Analysis of serum protein glycosylation by a differential lectin immunosorbant assay (dLISA)
© Li et al.; licensee BioMed Central Ltd. 2013
Received: 28 February 2013
Accepted: 8 August 2013
Published: 9 September 2013
Lectin immunosorbant assays (LISAs) have been widely used for analyzing protein glycosylation. However, the analysis of serum samples by LISAs could suffer from high sample-dependent background noise. The aim of this study is to develop a differential lectin immunosorbant assay (dLISA) with reduced background interferences.
For the analysis of protein glycosylation, dLISA establishes a dose–response curve for every serum sample. The sample is split into five aliquots. Four aliquots undergo differential removal of the glycoprotein of interest by immunoprecipitation. Then, all five aliquots are subject to two measurements: protein by immunoassay and protein glycans by LISA. A dose–response curve is established by plotting glycans signals on the y-axis and protein levels on the x-axis for all the aliquots. Slope of the curve, calculated by linear progression analysis and expressed as fluorescence per concentration of protein, is used for the measurement of protein glycosylation in the serum sample.
To demonstrate the feasibility of the dLISA approach, we used recombinant, fucosylated tissue inhibitor of metallopeptidase 1 (TIMP-1) as the target glycoprotein. Magnetic beads based TIMP1 immunoassay and TIMP-1 UEA LISA were developed for the measurement of TIMP1 protein and terminal α1, 2 fucosylated glycans on TIMP1, respectively. Serum samples supplemented with differentially fucosylated recombinant TIMP-1 were used to demonstrate that the slopes measured the TIMP-1 fucosylation, and were less prone to background interference.
Aberrant glycosylation of proteins has been implicated in many human diseases [1–3]. To aid in the diagnosis and prognosis, as well as in the understanding of these diseases at the molecular levels, there have been many research initiatives focusing on the development of analytical tools for effective analyses of subtle, yet biological significant, glycosylation changes [4–9]. One of the most informative and sensitive detection techniques for the analysis of glycans and glycoproteins is mass spectrometry (MS). Analytical tools combining MS with separation and enrichment techniques such as hydrophilic interaction chromatography (HILIC) and immunoaffinity enrichment are expected to provide a wealth of information on glycosylation changes that would allow a better understanding of the biological attributes of glycoproteins [10–13].
LISAs have been used in the discovery and validation of glycosylation of proteins as biomarkers [4, 6, 14–16]. Despite the potential advantages, LISAs have not been widely exploited. In LISAs, lectins could bind to antibodies since antibodies are glycoproteins, resulting in high backgrounds that could reduce sensitivity (Figure 1B). Several ways of blocking the lectin-antibody binding have been used in LISAs, such as (i) enzymatic release of the glycans off the antibody and (ii) oxidation of the glycans followed by derivatization with di-peptides. Although these treatments were able to reduce the background signals, reduction varied by lectin [4, 16, 18]. For some lectins (e.g., Pisum savitum Lectin and Lens Culinaris Agglutinin), background signals after blocking were still too high . Lectins could also bind to non-specific bound proteins (e.g., to the antibody and/or to solid surface that the antibody was adhered to) (Figure 1C-D). This could be problematic, especially when serum specimens are used, because the majority of serum proteins are glycosylated. The non-specific bound proteins could be reduced by the careful selection of blocking conditions. However, they could not be completely eliminated. Furthermore, the non-specific binding may vary from sample to sample. The sample-dependent background introduces significant variations in the signal generated by LISAs, making the comparison of glycosylation changes of proteins unreliable. In this study, we developed a dLISA approach for the analysis of protein glycosylation in serum that was less prone to background interference.
Materials and methods
TIMP-1 capture antibody, TIMP-1 biotinylated detection antibodies, and recombinant TIMP-1 protein were purchased from the R&D Systems (Minneapolis, MN). Bio-Plex Pro™ magnetic COOH beads, Amine Coupling Kits, and Cytokine Assay Kits were purchased from Bio-Rad Laboratories (Hercules, CA). Biotinylated Ulex europaeus agglutinin (UEA) was purchased from Vector Labs (Burlingame, CA). Dynabeads Antibody Coupling Kit was purchased from Invitrogen (Carlsbad, CA).
Serum samples from cancer and non-cancer patients (breast cancer, colon cancer, hepatocellular carcinoma, ovarian cancer, lung cancer, and prostate cancer) were obtained from the Serum Bank at the Center for Biomarker Discovery and Translation, Johns Hopkins University (Baltimore, MD) with the approval from the institutional review board.
TIMP-1 immunoassay was developed using the BioRad Cytokine Assay Kit (Hercules, CA). TIMP-1 capture antibody was coupled to Bio-Plex Pro™ magnetic COOH beads using the BioRad Amine Coupling Kit. The coupling was then validated using biotinylated goat anti-mouse IgG antibodies (Sigma-Aldrich, St. Louis, MO) to ensure binding of the capture antibody to the beads. After the validation, the beads were counted, and stored in the storage buffer at 4°C. For the TIMP-1 immunoassay, 2500 of the beads (per well for a 96-well plate) were incubated with 50 μL of a serum sample diluted in the Sample Diluent (provided in the BioRad Cytokine Assay Kit) at room temperature for 60 minutes. After the incubation, the beads were washed and incubated with 25 μL of 2 μg/mL biotinylated TIMP-1 detection antibody diluted in the Detection Antibody Diluent (provided in the Cyotkine Assay Kit) at room temperature for 30 minutes. Then the beads were washed again and incubated with 50 μL of 2 μg/mL streptavidin-phycoerytherin diluted in the Assay Buffer (provided in the Cyotkine Assay Kit) at room temperature for 10 minutes before analysis using the Bioplex 200 System (BioRad, Hercules, CA). Bioplex 200 uses 635 nm solid-state Laser to excite the fluorescent dyes inside the magnetic COOH beads to provide bead classification and assay identification information, and uses 532 nm Nd-Yag Laser to excite the phycoerytherin dye to generate a reporter signal. Recombinant TIMP-1 was used as standard. Concentrations of 0, 0.02, 0.1, 0.39, 1.56, 6.25, and 25 ng/mL of recombinant TIMP-1 were prepared in the Standard Diluent (provided in the Cytokine Assay Kit) as calibrators for establishment of a calibration curve for TIMP-1 protein quantification.
TIMP-1 UEA LISA
TIMP1 UEA LISA was established the same way as the TIMP1 immunoassay except that 20 μg/mL of biotinylated UEA was used for detection.
Immunoprecipitation of endogenous TIMP-1 in serum
Immunoprecipitation of endogenous TIMP-1 in serum was achieved using Dynabeads coupled with TIMP1 capture antibody. The coupling of the TIMP-1 capture antibody to the Dynabeads was performed using the Dynabeads Antibody Coupling Kit. For the immunoprecipitation, every 10 μL of serum was mixed with 5 μL of the coupled Dynabeads. The amount of beads (5 μL) was empirically determined to be the sufficient for the complete removal of endogenous TIMP-1 from serum samples. The mixture was then incubated on a rotator overnight at 4°C. For partial removal of the endogenous TIMP-1, fewer beads (e.g., 3uL) would be used.
Production of the differential fucosylated recombinant TIMP-1
Four microliter of the recombinant TIMP-1 (10 μg/mL prepared in PBS + 1% BSA buffer) was mixed with 4 μL of α1, 2 fucosidase (Catalog# P0724, New England Biolabs, Ipswich, MA). The mixture (8 μL) was then diluted 5 times in the 32 μL of the Reaction buffer (50 mM Sodium Citrate, 100 mM Sodium Chloride, pH 6.0) for incubation of 0, 15, 30, or 45 minutes with shaking at 37°C. After the incubation, each mixture was spiked into a serum aliquot. Presence of large quantities of glycoproteins in serum would stop the enzymatic digestion of the TIMP-1 by α1, 2 fucosidase. The serum aliquot did not contain endogenous TIMP-1, which was immuno-depleted. The immuno-depletion was confirmed using the TIMP-1 immunoassay.
Calibration curve for the TIMP-1 immunoassay was established using the 5-parameter nonlinear regression model in Bio-Plex Manager™ 6.0. Protein concentrations were calculated using the calibration curve and reported by Bio-Plex Manager™ 6.0. For linear regression and statistic analysis of the dose–response curves, we used Graphpad Prism 5.04. Limit of detection of the TIMP-1 immunoassay, determined using 3 times of standard deviation of the zero calibrator over six measurements, was 0.01 ng/mL. Reproducibility of the TIMP-1 immunoassays, determined using the coefficient variance of the calibrators at concentrations of 0.02, 0.1, 0.39, 1.56, 6.25, and 25 ng/mL, was less than 20%.
Overview of the dLISA approach
Although LISA assay appears to be similar to dLISA assay, they are different in several ways: (1) LISA assay is performed as a single measurement, whereas dLISA assay establishes a dose–response curve based on measurement of five aliquots; (2) LISA assay measures both protein abundance as well as protein glycosylation, both of which are combined together and reported as a single read out; on the other hand, the dLISA approach includes protein immunoassay, separating protein abundance from protein glycosylation, and therefore, specifically measures protein glycosylation; (3) Because it differential removed the target protein using immunoprecipitation, the dLISA approach is more specific for the target protein; (4) The dLISA approach also includes measurement of the background signal through the aliquot A5. This helps to identify serum samples with potential high background signal that interferes with LISA assay, and thus it demonstrates the advantages of the dLISA approach with reduced interferences.
Correlation of the slope with serum TIMP-1 fucosylation
Application of the dLISA approach to serum samples
Linear-regression slopes of the dose–response curves of four individual serum samples of normal, breast cancer, colon cancer, HCC, ovarian cancer, lung cancer, and prostate cancer
-0.1 ± 1.5
-0.9 ± 0.1
0.09 ± 0.13
-1 ± 0
2 ± 0
-0.4 ± 0.3
2 ± 1
-4 ± 1
-2 ± 0
0.07 ± 0.22
-1 ± 0
0.5 ± 0.5
-0.6 ± 0.5
1 ± 0
0.7 ± 0.0
-1 ± 0
-0.5 ± 0.2
-0.3 ± 0.0
2 ± 0
2 ± 0
0.51 ± 0.04
-0.3 ± 0.0
-0.8 ± 0.3
0.8 ± 0.2
0.5 ± 0.1
2 ± 1
-0.2 ± 0.2
0.1 ± 0.2
TIMP-1 belongs to a family of tissue inhibitors of matrix metalloproteinases (MMPs) whose inhibitory activities play important roles in cellular homeostasis, tissue remodeling, and oncogenesis . Expression of TIMP-1 was increased in cancer tissue in prostate . TIMP-1 was reported to have prognostic and predictive value in breast cancer [22, 23], and serum TIMP-1 was a predictor of survival outcomes in colorectal cancer . In addition, TIMP-1 has been implicated in MMP-independent actions, such as synaptic plasticity of the central nervous system .
Aberrant glycosylation of TIMP-1 was implicated in cancer progression. Increased β1, 6 branching of N-glycans of TIMP-1, induced by GnT-V N-acetylglucosaminyltransferase, was closely correlated with invasive/metastatic potential of colon cancer cell WiDr . Detection of TIMP-1 UEA α1, 2 fucosylation in prostate tissues was found to be superior to TIMP-1 protein in distinguishing aggressive and non-aggressive prostate cancer . Whether detection of TIMP-1 UEA fucosylation in serum could help identify aggressive prostate cancer remained to be determined. We measured TIMP-1 UEA fucosylation in sera of prostate cancer and found little UEA fucosyaltion. The seemingly contradictory findings between prostate tissues and sera could be due to the source differences of TIMP-1. Presence of TIMP-1 in serum may be a combination of physiology (e.g., for carrying out their functions in circulation), tissues leakages (e.g., as a result of cell death or damage), and/or aberrant secretions (e.g., released from tumors and other disease tissues, presumably not as a result of a functional requirement) . Given the potential sources of TIMP-1 in serum, if TIMP-1 from other sources were not fucosylated, the contribution of prostate tissue TIMP-1 would be too small to be detected. The same pattern was also observed in non-cancer sera and sera of other cancers such as breast, colon, lung, and ovarian, indicating that (1) TIMP-1 from other tissues may not be fucosylated, and (2) there was no difference in serum TIMP-1 fucosylation between non-cancer and cancers. This latter finding was consistent with Thaysen-Anderson et al., which showed no significant difference of TIMP-1 glycoprofiles between normal and colon cancer .
Ahn et al. showed that Leukocyte phytohemagglutinin (L-PHA) captured glycoforms of TIMP-1 in a pooled colon cancer serum was 5 times higher in abundance than that in a pooled non-cancer serum by mass spectrometric analysis . We did not try TIMP-1 L-PHA LISA in serum because the recombinant TIMP-1 did not have L-PHA bound glycans  and therefore could not be used as the standard protein for the feasibility study. This points out a bigger issue in the field of glycobiology. In the future, development of protein engineering technology that allows additions of glycans of interest to proteins and produces highly purified protein glycoforms  would help solve this problem.
Using recombinant TIMP-1 as the model, we determined a dLISA approach for the analysis of serum protein glycosylation that was less prone to potential interference of serum matrices. Applying the approach of analysis TIMP-1 fucosylation in serum samples, we found that serum TIMP-1 had little or no α1, 2 fucosylation in normal and many cancer conditions.
We thank the National Cancer Institute Early Detection and Research Network (NCI-EDRN) for providing grant support (U24CA115102). We thank Dr. Lori Sokoll and Debra Elliott at the Center for Biomarker Discover and Translation for providing the clinical specimens.
- Meany DL, Chan DW: Aberrant glycosylation associated with enzymes as cancer biomarkers. Clin Proteomics. 2011, 8: 7- 10.1186/1559-0275-8-7PubMed CentralView ArticlePubMedGoogle Scholar
- Lau KS, Dennis JW: N-Glycans in cancer progression. Glycobiology. 2008, 18: 750-760. 10.1093/glycob/cwn071View ArticlePubMedGoogle Scholar
- Ohtsubo K, Marth JD: Glycosylation in cellular mechanisms of health and disease. Cell. 2006, 126: 855-867. 10.1016/j.cell.2006.08.019View ArticlePubMedGoogle Scholar
- Meany DL, Zhang Z, Sokoll LJ, Zhang H, Chan DW: Glycoproteomics for prostate cancer detection: changes in serum PSA glycosylation patterns. J. Proteome Res. 2009, 8: 613-619. 10.1021/pr8007539PubMed CentralView ArticlePubMedGoogle Scholar
- Meany DL, Hackler L, Zhang H, Chan DW: Tyramide signal amplification for antibody-overlay lectin microarray: a strategy to improve the sensitivity of targeted glycan profiling. J. Proteome Res. 2011, 10: 1425-1431. 10.1021/pr1010873View ArticlePubMedGoogle Scholar
- Li Y, Tao SC, Bova GS: Detection and verification of glycosylation patterns of glycoproteins from clinical specimens using lectin microarrays and lectin-based immunosorbent assays. Anal. Chem. 2011, 83: 8509-8516. 10.1021/ac201452fPubMed CentralView ArticlePubMedGoogle Scholar
- Li Y, Tian Y, Rezai T: Simultaneous analysis of glycosylated and sialylated prostate-specific antigen revealing differential distribution of glycosylated prostate-specific antigen isoforms in prostate cancer tissues. Anal. Chem. 2011, 83: 240-245. 10.1021/ac102319gPubMed CentralView ArticlePubMedGoogle Scholar
- Ahn YH, Kim YS, Ji ES: Comparative quantitation of aberrant glycoforms by lectin-based glycoprotein enrichment coupled with multiple-reaction monitoring mass spectrometry. Anal. Chem. 2010, 82: 4441-4447. 10.1021/ac1001965View ArticlePubMedGoogle Scholar
- Thaysen-Andersen M, Thogersen IB, Lademann U: Investigating the biomarker potential of glycoproteins using comparative glycoprofiling - application to tissue inhibitor of metalloproteinases-1. Biochim. Biophys. Acta. 2008, 1784: 455-463. 10.1016/j.bbapap.2007.12.007View ArticlePubMedGoogle Scholar
- Mariño K, Bones J, Kattla JJ, Rudd PM: A systematic approach to protein glycosylation analysis: a path through the maze. Nature chemical biology. 2010, 6: 713-723.View ArticlePubMedGoogle Scholar
- Wuhrer M, de Boer AR, Deelder AM: Structural glycomics using hydrophilic interaction chromatography (HILIC) with mass spectrometry. Mass spectrometry reviews. 2009, 28: 192-206. 10.1002/mas.20195View ArticlePubMedGoogle Scholar
- Madian AG, Rochelle NS, Regnier FE: Mass-linked immuno-selective assays in targeted proteomics. Anal. Chem. 2013, 85: 737-748. 10.1021/ac302071kView ArticlePubMedGoogle Scholar
- Lund H, Lovsletten K, Paus E, Halvorsen TG, Reubsaet L: Immuno-MS based targeted proteomics: highly specific, sensitive, and reproducible human chorionic gonadotropin determination for clinical diagnostics and doping analysis. Anal. Chem. 2012, 84: 7926-7932. 10.1021/ac301418fView ArticlePubMedGoogle Scholar
- Mehta AS, Long RE, Comunale MA: Increased levels of galactose-deficient anti-Gal immunoglobulin G in the sera of hepatitis C virus-infected individuals with fibrosis and cirrhosis. J. Virol. 2008, 82: 1259-1270. 10.1128/JVI.01600-07PubMed CentralView ArticlePubMedGoogle Scholar
- Comunale MA, Wang M, Hafner J: Identification and development of fucosylated glycoproteins as biomarkers of primary hepatocellular carcinoma. J. Proteome Res. 2009, 8: 595-602. 10.1021/pr800752cPubMed CentralView ArticlePubMedGoogle Scholar
- Li D, Chiu H, Chen J, Zhang H, Chan DW: Integrated Analyses of Proteins and Their Glycans in a Magnetic Beads Based Multiplex Assay Format Clinical Chemistry. 2012,Google Scholar
- Li D, Chiu H, Gupta V, Chan DW: Validation of a multiplex immunoassay for serum angiogenic factors as biomarkers for aggressive prostate cancer. Clin. Chim. Acta. 2012, 413: 1506-1511. 10.1016/j.cca.2012.06.017PubMed CentralView ArticlePubMedGoogle Scholar
- Chen S, Haab BB: Analysis of glycans on serum proteins using antibody microarrays. Methods Mol. Biol. 2009, 520: 39-58. 10.1007/978-1-60327-811-9_4PubMed CentralView ArticlePubMedGoogle Scholar
- Thaysen-Andersen M, Thogersen IB, Nielsen HJ: Rapid and individual-specific glycoprofiling of the low abundance N-glycosylated protein tissue inhibitor of metalloproteinases-1. Mol. Cell. Proteomics. 2007, 6: 638-647. 10.1074/mcp.M600407-MCP200View ArticlePubMedGoogle Scholar
- Moore CS, Crocker SJ: An alternate perspective on the roles of TIMPs and MMPs in pathology. Am. J. Pathol. 2012, 180: 12-16. 10.1016/j.ajpath.2011.09.008View ArticlePubMedGoogle Scholar
- Liu AY, Zhang H, Sorensen CM, Diamond DL: Analysis of prostate cancer by proteomics using tissue specimens. J. Urol. 2005, 173: 73-78. 10.1097/01.ju.0000146543.33543.a3View ArticlePubMedGoogle Scholar
- Dechaphunkul A, Phukaoloun M, Kanjanapradit K: Prognostic significance of tissue inhibitor of metalloproteinase-1 in breast cancer. International journal of breast cancer. 2012, 2012: 290854-PubMed CentralView ArticlePubMedGoogle Scholar
- Zhu D, Zha X, Hu M: High expression of TIMP-1 in human breast cancer tissues is a predictive of resistance to paclitaxel-based chemotherapy. Med. Oncol. 2012, 29: 3207-3215. 10.1007/s12032-012-0239-3View ArticlePubMedGoogle Scholar
- Nielsen HJ, Brunner N, Jorgensen LN: Plasma TIMP-1 and CEA in detection of primary colorectal cancer: a prospective, population based study of 4509 high-risk individuals. Scandinavian journal of gastroenterology. 2011, 46: 60-69. 10.3109/00365521.2010.513060View ArticlePubMedGoogle Scholar
- Jourquin J, Tremblay E, Bernard A: Tissue inhibitor of metalloproteinases-1 (TIMP-1) modulates neuronal death, axonal plasticity, and learning and memory. The European journal of neuroscience. 2005, 22: 2569-2578. 10.1111/j.1460-9568.2005.04426.xView ArticlePubMedGoogle Scholar
- Kim YS, Hwang SY, Kang HY: Functional proteomics study reveals that N-Acetylglucosaminyltransferase V reinforces the invasive/metastatic potential of colon cancer through aberrant glycosylation on tissue inhibitor of metalloproteinase-1. Mol. Cell. Proteomics. 2008, 7: 1-14.View ArticlePubMedGoogle Scholar
- Anderson NL, Anderson NG: The human plasma proteome: history, character, and diagnostic prospects. Mol. Cell. Proteomics. 2002, 1: 845-867. 10.1074/mcp.R200007-MCP200View ArticlePubMedGoogle Scholar
- Ahn YH, Kim KH, Shin PM, Ji ES, Kim H, Yoo JS: Identification of low-abundance cancer biomarker candidate TIMP1 from serum with lectin fractionation and peptide affinity enrichment by ultrahigh-resolution mass spectrometry. Anal. Chem. 2012, 84: 1425-1431. 10.1021/ac2024987View ArticlePubMedGoogle Scholar
- Hang HC, Bertozzi CR: Chemoselective approaches to glycoprotein assembly. Accounts of chemical research. 2001, 34: 727-736. 10.1021/ar9901570View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.