-batch consistency. Glycobiology 2003;13:179–89. [34] Rose MP. Follicle stimulating hormone international standards and reference preparations for the calibration of immunoassays and bioassays. Clinica Chim Acta Int J Clin Chem 1998;273:103–17. [35] Birken S, Yershova O, Myers RV, et al. Analysis of human choriogonadotropin core 2 o-glycan isoforms. Mol Cell Endocrinol 2003;204:21–30. [36] Kovalevskaya G, Birken S, Kakuma T, et al. Differential expression of human chorionic gonadotropin (hCG) glycosylation isoforms in failing and continuing pregnancies: preliminary characterization of the hyperglycosylated hCG epitope. J Endocrinol 2002;172:497–506. [37] Keikkala E, Vuorela P, Laivuori H, et al. First trimester hyperglycosylated human chorionic gonadotrophin in serum – a marker of early-onset preeclampsia. Placenta 2013;34:1059–65. [38] Itkonen O, Parkkinen J, Stenman UH, et al. Preanalytical factors and reference intervals for serum hepcidin LC-MS/MS method. Clinica Chim Acta Int J Clin Chem 2012;413:696–701. [39] Lempiainen A, Stenman UH, Blomqvist C, et al. Free beta-subunit of human chorionic gonadotropin in serum is a diagnostically sensitive marker of seminomatous testicular cancer. Clin Chem 2008;54:1840–3. [40] Huillet C, Adrait A, Lebert D, et al. Accurate quantification of cardiovascular biomarkers in serum using protein standard absolute quantification (PSAQ) and selected reaction monitoring. Mol Cell Proteomics 2012;11:M111008235. [41] Gronowski AM. Clinical assays for human chorionic gonadotropin: what should we measure and how? Clin Chem 2009; 55:1900–4. [42] Directive 98/79/ec of the European parliament and of the council of 27 October 1998 on in vitro diagnostic medical devices. Off J Eur Communities 1998 Dec 7;L 331:1–37. [43] Pritchard C, Quaglia M, Mussell C, et al. traceable absolute protein quantification of somatropin that allows independent comparison of somatropin standards. Clin Chem 2009;55:1984–90. [44] Stenman UH. Pitfalls in hormone determinations. Best Pract Res Clin Endocrinol Metab 2013;27:629–30. [45] Fierens C, Stockl D, Thienpont LM, et al. Strategies for determination of insulin with tandem electrospray mass spectrometry: implications for other analyte proteins? Rapid Commun Mass Spectrom RCM 2001;15:1433–41. [46] Thomas A, Schanzer W, Delahaut P, et al. Immunoaffinity purification of peptide hormones prior to liquid chromatography-mass spectrometry in doping controls. Methods 2012;56:230–5.
