Printer Friendly

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.

The immunometric (two-site) assay format has gained increasing popularity in the routine clinical laboratory setting. However, behind the relative simplicity and high specificity and sensitivity of this methodology lies the ever-present problem of assay interference. Heterophilic and human anti-mouse antibodies (HAMAs) [1] are important sources of both positive and negative interference (1, 2). In a review, Kaplan and Levinson (3) distinguished between heterophilic antibodies and HAMAs on the basis of their origin and the type of assay interference they induce. HAMAs are produced as a result of exposure to murine immunoglobulins as occurs, for example, during diagnostic or therapeutic procedures. They display the properties of antibodies that have undergone affinity maturation, possessing high avidity and the ability to cause assay interference by cross-linking the capture reagent to the detection antibody in the presence or absence of analyte. In contrast, the origin of heterophilic antibodies is less clear. Heterophilic antibodies (antianimal) often also display rheumatoid factor (anti-human) reactivity (4, 5), making heterophilic antibodies and rheumatoid factors both overlapping and potentially confusing entities. Transient assay interference attributable to heterophilic immunoglobulins has been documented in case reports (6, 7) suggesting that some are produced as a result of an antigen-driven process. The corresponding antigens, however, are usually not known; the antibodies appear to be produced by B cells that have undergone little somatic mutation, and the antibodies are generally of low avidity (8, 9). It has been postulated that heterophilic antibodies may arise through immunization via the gastrointestinal tract or that they play a role in regulation of the immune system via idiotypic network interactions (2,10).

The frequency of interference from heterophilic antibodies has been investigated in several studies, which reported prevalences ranging from 0.5% to 12% (11-15). Two studies, however, both using one-step immunometric assays, found interference in as many as 40% (16) and 52% (17) of samples.

Several methods have been used to reduce interference in immunoassays. Attempts at removal of the interfering immunoglobulins by sample pretreatment with polyethylene glycol precipitation (18), affinity chromatography on protein A (10), or use of sulfhydryl agents and detergents (19) have been described, but these techniques are cumbersome and not suited to high-throughput assays. A more common approach is the use of a modified assay buffer containing "blocking agents" such as bovine immunoglobulins or irrelevant murine antibodies in either native or aggregated form (2). In a previous study (20), we found a 4.0% frequency of interference in our in-house immunoassay for carcinoembryonic antigen (CEA) before the addition of nonspecific mouse immunoglobulins to the assay buffer. The addition of an irrelevant mouse monoclonal antibody (MAK33) at a concentration of 15 mg/L had little effect on the frequency of interference (3.9%), whereas the inclusion of a similar concentration of heat-aggregated MAK33 reduced interference to 0.86%. The substitution of a [F(ab').sub.2] fragment of the capture monoclonal antibody (Mab), however, was sufficient to reduce interference to 0.1% even in the absence of irrelevant mouse immunoglobulin blocking agents. This observation is in agreement with previous studies suggesting that the majority of interfering immunoglobulins are targeted at the Fc portion of murine IgG (21,22).

The preparation of [F(ab').sub.2] fragments for roubne diagnostic assays is time-consuming and expensive and requires considerable technical expertise. The development of techniques for the molecular cloning of immunoglobulin variable domains has permitted their engineering into a wide variety of structural formats. The single-chain Fv fragment (scFv), consisting of the [V.sub.H] and [V.sub.L] domains connected by a short flexible linker peptide, has gained widespread use because of its stability, the ease of production in Escherichia coli, and the ability to modify recombinant antibodies by creating gene fusions (23). In the present study, we constructed a scFv, using mRNA from the well-characterized anti-CEA hybridoma cell line T84.66 (24, 25). The single-chain antibody is produced as a fusion protein containing a 17-amino acid C-terminal biotin acceptor domain that is specifically biotinylated during expression in E. coli. The in vivo-biotinylated anti-CEA scFv was purified from bacterial lysates by monovalent avidin chromatography and used to establish an immunofluorometric assay for CEA in streptavidincoated microplates. The incidence of assay interference in this novel assay was compared with the interference in two additional assays that use either chemically biotinylated T84.66 Mab or its corresponding [F(ab').sub.2] fragment.

Materials and Methods


General laboratory reagents were purchased from either Sigma or Merck unless otherwise noted. E. coli strains TG1 and HB2151 and the M13KO7 helper phage were from Amersham Pharmacia Biotech. The phage display vector pAK100 and the expression vector pAK400 were gifts from Dr. A. Phickthun (University of Zurich, Zurich, Switzerland). Enzymes used for recombinant DNA techniques were from New England Biolabs. Water purified by reversed osmosis followed by polishing in a MilliQ UF-PLUS system (Millipore Corporation) was used for the preparation of aqueous reagents. Protein concentrations were measured by the Bradford method (26) calibrated with a bovine serum albumin (BSA) calibrater. All pH measurements were performed at room temperature.

Mabs. Hybridoma cell lines T84.66 and 9E10 were obtained from American Type Culture Collection. The production and characterization of clones 12-140-1 and 12140-10 has been described previously (27). All cell cultures were maintained in RPMI 1640 (Invitrogen) containing sodium bicarbonate and supplemented with 10 mmol/L HEPES (pH 7.2), 1 mmol/L sodium pyruvate, 75 [micro]mol/L monothioglycerol, and 100 mL/L fetal calf serum (Biochrom KG). Monoclonal antibodies were purified from culture supernatants or ascites fluid by chromatography on protein A-Sepharose (Amersham Pharmacia Biotech) as described under the preparation of [F(ab').sub.2] fragments.

