Accuracy of Serum IgM and IgA Monoclonal Protein Measurements by Densitometry

Transcription

1 160 Annals of Clinical & Laboratory Science, vol. 33, no. 2, 2003 Accuracy of Serum IgM and IgA Monoclonal Protein Measurements by Densitometry C. Howard Tseng, 1 Chin-Yung Chang, 2 Kevin S. Liu, 3 and Frank J. Liu 2 1Department of Pathology, Chung Shan Medical University Hospital, Taichung, Taiwan; 2 Department of Laboratory Medicine, The University of Texas M.D. Anderson Cancer Center, Houston, Texas; 3 Department of Family Medicine, Scott & White Clinics, Texas A & M University College of Medicine, Waco, Texas Abstract. We previously reported that proper dilution of sera that contain IgG monoclonal proteins (Mproteins) is necessary for accurate quantitation of serum albumin and M-protein concentrations separated by serum protein electrophoresis (SPE) using the Beckman PARAGON agarose electrophoresis system. We now report the significance of pre-electrophoretic serum dilution for M-protein quantitation of sera from patients with IgA and IgM monoclonal gammopathy. We measured M-proteins by SPE in 82 serum samples from 29 patients with IgA and 72 samples from 23 patients with IgM monoclonal gammopathy. The serum M-protein concentrations (mean ± SD) at 1:5, 1:10, and 1:20 dilutions (v/v) for all samples of both types were 49.7 ± 12.9, 49.1 ± 13.1, and 47.8 ± 13.0 g/l, respectively. Thirty-two (20.8%) of 154 sera showed varying degrees of increase in M-protein concentrations with serum dilutions higher than 1:5; only 8 (5.2%) showed an increase 3 SDs. B y SPE, the M-protein concentration (mean ± SD) of these 8 sera at 1:5, 1:10, and 1:20 dilutions were 52.6 ± 7.8, 57.1 ± 7.2, and 57.6 ± 7.1 g/l, respectively; the albumin concentrations (mean ± SD) were 41.4 ± 4.4, 37.9 ± 3.8, and 37.1 ± 2.9 g/l, respectively. The corresponding albumin concentration (mean ± SD) was 36.8 ± 3.7 g/l, assayed by the bromcresol green dye-binding method. These 8 samples were obtained from 3 patients, 2 with IgM kappa and 1 with IgA lambda monoclonal gammopathy. On the electrophoresis membranes, the M-protein bands of these 8 samples were narrow, thin, and dense; upon scanning, they appeared taller and thinner than the corresponding albumin bands. The samples of this subset contained relatively high concentrations of M-protein and total serum protein. We conclude that a pre-electrophoretic dilution of 1:5 (v/v) is adequate for most sera with IgA or IgM M-proteins. However, 1:10 or 1:20 dilution is occasionally required for a subset of sera with IgA or IgM M-proteins that show an unusually thin, narrow, and dense M-protein band and have high total serum protein content. (received 29 October 2002; accepted 9 November 2002) Keywords: monoclonal gammopathy, IgA, IgM, M-protein, serum protein electrophoresis Introduction Monoclonal proteins (M-proteins) frequently appear in the sera of patients with multiple myeloma, other plasma cell dyscrasias, and some lymphoproliferative disorders [1]. Serial measurement of serum M-proteins can reflect both the tumor burden and the clinical course of disease after treatment Address correspondence to Frank J. Liu, M.D., 1910 Gray Birch Drive, Sugar Land, Texas 77479, USA; tel ; fax ; frankjliu_1139@yahoo.com. Dr. Chin- Yung Chang is now affiliated with the Department of Pathology, St. Mary s Medical Center, Hobart, IN, USA. [2,3]. M-proteins are quantified routinely in many laboratories by SPE after total serum protein has been determined by refractometry or protein-dye binding colorimetry [4,5]. We reported that undiluted sera evaluated for M-protein concentration by SPE on cellulose acetate membranes gave underestimation of the M-protein and overestimation of the albumin concentration [6]. Since then, we have used the Beckman SPE system with agarose gel and our experience with this system for IgG monoclonal gammopathy has also been published [7]. We recommended a 1:10 preelectrophoretic serum dilution for sera with total /03/0200/0158 $ by the Association of Clinical Scientists, Inc.

2 Quantitation of serum IgA and IgM M-proteins 161 protein in the range of g/l and a 1:20 dilution for those between g/l; all other samples with lower total protein concentrations can be analyzed with a 1:5 dilution, as recommended by the manufacturer [7]. The current investigation evaluated the dilution requirement for serum samples with IgA and IgM monoclonal gammopathies, which occur less frequently than their IgG counterpart. Our observations show that serum dilution beyond 1:5 is not required in most cases of these two categories. However, there is a small subset of IgA and IgM monoclonal gammopathy with an unusually thin, narrow, and dense M-protein band where a further dilution of 1:10 or 1:20 is required for accurate M- protein determinations. Possible explanations for such exceptions are provided and the different dilution requirements for samples of different monoclonal gammopathies are addressed. Materials and Methods Patients and serum samples. We studied 154 serum samples from 52 patients with IgA or IgM monoclonal gammopathy. There were 82 IgA samples from 29 patients and 72 IgM samples from 23 patients. All patients with IgA monoclonal gammopathy were diagnosed as multiple myeloma. Patients with IgM monoclonal gammopathy were diagnosed as Waldenström s macroglobulinemia, except for 2 patients with chronic lymphocytic leukemia (1 with kappa and 1 with lambda light chains). Total serum protein concentrations of all samples ranged from g/l; for samples with IgA kappa M-protein, g/l (median = 99.5); for samples with IgA lambda M-protein, g/l (median = 105); for samples with IgM kappa M-protein, g/l (median = 102.5); and for samples with IgM lambda M-protein, g/l (median = 101). Determination of serum albumin. Serum albumin concentrations were determined by using the bromocresol green dye-binding method reported by Rodkey [8] and modified in the Kodak Ektachem 700XR analyzer (Eastman Kodak Company, Rochester, NY). The reference value for serum albumin in our laboratory is g/l. Serum protein electrophoresis (SPE). Total serum protein concentration was determined by refractometry (American Optical TS Meter, Cambridge Instruments, Buffalo, NY). When total serum protein by refractometry is 140 g/l, or when the sample is lipemic, icteric, or contains high concentrations of urea nitrogen or glucose, total serum protein is determined by a biuret method on the Kodak Ektachem 750XRC analyzer (Eastman Kodak Company, Rochester, NY). Serum protein fractions were electrophoretically resolved on Beckman SPE agarose membranes into 5 fractions: albumin, alpha-1, alpha-2, beta, and gamma globulins. These fractions were denatured, stained with the PARAGON blue dye, scanned, and quantified with an APPRAISE densitometer (Beckman Instruments Inc., Fullerton, CA). The serum samples were diluted with isotonic saline at 1:5, 1:10 and 1:20 ratios (v/v) before electrophoresis and the dilutions were analyzed on the same strip. The electrophoretic patterns were inspected for the presence or absence of the M- protein band abnormality. The electrophoretic protein fractions were scanned and quantified densitometrically. The dilution was considered to be most appropriate that yielded the highest M- protein concentration and an albumin concentration closest to the albumin result obtained by the bromocresol green method. The reference values (g/l) in our laboratory are as follows: total serum protein = 60-82, albumin = 35-47, alpha-1 globulin = 2-3, alpha-2 globulin = 4-7, beta globulin = 5-10, and gamma globulin = M-protein bands usually appear on the agarose strip as an additional discrete and homogeneous band located in an area ranging from the alpha-1 globulin through the gamma globulin regions. Characterization of the M-proteins. Serum M- protein bands detected by SPE were typed by immunofixation (IFE) using the PARAGON system (Beckman Instruments Inc.) for at least one sample from each patient. The protein fractions separated by agarose gel electrophoresis were immunoprecipitated in the gel with anti-igg, anti- IgA, anti-igm, anti-bound kappa light chain, and anti-bound lambda light chain antisera.

3 162 Annals of Clinical & Laboratory Science Determination of interassay analytical precision. Five serum samples (2 with IgM kappa, 2 with IgA lambda, 1 with IgA kappa M-proteins) were selected for an interassay analytical precision study. Each sample was analyzed for M-protein concentration by SPE at its proper dilution once daily for 10 days. As shown in Table 1, the SDs were 0.5, 1.0, 1.3, 1.0, and 0.7 g/l, respectively, for the 5 samples with mean M-protein concentrations of 36.9, 42.9, 52.6, 54.8, and 66.2 g/l. The corresponding CVs for these samples were 1.38, 2.29, 2.46, 1.91, and 1.01%. The increase of M-protein concentration measured at a dilution >1:5 was considered significant if the percentage increase was 3 CVs [9]. To evaluate the analytical significance of an increase of the M-protein values by SPE after further serum dilution, the 1.38% CV of the sample with a mean M-protein concentration of 36.9 g/l was used for the sample group with IgM M-protein concentrations 45 g/l at 1:5 dilution; the 2.46% CV of the sample with a mean M-protein concentration of 52.6 g/l was used for the sample group with IgM M-protein 45 g/l at 1:5 dilution; the 2.29% CV of the sample with a mean M-protein concentration of 42.9 g/l was used for the sample group with IgA M-protein values 49 g/l at 1:5 dilution; the 1.91% CV of the sample with a mean M-protein concentration of 54.8 g/l was used for the sample group with IgA M-protein values between 49 and 60 g/l at 1:5 dilution; the 1.01% CV of the sample with a mean M-protein concentration of 66.2 g/l was used for the sample group with IgA M-protein concentrations 60 g/l at 1:5 dilution. Results The M-protein peak locations on the SPE strip for each light chain subtype of IgA or IgM monoclonal gammopathy in this study are shown in Table 2, along with the respective ranges of M-protein concentrations. The SPE from approximately 50% the IgA kappa or lambda M-protein subgroups contained peaks in the beta-globulin region. The M-protein peaks from most (80-90%) patients with the IgM kappa or lambda subtype were located in the beta-globulin, the fast gamma-globulin, or the mid-gamma-globulin regions. For purposes of comparison, the locations of IgG M-protein peaks from other patients tested in this laboratory during the period of this study were also evaluated (data not included in Table 2). In 81 patients, the IgG kappa M-protein peaks were all located between the beta and slow gamma-globulin regions of the strip, with 52 (64%) involving the slow gamma-globulin region. In 39 patients, the IgG lambda M-protein peaks were all present between the beta-globulin region and slow gamma-globulin regions, with 19 (49%) involving the slow gamma region. The number of M-protein peaks on the SPE strip was also evaluated in each patient. Seven (41%) of 17 patients with IgA kappa monoclonal gammopathy showed more than one M-protein peak of the same M-protein type on SPE. Similarly, 6 (50%) of 12 patients with the IgA lambda subtype showed more than one peak. On the contrary, all cases of IgM monoclonal gammopathy showed only 1 IgM peak, while only 2 (3%) of 81 patients with IgG kappa monoclonal gammopathy and none (0%) of 39 patients with the IgG lambda subtype showed multiple M-protein peaks of the same type on the SPE strip. Concentrations of serum albumin and M- protein assayed by SPE in patients with IgA and IgM monoclonal gammopathies, as well as in each individual category at various pre-electrophoretic dilutions, are listed in Table 3. The mean M-protein concentrations of the samples diluted 1:10 or 1:20 did not show any increase, compared to that of the recommended 1:5 diluted sera. The mean albumin concentrations did not show significant variations among the groups at different dilutions of samples. When the changes of M-protein concentration assayed by SPE of each sample were calculated between that measured at the 1:5 dilution and that at the 1:10 or 1:20 dilution, 38 of the154 samples showed no change, 84 showed small decreases, and 32 showed small increases. Altogether, 146 (94.8%) of the154 samples showed no analytically significant increase of M-protein concentration by SPE when the sera were diluted 1:10 or 1:20, compared to the 1:5 dilution. Only 8 samples showed a percentage increase 3 SD. The mean M-protein and albumin concentrations of these 8 samples measured at various

4 Quantitation of serum IgA and IgM M-proteins 163 Table 1. Inter-assay precision for quantitation of M-proteins by agarose gel electrophoresis, based on analyses of 5 serum samples in 10 successive assays. Parameters Serum #1 Serum #2 Serum #3 Serum #4 Serum #5 M-protein type IgM kappa IgA lambda IgM kappa IgA lambda IgA kappa Concentration (mean ±SD, g/l) 36.9 ± ± ± ± ± 0.7 Coefficient of variation (CV, %) Table 2. Locations of M-protein peaks and their concentration ranges in IgA and IgM monoclonal gammopathies. Parameters IgA kappa IgA lambda IgM kappa IgM lambda M-Protein conc at 1:5 dilution (g/l), median (range) 51 (32-85) 54 (21-90) 48 (28-67) 43 (24-67) Locations of M-protein peaks (number of patients) Beta Beta/gamma junction Fast gamma Mid-gamma Slow gamma Others 1* 2 2** 0 All regions * Fast and mid-gamma; 1 case at the alpha-2/beta junction and 1 case with 4 peaks (fast, mid-, and slow gamma regions); ** Both cases at the fast/mid-gamma junction. Table 3. Concentrations of M-proteins and albumin in the entire group of 154 serum samples from 52 patients with IgA or IgM gammopathies, based upon (a) agarose gel electrophoretic assays at pre-electrophoretic dilutions of 1:5, 1:10, and 1:20 (v/v) and (b) albumin assays by the Kodak BCG dye-binding assay. Monoclonal gammopathy M-protein concentration (g/l, mean ±SD) Albumin concentration (g/l, mean ± SD) 1:5 1:10 1:20 1:5 1:10 1:20 Kodak dilution dilution dilution dilution dilution dilution assay IgA (82 samples from 29 patients) 53.5± ± ± ± ± ± ±3.5 IgM (72 samples from 23 patients) 45.4± ± ± ± ± ± ±3.5 Both types (154 samples, 52 patients) 49.7± ± ± ± ± ± ±3.7 Table 4. Concentrations of M-proteins and albumin in a subgroup of 8 serum samples from 3 patients with IgA or IgM gammopathies, based upon (a) agarose gel electrophoretic assays at pre-electrophoretic dilutions of 1:5, 1:10, and 1:20 (v/v) and (b) albumin assays by the Kodak BCG dye-binding assay. Monoclonal gammopathy M-protein concentration (g/l, mean ±SD) Albumin concentration (g/l, mean ± SD) 1:5 1:10 1:20 1:5 1:10 1:20 Kodak dilution dilution dilution dilution dilution dilution assay Either IgA or IgM (8 samples from 3 patients) 52.6± ± ± ± ± ± ±3.7

5 164 Annals of Clinical & Laboratory Science Table 5. Concentrations of M-protein in 11 successive serum samples from a single patient with IgM kappa monoclonal gammopathy, based on agarose gel electrophoretic assays after pre-electrophoretic dilutions of 1:5, 1:10, and 1:20 (v/v). Dates of serum collection 2/2 2/13 2/21 3/2 3/7 4/11 5/3 6/10 6/21 7/27 8/15 Serum total protein concentration (g/l) Serum M-protein (g/l) at 1:5 dilution :10 dilution :20 dilution Maximal increase (%) at a higher dilution Increase as a multiple of the analytical CV dilutions are shown in Table 4. Immunophenotypically, the 8 samples were from 3 patients, 2 with an IgM kappa and 1 with an IgA lambda monoclonal gammopathy. Underestimation of M-protein concentrations and overestimation of albumin concentrations would occur in this small subgroup if the properly diluted sera were not tested by SPE. During the study, 11 serum samples were analyzed from 1 patient with IgM kappa monoclonal gammopathy. Table 5 shows the M-protein values by SPE at various pre-electrophoretic dilutions of the 11 samples, with the respective concentrations of serum total protein. Five of the samples (obtained on 2/13, 2/21, 3/2, 3/7 and 4/11) showed significant increase of the M-protein value ( 3 SD) at higher serum dilutions. The 5 samples all had total protein concentration >100 g/l; only 2 of the samples had total protein concentration <110 g/l. Six of the 11 samples from this patient did not show analytically significant increases at higher dilutions. The serum total protein concentrations of 5 of the 6 samples were < 110 g/l. The overall pattern of data in Table 5 suggested that the higher the serum total protein concentration, the higher the dilution factor that was required for accurate M-protein quantitation in this subset of samples. This is consistent with our previous report for IgG monoclonal gammopathy [7]. Underestimation of M-protein concentrations by SPE would occur in this small subgroup of samples if the properly diluted sera were not tested. In the 8 samples that required higher preelectrophoretic serum dilution for accurate quantitation of M-protein concentration, the M- protein bands were unusually narrow, with thin and dense morphology compared to the corresponding albumin bands. In contrast, samples that did not show any quantitative change at higher dilutions and samples that showed an analytically insignificant increase did not have unusually narrow, thin and dense M-protein bands on the SPE membrane. Discussion Serum M-protein and albumin concentrations are clinically useful in the diagnosis and evaluation of monoclonal gammopathies. These data are also clinically significant prognostic factors in patients with multiple myeloma [10]. Serial measurements of serum M-protein concentration in patients with monoclonal gammopathies are useful in monitoring the treatment response and disease progression [2,3]. SPE provides a simple, reliable method for assaying both the M-protein and albumin concentrations when analytical precautions are taken. In SPE, when any protein fraction exceeds a limiting concentration, the binding of dye is nonlinearly related to the concentration [11]. In addition, because of the narrow aperture of the densitometer, all protein bands must have an expected rectangular shape, similar length, and homogeneous staining for their accurate proportional estimation. Proper dilution

6 Quantitation of serum IgA and IgM M-proteins 165 of serum before SPE compensates for the distortion of the M-protein band, if present, and diminishes the degree of nonlinear dye binding when the density of the band is very high. The molecular weights of normal serum IgA and IgM molecules are larger than that of IgG [12]. A small portion of normal serum IgA exists as dimers and IgA M-proteins commonly exist in polymeric forms [13]. Serum IgM molecules are normally pentameric. A small fraction of IgM molecules may be monomeric; this is more frequent in some diseases, eg, Waldenström s macroglobulinemia [14]. In SPE separation of normal serum proteins by the Beckman PARAGON protocol, the gamma globulins and some beta globulins migrate cathodally from the application point because of electroendosmosis [15]. In the agarose matrix, smaller molecules of equal charge density migrate faster and further. The serum IgA and IgM molecules normally migrate in the fast gamma globulin and the mid-gamma globulin regions while a substantial portion of the IgG molecules migrate in the slow gamma region. Monoclonal proteins of different immunoglobulin classes show a similar distribution [16,] which was also observed in this study (Table 2). IgA M-proteins exist as polymers in almost half of the cases, based on the presence of multiple M-protein peaks of the same M-protein type. IgA M-proteins are also able to form complexes with other serum proteins [16] and often migrate in the beta globulin region, as shown in this study (Table 2). The shape of protein bands on the electrophoretic strip can be affected by electroendosmosis. [17]. In this laboratory, it is not uncommon to see a subtly curved band of IgG M-protein migrating in the slow gamma region. When the protein concentration is very high, the condensed IgG M- protein band may also appear with slightly bulging ends and be longer than the corresponding albumin band. On the other hand, IgA and IgM M-protein peaks migrating in the beta globulin or the fast gamma to the mid-gamma regions often show a rectangular shape with about the same length as the corresponding albumin bands. Because of the formation of polymers with larger molecular weight, IgA and IgM M-protein bands often appear broader than an IgG M-protein monomer at similar concentration, and much broader than the albumin band, in samples that do not show significant change at higher serum dilutions. At higher serum dilutions, IgA and IgM M- protein bands tend to be much thinner than IgG M-protein bands. This relatively narrow IgA and the IgM M-protein bands may partly contribute to the minor decreases of IgA or IgM M-protein values when measured by SPE at serum dilutions higher than 1:5 that was noted with certain samples in this study. Physicochemical and electrophoretic properties explain why a higher pre-electrophoretic serum dilution is commonly required for IgG M- proteins but not for IgA or IgM M-proteins. The higher sample dilution may compensate for the curving and dense character of IgG M-protein bands; this usually does not apply to IgA or IgM M-protein bands. Exceptional samples in this study (especially the subset of the IgM kappa type that showed a significant increase of the M-protein values 3 SD at higher sample dilution) showed bands located in the beta globulin region, with an M-protein band denser and narrower than the corresponding albumin band. The unusually narrow, thin and dense appearance of the IgM M-protein band was observed in 2 patients with the kappa light chain subtype. Theoretically, some tumor clones might produce an M-protein of one class (IgM) and banding characteristics of another class (IgG) because of altered genetic integrity for the truncated protein product. It is also possible that these M- proteins exist as IgM monomers, and hence, achieve more condensed bands on the strip because of their smaller size. One IgA lambda sample that showed a slightly but significantly higher M-protein concentration at higher sample dilution also showed single M-protein peak in the beta globulin region. Although the band was not extremely narrow, it was somewhat dense because of very high M-protein and serum total protein concentrations. Based on these considerations, we conclude that a 1:5 pre-electrophoretic serum dilution is adequate for SPE quantitation of IgA and IgM M-proteins using the Beckman PARAGON system, except for rare samples with very high serum total protein

7 166 Annals of Clinical & Laboratory Science concentrations and/or unusually dark and narrow M-protein bands, which can be roughly evaluated by visual comparisons with the corresponding albumin band on the same SPE membrane. In such samples, 1:10 or 1:20 dilutions should be assayed. References 1. Malpas JS. Clinical presentation and diagnosis. In: Myeloma, Biology and Management (Malpas JS, Bergsagel DE, Kyle RA, Eds), New York Oxford University Press, 1995, pp Salmon SE. Immunoglobulin synthesis and tumor kinetics of multiple myeloma. Semin Hematol 1973; 10: Austin GE, Check IJ, Hunter RL. Analysis of the reliability of serial paraprotein determinations in patients with plasma cell dyscrasias. Am J Clin Pathol 1983;79: Daniels JC, Vyvial TM, Levin WC, Ritzmann SE. Methodologic differences in values for M-proteins in serum, as measured by three techniques. Clin Chem 1975; 21: Riches PG, Hobbs JR. Laboratory investigation of paraproteinemia. Am J Clin Pathol. 1988;41: Liu FJ, Fritsche HA, Trujillo JM. Underestimation of monoclonal proteins by serum protein electrophoresis on cellulose acetate. Clin Chem. 1987;33: Chang CY, Fritsche HA, Glassman AB, McClure KC, Liu FJ. Underestimation of monoclonal proteins by agarose serum protein electrophoresis. Ann Clin Lab Sci. 1997;27: Rodkey FL. Direct spectrophotometric determination of albumin in human serum. Clin Chem. 1965;11: Reading MA. Probabilities for the standard Gaussian or normal distribution. In: Biostatistics in Clinical Medicine. 3rd ed (Ingelfinger JA, Mosteller F, Thibodeau LA, Ware JH, Eds), McGraw- Hill, New York, 1994, pp Boccadoro M, Pileri A. Prognostic factors in multiple myeloma. In: Myeloma, Biology and Management (Malpas JS, Bergsagel DE, Kyle RA, Eds), Oxford Univ Press, New York, 1995, pp Slater L, Carter PM, Hobbs JR. Measurement of albumin in sera of patients. Ann Clin Biochem 1975;12: Carayannopoulos L, Capra JD. Immunoglobulins: structure and function. In: Fundamental Immunology, 3rd ed (Paul WE, Ed), Raven Press, New York, 1993, Kyle RA, Bayrd ED, Immunoglobulin groups: structure, function, and synthesis. In: The Monoclonal Gammopathy, Multiple Myeloma and Related Plasma Cell Disorders (KyleRA, Bayrd ED, Eds), CC Thomas, Springfield, Il, 1976, pp Day ED. The second, tertiary and quaternary structures of assembled immunoglobulins. In: Advanced Immunochemistry, 2nd ed (Day ED, Eds), Wiley-Liss, New York,1990, pp McPherson RA. Specific proteins. In: Clinical Diagnosis and Management by Laboratory Methods, 19th ed. (Henry JB, Ed) WB Saunders, Philadelphia 1996, pp Waldenström J. Diagnosis and Treatment of Multiple Myeloma, Grune Stratton, New York, 1970, pp Kozulic B. Looking at bands from another side. Anal Biochem 1994;216: