Sensitive automated ELISA for measurement of vitamin D-binding protein (Gc) in human urine.
Gc is a 50- to 58-kDa multifunctional plasma protein synthesized mainly by hepatocytes and usually present in plasma in concentrations between 4 and 6 [micro]mol/L. The functions of Gc are diverse. Gc is an important player in the actin-scavenger system, which prevents the harmful consequences of actin in the blood stream during tissue injury. It binds actin monomers with high affinity, and Gc-actin complexes are rapidly cleared from the circulation (1). Gc also has functions in the immune system, acting as a co-chemotactic factor together with complement C5a (2) and, after deglycosylation, as a very potent macrophage-activating factor (Gc-MAF) also capable of activating osteoclasts (3,4). The name "vitamin D-binding protein" is derived from its ability to bind and transport vitamin D metabolites. Usually, <0.1% of 25-hydroxyvitamin D (25OHD) and <1% of 1,25-dihydroxyvitamin D [1,25[(OH).sub.2]D] circulate in their free forms (5). Gc is probably also important for the renal activation of 25OHD to 1,25[(OH).sub.2]D. Mice lacking the multifunctional receptor megalin lose Gc-25OHD complexes in the urine and are deficient in 1,25[(OH).sub.2]D (6). Thus, there is evidence that, generally, Gc and Gc-25OHD complexes are filtered in the glomerulus and reabsorbed by megalin-mediated endocytosis into the proximal tubular cells, where vitamin D activation takes place. This theory postulates that, under healthy conditions, only trace amounts of Gc should be excreted in the urine, whereas urinary loss of Gc is expected to increase with decreases in the capacity for reabsorption in the proximal tubules. To evaluate the fate of Gc in humans with various kidney diseases, there is a need for an assay sensitive enough to measure the minute amounts of Gc excreted in urine. This prompted us to develop a sensitive ELISA.
We used the principles described by Engbaek (7) to optimize the ELISA for Gc. The assay is based on an immobilized polyclonal antibody that captures the Gc, which subsequently binds a biotinylated monoclonal detection antibody that reacts with peroxidase-avidin-tetramethylbenzidine, producing a color reaction.
As capture antibody, we added to each well 8 [micro]g of rabbit anti-human Gc globulin (DakoCytomation Denmark A/S) in 100 [micro]L of 50 mmol/L sodium carbonate (pH 9.6). We incubated the ELISA plates (F96-Maxisorp Nunc immunoplates; Nunc A/S) at 4[degrees]C for 20 h before emptying the wells and adding 200 [micro]L of 1 mol/L ethanolamine (pH 8-9). After another 20 h at 4[degrees]C, the plates were stored at -20[degrees]C. We biotinylated the monoclonal mouse anti-human Gc globulin (AntibodyShop; SSI) after overnight dialysis of 200 [micro]L against 0.1 mol/L sodium bicarbonate (pH 8.3) by mixing gently for 4 h in the dark at room temperature with 10 [micro]L of 4.4 mmol/L biotin-amidocaproate-N-hydroxysuccinimide ester (Sigma). Subsequently, we added 10 [micro]L of 100 mmol/L lysine monohydrochloride (Fluka), waited for 15 min, and added 10 [micro]L of rabbit [gamma]-globulin (50 g/L; Calbiochem) and bovine [gamma]-globulin (100 g/L; Sigma) in 10 mmol/L sodium phosphate (pH 7.4). We then dialyzed the mixture for 48 h against 10 mmol/L sodium phosphate (pH 7.4) and for 24 h after addition of 1 g/L sodium azide (Merck). The biotinylated antibody was stored at -20[degrees]C and before use was diluted 1:10 000 in the assay buffer, 100 mmol/L sodium phosphate (pH 8.0) containing 1 g/L bovine albumin (Sigma). We used an 8-point calibration curve with purified human Gc-globulin, mixed type (Calbiochem), diluted in assay buffer (0, 1.25, 2.5, 5, 10, 20, 40, and 80 pmol/L) and obtained the calibration curve by cubic regression.
[FIGURE 1 OMITTED]
For the ELISA, we used an automated analyzer (BEP2000; Dade Behring) at 37[degrees]C. Between each step, we incubated the plates for 30 min and washed the wells three times with a washing buffer consisting of 10 mmol/L sodium phosphate (pH 7.4), 145 mmol/L sodium chloride, and 1 g/L Tween 20 (Merck). On the BEP-2000, controls and samples were diluted 1:10 in assay buffer, and 100-[micro]L duplicates of the calibrators and diluted controls and samples were added. The biotinylated detection antibody (100 [micro]L) was then added, followed by 100 [micro]L of peroxidase-avidin (DakoCytomation Denmark A/S) diluted 1:2000 in 10 mmol/L sodium phosphate (pH 7.4) containing 400 mmol/L sodium chloride and 0.2 g/L lysozyme (Sigma). The color reaction was started by addition of 100 [micro]L of TMB-One ready-to-use substrate (Kem-En-Tec Diagnostics A/S) and stopped after 9 min by addition of 100 [micro]L of 1 mol/L phosphoric acid to each well. Color development was measured photometrically at 620 run.
We checked for linearity by measuring 5 dilutions of each of 10 urine samples. The linear regression lines between measured and expected Gc concentrations all had intercepts not differing from 0, and 7 of 10 slopes did not differ from 1, whereas 3 slopes differed slightly (95% confidence intervals, 0.94-0.99, 1.02-1.36, and 1.01-1.10). Imprecision was estimated at 4 concentrations by 4 measurements on each plate and 2 plates a day for 6 days. At urinary Gc concentrations of 14, 84, 198, and 521 pmol/L, the within-plate, within-day, between-day, and total imprecisions calculated as recommended by Krouwer and Rabinowitz (8) were 1.8%-3.6%, 2.0%-4.1%, 1.4%-4.2%, and 2.6%-5.8%, respectively. Recovery was 95%-108% for 221 and 426 pmol/L added to five different urines and each measured four times. The detection limit, defined as the concentration corresponding to a signal 3 SD above the mean for the calibrator free of Gc, was 2.5 pmol/L, but in practice we used 10 pmol/L. The effect of storage for 3 and 8 months at -20 and -80[degrees]C is shown in Fig. 1. Contrary to our experiences with Gc in plasma (unpublished data), urine samples for Gc measurement cannot be stored at -20[degrees]C. However, when 2 vials of each of 15 urine samples were stored, one at -80[degrees]C for 1.5 months followed by 3 months at -20[degrees]C, the other at -80[degrees]C for 4.5 months, the Gc concentration did not differ. This indicates that the initial freezing temperature is crucial. We recommend either measurement of Gc in fresh urine or storage of urine samples at -80[degrees]C until measurement.
To examine the urinary loss of Gc as a function of kidney function, we enrolled 99 patients with various kidney diseases [dialysis (n = 0), glomerulonephritis (35), diabetic nephropathy (17), polycystic kidney disease (9), and other or nonspecified kidney disease (38)] and 12 healthy individuals (of those, 2 were excluded because of urinary albumin >30 mg/24-h) in a cross-sectional study approved by the local ethics committee (No. 20020055). Plasma and urine samples (24-h urine and second-void morning urine) were stored immediately at -80[degrees]C and, without thawing, moved to a -20[degrees]C freezer before measurement. We measured plasma Gc by our immunonephelometric method on a Behring Nephelometer 2 (Dade Behring) (9) and albumin and creatinine on an Integra 700 (Roche). We used Kruskal-Wallis and Mann-Whitney tests for comparisons between groups and the Spearman nonparametric method to look for correlations. We performed the calculations with SPSS 10.0.5 and set 0.05 as the significance level.
In the 99 patients, the median 24-h excretions of Gc and albumin did not differ significantly between patients in the various kidney disease groups but were significantly (P <0.001) higher in patients than in healthy controls (Table 1). Measurement of Gc in second-void morning urine gave a good estimate of the 24-h urinary excretion, equivalent to that given by the Gc/creatinine ratio, as seen from the high correlations in Table 1. Urinary Gc did not correlate with plasma Gc, but correlated with markers of kidney disease, particularly with urinary albumin excretion (Table 1). Linear regression with Gc as the dependent and albumin as the independent variable (after In transformation to obtain approximate gaussian distributions of residuals) showed a highly significant relationship in urine [r = 0.86; P <0.001; mean (95% confidence interval) slope, 1.2 (1.04-1.3); mean (95% confidence interval) intercept, -4.6 (-5.5 to -3.7)], but not in plasma (P = 0.23). The median urinary Gc/albumin ratio was significantly lower in patients with glomerulonephritis than in patients with polycystic, other, or nonspecified kidney diseases. The correlation between creatinine clearance and the Gc /albumin ratio in 24-h urine was high (r = -0.54; P <0.001) as was the correlation in second-void morning urine (r = -0.59; P <0.001). Longitudinal studies, however, are needed to test the utility of the urinary Gc/albumin ratio as a marker of kidney function.
In the kidneys, both Gc and albumin are believed to be filtered in the glomerulus and subsequently reabsorbed by megalin-cubilin-mediated endocytosis in the proximal tubule (10). Because of this shared fate, the high correlation between urinary Gc and urinary albumin was expected. However, contrary to the negative correlation between plasma and urinary albumin, we found no significant correlation between plasma and urinary Gc. The missing decrease in plasma Gc concentration is in agreement with findings in megalin-knockout mice (6) and in some studies of patients with kidney diseases (11, 12), but not in others (13, 14).
In conclusion, we have developed a sensitive automated ELISA capable of measuring very low Gc concentrations with low imprecision (functional detection limit, 0.01 nmol/L). In 99 patients, the urinary loss of Gc increased with increasing severity of kidney disease, but had no relationship with plasma Gc concentration.
We thank Nyreforeningen, the Institute of Clinical Medicine at Aarhus University, the County of Funen Research Foundation, and the Aarhus University Research Foundation for financial support and Lene Dabelstein Petersen for technical assistance.
(1.) Lee WM, Galbraith RM. The extracellular actin-scavenger system and actin toxicity. N Engl J Med 1992;326:1335-41.
(2.) Kew RR, Webster RO. Gc-globulin (vitamin D-binding protein) enhances the neutrophil chemotactic activity of C5a and C5a des Arg. J Clin Invest 1988;82:364-9.
(3.) Schneider GB, Benis KA, Flay NW, Ireland RA, Popoff SN. Effects of vitamin D binding protein-macrophage activating factor (DBP-MAF) infusion on bone resorption in two osteopetrotic mutations. Bone 1995;16:657-62.
(4.) Yamamoto N, Kumashiro R. Conversion of vitamin [D.sub.3] binding protein (group-specific component) to a macrophage activating factor by the stepwise action of [beta]-galactosidase of B cells and sialidase of T cells. J Immunol 1993;151:2794-802.
(5.) Bikle DD, Siiteri PK, Ryzen E, Haddad JG. Serum protein binding of 1,25-dihydroxyvitamin D: a reevaluation by direct measurement of free metabolite levels. J Clin Endocrinol Metab 1985;61:969-75.
(6.) Nykjaer A, Dragun D, Walther D, Vorum H, Jacobsen C, Herz J, et al. An endocytic pathway essential for renal uptake and activation of the steroid 25-(OH) vitamin [D.sub.3]. Cell 1999;96:507-15.
(7.) Engbaek F. A general procedure for optimizing concentrations of capture antibody, biotinylated detecting antibody, and enzyme-labeled avidin in ELISAs-application to assays for [alpha]-fetoprotein, prolactin, FSH and LH in serum. J Clin Immunoassay 1994;17:151-9.
(8.) Krouwer JS, Rabinowitz R. How to improve estimates of imprecision. Clin Chem 1984;30:290-2.
(9.) Lauridsen AL, Vestergaard P, Nexo E. Mean serum concentration of vitamin D-binding protein (Gc globulin) is related to the Gc phenotype in women. Clin Chem 2001;47:753-6.
(10.) Christensen EI, Birn H. Megalin and cubilin: synergistic endocytic receptors in renal proximal tubule. Am J Physiol Renal Physiol 2001;280:F562-F573.
(11.) St. John A, Thomas M, Dick I, Young P, Prince RL. Parathyroid function in mild to moderate renal failure: evaluation by oral calcium suppression test. J Clin Endocrinol Metab 1994;78:1436-8.
(12.) van Hoof HJ, de Sevaux RG, Van Baelen H, Swinkels LM, Klipping C, Ross HA, et al. Relationship between free and total 1,25-dihydroxyvitamin D in conditions of modified binding. Eur J Endocrinol 2001;144:391-6.
(13.) Auwerx J, De Keyser L, Bouillon R, De Moor P. Decreased free 1,25-dihydroxycholecalciferol index in patients with the nephrotic syndrome. Nephron 1986;42:231-5.
(14.) Grymonprez A, Proesmans W, Van Dyck M, Jans I, Goos G, Bouillon R. Vitamin D metabolites in childhood nephrotic syndrome. Pediatr Nephrol 1995;9:278-81.
Anna Lis Lauridsen,  * Michael Aarup,  Anna Lisa Christensen,  Bente Jespersen,  Kim Brixen,  and Ebba Nexo 
( Department of Clinical Biochemistry, Norrebrogade, Aarhus University Hospital, Aarhus, Denmark; Departments of  Nephrology and  Endocrinology, Odense University Hospital, Odense, Denmark; * address correspondence to this author at: Department of Clinical Biochemistry, Aarhus University Hospital, Norrebrogade 44, DK-8000 Aarhus C, Denmark; fax 45-8949-3060, e-mail email@example.com)
Table 1. Urinary and plasma concentrations of Gc, albumin, and creatinine in 10 healthy controls (28-66 years of age) and in 99 patients with various kidney diseases (20-88 years of age). Median (range) Healthy controls Patients Urinary Gc, nmol/24 h 0.62 (0.03-1.49) 27 (0.03-897) Gc in second-void 0.70 (0.07-1.11) 13 (0.04-375) morning urine, nmol/L Gc/Creatinine in 0.05 (0.01-0.09) 2.8 (0.01-62) second-void morning urine, nmol/mmol Urinary albumin, mg/24 h 17 (10-27) 734 (14-8204) Urinary creatinine, mmol/24 h 13 (10-22) 11 (4-20) Plasma Gc, mol/L 3.9-6.4 (e) 4.8 (3.4-7.7) Plasma albumin, g/L 36-51 (e) 41 (22-47) Plasma creatinine, mol/L 44-134 (e) 224 (59-712) Creatinine clearance, mL/min 72-138 (e) 37 (6-186) Correlation (a) with urinary Gc in patients Urinary Gc, nmol/24 h Gc in second-void 0.94 (b) morning urine, nmol/L Gc/Creatinine in 0.93 (b) second-void morning urine, nmol/mmol Urinary albumin, mg/24 h 0.86 (b) Urinary creatinine, mmol/24 h 0.05 Plasma Gc, mol/L 0.12 Plasma albumin, g/L -0.38 (b) Plasma creatinine, mol/L 0.32 (c) Creatinine clearance, mL/min -0.24 (d) (a) Spearman nonparametric correlation coefficients between 24-h urinary Gc and the other variables. (b-d) 24-h urinary Gc compared with the other variables: (b) P < 0.001; (c) P < 0.01; (d) P < 0.05. (e) 95% central reference interval.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Technical Briefs|
|Author:||Lauridsen, Anna Lis; Aarup, Michael; Christensen, Anna Lisa; Jespersen, Bente; Brixen, Kim; Nexo, Eb|
|Date:||Jun 1, 2005|
|Previous Article:||Use of low concentrations of human IgA anti-tissue transglutaminase to rule out selective IgA deficiency in patients with suspected celiac disease.|
|Next Article:||Serum erythropoietin measured by chemiluminescent immunometric assay: an accurate diagnostic test for absolute erythrocytosis.|