Preparation of [F(ab').sub.2] fragments of Mab T84.66. Bromelain cleavage as described by Bjerner et al. (20) was used to prepare [F(ab').sub.2] fragments of the T84.66 Mab. Antibody (13.6 mg) in 10 mL of buffer (50 mmol/L Tris-HCI, 100 mmol/L NaCl, 5 mmol/L EDTA, pH 7.0) was incubated with 0.68 mg (0.23 mL) of bromelain (ID-Diluent 1; Diamed) for 2 h at 37[degrees]C. The reaction was terminated by the addition of a 1/10 volume of 200 mmol/L N-ethylmaleimide. The digestion products were applied to a protein A column equilibrated in buffer containing 100 mmol/L Na[H.sub.2]P[O.sub.4] and 0.1 g/L Na[N.sub.3] (pH 8.2) and eluted with a linear gradient buffer composed of 25 mmol/L citric acid, 25 mmol/L NaH2PO4, 25 mmol/L NaCl, and 0.1 g/L Na[N.sub.3] (pH 3.2). Fractions containing [F(ab').sub.2] fragments were pooled and dialyzed against 150 mmol/L NaCl before biotinylation. It should be noted that this procedure has been tested for several mouse antibodies of IgG1 subclass but not for mouse antibodies of other subclasses.

Biotinylation of [F(ab').sub.2] and IgG capture antibodies. Antibodies were biotinylated with a fivefold molar excess of N-hydroxysuccinimidyl-6-biotinamido hexanoate (Vector). We added 58 nmol of biotinylation reagent in 49 [micro]L of dimethyl sulfoxide to 9.25 mg of whole-IgG T84.66 (in 3.8 mL of 0.15 mol/L NaCI) or 91 nmol of biotinylation reagent in 208 [micro]L of dimethyl sulfoxide to 9.06 mg of T84.66 [F(ab').sub.2] fragments (in 4.68 mL of 0.15 mol/L NaCI). Additionally, we added 0.5 mol/L sodium borate (pH 8.0) to a final concentration of 0.1 mol/L and incubated the solutions for 2 h at room temperature. Free biotin was removed by dialysis against 50 mmol/L Tris-HCI, 150 mmol/L NaCl, 0.5 g/L Na[N.sub.3] (pH 7.8).

Europium labeling of anti-CEA tracer antibody 12-140-1. We conjugated the Mab to a europium chelate (Perkin-Elmer Life Sciences) by use of a 12.5-fold molar excess of labeling reagent to antibody. Antibodies (10 mg) in 5.2 mL of 150 mmol/L NaCl were incubated for 48 h at room temperature with 0.75 [micro]mol of [Eu.sup.3+]-chelate (250 [micro]L) and 550 [micro]L of sodium borate (0.5 mol/L, pH 8.6). Excess label was removed by gel filtration on a PD 10 colurnn (Amersham Pharmacia Biotech) equilibrated with 50 mmol/L Tris-HC1,150 mmol/L NaCl, 0.5 g/L Na[N.sub.3] (pH 7.8). The stock conjugate was stored at 4 [degrees]C in the Tris buffer. Before assays, the stock conjugate was diluted to a working stock with a conjugate concentration of 25 mg/L in the same buffer, and diethylenetriamine pentaacetic acid-treated BSA (PerkinElmer Life Sciences/Wallac Oy) was added to a final concentration of 1 g/L.

Preparation and iodination of CEA. CEA was purified from liver metastases from colorectal carcinomas by perchloric acid extraction followed by ion-exchange and gel-filtration chromatography (28). The protein was iodinated by the indirect Iodogen method (Pierce) with [Na.sup.125]I (Amersham Pharmacia Biotech) at an equimolar ratio of protein to iodine. CEA for calibrators was diluted to concentrations of 0, 3.6, 9.6, 36, 180, and 900 [micro]L with a buffer containing 50 mmol/L Tris-HCI (pH 7.4), 100 mmol/L NaCl, 1 g/L Germall II (ISP Sutton Laboratories), and 60 g/L BSA.

Amplification of T84.66 V region genes and assembly into the scFv format for phage display. Amplification of immunoglobulin V regions, scFv assembly, and rescue of phages displaying scFv were performed essentially as described (23). Total RNA was extracted from T84.66 cells, and mRNA was isolated by use of oligo(dT) latex beads (Qiagen). Approximately 1 [micro]g of mRNA was reverse-transcribed with use of random hexamer primers (Amersham Pharmacia Biotech), and heavy- and light chain genes were amplified separately. The [V.sub.H] and [V.sub.L] PCR products were gel-purified and spliced, with overlap extension used to create a [V.sub.L] ([G.sub.4]S)4-[V.sub.H] format. After a subsequent round of PCR amplification with an outer primer set (scback/scfor), the gel-purified scFv fragments were digested with SfiI, ligated into the phage display vector pAK100, and transformed into electrocompetent TG-1 cells. After overnight incubation at 26 [degrees]C on 2xYT (16 g/L tryptone, 10 g/L yeast extract, 5 g/L NaCI) agar plates containing 10 g/L glucose and 25 mg/L chloramphenicol, phages displaying scFv were prepared for panning by rescue using the M13K07 helper phage.

Selection of phages binding CEA by immunopanning. Phage particles (~[10.sup.11] transducing units) were subjected to one round of panning in MaxiSorp immunotubes (Nunc) coated with 20 [micro]g of CEA in 2 mL of phosphate-buffered saline (PBS). Panning was performed as described previously (29). ARer extensive washing, bound phages were eluted from the tube surface with 100 mmol/L triethylamine and, after neutralization with 1 mol/L Tris-HCI (pH 7.4), were used to infect E. coli HB2151 cultures. Infected cells were plated on 2xYT agar containing 10 g/L glucose and 25 mg/L chloramphenicol and incubated overnight at 26 [degrees]C.

Screening for production of soluble anti-CEA single-chain antibody. Individual HB2151 colonies were inoculated in 96-well deep-well microplates. Each clone was cultured at 26 [degrees]C in 2xYT medium containing 1 g/L glucose and 30 mg/L chloramphenicol to mid-log phage and induced by the addition of isopropyl [beta]-D-thiogalactopyranoside to 1 mmol/L. ARer overnight culture at 26[degrees]C, cells were removed by centrifugation, and the supernatant was screened for anti-CEA single-chain antibody production by use of an [sup.125]I-labeled CEA capture assay. In brief, the myc-tagged single-chain antibodies were immobilized on well surfaces of Maxisorp 96-well microplates (Nunc) previously coated with 1 [micro]g/well of the anti-c-myc Mab 9E10. The plates were washed three times with wash buffer (PBS containing 0.5 mL/L Tween 20), and 100 [micro]L of PBS containing 10 g/L BSA and 100 000 cpm of [sup.125]I-labeled CEA was added to each well. After incubation with shaking for 1 h, the plates were again washed three times, and CEA binding was determined by a Model 1470 gamma counter (PerkinElmer Wallac).

Construction of the expression vector pAK400Bio-T84.66. pAK400Bio was constructed by replacing the his, coding region contained in the original pAK400 vector with a synthetic oligonucleotide duplex encoding the C-terminal biotinylation peptide GGGLNDIFEAQKIEWHE (30). Briefly, a 100-fold molar excess of a DNA cassette, produced by annealing two partially complementary oligonucleotides, Bio1 (5'-000CCGATGGTGGCGGTCTGAACGACATCTTCGAGGCTCAGAAAATCGAATGGCACG- AATAGTA-3') and Bio2 (5'-AGCTTACTATTCGT000ATTCGATTTTCTGAGCCTCGAAGATGTCGTTCAGACC- GCCACCATCGGCCCCCG-3'), was ligated to pAK400 that had been linearized previously by digestion with HindIII. After removal of excess oligonucleotide, the construct was digested with Sfil, and the linear pAK400Bio vector was purified by agarose gel electrophoresis. For expression of biotin-tagged single-chain antibodies, selected T84.66 scFv fragments were excised from the pAK100 phage-display vector by digestion with Sfil and ligated directly into the linearized pAK400Bio. Ligation products were transformed in an E. coli K12 strain (AVB100) containing a chromosomal copy of the biotin holoenzyme synthetase (birA) gene under the control of an L-arabinose-inducible promoter (Avidity Inc.).

Production and purification of in vivo-biotinylated T84.66 single-chain antibody. Transformed E. coli AVB100 cells were grown overnight at 28 [degrees]C in 2xYT broth supplemented with 10 g/L glucose and 30 mg/L chloramphenicol. This starter culture was used to inoculate 2 L of 2xYT broth containing 1 g/L glucose, 30 mg/L chloramphenicol, 50 mmol/L potassium phosphate (pH 7.2), and 5 mmol/L MgS[O.sub.4] to an initial absorbance ([A.sub.600]/cm) of 0.05. Cultures were then grown at 24[degrees]C with continuous shaking to an [A.sub.600]/cm of 0.8, and production of T84.66 scFv and biotin holoenzyme synthetase was induced by the addition of isopropyl [beta]-n-thiogalactopyranoside to 250 [micro]mol/L and L-arabinose to 1.5 [micro]mol/L in the presence of 50 [micro]mol/L n-biotin. Sixteen hours after induction, cells were harvested by centrifugation for 20 mm at 2800g. The bacterial cell pellet was resuspended in 32 mL of ice-cold 200 mmol/L Tris-HCI (pH 8.0) containing 0.5 mmol/L EDTA and 500 mmol/L sucrose (Tris-EDTA-sucrose), and a periplasmic extract was prepared by the subsequent addition of 48 mL of a 1:5 dilution of Tris-EDTA-sucrose buffer in water. After incubation on ice for 30 min, a supernatant fraction was harvested by centrifugation for 10 min at 40 000g, and 400 U of Benzonase nuclease (Novagen) was added. Extracts were buffer-exchanged into PBS by use of Sephadex G-25 (Amersham Pharmacia Biotech) chromatography and applied to a 1-mL column of monovalent avidin prepared essentially as described previously (31). The column was washed extensively with PBS, and the biotinylated T84.66 single-chain antibody was eluted with PBS containng 2 mmol/L n-biotin. Protein-containing fractions were pooled and exchanged, by use of disposable Sephadex G-25 columns (Amersham Pharmacia Biotech), into PBS containing 1 g/L BSA and 0.2 g/L sodium azide. The purified antibody was filter-sterilized and stored at 4[degrees]C.


Assay buffers. We used a basic assay buffer (BSA buffer) consisting of 50 mmol/L Tris-HCI (pH 7.4), 150 mmol/L NaCl, 20 mmol/L diethylenetriamine pentaacetic acid, 5 mg/L tartrazine (Aldrich), 1 g/L Germall II, 10 mg/L Triton X-100, and 5 g/L BSA. In certain assays, a modified buffer (blocking buffer) was used. This was prepared by supplementing the BSA buffer with 0.5 g/L bovine IgG (cat. no. G7516; Sigma) and 15 mg/L heat-aggregated nonspecific murine antibody MAK33 (Roche Molecular Biochemicals) prepared as described previously (20).

Plate washing buffer. Washing buffer consisted of 50 mmol/L Tris-HCI (pH 7.4), 150 mmol/L NaCl, 1 g/L Germall II, and 0.5 g/L Tween 20.

Assays. Streptavidin-coated DELFIA microtitration strips (PerkinElmer Life Sciences) were used for all CEA assays. Automated assays were performed on AutoDELFIA instruments (PerkinElmer Life Sciences). The solid phase was prepared by addition of 150 [micro]L/well of assay buffer containing the biotinylated capture antibody, and the plates were incubated for 30 min with continuous shaking. In the three assays, 0.24 [micro]g of IgG, 0.15 [micro]g of [F(ab').sub.2] fragments, or 0.08 [micro]g of scFv was used, representing 2.3-3 nmol of binding sites for each capture reagent. Plates were then washed three times, and 25-[micro]L duplicates of samples, calibrators, or controls and 125 [micro]L/well of buffer were pipetted into the microtitration plates. After 150 mm of incubation with continuous shaking, the plates were washed three times, and 150 [micro]L of europium-labeled tracer antibody 12-140-1 (75 ng) was added to each well. The plates were then incubated with shaking for 30 mm, and after six additional washes, 200 [micro]L/well of DELFIA enhancement solution (PerkinElmer Life Sciences) was added. After 10 mm of incubation, the time-resolved fluorescence was measured. Results were calculated with Multicalc software (PerkinElmer Life Sciences), using the spline curve fit. Manual assays were performed with the same reagents and under the same conditions as the automated assays with the exception that a Wallac Plate Shaker, a Wallac Plate Washer, and a Victor 1420 multilabel counter (all from PerkinElmer Life Sciences) were used.

Patient samples. We first selected serum samples from 390 individuals displaying interference in a previous study (20). In this number we included all samples available with interferences difficult to eliminate (types IV, V, or VI in the previous study). The samples selected are described in Table 1. We then selected serum samples from 179 individuals not displaying interference in the same study. CEA concentrations, as measured with the T84.66 [F(ab').sub.2] assay with buffer containing blockers, ranged from 0.2 to 527.5 [micro]L with a median concentration of 2.8 [micro]L. The interfering agents in our previous study had later been confirmed to be principally human IgM class heterophilic antibodies (Bjerner et al., submitted for publication). Of the 390 serum samples displaying interference, the specificities of the interfering agents were further characterized for 344 samples (20). Thirteen samples were found to contain heterophilic antibodies binding to the [F(ab').sub.2] fragment, whereas the remaining 331 samples contained heterophilic antibodies directed against the murine Fc fragment.

Controls and assay performance. Assay controls included a lyophilized serum control (Sero) and a serum pool prepared from interference-negative samples. Precision profiles were calculated with Multicalc software.

Statistical analysis. Because heterophilic antibodies binding scFv fragments can be expected to be a part of a larger polyclonal antibody response, we set the limits low, aiming to also detect low concentrations of scFv-binding antibodies. When comparing assay results, we therefore first corrected for the small differences between assays, using the results from control samples. We then calculated the combined standard deviation of the differences of assay results, using the precision profiles of compared assays. We finally chose the 99% level of statistical confidence (2.56 x SD). Differences <1 [micro]L were not considered even if statistically significant.


PRODUCTION OF IN VIVO-BIOTINYLATED T84.66 scFv Construction of T84.66 scFv phage library. We constructed a scFv phage-display library with mRNA extracted from the T84.66 hybridoma cell line. After PCR assembly, the [V.sub.L]-[([G.sub.4]S).sub.4]-[V.sub.H] gene fragments were cloned as SfiI cassettes upstream of the coding region for the gIII minor phage coat protein contained in the pAK100 phagemid display vector. After transformation into the E. coli supE strain TG1, recombinant virus displaying scFv-gIII fusion proteins were produced by phage rescue and subjected to one round of immunopanning on CEA-coated tubes. Phage particles binding to CEA were used to infect the E. coli supE- strain HB2151. In this nonsuppressor strain, the amber codon between the scFv construct and the gIII gene is not suppressed, which allows the production of soluble single-chain antibodies fused to a C-terminal c-myc tag. We subsequently screened individual HB2151 clones, using an [sup.125]I-CEA capture assay with microtiter plates previously coated with the anti-c-myc Mab 9E10. Of 96 clones tested, 87 produced soluble anti-CEA single-chain antibodies. One positive clone (scFvCEA10) was selected for subcloning of the [V.sub.L]-[([G.sub.4]S).sub.4]-[V.sub.H] cassette into the biotin-tagging vector pAK400Bio.

Production of soluble biotin-tagged anti-CEA scFv. The [V.sub.L] [([G.sub.4]S).sub.4]-[V.sub.H] cassette from scFvCEA10 was excised by use of SfiI and subcloned in frame with the coding region of a C-terminal biotin acceptor domain (GGGLNDIFEAQKIEWHE) contained in the vector pAK400Bio. The lysine residue of this domain is site-specifically biotinylated in E. coli through a reaction catalyzed by the enzyme biotin holoenzyme synthetase (BirA). Induction conditions were optimized for production of both scFv and BirA in E. coli strain AVB100 by titration of their respective inducing agents (data not shown). Recombinant biotinylated scFvCEA10 was purified from soluble bacterial periplasmic extracts by affinity chromatography on monovalent avidin-Sepharose. Purification yields for the scFv varied from 0.7 to 1.3 mg/L of culture with a purity of >95% (Fig. 1). The apparent molecular mass determined by sodium dodecyl sulfate--polyacrylamide gel electrophoresis under reducing conditions was ~34 kDa. The purified biotinylated scFvCEA10 antibody was found to be stable for at least 8 months when stored at 4 [degrees]C.


Assay optimization. The CEA assays were set up as three-step immunofluorometric assays. The microtiter wells were first coated with biotinylated T84.66 [F(ab').sub.2] fragments, and we found that 150 ng/well was optimal. To make a fair comparison between IgG and scFv, these capture reagents were used in equimolar amounts of binding sites, i.e., 240 ng of T84.66 IgG or 80 ng of scFv. This gave comparable dose responses as shown in Fig. 2. The rest of the procedure was similar to our routine CEA assay and used a common tracer antibody and calibrators. Nonspecific binding in assays was dependent on buffer additives. Median nonspecific binding was higher in assays with BSA as the only protein additive [1348, 1356, and 1350 cps for T84.66 whole-IgG, [F(ab').sub.2] and scFv-based assays, respectively] than in assays with bovine IgG and aggregated MAK33 added to the assay buffer [296 and 483 cps for T84.66 [F(ab').sub.2] and scFv-based assays, respectively]. Median detection limits, as calculated by Wiacalc, for assays were similar: 0.12 [micro]L for whole-IgG T84.66-based assays, 0.17 [micro]L for T84.66 [F(ab').sub.2] based assays, and 0.13 [micro]L for scFv-based assays.



Interference in an assay using whole-IgG T84.66 capture antibody (IgG assay). Because samples were originally selected for interference by use of our in-house CEA assay with capture Mab 12-140-10 (20), we first screened for heterophilic antibodies that also bound Mab T84.66. Samples were thus reanalyzed for interference with wholeIgG T84.66 used as capture reagent without buffer additives (BSA buffer) and compared with our "gold standard" assay format using [F(ab').sub.2] T84.66 with buffer additives [[F(ab').sub.2] assay with blocking buffer]. Of the 390 samples previously displaying interference with capture Mab 12-140-10, 360 still displayed significant interference. Of the previously interference-negative samples, 23 of the 179 samples displayed interference. The differences between the results obtained with the unprotected IgG assay and the [F(ab').sub.2] assays varied greatly (Fig. 3).

Interference in an assay using T84.66 scFv capture antibody (scFv assay). The test assay, which used single-chain T84.66 as capture antibody with only BSA added to buffer (scFv assay with BSA buffer), was compared with our gold standard assay, which uses [F(ab').sub.2] fragments from T84.66 as capture antibody and with bovine IgG and heat-aggregated MAK33 added to buffer (blocking buffer). Even in the absence of blocking agents, the single-chain assay gave discrepant CEA results for only 9 of the 569 samples. Reanalysis of these samples in the scFv assay but with blocking buffer replacing the BSA buffer gave assay results similar to these obtained by the [F(ab').sub.2] assay for eight of the nine samples. Only sample 233 (Fig. 4) gave a discrepant result and will be described in detail in the next section.


Cross-reactivity between heterophilic antibodies binding [F(ab').sub.2] and scFv fragments. Of the 569 samples, 344 had been previously tested for heterophilic antibodies interfering in a CEA assay using [F(ab').sub.2] fragments of Mab 12-140-10 as capture reagent. Thirteen of the 344 samples contained such interfering heterophilic antibodies. We first had to confirm that those heterophilic antibodies also interfered in the test assay using T84.66 [F(ab').sub.2] fragments as capture reagent. Three of the 13 samples did not show any interference in the CEA assay using T84.66 [F(ab').sub.2] fragments as capture reagent when the test assay was performed without blocking immunoglobulins added to buffer. Eight of the samples contained interfering heterophilic antibodies that could be neutralized by addition of bovine IgG and heat-aggregated MAK33 to the assay buffer in the T84.66 [F(ab').sub.2] fragment assay. Of these eight samples, seven samples gave normal assay results when scFv T84.66 was used as the capture antibody even without buffer additives, whereas one sample (135 in Fig. 4) needed the addition of heat-aggregated MAK33 and bovine IgG.


One sample, referred to as 797, displayed a high apparent concentration (18.4 [micro]L) and nonlinear sample dilution, indicating interference in the T84.66 [F(ab').sub.2] fragment assay even after addition of heat-aggregated MAK33 to the assay buffer. The T84.66 scFv assay gave a normal assay result (2.5 [micro]L) for this sample even without buffer additives. And finally, the last sample, referred to as 233 (Fig. 4), gave a much lower result (71.6 [micro]L) in the T84.66 scFv assay than in the T84.66 [F(ab').sub.2] fragment assay (170.5 [micro]L). Dilution of sample 233 gave nonlinear results in the T84.66 scFv assay, indicating in this case a falsely low value that could not be corrected by buffer additives.

We compared our results with results from a homogeneous time-resolved fluorescence resonance energy transfer (Kryptor) assay for CEA (Brahms) for samples 797 (2.0 [micro]L) and 233 (209 [micro]L). If we assume that the Kryptor technology used in the Brahms assay gives the "true value", sample 233 is our only example of a discrepant result obtained with the scFv assay with buffer additives.


Assay interference attributable to the presence of heterophilic antibodies, or the threat of such interference, will probably always be with us (32). However, a growing awareness of this problem has prompted the reformulation or redesign of most commercial and in-house immunoassays. At present, there is no simple technique that can completely eliminate interference from heterophilic antibodies. The most common practice is to include blocking agents such as nonspecific immunoglobulins in the assay buffer (33). An alternative or complementary approach is to redesign the assay, using Fab or [F(ab').sub.2] immunoglobulin fragments as capture or tracer antibodies (21, 22). This approach ameliorates heterophilic antibody interference caused by antibodies with specificity for the Fc portion of the immunoassay reagents. In a recent study, we described our attempts to improve the performance of an in-house two-site immunometric assay for CEA performed in streptavidin-coated microplates (20). Our observations indicated that use of a biotinylated [F(ab').sub.2] capture reagent and the inclusion of a heat-aggregated irrelevant mouse Mab in the assay buffer provided optimum assay protection. We now routinely use this format for all our in-house immunometric assays.

The preparation of biotinylated [F(ab').sub.2] immunoglobulin fragments is both time-consuming and technically demanding. In addition, for certain Mabs, the preparation of proteolytic fragments in adequate yields is not always straightforward and is occasionally impossible. In this regard, the development of techniques to genetically engineer and produce recombinant antibodies has permitted their evaluation as alternative reagents in diagnostic immunoassays. A human/mouse chimeric antibody has been used previously as a tracer in a two-site enzyme immunoassay for CEA (34). This reagent was produced by fusion of [V.sub.H] and [V.sub.L] genes from a murme hybridoma to the constant region heavy and light chain genes derived from a human myeloma cell line. The authors indicated that in an unprotected CEA assay, the false-positive rate was reduced from 29% to 2.8% when the murine monoclonal was substituted with the engineered chimeric antibody. Addition of 1% normal mouse serum to the buffer used in both assays reduced the false-positive rate to 6.2% and 2.6%, respectively. This indicated that use of the chimeric tracer antibody was more effective at reducing heterophilic antibody interference than the addition of blocking agent to an assay based on two conventional Mabs.

In the present study, we investigated the incidence of heterophilic antibody interference in a novel CEA assay that uses a scFv as the solid-phase capture reagent. This antibody format has gained popularity because it has several potential advantages over recombinant whole-antibody reagents. These include the relatively simple methodologies needed for their construction and the ability to produce protein rapidly and inexpensively in E. coli. An additional, theoretical advantage, which was tested in this study, is that the scFv format does not occur naturally and hence may display fewer epitopes reactive with heterophilic antibodies. We chose to use a scFv version of the anti-CEA Mab T84.66 because this antibody has been well characterized (24), possesses a high affinity, and has previously been produced in a variety of recombinant antibody formats (25, 35, 36).

Biotinylation of single-chain antibodies for use as solidphase reagents can potentially be problematic because of their relatively small size. For example, random chemical biotinylation can lead to either destruction of the antigen-binding site or the production of antibody fragments that bind to streptavidin in an orientation that is sterically unfavorable for antigen binding. Furthermore, the need to use low degrees of substitution typically leads to a heterogeneous mixture of biotinylated scFvs, causing problems with batch-to-batch reproducibility. In an attempt to ameliorate these potential problems, we chose to site-specifically biotinylate our capture scFv through a 17-amino acid peptide translationally fused to the carboxy terminus of the recombinant antibody fragment. Unlike random chemical biotinylation techniques, this methodology guarantees not only that each antibody fragment is labeled with only one biotin molecule but also that this residue is placed distally to the antigen-binding site.

In E. coli, the only biotinylated protein is the biotin carboxyl carrier protein (BCCP), a subunit of acetyl-CoA carboxylase. BCCP is biotinylated by BirA, which modifies a specific lysine residue in the carboxyl biotin acceptor domain of the molecule. Using a directed evolution approach, Schatz (30) identified a series of small 14- to 23-amino acid peptides that mimic the BCCP acceptor region and are themselves substrates for biotinylation at a specific lysine residue when expressed in E. coli.

We constructed a plasmid containing the coding sequence for a biotin acceptor peptide downstream of the T84.66 scFv [V.sub.L] [([G.sub.4]S).sub.4]-[V.sub.H] cassette and expressed the fusion protein in an E. coli strain that overproduces BirA. The in vivo-biotinylated scFv was purified from periplasmic lysates by monovalent-avidin chromatography. Using this approach, we were able to produce milligram quantities of uniformly biotinylated T84.66 single-chain antibody in standard laboratory shake-flask cultures.

Recombinant antibody fragments such as the scFv represent a step forward in the development of reagents for immunoassays. They have the potential to perform as highly specific binders without the biological limitations inherent to intact antibody molecules, i.e., they do not activate complement and lack many of the epitopes for common autoantibodies. In our comparison of assays using whole IgG, [F(ab').sub.2] fragments, or scFv as capture reagents, we found that the number of binding sites needed was similar for the three molecules when they were presented on a streptavidin-coated surface. This implies that the lower affinity of the scFv binding site (23) is compensated by the uniform and specific orientation of the molecule on the streptavidin surface. Further studies using T84.66 scFv in competitive immunoassays, in which affinity is more critical, will be needed to further evaluate the properties of this reagent.

In conclusion, in light of the present study, our assay format using [F(ab').sub.2] fragments combined with buffer containing aggregated mouse IgG is no longer the gold standard, and we intend to convert our other immunoassays to ones based on site-specific biotinylated scFv molecules. Other solutions, such as antbodies or aptamers, might also be novel alternatives; however, at present, the in vivo biotinylated scFv format is attractive because it combines the best properties of traditional antibody reagents, the ability to be immobilized in a defined orientation, and an apparent low reactivity to heterophilic antibodies.

We thank Prof. Bent Norgaard-Pedersen, Department of Clinical Biochemistry, National Serum Institute (Copenhagen, Denmark), for measuring the CEA concentration in two samples with the Kryptor CEA assay.

Received December 17, 2004; accepted February 1, 2005.

Previously published online at DOI: 10.1373/clinchem.2004.046979


(1.) Kricka U. Human anti-animal antibody interferences in immunological assays. Clin Chem 1999;45:942-56.

(2.) Levinson SS, Miller JJ. Towards a better understanding of heterophile (and the like) antibody interference with modern immunoassays. Clin Chim Acta 2002;325:1-15.

(3.) Kaplan IV, Levinson SS. When is a heterophile antibody not a heterophile antibody? When it is an antibody against a specific immunogen. Clin Chem 1999;45:616-8.

(4.) Hennig C, Rink L, Kirchner H. Evidence for presence of IgG4 anti-immunoglobulin autoantibodies in all human beings. Lancet 2000;355:1617-8.

(5.) Hennig C, Rink L, Fagin U, Jabs WJ, Kirchner H. The influence of naturally occurring heterophilic anti-immunoglobulin antibodies on direct measurement of serum proteins using sandwich ELISAs. J Immunol Methods 2000;235:71-80.

(6.) Wickus GG, Caplan RH, Mathews EA, Pehling GB. Sudden appearance and subsequent disappearance of interference in immunometric assays of thyrotropin neutralizable with purified mouse IgG. Clin Chem 1991;37:595-6.

(7.) Kazmierczak SC, Catrou PG, Briley KP. Transient nature of interference effects from heterophile antibodies: examples of interference with cardiac marker measurements. Clin Chem Lab Med 2000;38:33-9.

(8.) Hurez V, Kaveri SV, Kazatchkine MD. Expression and control of the natural autoreactive IgG repertoire in normal human serum. Eur J Immunol 1993;23:783-9.

(9.) Bouvet JP, Dighiero G. From natural polyreactive autoantibodies to a la carte monoreactive antibodies to infectious agents: is it a small world after all? Infect Immun 1998;66:1-4.

(10.) Hunter WM, Budd PS. Circulating antibodies to ovine and bovine immunoglobulin in healthy subjects: a hazard for immunoassays. Lancet 1980;2:1136.

(11.) Hawkins BR, Saueracker GC, Dawkins RL, Davey MG, O'Connor KJ. Population study of heterophile antibodies. Vox Sang 1980; 39:339-42.

(12.) Ward G, McKinnon L, Badrick T, Hickman PE. Heterophilic antibodies remain a problem for the immunoassay laboratory. Am J Clin Pathol 1997;108:417-21.

(13.) Thompson RJ, Jackson AP, Langlois N. Circulating antibodies to mouse monoclonal immunoglobulins in normal subjects-incidence, species specificity, and effects on a two-site assay for creatine kinase-MB isoenzyme. Clin Chem 1986;32:476-81.

(14.) Ismail AA, Walker PL, Barth JH, Lewandowski KC, Jones R, Burr WA. Wrong biochemistry results: two case reports and observational study in 5310 patients on potentially misleading thyroid stimulating hormone and gonadotropin immunoassay results. Clin Chem 2002;48:2023-9.

(15.) Ismail AA, Walker PL, Cawood ML, Barth JH. Interference in immunoassay is an underestimated problem. Ann Clin Biochem 2002;39:366-73.

(16.) Boscato LM, Stuart MC. Incidence and specificity of interference in two-site immunoassays. Clin Chem 1986;32:1491-5.

(17.) Frengen J, Kierulf B, Schmid R, Lindmo T, Nustad K. Demonstration and minimization of serum interference in flow cytometric two-site immunoassays. Clin Chem 1994;40:420-5.

(18.) Despres N, Grant AM. Antibody interference in thyroid assays: a potential for clinical misinformation. Clin Chem 1998;44:44054.

(19.) Muller W, Mierau R, Wohltmann D. Interference of IgM rheumatoid factor with nephelometric C-reactive protein determinations. J Immunol Methods 1985;80:77-90.

(20.) Bjerner J, Nustad K, Norum LF, Olsen KH, Bormer OP. Immunometric assay interference: incidence and prevention. Clin Chem 2002;48:613-21.

(21.) Vaidya HC, Beatty BG. Eliminating interference from heterophilic antibodies in a two-site immunoassay for creatine kinase MB by using [F(ab').sub.2] conjugate and polyclonal mouse IgG. Clin Chem 1992;38:1737-42.

(22.) Csako G, Weintraub BD, Zweig MH. The potency of immunoglobulin G fragments for inhibition of interference caused by antiimmunoglobulin antibodies in a monoclonal immunoradiometric assay for thyrotropin. Clin Chem 1988;34:1481-3.

(23.) Krebber A, Bornhauser S, Burmester J, Honegger A, Willuda J, Bosshard HR, et al. Reliable cloning of functional antibody variable domains from hybridomas and spleen cell repertoires em ploying a reengineered phage display system. J Immunol Methods 1997;201:35-55.

(24.) Wagener C, Yang YH, Crawford FG, Shively JE. Monoclonal antibodies for carcinoembryonic antigen and related antigens as a model system: a systematic approach for the determination of epitope specificities of monoclonal antibodies. J Immunol 1983; 130:2308-15.

(25.) Rodenburg CM, Mernaugh R, Bilbao G, Khazaeli M B. Production of a single-chain anti-CEA antibody from the hybridoma cell line T84.66 using a modified colony-lift selection procedure to detect antigen-positive ScFv bacterial clones. Hybridoma 1998;17:1-8.

(26.) Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Anal Biochem 1976;72:248-54.

(27.) Bormer OP, Nustad K. Selection of monoclonal antibodies for use in an immunometric assay for carcinoembryonic antigen. J Immunol Methods 1990;127:171-8.

(28.) Bormer 0. A direct assay for carcinoembryonic antigen in serum and its diagnostic value in metastatic breast cancer. Clin Biochem 1982;15:128-32.

(29.) Paus E, Almasbak H, Bormer OP, Warren DJ. A single-chain-Fvbased immunofluorometric assay specific for the CEA variant NCA-2. J Immunol Methods 2003;283:125-39.

(30.) Schatz PJ. Use of peptide libraries to map the substrate specificity of a peptide-modifying enzyme: a 13 residue consensus peptide specifies biotinylation in Escherichia coli. Biotechnology 1993;11: 1138-43.

(31.) Henrikson KP, Allen SH, Maloy WL. An avidin monomer affinity column for the purification of biotin-containing enzymes. Anal Biochem 1979;94:366-70.

(32.) Bjerner J, Bormer OP, Nustad K. The war on heterophilic antibody interference [Editorial]. Clin Chem 2005;51:9-11.

(33.) Boscato LM, Stuart MC. Heterophilic antibodies: a problem for all immunoassays. Clin Chem 1988;34:27-33.

(34.) Kuroki M, Matsumoto Y, Arakawa F, Haruno M, Murakami M, Kuwahara M, et al. Reducing interference from heterophilic antibodies in a two-site immunoassay for carcinoembryonic antigen (CEA) by using a human/mouse chimeric antibody to CEA as the tracer. J Immunol Methods 1995;180:81-91.

(35.) Hefta U, Wu AM, Neumaier M, Shively JE. Measuring antibody affinity using biosensors. In: McCafferty J, Hoogenboom HR, Chiswell DJ, eds. Antibody engineering: a practical approach. New York: Oxford University Press, 1996:98-117.

(36.) Shibaguchi H, Kuroki M, Kuroki M, Badran A, Hachimine K, Kinugasa T. Cloning and sequencing of variable region cDNAs of a novel human monoclonal antibody to carcinoembryonic antigen, and generation of a single chain variable fragmented antibody. Anticancer Res 2004;24:3355-60.

DAVID J. WARREN, * JOHAN BJERNER, ELISABETH PAUS, OLE P. BORMER, and KJELL NUSTAD Central Laboratory, The Norwegian Radium Hospital, Oslo, Norway.

[1] Nonstandard abbreviations: HAMA, human anti-mouse antibody; CEA, carcinoembryonic antigen; Mab, monoclonal antibody; scFv, single-chain fragment of variation; BSA, bovine serum albumin; PBS, phosphate-buffered saline; BirA, biofin holoenzyme synthetase; and BCCP, biofin carboxyl carrier protein.

* Address correspondence to this author at: Central Laboratory, The Norwegian Radium Hospital, N-0310 Oslo, Norway. Fax 47-22-73-07-25; e-mall
Table 1. Samples showing interference included in the study. (a)

 Frequency in
Type No. of nonimmunized
 patients population, (b) %

I Results normalized after 0 0.08 (0-0.15)
 removal of Fc fragment
 or after addition of 15
 mg/L native MAK33 or 15
 mg/L heat-aggregated

II Results normalized after 269 3.0 (2.5-3.6)
 removal of Fc fragment
 or after addition of 15
 mg/L heat-aggregated
 MAK33 but not after
 addition of 15 mg/L
 native MAK33

III Results normalized after 62 0.82 (0.58-1.09)
 removal of Fc fragment
 but not after addition
 of 15 mg/L
 heat-aggregated MAK33

IV Results normalized 4 0.06 (0-0.13)
 after addition of 15
 mg/L heat-aggregated
 MAK33 but not after
 removal of Fc fragment

V Results normalized 7 0.020 (0-0.062)
 only after both
 removal of Fc fragment
 and addition of15 mg/L
 heat-aggregated MAK33

VI Higher results in the 2 0.020 (0-0.062)
 original F[(ab').sub.2]
 assay than in the
 whole-IgG assay. Sample
 dilution indicates
 interference still
 after both removal of
 Fc fragment and
 addition of 15 mg/L
 heat-aggregated MAK33
 Not classified in previous 46

Total 390

(a) Frequencies and classifications from Bjerner et al. (20).

(b) Values in parentheses are the 95% confidence interval.
COPYRIGHT 2005 American Association for Clinical Chemistry, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Proteomics and Protein Markers
Author:Warren, David J.; Bjerner, Johan; Paus, Elizabeth; Bormer, Ole P.; Nustad, Kjell
Publication:Clinical Chemistry
Date:May 1, 2005
Previous Article:D-dimer concentrations in normal pregnancy: new diagnostic thresholds are needed.
Next Article:Negative interference in cardiac troponin I immunoassays by circulating troponin autoantibodies.

Related Articles
Development of a rapid and sensitive immunofluorometric assay for glutathione S-transferase A.
Streptavidin-polyvinylamine conjugates labeled with a europium chelate: applications in immunoassay, immunohistochemistry, and microarrays.
Dual-label time-resolved immunofluorometric assay of free and total prostate-specific antigen based on recombinant Fab fragments.
Automated assay for HER-2/neu in serum.
Human anti-animal antibody interferences in immunological assays.
Immunometric assay interference: incidence and prevention.
Miniature single-particle immunoassay for prostate-specific antigen in serum using recombinant fab fragments.
Interferences with two-site immunoassays by human anti-mouse antibodies formed by patients treated with monoclonal antibodies: comparison of...
Multicenter evaluation of an automated assay for troponin I.
Point-of-care time-resolved immunofluorometric assay for human pregnancy-associated plasma protein a: use in first-trimester screening for down...

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters