Highly Sensitive, Automated Immunoassay for Immunoglobulin Free ...

49 downloads 2205 Views 217KB Size Report
free light chain (FLC) molecules that are produced by malignant clones of B cells. BJ proteins are used as cancer markers for identifying and monitoring patients ...
Clinical Chemistry 47:4 673– 680 (2001)

Enzymes and Protein Markers

Highly Sensitive, Automated Immunoassay for Immunoglobulin Free Light Chains in Serum and Urine Arthur R. Bradwell,1* Hugh D. Carr-Smith,2 Graham P. Mead,2 Lian X. Tang,2 Paul J. Showell,2 Mark T. Drayson,1 and Roger Drew2

Background: Bence Jones proteins or monoclonal immunoglobulin ␬ and ␭ free light chains (FLCs) are important markers for identifying and monitoring many patients with B-cell tumors. Automated immunoassays that measure FLCs in urine and serum have considerable clinical potential. Methods: Sheep antibodies, specific for FLCs, were prepared by immunization with pure ␬ and ␭ molecules and then adsorbed extensively against whole immunoglobulins. The antibodies were conjugated onto latex particles and used to assay ␬ and ␭ FLCs on the Beckman IMMAGETM protein analyzer. Results: The unconjugated antibodies showed minimal cross-reactivity with intact immunoglobulins or other proteins. With latex-conjugated antibodies, ␬ and ␭ FLCs could be measured in normal sera and most normal urine samples. Patients with multiple myeloma had increased concentrations of the relevant serum FLC, whereas both FLCs were increased in the sera of patients with systemic lupus erythematosus. Conclusions: We developed sensitive, automated immunoassays for ␬ and ␭ FLC measurements in serum and urine that should facilitate the assessment of patients with light chain abnormalities.

eloma) and for characterizing light chain amyloidosis. Urine is the usual test fluid and typically is concentrated and then analyzed by protein electrophoresis (PE) or immunofixation electrophoresis (IFE) (1 ). Such techniques are time-consuming and can be inaccurate. Furthermore, the amounts of FLCs entering the urine are strongly influenced by renal tubular function. An alternative approach is to measure FLC concentrations in serum. Studies have shown that when urine concentrations are increased, there is a corresponding increase in the serum of the same FLC (2 ) and the latter may be diagnostically more accurate if patients have renal failure (3 ). However, immunoassays must be highly specific because serum concentrations of FLCs are several orders of magnitude lower than those of the light chains bound to intact immunoglobulins. Early serum immunoassays involved separating FLCs from intact immunoglobulins before analysis (4 – 6 ), and although accurate, they were impractical for routine use. Subsequent assays have used antibodies directed against the “hidden” epitopes of FLC molecules that are located at the interface between the light and heavy chains of intact immunoglobulins. Previous studies used RIAs and enzyme immunoassays with polyclonal antisera against FLCs to analyze urine samples, but the specificity was inadequate for serum measurements (7, 8 ), and variations in light chain polymerization caused measurement errors (9, 10 ). The use of monoclonal antibodies appears to be an obvious approach to improving specificity, but such antibodies have been difficult to develop (11–14 ), and their use has been restricted to RIAs and enzyme immunoassays, which are more complicated than turbidimetric

© 2001 American Association for Clinical Chemistry

Bence Jones (BJ)3 proteins are homogeneous populations of ␬ or ␭ free light chain (FLC) molecules that are produced by malignant clones of B cells. BJ proteins are used as cancer markers for identifying and monitoring patients with B-cell lineage tumors, (e.g., multiple my-

1 Department of Immunology, The Medical School, Edgbaston, Birmingham B15 2TT, United Kingdom. 2 The Binding Site, PO Box 4073, Birmingham B29 6AT, United Kingdom. *Author for correspondence. E-mail [email protected]. Received March 31, 2000; accepted January 4, 2001.

3 Nonstandard abbreviations: BJ, Bence Jones; FLC, free light chain; PE, protein electrophoresis; IFE, immunofixation electrophoresis; RID, radial immunodiffusion assay; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SLE, systemic lupus erythematosus; and RF, rheumatoid factor.

673

674

Bradwell et al.: Immunoassay for Free Light Chains in Serum and Urine

techniques and therefore not ideal for routine immunochemistry laboratories. Attempts were made to develop turbidimetric (15 ) and latex-enhanced nephelometric assays (16 ) using polyclonal antibodies, but they were insufficiently sensitive to detect serum or urine FLC concentrations within normal reference values, and crossreactivity with intact immunoglobulins was unacceptable. In this study, we describe the development and assessment of sensitive, latex-enhanced, turbidimetric, FLC assays for use on the Beckman IMMAGETM that can accurately measure ␬ and ␭ FLCs in serum and urine. The detection limit is compared with existing clinical laboratory tests for FLCs, and their potential in the clinical laboratory is discussed.

Materials and Methods preparation and assessment of flc antibodies Sheep were immunized with ␬ or ␭ molecules that had been purified from urine samples containing BJ proteins. The resulting antisera were adsorbed against purified myeloma IgG, IgA, and IgM proteins and then affinity purified against mixtures of the respective FLCs immobilized onto CNBr-activated Sepharose 4B (Amersham Pharmacia Biotech). Specificity was evaluated by (a) immunoelectrophoresis against BJ proteins and pooled normal human serum; (b) hemagglutination assays using sheep red blood cells sensitized with individual FLCs, purified polyclonal IgG, monoclonal IgA, or monoclonal IgM (17 ); (c) Western blot analysis comparing the staining of whole serum and urine blots produced by the polyclonal FLC antisera with monoclonal antibodies recognizing free and bound ␬ (clone GD12; The Binding Site Ltd.) and ␭ molecules (clone 312H; Department of Immunology, University of Birmingham, Birmingham, United Kingdom). To assess antibody reactivity against FLC monomers and dimers, urine samples containing both forms of ␬ and ␭ BJ proteins were separated on Sephadex G-100 (Amersham Pharmacia) in Tris-buffered saline, pH 7.4. Monomer and dimer fractions were then tested against the FLC antisera and whole light chain antisera by radial immunodiffusion (RID). In addition, FLC monomers and dimers in the urine samples were separated by nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). These gels were blotted and probed with the polyclonal FLC antisera. FLC-specific antisera were digested with pepsin to produce F(ab)2 fractions, which were adsorbed onto 184-nm polystyrene latex particles (18 ). The latex-conjugated reagents were tested for specificity by nephelometry on the Beckman IMMAGE, and antisera requiring further adsorption were recycled through the adsorption and testing procedures until satisfactory.

flc reference material

Monomeric, polyclonal FLC ␬ and ␭ molecules were prepared by reduction and acetylation of polyclonal IgG

that was purified from a serum pool prepared by combining sera collected from 200 apparently healthy donors. Acetylated FLCs were purified on a Sephadex G-100 column. ␬ and ␭ molecules were then separated using protein L (Actigen Ltd), which specifically binds only ␬ chains (19 ), and further purified by affinity chromatography against specific polyclonal antibodies. FLC purity was assessed using silver-stained SDS-PAGE gels, dot blot, and hemagglutination inhibition assays. Protein concentrations were determined by the BCATM protein assay (Pierce). Because of the limited amount of primary reference material, secondary reference materials were required, which were prepared from pools of nine ␬ and seven ␭ BJ proteins. However, these monoclonal proteins were not considered ideal for use as a working calibrator. Therefore, a third reference material was prepared that comprised serum from a patient with systemic lupus erythematosus (SLE) that contained increased polyclonal FLCs. The pure FLC preparations were used to assign ␬ and ␭ values to the secondary reference preparations by RID, which were then used to assign FLC values to the SLE serum preparation by nephelometry on the Beckman IMMAGE. Each stage of the value transfer was completed at three dilutions and repeated three times. All protein preparations were frozen and stored at ⫺20 °C until required.

nephelometric assay of FLCs on the beckman immage Both ␬ and ␭ assays used the noncompetitive, nearinfrared, particle immunoassay protocol on the instrument. A six-point calibration curve with a third-order polynomial curve fit was used. The assay components were as follows: 15 ␮L of serum sample for ␬ and 21 ␮L for ␭, 60 ␮L of latex reagent, 5 ␮L of 120 g/L polyethylene glycol 6000 (final concentration, 2 g/L), and 195 ␮L of phosphate-buffered saline. The reaction time for each assay was 6 min with a serum dilution of 1:10. Susceptibility to interference was assessed by adding known concentrations of purified whole immunoglobulins, triglycerides, hemoglobin, and bilirubin (Fig. 1) to base serum samples containing concentrations of FLCs just above the upper limit of normal (20 – 40 mg/L). Analysis was based on a paired difference protocol as proposed by the NCCLS (20 ). Each sample and the controls were analyzed three times, and the mean difference and 95% confidence intervals were calculated (Fig. 1). For rheumatoid factor (RF), a serum sample containing 1100 kIU/L RF activity was assayed for FLCs, undiluted and at different dilutions, and the results were compared. Assay precision was assessed by repeat assay of serum samples containing three different concentrations of FLCs. Intraassay precision was determined by repeat assay of the samples 15 times using a single calibration curve, interassay precision was determined by measuring the samples once using each of 10 separate calibration curves, and total assay

675

Clinical Chemistry 47, No. 4, 2001

Fig. 1. Interference data. Change of free light chain quantification on addition of the following substances: IgG (8.7 g/L; ⫻), IgA␬ (1.069 g/L; F), IgA␭ (2.138 g/L; 췦), IgM␬ (444 mg/L; 䡺), IgM␭ (888 mg/L; ⫹), bilirubin (200 mg/L; ⽧), triglycerides (5.95 g/L; f), and hemoglobin (3 g/L; Œ). Data points are means; bars, 95% confidence intervals.

precision was determined by measuring the samples five times using three separate calibration curves. Assay linearity was determined by serial dilution of serum samples, containing known amounts of FLCs, into Tris-buffered saline.

concentrations were measured in mid-stream, earlymorning urine samples from 36 healthy men (mean age, 35.1 year; range, 18 –54 years) and 30 healthy women (mean age, 33.1 year; range, 18 –52 years). Sodium azide (1 g/L) was added to all samples, which were stored at ⫺20 °C until analysis (21 ).

limits of detection of pe and ife for FLCs The detection limits of nephelometry, PE, and IFE for measuring FLCs were compared. For serum, different concentrations of purified monoclonal ␬ and ␭ FLCs were added to a normal serum that was subjected to PE and IFE. For urine, the same ␬ and ␭ proteins were added to a normal urine that had been concentrated 100-fold using a MiniconTM (Amicon), and the mixture was subjected to PE and IFE.

patient studies To produce a preliminary assessment of the assays in a clinical setting, the following groups were studied: 21 patients with multiple myeloma, 6 with Waldenstrom macroglobulinemia, 6 with FLC myeloma, and 12 patients with SLE.

Results antiserum specificity

comparison of urine flc measurements by nephelometry and rid Twenty-four urine samples containing ␬ and 22 samples containing ␭ BJ proteins were assayed for FLCs by RID and nephelometry. The antisera used in the RID assays reacted against whole light chains, but large amounts of anti-IgG, -IgA, and -IgM heavy chain antisera were added to prevent spurious whole immunoglobulin precipitate rings. In practice, this produced assays that were specific for FLCs in urine samples.

reference values FLC concentrations were measured by nephelometry in the serum of 50 male (mean age, 43.2 years; range, 19 –71 years) and 50 female (mean age, 42.4 years; range, 17–59 years) blood donors. For urine reference values, FLC

The antibodies against FLCs reacted with BJ proteins only and showed no immune precipitation reactions against whole immunoglobulin molecules in pooled normal serum by immunoelectrophoresis. By hemagglutination, ␬ FLC antibodies reacted with ␬-labeled cells at a dilution ⬎1:16 000 and at ⱕ1:2 against polyclonal IgG, monoclonal IgA␬ and IgM␬, and monoclonal ␭ FLC-coated cells. ␭ antibodies reacted with ␭-labeled cells at a dilution ⬎1:16 000 and at ⱕ1:2 against polyclonal IgG, monoclonal IgA␭ and IgM␭, and monoclonal ␬ FLC-coated cells. In Western blots, both FLC antisera reacted strongly with two closely migrating bands at 25 000 –30 000 kDa and weakly with several larger and smaller molecular mass bands. Similar staining patterns were observed with the monoclonal antibodies (Fig. 2). The polyclonal antibodies detected both monomers and dimers of FLCs by Western

676

Bradwell et al.: Immunoassay for Free Light Chains in Serum and Urine

CV was 3.8%, 2.1%, and 2.9%, respectively; the interassay CV was 9.2%, 5.0%, and 3.5%, respectively; and the total assay CV was 8.2%, 5.2%, and 2.7%, respectively. For free ␭ at 15.8, 62.4 and 227 mg/L, the intraassay CV was 4.7%, 1.9%, and 1.9%, respectively; the interassay CV was 7.6%, 2.9%, and 4.0%, respectively; and the total assay CV was 9.5%, 5.2%, and 4.1%, respectively. Linearity studies showed that the free ␬ assay was linear at 3.6 –172 mg/L (p ⫽ 0.9996; y ⫽ 1.03x ⫺ 1.6 mg/L) and the free ␭ assay was linear at 5.6 –268 mg/L (p ⫽ 0.9995; y ⫽ 1.00x ⫺ 0.5 mg/L). Fig. 2. Western blots showing the specificity of the polyclonal FLC antisera compared with monoclonal antibodies, and the reaction of polyclonal antisera against FLC monomers and dimers, separated by nonreducing SDS-PAGE. Lane 1, molecular mass markers. Lanes 2 and 3, urine containing ␬ FLCs probed with monoclonal and polyclonal anti-␬, respectively. Lanes 4 and 5, normal serum probed with monoclonal and polyclonal anti-␬, respectively. Lanes 6 and 7, urine containing ␭ FLC probed with monoclonal and polyclonal anti-␭, respectively. Lanes 8 and 9, normal serum probed with monoclonal and polyclonal anti-␭, respectively. Lanes 10 and 11, polyclonal FLC antisera reacted with monomers and dimers of ␬ and ␭, respectively.

blot (Fig. 2). By RID, both ␬ and ␭ FLC antisera produced precipitation rings against the monomers and dimers and in similar proportions.

flc reference materials

Each FLC preparation was found to be ⬎99% pure by silver-stained SDS-PAGE, and the other FLC was not detected by hemagglutination inhibition or dot blot. The data from the value transfers showed that the assays exhibited good linearity with all results being within 5.3% of the mean assigned values. The final FLC values for the reference preparations derived from the SLE serum were 46 mg/L for ␬ and 71.4 mg/L for ␭. These calibration values were used throughout the study.

assay conditions on the beckman immage When the assay conditions had been optimized (see Materials and Methods), the assay range was 3.6 –172 mg/L for serum free ␬ and 5.6 –268 mg/L for ␭. The detection limits for undiluted urine samples were 0.36 and 0.56 mg/L, respectively. The assays did not demonstrate antigen excess when tested up to 40 g/L for ␬ FLCs and to 60 g/L for ␭ FLCs. The interference studies for the ␬ FLC assay showed slight cross-reactivity with ␬ whole immunoglobulins, whereas all other substances had little effect on either assay (Fig. 1). There was modest interference by RF. The initial value for free ␬ in the undiluted RF sample (1100 kIU/L was 13.4 mg/L and changed to 13.7 and 12.5 mg/L at 1:2 and 1:4, respectively (corrected for dilution). The undiluted RF sample also contained 12.8 mg/L free ␭, which changed to 15.2 and 16.5 mg/L at 1:2 and 1:4, respectively (corrected for dilution). Assay imprecision, expressed as the CV, was as follows. For free ␬ at 7.3, 29.7, and 114 mg/L, the intraassay

comparison of flc detection by electrophoresis, ife, and rid A ␬ BJ protein migrating toward the anode (␤2 position) could be detected by PE at 2000 mg/L, and a ␭ BJ protein migrating toward the cathode (␥ mobility) could be detected at 500 mg/L. The differences in detection limits were attributable to masking of the monoclonal ␬ band by ␤ globulins (transferrin and complement component C3). The same BJ proteins added to urine concentrated 100fold showed similar detection limits, 20 and 5 mg/L, respectively, when compared with nonconcentrated urines. By IFE, the detection limits were 150 mg/L for ␬ and 100 mg/L for ␭, and were similar for serum and urine samples. Comparison of FLC measurement by RID and turbidimetry showed a high degree of correlation for both ␬ (p ⫽ 0.99; y ⫽ 1.15x ⫺ 15.2 mg/L) and ␭ assays (p ⫽ 0.98; y ⫽ 0.92x ⫺ 31.2 mg/L).

