Clinical Chemistry 54:5 919–922 (2008)
Automated Time-Resolved Immunofluorometric Assay for Progastrin-Releasing Peptide Marianne S. Nordlund,* David J. Warren, Kjell Nustad, Johan Bjerner, and Elisabeth Paus
Department of Medical Biochemistry at Radiumhospitalet, Rikshospitalet Medical Centre, Oslo, Norway; * address correspondence to this author at: Department of Medical Biochemistry at Radiumhospitalet, Rikshospitalet Medical Centre, Montebello, N-0310 Oslo, Norway. Fax ⫹47 22 73 07 25; e-mail
[email protected]. BACKGROUND: Small cell lung cancer accounts for approximately 20% of new cases of lung cancer, and advanced disease is prevalent at the time of diagnosis. Neuron-specific enolase (NSE) has been the primary tumor marker in small cell lung cancer but it has relatively low sensitivity in early-stage disease. Progastrinreleasing peptide (proGRP) is a promising alternative or complementary marker for NSE. We have previously described a time-resolved immunofluorometric assay (TR-IFMA) for proGRP that lacked the necessary sensitivity and robustness for use in the routine clinical laboratory. Herein we describe the development of an improved assay using a novel monoclonal antibody pair. METHODS:
Mice were immunized with different conjugated proGRP peptides, including residues 31–98, 1–98, and preproGRP(-23–125). Pair combinations of the resulting monoclonal antibodies (mAb) were tested. The improved TR-IFMA was compared with the only other available proGRP assay, the proGRP ELISA (IBL).
RESULTS:
A panel of 12 high-affinity mAbs was produced. The best assay combination was between our original E146 mAb as solid-phase antibody and the new mAb M16 as tracer. The new TR-IFMA had a linear dose-response curve, a wide dynamic range (13– 13 500 ng/L), and a limit of detection of 2.8 ng/L. Total CV was ⬍5.6% over the whole measuring range. Bland-Altman difference analysis indicated a significant positive bias between the IFMA and the ELISA.
CONCLUSIONS: We describe a sensitive and robust mAbbased TR-IFMA for proGRP. The assay is fully automated and displays high quality performance.
Progastrin-releasing peptide (ProGRP) is a promising tumor marker for small cell lung cancer (SCLC) (1–3 ).
Brief Communications The human gastrin-releasing peptide (GRP) gene encodes a 148 –amino acid prepropeptide (⫺23–125). Protease processing of the peptide results in several products: a signal peptide (residues ⫺23–1), the 2 mature GRP forms (residues 1–27 and 18 –27), and the proGRP moiety (residues 31–125) (4 ). Several cleaved products of the proGRP peptide have also been detected. To date, only 1 immunoassay for proGRP is commercially available, a manual method based on polyclonal and monoclonal antibodies (mAb) (5 ). We have previously produced a panel of mAbs against recombinant proGRP (31–98) and constructed an immunofluorometric assay (6 ). The tracer antibody (E149) had a lower affinity for the solid-phase complex than for the free peptide, resulting in loss of assay robustness, most likely due to a conformational change in the proGRP molecule on binding to the E146 solidphase antibody. Here we describe the development and testing of a time-resolved immunofluorometric assay (TR-IFMA) that uses a novel antibody pair to measure proGRP. We used pGEX-6P-3 constructs for Escherichia coli expression of human proGRP (31–98), proGRP (1– 98), and proGRP (1–125), and purified these products from cell lysates as described previously (6 ). ProGRP (1–98) (1 mg), was conjugated to 1 mg thyroglobulin using 1-ethyl-3-[3-(dimethylamino)-propyl] carbodiimide (Pierce) (7 ). Recombinant human preproGRP (⫺23–125) was expressed in a eukaryotic system as a C-terminal fusion with the Fc-part of an IgG2b immunoglobulin (8 ). Female BALB/c mice were immunized with each of the peptides as described previously (6 ). Handling of animals followed national guidelines. Specific immune response was monitored with an antigen-capture assay using 125I-labeled proGRP (31– 98), proGRP (1–98), or proGRP (1–125) (6 ). Sera from immunized mice were also screened for formation of an immunometric assay with Eu-labeled mAb E146 using sheep antimouse– coated Maxisorp microtiter plates as described elsewhere (9 ). E146 was conjugated to an Eu chelate using the Delfia Eu3⫹-Labeling Kit (PerkinElmer Life and Analytical Sciences) (6 ). Fluorescence was measured as described in a later section. Hybridomas were produced by polyethylene glycol–facilitated fusion of splenocytes from the immunized mice with NS0 cell-line mouse myeloma cells (10, 11 ). The hybridomas were cloned twice by limiting dilution. Screening for anti-proGRP mAbs was performed in sheep antimouse– coated microtiter plates and in an immunometric assay as described above, and culture medium from an SCLC cell line (NCI-H128) was used as a source of proGRP. Pair combinations of the mAbs were evaluated in 2-site immunoradiometric assays essentially as described earlier (12 ). 919
Brief Communications
Table 1. Evaluation of antibody pair combinations by immunoradiometric assays.a Tracer mAbs
a
Solid-phase mAbs
M10
M17
M11
M16
M7
M8
M9
M15
M18
M19
E149
M10
1.2
0.7
15.6
17.5
0.5
0.4
0.5
0.5
0.4
0.1
0.8
M17
0.9
0.5
0.8
0.8
0.3
0.3
0.3
0.3
0.3
0.0
0.5
E146
1.7
2.3
51.1
66.8
22.0
24.3
20.5
21.6
18.7
19.2
9.3
M11
1.5
2.3
1.3
0.7
22.6
24.0
22.4
25.6
21.0
20.4
8.7
M16
1.7
2.3
1.3
1.0
23.8
24.9
22.3
25.2
20.9
19.2
9.3
M7
1.1
0.7
16.4
19.3
0.6
0.3
0.3
0.4
0.4
0.1
0.7
M8
1.1
0.6
15.5
19.4
0.5
0.3
0.4
0.6
0.4
0.0
0.5
M9
1.3
0.6
15.6
21.2
0.4
0.4
0.3
0.5
0.4
0.0
0.7
M15
1.0
0.7
1.1
0.5
0.5
0.5
0.5
0.5
0.5
0.1
0.6
M18
1.1
0.6
18.0
23.5
0.4
0.4
0.4
0.6
0.3
0.1
0.6
M19
0.9
0.5
5.9
7.9
0.5
0.5
0.6
0.7
0.4
0.1
0.5
The numbers represent percentage binding of labeled mAbs. All antibodies were tested both as solid-phase antibodies and tracers, except for E146 and E149, which were tested only in their original orientation. Microtiter plates coated with mAbs were incubated with proGRP(31–98) (675 pg/well) before 125I-labelled tracer antibodies (50 000 cpm/well) were added.
Calibrators were prepared using recombinant proGRP (31–98). The peptide was diluted in calibration buffer (PBS containing 6% BSA, 10 mmol/L EDTA, and 0.05% sodium azide). The calibrators were standardized against those provided in the proGRP ELISA (IBL), divided into aliquots, and stored at ⫺30 °C. After thawing they were kept at 4 °C and used within a week. The final assay was automated using an AutoDelfia instrument (PerkinElmer). Biotinylated F(ab⬘)2 fragment (0.2 g/well) of mAb E146 in 200 L proGRP buffer [50 mmol/L Tris-HCl pH 7.8, 150 mmol/L NaCl, 20 mol/L diethylene triamine penta acetic acid, 1 g/L Germall, 0.1 g/L Tween 20, 3 mg/L tartrazine, 5 g/L BSA, 1 g/L bovine IgG, and 60 mg/L MAK33-IgG Poly (Roche Diagnostics GmbH)] was incubated in streptavidin-coated microtiter wells (PerkinElmer) with shaking for 30 min. After being washed 3 times with Delfia wash solution (50 mmol/L Tris-HCL pH 7.8, 150 mmol/L NaCl, 0.05% Tween20, 0.1% Germall), 100 L of calibrator or sample and 100 L proGRP buffer were added to duplicate wells, followed by shaking for 60 min. After 6 washes, 0.2 g/ well of Eu3⫹-labeled M16 in 200 L proGRP buffer was added, and the plates were incubated for another 60 min under continuous shaking. After 6 washes, 200 L/well of enhancement solution was added, followed by incubation with shaking at room temperature for 5 min, and fluorescence was measured in a time-resolved fluorometer. The minimum detectable concentrations were calculated as the mean of the zero calibrator ⫹ 3SD by the MultiCalcTM software (PerkinElmer) in 15 separate 920 Clinical Chemistry 54:5 (2008)
runs. All runs were done in duplicate. The within-run and total imprecision of the assay were determined with 495 samples in 15 separate runs. To determine the linearity on dilution, we diluted a serum sample with a high concentration of proGRP 2-, 4-, 16-, 64-, 256-, and 512-fold in calibration buffer and serum. The dilution was performed by weight. Surplus samples (n ⫽ 174) from the hospital routine laboratory with high values in the neuron-specific enolase (NSE) assay (13 ) were analyzed simultaneously with our assay and the proGRP ELISA, and the differences were evaluated by use of the Bland-Altman method (14 ). Immunization with the proGRP (1–98)-thyroglobulin conjugate resulted in 10 antibodies recognizing both recombinant proGRP (1–98) and proGRP secreted by the SCLC cell line H128. Another fusion with the ⫺23–125 preproGRP-Fc peptide gave 2 antibodies that reacted strongly to recombinant peptide but poorly to the H128-derived proGRP (data not shown). Our success in producing new antibodies may be attributable to the use of extended proGRP peptides as immunogens. The 1–98 and ⫺23–125 amino acid long peptides may be more naturally folded molecules. Consequently these peptides could be more resistant to proteolytic cleavage than shorter peptides (15 ), a characteristic that may in turn increase the chance of prolonged exposure and better presentation to the immune system. Pair combinations of the new and formerly characterized mAbs E146 and E149 were evaluated in 2-site immunoradiometric assays (Table 1). We have previously used E146 as the solid-phase antibody and E149
Brief Communications
Fig. 1. Calibration curve and imprecision profile for the proGRP TR-IFMA and comparison of the TR-IFMA with the ELISA method. (A), The calibration curve is linear over a range of 13–13 500 ng/L. The precision profile shows total CVs ⬍5.6% over the entire working range. Total imprecision was calculated as the CV of all measured concentrations in all assay runs (n ⫽ 495). (B), The Bland-Altman plot shows the difference between assay results from the TR-IFMA and the ELISA. The x axis shows the mean concentrations of proGRP in samples (n ⫽ 174) preselected for increased NSE concentrations determined by both assays. The y axis shows the percentage difference [(TR-IFMA – ELISA)/mean of both assays) ⫻ 100]. The solid line represents no difference. The TR-IFMA displays a positive bias over the whole proGRP concentration range.
as tracer (6 ); therefore they were tested only in this orientation. Antibodies M15 and M17 did not perform well as solid-phase reagents, and M10 and M17 did not perform well as tracers. M7, M8, M9, M15, M18, and M19 were good as tracers and outperformed our original tracer antibody, E149. mAbs M11 and M16 proved to be outstanding as both solid-phase antibodies and tracers. The best combination was E146 as solid phase and M16 as tracer. The BIAcore X optical biosensor system showed that mAb M16 had a significantly lower dissociation rate from the E146-proGRP (31–98) antibody-antigen complex than our original tracer E149 (see Fig. 1 in the Data Supplement that accompanies the online version of this Brief Communication at http://www.clinchem.org/content/vol54/issue5). Hence, the possible conformational change in proGRP when bound to E146 had no observed effect on the interaction with tracer antibody M16. The final assay was automated, and to minimize the threat of heterophilic antibody interference modifications were introduced, including the use of F(ab⬘)2 fragments as solid phase and irrelevant immunoglobulin buffer additives (11, 16 ). Patient serum diluted to 35.6 and 995 ng/L proGRP was used to determine optimal assay kinetics. Sera were obtained in accordance with institutional review board approval. The signals reached a plateau after a 30-min incubation for the F(ab⬘)2 capture reagent, 1 hour for the antigen, and a 1 hour final incubation with the Eu3⫹-labeled M16 antibody (data not shown). The assay was linear and dis-
played a wide dynamic range (Fig. 1). In-house calibrators (0 –13 500 ng/L) were standardized against those supplied in the ELISA reagent set (IBL). When both sets of calibrators were used as samples in the 2 assays, the measured values agreed well with the assigned values. The detection limit was determined in 15 separate runs and the median calculated as 2.8 ng/L. The variation between replicates of 495 samples was low, with a total imprecision of ⬍5.6% over the whole assay range (Fig. 1). Linearity on dilution resulted in apparent mean recoveries of 101%–108% of expected in zero calibrator and 73%–110% in serum. Concentrations of proGRP were determined simultaneously by TR-IFMA and IBL ELISA in samples (n ⫽ 174) preselected for high NSE values. Patient serum samples with high proGRP values in our assay also had increased values in the ELISA, hence we identified the same patient group. However, in 92% of the samples, TR-IFMA yielded higher results than the ELISA. This result is demonstrated in a Bland-Altman difference plot (Fig. 1). The regression equation was determined to be: y ⫽ 2.822x0.942 (see Fig. 2 in the online Data Supplement). A considerable discrepancy was observed between individual serum samples (⫺22% to ⫹184%). One possible explanation for the discordant results is that the TR-IFMA detects proGRP species not recognized by the antibody pair in the ELISA. The implications of this apparent broader specificity remain to be elucidated. Clinical Chemistry 54:5 (2008) 921
Brief Communications In a patient serum sample with high concentration of the peptide (8484 ng/L), proGRP decreased rapidly during storage at room temperature but were stable for 24 hours at 4 °C and for at least 1 month at ⫺20 °C (see Fig. 3 in the online Data Supplement). The instability of proGRP in serum has recently been described by Yoshimura et al. (17 ). Because we at present lack reference material of adequate size and storage history, we are unable to provide a definite cutoff value. However, proGRP values in the normal sera run to date (n ⫽ 184) agree well with the expected concentration range determined using the ELISA (5 ). In conclusion, we describe a sensitive and robust TR-IFMA, based on mAbs, for quantifying serum pro-
GRP in the routine clinical laboratory. The assay is fully automated, protected from heterophilic antibody interference, and displays a wide measuring range.
Grant/Funding Support: This work was supported by the Norwegian Cancer Society. Financial Disclosures: None declared. Acknowledgments: We thank Hans Christian Åsheim and Hanne Sagsveen Hjorthaug for help in expression of the preproGRP(-23–125) fusion protein and Tone Varaas and Anne-Marie Sauren for skilled technical assistance.
References 1. Takada M, Kusunoki Y, Masuda N, Matui K, Yana T, Ushijima S, et al. Pro-gastrin-releasing peptide (31–98) as a tumour marker of small-cell lung cancer: comparative evaluation with neuron-specific enolase. Br J Cancer 1996;73:1227–32. 2. Stieber P, Dienemann H, Schalhorn A, Schmitt UM, Reinmiedl J, Hofmann K, Yamaguchi K. Progastrin-releasing peptide (ProGRP): a useful marker in small cell lung carcinomas. Anticancer Res 1999;19:2673– 8. 3. Miyake Y, Kodama T, Yamaguchi K. Pro-gastrinreleasing peptide(31–98) is a specific tumor marker in patients with small cell lung carcinoma. Cancer Res 1994;54:2136 – 40. 4. Dumesny C, Patel O, Lachal S, Giraud AS, Baldwin GS, Shulkes A. Synthesis, expression and biological activity of the prohormone for gastrin releasing peptide (ProGRP). Endocrinology 2006;147: 502–9. 5. Aoyagi K, Miyake Y, Urakami K, Kashiwakuma T, Hasegawa A, Kodama T, Yamaguchi K. Enzyme immunoassay of immunoreactive progastrin-releasing peptide(31–98) as tumor marker for small-cell lung carcinoma: development and evaluation. Clin Chem 1995;41:537– 43. 6. Nordlund MS, Fermer C, Nilsson O, Warren DJ, Paus E. Production and characterization of monoclonal antibodies for immunoassay of the lung
922 Clinical Chemistry 54:5 (2008)
7.
8.
9.
10.
11.
12.
cancer marker proGRP. Tumour Biol 2007;28: 100 –10. Grant G.A. Synthetic peptides for production of antibodies that recognize intact proteins. In: Coligan JE, Kruisbeek AM, Margulies DH, Shevach EM, Strober W, editors. Current protocols in immunology. New York: John Wiley & Sons; 2003. p 9.2.1–19. Aasheim HC, Munthe E, Funderud S, Smeland EB, Beiske K, Logtenberg T. A splice variant of human ephrin-A4 encodes a soluble molecule that is secreted by activated human B lymphocytes. Blood 2000;95:221–30. Nustad K, Dowell BL, Davis GJ, Stewart K, Nilsson O, Roijer E et al. Characterization of monoclonal antibodies directed against squamous cell carcinoma antigens: report of the TD-10 Workshop. Tumour Biol 2004;25:69 –90. Galfre G, Milstein C. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol 1981;73:3– 46. Warren DJ, Bjerner J, Paus E, Bormer OP, Nustad K. Use of an in vivo biotinylated single-chain antibody as capture reagent in an immunometric assay to decrease the incidence of interference from heterophilic antibodies. Clin Chem 2005;51: 830 – 8. Paus E, Nustad K, Bormer OP. Epitope mapping
13.
14.
15.
16.
17.
and affinity estimation of 83 antibodies against prostate-specific antigen. Tumour Biol 1999; 20(Suppl 1):52– 69. Paus E, Nustad K. Immunoradiometric assay for alpha gamma- and gamma gamma-enolase (neuron-specific enolase), with use of monoclonal antibodies and magnetizable polymer particles. Clin Chem 1989;35:2034 – 8. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346: 1085–7. Kristensen P, Winter G. Proteolytic selection for protein folding using filamentous bacteriophages. Fold Des 1998;3:321– 8. Bjerner J, Nustad K, Norum LF, Olsen KH, Bormer OP. Immunometric assay interference: incidence and prevention. Clin Chem 2002;48:613–21. Yoshimura T., Fujita K., Kawakami S., Kitahara M., Dowell B. Stability of pro-gastrin releasing peptide in serum and plasma [Abstract]. Tumor Biol 2007;28(Suppl 1):109.
DOI: 10.1373/clinchem.2007.101436