normal serum and urine results

The mean (⫾ SD) concentration of serum free ␬ was 8.4 ⫾ 2.66 mg/L (n ⫽ 100; range, 3.6 –15.9 mg/L; 95% confidence interval, 4.2–13.1 mg/L), for free ␭, the mean (⫾ SD) concentration was 14.5 ⫾ 4.4 mg/L (n ⫽ 100; range, 8.1–33 mg/L; 95% confidence interval, 9.2–22.7 mg/L). The mean serum ␬:␭ ratio was 1:1.67 (95% confidence interval, 1:2.78 –1:0.99; Fig. 3). For urine, the mean (⫾ SD) free ␬ concentration was 5.4 ⫾ 4.95 mg/L (n ⫽ 66; range, 0.36 –20.3 mg/L; 95% confidence interval, 0.39 –15.1 mg/L), and the mean (⫾ SD) free ␭ concentration was 3.17 ⫾ 3.3 mg/L (n ⫽ 66; range, 0.81–17.3 mg/L; 95% confidence interval, 0.81–10.1 mg/L). The mean ␬:␭ ratio was 1:0.54 (95% confidence interval, 1:2.17–1:0.25). There were no differences between serum or urine free ␬ or ␭ results on the basis of age or sex. The mean, normal urine FLC excretion was 3.7 mg/g of creatinine for ␬ and 2.0 mg/g of creatinine for ␭. There was a positive but nonsignificant correlation of urine creatinine concentrations with ␬ (r ⫽ 0.22) and ␭ (r ⫽ 0.17) measurements.

patient samples Sera from patients with multiple myeloma or Waldenstrom macroglobulinemia who had intact monoclonal immunoglobulins contained increased concentrations of the relevant FLCs (Fig. 4). In all but one sample, the

677

Clinical Chemistry 47, No. 4, 2001

Fig. 3. Frequency distributions of free ␬ (A) and ␭ (B) and ␬:␭ ratios (C) in 100 normal sera and free ␬ (D) and ␭ (E) and ␬:␭ ratios (F) in 66 normal urines.

nonclonal light chains had normal or reduced concentrations. All of the sera from patients with BJ myeloma had increased serum concentrations of the relevant FLC and normal concentrations of the other FLC (Table 1). Two of these patients were in clinical remission with no FLCs detected in the urine by RID assay (⬍40 mg/L), but both had increased serum FLC concentrations with abnormal ␬:␭ ratios. Sera from patients with SLE showed increases of both FLC, and ␬:␭ ratios were within the reference interval (Fig. 4).

Discussion Critical aspects of the FLC assays include minimal crossreactivity with whole immunoglobulins and no reactivity

with the alternative light chain. The observation that suppression of the nonmalignant light chains can be detected in myeloma samples supports good specificity, although the ␬ FLC assay could be improved (Fig. 1). Such an improvement might produce a slight reduction of the ␬ reference interval. The concentrations of FLCs in normal sera were comparable with previously published values (Table 2), although ␭ values were higher than ␬. Because there are approximately twice as many ␬- as ␭-producing lymphoid cells, this result may seem rather surprising. However, a similar observation was made in a recent study (14 ), and it was suggested that in healthy individuals, ␬ FLC production might be less than ␭. An alternative explanation is that because ␬ molecules usually are monomeric (25 kDa), renal clearance is faster than for the dimeric ␭ molecules (50 kDa). In a study on the movement of dextran polymers across capillary membranes, it was shown that molecules of 20 kDa were cleared 3.2 times faster than 37-kDa molecules (22 ). Our results show a ␬:␭ clearance rate of 3 ⫻ [(␬ urine/␭ urine)/(␬ serum/␭

Table 1. Comparison of FLC concentrations in serum and urine of six patients with BJ myeloma. Myeloma type

Fig. 4. Comparison of free light chain concentrations (mean and 95% confidence interval) in 100 normal sera, sera from 27 patients with monoclonal gammopathies (IgG, F; IgA, Œ; and IgM, 䡺), and sera from 12 patients with SLE (E).

Kappa Kappa Kappa Lambda Lambda Lambda

1 2 3 1 2 3

Serum free ␬, mg/L

Serum free ␭, mg/L

Serum free ␬:␭ ratio

Urinary FLCs by RID, mg/L

163 868 259 10 9.2 21

6.09 7.6 23.4 43 86.6 17 161

23.6 114 11.1 0.233 0.106 0.001

500 4500 530 ⬍40 ⬍40 18 700

678

Bradwell et al.: Immunoassay for Free Light Chains in Serum and Urine

Table 2. Publications reporting normal serum and urine FLC concentrations (mg/L unless stated otherwise). Serum Publication

Date

Waldmann et al. (33 ) So¨lling (4 ) Hemmingsen and Skaruup (34 ) Robinson et al. (7 ) Brouwer et al. (8 ) Axiak et al. (12 ) Wakasugi et al. (35 ) Nelson et al. (13 ) Wakasugi et al. (16 ) Abe et al. (14 ) This study

1972 1975 1977 1982 1985 1987 1991 1992 1995 1998 2001



Urine



13.2 (⫾3.8)a,c

10.6 (⫾3.1)a,c

16.2 (9%)d,f 1.2 (0.8–7.5)e,g 7.1 (⫾4.3)c,d 10 (1.6–15.2)e,g 20.6 (⫾6)c,d,h 16.6 (⫾6.1)c,g c,d,h 8.4 (⫾2.66)

41.4 (10%)d,f 5.0 (⫾2.7)c,d e–g 3 (0.4–4.2) 16.2 (⫾8.6)c,d,h c,g 33.8 (⫾4.8) 14.5 (⫾4.4)c,d,h



3.2 (⫾1.2)a–c 1.2 (0.0–6.8)b,d,e 3.1 (1.9–7.1)b,d,e 1.8 (0.2–7.5)d 4.9 (3.0–8.0)e,g

2.96 (⫾1.84)c,g 5.5 (⫾4.95)c,d,h,i



2.3 (⫾1.1)a–c 1.1 (⫾0.4)a)–c 0.8 (0.0–2.4)b,d,e 1.45 (0.7–3.4)b,d,e 0.75 (0.15–2.1)d,e

1.07 (⫾0.69)c,g 3.17 (⫾3.3)c,d,h,i

a

Polyclonal antibody against total light chains. Urine FLCs in mg/24 h. c Mean (SD). d Polyclonal antibody against FLCs. e Mean (range). f Mean (CV). g Monoclonal antibody against FLCs. h Latex reagent. i Early-morning urine. b

serum)]. Although polysaccharides are different in charge, shape, and flexibility from globular proteins of similar molecular weight, differential filtration could account for our results. This may not have been observed in some of the early FLC studies because of the poor specificity of the antisera (see below). In urine, the concentrations of both FLCs were slightly higher than noted previously, whereas ␬:␭ ratios were similar to other studies. The higher FLC concentrations can be explained by our use of early-morning urine samples, which are typically two- to threefold more concentrated than 24-h urine samples. Table 2 shows that assays using monoclonal antibodies have produced rather varied results. One report showed rather low (12 ) and another high (14 ) FLC concentrations in normal sera. It is unclear whether the variation is attributable to specificity, calibration, or matrix differences. However, comparison must await internationally accepted reference materials. It is of note that the mean concentrations of the nonclonal FLCs in myeloma sera were lower than previous reports, suggesting improved specificity of the antisera (7, 13 ). One sample, containing an IgM monoclonal protein, had increased concentrations of the nonclonal FLCs. This result can be explained by a matrix effect, but a degree of cross-reactivity with whole immunoglobulins cannot be excluded. Polymerization of FLC molecules could lead to errors in quantification. Whereas So¨lling et al. (23 ) indicated that both monomers and dimers of FLCs were detected equally using antibodies against whole light chains, another study (10 ) showed that FLC antibodies failed to detect ␬ monomers satisfactorily, possibly because of

low-affinity antisera. Our preliminary experiments showed similar detection of both monomers and dimers, but further studies are required. It might be difficult to develop FLC assays that measure all forms of the molecules equally, so perfect quantification could remain elusive. The turbidimetric assays were at least 500 times more sensitive than serum or urine electrophoresis. Even with the use of high-resolution electrophoresis (24 ) and taking account of the focusing effect of monoclonal proteins on electrophoresis gels, the FLC immunoassays had detection limits ⬎50-fold lower than electrophoresis and perhaps 20-fold lower than IFE. These substantially lower detection limits should allow early identification of patients with abnormal FLC production.

clinical role of flc immunoassays Serum and urine FLC immunoassays should be useful for identifying and monitoring patients with BJ myeloma and nonsecretory myeloma (25 ). The assays might also find use in establishing light chain clonality and for monitoring patients with whole immunoglobulin-secreting multiple myeloma, light chain amyloidosis, and other diseases associated with excess monoclonal light chain production (26 ). FLC quantification may also be of interest in assessing patients with chronic B-cell activation. Polyclonal FLC concentrations are increased in autoimmune diseases such as SLE (27, 28 ) and insulin-dependent diabetes (29 ), and in chronic inflammatory diseases such as sarcoidosis and tuberculosis (30 ). They are also increased in the urine and cerebrospinal fluid patients with multiple sclerosis (31, 32 ). One potential drawback of quantitative FLC assays is

Clinical Chemistry 47, No. 4, 2001

that they cannot directly determine clonality, although altered ␬:␭ ratios are highly suggestive of clonality, particularly if extreme. On the other hand, ratios might be normal in early disease or during clinical remission, and increased polyclonal FLC concentrations will mask low concentrations of monoclonal FLCs. In addition, biclonal gammopathies of different FLC types could produce normal ␬:␭ ratios although the concentrations of both molecules might be increased. In all of these situations, clonality must be confirmed by electrophoresis. In spite of these limitations, the potential benefits of simple FLC immunoassays for assessing monoclonal gammopathies, in terms of improved sensitivity, accuracy, cost savings, and the use of serum as a test medium, are considerable. In conclusion, previously published immunoassays for FLCs have had little clinical impact either because the techniques were too complex or because the antibodies were nonspecific. The automated, turbidimetric immunoassays described here may allow serum and urine samples to be readily assessed for FLC concentrations in a routine, clinical laboratory setting.

This work was supported in part by a grant from the Department of Trade and Industry (Grant WMR/26799/ SP). We would like to acknowledge Phil Stubbs for assistance with the experiments and producing the electrophoresis gels, Beckman Coulter Inc. for allowing use of the Beckman IMMAGE, and Margaret Richards for technical assistance. Beckman IMMAGETM is a registered trademark of Beckman Coulter (Brea, CA).

References 1. Keren DF. Procedures for the evaluation of monoclonal immunoglobulins. Arch Pathol Lab Med 1999;123:126 –32. 2. So¨lling K. Lanng Nielsen J, So¨lling J, Ellegaard J. Free light chains of immunoglobulins in serum from patients with leukaemias and multiple myeloma. Scand J Haematol 1982;28:309 –18. 3. Sinclair D, Dagg JH, Smith JG, Stott DI. The incidence and possible relevance of Bence-Jones protein in the sera of patients with multiple myeloma. Br J Haematol 1986;62:689 –94. 4. So¨lling K. Free light chains of immunoglobulins in normal serum and urine determined by radioimmunoassay. Scand J Clin Lab Invest 1975;35:407–12. 5. So¨lling K. Normal values of free light chains in serum in different age groups. Scand J Clin Lab Invest 1977;37:21–5. 6. Cole PW, Durie BGM, Salmon SE. Immunoquantitation of free light chain immunoglobulins: applications in multiple myeloma. J Immunol Methods 1978;19:341–9. 7. Robinson EL, Gowland E, Ward ID, Scarffe JH. Radioimmunoassay of free light chains of immunoglobulins in urine. Clin Chem 1982;28:2254 – 8. 8. Brouwer J, Otting-van de Ruit M, Busking-van der Lely H. Estimation of free light chains of immunoglobulins by enzyme immunoassay. Clin Chim Acta 1985;150:257–74. 9. So¨lling K. Polymeric forms of free light chains in serum from normal individuals and from patients with renal diseases. Scand J Clin Lab Invest 1976;36:447–52.

679

10. Heino J, Rajamaki A, Irjala K. Turbidimetric measurement of Bence-Jones proteins using antibodies against free light chains of immunoglobulins. An artifact caused by different polymeric forms of light chains. Scand J Clin Lab Invest 1984;44:173– 6. 11. Ling NR, Lowe J, Hardie D, Evans S, Jefferis R. Detection of free ␬ chains in human serum and urine using pairs of monoclonal antibodies reacting with C␬ epitopes not available on whole immunoglobulins. Clin Exp Immunol 1983;52:234 – 40. 12. Axiak SM, Krishnamoorthy L, Guinan J, Raison RL. Quantitation of free k light chains in serum and urine using a monoclonal antibody based inhibition enzyme-linked immunoassay. J Immunol Methods 1987;99:141–7. 13. Nelson M, Brown RD, Gibson J, Joshua DE. Measurement of free ␬ and ␭ chains in serum and the significance of their ratio in patients with multiple myeloma. Br J Haematol 1992;81:223–30. 14. Abe M, Goto T, Kosaka M, Wolfenbarger D, Weiss DT, Solomon A. Differences in kappa and lambda (␬:␭) ratios of serum and urinary free light chains. Clin Exp Immunol 1998;111:457– 62. 15. Tillyer CR, Iqbal J, Raymond J, Gore M, McIlwain TJ. Immunoturbidimetric assay for estimating free light chains of immunoglobulins in urine and serum. J Clin Pathol 1991;44:466 –71. 16. Wakasugi K, Suzuki H, Imai A, Konishi S, Kishioka H. Immunoglobulin free light chain assay using latex agglutination. Int J Clin Lab Res 1995;25:211–5. 17. Ling, NR, Bishop S, Jefferis R. Use of antibody coated red cells for the sensitive detection of antigen and in rosette tests for cells bearing surface immunoglobulins J Immunol Methods 1977;15: 279 – 89. 18. Thakkar H, Davey CL, Medcalf EA, Skingle L, Craig AR, Newman DJ, Price CP. Stabilization of turbidimetric immunoassay by covalent coupling of antibody to latex particles. Clin Chem 1991;37: 1248 –51. 19. Nilson BHK, Logdberg L, Kastern W, Bjorck L, Akerstrom B. Purification of antibodies using protein L-binding framework structures in the light chain variable domain. J Immunol Methods 1993;164:33– 40. 20. National Committee for Clinical Laboratory Standards. Interference testing in clinical chemistry; proposed guideline. NCCLS Publication EP7-P. Villanova, PA: NCCLS, 1986. 21. Tencer J, Thysell H, Andersson K, Grubb A. Long-term stability of albumin, protein HC, immunoglobulin G, ␬ and ␭-immunoreactivity, orosomucoid and ␣1-antitrypsin in urine stored at ⫺20 °C. Scand J Urol Nephrol 1997;31:67–71. 22. Arfors K-E, Rutili G, Svensjo¨ E. Microvascular transport of macromolecules in normal and inflammatory conditions. Acta Physiol Scand Suppl 1979;463:93–103. 23. So¨lling K, So¨lling J, Lanng Nielsen J. Polymeric Bence Jones proteins in serum in myeloma patients with renal insufficiency. Acta Med Scand 1984;216:495–502. 24. Keren DF. Proteins in serum identified by high-resolution electrophoresis. In: High-resolution electrophoresis and immunofixation, 2nd ed. Boston: Butterworth-Heinemann, 1994:41– 89. 25. Drayson M, Tang LX, Drew R, Mead GP, Carr-Smith HD, Bradwell AR. Serum free light chain measurements for identifying and monitoring patients with nonsecretory myeloma. Blood 2001;in press. 26. Keren DF, Alexanian R, Goeken JA, Gorevic PD, Kyle RA, Tomar RH. Guidelines for clinical and laboratory evaluation of patients with monoclonal gammopathies. Arch Pathol Lab Med 1999;123: 106 –7. 27. Epstein WV, Tan M. Increase of L-chain proteins in the sera of patients with systemic lupus erythematosus and the synovial fluids of patients with peripheral rheumatoid arthritis. Arthritis Rheum 1966;9:713–9. 28. Hopper JE, Sequeira W, Martellotto J, Papagiannes E, Perna L,

680

29.

30.

31. 32.

Bradwell et al.: Immunoassay for Free Light Chains in Serum and Urine

Skosey JL. Clinical relapse in systemic lupus erythematosus: correlation with antecedent elevation of urinary free light-chain immunoglobulin. J Clin Immunol 1989;9:338 –50. Groop L, Ma¨kipernaa A, Stenman S, DeFronzo RA, Teppo A-M. Urinary excretion of ␬ light chains in patients with diabetes mellitus. Kidney Int 1990;37:1120 –5. So¨lling K, So¨lling J, Romer FK. Free light chains of immunoglobulins in serum from patients with rheumatoid arthritis, sarcoidosis, chronic infections and pulmonary cancer. Acta Med Scand 1981; 209:473–7. DeCarli C, Menegus MA, Rudick A. Free light chains in multiple sclerosis and infections of the CNS. Neurology 1987;37:1334 – 8. Mehta PD, Cook SD, Troiano RA, Coyle PK. Increased free light

chains in the urine from patients with multiple sclerosis. Neurology 1991;41:540 – 4. 33. Waldmann TA, Strober W, Mogielnicki RP. The renal handling of low molecular weight proteins. II. Disorders of serum protein catabolism in patients with tubular proteinuria, the nephrotic syndrome, or uremia. J Clin Invest 1972;51:2162–74. 34. Hemmingsen L, Skaarup P. Urinary excretion of ten plasma proteins in patients with febrile diseases. Acta Med Scand 1977;201:359 – 64. 35. Wakasugi K, Sasaki M, Suzuki M, Azuma N, Nobuto T. Increased concentrations of free light chain lambda in sera from chronic hemodialysis patients. Biomater Artif Cells Immobilization Biotechnol 1991;19:97–109.